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
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-161586970510.1186/1477-7827-3-16ResearchLepidium meyenii (Maca) increases litter size in normal adult female mice Ruiz-Luna Ana C [email protected] Stephanie [email protected] Norma J [email protected] Julio [email protected] Manuel [email protected] Gustavo F [email protected] Instituto de Investigaciones de la Altura, Universidad Peruana Cayetano Heredia, Lima, Perú2 Department of Biological and Physiological Sciences, Faculty of Sciences and Philosophy, Universidad Peruana Cayetano Heredia, Lima, Perú2005 3 5 2005 3 16 16 4 11 2004 3 5 2005 Copyright © 2005 Ruiz-Luna et al; licensee BioMed Central Ltd.2005Ruiz-Luna 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
Lepidium meyenii, known as Maca, grows exclusively in the Peruvian Andes over 4000 m altitude. It has been used traditionally to increase fertility. Previous scientific studies have demonstrated that Maca increases spermatogenesis and epididymal sperm count. The present study was aimed to investigate the effects of Maca on several fertility parameters of female mice at reproductive age.
Methods
Adult female Balb/C mice were divided at random into three main groups: i) Reproductive indexes group, ii) Implantation sites group and iii) Assessment of uterine weight in ovariectomized mice. Animals received an aqueous extract of lyophilized Yellow Maca (1 g/Kg BW) or vehicle orally as treatment. In the fertility indexes study, animals received the treatment before, during and after gestation. The fertility index, gestation index, post-natal viability index, weaning viability index and sex ratio were calculated. Sexual maturation was evaluated in the female pups by the vaginal opening (VO) day. In the implantation study, females were checked for implantation sites at gestation day 7 and the embryos were counted. In ovariectomized mice, the uterine weight was recorded at the end of treatment.
Results
Implantation sites were similar in mice treated with Maca and in controls. All reproductive indexes were similar in both groups of treatment. The number of pups per dam at birth and at postnatal day 4 was significantly higher in the group treated with Maca. VO day occurred earlier as litter size was smaller. Maca did not affect VO day. In ovariectomized mice, the treatment with Maca increased significantly the uterine weights in comparison to their respective control group.
Conclusion
Administration of aqueous extract of Yellow Maca to adult female mice increases the litter size. Moreover, this treatment increases the uterine weight in ovariectomized animals. Our study confirms for the first time some of the traditional uses of Maca to enhance female fertility.
==== Body
Background
Maca, is a cruciferous plant that grows exclusively between 4,000 and 4,500 m altitude in the Central Peruvian Andes [1]. The hypocotyl-root axis is the edible part of the plant, which is dried and can be stored in this way for years. This hypocotyl has been used for centuries as nutrient and by its fertility-enhancer property [2-4].
Scientific studies in male rats and mice have demonstrated that oral administration of aqueous or ethanolic extracts of Maca hypocotyls improved spermatogenesis [1,5-7]. Moreover, an improvement in sperm count was observed in men after four months of treatment [8].
Studies reported in the peer-reviewed scientific literature about the effects of this native plant in the female reproductive system are scarce. The only scientific evidence reported showed that Maca did not modify the rate of implantation. In that study, Maca was mixed in the food for 30 days [9]. For such reason, the dose administered was not controlled.
The present study was aimed to investigate the effects of an aqueous extract of Maca in female mice at reproductive age through several female reproductive indexes, uterine weight measurements, number of implanted embryos and vaginal opening (VO) day of the female pups with a fixed dose of Maca administered to the dams.
Methods
Animals
Three-month-old virgin female mice of the Balb/C strain obtained from the Animal House at the National Institute of Health (NIH) in Peru were used for this study. Mice were divided at random into three main groups: i) The first group was used to assess reproductive indexes; ii) the second group was aimed to study implantation sites and iii) the last group was aimed to evaluate the effect of Maca on uterine weight in ovariectomized mice.
Each of the three main groups was divided into two subgroups: the control group, which received 0.5 ml distilled water (vehicle) and the treated group, which received lyophilized aqueous extract of Yellow Maca orally by gavage. This study was conducted in accordance with Universidad Peruana Cayetano Heredia Guidelines and Ethics on Animal Experimentation.
Preparation of the Aqueous Extract of Maca
Dried Maca was provided by Santa Natura Company. The ecotype used for this study was Yellow Maca. Dried Maca hypocotyls were crushed in a mill, diluted 1:3 w/v in distilled water and boiled for approximately 2 hours. Afterwards, they were filtered and let at room temperature for cooling. Aqueous extract was freezed under -70°C for three days and then lyophilized. One gram of lyophilized Maca is equivalent to 2.19 g of dried Maca hypocotyls. Lyophilized Maca was diluted in distilled water to obtain a dose of 1 g/kg BW. This dose has been proved to be optimal in a dose-dependent study [10].
Implantation sites
Female mice were treated with lyophilized aqueous extract of Yellow Maca (1 g/Kg) or distilled water (vehicle) by oral route for 22–28 days (15 days prior mating and during the first seven days of gestation period). At day 15 of treatment with Maca or vehicle, each female mouse was mated with a single male of proven fertility. The presence of a vaginal plug was considered the first day of pregnancy. The females were killed by cervical dislocation at day 7 of gestation and the uterus was checked for implantation sites. The embryos were counted and weighed.
Reproductive indexes
Female mice were administered orally by gavage with distilled water (vehicle) or lyophilized aqueous extract of Maca (1 g/kg body weight) for 15 days prior to mating, during the whole period of gestation and 21 days after birth (lactation period). Body weight was recorded daily.
At day 15 of treatment with vehicle or Maca, each female mouse was housed independently of the estrous cycle with a sexually mature untreated male of proven fertility. The presence of a copulation plug in the vagina was regarded as successful copulation and was considered the first day of pregnancy. For pregnancy confirmation, mice were examined by vaginal smear to determine the presence of diestrus stage cells during a 7 day-period after the finding of a vaginal plug.
Upon delivery, the number of pups was counted. Both groups of dams continued their respective treatments during the whole lactation period (21 days). The number of pups alive at postnatal day 4 and 21, the VO day of the female pups and the sex ratio of both groups were also determined (male fraction).
The following reproductive indexes were calculated [11]: fertility index, defined as No pregnant females/No females with successful copulation × 100; gestation index, defined as No of females with alive pups/ No of pregnant females × 100; post-natal viability index, defined as No of pups alive on day 4 / No of alive pups × 100; and weaning viability index, defined as No of pups alive at day 21/ No of pups alive at day 4 × 100.
Ovariectomized mice
In order to assess a possible estrogenic role of Maca on uterine weight, 3 month-old mice were ovariectomized. Animals were shaved in the dorsal surgery area and anesthetized with ketamide (40 mg/kg body weight i.p). A dorsal incision was performed and the ovaries were exposed and removed. Three months after surgery, 15 animals were exposed orally by gavage to distilled water (n = 8; 0.5 ml distilled water) or Yellow Maca (n = 7; 0.5 g/kg body weight) for a 42-day period. At the end of the treatment, animals were sacrificed by cervical dislocation. Body and uterine wet weights were recorded.
Statistical analysis
Prior to the analysis, normality of the data was determined using the one sample Kolmogorov-Smirnov Test. Body weight, implantation sites, gestation length and litter size variables were analyzed using Student's t test. Multivariate analysis was used to analyze the influence of the Maca treatment on the number of pups and on the VO day. Reproductive indexes were analyzed using the Paired Proportion Test. Data are presented as mean ± S.E.M. Data without normal distribution, such as the vaginal opening day and the uterine weight were analyzed using the Mann Whitney U Test. Groups were considered significantly different if P < 0.05. All calculations and statistical analysis were generated in STATA v 8.0 (Stata Corporation, College Station, TX, USA).
Results
Indexes
In both Maca treated and control groups, the fertility index, the gestation index, the post-natal viability index and the weaning viability index were similar between the control and the Maca treated group (Table 1).
Table 1 Reproductive indexes in female mice receiving Maca (1 g/kg) distilled water
Indexes (%) Distilled water (n = 12) Maca (n = 11)
Fertility indexa 100.0 100.0
Gestation indexa 100.0 100.0
Post-natal viability indexa 94.3 97.1
Weaning viability indexa 96.3 96.0
a No statistical difference was found between both groups (P > 0.05).
Body Weight
In both the implantation study and the reproductive indexes study, the body weight did not differ significantly between control and Maca-treated groups. In the implantation study, at day 1, body weights were 34.68 ± 1.12 g in the control group and 34.13 ± 0.62 g in the Maca-treated group. After 17 days of treatment with lyophilized aqueous extract of Yellow Maca, body weight was similar to that observed in the control group (control group, 33.43 ± 1.18 g vs. Maca group, 34.98 ± 0.72 g). Body weight change between day 1 and day 17 of treatment was similar in the Maca treated animals and in controls (0.85 ± 0.30 g vs. -1.25 ± 1.17 g, respectively).
In the reproductive indexes study, body weight at day 1 of treatment with distilled water (control group) was 29.4 ± 0.57 g and with Maca was 29.94 ± 0.82 g. At day 15 of treatment, body weight of the control group was 30.77 ± 0.61 g, while the group treated with Maca showed a body weight of 32.07 ± 1.27 g. No differences between groups of treatment at day 1 or day 15 were observed.
Implantation sites
The implantation sites were similar in the Maca treated group and the controls (11.56 ± 2.13 vs. 9.67 ± 2.42, respectively).
Gestation length
The gestation length was similar in the Maca-treated group and the control group (19.88 ± 0.11 days and 19.92 ± 0.08 days, respectively).
Pups/ Litter size
Mean number of pups at birth and at postnatal day 4 in the Maca treated group was higher than their respective controls (P < 0.05) (Table 2). Multivariate analysis showed that treatment with Maca increase number of pups (3.18 ± 1.04; Coefficient of regression ± SE; P < 0.05).
Table 2 Effects of Maca (1 g/kg BW) on litter size, vaginal opening day and sex ratio.
Distilled water Maca
Litter size (n = 12) (n = 10)
Day 1 (birth) 7.25 ± 1.14 10.40 ± 1.32a
Day 4 6.83 ± 1.28 10.10 ± 1.30a
Day 21 6.58 ± 1.33 9.70 ± 1.30
Vaginal opening day (days) (n = 37) (n = 52)
Litter < 7 pups 26.1 ± 0.48 23.2 ± 0.14‡‡
Litter ≥ 7 pups 27.5 ± 0.27* 28.40 ± 0.46***
Sex ratio (male fraction) 0.54 0.50
aP < 0.05 with respect to distilled water group.
*P < 0.05, ***P < 0.001 with respect to values in the group with litter size below 7. ‡‡P < 0.01 with respect to distilled water group.
Sex ratio
The fraction of male pups in the control group and in the treatment group were not statistically different between Control and Maca group (0.54 and 0.50, respectively) (Table 2).
Vaginal opening
In order to assess the effect of Maca on female sexual maturation, the vaginal opening (VO) day was evaluated. During postnatal development of pups, it was noticed that those belonging to a bigger litter size developed and grew slower in size and weight in comparison to the smaller litters (data not shown). Statistical analysis showed that the VO day was indeed influenced by the litter size. Therefore, the litters were divided in two subgroups: litters with equal or more than seven pups and litters with less than seven pups. Litters with <7 pups showed lower VO day than litters with litters ≥ 7 (P < 0.01). Maca treated group produced an earlier VO day in litters with less than 7 pups than in the controls (P < 0.01). In litters with ≥ 7 pups, the VO day was similar in the Maca treated group and in the control group. The multivariate analysis showed that litter size but not treatment was related to VO day (R2 = 0.37; P < 0.01).
Ovariectomized mice
Oral administration of lyophilized aqueous extract of Yellow Maca for 42 days significantly increased uterine wet weight (65.56 ± 16.62 mg; mean ± SEM) when compared to their respective controls (29.98 ± 3.05 mg) (P < 0.05).
Discussion
Maca has been described to improve fertility since many centuries ago [3]. However, scientific evidences in peer reviewed journals were available since 2000 [5-8,12-14]. The studies demonstrated that Maca increases male sexual behavior [11-13] and increases sperm production [8].
In relation to female reproduction, there is only one study demonstrating that Maca did not increase implantation rate in mice [9]. We have demonstrated in the present study that implantation rate was not increased in mice after oral administration of 1 g/Kg BW of a lyophilized aqueous extract of Maca. However, interestingly, female mice that received Maca delivered more pups than the control group. This suggests that Maca may have a protective effect on the number of resorptions, having a less occurrence of them between day 7 of pregnancy and the subsequent days of pregnancy in the Maca treated group. Supplementation of Maca in diets improved growth rates and survival of rainbow trout Oncorhynchus mykiss (Walbaum) alevins and juveniles [15]. Some possible mechanisms through which Maca may act include increased uterine receptivity, altered immune functions and effects on vascular system [16]. It is known that functional lymphocytic progesterone binding sites are needed for the maintenance of normal pregnancy, and that progesterone-mediated immunosuppression is needed for the maintenance of normal gestation [17].
Litter size was higher at birth and at postnatal day 4 in the groups treated with Maca than in the controls. This effect seems to be due to a reduction in embryo resorption, since number of implantation sites were similar in mice treated with Maca than in controls. This effect seems not to be due to any estrogenic activity of Maca, since exposure to estradiol-17 beta or the proestrogen methoxychlor reduced litter size [18].
Sex ratio was not affected after treatment with Maca. In fish, the period of sex differentiation is the most sensitive to possible action of phytochemicals with steroid activity. However, there was no significant difference in sex ratios between the control and Maca treated group of rainbow trout. P-nonyphenol, an environmental toxicant with estrogenic properties did not affect the sex ratio of live pups [19]. All of these suggest that main effects of Maca were not due to increased estrogenic activity.
Sexual maturation in the litters was evaluated by the vaginal opening day. The pups were divided in two groups (< 7 pups and ≥ 7 pups). Fraternity size has a positive linear effect on age at vaginal opening. Lower litter size was related to early age at vaginal opening day. This has been observed previously [20]. This difference seems to be due to differences in dietary energy density [21]. When comparing both sub-groups with their respective controls, we observed that Maca had not effect on the VO day. Dietary phytoestrogens has been shown to accelerate the time of vaginal opening in immature CD-1 mice [22]. Since Maca does not affect the VO day, it is suggested that phytoestrogens present in Maca has little impact on reproductive parameters assessed in the present study.
Maca contains sterols, such as campesterols (27.3%), ergosterol (13.6%), brassicasterol (9.1%) Δ7,22-ergostadienol (4.5%) and sitosterol (45.5 %) [23], the latter having phytoestrogen activity. It has been suggested that β-sitosterol could cross the placenta [24]. In humans, it has been demonstrated that sitosterol can be excreted in milk and therefore, enter the neonate [25]. Maca chemical composition includes sitosterol, this compound would exert its effect on pups.
Treatment with Maca also increases male sexual behavior [12-14] and sperm production in mice and rats [7]. It is probable that main effects of Maca increasing uterine weight and litter size were not due to an estrogenic effect but to a progestin-like one, since Maca chemical composition include other sterols besides phytoestrogen sitosterol. Progesterone would certainly affect both, uterine weight via progesterone receptors, as well as prevent abortion. In mice and rats, progesterone is much more important in maintaining pregnancy than estrogen [16,17]. Moreover, physiological level of progesterone acts in conjunction with androgens to facilitate copulatory behavior in male rats and mice [26,27]. In addition, one of the most potent phytoestrogenic substance -6-(1,1-dimethylallyl) naringenin (6-DMAN)- did not have any effect on uterine wet weight in ovariectomized rats [28]. Our study demonstrates that Maca increases uterine weight. It is suggested that this activity may be produced by a progestin-like effect. However, more studies need to be performed to clarify this proposed mechanism.
Conclusion
To sum up, our results suggest that lyophilized aqueous extract of Yellow Maca increase the number of pups in normal adult female mice.
Authors' contributions
AR conceived of the study, participated in its design, coordination, execution, analysis and interpretation of the data, and drafted the manuscript.
SS participated in the design, coordination and execution of the experiment, analysis of the and interpretation of the data, and helped draft the manuscript of the study.
NA participated in the development of the experiment and revised the draft critically.
JR participated in the development of the experiment, carried out the analysis and interpretation of the data, and revised the draft critically.
MG participated in the development of the experiment.
GG participated in the analysis and interpretation of the data, helped draft the manuscript, revised it critically and gave the final approval of the version.
Acknowledgements
The authors would like to thank Sandra Yucra and Jessica Nieto for their technical support. Vicerrectorate of Investigation of Universidad Peruana Cayetano Heredia financially supported this study.
==== Refs
Gonzales GF Gasco M Cordova A Chung A Rubio J Villegas L Effect of Lepidium meyenii (Maca) on spermatogenesis in male rats acutely exposed to high altitude (4340 m) J Endocrinol 2004 180 87 95 14709147 10.1677/joe.0.1800087
Obregón LE Maca 1998 Lima; Instituto de Fitoterapia Americano
Cobo B Francisco de Mateos History of the New World 1956 Madrid: Ediciones Atlas
León J The "Maca" (Lepidium meyenii), a little-known food plant of Peru Econ Bot 1964 18 122 127
Gonzales GF Ruiz A Gonzales C Villegas L Cordova A Effect of Lepidium meyenii (Maca) roots on spermatogenesis of male rats Asian J Androl 2001 3 231 233 11561196
Gonzales GF Rubio J Chung A Gasco M Villegas L Effect of alcoholic extract of Lepidium meyenii (Maca) on testicular function in male rats Asian J Androl 2003 5 349 352 14695987
Bustos-Obregón E Yucra S Gonzales GF Lepidium meyenii (Maca) reduced spermatogenesis damage induced by a single dose of malathion in mice Asian J Androl 2005 7 71 76 15685355
Gonzales GF Cordova A Gonzales C Chung A Vega K Villena A Lepidium meyenii (Maca) improved semen parameters in adult men Asian J Androl 2001 3 301 303 11753476
Oshima M Gu Y Tsukada S Effects of Lepidium meyenii walp and Jatropha macrantha on blood levels of estradiol 17-β, progesterone, testosterone and the rate of embryo implantation in mice J Vet Med Sci 2003 65 1145 1146 14600359 10.1292/jvms.65.1145
Chung A Rubio J Gonzales C Gasco M Gonzales GF Dose-response effects of Lepidium meyenii (Maca) aqueous extract on testicular function and weight of different organs in adult rats J Ethnopharmacol 2005 98 143 147 15763375 10.1016/j.jep.2005.01.028
Ratnasooriya WD Jayakody JRAC Premakumara GAS Adverse pregnancy outcome in rats following exposure to a Salacia reticulata (Celastracea) root extract Braz J Med Biol Res 2003 36 931 935 12845381 10.1590/S0100-879X2003000700015
Zheng BL He K Kim CH Rogers L Shao Y Huang ZY Lu Y Yan SJ Qien LC Zheng QY Effect of a lipidic extract from Lepidium meyenii on sexual behavior in mice and rats Urology 2000 55 598 602 10736519 10.1016/S0090-4295(99)00549-X
Cicero AF Bandieri E Arletti R Lepidium meyenii Walp. improves sexual behaviour in male rats independently from its action on spontaneous locomotor activity J Ethnopharmacol 2001 75 225 229 11297856 10.1016/S0378-8741(01)00195-7
Cicero AF Piacente S Plaza A Sala E Arletti R Pizza C Hexanic Maca extract improves rat sexual performance more effectively than methanolic and chloroformic Maca extracts Andrologia 2002 34 177 179 12059814 10.1046/j.1439-0272.2002.00490.x
Lee K-J Dabrowski K Rinchard J Gomez C Guz L Vilchez C Supplementation of maca (Lepidium meyenii) tuber meal in diets improves growth rate and survival of rainbow trout Oncorhynchus mykiss (Walbaum) alevins and juveniles Aquac Res 2004 35 215 223 10.1111/j.1365-2109.2004.01022.x
Clark DA Quarrington C Banwatt D Manuel J Fupop G Spontaneous abortion in immunodeficient SCID mice Am J Reprod Immunol 1994 32 15 25 7945811
Szekeres-Bartho J Chaouat G Kinsky R A Progesterone-induced blocking factor corrects high resorption rates in mice treated with antiprogesterone Am J Obstet Gynecol 1990 163 1320 1322 2220944
Amstislavsky SY Kizilova EA Golubitsa AN Vasilkova AA Eroschenko VP Preimplantation exposures of murine embryos to estradiol or methoxychlor change postnatal development Reprod Toxicol 2004 18 103 108 15013069 10.1016/j.reprotox.2003.10.008
Kyselova V Peknicova J Buckiova D Boubelik M Effects of p-nonyphenol and resveratrol on body and organ weight and in vivo fertility of outbred CD-I mice Reprod Biol Endocrinol 2003 1 30 39 12749770 10.1186/1477-7827-1-30
Kirkpatrcik BW Rutledge JJ The influence of prenatal and postnatal fraternity size on reproduction in mice Biol Reprod 1987 36 907 914 3593855
Suzuki T Shimizu M Ishibashi T Effect of the energy density of non-purified diets on growth, gestation and lactation in mice Jikken Dobutsu 1991 40 305 314 1915597
Thigpen JE Haseman JK Saunders HE Setchel KD Grant MG Forsythe DB Dietary phytoestrogens accelerate the time of vaginal opening in immature CD-1 mice Comp Med 2003 53 607 615 14727808
Dini A Migliuolo G Rastrelli L Saturnino P Schettino O Chemical composition of Lepidium meyenii Food chemistry 1994 49 347 349 10.1016/0308-8146(94)90003-5
Doerge DR Churchwell MI Chang HC Newbold RR Delelos BK Placental transfer of soy isoflavone genistein following dietary and gavage administration to Sprague-Dawley rats Reprod Toxicol 2001 15 105 110 11297868 10.1016/S0890-6238(01)00108-3
Mellies MJ Ishikawa TT Glueck CJ Bove K Morrison J Phytosterols in aortic tissue in adults and infants J Lab Clin Med 1976 88 914 921 993644
Witt DM Young LJ Crews D Progesterone modulation of androgen-dependent sexual behavior in male rats Physiol Behav 1995 57 307 313 7716208 10.1016/0031-9384(94)00247-3
Witt DM Gao G Caldwell JD Testosterone and sexual experience alter levels of plasma membrane binding sites for progesterone in the male rat brain Horm Metab Res 2003 35 69 75 12734784 10.1055/s-2003-39060
Zierau O Geis RB Tischer S Schwab P Metz P Vollmer G Uterine effects of the phytoestrogen 6-(1,1-dimethylallyl) naringenin in rats Planta Med 2004 70 590 593 15254850 10.1055/s-2004-827178
| 15869705 | PMC1131923 | CC BY | 2021-01-04 16:37:13 | no | Reprod Biol Endocrinol. 2005 May 3; 3:16 | utf-8 | Reprod Biol Endocrinol | 2,005 | 10.1186/1477-7827-3-16 | oa_comm |
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-171587174710.1186/1477-7827-3-17ResearchOogenesis in cultures derived from adult human ovaries Bukovsky Antonin [email protected] Marta [email protected] Michael R [email protected] Laboratory of Development, Differentiation and Cancer, Department of Obstetrics and Gynecology, The University of Tennessee Graduate School of Medicine, Knoxville, Tennessee, USA2005 5 5 2005 3 17 17 17 2 2005 5 5 2005 Copyright © 2005 Bukovsky et al; licensee BioMed Central Ltd.2005Bukovsky 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.
Ten years ago, we reported that in adult human females the ovarian surface epithelium (OSE) is a source of germ cells. Recently, we also demonstrated that new primary follicles are formed by assembly of oocytes with nests of primitive granulosa cells in the ovarian cortex. The components of the new primary follicles, primitive granulosa and germ cells, differentiated sequentially from the OSE, which arises from cytokeratin positive mesenchymal progenitor cells residing in the ovarian tunica albuginea. In the present study, we investigated the possibility that the oocytes and granulosa cells may differentiate in cultures derived from adult human ovaries. Cells were scrapped from the surface of ovaries and cultured for 5 to 6 days, in the presence or absence of estrogenic stimuli [phenol red (PhR)]. The OSE cells cultured in the medium without PhR differentiated into small (15 micron) cells of granulosa phenotype, and epithelial, neural, and mesenchymal type cells. In contrast, OSE cells cultured in the presence of PhR differentiated directly into large (180 micron) cells of the oocyte phenotype. Such cells exhibited germinal vesicle breakdown, expulsion of the polar body, and surface expression of zona pellucida proteins, i.e. characteristics of secondary oocytes. These in vitro studies confirm our in vivo observations that in adult human ovaries, the OSE is a bipotent source of oocytes and granulosa cells. Development of numerous mature oocytes from adult ovarian stem cells in vitro offers new strategies for the egg preservation, IVF utilization, and treatment of female infertility. In addition, other clinical applications aiming to utilize stem cells, and basic stem cell research as well, may employ totipotent embryonic stem cells developing from fertilized oocytes.
==== Body
Background
The origin of oocytes (and primary follicles) in ovaries of adult mammalian females has been a matter of dispute for over one hundred years. The resulting dogma was that all oocytes in adult mammalian females persist from the fetal period of life [1]. However, from a phylogenetic viewpoint, it seems contradictory that mammalian females, including humans, would evolve a uniquely retrogressive reproductive mechanism, requiring preservation of their gametes from the fetal period for up to several decades. Such long lasting preservation could cause an accumulation of spontaneous or environmentally induced genetic alterations of oocytes in resting primary follicles. On the contrary, oogenesis has been demonstrated in cultured mouse embryonic stem cells [2], and mitotically active germ cells have been reported in ovaries of adult prosimian primates [3] and mice [4]. We have shown that mesenchymal cells in ovarian tunica albuginea (TA) differentiate into surface epithelium, a source of germ cells entering blood vessels and contributing to follicular renewal in adult human females [5,6]. These reports represent challenges to established dogma on the fetal origin of mammalian eggs [7,8].
Regarding follicular renewal in adult human females, our reports provide direct evidence that the ovarian surface epithelium (OSE) is a source of germ cells, and new primary follicles are formed by assembly of oocytes with nests of primitive granulosa cells in the ovarian cortex [5,6]. Components for the new primary follicles, primitive granulosa and germ cells, differentiate sequentially and de novo from mesenchymal progenitor cells residing in the ovarian TA. It appears that mesenchymal progenitor cells first contribute to the development of epithelial cells similar to granulosa cells, and these cells subsequently form epithelial nests descending into the deeper ovarian cortex. Oogenesis follows later. During this period, the mesenchymal progenitor cells differentiate into OSE cells with an embryonic character, lining either the ovarian surface or invaginated epithelial crypts. These cells are a source of germ cells, which assemble with nests of primitive granulosa cells to form primary follicles. Systemic hormonal concentrations may influence these processes. The data presented here indicate that granulosa cells are likely to develop during a period of low estrogenic stimuli (early and mid follicular phase). Differentiation of germ cells and oocytes may be associated with high estrogen concentrations (preovulatory period). The assembly of oocytes with granulosa cell nests (follicular renewal) may occur during the luteal phase [6]. Differentiation of primitive granulosa and germ cells from the bipotent mesenchymal cell precursors of TA in adult human ovaries possibly represents a sophisticated mechanism created during the evolution of female reproduction. This contrasts with the continuous preservation of germline stem cells seen in males [9], and female prosimian primates and mice [3,4].
In culture, the OSE cells undergo an epithelio-mesenchymal transition. These cells are initially cytokeratin (CK) positive, but lose CK expression with time and passage in culture [10]. However, upon introduction of E-cadherin, the mesenchymal type cells can differentiate backward into the epithelial phenotype [11]. Multipotential progenitor cells exhibiting a mesenchymal phenotype and capable of differentiation into distinct cell types have been derived from various other adult tissues [12-16]. Mesenchymal-epithelial and epithelial-mesenchymal transitions may reflect a plasticity of progenitor cells in a particular microenvironment. They may occur sequentially (mesenchymal-epithelial followed by epithelial-mesenchymal transitions) under the influence of the extracellular matrix, cytokines (transforming growth factor β, fibroblast growth factor, hepatocyte growth factor, epidermal growth factor, BMP2 and BMP4), adhesion molecules (integrins, E-cadherins), membrane receptors, intercellular junctions, signaling pathways (MAPK) or transcription factors (β-catenin), commonly produced in the embryo and less frequently in adult organisms. Such transitions are manifestations of cell plasticity and result in dramatic changes – in lineage commitments to certain cell types (reviewed in [17]).
In culture, in addition to their characteristic morphology, oocytes can also be identified by alkaline phosphatase activity, though nonspecific results have been described in various tissues [18]. However, zona pellucida (ZP) proteins and some ZP antigens, such as PS1 meiotically expressed carbohydrate antigen and heat-solubilized porcine zona (HSPZ) proteins, are more specific markers. In fact, they have been detected in the OSE cells of the rabbit, cat, monkey, baboon and human [6,19,20]. Hence, expression of ZP proteins in OSE cells suggests a relationship to oocytes. The oocytes may also express CK18, a marker of the Balbiani body, but not CK5,6,8,17 markers of OSE and granulosa cells (reviewed in [6]). Furthermore, oocytes also express the intermediate filament vimentin, a protein which plays an important role in the maturation and fertilization of eggs [21].
Our goal for the present study was to investigate mesenchymal-epithelial and epithelial-mesenchymal transitions in OSE cell cultures from adult human ovaries. An in vitro approach allows control of the conditions in the initial medium composition, e.g., presence or absence of estrogenic stimuli, and determination of the secretory products of cultured cells in the conditioned medium. The study of OSE cultures was initially a part of a larger project to establish primary fibroblast cultures from the pelvic floor (levator ani fascia) and uterine round ligaments. Our preliminary studies indicated that pelvic floor fibroblasts grow better in the presence of estrogenic stimuli. We also detected the presence of oocyte phenotype cells in some OSE cultures. Thus, in the present experiments, we cultured OSE cells in the presence or absence of estrogenic stimuli. Phase contract microscopy and antibodies against HSPZ, PS1, CK18, CK5,6,8,17, and vimentin were used in single and double color peroxidase immunohistochemistry to identify oocytes and granulosa cells in cultures derived from adult human ovaries.
Materials and methods
Collection and cultures of OSE cells
All chemicals and consumables, except where specified otherwise, were purchased from Sigma Chemical Co., St. Louis, MO. Cultured cells were collected from ovaries associated with fresh hysterectomy/bilateral salpingo-oophorectomy specimens (five women, ages 39 – 52 years). The surgery was performed for medical indications, including chronic pelvic pain, uterine fibroids, and/or uterine bleeding (severe dysmenorrhea) not responding to the conservative treatment. The study was approved by the Institutional Review Board and informed consent was obtained from the patients.
In order to prevent loss of the OSE cells, no attempts were made to clean the ovarian surface from blood contamination accompanying the surgery prior to their collection. Consequently, the marked contamination with erythrocytes, along with our inability to distinguish cell types at this time, hindered our ability to count cells until the medium was changes after 24 hours of culture. The surface of intact ovaries was first gently scraped in the aseptic laminar flow hood with a sterile stainless steel surgery knife blade No. 21 (Becton Dickinson, AcuteCare, Franklin Lakes, NJ), leaving the blade edge behind the knife pass (pertains to data in Figs 1,2,3). This procedure was selected with an intent to include OSE and some adjacent TA mesenchymal cells. In some instances, a second type of culture was also established, where the cells collected from the ovarian surface were enriched with cells spontaneously released (without scraping) from dissected ovaries (pertains to data in Fig. 4). The intent of this approach was to test the activity of the alternative pathway for the origin of germ cells from cortical epithelial crypts (see Fig. 5a and Ref. [6] for details). Such cells were passed through a sterile 70 μm nylon cell strainer (BD Biosciences, Bedford, MA) to prevent contamination with larger type cells and structures.
Figure 1 Cell phenotypes in 6-day OSE culture maintained in phenol red free DMEM/F12 medium. a) Live fibroblast type cells in phase contrast microscopy (PhC) show perinuclear accumulation of organelles (arrowhead). b), Immunostaining of the same cells with HSPZ antibody shows ZP+ nuclei and perinuclear organelles (arrowhead). c) A cluster of cells showing epithelial phenotype with ZP+ nuclei and perinuclear organelles (arrowhead). d) ZP+ neural type cell shows extensions toward the ZP+ perinuclear staining (arrowhead) of neighboring mesenchymal type cells. Inset shows similar type of cell in the live culture. e) Clusters (arrows) of small (15 μm) cells of granulosa cell phenotype with strong ZP expression. Arrowhead indicates mesenchymal type cells. f) Double staining for ZP/CK5,6,8,17 (blue substrate) shows dark-blue staining of small cell clusters (arrows), but mesenchymal type cells remain brown only (arrowhead).
Figure 2 Five-day OSE culture maintained in DMEM-HG medium with phenol red. a) Phase contrast microscopy of a live cell with a large oocyte phenotype (120 μm in diameter) shows a centrally located nucleus with nucleolus and perinuclear accumulation of cell organelles. b) Larger cell reaching 180 μm shows similar morphology. c) The cell presented in panel b stained for CK18 shows cytoplasmic staining with an accumulation of CK18 immunoexpression in the perinuclear space (arrowhead). d) The same cell in CK18 (brown) and ZP (blue) double color immunostaining (CK/ZP) shows high density of ZP expression in the perinuclear space (compare with panel c). e) Control staining shows no reactions in OSE cells when the anti-HLA DR antibody was used in place of the primary antibody. f) A cluster of cells exhibiting epithelial phenotype shows nuclear ZP expression (arrowhead) and differentiation into large cells (arrows). g) Another cluster subjected to the ZP/CK,5,6,8,17 double staining shows dark-blue staining of smaller epithelial cells (arrowhead), which diminishes in more differentiated larger cells (arrow). Bar in (a) for a-e.
Figure 3 Five-day OSE culture maintained in the DMEM-HG medium with phenol red as in Fig. 2. a) Cell of oocyte phenotype with two nuclei. The centrally located nucleus shows ZP immunostaining (arrowhead) but adjacent nucleus is unstained (arrow). Note ZP+ intermediate filaments. b) The same cell subjected to double color (ZP/CK5,6,8,17) immunohistochemistry shows no additional (dark blue) staining. c) Another large cell stained for vimentin shows expression in intermediate filaments and two unstained large nuclei (arrows). d) Double color CK18/ZP staining shows intermediate filaments and ZP+ centrally located nucleus (arrowhead) and unstained adjacent structure resembling the polar body (arrow). e) Cell with centrally located ZP+ nucleus (arrowhead) and unstained fragmented adjacent structure (arrow). Note a lack of ZP+ intermediate filaments (compare with panel a) and surface ZP expression (solid arrow; compare with black arrow, panel d). f) Nuclear expression (arrowhead) of meiotically expressed PS1 carbohydrate ZP antigen. Note focally enhanced (coarse) staining (arrow) in one of the two rounded structures. g) Two cells showing a lack of nuclear PS1 expression (arrowheads), but strong staining is associated with adjacent polar-like bodies (arrows).
Figure 4 Development of oocytes in day 3 to 5 mixed ovarian cultures. Undifferentiated (stem) cells persist in all day 3 cultures (a). Under PhR+ conditions (b-h), the 100 μm oocyte-like cells were found on day 5 (arrows, b) with developing zona pellucida layer (arrow, c) in the presence of associated fibroblasts (arrowheads). Association of granulosa type cells (arrowheads, d) appears to stimulate advanced oocyte growth. Thick zona pellucida layer is apparent in some oocytes accompanied by fibroblasts on day 4 (e), and staining for ZP proteins on day 5 shows no change in the oocyte size but a lack of surface ZP expression (arrow, f). Staining for CK5,6,8,17 (g) shows cytoplasmic expression in mesenchymal type cells (black arrowhead) and non-specific staining (peroxidase expression) in the nuclei of other cells (arrowheads). h) Isolated large (200 μm) cells showed characteristics of secondary oocytes, including poor separation between the nucleus and cytoplasm (black arrowhead) expulsion of polar body (white arrowhead) and surface expression of ZP proteins (black arrow). The latter was, however, not apparent in the surface segment close to the polar body (white arrow). Under PhR- conditions, rare giant oocytes (300 μm) were found on day 4 (i). They lacked the germinal vesicle breakdown and surface expression of ZP proteins on day 5 (arrow, j). Bar in (i) for a-f and h-j.
Figure 5 Comparison of human oogenesis in vivo and in vitro a) A working model of possible pathways for oogenesis and formation of primary follicles in adult human ovaries in vivo (adjusted from Ref. [6]). (1) Ovarian tunica albuginea (ta) stem cells (green color) first differentiate into the CK+ fibroblasts (blue color) and by mesenchymal-epithelial transition give rise to the SE cells directly covering the ovarian cortex (arched arrow). (2) Closing of TA results in the formation of epithelial cords in the upper ovarian cortex. (3) Fragmented epithelial cords give rise to the epithelial nests, which resemble primitive granulosa cells and descend into the lower ovarian cortex. (4) Depending on certain in situ (stromal and neuronal) and systemic (hormonal) influences, the TA progenitors differentiate into the SE cells covering TA, which may, by asymmetric division, give rise to the ZP+ germ cells. (5) These primitive germ cells symmetrically divide, descend into the ovarian cortex, and associate with adjacent cortical vessels (6). Intravascular transport (7) is associated with a substantial increase of germ cell size and with development of ZP+ anchors (green lines), which may serve to slow down the transport speed and signal the epithelial nests to associate with a particular vascular segment. 8) The intravascular germ cells differentiating into the oocytes are picked up by epithelial nests associated with the proper cortical vessels. Such oocyte-nest complexes show an "octopus-like" (9) formations during the early stage of assembly, and a formation of the Balbiani body during the intermediate stage (light blue body, 10). The Balbiani body persists in resting primary follicles (11), but diminishes upon the growth promoting signals, including Thy-1 dp signaling derived from the follicle-accompanying vessels (12, dashed line). An alternative pathway for the germ cell origin from TA precursors (4') consists of a constitution of cortical crypts formed by SE-like embryonal type cells, possibly originating from, but not necessarily connected with the deep SE invaginations, as evidenced from serial sections. The "alternative" pathway of germ cell origin may supply the oocytes directly to the neighboring nests (dashed arched arrows) and, via vascular transport (dotted arched arrow), saturate distant nests to form the primary follicles. The oocytes not utilized for follicular renewal accumulate in medullary vessels and degenerate (bottom right) [6]. b) In vitro, the CK5+ oocyte precursors may differentiate into CK5- cells (see Fig 2g), enlarge and undergo germinal vesicle breakdown followed by first meiotic division with a release of the polar body (pb) and ZP surface expression (brown color, see Fig. 3e and 4h).
The cells were collected into sterile petri dishes containing tissue culture medium supplemented with heat inactivated 20% fetal bovine serum (FBS; Gibco/BRL, Grand Island, NY) and antibiotics (50 μg/ml gentamycin, 100 U/ml penicillin, and 100 μg/ml streptomycin). The tissue culture media utilized was either Dulbecco's Modified Eagle Medium/Ham's F12, phenol red free (DMEM/F12; without estrogenic stimuli) or Dulbecco's Modified Eagle's Medium containing 25 mM HEPES, 4500 mg/L glucose, and phenol red (DMEM-HG; with estrogenic stimuli). The possible influence of additional steroid hormones or promoters of steroidogenesis present in the serum can not be ruled out. However, at the same serum concentration, there were apparent differences in the evolution and maturation of oocytes cultured in media with and without PhR (Fig. 4h–j). There was no other treatment imposed during the culture.
The cells were spun down (1000 × g, 5 min, 24°C), diluted in 0.75 – 1.5 ml of supplemented media, seeded in either 3 or 6 wells of a 24-well plate (250 μl per well) (Fisher Scientific, Pittsburgh, PA), and cultured in an humidified atmosphere with 5% CO2 at 37°C. The number of wells was chosen by the size of ovaries. Cells collected from small ovaries were seeded into 3 wells, and from larger ovaries into six wells. All ovaries involved in the experiment were anovulatory, as no corpora lutea were detected. The culture medium was changed once after 24 hours. This left only adherent (viable) cells in culture, and eliminated non-adherent cells and the majority of contaminating erythrocytes. The cell cultures were monitored daily by phase contrast microscopy and live cells evaluated by immunohistochemistry after 5–6 days from the initial seeding. Viability of cells was apparent from their active movement, changes in shape, and movement of their nuclei (compare Fig. 4e and 4f). The number of adherent cells in a single well of 24-well plate ranged between ~100 to 1000 during the late culture period (day 5 or 6).
Single and double color peroxidase immunohistochemistry
The medium was aspirated from wells of the 24-well plate, and cells were allowed to dry under a fan before placing them in a ventilated hood overnight. This is a procedure we have utilized for cryostat sections. The bottoms of the wells were dried in the upright position and found macroscopically dry within a few seconds. The cells were fixed with 96% ethanol for 5 minutes, allowed to dry again, and incubated overnight (4°C) with primary antibodies against ZP proteins: rabbit-anti heat solubilized porcine zona [19] (HSPZ, 1: 20) or mouse monoclonal PS1 antibody recognizing meiotically expressed ZP carbohydrate antigen [20,22] diluted 1:10 in phosphate buffered solution (PBS), pH 7.22. The HSPZ and PS1 antibodies were originally donated by Dr. Dunbar, and an additional HSPZ antibody by Dr. Prasad. We also used mouse-anti human CK18, clone CY90, Sigma (diluted 1:50), mouse-anti human CK5,6,8,17, clone MNF116, DAKO Corporation, Carpinteria, CA (diluted 1:50), and mouse anti-human vimentin, clone V9, DAKO Corporation (diluted 1:50).
As a control, the HLA-DR antibody donated by Drs. Hilgert and Horejsi (tissue culture supernatant diluted 1:5), which does not react with OSE cells, was used. We used an HLA-DR antibody not reacting with fibroblast, granulosa, epithelial, germ or oocyte cell types, in order to identify activated tissue macrophages, but did not find any such cells in day 5 or 6 cultures. This does not exclude the possibility that activated tissue macrophages may have been present during the initial stages of culture.
After several washes in PBS (room temperature), the cells were incubated with peroxidase labeled corresponding secondary antibodies – goat anti-rabbit IgG, pre-absorbed with human serum (Jackson Immunoresearch, West Grove, PA), diluted 1:50; or swine anti-mouse IgG (donated by Dr. Peknicova), diluted 1:50 and absorbed with rat kidney homogenate to remove background as described previously [6]. After additional washes in PBS (room temperature), the bound antibodies were visualized by diaminobenzidine substrate as described previously [6], but without hematoxylin counterstain, covered with PBS, and images captured as described below. The cells stained for CK18 or ZP were processed further for dual color immunohistochemistry to identify co-expression of other proteins, and visualized with blue chromogen substrate as described previously [6]. Finally, the washed cells were covered with PBS containing 0.01% azide as a preservative.
Image processing
Two independent observers evaluated the live cells by phase contrast and subsequent immunohistochemistry using an inverted microscope (NIKON, Nikon Inc., Instrument Division, Garden City, NY) equipped with a DEI-470 CCD Video Camera System (Optronics Engineering, Goleta, CA) with detail enhancement. The video images were captured by CG-7 color frame grabber (Scion Corporation, Frederick, MD) supported by Scion Image public software developed at the National Institutes of Health (Wayne Rasband, NIH, Bethesda, MD), and ported to Windows XP, 2002 release (Microsoft Corporation, Redmont, WA). To obtain figure panels, the captured video images were copied into Microsoft® Power-Point® 97 SR-2 (Microsoft Corporation). Each image (including controls) was further copied into Microsoft Photo Editor 3.0 (Microsoft Corporation), and blue saturation adjusted (brightness 70, contrast 70, gamma 0.30). In total, more than 100 images were captured and stored.
Results and discussion
Oogenesis in vitro from ovarian surface epithelium (OSE) cells
We report data from pure OSE cultures of two cases investigated either under PhR- (age 47) or PhR+ medium conditions (age 40), and pure OSE and mixed cultures (OSE + stromal compartments; one patient) studied under both conditions (age 39 years). Cultures from the remaining two patients (ages 42 and 52 years) consisted of mesenchymal cells only.
Experiment 1 (Fig. 1)
Primary culture of cells scrapped from the surface of ovaries of the 47 year old female was maintained for 6 days in the Dulbecco's Modified Eagle Medium/Ham's F12 phenol red free (DMEM/F12), which was supplemented with 20% fetal bovine serum (FBS) and antibiotics. The 5–6 day interval was chosen since, except for fibroblasts, other cell types are not phenotypically distinguishable prior to day 4 (see Fig. 4). Phase contrast microscopy (Fig. 1a) shows differentiation of cells toward the mesenchymal (fibroblast) phenotype with perinuclear accumulation of cytoplasmic organelles (arrowhead).
Immunohistochemistry shows ZP expression associated with the nuclei and perinuclear space of mesenchymal type cells (arrowhead, Fig. 1b). The culture also showed clusters of cells of epithelial phenotype, which exhibited nuclear and perinuclear ZP expression (Fig. 1c). Occasionally, ZP+ cells of neural phenotype with extensions toward the mesenchymal cells (arrowhead, Fig. 1d) were observed, including growing cultures (inset). Fig. 1e shows mesenchymal type cells (arrowhead) and clusters (arrows) of small (15 μm) cells of granulosa cell phenotype. Beside their characteristic size, these granulosa type cells showed strong ZP expression, which has been found to be shared between oocytes and granulosa cells [6,23]. Furthermore, double color ZP/CK5,6,8,17 immunohistochemistry (Fig. 1f) showed additional dark blue staining (arrows) characteristic of granulosa but not germ or mesenchymal cells [6]. Note only a brown color in mesenchymal type cells (arrowhead). These observations indicate that numerous small cells exhibiting the phenotype and biochemical markers of granulosa cells can differentiate in the OSE culture.
An interesting question is why other cell types, in addition to the mesenchymal cells, evolved in this culture. As we have noted, the multipotential progenitor cells exhibiting a mesenchymal phenotype are capable of differentiation into distinct cell types, similar to stem cell cultures derived from other adult tissues [12-16]. In this culture, groups of cells exhibited either epithelial, neural, of granulosa cell phenotypes. We speculate that if a single mesenchymal cell in the given group decides to be, for example, an epithelial cell, the adjacent mesenchymal cells follow the same pattern. If no such decision is made, the cells in the group remain mesenchymal (compare groups of cells in Fig 1b–f). Hence, in this instance, the result is not an artifact of the medium used, but may be a case of distinct differentiation between groups of cells in the same culture.
Experiment 2 (Figs 2 and 3)
In contrast, cells from 40 year old female cultured in the Dulbecco's Modified Eagle's Medium containing 25 mM HEPES, 4500 mg/L glucose and phenol red (DMEM-HG; supplemented with 20% FBS and antibiotics as above) showed a uniform pattern of epithelial type cells, while no other cell types, including mesenchymal and granulosa cells, were present. Phenol red is a weak estrogen with apparent biological effects, and the estrogen receptor alpha binding affinity in Dulbecco's Modified Eagle's Medium is equivalent to levels of 17β estradiol as high as 0.45 nM [24]. This estrogenic binding affinity of phenol red is close to that of 17β estradiol serum level during the preovulatory peak [25]. High estrogen concentrations prior to ovulation are not only predictive of high oocyte quality [26], but they may be essential for stimulating oocytes to enter the first meiotic division and form large secondary oocytes. Figs 2a and 2b show phase contrast of large cells exhibiting an oocyte phenotype in the OSE culture maintained in DMEM-HG medium with phenol red for 5 days. The cell in Fig. 2b reached 180 μm in diameter and showed a centrally located 40 μm nucleus with nucleolus, and perinuclear accumulation of organelles (arrowhead).
Fig. 2c shows immunohistochemical staining of the cell in panel 2b for CK18. The nascent primary follicles in adult human ovaries show the gradual formation of CK18+ Balbiani bodies, which are undetectable by antibodies against other cytokeratins expressed by granulosa cells – CK5,6,8,17 [5,6,27]. A study by Cox and Spradling indicates that during Drosophila oogenesis, follicular cells are a source of mitochondria, which enter the oocyte cytoplasm via the "ring canal" to form the Balbiani body, thereby supplying virtually all of the mitochondria of the oocyte [28]. Dispersion of Balbiani bodies releases mitochondria required for the progressive growth of the oocyte. In a study of turkey hens, no Balbiani bodies were detected in stage I oocytes, but they appeared in stage II oocytes, and diminished in the oocytes of growing follicles, coinciding with the dispersion of mitochondria throughout the ooplasm [29]. Similar observations were reported in human oocytes [30]. The staining of the large oocyte in Fig. 2c shows diffuse CK18 immunoexpression with a preserved accumulation of staining around the cell nucleus (arrowhead). The perinuclear space also exhibits enhanced staining for ZP proteins shown by double color immunohistochemistry (CK brown/ZP blue color) in the same cell (arrowhead, Fig. 2d, compare with 2c). Control immunohistochemical staining is shown in Fig. 2e. Fig. 2f shows a cluster of OSE cells with perinuclear ZP expression (arrowhead), which differentiates into the large cells (arrows) that can be released into the free space. This resembles oocytes leaving the ovary during the perinatal period [31]. Fig 2g shows another cluster of OSE cells subjected to ZP/CK5,6,8,17 staining. Note the dark blue color in smaller OSE cells (arrowhead) and a diminution in larger cells (arrow).
We also observed large cells with two nuclei, where the ZP expression was apparent in the centrally located nucleus only (arrowhead vs. arrow, Fig. 3a; note ZP+ intermediate filaments). These cells of oocyte phenotype did not express CK5,6,8,17 [6] (Fig. 3b). The cell in Fig. 3c has two 60 μm nuclei and expresses the intermediate filament vimentin, a protein which plays an important role in the maturation and fertilization of eggs [21]. This observation indicates endoreplication of the oocyte nuclei in the absence of cell division, which has also been described in Drosophila eggs [32,33]. Interestingly, these multinucleated cells were not observed in the cultures without phenol red. The cell in Fig. 3d was stained for CK18 (brown color) and ZP proteins (blue). The centrally located nucleus again shows ZP expression (arrowhead), while the nucleus to the side, which resembles a polar body, was unstained (white arrow). Note a lack of surface staining (black arrow) vs. Fig 3e. The cell in Fig. 3e shows a ZP+ centrally located nucleus (arrowhead), with an adjacent unstained and fragmented structure (white arrow). Notably, there was surface expression of ZP proteins (solid arrow), which is important for the sperm-egg interaction [34]. Note the diminution of ZP+ intermediate filaments vs. Fig. 3a. Fig. 3f shows nuclear staining (arrowhead) for PS1, a meiotically expressed lactosaminoglycan-associated carbohydrate antigen of all three ZP glycoproteins [22,35], with more intense expression in the one of two nuclear segments, as indicated by the arrow. Two cells in Fig. 3g show the PS1 unstained nuclei (arrowheads) and adjacent PS1+ structures resembling expelled polar bodies (arrows). It has to be noted that similar couples of cells were often observed in PhR+ cultures, one of which was usually more developed (left cell in panel g) compared to the other (right).
Oogenesis in "mixed" ovarian cultures
Experiment 3 (Fig. 4)
Next we decided to compare the PhR+ and PhR- conditions in pure OSE and mixed (OSE + stromal components) cultures derived from ovaries of a 39 year old female. It is important to note that all cells collected, including pure surface cultures, were passed through the cell strainer prior to processing and seeding, a procedure not applied in experiments 1 and 2. Utilization of a cell strainer ensured that mostly single cells, and not cell sheets (see Fig. 2f and 2g) were seeded. The mixed cultures were established to test the possible involvement of the alternative pathway for germ cell origin (Fig. 5a). Except for the occasional mesenchymal cells of the fibroblast phenotype, there was no evidence of cell commitment until day 4 of culture. However, the undifferentiated cells varied in the density of organelles and size, the latter ranging from 15 (arrowhead, Fig. 4a) to 50 μm in diameter (arrow).
On day 5 (end of the experiment) the cells collected from the ovarian surface did not progress beyond the state found on day 3, and no oocytes were detected regardless of the presence or absence of PhR. However, many cells of the oocyte phenotype were found in mixed cultures with PhR. They showed moderate (100 μm) size without zona pellucida (arrows, Fig. 4b), and were accompanied by fibroblasts (arrowheads). Larger oocytes (120 μm) accompanied by fibroblasts (arrowheads, Fig. 4c) showed a developing zona layer (arrow). On the other hand, an association of small round cells resembling granulosa cells (arrowheads, Fig. 4d) was characteristic for large oocytes (200 μm). The oocytes accompanied by fibroblasts (arrowheads, Fig. 4e) with thick zona pellucida (arrow) observed on day 4 showed no change in size on day 5 (Fig. 4f). Staining for ZP proteins showed a strong nuclear localization (black arrowhead, Fig. 4f) but no surface ZP expression (arrow). Note a well-defined separation between the nucleus and cytoplasm (a nuclear envelope, black arrowhead) of the oocyte. Also note the cellular and nuclear movements, when the Fig. 4e and 4f are compared. Fig. 4g shows cytoplasmic staining for CK 5,6,8,17 in mesenchymal cells (black arrowhead, note unstained nuclei) and moderate nonspecific staining (peroxidase expression) in the nuclei of other uncommitted cells (white arrowheads) in this culture. The nuclear peroxidase expression has been recently reported to increase in bovine oocytes and cumulus cells during in vitro maturation [36], and this may apply for many cells in this ovarian culture.
Some large cells in PhR+ mixed cultures showed characteristics of secondary oocytes with surface ZP expression (black arrow, Fig. 4h), expulsion of the polar body (white arrowhead) and poor nuclear/cytoplasmic separation (black arrowhead; see Ref. [37]). Interestingly, the surface ZP expression was absent in the oocyte segment toward where the polar body was extruded (white arrow).
Rare oocytes were detected in mixed cultures without PhR. The four day culture shows a giant (300 μm) oocyte with persisting germinal vesicle (Fig. 4i). Staining for ZP proteins (brown color) and vimentin (blue) on day 5 shows no surface ZP expression (arrow, Fig. 4j). This suggests that estrogenic stimuli may be required for the germinal vesicle breakdown and the development of secondary oocytes.
There are several issues to be addressed when experiments 1–3 are compared. For instance, experiment 3 provides evidence that ovarian fibroblasts may stimulate the oocytes to develop a thick zona pellucida layer, which allows their preservation until follicular growth is induced, when granulosa cells may contribute to oocyte enlargement. However, when compared to the experiment 2, no oocyte development was apparent in cultures derived from ovarian surface regardless of the presence of PhR. In experiment 3, however, the cells collected from the ovarian surface were passed through a cell strainer, which probably prevented the passage of the OSE cell sheets detected in the experiment 2 cultures (Fig. 2f and 2g). Apparently, cells derived from the OSE cell clusters are capable of transforming directly into secondary oocytes (Figs 2 and 3). On the other hand, experiment 3 has shown that an alternative origin of oocytes, e.g. from OSE cells in ovarian crypts (Fig. 5a), is possible in these mixed ovarian cultures in the presence of PhR. The likelihood that these mixed cultures are contaminated by migrating germ cells cannot be excluded, although very few oocytes were detected in PhR- mixed cultures. Nevertheless, many more oocytes of distinct sizes and stages of development were found in mixed cultures with PhR+. This suggests that estrogenic stimulation may play an important role in oocyte development. In addition, the age of the patients must be taken into account. Five day cultures of two patients (age 42 and 52) showed only mesenchymal cells. We suggested earlier that OSE in aging ovaries may be unable to produce germ cells, possibly due to the lack of local signals from activated tissue macrophages and neural cells, which may produce substances regulating the fate of progenitor (stem) cells [5] (see also Background).
Conclusion
Altogether, our observations show for the first time that granulosa cells (Figs 1e and 1f) and oocytes (Figs 2 and 3) may develop directly from cultured OSE cells derived from adult human ovaries. This confirms our in vivo observations that in adult human ovaries, the OSE is a bipotent source of oocytes and granulosa cells. In addition, the oocytes developed in vitro undergo the first meiotic division, (Fig. 5b), after which they become suitable for fertilization. Oocytes may also develop in cultures containing ovarian stromal components. We speculate that such oocytes may originate from migrating germ cells or OSE invaginations (cortical crypts, Fig. 5a). The development and maturation of oocytes appear to be stimulated by estrogens. Depending on culture conditions (type of media utilized), processing of the collected cells, age of the ovaries, commitment of neighboring cells, and other local and hormonal factors, the progenitor mesenchymal cells in ovarian cultures may differentiate into additional cell types, including granulosa cells, or persist unchanged.
The ability to produce mature mammalian eggs from adult ovaries in vitro has several potential applications in the human and animal reproduction. Firstly, compared to the collection of follicular oocytes, the technique is easier (scrapping of the ovarian surface, with or without cortical component) and the yield might be higher for IVF and veterinary medicine purposes, since differentiation of primary oocytes in vitro may provide a larger number of secondary (mature) eggs. Secondly, for IVF purposes, this technique may be successful in women with premature ovarian failure, who lack follicles in their ovaries. Thirdly, the development and differentiation of oocytes from OSE precursors in vitro may help to better understand the process of oocyte renewal in vivo, and the role of accompanying granulosa and mesenchymal cells in the regulation of oocyte maturation or preservation. Fourthly, frozen OSE cells (oocyte stem cells) from younger females may be preserved for later production of fresh eggs. This may prevent the occurrence of fetal genetic alterations, which are often associated with pregnancies in advanced maternal age, possibly due to the lack of follicular renewal in aging ovaries. In addition, a colonization of premenopausal ovaries with younger oocyte and granulosa stem cells may establish a new cohort of primary follicles. This may result in a 10- to 12-year delay of the onset of natural menopause. Finally, the ovarian stem cells may serve as progenitor cells for several cell types for stem cell research, and fertilization of evolved mature human oocytes could result in the production of totipotent embryonic stem cells for research purposes and therapeutic applications.
Acknowledgements
This research is supported by the Physicians' Medical Education and Research Foundation, Knoxville, TN, and the Department of Obstetrics and Gynecology UT G.S.M. funds. Nirmala B. Upadhyaya, M.D., and Pleas Copas, M.D., of the Department of Obstetrics and Gynecology, and Stewart E. Van Meter, M.D., of the Department of Pathology, UT G.S.M., provided ovaries and ovarian tissue samples from hystero-salpingo-oophorectomies. We thank Dr. Bonnie S. Dunbar and Dr. Sarvamangala V. Prasad of the Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, Texas, for the kind gift of the PS1 and HSPZ anti-ZP antibodies. Dr. Ivan Hilgert and Dr. Vaclav Horejsi of the Czech Academy of Sciences, Prague, Czech Republic, kindly supplied the HLA-DR antibody; and Dr. Jana Peknicova of the Czech Academy of Sciences, Prague, Czech Republic, provided the swine anti-mouse IgG peroxidase conjugate.
==== Refs
Franchi LL Mandl AM Zuckerman S Zuckerman S The development of the ovary and the process of oogenesis The Ovary 1962 London: Academic Press 1 88
Hubner K Fuhrmann G Christenson LK Kehler J Reinbold R De La FR Wood J Strauss JF IIIBoiani M Scholer HR Derivation of oocytes from mouse embryonic stem cells Science 2003 300 1251 1256 12730498 10.1126/science.1083452
Ioannou JM Oogenesis in adult prosimians J Embryol Exp Morphol 1968 17 139 145 6040545
Johnson J Canning J Kaneko T Pru JK Tilly JL Germline stem cells and follicular renewal in the postnatal mammalian ovary Nature 2004 428 145 150 15014492 10.1038/nature02316
Bukovsky A Keenan JA Caudle MR Wimalasena J Upadhyaya NB Van Meter SE Immunohistochemical studies of the adult human ovary: possible contribution of immune and epithelial factors to folliculogenesis Am J Reprod Immunol 1995 33 323 340 7546251
Bukovsky A Caudle MR Svetlikova M Upadhyaya NB Origin of germ cells and formation of new primary follicles in adult human ovaries Reprod Biol Endocrinol 2004 2 20 15115550 10.1186/1477-7827-2-20
Bazer FW Strong science challenges conventional wisdom: new perspectives on ovarian biology Reprod Biol Endocrinol 2004 2 28 15186496 10.1186/1477-7827-2-28
Gosden RG Germline stem cells in the postnatal ovary: is the ovary more like a testis? Hum Reprod Update 2004 10 193 195 15140866 10.1093/humupd/dmh023
Spradling AC Stem cells: more like a man Nature 2004 428 133 134 15014482 10.1038/428133b
Auersperg N Wong AS Choi KC Kang SK Leung PC Ovarian surface epithelium: biology, endocrinology, and pathology Endocr Rev 2001 22 255 288 11294827 10.1210/er.22.2.255
Auersperg N Pan J Grove BD Peterson T Fisher J Maines-Bandiera S Somasiri A Roskelley CD E-cadherin induces mesenchymal-to-epithelial transition in human ovarian surface epithelium Proc Natl Acad Sci U S A 1999 96 6249 6254 10339573 10.1073/pnas.96.11.6249
Jiang Y Vaessen B Lenvik T Blackstad M Reyes M Verfaillie CM Multipotent progenitor cells can be isolated from postnatal murine bone marrow, muscle, and brain Exp Hematol 2002 30 896 904 12160841 10.1016/S0301-472X(02)00869-X
Reyes M Dudek A Jahagirdar B Koodie L Marker PH Verfaillie CM Origin of endothelial progenitors in human postnatal bone marrow J Clin Invest 2002 109 337 346 11827993 10.1172/JCI200214327
Van Damme A Vanden Driessche T Collen D Chuah MK Bone marrow stromal cells as targets for gene therapy Curr Gene Ther 2002 2 195 209 12109216
Keene CD Ortiz-Gonzalez XR Jiang Y Largaespada DA Verfaillie CM Low WC Neural differentiation and incorporation of bone marrow-derived multipotent adult progenitor cells after single cell transplantation into blastocyst stage mouse embryos Cell Transplant 2003 12 201 213 12797375
Muguruma Y Reyes M Nakamura Y Sato T Matsuzawa H Miyatake H Akatsuka A Itoh J Yahata T Ando K In vivo and in vitro differentiation of myocytes from human bone marrow-derived multipotent progenitor cells Exp Hematol 2003 31 1323 1330 14662341 10.1016/j.exphem.2003.09.003
Prindull G Zipori D Environmental guidance of normal and tumor cell plasticity: epithelial mesenchymal transitions as a paradigm Blood 2004 103 2892 2899 15070660 10.1182/blood-2003-08-2807
Picher M Burch LH Hirsh AJ Spychala J Boucher RC Ecto 5'-nucleotidase and nonspecific alkaline phosphatase. Two AMP-hydrolyzing ectoenzymes with distinct roles in human airways J Biol Chem 2003 278 13468 13479 12560324 10.1074/jbc.M300569200
Dunbar BS Raynor BD Characterization of porcine zona pellucida antigens Biol Reprod 1980 22 941 954 6156718
Skinner SM Dunbar BS Localization of a carbohydrate antigen associated with growing oocytes and ovarian surface epithelium J Histochem Cytochem 1992 40 1031 1036 1607636
Payne C Rawe V Ramalho-Santos J Simerly C Schatten G Preferentially localized dynein and perinuclear dynactin associate with nuclear pore complex proteins to mediate genomic union during mammalian fertilization J Cell Sci 2003 116 4727 4738 14600259 10.1242/jcs.00784
Skinner SM Lee VH Kieback DG Jones LA Kaplan AL Dunbar BS Identification of a meiotically expressed carbohydrate antigen in ovarian carcinoma: I. Immunohistochemical localization Anticancer Res 1997 17 901 906 9137425
Carino C Prasad S Skinner S Dunbar B Chirinos M Schwoebel E Larrea F Dunbar B Localization of species conserved zona pellucida antigens in mammalian ovaries Reprod Biomed Online 2002 4 116 126 12470573
Berthois Y Katzenellenbogen JA Katzenellenbogen BS Phenol red in tissue culture media is a weak estrogen: implications concerning the study of estrogen-responsive cells in culture Proc Natl Acad Sci U S A 1986 83 2496 2500 3458212
Lobo RA Lobo RA, Mishell DR Jr, Paulson RJ, Shoupe D Endocrinology of pregnancy Mishell's Textbook of Infertility, Contraception, and Reproductive Endocrinology 1997 Malden, Massachusetts: Blackwell Science 183 206
Pena JE Chang PL Thornton MH Sauer MV Serum estradiol levels after 4 days of ovarian hyperstimulation in oocyte donors are predictive of embryo quality and clinical outcomes Gynecol Obstet Invest 2002 54 207 212 12592063 10.1159/000068384
Santini D Ceccarelli C Mazzoleni G Pasquinelli G Jasonni VM Martinelli GN Demonstration of cytokeratin intermediate filaments in oocytes of the developing and adult human ovary Histochemistry 1993 99 311 319 7684733 10.1007/BF00269104
Cox RT Spradling AC A Balbiani body and the fusome mediate mitochondrial inheritance during Drosophila oogenesis Development 2003 130 1579 1590 12620983 10.1242/dev.00365
Carlson JL Bakst MR Ottinger MA Developmental stages of primary oocytes in turkeys Poult Sci 1996 75 1569 1578 9000285
Motta PM Nottola SA Makabe S Heyn R Mitochondrial morphology in human fetal and adult female germ cells Hum Reprod 2000 15 129 147 11041520
Motta PM Makabe S Elimination of germ cells during differentiation of the human ovary: an electron microscopic study Eur J Obstet Gynecol Reprod Biol 1986 22 271 286 3770277
Hawkins NC Thorpe J Schupbach T Encore, a gene required for the regulation of germ line mitosis and oocyte differentiation during Drosophila oogenesis Development 1996 122 281 290 8565840
Maines JZ Stevens LM Tong X Stein D Drosophila dMyc is required for ovary cell growth and endoreplication Development 2004 131 775 786 14724122 10.1242/dev.00932
Prasad SV Skinner SM Carino C Wang N Cartwright J Dunbar BS Structure and function of the proteins of the mammalian Zona pellucida Cells Tissues Organs 2000 166 148 164 10729725 10.1159/000016730
Dunbar BS Timmons TM Skinner SM Prasad SV Molecular analysis of a carbohydrate antigen involved in the structure and function of zona pellucida glycoproteins Biol Reprod 2001 65 951 960 11514363
Leyens G Verhaeghe B Landtmeters M Marchandise J Knoops B Donnay I Peroxiredoxin 6 is upregulated in bovine oocytes and cumulus cells during in vitro maturation: role of intercellular communication Biol Reprod 2004 71 1646 1651 15240427 10.1095/biolreprod.104.030155
Peters H McNatty KP The Ovary. A Correlation of Structure and Function in Mammals 1980 Berkeley and Los Angeles, California: University of California Press
| 15871747 | PMC1131924 | CC BY | 2021-01-04 16:37:13 | no | Reprod Biol Endocrinol. 2005 May 5; 3:17 | utf-8 | Reprod Biol Endocrinol | 2,005 | 10.1186/1477-7827-3-17 | oa_comm |
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-251582901210.1186/1742-4690-2-25ReviewIntracellular immunity to HIV-1: newly defined retroviral battles inside infected cells Zheng Yong-Hui [email protected] B Matija [email protected] Departments of Medicine, Microbiology and Immunology, Rosalind Russell Arthritis Research Center, University of California, San Francisco, San Francisco, CA, 94143-0703, USA2005 13 4 2005 2 25 25 24 2 2005 13 4 2005 Copyright © 2005 Zheng and Peterlin; licensee BioMed Central Ltd.2005Zheng and Peterlin; 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.
Studies of the human immunodeficiency virus type 1 (HIV-1) continue to enrich eukaryotic biology and immunology. Recent advances have defined factors that function after viral entry and prevent the replication of proviruses in the infected cell. Some of these attack directly viral structures whereas others edit viral genetic material during reverse transcription. Together, they provide strong and immediate intracellular immunity against incoming pathogens. These processes also offer a tantalizing glimpse at basic cellular mechanisms that might restrict the movement of mobile genetic elements and protect the genome.
==== Body
Background
Although it is highly pathogenic in humans, HIV-1 cannot replicate in most other species [1]. This tropism is determined primarily by whether host cells express the required cofactors. For example, by lacking a functional receptor and appropriate transcriptional machinery, mouse cells do not support infection by HIV-1. Thus, the organism resists the pathogen via a cell-based incompatibility. However, a pathogen can also be restricted by the presence of dominant inhibitory factors. They attack the incoming virus directly and block its integration into the host genome. This situation also pertains to HIV-1 in mouse cells and represents true "intracellular immunity." Importantly, this host response is more rapid than either traditional innate or adaptive immunity and can prevent the establishment of the infection.
Recent advances in our understanding of intracellular immunity have identified two different proteins, the tripartite motif protein 5α (TRIM5α) [2] and the apolipoprotein B mRNA-editing enzyme catalytic-polypeptides 3B, 3F and 3G (APOBEC3B, APOBEC3F and APOBEC3G or A3B, A3F and A3G), which collectively inactivate several retroviruses including HIV-1, simian immunodeficiency virus (SIV), hepatitis B virus and some mouse mobile genetic elements [3-7]. This review highlights these recent developments and mentions briefly additional potential blocks to retroviral replication.
Interference and Restriction
Let us begin with some definitions and historical perspectives. Viral "interference" refers to the situation when cells, which are chronically infected with one virus or contain endogenous retroviruses, resist superinfection by other viruses bearing envelopes with a similar target specificity. This block usually results from the loss of the appropriate receptor on the cell surface. A good example of this interference is the Friend virus susceptibility factor 4 (Fv4), also known as Akvr-1, which controls the susceptibility of mice to infection by ecotropic but not other murine leukemia viruses (MLVs) [8]. This gene is located on mouse chromosome 12 [9] within an endogenous defective provirus and encodes a complete envelope [10] that shares very high sequence similarity with those from ecotropic Cas-Br-E virus and Moloney MLV [11]. This envelope then blocks the expression of the cationic amino acid transporter, which is the receptor for these MLVs, on the cell surface (Fig. 1) [12]. Of interest, MLV can only use the murine but not the human form of this receptor for entry.
Figure 1 Effective intracellular immunity targets incoming viruses. Whereas the ecotropic murine leukemia virus (MLV) (represented as a viral particle in blue) encounters Fv4 and Fv1 blocks, other retroviruses such as HIV-1, SIV and EIAV (represented as a viral particle in brown) encounter TRIM5α and cytidine deaminase blocks. Fv4 prevents the entry by ecotropic MLV by sequestering the viral receptor from the cell surface. Fv1 targets MLV CA and stops the nuclear import of the viral preintegration complex (PIC). TRIM5α also targets retroviral CA and blocks uncoating. hA3B, hA3F and hA3G deaminate cytidines on newly synthesized retroviral cDNA and disrupt viral replication. Capital red letters highlight the points of inhibition. Viral structural components, nucleic acids, RNA and DNA, and intracellular events are represented in different colors.
The term "restriction" refers to intracellular blocks to viral replication. Until now, the best example has been Fv1 [13]. Like Fv4 and the less well-characterized Fv3 and Fv2, Fv1 also confers resistance of mice to the infection by MLV (Figs. 1 and 2). The Fv1 gene is located on mouse chromosome 4 [14] and encodes a protein that resembles other endogenous retroviral structural group specific antigens (Gag) (Fig. 2) [15]. Of note, during the morphogenesis and release of progeny virions, retroviral Gag polyproteins are processed by the viral protease into distinct subunits, namely matrix (MA), capsid (CA) and nucleocapsid (NC). Whereas MA and CA form the outer shell and inner core of mature viral particles, NC packages viral genomic RNA into the core [16]. After entry and uncoating in newly infected cells, many structural proteins remain associated with viral enzymes (reverse transcriptase, RT and integrase, IN) and RNA in a large (2 mDa) preintegration complex (PIC).
Figure 2 Fv1 block in MLV infection. Ecotropic MLVs (e.g. Friend MLV) fall into two categories with respect to their host range: N-tropic strains infect NIH/Swiss mice (brown) much more efficiently than BALB/c mice (black), whereas B-tropic strains display the opposite preference. Based on their susceptibility to N- or B-tropic virus, mice were classified into Fv1n/n (NIH/Swiss) and Fv1b/b (Balb/c) strains. These viruses differ at a single residue at position 110 in CA (presented below the NIH/Swiss mouse). These changes are matched by residues at position 358 in Fv1n and Fv1b proteins (presented below the Balb/c mouse).
Alleles of Fv1 in Balb/c (Fv1b/b) and NIH/Swiss (Fv1n/n) mice result in resistance to N- and B-tropic strains of MLV, respectively, which maps to position 110 in CA (Fig. 2) [17]. A recent structural analysis revealed that this residue is located at the outer face of the core structure of CA with easy access to cellular proteins [18]. On Fv1, the key residue for this restriction was mapped to position 358 (Fig. 2) [19]. Although binding between CA and Fv1 has not been demonstrated, they could interact as higher order structures, especially since CA and Gag form oligomers, in the case of CA, hexagonal lattices of the viral core. As heterozygous Fv1n/b mice block infection by both viruses, resistance is dominant [20]. Conversely, NB-tropic MLV can infect all these mice. Of interest, this restriction is saturable with high levels of CA from either virus [21], implying that amounts of Fv1 or its cofactor/s are limiting. As described below, one of these cofactors could be TRIM5α [22]. As a result of these interactions, Fv1 is thought to block the disassembly of CA and the normal movement of the PIC into the nucleus (Fig. 1) [23,24].
TRIM5α
Fv1 is not the only genetic system conferring intracellular immunity against a retroviral infection. For example, the replication of N-tropic MLV and the equine infectious anemia virus (EIAV, a lentivirus) is also inhibited in human cells [25], as is that of the primate lentiviruses HIV-1 and SIV from rhesus macaques (SIVmac) in cells from different monkeys (Fig. 3) [26-29]. For example, HIV-1 does not grow in old world monkeys, which include African green monkeys and rhesus macaques, and SIVmac does not infect new world monkeys, which include squirrel monkeys and common marmosets (Fig. 3) [26-29].
Figure 3 Blocks to retroviral replication by different TRIM5α proteins from several species. The replication of HIV-1 is blocked by TRIM5α from old world monkeys and owl monkeys, but not human and new world monkeys (left top panel). The replication of SIV is blocked by TRIM5α proteins from new world monkeys, but not from humans, old world monkeys, and owl monkeys (right top panel). The replication of N-MLV is prevented by TRIM5α proteins from dogs, pigs, cows, old world monkeys, and humans, but not mice (left bottom panel). The replication of EIAV is blocked by TRIM5α from human and old world monkeys, but not horses (right bottom panel). In all cases, arrows indicate no inhibition.
Interestingly, these blocks resemble Fv1 restriction in several ways. First, viral replication is impaired at the step of reverse transcription [25-28]. Second, CA is also targeted. The residue at position 110 in CA also determines the restriction of N-tropic MLV in human cells [25] and that of HIV-1 in rhesus macaque cells is abrogated when its CA is replaced by that from SIVmac [29]. Third, because heterokaryons between non-restrictive and restrictive cells maintain the inhibition, this restriction is dominant [27,28]. Finally, these blocks are saturable. However, since no Fv1-related gene could be found in primate cells, blocks to N-tropic MLV and EIAV in human cells were thought to be due to the restriction factor 1 (Ref1) [25], and those to HIV-1 and SIV in monkey cells to the lentivirus susceptibility factor 1 (Lv1) (Fig. 3) [27].
Indeed, Ref1 and Lv1 share additional similarities in blocking retroviral replication. For examples, both restrictions can be attenuated by chemicals that disrupt the integrity of mitochondrial membranes [30,31], and they can be saturated by the same virus like particles (VLPs) [32]. Using a functional complementation assay, Lv1 was first identified as the rhesus macaque TRIM5α (macTRIM5α) gene [2]. Later, by eliminating TRIM5α transcripts from old world monkey and human cells with small interfering RNA (siRNA), the Lv1 and Ref1 blocks were also abrogated [33]. Further studies revealed that hTRIM5α (from humans), macTRIM5α and agmTRIM5α (from African green monkeys) restrict the replication of different viruses, which were assigned previously to Lv1 and Ref1 (Fig. 3) [22,33-35]. Thus, Ref1 and Lv1 are species-specific variants of TRIM5α.
The hTRIM5α protein contains 493 residues (Fig. 4) and belongs to the large tripartite motif (TRIM) family that consists of 37 genes, which include the promyelocytic leukemia (PML or TRIM19) protein [36]. By alternative RNA splicing, they produce 71 different transcripts. For example, the human TRIM5 gene is expressed as hTRIM5α, β, γ, σ, ε, and ζ. Although little is known of their function, they contain three distinctive structural motifs, a RING Zn++ finger, one or two B-box Zn++ finger, and an α-helical coiled-coil (CC) region (Fig. 4). For this reason, they are also called the RING finger:B box:Coiled-coil (RBCC) family proteins. The RING finger motif features a cysteine-rich consensus, which contains two interleaved Zn++-binding sites [37]. Many RING finger proteins act as E3 ubiquitin ligases and play key roles in protein degradation. For example, Ring-box-1 (Rbx1) is an essential component of the Skp1:cullin-1:F-box (SCF) complex. Additionally, TRIM5σ displays E3 ligase activity in vitro [38]. B-boxes, which consist of one Zn++-binding site and a B1 or B2 motif [39], orient the CC motif that mediates protein-protein interactions. Indeed, TRIM proteins form oligomers [36]. In addition, TRIM5α contains a SPRY domain at its C-terminus (Fig. 4). The SPRY domain was originally identified in the splA kinase of Dictyostelium and the rabbit ryanodine receptor [40], and belongs to the subclass of the B30.2 or RFP-like domains. In butyrophilin, the B30.2 domain, which contains 170 residues, is involved in ligand binding [41]. Of interest, TRIM proteins localize to particular cellular compartments where they form discrete structures. Whereas TRIM19 assembles discrete PML oncogenic domains (PODs) in the nucleus, TRIM5α can form cytoplasmic bodies [36].
Figure 4 Schematic representations of hTRIM5α, macTRIM5α, and omTRIM5a. CypA proteins. hTRIM5α contains 493 residues and four conserved motifs, whose positions are given. They are the RING domain, B box, coiled-coil and SPRY domains. The latter domain is required for species-specific restriction of primate lentiviruses and is diagrammed in red. A key residue in this domain, which is the arginine at position 332 in hTRIM5α, or the proline at position 334 in macTRIM5α, is responsible for its species-specific inhibition of lentiviral replication. omTRIM5α from owl monkeys contains an N-terminal omTRIM5α sequence to position 299, linked in-frame to the entire CypA gene (147 residues)(omTRIM5α.CypA).
Although they share 87% sequence similarity, only macTRIM5α but not hTRIM5α blocks HIV-1 replication [2]. This species-specific restriction was mapped to the SPRY domain [42-44]. Through genetic analysis, it was revealed that this SPRY domain has experienced dramatic mutations during primate evolution [45] and contains four variable regions V1, V2, V3, and V4 [46]. The change of a single residue (R332P) in V1 abolished the inhibition of HIV-1 replication by hTRIM5α [43,44], which suggests that the SPRY domain is responsible for its targeting of CA. Although no binding between the SPRY of rhTRIM5α and HIV-1 CA has been demonstrated, the findings with TRIM5α from owl monkeys (omTRIM5α) support such direct interactions [47,48]. A rather complicated story follows. Cyclophilin A (CypA) is an eukaryotic peptidyl-prolyl cis-trans-isomerase. It binds an exposed proline-rich loop in CA of HIV-1 and is critical for its replication in human cells [49,50]. In contrast, the ability to bind CypA restricts HIV-1 replication in owl monkey cells. Owl monkeys are atypical new world monkeys because their Lv1 inhibits HIV-1 but not SIVmac. Although this block to HIV-1 replication is abrogated when the interaction between CA and CypA is prevented by mutations in CA or by cyclosporin A treatment in owl monkey cells, the same manipulations increase effects of Ref1 on HIV-1 in human cells [51]. The explanation for these differences came with the cloning of the omTRIM5α gene. Instead of the SPRY domain, it contains the complete CypA gene [47,48] (Fig. 4). Thus, owl monkey cells express a fusion protein between omTRIM5α and CypA (omTRIM5α.Cyp), which most likely arose from a retrotransposition of the CypA gene into the omTRIM5α locus by the long interspersed nuclear elements-1 (LINE-1 or L1). In conclusion, CA and omTRIM5α interact via this CypA domain and restrict HIV-1 replication in owl monkeys.
These studies suggest that Fv1 and TRIM5α might interact directly with CA to block incoming viruses. However, in contrast to Fv1, which blocks nuclear entry and integration of the provirus [23,24], TRIM5α inhibits viral replication at a step before reverse transcription (Fig. 1). It is puzzling why such differences exist. An answer might lie in the observation that MLV, but not HIV-1, retains its CA in the reverse transcription complex [52,53]. Thus, the uncoating of HIV-1 could proceed much faster than that of MLV. Once the core structure is destroyed, the reverse transcription complex could become more susceptible to TRIM5α. TRIM5α could then trigger the proteasomal degradation of PIC. This model also offers an explanation of the enhancement of viral replication when target cells are treated with proteasomal inhibitors [54]. Further details await studies of other proteins that interact with Fv1 and TRIM5α, their enzymatic properties and trafficking in cells.
Cytidine deaminases
In addition to Fv1 and TRIM5α, host cells have developed additional mechanisms to protect themselves from viral invasion. The next important block involves nucleic acid editing of viral reverse transcripts. For a long time, it had been noted that retroviruses contain a high frequency of G to A transitions [55,56]. In certain strains of HIV-1, up to 60% of all guanidines are replaced by adenines [57]. Previously, this G to A hypermutation was attributed to the high error rate of reverse transcriptase and the imbalance in dCTP pools in cells [58]. However, we now know that host cellular cytidine deaminases are responsible.
In parallel, mutant HIV-1 lacking the viral infectivity factor (Vif) (HIV-1ΔVif) can only replicate in certain T cell lines, which are called "permissive" cells. In other "non-permissive" cells, only wild type HIV-1 but not HIV-1ΔVif can replicate [59,60]. Because heterokaryons between permissive and non-permissive cells do not support the replication of HIV-1ΔVif, there exists a dominant inhibitor in these non-permissive cells [61,62]. By subtractive cloning between non-permissive CEM and permissive CEM-SS T cells, the inhibitory factor was identified as the human A3G (hA3G) protein [3]. Later, its close relatives hA3F, and to a lesser degree, hA3B, were found to possess similar anti-viral activities [4-7].
The human APOBEC family comprises 10 proteins, among which are the founding member APOBEC1 (A1) and the activation induced deaminase (AID) [63]. They contain one (e.g. APOBEC1 and AID) or two (e.g. hA3F and hA3G) Zn++-binding deaminase motifs with the consensus sequence His-X-Glu-X23–28-Pro-Cys-X2–4-Cys (where X denotes any amino acid) [63]. They can target cytosines and convert them to uracils (C to U transitions) on DNA or RNA (e.g. A1) templates. During the second-strand DNA synthesis, these C to U transitions are then converted to those of G to A. For example, by changing C6666 to U6666, A1 introduces a stop codon at position 6666 into the apolipoprotein B100 mRNA, which is translated into the truncated apolipoprotein B48 (48 kDa) protein [64]. AID also directs the cytidine deamination at specific "hot spots" to direct somatic hypermutation and isotype class switching in B cells [65]. hA3F and hA3G block retroviral infection in hematopoietic cells. They share overall 70% sequence similarity and form homodimers as well as mixed oligomers [5]. Physiological functions of these proteins are not yet defined, except that hA3G also inhibits the movement of some mouse mobile genetic elements in cells [66]. Thus, they could contribute to the stability of the genome.
APOBEC proteins and viral replication
The mechanisms for antiviral activities of APOBEC proteins have been characterized extensively. In the absence of Vif, hA3F and hA3G are incorporated into virions. They are then transferred from producer to target cells by the virus. Following viral entry and uncoating, reverse transcription is initiated and viral minus-strand cDNA is synthesized. During this process, these APOBEC proteins attack newly synthesized minus-strand cDNAs and introduce C to U transitions [67-70], which block viral replication by several mechanisms [71]. First, since uracils are not tolerated in DNA, they are removed by uracil N-glycosidases (UNG) from DNA and these nicked DNA are further cleaved by the host DNA-repair enzymes like apurinic/apyrimidinic endonuclease-1 (APE1). Fragmented DNA neither integrates nor replicates. Second, should edited proviruses survive and integrate, the new G to A changes on the plus strand DNA also create havoc on viral transcripts. These changes could lead to alternate splicing and the production of nonfunctional proteins. To these ends, hA3G and hA3F have different sequence preferences. Whereas hA3G favors repeated deoxycytidines (GG on the opposite strand) [72], hA3F prefers deoxycytidines followed by deoxythymidines (GA on the opposite strand) [4,7].
The key step for the anti-viral activity of hA3G is its incorporation into virions. Although the encapsidation of hA3G requires NC of HIV-1 Gag, it is still controversial whether this interaction is mediated by RNA [73-76]. Since both NC and hA3G can bind RNA, this recruitment most likely reflects RNA-protein, as well as protein-protein, interactions [77,78]. Nevertheless, since hA3G blocks the replication of all primate lentiviruses in the absence of Vif (HIV-1, HIV-2, and SIV) [3,79,80], EIAV [69], HBV [81,82], and some mouse mobile genetic elements [66], these interactions must have broad specificities. For example, hA3F has the same effect against HIV-1, SIV and HBV and hA3B and hA3C block SIV (Table 1) [4-7,83,84]. Of interest, the rat but not human A1 proteins block HIV-1 by directly deaminating viral RNA [85].
Table 1 Abilities of APOBEC proteins to inhibit viruses and retrotransposons
Species APOBEC proteins HIV-1ΔVif HIV-2ΔVif SIVΔVif MLV EIAV HBV L1 IAP MusD
humans A3B + + -
A3C - + -
A3F + + +
A3G + + + + + + - +
old world monkeys A3G + + -
rats A1 + -
mice A3 + + - - +
(+) block, (-) do not block
Vif and APOBEC proteins
In contrast to HIV-1ΔVif, wild-type HIV-1 is not restricted in non-permissive cells. Thus, Vif counteracts the effects of hA3F and hA3G. Indeed, Vif binds and triggers the degradation of these APOBEC proteins in producer cells, thus blocking their incorporation into virions [86-88]. Initially, Vif was demonstrated to interact with cellular proteins Cul5, elonginB, elonginC, and Rbx1 to form a cullin-based E3 ubiquitin ligase complex [89], which displays striking similarities to SCF complex. Later, Vif was found to contain a conserved suppressor of cytokine signaling (SOCS) box-like motif (SLQ(Y/F)LA) that binds elongin C, which in turn recruits elongin B, cullin 5 and Rbx1, thus forming the ElonginB/C-Cul5-SOCS-box (ECS) E3 ubiquitin ligase complex [90,91]. As a consequence of these interactions, APOBEC proteins are polyubiquitylated and degraded [86-88]. In parallel, some groups observed that Vif triggers only a marginal degradation of hA3G, which suggested that Vif could sequester hA3G from encapsidation through a degradation-independent mechanism [79,92].
Although Vif blocks the antiviral activity of hA3G, its activity is highly species-specific (fig. 5). Additionally as presented in Table 2, Vif from HIV-1 blocks A3G proteins from humans and chimpanzees (hA3G and chA3G), but not from old world monkeys, Vif from SIVmac blocks all A3G isoforms from human and non-human primates, and Vif from SIVagm only blocks A3G proteins from monkeys [79]. In addition, hA3F can be inactivated by Vif proteins from HIV-1, HIV-2, and SIVmac, but not from SIVagm [4,5,84]. Moreover, the hA3C can be inactivated by Vif from SIVmac [84] and no Vif protein can inactivate the hA3B, rat A1, or mouse APOBEC3 (mA3) proteins [6]. Efforts have been made to uncover the molecular mechanisms of these species-specific differences, and agmA3G was chosen because it is most similar to hA3G. It was found that Vif from HIV-1 fails to inactivate agmA3G because it contains a lysine rather than aspartate at position 128 (K128D), which is found in hA3G. Although agmA3G neither binds Vif nor is excluded from virions, the mutant agmA3G protein bearing D128 becomes fully sensitive to Vif from HIV-1 [93-96]. In addition, the reciprocal exchange of D for K at position 128 in hA3G renders it resistant to Vif. Structural comparisons with the related cytidine deaminases from E.coli reveal that D128 maps to an α-helical turn on an exposed loop [96]. Since the same K128 residue also exists in A3G from rhesus macaque (macA3G), its sensitivity might also be altered with a similar K128D substitution. Although these studies established the correlation between the ability of Vif to neutralize APOBEC proteins and viral replication, it is unlikely that these species-specific susceptibilities of APOBEC proteins to Vif are responsible for the transmission of primate lentiviruses to new host species [97].
Figure 5 Species-specific inhibition of APOBEC3 (A3) proteins by Vif. HIV-1 and SIV can incorporate A3 proteins from all these species into virions. However, Vif from HIV-1 (VifHIV, brown), can only inhibit A3 proteins from humans and new world monkeys) but not those from old world monkeys and mice. In contrast, Vif from SIVmac (VifSIV, green) can inactivate A3 proteins from old and possibly new world monkeys, but not from humans and mice). In this drawing, stick figures represent sources of A3 proteins and not targets of infection.
Table 2 Species-specific susceptibility of APOBEC proteins to Vif
APOBEC proteins Species HIV-1 Vif HIV-2 Vif SIVagm Vif SIVmac Vif
A3G humans + + - +
chimpanzees + - +
African green monkeys - + +
rhesus macaques - + +
A3 mice - - -
A3F humans + + - +
A3B humans - - -
A3C humans - + +
A1 rats -
(+) susceptible, (-) not susceptible
Viruses that lack Vif and mobile genetic elements
How can viruses that do not encode Vif escape from this intracellular immunity?
As some APOBEC proteins also inhibit the replication of EIAV, HBV, and MLV, there must be additional mechanisms of escape. In the case of EIAV, it encodes an additional enzyme, which is named dUTPase [98]. This enzyme is also produced by herpesviruses, poxviruses, and some other retroviruses. Recently, it was demonstrated that dUTPase from caprine arthritis encephalitis virus (CAEV) could block the misincorporation of dUTP during HIV-1 reverse transcription [99]. Thus, its dUTPase could also protect EIAV from attack by APOBEC proteins. In the case of HBV, it replicates in tissues that do not express hA3G [63]. The situation for MLV is more complicated. Unlike human cells that express seven A3 proteins, the mouse genome contains only one A3 gene. Although mA3 blocks the replication of HIV-1 and SIV, it is less efficiently packaged into and does not inhibit MLV [100]. In contrast, hA3G blocks the replication of MLV. Thus, MLV has adapted to its natural host by a mechanism that remains poorly understood.
What is the situation with mobile genetic elements that resemble retroviruses? In humans and mice, there are several types of retrotransposons [101]. The most abundant are LINE-1 or L1 elements that do not contain long terminal repeats (LTRs). Up to one hundred human and several thousand mouse L1 elements are functional [101]. Although they require reverse transcription, they do not form VLPs and APOBEC proteins do not block their replication [66,102]. In contrast, LTR-containing retrotransposons, which represent up to 10% of the human genome, form VLPs, bud from intracellular organelles and behave similarly to incoming exogenous retroviruses. Although no active human endogenous retroviruses (HERVs) have been found, in the mouse, there are several hundred active intracisternal A-particles (IAPs) [103] and at least ten copies of MusD [104]. Indeed, sequences of IAPs and MusDs contain frequent G to A transversions in their genomes [66]. In addition, using transient expression assays in cells, hA3G and mA3 inhibit the retrotransposition of IAP and MusD (Table 1) [66]. Thus, APOBEC proteins also block the movement of some mobile genetic elements, most likely in germ cells and during embryogenesis, in mammals.
Other antiviral genes that contribute to intracellular immunity
Besides these predominant blocks to viral replication in cells, additional barriers have been described at levels of transcription and RNA stability, as well as assembly of progeny virions. However, since they do not block the integration of proviruses into the host genome, they play lesser roles in intracellular immunity. First, Murr1 blocks the activation of NF-κB in resting cells and thus the induction of HIV-1 replication [105]. Second, a sophisticated genetic screen looking for cells that survive attack by MuLV bearing the thymidine kinase (tk) gene (which would otherwise succumb to trifluorothymidine that is phosphorylated by tk) revealed the Zn++-finger antiviral protein ZAP that degrades rapidly MLV transcripts [106]. ZAP binds a specific sequence at the 3' end of viral, but not cellular, transcripts and leads to their rapid degradation in the exosome. This mechanism appears analogous to tristetraprolin, which binds AU-rich RNA species (e.g. those coding for cytokine genes) and targets them for rapid degradation in the cytoplasm. Apparently, not only are retroviral transcripts targeted by ZAP, but it destroys Ross River, Semliki, Sindbis and Venezualan equine encephalitis viruses, all of which belong to the alphavirus family [107]. Alghough ZAP is extremely efficient againt alphaviruses and MuLV, it is not clear what role, if any, it plays against primate retroviruses.
The final level of intracellular immunity deals with viral assembly and release. Again, HIV-1 encodes another accessory viral protein u (Vpu), which facilitates the release of progeny virions from infected cells [108]. Thus, analogous to the situation with Vif, some cells are "permissive" and others are "non-permissive" for viral replication in the absence of Vpu. Heterokaryons between them maintain the non-permissive phenotype, which is dominant. Thus, Vpu must counteract some dominant negative cellular factor, whose identity remains to be determined. Of interest, recent work suggests that Vpu counteracts the two-pore K+ (K2P) channel TASK-1, which inhibits the release of many viruses by an unknown mechanism, possibly by changing membrane fluidity [109]. Vpu also facilitates the release of other retroviruses. By mimicking a natural component of TASK-1, Vpu is incorporated into the channel, where it acts as a dominant negative effector. Vpu also binds βTRCP, an E3 ubiquitin ligase, which could accelerate the degradation of TASK-1 in the proteasome [110]. Thus, it is possible that levels and/or polymorphisms of TASK-1 are mostly responsible for this block in the assembly and release of progeny virions. However, additional experiments are required to make this connection and/or to reveal additional players in this last step of the viral replicative cycle in cells.
Intracellular immunity
Several themes emerge from these cell-intrinsic blocks to retroviral replication. First, the inhibition is broad. Thus, not only are retroviruses targeted, but other viruses as well, from HBV and alphaviruses to some mobile genetic elements, which once were viruses themselves. Second, multiple steps in the replicative cycles of these viruses are inhibited, most likely because each mechanism is not completely effective. This finding might reflect small differences between extracellular pathogens and normal cellular homeostatic mechanisms. Alternatively, it might reflect the vast spectrum of different pathogens, all of which must be targeted and destroyed. For retroviruses, the challenge is increased because of their rapid rate of mutations and their quick adaptation to the host. Third, these intracellular blocks are more pronounced and effective in zoonotic infections, where the virus jumps species. Finally, this inhibition is rapid and targets predominantly early steps in the replicative cycles of these viruses. Thus, it tries to prevent the integration of the viral genetic material into the host genome. Whether these inhibitors accomplish this task by targeting viral structures or genetic material to an endosome, exosome or proteasome, the end results are the same, i.e. the elimination of the virus. In this scenario, the outcome depends on the effectiveness of cellular proteins versus the protective armor of the virus.
Given these observations, one of the simplest new therapeutic interventions could be simply to increase intracellular levels of these antiviral proteins, e.g. TRIM5α, APOBEC proteins, ZAP and/or TASK-1. Thus, if we only understood their normal regulation, it is possible that we could augment their amounts and activities during active infections. Of course, as we do not know their other functions in cells, there are also many potential concerns. For example, would increased levels of APOBEC3G cause editing of genomic DNA during replication, thus facilitating oncogenic transformation? Likewise, would increased amounts of ZAP target critical cellular transcripts for accelerated degradation? Alternatively, one could try to block interactions between Vif and APOBEC proteins and TASK-1 and Vpu. Possibly, by studying their structures, one could design inhibitors for their protein-protein interactions. Moreover, all these processes can also be targeted by gene therapy, by introducing into cells their counterparts from different species and/or by changing binding surfaces of the host proteins so that they no longer interact with Vif or Vpu, for example. If not practical clinically, such genetic manipulation would yield important clues as to which restriction should be targeted by other therapeutic means.
Conclusion
It is remarkable how active are the processes that protect an organism from internal and external challenges. In humans, at least three layers of immunity have developed. Among them, intracellular and innate immune responses act primarily via pattern recognition, whereas adaptive immunity is very sequence and peptide-specific. Nevertheless, many pathogens break through and are integrated into the host genetic material. Thus, some of these defense mechanism must also survey the movements and effects of these mobile genetic elements. It appears that a fine line has been drawn between control and allowing for some escape as well. As mobile genetic elements contribute to evolution and fitness of all species, they must be kept in check, but not eliminated completely and it is possible that this intricate regulation of hA3G activity in cells reflects this requirement. On the other hand, there is also a high price to pay in terms of mistakes, be they developmental defects or cancer. Nevertheless, the study of these systems that fight extracellular pathogens is likely to reveal fundamental insights into a plethora of cellular processes that contribute to human health and disease.
Acknowledgements
We thank Bryan Cullen, Warner Greene, Lewis Lanier, Nika Lovsin, Peter Pesic and Olivier Schwartz for helpful comments on the manuscript. This work was supported by grants from the NIH (RO1 AI49104, RO1 A151165, P01 AI058708).
==== Refs
Gardner MB Luciw PA Animal models of AIDS Faseb J 1989 3 2593 2606 2556312
Stremlau M Owens CM Perron MJ Kiessling M Autissier P Sodroski J The cytoplasmic body component TRIM5alpha restricts HIV-1 infection in Old World monkeys Nature 2004 427 848 853 14985764 10.1038/nature02343
Sheehy AM Gaddis NC Choi JD Malim MH Isolation of a human gene that inhibits HIV-1 infection and is suppressed by the viral Vif protein Nature 2002 418 646 650 12167863 10.1038/nature00939
Zheng YH Irwin D Kurosu T Tokunaga K Sata T Peterlin BM Human APOBEC3F is another host factor that blocks human immunodeficiency virus type 1 replication J Virol 2004 78 6073 6076 15141007 10.1128/JVI.78.11.6073-6076.2004
Wiegand HL Doehle BP Bogerd HP Cullen BR A second human antiretroviral factor, APOBEC3F, is suppressed by the HIV-1 and HIV-2 Vif proteins Embo J 2004 23 2451 2458 15152192 10.1038/sj.emboj.7600246
Bishop KN Holmes RK Sheehy AM Davidson NO Cho SJ Malim MH Cytidine Deamination of Retroviral DNA by Diverse APOBEC Proteins Curr Biol 2004 14 1392 1396 15296758 10.1016/j.cub.2004.06.057
Liddament MT Brown WL Schumacher AJ Harris RS APOBEC3F Properties and Hypermutation Preferences Indicate Activity against HIV-1 In Vivo Curr Biol 2004 14 1385 1391 15296757 10.1016/j.cub.2004.06.050
Suzuki S FV-4: a new gene affecting the splenomegaly induction by Friend leukemia virus Jpn J Exp Med 1975 45 473 478 1232464
Ikeda H Sato H Odaka T Mapping of the Fv-4 mouse gene controlling resistance to murine leukemia viruses Int J Cancer 1981 28 237 240 6797957
Ikeda H Laigret F Martin MA Repaske R Characterization of a molecularly cloned retroviral sequence associated with Fv-4 resistance J Virol 1985 55 768 777 2991595
Masuda M Yoshikura H Construction and characterization of the recombinant Moloney murine leukemia viruses bearing the mouse Fv-4 env gene J Virol 1990 64 1033 1043 2304138
Taylor GM Gao Y Sanders DA Fv-4: identification of the defect in Env and the mechanism of resistance to ecotropic murine leukemia virus J Virol 2001 75 11244 11248 11602766 10.1128/JVI.75.22.11244-11248.2001
Lilly F Susceptibility to two strains of Friend leukemia virus in mice Science 1967 155 461 462 6015694
Rowe WP Humphrey JB Lilly F A major genetic locus affecting resistance to infection with murine leukemia viruses. 3. Assignment of the Fv-1 locus to linkage group 8 of the mouse J Exp Med 1973 137 850 853 4689339 10.1084/jem.137.3.850
Best S Le Tissier P Towers G Stoye JP Positional cloning of the mouse retrovirus restriction gene Fv1 Nature 1996 382 826 829 8752279 10.1038/382826a0
Freed EO HIV-1 gag proteins: diverse functions in the virus life cycle Virology 1998 251 1 15 9813197 10.1006/viro.1998.9398
Kozak CA Chakraborti A Single amino acid changes in the murine leukemia virus capsid protein gene define the target of Fv1 resistance Virology 1996 225 300 305 8918916 10.1006/viro.1996.0604
Mortuza GB Haire LF Stevens A Smerdon SJ Stoye JP Taylor IA High-resolution structure of a retroviral capsid hexameric amino-terminal domain Nature 2004 431 481 485 15386017 10.1038/nature02915
Bishop KN Bock M Towers G Stoye JP Identification of the regions of Fv1 necessary for murine leukemia virus restriction J Virol 2001 75 5182 5188 11333899 10.1128/JVI.75.11.5182-5188.2001
Pincus T Rowe WP Lilly F A major genetic locus affecting resistance to infection with murine leukemia viruses. II. Apparent identity to a major locus described for resistance to friend murine leukemia virus J Exp Med 1971 133 1234 1241 4325133 10.1084/jem.133.6.1234
Pincus T Hartley JW Rowe WP A major genetic locus affecting resistance to infection with murine leukemia viruses. IV. Dose-response relationships in Fv-1-sensitive and resistant cell cultures Virology 1975 65 333 342 165617 10.1016/0042-6822(75)90039-2
Keckesova Z Ylinen LM Towers GJ The human and African green monkey TRIM5alpha genes encode Ref1 and Lv1 retroviral restriction factor activities Proc Natl Acad Sci U S A 2004 101 10780 10785 15249687 10.1073/pnas.0402474101
Jolicoeur P Rassart E Effect of Fv-1 gene product on synthesis of linear and supercoiled viral DNA in cells infected with murine leukemia virus J Virol 1980 33 183 195 6245227
Yang WK Kiggans JO Yang DM Ou CY Tennant RW Brown A Bassin RH Synthesis and circularization of N- and B-tropic retroviral DNA Fv-1 permissive and restrictive mouse cells Proc Natl Acad Sci U S A 1980 77 2994 2998 6248878
Towers G Bock M Martin S Takeuchi Y Stoye JP Danos O A conserved mechanism of retrovirus restriction in mammals Proc Natl Acad Sci U S A 2000 97 12295 12299 11027299 10.1073/pnas.200286297
Besnier C Takeuchi Y Towers G Restriction of lentivirus in monkeys Proc Natl Acad Sci U S A 2002 99 11920 11925 12154231 10.1073/pnas.172384599
Cowan S Hatziioannou T Cunningham T Muesing MA Gottlinger HG Bieniasz PD Cellular inhibitors with Fv1-like activity restrict human and simian immunodeficiency virus tropism Proc Natl Acad Sci U S A 2002 99 11914 11919 12154227 10.1073/pnas.162299499
Munk C Brandt SM Lucero G Landau NR A dominant block to HIV-1 replication at reverse transcription in simian cells Proc Natl Acad Sci U S A 2002 99 13843 13848 12368468 10.1073/pnas.212400099
Owens CM Yang PC Gottlinger H Sodroski J Human and simian immunodeficiency virus capsid proteins are major viral determinants of early, postentry replication blocks in simian cells J Virol 2003 77 726 731 12477877 10.1128/JVI.77.1.726-731.2003
Berthoux L Sebastian S Sokolskaja E Luban J Lv1 inhibition of human immunodeficiency virus type 1 is counteracted by factors that stimulate synthesis or nuclear translocation of viral cDNA J Virol 2004 78 11739 11750 15479815 10.1128/JVI.78.21.11739-11750.2004
Berthoux L Towers GJ Gurer C Salomoni P Pandolfi PP Luban J As(2)O(3) enhances retroviral reverse transcription and counteracts Ref1 antiviral activity J Virol 2003 77 3167 3180 12584341 10.1128/JVI.77.5.3167-3180.2003
Hatziioannou T Cowan S Goff SP Bieniasz PD Towers GJ Restriction of multiple divergent retroviruses by Lv1 and Ref1 Embo J 2003 22 385 394 12554640 10.1093/emboj/cdg042
Hatziioannou T Perez-Caballero D Yang A Cowan S Bieniasz PD Retrovirus resistance factors Ref1 and Lv1 are species-specific variants of TRIM5alpha Proc Natl Acad Sci U S A 2004 101 10774 10779 15249685 10.1073/pnas.0402361101
Perron MJ Stremlau M Song B Ulm W Mulligan RC Sodroski J TRIM5alpha mediates the postentry block to N-tropic murine leukemia viruses in human cells Proc Natl Acad Sci U S A 2004 101 11827 11832 15280539 10.1073/pnas.0403364101
Yap MW Nisole S Lynch C Stoye JP Trim5alpha protein restricts both HIV-1 and murine leukemia virus Proc Natl Acad Sci U S A 2004 101 10786 10791 15249690 10.1073/pnas.0402876101
Reymond A Meroni G Fantozzi A Merla G Cairo S Luzi L Riganelli D Zanaria E Messali S Cainarca S Guffanti A Minucci S Pelicci PG Ballabio A The tripartite motif family identifies cell compartments Embo J 2001 20 2140 2151 11331580 10.1093/emboj/20.9.2140
Joazeiro CA Weissman AM RING finger proteins: mediators of ubiquitin ligase activity Cell 2000 102 549 552 11007473 10.1016/S0092-8674(00)00077-5
Xu L Yang L Moitra PK Hashimoto K Rallabhandi P Kaul S Meroni G Jensen JP Weissman AM D'Arpa P BTBD1 and BTBD2 colocalize to cytoplasmic bodies with the RBCC/tripartite motif protein, TRIM5delta Exp Cell Res 2003 288 84 93 12878161 10.1016/S0014-4827(03)00187-3
Torok M Etkin LD Two B or not two B? Overview of the rapidly expanding B-box family of proteins Differentiation 2001 67 63 71 11428128 10.1046/j.1432-0436.2001.067003063.x
Ponting C Schultz J Bork P SPRY domains in ryanodine receptors (Ca(2+)-release channels). Trends Biochem Sciences 1997 22 193 194 10.1016/S0968-0004(97)01049-9
Jack LJ Mather IH Cloning and analysis of cDNA encoding bovine butyrophilin, an apical glycoprotein expressed in mammary tissue and secreted in association with the milk-fat globule membrane during lactation J Biol Chem 1990 265 14481 14486 2387867
Sawyer SL Wu LI Emerman M Malik HS Positive selection of primate TRIM5alpha identifies a critical species-specific retroviral restriction domain Proc Natl Acad Sci U S A 2005 102 2832 2837 15689398 10.1073/pnas.0409853102
Stremlau M Perron M Welikala S Sodroski J Species-specific variation in the B30.2(SPRY) domain of TRIM5alpha determines the potency of human immunodeficiency virus restriction J Virol 2005 79 3139 3145 15709033 10.1128/JVI.79.5.3139-3145.2005
Yap MW Nisole S Stoye JP A single amino acid change in the SPRY domain of human Trim5alpha leads to HIV-1 restriction Curr Biol 2005 15 73 78 15649369 10.1016/j.cub.2004.12.042
Sawyer SL Emerman M Malik HS Ancient adaptive evolution of the primate antiviral DNA-editing enzyme APOBEC3G PLoS Biol 2004 2 E275 15269786 10.1371/journal.pbio.0020275
Song B Javanbakht H Perron M Park do H Stremlau M Sodroski J Retrovirus Restriction by TRIM5{alpha} Variants from Old World and New World Primates J Virol 2005 79 3930 3937 15767395 10.1128/JVI.79.7.3930-3937.2005
Nisole S Lynch C Stoye JP Yap MW A Trim5-cyclophilin A fusion protein found in owl monkey kidney cells can restrict HIV-1 Proc Natl Acad Sci U S A 2004 101 13324 13328 15326303 10.1073/pnas.0404640101
Sayah DM Sokolskaja E Berthoux L Luban J Cyclophilin A retrotransposition into TRIM5 explains owl monkey resistance to HIV-1 Nature 2004 430 569 573 15243629 10.1038/nature02777
Thali M Bukovsky A Kondo E Rosenwirth B Walsh CT Sodroski J Gottlinger HG Functional association of cyclophilin A with HIV-1 virions Nature 1994 372 363 365 7969495 10.1038/372363a0
Franke EK Yuan HE Luban J Specific incorporation of cyclophilin A into HIV-1 virions Nature 1994 372 359 362 7969494 10.1038/372359a0
Towers GJ Hatziioannou T Cowan S Goff SP Luban J Bieniasz PD Cyclophilin A modulates the sensitivity of HIV-1 to host restriction factors Nat Med 2003 9 1138 1143 12897779 10.1038/nm910
Fassati A Goff SP Characterization of intracellular reverse transcription complexes of human immunodeficiency virus type 1 J Virol 2001 75 3626 3635 11264352 10.1128/JVI.75.8.3626-3635.2001
Fassati A Goff SP Characterization of intracellular reverse transcription complexes of Moloney murine leukemia virus J Virol 1999 73 8919 8925 10515996
Schwartz O Marechal V Friguet B Arenzana-Seisdedos F Heard JM Antiviral activity of the proteasome on incoming human immunodeficiency virus type 1 J Virol 1998 72 3845 3850 9557668
Pathak VK Temin HM Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: substitutions, frameshifts, and hypermutations Proc Natl Acad Sci U S A 1990 87 6019 6023 2201018
Pathak VK Temin HM Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: deletions and deletions with insertions Proc Natl Acad Sci U S A 1990 87 6024 6028 2166940
Vartanian JP Henry M Wain-Hobson S Sustained G-->A hypermutation during reverse transcription of an entire human immunodeficiency virus type 1 strain Vau group O genome J Gen Virol 2002 83 801 805 11907329
Vartanian JP Meyerhans A Sala M Wain-Hobson S G-->A hypermutation of the human immunodeficiency virus type 1 genome: evidence for dCTP pool imbalance during reverse transcription Proc Natl Acad Sci U S A 1994 91 3092 3096 7512722
Gabuzda DH Lawrence K Langhoff E Terwilliger E Dorfman T Haseltine WA Sodroski J Role of vif in replication of human immunodeficiency virus type 1 in CD4+ T lymphocytes J Virol 1992 66 6489 6495 1357189
von Schwedler U Song J Aiken C Trono D Vif is crucial for human immunodeficiency virus type 1 proviral DNA synthesis in infected cells J Virol 1993 67 4945 4955 8331734
Madani N Kabat D An endogenous inhibitor of human immunodeficiency virus in human lymphocytes is overcome by the viral Vif protein J Virol 1998 72 10251 10255 9811770
Simon JH Gaddis NC Fouchier RA Malim MH Evidence for a newly discovered cellular anti-HIV-1 phenotype Nat Med 1998 4 1397 1400 9846577 10.1038/3987
Jarmuz A Chester A Bayliss J Gisbourne J Dunham I Scott J Navaratnam N An anthropoid-specific locus of orphan C to U RNA-editing enzymes on chromosome 22 Genomics 2002 79 285 296 11863358 10.1006/geno.2002.6718
Teng B Burant CF Davidson NO Molecular cloning of an apolipoprotein B messenger RNA editing protein Science 1993 260 1816 1819 8511591
Muramatsu M Kinoshita K Fagarasan S Yamada S Shinkai Y Honjo T Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme Cell 2000 102 553 563 11007474 10.1016/S0092-8674(00)00078-7
Esnault C Heidmann O Delebecque F Dewannieux M Ribet D Hance AJ Heidmann T Schwartz O APOBEC3G cytidine deaminase inhibits retrotransposition of endogenous retroviruses Nature 2005 433 430 433 15674295 10.1038/nature03238
Harris RS Bishop KN Sheehy AM Craig HM Petersen-Mahrt SK Watt IN Neuberger MS Malim MH DNA deamination mediates innate immunity to retroviral infection Cell 2003 113 803 809 12809610 10.1016/S0092-8674(03)00423-9
Lecossier D Bouchonnet F Clavel F Hance AJ Hypermutation of HIV-1 DNA in the absence of the Vif protein Science 2003 300 1112 12750511 10.1126/science.1083338
Mangeat B Turelli P Caron G Friedli M Perrin L Trono D Broad antiretroviral defence by human APOBEC3G through lethal editing of nascent reverse transcripts Nature 2003 424 99 103 12808466 10.1038/nature01709
Zhang H Yang B Pomerantz RJ Zhang C Arunachalam SC Gao L The cytidine deaminase CEM15 induces hypermutation in newly synthesized HIV-1 DNA Nature 2003 424 94 98 12808465 10.1038/nature01707
Gu Y Sundquist WI Good to CU Nature 2003 424 21 22 12840737 10.1038/424021a
Yu Q Konig R Pillai S Chiles K Kearney M Palmer S Richman D Coffin JM Landau NR Single-strand specificity of APOBEC3G accounts for minus-strand deamination of the HIV genome Nat Struct Mol Biol 2004 11 435 442 15098018 10.1038/nsmb758
Alce TM Popik W APOBEC3G is incorporated into virus-like particles by a direct interaction with HIV-1 Gag nucleocapsid protein J Biol Chem 2004 279 34083 34086 15215254 10.1074/jbc.C400235200
Cen S Guo F Niu M Saadatmand J Deflassieux J Kleiman L The interaction between HIV-1 Gag and APOBEC3G J Biol Chem 2004 279 33177 33184 15159405 10.1074/jbc.M402062200
Luo K Liu B Xiao Z Yu Y Yu X Gorelick R Yu XF Amino-terminal region of the human immunodeficiency virus type 1 nucleocapsid is required for human APOBEC3G packaging J Virol 2004 78 11841 11852 15479826 10.1128/JVI.78.21.11841-11852.2004
Svarovskaia ES Xu H Mbisa JL Barr R Gorelick RJ Ono A Freed EO Hu WS Pathak VK Human Apolipoprotein B mRNA-editing Enzyme-catalytic Polypeptide-like 3G (APOBEC3G) Is Incorporated into HIV-1 Virions through Interactions with Viral and Nonviral RNAs J Biol Chem 2004 279 35822 35828 15210704 10.1074/jbc.M405761200
Schafer A Bogerd HP Cullen BR Specific packaging of APOBEC3G into HIV-1 virions is mediated by the nucleocapsid domain of the gag polyprotein precursor Virology 2004 328 163 168 15464836 10.1016/j.virol.2004.08.006
Zennou V Perez-Caballero D Gottlinger H Bieniasz PD APOBEC3G incorporation into human immunodeficiency virus type 1 particles J Virol 2004 78 12058 12061 15479846 10.1128/JVI.78.21.12058-12061.2004
Mariani R Chen D Schrofelbauer B Navarro F Konig R Bollman B Munk C Nymark-McMahon H Landau NR Species-specific exclusion of APOBEC3G from HIV-1 virions by Vif Cell 2003 114 21 31 12859895 10.1016/S0092-8674(03)00515-4
Ribeiro AC Maia ESA Santa-Marta M Pombo A Moniz-Pereira J Goncalves J Barahona I Functional Analysis of Vif Protein Shows Less Restriction of Human Immunodeficiency Virus Type 2 by APOBEC3G J Virol 2005 79 823 833 15613310 10.1128/JVI.79.2.823-833.2005
Rosler C Kock J Malim MH Blum HE von Weizsacker F Comment on "Inhibition of hepatitis B virus replication by APOBEC3G" Science 2004 305 1403; author reply 1403 15353783 10.1126/science.1100464
Turelli P Mangeat B Jost S Vianin S Trono D Inhibition of hepatitis B virus replication by APOBEC3G Science 2004 303 1829 15031497 10.1126/science.1092066
Turelli P Jost S Mangeat B Trono D Response to comment on "Inhibition of Hepatitis B Virus Replication by APOBEC3G" Science 2004 305 1403b 15353783 10.1126/science.1099991
Yu Q Chen D Konig R Mariani R Unutmaz D Landau NR APOBEC3B and APOBEC3C are potent inhibitors of simian immunodeficiency virus replication J Biol Chem 2004
Bishop KN Holmes RK Sheehy AM Malim MH APOBEC-mediated editing of viral RNA Science 2004 305 645 15286366 10.1126/science.1100658
Marin M Rose KM Kozak SL Kabat D HIV-1 Vif protein binds the editing enzyme APOBEC3G and induces its degradation Nat Med 2003 9 1398 1403 14528301 10.1038/nm946
Sheehy AM Gaddis NC Malim MH The antiretroviral enzyme APOBEC3G is degraded by the proteasome in response to HIV-1 Vif Nat Med 2003 9 1404 1407 14528300 10.1038/nm945
Stopak K de Noronha C Yonemoto W Greene WC HIV-1 Vif blocks the antiviral activity of APOBEC3G by impairing both its translation and intracellular stability Mol Cell 2003 12 591 601 14527406 10.1016/S1097-2765(03)00353-8
Yu X Yu Y Liu B Luo K Kong W Mao P Yu XF Induction of APOBEC3G ubiquitination and degradation by an HIV-1 Vif-Cul5-SCF complex Science 2003 302 1056 1060 14564014 10.1126/science.1089591
Mehle A Goncalves J Santa-Marta M McPike M Gabuzda D Phosphorylation of a novel SOCS-box regulates assembly of the HIV-1 Vif-Cul5 complex that promotes APOBEC3G degradation Genes Dev 2004 18 2861 2866 15574592 10.1101/gad.1249904
Yu Y Xiao Z Ehrlich ES Yu X Yu XF Selective assembly of HIV-1 Vif-Cul5-ElonginB-ElonginC E3 ubiquitin ligase complex through a novel SOCS box and upstream cysteines Genes Dev 2004 18 2867 2872 15574593 10.1101/gad.1250204
Kao S Miyagi E Khan MA Takeuchi H Opi S Goila-Gaur R Strebel K Production of infectious human immunodeficiency virus type 1 does not require depletion of APOBEC3G from virus-producing cells Retrovirology 2004 1 27 15373943 10.1186/1742-4690-1-27
Mangeat B Turelli P Liao S Trono D A single amino acid determinant governs the species-specific sensitivity of APOBEC3G to Vif action J Biol Chem 2004 279 14481 14483 14966139 10.1074/jbc.C400060200
Xu H Svarovskaia ES Barr R Zhang Y Khan MA Strebel K Pathak VK A single amino acid substitution in human APOBEC3G antiretroviral enzyme confers resistance to HIV-1 virion infectivity factor-induced depletion Proc Natl Acad Sci U S A 2004 101 5652 5657 15054139 10.1073/pnas.0400830101
Bogerd HP Doehle BP Wiegand HL Cullen BR A single amino acid difference in the host APOBEC3G protein controls the primate species specificity of HIV type 1 virion infectivity factor Proc Natl Acad Sci U S A 2004 101 3770 3774 14999100 10.1073/pnas.0307713101
Schrofelbauer B Chen D Landau NR A single amino acid of APOBEC3G controls its species-specific interaction with virion infectivity factor (Vif) Proc Natl Acad Sci U S A 2004 101 3927 3932 14978281 10.1073/pnas.0307132101
Gaddis NC Sheehy AM Ahmad KM Swanson CM Bishop KN Beer BE Marx PA Gao F Bibollet-Ruche F Hahn BH Malim MH Further investigation of simian immunodeficiency virus Vif function in human cells J Virol 2004 78 12041 12046 15479843 10.1128/JVI.78.21.12041-12046.2004
Chen R Wang H Mansky LM Roles of uracil-DNA glycosylase and dUTPase in virus replication J Gen Virol 2002 83 2339 2345 12237414
Priet S Gros N Navarro JM Boretto J Canard B Querat G Sire J HIV-1-associated uracil DNA glycosylase activity controls dUTP misincorporation in viral DNA and is essential to the HIV-1 life cycle Mol Cell 2005 17 479 490 15721252 10.1016/j.molcel.2005.01.016
Kobayashi M Takaori-Kondo A Shindo K Abudu A Fukunaga K Uchiyama T APOBEC3G targets specific virus species J Virol 2004 78 8238 8244 15254195 10.1128/JVI.78.15.8238-8244.2004
Kazazian HHJ Mobile elements: drivers of genome evolution Science 2004 303 1626 1632 15016989 10.1126/science.1089670
Turelli P Vianin S Trono D The innate antiretroviral factor APOBEC3G does not affect human LINE-1 retrotransposition in a cell culture assay J Biol Chem 2004 279 43371 43373 15322092 10.1074/jbc.C400334200
Dewannieux M Dupressoir A Harper F Pierron G Heidmann T Identification of autonomous IAP LTR retrotransposons mobile in mammalian cells Nat Genet 2004 36 534 539 15107856 10.1038/ng1353
Ribet D Dewannieux M Heidmann T An active murine transposon family pair: retrotransposition of "master" MusD copies and ETn trans-mobilization Genome Res 2004 14 2261 2267 15479948 10.1101/gr.2924904
Ganesh L Burstein E Guha-Niyogi A Louder MK Mascola JR Klomp LW Wijmenga C Duckett CS Nabel GJ The gene product Murr1 restricts HIV-1 replication in resting CD4+ lymphocytes Nature 2003 426 853 857 14685242 10.1038/nature02171
Gao G Guo X Goff SP Inhibition of retroviral RNA production by ZAP, a CCCH-type zinc finger protein Science 2002 297 1703 1706 12215647 10.1126/science.1074276
Bick MJ Carroll JW Gao G Goff SP Rice CM MacDonald MR Expression of the zinc-finger antiviral protein inhibits alphavirus replication J Virol 2003 77 11555 11562 14557641 10.1128/JVI.77.21.11555-11562.2003
Varthakavi V Smith RM Bour SP Strebel K Spearman P Viral protein U counteracts a human host cell restriction that inhibits HIV-1 particle production Proc Natl Acad Sci U S A 2003 100 15154 15159 14657387 10.1073/pnas.2433165100
Hsu K Seharaseyon J Dong P Bour S Marban E Mutual functional destruction of HIV-1 Vpu and host TASK-1 channel Mol Cell 2004 14 259 267 15099524 10.1016/S1097-2765(04)00183-2
Margottin F Bour SP Durand H Selig L Benichou S Richard V Thomas D Strebel K Benarous R A novel human WD protein, h-beta TrCp, that interacts with HIV-1 Vpu connects CD4 to the ER degradation pathway through an F-box motif Mol Cell 1998 1 565 574 9660940 10.1016/S1097-2765(00)80056-8
| 15829012 | PMC1131925 | CC BY | 2021-01-04 16:36:40 | no | Retrovirology. 2005 Apr 13; 2:25 | utf-8 | Retrovirology | 2,005 | 10.1186/1742-4690-2-25 | oa_comm |
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-271585422910.1186/1742-4690-2-27ReviewHuman T-cell leukemia virus type I (HTLV-I) infection and the onset of adult T-cell leukemia (ATL) Matsuoka Masao [email protected] Institute for Virus Research, Kyoto University, Kyoto 606-8507, Japan2005 26 4 2005 2 27 27 29 3 2005 26 4 2005 Copyright © 2005 Matsuoka; licensee BioMed Central Ltd.2005Matsuoka; 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 clinical entity of adult T-cell leukemia (ATL) was established around 1977, and human T-cell leukemia virus type 1 (HTLV-I) was subsequently identified in 1980. In the 25 years since the discovery of HTLV-I, HTLV-I infection and its associated diseases have been extensively studied, and many of their aspects have been clarified. However, the detailed mechanism of leukemogenesis remains unsolved yet, and the prognosis of ATL patients still poor because of its resistance to chemotherapy and immunodeficiency. In this review, I highlight the recent progress and remaining enigmas in HTLV-I infection and its associated diseases, especially ATL.
==== Body
Background
In 1977, Takatsuki et al. reported adult T-cell leukemia (ATL) as a distinct clinical entity [1-3]. This disease is characterized by its aggressive clinical course, infiltrations into skin, liver, gastrointestinal tract and lung, hypercalcemia and the presence of leukemic cells with multilobulated nuclei (flower cell)(Figure 1). In 1980, Poiesz et al. discovered a human retrovirus in a cell line derived from a patient with ATL, and designated it human T-cell leukemia virus type I (HTLV-I) [4,5]. The linkage between ATL and HTLV-I was proven by Hinuma et al., who demonstrated the presence of an antibody against HTLV-I in patient sera [6]. Thereafter, Seiki et al. determined the whole sequence of HTLV-I and revealed the presence of a unique region, designated pX [7]. The pX region encodes several accessory genes, which control viral replication and the proliferation of infected cells [8]. In this review, I describe the recent advances in the field of HTLV-I and ATL research, with particular focus on the mechanism of leukemogenesis and therapeutic aspects.
Figure 1 Typical "flower cell" in the peripheral blood of an acute ATL patient. In the peripheral blood of an acute ATL patient, leukemic cells with multilobulated nuclei.
1. History of humans and HTLV-I
HTLV-I is a member of the Deltaretroviruses, which include HTLV-II, bovine leukemia virus and simian T-cell leukemia virus (STLV). The latter two viruses also cause lymphoid malignancies in the host, similar to the case with HTLV-I. HTLV and STLV are thought to originate from common ancestors, and share molecular, virological and epidemiological features. Therefore, they have been designated primate T-cell leukemia viruses (PTLVs). Phylogenetical analyses have revealed that HTLV-Ic first diverged from simian leukemia virus around 50,000 ± 10,000 years ago, while the spread of PTLV-I in Africa is estimated to have occurred at least 27,300 ± 8,200 years ago. Subsequently, HTLV-Ia, which is the most common subtype in Japan, diverged from the African strain 12,300 ± 4,900 years ago [9]. Thus, these viruses have had a long history with humans after the interspecies transmission. In contrast, human immunodeficiency virus type 1 (HIV-1) is thought to originate from simian immunodeficiency virus in chimpanzees (SIVCPZ) [10], and the interspecies transmission to humans is estimated to have occurred recently.
2. How does HTLV-I spread in humans?
There are approximately 10–20 million HTLV-I carriers in the world [11]. In particular, HTLV-I is endemic in Japan, parts of central Africa, the Caribbean basin and South America. In addition, epidemiological studies of HTLV-I have revealed high seroprevalence rates in Melanesia, Papua New Guinea and the Solomon islands, as well as among Australian aborigines [12]. In Japan, approximately 1.2 million individuals are estimated to be infected by HTLV-I, and more than 800 cases of ATL are diagnosed each year [13]. Moreover, this virus also causes the neurodegenerative disease, HTLV-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP) [14,15]. The cumulative risks of ATL among HTLV-I carriers in Japan are estimated to be about 6.6% for men and 2.1% for women, indicating that most HTLV-I carriers remain asymptomatic throughout their lives [16].
3. How does HTLV-I replicate and increase its copy number?
The HTLV-I provirus has a similar structure to other retroviruses: a long terminal repeat (LTR) at both ends and internal sequences such as the gag, pol and env genes. A characteristic of HTLV-I is the presence of the pX region, which exists between env and the 3'-LTR. This region encodes several accessory genes, which include the tax, rex, p12, p21, p30, p13 and HBZ genes. Among these, the tax gene plays central roles in viral gene transcription, viral replication and the proliferation of HTLV-I-infected cells. Tax enhances viral gene transcription from the 5'-LTR via interaction with cyclic AMP responsive element binding protein (CREB). Tax also interacts with cellular factors and activates transcriptional pathways, such as NF-κB, AP-1 and SRF [8,17-20]. For example, activation of NF-κB induces the transcription of various cytokines and their receptor genes, as well as anti-apoptotic genes such as bcl-xL and survivin [21-23]. The activation of NF-κB has been demonstrated to be critical for tumorigenesis both in vitro and in vivo [24,25]. On the other hand, Tax variant without activation of NF-κB has also been reported to immortalize primary T-lymphocytes in vitro [26], suggesting that mechanisms of immortalization are complex. In addition to NF-κB, activation of other transcriptional pathways such as CREB by Tax should be implicated in the immortalization and leukemogenesis.
Tax also interferes with the functions of p53, p16 and MAD1 [27-30]. These interactions enable HTLV-I-infected cells to escape from apoptosis, and also induce genetic instability. Although inactivation of p53 function by Tax is reported to be mediated by p300/CBP [27,28,31] or NF-κB activation [32], Tax can still repress p53's activity in spite of loss of p300/CBP binding or in cells lacking NF-κB activation [33], indicating the mechanism of p53 inactivation by Tax needs further investigation.
Although Tax promotes the proliferation of infected cells, it is also the major target of cytotoxic T-lymphocytes (CTLs) in vivo. Moreover, excess expression of Tax protein is considered to be harmful to infected cells. Therefore, HTLV-I has redundant mechanisms to suppress Tax expression. Rex binds to Rex-responsive element (RxRE) in the U3 and R regions of the 3'-LTR, and enhances the transport of the unspliced gag/pol and the singly spliced env transcripts. By this mechanism, double-spliced tax/rex mRNA decreases, resulting in suppressed expression of Tax [34]. On the other hand, p30 binds to tax/rex transcripts, and retains them in the nucleus [35]. The HBZ gene is encoded by the complementary strand of HTLV-I, and contains a leucine zipper domain. HBZ directly interacts with c-Jun or JunB [36], or enhances their degradation [37], resulting in the suppression of Tax-mediated viral transcription from the LTR.
Transforming growth factor-β (TGF-β) is an inhibitory cytokine that plays important roles in development, the immune system and oncogenesis. Since TGF-β generally suppresses the growth of tumor cells, most tumor cells acquire escape mechanisms that inhibit TGF-β signaling, including mutations in its receptor and in the Smad molecules that transduce the signal from the receptor. Tax has also been reported to inhibit TGF-β signaling by binding to Smad2, 3 and 4 or CBP/p300 [38,39]. Inhibition of TGF-β signaling enables HTLV-I-infected cells to escape TGF-β-mediated growth inhibition.
ATL cells have been reported to show remarkable chromosomal abnormalities [40], which should be implicated in the disease progression. Tax has been reported to interact with the checkpoint protein MAD1, which forms a complex with MAD2 and controls the mitotic checkpoint. This functional hindrance of MAD1 by Tax protein causes chromosomal instability, suggesting the involvement of this mechanism in oncogenesis [30]. Recently, Tax has been reported to interact with Cdc20 and activate Cdc20-associated anaphase-promoting complex, an E3 ubiquitin ligase that controls the metaphase-to-anaphase transition, thereby resulting in mitotic abnormalities [41].
In contrast to HTLV-I, HTLV-II promotes the proliferation of CD8-positive T-lymphocytes in vivo. Although it was first discovered in a patient with variant hairy cell leukemia, HTLV-II is less likely to have oncogenic properties since there is no obvious association between HTLV-II infections and cancers. Regardless of the homology of their tax sequences, the oncogenic potential of Tax1 (HTLV-I Tax) is more prominent than that of Tax2 (HTLV-II Tax). The most striking difference is that Tax2 lacks the binding motif at C-terminal end to PDZ domain proteins, while Tax 1 retains it [42]. When the PDZ domain of Tax1 is added to Tax2, the latter acquires oncogenic properties in the rat fibroblast cell line Rat-1, indicating that this domain is responsible for the transforming activity of HTLV-I [43].
To understand the pleiotropic actions of Tax protein more clearly, transcriptome analyses are essential. The transcriptional changes induced by Tax expression have been studied using DNA microarrays, which revealed that Tax upregulated the expression of the mixed-lineage kinase MLK3. MLK3 is involved in NF-κB activation by Tax as well as NIK and MEKK1 [44]. In addition to transcriptional changes, Tax is also well known to interact with cellular proteins and impair or alter their functions. For example, proteomic analyses of Tax-associated complexes showed that Tax could interact with cellular proteins, including the active forms of small GTPases, such as Cdc42, RhoA and Rac1, which should be implicated in the migration, invasion and adhesion of T-cells, as well as in the activation of the JNK pathway [45].
4. How does HTLV-I transmit and replicate in vivo?
Receptor and transmission of HTLV-I
HTLV-I can infect various types of cells, such as T-lymphocytes, B-lymphocytes, monocytes and fibroblasts [46]. Glucose transporter 1 (GLUT-1) has been identified as a receptor for HTLV-I and this receptor is ubiquitously expressed on cell surfaces [47]. However, the HTLV-I provirus is mainly detected in CD4-positive lymphocytes, with about 10% in CD8-positive T-lymphocytes [48]. This situation possibly arises because Tax mainly induces the increase of CD4-positive T-lymphocytes in vivo by enhanced proliferation and suppressed apoptosis.
In HTLV-I-infected individuals, no virions are detected in the serum. In addition, the infectivity of free virions is very poor compared with that of infected cells. These findings suggest that HTLV-I is spread by cell-to-cell transmission, rather than by free virions. In vitro analyses of HTLV-I-infected cells revealed that HTLV-I-infected cells form "virological synapses" with uninfected cells. Contact between an infected cell and a target cell induces the accumulation of the viral proteins Gag and Env, viral RNA and microtubules, and the viral complex subsequently transfers into the target cell [49]. HTLV-I also spreads in a cell-to-cell manner via such virological synapses in vivo.
HTLV-I is mainly transmitted via three routes: 1) mother-to-infant transmission (mainly through breast feeding) [50]; 2) sexual transmission (mainly from male-to-female); and 3) parenteral transmission (blood transfusion or intravenous drug use) [12]. In either route, HTLV-I-infected cells are essential for transmission. This was supported by the findings that fresh frozen plasma from carriers did not cause transmission [51] and freeze-thawing of breast milk reduced vertical transmission [52].
Provirus load and transmission
The provirus load varies more than 1000-fold among asymptomatic carriers [53]. Since most infected cells are considered to have one copy of the provirus, the provirus load indicates the percentage of infected cells among lymphocytes. The provirus load is relatively constant during the latent period [53]. Analysis of naive individuals who seroconvert after marrying an HTLV-I-seropositive spouse demonstrated that the proviral gp46 sequences are identical among married couples. This finding confirmed that HTLV-I is transmitted from a seropositive individual to an uninfected spouse. The provirus loads frequently differ between couples despite infection by the same HTLV-I virus, indicating that the number of infected cells is determined by host factors rather than virus itself [54].
Why does HTLV-I increase the number of infected cells by the pleiotropic actions of Tax? The provirus load in peripheral blood mononuclear cells (PBMCs) is well correlated with that in breast milk, and a higher provirus load in breast milk increases the risk of vertical transmission of HTLV-I [55,56]. Similarly, a higher provirus load in PBMCs may be associated with a higher risk of sexual transmission. Thus, an increase in the number of infected cells by the actions of accessory genes, especially tax, facilitates transmission. Therefore, HTLV-I has strategies that increase the number of HTLV-I-infected cells via the action of accessory gene products, thereby increasing the chance of transmission.
Clonal expansion of HTLV-I-infected cells
After HTLV-I infection, viral proteins such as Tax protein promote the proliferation of infected cells and also inhibit apoptosis by their pleiotropic actions. Since the HTLV-I provirus is randomly integrated into the host genome, the identification of integration sites enables to identify each infected clone, and to trace the kinetics of infected cells in vivo. Analyses using inverse PCR, which can identify the integration sites of the HTLV-I provirus, revealed that the proliferation of infected cells is oligoclonal, and that infected cells persistently survive in vivo [57-59]. Importantly, such clonal expansion in carriers is directly associated with the onset of ATL [60]. Thus, the viral strategies to increase the number of HTLV-I-infected cells work efficiently in most carriers without any adverse effects. However, the increased number of infected cells causes an excess immune reaction, leading to inflammatory diseases, HAM/TSP, infective dermatitis [61] or HTLV-I-associated uveitis [62]. Moreover, such prolonged proliferation of infected CD4-positive T-lymphocytes results in the onset of ATL in some carriers after a long latent period.
Inactivation of Tax expression in ATL cells
As mentioned above, Tax expression confers advantages and disadvantages on HTLV-I-infected cells. Although the proliferation of infected cells is promoted by Tax expression, CTLs attack the Tax-expressing cells since Tax is their major target [63]. In HTLV-I-infected cells, Rex, p30 and HBZ suppress Tax expression. On the other hand, loss of Tax expression is frequently observed in leukemic cells. Three mechanisms have been identified for inactivation of Tax expression: 1) genetic changes of the tax gene (nonsense mutations, deletions or insertions) [64,65]; 2) DNA methylation of the 5'-LTR [65,66]; and 3) deletion of the 5'-LTR (Figure 2) [67]. Among fresh leukemic cells isolated from ATL patients, about 60% of cases do not express the tax gene transcript. Interestingly, ATL cells with genetic changes of the tax gene expressed its transcripts, suggesting that ATL cells do not silence the transcription when the tax gene is abortive [65]. Loss of Tax expression gives ATL cells advantage for their survival since they can escape from CTLs.
Figure 2 Natural course of HTLV-I infection to onset of ATL. HTLV-I is transmitted via three routes, and infected cells are necessary in all three. After infection, HTLV-I promotes clonal proliferation of infected cells by pleiotropic actions of Tax. Tax expression is suppressed by viral accessory gene products, such as Rex, p30 and HBZ proteins. Proliferation of HTLV-I infected cells is controlled by cytotoxic T-cells in vivo. After a long latent period, ATL develops in about 5% of asymptomatic carriers. The expression of Tax is inactivated by several mechanisms, suggesting that Tax is not necessary in this stage. Alternatively, alternations in the host genome accumulate during the latent period, finally leading to onset of ATL.
Longer lifespan of HTLV-I-infected cells and cancer
Lymphoid malignancy with a T-cell origin is rare compared with B-cell malignancy. ATL shares hematological, pathological and immunological features with cutaneous T-cell lymphoma (CTCL; Sezary syndrome and Mycosis fungoides). The frequency of CTCL in Japan is estimated to be one/million/year. On the other hand, the frequency of ATL among carriers is estimated to be 1000/million/year. From these data, HTLV-I infection is estimated to increase the risk of T-cell malignancy by up to 1000-fold in carriers.
HTLV-I infection confers a long lifespan on the infected cells due to the pleiotropic actions of Tax, resulting in increased numbers of infected cells. Such infected cells are essential for the transmission of HTLV-I. This strategy to increase the number of infected cells in vivo is thought to increase the incidence of cancer in T-cells. What is the mechanism for this oncogenesis? DNA methylation is known to be associated with aging. Some genes are hypermethylated in older people, indicating that DNA hypermethylation is a physiological phenomenon in some genes. Under normal circumstances, T-lymphocytes survive for several years, and long-lived T-lymphocytes with disordered methylation should be replaced. However, HTLV-I-infected T-cells are considered to survive and accumulate abnormal methylation. The process of oncogenesis is similar to that of evolution [68]. The infected cells that are suitable for survival should be selected in vivo, and epigenetic and genetic changes of the genome play critical roles in this selection. Accumulating alterations of the host genome transform the HTLV-I-infected cells into ATL cells, and also enable ATL cells to proliferate in the absence of Tax expression (Figure 2). In the provirus, DNA methylation of the 5'-LTR silences viral transcription in leukemic cells, which facilitates the escape of ATL cells from the host immune system [65].
5. Somatic alterations in ATL cells
As described, some ATL cells can proliferate without functional Tax protein, suggesting that somatic (genetic and epigenetic) alterations cause transcriptional or functional changes to the host genes. The p53 gene is frequently mutated in various cancers, and these mutations are associated with disease progression and a poor prognosis. The mutation rate of the p53 gene in ATL cells has been reported to be 36% (4/11) and 30% (3/10) [69-71]. The p16 gene is an inhibitor of cyclin-dependent kinase 4/6, and blocks the cell cycle. Genetic changes in this gene (deletion in most cases) have been described in many types of cancer cells. Deletion of the p16 gene has also been reported in ATL cells [72]. Moreover, DNA methylation of the promoter region of the p16 gene is also implicated in the suppression of p16 [73]. In addition, genetic changes in the p27KIP1, RB1/p105 and RB2/p130 genes have been reported in ATL, although they are relatively rare: 2/42 (4.8%) for the p27KIP1 gene; 2/40 (5%) for the RB1/p105 gene; and 1/41 (2.4%) for the RB2/p130 gene) [74]. The fact that higher frequencies of genetic changes in these tumor suppressor genes are observed among aggressive forms of ATL suggests that such genetic changes are implicated in disease progression.
Fas antigen was the first identified death receptor. It transduces the death signal by binding of its ligand, Fas ligand (FasL). ATL cells highly express Fas antigen on their cell surface [75], and are highly susceptible to death signals mediated by agonistic antibodies to Fas antigen, such as CH-11. Genetic changes of Fas gene in ATL cells, which confer resistance to the Fas-mediated signal, have been reported [76,77]. Normal activated T-lymphocytes express FasL as well as Fas antigen. Apoptosis induced by autocrine mechanisms is designated activation-induced cell death (AICD) and this controls the immune response [78]. Although ATL cells express Fas antigen, they do not produce FasL, thereby enabling ATL cells to escape from AICD. Attempts to isolate hypermethylated genes from ATL cells identified the EGR3 gene as a hypermethylated gene compared to PBMCs from carriers [79]. EGR3 is a transcriptional factor with a zinc finger domain, that is essential for transcription of the FasL gene [80]. The finding that EGR3 gene transcription is silenced in ATL cells could account for the loss of FasL expression, and the escape of ATL cells from AICD. Thus, alterations of the Fas (genetic) and EGR3 (epigenetic) genes are examples of ATL cell evolution in vivo.
Disordered DNA methylation has been identified in the genome of ATL cells compared with that of PBMCs from carriers: hypomethylation is associated with aberrant expression of the MEL1S gene [81], while hypermethylation silences transcription of the p16 [73], EGR3 and KLF4 genes as well as many others [79]. It is reasonable to consider that other currently unidentified genes are involved in such alterations of the genome in ATL cells, and play roles in leukemogenesis.
Transcriptome analyses using DNA microarrays have revealed transcriptional changes that are specific to ATL cells. Among 192 up-regulated genes, the expressions of the tumor suppressor in lung cancer 1 (TSLC1), caveolin 1 and prostaglandin D2 synthase genes were increased more than 30-fold in fresh ATL cells compared with normal CD4+ and CD4+, CD45RO+ T-cells [82]. TSLC1 is a cell adhesion molecule that acts as a tumor suppressor in lung cancer. Although TSLC1 is not expressed on normal T-lymphocytes, all acute ATL cells show ectopic TSLC1 expression. Enforced expression of TSLC1 enhances both the self-aggregation and adhesion abilities to vascular endothelial cells in ATL cells. Thus, TSLC1 expression is implicated in the adhesion or infiltration of ATL cells. By screening a retrovirus cDNA library from ATL cells, a gene with oncogenic potency was identified in NIH3T3 cells, and designated the Tgat gene [83]. Ectopic expression of the Tgat gene is observed in aggressive forms of ATL, and in vitro experiments showed that its expression is associated with an invasive phenotype.
6. Immune control of HTLV-I infection
The host immune system, especially the cellular response, against HTLV-I exerts critical control over virus replication and the proliferation of infected cells [84]. CTLs against the virus have been extensively studied, and Tax protein was found to be the dominant antigen recognized by CTLs in vivo [63]. HTLV-I-specific CD8-positive CTLs are abundant and chronically activated. The paradox is that the frequency of Tax-specific CTLs is much higher in HAM/TSP patients than in carriers. Since the provirus load is higher in HAM/TSP patients, this finding suggests that the CTLs in HAM/TSP cannot control the number of infected cells. One explanation for this is that the CTLs in HAM/TSP patients show less efficient cytolytic activity toward infected cells, whereas CTLs in carriers can suppress the proliferation of infected cells [85]. Hence, the gene expression profiles of circulating CD4+ and CD8+ lymphocytes were compared between carriers with high and low provirus loads. The results revealed that CD8+ lymphocytes from individuals with a low HTLV-1 provirus load show higher expressions of genes associated with cytolytic activities or antigen recognition than those from carriers with a high provirus load [86]. Thus, CD8+ T-lymphocytes in individuals with a low provirus load successfully control the number of HTLV-I-infected cells due to their higher CTL activities. Thus, the major determinant of the provirus load is thought to be the CTL response to HTLV-I.
As mentioned above, the provirus load is considered to be controlled by host factors. Considering that the cellular immune responses are critically implicated in the control of HTLV-I infection, human leukocyte antigen (HLA) should be a candidate for such a host genetic factor. From analyses of HAM/TSP patients and asymptomatic carriers, HLA-A02, and Cw08 are independently associated with a lower provirus load and a lower risk of HAM/TSP. In addition, polymorphisms of other genes (TNF-α, SDF-1, HLA-B54, HLA-DRB-10101 and IL-15) are also associated with the provirus load, although their associations are not as significant compared with HLA-A02, and Cw08 [87,88]. Regarding the onset of ATL, only a polymorphism of TNF-α gene was reported to show an association [89]. However, familial clustering of ATL cases is a well-known phenomenon, strongly suggesting that genetic factors are implicated in the onset of ATL [90-92].
Spontaneous remission is more frequently observed in patients with ATL than those with other hematological malignancies [90,93]. Usually, this phenomenon is associated with infectious diseases, suggesting that immune activation of the host enhances the immune response against ATL cells. If the immune response against HTLV-I is implicated in spontaneous remission, this suggests the possibility of immunotherapy for ATL patients by the induction of an immune response to HTLV-I [94], for example via antigen-stimulated dendritic cells.
Immunodeficiency in ATL patients is pronounced, and results in frequent opportunistic infections by various pathogens, including Pneumocystis carinii, cytomegalovirus, fungus, Strongyloides and bacteria, due to the inevitable impairment of the T-cell functions [95]. To a lesser extent, impaired cell-mediated immunity has also been demonstrated in HTLV-I carriers [96]. Such immunodeficiency in the carrier state may be associated with the leukemogenesis of ATL by allowing the proliferation of HTLV-I-infected cells. A prospective study of HTLV-I-infected individuals found that carriers who later develop ATL have a higher anti-HTLV-I antibody and a low anti-Tax antibody level for up to 10 years preceding their diagnosis. This finding indicates that HTLV-I carriers with a higher anti-HTLV-I titer, which is roughly correlated with the HTLV-I provirus load, and a lower anti-Tax reactivity may be at the greatest risk of developing ATL [97]. The anti-HTLV-I antibody and soluble IL-2 receptor (sIL-2R) levels are correlated with the HTLV-I provirus load [53], and a high antibody titer and high sIL-2R level are risk factors for developing ATL among carriers [98]. Taken together, these findings suggest that a higher proliferation of HTLV-I-infected cells and a low immune response against Tax may be associated with the onset of ATL. Given these findings, potentiation of CTLs against Tax via a vaccine strategy may be useful for preventing the onset of ATL [99].
EBV-associated lymphomas frequently develop in individuals with an immunodeficient state associated with transplantation or AIDS. This has also been reported in an ATL patient [100]. Does such an immunodeficient state influence the onset of ATL? Among 24 patients with post-transplantation lymphoproliferative disorders (PT-LPDs) after renal transplantation in Japan, 5 cases of ATL have been reported. Considering that most PT-LPDs are of B-cell origin in Western countries, this frequency of ATL in Japan is quite high. Although the high HTLV-I seroprevalence is due to blood transfusion during hemodialysis, the immunodeficient state during renal transplantation apparently promotes the onset of ATL [101]. In addition, when experimental allogeneic transplantation was performed to 12 rhesus monkeys and immunosuppressive agents (cyclosporine, prednisolone or lymphocyte-specific monoclonal antibodies) were administered to prevent rejection, 4 of the 7 monkeys that died during the experiment showed PT-LPDs. Importantly, the STLV provirus was detected in all PT-LPD samples [102]. These observations emphasize that transplantation into HTLV-I-infected individuals or from HTLV-I positive donors require special attention.
Although the mechanism of immunodeficiency remains unknown, some previous reports have provided important clues. One mechanism for immunodeficiency is that HTLV-I infects CD8-positive T-lymphocytes, which may impair their functions [48]. Indeed, the immune response against Tax via HTLV-I-infected CD8-positive T-cells renders these cells susceptible to fratricide mediated by autologous HTLV-I-specific CD8-positive T-lymphocytes [103]. Fratricide among virus-specific CTLs could impair the immune control of HTLV-I. Another mechanism for immunodeficiency is based on the observation that the number of naive T-cells decreases in individuals infected with HTLV-I via decreased thymopoiesis [48]. In addition, CD4+ and CD25+ T-lymphocytes are classified as immunoregulatory T-cells that control the host immune system. Regulatory T-cells suppress the immune reaction via the expression of immunoregulatory molecules on their surfaces. The FOXP3 gene has been identified as a master gene that controls gene expressions specific to regulatory T-cells. FOXP3 gene transcription can be detected in some ATL cases (10/17; 59%) [104]. Such ATL cells are thought to suppress the immune response via expression of immunoregulatory molecules on their surfaces, and production of immunosuppressive cytokines.
6. Pathogenesis of HTLV-I infection
ATL cells are derived from activated helper T-lymphocytes, which play central roles in the immune system by elaborating cytokines and expressing immunoregulatory molecules. ATL cells are known to retain such features, and this cytokine production or surface molecule expression may modify the pathogenesis.
ATL is well known to infiltrate various organs and tissues, such as the skin, lungs, liver, gastrointestinal tract, central nervous system and bone [95]. This infiltrative tendency of leukemic cells is possibly attributable to the expressions of various surface molecules, such as chemokine receptors and adhesion molecules. Skin-homing memory T-cells uniformly express CCR4, and its ligands are thymus and activation-regulated chemokine (TARC) and macrophage-derived chemokine (MDC). CCR4 is expressed on most ATL cells. In addition, TARC and MDC are expressed in skin lesions in ATL patients. Thus, CCR4 expression should be implicated in the skin infiltration [105]. On the other hand, CCR7 expression is associated with lymph node involvement [106]. OX40 is a member of the tumor necrosis factor family, and was reported to be expressed on ATL cells [107]. It was also identified as a gene associated with the adhesion of ATL cells to endothelial cells by a functional cloning system using a monoclonal antibody that inhibited the attachment of ATL cells [108]. Thus, OX40 is also implicated in the cell adhesion and infiltration of ATL cells.
Hypercalcemia is frequently complicated in patients with acute ATL (more than 70% during the whole clinical course) [109]. In hypercalcemic patients, the number of osteoclasts increases in the bone (Figure 3). RANK ligand, which is expressed on osteoblasts, and M-CSF act synergistically on hematopoietic precursor cells, and induce the differentiation into osteoclasts [110]. ATL cells from hypercalcemic ATL patients express RANK ligand, and induced the differentiation of hematopoietic stem cells into osteoclasts when ATL cells were co-cultured with hematopoietic stem cells [111]. In addition, the serum level of parathyroid hormone-related peptide (PTH-rP) is also elevated in most of hypercalcemic ATL patients. PTH-rP indirectly increases the number of osteoclasts, as well as activating them [112,113], which is also implicated in mechanisms of hypercalcemia.
Figure 3 Increased number of osteoclasts in the bone of a hypercalcemic ATL patient. In a hypercalcemic patient, the number of osteoclast (arrows) increased in the bone, which accelerated bone resorption.
7. Treatment of ATL – the remaining mission and challenges
Regardless of intensive chemotherapies, the prognosis of ATL patients has not so improved. The median survival time of acute or lymphoma-type ATL was reported to be 13 months with the most intensive chemotherapy [114]. Such a poor prognosis might be due to: 1) the resistance of ATL cells to anti-cancer drugs; and 2) the immunodeficient state and complicated opportunistic infections as described above. Regarding the resistance to anti-cancer drugs, one mechanism is the activated NF-κB pathway in ATL cells [115], which increases the transcription of anti-apoptotic genes such as bcl-xL and survivin. A proteasome inhibitor, bortezomib, is currently used for the treatment of multiple myeloma. One of its mechanisms is suppression of the NF-κB pathway by inhibiting the proteasomal degradation of IκB protein. Several groups have shown that bortezomib is effective against ATL cells both in vitro and in vivo [116-119]. Since the sensitivity to bortezomib is well correlated with the extent of NF-κB activation, the major mechanism of the anti-ATL effect is speculated to be inhibition of NF-κB. In addition, an NF-κB inhibitor has also been demonstrated to be effective against ATL cells [120].
During chemotherapy for ATL, chemotherapeutic agents worsen the immunodeficient state of ATL patients. In this regard, antibody therapy against ATL cells has advantages due to its decreased adverse effects. A humanized monoclonal antibody to CD25 has been clinically administered to patients with ATL [121,122]. In addition, a monoclonal antibody to CD2 is at the preclinical stage [123]. As described above, most ATL cells express CCR4 antigen on their surfaces, and a humanized antibody against CCR4 is being developed as an anti-ATL agent [124].
Advances in the treatment of ATL were brought about by allogeneic bone marrow or stem cell transplantation [125,126]. Absence of graft-versus-host disease (GVHD) was linked with relapse of ATL, suggesting that GVHD or graft-versus-ATL may be implicated in the clinical effects of allogeneic stem cell transplantation [125]. Furthermore, 16 patients with ATL, who were over 50 years of age, were treated with allogeneic stem cell transplantation with reduced conditioning intensity (RIST) from HLA-matched sibling donors [127]. Among 9 patients in whom ATL relapsed after transplantation, 3 achieved a second complete remission after rapid discontinuation of cyclosporine A. This finding strongly suggests the presence of a graft-versus-ATL effect in these patients. In addition, Tax peptide-recognizing cells were detected by a tetramer assay (HLA-A2/Tax 11–19 or HLA-A24/Tax 301–309) in patients after allogeneic stem cell transplantation [128]. In 8 patients, the provirus became undetectable by real-time PCR. Among these, 2 patients who received grafts from HTLV-I-positive donors also became provirus-negative by real-time PCR after RIST. Since the provirus load is relatively constant in HTLV-I-infected individuals [53], this finding indicates an enhanced immune response against HTLV-I after RIST, which suppresses the provirus load. This may account for the effectiveness of allogeneic stem cell transplantation to ATL. However, Tax expression is frequently lost in ATL cells as described above. Many questions arise, such as whether the tax gene status is correlated with the effect of allogeneic stem cell transplantation, and whether the effectiveness of the anti-HTLV-I immune response is against leukemic cells or non-leukemic HTLV-I-infected cells. Nevertheless, these data suggest that potentiation of the immune response against viral proteins such as Tax may be an attractive way to treat ATL patients [94]. Such strategies may enable preventive treatment of high-risk HTLV-I carriers, such as those with familial ATL history, predisposing genetic factors to ATL, a higher provirus load, etc.
8. Two human retroviruses – HTLV-I and HIV-1
As described in the first section, HTLV-I has resided in humans for a long time. On the other hand, HIV-1 has only been recently transmitted to humans, probably from chimpanzees. Due to the comparatively small genomic differences between humans and chimpanzees, this virus can quickly adapt to human cells. These two human retroviruses are opposite in many aspects. HIV-1 vigorously replicates in vivo, and the maximum production of HIV-1 virions in the body can reach 1010 per day. Since reverse transcriptase is an error-prone enzyme due to its lack of proof-reading activity, it produces about one mistake per replication, resulting in tremendous errors in the proviral sequence during replication. Although most of these variations ruin the virus replication due to nonsense mutations or impairment of viral gene functions, some become capable of replicating under different circumstances such as the presence of anti-HIV drugs and activation of the host immune system. This can account for why HIV-1 acquires resistance against anti-HIV drugs, and escape from CTLs. On the other hand, HTLV-I increases its copy number in two ways, namely replication of HTLV-I itself and the proliferation of HTLV-I-infected cells in vivo. Although immune responses (antibodies, CTLs) against viral proteins suggest the presence of active viral replication in vivo, most of increased HTLV-I provirus load (the number of infected cells) is considered to be due to proliferation of infected cells since CTLs efficiently eliminate virus-expressing cells. Therefore, there is much less variation in the HTLV-I provirus sequence compared with HIV-1 [129]. However, this strategy by which HTLV-I increases the number of infected cells due to clonal expansion generates unfortunate side effects for both the host and the virus, namely oncogenesis of CD4-positive T-lymphocytes and the development of ATL.
Acknowledgements
I would like to thank my colleagues Jun-ichirou Yasunaga, Kisato Nosaka, Mika Yoshida, Yorifumi Satou, Yuko Taniguchi, Satoshi Takeda, Ken-ichirou Etoh and Sadahiro Tamiya for their excellent studies.
==== Refs
Takatsuki K Uchiyama T Sagawa K Yodoi J Seno S, Takaku F and Irino S Adult T cell leukemia in Japan. Topic in Hematology, the 16th International congress of Hematology 1977 Amsterdam, 73 77
Uchiyama T Yodoi J Sagawa K Takatsuki K Uchino H Adult T-cell leukemia: clinical and hematologic features of 16 cases. Blood 1977 50 481 492 301762
Takatsuki K Discovery of adult T-cell leukemia Retrovirology 2005 2 16 15743528 10.1186/1742-4690-2-16
Poiesz BJ Ruscetti FW Gazdar AF Bunn PA Minna JD Gallo RC Detection and isolation of type C retrovirus particles from fresh and cultured lymphocytes of a patient with cutaneous T-cell lymphoma Proc Natl Acad Sci U S A 1980 77 7415 7419 6261256
Gallo RC The discovery of the first human retrovirus: HTLV-1 and HTLV-2 Retrovirology 2005 2 17 15743526 10.1186/1742-4690-2-17
Hinuma Y Nagata K Hanaoka M Nakai M Matsumoto T Kinoshita KI Shirakawa S Miyoshi I Adult T-cell leukemia: antigen in an ATL cell line and detection of antibodies to the antigen in human sera Proc Natl Acad Sci U S A 1981 78 6476 6480 7031654
Seiki M Hattori S Hirayama Y Yoshida M Human adult T-cell leukemia virus: complete nucleotide sequence of the provirus genome integrated in leukemia cell DNA Proc Natl Acad Sci U S A 1983 80 3618 3622 6304725
Yoshida M Multiple viral strategies of htlv-1 for dysregulation of cell growth control Annu Rev Immunol 2001 19 475 496 11244044 10.1146/annurev.immunol.19.1.475
Van Dooren S Salemi M Vandamme AM Dating the origin of the African human T-cell lymphotropic virus type-i (HTLV-I) subtypes Mol Biol Evol 2001 18 661 671 11264418
Gao F Bailes E Robertson DL Chen Y Rodenburg CM Michael SF Cummins LB Arthur LO Peeters M Shaw GM Sharp PM Hahn BH Origin of HIV-1 in the chimpanzee Pan troglodytes troglodytes Nature 1999 397 436 441 9989410 10.1038/17130
Edlich RF Arnette JA Williams FM Global epidemic of human T-cell lymphotropic virus type-I (HTLV-I) J Emerg Med 2000 18 109 119 10645850 10.1016/S0736-4679(99)00173-0
Blattner WA Gallo RC Takahashi K Epidemiology of HTLV-I and HTLV-II infection Adult T-cell leukemia 1994 New York, Oxford University Press p.45 90
Tajima K Inoue M Takezaki T Ito M Ito SI Takatsuki K Ethnoepidemiology of ATL in Japan with special reference to the Mongoloid dispersal Adult T-cell leukemia 1994 New York, Oxford University Press p.91 112
Gessain A Jouannelle A Escarmant P Calender A Schaffar-Deshayes L de-The G HTLV antibodies in patients with non-Hodgkin lymphomas in Martinique Lancet 1984 1 1183 1184 6144909 10.1016/S0140-6736(84)91431-4
Osame M Usuku K Izumo S Ijichi N Amitani H Igata A Matsumoto M Tara M HTLV-I associated myelopathy, a new clinical entity Lancet 1986 1 1031 1032 2871307 10.1016/S0140-6736(86)91298-5
Arisawa K Soda M Endo S Kurokawa K Katamine S Shimokawa I Koba T Takahashi T Saito H Doi H Shirahama S Evaluation of adult T-cell leukemia/lymphoma incidence and its impact on non-Hodgkin lymphoma incidence in southwestern Japan Int J Cancer 2000 85 319 324 10652420 10.1002/(SICI)1097-0215(20000201)85:3<319::AID-IJC4>3.0.CO;2-B
Franchini G Fukumoto R Fullen JR T-cell control by human T-cell leukemia/lymphoma virus type 1 Int J Hematol 2003 78 280 296 14686485
Jeang KT Giam CZ Majone F Aboud M Life, death, and tax: role of HTLV-I oncoprotein in genetic instability and cellular transformation J Biol Chem 2004 279 31991 31994 15090550 10.1074/jbc.R400009200
Azran I Schavinsky-Khrapunsky Y Aboud M Role of Tax protein in human T-cell leukemia virus type-I leukemogenicity Retrovirology 2004 1 20 15310405 10.1186/1742-4690-1-20
Matsuoka M Jeang KT Human T-cell leukemia virus type I (HTLV-I): a progress report Cancer Research
Tsukahara T Kannagi M Ohashi T Kato H Arai M Nunez G Iwanaga Y Yamamoto N Ohtani K Nakamura M Fujii M Induction of Bcl-x(L) expression by human T-cell leukemia virus type 1 Tax through NF-kappaB in apoptosis-resistant T-cell transfectants with Tax J Virol 1999 73 7981 7987 10482545
Nicot C Mahieux R Takemoto S Franchini G Bcl-X(L) is up-regulated by HTLV-I and HTLV-II in vitro and in ex vivo ATLL samples Blood 2000 96 275 281 10891462
Kawakami H Tomita M Matsuda T Ohta T Tanaka Y Fujii M Hatano M Tokuhisa T Mori N Transcriptional activation of survivin through the NF-kappaB pathway by human T-cell leukemia virus type I tax Int J Cancer 2005
Yamaoka S Inoue H Sakurai M Sugiyama T Hazama M Yamada T Hatanaka M Constitutive activation of NF-kappa B is essential for transformation of rat fibroblasts by the human T-cell leukemia virus type I Tax protein Embo J 1996 15 873 887 8631308
Portis T Harding JC Ratner L The contribution of NF-kappa B activity to spontaneous proliferation and resistance to apoptosis in human T-cell leukemia virus type 1 Tax-induced tumors Blood 2001 98 1200 1208 11493471 10.1182/blood.V98.4.1200
Rosin O Koch C Schmitt I Semmes OJ Jeang KT Grassmann R A human T-cell leukemia virus Tax variant incapable of activating NF-kappaB retains its immortalizing potential for primary T-lymphocytes J Biol Chem 1998 273 6698 6703 9506967 10.1074/jbc.273.12.6698
Suzuki T Uchida-Toita M Yoshida M Tax protein of HTLV-1 inhibits CBP/p300-mediated transcription by interfering with recruitment of CBP/p300 onto DNA element of E-box or p53 binding site Oncogene 1999 18 4137 4143 10435595 10.1038/sj.onc.1202766
Ariumi Y Kaida A Lin JY Hirota M Masui O Yamaoka S Taya Y Shimotohno K HTLV-1 tax oncoprotein represses the p53-mediated trans-activation function through coactivator CBP sequestration Oncogene 2000 19 1491 1499 10734308 10.1038/sj.onc.1203450
Suzuki T Kitao S Matsushime H Yoshida M HTLV-1 Tax protein interacts with cyclin-dependent kinase inhibitor p16INK4A and counteracts its inhibitory activity towards CDK4 Embo J 1996 15 1607 1614 8612584
Jin DY Spencer F Jeang KT Human T cell leukemia virus type 1 oncoprotein Tax targets the human mitotic checkpoint protein MAD1 Cell 1998 93 81 91 9546394 10.1016/S0092-8674(00)81148-4
Van Orden K Yan JP Ulloa A Nyborg JK Binding of the human T-cell leukemia virus Tax protein to the coactivator CBP interferes with CBP-mediated transcriptional control Oncogene 1999 18 3766 3772 10391685 10.1038/sj.onc.1202703
Pise-Masison CA Mahieux R Jiang H Ashcroft M Radonovich M Duvall J Guillerm C Brady JN Inactivation of p53 by human T-cell lymphotropic virus type 1 Tax requires activation of the NF-kappaB pathway and is dependent on p53 phosphorylation Mol Cell Biol 2000 20 3377 3386 10779327 10.1128/MCB.20.10.3377-3386.2000
Miyazato A Sheleg S Iha H Li Y Jeang KT Evidence for NF-kB- and CBP-independent repression of p53's transcriptional activity by humna T-cell leukemia virus type I Tax in mouse embryo and primary humna fibroblasts Journal of Virology
Inoue J Yoshida M Seiki M Transcriptional (p40x) and post-transcriptional (p27x-III) regulators are required for the expression and replication of human T-cell leukemia virus type I genes Proc Natl Acad Sci U S A 1987 84 3653 3657 3035544
Nicot C Dundr M Johnson JM Fullen JR Alonzo N Fukumoto R Princler GL Derse D Misteli T Franchini G HTLV-1-encoded p30(II) is a post-transcriptional negative regulator of viral replication Nat Med 2004 10 197 201 14730358 10.1038/nm984
Basbous J Arpin C Gaudray G Piechaczyk M Devaux C Mesnard JM The HBZ factor of human T-cell leukemia virus type I dimerizes with transcription factors JunB and c-Jun and modulates their transcriptional activity J Biol Chem 2003 278 43620 43627 12937177 10.1074/jbc.M307275200
Matsumoto J Ohshima T Isono O Shimotohno K HTLV-1 HBZ suppresses AP-1 activity by impairing both the DNA-binding ability and the stability of c-Jun protein Oncogene 2005 24 1001 1010 15592508 10.1038/sj.onc.1208297
Mori N Morishita M Tsukazaki T Giam CZ Kumatori A Tanaka Y Yamamoto N Human T-cell leukemia virus type I oncoprotein Tax represses Smad-dependent transforming growth factor beta signaling through interaction with CREB-binding protein/p300 Blood 2001 97 2137 2144 11264182 10.1182/blood.V97.7.2137
Lee DK Kim BC Brady JN Jeang KT Kim SJ Human T-cell Lymphotropic Virus Type 1 Tax Inhibits Transforming Growth Factor-beta Signaling by Blocking the Association of Smad Proteins with Smad-binding Element J Biol Chem 2002 277 33766 33775 12097320 10.1074/jbc.M200150200
Kamada N Sakurai M Miyamoto K Sanada I Sadamori N Fukuhara S Abe S Shiraishi Y Abe T Kaneko Y Chromosome abnormalities in adult T-cell leukemia/lymphoma: a karyotype review committee report Cancer Res 1992 52 1481 1493 1540956
Liu B Hong S Tang Z Yu H Giam CZ HTLV-I Tax directly binds the Cdc20-associated anaphase-promoting complex and activates it ahead of schedule Proc Natl Acad Sci U S A 2005 102 63 68 15623561 10.1073/pnas.0406424101
Endo K Hirata A Iwai K Sakurai M Fukushi M Oie M Higuchi M Hall WW Gejyo F Fujii M Human T-cell leukemia virus type 2 (HTLV-2) Tax protein transforms a rat fibroblast cell line but less efficiently than HTLV-1 Tax J Virol 2002 76 2648 2653 11861831 10.1128/JVI.76.6.2648-2653.2002
Hirata A Higuchi M Niinuma A Ohashi M Fukushi M Oie M Akiyama T Tanaka Y Gejyo F Fujii M PDZ domain-binding motif of human T-cell leukemia virus type 1 Tax oncoprotein augments the transforming activity in a rat fibroblast cell line Virology 2004 318 327 336 14972558 10.1016/j.virol.2003.10.006
Ng PW Iha H Iwanaga Y Bittner M Chen Y Jiang Y Gooden G Trent JM Meltzer P Jeang KT Zeichner SL Genome-wide expression changes induced by HTLV-1 Tax: evidence for MLK-3 mixed lineage kinase involvement in Tax-mediated NF-kappaB activation Oncogene 2001 20 4484 4496 11494144 10.1038/sj.onc.1204513
Wu K Bottazzi ME de la Fuente C Deng L Gitlin SD Maddukuri A Dadgar S Li H Vertes A Pumfery A Kashanchi F Protein profile of tax-associated complexes J Biol Chem 2004 279 495 508 14530271 10.1074/jbc.M310069200
Koyanagi Y Itoyama Y Nakamura N Takamatsu K Kira J Iwamasa T Goto I Yamamoto N In vivo infection of human T-cell leukemia virus type I in non-T cells Virology 1993 196 25 33 8356797 10.1006/viro.1993.1451
Manel N Kim FJ Kinet S Taylor N Sitbon M Battini JL The ubiquitous glucose transporter GLUT-1 is a receptor for HTLV Cell 2003 115 449 459 14622599 10.1016/S0092-8674(03)00881-X
Yasunaga J Sakai T Nosaka K Etoh K Tamiya S Koga S Mita S Uchino M Mitsuya H Matsuoka M Impaired production of naive T lymphocytes in human T-cell leukemia virus type I-infected individuals: its implications in the immunodeficient state Blood 2001 97 3177 3183 11342446 10.1182/blood.V97.10.3177
Igakura T Stinchcombe JC Goon PK Taylor GP Weber JN Griffiths GM Tanaka Y Osame M Bangham CR Spread of HTLV-I between lymphocytes by virus-induced polarization of the cytoskeleton Science 2003 299 1713 1716 12589003 10.1126/science.1080115
Hino S Sugamura K, Uchiyama T, Matsuoka M and Kannagi M Primary prevention of ATL Two decades of adult T-cell leukemia and HTLV-I research 2003 Tokyo, Japan Sicentific Societies Press 241 251
Okochi K Sato H Hinuma Y A retrospective study on transmission of adult T cell leukemia virus by blood transfusion: seroconversion in recipients Vox Sang 1984 46 245 253 6328765
Ando Y Ekuni Y Matsumoto Y Nakano S Saito K Kakimoto K Tanigawa T Kawa M Toyama T Long-term serological outcome of infants who received frozen-thawed milk from human T-lymphotropic virus type-I positive mothers J Obstet Gynaecol Res 2004 30 436 438 15566458 10.1111/j.1447-0756.2004.00227.x
Etoh K Yamaguchi K Tokudome S Watanabe T Okayama A Stuver S Mueller N Takatsuki K Matsuoka M Rapid quantification of HTLV-I provirus load: detection of monoclonal proliferation of HTLV-I-infected cells among blood donors Int J Cancer 1999 81 859 864 10362130 10.1002/(SICI)1097-0215(19990611)81:6<859::AID-IJC4>3.0.CO;2-K
Iga M Okayama A Stuver S Matsuoka M Mueller N Aoki M Mitsuya H Tachibana N Tsubouchi H Genetic evidence of transmission of human T cell lymphotropic virus type 1 between spouses J Infect Dis 2002 185 691 695 11865428 10.1086/339002
Li HC Biggar RJ Miley WJ Maloney EM Cranston B Hanchard B Hisada M Provirus load in breast milk and risk of mother-to-child transmission of human T lymphotropic virus type I J Infect Dis 2004 190 1275 1278 15346338 10.1086/423941
Hisada M Maloney EM Sawada T Miley WJ Palmer P Hanchard B Goedert JJ Manns A Virus markers associated with vertical transmission of human T lymphotropic virus type 1 in Jamaica Clin Infect Dis 2002 34 1551 1557 12032888 10.1086/340537
Etoh K Tamiya S Yamaguchi K Okayama A Tsubouchi H Ideta T Mueller N Takatsuki K Matsuoka M Persistent clonal proliferation of human T-lymphotropic virus type I-infected cells in vivo Cancer Res 1997 57 4862 4867 9354450
Cavrois M Leclercq I Gout O Gessain A Wain-Hobson S Wattel E Persistent oligoclonal expansion of human T-cell leukemia virus type 1-infected circulating cells in patients with Tropical spastic paraparesis/HTLV-1 associated myelopathy Oncogene 1998 17 77 82 9671316 10.1038/sj.onc.1201906
Matsuoka M Human T-cell leukemia virus type I and adult T-cell leukemia Oncogene 2003 22 5131 5140 12910250 10.1038/sj.onc.1206551
Okayama A Stuver S Matsuoka M Ishizaki J Tanaka G Kubuki Y Mueller N Hsieh CC Tachibana N Tsubouchi H Role of HTLV-1 proviral DNA load and clonality in the development of adult T-cell leukemia/lymphoma in asymptomatic carriers Int J Cancer 2004 110 621 625 15122598 10.1002/ijc.20144
LaGrenade L Hanchard B Fletcher V Cranston B Blattner W Infective dermatitis of Jamaican children: a marker for HTLV-I infection Lancet 1990 336 1345 1347 1978165 10.1016/0140-6736(90)92896-P
Nakao K Ohba N Matsumoto M Noninfectious anterior uveitis in patients infected with human T-lymphotropic virus type I Jpn J Ophthalmol 1989 33 472 481 2576286
Kannagi M Harada S Maruyama I Inoko H Igarashi H Kuwashima G Sato S Morita M Kidokoro M Sugimoto M Predominant recognition of human T cell leukemia virus type I (HTLV-I) pX gene products by human CD8+ cytotoxic T cells directed against HTLV- I-infected cells Int Immunol 1991 3 761 767 1911545
Furukawa Y Kubota R Tara M Izumo S Osame M Existence of escape mutant in HTLV-I tax during the development of adult T-cell leukemia Blood 2001 97 987 993 11159527 10.1182/blood.V97.4.987
Takeda S Maeda M Morikawa S Taniguchi Y Yasunaga J Nosaka K Tanaka Y Matsuoka M Genetic and epigenetic inactivation of tax gene in adult T-cell leukemia cells Int J Cancer 2004 109 559 567 14991578 10.1002/ijc.20007
Koiwa T Hamano-Usami A Ishida T Okayama A Yamaguchi K Kamihira S Watanabe T 5'-long terminal repeat-selective CpG methylation of latent human T-cell leukemia virus type 1 provirus in vitro and in vivo J Virol 2002 76 9389 9397 12186921 10.1128/JVI.76.18.9389-9397.2002
Tamiya S Matsuoka M Etoh K Watanabe T Kamihira S Yamaguchi K Takatsuki K Two types of defective human T-lymphotropic virus type I provirus in adult T-cell leukemia Blood 1996 88 3065 3073 8874205
Tomlinson I Bodmer W Selection, the mutation rate and cancer: ensuring that the tail does not wag the dog Nat Med 1999 5 11 12 9883827 10.1038/4687
Sakashita A Hattori T Miller CW Suzushima H Asou N Takatsuki K Koeffler HP Mutations of the p53 gene in adult T-cell leukemia Blood 1992 79 477 480 1730092
Cesarman E Chadburn A Inghirami G Gaidano G Knowles DM Structural and functional analysis of oncogenes and tumor suppressor genes in adult T-cell leukemia/lymphoma shows frequent p53 mutations Blood 1992 80 3205 3216 1361372
Nishimura S Asou N Suzushima H Okubo T Fujimoto T Osato M Yamasaki H Lisha L Takatsuki K p53 gene mutation and loss of heterozygosity are associated with increased risk of disease progression in adult T cell leukemia Leukemia 1995 9 598 604 7723391
Hatta Y Hirama T Miller CW Yamada Y Tomonaga M Koeffler HP Homozygous deletions of the p15 (MTS2) and p16 (CDKN2/MTS1) genes in adult T-cell leukemia Blood 1995 85 2699 2704 7742529
Nosaka K Maeda M Tamiya S Sakai T Mitsuya H Matsuoka M Increasing methylation of the CDKN2A gene is associated with the progression of adult T-cell leukemia Cancer Res 2000 60 1043 1048 10706122
Morosetti R Kawamata N Gombart AF Miller CW Hatta Y Hirama T Said JW Tomonaga M Koeffler HP Alterations of the p27KIP1 gene in non-Hodgkin's lymphomas and adult T-cell leukemia/lymphoma Blood 1995 86 1924 1930 7655021
Nagata S Fas ligand-induced apoptosis Annu Rev Genet 1999 33 29 55 10690403 10.1146/annurev.genet.33.1.29
Tamiya S Etoh K Suzushima H Takatsuki K Matsuoka M Mutation of CD95 (Fas/Apo-1) gene in adult T-cell leukemia cells Blood 1998 91 3935 3942 9573032
Maeda T Yamada Y Moriuchi R Sugahara K Tsuruda K Joh T Atogami S Tsukasaki K Tomonaga M Kamihira S Fas gene mutation in the progression of adult T cell leukemia J Exp Med 1999 189 1063 1071 10190897 10.1084/jem.189.7.1063
Krueger A Fas SC Baumann S Krammer PH The role of CD95 in the regulation of peripheral T-cell apoptosis Immunol Rev 2003 193 58 69 12752671 10.1034/j.1600-065X.2003.00047.x
Yasunaga J Taniguchi Y Nosaka K Yoshida M Satou Y Sakai T Mitsuya H Matsuoka M Identification of aberrantly methylated genes in association with adult T-cell leukemia Cancer Res 2004 64 6002 6009 15342380
Mittelstadt PR Ashwell JD Cyclosporin A-sensitive transcription factor Egr-3 regulates Fas ligand expression Mol Cell Biol 1998 18 3744 3751 9632757
Yoshida M Nosaka K Yasunaga J Nishikata I Morishita K Matsuoka M Aberrant expression of the MEL1S gene identified in association with hypomethylation in adult T-cell leukemia cells Blood 2004 103 2753 2760 14656887 10.1182/blood-2003-07-2482
Sasaki H Nishikata I Shiraga T Akamatsu E Fukami T Hidaka T Kubuki Y Okayama A Hamada K Okabe H Murakami Y Tsubouchi H Morishita K Overexpression of a cell adhesion molecule, TSLC1, as a possible molecular marker for acute-type adult T-cell leukemia Blood 2005 105 1204 1213 15471956 10.1182/blood-2004-03-1222
Yoshizuka N Moriuchi R Mori T Yamada K Hasegawa S Maeda T Shimada T Yamada Y Kamihira S Tomonaga M Katamine S An alternative transcript derived from the trio locus encodes a guanosine nucleotide exchange factor with mouse cell-transforming potential J Biol Chem 2004 279 43998 44004 15308664 10.1074/jbc.M406082200
Bangham CR Human T-lymphotropic virus type 1 (HTLV-1): persistence and immune control Int J Hematol 2003 78 297 303 14686486
Wodarz D Hall SE Usuku K Osame M Ogg GS McMichael AJ Nowak MA Bangham CR Cytotoxic T-cell abundance and virus load in human immunodeficiency virus type 1 and human T-cell leukaemia virus type 1 Proc R Soc Lond B Biol Sci 2001 268 1215 1221 10.1098/rspb.2001.1608
Vine AM Heaps AG Kaftantzi L Mosley A Asquith B Witkover A Thompson G Saito M Goon PK Carr L Martinez-Murillo F Taylor GP Bangham CR The role of CTLs in persistent viral infection: cytolytic gene expression in CD8+ lymphocytes distinguishes between individuals with a high or low proviral load of human T cell lymphotropic virus type 1 J Immunol 2004 173 5121 5129 15470056
Jeffery KJ Usuku K Hall SE Matsumoto W Taylor GP Procter J Bunce M Ogg GS Welsh KI Weber JN Lloyd AL Nowak MA Nagai M Kodama D Izumo S Osame M Bangham CR HLA alleles determine human T-lymphotropic virus-I (HTLV-I) proviral load and the risk of HTLV-I-associated myelopathy Proc Natl Acad Sci U S A 1999 96 3848 3853 10097126 10.1073/pnas.96.7.3848
Vine AM Witkover AD Lloyd AL Jeffery KJ Siddiqui A Marshall SE Bunce M Eiraku N Izumo S Usuku K Osame M Bangham CR Polygenic control of human T lymphotropic virus type I (HTLV-I) provirus load and the risk of HTLV-I-associated myelopathy/tropical spastic paraparesis J Infect Dis 2002 186 932 939 12232833 10.1086/342953
Tsukasaki K Miller CW Kubota T Takeuchi S Fujimoto T Ikeda S Tomonaga M Koeffler HP Tumor necrosis factor alpha polymorphism associated with increased susceptibility to development of adult T-cell leukemia/lymphoma in human T-lymphotropic virus type 1 carriers Cancer Res 2001 61 3770 3774 11325850
Kawano F Tsuda H Yamaguchi K Nishimura H Sanada I Matsuzaki H Ishii M Takatsuki K Unusual clinical courses of adult T-cell leukemia in siblings Cancer 1984 54 131 134 6609759
Miyamoto Y Yamaguchi K Nishimura H Takatsuki K Motoori T Morimatsu M Yasaka T Ohya I Koga T Familial adult T-cell leukemia Cancer 1985 55 181 185 2981140
Yamaguchi K Yul LS Shimizu T Nozawa F Takeya M Takahashi K Takatsuki K Concurrence of lymphoma type adult T-cell leukemia in three sisters Cancer 1985 56 1688 1690 2992744
Shimamoto Y Kikuchi M Funai N Suga K Matsuzaki M Yamaguchi M Spontaneous regression in adult T-cell leukemia/lymphoma Cancer 1993 72 735 740 8101469
Kannagi M Ohashi T Harashima N Hanabuchi S Hasegawa A Immunological risks of adult T-cell leukemia at primary HTLV-I infection Trends Microbiol 2004 12 346 352 15223062 10.1016/j.tim.2004.05.005
Matsuoka M Takatsuki K Henderson ES, Lister TA and Greaves MF Adult T-cell leukemia Leukemia 2002 7th edition Philadelphia, Saunders 705 712
Welles SL Tachibana N Okayama A Shioiri S Ishihara S Murai K Mueller NE Decreased reactivity to PPD among HTLV-I carriers in relation to virus and hematologic status Int J Cancer 1994 56 337 340 8314320
Hisada M Okayama A Shioiri S Spiegelman DL Stuver SO Mueller NE Risk factors for adult T-cell leukemia among carriers of human T-lymphotropic virus type I Blood 1998 92 3557 3561 9808547
Arisawa K Katamine S Kamihira S Kurokawa K Sawada T Soda M Doi H Saito H Shirahama S A nested case-control study of risk factors for adult T-cell leukemia/lymphoma among human T-cell lymphotropic virus type-I carriers in Japan Cancer Causes Control 2002 13 657 663 12296513 10.1023/A:1019511224501
Hanabuchi S Ohashi T Koya Y Kato H Hasegawa A Takemura F Masuda T Kannagi M Regression of human T-cell leukemia virus type I (HTLV-I)-associated lymphomas in a rat model: peptide-induced T-cell immunity J Natl Cancer Inst 2001 93 1775 1783 11734593 10.1093/jnci/93.23.1775
Tobinai K Ohtsu T Hayashi M Kinoshita T Matsuno Y Mukai K Shimoyama M Epstein-Barr virus (EBV) genome carrying monoclonal B-cell lymphoma in a patient with adult T-cell leukemia-lymphoma Leuk Res 1991 15 837 846 1656151 10.1016/0145-2126(91)90468-9
Hoshida Y Li T Dong Z Tomita Y Yamauchi A Hanai J Aozasa K Lymphoproliferative disorders in renal transplant patients in Japan Int J Cancer 2001 91 869 875 11275994 10.1002/1097-0215(200002)9999:9999<::AID-IJC1125>3.0.CO;2-N
Stevens HP Holterman L Haaksma AG Jonker M Heeney JL Lymphoproliferative disorders developing after transplantation and their relation to simian T-cell leukemia virus infection Transpl Int 1992 5 Suppl 1 S450 3 14621843
Hanon E Stinchcombe JC Saito M Asquith BE Taylor GP Tanaka Y Weber JN Griffiths GM Bangham CR Fratricide among CD8(+) T lymphocytes naturally infected with human T cell lymphotropic virus type I Immunity 2000 13 657 664 11114378 10.1016/S1074-7613(00)00065-0
Karube K Ohshima K Tsuchiya T Yamaguchi T Kawano R Suzumiya J Utsunomiya A Harada M Kikuchi M Expression of FoxP3, a key molecule in CD4CD25 regulatory T cells, in adult T-cell leukaemia/lymphoma cells Br J Haematol 2004 126 81 84 15198736 10.1111/j.1365-2141.2004.04999.x
Yoshie O Fujisawa R Nakayama T Harasawa H Tago H Izawa D Hieshima K Tatsumi Y Matsushima K Hasegawa H Kanamaru A Kamihira S Yamada Y Frequent expression of CCR4 in adult T-cell leukemia and human T-cell leukemia virus type 1-transformed T cells Blood 2002 99 1505 1511 11861261 10.1182/blood.V99.5.1505
Hasegawa H Nomura T Kohno M Tateishi N Suzuki Y Maeda N Fujisawa R Yoshie O Fujita S Increased chemokine receptor CCR7/EBI1 expression enhances the infiltration of lymphoid organs by adult T-cell leukemia cells Blood 2000 95 30 38 10607681
Higashimura N Takasawa N Tanaka Y Nakamura M Sugamura K Induction of OX40, a receptor of gp34, on T cells by trans-acting transcriptional activator, Tax, of human T-cell leukemia virus type I Jpn J Cancer Res 1996 87 227 231 8613423
Imura A Hori T Imada K Ishikawa T Tanaka Y Maeda M Imamura S Uchiyama T The human OX40/gp34 system directly mediates adhesion of activated T cells to vascular endothelial cells J Exp Med 1996 183 2185 2195 8642328 10.1084/jem.183.5.2185
Kiyokawa T Yamaguchi K Takeya M Takahashi K Watanabe T Matsumoto T Lee SY Takatsuki K Hypercalcemia and osteoclast proliferation in adult T-cell leukemia Cancer 1987 59 1187 1191 2880656
Arai F Miyamoto T Ohneda O Inada T Sudo T Brasel K Miyata T Anderson DM Suda T Commitment and differentiation of osteoclast precursor cells by the sequential expression of c-Fms and receptor activator of nuclear factor kappaB (RANK) receptors J Exp Med 1999 190 1741 1754 10601350 10.1084/jem.190.12.1741
Nosaka K Miyamoto T Sakai T Mitsuya H Suda T Matsuoka M Mechanism of hypercalcemia in adult T-cell leukemia: overexpression of receptor activator of nuclear factor kappaB ligand on adult T-cell leukemia cells Blood 2002 99 634 640 11781248 10.1182/blood.V99.2.634
Fukumoto S Matsumoto T Watanabe T Takahashi H Miyoshi I Ogata E Secretion of parathyroid hormone-like activity from human T-cell lymphotropic virus type I-infected lymphocytes Cancer Res 1989 49 3849 3852 2544261
Watanabe T Yamaguchi K Takatsuki K Osame M Yoshida M Constitutive expression of parathyroid hormone-related protein gene in human T cell leukemia virus type 1 (HTLV-1) carriers and adult T cell leukemia patients that can be trans-activated by HTLV-1 tax gene J Exp Med 1990 172 759 765 2388034 10.1084/jem.172.3.759
Yamada Y Tomonaga M Fukuda H Hanada S Utsunomiya A Tara M Sano M Ikeda S Takatsuki K Kozuru M Araki K Kawano F Niimi M Tobinai K Hotta T Shimoyama M A new G-CSF-supported combination chemotherapy, LSG15, for adult T-cell leukaemia-lymphoma: Japan Clinical Oncology Group Study 9303 Br J Haematol 2001 113 375 382 11380402 10.1046/j.1365-2141.2001.02737.x
Mori N Fujii M Ikeda S Yamada Y Tomonaga M Ballard DW Yamamoto N Constitutive activation of NF-kappaB in primary adult T-cell leukemia cells Blood 1999 93 2360 2368 10090947
Tan C Waldmann TA Proteasome inhibitor PS-341, a potential therapeutic agent for adult T-cell leukemia Cancer Res 2002 62 1083 1086 11861386
Satou Y Nosaka K Koya Y Yasunaga JI Toyokuni S Matsuoka M Proteasome inhibitor, bortezomib, potently inhibits the growth of adult T-cell leukemia cells both in vivo and in vitro Leukemia 2004 18 1357 1363 15190257 10.1038/sj.leu.2403400
Mitra-Kaushik S Harding JC Hess J Ratner L Effects of the Proteasome Inhibitor PS-341 on Tumor Growth in HTLV-1 Tax Transgenic Mice and Tax Tumor Transplants Blood 2004
Nasr R El-Sabban ME Karam JA Dbaibo G Kfoury Y Arnulf B Lepelletier Y Bex F de The H Hermine O Bazarbachi A Efficacy and mechanism of action of the proteasome inhibitor PS-341 in T-cell lymphomas and HTLV-I associated adult T-cell leukemia/lymphoma Oncogene 2005 24 419 430 15543232 10.1038/sj.onc.1208212
Mori N Yamada Y Ikeda S Yamasaki Y Tsukasaki K Tanaka Y Tomonaga M Yamamoto N Fujii M Bay 11-7082 inhibits transcription factor NF-kappaB and induces apoptosis of HTLV-I-infected T-cell lines and primary adult T-cell leukemia cells Blood 2002 100 1828 1834 12176906 10.1182/blood-2002-01-0151
Waldmann TA White JD Goldman CK Top L Grant A Bamford R Roessler E Horak ID Zaknoen S Kasten-Sportes C The interleukin-2 receptor: a target for monoclonal antibody treatment of human T-cell lymphotrophic virus I-induced adult T-cell leukemia Blood 1993 82 1701 1712 8400227
Waldmann TA White JD Carrasquillo JA Reynolds JC Paik CH Gansow OA Brechbiel MW Jaffe ES Fleisher TA Goldman CK Radioimmunotherapy of interleukin-2R alpha-expressing adult T-cell leukemia with Yttrium-90-labeled anti-Tac Blood 1995 86 4063 4075 7492762
Zhang Z Zhang M Ravetch JV Goldman C Waldmann TA Effective therapy for a murine model of adult T-cell leukemia with the humanized anti-CD2 monoclonal antibody, MEDI-507 Blood 2003 102 284 288 12649132 10.1182/blood-2002-11-3601
Ishida T Iida S Akatsuka Y Ishii T Miyazaki M Komatsu H Inagaki H Okada N Fujita T Shitara K Akinaga S Takahashi T Utsunomiya A Ueda R The CC chemokine receptor 4 as a novel specific molecular target for immunotherapy in adult T-Cell leukemia/lymphoma Clin Cancer Res 2004 10 7529 7539 15569983
Borg A Yin JA Johnson PR Tosswill J Saunders M Morris D Successful treatment of HTLV-1-associated acute adult T-cell leukaemia lymphoma by allogeneic bone marrow transplantation Br J Haematol 1996 94 713 715 8826899 10.1046/j.1365-2141.1996.02338.x
Utsunomiya A Miyazaki Y Takatsuka Y Hanada S Uozumi K Yashiki S Tara M Kawano F Saburi Y Kikuchi H Hara M Sao H Morishima Y Kodera Y Sonoda S Tomonaga M Improved outcome of adult T cell leukemia/lymphoma with allogeneic hematopoietic stem cell transplantation Bone Marrow Transplant 2001 27 15 20 11244433 10.1038/sj.bmt.1702731
Okamura J Utsunomiya A Tanosaki R Uike N Sonoda S Kannagi M Tomonaga M Harada M Kimura N Masuda M Kawano F Yufu Y Hattori H Kikuchi H Saburi Y Allogeneic stem cell transplantation with reduced conditioning intensity as a novel immunotherapy and antiviral therapy for adult T-cell leukemia/lymphoma. Blood 2005 in press
Harashima N Kurihara K Utsunomiya A Tanosaki R Hanabuchi S Masuda M Ohashi T Fukui F Hasegawa A Masuda T Takaue Y Okamura J Kannagi M Graft-versus-Tax response in adult T-cell leukemia patients after hematopoietic stem cell transplantation Cancer Res 2004 64 391 399 14729650
Van Dooren S Pybus OG Salemi M Liu HF Goubau P Remondegui C Talarmin A Gotuzzo E Alcantara LC Galvao-Castro B Vandamme AM The low evolutionary rate of human T-cell lymphotropic virus type-1 confirmed by analysis of vertical transmission chains Mol Biol Evol 2004 21 603 611 14739252 10.1093/molbev/msh053
| 15854229 | PMC1131926 | CC BY | 2021-01-04 16:36:40 | no | Retrovirology. 2005 Apr 26; 2:27 | utf-8 | Retrovirology | 2,005 | 10.1186/1742-4690-2-27 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-361583310610.1186/1465-9921-6-36ResearchInhibitors of mitogen-activated protein kinases differentially regulate costimulated T cell cytokine production and mouse airway eosinophilia Chialda Ligia [email protected] Meixia [email protected] Kay [email protected] Andreas [email protected] Department of Experimental and Clinical Pharmacology and Toxicology, University of Erlangen-Nürnberg, Fahrstr. 17, D-91054 Erlangen, Germany2 Present Address Department of Clinical Pharmacology, Chinese Medical University, Shenyang, China2005 15 4 2005 6 1 36 36 28 7 2004 15 4 2005 Copyright © 2005 Chialda et al; licensee BioMed Central Ltd.2005Chialda 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
T cells play a dominant role in the pathogenesis of asthma. Costimulation of T cells is necessary to fully activate them. An inducible costimulator (ICOS) of T cells is predominantly expressed on Th2 cells. Therefore, interference of signaling pathways precipitated by ICOS may present new therapeutic options for Th2 dominated diseases such as asthma. However, these signaling pathways are poorly characterized in vitro and in vivo.
Methods
Human primary CD4+ T cells from blood were activated by beads with defined combinations of surface receptor stimulating antibodies and costimulatory receptor ligands. Real-time RT-PCR was used for measuring the production of cytokines from activated T cells. Activation of mitogen activated protein kinase (MAPK) signaling pathways leading to cytokine synthesis were investigated by western blot analysis and by specific inhibitors. The effect of inhibitors in vivo was tested in a murine asthma model of late phase eosinophilia. Lung inflammation was assessed by differential cell count of the bronchoalveolar lavage, determination of serum IgE and lung histology.
Results
We showed in vitro that ICOS and CD28 are stimulatory members of an expanding family of co-receptors, whereas PD1 ligands failed to co-stimulate T cells. ICOS and CD28 activated different MAPK signaling cascades necessary for cytokine activation. By means of specific inhibitors we showed that p38 and ERK act downstream of CD28 and that ERK and JNK act downstream of ICOS leading to the induction of various T cell derived cytokines. Using a murine asthma model of late phase eosinophilia, we demonstrated that the ERK inhibitor U0126 and the JNK inhibitor SP600125 inhibited lung inflammation in vivo. This inhibition correlated with the inhibition of Th2 cytokines in the BAL fluid. Despite acting on different signaling cascades, we could not detect synergistic action of any combination of MAPK inhibitors. In contrast, we found that the p38 inhibitor SB203580 antagonizes the action of the ERK inhibitor U0126 in vitro and in vivo.
Conclusion
These results demonstrate that the MAPKs ERK and JNK may be suitable targets for anti-inflammatory therapy of asthma, whereas inhibition of p38 seems to be an unlikely target.
==== Body
Background
Asthma is a chronic inflammatory disease of the airways. The inflammation characterizing asthma is complex and involves multiple cells and mediators. The cells involved include well-recognized immune and inflammatory cells, lymphocytes, macrophages, eosinophils, mast cells and neutrophils as well as resident lung cells [1,2]. The importance of allergen-specific CD4+ Th2 cells has been demonstrated. Th2-associated cytokines such as IL-4, IL-5, IL-9 and IL-13 are known to be involved in IgE production, airway eosinophilia, and airway hyperresponsiveness [3]. Consequently, the inhibition or modulation of allergen-specific Th2 cells and their cytokines has become an attractive target for novel therapeutic intervention strategies [4-6]. Downregulation of cytokine production has been achieved by glucocorticoids and immunosuppressants both in vitro and in vivo. As these agents suppressed a broad spectrum of immune function, more specific regulatory pathways of T cell need to be addressed.
In recent years, considerable effort has been mounted to dissect the signaling events in T cells. Full activation of T cells requires signaling through both the TCR/CD3 complex and the CD28 costimulatory receptor [7-10]. CD28 engagement by B7-1 and B7-2 on resting T cells provides a primary costimulatory signal critical for initial cell cycle progression, interleukin 2 production and clonal expansion [11]. Engagement of CTLA-4 by the same B7-1 or B7-2 ligands results in attenuation of T cell responses. Recently, molecular homologues of CD28 and CTLA-4 coreceptors and their B7-like ligands have been identified. These homologues presumably play an essential role in the acquisition of effector function and/or tolerance induction. One of the CD28-like molecules is induced during activation of T cells, thus it is referred as an inducible costimulator (ICOS) and has a unique B7-like ligand (B7-H2, B7 h or B7RP-1). PD-1 is an inhibitory CD28-like receptor, with two B7-like ligands (PD-L1 and PD-L2) [12,13]. It has been shown that the ICOS/B7-H2 pathway controls T cell dependent immune responses [14-17].
Several protein kinases such as the family of MAPKs are involved in the transmission of extracellular signals into the nucleus. [18]. MAPKs are serine/threonine kinases that include extracellular signal-regulated kinases (ERKs) [19], Jun NH2-terminal kinases (JNKs) [20] and p38 MAPK [21]. CD3 signaling alone has been shown to activate ERK, and the combination of CD3 and CD28 signaling can synergistically activate JNK in T cell lines and clones [22,23]. In these studies, CD28 signaling alone was shown to minimally activate ERK and JNK, while others have reported much more significant ERK activation [24]. Activation of p38 has also been observed in Jurkat cells and in mouse T cells treated with CD3/CD28 [25,26]. While it has become clear that productive T cell activation requires the cooperation of multiple signaling pathways, the composition and the relative contribution of these pathways to the activation process and cytokine production remain yet to be fully deciphered.
The predominant expression of ICOS on Th2 cells and its effect on the development of Th2 cells has been reported [15,27]. Therefore, interference of signaling pathways precipitated by ICOS may present new therapeutic options for Th2 dominated diseases such as asthma. However, these signaling pathways are poorly characterized. Recently, we reported that the ERK pathway plays a pivotal role in regulating IL-13 production in human CD4+ T cells after CD28 costimulation [28]. In the present study, we studied the activation of different MAP kinases in human CD4+ T cells stimulated by α-CD3/α-ICOS, their function in the production of cytokines from these cells and compared these results to α-CD3/α-CD28 stimulated T cells. Our results demonstrate, that in human CD4+ T cells, p42/p44 ERK, p38 and JNK differentially regulate the production of IL-2, IL-4, IL-13 and IFN-γ in response to α-CD3/α-CD28 or α-CD3/α-ICOS. Furthermore, inhibitors able to inhibit ICOS induced signaling proved to be effective in the murine asthma model of late phase eosinophilia.
Methods
Reagents
Oligonucleotides were synthesized by TIB Molbiol (Berlin, Germany). DMSO, TPA, ionomycin, dexamethasone, Histopaque-1077 was from Sigma Chemical Co (Deisenhofen, Germany). U0126, SP600125 and SB203580 were purchased from Biotrend Chemikalien GmbH (Köln, Germany). Purified anti-human CD3 and purified anti-human CD28 were from PharMingen Becton Dickinson Co (Heidelberg, Germany). RPMI 1640 medium was from Life Technologies (Heidelberg, Germany). Unless otherwise indicated, all other chemicals were purchased from the Sigma Chemical Co (Deisenhofen, Germany).
Preparation of CD4+ T cells
Buffy coats from healthy human volunteers were obtained from the Erlangen Blood Bank. PBMCs were isolated by density gradient centrifugation over Histopaque 1077 (Sigma, Deisenhofen, Germany), washed twice in Hanks buffer (Life Technologies, Heidelberg, Germany) and resuspended in RPMI 1640 medium (Roche Diagnostics, Penzberg, Germany). PBMCs were incubated with a hapten-antibody cocktail (containing monoclonal hapten-conjugated CD8, CD11b, CD16, CD19, CD36 and CD56 antibodies, Miltenyi Biotec, Bergisch Gladbach, Germany) and MACS anti-hapten microbeads (CD4+ T cell isolation kit, Miltenyi Biotec, Bergisch Gladbach, Germany). CD4+ T cells were isolated by negative selection on LS+ columns using a high gradient magnetic cell separation system MACS (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions. Purity of CD4+ T cells was assessed by flow cytometry and was 90–95% (FACScan, Becton Dickinson, Heidelberg, Germany). Additional cells were monocytes (4–8%) and B cells (1–2%), which did not respond to α-CD3 antibodies. Purified CD4+ T cells were resuspended in RPMI 1640 medium.
Cell culture
For cytokine production, CD4+ T cells were resuspended at 0.25 × 106 cells/ml and incubated in 500 μl aliquots in 24-well tissue culture plates (Falcon Becton Dickinson Labware, Heidelberg, Germany) at 37°C, 5%CO2. After preincubation with test substances for 30 min, cells were stimulated with α-CD3/α-CD28 Dynal beads (8*104/2 μl/well). B7-H1, B7-H2 and B7-H3 binding by their receptors was assessed by stimulation of the cultured cells with self coated Dynabeads M-450 Epoxy (Dynal Biotech GmbH, Hamburg, Germany). α-CD3 mAb (0.1 μg/μl) (Pharmingen, Heidelberg, Germany) together with 0.04 μg/μl recombinant fusion proteins of B7-H1, B7-H2 and B7-H3 respectively, (R&D, Wiesbaden-Nordenstadt, Germany), were added to the prewashed beads according to the manufacturer's recommendations. The beads were added at a ratio of 3:1 to the CD4+ T cells. At the indicated times (presented in figure legends), cells were sedimented by centrifugation, the supernatants were harvested and kept frozen at -80°C until cytokine protein determination; the cell pellet was lysed by RLT lysis buffer (Qiagen, Hilden, Germany) and frozen at -80°C until RNA isolation.
Enzyme-linked immunosorbent assay
Cytokine measurements in culture supernatants were done by sandwich ELISA using matched antibody pairs (BD Pharmingen, Heidelberg, Germany). ELISA plates (Maxisorb, Nunc) were coated overnight with anti-cytokine mAb in 0.1 M carbonate buffer, pH 9.5. After being washed, plates were blocked with Assay Diluent (Pharmingen, Heidelberg, Germany) for 1 h and washed again. Appropriately diluted supernatant samples and standards were distributed in duplicates and the plates were incubated for 2 h at room temperature. Plates were washed, incubated for 1 h with working detector (biotinylated anti-cytokine Ab and Avidin-horseradish peroxidase conjugate). After washing, substrate (TMB and hydrogen peroxide) was added. The reaction was stopped by addition of 1 M H3PO4. Plates were read at 450 nm (reference 570 nm) in a microplate reader (Dynatech). The results were expressed as a percentage of the control level of cytokines production by cells stimulated in the presence of the vehicle of the corresponding compound.
Determination of serum levels of total and OVA-specific IgE
Serum levels of OVA-specific IgE were measured by IgE ELISA (OptEIA, BD Pharmingen). Briefly, 96-well plates (Maxisorb, Nunc) were coated with OVA (100 μg/ml). After addition of serum samples, a sheep anti-IgE Ab was added to individual wells and its binding was detected with peroxidase-conjugated anti-sheep IgG. IgE concentrations were calculated by comparison with commercial mouse IgE standards (BD Pharmingen).
Analysis of cytokine mRNA expression by real-time RT-PCR
RNA was prepared from frozen lysates using Rneasy (QIAGEN, Hilden, Germany). One-tube RT-PCR was performed using Quantitect Probe RT-PCR Kit from QIAGEN (Hilden, Germany). Expression of cytokines were determined in relation to beta-actin by real time RT-PCR using TaqMan assay on a ABI Prism 7900. Human primers and probes have been described [6]. Murine primers and probes were Assay-On-Demand purchased from Applied Biosystems (Darmstadt, Germany). Quantity of mRNA was calculated using the ΔΔCT method (PE Applied Biosystems User Bulletin #2; ABI PRISM 7700 Sequence Detection System, 1997). For each RT-PCR the threshold cycle (CT) was determined, being defined as the cycle at which the fluorescence exceeds 10 times the standard deviation of the mean baseline emission for cycles 3 to 10. Cytokine mRNA levels were normalized to the housekeeping gene β-actin according to the following formula: ΔCT = CTβ-actin - CTcytokine. Subsequently, respective cytokine mRNA levels were calculated using the ΔΔCT method, i.e., ΔCT values representing mRNA from cells treated with stimulus in combination with a test compound were set in relation to the ΔCT value representing mRNA levels from cells treated with stimulus alone according to the following formula: ΔΔCT = ΔCT(drug) - ΔCT(vehicle). The relative mRNA level for the respective test compound was calculated as 2-ΔΔCT * 100% based on the results of control experiments with an efficiency of the PCR reaction of approximately 100%.
Western Blotting analysis of the MAPKs
Human CD4+ T cells were stimulated with Dynabeads CD3/CD28 T Cell Expander (Dynal Biotech GmbH, Hamburg, Germany) or with Dynal beads coated with α-CD3/B7-H2 Fc for 0, 5, 15, 60 or 120 min. After stimulation, cells were washed with ice-cold PBS and lysed in cell lysis buffer (Tris 50 mM, pH 7.2, containing 10 mM EDTA, 150 mM NaCl, 1% NaDOC, 0.1% SDS, 1% Triton X-100) supplemented with protease inhibitors (1 mM PMSF, 5 μg/μl Aprotinin and 5 μg/μl Leupeptin). Cells were left on ice for 20 min and then centrifuged at 14.000 rpm for 10 min. The supernatants were stored at -80°C until measurement of the MAPKs. Equal amounts of protein were resolved in 10% SDS-polyacrylamid gel electrophoresis and transferred onto PVDF membrane. Membranes were blocked overnight at 4°C with 0.2 % I-block (Applied Biosystems, Darmstadt, Germany) in TBS/T buffer and probed with rabbit primary antibodies specific for the dually phosphorylated active forms of ERK, p38 and JNK (Promega GmbH, Mannheim, Germany) for 2 h at room temperature. Blots were washed, incubated 1 h at room temperature with second antibody (donkey anti-rabbit horseradish peroxidase conjugated) (Promega GmbH, Mannheim, Germany). Immunoreactive bands were visualized by enhanced chemiluminiscence (ECL) according to the manufacturer's recommendations (Amersham Pharmacia Biotech, Freiburg, Germany). After detection of the active form of the MAPK, membranes were stripped and reprobed with the corresponding antibody that recognize both active and inactive forms of p38, JNK (Santa Cruz Biotechnology, Inc., Heidelberg, Germany) and ERK (Promega GmbH, Mannheim, Germany).
ELISA detection of activated JNK
The FACE (Fast Activated Cell-based ELISA) kit (Active Motif, Rixensart, Belgium) was used according to the provider's protocol. CD4+ T cells were cultured in precoated (poly-L-lysine) (Sigma, Deisenhofen, Germany) 96-well plates at a concentration of 1 × 106 /ml and were stimulated with α-CD3/α-CD28 and α-CD3/B7-H2 Fc for different time points (1 min, 5 min, 15 min, 60 min and 120 min) to induce the JNK pathway. Following stimulation the cells were fixed in order to preserve protein modifications, including phosphorylation. Corresponding wells were then incubated with a primary antibody specific for pJNK and total JNK respectively. Subsequent incubation with secondary HRP-conjugated antibody and developing solution was followed by a colorimetric detection. The signals were then normalized for cell number using Crystal Violet.
Murine model of late phase eosinophilia
Male Balb/c mice weighing 22–25 g were used. Animals were purchased from Harlan (Borchen, Germany). The animals were kept under constant environmental conditions (temperature: 18 ± 2°C, humidity: 40–60 %, light cycle: 7 am – 7 pm). They had free access to standardized food pellets (purchased from Altromin, Lage, Germany) and tap water. All animal studies were performed in accordance with the national animal protection rules and permitted by the local governmental authority (Regierung von Mittelfranken, Germany). The sensitization strategy was adapted from previously described mouse model [29]. Briefly, all Balb/c mice were actively immunized and boosted by intraperitoneal injections with 200 μl of a solution containing 100 μg ovalbumin (grade V, Sigma, Taufkirchen, Germany) adsorbed on 4 mg of an aqueous solution of aluminium hydroxide and magnesium hydroxide (Alum) (Perbio Science, Bonn, Germany) on days 1, 14 and 20. On day 26, the animals were used for experiments. Compounds, as suspension in 5% DMSO, were given intraperitoneally two hours prior to challenge once a day. Control animals (negative and positive) received only vehicle solution. Mice were challenged twice a day for 4 days by intranasal application of 50 μl of 2 μg/μl ovalbumin solution in NaCl to provoke an influx of inflammatory cells into the airways. Negative control mice were sensitized intraperitoneally with OVA/Alum and were challenged intranasally with saline solution. Positive control animals were sensitized as mentioned above and challenged intranasally with ovalbumin. On day 30 animals were sacrificed by putting them into a CO2 rich atmosphere and a bronchoalveolar lavage (BAL) was performed by flushing the lungs and the airways six times with 0.5 ml Hank's balanced solution (Gibco, Karlsruhe, Germany) supplemented with Na-EDTA and HEPES (Sigma, Taufkirchen, Germany). The number of eosinophils as well as the total cell number from the pooled BAL samples of one animal was counted using a haemocytometer (Sysmex microcellcounter F-300, Norderstedt, Germany). BAL cells were spun onto glass slides using a cytospin centrifuge and stained with Diff-Quik (Dade-Behring, Marburg, Germany). The percentage and number of eosinophils were determined microscopically using standard cytological (morphological and staining) criteria by counting 400 cells/slide. Each group of animals treated with compounds was compared with saline challenged (negative control) and vehicle-treated ovalbumin challenged (positive control) groups. For RNA analysis cells recovered from broncheoalveolar lavage were sedimented by centrifugation and RNA was prepared from the cell pellet.
Histological examination of lungs
Lungs removed from the chest were instilled intratracheally with 0.8 ml 4% buffered formalin solution and immersed in this fixative for 48 hours. Tissues were embedded in paraffin and sections of 5 μm were cut. Slices were stained with May-Grünwald Giemsa for examining inflammation and eosinophilic infiltration and periodic acid-Schiffs stain was done for measuring mucus production under the light microscope. To determine the severity of peribronchial inflammation a semiquantitatively score described by Myou et al. [30] was determined. Briefly, this score describes followingcategories: 0, no inflammation; 1, few inflammatory cells; 2, a ring of inflammatorycells 1 cell layer deep; 3, a ring of inflammatory cells 2–4 cells deep; 4, a ring of inflammatory cells of >4 cells deep. The numerical scores for the abundance of PAS-positive mucus-containingcell in each airway were adapted from the same reference as follows: 0, <0.5%PAS-positive cells; 1, 5–25%; 2, 25–50%; 3, 50–75%;4, >75%.
Data analysis
Data are expressed as means ± s.e.mean. Significant differences were statistically analyzed by the unpaired Student's t-test and by ANOVA. IC50 and ED50 values were calculated using the computer program PRISM 3.0 (GraphPad Software Inc., San Diego, CA, U.S.A).
Results
Influence of coreceptors on TCR induced cytokine expression in CD4+ T cells
The stimulation of human CD4+ T cells by beads coated with α-CD3 antibodies and various ligands for coreceptors was compared. For our studies, we developed a real-time RT-PCR method for accurate quantization of T cell cytokines. Cells were stimulated with beads coated with α-CD3 alone, α-CD3 plus α-CD28, α-CD3 plus B7-2 Fc fusion protein, α-CD3 plus B7-H2 Fc fusion protein, α-CD3 plus PD-1 ligand Fc fusion protein (PD-L1 or B7-H1) or α-CD3 plus B7-H3 Fc fusion protein. Naked beads alone were included as a negative control and did not activate CD4+ T cells. IL-5 mRNA was not induced by any stimulation condition. In contrast, IL-2, IL-4, IL-10, IL-13 and IFNγ mRNA expression was already stimulated by α-CD3 activation alone and maximally stimulated by α-CD3/α-CD28 activation, which was included as a positive control (Figure 1A). B7-2 and B7-H2 binding to their respective counterreceptors costimulated cytokine gene expression to elevated levels compared to α-CD3 activation alone. Costimulation strength by ICOS ligation was comparable to CD28 costimulation. In contrast, B7-H1 and B7-H3 mediated costimulation did not augment α-CD3 induced cytokine mRNA levels (Figure 1A). To substantiate these results, cytokine protein concentration of the prototypic cytokines IL-2, IL-4 and IFN was determined in stimulated CD4+ T cell supernatants (Figure 1B). The pattern was similar to the induction on the mRNA level. B7-2 and B7-H2 costimulation produced more cytokines than α-CD3 stimulation alone. B7-2 was more effective in costimulation that B7-H2 (Figure 1B). Similarly, B7-H1 and B7-H3 mediated costimulation did not augment α-CD3 induced cytokine protein levels (data not shown).
Figure 1 Induction of cytokines in CD4+ T cells by different costimulatory receptors. Human CD4+ T cells were stimulated with various stimuli. (A) Cytokine mRNA level were determined after eight hours stimulation using real-time RT-PCR. mRNA level were normalized to β-actin and CD3 stimulated cells were set to 1. Each column represents mean ± s.e.mean of three different donors. (B) Cytokine protein level was determined in the supernatant by ELISA after twenty-four hours of stimulation. Each column represents mean ± s.e.mean of three different volunteers. *P < 0.05 (versus CD3 stimulated samples).
Time course of cytokine induction in CD4+ T cells
Since CD28 and ICOS were the only stimulating coreceptors, they were chosen to study the time course of cytokine mRNA induction. As can be seen from figure 2A, α-CD3/B7-2 stimulation induced cytokine mRNA levels rapidly till two hours. Whereas IL-13 mRNA levels started to decline from that time point on, other cytokine mRNA levels still increased. Similar observations were made after α-CD3/B7-H2 stimulation (Figure 2B). Most remarkable, IL-13 and IFNγ were more strongly induced as compared to α-CD3/B7-2 stimulation, whereas IL-2 was much weaker induced. The induction of IL-10 was weak for both costimulatory receptors.
Figure 2 Time course of cytokine induction by costimulatory receptors. Human CD4+ T cells were stimulated by α-CD3/B7-2 (A) and α-CD3/B7-H2 (B) respectively and harvested after different time points. Cytokine mRNA level were determined using real-time RT-PCR. Levels were normalized to β-actin and unstimulated cells were set to 1. Each point and bar represents mean ± s.e.mean of duplicates measurements of one donor. Similar results were obtained from three independent donors.
Activation of MAPKs by costimulatory receptors
We speculated that differential activation of MAPK family members by CD28 and ICOS may contribute to their different cytokine production patterns. To this end, we investigated the activation of the MAPK family members as downstream signaling molecules. T cell receptor and B7-2 (Figure 3A) or B7-H2 (Figure 3B) cross-linking elicited the phosphorylation of ERK and p38 indicating the up-regulation of kinase activity. The activation reached a maximum 15 min after B7-2 costimulation and 2 min after B7-H2 costimulation for both kinases. It persisted over a period of two hours. Since we could not detect JNK by western blot, we analyzed JNK activation by a recently available ELISA method. A cell-based ELISA method was used to determine the JNK phosphorylation relative to the total JNK protein found in CD4+ T cells. Stimulation with α-CD3/α-CD28 and α-CD3/B7-H2 for different time points was assessed to induce the JNK pathway. No effect could be observed in cell costimulated with α-CD3/α-CD28. In CD4+ T cells costimulated with α-CD3/B7-H2 an increase in the ratio of phosphorylated JNK to total JNK was observed (Figure 3C).
Figure 3 Activation of MAPKs in human primary CD4+ T-cells by costimulatory receptors. Human CD4+ T cells were seeded at a concentration of 1 × 106/ml and stimulated with Dynal beads coated with α-CD3/B7-2 Fc (A) or α-CD3/B7-H2 Fc (B) for 0, 2, 5, 15, 60 or 120 min as indicated. Cells were harvested and cell extracts prepared. Cell extracts were analyzed by Western blot using either Anti-active pAb (upper row) or with corresponding antibodies that recognize both active and inactive forms of each subfamily of MAPK. Results are representative of three experiments with similar results. (C) Activated JNK was detected by ELISA. CD4+ T cells were cultured in precoated 96-well plates at a concentration of 1 × 106 /ml. Cells were stimulated with α-CD3/B7-2 Fc and α-CD3/B7-H2 Fc for different times (1, 5, 15, 60 and 120 min). Following stimulation the cells were fixed. The phosphorylated form of JNK as well as the amount of total JNK was detected using an ELISA method. Results shown are representative for three independent experiments.
Effect of MAPK inhibitors on costimulated CD4+ T cells
These results indicated a role of MAPKs in the signaling cascade precipitated from the costimulatory receptors CD28 and ICOS leading to cytokine expression. To elucidate which MAPK is important for cytokine induction, we first analyzed whether selective MAPK inhibitors were able to inhibit α-CD3/B7-2 induced gene expression. To this end, CD4+ T cells were preincubated with three selective MAPK inhibitors: SP600125 (selective inhibitor of c-Jun N-terminal kinase (JNK)), U0126 (selective inhibitor of MEK-1 and MEK2) and SB203580 (selective inhibitor of p38 MAPK). These compounds were chosen based on their pharmacological profile (high potency and specificity) for these kinases [31-33]. Since the cytokine induction behaved similar on the protein and mRNA level (see above), only mRNA was taken into consideration for determination of IC50 values for inhibition of cytokines. The JNK inhibitor did not affect induction of any cytokine analyzed. The p38 inhibitor inhibited IL-2, IL-13 and IFNγ with a similar IC50 of 3–7 μM, whereas IL-4 gene expression was not affected (Table 1). The MEK inhibitor U0126 inhibited IL-2 and IFNγ gene expression with IC50 of about 1 μM. IL-4 gene expression was not affected, whereas IL-13 was potently inhibited with an IC50 of 0.18 μM (Table 1). Since ICOS costimulation also induced the secretion of various cytokines and activated MAPKs, we further analyzed whether this induction is influenced by different pharmacological inhibitors of MAPKs in the same manner as for α-CD3/B7-2 activation. The selective p38 MAPK inhibitor (SB203580) affected cytokine synthesis only at very high concentrations (>30 μM). The JNK inhibitor SP600125 inhibited IL-2, IL-13 and IFNγ dose dependently at similar concentrations of about 1 μM whereas IL-4 was not inhibited by this compound (Table 1). The ERK inhibitor U0126 inhibited IL-2, IL-13 and IFNγ dose dependently at similar concentrations of 0.1 to 0.4 μM whereas IL-4 was inhibited only at ten-fold higher doses.
Table 1 IC50 values for the inhibition of cytokine mRNA synthesis by different MAPK inhibitors in CD28 or ICOS costimulated human CD4+ T cells. Human CD4+ T cells were preincubated with increasing concentrations of SP600125, U0126 and SB203580. The cells were then stimulated with beads coated with α-CD3/B7-2 fusion protein or α-CD3/B7-H2 fusion protein for 8 hours. Cytokine mRNA level were determined in triplicates using real time RT-PCR and were normalized to β-actin. Stimulated cells treated with DMSO were used as 100% control. IC50values were calculated for the different cytokines. Results represent mean ± s.e.m. for three different donors.
α-CD3/B7-2 stimulation
μM IL-2 IL-4 IL-13 IFNγ
U 0126 1.17 ± 0.97 no effect* 0.18 ± 0.14 0.72 ± 0.39
SB 203580 6.99 ± 4.57 no effect* 3.02 ± 1.14 4.15 ± 1.38
SP 600125 no effect* no effect* no effect* no effect*
α-CD3/B7-H2 stimulation
μM IL2 IL4 IL13 IFNγ
U 0126 0.11 ± 0.03 1.29 ± 0.96 0.31 ± 0.09 0.14 ± 0.08
SB 203580 no effect* no effect* no effect* 6.93 ± 2.47
SP 600125 1.25 ± 0.18 no effect* 1.33 ± 0.36 0.79 ± 0.35
*concentrations up to 30 μM were tested
Effects of MAPK inhibitors in vivo
The finding that the MAPKs inhibitors very effectively inhibited several cytokines in CD4+ T cells costimulated through CD28 and ICOS, prompted us to investigate whether these selective MAPKs inhibitors may be active in an animal model of allergic asthma. BALB/c mice were actively sensitized to ovalbumin and challenged by intranasal application of ovalbumin solution. Different doses of the three selective MAPKs inhibitors SP600125, U0126, SB203580 were given intraperitoneally 2 hours before challenge. After 4 days of challenge mice were killed and bronchoalveolar lavage (BAL) was performed. Antigen challenge caused a strong eosinophilic inflammation of the lung and this was reflected in the cellular composition of the lung lavage fluid. Compared with saline challenged animals, the allergen challenge in sensitized animals led to an increase in all leukocyte populations especially of eosinophils in the lavage fluid (Figure 4). On average, the total amount of leucocytes in the BAL of the positive control animals was 2.26 × 106 cells /ml (n = 7) and eosinophils accounted for 60.16 % of the total cell count, whereas in the saline challenged animals 0.16 × 106 cells /ml (n = 7) could be detected, which contained no granulocytes or lymphocytes. The treatment of the animals with SB203580 hardly affected leukocyte and eosinophils numbers in the lavage fluid, even at very high doses (20 mg/kg) (Figure 4A). In contrast, the administration of U0126 and SP600125 inhibited recruitment of the total cell number in the bronchoalveolar lavage dose dependently with an ED50 value of 2.5 mg/kg and 9 mg/kg respectively. Absolute eosinophil and lymphocyte numbers were also dose dependently inhibited by the ERK and JNK inhibitors, whereas neutrophil numbers were not affected. There were minor but not significant relative changes in the different cell populations. The cytokine protein levels in BAL fluid were below the detection limit of the ELISA kits used. Therefore cytokine mRNA levels were determined from BAL cells. SB203580 lacked also any effect on the cytokine mRNA induced in BAL cells (Figure 5A–C). In contrast, SP600125 and U0126 dose dependently inhibited IL-5 mRNA induced in BAL cells (Fig. 5D, G). Both compounds inhibited also IL-13 and IFNγmRNA at higher doses (Fig. 5 E, F, H, I).
Figure 4 Effects of MAPK inhibitors in vivo. Actively intraperitoneally sensitized mice were challenged by intranasally administration of ovalbumin. SB203580 (Figure A), SP600125 (Figure B) and U0126 (Figure C) were given intraperitoneally two hours prior to ovalbumin challenge. After 4 days of challenge bronchoalveolar lavage was performed. Total cell number in BAL and differential cell count on cytospin preparations was assessed. OVA challenged vehicle-treated animals were taken as a positive control. Each column represents mean ± s.e.mean. n = 5 mice per group. *p < 0.05 (drug treated versus OVA challenged vehicle (positive) control), **p < 0.005 (drug treated versus OVA challenged vehicle (positive) control). ++ p < 0.005 (OVA challenged (positive) versus saline challenged (negative) control).
Figure 5 Effects of MAPK inhibitors on BAL cytokine levels. Cells were recovered from bronchoalveolar lavage by centrifugation and RNA was prepared from the cell pellet. Cytokine mRNA level were determined using real-time RT-PCR and were normalized to β-actin. Each column represents mean ± s.e.mean. n = 5 mice per group. n.d. – not detectable. *p < 0.05 (versus OVA challenged positive control), +p < 0.05 (negative versus positive control)
Effects of MAPK inhibitors on serum OVA-specific IgE levels
We next determined whether MAPK inhibitors could modify an ongoing OVA-specificTh2 response in vivo by analyzing OVA-specific IgE levels. Serum was collected24 h after the last OVA challenge. Levels of OVA-specfic IgE, was determined using ELISA. Substantial elevation OVA-specific IgE was observed in serum from OVA-sensitized and -challengedmice as compared with untreated mice (Fig. 6). U0126 and SP600125 significantly lowered total OVA-specific IgE levels, whereas SB203580 did not affect OVA-specific IgE levels.
Figure 6 Effects of MAPK inhibitors on OVA-specific IgE antibody levels in the serum. Mouse serum was collected 24 h after the last OVA challenge. Samples were obtained from saline- (negative) or OVA-challenged mice in the absence (positive) or presence of MAPK inhibitors: SB203580 (20 mg/kg), U0126 (20 mg/kg), SP600125 (20 mg/kg) or of the combination of U0126 (5 mg/kg) plus SB203580 (5 mg/kg). Serum titers for OVA-specific IgE antibodies were measured by ELISA as described in methods. Each column represents mean ± s.e.mean. n = 5 mice per group. *p < 0.05 (versus OVA challenged positive control), +p < 0.05 (negative versus positive control)
Effects of MAPK inhibitors on lung inflammation in vivo
On separate experiments, lung tissue was collected 24 hours after the last challenge. Ovalbumin challenge induced marked infiltration of inflammatory cells into the peribronchial and perivasular tissue as compared with saline challenge (Figure 7 A, B). The majority of the infiltrated inflammatory cells were eosinophils as detected in the May-Günwald Giemsa stain (data not shown). MAPK inhibitors had different effects on peribronchial inflammation. The inflammatory score (calculated as described in Methods) decreased significantly after administration of 20 mg/kg ERK or JNK inhibitor, whereas inhibition of p38 had no influence on peribronchial inflammation (Figure 7G). The allergen challenge induced mucus secretion was detected with periodic acid-Schiff stain counterstained with hematoxylin (Figure 7A, B). The percentage of PAS-positive mucus-containing epithelial cells decreased as an effect of ERK or JNK inhibition (Figure 7D, E, G). The p38 inhibitor had no effect on mucus production (Figure 7C, G).
Figure 7 Histology of lung inflammation. Representative sections of the lungs are shown. Sections are periodic acid-Schiff stained and counterstained with haematoxylin for analysis of mucin-containing cells. Tissue was examined by light microscopy (original magnification 20×), zoomed inset delineates PAS+ epithelial cells (100×). Lung sections were obtained from saline- (A) or OVA- (B, C, D, E, F) challenged mice in the absence (A, B) or presence of MAPK inhibitors: SB 203580 (20 mg/kg) (C), U 0126 (20 mg/kg) (D), SP 600125 (20 mg/kg) (E) or of the combination of U 0126 (5 mg/kg) plus SB 203580 (5 mg/kg) (F). Peribronchial inflammation and the abundance of PAS-positive mucus-containing cells were determined by applying the scores as described in methods (G). The scores described in methods were applied to ten bronchi for each animal and mean and standard deviation were determined for all the animals in a group (n = 5). *P < 0.05 and **P < 0.01 compared with positive control group.
Effects of combined administration of MAPK inhibitors
Since the three MAPKs are part of different signaling cascades with different upstream and downstream mediators, we asked whether the combination of two MAPK inhibitors may have synergistic effects on the induction of cytokines in CD4+T cells. The simultaneous addition of SB203580 (5 μM) and SP600125 (3 μM) did not cause any synergistic effect. Unexpectedly, the simultaneous administration of increasing concentrations of U0126 (0.01 μM to 30 μM) and SB 203580 to α-CD3/α-B7-2 (Figure 8A, B) and α-CD3/B7-H2 (Figure 8C and 8D) stimulated CD4+ T cells led to a shift to the right of the dose-response curve for U0126 compared to the administration of U 0126 alone. The dose for SB203580 was 10 μM for ICOS costimulation and 1 μM for CD28 costimulation, respectively. This was reflected by an increase in the corresponding IC50 values (Table 2). In contrast, the addition of SP600125 (0.3 μM for ICOS costimulation and 1.5 μM for CD28 costimulation) did not change the dose response curve for U0126 (Figure 8). Since we observed for the combination SB203560 and U0126 only unforeseen effects in vitro, we analyzed this combination in vivo. The simultaneous treatment of the animals with 5 mg/kg SB203580 and 5 mg/kg U0126 (Figure 9) diminished the inhibition of late phase eosinophilia by the solely administration of the ERK inhibitor. There were minor and not significant change in the relative numbers of the populations counted in the BAL. This effect could also be observed measuring cytokine mRNA levels in the BAL cells. Treatment of the animals with SB203580 antagonized the inhibiting effect of U0126 on IL-5, IL-13 and IFNγ mRNA level (Figure 10). The effect of the combined administration of ERK inhibitor and p38 inhibitor on lung histology had the same effect as observed in vitro, as well as on the cellular composition of the BAL. The p38 inhibitor abolished the effect of ERK inhibition when administered simultaneously compared to administration of ERK inhibitor alone (Figure 7F, G). Furthermore, OVA-specific IgE Level in the serum of mice treated with the combination of the ERK and p38 inhibitor were higher than in mice treated with the ERK inhibitor alone (Figure 6).
Figure 8 Effects of combined administration of MAPK inhibitors in vitro. Human CD4+ T cells were preincubated either with increasing doses of U0126 alone (solid line) or in combination with SP 600125 (dashed line) of with SB 203580 (dotted line). Doses of the MAPK inhibitors are specitied in the text. After that, cells were stimulated with α-CD3/B7-2 Fc (A, B) or α-CD3/B7-H2 Fc (C, D) for 8 hours. IFNγ (A, C) and IL-13 (B, D) cytokine mRNA was measured by real-time RT-PCR. Dose-response curves were determined and IC50 values were calculated. Data are representative of three different donors.
Table 2 IC50 values for the inhibition of cytokine mRNA synthesis by combined administration of MAPK inhibitors in vitro. IC50values were calculated from data shown in Figure 8.
μM IFNγ (U0126) IFNγ (U0126+SB203580) IL-13 (U0126) IL-13 (U0126+SB203580)
α-CD3/B7-2 0.29 0.68 0.28 0.97
α-CD3/B7-H2 0.19 1.14 0.21 0.73
Figure 9 Effects of combined administration of MAPK inhibitors in vivo. Late phase eosinophilia was induced in mice pretreated either with 5 mg/kg U0126 and 5 mg/kg SB 203580 or with the combination thereof. BAL was performed and total cell number and differential leukocyte count was determined. OVA challenged vehicle-treated animals were taken as positive control. Each column represents mean ± s.e.mean. n = 5 mice per group. *p < 0.05 (versus OVA challenged vehicle control), +p < 0.05 (versus U0126 alone).
Figure 10 Effects of combined MAPK inhibitors on BAL cytokine levels. Cells were recovered from bronchoalveolar lavage by centrifugation and RNA was prepared from the cell pellet. Cytokine mRNA level were determined using real-time RT-PCR and were normalized to β-actin. Each column represents mean ± s.e.mean. n = 5 mice per group. *p < 0.05 (versus OVA challenged vehicle control), +p < 0.05 (versus U0126 alone).
Discussion
Costimulation is essential for the full activation of CD4+ T cells. In this study, we analyzed the induction and regulation of Th1 and Th2 cytokines in CD4+ T cells upon engagement of different costimulatory receptors. Beads with defined combinations of surface receptor stimulating antibodies and costimulatory receptor ligands were applied to induce cytokine synthesis. Ligation of two costimulatory receptors, CD28 and ICOS, augmented TCR activation and this in part is due to MAPK activation. The effects of pharmacological inhibitors of three different MAPKs on cytokines induced by α-CD3/B7-2 or α-CD3/B7-2 were investigated. Using these compounds, we were able to elucidate different MAPK signaling pathways leading to cytokines synthesis in dependence on the costimulatory receptor engaged. We tested these compounds in an animal model of asthma in which the ERK inhibitor U0126 and the JNK inhibitor SP600125 were able to reduce the influx of eosinophils into the lungs of sensitized and challenged mice. We found no synergistic effect of any combination of these MAPK inhibitors. Unexpectedly, SB203580 antagonized the in vitro and in vivo action of U0126.
For optimal activation, CD4+ T cells require specific antigen recognition by the TCR and additional signals (collectively called costimulatory signals), delivered by the same antigen-presenting cell [34]. In the absence of costimulation, lymphocytes fail to respond effectively and are rendered anergic [35]. CD28 is the best-characterized costimulatory receptor and constitutively present on the surface of T cells (reviewed in [12]). CD28 is activated by B7-1 and B7-2 counter-receptors on antigen-presenting cells. Its signaling contributes to the overall strength of T cell activation [36]. However, new B7- and CD28-like molecules have recently been discovered and new pathways have been delineated that seem to be important for regulating the responses of previously activated T cells [37]. Some B7 homologues have unknown receptors, indicating that other immunoregulatory pathways remain to be described. We established an in vitro test system to analyze to role of each coreceptor. To this end, we conjugated beads with defined combinations of α-CD3 and B7 counterreceptors. In line with previous reports described above we found that CD28 ligation augments the synthesis of all T cell derived cytokines. Similar results were obtained by ligating ICOS via its counter receptor ICOS-L (B7-H2). The effects of ICOS signals on T cell proliferation and IL-2 production were reported to be modest in comparison with those of CD28 [16]. However, ICOS costimulation is equivalent to that mediated by CD28 costimulation for the production of effector cytokines like IFNγ, IL-4, IL-10 and IL-13. IL-5 was not detected. It has been shown, that efficient in vitro IL-5 production needs elevated cAMP level in addition to TCR stimulation [38]. There was no difference in the strength of induction of Th2 compared with Th1 cytokines for CD28 and ICOS costimulation. Conflicting data has been reported about the function of PD-1. The engagement of PD-1 by its specific ligands, B7-H1 (PD-L1) or B7-DC (PD-L2) was reported to inhibit T and B cell proliferation and cytokine production [39,40]. In our test system engagement of PD-1 by ligation with B7-H1 neither augmented nor inhibited cytokine synthesis.
Activation of MAPK family members by CD28 and ICOS may contribute to their costimulatory activity. To analyze whether different signaling cascades are precipitated by the different coreceptors, their ability to activate different MAPK pathways was determined. Since only CD28 and ICOS activation augmented cytokine synthesis we focused our analyses on these receptors. We analyzed the signaling pathways of ERK1/2, JNK, and p38 MAP kinase, during the primary activation of human T-cells, costimulated with ICOS or CD28. ERK1/2 and p38 kinase were markedly activated by both CD28- and ICOS-mediated costimulation. In our cellular assay system the activation of ERK has a role in T-cell activation via CD28 and ICOS. The activation of ERK by CD28 is mediated by binding of SOS to the YMNM motif of the intracellular domain of CD28 [41]. The YMNM motif is not conserved in ICOS, the sequence being YMFM, and amino-acid substitutions in this motif results in a failure to associate with Grb2 [16,42]. Therefore it is currently unclear, how ICOS is able to activate ERK.
Previous studies of p38 MAP kinase in human purified T cells [43] and in the CD4+ subset [7] clearly demonstrated the involvement of p38 MAP kinase in the cell activation through TCR and CD28 costimulation signal pathways. However, little is known about this MAP kinase in ICOS costimulated T cells. In our cellular assay activation of p38 MAPK by ICOS was detected. This is in line with recent reports, in which ICOS ligation synergized with TCR signals for activation of the ERK and p38 MAP kinases [44,45].
While enhanced activityof JNK is an absolute necessity for regulation of IL-2 geneexpression in T cell lines [46], a defect in JNK signalingwas claimed to be involved in T cell differentiation, but notin T cell activation in vivo, i.e. in JNK-deficient animals [47]. This corresponds to our results with human T cells, where JNK was found not to be activated after CD28 costimulation. Conflicting reports appeared about the ability of ICOS to activate the JNK pathway. Parry et al. reported that only CD28 but not ICOS costimulation activated c-jun N-terminal kinase [48]. Arimura et al. reported that the cross-linking of ICOS induced much less phosphorylation of JNK than did the cross-linking of CD28 [49]. In our cellular assay system we found that ICOS activated the JNK pathway. However, this was only detected by using a sensitive ELISA based assay indicating a very low expression of JNK in primary human T cells. Alltogether, whereas p38 and ERK are consistently found to be activated after CD28 and ICOS costimulation, the activation of JNK appeared only after ICOS costimulation.
In order to elucidate which of the activated MAP kinases are important for cytokine induction after T cell activation and costimulation we employed specific inhibitors of the different MAPKs in our cellular assay system. These compounds were chosen based on their pharmacological profile (high potency and specificity) for these three kinases [31-33]. The p38 inhibitor SB203580 and the ERK inhibitor U0126 inhibited IL-2, IL-13 and IFNγ in CD28 costimulated T-cells. IL-4 induction, in contrast, was not affected by these inhibitors. In line with our observations, that JNK does not become activated by CD28 costimulation, the JNK inhibitor did not affect cytokine induction in CD28 costimulated T cells. Only one report analyzed the effect of a p38 inhibitor in primary human CD4+ T cells. There SB203580 blocked human CD4 T cell production of IL-4, IL-5, TNF-, and IFN-, but not IL-2, in response to CD3 and CD28 stimulation [7]. This difference is most likely due to different experimental protocols and the use of a different T cell population. While U0126 was never tested to inhibit human T cell cytokine production, in the first description of the JNK inhibitor SP600125, its effect on polarized Th1 and Th2 cells was described. There Th1 cytokines and some Th2 cytokines but not IL-4 was inhibited by this compound [32]. It is tempting to speculate, that the polarization of the T cells rendered them susceptible to JNK inhibition. Different observations were made in ICOS costimulated T cells. For human T cells, it has only been reported, that U0126 blocked the proliferation of T-cells and the transcription of IL-2 following costimulation by ICOS in human T cell [45]. We found, that the p38 inhibitor was not effective, but the JNK inhibitor inhibited IL-2, IL-13 and IFNγ. The ERK inhibitor U0126 inhibited all cytokines with similar efficiency as in CD28 costimulated T cells. Together these data indicates that in human T cells ICOS precipitates MAPK signaling cascades differentially compared to murine T cells. More interestingly, our inhibitor studies revealed that not every MAPK signaling cascade activated by costimulation is necessary for cytokine synthesis.
The sensitivity of CD4+ T cells to various MAPK inhibitors prompted us to analyze these compounds in an animal model of asthma. The p38 inhibitor SB203580 did not affect airway inflammation in our model. Conflicting data about the action of p38 inhibitors in asthma models have been reported in the literature. Similar to our results, one study found no effect of SB203580 in rats [50]. In contrast, Underwood et al. demonstrated anti-allergic activity using a novel second-generation p38 MAPK inhibitor [51]. Either increased specificity of this new inhibitor or better bioavailability of this compound may explain its effect in vivo. In contrast to the p38 inhibitor, the ERK as well as the JNK inhibitor dose dependently decreased allergen induced airway eosinophilia in our model. This anti-inflammatory effect of the ERK and the JNK inhibitor was also detected by measuring peribronchial inflammation and mucus production. Elevated serum IgE is a hallmark of a Th2 immune response. Our data showed that serum levels of OVA-specific IgE were also substantially reduced by U0126 and SP600125, whereas no significant inhibition was observed by SB203580. This extends our previous findings, where 10 mg/kg of U0126 were found to be effective in a rat asthma model [28]. During the preparation of the manuscript it was reported that 30 mg/kg SP600125 inhibited airway eosinophilia in rats [52]. This supports our finding, that targeting JNK reduces allergic airway inflammation.
The inhibitory effect could also be observed by the reduction of cytokine mRNA level. However, the inhibition was not selective for Th2 cytokines. The induction of IFNγ in BAL cells was also inhibited by U0126 and SP600125. This is in line with our findings described above, that costimulation of T cells induces Th1 as well as Th2 cytokines and that MAPK inhibitors are also effective in inhibiting IFNγ. This may be beneficial for the anti-inflammatory effect in the asthma model. It has been shown, that in IFNγ knockout mice less pronounced allergic symptoms such as antigen-specificIgE, eosinophilic infiltration, and airway hyperresponsiveness occur [53]. Therefore, it seems likely that endogenous IFNγ is necessaryfor optimal IgE production also during a secondary response.
The in vivo profile of MAPK inhibitors reflects the in vitro profile of ICOS costimulated T cells: ERK and JNK inhibitors are effective whereas the p38 inhibitor is not. In contrast, the inhibitory profile of CD28 costimulated T cells is different. There p38 is effective, while the JNK inhibitor is not. In line with this, it has been shown, that ICOS plays an important role in the Th2 effector responses in the lungs, in that inhibition of ICOS suppresses lung inflammation and Th2 cytokine production [54]. Similarly, ICOS-L-Fc augments lung inflammation and the production of cytokines [55]. Together these data indicate the importance of ICOS costimulation in the allergic lung response. Therefore, targeting signaling pathways precipitated by ICOS should dampen the allergic lung inflammation.
Since the three MAPKs are part of different signaling cascades with different upstream and downstream mediators, we asked whether the combination of two MAPK inhibitors may have synergistic effects on the induction of cytokines in CD4+T cells. Our in vitro studies did not reveal any such synergistic effect regardless of the combination used. However, we could observe that the simultaneous administration of increasing concentrations of U0126 and 1 μM SB203580 to α-CD3/α-CD28 and α-CD3/α-ICOS stimulated CD4+ T cells led to a right-shift of the dose-response curve for U0126 compared to the administration of U0126 alone. We analyzed this combination also in vivo. Whereas treatment of the animals with U0126 alone inhibited late phase eosinophilia markedly as in previous experiments the concomitant treatment with SB203580 diminished the effect of U0126 almost completely. To our knowledge no in vitro or in vivo studies reported the simultaneous addition of two MAPK inhibitors in T cell activation studies or asthma models. Currently, we investigate the mechanism for this antagonistic action.
Conclusion
Different MAPK signaling pathways are activated in T cells dependent on the costimulatory receptor engaged. The MAP kinase ERK is important for both stimuli. p38 MAPK activation is important for CD28 induced cytokine synthesis, whereas JNK is important for ICOS induced cytokine synthesis. The different regulation can be exploited to find new specifically targeted drugs aimed for diseases where different costimulatory molecules play an important pathophysiological role. We could demonstrate for the first time that inhibition of the JNK cascade is a therapeutic option for asthma. The specific JNK inhibitor SP 600125 reduced the influx of eosinophils in an animal model of asthma. The development of more specific MEK-ERK and JNK-targeted drugs would support this approach to treat T cell dominated diseases such as asthma.
List of abbreviations
BAL – bronchoalveolar lavage fluid
ERK – extracellular signal-regulated kinase
ICOS – inducible costimulatory
IL – interleukin
JNK – jun NH2-terminal kinase
MAPK – mitogen-activated protein kinase
OVA – ovalbumine
TCR – T cell receptor
Authors' contributions
LC carried out ICOS costimulation studies in vitro and all in vivo studies.
MZ carried out CD28 costimulation studies in vitro.
KB participated in study design and coordination.
AP conceived the study, and participated in its design and coordination.
All authors read and approved the final manuscript.
Acknowledgements
The authors thank Evi Hoier and Brigitte Neupert for her expert technical assistance and Alexandra Fischer for animal care. This work was supported by the Deutsche Forschungsgemeinschaft (Grant Pa 463/1-1, SFB 643) and a DAAD scholarship to Ligia Chialda.
==== Refs
Busse WW Lemanske RFJ Asthma N Engl J Med 2001 344 350 362 11172168 10.1056/NEJM200102013440507
Chung KF Barnes PJ Cytokines in asthma Thorax 1999 54 825 857 10456976
Yssel H Groux H Characterization of T cell subpopulations involved in the pathogenesis of asthma and allergic diseases Int Arch Allergy Immunol 2000 121 10 18 10686504 10.1159/000024292
Ray A Cohn L Altering the Th1/Th2 balance as a therapeutic strategy in asthmatic diseases Curr Opin Investig Drugs 2000 1 442 448 11249697
Pahl A Szelenyi I Asthma therapy in the new millennium Inflamm Res 2002 51 273 282 12088267
Pahl A Zhang M Torok K Kuss H Friedrich U Magyar Z Szekely J Horvath K Brune K Szelenyi I Anti-inflammatory effects of a cyclosporine receptor-binding compound, D-43787 J Pharmacol Exp Ther 2002 301 738 746 11961080 10.1124/jpet.301.2.738
Schafer PH Wadsworth SA Wang L Siekierka JJ p38 alpha mitogen-activated protein kinase is activated by CD28-mediated signaling and is required for IL-4 production by human CD4+CD45RO+ T cells and Th2 effector cells J Immunol 1999 162 7110 7119 10358155
Cefai D Cai YC Hu H Rudd C CD28 co-stimulatory regimes differ in their dependence on phosphatidylinositol 3-kinase: common co-signals induced by CD80 and CD86 Int Immunol 1996 8 1609 1616 8921441
Rudd CE Upstream-downstream: CD28 cosignaling pathways and T cell function Immunity 1996 4 527 534 8673699 10.1016/S1074-7613(00)80479-3
Weiss A Littman DR Signal transduction by lymphocyte antigen receptors Cell 1994 76 263 274 8293463 10.1016/0092-8674(94)90334-4
Bertram EM Tafuri A Shahinian A Chan VS Hunziker L Recher M Ohashi PS Mak TW Watts TH Role of ICOS versus CD28 in antiviral immunity Eur J Immunol 2002 32 3376 3385 12432568
Carreno BM Collins M The B7 family of ligands and its receptors: new pathways for costimulation and inhibition of immune responses Annu Rev Immunol 2002 20 29 53 11861596 10.1146/annurev.immunol.20.091101.091806
Rudd CE Schneider H Unifying concepts in CD28, ICOS and CTLA4 co-receptor signalling Nat Rev Immunol 2003 3 544 556 12876557 10.1038/nri1131
Tafuri A Shahinian A Bladt F Yoshinaga SK Jordana M Wakeham A Boucher LM Bouchard D Chan VS Duncan G Odermatt B Ho A Itie A Horan T Whoriskey JS Pawson T Penninger JM Ohashi PS Mak TW ICOS is essential for effective T-helper-cell responses Nature 2001 409 105 109 11343123 10.1038/35051113
Coyle AJ Lehar S Lloyd C Tian J Delaney T Manning S Nguyen T Burwell T Schneider H Gonzalo JA Gosselin M Owen LR Rudd CE Gutierrez-Ramos JC The CD28-related molecule ICOS is required for effective T cell-dependent immune responses Immunity 2000 13 95 105 10933398 10.1016/S1074-7613(00)00011-X
Hutloff A Dittrich AM Beier KC Eljaschewitsch B Kraft R Anagnostopoulos I Kroczek RA ICOS is an inducible T-cell co-stimulator structurally and functionally related to CD28 Nature 1999 397 263 266 9930702 10.1038/16717
Yoshinaga SK Zhang M Pistillo J Horan T Khare SD Miner K Sonnenberg M Boone T Brankow D Dai T Delaney J Han H Hui A Kohno T Manoukian R Whoriskey JS Coccia MA Characterization of a new human B7-related protein: B7RP-1 is the ligand to the co-stimulatory protein ICOS Int Immunol 2000 12 1439 1447 11007762 10.1093/intimm/12.10.1439
Davis RJ The mitogen-activated protein kinase signal transduction pathway J Biol Chem 1993 268 14553 14556 8325833
Schaeffer HJ Weber MJ Mitogen-activated protein kinases: specific messages from ubiquitous messengers Mol Cell Biol 1999 19 2435 2444 10082509
Davis RJ Signal transduction by the JNK group of MAP kinases Cell 2000 103 239 252 11057897 10.1016/S0092-8674(00)00116-1
Han J Ulevitch RJ Emerging targets for anti-inflammatory therapy Nat Cell Biol 1999 1 E39 40 10559893 10.1038/10032
Su B Jacinto E Hibi M Kallunki T Karin M Ben-Neriah Y JNK is involved in signal integration during costimulation of T lymphocytes Cell 1994 77 727 736 8205621 10.1016/0092-8674(94)90056-6
Li W Whaley CD Mondino A Mueller DL Blocked signal transduction to the ERK and JNK protein kinases in anergic CD4+ T cells Science 1996 271 1272 1276 8638107
Nunes JA Battifora M Woodgett JR Truneh A Olive D Cantrell DA CD28 signal transduction pathways. A comparison of B7-1 and B7-2 regulation of the map kinases: ERK2 and Jun kinases Mol Immunol 1996 33 63 70 8604225 10.1016/0161-5890(95)00121-2
Matsuda S Moriguchi T Koyasu S Nishida E T lymphocyte activation signals for interleukin-2 production involve activation of MKK6-p38 and MKK7-SAPK/JNK signaling pathways sensitive to cyclosporin A J Biol Chem 1998 273 12378 12382 9575191 10.1074/jbc.273.20.12378
Salmon RA Foltz IN Young PR Schrader JW The p38 mitogen-activated protein kinase is activated by ligation of the T or B lymphocyte antigen receptors, Fas or CD40, but suppression of kinase activity does not inhibit apoptosis induced by antigen receptors J Immunol 1997 159 5309 5317 9548470
McAdam AJ Chang TT Lumelsky AE Greenfield EA Boussiotis VA Duke-Cohan JS Chernova T Malenkovich N Jabs C Kuchroo VK Ling V Collins M Sharpe AH Freeman GJ Mouse inducible costimulatory molecule (ICOS) expression is enhanced by CD28 costimulation and regulates differentiation of CD4+ T cells J Immunol 2000 165 5035 5040 11046032
Pahl A Zhang M Kuss H Szelenyi I Brune K Regulation of IL-13 synthesis in human lymphocytes: implications for asthma therapy Br J Pharmacol 2002 135 1915 1926 11959794 10.1038/sj.bjp.0704656
Finotto S De Sanctis GT Lehr HA Herz U Buerke M Schipp M Bartsch B Atreya R Schmitt E Galle PR Renz H Neurath MF Treatment of allergic airway inflammation and hyperresponsiveness by antisense-induced local blockade of GATA-3 expression J Exp Med 2001 193 1247 1260 11390432 10.1084/jem.193.11.1247
Myou S Leff AR Myo S Boetticher E Tong J Meliton AY Liu J Munoz NM Zhu X Blockade of inflammation and airway hyperresponsiveness in immune-sensitized mice by dominant-negative phosphoinositide 3-kinase-TAT J Exp Med 2003 198 1573 1582 14623911 10.1084/jem.20030298
Davies SP Reddy H Caivano M Cohen P Specificity and mechanism of action of some commonly used protein kinase inhibitors Biochem J 2000 351 95 105 10998351 10.1042/0264-6021:3510095
Bennett BL Sasaki DT Murray BW O'Leary EC Sakata ST Xu W Leisten JC Motiwala A Pierce S Satoh Y Bhagwat SS Manning AM Anderson DW SP600125, an anthrapyrazolone inhibitor of Jun N-terminal kinase Proc Natl Acad Sci U S A 2001 98 13681 13686 11717429 10.1073/pnas.251194298
Murray BW Bennett BL Sasaki DT Analysis of pharmacologic inhibitors of Jun N-terminal kinases Methods Enzymol 2001 332 432 452 11305116
Baxter AG Hodgkin PD Activation rules: the two-signal theories of immune activation Nat Rev Immunol 2002 2 439 446 12093010
Schwartz RH T cell anergy Annu Rev Immunol 2003 21 305 334 12471050 10.1146/annurev.immunol.21.120601.141110
Manickasingham SP Anderton SM Burkhart C Wraith DC Qualitative and quantitative effects of CD28/B7-mediated costimulation on naive T cells in vitro J Immunol 1998 161 3827 3835 9780147
Sharpe AH Freeman GJ The B7-CD28 superfamily Nat Rev Immunol 2002 2 116 126 11910893 10.1038/nri727
Klein-Hessling S Jha MK Santner-Nanan B Berberich-Siebelt F Baumruker T Schimpl A Serfling E Protein kinase A regulates GATA-3-dependent activation of IL-5 gene expression in Th2 cells J Immunol 2003 170 2956 2961 12626547
Freeman GJ Long AJ Iwai Y Bourque K Chernova T Nishimura H Fitz LJ Malenkovich N Okazaki T Byrne MC Horton HF Fouser L Carter L Ling V Bowman MR Carreno BM Collins M Wood CR Honjo T Engagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to negative regulation of lymphocyte activation J Exp Med 2000 192 1027 1034 11015443 10.1084/jem.192.7.1027
Latchman Y Wood CR Chernova T Chaudhary D Borde M Chernova I Iwai Y Long AJ Brown JA Nunes R Greenfield EA Bourque K Boussiotis VA Carter LL Carreno BM Malenkovich N Nishimura H Okazaki T Honjo T Sharpe AH Freeman GJ PD-L2 is a second ligand for PD-1 and inhibits T cell activation Nat Immunol 2001 2 261 268 11224527 10.1038/85330
Schneider H Cai YC Prasad KV Shoelson SE Rudd CE T cell antigen CD28 binds to the GRB-2/SOS complex, regulators of p21ras Eur J Immunol 1995 25 1044 1050 7737275
Tezuka K Tsuji T Hirano D Tamatani T Sakamaki K Kobayashi Y Kamada M Identification and characterization of rat AILIM/ICOS, a novel T-cell costimulatory molecule, related to the CD28/CTLA4 family Biochem Biophys Res Commun 2000 276 335 345 11006126 10.1006/bbrc.2000.3466
Koprak S Staruch MJ Dumont FJ A specific inhibitor of the p38 mitogen activated protein kinase affects differentially the production of various cytokines by activated human T cells: dependence on CD28 signaling and preferential inhibition of IL-10 production Cell Immunol 1999 192 87 95 10087176 10.1006/cimm.1998.1448
Feito MJ Vaschetto R Criado G Sanchez A Chiocchetti A Jimenez-Perianez A Dianzani U Portoles P Rojo JM Mechanisms of H4/ICOS costimulation: effects on proximal TCR signals and MAP kinase pathways Eur J Immunol 2003 33 204 214 12594849 10.1002/immu.200390023
Okamoto N Tezuka K Kato M Abe R Tsuji T PI3-kinase and MAP-kinase signaling cascades in AILIM/ICOS- and CD28-costimulated T-cells have distinct functions between cell proliferation and IL-10 production Biochem Biophys Res Commun 2003 310 691 702 14550257 10.1016/j.bbrc.2003.09.065
Avraham A Jung S Samuels Y Seger R Ben-Neriah Y Co-stimulation-dependent activation of a JNK-kinase in T lymphocytes Eur J Immunol 1998 28 2320 2330 9710210 10.1002/(SICI)1521-4141(199808)28:08<2320::AID-IMMU2320>3.0.CO;2-K
Dong C Yang DD Tournier C Whitmarsh AJ Xu J Davis RJ Flavell RA JNK is required for effector T-cell function but not for T-cell activation Nature 2000 405 91 94 10811224 10.1038/35011091
Parry RV Rumbley CA Vandenberghe LH June CH Riley JL CD28 and inducible costimulatory protein Src homology 2 binding domains show distinct regulation of phosphatidylinositol 3-kinase, Bcl-xL, and IL-2 expression in primary human CD4 T lymphocytes J Immunol 2003 171 166 174 12816995
Arimura Y Kato H Dianzani U Okamoto T Kamekura S Buonfiglio D Miyoshi-Akiyama T Uchiyama T Yagi J A co-stimulatory molecule on activated T cells, H4/ICOS, delivers specific signals in T(h) cells and regulates their responses Int Immunol 2002 14 555 566 12039907 10.1093/intimm/dxf022
Escott KJ Belvisi MG Birrell MA Webber SE Foster ML Sargent CA Effect of the p38 kinase inhibitor, SB 203580, on allergic airway inflammation in the rat Br J Pharmacol 2000 131 173 176 10991908 10.1038/sj.bjp.0703605
Underwood DC Osborn RR Kotzer CJ Adams JL Lee JC Webb EF Carpenter DC Bochnowicz S Thomas HC Hay DW Griswold DE SB 239063, a potent p38 MAP kinase inhibitor, reduces inflammatory cytokine production, airways eosinophil infiltration, and persistence J Pharmacol Exp Ther 2000 293 281 288 10734180
Eynott PR Nath P Leung SY Adcock IM Bennett BL Chung KF Allergen-induced inflammation and airway epithelial and smooth muscle cell proliferation: role of Jun N-terminal kinase Br J Pharmacol 2003 140 1373 1380 14623764 10.1038/sj.bjp.0705569
Hofstra CL Van Ark I Hofman G Nijkamp FP Jardieu PM Van Oosterhout AJ Differential effects of endogenous and exogenous interferon-gamma on immunoglobulin E, cellular infiltration, and airway responsiveness in a murine model of allergic asthma Am J Respir Cell Mol Biol 1998 19 826 835 9806748
Gonzalo JA Tian J Delaney T Corcoran J Rottman JB Lora J Al-garawi A Kroczek R Gutierrez-Ramos JC Coyle AJ ICOS is critical for T helper cell-mediated lung mucosal inflammatory responses Nat Immunol 2001 2 597 604 11429543 10.1038/89739
Guo J Stolina M Bready JV Yin S Horan T Yoshinaga SK Senaldi G Stimulatory effects of B7-related protein-1 on cellular and humoral immune responses in mice J Immunol 2001 166 5578 5584 11313397
| 15833106 | PMC1131927 | CC BY | 2021-01-04 16:36:26 | no | Respir Res. 2005 Apr 15; 6(1):36 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-36 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-401585750610.1186/1465-9921-6-40ResearchDevelopment of a lung slice preparation for recording ion channel activity in alveolar epithelial type I cells Bourke Steven [email protected] Helen S [email protected] Zea [email protected] Kwang-Jin [email protected] Edward D [email protected] Paul J [email protected] Cardiff School of Biosciences, Museum Avenue, Cardiff CF10 3US, Wales, UK2 Will Rogers Institute Pulmonary Research Center, Division of Pulmonary and Critical Care Medicine, Keck School of Medicine, Los Angeles, CA 90033, USA2005 27 4 2005 6 1 40 40 21 12 2004 27 4 2005 Copyright © 2005 Bourke et al; licensee BioMed Central Ltd.2005Bourke 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
Lung fluid balance in the healthy lung is dependent upon finely regulated vectorial transport of ions across the alveolar epithelium. Classically, the cellular locus of the major ion transport processes has been widely accepted to be the alveolar type II cell. Although evidence is now emerging to suggest that the alveolar type I cell might significantly contribute to the overall ion and fluid homeostasis of the lung, direct assessment of functional ion channels in type I cells has remained elusive.
Methods
Here we describe a development of a lung slice preparation that has allowed positive identification of alveolar type I cells within an intact and viable alveolar epithelium using living cell immunohistochemistry.
Results
This technique has allowed, for the first time, single ion channels of identified alveolar type I cells to be recorded using the cell-attached configuration of the patch-clamp technique.
Conclusion
This exciting new development should facilitate the ascription of function to alveolar type I cells and allow us to integrate this cell type into the general model of alveolar ion and fluid balance in health and disease.
==== Body
Background
Of fundamental importance to the optimisation of gas exchange in health and disease is the role of the alveolar epithelium in regulation of the liquid sub-phase in the postnatal lung. There is now unequivocal evidence to support both constitutive and stimulated Na+-driven, active vectorial transport of water from lung lumen to interstitium in the neonatal and adult lung (see [1-5] for reviews). During the final days of fetal development, steroid hormone-induced transcriptional upregulation of the amiloride-sensitive epithelial Na+ channel (ENaC) ensures that, during labour, the huge surge in fetal adrenaline leads to channel opening and massive Na+ flux out of the fetal lumen [6-9]; this drives osmotically-linked fluid reabsorption in preparation for the neonatal lung to take on the gas exchange role previously undertaken by the placenta. Maturation of the response appears to be under control of a number of environmental factors, most importantly oxygen [10-12]. Key to understanding this mechanism was the observation that transgenic mice which lack the α subunit of the ENaC channel complex die shortly after birth of alveolar flooding [13]. However, although there is a general consensus concerning the cellular and molecular basis of the fluid transport which clears the neonatal lung at birth, controversy still surrounds the issue of postnatal absorptive mechanisms where the role of ENaC in lung fluid homeostasis is less clear. Indeed, evidence is now emerging which supports a significant role for amiloride-insensitive pathways in adult lung fluid reabsorption. Of note are whole tissue studies in several different species [14-16]. Norlin et al. [15] demonstrated that up to 70% of basal guinea pig lung fluid reabsorption is amiloride-insensitive whilst in adult rat [16] and lamb [14], the amiloride-insensitive component is 55% and 30%, respectively. Junor and colleagues, employing the in vivo lamb lung preparation, demonstrated a pharmacology which was consistent with the involvement of cyclic nucleotide-gated cation (CNG) channels [14], a notion supported recently by Norlin et al. [16] who have shown direct upregulation of amiloride-insensitive absorption with cGMP. Expression of this phenotype appears to be under developmental control since pimozide (a CNG channel blocker) is ineffective in fetal lamb lung [17]. Recently, such in vivo observations have been corroborated in a cellular system by demonstrating directly, amiloride-insensitive, cGMP-evoked whole-cell Na+ currents in postnatal alveolar epithelial cells [18].
Crucial information currently missing from the picture which defines alveolar fluid homeostasis concerns the potential differential location of key ion channel and transport proteins within the alveolar epithelium itself, i.e are they in alveolar type I cells or alveolar type II cells? Thus, until very recently, the processes which underlie physiological fluid reabsorption were believed to be physically situated exclusively in the alveolar type II cell. Although it is true that alveolar type II cells express the entire gamut of proteins believed to be sufficient for efficient vectorial ion, solute and water transport (see [19,20] for earliest reports of type II cell monolayer properties), the contribution to the reabsorptive response of alveolar type I cells has never been robustly investigated, due essentially to the difficulty in routinely isolating such cells. This problem has been confounded by the use of polarised monolayers of adult alveolar cells which have been consistently reported as being type II cultures but which by many parameters have characteristics of type I cells [21,22]. Several recent studies have suggested that type I cells are of importance to ion and water balance in the lung. Thus, employing immunohisto- and immunocytochemistry, two concurrent reports demonstrated expression of αENaC in both alveolar type I and type II cells [23,24], whilst aquaporin 5 was demonstrated to be exclusively localized to the apical membrane of the alveolar type I cell [23]. Using a relatively low purity cell preparation, alveolar type I cells became strongly implicated in ion transport by the demonstration that they transported Na+ in an amiloride-sensitive manner; quantitatively, this sodium transport was almost 2.5-fold larger than that afforded by alveolar type II cells [24]. Evidence for the involvement of alveolar type I cells in fluid transport came from an ex vivo observation that the ouabain-sensitivity of the largest proportion of lung fluid reabsorption closely matched that of the α2 subunit of Na,K-ATPase (a subunit that was shown by those authors to be localized to type I cells [25]).
A further disadvantage of cellular studies on isolated alveolar epithelial cells is that they provide little information about the physiological interactions that occur between type I and type II cells. The fact that such interactions occur and are of physiological import has recently become apparent with the development of a co-culture system. This system has demonstrated that functional gap junctions exist between alveolar epithelial cells and that the expression profile of connexins that form these gap junctions differs depending on the cell type which is interacting [26,27].
Although type I cells are now implicated in alveolar ion transport, direct electrophysiological evidence of specific ion channels is thus far completely lacking. There are two main reasons for such a gap in our knowledge. Firstly, highly pure alveolar type I cell monolayers have not been generated on permeable supports (making short-circuit current measurements unfeasible). Secondly, isolated alveolar type I cells appear to be overly fragile and, as a consequence, have been unamenable thus far to patch-clamp studies. Neither of these problems may prove insurmountable with further advances in production of relatively pure alveolar type I monolayers [28], but the fact remains that no electrophysiological information from either cell type in situ is currently available.
In order to address directly the most important outstanding question in adult alveolar ion handling, i.e. what cells express which channels, we have developed a novel lung slice preparation which can be utilised for electrophysiological study. Using this technique we show that: a) cells within the alveolar epithelium remain viable for a number of hours; b) positive identification of living type I cells is routinely possible using imunocytochemistry and; c) single ion channels can be reliably recorded from the apical membrane of identified type I cells in situ. This advance will allow us to ascribe function to specific cell types within the alveolar epithelium and represents a novel methodology for investigating ion transport in fetal and postnatal lungs. Some of this work has appeared previously in abstract form [29,30].
Methods
Reagents
All compounds were of analytical grade and were purchased from BDH Laboratory Supplies (Poole, Dorset, U.K.) unless otherwise stated. Sodium isethionate, N-(2-Hydroxyethyl)piperazine-N'-(2-ethanesulfonic acid) 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), tetraethylammonium chloride (TEA), amiloride, niflumic acid and bovine serum albumin (BSA) were obtained from Sigma-Aldrich (Poole, Dorset, U.K.). Potassium isethionate was obtained from Fisher Scientific (Loughborough, Leicestershire, U.K.).
Slice preparation
6–8 day old Wistar rats were killed in accordance with the Home Office guidelines. Heart and lungs were removed en bloc, and washed in a physiological solution containing (in mM): 140 NaCl, 5 KCl, 10 HEPES, 1 CaCl2, 1.2 MgCl2, 5 D-glucose, pH 7.4, 300 mOsM). Individual lung lobes were dissected and fixed directly to the jig of an Integraslice (7550 PSDS, Campden Instruments, Leicester, Leicestershire, UK) by cyanoacrylate adhesive. The Integraslice allows reproducible, thin slicing of delicate tissue without the need to fill with stabilizing agents such as low-melting point agarose; a manoeuvre which we have found to limit the ability of making patch-clamp recordings in living lung slices. 200 μm lung slices were sectioned on a transverse plane across the lobe. During sectioning, the temperature of the bathing physiological solution was maintained at 4°C using a temperature-controlled specimen bath (765 HP, Campden Instruments.). Lung tissue slices were then transferred to a 24-well plate containing ~2 ml of physiological solution and kept at 37°C in a humidified incubator gassed with 5% CO2 / 95% air prior to immunohistochemical staining and electrophysiological recording.
Live/Dead Staining
Lung tissue slices were treated with Live/dead® viability/cytotoxicity assay kit for animal cells (Molecular Probes, Eugene, Oregon, USA) which was stored at -20°C and allowed to warm to room temperature prior to experimentation. The live (Acetoxymethylester of calcein, (calcein-AM), 4 mM) and dead (Ethidium homodimer (Eth-D), 12 mM) stock reagents were diluted to their final working concentrations (1 μM and 4 μM, respectively) in physiological solution (see above). The slices were then incubated in these solutions at room temperature for 30 minutes. Images were collected using BioRad CellMap confocal system (BioRad, Hertfordshire, UK) mounted on an Olympus BX50WI microscope (Olympus Optical Co. (Europa) GmbH, Hamburg, Germany). In order to confirm the validity of the live/dead staining, lung tissue slices were also treated with 1% Triton X-100 for 15 minutes prior to staining with 1 μM calcein and 4 μM EthD-1 solution for 30 minutes and subsequent imaging.
Live immunostaining of alveolar type I cells
In order to identify positively living alveolar type I cells within the lung slice, the type I cell-specific, mouse monoclonal primary antibody, VIIIB2 [31], was employed. Lung slices were washed three times for five minutes with 2 ml of warmed PBS with Ca2+ and Mg2+ (Sigma-Aldrich, Poole, Dorset, U.K.). Lung slices were then incubated at 37°C for 1.5 hours in the presence of VIIIB2 (1:5 dilution). Lung slices were washed three times more for five minutes each in PBS and then incubated in a 5% solution of BSA for 15 minutes at 37°C, followed by washing three times for five minutes in warmed PBS. Slices were then labelled with aN anti-mouse IgG secondary antibody (Alexa Fluor 488, Molecular Probes, Strathclyde, Paisley, U.K.) at a 1:500 dilution and incubated for upwards of one hour at 37°C. Immediately prior to imaging, the lung slices were washed once again with warm PBS to remove any unbound antibody. Images were collected using BioRad CellMap confocal system mounted on an Olympus BX50WI microscope. In some experiments, lung slices were stained with either calcein-AM (1 μM, 30 minutes) or EthD-1 (4 μM, 30 minutes) following the usual VIIIB2 immunostaining protocol. In this cases, the secondary antibody was either a 1:500 dilution of mouse Alexa Fluor 546 secondary antibody (Molecular Probes) for calcein co-staining or a 1:500 dilution of mouse Alexa Fluor 488 for EthD-1 co-staining.
In a number of lung slices (from 5 animals), the mean number of dead cells per field (105 μM × 105 μM) was assessed (as EthD-1 positive nuclei). Such quantification was performed on untreated (n = 41), Triton-treated (n = 30) and VIIIB2-treated (n = 31) slices. Sample means were compared using one-way ANOVA followed by Bonferroni post-hoc test, with 0.05 taken as the level of significance.
Electrophysiology
Lung slices were placed in a recording chamber mounted on the Olympus BX50WI microscope and alveolar type I cells were identified prior to patch clamp using standard fluorescence microscopy. Slices were held in place by a platinum ring with nylon cross wires in order to stabilize it for patch clamp. Recording pipettes were pulled from borosilicate glass and fire-polished (Narishige MF-83 microforge) and coated with Sigmacote® (Sigma-Aldrich). Pipettes had resistances of 10-15MΩ when filled with pipette solution, which contained (in mM): 145 Na Isethionate, 10 HEPES, 1 CaCl2, 1.2 MgCl2, 5 D-Glucose, 10 mM TEA, 100 μM niflumic acid, 10 μM amiloride, with the pH adjusted to 7.4 with NaOH.
Following formation of a gigaohm seal, single-channel activity was recorded at room temperature in the cell-attached configuration of the patch-clamp technique. Cells were held at potentials ranging from +30 to -90 mV (Vm, assuming resting potential of -40 mV). Resistive-feedback voltage-clamp was achieved using an Axon mulitclamp 700A amplifier (Axon Instruments, Foster City, CA, USA). Voltage protocols were generated and currents recorded using pClamp 9.0 software, employing a Digidata 1322 A/D converter (Axon Instruments). Data were filtered (4-pole Bessel) at 2 kHz and digitized at 5 kHz. Data analyses were performed using the pClamp 9.0 suite of software (Axon Instruments).
Results
The lung slice preparation retains living cells
The viability of cells within lung slices was assayed using the Live/Dead® assay kit which differentially stains living or dead cells. Viable cells can be distinguished by the ability of their intracellular esterases to convert the non-fluorescent, cell-permeant calcein-AM ester to fluorescent calcein. Fluorescent calcein is retained within the cytoplasm of living cells and can be detected upon excitation at 488 nm (green emission). The damaged membranes of non-viable cells are permeable to EthD-1. Dead cells can be distinguished by nucleic acid-bound EthD-1 in their nuclei which can be detected upon excitation at 532 nm (red emission). The ability of this technique to distinguish between living and dead cells was confirmed using HEK 293 cells before and following treatment for 15 minutes with Triton-X-100 (data not shown). Following the successful optimization of the fluorescent live/dead assay in cultured cells, the same approach was utilized to determine the viability of the cells within the lung slices. As for the cell line, rat lung slices were loaded by incubation with 1 μM calcein and/or 4 μM EthD-1 in physiological solution for 30 minutes. Following this incubation, the majority of cells within a slice were positive for calcein (Figure 1A, C) whilst a small proportion of cells demonstrated EthD-1 positive nuclei (Figure 1B, C). The overlain images (Figure 1C) indicates that the majority of cells within the lung slice are viable. Pre-treatment with 1% Triton X-100 for 15 minutes resulted in a dramatic loss of calcein-positivity (Figure 1D, F) and a large increase in EthD-1 incorporation (Figure 1E, F). Quantification of the mean number of EthD-1 positive nuclei showed that Triton-X-100 treatment resulted in a significant, 4-fold increase in the number of dead cells per field from 18.0 ± 0.9 (n = 41) to 78.1 ± 3.3 (n = 30 slices; Figure 2G)). Thus, before such treatment, the majority of cells within the slice were alive whilst permeabilization with detergent resulted in cell death throughout the lung slice.
Figure 1 Confocal images of rat lung slice co-stained with the LIVE/DEAD® viability/cytotoxicity assay kit for animal cells. A) A 200 μm rat lung slice loaded with 1 μM calcein-AM/4 μM EthD-1 and excited with light of 488 nm (green emission) to show viable cells. B) The same 200 μm rat lung slice loaded with 1 μM calcein-AM/4 μM EthD-1 and excited with light of 532 nm, (red emission) to show the potential dead cells. C) Overlay of images (A) and (B) showing that the majority of the cells are alive. D) A 200 μm rat lung slice loaded with 1 μM calcein-AM/4 μM EthD-1 following a 15 minute treatment with 1% Triton X-100 and then excited with light of 488 nm (green emission) to show viable cells. E) The same 200 μm rat lung slice loaded with 1 μM calcein-AM/4 μM EthD-1 following a 15 minute treatment with 1% Triton X-100 and then excited with light of 532 nm, (red emission) to show the potential dead cells. F) Overlay of images (D) and (E) showing that no live cells remain following treatment with Triton X-100.
Figure 2 Confocal images of rat lung slice co-stained with either the LIVE® or DEAD® viability stain and the specific alveolar type I cell antibody, VIIIB2. A) A 200 μm rat lung slice incubated with the live stain (1 μM Calcein-AM) following live immunohistochemical staining with VIIIB2 primary antibody/Alexa Fluor 532 secondary antibody and excited at 488 nm (green emission) to show living cells. B) The same lung slice incubated with the live stain (1 μM Calcein-AM) following live immunohistochemical treatment with VIIIB2 primary antibody/Alexa Fluor 546 secondary antibody and excited at 532 nm (red emission) to show the Alexa Fluor 546 secondary antibody staining at alveolar type I cells. C) Overlay of images (A) and (B) showing that VIIIB2 immunoreactivity is clearly restricted to elongated, thin cells located at the edge of the alveolar space (indicated by the arrows), characteristic of ATI cells. D) A 200 μm rat lung slice incubated with the dead stain (4 μM EthD-1) following live immunohistochemical staining with VIIIB2 primary antibody/Alexa Fluor 488 secondary antibody and excited at 488 nm (green emission) to show secondary antibody staining of alveolar type I cells. E) The same lung slice incubated with the dead stain (4 μM EthD-1) following live immunohistochemical treatment with VIIIB2 primary antibody/Alexa Fluor 488 secondary antibody and excited at 532 nm (red emission) to show the EthD-1 positive, dead cell nuclei. F) Overlay of images (D) and (E) showing that VIIIB2 immunoreactivity and the dead cell stain do not co-localise. G) Graph showing quantification of dead cells per field (105 μM × 105 μM) following treatments indicated at the bottom of each data set.
Utilizing the VIIIB2 antibody (to label selectively alveolar type I cells [31]) in conjunction with calcein fluorescence allowed estimation of the viability of specific cells within the epithelium. Figure 2A, B and 2C show a typical results from a 200 μm lung slice which was treated with 1 μM calcein-AM for 30 minutes following the VIIIB2 live cell immunohistochemical protocol. Again, the majority of cells were calcein-positive (Figure 2A). The mouse VIIIB2 immunoreactivity was clearly restricted to elongated, thin cells located at the edge of the alveolar space (Figure 2B), a pattern of immunoreactivity characteristic of alveolar type I cells [31]. The overlain images (Figure 2C) show the green cytoplasm of viable alveolar type I cells in close proximity to the VIIIB2 immunoreactivity at the plasma membrane. Furthermore, using the VIIIB2 antibody in conjunction with EthD-1 fluorescence allowed investigation and quantification of the effect of the immunohistochemical staining protocol of cell death within the slice. Figure 2D shows again, but this time using the Alexa 488 secondary antibody, that VIIIB2 staining is localised to the alveolar type I apical membrane. Figure 2E shows the modest cell death which has occurred as a consequence of the live immunohistochemistry and Figure 2F indicates that cell death is not typically induced in alveolar type I cells following the immnostaining protocol. The modest, but significant 1.4-fold increase cell death evoked by live immunohistochemistry is quantified in Figure 2G which shows that VIIIB2 treatment results in 25.3 ± 1.7 dead cells per field (n = 31) compared with 18.0 ± 0.9 in untreated slices. Excluding the VIIIB2 from the immunohistochemical protocol or replacing VIIIB2 with an antibody known only to bind an intracellular epitope (BKα-subunit antibody from Santa Cruz) resulted in no fluorescent signal (data not shown).
Having shown that labelled type I cells within the lung slices were viable, the patch clamp technique was utilized to examine ion channel activity in such positively identified cells. Figure 3Ai shows a 200 μm lung slice immunostained with VIIIB2 to identify alveolar type I cells. Again, VIIIB2 immunoreactivity was restricted to elongated cells located at the edge of the alveolar space, characteristic of type I cells. In order to show the position of the recording pipette, the tip was filled with tetramethylrhodamine isothiocyanate (TRITC, 1:400) as shown in Figure 3Aii. Figure 3Aiii shows light-field view of the pipette and tissue and the overlain images (plus the blue transmitted light image, Figure 3Aiv) indicates clearly that the tip of the patch pipette can be manipulated onto the apical membrane of an identified alveolar type I cell. Using this cell-attached configuration, single channel currents could be recorded routinely from VIIIB2-immunoreactive cells. Figure 2B shows an exemplar current trace recorded at 0 mV (Vm), filtered at 2 kHz and with 10 mM TEA excluded from the pipette solution. Such "unfiltered" currents are extremely complex in nature when the broad-spectrum potassium blocker, TEA, is excluded from the extracellular recording solution indicating that alveolar type I cells express at least two populations of K+ channels – those which are TEA-sensitive and those which are not. However, when TEA was included in the pipette solution and the recordings were filtered at 200 Hz or lower, single channel openings could be resolved. A family of such currents recorded from an alveolar type I cell in the cell-attached configuration is shown in Figure 3C. Under the stringent recording conditions employed (where ENaC, the majority of potassium and all chloride channels are inhibited pharmacologically), changing the applied membrane potential from between -90 mV through +30 mV resulted in the activation of small conductance (Figure 3D), voltage-dependent channels (Figure 3E) whose currents reversed around -60 to -70 mV (Figure 3D), which is close to EK under the imposed ionic conditions. The current-voltage relationship (Figure 3D) of these channels recorded from three alveolar type I cells in separate lung slices gave a mean single channel conductance of 21 ± 3 pS. Figure 3E shows the product of the open state probability and the channel number (NPo) versus voltage relationship of the same channels and indicates that activity was voltage-dependent, activating at potentials more depolarized than -40 mV. These exemplar electrophysiological recordings show for the first time that ion channel activity can be successfully recorded from positively identified, living alveolar epithelial type I cells within an in situ lung preparation.
Figure 3 Cell-attached single-channel recordings from identified alveolar epithelial type I cells in a rat lung slice. A(i) A 200 μm rat lung slice immunostained stained with the mVIIIB2 antibody (Alexa Fluor 488 secondary antibody, green emission). (ii) Patch-clamp recording pipette in close proximity to an alveolar epithelial cells whose tip has been filled with TRITC (1:400). (iii) Transmitted light image. (iv) Overlay of images (i) and (ii) plus blue transmitted image to show that the patch recording pipette has formed a gigaohm seal on an VIIIB2-immunopositive alveolar type I cell. B) Exemplar current recording without 10 mM TEA in the pipette to demonstrate the complex nature of the current. Currents were recorded at 0 mV (Vm) and filtered at 2 kHz C) Representative family of cell-attached currents recorded from an alveolar epithelial type I cell, identified by live immunohistochemistry (as in (A)) in an acutely isolated rat lung slice. Single channel activity was recorded in the cell-attached configuration of the patch-clamp technique using a low chloride pipette solution with 10 μM amiloride, niflumic acid and 10 mM TEA. Currents were recorded at the potentials indicated to the left of each trace and were filtered at 200 Hz. L1, L2 and L3 indicate channel open levels 1, 2 and 3, respectively, whilst C indicates the channel closed level. D) Current voltage relationship of the single-channel activity recorded in 3 identified type I cells from 3 separate lung slice preparations. E) NPo (product of open state probability and number of channels) versus voltage plot of the single-channel activity shown in (E).
Discussion
Until now, cellular studies on alveolar epithelial cells have been almost exclusively limited either to direct observation from freshly purified and cultured cell populations or by inference from in vivo/ex vivo measurements of lung function. Such limitations notwithstanding, an enormous amount of important information has been gathered and a widely accepted integrated model of the ion and fluid homoestatic processes which clear the lung of fluid at birth and keep it essentially dry thereafter (in both health and disease) has emerged. Thus, it is commonly believed that alveolar type II cells are the principal cellular element responsible for generating the transepithelial osmotic driving force and that alveolar type I cells represent relatively passive players in the vectorial transport process. This almost universally accepted belief is based on a number of key observations from many different researchers employing a wide variety of techniques. Although not an exhaustive list, such crucial data include the observations that: a) alveolar type II cells in culture express ENaC subunits (e.g. [32]); b) manipulation of culture conditions to emulate the alveolar air/liquid interface induces functional ENaC expression in type II cells [33]; c) ENaC is necessary for effective fluid reabsorption at birth [13] and contributes one component to the reabsorptive response required for the resolution of postnatal pulmonary oedema [34] and; d) β-adrenoreceptors, believed to be necessary for catecholamine-evoked fluid reabsorption are expressed in type II cells [35]. Whilst all these observations are undoubtedly true, there are two important reasons why a model of lung fluid homeostasis which is so centred on alveolar type II cells may not represent the entire picture in vivo. Firstly, alveolar type II cells rapidly transdifferentiate into a type I cell phenotype in culture (see [36] for recent review). Secondly, only a very limited number of studies have actually investigated the cellular properties of alveolar type I cells, either in vivo or in vitro. However, new information is beginning to emerge which suggests that type I cells have the potential to contribute significantly to overall alveolar fluid homeostasis. For example, they express β-adrenoceptors [37] and other important ion transport proteins [23,38], have the capacity to generate a large ouabain-sensitive Na+ flux [25] and demonstrate significant amiloride-sensitive Na+ uptake [24], and currents (Kemp, Kim and Borok, unpublished observations).
Clearly, one way to resolve the current controversies and to generate a model which includes data from both epithelial cell populations is to study cellular function of identified cell types in situ. To this end, an elegant ex vivo study has demonstrated Ca2+ oscillations in alveolar type II cells which have been imaged in an isolated, perfused lung [39]. However, although this model has been used to puncture single alveoli with an injection microelectrode, it has not yet been amenable to single cell electrophysiology [40]. As a compromise, we report here the development of a technique which employs freshly cut lung slices to study the ion channel properties of identified alveolar type I cells in situ. The majority of cells within the lung slice are viable for up to 8 hours when bathed in physiological solution (Figure 1). The most important new aspect of the preparation is the ability to immunolocalize living alveolar type I cells within the slice. The most crucial tool for successful live immunostaining is a highly cell-specific antibody which is has been raised against an extracellular epitope; VIIIB2 is such an antibody and the data in Figure 2 show a pattern of live immunostaining which is unique to alveolar type I cells. The notion that this staining was cell specific was reinforced by the demonstration that employing a primary antibody known only to bind an intracellular epitope or removing the VIIIB2 from the protocol resulted in no fluorescent signal. A further important development was the adaptation of standard immunohistochemical techniques in order to reduce the amount of time that tissue is incubated with the various reagents. Thus, we optimised the staining protocols so that the entire procedure took less than three hours. This immunostaining was conducted entirely using physiological solutions and resulted in only a modest reduction in cell viability (see Figure 2). Crucially, this novel combination of lung slice and live cell immunohistochemistry has allowed us to identify viable alveolar type I cells in situ with sufficient accuracy to be able to manipulate patch-clamp electrodes onto their apical membranes. For the purposes of demonstrating the feasibility of the in situ, lung electrophysiological technique, we designed the pipette solution so that most ion channels within the membrane would be inhibited by employing an inhibitor cocktail which included TEA, niflumic acid, amiloride and low chloride. Under these conditions, we have shown that single channel currents can be recorded (Figure 3C) with a resolution sufficient to undertake basic biophysical characterization. Interestingly, employing a pipette solution which did not contain TEA resulted in recordings which were not readily amenable to robust analyses. Clearly, the challenge will now be to dissect each individual channel using a combination of pharmacology and molecular abrogation technologies in excised patches, manoeuvres outwith the scope of the current feasibility study. However, reducing the number of channel types within a cell-attached patch by employing a high concentration of TEA has allowed us to show that alveolar type I cells express a 23pS, voltage-activated conductance. The analyses of these channels shown in Figure 3 indicates that this channel is primarily permeable to potassium although a contribution from other ions cannot be completely rules out (Figure 3D). Further, this channel is TEA-insensitive and shows voltage-activation in the physiological range (Figure 3E). These characteristics suggest that the most likely potassium channel candidate underlying the recorded cell attached patch current belongs to Kv family, perhaps Kv1.7 (see [41] for review of known potassium channel characteristics). Potassium channels have also been reported in alveolar type II cells [42,43] where they are proposed to contribute to the regulation of driving forces across the epithelium. A parallel function may also hold true for such channels expressed in type I cells.
Finally, in addition to the feasibility of recording single ion channels expressed in alveolar type I cells, this technique has potentially much wider application and should, for example, allow investigation of Ca2+ and pH homeostasis using fluorescent indicator systems in individual cell populations and to study the interactions between alveolar type I and II cells in situ.
Conclusion
We report a development of the lung slice technique which allows immunolocalization of living type I cells within the alveolar epithelium. Using this novel positive identification, we have described a method which allows, for the first time, patch clamp analysis of single ion channels expressed on the apical membrane of alveolar epithelial type I cells in situ.
Authors' contributions
All authors contributed to the conception and design of the study and to the discussion of the data. SB developed the lung slicing and immunostaining protocols. SB, ZB and PJK developed the live staining to identify type I cells whilst SB, HSM and PJK developed the in situ electrophysiology. The initial manuscript was written by HSM, SB and PJK and the final editing was conducted jointly by all the authors.
Acknowledgements
The authors should like to acknowledge the Welcome Trust, Hastings Foundation and National Institutes of Health for funding of this project.
==== Refs
Matalon S O'Brodovich H Sodium channels in alveolar epithelial cells: molecular characterization, biophysical properties, and physiological significance Annu Rev Physiol 1999 61 627 661 10099704 10.1146/annurev.physiol.61.1.627
Folkesson HG Norlin A Baines DL Salt and water transport across the alveolar epithelium in the developing lung: Correlations between function and recent molecular biology advances Int J Mol Med 1998 2 515 531 9858647
Kemp PJ Olver RE G protein regulation of alveolar ion channels: implications for lung fluid transport Exp Physiol 1996 81 493 504 8737082
Matthay MA Folkesson HG Verkman AS Salt and water transport across alveolar and distal airway epithelia in the adult lung Am J Physiol 1996 270 L487 L503 8928808
Kemp PJ Kim KJ Spectrum of ion channels in alveolar epithelial cells: implications for alveolar fluid balance Am J Physiol Lung Cell Mol Physiol 2004 287 L460 L464 15308494 10.1152/ajplung.00191.2004
Barker PM Strang LB Walters DV The role of thyroid hormones in maturation of the adrenaline-sensitive lung liquid reabsorptive mechanism in fetal sheep J Physiol (Lond) 1990 424 473 485 2391659
Barker PM Walters DV Markiewicz M Strang LB Development of the lung liquid reabsorptive mechanism in fetal sheep: synergism of triiodothyronine and hydrocortisone J Physiol 1991 433 435 449 1841951
O'Brodovich HB Hannam V Seear M Mullen JBM Amiloride impairs lung water clearance in newborn guinea pigs J Appl Physiol 1990 68 1758 1762 2161411
Olver RE Ramsden CA Strang LB Walters DV The role of amiloride-blockable sodium transport in adrenaline-induced lung liquid reabsorption in the fetal lamb J Physiol 1986 376 321 340 3795077
Baines DL Ramminger SJ Collett A Haddad JJ Best OG Land SC Olver RE Wilson SM Oxygen-evoked Na+ transport in rat fetal distal lung epithelial cells J Physiol 2001 532 105 113 11283228 10.1111/j.1469-7793.2001.0105g.x
Rafii B Coutinho C Otulakowski G O'Brodovich H Oxygen induction of epithelial Na+ transport requires heme proteins Am J Physiol Lung Cell Mol Physiol 2000 278 L399 L406 10666125
Rafii B Tanswell AK Otulakowski G Pitkanen O Belcastro-Taylor R O'Brodovich H O2-induced ENaC expression is associated with NF-kB activation and blocked by superoxide scavenger Am J Physiol 1998 275 L764 L770 9755109
Hummler E Barker P Gatzy J Beerman F Verdumo C Schmidt A Boucher R Rossier BC Early death due to defective neonatal lung liquid clearance in aENaC-deficient mice Nature Genetics 1996 12 325 328 8589728 10.1038/ng0396-325
Junor RW Benjamin AR Alexandrou D Guggino SE Walters DV A novel role for cyclic nucleotide-gated cation channels in lung liquid homeostasis in sheep J Physiol 1999 520 255 260 10517816 10.1111/j.1469-7793.1999.00255.x
Norlin A Finley N Abedinpour P Folkesson HG Alveolar liquid clearance in the anesthetized ventilated guinea pig Am J Physiol 1998 274 L235 L243 9486208
Norlin A Lu LN Guggino SE Matthay MA Folkesson HG Contribution of amiloride-insensitive pathways to alveolar fluid clearance in adult rats J Appl Physiol 2001 90 1489 1496 11247951 10.1063/1.1382830
Junor RW Benjamin AR Alexandrou D Guggino SE Walters DV Lack of a role for cyclic nucleotide gated cation channels in lung liquid absorption in fetal sheep J Physiol 2000 523 493 502 10699091 10.1111/j.1469-7793.2000.t01-3-00493.x
Kemp PJ Kim KJ Borok Z Crandall ED Re-evaluating the nature of the Na+ conductance of adult alveolar type II epithelial cells: Evidence for the involvement of cGMP-activated cation channels. J Physiol 2001 536 693 701 11691865 10.1111/j.1469-7793.2001.t01-1-00693.x
Goodman BE Crandall ED Dome formation in primary cultured monolayers of alveolar epithelial cells Am J Physiol 1982 243 C96 100 7046469
Mason RJ Williams MC Widdicombe JH Sanders MJ Misfeldt DS Berry LCJ Transepithelial transport by pulmonary alveolar type II cells in primary culture Proc Natl Acad Sci U S A 1982 79 6033 6037 6964398
Cheek JM Evans MJ Crandall ED Type I cell-like morphology in tight alveolar epithelial monolayers Exp Cell Res 1989 184 375 387 2806398 10.1016/0014-4827(89)90337-6
Borok Z Lubman RL Danto SI Zhang XL Zabski SM King LS Lee DM Agre P Crandall ED Keratinocyte growth factor modulates alveolar epithelial cell phenotype in vitro: expression of aquaporin 5 Am J Respir Cell Mol Biol 1998 18 554 561 9533944
Borok Z Liebler JM Lubman RL Foster MJ Zhou B Li X Zabski SM Kim KJ Crandall ED Na transport proteins are expressed by rat alveolar epithelial type I cells Am J Physiol Lung Cell Mol Physiol 2002 282 L599 L608 11880284
Johnson MD Widdicombe JH Allen L Barbry P Dobbs LG Alveolar epithelial type I cells contain transport proteins and transport sodium, supporting an active role for type I cells in regulation of lung liquid homeostasis Proc Natl Acad Sci U S A 2002 99 1966 1971 11842214 10.1073/pnas.042689399
Ridge KM Olivera WG Saldias F Azzam Z Horowitz S Rutschman DH Dumasius V Factor P Sznajder JI Alveolar type 1 cells express the alpha2 Na,K-ATPase, which contributes to lung liquid clearance Circ Res 2003 92 453 460 12600893 10.1161/01.RES.0000059414.10360.F2
Isakson BE Seedorf GJ Lubman RL Boitano S Heterocellular cultures of pulmonary alveolar epithelial cells grown on laminin-5 supplemented matrix In Vitro Cell Dev Biol Anim 2002 38 443 449 12605538 10.1290/1071-2690(2002)038<0443:HCOPAE>2.0.CO;2
Isakson BE Seedorf GJ Lubman RL Evans WH Boitano S Cell-cell communication in heterocellular cultures of alveolar epithelial cells Am J Respir Cell Mol Biol 2003 29 552 561 12748060 10.1165/rcmb.2002-0281OC
Chen J Chen Z Narasaraju T Jin N Liu L Isolation of highly pure alveolar epithelial type I and type II cells from rat lungs Lab Invest 2004
Bourke S Mason HS Crandall ED Kim KJ Borok Z Kemp PJ Calcium homeostasis in identified alveolar epithelial cells: development of a novel in situ lun gslice model J Physiol 2004 552P P119 P119
Bourke S Mason HS Kim KJ Crandall ED Borok Z Kemp PJ Recording ion channels in alveolar type I cells utilizing a novel in situ lung slice model. 2004 Experimental Biology meeting abstracts FASEB J 2004 18 Abstract #469.1
Danto SI Zabski SM Crandall ED Reactivity of alveolar epithelial cells in primary culture with type I cell monoclonal antibodies Am J Respir Cell Mol Biol 1992 6 296 306 1540393
Danto SI Borok Z Zhang XL Lopez MZ Patel P Crandall ED Lubman RL Mechanisms of EGF-induced stimulation of sodium reabsorption by alveolar epithelial cells Am J Physiol 1998 275 C82 C92 9688838
Jain L Chen XJ Ramosevac S Brown LA Eaton DC Expression of highly selective sodium channels in alveolar type II cells is determined by culture conditions Am J Physiol Lung Cell Mol Physiol 2001 280 L646 L658 11238004
Egli M Duplain H Lepori M Cook S Nicod P Hummler E Sartori C Scherrer U Defective respiratory amiloride-sensitive sodium transport predisposes to pulmonary oedema and delays its resolution in mice J Physiol 2004 560 857 865 15308680 10.1113/jphysiol.2004.066704
Fabisiak JP Vesell ES Rannels DE Interactions of beta adrenergic antagonists with isolated rat alveolar type II pneumocytes. I. Analysis, characterization and regulation of specific beta adrenergic receptors J Pharmacol Exp Ther 1987 241 722 727 2883305
Williams MC Alveolar type I cells: molecular phenotype and development Annu Rev Physiol 2003 65 669 695 12428023 10.1146/annurev.physiol.65.092101.142446
Liebler JM Borok Z Li X Zhou B Sandoval AJ Kim KJ Crandall ED Alveolar epithelial type I cells express beta2-adrenergic receptors and G-protein receptor kinase 2 J Histochem Cytochem 2004 52 759 767 15150284 10.1369/jhc.3A6184.2004
Gonzalez R Yang YH Griffin C Allen L Tigue Z Dobbs L Freshly-isolated rat alveolar type I cells, type II cells, and cultured type II cells have distinct molecular phenotypes Am J Physiol Lung Cell Mol Physiol 2005 288 L179 L189 15447939 10.1152/ajplung.00272.2004
Ashino Y Ying X Dobbs LG Bhattacharya J [Ca2+]i oscillations regulate type II cell exocytosis in the pulmonary alveolus Am J Physiol Lung Cell Mol Physiol 2000 279 L5 13 10893197
Wang PM Ashino Y Ichimura H Bhattacharya J Rapid alveolar liquid removal by a novel convective mechanism Am J Physiol Lung Cell Mol Physiol 2001 281 L1327 L1334 11704526
Coetzee WA Amarillo Y Chiu J Chow A Lau D McCormack T Moreno H Nadal MS Ozaita A Pountney DJ Saganich M Vega- Saenz de Miera E Rudy B Molecular diversity of K+ channels Ann N Y Acad Sci 1999 868 233 285 10414301
Peers C Kemp PJ Boyd CAR Nye PCG Whole-cell K+ currents in type II pneumocytes freshly isolated from rat lung - pharmacological evidence for 2 subpopulations of cells Biochim Biophys Acta 1990 1052 113 118 2322587 10.1016/0167-4889(90)90064-K
DeCoursey TE Jacobs ER Silver MR Potassium currents in rat type II alveolar epithelial cells. J Physiol 1988 395 487 505 2457683
| 15857506 | PMC1131928 | CC BY | 2021-01-04 16:36:26 | no | Respir Res. 2005 Apr 27; 6(1):40 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-40 | oa_comm |
==== Front
Theor Biol Med ModelTheoretical Biology & Medical Modelling1742-4682BioMed Central London 1742-4682-2-171585750710.1186/1742-4682-2-17ResearchAn ETP model (exclusion-tolerance-progression) for multi drug resistance Kannan Subburaj [email protected] Division of Gastroenterology, School of Medicine, University of Pennsylvania, Philadelphia, PA, 19104, USA2005 27 4 2005 2 17 17 22 12 2004 27 4 2005 Copyright © 2005 Kannan; licensee BioMed Central Ltd.2005Kannan; 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 is known that sensitivity or resistance of tumor cells to a given chemotherapeutic agent is an acquired characteristic(s), depending on the heterogeneity of the tumor mass subjected to the treatment. The clinical success of a chemotherapeutic regimen depends on the ratio of sensitive to resistant cell populations.
Results
Based on findings from clinical and experimental studies, a unifying model is proposed to delineate the potential mechanism by which tumor cells progress towards multi drug resistance, resulting in failure of chemotherapy.
Conclusion
It is suggested that the evolution of multi drug resistance is a developmentally orchestrated event. Identifying stage-specific time windows during this process would help to identify valid therapeutic targets for the effective elimination of malignancy.
ETP ModelDrug resistanceDrug sensitivityPhysiological drug resistancePathological drug resistance
==== Body
Background
The phenomenon of drug resistance is a distinct and multifactorial entity culminating in the failure of therapeutic regimens in clinical oncology. From the clinical perspective, the emergence of drug resistance is determined by the rate of tumor growth, in conjunction with the remission index subsequent to chemotherapy. In contrast, experimental model(s) for studying drug resistance involve either homogeneous cell populations or co-culture models where the time frame ranges from a few days to a week at the most. It is obvious that the biochemical and collective physiological process that coexist in the cancer patient are totally distinct and do not warrant direct comparison with experimental data.
A tumor mass encompassing approximately 1011 to 1012 cells is considered a lethal tumor burden. Depending on the sensitivity of a given tumor cell population to anti-neoplastic drug(s) (chemotherapeutic agents), and the dose administered, the effectiveness of the therapy, referred to as "Cell Kill", is determined. "Cell Kill" depends on the inherent susceptibility of the tumor burden. The "Cell Kill" of a given tumor burden varies between 90 % to 99.99%. If we assume that, under the best therapeutic regimen, one round of chemotherapy or radiation therapy would be likely to achieve a 99.99% "cell kill" in a tumor burden of 1011cells, this would reduce the tumor burden to 108 cells. It should be noted that chemotherapy and radiation therapy exert considerable toxic effects on normal cells; because of this factor, the treatment regimen is staged in cycles. It is reasonable to presume that both the normal cells and the less sensitive tumor cells would be likely to proliferate with one or more defects after each cycle of the therapy.
Even if the most fortunate circumstances were to prevail for early detection followed by treatment, the size of the "tumor burden", the sensitivity of the tumor cells and the effectiveness of the therapy ("Cell kill") remain decisive in determining the outcome of the therapy. Clinical observation shows that relapse or recurrence of the tumor is always a possibility. This is partly due to the fact that diagnostic procedures are inadequate for detecting as few as 106 or 108 tumor cells in cancer patients. This technical shortcoming confers a growth advantage on both undetectable and insensitive tumor cells. Under these circumstances, should there be a relapse, the tumor burden would probably be composed of a heterogeneous population of tumor cells. These include both drug sensitive and drug resistant cells in either proliferating or dormant states. According to Gompertzian Kinetics, as the tumor burden increases, the number of proliferating cells would decrease. Also, it is known that in a given tumor burden, a considerable number of cells are in the resting phase; these are not sensitive to chemotherapy or radiation therapy [1].
As a result it would require a much higher dose of chemotherapeutic drugs or radiation therapy to achieve the maximal "Cell Kill" in a given tumor burden. It is known that higher drug doses are often correlated with increased response rates in terms of the effective and maximal "Cell Kill", thus offering a window of opportunity for cure (complete remission). However, there is no assurance that complete remission would be the immediate outcome after a given therapeutic regimen. It is known that the effectiveness of an anti-neoplastic drug depends on i) the half life of the drug in vivo, ii) the rate and iii) amount of the drug being absorbed (bioavailability), and iv) the toxicities of biologically active metabolites of the drug [2-4].
The biologically active form of a drug, and the effective range of radiation therapy, are not uniformly distributed, so the entire tumor burden is not reached. Therefore, in a given tumor burden, the entire cell population is not exposed to effective therapy, leaving a finite region insufficiently exposed to the drug or its biologically active metabolite(s). Considering these factors in the context of ongoing therapy, it is common practice to re-evaluate patients after 2 or 3 cycles of chemotherapy to determine its effectiveness. Depending on the toxicity profile and the rate of tumor progression versus "Cell Kill", either the therapy is continued or multiple drugs are used to achieve maximal "Cell Kill" to obtain complete remission.
Here, I advance a hypothesis for multi drug resistance based on the aforementioned factors: the ETP model. The founding factors are: 1. The sigmoidal curve depicting the toxicity-dose relationship for a given chemotherapeutic drug indicates that lower doses give lower toxicity but less "Cell Kill", whereas an increased dose would increase the toxicity with a better "Cell Kill" [5,6]. However the optimal dose required for a given therapeutic compound to achieve a maximal "Cell Kill" for a particular malignancy with minimum cytotoxicity is not well defined. This paves the way for cells within the tumor burden to acquire and evolve one or more mechanism(s) for survival under the drug-induced toxic environment during and after treatment. Therefore, it follows that the surviving tumor cells in a given "tumor burden" should be viewed as the best-suited or best-adapted for withstanding these toxic effects as a result of acquiring a relevant genotype and/or phenotype conferring "multi drug resistance" [7,8].
2. In terms of experimental data, I have consistently observed that periodic exposure to alkylating agents is a principal requirement for retaining the drug-resistance property of the drug-resistant variant of human ovarian carcinoma compared to the drug-sensitive tumor cells in vitro. If no drug treatment were provided, would the drug-resistant variant revert to a sensitive phenotype and subsequently die? Absence of drug exposure of resistant cells indeed causes reversion to a drug sensitive phenotype, which is intriguing as it implies that drug resistance is a transient phenomenon not an everlasting property, at least in vitro. Also, as the selection pressure is maintained with periodic drug treatment, the doubling rate of the drug resistant cells decreases compared to the drug-sensitive phenotype (Figure 1). Obviously, a considerable difference in biochemical properties between these two cell lines is to be expected; however, this is not the case (Kannan, Unpublished observations). On the contrary, if chemotherapeutic or radiation therapy is discontinued, relapse with or without an aggressive increase in tumor burden is observed in a cancer patient.
Figure 1 Schematic representation of an experimental drug resistance model.
Absence of therapy does not result in a complete remission paralleling the complete reversion to drug sensitivity that is observed in vitro. The phenomenon is rare, indeed scarcely possible, in the clinical scenario. As such, there is no direct correlation between clinical case reports and experimental data. This contradiction warrants critical appraisal of the published literature, which is seminal and has exerted a profound impact on our understanding of the most complex and still unattainable goal of complete recovery in any given cancer treatment. Based on extensive analysis of the published literature and experimental evidence, I propose a model to account for the incongruity or discrepancy between the experimental and clinical drug resistance data.
Although the biochemical basis of drug resistance has been extensively studied in both in vitro model systems and clinical specimens, no correlation has been forthcoming. Considering the disparity between experimental and clinical findings, it is clear that we currently lack an understanding of the physiological basis of the evolution of drug-resistant cells. The expression of proteins that play a pivotal role in cellular drug resistance, such as P-glycoprotein, MRP-1 and other members of the ATP-binding cassette family of efflux protein(s), has been shown to decline at the time of relapse compared to the initial presentation of a given type of cancer. However, this observation has not been corroborated in all types of cancer. MRP-1 mediated drug efflux has been shown to correlate with an elevated level of intracellular glutathione (GSH), GSH synthesis or glutathione S-transferase (GST) activity. GST-mediated glutathionyl S-conjugates are known to be removed from the intracellular milieu by an energy-dependent process, similar to that seen with MRP-1, which is mediated by ATP-binding cassette proteins. It is known that GST is expressed in four different isoforms (α, , π, ) but it has not yet been confirmed that expression of GST-π confers drug resistance or whether this isoform is present in drug-resistant cells [9].
Among the properties that differentiate tumor cells from normal cells are the presence of growth factors potentiating vascular growth, highly heterogeneous oxygen tension distributions, extreme acidic or alkaline pH, higher rates of glucose delivery and utilization, and finally, a state of hypoxia with an acidotic environment that is noncycling. This, in turn, endows the cell population with a lack of uniform sensitivity to different families of chemotherapeutic drugs. Oxidative stress has been demonstrated to induce genomic instability at a much higher rate than is seen in drug-sensitive cells. All these factors contribute to the development of drug resistance [10].
In addition, drug-resistant cells have evolved mechanisms for bypassing apoptosis (the controlled form of cell death due to dehydration, shrinkage, and fragmentation of the nucleus, eventually leading to phagocytosis by macrophages) and necrosis (a traumatic but passive form of cell death due to the dysfunction of ion-transporting proteins, cell swelling and lysis and associated with the release of inflammatory mediators) [11].
Rationale
Therefore, it is hypothesized that the formation of drug-resistant tumor cells occurs in at least two distinct stages, namely physiological drug resistance and pathological drug resistance. Physiological drug resistance denotes the stage during which the cells are afflicted with a variety of cellular stress signals and become more susceptible to the type of damage likely to be inflicted by chemotherapeutic drugs. Physiological drug resistance is characterized by uncontrolled proliferation; impairment of apoptosis; ability to repair DNA damage; and increasingly lower sensitivity to chemotherapeutic drugs and/or radiation. Thus, cells in this group are distinct from normal cells. Subsequent to chemotherapy, depending on its effectiveness, multi drug resistance tumor cells evolve within the tumor burden. After several courses of therapy, due to several contributing factors, the host is overwhelmed with predominantly multi drug-resistant cells (referred to as pathologically drug resistant tumor cells) which, in turn, confer unresponsiveness to chemotherapeutic agents and/or adjuvant therapeutic treatment, leading to mortality. Figures 2, 3 and 4 summarize the hypothesis advanced here; they depict the scheme of events following chemotherapeutic treatment and the significance of the two distinct tumor cell populations in leading to the failure of a therapeutic regimen.
Figure 2 Phase I: Evolution of physiologically drug resistant cells by "exclusion" from chemotherapy-sensitive tumor cells.
Figure 3 Phase II: Existence of multiple mechanism(s) in a tumor burden for efficient evolution of pathologically drug resistant cells and "toleration" of coexisting physiologically drug resistant tumor cells.
Figure 4 Phase III: Progression of pathologically drug resistant tumor cells leading to mortality.
Hypothesis
Phase I. Exclusion
A fundamental characteristic of malignancy is genetic instability, which leads to a heterogeneous cell population endowed with altered morphologies, invasiveness, drug resistance and neovascularization, properties acquired through genetic mutation and aberrant gene expression. In addition, chemosensitivity is most likely caused by genetic changes; it is an acquired feature that appears at one specific stage and may be lost as the tumor burden progresses. It is hypothesized that tumor cells progress through two distinct stages, namely 'physiological drug resistance' and 'pathological drug resistance'. Subsequent to chemotherapy, the tumor cells are in a state in which drug sensitivity is similar to that in the corresponding normal cells. Within the tumor burden, depending on their drug sensitivity, a more limited number of tumor cells than corresponding normal cells are damaged by chemotherapeutic agents.
As shown in Figure 2, following chemotherapy, the cells are subjected to oxidative stress, which has been shown to induce and sustain genomic and mitochondrial DNA damage. Notwithstanding the mechanisms for repairing such DNA damage, the loss of genomic DNA integrity, entailing the down-regulation of apoptotic suppressor proteins, drug-resistance suppressor proteins or cell cycle regulatory proteins, obviously programs the cells to undergo cell cycle arrest, culminating in apoptosis.
Tumor cells capable of repairing DNA damage but retaining the upregulation of cell cycle regulatory proteins and drug-resistance proteins are likely to survive both oxidative stress and apoptosis [12,13]. It is speculated that this defined sub-population is likely to represent the initially drug-resistant cells, where the phenomenon is referred as "physiological drug resistance".
It is also possible that a select population of tumor cells may retain drug resistance and increase the expression of pro-apoptotic genes while losing the cell cycle regulatory protein(s). Such cells would be likely to undergo apoptosis. Thus, the evolving tumor would most likely contain a defined population of cells tolerant of oxidative stress and toxic drug effects, and also capable of bypassing apoptosis. Therefore, this initial phase is an exclusion phase during which tumor cells that have lost cell cycle control and tumor suppressor proteins are excluded from the rest of the tumor burden as an evolving population endowed with physiological drug resistance.
However, this selection process does not ensure that the entire physiologically drug resistant fraction of the tumor burden has become resistant to all drugs (Figure 2). To support my argument that two distinct form of tumor cells progress and lead to drug resistance mediated failure of therapy, I invoke the process of "apomixis" [14] to erect a hypothesis that depends on the presence of two hypothetical types of cell in the same somatic tissue (tumor burden) and a potential exchange of genetic material between them. "Apomixis" is a process that accounts for asexualreproduction in higher forms of life where sexual reproduction is the norm. In essence, the successful evolution of pathologically drug-resistant cells may or may not follow a pattern. The foremost factors determining the success of this process are cellular genetic defects (mutations), and the amount of chemotherapy and/or radiation the patient will subsequently undergo. Together, these factors determine or influence either the physiologically or the pathologically drug resistant cells. In successive generations, resistant cells are likely to adopt at least one additional molecular mechanism for mounting an effective defense against adjuvant therapy, after the initial radiation or chemotherapy has failed, plausibly because of apomixis.
Phase II. Tolerance
Tumor cells that are not susceptible to the toxic effects of chemotherapeutic drugs and are tolerant of oxidative stress are expected to possess one or more molecular mechanisms to protect and maintain the proteins essential for survival. Such tumor cell populations show physiological drug resistance. The sustained cytotoxic effects of chemotherapeutic drugs would be likely neither to induce genomic instability nor to affect cell cycle progression in such a population. Also, insufficient exposure to the therapy would leave a defined fraction of the tumor cells to re-grow with the property of physiological drug resistance and remain in the tumor burden. In addition to these physiological possibilities, it seems reasonable to propose that expression of several cell cycle regulatory proteins will be lost and that cell cycle arrest will become dissociated from DNA damage. In turn, accumulated DNA damage and uncontrolled cell cycle progression with impaired apoptotic pathways will confer increased resistance to chemotherapeutic drugs and/or adjuvant therapeutic treatment, protecting against cell death and sustaining tumor cell proliferation.
Therefore, this subpopulation of tumor cells is selected to progress towards pathological drug resistance with or without a specific escape mechanism for multiple drugs (e.g. increased glutathione levels and altered DNA repair, loss of cell cycle check point kinases). Also, it suggested that pathological drug resistance probably represents a stage during which resistance to various cytotoxic insults increases markedly. It is also speculated that these distinct tumor cell populations would probably coexist in a tumor burden by "tolerating" each other (Figure 3).
A mechanistic working hypothesis is presented in Figure 3, in which physiological drug resistance progresses to increased insensitivity to multiple drugs and subsequently to multi drug resistance. With subsequent tumor progression, several defense mechanisms may be lost, including dissociation of cell cycle arrest from DNA damage. Further, clonal progression of the 'pathologically drug resistant' cells may emerge as a result of drug-specific escape mechanisms and the impairment of both triggering and effectors mechanisms of apoptosis. Essentially, failure of all options in a clinical chemotherapeutic regimen produces a shift from physiological drug resistance to pathological drug resistance. Advanced clinical stages, representing the failure of multiple episodes of a therapeutic regimen, would be more likely to contain pathologically drug-resistant cells that are beyond the chemosensitive window.
Overall, it is still a strong possibility that both physiologically and pathologically drug-resistant cells, together with yet unknown drug-sensitive cells, might coexist in a tumor burden, complicating any viable alternative approach to therapy. The promise of any therapeutic measure at this point would largely depend on the properties of the predominantly surviving cell population in the tumor burden. Regrettably, this would mean that the cancer patient is losing ground in therapy and reaching the clinical endpoint, which is not yet conclusive at this point of the treatment.
Do the pathologically drug resistant tumor cells secrete some unidentified factor(s) or adopt a novel mechanism(s) to transform physiologically drug resistant tumor cells to pathologically drug resistant ones? It has been suggested that such a "phase transition" is a strong possibility [15].
Phase III. Progression
As shown in Figure 4, should the pathologically drug resistant tumor cells overwhelm the tumor burden, with or without the concomitant presence of physiologically drug resistant cells, it is likely that the relapse may lead towards complete failure of any remaining therapy. Relapses in cases of metastatic tumor burdens cause deterioration of the clinical scenario; metastatic tumors are more aggressive, in particular with a pathologically drug-resistant tumor burden. Progression of pathologically drug-resistant cells would most likely occur because of clonal dominance under the selection pressure imparted by the chemotherapy. In summary, a tumor burden that already contains more pathologically resistant cells would make the most intensive therapeutic regimen a futile exercise. since the pathologically resistant tumor cells would be insensitive as well as resistant to all forms of therapy. At this point it is reasonable to conclude that the cancer patient has reached the end point, meaning mortality.
Conclusion
In this hypothesis, I have considered the multiple mechanism(s) of selection and proliferation in a distinct tumor cell population, namely pathologically drug resistant tumor cells, in the tumor burden, leading to the total failure of chemotherapy or an adjuvant therapeutic regimen. Furthermore, the following four characteristic properties of tumor cells may determine the pattern of drug resistance: 1. Absence of contact inhibition/uncontrolled proliferation; 2. Absence of apoptotic/necrotic mechanisms; 3. Multifactorial (epigenetic) up-regulation of drug resistance genes; 4. Sustained oxidative stress-mediated dysregulation of metabolic pathways. All the aforesaid factors would be likely to play pivotal roles in a developmental stage-specific manner, but not all at once. Delineating the specific molecular determinants conferring physiological versus pathological drug resistance genotype/phenotypes would be essential for providing an effective measure to attenuate the impact of multi drug resistance and clinical failure of the current therapeutic regimen.
Author's contributions
SK generated the hypothetical scheme and formulated the hypothesis after a careful review of the appropriate literature.
The author does not have any competing financial or intellectual property interests with any party or association with members of the pharmaceutical manufacturing industry. The author declares that nearly all of the sequences, and the methods for generating the diadenosine polyphosphate hydrolases (Hint, Fhit, and GalT) (Biochemistry. 2002 41(29):9003-14.) protein family, purine-/pyrimidine-receptor sequences and the ectonucleotidase protein family, are in the public domain as of April 25, 2005. Therefore, the modified reagents and/or probe(s) originating from published sequences of the aforesaid protein family in the context of the scheme proposed in this article are part of one or more manufacturer(s) which are part of the impending U.S. or International patent application(s).
Acknowledgements
J. Crater, D. Liebenthal and S. Boldogh of University of Texas Medical Branch, Galveston, P.J. Hastings and S. Rosenberg of Baylor College of Medicine, Houston, L. Panasci of National Jewish General Hospital, McGill University, Montreal, Quebec, S. Lees-Miller of University of Calgary, Canada and C.W. Anderson of Brookhaven National Laboratory, New York are gratefully acknowledged for their kind help in preparation of this manuscript. The draft version of this article was prepared during tenure of a post doctoral post at the University of Pennsylvania, supported by National Institutes of Health Grants (DK52216 and DK44237), and also at the University of Texas Medical Branch, supported in part by National Institutes of Health (CA84461) UTMB, Galveston, Texas. I also thank Drs. Robert H. Diamond and L. Grenbaum, of the University of Pennsylvania for their support during preparation of the first draft version.
==== Refs
Hande KR Lee GR, Foerster J, Lukens J, Paraskevas F, Greer JP, Rodgers GM Principles and Pharmacology of Chemotherapy Wintrobe's Clinical Hematology 1999 2 10 Baltimore, Maryland 21201-2436 USA: Williams & Wilkins 2076 2101
Jodrell DI Egorin MJ Canetta RM Langenberg P Goldbloom EP Burroughs JN Goodlow JL Tan S Wiltshaw E Relationships between carboplatin exposure and tumor response and toxicity in patients with ovarian cancer J Clin Oncol 1992 10 520 528 1548516
Frei E 3rdCanellos GP Dose: a critical factor in cancer chemotherapy Am J Med 1980 69 585 94 6999898 10.1016/0002-9343(80)90472-6
Hryniak WM DeVita VT, Hellman S, Rosenberg SA The importance of dose intensity in the outcome of chemotherapy Important advances in Oncology 1988 Philadelphia: JB Lippincott 121 142 3042601
Goldie JH Scientific basis of adjuvant and primary (neoadjuvant) chemotherapy Semin Oncol 1997 14 1 7 3823915
Hesketh PJ Kris MG Grunberg SM Beck T Hainsworth JD Harker G Aapro MS Gandara D Lindley CM Proposal for classifying the acute emetogenicity of cancer chemotherapy J Clin Oncol 1997 15 103 109 8996130
Gottesman MM Fojo T Bates SE Multidrug resistance in cancer: role of ATP-dependent transporters Nat Rev Cancer 2002 2 48 58 11902585 10.1038/nrc706
Gottesman MM Mechanisms of cancer drug resistance Annu Rev Med 2002 53 615 27 11818492 10.1146/annurev.med.53.082901.103929
McKenna SL Padua RA Multidrug resistance in leukemia Br J Haematol 1997 96 659 674 9074405 10.1046/j.1365-2141.1997.d01-2095.x
Teicher BA Physiologic mechanisms of therapeutic resistance. Blood flow and hypoxia Hematol Oncol Clin North Am 1995 9 475 506 7642474
Bernstein C Bernstein H Payne CM Garewal H DNA repair/pro-apoptotic dual-role proteins in five major DNA repair pathways: fail-safe protection against carcinogenesis Mutat Res 2002 511 145 178 12052432
Pusztai L Siddik ZH Mills GB Bast RC Jr Physiologic and pathologic drug resistance in ovarian carcinoma- a hypothesis based on a clonal progression model Acta Oncol 1998 37 629 640 10050979 10.1080/028418698429964
Tannock IF Tumor physiology and drug resistance Cancer Metastasis Rev 2001 20 123 132 11831640 10.1023/A:1013125027697
Stern C Sherwood ER The Origin of Genetics – a Mendel Source Book 1966 San Francisco and New York: W.H. Freeman and Company
Kannan S Molecular basis of the evolution of drug resistance: potential role of the transient state during infection/drug treatment Med Hypotheses 2004 63 71 72 15193350 10.1016/j.mehy.2003.04.001
| 15857507 | PMC1131929 | CC BY | 2021-01-04 16:39:25 | no | Theor Biol Med Model. 2005 Apr 27; 2:17 | utf-8 | Theor Biol Med Model | 2,005 | 10.1186/1742-4682-2-17 | oa_comm |
==== Front
J Transl MedJournal of Translational Medicine1479-5876BioMed Central London 1479-5876-3-171586013310.1186/1479-5876-3-17ReviewAdoptive T cell therapy: Addressing challenges in cancer immunotherapy Yee Cassian [email protected] Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N., D3-100, Seattle, WA 98109, USA2005 28 4 2005 3 17 17 13 1 2005 28 4 2005 Copyright © 2005 Yee; licensee BioMed Central Ltd.2005Yee; 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.
Adoptive T cell therapy involves the ex vivo selection and expansion of effector cells for the treatment of patients with cancer. In this review, the advantages and limitations of using antigen-specific T cells are discussed in counterpoint to vaccine strategies. Although vaccination strategies represent more readily available reagents, adoptive T cell therapy provides highly selected T cells of defined phenotype, specificity and function that may influence their biological behavior in vivo. Adoptive T cell therapy offers not only translational opportunities but also a means to address fundamental issues in the evolving field of cancer immunotherapy.
==== Body
Introduction
Augmentation of the immune response can be achieved through in vivo vaccination or ex vivo expansion of antigen-specific effectors followed by adoptive transfer. Both modalities share many features. For example, the antigen-presenting cell used for stimulating effector responses in vivo and in vitro represents a crucial element responsible for shaping the specificity and phenotype of the intended immune response. Therefore, preclinical studies that advance the engineering of robust antigen-presenting cells may be translated for use with either strategy. The cytokines necessary for augmentation and maintenance of the immune effector function and survival, the costimulatory factors required, and the regulatory and inhibitory mechanisms that must be overcome to achieve tumor eradication must be addressed whether vaccine strategies or adoptive T cell therapy is used. However, the behavior and ultimate fate of effectors generated in vivo can be substantially different from those generated in vitro. It would be naïve to assume that in vivo conditions could be reproduced completely by manipulating conditions in vitro and there may be effectors of desired phenotype and function that can only be generated or more easily generated in vivo than in vitro. On the other hand, when effectors can be generated in vitro, their specificity, magnitude, surface and functional phenotype can be far better defined than those generated following in vivo immunization. For this reason, the appeal of adoptive therapy is that the reasons for success or failure of a given strategy can be determined with greater precision than with in vivo vaccination. As more comprehensive and sensitive tools become available to monitor the immune response [1,2], this advantage may diminish; however it would be presumptuous to believe that immune monitoring can characterize induced vaccine-elicited effectors to the same degree as effectors elicited ex vivo. Although there is no guarantee that infused T cells will behave in vivo in the same manner as one would be able to predict in vitro, effector cells can be manipulated and selected ex vivo, prior to adoptive transfer, in a manner that can answer questions that cannot be addressed by vaccination strategies. When it is possible to generate an effector population of T cells of defined magnitude, the temptation however, is to ignore the role of qualitative differences by adopting a 'more is better' policy. The avidity, functional phenotype and in vivo 'survivability' are equally if not more relevant in mediating tumor eradication than numbers alone. With this in mind, the following commentary provides first a description of adoptive therapy strategies and then a point-by-point discussion of its various features and advantages in addressing challenges in immunotherapy.
Defining Adoptive Cellular Therapy
Adoptive therapy involves the transfer of ex vivo expanded effector cells as a means of augmenting the anti-tumor immune response. Depending on the method of ex vivo selection, stimulation and expansion, varying degrees of uniformity with respect to antigen-specificity and phenotype may be obtained. This can range from a diverse polyclonal population of effector cells to highly selected T cell clones of defined phenotype, specificity and tumor avidity. The following broad and somewhat arbitrary categories describing T cell expansion methods are listed in order of increasing antigen specificity:
1. Non-specific expansion of peripheral blood lymphocytes. Non-specific ex vivo expansion of peripheral blood T cells by triggering the T cell receptor and costimulatory molecules with antibodies and/or the use of cytokines to drive T cells have been used in a number of clinical studies for the treatment of patients with HIV and malignant diseases [3-6]. In spite of the absence of a specific in vitro stimulator, in vitro studies suggest that augmentation of existing antigen-specific immunity can be achieved.
2. Ex vivo expansion of Tumor infiltrating lymphocytes. Expansion of infiltrating lymphocytes harvested from tumor sites yields a polyclonal population of T cells with broad reactivity to a variety of autologous tumor antigens. Although some degree of tumor-reactivity can be achieved, there is little control over the specificity or phenotype of the effector population without further in vitro manipulation [7]
3. In vitro stimulation to elicit antigen-specific T cells from peripheral blood lymphocytes using cells engineered for antigen presentation (peptide pulsed, RNA transfected or viral transduced autologous stimulator cells or artificial antigen presenting cells [8,9]) provides the most precise control over the specificity, and phenotype of the intended immune response [10,11]. Greater uniformity of effector specificity and phenotype may be achieved using clonal T cells expanded ex vivo.
Although this discussion is limited to the treatment of solid tumor malignancies, it should be noted that the adoptive transfer of allogeneic effector cells including donor lymphocyte infusions [12], minor antigen-specific CTL [13] and strategies that exploit NK alloreactivity [14] have been used successfully for the treatment of leukemia following allogeneic stem cell transplant. In addition, more refined approaches for the treatment of patients with post-transplant lymphoproliferative disease or Hodgkin's disease using EBV-specific T cells [15-17] have led to durable complete responses that have yet to be achieved with any notable degree in solid tumor immunotherapy. These studies are instructive for the solid tumor immunotherapist in that they demonstrate the feasibility of targeting specific tumor-associated viral antigens and provide insight into the role that T cells can play in eradicating life-threatening disease. Since solid tumors, such as melanoma, in its advanced stages exhibit physical and immunologic barriers distinct from tumors of hematologic origin, the following is limited to a discussion of immunotherapeutic strategies for the treatment of solid tumors.
Different strategies in generating effector cells for adoptive therapy influence not only the antigen-specificity, tumor avidity and cellular phenotype, but also the behavior of T cells in vivo i.e., longevity, trafficking, anti-tumor efficacy. Although effector cells generated using the above protocols can be considered in the following discussion, as a counterpoint to vaccination strategies, it will be assumed that the prototypic effector cells for adoptive therapy will be ex vivo expanded antigen-specific T cells.
Antigen-specific Immunotherapy: Points to Consider
1. Magnitude and Persistence
The magnitude of the anti-tumor immune response has been demonstrated in murine models of immunotherapy to be a critical factor in tumor eradication [18]. Although the frequency of antigen-specific T cells required to mediate an anti-tumor effect in patients is not known and is likely to vary widely depending on the antigen target, tumor burden, stromal environment and many other factors, including, perhaps most importantly, qualitative features of the effector cell. However, it may be agreed that there is a threshold below which, it would be unreasonable to predict a response, especially against established tumors, but it is equally unlikely that that threshold need be as great as 90% of all circulating CD8 T cells (unless such large numbers of cells are required to compensate for a paucity of desired effectors or a general qualitative defect). Tumor-eradicating therapy in murine models suggest that a frequency of antigen-specific T cells of at least 1–10% of CD8 T cells is required. In patients, this translates to a dose of 2 to 20 × 109 cells. The use of non-specific expansion methods (cytokines, TCR and costimulatory molecule triggering) have been used to successfully expand unselected peripheral blood mononuclear cells to > 1010 cells in vitro over a period of 2–4 weeks [3,8]. In this case, however, the frequency of tumor-reactive T cells in the final infused product is often not known. Tumor-infiltrating lymphocytes after a 10–12 week period of in vitro culture with high-dose IL-2 yield 1010 - 1011 cells [7]; when expanded using anti-CD3 in combination with irradiated feeder cells, similar numbers can be achieved in less than half that time [19]. CD8 and CD4 T cell clones of defined antigen specificity and phenotype expand 500 to > 5000 fold over two weeks and can also achieve numbers > 1010 [11,20-22]. In vivo, T cell frequencies of up to 5 % of all CD8 T cells in an unmanipulated host can be achieved. Patients have received adoptively transferred antigen-specific T cells numbering > 1010 for a single infusion and often go on to receive multiple T cell infusions at intervals of days to weeks. In murine models, repeated infusions may play a role in 'attacking' the tumor in geographically distinct regions leading to tumor regression over time [23].
Equally important for successful therapy is the duration of in vivo persistence of transferred T cells which can vary from hours or days to weeks. This can depend in large part on the manner in which T cells were generated in vitro and under what conditions they are administered. Recent trials using adoptively transferred antigen-specific T cells are summarized in Table 1, Additional file 1 according to method of CTL generation and expansion, number of cells infused and immunomodulatory considerations such as prior lymphodepletion and dose of IL-2 administered. The methods for generating T cells varied with respect to the antigen presenting cell (Drosophila cells vs. autologous dendritic cells), in vitro dose of IL-2 ('T cell growth factor" to low-dose IL-2 at 10 U/ml to as much as 6000 U/ml), the number of cells infused and use of lymphodepleting regimens. In cases where polyclonal populations were used, as few as 108 antigen-specific CTL were infused; these T cells could not be detected in the peripheral blood (a Mitchell et al, Table 1, Additional file 1) [24]. The absence of detectable T cells may be attributed to the absence of requisite costimulatory signals not provided by the gene-modified insect cells, absence of co-administered IL-2, the relatively low cell dose and/or underestimation of the actual frequency due to the use of limiting dilution analysis instead of tetramer staining. Although up to ten-fold higher doses of antigen – specific CTL clones were administered in other studies (c Dudley et al) [25,26], these transferred cells also did not persist in vivo. In this case, T cells obtained from a previously vaccinated host failing peptide vaccine therapy, stimulated in vitro with the identical epitope and exposed to very high doses of IL-2 (c&d Dudley et al)[25,26] are likely to behave very differently from T cells generated from a non-vaccinated host under more physiologic conditions of cyclical antigen-stimulation and low-dose cytokines (b Meidenbauer et al; e Yee et al) [11,27,28]. In the former, adoptively transferred T cells experience a very short (< 48 hour) period of in vivo persistence possibly due to the requirement for supraphysiologic doses of IL-2 help in vivo and a starting population of T cells that may have reduced proliferative capacity due to prior in vivo vaccination. By contrast, T cells generated in vitro under more physiologic condition, can persist for more than 2 weeks in the presence of help (exogenous low-dose IL-2). Is this duration of in vivo persistence sufficient to mediate an anti-tumor response? While no clinical complete responses by RECIST (Response Evaluation Criteria In Solid Tumors) criteria [29] were noted in this study of patients with metastatic melanoma, what often escapes notice [30] is that patients experienced partial responses, significant tumor regression and stabilization of disease for an average of > 11 months and up to 29+ months – beyond what would be expected for patients with refractory disease following conventional therapy (median survival < 6 months) [11]. Since several of the patients who eventually progressed demonstrated evidence of outgrowth of antigen-loss tumor variants, it is suggested that such immunoselective pressure could not have occurred in the absence of effective antigen-specific immunotherapy [11].
One advantage of using ex vivo expanded T cells is that patients may have cells collected at an earlier stage of disease or prior to immunomodulatory therapy for later use. In one widely publicized study, TIL cells expanded using methods developed for clonal T cell expansion, were adoptively transferred to patients following lymphodepletion [19]. Melanoma-reactive T cells accumulated in the peripheral blood of these patients' reconstituting immune system to reach an astounding 97% of CD8+ T cells accompanied by tumor regression. Up to 50% of patients in an updated report demonstrated evidence of a clinical response and has been reported by its authors as a clear example of the role of adoptively transferred T cells, in the right setting to mediate dramatic clinical responses [31]. In corollary reports, the authors demonstrate that the level of T cell persistence observed in responding patients was significantly higher than that of nonresponding patients at early (5–15 days post infusion) and later timepoints (1–2 months post-infusion) suggesting that the duration of persistence of T cells derived from transferred, ex vivo expanded TIL play a role in the anti-tumor response [32]. Interestingly, although the infused product was comprised of several clonotypes, only a handful of T cell clones (among both responders and non-responders) persisted. This is an important study, but it leaves many questions unanswered and meaningful conclusions that advance the field are difficult to draw due to confounding variables in the design of this study. Without controlling for the specificity, dose or phenotype, it is unclear what type or dose of effector cells are required. Are CD4 T cells essential? What are the features of the subpopulation of T cells that experienced prolonged in vivo persistence and how do these differ from other T cells that were generated in vitro? Was their survival a result of greater (or lesser) avidity for their targets? Is it necessary to achieve the degree of lymphodepletion used in this study (which led to serious toxicities such as vision-threatening uveitis, and life-threatening PTLD)? Is selective depletion of regulatory cells or gentler preparative regimens to augment homeostatic mechanisms supportive of transferred T cells sufficient ? What role does high-dose IL-2 play in mediating clinical responses in this setting? Unfortunately, these answers cannot be divined from this study; rather, dissecting the contribution of these components to the anti-tumor response will be undertaken in carefully designed trials that exploit the advantages of using selectively expanded adoptively transferred T cells.
2. Phenotype
The effector cell phenotype can be described as either 1) surface markers that are associated with specific (and non-exclusive) effector function – for example, CD4+ T cells are more likely to provide cytokine help than CD8+ T cells which are more likely to be cytolytic; or 2), a functional phenotype, such as tumor cell killing or TCR affinity. Vaccination strategies may be directed towards the induction of CD4 or CD8 T cells on the basis of whether Class I or Class II-restricted epitopes are used or whether the method of antigen engineering or presentation favors Class I or Class II MHC loading. However, more precise selection of the intended phenotype can be achieved by in vitro selection or enrichment of CD4 or CD8 T cells by immunomagnetic bead selection for example. T cells generated following peptide vaccination may exhibit low avidity for tumor cells, possibly as a result of preferential expansion of lower affinity effectors by APC presenting supraphysiologic concentrations of peptide MHC [2,33]. The use of altered peptide ligands may be capable of inducing in vivo an effector population with greater affinity for the tumor targets [34], but such ligands have not been frequently described and T cells of defined affinity cannot be selected. Perhaps part of the reason for the disparity between T cell frequency and clinical response in earlier vaccination studies is that attempts at immune monitoring enumerated T cells regardless of avidity. The study using the altered peptide ligand of CEA is one of only a handful of vaccine studies that demonstrated correlation with clinical response [34]. By contrast, T cells generated ex vivo with altered or natural peptide ligands or any other tumor-derived APC can be selected on the basis of the affinity of their TCR, overall tumor avidity or any other measurable and selectable functional property, uniformly expanded and transferred at a desired T cell dose [35-37].
One property of T cells that may be gauged by surface expression of specific markers, is their proliferative capacity, an important feature that will no doubt receive greater attention in the design of clinical trials, and paradoxically, may be inversely correlated with more routine measures of cytolytic or effector capacity (Gattinoni L et al, ISBTC abstract, 2004). In this aspect, although clones demonstrate significant proliferative capacity under the right conditions (e.g. IL-15 [38]), early effectors clearly exhibit greater potential for durable in vivo persistence. Perhaps studies demonstrating significant tumor responses [19] are a result of the presence of some of these early effectors in a polyclonal infusate of T cells or the adoptive transfer of smaller numbers of carefully selected effectors on the basis of proliferative capacity rather than tumor killing. Whether such an approach will be more successful, remains to be seen but represents the type of question that could best be addressed by adoptive immunotherapy.
3. Specificity
As with vaccination strategies, the specificity of the intended immune response can be controlled and multivalent targeting can be achieved by adoptive therapy. However, eliciting responses by vaccination alone when the frequency of such responses is low as in the case of commonly shared tumor associated self antigens [39] or when such responses are represented by subdominant epitopes [40], may be limited by in vivo constraints. In this case ex vivo manipulation provides for a greater likelihood of generating T cells of desired specificity and magnitude by enriching for desired T cells in culture and careful selection of T cell clones. Alternatively, the ability to genetically modify T cells provides the opportunity to fashion T cells of defined specificity for adoptive therapy. T cell receptors cloned and sequenced from tumor-reactive lines and efficiently transferred into peripheral blood lymphocytes [41]. TCR-modified lymphocytes selected on the basis of in vitro markers or enriched by iterative stimulation have demonstrated the capacity to recognize and kill specific tumor target cells. In this way, patients for whom T cells of a given specificity are poorly represented by their immune repertoire, are not precluded from adoptive therapy when TCR-modified autologous lymphocytes can be used. Furthermore, mutant TCR can be designed to enhance affinity for the target MHC complex thus endowing genetically modified T cells with greater tumor avidity [42]. Finally, a chimeric T cell receptor comprised of an extracellular antibody binding a surface tumor target antigen coupled with intracellular signaling sequence (e.g. TCR-zeta) can also be used to endow peripheral blood lymphocytes with novel specificity [43,44].
4. Genetic modification as a safeguard mechanism, to facilitate tracking and to enhance function
Other advantages associated with the use of genetically modified T cells for adoptive therapy over vaccination strategies are the capacity to eradicate T cells in vivo through the use of drug-inducible 'suicide' genes and to track these cells using genetic markers [45,46]. Although T cells transduced with early generations of the inducible HSV thymidine kinase gene were effectively eliminated in vivo following ganciclovir administration, they also suffered from early peripheral clearance due to the induction of an endogenous anti – HSV-TK response[47]. Later generations of suicide genes utilizing Fas-Fas dimerization technology address this problem and are being evaluated in pre-clinical studies [48]. T cell tracking with a unique genetic marker, such as a resistance gene for example permits unequivocal evaluation of T cell frequency, and localization if a feasible biopsy can be obtained from tumor or lymph node sites. Such cells can be analyzed using fluorescent-tagged riboprobes corresponding to the unique transgene or quantitative real-time PCR [45]. Dynamic T cell tracking in vivo, without the requirement for serial biopsies may be achieved using TK-transduced T cells that are designed to preferentially sequester radiolabelled substrate and can then be analyzed by PET imaging [49].
Genetic modification may also be used to enhance T cell function, for example, by conferring a helper-independent phenotype to antigen-specific CD8+ T cells with the use of a chimeric IL-2 receptor [50] or restoration of CD28 expression [51] enabling antigen-driven autocrine proliferation.
5. Immune escape
Because T cells are isolated and expanded ex vivo, the clinical and immune state of the patient does not necessarily affect the ability to augment an immune response. Patients whose immune system may be crippled by tumor-suppressive factors or several rounds of chemotherapy and radiation may not be able to mount a robust immune response following vaccination. In these patients, ex vivo manipulation provides a means of isolating tumor-reactive T cells and expanding such cells for adoptive therapy. The mechanisms responsible for inhibiting an afferent response may limit the capacity of vaccines to generate functional T cells of sufficient magnitude. Ex vivo manipulation of T cells following exposure to immunomodulatory cytokines or selective depletion of regulatory cells (e.g. CD4, CD25+ T cells) may facilitate the isolation and expansion of tumor-reactive T cells for adoptive therapy. While it is possible to delete regulatory cells in vivo or co-administer immunomodulatory cytokines to augment a functional vaccine-elicited response [52,53], such strategies represent greater regulatory hurdles and can lead to unwanted effects. For example, the use of IL-2 to expand effector cells during the afferent phase of vaccine therapy can also lead to expansion of regulatory T cells in vivo [54] and depletion of regulatory T cells using anti-CD25 antibody can lead to the elimination of potentially beneficial CD25+ activated T cells.
6. Feasibility considerations
The issue addressed here is the potential for adoptive T cell therapy to become a clinically significant modality that participates in the standard treatment of patients with malignant disease. The isolation and expansion of antigen-specific T cells is time and labor intensive, requires infrastructure support to cultivate and qualify T cell products and can be prohibitively expensive in its current experimental phase. Isolation and expansion of T cells for adoptive therapy can take 4–16 weeks and for patients with progressive disease, this may not be feasible. However, adoptive therapy can achieve T cell frequencies that are equal and often greater than that possible over the same period of time for patients receiving vaccines since in vivo expansion may also require several weeks and repeated boost administrations. Unlike vaccine reagents, T cell products cannot be manufactured and distributed easily; cryopreservation, storage, transport and reconstitution / thawing are problematic with a cell product. Vaccines can be made readily available in some forms to many institutions thus facilitating recruitment for the large-scale Phase II and Phase III studies needed to demonstrate efficacy and superiority over conventional modalities. In many ways, vaccination strategies have and are likely to gain regulatory approval more readily. Limitations to large-scale production of antigen-specific effectors ex vivo are being addressed. Most of these advances are related to adaptation of closed bag systems to eliminate the labor and inefficiencies of handling large numbers of cells, and to the design of artificial antigen-presenting cells to eliminate variability and reduce quality control concerns associated with in vitro cultured autologous APCs [9]. Advances in the isolation of antigen-specific T cells by cell sorting or immunomagnetic bead selection that can expedite the process, specialized reagents and culture vessels that facilitate expansion and storage and quality control measures that ensure product fidelity are currently being developed and will decrease many of these cost-related, and logistical issues.
Conclusion
Our current understanding of the requirements for successful T cell-based therapy in the treatment of patients with solid tumors remains largely undeveloped Advances in this field will require judicious, step-wise translation of promising pre-clinical strategies into carefully designed clinical trials with discrete immunologic endpoints. This would be well-served by immunologic monitoring that encompasses not only a characterization of the biologic behavior of adoptively transferred or vaccine-elicited T cells in vivo but also a comprehensive analysis of immune escape mechanisms, especially those that develop with more and more effective strategies. Rather than claim clinical superiority of one modality over that of another on the basis of one or two early phase studies, it would be more instructive to exploit the individual advantages of vaccine or adoptive T cell therapy in designing clinical trials. Vaccine reagents can be easily produced and made readily available for widespread administration. This is particularly advantageous for later phase studies and multivalent approaches (especially where the antigen specificity is not known). Effectors may be more easily elicited in vivo; however, control over their desired features is less and the the burden for sophisticated immunologic monitoring much greater for vaccine strategies, where identification of T cells in vivo exhibiting such properties and correlating their presence with antitumor activity will be crucial. The implementation of adoptive therapy however belies its experimentalistic origins: in cases where a population of T cells of desired magnitude with defined phenotypic and functional properties is required, for example, to validate findings arising from vaccine studies or provide proof of principle for hypotheses based on pre-clinical studies, this represents the optimal strategy. In addition to extending exploratory research, there are translational opportunities afforded to the ability to manipulate effectors ex vivo that are otherwise not available to vaccines. In the end however, the difference in these two modalities can be considered largely arbitrary and there are complementary if not synergistic strategies utilizing both vaccination and adoptive T cell therapy [55] that will be essential for addressing challenges in cancer immunotherapy.
Acknowledgements
CY is a Damon Runyon-Lilly clinical investigator supported (in part) by the Damon Runyon Cancer Research Foundation. This work was also supported by funding from R01 CA 104711, R21 CA 94500 and The Fialkav Award.
Supplementary Material
Additional File 1
Recent trials using adoptively transferred antigen-specific T cells are summarized according to method of CTL generation and expansion, number of cells infused and immunomodulatory considerations such as prior lymphodepletion and dose of IL-2 administered.
Click here for file
==== Refs
Yee C Greenberg P Modulating T-cell immunity to tumours: new strategies for monitoring T-cell responses Nature Reviews Cancer 2002 2 409 419 12189383 10.1038/nrc820
Stuge TB Holmes SP Saharan S Tuettenberg A Roederer M Weber JS Lee PP Diversity and Recognition Efficiency of T Cell Responses to Cancer Plos Med 2004 1 e28 15578105 10.1371/journal.pmed.0010028
Laport GG Levine BL Stadtmauer EA Schuster SJ Luger SM Grupp S Bunin N Strobl FJ Cotte J Zheng Z Adoptive transfer of costimulated T cells induces lymphocytosis in patients with relapsed/refractory non-Hodgkin lymphoma following CD34+-selected hematopoietic cell transplantation Blood 2003 102 2004 2013 12763934 10.1182/blood-2003-01-0095
Rapoport AP Levine BL Badros A Meisenberg B Ruehle K Nandi A Rollins S Natt S Ratterree B Westphal S Molecular remission of CML after autotransplantation followed by adoptive transfer of costimulated autologous T cells Bone Marrow Transplant 2004 33 53 60 14578928 10.1038/sj.bmt.1704317
Grimm EA Mazumder A Zhang HZ Rosenberg SA Lymphokine-activated killer cell phenomenon. Lysis of natural killer-resistant fresh solid tumor cells by interleukin 2-activated autologous human peripheral blood lymphocytes Journal of Experimental Medicine 1982 155 1823 1841 6176669 10.1084/jem.155.6.1823
Thompson JA Figlin RA Sifri-Steele C Berenson RJ Frohlich MW A phase I trial of CD3/CD28-activated T cells (Xcellerated T cells) and interleukin-2 in patients with metastatic renal cell carcinoma Clin Cancer Res 2003 9 3562 3570 14506142
Topalian SL Muul LM Solomon D Rosenberg SA Expansion of human tumor infiltrating lymphocytes for use in immunotherapy trials Journal of Immunological Methods 1987 102 127 141 3305708
Maus MV Thomas AK Leonard DGB Allman D Addya K Schlienger K Riley JL June CH Ex vivo expansion of polyclonal and antigen-specific cytotoxic T lymphocytes by artificial APCs expressing ligands for the T-cell receptor, CD28 and 4-1BB Nature Biotechnology 2002 20 143 148 11821859 10.1038/nbt0202-143
Oelke M Maus MV Didiano D June CH Mackensen A Schneck JP Ex vivo induction and expansion of antigen-specific cytotoxic T cells by HLA-Ig-coated artificial antigen-presenting cells Nat Med 2003 9 619 624 12704385 10.1038/nm869
Yee C Gilbert MJ Riddell SR Brichard VG Fefer A Thompson JA Boon T Greenberg PD Isolation of tyrosinase-specific CD8+ and CD4+ T cell clones from the peripheral blood of melanoma patients following in vitro stimulation with recombinant vaccinia virus J Immunol 1996 157 4079 4086 8892642
Yee C Thompson JA Byrd D Riddell SR Roche P Celis E Greenberg PD Adoptive T cell therapy using antigen-specific CD8+ T cell clones for the treatment of patients with metastatic melanoma: In vivo persistence, migration, and antitumor effect of transferred T cells Proc Natl Acad Sci U S A 2002 99 16168 16173 12427970 10.1073/pnas.242600099
Kolb HJ Holler E Adoptive immunotherapy with donor lymphocyte transfusions Curr Opin Oncol 1997 9 139 145 9161791
Hambach L Goulmy E Immunotherapy of cancer through targeting of minor histocompatibility antigens Curr Opin Immunol 2005 17 202 210 15766682 10.1016/j.coi.2005.01.010
Ruggeri L Capanni M Urbani E Perruccio K Shlomchik WD Tosti A Posati S Rogaia D Frassoni F Aversa F Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic transplants Science 2002 295 2097 2100 11896281 10.1126/science.1068440
Bollard CM Aguilar L Straathof KC Gahn B Huls MH Rousseau A Sixbey J Gresik MV Carrum G Hudson M Cytotoxic T Lymphocyte Therapy for Epstein-Barr Virus+ Hodgkin's Disease J Exp Med 2004 200 1623 1633 15611290 10.1084/jem.20040890
Rooney CM Smith CA Ng CY Loftin SK Sixbey JW Gan Y Srivastava DK Bowman LC Krance RA Brenner MK Heslop HE Infusion of cytotoxic T cells for the prevention and treatment of Epstein-Barr virus-induced lymphoma in allogeneic transplant recipients Blood 1998 92 1549 1555 9716582
Khanna R Bell S Sherritt M Galbraith A Burrows SR Rafter L Clarke B Slaughter R Falk MC Douglass J Activation and adoptive transfer of Epstein-Barr virus-specific cytotoxic T cells in solid organ transplant patients with posttransplant lymphoproliferative disease Proc Natl Acad Sci U S A 1999 96 10391 10396 10468618 10.1073/pnas.96.18.10391
Greenberg PD Adoptive T cell therapy of tumors: mechanisms operative in the recognition and elimination of tumor cells Advances in Immunology 1991 49 281 355 1853786
Dudley ME Wunderlich JR Robbins PF Yang JC Hwu P Schwartzentruber DJ Topalian SL Sherry R Restifo NP Hubicki AM Cancer regression and autoimmunity in patients after clonal repopulation with antitumor lymphocytes Science 2002 298 850 854 12242449 10.1126/science.1076514
Ho WY Blattman JN Dossett ML Yee C Greenberg PD Adoptive immunotherapy: engineering T cell responses as biologic weapons for tumor mass destruction Cancer Cell 2003 3 431 437 12781360 10.1016/S1535-6108(03)00113-2
Riddell SR Greenberg PD The use of anti-CD3 and anti-CD28 monoclonal antibodies to clone and expand human antigen-specific T cells Journal of Immunological Methods 1990 128 189 201 1691237 10.1016/0022-1759(90)90210-M
Walter EA Greenberg PD Gilbert MJ Finch RJ Watanabe KS Thomas ED Riddell SR Reconstitution of cellular immunity against cytomegalovirus in recipients of allogeneic bone marrow by transfer of T-cell clones from the donor [see comments] New England Journal of Medicine 1995 333 1038 1044 7675046 10.1056/NEJM199510193331603
Kircher MF Allport JR Graves EE Love V Josephson L Lichtman AH Weissleder R In vivo high resolution three-dimensional imaging of antigen-specific cytotoxic T-lymphocyte trafficking to tumors Cancer Res 2003 63 6838 6846 14583481
Mitchell MS Darrah D Yeung D Halpern S Wallace A Voland J Jones V Kan-Mitchell J Phase I trial of adoptive immunotherapy with cytolytic T lymphocytes immunized against a tyrosinase epitope J Clin Oncol 2002 20 1075 1086 11844833 10.1200/JCO.20.4.1075
Dudley ME Adoptive transfer of cloned melanoma-reactive T lymphocytes for the treatment of patients with metastatic melanoma Journal of Immunotherapy 2001 24 363 373 11565838 10.1097/00002371-200107000-00012
Dudley ME Wunderlich JR Yang JC Hwu P Schwartzentruber DJ Topalian SL Sherry RM Marincola FM Leitman SF Seipp CA A phase I study of nonmyeloablative chemotherapy and adoptive transfer of autologous tumor antigen-specific T lymphocytes in patients with metastatic melanoma Journal of Immunotherapy 2002 25 243 251 12000866 10.1097/00002371-200205000-00007
Meidenbauer N Marienhagen J Laumer M Vogl S Heymann J Andreesen R Mackensen A Survival and tumor localization of adoptively transferred Melan-A-specific T cells in melanoma patients J Immunol 2003 170 2161 2169 12574389
Yee C Thompson JA Roche P Byrd DR Lee PP Piepkorn M Kenyon K Davis MM Riddell SR Greenberg PD Melanocyte destruction after antigen-specific immunotherapy of melanoma: direct evidence of t cell-mediated vitiligo J Exp Med 2000 192 1637 1644 11104805 10.1084/jem.192.11.1637
Therasse P Arbuck SG Eisenhauer EA Wanders J Kaplan RS Rubinstein L Verweij J Van Glabbeke M van Oosterom AT Christian MC Gwyther SG New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada J Natl Cancer Inst 2000 92 205 216 10655437 10.1093/jnci/92.3.205
Dudley ME Rosenberg SA Adoptive-cell-transfer therapy for the treatment of patients with cancer Nat Rev Cancer 2003 3 666 675 12951585 10.1038/nrc1167
Rosenberg SA Yang JC Restifo NP Cancer immunotherapy: moving beyond current vaccines Nat Med 2004 10 909 915 15340416 10.1038/nm1100
Robbins PF Dudley ME Wunderlich J El-Gamil M Li YF Zhou J Huang J Powell DJ JrRosenberg SA Cutting edge: Persistence of transferred lymphocyte clonotypes correlates with cancer regression in patients receiving cell transfer therapy J Immunol 2004 173 7125 7130 15585832
Alexander-Miller MA Leggatt GR Berzofsky JA Selective expansion of high- or low-avidity cytotoxic T lymphocytes and efficacy for adoptive immunotherapy Proceedings of the National Academy of Sciences of the United States of America 1996 93 4102 4107 8633023 10.1073/pnas.93.9.4102
Fong L Hou Y Rivas A Benike C Yuen A Fisher GA Davis MM Engleman EG Altered peptide ligand vaccination with Flt3 ligand expanded dendritic cells for tumor immunotherapy Proceedings of the National Academy of Sciences of the United States of America 2001 98 8809 8814 11427731 10.1073/pnas.141226398
Yee C Savage PA Lee PP Davis MM Greenberg PD Isolation of high avidity melanoma-reactive CTL from heterogeneous populations using peptide-MHC tetramers Journal of Immunology 1999 162 2227 2234
Becker C Pohla H Frankenberger B Schuler T Assenmacher M Schendel DJ Blankenstein T Adoptive tumor therapy with T lymphocytes enriched through an IFN-gamma capture assay Nat Med 2001 7 1159 1162 11590442 10.1038/nm1001-1159
Betts MR Brenchley JM Price DA De Rosa SC Douek DC Roederer M Koup RA Sensitive and viable identification of antigen-specific CD8+ T cells by a flow cytometric assay for degranulation J Immunol Methods 2003 281 65 78 14580882 10.1016/S0022-1759(03)00265-5
Lu J Giuntoli RL 2ndOmiya R Kobayashi H Kennedy R Celis E Interleukin 15 promotes antigen-independent in vitro expansion and long-term survival of antitumor cytotoxic T lymphocytes Clin Cancer Res 2002 8 3877 3884 12473603
Houghton AN Cancer antigens: immune recognition of self and altered self [comment] Journal of Experimental Medicine 1994 180 1 4 8006576 10.1084/jem.180.1.1
Makki A Weidt G Blachere NE Lefrancois L Srivastava PK Immunization against a dominant tumor antigen abrogates immunogenicity of the tumor Cancer Immun 2002 2 4 12747749
Clay TM Custer MC Sachs J Hwu P Rosenberg SA Nishimura MI Efficient transfer of a tumor antigen-reactive TCR to human peripheral blood lymphocytes confers anti-tumor reactivity J Immunol 1999 163 507 513 10384155
Holler PD Chlewicki LK Kranz DM TCRs with high affinity for foreign pMHC show self-reactivity Nat Immunol 2003 4 55 62 12469116 10.1038/ni863
Friedmann-Morvinski D Bendavid A Waks T Schindler D Eshhar Z Redirected primary T-cells harboring a chimeric receptor require co-stimulation for their antigen-specific activation Blood 2004 15626734
Wang J Press OW Lindgren CG Greenberg P Riddell S Qian X Laugen C Raubitschek A Forman SJ Jensen MC Cellular immunotherapy for follicular lymphoma using genetically modified CD20-specific CD8+ cytotoxic T lymphocytes Mol Ther 2004 9 577 586 15093188 10.1016/j.ymthe.2003.12.011
Brodie SJ Patterson BK Lewinsohn DA Diem K Spach D Greenberg PD Riddell SR Corey L HIV-specific cytotoxic T lymphocytes traffic to lymph nodes and localize at sites of HIV replication and cell death [see comments] Journal of Clinical Investigation 2000 105 1407 1417 10811848
Greenberg PD Yee C Warren EH Gavin M Topp M Cooper L Nelson B Ohlen C Riddell SR Therapy of human viral and malignant diseases with gene-modified T cell clones European Cytokine Network 2000 11 304 305
Riddell SR Elliott M Lewinsohn DA Gilbert MJ Wilson L Manley SA Lupton SD Overell RW Reynolds TC Corey L Greenberg PD T-cell mediated rejection of gene-modified HIV-specific cytotoxic T lymphocytes in HIV-infected patients [see comments] Nature Medicine 1996 2 216 223 8574968 10.1038/nm0296-216
Berger C Blau CA Huang ML Iuliucci JD Dalgarno DC Gaschet J Heimfeld S Clackson T Riddell SR Pharmacologically regulated Fas-mediated death of adoptively transferred T cells in a nonhuman primate model Blood 2004 103 1261 1269 14563634 10.1182/blood-2003-08-2908
Koehne G Doubrovin M Doubrovina E Zanzonico P Gallardo HF Ivanova A Balatoni J Teruya-Feldstein J Heller G May C Serial in vivo imaging of the targeted migration of human HSV-TK-transduced antigen-specific lymphocytes Nat Biotechnol 2003 21 405 413 12652311 10.1038/nbt805
Cheng LE Ohlen C Nelson BH Greenberg PD Enhanced signaling through the IL-2 receptor in CD8+ T cells regulated by antigen recognition results in preferential proliferation and expansion of responding CD8+ T cells rather than promotion of cell death Proc Natl Acad Sci U S A 2002 99 3001 3006 11867736 10.1073/pnas.052676899
Topp MS Riddell SR Akatsuka Y Jensen MC Blattman JN Greenberg PD Restoration of CD28 expression in CD28- CD8+ memory effector T cells reconstitutes antigen-induced IL-2 production J Exp Med 2003 198 947 955 12963692 10.1084/jem.20021288
Turk MJ Guevara-Patino JA Rizzuto GA Engelhorn ME Houghton AN Concomitant tumor immunity to a poorly immunogenic melanoma is prevented by regulatory T cells J Exp Med 2004 200 771 782 15381730 10.1084/jem.20041130
Sakaguchi S Naturally arising CD4+ regulatory t cells for immunologic self-tolerance and negative control of immune responses Annu Rev Immunol 2004 22 531 562 15032588 10.1146/annurev.immunol.21.120601.141122
Malek TR Bayer AL Tolerance, not immunity, crucially depends on IL-2 Nat Rev Immunol 2004 4 665 674 15343366 10.1038/nri1435
Lou Y Wang G Lizee G Kim GJ Finkelstein SE Feng C Restifo NP Hwu P Dendritic cells strongly boost the antitumor activity of adoptively transferred T cells in vivo Cancer Res 2004 64 6783 6790 15374997 10.1158/0008-5472.CAN-04-1621
| 15860133 | PMC1131930 | CC BY | 2021-01-04 16:39:28 | no | J Transl Med. 2005 Apr 28; 3:17 | utf-8 | J Transl Med | 2,005 | 10.1186/1479-5876-3-17 | oa_comm |
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-431587781210.1186/1743-422X-2-43ReviewAdeno-associated virus: from defective virus to effective vector Gonçalves Manuel AFV [email protected] Gene Therapy Section, Department of Molecular Cell Biology, Leiden University Medical Center, Wassenaarseweg 72, 2333 AL Leiden, the Netherlands2005 6 5 2005 2 43 43 8 4 2005 6 5 2005 Copyright © 2005 Gonçalves; licensee BioMed Central Ltd.2005Gonçalves; 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 initial discovery of adeno-associated virus (AAV) mixed with adenovirus particles was not a fortuitous one but rather an expression of AAV biology. Indeed, as it came to be known, in addition to the unavoidable host cell, AAV typically needs a so-called helper virus such as adenovirus to replicate. Since the AAV life cycle revolves around another unrelated virus it was dubbed a satellite virus. However, the structural simplicity plus the defective and non-pathogenic character of this satellite virus caused recombinant forms to acquire centre-stage prominence in the current constellation of vectors for human gene therapy. In the present review, issues related to the development of recombinant AAV (rAAV) vectors, from the general principle to production methods, tropism modifications and other emerging technologies are discussed. In addition, the accumulating knowledge regarding the mechanisms of rAAV genome transduction and persistence is reviewed. The topics on rAAV vectorology are supplemented with information on the parental virus biology with an emphasis on aspects that directly impact on vector design and performance such as genome replication, genetic structure, and host cell entry.
==== Body
Adeno-associated virus biology
Genome structure, DNA replication and virus assembly
The human adeno-associated virus (AAV) was discovered in 1965 as a contaminant of adenovirus (Ad) preparations [1]. AAV is one of the smallest viruses with a non-enveloped icosahedral capsid of approximately 22 nm (Fig. 1), the crystal structure of which has been recently determined to a 3-angstrom resolution [2]. Because a co-infecting helper virus is usually required for a productive infection to occur, AAV serotypes are ascribed to a separate genus in the Parvoviridae family designated Dependovirus. Despite the high seroprevalence of AAV in the human population (approximately 80% of humans are seropositive for AAV2) the virus has not been linked to any human illness. The AAV has a linear single-stranded DNA genome of approximately 4.7-kilobases (kb). The AAV2 DNA termini consist of a 145 nucleotide-long inverted terminal repeat (ITR) that, due to the multipalindromic nature of its terminal 125 bases, can fold on itself via complementary Watson-Crick base pairing and form a characteristic T-shaped hairpin structure (Fig. 2) [3]. According to the AAV DNA replication model [4] this secondary structure provides a free 3' hydroxyl group for the initiation of viral DNA replication via a self-priming strand-displacement mechanism involving leading-strand synthesis and double-stranded replicative intermediates (Fig. 3). The virus does not encode a polymerase relying instead on cellular polymerase activities to replicate its DNA [5]. The ITRs flank the two viral genes rep (replication) and cap (capsid) encoding nonstructural and structural proteins, respectively. The rep gene, through the use of two promoters located at map positions 5 (p5) and 19 (p19), and an internal splice donor and acceptor site, encode four regulatory proteins that are dubbed Rep78, Rep68, Rep52 and Rep40 on basis of their apparent molecular weights. The Rep78 and Rep68 proteins participate in the AAV DNA replication process via their interaction with Rep-binding element (RBE) and terminal resolution site (trs) sequences located within the ITRs (Fig. 2). In addition, in response to environmental cues such as presence or absence of a helper virus these proteins either positively or negatively regulate AAV gene expression, respectively [6]. The Rep52 and Rep40 proteins are involved in the generation and accumulation of single-stranded viral genomes from double-stranded replicative intermediates [7]. The resulting single-stranded genomes with plus and minus polarities are packaged with equal efficiency [8]. The economy displayed by AAV is staggering and derives not only from its overlapping genetic organization but also from the integration of various biochemical activities in each of its few gene products. For instance, Rep78 and Rep68 are site-specific DNA binding proteins, as well as strand- and site-specific endonucleases [9]. They also exhibit helicase and ATPase activities [10], which are shared by Rep52 [11] and by Rep40 [12].
Figure 1 Transmission electron microscopy of AAV2 and Ad5 particles in human cells. (A) AAV2 and Ad5 particles in the nucleus of a HeLa cell at 48 hours after co-infection. Magnification: × 15,000. (B) AAV2 virions in a HeLa cell at 48 hours after co-infection with Ad5. Magnification: × 40,000.
Figure 2 Secondary structure of the AAV2 ITR. The AAV2 ITR serves as origin of replication and is composed of two arm palindromes (B-B' and C-C') embedded in a larger stem palindrome (A-A'). The ITR can acquire two configurations (flip and flop). The flip (depicted) and flop configurations have the B-B' and the C-C' palindrome closest to the 3' end, respectively. The D sequence is present only once at each end of the genome thus remaining single-stranded. The boxed motif corresponds to the Rep-binding element (RBE) [119] where the AAV Rep78 and Rep68 proteins bind. The RBE consists of a tetranucleotide repeat with the consensus sequence 5'-GNGC-3'. The ATP-dependent DNA helicase activities of Rep78 and Rep68 remodel the A-A' region generating a stem-loop that locates at the summit the terminal resolution site (trs) in a single-stranded form [120,121]. In this configuration, the strand- and site-specific endonuclease catalytic domain of Rep78 and Rep68 introduces a nick at the trs. The shaded nucleotides at the apex of the T-shaped structure correspond to an additional RBE (RBE') [121] that stabilizes the association between the two largest Rep proteins and the ITR.
Figure 3 Schematic representation of the AAV DNA replication model. AAV DNA replication is thought to involve a self-priming single-strand displacement mechanism that is initiated by DNA polymerisation at the 3' hairpin primer of input single-stranded genomes. This leads to the formation of linear unit-length double-stranded molecules (duplex monomers, DMs) with one covalently closed end. These structures are resolved at the terminal resolution site (trs) by site-specific nicking of the parental strand opposite the original 3' end position (i.e., at nucleotide 125). The newly generated free 3' hydroxyl groups provide a substrate for DNA polymerases that unwind and copy the inverted terminal repeat (ITR). Finally, the palindromic linear duplex termini can renaturate into terminal hairpins putting the 3' hydroxyl groups in position for single-strand displacement synthesis. Next, single-stranded genomes and new DM replicative forms are made. When nicking does not occur, elongation proceeds through the covalently closed hairpin structure generating linear double-length double-stranded molecules (duplex dimers, DDs) with either a head-to-head or a tail-to-tail configuration. The DD replicative intermediates can be resolved to DMs through the AAV ITR sequences located at the axis of symmetry.
The cap gene is transcribed from a single promoter at map position 40 (p40). Alternative splicing at two acceptor sites originates two transcripts. The larger transcript encodes virion protein 1 (VP1), the biggest capsid protein subunit. The shorter mRNA possesses a noncanonical start codon (ACG), which is utilized to generate VP2, and a downstream conventional initiation codon (AUG) directing the synthesis of VP3. The VP1, VP2 and VP3 proteins differ from each other at their N terminus and have apparent molecular masses of 87, 72 and 62 kDa, respectively. Together they assemble into a near-spherical protein shell of 60 subunits with T = 1 icosahedral symmetry. At the 12 fivefold axes of symmetry lay narrow pores lately shown to be instrumental for virus infectivity and for genome packaging [13]. The molar ratio between VP1, VP2 and VP3 in AAV particles is 1:1:10. This stoichiometry is thought to reflect the relative abundance of the two cap gene transcripts and the relative efficiency of translation initiation at the three start codons for the structural proteins. A conserved phospholipase A2 (PLA2) motif, initially identified within the unique N-terminal region of the parvoviral VP1 proteins [14], was also reported to have a biological significance in AAV2 infection [15]. Specifically, although dispensable for capsid assembly, DNA packaging, and virion internalisation, the VP1-embedded PLA2 activity seems to play a key role at some stage between the translocation of the AAV genome from the endocytic to the nuclear compartment and the initiation of viral gene expression [15]. Lately, mutational analysis of amino acid residues involved in AAV2 capsid pore architecture indicate that conformational changes of the virion structure during infection lead the VP1 N termini to protrude through the capsid pores inducing the PLA2 enzymatic activity needed for successful infection [13]. At the level of virion formation, immunofluorescence data shows that the VP1 and VP2 proteins are found primarily in the nuclei of infected cells, whereas VP3 is nearly evenly distributed between the nucleus and the cytoplasm [16]. However, in the presence of VP1 and/or VP2, VP3 accumulates in the nucleus suggesting transport of the major capsid protein by association with the nuclear localization signal-bearing proteins VP1 and VP2 [17]. Immunofluorescence results suggest that capsid assembly is confined to the nucleoli of infected cells. The involvement of nucleolar chaperones in this process has been postulated [16]. Fully assembled AAV capsids enter the nucleoplasm in an AAV Rep-dependent manner. This redistribution of the structural proteins causes the co-localization of all ingredients necessary for infectious particle formation, i.e., capsids, Rep proteins and viral genomes. Indeed, the AAV DNA packaging process is though to take place in distinct regions of the nucleoplasm [16]. Selective AAV DNA encapsidation is presumably directed by protein-protein interactions between pre-formed empty capsids and complexes of Rep78 or Rep68 with the virus genome [18]. Next, the helicase domains of capsid-docked Rep52 and Rep40 proteins are proposed to act as molecular motors that unwind and transfer de novo synthesized single-stranded DNA into empty particles [19] through the pores located at the fivefold symmetry axes [13].
Host cell infection
AAV2 virions utilize as primary attachment receptor heparan sulphate proteoglycans [20] while internalisation is aided by the co-receptors αvβ5 integrin heterodimers [21], fibroblast growth factor receptor type 1 [22] and the hepatocyte growth factor receptor, c-Met [23]. The use of ubiquitous heparan sulphate proteoglycans as docking sites explains in part the well-known broad tropism of this virus that include, human, non-human primate, canine, murine and avian cell types. AAV5 and AAV4 also bind to charged carbohydrate moieties in the form of N- and O-linked sialic acids, respectively [24]. Expression profiling of AAV5 permissive and non-permissive cells with cDNA microarrays led to the identification of platelet-derived growth factor receptor as another cellular determinant involved in AAV5 infection [25].
The events and processes that regulate the trafficking of AAV particles into the nucleus are still not fully understood, however, some findings have been reported. For instance, infection experiments in HeLa cells expressing a dominant-negative form of dynamin significantly reduced AAV2 entry [26,27]. These results indicate that one route by which this virus can poke through the plasma membrane involves receptor-mediated endocytosis via the formation of clathrin-coated pits. In addition, lysomotropic agents and proton pump inhibitors greatly hamper AAV2 infection suggesting that internalised virions escape from endosomes and are released in the cytosol by a low pH-dependent process [27]. In addition, a powerful new imaging technique based on single-molecule labelling of discrete AAV particles enabled real-time monitoring of the trajectories of individual virions [28]. In these experiments, it was shown that each endosome carries a single AAV particle. Moreover, the abrogation of vectorial motion of virions in nocodazole-treated cells supported the involvement of microtubule assembly and motor proteins in active AAV intracellular transportation. Finally, it has been suggested that AAV particles due to their very small size can access the nucleus through the nuclear pore complex (NPC). However, recent research points to a nuclear entry process that is not dependent on NPC activity [29,30] whereas the issue of whether AAV capsids enter nuclei intact or remodelled seems to depend on the presence or absence, respectively, of co-infecting helper Ad particles [30].
Lytic and lysogenic pathways
After entry into the host cell nucleus, AAV can follow either one of two distinct and interchangeable pathways of its life cycle: the lytic or the lysogenic. The former develops in cells infected with a helper virus such as Ad or herpes simplex virus (HSV) whereas the latter is established in host cells in the absence of a helper virus. When AAV infects a human cell alone, its gene expression program is auto-repressed and latency ensues by preferential integration of the virus genome into a region of roughly 2-kb on the long arm (19q13.3-qter) of human chromosome 19 [31,32] designated AAVS1 [33]. Recent research showed that this locus is in the vicinity of the muscle-specific genes p85 [34], TNNT1 and TNNI3 [35]. Furthermore, the AAVS1 sequence lies in a chromosomal region with characteristics of a transcription-competent environment [36]. Interestingly, an insulator within this locus was recently identified [37]. The targeted integration of the AAV genome, a phenomenon unique among all known eukaryotic viruses, enables the provirus DNA to be perpetuated through host cell division. Moreover, the level of specificity of this process of AAV biology (a single preintegration region within the entire human genome) makes its exploitation highly attractive for achieving the ultimate goal of safe and stable transgene expression [38].
Even if working models for the targeted DNA integration mechanism remain sketchy [39,40], the viral components needed for the site-specific integration reaction have been identified. They are composed in cis by the AAV ITRs and in trans by either one of the two largest Rep proteins (i.e., Rep78 or Rep68). Recently, another cis-acting sequence was shown to be necessary for high-level site-specific DNA integration [41,42]. This sequence overlaps with the highly regulated p5 promoter and, like the ITR sequence, harbours an RBE.
Detailed genetic analyses using an AAVS1-containing episome system demonstrated that a 33-bp sequence containing elements related to the RBE and to the trs is sufficient for targeted DNA integration. Their functional relevance was demonstrated by the absence of targeted DNA integration into mutated substrates [39]. In addition, the AAVS1 region behaves as an origin of replication in the presence of Rep proteins both in vitro [43] and in vivo [44]. Finally, the AAVS1-specific RBE and trs are separated by a spacer element whose sequence and length affects the efficiency of the site-specific DNA integration reaction [45]. The human genome has numerous Rep binding sites. However, database searches have revealed that an RBE at a proper distance from a trs sequence occurs only in the AAVS1 locus, which is consistent with the specificity of the integration reaction revealed through biological assays [46]. Moreover, in vitro studies showed that via their interaction with the RBE sequences present in the AAV ITRs and in the AAVS1 locus, Rep78 and Rep68 proteins could tether viral to cellular DNA [47]. Although, as mentioned above, the actual mechanism evolved by AAV to target its DNA to the AAVS1 locus is currently unknown, taken together these observations provide at the molecular level an explanation for the specificity of the reaction and the requirement for RBE-containing sequences in cis and either one of the two largest Rep proteins in trans. Remarkably, only recently a study emerged directly addressing the AAV DNA integration efficiency and the correlation between random versus targeted integration levels [48]. Using a tissue culture system, the authors showed by clonal analyses of target cells and Southern blot hybridisations that 50% of infected cells were stably transduced by AAV when a multiplicity of infection of 100 was used. Raising the dose of virus increased neither the frequency of infected cells nor the integration levels. Although multiplicities of infection of 100 and 10 both yielded approximately 80% infected cells, the frequency of stably transduced cells was below 5% when employing the lower dose. Virtually all integration events targeted the AAVS1 locus. Finally, for each multiplicity of infection, the frequency of AAVS1 site disruption without accompanying DNA insertion was higher than the frequency of site-specific integration by a factor of 2.
When a latently infected cell is super-infected with a helper virus, the AAV gene expression program is activated leading to the AAV Rep-mediated rescue (i.e., excision) of the provirus DNA from the host cell chromosome followed by replication and packaging of the viral genome. Finally, upon helper virus-induced cell lysis, the newly assembled virions are released. The induction of the lytic phase of the AAV life cycle from a stably integrated provirus can also occur in the absence of a helper virus, though with a lower efficiency, when the host cell is subjected to metabolic inhibitors and to DNA damaging agents such as UV irradiation or genotoxic compounds [49]. Moreover, in differentiated keratinocytes of an epithelial tissue culture system modelling skin, AAV2 was shown to initiate and proceed through a complete replicative cycle in the absence of helper viruses or genotoxic agents [50]. Taken together, these phenomena indicate that AAV is not defective in absolute terms.
Adeno-associated virus vectorology
General principle
Historically, most recombinant AAV (rAAV) vectors have been based on serotype 2 (AAV2) that constitutes the prototype of the genus [51,52]. Important to those pursuing the use of rAAV for gene therapy applications is the defectiveness of the parental virus and its presumed non-pathogenic nature. The realization that a molecularly cloned AAV genome could in Ad-infected cells recapitulate the lytic phase of the AAV life cycle and give rise to infectious virions enabled not only the detailed genetic analyses of the virus but provided, in addition, a substrate to generate rAAV particles [53]. The latter task was facilitated by the fact that the AAV ITRs contain all cis-acting elements involved in genome rescue, replication and packaging. Furthermore, since the AAV ITRs are segregated from the viral encoding regions, rAAV design can follow the whole-gene-removal or "gutless" vector rational of, for instance, retrovirus-based vectors in the sense that the cis-acting elements involved in genome amplification and packaging are in linkage with the heterologous sequences of interest, whereas the virus encoding sequences necessary for genome replication and virion assembly are provided in trans (Fig. 4). Typically, rAAV particles are generated by transfecting producer cells with a plasmid containing a cloned rAAV genome composed of foreign DNA flanked by the 145 nucleotide-long AAV ITRs and a construct expressing in trans the viral rep and cap genes. In the presence of Ad helper functions, the rAAV genome is subjected to the wild-type AAV lytic processes by being rescued from the plasmid backbone, replicated and packaged into pre-formed AAV capsids as single-stranded molecules.
Figure 4 Overview of the initial recombinant AAV production system. The generation of the first infectious clones of AAV permitted functional dissection of the virus genome. This allowed the construction of plasmids encoding rAAV genomes in which the minimal complement of wild-type sequences necessary for genome replication and packaging (i.e., the AAV ITRs) frame a gene of interest (transgene) instead of the AAV rep and cap genes. When these constructs are transfected into packaging cells together with a rep and cap expression plasmid they lead to the production of rAAV particles. Helper activities required for the activation and support of the productive phase of the AAV life cycle were originally introduced by infection of the packaging cells with wild-type Ad as depicted. Current transfection-based production methods make use of recombinant DNA encoding the helper activities instead of Ad infection. Cellular DNA polymerase activities together with the Rep78 and Rep68 proteins lead to the accumulation of replicative intermediates both in the duplex monomer (DM) and duplex dimer (DD) forms. A fraction of this de novo synthesized DNA is incorporated in the single-stranded format into preformed empty capsids most likely through the catalytic activities of the Rep52 and Rep40 proteins. The resulting infectious rAAV virions are released from the producer cells together with helper Ad particles. Sequential heat treatment and buoyant density centrifugation allows the selective elimination of the helper virus from the final rAAV preparation.
Production and purification strategies
The Ad helper functions were originally supplied by infection of rAAV producer cells with a wild-type Ad (Fig. 4). Subsequent elimination of the helper virus from rAAV stocks relied on the distinct physical properties of AAV and Ad virions. In particular, differences in thermostability and density between AAV and Ad particles allowed the specific elimination of helper Ad virions by heat-inactivation (i.e., half-hour at 56°C) and isopycnic cesium chloride density ultracentrifugation. The finding that Ad helper functions are provided by expression of E1A, E1B, E2A, E4ORF6 and VA RNAs, enabled subsequent Ad-free production of rAAV vector stocks by incorporating VA RNAs, E2a and E4ORF6 sequences into a plasmid and transfecting it together with the rAAV DNA plus rep and cap templates into Ad E1A- and E1B-expressing cells [54-56]. During the testing of new packaging plamids for rAAV production it was also found that reduction of the expression levels of the two largest AAV Rep proteins leads to an increase in vector yields [56,57]. Although these methods improve rAAV production and avoid the need for Ad infection, they are difficult to scale up due to their dependence on DNA transfection. The development of up-scalable transfection-independent methods for rAAV production have been fiercely pursued by the requirement for large amounts of highly purified vector particles to perform experiments in large animal models and human clinical trials. One of these transfection-independent production strategies involves the generation of packaging cell lines having the AAV rep and cap genes stably integrated in their genomes. The establishment of effective, high-titer producer cell lines has proven difficult mainly due to the inhibitory effects of Rep proteins on cell growth [58] and the accumulation of low amounts of AAV gene products relative to a wild-type virus infection. Nonetheless, improvements in the control of rep expression through the development of stringent inducible gene expression systems can overcome the former hurdle [59] whereas in situ amplification of integrated rep and cap templates helps to minimize the latter problem [60,61]. Another transfection-independent approach to produce rAAV involves the delivery of the viral genes together with the rAAV DNA and the helper functions via infection of produced cells with recombinant viruses based on Ad [60], HSV [62] or baculovirus [63]. In parallel to new rAAV production platforms, insights into AAV biology are also leading to significant improvements in the quality and purity of vectors based on AAV2 as well as on those based on other serotypes. Specifically, knowledge on AAV receptor usage has permitted the implementation of up-scalable affinity column chromatography purification schemes [64,65]. In addition, a more broadly applicable column chromatography procedure, based on the ion-exchange principle, has recently been developed for the purification of rAAV2, rAAV4 and rAAV5 particles [66].
Tropism modification
An increasingly important area in the development of AAV as a vector concerns the engineering of altered cell tropisms to narrow or broaden rAAV-mediated gene delivery and to increase its efficiency in tissues refractory to AAV2 infection. Cells can be poorly transduced by prototype rAAV2 not only because of low receptor content but also owing to impaired intracellular virion trafficking and uncoating [67,68] or single-to-double strand genome conversion [69-71]. Thus, considering that these processes depend either directly or indirectly on capsid conformation, cell targeting strategies determine not only the cell type(s) with which the vector interacts but also critically affect the efficiency of the whole gene transfer process.
Several of these approaches rely on the modification by chemical, immunological or genetic means of the AAV2 capsid structure endowing it with ligands that interact with specific cell surface molecules [72]. The fact that the atomic structure of AAV2 has recently been determined [2] provides a significant boon to those pursuing the rational design of targeted AAV vectors. Another route to alter rAAV tropism exploits the natural capsid diversity of newly isolated serotypes by packaging rAAV2 genomes into capsids derived from other human or non-human AAV isolates [73]. To this end, up until now, most researches employ hybrid trans-complementing constructs that encode rep from AAV2 whereas cap is derived from the serotype displaying the cell tropism of choice. This pseudotyping approach may also be beneficial in evading neutralizing antibodies to capsid components in individuals seropositive for AAV2 or in those in need of vector readministration. Finally, experiments published recently using rAAV2 genomes pseudotyped with coats from AAV6 [74] and AAV8 [75] revealed stunning gene transfer efficiencies when these vectors were administered alone at high doses or in combination with a blood vessel permeating agent. The authors could demonstrate transduction of the entire murine striated muscle system (e.g., diaphragm, heart and skeletal muscles) and of virtually 100% of the hepatocytes after a single intravenous injection. These body-wide transduction efficiencies raise both great perspectives as well as caution since they open new therapeutic avenues for diseases that require widespread gene delivery (e.g., muscular dystrophies) while, simultaneously, beg for stringent tissue-specific transcriptional control to minimize potential deleterious effects due to transgene expression in non-target tissues. Moreover, assuming similar avidity of these serotypes for human tissues, translation of these protocols from mice to patients will require vastly greater amounts of vector particles.
Mechanisms of vector DNA persistence
Knowledge on the mechanisms at play following rAAV transduction is building steadily over recent years mainly because of its direct relevance to the application of rAAV in therapeutic gene transfer. DNA vectored through rAAV can persist long-term in organs such as in the liver and the striated muscles of mice and dogs. Most importantly, data showing prolonged and stable expression of an increasing variety of transgenes in numerous animal models without notable toxicity is accumulating [76]. It are in fact these attributes of rAAV-based gene transfer that turns it into one of the most promising methods for somatic gene therapy providing a rational for the entry of these vectors into the clinical trial arena. However, at the outset it is important to refer that this stability does not arise due to foreign DNA insertion into the parental virus pre-integration site since the absence of rep gene products prevents DNA targeting to the AAVS1 locus. Moreover, because rAAV vectors lack viral genes altogether, the molecular fate of the DNA once in the nucleus is dependent on host cell activities (though a role for the virion capsomers cannot be ruled out). These cellular activities, that only recently have started to be identified, depend on the type as well as on the physiological status of the target cell. Finally, it is also of note that the single-stranded nature of AAV genomes implies that, before transgene expression can occur, the incoming rAAV DNA needs to be converted into a transcriptionally functional double-stranded template.
A recent study indicates that free (i.e., unpackaged) single-stranded rAAV genomes have a very transient presence in the target cell [67] either because the majority is recognized by host enzymes as damaged DNA and degraded or because, under certain conditions, single-to-double strand conversion occurs readily following uncoating. There are two pathways by which rAAV DNA can be converted from the single- to the double-stranded form each of them with its own set of supporting experimental data. One possible route develops through de novo second-strand DNA synthesis from the hairpin at the 3' end of the genome (Fig. 2). Initial studies revealed that this step could be greatly enhanced by Ad E4ORF6 expression, UV irradiation or treatment of target cells with genotoxic chemicals [69,70]. Furthermore, a direct correlation between double-stranded template accumulation and gene expression was found. More recently, the phosphorylation status of a cellular protein named FKBP52 was shown to modulate the convertion of single-stranded rAAV DNA into double-stranded molecules both in tissue culture [77] and in murine hepatocytes [78]. FKBP52 phosphorylation by the epidermal growth factor receptor protein tyrosine kinase enables the molecule to bind the single-stranded AAV ITR D-sequence (Fig. 2). This binding activity correlates strongly with second-strand DNA synthesis inhibition. Conversely, in its dephosphorylated state, due to T-cell protein tyrosine phosphatase activity, FKBP52 does not bind vector genomes allowing synthesis of the complementary strand to occur with a subsequent increase in transgene expression levels.
As said before, single-stranded AAV genomes with sense (plus) and anti-sense (minus) orientations are packaged equally well. Therefore, another possible route involved in the generation of double-stranded DNA forms in target cells comprises the annealing of single-stranded molecules with opposing polarities. Evidence for the existence of this DNA synthesis-independent pathway came from experiments using rAAV genomes that were site-specifically methylated [71]. In these experiments restriction enzymes were used as probes to evaluate whether modified rAAV genomes extracted from murine livers were fully methylated (representing annealing products) or hemimethylated (corresponding to second-strand synthesis products). Thus, seemingly, a contention exits between advocates of DNA synthesis dependent and independent models. It is clear, however, that these two pathways are not necessarily mutually exclusive. In fact, recent experiments in cells under normal physiological conditions indicate that each of these pathways can contribute to the generation of transcriptionally active rAAV genomes [67]. For the latter experiments the authors resurrected a technique deployed to directly demonstrate that AAV is a single-stranded virus [8]. Exploiting the differential thymidine content of complementary polynucleotide chains they used incorporation of the thymidine analogue bromodeoxyuridine (BrdU) to physically separate plus- from minus-strand containing rAAV particles following buoyant density centrifugation. Infection of indicator cells with each vector type led to reporter gene expression signifying the involvement of second-strand DNA synthesis and precluding an absolute requirement for plus and minus strand annealing. However, co-infection with both vector types originated higher numbers of cells expressing the reporter gene indicating that strand annealing contributes to the accumulation of double-stranded genomes [67].
Subsequently, duplex rAAV genomes can, throught intra- or intermolecular recombination at the ITRs, originate circular forms or linear concatemers, respectively [71,79]. The circular episomes can also evolve into high-molecular-weight concatamers in a time-dependent manner [79]. The balance between linear versus circular forms is, at least in part, regulated by a complex containing DNA-dependent protein kinase (DNA-PK) [80]. This complex plays a vital role in the repair of double-stranded chromosomal breaks and in V(D)J recombination by non-homologous end-joining (NHEJ). The absence of the catalytic subunit of DNA-PK (DNA-PKcs) in severe combined immunodeficient (SCID) mice (DNA-PKcs-negative) allowed Song and colleagues to demonstrate its involvement in circular rAAV episome formation in skeletal muscle [80]. Subsequent studies in liver and skeletal muscle of SCID and normal (DNA-PKcs-positive) mice have extended the observation that DNA-PK enhances the formation of rAAV circular episomes over linear forms [81,82]. It has been postulated that free double-stranded rAAV DNA ends are substrates for the cellular double-stranded break repair machinery responsible for free-ended DNA removal through NHEJ ligation [80]. Notwithstanding their diverse topology and unit numbers, all these extrachromosomal DNA forms are transcription-competent templates. Furthermore, they are thought to be responsible for the stable maintenance of transgene expression both in skeletal muscles [79] and in the lungs [83]. In the liver it has been shown that, in addition to the aforesaid episomal forms, circa 10% of the double-stranded rAAV genomes can be found inserted in the chromosomal DNA [84].
Backed by the complete mouse genome sequence, researchers could establish that a significant proportion of rAAV DNA integration events occur in regions that are transcriptionally active in murine hepatocytes [85]. In some instances, sequence micro-homologies and unrelated nucleotides are found at the truncated ITR-chromosomal DNA junctions. Moreover, rAAV DNA insertion is consistently associated with host chromosomal deletions. These characteristics resemble the "fingerprints" following double-stranded DNA break repair through NHEJ recombination. Thus, taken together, these results point to the involvement of NHEJ in rAAV DNA integration in addition to its putative role in the removal of free rAAV DNA ends, as previously discussed. This interpretation is further supported by previous and newly acquired data. For instance, earlier tissue culture studies revealed a direct correlation between genomic instability due to DNA-damaging agents or genetic defects and stable transduction by rAAV [86,87]. Other results showed that proteins belonging to the NHEJ complex bind to linear rAAV DNA [88]. More recently, a genetic approach permitted the deliberate induction of double-stranded chromosomal breaks within a predefined site [89]. The experimental set up consisted of retrovirus vector-mediated expression of the I-SceI endonuclease in cells engineered with this enzyme's 18-bp recognition sequence. Following transduction of these cells with rAAV, the authors could demonstrate insertion of foreign DNA into I-SceI-induced double-stranded breaks. Characterization of vector-chromosome junctions revealed the telltale features observed after rAAV DNA integration into chromosomal breaks arising spontaneously at random sites. It is thus possible to speculate that rAAV proviral DNA is just another by-product of the mechanism the cell uses to eliminate free-ended substrates reminiscent of damaged DNA or invading nucleic acids (e.g., linear retroviral cDNA). As corollary, compared to the integrase-dependent retroviral genome integration, rAAV DNA insertion is a passive process that relies instead on pre-existent chromosomal breaks and host cell enzymes.
Chromosomal DNA integration with current vectors is a double-edged sword. On the one hand it provides a basis for permanent genetic correction while, on the other hand, raises safety issues related to insertional gene-inactivation and proto-oncogene deregulation. It is thus highly relevant for the clinical deployment of rAAV that these vectors do not create but instead insert into existing chromosomal breaks. The latter can be substrates for inaccurate NHEJ-mediated repair regardless of the presence of rAAV genomes. Therefore, concerns about insertional oncogenesis might be less for rAAV- than for retroviral vector-mediated gene transfer. Additionally, in contrast to retroviral vectors, rAAV vectors do not display "outward" promoter activity. Despite this, it is still conceivable that rAAV DNA insertion can lead to hazardous alteration of neighbouring gene(s) expression via vector-encoded regulatory sequences (e.g., enhancers). Thus, preventive measures such as judicious choice of transcriptional elements and use of insulators may turn out to be desirable or even indispensable in target tissues in which rAAV DNA is known to integrate at appreciable levels. Adding to the challenge these genetic elements have to be small enough to leave space needed to accommodate the gene of interest.
Emerging technologies
The small packaging capacity of AAV particles (about 4.7 kb) [90] is considered one of the main limitations of rAAV vectors since it excludes therapeutically important coding sequences (e.g., dystrophin cDNA) and potent regulatory elements (e.g., albumin promoter). As discussed above, incoming linear rAAV genomes can form concatamers in target cells through intermolecular recombination at their free ends. This phenomenon has been successfully exploited to assemble in target cells large genetic messages through the joining of two independently transduced rAAV genomes each of which encompassing a portion of a large transcriptional unit. mRNA molecules encoding a functional protein are generated from the rAAV DNA head-to-tail heterodimers by splicing out the AAV ITR sequences from the primary transcripts (Fig. 5) [91]. Although this split gene strategy allows expression of almost double-sized transgenes after rAAV-mediated gene delivery, its efficiency is consistently lower than that observed with a single control vector encoding the full-length transgene. Both vectors have to transduce the same cell and only heteroconcatamers with a head-to-tail organization will give rise to a functional full-length gene product. In addition, there are risks associated with the integration into host chromosomes of vectors encoding exclusively regulatory elements or truncated gene products. New work, however, suggests that some of these limitations and concerns can, at least partially, be addressed [92,93].
Figure 5 Diagram of the recombinant AAV split gene principle. An expression unit corresponding to a large gene is roughly divided in two halves. One of them consists of a promoter (solid box with arrowhead), the 5' half of the gene (open box) and a splice donor site (SD) while the other encodes a splice acceptor sequence (SA), the 3' portion of the gene (shaded box) and a polyadenylation signal (solid box). These fragments are independently cloned between two AAV ITRs. Vector stocks are then generated from the resulting shuttle plasmids and are used to co-transduce target cells. Head-to-tail heterodimerization via intermolecular recombination between the two vector DNA molecules restores the full-length expression unit and results in the synthesis of the desired protein after the splicing of the intervening AAV ITR sequences from the primary transcript.
Another development in rAAV design is the so-called self-complementary AAV vectors (scAAV) [94]. The scAAV approach builds on the ability of AAV to package replicons with half the size of the wild-type DNA in the form of single-stranded dimeric genomes with an inverted repeat configuration [95]. In the target cell, these self-complementary molecules can readily fold back into double-stranded forms without the need for de novo DNA synthesis or for the annealing of sense and antisense strands (Fig. 6). Ultimately, regardless of the mechanism(s) at play, scAAV lead to enhanced formation of transcription-competent double-stranded genomes thus improving the expression kinetics and yields of vector-encoded products. This scAAV method was subsequently perfected by mutagenesis of one of the two trs sequences to force the generation of dimeric over monomeric replicative forms (Fig. 6) [96]. The main disadvantage of this approach is the need to limit the size of the transgenes that can be delivered to approximately half the length of the already small AAV genome. It is conceivable that this drawback can be tackled by coupling scAAV with heterodimerization strategies. Alternatively, long double-stranded rAAV genomes can be transferred into target cells via capsids of larger viruses such as Ad [97-100], baculovirus [101] or HSV [102]. In some of these hybrid viral vector systems, integration of the rAAV DNA into the AAVS1 locus on human chromosome 19 was accomplished by transient expression of AAV Rep activities in the target cells [38]. Targeted DNA integration is advantageous since it dispels the insertional oncogenesis concerns discussed above.
Figure 6 Diagram of the generation and transduction of a self-complementary AAV vector as compared to that of a conventional recombinant AAV. Left panel: According to the AAV DNA replication scheme, full-length rAAV genomes of both polarities are generated from duplex monomeric (DM) and duplex dimeric (DD) replicative intermediates and individually packaged in AAV capsids. In the nucleus of transduced cells the single-stranded genomes can either be a target for degradation or be converted into transcriptionally active double-stranded templates. The single-to-double strand DNA conversion depends on complementary chain synthesis or on the recruitment of a complementary genome (i.e., intermolecular hybridization). Right panel: According to the same replication model, a rAAV genome with roughly half the size of the wild-type AAV DNA and with one trs mutated, generates DD replicative intermediates with an inverted repeat configuration containing wild-type ITRs at the extremities and mutated ITRs at the axis of symmetry. Single-stranded molecules derived from these DNA structures are packaged in AAV capsids. After uncoating in the target cell nucleus, these molecules can readily fold into double-stranded templates through intramolecular base pairing due to their self-complementary nature (i.e., intramolecular hybridization).
Site-specific or targeted DNA integration can also be achieved through homologous recombination (HR) between a transduced DNA fragment and an endogenous gene in the target cell genome. The ability to introduce precise genetic modifications in germ cells of mice combined with powerful selection markers has revolutionized mammalian genetics [103]. The same principle can be applied to achieve correction of defective genes in somatic human cells. In fact, targeted gene correction is conceptually an attractive alternative to gene addition since there is no strict need to transduce the entire gene and associated regulatory elements but only a fraction of the targeted gene sequence. In addition, the corrected gene remains in its chromosomal context thus being subject to the proper regulatory circuitry. However, gene targeting strategies are currently not practical mostly due to the inefficiency of HR after foreign DNA delivery (typical frequencies lie below 10-6). It has been demonstrated that rAAV can be tailored to introduce precise nucleotide alterations in the genome of human cells at frequencies approaching 1% when multiplicities of infection in the order of 105 to 106 infectious genomes per cell are used [104]. In these experiments, it was observed that for each targeted integration event 10 non-targeted DNA insertions occurred and that, in comparison with other methods, the HR process was less dependent on the extent of homology. More recently, this technology was successfully used in human mesenchymal stem cells to disrupt via HR a mutant COL1A1 allele coding for a dominant-negative type of collagen causing osteogenesis imperfecta [105].
Clinical trials
Data on safe and long-lasting rAAV-mediated transgene expression in organs of animal models of human disease such as lung, liver, central nervous system and eye, together with improvements in vector production and purification methods provided the rational for initiating clinical studies with rAAV vectors. Currently, these clinical trials are either in phase I or in phase II. The former studies aim at determining safety and often also maximum tolerable dose of the therapeutic agent, while the latter entail the assessment of its efficacy and have higher statistical significance to detect potential side effects. Ailments being targeted include Parkinson's disease, Canavan's disease, α1-antitrypsin deficiency, cystic fibrosis (cystic fibrosis transmembrane conductance regulator [CFTR] deficiency) and hemophilia B (blood clotting factor IX [FIX] deficiency). Cystic fibrosis and hemophilia B are two examples of which more information is available. In fact, more than one decade ago, cystic fibrosis patients were the first human individuals subjected to rAAV administration [106].
Cystic fibrosis is the most common autosomal recessive disorder among Caucasians. The CFTR gene encodes a chloride channel that is essential for the transport of chloride ions across the membranes of epithelial cells of the lungs, gastrointestinal tract and sweat glands. The CFTR aids in the physiological transport of other ions and water. The pathophysiology of cystic fibrosis in the lung is not settled [107]. However, it seems uncontroversial that in the absence of functional CFTR, mucus of high viscosity and abnormal ionic content covers the airway epithelium leading to the accumulation of infectious agents. Chronic inflammation results in lung tissue damage and loss of respiratory function. Early death ensues.
As said before, all clinical trials are based on preclinical data retrieved from experiments in animal models. Unfortunately, CFTR knockout mice display primarily intestinal defects as opposed to the lung deterioration typical of the human condition. Accordingly, New Zealand white rabbits [108] and rhesus monkeys [109] constituted the major preclinical models for rAAV-mediated CFTR cDNA transfer. Overall, these studies showed that transduction with AAV2-based vectors led to prolonged and dose-dependent CFTR cDNA expression in the respiratory tract after various modes of administration (e.g., direct bronchoscopic instillation and aerosol delivery). Importantly, no overt signs of vector-associated inflammation or toxicity were observed. Equally important, vector DNA was not detected in the gonads of any of the experimental animals tested, indicating that the risk of inadvertent germline transmission is very low. Initial clinical results showed rAAV2-mediated CFTR delivery to be well tolerated by human patients as well. It is also known from phase I dose-escalation studies that the aerosol method permits the delivery of vector DNA throughout the lung in a dose-dependent manner. Although vector sequences persisted for up to 90 days at the highest dose, vector-specific transcripts could not be detected in the samples tested [110]. A follow up placebo-controlled phase II study incorporated into its design repeated administration of aerosolized vector particles. In addition to safety monitoring, this trial included the evaluation of proinflammatory cytokine interleukine-8 (IL-8) levels and pulmonary function. The treatment was well tolerated and, at days 30 and 14, vector-treated patients showed evidence of improved lung function and reduced IL-8 concentrations in the sputum, respectively, when compared to placebo-treated individuals [111]. On the basis of these promising results new and expanded phase II clinical trials are currently underway.
In contrast to the mouse model of cystic fibrosis, FIX knockout mice and naturally occurring FIX-defective canines with missense and null mutations accurately mimic hemophilia B in humans. In addition, this X-linked coagulopathy has other features that turn it into an attractive target for gene transfer approaches. Firstly, the limited size of the FIX cDNA (i.e., 2.8 kb) allows the testing of a large variety of gene delivery systems including those with a small packaging capacity. Secondly, regulation of FIX expression is not needed because the encoded product has a broad therapeutic index and, importantly, concentrations above 1% of the physiological level start to be beneficial (i.e., < 1, 1 to 5, and > 5% correspond to severe, moderate and mild disease, respectively). Finally, although the liver is the normal site of FIX production, synthesis and secretion of a biologically active form of this protein can also be achieved from ectopic, easily accessible, tissues such as skeletal muscle. Indeed, sustained dose-dependent therapeutic levels of canine FIX expression were attained in hemophilic dogs after both portal vein [112] and intramuscular [113] injections of rAAV2 particles. Partial phenotypic correction could be unambiguously established in these studies by measurement of hemostatic parameters such as the whole blood clotting time (WBCT) and the activated partial thromboplastin time (aPTT) lending support for the testing of rAAV2 in patients. In 1999, a dose-escalation phase I trial consisting of three dose cohorts (i.e., 2.0 × 1011, 6.0 × 1011, and 1.8 × 1012 vector genomes per kilogram of body weight) with three patients each was initiated. The readily accessible vastus lateralis muscle was chosen as target tissue for safety reasons. Results from these first parenteral administrations of rAAV in human subjects showed safe transfer of FIX without evidence for the formation of inhibitory antibodies to FIX and for the presence of vector sequences in semen. Gene transfer was detected by PCR and Southern blot analyses, whereas immunohistochemical staining of muscle biopsies revealed sustained transgene expression distributed mainly in slow twitch fibers [114]. However, this trial also showed that the doses tested were too low to bring about FIX plasma concentrations decisively above 1% of the normal value. It became apparent that therapeutic doses required numerous injections with more particles being administered per site. Several issues, however, blocked this approach. Firstly, the number of injections needed rendered the procedure impractical. Secondly, it was considered that saturation of the AAV2 receptors and of the capacity of myocytes to secrete FIX with the correct posttranslational modifications [115] would curtail the effect of using very high particle concentrations. Finally, and most importantly, a correlation was observed between injection of very high dosages of rAAV2 into muscle and the development of FIX neutralizing antibodies [113].
The next phase I trial targeted the liver of individuals with missense mutations by systemic administration of FIX-encoding rAAV2. Unfortunately, this trial has been halted. Low vector doses were well tolerated but did not induce FIX levels above baseline, whereas high vector doses achieved only transient FIX expression and induced hepatotoxicity and immune responses against the vector and the transgene product [116]. Hopefully, new developments in rAAV technologies such as, vectors endowed with regulatory elements for high-level tissue-specific expression and higher liver and/or muscle tissue avidities will increase the therapeutic potency of rAAV-mediated FIX transfer in humans. Towards this goal, intraportal administration of an AAV8-based vector directing the synthesis of canine FIX through a liver-specific promoter achieved stable curative levels of the protein in naïve and in AAV2-preimmunized hemophilia B dogs (i.e., up to 26% and 16% of normal levels, respectively) [117]. The results obtained in AAV2-pretreated dogs are particularly significant if one considers that a significant proportion of humans have high AAV2 neutralizing antibody titers [118].
Conclusion
Important strides have recently been made in the optimisation of rAAV technology at the levels of production and performance. Insights from AAV biology have been instrumental in this process and are expected to continue to be the main catalyst behind the further development and efficacious deployment of rAAV. Most of the features initially identified in AAV as being highly desirable in a therapeutic gene carrier such as the seemingly nonpathogenic nature of the wild-type virus and its ability to infect, non-dividing, terminally differentiated cells remain valid and contribute to put rAAV at the forefront of all vector systems that aim at safe and sustained transgene expression in vivo. A notable exception of an AAV attribute not retained by rAAV concerns the loss of AAVS1-targeted DNA integration.
The number of promising reports documenting rAAV-mediated stable transgene expression in immunocompetent recipients is steadily increasing. However, the vast majority of these results have been obtained in inbred rodent models with relatively little genetic diversity. There are several indications (e.g., from research on rAAV-mediated FIX transfer) that the results obtained in mice cannot predict the outcome of experiments carried out in patients. This underscores the need not only for continuous improvement of the vectors themselves but also for deepening the knowledge about vector-host interactions outside the realm of rodent models. The ultimate goal of this research is to accomplish unequivocal clinical benefit by the identification of limitations and corresponding solutions to each particular disease-transgene-vector trilogy.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
I am grateful to Drs. Antoine A.F. de Vries, Shoshan-Knaän Shanzer and Maria Grazia Pau for their critical comments to this manuscript and to my lab colleagues for their enthusiasm and help. I thank Dr. Maria Grazia Pau and Maarten Holkers for making available the images depicted in figure 1 and 2, respectively. I am also thankful to the Fundação Portuguesa para a Ciência e Tecnologia and the Prinses Beatrix Fonds for neuromuscular diseases for previous (PRAXIS XXI/BD/9157/96) and current grants (MAR04-0222), respectively.
==== Refs
Atchison RW Castro BC Hammon WM Adenovirus-associated defective virus particles Science 1965 149 754 756 14325163
Xie Q Bu W Bhatia S Hare J Somasundaram T Azzi A Chapman MS The atomic structure of adeno-associated virus (AAV-2), a vector for human gene therapy Proc Natl Acad Sci USA 2002 99 10405 10410 12136130 10.1073/pnas.162250899
Koczot FJ Carter BJ Garon CF Rose JA Self-complementarity of terminal sequences within plus or minus strands of adenovirus-associated virus DNA Proc Natl Acad Sci USA 1973 70 215 219 4509654
Berns KI Parvovirus replication Microbiol Rev 1990 54 316 329 2215424
Ni T-H McDonald WF Zolotukhin I Melendy T Waga S Stillman B Muzyczka N Cellular proteins required for adeno-associated virus DNA replication in the absence of adenovirus coinfection J Virol 1998 72 2777 2787 9525597
Pereira DJ McCarty DM Muzyczka N The adeno-associated virus (AAV) Rep protein acts as both a repressor and an activator to regulate AAV transcription during a productive infection J Virol 1997 71 1079 1088 8995628
Chejanovsky N Carter BJ Mutagenesis of an AUG codon in the adeno-associated virus rep gene: effects on viral DNA replication Virology 1989 173 120 128 2554565 10.1016/0042-6822(89)90227-4
Berns KI Adler S Separation of two types of adeno-associated virus particles containing complementary polynucleotide chains J Virol 1972 9 394 396 5014934
Berns KI Giraud C Biology of adeno-associated virus Curr Top Microbiol Immunol 1996 218 1 23 8794242
Im DS Muzyczka N The AAV origin binding protein Rep68 is an ATP-dependent site-specific endonuclease with DNA helicase activity Cell 1990 61 447 457 2159383 10.1016/0092-8674(90)90526-K
Smith RH Kotin RM The Rep52 gene product of adeno-associated virus is a DNA helicase with 3'-to-5' polarity J Virol 1998 72 4874 4881 9573254
Collaco RF Kalman-Maltese V Smith AD Dignam JD Trempe JP A biochemical characterization of the adeno-associated virus rep40 helicase J Biol Chem 2003 278 34011 34017 12824181 10.1074/jbc.M301537200
Bleker S Sonntag F Kleinschmidt JA Mutational analysis of narrow pores at the fivefold symmetry axes of adeno-associated virus type 2 capsids reveals a dual role in genome packaging and activation of phospholipase A2 activity J Virol 2005 79 2528 2540 15681453 10.1128/JVI.79.4.2528-2540.2005
Zádori Z Szelei J Lacoste M-C Li Y Gariépy S Raymond P Allaire M Nabi IR Tijssen P A viral phospholipase A2 is required for parvovirus infectivity Dev Cell 2001 1 291 302 11702787 10.1016/S1534-5807(01)00031-4
Girod A Wobus CE Zádori Z Ried M Leike K Tijssen P Kleinschmidt JA Hallek M The VP1 capsid protein of adeno-associated virus type 2 is carrying a phospholipase A2 domain required for virus infectivity J Gen Virol 2002 83 973 978 11961250
Wistuba A Kern A Weger S Grimm D Kleinschmidt J Subcellular compartmentalization of adeno-associated virus type 2 assembly J Virol 1997 71 1341 1352 8995658
Ruffing M Zentgraf H Kleinschmidt JA Assembly of viruslike particles by recombinant structural proteins of adeno-associated virus type 2 in insect cells J Virol 1992 66 6922 6930 1331503
Im DS Muzyczka N Factors that bind to adeno-associated virus terminal repeats J Virol 1989 63 3095 3104 2542617
King JA Dubielzig R Grimm D Kleinschmidt JA DNA helicase-mediated packaging of adeno-associated virus type 2 genomes into preformed capsids EMBO J 2001 20 3282 3291 11406604 10.1093/emboj/20.12.3282
Summerford C Samulski RJ Membrane-associated heparan sulfate proteoglycan is a receptor for adeno-associated virus type 2 virions J Virol 1998 72 1438 1445 9445046
Summerford C Bartlett JS Samulski RJ αVβ5 integrin: a co-receptor for adeno-associated virus type 2 infection Nat Med 1999 5 78 82 9883843 10.1038/4768
Qing K Mah C Hansen J Zhou S Dwarki V Srivastava A Human fibroblast growth factor 1 is a co-receptor for infection by adeno-associated virus 2 Nat Med 1999 5 71 77 9883842 10.1038/8526
Kashiwakura Y Tamayose K Iwabuchi K Hirai Y Shimada T Matsumoto K Nakamura T Oshimi K Daida H Hepatocyte growth factor receptor is a coreceptor for adeno-associated virus type 2 infection J Virol 2005 79 609 614 15596854 10.1128/JVI.79.1.609-614.2005
Kaludov N Brown KE Walters RW Zabner J Chiorini JA Adeno-associated virus serotype 4 (AAV4) and AAV5 both require sialic acid binding for hemagglutination and efficient transduction but differ in sialic acid linkage specificity J Virol 2001 75 6884 6893 11435568 10.1128/JVI.75.15.6884-6893.2001
Di Pasquale G Davidson BL Stein CS Martins I Scudiero D Monks A Chiorini JA Identification of PDGFR as a receptor for AAV-5 transduction Nat Med 2003 9 1306 1312 14502277 10.1038/nm929
Duan D Li Q Kao AW Yue Y Pessin JE Engelhardt JF Dynamin is required for recombinant adeno-associated virus type 2 infection J Virol 1999 73 10371 10376 10559355
Bartlett JS Wilcher R Samulski RJ Infectious entry pathway of adeno-associated virus and adeno-associated virus vectors J Virol 2000 74 2777 2785 10684294 10.1128/JVI.74.6.2777-2785.2000
Seisenberger G Ried MU Endreß T Büning H Hallek M Bräuchle C Real-time single-molecule imaging of the infection pathway of an adeno-associated virus Science 2001 294 1929 1932 11729319 10.1126/science.1064103
Hansen J Qing K Srivastava A Infection of purified nuclei by adeno-associated virus 2 Mol Ther 2001 4 289 296 11592830 10.1006/mthe.2001.0457
Xiao W Warrington KH JrHearing P Hughes J Muzyczka N Adenovirus-facilitated nuclear translocation of adeno-associated virus type 2 J Virol 2002 76 11505 11517 12388712 10.1128/JVI.76.22.11505-11517.2002
Kotin RM Siniscalco M Samulski RJ Zhu XD Hunter L Laughlin CA McLaughlin S Muzyczka N Rocchi M Berns KI Site-specific integration by adeno-associated virus Proc Natl Acad Sci USA 1990 87 2211 2215 2156265
Samulski RJ Zhu X Xiao X Brook JD Housman DE Epstein N Hunter LA Targeted integration of adeno-associated virus (AAV) into human chromosome 19 EMBO J 1991 10 3941 3950 erratum 11:1228 1657596
Kotin RM Menninger JC Ward DC Berns KI Mapping and direct visualization of a region-specific viral DNA integration site on chromosome 19q13-qter Genomics 1991 10 831 834 1653762 10.1016/0888-7543(91)90470-Y
Tan I Ng CH Lim L Leung T Phosphorylation of a novel myosin binding subunit of protein phosphatase 1 reveals a conserved mechanism in the regulation of actin cytoskeleton J Biol Chem 2001 276 21209 21216 11399775 10.1074/jbc.M102615200
Dutheil N Shi F Dupressoir T Linden RM Adeno-associated virus site-specifically integrates into a muscle-specific DNA region Proc Natl Acad Sci USA 2000 97 4862 4866 10758163 10.1073/pnas.080079397
Lamartina S Sporeno E Fattori E Toniatti C Characteristics of the adeno-associated virus preintegration site in human chromosome 19: open chromatin conformation and transcription-competent environment J Virol 2000 74 7671 7677 10906224 10.1128/JVI.74.16.7671-7677.2000
Ogata T Kozuka T Kanda T Identification of an insulator in AAVS1, a preferred region for integration of adeno-associated virus DNA J Virol 2003 77 9000 9007 12885916 10.1128/JVI.77.16.9000-9007.2003
Owens RA Second generation adeno-associated virus type 2-based gene therapy systems with the potential for preferential integration into AAVS1 Curr Gene Ther 2002 2 145 159 12109212
Linden RM Winocour E Berns KI The recombination signals for adeno-associated virus site-specific integration Proc Natl Acad Sci USA 1996 93 7966 7972 8755586 10.1073/pnas.93.15.7966
Young SM JrSamulski RJ Adeno-associated virus (AAV) site-specific recombination does not require a Rep-dependent origin of replication within the AAV terminal repeat Proc Natl Acad Sci USA 2001 98 13525 13530 11707592 10.1073/pnas.241508998
Philpott NJ Giraud-Wali C Dupuis C Gomos J Hamilton H Berns KI Falck-Pedersen E Efficient integration of recombinant adeno-associated virus DNA vectors requires a p5-rep sequence in cis J Virol 2002 76 5411 5421 11991970 10.1128/JVI.76.11.5411-5421.2002
Philpott NJ Gomos J Berns KI Falck-Pedersen E A p5 integration efficiency element mediates Rep-dependent integration into AAVS1 at chromosome 19 Proc Natl Acad Sci USA 2002 99 12381 12385 12221283 10.1073/pnas.182430299
Urcelay E Ward P Wiener SM Safer B Kotin RM Asymmetric replication in vitro from a human sequence element is dependent on adeno-associated virus Rep protein J Virol 1995 69 2038 2046 7884849
Young SM JrMcCarty DM Degtyareva N Samulski RJ Roles of adeno-associated virus Rep protein and human chromosome 19 in site-specific recombination J Virol 2000 74 3953 3966 10756007 10.1128/JVI.74.9.3953-3966.2000
Meneses P Berns KI Winocour E DNA sequence motifs which direct adeno-associated virus site-specific integration in a model system J Virol 2000 74 6213 6216 10846109 10.1128/JVI.74.13.6213-6216.2000
Wonderling RS Owens RA Binding sites for adeno-associated virus Rep proteins within the human genome J Virol 1997 71 2528 2534 9032395
Weitzman MD Kyöstiö SRM Kotin RM Owens RA Adeno-associated virus (AAV) Rep proteins mediate complex formation between AAV DNA and its integration site in human DNA Proc Natl Acad Sci USA 1994 91 5808 5812 8016070
Hamilton H Gomos J Berns KI Falck-Pedersen E Adeno-associated virus site-specific integration and AAVS1 disruption J Virol 2004 78 7874 7882 15254160 10.1128/JVI.78.15.7874-7882.2004
Yalkinoglu AO Heilbronn R Burkle A Schlehofer JR zur Hausen H DNA amplification of adeno-associated virus as a response to cellular genotoxic stress Cancer Res 1988 48 3123 3129 2835153
Meyers C Mane M Kokorina N Alam S Hermonatt PL Ubiquitous human adeno-associated virus type 2 autonomously replicates in differentiating keratinocytes of a normal skin model Virology 2000 272 338 346 10873777 10.1006/viro.2000.0385
Hermonat PL Muzyczka N Use of adeno-associated virus as a mammalian DNA cloning vector: transduction of neomycin resistance into mammalian tissue cultured cells Proc Natl Acad Sci USA 1984 81 6466 6470 6093102
Tratschin JD West MH Sandbank T Carter BJ A human parvovirus, adeno-associated virus, as a eukaryotic vector: transient expression and encapsidation of the prokaryotic gene for chloramphenicol acetyltransferase Mol Cell Biol 1984 4 2072 2081 6095038
Samulski RJ Berns KI Tan M Muzyczka N Cloning of adeno-associated virus into pBR322: rescue of intact virus from the recombinant plasmid in human cells Proc Natl Acad Sci USA 1982 79 2077 2081 6281795
Grimm D Kern A Rittner K Kleinschmidt JA Novel tools for production of recombinant adenoassociated virus vectors Hum Gene Ther 1998 9 2745 2760 9874273
Matsushita T Elliger S Elliger C Podsakoff G Villarreal L Kurtzman GJ Iwaki Y Colosi P Adeno-associated virus vectors can be efficiently produced without helper virus Gene Ther 1998 5 938 945 9813665 10.1038/sj.gt.3300680
Xiao X Li J Samulski RJ Production of high-titer recombinant adeno-associated virus vectors in the absence of helper adenovirus J Virol 1998 72 2224 2232 9499080
Li J Samulski RJ Xiao X Role for highly regulated rep gene expression in adeno-associated virus vector production J Virol 1997 71 5236 5243 9188591
Saudan P Vlach J Beard P Inhibition of S-phase progression by adeno-associated virus Rep78 protein is mediated by hypophosphorylated pRb EMBO J 2000 19 4351 4361 10944118 10.1093/emboj/19.16.4351
Qiao C Wang B Zhu X Li J Xiao X A novel gene expression control system and its use in stable high-titer 293 cell-based adeno-associated virus packaging cell lines J Virol 2002 76 13015 13027 12438627 10.1128/JVI.76.24.13015-13027.2002
Gao GP Qu G Faust LZ Engdahl RK Xiao W Hughes JV Zoltick PW Wilson JM High-titer adeno-associated viral vectors from a Rep/Cap cell line and hybrid shuttle virus Hum Gene Ther 1998 9 2353 2362 9829534
Chadeuf G Favre D Tessier J Provost N Nony P Kleinschmidt J Moullier P Salvetti A Efficient recombinant adeno-associated virus production by a stable rep-cap HeLa cell line correlates with adenovirus-induced amplification of the integrated rep-cap genome J Gene Med 2000 2 260 268 10953917 10.1002/1521-2254(200007/08)2:4<260::AID-JGM111>3.0.CO;2-8
Conway JE Rhys CM Zolotukhin I Zolotukhin S Muzyczka N Hayward GS Byrne BJ High-titer recombinant adeno-associated virus production utilizing a recombinant herpes simplex virus type I vector expressing AAV-2 Rep and Cap Gene Ther 1999 6 986 993 10455400 10.1038/sj.gt.3300937
Urabe M Ding C Kotin RM Insect cells as a factory to produce adeno-associated virus type 2 vectors Hum Gene Ther 2002 13 1935 1943 12427305 10.1089/10430340260355347
Auricchio A Hildinger M O'Connor E Gao GP Wilson JM Isolation of highly infectious and pure adeno-associated virus type 2 vectors with a single-step gravity-flow column Hum Gene Ther 2001 12 71 76 11177544 10.1089/104303401450988
Auricchio A O'Connor E Hildinger M Wilson JM A single-step affinity column for purification of serotype-5 based adeno-associated viral vectors Mol Ther 2001 4 372 374 11592841 10.1006/mthe.2001.0462
Kaludov N Handelman B Chiorini JA Scalable purification of adeno-associated virus type 2, 4, or 5 using ion-exchange chromatography Hum Gene Ther 2002 13 1235 1243 12133276 10.1089/104303402320139014
Hauck B Zhao W High K Xiao W Intracellular viral processing, not single-stranded DNA accumulation, is crucial for recombinant adeno-associated virus transduction J Virol 2004 78 13678 13686 15564477 10.1128/JVI.78.24.13678-13686.2004
Thomas CE Storm TA Huang Z Kay MA Rapid uncoating of vector genomes is the key to efficient liver transduction with pseudotyped adeno-associated virus vectors J Virol 2004 78 3110 3122 14990730 10.1128/JVI.78.6.3110-3122.2004
Ferrari FK Samulski T Shenk T Samulski RJ Second-strand synthesis is a rate-limiting step for efficient transduction by recombinant adeno-associated virus vectors.1996 J Virol 1996 70 3227 3234 8627803
Fisher KJ Gao GP Weitzman MD DeMatteo R Burda JF Wilson JM Transduction with recombinant adeno-associated virus for gene therapy is limited by leading-strand synthesis J Virol 1996 70 520 532 8523565
Nakai H Storm TA Kay MA Recruitment of single-stranded recombinant adeno-associated virus vector genomes and intermolecular recombination are responsible for stable transduction of liver in vivo J Virol 2000 74 9451 9463 11000214 10.1128/JVI.74.20.9451-9463.2000
Rabinowitz JE Samulski RJ Building a better vector: the manipulation of AAV virions Virology 2000 278 301 308 11118354 10.1006/viro.2000.0707
Grimm D Kay MA From virus evolution to vector revolution: use of naturally occurring serotypes of adeno-associated virus (AAV) as novel vectors for human gene therapy Curr Gene Ther 2003 3 281 304 12871018
Gregorevic P Blankinship MJ Allen JM Crawford RW Meuse L Miller DG Russell DW Chamberlain JS Systemic delivery of genes to striated muscles using adeno-associated viral vectors Nat Med 2004 10 828 834 15273747 10.1038/nm1085
Nakai H Fuess S Storm TA Muramatsu S-i Nara Y Kay MA Unrestricted hepatocyte transduction with adeno-associated virus serotype 8 vectors in mice J Virol 2005 79 214 224 15596817 10.1128/JVI.79.1.214-224.2005
Buning H Nicklin SA Perabo L Hallek M Baker AH AAV-based gene transfer Curr Opin Mol Ther 2003 5 367 375 14513679
Qing K Hansen J Weigel-Kelley KA Tan M Zhou S Srivastava A Adeno-associated virus type 2-mediated gene transfer: role of cellular FKBP52 protein in transgene expression J Virol 2001 75 8968 8976 11533160 10.1128/JVI.75.19.8968-8976.2001
Zhong L Li W Yang Z Chen L Li Y Qing K Weigel-Kelley KA Yoder MC Shou W Srivastava A Improved transduction of primary murine hepatocytes by recombinant adeno-associated virus 2 vectors in vivo Gene Ther 2004 11 1165 1169 15164097 10.1038/sj.gt.3302283
Yang J Zhou W Zhang Y Zidon T Ritchie T Engelhardt JF Concatamerization of adeno-associated virus circular genomes occurs through intermolecular recombination J Virol 1999 73 9468 9477 10516055
Song S Laipis PJ Berns KI Flotte TR Effect of DNA-dependent protein kinase on the molecular fate of the rAAV2 genome in skeletal muscle Proc Natl Acad Sci USA 2001 98 4084 4088 11274433 10.1073/pnas.061014598
Nakai H Storm TA Fuess S Kay MA Pathways of removal of free DNA vector ends in normal and DNA-PKcs-deficient SCID mouse hepatocytes transduced with rAAV vectors Hum Gene Ther 2003 14 871 881 12828858 10.1089/104303403765701169
Duan D Yue Y Engelhardt JF Consequences of DNA-dependent protein kinase catalytic subunit deficiency on recombinant adeno-associated virus genome circularization and heterodimerization in muscle tissue J Virol 2003 77 4751 4759 12663782 10.1128/JVI.77.8.4751-4759.2003
Afione SA Conrad CK Kearns WG Chunduru S Adams R Reynolds TC Guggino WB Cutting GR Carter BJ Flotte TR In vivo model of adeno-associated virus vector persistence and rescue J Virol 1996 70 3235 3241 8627804
Nakai H Yant SR Storm TA Fuess S Meuse L Kay MA Extrachromosomal recombinant adeno-associated virus vector genomes are primarily responsible for stable liver transduction in vivo J Virol 2001 75 6969 6976 11435577 10.1128/JVI.75.15.6969-6976.2001
Nakai H Montini E Fuess S Storm TA Grompe M Kay MA AAV serotype 2 vectors preferentially integrate into active genes in mice Nat Genet 2003 34 297 302 12778174 10.1038/ng1179
Alexander IE Russell DW Miller AD DNA-damaging agents greatly increase the transduction of nondividing cells by adeno-associated virus vectors J Virol 1994 68 8282 8287 7966621
Sanlioglu S Benson P Engelhardt JF Loss of ATM function enhances adeno-associated virus transduction and integration through pathways similar to UV irradiation Virology 2000 268 68 78 10683328 10.1006/viro.1999.0137
Zentilin L Marcello A Giacca M Involvement of cellular double-stranded DNA break binding proteins in processing of the recombinant adeno-associated virus genome J Virol 2001 75 12279 12287 11711618 10.1128/JVI.75.24.12279-12287.2001
Miller DG Petek LM Russell DW Adeno-associated virus vectors integrate at chromosome breakage sites Nat Genet 2004 36 767 773 15208627 10.1038/ng1380
Dong JY Fan PD Frizzell RA Quantitative analysis of the packaging capacity of recombinant adeno-associated virus Hum Gene Ther 1996 7 2101 2112 8934224
Yan Z Ritchie TC Duan D Engelhardt JF Phillips MI Recombinant AAV-mediated gene delivery using dual vector heterodimerization Methods in Enzymology 2002 San Diego: Academic Press 334 357 [Abelson JN and Simon MI (Series Editors): Gene Therapy Methods, vol 346.] 11883078
Reich SJ Auricchio A Hildinger M Glover E Maguire AM Wilson JM Bennett J Efficient trans-splicing in the retina expands the utility of adeno-associated virus as a vector for gene therapy Hum Gene Ther 2003 14 37 44 12573057 10.1089/10430340360464697
Yan Z Zak R Zhang Y Engelhardt JF Inverted terminal repeat sequences are important for intermolecular recombination and circularization of adeno-associated virus genomes J Virol 2005 79 364 379 15596830 10.1128/JVI.79.1.364-379.2005
McCarty DM Monahan PE Samulski RJ Self-complementary recombinant adeno-associated virus (scAAV) vectors promote efficient transduction independently of DNA synthesis Gene Ther 2001 8 1248 1254 11509958 10.1038/sj.gt.3301514
De la Maza LM Carter BJ Molecular structure of adeno-associated virus variant DNA J Biol Chem 1980 255 3194 3203 6244311
McCarty DM Fu H Monahan PE Toulson CE Naik P Samulski RJ Adeno-associated virus terminal repeat (TR) mutant generates self-complementary vectors to overcome the rate-limiting step to transduction in vivo Gene Ther 2003 10 2112 2118 14625565 10.1038/sj.gt.3302134
Fisher KJ Kelley WM Burda JF Wilson JM A novel adenovirus-adeno-associated virus hybrid vector that displays efficient rescue and delivery of the AAV genome Hum Gene Ther 1996 7 2079 2087 8934222
Lieber A Steinwaerder DS Carlson CA Kay MA Integrating adenovirus-adeno-associated virus hybrid vectors devoid of all viral genes J Virol 1999 73 9314 9324 10516040
Recchia A Parks RJ Lamartina S Toniatti C Pieroni L Palombo F Ciliberto G Graham FL Cortese R La Monica N Colloca S Site-specific integration mediated by a hybrid adenovirus/adeno-associated virus vector Proc Natl Acad Sci USA 1999 96 2615 2620 10077559 10.1073/pnas.96.6.2615
Gonçalves MA Pau MG de Vries AA Valerio D Generation of a high-capacity hybrid vector: packaging of recombinant adenoassociated virus replicative intermediates in adenovirus capsids overcomes the limited cloning capacity of adenoassociated virus vectors Virology 2001 288 236 246 11601895 10.1006/viro.2001.1073
Palombo F Monciotti A Recchia A Cortese R Ciliberto G La Monica N Site-specific integration in mammalian cells mediated by a new hybrid baculovirus-adeno-associated virus vector J Virol 1998 72 5025 5034 9573272
Johnston KM Jacoby D Pechan PA Fraefel C Borghesani P Schuback D Dunn RJ Smith FI Breakefield XO HSV/AAV hybrid amplicon vectors extend transgene expression in human glioma cells Hum Gene Ther 1997 8 359 370 9048203
Capecchi MR Altering the genome by homologous recombination Science 1989 244 1288 1292 2660260
Russell DW Hirata RK Human gene targeting by viral vectors Nat Genet 1998 18 325 330 9537413 10.1038/ng0498-325
Chamberlain JR Schwarze U Wang P-P Hirata RK Hankenson KD Pace JM Underwood RA Song KM Sussman M Byers PH Russell DW Gene targeting in stem cells from individuals with osteogenesis imperfecta Science 2004 303 1198 1201 14976317 10.1126/science.1088757
Flotte TR Carter B Conrad C Guggino W Reynolds T Rosenstein B Taylor G Walden S Wetzel R A phase I study of an adeno-associated virus-CFTR gene vector in adult CF patients with mild lung disease Hum Gene Ther 1996 7 1145 1159 8773517
Guggino WB Cystic fibrosis and the salt controversy Cell 1999 96 607 610 10089875 10.1016/S0092-8674(00)80570-X
Flotte TR Afione SA Conrad C McGrath SA Solow R Oks H Zeitlin PL Guggino WB Carter BJ Stable in vivo expression of the cystic fibrosis transmembrane conductance regulator with an adeno-associated virus vector Proc Natl Acad Sci USA 1993 90 10613 10617 7504271
Conrad CK Allen SS Afione SA Reynolds TC Beck SE Fee-Maki M Barrazza-Ortiz X Adams R Askin FB Carter BJ Guggino WB Flotte TR Safety of single-dose administration of an adeno-associated virus (AAV)-CFTR vector in the primate lung Gene Ther 1996 3 658 668 8854091
Aitken ML Moss RB Waltz DA Dovey ME Tonelli MR McNamara SC Ginson RL Ramsey BW Carter BJ Reynolds TC A phase I study of aerosolized administration of tgAAVCF to cystic fibrosis subjects with mild lung disease Gene Ther 2001 12 1907 1916 10.1089/104303401753153956
Moss RB Rodman D Spencer LT Aitken ML Zeitlin PL Waltz D Milla C Brody AS Clancy JP Ramsey B Hamblett N Heald AE Repeated adeno-associated virus serotype 2 aerosol-mediated cystic fibrosis transmembrane regulator gene transfer to the lungs of patients with cystic fibrosis: a multicenter, double-blind, placebo-controlled trial Chest 2004 125 509 521 14769732 10.1378/chest.125.2.509
Snyder RO Miao C Meuse L Tubb J Donahue BA Lin H-F Stafford DW Patel S Thompson AR Nichols T Read MS Bellinger DA Brinkhous KM Kay MA Correction of hemophilia B in canine and murine models using recombinant adeno-associated viral vectors Nat Med 1999 5 64 70 9883841 10.1038/13518
Herzog RW Yang EY Couto LB Hagstrom JN Elwell D Fields PA Burton M Bellinger DA Read MS Brinhous KM Podsakoff GM Nichols TC Kurtzman GJ High KA Long-term correction of canine hemophilia B by gene transfer of blood coagulation factor IX mediated by adeno-associated viral vector Nat Med 1999 5 56 63 9883840 10.1038/4743
Manno CS Chew AJ Hutchison S Larson PJ Herzog RW Arruda VR Tal SJ Ragni MV Thompson A Ozello M LB Couto Leonard DG Johnson FA McClelland A Scallan C Skarsgard E Flake AW Kay MA High KA Glader B AAV-mediated factor IX gene transfer to skeletal muscle in patients with severe hemophilia B Blood 2003 101 2963 2972 12515715 10.1182/blood-2002-10-3296
Arruda VR Hagstrom JN Deitch J Heiman-Patterson T Camine RM Chu K Fields PA Herzog RW Couto LB Larson PJ High KA Posttranslational modifications of recombinant myotube-synthesized human factor IX Blood 2001 97 130 138 11133752 10.1182/blood.V97.1.130
High KA Manno CS Sabatino DE Hutchison S Dake M Razavi M Immune responses to AAV and to factor IX in a phase I study of AAV-mediated, liver-directed gene transfer for hemophilia B [abstract] Mol Ther 2004 5 s1002
Wang L Calcedo R Nichols TC Bellinger DA Dillow A Verma IM Wilson JM Sustained correction of disease in naïve and AAV2-pretreated hemophilia B dogs: AAV2/8-mediated, liver-directed gene therapy Blood 2005 105 3079 3086 15637142 10.1182/blood-2004-10-3867
Gao GP Alvira MR Wang L Calcedo R Johnston J Wilson JM Novel adeno-associated viruses fom rhesus monkeys as vectors for human gene therapy Proc Natl Acad Sci USA 2002 99 11854 11859 12192090 10.1073/pnas.182412299
Ryan JH Zolotukhin S Muzyczka N Sequence requirements for binding of Rep68 to the adeno-associated terminal repeats J Virol 1996 70 1542 1553 8627673
Snyder RO Im DS Ni T Xiao X Samulski RJ Muzyczka N Features of the adeno-associated virus origin involved in substrate recognition by the viral Rep protein J Virol 1993 67 6096 6104 8396670
Brister JR Muzyczka N Mechanism of Rep-mediated adeno-associated virus origin nicking J Virol 2000 74 7762 7771 10933682 10.1128/JVI.74.17.7762-7771.2000
| 15877812 | PMC1131931 | CC BY | 2021-01-04 16:38:58 | no | Virol J. 2005 May 6; 2:43 | utf-8 | Virol J | 2,005 | 10.1186/1743-422X-2-43 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-611587173710.1186/1471-2164-6-61Research ArticleComparison of standard exponential and linear techniques to amplify small cDNA samples for microarrays Wadenbäck Johan [email protected] David H [email protected] Deborah [email protected] Ronald [email protected] Gary F [email protected] Arnold Sara [email protected] Ulrika [email protected] Department of Plant Biology and Forest Genetics, Swedish University of Agricultural Sciences, P.O. Box 7080, Uppsala, Sweden2 Department of Forestry, North Carolina State University, Raleigh, NC 27695, USA3 School of Forest Resources and Conservation, University of Florida, Gainesville, FL 32611, USA4 Department of Forestry, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA2005 4 5 2005 6 61 61 29 10 2004 4 5 2005 Copyright © 2005 Wadenbäck et al; licensee BioMed Central Ltd.2005Wadenbäck 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 need to perform microarray experiments with small amounts of tissue has led to the development of several protocols for amplifying the target transcripts. The use of different amplification protocols could affect the comparability of microarray experiments.
Results
Here we compare expression data from Pinus taeda cDNA microarrays using transcripts amplified either exponentially by PCR or linearly by T7 transcription. The amplified transcripts vary significantly in estimated length, GC content and expression depending on amplification technique. Amplification by T7 RNA polymerase gives transcripts with a greater range of lengths, greater estimated mean length, and greater variation of expression levels, but lower average GC content, than those from PCR amplification. For genes with significantly higher expression after T7 transcription than after PCR, the transcripts were 27% longer and had about 2 percentage units lower GC content. The correlation of expression intensities between technical repeats was high for both methods (R2 = 0.98) whereas the correlation of expression intensities using the different methods was considerably lower (R2 = 0.52). Correlation of expression intensities between amplified and unamplified transcripts were intermediate (R2 = 0.68–0.77).
Conclusion
Amplification with T7 transcription better reflects the variation of the unamplified transcriptome than PCR based methods owing to the better representation of long transcripts. If transcripts of particular interest are known to have high GC content and are of limited length, however, PCR-based methods may be preferable.
==== Body
Background
The analysis of transcript abundance in samples of total RNA using standard techniques such as northern blotting or microarrays requires microgram quantities of total RNA. In our experience, a microarray analysis incorporating a loop design and reciprocal labeling with Cy™3 and Cy™5 dyes, requires 80 micrograms of total RNA per sample [1]. It is often inconvenient or impossible to obtain sufficient quantities without an amplification step, particularly if tissue sections are to be analyzed. Exponential amplification of cDNA by a standard PCR procedure [2] may result in the differential amplification of particular transcripts, since sequences differ in the rate with which they can be amplified by PCR [3]. To minimize this problem, the sequences to be amplified can be limited to about 300 nucleotides at the 3'-terminus of the cDNA; this can be achieved by ultrasound treatment [4] or by limiting the concentration of deoxynucleotides in the PCR reaction mixture [5]. These methods are promising but not yet in standard use. An alternative approach is linear amplification by in vitro transcription from a strong promoter such as a T7 phage promoter [6]. Linear amplification has been shown to retain the relative frequencies of transcripts with reasonable fidelity over a wide amplification range [7-10]. Many aspects of the high efficiency and reliability of linear and exponential amplification methods have been studied earlier. These deal mainly with comparisons between unamplified and amplified material and indirectly between different amplification methods [11,12]. The distorting effect in mRNA abundance of linear and exponential amplification techniques in relation to the sequence and GC content of the genes has been hypothesized [12], but very little evidence has been put forward to support this in relation to the characteristics of individual transcripts.
Commercial kits are available for both exponential (Super SMART™ from BD Biosciences Clontech) and linear (Message Amp™ from Ambion) amplification. Here we compare expression levels determined with cDNA microarrays hybridized with cDNA obtained from thin sections of secondary xylem tissue from Pinus taeda amplified by these two different strategies, or after using unamplified cDNA for hybridization. We have addressed the questions: how well do the results agree with each other in a direct comparison? What are the characteristics of the sequences showing preferential amplification by exponential or linear amplification?
Results and Discussion
Comparison of unamplified- and amplified targets
A typical sample for amplified Super SMART™ PCR-product yields a distribution of sizes from 500 bp-6000 bp with a peak centered at 900 bp (Clontech, Super Smart PCR cDNA Synthesis Kit User manual). A typical sample for amplified Message Amp™ aRNA yields a distribution of sizes from 250 nt-5500 nt with a peak centered at 1000–1500 nt (Ambion, Catalog #1752). The distributions of our amplified material agree well with the manufacturers' data (See Additional file 1 and 2). It has been reported that PCR amplification requires less RNA, is more reproducible and generates better target transcripts than linear amplification [5,13], at least if the sequences are limited to the 3'- end. Linear T7 amplification has however been widely used when starting material is limiting. Recently some researchers have reported bias in their data. In some studies the bias is said to be of minor importance, systematic and reproducible, affecting all the samples in the same way and therefore potentially controllable in the normalization (e.g. to calculate fold change) [10,14]. In other studies the bias from different amplification protocol is affecting the general ratios of gene expression [5,12]. Part of the bias may arise from the T7 RNA polymerase's intrinsic nucleolytic activity that appears during extended incubation [15]. Other bias is maybe introduced owing to the characteristics of the individual transcripts.
We have found a preferential amplification of certain nucleotide sequences by the Super SMART™ PCR relative to a nonamplified target in earlier membrane array experiments, where the targets were prepared from the samples of lignified planings and nonlignified xylem scrapings (data not shown). The correlation (R2 ) between transcript abundance using unamplified and Super SMART™ PCR amplified targets was 0.77 for scrapings and 0.68 for planings.
Comparing five lines of Picea abies shoots where the first biological replicate consisted of unamplified targets and the second biological replicate consisted of targets amplified with T7 transcription we obtained a correlation of R2 = 0.74 (data not shown). Ambion has reported R2 = 0.87 [16] between technical repeats.
Plots of the individual gene transcript abundance of unamplified versus amplified target should give a straight line of slope 1 if the overall expression is preserved. However, there is some nonlinear behavior in both cases. For unamplified versus PCR amplified target the curve is generally nonlinear and lower abundance transcripts are under-represented and highly expressed transcripts are amplified better than average. For the unamplified versus T7 amplified targets a very small minority of highly expressed transcripts do not follow the linear slope of around 1.
For the comparison between unamplified and PCR amplified targets the 95% confidence intervals for the fold-changes were as follows: For unamplified material: Downregulation, 2.3–8.0; upregulation: 1.1–1.3. For PCR-amplified material: Downregulation, 1.3–1.7; upregulation, 1.0–3.0. The differences between unamplified and PCR amplified targets were statistically significant.
For the highly significant (p < 0.0001) differentially expressed genes between lines in each of the ten comparisons of unamplified and T7 amplified targets, the 95% confidence intervals for the fold-changes were as follows. For unamplified material: Downregulation, 1.5–3.2 (all), and 2.7–7.5 (top); upregulation: 1.3–2.8 (all), and 2.6–4.4 (top). For T7-amplified material: Downregulation, 1.0–2.7 (all), and 2.5–4.2 (top); upregulation: 1.2–3.0 (all), and 2.3–4.3 (top). The differences between unamplified and T7 amplified targets are generally not statistically significant although the fold change for the unamplified targets were greater than for the T7 amplified targets indicating that some small bias may still exist when using T7 amplified relative to unamplified targets, especially for highly expressed transcripts.
However, in many situations there is no possibilty of using unamplified targets and amplification is required. Thus, starting with small amounts of secondary xylem tissue we compared PCR and T7 RNA polymerase amplification methods directly to investigate if, and how, the biases differ from each other.
Expression characteristics of transcripts amplified by PCR or T7 transcription
The two methods of amplification were compared to each other four times and twice to themselves in a fully balanced flip dye experimental design including technical repeats (Figure 1A). Only few spots were flagged as bad and excluded from further analysis. The percentage of detectable spots (above background) on each array and in each channel was 88% using T7 amplification and 71% using PCR amplification. The percentage of saturated spots was around 1% in all cases.
Figure 1 Global comparison of PCR and T7 amplification techniques. (A) Microarray experimental design. S1 and S2, and M1 and M2 are technical repeats of PCR (S) and T7 amplification (M) respectively. Each arrow represents one slide where the sample at the base of the arrow is labeled with Cy™3 and the sample at the tip of the arrow is labeled with Cy™5. (B-D) Correlation of results within and between amplification techniques; the values are least square means of expression of each of the genes represented on the array. (B) The correlation between two technical repeats of gene expression after S amplification. (C) The correlation between two technical repeats of gene expression after M amplification. (D) The correlation between gene expression after S and M amplification. Each amplification method produces highly consistent results (R2 = 0.98) whereas the correlation of the results given by the two different methods is considerably lower (R2 = 0.52) indicating bias in one or both amplification techniques.
After normalization the correlation of transcript abundance for each gene between technical repeats was very high, R2 = 0.98, after both PCR- (Figure 1B) and T7 amplification (Figure 1C). In contrast, the correlation between the two different amplification methods for both technical repeats was considerably lower, R2 = 0.52, (Figure 1D), indicating bias in one or both amplification techniques. As previously mentioned the correlation between unamplified and amplified transcript abundance was intermediate, indicating that both amplification methods have bias and that these biases are different from each other.
The genes present on the microarray were divided into two groups according to whether the PCR amplified transcripts (S') or the T7 amplified transcripts (M') were more abundant. The S' group was 9% larger than the M' group.
A relative frequency distribution plot of expression levels revealed a narrower peak for S' than for M' transcripts (Figure 2A). The arithmetic expression values showed a significantly greater mean for M' (1.76) than for S' (1.64) and a higher variance although the coefficient of variation was lower for M' (81.6%) than for S' (86.6%). The distribution of the data implies a broader population of transcript species present in the T7 amplified target.
Figure 2 Statistical analysis of microarray targets. Characteristics of genes (represented by Pinus taeda ESTs) showing preferential amplification by one method or the other. From the results of the microarray normalization, the genes were divided into two groups, those showing higher expression for PCR amplified transcripts (S') and those showing higher expression for T7 amplified transcripts (M'). (A) Distribution of expression for all the genes (2190 ESTs) in the S' and M' groups. (B) Transcript abundance of all the genes and selected genes (represented by 309 ESTs) in the S' and M' group. (C) Transcript lengths for the two groups, estimated by finding the Arabidopsis thaliana homologs either from the nucleotide sequence (BLASTn™) or the amino acid sequence (BLASTx™) of the Pinus taeda contigs. The variance of the transcript length is significantly smaller for the S' group than for the M' group for both the nucleotide and protein estimates. There is furthermore a significantly greater mean length for the M' group than for the S' group. (D) GC content sequenced ends of the selected genes of the S' and M' groups. The S' group is significantly more GC rich than the M' group, for both ESTs and contigs. Bars indicate the range; boxes extend from the 25th to the 75th percentile, with a horizontal line at the median.
Using the criteria for statistical significance described in methods, 309 ESTs (14%) showed different expression levels between the two amplification methods with 131 ESTs in the S' group and 178 ESTs in the M' group. The arithmetic mean of the S' group (3.40) was statistically higher than the M' group (2.95) and the S' group had higher variance (Figure 2B). The coefficient of variation was lower for M' (33.4%) than for S' (36.3%). The reason for the opposite trend observed for this subset of genes may reflect the differences in detectable spots and the amplification kinetics between PCR and T7 transcription.
Transcript characteristics amplified by PCR or T7 transcription
As shown above, out of the genes (309 ESTs) showing statistically significant abundance differences between the amplification methods, 36% more were found in the M' group than in the S' group. One possibility for why 36% more were found in the M' group is that the complexity of the T7 amplified transcripts is greater. To assess this we analyzed the length of the sequences on the array. Previous analyses of protein sequences showed about half of Pinus taeda ESTs on the array have an apparent homolog in Arabidopsis thaliana (increasing with length up to 90%). For these ESTs the sequence similarity is typically distributed over the full length of the contig indicating a substantial conservation of genes between these two species, suggesting a common functional genome [17]. From the BLASTn™ (nucleotide level) and BLASTx™ (amino acid level) searches relating the contig data to Arabidopsis thaliana homologs, the corresponding Pinus taeda full-length cDNAs were estimated. The contig lengths constitute on average about 45 % of the total cDNA lengths spotted on the array. For both the nucleotide and the amino acid levels there was a highly significant 60% greater variance in length of the M' group than of the S' group. At the amino acid level there was a significant 26.9% greater mean length of the M' group (1580 bp) than the S' counterpart (1245 bp). The maximum length of transcript present was also considerably greater in the M' group than the S' group (Figure 2C). In contrast to the contigs the singleton ESTs in the S' group (482 bp) had a significantly greater mean sequence length than those in the M' group (428 bp). The reason for this discrepancy is unclear but could reflect a difference in efficiency of the sequencing polymerase resulting from difference in the amount of secondary structures in the sequences from the two sets. The M' group contained 60% of the ESTs and contigs with nucleotide and amino acid homology to Arabidopsis thaliana reflecting both an initially greater transcript population as well as differences in transcript lengths. In conclusion, the possibility of getting transcripts of greater length and larger variability is considerably higher when using T7 amplification rather than PCR amplification.
Importance of GC content for amplification
Comparison of the selected genes (309 ESTs) differentially represented in the two amplification methods, the GC content of the ESTs, contigs and Arabidopsis thaliana cDNAs (on a nucleotide level) there was a significantly greater mean GC content for the sequences of the S' group than for those of the M' group. The difference was 2.7 percentage units for ESTs, and 1.4 percentage units for the corresponding contigs (Figure 2D). There was a similar difference for the cDNAs although only about 10% of the contigs were found to have a BLASTn™ score above 100 bits. Interestingly, for a smaller group of 80 contigs (40 from S' and 40 from M') showing the greatest fold changes between methods, the difference in GC content increased from 1.4 to 2.2 percentage units, due to an increase in GC content for the S' group. Additionally, the mean length of the 40 ESTs from the S' group (1428 bp) was significantly greater than the mean length of the 40 ESTs from the M' group (1275 bp). It appears that transcripts with a high GC content are amplified faster by PCR than by T7, often overriding the effect of length. If the GC content is nearer the average, long transcripts are favored by T7 amplification. The GC effect is presumably explained by the temperature of extension, which is 68–72°C for Taq polymerase and 37°C for T7 polymerase; high temperature favors polymerization through GC-rich areas. Evolution has in general tuned the cellular machinery, including polymerases, to fit the temperature environment of an organism. This might be reflected in the GC content and the temperature environment of the original organism for each polymerase. The GC content of a Pinus species genome is about 40%, which is considerably closer to the 48% GC content of T7 phage (or the 50% GC content of Escherichia coli, the typical host of T7 phage), than for the 67% GC content of Thermus aquaticus [18-20]. It implies that T7 transcription of the Pinus taeda transcriptome or consequently other transcriptomes with similar GC content in most cases is a better choice than PCR based techniques.
Conclusion
In summary, the two main approaches to amplification of small amounts of RNA for microarray studies, PCR and T7 transcription both introduce bias compared to the unamplified target and the nature of the bias is different for each method. Our results show that amplification by T7 RNA polymerase gives transcripts with a greater range of lengths, greater estimated mean length, and greater variation of expression levels, but lower average GC content, than those from PCR amplification. Amplification with T7 transcription would therefore better reflect the variation of the unamplified Pinus taeda transcriptome and other comparable transcriptomes than PCR based methods. If transcripts of particular interest are known to have high GC content and are of limited size, however, PCR based methods may be preferable. The results demonstrate the need to pay attention to possible biases introduced by the amplification methods and that in certain projects different amplification techniques should be tested and optimized before routine use.
Methods
Target extraction and amplification
Polyadenylated RNA was extracted from individual 30 μm cryotome sections (3 mm × 3 mm) through the cambial region of Pinus taeda L. using Dynabeads (Dynal Biotech, Oslo, Norway). The mRNA samples were reverse-transcribed and the resulting cDNA were amplified by a) exponential PCR amplification by Super SMART™ (BD Biosciences Clontech, Palo Alto, CA, USA), or b) linear T7 amplification through Message Amp™ aRNA kit (Ambion, Austin, TX, USA). The Super SMART™ cDNA products were directly labeled by Klenow with Cy™3 or Cy™5 dUTPs (Amersham Biosciences, Piscataway, NJ, USA). The Message Amp™ aRNA products were reverse-transcribed with aminoallyl-modified dUTPs (Sigma, St. Louis, MO, USA) and labeled by coupling to free Cy™3 or Cy™5 dye (Amersham Biosciences) (Table 1).
Table 1 Flow chart of the exponential- and linear amplification techniques with Klenow- and aminoallyl labeling respectively
Exponential Amplification with PCR DNA Polymerase Linear Amplification with T7 RNA Polymerase
Microarray hybridization and probe selection
Microarray hybridization and stringency washes have been described previously [21,22]. cDNA microarrays based on 2190 Pinus taeda ESTs from the NSF unigene set (Forest Biotechnology Group, NCSU, NC, USA) [17] were hybridized with the labeled targets. The PCR and T7 amplification methods were compared in a fully balanced, flip dye design encompassing eight microarray slides (Figure 1A). The microarray data is MIAME compliant [GEO:GPL1880].
Data normalization and analysis
The consistency of each method was assessed by dividing the samples into two technical repeats. The slides were scanned using a ScanArray® 4000 Microarray Analysis System (GSI Lumonics, Ottawa, Canada). Raw intensity values were collected with QuantArray® software (GSI Lumonics) and spots were visually inspected for spot morphology and background. No background subtraction was applied because backgrounds were low and subtraction can introduce bias. The microarray intensity data was normalized using a mixed model system [21,23-25] in SAS/STAT Software version 8 (SAS Institute Inc., Cary, NC, USA). The log2 fold change in abundance was used to divide the selected genes in two groups depending on sign.
The normalized log2 fold change (essentially a ratio of the least square means of Super SMART™- and Message Amp™-amplified transcript abundance derived from the mixed model) with a probability value of p < 0.001 and array- and array*dye interaction variance lower than 0.001 were used to select genes with significant changes in abundance (represented by 309 ESTs). The absolute values (i.e. a rescaling of the data disregarding the sign) of the log2 fold change abundance were then used in the subsequent statistical analysis. The abbreviations used are: S = abundance of Super SMART™-amplified transcripts; M = abundance of Message Amp™-amplified transcripts; S'= [log2(M/S)], S>M; and M' = [log2(M/S)], S<M. In all the comparisons the individual transcripts are represented by ESTs.
The lengths of the cDNAs represented on the microarray were estimated based on full length Arabidopsis thaliana [26] homolog sequences using the Pinus taeda ESTs and contigs [27]. The top Arabidopsis thaliana homolog cDNAs with a score greater than 100 bits were selected for Pinus taeda ESTs or contigs on nucleotide level (using BLASTn™ and the AGI transcripts (-introns, +UTRs) dataset) or amino acid level (using BLASTx™ and the AGI proteins dataset).
All the Pinus taeda ESTs and contigs including those subsets showing homology to Arabidopsis thaliana cDNAs were then analyzed for sequence length, GC content as well as log2 fold change abundance.
The corresponding groups in each subset were analyzed with Prism Software version 3 (GraphPad Software Inc., San Diego, CA, USA). F-tests were used for evaluating a group's compliance with Gaussian distribution. When the normal criteria were met for two groups, one-way ANOVA analysis (with Bonferroni post test) and unpaired t-tests with or without applicable Welch's correction (not assuming equal variances) were performed. When the normal criteria were not met the nonparametric Mann-Whitney test was performed.
Authors' contributions
JW, DHC, GFP and UE carried out the laboratory work. JW, DHC, DC, SvA and UE participated in the normalization and analysis of data. JW, DHC, DC, RS and UE conceived the study, and participated in its design. JW, DHC, RS, GFP, SvA and UE carried out the drafting of the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Size distribution of Super SMART™ amplified cDNAs (1% agarose gel)
Click here for file
Additional File 2
Size distribution of Message Amp™ amplified aRNAs (Electropherogram, LabChip)
Click here for file
Acknowledgements
The work was supported by grants to Sara von Arnold and Ronald Sederoff from the Swedish Foundation for International Cooperation in Research and Higher Education and the US Department of Agriculture (IFAFS Program). This research was also supported by the Florida Agricultural Experimental Station and approved for publication as Journal Series No. R-10840.
==== Refs
Brinker M van Zyl L Liu WB Craig D Sederoff RR Clapham DH von Arnold S Microarray analyses of gene expression during adventitious root development in Pinus contorta Plant Physiology 2004 3 1526 1539 15247392 10.1104/pp.103.032235
Saiki RK Gelfand DH Stoffel S Scharf SJ Higuchi R Horn GT Mullis KB Erlich HA Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase Science 1988 239 487 491 2448875
Lockhart DJ Winzeler EA Genomics, gene expression and DNA arrays Nature 2000 405 827 836 10866209 10.1038/35015701
Hertzberg M Aspeborg H Schrader J Andersson A Erlandsson R Blomqvist K Bhalerao R Uhlen M Teeri TT Lundeberg J Sundberg B Nilsson P Sandberg G A transcriptional roadmap to wood formation PNAS 2001 98 14732 14737 11724959 10.1073/pnas.261293398
Iscove NN Barbara M Gu M Gibson M Modi C Winegarden N Representation is faithfully preserved in global cDNA amplified exponentially from sub-picogram quantities of mRNA Nat Biotechnol 2002 20 940 943 12172558 10.1038/nbt729
van Gelder RN von Zastrow ME Yool A Dement WC Barchas JD Eberwine JH Amplified RNA synthesized from limited quantities of heterogenous cDNA Proc Natl Acad Sci USA 1990 87 1663 1167 1689846
Wang E Miller LD Ohnmacht GA Liu ET Marincola FM High-fidelity mRNA amplification for gene profiling Nature Biotechnology 2000 18 457 459 10748532 10.1038/74546
Baugh LR Hill AA Brown EL Hunter CP Quantitative analysis of mRNA amplification by in vitro transcription Nucleic Acid Res 2001 29 e29 11222780 10.1093/nar/29.5.e29
Gomes LI Silva RL Stolf BS Cristo EB Hirata R Soares FA Reis LF Neves EJ Carvalho AF Comparative analysis of amplified and nonamplified RNA for hybridization in cDNA microarray Analytical Biochemistry 2003 321 244 251 14511690 10.1016/S0003-2697(03)00466-4
Schneider J Buneß A Huber W Volz J Kioschis P Hafner M Poustka A Sültmann H Systematic analysis of T7 RNA polymerase based in vitro linear RNA amplification for use in microarray experiments BMC Genomics 2004 5 29 15119961 10.1186/1471-2164-5-29
Wang J Hu L Hamilton SR Coombes KR Zhang W RNA amplification strategies for cDNA microarray experiments BioTechniques 2003 34 394 400 12613262
Puskás LG Zvara Á Hackler L Van Hummelen P RNA amplification results in reproducible microarray data with slight ratio bias BioTechniques 2002 32 1330 1340 12074164
Klur S Toy K Williams M Certa U Evaluation of procedures for amplification of small-size samples for hybridization on microarrays Genomics 2004 83 508 517 14962677 10.1016/j.ygeno.2003.09.005
Wilson CL Pepper SD Hey Y Miller CJ Amplification protocols introduce systematic but reproducible errors into gene expression studies BioTechniques 2004 36 498 506 15038166
Spiess A-N Muller N Ivell R Amplified RNA degradation in T7-amplification methods results in biased microarray hybridizations BMC Genomics 2003 4 44 14606961 10.1186/1471-2164-4-44
Ambion Technotes 9(3): Microarray Analysis Gene Representation in Amplified vs. Unamplified RNA
Kirst M Johnson AF Baucom C Ulrich E Hubbard K Staggs R Paule C Retzel E Whetten R Sederoff R Apparent homology of expressed genes from wood-forming tissues of loblolly pine (Pinus taeda L.) with Arabidopsis thaliana PNAS 2003 100 7383 7388 12771380 10.1073/pnas.1132171100
Bogunic F Muratovic E Brown SC Siljak-Yakovlev S Genome size and base composition of five Pinus species from the Balkan region Plant Cell Rep 2003 22 59 63 12835994 10.1007/s00299-003-0653-2
Kunisawa T Kanaya S Kutter E Comparison of synonymous codon distribution patterns of bacteriophage and host genomes DNA Res 1998 5 319 326 10048480
Munster MJ Munster AP Woodrow JR Sharp RJ Isolation and preliminary taxonomic studies of Thermus strains isolated from Yellowstone National Park, USA J Gen Microbiol 1986 132 1677 1683 3806053
Stasolla C Belmonte MF van Zyl L Craig DL Liu W Yeung EC Sederoff RR The effect of reduced glutathione on morphology and gene expression of white spruce (Picea glauca) somatic embryos J Exp Bot 2004 55 695 709 14966213 10.1093/jxb/erh074
Hegde P Qi R Abernathy K Gay C Dharap S Gaspard R Hughes JE Snesrud E Lee N Quackenbush J A concise guide to cDNA microarray analysis BioTechniques 2000 29 548 562 10997270
Wolfinger RD Gibson E Wolfinger L Bennett H Hamadeh P Bushel C Afshari C Paules RS Assessing gene significance from cDNA microarray expression data via mixed models J Comput Biol 2001 8 625 637 11747616 10.1089/106652701753307520
Jin W Riley RM Wolfinger RD White KP Passador-Gurgel G Gibson G The contribution of sex, geneotype and age to transcriptional variance in Drosophila melanogaster Nat Genet 2001 29 389 395 11726925 10.1038/ng766
Brazma A Vilo J Gene expression data analysis FEBS Lett 2000 480 17 24 10967323 10.1016/S0014-5793(00)01772-5
The Arabidopsis Information Resource
Genomics of Wood Formation in Loblolly Pine, Nov2003 Pine Contig set
| 15871737 | PMC1134654 | CC BY | 2021-01-04 16:39:53 | no | BMC Genomics. 2005 May 4; 6:61 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-61 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-671588246110.1186/1471-2164-6-67Methodology ArticleComparison of array-based comparative genomic hybridization with gene expression-based regional expression biases to identify genetic abnormalities in hepatocellular carcinoma Furge Kyle A [email protected] Karl J [email protected] Coral [email protected] Xin [email protected] Bioinformatics Special Program, Van Andel Research Institute, 333 Bostwick Ave. NE, Grand Rapids, MI 49503, USA2 Dept of Pharmaceutical Sciences, University of California, S816 513 Parnassus Ave., San Francisco, CA 94143-0046, USA3 Cancer Center, University of California, S816 513 Parnassus Ave., San Francisco, CA 94143-0046, USA4 Liver Center, University of California, S816 513 Parnassus Ave., San Francisco, CA 94143-0046, USA2005 9 5 2005 6 67 67 10 1 2005 9 5 2005 Copyright © 2005 Furge et al; licensee BioMed Central Ltd.2005Furge 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
Regional expression biases (REBs) are genetic intervals where gene expression is coordinately changed. For example, if a region of the genome is amplified, often the majority of genes that map within the amplified region show increased expression when compared to genes located in cytogenetically normal regions. As such, REBs have the potential to act as surrogates for cytogenetic data traditionally obtained using molecular technologies such as comparative genomic hybridization. However as REBs are identified using transcriptional information, detection of REBs may also identify local transcriptional abnormalities produced by both genetic and epigenetic mechanisms.
Results
REBs were identified from a set of hepatocellular carcinoma (HCC) gene expression profiles using a multiple span moving binomial test and compared to genetic abnormalities identified using array-based comparative genomic hybridization (aCGH). In the majority of cases, REBs overlapped genetic abnormalities as determined by aCGH. For example, both methods identified narrow regions of frequent amplification on chromosome 1p and narrow regions of frequent deletion on 17q. In a minority of cases, REBs were identified in regions not determined to be abnormal via other cytogenetic technologies. Specifically, expression biases reflective of cell proliferation were frequently identified on chromosome 6p21-23.
Conclusion
Identification of REBs using a multiple span moving binomial test produced reasonable approximations of underlying cytogenetic abnormalities. However, caution should be used when attributing REBs identified on chromosome 6p to cytogenetic events in rapidly proliferating cells.
==== Body
Background
The parallel analysis of cytogenetic and transcriptional profiling data has revealed that changes in DNA copy number can have noticeable effects on gene expression. Studies comparing wild-type and mutant strains of yeast demonstrated that in regions of increased DNA copy number (i.e. genomic amplifications), the vast majority of genes that mapped within the amplified region had increased expression when compared to gene expression in non-amplified regions [1]. In this context, the unidirectional change in expression of a large number of adjacent genes can be termed a regional expression bias (REB). The dependence of gene expression on DNA copy number has also been observed with human derived samples, for example in a variety of aneuploid tumors and tumor derived cell lines, and in tissues obtained from patients with inherited trisomy disorders [2-13]. In these samples, ~40–70% of the genes that map to a cytogenetically abnormal region show corresponding expression changes; other genes within the region either do not change expression or, occasionally, change expression in the opposite direction of the cytogenetic abnormality. Nevertheless, as described in yeast, the majority of detectable regional gene expression biases in these mammalian tissues also coincide with chromosomal amplifications or deletions. As such, it is feasible to infer cytogenetic abnormalities by examination of high-density gene expression data. While the majority of REBs correspond to cytogenetic abnormalities, several groups have also identified a subset of regional gene expression biases that do not coincide with detectable DNA copy number changes [2,5,8,9,12,14]. While technical errors between DNA and expression-based approaches may account from some of these differences, it is also possible that other epigenetic factors could produce and regulate the appearance of REBs.
Partitioning gene expression data into subsets of adjacent genes and applying a summary function to each subset is a common method to identify REBs [1,5,11,15-19]. For example, a chromosome can be broken into consecutive, non-overlapping, 100 megabase (Mb) intervals and gene expression values that map to each interval tested for an expression biases using a variety of statistical/computational approaches. While partitioning approaches have been effective in identifying REBs, these approaches may be inherently limited due to the static nature of the partition span. Other, more dynamic, approaches to identify REBs utilizing run and scan statistics have also been reported [20]. However, the utility of these approaches for genome-wide scanning of expression biases is not well described.
Traditional data smoothing approaches ranging from simple moving averages to variable span local regression are common and straightforward methods that can be used to dampen variance and extract trends and patterns from ordered data series. For example, array comparative genomic hybridization (aCGH) data can be smoothed using an exponentially smoothed moving average to more easily identify abnormal chromosomal features [21]. While other approaches, such as hidden Markov models can also be utilized to analyze ordered genomic data [22], the complex nature of gene expression data may prevent the direct application of a subset of these types of analysis techniques. In this report, we outline an approach to identify REBs that summarizes the likelihood that each gene expression value measured lies within an regional expression bias using a multiple span moving binomial test. We use this approach to identify REBs in a set of hepatocellular carcinoma samples and compare the results to high resolution cytogenetic data obtained by aCGH. We also evaluated this approach using a set of clear cell renal cell carcinoma (ccRCC) gene expression profiles. In the majority of cases, dynamically determined REBs coincide with regions of DNA copy number change as determined by other molecular technologies. Interestingly, we identified a region on chromosome 6p where REBs are identified independent of apparent cytogenetic abnormalities. We show that the REBs in this region are produced in the most part by transcriptional responses to cellular proliferation.
Results
Identification of regional expression biases
To identify REBs, a modified version of a moving average is applied to two-color gene expression data obtained from the comparison of tumor HCC tissue to adjacent non-cancerous tissue (Figure 1a). Briefly, to calculate a moving average given a series of gene expression values ordered by genomic location and a window span that consists of five data points, the first five gene expression values would be collected, the average of this set determined, and the result stored as the first element of the moving average. The next span would include the second through the sixth gene expression values and the average of this span stored as the second element of the moving average. This process would continue until the end of the data series and the results of the moving average could be examined to identify trends. To identify REBs from ordered gene expression data, rather then a use an averaging function to evaluate each window span, an approximated binomial test is used to estimate of the probability, in terms of a z-score, that a gene expression bias exists within each span (see Materials and Methods). In this case, a positive z-score would indicate a disproportionate number of genes within the span show increased expression in the tumor profile when compared to the non-cancerous sample. Analogously, a negative z-score would indicate a disproportionate number of genes within the span show decreased expression in the tumor profile when compared to the non-cancerous samples. In addition, rather then collect data from a single window span, a data from a range of spans is collected and summarized (Figure 1b). In this case, the smallest window span used is 25, while the largest window size used is n/3 = 93. A minimum span of 25 assures the estimated z-scores are reasonably accurate (see Material and Methods) and a maximum span of n/3 prevents the generation of largely redundant data. Typical of many types of data smoothers, relatively small spans produce more variable REBs estimations while larger spans produce broader, more diffuse, REB estimations (Figure 1b). To estimate REB boundaries, for each gene loci the mean z-score derived from the range of window sizes is computed (Figure 1c). In addition, for plotting, the final REBs is masked so that only significant regions of bias are displayed. For simplicity, we term this approach IR-CGMA, for Improved Resolution-Comparative Genomic Microarray Analysis keeping in mind we have essentially described the application of an unweighted, multiple span, moving binomial test to identify REBs.
Figure 1 Identification of regional expression biases. A multiple span moving binomial test was applied to gene expression data to identify regional expression biases. A. Plot of log2-transformed tumor verses non-tumor expression ratios (m = 468) that map to chromosome 6 of sample HK1 organized from the p-arm telomere (left) to the q-arm telomere (right). B. Heatmap of the set of estimations generated by applying an approximated binomial function (see Materials and Methods) to the gene expression data using window spans of i = [25,...,m/3]. Genomic regions that contains a disproportionate number of relatively decreased expression values are shown in blue while genomic regions that show a disproportionate number of relatively increased expression values are shown in red. The color intensity indicates the significance of the expression bias. The highest intensity blue color indicates a z-score ≤ -4 while highest intensity red indicates a z-score ≥ 4 C. At each measured loci, an average z-score was computed from the set of estimations from each window span shown in B and plotted. Significantly down-regulated regional expression bias estimations are highlighted in blue (z ≤ -1.96, p ≈ 0.05) and up-regulated bias estimations highlighted in red (z ≥ 1.96, p ≈ 0.05).
Validation of IR-CGMA
To test the effectiveness of this method, we compared REBs identified by IR-CGMA to aCGH data derived from the same set of samples (Figure 2a, b). Both IR-CGMA and aCGH identified abnormalities that are commonly attributed to HCC such as +1q, -4q, -8p, +8q, -13q, -16q, -17p, and +17q [6]. To summarize the similarities and differences between IR.CGMA and aCGH, the predicted fractional allelic gain or loss was computed at each measured locus (Figure 3a). In the majority of cases, IR.CGMA identified frequent regional expression biases that corresponded to cytogenetic abnormalities as identified by aCGH. For example, on chromosome 1 both approaches identified a narrow region on the q-arm proximal to the centromere (1q21-23) that is frequently amplified. In addition, both approaches identified a region of frequent deletion on the distal tip of chromosome 17 (17p13). While in general REBs corresponded to features identified by aCGH there are regions of discrepancy. The most striking discrepancy between REBs and aCGH/CGH is located on chromosome 6p. Gain of chromosome 6p21-23 is not a frequently reported cytogenetic event in HCC either in this study or in other cytogenetic studies of HCC. However, chromosome 6p was frequently identified to be transcriptionally abnormal via REB scanning. Additionally, while gain of chromosome 17q frequently occurs in HCC, there is some discrepancy between the fraction of samples reported by IR-CGMA and aCGH.
Figure 2 Identification of REBs and DNA copy number abnormalities from individual HCC samples. 39 HCC samples were analyzed for REBs from gene expression data using IR-CGMA and for DNA copy number abnormalities from aCGH data using hidden Markov modelling. Corresponding chromosome ideograms for chromosomes 1, 4, 6, 8, 13, and 17 are also shown to scale. The red bars in the ideogram highlight the centromere. A. IR-CGMA estimations were plotted as a heatmap to indicate significant expression biases as described in Figure 1. For consistent plotting, z-scores > 4 and z-scores < -4 were set to 4 and -4 respectively. Scales ranging from 4 to -4 are shown adjacent to each graph. Data for all autosomal chromosomes for all samples was also generated [see Additional file 1]. B. aCGH predictions of genomic deletions (s ≤ -0.225, blue) and amplifications (s ≥ 0.225, red). The highest intensity blue color indicates a s ≤ -1 while highest intensity red indicates s ≥ 1. Scales ranging from 1 to -1 are shown adjacent to each graph. Data for all autosomal chromosomes was also generated [see Additional file 2].
Figure 3 Summary of REBs and DNA copy number changes in HCC. A summary of the data generated as described in Figure 2. A. For each genetic loci on the autosomal chromosomes, the fraction of HCC samples that contained significant upwards expression biases were plotted as a positive fraction and the fraction of samples that contained significant downwards expression bias were plotted as a negative fraction. DNA copy number data determined by aCGH was plotted in a similar manner. B. Data was plotted as in A., with the exception that genes involved in cell proliferation and nucleic acid metabolism were removed, as described in the text, before REBs were identified.
Given these discrepancies, to provide additional validation for the use of a multiple span binomial test to identify regional expression biases, REBs were also identified from a set of gene expression data derived from clear cell renal cell carcinomas (ccRCC). Like HCC, ccRCC presents with a consistent set of cytogenetic abnormalities including loss of 3p and gain of 5p [23]. Frequent gain of chromosome 12p has also been reported in some CGH studies of ccRCC [24]. While in this study, we do not have corresponding cytogenetic data for these specific samples to perform direct comparisons, IR-CGMA did identify abnormalities that overlap genetic abnormalities frequently identified in ccRCC, including loss of 3p and gain of 5p (Figure 4, 5). Interestingly, gain of chromosome 6p is not a frequent cytogenetic abnormality associated with ccRCC, however, like the HCC samples, this region was frequently identified as being abnormal via REB scanning. While technical effects associated with either aCGH, traditional, CGH, or IR-CGMA may be responsible a subset of these discrepancies, it is also possible that epigenetic transcriptional regulation could contributes to the REBs. Therefore, to determine if the transcriptional abnormalities reflected certain types of epigenetic effects, we examined the gene expression data in more detail.
Figure 4 Identification of REBs from individual ccRCC samples. 27 ccRCC samples were analyzed for REBs and plotted as described in Figure 2a with the exception that chromosomes 1, 2, 3, 5, 6, and 12 are shown. Chromosomes 1 and 2 are shown as representative regions that do not frequent REBs. Data for all autosomal chromosomes was also generated [see Additional file 3].
Figure 5 Summary of REBs in ccRCC. 27 ccRCC samples were analyzed for REBs and plotted as in Figure 3a.
Examination of chromosome 6p and 17q REBs
To evaluate the nature of the REBs on chromosomes 6p and 17q in HCC, misregulated genes within these regions were identified and partitioned based on Gene Ontology (Figure 6). Only two significantly enriched ontology's were identified (p < 0.005) from the upregulated genes in these regions: nucleic acid metabolism (GO:0006139) and cell proliferation (GO:0009607) [25,26]. While a small number of transcripts that had relatively increased expression in the tumor samples were identified as negative regulators of cell proliferation (GO:0008283), overall these results suggest that pronounced REBs on chromosome 6p and chromosome 17q reflect the active cell division of the tumor cells compared to non-cancerous cells. To test this hypothesis, up-regulated genes mapping to these ontologies were removed from the HCC gene expression dataset (154 of 8128 genes, 1.9%) and REBs recomputed (Figure 3b). The REBs on chromosome 6p were considerably diminished and the discrepancy on chromosome 17q was partially diminished. In contrast, REBs on chromosome 1q and 8q were not appreciably changed after removing the cell proliferation associated genes. Taken together, these results suggest that the transcriptional effects of active cell proliferation participate in the production of the REBs of 6p and 17q.
Figure 6 Functional classification of differentially expressed genes on 6p and 17q. Genes on chromosome 6p and chromosome 17q that are differentially expressed in HCC compared to adjacent non-cancerous tissue were identified as described in the Material and Methods. The t-statistic corresponding to each misregulated gene (p < 0.05) was plotted with respect to gene location. For consistent plotting, t-statistics > 10 and t-statistics < -10 were set to 10 and -10 respectively. Genes classified as nucleic acid metabolism, cell proliferation, and negative regulation of cell proliferation are highlighted orange, red, and cyan, respectively.
Discussion
In this paper, we describe the construction and application of a straightforward data smoothing approach to identify REBs from gene expression data. As evidence for the validity of this approach, we demonstrate that REBs overlap cytogenetic abnormalities as determined using other cytogenetic profiling methods in the majority of cases. Due to the dependence of gene expression on chromosome dosage, identification of REBs can often assist in the interpretation of gene expression data. For example, detection of REBs can rapidly determine if a potential cytogenetic abnormality associates with particular sample classification, for example a more aggressive tumor subtype [27]. Perhaps more importantly, the prevalent overlap of transcriptional and cytogenetic abnormalities support HCC tumorigenesis models that advocate that recurrent cytogenetic aberrations, via their significant influences on gene expression, play important roles in HCC pathogenesis. In addition, the correlation between REBs and specific DNA copy number variation can assist in identification of candidate genes that have important function during tumorigenesis in a specific chromosomal regions. For example, a narrow region on the q-arm of chromosome 1 proximal to the centromere (1q21-23) is predicted to be frequently amplified both by IR.CGMA and aCGH, suggesting that this region may harbour candidate oncogenes. Inspection of genes that are highly expressed in 1q21-23 included several signalling molecules (MDUSP12, SHC1) and transcriptional factors (MEF2D, ILF2, TCFL1). Particular interesting is ephrin-A1, the ligand of Eph receptor tyrosine kinase. Ephrin-A1 has been implicated in angiogenesis and therefore may contribute to HCC development [28]. Clearly, it is important to evaluate the functions of these genes in HCC and determine the extent in which their gene expression is regulated by DNA amplification.
We also demonstrate in this study that not all REBs corresponded with detectable cytogenetic abnormalities, particularly in the region of chromosome 6p. Therefore, it is appropriate to apply alternative molecular approaches before attributing cytogenetic abnormalities to regional expression biases located in this region. Classification of the differentially expressed genes in this region into Gene Ontologies suggests that the regional expression changes reflect aspects of tumor cell proliferation as evidenced by an enrichment of features classified in nucleic acid metabolism and cell proliferation GO categories. Another notable feature of chromosome 6p, particularly 6p21-23, is that the gene density in this region is unusually high and harbors gene clusters of several protein families [29]. The unusually high gene density may also contribute to the identification of this region as frequently abnormal by REB scanning. It has been suggested that regions of high gene density correlates with open chromatin fibers. This open chromatin structure may facilitate transcriptional activation if appropriate transcriptional signals are present [30]. Other possible explanations of the REBs include regional methylation or Histone deacetylation.
While the high variability of gene expression data may prevent the direct application of several data modelling approaches, this study suggests that application of traditional data smoothing methods are appropriate to infer cytogenetic abnormalities from gene expression data and are worth investigating further. One potential disadvantage of smoothing approaches can be difficulty determining an appropriate window span that balances overall smoothness with optimal feature identification. While cross-validation using training and test data sets could theoretically identify an optimal window span for regional expression bias identification, we could not derive a span that was appropriate for all chromosomal regions across multiple data sets (data not shown). However, the increase in computer processing power allows the utilization of more computational intensive multiple span approaches to partially compensate for single span effects.
Unlike traditional cytogenetic analysis approaches, the resolution of this technique has a complex dependency on gene density, gene coverage on the array platform used, and tissue-dependent expression patterns. On average, the genome contains about 10 genes per Mb and varies between regions that have gene densities of ~6 genes per Mb (chromosome 13) to regions that have gene densities of ~26 genes per Mb (chromosome 19) [31]. As the smoothing approach presented requires at least 25 gene expression values to make a prediction, theoretically, the resolution of a REB could average ~2.5 Mb across the genome and range between ~1 to 4 Mb. However, for this analysis the cDNA arrays used contained ~8500 features could be confidently mapped to predicted genes. Of these features ~6000 genes (70%) where expressed at measurable levels in the liver tissue. Assuming ~30,000 human genes, the resolution for this study would be about 5-fold lower then the theoretically limits or average ~12.5 Mb across the genome and range from ~5 Mb to ~20 Mb.
While not reported here, this approach is suitable for single channel gene expression data provided appropriate reference and test expression profiling data can be converted to log-transformed expression ratios. We have also successfully used this approach to infer cytogenetic abnormalities from other species, such as mice and rats.
Conclusion
In this report, we describe a method to identify regional expression biases using a multiple span moving binomial test. As evidence for the validity of this approach, we demonstrate that this methods identifies REBs that associate with cytogenetic abnormalities as determined by array CGH and traditional CGH in both hepatocellular carcinoma and clear cell renal cell carcinoma.
Methods
Pre-processing of gene expression data sets
Two-color gene expression profiles derived from 39 HCC tumor samples and corresponding non-cancerous liver samples [32], and 33 RCC and adjacent non-cancerous kidney tissue samples [33], were obtained from the Stanford Microarray Database [34]. In all cases, gene expression values were normalized using the within-print tip group normalization method as implemented in the BioConductor packages for the R environment [35,36]. Prior to normalization, R and G values were threshold such that R or G values <150 were set to 150. In these data sets, the cancerous and non-cancerous samples were compared to a pooled cell-line reference. To allow direct comparison of tumor to non-cancerous expression values, new gene expression ratios (R) were generated from tumor tissue ratio (T/U) and corresponding adjacent non-cancerous tissue ratios (N/U) such that R = log2(T/U) - log2(N/U) [2]. Sequence comparisons were used to map microarray probe sequences to predicted Ensembl transcripts (Ensembl version 19) [29]. Included in the Ensembl transcript annotations are chromosomal mapping locations at base-pair resolution. If multiple probes mapped to the same locus a mean gene expression value was utilized.
Pre-processing of array comparative genomic hybridization data sets
Two-color array CGH data for the HCC samples was generated essentially as described [37]. A manuscript describing the details of the HCC copy number data and initial analysis is in preparation. In all cases, copy number values were transformed into copy number states using an unsupervised hidden Markov model as implemented in the BioConductor packages for the R environment [22,36]. States in which the median copy number change was ≥ 0.225 were defined as region of DNA gains and states in which the median copy number change ≤ -0.225 were defined as regions of DNA loss [37].
Identification of regional expression biases (IR-CGMA method)
Gene expression values were separated into chromosome subsets and ordered by gene mapping location. A sliding window algorithm was applied to each ordered gene expression subset such that within each window span a binomial test was applied under the assumption that the probability (p) of the appearance of a positive relative gene expression value equals the probability (q) of the appearance of a negative relative expression value, p = q = 0.5, and a z-score for the span is computed using the normal approximation to the binomial distribution. The z-score can be converted to an approximate significance values based on the two-tailed z-statistic (za/2) critical values. Data was generated using multiple window spans and an average z-score at each gene location was computed. More formally, given a set of ordered gene expression values gj for genes j = 1, 2, ...m, let xij denote expression bias approximations for genes j = 1, 2, ...m using window spans i = 25, 26, 27, ...m/3 where n denotes the number of window spans examined. An empty matrix X[n#m] is populated such that for m-i+2 >j ≥ i where t denotes the number of non-zero and r the number of positive values within the span {gk, gk+1, ...gk+i-1}. To not discard regions, xij is tapered when j <i such that and analogously tapered when j ≥ m-i+2. Final regional expression bias estimates (bj) are computed such that . Performing IR-CGMA on the 39 HCC gene expression profiles took approximately five minutes on a 2.6 GHz Intel Pentium IV with 1 GB of RAM.
Identification of misregulated genes
Identification of misregulated genes from the HCC gene expression profiles occurred in two-steps. First, genes were filtered to ensure each gene was well measured across the data set using an exact binomial test (p < 0.05). In this case, data was required in 24 of 39 (64%) of samples. Next, a one-sample t-test assuming unequal variance was applied to determine if expression values were significantly misregulated (p < 0.05).
Authors' contributions
KF and KD designed and implemented the data analysis algorithms and performed the data analysis. XC and CH obtained the HCC samples and generated both the gene expression and array CGH data. KF and XC collaborated to write the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Regional expression biases for all chromosomes in the HCC samples
Click here for file
Additional File 2
aCGH states for all chromosomes in the HCC samples
Click here for file
Additional File 3
Regional expression biases for all chromosomes in the RCC samples
Click here for file
Acknowledgements
This work was supported by NIH grant R33-CA10113-01 to K.A.F, and NIH grant K01-CA096774 to X. C.
==== Refs
Hughes TR Roberts CJ Dai H Jones AR Meyer MR Slade D Burchard J Dow S Ward TR Kidd MJ Friend SH Marton MJ Widespread aneuploidy revealed by DNA microarray expression profiling Nat Genet 2000 25 333 337 10888885 10.1038/77116
Phillips JL Hayward SW Wang Y Vasselli J Pavlovich C Padilla-Nash H Pezullo JR Ghadimi BM Grossfeld GD Rivera A Linehan WM Cunha GR Ried T The consequences of chromosomal aneuploidy on gene expression profiles in a cell line model for prostate carcinogenesis Cancer Res 2001 61 8143 8149 11719443
Xu XR Huang J Xu ZG Qian BZ Zhu ZD Yan Q Cai T Zhang X Xiao HS Qu J Liu F Huang QH Cheng ZH Li NG Du JJ Hu W Shen KT Lu G Fu G Zhong M Xu SH Gu WY Huang W Zhao XT Hu GX Gu JR Chen Z Han ZG Insight into hepatocellular carcinogenesis at transcriptome level by comparing gene expression profiles of hepatocellular carcinoma with those of corresponding noncancerous liver Proc Natl Acad Sci USA 2001 98 15089 15094 11752456 10.1073/pnas.241522398
Virtaneva K Wright FA Tanner SM Yuan B Lemon WJ Caligiuri MA Bloomfield CD de La Chapelle A Krahe R Expression profiling reveals fundamental biological differences in acute myeloid leukemia with isolated trisomy 8 and normal cytogenetics Proc Natl Acad Sci USA 2001 98 1124 1129 11158605 10.1073/pnas.98.3.1124
Harding MA Arden KC Gildea JJ Perlman EJ Viars C Theodorescu D Functional genomic comparison of lineage-related human bladder cancer cell lines with differing tumorigenic and metastatic potentials by spectral karyotyping, comparative genomic hybridization, and a novel method of positional expression profiling. Cancer Res 2002 62 6981 6989 12460916
Crawley JJ Furge KA Identification of frequent cytogenetic aberrations in hepatocellular carcinoma using gene expression data Genome Biol 2002 3 RESEARCH0075 12537564 10.1186/gb-2002-3-12-research0075
Shaughnessy JD Barlogie B Integrating cytogenetics and gene expression profiling in the molecular analysis of multiple myeloma. Int J Hematol 2002 76 59 64 12430902
Haddad R Furge KA Miller J Schoumans J Haab B Teh B Barr L Webb C Genomic profiling and cDNA microarray analysis of human colon adenocarcinoma and associated peritoneal metastasis reveals consistant cytogenetic and transcriptional aberrations associated with progression of multiple metastases Appl Genom Proteom 2002 1 51 62
Pollack JR Sorlie T Perou CM Rees CA Jeffrey SS Lonning PE Tibshirani R Botstein D Borresen-Dale AL Brown PO Microarray analysis reveals a major direct role of DNA copy number alteration in the transcriptional program of human breast tumors Proc Natl Acad Sci USA 2002 99 12963 12968 12297621 10.1073/pnas.162471999
Platzer P Upender MB Wilson K Willis J Lutterbaugh J Nosrati A Willson JK Mack D Ried T Markowitz S Silence of chromosomal amplifications in colon cancer Cancer Res 2002 62 1134 1138 11861394
Mao R Zielke CL Zielke R Pevsner J Global up-regulation of chromsome 21 gene expression in the developing Down syndrome brain Genomics 2003 81 457 467 12706104 10.1016/S0888-7543(03)00035-1
Masayesva BG Patrick H Mayer-Garrett E Pilkington T Mao R Pevsner J Speed T Benoit N Moon CS Sidransky D Westra WH Califino J Gene expression alterations over large chromosomal regions in cancers include multiple genes unrelated to malignant progression Proc Natl Acad Sci USA 2004 101 8715 8720 15155901 10.1073/pnas.0400027101
Lindvall C Furge KA Bjorkholm M Guo X Blennow E Haab B Nordenskjold M Teh B Combined genetic- and transcriptional profiling of acute myeloid leukemia with complex and normal karyotypes. Haematologia 2004 89 1072 1081
Lu YJ Williamson D Clark J Wang R Tiffin N Skelton L Gordon T Williams R Allan B Jackman A Cooper C Prichard-Jones K Shipley J Comparative expressed sequence hybridization to chromosomes for tumor classification and identification of differential gene expression Proc Natl Acad Sci USA 2001 98 9197 9292 11481483 10.1073/pnas.161272798
van Eijk R Oosting J Sieben N van Wezel T Cleton-Jansen AM Visualization of regional gene expression biases by microarray data sorting. Biotechniques 2004 36 592 594 15088376
Fischer G Ibrahim SM Brockmann GA Pahnke J Bartocci E Thiesen HJ Fernadez-Serrano P Moller S Expressionview: visualization of quantitative trait loci and gene-expression data in Ensembl Genome Biol 2003 4 R77 14611663 10.1186/gb-2003-4-11-r77
Breitkreutz BJ Jorgensen P Breitkreutz A Tyers M AFM 4.0: a toolbox for DNA microarray analysis Genome Biol 2001 2
Kim J Chung HJ Park CH Park WY Kim JH ChromoViz: multimodal visualization of gene expression data onto chromosomes using scalable vector graphics Bioinformatics 2004 20 1191 1192 14764551 10.1093/bioinformatics/bth052
Midorikawa Y Tsutsumi S Nishimura K Kamimura N Kano M Sakamoto H Makuuchi M Aburatani H Distinct chromosomal bias of gene expression signatures in the progression of hepatocellular carcinoma. Cancer Res 2004 64 7263 7270 15492245
Husing J Zeschingk M Boes T Jockel KH Combining DNA expression with positional information to detect functional silencing of chromsomal regions Bioinformatics 2003 19 2335 2342 14668216 10.1093/bioinformatics/btg314
Awad IA Rees CA Hernandez-Boussard T Ball CA Sherlock G Caryoscope: an Open Source Java application for viewing microarray data in a genomic context. BMC Bioinformatics 2004 5 151 15488149 10.1186/1471-2105-5-151
Fridlyand J Snijders AM Pinkel D Albertson DG Jain AN Hidden Markov models approach to the analysis of CGH data JMVA 2004 90 132 153
Kovacs G Akhtar M Beckwith BJ Bugert P Cooper CS Delahunt B Eble JN Fleming S Ljungberg B Medeiros LJ Moch H Reuter VE Ritz E Roos G Schmidt D Srigley JR Storkel S van den Berg E Zbar B The Heidelberg classification of renal cell tumours J Pathol 1997 183 131 133 9390023 10.1002/(SICI)1096-9896(199710)183:2<131::AID-PATH931>3.3.CO;2-7
Verdorf I Hobisch A Hittmair A Duba HC Bartsch G Utermann G Erdel M Cytogenetic characterization of 22 human renal cell tumors in relation to a histopathological classification. Cancer Genet Cytogenet 1999 111 61 70 10326593 10.1016/S0165-4608(98)00217-9
Consortium TGO Gene Ontology: tool for the unification of biology Nat Genet 2000 25 25 29 10802651 10.1038/75556
Beissbart T Speed TP GOstat: Find statistically overrepresented Gene Ontologies within a group of genes Bioinformatics 2004 20 1464 1465 14962934 10.1093/bioinformatics/bth088
Furge KA Lucas KA Takahashi M Sugimura J Kort EJ Kanayama HO Kagawa S Hoekstra P Curry J Yang XJ Teh BT Robust classification of renal cell carcinoma based on gene expression data and predicted cytogenetic profiles. Cancer Res 2004 64 4117 4121 15205321
Brantley-Sieders DM Chen J Eph receptor tyrosine kinases in angiogenesis: from development to disease. Angiogenesis 2004 7 17 28 15302992 10.1023/B:AGEN.0000037340.33788.87
Clamp M Andrews D Barker D Bevan P Cameron G Chen Y Clark L Cox T Cuff J Curwen V Down T Durbin R Eyras E Gilbert J Hammond M Hubbard T Kasprzyk A Keefe D Lehvaslaiho H Iyer V Melsopp C Mongin E Pettett R Potter S Rust A Schmidt E Searle S Slater G Smith J Spooner W Stabenau A Stalker J Stupka E Ureta-Vidal A Vastrik I Birney E Ensembl 2002: accommodating comparative genomics Nucleic Acids Res 2003 31 38 42 12519943 10.1093/nar/gkg083
Gilbert N Boyle S Fiegler H Woodfine K Carter NP Bickmore WA Chromatin architecture of the human genome: gene-rich domains are enriched in open chromatin fibers. Cell 2004 118 555 566 15339661 10.1016/j.cell.2004.08.011
Semple C Deep genomics in shallow times: the finished sequence of human chromosome 13 and 19 European Journal of Human Genetics 2004 12 875 876 15489904 10.1038/sj.ejhg.5201254
Chen X Cheung ST So S Fan ST Barry C Higgins J Lai K Dudoit S Ng I vandeRijn M Bostein D Brown PO Gene expression patterns in human liver cancer Mol Bio Cell 2002 13 1929 1939 12058060 10.1091/mbc.02-02-0023.
Higgins JP Shinghal R Gill H Reese JH Terris M Cohen RJ Fero M Pollack JR Van De Rijn M Brooks JD Gene expression patterns in renal cell carcinoma assessed by complementary DNA microarray Am J Pathol 2003 162 925 932 12598325
Sherlock G Hernandez-Boussard T Kasarskis A Binkley G Matese JC Dwight SS Kaloper M Weng S Jin H Ball CA Eisen MB Spellman PT Brown PO Botstein D Cherry JM The Stanford Microarray Database Nucleic Acids Res 2001 29 152 155 11125075 10.1093/nar/29.1.152
Yang YH Dudoit S Luu P Lin DM Peng V Ngai J Speed TP Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation Nucleic Acids Res 2002 30 e15. 11842121
Gentleman RC Carey VJ Bates DM Bolstad B Dettling M Dudoit S Ellis B Gautier L Ge Y Gentry J Hornik K Hothorn T Huber W Iacus S Irizarry R Leisch F Li C Maechler M Rossini AJ Sawitzki G Smith C Smyth G Tierney L Yang JYH Zhang J Bioconductor: open software development for computational biology and bioinformatics Genome Biol 2004 5 R80 15461798 10.1186/gb-2004-5-10-r80
Veltman JA Fridlyand J Pejavar S Olshen AB Korkola JE DeVries S Carroll P Kuo WL Pinkel D Albertson D Cordon-Cardo-Carlos Jain AN Waldman FM Array-based comparative genomic hybridization for genome-wide screening of DNA copy number in bladder tumors Cancer Res 2003 63 2872 2880 12782593
| 15882461 | PMC1134655 | CC BY | 2021-01-04 16:39:54 | no | BMC Genomics. 2005 May 9; 6:67 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-67 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-681588244910.1186/1471-2164-6-68Research ArticlePredictive screening for regulators of conserved functional gene modules (gene batteries) in mammals Nelander Sven [email protected] Erik [email protected] Erik [email protected]ånsson Robert [email protected] Olle [email protected] Mikael [email protected] Petter [email protected] Per [email protected] Sahlgrenska Academy, Department of medical and physiological biochemistry Box 440, SE-405 30 Göteborg, Sweden2 Chalmers Technical University, Department of mathematical statistics, Eklandagatan 76, SE-412 96 Göteborg, Sweden3 Lund Strategic Research Center for Stem Cell Biology and Cell Therapy, BMC B10, Klinikgatan 26, SE-221 48 Lund, Sweden2005 9 5 2005 6 68 68 27 7 2004 9 5 2005 Copyright © 2005 Nelander et al; licensee BioMed Central Ltd.2005Nelander 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 expression of gene batteries, genomic units of functionally linked genes which are activated by similar sets of cis- and trans-acting regulators, has been proposed as a major determinant of cell specialization in metazoans. We developed a predictive procedure to screen the mouse and human genomes and transcriptomes for cases of gene-battery-like regulation.
Results
In a screen that covered ~40 per cent of all annotated protein-coding genes, we identified 21 co-expressed gene clusters with statistically supported sharing of cis-regulatory sequence elements. 66 predicted cases of over-represented transcription factor binding motifs were validated against the literature and fell into three categories: (i) previously described cases of gene battery-like regulation, (ii) previously unreported cases of gene battery-like regulation with some support in a limited number of genes, and (iii) predicted cases that currently lack experimental support. The novel predictions include for example Sox 17 and RFX transcription factor binding sites that were detected in ~10% of all testis specific genes, and HNF-1 and 4 binding sites that were detected in ~30% of all kidney specific genes respectively. The results are publicly available at .
Conclusion
21 co-expressed gene clusters were enriched for a total of 66 shared cis-regulatory sequence elements. A majority of these predictions represent novel cases of potential co-regulation of functionally coupled proteins. Critical technical parameters were evaluated, and the results and the methods provide a valuable resource for future experimental design.
==== Body
Background
To understand how gene expression is coordinated to produce hundreds of cell phenotypes from an identical complement of genes is a principal challenge in mammalian genome research. A commonly suggested model for terminal differentiation in metazoans is that the core features of the cellular phenotype are mediated by a set of genes that is regulated as a gene battery, i.e. a set of functionally coupled genes that are activated by similar cis- and trans-acting regulators [1-3]. Although the gene battery is an idealized concept, concrete examples of gene battery-like regulation have been found in for example muscle subtypes [4-7], megakaryocytes [8], the epidermis [9] and lymphocytes [10,11].
A key step in the elucidation of gene battery-like regulation is to detect and functionally test the cis regulatory elements that mediate the co-regulation. A number of computation-based methods have been proposed to do this. In micro-organisms, computational methods have proven useful to detect modules of co-regulated genes [12,13]. In mammals, predictive models based on assumed co-regulation at the cis level have been constructed for liver- and skeletal muscle-selective gene regulation [14,15], and general tools have been developed for the regulatory analysis of co-expressed genes [16-18].
The aim of this work is to screen the mouse and human genomes and transcriptomes for instances where sharing of cis-regulatory sequences is statistically coupled to conserved co-expression of genes, i.e. cases that fall within or near the idealized gene battery concept. Another aim is to critically investigate technical parameters in order to maximize the sensitivity by which co-regulation of co-expressed genes can be detected. In a screen that covered ~40 per cent of all protein-coding genes according to the latest Ensembl annotation, we identified 21 co-expressed gene clusters with 66 cases of statistically supported sharing of cis-regulatory sequence motifs. The predictive value of the assignment of transcription factor binding sites was experimentally evaluated on EBF binding sites in a set of B-cell expressed gene clusters. The predicted cases of co-regulation include several previously known prototype examples of tissue specific regulation, but also novel predictions. All data are made available to the research community in the form of an internet resource that may serve as a starting point for further analysis.
Results
The analysis was based on the assumption that homologous genes in mouse and human are equivalent in most aspects of regulation and function. In particular, we assumed that the transcriptional regulation is conserved for orthologous genes. For example, the mouse gene Myh1 and the human gene MYH1 are assumed to share expression pattern and to share important cis-regulatory sequences. Below, the term 'ortholog pair' will be applied as a two-species equivalent of 'gene', for which expression and sequence data where retrieved for both mouse and human. Ensembl gene annotations [19,20] were applied thoughout the analysis.
Co-expressed gene sets were defined from a compendium of mouse and human expression data and tested for functional coupling
Previous results by our group and others have shown that statistical analysis of gene expression profiles in a large compendium of expression data can predict targets of differentiation processes, and identify functionally coupled genes [12,21,22]. In the following analysis, we specifically focused on a compendium derived from the recently completed Novartis expression atlas (SymAtlas) [23]. These data contain transcription profiles for 140 mouse and human tissues generated by hybridization on customized Affymetrix chips, and cover a large fraction of the mouse and human protein-coding genes. Sequence annotation of the Novartis probes linked the mouse and human data to approximately 17.000 unique Ensembl gene identities in each species (Table 1). Between the two datasets, 13282 non-redundant ortholog pairs could be identified by linkage of reciprocal Ensembl homology assignments (Table 1). In later steps, we excluded genes for which regulatory sequence could not be extracted (Methods), leaving 9561 ortholog pairs for clustering (Table 1). The final dataset included ~40% of the mouse and human Ensembl annotated genes.
Table 1 Gene coverage of the analysis
MOUSE HUMAN BOTH
Sequence data:
Ensembl genes total: 23954 21961
Ensembl transcripts total: 34076 35685
Ortholog pairs of Ensembl genes: 20188
Ortholog pairs with upstream sequence extracted : 13272
Ortholog pairs with upstream sequence extracted (redundancies removed): 12239 *
Expression data:
Ensembl genes matching SymAtlas probes: 17552 16929
ortholog pairs with expression data in both species: 13282 **
Integrated dataset:
Two-species expression data AND regulatory sequence: 9561
Numbers in the 'MOUSE' and 'HUMAN' columns signify the number of unique Ensembl identifiers in each respective species. Numbers in column 'BOTH' signify ortholog pairs of Ensembl entries. The overlap between the nonredundant sequence database (*) and the nonredundant expression dataset (**) was 9561 ortholog pairs.
Clustering
We clustered the mouse/human ortholog pairs based on their expression profiles across the 140 mouse and human tissues (Methods). We computed clusters at cut-off levels ranging from Pearson's correlation coefficient (hereafter termed PCC) 0.61 to 0.99. At the lowest applied cut-off, 57% of all ortholog pairs in the data were members of a cluster. The cluster sizes were distributed in a skewed manner, with a predominant formation of small clusters (Figure 1A). All analyses hereafter were performed at a PCC = 0.75 cut-off, which generated 160 clusters with 2407 ortholog pairs. This relatively stringent cut-off was chosen to reduce the number of non-relevant genes in the clusters. A higher cut-off did not seem reasonable given the noise-level of the microarray experiments (as judged by the Pearson correlation between replicated samples and between probes that are annotated for the same gene, data not shown).
Figure 1 Cluster statistics A: Histogram showing the log number of clusters as a function of log cluster size, based on the clustering at Pearson correlation coefficient 0.75 cut-off. Numbers on the x axis denote cluster size intervals (2), (3–4), (5–8), (9–16),... B: Co-expression as a predictor for shared function, protein interaction and paralogy. We identified all gene pairs that correlated above or below a threshold T (X-axis). We measured the fraction of such pairs for which there was (i) a BIND database protein-protein interaction recorded in human, (ii) at least one shared gene ontology term, and (iii) evidence of paralogy. We then computed the relative probability for genes above T with this feature, compared to gene pairs below T. At expression correlation 0.80, co-expression was associated with a 100-fold relative probability for genes to encode protein interactors, a 10-fold probability for genes to share functional annotation, but only a 3-fold probability for genes to be paralogs. C: Fraction of clusters with at least one over-represented GO term (Y axis), as a function of cluster size (X axis). GO term over-representations were computed at a 10% false discovery rate.
Assessment of functional linkage
According to the definition, a gene battery should encode functionally linked proteins. We used Gene Ontology (GO) terms, protein-protein interactions (from the BIND database [24]), and manual curation to assess functional linkage within our expression clusters. First, we investigated the relationship between expression profile similarity of two ortholog pairs and their relative probability to share a functional annotation term or to encode interacting proteins (Figure 1B). There was a consistent correlation between co-expression and the relative probability for two genes to share a GO-term or to encode interacting proteins (Figure 1B).
Second, we studied the statistical over-representation of GO terms and interacting proteins inside the clusters (Methods). 30/32 clusters with ten or more ortholog pairs contained at least one over-represented GO term. The proportion of small clusters with over-represented GO terms was lower, which reflects a lack of statistical power in small clusters. Genes encoding interacting protein pairs were also over-represented inside clusters. 35 cases of protein-protein interaction between two genes in the same cluster were found. In contrast, 1000 simulations on permuted data revealed a median of 9 interactions (observations ranging between 5 and 21). The BIND data contained only 600 interactions that could be mapped to the dataset, which explains the seemingly low number of 35 interactions.
Finally, the clusters were annotated by manual curation. Six examples are shown in Figures 2, 3, 4, 5, 6, 7, and clusters with predicted regulators are listed in Table 2. For a full overview, see the web supplement . Several clusters clearly represent gene sets that mediate specialized features of different cell types, including smooth muscle specific genes (cluster 40, Figure 2), B lymphocyte-specific genes (cluster 16, Figure 3), and genes selectively expressed in the testis (cluster 5, part shown in Figure 4). Further, we detected clusters that were related to cellular processes or organelles, including endoplasmatic reticulum (cluster 13, Figure 5), protein synthesis (cluster 1, Figure 6), and modulators of transcriptional regulation (cluster 65, Figure 7). A grand majority of the clusters were defined by peaks at different amplitudes in several tissues. As an example, the endoplasmatic reticulum cluster (Figure 5) was defined by a highly variant profile with strong expression in, for example, exocrine glands. Generally, the cluster profiles were conserved between species, in the sense that clusters were defined by peaks in the same organs. This effect was more pronounced for clusters with expression in a single organ, such as the testis (data not shown).
Table 2 over-represented motifs detected at <10% false discovery rate
Cluster number FDR PFM Number PFM Annotation
1: Protein synthesis <2.5% 190 M00025:Elk-1, M00007:Elk-1
<2.5% 110 M00050:E2F, MA0024:E2F
<2.5% 57 M00108:NRF-2, MA0028:Elk-1, MA0062:NRF-2
<2.5% 181 MA0076:SAP-1
<10% 18 M00074:c-Ets-1(p54)
<10% 78 M00262:Staf
2: Oocyte / fertilized egg <2.5% 71 M00024:E2F
<2.5% 190 M00025:Elk-1, M00007:Elk-1
<2.5% 9 M00032:c-Ets-1(p54)
<2.5% 110 M00050:E2F, MA0024:E2F
<2.5% 57 M00108:NRF-2, MA0028:Elk-1, MA0062:NRF-2
<2.5% 181 MA0076:SAP-1
<10% 238 MA0088:Staf, M00264:Staf
3: Neural tissues <2.5% 99 M00189:AP-2
<2.5% 115 M00196:Sp1
<2.5% 141 M00256:NRSF
<10% 75 M00243:Egr-1
4: Lymphocytes <2.5% 143 MA0050:Irf-1, M00062:IRF-1, M00063:IRF-2
<10% 74 M00054:NF-kappaB, MA0061:NF-kappaB
<10% 28 M00258:ISRE
5: Testis / spermatogenesis <2.5% 109 M00281:RFX1
<2.5% 142 MA0078:SOX17
<10% 108 M00036:v-Jun
<10% 248 M00041:CRE-BP1/c-Jun
<10% 65 M00100:CdxA
6: Liver <2.5% 16 M00134:HNF-4
<2.5% 212 M00158:COUP-TF / HNF-4, MA0017:COUP-TF
<2.5% 33 M00206:HNF-1
<2.5% 203 MA0046:HNF-1, M00132:HNF-1
<2.5% 234 MA0047:HNF-3beta, M00131:HNF-3beta
<2.5% 113 MA0065:PPARgamma-RXRal
<10% 46 M00155:ARP-1
<10% 212 M00158:COUP-TF / HNF-4, MA0017:COUP-TF
<10% 146 MA0071:RORalfa-1, M00156:RORalpha1
8: ECM <10% 215 M00378:Pax-4
9: Cardiac muscle <2.5% 223 M00026:RSRFC4
<2.5% 144 M00152:SRF
<2.5% 59 M00231:MEF-2
<2.5% 222 M00232:MEF-2
<2.5% 161 M00252:TATA
<2.5% 259 M00418:TGIF, M00419:MEIS1
<2.5% 160 MA0052:MEF2
<10% 60 M00006:MEF-2
12: Skeletal muscle <2.5% 201 M00184:MyoD, M00001:MyoD
<10% 17 M00002:E47
<10% 59 M00231:MEF-2
13: Endoplasmatic reticulum <10% 190 M00025:Elk-1, M00007:Elk-1
<10% 57 M00108:NRF-2, MA0028:Elk-1, MA0062:NRF-2
<10% 181 MA0076:SAP-1
15: Erythrocyte <10% 209 M00128:GATA-1, M00127:GATA-1
<10% 122 M00203:GATA-X
<10% 198 M00413:AREB6
16: B lymphocyte <2.5% 133 MA0081:SPI-B
17: Kidney <2.5% 33 M00206:HNF-1
<2.5% 188 M00411:HNF-4alpha1
22: Cell cycle genes <10% 110 M00050:E2F, MA0024:E2F
24: Pancreas <10% 121 M00071:E47
<10% 193 M00080:Evi-1, M00082:Evi-1
30: Small intestine <10% 31 M00346:GATA-1, M00347:GATA-1, M00348:GATA-2
40: Smooth muscle <2.5% 144 M00152:SRF
<2.5% 245 M00186:SRF, M00215:SRF
<2.5% 88 MA0083:SRF
44: Retina <2.5% 196 M00087:Ik-2
45: Testis (mouse signal only) <2.5% 164 M00253:cap
49: Lung/endothelium (mouse signal only) <10% 66 M00199:AP-1, M00037:NF-E2
65: NfkappaB signalling <2.5% 235 M00051:NF-kappaB (p50), MA0105:p50
Over-represented motifs arranged by cluster number. FDR column: False Discovery Rate (estimated probability for the over-representation to be a spurious detection). Motifs are are shown both by their numerical identifiers (PFM number) and by their annotation (PFM annotation). In cases where a PFM is a composite based on more than one source, the components are given separated by commas. The data where generated from the PCC = 0.75 clustering, 2 kb sequence database, at 90% phylogenetic conservation.
Figure 2 A smooth muscle differentiation battery: The bar chart (left) illustrates the average expression level of cluster members (Y axis) across arbitrarily ordered tissues (X axis) in two species (red = mouse and blue = human). Three tables list over-represented functional terms (upper small table), over represented motifs (PFMs) (middle table), and cluster members (lower table).
Figure 3 B-lymphocyte differentiation battery: Tables and charts are organized as in figure 2.
Figure 4 Testis selective battery: Over representation of RFX and SOX17 motifs indicates new roles for these factors as coordinators of testis selective gene expression. Tables and charts are organized as in figure 2.
Figure 5 Endoplasmatic reticulum associated genes: Over representation of XBP-1, NRF and RTS motifs suggest novel functions for NRF and ETS family factors in the regulation of ER-related genes. Tables and charts are organized as in figure 2.
Figure 6 Ribosomal genes: Tables and charts are organized as in figure 2.
Figure 7 NF-kappaB pathway: Over representation of REL and NFkappaB motifs indicates feed back signalling. Tables and charts are organized as in figure 2.
In combination, the GO term enrichment, the protein-protein interaction, and the manual curation convincingly show that clustered genes are functionally linked.
Regulatory DNA and descriptions of transcription factor binding sites were extracted and pre-processed
In the next part of the analysis, individual ortholog pairs were scored for transcription factor binding sites. Binding motifs were represented in the form of Position Frequency Matrices (hereafter denoted PFMs). Based on a fixed amount of upstream DNA sequence in each ortholog pair, a statistical score was computed to predict the potential for a site in the sequence to bind the factor corresponding to a PFM (described in detail in Methods). We extracted upstream + intronic sequence from the Ensembl database, in amounts of 2, 6 or 15 kb per gene (see Methods for details on boundaries). A filter was applied that removed ortholog pairs for which the transcription start differed between the two species (>1000 bp difference, see Methods). Filtering was successful for 12239 unique ortholog pairs (Table 1). In a subsequent filter, DNA sequence that was not conserved between mouse and human was removed, so-called phylogenetic footprinting. Phylogenetic footprinting was applied at different stringency, to allow the following optimization of the protocol (below). Furthermore, all exon sequence was removed from the analysis (Methods). Finally, the sequences were matched with the expression data based on annotation, the overlap being 9561 ortholog pairs (Table 1).
322 vertebrate PFMs were downloaded from the TRANSFAC and JASPAR databases [25,26]. Since the databases appeared to contain redundant or equivalent entries, highly similar PFMs were grouped and merged using single linkage hierarchical clustering and a PFM distance measure defined in [27], which reduced the number of PFMs from 322 to 266 (Methods). This step reduced the redundancy, but did not merge all identically annotated PFMs (see for example the redundant serum response factor (SRF) PFMs in Table 2, cluster 40).
Design of a predictive scoring system for transcription factor binding
After the retrieval and preprocessing of both sequences and PFMs, all individual sequences were tested for PFM matches using the MAST software [28], a software for identifying single or multiple motifs in sequences. MAST was set to compute one p-value for each PFM with respect to each sequence (Methods). Based on the p-values obtained from the MAST software, a composite score was defined as the product of the p-value in the mouse and human sequences of an ortholog pair (Methods). A composite score close to 0 indicates that both the mouse and human promoter sequence in the ortholog pair contains sequence elements that are in very good agreement with a certain PFM.
To address the biological validity of the MAST composite scores within the context of a set of co-expressed genes, we screened 48 ortholog pairs present in B-cell expressed clusters for individual EBF sites. In all, 24 individual EBF sites in 15 different ortholog pairs were detected (supplementary data, additional file 1). To test the functionality (in terms of EBF binding) of these sites, we examined the potential of 22 basepair duplex oligo-nucleotides spanning the presumed sites to compete for protein binding in Electrophoretic Mobility Shift Assays (EMSA:s). EBF binding capacity was assayed using nuclear extracts from the pre-B cell line 40-EI, a labelled mouse mb-1 promoter high affinity EBF site [29], and competitor oligo-nucleotides covering the new potential binding sites. In the absence of competitor oligo-nucleotide, a prominent DNA/protein complex (mb-1/EBF) could be detected whereas this complex was undetectable after the addition of a 300 fold molar excess of the unlabeled binding site (mb-1). The identity of the protein and the specificity of the binding were verified by competition with a point mutated mb-1/EBF site and by the inclusion of an EBF specific antibody into the reaction mixture (Figure 8B). The point mutated EBF binding site was unable to abolish complex formation even in a 1000-fold molar excess (Figure 8B), indicating that we detect specific protein DNA interactions with this experimental set up. 18 out of the 24 new binding sites competed for complex formation when added in a 300- or 1000-fold molar excess, and thus have the ability to bind EBF in vitro (Figure 8A). We conclude that the large majority of binding sites were able to functionally interact with the predicted protein, and that the composite score in principle detected factor binding. It should however be emphasized that the quality of predictions is dependent on the quality of the binding site descriptions, and the result does not necessarily imply that other predicted factors bind.
Figure 8 EMSA validation of EBF binding sites: A: The figure displays EMSAs in which binding of EBF to a mb-1 promoter EBF site is competed for by the inclusion of 300 or 1000-fold molar excess of unlabelled oligonucleotides that correspond to the predicted motifs. The name of the gene and the position of the motif is given in the figure. (m) indicates mouse and (h) human. "EBF" shows the position of the DNA/protein complex, and "Probe" indicates the position of free DNA. See supplementary information for a detailed description of the sites. B: The mb-1 promoter EBF site interacts specifically with EBF protein in a pre-B cell nuclear extract. The figure displays an autoradiogram in which a labelled EBF binding site from the mb-1 promoter has been incubated with nuclear extracts from 40EI pre-B cells and competitors or antibodies as indicated. EBF denotes the bound EBF protein and EBF-SS the super-shifted complex obtained by the addition of the EBF reactive antibody to the reaction mixture.
A statistical procedure was used to detect enriched motifs in the clusters
To test whether the identified clusters represent potential gene batteries, i.e. contain shared cis-regulatory elements, we designed a procedure to detect significant over-representation of orthologs that match a PFM inside a cluster. The procedure is based on a modification of Fisher's exact test, which tests for dependency between two events (in this case cluster membership vs detection of a motif) [30]. We introduced a procedure to optimize the composite score thresholds for individual PFMs. In brief, we selected the threshold that gave the lowest Fisher test p-value in any one cluster. This was based on the assumption that non-random distribution of detections over clusters reflects biological function, as has been proposed in [31]. The Fisher test p-values with the optimized thresholds are hereafter termed p-scores.
The tests of multiple detection thresholds, multiple clusters and 266 PFMs led to a need to compensate for mass testing. This was done by estimating false discovery rates (FDR) based on simulations on randomized data. In brief, we compared the outcome when using permuted and when using observed data at different p-scores, and defined the FDR as the ratio between the two (Figure 9A). This procedure allowed us to choose a significance threshold with a controlled expected number of spurious detections. Simulations were repeated 100 times. Since optimization of detection thresholds were repeated in each simulation round, no bias in disfavour of the control case was introduced.
Figure 9 False discovery rate estimation. A: Illustration of how false discovery rates (FDRs) were estimated by use of simulations. Values on the y axis represent the number of motifs below a certain p-score (x axis). Triangles show results for observed data, filled circles show results for permuted data (error bars show the 90% confidence interval from 100 simulations). The FDR was calculated as the ratio between the simulated expectation and the observation. Red dotted line: The number of clusters with at least one enriched motif. Note that several motifs where over-represented in the same cluster. B: Removal of paralogous genes from each cluster did not affect the number of detected motifs. Consequently, co-expressed paralogs is not an important source of false positives. C: The amount of DNA used per gene, and the phylogenetic footprinting stringency has a strong effect on the number of detected over-represented motifs. The sensitivity is higher when the amount of DNA is reduced. Error bars in B and C were obtained by using the 5th and 95th percentiles in the simulation to define the FDR.
The detection of over-represented PFMs was affected by DNA amount and masking, but not affected by gene paralogy
Using the described algorithm, we analyzed the number of times a motif was over-represented when using different amounts of DNA sequence per gene (2, 6 or 15 kb) and different stringency in the phylogenetic footprinting (>0% (keep all sequence), >60%, >70%, >80% or >90% identity). The analysis was performed at 10% false discovery rate and with the clustering obtained at PCC = 0.75 clustering cut-off. At all DNA amounts, higher stringency phylogenetic footprinting appeared to be beneficial, and between DNA amounts, 2 kb and 4 kb compared favourably over 15 kb (Figure 9C). We conclude that optimal results are obtained when using a limited amount of sequence per gene (see Discussion).
A potential confounding factor in the analysis is that similarity in upstream sequence may be attributable to factors other than shared cis-regulatory elements. The most important such factor is likely to be gene paralogy, since over-represented motifs might simply represent matches to non-functional (not yet diverged) sequence in a co-expressed gene family. In total, 22343 cases of pairwise paralogy were detected in the dataset of 9561 ortholog pairs using BLAST. All genes with transcripts matches at a BLAST E-value less than 1.0 were defined as potential paralogs. Of the 2407 ortholog pairs that clustered, 219 genes had a paralog inside the same cluster. This was clearly an over-representation since filling clusters with random genes from the dataset produced an average of 122 genes with a paralog inside a cluster (values between 104 to 131 observed in 10 randomization rounds). In a side-by-side comparison, we therefore analyzed the effect of removing paralogs from a cluster before testing for over-represented motifs (see Methods), as opposed to not removing paralogs but the same number of randomly selected genes. The results showed no increase in the number of over-represented motifs when allowing paralogs in the same cluster (Figure 9B).
66 over-represented motifs were observed in 21 clusters
The final analysis was performed with 2 kb sequence and 80% phylogenetic footprinting (results generated with variations of these parameters are found in the web supplement). Again, the clustering generated at PCC = 0.75 was chosen for analysis. The over-representation algorithm was run in the above cases, using 100 simulations to estimate false discovery rate thresholds. Over-represented motifs with FDR:s less than 2.5% and 10% were recorded. Key features of these results are presented in Tables 2 and 3, and the complete results are available by web browser .
Table 3 Interpretation of over-represented motifs with respect to published evidence. (See footnote for definition of the categories.)
I: Expected cases
Number Function / expression Transcription factors
4 lymphocytes Irf1/Irf2/ISRE, NFKB [34-36]
6 liver HNF-1 alpha and beta, HNF4 alpha [32, 59-61]
9,12 cardiac and skeletal muscle MyoD/E47, MEF2 family factors, SRF [4, 6, 48]
15 erythroid cells GATA-1 [62]
16 B lymphocytes Spi-B, Oct-I* [37, 40, 63]
22 cell cycle E2F family factors [49, 50]
40 smooth muscle SRF [5, 64]
II: Extended
3 neural tissue NRSF [65]
5 testis SOX17, RFX2 (RFX1-RFX3) [41-43]
6 liver ARP-1/COUP-TF, PPARγ [66, 67]
13 ER XBP-1 [68]
15,16 erythroid cells/ B cells AREB6* [38]
17 kidney HNF-4alpha, HNF-1-alpha/beta [69, 70]
III: Unexpected
1 protein synthesis NRF and ETS family factors
2 oocyte E2F and ETS family factors
6 liver ROR-alpha
8 ECM Pax4
9 cardiac muscle TALE family factors TGIF and MEIS1
13 ER NRF and ETS family factors
24 pancreas E47
30 small intestine HNF4-alpha and GATA factors
44 retina/eye Ik-2
45 testis cap
Criteria for inclusion in the different groups: I. Existing evidence of co-regulation of cluster member by the factor, and/or a mutant mouse phenotype that affects differentiation of the cell type. II. Evidence of the factor acting on a limited number of genes in the tissue. Evidence of the transcription factor being expressed in the tissue. III. No or limited evidence in the literature. * = Detected at 20% FDR.
A predictive screen performed with these settings associated 66 motifs to co-expression in a total of 21 clusters at FDR thresholds 2.5% and 10% (Table 2). There was a clear tendency for motifs to be detected in most large clusters, and for smaller clusters to lack over-represented motifs. This can be directly explained by the fact that even a high degree of motif presence in a small cluster can be attributed to spurious detections, and that the Fisher test accounts for that. The over-represented motifs were validated against the literature, and fell into three main categories: 1) previously described cases of gene-battery-like regulation, 2)previously unreported cases of gene battery-like regulation with some support in a limited number of genes, and 3) hypothetical cases of gene battery regulation (Table 3, Discussion).
In the typical case, an over-represented motif did not cover all cluster members but rather a fraction. The average coverage was 15%, with observations ranging from 1% to 100%. The limited coverage can be exemplified by the relatively large (104 ortholog pairs) cluster of liver-selective genes, which contained 8 over-represented motifs at FDR<10%, where the PFM annotation implied HNF1, HNF4, ARP1, PPAR-γ, and COUP-TF as the binding factors (Table 2). These different motifs were detected in between 8 and 19 ortholog pairs, indicating a cluster coverage of less than 20%. An example of very high coverage was observed in the cluster that contained smooth muscle differentiation markers, in which 6/7 (86%) ortholog pairs were positive for a Serum Response Factor (SRF) motif. We compared our results for HNF-1 to a whole-genome experimental screen for targets of this factor in hepatocytes [32]. 28% (29/104) of the ortholog pairs inside the liver cluster contained experimentally verified HNF-1 sites, which indicates a high but not complete coverage. Cross-comparison of experimentally and computationally identified HNF-1 targets showed a 69% – 71% agreement (results for the two different HNF-1 PFMs at 2 kb/gene masked for 90% phylogenetic conservation).
As a complement to using database motifs, we applied de novo motif elicitation to screen the regulatory sequence of each cluster for over-represented DNA motifs that were not present in databases. A two-step procedure based on the MEME algorithm [33] was developed (Methods). The use of a cluster to define a PFM will lead to a bias when testing for over-representation. A simulation that involved motif elicitation from randomized clusters was used to account for this bias (Methods). 28 de novo motifs were identified as over-represented in relation to the null simulation (Supplementary data, additional file 1). Of these 28 motifs, 13 matched a described database motif (based on the PFM similarity score described in Methods and in footnote of supplementary table). 9/13 of the matching motifs corresponded to motifs that were over-represented in Table 3. The remaining 15 motifs may represent novel regulatory elements. These motifs, and a description of the procedure, are available in the supplementary information.
Discussion
The gene battery theory predicts that the core features of differentiated cells are encoded by gene batteries, i.e. groups of functionally coupled, co-expressed genes that are regulated by similar sets of cis-regulatory elements. The primary aim of this article was to survey the mouse and human genomes for cases that fall within or close to the idealized gene battery concept.
Hierarchical clustering of an extensive compendium of micro-array expression data in mouse and human identified large numbers of co-expressed gene groups. A grand majority of the larger clusters were significantly enriched for genes that shared one or several GO-terms, indicating that co-expressed genes are functionally coupled. Moreover, interactions were significantly more common between proteins that are encoded by clustered pairs compared to randomly chosen pairs. 21 clusters, finally, were significantly enriched for genes that shared potential cis-regulatory elements
Confirmation and extension of several cases of gene battery-like regulation
The predicted over represented motifs (Table 2) fell into three principal categories: (i) predictions in agreement with described gene batteries, (ii) predictions supported by a limited number of observations, where the analysis predicted gene battery-like regulation, and (iii) novel predictions of hypothetical gene batteries. The findings are summarized with respect to these categories in Table 3.
For type (i) predictions the method is useful for proposing new target genes. As an example, the smooth muscle cluster contains 4 validated serum response factor (SRF) targets (Acta2, Actg2, Myl11, Tagln) [5]. The method identifies two more genes in this cluster, Myl9 and Lpp as potential SRF targets in smooth muscle (Figure 2). The findings in two clusters that are relevant for lymphocytes (cluster 4 in the web supplement) and B-lymphocytes (cluster 16, Figure 3) are further examples of type (i) predictions. Cluster 4 contains genes that are expressed by various lymphocyte populations. Significantly enriched motifs were Irf1/Irf2, NF-kappaB, and ISRE. Irf1 and Ir2 appear to bind the same sites [34]. Irf1 knockout mice are impaired in their myelopoiesis [35]. NFkappaB is a well-known regulator of inflammation and immune functions [36]. Cluster 16 contains genes that are specifically expressed in B-lymphocytes. SpiB was identified as the most over-represented PFM followed by AREB6 and Oct-1. Knockout of SpiB caused specific defects in B-cell terminally differentiated functions [37]. AREB6 is a hematopoietic transcriptional repressor [38]. Whereas Oct-1 seems to be dispensable for B-cell development [39], it regulates B lymphocyte genes in combination with specific co-activators [40].
In type (ii) cases, a regulator with some degree of documentation is statistically supported to play a role in gene battery-like regulation. As an example, RFX family proteins have been shown to regulate a limited number of genes in spermatogenesis, and SOX17 has been shown to be expressed in the testis [41-43]. Here, we demonstrate that motifs corresponding to these factors are present in 8–11% of testis-selective genes (Figure 4). 29% of kidney-specific genes were covered by HNF-1 and HNF-4 sites, suggesting that these genes may be general regulators in the kidney.
Finally, type (iii) cases represent novel predictions which can be viewed as testable hypotheses of regulatory mechanisms. A complete list of PFMs in the different categories with references to the literature is presented in Table 3. Examples include a potential functional role for an Ikaros-like motif in retina-selective genes (cluster 44 in the web supplementary), and a Pax family factor acting on extracellular matrix genes with strong expression in arteries (cluster 8 in the web supplementary). Interestingly, a combination of Ets family factor motifs, Nuclear respiratory factor motifs, and E2F motifs was detected in several clusters of a housekeeping character (Table 3).
Incomplete coverage of cis-regulatory motifs indicates that other mechanisms than co-regulation may contribute to coordinated expression
The gene battery concept predicts that upstream transcription factors coordinate the expression of target genes through binding to similar cis-regulatory elements in the target genes. In a recent report, Alloco and co-workers [44] studied the relationship between co-expression in yeast (as determined by correlation in a set of 610 arrays) and the probability of sharing a transcription factor (as determined by experiments). In their experiments, an expression profile correlation of 0.85 implied a probability of transcription factor sharing of >0.5. The incomplete coverage in our analysis may similarly indicate that other mechanisms than co-regulation contribute to the coordinated expression. Importantly, it is technically difficult to accurately quantify the fraction of genes in a cluster that respond to a factor. The presence of false positive gene cluster members reduces the coverage. Uncertainty in the motif assignment to ortholog pairs may also reduce the coverage. In one case study, we could use data from an independent experimental screen to confirm that the liver-selective cluster had a limited coverage of HNF-1 responders (results). Further experiments are required to resolve to what extent the lack of coverage reflects alternative regulatory mechanisms or technical limitations.
Data limitations
The tissues in the dataset represent samples of limited morphological resolution at a fixed time point. As a consequence, co-expressed gene groups that are active under specific developmental phases, under specific environmental stresses, or in small and localized anatomical structures are likely to escape detection. Moreover, co-expressed gene groups with a peak in one tissue only are unresolved for single cell types. Low coverage of motifs in such clusters may reflect contaminating genes that derive from another cell type.
The Affymetrix technology that were used to generate the expression measurements has been validated and match results obtained with tag sampling for sufficiently abundant genes [45,46]. Low intensity signals did not correlate well, indicating that such genes are less likely to form clusters (data not shown). Cross hybridization between related genes may, in theory, contribute to correlating expression levels. We can exclude that clusters primarily form as a result of cross-hybridization, since (1) members of well-characterized gene families appear in different (the expected) clusters (eg smooth muscle and cardiac actins) and since (2) co-expressed genes tended to share GO terms or to encode interacting proteins to a higher degree than they tended to be paralogs (supplementary data, additional file 1). We cannot however exclude subtle effects related to cross-hybridization, since results with cDNA microarrays indicate this possibility [47].
Transcription factors in the same family often have similar DNA binding properties and bind to the same sites on target genes. This leads to an ambiguity in the interpretation of PFM annotations. The MEF2 motif, is for example a receptor for the mammalian MEF2A, MEF2B, MEFC and MEFD transcription factors [48]. Similarly, the E2F motif is a receptor for a family of 9 E2F family proteins that form heterodimers with another family of proteins, the DP proteins, in a way that affects the binding affinity [49]. Different E2F proteins act at different stages in the cell cycle [50], and this is not captured by our method.
One additional data limitation is that a fraction of Ensembl transcripts may lack sequence in their 5' ends. The accuracy of our transcription starts is therefore dependent on the quality of the Ensembl transcript database.
Method considerations and perspectives
Our approach is technically related to the Toucan and ConFac tools [17,51]. Important differences and extensions are the introduction of a composite scoring system, a procedure to optimize PFM thresholds, and the use of simulation at the level of the whole clustering to measure significances.
The performance of the method was not affected by the removal of paralogous genes from the clusters (Figure 9B), and clustering of such genes is clearly not a significant source of false positive motif prediction. Consequently, we decided to include the paralogs in the final analysis since we believe that co-expression of such genes may be a result of shared cis-regulatory elements.
We further evaluated the effects of using different DNA amounts in the analysis. In principal, it is motivated to include a substantial amount of DNA sequence per gene, since mammalian enhancers are frequently located far upstream in relation to the transcription start, or downstream in intronic DNA. However, the benefits of including as much sequence as possible must be balanced against the risk of introducing vast amounts of non-informative sequence into the analysis. The evaluation clearly favoured using limited amounts of DNA, and high stringency phylogenetic footprinting (Figure 9C). The optimal future alternative may be to assemble a small amount of DNA from a large area of the genome using algorithms that sort out regulatory DNA regions more efficient than phylogenetic footprinting.
The approach presented here needs to be extended to generate more complete models of gene battery regulation. Modelling of cis-regulatory element combinations and the relative position and spacing of elements are examples of such extensions. The use of PCC as a measure of co-expression may need to be reconsidered if extending the approach to very large datasets, since this measure may be sensitive to noninformative signals in a majority of samples. Further, the expression levels of potential trans-regulators and co-factors (based on protein interaction networks) can be introduced. Most likely, the method will benefit from a more accurate identification of regulatory regions (Crawford-04)
Conclusion
We screened the mouse and human genomes and transcriptomes for instances of gene battery-like regulation. Comparative clustering was highly predictive of gene function and protein interaction, which indicates that potential gene batteries could be identified this way. Based on a statistical composite score for motifs in ortholog pairs, and a simulation approach to determine significance levels, we found 21 instances of statistically supported gene battery-like regulation that were conserved between mouse and human. These included known cases of gene battery regulation in tissues such as muscle, lymphocytes, erythrocytes, and liver. A second category of predictions included regulators with some degree of documentation, e.g. in testis, kidney, and endoplasmatic reticulum. Finally, new candidate gene batteries with statistically enriched cis-regulatory motifs were listed.
The results of this investigation emphasizes the need to study differentiation in terms of larger transcriptional units, and extends the methodology for doing this.
Methods
Annotation and preprocessing of gene expression datasets
Target sequences for the Novartis Gene Atlas V2 mouse and human expression datasets [23] were matched against the Ensembl [20] collection of mouse and human transcripts using BLAST [52]. In cases where the E-value exceeded 10-20, the BLAST search was re-done against Ensembl gene sequences. If there was no match below 10-20, the probe set was excluded from further analysis. The resulting datasets covered 17552 mouse and 16929 human unique Ensembl genes (Table 1). Mouse/human orthologous gene pairs were formed using Ensembl homology maps. Redundant occurrences of the same gene in more than one ortholog pair were avoided according to the following procedure: the Ensembl human orthologs for each Ensembl mouse gene were identified. When more than one ortholog was assigned to a mouse gene, the one with the lowest positional disagreement d (defined below) was chosen. The procedure was repeated for all Ensembl human genes. Reciprocally matching ortholog pairs were identified, and others were excluded from further analysis. The mouse and human expression profiles of each ortholog pair were normalized with respect to mean and standard deviation and combined into a single larger profile, finally yielding an expression dataset with 13282 non-redundant ortholog pairs (Table 1).
Preparation of upstream DNA sequence
For each ortholog pair, mouse and human candidate regulatory sequence was extracted from Ensembl. The sequence extraction algorithm starts with an ortholog pair, localizes the 5' end of the transcript in the genome in each species, and computes a value that measures the positional disagreement between the transcript 5' ends in the two species. If the disagreement is too large, the ortholog pair is excluded from the analysis. Of the 12239 ortholog pairs in the expression dataset, 9561 satisfied this criterion (Table 1). A full description of the sequence extraction procedure is available in the online supplement.
Three different lengths of DNA were extracted: 2 kb, 6 kb and 15 kb. The 2 kb dataset contained nucleotide positions ranging from -2000 to -1 relative to the transcription start, the 6 kb dataset contained positions -4000 to +2000 and the 15 k dataset contained positions -10000 to +5000.
Phylogenetic footprinting
The sequence datasets were subjected to phylogenetic footprinting, i.e. removal of poorly conserved sequence. The mouse and human sequences of each ortholog pair were aligned by use of the LAGAN software [53] (standard settings). Similarity was defined as the number of identical nucleotides in a 20 bp window. Nucleotides in windows with similarity below the threshold were removed. In all, five different similarity threshold were applied: >0% (no footprinting – use all sequence), >60%, >70%, >80% and >90% identity.
Removal of exonic sequence
Each candidate regulatory sequence was aligned to the corresponding transcripts (pairwise BLAST, e-value threshold 0.01). Nucleotides aligning with one or more transcripts were removed.
Clustering
The set of 9561 orthologs pairs, for which both regulatory sequence and expression data could be assembled, were clustered with respect to expression pattern using hierarchical clustering [54] with average group linkage and Pearson's correlation coefficient as distance measure. In the average group linkage algorithm, cluster distances are defined as the distances between cluster means. We defined the cluster mean as the arithmetic mean of all cluster members. Correlation thresholds between 0.61 and 0.99 were applied in steps of 0.01.
Assembly of 266 non-redundant motif position weight matrices
Motifs represented as position frequency matrices (PFMs) were downloaded from the TRANSFAC [55] and JASPAR [25] databases. Non-vertebrate matrices were filtered out. Highly similar matrices were grouped and merged using single linkage hierarchical clustering, reducing the number of PFMs from 322 to 266. Distances between matrices were calculated using a probabilistic method [56]. Individual positions between matrices were compared using the chi square test, and p-values for all overlapping positions were combined using the geometric mean. Sense and antisense of motifs were compared for all possible frameshifts with at least 75% overlap. Clustered motifs (score > 0.5) were added together in overlapping positions, and flanking positions were discarded. PFMs were transformed into position weight matrices (PWM:s) to make them compatible with the MAST software (see below). The value of each matrix element was calculated according to the following formula:
where n is the raw count from the corresponding position in the PFM, N is the the number of observations (sum of each position/column in the PFM), and are pseudocounts and p(b) is the background frequency of the corresponding nucleotide.
Scoring of regulatory sequences for motif position weight matrices
Scoring of individual sequences
For each ortholog pair, both mouse and human regulatory sequences were scored for all motif PWM's using the MAST software [57]. MAST was set to compute whole-sequence p-values (-seqp setting), using a first order Markov chain background. The Markov chain background data where computed from unmasked genomic sequence -4000 upstream to +2000 downstream in all ortholog pairs.
MAST composite scores
The p-values reported by MAST for the mouse and human sequences of an ortholog pair were multiplied for each PWM. The result was treated as a composite score that reflects the overall "signal" for a certain binding site in the regulatory sequences of an ortholog pair. The composite score is a product of two p-values but should not be interpreted as a p-value, since the mouse and human sequences are highly dependent.
Algorithms to detect over-representation of GO terms and motifs
GO over-representation
Clusters were evaluated for over-representation of GO terms using Fisher's exact test [30]. Due to the large number of tests (the number of clusters times the number of GO terms), the resulting p-values were corrected using the Bonferroni method [30].
Motif over-representation
The test was applied to the 9561 ortholog pairs with both sequence and expression data. These constituted the population. For each PWM, genes with MAST composite scores below a threshold were defined as labeled and the others as unlabeled. Further, each cluster was considered as a sample from the population. The algorithm is briefly sketched here and is available in detail in the online supplement:
Step 1: Compute p-scores under the null hypothesis
First, ortholog pairs were permuted across the dataset, making each cluster a random selection of genes. Second, an optimal composite score threshold was found for each PWM. This was done by computing the Fischer test p-value for over-representation of labeled ortholog pairs for all clusters k at a range of detection thresholds ranging from 10-8 to 10-3 in stepwise increases by a factor of 100.5. The threshold chosen for each PWM was the one that gave the best p-value for that PWM in any cluster. P-values for over-representation of all motifs in all clusters were finally calculated using the optimized thresholds. We refer to this statistic as the p-score. This whole procedure was repeated for 100 iterations, which resulted in empirical estimates of how many over-represented motifs we could expect below a certain p-score under the null hypothesis.
Step 2: Compute p-scores for the observed data
This step was identical to step 1 but without permutations and repetitions.
Step 3. Compute the false discovery rate
After the simulation, we defined the false discovery rate (FDR) at p-score p as the expected number of over-represented motifs in the null simulation, divided by the corresponding value in the observed data.
Electrophoretic mobility shift assay (EMSA)
40-EI and 230–238 cells were grown in RPMI medium supplemented with 10% FCS, 10 mM HEPES, 2 mM pyruvate, 50 μM 2-mercaptoethanol and 50 μg gentamicin per ml (complete RPMI media). STAT activation in 230–238 cells was achieved by 5 hours of incubation with 0.5 ng/ml recombinant mouse interferon gamma (Immunokontact, Germany). Nuclear extracts were prepared according to Schreiber et al. [58]. DNA probes were labelled with γ[32P] ATP (Amersham Biosciences, UK) by incubation with T4 polynucleotide kinase (Roche Diagnostics, Mannheim, Germany), and purified on a mini Quick Spin Oligo Column (Roche Diagnostics, Sweden). Nuclear extracts were incubated with labelled probe (20,000 cpm, 3 fmol) for 30 min at room temperature in binding buffer (10 mM HEPES [pH 7.9], 70 mM KCl, 1 mM dithiothreitol, 1 mM EDTA, 2.5 mM MgCl2, 4% Glycerol) with 0.75 μg Poly(dI/dC) (Amersham Pharmacia Biotech, UK). When EBF-binding was investigated, 1 mM ZnCl2 was supplemented to the mixture. The samples were separated on a 6% polyacrylamide TBE gel, which was dried and subjected to autoradiography. In the supershift experiments (Figure 8B), DNA competitors or antibodies (anti-EBF SC-15333, anti-Actin SC-1616, Santa Cruz Biotech) were added 10 min before the addition of the DNA probe. To visualize the super-shifted complex, the unbound probe was run out of the gel. The following Oligonucleotides were used for EMSA: mb-1 sense: 5'-AGCCACCTCTCAGGGGAATTGTGG-3'; mb-1 antisense: 5'-CCACAATTCCCCTGAGAGGTGGCT-3'; mutated mb-1sense: 5'-AGCCACCTCTCAGCCGTTTTGTGG-3'; mutated mb-1 antisense: 5'-CCACAAAACGGCTGAGAGGTGGCT-3';
List of abbreviations
EBF Early B cell factor
EMSA Electrophoretic Mobility Shift Assay
GO term Gene Ontology term
PCC Pearson's correlation coefficient
PFM Position frequency matrix
PWM Position weight matrix
Authors' contributions
The overall computational strategy applied was conceived by SN, EL and PM together with EK, PL, and ON. SN and EL performed the bulk of the analysis and the manuscript was drafted by SN, EL and PL with contributions from all authors. The experimental validation of EBF sites was conceived and performed by RM and MS. The de novo detection of motifs was conceived and performed by EK.
Supplementary Material
Additional File 1
Supplementary.doc is a word file that contains all the supplementary information referred to in the text.
Click here for file
Acknowledgements
We thank Tore Samuelsson and Magnus Alm-Rosenblad for advice and for making the EVO high-performance computer available. We also thank Anne Uv, Klas Kullander and Christer Betsholtz for valuable comments on draft versions of the manuscript. The work was funded by the European Commission: The Sixth Framework Programme, The Swedish Cancer Foundation, and the Swedish Research Council for Medicine.
==== Refs
Davidson EH Genomic Regulatory Systems. Development and evolution. 2001 Academic Press
Morgan TH Embryology and Genetics 1934 New York , Columbia University Press
Britten RJ Davidson EH Gene regulation for higher cells: a theory Science 1969 165 349 357 5789433
Dias P Dilling M Houghton P The molecular basis of skeletal muscle differentiation Semin Diagn Pathol 1994 11 3 14 8202645
Owens GK Kumar MS Wamhoff BR Molecular regulation of vascular smooth muscle cell differentiation in development and disease Physiol Rev 2004 84 767 801 15269336 10.1152/physrev.00041.2003
Cripps RM Olson EN Control of cardiac development by an evolutionarily conserved transcriptional network Dev Biol 2002 246 14 28 12027431 10.1006/dbio.2002.0666
Firulli AB Olson EN Modular regulation of muscle gene transcription: a mechanism for muscle cell diversity Trends Genet 1997 13 364 369 9287492 10.1016/S0168-9525(97)01171-2
Wang X Crispino JD Letting DL Nakazawa M Poncz M Blobel GA Control of megakaryocyte-specific gene expression by GATA-1 and FOG-1: role of Ets transcription factors Embo J 2002 21 5225 5234 12356738 10.1093/emboj/cdf527
Ma S Rao L Freedberg IM Blumenberg M Transcriptional control of K5, K6, K14, and K17 keratin genes by AP-1 and NF-kappaB family members Gene Expr 1997 6 361 370 9495317
Arnone MI Davidson EH The hardwiring of development: organization and function of genomic regulatory systems Development 1997 124 1851 1864 9169833
Bartholdy B Matthias P Transcriptional control of B cell development and function Gene 2004 327 1 23 14960357 10.1016/j.gene.2003.11.008
Segal E Shapira M Regev A Pe'er D Botstein D Koller D Friedman N Module networks: identifying regulatory modules and their condition-specific regulators from gene expression data Nat Genet 2003 34 166 176 12740579
Simonis N van Helden J Cohen GN Wodak SJ Transcriptional regulation of protein complexes in yeast Genome Biol 2004 5 R33 15128447 10.1186/gb-2004-5-5-r33
Wasserman WW Fickett JW Identification of regulatory regions which confer muscle-specific gene expression J Mol Biol 1998 278 167 181 9571041 10.1006/jmbi.1998.1700
Krivan W Wasserman WW A predictive model for regulatory sequences directing liver-specific transcription Genome Res 2001 11 1559 1566 11544200 10.1101/gr.180601
Wasserman WW Palumbo M Thompson W Fickett JW Lawrence CE Human-mouse genome comparisons to locate regulatory sites Nat Genet 2000 26 225 228 11017083 10.1038/79965
Aerts S Thijs G Coessens B Staes M Moreau Y De Moor B Toucan: deciphering the cis-regulatory logic of coregulated genes Nucleic Acids Res 2003 31 1753 1764 12626717 10.1093/nar/gkg268
Frech K Werner T Specific modelling of regulatory units in DNA sequences Pac Symp Biocomput 1997 151 162 9390288
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 Database issue D468 70 14681459 10.1093/nar/gkh038
Hubbard T Barker D Birney E Cameron G Chen Y Clark L Cox T Cuff J Curwen V Down T Durbin R Eyras E Gilbert J Hammond M Huminiecki L Kasprzyk A Lehvaslaiho H Lijnzaad P Melsopp C Mongin E Pettett R Pocock M Potter S Rust A Schmidt E Searle S Slater G Smith J Spooner W Stabenau A Stalker J Stupka E Ureta-Vidal A Vastrik I Clamp M The Ensembl genome database project Nucleic Acids Res 2002 30 38 41 11752248 10.1093/nar/30.1.38
Nelander S Mostad P Lindahl P Prediction of cell type-specific gene modules: identification and initial characterization of a core set of smooth muscle-specific genes Genome Res 2003 13 1838 1854 12869577
Lee HK Hsu AK Sajdak J Qin J Pavlidis P Coexpression analysis of human genes across many microarray data sets Genome Res 2004 14 1085 1094 15173114 10.1101/gr.1910904
Su AI Wiltshire T Batalov S Lapp H Ching KA Block D Zhang J Soden R Hayakawa M Kreiman G Cooke MP Walker JR Hogenesch JB A gene atlas of the mouse and human protein-encoding transcriptomes Proc Natl Acad Sci U S A 2004 101 6062 6067 15075390 10.1073/pnas.0400782101
Bader GD Betel D Hogue CW BIND: the Biomolecular Interaction Network Database Nucleic Acids Res 2003 31 248 250 12519993 10.1093/nar/gkg056
Sandelin A Alkema W Engstrom P Wasserman WW Lenhard B JASPAR: an open-access database for eukaryotic transcription factor binding profiles Nucleic Acids Res 2004 32 D91 4 14681366 10.1093/nar/gkh012
Wingender E Chen X Hehl R 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
Zhang JC Kim S Helmke BP Yu WW Du KL Lu MM Strobeck M Yu Q Parmacek MS Analysis of SM22alpha-deficient mice reveals unanticipated insights into smooth muscle cell differentiation and function Mol Cell Biol 2001 21 1336 1344 11158319 10.1128/MCB.2001.21.4.1336-1344.2001
Bailey TL Gribskov M Combining evidence using p-values. application to sequence homology searches. Bioinformatics 1998 14 48 54 9520501 10.1093/bioinformatics/14.1.48
Hagman J Belanger C Travis A Turck CW Grosschedl R Cloning and functional characterization of early B-cell factor, a regulator of lymphocyte-specific gene expression Genes Dev 1993 7 760 773 8491377
Rice JA Mathematical statistics and data analysis 1995 2nd Belmont, CA , Duxbury Press xx, 602, A49 p.
Dieterich C Rahmann S Vingron M Functional inference from non-random distributions of conserved predicted transcription factor binding sites Bioinformatics 2004 20 Suppl 1 I109 I115 15262788 10.1093/bioinformatics/bth908
Odom DT Zizlsperger N Gordon DB Bell GW Rinaldi NJ Murray HL Volkert TL Schreiber J Rolfe PA Gifford DK Fraenkel E Bell GI Young RA Control of pancreas and liver gene expression by HNF transcription factors Science 2004 303 1378 1381 14988562 10.1126/science.1089769
Bailey TL Elkan C Fitting a mixture model by expectation maximization to discover motifs in biopolymers Proc Int Conf Intell Syst Mol Biol 1994 2 28 36 7584402
Kuo TC Calame KL B lymphocyte-induced maturation protein (Blimp)-1, IFN regulatory factor (IRF)-1, and IRF-2 can bind to the same regulatory sites J Immunol 2004 173 5556 5563 15494505
Testa U Stellacci E Pelosi E Sestili P Venditti M Orsatti R Fragale A Petrucci E Pasquini L Belardelli F Gabriele L Battistini A Impaired myelopoiesis in mice devoid of interferon regulatory factor 1 Leukemia 2004 18 1864 1871 15385939 10.1038/sj.leu.2403472
Aggarwal BB Nuclear factor-kappaB: the enemy within Cancer Cell 2004 6 203 208 15380510 10.1016/j.ccr.2004.09.003
Bartel FO Higuchi T Spyropoulos DD Mouse models in the study of the Ets family of transcription factors Oncogene 2000 19 6443 6454 11175360 10.1038/sj.onc.1204038
Turner J Crossley M Basic Kruppel-like factor functions within a network of interacting haematopoietic transcription factors Int J Biochem Cell Biol 1999 31 1169 1174 10582345 10.1016/S1357-2725(99)00067-9
Wang VE Tantin D Chen J Sharp PA B cell development and immunoglobulin transcription in Oct-1-deficient mice Proc Natl Acad Sci U S A 2004 101 2005 2010 14762167 10.1073/pnas.0307304101
Spiegelman BM Heinrich R Biological control through regulated transcriptional coactivators Cell 2004 119 157 167 15479634 10.1016/j.cell.2004.09.037
Wang R Cheng H Xia L Guo Y Huang X Zhou R Molecular cloning and expression of Sox17 in gonads during sex reversal in the rice field eel, a teleost fish with a characteristic of natural sex transformation Biochem Biophys Res Commun 2003 303 452 457 12659838 10.1016/S0006-291X(03)00361-9
Katoh M Molecular cloning and characterization of human SOX17 Int J Mol Med 2002 9 153 157 11786926
Horvath GC Kistler WS Kistler MK RFX2 is a potential transcriptional regulatory factor for histone H1t and other genes expressed during the meiotic phase of spermatogenesis Biol Reprod 2004 71 1551 1559 15229132 10.1095/biolreprod.104.032268
Allocco DJ Kohane IS Butte AJ Quantifying the relationship between co-expression, co-regulation and gene function BMC Bioinformatics 2004 5 18 15053845 10.1186/1471-2105-5-18
Ishii M Hashimoto S Tsutsumi S Wada Y Matsushima K Kodama T Aburatani H Direct comparison of GeneChip and SAGE on the quantitative accuracy in transcript profiling analysis Genomics 2000 68 136 143 10964511 10.1006/geno.2000.6284
Kim HL Comparison of oligonucleotide-microarray and serial analysis of gene expression (SAGE) in transcript profiling analysis of megakaryocytes derived from CD34+ cells Exp Mol Med 2003 35 460 466 14646601
Kuo WP Jenssen TK Butte AJ Ohno-Machado L Kohane IS Analysis of matched mRNA measurements from two different microarray technologies Bioinformatics 2002 18 405 412 11934739 10.1093/bioinformatics/18.3.405
Black BL Olson EN Transcriptional control of muscle development by myocyte enhancer factor-2 (MEF2) proteins Annu Rev Cell Dev Biol 1998 14 167 196 9891782 10.1146/annurev.cellbio.14.1.167
Tao Y Kassatly RF Cress WD Horowitz JM Subunit composition determines E2F DNA-binding site specificity Mol Cell Biol 1997 17 6994 7007 9372931
Attwooll C Denchi EL Helin K The E2F family: specific functions and overlapping interests Embo J 2004
Karanam S Moreno CS CONFAC: automated application of comparative genomic promoter analysis to DNA microarray datasets Nucleic Acids Res 2004 32 W475 84 15215433
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
Brudno M Do CB Cooper GM Kim MF Davydov E Green ED Sidow A Batzoglou S LAGAN and Multi-LAGAN: efficient tools for large-scale multiple alignment of genomic DNA Genome Res 2003 13 721 731 12654723 10.1101/gr.926603
Sokal RR Michener CD A Statistical Method for Evaluating Systematic Relationships Univ Kans Sci Bull 1958 38 1409 1438
Wingender E Recognition of regulatory regions in genomic sequences J Biotechnol 1994 35 273 280 7765063 10.1016/0168-1656(94)90041-8
Schones DE Sumazin P Zhang MQ Similarity of position frequency matrices for transcription factor binding sites Bioinformatics 2004
Bailey TL Baker ME Elkan CP An artificial intelligence approach to motif discovery in protein sequences: application to steriod dehydrogenases J Steroid Biochem Mol Biol 1997 62 29 44 9366496 10.1016/S0960-0760(97)00013-7
Schreiber E Matthias P Müller M Schaffner W Rapid detection of octamer binding proteins with "mini-extracts", prepared from a small number of cells Nucleic Acids Res 1989 17 6419 2771659
Li J Ning G Duncan SA Mammalian hepatocyte differentiation requires the transcription factor HNF-4alpha Genes Dev 2000 14 464 474 10691738
Ishiyama T Kano J Minami Y Iijima T Morishita Y Noguchi M Expression of HNFs and C/EBP alpha is correlated with immunocytochemical differentiation of cell lines derived from human hepatocellular carcinomas, hepatoblastomas and immortalized hepatocytes Cancer Sci 2003 94 757 763 12967472
Arrese M Karpen SJ HNF-1 alpha: have bile acid transport genes found their "master"? J Hepatol 2002 36 142 145 11804680 10.1016/S0168-8278(01)00298-7
Izraeli S Leukaemia -- a developmental perspective Br J Haematol 2004 126 3 10 15198727 10.1111/j.1365-2141.2004.04986.x
Hu CJ Rao S Ramirez-Bergeron DL Garrett-Sinha LA Gerondakis S Clark MR Simon MC PU.1/Spi-B regulation of c-rel is essential for mature B cell survival Immunity 2001 15 545 555 11672537 10.1016/S1074-7613(01)00219-9
Parmacek MS Transcriptional programs regulating vascular smooth muscle cell development and differentiation Curr Top Dev Biol 2001 51 69 89 11236716
Chen ZF Paquette AJ Anderson DJ NRSF/REST is required in vivo for repression of multiple neuronal target genes during embryogenesis Nat Genet 1998 20 136 142 9771705 10.1038/2431
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
Kang S Spann NJ Hui TY Davis RA ARP-1/COUP-TF II determines hepatoma phenotype by acting as both a transcriptional repressor of microsomal triglyceride transfer protein and an inducer of CYP7A1 J Biol Chem 2003 278 30478 30486 12777384 10.1074/jbc.M304201200
Lee AH Iwakoshi NN Glimcher LH XBP-1 regulates a subset of endoplasmic reticulum resident chaperone genes in the unfolded protein response Mol Cell Biol 2003 23 7448 7459 14559994 10.1128/MCB.23.21.7448-7459.2003
Ryffel GU Mutations in the human genes encoding the transcription factors of the hepatocyte nuclear factor (HNF)1 and HNF4 families: functional and pathological consequences J Mol Endocrinol 2001 27 11 29 11463573 10.1677/jme.0.0270011
Stoffel M Duncan SA The maturity-onset diabetes of the young (MODY1) transcription factor HNF4alpha regulates expression of genes required for glucose transport and metabolism Proc Natl Acad Sci U S A 1997 94 13209 13214 9371825 10.1073/pnas.94.24.13209
| 15882449 | PMC1134656 | CC BY | 2021-01-04 16:39:53 | no | BMC Genomics. 2005 May 9; 6:68 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-68 | oa_comm |
==== Front
BMC GeriatrBMC Geriatrics1471-2318BioMed Central London 1471-2318-5-71586971710.1186/1471-2318-5-7Research ArticleBrief cognitive assessment in a UK population sample – distributional properties and the relationship between the MMSE and an extended mental state examination Huppert Felicia A [email protected] Sara T [email protected] Fiona E [email protected] MRC Cognitive Function and Ageing Study (MRC CFAS) 1 Department of Psychiatry, University of Cambridge, Cambridge, UK2 Department of Public Health & Primary Care, Institute of Public Health, University of Cambridge, Cambridge, UK3 MRC Biostatistics Unit, Institute of Public Health, Cambridge, UK2005 4 5 2005 5 7 7 22 12 2004 4 5 2005 Copyright © 2005 Huppert et al; licensee BioMed Central Ltd.2005Huppert 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 MMSE's known flaws, it is still used extensively as both a screening instrument for dementia and a population measure of cognitive ability. The aim of this paper is to provide data on the distribution of MMSE scores in a representative sample from the UK population and to compare it with an extended cognitive assessment (EMSE) which covers a wider range of cognitive domains and provides a wider range of difficulty levels.
Methods
The MMSE and the EMSE were administered to over 12,000 participants at the screening stage of the MRC Cognitive Function and Ageing Study (MRC CFAS). MRC CFAS is a multi-centre population-based study in England and Wales with respondents aged 65 years and older.
Results
Normative values on the MMSE and EMSE are presented by age group, sex and level of education. There are very large differences between age groups, with smaller differences seen between the sexes and by level of education. The EMSE extends the scores at the high end of the ability range, but is no better than the MMSE at differentiating between dementia and non-dementia.
Conclusion
Population-derived norms are valuable for comparing an individual's score to the score that would be expected among the general population, given the individual's specific demographic characteristics.
==== Body
Background
The Mini-Mental State Examination (MMSE) was developed almost 30 years ago as a screen for dementia among hospital patients[1]. It remains the most widely used short cognitive test in clinical practice, clinical research, and epidemiological studies [2,3] However, its shortcomings are well known [4-6]. Principal amongst them are (a) very limited coverage of memory function, (b) a ceiling effect, resulting in inability to differentiate moderate from high functioning, and (c) absence of information about some aspects of cognitive function required for dementia diagnosis using internationally agreed criteria (DSM-IV [7]; ICD-10 [8]), namely perception and executive function. There is therefore a need for a short screening test that covers the range of cognitive processes required by diagnostic criteria, avoids the ceiling effect, and has an improved coverage of memory.
The Modified Mini-Mental State (3MS) Examination [4] was developed to extend the range of items in the MMSE and avoid ceiling and floor effects. It added eight extra items to the 19 items of the MMSE (date and place of birth, counting backwards from five to one, naming a body part, an easy repetition item, animal naming, a similarities item, and a delayed recall test), as well as a much extended scoring range for the original MMSE items, increasing the total score from 30 to 100. While the additional items extend the coverage of the MMSE by assessing additional aspects of cognitive function, i.e. remote memory (date and place of birth) and executive functioning (animal naming, similarities), the additional items are mainly designed for the low end of the ability range rather than the high end. Moreover, the 3MS still omits the assessment of perceptual ability which is required by diagnostic criteria to establish whether there is evidence of agnosia (DSM-IV [7]; ICD-10[8]). A large population-based study comparing the MMSE and the 3MS in a Canadian sample of people aged 65 years and over reported that the superiority of the 3MS over the MMSE appears more due to its extended scoring system than to its additional questions[9]. The 3MS has an additional drawback: it uses non-standard versions of some of the MMSE items and additional items. Specifically, for the memory task, rather than using three high frequency object nouns such as "apple – table – penny" which became standard in the community version of the MMSE[10], the 3MS uses the words "shirt – brown – honesty", which are more difficult to remember as they do not form a single visual image. Also the MMSE item asking subjects to write a sentence of their own choosing has been replaced in the 3MS by writing a sentence to dictation, which is far easier as it does not require the subject to generate a sentence. With regards to the additional items, the animal naming task gives the subject 30 seconds to name 4-legged animals, in contrast to the standard semantic fluency task which allows one minute to name any animals. Likewise, the 3MS uses a non-standard similarities question, ("In what way are an arm and a leg alike?"). These deviations mean that the 3MS is not strictly comparable either with the MMSE or with other standard cognitive tasks.
In September 1986, before the publication of the 3MS, the MRC convened an Alzheimer's Disease Workshop[11] whose aim was to establish the minimum dataset that should be collected in research studies on dementia. This included demographic data, history of physical and psychiatric disorder, alcohol and drug use, onset and duration of any difficulties, a physical examination, and a cognitive assessment. For the cognitive assessment, the MRC report recommended using the standardised administration and scoring instructions for the community version of the MMSE[10]. In addition, the MRC report recommended including the following items: category fluency (animal naming), recalling a name and address, assessment of remote memory, assessment of recent memory, ideational praxis, abstract thinking (similarities), and recognition of objects from unusual views. The specific aim of the additional items was to broaden the coverage of the MMSE in relation to both content and level of difficulty. The content requirement was to meet the needs for diagnostic criteria for dementia by including measures of perception and executive function; the difficulty requirement was to meet the need to differentiate between scores at the high end of the ability range. Individuals whose premorbid cognitive ability was high, might continue to obtain high scores on the MMSE and thus be missed on the MMSE screening test, even though their ability had in fact declined. It was hoped that the EMSE would measure more readily at higher cognitive abilities, and thereby differentiate between individuals at the high end of the ability range, therefore there would be scope to detect decline from an initially high level of functioning.
The MMSE items and most of the additional cognitive items recommended by the MRC report were included at the screening stage of the multi-centre MRC Cognitive Function and Ageing Study[12] and the findings are reported here. The screening test also included a measure of prospective memory (remembering to carry out an action), which has been reported elsewhere[13].
The aim of this paper is to describe the population distribution of performance on the MMSE and an extended cognitive assessment (EMSE) in a representative UK population.
Methods
Study design and population
The Medical Research Council Cognitive Function and Ageing Study (MRC CFAS) is a longitudinal population-based cohort study that involves six different study centres. The six centres were chosen because they represent the main national variation with regards to urban-rural differences, the north-south and east-west gradients, and variation in socio-economic levels and in known rates of chronic disease. Furthermore, all centres had existing researchers who were experienced in conducting population-based studies of the elderly. Urban sites included Liverpool, Newcastle, Nottingham, and Oxford, and rural sites included Cambridgeshire and Gwynedd, in North Wales. Liverpool was not included in this particular analysis because it was funded earlier than the other sites and had a different design without the same extended measurement of cognition. The full study design of the five identical centres is described in detail elsewhere and is explained briefly here[12].
Random samples of subjects over the age of 65 were selected from the Family Health Service Authority lists, giving an interviewed sample of approximately 2,500 people in each centre, stratified for equal numbers aged 65–74 years and 75 years and above. The study is longitudinal and this analysis focuses on information obtained at baseline – the prevalence (first) wave of the study. There were two phases at the prevalence wave. The first, screening, stage was used to establish level of cognitive performance and baseline risk factors on all individuals. A median of three months later, 20% of the subjects had a more detailed assessment interview to establish dementia diagnosis using the Geriatric Mental State (GMS) Automated Geriatric Examination Computer Assisted Taxonomy (AGECAT) diagnosis [14]. This group included the majority of individuals identified by the screening interview as potential cases of dementia, plus a random subset of the remaining population.
Cognitive measures
The screening interview included the Mini-Mental State Examination (MMSE) in the version developed for field surveys[10]. Spelling 'WORLD' backwards as an alternative to serial sevens was omitted to enhance standardisation. This version forms part of the Cambridge Cognitive Examination within the CAMDEX interview [15] and detailed administration and scoring instructions have been published [16,17]. The screening interview also included a selection of additional questions recommended by the MRC Alzheimer's Disease Workshop[11] as described previously. Coverage of language skills was extended by adding two objects to be named to the original MMSE objects. Praxis was extended by adding writing to dictation (a name and address) to the MMSE item writing a sentence, which requires the ability to write and to generate a sentence [4]. The coverage of memory was extended with three additional items: (a) asking subjects to recall the four objects they had named earlier in the session; (b) asking subjects to recall the name and address they had written earlier in the session; (c) a set of five questions assessing semantic memory or general knowledge. Executive function was assessed using a category fluency task (naming animals in one minute) and two similarities items. Perception was assessed by showing three photographs of familiar objects taken from unusual angles. These photographs and several of the other additional items (animal naming, writing to dictation, recalling the name and address, and one of the similarities items) were taken from the CAMCOG, and details of administration and scoring can be found elsewhere [16,17]. The additional items combined with the MMSE comprise the Extended Mental Status Exam (EMSE). The MMSE and the set of additional items each have a maximum score of 30, bringing the maximum score for the EMSE to 60. Details of the items in the EMSE (with the MMSE embedded) are shown in the Table 1. The MMSE takes around 7 minutes to administer, the additional EMSE items about 3 minutes.
Table 1 The Extended Mental State Exam
MMSE Question Physical Item Question Points
M What is the name of this place/what is the full address? 1
M What is the name of this city/town/village? 1
M What day of the week is it today? 1
M What is the date today (day)? 1
M What is the date today (month)? 1
M What is the date today (year)? 1
M What is the season? 1
M What is the country? 1
M Name two main streets nearby (or near your home). 1
M What floor of the building are we on? 1
M P What is this called? (pencil) 1
M P What is this called? (wristwatch) 1
P What are these called? (keys) 1
P What is this called? (envelope) 1
Name as many different animals as you can think of in one minute. 5
M Repeat 'No ifs, ands or buts' 1
What are the four things you were asked to name a few minutes ago?
P (pencil) 1
P (wristwatch) 1
P (keys) 1
P (envelope) 1
Who is the Prime Minister? 1
Who is the president of the United States of America? 1
What are the colours of the Union Jack (prompt: our national flag)? 1
Who was Neville Chamberlain? 1
Who was Guy Burgess? 1
M Repeat and remember these three words: apple, table, penny 3
M Serial sevens 5
What were the three words you were asked to repeat a little while ago?
M (apple) 1
M (table) 1
M (penny) 1
M P Read this and do what it says. (CLOSE YOUR EYES) 1
M P Copy this drawing. (a five sided figure) 1
M P Write a complete sentence. 1
Follow these instructions:
M P Take this piece of paper in your right hand 1
M P Fold the paper in half with both hands 1
M P Put the paper down on your lap 1
P Write down the following name address on this envelope: John Brown, 42 West Street, Bedford 2
In what way are an apple and a banana alike? 2
In what way are a boat and a car alike? 2
What was the name and address you were asked to remember a short while ago?
P (John) 1
P (Brown) 1
P (42) 1
P (West Street) 1
P (Bedford) 1
P What is the object in this picture? (shoe) 1
P What is the object in this picture? (spectacles) 1
P What is the object in this picture? (pipe) 1
Total 60
At screen, subjects were assigned an organicity score using the organic symptoms component of the AGECAT computerised algorithm [14]. The AGECAT at screen uses nine questions to obtain a level of organic symptoms. This algorithm is based mainly on interviewer ratings and the only items common to the MMSE and AGECAT are orientation items (place name and address, and current date – day, month and year). The only further item in common between AGECAT and the EMSE is naming the current UK Prime Minister. The organicity score ranges from O0 to O5, with O3 indicating mild organic symptoms, and O4 and O5 indicating probable dementia diagnosis.
Twenty percent of the screened sample went on to the diagnostic assessment interview. This included the majority of those who had a screen AGECAT organicity score of O3 and above. Individuals with a score of O3 at screen could have had dementia, but they are a mixed group with mild organic symptoms that could relate to dementia, mild cognitive impairment or depression. Of this group, those who went on to receive a diagnosis of dementia at the assessment interview were regarded as demented for this analysis. All other individuals who scored O3 and the interviewer reported moderate to severe memory impairment were excluded from the analysis for non-dementia norms.
Analysing performance on MMSE and EMSE
When describing performance on the MMSE and the EMSE, age was grouped into five-year bands (65–69, 70–74, 75–79, 80–84, 85 and above). Education was grouped into low level (9 years of schooling or less) and high level (greater than 9 years of full-time education). Those who had missing data about their educational attainment (n = 337) were placed in the low group.
Normative tables
To derive normative data on the MMSE and the EMSE, individuals classified as demented at screen or at assessment were excluded (n = 627). Tables by age group and sex, as well as by age, sex, and education are presented.
Missing data
Items that may have been missed due to sensory or motor impairment, called physical items, were recoded to 0 (i.e. treated as an incorrect answer). Such items include those involving writing or drawing, or those involving visual object recognition (see table 1 for details of items classified as being physical). Furthermore, in the MRC additional items, subjects are asked to recall an address that they have previously been asked to write. If the subject was physically unable to write the address, the instructions were to repeat twice by the interviewer and later the subject would be asked to recall it. There were a large number of missing values for these items, it is likely that interviewers omitted this recall in these individuals because of physical limitations. Therefore, the recall of the written address was categorised as a physical item and missing values were recoded to 0.
In general, individuals with items completely missing on the MMSE or the EMSE were not given scores for these tests. Many people were missing only one or two items from the MMSE or from the whole EMSE. Those missing two or fewer questions had their missing values recoded to 0 and were included in the analysis to ensure maximum use of the available data.
A sensitivity analysis to this missing data assumption has been undertaken using a pro-rata missing value score for the physical items. Individuals with missing data on physical items had a score generated for the proportion correct for their total score removing the physical question items from both the numerator and denominator.
Factors influencing missing scores
In order to have a clear picture of the sample on whom we have cognitive data, it is important to compare them with the sample from whom we were unable to obtain data. The interview was designed such that there was a small subset of questions deemed 'priority' questions to be answered by all individuals, if at all possible. These included the AGECAT organicity screen and the MMSE items. The interviewer could request 'priority mode' at any time or it was selected automatically if an individual was not orientated to time or place. Hence, there are missing data by design that need further investigation.
Several potential factors were investigated to describe the differences between those individuals who had a complete EMSE score, those who only had a complete MMSE score, and those who had neither test complete. These factors included the demographic variables, gender, age, education and social class (as defined either by the respondent's current or last occupation, or for some women, by their husband's current or last occupation). Other factors included dementia status, whether the subject appeared to be muddled, and whether the interview went into priority mode or had to be abandoned. Physical health was also analysed in relation to missing data, and included ADL impairment using the Townsend disability scale [18]. Interviewer-reported language problems or speech impairment, and an interviewer and self-reported evaluation of hearing impairment, visual impairment, and whether or not the subject was chairbound or bedfast were also included. Self-reported health problems including heart attack, transient ischaemic event, stroke, diabetes, Parkinson's disease, angina as measured by the Rose angina questionnaire[19], smoking status, and global self-reported health (excellent/good/fair/poor) were analysed.
Statistical Methods
Scores on the MMSE, and to a lesser degree the EMSE, do not follow a normal distribution. Hence, medians and other percentiles have been provided. For completeness, a logarithmic transformation (log (31-MMSE) or log(61- EMSE)) has been calculated and the estimate and reference ranges have been back-transformed to the original scale. Version 6.2 of the CFAS data has been used in this analysis. The analysis has been undertaken using STATAversion 8 [20].
Results
The total number of individuals screened was 13,004, representing a response rate of 80%. Table 2 shows the total number of subjects screened, and those with the MMSE complete and the EMSE complete by age and sex (12,680 (98%) had complete MMSE data and 12,394 (95%) had complete data on both MMSE and EMSE). Fewer subjects completed the entire EMSE compared to the MMSE because if a subject was unable to complete other sections of the interview, they entered priority mode where only a subset of questions were asked and most of the MRC additional items were skipped.
Table 2 Total number of people screened, MMSE complete, and EMSE complete, by sex, age, and dementia status
Men Women Total
Age (years) 65–74 75+ 65–74 75+
Total screened 2828 2329 3506 4341 13004
Non demented 2750 (97%) 2107 (90%) 3432 (98%) 3731 (87%) 12060 (93%)
Demented 78 (3%) 222 (10%) 74 (2%) 570 (13%) 944 (7%)
MMSE score† 2807 2272 3478 4123 12680
Non-demented 2735 (97%) 2076 (91%) 3410 (98%) 3664 (89%) 11885 (94%)
Demented 72 (3%) 196 (9%) 68 (2%) 459 (11%) 795 (6%)
EMSE score‡ 2788 2216 3455 3935 12394
Non-demented 2729 (98%) 2062 (93%) 3400 (98%) 3632 (92%) 11823 (95%)
Demented 59 (2%) 154 (7%) 55 (2%) 303 (8%) 571 (5%)
† Total number of respondents with complete MMSE score; ‡ Total number of respondents with complete EMSE score
For the sample as a whole, the median MMSE score is 27 (interquartile range 24 to 28) and the median EMSE score is 52 (interquartile range 47 to 55). The distributions of the two scores are presented in Figures 1 and 2, separately for the non-demented and demented groups. It can be seen that while the MMSE distribution is extremely skewed and truncated (Shapiro-Wilk W z = 18.5, skewness = -2.1), the skew is less marked for the EMSE (Shapiro-Wilk W z = 18.0, skewness = -1.6). The percentage of the sample who obtained maximum or near maximum scores (29 or 30) on the MMSE was 23.7% and the percentage who obtained maximum or near maximum scores (29 or 30) on the extra EMSE items was just 8.9%. Hence the EMSE shows better differentiation between scores at the high end of the scales. The Figures also show a relatively flat bell-shaped distribution of MMSE and EMSE scores for the group with dementia. The degree of overlap between the scores of demented and non-demented groups was similar in both MMSE and EMSE. A further illustration of the improved discrimination achieved by the EMSE at the high end of the distribution is depicted in Figure 3. This figure depicts the joint relationship between the high end of the MMSE scores (20–30) and the high end of the extra scores for the EMSE items (20–30). The contours show the increasing percentage of the study by score on MMSE and extra EMSE items. Whilst there is the expected relationship between ability on both parts of the EMSE scale the highest peak is centred just under 29 on the MMSE scale and 28 on the extra items. This shows the extent to which the EMSE successfully extends the scoring range for individuals who had already scored a maximum on MMSE.
Figure 1 MMSE scores in the demented and the non-demented
Figure 2 EMSE scores in the demented and non-demented
Figure 3 Joint relationship between top entries on the MMSE and extra items of the EMSE scale
Logarithmic means, standard deviations, medians, and various percentiles for both the MMSE and the EMSE in non-demented individuals are reported in Tables 2, 3, 4, 5, 6 and depicted graphically in figure 4. A logarithmic transformation [log(maximum score+1-score)] has been used to normalise the distributions of both measures, as they are both skewed. Both the MMSE and the EMSE have ceiling effects. The means and standard deviations were calculated on the log scale and the results were then back-transformed to the original scale for reporting purposes.
Table 3 Normative values for MMSE total score by age-group and sex (group without dementia)
Age-group (years)
65–69 70–74 75–79 80–84 85+
Men (total) 1407 1328 1080 700 296
Median 28 28 27 26 25
5th, 10th pcntle 23,25 22,24 21,23 21,23 18,20
25th, 75th pcntle 27,29 26,29 25,29 24,28 23,28
90th, 95th pcntle 30,30 30,30 29,30 29,29 28,29
Mean* 28.0 27.8 27.2 26.6 25.7
+/- 1 SD* 25–29 25–29 23–29 23–29 21–28
+/- 2 SDs* 20–30 18–30 16–30 16–30 11–30
Women (total) 1703 1707 1646 1257 761
Median 28 27 27 25 24
5th, 10th pcntle 22,23 21,23 21,22 19,20 17,19
25th, 75th pcntle 26,29 25,29 24,28 23,27 21,27
90th, 95th pcntle 30,30 30,30 29,30 29,29 28,29
Mean* 27.9 27.5 26.8 25.7 24.7
+/- 1 SD* 24–30 24–29 23–29 21–28 19–28
+/- 2 SDs* 17–30 16–30 14–30 11–29 9–29
* Back-transformed; pcntle Percentile of the distribution; SD Standard deviation
Table 6 Normative values for EMSE total score by age-group and sex, and educational level (low: ≤ 9 years, high: > 9 years)
Age (years): 65–69 70–74 75–79 80–84 85+
Educational level: L H L H L H L H L H
Men 913 491 812 513 693 379 425 277 199 94
Median 54 56 53 55 52 54 50 52 47 51
5th, 10th† 44,47 48,51 42,46 46,49 40,42 44,48 38,41 42,44 31,34 32,38
25th, 75th† 51,56 54,57 49,56 53,57 48,54 51,56 45,53 49,55 40,51 47,54
90th, 95th† 58,58 59,59 57,58 58,59 57,58 58,59 55,57 57,58 54,55 56,57
Mean* 54.2 55.8 53.3 55.4 51.8 54.5 50.2 52.5 47.2 51.0
+/- 1 SD* 48–57 51–58 46–57 50–58 44–56 48–58 42–55 46–56 36–53 43–55
+/- 2 SDs* 36–59 42–60 34–59 38–60 31–58 36–59 29–57 33–58 18–57 30–58
Women 1018 683 1011 688 1015 627 811 433 480 266
Median 53 55 52 55 51 53 47 51 45 49
5th, 10th† 42,45 46,49 41,44 46,48 38,42 42,46 33,37 39,42 29,33 35,38
25th, 75th† 50,56 53,57 48,55 52,57 46,54 50,55 43,51 47,54 40,50 43,53
90th, 95th† 58,58 59,59 57,58 58,59 56,57 57,58 54,55 56,57 53,54 56,57
Mean* 53.6 55.6 52.4 54.8 50.9 53.4 47.9 51.6 45.6 49.6
+/- 1 SD* 47–57 50–58 45–56 48–58 43–55 47–57 39–53 44–56 35–52 39–55
+/- 2 SDs* 33–59 39–60 31–59 35–59 29–58 34–59 23–57 30–58 17–56 20–58
† Percentiles of the distribution; SD Standard deviation; * Back-transformed
Figure 4 Percentiles for the MMSE and EMSE by age
Tables 3 and 4 present normative values for the MMSE and EMSE respectively, by age group and sex. All normative values are for individuals without dementia. There is a marked effect of age, with older subjects performing more poorly on both tests. The effect is particularly evident for the lowest percentiles (5th,10th) of the distribution, and for two standard deviations below the mean of the distributions. There is also a modest effect of gender, with women obtaining lower scores, which is particularly marked in the oldest age groups.
Table 4 Normative values for EMSE total score by age-group and sex (group without dementia)
Age-group (years)
65–69 70–74 75–79 80–84 85+
Men (total) 1404 1325 1072 697 293
Median 54 54 53 51 48
5th, 10th pcntle 44,48 43,46 41,44 38,42 31,35
25th, 75th pcntle 52,57 50,56 49,55 46,54 42,52
90th, 95th pcntle 58,59 58,59 57,58 56,57 55,56
Mean* 54.7 54.0 52.4 51.1 48.5
+/- 1 SD* 49–58 47–57 45–57 43–55 38–54
+/- 2 SDs* 37–59 34–59 31–59 30–58 20–57
Women (total) 1701 1699 1642 1244 746
Median 54 53 52 48 46
5th, 10th pcntle 43,46 42,45 40,43 34,38 30,35
25th, 75th pcntle 51,57 49,56 47,55 44,53 41,51
90th, 95th pcntle 58,59 57,58 56,58 55,56 54,55
Mean* 54.4 53.3 51.8 49.2 47.0
+/- 1 SD* 48–58 46–57 44–56 40–54 36–53
+/- 2 SDs* 34–59 32–59 30–58 23–57 16–57
* Back-transformed; pcntle Percentile of the distribution; SD Standard deviation
The effect of education on cognitive performance can be seen in Tables 5 and 6 where scores are broken down by age group, sex and level of education. Level of education has a marked effect on both MMSE and EMSE scores for all age groups and for both sexes. Therefore, when education level is known, users of these tables are advised to consult Tables 5 and 6, as they provide a better estimate of the individual's expected level of cognitive ability.
Table 5 Normative values for MMSE total score by age-group and sex, and educational level (low: ≤9 years, high: >9 years)
Age (years): 65–69 70–74 75–79 80–84 85+
Educational level: L H L H L H L H L H
Men 916 491 815 513 698 382 427 273 200 96
Median 28 29 27 28 27 28 26 27 25 27
5th, 10th† 22,24 24,26 22,23 23,25 21,22 23,24 20,22 22,24 18,20 19,21
25th, 75th† 26,29 27,29 26,29 27,29 25,28 26,29 24,28 25,28 22,27 25,28
90th, 95th† 30,30 30,30 30,30 30,30 29,30 30,30 29,29 29,30 28,29 29,29
Mean* 27.8 28.6 27.5 28.3 26.9 27.9 26.2 27.2 25.1 26.7
+/- 1 SD* 25–29 26–30 24–29 26–30 23–29 25–29 22–28 24–29 20–28 23–29
+/- 2 SDs* 19–30 21–30 18–30 20–30 15–30 20–30 16–29 18–30 11–29 14–30
Women 1019 684 1016 693 1018 628 815 442 491 270
Median 27 28 27 28 26 27 25 27 24 25
5th, 10th† 22,23 23,25 21,22 23,24 20,22 22,23 18,20 20,22 16,18 18,19
25th, 75th† 25,29 27,29 25,28 26,29 24,28 25,29 22,27 24,28 21,26 23,27
90th, 95th† 30,30 30,30 29,30 30,30 29,30 29,30 28,29 29,30 28,29 29,29
Mean* 27.6 28.4 27.1 28.0 26.5 27.5 25.2 26.7 24.2 25.7
+/- 1 SD* 24–29 26–30 23–29 25–30 22–29 24–29 20–28 23–29 19–27 21–28
+/- 2 SDs* 16–30 20–30 15–30 19–30 14–30 17–30 11–29 14–30 9–29 11–30
† Percentiles of the distribution; SD Standard deviation; * Back-transformed
Using selected cut-points from these normative tables, we examined how well these values were able to differentiate between demented and non-demented groups. These data are presented in Tables 6, 7, 8 and 9. Table 8 examines absolute cut-points, without taking account of socio-demographic characteristics. The next table presents the results adjusted for age and sex, which is useful for cases where the level of educational attainment is unknown. Table 8 presents the data adjusted for age, sex and education. From these tables, it can be seen that the EMSE is no better than the MMSE at distinguishing demented from non-demented individuals. Roughly the same percentage of demented subjects fall below the 5th percentile, the 10th percentile, 1 standard deviation of the mean, and 2 standard deviations of the mean for the MMSE and the EMSE. Furthermore, when comparing three tables, we find that adjusting for age and sex makes little difference to the percentage of demented subjects who fall below the given cut-points. Further adjusting for education seems to have no added benefit in this context. This suggests that the EMSE is primarily extending the description of higher functioning individuals rather than discriminating between the low functioning groups.
Table 7 Number and percentage of people below given cut-points on MMSE and EMSE by dementia status
MMSE Non-demented MMSE Demented EMSE Non-demented EMSE Demented
Below 5th pcntle (MMSE cut-point = 21, EMSE cut-point = 41) 618 (3%) 622 (73%) 810 (5%) 474 (77%)
Below 10th pcntle (MMSE cut-point = 23, EMSE cut-point = 45) 1443 (8%) 707 (84%) 1676 (10%) 524 (88%)
Below -1 SD (MMSE cut-point = 23.1, EMSE cut-point = 44.5) 2075 (13%) 739 (90%) 1676 (10%) 524 (88%)
Below -2 SD (MMSE cut-point = 12.9, EMSE cut-point-24.5) 5 (0%) 229 (24%) 36 (0%) 125 (19%)
pcntle Percentile of the distribution; SD Standard deviation of the mean
Table 9 Percentage of people below given cut-points on MMSE and EMSE by dementia status, adjusted for age, sex, and education
MMSE Non-demented MMSE Demented EMSE Non-demented EMSE Demented
5th pcntle 4 66 5 68
10th pcntle 8 78 9 80
-1 SD of the mean 16 87 12 87
-2 SD of the mean 1 39 1 33
pcntle Percentile of the distribution; SD Standard Deviation
Table 8 Percentage of people below given cut-points on MMSE and EMSE by dementia status, adjusted for age and sex
MMSE Non-demented MMSE Demented EMSE Non-demented EMSE Demented
5th pcntle 4 68 4 66
10th pcntle 8 80 9 83
-1 SD of the mean 14 87 12 88
-2 SD of the mean 1 38 1 31
pcntle Percentile of the distribution; SD Standard Deviation
Impact of missing data
Excluding the missing physical items responses from both the numerator and denominator caused little change in the results. The median and 5th percentile MMSE scores were at most one point higher. The EMSE median scores were not affected, however the 5th percentile was at most 2 points higher.
A comparison of the characteristics of those with complete MMSE scores, those with only complete EMSE scores and those with neither score complete has been undertaken. Those with no complete data or with only the MMSE complete were more likely to be older, female, in a manual occupation, and have a low level of education. They were also far more likely to have been classified as demented at screen. The vast majority of those who only had the MMSE complete (95%) went into priority mode during the screen interview. In general, those who completed both tests had fewer health problems than the other two groups.
Discussion
This paper presents the full normative values for the MMSE in a UK population sample aged 65 years and over, together with normative values for an extended cognitive assessment (EMSE), with a more complete coverage of cognitive domains than the MMSE and a wider difficulty range. The normative values have been calculated for the whole population sample, excluding those with probable or diagnosed dementia. The MMSE norms adds to the existing literature, in both English and other languages, providing norms based on the largest study to date for this age group [21-24].
This paper also presents normative values for a new scale – the Extended Mental State Exam (EMSE) which combines the MMSE with additional items recommended by a MRC Alzheimer's Disease Workshop [11]. Results show that the EMSE is more normally distributed than the MMSE and avoids the ceiling effect known to impair the usefulness of the MMSE in measuring population levels of cognitive ability. Compared with the MMSE, where almost a quarter of the normative sample achieved the two highest scores (29 or 30), only 9% of the sample achieved a similar level on the extra items in the EMSE.
We have examined how well selected MMSE and EMSE values differentiate between individuals with and without dementia. Both perform moderately well for this purpose. Like the EMSE, the Modified Mini-Mental State (3MS) Examination [4,25] extends the coverage of the MMSE and produces a much wider range of scores. However, the 3MS does not cover one of the domains of cognitive function required for a diagnosis of dementia, i.e. perception, and the additional items are geared more towards extending the low end of the ability range than the high end. Although the 3MS incorporates all the MMSE items, some have been modified, which makes it difficult to compare them with standard MMSE scores (see Introduction). In contrast, the EMSE incorporates the standardised field survey version of the MMSE [10] and the additional items are also presented in a standard way, thus enhancing the comparability between the EMSE and other measures. However unlike the EMSE the additional questions in the 3MS have been shown to assist in differentiating between individuals with and without dementia[9].
The values of the population norms were affected by level of education as in other studies [21,22,26] however a somewhat unexpected finding was that using age-sex and education cutpoints did not improve the discrimination between the normal and non-demented groups, as found previously using the 3MS [27], but in contrast to other studies that have used MMSE [28,29].
Other researchers have investigated the use of the MMSE in individuals with physical impairments and suggested improvements [30], but in this large sample we did not find that adjusting for missing had much impact on the distributions for either the MMSE or EMSE due to the coding of the physical items. This effect has been seen with MMSE in other studies[31].
The results show that the EMSE is comparable to the MMSE in its ability to differentiate between individuals with and without dementia. However, as described above, the EMSE is superior at providing data for individuals at the high end of the performance range, in a similar way to other tests (e.g. TICS-M and Hopkins Verbal Learning Test [32,33]).
The choice of a screening test for dementia depends on the type of population to be screened, and the aim of the screening procedure. If the population to be screened can be expected to perform poorly on a cognitive function measure (e.g. hospital patients), then we believe that the EMSE has few advantages over the shorter MMSE. This is also true if the aim of screening is to pick up definite cases of dementia. However if the purpose of the cognitive test is to examine whether individuals have early cognitive changes or mild cognitive impairment (MCI) or when the population to be measured includes many high performing individuals (e.g. population surveys), then the EMSE applied longitudinally has distinct advantages over the MMSE, and the additional 3 minutes of administration time may be regarded as worthwhile.
The data reported here are from the first cross-sectional wave of the MRC Cognitive Function & Ageing Study [12]. The EMSE has also been administered at later waves of the study, and later papers will examine longitudinal aspects of the EMSE and its ability to detect new cases of dementia.
Conclusion
Population-derived norms are valuable for comparing an individual's score to the score that would be expected among the general population, given the individual's specific demographic characteristics.
Role of funding source
The funding bodies have had no influence on the paper or decision to publish.
Conflict of interest
The author(s) declare that they have no competating interests.
Contributions
FH oversaw the clinical context of the paper, SC undertook the initial statistical analysis and co-wrote the paper, FM oversaw the analysis, undertook the final analysis and co-wrote the paper. MRC CFAS investigators undertook the fieldwork and oversaw the scientific integrity of the study and commented on the paper. All authors have seen and approved the final draft of the paper.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
MRC CFAS has been funded by programme grants from the Medical Research Council and Dept. of Health. It has since February 2001 been an MRC Co-operative. We would like to thank all the respondents, their families and primary care teams, from across the country for their continued participation in CFAS.
We thank the reviewers for their useful comments on our first submission.
==== Refs
Folstein MF Folstein SE McHugh PR Mini-mental state: a practical method for grading the cognitive state of patients for the clinician J Psychiatr Res 1975 313 1419 1420
Bleecker ML Bolla-Wilson K Kawas C Agnew J Age-specific norms for the Mini-Mental State Exam Neurology 1988 38 1565 1568 3419600
Bravo G Hebert R Age- and Education-Specific Reference Values for the Mini-Mental and Modified Mini-Mental State Examinations Derived from a Non-Demented Elderly Population Int J Geriatr Psychiatry 1997 12 1008 1018 9395933 10.1002/(SICI)1099-1166(199710)12:10<1008::AID-GPS676>3.0.CO;2-A
Teng EL Chui HC The Modified Mini-Mental State (3MS) examination J Clin Psychiatry 1987 48 314 318 3611032
Tombaugh TN McIntyre NJ The mini-mental state examination: a comprehensive review J Am Geriatr Soc 1992 40 922 935 1512391
Huppert FA Brayne C Gill C Paykel ES Beardsall L CAMCOG--a concise neuropsychological test to assist dementia diagnosis: socio-demographic determinants in an elderly population sample Br J Clin Psychol 1995 34 ( Pt 4) 529 541 8563660
American Psychiatric Association Diagnostic and Statistical Manual DSM-IV 1994 Edition IV revised Washington, DC, American Psychiatric Association
World Health Organisation ICD-10 Classification of Mental and Behavioural Disorders Diagnostic Criteria for Research 1993 Geneva, World Health Organisation
McDowell I Kristjansson B Hill GB Hébert R Community screening for dementia: the Mini Mental State Exam (MMSE) and Modified Mini-Mental State Exam (3MS) compared J Clin Epidemiol 1997 50 377 383 9179095 10.1016/S0895-4356(97)00060-7
Folstein M Anthony JC Parhad I Duffy B Gruenberg EM The meaning of cognitive impairment in the elderly J Am Geriatr Soc 1985 33 228 235 3989183
Medical Research Council Report form the MRC Alzheimer's Disease Workshop 1987 Medical Research Council
MRC CFAS Cognitive function and dementia in six areas of England and Wales: the distribution of MMSE and prevalence of GMS organicity level in the MRC CFA Study. The Medical Research Council Cognitive Function and Ageing Study (MRC CFAS) Psychol Med 1998 28 319 335 9572090 10.1017/S0033291797006272
Huppert FA Johnson AL Nickson J MRC CFAS High Prevalence of Prospective Memory Impairment in the Elderly and in Early-stage Dementia: Findings from a Population-based Study Applied Cognitive Psychology 2000 14 S63-S81 10.1002/acp.771
Copeland JR Dewey ME Griffiths-Jones HM A computerized psychiatric diagnostic system and case nomenclature for elderly subjects: GMS and AGECAT Psychol Med 1986 16 89 99 3515380
Roth M Tym E Mountjoy CQ Huppert FA Hendrie H Verma S Goddard R CAMDEX. A standardised instrument for the diagnosis of mental disorder in the elderly with special reference to the early detection of dementia Br J Psychiatry 1986 149 698 709 3790869
Roth M Huppert FA Tym E Mountjoy CQ CAMDEX: The Cambridge Examination for Mental Disoders of the Elderly 1988 Cambridge, Cambridge University Press
Roth M Huppert FA Mountjoy CQ Tym E CAMDEX-R: The Cambridge examination for mental disorders of the elderly - Revised 1999 Cambridge, Cambridge University Press
Townsend P Poverty in the United Kingdom 1979 UK, Pelican
Rose GA The Diagnosis of Ischaemic Heart Pain and Intermittent Claudication in Field Surveys. Bulletin of the World Health Organisation 1962 27 645 658
Statacorp Stata Statistical Software: Release 8.2 2003 College Station, Tx.
Grigoletto F Zappala G Anderson DW Lebowitz BD Norms for the Mini-Mental State Examination in a healthy population Neurology 1999 53 315 320 10430420 10.1159/000006604
Crum RM Anthony JC Bassett SS Folstein MF Population-based norms for the Mini-Mental State Examination by age and educational level JAMA 1993 269 2386 2391 8479064 10.1001/jama.269.18.2386
Kucukdeveci AA Kutlay S Elhan AH Tennant A Preliminary study to evaluate the validity of the mini-mental state examination in a normal population in Turkey Int J Rehabil Res 2005 28 77 79 15729101 10.1097/00004356-200503000-00011
Lechevallier-Michel N Fabrigoule C Lafont S Letenneur L Dartigues JF [Normative data for the MMSE, the Benton visual retention test, the Isaacs's set test, the digit symbol substitution test and the Zazzo's cancellation task in subjects over the age 70: results from the PAQUID Study] Rev Neurol (Paris) 2004 160 1059 1070 15602348
Tombaugh TN McDowell I Kristjansson B Hubley AM Mini-Mental State Examination (MMSE) and the Modified MMSE (3MS): A psychometric comparison and normative data 1996 US, American Psychological Assn.
Jones RN Gallo JJ Education bias in the mini-mental state examination Int Psychogeriatr 2001 13 299 310 11768377 10.1017/S1041610201007694
O'Connell ME Tuokko H Graves RE Kadlec H Correcting the 3MS for bias does not improve accuracy when screening for cognitive impairment or dementia J Clin Exp Neuropsychol 2004 26 970 980 15742546
Uhlmann RF Larson EB Effect of education on the mini-mental state examination as a screening test for dementia J Am Geriatr Soc 1991 39 876 880 1885862
Schmand B Lindeboom J Hooijer C Jonker C Relation between education and dementia: the role of test bias revisited J Neurol Neurosurg Psychiatry 1995 59 170 174 7629532
Busse A Sonntag A Bischkopf J Matschinger H Angermeyer MC Adaptation of dementia screening for vision-impaired older persons: administration of the Mini-Mental State Examination (MMSE) J Clin Epidemiol 2002 55 909 915 12393079 10.1016/S0895-4356(02)00449-3
Brayne C Gill C Paykel ES Huppert F O'Connor DW Cognitive decline in an elderly population--a two wave study of change Psychol Med 1995 25 673 683 7480446
de Jager CA Budge MM Clarke R Utility of TICS-M for the assessment of cognitive function in older adults Int J Geriatr Psychiatry 2003 18 318 324 12673608 10.1002/gps.830
Hogervorst E Combrinck M Lapuerta P Rue J Swales K Budge M The Hopkins Verbal Learning Test and screening for dementia Dement Geriatr Cogn Disord 2002 13 13 20 11731710 10.1159/000048628
| 15869717 | PMC1134657 | CC BY | 2021-01-04 16:30:32 | no | BMC Geriatr. 2005 May 4; 5:7 | utf-8 | BMC Geriatr | 2,005 | 10.1186/1471-2318-5-7 | oa_comm |
==== Front
BMC ImmunolBMC Immunology1471-2172BioMed Central London 1471-2172-6-91587782510.1186/1471-2172-6-9Research ArticleBMP-6 inhibits growth of mature human B cells; induction of Smad phosphorylation and upregulation of Id1 Kersten Christian [email protected] Einar A [email protected] Marit E [email protected] Lise [email protected] Erlend B [email protected] June H [email protected] Department of Immunology, Institute for Cancer Research, The Norwegian Radium Hospital, Montebello, 0310 Oslo, Norway2 Faculty Division The Norwegian Radium Hospital, University of Oslo, Norway2005 9 5 2005 6 9 9 23 12 2004 9 5 2005 Copyright © 2005 Kersten 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
Bone morphogenetic proteins (BMPs) belong to the TGF-β superfamily and are secreted proteins with pleiotropic roles in many different cell types. A potential role of BMP-6 in the immune system has been implied by various studies of malignant and rheumatoid diseases. In the present study, we explored the role of BMP-6 in normal human peripheral blood B cells.
Results
The B cells were found to express BMP type I and type II receptors and BMP-6 rapidly induced phosphorylation of Smad1/5/8. Furthermore, Smad-phosphorylation was followed by upregulation of Id1 mRNA and Id1 protein, whereas Id2 and Id3 expression was not affected. Furthermore, we found that BMP-6 had an antiproliferative effect both in naïve (CD19+CD27-) and memory B cells (CD19+CD27+) stimulated with anti-IgM alone or the combined action of anti-IgM and CD40L. Additionally, BMP-6 induced cell death in activated memory B cells. Importantly, the antiproliferative effect of BMP-6 in B-cells was completely neutralized by the natural antagonist, noggin. Furthermore, B cells were demonstrated to upregulate BMP-6 mRNA upon stimulation with anti-IgM.
Conclusion
In mature human B cells, BMP-6 inhibited cell growth, and rapidly induced phosphorylation of Smad1/5/8 followed by an upregulation of Id1.
==== Body
Background
Members of the transforming growth factor β (TGF-β) superfamily play central roles in controlling cellular proliferation, differentiation, migration and apoptosis [1]. These cytokines can be divided into three subgroups: TGF-β, the activins/inhibins, and the bone morphogenetic proteins (BMPs), of which the latter constitute the largest family. BMPs are 30–38 kDa hetero- or homodimeric proteins originally identified by their ability to induce ectopic cartilage and bone formation [2,3]. Several studies have demonstrated an essential role of these proteins during embryogenesis, and more recently, also in adult tissues [1]. TGF-β has been intensively studied in normal and malignant haematopoietic cells and is one of the most potent endogenous negative regulators known to date. [4]. In contrast, the effect of BMPs in the immune system has not been widely investigated. In that respect, BMP- 2, -4 and -7 have been found to control differentiation of hematopoietic stem cells [5] and early T cell development [6,7]. BMP-6 has been reported to reduce the number of cobblestone-area-forming cells of normal human haematopoietic cells [8]. Furthermore, BMP-2, -4, 6 and -7 had an antiproliferative and a proapoptotic effect on multiple myeloma cells [9-11]. In addition, by gene expression profiling, BMP-6 significantly increased the predictive value for a multi-gene signature test and was associated with a poor outcome in diffuse large B cell lymphomas (DLBCL) [12].
BMP-6, like the other BMP members, signals through ligation and heterodimerzation of BMP type I [activin-like-kinase (ALK)] and type II serine-threonine kinase receptors, which subsequently propagates the signal downstream by phosphorylating Smad proteins. BMP-6 can signal through the ligation of the type I receptors Act-RIA, BMP-RIA, and BMP-RIB and the type II receptors BMP-RII, Act-RIIA and Act-RIIB, which lead to the phosphorylation of the receptor Smads (Smad-1, Smad-5, and Smad-8). The R- Smads then form complexes with the co-Smad (Smad4) and are translocated into the nucleus where they exert gene regulation [1,13].
Given the reported role of BMP-6 in B-cell malignancies and haematopoietic progenitor cells, we wanted to explore its potential role in normal human B cells. We studied the effects of BMP-6 on proliferation and apoptosis on resting and stimulated B cells. Furthermore, the expression of BMP receptors and BMP-6 induced activation of the Smad signalling pathway with subsequent regulation of the target genes Id1–Id4, were resolved. Finally, we investigated whether B cells also were capable of producing BMP-6.
Results
BMP-6 inhibits anti-IgM induced proliferation of human B cells
The effects of BMP-6 on normal and neoplastic hematopoietic cells prompted us to investigate the effects of BMP-6 on normal human B cells. All experiments in this study were performed under serum-free conditions as FCS has been shown to interfere with BMP-signalling [14](own observations). To study the effect of BMP-6 on proliferation, B-cells from healthy volunteers were stimulated with anti-IgM and/or CD40L in the presence or absence of BMP-6 for three days. We found that BMP-6 led to a 35% mean reduction of anti-IgM- induced DNA synthesis (n = 8; p ≤ 0.0002, Figure 1A). Similar results were obtained for B cells treated with anti-IgM and CD40L (26% mean reduction, n = 6; p ≤ 0.023). The BMP-6-induced inhibition of proliferation was dose-dependent in both peripheral B cells (Figure 1B) and the Burkitt lymphoma cell line Ramos (40% reduction of DNA synthesis, Figure 1C). The BMP-6 effects could be reversed by addition of the extracellular inhibitor Noggin (Figure 1D). Similarly, a combination of the soluble BMP receptors BMP-RIB-Fc and BMP-RII-Fc also neutralized the effects of BMP-6 (data not shown). Next, we wanted to test whether BMP-6 had different effect on naïve and memory B cells. Naïve (CD19+CD27-) and memory (CD19+CD27+) B cells were isolated from peripheral blood by cell sorting of immunobead-isolated CD19+ B cells [15], and tested for their capacity to proliferate in the presence of BMP-6. However, BMP-6 inhibited anti-IgM induced DNA synthesis in the two subpopulations to a similar extent, with a mean reduction of DNA-synthesis of 45% (n = 5; p ≤ 0,004) for naïve B cells and 48% (n = 5; p ≤ 0,001) for memory B cells (Figure 1E).
BMP-6 induces cell death in human memory B cells and Ramos cells
Next, we wanted to establish whether BMP-6 also could affect the viability of normal B cells. Cell viability was determined by propidium iodide (PI) staining after culture with or without BMP-6 for 48 hours. Interestingly, BMP-6 showed a small, but reproducible mean increase of cell death from 17 to 23% (n = 5; p ≤ 0,003) in anti-IgM stimulated CD27+ memory B cells. Furthermore, Ramos cells showed a mean increase in cell death from 20 to 50% (n = 3, p < 0,001, figure 3) after BMP-6 treatment. In contrast, cell death of total CD19+ cells (n = 6; p ≤ 0,32; data not shown) or CD27-IgG- naïve B cells was not significantly affected (n = 5; p ≤ 0,65, figure 2).
Human B cells express BMP-6 receptors
Detailed knowledge regarding expression of different BMP receptors in B cells is currently not available. To further elucidate the role of BMPs in human B cells, we performed western blot analysis for type I and type II BMP receptors. This analysis revealed that the type I receptors Act-RIA, BMP-RIB and the type II receptors BMP-RII and Act-RIIb are expressed on resting human B-cells (Figure 4). Ramos cells expressed the type I receptors Act-RIA, weakly BMP-RIB and the type II receptor BMP-RII, but more weakly than normal B cells (Figure 4). HL60 cells were used for comparison and weakly expressed Act-RIA and BMP-RII.
Taken together, these data show that normal human B cells and Ramos cells express a set of BMP receptors, previously shown to bind BMP-6 [16].
BMP-6 induces phosphorylation of Smad1/5/8
Upon ligand binding, the type II receptor transphosphorylates and activates the type I receptor. Type I receptors can signal via several pathways. We examined the effect of BMP-6 on Smad phosphorylation, as the activation of Smad is considered to be a major signalling pathway for BMPs [17]. B cells were cultured in serum-free media over night and then treated with BMP-6 for various time points. Total protein lysates were prepared, and the amounts of the phosphorylated forms of Smad1/5/8 were determined by western blot analysis. Interestingly, treatment with 500 ng/ml BMP-6 induced phosphorylation of Smad. The BMP-6 induced phosphorylation was high at the earliest time point tested (15 minutes), and remained high for at least 48 hours (Figure 5). A similar phosphorylation was observed in Ramos cells, but not in HL60 cells (Figure 6). Furthermore, we also tested whether other known downstream signalling pathways of BMP-6 could be triggered by BMP-6 in human B cells. However, we did not observe any significant changes in the level of phospho-STAT3 or phospho-p38 upon BMP-6 treatment of B cells (data not shown).
BMP-6 induces upregulation of Id1
Next, we wanted to explore whether the BMP-6 induced phosphorylation of Smad 1/5/8 also could induce transcriptional changes of target genes. In this regard, the inhibitors of DNA binding proteins (Ids) are considered to be some of the major target genes for Smad-signalling [17]. Thus, B cells were pre-incubated over night in X-VIVO 15, and then cultured in medium alone or in the presence of BMP-6 for various time points before preparation of total RNA. The amount of Id1–Id4 mRNA was quantified by real-time RT-PCR. Interestingly, we observed a specific four-fold upregulation of Id1 mRNA in BMP-6-treated B cells (Figure 7). The up-regulation of Id1 mRNA was characteristic of an early inducible gene, with maximal upregulation two hours after the addition of BMP-6 and returned to baseline after 24 hours. In contrast, no significant changes were observed for Id2 and Id3 mRNA, whereas Id4-transcripts were not detectable (Figure 7, data not shown).
Western blot analysis revealed that the BMP-6-induced upregulation of Id1 mRNA also was correlated with upregulation of Id1 protein as well. The increase in Id1 protein level was detectable after one hour and increased until 24 hours after BMP-6 addition, showing a 16-fold upregulation compared with t0 (p ≤ 0.020, n = 4) (Figure 8 and 9). In line with the mRNA data, no consistent change in the amounts of Id2 and Id3 protein could be observed (Figure 8 and 9). We were able to block the Id1 specific band with a blocking peptide (data not shown). Taken together, these data suggest that Id1 could be a possible target gene for mediating the effects of BMP-6 in human B cells, whereas Id2 and Id3 not seem to be involved.
BMP-6 production in B cells
The fact that BMP-6 has been reported to act as an autocrine stimulator in chondrocytes [18] and ovarium [19], prompted us to investigate whether normal human B cells could produce BMP-6 upon stimulation. Ramos cells, which have been described to express BMP-6 mRNA endogenously [20], and the T cell line Jurkat, served as positive and negative controls, respectively. Endogenous BMP-6 mRNA levels in normal B cells were quantified by real-time RT-PCR after stimulation with anti-IgM for different time points. Interestingly, the up-regulation of BMP-6 mRNA was characteristic of an early-to intermediate inducible gene with maximal upregulation four hours after the addition of anti-IgM. The level of BMP-6 mRNA was back to baseline after 24 hours upon stimulation (Figure 10). Furthermore, both FCS and human AB-serum induced significant upregulation of BMP-6 mRNA (Figure 11). Interestingly, in a separate study we have found that normal human T cells do not express BMP-6 mRNA after activation (Sivertsen et al, manuscript in preparation). Next, we wanted to detect BMP-6 protein in normal B-cells and tested various commercially available antibodies. However, in our hands these anti-BMP-6 antibodies did only recognize the recombinant BMP-6 protein and not the native protein.
Discussion
Recent studies have demonstrated an important role for BMP superfamily members in hematopoietic stem cells, early thymocytes [6,7] and B-cell malignancies [8,11,12], but a role for BMPs in normal human B cells has previously not been reported. The present study demonstrated a significant antiproliferative effect of BMP-6 in peripheral blood CD19+ B cells. Additionally, BMP-6 induced cell death in CD27+ memory B cells as well as in a Burkitt lymphoma cell line (Ramos). Importantly, BMP-6 induced a rapid and marked increase in Smad-1/5/8 phosphorylation. Furthermore, the BMP-6 induced Smad phosphorylation was followed by a selective upregulation of Id1 mRNA and subsequent Id1 protein.
In the present study, the demonstrated antiproliferative effect of BMP-6 in anti-IgM treated B cells was significant and dose-dependent. Importantly, the anti-proliferative effect of BMP-6 could be completely neutralized by the use of a natural inhibitor, Noggin. This is in line with others, showing that Noggin can function as a BMP-6 antagonist [21,22]. In addition, the combination of soluble BMP-RIB-Fc and BMP-RII-Fc fusion proteins also neutralized the anti-proliferative effect of BMP-6 in human B cells. Interestingly, as for other TGF family members, bifunctional effects have also been demonstrated for BMPs. Whereas several of the BMPs have been shown to promote proliferation in various cell types including condrocytes [23], liver [24] and granulosa cells [25], antiproliferative effects and induction of apoptosis has been reported for B and T lineage cells. Similar effects as demonstrated for BMP-6 on human B cells in the present study, were demonstrated for BMP-2, 4, 6 and -7 in human myeloma cells [9-11]. Other members of the BMP-family have also been reported to induce apoptosis, including in mouse B lineage cells [26]. Additionally, BMP-4 inhibits thymocyte proliferation [6]. Taken together, these data suggest that the role of BMPs in the regulation of proliferation and apoptosis is highly cell type dependent.
To examine how BMP-6 exerts its functional effects in B cells, we analysed BMP receptor expression by western blot analysis. Human peripheral B cells were found to express the BMP type I receptors Act-RIA and BMP-RIB, and the type II-receptors BMP-RII and Act-RIIb, which signal after binding of several BMPs, including BMP-6 [16,13]. To further explore BMP-6 induced signalling, activation of several pathways is possible. The major signalling pathway known to date, is activation of R-Smads [13,27]. In that respect, BMPs have been shown to exert antiproliferative effects in B lineage cells via phosphorylation of R-Smad [11,28]. Furthermore, BMP-2 has been shown to induce activation of STAT3 in myeloma cells [9]. However, phosphorylation of R-Smad was not investigated in that study. BMP-2 has also been shown to induce phosphorylation of p38 [29]. Thus, phosphorylation of p38, STAT3 and Smad1/5/8 represent important BMP-signalling pathways that mediated the effects of BMPs and even cross-talk between these pathways has been reported [29,30]. In the present study, we were not able to detect BMP-6-induced changes in the phosphorylation status of STAT3 or p38 in human peripheral B cells. Instead, a rapid and marked phosphorylation of Smad1/5/8 was revealed. In a parallel study, we have found that other BMPs also induced phosphorylation of Smad1/5/8 in peripheral B cells (data not shown). We are currently pursuing microarray studies to identify the signalling pathways and target genes that are differently regulated by the various BMPs in human B cells.
Upregulation of Id1 via Smad1/5/8 phosphorylation is a known mechanism for BMP-6 signalling in other cell systems [31,32] and regulation of Id-proteins is thought to be an important mechanism for Smad-signalling [17]. In the present study, real-time RT-PCR experiments revealed a specific four-fold upregulation for Id1 in BMP-6-treated B cells, while the amount of Id2–Id4 remained unchanged. In agreement with this, western blot analysis demonstrated an upregulation of Id1 protein, while the amount of Id2 and Id3 protein levels remained unchanged. Previously, Id1 has been considered not to be expressed in later developmental stages than pro-B cells [33,34], and its constitutive expression has been reported to impair mouse B cell development [35]. Therefore, our demonstration of the time-dependent upregulation of Id1 mRNA and protein in mature normal human B cells is of particular interest. In that respect, it is noteworthy that TGF-β signalling in early and mature B cells induces both Id2 and Id3 expression [36,37], but not Id1 (data not shown). Interestingly, these results show that various members of the TGF-β family regulate Id proteins differently. Id2 and Id3 are considered to be the Id proteins mainly expressed in mature B cells [38]. The present study also found Id2 and Id3 protein in B cells to be more highly expressed than Id1 in resting B cells. However, BMP-6 did not induce significant changes in the protein expression of Id2 and Id3. It is believed that Id proteins block differentiation and promote proliferation in various cell types [39,33]. Id proteins act as dominant-negative inhibitors of E-proteins and Pax5 function by forming dimers with these proteins, making them unable to bind DNA. It has been proposed that the balance among E-proteins, Pax5 and Id proteins might have an important role in activated B cells [38]. In that respect, E-proteins have been implicated in both the promotion and inhibition of cell survival and growth at different points in lymphocyte development [40]. The antiproliferative and death inducing effect of BMP-6 in B cells with concomitant upregulation of Id1 protein is therefore in line with the view that Id proteins are required for the induction of growth arrest and apoptosis in B-lymphocyte progenitors by TGF-β [40].
Furthermore, Id proteins are known as important parts of signalling pathways involved in development, cell cycle and tumorigenesis [32]. It is well established that various members of the Id family are overexpressed in a range of human tumours and generally, Id1 appears to be the family member most widely overexpressed in a variety of human malignancies [41], including multiple myeloma [42,32]. Additionally, our findings that BMP-6 activates intracellular signalling pathways in human B cells might be of potential pathophysiological significance in lymphoma and inflammation. High BMP-6 mRNA expression in DLBCL has been shown to correlate to unfavourable outcome [12]. In this respect, it is of interest that targeted expression of Id1 to B-lymphocytes resulted in aberrant B cell development, massive apoptosis, and subsequent development of B cell lymphomas [35]. Moreover, BMP-6 has been suggested to play a role in rheumatoid arthritis (RA) [43,44] and elevated levels of Id1 and Id3 have been found in the synovia of RA-patients [45]. Altogether, these results point to an important role for Id proteins in the regulation of normal B cell homeostasis and in diseases, where B cells are involved. It will therefore be important to further elucidate the role of Id-1 in human B cells by selective over expression or inhibition of Id-1 gene expression.
Given the role of BMP-6 in mature human B cells demonstrated here, identification of BMP-6 producing cells in vivo with possibility of interaction with naïve and memory B cells might contribute to the understanding of mature B cell biology. High BMP-6 mRNA expression in DLBCL has been detected by gene expression profiling [12]. Furthermore, production of BMP-6 transcripts in normal activated B cells was detected in the same study. Of note, an autocrine BMP-6 loop has been reported by others in chondrocytes and in the ovarium [46,19,18]. Therefore, we wanted to explore the possibility for an autocrine BMP-6 loop in human B cells. We analysed the expression BMP-6 mRNA in peripheral blood B cells by real-time PCR, and report here the upregulation of endogenous BMP-6 transcripts after stimulation with FCS, human AB-serum and, most importantly, anti-IgM. However, our attempts to study BMP-6 protein levels were unsuccessful due to problems with unspecific binding of the anti-BMP-6 antibodies tested, and lack of specific staining in control cells known to express BMP-6 mRNA. In contrast, the recombinant protein was readily detected. In that respect, few investigators have detected BMP-6 protein in humans, especially in non-pathogenic tissue. The possibility of BMP-6 production in human B-cells is in line with a recent work that reported the production of BMP-6 in mouse B cells, infiltrating the bone marrow of mice with inflammatory arthritis [43]. In this study, a role for BMPs in the inflammatoric process of arthritis was suggested. The upregulation of the BMP-6-transcripts after IgM-crosslinking is of pathophysiologic interest [12]. A loss of TGF-β-responsiveness has been suggested to be a critical contribution to malignant transformation [47,48] and similar oncogenic mechanisms have been postulated for BMPs. Lines of evidence suggest [49] that at early stages of carcinogenesis, BMP-6 is not a tumour promoter, but suppresses benign and malignant tumour outgrowth. These findings are in good agreement with previous findings for other TGF-β family members, including TGF-β1 and BMP-4 [50], indicating that cellular context of the BMP target cell might define the various observed effects. In contrast to the upregulation of BMP-6 transcript in B cells, we were not able to detect BMP-6 transcripts in human peripheral blood CD4+ or CD8+ T cells (resting or stimulated with anti-CD3 and anti-CD28; data not shown), consistent with the findings in T cell lines [20] and T cells in mice [43]. Other potential BMP-6 sources for mature B cells in vivo might be other cells of the immune system or tissue with contact to the hematopoietic system. One well recognized source for BMP-6 production is the human bone and bone marrow stroma [51,8]. Furthermore, it is noteworthy that human umbilical vein endothelial cells (HUVEC) highly express BMP-6 mRNA [52], and vascular endothelium has been reported to produce BMP-6 [53]. These studies might imply a role for BMP-6 in transendothelial migration of B cells. BMP-6 mRNA has been demonstrated in murine macrophage cell lines, but not in humans [54]. In accordance with these findings, other human cell lines of neutrophil and monocytic origin have been described to be negative for the BMP-6 transcript [20]. To our knowledge, there is currently no report about BMP-6 production of human dendritic cells.
Conclusion
In conclusion, our results show that BMP-6 induces activation of intracellular Smad signalling in mature human B-cells with consecutive production of Id1 protein. Furthermore, we report that BMP-6 has an antiproliferative effect in B cells stimulated with anti-IgM alone or the combined action of anti-IgM and CD40L. Additionally, BMP-6 induces cell death in activated memory B cells and Ramos cells. Taken together, these results provide a rationale to further examine the role of BMP-6 signalling in normal B cell biology as well as in pathologic conditions like B cell malignancies and autoimmune disorders.
Methods
Cell culture
If not specified, all cells were cultured in X-VIVO 15™ (BioWhittaker, Verviers, Belgium) serum-free medium at 37°C and 5% CO2 in air.
Peripheral blood was provided by the Blood Bank at Buskerud Regional Hospital with formal agreement by the patients, and approval by the regional ethics committee. Highly purified resting human B-lymphocytes (CD19+ cells) were isolated from the peripheral blood by rosetting with immunomagnetic beads (Dynabeads M450; Dynal, Oslo, Norway) as described [55]. This procedure yields less than 0.5% T cells, 0.1% NK cells, and 0.5% monocytes as judged by indirect immunofluorescence staining.
The following cell lines from human lymphoid malignancies were maintained in RPMI 1640 (PAA Laboratories GmbH, Pasching, Austria) supplemented with 10% foetal bovine serum (FCS), 100 units/ml penicillin G, and 100 units/ml of streptomycin sulphate, but serum-starved for at least four hours and cultured in X-VIVO 15™ when included in experiments: EBV-negative BL cell lines Ramos (ECACC 85030802), HL60 (JCRB0085).
Growth factors/supplements
The following reagents were used at indicated concentrations: recombinant human (rhu) BMP-6 (1 μg/ml, if not specified otherwise), rhu BMP-RIB/ALK-6/Fc Chimera (5 μg/ml), rhu BMPR-II/Fc Chimera (5 μg/ml), and recombinant mouse Noggin (5 μg/ml) were purchased from R&D Systems (Abingdon, UK); Anti-IgM F(ab)2 fragments of rabbit polyclonal antibodies to human IgM heavy chain (37.5 μg/ml) was obtained from Dako, Copenhagen, Denmark and rhu CD40 ligand (CD40L, 10 ng/ml) was a gift from Immunex Corp. (Seattle, WA).
Antibodies used for flow cytometric analysis and immunoblot analysis
Antibodies against the human BMP-receptors Act-RIA, BMP-RIB, BMPR-II, Act-RIIA and Act-RIIb were purchased from R & D Systems (Abingdon, UK). Detection of the BMP-6-protein has been tried with the following antibodies: goat polyclonal anti-BMP-6 (Santa Cruz, San Diego, CA, USA), monoclonal mouse anti-BMP-6 and polyclonal goat anti-BMP-6 from R & D Systems (Abingdon, UK), and mouse monoclonal anti-BMP-6 (Chemicon International Inc, Temecula, CA, USA).
Characterisation of BMP-signalling pathways was done by use of anti-phospho-Smad1, -5, -and 8 polyclonal antibody (Chemicon, Temecula, CA, USA). Expression levels of Id1–3 proteins were detected with polyclonal rabbit antibody and detection was blocked with blocking peptide from Santa Cruz Biotechnology (Santa Cruz, San Diego, CA, USA). As secondary antibodies served anti-mouse, anti-goat or anti-rabbit IgG- horseradish peroxidase (HRP) from Dakocytomation AS (Copenhagen, Denmark) for immunoblot analysis. Anti-β-actin was from Santa-Cruz. From Becton Dickinson (San Jose, CA), we purchased anti-CD19-PE, anti-CD19-FITC. The antibodies used for cell sorting were anti-CD19 PC5 from Immunotech SA (Marseille, France) and anti-CD27 PE from Becton Dickinson, Biosciences Pharmingen (San Diego, CA, USA).
Cell sorting
Highly purified CD19+CD27- or CD19+CD27+ cells were obtained by staining CD19+ cells with anti-CD27 PE and CD19 PC5 mAbs for 30 minutes at 4°C, followed by washing with PBS and sorting on FACS DiVa from Becton Dickinson.
Western-blot analysis
B cells from peripheral blood or cultured cell-lines were lysed in lysis buffer (glycerol 10%, β-mercaptoethanol 5%, 0.0625 M Tris-HCL [pH 6.8], sodium dodecyl sulphate [SDS] 2.5%w/vol). Total protein (30–100 μg) from each sample was run on 10% or 12% SDS/polyacrylamide (SDS/PAGE) gels and blotted onto nitrocellulose filters (Protran; Schleicher &Schuell GmbH, Dassel, Germany). Blocking, washing and incubation of the filters with primary antibodies were done according to the manufacturer's protocols at room temperature (RT). After washing with TBS/0.1% Tween-20 (TBS-T), the filters were incubated with horseradish peroxidase (HRP) coupled to relevant secondary antibodies (see above) for 60 minutes at RT. Enzyme activity was visualised by the enhanced chemiluminescence system, ECL+PLUS (Amersham, Buckinghamshire, UK). Densitometric analysis was performed by scanning hyperfilms on a Personal Densitometer SI (Molecular Dynamics, Sunnyvale, CA). Quantification of Id1, Id2 and Id3 protein was calculated by normalizing the specific protein bands to β-actin using Image Quant 5.5 software (Molecular Dynamics).
Analysis of BMP-6 messenger RNA (mRNA) expression
Endogenous expression of the BMP-6 gene was examined by reverse transcription-polymerase chain reaction. Total RNA was isolated using Absolutely RNA™ RT-PCR Miniprep Kit (Stratagene Europe, Amsterdam, Netherland) according to the manufacturers instructions. Quantification of the isolated total RNA was achieved by using spectrophotometric OD260 measurements. Equal amounts of RNA were then reverse transcribed to cDNA with TaqMan® Reverse Transcription Reagents (Applied Biosystems, Foster City, CA, U.S.A). To measure mRNA expression of BMP6, Id1–Id4 and PGK1 PCR were carried out with TaqMan® universal master mix. Primers and probes were provided by Assay-on-Demand (Applied Biosystems). PCR reactions were carried out in a final volume of 25 μl (BMP-6) or 20 μl (ID1). The cDNA added to each reaction was equivalent to the input of 20 ng of total RNA. The gene expression was quantified using the standard curve method (BMP6), or the comparative CT method (Id1) as described in ABI7700 User Bulletin 2 (Applied Biosystems). The expression was then normalized to the expression level of PGK1. PGK1 was chosen, because it has been shown to have low expression variability among lymphocyte specimens [56]. Expression levels in B cells were then related to the expression levels in Ramos cells.
Cell proliferation
For estimation of DNA synthesis, CD19+ cells (7.5 × 104 cells/0.2 ml) or Ramos cells (1 × 104 cells/0.2 ml) were cultured in triplicate in microtiter wells. The cells were pulsed with 3.7 × 104Bq [3H]thymidine (Amersham, Buckinghamshire, UK) for the last 16 h of a 72-h incubation. The cells were harvested using an automated cell harvester (Packard Instrument Company, Meriden, CT, USA) and [3H]thymidine incorporation was determined in a scintillation counter (TopCount, Packard Instrument Company Inc., Meriden, CT).
Determination of cell death
Cell death was measured by vital dye exclusion test by staining cells with 5 μg/ml propidium iodide ([PI]; Calbiochem Corp.; La Jolla, CA; 5 mg/ml) for one minute on ice. At least 1,000 cells per sample were run on a BD FACSCalibur flow cytometer.
Statistical analysis
The statistical significance of differences between groups was determined using the paired two-tailed Wilcoxon nonparametric test, by applying SPSS10.1 software (SPSS Inc., Chicago, IL, USA). P values less than 0.05 were considered significant.
List of abbreviations
BMP, Bone morphogenetic protein
DLBCL, diffuse large B cell lymphoma
Id, inhibitor of dna binding
Smad, the name Smad originates from the Drosophila protein, MAD and the Caenorhabditis elegans proteins, sma-2, 3 and 4
STAT, signal transducer and activator of transcription
Authors' contributions
CK designed and conducted experiments, oversaw research, and wrote paper. EAS designed and conducted experiments, oversaw research. MEH designed and conducted experiments, oversaw research. LF designed and conducted experiments. EBS designed experiments, oversaw research, and wrote paper. JHM designed and conducted experiments, oversaw research, and wrote paper.
Acknowledgements
This work was supported by the Norwegian Cancer Society and the Norwegian Research Council.
Figures and Tables
Figure 1 BMP-6 inhibits proliferation of human B cells. A) CD19+ B cells were isolated from peripheral blood and stimulated in triplicates with anti-IgM (37.5 μg/ml) or anti-IgM and CD40 ligand (CD40L; 10 ng/ml) in the presence or absence of BMP-6 for 72 hours. DNA synthesis was measured by [3H]-thymidine incorporation for the last 18 hours. Data are given as relative proliferation obtained by normalizing the mean counts per minute (cpm) for each stimulation to the mean cpm obtained for anti-IgM stimulated cells ± SEM. (mean cpm = 25 352 for anti-IgM stimulated cells, *p ≤ 0.0002 (n = 8), **p ≤ 0.023 (n = 6). B) Dose dependent inhibition of BMP-6 of anti-IgM induced DNA-synthesis of CD19+ B cells (relative proliferation ± SEM, n = 3) and C) the Burkitt lymphoma cell line Ramos (relative proliferation ± SEM, n = 3). Ramos cells were cultured for 72 hours and [3H]-thymidine were added for the last 4 hours. D) Noggin (5 μg/ml) and BMP-6 (0,25 or 1 μg/ml) were preincubated for 1 h at 37°C and then added to the CD19+ B cells in the presence of anti-IgM(37.5 μg/ml). Cells were cultured for 72 h and DNA synthesis was measured by 3H-thymidine incorporation. One representative of three separate experiments is shown (mean cpm ± SD of triplicates). E) Highly purified CD19+CD27- or CD19+CD27+ cells were obtained by cell sorting of CD19+ cells and stimulated with anti-IgM in the presence or absence of BMP-6, as indicated for 72 hours. DNA synthesis was measuredby [3H]-thymidine incorporation. Data are given as relative proliferation obtained by normalizing the mean cpm for each stimulation to the mean cpm obtained for anti-IgM stimulated cells (mean cpm = 18 221 for CD19+CD27- naïve B cells, mean cpm = 8 930 for CD19+CD27+ memory B cells, n = 5; * p ≤ 0,004, **p ≤ 0.001).
Figure 2 BMP-6 induces cell death in B cells. CD19+CD27- naïve B cells or CD19+CD27+ memory B cells were cultured for 48 h with BMP-6 (1 μg/ml) with or without anti-IgM (37.5 μg/ml). Cell death (PI+ cells) was then measured by flow cytometry analysis. Data are shown as mean percentage PI+ cells from five independent donors (± SEM; p ≤ 0,003).
Figure 3 Ramos cells were cultured in the presence or absense of BMP-6 (1 μg/ml) for 48 h before analysis of cell death by PI staining. Data are shown as mean percentage PI+ cells (± SEM, n = 3, p ≤ 0,001).
Figure 4 CD19+ B cells express BMP type I and type II receptors. Total cell lysates were prepared from the myeloid cell line HL60, the cell line Ramos or CD19+ B cells isolated from peripheral blood and analyzed by western blot for expression of BMP type I and type II receptors. One representative experiment of four is shown.
Figure 5 BMP-6-induced phosphorylation of Smad1/5/8 in CD19+ B cells and Ramos, but not in HL60 cells. CD19+ B cells, or the cell lines Ramos or HL60
Figure 6 were cultured in X-vivo 15 over night before treatment with BMP-6 for 30 minutes, or for the indicated time points before total cell lysates were prepared. The amount of phosphorylated Smad 1/5/8 was determined by western-blot analysis. The membranes were reprobed for Smad1. One representative experiment of three is shown.
Figure 7 BMP-6 induced upregulation of Id1 at the mRNA and protein level. CD19+ B cells were cultured in X-vivo 15 over night before treatment with BMP-6 for the indicated time points. Total RNA was extracted and Id1, Id2 or Id3-expression was analysed by real-time RT-PCR; values are normalised to the expression level of PGK1 mRNA and expressed as relative quantification (2-average ΔΔCT – relative to Id1, Id2 or Id3 expression in the cell-line Ramos). One representative of three independent experiments is shown.
Figure 8 Id protein expression was determined by western-blot analysis. CD19+ B cells were cultured in X-vivo 15 over night before treatment with BMP-6 for the indicated time points and cell lysates were prepared. One representative experiment of three is shown. HeLa cells were used as a positive control for Id1 protein detection.
Figure 9 Relative protein expression of Id1, Id2 and Id3. Quantifications of Id1, Id2 and Id3 protein levels were performed using β-actin as normalization and expressed as mean ± SEM (Id1: n = 4, *p ≤ 0.020; Id2 and Id3: n = 3).
Figure 10 Anti-IgM rapidly upregulates BMP-6 mRNA expression in human B-cells. CD19-positive B cells were cultured in X-vivo 15 over night and stimulated with anti-IgM for the indicated time periods, before total RNA was extracted. The expression of BMP-6 mRNA by realtime RT-PCR; values are normalised to the expression level of PGK1 mRNA and expressed as relative quantification (2-average ΔΔCT – relative to BMP-6 expression in the cell-line Ramos. One representative of five independent experiments is shown.
Figure 11 Fetal calf serum and human AB serum upregulates BMP-6 mRNA expression in human B-cells. CD19-positive B cells were cultured in X-vivo 15 over night and with fetal calf serum or human AB serum at the indicated dilutions for four hours, before total RNA was extracted. The expression of BMP-6 mRNA by realtime RT-PCR; values are normalised to the expression level of PGK1 mRNA and expressed as relative quantification (2-average ΔΔCT – relative to BMP-6 expression in the cell-line Ramos. One representative of three independent experiments is shown.
==== Refs
Chen D Zhao M Harris SE Mi Z Signal transduction and biological functions of bone morphogenetic proteins Front Biosci 2004 9 349 358 14766372
Urist MR Bone: formation by autoinduction Science 1965 150 893 899 5319761
Wozney JM Rosen V Celeste AJ Mitsock LM Whitters MJ Kriz RW Hewick RM Wang EA Novel regulators of bone formation: molecular clones and activities Science 1988 242 1528 1534 3201241
Kim SJ Letterio J Transforming growth factor-beta signaling in normal and malignant hematopoiesis Leukemia 2003 17 1731 1737 12970772 10.1038/sj.leu.2403069
Bhatia M Bonnet D Wu D Murdoch B Wrana J Gallacher L Dick JE Bone morphogenetic proteins regulate the developmental program of human hematopoietic stem cells J Exp Med 1999 189 1139 1148 10190905 10.1084/jem.189.7.1139
Hager-Theodorides AL Outram SV Shah DK Sacedon R Shrimpton RE Vicente A Varas A Crompton T Bone morphogenetic protein 2/4 signaling regulates early thymocyte differentiation J Immunol 2002 169 5496 5504 12421925
Graf D Nethisinghe S Palmer DB Fisher AG Merkenschlager M The developmentally regulated expression of Twisted gastrulation reveals a role for bone morphogenetic proteins in the control of T cell development J Exp Med 2002 196 163 171 12119341 10.1084/jem.20020276
Ahmed N Sammons J Carson RJ Khokher MA Hassan HT Effect of bone morphogenetic protein-6 on haemopoietic stem cells and cytokine production in normal human bone marrow stroma Cell Biol Int 2001 25 429 435 11401330 10.1006/cbir.2000.0662
Kawamura C Kizaki M Yamato K Uchida H Fukuchi Y Hattori Y Koseki T Nishihara T Ikeda Y Bone morphogenetic protein-2 induces apoptosis in human myeloma cells with modulation of STAT3 Blood 2000 96 2005 2011 10979940
Hjertner O Hjorth-Hansen H Borset M Seidel C Waage A Sundan A Bone morphogenetic protein-4 inhibits proliferation and induces apoptosis of multiple myeloma cells Blood 2001 97 516 522 11154231 10.1182/blood.V97.2.516
Baade RT Utne HR Brenne AT Hjorth-Hansen H Waage A Hjertner O Sundan A Borset M Bone morphogenetic protein-5, -6 and -7 inhibit growth and induce apoptosis in human myeloma cells Oncogene 2004 23 3024 3032 14691444 10.1038/sj.onc.1207386
Rosenwald A Wright G Chan WC Connors JM Campo E Fisher RI Gascoyne RD Muller-Hermelink HK Smeland EB Giltnane JM Hurt EM Zhao H Averett L Yang L Wilson WH Jaffe ES Simon R Klausner RD Powell J Duffey PL Longo DL Greiner TC Weisenburger DD Sanger WG Dave BJ Lynch JC Vose J Armitage JO Montserrat E Lopez-Guillermo A Grogan TM Miller TP LeBlanc M Ott G Kvaloy S Delabie J Holte H Krajci P Stokke T Staudt LM The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma N Engl J Med 2002 346 1937 1947 12075054 10.1056/NEJMoa012914
Waite KA Eng C From developmental disorder to heritable cancer: it's all in the BMP/TGF-beta family Nat Rev Genet 2003 4 763 773 14526373 10.1038/nrg1178
Ying QL Nichols J Chambers I Smith A BMP induction of Id proteins suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with STAT3 Cell 2003 115 281 292 14636556 10.1016/S0092-8674(03)00847-X
Klein U Rajewsky K Kuppers R Human immunoglobulin (Ig)M+IgD+ peripheral blood B cells expressing the CD27 cell surface antigen carry somatically mutated variable region genes: CD27 as a general marker for somatically mutated (memory) B cells J Exp Med 1998 188 1679 1689 9802980 10.1084/jem.188.9.1679
Aoki H Fujii M Imamura T Yagi K Takehara K Kato M Miyazono K Synergistic effects of different bone morphogenetic protein type I receptors on alkaline phosphatase induction J Cell Sci 2001 114 1483 1489 11282024
Miyazawa K Shinozaki M Hara T Furuya T Miyazono K Two major Smad pathways in TGF-beta superfamily signalling Genes Cells 2002 7 1191 1204 12485160 10.1046/j.1365-2443.2002.00599.x
Grimsrud CD Romano PR D'Souza M Puzas JE Reynolds PR Rosier RN O'Keefe RJ BMP-6 is an autocrine stimulator of chondrocyte differentiation J Bone Miner Res 1999 14 475 482 10234567
Shepherd TG Nachtigal MW Identification of a putative autocrine bone morphogenetic protein-signaling pathway in human ovarian surface epithelium and ovarian cancer cells Endocrinology 2003 144 3306 3314 12865307 10.1210/en.2003-0185
Detmer K Steele TA Shoop MA Dannawi H Lineage-restricted expression of bone morphogenetic protein genes in human hematopoietic cell lines Blood Cells Mol Dis 1999 25 310 323 10660478 10.1006/bcmd.1999.0259
Winkler DG Yu C Geoghegan JC Ojala EW Skonier JE Shpektor D Sutherland MK Latham JA Noggin and sclerostin bone morphogenetic protein antagonists form a mutually inhibitory complex J Biol Chem 2004 279 36293 36298 15199066 10.1074/jbc.M400521200
Haudenschild DR Palmer SM Moseley TA You Z Reddi AH Bone morphogenetic protein (BMP)-6 signaling and BMP antagonist noggin in prostate cancer Cancer Res 2004 64 8276 8284 15548695
Xu WP Shiba H Mizuno N Uchida Y Mouri Y Kawaguchi H Kurihara H Effect of bone morphogenetic proteins-4, -5 and -6 on DNA synthesis and expression of bone-related proteins in cultured human periodontal ligament cells Cell Biol Int 2004 28 675 682 15516325 10.1016/j.cellbi.2004.06.004
Song JJ Celeste AJ Kong FM Jirtle RL Rosen V Thies RS Bone morphogenetic protein-9 binds to liver cells and stimulates proliferation Endocrinology 1995 136 4293 4297 7664647 10.1210/en.136.10.4293
Otsuka F Yao Z Lee T Yamamoto S Erickson GF Shimasaki S Bone morphogenetic protein-15. Identification of target cells and biological functions J Biol Chem 2000 275 39523 39528 10998422 10.1074/jbc.M007428200
Ishisaki A Yamato K Hashimoto S Nakao A Tamaki K Nonaka K ten Dijke P Sugino H Nishihara T Differential inhibition of Smad6 and Smad7 on bone morphogenetic protein- and activin-mediated growth arrest and apoptosis in B cells J Biol Chem 1999 274 13637 13642 10224135 10.1074/jbc.274.19.13637
Nohe A Keating E Knaus P Petersen NO Signal transduction of bone morphogenetic protein receptors Cell Signal 2004 16 291 299 14687659 10.1016/j.cellsig.2003.08.011
Nakahara T Tominaga K Koseki T Yamamoto M Yamato K Fukuda J Nishihara T Growth/differentiation factor-5 induces growth arrest and apoptosis in mouse B lineage cells with modulation by Smad Cell Signal 2003 15 181 187 12464389 10.1016/S0898-6568(02)00088-8
Kimura N Matsuo R Shibuya H Nakashima K Taga T BMP2-induced apoptosis is mediated by activation of the TAK1-p38 kinase pathway that is negatively regulated by Smad6 J Biol Chem 2000 275 17647 17652 10748100 10.1074/jbc.M908622199
Nakashima K Yanagisawa M Arakawa H Kimura N Hisatsune T Kawabata M Miyazono K Taga T Synergistic signaling in fetal brain by STAT3-Smad1 complex bridged by p300 Science 1999 284 479 482 10205054 10.1126/science.284.5413.479
Valdimarsdottir G Goumans MJ Rosendahl A Brugman M Itoh S Lebrin F Sideras P ten Dijke P Stimulation of Id1 expression by bone morphogenetic protein is sufficient and necessary for bone morphogenetic protein-induced activation of endothelial cells Circulation 2002 106 2263 2270 12390958 10.1161/01.CIR.0000033830.36431.46
Ruzinova MB Benezra R Id proteins in development, cell cycle and cancer Trends Cell Biol 2003 13 410 418 12888293 10.1016/S0962-8924(03)00147-8
Saisanit S Sun XH Regulation of the pro-B-cell-specific enhancer of the Id1 gene involves the C/EBP family of proteins Mol Cell Biol 1997 17 844 850 9001238
Wilson RB Kiledjian M Shen CP Benezra R Zwollo P Dymecki SM Desiderio SV Kadesch T Repression of immunoglobulin enhancers by the helix-loop-helix protein Id: implications for B-lymphoid-cell development Mol Cell Biol 1991 11 6185 6191 1944284
Sun XH Constitutive expression of the Id1 gene impairs mouse B cell development Cell 1994 79 893 900 8001126 10.1016/0092-8674(94)90078-7
Kee BL Rivera RR Murre C Id3 inhibits B lymphocyte progenitor growth and survival in response to TGF-beta Nat Immunol 2001 2 242 247 11224524 10.1038/85303
Sugai M Gonda H Kusunoki T Katakai T Yokota Y Shimizu A Essential role of Id2 in negative regulation of IgE class switching Nat Immunol 2003 4 25 30 12483209 10.1038/ni874
Sugai M Gonda H Nambu Y Yokota Y Shimizu A Role of Id proteins in B lymphocyte activation: new insights from knockout mouse studies J Mol Med 2004 2 592 599
Sikder HA Devlin MK Dunlap S Ryu B Alani RM Id proteins in cell growth and tumorigenesis Cancer Cell 2003 3 525 530 12842081 10.1016/S1535-6108(03)00141-7
Engel I Murre C The function of E- and Id proteins in lymphocyte development Nat Rev Immunol 2001 1 193 199 11905828 10.1038/35105060
Sikder HA Devlin MK Dunlap S Ryu B Alani RM Id proteins in cell growth and tumorigenesis Cancer Cell 2003 3 525 530 12842081 10.1016/S1535-6108(03)00141-7
Norton JD ID helix-loop-helix proteins in cell growth, differentiation and tumorigenesis J Cell Sci 2000 113 3897 3905 11058077
Gortz B Hayer S Redlich K Zwerina J Tohidast-Akrad M Tuerk B Hartmann C Kollias G Steiner G Smolen JS Schett G Arthritis induces lymphocytic bone marrow inflammation and endosteal bone formation J Bone Miner Res 2004 19 990 998 15125796
Lories RJ Derese I Ceuppens JL Luyten FP Bone morphogenetic proteins 2 and 6, expressed in arthritic synovium, are regulated by proinflammatory cytokines and differentially modulate fibroblast-like synoviocyte apoptosis Arthritis Rheum 2003 48 2807 2818 14558086 10.1002/art.11389
Sakurai D Yamaguchi A Tsuchiya N Yamamoto K Tokunaga K Expression of ID family genes in the synovia from patients with rheumatoid arthritis Biochem Biophys Res Commun 2001 284 436 442 11394898 10.1006/bbrc.2001.4974
Bobacz K Gruber R Soleiman A Erlacher L Smolen JS Graninger WB Expression of bone morphogenetic protein 6 in healthy and osteoarthritic human articular chondrocytes and stimulation of matrix synthesis in vitro Arthritis Rheum 2003 48 2501 2508 13130469 10.1002/art.11248
Schiemann WP Pfeifer WM Levi E Kadin ME Lodish HF A deletion in the gene for transforming growth factor beta type I receptor abolishes growth regulation by transforming growth factor beta in a cutaneous T-cell lymphoma Blood 1999 94 2854 2861 10515889
Yotnda P Garcia F Peuchmaur M Grandchamp B Duval M Lemonnier F Vilmer E Langlade-Demoyen P Cytotoxic T cell response against the chimeric ETV6-AML1 protein in childhood acute lymphoblastic leukemia J Clin Invest 1998 102 455 462 9664088
Wach S Schirmacher P Protschka M Blessing M Overexpression of bone morphogenetic protein-6 (BMP-6) in murine epidermis suppresses skin tumor formation by induction of apoptosis and downregulation of fos/jun family members Oncogene 2001 20 7761 7769 11753654 10.1038/sj.onc.1204962
Amendt C Schirmacher P Weber H Blessing M Expression of a dominant negative type II TGF-beta receptor in mouse skin results in an increase in carcinoma incidence and an acceleration of carcinoma development Oncogene 1998 17 25 34 9671311 10.1038/sj.onc.1202161
Liu Y Titus L Barghouthi M Viggeswarapu M Hair G Boden SD Glucocorticoid regulation of human BMP-6 transcription Bone 2004 35 673 681 15336603 10.1016/j.bone.2004.04.017
Su AI Cooke MP Ching KA Hakak Y Walker JR Wiltshire T Orth AP Vega RG Sapinoso LM Moqrich A Patapoutian A Hampton GM Schultz PG Hogenesch JB Large-scale analysis of the human and mouse transcriptomes Proc Natl Acad Sci U S A 2002 99 4465 4470 11904358 10.1073/pnas.012025199
Glienke J Schmitt AO Pilarsky C Hinzmann B Weiss B Rosenthal A Thierauch KH Differential gene expression by endothelial cells in distinct angiogenic states Eur J Biochem 2000 267 2820 2830 10785405 10.1046/j.1432-1327.2000.01325.x
Champagne CM Takebe J Offenbacher S Cooper LF Macrophage cell lines produce osteoinductive signals that include bone morphogenetic protein-2 Bone 2002 30 26 31 11792561 10.1016/S8756-3282(01)00638-X
Rasmussen AM Smeland EB Erikstein BK Caignault L Funderud S A new method for detachment of Dynabeads from positively selected B lymphocytes J Immunol Methods 1992 146 195 202 1371532 10.1016/0022-1759(92)90228-L
Lossos IS Czerwinski DK Wechser MA Levy R Optimization of quantitative real-time RT-PCR parameters for the study of lymphoid malignancies Leukemia 2003 17 789 795 12682639 10.1038/sj.leu.2402880
| 15877825 | PMC1134658 | CC BY | 2021-01-04 16:03:32 | no | BMC Immunol. 2005 May 9; 6:9 | utf-8 | BMC Immunol | 2,005 | 10.1186/1471-2172-6-9 | oa_comm |
==== Front
BMC Mol BiolBMC Molecular Biology1471-2199BioMed Central London 1471-2199-6-111588820210.1186/1471-2199-6-11Research ArticleInvestigation of factors responsible for cell line cytoplasmic expression differences Finn Jonathan D [email protected] Tabitha [email protected] Pieter R [email protected] Department of Biochemistry and Molecular Biology, University of British Columbia, Vancouver, B.C., Canada2 Inex Pharmaceuticals Corp., Burnaby, B.C., Canada2005 11 5 2005 6 11 11 5 1 2005 11 5 2005 Copyright © 2005 Finn et al; licensee BioMed Central Ltd.2005Finn 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
Previous work has described a novel cytoplasmic expression system that results in a 20-fold increase in the levels of gene expression over a standard CMV-based nuclear expression system, as compared with a 2–3 fold increase seen with previous similar systems. While this increase was seen with BHK and Neuro-2a cells, further studies revealed that some cell lines, such as COS-7, demonstrated relatively poor levels of cytoplasmic expression. The objective of this study was to determine what factors were responsible for the different expression levels between BHK (a high expressing cell line) and COS-7 (a low expressing cell line).
Results
The main findings of this work are that the individual elements of the cytoplasmic expression system (such as the T7 RNAP gene and Internal Ribosome Entry Sequence) are functioning similarly in both cell types. Both cell types were found to have the same amount of cytosolic nuclease activity, and that the cells appeared to have differences in the intra-cellular processing of DNA -cationic lipid complexes.
Conclusion
After exploring many factors, it was found that differences in the intra-cellular processing of the DNA-cationic lipid complex was the most probable factor responsible for the difference in cytoplasmic gene expression.
==== Body
Background
Previous work has described a novel cytoplasmic expression system that results in a 20-fold increase in the levels of gene expression over a standard CMV-based nuclear expression system [1], as compared with a 2–3 fold increase seen with previous similar systems [2]. While this increase was seen with BHK and Neuro-2a cells, further studies revealed that some cell lines, such as COS-7, demonstrated relatively poor levels of cytoplasmic expression.
Cytosolic nucleases have been discovered and characterized [3,4]. These proteins have found to be an important factor with respect to the cytosolic half-life of plasmid based vectors. These nucleases could be important to researchers using plasmid based delivery systems, especially cytoplasmic gene expression systems.
Internal Ribosome Entry Sequences (IRES) are elements that have been shown to drive expression of the second gene in a bi-cistronic mRNA transcript, as well as increase the translation of un-capped transcripts. IRES elements were first isolated from viral genomes, such as the encephalomyocarditis virus (EMCV), where they allow the translation of viral mRNA transcripts, made in the cytoplasm, and thus lacking the 5' cap structure essential for mRNA translation [5]. Various viral IRES elements, such as the EMCV, FMDV (foot and mouth disease virus), and other picornavirus based IRES elements share similar features. All are approx 450 bp long and share a conserved secondary structure, as well as a pyrimidine-rich tract that starts ~25 bp before the 3' end of the IRES [6-8]. The secondary structure is hypothesized to allow for proper alignment of ribosome subunits and other co-factors necessary for translation.
The objective of this study was to determine what factors were responsible for the different expression levels between BHK (a high expressing cell line) and COS-7 (a low expressing cell line). After investigating numerous factors, including the Internal Ribosome Entry Sequence (IRES) element, T7 RNAP activity, cytosolic nuclease levels and intra-cellular processing, it was found that differences in the intra-cellular processing of the DNA-cationic lipid complex was the most probable factor responsible for the difference in gene expression.
Results
Autogene transfection in COS-7 cells results in lower transgene expression than in BHK cells
In previous studies, it was found that the enhanced dual promoter autogene system demonstrated transgene expression levels that were 15–20 times higher than the equivalent nuclear control [1]. This was demonstrated for both BHK and Neuro-2a cells. However, when the system was tested in the COS-7 cell line, the autogene system resulted in levels of gene expression that were lower than the nuclear control (Figure 1). It had been hypothesized that the autogene system should produce high levels of gene expression in any transfected cell. Possible factors that would affect autogene expression levels including IRES function, T7 RNAP activity, cytosolic nuclease levels, and intra-cellular processing differences were therefore examined.
EMCV IRES is functional in COS-7 cells
The first parameter that was tested was whether or not the EMCV IRES was functional in COS-7 cells. In order to determine this, mRNA was synthesized in vitro and transfected directly into either BHK or COS-7 cells. As is observed in Figure 2, the pattern of gene expression is similar for both cell lines. This indicates that while the IRES containing transcripts are not translated as efficiently as capped transcripts [9], the EMCV IRES has the same relative expression pattern in both of the cell types, indicating that differences in IRES function are most likely not the reason for the poor expression levels in COS-7 cells. It is interesting to note that the while the pattern of gene expression is similar, the absolute levels of gene expression are much lower in the COS-7 cells than the BHK cells. While it is known that cell division is an important factor for plasmid transfection, the BHK and COS-7 cells were found to have similar doubling times (data not shown), ruling this out as a factor responsible for the low expression levels in the COS-7 cells. Another reason could be because COS-7 cells are transfected (nucleic acid taken up and delivered into the cytoplasm or nucleus and subsequently expressed) less efficiently than BHK cells, or the intracellular processing of the nucleic acid-lipid complexes are different in the two cell lines, with the BHK cells having a greater amount of intact nucleic acid being released into the cytoplasm than the COS-7 cells. This possible difference in nucleic acid:cationic lipid complex processing could be responsible for the low levels of cytoplasmic autogene expression. It should be noted that we have used both BHK and COS-7 cells extensively and have determined that both cell lines are transfected with similar efficiency (~20%, data not shown), and that this efficiency is not dependent on the plasmid used.
T7 RNAP is functional in COS-7 cells
The next parameter that was examined was whether the T7 RNA polymerase (that is the basis of the cytoplasmic expression system) is active in COS-7 cells. Therefore, in vitro synthesized T7 RNAP mRNA was transfected directly into BHK or COS-7 cells. Twenty-four hours later cells were transfected with a luciferase reporter plasmid that would only express luciferase if functional T7 RNAP was being produced in those cells. As can be seen in Figure 3, while the levels of luciferase expression were lower in the COS-7 cells, these lower levels are consistent with the overall lower expression levels also seen in Figure 2. Note that in the absence of T7 RNAP pretreatment, no luciferase was detected. This data demonstrates that the T7 RNAP was indeed functional in the COS-7 cells. Therefore, these results indicate that the activity of the T7 RNAP gene was most likely not the reason for the poor expression levels.
Effect of T7 RNAP mRNA pretreatment on autogene expression
Since it appeared as if all of the individual components of the autogene expression system were functioning in the COS-7 cells, it was suggested that perhaps the poor expression levels were due to temporal constraints. The cytoplasmic expression system is dependent on the initial source of T7 RNAP coming from the nucleus via the CMV promoter, which is then able to 'trigger' the cytoplasmic system. Perhaps by the time the first T7 RNAP proteins were produced via the nuclear transcripts (at least 8–12 hours post transfection), cytosolic nucleases had degraded the cytoplasmic plasmid, therefore leaving no substrate for the T7 RNAP. In order to test this hypothesis, BHK or COS-7 cells were transfected with T7 RNAP mRNA. Twenty-four hours post mRNA transfection cells were transfected with a plasmid encoding either the cytoplasmic expression system (R011) or the nuclear control (L053). As is apparent in Figure 4, the pretreatment of the COS-7 cells led to a 10 fold increase in autogene based expression that was not seen with the nuclear control. It was also observed that only about a 2–3 fold increase in expression was seen with the BHK cells (data not shown). This result appeared to support the hypothesis that timing was a major factor. If the cells were already primed with T7 RNAP protein prior to autogene expression, a substantial increase in gene expression was seen. These results suggest that certain factors affecting the amount of free cytoplasmic plasmid (such as cytoplasmic nuclease levels or the intracellular processing of complexes) are playing an important role in autogene expression differences between the BHK and COS-7 cells.
BHK and COS-7 cells have similar levels of cytosolic nucleases
In order to understand the difference between the high expressing (BHK) and low expressing (COS-7) cell lines, it was hypothesized that perhaps the BHK cells had lower levels of cytosolic nuclease, therefore more plasmid survived in the cytoplasm during the 8–12 h lag period before T7 RNAP was produced from the nuclear mRNA. Therefore, cytoplasmic extracts of both cell lines were obtained and subjected to nuclease assays. As can be seen in Figure 5, it appears as if the BHK cells did not have lower levels of cytoplasmic nuclease activity than the COS-7 cells, and may even appear to have slightly higher nuclease activity. Therefore, it appears as if the levels of cytosolic nuclease are not a major determining factor with respect to autogene based cytoplasmic expression.
Intracellular processing of DNA is different in BHK and COS-7 cells
While it had been determined that the low levels of autogene expression in COS-7 cells was not due to increased levels of cytosolic nucleases, it could occur due to differences in the intracellular processing of DNA-lipid complexes between the two cell lines. In order to investigate this, plasmid DNA was labeled with a fluorescent label, transfected into both BHK or COS-7 cells, and then florescence microscopy was performed at various time points. As can be seen in Figure 6, while there appears to be a similar level of plasmid taken up into both cell lines at early time points (24 h), at later time points (48 h), the BHK cells seem to have accumulated a significant amount of plasmid in punctate, perinuclear structures, while the COS-7 cells have lost a significant amount of their fluorescence. Thus, it appears as if the BHK cells contain a pool of plasmid DNA for a longer period than the COS-7 cells. This perinuclear pattern seen with the BHK cell is characteristic of DNA-lipid complexes, and previous work [10] has demonstrated that these structures are not associated with lysosomes, and appear to be large structures, presumably formed when multiple endosomes have fused into larger vesicles. These studies clearly show that there is a difference in the intracellular processing of the DNA-lipid complexes between the BHK and COS-7 cells, and this is most likely the reason behind the lower levels of cytoplasmic gene expression seen with the COS-7 cells.
Discussion
While previous work from our lab demonstrated a 15–20 fold increase in levels of gene expression using an autogene based, cytoplasmic expression system [1] in BHK and Neuro-2a cell lines, recent studies showed that the levels of cytoplasmic expression were significantly lower in some other cell lines, such as COS-7. In the present work, an attempt is made to determine what the reason for the difference was. Here various factors are discussed, including the activity of the various components of the autogene system in BHK and COS-7 cells, the finding that pretreating the COS-7 cells with T7 RNAP led to an increase in expression, and that the differences in expression between the two cell lines is most likely due to differences in intracellular processing of the DNA-lipid complexes.
Other studies have shown that IRES function is dependent on trans-acting cellular factors [11]. The pyrimidine tract binding protein (PTB) binds to the EMCV IRES [12], and functions by maintaining the IRES secondary structure, thereby facilitating translation through the IRES [13]. The La antigen has also been demonstrated to bind to the IRES [11,14]. There is also evidence for other proteins binding to various regions of IRES elements [15] that have been reported to have an effect on the IRES function. Therefore, it was thought that perhaps one of the differences between the BHK and COS-7 cells was the presence and/or abundance of various trans-acting cellular factors. However, transfection with various mRNA transcripts showed an identical pattern to that of the BHK cells, indicating that the EMCV IRES was functioning similarly in both of the cell lines, ruling out the IRES as the reason for the expression level differences.
It was then suggested that perhaps the T7 RNAP protein was not active in the COS-7 cells. This would also lead to poor expression levels. Transfecting the cells with capped mRNA encoding T7 RNAP, and then following 24 h later with a plasmid that will only express luciferase in the presence of T7 RNAP demonstrated that this was not the case. Even though the levels of expression were much lower than in BHK cells, a substantial amount of protein was expressed. It appears as if the COS-7 cells in general may not be very amenable to mRNA transfection, as both the transfections with the luciferase encoding mRNA transcripts (Figure 2) as well as the T7 RNAP encoding transcripts (Figure 3) gave much lower expression levels in the COS-7 cells than the BHK cells. This may be indicative of some difference in intracellular processing of the nucleic acid-lipid complex, allowing for less cytoplasmic delivery of mRNA.
As it appeared that the individual components of the autogene system were functioning, it was thought that perhaps the difference was due to a matter of timing. It is known that it takes at least 8–12 h post transfection in order to detect any expression from the nucleus, with expression usually peaking at around 24–48 h post-transfection. Perhaps in this time the majority of cytoplasmic plasmid in the COS-7 cells was degraded by cytosolic nucleases. By priming the cells with T7 RNAP (via pre-transfection of cells with T7 RNAP mRNA) prior to autogene transfection, a 10 fold increase in autogene expression was observed, as opposed to about a 2 fold increase with the nuclear control (most likely due to non-specific effects associated with increased lipid or nucleic acid delivery). Thus it was clear that the lag time between transfection and T7 RNAP expression from the nucleus was playing a role in the poor expression levels in COS-7 cells.
In order to further understand this phenomenon, it was proposed that the COS-7 cells had higher levels of a cytosolic nuclease, and this was degrading the cytoplasmic plasmid, decreasing the amount of intact plasmid left for cytoplasmic expression. However, nuclease assays on cytoplasmic extracts from BHK or COS-7 cells showed that in fact, the COS-7 cytoplasmic extracts had less nuclease activity than the BHK cells. Thus, increased levels of cytosolic nucleases in the COS-7 cells was not the reason for the decreased expression.
The observation that the mRNA transfections gave much lower levels of expression in COS-7 cells than BHK cells hinted at the fact that perhaps there were differences in the intracellular processing of the DNA-lipid complexes, with the COS-7 cells allowing for less free plasmid to be released into the cytoplasm.
By labeling the plasmid with a fluorescent label (FITC) and looking at the internal distribution of plasmid in both BHK and COS-7 cells over time, it was determined that there was a qualitative difference in not only the localization of the labeled plasmid, but also in the amount of plasmid. It appears as if in the COS-7 cells, the plasmids are degraded more quickly, shown by the less intense overall fluorescence as well as the diffuse green staining throughout the cell, indicative of plasmid degradation. This could be due to either a greater number of complexes being transported to lysosomes in COS-7 cells, or that the complexes are being released from the endosomes faster, and therefore being degraded by cytosolic nucleases before the cytoplasmic expression system can be triggered. In contrast, the BHK cells contained larger pools of plasmid DNA in discrete, non-lysosomal, perinuclear structures, even after 48 h. It is hypothesized that perhaps plasmid is being released into the cytoplasm over time, thereby allowing more free cytoplasmic plasmid to be available for cytoplasmic transcription at later time points.
Conclusion
In conclusion, it has been demonstrated that cell line differences in cytoplasmic expression (when compared to nuclear expression) are most likely due to differences in the intra-cellular processing of the nucleic acid-cationic lipid complexes, as opposed to differences in how the individual components (IRES, RNAP, etc.) function.
Methods
Plasmids
Plasmid R011 is a bi-cistronic plasmid consisting of a basic autogene cassette (containing the T7, T3 and SP6 RNAP promoters) driven by a CMV promoter and intron, as well as a downstream Photinus pyralis luciferase reporter gene cassette (for construction details see Finn et al, 2004). L053 consists of the CMV promoter (with intron) from NGVL3 and the Photinus pyralis luciferase gene. L059 consists of a pTRI-Amp (Ambion) backbone with EMCV IRES, Photinus pyralis luciferase and beta-globin poly-adenylation site derived from EMC-Luc, a gift from Dr. Jon Wolff (Waisman Center, Wisconsin). PT7-Luc (Promega) consists of the Photinus pyralis luciferase gene driven by a T7 RNAP promoter. L080 consists of PT7-Luc with the Luciferase gene driven by both the T7 and the SP6 promoter.
mRNA synthesis
mRNA synthesis was performed using a MEGAscript high yield transcription kit (Ambion). For all transcripts used for in vitro transfection, plasmid template (L059 for IRES-Luc and Cap-IRES-Luc mRNA; L080 for Cap-Luc mRNA; PT7-Luc for Luc mRNA) was linearized using EcoR 1 and 1 μg of each plasmid was used as per manufacturers protocols. mRNA was recovered by the LiCl procedure as outlined in the manufacturer's protocol. After LiCl precipitation, the RNA pellet was washed 3 × with 70% ethanol to completely remove any unincorporated NTPs.
Transfections
Lipoplexes were formed by mixing plasmid DNA with large unilamellar vesicles (LUVs) composed of equimolar amounts of DOPE:DODAC (50:50) on ice and incubated for 20 min prior to use. All transfections were performed at a cationic lipid to plasmid DNA charge ratio of 3:1. Lipoplexes were diluted with serum-containing media before addition directly to cell medium. BHK or COS-7 cells were plated at 30,000 cells per well in 24-well plates. The total mass of plasmid added was identical in all transfections. Equimolar transfections using experimental plasmids of different sizes were achieved through the addition of an empty vector (pPUC19) to normalize the total mass of DNA in each transfection. For mRNA transfections, TransMessenger reagent (Qiagen) was used to transfect differing amounts of various mRNA transcripts into BHK cells (30,000 cells per well in 24-well plates). An RNA to TransMessenger reagent ratio of 1:4 was used for all transfections. Transfections were performed as per manufacturer's instructions. Three hours post transfection the medium was removed and replaced with fresh serum containing medium. Cells were harvested and subjected to luciferase assays at time points indicated. For mRNA pretreatment studies, cells were transfected with mRNA as described above, followed by a DNA transfection 24 h later. All transfections were performed in triplicate. Data is presented as mean values +/- standard error.
Luciferase and BCA assays
Cells were washed twice with 1 mL PBS followed by the addition of 0.2 mL lysis buffer (PBS with 0.1 % Triton X-100) before being stored at -70°C. Cells were thawed and 5–20 μl of sample were assayed in duplicate on a 96-well plate. Samples were assayed using a Berthhold Centro LB960 Microplate Luminometer and Luciferase Assay System (Promega). Standard luciferase assays were performed and transfection data is reported as mass quantities of luciferase protein using a standard curve obtained from serial 10-fold dilutions of a 20 mg/mL Photinus pyralis luciferase standard (Promega). Total protein was quantified using a Pierce BCA assay kit as per manufacturer's instructions.
Cytoplasmic extract preparation
Cytoplasmic extracts were prepared using a hypotonic buffer solution. Cells (2 million) were washed with PBS (2 × 1 mL) before a 15 min incubation on ice in 500 μl hypotonic extraction buffer (10 mM Hepes-KOH pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, 0.2 mM PMSF, 0.6% NP-40). The samples were centrifuged at 13,000 × g for 3 min at 4°C and the supernatant was collected and saved as the cytoplasmic fraction. Protein levels were quantitated using a Pierce BCA assay (as described above) after exchanging the hypotonic extraction buffer with PBS using a centrifugal filter device with a MW cut off of 3000.
Nuclease assay procedure
Various amounts of cytoplasmic extract were incubated with plasmid DNA (50–300 ng extract/μg plasmid DNA) at 37°C. At time points indicated, aliquots were removed and flash frozen in liquid nitrogen and stored at -70°C until all assays were completed. Samples were then thawed and run on a 0.8% agarose gel and bands were visualized using ethidium bromide. Nuclease activity defined by the proportion of intact supercoiled plasmid left over time.
Plasmid labeling
R011 (cytoplasmic expression plasmid) was labeled using a Label IT Tracker Fluorescein Kit from Mirus Corporation following manufacturers instructions. A ratio of 1 μl Label IT Tracker Reagent per μg of plasmid DNA was used. Transfections were performed as previously described.
Fluorescence microscopy
At time points indicated, cells were washed 3 × with room temperature PBS before fixing for 5 min with 2% paraformaldeyde. Cells were further washed with 2 × 1 mL PBS before being treated with Prolong Antifade (Molecular Probes) and mounted. Pictures were taken using a Zeiss Axiovert S100 fluorescence microscope. In all cases, exposure times were kept constant to allow for ease of comparison of samples.
Authors' contributions
JF completed all of the experimental design and carried out 90% of the experimental work and wrote 80% of the manuscript. TW completed all of the tissue culture associated with the work. PC has written 20% of the manuscript and was responsible for critically revising it and giving the final verision approval for publication
Acknowledgements
These studies were funded by a grant from the Canadian Institutes of Health Research (CIHR), Inex Pharmaceuticals Corp. and Protiva Biotherapeutics Inc.
Figures and Tables
Figure 1 Autogene expression in COS-7 cells A) Diagrams of plasmids used. B) Comparison of cytoplasmic (R011) and nuclear (L053) expression plasmids in COS-7 cells. COS-7 cells were transfected with 0.75 μg/well of plasmid. Equimolar amounts of plasmids were added, and the total mass of DNA per transfection was kept equal by adding an unrelated plasmid (pPUC19). Transfections and luciferase assays were performed as described in Materials and Methods. Error bars indicate standard error.
Figure 2 Transfection of BHK or COS-7 cells with various mRNA species. Cells were transfected with 0.75 μg/well of mRNA. Transfections and luciferase assays were performed as described in Materials and Methods. Error bars indicate standard error.
Figure 3 T7 RNAP is functional in COS-7 cells. Transfection of BHK or COS-7 cells with 0.75 μg of capped T7-RNAP mRNA for 24 hours, followed by a transfection with 0.75 μg of a luciferase reporter construct (L059) mRNA species. Cells were harvested 20 h post plasmid transfection. Transfections and luciferase assays were performed as described in Materials and Methods. Error bars indicate standard error. Note that in the absence of T7 RNAP (no pre-transfection) no luciferase was detected.
Figure 4 Pretreatment of COS-7 cells with T7 RNAP mRNA. Transfection of COS-7 cells with 0.75 μg of capped T7-RNAP mRNA for 24 hours, followed by a transfection with 0.75 μg of R011 (cytoplasmic) or L053 (nuclear) expression plasmids. Cells were harvested 24 h post plasmid transfection. Transfections and luciferase assays were performed as described in Materials and Methods. Error bars indicate standard error
Figure 5 Cytosolic nuclease assays of BHK and COS-7 cytoplasmic extracts. 200 ng/μg plasmid DNA was incubated and aliquots were taken at time points indicated. No significant difference in nuclease activity between the BHK and COS-7 cytoplasmic extracts was observed.
Figure 6 Intracellular processing of complexes. Transfection of BHK or COS-7 cells with FITC labeled R011. Cells were transfected with 0.75 μg of FITC-R011. Cells were fixed at time points indicated and subjected to phase contrast and fluorescence microscopy.
==== Refs
Finn J Lee AC MacLachlan I Cullis P An enhanced autogene-based dual-promoter cytoplasmic expression system yields increased gene expression Gene Ther 2004 11 276 283 14737087 10.1038/sj.gt.3302172
Brisson M He Y Li S Yang JP Huang L A novel T7 RNA polymerase autogene for efficient cytoplasmic expression of target genes Gene Ther 1999 6 263 270 10435111 10.1038/sj.gt.3300827
Lechardeur D Sohn KJ Haardt M Joshi PB Monck M Graham RW Beatty B Squire J O'Brodovich H Lukacs GL Metabolic instability of plasmid DNA in the cytosol: a potential barrier to gene transfer Gene Ther 1999 6 482 497 10476208 10.1038/sj.gt.3300867
Pollard H Toumaniantz G Amos JL Avet-Loiseau H Guihard G Behr JP Escande D Ca2+-sensitive cytosolic nucleases prevent efficient delivery to the nucleus of injected plasmids J Gene Med 2001 3 153 164 11318114 10.1002/jgm.160
Jang SK Pestova TV Hellen CU Witherell GW Wimmer E Cap-independent translation of picornavirus RNAs: structure and function of the internal ribosomal entry site Enzyme 1990 44 292 309 1966843
Jackson RJ Hunt SL Gibbs CL Kaminski A Internal initiation of translation of picornavirus RNAs Mol Biol Rep 1994 19 147 159 7969103 10.1007/BF00986957
Jackson RJ Kaminski A Internal initiation of translation in eukaryotes: the picornavirus paradigm and beyond Rna 1995 1 985 1000 8595564
Agol VI The 5'-untranslated region of picornaviral genomes Adv Virus Res 1991 40 103 180 1957717
Finn J MacLachlan I Cullis P Factors limiting autogene-based cytoplasmic expression systems FASEB Journal 2004 in press
Zabner J Fasbender AJ Moninger T Poellinger KA Welsh MJ Cellular and molecular barriers to gene transfer by a cationic lipid J Biol Chem 1995 270 18997 19007 7642560 10.1074/jbc.270.32.18997
Belsham GJ Sonenberg N RNA-protein interactions in regulation of picornavirus RNA translation Microbiol Rev 1996 60 499 511 8840784
Kolupaeva VG Hellen CU Shatsky IN Structural analysis of the interaction of the pyrimidine tract-binding protein with the internal ribosomal entry site of encephalomyocarditis virus and foot-and-mouth disease virus RNAs Rna 1996 2 1199 1212 8972770
Kaminski A Jackson RJ The polypyrimidine tract binding protein (PTB) requirement for internal initiation of translation of cardiovirus RNAs is conditional rather than absolute Rna 1998 4 626 638 9622122 10.1017/S1355838298971898
Meerovitch K Svitkin YV Lee HS Lejbkowicz F Kenan DJ Chan EK Agol VI Keene JD Sonenberg N La autoantigen enhances and corrects aberrant translation of poliovirus RNA in reticulocyte lysate J Virol 1993 67 3798 3807 8389906
Hoffman MA Palmenberg AC Revertant analysis of J-K mutations in the encephalomyocarditis virus internal ribosomal entry site detects an altered leader protein J Virol 1996 70 6425 6430 8709275
| 15888202 | PMC1134659 | CC BY | 2021-01-04 16:22:25 | no | BMC Mol Biol. 2005 May 11; 6:11 | utf-8 | BMC Mol Biol | 2,005 | 10.1186/1471-2199-6-11 | oa_comm |
==== Front
BMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 1471-2288-5-151586212410.1186/1471-2288-5-15DebateAssessing subgroup effects with binary data: can the use of different effect measures lead to different conclusions? White Ian R [email protected] Diana [email protected] MRC Biostatistics Unit Institute of Public Health, Robinson Way, Cambridge CB2 2SR, UK2 Medical Statistics Unit London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK2005 29 4 2005 5 15 15 13 12 2004 29 4 2005 Copyright © 2005 White and Elbourne; licensee BioMed Central Ltd.2005White and Elbourne; 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
In order to use the results of a randomised trial, it is necessary to understand whether the overall observed benefit or harm applies to all individuals, or whether some subgroups receive more benefit or harm than others. This decision is commonly guided by a statistical test for interaction. However, with binary outcomes, different effect measures yield different interaction tests. For example, the UK Hip trial explored the impact of ultrasound of infants with suspected hip dysplasia on the occurrence of subsequent hip treatment. Risk ratios were similar between subgroups defined by level of clinical suspicion (P = 0.14), but odds ratios and risk differences differed strongly between subgroups (P < 0.001).
Discussion
Interaction tests on different effect measures differ because they test different null hypotheses. A graphical technique demonstrates that the difference arises when the subgroup risks differ markedly. We consider that the test of interaction acts as a check on the applicability of the trial results to all included subgroups. The test of interaction should therefore be applied to the effect measure which is least likely a priori to exhibit an interaction. We give examples of how this might be done.
Summary
The choice of interaction test is especially important when the risk of a binary outcome varies widely between subgroups. The interaction test should be pre-specified and should be guided by clinical knowledge.
==== Body
Background
Subgroup analysis
Any randomised controlled trial, however tight its inclusion and exclusion criteria, recruits individuals who differ in many observed and unobserved ways. Different individuals are rarely likely to respond to intervention in exactly the same way, so the intervention effect within carefully defined subgroups is of interest. Unfortunately, analysis of a trial within subgroups is usually underpowered: results often have wide confidence intervals and lack statistical significance, even when the trial intervention is beneficial. Further, repeating an analysis within several subgroups greatly increases the risk of false positive findings [1]. Subgroup analysis must therefore be treated cautiously, and "the overall trial result is usually a better guide to the direction of effect in subgroups than the apparent effect observed within a subgroup" [2].
When a trial outcome is binary, a number of different effect measures are available [3]: the risk difference or absolute benefit, which is easily translated into the clinically relevant number needed to treat [4]; the risk ratio, which is widely understood [5]; or the odds ratio, which has desirable statistical properties [6,7]. Further, the risk ratio for benefits gained differs from the risk ratio for harm avoided. Subgroups that are identical on one of these effect measures are not usually identical on a different effect measure. For example, if intervention halves the risk in each subgroup, and the subgroups themselves have different risks, then the risk differences differ between subgroups.
When subgroups have different risks, it is common to estimate an overall risk ratio and then, using the control group risk in each subgroup, to infer the subgroup-specific risk difference and number needed to treat [8]. In this approach, the estimated absolute benefit of intervention is proportional to the control group risk. However, this assumes the risk ratio is equal across subgroups. In making a treatment decision about a particular patient, therefore, the clinician makes best use of the evidence base by ignoring possible differences in the chosen measure of treatment effect between subgroups. It is important to have tools to indicate when this is inappropriate.
Interaction
The statistical test of interaction is a useful tool in this dilemma. In statistical language, interaction is the difference between the intervention effects in different subgroups, and the null hypothesis is that the intervention effect is equal across subgroups [9-12]. A statistically significant interaction supports placing more weight on subgroup-specific findings, especially if it arises from one of a small number of pre-defined subgroup analyses. On the other hand, a non-significant interaction suggests that the overall trial findings should inform individual intervention decisions. The clinical plausibility and importance of the subgroup-specific findings must also be taken into account [13].
It is useful to distinguish a qualitative interaction, in which intervention is beneficial in one subgroup but ineffective or harmful in another, from a quantitative interaction in which intervention is beneficial in all subgroups (or harmful in all) but the degree of benefit varies [14]. As an example of quantitative interaction, a large meta-analysis showed that tamoxifen has benefit in treating both oestrogen-receptor-positive and oestrogen-receptor-negative early breast cancer, but that the benefit is greater in the first group [15]. Such clear-cut statistically significant results are rare in single trials.
When the trial outcome is binary, discussion of interactions is further complicated by the variety of possible effect measures. Significance tests on the overall intervention effect are unaffected by the choice of measure, but the existence and strength of interactions depend on the effect measure used [16]. Quantitative interactions can usually be removed by changing the effect measure, but qualitative interactions cannot be removed in this way.
UK Hip trial
We explore these issues in the context of the UK Hip trial [17]. This trial aimed to show that diagnostic ultrasound in the management of infants with suspected developmental hip dysplasia reduced overall treatment (mainly splinting) without doubling the risk of operative treatment. Note that treatment is an outcome in this trial. Results were reported as risk ratios for operative treatment comparing ultrasound with no ultrasound. The observed risks were 21/314 (7%) and 25/315 (8%) respectively, so the overall risk ratio was 0.84 with a 95% confidence interval from 0.48 to 1.47, suggesting that risk of operative treatment was not doubled. In Table 1 this risk ratio for operative treatment is termed "risk ratio for harm".
Table 1 Operative treatment in the UK Hip trial
Level of clinical suspicion Ultrasound No Ultrasound Risk ratio for harm (95% CI) Risk ratio for benefit (95% CI) Odds ratio (95% CI) Risk difference (95% CI)
All 21/314 (7%) 25/315 (8%) 0.84 (0.48 to 1.47) 1.01 (0.97 to 1.06) 0.83 (0.46 to 1.52) -0.01 (-0.05 to 0.03)
Strong 7/95 (7%) 11/89 (12%) 0.60 (0.24 to 1.47) 1.06 (0.96 to 1.16) 0.56 (0.21 to 1.53) -0.05 (-0.14 to 0.04)
Moderate 14/219 (6%) 14/226 (6%) 1.03 (0.50 to 2.11) 1.00 (0.95 to 1.05) 1.03 (0.48 to 2.22) 0.00 (-0.04 to 0.05)
Test of
interaction1 P = 0.35 P = 0.29 P = 0.34 P = 0.30
1 Mantel-Haenszel test.
Infants fell into two subgroups defined by level of clinical suspicion prior to randomisation: strong suspicion, defined as "sufficient to warrant early prophylactic splinting", or moderate suspicion, defined as "sufficient to warrant further specialist examination". The risk ratio for the moderate-suspicion subgroup (Table 1) has a 95% confidence interval that includes a doubling of the risk of operative treatment. However, the risk ratios are not significantly different on an interaction test (P = 0.35). This suggests using the overall relative risk of 0.84, with its upper confidence limit of 1.47, as applying to both groups.
Alternative analyses based on the other effect measures are shown in the last three columns of Table 1. The risk ratios for benefit (the risk ratios for avoiding operative treatment) are all near 1, because the outcome event is rare. For the same reason, the odds ratio is numerically similar to the risk ratio. The risk difference is numerically very different. However, all four effect measures give very similar significance levels on the interaction test.
A second outcome measure in the UK Hip trial was the occurrence of any hip treatment (Table 2). Clinical suspicion is a strong prognostic factor for this outcome: in the no-ultrasound arm, 97% of the strong-suspicion group but only 32% of the moderate-suspicion group received hip treatment. The risk ratios, risk differences and odds ratios all show a larger effect in the strong-suspicion subgroup. However, the risk ratios for harm do not differ significantly between subgroups, yet the risk ratios for benefit, the risk differences and the odds ratios all have highly statistically significant differences between subgroups.
Table 2 Any hip treatment in the UK Hip trial
Level of clinical suspicion Ultrasound No Ultrasound Risk ratio for harm (95% CI) Risk ratio for benefit (95% CI) Odds ratio (95% CI) Risk difference (95% CI)
All 126/314 (40%) 159/315 (50%) 0.79 (0.67 to 0.95) 1.21 (1.05 to 1.40) 0.66 (0.48 to 0.90) -0.10 (-0.18 to -0.03)
Strong 65/95 (68%) 86/89 (97%) 0.71 (0.61 to 0.82) 9.37 (2.96 to 29.62) 0.08 (0.02 to 0.24) -0.28 (-0.38 to -0.18)
Moderate 61/219 (28%) 73/226 (32%) 0.86 (0.65 to 1.15) 1.07 (0.94 to 1.20) 0.81 (0.54 to 1.21) -0.04 (-0.13 to 0.04)
Test of
interaction1 P = 0.14 P < 0.001 P < 0.001 P < 0.001
1 Mantel-Haenszel test.
Discussion
Why do interaction tests differ?
Interaction tests on different scales differ because they are testing different null hypotheses. In the UK Hip trial, ultrasound reduced the risk of any hip treatment from 97% to 68% in the strong-suspicion group. Under the null hypothesis of a common risk ratio for hip treatment, the 32% risk in the moderate-suspicion group would be reduced to 23%, but the null hypothesis of a common risk difference implies a reduction to 4.1%, and the null hypothesis of a common odds ratio implies a reduction to 3.5%. The observed reduction to 28% is consistent only with the null hypothesis of a common risk ratio. A common risk ratio for avoiding hip treatment is not possible, since the risk ratio is over 9 in the strong-suspicion group, and multiplying the 68% risk of avoiding hip treatment in the strong-suspicion group by 9 would result in a risk over 100%.
When do interaction tests differ markedly?
A graph of the event fraction in the ultrasound group against the event fraction in the no-ultrasound group is helpful (Figure 1) [18]. Points below the diagonal line indicate a lower event fraction in the ultrasound group. The large dots represent the results for the two subgroups. The curved line shows results that have the same odds ratio as the strong-suspicion subgroup, while other lines show results that have the same risk difference or risk ratio. The moderate-suspicion subgroup lies closest to the line of a common risk ratio for harm. It is clear from Figure 1 that the choice of effect measure matters most when the subgroup risks differ markedly.
Figure 1 UK Hip trial: results for any treatment in strong-suspicion and moderate-suspicion subgroups. Points on the red diagonal line indicate lack of effect; points on the other lines indicate the same risk difference (RD), risk ratio for harm (RR(H)), risk ratio for benefit (RR(B)) or odds ratio (OR) as in the strong-suspicion subgroup.
Which interaction test is best?
Since the four interaction tests may differ markedly, it is important to make a careful choice. An intuitive approach is to perform the interaction analysis on the same scale on which the results are to be presented [19]. For example, clinical trial results are often presented on the risk ratio for harm scale, so the interaction analysis would test equality of these risk ratios. However, there are other considerations.
Deeks discussed selection of an effect measure or summary statistic in meta-analysis [16]. His aim was to find a summary statistic that is most plausibly equal in all trials, including those with different control group risks, in order to best predict the treatment benefit for various patient types. One way is to select an effect measure for which the subgroup-specific results are comparable, as judged by the interaction test. In meta-analysis, the effect measure minimising the Q (heterogeneity) statistic could be used. Using this approach in the UK Hip trial, the results for any hip treatment would have been reported on the risk ratio for harm scale, regardless of what had been planned. However, Deeks argues that this is problematic with the typically small number of trials in a meta-analysis. Instead, the choice of effect measure should use both clinical knowledge and empirical evidence. For example, given the clinical view that absolute benefit is likely to be greatest in those with greatest risk, the risk ratio for harm would appear the best effect measure. Empirically, Deeks shows that the risk ratio for harm and the odds ratio are more frequently homogeneous between trials than the risk ratio for benefit and the risk difference, supporting their wider use.
Deeks' arguments apply to meta-analysis, and we would not apply them to subgroup analysis within clinical trials. Instead, we view interaction tests as a check on the applicability of the trial results to all included subgroups. Investigators start with the belief that all subgroups recruited to the trial have qualitatively similar responses to intervention. It is reasonable to maintain that belief if it can be shown to be consistent with the data. We therefore propose that investigators should identify the effect measure that is most likely to be similar between subgroups. By carefully specifying this measure in advance, they ensure that the interaction test has maximum scientific validity.
For example, suppose that the investigators designing the UK Hip trial had predicted that 95% and 30% of the two subgroups would receive treatment in the absence of ultrasound. They could then have asked: if ultrasound reduced the proportion in the first subgroup from 95% to 70%, what effect is likely in the second subgroup? A reduction from 95% to 70% represents a risk ratio for harm of 0.74, which would reduce 30% to 22% in the second subgroup. A reduction from 95% to 70% also represents an odds ratio of 0.12 and a risk difference of 25 percentage points, both of which (by coincidence) would reduce 30% to 5% in the second subgroup. A common risk ratio for benefit is impossible with these numbers, as noted above. The choice among these possibilities will draw on investigators' knowledge and experience. If the investigators believed that all infants are equally able to be saved from treatment, then a common risk ratio for harm would be plausible. If on the other hand the lower treatment rate in the second subgroup implies less pathology and hence greater potential for avoiding treatment, then a common odds ratio might be more plausible.
As another example, consider a trial of a community intervention to promote vaccination. Suppose that the expected unvaccinated fractions in two subgroups are 20% and 80% without the intervention, and that the intervention is expected to halve the unvaccinated fraction in the first subgroup. If the difference between subgroups stems from a lack of previous vaccination campaigns, then all unvaccinated individuals would be equally likely to be vaccinated under the intervention, so the second subgroup would see a reduction to 40% – a common risk ratio of 0.5. But if the difference between subgroups stems from the second subgroup's greater suspicion of vaccination, then their likely reduction would be smaller – perhaps to the figure of 64% which represents a common odds ratio of 0.44.
We would usually prefer primary results to be presented on the scale selected for interaction testing. While it may be appropriate to present subgroup-specific results on a different scale, it would not be correct to use that scale for a single summary measure over the whole trial. For example, subgroups with equal odds ratios usually have unequal risk ratios, so a single summary risk ratio as proposed by Zheng and Yu [20] would generally be inappropriate. Just as a single risk ratio may be used to compute separate risk differences or numbers needed to treat, so a common odds ratio could be used to compute fitted risks in all subgroups and hence to compute appropriate risk ratios, risk differences or numbers needed to treat.
Finally, the choice of effect measure for the interaction test is important because it may affect conclusions about the applicability of the overall results of a clinical trial to all subgroups. This choice should therefore be specified in advance of data analysis. The best place to do this is in a trial protocol or statistical analysis plan.
Conclusion
A statistical test of interaction is important in deciding whether the overall results of a randomised trial apply to all subgroups. When the outcome is binary, different effect measures may lead to very different results on the test of interaction. The choice of effect measure for the test of interaction should therefore be specified before analysis of the data. The best choice of effect measure for the test of interaction is that which the investigators believe is most likely to be similar between subgroups.
Competing interests
DE received funding from the Department of Health, via the Medical Research Council, for the UK Hip Trial [17]. Apart from this, neither author has a competing interest.
Authors' contributions
The original idea arose in discussion between the authors. IRW wrote the first draft. Both authors contributed to subsequent drafts and approved the final version.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Simon Thompson, MRC Biostatistics Unit, and Chris Frost, London School of Hygiene and Tropical Medicine, for their comments on earlier drafts of this paper.
==== Refs
Assmann SF Pocock SJ Enos LE Kasten LE Subgroup analysis and other (mis)uses of baseline data in clinical trials Lancet 2000 355 1064 1069 10744093 10.1016/S0140-6736(00)02039-0
Yusuf S Wittes J Probstfield J Tyroler HA Analysis and interpretation of treatment effects in subgroups of patients in randomized clinical trials Journal of the American Medical Association 1991 266 93 98 2046134 10.1001/jama.266.1.93
Walter SD Choice of effect measure for epidemiological data Journal of Clinical Epidemiology 2000 53 931 9 11004419 10.1016/S0895-4356(00)00210-9
Cook RJ Sackett DL The number needed to treat: a clinically useful measure of treatment effect British Medical Journal 1995 310 452 4 7873954
Altman DG Deeks JJ Sackett DL Odds ratios should be avoided when events are common British Medical Journal 1998 317 1318 9804732
Cox DR Interaction International Statistical Review 1984 52 1 31
Senn S Rare Distinction and Common Fallacy British Medical Journal 1999 accessed 9 August 2004
Furukawa T Guyatt G Griffith L Can we individualize the "number needed to treat"? An empirical study of summary effect measures in meta-analyses International Journal of Epidemiology 2002 31 72 76 11914297 10.1093/ije/31.1.72
Altman DG Matthews JNS Interaction 1: Heterogeneity of effects British Medical Journal 1996 313 486 8776324
Matthews JNS Altman DG Interaction 2: Compare effect sizes not P values British Medical Journal 1996 313 808 8842080
Matthews JNS Altman DG Interaction 3: How to examine heterogeneity British Medical Journal 1996 313 862 8870577
Altman DG Bland JM Statistics Notes: Interaction revisited: the difference between two estimates British Medical Journal 2003 326 219 12543843
Oxman AD Guyatt GH A consumer's guide to subgroup analysis Annals of Internal Medicine 1992 116 78 84 1530753
Peto R Halnan KE Statistical aspects of cancer trials Treatment of Cancer 1982 London:Chapman and Hall 867 871
Cuzick J Powles T Veronesi U Forbes J Edwards R Ashley S Boyle P Overview of the main outcomes in breast-cancer prevention trials Lancet 2003 361 296 300 12559863 10.1016/S0140-6736(03)12342-2
Deeks JJ Issues in the selection of a summary statistic for meta-analysis of clinical trials with binary outcomes Statistics in Medicine 2002 21 1575 1600 12111921 10.1002/sim.1188
Elbourne D Ultrasonography in the diagnosis and management of developmental hip dysplasia (UK Hip trial) Lancet 2002 360 2009 2017 12504396 10.1016/S0140-6736(02)12024-1
Javier Jimenez F Guallar E Martin-Moreno JM A graphical display useful for meta-analysis Eur J Public Health 1997 7 101 105 10.1093/eurpub/7.1.101
Mehrotra DV Stratification issues with binary endpoints Drug Information Journal 2001 35 1343 1350
Zhang J Yu KF What's the relative risk? A method of correcting the odds ratio on cohort studies of common outcomes Journal of the American Medical Association 1998 280 1690 1 9832001 10.1001/jama.280.19.1690
| 15862124 | PMC1134660 | CC BY | 2021-01-04 16:32:51 | no | BMC Med Res Methodol. 2005 Apr 29; 5:15 | utf-8 | BMC Med Res Methodol | 2,005 | 10.1186/1471-2288-5-15 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-311586562610.1186/1471-2202-6-31Research ArticleCo-release of noradrenaline and dopamine in the cerebral cortex elicited by single train and repeated train stimulation of the locus coeruleus Devoto Paola [email protected] Giovanna [email protected] Pierluigi [email protected]à Mauro [email protected] Gian Luigi [email protected] Department of Neuroscience "B.B. Brodie" University of Cagliari, Italy2 Institute of Neuroscience, Consiglio Nazionale delle Ricerche, Section of Cagliari, Cagliari, Italy3 Centre of Excellence "Neurobiology of Addiction", University of Cagliari, Cagliari, Italy2005 2 5 2005 6 31 31 1 12 2004 2 5 2005 Copyright © 2005 Devoto 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
Previous studies by our group suggest that extracellular dopamine (DA) and noradrenaline (NA) may be co-released from noradrenergic nerve terminals in the cerebral cortex. We recently demonstrated that the concomitant release of DA and NA could be elicited in the cerebral cortex by electrical stimulation of the locus coeruleus (LC). This study analyses the effect of both single train and repeated electrical stimulation of LC on NA and DA release in the medial prefrontal cortex (mPFC), occipital cortex (Occ), and caudate nucleus. To rule out possible stressful effects of electrical stimulation, experiments were performed on chloral hydrate anaesthetised rats.
Results
Twenty min electrical stimulation of the LC, with burst type pattern of pulses, increased NA and DA both in the mPFC and in the Occ. NA in both cortices and DA in the mPFC returned to baseline within 20 min after the end of the stimulation period, while DA in the Occ reached a maximum increase during 20 min post-stimulation and remained higher than baseline values at 220 min post-stimulation. Local perfusion with tetrodotoxin (TTX, 10 μM) markedly reduced baseline NA and DA in the mPFC and Occ and totally suppressed the effect of electrical stimulation in both areas.
A sequence of five 20 min stimulations at 20 min intervals were delivered to the LC. Each stimulus increased NA to the same extent and duration as the first stimulus, whereas DA remained elevated at the time next stimulus was delivered, so that baseline DA progressively increased in the mPFC and Occ to reach about 130 and 200% the initial level, respectively.
In the presence of the NA transport (NAT) blocker desipramine (DMI, 100 μM), multiple LC stimulation still increased extracellular NA and DA levels.
Electrical stimulation of the LC increased NA levels in the homolateral caudate nucleus, but failed to modify DA level.
Conclusion
The results confirm and extend that LC stimulation induces a concomitant release of DA and NA in the mPFC and Occ.
The different time-course of LC-induced elevation of DA and NA suggests that their co-release may be differentially controlled.
==== Body
Background
Noradrenergic neurons in the locus coeruleus (LC) project homogeneously throughout the cerebral cortex, while dopaminergic afferents are mainly confined to discrete areas of the cortex, such as the medial prefrontal cortex (mPFC), anterior cingulate, rhinal and entorhinal cortices [1,2]. However, even in cortical areas where dopaminergic and noradrenergic innervations converge, such as the mPFC, tissue noradrenaline (NA) concentrations exceed those of dopamine (DA). Furthermore, in the cerebral cortex a consistent mismatch exists between DA receptor distribution and dopaminergic innervation, DA receptors being widely expressed in non-innervated areas [3-7].
Previous evidence from our laboratory has suggested the possibility that DA in the cerebral cortex may be released other than from dopaminergic terminals also from noradrenergic ones, where DA acts both as NA precursor and co-transmitter [8-12]. This hypothesis is supported by the fact that extracellular DA concentrations in cortical areas scarcely innervated by the meso-cortical dopaminergic pathway, such as the occipital (Occ) and parietal cortices, were found to be only slightly lower than in the mPFC, an area with massive dopaminergic afferents [8].
Moreover, extracellular DA in different cortices was found to be modified by drugs acting on noradrenergic transmission but not, or only modestly, by drugs modifying dopaminergic activity [9-11]. Thus, consistent with its ability to inhibit noradrenergic activity, the α2-adrenoceptor agonist clonidine produced a concomitant reduction in extracellular DA and NA both in the mPFC and Occ, whereas the α2-adrenoceptor antagonists idazoxan and RS 79948 produced a concomitant increase in extracellular NA and DA in both cortices [10,11]. Moreover, local infusion into the LC of carbachol, kainic acid or N-methyl aspartic acid increased both NA and DA in the mPFC, in accord with their ability to activate noradrenergic neurons [13].
In a previous study [12] we demonstrated that extracellular NA and DA in the cerebral cortex are increased by electrical stimulation of the LC in a frequency dependent manner. However, the higher frequency patterns tested elicited a robust behavioural activation, and even at low frequencies animals displayed a state of alert with fine muscle contractions and whisker vibrations. As stress has been demonstrated to increase both NA and DA in the cerebral cortex [14-16], to rule out the possibility that the effect of LC stimulation might be influenced by stress, we performed the present study in anaesthetized rats. Moreover, we studied whether the release of DA and NA might be differentiated by applying single train or repeated train stimulations.
Results
Positioning of the microdialysis probe is schematically depicted in fig. 1, while fig. 2 shows a microphotograph of electrode placement (A) and a schematic representation of the stimulation pattern used (B).
Mean extracellular basal levels of NA and DA, expressed as pg/sample (40 μl) injected on column were: in the mPFC (n = 25) 4.26 ± 0.49 and 4.34 ± 0.46, in the Occ (n = 26) 4.90 ± 0.65 and 4.22 ± 0.58, in the caudate nucleus (n = 6) 1.71 ± 0.16 and 27.35 ± 3.35 for NA and DA, respectively.
Fig 3 shows how LC stimulation increased extracellular NA by 78 and 56%, in the mPFC and Occ, respectively, the levels returning to baseline values within 20 min (ANOVA results: mPFC: F(8,40) = 16.103, P < 0.0001; Occ: F(8,40) = 11.349, P < 0.0001). Extracellular DA concentration in the mPFC increased by 29%, and returned to baseline within 20 min. In the Occ extracellular DA reached a maximal increase of about 54% at 20 min after the end of stimulation and remained elevated by about 30% throughout the experimental period (ANOVA results: mPFC: F(8,40) = 5.043, P = 0.0002; Occ: F(8,40) = 5.386, P = 0.0001).
In order to ensure that extracellular DA represented a nerve activity-mediated exocytotic process, the mPFC and Occ were perfused with the sodium channel blocker tetrodotoxin (TTX,10 μM). TTX markedly decreased extracellular NA and DA (to about 10 and 30% of baseline, respectively) and, as shown in fig. 4, totally suppressed the increase in both catecholamines produced by electrical stimulation.
A sequence of five stimulation trains was administered to a second group of rats, using the same burst pattern as for the single train stimulation, spaced out at 20 min interval (Fig 5). Each successive stimulation increased NA by approximately the same extent and duration as the first stimulation, NA returning to pre-stimulus level within 20 min after stimulation (ANOVA results: mPFC: F(12,48) = 12.329, P < 0.0001; Occ: F(12,72) = 13.086, P < 0.0001). In each interval sample NA levels differed significantly from those of the sample stimulated immediately previously, with the exception of the fifth pause in the mPFC and the second pause in the Occ (P < 0.05, Tukey-Kramer multiple comparison test).
On the other hand, following each stimulation DA remained elevated above the pre-stimulated level when the next stimulation was delivered, therefore the cumulative effect of the repeated stimulation was an increase in DA level by about 40 and 100% in the mPFC and Occ, respectively (ANOVA results: mPFC: F(12,48) = 17.402, P < 0.0001; Occ: F(12,72) = 18.714, P < 0.0001). In the mPFC, DA levels of the third, forth and fifth interval samples resulted significantly different from the sample stimulated previously, while in the Occ no significant difference was found between pauses and stimulations (P < 0.05, Tukey-Kramer multiple comparison test).
In a different group of rats, TTX was locally administered in the Occ after two stimulation periods, when DA level was elevated (Fig 6). This perfusion rapidly decreased extracellular NA and DA levels below basal concentrations, and the subsequent stimulations of LC totally failed to increase extracellular catecholamine levels (ANOVA results: NA: F(12,48) = 104.32, P < 0.0001; DA: F(12,48) = 40.806, P < 0.0001).
To rule out the possibility that electrical stimulation might have produced the concomitant stimulation of the midbrain DA neurons by current spread or trans-synaptically, the effect of repeated burst stimulation of the LC on extracellular DA and NA in the caudate nucleus and the mPFC was compared in a group of rats implanted with a microdialysis probe in the two areas. As shown in Fig. 7, each successive stimulation increased extracellular NA and DA in the mPFC in the same way as in single probe implanted rats (ANOVA results: NA: F(12,60) = 14.309, P < 0.0001; DA: F(12,60) = 7.392, P < 0.0001). During each pause, NA levels were significantly different from the previously stimulated sample, while no difference was found between DA levels (P < 0.05, Tukey-Kramer multiple comparison test).
At variance to the cortex, in the caudate nucleus only extracellular NA levels were increased by stimulations, returning to baseline during each post-stimulus interval, while DA displayed an oscillatory tendency towards a decrease that reached significant values during the last 3 pauses (ANOVA results: NA: F(12,60) = 10.759, P < 0.0001; DA: F(12,60) = 5.000, P < 0.0001). As described for the cortex, during pauses NA levels were significantly different from the previously stimulated levels, with the exception of the fifth interval sample. Moreover, DA levels displayed a significant difference only between the first stimulated sample and the subsequent pause, although no significant difference was evidenced respect to baseline (P < 0.05, Tukey-Kramer multiple comparison test).
A subset of experiments was performed during local perfusion of the NA transporter (NAT) blocker desipramine (DMI, 100 μM) into the mPFC and Occ. After at least 2 h of perfusion, extracellular NA and DA levels had increased stably, to 15.16 ± 1.93 and 8.27 ± 1.23 in the mPFC (n = 7), and to 19.60 ± 3,49 and 10.10 ± 3.53 pg/sample (n = 5) in the Occ, for NA and DA respectively. Multiple electrical stimulation of the LC further increased extracellular NA and DA levels, although in this condition not only DA but also NA failed to return to baseline during stimulation intervals, accumulating a significant increase of about 20% in the mPFC and 50% in the Occ (fig 8) (ANOVA results: mPFC, NA: F(12,48) = 20.743, P < 0.0001, DA: F(12,48) = 2.418, P = 0.0152; Occ, NA: F(12,48) = 21.852, P < 0.0001, DA: F(12,48) = 10.021, P < 0.0001). In the mPFC, each stimulated NA level was significantly different from the subsequent interval sample, while no significance was found in the Occ between stimulated and pause NA levels, as well as for DA levels in either the mPFC or Occ (P < 0.05, Tukey-Kramer multiple comparison test).
Fig 9 shows the results obtained in 3 rats whose data were discarded after histological examination, due to incorrect electrode positioning. In these animals neither extracellular NA nor DA were significantly increased by electrical stimulation, either in the mPFC (ANOVA results: NA: F(12,24) = 1.727, P > 0.05; DA: F(12,24) = 1.854, P < 0.05) or in the caudate nucleus (ANOVA results: NA: F(12,24) = 0.453, P > 0.05; DA: F(12,24) = 1.758, P > 0.05).
Discussion
In this research study on anaesthetised animals we found that extracellular basal levels of NA and DA in the mPFC and Occ were analogous to those previously detected in our lab in freely moving rats. Due to the different probe geometry, which may have affected in vivo recovery of catecholamines, comparison of data collected in the mPFC (by means of vertical probes) with data obtained from Occ (by means of horizontal probes) is not appropriate. However, considering each area separately, it appears that while in the mPFC DA and NA concentrations were of a similar magnitude, consistent with the dense dopaminergic innervation of the mPFC, in the Occ DA resulted slightly lower than NA, but higher than one could expect from the scarce, if any, dopaminergic innervation, confirming our previous observations [8].
Electrical stimulation of the LC produced an increase in both NA and DA concentrations in the cerebral cortex of anaesthetized rats, superimposable to that previously demonstrated by our group in freely moving animals [12]. These data exclude the possibility that stress is involved in electrical stimulation induced catecholamine release.
Our results show that single train and repetitive stimulation of the LC caused a concomitant increase not only of extracellular NA but also of DA in both the mPFC and Occ.
Basal catecholamine levels and stimulation-induced elevations appear to be produced by a nerve-impulse mediated exocytotic process since the local perfusion of TTX dramatically reduced basal levels and totally suppressed the effect of LC stimulation in both cortices. This result exclude the possibility that the increase of extracellular DA originates from a passive leakage of cytosolic DA.
Furthermore, LC stimulation failed to increase DA in the caudate nucleus, which argues against the possibility that midbrain DA neurons might be stimulated by current spread or transynaptically.
NAT displays a high affinity for DA in vitro, and actively reuptakes it from extracellular space [17,18]. The coupling of NA and DA changes in the mPFC has been explained with the hypothesis that DA changes are secondary to the changes in NA competing with DA for NAT [19-21]. However, against this hypothesis is the finding that following phasic stimulation, DA remained elevated much longer than NA both in the mPFC and to a much longer extent in the OCC.
Moreover, electrical stimulation of the LC produced an increase in DA output when NA uptake had been previously blocked by a NAT inhibitor, both in the mPFC and Occ.
On the other hand, our results indicate that the time course of the release of the two catecholamines may be differentially influenced by the frequency and pattern of noradrenergic activity. Thus, while NA returned to the pre-stimulus level within 20 min after either single train or repeated phasic stimulation, DA elevation persisted for a much longer time. Indeed, repeated stimulation produced a persistent elevation of basal DA both in the mPFC and, to a greater extent, in the OCC. A similar elevation of inter-stimulation levels is also observed for NA during DMI perfusion, leading to the speculation that re-uptake mechanism in vivo is more active in clearing NA than DA, especially in the Occ. In fact, when NAT is blocked by DMI, NA curve approaches the DA one that, on the contrary, displays a similar profile in the absence or in the presence of DMI. A scarce efficacy of NAT in the clearance of DA from extracellular space in the Occ has been already proposed by Valentini et al. [22].
Another possible explanation for the finding that the duration of DA increase in both cortices exceeded that of NA, is that LC stimulation activates DA synthesis in noradrenergic neurons in excess to that can be converted to NA, due to increase in tyrosine hydroxylase activity following nerve cell stimulation [23], and consequent possible partial saturation of DA-β-hydroxylase. However, there is no report of this type of saturation, even though evidence exists for a stimulation-induced increase in tyrosine hydroxylase with no change in DA-β-hydroxylase activity in noradrenergic neurons [24-26].
A difficult problem is posed by the finding that DA elevation in the mPFC was of shorter duration than in the parietal cortex. It might be suggested that in the mPFC the DA transporter (DAT), other than NAT, participates in the removal of DA from extracellular space, so that DA removal from mPFC would be more efficient than in the OCC, where DAT is scarce [9,27]. The latter hypothesis is difficult to test, as the selective DAT inhibitor GBR 12909 has been demonstrated to be ineffective in raising extracellular DA levels in the mPFC [11,28] even during dopaminergic activity stimulation by means of haloperidol administration [11]. This observation has been explained with the NAT capacity to totally compensate for DAT exclusion, due to the high affinity of DA for NAT [17,18]. Anyway, this consideration does not exclude that in the mPFC DAT actively reuptakes DA into dopaminergic terminals, thus contributing to decrease extracellular DA levels. Hence, the different profiles of the curves might imply differences in clearance mechanisms for the two catecholamines. MAO A displays higher affinity for NA than for DA [29], and in the Occ DOPAC is about 10% that found in the mPFC [11], thus the possibility exists that in the Occ released DA is cleared prevalently by diffusion, instead of being metabolized to DOPAC.
By means of the double probe preparation, we demonstrated that the NA increase produced by LC electrical stimulation is elicited both in the mPFC and caudate nucleus, while DA increase is present in the mPFC but not in the caudate nucleus. Furthermore, when stimulating electrodes were placed outside the LC, no catecholamine increase was observed. Together, these data indicate that DA increase is not due to dopaminergic neuron activation by current spread. Moreover they exclude the possibility that the DA increase observed in the cerebral cortex is due to direct or indirect stimulation of dopaminergic neurons by NA; on the contrary, the DA decrease observed in the caudate nucleus during prolonged stimulation is consistent with an inhibitory action of NA on DA neurons, as already suggested [30,31]. Such an inhibitory action might contribute to explain why DA elevation in the mPFC was of shorter duration than in the occipital cortex: it might suggest that DA released from noradrenergic terminals was compensated in the mPFC by a reduced DA release from dopaminergic terminals. Indeed, in the caudate nucleus multiple stimulation evidenced a decrease in DA levels.
Finally, the finding that LC stimulation produced the same elevation of DA in areas with high or low dopaminergic innervation, argues against the possibility that a robust portion of extracellular DA, at least in the occipital cortex, might originate from the scarce dopaminergic afferences in these areas.
On the other hand, the co-transmitter hypothesis poses the problem of whether the release of the two catecholamines may be differentially controlled. Indeed, our results indicate that the time course of the release of the two catecholamines may be differentially influenced by the frequency and pattern of noradrenergic activity.
Conclusion
In conclusion, results obtained in the present study provide further evidence that DA in the cerebral cortex satisfies the classic criteria for a candidate co-transmitter in NA neurons. Namely, DA is synthesized, stored, released and recaptured by NA neurons. Notably, the possibility that DA might be released from NA terminals may explain the presence of DA receptors in cortical areas where DA terminals are scarce or absent [3-7].
Our results also suggest that release of DA and NA from NA neurons might be differentially regulated depending on the pattern and frequency of neuronal activity. There is a great deal of evidence for a coupling between the release of DA and NA in the mPFC under different conditions including stress [14-16], antidepressants [32,33], psychostimulants [34,35], antipsychotic drugs [9,36,37], α2-adrenoceptor agonists and antagonists [8,10,11,38].
It would be of great interest to determine whether such coupling occurs in cortices outside the canonical mPFC, whether it represents the co-release of the two catecholamines from noradrenergic terminals, if the release of the two can be differentially controlled, and to understand the functional significance of such coupling.
Methods
The experiments were performed on male Sprague-Dawley rats (Harlan Italy, S. Pietro al Natisone, Italy) weighing 270–300 g, housed in groups of 5 per cage prior to surgery and singly afterwards, at standard conditions of temperature and humidity and artificial light from 8 a.m. to 8 p.m.; food and water were available ad libitum. Experiments were approved by the local ethical committee and performed according to the guidelines for care and use of experimental animals of the UE (EEC Council 86/609; D.L. 27/01/1992, n. 116). Each animal, under Equithesin anesthesia (0.97 g pentobarbital, 2.1 g MgSO4, 4.25 g chloral hydrate, 42.8 ml propylene glycol, 11.5 ml 90% ethanol, distilled water up to 100 ml), was stereotaxically implanted, according to the coordinates of the atlas by Paxinos and Watson [39], with one vertical microdialysis probe in the mPFC (A +3.0, L ± 0.7, V -6.5, 4 mm active membrane length), or one transversal microdialysis probe in the OCC (A -5.0, V-1.8 from bregma, 4 mm active membrane length, corresponding to one hemisphere), or with two probes, a vertical one in the mPFC and a horizontal one in the homolateral OCC. In another group of animals, two vertical probes were inserted, one in the mPFC and one in the homolateral caudate nucleus (A +0.5, L ± 3.0, V -6.5 from bregma, 4 mm active membrane length). Dialysis membrane was AN 69-HF, Hospal-Dasco, Bologna, Italy; the in vitro recovery ranged from 14 to 20% for both DA and NA.
The day after probe implantation, animals were anaesthetized with chloral hydrate (400 mg/kg i.p.) and placed in a Kopf stereotaxic apparatus, where they remained throughout the experiment. Body temperature was maintained at 37°C with a heating pad, and anesthesia was maintained with a continuous infusion of chloral hydrate (400 mg/5 ml, i.p. infused at 0.28 ml/h). A standard bipolar stimulant electrode (Rhodes SNEX-100, Harvard Apparatus, Edenbridge, UK) was implanted in the locus coeruleus (A -3.1, V -8.4, L ± 1.3 from lambda, entering at 15° angle), homolateral to the vertical probes or to the dialyzing portion of the horizontal probe. Stimulant electrodes were connected to a Digitimer (Welwyn Garden City, UK) D4030 stimulus generator and the stimuli were delivered through a DC constant current generator-isolation unit Digitimer DS3. Stimuli were delivered in bursts (trains of 3 spikes regularly spaced in 250 msec, one train per sec) at an intensity of 700 μA, each spike lasting 250 μsec. This stimulation pattern was chosen, on the basis of previous results [12], as it elicited a consistent catecholamine increase with the least behavioural activation. Each rat received 1 stimulation train lasting 20 min or a sequence of 5 stimulation trains, spaced with 20 min pauses.
An artificial cerebrospinal fluid (147 mM NaCl, 4 mM KCl, 1.5 mM CaCl2, pH 6–6.5) was pumped through the dialysis probes at a constant rate of 2.2 μl/min via a CMA/100 microinjection pump (Carnegie Medicine, Stockholm, Sweden). Tetrodotoxin (TTX) and desipramine (DMI), purchased from Sigma-Aldrich (Milan, Italy), were dissolved in the artificial cerebrospinal fluid and administered by inverse dialysis at 10 and 100 μM concentration, respectively.
Samples were collected every 20 min, and NA and DA simultaneously analysed by high performance liquid chromatography (HPLC) apparatus with electrochemical detection, equipped with a 3.0 × 150 mm C18 (5 μ) Symmetry column (Waters) and an ESA Coulochem II detector. The mobile phase consisted of 80 mM Na2HPO4, 0.27 mM EDTA, 0.58 mM sodium octyl sulphate, 12% methanol, pH 2.8 with H3PO4, delivered by a model 307 Gilson pump at 0.4 ml/min; the Coulochem analytical cell first electrode was set at +200 mV, the second at -300 mV. In these conditions, the detection limits (signal to noise ratio 3:1) were 0.3 pg of NA and DA on column.
The average concentration of three stable samples (less than 15% variation) before treatment was considered as the control and was defined as 100%. Thus, values are expressed as percentages of controls + S.E.M.
On termination of experiments, rats were killed with an overdose of chloral hydrate (1.2 g/kg), the brains were removed and stored with 10% formalin in 0.2 M phosphate buffer until histologically verified; coronal sections (40 μm thick) were made with a Vibratome, oriented according to the atlas of Paxinos and Watson [39], stained with fast cresyl violet stain and the location of probes and electrodes was verified according to the above atlas. While no probe resulted outside the correct sampling area, electrodes were found outside the locus coeruleus in 8 out of 59 rats, the data from which were discarded, with the exception of data obtained in 3 of these rats that are shown separately.
Statistical significance was evaluated by analysis of variance (ANOVA) for repeated measures, followed by Dunnett multiple comparison test.
Authors' contributions
PD designed and coordinated the study and participated in drafting the manuscript, GF and PS carried out the surgery and HPLC analysis, MF carried out the electrical stimulation and participated in the design of the study, GLG conceived the study and drafted the manuscript.
Figures and Tables
Figure 1 Schematic representations (redrawn from Paxinos and Watson's Atlas, 1997) of the cerebral areas implanted with dialysis probes.
Figure 2 A: microphotograph showing the electrode localization into the locus coeruleus. B: schematic representation of locus coeruleus stimulation parameters; each vertical line represent a stimulus (700 μA, 250 μs), burst were delivered in 250 ms.
Figure 3 Effect of locus coeruleus stimulation on extracellular NA and DA levels in the medial prefrontal and occipital cortex. Stimuli were delivered in bursts, as described in Methods, for a duration of 20 min (horizontal solid lines). Data are means ± SE, expressed as percent of mean basal value. Filled symbols indicate p < 0.05 with respect to basal values (Dunnett test)
Figure 4 Effect of locus coeruleus stimulation on extracellular NA and DA levels in the medial prefrontal and occipital cortex locally perfused with TTX. TTX perfusion started after collection of basal samples, as indicated by the dashed line. Burst stimuli were delivered 40 min (two samples) after the beginning of TTX perfusion, for a duration 20 min (horizontal solid line). Data are means ± SE, expressed as percent of mean basal value. Filled symbols indicate p < 0.05 with respect to basal values (Dunnett test).
Figure 5 Effect of repeated locus coeruleus stimulation on extracellular NA and DA levels in the medial prefrontal and occipital cortex. Stimuli were delivered in bursts, as described in Methods (horizontal solid lines). Data are means ± SE, expressed as percent of mean basal value. Filled symbols indicate p < 0.05 with respect to basal values (Dunnett test).
Figure 6 Effect of TTX perfusion during repeated locus coeruleus stimulation on extracellular NA and DA levels in the occipital cortex. Stimuli were delivered in bursts, as described in Methods (horizontal solid lines). TTX perfusion started with the third stimulation administration, as indicated by the dashed line. Data are means ± SE, expressed as percent of mean basal value. Filled symbols indicate p < 0.05 with respect to basal values (Dunnett test).
Figure 7 Effect of repeated locus coeruleus stimulation on extracellular NA and DA levels in the medial prefrontal cortex and caudate nucleus. Stimuli were delivered in bursts, as described in Methods (horizontal solid lines). Data are means ± SE, expressed as percent of mean basal value. Filled symbols indicate p < 0.05 with respect to basal values (Dunnett test).
Figure 8 Effect of repeated locus coeruleus stimulation on extracellular NA and DA levels in the medial prefrontal and occipital cortex locally perfused with DMI. 100 μM DMI was administered through reverse dialysis for at least 2 h before stimulation, as represented by the dotted line. Stimuli were delivered in bursts, as described in Methods, for a duration 20 min (horizontal solid lines). Data are means ± SE, expressed as percent of mean basal value. Filled symbols indicate p < 0.05 with respect to basal values (Dunnett test).
Figure 9 Effect of stimulation outside the locus coeruleus on extracellular NA and DA levels in the medial prefrontal cortex and caudate nucleus. Stimuli were delivered in bursts, as described in Methods (horizontal solid lines). Data are means ± SE, expressed as percent of mean basal value.
==== Refs
Descarries L Lemay B Doucet G Berger B Regional and laminar density of the dopamine innervation in adult rat cerebral cortex Neuroscience 1987 21 807 824 3627435 10.1016/0306-4522(87)90038-8
Seguela P Watkins KC Geffard M Descarries L Noradrenaline axon terminals in adult rat neocortex: an immunohistochemical analysis in serial thin sections Neuroscience 1990 35 249 264 2116602 10.1016/0306-4522(90)90079-J
Martres MP Bouthenet ML Sales N Sokoloff P Schwartz JC Widespread distribution of brain dopamine receptors evidenced with [125I]iodosulpiride, a highly selective ligand Science 1985 228 752 755 3838821
Lidow MS Goldman-Rakic PS Rakic P Innis RB Dopamine D2 receptors in the cerebral cortex: distribution and pharmacological characterization with [3H]raclopride Proc Natl Acad Sci USA 1989 86 6412 6416 2548214
Richfield EK Young AB Penney JB Comparative distribution of dopamine D1 and D2 receptors in the cerebral cortex of rats, cats and monkeys J Comp Neurol 1989 286 409 26 2506254 10.1002/cne.902860402
Goldsmith SK Joyce JN Dopamine D2 receptors in hippocampus and parahippocampal cortex of rat, cat and human in relation to tyrosine hydroxylase-immunoreactive fibers Hippocampus 1994 4 354 373 7842057
Wedzony K Chocyk A Mackowiak M Fijal K Czyrak A Cortical localization of dopamine D4 recetros in the rat brain – Immunocytochemical study J Physiol Pharmacol 2000 51 205 221 10898094
Devoto P Flore G Pani L Gessa GL Evidence for co-release of noradrenaline and dopamine from noradrenergic neurons in the cerebral cortex Mol Psychiatry 2001 6 657 664 11673793 10.1038/sj.mp.4000904
Devoto P Flore G Vacca G Pira L Arca A Casu MA Pani L Gessa GL Co-release of noradrenaline and dopamine from noradrenergic neurons in the cerebral cortex induced by clozapine, the prototype atypical antipsychotic Psychopharmacology (Berl) 2003 167 79 84 12632247
Devoto P Flore G Longu G Pira L Gessa GL Origin of extracellular dopamine from dopamine and noradrenaline neurons in the medial prefrontal and occipital cortex Synapse 2003 50 200 205 14515337 10.1002/syn.10264
Devoto P Flore G Longu G Pira L Gessa GL Alpha2-adrenoceptor mediated co-release of dopamine and noradrenaline from noradrenergic neurons in the cerebral cortex J Neurochem 2004 88 1003 1009 14756822
Devoto P Flore G Saba P Fà M Gessa GL Stimulation of the locus coeruleus elicits noradrenaline and dopamine release in the medial prefrontal and parietal cortex J Neurochem 2005 92 368 374 15663484 10.1111/j.1471-4159.2004.02866.x
Kawahara H Kawahara Y Westerink BHC The noradrenaline-dopamine interactions in the rat medial prefrontal cortex studied by multi-probe microdialysis Eur J Pharmacol 2001 418 177 186 11343687 10.1016/S0014-2999(01)00863-9
Kawahara Y Kawahara H Westerink BHC Comparison of the effects of hypotension and handling stress on the release of noradrenaline and dopamine in the locus coeruleus and medial prefrontal cortex of the rat Naunyn Schmiedebergs Arch Pharmacol 1999 360 42 49 10463333 10.1007/s002109900042
Feenstra MG Dopamine and noradrenaline release in the prefrontal cortex in relation to unconditioned and conditioned stress and reward Prog Brain Res 2000 126 133 163 11105645
Murphy EK Sved AF Finlay JM Corticotropin-releasing hormone receptor blockade fails to alter stress-evoked catecholamine release in prefrontal cortex of control or chronically stressed rats Neuroscience 2003 116 1081 1087 12617949 10.1016/S0306-4522(02)00565-1
Horn AS Structure-activity relations for the inhibition of catecholamine uptake into synaptosomes from noradrenergic and dopaminergic neurons in rat brain homogenates Br J Pharmacol 1973 47 332 338 4722047
Raiteri M Del Carmine R Bertollini A Levi G Effects of sympaticomimetic amines on the synaptosomal transport of noradrenaline, dopamine, 5-hydroxytryptamine Eur J Pharmacol 1977 41 133 143 832672 10.1016/0014-2999(77)90202-3
Carboni E Tanda GL Frau R Di Chiara G Blockade of the noradrenaline carrier increases extracellular dopamine concentrations in the prefrontal cortex: evidence that dopamine is taken up in vivo by noradrenergic terminals J Neurochem 1990 55 1067 1070 2117046
Pozzi L Invernizzi R Cervo L Vallebuona F Samanin R Evidence that extracellular concentrations of dopamine are regulated by noradrenergic neurons in the frontal cortex of rats J Neurochem 1994 63 195 200 8207428
Moron JA Brockington A Wise RA Rocha BA Hope BT Dopamine uptake through the norepinephrine transporter in brain regions with low levels of the dopamine transporter: evidence from knock-out mouse lines J Neurosci 2002 22 389 395 11784783
Valentini V Frau R Di Chiara G Noradrenaline transporter blockers raise extracellular dopamine in the medial prefrontal but not parietal and occipital cortex: differences with mianserine and clozapine J Neurochem 2004 88 917 927 14756813
Zigmond RE Schwarzschild MA Rittenhouse A Acute regulation of tyrosine hydroxylase by nerve activity and by neurotransmitters via phosphorilation Ann Rev Neurosci 1989 12 415 461 2564757 10.1146/annurev.ne.12.030189.002215
Anden N-E Grabowska-Anden M Synthesis and utilization of catecholamines in the rat superior cervical ganglion following changes in the nerve impulse flow J Neural Transm 1985 64 81 92 4078568 10.1007/BF01245970
Lewander T Joh TH Reis DJ Tyrosine hydroxylase: delayed activation in central noradrenergic neurons and induction in adrenal medulla elicited by stimulation of central cholinergic receptors J Pharmacol Exp Ther 1977 200 523 534 15097
Kilbourne EJ Sabban EL Differential effect of membrane depolarization on levels of tyrosine hydroxylase and dopamine β-hydroxylase mRNAs in PC12 pheochromocytoma cells Mol Brain Res 1990 8 121 127 1976198 10.1016/0169-328X(90)90056-J
Ciliax BJ Heilman C Demchyshyn LL Pristupa ZB Ince E Hersch SM Niznik HB Levey AI The dopamine transporter: immunochemical characterization and localization in brain J Neurosci 1995 15 1714 1723 7534339
Mazei MS Pluto CP Kirkbride B Pehek EA Effect of catecholamine uptake blockers in the caudate-putamen and subregions of the medial prefrontal cortex of the rat Brain Res 2002 936 58 67 11988230 10.1016/S0006-8993(02)02542-8
Trendelenburg U The uptake and metabolism of 3H-catecholamines in rat cerebral cortex slices Naunyn Schmiedebergs Arch Pharmacol 1989 339 293 297 2725706 10.1007/BF00173580
Grenhoff J Nisell M Ferré S Aston-Jones G Svensson TH Noradrenergic modulation of dopamine cell firing elicited by stimulation of the locus coeruleus in the rat J Neural Transm Gen Sect 1993 93 11 25 8373553 10.1007/BF01244934
Paladini CA Williams JT Noradrenergic inhibition of midbrain dopamine neurons J Neurosci 2004 24 4568 4575 15140928 10.1523/JNEUROSCI.5735-03.2004
Millan MJ Gobert A Rivet J-M Adhumeau-Auclair A Cussac D Newman-Tancredi A Dekeyne A Nicolas J-P Lejeune F Mirtazapine enhances frontocortical dopaminergic and corticolimbic adrenergic, but not serotonergic, transmission by blockade of α2-adrenergic and serotonin2C receptors: a comparison with citalopram Eur J Neurosci 2000 12 1079 1095 10762339 10.1046/j.1460-9568.2000.00982.x
Devoto P Flore G Pira L Longu G Gessa GL Mirtazapine-induced corelease of dopamine and noradrenaline from noradrenergic neurons in the medial prefrontal and occipital cortex Eur J Pharmacol 2004 487 105 111 15033381 10.1016/j.ejphar.2004.01.018
Roberts DC Zis A Fibiger HC Ascending catecholamine pathways and amphetamine-induced locomotor activity: importance of dopamine and apparent non-involvement of norepinephrine Brain Res 1975 93 441 454 1236760 10.1016/0006-8993(75)90182-1
Darracq L Blanc G Glowinski J Tassin J-P Importance of the noradrenaline-dopamine coupling in the locomotor activating effects of D-amphetamine J Neurosci 1998 18 2729 2739 9502830
Westerink BHC de Boer P de Vries JB Kruse CG Long SK Antipsychotic drugs induce similar effects on the release of dopamine and noradrenaline in the medial prefrontal cortex of the rat brain Eur J Pharmacol 1998 361 27 33 9851538 10.1016/S0014-2999(98)00711-0
Li XM Perry KW Wong DT Bymaster FP Olanzapine increases in vivo dopamine and norepinephrine release in rat prefrontal cortex, nucleus accumbens and striatum Psychopharmacology (Berl) 1998 136 153 161 9551772 10.1007/s002130050551
Gresch PJ Sved AF Zigmond MJ Finlay JM Local influence of endogenous norepinephrine on extracellular dopamine in rat medial prefrontal cortex J Neurochem 1995 65 111 116 7790854
Paxinos G Watson C The rat brain in stereotaxic coordinates 1997 3 Academic Press, San Diego, California
| 15865626 | PMC1134661 | CC BY | 2021-01-04 16:03:47 | no | BMC Neurosci. 2005 May 2; 6:31 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-31 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-321586970910.1186/1471-2202-6-32Research ArticleParallel organization of contralateral and ipsilateral prefrontal cortical projections in the rhesus monkey Barbas Helen [email protected] Claus C [email protected] Subhash [email protected] Caterina R [email protected] Joanna L [email protected] Department of Health Sciences, Boston University, Boston, MA, USA2 InternationalUniversity of Bremen, Bremen, Germany3 University of Crete, Iraklion, Crete, Greece2005 3 5 2005 6 32 32 18 1 2005 3 5 2005 Copyright © 2005 Barbas 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 neocortical commissures have a fundamental role in functional integration across the cerebral hemispheres. We investigated whether commissural projections in prefrontal cortices are organized according to the same or different rules as those within the same hemisphere, by quantitatively comparing density, topography, and laminar origin of contralateral and ipsilateral projections, labeled after unilateral injection of retrograde tracers in prefrontal areas.
Results
Commissural projection neurons constituted less than one third of the ipsilateral. Nevertheless, projections from the two hemispheres were strongly correlated in topography and relative density. We investigated to what extent the distribution of contralateral projections depended on: (a) geographic proximity of projection areas to the area homotopic to the injection site; (b) the structural type of the linked areas, based on the number and neuronal density of their layers. Although both measures were good predictors, structural type was a comparatively stronger determinant of the relative distribution and density of projections. Ipsilateral projection neurons were distributed in the superficial (II-III) and deep (V-VI) layers, in proportions that varied across areas. In contrast, contralateral projection neurons were found mostly in the superficial layers, but still showed a gradient in their distribution within cortical layers that correlated significantly with cortical type, but not with geographic proximity to the homotopic area.
Conclusion
The organization of ipsilateral and contralateral prefrontal projections is similar in topography and relative density, differing only by higher overall density and more widespread laminar origin of ipsilateral than contralateral projections. The projections on both sides are highly correlated with the structural architecture of the linked areas, and their remarkable organization is likely established by punctuated development of distinct cortical types. The preponderance of contralateral projections from layer III may be traced to the late development of the callosal system, whose function may be compromised in diseases that have their root late in ontogeny.
==== Body
Background
The primate cerebral cortex constitutes a vast communication network of ipsilateral and contralateral corticocortical connections. Although fewer in number, contralateral projection neurons, which course through the corpus callosum and the anterior commissure, have elaborate dendritic trees [1], and are critical for functional integration of the hemispheres [reviewed in [2-5]].
There is general agreement that commissural projections originate mostly from the homotopic area, and to a lesser extent from neighboring areas [e.g., [6-9]], and involve predominantly neurons in supragranular layers [reviewed in [10-12]]. It has been suggested that geographic distance is a determinant of the existence and relative laminar origin of ipsilateral corticocortical connections [13,14]. In an alternative hypothesis, the pattern of connections depends on the cortical type of the linked areas [15]. Categorical types of cortices can be determined by the number of cortical layers, thickness of layer IV, and density of neurons and other cellular markers [16]. Within this scheme, ipsilateral projections emanate from layers II-III when issued from areas with more layers, or denser layer IV, in comparison with the area of termination. In the reverse direction, projection neurons originate predominantly in layers V-VI. This hypothesis has received support in the ipsilateral connections of prefrontal areas with each other [15], and with distant sensory and association areas [17-20].
Here we tested whether geographic proximity or cortical type best explains the pattern of commissural projections linking prefrontal cortices. The prefrontal cortex is an ideal model system to investigate patterns of commissural projections because it has distinct lateral, orbitofrontal and medial sectors, which vary by distance, structural type, and pattern of ipsilateral interconnections [15]. For example, 'limbic' areas in posterior orbitofrontal and medial prefrontal regions have fewer layers and lower cell density than lateral prefrontal areas, which are eulaminate [16]. Accordingly, when limbic areas issue ipsilateral projections to eulaminate areas they do so overwhelmingly from layers V-VI [e.g. [15]]. In the reverse direction projections originate mostly in layers II-III [15].
Previous studies on patterns of commissural projections relied mostly on qualitative data, and focused on lateral prefrontal areas, or sensory and association cortices, all of which are eulaminate [e.g., [21-23]; reviewed in [10,24]]. Here we exploited the robust and consistent differences in the topography and laminar origin of ipsilateral projections arising from structurally distinct prefrontal areas, and used quantitative data to address the following questions: Are neighboring areas, or areas of similar structural type, more likely to be connected across the commissures than other areas? Are the robust laminar differences interconnecting different ipsilateral prefrontal areas also reflected in the commissural projections of these cortices?
Results
Injection sites
Evidence was provided from 12 cases, and a total of 16 prefrontal sites, which included cases with single injections of WGA-HRP and cases with injections of several distinct fluorescent dyes placed in orbital (n = 7), medial (n = 5), and lateral (n = 4) prefrontal areas, as depicted in a composite diagram of the surfaces of the prefrontal cortex in Figure 1. Further information about the injection sites is included in Additional file 1.
Comparison of the global topography of ipsilateral and contralateral projections
The overall number of projection neurons on the contralateral side was considerably lower than on the ipsilateral side (Fig. 2A), accounting for just 28% of ipsilateral labeled neurons (ncontra = 24,392; nipsi = 87,121). Single injections of tracer at each site labeled many cortical neurons in several ipsilateral areas (range 4–14 areas, median 12), as well as neurons in many of the same areas on the contralateral side (3–11 areas, median 9; Fig. 2B). In addition to this wide distribution of contralateral labeling, there were strong and significant correlations between the relative density patterns of ipsilateral and contralateral projections for all injection sites (Pearson's correlation, p < 0.05), except for injections in area 13 [13(1), 13 (2); cases AJb, ALb] that resulted in slightly higher probability values (p = 0.071, p = 0.051, respectively; Fig. 2C). This demonstrated that the areal distribution and density pattern of contralateral projections overlapped with a substantial subset of projections in the same areas on the ipsilateral hemisphere, despite the much lower absolute density of contralateral projection neurons. Detailed data showing the normalized density of projection neurons in different prefrontal areas, as well as the total number of projection neurons for each injection site are included in Table 1 (Additional file 2).
The general similarity of ipsilateral and contralateral projection patterns was confirmed by an NMDS analysis, arranging injection sites by the relative similarity of their afferent connection patterns (Fig. 3). The depicted fit of areal similarities in two dimensions was high, representing 97% of the variance for the ipsilateral patterns (alienation measure: 0.10), and 83% of the variance for contralateral patterns (alienation: 0.19). On the basis of ipsilateral projection origins, prefrontal cortices were separated into two principal clusters, of mainly mediodorsal areas on the left, and mainly basoventral areas on the right-hand side (Fig. 3A). Area 32 is situated in an intermediate position between mediodorsal and basoventral areas, and similarly occupied an intermediate position, particularly for ipsilateral projections, suggesting similarity with both mediodorsal and basoventral cortices. Within the main groups, the afferent projection patterns of areas varied systematically. Specifically, changes in projection patterns appeared to follow a parallel one-dimensional gradient in both groups. The patterns of contralateral projection origins showed a very similar picture, suggesting that the general organization of ipsilateral and contralateral projections follows similar principles (Fig. 3B).
Comparison of the regional and areal topography of ipsilateral and contralateral projections
When cortices were grouped by region, several trends emerged. Medial areas received the most widespread projections on the ipsilateral side (average = 12.4 areas), followed by orbitofrontal (average = 11.7 areas), and lateral prefrontal cortices with comparatively more restricted projections (average = 6.3 areas). The more widespread ipsilateral projections of posterior medial prefrontal and orbitofrontal 'limbic' areas, in comparison with lateral prefrontal areas, are consistent with previous findings [reviewed in [25]]. The differences noted on the ipsilateral side were also reflected in the number of areas with at least some labeled neurons on the contralateral side, with orbitofrontal areas leading (average of 9.3 areas), medial following (average 8.4 areas), and lateral cortices trailing (average 3.7 areas). The correlation in the areal pattern of projections between ipsilateral and contralateral areas did not reveal substantial regional variations. This finding reflects the overall high correlation in the areal pattern of ipsilateral and contralateral projections, described above.
In contrast to the generally high correlation in the topography of bilateral projections, there was considerable variation in the ratio of the absolute density of contralateral to ipsilateral labeled neurons among injection sites (Fig. 4). Medial prefrontal sites as a group had the highest ratio (average 0.42), followed by the orbitofrontal (average 0.20), and lateral sites (average 0.07), indicating regional trends (rank correlation ρ = -0.66, p < 0.01). Figure 5 shows examples of the distribution of ipsilateral and contralateral projection neurons in prefrontal areas in cases with orbitofrontal injections (areas OPro, cases AF, BCb; area 11, case AM), and Figure 6 in cases with medial (area 32, case AE; area M9, case AO), and lateral injections (area D46, case BFb). Details of the relative density of projection neurons in different prefrontal areas for each injection site are presented in Table 1 (Additional file 2).
Pattern in the topography of contralateral projections
We used two approaches to investigate whether there was a systematic pattern in the topography of contralateral projections. First, we investigated if contralateral projections were most prevalent with the homotopic area of the injection site in the opposite hemisphere, as well as with the corresponding neighboring areas. It has been suggested that neighboring areas are connected at a higher frequency than distant areas in the ipsilateral hemisphere [13]. For the present analysis, each area with projection neurons was given a distance rating based on the number of areal borders intervening between it and the homotopic area of the injection site. Thus, for the homotopic area the distance was 0, for an area sharing a border with the homotopic area it was 1, and for more distant areas the numbers were progressively higher.
Most contralateral projection neurons were found in the homotopic areas (border distance = 0) and their immediate neighbors (border distance = 1). Since the number of immediate neighbors was larger than that of the homotopic areas (just one for most cases), the frequency was shifted towards the next-neighbor origins, and projection neurons originated with decreasing frequency from contralateral cortices that were further away (Fig. 7A). A similar pattern was seen when the relative frequency of projections was considered (data not shown). Moreover, in terms of relative projection density, the densest contralateral projections also emanated from the immediate neighbors of the homotopic areas, showing a strong negative correlation between areal border distance and density of contralateral projections (Spearman's ρ = -0.87; p (two-tailed) < 0.0002; Fig. 7B). In all cases, the contralateral homotopic area included a significant number of projection neurons, and in most these included the densest contralateral projections (area OPro, cases BCb, ALy, AF, AG; area 13, case ALb; area 10, case ARb; medial area 9, cases AO, AQy; dorsal area 46, case BFb). In a minority of cases (cases AJb, AM, AE; Fig. 1), however, the densest projections emanated from heterotopic areas (cf. Table 1 in Additional file 2).
Heterotopic projections also originated from comparatively distant areas, including some that were separated by as many as 5 borders from the homotopic area. For example, projection neurons in substantial numbers were directed to orbitofrontal area OPro from the medially and dorsally situated areas 24, 32 and 9 (separated by 4–5 areas from the homotopic area; cases BCb, AF; Fig. 5). Similarly, relatively high numbers of projection neurons from area 24 on the medial surface were directed to orbitofrontal area 13 (separated by 3 borders; case AJb), and the orbitofrontal area OPro projected to medial area 32 (separated by 4 borders; case AE), as did orbital area 12 to medial area 9 (separated by 5 borders; case AO, Fig. 6, center).
What underlies the pattern of relatively dense projections in contralateral areas situated at a considerable distance from the homotopic area of the ipsilateral injection site? The second analysis addressed this question by considering the relationship of the structural type of pairs of connected cortices. This analysis was motivated by our previous findings that cortical type is highly correlated with the laminar pattern of ipsilateral corticocortical connections [15]. Structural type describes areas mainly by the number of identifiable layers and their relative neuronal density [16]. For this analysis the many distinct architectonic areas can be grouped into a few structural types, as shown graphically for all prefrontal areas in Figure 1. Similar to a previous study [15], we placed areas in one of five categories, the same number as the maximum number of borders between areas noted in the analysis of border distance (above). The cortical type categories ranged from 1 (agranular areas with three identifiable layers; Fig. 1, darkest grey), to 5 (eulaminate areas with six layers and the densest layer IV; Fig. 1, lightest grey). The five levels of cortical type were as follows: 1 (agranular areas OPAll and MPAll); 2 (dysgranular areas OPro, 13, 25, 24 and 32); 3 (areas 14, 11, orbital area 12, medial area 9); 4 (lateral areas 12 and 9, area 10, rostral part of area 46); and 5 (caudal part of area 46, area 8). In categories 3–5 the cortical type is eulaminate, but the density of layer IV increases from categories 3 to 5. For each pair of connected cortices we computed a value for Δ (Δ = origin level - termination level; origin was defined as the projection site, and termination as the injection site). The value of Δ for projection neurons in areas of the same type was 0 (e.g., 2-2), and so on. Type differences were treated as ordinal measures, in the same way as border distances.
The analysis showed that the great majority of projection neurons arose from areas that were of the same (|Δ| = 0) or very similar (|Δ| = 1) architectonic type as the injection site, while comparatively few projections possessed absolute Δ values greater than 2 (Fig. 7C). Moreover, the densest projections had associated Δ values of 0, and the density of all contralateral projections strongly related to the type similarity of contralateral origin and ipsilateral target area (Spearman's rank correlation, ρ = -0.96; p (two-tailed) < 0.0005; Fig. 7D). In all cases, relatively robust projections linking areas at a considerable distance (separated by >3 areal borders in the first analysis) could be explained by the similar structural relationship of the linked areas, that is, absolute type difference, |Δ|. Areas of similar type may be functionally similar as well [18], such as dorsal and ventral area 46. However, functional similarity cannot fully explain these findings, since some areas belonging to the same type have distinct functional attributes, such as anterior cingulate areas, and posterior orbitofrontal areas [reviewed in [26]].
Comparison of the laminar origin of ipsilateral and contralateral projections
There were marked differences in the laminar distribution of ipsilateral and contralateral projection neurons, as shown by their average origin in superficial and deep cortical layers (Fig. 8). As already presented in earlier studies [15,17], ipsilateral projection neurons had a broad distribution based on graded laminar patterns of projections on the ipsilateral side. We previously demonstrated that the percentage of projection neurons in superficial and deep layers of ipsilateral connections varied systematically with Δ, that is, the structural type difference between origin and target area [15]. For matched contralateral areas, however, we now found a preponderance of projection neurons in the upper layers (layers III and II, with most found in layer III). Figure 9 shows the contrasted patterns of the relative frequency of projection neurons in all cases and areas in layers II-III, which resembles a Gaussian curve for the ipsilateral side (Fig. 9A), but shows a marked bias for layers II-III on the contralateral side (Fig. 9B). However, even within this restricted range of origins, the patterns were still significantly correlated with the differences in architectonic type between the area of the projection target and the area of origin, i.e., Δ (rank correlation coefficient, Spearman's ρ = 0.70; p < 0.04; Fig. 10A). Thus, the type difference between a given cortical origin and its contralateral target shaped the relative number of projection neurons arising from superficial and deep layers. A similar analysis relating laminar origin of contralateral projection neurons with geographic proximity to the homotopic area showed no correlation ρ = -0.04; p > 0.91; Fig. 10B). The comparison of the laminar distribution of ipsilateral and contralateral projections was consistent across cases, as shown in examples from individual cases (Fig. 11), and described in detail in Additional file 1 (see Laminar distribution of projection neurons: individual cases).
Discussion
Patterns of contralateral projections to prefrontal cortices were highly correlated with the ipsilateral, both in topography and relative density. Although sparser in overall density than the ipsilateral, widespread commissural projections reached all prefrontal areas studied, differing regionally in density: medial prefrontal areas received the densest projections, followed by orbitofrontal, and then lateral prefrontal areas. Our findings suggest that the number of areas issuing commissural projections to prefrontal cortices surpasses the sensory areas [reviewed in [10,24], the primary motor cortex [e.g., [23,27-32]], or even neighboring premotor areas [8,22,33,34].
Global comparison of prefrontal projections suggested that the organization of ipsilateral and contralateral projections follows very similar rules. Thus, mediodorsal and basoventral prefrontal cortices segregated into two groups, both by their ipsilateral and contralateral projection patterns. Moreover, individual prefrontal areas were distinguished by their characteristic afferent projections, which were similar for ipsilateral and contralateral inputs.
Our finding that prefrontal projections group into two main clusters by their projections is consistent with previous studies on ipsilateral prefrontal projections [e.g., [35,36]]. The somewhat different appearance of the scaling and cluster diagrams in the previous and present studies is likely due to methodological differences. For example, the previous studies were based on evaluation of qualitative data from the published literature which were obtained using several different approaches [35,36]. In contrast, our analysis was based on a uniform approach by one laboratory, using quantitative data. Finally, the previous studies were based on the map of Walker [37], whereas the present study used the map of Barbas and Pandya [38], as modified from the map of Walker.
Determinants of topography, density, and laminar origin of commissural projections
Previous studies have indicated that commissural projections of a given area are densest in the homotopic area, and, in addition, arise from several neighboring areas which also issue ipsilateral projections to the same area [discussion in [6-8]]. Our results are generally consistent with these findings, as demonstrated by a geographic neighbor analysis of projections. However, the relationship of the cortical type of linked areas (Δ) was an even better indicator for the topography, density, and laminar origin of commissural projections. It should be noted that geographic distance correlates with similarity in structural type, since areas having similar structure are frequently neighbors (Fig. 1). However, the existence and density of projections found at a considerable distance from the injection site, in either hemisphere, is consistent with the rules of the 'structural model' [15], but not a 'distance model' [14]. This result supports a broader application of the structural model for understanding organizing features of corticocortical connections [15,17,18].
There is evidence that commissural projections originate mostly from layer III in sensory [e.g., [39-43]], inferior parietal, prefrontal, and temporal cortices [e.g., [21,44]; reviewed in [24]]. The deep layers participate in the commissural system to a lesser extent, though they may have a bigger role in some premotor areas [8,45,46], and in rodents [e.g., [47,48]].
Studies based on qualitative comparisons of ipsilateral and contralateral projections showed similar laminar distributions, with minor variations in different areas or species [e.g., [49,56]]. Most previous comparisons, however, were conducted on eulaminate areas. The present study was based on a broad spectrum of prefrontal cortices which show marked differences in the laminar pattern of their ipsilateral projections. Specifically, ipsilateral projections emanate mostly from layers V-VI of caudal medial or orbitofrontal (limbic) areas, when their destination is a eulaminate cortex. In contrast, projections in the reverse direction originate mostly from layers II-III [15]. The contrasted pattern in the laminar origin of projections in limbic prefrontal cortices provided a striking example of differences across the hemispheres, demonstrating a predominant origin in layers V-VI on the ipsilateral side, but mostly from layer III on the contralateral side. On the other hand, differences in the laminar origin of projections in eulaminate prefrontal areas of the two hemispheres were subtle, consistent with previous findings [e.g., [22]].
The pattern of commissural projections emerges in development
The preponderance of contralateral projection neurons in layer III can be traced to the late development of this system. Even though the corpus callosum emerges between embryonic days E60-65 in rhesus monkeys, at a time when only cortical layers V and VI are present [57], the establishment of callosal projections in prefrontal cortices occurs considerably later, between embryonic days E89 and E111 [58], at a time when the superficial layers are undergoing rapid growth [reviewed in [59]]. The development of this projection system late in gestation is also seen in other areas and other mammalian species [e.g. [60-63]]. Moreover, in reeler mice, where layers develop in the reverse order [64], callosal neurons are accordingly malpositioned in the deep layers, but, as in normal mice, they have the same morphology of medium-sized pyramidal neurons which develop late in ontogeny [64,65]. Further, rats irradiated late in gestation, and acallosal mice show reduction of the supragranular layers and in the size of the corpus callosum [62,66].
Notwithstanding the preponderance of commissural projection neurons in layer III, we still could detect a significant gradient of regional variation in their laminar origin, suggesting that cortical structure is a powerful determinant of the pattern of contralateral as well as ipsilateral corticocortical connections [reviewed in [67]]. This evidence raises the question of how differences in structure emerge, and by extension, how systematic patterns of connections arise.
Variation in structure among prefrontal areas is consistent with differences in the timing of their development [16]. There is evidence that cognitive abilities that rely on the orbitofrontal cortex develop earlier than those dependent on lateral prefrontal areas [68]. Importantly, in rhesus monkeys the upper layers of limbic area 24 complete their development before the same layers in area 11, which, in turn, are completed before the upper layers of area 46 [69]. These temporal differences in development explain the lower density of the upper layers in area 24, than in area 11, which, in turn, is lower than in area 46 [16]. Since projections arise as different layers are generated, a punctuated course of development of prefrontal areas could provide a mechanism for the graded laminar distribution of projections that we showed for ipsilateral connections [15] and for contralateral projections here. Moreover, axons from topographically distant prefrontal areas belonging to the same structural type share the same position within the corpus callosum, suggesting that they develop at the same time [70]. Finally, functional studies indicate that neurons participating in the tangential organization of the cortex and which develop at the same time have correlated firing patterns and likely become connected [for review [71]].
Implications for normal function and neuropathology
Processing across the commissures may be critical for prefrontal areas, which rely on selecting relevant information and suppressing irrelevant information to guide behavior. Evidence of direct involvement of prefrontal commissural projections emerged from physiologic studies in behaving non-human primates performing cognitive tasks within the visual domain. Thus, in macaque monkeys with posterior commissurectomy that prevented the inferior temporal cortex of one hemisphere from receiving bottom-up input from the opposite visual field, single neurons in the inferior temporal cortex in the 'blind' hemisphere responded selectively to task-related stimuli [72]. Visual information relevant to the task at hand apparently followed a circuitous route from the contralateral inferior temporal cortex to the contralateral prefrontal cortex, across the corpus callosum to the opposite prefrontal cortex, and then to the inferior temporal cortex on the side of single cell recording [72,73].
Commissural projections likely have a variety of effects at the site of termination, though most commissural axons arise from excitatory pyramidal neurons [reviewed in [10]], and synapse on spines, which are enriched in pyramidal neurons [74]. The extent of contribution of inhibitory mechanisms across the commissures is not clear, though a small number of callosal axons synapse with GABAergic dendrites in rats [75,76], and a small but significant number of callosal axons originate from inhibitory interneurons [[77,78]; reviewed in [79]]. At the functional level, there is evidence of disynaptic inhibition across the corpus callosum [reviewed in [10]], as well as monosynaptic activation of basket neurons in layers III and IV [80]. Moreover, there is evidence that callosal projection neurons are modulated by dopamine [81], a neurotransmitter system affecting behavior in prefrontal cortex [e.g., [82-85]]. Interestingly, congenitally acallosal mice are hyperactive, and, unlike normal mice, have a tendency to stay off the walls, a behavior associated with a decrease in a metabolic indicator in frontal cortex [86].
The predominance of contralateral projections in layer III may be traced to the late development of the callosal system, which also has a protracted course of myelination [57,87]. Consequently, the commissural system may be affected in disorders that have their root late in development. In schizophrenia, for example, there is evidence of ineffective transmission of information across the hemispheres for cognitive tasks [reviewed in [88]]. In addition, the corpus callosum is preferentially affected with age and in Alzheimer's disease [89,90].
Conclusion
The organization of ipsilateral and contralateral prefrontal projection systems is remarkably similar in terms of topography and relative density of projections. The two systems vary in absolute density, with contralateral projections constituting about a third of the ipsilateral. In addition, ipsilateral projections stem from layers II-III and V-VI, whereas contralateral projections have a narrow range of laminar origins, involving mostly layer III. In both systems there is a tight correlation of projection origins with the structural similarity of the linked areas. This correlation emerges even though the laminar origins of projections in the two hemispheres differ considerably. Overall, structural similarity is a better predictor of existence/absence, relative density, and laminar origin of contralateral projections than border distance of the linked areas.
The organization of ipsilateral and contralateral connections is likely established by punctuated developmental events in structurally and functionally diverse prefrontal cortices. The preponderance of commissural projections from layer III is consistent with the late development of this system, and is likely affected preferentially in diseases that have their root late in development. Contralateral projections in the prefrontal cortical system are richer in topographic origin than in sensory or motor systems. These extensive contralateral projections may be associated with the executive functions of the prefrontal cortex in the selection and synthesis of diverse signals in behavior.
Methods
Surgical and histological procedures
Experiments were conducted on rhesus monkeys (Macaca mulatta) according to the NIH guide for the Care and Use of Laboratory Animals (NIH publication 86–23, revised, 1987). The animals were anesthetized with ketamine hydrochloride (10 mg/kg, i.m.) followed either by sodium pentobarbital administered intravenously through a femoral catheter (to effect, cumulative dose approximately 30 mg/kg), or gas anesthetic (isoflurane) after intubation, until a surgical level of anesthesia was achieved. The monkey's head was firmly positioned in a holder that left the cranium unobstructed for surgical approach. A craniotomy was made and the dura retracted to expose the cortex. All injections were made with a microsyringe (Hamilton, 5 μl) mounted on a microdrive. In each animal we injected a single prefrontal site with HRP-WGA (Sigma, St. Louis, MO; 0.05–0.1 μl, 8%), or one or more prefrontal sites with distinct fluorescent tracers [(diamidino yellow, Sigma, 3% 0.5 μl; or 10% 0.25–4 μl fast blue, Sigma; or fluororuby (dextrantetramethylrhodamine, Molecular Probes); or fluoroemerald (dextran fluorescein, Molecular Probes)].
In the HRP experiments, the monkeys were given an overdose of anesthetic (sodium pentobarbital, intravenously, to effect) 40–48 hours after injection and perfused through the heart with saline followed by 2 liters of fixative (1.25% glutaraldehyde, 1% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4), followed by 2 liters of cold (4°C) phosphate buffer (0.1 M, pH 7.4). The brains were then removed from the skull, photographed, and cryoprotected in glycerol phosphate buffer (10% glycerol and 2% DMSO in 0.1 M phosphate buffer at pH 7.4) for 1 day and in 20% glycerol phosphate buffer for another 2 days. The brains were then frozen in -75°C isopentane, transferred to a freezing microtome, and cut in the coronal plane at 40 μm to produce ten series. One series of sections was treated to visualize HRP [91]. The tissue was mounted, dried, and counterstained with neutral red.
In animals injected with fluorescent dyes, the survival period was 10–18 days. The animals were given an overdose of anesthetic (sodium pentobarbital, intravenously, to effect) and perfused with saline followed by 4% formalin. The brains were removed, photographed, cryoprotected in increasing concentrations of sucrose (10–30%), frozen, and cut in the coronal plane at 40 or 50 μm thickness. Architectonic areas and their borders were determined by staining with thionin, acetylcholinesterase (AChE), or myelin.
Data analysis
Outlines of brain sections, the location of the injection site, and the areal distribution of labeled neurons, were transferred from the slides onto paper by means of a digital plotter (Hewlett Packard, 7475A) electronically coupled to the stage of the microscope and to a PC computer. In this system movements of the microscope stage are recorded as analog signals with linear potentiometers (Vernitech, Axsys, San Diego, CA) mounted on the X and Y axes of the stage of the microscope and coupled to a power supply. The analog signals are then converted to digital signals through an analog-to-digital converter (Data Translation, Marlboro, MA) in the computer. Every other section through the cortex in one series was examined and charted.
Reconstruction of injection sites
The cortical regions containing the injection sites were reconstructed serially by using the sulci as landmarks, as described previously [92]. References to architectonic areas of the prefrontal cortex are according to a previous study [38]. Each injection extended through the depth of the cortex to include all layers.
Statistical analysis
For topological analyses we considered, for each injection site: the number of areas with projection neurons; absolute number of labeled projection neurons; and relative projection densities. The latter measures were derived by normalizing the number of projection neurons directed to an injection site by the total number of prefrontal projection neurons in a case for each side. To determine the laminar pattern of labeled neurons, we normalized data so that the density in the upper layers (layers II-III) and the lower layers (V-VI) was expressed as a complementary percentage of the total number of labeled neurons in that area.
In order to increase statistical reliability, we applied exclusion criteria for the different analytical approaches. Projections from a given area had to consist of at least 20 neurons to be included in the laminar analyses. Since contralateral projections were generally more sparse than the ipsilateral (see Results), the laminar analyses were based on a larger number of individual ipsilateral compared to contralateral data. No general exclusion threshold was applied for other analyses. In one case (BFg), contralateral labeling did not produce sufficient data for analysis, so both ipsilateral and contralateral data were excluded from the topological investigations.
For the investigation of principles underlying connectivity, we also determined the 'border distance' between areas, that is the number of areal borders intervening between the area homotopic to the injection site and each area with projection neurons on the contralateral side. There is little information on the exact spatial trajectory of corticocortical or commissural pathways. Calculations here followed earlier approaches [13,93] and were based on the number of areal borders underneath which fibers from the injection site must have traversed to the area of labeled neurons, assuming they took the shortest path in 3D. We also calculated the architectonic type difference, delta, Δ. The latter is based on a grouping of prefrontal cortices into five structural types, defined by the structure of areas, specifically the number and neuronal density of cortical layers [15,16]. For each pair of connected cortices we computed a value for Δ (Δ = origin level - termination level), with origin defined as the projection site, and termination as the injection site. Type differences ranged from 1 (agranular areas that lack layer IV; Fig. 1, darkest grey) to 5 (eulaminate areas with the densest layer IV; Fig. 1, lightest grey), and were treated as ordinal measures, in the same way as border distances.
In instances where projections originated from two subdivisions of an area, border distance or Δ values were computed on the basis of the average of the two subdivisions (e.g., distance from lateral area 9 to dorsal area 46 = 1; distance from medial area 9 to dorsal area 46 = 2; for projections from area 9 (both divisions) to dorsal area 46, the distance was = [1+2]/2 = 1.5). This procedure affected several prefrontal areas, including the dorsal and ventral part of areas 46, 8, or 24; the orbital and lateral parts of areas 12, 14, or 25; and the medial and lateral parts of area 9. The same method was used in one case where the injection impinged on two distinct architectonic areas (areas OPAll, OPro; case AG). For graphic presentation data were pooled over adjacent intervals, so that a distance or Δ of 2, for instance, included data for distance/Δ of 2 proper as well as for 2.5.
Analyses included calculation of ratios of contralateral to ipsilateral properties, as well as correlations and non-metric multidimensional scaling (NMDS). NMDS analyses and non-parametric correlations were performed in SYSTAT V9 and SPSS V11 for Windows (SPSS Inc.), respectively. Rank correlations (using Spearman's ρ) were employed to investigate the relationship of ordinal measures, such as type differences and border distances, with metric measures of relative projection density, or with relative projection origin within the superficial cortical layers.
The NMDS analysis used to compare the global topology of ipsilateral and contralateral projections, arranges objects in two-dimensional space, based on the high-dimensional similarities among the objects [94]. The relative proximity between items in an NMDS diagram thus represents their relative similarity. In the present study, we assessed the similarity of all injection sites based on the relative ipsilateral and contralateral projection patterns that the injections uncovered. Relative density patterns were metrically correlated for all combinations of injection sites, to derive ipsilateral and contralateral matrices of pairwise injection site similarities. The two-dimensional NMDS analyses in SYSTAT used the Guttman loss function in order to represent these similarity matrices.
Authors' contributions
HB designed the study and drafted the manuscript. HB, SS, CRD and JLS participated in mapping pathways. CCH conducted data analyses. All authors read and approved the manuscript.
Supplementary Material
Additional File 1
Injection sites; and distribution of ipsilateral and contralateral projection neurons in prefrontal cortices in individual cases.
Click here for file
Additional File 2
Table 1 Distribution density of labeled neurons in ipsilateral and contralateral prefrontal cortices.
Click here for file
Acknowledgements
We thank Piro Lera for help with graphics, and Ola Alade and Karen Trait for technical assistance. This work was supported by NIH grants from NINDS and NIMH.
Figures and Tables
Figure 1 Composite of injection sites shown on the medial (A), lateral (B) and orbital (C) surfaces of the right cerebral hemisphere. The injection sites are superimposed on an architectonic map of the prefrontal cortex [38]. Different cortical types are depicted in shades of grey as follows: 1 (darkest grey) agranular areas with three distinguishable layers; 2, dysgranular areas with four distinguishable layers, including a poorly developed layer IV; 3–5, eulaminate areas with increasing cellular density and thickness of layer IV from levels 3 to 5. In A-C, small dashed lines demarcate architectonic areas indicated by numbers; large dashed lines depict the cortex buried in sulci. MPAll, OPAll, OPro, OLF indicate architectonic areas. Letters before architectonic areas designated by letters or numbers denote: C, caudal; L, lateral; M, medial; O, orbital; R, rostral. Other letter combinations refer to cases. Abbreviations: A, arcuate sulcus; Cg, cingulate sulcus; LF, lateral fissure; ST, superior temporal sulcus.
Figure 2 Comparison of contralateral to ipsilateral projections. Cases on the abscissa are identified by the area of the injection of neural tracer. Multiple cases with injection of tracers in the same area are identified by the number in parenthesis after each case. (A) Ratios of contralateral to ipsilateral projection densities. Values were calculated as the total number of contralateral neurons retrogradely labeled by an injection, divided by all labeled neurons ipsilaterally. (B) Ratios of contralateral to ipsilateral projection frequencies. Values were determined as the total number of areas with projection neurons on the contralateral side, divided by all areas with projection neurons in the ipsilateral hemisphere. Projections were included independently of their density and matching origins in the two hemispheres. (C) Strength of correlation between contralateral and ipsilateral projection patterns.
Figure 3 Non-metric multidimensional scaling (NMDS), assessing the pairwise similarities of projection patterns among cases. Cases are identified by the area of tracer injection. Multiple cases with injection of tracer in the same area are identified by the number in parenthesis after each case. Since NMDS configurations are invariant to rotation, the coordinate axes provide a scale of relative similarity, but do not prescribe specific dimensions, and are left unlabeled. (A) ipsilateral cases; (B) contralateral cases. Similarity was defined as the correlation of relative retrograde projection patterns resulting from the injections. The two main projection systems apparent in the diagram consist of predominantly mediodorsal areas (to the left), and basoventral areas (to the right) for both ipsilateral (A) and contralateral (B) projections.
Figure 4 Regional average for ratios of contralateral to ipsilateral projection densities. The apparent trend of progressively smaller contralateral projection densities in medial prefrontal, orbital and lateral prefrontal cortices was significant (rank correlation ρ, p < 0.01).
Figure 5 Bilateral distribution of projections neurons directed to orbitofrontal cortices. Coronal sections through the prefrontal cortex showing labeled neurons in ipsilateral (ipsi) and contralateral (contra) hemispheres after infection of tracers in three cases. Rostral to caudal sections are shown from left to right (case AF), and top to bottom (cases BCb and AM). (Top) The injection of HRP was in area OPro (A, B, black area) and labeled neurons are seen in layers II-III (blue dots) and layers V and VI (red dots); (Left) Injection of fast blue was in area OPro (C), blue area). (Right) Injection of HRP was in the rostral part of area 11 (not shown). Dotted lines through the cortex in the coronal sections show the bottom of layer IV. Architectonic areas indicated by letters include: OPAll, orbital periallocortex; OPro, orbital proisocortex. Other architectonic areas are indicated by numbers. Abbreviations: A, arcuate sulcus; Cg, cingulate sulcus; LF, lateral fissure; LO, lateral orbital sulcus; MO, medial orbital sulcus; P, principal sulcus.
Figure 6 Bilateral distribution of projection neurons directed to medial prefrontal and lateral prefrontal cortices. Coronal sections through the prefrontal cortex showing labeled neurons in ipsilateral (ipsi) and contralateral (contra) hemispheres after infection of tracers in three cases. Labeled neurons were found in layers II-III (blue dots) and in layers V-VI (red dots). In all cases rostral sections are shown on the left, and caudal on the right. (Top) Injection of HRP-WGA (A, black area) was in medial area 32. (Center) Injection of HRP-WGA was in medial area 9 (not shown). (Bottom) Injection of fast blue was in dorsal area 46 (not shown). Dotted lines through the cortex in the coronal sections show the bottom of layer IV. Abbreviations as in Figure 5.
Figure 7 Comparison of border distance and structural type (Δ) as determinants of the pattern of contralateral projections. (A) Frequency of contralateral projections depending on border distance, defined as the number of borders between the contralateral area of projection origin and the area homotopic to the injection site. (B) The relationship of the density of contralateral projections to border distance (Spearman's ρ = -0.87; p (two-tailed) < 0.0002). (C) Frequency of contralateral projections as a function of the absolute structural type difference of the linked areas, delta (|Δ|). (D) The relationship of the density of contralateral projections to the type similarity between the contralateral area of origin and the ipsilateral target area, |Δ| (Spearman's ρ = -0.96; p (two-tailed) < 0.0005). Note that in panels A and C data were pooled over adjacent intervals to avoid artificial variance from sparsely filled categories. In these cases, a distance or Δ of '2', for instance, included data for distance or Δ for 2 proper as well as for 2.5.
Figure 8 Average proportion of laminar origins in superficial layers II-III for ipsilateral and contralateral projections. Data from all cases were included, except for instances where a projection resulted in fewer than 20 neurons. The black and white bars, respectively, add up to 1.0.
Figure 9 Comparison of the relative origin of projection neurons in superficial cortical layers II-III in ipsilateral and contralateral areas. Relative laminar origin of: (A) ipsilateral; (B) contralateral projections. The x-axis shows the proportion of projection neurons found in the upper cortical layers II-III, plotted into intervals of 10% (e.g., in the interval '.8,.9', 80%–90% of projection neurons were located in layers II-III, and the remaining 10%–20% were found in the deep layers, V-VI). Projections originating predominantly from layers V-VI are shown on the left, and projections originating in layers II-III are shown on the right. Projection neurons found in approximately equal proportions in superficial and deep layers are shown in the center. Ipsilateral projections (A) showed a broad Gaussian distribution of laminar origin patterns. By contrast, the distribution of contralateral projection origins (B) was more restricted and skewed towards the superficial layers.
Figure 10 Comparison of the relationship of laminar origin of contralateral projection neurons to structural type and border distance. (A) The x-axis represents the type level difference, Δ), calculated as level (projection origin) – level (projection target, that is, injection site). Normalized origins of contralateral projection neurons varied significantly with Δ (Spearman's ρ, = 0.70; p < 0.04). (B) The x-axis represents the border distance from the injection site on the contralateral side (Spearman's ρ, = -0.04, p > 0.91).
Figure 11 Comparison of the distribution of projection neurons in the supragranular layers II-III in matched areas on the contralateral and ipsilateral hemispheres for individual cases. (A-C) Areal distribution of projection neurons in cases with injection of tracer in orbitofrontal areas (area OPro, case ALy; area 13, case ALb; area 11, case AM); (D) In a case with injection of tracer in medial area 9 (case AO); (E) In a case with injection in dorsal area 46 (case BFb). In all cases the prevalence of projection neurons in layers II-III in most areas is higher on the contralateral side (silhouette bars) than on the ipsilateral side (black bars).
==== Refs
Soloway AS Pucak ML Melchitzky DS Lewis DA Dendritic morphology of callosal and ipsilateral projection neurons in monkey prefrontal cortex Neuroscience 2002 109 461 471 11823059 10.1016/S0306-4522(01)00507-3
Trevarthen C Boller F, Grafman J Integrative functions of the cerebral commissures Handbook of neuropsychology 1990 Amsterdam: Elsevier 49 83
Jeeves MA Lepore F, Ptito M, Jasper HH Callosal agenesis: neuronal and developmental adaptations Two hemispheres, one brain: functions of the corpus callosum 1986 New York: Liss 403 422
Sperry R Lepore F, Ptito M, Jasper HH Consciousness, personal identity, and the divided brain Two hemispheres, one brain: functions of the corpus callosum 1986 New York: Liss 3 20
Gazzaniga MS Cerebral specialization and interhemispheric communication: does the corpus callosum enable the human condition? Brain 2000 123 1293 1326 10869045 10.1093/brain/123.7.1293
Pandya DN Seltzer B Lepore F, Ptito M, Jasper HH The topography of commissural fibers Two hemispheres, one brain: functions of the corpus callosum 1986 New York: Liss 47 73
Hedreen JC Yin TCT Homotopic and heterotopic callosal afferents of caudal inferior parietal lobule in Macaca mulatta J Comp Neurol 1981 197 605 621 7229130 10.1002/cne.901970405
Liu J Morel A Wannier T Rouiller EM Origins of callosal projections to the supplementary motor area (SMA): A direct comparison between pre-SMA and SMA-proper in macaque monkeys J Comp Neurol 2002 443 71 85 11793348 10.1002/cne.10087
Hackett TA Stepniewska I Kaas JH Callosal connections of the parabelt auditory cortex in macaque monkeys Eur J Neurosci 1999 11 856 866 10103079 10.1046/j.1460-9568.1999.00492.x
Innocenti GM Jones EG, Peters A The general organization of callosal connections Cerebral Cortex 1986 New York: Plenum 291 353
Payne BR Lepore F, Ptito M, Jasper HH Role of callosal cells in the functional organization of cat striate cortex Two hemispheres, one brain: functions of the corpus callosum 1986 New York: Liss 231 254
Colonnier M Lepore F, Ptito M, Jasper HH Notes on the early history of the corpus callosum with an introduction to the morphological papers published in this festschrift Two hemispheres, one brain: functions of the corpus callosum 1986 New York: Liss 37 45
Young MP Objective analysis of the topological organization of the primate cortical visual system Nature 1992 358 152 155 1614547 10.1038/358152a0
Barone P Batardiere A Knoblauch K Kennedy H Laminar distribution of neurons in extrastriate areas projecting to visual areas V1 and V4 correlates with the hierarchical rank and indicates the operation of a distance rule J Neurosci 2000 20 3263 3281 10777791
Barbas H Rempel-Clower N Cortical structure predicts the pattern of corticocortical connections Cereb Cortex 1997 7 635 646 9373019 10.1093/cercor/7.7.635
Dombrowski SM Hilgetag CC Barbas H Quantitative architecture distinguishes prefrontal cortical systems in the rhesus monkey Cereb Cortex 2001 11 975 988 11549620 10.1093/cercor/11.10.975
Barbas H Pattern in the laminar origin of corticocortical connections J Comp Neurol 1986 252 415 422 3793985 10.1002/cne.902520310
Rempel-Clower NL Barbas H The laminar pattern of connections between prefrontal and anterior temporal cortices in the rhesus monkey is related to cortical structure and function Cereb Cortex 2000 10 851 865 10982746 10.1093/cercor/10.9.851
Barbas H Ghashghaei H Dombrowski SM Rempel-Clower NL Medial prefrontal cortices are unified by common connections with superior temporal cortices and distinguished by input from memory-related areas in the rhesus monkey J Comp Neurol 1999 410 343 367 10404405 10.1002/(SICI)1096-9861(19990802)410:3<343::AID-CNE1>3.0.CO;2-1
Barbas H Medalla M Alade O Suski J Zikopoulos B Lera P Relationship of prefrontal connections to inhibitory systems in superior temporal areas in the rhesus monkey Cereb Cortex 2005
Jacobson S Trojanowski JQ Prefrontal granular cortex of the rhesus monkey II. Interhemispheric cortical afferents Brain Res 1977 132 235 246 407976 10.1016/0006-8993(77)90418-8
McGuire PK Bates JF Goldman-Rakic PS Interhemispheric integration: I. symmetry and convergence of the corticocortical connections of the left and the right principal sulcus (PS) and the left and the right supplementary motor area (SMA) in the rhesus monkey Cereb Cortex 1991 1 390 407 1726605
Pandya DN Vignolo LA Intra- and interhemispheric projections of the precentral, premotor and arcuate areas in the rhesus monkey Brain Res 1971 26 217 233 4993847 10.1016/0006-8993(71)90215-0
Jones EG Lepore F, Ptito M, Jasper HH Corpus callosum- one system, two approaches Two hemispheres, one brain: functions of the corpus callosum 1986 New York: Liss 149 164
Barbas H Anatomic basis of cognitive-emotional interactions in the primate prefrontal cortex Neurosci Biobehav Rev 1995 19 499 510 7566750 10.1016/0149-7634(94)00053-4
Barbas H Ghashghaei H Rempel-Clower N Xiao D Grafman J Anatomic basis of functional specialization in prefrontal cortices in primates Handbook of Neuropsychology 2002 Amsterdam: Elsevier Science B.V 1 27
Karol EA Pandya DN The distribution of the corpus callosum in the rhesus monkey Brain 1971 94 471 486 4106844
Jones EG Wise SP Size, laminar and columnar distribution of efferent cells in the sensory-motor cortex of monkeys J Comp Neurol 1977 175 391 438 410849 10.1002/cne.901750403
Jenny AB Commissural projections of the cortical hand motor area in monkeys J Comp Neurol 1979 188 137 145 115906 10.1002/cne.901880111
Rouiller EM Moret V Tanne J Boussaoud D Evidence for direct connections between the hand region of the supplementary motor area and cervical motoneurons in the macaque monkey Eur J Neurosci 1996 8 1055 1059 8743753
Gould HG Cusick CG Pons TP Kaas JH The relationship of corpus callosum connections to electrical stimulation maps of motor, supplementary motor, and frontal eye fields in owl monkeys J Comp Neurol 1986 247 297 325 3722441 10.1002/cne.902470303
Jones EG Coulter JD Wise SP Commissural columns in the sensory-motor cortex of monkeys J Comp Neurol 1979 188 113 136 115905 10.1002/cne.901880110
Rouiller EM Babalian A Kazennikov O Moret V Yu XH Wiesendanger M Transcallosal connections of the distal forelimb representations of the primary and supplementary motor cortical areas in macaque monkeys Exp Brain Res 1994 102 227 243 7705502 10.1007/BF00227511
Marconi B Genovesio A Giannetti S Molinari M Caminiti R Callosal connections of dorso-lateral premotor cortex Eur J Neurosci 2003 18 775 788 12925004 10.1046/j.1460-9568.2003.02807.x
Kotter R Hilgetag CC Stephan KE Connectional characteristics of areas in Walker's map of primate prefrontal cortex Neurocomputing 2001 38–40 741 746 12194188 10.1016/S0925-2312(01)00397-6
Passingham RE Stephan KE Kotter R The anatomical basis of functional localization in the cortex Nat Rev Neurosci 2002 3 606 616 12154362
Walker AE A cytoarchitectural study of the prefrontal area of the macaque monkey J Comp Neurol 1940 73 59 86 10.1002/cne.900730106
Barbas H Pandya DN Architecture and intrinsic connections of the prefrontal cortex in the rhesus monkey J Comp Neurol 1989 286 353 375 2768563 10.1002/cne.902860306
Kennedy H Dehay C Functional implications of the anatomical organization of the callosal projections of visual areas V1 and V2 in the macaque monkey Behav Brain Res 1988 29 225 236 3166700 10.1016/0166-4328(88)90027-7
Hof PR Ungerleider LG Adams MM Webster MJ Gattass R Blumberg DM Callosally projecting neurons in the macaque monkey V1/V2 border are enriched in nonphosphorylated neurofilament protein Vis Neurosci 1997 14 981 987 9364733
Manzoni T Conti F Fabri M Callosal projections from area SII to SI in monkeys: anatomical organization and comparison with association projections J Comp Neurol 1986 252 245 263 3782508 10.1002/cne.902520208
Caminiti R Sbriccoli A The callosal system of the superior parietal lobule in the monkey J Comp Neurol 1985 237 85 99 3930578 10.1002/cne.902370107
Wise SP Jones EG Developmental studies of thalamocortical and commissural connections in the rat somatic sensory cortex J Comp Neurol 1978 178 187 208 627623
Jouandet ML Gazzaniga MS Cortical field of origin of the anterior commisure of the Rhesus monkey Exp Neurol 1979 66 381 397 114407 10.1016/0014-4886(79)90088-8
Johnson PB Angelucci A Ziparo RM Michiacchi D Bentivoglio M Caminiti R Segregation and overlap of callosal and association neurons in frontal and parietal cortices of primates: A spectral coherency analysis J Neurosci 1989 9 2313 2326 2746330
Nimchinsky EA Hof PR Young WG Morrison JH Neurochemical, morphologic, and laminar characterization of cortical projection neurons in the cingulate motor areas of the macaque monkey J Comp Neurol 1996 374 136 160 8891952 10.1002/(SICI)1096-9861(19961007)374:1<136::AID-CNE10>3.0.CO;2-S
Olavarria J Van Sluyters RC Widespread callosal connections in infragranular visual cortex of the rat Brain Res 1983 279 233 237 6640342 10.1016/0006-8993(83)90182-8
Koralek KA Olavarria J Killackey HP Areal and laminar organization of corticocortical projections in the rat somatosensory cortex J Comp Neurol 1990 299 133 150 2172324 10.1002/cne.902990202
Schwartz ML Goldman-Rakic PS Callosal and intrahemispheric connectivity of the prefrontal association cortex in rhesus monkey: Relation between intraparietal and principal sulcal cortex J Comp Neurol 1984 226 403 420 6747030 10.1002/cne.902260309
Keller G Innocenti GM Callosal connections of suprasylvian visual area in the cat Neuroscience 1981 6 703 712 6165925 10.1016/0306-4522(81)90154-8
Segraves MA Rosenquist AC The distribution of the cells of origin of callosal projections in cat visual cortex J Neurosci 1982 2 1079 1089 6180149
Van Essen DC Newsome WT Bixby JL The pattern of interhemispheric connections and its relationship to extrastriate visual areas in the macaque monkey J Neurosci 1982 2 265 283 7062108
Jouandet ML Garey LJ Lipp H-P Distribution of the cells of origin of the corpus callosum and anterior commissure in the marmoset monkey Anat Embryol 1984 169 45 59 6426345 10.1007/BF00300586
Beck PD Kaas JH Interhemispheric connections in neonatal owl monkeys (Aotus trivirgatus) and galagos (Galago crassicaudatus) Brain Res 1994 651 57 75 7922590 10.1016/0006-8993(94)90680-7
Cusick CG Kaas JH Lepore F, Ptito M, Jasper HH Interhemispheric connections of cortical sensory and motor representations in primates Two hemispheres, one brain: functions of the corpus callosum 1986 New York: Liss 83 102
Wise SP Jones EG The organization and postnatal development of the commissural projection of the rat somatic sensory cortex J Comp Neurol 1976 168 313 344 950383 10.1002/cne.901680302
LaMantia A-S Rakic P Axon overproduction and elimination in the corpus callosum of the developing rhesus monkey J Neurosci 1990 10 2156 2175 2376772
Schwartz ML Goldman-Rakic PS Prenatal specification of callosal connections in rhesus monkey J Comp Neurol 1991 307 144 162 1713225 10.1002/cne.903070113
Darlington RB Dunlop SA Finlay BL Neural development in metatherian and eutherian mammals: variation and constraint J Comp Neurol 1999 411 359 368 10413772 10.1002/(SICI)1096-9861(19990830)411:3<359::AID-CNE1>3.0.CO;2-J
Angevine JB JrSidman RL Autoradioraphic study of cell migration during histogenesis of cerebral cortex in the mouse Nature 1961 192 766 786 13907727
Rakic P Neurons in rhesus monkey visual cortex: systematic relation between time of origin and eventual disposition Science 1974 183 425 426 4203022
Jensen KF Altman J The contribution of late-generated neurons to the callosal projection in rat: a study with prenatal x-irradiation J Comp Neurol 1982 209 113 122 7130449 10.1002/cne.902090202
Killackey HP Chalupa LM Ontogenetic change in the distribution of callosal projection neurons in the postcentral gyrus of the fetal rhesus monkey J Comp Neurol 1986 244 331 348 3958231
Caviness VS JrSidman RL Time of origin of corresponding cell classes in the cerebral cortex of normal and reeler mutant mice: an autoradiographic analysis J Comp Neurol 1973 148 141 152 4700506 10.1002/cne.901480202
Caviness VSJr Yorke CHJr Interhemispheric neocortical connections of the corpus callosum in the reeler mutant mouse: a study based on anterograde and retrograde methods J Comp Neurol 1976 170 449 460 63471 10.1002/cne.901700405
Abreu-Villaca Y Silva W Manhaes A Schmidt S The effect of corpus callosum agenesis on neocortical thickness and neuronal density of BALB/cCF mice Brain Res Bull 2002 58 411 416 12183019 10.1016/S0361-9230(02)00812-2
Barbas H Hilgetag CC Rules relating connections to cortical structure in primate prefrontal cortex Neurocomputing 2002 44–46 301 308 12744262 10.1016/S0925-2312(02)00356-9
Goldman-Rakic PS Isseroff A Schwartz ML Bugbee NM Mussen P The neurobiology of cognitive development Handbook of child psychology: biology and infancy development 1983 New York: Wiley 281 344
Rakic P Neurogenesis in adult primate neocortex: an evaluation of the evidence Nat Rev Neurosci 2002 3 65 71 11823806 10.1038/nrn700
Barbas H Pandya DN Topography of commissural fibers of the prefrontal cortex in the rhesus monkey Exp Brain Res 1984 55 187 191 6745350 10.1007/BF00240516
Goodman CS Shatz CJ Developmental mechanisms that generate precise patterns of neuronal connectivity Cell 1993 72 77 98 8428376 10.1016/S0092-8674(05)80030-3
Tomita H Ohbayashi M Nakahara K Hasegawa I Miyashita Y Top-down signal from prefrontal cortex in executive control of memory retrieval Nature 1999 401 699 703 10537108 10.1038/44372
Hasegawa I Fukushima T Ihara T Miyashita Y Callosal window between prefrontal cortices: cognitive interaction to retrieve long-term memory Science 1998 281 814 818 9694651 10.1126/science.281.5378.814
Peters A Palay SL Webster HD The fine structure of the nervous system Neurons and their supporting cells 1991 3 New York: Oxford University Press
Carr DB Sesack SR Callosal terminals in the rat prefrontal cortex: synaptic targets and association with GABA-immunoreactive structures Synapse 1998 29 193 205 9635889 10.1002/(SICI)1098-2396(199807)29:3<193::AID-SYN1>3.0.CO;2-1
Vogt BA Gorman AL Responses of cortical neurons to stimulation of corpus callosum in vitro J Neurophysiol 1982 48 1257 1273 6296326
Buhl EH Singer W The callosal projection in cat visual cortex as revealed by a combination of retrograde tracing and intracellular injection Exp Brain Res 1989 75 470 476 2744105 10.1007/BF00249898
Peters A Payne BR Josephson K Transcallosal non-pyramidal cell projections from visual cortex in the cat J Comp Neurol 1990 302 124 142 2086610 10.1002/cne.903020110
Conti F Manzoni T The neurotransmitters and postsynaptic actions of callosally projecting neurons Behav Brain Res 1994 64 37 53 7840891 10.1016/0166-4328(94)90117-1
Somogyi P Kisvarday ZF Martin KA Whitteridge D Synaptic connections of morphologically identified and physiologically characterized large basket cells in the striate cortex of cat Neuroscience 1983 10 261 294 6633861 10.1016/0306-4522(83)90133-1
Carr DB Sesack SR Dopamine terminals synapse on callosal projection neurons in the rat prefrontal cortex J Comp Neurol 2000 425 275 283 10954845 10.1002/1096-9861(20000918)425:2<275::AID-CNE9>3.0.CO;2-Z
Arnsten AF Goldman-Rakic PS Noise stress impairs prefrontal cortical cognitive function in monkeys: evidence for a hyperdopaminergic mechanism Arch Gen Psychiatry 1998 55 362 368 9554432 10.1001/archpsyc.55.4.362
Mrzljak L Bergson C Pappy M Huff R Levenson R Goldman-Rakic PS Localization of dopamine D4 receptors in GABAergic neurons of the primate brain Nature 1996 381 245 248 8622768 10.1038/381245a0
Murphy BL Arnsten AFT Goldman-Rakic PS Roth RH Increased dopamine turnover in the prefrontal cortex impairs spatial working memory performance in rats and monkeys Proc Natl Acad Sci U S A 1996 93 1325 1329 8577763 10.1073/pnas.93.3.1325
Sawaguchi T Goldman-Rakic PS D1 dopamine receptors in prefrontal cortex: involvement in working memory Science 1991 251 947 950 1825731
Magara F Ricceri L Wolfer DP Lipp HP The acallosal mouse strain I/LnJ: a putative model of ADHD? Neurosci Biobehav Rev 2000 24 45 50 10654660 10.1016/S0149-7634(99)00051-2
Yakovlev PI Lecours A-R Minowski A The myelogenetic cycles of regional maturation of the brain Regiona Development of the Brain in Early Life 1967 Oxford and Edinburgh: Blackwell Scienctific Publications 3 70
Butler SR Gruzelier J, Flor-Henry P Interhemispheric relation in schizophrenia Hemisphere asymmetries of function in psychopathology 1979 New York: Elsevier/North-Holland Biomedical Press 47 58
Peters A Sethares C Aging and the myelinated fibers in prefrontal cortex and corpus callosum of the monkey J Comp Neurol 2002 442 277 291 11774342 10.1002/cne.10099
Hampel H Teipel SJ Alexander GE Horwitz B Teichberg D Schapiro MB Corpus callosum atrophy is a possible indicator of region- and cell type-specific neuronal degeneration in Alzheimer disease: a magnetic resonance imaging analysis Arch Neurol 1998 55 193 198 9482361 10.1001/archneur.55.2.193
Mesulam MM Hegarty E Barbas H Carson ECG Mufson EJ Additional factors influencing sensitivity in the tetramethyl benzidine method for horseradish peroxidase J Histochem Cytochem 1980 28 1255 1259 6159394
Barbas H Anatomic organization of basoventral and mediodorsal visual recipient prefrontal regions in the rhesus monkey J Comp Neurol 1988 276 313 342 3192766 10.1002/cne.902760302
Klyachko VA Stevens CF Connectivity optimization and the positioning of cortical areas Proc Natl Acad Sci U S A 2003 100 7937 7941 12796510 10.1073/pnas.0932745100
Kruskal JB Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis Psychometrika 1964 29 1 27
| 15869709 | PMC1134662 | CC BY | 2021-01-04 16:03:47 | no | BMC Neurosci. 2005 May 3; 6:32 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-32 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-351589007410.1186/1471-2202-6-35Research ArticleThe NMDA antagonist memantine affects training induced motor cortex plasticity – a study using transcranial magnetic stimulation [ISRCTN65784760] Schwenkreis Peter [email protected] Katja [email protected] Burkhard [email protected] Jean-Pierre [email protected] Martin [email protected] Department of Neurology, Ruhr-University Bochum, BG-Kliniken Bergmannsheil, Buerkle-de-la-Camp-Platz 1, D-44789 Bochum, Germany2005 12 5 2005 6 35 35 19 4 2005 12 5 2005 Copyright © 2005 Schwenkreis 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
Training of a repetitive synchronised movement of two limb muscles leads to short-term plastic changes in the primary motor cortex, which can be assessed by transcranial magnetic stimulation (TMS) mapping. We used this paradigm to study the effect of memantine, a NDMA antagonist, on short-term motor cortex plasticity in 20 healthy human subjects, and we were especially interested in possible differential effects of different treatment regimens. In a randomised double-blinded cross over study design we therefore administered placebo or memantine either as a single dosage or as an ascending dosage over 8 days. Before and after one hour of motor training, which consisted of a repetitive co-contraction of the abductor pollicis brevis (APB) and the deltoid muscle, we assessed the motor output map of the APB muscle by TMS under the different conditions.
Results
We found a significant medial shift of the APB motor output map after training in the placebo condition, indicating training-induced short-term plastic changes in the motor cortex. A single dosage of memantine had no significant effect on this training-induced plasticity, whereas memantine administered in an ascending dosage over 8 days was able to block the cortical effect of the motor training. The memantine serum levels after 8 days were markedly higher than the serum levels after a single dosage of memantine, but there was no individual correlation between the shift of the motor output map and the memantine serum level. Besides, repeated administration of a low memantine dosage also led to an effective blockade of training-induced cortical plasticity in spite of serum levels comparable to those reached after single dose administration, suggesting that the repeated administration was more important for the blocking effect than the memantine serum levels.
Conclusion
We conclude that the NMDA-antagonist memantine is able to block training-induced motor cortex plasticity when administered over 8 days, but not after administration of a single dose. This differential effect might be mainly due to the prolonged action of memantine at the NMDA receptor. These findings must be considered if clinical studies are designed, which aim at evaluating the potency of memantine to prevent "maladaptive" plasticity, e.g. after limb amputation.
==== Body
Background
Determination of motor output maps by transcranial magnetic stimulation (TMS mapping) proved to be a useful tool allowing the study of the cortical representation of various muscles [1-5], and showing a high map stability and reproducibility [6-8]. Serial TMS mappings can be used to assess short-term plastic changes of the motor cortex induced by the repetitive performance of a motor task consisting of a synchronised movement of two limb muscles. This was demonstrated by a shift of the centre of gravity (COG) of the motor output map derived from a small hand muscle towards the representation of the co-contracted shoulder [9] or leg muscle [10]. Hence this model was used to study the effect of different central acting drugs on short-term motor cortex plasticity: The GABAA agonist lorazepam and the N-methyl-D-aspartate (NMDA)-antagonist amantadine were found to block such cortical plastic changes [11], whereas the indirect dopaminergic and adrenergic agonist amphetamine and the serotonin reuptake inhibitor fluoxetine enhanced training induced cortical plasticity [12,13]. Similar effects of GABAA agonists, NMDA antagonists and amphetamine were reported in series of studies using different paradigms to study training-induced motor cortex plasticity by TMS [14-16], and coactivation-induced plasticity in the primary somatosensory cortex [17-19]. These results support the view that short-term plastic changes in the motor cortex are based on functional changes of synaptic activity, requiring removal of local (presumably GABAA mediated) inhibition, as well as long-term potentiation (LTP)-like changes which are mediated through NMDA receptor activation [20-22].
Motor cortex plasticity, which occurs not only after motor training, but also after peripheral or central lesions [23], might not necessarily be adaptive or beneficial: Possible maladaptive consequences such as phantom limb pain after limb amputation are also discussed to be related to such cortical plastic changes [24]. Hence it could be of therapeutical interest to prevent cortical plasticity, e.g. by using drugs such as memantine, which is a non-competitive antagonist of glutamate and other excitatory amino acids at the MK-801-binding site of the NMDA receptor [25-27].
Here, we used the previously introduced muscle co-contraction paradigm (co-contraction of the deltoid and of the abductor pollicis brevis (APB) muscle) combined with TMS mapping of the APB muscle before and after training in order to evaluate the effect of memantine on short-term motor cortex plasticity in healthy subjects, and on motor performance. Previous studies generally used single dose administration of different drugs like NMDA antagonists or GABA agonists in order to pharmacologically modulate cortical plasticity [15,28,29]. Here, we were particularly interested in possible differential effects of different treatment regimens of memantine, and therefore administered placebo or memantine either as a single dosage or as an ascending dosage over 8 days to healthy humans in a randomised double-blinded cross over study design.
Results
Training effect on mapping parameters and map reproducibility
Looking at the placebo sessions of all 15 subjects participating in the placebo-controlled experiments, we found a significant medial shift of the amplitude-weighted centre of gravity (COG) of the motor output map of the APB muscle (y coordinate before training -5.12 ± 0.64 cm mean ± SD, after training -5.02 ± 0.58 cm, paired t-test, p = 0.022) after motor training. This COG shift was accompanied by a significant lowering of the motor threshold (MT before training 37.4 ± 6.0%, after training 36.7 ± 5.5%, p = 0.019). Other mapping parameters like the x coordinate of the COG, the area of the map, the number of hotspots of the map, the mean MEP amplitude and the sum of amplitudes (SOA) of the map did not change significantly during training.
Comparing the TMS mappings before motor training in the individual subjects between memantine and placebo sessions, there was no significant difference in one of the mapping parameters, indicating a high reproducibility of the mapping procedure.
Memantine effect (single dosage experiment; Table 1)
In the eight subjects who participated in this experiment, we found a significant shortening of the latency differences between the onset of the APB and deltoid muscle contraction during the course of the motor training as revealed by ANOVA for repeated measurements (significant influence of the factor "training duration", F(3.14,43.95) = 17.729, p < 0.001; Fig. 1A). Memantine did not affect this training effect, as shown by the non-significant influence of the factor "drug" (F(1,14) = 1.876, p = 0.192), and by the non-significant interaction between both factors (F(3.14,43.95) = 1.646, p = 0.191).
Memantine serum levels, which were assessed in seven subjects, reached a plateau about 2 hours after drug intake and remained almost stable during the following 5 hours (Fig. 2). Among all subjects, the highest memantine serum level that was reached was 56.16 ng/ml of memantine free base (3 hours after drug intake). Mean memantine serum level 5 hours after drug intake (i.e., immediately after motor training, before starting the second mapping procedure) was 36.1 ± 6.1 ng/ml.
ANOVA revealed a significant lowering of MT after training (F(1,14) = 9.000, p = 0.010 for the factor "training"), but without significant influence of the factor "drug" (F(1,14) = 0.072, p = 0.793), and without significant interaction between both factors (F(1,14) = 0.111, p = 0.744). For other mapping parameters, no significant influence of the factors "training" or "drug", and no significant interaction between both factors was detected.
Memantine effect (ascending dosage experiment; Table 1)
Similar to the results of the single dosage session, we found a significant shortening of the latency differences between the onset of the APB and deltoid muscle contraction during training as revealed by ANOVA for repeated measurements (significant influence of the factor "training duration", F(1.79,21.46) = 13.528, p < 0.001; Fig. 1B) in the seven subjects who participated in this experiment. Again, memantine did not affect this training effect, as shown by the non-significant influence of the factor "drug" (F(1,12) = 0.422, p = 0.528), and by the non-significant interaction between both factors (F(1.79,21.46) = 0.476, p = 0.607).
Mean serum level in the memantine session was 83.1 ± 17.4 ng/ml of memantine free base, with a maximum of 105.4 ng/ml in one subject.
Regarding the y coordinate of the COG, there was a medial shift of the y coordinate under placebo (+0.12 ± 0.18 cm), whereas under memantine a slight lateral shift was observed (-0.08 ± 0.17 cm). ANOVA revealed a significant interaction of the factors "training" and "drug" (F(1,12) = 4.789, p = 0.049), whereas the factors "training" (F(1,12) = 0.150, p = 0.705) and "drug" (F(1,12) = 0.000, p = 1.000) alone did not influence the y coordinate, indicating a significant difference between the shifts of the y coordinate under the different experimental conditions (Fig. 3). For other mapping parameters, no significant influence of the factors "training" or "drug", and no significant interaction between both factors was found.
In order to assess the effect of memantine alone on mapping parameters, without motor training, we additionally compared baseline TMS maps in the placebo and in the memantine condition using Student's paired t-test. For none of the mapping parameters, including the y-coordinate of the COG (-5.460 ± 0.692 with placebo, -5.359 ± 0.503 with memantine, p = 0.301), a significant difference could be observed between the baseline maps (Table 1), indicating that memantine alone had no effect on the mapping parameters.
Influence of memantine serum levels
In order to determine whether the differential effect of memantine in the two experiments might be more related to the different serum levels or to the repeated dosage, we administered a low dosage of memantine (10 mg/d) over 8 days in 5 additional healthy subjects. Mean memantine serum level in this control experiment was 32.2 ± 5.4 ng/ml of memantine free base, with a maximum of 39.9 ng/ml in one subject, and therefore similar to the mean memantine level in the single dosage experiment. The training effect on motor performance was comparable to the other two experiments, with a significant shortening of the latency differences between the onset of the APB and deltoid muscle contraction during training as revealed by ANOVA for repeated measurements (significant influence of the factor "training duration", F(5,20) = 8.907, p < 0.001). However, for none of the mapping parameters, a significant pre-post difference could be observed after motor training (Table 1). Similarly to the ascending dosage experiment, the y co-ordinate of the COG showed a tendency rather to a lateral than to a medial shift, indicating an effective prevention of training-induced cortical plastic changes.
Looking separately at the different experiments, no significant individual correlation was found between memantine serum levels and the pre-post differences of one of the mapping parameters. With the data of the memantine conditions of all experiments pooled together, also no significant correlation with the memantine serum level was detected, neither for the pre-post difference of the y coordinate (Fig. 4), nor for one of the other mapping parameters.
Relationship between motor performance and TMS mapping
For none of the pre-post differences of the mapping parameters, a significant correlation to the improvement of the motor performance could be observed.
Discussion
The main finding of this study is that the repetitive performance of a motor task consisting of a synchronised contraction of the APB and deltoid muscle in the placebo condition leads to a directional shift of the COG of the motor output map of the APB muscle towards the more medial representation of the co-contracted muscle, and that this medial shift is blocked by memantine administered over a period of 8 days, but not by a single dose of memantine.
Our results confirm the results of previous studies, which demonstrated a similar directional shift by means of TMS mapping using the same motor training paradigm [11-13]. In addition, a training induced reduction of motor threshold was observed, which did not differ between placebo and memantine sessions. This lack of influence of memantine on motor threshold is in line with previous findings, suggesting that motor threshold mainly reflects the excitability of the neuronal membrane, and therefore can be influenced by drugs that block voltage-gated sodium channels like lamotrigine or phenytoin [30,31], whereas drugs that influence synaptic transmission like memantine have little influence [32].
Changes of motor threshold and area may be simply due to changes of local motor excitability and not necessarily linked to "true" plasticity in the underlying networks [33]. In contrast, the shift of the COG of the motor output map must be considered as a marker for organisational changes in the representation of movements within the primary motor cortex, and therefore as a correlate of "true" motor cortex plasticity [33-35]. In previous studies, it has been extensively demonstrated that drugs affecting cortical excitability (including NMDA-antagonists) are not able to induce "true" cortical plasticity on their own, suggesting that cortical excitability changes might constitute a necessary, but not sufficient factor in the induction of cortical plasticity [12,15]. Thus it is not conceivable that memantine, which is also known to affect motor cortex excitability [32], induces a (lateral) directional shift of the COG on its own, and therefore only masks the training induced medial shift of the COG. Such an effect of memantine alone on the COG could additionally be excluded by comparing the baseline maps obtained prior to the motor learning with placebo and memantine in the ascending dosage experiment, which did not significantly differ from each other. Our results therefore strongly suggest that memantine is able to block training-induced motor cortex plasticity.
Similar use-dependent alterations of movement representations were observed in the primary motor cortex of adult squirrel monkeys by Nudo et al. [36]. They discovered that after training there was an area expansion of dual-response representations, i.e., cortical sectors over which stimulation produced movements over two or more joints. This may explain the shift of the APB motor output map towards the representation of the co-contracted deltoid muscle in our study. This representational shift can be referred to a principle presented by Hebb, who suggested that individual neurons could participate in different cell assemblies and be involved in multiple functions and representations [37]. The synchronisation or pairing of impulses would then lead to an increase of the excitability of specific neuronal populations, and to a strengthening of the efficacy of their synaptic pathways [10]. This strengthening of synaptic efficacy would involve LTP-like mechanisms, and require the activation of NMDA receptors [21,22]. The fact that the NMDA antagonist memantine was able to prevent cortical plasticity supports this view and confirms previous findings [11,15].
A significant effect of memantine on motor cortex plasticity could be observed in the ascending dosage experiment, whereas the single dosage experiment failed to show a significant difference between memantine and placebo. Since memantine serum levels were markedly higher in this group, we cannot completely rule out that the differential effects of the two administration regimens are linked to these differences in the serum levels. However, in an additional control experiment, an effective blockade of training-induced cortical plasticity was also reached by a repeated administration of a low memantine dosage, with memantine serum levels comparable to those measured after administration of a higher single dosage. Besides, correlation analysis failed to detect a linear correlation between memantine serum levels and COG shifts in individual subjects. These additional findings suggest that the repeated administration of memantine over 8 days was more important for the blocking effect than the memantine serum levels. In previous studies, it has been demonstrated that after subchronic NMDA receptor blockade over seven days, regulatory mechanisms occur at a cellular level, including a down-regulation of cortical NMDA receptors due to a reduction of the glutamate binding site [38,39]. Such mechanisms might also have occurred in our study after 8 days of memantine administration in both the ascending dosage experiment and the control experiment using a repeated low dosage, and therefore have contributed to the effective blockade of training-induced cortical plasticity regardless of memantine serum levels.
During one hour of training, a significant improvement was observed in motor performance, which was not affected by memantine administration, neither in the single dosage nor in the ascending dosage experiment. There was also no significant correlation between the performance improvement and the changes of one of the mapping parameters. This lack of correlation also corresponds to the results reported in previous studies [10,12], which did not reveal a correlation between the improvement of motor performance and the shift of the motor output map. It supports the view that the repetition of a synchronised movement is more important for the induction of cortical plasticity than the improvement of the motor performance.
Conclusion
Previous clinical studies, which used the NMDA antagonist memantine to treat patients e.g. with chronic phantom limb pain, failed to prove any therapeutic effect of memantine in these patients [40-42]. This might be due to the fact that pain-related reorganisation in these chronic patients was already fixed and based on structural changes, hence no longer dependent on NMDA receptor activation. However, from a theoretical point of view, preventing cortical plasticity by blocking NMDA receptors might be a promising clinical approach for the prevention of phantom limb pain if the NMDA antagonist is administered before amputation. The present study suggests that memantine is a drug that could be effective for this purpose, and should be evaluated in further clinical studies. It also emphasizes the importance of an adequate dosage regime, since it would be probably ineffective if administered as a single dosage. Repeated administration seems to be more important than the reached serum levels, possibly allowing the use of lower dosages which are better tolerated.
Methods
Subjects and study design
We investigated 20 healthy right-handed subjects (12 men and eight women, aged 19 – 41 years, mean age 27.7 ± 6.8 years), who were all unrelated to the medical field. They all gave their written informed consent, and the protocol was approved by the local ethical committee. All subjects participated either in the single dosage experiment, or in the ascending dosage experiment, or in the control experiment with a repeated low dosage.
Single dosage experiment: Eight subjects (four men and four women aged between 20 and 34 years) had to participate in two experimental sessions using a randomised double-blinded cross-over study design, with both the researcher and the participating subject being blinded to the experimental condition. They were randomly split into two groups. The first group, made up of four subjects, started with a session including the administration of a single dose of memantine (20 mg to 35 mg, adapted to the individual body weight), the other group with placebo administration. This process was reversed after an interval of at least 14 days. Each session started with a TMS mapping of the right abductor pollicis brevis (APB) muscle as described below. The drug was administered immediately after finishing this mapping procedure, and blood samples were drawn every hour afterwards in order to monitor memantine serum levels. Four hours after drug intake, subjects had to perform a repetitive motor task as described below. Immediately after this motor training, a second TMS mapping of the right APB muscle was performed in order to assess training-induced changes of the APB motor output map.
Ascending dosage experiment: Again a randomised double-blinded cross-over study design was used to study seven other subjects (four men and three women) in this experiment, with the researcher and the participating subject being blinded to the experimental condition. Subjects were randomly split into 2 groups, one group receiving placebo in a first and memantine in a second session, and one group receiving memantine in a first and placebo in a second session, each over a period of 8 days. There was a wash out phase of at least 14 days between both sessions in each subject. Memantine was initially given with a dosage of 10 mg daily, and augmented in steps of 10 mg every 3 days up to 30 mg daily. At day 8, memantine serum levels were determined, and a TMS mapping of the APB muscle was performed twice, immediately before and after one hour of repetitive motor training. Mapping procedures and motor training were identical as in experiment 1.
Control experiment with a repeated low dosage: Five subjects (four men and one woman) participated in this control experiment, receiving a daily dosage of 10 mg memantine over a period of 8 days. In this control experiment, there was no placebo condition, and the researcher as well as the subject was aware of the administered drug. At day 8, memantine serum levels were determined, and a TMS mapping of the APB muscle was performed twice, immediately before and after one hour of repetitive motor training. Mapping procedures and motor training were identical as in the other experiments. This experiment was designed to achieve memantine serum levels comparable to the single dosage experiment, in order to determine whether the memantine serum levels or the repeated administration might be more important for the drug's effect on training- induced cortical plasticity.
Memantine serum level
Immediately after each blood withdrawal, the sample was centrifuged, and the serum was frozen at -70°C. After finishing the study, all frozen samples were shipped together to the Department of Pharmacological Research, Merz Pharmaceuticals, Frankfurt am Main, Germany, where memantine serum levels were assayed with a gas-chromatographic system coupled to a mass selective detector (for details, see [43-45]). Results were expressed as ng/ml concentrations of memantine free base, and not as concentrations of memantine hydrochloride.
Transcranial magnetic stimulation (TMS) mapping
TMS was performed with a Magstim 200 HP device (The Magstim Company) and a figure-of-eight shaped coil (outside diameter 8.7 cm, peak magnetic field strength 2.2T, peak electric field strength 660 V/m), which predominantly stimulates neural structures under its centre. During the whole mapping procedure the coil was held tangentially to the head in an anterior-posterior direction, with the grip pointing backwards. Motor evoked potentials (MEP) were recorded with surface electrodes from the right abductor pollicis brevis muscle (APB) and stored on an EMG device (Neuropack 8, Nihon Kohden). The band pass was 20 Hz to 2 kHz, the gain 0.1 to 1 mV/D. The magnetic stimuli were delivered while the subjects were seated comfortably in a chair. During the whole examination, muscle relaxation was monitored with surface electrodes by EMG (gain 0.1 mV/D). Motor threshold (MT) was determined at rest to the nearest 1% of the stimulator output, and was defined as the minimum intensity which produced five motor evoked potentials >50 μV out of ten trials [46]. Threshold was determined over the scalp position were TMS previously elicited the highest amplitude. Starting at this scalp position and using a stimulation intensity of 110 % of the motor threshold, the motor cortex was examined systematically in rostral, dorsal, medial and lateral directions in steps of 1 cm until no further MEP could be elicited. The positions were identified with the help of a tight fitting cap with a coordinate system drawn on it (1 × 1 cm width). Cz was identified as the intersection of the interaural line and the connection between nasion and inion, which made it possible to localize the coordinates relative to Cz. The x-coordinate was used to indicate the distance in anterior-posterior direction relative to Cz, and the y-coordinate to indicate the distance in medio-lateral direction. Coordinates of Cz were defined as 0/0. Eight stimuli were applied to each position of the grid, and the averaged peak-to-peak MEP amplitude was considered for further statistical analysis. Averaged amplitudes smaller than 10 μV were regarded as zero. Afterwards, we calculated the sum of all MEP amplitudes of the motor output map (SOA), and its amplitude-weighted centre of gravity (COG). The x and y-coordinates of the centre of gravity (COG) are derived from the distribution of MEP amplitudes within the motor output area. They were calculated according to the formula [Σ(x*z)/Σz] or [Σ(y*z)/Σz], where x or y is the position along the x or y axis, and z is the amplitude at this position [10].
As additional mapping parameters were assessed:
- the number of active stimulation sites, i.e. the number of scalp sites from which MEPs could be elicited was used as a marker for the area size of the motor output map, each position equaling 1 cm2,
- the number of "hotspots" of the motor output map, i.e., the number of scalp sites on the coordinate system where TMS elicited MEPs with a mean MEP amplitude of >150 μV, and
- the mean MEP amplitude across all stimulation sites, which was calculated by dividing the SOA by the number of active stimulation sites.
Motor training
In each experimental session, the subjects had to perform a repetitive motor task. This motor task consisted of a synchronised contraction of the deltoid and abductor pollicis brevis (APB) muscle. The participants were instructed to make brisk and short movements of both muscles as synchronously as possible. Approximately three co-contractions per minute had to be performed over one hour. After each single co-contraction, the latency difference between the onsets of muscle contractions was determined using EMG-monitoring with surface electrodes on both muscles. These latency differences of voluntary EMG-activity allowed us to evaluate motor performance. The subjects were informed about the results of their performance and encouraged to improve it [11].
Statistical analysis
Student's paired t-test served to assess intraindividual changes in mapping parameters before and after motor training without memantine intake, to evaluate intrasubject reproducibility of the neurophysiologic data by comparing the maps of individual subjects obtained prior to motor learning in different sessions, and to examine the effect of memantine alone by comparing the baseline maps between the memantine and the placebo condition in the ascending dosage experiment. Student's paired t-test was also used in the control experiment to assess changes in mapping parameters before and after motor training after repeated administration of a low dosage of memantine. To evaluate differences of the TMS mapping parameters between placebo and memantine sessions we used an ANOVA for repeated measurements (main factors "training", i.e., pre vs. post, and "drug" i.e., memantine vs. placebo). The Greenhouse-Geisser procedure was used with epsilon-corrected degrees of freedom, where data showed significant deviations from sphericity. To evaluate the effect of repetitive co-contraction on motor performance, the mean latency differences between the onsets of both muscles for the intervals 0–10 min, 10–20 min, 20–30 min, 30–40 min, 40–50 min and 50–60 min were calculated in each subject and for each session. ANOVA for repeated measurements (main factors: "training duration" and "drug") was performed. Pearson's correlation coefficient r was calculated in order to detect a possible relationship between the pre-post differences of the mapping parameters, memantine serum levels and the improvement of motor task performance, defined as the difference between the mean onset latency difference in the 0–10 min and the 50–60 min interval. For this correlation analysis, only the data obtained after memantine intake were used. For all tests, significance was assumed at the 5 % level.
Authors' contributions
PS and KW conducted the TMS mapping experiments, performed the statistical analysis and drafted the manuscript. BP, JPM and MT participated in the design and coordination of the study, and in the discussion of the results. All authors read and approved the final manuscript.
Acknowledgements
We thank Dr. G. Quack and his colleagues from the Department of Pharmacological Research, Merz Pharmaceuticals GmbH, Frankfurt am Main, Germany, for measurement of memantine serum levels. We also thank L. Schaller for correction and improvement of the language.
Figures and Tables
Figure 1 Motor performance. Comparison of motor performance with placebo (filled circles) and with memantine (open circles). Shown are the results of the single dosage (A) and ascending dosage (B) experiment. The mean latency difference between the onset of muscle contractions (abductor pollicis brevis and deltoid muscle) is shown at different motor training durations.
Figure 2 Memantine serum levels (single dosage). Time course of the memantine serum level after administration of a single dosage. The serum levels of individual subjects, as well as the mean serum level at the different measurement points are shown.
Figure 3 TMS mapping (individual subject). Example of the TMS mapping results in one individual subject who participated in the ascending dosage experiment. Different colours represent the MEP amplitudes after stimulation at the different scalp positions. Note the medial shift of the motor output map of the abductor pollicis brevis muscle after training in the placebo condition, which is no longer present in the memantine condition.
Figure 4 Relationship between memantine serum levels and COG shift. Memantine serum levels are plotted against the shift of the y coordinate of the COG after training. Positive values indicate a shift in a medial, negative values a shift in a lateral direction. Linear correlation analysis reveals no significant correlation. Subjects who participated in the single dosage experiment are red-labelled, subjects who participated in the ascending dosage experiment are blue-labelled, and subjects who participated in the control experiment (repeated low dosage) are green-labelled. Memantine serum levels of the single dosage experiment refer to the values obtained 5 hours after drug administration.
Table 1 TMS mapping results. Mean ± standard deviation for the different TMS parameters assessed before and after one hour of synchronized movements of the right thumb and shoulder under different conditions
Single dosage (n= 8)
Placebo Memantine
Before training After training Before training After training
x coordinate COG (cm) 1.58 ± 0.61 1.59 ± 0.67 1.43 ± 0.72 1.58 ± 0.62
y coordinate COG (cm) -4.82 ± 0.44 -4.74 ± 0.41 -4.98 ± 0.56 -4.97 ± 0.65
Motor threshold (%) 34.6 ± 3.0 34.1 ± 2.7 35.0 ± 1.7 34.4 ± 1.9
Area (cm2) 14.6 ± 4.4 13.8 ± 2.8 15.5 ± 5.2 16.1 ± 2.8
Sum of amplitudes (μV) 1584.6 ± 999.6 1412.0 ± 797.8 2022.4 ± 1193.6 1733.0 ± 756.6
Number of "hotspots" 3.4 ± 2.4 2.7 ± 2.6 4.2 ± 2.8 3.4 ± 2.3
Mean amplitude (μV) 116.1 ± 94.9 100.4 ± 49.9 161.9 ± 180.6 111.4 ± 53.9
Ascending dosage over 8 days (n = 7)
Placebo Memantine
Before training After training Before training After training
x coordinate COG (cm) 1.21 ± 0.60 1.38 ± 0.59 1.49 ± 0.64 1.53 ± 0.66
y coordinate COG (cm) -5.46 ± 0.69 -5.34 ± 0.61 -5.36 ± 0.50 -5.44 ± 0.59
Motor threshold (%) 40.6 ± 7.1 39.7 ± 6.5 41.3 ± 7.3 40.9 ± 7.0
Area (cm2) 16.6 ± 3.8 15.7 ± 3.6 15.9 ± 3.6 16.3 ± 3.7
Sum of amplitudes (μV) 1901.6 ± 1515.8 1535.1 ± 939.3 1744.4 ± 1086.0 1726.7 ± 808.6
Number of "hotspots" 4.4 ± 3.5 3.6 ± 2.6 3.9 ± 3.1 3.6 ± 2.8
Mean amplitude (μV) 120.5 ± 95.4 104.3 ± 65.0 107.2 ± 52.7 107.6 ± 45.2
Low dosage (10 mg/d) over 8 days (n = 5)
Memantine
Before training After training
x coordinate COG (cm) 0.80 ± 0.46 0.62 ± 0.71
y coordinate COG (cm) -5.12 ± 0.46 -5.20 ± 0.53
Motor threshold (%) 38.4 ± 8.3 38.8 ± 12.1
Area (cm2) 12.0 ± 4.1 14.6 ± 4.0
Sum of amplitudes (μV) 1852.2 ± 2169.4 1984.6 ± 1686.1
Number of "hotspots" 3.8 ± 3.8 4.0 ± 4.3
Mean amplitude (μV) 131.2 ± 106.1 122.0 ± 92.5
==== Refs
Wassermann EM McShane LM Hallett M Cohen LG Noninvasive mapping of muscle representations in human motor cortex Electroencephalogr Clin Neurophysiol 1992 85 1 8 1371738
Wassermann EM Pascual-Leone A Valls-Sole J Toro C Cohen LG Hallett M Topography of the inhibitory and excitatory responses to transcranial magnetic stimulation in a hand muscle Electroencephalogr Clin Neurophysiol 1993 89 424 433 7507429 10.1016/0168-5597(93)90116-7
Liepert J Tegenthoff M Malin JP Changes of cortical motor area size during immobilization Electroencephalogr Clin Neurophysiol 1995 97 382 386 8536589 10.1016/0924-980X(95)00194-P
Classen J Knorr U Werhahn KJ Schlaug G Kunesch E Cohen LG Seitz RJ Benecke R Multimodal output mapping of human central motor representation on different spatial scales J Physiol 1998 512 163 179 9729626 10.1111/j.1469-7793.1998.163bf.x
Schwenkreis P Witscher K Janssen F Pleger B Dertwinkel R Zenz M Malin JP Tegenthoff M Assessment of reorganization in the sensorimotor cortex after upper limb amputation Clin Neurophysiol 2001 112 627 635 11275535 10.1016/S1388-2457(01)00486-2
Wilson SA Thickbroom GW Mastaglia FL Transcranial magnetic stimulation mapping of the motor cortex in normal subjects. The representation of two intrinsic hand muscles J Neurol Sci 1993 118 134 144 8229061 10.1016/0022-510X(93)90102-5
Mortifee P Stewart H Schulzer M Eisen A Reliability of transcranial magnetic stimulation for mapping the human motor cortex Electroencephalogr Clin Neurophysiol 1994 93 131 137 7512919 10.1016/0168-5597(94)90076-0
Uy J Ridding MC Miles TS Stability of maps of human motor cortex made with transcranial magnetic stimulation Brain Topogr 2002 14 293 297 12137362 10.1023/A:1015752711146
Cohen LG Gerloff C Ikoma K Hallett M Plasticity of motor cortex elicited by training of synchronous movements of hand and shoulder Soc Neurosci Abstr 1995 21 517
Liepert J Terborg C Weiller C Motor plasticity induced by synchronized thumb and foot movements Exp Brain Res 1999 125 435 439 10323289 10.1007/s002210050700
Tegenthoff M Witscher K Schwenkreis P Liepert J Pharmacological modulation of training-induced plastic changes in human motor cortex Electroencephalogr Clin Neurophysiol Suppl 1999 51 188 196 10590951
Tegenthoff M Cornelius B Pleger B Malin JP Schwenkreis P Amphetamine enhances training-induced motor cortex plasticity Acta Neurol Scand 2004 109 330 336 15080859 10.1046/j.1600-0404.2003.00235.x
Pleger B Schwenkreis P Grunberg C Malin JP Tegenthoff M Fluoxetine facilitates use-dependent excitability of human primary motor cortex Clin Neurophysiol 2004 115 2157 2163 15294219 10.1016/j.clinph.2004.04.015
Classen J Liepert J Wise SP Hallett M Cohen LG Rapid plasticity of human cortical movement representation induced by practice J Neurophysiol 1998 79 1117 1123 9463469
Bütefisch CM Davis BC Wise SP Sawaki L Kopylev L Classen J Cohen LG Mechanisms of use-dependent plasticity in the human motor cortex Proc Natl Acad Sci USA 2000 97 3661 3665 10716702 10.1073/pnas.050350297
Bütefisch CM Davis BC Sawaki L Waldvogel D Classen J Kopylev L Cohen LG Modulation of use-dependent plasticity by d-amphetamine Ann Neurol 2002 51 59 68 11782985 10.1002/ana.10056
Pleger B Dinse HR Ragert P Schwenkreis P Malin JP Tegenthoff M Shifts in cortical representations predict human discrimination improvement Proc Natl Acad Sci U S A 2001 98 12255 12260 11593042 10.1073/pnas.191176298
Dinse HR Ragert P Pleger B Schwenkreis P Tegenthoff M Pharmacological modulation of perceptual learning and associated cortical reorganization Science 2003 301 91 94 12843392 10.1126/science.1085423
Dinse HR Ragert P Pleger B Schwenkreis P Tegenthoff M GABAergic mechanisms gate tactile discrimination learning Neuroreport 2003 14 1747 1751 14512850 10.1097/00001756-200309150-00018
Jacobs KM Donoghue JP Reshaping the cortical motor map by unmasking latent intracortical connections Science 1991 251 944 947 2000496
Hess G Donoghue JP Long-term potentiation of horizontal connections provides a mechanism to reorganize cortical motor maps J Neurophysiol 1994 71 2543 2547 7931533
Garraghty PE Muja N NMDA receptors and plasticity in adult primate somatosensory cortex J Comp Neurol 1996 367 319 326 8708013 10.1002/(SICI)1096-9861(19960401)367:2<319::AID-CNE12>3.0.CO;2-L
Liepert J TMS mapping studies in peripheral and central lesions Electroencephalogr Clin Neurophysiol Suppl 1999 51 151 156 10590946
Karl A Birbaumer N Lutzenberger W Cohen LG Flor H Reorganization of motor and somatosensory cortex in upper extremity amputees with phantom limb pain J Neurosci 2001 21 3609 3618 11331390
Kornhuber J Bormann J Hubers M Rusche K Riederer P Effects of the 1-amino-adamantanes at the MK-801-binding site of the NMDA-receptor-gated ion channel: a human postmortem brain study Eur J Pharmacol 1991 206 297 300 1717296 10.1016/0922-4106(91)90113-V
Danysz W Parsons CG Kornhuber J Schmidt WJ Quack G Aminoadamantanes as NMDA receptor antagonists and antiparkinsonian agents--preclinical studies Neurosci Biobehav Rev 1997 21 455 468 9195603 10.1016/S0149-7634(96)00037-1
Parsons CG Danysz W Quack G Memantine is a clinically well tolerated N-methyl-D-aspartate (NMDA) receptor antagonist - a review of preclinical data Neuropharmacology 1999 38 735 767 10465680 10.1016/S0028-3908(99)00019-2
Schugens MM Egerter R Daum I Schepelmann K Klockgether T Loschmann PA The NMDA antagonist memantine impairs classical eyeblink conditioning in humans Neurosci Lett 1997 224 57 60 9132691 10.1016/S0304-3940(97)13452-8
Stefan K Kunesch E Benecke R Cohen LG Classen J Mechanisms of enhancement of human motor cortex excitability induced by interventional paired associative stimulation J Physiol 2002 543 699 708 12205201 10.1113/jphysiol.2002.023317
Ziemann U Lonnecker S Steinhoff BJ Paulus W Effects of antiepileptic drugs on motor cortex excitability in humans: a transcranial magnetic stimulation study Ann Neurol 1996 40 367 378 8797526 10.1002/ana.410400306
Chen R Samii A Canos M Wassermann EM Hallett M Effects of phenytoin on cortical excitability in humans Neurology 1997 49 881 883 9305361
Schwenkreis P Witscher K Janssen F Addo A Dertwinkel R Zenz M Malin JP Tegenthoff M Influence of the N-methyl-D-aspartate antagonist memantine on human motor cortex excitability Neurosci Lett 1999 270 137 140 10462113 10.1016/S0304-3940(99)00492-9
Ridding MC Rothwell JC Stimulus/response curves as a method of measuring motor cortical excitability in man Electroencephalogr Clin Neurophysiol 1997 105 340 344 9362997 10.1016/S0924-980X(97)00041-6
Mano Y Nakamuro T Tamura R Takayanagi T Kawanishi K Tamai S Mayer RF Central motor reorganization after anastomosis of the musculocutaneous and intercostal nerves following cervical root avulsion [see comments] Ann Neurol 1995 38 15 20 7611718 10.1002/ana.410380106
Traversa R Cicinelli P Bassi A Rossini PM Bernardi G Mapping of motor cortical reorganization after stroke. A brain stimulation study with focal magnetic pulses Stroke 1997 28 110 117 8996498
Nudo RJ Milliken GW Jenkins WM Merzenich MM Use-dependent alterations of movement representations in primary motor cortex of adult squirrel monkeys J Neurosci 1996 16 785 807 8551360
Hebb DO The organization of behaviour 1949 New York, Wiley and Sons
Manallack DT Lodge D Beart PM Subchronic administration of MK-801 in the rat decreases cortical binding of [3H]D-AP5, suggesting down-regulation of the cortical N-methyl-D-aspartate receptors Neuroscience 1989 30 87 94 2568602 10.1016/0306-4522(89)90355-2
Beart PM Lodge D Chronic administration of MK-801 and the NMDA receptor: further evidence for reduced sensitivity of the primary acceptor site from studies with the cortical wedge preparation J Pharm Pharmacol 1990 42 354 355 1976785
Nikolajsen L Gottrup H Kristensen AG Jensen TS Memantine (a N-methyl-D-aspartate receptor antagonist) in the treatment of neuropathic pain after amputation or surgery: a randomized, double-blinded, cross-over study Anesth Analg 2000 91 960 966 11004057 10.1097/00000539-200010000-00036
Maier C Dertwinkel R Mansourian N Hosbach I Schwenkreis P Senne I Skipka G Zenz M Tegenthoff M Efficacy of the NMDA-receptor antagonist memantine in patients with chronic phantom limb pain - results of a randomized double-blinded, placebo-controlled trial Pain 2003 103 277 283 12791434 10.1016/S0304-3959(02)00456-6
Wiech K Kiefer RT Topfner S Preissl H Braun C Unertl K Flor H Birbaumer N A placebo-controlled randomized crossover trial of the N-methyl-D-aspartic acid receptor antagonist, memantine, in patients with chronic phantom limb pain Anesth Analg 2004 98 408 13, table of contents 14742379 10.1213/01.ANE.0000096002.53818.BD
Kornhuber J Quack G Cerebrospinal fluid and serum concentrations of the N-methyl-D-aspartate (NMDA) receptor antagonist memantine in man Neurosci Lett 1995 195 137 139 7478269 10.1016/0304-3940(95)11785-U
Danysz W Gossel M Zajaczkowski W Dill D Quack G Are NMDA antagonistic properties relevant for antiparkinsonian-like activity in rats?--case of amantadine and memantine J Neural Transm Park Dis Dement Sect 1994 7 155 166 7710668
Spanagel R Eilbacher B Wilke R Memantine-induced dopamine release in the prefrontal cortex and striatum of the rat--a pharmacokinetic microdialysis study Eur J Pharmacol 1994 262 21 26 7813574 10.1016/0014-2999(94)90023-X
Rothwell JC Hallett M Berardelli A Eisen A Rossini P Paulus W Magnetic stimulation: motor evoked potentials. The International Federation of Clinical Neurophysiology Electroencephalogr Clin Neurophysiol Suppl 1999 52 97 103 10590980
| 15890074 | PMC1134663 | CC BY | 2021-01-04 16:03:48 | no | BMC Neurosci. 2005 May 12; 6:35 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-35 | oa_comm |
==== Front
BMC Pulm MedBMC Pulmonary Medicine1471-2466BioMed Central London 1471-2466-5-71588514710.1186/1471-2466-5-7Research ArticleUsing the ecology model to describe the impact of asthma on patterns of health care Yawn Barbara P [email protected] George E [email protected] Robert L [email protected] Susan M [email protected] David [email protected] Larry A [email protected] Department of Research, Olmsted Medical Center, Rochester, MN. 55904, USA2 Robert Graham Policy Center, Washington, DC 20036, USA3 Center for Primary Care, Agency for Healthcare Research and Quality, Washington, DC, 20850, USA2005 10 5 2005 5 7 7 22 11 2004 10 5 2005 Copyright © 2005 Yawn et al; licensee BioMed Central Ltd.2005Yawn 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
Asthma changes both the volume and patterns of healthcare of affected people. Most studies of asthma health care utilization have been done in selected insured populations or in a single site such as the emergency department. Asthma is an ambulatory sensitive care condition making it important to understand the relationship between care in all sites across the health service spectrum. Asthma is also more common in people with fewer economic resources making it important to include people across all types of insurance and no insurance categories. The ecology of medical care model may provide a useful framework to describe the use of health services in people with asthma compared to those without asthma and identify subgroups with apparent gaps in care.
Methods
This is a case-control study using the 1999 U.S. Medical Expenditure Panel Survey. Cases are school-aged children (6 to 17 years) and young adults (18 to 44 years) with self-reported asthma. Controls are from the same age groups who have no self-reported asthma. Descriptive analyses and risk ratios are placed within the ecology of medical care model and used to describe and compare the healthcare contact of cases and controls across multiple settings.
Results
In 1999, the presence of asthma significantly increased the likelihood of an ambulatory care visit by 20 to 30% and more than doubled the likelihood of making one or more visits to the emergency department (ED). Yet, 18.8% of children and 14.5% of adults with asthma (over a million Americans) had no ambulatory care visits for asthma. About one in 20 to 35 people with asthma (5.2% of children and 3.6% of adults) were seen in the ED or hospital but had no prior or follow-up ambulatory care visits. These Americans were more likely to be uninsured, have no usual source of care and live in metropolitan areas.
Conclusion
The ecology model confirmed that having asthma changes the likelihood and pattern of care for Americans. More importantly, the ecology model identified a subgroup with asthma who sought only emergent or hospital services.
==== Body
Background
Asthma is a common chronic disease affecting 5–13% of U.S. children and 3–5% of U.S. adults. [1-6] Due to its high and increasing prevalence and resulting morbidity, mortality and high cost of care, asthma is considered a priority condition by the Agency for Health Care Research and Quality (AHRQ)[7] and a major focus of Healthy People 2010 in the United States.[8]
Several studies have reported on health care utilization data for people with asthma, often focusing on emergency and hospital based care. [5,9-16] While these data are important, the designation of asthma as an ambulatory care sensitive condition demands that urgent and emergent care must be studied in the context of ambulatory visits. The ecology of medical care model is a graphical model to display the use of all types of health care by a representative group of 1000 people over a specified period of time, usually one year. The model displays the proportion of those 1000 representative people (in this case US adults and children) who have had at least one contact with the individual sites of health care services such as the emergency department, a tertiary care hospital, a clinic office or home health care services. The model does not account for the number of visits for each person but only whether or not they had any visits in that site. Each individual could be counted once in each site.[17,18]
The ecology model forces examination of the total context of health care giving the clinician, administrator or policy maker the necessary foundation for interpreting ED and hospitalization data. To date the ecology model has not been used to study the impact of a chronic disease on the interaction or sites of interaction of people and health care services. [17-19]
Using the ecology model, the 1999 Medical Expenditures Panel Survey (MEPS) provides the data needed to compare the pattern of health care contact in two population groups: people with asthma and people without any "priority" (that is, chronic, life threatening or mental health) medical conditions. The ecology model is used to demonstrate the impact of asthma on health care patterns and attempt to identify people with inappropriate patterns of care or who suffer from health care disparities.
Methods
This is a case-control study comparing the ecology of medical care for adults and school-aged children with asthma to the ecology of care for adults and school-aged children without any known chronic, life threatening or mental health conditions. The cases and controls were selected from participants in the 1999 MEPS.
Data Source
MEPS is sponsored by the Agency for HealthCare Quality and Research [20] and gathers information from U.S. civilian, non-institutionalized people who participated in National Health Interview Survey (NHIS).[18,19] In 1999, 23,565 persons participated in MEPS computer-assisted personal interviews.
In this study, household component records that query respondents about health care encounters in 1999 were linked with data from the MEPS Condition File describing health problems during that year. All children 6–17 and adults 18–44 years of age with self-reported asthma at the beginning of 1999 were selected as the childhood and adult asthma cases. Children less than 6 years of age were not included in this study due to the difficulty in diagnosing asthma in young children.[21] Adults 45 and older were excluded to avoid the co-mingling of people with asthma and chronic obstructive pulmonary disease. For purposes of study comparisons, adult and childhood control groups were selected that included all persons within the selected age categories (children 6–17 and adults 18–44) with no reported "priority conditions", including no asthma.
Study Variables
Comparisons were made of the proportion of cases and controls who in 1999 had at least one contact with the health care settings assessed in MEPS including: The office of a health professional (excluding dentists and optometrists); an outpatient department; an inpatient hospital service; an emergency department (ED) or who filled a medication prescription. Personal characteristics of the cases and controls were summarized including: gender, family income, self-reported overall health, usual source of care and insurance status.
Analytical Strategy
Descriptive analyses were used to develop the ecology boxes for cases and control subjects, analyzing adults and children separately. Analyses focused on individuals and whether they received care in each of the study settings, not the number of times seen in that setting. Relative risk ratios were calculated for visiting each of the sites, comparing patterns of use of children or adults with and without asthma. Comparisons were made within the cases (children and adults separately) stratified by the demographic and health care arrangement variables.
Results
Asthma was reported to be present in 5% of the school-aged children (6% of boys and 4% of girls) and 3% of the adults (2% of men and 5% of women). Children and adults with asthma were more likely than those without asthma to be insured, Black, live in households at or below 300% of poverty, and report a lower overall health status. (Table 1) Adults with asthma were more likely than children with asthma to remain uninsured (12.8% versus 6.5%, P < 0.04).
Table 1 Demographics of children and adults with asthma using 1999 MEPS data
Children ages 6 through 17 years Adults 18 through 45 years
Asthma % No priority one conditions % P value* Asthma % No priority one conditions % P value*
Education
<HS 18.9 17.3
HS NA NA NA 49.3 52.4 N
>HS 32.8 30.3 S
Rural
Non-MSA 18.1 19.3 NS 17.1 16.2 NS
Gender NS
Male 6 94 2 98
Female 4 96 0.09 5 95
Race
Black 7.9 92.3 3.6 96.4 NS
White 4.4 95.6 <0.001 2.6 97.4
Ethnicity
Hispanic 3.6 96.4 1.5 98.6 <0.001
Non-Hisp 5.2 94.8 0.05 3.0 97.0
Income level
Poor/Near 30.6 22.13 NS 20.1 13.5
Low 18.2 15.9 8.1 13.6 0.03
Middle 26.1 31.9 32.7 32.8
High 25.1 30.1 39.2 40.1
USC
Yes 95.7 88.2 <0.01 81.7 67.2 <0.001
Perceived Health Status
Excellent 22.2 55.4 11.1 40.2 <0.01
Good/very 62.3 53.0 70.2 56.2
Poor/Fair 15.5 1.6 <0.01 18.7 3.6
Insurance
Private 65.0 73.4 75.7 76.6
Public 28.5 17.3 0.01 11.5 5.7 0.01
No 6.5 9.3 12.8 17.7
*p-value for comparison of those with self-reported asthma compared to those with no priority conditions
People with asthma (children 95.7% and adults 81.7%) were more likely than those without asthma to have a usual source of care (Table 1). Less than 1% of adults and children with asthma listed an ED as their usual source of care. Of the one-third who reported a physician rather than a facility as their usual source of care, 66% of adults and 37% of children and teens reported that the physician was a family physician. Pediatricians were listed as the usual source of care for 56% of children and internists for 21% of adults with asthma.
As anticipated, the presence of self-reported asthma increased the likelihood of contact with the health care system in 1999 (Table 2). For example, children and teens with asthma were 1.3 times more likely to have at least one office visit during 1999 compared to the same age group without a priority condition. The comparable difference in office visits for adults was 1.6 fold.
Table 2 Differences in the Ecology of Medical care Associated with Self-Reported Asthma.
Children and teens (6–17) Young adults (18 – 45)
Service site Self-reported Asthma % No priority conditions % Ratio Self-reported Asthma % No priority conditions % Ratio
Office 80.8 61.0 1.3 84.4 54.4 1.6
Out Pt 7.8 4.6 1.7 16.1 6.9 2.3
ED 18.5 8.2 2.2 27.2 9.4 2.9
Hospital 4.7 1.3 3.6 13.5 4.7 2.9
Hospital <24 hrs 0.6 0.2 3.2 0.6 0.2 2.5
Any prescription filled 87.3 41.2 2.1 90.5 46.9 1.9
Figures 1 and 2 use the ecology of medical care model to graphically illustrate the medical ecology for school-aged children and young adults and show the number of people out of 1000 who made a contact with each health care setting during 1999. The risk ratios compare those with asthma to those without chronic, life threatening or mental health conditions. Similar to previous ecology studies, [21,22] the ambulatory setting was the most common setting for care. For both children and adults with asthma, the ratio of those with at least one visit to a site was higher for the hospital (3.2 and 2.9 for children and adults respectively) and the ED (2.2 and 2.9 for children and adults, respectively) than for the office setting, suggesting that asthma has a greater impact on use of healthcare settings that provide urgent and intensive care than on ambulatory care. This relatively large difference in proportion of people contacting the ED and hospital was seen despite a significantly higher proportion of people also obtaining at least some ambulatory care. In both children and adults with asthma, filling at least one prescription was even more common than making an office visit suggesting that several people received all prescriptions from a site other than the office. This finding confirms that a group of people appears to have barriers to ambulatory care and obtains necessary drug therapy from other sites or by telephone without any office follow-up.
Figure 1 Ecology of Asthma. School-Aged Children (6 through 17 years)
Figure 2 Ecology of Asthma. Young Adults (18 through 45 years)
African American children with asthma were most likely to visit an ED or be hospitalized. They were also less likely than non-Black and Hispanic children with asthma to have any ambulatory care contact during the year. Combining office and outpatient visits, Hispanic and White (non-Hispanic) children with asthma had similar proportions of contact with ambulatory, ED and hospital care.
Some children (13.6%) and young adults (10.9%) with self reported asthma had no contact with any of the studied health care sites in 1999. Overall these people had the same average income, and racial distribution as those making at least one ambulatory visit but were more likely to be uninsured (p < 0.01) In addition, 5.2% of children and 3.6% of young adults with asthma visited the ED or were hospitalized but had no ambulatory care visits during 1999. This group was more likely to be uninsured, to have no usual source of care and to live in a metropolitan area than those with ED or hospital visits plus ambulatory visits during 1999 (p < 0.05 for each characteristic). The adults with only ED or hospital care visits also had a lower self-reported health status than adults with asthma having ambulatory visits as well as ED or hospital visits.
Discussion
Using the ecology of medical care model to analyze MEPS data affirms that Americans with asthma are more likely to visit all healthcare sites than Americans without a chronic condition. The additional contact for the people with asthma is primarily in the ambulatory and primary care ambulatory setting. Physicians' offices appear to be an appropriate foci of care since ambulatory care visits provide the opportunity to increase the patient's and family's self-management skills, provide asthma related education and assess current asthma control. [15,16,22,23] Twenty percent of children and 16% of adults (up to 400,000 US children and 460,000 US adults) with self-reported asthma failed to make any ambulatory visit during 1999 missing these educational and monitoring opportunities. While some of these children and adults may have mistakenly reported currently active asthma, the data suggests that almost a million Americans may not be receiving adequate asthma care, if adequate care includes at least yearly visits for active asthma.
The required frequency of asthma visits when no acute exacerbations have occurred is unknown. Studies by Tirimanna and colleagues and Boom and colleagues in the Netherlands attempted to identify the prevalence of undiagnosed asthma and the required frequency of visits to minimize health care utilization.[24,25] The findings varied widely but suggested that at least once a year visits appeared beneficial.
The reasons for not making at least one ambulatory visit during a year may be enlightened by other study results from the Netherlands. In a population- based study Grunsven and colleagues found that many asthma patients were not willing to accept asthma treatment. [26] Part of the reluctance to accept therapy was a "steroid" fear [26] also reported by parents of children with asthma in the US. [27] This may explain the lack of ambulatory visits for part of this 20% of children and 16% of adults; they may quit seeking care because they do not accept the treatment recommended. The cost of co-pays to visit when the asthma is not a major problem may also discourage some less urgent asthma visits.
The lack of ambulatory visits is a special concern for a smaller group of respondents, the 5.2% of children and 3.6% of adults (approximately 130,000 U.S. children and 140,000 U.S. adults) with asthma who made an ED visit or had a hospitalization but made no ambulatory visits during the year. This group is less likely to have insurance or a usual source of care and appears to rely on urgent care sites for asthma management. The ecology model clearly highlights this group, whereas assessments based on membership in a health plan or audits of ambulatory clinics will miss this important subgroup. This is a group that may be uncommon in other countries where there all citizens are assigned a health care clinician and office.
Self-reported asthma is associated with more than twice as many people who make ambulatory visits plus visit the ED or are hospitalized compared to the control group. The reliance on ED or hospital care is most pronounced for young adults with asthma. High rates of ED visits for people with asthma have been reported previously. [9-16,21,28-32] However, few studies have attempted to anchor asthma ED visits in the context of overall pattern of care, sites visited, the presence of a usual source of care, and insurance status. This context of care is particularly important to understand since continuity ambulatory asthma management has been shown to decrease asthma-related visits to the ED and hospitalizations. [33-36] The large group with both ambulatory care and ED visits or hospitalizations may reflect people with severe and difficult to control asthma but may also identify a group without adequate ambulatory care. The ecology of medical care model is an appropriate tool to facilitate this broader view of the impact of asthma on a person's or groups' health care contacts.[17,18] The model also highlights the fact that most of the people making ED visits had a usual source of care other than the ED. This demands that we develop systems that better link ED and ambulatory care, encouraging the potential continuity of care from the ED to the office and vice versa. Several US programs are currently being studied through funding by the CDC. Results of those attempts to link the ED and the office may provide interesting and useful data.
The application of the medical ecology model to the MEPS data has limitations as well as strengths. In MEPS, asthma is self-reported but consistent with published rates of asthma prevalence. [1-3,5,41-43] In addition, the population-based data assures that the asthma experiences represent the full spectrum of disease, rather than only the moderate or severe persistent asthma that is the focus of many published studies.
The data on healthcare system encounters is also self-reported. However, the ecology model uses only the presence or absence of a health care contact. Simply remembering an encounter occurred is less subject to recall bias than remembering the timing and number of encounters.[44,45] MEPS data do not allow assessment of care appropriateness but an asthma ED or hospital visit should be followed by an ambulatory care visit. [46-49] The ecology model highlights the 3 to 5% of people with asthma who fail to access ambulatory care but use more intensive and expensive ED and hospital services.
Conclusion
The ecology of medical care for school-aged children and young adults with self-reported asthma reveals a pattern of health care contacts that is distinctly different from those without priority conditions and identifies a group that may be the victim of health care access disparities.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
BY-initial conceptualization and design, oversaw analysis, drafted manuscript and revision, gave final approval. GF-acquired data, performed analyses, critically reviewed manuscript and revisions, gave final approval. SD-same as RP. RP participated in design and critical review of manuscript and gave final approval. DL- same as RP. LAG-same as RP & DL
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Mannino DM Homa DM Akinbami LJ Moorman JE Gwynn C Redd SC Surveillance for asthma -United States, 1980–1999 MMWR CDC Surveill Summ 2002 51 1 13
Yawn BP Wollan P Kurland M Scanlon P A longitudinal study of the prevalence of asthma in a community population of school -age children J Pediatr 2002 140 576 581 12032525 10.1067/mpd.2002.123764
Newacheck PW Halfon N Prevalence, impact, and trends in childhood disability due to asthma Arch Pediatr Adolesc Med 2000 154 287 93 10710030
Woodruff PG Fahy JV Asthma: Prevalence, Pathogenesis, and Prospects for Novel Therapies JAMA 2001 286 395 398 11466111 10.1001/jama.286.4.395
American Lung Association Epidemiology and Statistics Unit, Best Practices and Program Services Trends in Morbidity and Mortality 2001
Lwebuga-Mukasa JS Dunn-Georgiou E The prevalence of asthma in children of elementary school age in western New York J Urban Health 2000 77 745 61 11194314
Agency for Healthcare Research and Quality Medical Expenditure Panel Survey accessed 7/6/04.
Healthy People 2010 Goals (U.S. Department of Health and Human Services Healthy People 2010 (Conference Edition, in two volumes) Washington, DC 2000
Donahue JG Fuhlbrigge AL Finkelstein JA Fagan J Livingston JM Lozano P Platt R Weiss ST Weiss KB Asthma pharmacotherapy and utilization by children in 3 managed care organizations. The Pediatric Asthma Care Patient Outcomes Research Team J Allergy Clin Immunol 2000 106 1108 14 11112894 10.1067/mai.2000.111432
Christakis DA Mell L Koepsell TD Zimmerman FJ Connell FA Association of lower continuity of care with greater risk of emergency department use and hospitalization in children Pediatrics 2001 107 524 9 11230593 10.1542/peds.107.3.524
Department of Health and Human Services Actions Against Asthma US Department of Health and Human Services Washington DC 2000
Persky VW Slezak J Contreras A Becker L Hernandez E Ramakrishnan V Piorkowski J Relationships of race and socioeconomic status with prevalence, severity, and symptoms of asthma in Chicago school children Ann Allergy Asthma Immunol 1998 81 266 71 9759806
Yawn BP Mainous AG 3rdLove MM Hueston W Do rural and urban children receive similar therapy for asthma? Journal of Rural Health 2000 4 1078 1086
Nelson DA Johnson CC Divine GW Strauchman C Joseph CL Ownby DR Ethnic differences in the prevalence of asthma in middle class children Ann Allergy Asthma Immunol 1997 78 21 6 9012615
Rand CS Butz AM Kolodner K Huss K Eggleston P Malveaux F Emergency department visits by urban African American children with asthma J Allergy Clin Immunol 2000 105 83 90 10629457 10.1016/S0091-6749(00)90182-9
Legorreta AP Liu X Zaher CA Jatulis DE Variation in Managing Asthma: Experience at the Medical Group Level in California Am J Man Care 2000 6 445 453
White K Williams TF Greenberg B The ecology of medical care N Engl J Med 1961 265 885 892 14006536
Green LA Fryer GE JrYawn BP Lanier D Dovey SM The New Ecology of Medicine N Engl J Med 2001 344 2021 2025 11430334 10.1056/NEJM200106283442611
Dovey S Weitzman M Fryer G Green L Yawn B Lanier D Phillips R The ecology of medical care for children Pediatrics 2003 111 1024 9 12728083 10.1542/peds.111.5.1024
DeVoe J Fryer GE Phillips RL Green L Comparing receipt of preventive care among patients based on insurance status and/or usual source of care American Journal of Public Health 2003 93 786 791 12721145
National Asthma Education and Prevention Program (NAEPP), Expert Panel Report 2 Guidelines for the Diagnosis and Management of Asthma Bethesda, MD National Institutes of Health National Heart, Lung, and Blood Institute, NIH Publication No 97-4051 1997
Yawn BP Zyzanski SJ Goodwin MA Gotler RS Stange KC The Anatomy of Asthma Care Visits in Community Family Practice J Asthma 2002 39 719 727 12507192 10.1081/JAS-120015795
Institute of Medicine Crossing the Quality Chasm A New Health System for the 21st Century 2001 National Academy Press
Tirimanna PRS van Schayck CP Otter den JJ Weel C van van Herwaarden CLA van den Boom G van Grunsven PM Bosch vanden WJHM Prevalence of asthma and COPD in general practice since 1992: Has it changed since 1977? Brt J Gen Pract 1996 46 27 281
vanden Boom G van Schayck CP Rutten-van MA Iken MPMH Tirimanna PRS den Otter JJ van Grunsven PM Buitendijk MJ van Herwaarden CLA van Weel C Active detection of COPD and asthma in the general population: results and economic consequences of the DIMCA programme Am J Resp Crit Care 1998 158 1730 1738
van Grunsven PM van Schayck CP van Kollenburg HJM van Bosheide K van den Hoogen HJM Molema J van Weel C The role of "fear of corticosteroids" in nonparticipation in early intervention with inhaled corticosteroids in asthma and COPD in general practice Eur Respir J 1998 11 1178 1181 9648974 10.1183/09031936.98.11051178
Yawn BP The impact of childhood asthma on daily life of the family GPIAG 2003 12 82 85
Cabana MD Rand CS Becher OJ Rubin HR Reasons for pediatrician nonadherence to asthma guidelines Arch Pediatr Adolesc Med 2001 155 1057 62 11529809
National Asthma Education and Prevention Program (NAEPP), Expert Panel Report Guidelines for the Diagnosis and Management of Asthma Bethesda, MD National Institutes of Health National Heart, Lung, and Blood Institute, NIH Publication No 91-3042 1991
National Asthma Education and Prevention Program Expert Panel Report Guidelines for the Diagnosis and Management of Asthma Update on Selected Topics Updates to NIH publication 97-4051 2002
Lang DM Sherman MS Polansky M Guidelines and Realities of Asthma Management. The Philadelphia Story Arch Intern Med 1997 157 1193 1200 9183230 10.1001/archinte.157.11.1193
Laumann JM Bjornson DC Treatment of Medicaid Patients with Asthma: Comparison with Treatment Guidelines Using Disease-Based Drug Utilization Review Methodology Annals of Pharmacotherapy 1998 32 1290 94 9876808 10.1345/aph.18147
Suissa S Ernst P Inhaled corticosteroids: impact on asthma morbidity and mortality J Allergy Clin Immunol 2001 107 937 44 11398069 10.1067/mai.2001.115653
Warman KL Silver EJ Stein RE Asthma symptoms, morbidity, and antiinflammatory use in inner-city children Pediatrics 2001 108 277 82 11483788 10.1542/peds.108.2.277
Liu S Wen SW Demissie K Marcoux S Kramer MS Maternal asthma and pregnancy outcomes: A retrospective cohort study Am J Obstet Gynecol 2001 184 90 6 11174486 10.1067/mob.2001.108073
Olesen C Thrane N Nielsen GL Sorensen HT Olsen J EuroMAP Group A population based prescription study of asthma drugs during pregnancy: changing the intensity of asthma therapy and perinatal outcomes Respiration 2001 68 256 61 11416245 10.1159/000050507
Grant EN Lyttle CS Weiss KB The relation of socioeconomic factors and racial/ethnic differences in US asthma mortality Am J Public Health 2000 90 1923 5 11111268
Guyatt GH Juniper EF Griffith LE Feeny DH Ferrie PJ Children and Adult Perceptions of Childhood Asthma Pediatrics 1997 99 165 8 9024440 10.1542/peds.99.2.165
Reznik M Ozuah PO Franco K Cohen R Motlow F Use of complementary therapy by adolescents with asthma Arch Pediatr Adolesc Med 2002 156 1042 4 12361452
Davis K Schoen C Creating Consensus On Coverage Choices Health Affairs Web Exclusive Project HOPE-the People-to-People Health Foundation, Inc April 23 accessed 11/19/04
Yeatts K Shy C Wiley J Music S Statewide adolescent asthma surveillance J Asthma 2000 37 425 34 10983620
de Marco R Locatelli F Sunyer J Burney P Differences in incidence of reported asthma related to age in men and women. A retrospective analysis of the data of the European Respiratory Health Survey Am J Respir Crit Care Med 2000 162 68 74 10903222
Akinbami LJ LaFleur BJ Schoendorf KC Racial and income disparities in childhood asthma in the United States Ambul Pediatr 2002 2 382 7 12241134 10.1367/1539-4409(2002)002<0382:RAIDIC>2.0.CO;2
Yawn BP Suman VJ Jacobsen SJ Maternal Recall of Distant Pregnancy Events: Accuracy and Recall Bias J Clin Epidemiol 1998 51 399 405 9619967 10.1016/S0895-4356(97)00304-1
Chouinard E Walter S Recall bias in case-control studies J Clin Epidemiol 1995 48 245 254 7869070 10.1016/0895-4356(94)00132-A
Kuttner R The American health care system – health insurance coverage N Engl J Med 1999 340 163 168 9887170 10.1056/NEJM199901143400226
Bodenheimer T The American health care system – physicians and the changing medical marketplace N Engl J Med 1999 340 584 588 10021483 10.1056/NEJM199902183400725
Iglehart JK The American health care system – expenditures N Engl J Med 1999 340 70 76 9878649 10.1056/NEJM199901073400122
Bodenheimer T Wagner EH Grumbach K Improving primary care for patients with chronic illness JAMA 2002 288 1775 9 12365965 10.1001/jama.288.14.1775
| 15885147 | PMC1134664 | CC BY | 2021-01-04 16:30:12 | no | BMC Pulm Med. 2005 May 10; 5:7 | utf-8 | BMC Pulm Med | 2,005 | 10.1186/1471-2466-5-7 | oa_comm |
==== Front
BMC Womens HealthBMC Women's Health1472-6874BioMed Central London 1472-6874-5-61589007810.1186/1472-6874-5-6Research ArticleRace and satisfaction in general OB/GYN clinics Rohrer James E [email protected] Jon D [email protected] Susan [email protected] Department of Family and Community Medicine, Texas Tech University Health Sciences Center, Amarillo TX, USA2 Department of Obstetrics and Gynecology, Texas Tech University Health Sciences Center, Amarillo TX, USA3 Department of OB-GYN, Naval Medical Center San Diego, San Diego CA, USA2005 12 5 2005 5 6 6 11 1 2005 12 5 2005 Copyright © 2005 Rohrer et al; licensee BioMed Central Ltd.2005Rohrer 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 purpose of this study was to test the hypothesis that racial differences in satisfaction can be found among OB/GYN patients on a US naval base.
Methods
Cross-sectional surveys assessing satisfaction with services were obtained from 838 patients who were seen in one of the two general OB/GYN clinics (455 in the base hospital clinic and 391 in a satellite clinic). Multiple logistic regression analysis was used to identify subgroups of patients who were not very satisfied with care received.
Results
When the patients seen in the base hospital were analyzed separately, Asian respondents had significantly lower odds of being very satisfied relative to non-Hispanic white respondents (AOR = .33, p < .01).
Conclusion
Asian patients may be less satisfied than others when treated at a larger facility. Patients treated at a satellite clinic tended to be more satisfied than patients seen at the base hospital.
==== Body
Background
Patient satisfaction has become widely regarded as an important performance indicator for health systems [1]. Patients are the best judge of those aspects of care that matter the most to them [2]. Therefore, patient-centered health systems seek to achieve high levels of patient satisfaction. However, fewer studies reported in the OB/GYN literature focus on satisfaction than might be expected. Hospital maternity care has been assessed using patient satisfaction [3,4] and patient satisfaction has been used to compare OB/GYN to other providers or specialties [5,6]. An increasing number of studies have evaluated the benefits of particular procedures using patient satisfaction as an outcome measure [7-13].
This report adds to the limited fund of information about the determinants of patient satisfaction among OB/GYN clinic patients. Studies of other types of medical care have found racial disparities in satisfaction [14-16] and better satisfaction with midlevel providers than with MDs in either pediatrics or adult medicine clinics [17]. Organizational issues appear to be important as well: women seen in women's clinics are more satisfied than those attending mixed-gender clinics [18], patients seeing providers of the same race are more satisfied than those seeing a physician of a different race [19], free-standing clinics may be scored higher than hospital based clinics by patients [1], and clinics achieving a higher level of provider continuity may also attain higher levels of patient satisfaction, provided that patients may change providers when they so desire [20]. Whether any of these relationships hold true for OB/GYN care is not known. Racial disparities in health status, access to care and quality have been demonstrated in many venues; these, therefore, will be the focus of this report.
The purpose of this study was to test the hypothesis that racial differences in satisfaction can be found among patients seen in general OB/GYN clinics operated by the Naval Medical Center in San Diego. By using just one health system, most organizational characteristics were ruled out as possible causes of differences in satisfaction. In addition, cost to the patient was not a factor since all care was free to the user. In effect, the military medical care system in a single city provides a natural laboratory for assessing the effects of race on satisfaction with minimal confounding from other variables such as poverty and variations in the availability of providers.
Methods
Over a two-month period, a convenience sample of 1544 women receiving OB/GYN services at a large military hospital completed a two-page questionnaire. The survey was approved by the Institutional Review Board. No written consents were required.
The questionnaire was formatted on a two-sided standard automated data form, which was distributed randomly to patients receiving OB/GYN care at all department clinics, as well as to antepartum and post-partum inpatients. With receipt of the survey, each patient also received written and verbal explanations and instructions from clinic staff. Patients previously completing a questionnaire at any location were excluded from repeat sampling.
A total of 1544 patients returned the form. Eight hundred forty six were seen in one of the two general OB/GYN clinics. The base hospital clinic served 455 of these and 391 were seen in a satellite clinic. Of the patients seen in general clinics, 838 answered the question about satisfaction with services.
The dependent variable, patient satisfaction, was measured by asking "Please indicate your overall satisfaction with OB/GYN care received at the Naval Medical Center San Diego. (NTC Clinic included). Possible answers were Not satisfied...would seek care elsewhere if possible, Satisfied, or Very satisfied. The first two answers were combined to form a variable that measured very satisfied versus not very satisfied.
Predictor variables included age, marital status, duty status (active versus retired), relation (service member versus family member), race (non-Hispanic white, Hispanic, Asian, black, or other/missing), rank and clinic location (main hospital vs satellite clinic). Age and duty status were strongly related. Since duty status had a stronger independent relationship with satisfaction, age was dropped from the multivariate analysis. The lower ranks are indicated by 'E' for enlisted, with E1 being the lowest. Officers are higher ranking and they are indicated by an 'O'. Warrant officers are between enlisted and officer ranks.
Univariate associations between being very satisfied and the predictor variables were tested using chi-square. Multiple logistic regression analysis was used to test the unconditional relationship of each independent variable with satisfaction.
Results
Over half (56.1 percent) of the users of the general OB/GYN clinics were very satisfied with the services they received (see Table 1). Race, relation and marital status were not significantly related to percent very satisfied. Retired respondents were more likely to be very satisfied than active duty respondents (68.5 percent vs 51.6 percent, p = .0002). Percent very satisfied increased with rank (p = .0002). Women seen at the satellite facility were more likely to report being very satisfied than those seen at the base hospital clinic (62.6 percent vs 49.1 percent, p = .0005).
Table 1 Descriptive statistics comparing satisfied (yes) vs. non-satisfied (no) patients
Yes No P
N (%) N (%)
Race .1787
Non-Hispanic white 212 (57.6) 156 (42.6)
Asian 40 (49.4) 41 (50.6)
Black 42 (48.3) 45 (51.7)
Hispanic 57 (64.0) 32 (36.0)
Other or missing 32 (55.2) 26 (44.8)
Duty status .0002
Active 263 (51.6) 247 (48.4)
Retired 111 (68.5) 51 (31.5)
Marital status .3246
Divorced/separated 22 (68.8) 10 (31.3)
Married 313 (55.3) 253 (44.7)
Single 45 (57.0) 34 (43.0)
Rank .0002
E1–E4 84 (48.0) 91 (52.0)
E5–E6 121 (51.9) 112 (48.1)
E7–E9 59 (58.4) 42 (41.6)
Warrant-O3 30 (54.5) 25 (45.5)
O4 and above 75 (76.5) 23 (23.5)
Missing 14 (66,7) 7 (33.3)
Relation .4727
Self 99 (53.8) 85 (46.2)
Wife or daughter 281 (56.9) 213 (43.1)
Location .0005
Main hospital 162 (49.1) 168 (50.9)
Satellite 221 (62.6) 132 (37.4)
* number of cases
** significance level
Multivariate analysis of all general clinic patients (Table 2) revealed that retired personnel had an increased odds of being very satisfied compared to active duty respondents (adjusted odds ratio = 1.65). Race was not significant. However, the highest ranking respondents (or women sponsored by high ranking persons) had more than double the odds of being very satisfied than persons of the lowest ranks (AOR = 2.71). Being seen at the satellite clinic independently increased the odds of being satisfied to 1.5 in comparison to being seen in the base hospital clinic.
Table 2 Multiple Logistic Regression analysis of Very Satisfied vs Not Very Satisfied (Adjusted Odds Ratios and Confidence Intervals)
All general patients Base Hospital Satellite clinic
N 672 391 347
OR (Conf Interv) p OR (Conf Interv) p OR (Conf Interv) p
Duty status
Active 1.0 1.0 1.0
Retired 1.645 (1.07–2.54) .025 1.73 (0.80–3.76) .166 1.59 (0.93–2.71) .091
Race/ethnicity
Non-Hispanic White 1.0 1.0 1.0
Asian 0.77 (0.47–1.28) .318 0.33 (0.14–0.74) .006 1.56 (0.77–3.30) .205
Black 0.80 (0.49–1.30) .358 0.61 (0.30–1.21) .155 1.05 (0.52–2.14) .884
Hispanic 1.64 (0.99–2.72) .056 1.48 (0.76–2.90) .251 1.79 (0.83–3.88) .139
Other/ missing 0.98 (0.55–1.75) .955 1.09 (0.47–2.52) .840
Rank
E1–E4 1.0 1.0 1.0
E5–E6 1.06 (0.70–1.59) .787 1.20 (0.69–2.07) .524 0.95 (0.50–1.81) .881
E7–E9 1.16 (0.65–2.04) .618 1.48 (0.60–3.65) .395 0.98 (0.45–2.15) .963
Warrant-O3 1.20 (0.64–2.25) .577 1.57 (0.66–3.72) .310 0.84 (0.33–2.17) .723
O4 and above 2.71 (1.49–4.90) .001 2.67 (1.11–6.45) .029 2.76 (1.19–.39) .018
Missing 0.94 (0.29–3.04) .923 1.41 (0.27–7.36) .681 0.56 (.10–3.14) .509
Location
Base hospital 1.0
Satellite 1.49 (1.07–2.06) .017
Stratification of the data produced additional insights. When the patients seen in the base hospital were analyzed separately, rank remained important but duty status was no longer significant. Asian respondents had significantly lower odds of being very satisfied relative to non-Hispanic white respondents (AOR = .33, p = .0077) in this subset of the data. Separate analysis of the patients seen in the satellite clinic produced only one significant predictor: persons in the highest rank group had an adjusted odds of 2.76 relative to the lowest rank (p = .0181)
Discussion
An increasing number of research reports that address patient satisfaction is appearing in the OB/GYN literature. These studies do not employ standard methods for either measurement or analysis and study designs are varied as well.
A critical issue has to do with the measurement of satisfaction. Some studies use mean satisfaction scores [9,10] while others divide subjects into very satisfied versus not very satisfied, as we did [8]. Skewed distributions are the norm in patient satisfaction surveys, due to reluctance on the part of patients to criticize their health care providers, so satisfaction scales usually are dichotomized and analyzed using logistic regression analysis.
Our study differs from many in that it has a large sample size and measures satisfaction using a single item that was dichotomized into satisfied versus not satisfied. It adds to the OB/GYN literature by showing that, as expected, Asian patients are less likely to be very satisfied with care received. This is consistent with findings relating to other patient groups [14-16]. However, we add the proviso that Asian patients are less likely to be satisfied in the base hospital clinic, but not in the satellite clinic. We do not know why this is the case. Approximately the same numbers of Asian patients attended both clinics (43/353 in the satellite and 38/330 in the hospital clinic). Further investigation of this issue is needed so that disparities can be eliminated and so that we can learn of any particular clinic characteristics that Asian patients especially appreciate or dislike.
The second interesting finding in this study is the importance of rank. If the OB/GYN care was being provided by private sector clinics, we might assume that higher ranking officers have higher incomes and thus would receive more attention in a system that is driven by profit. However, the military hospital does not bill patients and so had no direct financial incentive to give special treatment to the 98 higher ranker officers (or spouses) despite their higher incomes. We speculate that rank is important simply because it denotes higher social class and, perhaps, political influence.
Finally, our discovery that the satellite clinic has higher satisfaction levels than the base hospital clinic is worthy of note. We take as a matter of course that managers should monitor satisfaction levels at particular clinics so as to assure that local performance does not drop below norms. Friendliness, patient centered styles of communication [19], shorter wait times, midlevel providers [5,17], seeing the same physician [20], female physicians [21], and racially-concordant physicians [19] may be more common at the satellite clinic while language barriers may be less common; the hospital clinic could take steps to improve these aspects of their services. The base hospital outpatient clinic suffers from some particular disadvantages, including parking problems and a more hectic atmosphere. Further decentralization of OB/GYN services is worthy of consideration as a strategy for addressing these issues. A national study of veteran patients found that veterans seen in free-standing "community-based" outpatient clinics were more satisfied than those seen in traditional hospital outpatient clinics, though the effect was not strong [1]. Easier parking, less travel time and shorter wait times may be at work to produce this result. A study of patients in the Israel Defense Forces found an inverse relationship between satisfaction and clinic size [22]. An observational study of 60 general practices in England found that satisfaction with access to care was better in small practices. Scores for overall satisfaction, continuity of care, and access to care were higher in practices were staff reported a better team climate [23]. If smaller clinics are better able to develop a team spirit, this could explain higher satisfaction levels. Additional research into the relationships between clinic size, team practice, staff morale, and patient satisfaction are needed.
Conclusion
The results of this study should be treated with caution. The sample may not have been representative of the clinic population. Some cases were dropped from the multivariate analysis due to missing data on one or more variables. Furthermore, San Diego's Naval Medical Center may not be typical of all Naval Medical Centers. Certainly, findings drawn from military clinics may not be generalizable to civilian settings. And, of course, conclusions about patient satisfaction do not necessarily apply to quality of care, since satisfaction reflects lay judgments about quality and may not be accurate on technical matters.
Despite these limitations, we believe that our results contribute to the OB/GYN literature and have practical implications. Clinic managers and physicians should seek to enhance patient satisfaction, especially for patients seen in larger clinic settings and patients in the lower ranks. The reasons why Asian American patients might be less satisfied than other patients should be explored. The mounting evidence in favor of smaller clinics suggests that it is time for further decentralization of services. The purpose behind these suggestions is to make care more personalized and patient-centered.
Our findings suggest a need for additional research. Knowing that Asian patients may be less satisfied begs for further investigation. Lower ratings from Asian Americans may reflect different response tendencies rather than less satisfaction with care. In addition, learning that a satellite clinic produces more satisfied patients suggests that a larger study should be conducted comparing satellites to base hospital clinics.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JR analyzed the data and wrote the first draft of the paper. JL conceived the survey and planned it. SG organized data collection and data entry and critiqued the paper.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The cooperation of the US Naval Base in San Diego was essential to the completion of this study. The conclusions do not represent the official policy of the US government and are solely the responsibility of the authors.
==== Refs
Borowsky SJ Nelson DB Fortney JC Hedeen AN Bradley JL Chapko MK VA community-based outpatient clinics: performance measures based on patient perceptions of care Med Care 2002 40 578 586 12142773 10.1097/00005650-200207000-00004
Donabedian A The Definition of Quality and Approaches to its Assessment 1980 Ann Arbor, Health Administration Press
Finkelstein BS Harper DL Rosenthal GE Patient assessments of hospital maternity care: a useful tool for consumers? Health Serv Res 1999 34 623 640 10357293
Finkelstein BS Singh J Silvers JB Neuhauser D Rosenthal GE Patient and hospital characteristics associated with patient assessments of hospital obstetrical care Med Care 1998 36 AS68 78 9708584 10.1097/00005650-199808001-00008
Tucker JS Hall MH Howie PW Reid ME Barbour RS Florey CD McIlwaine GM Should obstetricians see women with normal pregnancies? A multicentre randomised controlled trial of routine antenatal care by general practitioners and midwives compared with shared care led by obstetricians BMJ 1996 312 554 559 8595287
Grumbach K Selby JV Schmittdiel JA Quesenberry CP Quality of primary care practice in a large HMO according to physician specialty Health Serv Res 1999 34 485 502 10357286
Clemons JL Aguilar VC Tillinghast TA Jackson ND Myers DL Patient satisfaction and changes in prolapse and urinary symptoms in women who were fitted successfully with a pessary for pelvic organ prolapse Am J Obstet Gynecol 2004 190 1025 1029 15118635 10.1016/j.ajog.2003.10.711
Davis TL Lukacz ES Luber KM Nager CW Determinants of patient satisfaction after the tension-free vaginal tape procedure Am J Obstet Gynecol 2004 191 176 181 15295361 10.1016/j.ajog.2004.03.085
Smith WJ Upton E Shuster EJ Klein AJ Schwartz ML Patient satisfaction and disease specific quality of life after uterine artery embolization Am J Obstet Gynecol 2004 190 1697 1703 discussion 1703-1696 15284774 10.1016/j.ajog.2004.02.043
Jensen JT Bird M Leclair CM Patient satisfaction after the treatment of vulvovaginal erosive lichen planus with topical clobetasol and tacrolimus: a survey study Am J Obstet Gynecol 2004 190 1759 1763 discussion 1763-1755 15284791 10.1016/j.ajog.2004.02.061
Basen-Engquist K Shinn EH Warneke C de Moor C Le T Richards-Kortum R Follen M Patient distress and satisfaction with optical spectroscopy in cervical dysplasia detection Am J Obstet Gynecol 2003 189 1136 1142 14586367 10.1067/S0002-9378(03)00540-4
Yost NP Bloom SL Sibley MK Lo JY McIntire DD Leveno KJ A hospital-sponsored quality improvement study of pain management after cesarean delivery Am J Obstet Gynecol 2004 190 1341 1346 15167840 10.1016/j.ajog.2003.10.707
Main EK Bloomfield L Hunt G Sutter Health FPaDCIC Development of a large-scale obstetric quality-improvement program that focused on the nulliparous patient at term Am J Obstet Gynecol 2004 190 1747 1756 discussion 1756-1748 15284788 10.1016/j.ajog.2004.02.055
Taira DA Safran DG Seto TB Rogers WH Kosinski M Ware JE Lieberman N Tarlov AR Asian-American patient ratings of physician primary care performance J Gen Intern Med 1997 12 237 242 9127228 10.1046/j.1525-1497.1997.012004237.x
Weech-Maldonado R Morales LS Elliott M Spritzer K Marshall G Hays RD Race/ethnicity, language, and patients' assessments of care in Medicaid managed care Health Serv Res 2003 38 789 808 12822913 10.1111/1475-6773.00147
Seid M Stevens GD Varni JW Parents' perceptions of pediatric primary care quality: effects of race/ethnicity, language, and access Health Serv Res 2003 38 1009 1031 12968814 10.1111/1475-6773.00160
Roblin DW Becker ER Adams EK Howard DH Roberts MH Patient satisfaction with primary care: does type of practitioner matter? Med Care 2004 42 579 590 15167326 10.1097/01.mlr.0000128005.27364.72
Bean-Mayberry BA Chang CC McNeil MA Whittle J Hayes PM Scholle SH Patient satisfaction in women's clinics versus traditional primary care clinics in the Veterans Administration J Gen Intern Med 2003 18 175 181 12648248 10.1046/j.1525-1497.2003.20512.x
Cooper LA Roter DL Johnson RL Ford DE Steinwachs DM Powe NR Patient-centered communication, ratings of care, and concordance of patient and physician race Ann Intern Med 2003 139 907 915 14644893
Morgan ED Pasquarella M Holman JR Continuity of care and patient satisfaction in a family practice clinic J Am Board Fam Pract 2004 17 341 346 15355947
Howell EA Gardiner B Concato J Do women prefer female obstetricians? Obstet Gynecol 2002 99 1031 1035 12052594 10.1016/S0029-7844(02)01980-4
Mandel D Zimlichman E Wartenfeld R Vinker S Mimouni FB Kreiss Y Primary care clinic size and patient satisfaction in a military setting Am J Med Qual 2003 18 251 255 14717383
Campbell SM Hann M Hacker J Burns C Olivere D Thapar A Mean N Safran DG Roland MO Identifying predictors of high quality care in English general practice: observational study BMJ 2001 323 784 787 11588082 10.1136/bmj.323.7316.784
| 15890078 | PMC1134665 | CC BY | 2021-01-04 16:30:36 | no | BMC Womens Health. 2005 May 12; 5:6 | utf-8 | BMC Womens Health | 2,005 | 10.1186/1472-6874-5-6 | oa_comm |
==== Front
Nutr Metab (Lond)Nutrition & Metabolism1743-7075BioMed Central London 1743-7075-2-111586970310.1186/1743-7075-2-11ReviewMechanisms of lipid malabsorption in Cystic Fibrosis: the impact of essential fatty acids deficiency Peretti N [email protected] V [email protected] E [email protected] E [email protected] Department of Nutrition, CHU-Sainte-Justine, Université de Montréal, Montréal, Québec, Canada2 Department of Pediatrics, CHU-Sainte-Justine, Université de Montréal, Montréal, Québec, Canada2005 3 5 2005 2 11 11 19 2 2005 3 5 2005 Copyright © 2005 Peretti et al; licensee BioMed Central Ltd.2005Peretti 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.
Transport mechanisms, whereby alimentary lipids are digested and packaged into small emulsion particles that enter intestinal cells to be translocated to the plasma in the form of chylomicrons, are impaired in cystic fibrosis. The purpose of this paper is to focus on defects that are related to intraluminal and intracellular events in this life-limiting genetic disorder. Specific evidence is presented to highlight the relationship between fat malabsorption and essential fatty acid deficiency commonly found in patients with cystic fibrosis that are often related to the genotype. Given the interdependency of pulmonary disease, pancreatic insufficiency and nutritional status, greater attention should be paid to the optimal correction of fat malabsorption and essential fatty acid deficiency in order to improve the quality of life and extend the life span of patients with cystic fibrosis.
==== Body
I Introduction
Cystic fibrosis (CF) is the most common autosomal recessive genetic disease observed in the Caucasian population, affecting about 1 in 2,500 newborns. It is caused by mutations in the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) gene discovered in 1989. The gene was cloned from chromosome 7q21-31, is composed of 27 exons and encodes a protein that functions as a chloride channel [1-4]. CFTR has 1,480 amino acids and the molecular weight varies from 140 kDA to 170 kDA, depending upon the degree of posttranslational glycation in the Golgi complex, which appears to vary somewhat according to cell type and genotype. Defective cAMP-dependent chloride ion conductance leads to an imbalance in fluid flow across epithelia, thickened mucus and blocked exocrine ducts in the affected tissues. In addition to respiratory-tract manifestations that represent the major cause of death in CF, the vast majority of young CF subjects have a wide variability of clinical expression, including gastrointestinal, metabolic and nutritional problems [5,6]. Advances in medical management have led to a continual improvement of life expectancy for CF patients. However, intestinal fat malabsorption remains a persistent feature, given the difficulty of achieving complete correction in clinical practice despite the remarkable benefits resulting from exogenous pancreatic enzyme replacement therapy. [7-11]. The purpose of this article is to present a systematic critical review of literature and data sources pertaining to fat malabsorption in CF, emphasizing the role of abnormal, intraluminal and intracellular factors. Furthermore, the relationship between essential fatty acid deficiency (EFAD) and intestinal fat transport is particularly underscored given its implication in gastrointestinal pathophysiology. To facilitate the reader's appreciation of major gastroenterological problems in CF, we first describe the normal digestive and absorptive processes before addressing the issue of various defective mechanisms in CF.
II Lipids and Essential Fatty Acids (EFAs)
Dietary fat intake represents 35% (100 to 150 g per day) of total calories consumed in the North American diet [12]. Dietary fats are composed mainly of triacylglycerol (TG) with 92–96% long-chain fatty acids [13]. In addition to dietary intake, lipids are derived from bile and desquamated mucosal cells.
Lipids have varied as well as essential functions in organisms. TGs are the principal reserve supply of energy, while phospholipids (PLs) and cholesterol are crucial for the elaboration of cell membranes, and play a functional role in steroid hormone and biliary acid (BA) synthesis. Linoleic (18:2n-6) and linolenic (18:3n-3) acids have recognized functions in growth, the development of the central nervous system, immune and vascular functions, eicosanoid production, membrane fluidity and the control of lipid homeostasis. Adequate EFA levels depend entirely on adequate dietary intake and absorption because mammalian cells cannot synthesize de novo EFA efficiently. Once these EFAs are appropriately absorbed, they can be converted into long-chain polyunsaturated fatty acids (PUFAs) such as arachidonic acid (AA, 20:4n-6), eicosapentaenoic (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3). Conditions leading to fat malabsorption, as in CF, have been associated with a high incidence of EFAD. However, diet-induced EFAD impairs dietary lipid absorption through various mechanisms, which will be thoroughly discussed sin this review.
II.1. CF Patients and EFAD
A high EFAD incidence (85%) has been frequently reported in CF patients [14-19]. EFAD is most marked at infancy [20] and EFA impairment has been observed most significantly in the cholesteryl ester (CE) fraction [15,21]. The most often reported abnormalities in plasma are a decrease in linoleic acid and AA, its major metabolites, as well as DHA [22]. This EFA diminution is associated with an elevation in eicosatrienoic acid (20:3n-9), oleic acid (18:1n-9) and palmitoleic acid (16:1n-7) [14,23,24], resulting in a rise in the triene/tetraene ratio. It has been suggested that the (20:3n-9/20:4n-6) ratio is a very sensitive and reliable index of EFAD. A value of the ratio above 0.2 indicates an appropriate cutoff for the diagnosis of EFAD [25]. However, results obtained from CF studies vary, possibly as a consequence of different nutritional status, the degree of malabsorption and the severity of the CFTR mutation.
Abnormal EFA status in CF patients does not seem to be the consequence of intestinal malabsorption only, since reduced EFA values are found as early as in the first weeks of life in infants with CF [22] and they are also present in well-nourished young CF patients who do not receive a low-fat diet and do not present with fat malabsorption [17,24,26-28] even with regular pancreatic enzyme therapy [21]. Among other mechanisms related to EFAD, one can mention the excessive oxidation of EFA as an energy source [29], the exaggerated utilization of eicosanoids as precursors of inflammatory response [30,31], the higher rate of lipid turnover in cell membranes [32], the impaired metabolism of EFA with a defect in plasma membrane incorporation, the decreased activity of desaturases [14,33] and lipid peroxidation [34].
II.2. Relationship between CFTR and EFAD
CFTR primarily functions as a chloride channel in the apical membrane of the respiratory and gastro-intestinal tracts. An elevation of cAMP within the cell results in increased chloride transport through the CFTR. Movement of water is linked osmotically to such ion transport and chloride secretion represents one likely means of hydrating the mucosal surface of these organs. Other important studies have implicated the CFTR in additional functions, including the regulation of distinct Cl- channel proteins, the modulation of intracellular vesicle and Golgi acidification, and the control of vesicle trafficking [35-40]. Furthermore, it has been suggested that the CFTR may regulate membrane PL composition. In human airway epithelial cultured cells, the incorporation of fatty acids (FAs) into membrane PLs is decreased when chloride conductance channels are chemically blocked [41]. A defective CFTR reduces the incorporation of linoleic acid in the PL of CF cells and transfecting these cells with the normal gene increases linoleic acid incorporation [42]. Thus, a relationship is possible between chloride transport conductance and EFA metabolism. To date, these studies are limited and additional investigation is needed to establish the role of the chloride channel function in EFA movement and metabolism.
CFTR knock-out mice demonstrate EFAD. But the EFA composition of the cell membrane varies with different animal models. In cftr-/- mouse, the pancreas, lungs and ileum were deficient in (n-3) EFA [43], whereas in cftr tm1HGU/tm1HGU mice, the deficiency was in (n-6) EFA in plasma, lung and pancreatic PL [44]. The PUFA differences between the two mouse models emphasize the importance of determining the exact composition of membrane PL from healthy men and CF patients according to genotype in order to tailor appropriate EFA nutritional intervention.
In CF, AA levels suggest that abnormal EFA status may result from an impaired EFA metabolism. High levels of AA are also reported in the cell membrane of the ileum of CFTR knock-out mice [43]. The same AA imbalance has been reported in the PL fraction deriving from bronchial alveolar lavage fluid, plasma and tissues [31,45-48] of CF patients. The increased AA level in tissues may result from either overproduction or diminished catabolism. In CF cells (CFPAC-1), the conversion of 18:2n-6 into 20:4n-6 by desaturation and elongation processes is enhanced 1.8-fold compared with CF cells transfected with the normal gene [42]. AA overproduction is compatible with an intrinsic increase in AA release by CFTR-mutated respiratory epithelial cell lines [49]. A decreased metabolism of AA through cyclooxygenase, lipooxygenase or cytochrome P450 pathways could also explain high levels of AA in tissues, but this hypothesis should be thoroughly tested in CF because of a) the high rates of inflammatory species originating from AA that characterize CF patients; and b) the enhanced turnover of endogenous and exogenous AA in polymorph neutrophil granulocytes, which is possibly due to enhanced phospholipase A2 (PLA2) activity [50]. Then, AA catabolism is not obviously diminished, but the ratio between production and degradation may be unbalanced in favor of overproduction. Accordingly, it has been suggested that AA participates in CF physiopathology, since AA has been shown to inhibit CFTR Cl- currents when applied to the cytoplasmic face of excised membrane patches [51,52]. Other FAs may also block CFTR pores with various levels of efficiency. For example, cis-unsaturated FAs are more effective than trans-unsaturated FAs and saturated FAs in interfering with CFTR function [52]. Furthermore, in respiratory epithelial cells, CFTR mutation ΔF508 is associated with an intrinsic increase in AA release [49] suggesting that the mechanism of chronic inflammation in CF, at least in part, involves this abnormality. Importantly, Chlore ion (Cl-) secretion across epithelial cells is induced by AA [53,54]. On the other hand, AA decreases Cl- secretion across tracheal epithelium when its metabolism is impaired by cyclooxygenase and lipooxygenase inhibitors [54] suggesting that enhanced Cl- secretion results from the action of its metabolites [52]. Furthermore, there is a correlation between EFAD and CF genotype in homozygotes for ΔF508 and heterozygotes for 394 delTT have displayed significantly lower concentrations of linoleic and DHA than the other mutations [55]. Thus, the relationship between AA metabolism abnormalities and specific mutations may explain the discrepancies observed with n-6 EFA supplementation [19,22,56-58]. Evidently, extensive studies are required to clarify the link between the CFTR and FAs, especially in view of the contribution of EFAD to CF pathophysiology.
In animals, EFAD induces symptoms similar to those in CF: defective Na+ transport, increased bacterial colonization of airways, formation of clusters of lipid-laden macrophages in lungs [59], impaired alveolar macrophage function [60], alteration of surfactants [61], liver steatosis, decreased insulin secretion and increased caloric needs [62]. Interestingly, a membrane lipid imbalance plays a role in the phenotype expression in CFTR knock-out mice and feeding these animals with DHA has normalized disease-related changes that occur in epithelial cells and intestinal mucosa [63].
Of note is the finding that altered PL composition and structure can impair membrane fluidity [64-66], thereby modifying CFTR cell membrane localization and function. Membrane PL composition in CFTR knock-out mice is different from controls [44]. Apparently, lipid membrane composition regulates membrane protein activity through the association between lipid rafts and these proteins [67]. The transient residency of proteins in rafts emerges as a regulatory mechanism responsible for protein biological activity. Recently, it has been reported that lipid raft localization for CFTR is required for signaling in response to Pseudomonas aeruginosa infection, underlying the interaction between the lipid composition of plasma membrane and the CFTR protein [68]. It seems, therefore, that a correction of PL composition may improve CFTR activity in vivo, a hypothesis that calls for direct experimental validation in intestinal cells.
II.3. Nutritional Status and Prognosis in CF
Various studies emphasize a correlation between a low nutritional status and an unfavorable prognosis in patients with CF [69]. Chronic undernutrition is associated with weight and height retardation in CF children. Early studies showed a correlation between the degree of malnutrition and the severity of pulmonary disease [69-71] and a slower rate of deterioration of pulmonary function was found in CF children without steatorrhea [72]. However, low values of EFA (linoleic acid and DHA) are not always correlated with anthropometric data or lung function [55]. CF survival and well-being are also correlated with malnutrition in CF [73]. This positive correlation between good nutritional status and long-term survival underscores the optimization of energy balance, nutrient intake, as well as intestinal absorption for CF patients.
Since EFA levels are abnormal in CF patients and EFAD may contribute to CF pathophysiology in animal models, it has been suggested that EFA supplementation has a beneficial effect in this disease. Nutritional studies demonstrate that oral and intravenous supplementations with n-3 FAs are efficient in increasing EPA and DHA incorporation in plasma, erythrocyte and platelet membrane. [56-58]. Oral supplementation with n-6 FAs increases plasma and erythrocyte membrane AA levels [22,74-76]. The clinical effects of EFA supplementation vary among studies: linoleic acid increases weight for height in CF children [19,22] and an improvement in pulmonary function is also observed with n-3 FA intake. Nevertheless, this valuable effect is not reported in all studies. [56-58] and the differences may be explained by the limited number of patients (under 20) and the short duration (between 4 and 6 weeks) that prevent the proper evaluation of any clinical benefit. Undoubtedly, further studies should focus on the relationship between dose recommendations, the degree of correction of plasma EFAs and clinical outcome. In particular, long-term lipid dosage adjustments are necessary to define the optimal n-3/n-6 ratio required in CF patients to prevent clinical manifestations.
In contrast to carbohydrates and proteins, lipids are poorly soluble in water. Their lipolysis into the aqueous phase of the intestinal lumen, as well as their transport into plasma, requires the formation of soluble complexes such as BAs and lipoproteins (LPs). Any deficiency or variation in these processes may impact on fat digestion and transport. Fat malabsorption in CF may result from the inefficiency of sequential events: the lipolysis of alimentary lipids, the effect of BAs on micellar solubilization, the integrity of intestinal mucus and enterocytes, multiple intracellular processes and LP secretion. These lipid digestive and absorptive steps will be discussed chronologically.
III. Intraluminal Abnormalities
III.1.1. Lipolytic Phase
TGs cannot be transported into the cell because of their size and their hydrophobic characteristics. The lipolysis process is required to produce smaller and more hydrophilic molecules, which can thus be absorbed. TGs can be hydrolyzed by two lipolytic enzymes in humans: gastric lipase and pancreatic lipase [77,78]. The enzyme postulated to be the main participant in gastrointestinal lipid digestion is pancreatic lipase. However, intra-gastric lipolysis accounts for about 20–30% of total gastrointestinal lipolysis in vivo and it has been suggested that it is increased in CF pancreatic insufficiency.
III.1.1 Gastric Lipase
Human gastric lipase is the key acid enzyme that is secreted by the chief cells located in the fundic region of the stomach. Human gastric lipase is stable in gastric juice at pH values ranging from 2.0 to 7.0 and is the most stable acid lipase [79].
In the past, it was suggested that gastric lipase played only a minor role in the digestion of dietary fat, because of its specificity for medium-chain FAs, which are less common than long-chain FAs in dietary fat [80]. In fact, human gastric lipase also exerts an important activity on long-chain TGs in an acidic pH. When pH levels are between 5 and 6, it hydrolyzes medium-chain TGs more efficiently than longer ones [81]. This lipase is quantitatively important, contributing 20–30% of TG acyl chain hydrolysis [82-84]. Moreover, initial digestion of dietary fat in the stomach was found to be an essential step for optimal intestinal lipolysis for three reasons: releasing long-chain PUFA in the stomach induces cholecystokinine output, which stimulates pancreatic lipase secretion [85]; the efficiency of pancreatic lipase is increased with partially hydrolyzed TGs [86]; and free fatty acids (FFAs) enhance lipase binding to colipase [85].
Acid lipase activity exists in CF children [87,88], but its level, compared with normal controls, varies among studies [89,90]. Basal lipase activity in the stomach and in the duodenum is similar in CF patients and normal controls [91], but its lipolytic activity is almost three times higher in CF patients after a test meal [90]. There are two main explanations for the maintained high activity of acid lipase in CF patients. First, contrary to pancreatic lipase, the low postprandial pH in the upper small intestine, due to bicarbonate deficiency, does not inactivate gastric lipase activity [92,93] and the latter is preserved for a longer period than in normal subjects. Second, the inactivation of gastric lipase is less effective, because of the low concentration of BAs and pancreatic proteolytic enzymes [94,95]. Compensatory lipolytic activity by gastric lipase may account for 40–70% of dietary fat digestion [87,96]. In CF patients, preduodenal activity amounts to 90% of the total lipase activity at the ligament of Treitz in the postprandial state [96]. However, acid-resistant enzyme supplements, such as fungal lipase, fail to prove efficacy in CF [97]. No correlation between gastric lipase and EFA status is reported in the literature.
III.1.2. Pancreatic Lipase and Colipase
Amphipathic molecules, monoacylglycerols (MG) and FA, produced during TG hydrolysis, stabilize dietary fat emulsion. Most of the fat digestion occurs in the upper small intestine following the action of pancreatic lipase and BAs. The hydrolysis of long-chain TGs is achieved by pancreatic lipase. Its optimum pH is 8 and the enzyme becomes inactive at a pH under 6 and is irreversibly inactivated at a pH less than 4.5. Colipase is required for pancreatic lipase stabilization on lipid droplets and for optimal lipolytic activity. At least, two other hydrolytic enzymes are needed to release FAs: PLA2 and cholesterol esterase (CET).
Fat malabsorption is present only when pancreatic enzyme secretion is under 10% of normal [98]. On diagnosis, only 15% of CF patients have sufficient pancreatic function for normal fat digestion [72]. The pancreatic function decreases with time, since 37% of CF newborns show a substantial preservation of pancreatic function [99] and more than 90% of CF infants and younger children display pancreatic insufficiency. The physiopathology of pancreatic insufficiency results from several abnormalities. In fact, pancreatic changes are caused by the obstruction of small ducts by thick, sticky secretions and cellular debris. A primary defect of fluid secretion (water and bicarbonate) leads to high protein concentration in pancreatic juice [100]. Then, protein precipitation blocks small pancreatic ducts and promotes pancreatic acinar atrophy and fibrosis. The ratio of acinar cells to connective tissue decreases progressively in the pancreas of CF children until the normal structure of the pancreas is lost. The role of the CFTR for normal proliferation and differentiation of secretory cell populations has been demonstrated in the lung and intestine of transgenic mice [101,102] and, interestingly, the CFTR appears as a valuable marker of human pancreatic duct cell development and differentiation [103]. CFTR mutation may, therefore, affect the proliferation of pancreatic secretory cells, thereby favoring pancreatic insufficiency and physiopathology.
Not only quantitative but also qualitative abnormalities characterize pancreatic lipase activity. Pancreatic lipase supplements are inactivated by acidic pH levels in the small bowel: only 8% of active enzyme reaches the ligament of Treitz [92,93]. The severity of pancreatic involvement appears to increase with age and, even with maximum pancreatic enzyme replacement therapy, a substantial proportion of ingested energy is probably malabsorbed.
There is little information about colipase in CF. Although, colipase has been reported to be deficient in this disease [104], its secretion is extremely variable in CF patients [105]. More investigations are necessary to clarify this aspect.
End products of gastric lipases are mainly one FFA and one 1-3-diacylglycerol, while the lipolytic products of pancreatic lipase are two FFAs (Sn-1 and Sn-3 postions) and one 2-monoacylglycerol [106]. Furthermore, in contrast to pancreatic lipase, gastric lipase has been shown to cleave not only the external ester bonds (n-1 and n-3 position), but also the ester bond in the Sn-2 position [107]. This difference is important for nutritional intervention with enteral defined structured TGs, which provide medium-chain FA as a source of immediate energy. In Caco-2 cell cultures, the addition of structured TGs containing only octanoic acid, a non-EFA, in the Sn-1 and Sn-3 position, but linoleic acid in the Sn-2 position in the presence of gastric lipase, enhances the cellular uptake of TGs, improves cellular EFAD and exhibits a beneficial effect on lipid incorporation and LP production [108]. In the rat, EFAs at the Sn-2 position appear faster in lymph than EFAs at the terminal positions of TGs [109] and are more readily recovered in lymph [110]. Defined TG with EFA linoleic acid at the Sn-2 position increases the absorption of EFA in rats with fat malabsorption induced by biliary and pancreatic diversion [111]. These results suggest that structured TG, in the presence of gastric lipase, could improve EFAD, despite the absence of pancreatic lipase activity. In patients who are not supplemented with pancreatic enzymes, an organized lipid matrix containing lysophosphatidylcholine (Lyso-PC), MG and FFA could be effective in increasing TG and retinyl-palmitate absorption and in improving clinical outcome [112].
In animal models, EFAD does not affect the lipolytic phase: there is no difference between EFAD and control rodents in the lipolytic step [113,114]. Furthermore, animal models suggest that EFAD is likely to be of minor importance in exocrine pancreas function in CF [113,115]. Furthermore, EFAD does not seem to induce any significant changes in the structure and composition of pancreas [116]. Overall, these studies show that structured TG can effectively supply functional FAs distributed throughout the glycerol skeleton. In particular, structured TG with EFAs at the Sn-2 position and medium chain FAs at the Sn-1 and Sn-2 positions of the glycerol molecule can be used for distinct nutritional and clinical purposes in CF. Although, it is obvious that structured TG are far superior to conventional oils, it is quite difficult to appreciate their influence on lipolysis in CF patients with EFAD in view of the paucity of information available in the literature. Probably, the difficulty originates from the enormous cost of producing highly purified structured TG.
III.1.3. Phospholipase A2 (PLA2) and Cholesterol Esterase (CET)
The phospholipid-hydrolyzing enzymes include phospholipase A1 (PLA1), PLA2 and CET [117]. PLA1 and PLA2 are secreted in their zymogen form and activated by trypsin on entering the duodenum [118]. PLA2 preferentially hydrolyzes FA in phosphatidylcholine (PC), the most abundant dietary PL, at position 2 to produce more hydrosoluble components (FA and 2-lyso-PC). CET hydrolyzes CE, retinyl ester and lyso-PC in the small bowel [117], but it is 100 times less effective than pancreatic lipase in the TG digestion process.
In CF patients, PLA1 and PLA2 activity is correlated with the pancreatic function [119]. It is of particular interest to note that, despite the total absence of PLA2, around 30% of PLs are hydrolyzed [120]. The decreased activity of PL hydrolyzing enzymes may interact with lipid absorption in several ways. First, since PLs are amphipathic molecules, they adsorb to the surface of lipid droplets, preventing contact between the lipase-colipase complex and the TG substrate [118]. PC hydrolysis will allow desorption of lyso-PL from lipid droplets, which are relatively hydrosoluble molecules, and will thus facilitate the interaction between TG and pancreatic lipase. Secondly, unhydrolyzed PLs impair the absorption of cholesterol and FAs [121,122]. Finally, they impair bile acid absorption through their receptors in the terminal ileum, which increases bile acid losses in the stool.
In CF children, CET activity is dramatically decreased [123] and parallels pancreatic lipase activity. The reduction in BAs, which are essential cofactors for optimal esterase activity, results in a failure to solubilize and thus absorb cholesterol and their esters.
In the enterocyte, lyso-PC is an essential source of PLs for LP formation [124], the transport of TGs from the enterocyte into the blood [125] and the PL turnover induced by exocytosis [126]. These phenomena may be impaired by phospholipase or CET defects.
No data are provided about the impact of EFAD on PLA1, PLA2 and CET activity. However, AA and DHA decrease the secretion of bile-salt dependent lipase, but without directly impairing the biosynthesis of this enzyme. AA alters the transport of the enzyme toward the cytosol, leading to the retention of bile-salt dependent lipase in microsomes [127]. It is likely that much attention must be paid to phospholipases and CET in CF in view of their role in fat absorption. Future studies should be designed to test whether their supplementation may increase lipolysis efficiency and alleviate EFAD in CF.
III.1.4. Intraluminal pH
Duodenum pH results from an interaction among food buffering, gastric acid production and pancreatic bicarbonate secretion. In humans, physiological pH is 2–4 in the stomach and it varies from 6–7 in the upper duodenum, but rarely dips below 5.
In CF patients, pancreatic, intestinal and biliary HCO3- secretion is decreased to below 10% of the normal values [128]. Consequently, a low pH is observed in the intestinal lumen of the duodenum and even during postprandial time. Besides, basal and postprandial pH levels in the stomach are similar in CF patients and normal subjects [91,96]. In CFTR(-/-) mouse intestine, HCO3- secretion is impaired too [129,130]. The acidic upper small intestine in CF may contribute to fat malabsorption given: 1) the inactivation of trypsin, which induces an impairment of active enzyme molecules from their zymogen form (pancreatic lipase, PLA2); 2) the inactivation of pancreatic lipase [131]; 3) BA precipitation, decreasing the BA pool and diminishing its availability to form micelles; and 4) early protonation, which impairs the micellar dispersion of lipolytic products [132]. The ability of acid suppressant therapy (cimetidine, ranitidine or proton pump inhibitors) to improve fat absorption is controversial. Recently, a metanalysis of randomized trials involving agents that reduce gastric acidity has been carried out and it showed that treatment failed to improve nutritional status and had little impact on fat absorption [133]. Differences between studies could be due to the small number of subjects, differences in dietary intake and the degree of pancreatic insufficiency, the important physiologic range in fat intestinal absorption and variability in intestinal drug absorption or interindividual variation in medication bioavailability, pH measurement and fat absorption evaluation.
III.2. Micellar Phase and Biliary Abnormalities
BAs are amphipathic molecules synthesized by the liver. They enhance the solubility of lipolytic products in the aqueous intestinal phase [134,135]. In fact, micellar solubilization increases the aqueous concentration of MG and FA 100 to 1000 times. The importance of this phase is underscored by a 30% reduction in dietary lipid absorption in patients with biliary atresia [136]. Similarly, rats with biliary drainage displayed reduced linoleic acid absorption [137].
III.2.1. BA Excretion
Outside the critical micellar concentration range, the amount of lipid solubilized is significantly reduced. Consequently, steatorrhea appears [138] and FA absorption decreases [139]. This situation is found in CF patients where total bile salt secretion is impaired [140]. About 36% of CF children showed reduced BA secretion into the duodenum [141]. Despite this low BA secretion, impaired water secretion leads to high concentrations of bile salts in hepatobiliary secretions [142], which may contribute to the cholelithiasis reported in CF patients [94,143].
EFAs regulate BA excretion: PUFAs are reported to induce the excretion of BA in rats [144] and humans [145], but this effect depends on the EFA family, since n-3 FA increases bile flow more than n-6 FA [146]. The EFAD effect on bile flow and BA secretion varies among animal species: decreasing in rats [147,148], having no effect in mice [114] and increasing in hamsters [114]. In humans, the impact of EFAD on BA secretion and composition remains unclear. However, prostaglandins, which are AA metabolites, alter hepatic bile flow [149] indicating that disturbances of AA status may affect choleretic response in CF patients.
III.2.2. BA Composition
The nature of BAs influences their ability to solubilize lipids. Cholesterol absorption by enterocytes is greater with cholyl-taurine than with chenodeoxycholyl-taurine [150-152] and BAs with conjugated taurine are more effective than glycine-conjugated BAs at solubilizing lipids. The effect of EFA on BA composition variation and fat absorption remains enigmatic in humans, since there are no available data.
In CF, oral taurine load appears to be normal, but excessive fecal loss [153] increases the ratio of glycine/taurine-conjugated BA [94,140,141]. Oral taurine supplementation in CF children is effective in decreasing the glycine/taurine ratio in duodenal fluid [153], but its ability to improve fat malabsorption is controversial [154-157]. Apparently, taurine supplementation improves BA malabsorption, mainly in patients with a high degree of steatorrhea [158-160].
Studies on animals show that EFAD leads to impaired biliary excretion of taurocholate [161] and to reduced EFA content (linoleic and arachidonic acid) in biliary PC that is essential for micelle formation [162]. However, BA and bile lipid composition vary across species. In rats and hamsters with EFAD, bile composition is markedly impaired [147,161]. On the other hand, BA composition appears to be similar in EFA-deficient and EFA-sufficient mice [114].
The relationship between BA and EFA was investigated in CF-EFAD children and it showed that ursodeoxycholic acid supplementation improves the hepatic metabolism of EFA. After 1 year, ursodeoxycholic acid supplementation led to an improvement of EFA status (reduction of triene/tetraene FA ratio) [163].
III.2.3. BA Malabsorption
BAs are recycled through the enterohepatic cycle with remarkable efficiency (95% reabsorbed) [164,165]. The enterohepatic cycle is often interrupted in CF because of excessive fecal losses of BA, [95,166-170], which diminish the BA pool [171,172]. Excessive BA losses may be attributed to: 1) BA irreversibly bound to unhydrolyzed TG and PL within the intestinal lumen [173] ; 2) BA precipitation due to acidic pH in the duodenum [174]; and 3) intestinal bacterial overgrowth, which is present in 40% of CF patients and results in BA deconjugation and dehydratation [175]. BA loss may also result from a primary cell defect in the active absorption of BA in the ileum. In vitro studies using brush border membrane vesicles from CF patients have shown that total ileal BA uptake is diminished [176,177]. Surprisingly, a study of Ileal Biliary Acid Transporter (IBAT) in CFTR knock-out mice shows a BA uptake rate four-fold that of wild-type mice [178]. The increase in IBAT protein and BA uptake can be interpreted as an up-regulation in response to a low BA rate. This result should lead to a renewed interest in intraluminal events, especially for the implication of the thick mucus barrier in CF pathophysiology.
III.3. Gastric and Intestinal Transit Time
Fat digestion begins in the stomach with the action of acid lipase. Following hydrolysis by gastric lipase, medium-chain FAs are partly absorbed by the stomach [179]. Thus, the stomach plays an essential role in fat digestion, especially in pancreatic insufficiency [147].
In CF patients with pancreatic insufficiency, altered motility with an increase in gastric emptying [180] and small bowel transit time has been described [181-184]. It has been reported that slow gastric emptying reduces the success of pancreatic enzyme replacement therapy in improving TG hydrolysis, which could explain in part the variation in pancreatic enzyme replacement therapy efficiency [180,185]. However, other studies have shown that gastric emptying time is similar in both CF patients and healthy controls [90,182].
In EFAD rats, AA and linoleic acid are emptied from the stomach at similar rates and these rates do not differ from controls [186]. However, EFA-deficient mice show that the motility of epithelial cells is increased in the jejunum [187].
III.4. Intestinal Mucosa Trophicity
III.4.1. Quantitative and Qualitative Mucosal Abnormalities
In the majority of CF patients, histological brush border studies reveal normal morphology [188]. However, in some cases, abnormalities are reported, such as ileal hypertrophy [189] or partial villous atrophy in the small intestine [190,191]. This atrophic mucosa, which also occurs in knock-out mice [189], may result from acidic aggression by unbuffered stomach chyme, chronic inflammation or denutrition. Furthermore, a thick mucinous layer covering the brush border and large areas of the microvilli is revealed by electron microscopic examination of biopsies from CF patients [188], which may contribute to malabsorption in CF patients. The viscosity of the intact CF glycoprotein is almost two-fold that of normal glycoprotein [192]. Elevated viscosity may be caused by defective chloride transport, mucin hyperglycosylation or a high level of disulfide-linked peptides [193,194]. Obviously, the highly acidic properties of surface components related to oversulfatation could modify the micro acidic climate on the intestinal epithelium surface and influence interaction between micelles and enterocytes, impairing the protonation of FAs.
In EFAD, histological and biochemical alterations of the intestinal mucosa are described [195]. In rats and mice with EFAD, the height of villi is decreased leading to a diminished absorption area. Cellular differentiation was also found to be impaired [196], highlighting the role of EFA in the formation of new tissues, such as the maintenance of tissue and cell structure [197]. Furthermore, the abnormal structure of mitochondria and microvilli is correlated to decreased fat absorption [196]. It seems difficult to reconcile jejunal hypertrophy in CF patients and microvilli in EFAD animals. However, EFAD seems to induce histological abnormalities in the CFTR function. In CFTR knock-out mice, jejunal hypertrophy is corrected with an oral administration of high doses of DHA [198], which are associated with an inhibitory proliferative cell effect [199]. Then, more than in typical EFAD, an abnormal EFA imbalance could lead to hypertrophic villosities in CF patients. Hypertrophy is only localized in the jejunal portion. Accordingly, the FA composition of jejunum mucosa, the main segment for optimal lipid absorption, is markedly different from ileal and colonic mucosa [195]. Furthermore, it is interesting to note that dietary influences are tissue specific, since serum or red blood cell membranes do not reflect local changes in any of the different intestinal segments [195]. At present, no studies have correlated intestinal morphology in CF patients with their EFAD status.
III.4.2. Intestinal Permeability
Recently, studies have suggested that pathological modifications in EFAD may be the consequence of cell adhesion disorders [200].
In CF patients, the dual sugar permeability test [201] demonstrates that the paracellular pathway is more permeable to large molecules, while the passive transcellular uptake of small molecules is normal [182]. Intestinal permeability in CF is related to patient genotype: patients homozygous or heterozygous for ΔF508 mutation exhibit significantly increased intestinal permeability compared with patients with unidentified genotypes or controls [202]. Moreover, abnormalities in the tight junction, an essential structure for the control of intestinal permeability, have been reported in the intestinal epithelium of fetuses with CF [203].
EFAs are able to modify cells ultrastructurally and to alter intestinal permeability. In the culture of enterocytes from EFA-deficient CF, the lateral surfaces between cells are fairly straight, a consequence of the absence of complex interdigitations that play an essential intercellular cohesive role [196]. Moreover, a reduction in the number of desmosomes has been reported in the intestinal tract of EFAD rats [59,204]. Studies in endothelial cells suggest a possible mechanism of EFA-modulating cell adhesion. However, this EFA function has to be demonstrated in intestinal cells. In endothelial cells, some PUFAs such as γ-linolenic acid or EPA increase transepithelial electrical resistance and reduce paracellular permeability. Studies involving CF patients are definitely needed to assay intestinal cell adhesion molecules and their relation to FA status.
IV. Enterocyte Phase Abnormalities
IV.1. Enterocyte Lipid Uptake
IV.1.1. Physiology
EFAs may be absorbed by enterocytes, mostly in the form of EFA and MG. Until recently, it was assumed that these lipids diffused passively through the enterocyte brush border membrane [205]. Indeed, earlier studies reported that the uptake of FAs occurred at 0°C, implying that the process is strictly passive. In particular, the intestinal absorption mechanism of linoleate was noted to depend on its intraluminal concentration, showing a passive diffusion at high concentrations. However, a transporter was required at weaker concentrations [206]. Accordingly, the absorption of FFA was found to be a saturable phenomenon that can be inhibited through competition with long-chain PUFA [207]. These observations suggest that FA uptake is a concentration-dependent dual transport mechanism involving both passive diffusion and a carrier-dependent process. Recently, several membrane transport systems have been identified and they seem to be involved in the enterocyte absorption of lipids: membrane FA binding protein (FABPm), capable of facilitating transmembrane passage mainly of FAs, but other lipids as well [208-210]; FA translocase/cluster determinant 36 (FAT/CD36) implicated in long-chain FA transport. [211-213]; scavenger Receptor class B type I (SR-BI) that plays a role especially in cholesterol movement (absorption and/or efflux) at the enterocyte level [214]; ATP Binding Cassette transporter family that provides several cholesterol carriers: ABCG5 and ABCG8 cooperate to limit sterol intestinal absorption, rather facilitating cholesterol efflux toward the intestinal lumen and their mutations predispose to sitostrolemia [215,216]; and ABCA1 expressed in the enterocyte, which could partially control cholesterol efflux toward the intestinal lumen [217], although its exact role in the brush border membrane remains controversial [218,219]. Recently, a new protein called Nieman Pick C1-Like1 was identified in the small intestine. It seems closely involved in intestinal cholesterol absorption, a pathway sensitive to sterol absorption inhibitors such as ezetimibe [220,221].
IV.1.2. CFTR and Lipid Transporters
No study to our knowledge has investigated the interaction that may exist between the CFTR and these lipid transporters. Yet, the CFTR is known to modulate the activity of other carriers as well as certain ionic channels, for instance [37,222]. The CFTR regulation of other intestinal ionic transporters is effectively diminished in CF patients [223]. Furthermore, a microarray study on pulmonary tissue from knock-out CFTR mice shows that membrane transporters specific for ligands as different as glutamate, hormones or neurotransmitters have their expression influenced by CFTR [224]. The completion of the same type of study on intestinal tissue would likely offer interesting tracks to target lipid transport proteins capable of being influenced by the CFTR. The large diversity of transporters interacting with the CFTR could lead to impairment in enterocyte lipid uptake and trafficking in CF, which would represent another cause for nutrient malabsorption.
The existence of anomalies in the enterocyte uptake to EFA is controversial. In effect, EFA intestinal absorption in patients does not always appear to be impaired. Some studies report that the rate of linoleic acid absorption is normal when pancreatic enzyme supplementation is given at sufficient doses [225]. Apparently, even the presence of steatorrhea was not accompanied by diminished EFA absorption [186] and no correlation has been established between steatorrhea and EFAD in preadolescents with CF [226]. On the other hand, a recent study has shown that patients undergoing pancreatic enzyme treatment display a reduction in the intestinal uptake of long-chain FAs [9]. The differences between these studies may indicate either that the intestinal malabsorption does not in itself explain the EFA deficit or that the severity of CFTR mutation could influence the enterocyte absorption of lipids. Unfortunately, genotyping analysis has not been carried out in most of these investigations. Future developments will get to the bottom of these major unsolved questions by tracing the defects in enterocyte lipid uptake, the status of lipid transporters, the relationship with CFTR in CF patients and fat malabsorption.
IV.1.3. EFAD and Lipid Transporters
Early studies on EFAD did not notice an anomaly in the enterocyte transport of EFA [186,196,227]. However, these data have not been confirmed. The studies are somewhat outdated and the sensitivity of the techniques used may be called into question. Furthermore, several elements suggest that EFAD may be implicated, at least to some extent, in lipid intestinal malabsorption. Indeed, EFAD can affect the lipid composition of the enterocyte membrane and modify its fluidity, which may directly disturb the functioning of the transporters that can be found there. Moreover, some transporters like the SR-BI or CD36 could act as transporters, not directly, but by creating a special micro-environment in the neighbouring membrane lipids. This local change favors the transfer of their ligands [228]. In this model, it appears very likely that a modification of the physiochemical properties of the membrane, as is the case in EFAD, may impact on the transporter lipid transfer abilities. EFAD could also influence these lipid transporters more directly. In effect, most of these transporters are regulated by long-chain FAs: the expression of the FABP and CD36 genes is increased by long-chain FAs [229]; the SR-BI protein is also regulated according to the type of long-chain FAs (unpublished personal data); and finally, the polyunsaturated FAs trigger a decrease in the ABCA1 protein, thereby reducing the basolateral efflux of cholesterol in human Caco-2 cells [230]. This regulation could be drilled through a reduction in ABCA1 expression [231] or an increase in the degradation of the protein [232].
FAs can also act directly on the level of expression of the CFTR protein. In this way, a short-chain FA, the butyrate, increases CFTR expression significantly in animal epithelium cell cultures [233]. It is interesting to note that this same FA modulates lipid synthesis, the biogenesis of apolipoproteins (apos) and the assembly of LP in the enterocyte [234]. To our knowledge, there are no similar studies with long-chain FAs or EFAs. The potential role for the CFTR in all the enterocyte lipid synthesis steps underlines the need for new investigations, which may lead to new therapeutic strategies.
IV.2. Enterocyte Lipid Trafficking
IV.2.1. Physiology
After crossing the brush border membrane, the lipids must be processed by cytosolic proteins for their intracellular trafficking toward various compartments, including the endoplasmic reticulum (ER), where their reesterification may take place. However, the precise mechanisms behind this transport remain unclear. Certain transport proteins have been identified, but their roles often remain hypothetical: the Sterol Carrier Protein (SCP-2) and two cytosolic FABP capable of linking FA in particular, but also PL. According to our unpublished and preliminary data, intestinal-FABP (I-FABP), which is restricted to the intestine, could direct the FAs to membranes for lipid cycling or to a degradation pathway (peroxisome or mitochondria), whereas the liver-FABP (L-FABP), also found in the liver, could guide them to the ER to be assembled into LP. I-FABP and L-FABP bind differently according to lipid class, but both with a greater affinity for unsaturated FAs than saturated FAs.
IV.2.2. CFTR and Enterocyte Lipid Trafficking
No study has examined the interactions that can exist between the various proteins of the cytoplasmic transport of lipids and the CFTR. However, it has been clearly demonstrated that anomalies in CF exist in the intracellular movement of the CFTR itself. Various physiochemical conditions, such as low temperatures [233] or chemical agents [235], are able to increase the stability of the mutated CFTR protein and avoid its abnormal retention in the ER [236,237]. The latter would, in part, be attributable to a defective interaction between the CFTR and certain PLs that play the role of lipid chaperones. Obviously, the mutation of ΔF508 causes the CFTR protein to lose its ability to bind preferentially with phosphatidylserine rather than with PC. The replacement of PC with non charged analogues in mutated cell cultures increases CFTR expression and the quantity of its mature form. Therefore, certain PLs through their lipid chaperone function seem important for the intracellular trafficking of the CFTR [238]. Moreover, the invalidation of the CFTR initiates changes in gene expression as well as protein degradation via ubiquitin-dependent proteasome, which may modify several transport proteins [224]. Overall, these studies support the concept that CFTR is modulated by PLs and indicate potential relationships between CFTR and other local transporters. Unfortunately, it remains unknown how EFAs contained in PLs directly alter the activity of CFTR and whether a "partnership" exists between CFTR and intracellular lipid transporters in the enterocyte. Forthcoming investigations will highlight new links between the CFTR, EFAs and cellular processes, which may identify important factors that play a role in networks of lipid signalling and transport.
IV.2.3. EFA and Enterocyte Lipid Trafficking
The assembly of microtubules is critical for intracellular trafficking and chylomicron transport [239,240]. Early reports had indicated that the administration of microtubule inhibitors led to a decrease in the conveyance of radioactive lipids in rat enterocytes [241]. Certain EFAs, such as γ-linoleic acid or AA, regulate the microtubule polymerization [242,243]. Thus, EFAD could hinder intraenterocyte trafficking through an alteration of microtubule polymerization, synthesis or function.
IV.3. Lipid Esterification and Lipoprotein Synthesis
IV.3.1. Physiology
The products of lipolysis once absorbed and transported to the ER are resesterified to form TG, PL and CE through specific enzymatic pathways. This reesterification involves several enzymes monoacylglycerol acyltransferase (MGAT), diacylglycerol acyltransferase (DGAT), glycerophosphate acyltransferase, phosphatidate phosphodiesterase, lyso-PC acyltransferase, CE and acyl-coenzyme A:cholesterol acyltransferase (ACAT). After their synthesis, hydrophobic lipids must be associated with proteins or apolipoproteins, in order to allow their solubilization in the blood circulation, thereby forming complexes known as LP. The intestine is capable of secreting most lipoprotein classes (chylomicrons, VLDL, HDL), but chylomicrons represent the specific and most abundant class in the enterocyte.
Note that apo B is a component that is essential for LP assembly and secretion by the intestine. It exists in two forms: apo B-100, present particularly in the liver, but to a limited extent in the intestine, and apo B-48, which is specific to the enterocyte that results from the posttranscriptional modification of the apo B-100 mRNA, or editing, involving an enzymatic complex called APOBEC-I (apo B mRNA-editing catalytic subunit-1) [244]. Other apos are produced in the enterocyte, mainly apo A-I and apo A-IV that is exclusively of intestinal origin in humans [245]. During its synthesis in the ER, apo B must undergo lipidation that protects it from degradation by the proteasome [246]. Lipid transfer from the endoplasmic reticulum to apo B requires microsomal triglyceride transfer protein (MTP) intervention. This step is crucial for chylomicron assembly as noted in abetalipoproteinemia, an illness brought on by the mutation of the MTP gene, where there is defective lipoprotein secretion [247].
Pre-chylomicrons are exported to the Golgi apparatus where they undergo their final maturation (glycosylation of the apos, modification of certain PLs, etc.) before being secreted through the basolateral membrane into the lymphatic milieu. It is important to note the role of cargo proteins in the secretion step, which has been highlighted in chylomicron retention disease: the mutation of the Sar-1 GTPase protein prevents the secretion of chylomicrons through the dependent COPII vesicles [248]. TG transfer from the ER to the Golgi apparatus seems to be a limiting step in fat absorption [249]. After being secreted into the lymphatic capillaries, intestinal LPs are discharged into the systemic blood circulation through the thoracic canal.
IV.3.2. Role of the CFTR in the Intracellular Phase
In CF patients, lipid composition, concentration and size are irregular [250]. At present, the role played by CFTR anomalies is unknown. To our knowledge, no study has focused on the possible interactions of the CFTR with the implicated enzymes in lipid reesterification, apo B biosynthesis, MTP activity or the relationship with Sar-1 GTPase protein expression.
It is during this LP secretion step that interactions with the CFTR may be the most likely. In fact, CFTR mutations alter the secretion not only of electrolytes, but also of substances as different as BAs by the hepatocyte [172], γ-light chain antibodies by lymphocytes [251], INFγ by monocytes [252] or neurotransmitters by pulmonary neuroendocrine cells [253]. Since the enterocyte represents a key cell in the physiopathology of CF and given the numerous lipid aberrations observed in this disease, it appears essential to study the secretion abilities of lipids by epithelial cells instead of focusing only on the digestive mechanism in CF as is presently the case. For example, the recent analysis of RNA, influenced by CFTR knock-out, identified a number of proteins involved in LP metabolism, including proteasome 26S [224]. The latter subunit is the major proteolytic component of the ubiquitin-dependent proteasome. As mentioned before, the ER-localized ubiquitin-proteasome pathway is primarily involved in the intracellular degradation of apo B. Its alteration in the absence of CFTR [224] indicates potential relationships between CFTR and the apo B recovery/degradation pathway, thereby determining LP assembly and secretion. Furthermore, there is increasing evidence that CFTR regulates endosomal fusion and vesicular trafficking [254], indicating potential relationships between the CFTR and ADP-ribosylation factor [254], which has a central role in VLDL assembly [255]. Finally, the Sar-1/COPII complex is necessary when apo B-100 exits the hepatic ER [256] and apo B-48 containing chylomicrons are exported from the enterocyte [248]. However, the Sar-1/COPII complex is also implicated in the exiting of CFTR from the ER [257] for its entry into the proteasome degradation pathway [258]. Overall, these observations demonstrate the existence of possible interactions among apo B elaboration, chylomicron packaging and CFTR function. Hence, CFTR mutations may significantly affect lipid transport and obviously studies are needed to clarify this relationship.
IV.3.3. Possible Implications of EFAs in Intracellular Lipid Transport
It is important to underline that the anomalies in the reported LP during CF are most marked in patients with EFAD [250]. These defects may either be the direct consequence of EFAD on LP synthesis or may only reflect an association between the type of CFTR mutation and lipid metabolism. A phenotypic classification of LP profiles according to genotype would allow us to elucidate this question. In fact, EFAD could cause interference at several levels within the process of LP synthesis and secretion.
TG reesterification rates are diminished in rats during EFAD [259,260]. Lipid membrane modifications, notably long-chain FAs, are known to change the functioning of reesterification key enzymes such as MGAT [261]. Similarly, intake rich in (n-3) FA lowers the ratio (n-6)/(n-3) and leads to an activation of DGAT and ACAT [262]. Therefore, an unbalancing of the EFA status could affect the lipid reesterification step.
Biosynthesis through the enterocytes of several apos, such as apo B or apo A-IV, is regulated in a specific way by certain EFAs [263-265]. Experimental studies confirm investigations carried out on humans [266]. Thus, EFAD is possibly deleterious for the synthesis of the main intestinal LP. Accordingly, enterocyte secretion of synthesized lipids is impaired during EFAD in mice and rats, which translates into an accumulation of large lipid droplets in the intercellular space [196]. Notably, in EFAD, the balance between the various EFAs rather than directly the absence of EFAs could impair the exocytosis mechanism. In effect, in enterocyte cultures, certain EFAs, such as EPA, decrease TG esterification and PL transfer from the ER to the Golgi and the mechanisms responsible for these processes have not yet been specified [267]. Similarly, EFAD may affect microtubules [242-268], as we previously mentioned, which could impact on the assembly and secretion of chylomicrons [239,240,242].
V. Conclusion
It is well established that defective digestive processes in CF patients are secondary to pancreatic insufficiency. However, it is difficult to reconcile the failure of appropriate pancreatic enzyme replacement therapy with the persistent fat malabsorption. Since mutations in CFTR result in impaired intracellular pH organelle, glycosylation and sialylation in mammalian cells, it is possible that disturbances in intestinal CF lipid transport may also be associated with cause-related changes in the second step of fat absorption, i.e. the intracellular phase leading to lipolytic product uptake and esterification, apolipoprotein synthesis and processing, and nascent lipoprotein assembly and secretion following the fusion of Golgi vesicles with the basolateral plasma membrane. This hypothesis is reinforced by the findings that CFTR dysfunction alters further intracellular pathways, crucial for lipoprotein packaging and delivery, such as ubiquitin-proteasome complexes, endosomal fusion and vesicular trafficking, as well as Sar1/COP II and ADP-ribosylation factor 1/COP I systems. Additionally, in keeping with substantial growing evidence from the available literature, it is reasonable to put forward that EFAD contributes to CF malabsorption by interfering with intra-enterocyte lipid transport. If previous studies had entirely attributed EFAD to diminished EFA intake and malabsorption, it is likely today that a mutated CFTR may also decrease the incorporation of EFAs in PL [42], increases the release of AA [49], lowers the concentrations of linoleic acid and DHA, and disrupts EFA metabolism [42,50]. An interrelation between CFTR and EFAs is demonstrated when the chloride channels are blocked chemically [41]. Nevertheless, the exact relationship between CFTR and EFAs remains unclear and studies in this fundamental direction would shed considerable light on our understanding of the mechanisms responsible for EFAD in CF. This aspect is particularly important given the suggestions that EFAD is related to a basic defect in FA metabolism. It seems imperative that rigorous and long-term studies be conducted on EFA supplementation to normalize EFA status in CF patients, otherwise EFAD would continue to affect intestinal lipid transport and to simultaneously exacerbate the poor clinical course of the CF patients.
Abbreviations
AA (20:4n-6): Arachidonic Acid
ABCA1: ATP Binding Cassette transporter A1
ACAT: Acyl-coenzyme A:cholesterol acyltransferase
Apo: Apolipoprotein
BA: Biliary Acid
CD36: Cluster Determinant 36
CE: Cholesterol-Ester
CET: Cholesterol esterase
CF: Cystic Fibrosis
CFTR: Cystic Fibrosis Transmembrane Conductance Regulator
DGAT: Diacylglycerol acyltransferase
DHA (22:6n-3): Docosahexaenoic Acid
EFA: Essential Fatty Acid
EFAD: Essential Fatty Acid Deficiency
EPA (20:5n-3): Eicosapentaenoic acid
ER: Endoplasmic Reticulum
FA: Fatty Acid
FABP: Fatty Acid Binding Protein
FFA: Free Fatty Acids
IBAT: Ileal Biliary Acid Transporter
LP: Lipoprotein
Lyso-PC: Lyso-Phosphatidylcholine
MG: Monoacylglycerol
MGAT: Monoacylglycerol Acyltransferase
MTP: Microsomal Triglyceride Transfer Protein
PC: Phosphatidylcholine (lecithin)
PL: Phospholipid
PLA2: Phospholipase A2
PUFA: Polyunsaturated Fatty Acid
SR-BI: Scavenger Receptor class B type I
TG: Triacylglycerol
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
All the authors have contributed to the preparation of this review. The authors thank Schohraya Spahis for her expert technical assistance. This paper was supported by research grants from the Canadian Institutes of Health Research (CIHR) (MT 10584 and MOP 49433)
==== Refs
Smith AE Treatment of cystic fibrosis based on understanding CFTR J Inherit Metab Dis 1995 18 508 516 7494409 10.1007/BF00710062
Anderson MP Rich DP Gregory RJ Smith AE Welsh MJ Generation of cAMP-activated chloride currents by expression of CFTR Science 1991 251 679 682 1704151
Bear CE Li CH Kartner N Bridges RJ Jensen TJ Ramjeesingh M Riordan JR Purification and functional reconstitution of the cystic fibrosis transmembrane conductance regulator (CFTR) Cell 1992 68 809 818 1371239 10.1016/0092-8674(92)90155-6
Drumm ML Pope HA Cliff WH Rommens JM Marvin SA Tsui LC Collins FS Frizzell RA Wilson JM Correction of the cystic fibrosis defect in vitro by retrovirus-mediated gene transfer Cell 1990 62 1227 1233 1698126 10.1016/0092-8674(90)90398-X
Roy CC Weber AM Lepage G Smith L Levy E Digestive and absorptive phase anomalies associated with the exocrine pancreatic insufficiency of cystic fibrosis J Pediatr Gastroenterol Nutr 1988 7 Suppl 1 S1 S7 3042936
Welsh MJ Fick RB Cystic fibrosis J Clin Invest 1987 80 1523 1526 3316277
Carroccio A Pardo F Montalto G Iapichino L Soresi M Averna MR Iacono G Notarbartolo A Use of famotidine in severe exocrine pancreatic insufficiency with persistent maldigestion on enzymatic replacement therapy. A long-term study in cystic fibrosis Dig Dis Sci 1992 37 1441 1446 1505293 10.1007/BF01296016
Couper RT Corey M Durie PR Forstner GG Moore DJ Longitudinal evaluation of serum trypsinogen measurement in pancreatic- insufficient and pancreatic-sufficient patients with cystic fibrosis J Pediatr 1995 127 408 413 7658271
Kalivianakis M Minich DM Bijleveld CM van Aalderen WM Stellaard F Laseur M Vonk RJ Verkade HJ Fat malabsorption in cystic fibrosis patients receiving enzyme replacement therapy is due to impaired intestinal uptake of long-chain fatty acids Am J Clin Nutr 1999 69 127 134 9925134
Kristidis P Bozon D Corey M Markiewicz D Rommens J Tsui LC Durie P Genetic determination of exocrine pancreatic function in cystic fibrosis Am J Hum Genet 1992 50 1178 1184 1376016
Zentler-Munro PL Fine DR Batten JC Northfield TC Effect of cimetidine on enzyme inactivation, bile acid precipitation, and lipid solubilisation in pancreatic steatorrhoea due to cystic fibrosis Gut 1985 26 892 901 3849459
Posner BM Cupples LA Franz MM Gagnon DR Diet and heart disease risk factors in adult American men and women: the Framingham Offspring-Spouse nutrition studies Int J Epidemiol 1993 22 1014 1025 8144282
Carey MC Hernell O Digestion and absorption of fat Semin Gastrointest Dis 1992 3 189 208
Farrell PM Mischler EH Engle MJ Brown DJ Lau SM Fatty acid abnormalities in cystic fibrosis Pediatr Res 1985 19 104 109 3918287
Hubbard VS Dunn GD di Sant'Agnese PA Abnormal fatty-acid composition of plasma-lipids in cystic fibrosis. A primary or a secondary defect? Lancet 1977 2 1302 1304 74723 10.1016/S0140-6736(77)90359-2
Lepage G Levy E Ronco N Smith L Galeano N Roy CC Direct transesterification of plasma fatty acids for the diagnosis of essential fatty acid deficiency in cystic fibrosis J Lipid Res 1989 30 1483 1490 2614252
Parsons HG O'Loughlin EV Forbes D Cooper D Gall DG Supplemental calories improve essential fatty acid deficiency in cystic fibrosis patients Pediatr Res 1988 24 353 356 3211621
Rosenlund ML Kim HK Kritchevsky D Essential fatty acids in cystic fibrosis Nature 1974 251 719 4427669
van Egmond AW Kosorok MR Koscik R Laxova A Farrell PM Effect of linoleic acid intake on growth of infants with cystic fibrosis Am J Clin Nutr 1996 63 746 752 8615359
Lloyd-Still JD Bibus DM Powers CA Johnson SB Holman RT Essential fatty acid deficiency and predisposition to lung disease in cystic fibrosis Acta Paediatr 1996 85 1426 1432 9001653
Benabdeslam H Garcia I Bellon G Gilly R Revol A Biochemical assessment of the nutritional status of cystic fibrosis patients treated with pancreatic enzyme extracts Am J Clin Nutr 1998 67 912 918 9583849
Steinkamp G Demmelmair H Ruhl-Bagheri I von der HH Koletzko B Energy supplements rich in linoleic acid improve body weight and essential fatty acid status of cystic fibrosis patients J Pediatr Gastroenterol Nutr 2000 31 418 423 11045840
Chase HP Dupont J Abnormal levels of prostaglandins and fatty acids in blood of children with cystic fibrosis Lancet 1978 2 236 238 79031 10.1016/S0140-6736(78)91746-4
Roulet M Frascarolo P Rappaz I Pilet M Essential fatty acid deficiency in well nourished young cystic fibrosis patients Eur J Pediatr 1997 156 952 956 9453380 10.1007/s004310050750
Holman RT The ratio of trienoic: tetraenoic acids in tissue lipids as a measure of essential fatty acid requirement J Nutr 1960 70 405 410 14402760
Chan GM Performance of dual-energy x-ray absorptiometry in evaluating bone, lean body mass, and fat in pediatric subjects J Bone Miner Res 1992 7 369 374 1609625
Clandinin MT Zuberbuhler P Brown NE Kielo ES Goh YK Fatty acid pool size in plasma lipoprotein fractions of cystic fibrosis patients Am J Clin Nutr 1995 62 1268 1275 7491891
Lloyd-Still JD Johnson SB Holman RT Essential fatty acid status in cystic fibrosis and the effects of safflower oil supplementation Am J Clin Nutr 1981 34 1 7 7446449
Landon C Kerner JA Castillo R Adams L Whalen R Lewiston NJ Oral correction of essential fatty acid deficiency in cystic fibrosis JPEN J Parenter Enteral Nutr 1981 5 501 504 6801283
Strandvik B Bronnegard M Gilljam H Carlstedt-Duke J Relation between defective regulation of arachidonic acid release and symptoms in cystic fibrosis Scand J Gastroenterol Suppl 1988 143 1 4 2455328
Strandvik B Svensson E Seyberth HW Prostanoid biosynthesis in patients with cystic fibrosis Prostaglandins Leukot Essent Fatty Acids 1996 55 419 425 9014220 10.1016/S0952-3278(96)90125-8
Rogiers V Dab I Michotte Y Vercruysse A Crokaert R Vis HL Abnormal fatty acid turnover in the phospholipids of the red blood cell membranes of cystic fibrosis patients (in vitro study) Pediatr Res 1984 18 704 709 6472942
Levy E Roy C Lacaille F Lambert M Messier M Gavino V Lepage G Thibault L Lipoprotein abnormalities associated with cholesteryl ester transfer activity in cystic fibrosis patients: the role of essential fatty acid deficiency Am J Clin Nutr 1993 57 573 579 8460614
Lepage G Kneepkens CM Roy C Dysfonction peroxysomiale associée à la déficience en acides gras essentiels Médecine/Sciences 1993 9 571 576
Egan M Flotte T Afione S Solow R Zeitlin PL Carter BJ Guggino WB Defective regulation of outwardly rectifying Cl- channels by protein kinase A corrected by insertion of CFTR Nature 1992 358 581 584 1380129 10.1038/358581a0
Guggino WB Outwardly rectifying chloride channels and CF: a divorce and remarriage J Bioenerg Biomembr 1993 25 27 35 7680029 10.1007/BF00768065
Gabriel SE Clarke LL Boucher RC Stutts MJ CFTR and outward rectifying chloride channels are distinct proteins with a regulatory relationship Nature 1993 363 263 268 7683773 10.1038/363263a0
Barasch J Kiss B Prince A Saiman L Gruenert D al Awqati Q Defective acidification of intracellular organelles in cystic fibrosis Nature 1991 352 70 73 1712081 10.1038/352070a0
Bradbury NA Jilling T Berta G Sorscher EJ Bridges RJ Kirk KL Regulation of plasma membrane recycling by CFTR Science 1992 256 530 532 1373908
Jilling T Cunningham S Barker PE Green MW Frizzell RA Kirk KL Genetic complementation in cystic fibrosis pancreatic cells by somatic cell fusion Am J Physiol 1990 259 C1010 C1015 1701965
Kang JX Man SF Brown NE Labrecque PA Clandinin MT The chloride channel blocker anthracene 9-carboxylate inhibits fatty acid incorporation into phospholipid in cultured human airway epithelial cells Biochem J 1992 285 ( Pt 3) 725 729 1323271
Bhura-Bandali FN Suh M Man SF Clandinin MT The deltaF508 mutation in the cystic fibrosis transmembrane conductance regulator alters control of essential fatty acid utilization in epithelial cells J Nutr 2000 130 2870 2875 11110839
Freedman SD Katz MH Parker EM Laposata M Urman MY Alvarez JG A membrane lipid imbalance plays a role in the phenotypic expression of cystic fibrosis in cftr(-/-) mice Proc Natl Acad Sci U S A 1999 96 13995 14000 10570187 10.1073/pnas.96.24.13995
Dombrowsky H Clark GT Rau GA Bernhard W Postle AD Molecular species compositions of lung and pancreas phospholipids in the cftr(tm1HGU/tm1HGU) cystic fibrosis mouse Pediatr Res 2003 53 447 454 12595593 10.1203/01.PDR.0000049937.30305.8A
Carlstedt-Duke J Bronnegard M Strandvik B Pathological regulation of arachidonic acid release in cystic fibrosis: the putative basic defect Proc Natl Acad Sci U S A 1986 83 9202 9206 3097647
Freedman SD Blanco PG Zaman MM Shea JC Ollero M Hopper IK Weed DA Gelrud A Regan MM Laposata M Alvarez JG O'Sullivan BP Association of cystic fibrosis with abnormalities in fatty acid metabolism N Engl J Med 2004 350 560 569 14762183 10.1056/NEJMoa021218
Gilljam H Strandvik B Ellin A Wiman LG Increased mole fraction of arachidonic acid in bronchial phospholipids in patients with cystic fibrosis Scand J Clin Lab Invest 1986 46 511 518 3775237
Stead RJ Muller DP Matthews S Hodson ME Batten JC Effect of abnormal liver function on vitamin E status and supplementation in adults with cystic fibrosis Gut 1986 27 714 718 3721295
Miele L Cordella-Miele E Xing M Frizzell R Mukherjee AB Cystic fibrosis gene mutation (deltaF508) is associated with an intrinsic abnormality in Ca2+-induced arachidonic acid release by epithelial cells DNA Cell Biol 1997 16 749 759 9212168
Keicher U Koletzko B Reinhardt D Omega-3 fatty acids suppress the enhanced production of 5-lipoxygenase products from polymorph neutrophil granulocytes in cystic fibrosis Eur J Clin Invest 1995 25 915 919 8719931
Anderson MP Welsh MJ Fatty acids inhibit apical membrane chloride channels in airway epithelia Proc Natl Acad Sci U S A 1990 87 7334 7338 1698296
Linsdell P Inhibition of cystic fibrosis transmembrane conductance regulator chloride channel currents by arachidonic acid Can J Physiol Pharmacol 2000 78 490 499 10914639 10.1139/cjpp-78-6-490
Barrett KE Bigby TD Involvement of arachidonic acid in the chloride secretory response of intestinal epithelial cells Am J Physiol 1993 264 C446 C452 8447375
Shiota M Kurano Y Mochizuki Y Kimura K Ohta M Sugano T Effects of nerve stimulation and zymosan on glycogenolysis in perfused livers from cold-exposed rats Am J Physiol 1992 263 G353 G359 1415548
Strandvik B Gronowitz E Enlund F Martinsson T Wahlstrom J Essential fatty acid deficiency in relation to genotype in patients with cystic fibrosis J Pediatr 2001 139 650 655 11713441 10.1067/mpd.2001.118890
Henderson WRJ Astley SJ McCready MM Kushmerick P Casey S Becker JW Ramsey BW Oral absorption of omega-3 fatty acids in patients with cystic fibrosis who have pancreatic insufficiency and in healthy control subjects J Pediatr 1994 124 400 408 8120709
Katz DP Manner T Furst P Askanazi J The use of an intravenous fish oil emulsion enriched with omega-3 fatty acids in patients with cystic fibrosis Nutrition 1996 12 334 339 8875517 10.1016/S0899-9007(96)80056-6
Kurlandsky LE Bennink MR Webb PM Ulrich PJ Baer LJ The absorption and effect of dietary supplementation with omega-3 fatty acids on serum leukotriene B4 in patients with cystic fibrosis Pediatr Pulmonol 1994 18 211 217 7838619
Eynard AR Cejas V Silva R Quiroga P Munoz S Histopathology of essential fatty acid-deficient mice Nutrition 1992 8 37 40 1562787
Harper TB Chase HP Henson J Henson PM Essential fatty acid deficiency in the rabbit as a model of nutritional impairment in cystic fibrosis. In vitro and in vivo effects on lung defense mechanisms Am Rev Respir Dis 1982 126 540 547 7125341
Galabert C Filliat M Chazalette JP Fatty-acid composition of serum-lecithins in cystic-fibrosis patients without steatorrhoea Lancet 1978 2 903 81456 10.1016/S0140-6736(78)91620-3
Akpan JO Hurley MC Lands WE Insulin and glucagon secretion in essential fatty acid deficient rats Acta Diabetol Lat 1981 18 147 156 7018148
Daveloose D Linard A Arfi T Viret J Christon R Simultaneous changes in lipid composition, fluidity and enzyme activity in piglet intestinal brush border membrane as affected by dietary polyunsaturated fatty acid deficiency Biochim Biophys Acta 1993 1166 229 237 8443241
Gaggiotti G Taus M Spazzafumo L Tesei M La Rocca R Mazzanti L Modifications of functional and physico-chemical properties of rat ileal plasma membranes Biochem Mol Biol Int 1995 35 851 854 7627134
Kaur M Kaur J Ojha S Mahmood A Dietary fat effects on brush border membrane composition and enzyme activities in rat intestine Ann Nutr Metab 1996 40 269 276 9001687
Lucero HA Robbins PW Lipid rafts-protein association and the regulation of protein activity Arch Biochem Biophys 2004 426 208 224 15158671 10.1016/j.abb.2004.03.020
Kowalski MP Pier GB Localization of cystic fibrosis transmembrane conductance regulator to lipid rafts of epithelial cells is required for Pseudomonas aeruginosa-induced cellular activation J Immunol 2004 172 418 425 14688350
Ramsey BW Farrell PM Pencharz P Nutritional assessment and management in cystic fibrosis: a consensus report. The Consensus Committee Am J Clin Nutr 1992 55 108 116 1728810
Kraemer R Rudeberg A Hadorn B Rossi E Relative underweight in cystic fibrosis and its prognostic value Acta Paediatr Scand 1978 67 33 37 626067
SPROUL A Huang N GROWTH PATTERNS IN CHILDREN WITH CYSTIC FIBROSIS J Pediatr 1964 65 664 676 14221167
Gaskin K Gurwitz D Durie P Corey M Levison H Forstner G Improved respiratory prognosis in patients with cystic fibrosis with normal fat absorption J Pediatr 1982 100 857 862 7086584
Corey M McLaughlin FJ Williams M Levison H A comparison of survival, growth, and pulmonary function in patients with cystic fibrosis in Boston and Toronto J Clin Epidemiol 1988 41 583 591 3260274 10.1016/0895-4356(88)90063-7
Dodge JA Custance JM Goodchild MC Laing SC Vaughan M Paradoxical effects of essential fatty acid supplementation on lipid profiles and sweat electrolytes in cystic fibrosis Br J Nutr 1990 63 259 271 2185826
Hubbard VS McKenna MC Absorption of safflower oil and structured lipid preparations in patients with cystic fibrosis Lipids 1987 22 424 428 3613873
Christophe A Robberecht E Franckx H De Baets F van de PM Effect of administration of gamma-linolenic acid on the fatty acid composition of serum phospholipids and cholesteryl esters in patients with cystic fibrosis Ann Nutr Metab 1994 38 40 47 8185286
Hamosh M Scanlon JW Ganot D Likel M Scanlon KB Hamosh P Fat digestion in the newborn. Characterization of lipase in gastric aspirates of premature and term infants J Clin Invest 1981 67 838 846 7204558
Roberts IM Montgomery RK Carey MC Rat lingual lipase: partial purification, hydrolytic properties, and comparison with pancreatic lipase Am J Physiol 1984 247 G385 G393 6496677
Ville E Carriere F Renou C Laugier R Physiological study of pH stability and sensitivity to pepsin of human gastric lipase Digestion 2002 65 73 81 12021480 10.1159/000057708
Cohen M Morgan RG Hofmann AF Lipolytic activity of human gastric and duodenal juice against medium and long chain triglycerides Gastroenterology 1971 60 1 15 5544089
Liao TH Hamosh P Hamosh M Fat digestion by lingual lipase: mechanism of lipolysis in the stomach and upper small intestine Pediatr Res 1984 18 402 409 6728567
Carriere F Barrowman JA Verger R Laugier R Secretion and contribution to lipolysis of gastric and pancreatic lipases during a test meal in humans Gastroenterology 1993 105 876 888 8359655
Armand M Pasquier B Andre M Borel P Senft M Peyrot J Salducci J Portugal H Jaussan V Lairon D Digestion and absorption of 2 fat emulsions with different droplet sizes in the human digestive tract Am J Clin Nutr 1999 70 1096 1106 10584056
Armand M Borel P Dubois C Senft M Peyrot J Salducci J Lafont H Lairon D Characterization of emulsions and lipolysis of dietary lipids in the human stomach Am J Physiol 1994 266 G372 G381 8166277
Carey MC Small DM Bliss CM Lipid digestion and absorption Annu Rev Physiol 1983 45 651 677 6342528 10.1146/annurev.ph.45.030183.003251
Patton JS Johnson LR Gastrointestinal lipid digestion Physiology of the gastrointestinal tract 1981 New York, Raven Press 1123
Fredrikzon B Blackberg L Lingual lipase: an important lipase in the digestion of dietary lipids in cystic fibrosis? Pediatr Res 1980 14 1387 1390 7208157
Moreau H Sauniere JF Gargouri Y Pieroni G Verger R Sarles H Human gastric lipase: variations induced by gastrointestinal hormones and by pathology Scand J Gastroenterol 1988 23 1044 1048 3247586
Hamosh M A review. Fat digestion in the newborn: role of lingual lipase and preduodenal digestion Pediatr Res 1979 13 615 622 382069
Roulet M Weber AM Paradis Y Roy CC Chartrand L Lasalle R Morin CL Gastric emptying and lingual lipase activity in cystic fibrosis Pediatr Res 1980 14 1360 1362 7208153
Abrams CK Hamosh M Hubbard VS Dutta SK Hamosh P Lingual lipase in cystic fibrosis. Quantitation of enzyme activity in the upper small intestine of patients with exocrine pancreatic insufficiency J Clin Invest 1984 73 374 382 6699170
DiMagno EP Malagelada JR Go VL Moertel CG Fate of orally ingested enzymes in pancreatic insufficiency. Comparison of two dosage schedules N Engl J Med 1977 296 1318 1322 16213
Meyer JH The ins and outs of oral pancreatic enzymes N Engl J Med 1977 296 1347 1348 859534
Roy CC Weber AM Morin CL Combes JC Nussle D Megevand A Lasalle R Abnormal biliary lipid composition in cystic fibrosis. Effect of pancreatic enzymes N Engl J Med 1977 297 1301 1305 917086
Weber AM Roy CC Morin CL Lasalle R Malabsorption of bile acids in children with cystic fibrosis N Engl J Med 1973 289 1001 1005 4742200
Hamosh M Lingual and gastric lipases Nutrition 1990 6 421 428 2134569
Zentler-Munro PL Assoufi BA Balasubramanian K Cornell S Benoliel D Northfield TC Hodson ME Therapeutic potential and clinical efficacy of acid-resistant fungal lipase in the treatment of pancreatic steatorrhoea due to cystic fibrosis Pancreas 1992 7 311 319 1594552
DiMagno EP Go VL Summerskill WH Relations between pancreatic enzyme ouputs and malabsorption in severe pancreatic insufficiency N Engl J Med 1973 288 813 815 4693931
Waters DL Dorney SF Gaskin KJ Gruca MA O'Halloran M Wilcken B Pancreatic function in infants identified as having cystic fibrosis in a neonatal screening program N Engl J Med 1990 322 303 308 2296272
Durie PR The pathophysiology of the pancreatic defect in cystic fibrosis Acta Paediatr Scand Suppl 1989 363 41 44 2701923
Cohen JC Morrow SL Cork RJ Delcarpio JB Larson JE Molecular pathophysiology of cystic fibrosis based on the rescued knockout mouse model Mol Genet Metab 1998 64 108 118 9705235 10.1006/mgme.1998.2683
Larson JE Delcarpio JB Farberman MM Morrow SL Cohen JC CFTR modulates lung secretory cell proliferation and differentiation Am J Physiol Lung Cell Mol Physiol 2000 279 L333 L341 10926557
Hyde K Reid CJ Tebbutt SJ Weide L Hollingsworth MA Harris A The cystic fibrosis transmembrane conductance regulator as a marker of human pancreatic duct development Gastroenterology 1997 113 914 919 9287984
Gaskin KJ Durie PR Lee L Hill R Forstner GG Colipase and lipase secretion in childhood-onset pancreatic insufficiency. Delineation of patients with steatorrhea secondary to relative colipase deficiency Gastroenterology 1984 86 1 7 6689652
Blackberg L Hernell O Bengtsson G Olivecrona T Colipase enhances hydrolysis of dietary triglycerides in the absence of bile salts J Clin Invest 1979 64 1303 1308 500812
Borgstrom B On the interactions between pancreatic lipase and colipase and the substrate, and the importance of bile salts J Lipid Res 1975 16 411 417 446
Levy E Goldstein R Freier S Shafrir E Characterization of gastric lipolytic activity Biochim Biophys Acta 1981 664 316 326 7248327
Spalinger JH Seidman EG Lepage G Menard D Gavino V Levy E Uptake and metabolism of structured triglyceride by Caco-2 cells: reversal of essential fatty acid deficiency Am J Physiol 1998 275 G652 G659 9756493
Ikeda I Tomari Y Sugano M Watanabe S Nagata J Lymphatic absorption of structured glycerolipids containing medium-chain fatty acids and linoleic acid, and their effect on cholesterol absorption in rats Lipids 1991 26 369 373 1895884
Brink EJ Haddeman E de Fouw NJ Weststrate JA Positional distribution of stearic acid and oleic acid in a triacylglycerol and dietary calcium concentration determines the apparent absorption of these fatty acids in rats J Nutr 1995 125 2379 2387 7666256
Christensen MS Mullertz A Hoy CE Absorption of triglycerides with defined or random structure by rats with biliary and pancreatic diversion Lipids 1995 30 521 526 7651079
Lepage G Yesair DW Ronco N Champagne J Bureau N Chemtob S Berube D Roy CC Effect of an organized lipid matrix on lipid absorption and clinical outcomes in patients with cystic fibrosis J Pediatr 2002 141 178 185 12183711 10.1067/mpd.2002.124305
Levy E Garofalo C Thibault L Dionne S Daoust L Lepage G Roy CC Intraluminal and intracellular phases of fat absorption are impaired in essential fatty acid deficiency Am J Physiol 1992 262 G319 G326 1539663
Werner A Minich DM Havinga R Bloks V Van Goor H Kuipers F Verkade HJ Fat malabsorption in essential fatty acid-deficient mice is not due to impaired bile formation Am J Physiol Gastrointest Liver Physiol 2002 283 G900 G908 12223350
Hjelte L Ahren B Andren-Sandberg A Bottcher G Strandvik B Pancreatic function in the essential fatty acid deficient rat Metabolism 1990 39 871 875 1695988 10.1016/0026-0495(90)90135-Y
Hjelte L Strandvik B Muller RM Sagstrom S Roomans GM Essential fatty acid deficient rats in the study of cystic fibrosis: an X-ray microanalytical and ultrastructural study in chronically reserpinized rats J Submicrosc Cytol Pathol 1990 22 409 414 2390763
Duan RD Borgstrom B Is there a specific lysophospholipase in human pancreatic juice? Biochim Biophys Acta 1993 1167 326 330 8481395
Borgstrom B Importance of phospholipids, pancreatic phospholipase A2, and fatty acid for the digestion of dietary fat: in vitro experiments with the porcine enzymes Gastroenterology 1980 78 954 962 7380202
Nouri-Sorkhabi MH Chapman BE Kuchel PW Gruca MA Gaskin KJ Parallel secretion of pancreatic phospholipase A(2), phospholipase A(1), lipase, and colipase in children with exocrine pancreatic dysfunction Pediatr Res 2000 48 735 740 11102539
Weber AM Roy CC Intraduodenal events in cystic fibrosis J Pediatr Gastroenterol Nutr 1984 3 Suppl 1 S113 S119 6502389
O'Connor PJ Loiudice TA Bochenek W Rodgers JB Effect of diester and diether phosphatidylcholine on intestinal absorption of neutral and acidic sterols Am J Dig Dis 1978 23 316 320 665626 10.1007/BF01072413
Saunders DR Sillery J Lecithin inhibits fatty acid and bile salt absorption from rat small intestine in vivo Lipids 1976 11 830 832 1011937
Mathias PM Harries JT Muller DP Optimization and validation of assays to estimate pancreatic esterase activity using well-characterized micellar solutions of cholesteryl oleate and tocopheryl acetate J Lipid Res 1981 22 177 184 7217783
Clark SB Chylomicron composition during duodenal triglyceride and lecithin infusion Am J Physiol 1978 235 E183 E190 686165
Nilsson A Borgstrom B Absorption and metabolism of lecithin and lysolecithin by intestinal slices Biochim Biophys Acta 1967 137 240 254 6051560
Tso P Lam J Simmonds WJ The importance of the lysophosphatidylcholine and choline moiety of bile phosphatidylcholine in lymphatic transport of fat Biochim Biophys Acta 1978 528 364 372 638162
Petit-Thevenin J Bruneau N Nganga A Lombardo D Verine A Effects of arachidonic and docosahexaenoic acids on secretion and degradation of bile salt-dependent lipase in AR4-2J cells J Lipid Res 2001 42 1220 1230 11483623
Sinaasappel M Bouquet J Neijens HJ Problems in the treatment of malabsorption in CF Acta Paediatr Scand Suppl 1985 317 22 27 3864358
Clarke LL Harline MC Gawenis LR Walker NM Turner JT Weisman GA Extracellular UTP stimulates electrogenic bicarbonate secretion across CFTR knockout gallbladder epithelium Am J Physiol Gastrointest Liver Physiol 2000 279 G132 G138 10898755
Seidler U Blumenstein I Kretz A Viellard-Baron D Rossmann H Colledge WH Evans M Ratcliff R Gregor M A functional CFTR protein is required for mouse intestinal cAMP-, cGMP- and Ca(2+)-dependent J Physiol 1997 505 ( Pt 2) 411 423 9423183 10.1111/j.1469-7793.1997.411bb.x
Robinson PJ Smith AL Sly PD Duodenal pH in cystic fibrosis and its relationship to fat malabsorption Dig Dis Sci 1990 35 1299 1304 2120019 10.1007/BF01536423
Regan PT Malagelada JR DiMagno EP Go VL Reduced intraluminal bile acid concentrations and fat maldigestion in pancreatic insufficiency: correction by treatment Gastroenterology 1979 77 285 289 447041
Ng SM Jones AP Wiley JS Drug Therapies for Reducing Gastric Acidity in People with Cystic Fibrosis 2004 Chichester, UK 1 21
Westergaard H Dietschy JM The mechanism whereby bile acid micelles increase the rate of fatty acid and cholesterol uptake into the intestinal mucosal cell J Clin Invest 1976 58 97 108 932213
Wilson FA Sallee VL Dietschy JM Unstirred water layers in intestine: rate determinant of fatty acid absorption from micellar solutions Science 1971 174 1031 1033 5120088
Kobayashi A Ohbe Y Yonekubo A Fat absorption in patients with surgically repaired biliary atresia Helv Paediatr Acta 1983 38 307 314 6654680
Minich DM Kalivianakis M Havinga R Van Goor H Stellaard F Vonk RJ Kuipers F Verkade HJ Bile diversion in rats leads to a decreased plasma concentration of linoleic acid which is not due to decreased net intestinal absorption of dietary linoleic acid Biochim Biophys Acta 1999 1438 111 119 10216285
Ros E Garcia-Puges A Reixach M Cuso E Rodes J Fat digestion and exocrine pancreatic function in primary biliary cirrhosis Gastroenterology 1984 87 180 187 6724261
Narayanan VS Storch J Fatty acid transfer in taurodeoxycholate mixed micelles Biochemistry 1996 35 7466 7473 8652524 10.1021/bi952979k
Harries JT Muller DP McCollum JP Lipson A Roma E Norman AP Intestinal bile salts in cystic fibrosis: studies in the patient and experimental animal Arch Dis Child 1979 54 19 24 420518
Robb TA Davidson GP Kirubakaran C Conjugated bile acids in serum and secretions in response to cholecystokinin/secretin stimulation in children with cystic fibrosis Gut 1985 26 1246 1256 4065698
Weizman Z Durie PR Kopelman HR Vesely SM Forstner GG Bile acid secretion in cystic fibrosis: evidence for a defect unrelated to fat malabsorption Gut 1986 27 1043 1048 3758817
Roy CC Weber AM Morin CL Lepage G Brisson G Yousef I Lasalle R Hepatobiliary disease in cystic fibrosis: a survey of current issues and concepts J Pediatr Gastroenterol Nutr 1982 1 469 478 7186061
Knox R Stein I Levinson D Tso P Mansbach CM Effect of fat pre-feeding on bile flow and composition in the rat Biochim Biophys Acta 1991 1083 65 70 2031939
Connor WE Witiak DT Stone DB Armstrong ML Cholesterol balance and fecal neutral steroid and bile acid excretion in normal men fed dietary fats of different fatty acid composition J Clin Invest 1969 48 1363 1375 5796351
Smit MJ Temmerman AM Wolters H Kuipers F Beynen AC Vonk RJ Dietary fish oil-induced changes in intrahepatic cholesterol transport and bile acid synthesis in rats J Clin Invest 1991 88 943 951 1885779
Levy E The 1991 Borden Award Lecture. Selected aspects of intraluminal and intracellular phases of intestinal fat absorption Can J Physiol Pharmacol 1992 70 413 419 1498710
Levy E Garofalo C Rouleau T Gavino V Bendayan M Impact of essential fatty acid deficiency on hepatic sterol metabolism in rats Hepatology 1996 23 848 857 8666341
Ekelund K Johansson C Nylander B Effects of 16,16 dimethyl prostaglandin E2 on food-stimulated pancreatic secretion and output of bile in man Scand J Gastroenterol 1977 12 457 460 329405
Cohen BI Raicht RF Mosbach EH Sterol metabolism studies in the rat. Effects of primary bile acids (sodium taurochenodeoxycholate and sodium taurocholate) on sterol metabolism J Lipid Res 1977 18 223 231 845503
Gallo-Torres HE Miller ON Hamilton JG Further studies on the role of bile salts in cholesterol esterification and absorption from the gut Arch Biochem Biophys 1971 143 22 36 5561745 10.1016/0003-9861(71)90182-2
Reynier MO Montet JC Gerolami A Marteau C Crotte C Montet AM Mathieu S Comparative effects of cholic, chenodeoxycholic, and ursodeoxycholic acids on micellar solubilization and intestinal absorption of cholesterol J Lipid Res 1981 22 467 473 7240971
Thompson GN Excessive fecal taurine loss predisposes to taurine deficiency in cystic fibrosis J Pediatr Gastroenterol Nutr 1988 7 214 219 3351706
De Curtis M Santamaria F Ercolini P Vittoria L De Ritis G Garofalo V Ciccimarra F Effect of taurine supplementation on fat and energy absorption in cystic fibrosis Arch Dis Child 1992 67 1082 1085 1417050
Merli M Bertasi S Servi R Diamanti S Martino F De Santis A Goffredo F Quattrucci S Antonelli M Angelico M Effect of a medium dose of ursodeoxycholic acid with or without taurine supplementation on the nutritional status of patients with cystic fibrosis: a randomized, placebo-controlled, crossover trial J Pediatr Gastroenterol Nutr 1994 19 198 203 7815243
Smith LJ Lacaille F Lepage G Ronco N Lamarre A Roy CC Taurine decreases fecal fatty acid and sterol excretion in cystic fibrosis. A randomized double-blind trial Am J Dis Child 1991 145 1401 1404 1669669
Thompson GN Failure of taurine to improve fat absorption in cystic fibrosis J Inherit Metab Dis 1988 11 Suppl 2 158 160 3141698
Belli DC Levy E Darling P Leroy C Lepage G Giguere R Roy CC Taurine improves the absorption of a fat meal in patients with cystic fibrosis Pediatrics 1987 80 517 523 3658570
Carrasco S Codoceo R Prieto G Lama R Polanco I Effect of taurine supplements on growth, fat absorption and bile acid on cystic fibrosis Acta Univ Carol [Med ] (Praha) 1990 36 152 156 2130680
Colombo C Setchell KD Podda M Crosignani A Roda A Curcio L Ronchi M Giunta A Effects of ursodeoxycholic acid therapy for liver disease associated with cystic fibrosis J Pediatr 1990 117 482 489 2391610
Sarfeh IJ Beeler DA Treble DH Balint JA Studies of the hepatic excretory defects in essential fatty acid deficiency. Their possible relationship to the genesis of cholesterol gallstones J Clin Invest 1974 53 423 430 11344556
Trewhella MA Collins FD A comparison of the relative turnover of individual molecular species of phospholipids in normal rats and in rats deficient in essential fatty acids Biochim Biophys Acta 1973 296 34 50 4693511
Lepage G Paradis K Lacaille F Senechal L Ronco N Champagne J Lenaerts C Roy CC Rasquin-Weber A Ursodeoxycholic acid improves the hepatic metabolism of essential fatty acids and retinol in children with cystic fibrosis J Pediatr 1997 130 52 58 9003851
Hofmann AF The enterohepatic circulation of bile acids in man Adv Intern Med 1976 21 501 534 766594
Schreiber AJ Simon FR Overview of clinical aspects of bile salt physiology J Pediatr Gastroenterol Nutr 1983 2 337 345 6348231
Goodchild MC Murphy GM Howell AM Nutter SA Anderson CM Aspects of bile acid metabolism in cystic fibrosis Arch Dis Child 1975 50 769 778 1236566
O'Brien S Mulcahy H Fenlon H O'Broin A Casey M Burke A FitzGerald MX Hegarty JE Intestinal bile acid malabsorption in cystic fibrosis Gut 1993 34 1137 1141 8174969
Smalley CA Brown GA Parkes ME Tease H Brookes V Anderson CM Reduction of bile acid loss in cystic fibrosis by dietary means Arch Dis Child 1978 53 477 482 686773
Walters MP Littlewood JM Faecal bile acid and dietary residue excretion in cystic fibrosis: age group variations J Pediatr Gastroenterol Nutr 1998 27 296 300 9740200 10.1097/00005176-199809000-00005
Watkins JB Tercyak AM Szczepanik P Klein PD Bile salt kinetics in cystic fibrosis: influence of pancreatic enzyme replacement Gastroenterology 1977 73 1023 1028 332576
Vlahcevic ZR Miller JR Farrar JT Swell L Kinetics and pool size of primary bile acids in man Gastroenterology 1971 61 85 90 5088530
Weber AM Roy C Lepage G Chartrand L Lasalle R Interruption of the enterohepatic circulation of bile acids in cystic fibrosis Gastroenterology 1975 68 1066
Jonas A Diver-Haber A Bile acid sequestration by the solid phase of stools in cystic fibrosis patients. Role of pancreatic enzymes Dig Dis Sci 1988 33 724 731 3371142 10.1007/BF01540437
Isenberg JN Hendrix PY Cox KL Effect of short-term cimetidine administration on fecal bile acid losses in cystic fibrosis J Pediatr Gastroenterol Nutr 1983 2 447 451 6620052
Lefebvre D Ratelle S Chartrand L Roy CC Reduced microbial transformation of bile acids in cystic fibrosis Experientia 1977 33 616 618 862793
de Rooij FW van den Berg JW Sinaasappel M Bosman-Jacobs EP Touw-Blommesteijn AC Bile acid malabsorption in cystic fibrosis; membrane vesicles, a tool for revealing the role of the ileal brush border membrane Acta Paediatr Scand Suppl 1985 317 28 30 3864359
Fondacaro JD Heubi JE Kellogg FW Intestinal bile acid malabsorption in cystic fibrosis: a primary mucosal cell defect Pediatr Res 1982 16 494 498 7099766
Stelzner M Somasundaram S Lee SP Kuver R Ileal mucosal bile acid absorption is increased in Cftr knockout mice BMC Gastroenterol 2001 1 10 11696242 10.1186/1471-230X-1-10
Levy E Goldstein R Stankievicz H Hager E Freier S Gastric handling of medium-chain triglycerides and subsequent metabolism in the suckling rat J Pediatr Gastroenterol Nutr 1984 3 784 789 6502379
Symonds EL Omari TI Webster JM Davidson GP Butler RN Relation between pancreatic lipase activity and gastric emptying rate in children with cystic fibrosis J Pediatr 2003 143 772 775 14657826 10.1067/S0022-3476(03)00581-X
Bali A Stableforth DE Asquith P Prolonged small-intestinal transit time in cystic fibrosis Br Med J (Clin Res Ed) 1983 287 1011 1013 6412927
Dalzell AM Freestone NS Billington D Heaf DP Small intestinal permeability and orocaecal transit time in cystic fibrosis Arch Dis Child 1990 65 585 588 2116115
Mack DR Flick JA Durie PR Rosenstein BJ Ellis LE Perman JA Correlation of intestinal lactulose permeability with exocrine pancreatic dysfunction J Pediatr 1992 120 696 701 1578303
Seal S McClean P Walters M Wolfe SP Harding M Coward W Littlewood JM Stable isotope studies of pancreatic enzyme release in vivo Postgrad Med J 1996 72 Suppl 2 S37 S38 8869181
Taylor CJ Hillel PG Ghosal S Frier M Senior S Tindale WB Read N Gastric emptying and intestinal transit of pancreatic enzyme supplements in cystic fibrosis Arch Dis Child 1999 80 149 152 10325730
Hjelte L Melin T Nilsson A Strandvik B Absorption and metabolism of [3H]arachidonic and [14C]linoleic acid in essential fatty acid-deficient rats Am J Physiol 1990 259 G116 G124 2115302
Snipes RL Cellular dynamics in the jejunum of essential fatty acid deficient mice Anat Rec 1967 159 421 429 5586291 10.1002/ar.1091590411
FREYE HB KURTZ SM SPOCK A CAPP MP LIGHT AND ELECTRON MICROSCOPIC EXAMINATION OF THE SMALL BOWEL OF CHILDREN WITH CYSTIC FIBROSIS J Pediatr 1964 64 575 579 14144756
Kent G Oliver M Foskett JK Frndova H Durie P Forstner J Forstner GG Riordan JR Percy D Buchwald M Phenotypic abnormalities in long-term surviving cystic fibrosis mice Pediatr Res 1996 40 233 241 8827771
Cox KL Isenberg JN Ament ME Gastric acid hypersecretion in cystic fibrosis J Pediatr Gastroenterol Nutr 1982 1 559 565 7186068
Van Biervliet S Eggermont E Marien P Hoffman I Veereman G Combined impact of mucosal damage and of cystic fibrosis on the small intestinal brush border enzyme activities Acta Clin Belg 2003 58 220 224 14635529
Slomiany A Liau YH Carter SR Newman LJ Slomiany BL Mucus glycoprotein fatty acyltransferase in patients with cystic fibrosis: effect on the glycoprotein viscosity Biochem Biophys Res Commun 1985 132 299 306 4062931 10.1016/0006-291X(85)91022-8
Schutt WH Isles TE Protein in meconium from meconium ileus Arch Dis Child 1968 43 178 181 5689531
Wesley A Forstner J Qureshi R Mantle M Forstner G Human intestinal mucin in cystic fibrosis Pediatr Res 1983 17 65 69 6835717
Korotkova M Strandvik B Essential fatty acid deficiency affects the fatty acid composition of the rat small intestinal and colonic mucosa differently Biochim Biophys Acta 2000 1487 319 325 11018483
Snipes RL The effects of essential fatty acid deficiency on the ultrastructure and functional capacity of the jejunal epithelium Lab Invest 1968 18 179 189 5644532
AAES-JORGENSEN E Essential fatty acids Physiol Rev 1961 41 1 51 13680857
Pandalai PK Pilat MJ Yamazaki K Naik H Pienta KJ The effects of omega-3 and omega-6 fatty acids on in vitro prostate cancer growth Anticancer Res 1996 16 815 820 8687134
Jiang WG Eynard AR Mansel RE The pathology of essential fatty acid deficiency: is it cell adhesion mediated? Med Hypotheses 2000 55 257 262 10985920 10.1054/mehy.1998.1191
Juby LD Rothwell J Axon AT Cellobiose/mannitol sugar test--a sensitive tubeless test for coeliac disease: results on 1010 unselected patients Gut 1989 30 476 480 2497056
Hallberg K Grzegorczyk A Larson G Strandvik B Intestinal permeability in cystic fibrosis in relation to genotype J Pediatr Gastroenterol Nutr 1997 25 290 295 9285379 10.1097/00005176-199709000-00008
Gosden CM Gosden JR Fetal abnormalities in cystic fibrosis suggest a deficiency in proteolysis of cholecystokinin Lancet 1984 2 541 546 6147603 10.1016/S0140-6736(84)90765-7
Eynard AR Monis B Kalinec F Leguizamon RO Increased proliferation of the epithelium of the proximal alimentary tract of EFA-deficient rats: a light and electron microscopy study Exp Mol Pathol 1982 36 135 143 7060719 10.1016/0014-4800(82)90088-0
Thomson AB Keelan M Garg ML Clandinin MT Intestinal aspects of lipid absorption: in review Can J Physiol Pharmacol 1989 67 179 191 2663123
Chow SL Hollander D A dual, concentration-dependent absorption mechanism of linoleic acid by rat jejunum in vitro J Lipid Res 1979 20 349 356 36439
Trotter PJ Ho SY Storch J Fatty acid uptake by Caco-2 human intestinal cells J Lipid Res 1996 37 336 346 9026531
Baier LJ Bogardus C Sacchettini JC A polymorphism in the human intestinal fatty acid binding protein alters fatty acid transport across Caco-2 cells J Biol Chem 1996 271 10892 10896 8631905 10.1074/jbc.271.20.11689
Clarke SD Armstrong MK Cellular lipid binding proteins: expression, function, and nutritional regulation FASEB J 1989 3 2480 2487 2680704
Ockner RK Manning JA Fatty acid-binding protein in small intestine. Identification, isolation, and evidence for its role in cellular fatty acid transport J Clin Invest 1974 54 326 338 4211161
Sfeir Z Ibrahimi A Amri E Grimaldi P Abumrad N Regulation of FAT/CD36 gene expression: further evidence in support of a role of the protein in fatty acid binding/transport Prostaglandins Leukot Essent Fatty Acids 1997 57 17 21 9250603 10.1016/S0952-3278(97)90487-7
Stahl A Hirsch DJ Gimeno RE Punreddy S Ge P Watson N Patel S Kotler M Raimondi A Tartaglia LA Lodish HF Identification of the major intestinal fatty acid transport protein Mol Cell 1999 4 299 308 10518211 10.1016/S1097-2765(00)80332-9
Lobo MV Huerta L Ruiz-Velasco N Teixeiro E de la CP Celdran A Martin-Hidalgo A Vega MA Bragado R Localization of the lipid receptors CD36 and CLA-1/SR-BI in the human gastrointestinal tract: towards the identification of receptors mediating the intestinal absorption of dietary lipids J Histochem Cytochem 2001 49 1253 1260 11561009
Levy E Menard D Suc I Delvin E Marcil V Brissette L Thibault L Bendayan M Ontogeny, immunolocalisation, distribution and function of SR-BI in the human intestine J Cell Sci 2004 117 327 337 14676281 10.1242/jcs.00856
Berge KE Tian H Graf GA Yu L Grishin NV Schultz J Kwiterovich P Shan B Barnes R Hobbs HH Accumulation of dietary cholesterol in sitosterolemia caused by mutations in adjacent ABC transporters Science 2000 290 1771 1775 11099417 10.1126/science.290.5497.1771
Lee MH Lu K Hazard S Yu H Shulenin S Hidaka H Kojima H Allikmets R Sakuma N Pegoraro R Srivastava AK Salen G Dean M Patel SB Identification of a gene, ABCG5, important in the regulation of dietary cholesterol absorption Nat Genet 2001 27 79 83 11138003 10.1038/87170
Repa JJ Turley SD Lobaccaro JA Medina J Li L Lustig K Shan B Heyman RA Dietschy JM Mangelsdorf DJ Regulation of absorption and ABC1-mediated efflux of cholesterol by RXR heterodimers Science 2000 289 1524 1529 10968783 10.1126/science.289.5484.1524
Mulligan JD Flowers MT Tebon A Bitgood JJ Wellington C Hayden MR Attie AD ABCA1 is essential for efficient basolateral cholesterol efflux during the absorption of dietary cholesterol in chickens J Biol Chem 2003 278 13356 13366 12551945 10.1074/jbc.M212377200
Murthy S Born E Mathur SN Field FJ LXR/RXR activation enhances basolateral efflux of cholesterol in CaCo-2 cells J Lipid Res 2002 43 1054 1064 12091489 10.1194/jlr.M100358-JLR200
Iyer SP Yao X Crona JH Hoos LM Tetzloff G Davis HRJ Graziano MP Altmann SW Characterization of the putative native and recombinant rat sterol transporter Niemann-Pick C1 Like 1 (NPC1L1) protein Biochim Biophys Acta 2005
Klett EL Patel SB Biomedicine. Will the real cholesterol transporter please stand up Science 2004 303 1149 1150 14976303 10.1126/science.1095519
Briel M Greger R Kunzelmann K Cl- transport by cystic fibrosis transmembrane conductance regulator (CFTR) contributes to the inhibition of epithelial Na+ channels (ENaCs) in Xenopus oocytes co-expressing CFTR and ENaC J Physiol 1998 508 ( Pt 3) 825 836 9518736 10.1111/j.1469-7793.1998.825bp.x
Mall M Bleich M Kuehr J Brandis M Greger R Kunzelmann K CFTR-mediated inhibition of epithelial Na+ conductance in human colon is defective in cystic fibrosis Am J Physiol 1999 277 G709 G716 10484398
Xu Y Clark JC Aronow BJ Dey CR Liu C Wooldridge JL Whitsett JA Transcriptional adaptation to cystic fibrosis transmembrane conductance regulator deficiency J Biol Chem 2003 278 7674 7682 12482874 10.1074/jbc.M210277200
McKenna MC Hubbard VS Bieri JG Linoleic acid absorption from lipid supplements in patients with cystic fibrosis with pancreatic insufficiency and in control subjects J Pediatr Gastroenterol Nutr 1985 4 45 51 3981368
Thompson GN Relationships between essential fatty acid levels, pulmonary function and fat absorption in pre-adolescent cystic fibrosis children with good clinical scores Eur J Pediatr 1989 148 327 329 2707278 10.1007/BF00444126
Clark SB Ekkers TE Singh A Balint JA Holt PR Rodgers JBJ Fat absorption in essential fatty acid deficiency: a model experimental approach to studies of the mechanism of fat malabsorption of unknown etiology J Lipid Res 1973 14 581 588 4729974
Kellner-Weibel G Llera-Moya M Connelly MA Stoudt G Christian AE Haynes MP Williams DL Rothblat GH Expression of scavenger receptor BI in COS-7 cells alters cholesterol content and distribution Biochemistry 2000 39 221 229 10625497 10.1021/bi991666c
Amri EZ Bonino F Ailhaud G Abumrad NA Grimaldi PA Cloning of a protein that mediates transcriptional effects of fatty acids in preadipocytes. Homology to peroxisome proliferator-activated receptors J Biol Chem 1995 270 2367 2371 7836471 10.1074/jbc.270.5.2367
Murthy S Born E Mathur SN Field FJ Liver-X-receptor-mediated increase in ATP-binding cassette transporter A1 expression is attenuated by fatty acids in CaCo-2 cells: effect on cholesterol efflux to high-density lipoprotein Biochem J 2004 377 545 552 14604434
Uehara Y Engel T Li Z Goepfert C Rust S Zhou X Langer C Schachtrup C Wiekowski J Lorkowski S Assmann G von Eckardstein A Polyunsaturated fatty acids and acetoacetate downregulate the expression of the ATP-binding cassette transporter A1 Diabetes 2002 51 2922 2928 12351428
Wang Y Oram JF Unsaturated fatty acids inhibit cholesterol efflux from macrophages by increasing degradation of ATP-binding cassette transporter A1 J Biol Chem 2002 277 5692 5697 11741998 10.1074/jbc.M109977200
Heda GD Marino CR Surface expression of the cystic fibrosis transmembrane conductance regulator mutant DeltaF508 is markedly upregulated by combination treatment with sodium butyrate and low temperature Biochem Biophys Res Commun 2000 271 659 664 10814518 10.1006/bbrc.2000.2684
Marcil V Delvin E Seidman E Poitras L Zoltowska M Garofalo C Levy E Modulation of lipid synthesis, apolipoprotein biogenesis, and lipoprotein assembly by butyrate Am J Physiol Gastrointest Liver Physiol 2002 283 G340 G346 12121881
Brown CR Hong-Brown LQ Biwersi J Verkman AS Welch WJ Chemical chaperones correct the mutant phenotype of the delta F508 cystic fibrosis transmembrane conductance regulator protein Cell Stress Chaperones 1996 1 117 125 9222597 10.1379/1466-1268(1996)001<0117:CCCTMP>2.3.CO;2
Venglarik CJ Bridges RJ Frizzell RA A simple assay for agonist-regulated Cl and K conductances in salt-secreting epithelial cells Am J Physiol 1990 259 C358 C364 1696431
Zhang XM Wang XT Yue H Leung SW Thibodeau PH Thomas PJ Guggino SE Organic solutes rescue the functional defect in delta F508 cystic fibrosis transmembrane conductance regulator J Biol Chem 2003 278 51232 51242 14532265 10.1074/jbc.M309076200
Eidelman O BarNoy S Razin M Zhang J McPhie P Lee G Huang Z Sorscher EJ Pollard HB Role for phospholipid interactions in the trafficking defect of Delta F508-CFTR Biochemistry 2002 41 11161 11170 12220181 10.1021/bi020289s
Arreaza-Plaza CA Bosch V Otayek MA Lipid transport across the intestinal epithelial cell. Effect of colchicine Biochim Biophys Acta 1976 431 297 302 938655
Glickman RM Perrotto JL Kirsch K Intestinal lipoprotein formation: effect of cholchicine Gastroenterology 1976 70 347 352 765186
Pavelka M Gangl A Effects of colchicine on the intestinal transport of endogenous lipid. Ultrastructural, biochemical, and radiochemical studies in fasting rats Gastroenterology 1983 84 544 555 6822325
Chun J Auer KA Jacobson BS Arachidonate initiated protein kinase C activation regulates HeLa cell spreading on a gelatin substrate by inducing F-actin formation and exocytotic upregulation of beta 1 integrin J Cell Physiol 1997 173 361 370 9369949 10.1002/(SICI)1097-4652(199712)173:3<361::AID-JCP8>3.0.CO;2-L
Jiang WG Bryce RP Horrobin DF Mansel RE Regulation of tight junction permeability and occludin expression by polyunsaturated fatty acids Biochem Biophys Res Commun 1998 244 414 420 9514943 10.1006/bbrc.1998.8288
Anant S Davidson NO Molecular mechanisms of apolipoprotein B mRNA editing Curr Opin Lipidol 2001 12 159 165 11264987 10.1097/00041433-200104000-00009
Le Beyec J Chauffeton V Kan HY Janvier PL Cywiner-Golenzer C Chatelet FP Kalopissis AD Zannis V Chambaz J Pincon-Raymond M Cardot P The -700/-310 fragment of the apolipoprotein A-IV gene combined with the -890/-500 apolipoprotein C-III enhancer is sufficient to direct a pattern of gene expression similar to that for the endogenous apolipoprotein A-IV gene J Biol Chem 1999 274 4954 4961 9988739 10.1074/jbc.274.8.4954
Hussain MM Kancha RK Zhou Z Luchoomun J Zu H Bakillah A Chylomicron assembly and catabolism: role of apolipoproteins and receptors Biochim Biophys Acta 1996 1300 151 170 8679680
Wetterau JR Aggerbeck LP Bouma ME Eisenberg C Munck A Hermier M Schmitz J Gay G Rader DJ Gregg RE Absence of microsomal triglyceride transfer protein in individuals with abetalipoproteinemia Science 1992 258 999 1001 1439810
Jones B Jones EL Bonney SA Patel HN Mensenkamp AR Eichenbaum-Voline S Rudling M Myrdal U Annesi G Naik S Meadows N Quattrone A Islam SA Naoumova RP Angelin B Infante R Levy E Roy CC Freemont PS Scott J Shoulders CC Mutations in a Sar1 GTPase of COPII vesicles are associated with lipid absorption disorders Nat Genet 2003 34 29 31 12692552 10.1038/ng1145
Mansbach CM Nevin P Intracellular movement of triacylglycerols in the intestine J Lipid Res 1998 39 963 968 9610762
Levy E Thibault L Garofalo C Messier M Lepage G Ronco N Roy CC Combined (n-3 and n-6) essential fatty deficiency is a potent modulator of plasma lipids, lipoprotein composition, and lipolytic enzymes J Lipid Res 1990 31 2009 2017 2086700
Bubien JK CFTR may play a role in regulated secretion by lymphocytes: a new hypothesis for the pathophysiology of cystic fibrosis Pflugers Arch 2001 443 Suppl 1 S36 S39 11845300 10.1007/s004240100641
Moss RB Hsu YP Olds L Cytokine dysregulation in activated cystic fibrosis (CF) peripheral lymphocytes Clin Exp Immunol 2000 120 518 525 10844532 10.1046/j.1365-2249.2000.01232.x
Pan J Bear C Farragher S Cutz E Yeger H Cystic fibrosis transmembrane conductance regulator modulates neurosecretory function in pulmonary neuroendocrine cell-related tumor cell line models Am J Respir Cell Mol Biol 2002 27 553 560 12397014
Biwersi J Emans N Verkman AS Cystic fibrosis transmembrane conductance regulator activation stimulates endosome fusion in vivo Proc Natl Acad Sci U S A 1996 93 12484 12489 8901608 10.1073/pnas.93.22.12484
Asp L Claesson C Boren J Olofsson SO ADP-ribosylation factor 1 and its activation of phospholipase D are important for the assembly of very low density lipoproteins J Biol Chem 2000 275 26285 26292 10843997 10.1074/jbc.M003520200
Gusarova V Brodsky JL Fisher EA Apolipoprotein B100 exit from the endoplasmic reticulum (ER) is COPII-dependent, and its lipidation to very low density lipoprotein occurs post-ER J Biol Chem 2003 278 48051 48058 12960170 10.1074/jbc.M306898200
Yoo JS Moyer BD Bannykh S Yoo HM Riordan JR Balch WE Non-conventional trafficking of the cystic fibrosis transmembrane conductance regulator through the early secretory pathway J Biol Chem 2002 277 11401 11409 11799116 10.1074/jbc.M110263200
Fu L Sztul E Traffic-independent function of the Sar1p/COPII machinery in proteasomal sorting of the cystic fibrosis transmembrane conductance regulator J Cell Biol 2003 160 157 163 12538638 10.1083/jcb.200210086
Becker W Mansson JE Incorporation of 14C into tissue lipids after oral administration of [1-14C]linoleic acid in rats fed different levels of essential fatty acids J Nutr 1985 115 1248 1258 4045568
Williams MA Tinoco J Yang YT Bird MI Hincenbergs I Feeding pure docosahexaenoate or arachidonate decreases plasma triacylglycerol secretion in rats Lipids 1989 24 753 758 2573813
Coleman RA Wang P Bhat BG Diradylglycerols alter fatty acid inhibition of monoacylglycerol acyltransferase activity in Triton X-100 mixed micelles Biochemistry 1998 37 5916 5922 9558325 10.1021/bi9802972
Chautan M Charbonnier M Leonardi J Andre M Lafont H Nalbone G Modulation of lipid chylomicron-synthesizing enzymes in rats by the dietary (n-6):(n-3) fatty acid ratio J Nutr 1991 121 1305 1310 1880608
Stan S Delvin EE Seidman E Rouleau T Steinmetz A Bendayan M Yotov W Levy E Modulation of apo A-IV transcript levels and synthesis by n-3, n-6, and n-9 fatty acids in CACO-2 cells J Cell Biochem 1999 75 73 81 10462706 10.1002/(SICI)1097-4644(19991001)75:1<73::AID-JCB8>3.0.CO;2-K
Wang H Lu S Du J Yao Y Berschneider HM Black DD Regulation of apolipoprotein secretion by long-chain polyunsaturated fatty acids in newborn swine enterocytes Am J Physiol Gastrointest Liver Physiol 2001 280 G1137 G1144 11352806
Murthy S Albright E Mathur SN Davidson NO Field FJ Apolipoprotein B mRNA abundance is decreased by eicosapentaenoic acid in CaCo-2 cells. Effect on the synthesis and secretion of apolipoprotein B Arterioscler Thromb 1992 12 691 700 1591230
Kratz M Wahrburg U von Eckardstein A Ezeh B Assmann G Kronenberg F Dietary mono- and polyunsaturated fatty acids similarly increase plasma apolipoprotein A-IV concentrations in healthy men and women J Nutr 2003 133 1821 1825 12771323
Yao Y Eshun JK Lu S Berschneider HM Black DD Regulation of triacylglycerol and phospholipid trafficking by fatty acids in newborn swine enterocytes Am J Physiol Gastrointest Liver Physiol 2002 282 G817 G824 11960778
Jiang WG Bryce RP Horrobin DF Mansel RE gamma-Linolenic acid blocks cell cycle progression by regulating phosphorylation of p27kip1 and p57kip2 and their interactions with other cycle regulators in cancer cells Int J Oncol 1998 13 611 617 9683802
| 15869703 | PMC1134666 | CC BY | 2021-01-04 16:37:46 | no | Nutr Metab (Lond). 2005 May 3; 2:11 | utf-8 | Nutr Metab (Lond) | 2,005 | 10.1186/1743-7075-2-11 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1590120810.1371/journal.pbio.0030192Research ArticleCell BiologyInfectious DiseasesMicrobiologyGastroenterology/HepatologyPlasmodiumRattus (Rat)Mus (Mouse)Intravital Observation of Plasmodium berghei Sporozoite Infection of the Liver Dynamics of Plasmodium Sporozoite Liver InfectionFrevert Ute [email protected]
1
Engelmann Sabine
2
Zougbédé Sergine
1
Stange Jörg
1
Ng Bruce
3
Matuschewski Kai
2
Liebes Leonard
3
Yee Herman
4
1Department of Medical and Molecular Parasitology, New York University School of MedicineNew York, New YorkUnited States of America2Department of Parasitology, Heidelberg University School of MedicineHeidelbergGermany3Department of Medical Oncology, New York University School of MedicineNew York, New YorkUnited States of America4Department of Pathology, New York University School of MedicineNew York, New YorkUnited States of AmericaEgwang Thomas Academic EditorMedical Biotechnology LabsUganda6 2005 24 5 2005 24 5 2005 3 6 e1927 12 2004 30 3 2005 Copyright: © 2005 Frevert 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.
Tracking a Killer: In Vivo Microscopy Reveals Details on the Life Cycle of Malarial Parasites
Plasmodium sporozoite invasion of liver cells has been an extremely elusive event to study. In the prevailing model, sporozoites enter the liver by passing through Kupffer cells, but this model was based solely on incidental observations in fixed specimens and on biochemical and physiological data. To obtain direct information on the dynamics of sporozoite infection of the liver, we infected live mice with red or green fluorescent Plasmodium berghei sporozoites and monitored their behavior using intravital microscopy. Digital recordings show that sporozoites entering a liver lobule abruptly adhere to the sinusoidal cell layer, suggesting a high-affinity interaction. They glide along the sinusoid, with or against the bloodstream, to a Kupffer cell, and, by slowly pushing through a constriction, traverse across the space of Disse. Once inside the liver parenchyma, sporozoites move rapidly for many minutes, traversing several hepatocytes, until ultimately settling within a final one. Migration damage to hepatocytes was confirmed in liver sections, revealing clusters of necrotic hepatocytes adjacent to structurally intact, sporozoite-infected hepatocytes, and by elevated serum alanine aminotransferase activity. In summary, malaria sporozoites bind tightly to the sinusoidal cell layer, cross Kupffer cells, and leave behind a trail of dead hepatocytes when migrating to their final destination in the liver.
Tiny parasites of the Plasmodium genus cause malaria. New details of the parasites' life cycle are uncovered through the use of intravital microscopy to observe the parasites' infiltration of the liver.
==== Body
Introduction
After transmission by an infected mosquito, malaria sporozoites enter a blood vessel at the bite site [1] and travel via the bloodstream to the liver, their initial site of replication in the mammalian host. But to gain access to hepatocytes, where they replicate, sporozoites must first recognize and arrest in the liver and then cross the sinusoidal cell layer, which is composed of specialized, highly fenestrated endothelia and Kupffer cells, the resident macrophages of the liver (Figure 1).
Figure 1 Architecture of the Liver Sinusoid
Liver sinusoids (S) are lined by fenestrated endothelia (EC) and interspersed Kupffer cells (KC), the resident macrophages of the liver. Stellate cells (SC), the major producers of liver ECM, are located inside the narrow space of Disse (D), which is formed by the sinusoidal cell layer and cords of hepatocytes (H).
Initial data had indicated that arrest in the liver is mediated by binding of Plasmodium sporozoites via their major surface proteins, circumsporozoite protein (CSP) and thrombospondin-related adhesive protein, to the unique heparin-like oligosaccharides in hepatic heparan sulfate proteoglycans [2–10]. Analysis of the individual glycosaminoglycans produced by the major liver cell types revealed that stellate cells, perisinusoidal fat-storing cells of a dendritic shape [11], synthesize eight times more sulfated proteoglycans than hepatocytes and secrete most of these proteoglycans into the extracellular matrix (ECM) [12,13]. Recent work showed that CSP interacts with large secreted proteoglycans from stellate cells, distinct chondroitin and heparan sulfate proteoglycans on the surface of Kupffer cells, and heparan sulfate proteoglycans expressed on the surface of hepatocytes, but not with proteoglycans synthesized by sinusoidal endothelia [4]. Our previously published data indicate that the initial arrest of malaria sporozoites in the liver is mediated by the large stellate-cell-derived ECM proteoglycans that protrude from the space of Disse through the endothelial sieve plates into the sinusoidal lumen [4,14]. The parasites then glide along the endothelial cell layer until they recognize proteoglycans expressed on the surface of Kupffer cells. The best evidence obtained from fixed specimens and biochemical and physiological data suggests that sporozoites enter the liver by passing through Kupffer cells [4,14–21], although it has been suggested they cross directly through the fenestration of sinusoidal endothelia [22]. Support for Kupffer cells being the portal to the liver parenchyma has come from an in vitro model based on purified P. berghei and P. yoelii salivary gland sporozoites and Kupffer cells isolated from rat livers [14] showing that sporozoites actively invade Kupffer cells, enter a nonfusogenic vacuole, and exit the phagocytes unharmed. Further support for Kupffer cells being the gate to the liver derives from mutant P. berghei strains deficient in either SPECT1, a micronemal protein of unknown function, or SPECT2, a sporozoite protein with a putative membrane attack complex domain, which exhibit greatly diminished infectivity [23,24]. However, the mechanism of action of these proteins, their role in sporozoite entry into the liver, and the fate of the mutant sporozoites in the infected host remain to be established. Our hypothesis is that sporozoites actively enter Kupffer cells, traverse them by forming a nonfusogenic vacuole, exit them unharmed toward the space of Disse [4,14,16,17], and then migrate through several hepatocytes before eventually settling down in a final one for multiplication and differentiation to thousands of merozoites [25].
As stated above, the current model is based on a collection of indirect data. Direct confirmation of this model is essential prior to further development of our understanding of the molecular events involved in this process. However, direct, real-time observation of malaria sporozoite invasion within the liver is difficult, because the liver is a dense internal organ and the chance of observing a sporozoite in a limited viewing range requires inoculation of large numbers of sporozoites. Although large numbers of sporozoites can be routinely isolated from mosquito salivary glands, only a minority of these is infective in vitro. Even with optimized conditions, P. berghei invasion rates reach only 10%–20% in vitro [26], perhaps because a small percentage of the sporozoites in mosquito salivary glands are infectious [27,28]. Here, we present an approach that has overcome these obstacles through use of intravital microscopy for direct observation of fluorescent P. berghei sporozoites transmitted naturally by the bite of infected Anopheles stephensi mosquitoes. We were able to directly observe within the livers of live mice and rats the process of sporozoites being arrested within liver sinusoids, migrating along the sinusoid, entering and passing through Kupffer cells, traversing and damaging hepatocytes, and, finally, entering and remaining within hepatocytes.
Results
Intravital Studies
Mice or rats were placed ventral side down on the stage of an inverted digital microscope with the liver exposed and immobilized so as to limit to a minimum motion caused by breathing. Typically, this allowed a 0.5–1-cm2 area of the liver surface to be examined by intravital microscopy. We were aided by the use of transgenic Tie2-GFP and lys-EGFP-ki mice, which express green fluorescent protein (GFP) in endothelia and in phagocytes (including Kupffer cells), respectively [29,30], thus allowing us to identify sinusoidal cell types in the liver. GFP was excited at 488 nm and documented by confocal or digital epifluorescence microscopy. The natural autofluorescence emission of the tissue allowed simultaneous revelation of the liver architecture. Confocal laser scanning as well as conventional epifluorescence microscopy permitted viewing through the intact liver capsule into the parenchyma to a depth of up to 50 μm so that we could observe the entire peripheral layer of hepatocytes and sinusoidal capillaries. In the Tie2-GFP mice, GFP was expressed predominantly by endothelia of the periportal zone of the liver lobule (Video S1; Figure 2A and 2B), while in the lys-EGFP-ki mice, Kupffer cells throughout the entire lobule exhibited equal GFP signals (Video S2; Figure 2C and 2D). Neutrophil granulocytes could be distinguished by their shape and by a GFP signal brighter than that of Kupffer cells.
Figure 2 Distribution of GFP-Expressing Endothelia and Kupffer Cells in the Liver
(A) Confocal microscopy demonstrates the GFP distribution in sinusoidal endothelia from a Tie2-GFP mouse. GFP is most prominent in the perinuclear region (arrowheads) of endothelia located in the periphery of the liver lobule.
(B) A still image from an intravital movie shows GFP-expressing endothelia lining the sinusoids of a Tie2-GFP mouse. Kupffer cells can be identified by their orange autofluorescent lysosomes (arrowheads).
(C) Star-shaped Kupffer cells (arrowheads) are located in sinusoids of a lys-EGFP-ki mouse liver.
(D) Round blood granulocytes (arrows), traveling with the bloodstream or crawling along the sinusoidal cell layer, exhibit a stronger GFP signal than Kupffer cells (still image extracted from an intravital movie). Note the orange autofluorescence of the Kupffer cell lysosomes (arrowheads).
Bar = 10 μm. See Videos S1 and S2.
Sporozoite infection for intravital examination was done exclusively by mosquito bite for the following reasons. First, the continuing biting activity of the mosquitoes and the gradual escape of the parasites from the skin [1,22,31] allowed examination of sporozoites entering the liver over a period of at least 3 h. Intravenous sporozoite injection would have narrowed the window of analysis to a few minutes [22]. Second, electron microscopic examination revealed that routine sporozoite preparations are highly contaminated with salivary gland debris (U. Frevert, unpublished data). Intravenous inoculation would have caused most of this mosquito debris to be cleared from the bloodstream by Kupffer cells. The consequences of extensive phagocytosis of foreign material for sporozoite passage would have been unpredictable. Third, in agreement with the notion that only a small pool of the Plasmodium sporozoites in salivary glands may be infectious [27], the vast majority of purified salivary gland sporozoites are unable to infect hepatoma cells in vitro, even under optimal conditions [26]. Inoculation of a mixture of infectious and noninfectious parasites would have obscured the difference between active sporozoite invasion and phagocytosis of dead or damaged parasites. We reasoned that sporozoites that are able to leave the skin at the mosquito bite site [1] compose in all likelihood the most viable parasite pool and therefore reflect the onset of the liver phase of the malaria life cycle most accurately.
In the 26 mice examined, a total of 86 P. berghei sporozoites were analyzed, ranging from one to 14 parasites per mouse. P. berghei sporozoites were detected at various stages of the liver infection cascade. The dynamic aspects of the events are critical to the proper interpretation of the data (see Videos S1–S13). Eleven (12.8%) sporozoites were observed while they were gliding along the sinusoidal cell layer (Videos S3 and S4; Figure 3A and 3B) and then crossing it. Four of these observations were made in lys-EGFP-ki mice, and in all four cases, passage occurred through a Kupffer cell (Videos S5 and S6; Figure 4). Another seven sporozoites were found crossing this barrier in Tie2-GFP or Balb/c mice or in Brown Norway rats in a similar fashion (data not shown). In these animals, Kupffer cells were identified by their orange autofluorescent lysosomes (see, e.g., Figure 2D). Forty-six (53.5%) sporozoites had already entered the liver tissue at the time of detection and were in the process of transmigrating through hepatocytes (Videos S7 and S8; Figure 5). Ten (11.6%) sporozoites were sitting still inside a hepatocyte and had apparently completed the transmigration phase. Fifteen (17.4%) parasites were caught in the process of gliding along the sinusoidal cell layer but failed to pass into the liver tissue (Video S9; Figure 6A), and four (4.7%) sporozoites initially migrated in the tissue, but then re-entered the bloodstream and were flushed out of the liver lobule (Video S10; Figure 6B). The maximum time of sporozoite migration observed was approximately 15 min, and the longest covered distance was an estimated 400–500 μm (i.e., 40 to 50 times the length of a sporozoite).
Figure 3 Sporozoite Gliding along the Sinusoidal Endothelium
(A) A P. berghei sporozoite expressing fluorescent RedStar protein glides with and against the bloodstream inside a liver sinusoid of a Tie2-GFP mouse. The arrow indicates the overall movement of the parasite.
(B) The projection through the same area of the liver visualizes the outline of the highly branched sinusoids. The direction of the blood flow is indicated by the dashed arrows.
Bar = 10 μm. See Videos S3 and S4.
Figure 4 Sporozoite Passage into the Liver Parenchyma
(A–E) show individual frames extracted from an intravital movie; (F) is a projection visualizing the transmigration path of the GFP P. berghei sporozoite in a Tie2-GFP mouse liver. (G–I) show projections from an intravital movie demonstrating the path of the parasite; its overall direction is indicated by arrows (dotted lines).
(A) After gliding along a sinusoid, a sporozoite has encountered a Kupffer cell, which it faces with its apical cell pole.
(B and C) Following a pause, the parasite slowly enters the Kupffer cell.
(D and E) Sporozoite passage into the liver parenchyma occurs at a slow speed and involves the formation of a constriction in the parasite (arrow).
(F) Once inside the liver tissue, the sporozoite increases its speed and transmigrates through several hepatocytes.
(G) A projection from an intravital movie shows the path of a GFP P. berghei sporozoite gliding against the bloodstream along a sinusoid in a lys-EGFP-ki mouse liver. Eventually, the parasite encounters a Kupffer cell.
(H) The sporozoite stops, facing the phagocyte with its apical cell pole. The outline of the two Kupffer cells in the image is indicated by dotted lines.
(I) After slowly passing through the Kupffer cell, the sporozoite enters the liver parenchyma and migrates through several hepatocytes.
Bars = 10 μm. See Videos S5 and S6.
Figure 5
P. berghei Sporozoite Transmigration Is Independent of the Species of the Infected Host
Projections of typical GFP P. berghei sporozoite paths show that the parasites transmigrate in a similar fashion through many hepatocytes in mouse ([A] Tie2-GFP mouse; [B] lys-EGFP-ki mouse) and also in rat livers (C and D).
Bars = 10 μm. See Videos S7 and S8.
Figure 6 Unsuccessful Attempts of Liver Infection
(A) The composite image of ten selected frames from an intravital movie shows a paralyzed or dead P. berghei sporozoite that is eventually dislodged and flushed out of the liver lobule (short arrows). The parasite maintains a fixed crescent shape, fails to cling to the sinusoidal cell layer, and makes no attempt to glide against the bloodstream. The long arrow (dotted line) indicates the direction of the blood flow.
(B) Projection composed of 14 selected frames extracted from an intravital movie showing a GFP P. berghei sporozoite that initially transmigrates in the liver parenchyma (dotted lines), but then reenters a sinusoid (arrow) and is swept away with the bloodstream (solid lines).
Bars = 10 μm. See Videos S9 and S10.
When being transported passively with the bloodstream, sporozoites travel at a speed of 11.2 ± 5.7 μm/s (Table 1). For comparison, the sinusoidal velocity of neutrophil blood granulocytes in lys-EGFP-ki mice was 29.0 ± 13.9 μm/s. The analysis was restricted to phagocytes traveling free in the bloodstream; cells crawling along the sinusoidal cell layer were excluded. We observed sporozoites entering a liver lobule to be abruptly arrested by binding to the sinusoidal cell layer (data not shown). After a short pause, these sporozoites began gliding along the sinusoidal cell layer at a speed of 1.7 ± 0.7 μm/s (Table 1). Neither the shear force of the blood plasma nor the repeated encounters with passing blood cells released attached parasites from the sinusoidal cell wall, suggesting adhesion with high affinity. This conclusion is supported by our observation that sporozoites glided on sinusoidal cell layer equally well with and against the direction of blood flow. Paralyzed or dead parasites were characterized by a fixed crescent shape and no movement against the bloodstream (see Video S9; Figure 6A). They seemed not to adhere as well, since they were frequently released from the sinusoidal cell layer and eventually flushed out of the liver lobule.
Table 1
Plasmodium Sporozoite Migration Speed
Prior to crossing the sinusoidal cell layer, sporozoites adhered with their apical end to a Kupffer cell. After a pause of 99.7 ± 49.5 s (see Videos S5 and S7; Figure 4A–4I; Table 1), the parasites slowly pushed, at a speed of 0.3 ± 0.2 μm/s, through a clearly visible constriction (see Video S5; Figure 4D and 4E). The time required for completing the passage was 51.7 ± 52.9 s. Compared with the velocities of sporozoite gliding along the sinusoidal cell surface and sporozoite transmigration through hepatocytes, which did not differ significantly from each other or from sporozoite gliding in Matrigel (Video S11; Table 1), the sporozoite passage through Kupffer cells was slower by an order of magnitude (p ≤ 0.0001). Since the gliding speed in Matrigel (1.8 ± 0.6 μm/s) is similar to the speed within the liver and since the migration pattern in Matrigel resembles that in skin and liver, we think that gliding in Matrigel offers a reasonable model for the study of sporozoite gliding.
Once inside the liver parenchyma, sporozoites increased their speed to 1.6 ± 0.5 μm/s (Table 1), seeming to effortlessly breach the plasma and intracellular membranes of several hepatocytes (see Video S6; Figure 4I) and most likely causing substantial damage to the subcellular architecture (Video S12). Sporozoites were observed transmigrating through the liver parenchyma for up to 15 min before eventually settling down in a final hepatocyte. Sporozoite transmigration through hepatocytes differed clearly from Kupffer cell passage in that this mode of invasion did not involve the formation of a constriction, a pause before entry, or a decrease in migration speed. In contrast to in vitro gliding, the parasites were highly flexible in vivo and frequently changed their direction. In some cases, sporozoites passed across tissue barriers in a manner that cannot be explained solely by gliding motility: parasites were observed to compress or coil up before crossing the barrier and to use their posterior end to propel themselves across the obstacle (Video S13).
To document the path of migrating sporozoites, projections were generated from digital movies. The parasites generally followed the architecture of the liver lobule, i.e., the sinusoidal lumen (see Videos S3, S6, S9, and S10; Figures 3A, 4G, 6A, and 6B) or the hepatocyte chords (see Videos S6, S7, S8, and S10; Figures 4I, 5A–5D, and 6B). Overall, 77.9% of the P. berghei sporozoites traveling in the sinusoidal blood were able to enter the liver parenchyma of mice—which is not to say, however, that all these sporozoites would have matured to exoerythrocytic forms (EEFs). A few sporozoites were observed re-entering the bloodstream after successful transmigration in the liver parenchyma and being flushed out of the liver lobule (see Video S10; Figure 6B).
Compared with Balb/c mice, young Brown Norway rats are considerably more susceptible to infection with P. berghei sporozoites, coinciding with a markedly less pronounced inflammatory response [32]. Since the inflammatory response of the host is thought to eliminate a large proportion of the EEFs in mice, we reasoned that a shorter transmigration phase, resulting in a lower number of dead hepatocytes, would explain the higher sensitivity of rats to P. berghei infection. To test this hypothesis, we infected young Brown Norway rats with green fluorescent P. berghei sporozoites. Surprisingly, intravital examination of the livers of these animals demonstrated the same general sporozoite migration behavior as in mice (see Videos S7 and S8; Figure 5A–5D), suggesting that the extent of hepatocyte transmigration does not directly correlate with the degree of hepatocyte death and/or inflammation.
Histopathology
In fixed liver sections, focal hepatocyte necrosis was found whether the animals were infected by intravenous sporozoite inoculation or by mosquito bite. Starting a few hours after sporozoite infection, mouse livers contained individual or small clusters of hepatocytes that exhibited signs of necrosis, from hydropic swelling to cell death and complete cell disintegration (Figure 7). At 3 h after intravenous inoculation of 5 × 106
P. berghei sporozoites (Figure 7A) and more pronounced at 4 h and 6 h after inoculation with 5 × 106
P. yoelii sporozoites (Figure 7B and 7C), we found individual or small groups of hepatocytes exhibiting various stages of necrotic cell death that were in intimate contact with neutrophil granulocytes and mononuclear inflammatory cells. Electron microscopic examination of a mouse liver 6 h after infection by bite of approximately 50 P. yoelii–infected mosquitoes showed occasional sporozoites surrounded by a parasitophorous vacuole membrane inside morphologically intact hepatocytes (Figure 8A), with neighboring hepatocytes clearly undergoing necrosis (Figure 8B and 8C). Electron microscopy failed to reveal evidence of apoptosis in any of the various liver cell types. The appearance of necrotic hepatocytes was strictly dependent on sporozoite infection; there was no parenchymal cell damage in control livers from animals exposed to uninfected mosquito bites (data not shown). The same general histopathological pattern was observed in mouse livers infected with P. berghei or P. yoelii and also in P. berghei–infected rat livers. At 40 h after inoculation of 5 × 106
P. yoelii sporozoites, shortly before the release of merozoites, we found that inflammatory cells had infiltrated the mouse livers in a focal fashion; this was more pronounced at 52 h (see Figure 7D and 7E). Maturing EEFs, however, were always clear of any inflammatory reaction (see Figure 7D).
Figure 7 Transmigrating Plasmodium Sporozoites Leave Behind a Trail of Dead Hepatocytes
(A) Three hours after infection with 5 × 106
P. berghei salivary gland sporozoites, a mouse liver contains small clusters of necrotic hepatocytes that have been infiltrated by inflammatory cells (arrows).
(B) Four hours after intravenous infection with 5 × 106
P. yoelii sporozoites, a mouse liver contains individual or small clusters of necrotic hepatocytes (arrows).
(C) Six hours after infection, the signs of hepatocytic damage appear more severe in another mouse liver (arrow).
(D) Forty hours after inoculation of 2 × 106
P. yoelii sporozoites, small infiltrates of inflammatory cells (arrows) block the lumina of some sinusoids, while maturing EEFs are free of any inflammatory reaction.
(E) Fifty hours after infection with 2 × 106
P. yoelii sporozoites, the size of the inflammatory infiltrates (arrows) has increased.
Stains used: (A) paraffin section stained with H&E, (B and C) frozen sections stained with Evans blue, (D and E) semithin Epon sections stained with Toluidine blue. Bars = 20 μm.
Figure 8 Sporozoites, Surrounded by a Parasitophorous Vacuole Membrane, Can Be Found in Intact Hepatocytes
(A) Six hours after infection by bite of 50 P. yoelii–infected mosquitoes, electron microscopic examination of a mouse liver shows a sporozoite inside an intact hepatocyte. Note that the parasite is enclosed in a parasitophorous vacuole (insert). The neighboring hepatocyte shows signs of cytoplasmic swelling.
(B and C) P. yoelii–infected mouse livers contain hepatocytes that exhibit various degrees of necrosis, ranging from hydropic swelling to near-complete disintegration of the cell, adjacent to parenchymal cells with a normal ultrastructure (arrows). The normal hepatocytes did not contain a sporozoite in the plane of the section.
L, lipid droplet; M, mitochondrium; N, nucleus; S, sinusoid. Bars = 1 μm.
Compared with uninfected control livers (not shown), livers from mice that had been exposed on two consecutive days to the bite of approximately 150 P. yoelii–infected mosquitoes and fixed 2 h after the second infection showed only minor alterations (Figure 9A–9D). More dramatic histopathological changes became apparent at later stages of the liver infection. This was found by infecting mice each day for a week by bite of approximately 150 P. yoelii–infected mosquitoes per day. Proliferation of blood stages was prevented by quinine treatment. Hematoxylin-eosin (H&E)–stained liver sections of these mice showed large numbers of nonparenchymal cells lining the sinusoids 7 d after infection (Figure 9E). Masson's trichrome staining revealed focal deposition of collagen in some spaces of Disse (Figure 9F). Immunohistochemistry using monoclonal antibody (mAb) PC10, directed against proliferating cell nuclear antigen, revealed hepatocyte proliferation and marked Kupffer cell hyperplasia (Figure 9G). An increased deposition of alpha-smooth muscle actin in the space of Disse (Figure 9H) was found after labeling with mAb HHF35, suggesting focal stellate cell activation [33].
Figure 9 Sporozoite Transmigration Causes Histopathological Changes in the Liver
Mouse livers were removed 2 d (A–D) or 7 d (E–H) after daily infection with P. yoelii by bite of 150 mosquitoes and stained with H&E (A and E) or Masson's trichrome (B and F). Other sections were subjected to immunohistochemistry using mAb PC10 against proliferating cell nuclear antigen (C and G) or mAb HHF35 against smooth muscle actin (D and H). In contrast to the livers fixed after 2 d of infection, in which only a few cells reacted with mAb PC10 and mAb HHF35 (arrows in C and D), livers examined after 7 d of infection showed (E) increased numbers of nonparenchymal cells lining the sinusoids (arrow), (F) a focal deposition of collagen (blue) in some spaces of Disse, (G) large numbers of proliferating nonparenchymal cells and hepatocytes (brown, arrows), and (H) a focal increase in the concentration of smooth muscle actin (brown, arrow).
Liver Transaminases
Alanine aminotransferase (ALT) is commonly used as a serum marker of liver damage, and we used this to detect hepatic injury after sporozoite infection. Intravenous inoculation into mice of 0.7, 1.2, or 1.8 × 106
P. yoelii sporozoites resulted in a significant increase of the ALT serum levels compared with the control sera collected before infection (Figure 10A). The degree of the increase depended on the number of injected sporozoites, and the elevated levels persisted over the entire observation period of 52 h. Because the sporozoite preparations contained debris from 60, 90, and 150 salivary glands, respectively, we also evaluated the effect of mosquito debris on the ALT activity. Extracts from uninfected salivary glands were prepared in an identical fashion, and three mice were injected intravenously with extract from 100 salivary glands each. The serum ALT levels increased temporarily at 9 h, but returned to base levels at 24 h and remained low until 52 h after inoculation (Figure 10A). The extent of this temporary increase resembled that observed after inoculation of 1.2 × 106 sporozoites, which were purified from 90 salivary glands. Debris from 60 glands (containing 0.7 × 106 sporozoites) had no effect at 9 h, while extract from 150 glands (1.8 × 106 sporozoites) resulted in a strong elevation of the ALT levels at 9 h. Mice infected with P. berghei (Figure 10B) or P. yoelii (data not shown) by a single exposure to the bite of about 150 mosquitoes per animal did not exhibit any measurable increase in the ALT activity in the serum.
Figure 10 Sporozoite Infection Increases the Serum ALT Activity
(A) Three mice were inoculated with salivary gland extract from 100 uninfected mosquitoes each (labeled “0” on the x-axis). Another three mice were infected by intravenous inoculation into the tail vein of 0.7 × 106, 1.2 × 106, or 1.8 × 106 purified P. yoelii sporozoites (indicated as “0.7,” “1.2,” and “1.8” on the x-axis). The ALT activity in the serum was determined before and after infection at the indicated time points. In comparison to the control serum drawn before infection, the ALT levels increased significantly and depended on the number of inoculated sporozoites in all mice during the observation period of 52 h. Uninfected salivary gland extract had only a temporary effect on the serum ALT level (9 h). The indicated values represent the average ± standard deviation of triplicate measurements. *, p < 0.005 in relation to the corresponding control sera.
(B) No change in the serum ALT activity was detectable when three mice were infected with P. berghei by bite of 150 mosquitoes. The indicated values represent the average ± standard deviation of duplicate measurements. For ready comparison, the data were normalized for each animal prior to statistical analysis.
Discussion
Intravital microscopy enabled us to observe and document the cascade of events leading to malaria sporozoite infection of the liver. Our recordings show that sporozoites entering the liver lobule are abruptly arrested by binding to the sinusoidal cell layer, presumably by recognizing liver-specific ECM proteoglycans protruding through the endothelial fenestration into the sinusoidal lumen [4]. Sporozoites then glide along the sinusoid until they encounter a Kupffer cell, where they stop with their apical end attached to the phagocyte. After a pause, they push slowly across the sinusoidal cell barrier, showing a point of constriction in their outline. Once inside the liver parenchyma, they continue to migrate for many minutes, fatally wounding several hepatocytes in the process. Eventually, the parasites settle down in one final hepatocyte, become surrounded by a parasitophorous vacuole membrane, and differentiate to EEFs (model in Figure 11).
Figure 11 Model of Plasmodium Sporozoite Infection of the Mammalian Liver
The dual blood supply of the liver, consisting of branches of the portal vein and the hepatic artery, merges upon entry into the liver lobule at the portal field. The blood flows along the sinusoid and exits at the central vein. First sporozoites enter the liver lobule either via the portal vein or the hepatic artery, and then are abruptly arrested by binding to the sinusoidal cell layer. The initial binding is presumably mediated by stellate-cell-derived ECM proteoglycans that protrude from the space of Disse across the endothelial sieve plates into the sinusoidal lumen. After a pause, the parasites begin to glide along the sinusoid, frequently moving against the bloodstream, until they then encounter a Kupffer cell, on the surface of which they recognize selected chondroitin and heparan sulfate proteoglycans. Sporozoites position themselves with their apical cell pole facing the phagocyte. After a considerable pause, they slowly pass through the Kupffer cell and cross the space of Disse beyond it, exhibiting a clearly visible constriction. Once inside the liver parenchyma, the parasites increase their velocity and migrate for many minutes through several hepatocytes, before they eventually settle down in a final one for EEF development. Sporozoite transmigration results in a trail of necrotic hepatocytes, whose remains are subsequently removed by infiltrating inflammatory cells.
The overall velocities of sinusoidal gliding and hepatocyte transmigration are similar to sporozoite gliding in Matrigel (approximately 1.8 μm/s; see Table 1), on artificial surfaces in vitro (1–3 μm/s) [34], and in mosquito salivary glands (≤ 2 μm/s) [27], but migration through Kupffer cells is slower by an order of magnitude. The speeds of sinusoidal gliding and hepatocyte transmigration (Table 2) fall into the general speed range of 1–10 μm/s found for various other apicomplexan parasites [35,36], including Toxoplasma gondii tachyzoites [37], Eimeria tenella sporozoites [38], and Gregarina polymorpha [35]. Interestingly, the velocity of Plasmodium ookinetes (0.08–0.25 μm/s) is slower than that of sporozoites [39,40] and similar to that of sporozoites traversing Kupffer cells, perhaps reflecting an adaptation to the environment of the mosquito midgut.
Table 2 Gliding Speeds of Apicomplexan Parasites
The process of Kupffer cell passage differs from passage through hepatocytes by more than speed. Entry into hepatocytes does not involve a pause before entry, a constriction in the parasite during passage, or a change in migration speed (see Table 1). Perhaps the slower speed relates to the need to form a vacuole or duct for passage through Kupffer cells [14,15,41] in contrast to direct breaching of the cell membrane during transmigration through hepatocytes [25,42]. Alternatively, the barrier that causes the constriction and slows the parasites could be the ECM in the space of Disse just beyond the Kupffer cell. A similar phenomenon has been observed for P. gallinaceum ookinetes, which cross the microvilli-associated network of Aedes aegypti midgut epithelia at roughly 10-fold slower a speed compared with the subsequent migration on or through the cells [40,43]. However, since host cell entry can also cause the formation of a constriction in apicomplexan parasites [44−46], additional work is required to identify the nature of the barrier presented to sporozoites leaving the sinusoid.
In contrast to motility on artificial surfaces in vitro, where sporozoites maintain a fixed crescent shape and move in circles or spirals [34,47], sporozoite migration in vivo and in Matrigel is characterized by a more linear migration pattern, high parasite flexibility, and frequent changes in direction, most likely guided by structural tissue components. The general movement pattern in the liver resembles that of sporozoites gliding in the skin after transmission by mosquito bite [1]. P. berghei sporozoites in An. stephensi salivary glands also move according to tissue architecture and follow the outline of the acinar epithelia and the secretory duct [27]. It appears that Plasmodium sporozoites, in general, are guided by the three-dimensional arrangement of matrix structures to reach their next destination in a tissue.
We were surprised to find that P. berghei sporozoites traverse a similar number of hepatocytes in livers from mice and rats. P. berghei is much less infectious to mice than to its natural host, the African wild rat Thamnomys surdaster [48,49], and induces a more pronounced inflammatory reaction in these rodents, while young rats exhibit intermediate levels of infection and inflammation [50–52]. Based on our data, it appears that sporozoite transmigration is an intrinsic feature of Plasmodium sporozoites, which occurs irrespective of the host–parasite combination. Thus, it may be the degree of the mismatch between parasite and host that determines the extent of the inflammatory response to hepatocyte necrosis.
Malaria sporozoite migration can cause significant injury to the liver tissue, as demonstrated by the inoculation-size-dependent increase in the serum ALT levels for at least 2 d after intravenous inoculation of P. yoelii sporozoites (see Video S12). Uninfected gland extracts had a significant, albeit temporary, effect, suggesting that the sustained ALT elevation observed 24 and 52 h after sporozoite inoculation was indeed a consequence of parasite migration. Although the ALT levels measured here are well below those observed in experimental fulminant hepatitis caused by murine hepatitis virus 3 infection [53], the sporozoite-induced damage was large enough to induce deposition of detectable amounts of collagen in the space of Disse, focal Kupffer cell hyperplasia, and groups of proliferating hepatocytes at later time points. The liver damage induced by bite of infected mosquitoes, although morphologically detectable, remained below the detection limit of the ALT assay, most likely because the number of sporozoites reaching the liver under natural transmission conditions is too low to induce a significant hepatotoxic effect. Mosquitoes generally transmit less than 100 sporozoites per bite [1,50,54,55], so under optimal conditions 150 mosquitoes would have transmitted roughly 15,000 sporozoites per mouse, only part of which leave the skin and travel to the liver (J. Vanderberg and U. Frevert, unpublished data). These findings also suggest either that mosquitoes do not inject significant amounts of saliva into the bloodstream while probing the skin or that saliva has no adverse effect on the liver parenchyma. In conclusion, while experimental infection with large numbers of purified sporozoites can inflict measurable liver injury, natural infection by mosquito bite poses no risk to liver function, even when small rodents are exposed to large numbers of well-infected mosquitoes under laboratory conditions. These data confirm that sporozoite-induced liver injury is of no concern to malaria-infected individuals living in endemic areas.
A contradictory model of sporozoite entry into the liver was recently presented by Yuda and colleagues [23]. The group generated P. berghei parasites deficient in the micronemal proteins SPECT1 and SPECT2, which exhibit greatly diminished liver infectivity after intravenous inoculation into mice. Experimental data on the function of the SPECT proteins are lacking to date, but SPECT2 has been proposed to contain a membrane attack complex that mediates membrane wounding and, consequently, sporozoite transmigration through cells. In vitro results obtained from permanent nonphagocytic cell lines supported this hypothesis, but the authors then extrapolated from these data to Kupffer cells, hypothesizing that SPECT mutants have lost their ability to wound the Kupffer cell membrane and therefore cannot pass through the cytoplasm of these phagocytes. This is in contrast to Kupffer cell passage by membrane invagination as documented by others [14,15,41]. To explain this discrepancy and taking into account the putative membrane attack complex in SPECT2, one could speculate that SPECT mutants accumulate inside Kupffer cells because they have lost the ability to escape the vacuole surrounding them in these cells.
In another study by Ishino and colleagues, Kupffer cell elimination more than doubled the rate of liver infection by P. berghei [24]; this would not be possible at the approximately 80% infection rate presented here. The most likely explanation for this discrepancy is that our results are based on natural sporozoite transmission by mosquito bite, while Ishino and colleagues used intravenous injection, which results in markedly lower liver infection efficiencies relative to the size of the inoculum [56]. It is generally believed that preparations of purified sporozoites contain a considerable percentage of noninfectious parasites, the majority of which is likely cleared from the bloodstream by Kupffer cells. Kupffer cells may become activated upon phagocytosis of dead sporozoites, and the resulting inflammatory microenvironment would clearly inhibit EEF development [57]. This would not occur in the absence of Kupffer cells [58], thus explaining the large increase in the liver infection rate in clodronate-treated mice. In addition, clodronate released by leakage from liposomes or by dying Kupffer cells has been suggested to suppress inflammatory cytokines [58], which would also enhance EEF survival.
Plasmodium sporozoites continuously release vesicles covered with CSP and thrombospondin-related adhesive protein from their cell surface membrane [47,59–61]. When gliding on artificial surfaces such as glass, sporozoites leave these vesicles behind in the shape of a trail, but when migrating on cultured cells or in tissues such as mosquito salivary glands, the released CSP translocates across membranes; it initially spreads across the cytoplasm of the affected cells and inhibits protein synthesis and later redistributes to the perinuclear space [62,63]. Since sporozoites release CSP into Kupffer cells in vitro [14], we expect that this also occurs with liver endothelia and Kupffer cells in vivo. The significance of this is that both sinusoidal endothelia and Kupffer cells are fully mature antigen-presenting cells that express major histocompatibility complex class I and II molecules as well as the co-stimulatory molecules CD80 and CD86, they are able to prime naïve CD4+ T cells in vitro, and they can cross-present alloantigen to CD8+ T cells [64–67]. Intimate contact between sporozoites and nonparenchymal cells could be expected to lead to the presentation of parasite antigen—in particular, CSP-derived peptides. However, due to its ribotoxic effect, translocated CSP may interfere with antigen processing. In addition, the unique microenvironment of the liver generally favors the development of tolerance rather than inflammation and immunity, a property thought to have evolved to avoid overreaction to the continuous influx into the portal circulation of foreign materials such as bacteria and endotoxins from the intestines [68,69]. Portal vein tolerance [70] is predominantly mediated by Kupffer cells [71], can occur irrespective of nature and origin of the antigen and has been implicated in the acceptance of liver allografts, the reduced rejection of allografts when preceded by portal venous administration of donor antigen, the high frequency of tumor metastases in the liver, and the persistence of intrahepatic pathogens such as hepatitis C and hepatitis B virus [72,73]. Does this unusual route of entry, together with translocation of CSP by migrating sporozoites, enable malaria sporozoites to take advantage of the unique tolerogenic nature of the liver? Can Plasmodium, one of the deadliest infectious agents worldwide, avoid the generation of an effective immune response against the continuous influx of sporozoites in endemic areas so successfully because it begins its life cycle in the mammalian host with a round of multiplication in the liver? The liver may be the preferred site for multiplication because of its unique nature as an immune organ as opposed to other tissues that could have possibly supported replication but do not possess this specialized property.
Materials and Methods
Parasites.
An. stephensi mosquitoes were used to propagate (1) P. berghei (NK65), (2) P. yoelii (17 XNL), or (3) a P. berghei clone that expresses EGFP at the sporozoite stage under control of the CSP promoter (termed here GFP P. berghei) [74]. Another P. berghei clone was generated that expresses an improved version of the red fluorescent protein drFP583/DsRed/RFP, RedStar, under the control of the CSP promoter (termed here RedStar P. berghei); RedStar was chosen because of its 10–20× enhanced brightness when expressed in mammalian cells [75]. The transfection vector pSE-26 contains a RedStar expression cassette and is a derivative of the b3Djournal-pbio-0030192-transfection vector [76] that confers resistance to pyrimethamine. The RedStar open reading frame was amplified from the plasmid PRS415-Gal1-RedStar [75] (a gift from M. Knop, Heidelberg) with primers RFPfor (5′ CGGGATCCAAAATGAGTAGATCTTCTAAGAAC 3′; BamHI site is underlined) and RFPrev (5′ GGACTAGTTTACAAGAACAAGTGGTGTCTACC 3′; SpeI site is underlined). The 3′ UTR region of PbDHFR was amplified from pExpEF (a gift from A. P. Waters, Leiden) with primers SEb3DforXbaI (5′ TGCTCTAGACGTTTTTCTTACTTATATAT 3′; XbaI site is underlined) and SEb3CrevSacII (5′ TCCCCGCGGCGGTGTGAAATACCGCACAGA 3′; SacII site is underlined). To drive stage-specific expression, the P. berghei CSP promoter was amplified using primers p5′CSEcoRIfor (5′ CCGGAATTCACATAAAAGGGAATATGGAATATACTAGC 3′; EcoRI site is underlined) and p5′CSBamHIrev (5′ CGCGGATCCAAATATATGCGTGTATATATAGATTTTG 3′; BamHI site is underlined) and genomic PbDNA. Parasite transfection and selection was performed as described previously [76]. Serial dilution of the parental pyrimethamine-resistant parasites resulted in three independent clonal parasite lines that stably express RedStar during sporozoite stages.
Animals
Mice were (1) Balb/c (Taconic, Germantown, New York, United States); (2) Swiss Webster (Taconic); (3) a transgenic strain expressing GFP in vascular endothelial cells under control of the Tie2 promoter [29]; (4) lys-EGFP-ki mice that express EGFP in myelomonocytic cells—in particular, neutrophil granulocytes—and tissue macrophages including Kupffer cells [30]; or (5) Tie2-GFP × lys-EGFP-ki hybrid mice. The lys-EGFP-ki mice were a kind gift from Dr. Thomas Graf, Albert Einstein College of Medicine, Bronx, New York. Brown Norway rats were purchased from Charles River (Wilmington, Massachusetts, United States).
Surgery and intravital microscopy
Mice were anesthetized by intraperitoneal injection of a cocktail of 50 mg/kg of ketamine (Ketaset, Fort Dodge Animal Health, Overland Park, Kansas, United States), 10 mg/kg of xylazine (AnaSed, Ben Venue Laboratories, Bedford, Ohio), and 1.7 mg/kg of acepromazine (Boehringer Ingelheim Vetmedica, St. Joseph, Missouri, United States). A portion of the right abdominal skin was removed and the peritoneal cavity opened along the rib cage. The mouse was then placed on the stage of an inverted Nikon (Tokyo, Japan) Diaphot microscope equipped with a Cooke SensiCam digital camera (Cooke, Romulus, Michigan, United States). The liver was immobilized with gauze and kept moist with warm PBS. Sporozoite infection was done on the microscope stage by bite of 100−200 infected mosquitoes. Images of sporozoites entering the liver were captured using either a fluorescein isothiocyanate long-pass filter combination or a GFP/DsRed dual band filter set (Chroma Technology, Rockingham, Vermont, United States) and imported into Image-Pro Plus software (Media Cybernetics, Silver Spring, Maryland, United States). Typical exposures times were 100 ms per image for gray tone and 300 ms for RGB images. Phototoxicity was limited by reducing light transmission to 20% with a neutral density filter.
Infection of mouse livers for histopathology.
Mice were intravenously inoculated with 2–5 × 106 purified P. berghei or P. yoelii salivary gland sporozoites, and the livers were removed 0.5, 1, 2, 4, 6, 12, 24, or 48 h after infection. Other mice were infected with P. berghei or P. yoelii by bite of 100–200 mosquitoes. For observation periods longer than 2 d, blood infection was prevented by daily administration of 70 mg/kg of quinine hydrochloride (Sigma, St. Louis, Missouri, United States). The tissue was fixed 2 h after the last infection with PBS containing 4% paraformaldehyde and with PBS containing 4% paraformaldehyde plus 1% glutaraldehyde for confocal and electron microscopic examination, respectively [5].
Immunofluorescence microscopy
Ten-micrometer frozen sections were prepared with a Reichert-Jung (Arnsberg, Germany) Frigocut cryostat, and 1-μm cryosections were cut with an RMC (Tucson, Arizona, United States) MT-7 ultramicrotome equipped with a MRC-21 cryochamber. Sporozoites were labeled with mAb 3D11 [77] or mAb NYS1 [78] directed against the P. berghei and P. yoelii CSP, respectively, followed by protein A conjugated to fluorescein isothiocyanate. The liver tissue was counterstained with 0.1% of Evans blue in PBS.
Confocal microscopy
Immunofluorescence-labeled frozen liver sections were examined with a Zeiss (Oberkochen, Germany) LSM 510 laser scanning microscope. Whole unfixed livers were examined for GFP-expressing parasites and liver cells by scanning through the intact capsule of the organ. The natural autofluorescence of the tissue was excited at 543 nm and recorded at 560–615 nm to visualize the microarchitecture of the organ.
Histopathology and immunohistochemistry
Paraffin-embedded liver tissue was processed for H&E and Masson's trichrome staining. Immunohistochemistry was performed on deparaffinized mouse liver sections using antibody PC10 against proliferating cell nuclear antigen and antibody HHF35 against muscle-specific actin as a marker specific for dedifferentiating stellate cells (both from Ventana Medical Systems, Tucson, Arizona, United States).
Electron microscopy
Mouse liver tissue was postfixed with 1% osmium tetroxide, dehydrated, and embedded in Epon as described [62]. Thin sections were cut with an RMC MT-7 ultramicrotome and post-stained with uranyl acetate and lead citrate. Photographs were taken with a Zeiss EM 910 electron microscope.
Image processing
Electron microscopy negatives were scanned with an Agfa (Mortzel, Belgium) Horizon Plus flatbed scanner. Digital, electron, or confocal microscopy images were processed using Image-Pro Plus, Adobe Photoshop (Adobe Systems, San Jose, California, United States), AutoDeBlur (AutoQuant Imaging, Troy, New York, United States), and Microsoft (Redmond, Washington, United States) PowerPoint software.
Liver transaminases.
Mice were anesthetized and infected with P. berghei or P. yoelii by bite of 150 mosquitoes per animal. An. stephensi mosquitoes were allowed to probe the skin of the mice for 5 × 2 min. Blood was drawn from the tail vein before and at various time points for up to 52 h after infection. Other mice were infected by inoculation into the tail vein of 0.7 × 106, 1.2 × 106, and 1.8 × 106
P. yoelii sporozoites, which were purified from 60, 90, and 150 salivary glands, respectively. Control mice were intravenously injected with identically prepared extracts from 100 uninfected salivary glands each. The serum ALT levels were measured at the indicated time points using a GPT/ALT assay (Wako Chemicals, Richmond, Virginia, United States) according to manufacturer's instructions. The values represent the average of triplicate measurements ± standard deviation.
Supporting Information
Digital movies document the individual steps of the P. berghei sporozoite infection cascade of the mouse or rat liver. Replay speed = 50×.
Video S1 Intravital Microscopy of the Sinusoidal Blood Flow in a Tie2-GFP Mouse Liver
Note the dark blood cells traveling inside the sinusoids, which are lined with GFP-expressing endothelia. GFP is particularly prominent in the perinuclear region of the endothelia.
(656 KB ZIP).
Click here for additional data file.
Video S2 Intravital Microscopy of the Sinusoidal Blood Flow in a lys-EGFP-ki Mouse Liver
Bright green fluorescent blood granulocytes are traveling rapidly inside the sinusoids. Kupffer cells are stationary and can be identified by their star-like shape and their lower fluorescence signal.
(656 KB ZIP).
Click here for additional data file.
Video S3 Intravital Microscopy of RedStar Sporozoite Gliding along Liver Sinusoids
P. berghei sporozoites entering the liver lobule are abruptly arrested by binding to the sinusoidal endothelium. After a brief pause, the parasites begin to glide along the sinusoidal cell layer. Periodically, the parasites lose their grip to the sinusoidal cell layer and are swept with the bloodstream for a short distance before they adhere again and continue to glide. This movie clip shows a RedStar P. berghei sporozoite gliding with or against the bloodstream in a Tie2-GFP mouse liver.
(1.5 MB ZIP).
Click here for additional data file.
Video S4 Intravital Microscopy of GFP Sporozoite Gliding along Liver Sinusoids
This movie clip shows a GFP P. berghei sporozoite gliding with or against the bloodstream in a Tie2-GFP mouse liver.
(1.1 MB ZIP).
Click here for additional data file.
Video S5 Intravital Microscopy of Sporozoites Passing through Kupffer Cells
After encountering a Kupffer cell in a lys-EGFP-ki mouse liver, a GFP P. berghei sporozoite pauses for 1–2 min, facing the phagocyte with its apical cell pole. It then passes through the Kupffer cell at a slow speed, pushing through a narrow constriction. Once inside the liver parenchyma, the sporozoite increases its speed and continues its migration through several hepatocytes.
(2.6 MB ZIP).
Click here for additional data file.
Video S6 Intravital Microscopy of a Kupffer Cell Passage
Another sporozoite glides inside a sinusoid before it encounters a Kupffer cell in a lys-EGFP-ki mouse liver. After a pause, it traverses the phagocyte and continues its migration in the liver tissue for many minutes.
(2.1 MB ZIP).
Click here for additional data file.
Video S7 Sporozoite Transmigration through Hepatocytes Occurs in Mouse and Rat Livers
Typical transmigration behavior of GFP P. berghei sporozoites in a Tie2-GFP mouse liver.
(907 KB ZIP).
Click here for additional data file.
Video S8 Sporozoites Migrate through Several Hepatocytes in Mouse and Rat Livers
Typical transmigration behavior of GFP P. berghei sporozoites in the livers of young Brown Norway rats.
(1.2 MB ZIP).
Click here for additional data file.
Video S9 All Sporozoites That Adhere to the Sinusoidal Cell Layer Do Not Enter the Liver
The GFP P. berghei sporozoite does not bind tightly to the surface of the sinusoidal cell layer. It retains a fixed crescent shape and appears paralyzed or dead. Eventually, the parasite is displaced from the sinusoidal cell layer by the shear force of the bloodstream and flushed out of the liver lobule.
(2.5 MB ZIP).
Click here for additional data file.
Video S10 Sporozoites Reentering a Sinusoid Are Swept Away with the Bloodstream
This movie clip shows a GFP P. berghei sporozoite that has successfully entered the liver tissue and is in the process of transmigrating in the liver parenchyma. However, the parasite accidentally reenters a sinusoid and is flushed out of the liver lobule.
(893 KB ZIP).
Click here for additional data file.
Video S11 Intracellular Sporozoite Migration Affects the Subcellular Structure
This confocal microscopic movie clip shows a GFP P. berghei sporozoite circling around the nucleus of a CHO-K1 cell in vitro. The clearly visible displacement of cell organelles by the parasite suggests that transmigrating sporozoites can disrupt vital functions in hepatocytes as a consequence of the destruction of the cellular architecture—in particular, the abundant, well-developed rough endoplasmic reticulum.
(839 KB ZIP).
Click here for additional data file.
Video S12 Sporozoite Passage across Tissue Barriers
Occasionally, GFP P. berghei sporozoites can be observed crossing unidentified tissue barriers by propelling themselves forward in a manner that cannot solely be explained by gliding motility.
(662 KB ZIP).
Click here for additional data file.
Video S13 Sporozoite Migration in Matrigel
The sporozoite migration pattern in Matrigel resembles that in the liver or the skin. GFP P. berghei sporozoites were purified from salivary glands, suspended in Matrigel, and incubated at 37 °C for 30 min before recording.
(149 KB ZIP).
Click here for additional data file.
Accession Numbers
The GenBank (http://www.ncbi.nlm.nih.gov/Genbank/) accession number for red fluorescent protein drFP583/DsRed/RFP is AAF03369.
We thank Dabeiba Bernal-Rubio and Rita Altszuler for expert help with maintenance of the Plasmodium cycles and sporozoite purification. Many thanks to Drs. Mauricio Calvo-Calle and Jerome Vanderberg (New York University School of Medicine), as well as Gabriele Pradel (Cornell University, New York, New York) for critically reading the manuscript. We are grateful to Dr. Thomas Graf (Albert Einstein College of Medicine, Bronx, New York) for a gift of lys-EGFP-ki mice. The work was supported by the National Institutes of Health (grant AI51656).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. UF, KM, LL, and HY conceived and designed the experiments. UF, SE, SZ, JS, BN, and HY performed the experiments. UF, KM, and HY analyzed the data. UF, KM, and HY contributed reagents/materials/analysis tools. UF wrote the paper.
Citation: Frevert U, Engelmann S, Zougbédé S, Stange J, Ng B, et al. (2005) Intravital observation of Plasmodium berghei sporozoite infection of the liver. PLoS Biol 3(6): e192.
Abbreviations
ALTalanine aminotransferase
CSPcircumsporozoite protein
ECMextracellular matrix
EEFexoerythrocytic form
GFPgreen fluorescent protein
H&Ehematoxylin-eosin
mAbmonoclonal antibody;
==== Refs
References
Vanderberg JP Frevert U Intravital microscopy demonstrating antibody-mediated immobilization of Plasmodium berghei sporozoites injected into skin by mosquitoes Int J Parasitol 2004 34 991 996 15313126
Cerami C Frevert U Sinnis P Takacs B Clavijo P The basolateral domain of the hepatocyte plasma membrane bears receptors for the circumsporozoite protein of Plasmodium falciparum sporozoites Cell 1992 70 1021 1033 1326407
Ying P Shakibaei M Patankar MS Clavijo P Beavis RC The malaria circumsporozoite protein: Interaction of the conserved regions I and II-plus with heparin-like oligosaccharides in heparan sulfate Exp Parasitol 1997 85 168 182 9030667
Pradel G Garapaty S Frevert U Proteoglycans mediate malaria sporozoite targeting to the liver Mol Microbiol 2002 45 637 651 12139612
Frevert U Sinnis P Cerami C Shreffler W Takacs B Malaria circumsporozoite protein binds to heparan sulfate proteoglycans associated with the surface membrane of hepatocytes J Exp Med 1993 177 1287 1298 8478608
Rathore D McCutchan TF Garboczi DN Toida T Hernaiz MJ Direct measurement of the interactions of glycosaminoglycans and a heparin decasaccharide with the malaria circumsporozoite protein Biochemistry 2001 40 11518 11524 11560500
Ancsin JB Kisilevsky R A binding site for highly sulfated heparan sulfate is identified in the aminojournal-pbio-0030192-terminus of the circumsporozoite protein: Significance for malarial sporozoite attachment to hepatocytes J Biol Chem 2004 279 21824 21832 15007056
Pinzon-Ortiz C Friedman J Esko J Sinnis P The binding of the circumsporozoite protein to cell surface heparan sulfate proteoglycans is required for Plasmodium sporozoite attachment to target cells J Biol Chem 2001 276 26784 26791 11352923
Sinnis P Clavijo P Fenyö D Chait BT Cerami C Structural and functional properties of region-II plus of the malaria circumsporozoite protein J Exp Med 1994 180 297 306 8006589
Lyon M Denkin JA Gallagher JT Liver heparan sulfate structure. A novel molecular design J Biol Chem 1994 269 11208 11215 8157650
Wake K Perisinusoidal stellate cells (fat-storing cells, interstitial cells, lipocytes), their related structure in and around the liver sinusoids, and vitamin A-storing cells in extrahepatic organs Int Rev Cytol 1980 66 303 353 6993411
Gressner AM Schäfer S Comparison of sulphated glycosaminoglycan and hyaluronate synthesis and secretion in cultured hepatocytes, fat storing cells, and Kupffer cells J Clin Chem Clin Biochem 1989 27 141 149 2708943
Soroka CJ Farquhar MG Characterization of a novel heparan sulfate proteoglycan found in the extracellular matrix of liver sinusoids and basement membrane J Cell Biol 1991 113 1231 1241 2040650
Pradel G Frevert U
Plasmodium sporozoites actively enter and pass through Kupffer cells prior to hepatocyte invasion Hepatology 2001 33 1154 1165 11343244
Meis JF Verhave JP Brouwer A Meuwissen JH Electron microscopic studies on the interaction of rat Kupffer cells and Plasmodium berghei sporozoites Z Parasitenkd 1985 71 473 483 3895767
Pradel G Garapaty S Frevert U Kupffer and stellate cell proteoglycans mediate malaria sporozoite targeting to the liver Comp Hepatol 3 (Suppl 1) 2004 S47
Frevert U Sneaking in through the back entrance: The biology of malaria liver stages Trends Parasitol 2004 20 417 424 15324732
Vreden SG The role of Kupffer cells in the clearance of malaria sporozoites from the circulation Parasitol Today 1994 10 304 308 15275428
Vreden SG Sauerwein RW Verhave JP Rooijen N Meuwissen JH Kupffer cell elimination enhances development of liver schizonts of Plasmodium berghei in rats Infect Immun 1993 61 1936 1939 8386704
Sinden RE Smith JE The role of the Kupffer cell in the infection of rodents by sporozoites of Plasmodium Uptake of sporozoites by perfused liver and the establishment of infection in vivo Acta Trop 1982 39 11 27 6122359
Smith JE Alexander J Evasion of macrophage microbicidal mechanisms by mature sporozoites of Plasmodium yoelii yoelii
Parasitology 1986 93 33 38 3748614
Shin SC Vanderberg JP Terzakis JA Direct infection of hepatocytes by sporozoites of Plasmodium berghei
J Protozool 1982 29 448 454 6752394
Ishino T Yano K Chinzei Y Yuda M Cell-passage activity is required for the malarial parasite to cross the liver sinusoidal cell layer PLoS Biol 2: e4 2004 10.1371/journal.pbio.0020004
Ishino T Chinzei H Yuda M A Plasmodium sporozoite protein with a membrane attack complex domain is required for breaching the liver sinusoidal cell layer prior to hepatocyte infection Cell Microbiol 2 2004 E4
Mota MM Pradel G Vanderberg JP Hafalla JC Frevert U Migration of Plasmodium sporozoites through cells before infection Science 2001 291 141 144 11141568
Sinden RE Suhrbier A Davies CS Fleck SL Hodivala K The development and routine application of high-density exoerythrocytic-stage culture of Plasmodium berghei
Bull World Health Organ 68 (Suppl) 1990 115 125
Frischknecht F Baldacci P Martin B Zimmer C Thiberge S Imaging movement of malaria parasites during transmission by Anopheles mosquitoes Cell Microbiol 2004 6 687 694 15186404
Sterling CR Aikawa M Vanderberg JP The passage of Plasmodium berghei sporozoites through the salivary glands of Anopheles stephensi An electron microscope study J Parasitol 1973 59 593 605 4578977
Motoike T Loughna S Perens E Roman BL Liao W Universal GFP reporter for the study of vascular development Genesis 2000 28 75 81 11064424
Faust N Varas F Kelly LM Heck S Graf T Insertion of green fluorescent protein into the lysozyme gene creates mice with green fluorescent granulocytes and macrophages Blood 2000 96 719 726 10887140
Sidjanski S Vanderberg JP Delayed migration of Plasmodium sporozoites from the mosquito bite site to the blood Am J Trop Med Hyg 1997 57 426 429 9347958
Vanderberg JP Nussenzweig RS Most H Further studies on the Plasmodium berghei -Anopheles stephensi -rodent system of mammalian malaria J Parasitol 1968 54 1009 1016 4919047
Hautekeete ML Geerts A The hepatic stellate (Ito) cell: Its role in human liver disease Virchows Arch 1997 430 195 207 9099976
Vanderberg JP Studies on the motility of Plasmodium sporozoites J Protozool 1974 21 527 537 4138523
King CA Cell motility of sporozoan protozoa Parasitol Today 1988 4 315 319 15463014
Sibley LD Intracellular parasite invasion strategies Science 2004 304 248 253 15073368
Hakansson S Morisaki H Heuser J Sibley LD Time-lapse video microscopy of gliding motility in Toxoplasma gondii reveals a novel, biphasic mechanism of cell locomotion Mol Biol Cell 1999 10 3539 3547 10564254
Russell DG Sinden RE The role of the cytoskeleton in the motility of coccidian sporozoites J Cell Sci 1981 50 345 359 7033252
Zieler H Garon CF Fischer ER Shahabuddin M A tubular network associated with the brush-border surface of the Aedes aegypti midgut: Implications for pathogen transmission by mosquitoes J Exp Med 2000 203 1599 1611
Vlachou D Zimmermann T Cantera R Janse CJ Waters AP Real-time, in vivo analysis of malaria ookinete locomotion and mosquito midgut invasion Cell Microbiol 2004 6 671 685 15186403
Meis JF Verhave JP Jap PH Meuwissen JH An ultrastructural study on the role of Kupffer cells in the process of infection by Plasmodium berghei sporozoites in rats Parasitology 1983 86 231 242 6343960
Mota MM Hafalla JC Rodriguez A Migration through host cells activates Plasmodium sporozoites for infection Nat Med 2002 8 1318 1322 12379848
Zieler H Dvorak JA Invasion in vitro of mosquito midgut cells by the malaria parasite proceeds by a conserved mechanism and results in death of the invaded midgut cells Proc Natl Acad Sci U S A 2000 97 11516 11521 11027351
Dubremetz JF Host cell invasion by Toxoplasma gondii
Trends Microbiol 1998 6 27 30 9481821
Nichols BA O'Connor GR Penetration of mouse peritoneal macrophages by the protozoon Toxoplasma gondii
Lab Invest 1981 44 324 335 7206629
Morisaki JH Heuser JE Sibley LD Invasion of Toxoplasma gondii occurs by active penetration of the host cell J Cell Sci 1995 108 2457 2464 7673360
Stewart MJ Vanderberg JP Malaria sporozoites leave behind trails of circumsporozoite protein during gliding motility J Protozool 1988 35 389 393 3054075
Vincke IH Lips M Un nouveau Plasmodium d'un rongeur sauvage du Congo, Plasmodium berghei n. sp Ann Soc Belg Med Trop 1948 28 97 104
Chatterjee S Francois G Druilhe P Timperman G Wéry M Immunity to Plasmodium berghei exoerythrocytic forms derived from irradiated sporozoites Pharmacol Rev 1996 82 297 303
Vanderberg JP
Plasmodium berghei Quantitation of sporozoites injected by mosquitoes feeding on a rodent host Exp Parasitol 1977 42 169 181 324786
Nussenzweig R Herman R Vanderberg JP Yoeli M Most H Studies on sporozoite-induced infections of rodent malaria. III. The course of sporozoite-induced Plasmodium berghei in different hosts Am J Trop Med Hyg 1966 15 684 689 5917626
Scheller LF Stump KC Azad AF
Plasmodium berghei Production and quantitation of hepatic stages derived from irradiated sporozoites in rats and mice J Parasitol 1995 8 58 62
Abecassis M Falk JA Makowka L Dindzans VJ Falk RE 16,16 dimethyl prostaglandin E2 prevents the development of fulminant hepatitis and blocks the induction of monocyte/macrophage procoagulant activity after murine hepatitis virus strain 3 infection J Clin Invest 1987 80 881 889 3624490
Ponnudurai T Lensen AH van Gemert GJ Bolmer MG Meuwissen JH Feeding behaviour and sporozoite ejection by infected Anopheles stephensi
Trans R Soc Trop Med Hyg 1991 85 175 180 1887464
Rosenberg R Wirtz RA Schneider I Burge R An estimation of the number of malaria sporozoites ejected by a feeding mosquito Trans R Soc Trop Med Hyg 1990 84 209 212 2202101
Vaughan JA Scheller LF Wirtz RA Azad AF Infectivity of Plasmodium berghei sporozoites delivered by intravenous inoculation versus mosquito bite: Implications for sporozoite vaccine trials Infect Immun 1999 67 4285 4289 10417207
Hollingdale M Krzych U Perlman P Troye-Blomberg M Immune responses to liver-stage parasites: Implications for vaccine development Malaria immunology, 2nd ed 2002 Basel Karger 97 124
van Rooijen N Sanders A Elimination, blocking, and activation of macrophages: Three of a kind? J Leukocyte Biol 1997 62 702 709 9400810
Stewart MJ Vanderberg JP Electron microscopic analysis of circumsporozoite protein trail formation by gliding malaria sporozoites J Protozool 1992 39 663 671 1453354
Stewart MJ Vanderberg JP Malaria sporozoites release circumsporozoite protein from their apical end and translocate it along their surface J Protozool 1991 38 411 421 1787427
Spaccapelo R Naitza S Robson KJH Crisanti A Thrombospondin-related adhesive protein (TRAP) of Plasmodium berghei and parasite motility Lancet 1997 350 335 9251640
Hügel FU Pradel G Frevert U Release of malaria circumsporozoite protein into the host cell cytoplasm and interaction with ribosomes Mol Biochem Parasitol 1996 81 151 170 8898331
Frevert U Galinski MR Hügel FU Allon N Schreier H Malaria circumsporozoite protein inhibits protein synthesis in mammalian cells EMBO J 1998 17 3816 3826 9669999
Bertolino P McCaughan GW Bowen DG Role of primary intrahepatic T-cell activation in the ‘liver tolerance effect' Immunol Cell Biol 2002 80 84 92 11869365
Knolle PA Gerken G Local control of the immune response in the liver Immunol Rev 2000 174 21 34 10807504
Lohse AW Knolle PA Bilo K Uhrig A Waldmann C Antigen-presenting function and B7 expression of murine sinusoidal endothelial cells and Kupffer cells Gastroenterology 1996 110 1175 1181 8613007
Squiers EC Brunson ME Salomon DR Kupffer cells can present alloantigen in vitro: An effect abrogated by gadolinium J Surg Res 1993 55 571 574 8246488
Freudenberg MA Freudenberg N Galanos C Time course of cellular distribution of endotoxin in liver, lungs and kidneys of rats Br J Exp Pathol 1982 63 56 65 7039654
Lumsden AB Henderson JM Kutner MH Endotoxin levels measured by a chromogenic assay in portal, hepatic and peripheral venous blood in patients with cirrhosis Hepatology 1988 8 232 236 3281884
Cantor HM Dumont AE Hepatic suppression of sensitization to antigen absorbed into the portal system Nature 1967 215 744 745 6059543
Callery MP Kamei T Flye MW Kupffer cell blockade inhibits induction of tolerance by the portal venous route Transplantation 1989 47 1092 1094 2734830
Crispe IN Hepatic T cells and liver tolerance Nat Rev Immunol 2003 3 51 62 12511875
Mowat AM Crispe IN Induction of peripheral tolerance by portal vein administration of antigen T lymphocytes in the liver: Immunobiology, pathology, and host defence 2002 New York Wiley and Sons 101 115
Natarajan R Thathy V Mota MM Hafalla JC Ménard R Fluorescent Plasmodium berghei sporozoites and pre-erythrocytic stages: A new tool to study mosquito and mammalian host interactions with malaria parasites Cell Microbiol 2001 3 371 379 11422080
Knop M Barr F Riedel CG Heckel T Reichel C Improved version of the red fluorescent protein (drFP583/DsRed/RFP) Biotechniques 2002 33 592 602 12238769
Thathy V Menard R Gene targeting in Plasmodium berghei
Methods Mol Med 2002 72 317 331 12125129
Yoshida N Nussenzweig RS Potocnjak P Nussenzweig V Aikawa M Hybridoma produces protective antibodies directed against the sporozoite stage of malaria parasite Science 1980 207 71 73 6985745
Charoenvit Y Leef MF Yuan LF Sedegah M Beaudoin RL Characterization of Plasmodium yoelii monoclonal antibodies directed against stage-specific sporozoite antigens Infect Immun 1987 55 604 608 2434426
King CA Cell surface interaction of the protozoan Gregarina with concanavalin A beads—Implications for models of gregarine gliding Cell Biol Int Rep 1981 5 297 305 6783325
| 15901208 | PMC1135295 | CC BY | 2021-01-05 08:28:14 | no | PLoS Biol. 2005 Jun 24; 3(6):e192 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030192 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1590715510.1371/journal.pbio.0030196Research ArticleBioinformatics/Computational BiologyBotanyEcologyEvolutionGenetics/Genomics/Gene TherapyPlant ScienceStatisticsArabidopsis (Thale Cress)The Pattern of Polymorphism in Arabidopsis thaliana
The Pattern of Polymorphism in A. thalianaNordborg Magnus [email protected]
1
Hu Tina T
1
Ishino Yoko
1
Jhaveri Jinal
1
Toomajian Christopher
1
Zheng Honggang
1
Bakker Erica
2
Calabrese Peter
1
Gladstone Jean
2
Goyal Rana
1
Jakobsson Mattias
3
Kim Sung
1
Morozov Yuri
4
Padhukasahasram Badri
1
Plagnol Vincent
1
Rosenberg Noah A
1
Shah Chitiksha
1
Wall Jeffrey D
1
Wang Jue
2
Zhao Keyan
1
Kalbfleisch Theodore
4
Schulz Vincent
4
Kreitman Martin
2
Bergelson Joy
2
1 Molecular and Computational Biology, University of Southern California, Los Angeles, California, United States of America2 Department of Ecology and Evolution, University of Chicago, Chicago, Illinois, United States of America3 Department of Cell and Organism Biology, Lund University, Lund, Sweden4 Genaissance Pharmaceuticals, New Haven, Connecticut, United States of AmericaMitchell-Olds Tom Academic EditorMax Planck Institute of Chemical EcologyGermany7 2005 24 5 2005 24 5 2005 3 7 e19621 11 2004 31 3 2005 Copyright: © 2005 Nordborg 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.
Patterns of Genetic Variation Reveal Plant's Evolutionary Roots
We resequenced 876 short fragments in a sample of 96 individuals of Arabidopsis thaliana that included stock center accessions as well as a hierarchical sample from natural populations. Although A. thaliana is a selfing weed, the pattern of polymorphism in general agrees with what is expected for a widely distributed, sexually reproducing species. Linkage disequilibrium decays rapidly, within 50 kb. Variation is shared worldwide, although population structure and isolation by distance are evident. The data fail to fit standard neutral models in several ways. There is a genome-wide excess of rare alleles, at least partially due to selection. There is too much variation between genomic regions in the level of polymorphism. The local level of polymorphism is negatively correlated with gene density and positively correlated with segmental duplications. Because the data do not fit theoretical null distributions, attempts to infer natural selection from polymorphism data will require genome-wide surveys of polymorphism in order to identify anomalous regions. Despite this, our data support the utility of A. thaliana as a model for evolutionary functional genomics.
A systematic global survey of genomic DNA sequence polymorphism in Arabidopsis thaliana reveals that standard genetic tests for selection do not apply to this species but supports its status as a model organism.
==== Body
Introduction
The field of population genetics has always been heavily influenced by mathematical models. Ever since molecular polymorphism data started to become available, in the form of allozymes [1] or DNA sequences [2], population geneticists have searched for footprints of selection by comparing the patterns of polymorphism in particular genes with the pattern expected under standard neutral models [3,4]. Considerable intellectual effort has gone into estimating model parameters such as the mutation rate θ, the recombination rate ρ, and the effective population size, Ne [3,5]. However, because of the limited availability of data, it has been difficult to determine whether the underlying models are appropriate. For example, demographic factors such as population structure and growth can cause the genome-wide pattern of variation to deviate from standard neutral models in ways that mimic selection [4,6]. Thus, without knowing whether a standard neutral model describes the pattern of variation in most of the genome, it is difficult to conclude that a particular gene has been under selection.
With the advent of high-throughput genotyping and sequencing, sufficient data for the critical appraisal of standard models are starting to become available, especially in humans [7–9]. Here we report our findings from a systematic survey of genomic DNA sequence polymorphism in Arabidopsis thaliana, one of the first in any organism. Our goal was to investigate the pattern of polymorphism in a large sample of individuals, using sufficiently densely spaced loci to obtain insight into the genome-wide haplotype structure of the species.
The scale of our study allows us to describe the pattern of polymorphism with unprecedented accuracy. We begin by describing how variation is distributed, with respect to space (i.e., population structure) as well as with respect to haplotypes (i.e., linkage disequilibrium [LD]). Our set of 96 individuals contained hierarchical population samples in addition to a worldwide collection of stock center accessions (Tables S1 and S2): because of this and the large number of polymorphisms, we are able to answer a number of questions about population structure that previous studies have not been able to address.
In the second part of the paper, we compare the pattern of variation to predictions made by standard population genetics models. The number of loci sequenced is sufficient to investigate the distribution of important summary statistics across the genome rather than simply looking at averages (as is usually done).
Results/Discussion
Sequencing
A total of 876 reliable alignments, representing 0.48 Mbp of the genome (or a total over all individuals of ~44 Mbp) was generated. The average sequence length is 583 bp; the average sample size across alignments is 89. Based on the A. thaliana genome annotation, the composition of our data, which includes more than 17,000 single nucleotide polymorphisms (SNPs) and insertion/deletion polymorphisms, is 15% intergenic, 55% exon, 22% intron, 4% UTR, and 5% pseudogene (see Materials and Methods). The majority of fragments, 67%, contain both coding and noncoding sequence.
Population Structure
Overall levels of polymorphism
Our estimates of the level of polymorphism are broadly comparable to what has previously been found in A. thaliana and other species, both in terms of overall levels of polymorphism, and in the degree of constraint on different kinds of sites (Figure 1; cf. [3]). The highly selfing A. thaliana does not have unusually low levels of polymorphism: the observed values are somewhat lower than for Drosophila melanogaster, and are considerably higher than for humans.
Figure 1 Levels of Polymorphism for Different Classes of Sites
Levels of polymorphism were quantified using two different estimators of the neutral mutation rate θ: θ^S
, which uses the number of polymorphic sites, and θ^P
, which uses the average number of pairwise differences [3].
A standard way of summarizing the geographical distribution of this variation is through the statistic FST, which, loosely speaking, measures the fraction of the observed genetic variation that is due to population structure [10]. Our sample contains 40 individuals that were hierarchically sampled in pairs from four populations in each of five regions (Table S1). A hierarchical analysis of variation in these individuals reveals that 33% of the global variation is segregating among individuals within populations, 35% is segregating among local populations within regions, and 26% is segregating among regions. Only 6% of the variation in the global sample is not captured by these 40 individuals. Even though only two individuals were sampled per population, and even though our estimates of within-population variance are upwardly biased (individuals were prescreened to avoid sequencing identical individuals; see Materials and Methods), our data clearly show that individual populations harbor much of the variation present species-wide. At the same time, there is strong population structure.
Global geographic structure
Studies of variation in A. thaliana have typically not found any correlation between the genotype and geographic origin of accessions. This has been attributed to a recent expansion of the species, perhaps in combination with human disturbance. However, early studies had little power to detect population structure, and a more recent, larger survey revealed weak isolation by distance [11]. Our study has several orders of magnitude more markers than any previous study known to us, and we find clear global population structure.
We used a model-based clustering algorithm, implemented in the computer program Structure, to cluster our accessions on the basis of genotype [12]. Loosely speaking, the algorithm attempts to identify a predetermined number of clusters, K, that have distinctive allele frequencies, and assigns portions of individual genomes to these clusters. A genome can have membership in several clusters, and the algorithm reports the probability distribution of the assignment of each section of the genome.
We analyzed the data by successively increasing K from two to eight (Figure 2). For K = 2, we see an East–West gradient, potentially attributable to post-glaciation colonization routes [11]. When we increaseK to three, all accessions from northern Sweden and Finland are assigned to a single cluster together with (to varying degrees) accessions from Eastern Europe, Russia, and Central Asia. While a relationship between northern Sweden and Tajikistan may seem far-fetched, several species are known to have colonized the Scandinavian Peninsula from both north (from Russia via Finland) and south (from Europe via Denmark) after the last glaciation [13].
Figure 2 Population Structure and Genomic Distributions of Various Statistics
(A) Results from Structure under different assumptions about the number of clusters (K = 2,…, 8). Each individual is represented by a line, which is partitioned into K colored segments according to the individual's estimated membership fractions in each of the K clusters. The assignment of each individual is the average across the genome.
(B) Results from Structure across Chromosome 1 for K = 3. Each chromosomal segment is colored according to the cluster in which it had the highest probability of membership.
(C) A plot showing those fragments that appear to be monophyletic with respect to each of the three clusters identified by Structure.
(D) FST with respect to the same three clusters (blue solid line) and the lower 95th percentile of FST obtained through 1,000 random permutations of the accessions (red dotted line).
(E) θ^P
within each of the three clusters.
(F) Tajima's D statistic within each of the three clusters.
(G) Results from Structure across Chromosome 1 for K = 8.
(H) A plot showing those fragments that appear to be monophyletic with respect to each of these eight clusters.
(I) FST with respect to these eight clusters.
As we increase K from three to eight, each new cluster splits a previously existing one along plausible geographical boundaries (and identifies some accessions as mixed). Thus K = 4 separates central European (Czech, Austrian, and Croatian) accessions from the main European cluster, K = 5 separates a subset of the United States accessions from the rest of Western Europe, K = 6 separates the Central Asian and Russian accessions from northern Sweden and Finland, K = 7 separates many German and southern Swedish accessions from the rest of Europe, and K = 8 identifies a Catalan cluster. Clusters identified for K > 8 did not contain the majority of any single individual genome.
When these results are superimposed on a map (Figure 3), the pattern of isolation by distance becomes obvious. Individuals are, by and large, more similar to individuals that grow nearby than to individuals from far away. Although A. thaliana commonly occurs as a weed and human commensal, this has not been sufficient to erase population structure.
Figure 3 Population Structure in A. thaliana
Each pie chart represents an accession, and is placed on the map according to origin (some of the population samples were too densely sampled and have been shifted for clarity). Accessions sampled outside Europe have been placed at the correct latitude. The exact origin of the standard lab accession Col-0 is not known. The colors and proportions within each pie chart correspond to the output of Structure in Figure 2. (A) K = 3; (B) K = 8.
Population structure is also evident in the distribution of pairwise differences between individuals. In the absence of population structure, all individuals should be equally closely related on average. As shown in Figure 4, this is clearly not the case. Not only is there generally a wider range of variation than would be expected in the absence of population structure, but there are also clear outliers. Some individuals are extremely closely related: these are typically from the same local population (and will be further discussed below). Perhaps more surprising is that two stock center accessions, Cvi-0 (Cape Verde Islands) (cf. [14]) and Mr-0 (Italy), are very different from all others (and each other).
Figure 4 The Distribution of Pairwise Differences (SNPs Only) between All Pairs of Accessions
(A) An example of the distribution we would expect to see in the absence of population structure, obtained by randomizing genotypes with respect to individuals for each sequenced fragment.
(B) The observed distribution.
(C) The observed distribution with accessions Cvi-0 and Mr-0 removed.
The distribution of pairwise differences can be conveniently summarized using hierarchical clustering. The population structure revealed by the resulting tree (Figure S1) generally agrees with the output of Structure.
Variation within and between populations
Because it is highly selfing, A. thaliana has often been considered a collection of asexual lineages, or “ecotypes.” This view is completely false. Indeed, even the first sequencing survey of variation showed clear evidence of recombination between accessions [15], and a subsequent study has shown that recombination has generally been sufficient to erode genome-wide LD on a very fine scale [16]. Consequently, there is no “phylogeny” of ecotypes.
However, this still leaves open the possibility that local population structure is tree-like, with individuals within the same population being much more closely related to each other than to individuals from other populations. Indeed, since A. thaliana is a selfer, it is perfectly possible for a local population to consist of a single inbred sibship.
We find that this is typically not the case; most sampled populations were polymorphic (even though this part of our study had relatively little power; see Materials and Methods), and when we consider the genome-wide data, the pattern that emerges is far from tree-like. As shown in Figure 2C and 2H, only a small fraction of all sequenced fragments have patterns of polymorphism compatible with monophyly in the sense that, with respect to that fragment, the members of a particular cluster are all more closely related to each other than to members of other clusters (the yellow “US” cluster is an exception to which we return below). Instead, just as is the case for human populations [17], most polymorphisms are shared between clusters, and levels of polymorphism are in general comparable within all clusters (see Figure 2E; for clarity, only values for K = 3 are shown).
The same is true with respect to the local populations, although there is great variation between regions. An interesting way to describe the relationship between individuals is to try to identify chromosomal regions shared identical by descent by looking for long identical haplotypes [18–22]. The resulting patterns clearly reveal the difference in structure between geographic regions (Figure 5). In northern Sweden, pairs of accessions sampled from the same population are invariably more closely related to each other than to accessions from other populations; however, even here populations are far from monophyletic in the sense used above (i.e., for a particular locus, the closest relative may well come from a different population). In most regions (exemplified in Figure 5 by central Europe), haplotype sharing is moderate, and is not much greater within than between populations. The US sample is again different: here, pairs of accessions often appear to share entire chromosomes, and are equally likely to do so between populations.
Figure 5 Haplotype Sharing on Chromosome 4 among Pairs of Individuals in the Population Samples from Northern Sweden, Central Europe, and the US
The lines indicate regions where the particular pair of accessions share at least five identical adjacent fragments. Within-population comparisons are highlighted in red. The patterns in southern Sweden and in the UK are similar to that in central Europe.
Regional variation in the level of population structure is also evident in the distribution of pairwise differences (see Figures 4 and S1). Individuals that are extremely closely related (less than ten differences) are almost always pairs from the same local population (two pairs from northern Sweden, one pair from Finland, and one from Germany). The one exception is a trio of nearly identical individuals from different US populations.
Figure 2D and 2I illustrate the variation in FST across the genome for K = 3 and K = 8. This pattern is of interest in that regions with extremely high or low values may be seen as candidates for harboring selectively important loci [23,24].
Structure summarized
The picture that emerges is that of a single, large, globally distributed population with historical gene flow sufficient to ensure that variation is shared worldwide, yet limited enough to cause considerable population structure. Genetic exchange is not only geographic; there has been enough outcrossing to ensure that LD decays within 25–50 kb on average (Figure 6), which is comparable to what has been observed in humans. All of this may seem surprising given the highly selfing nature of A. thaliana, but it is in fact completely compatible with theoretical predictions [16,25].
Figure 6 The Decay of LD as a Function of Distance between the Polymorphisms
The only exception to this pattern comes from the US. Our sample from the US Midwest is clearly a heterogeneous collection, characterized by genome-wide LD and haplotype sharing. Especially notable is extensive haplotype sharing with accessions from other regions, in particular with the United Kingdom. All this strongly suggests that A. thaliana is a recent human introduction to the New World, and that the introduction severely reduced haplotype variation through bottleneck effects, causing genome-wide LD [16]. Since we have population samples from only the Midwest, we cannot rule out the possibility that the pattern is different in other parts of the US; however, we note that our one non-Midwestern US accession, Yo-0 (from Yosemite, California), is almost identical to some of our Midwestern accessions, and that a recent survey of variation in 53 US populations found no evidence of differentiation across the continent [26].
It should be emphasized that we view both Structure and hierarchical clustering as tools for exploring the data. The results should not be taken literally. For example, we do not believe that there are K = 8 random-mating populations in A. thaliana, as might be suggested by Figures 2 and 3, nor do we believe that populations are related in a tree-like manner, as might be suggested by Figure S1.
Global Patterns
Allele frequency distribution
Note that θ^S
is consistently higher than θ^P
(see Figure 1). This is typically caused by an unusually high ratio of rare to common alleles, compared to what is expected under standard population genetics models [27]. A closer look at the distribution of allele frequencies reveals that this is indeed the case (Figure 7A). The observed distribution is skewed toward rare alleles compared to standard neutral expectations. A possible explanation for this is recent population growth [28]; however, the skew is much greater for nonsynonymous than for synonymous polymorphisms, suggesting that selective factors must also be involved.
Figure 7 Characteristics of the Pattern of Polymorphism
(A) The allele frequency distribution for synonymous and nonsynonymous SNPs using a sample size of 90 individuals (loci with less than 90 individuals were not used; loci with greater than 90 individuals were randomly culled). For a sample of size n, the expected frequency of SNP loci with a minor allele frequency of i under a standard constant-size population genetics model is
. The excess of rare alleles is largely limited to frequencies one and two.
(B) The distribution of Tajima's D statistic [27] across the sequenced fragments, along with its expected distribution in a constant population (estimated by simulating 1,000 datasets matching the real one in terms of exon/nonexon composition and sample size).
(C) The distribution of the level of polymorphism (θ^S
) across the sequenced fragments along with its expected distribution (estimated the same way).
(D) The level of polymorphism in nonexon sequences as a function of the local gene density (measured in open reading frames per centimorgan).
(E) The level of polymorphism in nonexon sequences as a function of the degree of duplication in each fragment (measured as the negative log10 of the BLAST significance for the second-best hit in the genome).
The patterns in (D) and (E) are also seen in exons.
The effect of this genome-wide deviation from standard neutral models on “tests of selection” can be dramatic. These tests typically assume that the standard neutral model describes most of the genome, and interpret deviations at particular loci as signs of selection [3]. Our results show clearly that this procedure is not appropriate for A. thaliana (see also [29]).
It is, of course, not a new finding that demographic history can invalidate tests of selection, but our results provide a striking illustration of the potential seriousness of the problem. For example, the mean value of one popular statistic, Tajima's D [27], is −0.8 rather than the (approximately) zero expected under simple neutral models, and the variance is also larger than predicted (Figure 7B). Positive values of Tajima's D are typically attributed to balancing selection: 2% of our fragments are significantly positive at the 1% level. Negative values are typically attributed to directional selection: 15% of our fragments are significantly negative at the 1% level. Although some of these deviations may, of course, actually be due to selection, our data suggest that tests based on standard cutoff values are anticonservative in both tails of the distribution. Consistent with this interpretation, a much higher fraction of studied genes have been reported to be under selection in A. thaliana than in other species [30].
Variation in the level of polymorphism
While it is straightforward to fit a neutral model with growth to the observed distribution of Tajima's D (Figure 7B), the value of this exercise is doubtful. First, it is clear from Figure 7A that selection must be part of the reason for the skewed allele frequency distribution. Second, a model with growth would, in fact, fit other aspects of the data less well. In particular, population growth tends to reduce the variability in coalescence times across the genome compared to models with constant population size, resulting in less variation in the level of polymorphism between loci. We see the opposite: the variance between loci is considerably greater than expected under a standard neutral model with constant population size (Figure 7C). In addition, the distribution is heavily skewed and displays a long tail of extremely high values.
There are several reasons to expect a poor fit to a simple neutral model. One is variation across the genome in the “neutral” mutation rate, θ, either due to variation in the level of selective constraint or due to variation in the actual, underlying mutation rate. Since the excess variability is observed equally for coding and noncoding DNA, we would have to invoke the latter. The extent to which variation in the mutation rate contributes to the pattern in Figure 7C can be estimated as soon as divergence data from a closely related outgroup species become available.
Another factor likely to contribute to the variation in the level of polymorphism is population structure such as that described in the first part of the paper. It is well known that population structure can inflate the variance of coalescence times as well as induce a tail of very large values that would result in patterns of variability such as the one observed [6]. However, strong population structure is generally expected to push the distribution of Tajima's D toward positive values, which is the opposite of what we observe. It is possible that some model that involves growth (perhaps in conjunction with bottlenecks), structure, and finite sampling [7] could explain the pattern observed in A. thaliana, but we have not been able to find such a model (see also [29]).
Genomic patterns of polymorphism
It turns out that the pattern of polymorphism is affected by not only demographic forces, but also factors intrinsic to the genome. Figure 7D shows that polymorphism in noncoding regions is negatively correlated with local gene density. This mirrors the positive correlation between polymorphism and local recombination rates that was first noted in Drosophila [31] and has since been observed in a wide range of organisms [32]. A possible explanation is that recombination itself is mutagenic: this appears to explain the correlation observed in humans, but is not sufficient to explain the phenomenon in general [32,33]. Instead, it has been proposed that variation is reduced in low-recombination regions because of a “hitchhiking effect” due to selection on linked sites [34,35], in the form of either positive selection (“selective sweeps”) [36] or purifying selection (“background selection”) [37]. Such hitchhiking effects would be stronger in low-recombination regions because sites in these regions are affected by selection on larger pieces of the chromosome (i.e., more genes). Recombination is thus used as a proxy for gene density: the real prediction of these models is that polymorphism should decrease with gene density. This is precisely what we observe. The level of polymorphism is insignificantly positively correlated with recombination (suggesting that although recombination may well be mutagenic, this does not explain the phenomenon), but is strongly negatively correlated with gene density.
Two factors suggest that background selection rather than selective sweeps is responsible for the correlation. First, unlike background selection, selective sweeps are expected to skew Tajima's D toward negative values. However, we find no correlation between Tajima's D and gene density. Second, it is clear from Figure 7A that a significant load of deleterious mutations (as is required by background selection) exists in A. thaliana.
Figure 7E reveals that polymorphism is also positively correlated with segmental duplication, similar to what has been observed in humans [38–40]. In humans, the phenomenon appears to be largely due to misclassification of paralogous sequence variants as SNPs. Distinguishing between paralogous sequence variants and true SNPs is difficult for human data, where putative SNPs are typically detected in small samples of highly heterozygous individuals. In contrast, our data consist of high-quality sequences from a large sample of almost completely homozygous individuals, and we are therefore confident that nearly all of our polymorphisms are genuine (see Materials and Methods); fragments that have a close match elsewhere in the genome thus appear to be more variable than fragments that do not. We hypothesize that this is caused by a low level of intergenic gene conversion that serves to “shuffle” variation between loci. Such gene conversion has long been known to occur in large multigene families (“concerted evolution”; [41]): our results suggest that it may be a general phenomenon.
Recombination and gene conversion
We noted above (see Figure 6) that LD decays within 25–50 kb, somewhat faster than has previously been suggested [16]. At least 25% of our sequenced fragments show evidence of recombination (using the four-gamete test; [42]). Estimates based on coalescent models suggest an effective population recombination rate (e.g., [6]) of approximately ρ = 2 × 10−4 per basepair (V. P., B. P., P. Marjoram, J. W., and M. N., unpublished data). Given our estimates of the mutation rate θ (see Figure 1), this implies a ratio θ/ρ of about 20.
The short-range pattern of LD in several species is incompatible with the long-range pattern; there is too little of the former relative to the latter for a simple recombination model that includes only crossing over to explain the data [43–48]. Possible explanations include gene conversion and multiple mutations (i.e., each SNP not being due to a unique mutation event), both of which will erode short-range LD [46]. There is clear evidence for both phenomena in our data. We observed a total of 315 tri-allelic SNPs. Since less than 50% of all multiple-hit mutations will result in more than two distinct alleles, this suggests that a total of more than 600 of our SNPs are, in fact, the product of multiple mutations. We also observed three fragments that show clear evidence for gene conversion in that a single gene conversion event (i.e., a double cross-over within 500 bp, as would result from the resolution of a single Holiday junction) suffices to explain a complicated pattern of polymorphisms based on multiple SNPs in the fragment. Coalescent-based analyses based on the fine-scale pattern of polymorphism suggest that gene conversion is about five times more common than crossing over, in agreement with previous population genetic analyses [48], as well as with direct estimates based on tetrad analysis [49].
Concluding Remarks
We have shown that the pattern of polymorphism in A. thaliana, a selfing human commensal, generally agrees with what would be expected for a widely distributed sexually reproducing species. Although there is significant population structure, polymorphism is shared worldwide. As predicted by population genetics theory [25], the only clear indications of selfing in the pattern of polymorphism are that individuals are typically homozygous, and that LD is unusually extensive.
The scale of our study allows us to consider the genomic distribution of statistics commonly used to summarize polymorphism data. We find that these distributions generally deviate significantly from what is assumed by standard population genetics models. This highlights the danger of using highly parameterized models based on untested assumptions for inference in population genetics. Commonly used “tests of selection” are simply not valid in A. thaliana (cf. [29,30]). Large-scale analyses in other organisms have similarly found genome-wide deviations from standard models (e.g., [43,50,51]). As data continue to accumulate, the focus of population geneticists will surely have to shift from rejecting null models that do not fit particular loci to finding models that actually do explain the bulk of the data.
Genomic polymorphism data are required to develop more robust inference methods, and will enable us to study phenomena that are intrinsically genomic (e.g., the correlations in Figure 7D and 7E). More importantly, however, these data will help identify the functional polymorphisms that underlie phenotypic variation. The pattern of polymorphism in A. thaliana, characterized by humanlike levels of LD but much higher SNP density, coupled with the availability of naturally occurring inbred lines, makes the species ideal for LD mapping. Although the strong population structure is likely to cause a high rate of spurious genotype–phenotype associations, these problems can easily be overcome through direct experimental verification using crosses or transgenics. This significantly strengthens the position of A. thaliana as a model for evolutionary functional genomics.
Materials and Methods
Sampling
The sample of 96 individuals included pairs of individuals from 25 local “populations” (typically sampled within a few hundred meters of each other, often much closer) as well as a worldwide survey of commonly used stock center accessions (Tables S1 and S2). Where possible, four populations were sampled from each of several regions.
The sample was generated by screening a larger set of accessions with a small number of markers to avoid inbred siblings or extensively heterozygous individuals (E. B., E. Stahl, C. T., M. N., M. K., and J. B., unpublished data). Accessions were genotyped using 11 unlinked markers (five microsatellites, two indel R-genes, and four housekeeping genes with previously identified polymorphisms). To ensure that individuals sampled from local populations were not part of inbred sibships, four (three in one case) individuals from each of 37 populations were tested. Polymorphism was found in 25 of these populations, and a pair of nonidentical individuals was selected at random from each (Table S1). Some accessions not from the same population were also found to be identical with respect to these markers (Col-0 and Lp2-2; Ts-1 and Shahdara), but these were included nonetheless. Five accessions were found to be heterozygous and were eliminated. Four of these were from the population samples, and one, Ms-0, was from the stock center. Further testing of two additional Ms-0 lines revealed one more heterozygote and one homozygote, which was included. In spite of these precautions, one sequenced stock center accession, Van-0, turned out to be extensively heterozygous and was eliminated from the analyses in this paper (bringing the sample size to 95).
Data generation
We used direct, PCR-based sequencing of genomic DNA, with primers designed from the A. thaliana reference sequence to cover the genome relatively uniformly. To achieve uniform density of our fragments, the reference genome (releases January 7, 2002, and April 17, 2003) was first divided into equally spaced regions. The last 10 kb of each region then served as an input record to Primer3 (v. 0.6). The designed primer pairs returned from Primer3 for each region were then screened for uniqueness and quality. To screen for uniqueness, all primer pairs were BLASTed (BLAST v. 2.2.3) against both the reference genome as well as BAC datasets downloaded from the Arabidopsis Information Resource (http://www.arabidopsis.org/). Any primer pair that produced a hit in the same region (≤2,300 bps) was removed. Self-amplifying primers were also removed based on this same criterion. Additionally, primers with more than five BLAST hits against the reference were also discarded. To improve the quality of each fragment, any primer pair that amplified a target sequence that contained a homonucleotide run of nine bases or more was removed.
All sequencing was done using ABI 3700 automated sequencers (Applied Biosystems, Foster City, California, United States). All fragments were sequenced in both directions.
Chromatograms were initially base-called with Phred (v. 0.020425.c) and trimmed based on quality value. The start and end of each read was trimmed until the average quality value was 25 in a window of ten bases, and internal bases were converted to missing data when their quality value was below ten. Accessions missing one read of data were trimmed more severely (different setting were used). A combination of Phrap (v. 0.020425.c) and ClustalW (v. 1.82) was used for producing alignments using a modified weight matrix that allowed us to incorporate quality values into the ClustalW algorithm. Alignments were then visually inspected and adjusted as necessary using Consed (v. 13.0). Polyphred (v. 4.20) was used to flag potential heterozygotes, which were confirmed by visual inspection of chromatograms.
Additional trimming was performed as necessary for accessions with multiple false polymorphisms and low-quality sequence after a visual inspection of chromatograms. Whenever two reads from the same accession disagreed, the final call was made by visually inspecting chromatograms unless the difference in quality value made the final call obvious.
Potential polymorphisms in each alignment were then verified by a second person. All alleles found only once or twice in the sample were verified by visually inspecting the chromatograms. When this inspection did not reveal a chromatogram peak clearly different from the other accessions, the base was changed to missing data. This would, if anything, produce a slight underestimate in alleles of frequency one and two. Higher-frequency polymorphisms with generally low-quality values (20 or lower) were also verified by checking the chromatograms.
A total of 876 high-quality fragment alignments were obtained from 979 PCR primers and used for the analyses in this paper. Of the remaining PCR primers, some failed at the stage of PCR amplification and sequencing, while some produced sequencing output that could not be base-called with certainty when the sequence quality was particularly low or when there was evidence that the primer pairs amplified two or more different products in some of the accessions.
To calculate genetic distances, we used a set of markers that have been genetically mapped to the Lister and Dean recombinant inbred lines and that also can be mapped to the AGI reference genome. Some markers were removed so that both physical position and genetic position were monotonically increasing functions.
All data are publicly available through our Web site (http://walnut.usc.edu/2010, and also as Dataset S1.
Population structure
To infer population structure and assign accessions to populations, we used a model-based clustering algorithm implemented in Structure v. 2.0 [12]. Since A. thaliana is largely homozygous, we used a haploid setting. We used the “linkage model” with “correlated allele frequencies” in Structure, where genetic distances (calculated by fitting a third-order polynomial to the Lister and Dean recombinant inbred mapping data) were used to indicate locus proximity. The algorithm was run with a burn-in length of 50,000 MCMC iterations and then 20,000 iterations for estimating the parameters. This was repeated ten times for each K (ranging from one to 17). In these analyses, each fragment-haplotype was treated as a marker at a multiallelic locus, so that two accessions had a different type if they differed at any site in the fragment.
The likelihood of the data increases with K from K = 1 until K = 7 (using the Wilcoxon two-sample test to compare the ten runs for each K; two-sided p = 0.001 for K = 7 versus K = 6). The likelihoods of K = 7 and K = 8 were similar (two-sided p = 0.97). For K > 7, the likelihoods of different runs were more variable than for K ≤ 7, with the added variability caused only by runs with lower likelihoods. Moreover, the additional clusters for K > 8 do not have a majority of the genome for any of the accessions. These observations taken together indicate that it is less meaningful to choose K > 8.
In displaying the output from Structure, we computed an average of the ten runs for each K. Because there are K! distinct permutations of the clusters that all correspond to equivalent assignments of membership coefficients to accessions, and because independent runs may produce different permutations, to compute an average we first permuted the clusters to align the solutions. For R runs, there are (K!)R
− 1 ways of aligning clusters across runs. To determine which of the clusters of each of the other runs corresponds to a specific cluster in a given run, the symmetric similarity coefficient (SSC) was used with the matrices of membership coefficients (based on the genome-wide average). For a given K, the SSC was calculated for all combinations of pairs of runs:
where Qi and Qj are the membership matrices of runs i and j (i ≠ j), P is a permutation; the minimum is taken over all permutations, S is a probability matrix of K columns where all elements equal 1/K, and A F is the Frobenius matrix norm [52]. This is a slight adaptation of the asymmetric similarity coefficient used in previous work [17].
For K = 2, the runs were permuted to the arrangement that maximizes the sum of SSC across pairs of runs, and an average of the membership matrices across runs was then taken. For K > 2, it was not feasible to test all possible arrangements; therefore, the following greedy algorithm was used. (1) Fix a permutation, P
1, of one (randomly chosen) run,
. (2) Randomly choose a second run, Q2, and fix the permutation, P
2, that maximizes
. (3) Continue sequentially with each remaining run, Qx, where x = 3,…, R, and fix the permutation, Px, that maximizes
for the current run, Qx. Because the choice of starting run can affect the result, we tested all ten possibilities for the starting run. For K = 2 to K = 8, there were thus 70 possible ways of starting the algorithm, and in only two of 70 possible cases was a different result obtained. These two solutions differed from the common solution by switching one pair of clusters in one run (2.5% of the clusters differed from the common solution), and switching one pair of clusters in two different runs (5%).
We tested for monophyly as follows. For every variable site in a fragment, each cluster was checked for the presence of both alleles as well as for the presence of both alleles outside the cluster. If a variable site in a fragment had both alleles within the cluster as well as outside the cluster, then the whole fragment was deemed nonmonophyletic for that specific cluster. Clusters that failed to show nonmonophyly for a fragment were considered monophyletic for that fragment. Fragments with less than five variable sites and clusters with less than five accessions were always considered to be nonmonophyletic.
FST for the inferred clusters was computed as:
where θ^P, total
is the average number of pairwise differences per site for all pairs of accessions, and θ^P,
within
is the average number of pairwise differences per site for all pairs within cluster i.
Of the total of 95 accessions, 40 were hierarchically sampled in pairs from four populations in each of five regions (Table S1). The total amount of variation among these 40 accessions, θ^P,
among40
, was computed by taking the total average pairwise difference for all pairs of the 40 accessions, whereas the amount of variation within populations, θ^P,
withinpop
, was calculated by taking the mean of the total average pairwise difference for the pairs of accessions in the 20 populations. The level of variation among geographical regions, θ^P,
amongreg
, was computed as the difference between θ^P,
among40
and the mean of the total average pairwise differences for all pairs of accessions within regions. The level of variation among populations, θ^P,
amongpop
, was calculated from the following expression:
Genomic patterns of polymorphism
Correlations were identified between levels of polymorphism and local gene density or degree of duplication (Figure 7). The local gene density was measured as open reading frames per centimorgan in windows of size greater than or equal to 1 Mb (using genetically mapped markers from the Lister and Dean recombinant inbred data as endpoints). The number of open reading frames (excluding pseudogenes and RNA genes) from the annotated reference sequence that fell between these window endpoints was counted, and length in centimorgans of each window was estimated from the genetic distance of the markers used as window endpoints.
Correlations were quantified using Spearman's rank correlation, and the significance of the observed values was evaluated using 50,000 permutations that maintained the chromosomal order of all observations but that shuffled the relative positions of the two variables. (For each variable, the lists representing the consecutive values within each chromosome were concatenated in random order and direction to form a circle. The two circles were then randomly aligned with each other.) This is necessary to avoid inflated significance values due to autocorrelations along the chromosomes (of both variables). Using this procedure, the rank correlation between θ^S
in nonexon sequences and gene density is −0.27 (p = 0.0014), and the rank correlation between θ^S
in nonexon sequences and the negative log of the second-best BLAST e-value is 0.13 (p = 0.0018).
To investigate the effect of population structure, all analyses (except those of population structure) were repeated with the outliers in Figure 4 removed (Cvi-0, Mr-0, and all but one randomly chosen member of each closely related group). All conclusions remain qualitatively the same.
Supporting Information
Dataset S1 All Data Used in the Paper
(912 KB ZIP).
Click here for additional data file.
Figure S1 Hierarchical Clustering of Individuals Based on Pairwise Differences
(12 MB EPS).
Click here for additional data file.
Table S1 The Population Samples Used in the Project
(8 KB PDF).
Click here for additional data file.
Table S2 The Individual Accessions Used in the Project
(8 KB PDF).
Click here for additional data file.
This work was mainly supported by the National Science Foundation (NSF 2010 grant DEB-0115062 to MN, MK, and JB). In addition, HZ was supported by a grant from the Fletcher Jones Foundation (Simon Tavaré, PI), EB by a fellowship from the Dropkin Foundation, NAR by a Burroughs-Wellcome Fund Career Award in the Biomedical Sciences, and CT by a National Institutes of Health postdoctoral grant. C. Alonso-Blanco, I. Al-Shebaz, J. Borevitz, I. Cetl, S. Holm, M. Jönsson, M. Koch, J. Relichová, G. Röbbelen, T. Säll, and S. Somerville contributed seed or helped with seed collection.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. M. Nordborg, M. Kreitman, and J. Bergelson conceived and designed the experiments. T. T. Hu, Y. Ishino, J. Jhaveri, C. Toomajian, H. Zheng, E. Bakker, J. Gladstone, Y. Morozov, C. Shah, J. Wang, T. Kalbfleisch, and V. Schulz performed the experiments and contributed reagents/materials/analysis tools. M. Nordborg, T. T. Hu, C. Toomajian, E. Bakker, P. Calabrese, M. Jakobsson, S. Kim, B. Padhukasahasram, V. Plagnol, N. A. Rosenberg, J. Wall, K. Zhao, M. Kalbfleisch, M. Kreitman, and J. Bergelson analyzed the data and wrote the paper.
Citation: Nordborg M, Hu TT, Ishino Y, Jhaveri J, Toomajian C, et al. (2005) The pattern of polymorphism in Arabidopsis thaliana. PLoS Biol 3(7): e196.
Abbreviations
LDlinkage disequilibrium
SNPsingle nucleotide polymorphism
SSCsymmetric similarity coefficient
==== Refs
References
Lewontin RC Hubby JL A molecular approach to the study of genetic heterozygosity in natural populations. II. Amount of variation and degree of heterozygosity in natural populations of Drosophila pseudoobscura
Genetics 1966 54 595 609 5968643
Kreitman M Nucleotide polymorphism at the alcohol dehydrogenase locus of Drosophila melanogaster
Nature 1983 304 412 417 6410283
Li WH Molecular evolution 1997 Sunderland (Massachusetts) Sinauer Associates 487
Kreitman M Methods to detect selection in populations with applications to the human Annu Rev Genomics Hum Genet 2000 1 539 559 11701640
Stephens M Balding DJ Bishop MJ Cannings C Inference under the coalescent Handbook of statistical genetics 2001 Chichester (United Kingdom) John Wiley and Sons 213 238
Nordborg M Balding DJ Bishop MJ Cannings C Coalescent theory Handbook of statistical genetics 2001 Chichester (United Kingdom) John Wiley and Sons 179 212
Ptak SE Przeworski M Evidence for population growth in humans is confounded by fine-scale population structure Trends Genet 2002 18 559 563 12414185
Innan H Padhukasahasram B Nordborg M The pattern of polymorphism on human chromosome 21 Genome Res 2003 13 1158 1168 12799351
Wall JD Pritchard JK Haplotype blocks and linkage disequilibrium in the human genome Nature Rev Genet 2003 4 587 597 12897771
Excoffier L Balding DJ Bishop MJ Cannings C Analysis of population subdivision Handbook of statistical genetics, 2nd ed 2003 Chichester (United Kingdom) John Wiley and Sons 713 750
Sharbel TF Haubold B Mitchell-Olds T Genetics isolation by distance in Arabidopsis thaliana Biogeography and postglacial colonization of Europe Mol Ecol 2000 9 2109 2118 11123622
Falush D Stephens M Pritchard JK Inference of population structure using multilocus genotype data: Linked loci and correlated allele frequencies Genetics 2003 164 1567 1587 12930761
Andersen BG Borns HW The ice age world 1997 Oslo Scandinavian University Press 208
Schmid KJ Rosleff Sörensen T Stracke R Törjék O Altmann T Large-scale identification and analysis of genome-wide single-nucleotide polymorphisms for mapping in Arabidopsis thaliana
Genome Res 2003 13 1250 1257 12799357
Hanfstingl U Berry A Kellogg E Costa JT Rüdiger W Haplotypic divergence coupled with lack of diversity at the Arabidopsis thaliana alcohol dehydrogenase locus: Roles for balancing and directional selection Genetics 1994 138 811 828 7851777
Nordborg M Borevitz JO Bergelson J Berry CC Chory J The extent of linkage disequilibrium in Arabidopsis thaliana
Nature Genet 2002 30 190 193 11780140
Rosenberg NA Pritchard JK Weber JL Cann HM Kidd KK Genetic structure of human populations Science 2002 298 2381 2385 12493913
Franklin IR The distribution of the proportion of the genome which is homozygous by descent in inbred individuals Theor Popul Biol 1977 11 60 80 404725
Stam P The distribution of the fraction of the genome identical by descent in finite random mating populations Genet Res 1980 35 131 155
Donnelly KP The probability that related individuals share some section of the genome identical by descent Theor Popul Biol 1983 23 34 63 6857549
Broman KW Weber JL Long homozygous chromosomal segments in reference families from the Centre d'Étude du Polymorphisme Humain Am J Hum Genet 1999 65 1493 1500 10577902
Clark AG The size distribution of homozygous segments in the human genome Am J Hum Genet 1999 65 1489 1492 10577901
Lewontin RC Krakauer J Distribution of gene frequency as a test of the theory of the selective neutrality of polymorphisms Genetics 1973 74 175 195 4711903
Akey JM Zhang G Zhang K Jin L Shriver MD Interrogating a high-density SNP map for signatures of natural selection Genome Res 2002 12 1805 1812 12466284
Nordborg M Linkage disequilibrium, gene trees, and selfing: An ancestral recombination graph with partial self-fertilization Genetics 2000 154 923 929 10655241
Jørgensen S Mauricio R Neutral genetic variation among wild North American populations of the weedy plant Arabidopsis thaliana is not geographically structured Mol Ecol 2004 13 3403 3413 15487999
Tajima F Statistical method for testing the neutral mutation hypothesis by DNA polymorphism Genetics 1989 123 585 595 2513255
Innan H Terauchi R Miyashita NT Microsatellite polymorphism in natural populations of the wild plant Arabidopsis thaliana
Genetics 1997 146 1441 1452 9258686
Schmid KJ Ramos-Onsins S Ringys-Beckstein H Weisshaar B Mitchell-Olds T A multilocus sequence survey in Arabidopsis thaliana reveals a genome-wide departure from the standard neutral model of DNA sequence polymorphism Genetics 2005 169 1601 1615 15654111
Wright SI Gaut BS Molecular population genetics and the search for adaptive evolution in plants Mol Biol Evol 2005 22 506 519 15525701
Begun DJ Aquadro CF Levels of naturally occurring DNA polymorphism correlate with recombination rates in D. melanogaster
Nature 1992 356 519 520 1560824
Cutter AD Payseur BA Selection at linked sites in the partial selfer Caenorhabditis elegans
Mol Biol Evol 2003 20 665 673 12679551
Hellman I Ebersberger I Ptak SE Pääbo S Przeworski M A neutral explanation for the correlation of diversity with recombination rates in humans Am J Hum Genet 2003 72 1527 1535 12740762
Hill WC Robertson A Linkage disequilibrium in finite populations Theor Appl Genet 1968 38 226 231 24442307
Maynard Smith J Haigh J The hitchhiking effect of a favourable gene Genet Res 1974 23 23 35 4407212
Kaplan NL Hudson RR Langley CH The “hitch-hiking” effect revisited Genetics 1989 123 887 899 2612899
Charlesworth B Morgan MT Charlesworth D The effect of deleterious mutations on neutral molecular variation Genetics 1993 134 1289 1303 8375663
Bailey JA Gu Z Clark RA Reinert K Samonte RV Recent segmental duplications in the human genome Science 2002 297 1003 1007 12169732
Estivill X Cheung J Pujana MA Nakabayashi K Scherer SW Chromosomal regions containing high-density and ambiguously mapped putative single nucleotide polymorphisms (SNPs) correlate with segmental duplications in the human genome Hum Mol Genet 2002 11 1987 1995 12165560
Fredman D White SJ Potter S Eichler EE Den Dunnen JT Complex SNP-related sequence variation in segmental genome duplications Nature Genet 2004 36 861 866 15247918
Arnheim N Nei M Koehn RK Concerted evolution of multigene families Evolution of genes and proteins 1983 Sunderland (Massachusetts) Sinauer 38 61
Hudson RR Kaplan NL Statistical properties of the number of recombination events in the history of a sample of DNA sequences Genetics 1985 111 147 164 4029609
Andolfatto P Przeworski M A genome-wide departure from the standard neutral model in natural populations of Drosophila
Genetics 2000 156 257 268 10978290
Ardlie KG Liu-Cordero SN Eberle MA Daly M Barrett J Lower-than-expected linkage disequilibrium between tightly linked markers in humans suggests a role for gene conversion Am J Hum Genet 2001 69 582 589 11473344
Frisse L Hudson RR Bartoszewicz A Wall JD Donfack J Gene conversion and different population histories may explain the contrast between polymorphism and linkage disequilibrium levels Am J Hum Genet 2001 69 831 843 11533915
Przeworski M Wall JD Why is there so little intragenic linkage disequilibrium in humans? Genet Res 2001 77 143 151 11355570
Hagenblad J Nordborg M Sequence variation and haplotype structure surrounding the flowering time locus FRI in Arabidopsis thaliana
Genetics 2002 161 289 298 12019242
Haubold B Kroymann J Ratzka A Mitchell-Olds T Wiehe T Recombination and gene conversion in a 170-kb genomic region of Arabidopsis thaliana
Genetics 2002 161 1269 1278 12136029
Copenhaver GP Housworth EA Stahl FW Crossover interference in Arabidopsis
Genetics 2002 160 1631 1639 11973316
Stephens JC Schneider JA Tanguay DA Choi J Acharya T Haplotype variation and linkage disequilibrium in 313 human genes Science 2001 293 489 493 11452081
Akey JM Eberle MA Rieder MJ Carlson CS Shriver MD Population history and natural selection shape patterns of genetic variation in 132 genes PLoS Biol 2004 2 e286 10.1371/journal.pbio.0020286 15361935
Golub GH Van Loan CF Matrix computations, 3rd ed 1996 Baltimore Johns Hopkins University Press 694
| 15907155 | PMC1135296 | CC BY | 2021-01-05 08:21:23 | no | PLoS Biol. 2005 Jul 24; 3(7):e196 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030196 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030215SynopsisCell BiologyInfectious DiseasesMicrobiologyGastroenterology/HepatologyMus (Mouse)PlasmodiumRattus (Rat)Tracking a Killer: In Vivo Microscopy Reveals Details on the Life Cycle of Malarial Parasites Synopsis6 2005 24 5 2005 24 5 2005 3 6 e215Copyright: © 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.
Intravital Observation of Plasmodium berghei Sporozoite Infection of the Liver
==== Body
In a perfect world, anyone frequenting the habitat of malaria-carrying mosquitoes would sleep beneath a mosquito net, take prophylactic drugs, and maybe even quaff quinine-laced tonic water. But life is rarely perfect and even though governments and the World Health Organization have combined such prevention measures with widespread efforts to eradicate malaria-carrying mosquitoes, the disease kills more than a million people each year.
The microbes that cause the disease, tiny parasites of the genus Plasmodium, have a complex life cycle that involves several distinct phases and habitats. When a person is bitten by an infected mosquito, highly mobile Plasmodium cells (called sporozoites) migrate from the skin into the bloodstream, which carries them to the liver, where they set up shop in liver cells (called hepatocytes) and multiply asexually as merozoites. Eventually the parasites leave the liver and reenter the bloodstream, where they invade red blood cells, multiply again, and differentiate into new merozoites. Ultimately, a red blood cell will become so chock-full of merozoites it bursts, releasing more merozoites to infect other red blood cells. Some of these merozoites will differentiate into male and female sex cells (called gametocytes) that hitchhike along with red blood cells when a new mosquito takes a blood meal from an infected person. The gametocytes then breed within the mosquito and produce sporozoites, which reside in the insect's gut—and then the cycle begins all over again.
Although much is already known about the Plasmodium life cycle, many details—including the discrete steps that facilitate sporozoites' invasion of the liver—have escaped direct observation until now. In a new study, Ute Frevert et al. literally take a closer look at this process with intravital microscopy—which allows direct observation of cell movement in a living animal—to see how the parasites gain entry into the liver.
Tagged with red fluorescent protein, the malarial parasites in the sporozoite stage can be seen migrating along the sinusoids in a mouse liver
To visualize this process, the authors used genetically engineered Plasmodium parasites to express fluorescent tags. They introduced these fluorescent parasites to mice and rats the natural way—through mosquito bites—and then watched for the arrival of the parasites in the livers of the test animals. They observed that parasites were carried into the liver by the bloodstream, and then attached to the sinusoidal walls (sinusoids take the place of capillaries in the liver). They watched as the parasites crawled along the interior of the sinusoids—sometimes against the direction of blood flow—until they reached a specialized cell called a Kupffer cell. These star-shaped cells line the liver sinusoids and clean the blood of particulate debris and dead blood cells. But they also, apparently, serve as the parasites' portal of entry into the liver.
Earlier work had suggested that sporozoites might use Kupffer cells to access the liver, but Frevert et al. watched sporozoites traverse Kupffer cells to reach the liver interior. They observed some interesting details in this process: as sporozoites enter a Kupffer cell, they first pause, then undergo a slow constriction as they insinuate their way through the cell (rather like a napkin being drawn through a napkin ring). Sporozoites traverse the Kupffer cell at a speed much slower than they could crawl, indicating that this traversal involves more than mere parasite locomotion. Upon exiting the Kupffer cell on the other side, the sporozoites wreaked havoc in the liver, leaving a path of destruction and dead cells behind them as they moved through several consecutive hepatocytes before finally settling down in one to begin reproducing.
Frevert et al. had to make multiple attempts and track several different sporozoites at each stage in order to gain a comprehensive picture of this part of the parasite's life cycle. But thanks to the visualization advantages provided by using the fluorescent parasites and intravital microscopy, the authors show that it is now possible to directly observe events in the Plasmodium life cycle that had only been inferred before.
| 0 | PMC1135297 | CC BY | 2021-01-05 08:21:24 | no | PLoS Biol. 2005 Jun 24; 3(6):e215 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030215 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1590715610.1371/journal.pbio.0030201Research ArticleCell BiologyDevelopmentGenetics/Genomics/Gene TherapyDrosophilaA Role for Adenosine Deaminase in Drosophila Larval Development Adenosine Deaminase in DrosophilaDolezal Tomas
1
Dolezelova Eva
2
Zurovec Michal
2
Bryant Peter J [email protected]
1
1Developmental Biology Center, University of California, Irvine, California, United States of America2Institute of Entomology and University of South Bohemia, Ceske Budejovice, Czech RepublicKrasnow Mark Academic EditorStanford University School of MedicineUnited States of America7 2005 24 5 2005 24 5 2005 3 7 e20122 11 2004 29 3 2005 Copyright: © 2005 Dolezal 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.
Drosophila Larval Development and Human Immunodeficiency: The Adenosine Deaminase Connection
Adenosine deaminase (ADA) is an enzyme present in all organisms that catalyzes the irreversible deamination of adenosine and deoxyadenosine to inosine and deoxyinosine. Both adenosine and deoxyadenosine are biologically active purines that can have a deep impact on cellular physiology; notably, ADA deficiency in humans causes severe combined immunodeficiency. We have established a Drosophila model to study the effects of altered adenosine levels in vivo by genetic elimination of adenosine deaminase-related growth factor-A (ADGF-A), which has ADA activity and is expressed in the gut and hematopoietic organ. Here we show that the hemocytes (blood cells) are the main regulator of adenosine in the Drosophila larva, as was speculated previously for mammals. The elevated level of adenosine in the hemolymph due to lack of ADGF-A leads to apparently inconsistent phenotypic effects: precocious metamorphic changes including differentiation of macrophage-like cells and fat body disintegration on one hand, and delay of development with block of pupariation on the other. The block of pupariation appears to involve signaling through the adenosine receptor (AdoR), but fat body disintegration, which is promoted by action of the hemocytes, seems to be independent of the AdoR. The existence of such an independent mechanism has also been suggested in mammals.
Adenosine deaminase is critically important to survival; congenital deficiency in humans leads to severe immunodeficiency. Here, the authors demonstrate that adenosine deaminase deficiency in flies results in severe developmental defects.
==== Body
Introduction
Adenosine deaminase (ADA) is an enzyme present in all organisms that catalyzes the irreversible deamination of adenosine and deoxyadenosine to inosine and deoxyinosine. It is a critically important enzyme for human survival because its congenital absence causes severe combined immunodeficiency disease (SCID). ADA deficiency accounts for about 20% of all types of SCID [1]. It is one of the most severe human immunodeficiencies and is associated with depletion of all three major categories of lymphocytes: T cells, B cells, and natural killer cells, resulting in impaired cellular immunity and decreased production of immunoglobulins [2]. Without intervention, the affected individuals die from opportunistic infections within the first few months of life.
ADA occurs as a soluble monomer in all human cells, but also exists as “ecto-ADA,” bound to the membrane glycoprotein CD26/dipeptidyl peptidase IV, and it has been suggested that this form of ADA regulates extracellular adenosine levels [3]. ADA deficiency is accompanied by greatly elevated levels of the ADA substrates adenosine and deoxyadenosine, both of which are biologically active purines that can have a deep impact on cellular physiology. Adenosine is not just a metabolite; it is also a signaling molecule that regulates numerous cellular functions by binding to G protein-coupled adenosine receptors (A1, A2a, A2b, and A3 in mammals) that can regulate intracellular cyclic adenosine monophosphate [4]. Deoxyadenosine is a cytotoxic metabolite released by various cell populations that undergo programmed cell death; it can kill cells through a mechanism that includes disturbances in deoxynucleotide metabolism [5].
Extracellular adenosine is now considered an important stress hormone that is released in excessive amounts in the vicinity of immune cells during both systemic and cellular stress [6]. The predominant source of extracellular adenosine during systemic activation of the stress system is the sympathetic nervous system [7]. Specific inflammatory stimuli such as bacterial products are also capable of triggering adenosine release from immune cells [8]. These data are in line with evidence demonstrating a dramatic increase in extracellular adenosine levels under conditions associated with multiple organ failure, which is the cause of 50%–80% of all deaths in surgical intensive care units [6].
ADA is not the only adenosine deaminase in mammalian cells. Recently, the cat eye syndrome critical region protein 1 (CECR1) gene was identified and shown to encode a protein representing a subfamily of proteins related to but distinct from classical ADAs [9]. The duplication of a small region of chromosome 22 containing this gene is associated with “cat eye syndrome,” a disorder characterized by hypoplastic kidneys, congenital heart malformation, and anomalous pulmonary venous connections. The founding member of this subfamily is encoded by insect-derived growth factor
(IDGF) [10], and homologs have been described in various organisms [11–14].
We have previously found six Drosophila genes with sequence similarity to the CECR1 subfamily [15]. Their products are mitogenic on Drosophila cells, and at least two of them (ADGF-A and ADGF-D) exhibit strong ADA activity, which is necessary for their mitogenic function. We therefore named them adenosine deaminase-related growth factors (ADGFs). We also demonstrated that adenosine functions as a negative signal for cell proliferation and concluded that ADGFs stimulate cell growth in vitro by depletion of extracellular adenosine [16]. Drosophila also contains a gene, termed Ada, with sequence similarity to human ADA, but as we have previously shown the product of this gene is most likely not an active ADA [16].
In this report we show that a null mutation in Drosophila ADGF-A gene leads to dramatically increased levels of adenosine and deoxyadenosine in the larval hemolymph. This increase leads to larval death associated with the disintegration of fat body and the development of melanotic tumors. We present a detailed analysis of the hematopoietic defects associated with the adgf-a mutation, show a genetic interaction of this mutation with signaling through the Drosophila adenosine receptor (AdoR, encoded by the gene CG9753) and with regulation of premetamorphic changes by ecdysone, as well as a genetic interaction of ADGF-A with a major innate immunity regulator—the Toll signaling pathway.
Results
Mutation in the ADGF-A Gene Causes Larval Death and Melanotic Tumors
We produced mutations in five of the six ADGF genes by homologous recombination mutagenesis [17] and showed that loss of the most abundantly expressed gene, ADGF-A, leads to death in the larval or pupal stage. Under optimal conditions (20–30 isolated homozygous larvae per vial), about 60% of larvae homozygous for the adgf-a mutation reach the third instar. Development during the third larval instar is significantly delayed, and wandering homozygous larvae usually appear 2 d after their heterozygous siblings, which start wandering at about 5 d of development. Some homozygous third-instar larvae can be found alive in the vial even after 10 d of development. Mutant third-instar larvae show fat body disintegration (Figure 1A and 1B) and multiple melanotic tumors (Figure 1C), predominantly in the caudal part of the body and accompanied by disintegration of the fat body. Melanization of the lymph glands was never observed in these larvae, and the imaginal discs and brain appear normal. Less than 30% of homozygotes eventually pupate. Homozygous pupae usually die soon after pupariation; in some cases they develop normal head and thorax imaginal structures; however, abdominal parts usually do not develop. There is also an abnormal curvature (to the right) of the pupal abdomen (Figure 1D). Less than 2% of mutant pupae develop normally and eventually emerge as adults without any obvious abnormalities besides the abdominal curvature; some of them are sterile.
Figure 1 adgf-a Mutant Phenotype
(A and B) Fat body disintegration visualized by GFP expression driven by Cg-Gal4 driver in the fat body. While adgf-a/+ heterozygous third instar larvae have normal flat layers of fat body (A),adgf-a mutant showed extensive fat body disintegration into small pieces of tissue (B).
(C) Multiple melanotic tumors present in adgf-a mutant third-instar larva.
(D) An adgf-a mutant pupa with typical abdominal curvature.
To confirm that the mutant phenotype is caused solely by a mutation in the ADGF-A gene, we created transgenic flies carrying the ADGF-A gene under a heat-shock promoter (HS-ADGF-A). The adgf-a homozygous flies carrying the HS-ADGF-A construct showed survival rates significantly higher than adgf-a even without heat shock, probably due to leaky expression of the HS-ADGF-A construct (Figure 2A). However, while non-heat shocked animals still produced many melanotic tumors, only 22% of animals that were heat shocked as late embryos/early first instar developed these tumors (Figure 2B). This result confirms that the mutant phenotype is caused by the mutation in the ADGF-A gene. This conclusion is further supported by the even more efficient rescue achieved by expression of transgenically provided ADGF-A in the lymph glands using the Gal4/UAS system (see below).
Figure 2 Rescue of the adgf-a Mutant Phenotype by Expression of ADGF-A in Different Tissues
(A) Percentage of pupae (blue bars) and adult flies (purple bars) demonstrating the larval and pupal survival, respectively, of the adgf-a mutant flies rescued by expression of transgenic ADGF-A in different tissues. Along the x-axis (which is shared with [B]), the rescue experiments are shown (marked by the Gal4 driver used for expression of ADGF-A except for first three sets of bars—the first set presents only an adgf-a mutant, the second an adgf-a mutant carrying HS-ADGF-A construct without heat shock, and the third with heat shock) and the y-axis represents percentage of pupae and adult flies out of the total number of transferred first-instar larvae of particular genotype. Each experiment was repeated at least four times (with 20–30 animals in each vial) and the standard error is shown.
(B) Percentage of late third-instar larvae with melanotic tumors. The x-axis is shared with (A) (described above). The y-axis shows the percentage of larvae with tumors out of all larvae of each genotype examined for (A).
The adgf-a Mutant Phenotype Is Associated with Elevated Levels of Adenosine and/or Deoxyadenosine
Using liquid chromatography and mass spectrometry of deproteinated hemolymph samples, we measured adenosine concentrations in hemolymph of mutant and wild-type third-instar larvae. The adenosine concentration in the adgf-a mutant was 1.14 ± 0.26 μM compared to less than 0.08 μM in the wild type, and the deoxyadenosine concentration in mutants was 1.66 ± 0.99 μM compared to an undetectable level in the wild type.
The Catalytic Activity of ADGF-A Is Required for Its Function
To test whether the function of ADGF-A in vivo is also dependent on its catalytic activity, we produced two versions of the UAS-ADGF-A construct [18]: one carrying wild-type cDNA of ADGF-A and one carrying an ADGF-A cDNA with a mutation causing a substitution of two amino acids (H386G and A387E) in the catalytic domain [16]. Two different lines carrying the wild-type UAS-ADGF-A expression construct together with an Actin-Gal4 driver (providing ubiquitous expression) both completely rescued the mutant phenotype, whereas larvae with UAS-ADGF-A but without the driver showed the typical mutant phenotype. However, neither of the two lines carrying the mutated version of the UAS-ADGF-A (producing full-length protein detected by anti-myc antibody; see Materials and Methods) showed any rescue of the mutant phenotype. This result therefore demonstrates that the catalytic activity of ADGF-A is required for its function in vivo.
Hemocyte Development Is Affected in the adgf-a Mutant
We investigated the number and morphology of hemocytes (blood cells) in the hemolymph of the adgf-a late third-instar larvae (Figures 3 and 4). These larvae contain an average of seven-fold more hemocytes in circulation than wild-type larvae (Figure 3). In contrast to normal larval plasmatocytes, which remain rounded after settling down on the substrate (Figure 4A), most of the cells in the adgf-a mutant (more than 75%) are strongly adhesive and, after they are deposited in a drop of hemolymph on a microscope slide, develop filamentous and membranous extensions (Figure 4B–4D). An average of 7% of hemocytes in the adgf-a mutant are lamellocytes (Figures 3 and 4E), large flat cells that are not present in circulation of wild-type larvae under normal conditions [19]. Crystal cells were also detected in excess, with mutant larvae carrying several hundred while there are fewer than a hundred of these cells in the wild type (Figure 5). The lymph glands normally do not release hemocytes into the hemolymph before metamorphosis [20]; instead, they are released during metamorphosis when the lymph glands disperse [19]. However, the lymph glands of adgf-a mutant larvae are already dispersed in the late third instar. This process is similar to normal metamorphic changes, in which the hemocytes are first released from the front lobes, and the posterior lobes disperse later.
Figure 3 Number of Circulating Hemocytes in Late Third-Instar Larvae
Genotypes are shown along the x-axis, and the number of hemocytes/larva along the y-axis. Each bar shows the number of all circulating hemocytes, and the gray part of the bars represent the lamellocyte population. Each count was repeated five to ten times and the standard error is shown.
Figure 4 Hemocyte Abnormalities in adgf-a Mutant Larvae
(A–E) Differential interference contrast microscopy of living circulating hemocytes (magnification 200×; scale bar, 10 μm). Round, nonadhesive plasmatocytes from wild-type larva (A). Hemocytes from the adgf-a mutant developing filamentous extensions (B and C) or membranous extension surrounding the cell (D). Large flat lamellocyte from the adgf-a mutant (E).
(F and G) Differential interference contrast and fluorescent microscopy (merged image) of living circulating hemocytes stained by the Hml-GFP marker (magnification 100×; scale bar, 10 μm). While most of the cells from wild-type larvae are GFP-positive (F), just few of the cells from late third instar adgf-a larvae are stained by GFP at this stage (G).
(H–J) Fluorescence microscopy of living larvae with Hml-GFP stained hemocytes (magnification 40×; scale bar, 100 μm). Posterior part of late third-instar wild-type larva (H). Middle sections of early third-instar larvae of wild type (I) and adgf-a mutant (J).
Figure 5 Crystal Cells in Late Third Instar Larvae
Crystal cells were visualized by heating larvae of different genotypes at 60 °C for 10 min [46]. (A) Wild-type larva, (B) adgf-a single mutant, (C) adoR adgf-a double mutant (scale bar, 0.5 mm).
To analyze hemocytes in living larvae, we used the Hemolectin marker (Hml) [21]. We compared the number and distribution of hemocytes stained by GFP in flies carrying hml-Gal4 UAS-GFP in wild-type and mutant backgrounds. While there are relatively few hemocytes, mostly free-floating in the hemolymph, in early third-instar wild-type larvae (see Figure 4I), a much higher number of hemocytes, which are mostly attached to the tissues under the integument (described as sessile hemocytes in [19]), was observed in mutant larvae (see Figure 4J). A similar behavior was detected later in wild-type larvae, toward the end of the third instar (see Figure 4H). At this stage, the Hml marker disappeared from the most of the hemocytes in mutants (see Figure 4F and 4G).
The adgf-a Mutant Phenotype Is Rescued by Expression of ADGF-A in the Lymph Glands
To distinguish which tissues require ADGF-A expression for proper development, we tested for rescue of adgf-a lethality by expressing ADGF-A in specific subsets of larval tissues. A transgenic line carrying the UAS-ADGF-A construct on Chromosome II was crossed to lines expressing the Gal4 driver [18] in different tissues (Table 1). Since ADGF-A is normally expressed in the larval lymph glands [16], and the mutant phenotype is characterized by abnormal hemocyte development, special consideration was given to lines expressing the Gal4 driver in the lymph glands and/or circulating hemocytes. No line expressing the Gal4 driver exclusively in the lymph glands has been reported, so we used a combination of lines sharing in common the feature of Gal4 driver expression in the lymph glands. The results (see Figure 2 and Table 1) clearly demonstrate that expression of ADGF-A in the lymph glands (driven by Cg-Gal4,
e33C-Gal4, or c564-Gal4), but not in any other tissue examined, is necessary and sufficient to fully rescue the adgf-a lethality. In e33C-Gal4/UAS-ADGF-A, strong expression of ADGF-A in all lobes of developing lymph glands (but not in circulating hemocytes) reduces the number of hemocytes in the hemolymph to almost normal levels (see Figure 3). The number of hemocytes is also reduced, but to a lesser extent in larvae rescued by Cg-Gal4/UAS-ADGF-A. However, when assayed by survival rate and melanotic tumor formation, the rescue by Cg-Gal4 is full and similar to that of e33C-Gal4 (see Figure 2). The difference in effectiveness may be explained by the different expression patterns of the drivers. Cg-Gal4 is expressed only in certain compartments of lymph gland lobes containing relatively mature hemocytes, and strongly in most circulating hemocytes [22, 23]. The C564-Gal4 driver is not expressed as strongly as e33C-Gal4, but is still uniformly expressed in the lymph glands; it also fully rescued the mutant phenotype. We have tried two different insertions of the Dot-Gal4 construct. The Dot-Gal411C on Chromosome II, which shows weak expression [24], did not rescue the phenotype, but a Dot-Gal443A insertion on Chromosome X, which shows stronger expression, rescued approximately half of the mutant animals (Figure 2). Nearly all rescued individuals were males, suggesting that expression of the Gal4 driver was influenced by X-chromosome dosage compensation, and expression in females heterozygous for Dot-Gal4 was not strong enough for rescue.
Table 1 Gal4 Drivers—Expression Pattern and Rescue of the adgf-a Phenotype
Expression of ADGF-A in salivary glands and fat body (as well as in other tissues) is not required for full rescue, as demonstrated by use of the Cg-Gal4,
Dot-Gal4, but especially by e33C-Gal4 driver, and is also not sufficient to rescue the phenotype at all, as demonstrated by T110-Gal4 and Lsp2-Gal4 (Table 1).
Since ADGF-A is strongly expressed in embryonic mesoderm [16], we have tried to rescue the phenotype by the expression of ADGF-A in embryonic and larval muscle cells using the Dmef2-Gal4 driver [25]. No rescue of the phenotype, including body shape of escaping pupae, was observed.
The only line showing significant (but not complete) rescue of adgf-a survival without expression in the lymph glands was GawB5015 (see Figure 2), which expresses the Gal4 driver very strongly and specifically in the ring gland and salivary glands (as well as very weak and spotty expression in imaginal discs [unpublished data]). However, expression of ADGF-A driven by GawB5015 does not prevent the formation of melanotic tumors (see Figure 2B).
Ablation of Hemocytes in Mutant Larvae Reduces Fat Body Disintegration and Melanotic Tumor Formation
The l(3)hematopoiesis missing
(l[3]hem) mutation reduces cell division in larval proliferating tissues and thus dramatically reduces the number of hemocytes in larvae. It also suppresses the hemocyte overproliferation and associated defects observed in the hopscotchTumorous-lethal mutant [26]. We therefore used the l(3)hem1 mutation to test whether the reduction of hemocyte number in the adgf-a mutant affects the phenotype. We recombined this mutation onto the chromosome containing the adgf-a mutation and found that in homozygous l(3)hem1,
adgf-a double mutants the number of hemocytes is significantly reduced compared to the adgf-a single mutants (see Figure 3). Furthermore, while 90% of adgf-a mutant larvae showed disintegration of fat body, only 40% of l(3)hem1,
adgf-a double mutants (total number of counted animals was 82) show the disintegration (Figure 6A). Similarly, melanotic tumor formation is significantly suppressed by l(3)hem1, with only 55% of double mutants showing melanotic tumors compared to more than 83% in adgf-a (Figure 6A). However, the delay in development and block of pupariation (Figure 6B), as well as the pupal body shape, were not influenced by this mutation. This shows that the effect on hemocyte development is related to only one other aspect of the adgf-a phenotype—namely, fat body disintegration—and the developmental arrest of adgf-a mutants is probably independent of this process.
Figure 6 Suppression of the adgf-a Mutant Phenotype by Mutations in Other Genes
(A) Percentage of late third-instar larvae with melanotic tumors (black bars) and fat body disintegration (green bars). The x-axis (which is shared with [B]), shows the genotype. The y-axis shows the percentage of larvae with tumors and fat body disintegration.
(B) Survival rate of double mutants compared to single adgf-a mutant. The y-axis shows the percentage of the pupae (blue bars) and adult flies (purple bars) demonstrating the larval and pupal survival, respectively. Each experiment was repeated at least four times (with 20–30 animals in each vial) and the standard error is shown.
Block in Activation of Macrophages Suppresses Disintegration of Fat Body
Previous results suggest that fat body disintegration might be caused by the action of hemocytes. Embryonic macrophages express the scavenger receptor encoded by croquemort
(crq), which allows them to bind and remove apoptotic corpses [27]. We therefore tested whether a mutation in the crq gene would block the suggested interaction between hemocytes and fat body in adgf-a mutant larvae. We used the mutation crqKG01679, caused by a P-element insertion in the first untranslated exon of crq, which leads to pupal lethality. The number of crystal cells was not increased and lamellocytes were not detected in crq, adgf-a double mutants (see Figure 3). The double mutants showed a lower number of circulating hemocytes than the single mutant, but there was still a significant increase in this number compared to wild type (see Figure 3), and the cells showed increased clumping. None of the double-mutant larvae showed either disintegration of fat body or melanotic tumor formation (Figure 6A). Even the adgf-a mutant larvae heterozygous for the crq mutation (crq/CyO GFP; adgf-a/adgf-a) showed significant suppression of the fat body disintegration, with most of the tissue staying compact in bigger pieces and never disintegrating to single adipose cells; melanotic tumors were rarely observed. This shows that the block of the putative interaction between fat body and macrophage-like cells (which are still present in double mutants) suppresses the fat body disintegration, further strengthening the hypothesis that the disintegration is caused by hemocytes. In addition, the absence of lamellocytes and the normal number of crystal cells in the double mutant strongly suggest that the differentiation of these cells and thus melanotic tumor formation is a secondary reaction to fat body disintegration, rather than a primary effect of the adgf-a mutation.
Mutation in a Putative Adenosine Receptor Suppresses the Block of Pupariation in adgf-a
We have identified a putative homolog of the mammalian adenosine receptor family in the Drosophila genome, AdoR, and produced a null mutation in this gene using homologous recombination (adoR; ED, unpublished data). The adoR mutants are fully viable. We used this mutant to test the hypothesis that the increased level of adenosine in the adgf-a mutant contributes to the mutant phenotype by its effect on signaling through the adenosine receptor. The results show that introducing the adoR mutation into the adgf-a background significantly increases pupariation, as well as adult emerging rate, compared to the adgf-a single mutant (Figure 6B). When the earlier lethality was avoided by picking up larvae after molt to the third instar, the pupariation rate of adoR, adgf-a double mutant was comparable to wild type as well as to the single adgf-a mutant treated with ecdysone (Figure 7A). Development during the third instar is much less delayed in the double mutant, with most of the larvae pupariating within 1 d after their heterozygous siblings (Figure 7A).
Figure 7 Ecdysone Regulation of Development in adgf-a
(A) Larvae of different genotypes were collected after L2/L3 molt, and the number of puparia was counted at different time points (x-axis: hours after egg laying). The y-axis shows the percentage of puparia out of all collected third-instar larvae (three vials each with 30 animals; the standard error is shown).
(B and C) Ring gland morphology in arrested adgf-a larvae. Approximately 8-d old mutant larva (i.e., 3 d after normal pupariation) with very extensive fat body disintegration (note the transparency of larva in the middle part with small white pieces of fat body) (B). The ring gland dissected from this larva (C) shows morphology of the normal ring gland before the degenerative changes of prothoracic gland starts (compare to schematic diagram to the left of [C], from [28]).
(D–F) Expression of GFP-marked glue protein (SgsΔ3-GFP) in salivary gland of the adgf-a mutant larvae and pupae. All late third-instar larvae express the glue protein as shown on dissected salivary gland (D). Some mutants show typical expulsion from the glands with GFP totally external to the puparial case (E), while others do not expel glue proteins even after puparium formation (F).
The adoR mutation also significantly reduced melanotic tumor formation in the adgf-a mutant (see Figure 6A), but disintegration of the fat body appeared at the same rate as in the single mutant (see Figure 6A). While the number of macrophage-like cells in circulation is not significantly changed in the double mutant, the number of lamellocytes is decreased (see Figure 3), but the number of crystal cells is normal (see Figure 5A and 5C).
These results demonstrate that adenosine signaling through the adenosine receptor is involved in the developmental arrest of adgf-a mutant, but that it does not play a role in fat body disintegration and macrophage differentiation.
Hormonal Regulation in the adgf-a Mutant
The delayed development and low pupariation rate in the adgf-a mutant larvae (see Figures 2A and 7A) could be caused by an effect on hormonal regulation of development. The main source of developmental hormones in the Drosophila larva is the ring gland, composed of the prothoracic gland, corpus allatum, and corpus cardiacum [28]. The prothoracic gland releases the steroid molting hormone ecdysone, which is converted to an active form, 20-hydroxyecdysone (20E), by the fat body as well as some of the target organs [29]. The block of pupariation in the adgf-a mutant suggested that the level of ecdysone in these larvae might not be sufficient to initiate pupariation. To test this possibility, we tried to rescue the phenotype by feeding mutant larvae 20E, which can initiate pupariation in the ecd1 mutant, which has an extremely low level of ecdysone [30,31]. The results (Figure 7A) clearly demonstrate that the adgf-a mutant larvae are responsive to ecdysone and that this treatment restores the pupariation frequency to almost wild-type level. The delay in development is also significantly reduced (Figure 7A).
Since the adgf-a mutant shows certain precocious metamorphic changes (macrophage differentiation and fat body disintegration), we speculated that a reduced ecdysteroid level could be caused by precocious degeneration of the prothoracic part of the ring gland. However, the overall structure of the ring gland is not visibly affected even in the oldest larvae (10 d, i.e., 5 d after the heterozygous siblings pupariated) with a fully disintegrated fat body (Figure 7B and 7C).
We also used a transgenic line carrying the SgsΔ3-GFP construct, which was previously used to monitor the effects of ecdysteroid levels on glue protein expression in salivary glands [32]. All analyzed adgf-a mutant larvae carrying the SgsΔ3-GFP construct showed normal expression of Sgs-GFP in salivary glands (Figure 7D). Mutants that pupariated usually showed typical GFP expectoration, indicating the presence of a high premetamorphic peak of ecdysteroids (Figure 7E). In some cases, GFP was secreted into the lumen of salivary glands, but was not expectorated (Figure 7F), which is similar to the defect seen in animals expressing the dominant-negative form of ecdysone receptor driven by the Sgs3-Gal4 driver [33]. These results demonstrate that the target tissues of adgf-a mutants are normally responsive to ecdysteroids and that they are probably capable of releasing ecdysteroids, although the level of ecdysteroids might vary.?>
ADGF-A Genetically Interacts with Toll Signaling Pathway
The antimicrobial response of Drosophila includes at least two distinct signaling pathways [34]—the Toll signaling pathway, which leads to the activation of two nuclear factor kappa B (NF-κB) factors, Dorsal-related immunity factor (DIF) and dorsal (DL); and the immune deficiency protein pathway activating the third NF-κB factor, Relish (REL). A zygotic null mutation in cactus (cact; a Drosophila inhibitor of NF-κB) leads to hyperproliferation of hemocytes, melanotic tumor formation, disintegration of fat body, and slower larval development, with 60% larval lethality, as well as a thin body-shape phenotype [35]. All of these phenotypes are strikingly similar to the abnormalities seen in adgf-a mutants, which was our first clue as to a possible interaction of ADGF-A with the Toll signaling pathway. We hypothesized that the activity of ADGF-A is suppressed by Toll signaling, resulting in similar phenotypes of the adgf-a mutation and constitutive activation of Toll pathway.
To test this hypothesis, we crossed transgenic flies carrying ADGF-A gene under the control of a heat-shock promoter on Chromosome II (HS-ADGF-A) with cactE8 (a lethal allele of cact on Chromosome II, which, in combination with cactD13, results in a zygotic null combination, or, with cactIIIG, results in zygotic hypomorphic combination).
Overexpression of ADGF-A in animals with a hypomorphic cact combination (cactE8/cactIIIG) increased the adult survival rate almost 4-fold (Figure 8A). The rescue could be increased by multiple heat shocks before pupariation to 7-fold (unpublished data). The suppression of melanotic tumor formation is also significant (from more than 80% down to 26%, Figure 8B). The most severe cact null mutation (cactE8/cactD13), leading to developmental arrest in larvae (less than 8% pupate), is partially rescued in animals with overexpression of ADGF-A when the pupariation rate is increased 3-fold (Figure 8A).
Figure 8 Genetic Interactions of Toll Signaling and ADGF-A
Survival rate and melanotic tumor formation were compared in mutants in the Toll signaling pathway and in similar mutants with overexpression of ADGF-A using the HS-ADGF-A construct.
(A) The bar graph shows the percentage of the pupae (blue bars) and adult flies (purple bars) demonstrating the larval and pupal survival of each genotype. The x-axis shows the genotypes and is shared with (B). Flies heterozygous for the cact mutation were used as a control.
(B) Percentage of late third instar larvae presenting melanotic tumor formation.
These results demonstrate that ADGF-A overexpression can partially rescue the effects of constitutively active Toll signaling in larvae, mainly the developmental arrest, but also the melanotic tumor formation, in the case of hypomorphic cact mutants.
Discussion
ADA Deficiency in Drosophila Causes Abnormal Hemocyte Development, Melanotic Tumor Formation, Fat Body Degeneration, and Delayed Development
We have established an ADA deficiency model in Drosophila in order to study the effects of altered adenosine levels in vivo. We produced a loss-of-function mutation in the ADGF-A gene, which produces a product (ADGF-A) with ADA activity. When homozygous, the mutation causes abnormal hemocyte development, leading to melanotic tumor formation [36], as well as fat-body disintegration associated with death during the larval stage or delayed transition to the pupal stage of development. In agreement with our previous study using cells cultured in vitro [16], here we have shown that ADA enzymatic activity is essential for ADGF-A function in vivo, when this function is assayed by testing for rescue of the mutant phenotype. Just as increased levels of both ADA substrates, adenosine and deoxyadenosine, are found in blood of SCID patients [5], adgf-a mutant larvae also have elevated levels of adenosine and deoxyadenosine, indicating that the mutant phenotype is caused by disturbance in the turnover of these nucleosides.
Expression of ADGF-A only in the lymph glands is sufficient to fully rescue the mutant phenotype, indicating that the hemocytes within the lymph glands play a major role in regulation of adenosine levels in the hemolymph. A similar regulatory role has also been attributed to blood cells in humans [5]. This suggests a function for ADGF-A within the lymph gland. However, ADGF-A behaves as a soluble growth factor and could be released from the lymph gland to activate targets elsewhere in the larval body. Our results show that ADGF-A functions by limiting the level of extracellular adenosine, and in this way the protein could have a systemic function even if it were restricted to its tissue of origin. Although our tests did not exclude a role for ADGF-A in circulating hemocytes (which constitute a separate lineage from the lymph gland hemocytes [20]), we showed that expression of ADGF-A in circulating hemocytes is not required for rescue of the adgf-a mutant phenotype, since e33C-Gal4/UAS-ADGF-A—which expresses ADGF-A in the lymph gland but not in circulating hemocytes—fully rescued the phenotype.
ADGF-A Is Involved in Hemocyte Differentiation in the Lymph Glands
Late third-instar larvae homozygous for the adgf-a mutation contain, on average, seven times more hemocytes in circulation than wild-type larvae, and most of these cells show strong adhesive properties compared to normal larval plasmatocytes, which remain rounded after settling down on the substrate. Although these cells share other characteristics with plasmatocytes, they are normally not seen in circulation until they are released from the lymph glands at the onset of metamorphosis under the regulation of ecdysone to serve as phagocytes for histolysing tissues during metamorphosis—thus, they are referred to as pupal macrophages [19]. In agreement with the presence of these cells in circulation, at least the first lobes of the lymph glands are usually completely dispersed in late third-instar mutant larvae. This indication of precocious metamorphic changes [36] in the mutant is further supported by the finding that hemocytes aggregate in a segmental pattern in early rather than late third instar (see Figure 4H–4J), and that the hemocytes lose expression of Hemolectin in late third-instar larvae rather than at the onset of metamorphosis (see Figure 4G) [21].
Recent studies show that the Toll signaling pathway, which is already known to be involved in the control of innate immunity of both Drosophila and mammals [34], may also be involved in the control of hemocyte differentiation in the Drosophila larva. Constitutive activation of Toll signaling leads to developmental arrest and hematopoietic defects associated with melanotic tumor formation [35], similar to the phenotype of the adgf-a mutant. Our work also shows that forced expression of the ADGF-A gene can rescue the effects of overactive Toll signaling, suggesting that ADGF-A might function downstream of Toll signaling to control its effects. This conclusion is consistent with the existence of a putative binding site for Dorsal (one of two known effectors of Toll signaling) in the ADGF-A promoter (Figure 9). It will be important to explore this connection further, since recent studies suggest an interaction between adenosine signaling and the NF-κB signaling pathway, which is the mammalian counterpart of the Toll pathway [37].
Figure 9 Schematic Map of the ADGF-A Gene with Promoter Analysis
The ADGF-A gene contains four exons and two transcriptional starts [17,47]. We analyzed sequences preceding both transcriptional starts for the presence of known transcriptional factor binding sites using the software program Gene2Promoter (Genomatix Software GmbH). Selected sites are represented by color bars in approximate positions of promoter regions. The legend under the sequence show the names of transcription factors binding to matching colored binding sites.
Precocious Fat-Body Disintegration Caused by Mutant Hemocytes
One of the most remarkable features of the adgf-a mutant phenotype is the disintegration of the fat body in third-instar larvae, another indication of precocious metamorphic changes since the disintegration normally occurs much later, during pupal life. Furthermore, our study of this mutant provides strong evidence that the fat body disintegration is promoted by the action of hemocytes. Fat body disintegration was significantly suppressed when the hemocyte number was reduced using the l(3)hem1 mutation [26], and fully blocked by the croquemort (crq) mutation [27] which affects a CD36-related receptor (Croquemort) expressed on macrophages and required in phagocytosis of apoptotic cells. Human CD36 is a scavenger receptor which, in combination with the macrophage vitronectin receptor and thrombospondin, binds apoptotic cells. A similar role of Croquemort for removing histolysing tissues during Drosophila metamorphosis has not yet been tested, but seems likely since the crq mutant used in this study (crqKG01679) is lethal in pupae.
The idea that hemocytes are involved in fat body dissociation in Drosophila is further supported by work on the flesh fly Sarcophaga. Natori's group showed that proteinase cathepsin B was released from pupal hemocytes when they interacted with the fat body, and that this enzyme digested the basement membrane of the fat body, causing the tissue to dissociate [38,39]. They also showed that the interaction of hemocytes with the fat body is mediated by a 120-kDa membrane protein localized specifically on pupal hemocytes [40]. This protein was suggested to be a scavenger receptor, but it does not seem to be homologous to Drosophila Croquemort (unpublished data). Work by Franc et al. [27] is consistent with the idea that more than one scavenger receptor is involved in this process.
Possible Signaling Role for Adenosine
The precocious metamorphic changes that appear to occur in response to elevated adenosine in the adgf-a mutant larvae lead to the suggestion that adenosine may act as a regulatory signal for these processes during normal development. One possibility is that adenosine acts as a downstream effector of ecdysone-regulated prepupal changes, and that the increase in adenosine concentration is mediated by ecdysone-induced down-regulation of ADGF-A expression. This is supported by the presence of multiple sites for ecdysone-inducible transcription regulators in the ADGF-A promoter (Figure 9). Adenosine could serve as a signal for macrophage differentiation, and the lack of adenosine deaminase activity due to the adgf-a mutation could cause precocious differentiation of these cells in mutant larvae. We are now carrying out direct tests of the idea that the differentiation of hemocytes in mutant larvae is caused by elevated adenosine. If confirmed, this effect would have general significance, since in ADA-deficient mice, inflammatory changes in the lungs include an accumulation of activated alveolar macrophages [41], and this could also be mediated by elevated adenosine.
Elevated Adenosine Delays Development and Inhibits Pupariation
The elevated adenosine in the adgf-a mutant larvae leads to precocious changes (hemocyte differentiation and fat body disintegration) resembling those normally occurring at the time of metamorphosis, but it also is associated with an apparently opposite effect, in that it causes a significant delay in progress through the third larval instar and a decrease in the frequency of successful pupariation (formation of the puparium from the larval cuticle), which is one of the earliest steps in metamorphosis. We conclude that the mutation has additional effects on the hormonal regulation of development.
One possible explanation for the developmental delay and failure to pupariate is that the adgf-a mutation affects the production or release of ecdysteroid hormones from the major endocrine organ of the Drosophila larva—the ring gland. This is supported by the fact that pupariation rate and survival of the adgf-a mutant can be significantly improved by expression of transgenic ADGF-A in the ring gland and salivary glands. We suggest that this somehow interferes with the regulation of hormone release. Other mutants with hormonal dysregulation show delayed larval development and failure to pupariate [42,43]. Presumably the elevated adenosine in the adgf-a mutant blocks the production or release of ecdysone from the ring gland by an unknown mechanism. This idea is supported by our finding that both pupariation rate and survival of the adgf-a mutant can also be improved by feeding the mutant larvae with 20E in the diet (see Figure 7A). Thus it is clear that the adgf-a mutant is arrested in development due to an effect of the mutation on hormone production from the ring gland.
The arrest of development in the adgf-a mutants was significantly suppressed by loss of the adenosine receptor caused by the adoR mutation: larvae simply homozygous for adgf-a pupated after two or more days, whereas larvae also homozygous for adoR pupated within 1 d after their heterozygous siblings (see Figure 7A). Therefore, adenosine signaling through the AdoR must play a role in the developmental arrest of the adgf-a mutant, and this is most likely mediated by signaling to the ring gland, where AdoR is expressed (ED, unpublished data). The mutation in AdoR does not block macrophage differentiation and fat-body disintegration, so this effect must involve another, as yet uncharacterized mechanism independent of AdoR signaling. Work using adenosine-receptor deficient mammalian cells also suggested the existence of a novel, undefined adenosine signaling mechanism [44]. However, we cannot exclude the role of elevated deoxyadenosine in these effects. Drosophila, now with the advantage of the well-characterized adgf-a mutant, could serve as an ideal model system in which to investigate this mechanism.
Concluding Remarks
In our previous work using cells cultured in vitro, we showed that, as in mammals, adenosine can block proliferation and/or survival of some Drosophila cell types [16]. In the present work, we have established a Drosophila model to study altered levels of adenosine and deoxyadenosine in vivo, and we have shown that loss of ADGF-A function causes an increase of these nucleosides in larval hemolymph. Although the adgf-a mutation leads to larval or pupal death, we have shown that this is not due to the adenosine or deoxyadenosine simply blocking cellular proliferation or survival, as the experiments in vitro would suggest. Rather, this mutation leads to an increase in number of hemocytes at the end of larval development due to the differentiation and release of hemocytes from the lymph glands. Hemocytes also differentiate and are released from the lymph glands during systemic infection [19]. Together with our result suggesting an interaction between Toll signaling and ADGF-A, this leads to the hypothesis that adenosine controls hemocyte differentiation in response to infection, and that it signals through the adenosine receptor to postpone the next developmental step, metamorphosis. This would be consistent with the role of adenosine as a “stress hormone” in mammals [6]. A similar process of hemocyte differentiation and release from the lymph glands normally takes place at the onset of metamorphosis, when pupal macrophages remove histolyzing tissues. The ADGF-A promoter contains consensus binding sites for effectors of both Toll and ecdysone signaling. This raises the possibility that adenosine plays a role in the control of metamorphosis as well as in the response to stress.
Materials and Methods
Fly strains and genetics
For standard procedures, flies were raised at 25 °C on a standard cornmeal-agar-yeast-molasses diet supplemented with 0.3% Nipagin to retard mold growth. Oregon flies were used as the wild-type Drosophila strain, but in most cases the y w strain was used as a control since most mutations were carried in the y w background. A mutation in the ADGF-A gene on Chromosome III was obtained as described earlier [17]. In this study, the mutation described as adgf-akarel was used in all experiments and is referred to here as adgf-a. A mutation in the adenosine receptor gene on Chromosome III was produced by the ends-out targeting method (ED, unpublished data) and is referred to here as adoR. Transgenic flies carrying HS-ADGF-A,
UAS-ADGF-Amyc, and UAS-mutADGF-Amyc construct (see description below) were produced by a modified P-element transformation method [45]. HS-ADGF-A,
UAS-ADGF-Amyc[2A],
UAS-ADGF-Amyc[7A],
UAS-mutADGF-Amyc[1A], andUAS-mutADGF-Amyc[3B], all insertions on Chromosome II, were isolated and used in this work. The following markers and mutations were obtained from the Bloomington stock center, accessible at http://fly.bio.indiana.edu/ (stock numbers provided in parentheses): Hml-GFP marker (Hml-Gal4/UAS-GFP) expressing GFP in embryonic and larval hemocytes on Chromosome II (BL-6397), the l(3)hem1 mutation on Chromosome III (BL-6184), and the crqKG01679 mutation in the crq gene on Chromosome II (BL-14900). Mutants in Toll signaling pathway were obtained from Dr. S. Govind: cactE8,
cactIIIG, and cactD13 mutations in the cact gene on Chromosome II. The Gal4/UAS [18] system was used for protein misexpression. The following were obtained from the Bloomington stock center (stock numbers in parenthesis): Cg-Gal4 on Chromosome II (BL-7011), Pw+mW.hs=GawB5015 on II (BL-2721), Pw+mW.hs=GawBc564 on II (BL-6982), Pw+mW.hs=GawBT110 on II (BL-6998), Hml-Gal4 on II (BL-6396), Dot-Gal443A on X (BL-6903), Dot-Gal411C on II (BL-6902), and Lsp2-Gal4 (BL-6357) on III. The Pen2.4-GAL4e33C lethal insertion on Chromosome III was obtained from Dr. N. Perrimon's lab, and the Dmef2-Gal4 driver on II from Dr. A. Michelson. Expression information of these Gal4 drivers is provided in Table 1. A stock carrying the ubiquitous actin-Gal4 driver (P-actin-Gal4 UAS-GFP/CyO; lethal insertion on Chromosome II) was obtained from Dr. R. Sousa. To recognize homozygous larvae, balancer chromosomes with the GFP marker were used: CyO Pw+mW.hs=Ubi-GFP.S65TPAD1 (BL-4559) and TM3 Pw+mC=ActGFPJMR2 Ser (BL-4888). Transgenic flies SgsGFP-1 (insertion on Chromosome X) and SgsGFP-2 (insertion on Chromosome II) containing the chimeric gene construct SgsΔ3-GFP were obtained from Dr. A. J. Andres. For expression of ADGF-A using the HS-ADGF-A construct, flies were heat shocked as late embryos/early first instars at 37 °C for 30 min. In all rescue experiments, 30 freshly hatched homozygous first-instar larvae were selected using a GFP dissecting microscope and transferred into fresh vials (at least four vials for each variant). They were left to develop at 25 °C and examined as wandering third-instar larvae, pupae, and adults.
Ecdysone treatment
Mutant larvae were raised on plates with yeast paste at 25 °C and transferred to vials with glucose-yeast medium (control) or with glucose-yeast medium containing 20-hydroxyecdysone (H-5142; Sigma-Aldrich, St. Louis, Missouri, United States) at a concentration of 0.5 mg/ml shortly after the L2/L3 molt. Numbers of puparia were counted at 12-h intervals after the 120-h time point (when the first control larvae start to pupariate). The ecd1 flies (Bloomington stock BL-218) served as a control for the functional 20E diet [31]: flies were raised at 22 °C (permissive temperature for the temperature-sensitive ecd1 mutation) and transferred to vials with control or 20E-containing diet and raised at 29 °C (restrictive temperature).
Fat body observation
Living late third-instar larvae were washed and examined in PBS using a standard dissecting microscope with transmitted light. For finer analysis, the fat body was dissected from larvae in PBS and observed using a dissecting microscope. GFP-stained fat body was observed in living, etherized larvae in PBS solution on a standard microscopic slide with a coverslip under a fluorescence microscope.
Hemocyte counts and observations
Circulating hemocytes were obtained by opening two late third-instar larvae in 30 μl of PBS. This allowed us to collect all hemolymph from the larvae in a defined volume. The solution with circulating hemocytes was mixed by gently pipetting, and part was transferred into the chamber of an improved Neubauer hemocytometer. Cell number was recounted to one animal equivalent. Hemocyte morphology was observed by differential interference contrast microscopy of living cells in Shields and Sang Insect Medium (Sigma-Aldrich) obtained by the same procedure as for counting. To observe hemocyte morphology, samples were analyzed at least 10 min after the deposition of solution with hemocytes, in order to allow the cells to adhere to the surface of the slide. Crystal cells were visualized by heating larvae at 60 °C for 10 min in a water bath [46]. GFP-stained hemocytes were observed in living, etherized larvae in PBS solution on a standard microscopic slide with a coverslip under the fluorescence microscope or by deposition of hemocytes in PBS as for counting and observing under the fluorescence microscope.
Transgenic constructs
Wild-type cDNA for ADGF-A was amplified by PCR using proofreading DNA polymerase (ProofStart; Invitrogen, Carlsbad, California, United States) from a pOT2 vector containing the ADGF-A EST-clone (GH08276) using the following primers: 5′-
CGTCTAGAATGTCGCCAGTCATCCGCC-3′ (5′ end primer with XbaI tail) and 5′-
GCTGATCATCAATCGATCCGTTGACTGGGGGA-3′ (3′ end primer with BclI tail). The PCR product was cloned into the pGEM-T Easy vector (Promega, Madison, Wisconsin, United States), and the resulting plasmid (ADGF-A-pGEM) was cut by NotI/SpeI restriction enzymes. The ADGF-A fragment was then cloned into the pCaSpeR-hs) vector cut by NotI/XbaI to obtain the HS-ADGF-A construct. The myc tag was added to the C terminus of the ADGF-A protein for detection by anti-c-Myc antibody (Sigma-Aldrich). To produce a UAS-ADGF-Amyc construct, the ADGF-A fragment was amplified (by ProofStart from pOT2 vector) using the following primers: 5′-
AATCTCGAGCTCATCATGTCGCCAGTCATC-3′ (5′ end with XhoI tail) and 5′-
TATCTAGATCGATCCGTTGACTGGGGG-3′ (3′ end with XbaI tail). The fragment was cut by XhoI/XbaI and cloned into the pUAST vector modified by MZ. The sequence encoding the myc-tag 5′-
GAGCAAAAGCTCATTTCTGAAGAGGACTTG-3′ plus a stop codon was inserted into XbaI site of pUAST (using the XbaI site on the 5′ end and the NheI site on the 3′ end) cut by XhoI/XbaI. A mutated version—UAS-mutADGF-Amyc—was prepared in the same way as UAS-ADGF-Amyc, but pBLUESCRIPT containing mutated ADGF-A cDNA was used as a template. The mutated version of ADGF-A (carrying a mutation causing the substitution of two amino acids—His386 and Ala387 for Glu and Leu, respectively) in the catalytic domain, shown to abolish adenosine deaminase activity [16], was prepared by recombinant PCR using the following recombinant primers: 5′-
TCTACTTCGAGCTCGGAGAAACAAACTGGTTCGGT-3′ and 5′-
CTCCGAGCTCGAAGTAGAAATCAATGTCATCG-3′ and the same 5′ and 3′ end primers as above.
Adenosine and deoxyadenosine concentrations measurement
The detection method used liquid chromatography and mass spectrometry (LC/MS method) of deproteinated hemolymph samples. Larval hemolymph was collected from several larvae and centrifuged to pellet the hemocytes. 1 μl of hemolymph was diluted in 99 μl of buffer. The sample was introduced in CH3CN-0.05% TFA (50:50) either via a syringe pump at 3 μl/min or via an RP-C18 150 mm × 1 mm Symmetry C8 column at 50 μl/min employing an LCQ electrospray ion source operated at 4.2 kV. The peaks were then identified using the electrospray MSN mass spectra obtained by the collision-induced decomposition of the MH+ ion and its product ions in a series of MSN experiments that were performed with the ion trap mass spectrometer. The sugar moiety was cleaved off the adenosine molecule and produced ion with a molecular weight of 136 (adenine), which was then detected by MS.
Supporting Information
Accession Numbers
The FlyBase (http://flybase.bio.indiana.edu/) accession numbers for the genes and proteins discussed in this paper are: Ada (FBgn0037661), ADGF-A (FBgn0036752), ADGF-A cDNA (FBtp0018801), adgf-a mutation (FBal0157461), adgf-akarel (FBab0038650), ADGF-D (FBgn0038172), AdoR (the CG9753 gene; FBgn0039747), c564-Gal4 (FBti0002592), cactD13 (FBal0001509), cactE8 (FBal0030706), cactIIIG (FBal0001515), cactus (FBgn0000250), Cg-Gal4 (FBtp0012452) , croquemort (FBgn0015924), crqKG01679 (FBal0147219), Dmef2-Gal4 (FBtp0006434), Dorsal (FBgn0000462), Dorsal-related immunity factor (FBgn0011274), Dot-Gal411C (FBti0024024), Dot-Gal443A (FBti0024023), e33C-Gal4 (FBti0002599), ecd1 (FBal0003500), GawB5015 (FBti0001256 ), Hemolectin (FBgn0029167), HS-ADGF-A (FBtp0018800), l(3)hem1 (FBal0010873), Lsp2-Gal4 (FBti0018531), Relish (FBgn0014018), SgsΔ3-GFP (FBtp0013370), Sgs3-Gal4 (FBtp0016397), and T110-Gal4 (FBti0002605).
The GenBank Nucleotide database accession numbers for the genes and proteins discussed in this papers are: cathepsin B (496316), IDGF (1402633), and pupal hemocyte 120-kDa membrane protein (7023974).
HUGO Gene Nomenclature Committee (HGNC; http://www.gene.ucl.ac.uk/nomenclature/) accession numbers for the genes and proteins discussed in this paper are: CECR1 (1839) and human CD36 (1663).
Online Mendelian Inheritance in Man (OMIM; http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM) accession numbers for the conditions discussed in this paper are: cat eye syndrome (115470) and SCID (102700).
We are especially grateful to Petr Simek and his lab for performing the measurement of adenosine concentration. We thank Rui Sousa, Shubha Govind, Andrew J. Andres, and the Bloomington Stock Center for fly stocks; and Michal Gazi and other colleagues for critical reading of the manuscript and discussions. We are also grateful to our research technician Ruzenka Kuklova for her help. This work was supported by grants from the United States National Science Foundation (grant number 440860-21565), the Grant Agency of the Czech Republic (grant number 204/04/1205), the Grant Agency of the Czech Academy of Sciences (grant number A5007107), and the Ministry of Education, Youth, and Sports of the Czech Republic (grant number MSM6007665801).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. TD, ED, and MZ conceived and designed the experiments. TD and ED performed the experiments. TD analyzed the data. MZ and PJB contributed reagents/materials/analysis tools. TD, ED, MZ, and PJB wrote the paper.
Citation: Dolezal T, Dolezelova E, Zurovec M, Bryant PJ (2005) A role for adenosine deaminase in Drosophila larval development. PLoS Biol 3(7): e201.
Abbreviations
20E20-hydroxyecdysone
ADAadenosine deaminase
ADGF-Aadenosine deaminase-related growth factor-A
AdoRDrosophila adenosine receptor
cactcactus
crqcroquemort
GFPgreen fluorescent protein
HMLHemolectin
l(3)heml(3)hematopoiesis missing gene
NF-κBnuclear factor kappa B
SCIDsevere combined immunodeficiency disease
==== Refs
References
Aldrich MB Blackburn MR Kellems RE The importance of adenosine deaminase for lymphocyte development and function Biochem Biophys Res Commun 2000 272 311 315 10833410
Buckley RH Schiff RI Schiff SE Markert ML Williams LW Human severe combined immunodeficiency: Genetic, phenotypic, and functional diversity in one hundred eight infants J Pediatr 1997 130 378 387 9063412
Richard E Alam SM Arredondo-Vega FX Patel DD Hershfield MS Clustered charged amino acids of human adenosine deaminase comprise a functional epitope for binding the adenosine deaminase complexing protein CD26/dipeptidyl peptidase IV J Biol Chem 2002 277 19720 19726 11901152
Latini S Pedata F Adenosine in the central nervous system: Release mechanisms and extracellular concentrations J Neurochem 2001 79 463 484 11701750
Hershfield MS Mitchell BS Vogelstein B Immunodeficiency diseases caused by adenosine deaminase deficiency and purine nucleoside phosphorylase deficiency The metabolic and molecular bases of inherited disease 2001 New York McGraw-Hill 2585 2611
Hasko G Deitch EA Szabo C Nemeth ZH Vizi ES Adenosine: A potential mediator of immunosuppression in multiple organ failure Curr Opin Pharmacol 2002 2 440 444 12127878
Sperlagh B Doda M Baranyi M Hasko G Ischemic-like condition releases norepinephrine and purines from different sources in superfused rat spleen strips J Neuroimmunol 2000 111 45 54 11063820
Bodin P Burnstock G Increased release of ATP from endothelial cells during acute inflammation Inflamm Res 1998 47 351 354 9754870
Riazi MA Brinkmann-Mills P Nguyen T Pan H Phan S The human homolog of insect-derived growth factor, CECR1, is a candidate gene for features of cat eye syndrome Genomics 2000 64 277 285 10756095
Homma K Matsushita T Natori S Purification, characterization, and cDNA cloning of a novel growth factor from the conditioned medium of NIH-Sape-4, an embryonic cell line of Sarcophaga peregrina (flesh fly) J Biol Chem 1996 271 13770 13775 8662793
Charlab R Rowton ED Ribeiro JM The salivary adenosine deaminase from the sand fly Lutzomyia longipalpis
Exp Parasitol 2000 95 45 53 10864517
Li S Aksoy S A family of genes with growth factor and adenosine deaminase similarity are preferentially expressed in the salivary glands of Glossina m. morsitans
Gene 2000 252 83 93 10903440
Matsushita T Fujii-Taira I Tanaka Y Homma KJ Natori S Male-specific IDGF a novel gene encoding a membrane-bound extracellular signaling molecule expressed exclusively in testis of Drosophila melanogaster
J Biol Chem 2000 275 36934 36941 10967093
Akalal DB Nagle GT Mollusk-derived growth factor: Cloning and developmental expression in the central nervous system and reproductive tract of Aplysia
Brain Res Mol Brain Res 2001 91 163 168 11457505
Zurovec M Dolezal T Gazi M Bryant PJ ADGFs—Growth factors with enzymatic activity A Dros Res Conf 2001 42 69
Zurovec M Dolezal T Gazi M Pavlova E Bryant PJ Adenosine deaminase-related growth factors stimulate cell proliferation in Drosophila by depleting extracellular adenosine Proc Natl Acad Sci U S A 2002 99 4403 4408 11904370
Dolezal T Gazi M Zurovec M Bryant PJ Genetic analysis of the ADGF multigene family by homologous recombination and gene conversion in Drosophila
Genetics 2003 165 653 666 14573477
Brand AH Perrimon N Targeted gene expression as a means of altering cell fates and generating dominant phenotypes Development 1993 118 401 415 8223268
Lanot R Zachary D Holder F Meister M Postembryonic hematopoiesis in Drosophila
Dev Biol 2001 230 243 257 11161576
Holz A Bossinger B Strasser T Janning W Klapper R The two origins of hemocytes in Drosophila
Development 2003 130 4955 4962 12930778
Goto A Kadowaki T Kitagawa Y
Drosophila hemolectin gene is expressed in embryonic and larval hemocytes and its knock down causes bleeding defects Dev Biol 2003 264 582 591 14651939
Asha H Nagy I Kovacs G Stetson D Ando I Analysis of ras-induced overproliferation in Drosophila hemocytes Genetics 2003 163 203 215 12586708
Evans CJ Banerjee U Molecular genetic analysis of the Drosophila lymph gland A Dros Res Conf 2004 45 634
Kimbrell DA Hice C Bolduc C Kleinhesselink K Beckingham K The Dorothy enhancer has Tinman binding sites and drives hopscotch-induced tumor formation Genesis 2002 34 23 28 12324942
Ranganayakulu G Schulz RA Olson EN Wingless signaling induces nautilus expression in the ventral mesoderm of the Drosophila embryo Dev Biol 1996 176 143 148 8654890
Leung SS Govind S The l(3)hem1 mutation blocks proliferation and development of hematopoietic precursors in the Drosophila larva A Dros Res Conf 2003 44 A321
Franc NC Heitzler P Ezekowitz RA White K Requirement for croquemort in phagocytosis of apoptotic cells in Drosophila
Science 1999 284 1991 1994 10373118
Dai JD Gilbert LI Metamorphosis of the corpus allatum and degeneration of the prothoracic glands during the larval-pupal-adult transformation of Drosophila melanogaster A cytophysiological analysis of the ring gland Dev Biol 1991 144 309 326 1901285
Riddiford LM Bate M Arias AM Hormones and Drosophila development The development of Drosophila melanogaster 1993 Plainview Cold Spring Harbor Laboratory Press 899 939
Garen A Kauvar L Lepesant J-A Roles of ecdysone in Drosophila development Proc Natl Acad Sci U S A 1977 74 5099 5103 16592466
Gaziova I Bonnette PC Henrich VC Jindra M Cell-autonomous roles of the ecdysoneless gene in Drosophila development and oogenesis Develop 2004 131 2715 2725
Biyasheva A Do TV Lu Y Vaskova M Andres AJ Glue secretion in the Drosophila salivary gland: A model for steroid-regulated exocytosis Dev Biol 2001 231 234 251 11180965
Cherbas L Hu X Zhimulev I Belyaeva E Cherbas P EcR isoforms in Drosophila Testing tissue-specific requirements by targeted blockade and rescue Development 2003 130 271 284 12466195
Hoffmann JA Reichhart JM
Drosophila innate immunity: An evolutionary perspective Nat Immunol 2002 3 121 126 11812988
Qiu P Pan PC Govind S A role for the Drosophila Toll/Cactus pathway in larval hematopoiesis Development 1998 125 1909 1920 9550723
Rizki MTM Tumor formation in relation to metamorphosis in Drosophila melanogaster
J Morphol 1957 100 459 472
Hershfield MS New insights into adenosine-receptor-mediated immunosuppression and the role of adenosine in causing the immunodeficiency associated with adenosine deaminase deficiency Eur J Immunol 2005 35 25 30 15580654
Kurata S Saito H Natori S The 29-kDa hemocyte proteinase dissociates fat body at metamorphosis of Sarcophaga
Dev Biol 1992 153 115 121 1516741
Yano T Takahashi N Kurata S Natori S Regulation of the expression of cathepsin B in Sarcophaga peregrina (flesh fly) at the translational level during metamorphosis Eur J Biochem 1995 234 39 43 8529666
Hori S Kobayashi A Natori S A novel hemocyte-specific membrane protein of Sarcophaga (flesh fly) Eur J Biochem 2000 267 5397 5403 10951197
Blackburn MR Volmer JB Thrasher JL Zhong H Crosby JR Metabolic consequences of adenosine deaminase deficiency in mice are associated with defects in alveogenesis, pulmonary inflammation, and airway obstruction J Exp Med 2000 192 159 170 10899903
Zhou X Riddiford LM Broad specifies pupal development and mediates the “status quo” action of juvenile hormone on the pupal-adult transformation in Drosophila and Manduca
Development 2002 129 2259 2269 11959833
Zhou X Zhou B Truman JW Riddiford LM Overexpression of broad: A new insight into its role in the Drosophila prothoracic gland cells J Exp Biol 2004 207 1151 1161 14978057
Apasov S Chen JF Smith P Sitkovsky M A(2A) receptor dependent and A(2A) receptor independent effects of extracellular adenosine on murine thymocytes in conditions of adenosine deaminase deficiency Blood 2000 95 3859 3867 10845921
Park S Lim JK A microinjection technique for ethanol-treated eggs and a mating scheme for detection of germ line transformants Drosoph Inf Serv 1995 76 187 189
Rizki TM Rizki RM Grell EH A mutant affecting the crystal cells in Drosophila melanogaster
Wilhelm Roux's Arch 1980 188 91 99
Maier SA Podemski L Graham SW McDermid HE Locke J Characterization of the adenosine deaminase-related growth factor (ADGF) gene family in Drosophila
Gene 2001 280 27 36 11738815
| 15907156 | PMC1135298 | CC BY | 2021-01-05 08:21:24 | no | PLoS Biol. 2005 Jul 24; 3(7):e201 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030201 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030231SynopsisBioinformatics/Computational BiologyBotanyEcologyEvolutionGenetics/Genomics/Gene TherapyPlant ScienceStatisticsArabidopsis (Thale Cress)Patterns of Genetic Variation Reveal Plant's Evolutionary Roots Synopsis7 2005 24 5 2005 24 5 2005 3 7 e231Copyright: © 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.
The Pattern of Polymorphism in Arabidopsis thaliana
==== Body
Biologists have developed ever more sophisticated ways to find molecular traces of natural selection. These traces—which occur as variations in DNA sequence, or polymorphisms, between and within species—are thought to harbor the genetic basis of adaptive events.
The study of natural selection at the molecular level has long been dominated by Kimura's theory of neutral evolution, which argues that most polymorphisms (in both DNA and protein sequence) have minor or no selective effect, and are governed by random, not selective, processes. The strength of this theory is that it leads to clear predictions that can be tested to identify those polymorphisms that really are subject to selection. Even though there's a large body of literature devoted to the statistical testing of selective neutrality, these tests are generally based on theoretical models, the assumptions of which have largely been untested. Only recently have the necessary quantities of data for testing the merits of these models become available, thanks to high-throughput genotyping and sequencing technologies.
Working with Arabidopsis thaliana, the first plant genome sequenced, Magnus Nordborg, Joy Bergelson, and their colleagues investigate a global survey of polymorphism patterns in the genomes of 96 plants. The scale of their study affords robust insights into the genomic pattern of polymorphism of the plant and sheds light on its demographic history. The results also lay the foundation for future work on the genetic basis of A. thaliana variation while challenging the assumptions of standard mathematical models for determining whether a gene is under natural selection in the plant.
Nordborg and colleagues sequenced 876 short genome fragments of 96 A. thaliana plants from both worldwide natural populations and laboratory stocks. In total, they described 44,000,000 DNA bases of genetic material, which revealed 17,000 polymorphisms, either in the form of single changes in DNA sequence (single nucleotide polymorphisms, or SNPs) or as losses or additions of DNA sequence between individual plants.
The level of polymorphism in A. thaliana is unexpectedly high for a plant that is highly self-fertilizing. To see if these polymorphisms were uniformly shared across plant populations, or had a distinct structure, Nordborg and colleagues grouped a subset of the A. thaliana plants into populations based on their geographic origin. Although they found that individuals within a population harbor much of the variation that is typical of the species worldwide, it appeared that some of the variation was specific to particular geographic regions. Furthermore, closely related plants were almost always from the same local region. The authors suggest that is what would be expected for a sexually reproducing species found worldwide.
The global geographic structure of variation in Arabidopsis is shown by the clusters of similar pie charts signifying patterns of isolation (each chart represents an Arabidopsis accession)
With such a large data set, it also becomes possible to see if the underlying assumptions of mathematical models commonly used for determining whether a gene is under selection are appropriate for A. thaliana. Nordborg and colleagues found that the patterns they observe do not fit the standard neutral model of evolution, which is expected to explain most genetic variation. This model is the benchmark against which researchers pinpoint the signature of selection at particular genes. The authors caution that “commonly used ‘tests of selection’ are simply not valid in A. thaliana.”
Nordborg and colleagues have provided a wealth of detail to our understanding of genetic variation in Arabidopsis on a genome-wide scale. Future research can now begin to use this Arabidopsis genetic footprint to find the exact variations that contribute to useful plant traits—and plumb its genome for evolutionary clues.
| 0 | PMC1135299 | CC BY | 2021-01-05 08:28:15 | no | PLoS Biol. 2005 Jul 24; 3(7):e231 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030231 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030232SynopsisCell BiologyDevelopmentGenetics/Genomics/Gene TherapyDrosophila
Drosophila Larval Development and Human Immunodeficiency: The Adenosine Deaminase Connection Synopsis7 2005 24 5 2005 24 5 2005 3 7 e232Copyright: © 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.
A Role for Adenosine Deaminase in Drosophila Larval Development
==== Body
For most healthy individuals, infection triggers a rapid immune response that repels the invaders. But for those rare individuals born without the immune system cells (lymphocytes) that recognize and kill pathogens, bacterial, viral, or fungal encounters can result in recurrent infections that are more life-threatening and less responsive to treatment than similar infections in normal infants. In the past, all that could be done for children with severe combined immunodeficiency (SCID) was to protect them from infections by cocooning them in sterile plastic bubbles, which gave the disease its common name: bubble-boy syndrome. Nowadays, the treatment of choice, provided a suitable donor is available, is bone-marrow or stem-cell transplantation, which provides SCID children with a functioning immune system.
Mutations in at least nine genes can cause human SCID, but 20% of cases are caused by a deficiency of the enzyme adenosine deaminase. This enzyme, which is present in all organisms, converts adenosine and deoxyadenosine to inosine and deoxyinosine, respectively. When adenosine deaminase is missing, its substrates (adenosine and deoxyadenosine) accumulate, and this is thought to cause the complete breakdown in immune defense characteristic of SCID.
To explore the role of adenosine deaminase in a tractable model system, Peter Bryant and his colleagues have now developed a Drosophila model by disabling the expression of a protein—called adenosine deaminase-related growth factor A (ADGF-A)—that serves as a major adenosine deaminase in the fly. In flies lacking ADGF-A enzymatic activity, adenosine and deoxyadenosine concentrations increase in the larval hemolymph, the circulatory fluid or “blood” of insects. Lack of the enzyme, the researchers report, caused larval death associated with the disintegration of the fat body (the adipose tissue spread throughout the body of the insect), melanotic tumors, and delays and defects in development.
The first two effects, fat body disintegration and the presence of melanotic tumors, are directly attributable to dysregulation of hemocytes (fly blood cells) in the mutant animals. It turns out that in adgf-a-mutant larvae, hemocytes are released prematurely from the lymph glands (the organs where hemocytes are produced and stored). These prematurely released hemocytes then cause fat body disintegration and formation of melanotic tumors; however, if ADGF-A expression is selectively restored in the lymph glands, then hemocytes are not prematurely released, and the larvae survive and develop without the tumors or fat body disintegration.
Bryant and his colleagues reasoned that the elevated adenosine might have direct effects on fly development aside from the dysregulation of hemocytes, so they examined the development of adgf-a mutant flies that also lacked a functional adenosine receptor (adoR mutants). They found that adgf-a/adoR mutant larvae were able to survive and continue development to adulthood, although these animals still experienced fat body disintegration and some melanotic tumors. These results suggest that the second consequence of ADGF-A deficiency, delayed development, is caused by the elevated adenosine in the animals signaling through adenosine receptors.
Drosophila
adgf-a mutant larvae with melanotic tumors in their body cavities
Altogether, these results establish adgf-a flies as a useful model system for unraveling the many effects that adenosine and deoxyadenosine have on cellular physiology in general and on the immune system in particular. Because hemocyte release from lymph glands and delays in development also occur in response to infection, the authors hypothesize that adenosine might be involved in controlling hemocyte release and postponing development when fly larvae are challenged by microbial attacks. Future experiments in this model system should provide important clues to the pathology of adenosine deaminase deficiency–associated SCID and should also advance our understanding of how adenosine acts as a stress hormone during infections in individuals with normal immune systems.
| 0 | PMC1135300 | CC BY | 2021-01-05 08:28:16 | no | PLoS Biol. 2005 Jul 24; 3(7):e232 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030232 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-841580788910.1186/1471-2105-6-84Research ArticleScoring functions for transcription factor binding site prediction Friberg Markus [email protected] Rohr Peter [email protected] Gaston [email protected] Institute of Computational Science, ETH, 8092 Zurich, Switzerland2005 4 4 2005 6 84 84 12 11 2004 4 4 2005 Copyright © 2005 Friberg 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
Transcription factor binding site (TFBS) prediction is a difficult problem, which requires a good scoring function to discriminate between real binding sites and background noise. Many scoring functions have been proposed in the literature, but it is difficult to assess their relative performance, because they are implemented in different software tools using different search methods and different TFBS representations.
Results
Here we compare how several scoring functions perform on both real and semi-simulated data sets in a common test environment. We have also developed two new scoring functions and included them in the comparison. The data sets are from the yeast (S. cerevisiae) genome.
Our new scoring function LLBG (least likely under the background model) performs best in this study. It achieves the best average rank for the correct motifs. Scoring functions based on positional bias performed quite poorly in this study.
Conclusion
LLBG may provide an interesting alternative to current scoring functions for TFBS prediction.
==== Body
Background
The TFBS prediction problem can be defined as follows: Given N hypothetically co-regulated genes and their promoter sequences S = {S1, S2, ..., SN} (typically 1000 bp upstream of each gene, although they can be much longer in higher eukaryotes), search for motifs that are overrepresented in S compared to the set A of all promoter sequences in the genome. Ideally, the most overrepresented motif is the TFBS. A recent review on both biological and computational aspects of TFBS prediction is [1]. Another review focusing more on the computational aspects is also available [2].
Many software tools exist for TFBS prediction, e.g. Consensus [3], MEME [4,5], AlignACE [6], BioProspector [7], and MDscan [8]. These tools can be classified according to three criteria:
1. TFBS representation: How a putative TFBS is represented, e.g. consensus sequence [9,10], PSFM (position specific frequency matrix) [7], Bayesian network [11] and HMM [12].
2. Search method: How promoter sequences are searched for putative TFBSs, e.g. greedy search [3], Gibbs sampling [13] and deterministic iterative search [8].
3. Scoring function: How a newly found PSFM (or any other TFBS representation) is scored to distinguish real binding sites from background noise.
In this paper we will focus on the scoring function, which is a crucial part of any TFBS prediction software. Many scoring functions for TFBS prediction have been proposed in the literature. Unfortunately, it is difficult to assess their relative performance because they are implemented in different software tools that:
1. use different TFBS representations.
2. use different search methods.
3. are tested on different data sets in the original papers.
We compare how several scoring functions perform on both real and semi-simulated data sets in a common test environment. We also develop two new scoring functions and include them in the comparison.
Results
Scoring function performance
Six scoring functions (described in detail in the Methods section) have been evaluated in this study. The scoring functions were tested on eight different yeast data sets (Fig. 1). In order to compare the performance of the different scoring functions, the rank of the correct motif is shown. Lower rank is better, since the rank is the position of the correct motif in the list of all potential motifs, sorted according to the score from each scoring function.
MAP is performing quite well on most data sets, except for mac1. The same holds for Group Specificity. Positional Bias performs poorly for most data sets, except reb1, abf1 and rap1. Local Positional Bias is clearly better than Positional Bias, but has problems with abf1 and mig1. A closer look at the poor performance of Local Positional Bias for abf1 reveals that the positions of the correct binding sites are not clearly localized to a certain region of the promoter sequences. There is only a weak local positional bias (0.18), and many random motifs show a higher local positional bias.
LLBG performs well for all data sets and is the best scoring function in this comparison. The linear combination of LLBG and Local Positional Bias performs well in general, but has some problems with mac1, where it interestingly performs worse than LLBG alone. The reason is that Local Positional Bias performs significantly worse than LLBG for this data set. Please note that ranks are shown in the graph, so the combined score is not a linear function of the bar heights of LLBG and Local Positional Bias, but a linear function of the actual score values (data not shown) of the LLBG and Local Positional Bias.
In order to avoid overfitting, the value of the Local Positional Bias weight a2 was estimated based on all data sets but the current one. The value of a2 was around 0.5 for all data sets (the LLBG weight a1 was fixed to 1).
Addition of noise
Generally, it is not realistic that all sequences in a data set contain the binding site of interest. Often, when the set of potentially co-regulated genes is defined by microarray experiments (gene expression profiling), there are false positives (genes where the TFBS of interest is not present) in the data set. In order to evaluate the scoring functions in the presence of this biologically relevant noise, between 10 and 30 promoter sequences were randomly selected from the genome and added to the reb1 (Fig. 2) and mig1 (Fig. 3) data sets (in the original comparison without noise, reb1 was the easiest data set to predict, and mig1 was a more difficult one, for which the best scoring functions performed about equally well). The results shown are the average of ten independent runs, with different randomly selected promoter sequences added in each run. Table 4 contains a summary of the results.
reb1
In the reb1 data set (Fig. 2) the MAP score performs well for reb1+10 (the original reb1 data set plus 10 randomly selected upstream sequences). However, for reb1+20 and reb1+30, performance decreases quickly. Group Specificity shows a similar trend, but is clearly better than MAP for reb1+20 and reb1+30. Interestingly, Positional Bias performs extremely well on this data set. However, because of its general bad performance (Fig. 1), we should not put too much confidence into this scoring function. Local Positional Bias performs consistently poorly on this data set. LLBG does extremely well on reb1+10 and reb1+20, and for reb1+30 it also shows a good result. The combined score performs quite well, but consistently worse than LLBG alone. The reason, as we can see, is that Local Positional Bias does not perform well on this data set.
mig1
The mig1 data set (Fig. 3) was more difficult. The reason seems to be that the mig1 motif shows higher variability between the different promoter sequences than does the reb1 motif. MAP performed quite poorly, especially on mig1+20 and mig1+30. Group Specificity, Positional Bias and Local Positional Bias failed already at mig1+10. LLBG performed best in every case, slightly better than the combined scoring function on mig1+10 and mig1+20, and significantly better on mig1+30.
All in all, LLBG seems to be the best scoring function in this study.
Discussion
Choice of search method
The iterative deterministic search method was used in this study because it has been shown [8,14] to suffer less from local optima than e.g. Gibbs sampling. However, the scoring functions tested here can score any arbitrary set of candidate words, no matter how these words are selected, so the relative performance of the different scoring functions should not depend on the choice of search method. The only interaction between the search method and the scoring function is that the search method provides the scoring function with several sets of candidate motifs to score.
Determination of parameters
Most scoring functions have one parameter where the value is not directly determined by the data: MAP (Markov model order), Group Specificity (s1), Local Positional Bias (Lw) and LLBG (Markov model order). In this respect, these scoring functions are similar. The exception is Positional Bias, which has two parameters (tm and Lw). We also note that the cardinality of the parameters are different. For example, only a few discrete values are reasonable for the Markov model order. On the other hand, e.g. the Group Specificity s1 parameter has a larger range of possible values.
We have not tried to find the optimal parameters of each scoring function, but used the values proposed in the original papers. Since yeast is often used as a model organism for TFBS prediction studies, we assumed that the default parameters are reasonable for the yeast data sets in this study. Also, to make the comparison as fair as possible, we used the same Markov model order (3) in LLBG as was used in MAP.
Positional bias
Because of the high variance in performance of the positional bias based scoring functions, it seems that positional bias is a feature of only a few of the data sets in this study. For others, it seems that the positions of the TFBSs do not deviate strongly from a random distribution. This has also been observed previously [6]. When a large number of long promoter sequences are searched for motifs, many candidate motifs have to be considered. If a scoring function only deviates slightly from the random distribution, many false positives will be found, which is the case of Positional Bias and Local Positional Bias.
Advantages of LLBG
Robustness is an important property of TFBS scoring functions. The tests performed on reb1 and mig1 with added noise indicate that LLBG is quite robust against this form of biologically realistic noise, more robust than the other scoring functions in this test.
Many software tools for TFBS prediction require the user to specify the motif width w as a parameter. This is of course difficult when the motif is unknown and makes these tools impractical to use. The MAP scoring function is normalized by the motif width, which should make it comparable for motifs of different widths [8]. However, scaling the entropy part of MAP is problematic, as pointed out by [14]. The LLBG score does not have entropy as a part of its function, and hence should not suffer from this problem.
Possible extensions
The LLBG score, as it is currently defined, measures the probability of a motif occurring at least once in the promoter sequence. It is possible to extend it to how many times a motif occurs in a sequence, which would increase its performance on data sets with several TFBSs per sequence. However, initial experiments have indicated that a multi-motif-per-sequence version of LLBG did not improve the results for the yeast data sets that we have been working with so far.
Even though we have here treated only the problem of finding a single motif, it is possible to generalize the LLBG scoring function to clusters of different motifs. This is of special interest in higher eukaryotes and will be subject to future work.
Currently, the LLBG is based on the discrete distance measure of number of mismatches between a candidate word and the PSFM consensus. Future research will go into using a continuous distance between a candidate word and the PSFM (not the PSFM consensus). This should make the score more robust, especially for long motifs with many uninformative positions, and it should further improve the performance of this scoring function.
We have here focused on the problem of de novo prediction of TFBSs. The related problem of TFBS recognition, where a library of known TFBSs is used to search for similar motifs in S, has not been considered here. Clearly, these libraries may improve TFBS prediction if the TFBS of interest happens to be similar to a TFBS already documented. Conceptually, this can easily be incorporated in the LLBG scoring function by studying the likelihood ratio between the TFBS library model and the background model.
Limitations of current models
It should be noted that although the best scoring functions perform reasonably on these yeast data sets (with promoter regions of 1000 bp), the problem becomes much more difficult when dealing with higher eukaryotes (with promoter regions of more than 10000 bp). In that case, all of these scoring functions are likely to have problems (because of low signal-to-noise ratio), and it becomes more important to extend the models by including other sources of information, such as ChIP-chip and phylogenetic footprinting data. Since nature is able to find TFBSs with higher precision than any of the scoring functions reviewed here, we believe that the current computational models are missing some fundamental part of the transcription regulation mechanism. Future research will go into investigating the structural properties of DNA that enables transcription [15]. Interesting progress on work in this direction has been done recently for prokaryotes [16], and the related histone code has been suggested for eukaryotes [17].
Conclusion
The time requirements of the scoring functions in this study are very different. LLBG, MAP and Local Positional Bias are relatively fast to evaluate. Group Specificity and Positional Bias are significantly more time consuming, since they require a search of a PSFM in all intergenic sequences. Since the two slower scoring functions do not perform better than three faster ones, their longer computation time does not seem to be justified.
The Positional Bias and Local Positional Bias are scoring functions that perform quite poorly for several data sets (e.g. mig1), but quite well for others (e.g. reb1 without the added noise). In other words, it seems that this feature is not relevant for some data sets (that the position of the binding site in the upstream sequence is clearly different between genes), but that it clearly matters for other data sets. This makes these scoring functions difficult to use for de novo TFBS prediction, since we cannot know beforehand whether an unknown TFBS is positionally biased.
LLBG is the scoring function that performs best in this test. The other scoring functions perform well on some data sets and poorly on others. Combining LLBG and Local Positional Bias results in a scoring function that on average performs slightly worse than LLBG alone. Since there is no clear improvement in combining the scoring functions (Fig 1, 2 and 3), the simpler solution of using only LLBG should be preferred.
A software tool using the LLBG scoring function is currently being developed.
Methods
TFBS representation
The PSFM representation is used for all comparisons, since it provides a good approximation of the specific protein-DNA interactions [18], and since it seems to be the representation most commonly used in the literature. The PSFM is a matrix consisting of the frequency of each nucleotide at each motif position.
Search method
A deterministic iterative search method similar to [8,14] was used, since it thoroughly searches the promoter sequences in quite reasonable time (less than a minute for most data sets of typical size, around 20 sequences of 1000 bp each). Furthermore, it does not suffer from the problem of local minima to the same degree as Gibbs sampling, as pointed out in [8,14]. In short, it works as follows:
The first word Wb of width w in S1 (position 1..w) is chosen as base word. A candidate set of words is collected for which the hamming distance to the base word is at most m, a threshold which is determined empirically as a function of w [14]. Each sequence Si may contribute zero, one or several words to the set of candidate words. A PSFM is created from all the Nc candidate words by computing the frequency of each nucleotide at each position. This PSFM is iteratively refined by removing words until a scoring function is maximized. This is repeated for all base words Wb in S1 (position 2..w + 1, 3..w + 2, ...), generating one PSFM for each Wb. In order to improve the results, a search is done using each PSFM against S to define a new set of candidate words, which are again iteratively refined a maximum number of times or until convergence. For the purpose of searching for a PSFM in S, pseudo-counts are used as described in [13].
The data sets in this study were known to have a motif occurrence in S1, so we only used base words from S1 for efficiency reasons, as was done in [14]. However, in general this cannot be assumed, so base words should normally be chosen from several (or all) sequences in S. Base words were collected from both strands.
Scoring functions
The following scoring functions were tested: MAP [8], Group Specificity and Positional Bias [6]. We also introduced two new scoring functions: LLBG and Local Positional Bias, and included them in the comparison.
MAP
The MAP (maximum a posteriori probability) score is used in MDscan [8]. It is a combination of the negative entropy of the PSFM and the rareness of the PSFM according to a 3rd order Markov model estimated from all intergenic regions of a genome:
where w is the width of the motif, xm is the number of candidate words (m-matches) in the PSFM, pij is the frequency of nucleotide j at position i of the PSFM and p0(s) is the probability of generating the candidate word s from the background model. We computed p0 using a 3rd order Markov model in the following way (for the example word ACAGT):
p0(ACAGT) = p(ACA)p(G|ACA)p(T|CAG) (2)
The first part of the MAP score is the negative entropy, which is higher for PSFMs with more similar candidate words. A PSFM with identical words has maximum negative entropy (0), and a PSFM where all nucleotides are equally frequent at each position has minimum negative entropy (-2w). Naturally, true TFBSs are expected to be similar (and in some rare cases even identical) words.
Group specificity
This score is used in AlignACE [6]. It measures how well a given motif is localized to the set of input sequences S compared to all non-coding sequences. The rationale is that the true TFBS is a motif that is clearly more frequent in the selected promoter sequences than in all promoter sequences.
All promoter sequences in the genome are searched for the motif PSFM. The set of sequences from the top s1 hits are intersected with S, and the probability that these two sets would have the observed intersection or greater (Group Specificity score) is calculated:
where T is the total number of promoter sequences, s1 is the top number of genes (typically 100), s2 is the number of sequences in S, and x is the number of ORFs in the intersection of the two lists. This score has the advantage (compared to a kth-order Markov model) that it estimates how rare the motif PSFM is, not the rareness of parts of each candidate word in the PSFM.
Positional Bias
Like Group Specificity, this score was proposed in AlignACE [6]. It measures the concentration of motifs within a certain distance from the transcriptional start site. The rationale is that TFBSs tend to be located at the same distance from the transcriptional start site of each gene. Since the transcriptional start site is difficult to map, the translational start site is used as a reasonable approximation. The top tm (typically 200) PSFM hits in the genome for a given motif are found and their positions relative to the nearest ORF start are extracted. Among these, the t PSFM hits that are found within L bp upstream of some ORF are considered further. Let mw be the largest number of hits found in any Lw bp window of the upstream sequences. The probability (Positional Bias) of observing mw or more sites out of the maximum possible t is determined by a binomial distribution:
where
In the original paper, Lw = 50 and L = 600 were used. Because of longer upstream sequences in the data sets used in this paper (L = 1000), we chose a larger window (Lw = 100).
Local Positional Bias
Initial tests showed that the original Positional Bias performed poorly on some data sets, so we decided to try a modified version of it. We call this new scoring function Local Positional Bias. It differs from Positional Bias in two ways:
1. The positional bias is measured in the input sequences S instead of the whole genome.
2. The bias of all windows is considered (using a χ2-test) instead of only the window with the largest number of sites. This makes it less sensitive to noise.
The Local Positional Bias Ploc is defined as the probability (according to a χ2-test) that the positions of the sampled motifs were generated from a model where all positions are equally probable. For example, consider Fig. 4 with 16 motif occurrences distributed over w = 10 windows. The test statistic is computed as:
where Ci is the number of occurrences in window i and E(Ci) is the expected number of occurrences in each window, in this example 16/10 = 1.6. The χ2 test statistic has f = w - 1 degrees of freedom, in our case f = 10 - 1 = 9. In our example, χ2 = 25.25, which leads to a probability (Local Positional Bias) of 0.0027 according to the cumulative χ2-distribution.
LLBG
In the LLBG (least likely under the background model) score we consider the likelihood that a motif occurs at least once in M promoter sequences out of N under a background model. The idea is that the TFBS is the motif that is least likely to have been produced in these promoter sequences by the background model. The trade-off between more motif occurrences and lower probability according to the background model is treated in a probabilistic manner.
Given a set of candidate words (and a PSFM created from these words), we define the worst candidate word Ww as the word with the largest Hamming distance dmax to the PSFM consensus sequence Wc. Let peb be the probability that a randomly chosen word is at most dmax from the consensus:
peb = Pr[d(W, Wc) ≤ dmax] (7)
where d is the Hamming distance function and W is a random word from the 3rd order Markov model of all intergenic sequences. In the trivial case where all candidate words are identical (dmax = 0), there is only one word for which d(W, Wc) ≤ dmax. In this case, peb = p0(Ww), where p0 is computed like in equation (2). If Ww is at distance 1 from Wc, we have to sum the p0 of 1 + 3w different words. In the general case, we have to consider nw words:
and sum the p0 of all nw words to get peb. However, it becomes intractable to sum over all the possibilities for many mismatches in long motifs. Instead, we use the approximation:
where pavg is the average p0 of the total set of words with Hamming distance at most dmax from Wc and is the average p0 of the candidate words. While pavg is often intractable to compute, can be computed much faster. In practice, the candidate words serve as a good representation of the total set of words of interest. Furthermore, the variance of p0 of different words within these sets is usually low, which makes very close to pavg.
Having defined peb, the probability of the motif occurring at least once in a promoter sequence of length L (assuming that the motif has equal probability to occur at all positions) is:
p1s = 1 - (1 - peb)L-w+1 (10)
For small values of peb, we can approximate equation (10) using Maclaurin polynomials:
p1s ≈ (L - w + 1)peb (11)
Holding the first occurrence of the motif fixed, the probability of the motif occurring in at least M - 1 additional promoter sequences out of the total N - 1 is:
which is our LLBG score. Since each occurrence is compared to the consensus, it can be argued that we should consider M occurrences. The reason that we consider M - 1 instead of M additional promoter sequences is that the latter introduces a bias when comparing scores from motifs of different lengths (long motifs with very few occurrences get too high scores). In the extreme case, consider a very long motif with only one occurrence: If M were used, this motif would get higher scores than all the biologically relevant motifs. By removing one occurrence (the most consensus like) and only considering the remaining M - 1 occurrences, this problem is solved.
LLBG + Local Positional Bias
In addition, we considered a combined scoring function of LLBG and Local Positional Bias. The correlation coefficient between these scoring functions is low (Table 1) and they are independent (within one standard deviation) according to a χ2-test (Table 2).
Generally, scores can be combined by converting them to normal deviates (using the z-transform) and summing them together. However, LLBG and Local Positional Bias deviate quite clearly from a normal distribution (data not shown), so we do not consider this option. Instead, since these scores are probabilities, we can combine them by adding their logarithms together:
comb = log pms + log Ploc (13)
However, as pointed out in [19], giving the scores equal weight often causes problems since one scoring function may be so dominating that the contribution of the other is practically ignored. Instead, we give each scoring function a unique weight:
comb = a1 log pms + a2 log Ploc (14)
In practice, we are only interested in the relative weighting between the two scores, so we fix one (a1 = 1) and find the optimal value of the other. In order to avoid overfitting, we do an n-fold cross validation (e.g. when combining the scores for the reb1 data set, the parameters are optimized on all data sets except reb1).
As was shown in the results section, this combined scoring function did not perform better than LLBG alone.
Data sets
The scoring functions were tested on data sets from [14]. These data sets consist of the promoter sequences (1000 bp upstream regions) of genes regulated by a certain transcription factor in yeast (Table 3). The 1000 bp upstream regions were used even in those cases where it overlaps another ORF. Each promoter sequence has at least one (putative or biologically verified) binding site of the transcription factor of interest. Most data sets are compiled from different molecular biology studies, where each binding site is biologically verified (Table 3). It is possible (and likely) that the promoter sequences also contain some binding sites of other transcription factors, but this is not considered here, as we do not have sufficient information on this. The width of the correct motif was given to the search algorithm. This is not needed by the search algorithm or scoring functions used in this comparison, but it facilitates the definition of what should be regarded as the correct motif. This simplification has also been used in comparisons of TFBS prediction software [14].
Each of these data sets were analyzed using each of the scoring functions. All PSFMs generated by the search method were scored using each of the scoring functions, and the complete list of PSFMs was sorted according to the score (highest first). If the consensus motifs of several PSFMs were identical, the lower scoring duplicate PSFMs were removed from the list. The rank (position in the sorted list) of the correct motif was compared between the different scoring functions for each data set.
As 'correct motif' we considered the known TFBS consensus (Table 3), either exact or shifted one position. Long motifs (i.e. at least ten informative positions) are allowed to have one mismatch or be shifted up to two positions. The reason for this relaxed definition is that the reported TFBS consensus motif for some data sets is not absolutely correct. For example, the given GAL4 binding site differs from recent findings from ChIP-chip experiments, as pointed out by [14]. Also, sometimes the first or last positions of the consensus are about as significant as the positions outside of these two (and hence not contained in the consensus sequence). This relaxed definition makes the comparison less sensitive to random fluctuations.
Authors' contributions
GG introduced the LLBG scoring function. PvR introduced statistical ideas and performed statistical analysis. MF further developed the LLBG scoring function, implemented the scoring functions and search method, and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Adrian Schneider and Daniel Margadant for helpful discussions. We also thank the anonymous reviewers, whose constructive comments improved the manuscript.
Figures and Tables
Figure 1 Comparison of scoring functions on eight different data sets (lower rank is better)
Figure 2 Comparison of scoring functions on the reb1 data set with different amounts of added noise. The average of ten independent runs is shown (lower average rank is better).
Figure 3 Comparison of scoring functions on the mig1 data set with different amounts of added noise. The average of ten independent runs is shown (lower average rank is better).
Figure 4 Example of Local Positional Bias calculation: A promoter sequence of 1000 bp is split into windows of 100 bp each. 16 motif occurrences are distributed over the 10 windows.
Table 1 Correlation coefficients between scores
GroupSpec PosBias LocPosBias LLBG
MAP 0.36 0.06 0.02 0.43
GroupSpec -0.19 -0.02 0.26
PosBias 0.01 0.02
LocPosBias 0.04
Table 2 χ2 independence test (standard deviations)
GroupSpec PosBias LocPosBias LLBG
MAP 8.48 0.66 1.25 9.00
GroupSpec 3.86 1.09 5.62
PosBias 0.47 1.27
LocPosBias 0.99
Table 3 Data sets of promoter sequences of genes regulated by different transcription factors. By 'molecular biology approaches' we mean methods like DNAse footprinting and methylation interference. 'AlignACE' stands for functional group data from the AlignACE web server. '#seqs' stands for number of promoter sequences in the data set.
TF TFBS consensus motif #seqs source and type of evidence
abf1 CGTNNNNNNTGA 20 molecular biology approaches [20]
gal4 CGGNNNNNNNNNNNCCG 10 molecular biology approaches [20] and AlignACE [6]
mac1 TTTGCTCA 6 microarray [21]
mcm1 TTTCCCAAANNGGAAA 24 molecular biology approaches [20]
mig1 AAAAATCTGGGG 11 molecular biology approaches [22]
pdr TCCGCGGA 11 AlignACE [6]
rap1 TACACCCATACATT 44 molecular biology approaches [23] [24]
reb1 TTACCCG 13 molecular biology approaches [20]
Table 4 Schematic interpretation of the results for the reb1 and mig1 data sets with added noise, good: top 3, ok: top 10, bad: worse than top 10
MAP GroupSpec PosBias LocPosBias LLBG LLBG LocPosBias
reb1 good good good ok good good
reb1+10 good good good bad good good
reb1+20 bad ok good bad good ok
reb1+30 bad ok good bad ok ok
mig1 good good bad bad good good
mig1+10 ok bad bad bad good ok
mig1+20 bad bad bad bad ok bad
mig1+30 bad bad bad bad bad bad
==== Refs
Bulyk ML Computational prediction of transcription-factor binding site locations Genome Biol 2003 5 201 14709165 10.1186/gb-2003-5-1-201
Pavesi G Mauri G Pesole G In silico representation and discovery of transcription factor binding sites Brief Bioinform 2004 5 217 236 15383209
Hertz GZ Stormo GD Identifying DNA and protein patterns with statistically significant alignments of multiple sequences Bioinformatics 1999 15 563 577 10487864 10.1093/bioinformatics/15.7.563
Bailey TL Elkan C Fitting a mixture model by expectation maximization to discover motifs in biopolymers Proc Int Conf Intell Syst Mol Biol 1994 2 28 36 7584402
Bailey TL Elkan C The value of prior knowledge in discovering motifs with MEME Proc Int Conf Intell Syst Mol Biol 1995 3 21 29 7584439
Hughes JD Estep PW Tavazoie S Church GM Computational identification of cis-regulatory elements associated with groups of functionally related genes in Saccharomyces cerevisiae J Mol Biol 2000 296 1205 1214 10698627 10.1006/jmbi.2000.3519
Liu X Brutlag DL Liu JS BioProspector: discovering conserved DNA motifs in upstream regulatory regions of co-expressed genes Pac Symp Biocomput 2001 127 138 11262934
Liu XS Brutlag DL Liu JS An algorithm for finding protein-DNA binding sites with applications to chromatin-immunoprecipitation microarray experiments Nat Biotechnol 2002 20 835 839 12101404
van Helden J Rios AF Collado-Vides J Discovering regulatory elements in non-coding sequences by analysis of spaced dyads Nucleic Acids Res 2000 28 1808 1818 10734201 10.1093/nar/28.8.1808
Price A Ramabhadran S Pevzner PA Finding subtle motifs by branching from sample strings Bioinformatics 2003 19 II149 II155 14534184 14534184
Barash Y Elidan G Friedman N Kaplan T Modeling Dependencies in Protein-DNA Binding Sites Proc Seventh Annual Inter Conf on Computational Molecular Biology (RECOMB) 2003
Yada T Totoki Y Ishikawa M Asai K Nakai K Automatic extraction of motifs represented in the hidden Markov model from a number of DNA sequences Bioinformatics 1998 14 317 325 9632826 10.1093/bioinformatics/14.4.317
Lawrence CE Altschul SF Boguski MS Liu JS Neuwald AF Wootton JC Detecting subtle sequence signals: a Gibbs sampling strategy for multiple alignment Science 1993 262 208 214 8211139
Narasimhan C LoCascio P Uberbacher E Background rareness-based iterative multiple sequence alignment algorithm for regulatory element detection Bioinformatics 2003 19 1952 1963 14555629 10.1093/bioinformatics/btg266
Hoglund A Kohlbacher O From sequence to structure and back again: approaches for predicting protein-DNA binding Proteome Sci 2004 2 3 15202939 10.1186/1477-5956-2-3
Wang H Noordewier M Benham CJ Stress-induced DNA duplex destabilization (SIDD) in the E. coli genome: SIDD sites are closely associated with promoters Genome Res 2004 14 1575 1584 15289476 10.1101/gr.2080004
Jenuwein T Allis CD Translating the histone code Science 2001 293 1074 1080 11498575 10.1126/science.1063127
Benos PV Bulyk ML Stormo GD Additivity in protein-DNA interactions: how good an approximation is it? Nucleic Acids Res 2002 30 4442 4451 12384591 10.1093/nar/gkf578
De Hoon MJ Makita Y Imoto S Kobayashi K Ogasawara N Nakai K Miyano S Predicting gene regulation by sigma factors in Bacillus subtilis from genome-wide data Bioinformatics 2004 20 101 108 10.1093/bioinformatics/bth927
Zhu J Zhang MQ SCPD: a promoter database of the yeast Saccharomyces cerevisiae Bioinformatics 1999 15 607 611 10487868 10.1093/bioinformatics/15.7.607
Gross C Kelleher M Iyer VR Brown PO Winge DR Identification of the copper regulon in Saccharomyces cerevisiae by DNA microarrays J Biol Chem 2000 275 32310 32316 10922376 10.1074/jbc.M005946200
Klein CJ Olsson L Nielsen J Glucose control in Saccharomyces cerevisiae: the role of Mig1 in metabolic functions Microbiology 1998 144 13 24 9467897
Pavlidis P Furey TS Liberto M Haussler D Grundy WN Promoter region-based classification of genes Pac Symp Biocomput 2001 151 163 11262936
Lascaris RF Mager WH Planta RJ DNA-binding requirements of the yeast protein Rap1p as selected in silico from ribosomal protein gene promoter sequences Bioinformatics 1999 15 267 277 10320394 10.1093/bioinformatics/15.4.267
| 15807889 | PMC1140076 | CC BY | 2021-01-04 16:02:50 | no | BMC Bioinformatics. 2005 Apr 4; 6:84 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-84 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591338710.1371/journal.pmed.0020166Policy ForumEpidemiology/Public HealthHealth PolicyHealth PolicyPublic HealthPathways to “Evidence-Informed” Policy and Practice: A Framework for Action Policy ForumBowen Shelley *Zwi Anthony B Shelley Bowen is a doctoral candidate and Anthony B. Zwi is Professor and Head at the School of Public Health and Community Medicine at the University of New South Wales, Randwick, Australia.
Competing Interests: The authors declare that they have no competing interests.
*To whom correspondence should be addressed. E-mail: [email protected] 2005 31 5 2005 2 7 e166Copyright: © 2005 Bowen and Zwi.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.Bowen and Zwi propose a new framework that can help researchers and policy makers to navigate the use of evidence.
==== Body
The contemporary public health effort sees much debate about the concepts of “evidence” and “the evidence base”, and the usefulness and relevance of such terms to both policymaking and practice. A key challenge to public health is to better contextualize evidence for more effective policymaking and practice. Theory on the translation of research findings into policy and practice, and on knowledge utilization, offers only part of the solution to this complex task. The policymaking context is highly political and rapidly changing, and depends on a variety of factors, inputs, and relationships.
In this article, we propose that an “evidence-informed policy and practice pathway” can help both researchers and policy actors navigate the use of evidence (Figure 1). The pathway illustrates different types of evidence and their uses in health policymaking, and proposes that specific capacities, such as an individual's skills, experience, and participation in networks, influence the adoption and adaptation of evidence in practice.
Figure 1 The Evidence-Informed Policy and Practice Pathway
The pathway to “evidence-informed” policy and practice involves three active stages of progression, influenced by the policy context. The three stages are (1) sourcing the evidence, (2) using the evidence, and (3) implementing the evidence. The pathway also involves decision-making factors and a process which we have termed “adopt, adapt, and act”. Once adopted, evidence about implementation is usually adapted or changed before use in the policy context. Policy actors and practitioners rightfully need to understand and decide how best this evidence should be acted upon in each circumstance. Each stage in this pathway is underpinned by a variety of individual-, organizational-, and system-level values.
To formulate the evidence-informed policy and practice pathway presented in this paper, we reviewed relevant literature from health, public policy, and the social sciences.
Diffusion of Innovations: An Underlying Theory
Fundamental to the transfer of evidence into policy and practice is diffusion, the process by which an innovation is communicated over time among members of a social system (classical diffusion) [1]. In this paper we consider innovation to be the policy idea. Studies of innovation in health-care organizations by Lennarson Greer [2] proposed that diffusion theory helps us understand the following: (1) how individuals within an organization receive, adopt, and adapt evidence; (2) the organizational factors that constrain or facilitate the adoption or implementation of the evidence; and (3) the interests and values at play within organizations that influence responses to the evidence/policy issue.
The success of diffusion of evidence into policy and practice rests largely with the characteristics at play at each stage of the adoption process [3,4]. Information passes through an “adopt, adapt, and act” cycle. Characteristics of the individuals involved, the innovation itself, and the organizations in which they are considered affect decisions made about evidence in terms of the perceived value, priority given, and seriousness of response. The extent to which individual-, organizational-, and system-level values influence a decision to accept or reject the policy-related evidence is largely unexplored in the literature [5]. For example, the importance of values as a factor that influences the lack of action on health inequity has been poorly researched.
Stage 1: Sourcing the Evidence
Evidence-informed policymaking sees the use of different types of information in a variety of forms and from a variety of sources, reflective of, and responsive to, the policy and practice context. Types of evidence that inform the policy process can be grouped as research, knowledge/information, ideas/interests, politics, and economics (see Table 1) [6]. Evidence is usually sought to show effectiveness (“it works”), show the need for policy action (“it solves a problem”), guide effective implementation (“it can be done”), and show cost effectiveness (“it is feasible and may even save money”).
Table 1 Types of Evidence and How They Are Used in Policy Making
The term evidence-based policy is used in the literature, yet largely relates to only one type of evidence—research. Using the term “evidence-influenced” or “evidence-informed” reflects the need to be context sensitive and consider use of the best available evidence when dealing with everyday circumstances [7–9]. A variety of distinct pieces of evidence and sources of knowledge inform policy, such as histories and experience, beliefs, values, competency/skills, legislation, politics and politicians, protocols, and research results [10,11]. Policy analysis theory proposes that evidence is information (information is data that has meaning) “that affects existing beliefs of important people about significant features of the problem under study and how it might be solved or mitigated” [12]. A case study of applying evidence to policy and practice in the real world is described in Box 1.
Box 1. Case Study of Applying Evidence to Policy and Practice in the Real World
The Kings Fund recently reviewed investments by the United Kingdom government into major social programmes. The Fund's report, called Finding Out What Works, asks to what extent these programmes are evidence based, what is being done to find out if they work, and whether the evaluations are helping to inform policy and practice in the future [54].
The report concludes that programmes like Sure Start (a government programme involving early interventions to improve children's social and educational welfare; see http://www.surestart.gov.uk) are largely driven by “informed guesswork, expert hunches, political and other imperatives”.
In this case, the application of evidence to policymaking was hindered by a lack of good-quality, synthesized evidence, capacity to apply the evidence, and organizational support and resources to make evidence-based decisions.
A real conflict exists between local control of decision-making and the idea of evidence-informed decision-making and evidence-based policy and practice. The report's authors state that “decisions are guided by common sense and experience rather than the formal evidence base”.
The way in which research evidence is combined with other forms of information is key to understanding the meaning and use of evidence in policy development and practice. Current literature on evidence-based health care is often limited by inadequate attention to context [4,13,14]. A major challenge to contextualizing evidence for policymaking is recognition that a broad information base is required [4].
The Policy Context
Considering the evidence within the context in which it will be used is critical for effective policymaking and practice. The context is the environment or setting in which the policy is being developed and implemented [15], incorporating the historic, cultural, health services, system, and resource contexts. The social and political context and the many forces at work in the policy environment provide challenges to integrating evidence into policy and practice. Researchers often do not see or recognize these factors. The political, ideological, and economic factors influencing policy development and decision-making often gain strength at the expense of the research evidence; a recent example is the political commitment in Australia to establish a range of new medical schools rather than enable existing schools to train more people, which would likely be far more cost-effective.
The ways in which the evidence is used in the policy process are largely determined by the beliefs and values of policymakers, as well as by considerations of timing, economic costs, and politics [2,18–22]. The development of both the Black report [16] and the Independent Inquiry into Inequalities in Health Report in England [17] are illustrative. The Black report was dismissed by the Conservative government of the time despite extensive evidence on inequality. Almost 20 years later the Independent Inquiry into Inequalities in Health Report and subsequent policy actions were largely driven by government. How and when evidence is used often depends upon the political agenda and ideology of the government of the day, not on the nature of the evidence, however compelling [20,23].
Policy networks provide a useful lens through which to analyze the context of policymaking and research utilization. A policy network focuses on the relationships that shape the policy agenda and decision-making process. Networks can shape the way policy is formulated, and in particular the way in which evidence is gathered and presented in policy formulation [24,25]. Epistemic communities are formal and informal groups of technical “experts” who purvey information and share ideas about research data, knowledge, and experience. These communities gather, synthesize, and disseminate information about a policy issue, as well as advocate for knowledge transfer across social systems and government [26,27].
Factors in Decision-Making
The usefulness of the innovation
Decisions about the usefulness of an innovation itself are often based on relative advantage (is the innovation better than previous approach?); complexity (is the innovation understandable?); compatibility with values and past experiences; and cost and flexibility, reversibility, trialability, and revisability (is there opportunity to trial and change?) [1,2,28–30]. Potential adopters who see the innovation as compatible with their values and those of their organizations are more likely to adopt than those who do not [3]. The literature also suggests that organizations that are close to each other—geographically or in communication—will adopt innovations because of the “bandwagon effect” [3,31].
The rapid diffusion of new hospital equipment [32] versus the slow diffusion of ideas on sudden infant death syndrome [33] or policy to tackle health inequalities in health services [34] provide examples of how different ideas and forms of knowledge on different issues can result in different diffusion efforts and successes.
The influence of the individual
Individuals are key participants in decisions about use of evidence throughout the policy and practice pathway, as it is individuals who decide whether to accept or reject something new. Individual decisions are influenced by a variety of personal qualities and capacities such as values and beliefs, leadership, knowledge and skills, resources, organizational support, partnership links, and networking. Additionally, individuals are influenced by the perceived benefit of change, and, once again, by the complexity of the innovation itself. Individuals often avoid change, reinforcing organizational inertia [32].
Classical diffusion theory identifies and categorizes adopters. Early adopters are defined as venturesome innovators, active seekers of new ideas, favourable to change, willing to take risks, part of a highly interconnected social system and networks, and cosmopolitan [2,3]. The early adopters are deliberate, the late majority sceptical, and the belated adopters traditional [1]. Late adopters are often influenced most strongly by local experience and interpersonal contact. Greenhalgh and colleagues [35] advocate that diffusion is influenced more by broader organizational and environmental factors and less by an individual's adoption style [35].
The influence of the organization.
In health systems, groups of individuals, the structure of the organization they are a part of, and the broader policy context influence decision-making and the diffusion of ideas [32]. An organization's structure, function, composition, and socioeconomic context are primary influences on both what decisions are made and how they are made [2,32,36]. As an example, centralization or formalization of decision-making processes can affect adoption, affecting information flow. Organizational composition, the nature of staff, and the degree of skills and training can have a direct relationship to acceptance and change [36].
The extent to which change, new concepts, and new ideas are valued by management and leadership figures influences rates of adoption and adaptation [2,36]. Dealing with and accepting change, such as using research evidence in practice, calls for application of change theory, which proposes ideas, adoption, and implementation stages [2]. The ideas stage calls for flexibility and creativity, the adoption stage focuses on motivation, resource allocation, and negotiation, and the implementation stage is based on perceptions of legitimacy and an environment of trust [2,30,37,38].
Stage 2: Using Evidence in Policymaking
A number of studies have considered the social and political environment in which evidence is used in policymaking, offering a series of models starting with problem identification through to collaborative interpretation, solution, and application [39–41]. Staged models, whilst insightful, can suggest that policymaking is a logical, rational, and linear process. It is difficult for evidence to remain intact through the process given the policy context, decision-making factors, and the need to adapt. This indicates two things, that the evidence interacts with “context” before it is fully adopted in policy and practice, and/or that different types of evidence are useful at different times in the policy process.
The literature identifies at least three key stages of knowledge utilization: introduction, interpretation, and application [39–42]. Table 2 suggests a variety of considerations as research evidence passes through three stages of use, sourced from the work of Dobrow and colleagues [40].
Table 2 Stages of Research Utilization
Effective knowledge transfer is not a “one off” event, rather it is a powerful and continuous process in which knowledge accumulates and influences thinking over time [43]. The ability to sustain this process and a focus on human interactions is essential [43,44]. Differences in conceptual understanding, scientific uncertainty, timing, and confusion influence the response to evidence. There is no shortage of great ideas presented to policymakers in which the evidence might be either insufficient or overrepresented, often leading governments into decision-making with inadequate information [4,23].
Understanding knowledge utilization in policymaking requires an understanding of what drives policy. A variety of policy processes may be operating that influence the climate for accepting different types of evidence. As proposed by Weiss in the late 1970s [45], policy models influence where, when, and if evidence is used. A combination of the models presented in Box 2 best describes the policymaking process.
Box 2. Policymaking Models and the Use of Research Evidence
The Knowledge-Driven Model: This model suggests that emergent research about a social problem will lead to direct application to policy; it relies on effective strategies for the transfer of research evidence into practice.
The Problem-Solving Model: This model expects research to provide empirical evidence and conclusions that help solve a policy problem; it assumes that evidence is systematically gathered and applied in the policy process.
The Interactive Model: This model suggests that the search for knowledge moves beyond research to include a variety of sources such as politics and interests; it aims to reflect the complexity of the policymaking process.
The Political Model: In this model, decision-makers are not receptive to research unless it serves political gain, that is, demonstrates proof for a predetermined decision; evidence is sought to justify the problem.
The Enlightenment Model: This model suggests that cumulative research shapes concepts and perspectives that permeate the policy process over time, influencing how people think about social issues.
The Tactical Model: This model sees evidence used to support and justify government inaction, or rejection of and delay in commitment to a policy issue [23,45].
Stage 3: Capacity for Implementation
Determining capacity to act on evidence is a neglected area of policy analysis and research efforts to date. This gap exists largely because capacity is a difficult concept to define and subsequently to assess or measure [46–49]. Capacity in the health sector refers to the ability to carry out stated objectives; it is the expertise and resources at individual, organizational, and system levels for the production and application of new knowledge to health problems [50,51]. At the individual and organizational levels, capacity is often visible as skills and competencies, leadership, partnerships, the development of appropriate workforce and organizational structures, and the ability to mobilize and allocate resources [52,53]. Key at a system level are processes, policies, politics, and people (see Table 3). A case study showing how capacity is required to implement an idea informed by evidence is shown in Box 3.
Box 3. Case Study Showing How Capacity Is Required to Implement an Idea Informed by Evidence
McKee and colleagues [33] examined the diffusion of strategies for preventing sudden infant death syndrome internationally. Their analysis provides an example of a policy idea well supported by evidence, values, and cost effectiveness data, but poorly implemented. A few simple strategies have been found to reduce the risk of the syndrome, including sleeping in the supine position, avoiding exposure to tobacco smoke, breast feeding where possible, and avoiding overheating.
The authors found that although evidence about the role of sleeping position began to become available in the early 1980s, it was several years before it was acted upon, initially in the Netherlands and subsequently in New Zealand, the United Kingdom, and Scandinavia. Several countries have mounted major national preventive campaigns, but others have not. This case highlights the importance of considering implementation and systems/policy factors alongside the strength of the evidence.
Table 3 Capacities Required for Policy Adoption and Adaptation
The literature on capacity and capacity building adds value to what is already known about mechanisms for optimizing conditions for integrating research with policy and practice. Capacity theory offers something practical and operational, and calls for capacity to “adopt, adapt, and act” on the evidence in informing policy issues, otherwise policy remains idle [33,34,51]. This literature offers a more concentrated focus on individual-, organization-, and system-level factors as key to adoption, adaptation, and action in both developing and implementing evidence-informed policy. Capacity thinking both asks and answers the question of what needs to be in place to support evidence uptake in policy and practice across a variety of settings.
Why This Framework?
The purpose of this framework is to describe the myriad of changing influences in achieving evidence-informed policy and practice. The framework encourages research and planning in the area of how to “adopt, adapt, and act” on the evidence and in capacity for implementation as part of the evidence-informed policy development process. The visual presentation and descriptive mapping of the stages and features offers opportunity for deepening our understanding of the connectedness (or non-connectedness) between these factors. It also helps identify potential interventions. The framework emphasizes the policy context and its influence on each stage of interaction between research, other forms of evidence, and the policy process. Defining different types of evidence helps to both value and guide the sourcing of a broad range of information for policymaking.
Conclusion
Understanding how evidence informs policy and practice is critical in promoting effective and sustained public health action. The debate on evidence in public health has largely focussed on the linear use of research evidence in a programmatic rather than policy context. The starting point for navigating the use of evidence in policy and practice is understanding diffusion (how ideas spread throughout systems), how decisions are made, how policy is developed, and how capacity is required to effectively use evidence in this process.
The ideas behind this paper and the framework itself were jointly developed and elaborated by the authors. SB searched the literature and undertook the synthesis of materials, theories, and ideas and wrote the first draft of the paper. ABZ commented on each successive version and assisted in shaping the paper for publication. This study is being conducted with a National Health and Medical Research Council Post-Graduate Public Health Scholarship. SB's Ph.D. is being conducted under the supervision of ABZ and co-supervised by Professors Peter Sainsbury (University of Sydney, Australia) and Margaret Whitehead (University of Liverpool, England).
Citation: Bowen S, Zwi AB (2005) Pathways to “evidence-informed” policy and practice: A framework for action. PLoS Med 2(7): e166.
==== Refs
References
Rogers E Diffusion of innovations, 3rd ed 1983 New York Free Press 453
Lennarson Greer A Advances in the study of diffusion of innovation in health care organisations Millbank Mem Fund Q Health Soc (Fall) 1977 505 532
Dobbins M Ciliska D Cockerill R Barnsley J DiCenso A A framework for the dissemination and utilization of research for health-care policy and practice Online J Knowl Synth Nurs 2002 9 7 12439759
Nutley S Walter I Davies H From knowing to doing Evaluation 2003 9 125 148
Biller-Andorno N Lie RK ter Meulen R Evidence-based medicine as an instrument for rational health policy Health Care Anal 2002 10 261 275 12769414
The Cabinet Office Professional policy-making for the 21st century London: The Cabinet Office 1999 78 http://www.policyhub.gov.uk/docs/profpolicymaking.pdf
Hayward S Ciliska D DiCenso A Thomas H Underwood EJ Evaluation research in public health: Barriers to the production and dissemination of outcomes data Can J Public Health 1996 87 413 417 9009402
Nutbeam D Improving the fit between research and practice in health promotion: Overcoming structural barriers Can J Public Health 1996 87 18 23
Sackett DL Rosenberg WM Gray JA Haynes RB Richardson WS Evidence based medicine: What it is and what it isn't BMJ 1996 312 71 72 8555924
Sibbald B Roland M Getting research into practice J Eval Clin Pract 1997 2 15 21
Elliot H Popay J How are policy makers using evidence? Models of research utilisation and local NHS policy making J Epidemiol Community Health 2000 54 461 468 10818123
Bardach E A practical guide for policy analysis: The eightfold path to more effective problem solving 2000 New York Seven Bridges Press 102
Rychetnik L Hawe P Waters E Barratt A Frommer M A glossary for evidence based public health J Epidemiol Community Health 2004 58 538 545 15194712
Nutbeam D Evidence based public policy for health: matching research to policy need 1996 15 19 Evidence pp
Kitson A Ahmed LB Harvey G Seers K Thompson DR From research to practice: One organizational model for promoting research-based practice J Adv Nurs 1996 23 430 440 8655816
Black D Inequalities in health: Report of a research working group 1980 London DHSS 233
Acheson D Independent inquiry into inequalities in health report 1999 London The Stationery Office http://www.archive.official-documents.co.uk/document/doh/ih/contents.htm
Kingdon JW Agendas, alternatives and public policies, 2nd ed 1995 New York Longman 254
Whitelaw A Williams J Relating health education research to health policy Health Educ Res 1994 9 519 526 10150461
Pappaioanou M Malison M Wilkins K Otto B Goodman RA Strengthening capacity in developing countries for evidence-based public health: The data for decision-making project Soc Sci Med 2003 57 1925 1937 14499516
Johnson JL Green LW Frankish CJ Maclean DR Stachenko S A dissemination research agenda to strengthen health promotion and disease prevention Can J Public Health 1996 87 S5 S10 9002336
Milio N Making healthy public policy; developing the science by learning the art: An ecological framework for policy studies Health Promot 1987 2 263 274 10303001
Nutbeam D How does evidence influence public health policy? Tackling health inequalities in England Health Promot J Australia 2003 14 154 158
Davies H Nutley S Smith P What works? Evidence-based policy and practice in public services 2000 Bristol The Policy Press 396
Nutley S Webb J Davies H Nutley S Smith P Evidence and the policy process What works? Evidence-based policy and practice in public services 2000 Bristol The Policy Press 13 41
Haas PM Introduction: Epistemic communities and international policy coordination Int Organ 1992 46 1 21
Adler E Haas PM Conclusion: Epistemic communities, world order, and the creation of a reflective research program Int Organ 1992 46 390
Orlandi MA The diffusion and adoption of worksite health promotion innovations: An analysis of barriers Prev Med 1986 15 522 536 3774782
Rogers EM Diffusion of innovations, 4th ed 1995 New York Free Press 519
Berwick D Disseminating innovations in health care JAMA 2003 289 1969 1975 12697800
Abrhamson E Rosenkoff E Institutional and competitive bandwagons: Using mathematical modelling as a tool to explore innovation diffusion Acad Manage Rev 1993 18 487 517
Stocking B Initiative & inertia: Case studies in the NHS 1985 London Nuffield Provincial Hospitals Trust 236
McKee M Fulop N Bouvier P Hort A Brand H Preventing sudden infant deaths—The slow diffusion of an idea Health Policy 1996 37 117 135 10162643
Exworthy M Berney L Powell M ‘How great expectations in Westminster may be dashed locally’: The local implementation of national policy on health inequalities Policy Polit 2002 29 79 96
Greenhalgh T Robert G Macfarlane F Bate P Kyriakidou O Diffusion of innovations in service organisations: Systematic review and recommendations Milbank Q 2004 82 581 629 15595944
Kaluzny AD Gentry JT Veney JE Innovations of health services: A comparative study of hospitals and health departments Millbank Q 1974 52 51 58
Lomas J Diffusion, dissemination and implementation: Who should do what? Ann N Y Acad Sci 1993 703 226 235 8192299
Lomas J Sick JE Stocking B From evidence to practice in the United States, the United Kingdom and Canada Millbank Q 1993 71 405 410
Newman K Pyne T Leigh S Rounce K Cowling A Personal and organizational competencies requisite for the adoption and implementation of evidence-based healthcare Health Serv Manage Res 2000 13 97 110 11184014
Dobrow MJ Goel V Upshur REG Evidence-based health policy: Context and utilisation Soc Sci Med 2004 58 207 217 14572932
Hanney SR Gonzalez-Block MA Buxton MJ Kogan M The utilisation of health research in policy-making: Concepts, examples and methods of assessment Health Res Policy Syst 2003 1 2 Available: http://www.health-policy-systems.com/content/pdf/1478-4505-1-2.pdf . Accessed 28 April 2005 12646071
Kothari A Birch S Charles C “Interaction” and research utilisation in health policies and programs: Does it work? Health Policy 2005 71 117 125 15563998
Walt G How far does research influence policy? Eur J Public Health 1994 4 233 235
Yin RK Gwaltney MK Knowledge utilization as a networking process Knowledge: Creation, Diffusion, Utilization 1981 2 555 580
Weiss CH The many meanings of research utilization Public Adm Rev 1979 426 431
Eade D Capacity-building an approach to people-centred development 1997 Great Britain OXFAM 160
Bush R Mutch A District health development: Capacity audit 1997 Brisbane (Australia) Centre for Primary Health Care, University of Queensland
Hawe P King L Noort M Gifford S Lloyd B Working invisibly: Health workers talk about capacity-building in health promotion Health Promot Int 1998 13 285 295
Crisp B Swerrissen H Duckett SJ Four approaches to capacity building: Consequences for measurement and accountability Health Promot Int 2000 15 99 107
LaFond AK Brown L Macintyre KK Mapping capacity in the health sector: A conceptual framework Int J Health Plann Manage 2002 17 3 22 11963442
Mills A Bennett S Russell S The challenge of health sector reform: What must governments do? 2001 London Palgrave
NSW Health Department A Framework for building capacity to improve health NSW Health Department 2001 Available: http://www.health.nsw.gov.au/pubs/f/pdf/frwk_improve.pdf . Accessed 28 April 2005
Brijlal V Gilson L Makan B McIntyre D District financing in support of equity 1997 Johannesburg Center for Health Policy, Community Health, University of Witwatersrand
Coote A Allen J Woodhead D Finding out what works. Building knowledge about complex, community-based initiatives 2004 London Kings Fund 85
| 15913387 | PMC1140676 | CC BY | 2021-01-05 10:40:11 | no | PLoS Med. 2005 Jul 31; 2(7):e166 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020166 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591341510.1371/journal.pmed.0020208EditorialOtherPharmacology/Drug DiscoveryClinical PharmacologyEpidemiology/Public HealthHealth PolicyGeneral MedicineRegulationGuidelinesMedical journalsHealth PolicyPublic HealthWhy PLoS Sponsored a Roundtable of Medical Whistleblowers EditorialThe PLoS Medicine Editors 7 2005 27 5 2005 2 7 e208Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Public Domain declaration, which stipulates that, once placed in the public domain, this work may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose.On May 15, 2005, PLoS and the Government Accountability Project co-sponsored a roundtable of some of the most high profile medical whistleblowers of recent times.
==== Body
On May 15, 2005, the Public Library of Science and the Government Accountability Project, a public interest legal group that advises and supports whistleblowers (www.whistleblower.org), co-sponsored a private meeting near the Capitol Building in Washington, D. C. In the room were four of the most high profile medical whistleblowers of recent times. All four have gone public with information about practices in medicine and medical research that they believe are risking the public's health or safety [1–5]. One of them was David Graham, Associate Director in the United States Food and Drug Administration's (FDA's) Office of Drug Safety, whose research on rofecoxib (Vioxx) pointed to the serious cardiovascular risks of the drug [1]. Graham was speaking in his own capacity and was not representing the FDA. An anonymous fifth whistleblower—a research scientist at a major drug company—participated by phone.
The whistleblowers took turns to share their stories, including their accounts of retaliations they said they had faced from their employers on raising their concerns, which led to lawsuits by at least two of the whistleblowers [2,3]. The picture that emerged from these accounts—a picture of American medicine's inappropriate ties with the pharmaceutical industry—was deeply troubling.
As the investigative medical journalist Jeanne Lenzer reports in her Essay for PLoS Medicine [6], the whistleblowers spoke of public regulatory agencies that are putting the interests of drug companies ahead of the safety of patients, and of pharmaceutical companies that allow their marketing departments to knowingly downplay serious side effects when promoting their drugs. And they spoke of the woefully inadequate protection offered to those in the medical community who feel morally compelled to blow the whistle.
It was Lenzer who conceived the idea for the meeting. She believed that important lessons would emerge from having these medical whistleblowers, who come from very different professional backgrounds, together in one room to share their experiences. It took her many months of planning. In particular, she needed to gain the trust of the industry research scientist, so that the scientist could feel sure that anonymity would be preserved. But all of her planning nearly came to nothing. At the last moment the original journal sponsor pulled out on the advice of its lawyers.
The picture that emerged from these accounts was deeply troubling.
Lenzer's phone call to PLoS, enquiring whether we might step in, came just ten days before the event was due to happen. We took our own legal advice and then agreed to sponsor the roundtable. PLoS was eager to support this event, and willing to accept any small legal risks, because we believe that the issues raised will be of huge interest to the medical community, to the press, to patients, and to the broader public. Further, the event fits well with our own public service mission of making all scientific and medical research results freely and publicly available, and with our belief that transparency in the conduct and publication of research is important for public trust. And the topic of the roundtable was in line with other articles we have published highlighting the many ways in which medicine has become tightly entangled with industry, to the great detriment of patients [7–9].
The risks to a journal in sponsoring such an event are, of course, much smaller than the risks that the whistleblowers at the roundtable faced in going public with their stories. Studies have shown that whistleblowers in both public service and private industry almost always experience retaliation from their employers, with those employed longer experiencing greater retaliation [10,11]. They risk loss of earnings, intimidation, harassment, victimization, and personal abuse, and they traditionally receive little help from statutory authorities [12,13].
The Washington whistleblowers' stories illustrate these issues. Psychiatrist Stefan Kruszewski described how he was fired from his job at the Pennsylvania Department of Public Welfare (DPW) after alerting his seniors to prescribing practices across the state that he considered to be alarming and dangerous [2]. “I was fired in a demeaning manner,” said Kruszewski, who has sued DPW over his firing. “My two offices were emptied and the contents of these offices were put in the gutter.” David Graham—who testified at a US Senate Finance Committee hearing on rofexocib (Vioxx), the FDA, and Merck [1]—said that there was a conspiracy by senior management at the FDA “to intimidate me ahead of the Senate testimony.” Both of these individuals contend that pharmaceutical industry influence over their employers (a state and a federal regulatory agency, respectively) played a part in the difficulties these individuals faced in getting their concerns heard [1,2].
Lenzer's report will, we hope, spark discussion and debate about how American medicine—clinicians, researchers, regulatory agencies, and medical journals—can disentangle itself from the influence of the pharmaceutical industry. In the past, medical journals and their editors have played an important part in exposing the complex relationships between the pharmaceutical industry and medicine [14–16], including between industry and the medical journals themselves [9]. PLoS Medicine will continue to look critically at these relationships. A common theme at the roundtable was that, armed with information, the public too could have an important role in unpicking these ties. “The pharma–FDA complex has to be dismantled,” said Graham, “and the American people have to insist on that, otherwise we're going to have disasters like Vioxx that happen in the future.” Patients, health professionals, and even the industry itself all surely stand to gain from disentanglement.
We would like to thank the Government Accountability Project for co-sponsoring the roundtable.
==== Refs
References
Lenzer J FDA is incapable of protecting US “against another Vioxx.” BMJ 2004 329 1253 15564236
Lenzer J Whistleblower charges medical oversight bureau with corruption BMJ 2004 329 69
Lenzer J Whistleblower removed from job for talking to the press BMJ 2004 328 1153
Lenzer J Bush plans to screen whole US population for mental illness BMJ 2004 328 1458 15205287
Lenzer J Confessions of a drug rep BMJ 2005 330 911
Lenzer J What can we learn from medical whistleblowers? PLoS Med 2005 2 e209 10.1371/journal.pmed.0020209 15913416
Barbour V Butcher J Cohen B Yamey G Prescription for a healthy journal PLoS Med 2004 1 e22 10.1371/journal.pmed.0010022 17523248
Caulfield T The commercialisation of medical and scientific reporting PLoS Med 2004 1 e38 10.1371/journal.pmed.0010038 15630461
Smith R Medical journals are an extension of the marketing arm of pharmaceutical companies PLoS Med 2005 2 e138 10.1371/journal.pmed.0020138 15916457
Soeken K Soeken D A survey of whistleblowers: Their stressors and coping strategies 1987 Laurel (Maryland) Association of Mental Health Specialities
Yamey G Protecting whistleblowers BMJ 2000 320 70 71 10625244
Lennane KJ “Whistleblowing”: A health issue BMJ 1993 307 667 670 8401056
Devine T The whistleblower's survival guide: Courage without martyrdom 1997 Washington (DC) Fund for Constitutional Government 174
Horton R Lotronex and the FDA: A fatal erosion of integrity Lancet 2001 357 1544 1545 11377636
Abbasi K Smith R No more free lunches BMJ 2003 326 1155 1156 12775587
Angell M The truth about the drug companies: How they deceive us and what to do about it 2004 Random House, 2004 336
| 15913415 | PMC1140677 | CC0 | 2021-01-05 10:40:20 | no | PLoS Med. 2005 Jul 27; 2(7):e208 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020208 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591341610.1371/journal.pmed.0020209EssayOtherClinical PharmacologyEpidemiology/Public HealthHealth PolicyGeneral MedicineRegulationGuidelinesMedical errors/patient safetyMedical journalsPublic HealthWhat Can We Learn from Medical Whistleblowers? EssayLenzer Jeanne Jeanne Lenzer is a freelance medical investigative journalist based in Kingston, New York, United States of America. E-mail: [email protected]
Competing Interests: The author declares that she has no competing interests.
7 2005 27 5 2005 2 7 e209Copyright: © 2005 Jeanne Lenzer.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.High-profile medical whistleblowers tell their stories of how American medicine's close ties with the pharmaceutical industry may be putting patients' safety at risk.
Their experiences paint a troubling picture of American medicine's ties with the pharmaceutical industry
==== Body
A year ago, I received an E-mail from a research scientist at a major pharmaceutical company. The scientist had read my articles on whistleblowers who had raised concerns about the undue influence of the pharmaceutical industry on American medicine My industry source had information for me about drug company practices, but—out of fear of career ruin—would only talk on the condition that I would conceal the scientist's identity.
For the next year or so, I had repeated contacts with the scientist. As I listened to this researcher—and to the other medical whistleblowers that I continued to interview—it occurred to me that each whistleblower was like the proverbial blind man with a hand on the elephant. Each could describe one piece of the puzzle, but the full picture could only emerge by bringing these whistleblowers together.
With an eye to focusing on the systemic problems that have allowed American medicine to be unduly influenced by industry, on May 15, 2005, I brought together five whistleblowers in Washington, D. C. I asked them each to tell their story and to suggest ways to restore objectivity to medicine and medical research.
The Whistleblowers
Four whistleblowers attended in person, and the anonymous industry scientist participated via speakerphone. The whistleblowers came from an extraordinary variety of different professional backgrounds.
David Graham
This Food and Drug Administration (FDA) safety officer raised concerns about the cardiovascular side effects of rofecoxib (Vioxx) and other Cox-2 inhibitors. He testified at a United States Senate Finance Committee hearing on rofexocib, the FDA, and Merck [1,2]. Graham attended the roundtable in his own personal capacity and was not representing the FDA.
Each whistleblower was like the proverbial blind man with a hand on the elephant.
Allen Jones
This investigator at the Pennsylvania Office of the Inspector General led an investigation into an off-the-books account, funded in part by drug companies, from which payments were made to state employees to develop a medication treatment algorithm. He filed a civil rights lawsuit against the Pennsylvania Office of the Inspector General to protect his right to publicly discuss his findings, and was later fired from his job for talking to the press [3–6].
Stefan Kruszewski
This Harvard-trained psychiatrist was hired by the Bureau of Program Integrity in the Pennsylvania Department of Public Welfare to oversee the state's mental health and substance misuse programs. He filed a law suit in a federal court in the Middle District of Pennsylvania, charging that he was fired after uncovering widespread abuse and fraud in the bureau [7,8].
Kathleen Slattery-Moschkau
This former drug representative left the pharmaceutical industry after witnessing marketing practices that she found disturbing. She wrote and directed the movie Side Effects, a fictionalized account of her experiences [9,10].
The anonymous research scientist
This is an industry insider who said to me, ahead of the roundtable, that the culture of secrecy at drug companies too often results in claims that are closer to “propaganda” than science.
Lessons Learned from the Roundtable
Ties between drug regulators and industry may influence new drug approval
David Graham described the frustrations he had felt in his almost 20 years of experience as an FDA drug safety officer. Although he was instrumental, he said, in getting ten drugs off the market because of safety concerns, his experience was like a salmon swimming upstream—“a single individual…against the tide.” The tide, he said, “is an entire institution whose mission is to approve drugs and make industry happy.”
The FDA, said Graham, is in a “collaborative relationship” with industry. The FDA gets money from drug companies through the Prescription Drug User Fee Act of 1992 (see http://www.fda.gov/cber/pdufa.htm) “to approve new drugs and approve them more quickly.” The mindset at the FDA, he said, is that “we will find a reason to approve a drug no matter how small the indication for the drug.” Graham explained that a senior official at the FDA had told him: “industry is our client.“
When the FDA knows there is a serious problem with a new drug, he said, the FDA deals with this by saying, “well, we'll handle it in labeling” even though, said Graham, “FDA knows labeling doesn't work.”
“There is no independent voice for drug safety in the United States,” he said. The upper-level managers in the FDA's Office of Drug Safety are appointed from the FDA's Office of New Drugs, which approves new medicines. This makes the Office of Drug Safety “captive,” he said, to the Office of New Drugs.
The anonymous scientist said that in order to speed up drug approval, companies “don't measure things like whether we are really curing the disease, or prolonging life, or preventing hospitalization, or whether a patient is truly more functional. Oftentimes, we're measuring intermediate, lesser things, markers, predictors—we hope—of these clinical endpoints, but they may or may not be accurate.”
And the FDA, said the scientist, requires just two positive studies to grant approval to a new drug, but there is no limitation on how many negative studies can be done before one or two positive studies are produced. This can lead to approval of a drug even when most studies are negative or show no effect.
Both Graham and the anonymous scientist suggested putting an end to the Prescription Drug User Fee Act, and Graham argued that there needs to be independent authority for those in charge of drug safety. They indicated that two bills in Congress, introduced by Senator Grassley and by Congressman Hinchey, at least partly address these concerns.
“The pharma–FDA complex has to be dismantled,” said Graham, “and the American people have to insist on that, otherwise we're going to have disasters like Vioxx that happen in the future.”
The race to approve new drugs without proper safety testing may be compromising the public's health
“Drug companies assiduously avoid acquiring information about side effects,” said the industry scientist. “Drug companies will not conduct safety studies unless they have to—meaning basically that they're required by a regulator—and that rarely happens.” High-risk patients who might have a bad reaction to a drug, said the scientist, “are excluded from studies deliberately, even though, when the drug is approved, these patients will be targeted for sales.” When a safety study is proposed within the industry, said the scientist, “a typical response will be that if we conducted a study to find out if there was a safety problem, people would learn about it and think we had a problem [which] would destroy the image of safety that has been so carefully constructed.”
“There is no independent voice for drug safety in the United States.”
Studies are too small and are conducted over too brief a period to properly assess safety: “The largest studies—the phase three studies, [which] might be several thousand people—last for a few months. If drugs kill one in several thousand per year, this would be a public health catastrophe. A blockbuster drug with that kind of hazard associated with it could be associated with tens of thousands of deaths a year, and it would never be detected in studies of the kind that we routinely submit and are the basis for approval.” These drugs, said the scientist, and these kinds of risks, are “essentially out there now, unlabeled, unnoticed, all beneath the radar.”
The scientist said that, “to ensure that safety problems will go unnoticed, we compound the problem of conducting small studies by setting a statistical threshold for acknowledging the safety problem that is so high that you know in advance it could never be reached for any serious side effect, like myocardial infarction.” This practice, said the scientist, “virtually ensures that if a bad side effect happens to show up, it's not going to reach the arbitrary level that we call statistically significant, and the company can maintain that it's just bad luck.” And if a bad result does happen, “typically a company is not going to publish the study at all. If they do publish it, the bad result can be omitted as ‘not statistically important.’”
The funding of state officials by industry may be affecting prescribing patterns
Allen Jones described how he believed that drug companies were acting at the state level to influence the prescribing of psychiatric medications.
“I began to investigate an account into which pharmaceutical companies were paying money that was being accessed by state employees,” he said. “Additionally, I found that various pharmaceutical companies were paying state employees directly—also giving them trips, perks, lavish meals, transportation, honorariums up to $2,000 for speaking in their official capacities at drug company events. They were given unrestricted educational grants that were deposited into an off-the-books account—unregistered, unmonitored, literally operated out of a drawer.”
These same state officials, he said, were responsible for dictating clinical policy and writing guidelines for the treatment of patients in the state system. These officials were, he said, receiving money from companies with a stake in the guidelines. “The protocol they [the officials] were developing was called the Texas Medication Algorithm Project, TMAP, which began in Texas in the mid-90s. It outlined detailed medication guidelines for schizophrenia, depression, and bipolar disorder. It recommends almost exclusive usage of newer, patented, very expensive atypical antipsychotics, SSRIs [selective serotonin uptake inhibitors], and mood stabilizers.” The Texas Medication Algorithm Project, said Jones, was based on “expert consensus” from industry-supported meetings.
Jones said that when he wanted to investigate these findings, he was shut down. “I was told point black, ‘Look, drug companies write checks to politicians, they write checks to politicians on both sides of the aisle—back off.’” He was told, he said, to “quit being a salmon, quit swimming against a stream.” He wouldn't back down from his investigation, he said, and was demoted. On November 22, 2002, he filed a civil rights lawsuit “to preserve my job and my right to speak out.” His employer, he said, took him off investigative duties altogether.
Stefan Kruszewski, who has filed a law suit in a federal court in Pennsylvania, raised concerns to his seniors in the Pennsylvania Department of Public Welfare about prescribing practices in the state that he did not feel were evidence based, and said he lost his job for raising his concerns. For example, he alerted his seniors to the off-label prescribing of the anticonvulsant gabapentin (Neurontin) for mood disorders and addictive disorders.
“The pharmaceutical industry is the single most powerful lobbying group on Capitol Hill—outspending even the oil and banking industries,” said Jones. “It should come as no surprise that the ties go far beyond just the mental health officials who wrote the guidelines, but extend to many of the politicians who, in the end, allowed an investigation into pharma corruption to be dropped, and the investigator—me—to be fired.”
Efforts to detect and deter fraud and abuse due to these conflicts, he said, “will be likely to be undermined as long as those charged with detecting fraud and abuse, like the [Pennsylvania] Inspector General, are appointed by politicians who are themselves beholden to the drug industry. Such positions should instead be filled by career civil servants and not political appointees.”
Regulatory agencies are not being held accountable
In comments that echoed his testimony to the US Senate Finance Committee, Graham said that, “FDA was the single greatest obstacle to doing anything effective with Vioxx. As a result, nearly 60,000 people probably died from that drug. That's as many of our soldiers that were killed in the Vietnam war [who] died as a result of Vioxx use. And FDA had the opportunity, the responsibility, to stop that and didn't. In fact, FDA allowed it to continue. In my book, FDA shares in the responsibility for those deaths and yet it's not being held accountable by Congress.” Congress itself, added Graham, is deeply beholden to the drug industry since many politicians receive “often quite a bit of campaign contributions” from the industry.
Kruszewski reflected upon the problems he said he had encountered in Pennsylvania, saying that “there is no accountability in the system for oversight [agencies].” He has become “a stronger advocate than ever for a federal patient bill of rights.”
Marketing departments can influence doctors’ prescribing habits
The research scientist said that the job was attractive because of the “many excellent drugs” developed, such as drugs to treat HIV, but the scientist “also saw drugs marketed in a way that will exaggerate the benefits and conceal the risks.”
Kathleen Slattery-Moschkau gave an insider's view of drug marketing practices, from her former experiences as a drug rep. She clutched her head in disbelief as she told the roundtable that doctors would come up to her with patients' charts asking her for advice on treating patients. Slattery-Moschkau, like most of the drug representatives she came to know over the years, had no science background at all.
“Drug companies assiduously avoid acquiring information about side effects.”
The various techniques drug representatives were trained in to “educate doctors” eventually proved to be not just “comical” but “also scary,” she said. “Whether it was hiring, training, what we were told to say about drugs and what we were told not to say,” it was marketing, not science, that dominated. One of the techniques used by drug companies was to buy doctors' prescribing records so drug representatives knew “to the dime” what drugs doctors were prescribing and could tailor their marketing to them. Drug representatives developed “personality profiles” on doctors and were taught to pitch their sales to specific personality types. Representatives were compensated, she said, by “how many prescriptions we could encourage.”
Both Slattery-Moschkau and the industry scientist described tensions within drug companies between marketing departments and industry scientists. “The marketing spin on things,” said the scientist, “carries the day.”
The published medical literature contains many biases
“When studies are published,” said the scientist, “they are frequently written not by the trained research scientist, who might have designed and analyzed the study, but by a designated medical writer with little if any background in research, but who is trained instead to craft the findings of the study in the best possible way for the company.”
The body of literature available to the public, said the scientist, “is a biased sample of what companies want people to see.” The research scientist described “a culture of secrecy,” which makes it hard even for industry scientists tasked with ensuring drug safety to obtain the full datasets needed to genuinely understand a drug's risk–benefit profile.
Conclusion
Whistleblowers have been compared to bees—they have just one sting to use and using it may lead to career suicide [11]. Many of the whistleblowers at the roundtable said they had experienced retaliation from their employers for raising concerns, but all had felt obligated to speak out about practices in medicine and medical research that they believe are risking the public's health or safety. Graham said he felt “trapped by the truth” and had to act. “There are bigger issues here,” said Kruszewski. “I felt right from the start [that] if I wallowed in self-pity about being fired and having my belongings piled in the gutter that I would never understand why all these things were happening. The bigger issue is that we've got people in the pharmaceutical industry and the health-care industry all acting in synchrony.”
Each of these whistleblowers, in very different ways—from making a satiric film to speaking out in Congress—has shone light on how this “synchrony” may be compromising the integrity of American medicine. We should not have to rely on medical whistleblowers to alert us to these fault lines. If we are to restore objectivity to drug development, prescribing, and safety monitoring, we must be willing to examine and change all of the institutions that allow this synchrony to occur.
The roundtable was co-sponsored by the Public Library of Science and the Government Accountability Project (www.whistleblower.org), a public interest group that helps whistleblowers in order to promote governmental and corporate accountability.
Citation: Lenzer J (2005) What can we learn from medical whistleblowers? PLoS Med 2(7): e209.
Abbreviation
FDAFood and Drug Administration
==== Refs
References
Kaufman M FDA official alleges pressure to suppress Vioxx findings. Washington Post 2004 October 8 Available: http://www.washingtonpost.com/wp-dyn/articles/A16546-2004Oct7.html . Accessed 18 May 2005
Lenzer J FDA is incapable of protecting US “against another Vioxx” BMJ 2004 329 1253 15564236
Lenzer J Whistleblower removed from job for talking to the press BMJ 2004 328 1153
Lenzer J Bush plans to screen whole US population for mental illness BMJ 2004 328 1458 15205287
Moynihan R Drug company targets US state health officials BMJ 2004 328 306
Petersen M Making drugs, shaping the rules. New York Times 2004 February 1 Available: http://psychrights.org/Articles/NYTMakingDrugsShapingtheRules.htm . Accessed 18 May 2005
Lenzer J Whistleblower charges medical oversight bureau with corruption BMJ 2004 329 69
Weisenseegan N Lawsuit: State fired shrink for exposing abuse. Philadelphia Daily News 2004 July 7 Available: http://www.philly.com/mld/dailynews/news/special_packages/phillycom_front_dn/10352527.htm?template=contentModules/printstory.jsp . Accessed 18 May 2005
Lenzer J Confessions of a drug rep BMJ 2005 330 911
Associated Press Film exposes pharmaceutical secrets Associated Press 2005 March 10 Available: http://www.forbes.com/associatedpress/feeds/ap/2005/03/11/ap1878239.html . Accessed 18 May 2005
Vinten G Whistle while you work in the health-related professions? J R Soc Health 1994 114 256 262 7844789
| 15913416 | PMC1140678 | CC BY | 2021-01-05 10:40:23 | no | PLoS Med. 2005 Jul 27; 2(7):e209 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020209 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1591342010.1371/journal.pbio.0030203Research ArticleEcologyEvolutionGenetics/Genomics/Gene TherapyInfectious DiseasesNoneHost–Parasite Interactions and the Evolution of Gene Expression Evolution of Gene ExpressionNuismer Scott L [email protected]
1
Otto Sarah P
2
1 Department of Biological Sciences, University of Idaho, Moscow, Idaho, United States of America,2 Department of Zoology, University of British Columbia, Vancouver, British Columbia, CanadaKingsolver Joel Academic EditorUniversity of North CarolinaUnited States of America7 2005 31 5 2005 31 5 2005 3 7 e20313 12 2004 5 4 2005 Copyright: © 2005 Nuismer and Otto.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.
Host-Parasite Battles Shed Light on the Evolution of Gene Expression
Interactions between hosts and parasites provide an ongoing source of selection that promotes the evolution of a variety of features in the interacting species. Here, we use a genetically explicit mathematical model to explore how patterns of gene expression evolve at genetic loci responsible for host resistance and parasite infection. Our results reveal the striking yet intuitive conclusion that gene expression should evolve along very different trajectories in the two interacting species. Specifically, host resistance loci should frequently evolve to co-express alleles, whereas parasite infection loci should evolve to express only a single allele. This result arises because hosts that co-express resistance alleles are able to recognize and clear a greater diversity of parasite genotypes. By the same token, parasites that co-express antigen or elicitor alleles are more likely to be recognized and cleared by the host, and this favours the expression of only a single allele. Our model provides testable predictions that can help interpret accumulating data on expression levels for genes relevant to host−parasite interactions.
A model reveals that hosts should evolve co-expression of resistance alleles to recognize a range of parasites, but the parasite shouldn't evolve co-expression of infection alleles because it enhances recognition by the host.
==== Body
Introduction
Hosts and parasites are locked in a continual co-evolutionary race, which generates
persistent selection for resistant hosts and infectious parasites. Understanding the direct
effects of this process on spatial patterns of local adaptation [1–5
], the evolution of virulence/pathogenicity [6–8
], and the spread of infectious disease [9–11
] has been a central focus of research into host−parasite interactions. Yet
host−parasite interactions also generate indirect selection on a variety of other
features of the interacting species. The classical example of indirect selection imposed by
host−parasite interactions is on the mode of reproduction [12–16]. Host−parasite interactions can select for sexual rather than asexual
reproduction, although they tend to do so only when selection is strong and sex is rare
[17]. Recently, we have shown
that indirect selection also acts on genome size (ploidy level), with selection favouring
diploidy more often among host species and haploidy more often among parasite species [
18]. There are a variety of other
genomic features besides ploidy level that should experience indirect selection in response
to host−parasite interactions. Here, we examine the evolution of expression levels
using a model that is structurally similar to models of the evolution of dominance (as in
the classic papers by Fisher [19,
20], Wright [21–
23], and Haldane [24,
25], and more recent papers
reviewed in Otto and Bourguet [26 ]).
The Model
To explore the evolution of expression levels, we assumed that infection/resistance was determined by a single gene in the host with alleles A and a, and a single gene in the parasite with alleles B and b. We then tracked changes in allele frequency at a single modifier locus, whose alleles (M and m) altered the pattern of expression in heterozygotes at the A locus (if in hosts) or B locus (if in parasites). Thus, we refer to this modifier locus, M, as a regulatory locus. To simplify the analysis and interpretation, we allowed expression levels to evolve in only one species (the “focal species”) at a time.
Determining how expression patterns evolve during the course of host−parasite co-evolution requires that we relate expression patterns to the phenotype expressed by heterozygous genotypes. We assumed that a heterozygous individual of species j could express the phenotype of homozygotes carrying allele A (or B) with probability ρ
1,j
,
A and a (or B and b) with probability ρ
2,j
, and a (or b) with probability ρ
3,j
, where the terms in parentheses are appropriate when the focal species is the parasite. These probabilities were assumed to sum to one (ρ
1,j + ρ
2,j + ρ
3,j = 1), for both hosts (j = h) and parasites (j = p). This constraint prevents heterozygotes from having fitness greater than the best homozygous genotype in any given encounter between host and parasite genotypes. An implicit assumption of this mapping between genotype and phenotype is that heterozygotes can, if ρ
2 = 1, co-express both alleles without decreasing the function of either allele. To take a concrete example, our mapping of phenotype onto genotype assumes that Aa hosts could express receptor A as effectively as AA hosts and also express receptor a as effectively as aa hosts. The model is easily generalized, however, to relax this assumption (results available upon request). Alleles at the regulatory locus, M, were allowed to alter the pattern of expression in heterozygotes by altering the probabilities, ρi
,j. Because an individual's genotype at the regulatory locus determines these probabilities, we specify the genotype in square brackets (e.g., ρi,j[MM]). When exposed to selection induced by the interacting species, alleles at the regulatory locus might evolve to upregulate one allele over the other or to express both alleles equally (co-expression), as illustrated in Figure 1.
Figure 1 Expression Levels Are Allowed to Evolve toward Any Point in the Triangle
For example, the circle corresponds to the additive case, where heterozygotes are equally likely to express either A only or a only and so have fitness halfway between the fitnesses of AA and aa individuals. The evolution of expression levels predicted by the quasi-linkage equilibrium analysis is indicated by the direction of arrows. Double-headed arrows indicate that the quasi-linkage equilibrium analysis predicts an outcome that depends on allele frequencies. Results from numerical simulations are shown as percentages of total parameter combinations that resulted in evolution of expression levels in the direction shown. Entries labelled “neutral” are cases where no change in modifier frequency occured. The range of parameter values used in these simulations is described in the main text. Predicted patterns for the host are shown in (A), and those for the parasite are shown in (B).
We incorporated host−parasite co-evolution into the modifier framework described above by considering the following well-studied genetic interactions. In the gene-for-gene (GFG) model [27], avirulence alleles in the parasite produce signal molecules that elicit a defence response in resistant hosts, whereas parasites carrying virulence alleles fail to produce the signal molecule and cannot be detected by any host. GFG interactions are considered to be prevalent in plant–pathogen interactions [28]. Costs of resistance and virulence alleles have been demonstrated in some GFG systems [29,30], so we let Ch be the fitness cost of expressing only the resistant allele in hosts, and Cp be the fitness cost of expressing only the virulence allele in parasites. Co-expressing both alleles might reduce these costs, particularly when the susceptible allele in the host or the avirulent allele in the parasite performs a beneficial function. The fitness costs experienced by heterozygotes expressing both alleles were thus set to ch in hosts and cp in parasites. The matching-alleles (MA) model is predicated upon a system of self/non-self recognition. Hosts can successfully defend against attack by a parasite whose genotype does not match their own. Such recognition systems have been observed in invertebrates [31] and vertebrates [32]. Finally, in the inverse-matching-alleles (IMA) model, host defence involves an array of recognition molecules (e.g., antibodies) that are able to recognize specific antigens and resist attack by parasites carrying these antigens [32]. Following the rules imposed by each of these modes of co-evolution allowed us to create a matrix that describes the outcome of an interaction between any two phenotypes (Table 1). In all cases, we assumed that infection results in a loss of host fitness but an increase in parasite fitness.
Table 1 An Interaction between a Host and a Parasite Results in Either Infection or Resistance, Depending on the Phenotype of the Interacting Species
We assumed a life cycle where selection due to interactions between host and parasite was followed by sexual reproduction. Species interactions are assumed to depend on loci: a regulatory locus with alleles M and m and an interaction locus with alleles A and a if the focal species is the host, or B and b if the focal species is the parasite. Thus, there are four chromosome types in each species: MA (MB), Ma (Mb), mA (mB), and ma (mb), where the terms in parentheses correspond to cases where the focal species is the parasite. We track evolution at the regulatory locus in only one species at a time and assume that the regulatory locus is fixed on M in the non-focal species. The non-focal species is assumed to be diploid, although results derived with a haploid non-focal species were similar. Species j is assumed to undergo sexual reproduction with random mating with probability sexj and to reproduce asexually with probability (1 − sexj). During sexual reproduction, the two loci are allowed to recombine at rate rj.
Genotype frequencies after one round of selection can be determined using standard population genetic equations once the fitnesses of genotypes have been determined. We assume that encounters between species occur at random and that at most one interaction occurs per generation per individual. When interacting with genotype k in species j¯
, the fitness of genotype i in species j is denoted by W
i.j↔k.j¯
, where j = h and j¯
= p when the focal species is a host, and j = p and j¯
= h when the focal species is a parasite. The average fitness of genotype i in species j is given by its fitness in the presence of genotype k in the interacting species, weighted by the frequency of genotype k, summed over all k:
Thus, we assume that fitness depends upon genotype frequencies but is independent of the population sizes of the interacting species (e.g., [33,34]). The mean fitness in species j is calculated as the weighted sum of equation 1 over all genotypes in species j:
W
i.j↔k.j¯
can be calculated using Table 1 and the probabilities, ρi,j, that heterozygous hosts (parasites) express a particular phenotype. The encounter rate between hosts and parasites is implicitly incorporated in W
i.j↔k.j¯
; when hosts and parasites rarely encounter one another, the fitnesses will be more similar to one another, all else being equal.
Assuming an infinite population size and ignoring mutation, we can write down recursions for the frequency, Xi,j, of each diploid genotype (e.g., i = MA/Ma or MB/Mb) in species j after one round of selection followed by reproduction. For example, the first four recursions for host genotypes are given by
where primes indicate post-selection genotype (X′i,j)
and gamete (p′i,j)
frequencies. Specifically, the frequency of genotype i after selection is given by X′i,j = Xi,j
Wi,j/W¯j,
and the gametes produced by the surviving hosts are in the following frequencies:
Recursions for the parasite species are identical, with the exceptions of the subscripts A, a, and h, which are replaced by B, b, and p, respectively.
Results
To analyze the model, we assumed that selection was weak relative to the rate of recombination between the modifier locus and the locus determining infection/resistance. This allowed us to derive very general conditions for the evolution of expression levels in the focal species using quasi-linkage equilibrium approximations [35,36]. In short, the frequency of sex and recombination are assumed to be high enough relative to the strength of selection that the disequilibrium between the regulatory and interaction locus (Dh = freq(MA)hfreq(ma)h − freq(Ma)hfreq(mA)h
in hosts) reaches a steady-state value that depends on the current allele frequencies in the host and parasite. Solving for this disequilibrium then allows us to calculate the rate of allele frequency change at the regulatory locus to leading order in the selection coefficients (Protocol S1).
When the host was the focal species, the frequency of allele M at the regulatory locus changed at a per-generation rate of:
where pi
,j is the frequency of allele i in species j and where s
model,h
depends on the model of host−parasite interactions and, for the GFG, MA, and IMA models, is given by
In equations 12–14, γi, ξi, and αi measure the strength of selection acting on species i due to GFG interactions, MA interactions, and IMA interactions, respectively (see Table 1), and Δρi
,j represents the average effect of allele M on the probability of expression pattern i in species j:
Assuming weak selection, equation 11 is equivalent to the allele frequency change in a standard one-locus model with a selection coefficient given by the selection term s
model,h multiplied by the frequency of Aa hosts (2pA,hpa,h). As discussed in greater detail in Protocol S1, the selection coefficient s
model,h can be readily interpreted on the basis of how changes in expression pattern alter the likelihood that a heterozygous host will be infected.
Similarly, when the parasite was the focal species, the frequency of allele M at the regulatory locus changed at a per-generation rate of
where now
To the order of these approximations, genetic associations (Dh, Dp) had no influence on the frequency of the alleles at the regulatory locus, M. Instead, frequency change at the regulatory locus resulted from the direct effect of altered expression levels on fitness. Indeed, to leading order in the selection coefficients, equations 12–14 in hosts and 17–19 in parasites describe the change in fitness expected if a randomly chosen m allele were replaced by an M allele within an Aa heterozygote.
Examining the signs of equations 12–14 and 17–19 allows us to predict the directions in which expression levels should evolve in heterozygotes. These results are summarized in Figure 1A for the host and Figure 1B for the parasite. As is clear from Figure 1, selection typically favours the evolution of co-expression among hosts but rarely favours co-expression among parasites. These results are conceptually similar to recent findings on the evolution of ploidy levels [18]. In order to recognize and clear a wide array of parasites, selection favours hosts with a broader arsenal of recognition molecules, thus favouring diploid life cycles and the co-expression of alleles in heterozygotes. In contrast, in order to evade a host's immune system or defence response, selection favours parasitic individuals that express a narrow array of antigens and elicitors, thus favouring haploid life cycles or expression of only one allele in heterozygotes. Exceptions to these general rules arise when selection acts in ways other than recognition and evasion. In the MA model, hosts are more likely to survive if they are difficult to mimic, which selects for a narrow expression pattern of only one allele. Furthermore, when costs are added to the GFG model, there are periods of time when selection favours expression of only the least costly allele (i.e., expression of the susceptible allele in hosts when virulence is common among parasites [see equation 12] or the expression of the avirulent allele in parasites when resistance is rare among hosts [see equation 17).
To evaluate whether our analytical results are robust to violations of the assumption that recombination is frequent and selection is weak, we numerically iterated the exact recursions. For each genetic model of co-evolution, we considered both focal hosts and focal parasites, and modifiers that altered the expression probabilities ρ
1, ρ
2, and ρ
3 (Protocol S1). In each case, we considered all combinations of the following selection intensities (0.005, 0.05, and 0.50) and recombination rates (0.005, 0.05, and 0.50) and ran five simulations with randomly chosen initial allele frequencies. In the GFG model, we considered six levels of the costs of expressing only the resistance allele (Ch) or only the virulence allele (Cp): 15%, 30%, 45%, 60%, 75%, or 90% of the value of the fitness cost of infection in hosts, γh, and the fitness cost of resistance in parasites, γp, respectively. In addition, the costs of co-expression (ch or cp) were set to 33%, 66%, or 100% of the full costs of resistance or virulence (Ch or Cp). In all simulations, the modifier was introduced at an initial frequency of 0.5 after a 1,000 generation burn-in period had elapsed. All simulations were then run for an additional 4,000 generations, and the modifier was considered to have changed in frequency if its final frequency differed from its initial frequency by an amount greater than 10−13. This minimum threshold was set to eliminate false positives due to numerical imprecision and was based upon the maximum change in frequency observed for a modifier with no effect. The simulation results always coincided with the analytical predictions (Figure 1).
Taken together, our analytical and simulation results suggest that heterozygous hosts should generally evolve to co-express resistance alleles but heterozygous parasites should evolve to express only a single infection allele (Figure 1). It is not clear from the analytical results, however, which allele (B orb), will ultimately be expressed in heterozygous parasites. Specifically, our analytical results suggest that expression of the B allele is favoured at some host allele frequencies, whereas expression of the b allele is favoured at others (see equations 17–19). Thus, the potential exists for patterns of parasite gene expression in heterozygotes to cycle over evolutionary time. Results from numerical simulations demonstrate that this is indeed the case. Cycles in parasite gene expression, where allele B was expressed during some periods of time and allele b at others, were frequently observed in IMA interactions and occasionally in GFG interactions with a cost of resistance (Figure 2). In contrast, cyclical patterns are less likely to persist in host species over long periods of evolutionary time because modifiers that increase co-expression generally spread to fixation (see Figure 1A). Only in the MA model do we expect long-term cycles in levels of dominance to potentially occur in both host and parasite.
Figure 2 Co-Evolutionary Dynamics of Parasite Gene Expression
In both panels, the frequency of a modifier allele that increases the expression of the B allele in the parasite population is shown in orange. The frequency of the B allele in the parasite population is shown in blue, and the frequency of the A allele in the host population is shown in maroon. Both panels considered an expression modifier introduced at an initial frequency of 0.5 with the following effects: ρ
1[MM] = 0.75, ρ
1[Mm] = 0.50, ρ
1[mm] = 0.25, ρ
3[MM] = 0.25, ρ
3[Mm] = 0.5, ρ
3[mm] = 0.75, and ρ
2[i] = 1 − ρ
1[i] − ρ
3[i]. Parameters for the IMA model (A) were αh = 0.15, αp = 0.20, rh = 0.25, and rp = 0.25. Parameters for the GFG model (B) were γh = 0.15, γp = 0.20, Ch = 0.075, Cp = 0.10, ch = 0.0075, cp = 0.01, rh = 0.25, and rp = 0.25. In both panels, the initial frequency of the A allele and the B allele was 0.55.
Discussion
Our results demonstrate that co-evolution between hosts and parasites favours co-expression of alleles more often in hosts than in parasites. This predicted pattern is particularly striking among the models with the greatest empirical support (GFG and IMA) and helps explain observed patterns of expression at loci governing infection/resistance in hosts and parasites. Co-expression of resistance alleles has been observed in both the R gene family in plant hosts [37,38] and the major histocompatibility complex and immunoglobulin gene families in animal hosts [39]. In contrast, many parasites typically express only one of many antigen alleles encoded by large gene families. For instance, trypanosomes typically express only one of thousands of variant surface glycoprotein genes [40,41]; Giardia express only one of 30 to 150 variant-specific surface protein genes [41,42]; ciliates also express only one of many genes encoding surface antigens [43].
Although our modelling framework is quite general in many ways, it makes several important assumptions. First, we have assumed that infection and resistance are mediated by a single genetic locus with only two alleles. Adding additional alleles or loci could conceivably alter our results by changing co-evolutionary dynamics in such a way that polymorphism is either more or less likely to be maintained (e.g., [34]). Because the maintenance of genetic polymorphism is crucial for the evolution of gene-expression modifiers, these effects could be quantitatively important, although we would not expect a qualitative effect. Second, we have not considered limitations on the evolution of increased gene expression that may arise from selection imposed by autoimmune reactions. Increasing the number of parasite-recognition molecules expressed in an IMA or GFG system might increase the likelihood of an autoimmune response. This phenomenon has been demonstrated for the adaptive immune system of vertebrates, where it is thought to select for an intermediate number of antigen receptors [44].
As we have argued, host−parasite interactions provide a theoretical framework in which to understand and interpret the evolution of genetic systems. While we had previously explored the evolution of ploidy levels in hosts and parasites [18], ploidy levels are often relatively stable over evolutionary time and have wide-ranging effects on phenotype beyond their effect on host−parasite interactions [45]. In contrast, expression levels are known to be evolutionarily labile [46] and should be much less constrained by pleiotropy, especially when cis-regulated [47]. As a consequence, we expect the results developed within this paper to yield accurate predictions over a broader range of taxa and types of interactions. Accumulating data on patterns of heterozygous gene expression at loci responsible for infection/resistance will be critical for evaluating this expectation.
Supporting Information
Protocol S1 Supporting Model Description
(85 KB DOC).
Click here for additional data file.
We thank Aneil Agrawal, Thomas Lenormand, Danny Browning, Mark Dybdahl, Dave Hall, Sylvain Gandon, and an anonymous reviewer for helpful comments. Funding was provided by the National Science Foundation (DEB-0343023 to SLN) and by the Natural Sciences and Engineering Research Council of Canada (to SPO).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. SLN and SPO conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, and wrote the paper.
Citation: Nuismer SL, Otto SP (2005) Host-parasite interactions and the evolution of gene expression. PLoS Biol 3(7): e203.
Abbreviations
GFGgene-for-gene
IMAinverse-matching-alleles
MAmatching-alleles
==== Refs
References
Nuismer SL Thompson JN Gomulkiewicz R Coevolutionary clines across selection mosaics Evolution Int J Org Evolution 2000 54 1102 1115
Gandon S Capowiez Y Dubois Y Michalakis Y Olivieri I Local adaptation and gene-for-gene coevolution in a metapopulation model Proc R Soc Lond B Biol Sci 1996 263 1003 1009
Thrall PH Burdon JJ Bever JD Local adaptation in theLinum marginale -Melampsora lini host-pathogen interaction Evolution Int J Org Evolution 2002 56 1340 1351
Ebert D Zschokke-Rohringer CD Carius HJ Within- and between-population variation for resistance of Daphnia magna to the bacterial endoparasite Pasteuria ramosa
Proc R Soc Lond B Biol Sci 1998 265 2127 2134
Thompson JN The evolution of species interactions Science 1999 284 2116 2118 10381869
Boots M Hudson PJ Sasaki A Large shifts in pathogen virulence relate to host population structure Science 2004 303 842 844 14764881
Day T Burns JG A consideration of patterns of virulence arising from host-parasite coevolution Evolution Int J Org Evolution 2003 57 671 676
Fellowes MDE Travis JMJ Linking the coevolutionary and population dynamics of host-parasitoid interactions Popul Ecol 2000 42 195 203
Nuismer SL Kirkpatrick M Gene flow and the coevolution of parasite range Evolution Int J Org Evolution 2003 57 746 754
Thrall PH Burdon JJ The spatial scale of pathogen dispersal: Consequences for disease dynamics and persistence Evol Ecol Res 1999 1 681 701
Woolhouse MEJ Webster JP Domingo E Charlesworth B Levin BR Biological and biomedical implications of the co-evolution of pathogens and their hosts Nat Genet 2002 32 569 577 12457190
Bell G The masterpiece of nature: The evolution and genetics of sexuality 1982 Berkeley (California) University of California Press 635
Hamilton WD Sex vs. non-sex vs. parasite Oikos 1980 35 282 290
Howard RS Lively CM The ratchet and the red queen: The maintenance of sex in parasites J Evol Biol 2002 15 648 656
Lively CM Trematode infection and the distribution and dynamics of parthenogenetic snail populations Parasitology 2001 123 S19 S26 11769283
Peters AD Lively CM The red queen and fluctuating epistasis: A population genetic analysis of antagonistic coevolution Am Nat 1999 154 393 405 10523486
Otto SP Nuismer SL Species interactions and the evolution of sex Science 2004 304 1018 1020 15143283
Nuismer SL Otto SP Host-parasite interactions and the evolution of ploidy Proc Natl Acad Sci U S A 2004 101 11036 11039 15252199
Fisher RA The possible modification of the response of the wild type to recurrent mutations Am Nat 1928 62 115 126
Fisher RA The evolution of dominance: Reply to Professor Sewall Wright Am Nat 1929 63 553 556
Wright S The evolution of dominance: Comment on Dr. Fisher's reply Am Nat 1929 63 556 561
Wright S Fisher's theory of dominance Am Nat 1929 63 274 279
Wright S Physiological and evolutionary theories of dominance Am Nat 1934 68 24 53
Haldane JBS A note on Fisher's theory of the origin of dominance, and on a correlation between dominance and linkage Am Nat 1930 64 87 90
Haldane JBS The theory of selection for melanism in Lepidoptera
Proc R Soc Lond B Biol Sci 1956 145 303 306
Otto SP Bourguet D Balanced polymorphisms and the evolution of dominance Am Nat 1999 153 561 574
Flor HH The complementary genetic systems in flax and flax rust Adv Genet 1956 8 29 54
Thompson JN Burdon JJ Gene-for-gene coevolution between plants and parasites Nature 1992 360 121 125
Tian D Traw MB Chen JQ Kreitman M Bergelson J Fitness costs of R-gene-mediated resistance in Arabidopsis thaliana
Nature 2003 423 74 77 12721627
Thrall PH Burdon JJ Evolution of virulence in a plant host-pathogen metapopulation Science 2003 299 1735 1737 12637745
Grosberg RK Hart MW Mate selection and the evolution of highly polymorphic self/nonself recognition genes Science 2000 289 2111 2114 11000110
Frank SA Immunology and the evolution of infectious disease 2002 Princeton (New Jersey) Princeton University Press 348
Jayakar SD A mathematical model for interaction of gene frequencies in a parasite and its host Theor Popul Biol 1970 1 140 164 5527629
Seger J Dynamics of some simple host-parasite models with more than two genotypes in each species Philos Trans R Soc Lond B Biol Sci 1988 319 541 555 2905491
Kirkpatrick M Johnson T Barton N General models of multilocus evolution Genetics 2002 161 1727 1750 12196414
Nagylaki TJ The evolution of multilocus systems under weak selection Genetics 1993 134 627 647 8325492
Young ND The genetic architecture of resistance Curr Opin Plant Biol 2000 3 285 290 10873848
Beynon JL Crute IR Holub EB Burdon JJ Molecular genetics of disease resistance: An end to the ‘gene-for-gene' concept? The gene-for-gene relationship in plant-parasite interactions 1997 New York CAB International 359 378
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
Donelson JE Mechanisms of antigenic variation in Borrelia hermsii and African trypanosomes J Biol Chem 1995 270 7783 7786 7713867
Barbour AG Restrepo BI Antigenic variation in vector-borne pathogens Emerg Infect Dis 2000 6 449 457 10998374
Svard SG Meng TC Hetsko ML McCaffery JM Gillin FD Differentiation-associated surface antigen variation in the ancient eukaryote Giardia lamblia
Mol Microbiol 1998 30 979 989 9988475
Kusch J Schmidt HJ Genetically controlled expression of surface variant antigens in free-living protozoa J Membr Biol 2001 180 101 109 11318094
Nowak MA Tarczy-Hornoch K Austyn JM The optimal number of major histocompatibility complex molecules in an individual Proc Natl Acad Sci U S A 1992 89 10896 10899 1438295
Mable BK Otto SP The evolution of life cycles with haploid and diploid phases Bioessays 1998 20 453 462
Meiklejohn CD Parsch J Ranz JM Hartl DL Rapid evolution of male-biased gene expression in Drosophila
Proc Natl Acad Sci U S A 2003 100 9894 9899 12907700
Stern SL Perspective: Evolutionary developmental biology and the problem of variation Evolution Int J Org Evolution 2000 54 1079 1091
| 15913420 | PMC1140679 | CC BY | 2021-01-05 08:21:23 | no | PLoS Biol. 2005 Jul 31; 3(7):e203 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030203 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1600001810.1371/journal.pbio.0030223Research ArticleDevelopmentEcologyEvolutionNutritionDrosophilaCalories Do Not Explain Extension of Life Span by Dietary Restriction in Drosophila
Calorie Intake and Life Span in DrosophilaMair William
1
Piper Matthew D. W
1
Partridge Linda [email protected]
1
1Centre for Research on Ageing, University College London, Department of Biology, LondonUnited KingdomKirkwood Thomas Academic EditorUniversity of Newcastle upon TyneUnited Kingdom7 2005 31 5 2005 31 5 2005 3 7 e22320 12 2004 21 4 2005 Copyright: © 2005 Mair 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.
For Long-Lived Flies, It's Calorie Quality, Not Quantity, That Matters
Dietary restriction (DR) extends life span in diverse organisms, including mammals, and common mechanisms may be at work. DR is often known as calorie restriction, because it has been suggested that reduction of calories, rather than of particular nutrients in the diet, mediates extension of life span in rodents. We here demonstrate that extension of life span by DR in Drosophila is not attributable to the reduction in calorie intake. Reduction of either dietary yeast or sugar can reduce mortality and extend life span, but by an amount that is unrelated to the calorie content of the food, and with yeast having a much greater effect per calorie than does sugar. Calorie intake is therefore not the key factor in the reduction of mortality rate by DR in this species.
Experimental evidence reveals that specific nutritional components, rather than reducing calorie intake per se, are responsible for extending lifespan via dietary restriction in Drosophila melanogaster.
==== Body
Introduction
Dietary restriction (DR), the extension of life span by reduction of nutrient intake without malnutrition, is often used as a benchmark comparison for interventions that extend life span [1–3]. Since McCay's pioneering experiments in rats 70 years ago [4], some form of food restriction has been shown to increase life span in commonly used model organisms such as yeast [5,6], nematodes [7], fruit flies [8,9], and mice [10], along with many species less often used for laboratory research such as water fleas, spiders, fish (see [3] for review), and dogs [11]. Preliminary data also suggest that DR may extend life span in nonhuman primates [12,13] and potentially give health benefits in humans [14]. Despite the finding that restricting diet increases longevity in such a diversity of species, the mechanisms responsible remain to be fully elucidated in any of them. It is therefore as yet unclear whether these mechanisms are evolutionarily conserved across taxa or if instead life extension during DR is an example of convergent evolution.
DR is often termed ‘calorie restriction' because, in rodents, daily calorie intake per se has been implicated as the key determinant of life span, with the source of these calories (i.e., carbohydrate, protein, or fat) being considered irrelevant [1]. Evidence for this point of view came from two types of experiment on rats: (1) restriction of calorie intake without reduction of protein intake resulted in life-span extension [15]; (2) no life-span extension was seen in rats fed isocaloric diets in which either the fat or mineral components had been reduced [16]. However, in other experiments, rats fed isocaloric diets with altered nutritional composition [17,18] or reduced protein [19] showed life-span extension. Furthermore, reducing just one amino acid (methionine) increases life span in both mice (R. Miller, personal communication) and rats [20]. Hence, it seems that reducing the level of ingested calories may not always be critical for life-span extension by DR in rodents. Here we address this issue in the fruit fly Drosophila melanogaster.
Results
DR can be applied in Drosophila by the simultaneous dilution of the nutrients in a standard sugar yeast (SY) food medium [9] in which the yeast is the only source of protein and lipid. As food concentration declines from maximum, life span first increases in response to DR, becoming greatest at an intermediate food concentration, before declining due to starvation at lower concentrations [9,21]. We tested the separate effects of sugar and yeast on life span at the concentrations that maximise life span (DR) and under full feeding (control).
Feeding Rates of Flies on Different Food Types
Because flies may respond to changes in dietary composition by altering their feeding behaviour, thereby potentially compensating, we determined the effect of food composition on the amount of time that the flies spent feeding on different diets. Varying the proportions of sugar or dead yeast fed to adult Drosophila females did not have a significant effect on feeding behaviours (Figure 1; p > 0.01 in all cases, chi-squared test, Bonferroni adjustment for multiple comparisons). A significant difference was seen on day 17 (chi-squared, p = 0.0068) with flies on DR yeast/control sugar food eating less. However, this difference was in the opposite direction to that expected if flies on low-nutrient diets compensated by increasing feeding rates. Hence, the flies did not compensate for decreased nutrient content of the food medium by increasing the time that they spent feeding.
Figure 1 Feeding Rates of Female Drosophila on Food Media with Different Nutrient Concentrations
Feeding rates were recorded by direct observation as the proportion of time flies spent on the surface of the media with their proboscis extended and touching the food (y-axis). Replicate measurements of the proportion of females feeding versus those not feeding were recorded during a 2-h period on the days shown. No significant different was seen between flies fed different diets on days 3, 7, 11, and 24 as assessed by chi-squared tests (p > 0.01, Bonferroni correction for multiple comparisons). There was a significant difference in feeding rates on day 17 (p = 0.0068) with flies on the DR yeast/control sugar media eating less. These data show that Drosophila does not exhibit compensatory feeding behaviour for the DR regime imposed.
Caloric Content of Dead Yeast/Sucrose
Values for yeast biomass components were taken from Lange and Heijnen [22] and estimations of the caloric content of protein, carbohydrate and lipid from Southgate and Durnin [23]. This allowed estimation of the caloric content per gram of sucrose and autolysed yeast powder, the only sources of nutrients in the Drosophila food medium. These values were 4 kcal/g sucrose and 4.02 kcal/g autolysed yeast powder. Since these values are virtually identical, changing either the sugar or yeast content of the foods between the DR and control concentrations generated food types with similar caloric values but with different nutritional compositions (see Table 1).
Table 1 Nutritional Composition and Caloric Content of Experimental Food Types
Life Span of Female Drosophila Given Foods of Different Caloric Value
Life span of female Drosophila was extended much more by reduction of yeast from control to DR concentration than by the equivalent reduction in sugar (Figure 2; Table 2), and median life span therefore did not correlate with caloric content of the food medium to which the flies were exposed (Figure 3). In two independent experiments, reducing yeast concentration from control to DR levels whilst keeping sugar levels constant significantly increased life span (p < 0.0001 in both cases, log-rank test). Lowering caloric content to the same extent by reducing sugar from control to DR levels increased life span at DR yeast levels in both experiments (p < 0.0001 in both cases, log-rank test), but the effect on median life span was much less than that of changing yeast levels (Figure 3; Table 2). Reducing sugar from control to DR concentrations whilst keeping yeast at control levels significantly increased life span in experiment 1 (p < 0.0001, log-rank test), but again the effect on median life span was much less than that of changing yeast levels (Figure 3; Table 2). Reducing sugar from control to DR concentrations whilst maintaining yeast at control levels increased median life span in experiment 2 (Figure 3), but the effect on life span was not significant (p > 0.05, log-rank test).
Figure 2 Survivorship (lx) Analysis of Life Span of Female Drosophila on Different Food Regimes
Colour/Symbol of the curves shows yeast level while the line type represents sugar levels in the respective foods. (A) and (B) are independent repeats. In both cases, changing caloric content of the food by altering yeast levels had a much greater effect on life span than that seen when the same change in caloric content was brought about by manipulating sugar levels.
Figure 3 Plot of Median Life Span of Female Drosophila against the Estimated Caloric Content of the Food Medium
(A) and (B) represent independent repeats. Red arrows link pairs of food types where differences in caloric content are due to different yeast concentrations. Blue arrows link pairs of food types where differences in caloric content are due to different sugar concentrations. Green arrow links food types where differences in caloric content are due to both different sugar and yeast concentrations. Life span is extended to a greater extent per calorie by reducing yeast concentration from control to DR levels than by reducing sugar. This is in contrast to what would be predicted if calorie intake were the key mediator of life-span extension by DR in fruit flies.
Table 2 Median and Maximum Life Span of Flies Fed Different Food Media as Adults
Effect of Bacteria on Response of Life Span to Diet
To test if different levels of bacteria in the food medium could account for effects of nutrient composition on life span, we tested the effect of an antibiotic. The addition of the antibiotic tetracycline to the food media did not have a significant effect on life span on either control or DR food medium (Figure 4, p > 0.05; log-rank test in each case), and the life-span extension seen when sugar and yeast levels were reduced from control to DR concentrations was therefore not blocked or modified by the addition of antibiotic to the food medium.
Figure 4 Effect of Tetracycline on Life Span of Female D. melanogaster
The addition of the antibiotic tetracycline to the food media did not have a significant effect on life span at either control or DR concentration food media.
Effects on Mortality of Switching Yeast and Sugar
The effect of DR on mortality in Drosophila is acute; within 48 h flies switched between DR and control diets adopt the mortality rates characteristic of flies chronically exposed to the nutritional regime that the switched flies have joined [24]. We therefore measured the acute effects on mortality of switching the yeast and sugar components of the diet separately. When yeast was switched, mortality rates responded similarly to the responses to switches between control and DR SY food medium. Forty-eight hours after being switched from control SY medium to DR yeast/control sugar medium at day 25, flies were no more likely to die than those maintained on DR yeast/control sugar medium throughout adulthood (Cox regression; p = 0.22; n DR yeast/control sugar chronic group = 626; n switch group = 475; risk ratio = 0.96 [95% confidence interval CI: 0.91, 1.02]) (Figure 5A). In the reciprocal switch, flies moved from DR SY medium to control yeast/DR sugar medium showed a rapid increase in mortality rate, although this did not quite reach the level seen in flies that had been on control yeast/DR sugar medium throughout adult life (Cox regression; p < 0.05; n control yeast/DR sugar chronic group = 480; n switch group = 668; risk ratio = 0.88 [95% CI: 0.83, 0.93]) (Figure 5A).
Figure 5 The Acute Effects on Age-Specific Mortality in Drosophila of Changes in Nutritional Content of the Food Midway through Life
Vertical line represents switch day. Mortality trajectories were truncated when n < 40.
(A) Switching between control and DR yeast (Y) diets midway though life results in rapid changes in age-specific mortality rates within 48 h similar to those seen previously for whole food dilutions [24]. Control yeast intake caused no irreversible damage since flies switched from control yeast to DR yeast at day 25 rapidly became no more likely to die than those flies given DR yeast levels throughout adulthood. Flies with a history of DR yeast levels showed rapid increases in mortality rate when moved to control yeast levels at day 25, but mortality rates did not become as high as those of flies that had been maintained on control yeast levels permanently.
(B) Changing caloric intake to the same extent via changes to sugar (S) levels rather than yeast did not cause rapid changes in mortality rate. Despite flies chronically fed control sugar and DR yeast having increased mortality rate compared to the DR control, switching from DR to control sugar late in life did not increase mortality rate.
In contrast, switching of sugar had no significant effect on mortality. From 48 h after being switched from control SY medium to control yeast/DR sugar at day 25, no significant difference was seen between the mortality of switched flies and the unswitched group maintained on control SY medium (Cox regression; p = 0.34; n control SY = 427; n switch group = 440; risk ratio = 0.97 [95% CI: 0.91, 1.04]) (Figure 5B). Similarly, flies switched to DR yeast/control sugar from DR SY medium at day 25 did not show increased mortality in comparison to unswitched controls (Cox regression; p = 0.41; n DR group = 615; n switch group = 676; risk ratio = 0.98 [95% CI: 0.93, 1.03]) (Figure 5B). A second experiment that was terminated 4 d after the switch in diet gave the same result (see Figure S1). These data show that the rapid switch in mortality rates upon changes between DR and control food medium are overwhelmingly attributable to the yeast rather than to the sugar component of the diet.
Discussion
Life Span Is Not Related to Calorie Intake
Flies fed food media with very similar caloric content showed marked differences in their life spans (see Figure 3). This finding is in direct contrast to what would be predicted if ingested calories were the key mediator of life span in D. melanogaster and demonstrates that the nutritional composition of the diet affects life-span extension by DR in this species. Reduction in the concentration of either sugar or yeast levels increased life span (see Figures 2 and 3). However, the magnitude of the effects on life span when the caloric content of the food was changed via altering yeast concentration far exceeded that seen when calories were changed to the same extent via manipulation of sugar levels, suggesting that protein/lipid levels have a greater effect on Drosophila survival than does carbohydrate. The differing effect of sugar and yeast on mortality in Drosophila could occur if different pathways sense nutrients during DR, possibly with different outputs affecting life span. Sir2 [25,26], Rpd3 [27], the insulin/IGF-like signalling [28], and target of rapamycin pathways [29] have all been implicated in mediating the response of life span to DR in Drosophila, with the latter two suggested to interact in the fly to control growth in response to nutrient levels [30]. The role of these and other candidate pathways in mediating the response of life span to specific nutrients should be investigated further. Sugar and yeast could affect mortality rates differently if they differentially modulate metabolic or other processes that increase risk of death.
Experimentally increased reproduction has been shown to decrease life span in a variety of species [31–35] and the level of dietary yeast and egg production are positively correlated in Drosophila [8,9]. Therefore an obvious hypothesis as to why there is a greater response of life span in Drosophila to changes in yeast than in sugar is that the increased mortality on control yeast levels represents the cost of reproduction, which correlates with yeast intake and not with sugar. However, since life-span extension via DR in Drosophila occurs normally when egg production or vitellogenesis are blocked either by X-irradiation or genetically [36], the greater response of life span to changes in yeast is not directly attributable to the reduction of reproductive output. Furthermore, although the magnitude of the response to DR in male Drosophila is less than that of females [21], males do live longer if nutrient levels are reduced, and they show the same rapid changes in mortality as females when dietary regime is changed [24], yet they do not suffer the high costs of producing eggs on high yeast.
Rapid Changes in Mortality in Response to DR Are Attributable Solely to Yeast Content
DR acts acutely to extend life span in Drosophila; it does not slow the accumulation of irreversible damage with age [24]. Flies subjected to DR for the first time in midlife rapidly become no more likely to die than those that have been under DR throughout adulthood [24]. We investigated the roles of the sugar and yeast components of the diet in producing this rapid change in mortality rate in flies switched between DR and control conditions. When flies previously subjected to control SY food were switched to DR yeast levels, there was a rapid (within 48 h) drop in mortality rates to those seen in the flies chronically exposed to DR yeast/control sugar food (see Figure 5A). A similar rapid increase in mortality rates was seen when flies exposed to DR food were switched to control yeast levels (Figure 5A), although, as seen previously using whole food dilutions [24], a history of low yeast gave slight protection to female Drosophila moved to control yeast late in life.
However, when caloric content of the food given to flies was changed to the same extent midway though life by changing sugar rather than yeast levels, no change in mortality rate was seen (Figure 5B). Therefore the acute mortality ‘switch' phenotype in response to dietary restriction is attributable to changes in the level of the dietary yeast alone. That chronically reducing sugar intake of flies can extend life span, yet reducing sugar intake late in life does not cause rapid changes to mortality rates, suggests the deleterious effects of sugar may occur mainly early in adult life. The mortality trajectories in Figure 5 support this conclusion, by showing that the lowering of mortality rate in response to sugar is most obvious early in the trajectory, when mortality rates in all groups of flies are low. More work is needed using accurately defined media to investigate this effect. Rapid reductions in mortality rate have been seen previously in Drosophila by altering the intake of yeast only [37]. However, the results of the previous study differ from those here in that reduced mortality was achieved by increasing the nutrient intake of flies that had previously been deprived of yeast, rather than by reducing the nutrient intake of control-fed flies.
Feeding Rates of Flies on Different Food Types
Unlike in rodents, where DR can be achieved by directly reducing the quantity of food eaten in comparison to animals given ad libitum access [1], DR is achieved in Drosophila by reducing the quality (nutrient concentration) of the food given to the flies [9] with the quantity maintained in excess of that which they can consume. Despite the fact that fecundity correlates with food medium concentration [9], it has been suggested that flies may be able to compensate when faced with reduced nutrients by increasing feeding rates, and therefore they may not be dietarily restricted [38]. However, our results suggest that flies on low-quality media do not compensate by eating more, as measured by time spent on the food with the proboscis extended. It is possible Drosophila can alter the rate of food uptake per unit time that the proboscis is extended, in which case our indirect measurements would not detect these changes. More direct approaches to quantify feeding rates require radio-labelling the food [39] or the addition of coloured food dye [40], with uptake rates assessed upon short-term exposure to labelled food. However, our own unpublished observations show that flies moved to fresh food medium display elevated feeding behaviour that is unrepresentative of the steady-state situation and that leads to a highly nonlinear relationship between time and uptake of the food label. We hence used the behavioural measure described here, which better represents the normal feeding of the flies. Our feeding assay results, in combination with the reduced fecundity seen as food nutrient concentration is reduced, suggest that diluting the food medium results in a co-ordinate reduction in the intake of nutrients in Drosophila and therefore is a robust protocol for DR in this species.
Effect of Tetracycline on Life Span
It has been suggested that higher nutrient concentrations in fly food may lead to higher proliferation rates of bacteria on the media, which in turn could increase mortality of D. melanogaster in a mechanism that is unrelated to ingestion of different amounts of nutrients [38]. If this were the case then we would expect that (1) flies fed antibiotics would live longer, and (2) the life-span extension seen when nutrient concentration is reduced would be blocked when antibiotics are present. Tetracycline did not extend the life span of flies in our experiments, nor did it block the DR response, meaning either that reduced bacterial challenge is not the mechanism by which diluting food media extends life span in Drosophila, or that the relevant microorganisms are tetracycline resistant.
Conclusions
The response of Drosophila life span to nutrition is not governed by calories, but rather by specific nutritional components of the food. This finding represents a departure from the generally accepted model in rodents, where it has been suggested that the level of calorie intake per se, not the source of calories, is critical for life-span extension [1]. The apparent disparity between the factors in the diet that affect life span in fruit flies and rodents leads to two possible conclusions. First, the mechanisms by which these organisms respond to food shortage could be different. Second, the long-held view that calorie intake is the critical variable in the response of mammalian life span to DR may require further evaluation.
Despite some reports in the literature that DR did not extend life span [38,41,42], the overwhelming majority of data support the idea that DR in some form extends life span across diverse taxa. However, it is still unknown if life-span extension under DR is achieved through common mechanisms in different species. A case for conservation of the mechanisms by which DR extends life span can be made from evolutionary considerations. It has been suggested that, during times of famine, diversion of resources away from reproduction towards somatic maintenance will increase the chances of an organism surviving to more plentiful times and thus increase long-term reproductive success [43–46]. The selective advantage of shifting resources from reproduction to maintenance when food is restricted could be the “public” factor shared between diverse organisms. However, the mechanisms by which extension of life span is achieved could be an example of convergent evolution, producing the same plasticity of life span in response to food shortage through mechanisms at least to some extent specific to different organisms, dependent upon their diet, experience of food shortages, and life history. More work is needed to elucidate the precise relationship between the composition of the diet and life span in different organisms, including mammals. Our results suggest that it may be possible to obtain the full extension of life span by DR by reducing critical nutrients in the food without any reduction in overall calorie intake.
Materials and Methods
Fly stocks and husbandry
The wild-type stock used in all experiments was collected in Dahomey (now Benin) in 1970 and has since been maintained in large population cages with overlapping generations on a 12:12-h light:dark cycle at 25 °C. This culturing method has been shown to maintain life span and fecundity at levels similar to those in freshly collected flies [47].
Feeding rates of flies on different food types
To measure feeding rates in Drosophila we observed behaviour of age-matched, once-mated Dahomey females on each of the four food types. This approach was adopted in preference to direct quantification of ingested food [40] because DR flies transiently elevate their feeding rate following transfer onto new food (unpublished observations). In the present assay, 30 female flies were individually allocated to a vial containing either control SY, control Y/DR S, DR Y/control S, or DR SY and placed at 25 °C overnight to adopt their undisturbed pattern of feeding. The following day, 1 h after lights on, observations were taken for a 2-h period, and flies were scored as eating if they were on the food with their proboscis extended and touching the food surface. During this time, 360 observations of flies in each treatment were made (12 observations of 30 flies) except on day 24 when 18 observations were made of each treatment set. The final data are the proportion of flies feeding out of the feeding opportunities given (total observations). Differences between treatments at a given time point were assessed using the chi-squared test.
Effect of tetracycline on life span
Tetracycline is a general antibiotic that inhibits ribosomal translocation and acts on both Gram-positive and negative bacteria [48]. A tetracycline solution was made up in 70% ethanol and added to the food media after it had been boiled and cooled to 60 °C. The final concentration of tetracycline in the media was 0.025% weight/volume [49], five times more than that used when tetracycline resistance is utilised as a selectable marker for bacterial transformation [50]. The wild-type stock Dahomey is infected by the cytoplasmic bacteria Wolbachia (unpublished). A 0.025% tetracycline solution is sufficient to remove bacteria such as Wolbachia from Drosophila stocks if fed to larvae [49] and can suppress Wolbachia in other insects when fed to adults only [51]. Therefore flies fed tetracycline media as adults may not only have reduced exposure to external microorganisms on the food surface compared to controls, but may also have reduced Wolbachia infection. Seven millilitres of food was poured into 30-ml glass vials and the life span of flies measured with 92–101 flies per treatment and 10 flies per vial. Fresh food was prepared once a week and flies moved onto new media three times per week.
Life span experiments
Experimental flies were raised at a standard density of 400–450 eggs per 200-ml bottle [52] on standard SY medium (1,000 ml distilled water, 100 g autolysed yeast powder, 100 g sucrose, 20 g agar, 30 ml Nipagin (100 gl–1), 3 ml propionic acid). Adults were collected over a 24-h period and transferred without anaesthesia to fresh SY food for 48 h and allowed to mate. Females were then collected using light CO2 anaesthesia and assigned randomly to the food regimes (Table S1). All experiments were done with mated females. Flies were kept on 35 ml of food at an initial density of 100 individuals per 200-ml bottle and transferred without anaesthesia to fresh food every 2–3 d. Deaths were scored 5–6 d a week and initial sample sizes (n
0) were calculated as the summed death and censor observations over all ages. To minimise any density effects on mortality, two bottles within cohorts were merged when the density of flies reached 50 ± 10. To standardise the effects of parental age on offspring fitness [53], parents of experimental flies were of the same age and reared at a constant density.
Statistical analysis
Age-specific mortality (μx) was estimated as μx = −ln(px), where px is the probability of an individual alive at age x − 1 surviving to age x [54]. log-rank tests [55] were used for survivorship analysis. All statistical analysis was performed using JMP. 5.0 statistical software (SAS Institute Inc., Cary, North Carolina, United States).
Supporting Information
Figure S1 The Acute Effects on Age-Specific Mortality in Drosophila of Changes in Nutritional Content of the Food Midway through Life
Vertical line represents switch day. Experiment was terminated 4 d after the switch.
(A) Similar to the experiment shown in Figure 5, switching between control and DR yeast (Y) diets midway though life results in rapid changes in age-specific mortality rates within 48 h similar to those seen previously for whole food dilutions [24].
(B) Changing caloric intake to the same extent via changes to sugar (S) levels rather than yeast did not cause rapid changes in mortality rate.
(839 KB TIF).
Click here for additional data file.
Table S1 Sample Sizes and Treatments
These represent the number of flies switched between treatments (i.e., n
25) and were sampled from the original chronic controls (control SY or DR SY) and censored from the life-span data of these treatments at day 25.
(27 KB DOC).
Click here for additional data file.
We thank P. Martinez, F. C. F. Calboli, Y. Driege, and T. Magwere for experimental assistance, G. D. D. Hurst for help and advice with the antibiotic experiment, S. D. Pletcher and members of the Partridge lab for valuable discussion, and three reviewers for helpful comments. We also thank T. G. Standish for the use of his EM Drosophila picture. This work was supported by the Biotechnology and Biological Research Council and the Wellcome Trust.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. LP and WM conceived and designed the experiments. WM performed the experiments and analyzed the data. LP, WM, and MDWP wrote the paper.
Citation: Mair W, Piper MDW, Partridge L (2005) Calories do not explain extension of life span by dietary restriction in Drosophila. PLoS Biol 3(7): e223.
Abbreviations
CIconfidence interval
DRdietary restriction
SYsugar yeast
==== Refs
References
Masoro EJ Caloric restriction: A key to understanding and modulating aging 2002 Amsterdam Elsevier 183
Sohal RS Weindruch R Oxidative stress, caloric restriction, and aging Science 1996 273 59 63 8658196
Weindruch R Walford RL The retardation of aging and disease by dietary restriction 1988 Springfield (Illinois) Thomas 436
McCay CM Crowell MF Maynard LA The effect of retarded growth upon the length of life span and upon the ultimate body size J Nutr 1935 10 63 79
Jiang JC Jaruga E Repnevskaya MV Jazwinski SM An intervention resembling caloric restriction prolongs life span and retards aging in yeast FASEB J 2000 14 2135 2137 11024000
Lin SJ Kaeberlein M Andalis AA Sturtz LA Defossez PA Calorie restriction extends Saccharomyces cerevisiae life span by increasing respiration Nature 2002 418 344 348 12124627
Klass Aging in the nematode Caenorhabditis elegans Major biological and environmental factors influencing life span Mech Ageing Dev 1977 6 413 429 926867
Chippindale AK Leroi AM Kim SB Rose MR Phenotypic plasticity and selection in Drosophila life-history evolution. 1. Nutrition and the cost of reproduction J Evol Biol 1993 6 171 193
Chapman T Partridge L Female fitness in Drosophila melanogaster An interaction between the effect of nutrition and of encounter rate with males Proc R Soc Lond B Biol Sci 1996 263 755 759
Weindruch R Walford RL Dietary restriction in mice beginning at 1 year of age: Effect on life-span and spontaneous cancer incidence Science 1982 215 1415 1418 7063854
Kealy RD Lawler DF Ballam JM Mantz SL Biery DN Effects of diet restriction on life span and age-related changes in dogs J Am Vet Med Assoc 2002 220 1315 1320 11991408
Roth GS Ingram DK Lane MA Calorie restriction in primates: Will it work and how will we know? J Am Geriatr Soc 1999 47 896 903 10404938
Lane MA Tilmont EM De Angelis H Handy A Ingram DK Short-term calorie restriction improves disease-related markers in older male rhesus monkeys (Macaca mulatta)
Mech Ageing Dev 2000 112 185 196 10687924
Fontana L Meyer TE Klein S Holloszy JO Long-term calorie restriction is highly effective in reducing the risk for atherosclerosis in humans Proc Natl Acad Sci U S A 2004 101 6659 6663 15096581
Masoro EJ Iwasaki K Gleiser CA McMahan CA Seo EJ Dietary modulation of the progression of nephropathy in aging rats—An evaluation of the importance of protein A J Clin Nutr 1989 49 1217 1227
Iwasaki K Gleiser CA Masoro EJ McMahan CA Seo EJ Influence of the restriction of individual dietary-components on longevity and age-related disease of Fischer rats—The fat component and the mineral component J Gerontol 1988 43 B13 B21 3335742
Dalderup LM Visser W Influence of extra sucrose in the daily food on the life-span of Wistar albino rats Nature 1969 1050 1052
Iwasaki K Gleiser CA Masoro EJ McMahan CA Seo EJ The influence of dietary-protein source on longevity and age-related disease processes of Fischer rats J Gerontol 1988 43 B5 B12 3335746
Yu BP Masoro EJ McMahan CA Nutritional influences on aging of Fischer-344 rats. 1. Physical, metabolic, and longevity characteristics J Gerontol 1985 40 657 670 4056321
Zimmerman JA Malloy V Krajcik R Orentreich N Nutritional control of aging Exp Gerontol 2003 38 47 52 12543260
Magwere T Chapman T Partridge L Sex differences in the effect of dietary restriction on life span and mortality rates in female and male Drosophila melanogaster
J Gerontol A Biol Sci Med Sci 2004 59 3 9 14718480
Lange HC Heijnen JJ Statistical reconciliation of the elemental and molecular biomass composition of Saccharomyces cerevisiae
Biotechnol Bioeng 2001 75 334 344 11590606
Southgate DAT Durnin JVGA Calorie conversion factors. An experimental reassessment of the factors used in the calculation of the energy value of human diets Br J Nutr 1970 24 517 535 5452702
Mair W Goymer P Pletcher SD Partridge L Demography of dietary restriction and death in Drosophila
Science 2003 301 1731 1733 14500985
Rogina B Helfand SL Sir2 mediates longevity in the fly through a pathway related to calorie restriction Proc Natl Acad Sci U S A 2004 101 15998 16003 15520384
Wood JG Rogina B Lavu S Howitz K Helfand SL Sirtuin activators mimic caloric restriction and delay ageing in metazoans Nature 2004 430 686 689 15254550
Rogina B Helfand SL Frankel S Longevity regulation by Drosophila Rpd3 deacetylase and caloric restriction Science 2002 298 1745 12459580
Clancy DJ Gems D Hafen E Leevers SJ Partridge L Dietary restriction in long-lived dwarf flies Science 2002 296 319 11951037
Kapahi P Zid BM Harper T Koslover D Sapin V Regulation of lifespan in Drosophila by modulation of genes in the TOR signaling pathway Curr Biol 2004 14 885 890 15186745
Colombani J Raisin S Pantalacci S Radimerski T Montagne J A nutrient sensor mechanism controls Drosophila growth Cell 2003 114 739 749 14505573
Williams GC Natural selection, the cost of reproduction and a refinement of Lack's principle Am Nat 1966 100 687 690
Rose MR Laboratory evolution of postponed senescence in Drosophila melanogaster
Evolution 1984 38 1004 1010
Gustafsson L Part T Acceleration of senescence in the collared flycatcher Ficedula-Albicollis by reproductive costs Nature 1990 347 279 281
Nager RG Monaghan P Houston DC The cost of egg production: Increased egg production reduces future fitness in gulls J Avian Biol 2001 32 159 166
Barnes AI Partridge L Costing reproduction Anim Behav 2003 66 199 204
Mair W Sgro CM Johnson AP Chapman T Partridge L Lifespan extension by dietary restriction in female Drosophila melanogaster is not caused by a reduction in vitellogenesis or ovarian activity Exp Gerontol 2004 39 1011 1019 15236760
Good TP Tatar M Age-specific mortality and reproduction respond to adult dietary restriction in Drosophila melanogaster
J Insect Physiol 2001 47 1467 1473 12770153
Cooper TM Mockett RJ Sohal BH Sohal RS Orr WC Effect of caloric restriction on life span of the housefly, Musca domestica
FASEB J 2004 18 1591 1593 15319362
Brummel T Ching A Seroude L Simon AF Benzer S
Drosophila lifespan enhancement by exogenous bacteria Proc Natl Acad Sci U S A 2004 101 12974 12979 15322271
Edgecomb RS Harth CE Schneiderman AM Regulation of feeding-behavior in adult Drosophila melanogaster varies with feeding regime and nutritional state J Exp Biol 1994 197 215 235 7852903
Forster MJ Morris P Sohal RS Genotype and age influence the effect of caloric intake on mortality in mice FASEB J 2003 17 690 692 12586746
Le Bourg E Minois N Failure to confirm increased longevity in Drosophila melanogaster submitted to a food restriction procedure J Gerontol A Biol Sci Med Sci 1996 51 B280 B283 8680993
Harrison DE Archer JR Natural selection for extended longevity from food restriction Growth Dev Aging 1988 52 65 65
Holliday R Food, reproduction and longevity—Is the extended lifespan of calorie-restricted animals an evolutionary adaptation Bioessays 1989 10 125 127 2730632
Masoro EJ Austad SN The evolution of the antiaging action of dietary restriction: A hypothesis J Gerontol A Biol Sci Med Sci 1996 51 B387 B391 8914486
Shanley DP Kirkwood TBL Calorie restriction and aging: A life-history analysis Evolution 2000 54 740 750 10937249
Sgro CM Partridge L Evolutionary responses of the life history of wild-caught Drosophila melanogaster to two standard methods of laboratory culture Am Nat 2001 156 341 353
Davies J Smith DI Plasmid-determined resistance to antimicrobial agents Annu Rev Microbiol 1978 32 469 518 360974
Hurst GDD Johnson AP Schulenburg JH Fuyama Y Male-killing Wolbachia in Drosophila A temperature sensitive trait with a threshold density Genetics 2000 156 699 709 11014817
Sambrook J Fritsch EF Maniatis T Molecular cloning—A laboratory manual 1989 Cold Spring Harbor (New York) Cold Spring Harbor Laboratory Press 3 vol
Hurst GDD Majerus MEN Walker LE Cytoplasmic male killing elements in Adalia bipunctata (Linnaeus) (Coleoptera: Coccinellidae) Heredity 1992 69 84 91
Clancy DJ Kennington WJ A simple method to achieve consistent larval density in bottle cultures Drosoph Inf Serv 2001 84 168 169
Priest NK Mackowiak B Promislow DEL The role of parental age effects on the evolution of aging Evolution 2002 56 927 935 12093028
Lee ET Statistical methods for survival data analysis 1992 New York Wiley 513
Peto R Peto J Asymptomatically efficient rank invariant procedures J R Stat Soc [Ser A] 1972 135 185 207
| 16000018 | PMC1140680 | CC BY | 2021-01-05 08:21:23 | no | PLoS Biol. 2005 Jul 31; 3(7):e223 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030223 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030237SynopsisDevelopmentEcologyEvolutionNutritionDrosophilaFor Long-Lived Flies, It's Calorie Quality, Not Quantity, That Matters Synopsis7 2005 31 5 2005 31 5 2005 3 7 e237Copyright: © 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.
Calories Do Not Explain Extension of Life Span by Dietary Restriction in Drosophila
==== Body
In April, the United States Centers for Disease Control and Prevention released a study challenging the conventional wisdom that eating less promotes longevity. The study found that the very thin run roughly the same risk of early death as the overweight. And now the tide seems to be turning against a common explanation for the long-standing observation that restricting food in lab organisms from yeast to mice prolongs life.
Many studies have indicated that it's calorie reduction, rather than the specific source of calories, that increases longevity. That this effect occurs in such diverse organisms suggests a common mechanism may be at work, though none has been definitively characterized. And while calorie restriction enhances longevity in mice, it has not always done so in rats. In a new study, William Mair, Matthew Piper, and Linda Partridge show that flies can live longer without reducing calories but by eating proportionally less yeast, supporting the notion that calorie-restriction-induced longevity may not be as universal as once thought.
Dietary restriction in Drosophila involves diluting the nutrients in the fly's standard lab diet of yeast and sugar to a level known to maximize life span. Since both yeast (which contributes protein and fat) and sugar (carbohydrates) provide the same calories per gram, the authors could adjust nutrient composition without affecting the calorie count, allowing them to separate the effects of calories and nutrients. The standard restricted diet had equivalent amounts of yeast and sugar (65 grams each) and an estimated caloric content of 521, while the yeast-restricted (65 g yeast/150 g sugar) and sugar-restricted (65 g sugar/150 g yeast) diets each had just over 860 calories. The control diet for the flies had equivalent amounts of sugar and yeast (150 grams), amounting to an estimated 1,203 calories.
Contrary to popular belief, life span extension by dietary restriction in Drosophila is not explained by calories
First, the authors had to make sure the flies didn't change their eating behavior to make up for a less nutritious diet. (They didn't.) Reducing both nutrients increased the flies’ life spans, but yeast had a much greater effect: reducing yeast from control to dietary restriction levels increased median life span by over 60%.
In a previous study, Mair et al. showed that flies that were switched from dietary-restricted diets to control diets soon began to die at the same rates as flies accustomed to the control diet. In this study, the authors studied the effects of switching yeast and sugar. Forty-eight hours after being switched from normal diets to yeast-restricted diets, flies were no more likely to die than flies fed the yeast-restricted diet from the beginning. In contrast, those switched from the standard restriction diet to the sugar-restricted diet began to die at the same rate as flies on the control diet.
The authors also ruled out the possibility that bacteria—attracted to high-nutrient food—might be influencing fly survival. Altogether these results make a strong case that calories per se are not the salient factor in prolonging life—at least in fruitflies. The dramatic impact of reducing yeast suggests that protein or fat plays a greater role in fly longevity than sugar. This in turn suggests, the authors argue, that yeast and sugar trigger different metabolic pathways with different effects on life span.
Why might different factors promote longevity in flies and rats? It could be that the caloric-restriction/longevity paradigm needs more rigorous review—though a vast body of literature does support it. Or it may be that the animals use the same strategy for dealing with food shortages—shifting resources from reproduction to survival, for example—but have evolved different mechanisms for doing so that reflect each species's life history, diet, and environment. Whatever explains the disparity, this study should give researchers interested in caloric restriction plenty to chew on.
| 0 | PMC1140681 | CC BY | 2021-01-05 08:28:15 | no | PLoS Biol. 2005 Jul 31; 3(7):e237 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030237 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030252SynopsisEcologyEvolutionGenetics/Genomics/Gene TherapyInfectious DiseasesNoneHost–Parasite Battles Shed Light on the Evolution of Gene Expression Synopsis7 2005 31 5 2005 31 5 2005 3 7 e252Copyright: © 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.
Host-Parasite Interactions and the Evolution of Gene Expression
==== Body
In human–pathogen encounters, the battle for advantage plays out at the level of gene expression. Hosts will stand a better chance of surviving if their genes confer resistance to diverse pathogens, while pathogens need genes that promote virulence and infection. To understand this game of evolutionary one-upsmanship, biologists study the genetic basis of resistance and infection by investigating how changes in an organism's genetic makeup, or genotype, affect its physiological and physical makeup, or phenotype.
In a new study, Scott Nuismer and Sarah Otto detail how host–parasite interactions shape the changes in gene expression that alter an organism's ability to induce or resist infection. They find that gene expression for host and parasite follows quite different evolutionary paths: hosts express as many different gene variants, or alleles, as possible, while parasites express very few alleles. It's in the host's interest to have as many genetic weapons as possible that can recognize a foreign invader, while it's in the parasite's interest to reduce the number of recognizable molecules for a host to latch on to and destroy. Even though these results are intuitive, this phenomenon has not been shown before.
To model the evolution of gene expression levels in a host–parasite interaction, Nuismer and Otto started with a single gene A in hosts and a single gene B in parasites with two alleles each (A or a, and B or b) that are involved respectively in resistance in the host and promoting infection by the parasite. Nuismer and Otto's model allows gene expression levels to be regulated by an additional modifier gene in hosts and parasites. The two variants of the modifier gene (M and m) alter the expression of the host gene A or parasite gene B. As a result of host–parasite interactions, the alleles at the modifier gene either evolve to increase expression of only one allele or to co-express both alleles.
Interactions between hosts and parasites shape changes in gene expression, potentially maximizing parasite recognition for hosts while minimizing detection for parasites (such as red blood cells and trypanosomes)
When parasite and host are allowed to interact, the model shows that host resistance alleles typically evolve toward co-expression while parasite infection alleles evolve toward single expression. By expressing more than one gene at a time, the host can recognize a greater diversity of parasites. But what's good for the host is bad for the parasite. Hosts benefit from a wider array of parasite recognition systems, while parasites benefit from expressing a narrow range of antigens to evade the host recognition system.
Human immune cells, for example, can recognize billions of different antigens, which then triggers an immune response against the foreign substance and increases the chance of surviving the infection. Parasites, however, generally express only one of many possible antigen alleles. The parasite responsible for African sleeping sickness expresses only one of thousands of surface receptor genes, which offers the host fewer opportunities for detection.
Nuismer and Otto's model provides a framework for understanding empirical observations of allele expression in known host–parasite interactions and may well help explain similar modifications in allele expression in other systems. Because the model also provides testable predictions, it should be useful in interpreting data from a wider range of species and interactions, furthering our understanding of the evolutionary forces that shape infection and resistance and ultimately influence how genes evolve.
| 0 | PMC1140682 | CC BY | 2021-01-05 08:21:23 | no | PLoS Biol. 2005 Jul 31; 3(7):e252 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030252 | oa_comm |
==== Front
BMC GenetBMC Genetics1471-2156BioMed Central London 1471-2156-6-221588513810.1186/1471-2156-6-22Research ArticleCommon polymorphism in H19 associated with birthweight and cord blood IGF-II levels in humans Petry Clive J [email protected] Ken K [email protected] Bryan J [email protected] Diane [email protected] Heather J [email protected] Susan M [email protected] Marcus E [email protected] ALSPAC Study Team [email protected] Wolf [email protected] John A [email protected] David B [email protected] Department of Paediatrics, University of Cambridge, Addenbrooke's Hospital Level 8, Box 116, Cambridge CB2 2QQ, UK2 Juvenile Diabetes Research Foundation/Wellcome Trust Diabetes and Inflammation Laboratory, Cambridge Institute for Medical Research, University of Cambridge, Addenbrooke's Hospital, Cambridge, UK3 Unit of Paediatric and Perinatal Epidemiology, University of Bristol, 24 Tyndall Avenue, Bristol, UK4 Clinical and Molecular Genetics Unit, Institute of Child Health, London, UK5 Laboratory of Developmental Genetics and Imprinting, Babraham Institute, Cambridge, UK2005 10 5 2005 6 22 22 19 10 2004 10 5 2005 Copyright © 2005 Petry et al; licensee BioMed Central Ltd.2005Petry 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
Common genetic variation at genes that are imprinted and exclusively maternally expressed could explain the apparent maternal-specific inheritance of low birthweight reported in large family pedigrees. We identified ten single nucleotide polymorphisms (SNPs) in H19, and we genotyped three of these SNPs in families from the contemporary ALSPAC UK birth cohort (1,696 children, 822 mothers and 661 fathers) in order to explore associations with size at birth and cord blood IGF-II levels.
Results
Both offspring's and mother's H19 2992C>T SNP genotypes showed associations with offspring birthweight (P = 0.03 to P = 0.003) and mother's genotype was also associated with cord blood IGF-II levels (P = 0.0003 to P = 0.0001). The offspring genotype association with birthweight was independent of mother's genotype (P = 0.01 to P = 0.007). However, mother's untransmitted H19 2992T allele was also associated with larger birthweight (P = 0.04) and higher cord blood IGF-II levels (P = 0.002), suggesting a direct effect of mother's genotype on placental IGF-II expression and fetal growth. The association between mother's untransmitted allele and cord blood IGF-II levels was more apparent in offspring of first pregnancies than subsequent pregnancies (P-interaction = 0.03). Study of the independent Cambridge birth cohort with available DNA in mothers (N = 646) provided additional support for mother's H19 2992 genotype associations with birthweight (P = 0.04) and with mother's glucose levels (P = 0.01) in first pregnancies.
Conclusion
The common H19 2992T allele, in the mother or offspring or both, may confer reduced fetal growth restraint, as indicated by associations with larger offspring birth size, higher cord blood IGF-II levels, and lower compensatory early postnatal catch-up weight gain, that are more evident among mother's smaller first-born infants.
==== Body
Background
The maternal-uterine environment has a major influence on fetal growth and size at birth [1]. Early cross-breeding experiments in animals demonstrated that fetal growth could be restrained or enhanced from its genetic potential, according to the size of the mother [2,3]. In a normal human birth cohort study, we demonstrated that over 50% of all offspring may have experienced clinically significant restraint or enhancement of growth in utero, followed by compensatory catch-up or catch-down growth respectively during the first 2 years of postnatal life [4].
Birth order is an important common determinant of maternal-uterine restraint, as first-born infants are more likely to be restrained in utero and show compensatory postnatal catch-up growth, compared to subsequent-born infants [5]. Heritable or genetic factors may also contribute to degree of maternal restraint of fetal growth; from studies of birthweights in extended human families it has been suggested that low birthweight may be maternally transmitted [6]. We previously reported association between thinness at birth and the common 16189 variant in mitochondrial DNA, which is maternally transmitted [7]. Another possible mechanism to explain maternal inheritance of fetal growth restraint is common variation in genes that are imprinted and exclusively maternally expressed.
Many of the genes known to be imprinted and paternally or maternally expressed influence fetal growth [8]. An imprinted region on chromosome 11p15.5 in humans regulates the major fetal growth factor "insulin-like growth factor-II" (IGF-II). The IGF2 gene is imprinted and paternally expressed in both the mouse and human fetus [9,10]. The mechanisms of imprinting and expression of Igf2 have been thoroughly studied in the mouse. The gene is regulated by DNA elements close to Igf2, such as silencers [11] and activators [12], and particularly by enhancers located distal to the neighbouring maternally expressed gene H19 [13,14]. Access to these enhancers is restricted by an epigenetically controlled insulator upstream of H19 [15-18]. Finally, the maternally expressed H19 gene itself does not code for a protein, but the RNA has growth suppressing functions [19], potentially through inhibiting translation of IGF2 RNA [20].
We therefore aimed to demonstrate whether common genetic variation in human H19 may be associated with fetal growth restraint in a large representative birth cohort. We identified a common H19 2992 C>T SNP that was associated with offspring birth size, cord blood IGF-II levels, and infancy catch-up weight gain. These associations were more evident among first-born infants, and we propose that the common H19 2992T allele may confer reduced fetal growth restraint.
Results
Of the ten SNPs that we identified in the H19 region, we selected three SNPs (2992 C/T; 1737 A/G; and 3238 A/G) that marked different common H19 haplotypes for genotyping in ALSPAC offspring DNA samples for genetic association studies.
Associations with offspring H19 genotypes
Offspring H19 2992 SNP genotypes showed significant associations with size at birth (Table 1). However, neither H19 1737 nor 3231 genotypes were associated with birth size (data not shown).
H19 2992, 1737 and 3231 genotypes were in significant linkage disequilibrium, and haplotype analysis revealed only 7 genotype combinations with frequency >1%. Only those two haplotypes that included a 2992T allele had birthweights that were above the average [see Additional file 1].
Table 1 Birth size and cord IGF-II levels by offspring H19 2992 genotype (ALSPAC cohort)
CC CT TT PA PD
(n = 1075) (n = 518) (n = 54)
Weight (gm) 3468 ± 488 3522 ± 486 3524 ± 587 0.03 0.02
Length (cm) 50.7 ± 2.0 50.9 ± 2.0 50.9 ± 2.1 0.08 0.06
Head circ. (cm) 34.9 ± 1.2 34.9 ± 1.3 35.0 ± 1.6 0.3 0.4
(n = 226) (n = 110) (n = 13)
Cord IGF-II (ng/dl) 275 ± 78 277 ± 79 337 ± 91 0.1 0.3
Mean ± SD, adjusted for sex, gestation and parity.
PA = P value for additive genetic model (per T allele)
PD = P value for dominant genetic model (CC versus T/*)
Associations with mother's H19 2992 genotype
The H19 2992 SNP was subsequently genotyped in DNA samples collected from mothers and fathers, for which the family relationships had been validated by genotyping. H19 2992 genotype in mothers, but not in fathers, showed associations with size at birth, and also with cord blood IGF-II levels (Table 2). Mother's H19 2992 genotype associations were independent of mother's pre-pregnancy weight and height (additive models: birthweight: P = 0.001; cord IGF-II levels: P = 0.0005). The association between mother's H19 2992 genotype and IGF-II levels appeared to vary with mother's parity (birth order), being evident only in first pregnancies (Figure 1), although formal test for interaction did not reach significance (P = 0.06). When both mother's and offspring's genotypes were included in a multivariate analysis, birthweight was significantly associated with both mother's H19 2992 genotype (P < 0.05) and also offspring's H19 2992 genotype (P = 0.01), suggesting that there were separate effects of mother's and offspring's genotype on birthweight.
Table 2 Birth size by mothers or fathers H19 2992 genotype (ALSPAC cohort)
Mothers genotype CC CT TT PA PD
(n = 550) (n = 242) (n = 30)
Weight (gm) 3457 ± 474 3543 ± 508 3595 ± 485 0.003 0.003
Length (cm) 50.7 ± 2.0 51.0 ± 1.9 51.1 ± 1.8 0.04 0.04
Head circ. (cm) 34.8 ± 1.2 35.1 ± 1.3 34.9 ± 1.4 0.07 0.03
(n = 164) (n = 61) (n = 9)
Cord IGF-II (ng/dl) 266 ± 72 312 ± 99 312 ± 92 0.0003 0.0001
Fathers genotype CC CT TT PA PD
(n = 442) (n = 201) (n = 18)
Weight (gm) 3492 ± 491 3474 ± 472 3537 ± 519 0.9 0.7
Length (cm) 50.8 ± 1.9 50.8 ± 1.9 51.1 ± 1.7 0.8 0.9
Head circ. (cm) 34.9 ± 1.2 34.9 ± 1.2 35.0 ± 1.2 0.9 0.7
(n = 76) (n = 37) (n = 3)
Cord IGF-II (ng/dl) 270 ± 75 290 ± 93 245 ± 96 0.5 1.0
Mean ± SD, adjusted for sex, gestation and parity.
PA = P value for additive genetic model (per T allele)
PD = P value for dominant genetic model (CC versus T/*)
Figure 1 Birthweight SD score (A) and cord blood IGF-II levels at birth (B) in the ALSPAC cohort, by mother's H19 2992 genotype, and stratified by birth order ("Primip" = mother's first child; "Non-primip" = second or subsequent child). Mean ± 95% CI. First-born infants had lower birthweights than infants of subsequent pregnancies (Ref. 5). Associations with mother's genotype (CC vs. T* [CT or TT]) were only seen in first pregnancies.
Mother's untransmitted H19 2992 allele effects
Effects of mother's genotype, independent of transmission to the offspring, were examined by looking at association with mother's H19 2992 CC versus CT genotype within CC genotype offspring. Mother's untransmitted T versus C allele was associated with larger birthweight (P = 0.04) and higher cord blood IGF-II levels (P = 0.002; Table 3). As suggested by the overall associations with mother's genotype (in Figure 1), the effect of mother's untransmitted allele on IGF-II levels was clearer in first pregnancies (P = 0.0002) than in subsequent pregnancies (P = 0.36; P-interaction between mother's untransmitted allele and birth order = 0.03).
Table 3 Birthweight and IGF-II levels by maternal H19 2992 genotype, within CC genotype offspring (i.e. effect of mother's untransmitted allele), (ALSPAC cohort)
Mother's Genotype ANOVA
CC CT
Birthweight (gm) 3431 ± 474 3533 ± 505 P = 0.04
(n = 407) (n = 96)
Cord IGF-II (ng/dl) 264 ± 72 320 ± 108 P = 0.002
(n = 124) (n = 22)
Mean ± SD, adjusted for sex, gestation and parity.
Those results indicate that the H19 2992 SNP associations may depend on mother's genotype. However, offspring H19 2992 SNP associations were also seen independent of mother's genotype; among CC genotype mothers, offspring CC vs. CT genotype was also associated with birthweight (P = 0.007, Table 4).
Table 4 Birthweight and IGF-II levels by offspring H19 2992 genotype, within CC genotype mother's (i.e. effect of fetal genotype), (ALSPAC cohort)
Offspring Genotype ANOVA
CC CT
Birthweight (gm) 3425 ± 490 3535 ± 495 P = 0.007
(n = 408) (n = 126)
Cord IGF-II (ng/dl) 262 ± 80 266 ± 98 P = 0.8
(n = 125) (n = 36)
Mean ± SD, adjusted for sex, gestation and parity.
H19 2992 allele transmission and parent-of-origin effects
Among the validated parent-offspring trios, comparison of allele transmission from heterozygous parents to their offspring showed that T versus C allele transmission was not associated with birthweight (P = 0.22, number of observations = 204), nor with cord blood IGF-II levels (P = 0.52, number of observations = 38). Nor was there any difference between paternal and maternal T allele transmission (parent-of-origin effects) on birthweight (P = 0.92, number of observations = 87), or on IGF-II levels (P = 0.81, number of observations = 15). Data on mother's parents' genotypes were unavailable.
Genotype associations with postnatal weight gain
1,449 ALSPAC children were followed-up to age 3 years. Neither offspring nor parental H19 genotypes showed association with offspring body size at age 3 years (not shown). The apparent loss of T allele-related size advantage between birth and age 3 years was explained by slower rates of weight gain during this period (Table 5). As with birthweight and IGF-II levels, the H19 2992 genotype associations with postnatal weight gain were more apparent in offspring of first pregnancies (Figure 2).
Table 5 Size at 3 years, and growth from birth, by offspring H19 2992 genotype, (ALSPAC cohort)
Offspring genotype CC CT TT PA PD
(n = 946) (n = 462) (n = 41)
Weight at 3 y (kg) 15.0 ± 1.7 14.9 ± 1.7 14.8 ± 1.7 0.4 0.4
Height at 3 y (cm) 95.6 ± 3.6 95.3 ± 3.5 95.2 ± 3.7 0.2 0.2
Change in Weight SDS 0–3 y 0.01 ± 1.17 -0.13 ± 1.07 -0.10 ± 1.29 0.01 0.02
Change in Height SDS 0–3 y 0.06 ± 1.12 -0.10 ± 1.02 -0.17 ± 1.26 0.01 0.01
Mean ± SD, adjusted for sex and age
PA = P value for additive genetic model (per T allele)
PD = P value for dominant genetic model (CC versus T/*)
Figure 2 Postnatal weight gain (change in weight SD score 0–3 years) in the ALSPAC cohort, by offspring's (A) or mother's (B) H19 2992 genotype, and stratified by birth order ("Primip" = mother's first child; "Non-primip" = second or subsequent child). Mean ± 95% CI. The overall association between weight gain and offspring genotype (CC vs. T* [CT or TT], P = 0.01; Table 5) was only apparent in first-born offspring.
Replication of mother's H19 2992 genotype associations in the Cambridge cohort
In the independent Cambridge birth cohort (n = 646 births analysed), we obtained additional support for the association between mother's H19 2992T allele and larger birthweight in offspring of first pregnancies (P = 0.04, Figure 3A). Mother's glucose levels one hour post oral glucose at 27 to 32 weeks gestation (mean ± SD: 5.5 ± 1.2 mmol/l) were positively related to offspring birthweight (r = 0.14, P = 0.01), and mother's H19 2992 T/* genotype was also associated with higher mother's glucose levels during first pregnancies (P = 0.01, Figure 3B).
Figure 3 Birthweight SD score (A) and mother's glucose levels one hour post-oral glucose load at 27 to 32 weeks gestation (B) in the Cambridge cohort, by mother's H19 2992 genotype, and stratified by birth order ("Primip" = mother's first child; "Non-primip" = second or subsequent child). Mean ± 95% CI. Associations with mother's genotype (CC vs. T* [CT or TT]) were only seen in first pregnancies.
Discussion
In two separate contemporary birth cohorts we have identified associations between a common H19 2992 genotype and birthweight. This SNP is located in H19 exon 5 [see Additional file 2], however its effect on mRNA structure or function is unknown and H19 has no protein product. An insulator upstream of H19 [15-18] and enhancers downstream [13,14] regulate imprinting of IGF2, which encodes a major fetal growth factor [21]. In addition H19 RNA may suppress IGF2 expression in trans [20] and has tumour suppressor functions in cell transfection studies [19]. Consistent with the birthweight association, H19 2992 genotype was also associated with cord blood IGF-II protein levels at birth.
Birthweight associations were found with both mother's and offspring's H19 2992 genotypes, and it is not clear whether our findings represent direct effects of mother's genotype, offspring genotype, or both. As in other population studies, the low number of available DNA samples and genotypes from fathers contributed to relatively lower statistical power to formally detect effects of H19 2992 allele transmission. However, we did observe significant birthweight and cord blood IGF-II associations with mother's untransmitted allele, suggesting direct effects of mother's genotype. While in mice H19 and IGF2 are not expressed postnatally [22], in humans IGF2 expression continues into adulthood [23]. In most adult tissues IGF2 expression remains monoallelic as in the fetus, however biallelic expression is observed in adult human liver [23]. It is possible that mother's H19 2992 genotype may regulate IGF2 expression in maternally-derived placental tissues. Alternatively, H19 regulation of maternal IGF2 expression could influence mother's glucose metabolism [24] and thereby influence placental glucose transfer and fetal growth [25]. Offspring H19 2992 genotype was associated with birthweight, but not with IGF-II levels. It is possible that this discrepancy could relate to reduced power due to the smaller number of IGF-II samples available, or alternatively it could be possibly explained by a non-IGF-II mediated effect of H19 on birthweight. The association between birthweight and IGF-II levels in cord blood at birth is weak, and it is possible that the effects of IGF-II on fetal growth may be specific to certain tissues or developmental stages in fetal life [26].
The H19 2992 genotype associations with birthweight and IGF-II were independent of mother's body size and were more apparent in mother's first pregnancies. First-born offspring are more growth-restrained in utero than subsequent children, as evident by smaller birth size and compensatory rapid postnatal weight gain within the first 2 to 3 years [5]. The mechanism of growth restraint seen in first pregnancies is unknown, but could reflect a maternally inherited genetic trait with subsequent relaxation of restraint in later pregnancies [27]. We propose that the common H19 2992T allele, either in the mother or offspring, confers reduced fetal growth restraint particularly among first-born infants as indicated by larger birth size and less postnatal compensatory catch-up weight gain.
In subsequent pregnancies and in larger babies, a mendelian pattern of inheritance of birthweight is observed [27]. It was in these second and subsequent pregnancies that we saw association between size at birth and common allele class at the insulin gene minisatellite (INS VNTR) [28], which also regulates IGF2 expression [29]. Offspring INS VNTR class III alleles were associated with larger head size and higher IGF-II levels at birth, and this effect was masked in growth-restricted first pregnancies [28]. Thus, there may be important interaction between maternal parity and imprinted and non-imprinted genes involved in the regulation of IGF-II expression and fetal growth.
Altogether we identified ten SNPs in H19. Given that there is no confirmed function for the H19 transcript in vivo [30], it is difficult to assess the effect of the H19 2992 SNP. It is possible that this variant: (a) may alter the RNA itself, and might influence the putative growth suppressing function of the H19 RNA; (b) may influence levels of H19 transcripts, leading to an effect as above; (c) may be in linkage with an unidentified regulatory element in H19 acting on IGF2 in cis; (d) may be in linkage with variants in known cis regulatory elements for IGF2, such as the insulator or the enhancers. Further work, including systematic genotyping of other SNPs in the extended region and functional studies, will be required to identify the causal variant(s) and, hence, possible mechanisms by which they may influence birthweight and cord blood IGF-II levels. The identification of separate effects of H19 on fetal growth, by acting both within the mother and also within the fetus, could be confirmed by studying a larger population to provide greater power to detect an interaction between mother's and offspring's genotypes, and by exploration of grand-parental genotype and allele transmission to the mother. Intriguingly, while we confirmed that the offspring 2992 genotype (CC vs. CT) association with birthweight was independent of mother's genotype (within CC mothers; Table 4), that finding also raises the surprising suggestion of a paternal H19 allele transmission effect, however, we did not detect any formal association with paternal transmission. An alternative explanation would be that the H19 2992 is in LD with another paternally transmitted determinant of birthweight. Further studies with larger numbers of complete informative trios will be needed to clarify this issue.
Conclusion
In conclusion, the common H19 2992T allele, in the mother or offspring or both, may confer reduced fetal growth restraint, as indicated by associations with larger offspring birth size, higher cord blood IGF-II levels, and lower compensatory early postnatal catch-up weight gain. These observations may have implications for the early origins of adult disease hypothesis [31]. Following the original observations of association between small size at birth and increased later risks for cardiovascular disease and diabetes [32], recent studies have highlighted that these risks are strongest in subjects with the combination of fetal growth restraint followed by rapid postnatal catch-up growth [33]. Common factors that lead to fetal growth restraint in humans include mother's smoking and 1st pregnancies [5]. Common variants in exclusively maternally-expressed genes (such as H19) or maternally transmitted genes (such as in mitochondrial DNA) [7] could also contribute to fetal growth restraint and compensatory postnatal catch-up growth.
Methods
Subjects
The Avon Longitudinal Study of Parents and Children (ALSPAC) birth cohort numbers 14,185 children comprising over 80% of all births in three Bristol based health authorities between April 1991 and December 1992 [34]. Our studies are confined to a 10% sub-cohort ("Children in Focus") [4] and a second (control) cohort for which there were detailed data on size at birth and subsequent childhood growth. In these families, offspring DNA was extracted from cord blood samples collected at birth, mothers DNA from venous blood samples in pregnancy, and fathers DNA from mouthswab samples. H19 genotypes were analysed in available DNA samples from 1,696 children, 822 mothers and 661 fathers. Cord blood samples were collected at birth in 338 of these infants and assayed for IGF-II levels as previously reported [26]; these infants had larger birthweight (mean = 3546 g) than infants for whom cord blood was unavailable (3467 g), but were no different in H19 +2992 genotype (chi-square: P = 0.7)
The Cambridge (Wellbeing) birth cohort was studied to provide independent confirmation of mother's H19 genotype associations. This cohort was recruited from full-term, singleton deliveries at the Rosie Maternity Unit, Cambridge, 1999 to 2000. Routinely collected clinical data were available on offspring birthweight and mother's whole blood glucose levels measured at one hour following a standard dose oral glucose load at 27 to 32 weeks of gestation. Mother's were selected of full-term, singleton infants. Exclusion criteria were mother's hypertension and treatment for diabetes during pregnancy. We sought permission from the mother's General Practitioner to approach the mother to collect her DNA sample by postal mouthswab kits and questionnaire data on her other children. Mother's DNA was extracted and genotyped for the H19 2992 genotype only. Local ethical committee approval and signed informed consent from mothers were obtained.
SNP identification
In an independent cohort of 20 subjects, ten SNPs were identified in the H19 region using denaturing high performance liquid chromatography (Wave™) and were characterised by sequencing of PCR products (SNPs submitted to dbSNP). Confirmation and assessment of allele frequency was obtained by genotyping a second independent cohort of 100–200 subjects. We selected three H19 SNPs as these SNPs marked different haplotypes (2992 C/T [rs217727], 1737 A/G [rs2067051], and 3238 A/G [rs2839703]; the SNP numbers relate to their position in accession number M32053). We genotyped these three SNPs in offspring DNA samples from the ALSPAC cohort for genetic association studies. The H19 2992 SNP was subsequently genotyped in DNA samples from ALSPAC parents, and from mothers in the Cambridge cohort.
H19 2992 C/T genotyping
Genomic DNA (40 ng) was amplified in a reaction mixture containing 1x ammonium reaction buffer, 50 μM each dNTP, 500 μM magnesium, 225 ng each primer (forward: 5'-aaagacaccatcggaacagc-3', reverse: 5'-agcttccagact aggcgagg-3'), 10% (v/v) dimethylsulphoxide and 0.6 units DNA polymerase (Bioline, London, UK) in a 45 μL reaction volume. After an initial 5 min. incubation at 94°C, twenty amplification cycles of 94°C (65 sec.), 58°C (65 sec., dropping 0.5°C per cycle) and 72°C (140 sec.) were performed. This was followed by 14 cycles of 94°C (65 sec.), 48°C (65 sec.) and 72°C (140 sec.) and a final incubation of 72°C for 10 min. After overnight digestion with Fnu4HI (2 units per reaction; New England Biolabs, Hitchin, UK) at 37°C, PCR products were separated by agarose gel electrophoresis. This gave a 351 bp band for the 'T' allele and a 304 bp band for the 'C' allele. The observed genotype frequencies: CC: 65.3%, CT: 31.5%, TT: 3.3% (Table 1) were in Hardy Weinberg equilibrium (chi-square test: P = 0.99). The H19 2992 SNP was also genotyped in the parents DNA samples.
H19 1737 A/G genotyping
The reaction mix for the H19 1737 genotyping was the same as for the 2992 SNP, with the exception of the oligonucleotide primer sequences (forward: 5'-aaggtgacatcttctcgggg-3'; reverse: 5'-tgagagctcattcactccgc-3'). The amplification conditions were the same as those for the 2992 SNP except that the annealing temperatures were 4°C higher throughout. Overnight digestion with Tsp45I (2 units per reaction; New England Biolabs) produced 474 and 131 bp bands for the 'A' allele and 329, 145 and 131 bp fragments for the 'G' allele. The observed genotype frequencies (AA: 23.1%, AG: 52.6%, GG: 24.4%) were in Hardy Weinberg equilibrium (chi-square test: P = 0.5).
H19 3238 A/G genotyping
The reaction mix for the H19 3238 genotyping was the same as for the 2992 SNP, with the exception of the oligonucleotide primer sequences (forward: 5'-aaagacaccatcggaacagc-3'; reverse: 5'-agcttccagactaggcgagg-3'). Amplification conditions were the same as those for the 2992 SNP except that the annealing temperatures were 4°C higher throughout. Overnight digestion with DdeI (2 units per reaction; New England Biolabs) produced 137 and 44 bp bands and multiple small bands (116, 76, 60 and 26 bp) for the 'A' allele and a 181 bp band along with the same multiple small bands for the 'G' allele. The observed genotype frequencies (AA: 38.0%, AG: 48.5%, GG: 13.5%) were in Hardy Weinberg equilibrium (chi-square test: P = 0.8).
Microsatellite validation of parental DNA samples
Parental DNA samples were validated using amplification of markers D16S539, D7S820, D13S317 and D5S818 (Geneprint Multiplex-GammaSTR kit, Promega, Southampton, UK) on an ABI 3700 instrument as previously described [28]. We had DNA samples from 597 complete and validated parent-offspring trios.
Calculations and statistics
Postnatal weight and length gains were calculated by transforming weight and length data into SD scores allowing for variation relating to (gestational) age and gender, using the formula [SD score = individual value – group mean / group SD] [35], and then calculating changes in weight and length SD score between birth to 3 years.
Associations between offspring or mother's genotype and size at birth were determined by ANOVA. Multivariate analysis (general linear models) was used to identify independent effects of mother's and offspring genotypes. Associations with H19 2992 allele transmission were assessed by TDT using both quantitative and qualitative methods. Parent-of-origin effects were sought using quantitative and qualitative methods, allowing for possible confounding by the effects of non-transmission of mother's alleles [36,37].
Abbreviations
ALSPAC – Avon Longitudinal Study of Parents and Children
IGF – insulin-like growth factor
Primip – primiparous
SNP – single nucleotide polymorphism
Authors' contributions
CP performed genotyping and drafted the manuscript. KO contributed to the study design, and co-ordination, performed statistical analyses and drafted the manuscript. BB performed SNP identification, and contributed to the study design. DW performed Cambridge cohort recruitment, DNA preparation and genotyping. HC performed statistical analyses. SR supervised preparation of ALSPAC DNA samples and validation of genotyping results. MR participated in the design and co-ordination of the ALSPAC study, and supervised DNA management. WR participated in the conception and design of the study. JT supervised SNP identification, genotyping and data analyses. DD was responsible for the study conception, and contributed to the study coordination, analyses and drafting the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Birthweight SD score (adjusted for sex and gestational age) by combination of H19 2992/1737/3238 genotypes (for each SNP, 1 = more common homozygote, 2 = heterozygote, 3 = less common homozygote).
Click here for file
Additional File 2
Schematic map of H19 indicating exons, SNPs published on dbSNP and also published SNPs identified by the Juvenile Diabetes Research Foundation/Wellcome Trust Diabetes and Inflammation Laboratory (DIL) T1Dbase . H19 2992 is indicated by the labels rs217727 and DIL1049.
Click here for file
Acknowledgements
We are extremely grateful to all the children and parents who took part in the ALSPAC and Cambridge cohorts and to the study teams, nurses and midwives who supported these studies. The whole ALSPAC study team comprises interviewers, computer technicians, laboratory technicians, clerical workers, research scientists, volunteers, managers and also the staff of the ALSPAC research clinics. We also thank Miguel Constancia for helpful comments.
ALSPAC is supported by the Medical Research Council (MRC), the Wellcome Trust, the Department of Health, the Department of the Environment and many others. The Cambridge cohort is supported by Wellbeing, UK. KKO was supported by a MRC Clinical Training Fellowship. HJC, JAT and DBD are supported by the Wellcome Trust and the Juvenile Diabetes Research Foundation. WR is supported by the Biotechnology and Biological Sciences Research Council and the European Union.
==== Refs
Ounsted M Ounsted C Maternal regulation of intrauterine growth Nature 1966 111 995 997
Walton A Hammond J The maternal effects on growth and conformation in Shire horse-Shetland pony crosses Proceedings of the Royal Society of London 1938 125 311 335
Joubert DM Hammond J A cross breeding experiment with cattle, with special reference to the maternal effect in South Devon-Dexter crosses. Journal of Agricultural Science 1958 51 325 341
Ong KK Ahmed ML Emmett PM Preece MA Dunger DB the-ALSPAC-Study-Team Association between postnatal catch-up growth and obesity in childhood: prospective cohort study BMJ 2000 320 967 971 10753147 10.1136/bmj.320.7240.967
Ong KK Preece MA Emmett PM Ahmed ML Dunger DB Size at birth and early childhood growth in relation to maternal smoking, parity and infant breast-feeding: longitudinal birth cohort study and analysis Pediatric Research 2002 52 863 867 12438662 10.1203/01.PDR.0000036602.81878.6D
Ounsted M Scott A Ounsted C Transmission through the female line of a mechanism constraining human fetal growth Annals of Human Biology 1986 13 143 151 3707043
Casteels K Ong KK Phillips DI Bednarz A Bendall H Woods KA Sherriff A ALSPAC_Study_Team Golding J Pembrey ME Poulton J Dunger DB Mitochondrial 16189 variant, thinness at birth and type 2 diabetes Lancet 1999 353 1499 1500 10232327 10.1016/S0140-6736(98)05817-6
Reik W Walter J Genomic imprinting: parental influence on the genome Nature Reviews Genetics 2001 2 21 32 11253064 10.1038/35047554
Giannoukakis N Deal C Paquette J Goodyer CG Polychronakos C Parental genomic imprinting of the human IGF2 gene Nature Genetics 1993 4 98 101 8099843 10.1038/ng0593-98
Ohlsson R Nystrom A Pfeifer-Ohlsson S Tohonen V Hedborg F Schofield P Flam F Ekstrom TJ IGF2 is parentally imprinted during human embryogenesis and in the Beckwith-Wiedemann syndrome Nature Genetics 1993 4 94 97 8513333 10.1038/ng0593-94
Constancia M Dean W Lopes S Moore T Kelsey G Reik W Deletion of a silencer element in Igf2 results in loss of imprinting independent of H19 Nature Genetics 2000 26 203 206 11017078 10.1038/79930
Murrell A Heeson S Bowden L Constancia M Dean W Kelsey G Reik W An intragenic methylated region in the imprinted Igf2 gene augments transcription EMBO Reports 2001 2 1101 1106 11743023 10.1093/embo-reports/kve248
Leighton PA Saam JR Ingram RS Stewart CL Tilghman SM An enhancer deletion affects both H19 and Igf2 expression Genes & Development 1995 9 2079 2089 7544754
Davies K Bowden L Smith P Dean W Hill D Furuumi H Sasaki H Cattanach B Reik W Disruption of mesodermal enhancers for Igf2 in the minute mutant Development 2002 129 1657 1668 11923202
Murrell A Heeson S Reik W Interaction between differentially methylated regions partitions the imprinted genes Igf2 and H19 into parent-specific chromatin loops Nature Genetics 2004 36 889 893 15273689 10.1038/ng1402
Thorvaldsen JL Duran KL Bartolomei MS Deletion of the H19 differentially methylated domain results in loss of imprinted expression of H19 and Igf2 Genes & Development 1998 12 3693 3702 9851976
Hark AT Schoenherr CJ Katz DJ Ingram RS Levorse JM Tilghman SM CTCF mediates methylation-sensitive enhancer-blocking activity at the H19/Igf2 locus Nature 2000 405 486 489 10839547 10.1038/35013106
Bell AC Felsenfeld G Methylation of a CTCF-dependent boundary controls imprinted expression of the Igf2 gene Nature 2000 405 482 485 10839546 10.1038/35013100
Hao Y Crenshaw T Moulton T Newcomb E Tycko B Tumour-suppressor activity of H19 RNA Nature 1993 365 764 767 7692308 10.1038/365764a0
Li YM Franklin G Cui HM Svensson K He XB Adam G Ohlsson R Pfeifer S The H19 transcript is associated with polysomes and may regulate IGF2 expression in trans Journal of Biological Chemistry 1998 273 28247 28252 9774446 10.1074/jbc.273.43.28247
Leighton PA Ingram RS Eggenschwiler J Efstratiadis A Tilghman SM Disruption of imprinting caused by deletion of the H19 gene region in mice Nature 1995 375 34 39 7536897 10.1038/375034a0
Stylianopoulou F Efstratiadis A Herbert J Pintar J Pattern of the insulin-like growth factor II gene expression during rat embryogenesis Development 1988 103 497 506 3246220
Wu HK Squire JA Song Q Weksberg R Promoter-dependent tissue-specific expressive nature of imprinting gene, insulin-like growth factor II, in human tissues Biochemical and Biophysical Research Communications 1997 233 221 226 9144427 10.1006/bbrc.1997.6431
Zapf J Role of insulin-like growth factor II and IGF binding proteins in extrapancreatic tumor hypoglycemia Hormone Research 1994 42 20 26 7525443
Catalano PM Thomas AJ Huston LP Fung CM Effect of maternal metabolism on fetal growth and body composition Diabetes Care 1998 21 Suppl 2 B85 90 9704233
Ong KK Kratzsch J Kiess W the-ALSPAC-Study-Team Costello M Scott CD Dunger DB Size at birth and cord blood levels of insulin, insulin-like growth factor I (IGF-I), IGF-II, IGF-binding protein-1 (IGFBP-1), IGFBP-3, and the soluble IGF-II/mannose-6-phosphate receptor in term human infants. Journal of Clinical Endocrinology & Metabolism 2000 85 4266 4269 11095465 10.1210/jc.85.11.4266
Ounsted M Scott A Moar VA Constrained and unconstrained fetal growth: associations with some biological and pathological factors Annals of Human Biology 1988 15 119 129 3281574
Ong KK Petry CJ Barratt BJ Ring S Cordell HJ Wingate DL Pembrey ME Todd JA Dunger DB Maternal-fetal interactions and birth order influence insulin variable number of tandem repeats allele class associations with head size at birth and childhood weight gain Diabetes 2004 53 1128 1133 15047631
Paquette J Giannoukakis N Polychronakos C Vafiadis P Deal C The INS 5' variable number of tandem repeats is associated with IGF2 expression in humans Journal of Biological Chemistry 1998 273 14158 14164 9603916 10.1074/jbc.273.23.14158
Jones BK Levorse JM Tilghman SM Igf2 imprinting does not require its own DNA methylation or H19 RNA Genes & Development 1998 12 2200 2207 9679064
Barker DJ Fetal growth and adult disease B J Obstet Gynaecol 1992 99 275 276
Hales CN Barker DJ Clark PM Cox LJ Fall C Osmond C Winter PD Fetal and infant growth and impaired glucose tolerance at age 64 BMJ 1991 303 1019 1022 1954451
Eriksson JG Forsen T Tuomilehto J Winter PD Osmond C Barker DJ Catch-up growth in childhood and death from coronary heart disease: longitudinal study BMJ 1999 318 427 431 9974455
Golding J Pembrey ME Jones R ALSPAC--the Avon Longitudinal Study of Parents and Children. I. Study methodology Paediatric and Perinatal Epidemiology 2001 15 74 87 11237119 10.1046/j.1365-3016.2001.00325.x
Cole TJ Do growth chart centiles need a face lift? BMJ 1994 308 641 642 8148716
Whittaker JC Gharani N Hindmarsh P McCarthy MI Estimation and testing of parent-of-origin effects for quantitative traits American Journal of Human Genetics 2003 72 1035 1039 12644965 10.1086/374382
Gauderman WJ Candidate gene association analysis for a quantitative trait, using parent-offspring trios Genetic Epidemiology 2003 25 327 338 14639702 10.1002/gepi.10262
| 15885138 | PMC1140752 | CC BY | 2021-01-04 16:38:18 | no | BMC Genet. 2005 May 10; 6:22 | utf-8 | BMC Genet | 2,005 | 10.1186/1471-2156-6-22 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-631587635510.1186/1471-2164-6-63Research ArticleThree microarray platforms: an analysis of their concordance in profiling gene expression Petersen David [email protected] GVR [email protected] Joel [email protected] Joanne [email protected] Jonathon [email protected] Chang Hee [email protected] David [email protected] Lisa [email protected] Jing [email protected] Raj [email protected] Lou [email protected] John [email protected] J Carl [email protected] Jeffrey [email protected] Ernest S [email protected] Advanced Technology Center, Center for Cancer Research, National Cancer Institute, Gaithersburg, MD 20877 USA2 Laboratory of Molecular Technology, SAIC Frederick, Frederick, MD 21701 USA3 Center for Biologics Evaluations & Research, Food & Drug Administration, Bethesda, MD 20892 USA4 Center for Cancer Research, National Cancer Institute, Bethesda, MD 20892 USA2005 5 5 2005 6 63 63 22 10 2004 5 5 2005 Copyright © 2005 Petersen et al; licensee BioMed Central Ltd.2005Petersen 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
Microarrays for the analysis of gene expression are of three different types: short oligonucleotide (25–30 base), long oligonucleotide (50–80 base), and cDNA (highly variable in length). The short oligonucleotide and cDNA arrays have been the mainstay of expression analysis to date, but long oligonucleotide platforms are gaining in popularity and will probably replace cDNA arrays. As part of a validation study for the long oligonucleotide arrays, we compared and contrasted expression profiles from the three formats, testing RNA from six different cell lines against a universal reference standard.
Results
The three platforms had 6430 genes in common. In general, correlation of gene expression levels across the platforms was good when defined by concordance in the direction of expression difference (upregulation or downregulation), scatter plot analysis, principal component analysis, cell line correlation or quantitative RT-PCR. The overall correlations (r values) between platforms were in the range 0.7 to 0.8, as determined by analysis of scatter plots. When concordance was measured for expression ratios significant at p-values of <0.05 and at expression threshold levels of 1.5 and 2-fold, the agreement among the platforms was very high, ranging from 93% to 100%.
Conclusion
Our results indicate that the long oligonucleotide platform is highly suitable for expression analysis and compares favorably with the cDNA and short oligonucleotide varieties. All three platforms can give similar and reproducible results if the criterion is the direction of change in gene expression and minimal emphasis is placed on the magnitude of change.
==== Body
Background
Completion of the human genome sequence has made it possible to study expression of the entire complement of 20,000–30,000 genes in a single assay. The two most common array platforms are based on collections of cDNA clones [1] or short (25 base) oligonucleotides synthesized in situ by photolithographic methods (i.e., by Affymetrix, Inc.) [2]. Partly because they are easy to use, microarrays are the most extensively used technology for studying gene expression on a global scale [3,4]. Thousands of expression studies employ one or the other microarray platform, but comparison of results between platforms has been difficult because of inherent differences in the array technologies. The situation became more complex as investigators began using long oligonucleotide arrays for expression profiling [5-9].
Because long oligonucleotide arrays for expression profiling are relatively new, we wished to validate them in relation to the cDNA and short oligonucleotide platforms, both of which have been used extensively in our laboratories over a number of years. The three platforms were evaluated using RNAs isolated from six cell lines and tested against a universal reference RNA. Sufficient RNA was isolated in a single harvest to supply labeling template for all experiments, so variability of RNA isolation was not an issue. However, no attempt was made to eliminate other sources of variation such as differences between lots of fluorescent dye label, microarray batch, operator, etc. We conducted these experiments under "normal" laboratory conditions so that one would not need to go to extreme lengths to reproduce the results. In almost all cases, results from the three platforms correlated reasonably well with each other. The Pearson correlation coefficients (r) ranged from 0.7 to 0.8. Because of different labeling methods and analysis algorithms, comparison of the cDNA and long oligonucleotide platforms with the short oligonucleotide system was not as straightforward, but in general all of the platforms were in reasonable agreement.
Results
This study was carried out to compare cDNA (Incyte), long oligonucleotide (Operon 70-mer), and short 25-mer (Affymetrix) array platforms, with the goal of qualifying the 70-mer arrays for general use at the National Cancer Institute. More specifically, we compared the Incyte Unigem2 set of human cDNAs (~9900 genes), the Operon human Version 2.0 set of long oligonucleotides (~21,329 genes), and Affymetrix HG-U133A arrays (~22,200 genes). RNA preparations from cell lines MCF10A, LNCaP, Jurkat, L428, SUDHL6, and OCI-Ly3 were used as probe templates, and the expression of each gene was compared directly with that of the same gene in the Human Universal Reference (HUR) RNA from Stratagene.
Genes in common across platforms
Only genes common to all three platforms were used in the comparison. Genes were matched by UniGene Cluster (UniGene Build #161), and unique cluster memberships were determined for each array type, as listed in Table 1 and enumerated in the Venn diagram in Figure 1. The intersection for the three platforms consisted of 6430 UniGene clusters, and all analyses were carried out with all of these genes or a subset of them.
Table 1 Overlapping gene sets represented in 3 microarray platforms
Incyte UniGEM2 Hs-Operon V2 HG-U133A
Total features / probe sets: 9128 21522 22215
Distinct UniGene clusters: 8097 19179 13899
UniGem2 & Operon V2 UniGEM2 & HG-U133A Operon V2 & HG-U133A
Genes in common 7082 6593 12999
Genes in common in all arrays 6430 6430 6430
Figure 1 Venn diagram with number of genes present in each platform, genes in common between platforms, and genes in common among all three platforms.
Comparison of expression ratios
An estimate of the concordance of the platforms was provided by the percentage of genes that showed expression ratios "in the same direction." All matched gene sets that had expression ratios (relative to the Universal Standard) greater than 1.0, irrespective of magnitude, were considered to be in the same direction, and the same for expression ratios less than 1.0. As seen in Table 2, the agreement was reasonably good, ranging from 74% to 82%. For comparison, ~50% concordance would be expected by chance. Many of the discordant values were at what might be considered "trivial" levels because a large proportion of ratios are near unity. For example, approximately 48%, 28% and 13% of the genes from Incyte, Operon and Affymetrix respectively, exhibit expression ratios within the range of 1.1 to 0.9. Thus, a ratio difference of only 0.95 versus 1.05 is designated a mismatch, which normally would be considered very concordant in the biological sense.
Table 2 Proportion of genes expressed in same direction between platforms ignoring expression level. The numbers of genes in matched and mismatched directions are given for each cell line and each platform pair, and agreement is given in percentages.
Array Platform: Cell line: Direction-matched Direction-mismatched (ignoring expression level) Total number of genes used in calculation Percent agreement in direction
GEM2 & Operon V2 Jurkat 4597 1373 5970 77 %
L428 3535 1251 4786 74 %
SUDHL 4608 1525 6133 75 %
Ocily3 4094 1368 5462 75 %
LNCaP 4189 1333 5522 76 %
MCF10A 4469 1429 5898 76 %
Total 25492 8279 33771 75 %
GEM2 & HG-U133A Jurkat 3657 906 4563 80 %
L428 3203 918 4121 78 %
SUDHL 3541 809 4350 81 %
Ocily3 3826 984 4810 80%
LNCaP 3747 1031 4778 78 %
MCF10A 3584 950 4534 79 %
Total 21558 5598 27156 79 %
Operon V2 & HG-U133A Jurkat 3688 795 4483 82 %
L428 2909 738 3647 80 %
SUDHL 3520 776 4296 82 %
Ocily3 3509 947 4456 79 %
LNCaP 3607 860 4467 81 %
MCF10A 3511 898 4409 80 %
Total 20744 5014 25758 81 %
Statistical tests and thresholds on fold change are commonly used for the identification of altered gene expression by microarray assays. Hence, we estimated the statistical significance of differential expression of each RNA from the HUR by one sample Student-t test for a change from ratio of 1. Comparison between platforms using expression values significant at p < 0.05 shows concordance in direction of 93%–98% (Table 3 – no threshold). At the 1.5 and 2-fold threshold level, the agreement between platforms approaches 100% indicating that for reproducible data all three platforms are highly concordant in direction.
Table 3 Concordance between platforms using statistically significant expression ratios at p-value < 0.05 and at 1.5 and 2-fold threshold levels.1
No threshold 1.5-fold 2-fold
GEM2 vs. Operon GEM2 vs. U133A Operon vs. U133A GEM2 vs. Operon GEM2 vs. U133A Operon vs. U133A GEM2 vs. Operon GEM2 vs. U133A Operon vs. U133A
Number of comparisons2
Jurkat 1471 1678 1740 1096 1088 1220 588 614 749
L428 651 1394 750 580 1051 632 430 678 469
SUDHL 1400 1527 1707 1081 1177 1277 698 741 831
OCI-Ly3 1484 1834 1698 1285 1428 1333 814 900 903
LNCaP 1167 1170 2062 737 751 1252 369 404 696
MCF10A 1555 1429 1880 899 812 1118 405 409 559
Percent of concordance
Jurkat 97% 96% 98% 98% 99% 100% 99% 99% 100%
L428 96% 96% 98% 98% 98% 99% 99% 99% 100%
SUDHL 95% 97% 98% 97% 98% 99% 99% 99% 100%
OCI-Ly3 94% 96% 96% 95% 98% 98% 98% 99% 99%
LNCaP 93% 95% 96% 97% 98% 99% 99% 100% 100%
MCF10A 94% 95% 96% 98% 98% 99% 99% 99% 100%
Total number of comparisons
7728 9032 9837 5678 6307 6832 3304 3746 4207
Total number of concordances
7334 8650 9540 5509 6191 6755 3257 3719 4190
Percent agreement
94.9% 95.8% 97.0% 97.0% 98.2% 98.9% 98.6% 99.3% 99.6%
Percent of data used3
54.0% 57.5% 60.4% 39.7% 40.2% 42.0% 23.1% 23.9% 25.8%
1GEM2: Incyte GEM2 arrays; Operon: Operon V2 arrays; U133A: Affymetrix Human Genome U 133A GeneChip arrays.
2Comparisons between two platforms were done only when the expressions of both platforms met the p-value and threshold criteria.
3Fraction of common expressions between two platforms to the total number of expressions found at p < 0.05 in both platforms.
Correlation between platforms
The scatter plots shown in Figures 2a–c indicate graphically the correlation between the platforms for Jurkat RNA. Values of r for all platforms and all cell lines are given in Table 4. The values ranged from 0.7 to 0.8. Operon and Affymetrix were better correlated than were Operon and Incyte. The values were intermediate for the relationship between Incyte and Affymetrix. Figures 2d–f and Table 5 indicate that the correlation coefficients between replicates within the same platform are closer, as expected, when compared to the correlations between platforms. In general, however, there was reasonably good concordance among the platforms.
Table 4 Cross platform correlation coefficients. Pearson correlation coefficients are given for each platform pair and cell line, calculated over the genes common to all platforms. Values were obtained from scatter plots as shown in Figure 2a-c.
Platform Jurkat L428 SUDHL OCI-Ly3 LNCaP MCF10A Median
Incyte/Operon 0.727 0.707 0.708 0.724 0.710 0.708 0.709
Incyte/Affy 0.767 0.777 0.781 0.780 0.741 0.744 0.772
Operon/Affy 0.813 0.784 0.783 0.790 0.796 0.782 0.787
Table 5 Median correlation coefficients of replicates within same platform. Calculations and values obtained as in Table 4.
Platform Jurkat L428 SUDHL OCI-Ly3 LNCaP MCF10A
GEM2 0.859 0.919 0.902 0.873 0.827 0.845
Operon V2 0.916 0.805 0.849 0.828 0.951 0.928
HG-U133A 0.956 0.957 0.960 0.952 0.963 0.954
Figure 2 a-f. Scatter plot analysis to determine correlation coefficients between and within platforms using Jurkat RNA as an example. Correlations for all cell lines are given in Table 4. (a) Operon versus Incyte (b) Affymetrix versus Incyte (c) Affymetrix versus Operon (d) GEM2 versus GEM2 replicate correlation (e) Operon versus Operon (f) HG-U133A versus HG-U133A
Principal Component Analysis (PCA)
A projection on the first three principal components, which together explain 48.8 % (21%, 14%, and 13%) of the total variance, is shown in Figure 3. Close clustering of the cell samples is observed in this projection, indicating appreciable agreement among array platforms.
Figure 3 Principal Component Analysis (PCA) of the three microarray platforms and six cell lines using expression of the 3186 genes with signals above background.
"Correlation of Correlations" analysis
The global concordance of the three platforms across all of the cell lines was estimated by calculating the "correlation of correlations" coefficient [10,11]. As seen in Figure 4a–c, the correlations for the three platforms across all cells lines were quite good. The Pearson correlation of correlation coefficients was 0.965 between Operon and Incyte, 0.995 between Affymetrix and Incyte, and 0.956 between Operon and Affymetrix.
Figure 4 a-c. Correlation of correlations of platforms for all cell lines. Correlation values R for each pair of platforms are given in the figures. (a) Operon versus Incyte (b) Affymetrix versus Incyte (c) Affymetrix versus Operon.
Clustered Image Map (CIM) visualization and analysis [12]
We used hierarchical clustering to demonstrate graphically the relationships among platforms, among cell lines, and among genes. 909 genes expressed at two times background or more in at least two of the six cell lines were included in the analysis. The resulting CIM ("heat map") is shown in Figure 5a–b. All three platforms cluster together for all six cell lines, as one would wish to find, and almost all of the gene expression values show reasonable concordance. Only a few exceptions can be seen in the cluster shown in Figure 5b. Some of the mismatches may be due to simple errors in gene identification, rather than to the technologies of the platforms. The Incyte library is guaranteed by the manufacturer to be only 90% correct, and an unknown percentage of the Operon and Affymetrix oligonucleotides may have been designed on the basis of incorrect sequences in the public databases. Indeed, we found one oligonucleotide in the Operon set that was apparently designed from an EST sequence that has since been withdrawn from the UniGene database (see RT-PCR studies below). In any case, the concordance is quite high across all platforms with this method of analysis as well as with the others.
Figure 5 a-b. Clustered image maps showing patterns of expression relationship among genes, platforms, and cell lines. The axes were ordered by hierarchical clustering using an uncentered correlation and the average linkage algorithm for 909 genes expressed at a two-fold or greater level in at least two of the six cell lines. (a) Clustering of all 909 genes (b) A subcluster of 41 genes to show correct clustering and congruence of expression values. As indicated by the cluster trees, all three platforms gave essentially the same relationships among the six cell lines.
Quantitative real-time RT-PCR
In a pilot study with the three platforms, we compared and contrasted gene expression values for only the cell lines MCF10A and LNCaP. RT-PCR data for twelve genes are shown in Figure 6. Most of the values are in reasonable agreement except that there are differences in the magnitudes of the expression ratios. As found in other studies, the RT-PCR values are generally higher, probably because ratios are "flattened" with the microarray platforms. Affymetrix ratios are sometimes higher, but that may simply reflect the method of quantitation used in their analysis. The cDNA array ratios are generally lower than those from other platforms. Because the cDNA fragments are longer and double-stranded, they are more prone to non-specific hybridization and can cross-hybridize more easily to related sequences. These characteristics of the probes may result in higher background signal and concomitant reduction in dynamic range of the ratios. In general, we have found that the long oligonucleotide arrays have a larger dynamic range than the do the cDNA arrays.
Figure 6 Quantitative real-time RT-PCR analysis of 12 genes matched for direction of expression relative to the reference RNA for all three platforms. Log2 ratios are given in table below the graph. This example is a comparison between LnCaP and MCF-10A.
For further RT-PCR analysis, we chose a set of ten genes to test the accuracy of the three array platforms for all six cell lines. Those genes were chosen because there appeared to be a discrepancy among platforms in the direction of their ratios (i.e., whether they ratios were greater or less than unity). The results are shown in Figures 7a–f. Of special interest was gene ETR101, in which the Operon array was in disagreement across all cell lines. Further inquiry revealed that the sequence had been found to be incorrect and had been removed from the UniGene database. Since the oligonucleotide had been designed from the incorrect sequence, it is not surprising that it gave a different value. Other discrepancies may be due to similar sequence errors, as even the most up-to-date databases are still being corrected and modified. In the case of AGL, the RT-PCR assay is in disagreement in several cases with two out of three of the array platforms; it appears to demonstrate an upregulation of the mRNA, whereas the arrays, with the exception of Incyte, point to a downregulation. Although RT-PCR is supposed to be the "gold standard" for measuring gene expression, this result shows that caution is indicated in interpreting results with even the PCR technology.
Figure 7 a-f. Quantitative RT-PCR analysis of 10 mismatched genes in the six cells lines for all three platforms. (a) MCF10A, (b) LnCaP, (c) OCI-Ly3, (d) Jurkat, (e) SUDHL-6 and (f) L428.
Discussion
The purpose of this study was to compare and contrast the three major microarray platforms, with the goal of qualifying the long oligonucleotide platform for general use in our laboratories. Our principal findings were that the magnitude of any particular expression ratio may differ from one platform to the other but that the "direction" of gene expression difference for genes with sufficient intensity of signal appears to be well correlated across platforms. The differences in apparent magnitude of expression were not unexpected. The platforms differ widely in many characteristics, including size of targets, number of targets, concentration of targets, labeling protocol, and quantitation algorithms. Nevertheless, the overall concordance of the three platforms was reasonably good, and one should be able to compare experimental data between and among the different array types. That is perhaps not surprising if one considers microarrays simply as miniaturized, multiplexed dot blots.
A survey of the literature on platform comparisons reveals a mix of results. Several studies suggest disagreement in expression between platforms [13-20], and several demonstrate agreement [11,21-34]. The explanation for the discrepancies among these reports is not altogether clear. However, we think the following points should be remembered when designing such studies:
The same RNA templates should be used throughout the entire experiment
RNA preparations from similar, but non-identical, biopsy samples can yield dissimilar results [35], and RNA from different versions of the same cell line can yield widely different expression profiles [36,37].
Only genes common to all platforms should be used in the analysis
That may seem an obvious requirement, but it is not always easy to meet in practice. For example, careful study of sequences present in various cDNA arrays indicates that 20–30% of spots may contain the wrong clone or a misidentified one [10,14,38-40]. A similar situation exists for the short oligonucleotide platform (Affymetrix) because more than 19% of the sequences have been reported not to correspond to the appropriate mRNA Reference Sequence [41,42]. In these studies, measurements of cross-platform consistency were significantly improved when only sequence-matched genes were used. Similarly, discrepant results in studies using cDNA and Affymetrix platforms could be explained by errors in gene sequences [43]. The percentage of incorrect or misidentified sequences in the long oligonucleotide libraries is unknown, but we appear to have found one sequence in error during our own RT-PCR studies. Overall, then, significant disagreement in expression ratios among platforms may be due to sequence errors, not to intrinsic differences among array systems.
Only spots with signals significantly above background (and that pass stringent quality-control filters) should be used in the comparisons
With most high-density arrays, a large proportion of the genes can be expected to have signals near background, as probably fewer than half of all human transcripts are expressed in any particular cell type or tissue [44]. Although stringent filtering decreases the number of measurements included in the analysis, the outcome will be more meaningful.
When comparing results from the three platforms, the magnitudes of ratios were often quite different, but there was generally good agreement in the direction of differences in expression (Figs. 6 &7, Tables 2 &3). Thus, it may be less productive to correlate absolute changes in expression than it is to look for agreement in direction.
Overall, the two oligonucleotide platforms were the most concordant pair. Possible reasons are (i) that the array targets are single stranded, all of the same size (25-mer or 70-mer), with approximately the same melting temperature; (ii) the array elements have a much higher molarity of gene-specific sequences than do the cDNA arrays; (iii) that oligonucleotides tend to be more specific in terms of sequence and less prone to cross-hybridization than are cDNAs.
Conclusion
As the efficiency and economy of oligonucleotide arrays improves, they will probably become the platforms of choice for gene expression analysis, replacing the cDNA type entirely. Long-oligonucleotide arrays are being produced by ink-jet [6,45] and pin-spotting methods [26,27,31]. In these studies, where experiments specifically compared long-oligonucleotide arrays with the more "standard" platforms, correlations were good, and the long-oligonucleotide arrays performed as well as, or better than, the cDNA or short oligonucleotide variety. The data reported here confirm and extend those results, with the added advantage of comparing all three major platforms at one time and in the same place – something that, to our knowledge, has not been done before. As a note of caution, however, we have used materials from three commercial array platforms, and our results and conclusions may not necessarily translate to other platforms and manufacturers.
Since this study began, many articles have appeared reporting the "meta-analysis" of microarray data from unrelated laboratories using several different array platforms [46-54]. Those studies have reported useful clinical and diagnostic findings when the data were stringently filtered before analysis. Eventually, standardization and simplification of the systems may lead to a situation in which only one or two robust methods are used in all laboratories, with a concomitant improvement in the accuracy of gene expression data.
Methods
Experimental design
Three array platforms were tested. Incyte cDNA arrays and Operon long oligonucleotide (70-mer) arrays were printed in house, and Affymetrix 25-mer HG U133A arrays were purchased from the manufacturer. The cDNA and long oligonucleotide arrays were assayed in quadruplicate, two each of red/green and green/red for dye "flips" (reverse fluor experiments) to help eliminate dye bias (except for L428 that had 3 replicates and OCI-Ly3 with 5 replicates). The Affymetrix arrays were tested in duplicate. RNA preparations from six cell lines were tested with each platform using a universal reference RNA (Stratagene) as the standard.
Cell lines and RNAs
Cell lines were grown and RNA isolated at the core Gene Expression Laboratory, NCI-Frederick. MCF10A (benign mammary epithelial), LNCaP (prostate carcinoma), Jurkat (T-cell lymphoma), SUDHL6 (germinal center B-cell like diffuse large B-cell lymphoma), OCI-Ly3 (activated B-cell like diffuse rare B-cell lymphoma and L428 (Hodgkin's lymphoma) were grown under standard conditions [55], and RNA was isolated from the cells using TriReagent following the manufacturer's protocol (Molecular Research Center, Inc., Cincinnati, OH). Integrity of the RNA was confirmed by analysis with the Agilent 2100 Bioanalyzer (Palo Alto, CA) using the RNA 6000 LabChip® kit. For use as the index standard, Human Universal Reference RNA (HUR RNA) was purchased from Stratagene (La Jolla, CA).
Preparation of cDNA and oligonucleotide arrays
Arrays with 10,000 cDNAs were prepared from ready-to-print UniGEM2 libraries obtained from Incyte, Inc. (Wilmington, DE). Version 2.0 libraries containing 22,000 oligonucleotides of 70 bases in length were obtained from Operon, Inc. (Alameda, CA). Arrays were printed by standard protocols on Corning Ultra-GAPS II slides (Corning, NY) using a GeneMachine® (San Carlos, CA) instrument. cDNAs were suspended at a concentration of 100 μg/ml and oligonucleotides at 25 μM in 3XSSC buffer, and the arrays were printed using SMP3 pins from Telechem International (Sunnyvale, CA). The spotted nucleic acids were fixed to the slides using protocols supplied by the manufacturers. The 25-mer oligonucleotide HG U133A Genechip® arrays were purchased from Affymetrix, Inc (Santa Clara, CA).
Labeling and purification of targets
Labeled cDNAs from cell samples for hybridization to the long-oligonucleotide and cDNA arrays were synthesized and labeled by the indirect amino-allyl method, using reagents and protocols supplied by the manufacturer. For cDNA synthesis, we used Stratascript reagents (Stratagene, La Jolla, CA) and Cy3 and Cy5 fluorophore amino-allyl reagents from Amersham (Piscataway, NJ). Twenty micrograms of total RNA was used for each synthesis. Labeled cDNA targets were purified using Minelute purification kits (Qiagen, Valencia, CA). cRNA targets for the Affymetrix arrays were synthesized, labeled, and purified according to vendor's (Affymetrix) instructions. Briefly, 10 μg of total RNA was used to make double-stranded cDNA using reagents and protocols obtained from Invitrogen (Carlsbad, CA). Linear amplification was carried out by a modification of the Eberwine T7 method [56], and biotin was incorporated into the cRNA using the Enzo High Yield RNA Transcript labeling Kit (EnzoDiagnostics, Farmingdale, NY).
Hybridization and washing of arrays
The cDNA and long-oligonucleotide microarrays were prehybridized in 40 μl of 5XSSC with 0.1% SDS and 1% BSA at 42°C for 30 minutes. The prehybridization solution was removed, and arrays were hybridized for 16 hours at 42°C in 5XSSC buffer containing Cy3/Cy5 labeled targets, 25% formamide, 0.1% SDS, 1 μg Cot-1 DNA, and 1 μg poly A RNA. The cDNA arrays were washed at room temperature in 2XSSC, 0.1% SDS for 2 minutes, 1XSSC for 2 minutes, 0.2XSSC for 2 minutes and 0.05XSSC for 1 minute. The long-oligonucleotide arrays were treated the same except that the last wash step was omitted. The slides were dried by centrifugation at 650 rpm for 3 minutes. The Affymetrix arrays were hybridized and washed using the manufacturer's protocol. The arrays were then stained with streptavidin-phycoerythrin using the standard antibody amplification protocol (GeneChip® Expression Manual, Affymetrix, Inc., Santa Clara, CA).
Array scanning and image processing
Long-oligonucleotide and cDNA arrays were scanned using an Axon 4000B scanner at 10-micron resolution. Images were processed, and signals from spotted arrays were quantitated using Genepix 3.0 software (Axon Instruments, Union City, CA). The Genepix result files, including signal, background, standard deviation, pixel statistics and quality parameters for both channels were deposited in the microarray database (mAdb) maintained by the NCI/CIT bioinformatics group [57]. The data were filtered on the basis of signal levels and spot quality. Local background values were subtracted from spot intensities to obtain signal values. Data were included if the signal-to-background ratio was ≥ 2, the signal intensity was >100, the spot diameter was between 50 and 180 microns, at least 70% of the pixels were above their standard deviation and the spot was not flagged as "bad" visually. Arrays were normalized by median-centering the logarithmic ratios so that the median ratio of all genes that passed through the filters was equal to 1. For cDNA arrays, normalized expression ratios of 9050 genes were calculated, and the same procedure was applied to long-oligonucleotide arrays for the expression of 20,799 genes. Affymetrix HG-U133A arrays were scanned with the Affymetrix GeneArray scanner at 488 nm and 3-micron resolution. The images were analyzed using Microarray Suite 5.0 software (MAS5; Affymetrix Inc., Santa Clara, CA). Cell-line to HUR expression ratios were computed by comparative analysis of MAS5 values. The data were filtered using MAS5's signal detection and change calls generated at recommended default settings. The ratios included were those that had present calls for signal detection or an increase or decrease call associated with the ratio. The filtered data contained 17,647 genes. For all statistical calculations, logarithmic values of ratios to the base 2 were used.
Determination of genes in common among all platforms
Genes were matched by UniGene cluster methods [11,58], and expression levels were compared for only the 6,430 genes common to all platforms (Table 1 and Figure 1). UniGene clusters (Homo sapiens: UniGene Build #161) of probes of all three platforms were determined by the NCI/CIT Bioinformatics group [57] using the BLAST program from the National Center for Biotechnology (NCBI, Bethesda, MD 20894). There were multiple probes for some of the UniGene clusters. This resulted in matching of two or more probes of one platform to one or more probes of another platform. All possible combinations across three platforms were considered for each UniGene cluster. One probe from each of the platforms was selected as follows. Initially, for a given combination, the replicates were averaged to obtain expression patterns of six RNAs on three platforms. The sum correlation coefficients of these three patterns to their mean pattern was determined. The combination having the highest sum was selected for further analysis. This method relies on the assumption that the probes specific for a gene yield similar expression patterns independent of the platform. The matched expression ratios will be made available at our website .
Estimation of total matched versus mismatched expression values
Expression ratios for genes in common across all cell lines and platforms were determined. If the ratio of a gene was ≥ 1 or ≤ 1 for both platforms being compared, the expression was considered matched irrespective of the magnitude. Otherwise, the ratios were considered to be mismatched (i.e., in opposite directions). These values give a rough, binary indication of the correlation between platforms (Table 2). Concordance between platforms using significant expression ratios at p-value <0.05 and 1.5 to 2-fold threshold levels are given in Table 3.
Correlation of expression among all platforms
To determine how well the data from the three platforms coincided, correlation coefficients were obtained from analysis of scatter plots of the mean expression values from the three array formats. Figure 2a–c are examples of scatter plot analyses of the three platforms using data from the Jurkat cell line RNA. The boundaries for ratio values greater than two-fold are delineated by the external lines. Correlation coefficients for all three pairs of platforms for all six cell lines are listed in Tables 4 &5).
Principal Component Analysis (PCA)
The global gene expression patterns of the six cell lines in all three platforms were studied by principal component analysis (PCA) [59]. All genes (3186) with signals above background were included. Differences in signal magnitude among platforms were nulled out by normalizing the data from each to unit standard deviation. Individual platform variations were accounted for by employing unit variance. A projection on the first three principal components, which explained 48.8% (21%, 14%, and 13%) of the total variance, is shown in Fig. 3. The calculations were done using Partekpro 5.0 software (Partek Inc., St. Charles, MO)
Platform concordance by "correlation of correlations" coefficient
The global concordance of expression levels of the three platforms can be expressed in terms of the 'correlations of correlations' coefficient described previously [10,11]. To perform the computations, step 1 was calculation of the Pearson correlation coefficient across all matched genes for each of the 15 possible pairs of cell lines for each platform. Step 2 was calculation of the Pearson correlation coefficients of those correlation coefficients for the three possible ways of pairing the three platforms. The results are shown in Figures 4a–c.
Hierarchical cluster analysis and Clustered Image Maps (CIMs)
Hierarchical clustering of individual replicates (data not shown), including 3662 genes detected in 80% of the arrays revealed a grouping of RNA samples independent of the type of array platform. A set of 909 genes expressed at >two-fold levels in all platforms in at least two cell lines was used for hierarchical cluster analysis to determine how closely the genes, cell lines and platforms corresponded in their expression values. As a distance metric, we used 1-r, where r is the Pearson correlation coefficient [60]. Cluster nodes were determined using an average linkage algorithm. In the resulting CIM (heat map) [14a], up- and down-regulation with respect to the reference RNA are color-coded as red and green, respectively (Figure 5a–b).
Quantitative real-time RT-PCR
RNA preparations from each of the six cell lines and Stratagene Human Universal Reference RNA were converted into single-stranded cDNA using the Applied Biosystems High-Capacity cDNA Archive kit (ABI, Foster City, CA). Primer and probe sets were obtained as ABI Assays-on-Demand™ Gene Expression Products (TaqMAN® MGB probes, FAM™ dye-labeled) for a set of genes to be studied, as well as GAPDH and BACT, which were used as comparative controls. All quantitative PCR reactions were performed in quadruplicate, with two carried out in an ABI Prism® 7000 sequence detection system and two in a Corbett Research R-300 instrument (Corbett Research, Sidney, Australia). The results were analyzed using the "Relative Quantitation of Gene Expression" method described in ABI Prism 7700 Sequence Detection System User Bulletin #2, Rev B. An initial study comparing only two cell lines, LnCaP and MCF10A, was carried out for the three platforms. Twelve genes that matched in direction of change were chosen for Q-PCR analysis, and the results are shown in Figure 6. Subsequently, ten genes were chosen for analysis using all six cell lines and three platforms. Genes in that set were deliberately chosen for mismatched ratios to determine if any platform was in error more often than the others. The results are shown in Figures 7a–f.
Abbreviations
HUR RNA: Human Universal Reference RNA
mAdb: microarray data base
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DP, JG, JH, JP, JH, and RP worked on developing the long oligonucleotide microarray system and provided the raw data and initial array analysis for both the cDNA and long oligonucleotide arrays. CHK, DM and LG provided input for the long oligonucleotide development and provided the Affymetrix chip data and analysis. GVRC provided statistical analysis of microarray data for the cross platform comparisons of this study. ESK drafted the original manuscript and JW was the main critical reviewer of the manuscript. LS, JCB and JeG provided valuable input into the design of the original experiments and were crucial in support of the new array development. All authors read and approved the manuscript.
Acknowledgements
GVRC acknowledges the critical bioinformatic support provided by John Powell (NCI/CIT Bioinformatics group) for conducting BLAST runs to match the probe sequences to the UniGene clusters. The authors thank Dr. Ting Qiu and Inna Dzekunova (Advanced Technology Institute/National Cancer Institute) for conducting some of the preliminary microarray experiments with the three platforms.
==== Refs
Schena M Shalon D Davis RW Brown PO Quantitative monitoring of gene expression patterns with a complementary DNA microarray Science 1995 270 368 371 7569988
Lockhart DJ Dong H Byrne MC Follettie MT Gallo MV Chee MS Mittmann M Wang C Kobayashi M Horton H Brown EL Expression monitoring by hybridization to high-density oligonucleotide arrays Nature Biotechnol 1996 14 1675 1680 9634850 10.1038/nbt1296-1675
Barrett JC Kawasaki EK Microarrays: the use of oligonucleotides and cDNA for the analysis of gene expression Drug Disc Today 2003 8 134 141 10.1016/S1359-6446(02)02578-3
Hardiman G Microarray platforms – comparisons and contrasts Pharmacogenomics 2004 5 487 502 15212585 10.1517/14622416.5.5.487
Blanchard AP Kaiser RJ Hood LE High density oligonucleotide arrays Biosens Bioelectron 1996 11 687 690 10.1016/0956-5663(96)83302-1
Hughes TR Mao M Jones AR Burchard J Marton MJ Shannon KW Lefkowitz SM Ziman M Schelter JM Meyer MR Kobayashi S Davis C Dai H He YD Stephaniants SB Cavet G Walker WL West A Coffey E Shoemaker DD Stoughton DD Stoughton R Blanchard AP Friend SH Linsley PS Expression profiling using microarrays fabricated by an ink-jet oligonucleotide synthesizer Nature Biotechnol 2001 19 342 347 11283592 10.1038/86730
Van't Veer LJ Dai H van de Vijver MJ He YD Hart AAM Mao M Peterse HL van der Kooy K Marton MJ Witteveen AT Schreiber GJ Kerkhoven RM Roberts C Linsley PS Bernards R Friend SH Gene expression profiling predicts clinical outcome of breast cancer Nature 2002 415 530 536 11823860 10.1038/415530a
Dudley AM Aach J Steffen MA Church GM Measuring absolute expression with microarrays with a calibrated reference sample and an extended signal intensity range Proc Natl Acad Sci USA 2002 99 7554 7559 12032321 10.1073/pnas.112683499
Relogio A Schwager C Richter A Ansorge W Valcarcel J Optimization of oligonucleotide-based DNA microarrays Nucleic Acids Res 2002 30 e51 12034852 10.1093/nar/30.11.e51
Scherf U Ross DT Waltham M Smith LH Lee JK Kohn KW Reinhold WC Myers TG Andrews DT Scudiero DA Eisen MB Sausville EA Pommier Y Botstein D Brown PO Weinstein JN A cDNA microarray gene expression database for the molecular pharmacology of cancer Nature Genetics 2000 24 236 244 10700175 10.1038/73439
Lee JK Bussey KJ Gwadry FG Reinhold W Riddick G Pelletier SL Nishizuka S Szakacs G Annereau J-P Shankavaram U Lababidi S Smith LH Gottesman MM Weinstein JN Comparing cDNA and oligonucleotide data: concordance of gene expression across platforms for the NCI-60 cancer cells Genome Biol 2003 4 R82 14659019 10.1186/gb-2003-4-12-r82
Weinstein JN Myers TG O'Connor PM Friend SH Fornace AJ Kohn KW Fojo T Bates SE Rubinstein LV Anderson NL Buolamwini JK van Osdol WW Monks AP Scudiero DA Sausville EA Zaharevitz DW Bunow B Viswanadhan VN Johnson GS Wittes RE Paull KD An information-intensive approach to the molecular pharmacology of cancer Science 1997 275 343 349 8994024 10.1126/science.275.5298.343
Kuo WP Jenssen T-K Butte AJ Ohno-Machado L Kohane IS Analysis of matched mRNA measurements from two different microarray technologies Bioinformatics 2002 18 405 412 11934739 10.1093/bioinformatics/18.3.405
Li J Pankratz M Johnson JA Differential gene expression patterns revealed by oligonucleotide versus long cDNA arrays Toxicological Sciences 2002 69 383 390 12377987 10.1093/toxsci/69.2.383
Kothapalli R Yoder SJ Mane S Loughran TP Jr Microarray results: how accurate are they? BMC Bioinformatics 2002 3 22 12194703 10.1186/1471-2105-3-22
Rogojina AT Orr WE Song BK Geisert EE Jr Comparing the use of Affymetrix to spotted oligonucleotide microarrays using two retinal pigment epithelium cell lines Molecular Vision 2003 9 482 496 14551534
Carter MG Hamatani T Sharov AA Carmack CE Qian Y Aiba K Ko NT Dudekula DB Brzoska PM Hwang SS Ko MSH In situ-synthesized novel microarray optimized for mouse stem cell and early developmental expression profiling Genome Res 2003 13 1011 1021 12727912 10.1101/gr.878903
Tan PK Downey TJ Spitznagel EL JrXu P Fu D Dimitrov DS Lempicki RA Raaka BM Cam MC Evaluation of gene expression measurements from commercial microarray platforms Nucleic Acids Res 2003 31 5676 5684 14500831 10.1093/nar/gkg763
Lenburg ME Liou LS Gerry NP Frampton GM Cohen HT Christman MF Previously unidentified changes in renal cell carcinoma gene expression identified by parametric analysis of microarray data BMC Cancer 2003 3 31 14641932 10.1186/1471-2407-3-31
Mah N Thelin A Lu T Nikolaus S Kuhbacher T Gurbuz Y Eickhoff H Kloppel G Lehrach H Mellgard B Costello CM Schreiber S A comparison of oligonucleotide and cDNA-based microarray systems Physiol Genomics 2004 16 361 370 14645736 10.1152/physiolgenomics.00080.2003
Kane MD Jatkoe TA Stumpf CR Liu J Thomas JD Madore SJ Assessment of the sensitivity and specificity of oligonucleotide (50 mer) microarrays Nucleic Acids Res 2000 28 4552 4557 11071945 10.1093/nar/28.22.4552
Taniguchi M Miura K Iwao H Yamanaka S Quantitative assessment of DNA microarrays – Comparison with Northern blot analyses Genomics 2001 71 34 39 11161795 10.1006/geno.2000.6427
Hughes TR Mao M Jones AR Burchard J Marton MJ Shannon KW Lefkowitz SM Ziman M Schelter JM Meyer MR Kobayashi S Davis C Dai H He YD Stephaniants SB Cavet G Walker WL West A Coffey E Shoemaker DD Stoughton R Blanchard AP Friend SH Linsley PS Expression profiling using microarrays fabricated by an ink-jet oligonucleotide synthesizer Nature Biotechnol 2001 19 342 347 11283592 10.1038/86730
Guckenberger M Kurz S Aepinus C Theiss S Haller S Leimbach T Panzner U Weber J Paul H Unkmeir A Frosch M Dietrich G Analysis of heat shock response of Neisseria meningitides with cDNA- and oligonucleotide-based DNA microarrays J Bacteriol 2002 184 2546 2551 11948171 10.1128/JB.184.9.2546-2551.2002
Yuen T Wurmbach E Pfeffer RL Ebersole BJ Sealfon SC Accuracy and calibration of commercial oligonucleotide and custom cDNA microarrays Nucleic Acids Res 2002 30 e48 12000853 10.1093/nar/30.10.e48
Barczak A Rodriguez MW Hanspers K Koth LL Tai YC Bolstad BM Speed TP Erle DJ Spotted long oligonucleotide arrays for human gene expression analysis Genome Res 2003 13 1775 1785 12805270 10.1101/gr.1048803
Wang H-Y Malek RL Kwitek AE Greene AS Luu TV Behbahani B Frank B Quackenbush J Lee NH Assessing unmodified 70-mer oligonucleotide probe performance on glass-slide microarrays Genome Biol 2003 4 R5 12540297 10.1186/gb-2003-4-1-r5
Dabrowski M Aerts S Hummelen PV Craessaerts K De Moor B Annaert W Moreau Y De Strooper B Gene profiling of hippocampal neuronal culture J Neurochem 2003 85 1279 1288 12753086 10.1046/j.1471-4159.2003.01753.x
Bloom G Yang IV Boulware D Kwong KY Coppola D Eschrich S Quackenbush J Yeatman TJ Multi-platform, multi-site, microarray-based human tumor classification Am J Pathol 2004 164 9 16 14695313
Jarvinen A-K Hautaniemi S Edgren H Auvinen P Saarela J Kallioniemi Olli-P Monni O Are data from different gene expression platforms comparable? Genomics 2004 83 1164 1168 15177569 10.1016/j.ygeno.2004.01.004
Lee H-S Wang J Tian L Jiang H Black MA Madlung A Watson B Lukens L Pires JC Wang JJ Comai L Osborn TC Doerge RW Chen ZJ Sensitivity of 70-mer oligonucleotides and cDNAs for microarray analysis of gene expression in Arabidopsis and its related species Plant Biotech J 2004 2 45 57 10.1046/j.1467-7652.2003.00048.x
Thompson KL Afshari CA Amin RP Bertram TA Car B Cunningham M Kind C Kramer JA Lawton M Mirsky M Naciff JM Oreffo V Pine PS Sistare FD Identification of platform-independent gene expression markers of cisplatin nephrotoxicity Environ Health Perspect 2004 112 488 494 15033599
Ulrich RG Rockett JC Gibson GG Pettit SD Overview of an interlaboratory collaboration on evaluating the effects of model hepatotoxicants on hepatic gene expression Environ Health Perspect 2004 112 423 427 15033591
Parmigiani G Garrett-Mayer ES Anbazhagan Gabrielson E A cross-study comparison of gene expression studies for the molecular classification of lung cancer Clin Cancer Res 2004 10 2922 2927 15131026
Catherino WH Segars JH Microarray analysis in fibroids: which gene list is correct? Fertil Steril 2003 80 293 294 12909489 10.1016/S0015-0282(03)00958-0
Nugoli M Chuchana P Vendrell J Orsetti B Ursule L Nguyen C Birnbaum D Douzery EJP Cohen P Theillet C Genetic variability in MCF-7 sublines: evidence of rapid genomic and RNA expression profile modifications BMC Cancer 2003 3 13 12713671 10.1186/1471-2407-3-13
Freedland SJ Pantuck AJ Paik SH Zisman A Graeber TG Eisenberg D McBride WH Nguyen D Tso C-L Belldegrun AS Heterogeneity of molecular targets on clonal cancer lines derived from a novel hormone-refractory prostate cancer tumor system The Prostate 2003 55 299 307 12712409 10.1002/pros.10226
Ross DT Scherf U Eisen MB Perou CM Rees C Spellman P Iyer V Jeffrey SS Van de Rijn M Waltham M Pergamenschikov A Lee JCF Lashkari D Shalon D Myers TG Weinstein JN Botstein D Brown PO Systematic variation in gene expression patterns in human cancer cell lines Nature Genetics 2000 24 227 235 10700174 10.1038/73432
Halgren RG Fielden MR Fong CJ Zacharewski TR Assessment of clone identity and sequence fidelity for 1189 IMAGE cDNA clones Nucleic Acids Res 2001 29 582 588 11139630 10.1093/nar/29.2.582
Taylor E Cogdell D Coombes K Hu L Ramdas L Tabor A Hamilton S Zhang W Sequence verification as quality-control step for production of cDNA arrays Biotechniques 2001 31 62 65 11464521
Mecham BH Klus GT Strovel J Augustus M Byrne D Bozso P Wetmore DZ Mariani TJ Kohane IS Szallasi Z Sequence-matched probes produce increased cross-platform consistency and more reproducible biological results in microarray-based gene expression measurements Nucleic Acids Res 2004 32 e74 15161944 10.1093/nar/gnh071
Mecham BH Wetmore DZ Szallasi Z Sadovsky Y Kohane I Mariani TJ Increased measurement accuracy for sequence-verified microarray probes Physiol Genomics 2004
Goodsaid FM Smith RJ Rosenblum Quantitative PCR deconstruction of discrepancies between results reported by different hybridization platforms Environ Health Perspect 2004 112 456 459 15033595
Jongeneel CV Iseli C Stevneson BJ Riggins GJ Lal A Mackay A Harris RA O'Hare MJ Neville AM Simpson AJG Strausberg RL Comprehensive sampling of gene expression in human cell lines with massively parallel signature sequencing Proc Natl Acad Sci USA 2003 100 4702 4705 12671075 10.1073/pnas.0831040100
Carter MG Hamatami T Sharov AA Carmack CE Qian Y Aiba K Ko NT Dudekula DB Brzoska PM Hwang SS Ko MSH In situ-synthesized novel microarray optimized for mouse stem cell and early developmental expression profiling Genome Res 2003 13 1011 1021 12727912 10.1101/gr.878903
Rhodes DR Barrette TR Rubin MA Ghosh D Chinnaiyan AM Meta-analysis of microarrays: Interstudy validation of gene expression profiles reveals pathway dysregulation in prostate cancer Cancer Res 2002 62 4427 4433 12154050
Moreau Y Aerts S De Moor B De Strooper B Dabrowski M Comparison and meta-analysis of microarray data: from the bench to the computer desk Trends Genet 2003 19 570 577 14550631 10.1016/j.tig.2003.08.006
Wright G Tan B Rosenwald A Hurt EH Wiestner A Staudt LM A gene expression-based method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma Proc Natl Acad Sci USA 2003 100 9991 9996 12900505 10.1073/pnas.1732008100
Ferl GZ Timmerman JM Witte ON Extending the utility of gene profiling data by bridging microarray platforms Proc Natl Acad Sci USA 2003 100 10585 10587 12963810 10.1073/pnas.2034937100
Culhane AC Perriere G Higgins DG Cross-platform comparison and visualization of gene expression data using co-inertia analysis BMC Bioinformatics 2003 4 59 14633289 10.1186/1471-2105-4-59
Lee HK Hsu AK Sajdak J Qin J Pavlidis P Coexpression analysis of human genes across many microarray data sets Genome Res 2004 14 1085 1094 15173114 10.1101/gr.1910904
Bloom G Yang IV Boulware D Kwong KY Coppola D Eschrich S Quackenbush J Yeatman TJ Multi-platform, multi-site, microarray-based human tumor classification Am J Pathol 2004 164 9 16 14695313
Rhodes DR Yu J Shanker K Deshpande N Varambally R Ghosh D Barrette T Pandey A Chinnaiyan AM Large-scale meta-analysis of cancer microarray data identifies common transcriptional profiles of neoplastic transformation and progression Proc Natl Acad Sci USA 101 9309 9314 15184677 10.1073/pnas.0401994101
Wang J Coombes KR Highsmith WE Keating MJ Abruzzo LV Differences in gene expression between B-cell chronic lymphocytic leukemia and normal B cells: a meta-analysis of three microarray studies Bioinformatics 2004
Chen SL Maroulakou IG Green JE Romano-Spica V Modi W Lautenberger J Bhat NK Isolation and characterization of a novel gene expressed in multiple cancers Oncogene 1996 12 741 751 8632896
Van Gelder RN von Xastrow ME Yool A Dement DC Barchus JD Eberwine JH Amplified RNA synthesized from limited quantities of heterogeneous cDNA Proc Natl Acad Sci USA 1990 87 1663 1667 1689846
Greene JM Asaki E Bian X Bock C Castillo S Chandramouli G Martell R Meyer K Ruppert T Sundaram S Tomlin J Yang L Powell J The NCI/CIT microarray database (mAdb) system – bioinformatics for the management and analysis of Affymetrix and spotted gene expression microarrays Proc AMIA Symp 2003 1066 14728569
Bussey KJ Kane D Sunshine M Narasimhan S Nishizuka S Reinhold WC Zeeberg B Ajay Weinstein JN MatchMiner A tool for batch navigation among gene and gene product identifiers Genome Research 2003 4 R27
Mardia KV Kent JT Bibby JM Multivariate Analysis 1979 London Academic Press
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
| 15876355 | PMC1140753 | CC BY | 2021-01-04 16:39:53 | no | BMC Genomics. 2005 May 5; 6:63 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-63 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-381585049410.1186/1465-9921-6-38ResearchAssociation of current smoking with airway inflammation in chronic obstructive pulmonary disease and asymptomatic smokers Willemse Brigitte WM [email protected] Hacken Nick HT [email protected] Bea [email protected] Dirkje S [email protected] Wim [email protected] Department of Pathology, University Medical Centre Groningen, Groningen, The Netherlands2 Department of Pulmonology, University Medical Centre Groningen, Groningen, The Netherlands2005 25 4 2005 6 1 38 38 14 11 2004 25 4 2005 Copyright © 2005 Willemse et al; licensee BioMed Central Ltd.2005Willemse 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
Inflammation in the airways and lung parenchyma underlies fixed airway obstruction in chronic obstructive pulmonary disease. The exact role of smoking as promoting factor of inflammation in chronic obstructive pulmonary disease is not clear, partly because studies often do not distinguish between current and ex-smokers.
Methods
We investigated airway inflammation in sputum and bronchial biopsies of 34 smokers with chronic obstructive pulmonary disease (9 Global initiative for Chronic Obstructive Lung Disease stage 0, 9 stage I, 10 stage II and 6 stage III) and 26 asymptomatic smokers, and its relationship with past and present smoking habits and airway obstruction.
Results
Neutrophil percentage, interleukin-8 and eosinophilic-cationic-protein levels in sputum were higher in chronic obstructive pulmonary disease (stage I-III) than asymptomatic smokers. Inflammatory cell numbers in bronchial biopsies were similar in both groups. Current smoking correlated positively with macrophages: in bronchial biopsies in both groups, and in sputum in chronic obstructive pulmonary disease. Pack-years smoking correlated positively with biopsy macrophages only in chronic obstructive pulmonary disease.
Conclusion
Inflammatory effects of current smoking may mask the underlying ongoing inflammatory process pertinent to chronic obstructive pulmonary disease. This may have implications for future studies, which should avoid including mixed populations of smokers and ex-smokers.
current smokingbronchial biopsiessputum
==== Body
Background
Chronic obstructive pulmonary disease (COPD) is one of the most important causes of death and its prevalence is still increasing [1]. The major risk factor in the development and progression of COPD is cigarette smoking. COPD is characterised by fixed airway obstruction and respiratory symptoms, i.e. chronic cough, sputum production and dyspnoea. COPD patients have, just like so called healthy smokers, an inflammatory reaction involving the entire tracheobronchial tree [2,3].
As compared to healthy non-smokers the degree of airway inflammation seems higher in COPD patients. For example, higher numbers of CD8 positive T-cells, macrophages, neutrophils, and mast cells, both in central and peripheral airways have been found in COPD patients, irrespective whether these patients were current smokers or ex-smokers [4-10]. In addition, the percentage of neutrophils and IL-8 levels in sputum and broncho-alveolar lavage of COPD patients were higher [7,11-16]. As compared to healthy smokers, the differences with COPD patients are less clear cut. For example, higher numbers of neutrophils, macrophages and CD8 positive T-cells in the peripheral airways of COPD patients were found as compared to smokers [10,17-19], whereas others did not [10,19,20]. Two studies showed a higher percentage of neutrophils and higher IL-8 levels in broncho-alveolar lavage of COPD patients [13,21], whereas Linden et al. found no differences [7]. A few studies showed higher numbers of neutrophils [22], CD3, CD4[23] CD8 positive T-cells [23,24] in bronchial biopsies, whereas other studies found no differences in neutrophils [24], CD3, CD4 [22,24] and CD8 positive T-cels [22], macrophages, eosinophils and mast cells [22,24]. In conclusion, COPD patients have a higher degree of airway inflammation compared to non-smokers, however it remains unclear whether this is also true comparing COPD patients with so called healthy smokers.
Definite conclusions about the exact role of cigarette smoking in COPD are difficult to draw for a number of reasons. First, most studies investigated smokers combined with ex-smokers. Second, many studies investigated COPD patients combined with patients with chronic bronchitis. Third, many studies investigated only one aspect of inflammation, or only one compartment (sputum, broncho-alveolar lavage, bronchial biopsies, peripheral airways), which may be insufficient to obtain a full view. Fourth, remodelling in COPD may itself generate and maintain an inflammatory process, independent of cigarette smoking [25].
In order to elucidate the role of smoking on inflammation in COPD we have investigated airway inflammation in sputum and bronchial biopsies of asymptomatic smokers and smokers with COPD. Furthermore, we assessed whether airway inflammation is related to the number of cigarettes smoked per day, to pack-years smoking and to severity of airway obstruction.
Methods
Subjects
Subjects were recruited from the pulmonary outpatient clinic of the Groningen University Hospital and by advertisements in local newspapers. 34 smokers with COPD and 26 smokers without COPD were included according to the ERS criteria [26]. Smokers with COPD had chronic cough and sputum production for at least 3 months for 2 successive years, and an forced expiratory volume in one second (FEV1)/ vital capacity (VC) ≤ 88% of predicted for males and ≤ 89% of predicted for females. Asymptomatic smokers without COPD had no chronic respiratory symptoms, and FEV1/VC >88% of predicted for males and >89% of predicted for females and an FEV1 >85% of predicted. To detect respiratory symptoms to delineate the group of symptomatic smokers without COPD we used the questions about respiratory symptoms and smoking from the Dutch version of the British Medical Research Council's standardised questionnaire [27]. These data were collected by interviewing the participants at the first visit. All participants had to meet the following criteria: age between 45–75 years, minimum of 10 pack-years smoking, actual smoking ≥ 10 cigarettes per day, reversibility to salbutamol < 9% of the predicted FEV1, no use of inhaled or oral corticosteroids in the previous 6 months, no atopy (no positive skin prick test for 10 common aeroallergens and serum total IgE < 200 IU), no respiratory tract infection 1 month prior to the study. After inclusion, subjects were categorized according to the Global Initiative for Chronic Obstructive Lung Disease, GOLD stage 0-IV [28]. GOLD stage 0 (symptomatic smokers): 'at risk' with normal spirometry but chronic symptoms (cough, sputum production); GOLD stages I-IV: FEV1/FVC post bronchodilator (post BD) < 70% and GOLD stage I: FEV1 post BD ≥ 80% predicted; GOLD stage II: 50% ≤ FEV1 post BD < 80% predicted; GOLD stage III: 30% ≤ FEV1 post BD <50% predicted and GOLD stage IV: 30% ≤ FEV1 post BD or FEV1<50% predicted plus respiratory failure. Current smoking was confirmed by urinary cotinine levels > 25 ng/ml. Before each measurement subjects were asked not to use long or short-acting β2 agonists and/or ipratropium at least 12 hours before the test. The local medical ethics committee approved the study protocol and all subjects gave their written informed consent.
Study Design
All subjects visited the hospital on 5 separate days, at least one week apart. Lung function tests (flow-volume curves, reversibility, airway conductance), airway hyperresponsiveness (AHR) to methacholine and to adenosine-5'-monophosphate (AMP), and sputum induction (twice) were performed and all subjects underwent bronchoscopy.
Lung Function
Lung function (FEV1, FEV1/VC) was measured using dry wedge spirometry (Masterscope, Jaeger, Breda, The Netherlands) according to standardized guidelines [29]. Specific airway conductance (sGaw) was measured by body plethysmography (Masterscope, Jaeger, Breda, The Netherlands). Provocation tests were performed with a 2-minute tidal breathing method adapted from Cockcroft and co-workers [30]. After an initial nebulised saline (0.9%) challenge, subjects inhaled doubling concentrations, ranging from 0.038 to 39.2 mg/ml of methacholine-bromide (Sigma Chemical Co. St Louis, MO) and from 0.04 to 320 mg/ml of AMP (Sigma Chemical Co. St Louis, MO) at 5-minute intervals. Tests were terminated when FEV1 had fallen 20% or more from its baseline value (PC20).
Sputum Induction and Sputum Processing
Sputum was induced by inhalation of hypertonic saline aerosol and processed as described previously [31]. Briefly 15 minutes after salbutamol (400 μg) inhalation, hypertonic saline (3%, 4% and 5% w/v) was nebulised and inhaled for each concentration over a period of 7 minutes. Whole sputum samples were processed within 2 hours after termination of the induction. Two sputum cytospin slides were stained with May-Grünwald-Giemsa for differential cell counts. Counting of 600 non-squamous cells in a blinded way by one technician (B.R.). Sputum samples containing > 80% of squamous cells were excluded from analysis as indication of poor cytospin quality. Interleukin 8 (IL-8) concentration was measured using ELISA (CLB, Amsterdam, the Netherlands) and eosinophil cationic protein (ECP) concentration by a fluorenzyme immunoassay (ImmunoCAP ECP, Pharmacia, Uppsala, Sweden).
Bronchoscopy and biopsy processing
Subjects were not allowed to drink or eat at least 4 hours prior to the bronchoscopy. Smoking was not allowed before the bronchoscopy. On arrival, FEV1 was measured before and 15 minutes after 400 μg salbutamol. Hereafter subjects gargled with 5 ml of 2% lidocaine and had 2% lidocaine sprayed on the posterior pharynx, dripped onto the vocal cords and into the trachea, with a maximum dose of 3 mg/kg lidocaine. A flexible fiberoptic bronchoscope (Olympus B1 IT10, Olympus Optical, Tokyo, Japan) was introduced and preferably 6 bronchial biopsies were taken from the subcarinae of the right middle or lower lobe using a fenestrated cup forceps (Olympus FB-21C, Olympus Optical Tokyo, Japan) [32]. Biopsies were collected into sterile PBS on ice. Two biopsies were directly embedded in Tissue Tek (Bayer Corporation, Elkhart, Indiana, USA), snap-frozen in liquid isopentane and stored at -80°C, 4 biopsies were fixed in 4% paraformaldehyde, processed and embedded in paraffin.
Serial sections were cut from the paraffin biopsies with a thickness of 4 μm and stored at room temperature. Selection of morphological optimal tissue was based on a hematoxylin and eosin stained slide. Tissue slides were deparaffinised with xylene (15 min) and dehydrated before staining. Immunohistochemical staining was performed with monoclonal antibodies against: CD3 (A0452, DAKO, Copenhagen, Denmark) CD4 (CD4-368, Novacastra, UK), CD8 (M7103, DAKO), B cells (CD20 L26, M0755, DAKO), mast cell tryptase (AA1, M7052, DAKO), neutrophil elastase (NP57, M0752, DAKO), macrophages (CD68, M0814, DAKO) and secreted form of eosinophilic cationic protein EG2 (Pharmacia Diagnostics, Sweden). Negative controls were obtained by omission of the primary antibody. Slides were pre-treated with 1 mM EDTA buffer pH = 8 (CD4, CD8), 0.1 mM tris-HCL buffer pH = 9.0 (CD20) in the microwave for 8 or 30 minutes respectively or with 1% protease for 30 minutes at room temperature (CD68, NP57, AA1, EG2). CD3 slides were incubated overnight at 80°C with tris/HCL buffer pH = 9.0. All stainings were performed in an automated system using the Dako Autostainer (DAKO, Copenhagen, Denmark), except for CD4 that was done manually.
The dilutions used were: CD3 1:100; CD4 1:25; CD8 1:100; CD68 1:50; EG2 1:200; NE 1:200; AA1 1:100; CD20 1:400. As detection system we used labelled streptavidin-biotin (LSAB+, K0690, DAKO, Copenhagen, Denmark) except for CD4 where the Envision system (K5007, DAKO, Copenhagen, Denmark) was used. 3-amino-9- Ethyl Carbazole (AEC) (K3469, DAKO, Copenhagen, Denmark), or Nova Red (SK4800, Vector, USA) for CD4, was used as a chromogen (substrate) giving a reddish-brown reaction product. Hydrogen peroxide was used for blocking endogenous peroxidase and haematoxylin was used as a counterstain. For each antigen, all slides were stained simultaneously.
For each immunohistochemical staining 2 sections of 2 different bronchial biopsies were quantified by computer-assisted image analysis at magnification of 200× (Qwin, Leica Microsystems Imaging Solutions Ltd, Cambridge, UK). Automated image analysis to quantify immunopositivity was performed using the next algorithm: first the intensity of the positive area (cells) was appointed in each biopsy by the observer, followed by the intensity of the total area of the biopsy, based on the red-green-blue (RGB) color model [33,34]. After this, all images of the biopsy were analyzed. Excluded were epithelium, submucosal glands, airway smooth muscle tissue and damaged tissue. Afterwards, the algorithm determined the immuno-positive area and the measured area of the biopsy, leading to the percentage positive area per biopsy. A positive area was at least 11.8 μm2, to exclude false positive areas. In this manner, the total positive area and the total measurable area of the biopsy were quantified and the percentage positive area per biopsy was calculated. The smallest evaluable area per section (after exclusion of epithelium, submucosal glands, airway smooth muscles and damaged tissue) was 0.4 mm2. The mean percentage positive area of two biopsies was used. Measurements were performed in a blinded way by 2 observers (B.R. and B.W.).
Data analysis
Analyses were performed using SPSS for Windows 10.0 (SPSS Inc., Chicago, IL). Values of p < 0.05 were considered statistically significant. Clinical data were expressed in means (± SD) or geometric means (minimum-maximum); inflammatory data were expressed in medians (minimum-maximum). Differences between asymptomatic smokers, symptomatic smokers (GOLD 0) and smokers with COPD (GOLD stage I, II and III) were analysed using the Kruskall-Wallis test, a non-parametric equivalent to one-way ANOVA. Only, when the Kruskall-Wallis test was significant the Mann Whitney U test was used to analyse the differences between the 3 groups. Differences between GOLD stages 0, I, II, and III were analysed using the Kruskall-Wallis test, when this test was significant the Mann Whitney U test was used to analyse the differences between the different GOLD stages.
Correlations between smoking characteristics and lung function parameters were calculated with Pearson correlation test. Correlations between inflammatory cells and mediators in sputum and/or bronchial biopsies and smoking characteristics or lung function parameters were calculated with Spearman's rank correlation test. The subjects with GOLD stages I-III were used to investigate the correlations in COPD patients.
Results
Asymptomatic smokers versus smokers with GOLD stage 0-III
The 34 smokers with COPD were categorised into GOLD stage 0 'symptomatic smokers' (n = 9), GOLD stage I (n = 9), stage II (n = 10) and stage III (n = 6); none of the patients fulfilled the criteria for GOLD stage IV. The clinical characteristics of all subjects are presented in (see Additional file 1). Symptomatic smokers (GOLD stage 0) had significantly decreased lung function and more severe hyperresponsiveness to AMP than asymptomatic smokers had. COPD patients in GOLD stages I-III were older, had significantly more pack-years smoking, lower airway conductance and more severe hyperresponsiveness to AMP and methacholine than asymptomatic and symptomatic (GOLD 0) smokers.
Sputum
Two asymptomatic smokers could not produce sputum. The median (range) percentage non-squamous cells was 94 (75–99)% in COPD patients and 88 (64–99)% in asymptomatic smokers (table 1). Symptomatic smokers (GOLD 0) had higher percentage of sputum neutrophils than asymptomatic smokers. Smokers with COPD (GOLD stage I-III) had higher percentage of neutrophils, IL-8 and ECP levels in sputum than asymptomatic smokers, and higher IL-8 levels in sputum than symptomatic smokers. The percentage of macrophages was lower (table 1). In the separate GOLD stages, GOLD stage II had a higher percentage of sputum neutrophils compared with the asymptomatic smokers (70% and 60% respectively) and higher IL-8 and ECP levels in sputum than GOLD stage 0 and I (21.4 ng/ml versus 8.7 and 8.5 ng/ml respectively, and 291 μg/L versus 120 and 99 μg/L respectively). GOLD stage III had higher levels of IL-8 than GOLD stage 0 (27.7 ng/ml and 8.7 ng/ml respectively) and lower ECP levels than GOLD stage II (87 μ/L versus 291 μg/L).
Table 1 Sputum inflammation in smokers with COPD, symptomatic smokers and asymptomatic smokers
COPD Symptomatic smokers Asymptomatic smokers
GOLD I-III GOLD 0
Sputum, n 25 9 24
Volume, ml 4.1 (0.7–14.3) 3.1 (0.3–10.0) 2.3 (0.6–10.8)*
Total cells, 106 6.7 (1.4–54.5) 4.1 (1.1–15.3) 3.5 (0.2–23)*
Cell conc., 103/ml 1507 (484–9620) 2134 (534–4146) 1445 (303–4592)
Nonsquamous cells, % 94 (75–99.7) 92 (81–96) 88 (64–99.5)
Eosinophils, 103/ml 15 (0–106) 20 (0–135) 13 (0–235)
% 1.4 (0–4.0) 1.8 (0–4.1) 0.8 (0–12.6)
Neutrophils, 103/ml 870 (235–7608) 1575 (434–2558) 661 (164–2856)
% 72.6 (45–89) 66 (39–81) 60.1 (31.5–92.6)* †
Macrophages, 103/ml 407 (89–2615) 535 (89–2422) 568 (22–1488)
% 25.4 (8.2–52.7) 28.7 (16.6–58.4) 36 (6.5–62.4)*
Lymphocytes, 103/ml 14 (0–77) 15 (0–62) 11 (1–161)
% 0.8 (0.1–4) 0.8 (0.4–1.6) 0.7 (0.1–3.8)
Epithelial cells, 103/ml 10 (0–107) 0.4 (0–84) 10 (0–55)
% 0.5 (0–11) 0.5 (0–2.5) 0.8 (0–6.6)
Basophils, 103/ml 0 (0–4) 0 (0–12) 0 (0–8)
% 0 (0–0.1) 0 (0–0.3) 0 (0–0.3)
IL-8, ng/ml 16.8 (2.1–161)† 8.7 (0.1–25.7) 5.3 (0–25)*
ECP, μg/L 157 (32–2700) 119.8 (13.3–238) 66 (4.7–1282)*
Values expressed in median (range). Abbreviations: cell conc. = cell concentration; IL-8 = interleukin-8; ECP = eosinophilic cationic protein.
* p < 0.05 asymptomatic smokers versus total COPD (I-III), Mann-Whitney -U test, † p < 0.05 versus GOLD stage 0, Mann-Whitney-U test.
Bronchial biopsies
Bronchial biopsies could not be collected or were of insufficient quality in 2 asymptomatic smokers, in 1 subject GOLD stage 0, and in 6 COPD patients. The percentage positive area of inflammatory cells in bronchial biopsies (CD3, CD4, CD8, CD20, neutrophils, macrophages, eosinophils and mast cells) did not differ between COPD (GOLD I-III), symptomatic smokers (GOLD 0) and asymptomatic smokers (table 2). Only COPD patients with GOLD stage II had a higher percentage positive CD3 area than asymptomatic smokers (1.84 (0.24–9.24) and 0.76 (0.17–2.4) respectively).
Table 2 Inflammation in bronchial biopsies from smokers with COPD and asymptomatic smokers
Total COPD Symptomatic smokers Asymptomatic smokers
GOLD I-III GOLD 0
Biopsies, n 19 8 24
CD3, %positive area 1.05 (0.2–9.24) 0.68 (0.19–1.7) 0.76 (0.17–2.4)
CD4, %positive area 0.041 (0.01–0.57) 0.073 (0–0.18) 0.04 (0–0.15)
CD8, %positive area 0.27 (0.03–2.55) 0.19 (0.02–1.53) 0.33 (0.3–1.25)
CD4/CD8 ratio 0.19 (0.1–4.4) 0.39 (0.04–1.2) 0.18 (0–0.91)
CD20, %positive area 0.003 (0–3.40) 0.003 (0–0.23) 0.005 (0–0.12)
NP57, %positive area 0.025 (0–0.13) 0.05 (0–0.23) 0.021 (0–0.36)
CD68, %positive area 0.035 (0–0.21) 0.056 (0–0.16) 0.041 (0–0.32)
EG2, %positive area 0.021 (0–0.31) 0.049 (0–0.15) 0.063 (0–0.59)
AA1, %positive area 0.15 (0.01–0.91) 0.22 (0–0.41) 0.22 (0.1–1.16)
Values expressed in median (range). Abbreviations: CD20 = B-cell marker, NP57 = neutrophil elastase, CD68 = macrophages, EG2 = eosinophils, AA1 = mast cell tryptase.
Correlations of lung function with smoking and airway inflammation
FEV1 post BD (% predicted) correlated negatively with the number of pack-years smoking (r = -0.51, p = 0.03) in COPD, but not significantly with the number of cigarettes smoked per day.
FEV1 post BD correlated negatively with IL-8 levels in sputum and positively with macrophages in sputum and mast cells in bronchial biopsies of patients with COPD (table 3). The latter correlation was mainly caused by 4 patients with low mast cell positive areas. In asymptomatic smokers, no significant correlations were found between lung function and airway inflammation (table 3).
Table 3 Spearman's rank correlations between current smoking and airway obstruction and airway inflammation.
COPD GOLD I-III (n = 19) Asymptomatic smokers (n = 26)
rho p-value rho p-value
Cigarettes/day
Neutrophils sputum, % -0.46 0.02 -0.24 NS
Macrophages sputum, % 0.44 0.027 0.35 0.095
Eosinophils sputum, % -0.19 NS -0.39 0.057
Eosinophils sputum, 106/ml -0.11 NS -0.44 0.029
CD68 biopsy, % pos. area 0.69 0.002 0.46 0.029
EG2 biopsy, % pos. area -0.79 0.001 0.02 NS
Pack-years smoking
CD68 biopsy, %pos. area 0.48 0.043 0.21 NS
FEV1 post BD, %pred.
Neutrophils sputum, % -0.31 NS 0.15 NS
Macrophages sputum, % 0.41 0.046 -0.22 NS
IL-8 sputum, ng/ml -0.54 0.007 0.15 NS
CD68 biopsy, %pos. area 0.33 NS 0.18 NS
AA1 biopsy, %pos. area 0.53 0.02 0.09 NS
CD68 = macrophages; EG2 = eosinophils; % pos. area = percentage positive area; FEV1 = forced expiratory volume in one second; post BD = post bronchodilator (15 minutes after 400 μg salbutamol); IL-8 = interleukin 8; AA1 = mast cells; NS = not significant
AHR did not correlate with number of cigarettes smoked per day, number of pack-years smoking or airway inflammation in sputum or bronchial biopsies in both asymptomatic smokers and COPD patients (data partially presented and discussed earlier: Willemse et al, [35]).
Correlations of current smoking with airway inflammation
The number of cigarettes smoked per day correlated negatively with neutrophils and positively with macrophages in sputum, which was significant in COPD (table 3, figure 1). The number of cigarettes smoked per day correlated positively with macrophages in bronchial biopsies, in both groups (table 3, figure 2). In asymptomatic smokers, the number of cigarettes per day correlated negatively with the number and percentage of eosinophils in sputum. In COPD the number of cigarettes smoked per day correlated negatively with eosinophil area in bronchial biopsies (table 3).
Figure 1 Spearman's rank correlation: Cigarettes smoked per day and macrophages in induced sputum. COPD (■, ): rho = 0.44 p = 0.03 and asymptomatic smokers(□, ): rho = 0.35 p = 0.1.
Figure 2 Spearman's rank correlation: Cigarettes smoked per day and percentage of macrophages in biopsies. COPD (■, ): rho = 0.69 p = 0.002 and asymptomatic smokers(□, ): rho = 0.46 p = 0.03.
Correlations of pack-years smoking with airway inflammation
In COPD patients pack-years smoking was positively correlated with the macrophage percentage positive area (table 3). Otherwise no significant correlations were found.
Discussion
This study shows that asymptomatic smokers, symptomatic smokers (GOLD stage 0), and smoking patients with COPD have a large overlap in inflammation as assessed in sputum and airway wall biopsies. Patients with stage GOLD I-III had a higher percentage of neutrophils, and higher ECP and IL-8 levels in sputum than asymptomatic smokers, and higher IL-8 levels than symptomatic smokers. In symptomatic smokers percentage sputum neutrophils were higher than in asymptomatic smokers.
Whereas current smoking was associated with higher numbers of inflammatory cells in both asymptomatic smokers and COPD patients, pack-years smoking was only associated with higher airway wall macrophages in COPD and to the severity of airway obstruction. More severe airway obstruction in its turn was associated with lower percentage of sputum macrophages in smokers with COPD. Thus, the small difference in airway inflammation found between smokers with and without COPD may be due to the interference of current cigarette smoking.
This study demonstrates that a higher number of daily smoked cigarettes is associated with a higher percentage of macrophages in bronchial biopsies and sputum, both in smokers with COPD and asymptomatic smokers. In addition, eosinophils and neutrophils in sputum were negatively correlated to current smoking. Only few studies have provided data on correlations between airway inflammation and current smoking since smokers and ex-smokers were generally investigated together as one group. Two studies reported a positive correlation between neutrophils in bronchoalveolar lavage and the number of cigarettes smoked per day when asymptomatic smokers, chronic bronchitis patients and COPD patients were analysed together [7,13]. One study in asymptomatic smokers reported that the number of cigarettes smoked per day correlated positively with macrophages and IL-8 levels in bronchoalveolar lavage [36]. Macrophages in the central airways of smokers with and without COPD may be a direct inflammatory reflection of current smoking. On the other hand, it is not likely that current smoking is the only factor responsible for the accumulation of macrophages, since they are also increased in bronchial biopsies of ex-smokers with COPD [37]. Furthermore, we show that not only current smoking but also a higher number of pack-years smoking is associated with higher number of macrophages in COPD. This suggests that effects of current smoking are superimposed upon the underlying macrophage infiltration, which is part of the ongoing inflammatory process in COPD. This is important to realise when investigating the inflammatory and remodelling processes in smokers and ex-smokers with or without COPD. We therefore strongly suggest to avoid including mixed populations of smokers and ex-smokers in future studies on inflammatory processes in COPD
Current smoking was negatively related to eosinophils, i.e. the more cigarettes smoked per day the fewer eosinophils were present in sputum of asymptomatic smokers and in bronchial biopsies of patients with COPD. It may be that smoking has an anti-inflammatory effect on eosinophils or may influence cell kinetics. It has been suggested that carbon monoxide (CO) present in cigarette smoke has an anti-inflammatory potential [38,39], at least with respect to certain cell types and/or subsets. The extent and relevance of this supposed anti-inflammatory effect in humans remains to be established, but in guinea pigs it has been shown that acute cigarette smoke exposure suppresses the number of eosinophils after 6, 12 and 24 hours [40]. This may indicate that even the cigarettes smoked 24 hours before sputum induction or bronchoscopy could have induced this inverse relationship between current smoking and eosinophilic inflammation, since our participants refrained from smoking for 8 hours before the bronchoscopy. Nevertheless, it is well known that repetitive smoking for several years causes extensive damaging effects, indicating that the long-term overall effects of cigarette smoke dominate the anti-inflammatory effects.
Macrophages in bronchial biopsies of smokers with COPD were positively associated with pack-years smoking. No other relationships between pack-years smoking and airway inflammation were found in our study. This is in agreement with previous studies which either did not find any correlations [41] or did not investigate this [11,15,42]. Only Lams et al. [24] reported a positive correlation between CD8+ cells in bronchial biopsies and pack-years smoking and a negative correlation between neutrophils in bronchial biopsies and pack-years smoking, when all smokers (COPD and asymptomatic smokers) were analysed. In broncho-alveolar lavage percentage neutrophils was positively associated with pack-years smoking when all smokers and ex-smokers with and without COPD were analysed together [7,13].
One would expect that in COPD patients inflammatory markers would be more related to pack-years smoking instead of the number of cigarettes smoked per day. However, only macrophages in bronchial biopsies showed a positive correlation with pack-years smoking whereas macrophages, eosinophils and neutrophils were related to the number of cigarettes smoked. This may indicate that some of the inflammation due to cumulative smoke exposition is overruled by inflammation caused by current smoking. Neutrophils and eosinophils are "fast moving, or transient" inflammatory cells, whereas macrophages remain much longer in the lung tissue. This stresses the importance of macrophages in the development and progression of COPD.
This study shows that the percentage of neutrophils in sputum is higher in smokers with COPD (median 72.6%) than in asymptomatic smokers (median 60.1%), especially in GOLD stage II. This is completely in line with results of previous studies, which showed that smokers with moderate to severe COPD had higher total cell numbers and percentages of neutrophils in sputum than asymptomatic smokers [11,15,42]. Thus our finding suggests that this aspect of inflammation is associated with disease severity.
In symptomatic smokers (GOLD stage 0) the percentage of neutrophils in sputum was higher than in asymptomatic smokers, but similar to COPD patients. This has not been investigated in induced sputum before, however in broncho-alveolar lavage neutrophils show the same pattern [12]. No other differences were found in airway inflammation between symptomatic smokers and asymptomatic smokers. This is in contrast to the study of Sun et al [43], who investigated smokers with chronic bronchitis and found not only an increased number of neutrophils in broncho-alveolar lavage, but also increased eosinophils, mast-cells, CD4 positive and CD8 positive T cells compared to "healthy" smokers. This suggests that chronic bronchitis is better reflected by broncho-alveolar lavage than by induced sputum or bronchial biopsies.
In the present study, IL-8 levels in sputum were significantly higher in smokers with COPD than in asymptomatic and symptomatic smokers. In addition, higher IL-8 levels strongly correlated with more severe airway obstruction in smokers with COPD. This is in line with the data of Keatings et al. who showed that both IL-8 and percentage of neutrophils in sputum were increased in patients with moderate COPD as compared to asymptomatic smokers [15]. This may suggest that IL-8, a chemoattractant of neutrophils and an activator of MMP-9, plays a role in the development of airway obstruction. Alternatively, this may reflect the airway obstruction present.
Inflammatory cell density in bronchial biopsies did not significantly differ between smokers with COPD (GOLD I-III), asymptomatic smokers and symptomatic smokers. Only CD3 percentage positive areas in bronchial biopsies were higher in smokers with COPD stage II than in asymptomatic smokers. In agreement with our findings, other studies [24,41] investigating smokers with and without COPD, found no differences in neutrophils, macrophages, eosinophils, CD4 positive cells or CD4/CD8 ratio in bronchial biopsies. In contrast, one previous study demonstrated a higher number of CD8+ cells in smokers with predominantly moderate COPD compared to asymptomatic smokers [24]. In addition, two other studies demonstrated that CD3+ and CD8+ cell numbers were lower and macrophages and neutrophils were higher in smokers with severe COPD [22,41]. It may thus well be that differences between smokers with and without COPD become only apparent in case of severe COPD. Unfortunately the number of patients with evaluable biopsies was too small in our study population (n = 4) to investigate whether this indeed is the case.
A factor that should be taken into consideration is the age difference between the COPD patients and asymptomatic smokers under study. Previous studies investigated younger (mean age 35 years) asymptomatic smokers than our participants (mean age 50 years) [11,15,42]. The composition of sputum may differ between older and younger healthy subjects, as shown in bronchoalveolar lavage where the number of total cells and neutrophils increase with age [44]. Since we investigated COPD patients and asymptomatic smokers of almost similar age, our data are not hampered by age differences.
Conclusion
Smoking COPD patients with GOLD stage I-III had almost similar airway wall and sputum inflammation as asymptomatic and symptomatic smokers without airway obstruction. Current smoking was associated with airway inflammation in patients with COPD and in asymptomatic smokers, whereas this was not the case for the cumulative pack-years smoked. In contrast, cumulative pack-years smoking was associated with the level of airway obstruction in COPD, suggesting that cumulative smoking induces chronic inflammation with subsequent sequelae of airway obstruction. Our results indicate that inflammatory effects of current smoking may mask findings of chronic inflammation in COPD, since numbers of inflammatory cells in bronchial biopsies and sputum are comparable in smokers with mild COPD and asymptomatic smokers.
Authors' contributions
BW carried out the data collection and its coordination, immunohistochemical staining and quantification of the bronchial biopsies, performed the statistical analysis and interpretation of the data and drafted and revised the manuscript. NtH contributed to the conception and design of the study, the data collection and the interpretation of the data and revised the manuscript. BR carried out the sputum processing and immunoassays and revised the manuscript. DP contributed to the conception and design of the study, the data collection and the interpretation of the data and revised the manuscript. WT contributed to the conception and design of the study, the data collection and the interpretation of the data and revised the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Description of the clinical characteristics of the participating subjects
Click here for file
Acknowledgements
This project was funded by the Dutch Asthma Foundation (NAF 97.74). The authors would like to thank Mrs A.A. Smidt for her assistance with the bronchial biopsies and Mrs I. Barta-Sloots for her help with the ECP measurements.
==== Refs
Murray CJ Lopez AD Alternative projections of mortality and disability by cause 1990-2020: Global Burden of Disease Study Lancet 1997 349 1498 1504 9167458 10.1016/S0140-6736(96)07492-2
Barnes PJ Shapiro SD Pauwels RA Chronic obstructive pulmonary disease: molecular and cellular mechanisms Eur Respir J 2003 22 672 688 14582923
Saetta M Turato G Maestrelli P Mapp CE Fabbri LM Cellular and Structural Bases of Chronic Obstructive Pulmonary Disease Am J Respir Crit Care Med 2001 163 1304 1309 11371392
Saetta M Di Stefano A Maestrelli P Ferraresso A Drigo R Potena A Ciaccia A Fabbri LM Activated T-lymphocytes and macrophages in bronchial mucosa of subjects with chronic bronchitis Am Rev Respir Dis 1993 147 301 306 8430952
O'Shaughnessy TC Ansari TW Barnes NC Jeffery PK Inflammation in bronchial biopsies of subjects with chronic bronchitis: inverse relationship of CD8+ T lymphocytes with FEV1 Am J Respir Crit Care Med 1997 155 852 857 9117016
Turato G Di Stefano A Maestrelli P Mapp CE Ruggieri MP Roggeri A Fabbri LM Saetta M Effect of smoking cessation on airway inflammation in chronic bronchitis Am J Respir Crit Care Med 1995 152 1262 1267 7551380
Linden M Rasmussen JB Piitulainen E Tunek A Larson M Tegner H Venge P Laitinen LA Brattsand R Airway inflammation in smokers with nonobstructive and obstructive chronic bronchitis Am Rev Respir Dis 1993 148 1226 1232 8239158
Saetta M Di Stefano A Turato G Facchini FM Corbino L Mapp CE Maestrelli P Ciaccia A Fabbri LM CD8+ T-lymphocytes in peripheral airways of smokers with chronic obstructive pulmonary disease Am J Respir Crit Care Med 1998 157 822 826 9517597
Pesci A Rossi GA Bertorelli G Aufiero A Zanon P Olivieri D Mast cells in the airway lumen and bronchial mucosa of patients with chronic bronchitis Am J Respir Crit Care Med 1994 149 1311 1316 8173772
Saetta M Turato G Baraldo S Zanin A Braccioni F Mapp CE Maestrelli P Cavallesco G Papi A Fabbri LM Goblet cell hyperplasia and epithelial inflammation in peripheral airways of smokers with both symptoms of chronic bronchitis and chronic airflow limitation Am J Respir Crit Care Med 2000 161 1016 1021 10712357
Balzano G Stefanelli F Iorio C De Felice A Melillo EM Martucci M Melillo G Eosinophilic inflammation in stable chronic obstructive pulmonary disease. Relationship with neutrophils and airway function Am J Respir Crit Care Med 1999 160 1486 1492 10556110
Lacoste JY Bousquet J Chanez P Van Vyve T Simony-Lafontaine J Lequeu N Vic P Enander I Godard P Michel FB Eosinophilic and neutrophilic inflammation in asthma, chronic bronchitis, and chronic obstructive pulmonary disease J Allergy Clin Immunol 1993 92 537 548 8409114
Thompson AB Daughton D Robbins RA Ghafouri MA Oehlerking M Rennard SI Intraluminal airway inflammation in chronic bronchitis. Characterization and correlation with clinical parameters Am Rev Respir Dis 1989 140 1527 1537 2604284
Traves SL Culpitt SV Russell RE Barnes PJ Donnelly LE Increased levels of the chemokines GROalpha and MCP-1 in sputum samples from patients with COPD Thorax 2002 57 590 595 12096201 10.1136/thorax.57.7.590
Keatings VM Collins PD Scott DM Barnes PJ Differences in interleukin-8 and tumor necrosis factor-alpha in induced sputum from patients with chronic obstructive pulmonary disease or asthma Am J Respir Crit Care Med 1996 153 530 534 8564092
Keatings VM Barnes PJ Granulocyte activation markers in induced sputum: comparison between chronic obstructive pulmonary disease, asthma, and normal subjects Am J Respir Crit Care Med 1997 155 449 453 9032177
Baraldo S Turato G Badin C Bazzan E Beghe B Zuin R Calabrese F Casoni G Maestrelli P Papi A Fabbri LM Saetta M Neutrophilic infiltration within the airway smooth muscle in patients with COPD Thorax 2004 59 308 312 15047950 10.1136/thx.2003.012146
Saetta M Turato G Facchini FM Corbino L Lucchini RE Casoni G Maestrelli P Mapp CE Ciaccia A Fabbri LM Inflammatory cells in the bronchial glands of smokers with chronic bronchitis Am J Respir Crit Care Med 1997 156 1633 1639 9372687
Saetta M Baraldo S Corbino L Turato G Braccioni F Rea F Cavallesco G Tropeano G Mapp CE Maestrelli P Ciaccia A Fabbri LM CD8+ve cells in the lungs of smokers with chronic obstructive pulmonary disease Am J Respir Crit Care Med 1999 160 711 717 10430750
de Boer WI Sont JK van Schadewijk A Stolk J van Krieken JH Hiemstra PS Monocyte chemoattractant protein 1, interleukin 8, and chronic airways inflammation in COPD J Pathol 2000 190 619 626 10727989 10.1002/(SICI)1096-9896(200004)190:5<619::AID-PATH555>3.0.CO;2-6
Pesci A Balbi B Majori M Cacciani G Bertacco S Alciato P Donner CF Inflammatory cells and mediators in bronchial lavage of patients with chronic obstructive pulmonary disease Eur Respir J 1998 12 380 386 9727789 10.1183/09031936.98.12020380
Di Stefano A Capelli A Lusuardi M Balbo P Vecchio C Maestrelli P Mapp CE Fabbri LM Donner CF Saetta M Severity of airflow limitation is associated with severity of airway inflammation in smokers Am J Respir Crit Care Med 1998 158 1277 1285 9769292
Fournier M Lebargy F Le Roy LF Lenormand E Pariente R Intraepithelial T-lymphocyte subsets in the airways of normal subjects and of patients with chronic bronchitis Am Rev Respir Dis 1989 140 737 742 2528933
Lams BE Sousa AR Rees PJ Lee TH Subepithelial immunopathology of the large airways in smokers with and without chronic obstructive pulmonary disease Eur Respir J 2000 15 512 516 10759445 10.1034/j.1399-3003.2000.15.14.x
Rutgers SR Timens W Kauffman HF Postma DS Markers of active airway inflammation and remodelling in chronic obstructive pulmonary disease Clin Exp Allergy 2001 31 193 205 11251621 10.1046/j.1365-2222.2001.01004.x
Siafakas NM Vermeire P Pride NB Paoletti P Gibson J Howard P Yernault JC Decramer M Higenbottam T Postma DS et Optimal assessment and management of chronic obstructive pulmonary disease (COPD). The European Respiratory Society Task Force [see comments] Eur Respir J 1995 8 1398 1420 7489808 10.1183/09031936.95.08081398
van der Lende R Orie NG The MRC-ECCS questionnaire on respiratory symptoms (use in epidemiology) Scand J Respir Dis 1972 53 218 226 4640659
Global initiative for Chronic Obstructive Lung Disease NIH Definitions chapter 1 in Global Initiative for Chronic Obstructive Lung Disease GOLD 2003 5 10
Quanjer PH Tammeling GJ Cotes JE Pedersen OF Peslin R Yernault JC Lung volumes and forced ventilatory flows. Report Working Party Standardization of Lung Function Tests, European Community for Steel and Coal. Official Statement of the European Respiratory Society Eur Respir J Suppl 1993 16 5 40 8499054
Cockcroft DW Killian DN Mellon JJ Hargreave FE Bronchial reactivity to inhaled histamine: a method and clinical survey Clin Allergy 1977 7 235 243 908121
Rutgers SR Timens W Kaufmann HF van der Mark TW Koeter GH Postma DS Comparison of induced sputum with bronchial wash, bronchoalveolar lavage and bronchial biopsies in COPD Eur Respir J 2000 15 109 115 10678630
Aleva RM Kraan J Smith M ten Hacken NH Postma DS Timens W Techniques in human airway inflammation: quantity and morphology of bronchial biopsy specimens taken by forceps of three sizes Chest 1998 113 182 185 9440588
Sont JK de Boer WI van Schadewijk WA Grunberg K van Krieken JH Hiemstra PS Sterk PJ Fully automated assessment of inflammatory cell counts and cytokine expression in bronchial tissue Am J Respir Crit Care Med 2003 167 1496 1503 12770855 10.1164/rccm.2205003
Puddicombe SM Polosa R Richter A Krishna MT Howarth PH Holgate ST Davies DE Involvement of the epidermal growth factor receptor in epithelial repair in asthma FASEB J 2000 14 1362 1374 10877829 10.1096/fj.14.10.1362
Willemse BWM ten Hacken NHT Rutgers B Lesman-Leegte IGAT Timens W Postma DS Smoking cessation improves direct and indirect airway hyperresponsiveness in COPD Eur Respir J 2004 24 1 6 15293595 10.1183/09031936.04.00117603
Kuschner WG D'Alessandro A Wong H Blanc PD Dose-dependent cigarette smoking-related inflammatory responses in healthy adults Eur Respir J 1996 9 1989 1994 8902455 10.1183/09031936.96.09101989
Rutgers SR Postma DS ten Hacken NH Kauffman HF Der Mark TW Koeter GH Timens W Ongoing airway inflammation in patients with COPD who do not currently smoke Thorax 2000 55 12 18 10607796 10.1136/thorax.55.1.12
Chapman JT Otterbein LE Elias JA Choi AM Carbon monoxide attenuates aeroallergen-induced inflammation in mice Am J Physiol Lung Cell Mol Physiol 2001 281 L209 L216 11404264
Melgert BN Postma DS Geerlings M Luinge MA Klok PA Van Der Strate BW Kerstjens HA Timens W Hylkema MN Short-term smoke exposure attenuates ovalbumin-induced airway inflammation in allergic mice Am J Respir Cell Mol Biol 2004 30 880 885 14722223 10.1165/rcmb.2003-0178OC
Hulbert WC Walker DC Jackson A Hogg JC Airway permeability to horseradish peroxidase in guinea pigs: the repair phase after injury by cigarette smoke Am Rev Respir Dis 1981 123 320 326 7224343
Di Stefano A Capelli A Lusuardi M Caramori G Balbo P Ioli F Sacco S Gnemmi I Brun P Adcock IM Balbi B Barnes PJ Chung KF Donner CF Decreased T lymphocyte infiltration in bronchial biopsies of subjects with severe chronic obstructive pulmonary disease Clin Exp Allergy 2001 31 893 902 11422154 10.1046/j.1365-2222.2001.01098.x
Takanashi S Hasegawa Y Kanehira Y Yamamoto K Fujimoto K Satoh K Okamura K Interleukin-10 level in sputum is reduced in bronchial asthma, COPD and in smokers Eur Respir J 1999 14 309 314 10515406 10.1034/j.1399-3003.1999.14b12.x
Sun G Stacey MA Vittori E Marini M Bellini A Kleimberg J Mattoli S Cellular and molecular characteristics of inflammation in chronic bronchitis Eur J Clin Invest 1998 28 364 372 9650009 10.1046/j.1365-2362.1998.00285.x
Meyer KC Ershler W Rosenthal NS Lu XG Peterson K Immune dysregulation in the aging human lung Am J Respir Crit Care Med 1996 153 1072 1079 8630547
| 15850494 | PMC1140754 | CC BY | 2021-01-04 16:36:26 | no | Respir Res. 2005 Apr 25; 6(1):38 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-38 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646110.1371/journal.pmed.0020089Case ReportDiabetes/Endocrinology/MetabolismEndocrinologyAutoimmune Thyroid Disease with Fluctuating Thyroid Function Case ReportAlzahrani Ali S *Aldasouqi Saleh Abdel Salam Suzan Sultan Ali Department of Medicine, King Faisal Specialist Hospital and Research Centre, RiyadhSaudi Arabia
Competing Interests: The authors have declared that no competing interests exist.
Author Contributions: ASA managed the patient and wrote the manuscript. SA reviewed the manuscript and provided editorial assistance. SAS and AS reviewed the literature.
* To whom correspondence should be addressed. E-mail: [email protected] 2005 31 5 2005 2 5 e8925 10 2004 8 2 2005 Copyright: © 2005 Alzahrani 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.The authors describe the case of a woman with autoimmune thyroid disease, who presented with hypothyroidism but went on to develop hyperthyroidism.
==== Body
PRESENTATION OF CASE
A 38-y-old East Asian woman presented in 1985 with weight gain and cold intolerance. She was found to have a small goiter and an elevated thyroid-stimulating hormone (TSH) level of 58 mU/l (normal range, 0.5–5). She was diagnosed with primary hypothyroidism and was treated with levothyroxine (L-T4) 0.1 mg/d. Her symptoms promptly improved. Between 1985 and 1993, she felt generally well. Her adherence to treatment was variable, and when she was reviewed her TSH levels were raised on several occasions (in the range of 15–38 mU/l). In her past medical history, she had an ectopic pregnancy in 1990 and bilateral oophorectomy in 2000 for benign cystadenomas. There was no family history of thyroid or autoimmune diseases.
While the patient was on vacation in the summer of 1993, she stopped taking L-T4 altogether. Upon her return, thyroid function tests (TFTs) showed a TSH level of 0.46 mU/l, free thyroxine (FT4) at 15 pmol/l (normal range, 11–24), and thyroid hormone T3 at 1.9 nmol/l (normal range, 1.0–2.5). Antimicrosomal antibody titer was 1:400. Between June 1994 and November 1998, she continued to feel well on no treatment. TSH was measured annually and remained normal (range, 0.8–1.6 mU/l).
On 3 June 1999, the patient presented with palpitations, heat intolerance, excessive sweating, and anxiety. She also reported eye staring but no diplopia, proptosis (bulging eyes), or eye pain. She had no neck pain, dysphagia, or change in voice. She was clinically thyrotoxic with tachycardia, tremor, lid lag, and retraction. She had a symmetrical goiter, 40 g in size, that was firm, non-tender, and mobile with no distinct nodules or bruits. Hertel exophthalmometric measurements were 18 mm in both eyes (normal is up to 18 mm). The rest of the examination was unremarkable. TSH was less than 0.06 mU/l and FT4 was 31 pmol/l (Figure 1). Liver enzymes were normal. Radioiodine uptake was 41% (normal range, 25%–35%) at 24 h with homogeneous distribution of the tracer.
Figure 1 Alternating Thyroid Function over Time
Numbered time points: (1) initial diagnosis of hypothyroidism; (2) off L-T4 for 5 wk while on vacation; (3) presentation with Graves' disease; (4) I131 treatment; (5) severe hyperthyroidism 5 wk after I131 treatment; (6) a few days after treatment with propylthiouracil, propranalol, and Lugol's solution; (7) hypothyroidism about 4 mo after I131 treatment off all medications; (8) hypothyroidism confirmed with another set of TFTs and L-T4 started; (9) relapse of hyperthyroidism and L-T4 discontinued; (10) hyperthyroidism confirmed and methimazole started; (11) hypothyroidism developed and L-T4 resumed.
On 8 June 1999, the patient was started on methimazole 20 mg/d. By 2 July 1999, she was feeling well, and TFTs showed FT4 at 24 pmol/l, T3 at 2.2 nmol/l, and TSH at less than 0.06 mU/l. On 25 July 1999, she was treated with 14 mCi of I131. On 2 September 1999, the patient presented with severe symptoms of palpitations, heat intolerance, tremor, headache, nausea, and vomiting. Clinically, she was severely thyrotoxic, with a regular pulse of 110/min, and an oral temperature of 37.4 °C. She was irritable but fully conscious and oriented. She had lid lag and retraction and mild proptosis of the left eye (20 mm). The goiter was the same size (40 g) and consistency as during her presentation on 3 June1999, but it was now moderately tender. TFTs showed FT4 at more than 77 pmol/l, T3 at 6.4 nmol/l, and TSH at less than 0.06 mU/l. Methimazole was increased to 15 mg three times a day. Because of the severity of her hyperthyroidism, Lugol's solution was added at five drops, three times a day, for 14 d. The patient declined corticosteroids.
After a few days, she felt better and repeat FT4 was 42 pmol/l and TSH was less than 0.06 mU/l. Liver enzymes, however, were now raised: AST at 95 U/l, ALP at 372 U/l, and total bilirubin at 10 μmol/l. The methimazole was discontinued and replaced 2 wk later with propylthiouracil at a dose of 100 mg twice daily. The patient continued to improve over the following several weeks. Liver enzymes slowly normalized after about 2.5 mo. On 3 November 1999, the propylthiouracil was discontinued. TFTs done 2 and 4 wk later showed FT4 at 6 and 5 pmol/l and TSH at 18 and 30 mU/l, respectively. It was thought that this might be a transient hypothyroidism after I131 therapy or a permanent post-ablation hypothyroidism. The latter was considered more likely, and therefore the patient was started on L-T4. Five weeks later, she was doing well, and was clinically euthyroid with FT4 of 16 pmol/l and TSH of 0.8 mU/l.
In February 2000, the patient presented with symptoms of hyperthyroidism and her TSH was low, at 0.05 mU/l. L-T4 was discontinued, but she remained hyperthyroid 2 mo later. She was then started on methimazole 10 mg once daily. In May 2000, a repeat 24-h radioiodine uptake was elevated at 65%. The patient was continued on methimazole until December 2000, when the drug was discontinued because the patient was found to be euthyroid.
She remained euthyroid off all treatment for 2 mo, but in March 2001, she started to have symptoms of hypothyroidism and her TSH rose to 22 mU/l, with a FT4 of 9.2 pmol/l. She was started on L-T4 50 μg once daily, on which she remained euthyroid with repeatedly normal TFTs. Between March 2001 and December 2004, she remained euthyroid on L-T4 50 μg once daily (TSH ranged between 1.2 and 2.8 mU/l). She was last seen on 11 December 2004, when she was clinically and biochemically euthyroid with FT4 of 18.4 pmol/l and TSH of 2.8 mU/l.
DISCUSSION
The two main disorders that comprise autoimmune thyroid disease are Hashimoto thyroiditis and Graves disease. The former is the most common cause of hypothyroidism, whereas the latter is a major cause of hyperthyroidism. Occasionally, a patient may present with features of one of these disorders at one time and features of the other at another time. The usual sequence is hyperthyroidism followed by hypothyroidism [1]. However, cases of hypothyroidism followed by hyperthyroidism, such as in our patient, have also been described [2]. Our patient developed Graves disease after many years of hypothyroidism followed by a similar period of euthyroidism. The initial phase of hypothyroidism was secondary to Hashimoto thyroiditis since the patient had goiter and high antimicrosomal antibodies. The hypothyroidism did not represent a hypothyroid phase of an episode of thyroiditis since the patient repeatedly had elevated TSH levels over several years when she was not adhering to L-T4 therapy. Interestingly, following treatment with I131, she developed multiple alternating phases of hypothyroidism and hyperthyroidism (Figure 1). It is possible that the thyroiditis caused by I131 induced some immune reaction, with the formation of stimulating and inhibiting antibodies leading later to alternating phases of hypo- and hyperthyroidism.
Takasu et al. described eight cases of autoimmune thyroid disease with an alternating pattern of thyroid function [2]. In five cases, hypothyroidism was followed by transient hyperthyroidism. In two other cases, hypothyroidism was followed by persistent hyperthyroidism, and in one case hypothyroidism was associated with thyroid-stimulating antibodies, a characteristic finding in Graves disease. Tamai et al. reported the development of spontaneous hypothyroidism in 26 patients with Graves disease treated with antithyroidal drugs [1]. In another reported case, a patient underwent three cycles of transition from hypo- to hyperthyroidism and back to hypothyroidism, with corresponding changes in stimulating and blocking TSH receptor antibodies [3]. In our patient, we did not have serial measurements of stimulating and blocking thyroid antibodies. However, it is likely that the alternating thyroid function was associated with an alternating pattern of antibodies.
Recovery from hypothyroidism in chronic autoimmune thyroiditis is a rare but recognized phenomenon [4]. This is more likely to happen after iodide restriction in countries with high dietary iodide intake [5]. Predictors of spontaneous recovery from hypothyroidism in Hashimoto thyroiditis include the presence of goiter and high radioiodine uptake [5–7]. Spontaneous recovery also has been related to the disappearance of TSH-receptor-blocking antibodies [8].
Our patient had severe hyperthyroidism 5 wk after I131 treatment, while she was taking a relatively large dose of methimazole. This was most likely secondary to I131-induced thyroiditis as there was tenderness of the thyroid gland and FT4 increased proportionally much more than T3. The use of thionamide drugs to establish euthyroidism prior to I131 treatment is controversial, especially in a young patient without cardiovascular disease [9,10]. While rare, severe hyperthyroidism and even thyroid storm have been described following I131 treatment of Graves disease [11].
The patient also developed elevated liver enzymes. This could be secondary to the severe hyperthyroidism per se or due to the high dose of methimazole. Elevated liver enzymes are occasionally seen as a manifestation of hyperthyroidism and are also a well-recognized complication of antithyroid drugs [12–14].
Learning Points
• Rarely, autoimmune hyperthyroidism can occur after many years of hypothyroidism in patients with Hashimoto thyroiditis.
• When a patient on L-T4 for Hashimoto hypothyroidism presents with hyperthyroidism, over-replacement with L-T4 is the likely cause, but the possibility of endogenous hyperthyroidism should also be considered.
• Although radioactive iodine therapy for Graves disease is generally safe, especially in an otherwise healthy patient, severe exacerbations of hyperthyroidism can occur, and close monitoring in the first few weeks after therapy is prudent.
Citation: Alzahrani AS, Aldasouqi S, Salam SA, Sultan A (2005) Autoimmune thyroid disease with fluctuating thyroid function. PLoS Med 2(5): e89.
Abbreviations
FT4free thyroxine
L-T4levothyroxine
TFTthyroid function test
TSHthyroid-stimulating hormone
==== Refs
References
Tamai H Kasagi K Takaichi Y Takamatsu J Komaki G Development of spontaneous hypothyroidism in patients with Graves' disease treated with antithyroidal drugs: Clinical, immunological, and histological findings in 26 patients J Clin Endocrinol Metab 1989 69 49 53 2471711
Takasu N Yamada T Sato A Nakagawa M Komiya I Graves' disease following hypothyroidism due to Hashimoto's disease: Studies of eight cases Clin Endocrinol (Oxf) 1990 33 687 698 1982861
Kraiem Z Baron E Kahana L Sadeh O Sheinfeld M Changes in stimulating and blocking TSH receptor antibodies in a patient undergoing three cycles of transition from hypo to hyper-thyroidism and back to hypothyroidism Clin Endocrinol (Oxf) 1992 36 211 214 1637398
Nikolai TF Recovery of thyroid function in primary hypothyroidism Am J Med Sci 1989 297 18 21 2913797
Okamura K Sato K Ikenoue H Yoshinari M Nakagawa M Reevaluation of the thyroidal radioactive iodine uptake test, with special reference to reversible primary hypothyroidism with elevated thyroid radioiodine uptake J Clin Endocrinol Metab 1988 67 720 726 2843560
Comtois R Faucher L Lafleche L Outcome of hypothyroidism caused by Hashimoto's thyroiditis Arch Intern Med 1995 155 1404 1408 7794089
Takasu N Komiya I Asawa T Nagasawa Y Yamada T Test for recovery from hypothyroidism during thyroxine therapy in Hashimoto's thyroiditis Lancet 1990 336 1084 1086 1977978
Takasu N Yamada T Takasu M Komiya I Nagasawa Y Disappearance of thyrotropin-blocking antibodies and spontaneous recovery from hypothyroidism in autoimmune thyroiditis N Engl J Med 1992 326 513 518 1732791
Burch HB Solomon BL Wartofsky L Burman KD Discontinuing antithyroid drug therapy before ablation with radioiodine in Graves disease Ann Intern Med 1994 121 553 559 7521992
Burch HB Solomon BL Cooper DS Ferguson P Walpert N The effect of antithyroid drug pretreatment on acute changes in thyroid hormone levels after (131)I ablation for Graves' disease J Clin Endocrinol Metab 2001 86 3016 3021 11443161
McDermott MT Kidd GS Dodson LE Hofeldt FD Radioiodine-induced thyroid storm. Case report and literature review Am J Med 1983 75 353 359 6349350
Arab DM Malatjalian DA Rittmaster RS Severe cholestatic jaundice in uncomplicated hyperthyroidism treated with methimazole J Clin Endocrinol Metab 1995 80 1083 1085 7714072
Fischer MG Nayer HR Miller A Methimazole-induced jaundice JAMA 1973 223 1028 1029 4739295
Williams KV Nayak S Becker D Reyes J Burmeister LA Fifty years of experience with propylthiouracil-associated hepatotoxicity: What have we learned? J Clin Endocrinol Metab 1997 82 1727 1733 9177371
| 15916461 | PMC1140938 | CC BY | 2021-01-05 10:39:50 | no | PLoS Med. 2005 May 31; 2(5):e89 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020089 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591645610.1371/journal.pmed.0020107EssayOtherMedicine in Developing CountriesResearch MethodsEpidemiologyIs Evidence-Based Medicine Relevant to the Developing World? EssayChinnock Paul *Siegfried Nandi Clarke Mike Paul Chinnock is Managing Editor of the Cochrane Injuries Group (London, United Kingdom). Nandi Siegfried is a South African Nuffield Medical Fellow at the University of Oxford (Oxford, United Kingdom). Mike Clarke is Director of the United Kingdom Cochrane Centre (Oxford, United Kingdom). The authors are involved in the work of the Cochrane Collaboration, but the views expressed in this article are their own and are not necessarily those of the Cochrane Collaboration.
Competing Interests: The authors' employment depends upon the value and importance given to systematic reviews.
*To whom correspondence should be addressed: E-mail: [email protected] 2005 31 5 2005 2 5 e107Copyright: © 2005 Chinnock 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.Systematic reviews have yet to achieve their potential as a resource for practitioners in developing countries, argue Chinnock et al.
Systematic reviews have yet to achieve their potential as a resource for practitioners in developing countries
==== Body
Although there is still some resistance to the evidence-based medicine movement, evidence-based health care has now become widely accepted and adopted. Systematic reviews of the effectiveness of health care interventions are the engine room of evidence-based health care; much has been written about how these reviews should be conducted and what they can achieve [1,2]. If the case for the use of systematic reviews is good in developed countries—and we think it is—then the case is even stronger in the developing world. Wherever health care is provided and used, it is essential to know which interventions work, which do not work, and which are likely to be harmful. This is especially important in situations where health problems are severe and the scarcity of resources makes it vital that they are not wasted [3].
But are the systematic reviews that have so far been published relevant and of practical use to those who provide health care in “the majority world” (i.e., in developing countries? In our view, the relevance of systematic reviews to frontline health care workers in developing countries has so far been limited, for a number of reasons.
Reasons Why the Relevance Is Limited
Conditions.
Most of the reviews produced to date address health conditions that are priorities in the developed world [4]. Many major health concerns in developing nations have yet to be made the subject of a review, although there are signs that this may be changing [5]. The introductory discussions of most reviews focus on the impact of conditions in the United States and Western Europe. This may be an indication of the authors' own priorities and experience, or it may be because they have made assumptions about the priorities of journal editors and readers.
Health care practitioners in developing countries need the most appropriate evidence to guide their practice
(Photo: World Health Organization/P. Virot)
Interventions.
Health care professionals in developing countries sometimes wonder whether their reliance on older, cheaper, “lower-tech” approaches has made their practice quite distinct from that of their colleagues in richer regions [6]. Yet the authors of systematic reviews seem, by and large, to prefer to take on the task of assessing the evidence for more recent (and generally more expensive) technologies. This is not to say that reviewers should avoid high-tech interventions. Again, it is a question of setting priorities, and of recognising the urgent need for more reviews on interventions that are feasible in the majority world.
Exclusion of studies from the developing world.
Systematic reviews are based largely on research that has been done in rich countries. One of the reasons for this is the relative lack of research in developing countries. However, even when research has been conducted in these countries, it might not be published [7]—or if it is published, it might not be in a journal that is indexed in the widely used bibliographic databases such as MEDLINE and EMBASE. Thus, despite the best efforts of many reviewers, relevant studies may easily be missed. Excluding studies on the basis of language or region is generally not considered good practice in systematic reviewing [8], but the difficulties of identifying and assessing such studies can make finding them and including them in a review an unrealistic expectation.
Quality of studies from the developing world.
Once studies have been found, they are assessed for quality by the reviewers. Only when the quality meets the criteria specified in the review protocol (in most cases, this specifies randomised controlled trials only) are they included in the analysis. The difficulties of conducting randomized controlled trials in resource-poor situations result in the exclusion of many developing country studies. Some have suggested that the “quality threshold” should be lowered, so that more studies from developing countries can be included in systematic reviews. This question is contentious, and indeed divides the authors of this essay, but it needs to be recognised and debated openly.
Transferability.
Practitioners in low-income countries have questioned the “transferability” of evidence derived from studies conducted in richer nations [9]. The basis of their concern is their awareness that there can be many differences between patient populations and in the delivery of health care. Forjuoh et al. have pointed out that some injury prevention interventions will have broad transferability, while others will not [10]. They went on to make suggestions as to which intervention would be transferable, but they did so on theoretical grounds without any supporting data.
Features of the typical health care experience of a patient living in the developing world, as compared with features of the typical health care experience of a patient in a clinical trial in a developed country, are shown in Box 1.
Box 1. Comparison of the Health Care Experiences of Patients in the Less Developed and Developed Worlds
Features of the typical health care experience of a patient living in the less developed world include
late presentation
self-medication of “prescription” drugs or traditional treatments
poor facilities may delay diagnosis
referral (if needed) not easily arranged
if a child, may be malnourished
if a woman, may be anaemic
will experience problems because of shortages of trained staff
…and because of poor infection control
…and because of a lack of follow-up care
patient may be unable (e.g., because of lack of funds) to fully adhere to treatment.
Features of the typical health care experience of a patient in a clinical trial in a developed country include
none of the above
There are also important differences in the way in which care is delivered in developing and developed countries. In developing countries, treatments that would be delivered by doctors elsewhere are often delivered by medical assistants or clinical officers. This may or may not have an impact on the effectiveness of the treatment. Similarly, legislation can be considered a health care intervention for the prevention of road traffic injury, but the “delivery” of such legislation (i.e., its enforcement) is often harder to achieve in developing countries for a multitude of reasons.
As a result of such differences, the most effective treatment in a randomised controlled trial may not be the most effective treatment when provided in the developing world. Some treatments will retain much of their effectiveness in a resource-poor context; others will not.
One recently updated Cochrane review on the primary repair of penetrating colon injuries is a case in point [11]. The update involved the addition of data from one study, which had been completed since the original version of the review had been published. This addition introduced a much greater level of heterogeneity. The likely explanation for this, in the opinion of the reviewers, was that the new study was the only one in which the intervention had been applied in a developing country, which had imposed a number of limitations on its delivery.
Rather than implying that a review's conclusions are globally applicable, perhaps this is one of those circumstances where it would be more appropriate if reviewers concluded with statements such as, “There is evidence for the effectiveness of this intervention in the countries and setting where the included studies were conducted, and in places that are similar in terms of the resources available.”
What Can Be Done?
It is, of course, vital that more research of quality and relevance is conducted in developing countries, but the writers of systematic reviewers also have much to do. We need to find ways to make a good product better, and we must do more to make sure that people in the majority world are able to access the reviews that are published. In order for progress to be made, the following questions require more attention than they have received up to now.
Authors.
How can we involve more people from developing countries in the writing and peer reviewing of systematic reviews? For example, how can we continue to build on progress made on international activity within the Cochrane Collaboration [12] (see Table 1)?
Table 1 Number of People Actively Involved in Cochrane Collaborative Review Groups
Data used in this table are from [12].
Titles.
How can we get more reviews written on (a) health problems that are priorities, and (b) interventions that are affordable and feasible in the majority world?
Context.
Should reviews focus on specific contexts in relation to the location of the condition and the delivery of the intervention?
Background sections.
How can we encourage reviewers to look at conditions/interventions globally, and not just as they affect the United States and Western Europe?
Search for studies.
How can we make it easier to find and review data from research done in developing countries?
Analysis.
Should reviewers be encouraged to consider whether heterogeneity between study results might be due to differences in underlying resources?
Conclusions.
Should conclusions address whether any recommendations apply everywhere, or just in settings similar to those in which the included studies were done? Or is this beyond the recommendations of a review?
Dissemination of the findings of reviews.
Is this best done by circulating the reviews themselves, or are reviews merely a stage in the production of more accessible evidence-based health information materials? For example, the World Health Organization's Reproductive Health Library, available on CD-ROM, includes selected Cochrane reviews but also summaries and commentaries that have been specially prepared to provide a developing world perspective. The BMJ's Clinical Evidence produces other summaries of the evidence (for example, often integrating the findings of Cochrane Reviews into answers to clinical questions), and aims to prepare these in user-friendly formats and languages. Are more initiatives like these needed?
Research.
Research is needed on the impact of systematic reviews on practice in the developing world. We need to assess: What proportion of reviews are relevant to health care in low-resource settings? Are evidence-based sources used to set policy in different countries? How widely are the Cochrane Library and/or Cochrane reviews used by health care workers, and what are the barriers to use? How widely are these resources used by other people involved in decisions about health care, including patients, their carers, and policy makers? Has the use of Cochrane evidence influenced practice? What do these users and potential users think would make reviews more useful?
Conclusion
When so-called developing countries first gained freedom from their colonial oppressors, Ernst Schumacher pointed out that there was a need, not for the “best” technology, but for “appropriate” technology [13]. When it comes to health care, practitioners and patients of these countries need and deserve nothing less than the most “appropriate evidence”.
With thanks to Fred Bukachi, Luis Gabriel Cuervo, and Neil Pakenham-Walsh for their comments on previous drafts of this essay.
Citation: Chinnock P, Siegfried N, Clarke M (2005) Is evidence-based medicine relevant to the developing world? PLoS Med 2(5): e107.
==== Refs
References
Smith GD Egger M Meta-analysis: Unresolved issues and future developments BMJ 1998 316 221 225 9468698
Egger M Ebrahim S Smith GD Where now for meta-analysis? Int J Epidemiol 2002 31 1 5 11914281
Volmink J Swingler G Siegfried N Where to practise evidence-based medicine? Lancet 2001 357 724
Swingler GH Volmink J Ioannidis JPA Number of published systematic reviews and global burden of disease: Database analysis BMJ 2003 327 1083 1084 14604930
Waters E Doyle J Systematic reviews of public health in developing countries are in train BMJ 2004 328 585
Lowe M Evidence-based medicine—The view from Fiji Lancet 2000 356 1105 1107 11009161
Horton R Medical journals: Evidence of bias against the diseases of poverty Lancet 2003 361 712 713 12620731
Egger M Smith GD Meta-analysis: Bias in location and selection of studies BMJ 1998 316 61 66 9451274
Maharaj R Evidence-based medicine [Letter] Caribb Health 1999 2 2
Forjuoh SN Li G A review of successful transport and home injury interventions to guide developing countries Soc Sci Med 1996 43 1551 1560 8961399
Nelson R Singer M Primary repair for penetrating colon injuries Cochrane Database Syst Rev 2003 2003 CD002247
Allen C Clarke M International activity within collaborative review groups 2004 Ottawa The Cochrane Collaboration 102 In: 12th Cochrane Colloquium, 2004
Schumacher EF Small is beautiful: Economics as if people mattered 1973 London Blond and Briggs 288
| 15916456 | PMC1140939 | CC BY | 2021-01-05 11:13:38 | no | PLoS Med. 2005 May 31; 2(5):e107 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020107 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591645810.1371/journal.pmed.0020109PerspectivesEpidemiology/Public HealthHealth PolicyMedicine in Developing CountriesInternational healthResource allocation and rationingHealth Services Administration/ManagementTaking More Than a Fair Share? The Migration of Health Professionals from Poor to Rich Countries PerspectivesDovlo Delanyo Delanyo Dovlo is a physician and specialist in human resources for health based in Accra, Ghana. E-mail: [email protected]
Competing Interests: The author is from Ghana, one of the countries with high physician flows to the United States.
5 2005 31 5 2005 2 5 e109Copyright: © 2005 Delanyo Dovlo.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.The migration of health professionals from developing countries undermines the ability of these countries to meet the Millennium Development Goals.
==== Body
Brain Drain: A Global Health Problem
The international migration of health workers, especially of physicians and nurses but also increasingly of other health workers, has become a major global health concern. Recent meetings, such as the World Health Assembly of 2004 [1] and the High-Level Forum on the Millennium Development Goals in December 2004, as well as a number of publications have highlighted the severe shortage of health personnel in poorer parts of the world and the rise in demand for health workers in rich countries. The 2005 World Health Assembly, being held this month, is expected to discuss how to limit the adverse effects of the migration and to promote fairer recruitment tactics by developed countries as a follow-up to a resolution passed in 2004.
There is now considerable interest in measuring and managing the migratory flow of health workers; in seeking reparations, payments or remittances; and in training “substitute health workers”—groups who have taken on jobs, functions, and roles that are normally the tasks of internationally recognized professionals such as doctors, nurses, and pharmacists [2]. Indeed, words such as “slavery” and “human rights” underlie the debate's emotional underpinnings (such words have appeared, for example, in listserv discussions about health worker migration).
Though some migration occurs between rich countries (and also between poor countries), most of the migration of health professionals is occurring from countries with physician densities of about 17 per 100,000 population to countries with densities of 300 per 100,000 population (see p. 16 of [3]). This is a good example of the “inverse care law”—that countries that need the most health care resources are getting the least (Figure 1). Why does this migration occur when there appears to be a glut of physicians in the recipient countries? One of the reasons is that pay levels are up to 24 times higher in recipient countries than they are in source countries [4].
Figure 1 Global Variation in the Density of Health Workers
In a report by the Joint Learning Initiative, 186 countries were designated as having low, medium, and high worker density clusters (below 2.5, between 2.5 and 5.0, and above 5.0 workers per 1,000 population, respectively), with the low- and high-density clusters further subdivided according to high and low under-five mortality [9]. Among low-density countries, 45 are in the low-density/high-mortality cluster; these are predominantly sub-Saharan countries experiencing rising death rates and weak health systems. (Illustration: Giovanni Maki, adapted from [9])
Countries such as India, the Philippines, and Nigeria—highly populated countries that train large numbers of health professionals and have a long-standing remittance culture, in which professionals working outside the country send money back home to relatives—have produced doctors and nurses for “export.” However, demand for different kinds of health professionals can fluctuate. For example, recent rises in demand for nurses in the United States led to reports of immigrant doctors in Florida who took up nursing to get into the job market [5]. This doubly wastes the resources poor countries invest into training physicians; indeed, other anecdotes suggest that many immigrant physicians and nurses take on jobs completely unrelated to their training.
The Impact of Migration
For source countries with few physicians, the loss of even a single doctor often has a major impact on the health service [6]. In their analysis of the impact of the migration of health professionals, Martineau and colleagues state: “The ultimate losers tend to be health services (and their users) in the remoter rural areas, as they come lowest in the pecking order of people's preferred working location” [6]. And since it is the poorest citizens who live in the remoter areas, say the authors, it is they who are affected most by migration.
Often, health professionals leave to undertake training only obtainable (or seductively offered as “scholarships”) in rich recipient countries. Several years later, the metamorphosis from student to migrant is complete, but the migrant professionals may remain on payrolls in their home countries for several years. Thus, barely affordable initiatives towards capacity-building result instead in further losses of capacity.
Policy-makers in sub-Saharan Africa must feel helpless when they are completely unable to match either the remuneration or the working conditions found in recipient countries. A self-sustaining cycle ensues: as more and more physicians migrate from a country, they create an environment (a “home from home”) that entices newer migrants from the same country. The helplessness is reinforced by a lack of information on workforce losses for policy-making.
A New Study on Physician Migration
A new study by Hagopian and colleagues analyzes the numbers, characteristics, and trends in the migration to the United States of physicians trained in sub-Saharan Africa [7]. The US is perhaps the world's largest “consumer” of health workers from the developing world.
The authors did a cross-sectional study using the 2002 American Medical Association Physician Masterfile, which contains detailed information on all 771,491 active physicians who were licensed to practice medicine in the year 2002 (excluding those physicians employed by federal entities such as the Veterans Administration, federal prisons, or the military). The authors reviewed these data for all physicians in the US who received their training in sub-Saharan Africa.
What they found was that more than 23% of America's 771,491 physicians received their medical training outside the country, mostly (64%) in low-income or lower-middle-income countries. A total of 5,334 physicians from sub-Saharan Africa are in that group, a number that represents more than 6% of the physicians practicing in sub-Saharan Africa now. Nearly 86% of these Africans practicing in the US originate from only three countries: Nigeria, South Africa, and Ghana. Of these, 79% were trained at just ten medical schools.
Implications of the Study
This study suggests that selection of particular types of medical graduates is occurring, either for their training or language skills or, as mentioned earlier, by the presence of previous migrants from the same schools. Hagopian and colleagues state that there are several federal agencies and state health departments who “sponsor” physicians who have completed their residency training in the US on a “student” (J-1) visa. These sponsorships allow physicians who are foreign nationals to gain approval from the State Department and US Citizenship and Immigration Services to waive J-1 visa requirements that would otherwise require them to return to home countries for at least two years. In exchange for this waiver, physicians find employment with a health agency or private physician in a health professional shortage area. Such waivers raise the possibility of deliberate selection of the best graduates.
A common argument against any restriction of migration of health workers is that the right to live anywhere one chooses is a basic human right. This argument is particularly used by recipient countries. But even when we acknowledge such human rights concerns, the study by Hagopian and colleagues rightly notes that the migration tide subverts the development potential of poor countries. Migration further undermines the right of the people who contributed to the training of their country's health professionals to benefit directly from their investment.
The study also alludes to a lack of data about migration from poorer countries, such that accurate data on who has left, and to which countries they emigrated, are more easily obtained in the recipient country. Departures from the workforce need to be more properly documented in source countries. This documentation may be particularly difficult for countries that pay “ghost workers” (names on a payroll for employees that do not actually exist) [8], obscuring the real losses from the health system.
Physician shortages exist worldwide, but this study shows that the numbers of medical schools are very low in sub-Saharan Africa. Conventional medical training in tertiary hospitals is very expensive. Yet, instead of training new types of workers to match local needs, many countries in sub-Saharan Africa continue to adhere to such training out of professional pride—they are reluctant to been seen to be using less than the “best” type of health worker. Furthermore, professionals in these countries wish to retain reciprocal recognition of their qualifications by the developed world—another obstacle to training new types of workers, who may not be internationally recognized. Professional associations and regulatory councils have resisted or limited the introduction of substitute health workers. Production of enrolled and auxiliary nurses, for example, was banned in some of the poorest countries (such as Zambia and Ghana), ostensibly to enhance the status of the professions even as health and economic indices receded and remaining professionals fled.
International Arrangements to Manage Flows?
Although there is a moral argument that richer recipient countries should help to mitigate migration, this argument is usually derided even by source countries, perhaps because they are powerless and are often dependent on the same recipient countries for aid.
Unlike the United Kingdom, the US has a multifaceted health system, which makes it difficult to imagine a situation in which it is feasible to agree on control of physician migration from countries that have little political or economic clout. Such source countries have little influence when it comes to negotiating over the poaching of their health professionals by powerful countries such as the US.
The reality, though, is that facilitating the migration of health workers from poor countries contributes to worse health outcomes in these countries. Human Resources for Health: Overcoming the Crisis [9], a strategic report of the Joint Learning Initiative (a consortium of over 100 leaders in health), recently analyzed the global workforce [9]. The analysis considered the impact that the global distribution of health workers will have on reaching the health-related Millennium Development Goals. The report suggested that the low health worker density in some countries has already had a major impact on maternal and child mortality (Figure 2) [9]. For example, the report states: “the prospects for achieving 80 percent coverage of measles immunization and skilled attendants at birth are greatly enhanced where worker density exceeds 2.5 workers per 1,000 population. Seventy-five countries with 2.5 billion people are below this minimum threshold.” The report suggests that low health worker density has a particularly marked effect on maternal deaths: a 10% increase in the density of the health workforce is correlated with about a 5% decline in maternal mortality [9]. This strong effect of worker density on maternal health may be due to the fact that highly trained personnel are essential for emergency obstetric services.
Figure 2 Association between Mortality and Health Worker Density
(Illustration: Giovanni Maki, adapted from [9])
Thus, policies and actions that reduce medical and nursing school intake in poor countries while facilitating entry visas into rich countries for physicians and nurses from these same poor countries may be responsible for the deaths of thousands of African children and women. Moral arguments must therefore be used to create policies that moderate the loss of trained health workers from poor countries and stop the medical training subsidies they make to rich countries.
Recruiting countries are much better equipped than source countries to increase their own physician supply and moderate the level of migration. Discussions and debates about health worker migration have, however, also pointed to poor governance and management in source countries as major “push” factors. Strong “pushes” will still exist, even if richer countries control health worker migration in the same way as they have tried to control general immigration.
There are many different policy options for tackling the migration of health professionals', but what works for one country may not work for another, and indeed, what works for nurses may not work for physicians. For example, a Jamaican nursing agency opened an office in Ghana in the 1990s, and Ghana was able to negotiate and agree on the numbers of Ghanaian nurses that Jamaica could recruit—which mitigated the effect of migration. In contrast, doctors tend to seek jobs abroad independently.
Effective international agreements on managing recruitment seem only to work when both source and recipient countries are developing countries. For example, South Africa has been successful at banning recruitment from within Africa, but richer countries opt for voluntary “codes of conduct” that are often quite ineffective [10].
Conclusion
The migration of physicians and other trained health professionals undermines the ability of developing countries to meet agreed Millennium Development Goals and creates untenable health conditions for the poorer sections of their populations. Developing countries on their own cannot achieve effective moderation of migration and secure the integrity of health services without the cooperation and collaboration of the countries that receive their health workers. An international regimen is needed to manage and moderate the migration of health workers in order to minimize the deleterious effects this has on underdeveloped countries.
Developing countries, on the other hand, need to evolve strategies that reflect their internal needs. This must include designing cadres of health professionals that are trained mainly for the purposes of local needs and are less prone to the attractions of migration.
Countries have different experiences, and each country must develop strategies that reflect the needs of their particular situation. However, the appropriate international environment for managing human resources is necessary if the strategies of developing countries are to achieve meaningful results.
Citation: Dovlo D (2005) Taking more than a fair share? The migration of health professionals from poor to rich countries. PLoS Med 2(5): e109.
==== Refs
References
Fifty-Seventh World Health Assembly (WHA57.19) International migration of health personnel: A challenge for health systems in developing countries 2004 Available: http://www.who.int/gb/ebwha/pdf_files/WHA57/A57_R19-en.pdf . Accessed 3 March 2005
Dovlo D Using mid-level cadres as substitutes for internationally mobile health professionals in Africa A desk review 2004 Available: http://www.human-resources-health.com/content/2/1/7 . Accessed 3 March 2005
ICN The global shortage of registered nurses – An overview of issues and actions 2004 Available: http://www.icn.ch/global/shortage.pdf . Accessed 3 March 2005
Vujicic M Zurn P Diallo P Adams O Dal Poz M The role of wages in slowing the migration of health care professionals from developing countries 2004 Available: http://www.human-resources-health.com/content/2/1/3 . Accessed 3 March 2005
McVicar N Foreign physicians take up nursing as they resettle in U.S 2002 Available: http://www.centredaily.com/mld/centredaily/news/4642809.htm . Accessed 3 March 2005
Martineau T Decker K Briefing note on international migration of health professionals: Levelling the playing field for developing country health systems 2002 Available: http://www.liv.ac.uk/lstm/research/documents/InternationalMigrationBriefNote.pdf . Accessed 3 March 2005
Hagopian A Thompson MJ Fordyce M Johnson KE Hart LG The migration of physicians from sub-Saharan Africa to the United States of America: Measures of the African brain drain 2004 Available: http://www.human-resources-health.com/content/2/1/17 . Accessed 3 March 2005
Safo A Ghost names killing Ghana softly 2002 Available: http://www.newsfromafrica.org/newsfromafrica/articles/art_7901.html . Accessed 3 March 2005
Joint Learning Initiative Human resources for health – Overcoming the crisis 2004 Available: http://www.globalhealthtrust.org/Report.html . Accessed 3 March 2005
Commonwealth Secretariat, London Commonwealth code of practice for the international recruitment of health workers 2003 Available: http://www.thecommonwealth.org/shared_asp_files/uploadedfiles/%7B7BDD970B-53AE-441D-81DB-1B64C37E992A%7D_CommonwealthCodeofPractice.pdf . Accessed 3 March 2005
| 15916458 | PMC1140940 | CC BY | 2021-01-05 10:39:46 | no | PLoS Med. 2005 May 31; 2(5):e109 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020109 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646210.1371/journal.pmed.0020122Health in ActionInfectious DiseasesEpidemiology/Public HealthHealth PolicyHIV/AIDSTuberculosisHIV Infection/AIDSMedicine in Developing CountriesPublic HealthForensic MedicineTackling Tuberculosis in Latvia Health in ActionZipperer Melanie Melanie Zipperer is Communication Adviser for HIV/AIDS, Tuberculosis, and Malaria at the World Health Organization in Geneva, Switzerland. E-mail: [email protected]
Competing Interests: Melanie Zipperer is employed by the World Health Organization. The World Health Organization fully supports DOTS, the internationally recommended tuberculosis control strategy, and provides technical assistance to its member states to implement DOTS.
5 2005 31 5 2005 2 5 e122Copyright: © 2005 Melanie Zipperer.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.After the fall of the Soviet Union, Latvia faced rising unemployment, poverty, and alcoholism, conditions that fueled a surge in TB. Zipperer discusses how Latvian health professionals are tackling the problem.
==== Body
With a forceful gesture, the delicate blond unlocks the heavy door of a cell in Riga's Central Prison clinic (Figure 1). Her name is Dr. Inga Nagele. She is the head of the clinic's tuberculosis (TB) unit. She opens the door to check on three female prisoners who have just been admitted. The conditions in the cell are bleak. Washing hangs on a line that stretches from one wall to the other. The bars on the small window hardly let any light into the room, and a bucket in the corner serves as a toilet. All the women were infected with TB at the prison.
Corridor of the clinic in Riga's Central Prison
(Photo: Melanie Zipperer)
“Living conditions in Latvia's prisons are miserable. This is ideal for a relentless spread of the life-threatening lung bacteria and other communicable diseases”, says Dr. Nagele. Because of permanent overcrowding, poor hygiene, and frequent relocation of prisoners within Latvia's 15 prisons, two-thirds of inmates infected with TB developed the disease once they were locked up, she says.
Latvia has 352 prisoners per 100 000 individuals in the population [1]. This number makes it one of the countries with the highest proportion of prisoners in the European Union, according to Mihails Azarenko, Chief of Riga's Central Prison. It is particularly high when compared with Denmark, for example, which has only 64 prisoners per 100 000 individuals [1].
Rising TB Rate after the Breakdown of the Soviet System
Outside Latvia's prisons, the conditions were just as favourable for an unchecked spread of infectious diseases, and TB in particular. The fall of the Soviet Union in 1991, and with it the collapse of a centralised public-health system, left the population in the newly independent states without proper access to health care. Latvia struggled against soaring unemployment, poverty, and alcoholism. These transition problems led to a rapid increase of TB and the development of multidrug-resistant TB (MDR-TB) strains.
MDR-TB is a form of the disease resistant to at least two of the most powerful conventional antibiotics that can treat TB (rifampicin and isoniazid). It develops when public-health programs fail to deliver regular, reliable treatment to patients. While the common strain of TB can be cured at a cost of US$10 for a six-month treatment course, MDR-TB can cost 100 times as much and takes two years to treat.
According to the World Health Organization's (WHO's) “Global Tuberculosis Control” report [2], reported notifications of TB in Latvia peaked in 1998 with 90 cases per 100 000 individuals. This made Latvia the country with the highest TB incidence rate in the European Union, compared with an average rate of 13 per 100 000 in western Europe. It was also considerably higher than in the other two Baltic states, Estonia (48 per 100 000) and Lithuania (70 per 100 000). Rates of MDR-TB in Latvia are among the highest in the world. According to a report by the WHO and the International Union Against Tuberculosis and Lung Disease, patients in parts of eastern Europe and central Asia are ten times more likely to have MDR-TB than in the rest of the world [3].
Rinalds Mucins, former Health Minister of Latvia, sees the main reason for Latvia's health problems as the difficult transition process of the health system. He was part of a government that resigned on 28 October 2004, when parliament rejected the prime minister's draft budget. “Over the past ten years we have tried every health system from Semashko to Bismarck. We changed our system too many times and too rapidly and couldn't provide patients and doctors with any stability”, he says. “Now we are working on some long-term programs; this is definitely a good start.”
Implementing DOTS and DOTS-Plus
This good start has already yielded positive results. Since 1995, Latvia has begun to make great strides in reducing the threat of TB and has started to rigorously implement DOTS, the internationally recommended strategy for controlling TB (see Sidebar).
The government backed this up with a strong political and financial commitment, resulting in TB cure rates rising from 64% in 1997 to 73% in 2002, according to the WHO European Regional Office centralised information system for infectious diseases. As a response to the serious MDR-TB situation, the DOTS-Plus project was introduced a couple of years later in close collaboration with the WHO and the United States Centers for Disease Control and Prevention. The DOTS-Plus project was also approved by the Green Light Committee. This committee is the technical review panel of the Green Light Committee mechanism, the WHO's mechanism to enable access to quality-assured second-line drugs at a reduced cost and technical assistance in the management of MDR-TB [4].
DOTS-Plus targets MDR-TB and builds on the elements of the DOTS strategy. However, it also takes into account specific issues, such as the use of second-line drugs, that need to be addressed in areas with a significant prevalence of MDR-TB. “DOTS-Plus has been very effective so far. According to the State Centre of Tuberculosis and Lung Diseases the number of MDR-TB cases decreased by 50%, from 335 cases in 1997 to 166 in 2003”, says Dr. Aiga Rurane, head of the WHO Liaison Office in Riga, “However, much more remains to be done to fully control the spread of MDR-TB.”
In 1997, TB control in prisons became part of Latvia's National TB Control Program. Since then, all prisoners undergo regular examinations and sputum tests. With an 85.9% cure rate, the number of patients with TB who are cured in prisons is now higher than outside. According to official data provided by the State Centre of Tuberculosis and Lung Diseases (http://www.tuberculosis.lv), the number of cases of MDR-TB in prisons was reduced from 19 in 2000 to nine in 2003 [5]. “The reason for these good cure rates is that patients are under permanent surveillance and must swallow the prescribed drugs,” says Dr. Nagele. “It also helps that they get a warm meal and regular sleeping hours.”
Dr. Gunta Dravniece, head of the MDR-TB ward in Riga's State Agency for TB and Lung Diseases, believes that controlling TB inside prisons is key to preventing the spread of the disease in the rest of the country. The State Agency for TB and Lung Diseases is located in a forest, a half-hour drive outside Riga. It was built in the 1950s by prisoners and designed as an isolation hospital for Riga's patients with TB. It has separate sections for patients with TB and MDR-TB.
At least one-third of Dr. Dravniece's patients in the MDR-TB ward are ex-prisoners. The others are often socially disadvantaged, homeless, and suffering from chronic alcoholism. “Many of them don't have even a passport,” she says. “When they get infected with MDR-TB, they want to stay as long as possible here at the state agency clinic, because they have a bed and something to eat.”
Long-term treatment, which is a necessity for MDR-TB, is difficult for both patients and health workers. Patients have to take up to 25 pills every day. Nurse Valentina carefully watches that each of the 55 patients with MDR-TB in the ward takes them. “They don't like these drugs, because often the pills provoke sickness and are difficult to digest. So I have to explain to them over and over why it is so important to take the medication regularly”, she explains.
Once patients are not infectious anymore, they can leave the ward and go back to their hometowns. There they receive almost daily follow-up treatment by local doctors and nurses. “Patients who come for their medical checks get food and are reimbursed for transportation. This is cheaper for the public sector than the hospital cost and motivates them to come back”, says Dr. Dravniece.
Training Centre to Disseminate Lessons Learned
The state agency also includes a training centre. It offers research and training courses to TB experts from eastern Europe and central Asia. Since 2000, the training centre has held countless courses for program managers, social workers, nurses, and laboratory staff across Europe. Training modules focus on key issues such as improvement of laboratory diagnostic methods, patient education, and MDR-TB in prisons. So far 156 specialists have been trained.
In November 2004, the training centre was officially awarded the status of a WHO Collaborating Centre, the first one on MDR-TB in the European region. “We accumulated a lot of experience in the past years while implementing DOTS and DOTS-Plus”, says Dr. Vaira Leimane, director of the training centre. “We received a lot of help from donors. Now it is time for us to give back. We want to show other countries with similar problems what we learned, so that they can improve their capacity to effectively treat and control the disease.”
Co-Infection with TB and HIV
However, the battle is not won yet. A new public-health challenge is posed by the increasing number of patients with HIV/TB co-infection. Eastern Europe and central Asia are home to some of the fastest-growing epidemics, mostly among young people. People who are HIV-positive are more likely to become sick with TB and MDR-TB than others. In Latvia, according to the State Centre of TB and Lung Diseases, the number of patients co-infected with HIV and TB has risen from one in 1994 to 40 in 2003.
Worldwide, in 2000, 11% of all new cases of TB in adults occurred in people infected with HIV, and 9% of the cases were directly attributable to HIV. The contribution of TB to AIDS-related deaths is substantial: 11% of AIDS-related deaths occur primarily from TB [5].
Dr. Jack C. Chow, the WHO's Assistant Director-General for HIV/ AIDS, TB, and Malaria, stresses that TB and HIV control programs must collaborate: “There are many ways how TB and HIV programs can closely work together. HIV-positive people can be given prophylactic treatment to prevent development of active TB disease. TB patients can be offered an HIV test.”
Without prompt action in this direction, years of Latvia's successful TB control efforts could be undermined.
DOTS and DOTS-Plus
DOTS (for “directly observed treatment, short course”) is the internationally recommended TB control strategy that combines five elements: sustained political commitment; access to quality-assured TB sputum microscopy; standardized short-course chemotherapy to all cases of TB under proper case-management conditions; uninterrupted supply of quality-assured drugs; and a recording and reporting system enabling outcome assessment.
DOTS-Plus takes into account specific additional issues, such as the use of second-line drugs, that need to be addressed in areas where there is high prevalence of MDR-TB. DOTS-Plus works as a supplement to the standard DOTS strategy. By definition, it is impossible to conduct DOTS-Plus in an area without having an effective DOTS-based TB control program in place.
Eva Nathanson, Technical Officer at the WHO's STOP TB Department, helped to find references and prevalence data for this article.
Citation: Zipperer M (2005) Tackling tuberculosis in Latvia. PLoS Med 2(5): e122.
Abbreviations
MDR-TBmultidrug-resistant tuberculosis
TBtuberculosis
WHOWorld Health Organization
==== Refs
References
United Kingdom's Home Office, Research, Development and Statistics Directorate World prison population list, 5th edition 2003 Available: http://www.homeoffice.gov.uk/rds/pdfs2/r234.pdf . Accessed 28 March 2005
World Health Organization Global tuberculosis control—Surveillance, planning, financing WHO/HTM/TB/2004.331 2004 Geneva World Health Organization Available: http://www.who.int/tb/publications/global_report/2004/en/ . Accessed 28 March 2005
WHO/IUATLD Global Project on Anti-Tuberculosis Drug Resistance Surveillance Anti-tuberculosis drug resistance in the world—Third global report 2004 Geneva World Health Organization Available: http://www.who.int/tb/publications/who_htm_tb_2004_343/en/ . Accessed 29 March 2005
Gupta R Cegielski JP Espinal MA Henkens M Kim JY Increasing transparency in partnerships for health—Introducing the Green Light Committee Trop Med Int Health 2002 7 970 976 12390604
Corbett EL Watt CJ Walker N Maher D Williams BG The growing burden of tuberculosis: Global trends and interactions with the HIV epidemic Arch Intern Med 2003 163 1009 1021 12742798
| 15916462 | PMC1140941 | CC BY | 2021-01-05 10:39:47 | no | PLoS Med. 2005 May 31; 2(5):e122 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020122 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646310.1371/journal.pmed.0020123Research ArticlePharmacology/Drug DiscoveryCardiology/Cardiac SurgeryCardiovascular MedicinePharmacology and toxicologyNational Trends in Statin Use by Coronary Heart Disease Risk Category Statin Use by CHD RiskMa Jun
1
Sehgal Niraj L
1
Ayanian John Z
2
Stafford Randall S
1
*1Stanford Prevention Research Center, Stanford University School of MedicineStanford, CaliforniaUnited States of America2Division of General Medicine, Brigham and Women's Hospital and Department of Health Care PolicyHarvard Medical School, Boston, MassachusettsUnited States of AmericaColhoun Helen Academic EditorUniversity College DublinIreland
Competing Interests: The authors have declared that no competing interests exist.
Author Contributions: JM, NLS, and RSS designed the study. JM, NLS, and RSS analyzed the data. JM, JZA, NLS, and RSS contributed to writing the paper.
*To whom correspondence should be addressed. E-mail: [email protected] 2005 31 5 2005 2 5 e12317 9 2004 4 3 2005 Copyright: © 2005 Ma 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.
Managing Cardiovascular Risk Factors: The Gap between Evidence and Practice
Patterns of Statin Prescribing
Background
Only limited research tracks United States trends in the use of statins recorded during outpatient visits, particularly use by patients at moderate to high cardiovascular risk.
Methods and Findings
Data collected between 1992 and 2002 in two federally administered surveys provided national estimates of statin use among ambulatory patients, stratified by coronary heart disease risk based on risk factor counting and clinical diagnoses. Statin use grew from 47% of all lipid-lowering medications in 1992 to 87% in 2002, with atorvastatin being the leading medication in 2002. Statin use by patients with hyperlipidemia, as recorded by the number of patient visits, increased significantly from 9% of patient visits in 1992 to 49% in 2000 but then declined to 36% in 2002. Absolute increases in the rate of statin use were greatest for high-risk patients, from 4% of patient visits in 1992 to 19% in 2002. Use among moderate-risk patients increased from 2% of patient visits in 1992 to 14% in 1999 but showed no continued growth subsequently. In 2002, 1 y after the release of the Adult Treatment Panel III recommendations, treatment gaps in statin use were detected for more than 50% of outpatient visits by moderate- and high-risk patients with reported hyperlipidemia. Lower statin use was independently associated with younger patient age, female gender, African American race (versus non-Hispanic white), and non-cardiologist care.
Conclusion
Despite notable improvements in the past decade, clinical practice fails to institute recommended statin therapy during many ambulatory visits of patients at moderate-to-high cardiovascular risk. Innovative approaches are needed to promote appropriate, more aggressive statin use for eligible patients.
The prescribing of statins shows great variation among different types of patients, with many who would benefit receiving neither lifestyle counseling nor appropriate drugs.
==== Body
Introduction
Coronary heart disease (CHD) remains the leading cause of morbidity and mortality in the United States and is associated with substantial economic cost [1]. Hyperlipidemia represents an important modifiable risk factor in the development and progression of CHD. Estimates indicate that nearly 100 million American adults have total blood cholesterol levels of greater than 5.17 mmol/l (200 mg/dl) with 40% having levels greater than 6.21 mmol/l (240 mg/dl) [2]. Identification and treatment of patients with hyperlipidemia play an essential role in the primary and secondary prevention of CHD.
Currently, evidence-based practice guidelines focus on low-density lipoprotein cholesterol (LDL-C) as the primary target for risk reduction therapy and recommend that the intensity and target goals of LDL-C-lowering therapy should be adjusted to individual absolute risk for CHD [3]. Absolute CHD risk is categorized as low, moderate, or high based on the presence or absence of CHD, CHD-equivalent conditions, and major risk factors other than LDL-C. While therapeutic lifestyle changes are integral to general risk reduction, drug treatment proves necessary for selected patients whose absolute risk is high and/or whose LDL-C is inadequately controlled with lifestyle modifications alone. Among existing drug therapies, 3-hydroxy-3-methyl-glutaryl coenzyme A reductase inhibitors, more commonly known as statins, provide a generally well-tolerated and effective option for lowering LDL-C levels and decreasing the likelihood of subsequent CHD events [3,4].
Despite the compelling evidence of statins' therapeutic benefits, the literature abounds with documentation of wide treatment gaps in clinical practice [5–10]. Available research, however, offers only a limited understanding of how statin therapy varies by CHD risk, particularly for statin-eligible patients in the moderate-risk group. Also, national data are limited regarding recent changes in statin use.
Using serial cross-sectional data from 1992 through 2002, we tracked trends in statin use in the United States during ambulatory visits categorized by CHD risk, with or without a diagnosis of hyperlipidemia. In addition, we analyzed the independent associations of patient and physician characteristics with statin use for insights as to how to target interventions to improve statin use.
Methods
Data Sources
Annual data from1992 through 2002 were obtained from the National Ambulatory Medical Care Survey (NAMCS) and the outpatient department component of the National Hospital Ambulatory Medical Care Survey (NHAMCS). The National Center for Health Statistics provides complete descriptions of both surveys and yearly data at http://www.cdc.gov/nchs/about/major/ahcd/ahcd1.htm. These surveys, particularly NAMCS, have been validated against other data sources [11,12], and have also been utilized in past research of cholesterol management [13].
In brief, NAMCS captures health-care services provided by office-based physicians, while NHAMCS assesses services offered at hospital outpatient departments. Both surveys utilize multistage probability sampling procedures, enabling the generation of nationally representative estimates. Between 1992 and 2002, annual participation rates among physicians selected for NAMCS averaged 70%, while the participation rate in NHAMCS by selected hospitals with outpatient departments was 90%. In our study, we combined NAMCS and NHAMCS data to obtain a wider range of outpatient settings and a broader socioeconomic spectrum of patients seeking ambulatory care.
Standard encounter forms were completed for a systematic random sample of patient visits during randomly assigned reporting periods. Item nonresponse rates were mostly 5% or less in both surveys for all years. Yearly encounter forms varied slightly between NAMCS and NHAMCS and were revised every two years. Our analysis focused on domains of data that were consistently collected in both NAMCS and NHAMCS for the time period 1992–2002, including patient demographic and geography characteristics, reasons for visit (up to three), diagnoses (up to three), new and continuing medications (up to five in 1992–1994 and six in 1995–2002), and lifestyle counseling services provided or ordered at the visit.
Participants
CHD risk categorization
We estimated CHD risk for adults aged 20 y and older based on risk factor counting. CHD risk was mutually exclusively categorized as low (0–1 risk factors), moderate (2+ risk factors), or high (CHD, other atherosclerotic diseases, or diabetes). The moderate-risk group included visits by patients without CHD or equivalent but with at least two of the following risk factors: age (for men, >45 y; for women, >55 y), cigarette smoking, or a physician-reported diagnosis of hypertension. Unfortunately, the other two major CHD risk factors—high-density lipoprotein cholesterol levels and family history of premature CHD—were not captured in either data source. Also, neither data source provided actual cholesterol measurements. Disease conditions were identified by International Classification of Disease (ICD-9) codes, as well as by the appropriate reason-for-visit codes that are specific to NAMCS and NHAMCS. For instance, we identified patients as having hyperlipidemia if their encounter forms contained an International Classification of Disease code within 272.0–272.4. For the sake of this study, patients whose encounter forms did not indicate the presence of a condition were assumed to not have that condition.
Patient visit characteristics.
Nonclincal characteristics included patient age, gender, race/ethnicity, medical insurance, visit status, United States census region, metropolitan area status, physician specialty, and practice setting. Medical insurance was classified as private/commercial, public (i.e., Medicare and Medicaid), or other (e.g., workers' compensation or self-pay). Visit status distinguished first-time visits from return visits to a practice. Physician specialty was available only from NAMCS, which contributed more than 90% of the total visits for each of the study years. We categorized physician specialties as cardiology, internal medicine, general and family practice, or other.
Measures
Of primary interest were the rate of statin use relative to CHD risk and the relationship of statin use to patient visit characteristics. The rate of statin use was calculated as the proportion of patient visits where a statin was reported (i.e., atorvastatin, lovastatin, pravastatin, simvastatin, or fluvastatin). Before its removal from the market in 2001, cerivastatin was used scarcely (<2% among visits by patients with hyperlipidemia) and therefore is not reported in this study. Measuring the rate of statin use by CHD risk category provided a relative indicator of appropriate prescribing patterns, that is, the prevalence of statin use should be highest among high-risk patients, for whom secondary prevention is a priority. Variations of statin use by patient visit characteristics, if detected, would reflect a lack of equity in processes of care in that uniform practices are expected unless evidence-based guidelines recommend otherwise.
Analyses
Statistical analyses were performed using SAS for Windows software (SAS Institute, Cary, North Carolina, United States) and SAS-callable SUDAAN software (RTI, Research Triangle Park, North Carolina, United States) to account for sampling weights and the complex survey design. The unit of analysis is the patient visit. We report national annual means of the rate of statin use by CHD risk category and corresponding 99% confidence intervals for the years 1992 through 2002. χ2 tests examined the association of statin use with individual patient visit characteristics for combined 1995–2002 NAMCS and NHAMCS data. The independent effect of each patient visit characteristic on statin use after controlling for all other characteristics was assessed with multivariate logistic regression.
Results
In 2002, visits by patients at moderate or high risk involved higher proportions of older patients (mean age 65 y) than low-risk patient visits (mean age 51 y), and consequently were more likely to be covered by public insurance, particularly Medicaid (Table 1). Moderate- and high-risk patient visits also were made up of more men and return patients. In addition, a greater percentage of high-risk patient visits (11%) were seen by cardiologists than patient visits at low and moderate risk (2% and 4%, respectively). Internists and general and family practitioners played a dominant role in the care of moderate- and high-risk patients, accounting for 69% of visits by moderate-risk patients and 58% of visits by high-risk patients. Distributions by race/ethnicity, geographic region, residence area, and practice setting did not differ by CHD risk. Overall, the majority of patient visits were return visits to office-based physicians made by non-Hispanic whites and residents living within metropolitan statistic areas. Patient visits were distributed similarly across the four geographic regions, with a slightly higher proportion from the southern region.
Table 1 Differences in Patient, Physician, and Visit Characteristics by CHD Risk in 2002
Sample sizes, with weighted sample sizes in parentheses, are as follows: low-risk group, 40,127 (588,000,000); intermediate-risk group, 4,112 (87,000,000); high-risk group, 4,773 (79,000,000).
a Number in parentheses is standard error.
NS, non-significant.
Throughout the study period, statins were primarily used among patients whose visit involved reported hyperlipidemia, representing 97% of all statin use in 1992 and 91% in 2002. Statin use increased nearly 5-fold from 9% (99% confidence interval: 7%–12%) of all visits with reported hyperlipidemia in 1992 to 49% (42%–55%) in 2000, but then declined to 36% (31%–42%) in 2002 (Figure 1). Of note, however, the annual rate of increase in frequency of patient visits with reported hyperlipidemia was 34% in 2001 and 21% in 2002, while it averaged only 12% through 2000. The dominance of statins as lipid-lowering agents grew markedly from 47% of all lipid-lowering medications in 1992 to 87% in 2002 (Figure 1). Among available statins, lovastatin remained the therapeutic choice through 1996, after which it was surpassed by other statins, particularly simvastatin and then atorvastatin (Figure 2). Atorvastatin constituted 51% (46%–56%) and simvastatin 32% (27%–36%) of all statin use in 2002.
Figure 1 Use of Statins and All Lipid-Lowering Medications among United States Ambulatory Visits by Patients Diagnosed with Hyperlipidemia
Data from NAMCS and NHAMCS, 1992–2002.
Figure 2 Share of Total Statin Use among United States Ambulatory Visits by Individual Statin Medications
Data from NAMCS and NHAMCS, 1992–2002.
As expected, high CHD risk patient visits resulted in greater statin use, and the divergence in statin use among the three risk categories has grown in recent years. Absolute increases in the rate of statin use were greatest for high-risk patient visits with or without reported hyperlipidemia—a 15 percentage-point increase from 4% of all visits in 1992 to 19% in 2002—followed by a nine percentage-point increase (2% to 11%) for moderate-risk patient visits and a 2.5 percentage-point increase (0.3% to 2.8%) for low-risk patient visits (Figure 3). Statin use in the moderate-risk group peaked at 14% (10%–17%) in 1999. Similarly, the rate of statin use in the high-risk group declined slightly from 2001 to 2002.
Figure 3 Use of Statins among United States Ambulatory Visits, by CHD Risk Category
Data from NAMCS and NHAMCS, 1992–2002.
Among patient visits with reported hyperlipidemia, statins were used in 14% (8%–19%) of high-risk visits and 9% (5%–14%) of moderate-risk visits in 1992. The high-risk group's statin use rate rose to 60% (49%–71%) in 2000 and was 50% (40%–61%) in 2002. Likewise, the rate in the moderate-risk group climbed to 56% (42%–70%) in 1999 and stabilized at 44% (32%–57%) in 2002. In addition, lifestyle counseling (i.e., regarding diet, exercise, or smoking cessation) occurred in only 43% (32%–53%) of new and general medical examination visits in 2002 for patients who had moderate CHD risk and were diagnosed with hyperlipidemia. Improvements over time in counseling rates were minimal.
The increase in statin use with CHD risk and with the year of study persisted after controlling for physician-reported hyperlipidemia, number of medications, and nonclinical patient visit characteristics (Table 2). Moderate- to high-risk patient visits had a 1.2- to 2.5-fold greater likelihood of taking a statin relative to visits by patients at low risk. Statin use was approximately three times as likely in 2001 and 2002 as in 1995 and 1996. Additionally, lower statin use was independently associated with younger patient age, female gender, African American background (versus non-Hispanic white), non-cardiologist care, and fewer total reported medications.
Table 2 Predictors of Statin Use with Combined 1995–2002 NAMCS and NHAMCS Data
a Numbers in parentheses are standard errors.
b The odds ratio for each variable was adjusted for all other variables listed in the table.
c Numbers in parentheses are 99% confidence intervals.
Discussion
Despite significant increases from 1992 to 2002 in use of statins associated with hyperlipidemic patient visits, the magnitude of increases is smaller than expected and the rate of use remains suboptimal according to the best available evidence. The underuse of statins is most prominent among visits by patients at high or moderate risk of CHD who do not have a physician-noted diagnosis of hyperlipidemia but may nonetheless be eligible for lipid-lowering drug therapy. Previous research reports that physicians are more likely to diagnose hyperlipidemia if laboratory reports show abnormal lipid levels [14]. However, the normal ranges of lipid levels on many laboratory reports do not take into account individual patients' absolute risk.
When evaluating statin use across different CHD risk categories, the observed trends raise several issues. Both the rate of statin use and the absolute increases in the rate over time were positively associated with the level of CHD risk, which appropriately conforms to the notion of risk stratification. The associations persisted after adjusting for potentially confounding factors such as a hyperlipidemia diagnosis and nonclinical patient visit characteristics. Even so, in 2002, one year after the publication of Adult Treatment Panel III [3], statins were reportedly used in only 19% of patient visits with established CHD or its equivalents, and the average rate was no higher than 50% among high-risk visits where a diagnosis of hyperlipidemia also was noted. These data suggest a dramatic treatment gap. Another analysis based on national data estimated that 72% of Americans with existing CHD would benefit from drug therapy to achieve the target LDL-C goal of 2.59 mmol/l (100 mg/dl) or less, assuming a 10% LDL-C reduction with diet [15]. However, only 11% of those eligible individuals received lipid-lowering drug therapy, suggesting a gap of 89% [15]. These obvious treatment gaps are disconcerting, especially in light of the recent Adult Treatment Panel III update[16] that supports more intensive lipid-lowering drug therapy for patients at high and moderately high risk for a heart attack. Barriers to adequate treatment of high-risk patients may stem from the patient (e.g., lack of drug adherence, concern about adverse effects, inadequate knowledge of their hyperlipidemia, and drug cost), the physician (e.g., lack of guideline awareness, failure to measure lipid levels, and overestimation of actual treatment), and the health-care system (e.g., lack of monitoring and follow-up and emphasis on acute medical problems) [10]. If the current practice continues, the observed treatment gaps are expected to persist or even widen.
While statins deliver the greatest benefits when used for secondary prevention, evidence continues to accumulate that suggests an important role of statins in the primary prevention of cardiovascular events, particularly for patients at increased risk [6,17]. Our data show an increase in statin use from 2% of moderate-risk patient visits in 1992 to 14% in 1999, but without continued growth subsequently. Optimal proportions could not be determined because of the lack of detailed clinical data. Nonetheless, National Health and Nutrition Examination Survey III data showed that 60% of 38.5 million adult Americans without CHD who had two or more risk factors had an LDL-C level above the recommended 3.36 mmol/l (130 mg/dl) and that 45% would remain eligible for drug therapy even after a 10% decrease in LDL-C with diet [15]. In addition, Fedder and colleagues found a doubling effect in the number eligible for primary prevention drug therapy by switching to Framingham risk scoring [18]. Other researchers have reported that the proportions of treatment-eligible primary prevention patients who received no drug therapy reached as high as 97% [5,19]. In our study, statin use was reported in only 44% of moderate-risk patient visits for which a diagnosis of hyperlipidemia was noted, which is surprisingly low given that the entire group would be expected to benefit from statin therapy. We also concur with other researchers who have discussed the role that inadequate lifestyle counseling plays in the existing cholesterol treatment gaps [10,13]. Our data show that lifestyle counseling occurred during fewer than 50% of new and general medical examination visits by moderate-risk patients, even though these types of visits arguably represent better opportunities for counseling services than return, illness-focused visits.
It is intriguing to note that earlier increases in statin use were not sustained in 2001 and 2002. Studies using alternative data sources are needed to corroborate this observation, and detailed market research is necessary for understanding the underlying causes of this unexpected decline in use. We speculate that the observed trends may be partially explained by discordant rates of increase in the diagnosis of hyperlipidemia versus the prescribing of statins. Also, NAMCS and NHAMCS data released after 2002 will help determine whether the noted declines are due to random fluctuations in data reporting.
Wide gaps between evidence-based lipid-lowering therapy and physician practice were reported in many other western countries as well. For instance, a survey conducted in nine European countries found that only 32% of patients with confirmed CHD received lipid-lowering medications [20]. Likewise, in a population-based study from the Netherlands, merely 16% of individuals eligible for lipid-lowering drugs were actually treated [7].
In spite of being clearly underused, statins increasingly dominate lipid-lowering drug therapy, accounting for 92% of all lipid-lowering medications used in 2002, which confirms the trends seen in United States retail pharmacy dispensing data [2]. Also, in concert with other researchers [21], we observed a shift in the leading statin prescribed over time, from lovastatin to simvastatin and then to atorvastatin, corresponding to their market entry. Atorvastatin accounted for over half of all statin use in 2002. Even though most statins share similar tolerability, some evidence shows that atorvastatin has greater dose-specific potency for lowering LDL-C and total cholesterol [22].
Additionally, our data add support to available literature documenting inequities in use of statins for patients with different social and clinical characteristics [23–26]. Of particular note are the lower rates of statin use in at-risk younger patients, females, African-Americans, and patients cared for by non-cardiologists. These findings may be useful for guiding targeted interventions that aim to bring physician practice into agreement with published guidelines for cardiovascular risk reduction.
Our findings must be interpreted in the context of data limitations. Although both NAMCS and NHAMCS are designed to produce nationally representative estimates, these estimates are not linked to individuals but to patient visits. As a result, reported statin use may overestimate the actual administration because patients prescribed drug therapy likely make more visits because of greater disease severity and/or the need of frequent follow-ups. Also, we are missing people with risk factors who have not been seen by a physician or whose risk factors failed to be recorded. On the other hand, underestimation is also possible, for example, because of physicians' lack of awareness or incomplete reporting of patient medication uses. However, the failure to inquire or report an important agent such as a statin may be a clinical oversight in itself and contribute to therapeutic gaps. The degree of inaccuracy in our estimates is perhaps small, however, as suggested by the comparability of the current results to previous reports.
Lack of detailed clinical data prohibits accurate risk assessment based on Framingham risk scoring. While the risk factor counting algorithm that we used may simulate practical risk estimation by many physicians, it precludes the assessment of appropriateness of statin use in relation to the latest lipid-lowering guidelines. This creates difficulty interpreting the rate of statin use observed for the moderate-risk group. In particular, adequate information is not available to differentiate varying levels of absolute risk among the moderate-risk group. We likely misclassified some patients as moderate risk when they may have actually been high risk despite the absence of CHD or CHD equivalents. On the other hand, indications for statins might be marginal for some young patients with modestly elevated risk factors. A final caveat is that neither NAMCS nor NHAMCS captures patient compliance or outcomes, although these are perhaps separate issues from physician adherence to evidence-based medicine.
Despite the acknowledged limitations, NAMCS and NHAMCS cover a longer consecutive time span and provide more complete information about disease-specific physician activities than many other national data bases, e.g., the Medical Expenditures Panel Survey, the National Health Assessment Nutrition Examination Survey, and the National Health Interview Survey.
In conclusion, persistent gaps in statin therapy suggest a continued need for improved CHD risk stratification of all patients, and treatment with statins when indicated. Information technology and broader national policy around quality measurement and reporting are just two potential strategies that could be used to improve current practice. Patient-centered interventions should strengthen patient education and improve patient access to different treatment options. Interventions should be targeted to at-risk patients whose drug regimens need to be reassessed and to physicians, particularly non-cardiologists, whose practices need be improved. Guidelines for cardiovascular risk reduction treatment and determination of the specific patients who can benefit from statin therapy will continue to evolve. Indications for use in primary CHD prevention are likely to expand for statins. Given the observed practice shortfalls, drug therapy in moderate-risk patients remains an important priority for improvement.
Patient Summary
Background
Statins, a group of drugs that block a part of the pathway essential in forming lipids, have been shown to be very effective in reducing the chances of heart attacks and stroke both in people who have already had such an illness, and also in otherwise healthy people who have multiple risk factors. Long-term data are limited regarding whether people who might most benefit from such drugs are in fact receiving them.
What Did the Investigators Do?
Using two large United States outpatient care surveys that have run for many years, they looked at information on how doctors prescribed these drugs to patients. They found that between 1992 and 2002 the proportion of visits by patients with high lipid levels during which statins were prescribed rose from 9% in 1992 to 49% in 2000 and then fell to 36%. By 2002, statins were the dominant drug used for people with high lipid levels, accounting for 87% of all prescribed lipid-lowering medications. Although people at highest risk of cardiovascular diseases (with or without high lipid levels) were most likely to be prescribed a statin, by 2002, statins were prescribed in only 19% of visits by these people. In addition, only 11% of visits by people with multiple risk factors received a statin. Less than half the patient visits that arguably represent optimal opportunities for counseling services received counseling about how they might improve their cardiovascular health through lifestyle changes. Other factors that led to lower prescribing of statins were younger patient age, female gender, African American background, and care by non-cardiologists.
What Do These Findings Mean?
It seems that many people who could benefit from statins are not receiving them. Equally importantly, many people also did not receive information about how they could modify their lifestyle.
Where Can I Get More Information?
MedlinePlus has information on high lipids: go to http://www.nlm.nih.gov/medlineplus and search for “hyperlipidemia”
The British Heart Foundation has a number of pages on lipids: go to http://www.bhf.org.uk and search for “lipids”
This research was sponsored by an unrestricted grant from Merck. Additional support was provided by a research award from the Agency for Healthcare Research and Quality (R01-HS11313). Both funding sources had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Citation: Ma J, Sehgal NL, Ayanian JZ, Stafford RS (2005) National trends in statin use by coronary heart disease risk category. PLoS Med 2(5): e123.
Abbreviations
CHDcoronary heart disease
LDL-Clow-density lipoprotein cholesterol
NAMCSNational Ambulatory Medical Care Survey
NHAMCSNational Hospital Ambulatory Medical Care Survey
==== Refs
References
American Heart Association Heart disease and stroke statistics—2003 Update. Dallas Texas: American Heart Association. Available: http://www.americanheart.org/downloadable/heart/10590179711482003HDSStatsBookREV7–03.pdf
2002 Accessed 23 March 2005
Siegel D Lopez J Meier J Use of cholesterol-lowering medications in the United States from 1991 to 1997 Am J Med 2000 108 496 499 10781783
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults 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
Law MR Wald NJ Rudnicka AR Quantifying effect of statins on low density lipoprotein cholesterol, ischaemic heart disease, and stroke: Systematic review and meta-analysis BMJ 2003 326 1423 12829554
Whincup PH Emberson JR Lennon L Walker M Papacosta O Low prevalence of lipid lowering drug use in older men with established coronary heart disease Heart 2002 88 25 29 12067936
Gotto AM Jr Lipid management in patients at moderate risk for coronary heart disease: Insights from the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS) Am J Med 1999 107 36S-39S
Mantel-Teeuwisse AK Verschuren WM Klungel OH Kromhout D Lindemans AD Undertreatment of hypercholesterolaemia: A population-based study Br J Clin Pharmacol 2003 55 389 397 12680888
Sueta CA Chowdhury M Boccuzzi SJ Smith SC Alexander CM Analysis of the degree of undertreatment of hyperlipidemia and congestive heart failure secondary to coronary artery disease Am J Cardiol 1999 83 1303 1307 10235085
Pearson TA Laurora I Chu H Kafonek S The lipid treatment assessment project (L-TAP): A multicenter survey to evaluate the percentages of dyslipidemic patients receiving lipid-lowering therapy and achieving low-density lipoprotein cholesterol goals Arch Intern Med 2000 160 459 467 10695686
Olson KL Bungard TJ Tsuyuki RT Cholesterol risk management: A systematic examination of the gap from evidence to practice Pharmacotherapy 2001 21 807 817 11444577
Zell ER McCaig LF Kupronis BA Besser RE Schuchat A A comparison of the National Disease and Therapeutic Index and the National Ambulatory Medical Care Survey to evaluate antibiotic usage. In: Proceedings of the survey research methods section, American Statistical Association (2000). Alexandria (Virginia): American Statistical Association. Available: http://www.amstat.org/sections/srms/Proceedings/papers/2000_143.pdf
2000 Accessed 1 April 2005
Gilchrist VJ Stange KC Flocke SA McCord G Bourguet CC A comparison of the National Ambulatory Medical Care Survey (NAMCS) measurement approach with direct observation of outpatient visits Med Care 2004 42 276 280 15076827
Stafford RS Blumenthal D Pasternak RC Variations in cholesterol management practices of U.S. physicians J Am Coll Cardiol 1997 29 139 146 8996306
Madlon-Kay DJ Family physician recognition and treatment of severe hypercholesterolemia J Fam Pract 1987 24 54 56 3794614
Jacobson TA Griffiths GG Varas C Gause D Sung JC Impact of evidence-based “clinical judgment” on the number of American adults requiring lipid-lowering therapy based on updated NHANES III data. National Health and Nutrition Examination Survey Arch Intern Med 2000 160 1361 1369 10809042
Grundy SM Cleeman JI Merz CN Brewer HB Clark LT Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III guidelines Circulation 2004 110 227 239 15249516
Goldman L Coxson P Hunink MG Goldman PA Tosteson AN The relative influence of secondary versus primary prevention using the National Cholesterol Education Program Adult Treatment Panel II guidelines J Am Coll Cardiol 1999 34 768 776 10483959
Fedder DO Koro CE L'Italien GJ New National Cholesterol Education Program III guidelines for primary prevention lipid-lowering drug therapy: Projected impact on the size, sex, and age distribution of the treatment-eligible population Circulation 2002 105 152 156 11790693
Hoerger TJ Bala MV Bray JW Wilcosky TC LaRosa J Treatment patterns and distribution of low-density lipoprotein cholesterol levels in treatment-eligible United States adults Am J Cardiol 1998 82 61 65 9671010
EUROASPIRE A European Society of Cardiology survey of secondary prevention of coronary heart disease: principal results. EUROASPIRE Study Group. European Action on Secondary Prevention through Intervention to Reduce Events Eur Heart J 1997 18 1569 1582 9347267
Wang TJ Stafford RS Ausiello JC Chaisson CE Randomized clinical trials and recent patterns in the use of statins Am Heart J 2001 141 957 963 11376310
Jones P Kafonek S Laurora I Hunninghake D Comparative dose efficacy study of atorvastatin versus simvastatin, pravastatin, lovastatin, and fluvastatin in patients with hypercholesterolemia (the CURVES study) Am J Cardiol 1998 81 582 587 9514454
Ayanian JZ Landon BE Landrum MB Grana JR McNeil BJ Use of cholesterol-lowering therapy and related beliefs among middle-aged adults after myocardial infarction J Gen Intern Med 2002 17 95 102 11841524
Ayanian JZ Landrum MB McNeil BJ Use of cholesterol-lowering therapy by elderly adults after myocardial infarction Arch Intern Med 2002 162 1013 1019 11996611
Reid FD Cook DG Whincup PH Use of statins in the secondary prevention of coronary heart disease: Is treatment equitable? Heart 2002 88 15 19 12067933
Yang CC Jick SS Testa MA Who receives lipid-lowering drugs: The effects of comorbidities and patient characteristics on treatment initiation Br J Clin Pharmacol 2003 55 288 298 12630980
| 15916463 | PMC1140942 | CC BY | 2021-01-05 10:39:50 | no | PLoS Med. 2005 May 31; 2(5):e123 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020123 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646410.1371/journal.pmed.0020126Correspondence and Other CommunicationsOtherMedical JournalsEditorial Policies (Including Conflicts of Interest)Unjustified Restrictions on Letters to the Editor CorrespondenceAltman Douglas G Cancer Research UK/NHS Centre for Statistics in MedicineOxfordUnited KingdomE-mail: [email protected]
Competing Interests: The author has declared that no competing interests exist.
5 2005 31 5 2005 2 5 e126Copyright: © 2005 Douglas G. Altman.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.
Editors' Reply
==== Body
Editors of medical journals accept that published research should be open to comment and correction in published correspondence ([1]; Box 1). “Post-publication peer review” enables comments on, clarifications of, and corrections to published research. All journals should have a correspondence page for this purpose.
Box 1. “Editors Should Promote Self-Correction in Science and Participate in Efforts to Improve the Practice of Scientific Investigation by:
“Publishing corrections, retractions, and letters critical of articles published in their own journal.
“Playing an active role in investigating and preventing fraud.
“Taking responsibility for improving the level of scientific investigation and medical writing in the larger community of potential authors.
“Giving authors an opportunity to review and approve edited manuscripts before they are published.
“Participating in efforts to detect and prevent publication bias—for example, by collaborating with registries of controlled trials and publishing protocols.”
Source: [1].
I previously criticised the effective “statute of limitations” in several leading general medical journals “whereby authors of papers are immune to disclosure of methodological weaknesses once some arbitrary (short) period has elapsed” [2]. Such a time limit discourages post-publication peer review, with potential correspondents deterred by the short and unambiguous deadline. I suggested that journals with such a policy should reconsider. The word limit on that article precluded additional adverse comments on journals' word limits for letters, although they were presented in Table 2 of that article [2].
Subsequently, three of the six journals did revise their instructions [3–5], but each imposed tougher restrictions on letters, reducing either the maximum time limit, the maximum length, or both. The strictest current requirements are a two-week limit by The Lancet and a 175-word limit by the New England Journal of Medicine.
Editors are seemingly falling over themselves to speed up and shorten letters, but this behaviour is inappropriate for a scientific journal. The key characteristic of science is not its infallibility, a quality it clearly does not and cannot have, but its self-correcting ability. The decision by medical editors to stifle debate is misguided [2,6]. A time limit, especially a very short one, signals that speed is more important than content, that convenience takes precedence over science. While it is reasonable to encourage early comments, there should be no time limit on comments aimed at clarifying or criticising study methodology. Likewise, it will often be impossible to explain the subtleties of methodological problems in 400 words, and impossible in only 175. Additional restrictions on the number of authors and references are also questionable.
I am disappointed that PLoS Medicine has imposed a time limit of four weeks on correspondence. As explained above, I believe that such a limit is mistaken. The word limit of 750 words is generous by comparison to established general medical journals, but even this should be open to flexibility should the circumstances merit it.
In this world of Web-based journals and Web pages for print journals there is no real cost to permitting longer and later letters on the web while keeping the print version timely and terse.
Citation: Altman DG (2005) Unjustified restrictions on letters to the editor. PLoS Med 2(5): e126.
==== Refs
References
World Association of Medical Editors Report of the World Association of Medical Editors (WAME): An agenda for the future 2001 Available at http://www.wame.org/bellagioreport_1.htm . Accessed 4 February 2005
Altman DG Poor-quality medical research: What can journals do? JAMA 2002 287 2765 2767 12038906
Mullan Z Lancet correspondence: Old letters, new rules Lancet 2003 361 12
Curfman GD Graham A Lindenfelser L Anderson KR Drazen JM Innovations in correspondence N Engl J Med 2003 348 344
Davies S New edicts for letters to the editor BMJ 2003 326 63 64
Metcalfe S Old letters, new rules Lancet 2003 361 706
| 15916464 | PMC1140943 | CC BY | 2021-01-05 10:39:50 | no | PLoS Med. 2005 May 31; 2(5):e126 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020126 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646510.1371/journal.pmed.0020127Policy ForumInfectious DiseasesAllergy/ImmunologyHealth PolicyMedicine in Developing CountriesPediatricsImmunology and allergyHealth PolicyInternational healthVaccine Policy in India Policy ForumMadhavi Yennapu Yennapu Madhavi holds a PhD in life sciences and works as a scientist at the National Institute of Science Technology and Development Studies (New Delhi, India). She specialises in science and technology policy issues in the area of public health in general, and vaccines in particular. The views expressed in this article are those of the author rather than her institution. E-mail: [email protected]
Competing Interests: The author declares that she has no competing interests.
5 2005 31 5 2005 2 5 e127Copyright: © 2005 Yennapu Madhavi.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.India enjoyed early initial successes in vaccine development and indigenous production of vaccines in the public sector. But the country now faces a growing gap between the demand for and supply of essential vaccines.
==== Body
Vaccines are important preventive medicines for primary health care, and are a critical component of a nation's health security. Although international agencies such as the World Health Organization (WHO) and the United Nations Children's Fund (UNICEF) promote global immunisation drives and policies, the success of an immunisation programme in any country depends more upon local realities and national policies (Box 1). This is particularly true for a huge and diverse developing country such as India, with its population of more than 1 billion people, and 25 million new births every year.
Box 1. Local Realities and National Policies That Affect the Success of a Country's Immunisation Program
disease surveillance
pathogen variations
incidence levels that qualify for mass vaccination
development and/or procurement of vaccines
choice of technologies
choice of selective vs. universal vaccination (even among childhood vaccines)
logistics, cost-benefit analyses, and resource mobilisation
The current Indian market for vaccines is estimated to be about US$260 million [1]. India is among the major buyers and makers of vaccines, locally as well as globally, and has traditionally aimed at self-reliance in vaccine technologies and production. This article explores the trajectory of vaccine policy in India through its historical roots and institutional development, the gaps in demand and supply, the changing nature of the vaccine industry, and the emerging challenges in meeting national immunisation targets.
A nurse at Malipur Maternity Home (Delhi, India) prepares to vaccinate a child
(Photo: the WHO/P. Virot)
Early Origins
The history of vaccine research and production in India is almost as old as the history of vaccines themselves. During the latter half of the 19th century, when institutions for vaccine development and production were taking root in the Western world [2], the British rulers in India, concerned by the large number of their personnel dying from tropical diseases, promoted research on these diseases and established about fifteen vaccine institutes beginning in the 1890s. Prior to the establishment of these institutions, there were no dedicated organisations for medical research in India.
Haffkine's development of the world's first plague vaccine in 1897 (which he developed at the Plague Laboratory (Mumbai, India), later named the Haffkine Institute) and Manson's development of an indigenous cholera vaccine at Kolkata during the same period bear testimony to the benefits of the early institutionalisation of vaccine research and development in India [3]. Soon, Indian vaccine institutes were also producing tetanus toxoid (TT), diphtheria toxoid (DT), and diphtheria, pertussis, and tetanus toxoid (DPT).
However, the benefits of this early institutionalisation did not last long. The policies of the colonial government ensured that Indian scientists were not a significant part of this intellectual legacy. By the time Indians inherited the leadership of the above institutions in the early 20th century, research and technological innovation were sidelined as demands for routine vaccine production took priority [3]. By the time India gained independence in 1947, the Indian vaccine research and development (R&D) institutions were no longer on a par with vaccine technology development centres elsewhere. This is reflected in the fact that improved techniques for bacterial vaccines were introduced in India almost a decade after their introduction elsewhere in the world (Table 1) [4].
Table 1 The Introduction of Vaccine Technologies in India and Elsewhere in the World
Source: [4].
What were the factors that led to the stagnation in vaccine development efforts between the time of Haffkine's success and India's independence? These included the pressures of routine production and service functions, financial constraints, lack of institutional mechanisms to foster and link up research and technology development, and the absence of an interdisciplinary approach. All these factors posed a threat to India's vaccine development efforts [5].
Vaccine Policy in Independent India
One year after its independence in 1947, India became a member country of the WHO and eagerly aligned itself to the policies of the WHO and UNICEF. Many new Indian institutions were established with partial support from international organisations during the period 1950–1970.
However, after independence, it took three decades for India to articulate its first official policy for childhood vaccination, a policy that was in alignment with the WHO's policy of “Health for All by 2000” (famously announced in 1978 at Alma Atta, Kazakhstan). The WHO's policy recommended universal immunisation of all children to reduce child mortality under its Expanded Programme of Immunization (EPI). In line with Health for All by 2000, in 1978 India introduced six childhood vaccines (Bacillus Calmette-Guerin, TT, DPT, DT, polio, and typhoid) in its EPI. Measles vaccine was added much later, in 1985, when the Indian government launched the Universal Immunization Programme (UIP) and a mission to achieve immunisation coverage of all children and pregnant women by the1990s.
Gaps in Vaccine Technology and Production: The Declining Role of the Public Sector
Vaccine requirements for India's EPI have been met mainly through the public-sector vaccine institutions, as was the case in most parts of the world until the 1980s. However, the Indian public sector failed to introduce new technologies of production (such as production of TT, DT, or DTP) or to expand production to become self-reliant in producing oral polio vaccine (OPV) or the measles vaccine [6]. Thus, even though successive governments have adopted self-reliance in vaccine technology and self-sufficiency in vaccine production as policy objectives in theory, the growing gap between demand and supply meant that in practice, India had increasingly to resort to imports.
In some cases, indigenously manufactured vaccines were stopped in favour of imported vaccines. For example, the Pasteur Institute of India in Coonoor indigenously produced polio vaccine during the period 1967–1977 with the help of seed virus from Dr. A. B. Sabin (who developed OPV) and with the approval of the WHO. However, the Indian government discouraged its production in 1977, alleging that one of the batches was virulent, and since then OPV has become one of India's major imports [4]. Subsequently, the Haffkine Institute was able to produce OPV indigenously, but this was mysteriously discontinued [5].
In 1987, the Union (federal) government's Department of Biotechnology established a new public sector unit, Bharat Immunologicals and Biologicals Corporation Ltd. (Bulandshar, India) with technology transferred from the Institute of Poliomyelitis and Viral Encephalitis (Moscow, Russia). The first phase of production was based on repackaging OPV imported in bulk from Russia. The aim was for OPV production to be completely indigenised in the second phase, within five years. However, the first phase continued (with imports) until the year 2000, when such importation supplied 70 million doses of OPV to UNICEF and earned a net profit of Rs 8 million. Yet, in 2000, the government declared Bharat Immunologicals and Biologicals a sick unit (a loss-making unit that is financially unviable), and its revival remains uncertain.
It is strange that there is no published analysis as to why OPV production in the Indian public sector has failed repeatedly while Panacea Biotec, an Indian private sector firm, has recently secured a comfortable position as a WHO pre-qualified supplier of OPV for UNICEF. Panacea repackages its OPV from imported bulk OPV obtained from Biopharma (Bandung, Indonesia) and Chiron (Siena, Italy). Some argue that even though India always had an effective indigenous injectable polio vaccine (IPV), OPV was recommended in developing countries because international organizations were trying to find new markets for United States multinational corporations (since market demand for OPV ceased to exist in the US and other developed countries by the end of the 20th century) [7].
The failure of the Indian public sector in vaccine production was not limited to OPV. In 1984, the government took over Bengal Immunity Ltd. (Kolkata, India), a loss-making private company, and revived it so that it could supply TT, DT, DPT and other products to the government. But within a decade, the government declared that the company was financially unviable and eventually closed it. Similarly, in 1989, the Union government's Department of Biotechnology established a new public sector unit, Indian Vaccine Corporation Ltd. (IVCOL) at Gurgaon, for the indigenous production of measles vaccine with technology transferred from Institut Merieux, a public sector company based in Lyons, France. However, the technology transfer never materialised, as the private sector took over the French public sector firm and denied the technology transfer to IVCOL [8]. IVCOL was eventually closed down, and India's entire measles vaccination requirement was met through imports until an Indian private company based in Pune, the Serum Institute of India (SII), started its supply to the EPI in 1992.
The inability of the Indian public sector to recover from its mounting failures to achieve self-sufficiency and self-reliance in primary vaccines is also related to the liberalisation and globalisation of the Indian economy. It is not a coincidence that these failures and closures, and the preference for imports (while paying lip service to self-reliance), happened after the Indian government liberalised its economy in 1991 as prescribed by the International Monetary Fund and the World Bank. It is no longer fashionable to produce vaccines in the public sector in India, let alone to try and revive failing public sector units, even if essential vaccines are not available from the private sector.
The Increased Role of the Private Sector: Distorted Prioritisation of Vaccine R&D/Production
One of the main reasons for the growing gap in demand for and supply of primary vaccines in India is that while public sector production is on the decline (Figure 1), vaccine availability from the private sector (Figures 1 and 2) or through the UNICEF procurement mechanism (based on global tenders from suppliers pre-approved by the WHO) has not improved. This is a part of a worrisome global trend that has been acknowledged by UNICEF (http://www.unicef.org/supply/index_vaccine_security.html).
Figure 1 Primary Vaccine Suppliers to the Indian EPI in the Last Four Decades
The data were compiled from the annual reports of Health Information of India (1970–1971 to 2001–2002), and the Ministry of Health and Family Welfare, Government of India, New Delhi.
Figure 2 The Growth of the Private Sector in the Indian Vaccine Market
The data were compiled from the annual reports of Health Information of India (1970–1971 to 2001–2002), the Ministry of Health and Family Welfare (Government of India) and MIMS India (www.mims-india.com), Nov 2001, New Delhi.
Shortages of primary vaccines in developing countries began to emerge in the late 1990s. These shortages were due to the introduction of new, more sophisticated, more expensive vaccines in industrialised country markets, leading to manufacturers phasing out the production of the traditional, less expensive vaccines used in developing countries. Between 1998 and 2001, ten out of 14 major manufacturers partially or totally stopped production of traditional vaccines. Eight of these firms were the main suppliers of vaccines to UNICEF. Of these eight, six were involved in mergers between larger pharmaceutical companies. The overall outcome of these developments is that the availability of primary vaccines to UNICEF has dramatically decreased, while the prices have increased (http://www.unicef.org/publications/index_4442.html).
Indeed, the rapid growth (8%–10% per annum) of India's current human vaccine market is mainly attributed to the new, high-priced vaccines (Figure 2 and Table 2) such as Hepatitis B that have been launched since the 1990s. There has been pressure from the industry to include these new vaccines in the government's UIP, even though the clinical and epidemiological justification for their inclusion is controversial [9,10]. With epidemiology taking a backseat, government decisions on vaccination are increasingly determined by price competition and supply “push” (by the companies) rather than “pull” (demand) from proven public health needs [9].
Table 2 Cost of Full Immunisation with Each Vaccine (in US dollars)
Data compiled from Monthly Index of Medical Specialities (MIMS) India, New Delhi, November 2001.
Many western countries have included several other new vaccines (such as influenza type B, meningitis, measles-mumps-rubella, and chickenpox) in their regular immunisation programmes [11]. These trends are used as a justification by the industry to include these vaccines in the Indian UIP in the future. Aggressive promotional campaigns for the new vaccines and their quick adoption by industry-friendly private medical practitioners have already made these vaccines akin to fast-moving consumer goods. The industry, which enjoys all the benefits of economic liberalisation, sees no contradiction in seeking a captive market for its new vaccines through the government-sponsored UIP while at the same time failing to meet its social responsibility to meet the shortfall in production of existing UIP vaccines.
There is another serious contradiction that grips the global drug and vaccine industry. For curative medicine, the pharmaceutical industry places increasing emphasis on the use of genomics and bioinformatics to move toward customised medicine to suit different populations. And yet in vaccines, the tendency is to move toward a “one vaccine fits all” regime. This would be fine if the vaccines were specifically designed for universal use, but there was no attempt to conclusively establish that the imported vaccines actually suited the Indian strains of the pathogens before they were adopted. Doubts over suitability that have subsequently emerged have not been adequately addressed. With the decline of epidemiology and disease surveillance in India, and the main emphasis being on the statistics of vaccine “coverage” rather than the immune protection achieved, it seems that spending money on vaccines is more important than actual disease prevention. If these trends continue unabated, they will lead to serious distortions in the vaccination programmes of India and other developing countries facing a similar situation.
Conclusions and Recommendations
India enjoyed the advantages of early initial successes in vaccine R&D and indigenous production in the public sector, but the country is increasingly unable to cope with the growing gap in the demand and supply of UIP vaccines [6]. The availability of UIP vaccines from the private sector is also on the decline in India and abroad, in favour of more expensive new vaccines and combination vaccines, whose public health need has not been unequivocally established in India with sound epidemiological and cost-benefit data [9,12]. Therefore, India (and indeed, every country) must evolve its own national strategies to meet its vaccination needs within its budgetary constraints. To do so will require four key actions.
The first and foremost element in this strategy must be the decisive intervention of the Indian government to meet the shortfall in the UIP vaccines. This may be done either by strengthening the public sector wherever possible, or by taking suitable (and transparent) measures to encourage the indigenous private sector on a case-by-case basis to make safe and effective vaccines available at affordable prices. The suitability of imported vaccines to deal with Indian pathogenic strains also needs to be conclusively established wherever necessary. The health security of a nation of India's size cannot be left to the vagaries of global market forces. With a strong will and a small amount of planning, the current situation in India can be reversed, and India can even play a major role in meeting the global shortfall in the vaccines procured by UNICEF.
Secondly, India needs to strengthen epidemiology and revive the collapsing disease surveillance system. This would help to decide between universal or selective immunisation based on unequivocal scientific evidence, as well as to respond to the changing disease prevalence scenario on the ground, which may call for a move from universal to selective immunisation or vice versa. Some diseases may not need vaccinating against at all, and may be better controlled by other strategies, such as better sanitation, vector control, quarantine, and curative medicines. National immunisation programmes must be led by scientifically established public health needs and not by the mere availability of a vaccine in the market.
Thirdly, a strong emphasis on in-house R&D is needed in order to ensure that our production technologies are in tune with the times, and to negotiate strategic partnerships with outside scientists or institutions and companies.
Last but not least, the Indian government should actively encourage independent policy research, cost-benefit studies, and wider national consultations on various aspects of vaccination and public health so that it can take more informed decisions on such matters.
Citation: Madhavi Y (2005) Vaccine policy in India. PLoS Med 2(5): e127.
Abbreviations
DPTdiphtheria
DTdiphtheria toxoid
EPIExpanded Programme of Immunization
IPVinjectable polio vaccine
IVCOLIndian Vaccine Corporation Ltd.
OPVoral polio vaccine
SIISerum Institute of India
TTtetanus toxoid
UIPUniversal Immunization Programme
UNICEFUnited Nations Children's Fund
WHOWorld Health Organization
==== Refs
References
Lohray BB Medical biotechnology in India Adv Biochem Eng Biotechnol 2003 85 215 281 12930097
Worboys M Lemaine G MacLeod R Mulkay M Weingart P The emergence of tropical medicine: A study in the establishment of a scientific specialty Perspectives on the emergence of scientific disciplines 1976 Moulton The Hague 281
Kumar A Medicine and the Raj, British medical policy in India, 1835–1911 1998 New Delhi (India) Sage Publications 246
Madhavi Y The issue of equity in primary vaccine technology development and its implications on the implementation of vaccine policy in India Social Sciences and Health News Letter 2001 2 5 17
Madhavi Y Research & production in the Haffkine Institute: A century and beyond Association for Consumers Action on Safety and Health (ACASH) News 2000 14 13 15
Madhavi Y Vaccine research: A case for national innovation strategy Curr Sci 1997 73 25 30
Bhargava PM An extract from a text book of history on planet Eurotopea in 2503 The Hindu (Chennai) 2003 August 26 10
Ramachandran R Vaccine project In need of a booster 1995 July 6 (New Delhi) Economic Times 10
Madhavi Y The manufacture of consent? Hepatitis B vaccination 2003 Available: http://www.epw.org.in/showArticles.php?root=2003&leaf=06&filename=5931&filetype=pdf . Accessed 24 March 2005
Phadke A Kale A Some critical issues in the epidemiology of Hepatitis-B in India Indian J Gastroenterol 2000 19 Suppl 3 c76 c77
Centre for Disease Control [CDC] Notice to readers, recommendations of the Advisory Committee on Immunization Practices: Revised recommendations for routine poliomyelitis vaccination Morb Mortal Wkly Rep 1999 48 590
Addlakha R Grover R User configuration and perspective: Hepatitis B introductory trial in East Delhi 2000 Available: http://www.epw.org.in/showArticles.php?root=2003&leaf=06&filename=5931&filetype=pdf . Accessed 24 March 2005
| 15916465 | PMC1140944 | CC BY | 2021-01-05 11:13:38 | no | PLoS Med. 2005 May 31; 2(5):e127 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020127 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646610.1371/journal.pmed.0020128Research ArticleInfectious DiseasesInfectious DiseasesMalariaAn Immune Basis for Malaria Protection by the Sickle Cell Trait Malaria Protection by the Sickle Cell TraitWilliams Thomas N
1
2
3
*Mwangi Tabitha W
1
Roberts David J
4
5
Alexander Neal D
6
Weatherall David J
7
Wambua Sammy
1
Kortok Moses
1
Snow Robert W
1
2
Marsh Kevin
1
2
1Kenya Medical Research Institute/Wellcome Trust Programme, Centre for Geographic Medicine ResearchCoast, Kilifi District Hospital, KilifiKenya2Nuffield Department of Medicine, John Radcliffe HospitalOxfordUnited Kingdom3Department of Paediatrics, John Radcliffe HospitalOxfordUnited Kingdom4Blood Research Laboratory, National Blood Service—OxfordJohn Radcliffe Hospital, OxfordUnited Kingdom5Nuffield Department of Clinical Laboratory Sciences, John Radcliffe HospitalOxfordUnited Kingdom6Medical Research Council Tropical Epidemiology Group, London School of Hygiene and Tropical MedicineLondonUnited Kingdom7Weatherall Institute of Molecular Medicine, John Radcliffe HospitalOxfordUnited KingdomModiano David Academic EditorUniversity of RomeItaly
Competing Interests: The authors have declared that no competing interests exist.
Author Contributions: TNW, TWM, DJR, DJW, and KM designed the study. TNW, NDA, SW, and MK analyzed the data. TNW wrote the paper.
*To whom correspondence should be addressed. E-mail: twilliams@ kilifi.mimcom.net5 2005 31 5 2005 2 5 e12823 11 2004 18 3 2005 Copyright: © 2005 Williams 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.
Sickle Cell and Malaria
Background
Malaria resistance by the sickle cell trait (genotype HbAS) has served as the prime example of genetic selection for over half a century. Nevertheless, the mechanism of this resistance remains the subject of considerable debate. While it probably involves innate factors such as the reduced ability of Plasmodium falciparum parasites to grow and multiply in HbAS erythrocytes, recent observations suggest that it might also involve the accelerated acquisition of malaria-specific immunity.
Methods and Findings
We studied the age-specific protection afforded by HbAS against clinical malaria in children living on the coast of Kenya. We found that protection increased with age from only 20% in the first 2 y of life to a maximum of 56% by the age of 10 y, returning thereafter to 30% in participants greater than 10 y old.
Conclusions
Our observations suggest that malaria protection by HbAS involves the enhancement of not only innate but also of acquired immunity to the parasite. A better understanding of the underlying mechanisms might yield important insights into both these processes.
In a Kenyan population, protection against malaria by sickle cell trait increased over the first 10 years of life, suggesting that the mechanism of protection involves acquired immunity to the parasite.
==== Body
Introduction
Sickle cell trait (genotype HbAS) confers a high degree of resistance to severe and complicated malaria [1–4] yet the precise mechanism remains unknown. To some extent it almost certainly relates to the peculiar physical or biochemical properties of HbAS red blood cells: invasion, growth, and development of Plasmodium falciparum parasites are all reduced in such cells under physiological conditions in vitro [5,6], and parasite-infected HbAS red blood cells also tend to sickle [5,7,8], a process that may result in their premature destruction by the spleen [5,9]. Nevertheless, while such factors appear to be important, recent observations suggest that the mechanism might also involve an immune component. For example, in a study conducted in Gambia, we found that the immune recognition of P. falciparum–infected red blood cells was enhanced in HbAS children [10], and up-regulation of malaria-specific cell-mediated immune responses has also been observed in HbAS individuals in Sudan [11,12]. While potentially important, such observations could represent epi-phenomena, rather than proximate effects of the HbAS red cell phenotype. Establishing whether or not immune processes are involved may prove useful in learning about malaria protection more generally. We have therefore investigated this question by studying the age-specific pattern of malaria disease in children living on the coast of Kenya. We reasoned that if malaria protection by HbAS was predominantly innate, it should be independent of malaria exposure and therefore remain constant with age. Conversely, if immune mechanisms were involved, the degree of protection should increase with age up until the age when children generally become functionally immune to malaria, at which time any additional immunological advantage should be lost.
Methods
Patients and Methods
The study was conducted in a cohort of children and adults living within the Ngerenya and Chonyi areas of Kilifi District on the coast of Kenya as described in detail previously [13,14]. Briefly, participants were recruited from an age-stratified population sample weighted towards children less than 10 y old, and between September 1998 and March 2004 study participants were monitored by active surveillance for clinical events with a focus on malaria. Children born into study households during the course of the program were recruited at birth, and participants exited from the study if informed consent was withdrawn, if they moved out of the study area for more than 2 mo, or if they died. Hemoglobin types were available for 1,054 of 1,795 total cohort members who attended periodic cross-sectional surveys conducted throughout the study. We have previously defined malaria as a fever (axillary temperature >37.5 °C) in association with malaria parasitaemia of any density in children less than 1 y old or at a density of greater than 2,500 parasites/μl in older children [13]. However, HbAS has a significant effect on the densities of incident malaria infections, and we have no data that allow us to confirm whether or not this definition is also appropriate for such children. For the purposes of this analysis, we have therefore used a conservative definition of malaria—fever in association with a slide positive for blood stage asexual P. falciparum parasites at any density. Incident malaria infections were treated with sulphadoxine–pyrimethamine according to local guidelines.
Laboratory Procedures and Statistical Analysis
Blood films were stained and examined for malaria parasites by standard methods, and haemoglobin types were characterized by electrophoresis. We compared the incidence of malaria in HbAS individuals versus individuals without the sickle cell allele (genotype HbAA) by Poisson regression (with malaria as the dependent variable) both with and without adjustments for the following confounding variables: season (defined as 90-d blocks), study area (Ngerenya or Chonyi), ethnic group, and age (in 2-y bands until the age of 10 y, older participants being classified in the top band as described in Table 1). Participants were considered not to be at risk of malaria and were dropped from both numerator and denominator populations for 21 d after receiving treatment with an anti-malarial drug. Because all infants are relatively resistant to malaria during the first 3 mo of life [15], we excluded children less than 3 mo old from our analyses. Because the study was conducted over a prolonged period, most participants contributed data to more than one age stratum. We took account of potential within-person clustering of malaria events, both within and between age strata, by using the “sandwich” estimator as described by Armitage and colleagues [16], which inflates confidence intervals and adjusts significance values as appropriate. We have expressed our comparisons as adjusted incidence rate ratios (IRRs). We investigated the possibility that age might be acting as an effect modifier in the association between malaria and haemoglobin type by comparing models that included or excluded interaction terms between haemoglobin type and age using the Wald test. All analyses were conducted using STATA version 8.0 (StataCorp, Timberlake, London, United Kingdom).
Table 1 Incidence of Clinical Malaria by Age and Haemoglobin Type
a Because this was a longitudinal cohort study, some participants contributed data to more than one age stratum (see text for details).
b See text for definitions.
c Crude incidence of malaria per person-year of observation.
d Adjusted IRRs for malaria in HbAS individuals compared to baseline (HbAA individuals) were calculated individually for each age stratum using a Poisson regression model that included the confounding variables season, study area, ethnic group, and bed net usage. Confidence intervals and significance values were adjusted to take account of clustering of events within individual participants using the “sandwich” estimator as described by Armitage and colleagues [16].
e All participants greater than 10 y old were considered functionally immune and were therefore included as a single category. This group included participants 10–84 y old.
Ethical permission for the study was granted by the Kenya Medical Research Institute National Ethical Review Committee. Individual written informed consent was provided by all study participants or their parents.
Results
Overall, HbAS was almost 40% protective against mild clinical malaria (IRR = 0.62; 95% confidence interval 0.51–0.76; p < 0.001); however, protection appeared to vary with age, increasing from only 20% to almost 60% over the first 10 y of life and returning to around 30% thereafter (Table 1; Figure 1). A similar pattern was seen when data from each of the study areas were analyzed separately. Although we were not able to prove statistically an overall interaction between age and protection over the full range of ages (χ2
5 = 6.46; p = 0.26), the data support the strong impression of acquired protection with age.
Figure 1 The IRR for Malaria in HbAS versus HbAA Children by Age and Genotypic Group
Infants less than 3 mo old were excluded from the baseline group.
Discussion
The mechanism by which HbAS protects against malaria has been the subject of speculation for more than 50 y. While to some extent it probably relates to the physical characteristics of HbAS erythrocytes, a number of studies suggest that HbAS may also enhance the acquisition of natural immunity [10,17–19]; however, establishing this relationship is difficult because immunity to malaria is hard to measure.
To date, no single immune response has been described that reliably predicts protective immunity. As a result, immunity to malaria is usually defined as the ability to control new infections to a level at which they fail to reach a clinical threshold. We therefore reasoned that the best way to find out whether malaria protection by HbAS involves a significant immune component was to see whether protection varies with age. Of the cohort studies that have been reported to date, most have involved repeated cross-sectional sampling rather than active monitoring for clinical events. Moreover, of the studies that have investigated the genotype-specific incidence of mild malaria [10,20–23], all have been either too small, have involved a restricted age range of participants, or have been conducted over too short a period to make it possible to address this important question. Our study is, to our knowledge, the first with sufficient power to observe the protective effect of HbAS over a broad age range. We found that HbAS protection increases throughout the first 10 y of life, returning thereafter to baseline. While it is possible that this observation could result from any factor that both affects malaria risk and varies with age, accelerated immune acquisition seems by far the most likely explanation.
So how might HbAS result in the accelerated acquisition of malaria-specific immunity? A number of mechanisms have been proposed. In common with other red cell genetic defects, enhanced phagocytosis of HbAS erythrocytes infected with ring stage P. falciparum has been demonstrated in vitro [24], a process that appears to be mediated by a mechanism essentially similar to that involved in the phagocytosis of senescent or damaged normal erythrocytes. Experimental data suggest that this process is initiated by enhanced oxidant damage to the erythrocyte membrane and that this leads to the aggregation of band 3 protein and the binding of autologous IgG and complement [24], a mechanism similar to that previously proposed for α thalassaemia [25]. It therefore seems plausible that enhanced immunity could be mediated by the accelerated acquisition of antibodies to altered host antigens expressed on the parasite-infected red cell surface, such as band 3 protein [26]. On the other hand, parasite-derived proteins such as the variant surface antigen P. falciparum erythrocyte membrane protein-1 might represent an alternative target, an hypothesis supported by the raised titres of antibodies directed towards variant antigens seen in HbAS children living in Gambia [10]. These mechanisms need not be mutually exclusive.
As an alternative explanation, it seems possible that by controlling parasite densities during malaria infections [27] innate processes might paradoxically increase the chronicity of individual infections. This hypothesis is supported by the greater number of strains of P. falciparum parasites found in HbAS than HbAA children at cross-sectional survey [28]. By increasing the duration of individual malaria infections HbAS might paradoxically increase host exposure to a variety of antigens capable of inducing malaria-specific immunity. Determining which if any of these mechanisms are involved could lead to a better understanding of malaria immunity more generally.
In our current study we have focused on mild clinical malaria. For accelerated malaria-specific immunity to be relevant to HbAS selection it would have to operate within the period of maximum risk for severe and fatal malaria. In Kilifi, this risk is greatest in children less than 5 y old [29]. It is clear from a recent study conducted in western Kenya [4] that HbAS is strongly protective against severe and fatal malaria within this age range; however, protection by HbAS against both severe malaria anaemia and all-cause mortality was only seen in the age range 2–16 mo. The authors suggested that this may have reflected early protection by maternally transferred immunoglobulins followed by the general acquisition of protective immunity after the age of 16 mo; however, they presented no data regarding the effect of age between these extremes. Given the level of protection conferred by HbAS against severe malaria, it is possible that their study was not sufficiently powered to address this question. The relevance of our observations in mild clinical malaria to the protection afforded by HbAS against severe and fatal malaria therefore remains unknown. While immunity against severe malaria develops significantly more rapidly than immunity to mild clinical attacks, the determinants of each remain poorly understood. We suggest that establishing the role of HbAS in each of these processes may be one route to learning more about the mechanisms involved.
Patient Summary
Background
Sickle cell anemia, which is caused by having two copies of an abnormal gene (hemoglobin S—HbS) that causes red cells to deform easily, occurs more frequently in populations exposed to malaria. Previous work has shown that carrying one normal copy of the gene (HbA) and one copy of the version responsible for sickle cell disease (the combination is called HbAS) may protect against getting malaria; hence, this abnormal gene provides an advantage to some people who carry it. How this protection happens is unclear, but may be due to changes in the way that people with HbAS develop immunity to malaria.
What Did the Authors Do?
One way of working out whether acquired immunity is important in how HbAS protects against malaria is to look at a large population with many different age ranges, all exposed to malaria, and measure how often these individuals get malaria. The authors of this paper looked at 1,054 people in Kenya with an age range from birth up to 84 years, but predominantly aged less than 10 years, who either had HbAS or normal hemoglobin (HbAA). They found that protection of HbAS against mild malaria increased with age from 20% in the first two years of life to a maximum of 56% by the age of ten years, and then decreased to 30% in people older than ten years.
What Do These Findings Mean?
The presence of HbAS is associated with increased acquired immunity to mild malaria. Further work will need to be done to work out how this change in immunity occurs. It is not yet known whether these results are also true for protection against severe malaria, and in any case the protection is only partial; hence, treatment of anyone with malaria, whatever their sickle cell status, is essential.
Where Can I Get More Information?
The Centers for Disease Control and Prevention publish reviews of various conditions. The one on sickle cell disease includes links to other sources of information: http://www.cdc.gov/genomics/hugenet/reviews/sickle.htm
The World Health Organization has a Web page on malaria: http://www.who.int/topics/malaria/en/
TNW and KM are funded by the Wellcome Trust. The sponsors of the study had no involvement in the analysis or interpretation of the data. We thank the study participants and their families for their co-operation. This paper is dedicated to Professor Steve Bennett, a key collaborator who died before completion of the study in April 2003. This paper is published with the permission of the Director of the Kenya Medical Research Institute.
This study received financial support from the Wellcome Trust. TNW and KM are supported by the Wellcome Trust. DJW is supported by a grant from the Leverhulme Trust. DJR is a Howard Hughes International Research Scholar and is supported by the National Blood Transfusion Service, United Kingdom.
Citation: Williams TN, Mwangi TW, Roberts DJ, Alexander ND, Weatherall DJ, et al. (2005) An immune basis for malaria protection by the sickle cell trait. PLoS Med 2(5): e128.
Abbreviations
IRRincidence rate ratio
==== Refs
References
Allison AC Polymorphism and natural selection in human populations Cold Spring Harb Symp Quant Biol 1964 24 137 149
Willcox M Bjorkman A Brohult J Pehrson PO Rombo L A case-control study in northern Liberia of Plasmodium falciparum malaria in haemoglobin S and beta-thalassaemia traits Ann Trop Med Parasitol 1983 77 239 246 6354114
Hill AV Allsopp CE Kwiatkowski D Anstey NM Twumasi P Common west African HLA antigens are associated with protection from severe malaria Nature 1991 352 595 600 1865923
Aidoo M Terlouw DJ Kolczak MS McElroy PD ter Kuile FO Protective effects of the sickle cell gene against malaria morbidity and mortality Lancet 2002 359 1311 1312 11965279
Friedman MJ Erythrocytic mechanism of sickle cell resistance to malaria Proc Natl Acad Sci U S A 1978 75 1994 1997 347452
Pasvol G Weatherall DJ Wilson RJ Cellular mechanism for the protective effect of haemoglobin S against P. falciparum malaria Nature 1978 274 701 703 353566
Luzzatto L Nwachuku-Jarrett ES Reddy S Increased sickling of parasitised erythrocytes as mechanism of resistance against malaria in the sickle-cell trait Lancet 1970 1 319 321 4189578
Roth EF Friedman M Ueda Y Tellez I Trager W Sickling rates of human AS red cells infected in vitro with Plasmodium falciparum malaria Science 1978 202 650 652 360396
Shear HL Roth EF Fabry ME Costantini FD Pachnis A Transgenic mice expressing human sickle hemoglobin are partially resistant to rodent malaria Blood 1993 81 222 226 8417791
Marsh K Otoo L Hayes RJ Carson DC Greenwood BM Antibodies to blood stage antigens of Plasmodium falciparum in rural Gambians and their relation to protection against infection Trans R Soc Trop Med Hyg 1989 83 293 303 2694458
Abu-Zeid YA Abdulhadi NH Theander TG Hviid L Saeed BO Seasonal changes in cell mediated immune responses to soluble Plasmodium falciparum antigens in children with haemoglobin AA and haemoglobin AS Trans R Soc Trop Med Hyg 1992 86 20 22 1566293
Bayoumi RA Abu-Zeid YA Abdulhadi NH Saeed BO Theander TG Cell-mediated immune responses to Plasmodium falciparum purified soluble antigens in sickle-cell trait subjects Immunol Lett 1990 25 243 249 2283154
Mwangi TW Clinical epidemiology of malaria under differing levels of transmission 2003 Oxford Open University 320
Polley SD Mwangi T Kocken CH Thomas AW Dutta S Human antibodies to recombinant protein constructs of Plasmodium falciparum Apical Membrane Antigen 1 (AMA1) and their associations with protection from malaria Vaccine 2004 23 718 728 15542195
Snow RW Nahlen B Palmer A Donnelly CA Gupta S Risk of severe malaria among African infants: Direct evidence of clinical protection during early infancy J Infect Dis 1998 177 819 822 9498474
Armitage P Berry G Matthews JNS Statistical methods in medical research, 4th ed 2001 Oxford Blackwell Scientific Publications 817
Guggenmoos-Holzmann I Bienzle U Luzzatto L
Plasmodium falciparum malaria and human red cells. II. Red cell genetic traits and resistance against malaria Int J Epidemiol 1981 10 16 22 7016776
Edozien JC Boyo AE Morley DC The relationship of serum gamma-globulin concentration to malaria and sickling J Clin Pathol 1960 13 118 123 13819366
Cornille-Brogger R Fleming AF Kagan I Matsushima T Molineaux L Abnormal haemoglobins in the Sudan savanna of Nigeria. II. Immunological response to malaria in normals and subjects with sickle cell trait Ann Trop Med Parasitol 1979 73 173 183 386969
Stirnadel HA Stockle M Felger I Smith T Tanner M Malaria infection and morbidity in infants in relation to genetic polymorphisms in Tanzania Trop Med Int Health 1999 4 187 193 10223213
Jakobsen PH Riley EM Allen SJ Larsen SO Bennett S Differential antibody response of Gambian donors to soluble Plasmodium falciparum antigens Trans R Soc Trop Med Hyg 1991 85 26 32 2068750
Lell B May J Schmidt-Ott RJ Lehman LG Luckner D The role of red blood cell polymorphisms in resistance and susceptibility to malaria Clin Infect Dis 1999 28 794 799 10825041
Le Hesran JY Personne I Personne P Fievet N Dubois B Longitudinal study of Plasmodium falciparum infection and immune responses in infants with or without the sickle cell trait Int J Epidemiol 1999 28 793 798 10480713
Ayi K Turrini F Piga A Arese P Enhanced phagocytosis of ring-parasitized mutant erythrocytes. A common mechanism that may explain protection against falciparum-malaria in sickle-trait and beta-thalassemia-trait Blood 2004 104 3364 3371 15280204
Williams TN Weatherall DJ Newbold CI The membrane characteristics of Plasmodium falciparum -infected and -uninfected heterozygous alpha(0)thalassaemic erythrocytes Br J Haematol 2002 118 663 670 12139762
Hebbel RP Sickle hemoglobin instability: A mechanism for malaria protection Redox Rep 2003 8 238 240 14962356
Roberts DJ Williams TN Haemoglobinopathies and resistance to malaria Redox Rep 2003 8 294 300
Ntoumi F Mercereau-Puijalon O Ossari S Luty A Reltien J
Plasmodium falciparum Sickle-cell trait is associated with higher prevalence of multiple infections in Gabonese children with asymptomatic infections Exp Parasitol 1997 87 39 46 9287956
Marsh K Forster D Waruiru C Mwangi I Winstanley M Indicators of life-threatening malaria in African children N Engl J Med 1995 332 1399 1404 7723795
| 15916466 | PMC1140945 | CC BY | 2021-01-05 10:39:45 | no | PLoS Med. 2005 May 31; 2(5):e128 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020128 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591645910.1371/journal.pmed.0020131PerspectivesCardiology/Cardiac SurgeryPrimary CareIschemic heart diseaseGeneral Practice/Family Practice/Primary CareGuideline adherenceManaging Cardiovascular Risk Factors: The Gap between Evidence and Practice PerspectivesTurnbull Fiona Fiona Turnbull is at the George Institute for International Health, University of Sydney, Australia. E-mail: [email protected]
Competing Interests: The author declares that she has no competing interests.
5 2005 31 5 2005 2 5 e131Copyright: © 2005 Fiona Turnbull.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.
National Trends in Statin Use by Coronary Heart Disease Risk Category
There are clear evidence-based guidelines for managing patients at risk of cardiovascular disease, and yet many doctors don't follow these guidelines.
==== Body
Many large randomised trials have provided an abundance of high-quality evidence for the benefits of managing two major cardiovascular risk factors—high blood pressure and high cholesterol. National guidelines in turn distil this evidence and aid clinicians in making decisions about how best to manage their patients' care. However, despite the existence of clear evidence-based guidelines, many patients at risk of cardiovascular disease who should be receiving treatment are not, while many others who are on treatment are not receiving treatment in line with recommendations. A paper in last month's PLoS Medicine and another in this issue of the journal illustrate such “treatment gaps”.
Two New Studies on the “Treatment Gap”
One of the papers, by Ma et al. [1], uses United States national survey data collected over a ten-year period to obtain estimates of statin use among patients categorized by their risk of coronary heart disease and presence/ absence of hyperlipidaemia. Not surprisingly, the results show a more than 2-fold increase in the proportion of patients with hyperlipidaemia treated with lipid lowering agents between 1992 and 2002— statins accounting for most of this increase. However, even at the point of highest treatment uptake, only around half of patients with hyperlipidaemia were receiving treatment.
Even more striking are the results for the use of statins in patients categorized by their cardiovascular risk. Among patients at high risk, the absolute maximum proportion of individuals receiving treatment at the end of the ten-year review period (i.e., in 2002) was only 19%. Additional analyses suggest that lower statin use in at-risk patients was associated with younger age, female gender, African-American background, and care by non-cardiologists. The authors appropriately conclude that statins remain underused —particularly among patients who have normal lipid levels but who are otherwise at high cardiovascular risk [2].
A similar evidence–practice gap, this time for blood pressure, is highlighted in the other article, by Morgan et al. [3]. In this paper, data from public, medical, hospital, and pharmaceutical programs in British Columbia are used to determine trends in the use of thiazide diuretics compared with other, more costly agents as a first-line treatment to lower blood pressure among older, newly treated patients with hypertension. The results show that only around one-third of patients received thiazide diuretics. Furthermore, even in the absence of certain comorbidities—such as diabetes, which might influence a clinician to choose an alternative agent—thiazides were used in no more than 45% of older eligible patients.
Compared with newer agents such as angiotensin receptor blockers and calcium antagonists, which cost upwards of US$1.00/day, thiazides remain the cheapest blood pressure lowering agents, costing less than $0.01/day. The authors reasonably argue that as long as thiazides remain at least equivalent to other blood pressure lowering agents in terms of reducing cardiovascular mortality and morbidity [4], their preferential use as a first-line agent can be justified on the basis of their low cost.
Narrowing the Gap
Why do such gaps between evidence and practice exist? In 2002, around 800 primary care physicians in five European countries were surveyed to assess the acceptance and or implementation of treatment guidelines for high cholesterol and coronary heart disease (the Reassessing European Attitudes about Cardiovascular Treatment survey) [5]. Although most (89%) of those interviewed acknowledged the need for formal guidelines, and a similar proportion agreed with the content of current guidelines, only 18% of physicians believed that guidelines were being implemented to a major extent, indicating a problem with either their understanding or implementation. The barriers to implementation that were most commonly cited by physicians in the survey are shown in Table 1.
Table 1 The Reassessing European Attitudes about Cardiovascular Treatment Survey: Most Commonly Cited Barriers to Implementation of Coronary Heart Disease Guidelines
Data from [5]
Perhaps the two most important means by which improved use of treatment guidelines can be achieved are (1) improving the understanding of the basic concepts that underpin them and (2) reducing the number and complexity of the main messages. In terms of addressing the first of these, an understanding of the concept of “absolute risk”—the probability of a patient developing a cardiovascular event over a specified time period—is crucial.
An absolute risk approach to cardiovascular prevention acknowledges that the presence of small or moderate elevations of multiple risk factors often confer greater risk of cardiovascular disease than an extreme elevation of a single risk factor. Furthermore, the nature of the association between blood pressure, cholesterol, and cardiovascular disease implies that a given reduction in the level of the risk factor, regardless of baseline level, will reduce cardiovascular risk by a constant proportion. Therefore, the goal of blood pressure lowering and lipid lowering is not to “normalize” levels but to reduce them as much as possible, and this means targeting everyone at high risk as determined by age or known cardiovascular disease rather than by the level of the risk factors [6]. This approach (“the lower, the better”) to both blood pressure and cholesterol management in high-risk individuals has been supported by recent meta-analyses and large trials [7,8].
Adopting an absolute-risk-based approach requires a paradigm shift and challenges the way doctors have traditionally made treatment decisions based on single risk factor levels. Although there is some evidence that clinical practice is conforming to the notion of risk stratification, other studies suggest that a large proportion of clinicians, particularly those in primary care, still do not use risk charts [5]. While it is acknowledged that the absolute risk approach has some limitations, integration of this approach into clinical care will be the key to future major gains in the prevention of cardiovascular disease.
Citation: Turnbull F (2005) Managing cardiovascular risk factors: The gap between evidence and practice. PLoS Med 2(5): e131.
==== Refs
References
Ma J Sehgal NL Ayanian JZ Stafford RS National trends in statin use by coronary heart disease risk category PLoS Med 2005 2 e123 15916463
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. 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
Morgan S Bassett KL Wright JM Yan L First-line first? Trends in thiazide prescribing for hypertensive seniors PLoS Med 2005 2 e80 15839739
Whitworth JA World Health Organization, International Society of Hypertension Writing Group World Health Organization (WHO)/International Society of Hypertension (ISH) statement on management of hypertension J Hypertens 2003 21 1983 1992 14597836
Hobbs F Erhardt L Acceptance of guideline recommendations and perceived implementation of coronary heart disease prevention among primary care physicians in five European countries: The Reassessing European Attitudes about Cardiovascular Treatment (REACT) survey Fam Pract 2002 19 596 604 12429661
Law MR Wald NJ Risk factor thresholds: Their existence under scrutiny BMJ 2002 324 1570 1576 12089098
Blood Pressure Lowering Treatment Trialists' Collaboration Effects of different blood pressure lowering regimens on major cardiovascular events: Second cycle of prospectively designed overviews Lancet 2003 362 1527 1535 14615107
Larosa JC Grundy SM Waters DD Shear C Barter P Intensive lipid lowering with atorvastatin in patients with stable coronary disease N Engl J Med 2005 In press
| 15916459 | PMC1140946 | CC BY | 2021-01-05 10:39:47 | no | PLoS Med. 2005 May 31; 2(5):e131 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020131 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646810.1371/journal.pmed.0020135Research ArticleCardiology/Cardiac SurgeryNutritionCardiovascular MedicineNutrition and MetabolismNutrition policyEffect of Homocysteine-Lowering Nutrients on Blood Lipids: Results from Four Randomised, Placebo-Controlled Studies in Healthy Humans Homocysteine Lowering and Blood LipidsOlthof Margreet R
1
*van Vliet Trinette
2
Verhoef Petra
1
Zock Peter L
1
Katan Martijn B
1
1Wageningen Centre for Food Sciences and Division of Human Nutrition, Wageningen UniversityWageningenthe Netherlands2TNO Quality of LifeZeistthe NetherlandsLudwig David Academic EditorChildren's Hospital BostonUnited States of America
Competing Interests: The Wageningen Centre for Food Sciences (WCFS) is an alliance of six industrial partners (AVEBE, Cosun, CSM, DSM, Unilever, and NZO) and four research organizations (Wageningen University and Research Centre, Maastricht University, NIZO Food Research, and TNO Quality of Life). Forty-seven percent of funding comes from the Dutch government, 29% from the industrial partners, and 24% from the research institutes. The partners have no say over the mode or content of publications resulting from work at WCFS but have the right to seek patent protection so that findings can be commercialized. MBK is a member of the editorial board of PLoS Medicine.
Author Contributions: MRO, TvV, PV, and MBK designed the study. MRO analyzed the data. MRO and TvV enrolled patients. MRO, TvV, PV, PLZ, and MBK contributed to writing the paper.
*To whom correspondence should be addressed. E-mail: [email protected] 2005 31 5 2005 2 5 e13528 7 2004 24 3 2005 Copyright: © 2005 Olthof 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.
Comparison of Homocysteine-Lowering Drugs
Background
Betaine (trimethylglycine) lowers plasma homocysteine, a possible risk factor for cardiovascular disease. However, studies in renal patients and in obese individuals who are on a weight-loss diet suggest that betaine supplementation raises blood cholesterol; data in healthy individuals are lacking. Such an effect on cholesterol would counteract any favourable effect on homocysteine. We therefore investigated the effect of betaine, of its precursor choline in the form of phosphatidylcholine, and of the classical homocysteine-lowering vitamin folic acid on blood lipid concentrations in healthy humans.
Methods and Findings
We measured blood lipids in four placebo-controlled, randomised intervention studies that examined the effect of betaine (three studies, n = 151), folic acid (two studies, n = 75), and phosphatidylcholine (one study, n = 26) on plasma homocysteine concentrations. We combined blood lipid data from the individual studies and calculated a weighted mean change in blood lipid concentrations relative to placebo. Betaine supplementation (6 g/d) for 6 wk increased blood LDL cholesterol concentrations by 0.36 mmol/l (95% confidence interval: 0.25–0.46), and triacylglycerol concentrations by 0.14 mmol/l (0.04–0.23) relative to placebo. The ratio of total to HDL cholesterol increased by 0.23 (0.14–0.32). Concentrations of HDL cholesterol were not affected. Doses of betaine lower than 6 g/d also raised LDL cholesterol, but these changes were not statistically significant. Further, the effect of betaine on LDL cholesterol was already evident after 2 wk of intervention. Phosphatidylcholine supplementation (providing approximately 2.6 g/d of choline) for 2 wk increased triacylglycerol concentrations by 0.14 mmol/l (0.06–0.21), but did not affect cholesterol concentrations. Folic acid supplementation (0.8 mg/d) had no effect on lipid concentrations.
Conclusions
Betaine supplementation increased blood LDL cholesterol and triacylglycerol concentrations in healthy humans, which agrees with the limited previous data. The adverse effects on blood lipids may undo the potential benefits for cardiovascular health of betaine supplementation through homocysteine lowering. In our study phosphatidylcholine supplementation slightly increased triacylglycerol concentrations in healthy humans. Previous studies of phosphatidylcholine and blood lipids showed no clear effect. Thus the effect of phosphatidylcholine supplementation on blood lipids remains inconclusive, but is probably not large.
Folic acid supplementation does not seem to affect blood lipids and therefore remains the preferred treatment for lowering of blood homocysteine concentrations.
Lowering homocysteine might reduce the risk for heart disease. Betaine seems to have an adverse effect on blood lipids. This would make it less suitable than folic acid, which does not affect blood lipids.
==== Body
Introduction
Cardiovascular disease (CVD) is a major cause of morbidity and mortality in Western societies. Diet may play a crucial role in the aetiology of CVD. However, clinical trials that test the effects of diet on CVD are complicated to conduct. Hence, risk factors that are predictive of CVD risk are frequently used to test the effects of diet on disease prevention. Examples of such risk factors are homocysteine concentrations and cholesterol concentrations. A limitation of studying effects of dietary changes on risk factors is that often a single factor is measured, whereas multiple factors might be affected by changes in diet. Whether homocysteine lowering indeed lowers risk of CVD is not yet established, but evidence for a causal relationship is accumulating [1–3]. The effects of increased blood low-density lipoprotein (LDL) cholesterol concentrations on CVD risk are well established.
Betaine is used as therapy to lower plasma homocysteine in hyperhomocysteinemic patients with genetic defects in their homocysteine metabolism who are unresponsive to pyridoxine, folic acid, and vitamin B12 [4–6]. It also lowers plasma homocysteine in healthy humans with homocysteine concentrations in the normal range [7–9]. Betaine occurs naturally in the diet, or it can be produced endogenously through oxidation of choline. Dietary intake of betaine is estimated at 0.5–2 g/d generally, and major food sources are wheat products (e.g., bread), beets, and spinach [10]. Recently we showed that supplementation with 1.5 g/d of betaine, which is in the range of daily intake, lowers fasting and post-methionine plasma homocysteine concentrations in healthy humans [8]. The homocysteine-lowering properties of betaine would predict a decrease in risk of CVD. However, betaine and its precursor choline are also involved in lipid metabolism. In animal studies betaine supplementation increases serum cholesterol concentrations relative to control [11,12]. Studies in humans have also found that betaine supplementation increases serum cholesterol in obese individuals who are on a weight-loss diet, and in patients with chronic renal failure [9,13]. However, in these studies betaine was not the only intervention, so the observed effects on lipid concentrations could have been due to factors other than betaine supplementation. Furthermore, in the study with obese individuals, weight changes might have contributed to the changes in serum lipids [9]. The effects of betaine on lipid metabolism in healthy humans are not well established.
Choline, the precursor of betaine, also occurs naturally in foods, mainly as phosphatidylcholine. Daily intake of choline is generally estimated to be 0.3–1 g [14]. Phosphatidylcholine is an essential component of very low density lipoproteins (VLDLs) and is therefore involved in lipid metabolism. Animal and human studies indicate that choline deficiency decreases serum cholesterol, which can be prevented by addition of choline [15–17]. Data on the effects of choline supplementation on blood lipids in humans are scarce.
Increases in blood lipid concentrations due to betaine or choline supplementation might offset the potential beneficial effects of their homocysteine-lowering properties. Folic acid is the most commonly used homocysteine-lowering agent. The betaine remethylation and folic acid remethylation pathways are interrelated (Figure 1), so effects of folic acid supplementation on lipid metabolism cannot be excluded [18].
Figure 1 Role of Folic Acid, Choline, and Betaine in Homocysteine Metabolism and in Phosphatidylcholine Metabolism
Phosphatidylcholine is necessary for synthesis of VLDL, which exports lipids from the liver.
In this study we investigated the effects of supplementation of three homocysteine-lowering nutrients, betaine, choline (as phosphatidylcholine), and folic acid, on blood lipid concentrations in healthy humans. For this, we combined data of four placebo-controlled studies in healthy humans, performed by our group. These are the only available randomized, controlled studies in healthy humans on this matter that we are aware of.
Methods
Data were derived from four placebo-controlled intervention studies (studies 1, 2, 3 and 4) into the effects of betaine, phosphatidylcholine, and folic acid on homocysteine concentrations (Table 1). All studies are reported in accordance with the CONSORT guidelines (Tables S1–S4).
Table 1 Overview of Study Designs into the Effects of Betaine, Folic Acid, and Phosphatidylcholine on Blood Lipid Concentrations
Study 1
Study participants and design
The primary endpoint of this study was plasma homocysteine concentrations, and these results have been published [7]. Blood lipid measurements were done post-hoc. Study 1 was conducted according to Good Clinical Practice guidelines at TNO Quality of Life (Zeist, the Netherlands). The local medical ethics committee approved the protocol, and all volunteers gave their written informed consent. Volunteers were recruited from April to May 2000. Eligible volunteers were healthy as assessed by routine medical screening and a general health questionnaire, had plasma total homocysteine concentrations below 25 μmol/l, had no history of cardiovascular disease, and had not used vitamin B supplements more than once a week in the 3 mo before entering the study. Out of the eligible participants, the 36 participants (15 males and 21 females, age 30–50 y) with the highest plasma total homocysteine concentrations (range 8.9 to 21.0 μmol/l) were included in this placebo-controlled, double-blind parallel study (Figure 2). Participants were stratified by gender and plasma homocysteine concentration, and then randomly assigned to one of three treatments for 6 wk: (a) 6 g/d of betaine (n = 12) (BUFA Pharmaceutical Products, Uitgeest, the Netherlands), (b) 0.8 mg/d of folic acid (n = 12), or (c) placebo (n = 12). One person at the local pharmacy assigned codes to the study treatments and provided the key in a sealed envelope to the principal investigator at TNO. The statistician then randomly allocated treatment codes to participant numbers. Randomization was done using a computerized procedure that produced combinations based on random seed numbers. The statistician supplied the medical investigator with sealed envelopes with the treatment allocation per participant. The participants and all others involved in this study were unaware of treatment allocation. The principal investigator performed unblinding of the treatment allocation after the study had ended, laboratory analyses were complete, and datasets were locked. The study supplements were dissolved in water and ingested on two daily occasions, one half of the daily dose after breakfast and the other half after the evening meal. Throughout the study, participants were asked to refrain from eating products based on animal liver and to consume no more than two eggs per week. (Eggs and liver are major sources of betaine and of choline, the dietary precursor for betaine.)
Figure 2 Flow Diagram of Participant Progress through Study 1
Blood sampling and blood lipid analysis
Venous blood was taken from an anticubital vein in the forearm following an overnight fast at baseline and after 2 wk and 6 wk of intervention. Immediately after collection, blood was mixed well and put on ice. Within 30 min samples were centrifuged for 10 min at 2,000 g at 4 °C. Samples were stored below −20 °C. Samples were coded to hide the identity and treatment of participants. All samples obtained from one participant were analysed in the same run. We measured blood lipid concentrations in fasting EDTA plasma samples at baseline and after 2 wk of intervention, and in fasting serum samples collected at baseline and after 6 wk of intervention. Concentrations of total cholesterol, high-density lipoprotein (HDL) cholesterol, and triacylglycerol were measured with a Hitachi (Tokyo, Japan) 911 analyzer and enzymatic assays of Roche (Basel, Switzerland). LDL cholesterol concentrations were calculated with the formula of Friedewald et al. [19].
Study 2
Participants and design
The primary endpoint of this study was plasma homocysteine concentrations, and these results have been published [8]. Blood lipid measurements were done post-hoc. Study 2 was conducted according to Good Clinical Practice guidelines at TNO Quality of Life (Zeist, the Netherlands). The local medical ethics committee approved the protocol, and all volunteers gave their written informed consent. Volunteers were recruited from September to October 2001. Eligible volunteers were healthy as assessed by a general health and lifestyle questionnaire, blood pressure measurement, and blood analyses of haematology, homocysteine, B vitamins, liver enzymes, creatinine, glucose, and lipids. Plasma total homocysteine concentrations were below 25 μmol/l. Volunteers had no history of CVD, and had not used vitamin B supplements more than once a week in the 3 mo before entering the study. Out of the eligible participants, the 76 participants (44 males and 32 females, age 30–50 y) with the highest plasma total homocysteine concentrations (range 8.4 to 22.2 μmol/l) were included in this placebo-controlled, double-blind parallel trial (Figure 3). Volunteers were stratified by gender, plasma homocysteine concentration, blood pressure, and smoking (yes or no), and then randomly assigned to one of four treatment groups for 6 wk: (a) 1.5 g/d of anhydrous betaine (BUFA) (n = 19), (b) 3 g/d of betaine (n = 19), (c) 6 g/d of betaine (n = 19), or (d) placebo (n = 19). One person at the local pharmacy assigned codes to the study treatments and provided the key in a sealed envelope to the principal investigator at TNO. The statistician then randomly allocated treatment codes to the participant numbers. Randomization was done using a computerized procedure that produced combinations based on random seed numbers. The statistician supplied the principal investigator with sealed envelopes with the treatment allocation per participant. The participants and all others involved in this study were unaware of treatment allocation. The statistician performed unblinding of the treatment allocation after the study had ended, laboratory analyses were complete, and datasets were locked. The study supplements were dissolved in water and ingested on two daily occasions, one half of the daily dose after breakfast and the other half after the evening meal. Throughout the study participants were asked not to consume liver products more than twice a week, and not to consume more than two eggs per week. Six grams per day of betaine is the lowest therapeutic dose used to lower plasma homocysteine in hyperhomocysteinemic patients with genetic defects in their homocysteine metabolism [20]. We chose to study doses of betaine lower than 6 g/d, e.g., 1.5 g/d and 3 g/d, because these are within the range of intake of betaine with foods. Intake of betaine with foods is generally estimated at 0.5–2 g/d.
Figure 3 Flow Diagram of Participant Progress through Study 2
Blood sampling and blood lipid analysis
Venous blood was taken from the anticubital vein in the forearm following an overnight fast at baseline and after 2 wk and 6 wk of intervention. Samples were mixed and put on ice immediately after collection. Within 30 min samples were centrifuged for 10 min at 2,000 g at 4 °C. Samples were stored below −18 °C. Samples were coded to hide the identity and treatment of participants. All samples obtained from one participant were analyzed in the same run. We measured blood lipid concentrations in fasting EDTA plasma samples at baseline and after 2 wk of intervention, and in fasting heparin plasma samples collected at baseline and after 6 wk of intervention. Concentrations of total cholesterol, HDL cholesterol, and triacylglycerol were measured with a Hitachi 911 analyzer and enzymatic assays of Roche. LDL cholesterol concentrations were calculated with the formula of Friedewald et al. [19].
Study 3
Participants and design
The primary endpoints of this study were plasma homocysteine concentrations and vascular function. Blood lipid measurements were planned before the study took place, but power analysis was based on changes in the primary endpoint vascular function, and not on changes in blood lipids. The original study protocol of this study can be found in Protocol S1, and the trial was registered at clinicaltrials.gov under identifier NCT00102843. The study was conducted at the division of Human Nutrition, Wageningen University (Wageningen, the Netherlands). The local medical ethics committee approved the protocol, and all volunteers gave their written informed consent. Volunteers were recruited from June to September 2002. Eligible volunteers were healthy as assessed by routine medical screening and a general health questionnaire, had plasma total homocysteine concentrations below 26 μmol/l, had no history of CVD; and had not used vitamin B supplements more than once a week in the 3 mo before entering the study. Out of the eligible participants, the 40 participants (23 males and 17 females, age 50–70 y) with the highest plasma total homocysteine concentrations (range 10.2 to 21.7 μmol/l) were included in this placebo-controlled, double-blind cross-over trial (Figure 4). Participants were randomly assigned to one out of six treatment orders, and they received each of the following supplements for 6 wk, with a 6-wk wash out in between: (a) 6 g/d of betaine (BUFA), (b) 0.8 mg/d of folic acid, or (c) placebo. A person not further involved in the study assigned codes to the study treatments, and randomly allocated the selected participants to one out of six treatment orders, according to a computer-generated randomization list, and kept the key in a sealed envelope. The participants and all others involved in this study were unaware of treatment allocation. The principal investigator performed unblinding of the treatment allocation after the study had ended and laboratory analyses were complete. The study supplements were dissolved in water and ingested twice per day, one half of the daily dose after breakfast and the other half after the evening meal. During the study participants were not allowed to consume supplements containing B vitamins, antioxidant vitamins (A, beta-carotene, C, and E), or omega-3 fatty acids/fish oil supplements.
Figure 4 Flow Diagram of Participant Progress through Study 3
Blood sampling and laboratory analyses
Venous blood was taken from the antecubital vein following an overnight fast on days 41 and 43 of each treatment period. Blood was collected in vacutainer tubes containing clot activator and a gel to separate serum and cells. About 30 min after collection, samples were centrifuged for 15 min at 2,000 g at 4 °C. All serum samples were stored below −70 °C. Samples were coded to hide the identity and treatment of participants. All samples obtained from one participant were analyzed in the same run. In serum samples we measured concentrations of total cholesterol, HDL cholesterol, and triacylglycerols on a Synchron LX 20 system (Beckman Coulter, Mijdrecht, the Netherlands). LDL cholesterol concentrations were calculated with the formula of Friedewald et al. [19].
Study 4
Participants and design
The primary endpoint of study 4 was plasma homocysteine concentrations. Blood lipid measurements were planned before the study took place, but power analysis was based on changes in the primary endpoint homocysteine concentrations, and not on changes in blood lipids. The original study protocol of this study can be found in Protocols S2–S4, and the trial was registered at clinicaltrials.gov under identifier NCT00102232. This study was conducted according to Good Clinical Practice guidelines at TNO Quality of Life (Zeist, the Netherlands). The local medical ethics committee approved the protocol, and all volunteers gave their written informed consent. Volunteers were recruited from March to May 2003. Eligible volunteers were healthy as assessed by physical examination, a general health and lifestyle questionnaire, blood pressure measurement, routine clinical laboratory tests, and blood analyses of homocysteine and B vitamins. Plasma homocysteine concentrations were below 26 μmol/l. Volunteers had no history of CVD, and had not used vitamin B supplements, lecithin, or supplements containing choline, choline derivatives, or betaine more than once a week during the 1 mo before screening. Out of the eligible men, the 26 men between 50–71 y of age with the highest plasma homocysteine concentrations (range 11.0 to 23.1 μmol/l) were included in this placebo-controlled, double-blind cross-over trial (Figure 5). Participants were stratified by plasma homocysteine concentrations at screening and by smoking habits and then randomly assigned to one of two treatment orders; they received each of the following supplements for 2 wk, with a 2-wk wash out in between: (a) 34.0 g of a soybean lecithin preparation, in which phosphatidylcholine is the only phospholipid (PhosChol Nutrasal, Oxford, Connecticut, United States) and (b) placebo oil, which consisted of 25.5 g of a mixture of edible oils that mimicked the fatty acid composition of the phosphatidylcholine supplement (provided by Unilever Research Laboratory, Vlaardingen, the Netherlands) (Table 2). A person not further involved in the study assigned codes to the study treatments and provided the key in a sealed envelope to the principal investigator at TNO. The statistician randomly allocated the selected participants to one of the two treatment orders. Randomization was done using a computerized procedure that produced combinations based on random seed numbers. The statistician supplied the principal investigator with sealed envelopes with the treatment allocation per participant. The participants and all others involved in this study were unaware of treatment allocation. The statistician performed unblinding of the treatment allocation after the study had ended, laboratory analyses were complete, and datasets were locked. Phosphatidylcholine and placebo supplements were matched for fat content and fatty acid composition (Table 2). The amount of choline in 34 g of the phosphatidylcholine supplement was 2.6 g, as measured by Koc et al. [21] and by TNO. The daily dose of 2.6 g of choline is well below the current tolerable upper intake level of 3.5 g of choline per day for adults [22]. Participants ingested half of the daily supplement dose two times per day (i.e., at breakfast and at dinner). The individual portions of half the daily dose of the supplements were mixed with 200 ml of custard. Participants returned 99.5% of the bowls empty, which indicated good compliance. From 2 wk before the start of the study until the end of the study participants were not allowed to consume food products rich in lecithin or betaine.
Figure 5 Flow Diagram of Participant Progress through Study 4
Table 2 Composition of Treatment Supplements in Study 4
a PhosChol consisted of 80% of a soybean lecithin extract (“Phosal 75A”), 18% medium-chain triglyceride mix, and traces of D alpha-tocopherol and paraben blend. The soybean lecithin extract consisted of 8% moisture, 3% ash, 10% protein, 71% fat (including phosphatidylcholine), 9% carbohydrates, and traces of vitamins and minerals. Folic acid content of the soybean lecithin extract was 195 μg/kg, which is approximately 5 μg per 34 g of phosphatidylcholine oil (calculated from data provided by the manufacturer). This amount is negligible relative to daily intake of folic acid (100–200 μg/d).
b Placebo consisted of a mixture of linola (low-linolenic and high-linoleic linseed) oil (70%), rapeseed oil (21%), and medium-chain triglycerides mix (9%).
c The choline content of PhosChol was measured in two different labs. The lab of S. H. Zeisel [21] measured choline as phosphatidylcholine in two samples of PhosChol. Concentrations of phosphatidylcholine in the two samples were 800 and 801 μmol of phosphatidylcholine per gram of sample: this is approximately 20.6 g of phosphatidylcholine (molecular weight 758), or 2.8 g of choline in 34 g of PhosChol. Total choline content was also measured in four samples of PhosChol at TNO (see Methods). Mean choline content was 74.8 mg/g: this is 2.5 g of choline in 34 g of PhosChol. The mean choline content from all six samples analyzed was 2.6 g/d.
d Fatty acids were measured by gas chromatography of fatty acid methyl esters [50].Trinonadecanoin (C19:0) was used as a reference compound to calculate the amounts of individual fatty acids.
Blood sampling and laboratory analyses
Venous blood was taken from the antecubital vein following an overnight fast on days 13 and 15 of each treatment period. Blood was collected in vacutainer tubes containing clot activator and a gel to separate serum and cells. About 30 min after collection, samples were centrifuged for 15 min at 2,000 g at 4 °C. All serum samples were stored below −70 °C. Samples were coded to hide the identity and treatment of participants. All samples obtained from one participant were analyzed in the same run. In serum samples we measured concentrations of total cholesterol, HDL cholesterol, and triacylglycerols with a Hitachi 911 analyzer and enzymatic assays of Roche. LDL cholesterol concentrations were calculated with the formula of Friedewald et al. [19]. For the measurement of choline at TNO, choline was liberated from the sample by boiling the sample for 4 h with nitric acid (17 v/v%). After filtration of the sample, the pH of the extract was raised to 9.0 with sodium hydroxide solution. To an aliquot part of the extract a solution of ammonium reineckate in methanol was added, which formed the insoluble choline reineckate. The precipitate was filtered off and dissolved in acetone. The color of the acetone solution was determined with a spectrophotometer at a wavelength of 530 nm.
Statistics
The sample size of each study was based on power analysis for the primary outcome measure. To exclude effects of the matrix in which lipids were measured [23], all changes in lipid concentrations within persons or studies were based on measurements within the same matrix. So choice of matrix could not have been responsible for differences in results between the studies. First we calculated the mean change in lipid concentrations relative to placebo after supplementation with the treatments for each study. Means were compared with the General Linear Models procedure in SAS (ANOVA; SAS version 6.12, SAS Institute, Cary, North Carolina, United States). For the parallel studies 1 and 2, Dunnett's two-tailed t-test was used to test whether any of the treatments significantly differed from placebo. For the cross-over studies 3 and 4, a paired Student's t-test was used to test whether any of the treatments significantly differed from placebo. Then we combined the lipid data of the studies with betaine or folic acid supplementation. For this we calculated a combined effect of all four studies, which was a weighted mean and 95% confidence intervals of the change in lipid concentrations relative to placebo after betaine supplementation (studies 1, 2, and 3), and after folic acid supplementation (studies 1 and 3) according to the method used by Curtin et al. [24] (Tables 3 and 4). We tested whether the weighted mean changes differed from zero with a t-test. For all tests mean differences were considered statistically significant at p < 0.05.
Table 3 Mean Changes in Blood Lipid Concentrations and the Ratio of Total to HDL Cholesterol in Fasting State in Studies 1, 2, and 3, and the Combined Effect of Supplementation with 6 g/d of Betaine
Study 1 is a parallel study with 12 participants per group; study 2 is a parallel study with 19 participants per group; study 3 is a cross-over study with 39 participants.
a Significantly different from placebo, p < 0.05.
Table 4 Mean Changes in Blood Lipid Concentrations and the Ratio of Total to HDL Cholesterol in Fasting State in Studies 1 and 3, and the Combined Effect of Supplementation with 0.8 mg/d of Folic Acid
Study 1 is a parallel study with 12 participants per group; study 3 is a cross-over study with 39 participants. Lipid concentrations of the placebo group in study 1 are presented in Table 3.
a Significantly different from placebo, p < 0.05
Results
All participants completed the study, except for one male participant who withdrew from study 3 due to emigration. No serious adverse events were reported in any of the studies. Non-serious adverse events are shown in Table S5.
Supplementation of 6 g/d of betaine for 6 wk increased total cholesterol concentrations by 0.42 mmol/l (8%) relative to placebo (see Table 3). This increase was almost completely accounted for by an increase in LDL cholesterol concentrations of 0.36 mmol/l (11%). Betaine increased triacylglycerol concentrations by 0.14 mmol/l (13%). HDL cholesterol concentrations did not change, but the ratio of total to HDL cholesterol concentration increased by 0.23 (6%). The increases in LDL cholesterol concentrations in the groups that ingested 1.5 g/d of betaine and 3 g/d of betaine for 6 wk were higher than in the placebo group, but these changes did not reach statistical significance (Figure 6). The effects of betaine supplementation on LDL cholesterol concentrations were already evident after 2 wk of supplementation in studies 1 and 2.
Figure 6 Mean Change Relative to Placebo in LDL Cholesterol Concentrations
Compared are the values after participants had ingested 1.5 g/d of betaine (study 2, n = 19), 3 g/d of betaine (study 2, n = 18), or 6 g/d of betaine (studies 1 and 2 combined, n = 31) after 2 wk (+95% confidence interval) and 6 wk (–95% confidence interval). In the group that ingested 3 g/d of betaine one participant missed the blood collection before treatment.
Supplementation of 0.8 mg/d of folic acid for 6 wk did not affect lipid concentrations in healthy volunteers (see Table 4).
Supplementation with 2.6 g/d of choline (provided as phosphatidylcholine) for 2 wk increased serum triacylglycerols by 0.14 mmol/l (8%), but did not affect cholesterol concentrations (Table 5).
Table 5 Concentrations of Serum Lipids (mmol/l) and the Ratio of Total to HDL Cholesterol in Fasting State after 26 Healthy Men Had Ingested Phosphatidylcholine Providing ~2.6 g/d of Choline or Placebo for 2 Wk
a Data from two participants were missing because serum triacylglycerol concentrations were >4.5 mmol/l and under these circumstances calculation of LDL concentrations with the Friedewald formula [19] gives unreliable results.
b Results were based on log-transformed values, because values were not normally distributed.
c Significantly different from placebo, p < 0.05.
Discussion
We found that supplementation of 6 g/d of betaine for 6 wk consistently increased total cholesterol concentrations, mainly LDL cholesterol, and triacylglycerol concentrations. Phosphatidylcholine supplementation for 2 wk increased only serum triacylglycerol concentrations. Folic acid supplementation did not affect lipid concentrations.
The three studies on betaine supplementation were not originally designed to investigate changes in blood lipids. Therefore, we did not perform power calculations for changes in blood lipids for each study beforehand. We combined the lipid data of the three studies to have a better estimate of the effects of betaine on blood lipids, and thereby increase the statistical power afterwards. Similarly we did not do power calculations for effects of folic acid on blood lipids. The increase in LDL cholesterol concentrations with betaine supplementation that we found in healthy humans is in accordance with results from previous studies in renal patients [13] and obese individuals who were on a weight-loss diet [9]. Doses of betaine lower than 6 g/d also raised LDL cholesterol, although this did not reach statistical significance (Figure 6). Our study was not designed to investigate the dose–response relationship between betaine doses and changes in LDL cholesterol. Establishing the dose–response relationship between betaine and LDL cholesterol might require more than three betaine dose groups. Furthermore, betaine supplementation affected blood lipids already after 2 wk of intervention (Figure 6). The increase in triacylglycerol concentrations with betaine supplementation is well in line with the results of Schwab et al. in obese individuals [9]. In the obese individuals betaine supplementation increased triacylglycerol concentrations by approximately 12%, although this increase did not reach statistical significance.
Our observations should be of specific concern for clinical practice as well as for the general population. Patients with inborn errors in the enzymes involved in homocysteine metabolism and who are not or only partially responsive to pyridoxine are generally prescribed betaine in doses of 6 g/d or higher in order to lower their plasma homocysteine [25]. Our data imply that betaine treatment in these patients may have adverse effects on lipid concentrations. No one has systematically reported changes in serum cholesterol concentrations in hyperhomocysteinemic patients who are on betaine treatment, although this may be warranted. Clearly in these patients, the benefit of the substantial homocysteine-lowering effect of betaine outweighs the effect of any modest lipid-related changes. Furthermore, our data imply that among the general population intake of extra betaine, e.g., from supplements or through increased intake of betaine-rich foods, may not bring the expected benefits for CVD prevention predicted from the homocysteine-lowering effect of betaine.
Phosphatidylcholine supplementation increased serum triacylglycerols, but did not affect serum cholesterol concentrations. We supplemented phosphatidylcholine for only 2 wk, and it is possible that this was too short to induce changes in serum cholesterol, although betaine supplementation did increase blood cholesterol concentrations within 2 wk. We carefully matched the fat content and fatty acid composition of the control and the phosphatidylcholine supplements. Previous studies that also attempted to control for the fatty acid component of lecithin (the trivial name for phosphatidylcholine) [26–30], or that tested choline bitartrate [31], suggest that there is little or no effect of the choline moiety of lecithin on blood lipids, provided that the fatty acids in lecithin are balanced in the placebo treatment. Only two of the studies that investigated phosphatidylcholine included a concurrent control group [27,30]. In the randomized cross-over study of Childs et al. [27] triacylglycerol concentrations in 12 normolipidemic individuals were 0.01 mmol/l lower after they had consumed lecithin for 3 wk than after they had consumed placebo. Oosthuizen et al. [30] found in a parallel study with seven hyperlipidemic men per group that the median triacylglycerol concentration decreased by 0.19 mmol/l in the lecithin group relative to the placebo group after 2 wk of intervention. After 4 wk of intervention this difference had disappeared. The increase in triacylglycerol concentrations with phosphatidylcholine treatment in our study could be a chance finding. However, our study was larger than previous studies, we used a cross-over design with a wash out period, and participants received supplements in random order. Moreover, the dose of phosphatidylcholine we used was higher than in the previous studies. Therefore, our study leaves open a real possibility that the choline moiety of phosphatidylcholine in high doses raises serum triacylglycerol concentrations.
We propose that the mechanism by which betaine and phosphatidylcholine supplementation increase blood lipid concentrations involves increased export of lipids from the liver into the circulation. Lipids are exported from the liver through VLDL particles, which consist of a triglyceride core packaged in an amphipatic shell of phospholipids, mostly phosphatidylcholine, and apolipoprotein B-100. More VLDL will increase transport of lipids from the liver into the blood, and concentrations of lipids in the liver decrease accordingly [11]. Betaine and phosphatidylcholine supplementation could increase VLDL synthesis via increased availability of phosphatidylcholine, an essential component of VLDL (see Figure 1) [32]. Furthermore, betaine might increase phosphatidylcholine synthesis owing to increased availability of S-adenosylmethionine via increased remethylation of homocysteine into methionine, because S-adenosylmethionine is the methyl donor for the formation of phosphatidylcholine out of phosphatidylethanolamine [33]. In addition, betaine supplementation could promote phosphatidylcholine synthesis through increasing the activity of the enzyme betaine–homocysteine methyltransferase [34]. Betaine–homocysteine methyltransferase catalyzes the transfer of a methyl group from betaine to homocysteine, but it also catalyzes the first step in the three-enzyme pathway that promotes methylation of phosphatidylethanolamine into phosphatidylcholine [35,36].
We speculate that betaine and phosphatidylcholine supplementation increase lipid concentrations in blood because of increased synthesis and availability of phosphatidylcholine, which promotes VLDL secretion from the liver into plasma, resulting in higher lipid concentrations in blood [32,37].
Our data indicate that folic acid supplementation does not affect serum lipid concentrations, which is in agreement with other studies [38–41].
The consequences of betaine or phosphatidylcholine supplementation for risk of CVD are difficult to predict. In healthy participants, betaine supplementation lowers plasma homocysteine after methionine loading by 20%–40%. Betaine as well as phosphatidylcholine supplementation lower fasting plasma homocysteine by 12%–20% (Table 6) [7,8]. This is estimated to lower CVD risk by 10%–15% [1,42]. However, it is not yet firmly established that homocysteine lowering indeed lowers risk of CVD, although evidence for a causal relationship is accumulating [1–3,43]. We now find that betaine not only decreases homocysteine, but also increases LDL cholesterol by approximately 10% and the total/HDL cholesterol ratio by approximately 6%. This is estimated to increase CVD risk by approximately 10% [44–46]. Betaine (6 g/d) and phosphatidylcholine (providing approximately 2.6 g/d of choline) supplementation increase triacylglycerol concentrations by up to 13%, which is estimated to increase risk of CVD by approximately 6%, after adjustment for other risk factors [47–49]. This implies that betaine supplementation is expected to increase risk of CVD if homocysteine is not a cause of CVD, or does not affect risk if homocysteine is causally involved. Whether phosphatidylcholine affects CVD risk remains uncertain so far.
Table 6 Percentage Change in Fasting Concentrations of Plasma Homocysteine after Participants Had Ingested Betaine and Folic Acid for 6 Wk
Sources: [7,8].
a The magnitude of the effect of 2.6 g/d of phosphatidylcholine supplementation on fasting plasma homocysteine is comparable to the effect of betaine on plasma homocysteine in humans [51].
Observational studies could help in assessing the overall effect of phosphatidylcholine and betaine consumption on CVD risk. Recently, data on betaine and choline contents in foods became available [10], so it should now be feasible to assess the relationship between betaine and phosphatidylcholine intake and risk of CVD. In summary, the adverse effects of betaine on blood lipids make it less suitable as a homocysteine-lowering agent in healthy humans. Phosphatidylcholine may be more suitable because it only slightly increases serum triacylglycerol concentrations and does not affect LDL cholesterol. Folic acid has no adverse side effects on blood lipids, and therefore folic acid supplementation should remain the preferred homocysteine-lowering treatment in healthy humans.
Supporting Information
Protocol S1 Protocol for Study 3
(95 KB DOC).
Click here for additional data file.
Protocol S2 Protocol for Study 4
(840 KB PDF).
Click here for additional data file.
Protocol S3 Amendment 1 of Protocol for Study 4
(585 KB PDF).
Click here for additional data file.
Protocol S4 Amendment 2 of Protocol for Study 4
(470 KB PDF).
Click here for additional data file.
Table S1 Consort Checklist for Study 1
(47 KB DOC).
Click here for additional data file.
Table S2 Consort Checklist for Study 2
(47 KB DOC).
Click here for additional data file.
Table S3 Consort Checklist for Study 3
(47 KB DOC).
Click here for additional data file.
Table S4 Consort Checklist for Study 4
(38 KB DOC).
Click here for additional data file.
Table S5 Overview of Non-Serious Adverse Events per Study
(25 KB DOC).
Click here for additional data file.
Trial Registration
Study 3 is registered at clinicaltrials.gov under identifier NCT00102843. Study 4 is registered at clinicaltrials.gov under identifier NCT00102232.
Patient Summary
Background
Homocysteine is an amino acid in the blood (amino acids are the building blocks of proteins). Too much homocysteine in the blood is related to a higher risk of stroke and heart disease. Mildly elevated levels (which are quite common) might promote atherosclerosis (furring up of the arteries), but this link has not been proven yet. Studies are currently under way to test whether reducing homocysteine levels in people who have mildly elevated levels could prevent heart disease or stroke. Again, the results are not yet known.
Why Was This Study Done?
In people who have dramatically elevated levels of homocysteine, different food supplements have been shown to bring down homocysteine concentrations and to reduce the risk of heart disease and strokes. These supplements include folic acid, vitamin B12, vitamin B6, betaine, and phosphatidylcholine. This study looked at the effects of betaine, folic acid, and phosphatidylcholine on blood lipids in healthy people. The researchers wanted to study blood lipids, especially “bad” cholesterol (LDL cholesterol), because they are known to influence the risk of heart disease.
What Did the Researchers Do?
They analyzed data from four different studies that looked at the ability of the three substances to lower homocysteine levels. They analyzed the blood samples (which had previously been used to measure homocysteine) for lipid levels.
What Did They Find?
They found that betaine increased the level of “bad” cholesterol. Folic acid did not affect lipid levels. The data for phosphatidylcholine were not conclusive.
What Does This Mean?
This suggests that the beneficial effects of betaine (which lowers homocysteine) might be undone at least in part by its negative effects on blood lipids. Based on these results, folic acid would be a better choice for people who want to lower their homocysteine levels, since folic acid doesn't cause a rise in “bad” cholesterol.
More Information Online
More information on the link between homocysteine and heart disease can be found at the following Web sites. Web site of the American Heart Association (search for “homocysteine” or “cholesterol”): http://www.americanheart.org/presenter.jhtml?identifier = 1200000
Web site of the Australian Heart Association (search for “homocysteine,” “cholesterol,” and “cardiovascular risk”): http://www.heartfoundation.com.au/
Web site of the British Heart Foundation (search for “homocysteine,” “cholesterol,” and “cardiovascular risk”): http://www.bhf.org.uk/
WebMD Web pages on homocysteine, cholesterol, and heart disease (or search http://webmd.com for “homocysteine” or “cholesterol”): http://my.webmd.com/content/pages/9/1675_57859; http://my.webmd.com/content/pages/9/1675_57815
We thank all volunteers for their participation, and our colleagues at TNO Quality of life and at the Division of Human Nutrition, Wageningen University, for their help. We thank J. Burema, MSc, for statistical advice, and Dr. H. J. Verkade, Academic Hospital Groningen, for advice on phosphatidylcholine metabolism.
Funding was provided by the Wageningen Centre for Food Sciences, an alliance of major Dutch food industries, Maastricht University, TNO Quality of Life, and Wageningen University and Research Centre, with financial support from the Dutch government. The funders approved the design of the studies, but had no role in data collection, analysis, and interpretation, in preparation of the manuscript, or the decision to publish
Citation: Olthof MR, van Vliet T, Verhoef P, Zock PL, Katan MB (2005) Effect of homocysteine-lowering nutrients on blood lipids: Results from four randomised, placebo-controlled studies in healthy humans. PLoS Med 2(5): e135.
Abbreviations
CVDcardiovascular disease
HDLhigh-density lipoprotein
LDLlow-density lipoprotein
VLDLvery low density lipoprotein
==== Refs
References
Homocysteine Studies Collaboration Homocysteine and risk of ischemic heart disease and stroke: A meta-analysis JAMA 2002 288 2015 2022 12387654
Schnyder G Roffi M Pin R Flammer Y Lange H Decreased rate of coronary restenosis after lowering of plasma homocysteine levels N Engl J Med 2001 345 1593 1600 11757505
Schnyder G Roffi M Flammer Y Pin R Hess OM Effect of homocysteine-lowering therapy with folic acid, vitamin B(12), and vitamin B(6) on clinical outcome after percutaneous coronary intervention: The Swiss Heart study: A randomized controlled trial JAMA 2002 288 973 979 12190367
Wilcken DE Wilcken B Dudman NP Tyrrell PA Homocystinuria—The effects of betaine in the treatment of patients not responsive to pyridoxine N Engl J Med 1983 309 448 453 6877313
Smolin LA Benevenga NJ Berlow S The use of betaine for the treatment of homocystinuria J Pediatr 1981 99 467 472 7264811
Bostom AG Shemin D Nadeau MR Shih V Stabler SP Short term betaine therapy fails to lower elevated fasting total plasma homocysteine concentrations in hemodialysis patients maintained on chronic folic acid supplementation Atherosclerosis 1995 113 129 132 7755649
Steenge GR Verhoef P Katan MB Betaine supplementation lowers plasma homocysteine in healthy men and women J Nutr 2003 133 1291 1295 12730412
Olthof MR van Vliet T Boelsma E Verhoef P Low dose betaine supplementation leads to immediate and long term lowering of plasma homocysteine in healthy men and women J Nutr 2003 133 4135 4138 14652361
Schwab U Torronen A Toppinen L Alfthan G Saarinen M Betaine supplementation decreases plasma homocysteine concentrations but does not affect body weight, body composition, or resting energy expenditure in human subjects Am J Clin Nutr 2002 76 961 967 12399266
Zeisel SH Mar MH Howe JC Holden JM Concentrations of choline-containing compounds and betaine in common foods J Nutr 2003 133 1302 1307 12730414
Hayes KC Pronczuk A Cook MW Robbins MC Betaine in sub-acute and sub-chronic rat studies Food Chem Toxicol 2003 41 1685 1700 14563394
Matthews JO Southern LL Higbie AD Persica MA Bidner TD Effects of betaine on growth, carcass characteristics, pork quality, and plasma metabolites of finishing pigs J Anim Sci 2001 79 722 728 11263833
McGregor DO Dellow WJ Robson RA Lever M George PM Betaine supplementation decreases post-methionine hyperhomocysteinemia in chronic renal failure Kidney Int 2002 61 1040 1046 11849459
Zeisel SH Dietary choline: Biochemistry, physiology, and pharmacology Annu Rev Nutr 1981 1 95 121 6764726
Zeisel SH Da Costa KA Franklin PD Alexander EA Lamont JT Choline, an essential nutrient for humans FASEB J 1991 5 2093 2098 2010061
Olson RE Vester JW Gursey D Davis N Longman D The effect of low-protein diets upon serum cholesterol in man Am J Clin Nutr 1958 6 310 324 13533316
Olson RE Jablonski JR Taylor E The effect of dietary protein, fat, and choline upon the serum lipids and lipoproteins of the rat Am J Clin Nutr 1958 6 111 118 13520640
Niculescu MD Zeisel SH Diet, methyl donors and DNA methylation: Interactions between dietary folate, methionine and choline J Nutr 2002 132 2333S-2335S
Friedewald WT Levy RI Fredrickson 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
Brenton DP Cusworth DC Dent CE Jones EE Homocystinuria. Clinical and dietary studies Q J Med 1966 35 325 346 5956442
Koc H Mar MH Ranasinghe A Swenberg JA Zeisel SH Quantitation of choline and its metabolites in tissues and foods by liquid chromatography/electrospray ionization-isotope dilution mass spectrometry Anal Chem 2002 74 4734 4740 12349977
Institute of Medicine Food and Nutrition Board Dietary references intakes for thiamin, riboflavin, niacin, vitamin B6 , folate, vitamin B12 , pantothenic acid, biotin, and choline. Washington (DC): National Academies Press. Available: http://www.nap.edu/books/0309065542/html/index.html
1998 Accessed 4 April 2005
Beheshti I Wessels LM Eckfeldt JH EDTA-plasma vs serum differences in cholesterol, high-density-lipoprotein cholesterol, and triglyceride as measured by several methods Clin Chem 1994 40 2088 2092 7955384
Curtin F Altman DG Elbourne D Meta-analysis combining parallel and cross-over clinical trials. I: Continuous outcomes Stat Med 2002 21 2131 2144 12210629
Wilcken DE Wilcken B The natural history of vascular disease in homocystinuria and the effects of treatment J Inherit Metab Dis 1997 20 295 300 9211201
Greten H Raetzer H Stiehl A Schettler G The effect of polyunsaturated phosphatidylcholine on plasma lipids and fecal sterol excretion Atherosclerosis 1980 36 81 88 7387779
Childs MT Bowlin JA Ogilvie JT Hazzard WR Albers JJ The contrasting effects of a dietary soya lecithin product and corn oil on lipoprotein lipids in normolipidemic and familial hypercholesterolemic subjects Atherosclerosis 1981 38 217 228 7193460
Prack M Sanborn T Waugh D Simkin H Bennett Clark S Perkins EG Visek WJ Effects of polyunsaturated lecithin on plasma and lipoprotein cholesterol and fatty acids in normal men Dietary fats and health 1983 Champaign (Illinois) American Oil Chemist's Society 689 697
29. Kesaniemi YA Grundy SM Effects of dietary polyenylphosphatidylcholine on metabolism of cholesterol and triglycerides in hypertriglyceridemic patients Am J Clin Nutr 1986 43 98 107 3942098
Oosthuizen W Vorster HH Vermaak WJ Smuts CM Jerling JC Lecithin has no effect on serum lipoprotein, plasma fibrinogen and macro molecular protein complex levels in hyperlipidaemic men in a double-blind controlled study Eur J Clin Nutr 1998 52 419 424 9683394
Dodson WL Sachan DS Choline supplementation reduces urinary carnitine excretion in humans Am J Clin Nutr 1996 63 904 910 8644685
Yao ZM Vance DE The active synthesis of phosphatidylcholine is required for very low density lipoprotein secretion from rat hepatocytes J Biol Chem 1988 263 2998 3004 3343237
Ridgway ND Vance DE Kinetic mechanism of phosphatidylethanolamine N-methyltransferase J Biol Chem 1988 263 16864 16871 3182819
Emmert JL Webel DM Biehl RR Griffiths MA Garrow LS Hepatic and renal betaine-homocysteine methyltransferase activity in pigs as affected by dietary intakes of sulfur amino acids, choline, and betaine J Anim Sci 1998 76 606 610 9498371
Shibata T Akamine T Nikki T Yamashita H Nobukuni K Synthesis of betaine-homocysteine S-methyltransferase is continuously enhanced in fatty livers of thyroidectomized chickens Poult Sci 2003 82 207 213 12619796
Sehayek E Wang R Ono JG Zinchuk VS Duncan EM Localization of the PE methylation pathway and SR-BI to the canalicular membrane: Evidence for apical PC biosynthesis that may promote biliary excretion of phospholipid and cholesterol J Lipid Res 2003 44 1605 1613 12810817
Sowden MP Collins HL Smith HC Garrow TA Sparks JD Apolipoprotein B mRNA and lipoprotein secretion are increased in McArdle RH-7777 cells by expression of betaine-homocysteine S-methyltransferase Biochem J 1999 341 639 645 10417327
Holven KB Haugstad TS Holm T Aukrust P Ose L Folic acid treatment reduces elevated plasma levels of asymmetric dimethylarginine in hyperhomocysteinaemic subjects Br J Nutr 2003 89 359 363 12628031
Liem A Reynierse-Buitenwerf GH Zwinderman AH Jukema JW van Veldhuisen DJ Secondary prevention with folic acid: Effects on clinical outcomes J Am Coll Cardiol 2003 41 2105 2113 12821232
Verhaar MC Wever RM Kastelein JJ van Loon D Milstien S Effects of oral folic acid supplementation on endothelial function in familial hypercholesterolemia. A randomized placebo-controlled trial Circulation 1999 100 335 338 10421591
Toole JF Malinow MR Chambless LE Spence JD Pettigrew LC Lowering homocysteine in patients with ischemic stroke to prevent recurrent stroke, myocardial infarction, and death: The Vitamin Intervention for Stroke Prevention (VISP) randomized controlled trial JAMA 2004 291 565 575 14762035
Wald DS Law M Morris JK Homocysteine and cardiovascular disease: Evidence on causality from a meta-analysis BMJ 2002 325 1202 1206 12446535
Klerk M Verhoef P Clarke R Blom HJ Kok FJ MTHFR 677C→T polymorphism and risk of coronary heart disease: A meta-analysis JAMA 2002 288 2023 2031 12387655
Natarajan S Glick H Criqui M Horowitz D Lipsitz SR Cholesterol measures to identify and treat individuals at risk for coronary heart disease Am J Prev Med 2003 25 50 57 12818310
Kinosian B Glick H Preiss L Puder KL Cholesterol and coronary heart disease: Predicting risks in men by changes in levels and ratios J Investig Med 1995 43 443 450
Stampfer MJ Sacks FM Salvini S Willett WC Hennekens CH A prospective study of cholesterol, apolipoproteins, and the risk of myocardial infarction N Engl J Med 1991 325 373 381 2062328
Hokanson JE Austin MA Plasma triglyceride level is a risk factor for cardiovascular disease independent of high-density lipoprotein cholesterol level: A meta-analysis of population-based prospective studies J Cardiovasc Risk 1996 3 213 219 8836866
Jeppesen J Hein HO Suadicani P Gyntelberg F Triglyceride concentration and ischemic heart disease: An eight-year follow-up in the Copenhagen Male Study Circulation 1998 97 1029 1036 9531248
Stampfer MJ Krauss RM Ma J Blanche PJ Holl LG A prospective study of triglyceride level, low-density lipoprotein particle diameter, and risk of myocardial infarction JAMA 1996 276 882 888 8782637
Metcalfe LD Schmitz AA Pelka JR Rapid preparation of fatty acid esters from lipids for gas chromatographic analysis Anal Chem 1966 38 514 515
Olthof MR Brink EJ Katan MB Verhoef P Choline supplemented as phosphatidylcholine decreases fasting and post-methionine plasma homocysteine concentrations in healthy men Am J Clin Nutr 2005 In press
| 15916468 | PMC1140947 | CC BY | 2021-01-05 10:39:46 | no | PLoS Med. 2005 May 31; 2(5):e135 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020135 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646910.1371/journal.pmed.0020137Correspondence and Other CommunicationsInfectious DiseasesEpidemiology/Public HealthHealth PolicyHIV/AIDSSexual HealthHIV Epidemiology in Africa: Weak Variables and Tendentiousness Generate Wobbly Conclusions CorrespondenceBrody Stuart
1
Potterat John J
2
1University of PaisleyPaisleyScotland2Colorado Springs, Colorado, United States of AmericaE-mail: [email protected]
Competing Interests: The authors have declared that no competing interests exist.
5 2005 31 5 2005 2 5 e137Copyright: © 2005 Brody and Potterat.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.
Individual Level Injection History: A Lack of Association with HIV Incidence in Rural Zimbabwe
Authors' Reply: Don't Let the Hypothesis Slip
==== Body
In their attempt to defend heterosexual transmission as the driving force for HIV epidemics in sub-Saharan Africa and dismiss the evidence for unsafe injections (or other percutaneous exposures) as a major source of HIV transmission [1,2], Lopman and colleagues [3] assert that “unsafe medical injections can be confidently excluded as a major source of HIV infection” in Manicaland, Zimbabwe. Their methods and results are glaringly insufficient to generate the confidence they seek. Confidence requires thorough evaluation, which was not done in their study [3].
Although the authors quantified the number of sexual partners, they failed to quantify the number of injections [3], which undermines the likelihood of detecting an effect of injections, as does adding the statistical noise of “needle pricks” (how was that question phrased?) to injections. We find it difficult to believe that the authors were unaware of either the dose-dependency issue or the implications of using a weak measure of a vector that they are transparently invested in dismissing. In addition, the recall interval was inordinately long (up to three years). It is unreasonable to expect accurate data from subjects who are not multiply prompted to stimulate recall.
As a secondary issue, their adjustment of the risk ratio associating HIV with injections against age is problematic. Adjusting a causal variable against age (which is a proxy for causal variables) may reduce the true association. Their Table 1 shows that injections in women vary by age in the same pattern that HIV incidence varies by age. It might be that when they adjust for age, the risk ratios for age capture the association, leaving a diminished association for injections.
Confidence in their report is also undermined by their approach to their own finding that sexual behavior was unrelated to risk of incident HIV. Indeed, using the same standard of evidence that the authors applied to their null medical-injection finding, the abstract, discussion, and press release information should also have proclaimed that sex was “confidently excluded” as a risk for HIV. Indeed, this would be consistent with the largest studies of HIV risk in Africa, including one in their own backyard (Manicaland) [4], which found little association of measured sexual behaviors with HIV risk. It would also be consistent with the many intervention studies reporting no benefit from condom-promotion programs [5], as well as with observation of the opposite trajectories of the HIV and STD epidemics observed in Zimbabwe [6].
Of additional concern is Lopman et al.'s finding that 13 of 67 individuals who seroconverted reported no sexual partners in the long inter-survey period. Indeed, their Table 2 data show that women with no reported sexual partners have higher HIV incidence than women who report any partner during the three-year interval (1.56 HIV cases per 100 person-years [12/770] for the former compared to 1.21 [36/2975] for the latter). They reveal their a priori conviction by forcing this datum into the procrustean sexual bed, inferring that it is explainable by unreported sexual activity rather than unreported or unmeasured percutaneous exposure. The authors blame underreporting of sexual behavior without using techniques shown to dramatically improve valid reporting [7].
Despite this remarkable lack of association in their women respondents between HIV incidence and number of sexual partners, they counterintuitively suggest (in the patient summary) that the important issue is sexual transmission. It seems to us that if sex doesn't appear to explain high HIV incidence, then one should recommend looking for what does, which the authors do not do. Rather than relying on the case-control approach, what is needed is intensive contact tracing (with viral sequencing of HIV specimens from index cases and their infected contacts to elucidate transmission relationships [8]) and a rigorous inventory of possible exposures and vectors [7]. The contact tracing and conscientious environmental risk probing of recent public-health responses to avian influenza and severe acute respiratory syndrome are an example of a useful and superior step to the case-control approach.
In brief, what is truly driving HIV transmission in sub-Saharan Africa will not be resolved by hastily implemented weak variables or by dismissive comments about Africans' self-reports of sexual behavior, especially comments unencumbered by data. Africans, science, and public health deserve better research [7].
Citation: Brody S, Potterat JJ (2005) HIV epidemiology in Africa: Weak variables and tendentiousness generate wobbly conclusions. PLoS Med 2(5): e137.
==== Refs
References
Brewer DD Brody S Drucker E Gisselquist D Potterat JJ Mounting anomalies in the epidemiology of HIV in Africa: Cry the beloved paradigm Int J STD AIDS 2003 14 144 147 12665436
Gisselquist D Potterat JJ Brody S Vachon F Let it be sexual: How health care transmission of AIDS in Africa was ignored Int J STD AIDS 2003 14 148 161 12665437
Lopman BA Garnett GP Mason PR Gregson S Individual level injection history: A lack of association with HIV incidence in rural Zimbabwe PloS Med 2005 2 e37 10.1371/journal.pmed.0020037 15736994
Potterat JJ Gisselquist D Brody S Still not understanding the uneven spread of HIV within Africa Sex Transm Dis 2004 31 365 15167647
Brody S Declining HIV rates in Uganda: Due to cleaner needles, not abstinence or condoms Int J STD AIDS 2004 15 440 441 15228727
Potterat JJ Brody S HIV epidemicity in context of STI declines: A telling discordance Sex Transm Infect 2002 78 467
Brody S Potterat JJ Establishing valid AIDS monitoring and research in countries with generalized epidemics Int J STD AIDS 2004 15 1 6 14769163
Brewer DD Rothenberg RB Potterat JJ Brody S Gisselquist D HIV epidemiology in Africa: Rich in conjecture, poor in data Int J STD AIDS 2004 15 63 65
| 15916469 | PMC1140948 | CC BY | 2021-01-05 10:39:46 | no | PLoS Med. 2005 May 31; 2(5):e137 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020137 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591645710.1371/journal.pmed.0020138EssayOtherHealth PolicyMedical EducationMedical journalsEditorial policies (including conflicts of interest)Health PolicyMedical EducationMedical Journals Are an Extension of the Marketing Arm of Pharmaceutical Companies EssaySmith Richard Richard Smith is Chief Executive of UnitedHealth Europe, London, United Kingdom. E-mail: [email protected]
Competing Interests: RS was an editor for the BMJ for 25 years. For the last 13 of those years, he was the editor and chief executive of the BMJ Publishing Group, responsible for the profits of not only the BMJ but of the whole group, which published some 25 other journals. He stepped down in July 2004. He is now a member of the board of the Public Library of Science, a position for which he is not paid.
5 2005 17 5 2005 2 5 e138Copyright: © 2005 Richard Smith.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.Medical journals have become dependent on the pharmaceutical industry for their survival, which can have a corrupting influence on their content, argues Smith, the former editor of the BMJ.
==== Body
“Journals have devolved into information laundering operations for the pharmaceutical industry”, wrote Richard Horton, editor of the Lancet, in March 2004 [1]. In the same year, Marcia Angell, former editor of the New England Journal of Medicine, lambasted the industry for becoming “primarily a marketing machine” and co-opting “every institution that might stand in its way” [2]. Medical journals were conspicuously absent from her list of co-opted institutions, but she and Horton are not the only editors who have become increasingly queasy about the power and influence of the industry. Jerry Kassirer, another former editor of the New England Journal of Medicine, argues that the industry has deflected the moral compasses of many physicians [3], and the editors of PLoS Medicine have declared that they will not become “part of the cycle of dependency…between journals and the pharmaceutical industry” [4]. Something is clearly up.
The Problem: Less to Do with Advertising, More to Do with Sponsored Trials
The most conspicuous example of medical journals' dependence on the pharmaceutical industry is the substantial income from advertising, but this is, I suggest, the least corrupting form of dependence. The advertisements may often be misleading [5,6] and the profits worth millions, but the advertisements are there for all to see and criticise. Doctors may not be as uninfluenced by the advertisements as they would like to believe, but in every sphere, the public is used to discounting the claims of advertisers.
The much bigger problem lies with the original studies, particularly the clinical trials, published by journals. Far from discounting these, readers see randomised controlled trials as one of the highest forms of evidence. A large trial published in a major journal has the journal's stamp of approval (unlike the advertising), will be distributed around the world, and may well receive global media coverage, particularly if promoted simultaneously by press releases from both the journal and the expensive public-relations firm hired by the pharmaceutical company that sponsored the trial. For a drug company, a favourable trial is worth thousands of pages of advertising, which is why a company will sometimes spend upwards of a million dollars on reprints of the trial for worldwide distribution. The doctors receiving the reprints may not read them, but they will be impressed by the name of the journal from which they come. The quality of the journal will bless the quality of the drug.
(Illustration: Margaret Shear, Public Library of Science)
Fortunately from the point of view of the companies funding these trials—but unfortunately for the credibility of the journals who publish them—these trials rarely produce results that are unfavourable to the companies' products [7,8]. Paula Rochon and others examined in 1994 all the trials funded by manufacturers of nonsteroidal anti-inflammatory drugs for arthritis that they could find [7]. They found 56 trials, and not one of the published trials presented results that were unfavourable to the company that sponsored the trial. Every trial showed the company's drug to be as good as or better than the comparison treatment.
By 2003 it was possible to do a systematic review of 30 studies comparing the outcomes of studies funded by the pharmaceutical industry with those of studies funded from other sources [8]. Some 16 of the studies looked at clinical trials or meta-analyses, and 13 had outcomes favourable to the sponsoring companies. Overall, studies funded by a company were four times more likely to have results favourable to the company than studies funded from other sources. In the case of the five studies that looked at economic evaluations, the results were favourable to the sponsoring company in every case.
The evidence is strong that companies are getting the results they want, and this is especially worrisome because between two-thirds and three-quarters of the trials published in the major journals—Annals of Internal Medicine, JAMA, Lancet, and New England Journal of Medicine—are funded by the industry [9]. For the BMJ, it's only one-third—partly, perhaps, because the journal has less influence than the others in North America, which is responsible for half of all the revenue of drug companies, and partly because the journal publishes more cluster-randomised trials (which are usually not drug trials) [9].
Why Do Pharmaceutical Companies Get the Results They Want?
Why are pharmaceutical companies getting the results they want? Why are the peer-review systems of journals not noticing what seem to be biased results? The systematic review of 2003 looked at the technical quality of the studies funded by the industry and found that it was as good—and often better—than that of studies funded by others [8]. This is not surprising as the companies have huge resources and are very familiar with conducting trials to the highest standards.
The companies seem to get the results they want not by fiddling the results, which would be far too crude and possibly detectable by peer review, but rather by asking the “right” questions—and there are many ways to do this [10]. Some of the methods for achieving favourable results are listed in the Sidebar, but there are many ways to hugely increase the chance of producing favourable results, and there are many hired guns who will think up new ways and stay one jump ahead of peer reviewers.
Then, various publishing strategies are available to ensure maximum exposure of positive results. Companies have resorted to trying to suppress negative studies [11,12], but this is a crude strategy—and one that should rarely be necessary if the company is asking the “right” questions. A much better strategy is to publish positive results more than once, often in supplements to journals, which are highly profitable to the publishers and shown to be of dubious quality [13,14]. Companies will usually conduct multicentre trials, and there is huge scope for publishing different results from different centres at different times in different journals. It's also possible to combine the results from different centres in multiple combinations.
These strategies have been exposed in the cases of risperidone [15] and odansetron [16], but it's a huge amount of work to discover how many trials are truly independent and how many are simply the same results being published more than once. And usually it's impossible to tell from the published studies: it's necessary to go back to the authors and get data on individual patients.
Peer Review Doesn't Solve the Problem
Journal editors are becoming increasingly aware of how they are being manipulated and are fighting back [17,18], but I must confess that it took me almost a quarter of a century editing for the BMJ to wake up to what was happening. Editors work by considering the studies submitted to them. They ask the authors to send them any related studies, but editors have no other mechanism to know what other unpublished studies exist. It's hard even to know about related studies that are published, and it may be impossible to tell that studies are describing results from some of the same patients. Editors may thus be peer reviewing one piece of a gigantic and clever marketing jigsaw—and the piece they have is likely to be of high technical quality. It will probably pass peer review, a process that research has anyway shown to be an ineffective lottery prone to bias and abuse [19].
Furthermore, the editors are likely to favour randomised trials. Many journals publish few such trials and would like to publish more: they are, as I've said, a superior form of evidence. The trials are also likely to be clinically interesting. Other reasons for publishing are less worthy. Publishers know that pharmaceutical companies will often purchase thousands of dollars' worth of reprints, and the profit margin on reprints is likely to be 70%. Editors, too, know that publishing such studies is highly profitable, and editors are increasingly responsible for the budgets of their journals and for producing a profit for the owners. Many owners—including academic societies—depend on profits from their journals. An editor may thus face a frighteningly stark conflict of interest: publish a trial that will bring US$100 000 of profit or meet the end-of-year budget by firing an editor.
Journals Should Critique Trials, Not Publish Them
How might we prevent journals from being an extension of the marketing arm of pharmaceutical companies in publishing trials that favour their products? Editors can review protocols, insist on trials being registered, demand that the role of sponsors be made transparent, and decline to publish trials unless researchers control the decision to publish [17,18]. I doubt, however, that these steps will make much difference. Something more fundamental is needed.
Firstly, we need more public funding of trials, particularly of large head-to-head trials of all the treatments available for treating a condition. Secondly, journals should perhaps stop publishing trials. Instead, the protocols and results should be made available on regulated Web sites. Only such a radical step, I think, will stop journals from being beholden to companies. Instead of publishing trials, journals could concentrate on critically describing them.
Examples of Methods for Pharmaceutical Companies to Get the Results They Want from Clinical Trials
Conduct a trial of your drug against a treatment known to be inferior.
Trial your drugs against too low a dose of a competitor drug.
Conduct a trial of your drug against too high a dose of a competitor drug (making your drug seem less toxic).
Conduct trials that are too small to show differences from competitor drugs.
Use multiple endpoints in the trial and select for publication those that give favourable results.
Do multicentre trials and select for publication results from centres that are favourable.
Conduct subgroup analyses and select for publication those that are favourable.
Present results that are most likely to impress—for example, reduction in relative rather than absolute risk.
This article is based on a talk that Richard Smith gave at the Medical Society of London in October 2004 when receiving the HealthWatch Award for 2004. The speech is reported in the January 2005 HealthWatch newsletter [20]. The article overlaps to a small extent with an article published in the BMJ [21].
Citation: Smith R (2005) Medical journals are an extension of the marketing arm of pharmaceutical companies. PLoS Med 2(5): e138.
==== Refs
References
Horton R The dawn of McScience New York Rev Books 2004 51 4 7 9
Angell M The truth about drug companies: How they deceive us and what to do about it 2005 New York Random House 336
Kassirer JP On the take: How medicine's complicity with big business can endanger your health 2004 New York Oxford University Press 251
Barbour V Butcher J Cohen B Yamey G Prescription for a healthy journal PLoS Med 2004 1 e22 10.1371/journal.pmed.0010022 17523248
Wilkes MS Doblin BH Shapiro MF Pharmaceutical advertisements in leading medical journals: Experts' assessments Ann Intern Med 1992 116 912 919 1580449
Villanueva P Peiro S Librero J Pereiro I Accuracy of pharmaceutical advertisements in medical journals Lancet 2003 361 27 32 12517463
Rochon PA Gurwitz JH Simms RW Fortin PR Felson DT A study of manufacturer-supported trials of nonsteroidal anti-inflammatory drugs in the treatment of arthritis Arch Intern Med 1994 154 157 163 8285810
Lexchin J Bero LA Djulbegovic B Clark O Pharmaceutical industry sponsorship and research outcome and quality BMJ 2003 326 1167 1170 12775614
Egger M Bartlett C Juni P Are randomised controlled trials in the BMJ different? BMJ 2001 323 1253
Sackett DL Oxman AD HARLOT plc: An amalgamation of the world's two oldest professions BMJ 2003 327 1442 1445 14684640
Thompson J Baird P Downie J The complete text of the independent inquiry commissioned by the Canadian Association of University Teachers The Olivieri report 2001 Toronto Lorimer 584
Rennie D Thyroid storm JAMA 1997 277 1238 1243 9103350
Rochon PA Gurwitz JH Cheung M Hayes JA Chalmers TC Evaluating the quality of articles published in journal supplements compared with the quality of those published in the parent journal JAMA 1994 272 108 113 8015117
Cho MK Bero LA The quality of drug studies published in symposium proceedings Ann Intern Med 1996 124 485 489 8602706
Huston P Moher D Redundancy, disaggregation, and the integrity of medical research Lancet 1996 347 1024 1026 8606568
Tramèr MR Reynolds DJM Moore RA McQuay HJ Impact of covert duplicate publication on meta-analysis: A case study BMJ 1997 315 635 640 9310564
Davidoff F DeAngelis CD Drazen JM Hoey J Hojgaard L Sponsorship, authorship, and accountability Lancet 2001 358 854 856 11567695
De Angelis C Drazen JM Frizelle FA Haug C Hoey J Clinical trial registration: A statement from the International Committee of Medical Journal Editors Lancet 2004 364 911 912 15364170
Godlee F Jefferson T Peer review in health sciences, 2nd ed 2003 London BMJ Publishing Group 367
Garrow J HealthWatch Award winner HealthWatch 2005 January 56 4 5
Smith R Medical journals and pharmaceutical companies: Uneasy bedfellows BMJ 2003 326 1202 1205 12775625
| 15916457 | PMC1140949 | CC BY | 2021-01-05 10:39:45 | no | PLoS Med. 2005 May 17; 2(5):e138 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020138 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591647010.1371/journal.pmed.0020139Correspondence and Other CommunicationsHIV/AIDSHIV Infection/AIDSMedicine in Developing CountriesAssociation between Injections and HIV Incidence CorrespondenceZafar Janjua Naveed Ahmad Khabir Altaf Arshad Imran Khan Mohammad Bin Hamza Hasan Aga Khan UniversityKarachiPakistanE-mail: [email protected]
Competing Interests: The authors have declared that no competing interests exist.
5 2005 31 5 2005 2 5 e139Copyright: © 2005 Janjua 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.
Individual Level Injection History: A Lack of Association with HIV Incidence in Rural Zimbabwe
Authors' Reply: Don't Let the Hypothesis Slip
==== Body
The article by Lopman and colleagues [1] has created more controversy than it's cured. Their conclusion that unsafe injections “do not play a major role in the transmission of HIV in rural Zimbabwe” cannot be based on the data they have presented. There are many major problems with their conclusions.
First, they did not assess at all whether injections received by the participants in the study were safe or unsafe. All they assessed was whether participants had received an injection or had been pricked by a needle during the past three years. The possibility that most of the injections in the setting were “safe” cannot be ruled out, and the data presented did not in any way address the issue of an association between “unsafe injection” practices and HIV infection.
Second, the recall period used by the investigators was too long. Cognizant of this problem, the World Health Organization has been proposing a three-month recall period for assessing frequency of injections. Although this recall period is 12 times shorter than the one used by the authors, we found in our studies in Pakistan that, even then, people had great difficulties recalling injections. It must be mentioned that injections are very frequent procedures in Pakistan [2].
Third, it seems that aside from injections, the investigators have included only data from sexual histories in their study; hence, other potential sources of exposure such as minor and major surgical procedures, dental instrumentation, and tattooing or other traditional practices involving scarification have been missed.
Fourth, it is not clear how “needle prick” was defined. Solid needles are also considered needles, but injury caused by these is less likely to transmit HIV than that caused by hollow bore needles.
Fifth, the authors have failed to quantify exposure. The risk of contracting HIV increases as the number of unsafe injections increases [3]. The incidence of disease among people who received one injection during follow-up compared with those who received 20 injections would clearly be different. This relationship has been clearly seen in the case of hepatitis C infection [4].
Sixth, the authors recommend that policymakers “should concentrate more on trying to prevent infection from unsafe sex” than on injections. But they have failed to assess whether the sex was unsafe or otherwise.
Further, we believe that even if the methodology is considered absolutely flawless, the current conclusion can only apply to a particular population and geographic area because the proportion of disease attributable to various exposures depends on the relative distribution of exposures in the population. For example, it is argued that in India (with the world's second largest HIV/AIDS population, more than 5 million) the HIV epidemic started as a result of high-risk sexual behaviors, but the number of injections per person is high and reuse of syringes in the health-care sector is widespread (N. K. Arora, personal communication). Therefore, unsafe medical injections have the potential to propagate this epidemic. Also, injections transmit many other pathogens like hepatitis B virus and hepatitis C virus. The infections that they cause have a very high morbidity and mortality. Hence, the need and urgency of intervention to decrease the overuse of injections and improve the safety of desired injections should not be questioned. Each country should make appropriate allocation of resources according to its own needs.
Citation: Janjua NZ, Ahmad K, Altaf A, Khan MI, Hamza HB (2005) Association between injections and HIV incidence. PLoS Med 2(5): e139.
==== Refs
References
Lopman BA Garnett GP Mason PR Gregson S Individual level injection history: A lack of association with HIV incidence in rural Zimbabwe PLoS Med 2005 2 e37 10.1371/journal.pmed.0020037 15736994
Janjua NZ Akhtar S Hutin Y Injection use in two districts of Pakistan: Implications for disease prevention Int J Qual Health Care 2005 In press
Hauri AM Armstrong GL Hutin YJ The global burden of disease attributable to contaminated injections given in health-care settings Int J STD AIDS 2004 15 7 16 14769164
Luby SP Qamruddin K Shah AA Omair A Pahsa O The relationship between therapeutic injections and high prevalence of hepatitis C infection in Hafizabad, Pakistan Epidemiol Infect 1997 119 349 356 9440439
| 15916470 | PMC1140950 | CC BY | 2021-01-05 10:39:45 | no | PLoS Med. 2005 May 31; 2(5):e139 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020139 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591647110.1371/journal.pmed.0020140Learning ForumDiabetes/Endocrinology/MetabolismEmergency MedicineSurgeryDiabetesSurgeryEmergency MedicineAn 18-Year-Old Patient with Type 1 Diabetes Undergoing Surgery Learning ForumPietropaolo Massimo Massimo Pietropaolo is at the Division of Immunogenetics, Diabetes Institute, Department of Pediatrics, Rangos Research Center, Children's Hospital of Pittsburgh, University of Pittsburgh School of Medicine, and at the Center for Diabetes and Endocrinology, Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, United States of America. E-mail: [email protected]
Competing Interests: The author declares that he has no competing interests.
5 2005 31 5 2005 2 5 e140Copyright: © 2005 Massimo Pietropaolo.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.A young patient with type 1 diabetes needs an elective operation under general anesthesia. How will you manage his diabetes before, during, and after the surgery?
==== Body
DESCRIPTION of CASE
An 18-year-old Caucasian male with type 1 diabetes presented to the emergency department complaining of severe left knee pain and swelling after sustaining a knee injury that occurred during a high school football match. Joint effusions were visible and palpable above the left knee, and there was significant loss of smooth motion of the knee, passively performed. Plain X rays showed no signs of fractures. The patient had had type 1 diabetes for six years, and his insulin regimen consisted of insulin glargine, 35 units at 8:00 p.m., and insulin lispro, 23 units at 8:00 a.m. and 16 units at 8:00 p.m. The patient had no apparent complications related to type 1 diabetes.
On examination he was alert, his pulse was 76 bpm regular, and his blood pressure was 118/66 mm Hg. Recently, the patient had had frequent episodes of both hyperglycemia and hypoglycemia. However, he had never developed diabetic ketoacidosis (DKA). His recent HbA1c was 9.5%, demonstrating inadequate glycemic control.
The patient was referred to an orthopedic surgeon, and arthroscopy was scheduled a few days later. A complex tear of the medial meniscus extending to the articular surfaces was diagnosed. Partial meniscectomy was recommended. (This procedure usually takes about one hour—nonetheless, the preoperative preparation for general anesthesia and the postoperative recovery may add several hours to this time.)
When Would You Have This Patient Report to the Hospital? The Day before Surgery or the Morning of Surgery?
This patient should be hospitalized no later than the evening before surgery, given his history of frequent episodes of hypo- and hyperglycemia and his poor glycemic control. This should allow for final optimization of glucose control before surgery. Ideally, frequent contact with the patient and adjustment of the insulin regimen prior to surgery should lead to an excellent glycemic control. This 18-year-old patient on glargine might benefit from additional doses of lispro given before lunch and dinner (rather than at 8 p.m., unless this is before dinner).
A patient with poorly controlled diabetes should not undergo elective surgery until glycemic levels are reasonably controlled. For example, bringing the patient into the hospital with blood glucose levels of 450 mg/dl will likely result in a canceled surgery. If a patient is scheduled for day of admission surgery, one way to avoid having to send the patient out without being able to do the procedure is to perform a finger stick blood sugar on arrival so one can cancel before admission.
Patients undergoing minor surgical procedures (e.g., arthroscopy) may be brought to the hospital on the morning of surgery. As the patient is usually NPO (i.e., has been given orders to fast before the surgery), a lower dose of the intermediate- or long-acting insulin is administered, and the regular insulin is withheld (Box 1). Insulin therapy should never be withheld in a patient with type 1 diabetes as this can result in DKA.
Box 1. Management of Type 1 Diabetes during Minor Surgery Procedures
Day of procedure if the patient is NPO:
Withhold morning dose of insulin.
Measure capillary glucose levels prior to procedure every 2–4 hours.
Administer short-acting or fasting insulin subcutaneously every 2–4 hours as indicated in Table 1.
Administer the usual afternoon dose of insulin.
Day of procedure if breakfast is allowed:
Administer morning dose of insulin.
Measure glycemia before and after procedure.
Give supplemental four units of short- or fast-acting insulin subcutaneously if glycemic levels are >250 mg/dl.
Give usual afternoon insulin dose.
Modified from [6].
How Should His Insulin Be Managed before Surgery?
The degree of metabolic control should be carefully evaluated before surgery; the goal is to improve the patient's blood glucose readings on an outpatient basis before undergoing surgery. If hyperglycemia has been present for a prolonged period of time before surgery, this could result in dehydration, which is commonly associated with electrolytic abnormalities such as sodium and potassium loss and possibly intravascular volume depletion. Prolonged hyperglycemia will delay healing and increase the risk of ischemia. A number of observations have indicated that hyperglycemia impairs collagen formation and causes a decrease in the tensile strength of surgical wounds [1,2]. Simply avoiding hyperglycemia can prevent these consequences. The reduction of glucose levels to below 200 mg/dl has been shown to improve granulocyte adherence and granulocytosis, both key components of the innate immunity and the defense against bacterial infections. Studies in both humans and animals suggest that high glucose levels might exacerbate ischemic brain damage [3].
Admission to the hospital is recommended for all patients with type 1 diabetes, and a stabilization period of 12–16 hours is also recommended for urgent procedures if severe hyperglycemia is present [4,5]. However, the widespread use of home blood glucose monitoring makes the improvement of glycemic control possible prior to admission. Traditionally, long-acting (e.g. ultralente) insulin is discontinued 2–3 days before surgery, and the patient is stabilized on a regimen of intermediate-acting (neutral protamine hagedorn [NPH] or lente) and short-acting (regular or humalog) insulin twice a day, or regular insulin before meals and intermediate-acting insulin at bedtime. However, if the glycemic control is good and the patient is being treated with glargine, it is acceptable to continue the regimen until the day of surgery [3]. Alternatively, 1/2–2/3 of the usual insulin regimen is given on the day of the procedure [3]. On the morning of surgery patients with type 1 diabetes should receive the insulin regimen that is used intraoperatively (Box 2) [6].
Box 2. IV Insulin-Glucose Infusion for Surgery
Preoperative Days
Attempt to obtain reasonable glycemic control. The goal is to achieve preprandial glycemic levels between 70 and 150 mg/dl (3.9–8.3 mmol/l).
Operative Day
Four to eight hours before surgery, keep patient NPO, discontinue SC insulin, and insert IV infusion line.
Measure capillary glucose levels at one-hour intervals.
Infuse D5W intravenously via an IV regulator pump.
If renal function and serum potassium concentration are normal, add 20 mEq KCl/l.
Based on hourly blood glucose determination, adjust each infusion as indicated in Table 2.
Modified from [6].
Evaluation of metabolic homeostasis, lipid profile, and kidney and myocardial function must be completed before surgery. The presence of diabetic autonomic neuropathy should also be assessed prior to surgery because this condition predisposes to perioperative hypotension. In such patients meticulous monitoring of blood pressure and volume status is essential during the perioperative period [7].
The common agreement is that 4–8 hours before surgery, the patient should be kept NPO, subcutaneous (SC) insulin should be discontinued, and an intravenous (IV) infusion line should be inserted. It should be emphasized that if the elective surgical procedure can be scheduled for early morning hours and the procedure lasts at least four hours, this limits the NPO to about 4–6 hours and allows administration of usual or 1/2–2/3 usual insulin and use of a glucose infusion to replace breakfast and supplemental insulin either subcutaneously or intravenously. If both glucose and insulin are infused, it should be pointed out that these must be two separately controlled infusions, so glucose and insulin infusion can be varied independently.
How Should His Insulin Be Managed during Surgery?
Major surgery and general anesthesia can cause severe metabolic abnormalities in patients with type 1 diabetes. Given the limitations of SC insulin therapy, such as unpredictable absorption and variable plasma insulin levels, constant infusion of insulin is recommended [3]. Anesthesia induces complex neuroendocrine stress responses and activates the sympathetic nervous system. The abnormal release of growth hormone, cortisol, and epinephrine leads to impaired insulin secretion, and causes insulin resistance and hyperglycemia due to increased glycogenolysis, gluconeogenesis, and decreased glucose disposal. Many other conditions may cause severe insulin resistance in a patient with diabetes (Box 3).
Box 3. Illnesses and Conditions Frequently Associated with Increased Insulin Requirement
Recent DKA
Poor glycemic control
Sepsis
Steroid therapy
Liver disease
Obesity
For patients scheduled for elective surgery, IV insulin and glucose infusion are usually started several hours preoperatively, and glucose levels should be maintained between 100 and 125 mg/dl. Slightly higher targets (100–150 mg/dl) have been recommended by some diabetologists to minimize the risk of hypoglycemia. The maintenance of glucose levels below 200 mg/dl has been shown to prevent bacterial infections and ischemic brain damage [3,8]. Suggested guidelines for management of patients with type 1 diabetes by use of insulin are outlined in Box 2. Since IV regular insulin has a short half-life (ten minutes), hypoglycemia is of little concern, as the infusion can be decreased and the IV glucose rate increased. The infusion rate can be adjusted by a floor nurse before surgery and by the anesthesiologist intraoperatively. An insulin infusion algorithm is constructed to allow easy titration of the insulin dose, and IV glucose must also be infused to obtain glucose levels within target (Box 2). This type of algorithm is effective in the majority of patients; however, it is based on the “average” patient and might require individualization. If the patient has a coexisting condition associated with increased insulin requirement, IV insulin doses need to be increased. Glycemic levels must be monitored at hourly intervals to keep glucose levels between 100 and 125 mg/dl.
Patients with type 1 diabetes who are treated with continuous SC insulin infusion by an insulin pump should be easily converted to IV regular insulin infusion just before surgery. Continuous SC insulin infusion is an acceptable regimen for surgical procedures requiring local anesthesia.
How Should Fluids and Electrolytes Be Managed in This Patient?
An adult without diabetes requires a minimum of 100 to 125 grams (400 to 500 calories) of glucose per day to prevent protein catabolism and the development of ketosis. Hence, this patient with types 1 diabetes should be treated with 5–10 grams of glucose per hour (1.2 to 2.4 mg/kg/min in a 70-kilogram subject) to provide sufficient basal energy requirement and prevent hypoglycemia during surgery. The dextrose concentration of the IV solution (Box 2) is adjusted based on the expected length of surgery. Thus, in this case 5% of dextrose in water (D5W) can be administered intravenously via infusion pump. For longer surgical procedures (intraabdominal or intrathoracic surgery), 10% of dextrose should be used to avoid excessive fluid administration. A 20% or 50% dextrose solution can be infused through a central venous catheter if fluid restriction is critical. If additional fluids are required, for instance, to replace unexpected intraoperative blood losses, non-glucose-containing solutions should be administered.
As a general rule, normal serum potassium levels do not necessarily imply that the total body potassium content is normal, as only 2% of total body potassium stores are extracellular [9]. In patients with diabetes, the metabolic homeostasis can rapidly be altered and many factors may influence serum potassium levels and total body potassium stores. These factors are (a) insulin, which increases potassium uptake by cells; (b) acidemia, which causes hyperkalemia as a result of the exchange of intracellular potassium for hydrogen ions; and (c) hyperosmolarity, which causes a rearrangement of potassium and fluid from intracellular to extracellular compartments. In patients with diabetes who have normal renal function and normal serum potassium concentration, 10 to 20 mEq of potassium should be added to each liter of dextrose-containing fluid. A higher dose is required in patients with hypokalemia. If serum potassium levels are greater than 5.5 mEq/l, potassium therapy should be withheld from the IV fluids and potassium serum levels should be monitored closely.
How Would the Management of This Patient Be Different for Emergency Rather Than Elective Surgery?
It has been estimated that as many as 5% of all patients with diabetes require surgery at some point during their lives [3]. Many of these patients undergo emergency surgery as a result of lower extremity infections requiring incision and drainage or even lower limb amputations. More than 50% of lower limb amputations in the United States occur among people with diabetes. The majority of patients with diabetes admitted for emergency procedures have a poor glyco-metabolic control and some of them may have coexisting DKA. The first step in management is to assess glycemic, electrolyte, acid-base, and volume status. An IV saline infusion should be started while waiting for laboratory tests to correct possible volume loss. Insulin infusion should be started at an appropriate rate (Box 2), and frequently insulin requirement might increase during emergency surgery. It is crucial that volume losses and electrolyte abnormalities are corrected prior to surgery.
If DKA is diagnosed, immediate treatment is indicated, and, if possible, surgery should be delayed until the glyco-metabolic control is corrected and stabilized. If emergency surgery procedures cannot be delayed, DKA can be treated concurrently with surgery.
How Should the Postoperative Glyco-Metabolic Management Be Handled in This Patient?
Both IV insulin and glucose (D5W, 0.45% normal saline) infusion should be continued until the glycol-metabolic control is stable and until 1–2 hours after the patient is able to resume oral feeding without difficulty. If postoperative nausea and vomiting are present, IV insulin and glucose infusion should not be discontinued. Furthermore, in patients with type 1 diabetes ketonuria could be an early sign of impending DKA, which could be triggered by starvation. Capillary glucose should be monitored every 1–2 hours at the bedside, and the variable insulin and glucose infusion should be adjusted to maintain blood glucose levels between 100 and 150 mg/dl (Box 4). Serum electrolytes should be measured immediately after surgery. Hypo- and hyperkalemia are fairly common in the postoperative period and should be corrected without delay. The presence of a widened anion gap suggests the possibility of DKA or lactic acidosis that might be caused by systemic infections or hypoperfusion.
Box 4. Type 1 Diabetes Management after Surgery
Continue IV insulin and glucose (D5W) infusion until two hours after oral feeding is resumed.
Measure capillary blood glucose before meals, at 10:00 p.m., and at 3:00 a.m.a
Maintain blood glucose levels between 100 and 150 mg/dl (5.5–8.3 mmol/l).
Provide three meals and three snacks (20–30 kcal/kg/day).
Administer preprandial short- or fast-acting insulin subcutaneously according to the schedule in Table 3.
For patients treated with continuous SC insulin infusion, resume basal rate and give bolus doses according to the carbohydrate counts.
aIf hypoglycemia is present at 3:00 a.m., reduce the 10:00 p.m. insulin dose.
Modified from [6].
If the patient is stable and can tolerate oral feeding, the regular home dose of insulin may be administered 20–30 minutes before the meal and the insulin infusion stopped 15–20 minutes after the meal. Of note, insulin infusion should be discontinued only after the SC insulin regimen is started to avoid any gaps in plasma insulin levels that may lead to a loss of metabolic control.
Resuming a sliding scale SC insulin treatment postoperatively has been advocated to improve metabolic control in patients with type 1 diabetes. However, there are several drawbacks with this approach. First of all, this approach is based on a retrospective treatment for hyperglycemia, which reflects the degree of insulin sensitivity and glucose disposal rate in the preoperative period. The use of a SC insulin treatment postoperatively tends to create fluctuations in blood glucose levels that could be difficult to control. Moreover, there is a risk of exposing the patient to the risk of hypoglycemia if excessive doses of insulin are administered. More importantly, the use of a sliding scale treatment might predispose to DKA in insulin-deficient patients before the development of hyperglycemic levels.
What Are the General Principles for Surgery in a Patient with Type 2 Diabetes?
Patients with type 2 diabetes require good blood glucose control prior to undergoing surgery. Although these patients seldom develop DKA, the same adverse effects of poor glycemic control may develop. For procedures that require general anesthesia, specific treatment, other than hourly glycemic monitoring, is not required for patients with type 2 diabetes whose diabetes is well controlled with diet (i.e., fasting blood glucose less than 125 mg/dl). However, insulin should still be administered as described in Box 2, as hyperglycemic responses may still occur intraoperatively in these patients. Targeted glycemic levels are identical to those for patients with type 1 diabetes. For elective procedures requiring general anesthesia, insulin infusion is usually required to control hyperglycemic levels that occur during surgery. Frequent intraoperative glucose monitoring is imperative in this setting to avoid complications resulting from a poor glycemic control. Patients with type 2 diabetes who require insulin and who are scheduled for surgery should be managed similarly to patients with type 1 diabetes.
As a general rule, SC insulin should not be used in patients with type 2 diabetes requiring surgery and general anesthesia, since insulin absorption from the SC tissue could be quite variable, particularly in obese individuals.
With regard to patients with type 2 diabetes who require general anesthesia and whose diabetes is well controlled with sulfonylureas, there is some degree of disagreement. Many diabetologists recommend withholding the sulfonylurea the morning of surgery for procedures requiring general or local anesthesia and using continuous insulin infusion.
Any patient treated with metformin should discontinue this medication at least 48 hours before surgery. The drug should be completely cleared after discontinuing metformin 48 hours preceding surgery. This is a prophylactic measure in an effort to lower the risk of lactic acidosis that could be secondary to complications of surgical procedures such as hypotension, myocardial infarction, or septic shock.
Treatment decisions for patients with type 2 diabetes receiving local anesthesia and undergoing minor surgical procedures are similar to those described above for patients with type 1 diabetes.
DISCUSSION
Patients with type 1 diabetes undergoing elective or emergency surgical procedures have a higher degree of morbidity and mortality than those without diabetes, as a result of impaired glyco-metabolic homeostasis and electrolyte balance. The magnitude of the catabolic responses is not only related to the severity of surgical and postsurgical complications but also to the effects of inadequate perioperative management on metabolic control. Health-care providers should be very familiar with the perioperative management of type 1 diabetes; with individualized insulin and glucose variable infusions, young patients affected by type 1 diabetes can undergo surgery with a minimal risk.
The goal of glycemic management in these situations is to maintain normal glucose homeostasis and normal metabolism. As insulin resistance and gluconeogenesis increase during surgery-related stress and anesthesia, additional insulin will be needed to prevent excessive hepatic glucose release. An important issue is maintaining a physiological fluid and electrolyte balance.
Key Learning Points
Patients with type 1 diabetes undergoing elective or emergency surgical procedures have a higher degree of morbidity and mortality than those without diabetes.
Perioperative hyperglycemia is associated with an increased risk of infections, delayed wound healing, and an increased risk of ischemia.
A patient with poorly controlled diabetes should not undergo elective surgery until glycemic levels are reasonably controlled.
Patients undergoing minor surgical procedures, such as arthroscopy, may be brought to the hospital on the morning of surgery.
Given the limitations of subcutaneous insulin therapy, such as unpredictable absorption and variable plasma insulin levels, constant infusion of insulin is recommended during surgery.
Perioperative hyperglycemia increases the risk of infections, delayed wound healing, and ischemia. Achieving preprandial glycemic levels between 70 and 150 mg/dl before surgery in the preoperative period and maintaining plasma glucose levels between 100 and 125 mg/dl during surgery and between 100 and 150 mg/dl after surgery using simple and safe algorithms (Boxes 1, 2, and 4) can significantly decrease the risk of these complications.
Table 1 Administering Insulin during Minor Surgery Procedures
Table 2 Infusing Insulin-Glucose during Surgery
a Administer 10 ml D5W intravenously and repeat glycemic levels 15 minutes later.
Table 3 Administering Insulin after Surgery
Citation: Pietropaolo M (2005) An 18-year-old patient with type 1 diabetes undergoing surgery. PLoS Med 2(5): e140.
Abbreviations
D5W5% percent of dextrose in water
DKAdiabetic ketoacidosis
IVintravenous
SCsubcutaneous
==== Refs
References
Marhoffer W Stein M Maeser E Federlin K Impairment of polymorphonuclear leukocyte function and metabolic control of diabetes Diabetes Care 1992 15 256 260 1547682
Rosenberg CS Wound healing in the patient with diabetes mellitus Nurs Clin North Am 1990 25 247 261 2179891
Clement S Braithwaite SS Magee MF Ahmann A Smith EP Management of diabetes and hyperglycemia in hospitals Diabetes Care 2004 27 553 591 14747243
van den Berghe G Wouters P Weekers F Verwaest C Bruyninckx F Intensive insulin therapy in the critically ill patients N Engl J Med 2001 345 1359 1367 11794168
Avilès-Santa L Raskin P Lebovitz HE Therapy for diabetes mellitus and its related disorders Surgery and anesthesia 2004 Alexandria (Virginia) American Diabetes Association/Port City Press 247 258
Rosenstock J Raskin P Surgery! Practical guidelines for diabetes management Clin Diabetes 1987 5 61
Burgos LG Ebert TJ Asiddao C Turner LA Pattison CZ Increased intraoperative cardiovascular morbidity in diabetics with autonomic neuropathy Anesthesiology 1989 70 591 597 2929996
Gill GV Alberti KGMM Alberti KGMM Zimmet PZ DeFronzo RA International textbook of diabetes mellitus The care of the diabetic patient during surgery 1997 Chichester (United Kingdom) Wiley 1243 1254
Levetan CS Magee MF Hospital management of diabetes Endocrinol Metab Clin North Am 2000 29 745 770 11149160
| 15916471 | PMC1140951 | CC BY | 2021-01-05 10:39:48 | no | PLoS Med. 2005 May 31; 2(5):e140 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020140 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591647210.1371/journal.pmed.0020141Research ArticleEpidemiology/Public HealthMental HealthPsychiatryEpidemiologyInternational healthPublic HealthSchizophrenia and other psychotic disordersA Systematic Review of the Prevalence of Schizophrenia The Prevalence of SchizophreniaSaha Sukanta
1
Chant David
1
2
Welham Joy
1
McGrath John
1
2
*1Queensland Centre for Mental Health Research, The Park Centre for Mental HealthWacolAustralia2Department of Psychiatry, University of QueenslandSt. LuciaAustraliaHyman Steven E. Academic EditorHarvard UniversityUnited States of America
Competing Interests: The authors have declared that no competing interests exist.
Author Contributions: JM designed the study. SS, DC, JW, and JM analyzed the data and contributed to writing the paper.
*To whom correspondence should be addressed. E-mail: [email protected] 2005 31 5 2005 2 5 e14115 2 2005 29 3 2005 Copyright: © 2005 Saha 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.
How Prevalent Is Schizophrenia?
Background
Understanding the prevalence of schizophrenia has important implications for both health service planning and risk factor epidemiology. The aims of this review are to systematically identify and collate studies describing the prevalence of schizophrenia, to summarize the findings of these studies, and to explore selected factors that may influence prevalence estimates.
Methods and Findings
Studies with original data related to the prevalence of schizophrenia (published 1965–2002) were identified via searching electronic databases, reviewing citations, and writing to authors. These studies were divided into “core” studies, “migrant” studies, and studies based on “other special groups.” Between- and within-study filters were applied in order to identify discrete prevalence estimates. Cumulative plots of prevalence estimates were made and the distributions described when the underlying estimates were sorted according to prevalence type (point, period, lifetime, and lifetime morbid risk). Based on combined prevalence estimates, the influence of selected key variables was examined (sex, urbanicity, migrant status, country economic index, and study quality).
A total of 1,721 prevalence estimates from 188 studies were identified. These estimates were drawn from 46 countries, and were based on an estimated 154,140 potentially overlapping prevalent cases. We identified 132 core studies, 15 migrant studies, and 41 studies based on other special groups. The median values per 1,000 persons (10%–90% quantiles) for the distributions for point, period, lifetime, and lifetime morbid risk were 4.6 (1.9–10.0), 3.3 (1.3–8.2), 4.0 (1.6–12.1), and 7.2 (3.1–27.1), respectively. Based on combined prevalence estimates, we found no significant difference (a) between males and females, or (b) between urban, rural, and mixed sites. The prevalence of schizophrenia in migrants was higher compared to native-born individuals: the migrant-to-native-born ratio median (10%–90% quantile) was 1.8 (0.9–6.4). When sites were grouped by economic status, prevalence estimates from “least developed” countries were significantly lower than those from both “emerging” and “developed” sites (p = 0.04). Studies that scored higher on a quality score had significantly higher prevalence estimates (p = 0.02).
Conclusions
There is a wealth of data about the prevalence of schizophrenia. These gradients, and the variability found in prevalence estimate distributions, can provide direction for future hypothesis-driven research.
Analysis of 188 studies published between 1965 to 2002 yields no difference between men and women and suggests that current text-book estimates for lifetime morbid risk are too high.
==== Body
Introduction
Schizophrenia is a disabling group of brain disorders characterized by symptoms such as hallucinations, delusions, disorganized communication, poor planning, reduced motivation, and blunted affect. While the incidence of the disorder is relatively low (median value 15.2 per 100,000 persons per year) [1], the condition is one of the major contributors to the global burden of disease [2]. The substantial burden of disease is a reflection of two features of schizophrenia: (a) the disorder usually has its onset in early adulthood, and (b) despite optimal treatment, approximately two-thirds of affected individuals have persisting or fluctuating symptoms [3].
Understanding the “epidemiological landscape” of schizophrenia requires many different types of descriptive studies [4]. Studies that estimate the incidence of schizophrenia are required in order to identify gradients across time and/or place. These gradients allow us to generate candidate risk factors that may underlie variations in the disorder. However, studies that report the prevalence of a disorder are also important. Estimating the proportion of a population affected with schizophrenia is central to health service planning. With respect to estimating the burden of disorder, prevalence proportions can provide insights into how incidence rates are refracted via different trajectories (e.g., recovery, chronicity, or early death). The statement “prevalence = incidence + course of illness” oversimplifies the dynamic matrix of factors influencing each component of the equation. Nevertheless, prevalence proportions can help us chart contours on the still-incomplete epidemiological map of schizophrenia.
Several scholarly narrative reviews of the prevalence of schizophrenia have been published in recent decades [4–8]. The sheer volume of data available on the prevalence of schizophrenia now requires a more systematic and orderly approach. As with many fields of medical knowledge, there is a growing appreciation that reviews should be based on data that are as complete and as free of bias as possible [9]. Systematic reviews have prespecified methods for locating studies and for extracting and synthesizing the data. Not all systematic reviews are accompanied by meta-analysis (i.e., pooling the data to provide one summary value) [10]. Even without pooling of data, the orderly sorting of data using meta-analytic techniques can provide useful insights into the structure of the relevant literature [11].
One systematic review of the prevalence of schizophrenia has been published to date [12]. This review (based solely on census and/or community survey data) identified 18 studies that provided estimates of either period and/or lifetime prevalence of schizophrenia. Goldner and colleagues reported pooled estimates for 1-y and lifetime prevalence of 3.4 and 5.5 per 1,000 persons, respectively. The authors commented on the heterogeneity of the data and suggested that this reflected “real variation” in the distribution of schizophrenia around the world.
Recently, we published a systematic review of the incidence of schizophrenia [1]. In brief, we found that the incidence of schizophrenia varied widely between sites (persons, median = 15.2 per 100,000; 10%–90% quantiles = 7.7–43.0). In addition, the study identified that (a) males were more likely to develop schizophrenia than females (median male:female risk ratio = 1.4); (b) migrants were more likely to develop schizophrenia than native-born individuals (median risk ratio = 4.6); and (c) individuals in urban sites had a higher risk of developing schizophrenia than those in mixed urban/rural sites. Regardless of the factors that underpin these incidence gradients, would these same gradients also be found in the prevalence of schizophrenia? If so, then it might suggest, for example, that factors influencing the course of the illness were more evenly distributed across these groups than factors influencing the incidence of the disorder. If the prevalence gradients are not congruent with the incidence gradients, then we are faced with the challenging task of unraveling the factors that could influence the differential course of schizophrenia between risk groups.
In this paper we continue our cartography of the epidemiological landscape of schizophrenia by presenting a systematic review of the prevalence of this disorder.
Ways to Measure the Prevalence of Schizophrenia
Prevalence measures the proportion of individuals who manifest a disorder at a specified time, or during a specified period. Generally prevalence estimates are calculated as a proportion, by dividing the total number of individuals who manifest a disorder (the numerator) by the total population at risk, including those with the disorder (the denominator). Prevalence proportions vary according to temporal criteria (e.g., point, period, or lifetime), but are not reported as an index of events over time (i.e., they are not like incidence rates that report the number of new cases per background population per year). Prevalence proportions are often loosely referred to as “rates”; however, in this review we will refer to them as “prevalence estimates” or “estimates.” Tables S1 and S2 define the types of prevalence estimates used in this study, and provide descriptions of the variables that we have used to describe the studies.
Point prevalence is the proportion of individuals who manifest a disorder at a given point in time (e.g., 1 d or 1 wk), while period prevalence measures the proportion of individuals who manifest a disorder during a specified period of time (e.g., 1 y). Given that the course of schizophrenia extends over months to decades, estimates of point prevalence based on 1 d are comparable to those based on 1 mo [5]. Thus, in this review we have combined all estimates based on temporal criteria of 1 mo or less in “point prevalence,” while studies that reported prevalence estimates between 1 mo and 12 mo are included under the heading “period prevalence.”
“Lifetime prevalence” is the proportion of individuals in the population who have ever manifested a disorder, who are alive on a given day. It is important to emphasize that lifetime prevalence needs to be clearly distinguished from “lifetime morbid risk” (LMR; also described elsewhere as morbid risk or expectancy). LMR differs from lifetime prevalence in that it attempts to include the entire lifetime of a birth cohort both past and future, and includes those deceased at the time of the survey [13]. LMR is the probability of a person developing the disorder during a specified period of their life or up to a specified age. There are various ways to calculate LMR [14,15]. The reviews of Odegaard [15], and Larsson and Sjogren [16] noted that, for low-incidence disorders such as schizophrenia, summation of age-specific incidence rates gives almost the same result as other more complicated methods of calculation [17]. The World Health Organization ten-country study [18] used this so-called “summation method” for the approximation of LMR. If one were to apply Linnean principles in order to design a taxonomy of frequency measures of disease, prevalence measures such as point, period, and lifetime would be closely related species within the same genus. However, there is a case to allocate LMR to the Genus “Incidence” rather than the Genus “Prevalence.” Conceptually (but not mathematically), LMR is closely related to cumulative incidence proportions derived from birth cohort studies [19].
Traditional prevalence studies (henceforth referred to as “core” studies) generate an estimate based on the population residing within a defined catchment area. However, it should be noted that the boundaries chosen for epidemiological studies (e.g., health districts, cities, states, or nations) may not be optimal for the detection of variations of the disorder within or between various populations. Lumping populations into large but convenient administrative areas can obscure informative, fine-grained gradients. With respect to prevalence estimates, factors such as the age structure of the population, mortality rates, and migration patterns can influence the estimates, and these may vary within and between sites.
Apart from catchment-area-based studies of the general population, there are many studies that report prevalence estimates for subgroups of the population. These may include groups defined by narrow age strata (e.g., the elderly or children), migrant status, ethnic or religious status, or twin status, to name but a few. A recent paper has systematically reviewed the prevalence of schizophrenia in prison settings [20]; however, this will not be included in this review. Migrant studies will be collated separately for analysis, while the remaining subgroup prevalence estimates will be included in “other special groups.”
Some studies report inpatient census data over a period of time (e.g., 1 y) and use the count of unique individuals with schizophrenia to generate a proportion based on general population figures. While these studies may be useful for administrative purposes, it is important not to mistake these estimates as “true” prevalence proportions. Very few patients require prolonged and continuous inpatient care; therefore, prevalence proportions based on inpatient data alone grossly underestimate true prevalence proportions. This review will collate these studies separately (henceforth referred to as “inpatient-census-derived” data); however, they will not be included in any of the main analyses.
Key Research Questions about the Prevalence of Schizophrenia
First there is a need to examine the degree of variation in the prevalence estimates of schizophrenia between sites. The companion review on the incidence of schizophrenia [1] found that within the central 80% of incidence rates, the difference ranged from 7.7 to 43 per 100,000 (over a 5-fold difference). While there has been debate within the schizophrenia research community about whether this range of rates is “narrow” or “prominent” (see review [21]), variations in prevalence estimates have not been a focus of controversy. The World Health Organization ten-country study commented that the prognosis of schizophrenia [18] was better in developing than in developed nations, a finding that has been “clear and consistent” in general [22]. The present review will describe the distribution of the different types of prevalence rates, and specifically examine whether the “developed versus developing” status of the sites influences the distribution of estimates.
Are the gradients that were identified in the incidence of schizophrenia also reflected in the prevalence of the disorder? For example, based on the previous finding that males have a significantly higher incidence of schizophrenia [1,23], it would be predicted that this sex difference might also be reflected in prevalence estimates. In addition, a recent study from China [24,25] highlighted an apparently unusual higher prevalence of schizophrenia in females in this country. In light of this issue, the male:female prevalence ratio will also be compared when the sites are sorted by a measure of “developed versus developing” status. Similarly, the incidence review identified significantly higher rates for (a) urban place of residence when compared to mixed urban/rural sites, and (b) migrant groups when compared to native-born individuals. These gradients will also be explored regarding the prevalence of schizophrenia.
Finally, systematic reviews can explore possible sources of heterogeneity in data by sorting the data according to methodological features. We will compare the distributions of estimates based on the quality of the study (as assessed by design features and thoroughness of reporting).
Methods
Identification of Studies
This systematic review conforms to the guidelines outlined by the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) recommendations [26]. The search methodology for this review was identical to that of our previous review paper on incidence of schizophrenia [1]. As a first step, a broad (free text) search string ([schizo* OR psych*] AND [incidence OR prevalence]) was used in MEDLINE, PsychINFO, EMBASE, and LILACS. Potentially relevant papers (in all languages) were accessed in order to review the full text. The references cited by each potentially relevant paper, review, and book chapter were scrutinized in order to locate additional potential papers. Posters were presented at two international schizophrenia conferences [27,28] in order to encourage researchers to contribute studies, especially studies from the “grey literature” (e.g., conference reports, theses, government reports, and unpublished studies). Subsequently, letters or E-mails were sent to the senior authors of papers that met the inclusion criteria. These authors were provided with an interim list of included papers and asked to nominate missing studies.
Included Studies
We included studies that reported primary data on the prevalence of schizophrenia first published between January 1965 and December 2002. Where multiple publications presented identical data, the most “informative version” of the study was included. Studies published in a language other than English were translated, and relevant papers were included.
Excluded Studies
Studies that reported prevalence data on prison or forensic populations were excluded (see recent systematic review of these studies [20]). We did not include genetic epidemiological studies that reported prevalence estimates in family members of affected (index) probands. Some studies report the LMR within large, multiplex families. These were not included; however, if the prevalence estimates were based on the entire population within a catchment area (e.g., an isolated population living in a village), then they were included.
Potential studies that had not been located at the time of submission were allocated to the “awaiting assessment” category. Studies based on inpatient-census-derived proportions are presented in the tables and summarized for comparative purposes, but were excluded from the main analyses.
Data Extraction
Once a study was included, data were extracted and entered into a three-level normalized database (i.e., only the unique prevalence estimate identifier was allowed to occur in more than one level) that included study-level variables (e.g., authors, year of publication, and site), middle-level variables (e.g., urban/rural status, age group, recruitment duration, case finding method, and diagnostic criteria), and rate-level variables (e.g., sex-specific rates for persons, males, and females). Two or more of the authors checked all data used in the analysis. When disagreements arose, these were resolved by consensus. If required, we contacted the original authors for clarification of issues. The full electronic dataset is available as Dataset 1.
Consistent with our previous systematic review of the incidence of schizophrenia [1], studies were given “quality points” based on operationalized features related to (a) optimal research design (e.g., higher scores for greater coverage, face-to-face interview versus chart diagnosis, and reliability of instruments), and (b) quality of reporting (e.g., provision of numerator and denominator, and description of diagnostic criteria). Details of the quality scores used in this review are provided in Table S3.
Sorting Prevalence Estimates by the Application of Sequential Filters
In systematic reviews, it is important that individuals are not “double counted” by the same or different studies. Thus, a key feature of this study is the application of sequential filters in order to identify discrete prevalence estimates. We applied a similar sorting algorithm as in our previous review of incidence of schizophrenia [1]. Briefly, the first filter parsed prevalence estimates from the included studies into three groups: core, migrant, and other special groups. Next, as the second filter, the estimates were sorted into six main types: point (1 mo or less), period (between 1 and 12 mo), lifetime, LMR, not otherwise specified (NOS), and inpatient-census-derived data.
A third, study-level filter was applied in order to isolate discrete data from multiple studies that overlapped in both time and place. This third filter was used to select one representative prevalence estimate for inclusion in the cumulative distribution using the “most informative” rule. For example, if one study presented multiple overlapping estimates, the estimate based on the largest sample was preferred (e.g., the widest age range was preferred over narrower age strata). Furthermore, filter rules were defined in order to select discrete estimates such that they allowed the greatest number of estimates to be included.
Presentation and Analyses of the Data
Key details of the included studies are presented in tables sorted by country, year of publication and first author (Tables S4, S5, and S6). The distributions of prevalence estimates are presented in cumulative plots, with every estimate contributing to the distribution. The distribution of the data is shown in rank order for prevalence estimate (lowest to highest ranks) with the cumulative percent of estimates shown on the vertical axis. The plots show horizontal reference lines indicating the 50% (median), and 25% and 75% quantiles (between which lies the interquartile range). In order to aid visual interpretation, some plots have been truncated, excluding very high estimates. Key features of these distributions are presented in tables (e.g., median, mean, harmonic mean, standard deviation, and quantiles at 10%, 25%, 50%, 75%, and 90%). These summary characteristics are based on the entire distributions. Results are presented as prevalence estimates per 1,000. In plots of prevalence ratios (e.g., male:female ratio), a vertical reference at the line of unity is shown.
We wish to draw attention to several features of the graphs used in this review. First, the central, near-linear segment of the cumulative distributions may extend beyond the interquartile range (e.g., from the 10%–90% quantiles), thus shape features (where the tails start or the range of the linear central segment) can be more informative than traditional interquartile ranges. Second, steeper segments of the cumulative plots are underpinned by estimates that have a narrow distribution, while flatter (i.e., more horizontal) segments of the distribution are underpinned by data that are relatively more dispersed. Finally, some distributions are derived from more data than others. Regardless of slope (i.e., steep or flat), if many estimates underpin segments of the distributions, then inferences based on these segments are probably more reliable than those based on segments underpinned by less data.
Meta-analyses often display data points with confidence intervals, and formal tests of heterogeneity are usually applied before combining data. For several reasons, the data in this review do not lend themselves to this type of analysis. Among the discrete core studies (see below), no study provided confidence limits to accompany the prevalence estimate. One study, which was allocated to “other special groups,” did provide confidence limits [29]. Where studies provided the corresponding numerator and denominator for a prevalence estimate, we were able to derive standard errors. However, we were able to impute standard errors for only 26% of the prevalence estimates, which were drawn from less than half (45%) of the discrete core studies. Faced with such a restricted pool of standard errors, the ability to assess the heterogeneity of the estimates in a manner generalizable across all core studies is compromised. In addition, the issues that underlie the decision to combine data from randomized controlled trials or risk factor epidemiological studies are of less relevance to prevalence estimates, where estimates based on very large populations should not necessarily carry more weight than estimates based on small populations. Based on first principles, there is no reason to assume that prevalence estimates for a disease remain static across time or place. Thus, forcing individual prevalence estimates into one pooled estimate loses important information. In this review we wish to draw attention to several characteristics of the distribution of estimates (e.g., central tendency, shape and width of the distribution, and density of data), rather than provide one pooled estimate.
In keeping with our systematic review of the incidence of schizophrenia [1], we supplement the graphical presentation of the prevalence estimates with statistical analyses. These analyses take into account (a) the need to control for within-study variation (estimates drawn from the same study tend to be more alike than estimates drawn from different studies), and (b) the use of a log transformation of the data in order to analyze distributions that are often positively skewed. Note that the median value is more informative than the arithmetic mean to assess central tendency in a skewed distribution, as is the harmonic mean (which is calculated as the exponential of the arithmetic mean of the log-transformed data, also known as the geometric mean). The analyses were carried out in SAS 9.1 using proc univariate (for medians and other quantiles of the raw data) and proc mixed for comparisons of harmonic means (because one study may provide more than one estimate, it is important to control for within-study variation).
Faced with a large quantity of data, systematic reviewers need to keep a tight rein on the number of comparisons undertaken on the data [30]. While it is tempting to reanalyze data in the light of findings that emerge from the data, such reanalyses should be kept to a minimum. The analysis of prevalence estimates is particularly challenging because of the many different prevalence types (e.g., point, period, lifetime, and LMR). Thus, in order to minimize the number of statistical comparisons in the current review, we restricted the analyses to a limited set of planned sensitivity analyses, each with a priori directional hypotheses, and, for post hoc analyses, applied multiple comparison corrections to the nominal significance levels by a Bonferroni correction. Furthermore, these analyses were based on hybrid distributions, which merged four different prevalence estimate types (point, period, lifetime, and NOS; henceforth referred to as “combined prevalence estimates”). Apart from the specific analyses related to sex differences, we undertook these analyses on distributions for persons only (i.e., males and females combined).
Hypotheses
Based on first principles, we predicted that the estimates for known prevalence types that include different temporal criteria would be significantly different. More specifically, we predicted the following: (a) prevalence estimates for persons would differ between lifetime, period and point (point being the lowest), and (b) LMR estimates would be higher than lifetime estimates.
There is now strong evidence that males have an increased risk of developing schizophrenia [1,23]. We compared the distribution for males versus females on the combined prevalence estimates, predicting that males would have distributions derived from higher estimates (i.e., distributions for males would be right-shifted compared to distributions for females).
In order to explore the influence of urbanicity of site on the prevalence of schizophrenia, we divided the combined prevalence estimates for persons into three categories (urban, rural, and mixed urban/rural). Allocation was based on the study descriptions of the area or, in the absence of these descriptors, the review authors' best estimate of this variable. There are several reasons to predict that the prevalence of schizophrenia would be higher in urban regions than in rural regions. First, the incidence of schizophrenia is higher in urban sites than mixed urban/rural sites [1]. Second, the “social drift” hypothesis suggests that the individuals with schizophrenia are more likely to move into urban regions in response to various factors related to poverty, the availability of services, and easier access to cheap accommodation [31]. Finally, some commentators suggest that less industrialized settings (e.g., rural regions and/or developing countries) may facilitate recovery via social connectedness and easier access to work [32]. Thus, we predicted that the prevalence of schizophrenia would be higher in urban sites than in rural or mixed urban/rural.
Migrants have a significantly increased risk of developing schizophrenia [1,33]. Assuming that the course of the illness does not vary according to migrant status, based on combined prevalence estimates for persons, we predicted that the prevalence of schizophrenia would be higher in migrants than in native-born individuals.
While there is a lack of evidence addressing whether the incidence of schizophrenia varies with the economic status of nations, there is solid evidence showing that people with schizophrenia from developing countries tend to have better outcomes than individuals in developed nations [18,22]. Mindful that there is a lack of consensus on how best to define the multidimensional concept of economic development, we have sorted prevalence estimates according to the per capita gross national product of the study site (2004 data) [34], and used standard World Bank definitions [35]: (a) least developed countries, = mean income of less than US$2,995; (b) emerging economy countries, = mean income between US$2,995 and $9,266; and (c) developed countries, = mean income of greater than US$9,266. Thus, based on combined prevalence estimates for persons, we predicted that the prevalence of schizophrenia would be significantly different across the three economic categories, and that the prevalence of schizophrenia would be significantly lower in least developed countries than in developed countries. Furthermore, a recent commentary drew attention to the apparent female excess in the prevalence of schizophrenia in developing nations, in contrast to the male excess thought to characterize the developed world [25]. Thus, based on combined prevalence estimates, we compared the male:female ratio when the prevalence estimates were classified by the three economic categories. We predicted that the ratio would be significantly different between the three economic levels, and specifically, that the male:female ratio in developed nations would be significantly higher than that of least developed countries.
Finally, methodological features can influence prevalence estimates. For example, studies that use comprehensive case ascertainment methods (e.g., “door-knock” surveys, inpatient and outpatient records, general practitioner surveys, and/or surveys based on other community sources), should identify more cases than those that rely on fewer recruitment sources. Based on the combined prevalence estimates for persons, we divided the estimates into quality score terciles. We predicted that the prevalence estimates would be significantly different when assessed by quality score. More specifically, we predicted that prevalence estimates from studies with the highest quality score tercile would be higher than those from the lowest tercile.
Results
The “Epidemiology” of Prevalence Estimates
The results of the search strategy, including source of the studies, subsequent culling, and final distribution of the papers, are shown in Figure 1. The electronic search identified 1,112 papers (85% of the total papers included in the study), while manual reference checking identified an additional 142 references (11%). We received responses from 31 authors (see Acknowledgments for full list), who provided an additional 53 references (4%). We identified 98 studies that were published in languages other than English. After translation 17 of these studies were included in this review.
Figure 1 Flow Diagram (Selection Strategy) of Included Studies
Double asterisk indicates exclusion categories (number studies excluded in parentheses). Double asterisk indicates numbers that are not mutually exclusive. A few studies provided rates for more than one group (11 studies provided data for both core and migrant [n = 3] or both core and other special groups [n = 8]; details in Results). LOTE, language other than English.
The list of references arranged by various criteria can be found in Tables 1–4. The systematic review identified 188 studies that provided prevalence estimates [18,29,36–223]. These studies provided 1,721 estimates and were drawn from 46 countries. There were 132 core studies, 15 migrant studies (of which three overlap with discrete core), and 41 studies that reported the prevalence of schizophrenia in other special groups.
Table 1 References by Type of Study
Table 4 References by Urbanicity of Sites
Table 2 References by Type of Prevalence Estimate
Table 3 References by Country
Key features of these core, migrant, and special groups are provided in Tables S4–S6. We excluded 26 studies that were completely overlapping by time and place, and 19 studies that reported prevalence data on prison populations (see Figure 1). However, ten partially overlapping studies were included that provided at least one discrete rate for this review [37,40,45,60,64,100,148,153,187,220].
The prevalence estimates were based on an estimated total of 154,140 potentially overlapping cases. The 132 core studies provided from one to 13 prevalence estimates per study. Four studies [59,120,169,224] reported prevalence only within narrow age strata without providing an overall rate. These studies were not included in the discrete core analyses.
Of the 132 core studies, we identified 21 studies for point prevalence, 34 studies for period prevalence, and 24 studies for lifetime prevalence. Thirty-two studies provided no information on the type of prevalence they reported—these were allocated to NOS prevalence. There were nine studies that reported LMR. Finally there were 44 studies that reported inpatient-census-derived data.
The Distribution of Prevalence Estimates
Figures 2–7 and Tables 5–7 show the distribution of the different types of prevalence estimates, and quantiles and moments for persons, males, and females.
Figure 2 Cumulative Plots of the Point Prevalence Estimates per 1,000 by Sex
Figure 7 Cumulative Plots of the Inpatient-Census-Derived Prevalence Estimates per 1,000 by Sex
Table 5. Quantiles and Moments of Point, Period, Lifetime, and NOS Prevalence per 1,000 by Sex
Table 7 Quantiles and Moments of Studies with Inpatient-Census-Derived Data per 1,000 by Sex
Data from 44 studies.
Figure 3 Cumulative Plots of the Period Prevalence Estimates per 1,000 by Sex
Figure 4 Cumulative Plots of the Lifetime Prevalence Estimates per 1,000 by Sex
Figure 5 Cumulative Plots of the LMR Estimates per 1,000 by Sex
Figure 6 Cumulative Plots of the NOS Prevalence Estimates per 1,000 by Sex
The median point prevalence for persons (based on 23 estimates) was 4.6 per 1,000, and the 10% and 90% quantiles ranged from 1.9 to 10.0 per 1,000 (a 5-fold difference). The median period prevalence for persons (based on 42 estimates) was 3.3 per 1,000, and the 10% and 90% quantiles ranged from 1.3 to 8.2 per 1,000 (a 6.5-fold difference). The median lifetime prevalence for persons (based on 29 estimates) was 4.0 per 1,000, and the 10% and 90% quantiles ranged from 1.8 to 11.6 per 1,000 (a 6.4-fold difference).
There were 32 prevalence estimates that could not be classified to the above criteria (NOS). Based on the distribution of these prevalence estimates, the median prevalence was 2.7 per 1,000 for persons, and the 10% and 90% quantiles ranged from 1.4 to 4.8 per 1,000 (a 3.4-fold difference).
The median LMR for persons (based on 27 estimates) was 7.2 per 1,000, and the 10% and 90% quantiles ranged from 3.1 to 27.1 per 1,000 (a 8.7-fold difference) (see Table 6).
Table 6 Quantiles and Moments of Studies with LMR per 1,000 by Sex
Data from nine studies.
The review identified 108 estimates based on inpatient-census-derived data. Based on the distribution of these estimates for persons, the median value was 2.4 per 1,000, and the 10% and 90% quantiles ranged from 0.07 to 10.0 per 1,000 (a 154-fold difference) (see Table 7). Inpatient-census-derived prevalence is not included for any subsequent analyses.
When point, period, and lifetime estimates were compared, the distributions were not significantly different (F
2,75 = 2.48, p = 0.09) . Estimates based on LMR were significantly higher than estimates based on lifetime estimates (F
1,25 = 4.53, p = 0.04).
Male Versus Female Prevalence
Table 8 shows the moments and quantiles for the combined prevalence estimates for persons, males, and females, and for a ratio derived from male:female estimates. Figure 8 shows the distribution of these data for males and females—these distributions were not significantly different (F
1,72 = 0.68, p = 0.41). For the male:female estimate ratio (based on 57 ratios), the median value was 1.11, and the 10% and 90% quantiles were 0.50 to 1.70 (approximately a 3.4-fold difference) (see Figure 9).
Figure 8 Cumulative Plots of Combined Prevalence Estimates per 1,000 by Sex
Figure 9 Cumulative Plots of the Male:Female Prevalence Estimate Ratio of Schizophrenia
Table 8 Quantiles and Moments of Combined Prevalence Estimates per 1,000 by Sex, and Male:Female Prevalence Estimate Ratio
Data from 93 studies.
Urbanicity of Sites
We identified 31 discrete-core studies with 73 rates from urban sites (see Table 4), 24 studies with 48 rates from rural sites, and 45 studies with 137 mixed urban/rural rates. There were four discrete-core studies providing rates for both urban (n = 12) and rural (n = 10) categories. Figure 10 and Table 9 show the distribution of overall prevalence based on rural, urban, and mixed urbanicity status for persons. While the mixed urban/rural estimates were higher than urban and rural rates, this difference was not statistically significant (F
2,235 = 1.63, p = 0.20), nor were urban estimates significantly different from rural estimates (F
1,120 = 0.95, p = 0.33).
Figure 10 Cumulative Plots of Combined Prevalence Estimates per 1,000 for Persons by Urbanicity
Table 9 Quantiles and Moments of Combined Prevalence Estimates per 1,000 for Persons by Urbanicity
Data from 93 studies.
Table 10 Quantiles and Moments of Combined Prevalence Estimates per 1,000 for Persons by Migrant Status, and Migrant:Native-Born Prevalence Estimate Ratio
Data from six studies.
Table 11 Quantiles and Moments of Combined Prevalence Estimates per 1,000 for Persons by Economic Status of Country
Data from 85 studies.
Migrant Status
We identified 15 migrant studies from eight countries: Australia (n = 2; [55,207]), Germany (n = 1; [92]), India (n = 1; [175]), Israel (n = 1; [200]), Taiwan (n = 1; [129]), the Netherlands (n = 1; [173]); United Kingdom (n = 7; [39,43,65–68,133]); and United States (n = 1; [180]).
Table S5 presents a detailed list of migrant studies with key descriptive variables, prevalence rates, and within-study migrant:native-born estimate ratios.
The number of different migrant groups in one study ranged between one and 38. There were six studies that derived data from inpatient-census-derived prevalence [43,55,65–67,92] and thus could not used in this analysis. In addition, four migrant studies did not present data for native-born populations [92,133,173,200]. Therefore, our analysis was limited to five papers only [39,129,175,180,207]. Based on 22 prevalence ratios, the median migrant:native-born prevalence ratio was 1.84 and the 10% and 90% quantiles were 0.86 to 6.41 (approximately a 7.5-fold difference) (see Table 10; Figure 11). When the migrant versus the native-born prevalence estimates were compared, there was a significant difference (F
1,2 = 5.57, p = 0.04).
Figure 11 Cumulative Plots of the Migrant:Native-Born Prevalence Estimate Ratio for Persons
Economic Status of Sites
Based on the three economic categories, we identified 19 estimates from least developed countries, 22 estimates from emerging economy countries, and 96 estimates from developed countries (see Table 11; Figure 12). When divided by this criterion, the prevalence estimate distributions were significantly different (F
2,85 = 3.57, p = 0.03), with the difference attributed to the lower prevalence estimate distribution for the less developed economies (developed versus least developed, F
1,74 = 6.55, p = 0.04). Table 12 also shows the male:female prevalence estimate ratio when subdivided by economic status. The distributions of these ratios (see Figure 13
), were not significantly different (F
2,42 = 0.44, p = 0.44).
Figure 12 Cumulative Plots of the Combined Prevalence Estimates per 1,000 for Persons by Economic Status of Country
Figure 13 Cumulative Plots of the Male:Female Prevalence Estimate Ratio of Schizophrenia by Economic Status of Country
Table 12 Quantiles and Moments of Male:Female Prevalence Estimate Ratio by Economic Status of Country
Data from 42 studies.
Quality Score
When the combined prevalence estimates for persons were divided into quality score terciles, the prevalence estimate distributions were significantly different (F
2,105 = 4.79, p = 0.01), with the highest quality studies reporting significantly higher prevalence estimates than the other two terciles (highest versus lowest quality scores, p = 0.02) (Table 13; Figure 14).
Figure 14 Cumulative Plots of Combined Prevalence Estimates per 1,000 for Persons by Tercile of Quality Score
Table 13 Quantiles and Moments of Combined Prevalence Estimates per 1,000 for Persons by Tercile of Quality Score
Data from 85 studies.
Other Special Group Studies
Details of these studies can be found in Table S6. We identified 41 studies that reported the prevalence of schizophrenia in other special groups. These studies came from 14 countries: Australia (n = 4), Canada (n = 4), Denmark (n = 3), Finland (n = 1), Germany (n = 3), India (n = 2), Israel (n = 1), Japan (n = 3), Romania (n = 1), Spain (n = 1), Sweden (n = 2), Taiwan (n = 1), United Kingdom (n = 2), and United States (n = 5).
Prevalence estimates were obtained from a range of population subgroups including elderly individuals (n = 10; [52,70,71,101,108,121,149–151,159]), ethnic groups (n = 8; [58,134,139,140,166,199,213,218]), Aborigines (n = 4; [105,106,115,164]), religious groups (n = 5; [29,80,128,182,191]), homeless individuals (n = 4; [118,161,192,194]), children and adolescents (n = 3; [57,185,189]), students (n = 2; [147,178]), twins (n = 1; [61]), industrial workers (n = 1; [172]), different castes (n = 1; [145]), and an isolate pedigree (n = 1; [99]).
The marked heterogeneity of these data does not make them suitable for combining. However, we note that prevalence estimates in some homeless populations were very high—300 per 1,000 persons for Sydney homeless individuals [194] and 131 per 1,000 persons for Los Angeles homeless individuals [118]. Conversely, some religious groups had very low prevalence estimates—0.36 per 1,000 persons for Amish individuals [80] and 1.29 per 1,000 persons for Hutterite individuals [29].
Discussion
There is a wealth of data available on the prevalence of schizophrenia—a total of 1,721 estimates from 188 studies were identified in this systematic review. These estimates were drawn from 46 countries, and were based on an estimated 154,140 potentially overlapping prevalent cases.
The median prevalence estimates for persons were 4.6 per 1,000 for point prevalence, 3.3 for period prevalence, 4.0 for lifetime prevalence, and 7.2 for LMR. These estimates are congruent with an earlier narrative review of 70 studies by Torrey [8], who reported an overall prevalence estimate of 4.6 per 1,000. Key policy documents have correctly estimated the point prevalence of schizophrenia at about four per 1,000 [2,225]; however, the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) [3], reported that the lifetime prevalence of schizophrenia is “usually estimated to be between 0.5% and 1%.” This overestimate is often repeated in textbooks [226]. As with the misunderstandings about the incidence of schizophrenia [21], this is another example where the research community needs to review their belief systems in the face of data. It is reasonable to assume that lifetime prevalence estimates for schizophrenia would be higher than point estimates. Surprisingly, the data in this review do not support this assumption. While outside the scope of the current review, the findings raise interesting research questions about factors that may influence prevalence (e.g., recovery, suicide, or other forms of early mortality). Indeed, it is curious that the identification of the onset of psychotic disorders has received so much recent attention [227,228], while we still struggle to understand the offset of schizophrenia. Point and period prevalence estimates assume that we can identify when someone has recovered from an illness. Recovery from schizophrenia clearly occurs [229–231], but it is unclear whether those who are free of positive symptoms but who have mild residual disability should be counted as “active” cases or not. The definitions of recovery versus persistence are multidimensional, and future prevalence studies will benefit if these definitions can be operationalized.
The median LMR estimate was 7.2 per 1,000, which is consistent with two other narrative reviews. Fremming [232], who reviewed 18 studies conducted in central Europe between 1926 and 1938, reported a mean LMR of 7.4 per 1,000, while Gottesman and Shields [233] reported a mean LMR of 8.0 per 1,000 in their classic review. As predicted, LMR estimates were significantly higher than lifetime estimates, which reflects the different heritage of these two indices. It is reasonable to assume that the oft-quoted statistic that “schizophrenia affects about one in a hundred” derives from LMR data (see [234]). However, one in a hundred is an overestimate—our systematic review agrees with two previous reviews showing that the LMR for schizophrenia is between seven and eight per 1,000. While the arithmetic mean value of 11.9 per 1,000 is more consistent with the “one in a hundred” dogma, the median is a more appropriate measure of central tendency for this skewed distribution. If we wish to provide the general public with a measure of the likelihood that individuals will develop schizophrenia during their lifetime, then a more accurate statement would be that “about seven to eight individuals per 1,000 will be affected.”
While there has been considerable debate about whether or not the incidence of schizophrenia varies between sites [21], there is a tacit understanding that the prevalence of schizophrenia is variable. For example, in an earlier review by Eaton [5], a 12-fold variation in point and a 10-fold variation in lifetime prevalence were noted. A recent systematic review by Goldner et al. [12] also observed a 13-fold variation in lifetime prevalence of schizophrenia. Based on the central 80% of the estimates (10% to 90% quantiles), the present review found that the different types of prevalence estimates had from 3.4-fold (point) to 4.6-fold (period) variation. The use of the 10% and 90% quantiles to define the central segment of the distribution means that our reporting of the variability of estimates is more conservative than other commentators (i.e., we have ignored 20% of the distribution in the tails). If we had included all data points, the range of prevalence estimates would have been much higher. Regardless of whether this variability is labeled “narrow” or “prominent,” the task for the researchers is to determine how much of this variation is a function of measurement error versus “true” underlying variation. With respect to measurement error, it should be noted that this study found that quality of the study does significantly influence prevalence estimates. Future studies could explore the impact of quality on the variation in prevalence estimates.
Sex and Schizophrenia
One of the unexpected findings of this review was that there was no statistically significant difference in prevalence estimates between males and females. In our previous study of incidence of schizophrenia we found a male:female risk ratio of 1.40 [1]. Because narrative reviews conclude that the course of the illness tends to be more severe in men than in women [235], we assumed that this would be reflected in a higher prevalence in males than females. The lack of coherence between (a) the sex differences found in the incidence of schizophrenia, (b) the presumed difference in course of illness, and (c) the identified lack of difference in prevalence warrants closer scrutiny.
Economic Status and Schizophrenia
In keeping with our hypothesis, the prevalence of schizophrenia is lower in developing nations than in developed nations. However, we urge caution in the interpretation of these data. The use of a single economic variable is a crude way to assess a complex and multidimensional concept. Furthermore, the median prevalence estimates for emerging economies are numerically higher than those for the richest countries. While not statistically significant, the results did identify many prevalence studies from the developing world where females outnumbered males. Recently, a study from China examined whether this unexpected sex ratio was due to differential suicide rates in males with schizophrenia [24]; however, this did not seem to explain the female excess. Our findings lend weight to the commentary by Ran and Yu-Hai Chen [25], drawing attention to the different features of schizophrenia in the developing world. Overall, the findings suggest that factors that influence the course of illness of schizophrenia in men and women differ around the world. Regardless of the mechanisms underlying this possibility, the findings highlight the importance of using systematic techniques to identify data; 17 studies included in this review were only available in languages other than English. We speculate that the results of past narrative reviews may have been biased towards data from developed nations. From a wider perspective, the findings reinforce the importance of encouraging more research from poorer countries [236].
Urbanicity and the Prevalence of Schizophrenia
In the previous systematic review of the incidence of schizophrenia, we found that urban sites had significantly higher incidence rates of schizophrenia than mixed urban/rural sites (there were too few pure rural sites to make the direct urban versus rural comparison) [1]. Contrary to our expectations, the prevalence of schizophrenia did not differ according to urbanicity. While Figure 10 suggests that mixed urban/rural sites have higher prevalence estimates than pure urban and rural sites, this study found, in fact, that there was no significant difference between urban, rural, and mixed sites. Perhaps the inclusion of many sites from the developing world in this review has confounded the expected urban/rural gradient. This will be examined in more detail in future analyses.
Migrant Status and the Prevalence of Schizophrenia
As predicted, prevalence estimates for migrant groups tend to be higher than estimates for native-born populations. This finding is consistent with past systematic reviews of the incidence of schizophrenia [1,33]. Migrant studies are prone to a range of methodological issues (e.g., differential pathways to care, diagnostic inaccuracies due to language and cultural practices, and uncertainty about the denominator required for the calculation of proportions). While the prevalence estimates included in this systematic review may share common biases, the increased prevalence of schizophrenia in migrant groups found in this study adds weight to the argument that migrant status is an important risk factor for schizophrenia.
Quality Scores and Other Special Groups
Reassuringly, studies that had higher overall quality scores tended to identify more cases, and thus generate higher prevalence estimates than lower quality studies. Future studies will explore whether the findings based on the overall studies persist in the subgroup of studies in the highest quality tercile.
With respect to the studies included in the category “other special groups,” the estimates are not readily comparable, but it is interesting to note that these studies reported a wide range of prevalence estimates (e.g., high in homeless populations and low in certain religious groups). Future publications will examine these groups in more detail.
Caveats
Based on our experience with previous systematic reviews, we acknowledge that we may have missed studies and/or made data entry errors. We encourage readers to inform us of missing studies or errors in the data. Updated lists of relevant studies and raw data will be available from the authors. Furthermore, in the absence of clear guidelines on how to synthesize descriptive studies [26,237], many of the rules we used to filter studies and extract data were necessarily ad hoc. In the future, researchers may wish to reanalyze the dataset using different criteria, and perform sensitivity analyses related to these choices.
Two of the prevalence types (LMR and inpatient-census-derived data) had distributions for persons that were higher than distributions for both males and females separately. This pattern, which is difficult to explain, was also noted in some of the previously published incidence distributions [1]. The impact of quality scores on this pattern will be assessed in future studies.
The planned sensitivity analyses were conducted on combined data, a strategy that reduced the number of comparisons substantially (one combined analysis versus five analyses on each of point, period, lifetime, LMR, and NOS data). However, the combined prevalence estimate included studies that contributed more than one prevalence type (e.g., one study could contribute both point and period prevalence estimates). Of the 94 studies, eight contributed more than one prevalence type to the combined prevalence estimates. While the analytic technique controlled for within-study variance, the combined dataset is not based on discrete data (in contrast to the prevalence-type-specific analyses).
It was disappointing that standard errors could be allocated to so few prevalence estimates (26%). Despite this, in the future we plan to undertake a traditional meta-analysis based on this subset of estimates in order to compare the pooled estimate values with those presented in the current study.
Concerning the analyses for urbanicity, the estimates from mixed urban/rural studies are likely to be very heterogeneous. Indeed, we allocated studies to the mixed category if there was any possibility that rural sectors were included. This bias would have made any true difference between urban versus mixed urban/rural more difficult to detect. There are good reasons to review the findings for both urbanicity and sex ratio more closely when categorized by economic status. Such analyses may help generate hypotheses for future analyses, but researchers need to be extremely cautious when systematic reviews are subjected to excessive data analyses (i.e., “data torturing” [238]). The contributing studies were not designed to test many of the hypotheses examined in this review, therefore researchers must be frugal in the use of planned sensitivity analyses, and cautious in the interpretation of the results. However, researchers are encouraged to freely explore the full data to examine additional research questions.
Conclusions
While there is substantial variation between sites, generally the prevalence of schizophrenia ranges from four to seven per 1,000 persons, depending on the type of prevalence estimate used. Countries from the developing world have a lower prevalence of schizophrenia. Overall, the prevalence of schizophrenia does not vary between the sexes; however, the data suggest that sex ratio of prevalence estimates may vary between sites more than previously believed. While the incidence of schizophrenia is higher in urban than rural settings, this is not reflected in the overall prevalence data. The prevalence of schizophrenia is higher in migrants than native-born individuals.
Regardless of the exact magnitude and precision of prevalence estimates, the numbers speak to a deeper, human dimension. Many people with schizophrenia have persisting symptoms, despite the best mix of interventions we can offer. This sobering reality has also emerged from research about “best buys” with respect to the cost of averting disability [239]. For schizophrenia, with the current mix of interventions we can only reduce 13% of the burden. If we improve efficiencies within the current services, we can do somewhat better (22%). In a utopian world, even if unlimited funding were available, three-quarters of the burden of schizophrenia would remain unavoidable [240]. This is a powerful argument for investing in applied and basic research.
As with its companion study on the incidence of schizophrenia [1], we hope that the current review will populate the “epidemiological landscape” with data, and that this enriched environment will select the fittest (most heuristic) hypotheses [21]. The epidemiological landscape of schizophrenia is no longer terra incognita—many of its contours have been mapped out. We can gain traction on this landscape and use the identified gradients to generate candidate risk factors for future research [241]. Equally, these systematic reviews have brought into focus the gaps in our knowledge—parts of the map “do not fit.” Paradoxes such as these can be powerful catalysts for advancing knowledge.
Supporting Information
Dataset S1 Access Dataset of Prevalence Studies
(272 KB ZIP).
Click here for additional data file.
Table S1 Definitions of Prevalence Estimate Types
(32 KB DOC).
Click here for additional data file.
Table S2 Definitions for the Variables Used to Characterize the Prevalence Studies
(43 KB DOC).
Click here for additional data file.
Table S3 Quality Score Criteria
(38 KB DOC).
Click here for additional data file.
Table S4 Characteristics of Core Prevalence Studies
(644 KB DOC).
Click here for additional data file.
Table S5 Characteristics of Migrant Prevalence Studies
(136 KB DOC).
Click here for additional data file.
Tables S6 Characteristics of Other Special Groups Prevalence Studies
(243 KB DOC).
Click here for additional data file.
Patient Summary
Background.
Schizophrenia is a very serious mental illness and a major contributor to the global burden of disease. The topic of this study is the question of how common schizophrenia is among different groups and in different countries around the world. “Prevalence” means the number of people who have the disease at a particular time. The study itself is a so-called systematic review, which means the researchers used prespecified methods for finding individual studies and for extracting and summarizing the data from these individual studies in as objective a way as possible.
Why Was This Study Done?
Health care planning is based on prevalence estimates, and as a result, many studies on schizophrenia prevalence have been done by researchers around the world. The authors decided to do a systematic review of these studies to come up with a scientifically sound view of the big picture.
What Did the Researchers Do?
They looked at a total of 1,721 estimates of the prevalence of schizophrenia from 188 studies and covering 46 countries. They then calculated median prevalence estimates (that is, the middle value of all estimates) over a variety of time periods (see below).
What Did They Find?
The take-home message from their study is that about seven to eight individuals out of 1,000 will be affected by schizophrenia. To be more precise, the researchers found the following median estimates for the prevalence of schizophrenia: 4.6 out of 1,000 people have the disease at a specific time point; 3.3 per 1,000 have the disease within a surveillance period one to 12 months long; the lifetime prevalence (the number of people in the population who have ever manifested the disease) is 4.0 per 1,000; and the lifetime morbid risk (the likelihood that a particular individual will develop schizophrenia in their lifetime) is 7.2 per 1,000. While previous research has shown that men have a higher risk of developing schizophrenia, the researchers found that the prevalence of schizophrenia was the same in men and women (suggesting that the course of the illness differs between the sexes). The prevalence of schizophrenia was lower in poorer countries than in richer countries.
What Does This Mean?
Based on these estimates, our textbook numbers on lifetime prevalence and overall risk for an individual to develop schizophrenia are probably too high. Taken together with estimates on the incidence of schizophrenia (that is, the annual number of new cases), it is also clear that current treatments fail to cure most patients with schizophrenia.
More Information Online.
Additional information on schizophrenia can be found at the following sources.
United States National Institutes of Mental Health (search for “schizophrenia”): http://www.nimh.nih.gov/
Schizophrenia.com, a not-for-profit Web site providing information and education on schizophrenia: http://www.schizophrenia.com
For an explanation of systematic reviews: http://www.shef.ac.uk/scharr/ir/units/systrev/definitions.htm; http://www.cochrane.org/index0.htm
For definitions of incidence and prevalence: http://www.wrongdiagnosis.com/admin/preval.htm
For more information about the systematic reviews of the incidence and prevalence of schizophrenia: http://www.qcmhr.uq.edu.au/epi/
The authors wish to express their gratitude to the following colleagues who assisted in the search for data and translation of the studies: R. C. Bland, D. Blazer, S. Caleo, G. Canino, B. Cooper, J. Copeland, J. Cullberg, H. Dominique, O. El Saadi, M. Fichter, R. Grawe, S. C. Gupta, H. Herrman, A. Isailovic, E. Jacko, F. Jacobi, H. Katchadourian, K. Kendler, B. Moreno-Kustner, I. Levav, C. MacCauley, D. McLean, R. McCreadie, P. Munk-Jorgensen, A. Preti, P. Rabins, J. Robertson, A. Robinson, S. Scheurer, P. Shrout, L. Teplin, P. Thomsen, E. F. Torrey, M. Von Korff, J. Waddington, A. Weeke, M. Weingarten, M. Weissman, Z. Welham, E. Wells, and H. Wittchen. The Stanley Medical Research Institute supported this project. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Citation: Saha S, Chant D, Welham J, McGrath J (2005) A systematic review of the prevalence of schizophrenia. PLoS Med 2(5): e141.
Abbreviations
LMRlifetime morbid risk
NOSnot otherwise specified
==== Refs
References
McGrath J Saha S Welham J El Saadi O MacCauley C A systematic review of the incidence of schizophrenia: The distribution of rates and the influence of sex, urbanicity, migrant status and methodology BMC Med 2004 2 13 15115547
Murray CJ Lopez AD The global burden of disease: A comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020 1996 Boston Harvard School of Public Health 990
American Psychiatric Association Diagnostic and statistical manual of mental disorders: DSM-IV, 4th ed 1994 Washington (DC) American Psychiatric Association 886
Jablensky A Hirsch SR Weinberger DR Schizophrenia: The epidemiological horizon 2003 Schizophrenia Oxford Blackwell Science 203 231
Eaton WW Epidemiology of schizophrenia Epidemiol Rev 1985 7 105 126 3902492
Eaton WW Tien AY Poeschla BD Den Boer JA Westenberg HGM van Praag HM Epidemiology of schizophrenia Advances in the neurobiology of schizophrenia 1995 Chichester (New York) John Wiley and Sons 469
Jablensky A Prevalence and incidence of schizophrenia spectrum disorders: Implications for prevention Aust N Z J Psychiatry 2000 34 S26 S34 11129313
Torrey EF Prevalence studies in schizophrenia Br J Psychiatry 1987 150 598 608 3307980
Breslow RA Ross SA Weed DL Quality of reviews in epidemiology Am J Public Health 1998 88 475 477 9518988
Egger M Smith GD Meta-Analysis. Potentials and promise BMJ 1997 315 1371 1374 9432250
Egger M Smith GD Phillips AN Meta-analysis: Principles and procedures BMJ 1997 315 1533 1537 9432252
Goldner EM Hsu L Waraich P Somers JM Prevalence and incidence studies of schizophrenic disorders: A systematic review of the literature Can J Psychiatry 2002 47 833 843 12500753
Kleinbaum DG Kupper LL Morgenstern H Epidemiologic research: Principles and quantitative methods 1982 Belmont (California) Lifetime Learning Publications 560
Stromgren E Zum Ersatz des Weinbergschen ‘abgekurzten Verfahren'. Zugleich ein Beitrag zur Frage der Erblichkeit des Erkrankungsalters bei der Schizophrenie Z Gesamte Neurol Psychiatr 1935 153 784 797
Odegaard O A statistical investigation of the incidence of mental disorder in Norway Psychiatr Q 1946 20 381 401 21002015
Larsson T Sjogren T A methodological, psychiatric and statistical study of a large Swedish rural population Acta Psychiatr Neurol Scand 1954 89 11 250
Alda M Dvorakova M Zvolsky P Papezova H Posmurova M Genetic aspects in chronic schizophrenia. Morbidity risks and contributory factors Schizophr Res 1989 2 339 344 2487175
Jablensky A Sartorius N Ernberg G Anker M Korten A Schizophrenia: Manifestations, incidence and course in different cultures. A World Health Organization ten-country study Psychol Med Monogr Suppl 1992 20 1 97 1565705
Cox DR Oakes D Analysis of survival data 1984 London Chapman and Hall 201
Fazel S Danesh J Serious mental disorder in 23000 prisoners: A systematic review of 62 surveys Lancet 2002 359 545 550 11867106
McGrath J Myths and plain truths about schizophrenia—The NAPE lecture 2004 Acta Psychiatr Scand 2005 111 4 11 15636588
Bresnahan M Menezes P Varma V Susser E Murray RM> Jones PB Susser E van Os J Cannon M Geographical variation in incidence, course and outcome of schiozphrenia: A comparison of developing and developed countries The epidemiology of schizophrenia 2003 Cambridge Cambridge University Press 18 33
Aleman A Kahn RS Selten JP Sex differences in the risk of schizophrenia: Evidence from meta-analysis Arch Gen Psychiatry 2003 60 565 571 12796219
Phillips MR Yang G Li S Li Y Suicide and the unique prevalence pattern of schizophrenia in mainland China: A retrospective observational study Lancet 2004 364 1062 1068 15380965
Ran MS Yu-Hai Chen E Suicide and schizophrenia in China Lancet 2004 364 1016 1017 15380949
Stroup DF Berlin JA Morton SC Olkin I Williamson GD Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis of Observational Studies in Epidemiology (MOOSE) group JAMA 2000 283 2008 2012 10789670
El-Saadi O Welham J Saha S MacCaulay C Collingwood L The incidence and prevalence of schizophrenia: Preliminary results from a systematic review Schizophr Res 2002 53 32
Saha S Welham J MacCaulay C El-Saadi O Chant D The incidence of schizophrenia: A systematic review Schizophr Res 2003 60 49
Nimgaonkar VL Fujiwara TM Dutta M Wood J Gentry K Low prevalence of psychoses among the Hutterites, an isolated religious community Am J Psychiatry 2000 157 1065 1070 10873912
Berlin JA Invited commentary: Benefits of heterogeity in meta-analysis of data from epidemiologic studies Am J Epidemiol 1995 142 383 387 7625402
Goldberg EM Morrison SL Schizophrenia and social class Br J Psychiatry 1963 109 785 802 14080574
Warner R Recovery from schizophrenia: Psychiatry and political economy 1985 London Routledge and Kegan Paul 380
Cantor-Graae E Selten JP Schizophrenia and migration: A meta-analysis and review Am J Psychiatry 2005 162 12 24 15625195
Central Intelligence Agency CIA World Factbook, 2004 2004 Washington (DC) Central Intelligence Agency, Office of Public Affairs 730
Soubbotina TP Beyond economic growth: An introduction to sustainable development, 2nd ed 2004 Washington (DC) World Bank 205
Astrup C The Berlevag project from 1939 through 1976 Acta Psychiatr Scand Suppl 1989 348 79 83 2801182
Awas M Kebede D Alem A Major mental disorders in Butajira, southern Ethiopia Acta Psychiatr Scand Suppl 1999 397 56 64 10470356
Babigian H Kaplan H Freedman A Sadock B Schizophrenia: Epidemiology Comprehensive textbook of psychiatry III 1980 Baltimore Williams and Wilkins 1113 1121
Bagley C The social aetiology of schizophrenia in immigrant groups Race Today 1969 1 170 174
Bamrah JS Freeman HL Goldberg DP Epidemiology of schizophrenia in Salford, 1974–84. Changes in an urban community over ten years Br J Psychiatry 1991 159 802 810 1790448
Bash KW Bash-Liechti J Studies on the epidemiology of neuropsychiatric disorders among the population of the city of Shiraz, Iran Soc Psychiatry 1974 9 163 171
Bates CE Van Dam CH Low incidence of schizophrenia in British Columbia coastal Indians J Epidemiol Community Health 1984 38 127 130 6431048
Bebbington PE Hurry J Tennant C Psychiatric disorders in selected immigrant groups in Camberwell Soc Psychiatry 1981 16 43 51
Ben-Tovim DI Cushnie JM The prevalence of schizophrenia in a remote area of Botswana Br J Psychiatry 1986 148 576 580 3779228
Bijl RV Ravelli A van Zessen G Prevalence of psychiatric disorder in the general population: Results of The Netherlands Mental Health Survey and Incidence Study (NEMESIS) Soc Psychiatry Psychiatr Epidemiol 1998 33 587 595 9857791
Bland RC Demographic aspects of functional psychoses in Canada Acta Psychiatr Scand 1977 55 369 380 868568
Bland RC Orn H 14-year outcome in early schizophrenia Acta Psychiatr Scand 1978 58 327 338 717003
Bland RC Long term mental illness in Canada: An epidemiological perspective on schizophrenia and affective disorders Can J Psychiatry 1984 29 242 246 6529715
Bland RC Newman SC Orn H Schizophrenia: Lifetime co-morbidity in a community sample Acta Psychiatr Scand 1987 75 383 391 3495958
Bland RC Orn H Newman SC Lifetime prevalence of psychiatric disorders in Edmonton Acta Psychiatr Scand Suppl 1988 338 24 32 3165592
Bojholm S Stromgren E Prevalence of schizophrenia on the island of Bornholm in 1935 and in 1983 Acta Psychiatr Scand Suppl 1989 348 157 166 2801177
Bollerup TR Prevalence of mental illness among 70-year-olds domiciled in nine Copenhagen suburbs Acta Psychiatr Scand 1975 51 327 339 1146591
Bondestam S Garssen J Abdulwakil AI Prevalence and treatment of mental disorders and epilepsy in Zanzibar Acta Psychiatr Scand 1990 81 327 331 2343759
Brunetti PM Rural Vaucluse: Two surveys on the prevalence of mental disorders. Summary of data Acta Psychiatr Scand Suppl 1975 263 12 15 1061489
Bruxner G Burvill P Fazio S Febbo S Aspects of psychiatric admissions of migrants to hospitals in Perth, Western Australia Aust N Z J Psychiatry 1997 31 532 542 9272263
Bulayeva K Pavlova T Kurbanov R Bulayev O Mapping genes of complex disease in genetic isolates of Daghestan Russ J Genet 2002 38 1539 1548
Burd L Kerbeshian J A North Dakota prevalence study of schizophrenia presenting in childhood J Am Acad Child Adolesc Psychiatry 1987 26 347 350 3496327
Burnam MA Hough RL Escobar JI Karno M Timbers DM Six-month prevalence of specific psychiatric disorders among Mexican Americans and non-Hispanic whites in Los Angeles Arch Gen Psychiatry 1987 44 687 694 3498452
Statistics Canada Mental Health Statistics, Volume II 1979 Ottawa Ministry of Industry, Trade and Commerce 123
Canino GJ Bird HR Shrout PE Rubio-Stipec M Bravo M The prevalence of specific psychiatric disorders in Puerto Rico Arch Gen Psychiatry 1987 44 727 735 3498456
Cannon TD Kaprio J Lonnqvist J Huttunen M Koskenvuo M The genetic epidemiology of schizophrenia in a Finnish twin cohort. A population-based modeling study Arch Gen Psychiatry 1998 55 67 74 9435762
Chattopadhyay O Gill JS Bali P Wig NN Psychotic disorders in the adult population of an urban slum Indian J Public Health 1989 33 37 2641743
Chen CN Wong J Lee N Chan-Ho MW Lau JT The Shatin community mental health survey in Hong Kong. II. Major findings Arch Gen Psychiatry 1993 50 125 133 8427552
Chen C Zhang WX Epidemiological survey on schizophrenia in 7 areas of China Chin J Psychiatry 1998 31 72 74
Clare AW Mental illness in the Irish emigrant J Ir Med Assoc 1974 67 20 24 4811395
Cochrane R Mental illness in immigrants to England and Wales: An analysis of mental hospital admissions, 1971 Soc Psychiatry 1977 12 25 35
Cochrane R Bal SS Mental hospital admission rates of immigrants to England: A comparison of 1971 and 1981 Soc Psychiatry Psychiatr Epidemiol 1989 24 2 11 2496474
Commander MJ Odell S Sashidharan SP Surtees PG Psychiatric morbidity in people born in Ireland Soc Psychiatry Psychiatr Epidemiol 1999 34 565 569 10651174
Cooper B Sosna U Psychische Erkrankung in der Altenbevoelkerung [Psychiatric illness in the elderly population] Nervenarzt 1983 54 239 249 6866167
Cooper B Home and away: The disposition of mentally ill old people in an urban population Soc Psychiatry 1984 19 187 196 6494987
Copeland JR Dewey ME Scott A Gilmore C Larkin BA Schizophrenia and delusional disorder in older age: Community prevalence, incidence, comorbidity, and outcome Schizophr Bull 1998 24 153 161 9502553
Crocetti GM Lemkau PV Kulcar Z Kesic B Selected aspects of the epidemiology of psychoses in Croatia, Yugoslavia. 3. The cluster sample and the results of the pilot survey Am J Epidemiol 1971 94 126 134 5568602
Dale PW Prevalence of schizophrenia in the Pacific Island populations of Micronesia J Psychiatr Res 1981 16 103 111 7265007
Di Marco G Prevalencia de desordenes mentales en el area metropolitana de la Republica Argentina Acta Psiquiat Psicol Amer Lat 1982 28 93 102
Dilling H Weyerer S Prevalence of mental disorders in the small-town—Rural region of Traunstein (Upper Bavaria) Acta Psychiatr Scand 1984 69 60 79 6702472
Dilling H Weyerer S Fichter M The upper Bavarian studies Acta Psychiatr Scand Suppl 1989 348 113 139 2801175
Dourado A Azevedo MH Macedo A Coelho I Valente J Reduced prevalence of psychoses in Santa Maria Island, Azores, Portugal Am J Med Gen 2000 96 513
Dube KC Mental disorder in Agra Soc Psychiatry 1968 3 139 143
Dube KC Kumar N An epidemiological study of schizophrenia J Biosoc Sci 1972 4 187 195 5030377
Egeland JA Hostetter AM Amish Study, I: Affective disorders among the Amish, 1976–1980 Am J Psychiatry 1983 140 56 61 6847986
Elnagar MN Maitra P Rao MN Mental health in an Indian rural community Br J Psychiatry 1971 118 499 503 5580365
Fink P Mental illness and admission to general hospitals: A register investigation Acta Psychiatr Scand 1990 82 458 462 2291415
Folnegovic Z Folnegovic-Smalc V Schizophrenia in Croatia: Interregional differences in prevalence and a comment on constant incidence J Epidemiol Community Health 1992 46 248 255 1645081
Freeman H Alpert M Prevalence of schizophrenia in an urban population Br J Psychiatry 1986 149 603 611 3814952
Fugelli P Mental health and living conditions in a fishing community in northern Norway Acta Psychiatr Scand Suppl 1975 263 39 42 1061493
Fujita T [Trends of psychiatric in and out patients] Nippon Koshu Eisei Zasshi 1991 38 233 245 1958872
Giel R Van Luijk JN Psychiatric morbidity in a small Ethiopian town Br J Psychiatry 1969 115 149 162 5778192
Goldacre M Shiwach R Yeates D Estimating incidence and prevalence of treated psychiatric disorders from routine statistics: The example of schizophrenia in Oxfordshire J Epidemiol Community Health 1994 48 318 322 8051535
Goodman AB Siegel C Craig TJ Lin SP The relationship between socioeconomic class and prevalence of schizophrenia, alcoholism, and affective disorders treated by inpatient care in a suburban area Am J Psychiatry 1983 140 166 170 6849427
Gorwood P Leboyer M Jay M Payan C Feingold J Gender and age at onset in schizophrenia: Impact of family history Am J Psychiatry 1995 152 208 212 7840353
Grawe RW Pedersen PB Widen JH Changes in prevalence and comorbidity in a total population of patients with psychotic disorders in Norwegian psychiatric hospitals Nord J Psychiatry 1997 51 127 132
Haasen C Lambert M Mass R Krausz M Impact of ethnicity on the prevalence of psychiatric disorders among migrants in Germany Ethn Health 1998 3 159 165 9798114
Hagnell O Prevalence of mental disorders Scand J Soc Med 1994 50 1 77
Halldin J Prevalence of mental disorder in an urban population in central Sweden Acta Psychiatr Scand 1984 69 503 518 6611017
Harvey CA Pantelis C Taylor J McCabe PJ Lefevre K The Camden schizophrenia surveys. II. High prevalence of schizophrenia in an inner London borough and its relationship to socio-demographic factors Br J Psychiatry 1996 168 418 426 8730937
Helgason L Psychiatric services and mental illness in Iceland. Incidence study (1966–1967) with 6–7 year follow-up Acta Psychiatr Scand Suppl 1977 1 140
Hinterhuber H Liensberger D idemiology of schizophrenic disorders] Wien Med Wochenschr 1998 148 253 258 9746967
Hodiamont P Peer N Syben N Epidemiological aspects of psychiatric disorder in a Dutch health area Psychol Med 1987 17 495 505 3496622
Hovatta I Terwilliger JD Lichtermann D Makikyro T Suvisaari J Schizophrenia in the genetic isolate of Finland Am J Med Genet 1997 74 353 360 9259368
Hwu HG Yeh EK Chang LY Prevalence of psychiatric disorders in Taiwan defined by the Chinese Diagnostic Interview Schedule Acta Psychiatr Scand 1989 79 136 147 2923007
Ichinowatari N Tatsunuma T Makiya H Epidemiological study of old age mental disorders in the two rural areas of Japan Jpn J Psychiatry Neurol 1987 41 629 636 3502586
Jay M Gorwood P Feingold J Leboyer M A one year prevalence study of schizophrenia on Reunion Island Eur Psychiatry 1997 12 284 288 19698537
Jayasundera MG Lin WCT Mental health surveys in Ceylon Mental health research in Asia and the Pacific 1969 Honolulu East-West Center Press 54 64
Jeffreys SE Harvey CA McNaught AS Quayle AS King MB The Hampstead Schizophrenia Survey 1991. I: Prevalence and service use comparisons in an inner London health authority, 1986–1991 Br J Psychiatry 1997 170 301 306 9246245
Jones IH Psychiatric disorders among Aborigines of the Australian Western Desert. II Soc Sci Med 1972 6 263 267 5028853
Jones IH Horne DJ Psychiatric disorders among aborigines of the Australian western desert. Further data and discussion Soc Sci Med 1973 7 219 228 4708234
Jorda-Moscardo E Munk-Jorgensen P A comparative census study of Danish schizophrenic patients in 1977 and 1982 Eur Arch Psychiatry Neurol Sci 1986 235 323 327 3732344
Junginger J Phelan E Cherry K Levy J Prevalence of psychopathology in elderly persons in nursing homes and in the community Hosp Community Psychiatry 1993 44 381 383 8462948
Katchadourian HA Churchill CW Social class and mental illness in urban Lebanon Soc Psychiatry 1969 4 49 55
Kay RW Prevalence of psychotic mental disorders in the Commonwealth of Dominica West Indian Med J 1989 39 30 31
Keatinge C Schizophrenia in rural Ireland: A case of service overutilisation Int J Soc Psychiatry 1987 33 186 194 3679722
Kebede D Alem A Major mental disorders in Addis Ababa, Ethiopia. I. Schizophrenia, schizoaffective and cognitive disorders Acta Psychiatr Scand Suppl 1999 397 11 17 10470349
Kendler KS Gallagher TJ Abelson JM Kessler RC Lifetime prevalence, demographic risk factors, and diagnostic validity of nonaffective psychosis as assessed in a US community sample. The National Comorbidity Survey Arch Gen Psychiatry 1996 53 1022 1031 8911225
Keown P Holloway F Kuipers E The prevalence of personality disorders, psychotic disorders and affective disorders amongst the patients seen by a community mental health team in London Soc Psychiatry Psychiatr Epidemiol 2002 37 225 229 12107714
Kidson MA Jones IH Psychiatric disorders among Aborigines of the Australian Western Desert Arch Gen Psychiatry 1968 19 413 417 5672797
Kinzie JD Leung PK Boehnlein J Matsunaga D Johnson R Psychiatric epidemiology of an Indian village. A 19-year replication study J Nerv Ment Dis 1992 180 33 39 1538204
Kjellin L Compulsory psychiatric care in Sweden 1979–1993. Prevalence of committed patients, discharge rates and area variation Soc Psychiatry Psychiatr Epidemiol 1997 32 90 96 9050350
Koegel P Burnam MA Farr RK The prevalence of specific psychiatric disorders among homeless individuals in the inner city of Los Angeles Arch Gen Psychiatry 1988 45 1085 1092 2461690
Kracuk E Materialy K Sravnitel'noi epidemiologii shizofreni Zh Nevropatol Psikhiatr Im S S Korsakova 1965 65 608 616 4222251
Kramer M von Korff M Kessler L The lifetime prevalence of mental disorders: Estimation, uses and limitations Psychol Med 1980 10 429 435 7443897
Kramer M German PS Anthony JC Von Korff M Skinner EA Patterns of mental disorders among the elderly residents of eastern Baltimore J Am Geriatr Soc 1985 33 236 245 3989184
Kulcar Z Crocetti GM Lemkau PV Kesic B Selected aspects of the epidemiology of psychoses in Croatia, Yugoslavia. II. Pilot studies of communities Am J Epidemiol 1971 94 118 125 5568601
Leaf PJ Myers JK McEvoy LT Robins LN Regier DA Procedures used in the epidemiologic catchment area study Psychiatric disorders in America 1991 New York The Free Press 11 32
Lee CK Kwak YS Rhee H Kim YS Han JH The nationwide epidemiological study of mental disorders in Korea J Korean Med Sci 1987 2 19 34 3269242
Lee CK Kwak YS Yamamoto J Rhee H Kim YS Psychiatric epidemiology in Korea. Part II: Urban and rural differences J Nerv Ment Dis 1990 178 247 252 2181056
Lehtinen V Lindholm T Veijola J Vaisanen E The prevalence of PSE-CATEGO disorders in a Finnish adult population cohort Soc Psychiatry Psychiatr Epidemiol 1990 25 187 192 2399475
Lehtinen V Joukamaa M Lahtela K Raitasalo R Jyrkinen E Prevalence of mental disorders among adults in Finland: Basic results from the Mini Finland Health Survey Acta Psychiatr Scand 1990 81 418 425 2356764
Levav I Kohn R Dohrenwend BP Shrout PE Skodol AE An epidemiological study of mental disorders in a 10-year cohort of young adults in Israel Psychol Med 1993 23 691 707 8234576
Lin TY Rin H Yeh EK Hsu CC Chu HM Caudill W Lin T Mental disorders in Taiwan, fifteen years later Mental health research in Asia and the Pacific 1969 Honolulu East-West Center Press 66 91
Lindstrom E Widerlov B von Knorring L The ICD-10 and DSM-IV diagnostic criteria and the prevalence of schizophrenia Eur Psychiatry 1997 12 217 223
Magzumova S Prevalence of psychic disorders in the Republic of Uzbekistan World J Biol Psychiatry 2001 2 203S
Mavreas VG Beis A Mouyias A Rigoni F Lyketsos GC Prevalence of psychiatric disorders in Athens: A community study Soc Psychiatry 1986 21 172 181 3787312
Mavreas VG Bebbington PE Psychiatric morbidity in London's Greek-Cypriot immigrant community. I. Associations with sociodemographic variables Soc Psychiatry 1987 22 150 159 3498222
McCreadie RG Leese M Tilak-Singh D Loftus L MacEwan T Nithsdale, Nunhead and Norwood: Similarities and differences in prevalence of schizophrenia and utilisation of services in rural and urban areas Br J Psychiatry 1997 170 31 36 9068772
Mehta P Joseph A Verghese A An epidemiologic study of psychiatric disorders in a rural area in Tamilnadu Indian J Psychiatry 1985 27 153 158 21927091
Moreno-Kustner B Rosales-Varo C Torres-Gonzalez F Emmett C The treated prevalence of schizophrenia in Granada. Data from the accumulative case register Acta Psychiatr Scand 2002 105 24
Munk-Jorgensen P Weeke A Jensen EB Dupont A Stromgren E Changes in utilization of Danish psychiatric institutions. II. Census studies 1977 and 1982 Compr Psychiatry 1986 27 416 429 3757490
Munk-Jorgensen P Lutzhoft JH Jensen J Stromgren E Trends in psychiatric hospitalization in Denmark: A 10-year register-based investigation Acta Psychiatr Scand 1992 86 79 83 1414407
Murphy HB Lemieux M Quelques considerations sur le taux eleve de schizophrenie dans un type de communaute Canadienne–Francaise Can Psychiatr Assoc J 1967 12 S71 S81
Murphy HB Differences between mental disorders of French Canadians and British Canadians Can Psychiatr Assoc J 1974 19 247 257 4843000
Murthy RS Kala R Wig NN Mentally ill in a rural community: Some initial experiences in case identification and management Indian J Psychiatry 1978 20 143 147
Myles-Worsley M Coon H Tiobech J Collier J Dale P Genetic epidemiological study of schizophrenia in Palau, Micronesia: Prevalence and familiality Am J Med Genet 1999 88 4 10 10050960
Nakamura Y Ojima T Oki I Tanihara S Yanagawa H Estimation of the future numbers of patients with mental disorders in Japan based on the results of National Patient Surveys J Epidemiol 1997 7 214 220 9465546
Nandi DN Ajmany S Ganguli H Banerjee G Boral GC Psychiatric disorders in a rural community in West Bengal—An epidomiological study Indian J Psychiatry 1975 17 87 99
Nandi DN Mukherjee SP Boral GC Banerjee G Ghosh A Socio-economic status and mental morbidity in certain tribes and castes in India—A cross-cultural study Br J Psychiatry 1980 136 73 85 7357224
Neehall J An analysis of psychiatric inpatient admissions from a defined geographic catchment area over a one-year period West Indian Med J 1991 40 16 21 1858369
Nica-Udangiu S Epidemiologic aspects of schizophrenias, schizotypal disorders and schizoaffective psychoses in the student population Neurol Psychiatr (Bucur) 1983 21 115 120 6867581
Nielsen J The Samso project from 1957 to 1974 Acta Psychiatr Scand 1976 54 198 222 970197
Nielsen JA Nielsen J Prevalence investigation of mental illness in the aged in 1961, 1972 and 1977 in a geographically delimited Danish population group Acta Psychiatr Scand Suppl 1989 348 95 104 2801184
Nilsson LV Prevalence of mental disorders in a 70-year-old urban sample—A cohort comparison J Clin Exp Gerontol 1983 5 101 120
Nilsson LV Persson G Prevalence of mental disorders in an urban sample examined at 70, 75 and 79 years of age Acta Psychiatr Scand 1984 69 519 527 6741600
O'Hare A Walsh D First admissions to Irish psychiatric hospitals 1965–1969. Preliminary findings J Ir Med Assoc 1972 65 189 192 5024779
Oakley-Browne MA Joyce PR Wells JE Bushnell JA Hornblow AR Christchurch Psychiatric Epidemiology Study, Part II: Six month and other period prevalences of specific psychiatric disorders Aust N Z J Psychiatry 1989 23 327 340 2803145
Ouspersicaya LY [Problems of methodology of comparative epidemiological studies] Zh Nevropatol Psikhiatr Im S S Korsakova 1978 78 742 749 665073
Padmavathi R Rajkumar S Kumar N Manoharan A Kamath S Prevalence of schizophrenia in an urban community in Madras Indian J Psychiatry 1988 30 233 239 21927316
Padmavathi R Rajkumar S Srinivasan TN Schizophrenic patients who were never treated—A study in an Indian urban community Psychol Med 1998 28 1113 1117 9794018
Pantelis C Taylor J Campbell PG The South Camden schizophrenia survey—An experience of community-based research Psychiatr Bull R Coll Psychiatr 1988 12 98 101
Pfeiffer WM [Psychoses in a tradition determined culture (Nias-Indonesia)] Confin Psychiatr 1974 17 15 41 4842124
Rabins PV Black B German P Roca R McGuire M The prevalence of psychiatric disorders in elderly residents of public housing J Gerontol A Biol Sci Med Sci 1996 51 M319 M324 8914505
Ran M Xiang M Li SX Shan Y Huang M Prevalence and outcome of schizophrenia in a Chinese rural area: An epidemiological study Aust N Z J Psychiatry 2002 37 452 457
Reker T Eikelmann B Folkerts H [Prevalence of psychiatric disorders and outcome of social integration of homeless men] Gesundheitswesen 1997 59 79 82 9156629
Repetto F Formigaro F Ferrari P Frascaroli G Lora A [Estimate of the hospital incidence of schizophrenia in Lombardy] Epidemiol Prev 1988 10 20 25
Ries R Bokan J Schuckit MA Modern diagnosis schizophrenia in hospitalized psychiatric patients Am J Psychiatry 1980 137 1419 1421 7435676
Rin H Lin TY Mental illness among Formosan Aborigines as compared with the Chinese in Taiwan J Ment Sci 1962 108 134 146 14492321
Robinson AD A century of delusions in south west Scotland Br J Psychiatry 1988 153 163 167 3076491
Roy C Choudhuri A Irvine D The prevalence of mental disorders among Saskatchewan Indians J Cross Cult Psychol 1970 1 383 392
Rumble S Swartz L Parry C Zwarenstein M Prevalence of psychiatric morbidity in the adult population of a rural South African village Psychol Med 1996 26 997 1007 8878332
Sachdeva J Singh S Sidhu BS Goyal RKD Sing J An epidemiological study of psychiatric disorders in rural Faridkot Indian J Psychiatry 1986 28 317 323 21927196
Sadoun R Quemada N Chevalier A Une approche epidemiologique des psychoses J L'Hopital Sainte-Anne 1979 5 1 113
Salan R Epidemiology of schizophrenia in Indonesia (the Tambora I study) ASEAN J Psychiatry 1992 2 52 57
Sampath HM Prevalence of psychiatric disorders in a southern Baffin Island Eskimo settlement Can Psychiatr Assoc J 1974 19 363 367 4426018
Satija DC Patni SK Nathawat SS Mental morbidity in industrial workers of Khetri copper complex Indian J Psychiatry 1984 26 147 155 21965974
Schrier AC van de Wetering BJ Mulder PG Selten JP Point prevalence of schizophrenia in immigrant groups in Rotterdam: Data from outpatient facilities Eur Psychiatry 2001 16 162 166 11353594
Sethi BB Gupta SC Kumar R Kumari P A psychiatric survey of 500 rural families Indian J Psychiatry 1972 14 183 196
Sethi BB Gupta SC Mahendru RK Kumari P Migration and mental health Indian J Psychiatry 1972 14 115 121
Shah AV Goswami UA Maniar RC Hajariwala DC Sinha BK Prevalence of psychiatric disorders in Ahmedabad (an epidemiological study) Indian J Psychiatry 1980 22 384 389 22058505
Shen YC Zhang WX Shu L Yang XL Cui YH Investigation of mental disorders in Beijing suburban district Chin Med J (Engl) 1981 94 153 156 6785030
Shingu K Sato M Miyoshi A Psychiatric diagnosis in a Japanese university population—Using DSM-III J Am Coll Health 1982 31 67 72 7153434
Shmaonova LM [Possibilities of an epidemiological method and some results of a population study of schizophrenia] Zh Nevropatol Psikhiatr Im S S Korsakova 1983 83 707 716 6880499
Shrout PE Canino GJ Bird HR Rubio-Stipec M Bravo M Mental health status among Puerto Ricans, Mexican Americans, and non-Hispanic whites Am J Community Psychol 1992 20 729 752 1302447
Sikanartey T Eaton WW Prevalence of schizophrenia in the Labadi District of Ghana Acta Psychiatr Scand 1984 69 156 161 6608212
Spencer J The mental health of Jehovah's Witnesses Br J Psychiatry 1975 126 556 559 1174772
Stefansson JG Lindal E Bjornsson JK Guomundsdottir A Lifetime prevalence of specific mental disorders among people born in Iceland in 1931 Acta Psychiatr Scand 1991 84 142 149 1950608
Sundaram D Sathyavathi K Murthy HN Ecological aspects in schizophrenia. Transactions of All-India Institute of Mental Health 1967 Bangladore India Institute of Mental Health 43 53
Suzuki M Morita H Kamoshita S idemiological survey of psychiatric disorders in Japanese school children. Part III: Prevalence of psychiatric disorders in junior high school children] Nippon Koshu Eisei Zasshi 1990 37 991 1000 2132376
Tansella M Balestrieri M Meneghelli G Micciolo R Trends in the provision of psychiatric care 1979–1988 Psychol Med 1991 19 5 54
Temkov I Jablensky A Boyadjieva M Use of reported prevalence data in cross-national comparisons of psychiatric morbidity Soc Psihijatr 1975 3 111 117
Thacore VR Gupta SC Suraiya M Psychiatric morbidity in a north Indian community Br J Psychiatry 1975 126 365 369 1148570
Thomsen PH Schizophrenia with childhood and adolescent onset—A nationwide register-based study Acta Psychiatr Scand 1996 94 187 193 8891086
Torrey EF Torrey BB Burton-Bradley BG The epidemiology of schizophrenia in Papua New Guinea Am J Psychiatry 1974 131 567 573 4819051
Torrey EF Prevalence of psychosis among the Hutterites: A reanalysis of the 1950–53 study Schizophr Res 1995 16 167 170 7577770
Vazquez C Munoz M Sanz J Lifetime and 12-month prevalence of DSM-III-R mental disorders among the homeless in Madrid: A European study using the CIDI Acta Psychiatr Scand 1997 95 523 530 9242848
Verghese A Beig A Senseman LA Rao SS Benjamin V A social and psychiatric study of a representative group of families in Vellore town Indian J Med Res 1973 61 608 620 4752395
Virgona A Buhrich N Teesson M Prevalence of schizophrenia among women in refuges for the homeless Aust N Z J Psychiatry 1993 27 405 410 8250783
Von Korff M Nestadt G Romanoski A Anthony J Eaton W Prevalence of treated and untreated DSM-III schizophrenia. Results of a two-stage community survey J Nerv Ment Dis 1985 173 577 581 3875685
Waddington JL Youssef HA Evidence for a gender-specific decline in the rate of schizophrenia in rural Ireland over a 50-year period Br J Psychiatry 1994 164 171 176 8173821
Waldo MC Schizophrenia in Kosrae, Micronesia: Prevalence, gender ratios, and clinical symptomatology Schizophr Res 1999 35 175 181 9988854
Warthen FJ Klee GD Bahn AK Gorwitz K Diagnosed schizophrenia in Maryland Psychiatr Res Rep Am Psychiatr Assoc 1967 22 149 170 6028376
Weiner BP Marvit RC Schizophrenia in Hawaii: Analysis of cohort mortality risk in a multi-ethnic population Br J Psychiatry 1977 131 497 503 588867
Weingarten MA Orron DE Schizophrenia in a Yemenite immigrant town in Israel Int J Soc Psychiatry 1983 29 249 254 6642917
Weissman MM Myers JK Harding PS Psychiatric disorders in a U.S. urban community: 1975–1976 Am J Psychiatry 1978 135 459 462 637143
Weissman MM Myers JK Psychiatric disorders in a U.S. community. The application of research diagnostic criteria to a resurveyed community sample Acta Psychiatr Scand 1980 62 99 111 7468289
Wells JE Bushnell JA Hornblow AR Joyce PR Oakley-Browne MA Christchurch Psychiatric Epidemiology Study, Part I: Methodology and lifetime prevalence for specific psychiatric disorders Aust N Z J Psychiatry 1989 23 315 326 2803144
Widerlov B Borga P Cullberg J Stefansson CG Lindqvist G Epidemiology of long-term functional psychosis in three different areas in Stockholm County Acta Psychiatr Scand 1989 80 40 46 2788354
Widerlov B Lindstrom E von Knorring L One-year prevalence of long-term functional psychosis in three different areas of Uppsala Acta Psychiatr Scand 1997 96 452 458 9421342
Wijesinghe CP Dissanayake SA Dassanayake PV Survey of psychiatric morbidity in a semi-urban population in Sri Lanka Acta Psychiatr Scand 1978 58 413 441 717008
Wijesinghe CP Clancy DJ Schizophrenia in migrants living in the western region of Melbourne Aust N Z J Psychiatry 1991 25 350 357 1958159
Wing L Wing JK Hailey A Bahn AK Smith HE The use of psychiatric services in three urban areas: An international case register study Soc Psychiatry 1967 2 158 167
Wittchen HU Essau CA von Zerssen D Krieg JC Zaudig M Lifetime and six-month prevalence of mental disorders in the Munich Follow-Up Study Eur Arch Psychiatry Clin Neurosci 1992 241 247 258 1576182
Wooff K Freeman HL Fryers T Psychiatric service use in Salford: A comparison of point-prevalence ratios 1968 and 1978 Br J Psychiatry 1983 142 588 597 6882982
Woogh C Is schizophrenia on the decline in Canada? Can J Psychiatry 2001 46 61 67 11221491
Xia Z Yan H Wang C Mental health care in Shanghai Int J Ment Health 1988 16 81 85
Xu WS [Analysis of the cultural and mental diseases rates of the three nationalities from She, Hui and Mongolian peoples] Chung Hua Shen Ching Ching Shen Ko Tsa Chih 1991 24 87 89, 124
Yang J [An epidemiologic survey of mental diseases in Xinjiang] Chung Hua Shen Ching Ching Shen Ko Tsa Chih 1989 22 366 368, 384
Youssef HA Kinsella A Waddington JL Evidence for geographical variations in the prevalence of schizophrenia in rural Ireland Arch Gen Psychiatry 1991 48 254 258 1996920
Youssef HA Scully PJ Kinsella A Waddington JL Geographical variation in rate of schizophrenia in rural Ireland by place at birth vs place at onset Schizophr Res 1999 37 233 243 10403195
Yursinova MS [Some results of an epidemiological study of schizophrenia in the city of Samarkand] Zh Nevropatol Psikhiatr Im S S Korsakova 1982 9 93 98
Zhang A Snowden LR Ethnic characteristics of mental disorders in five U.S. communities Cultur Divers Ethnic Minor Psychol 1999 5 134 146 15605683
Zharikov NM Epidemiological study of mental illness in the U.S.S.R Soc Psychiatry 1968 3 135 138
de Salvia D Barbato A Salvo P Zadro F Prevalence and incidence of schizophrenic disorders in Portogruaro. An Italian case register study J Nerv Ment Dis 1993 181 275 282 8501442
Ni Nuallain M O'Hare A Walsh D The prevalence of schizophrenia in three counties in Ireland Acta Psychiatr Scand 1990 82 136 140 2239357
Kendler KS McGuire M Gruenberg AM Walsh D An epidemiologic, clinical, and family study of simple schizophrenia in County Roscommon, Ireland Am J Psychiatry 1994 151 27 34 8267130
Erlenmeyer-Kimling L Squires-Wheeler E Adamo UH Bassett AS Cornblatt BA The New York High-Risk Project. Psychoses and cluster A personality disorders in offspring of schizophrenic parents at 23 years of follow-up Arch Gen Psychiatry 1995 52 857 865 7575106
O'Hare A Walsh D Activities of Irish Psychiatric Hospitals and Units 1965–1969, and 1970 1972 Dublin The Medico-Social Research Board 33
World Health Organization World Health Report 2001—Mental health: New understanding, new hope. Geneva: World Health Organization. Available: http://www.who.int/whr/2001/en/whr01_en.pdf
2001 Accessed 5 April 2005
Murray RM Murray R Hill P McGuffin P Schizophrenia The essentials of postgraduate psychiatry, 3rd ed 1997 Cambridge Cambridge University Press 281 309
Yung AR McGorry PD The prodromal phase of first-episode psychosis: Past and current conceptualizations Schizophr Bull 1996 22 353 370 8782291
McGlashan TH Early detection and intervention of schizophrenia: Rationale and research Br J Psychiatry Suppl 1998 172 3 6 9764119
Sartorius N Jablensky A Korten A Ernberg G Anker M Early manifestations and first-contact incidence of schizophrenia in different cultures. A preliminary report on the initial evaluation phase of the WHO Collaborative Study on determinants of outcome of severe mental disorders Psychol Med 1986 16 909 928 3493497
Harding CM Zubin J Strauss JS Chronicity in schizophrenia: Fact, partial fact, or artifact? Hosp Community Psychiatry 1987 38 477 486 3297969
Harding CM Course types in schizophrenia: An analysis of European and American studies Schizophr Bull 1988 14 633 643 3064287
Fremming KH Morbidity risk of mental disease and other mental abnormalities in an average Danish population 1947 Copenhagen Munksgaard 278
Gottesman II Shields J Schizophrenia: The epi-genetic puzzle 1982 Cambridge Cambridge University Press 258
Spering MK Schizophrenia. NIH Publication No. 3517. Bethesda (Maryland): National Institute of Mental Health. Available: http://www.nimh.nih.gov/publicat/NIMHschizoph.pdf
2004 Accessed 5 April 2005
Leung A Chue P Sex differences in schizophrenia, a review of the literature Acta Psychiatr Scand Suppl 2000 401 3 38 10887978
Alem A Kebede D Conducting psychiatric research in the developing world: Challenges and rewards Br J Psychiatry 2003 182 185 187 12611776
von Elm E Egger M The scandal of poor epidemiological research BMJ 2004 329 868 869 15485939
Mills JL Data torturing N Engl J Med 1993 329 1196 1199 8166792
Andrews G Sanderson K Corry J Issakidis C Lapsley H Cost-effectiveness of current and optimal treatment for schizophrenia Br J Psychiatry 2003 183 427 435 14594918
Andrews G Issakidis C Sanderson K Corry J Lapsley H Utilising survey data to inform public policy: Comparison of the cost-effectiveness of treatment of ten mental disorders Br J Psychiatry 2004 184 526 533 15172947
McGrath JJ Invited commentary: Gaining traction on the epidemiologic landscape of schizophrenia Am J Epidemiol 2003 158 301 304 12915494
| 15916472 | PMC1140952 | CC BY | 2021-01-05 10:39:48 | no | PLoS Med. 2005 May 31; 2(5):e141 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020141 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591647310.1371/journal.pmed.0020142Research in TranslationHIV/AIDSHIV Infection/AIDSMedicine in Developing CountriesThe Development of Vaginal Microbicides for the Prevention of HIV Transmission Research in TranslationWeber Jonathan *Desai Kamal Darbyshire Janet on behalf of the Microbicides Development Programme Jonathan Weber and Kamal Desai are at the Wright-Fleming Institute, Imperial College London, United Kingdom. Janet Darbyshire is at the Medical Research Council Clinical Trials Unit, London, United Kingdom.
Competing Interests: JW and JD are principal investigators on the Microbicides Development Programme, which is trialling PRO 2000.
*To whom correspondence should be addressed. E-mail: [email protected] 2005 31 5 2005 2 5 e142Copyright: © 2005 Weber 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.Given that we still do not have an effective vaccine against HIV, the development of novel biomedical methods for preventing HIV transmission remains a top priority in controlling the HIV pandemic.
==== Body
Microbicides are chemical agents used topically by women within the vagina in order to prevent infection by HIV and potentially by other enveloped viruses and sexually transmitted pathogens. Prototype microbicides are designed to be inserted prior to each act of sexual intercourse and could also be contraceptive, although most current potential microbicides are not. Several proof-of-principle phase III trials of candidate microbicides are currently in progress or are shortly to commence, and a definitive answer to their efficacy and safety is anticipated by 2008.
Why Develop Chemical Barriers against HIV?
It is the unequivocal experience of the 20th century that the control of viral infections globally can best be achieved through the wide application of preventative vaccines. As discussed recently in PLoS Medicine, there is to be increased coordination and investment in a “Global Enterprise” to create an AIDS vaccine [1]. Disappointingly, however, in spite of much effort and investment, there is currently no proven vaccine against HIV, and no genuinely plausible candidate on the horizon. Richard Horton, editor of The Lancet, commented recently in the New York Review of Books that “one would be wise to plan the control of HIV infection globally over the next decade without assuming that a vaccine will be available” [2].
In the absence of a vaccine, novel biomedical methods for the prevention of HIV transmission assume far greater priority. Physical barriers such as the condom have been repeatedly demonstrated to have a high efficacy in the prevention of HIV transmission during sexual intercourse [3]. However, condoms remain almost exclusively under the control of the male partner, and in many societies women simply cannot negotiate condom use. A female-controlled method to prevent or reduce HIV risk is highly desirable, especially as this would allow sex education to promote mechanisms for both males and females, with the potential for an additive effect on the reduction of HIV transmission if used with a condom or other barrier methods. To be successful in reducing HIV transmission through being widely accepted and available, microbicides will need to be proven effective in phase III trials, and will need to be non-toxic and well tolerated, odour-less and colour-less, easily administered, and cheap to manufacture.
Women at a Microbicides Development Programme phase III trial site
(Photo: Frank Herholdt; Copyright: © 2005 Microbicides Development Programme. This is an open-access photo 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.)
First-Generation Microbicides: Surfactants
The first vaginal microbicide to be studied was nonoxynol-9 (N9), an anionic surfactant initially developed in the 1960s as a contraceptive spermicide with lubricant properties, and latterly used extensively to coat latex condoms. N9 destroys the integrity of lipid bilayer membranes, and so has a virucidal action through disrupting the viral envelope.
The antiviral activity of N9 in vitro was first recognised in 1985 [4]. Evidence of activity in vivo was reported by Miller et al. in 1992 [5], with local N9 administration leading to reduction in transmission of simian immunodeficiency virus in a vaginal challenge macaque model. N9 has an IC50 of 2–5 ug/ml for both primary and laboratory-adapted HIV-1 strains (IC50 is the quantity of a substance that reduces HIV infection of cells in culture by 50%), and it is active against diverse HIV genotypes in vitro. However, the antiviral effect of N9 is non-specific, with the CC50 (the quantity of a substance required to damage cells in culture such that active uptake of tritiated thymidine or other biomarkers is reduced by 50%) and IC50 occurring at similar concentrations.
Early experience in placebo-controlled field studies by Joan Kreiss et al. in 1992 in Kenya [6] and by Ron Roddy et al. in 1998 among commercial sex workers in Cameroon [7] suggested that N9 may be associated with local vaginal toxicity, including ulcerations, without clear evidence of efficacy against HIV transmission. A phase I safety study of N9 in healthy women volunteers in London showed that regular administration of N9 led to histological evidence of vaginal inflammation, including increased CD4+ T lymphocytes and macrophages [8]. The final phase III multi-centre randomised placebo-controlled trial (COL 1492) of N9, undertaken by the United Nations Joint Programme on HIV/AIDS in 2002, showed that N9 had no efficacy in preventing HIV transmission [9]. Indeed, the transmission rate was marginally higher in the N9 group, and it was considered that this might be related to the local vaginal toxicity seen in earlier studies.
More recently, the company Biosyn has developed a new surfactant microbicide, SAVVY (Figure 1), which is thought to be significantly less cytotoxic than N9 while retaining antiviral activity, and which is currently in phase III trials in sites in West Africa. SAVVY is believed to exert its virucidal activity through disruption of the HIV envelope lipid bilayer via the same mechanism as N9, although the basis of its lower cytotoxicity is not reported [10].
Figure 1 Sites of Action of Candidate Microbicides
(Illustration: Giovanni Maki)
Second-Generation Microbicides: Blocking HIV Binding
The problems with N9 strongly suggested that for new vaginal microbicides to be effective, these agents would need to be largely devoid of local vaginal toxicity. The first class of agents to be systematically explored (in 1992) were agents that blocked the binding of HIV-1 to target cells in vitro [11]. The anti-HIV activity of polyanions had first been described in 1987 [12], but these agents had too high a molecular weight to be orally absorbed, and interfered with clotting if given parenterally, and hence they were not developed for antiviral therapy.
Candidate agents in this class (i.e., agents that block HIV binding) now include several high molecular weight anionically charged sulphated polymers such as PRO 2000 (a naphthalene sulphonate polymer), carageenan (a naturally occurring sulphated sugar polymer), and cellulose sulphate (Figure 1). These three agents have potent anti-HIV activity in vitro against the two major classes of HIV-1 isolates using both the CCR5 and the CXCR4 co-receptors. Unlike N9, these agents have a very low toxicity, and the therapeutic index (ratio of CC50 to IC50) is above 10,000.
PRO 2000 has been extensively studied in vivo in the simian-human immunodeficiency virus macaque vaginal challenge model, with 50%–90% protection against infection depending on experimental conditions [13]. The absence of absorption of these high molecular weight agents is a positive benefit for intra-vaginal use, as very high local concentrations of these polyanions have been shown to be tolerable in animal models and in human phase II trials [14]. PRO 2000, carageenan, and cellulose sulphate are all in, or will shortly be in, phase III trials in women at risk of HIV infection in Africa (see Table 1).
Table 1 Status of Clinical Trials of Current Microbicide Compounds
Source: Alliance for Microbicide Development
The healthy vagina is maintained at low pH and microbially populated with hydrogen-peroxide-producing lactobacilli. There is some evidence that maintenance of this healthy vaginal milieu may protect against infection by HIV and other sexually transmitted infections; all current microbicides are buffered to pH 4.5 in order to promote vaginal health, and all have been investigated for their ability to sustain vaginal lactobacilli in vivo [15]. One candidate microbicide, “BufferGel”, whose mechanism of action is specifically based on the maintenance of a low vaginal pH, is being assessed in parallel with PRO 2000 in a phase IIb trial ([16]; see Table 1).
Third-Generation Microbicides: Topical Antiretrovirals
The polyanions meet many of the desirable properties of an ideal vaginal microbicide, being cheap to manufacture, active in vitro and in macaque vaginal challenge experiments, and safe and tolerable in vivo in phase II trials. However, they rely on charge for their effect, and they are non-specific for enveloped viruses and may all be slightly less active against CCR5-using primary virus isolates, the major transmitted variant of HIV-1. Ongoing phase III trials will ultimately determine whether their activity and tolerability translate into efficacy. It is too premature to assume success with the second-generation agents, and HIV-specific microbicide agents are currently being developed in phase I and II clinical trials.
Tenofovir, a nucleotide analogue reverse transcriptase inhibitor antiretroviral drug with a long half-life, has been successfully introduced as an oral agent for HIV infection. Studies of tenofovir in macaques suggest that pre-dosing may prevent simian immunodeficiency virus infection, and a gel formulation has been developed for vaginal use. A multi-centre phase II trial of vaginal tenofovir gel is planned for 2005.
Non-nucleoside reverse transcriptase inhibitors (NNRTI) represent a range of chemical structures with direct binding to the active site of HIV-1, and efavirenz and nevirapine are very successful NNRTIs used in combination antiretroviral therapy. Some identified NNRTIs are highly lipophilic, unsuitable for oral administration but potentially useful for intra-vaginal use as microbicides. Of these, UC781 and TMC120 are being formulated and developed as vaginal gels (Figure 1), with the aim of entering phase III trials in 2006 or 2007.
Finally, data have recently been reported on the prevention of simian-human immunodeficiency virus transmission in the macaque vaginal challenge model by the use of monoclonal antibodies to HIV-1 [17] and by a topical small-molecule CCR5 inhibitor, PSC-RANTES [18]. In both cases, the quantity of the agent required to prevent transmission in vivo was considerably greater than that required in vitro. The use of these highly specific agents in combination with other second- or third-generation microbicides may address this issue. However, monoclonal antibodies and peptides alone are never likely to meet the requirement for a microbicide to be cheaply manufactured.
Phase III Trials of Vaginal Microbicides
Clinical trials of the efficacy of microbicides present particular problems over and above those encountered for trials of preventative HIV vaccines. Trial populations must be identified with an HIV incidence of at least 2% if sample size is to be feasible. However, this high level of HIV transmission must be vigorously addressed though counselling, screening for and treatment of sexually transmitted infections, and provision of condoms and health education. Adherence to use of the microbicide will be central to a successful outcome, and these agents will need to be used to protect every risky act of intercourse. As with male condoms, adherence is unlikely to be complete, and may reduce over time. Thus, trials designed to assess proof of efficacy should be for as short an intervention period as possible in order not to dilute any antiviral effect of the microbicide through the likely reduction in adherence over time.
Further, non-vaginal exposure to HIV such as through frequent anal intercourse or the use of non-sterile needles, will reduce the effect of a microbicide. Thus, considerable behavioural data will need to be collected at these phase III trials for the results to be fully interpreted and understood. These trials will therefore be large, requiring multiple sites in high-incidence regions, and they will be long in duration and hence expensive.
Phase III trials of novel microbicide products undertaken in resource-poor regions in large numbers of women are financially challenging, inevitably costing upwards of $US50 million depending on sample size. In the absence of commercial investment, this burden falls on the charitable and/or public sector. It is highly likely that because of the high costs, phase III trials will be the major rate-limiting step to new product development and testing.
In the long term, the effectiveness of microbicides will be highly dependent on continued high-level adherence, and this will need to examined in longer term trials As the safety and tolerability of these new materials will continue to be paramount, some participants even in shorter trials will need longer follow-up for safety. Ultimately, it is possible that adherence could be enhanced through novel depot methods for the sustained delivery of microbicides within the vagina without the need for dosing prior to each act of intercourse. For example, the development of resident vaginal rings allowing sustained slow release of the active agent over weeks or months would overcome the key issue of adherence.
Advocacy, Funding, and Public–Private Partnerships
To date, there has been little interest from the pharmaceutical industry, outside of biotech companies, in the development of vaginal microbicides, and the burden of funding has fallen on the public and charitable sectors. This reliance on non-private funding has led to the new agents being developed in the academic sector, which has probably slowed progress even though the approach has been highly cost effective. However, microbicide research has attracted considerable political attention because of the urgency of the HIV epidemic, the plight of vulnerable women in high-incidence regions, and the delays in progress towards an HIV vaccine. A product-development public–private partnership has been established, the International Partnership for Microbicides, which is establishing clinical trial sites and is developing TMC120. Other advocacy groups include the Global Campaign for Microbicides and the Alliance for Microbicides, which ensure that pressure is maintained for funding the development of an effective microbicide.
Although it is hoped that one or more of the second-generation potential microbicides will have a positive risk–benefit ratio, they may have only partial efficacy, perhaps being as low as 35% effective. Indeed, the effect of these second-generation agents may be even lower when they are used outside of a clinical trial. Alternative approaches that require less attention and action from users—such as vaginal rings, more potent, longer acting products, or combinations of agents that might need to be used only once a day—could increase the effectiveness of microbicides and therefore have a much greater effect on HIV transmission.
However, modelling suggests that even a microbicide shown to be only partially effective, if used by women in concert with the promotion of condoms for men, would have a beneficial effect on the reduction of HIV transmission at a population level. In a sub-Saharan setting where endemic HIV prevalence is currently 10.8%, the introduction of a microbicide of 50% efficacy covering 50% of sex acts in high-risk women could achieve a population-wide reduction in HIV prevalence to 8.1% after 20 years. Concurrent promotion of condoms additionally covering 50% of sex acts in high-risk men could potentially achieve a prevalence as low as 1.4% (see simulations; Table 2). In the continued absence of an effective HIV vaccine, a partially active microbicide would still be a highly desirable intervention.
Table 2 Modelling Microbicide Effectiveness
This table relates to a model of the future effect of a partially active microbicide on the HIV epidemic in a country with a current prevalence of 10.8%, and the additive effect of promoting concurrent condom use. It is assumed in simulations that condoms prevent both male to female and female to male infection while the microbicide is assumed to protect only the female. Condoms are assumed to have 95% efficacy.
Citation: Weber J, Desai K, Darbyshire J (2005) The development of vaginal microbicides for the prevention of HIV transmission. PLoS Med 2(5): e142.
Abbreviations
N9nonoxynol-9
NNRTInon-nucleoside reverse transcriptase inhibitor
==== Refs
References
Coordinating Committee of the Global HIV/ AIDS Vaccine Enterprise The Global HIV/AIDS Vaccine Enterprise: Scientific strategic plan PLoS Med 2005 2 e25 15740411
Horton R AIDS: The elusive vaccine New York Rev Books 2003 September 23 51 14 53 58
Weller S Davis K Condom effectiveness in reducing heterosexual HIV transmission Cochrane Database Syst Rev 2002 2002 CD003255
Hicks DR Martin LS Getchell JP Heath JL Francis DP Inactivation of HTLV-III/LAV-infected cultures of normal human lymphocytes by nonoxynol-9 in vitro Lancet 1985 2 1422 1423 2867413
Miller CJ Alexander NJ Gettie A Hendrickx AG Marx PA The effect of contraceptives containing nonoxynol-9 on the genital transmission of simian immunodeficiency virus in rhesus macaques Fertil Steril 1992 57 1126 1128 1315297
Kreiss J Ngugi E Holmes K Ndinya-Achola J Waiyaki P Efficacy of nonoxynol 9 contraceptive sponge use in preventing heterosexual acquisition of HIV in Nairobi prostitutes JAMA 1992 268 477 482 1320133
Roddy RE Zekeng L Ryan KA Tamoufe U Weir SS A controlled trial of nonoxynol 9 film to reduce male-to-female transmission of sexually transmitted diseases N Engl J Med 1998 339 504 510 9709043
Stafford MK Ward H Flanagan A Rosenstein IJ Taylor-Robinson D Safety study of nonoxynol-9 as a vaginal microbicide: Evidence of adverse effects J Acquir Immune Defic Syndr Hum Retrovirol 1998 17 327 331 9525433
Van Damme L Ramjee G Alary M Vuylsteke B Chandeying V Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: A randomised controlled trial Lancet 2002 360 971 977 12383665
Krebs FC Miller SR Catalone BJ Welsh PA Malamud D Sodium dodecyl sulfate and C31G as microbicidal alternatives to nonoxynol 9: Comparative sensitivity of primary human vaginal keratinocytes Antimicrob Agents Chemother 2000 44 1954 1960 10858360
McClure MO Moore JP Blanc DF Scotting P Cook GM Investigations into the mechanism by which sulfated polysaccharides inhibit HIV infection in vitro AIDS Res Hum Retroviruses 1992 8 19 26 1346567
Mitsuya H Popovic M Yarchoan R Matsushita S Gallo RC Suramin protection of T cells in vitro against infectivity and cytopathic effect of HTLV-III Science 1984 226 172 174 6091268
Weber J Nunn A O'Connor T Jeffries D Kitchen V ‘Chemical condoms’ for the prevention of HIV infection: Evaluation of novel agents against SHIV(89.6PD) in vitro and in vivo AIDS 2001 15 1563 1568 11504989
Mayer KH Karim SA Kelly C Maslankowski L Rees H Safety and tolerability of vaginal PRO 2000 gel in sexually active HIV-uninfected and abstinent HIV-infected women AIDS 2003 17 321 329 12556685
Rosenstein IJ Stafford MK Kitchen VS Ward H Weber JN Effect on normal vaginal flora of three intravaginal microbicidal agents potentially active against human immunodeficiency virus type 1 J Infect Dis 1998 177 1386 1390 9593030
Mayer KH Peipert J Fleming T Fullem A Moench T Safety and tolerability of BufferGel, a novel vaginal microbicide, in women in the United States Clin Infect Dis 2001 32 476 482 11170957
Veazey RS Shattock RJ Pope M Kirijan JC Jones J Prevention of virus transmission to macaque monkeys by a vaginally applied monoclonal antibody to HIV-1 gp120 Nat Med 2003 9 343 346 12579198
Lederman MM Veazey RS Offord R Mosier DE Dufour J Prevention of vaginal SHIV transmission in rhesus macaques through inhibition of CCR5 Science 2004 306 485 487 15486300
| 15916473 | PMC1140953 | CC BY | 2021-01-05 10:39:53 | no | PLoS Med. 2005 May 31; 2(5):e142 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020142 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591647410.1371/journal.pmed.0020144Correspondence and Other CommunicationsInfectious DiseasesEpidemiology/Public HealthInfectious DiseasesFundamental Limits to the Precision of Early Warning Systems for Epidemics of Infectious Diseases CorrespondenceDrake John M University of California at Santa BarbaraSanta Barbara, CaliforniaUnited States of AmericaE-mail: [email protected]
Competing Interests: The author has declared that no competing interests exist.
5 2005 31 5 2005 2 5 e144Copyright: © 2005 John M. Drake.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.
Climate Drives the Meningitis Epidemics Onset in West Africa
==== Body
The development of early warning systems (EWSs) for epidemics of infectious diseases based on recurrent statistical patterns in other kinds of information, particularly data on climate, is an active area of research [1,2]. Judging from the estimated burden of diseases for which EWSs might be developed, such systems, if effective, would contribute greatly to human welfare and could potentially save many lives [2]. According to a recent report [2], EWSs have two principal aims: (i) to identify whether an epidemic will occur and (ii) to predict the number of cases that will result from it. For directly transmitted diseases, this second aim may be unattainable at the desired levels of precision, regardless of the quality of information.
As an example, in a recent report on the relationship between climate and outbreaks of meningococcal meningitis, the authors found that the timing of epidemics is highly predictable from information on the dynamics of a seasonal weather pattern, the Harmattan winds, but that the final epidemic size is not [1]. This finding is not surprising. The characteristics of disease outbreaks, particularly outbreaks of emerging diseases to which human populations are highly susceptible, prevent highly precise forecasts.
The reason that precise estimates of the final epidemic size cannot be obtained can be understood intuitively. Consider the following description of a typical outbreak. Characteristically, an outbreak begins with a small number of initially infectious individuals. Subsequent infectious contacts are mediated by a wide range of social interactions—contacts within and among households and communities—so that even individuals that are virtually identical can differ considerably in the number of secondary infections they cause. This is a micro-scale cause of variation compared with macro-scale, population-level sources of variation. The important implication for EWSs is that in such situations, especially where the basic reproductive ratio of infections (R0) is initially very high but is rapidly reduced (perhaps by public-health interventions), small deviations in the realized number of infectious contacts are amplified, resulting in relatively large variation in the final size of the outbreak. Because this variation reflects differences in individual behavior and not macroscopic characteristics of epidemic spread, it is unlikely that climate or other data contain any information about this source of variation (though such data do contain information about macroscopic variation).
A more formal explanation of this phenomenon can be formulated based on a simple model of an epidemic in which an EWS captures all macroscopic causes of variation in the final epidemic size but no microscopic causes. Obviously, an EWS cannot realistically be expected to capture even all the macroscopic information. Thus, this limit to precision is a fundamental limit and should be interpreted as a theoretical upper bound on forecast precision. The simplest case considers a disease with only two macroscopic epidemiological characteristics, an infection rate and a removal rate, which may change over time as in the case of meningococcal meningitis. In particular, we assume that there is no immunity in the population and that infection and removal are independent in time. This model of disease dynamics belongs to a class of stochastic processes known as nonhomogenous birth–death processes, which, conveniently, turn out to be reasonably tractable. More than 50 years ago, Kendall [3] showed how models for the mean and the variance in the final epidemic size are affected by these parameters. The variance can be interpreted as a measure of the precision with which the final epidemic size can be predicted. Kendall's results can be broken down to show that this quantity is equal to the sum of the average final epidemic size and another quantity (x) minus one. For most realistic epidemiological parameters, this other quantity, which is related to the covariance between final epidemic size and the size of the infected population, will be much greater than one. In these cases, the variance in the final epidemic size will be much greater than the average final epidemic size itself.
This fundamental limit to the precision of forecasts does not imply that EWSs cannot be used effectively to plan a response to outbreaks. Rather, it suggests what expectations of EWSs are reasonable. Further, since the precision with which forecasts of the final epidemic size can be obtained will depend on many disease-specific properties and maybe other factors, too, case studies of the potential effectiveness of EWSs for different diseases are needed. These studies should exploit recent advances in modeling birth–death processes [4] to gain further understanding of the differences among diseases and of the causes of geographic variation in the intensity of epidemics. Finally, notwithstanding limits to precision, the benefits to be obtained from estimates of the average final epidemic size and the timing of epidemics alone may warrant considerable investment in EWSs.
Citation: Drake JM (2005) Fundamental limits to the precision of early warning systems for epidemics of infectious diseases. PLoS Med 2(5): e144.
==== Refs
References
Sultan B Labadi K Guégan JF Janicot S Climate drives the meningitis epidemics onset in West Africa PLoS Med 2005 2 43 49 10.1371/journal.pmed.0020006
World Health Organization Using climate to predict infectious disease outbreaks: A review WHO/SDE/OEH/04.01 2004 Available: http://www.who.int/globalchange/publications/en/oeh0401.pdf . Accessed 14 April 2005
Kendall DG On the generalized “birth-and-death” process Ann Math Stat 1948 19 1 15
Dorman KS Sinsheimer JS Lange K In the garden of branching processes SIAM Rev 2004 46 202 229
| 15916474 | PMC1140954 | CC BY | 2021-01-05 10:39:51 | no | PLoS Med. 2005 May 31; 2(5):e144 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020144 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 10.1371/journal.pmed.0020147Correspondence and Other CommunicationsHIV/AIDSHIV Infection/AIDSMedicine in Developing CountriesAuthors' Reply: Don't Let the Hypothesis Slip CorrespondenceLopman Ben A Garnett Geoff P
1
Gregson Simon
2
Gregson Simon
3
Mason Peter R
4
1Imperial College LondonLondonUnited Kingdom2Imperial College LondonLondonUnited Kingdom3Biomedical Research and Training InstituteHarareZimbabwe4Biomedical Research and Training InstituteHarareZimbabweE-mail: [email protected]
Competing Interests: GPG has acted as a consultant for and/or received grants from GlaxoSmithKline, Aventis Pasteur, Merck, and Abbott Pharmaceuticals. GPG also chaired a meeting of the World Health Organization in 2003 to develop a consensus on the importance of unsafe injections in HIV epidemiology. SG owns shares in GlaxoSmithKline Beecham and Astra Zeneca.
5 2005 31 5 2005 2 5 e147Copyright: © 2005 Lopman 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.
Individual Level Injection History: A Lack of Association with HIV Incidence in Rural Zimbabwe
Association between Injections and HIV Incidence
HIV Epidemiology in Africa: Weak Variables and Tendentiousness Generate Wobbly Conclusions
==== Body
In 2003, Brody and colleagues called for researchers to publish analyses investigating the hypothesised importance of medical injections in the transmission of HIV in Africa [1,2]. Considering the general failure of HIV/AIDS control programs and the neglect of this subject, we believed they were right to raise this controversial hypothesis, so we added a question to our field survey and performed a fresh analysis to test the strength of association between injections and HIV incidence [3]. Therefore, we are disappointed that Brody and Potterat think that we are “transparently invested in dismissing” the hypothesis [2].
Now that pertinent incidence data (where the timing of exposure and event can be determined) have been published for Manicaland, Zimbabwe, and Rakai, Uganda, and have shown a lack of association with injections, we think it is unfair to belittle the difficulty of collecting data and to claim that we have not gone to great lengths to collect high-quality data on sexual behaviour [3,4]. On the contrary, because of the general problems in generating reliable responses to questions about sexual behaviour [5,6], the Manicaland HIV/ STD Prevention study has pioneered the use of informal, confidential voting interview methods [6]. Brody and Potterat state that we found “that sexual behaviour was unrelated to risk of incident HIV”. However, in women, having a history of STD symptoms, having multiple sexual partners, or being widowed/separated/divorced (a proxy that a previous sexual partner died of HIV) were associated with HIV incidence. In men, the associations of HIV and sexual behaviour did not reach statistical significance because of the small number of seroconversions.
It is true that women with one reported sex partner did not have a higher incidence than women with no reported partners. However, Brody and Potterat fail to point out that women with multiple sex partners had the highest incidence (31.3 cases per 1,000 person-years) and that rates were lower in men with no sex partners (3.1) than in those with one sex partner (13.6) or multiple sex partners (14.9). These analyses were performed on only a subset of the Manicaland cohort, but other publications have demonstrated the role of sexual behaviours as risk factors for HIV in this population [7].
We agree that our measure of injections was not perfect, and Brody and Potterat reiterate many of the limitations discussed in our paper: we used a binary (yes/no) measure of exposure, which did not capture the number of injections and had a fairly long follow-up period of three years. These dimensions are being measured in the next round of the cohort study. In the published data, it is possible that some cases had their exposure misclassified, but as many as 40 (60%) of the individuals who seroconverted reported not to have received an injection. Post hoc power calculations (Figure 1) demonstrate that if there was a risk of 2.27 associated with injections, the finding would have been statistically significant. The crude rate ratio for both sexes was 1.1 (95%; confidence interval: 0.7–1.8), which is not evidence that injections are a major transmission route of HIV.
Figure 1 Plot of the Detectable Rate Ratio as a Function of Sample Size
Assuming power of 90% at the 95% significance level [11], a rate ratio of 2.27 or greater would have achieved a significant result given the 5,500 years of observation in the sample:
where u = 1.28, which is the one-sided percentage point of the normal distribution corresponding to 100% minus the power; v = 1.96, which is the percentage point of the normal distribution corresponding to the two-sided significance level; and μ1 and μ2 are the seroconversion rates in exposed and unexposed individuals, respectively.
Brody and Potterat also claim that our statistics are flawed because we controlled for age in the analysis. This is a moot point. We presented both univariable and age-adjusted rate ratios of injection exposure—neither showed an association.
We find it strange that Brody and Potterat reference themselves for a study performed in our “own backyard”, which was actually the baseline survey for our current study, and then mislead by saying that it shows little association between sexual behaviour and HIV risk. Lifetime number of sexual partners was in fact a very strong determinant of HIV status in this population [6,7].
In their separate response, Naveed Zafar Janjua and colleagues point out a number of important aspects concerning injection epidemiology and health care–associated infections [8]. First, there is, by definition, a difference between safe and unsafe injections. Second, heightened risks may be associated with “minor and major surgical procedures, dental instrumentation, and tattooing or other traditional practices involving scarification”. (Although not part of our original report, 16 HIV-negative individuals reported to have received a blood transfusion in the follow-up period. None of them seroconverted.) Third, a “needle prick” is a general term that captures lacerations with solid needles as well as those with a borehole. And fourth, the risk associated with receiving one injection is not the same as that for multiple injections, with certain types of injections carrying more risk than others.
However, these concerns expressed by Janjua et al. are not pertinent to the hypothesis that we were testing: are injections a major route of transmission of HIV in this population in Manicaland Province in Zimbabwe? This analysis was motivated by the arguments of Gisselquist et al. that injections are the main driver of HIV transmission in southern Africa [1]. To be clear—we were not testing whether exposure to contaminated needles is a risk factor (clearly it is), whether certain types of injections carry more risk than others (clearly they do), or whether needles are a driver of the epidemic in certain populations in the world (clearly they are). The fact that there is no evidence of association between receipt of injections (of any number) and HIV incidence, before and after controlling for confounding variables, allows us to conclude that injections “do not play a major role in the transmission of HIV in rural Zimbabwe”[3].
The global HIV problem is not a single epidemic. In eastern Europe, over 50% of HIV infections are among users of injection drugs [9]; in Pakistan, people receive on average eight injections per year compared with about one in sub-Saharan Africa [10]. Our findings apply to Manicaland and may be relevant for similar epidemic patterns in southern Africa. They are not generalisable to all locations, but they do refute the hypothesis that HIV is transmitted through medical injections in the study population.
Citation: Lopman BA, Garnett GP, Mason PR, Gregson S (2005) Authors' reply: Don't let the hypothesis slip. PLos Med 2(5): e147.
==== Refs
References
Gisselquist D Potterat JJ Brody S Vachon F Let it be sexual: How health care transmission of AIDS in Africa was ignored Int J STD AIDS 2003 14 148 161 12665437
Brody S Potterat JJ HIV epidemiology in Africa: Weak variables and tendentiousness generate wobbly conclusions PLoS Med 2005 2 e137 10.1371/journal.pmed.0020137 15916469
Lopman BA Garnett GP Mason PR Gregson S Individual level injection history: A lack of association with HIV incidence in rural Zimbabwe PLoS Med 2005 2 e37 10.1371/journal.pmed.0020037 15736994
Kiwanuka N Gray RH Serwadda D Li X Sewankambo NK The incidence of HIV-1 associated with injections and transfusions in a prospective cohort, Rakai, Uganda AIDS 2004 18 342 344 15075560
Cleland J Boerma JT Carael M Weir SS Monitoring sexual behaviour in general populations: A synthesis of lessons of the past decade Sex Transm Infect 2004 80 Suppl 2 1 7
Gregson S Mushati P White PJ Mlilo M Mundandi C Informal confidential voting interview methods and temporal changes in reported sexual risk behaviour for HIV transmission in sub-Saharan Africa Sex Transm Infect 2004 80 Suppl 2 36 42
Gregson S Nyamukapa CA Garnett GP Mason PR Zhuwau T Sexual mixing patterns and sex-differentials in teenage exposure to HIV infection in rural Zimbabwe Lancet 2002 359 1896 1903 12057552
Janjua NZ Ahmad K Altaf A Khan MI Hamza HB Association between injections and HIV incidence PLoS Med 2005 2 e139 10.1371/journal.pmed.0020139 15916470
Dehne KL Khodakevich L Hamers FF Schwartlander B The HIV/ AIDS epidemic in eastern Europe: Recent patterns and trends and their implications for policy-making AIDS 1999 13 741 749 10357372
Simonsen L Kane A Lloyd J Zaffran M Kane M Unsafe injections in the developing world and transmission of bloodborne pathogens: A review Bull World Health Organ 1999 77 789 800 10593026
Kirkwood BR Sterne JAC Essential medical statistics 2003 Oxford Blackwell Publishing 288
| 0 | PMC1140957 | CC BY | 2021-01-05 11:13:38 | no | PLoS Med. 2005 May 31; 2(5):e147 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020147 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1591646010.1371/journal.pmed.0020151PerspectivesMental HealthSchizophrenia and other psychotic disordersEpidemiologyMedicine in Developing CountriesThe Global Prevalence of Schizophrenia PerspectivesBhugra Dinesh Dinesh Bhugra is Professor of Mental Health and Cultural Diversity, Institute of Psychiatry, Kings College London, United Kingdom. E-mail: [email protected]
Competing Interests: The author declares that he has no competing interests.
5 2005 31 5 2005 2 5 e151Copyright: © 2005 Dinesh Bhugra.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.Bhugra discusses the implications of a new study in PLoS Medicine that challenges widely held assumptions about the epidemiology of schizophrenia.
==== Body
There is no doubt that the global burden of schizophrenia, a chronic serious mental illness, is massive. It is therefore essential that any intervention is appropriate, cost-effective, and efficacious.
To reduce the burden, a clear epidemiologically based dataset is required. The first step in such a venture is provided by Saha et al. in their systematic review of prevalence data on schizophrenia across cultures, published in this month's PLoS Medicine [1]. Using a number of strategies these authors have distilled the findings from just under 200 studies from 46 nations.
Defining the Prevalence of the Disease
The prevalence rates of schizophrenia depend upon a whole range of factors, such as the availability of and response to treatment. The prevalence of schizophrenia, as with other mental disorders, can be calculated from a number of sources—from case registers to field surveys. The latter lend themselves more readily to estimation of period prevalence (see Glossary) than point prevalence, while case register data can provide point prevalence more readily. The denominator can be the whole population or only a small defined population.
Saha et al. quite rightly differentiate between traditional prevalence, or “core”, studies (these generate an estimate based on the population residing within a defined catchment area), and studies in specific sub-groups (which they divide into migrant studies and studies in other special groups). Using sequential filters they were able to isolate discrete data from multiple studies, and they used other strategies to ensure that the largest groups were counted.
The authors had hypothesised that prevalence estimates would differ between lifetime, period, and point prevalence and that lifetime prevalence would be higher than lifetime estimates (estimates are calculated as a proportion by dividing the total number of individuals who manifest a disorder [the numerator] by the total population at risk including those with the disorder [the denominator]). They also predicted that the estimates would be higher for males, those from urban areas, and migrants. Case ascertainment methods and sample selection do influence prevalence rates, so Saha et al. chose studies that used comprehensive case ascertainment methods.
Findings of the Study
Of the 132 core studies, 21 studies reported point prevalence, 34 reported period prevalence, and 24 reported lifetime prevalence. The median prevalence of schizophrenia was 4.6/1,000 for point prevalence, 3.3/1,000 for period prevalence, 4.0 for lifetime prevalence, and 7.2 for lifetime morbid risk.
There were no significant differences between males and females, nor between urban, rural, and mixed sites, although migrants and homeless people had higher rates of schizophrenia and, not surprisingly, developing countries had lower prevalence rates (the lower prevalence of schizophrenia in developing countries has been previously documented). It is well known from other studies that migrants have higher than expected rates of schizophrenia [2–8], although definitions of migrants in these studies have been variable and the studies have suffered from a series of other methodological problems.
Implications of the Study
Several important findings emerge from Saha and colleagues' analysis. For clinicians, the analysis indicates clearly that lifetime prevalence is 4.0/1,000 and not 1%, as reported in the Diagnostic and Statistic Manual of Mental Disorders, fourth edition [9], and other textbooks.
The study also challenges the widely held view that schizophrenia is much more common in men. Saha et al.'s finding that schizophrenia was just as common in women has clear implications for developing services, since it means that not only must we develop and provide culturally appropriate services but also services that are gender sensitive (as the number of cases in women are higher than expected, gender becomes a more important factor). Furthermore, if the prodrome of the illness is long, this will affect the number of new cases appearing in the population, and longer delays in treatment will also affect the rates.
Psychiatric Research by Ted Watson—this painting, representing collaborative research between people with schizophrenia and mental health professionals, is by an aboriginal mental health service user and was commissioned by the Queensland Centre for Mental Health Research, Australia
An important question for the researchers is whether cases counted as positive included patients with positive and/or negative symptoms, those who may have had underlying cognitive deficits, and those whose illness was treatment resistant.
Detailed economic measures must be included in analyses of the prevalence of schizophrenia to determine whether countries are developed or developing. Saha et al. themselves acknowledge that they have used a single measure of World Bank definitions relying on per capita income (whereas in any country there will be marked geographical variation in social and economic classes) for assessing a complex and multi-dimensional concept, which is a weak point of their systematic review. The impact of urbanisation must be studied especially, as in many developing or low-income countries the migration into urban areas adds a tremendous amount of variation that must be taken into account in future ecological studies.
The authors acknowledge that their systematic review may have missed some studies, and they encourage readers who know about missing studies to contact them.
Schizophrenia across Cultures
Studies have shown that the outcome of schizophrenia is better in developing countries [10–12], and therefore the point prevalence in these countries should be lower. Despite this clear difference in the course of schizophrenia in different cultures, cross-cultural research in psychiatry focuses on similarities rather than differences. Kleinman suggests that there is a very strong bias towards discovering universals in mental disorder [13]. Both the International Pilot Study of Schizophrenia [10] and the Determinants of Outcome of Severe Mental Disorders study [12] used a template of symptoms of psychosis across cultures to identify groups of patients who seemed similar, but these studies left out all those patients who failed to fit the template. It is these excluded patients that Kleinman suggests are of greater interest from a cultural perspective simply because they are the ones who would reveal the greatest amount of cultural diversity.
In the International Pilot Study of Schizophrenia [10] and the Determinants of Outcome of Severe Mental Disorders study [12], catatonia (a form of schizophrenia characterized by a tendency to remain in a fixed stuporous state for long periods) was diagnosed in 10% of cases in developing countries compared with less than 1% in developed countries. Hebephrenia (a form of schizophrenia characterized by severe disintegration of personality) was present in 13% of cases in developed countries and 4% in developing countries. These differences in the disease in developed versus developing countries indicate that there is more to the prevalence of schizophrenia than simple epidemiological data. Better prognosis in developing countries may indicate different sets of aetiological and perpetuating factors [14].
Cohen [15] argued that although the case-finding method in both these studies was accurate, the vast majority of cases were identified in Western-type facilities, and therefore the numbers of true cases of schizophrenia may be an underestimate. He also pointed out that in developing countries the proportion of cases with acute onset schizophrenia was twice as high as in developed countries. Such variations may indicate a real difference in the cross-cultural manifestations of schizophrenia—hospital-based data collection reflects cultural processes that have little to do with the true prevalence and incidence rates of schizophrenia.
While prevalence studies can help contribute to an understanding of the aetiology of schizophrenia, psychodynamic issues—such as cultural identity and attachment—must also be studied, especially among migrant groups, as cultural congruity and ethnic density may influence the presentation of suffering individuals to psychiatric services [16,17].
Glossary
Point prevalence: The proportion of individuals who manifest a disorder at a given point in time.
Period prevalence: The proportion of individuals who manifest a disorder over a specific period of time (e.g., over one year).
Lifetime prevalence: The proportion of individuals in the population who have ever manifested a disorder, who are alive on a given day.
Lifetime morbid risk: The probability of a person developing the disorder during a specified period of their life or up to a specified age (lifetime morbid risk differs from lifetime prevalence in that it attempts to include the entire lifetime of a birth cohort both past and future, and includes those deceased at the time of the survey).
Citation: Bhugra D (2005) The global prevalence of schizophrenia. PLoS Med 2(5): e151.
==== Refs
References
Saha S Chant D Welham J McGrath J A systematic review of the prevalence of schizophrenia PLoS Med 2005 2 e141 15916472
Bagley C Mental illness in immigrant communities in London J Biosoc Sci 1971 3 449 459 5129300
Bebbington P Hurry J Tennant C Psychiatric disorders in selected immigrant groups in Camberwell Soc Psychiatry 1981 16 43 51
Bhugra D Leff J Mallet M Der G Corridan B Incidence and outcome of schizophrenia in whites, African-Caribbeans and Asians in London Psych Med 1987 27 791 798
Birchwood M Cochrane R MacMillan F Copestake S Kucharska J The influence of ethnicity and family structure on relapse Br J Psychiatry 1992 161 783 793 1483164
Cochrane R Mental illness in immigrants in England and Wales Soc Psychiatry 1977 12 25 35
Harrison G Glazebrook C Brewin J Cantwell R Dalkin T Increased incidence of psychotic disorders in migrants from the Caribbean in the UK Psych Med 1997 27 799 807
Callan AF Schizophrenia in Afro-Caribbean immigrants J R Soc Med 1996 89 253 256 8778431
American Psychiatric Association Diagnostic and statistical manual of mental disorders: DSM-IV, 4th ed 1994 Washington (DC) American Psychiatric Association 886
World Health Organization International pilot study of schizophrenia 1973 Geneva World Health Organization 427
World Health Organization Schizophrenia: an international follow-up study 1979 Chichester (United Kingdom) Wiley 438
Jablensky A Sartorius N Ernberg G Anker M Korten A Schizophrenia: manifestations, incidence and course in different cultures. A World Health Organization ten-country study Psychol Med Monogr Suppl 1992 20 1 97 1565705
Kleinman A Anthropology and psychiatry. The role of culture in cross-cultural research on illness Br J Psychiatry 1987 151 447 454 3447661
Karuo M Jenkins J Cross cultural issues in the course and treatment of schizophrenia Psychiatr Clin North Am 1996 16 339 350
Cohen A Prognosis for schizophrenia in the Third World: A re-evaluation of cross-cultural research Cult Med Psychiatry 1992 16 53 75 1537236
Bhugra D Migration, distress and cultural identity Br Med Bull 2004 69 129 142 15226202
Bhugra D Cultural identities and cultural congruency: A new model for evaluating mental distress in immigrants Acta Psychiatr Scand 2005 111 84 93 15667427
| 15916460 | PMC1140960 | CC BY | 2021-01-05 11:13:38 | no | PLoS Med. 2005 May 31; 2(5):e151 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020151 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 10.1371/journal.pmed.0020152Correspondence and Other CommunicationsOtherMedical JournalsEditorial Policies (Including Conflicts of Interest)Editors' Reply CorrespondenceBarbour Virginia Cohen Barbara Yamey Gavin E-mail: [email protected] 2005 31 5 2005 2 5 e152Copyright: © 2005 Barbour 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.
Unjustified Restrictions on Letters to the Editor
==== Body
We agree with Doug Altman that correspondence is an essential part of medical publishing—acting as post-publication peer review. Hence, we encourage submission of letters in response to articles we publish, preferably as e-letters via our Web site (www.plosmedicine.org) rather than via our manuscript submission site. The four-week limit we originally suggested in our author guidelines was to encourage timely correspondence; however, we accept that there will be occasions when it will be appropriate for letters to be published later than that, and we will accept such submissions. The word limit is more debatable. We think that 750 words is a reasonable limit; any more, and letters are likely to stray off the topic.
Citation: Barbour V, Cohen B, Yamey G (2005) Editors' reply. PLoS Med 2(5): e152.
| 0 | PMC1140961 | CC BY | 2021-01-05 11:13:39 | no | PLoS Med. 2005 May 31; 2(5):e152 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020152 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1752608010.1371/journal.pmed.0020158EditorialOtherMedical JournalsEditorial Policies (Including Conflicts of Interest)Meeting the Needs of Different Research Communities EditorialThe PLoS Medicine Editors 5 2005 31 5 2005 2 5 e158Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Public Domain declaration, which stipulates that, once placed in the public domain, this work may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose.The Public Library of Science is launching three new journals in 2005, led by academics and aimed at the needs of different research communities.
PLoS launches three new community journals in 2005
==== Body
PLoS was founded in October 2000 not as a publisher, but as a grassroots movement of researchers who believed that the results of the global research enterprise should be a freely available public resource. So why did PLoS become a publisher, and what are we doing in this role to serve the world's varying biomedical researchers and clinicians?
At its inception, the first action by PLoS was to encourage scientific and medical journal publishers to make the archival research literature freely available. An open letter, signed by almost 34,000 scientists from 180 countries, urged publishers to deposit copies of their research articles in a full text public repository, such as PubMed Central, within six months of publication. Sadly, the vast majority of publishers declined to deposit their works—a depressing situation that continues to this day.
We concluded that the only way forward was to publish our own journals. These would provide an alternative, open-access venue for important discoveries in science and medicine and would serve as a model for showing that open-access publication is viable.
The first phase of our life as a publisher involved launching our two flagship journals—PLoS Biology in October 2003 and PLoS Medicine in October 2004. These two journals provide an open-access alternative to the best subscription journals in the life sciences and medicine, respectively.
These first journals have helped to put open access on the map. In the 18 months since we started as a publisher not only have many researchers embraced these fledgling journals as a top tier “home” for their work, but also there has been growing international support for open-access initiatives.
For example, 55 institutions worldwide have so far signed the Berlin Declaration on Open Access to Knowledge in the Science and Humanities. Beginning this month, researchers who are funded by the United States National Institutes of Health are being asked by the agency to deposit a copy of their accepted research papers into PubMed Central. And in the United Kingdom, the Wellcome Trust is making it a requirement of its grant conditions that Wellcome Trust–funded researchers deposit an electronic version of their manuscripts in a UK portal of PubMed Central within six months of publication.
PLoS is now entering its second phase as a publisher, in which we launch the first three PLoS community journals. The case for launching these journals was compelling—we wanted to serve three research communities that had few open-access alternatives to the subscription journals in their field. As a result, we are launching PLoS Computational Biology (www.ploscompbiol.org), a collaboration between PLoS and the International Society for Computational Biology, scheduled to start publishing in June 2005; PLoS Genetics (www.plosgenetics.org), scheduled for July 2005; and PLoS Pathogens (www.plospathogens.org), scheduled for September 2005.
With the arrival of the community journals, we are providing a greater range of open-access venues for researchers who wish to ensure that anyone can read, use, and build on their work. Unlike PLoS Biology and PLoS Medicine, which are run by PLoS editorial staff, each community journal is run by the community itself—that is, by an academic editor-in-chief and editorial board, with production support from PLoS staff. The community-led nature of the new journals, coupled with a business model in which publication costs are borne largely by publication charges, provides an example for other journals that wish to transition to open access.
When submitting their work to PLoS, how do researchers and clinicians distinguish between PLoS Medicine and the PLoS community journals? PLoS Medicine remains committed to publishing the best medical research that is relevant to a broad international community of clinicians and researchers. The community journals are aimed at a more specialized audience. PLoS Pathogens, for example, will publish basic scientific research that “significantly advances the understanding of pathogens and how they interact with their host organisms.” But we also think that PLoS Medicine readers will find much of interest in the community journals, and vice versa, and we will cross-link between articles in the different journals.
The different PLoS journals will be editorially independent and submissions will remain confidential to each journal. But if an author would like a manuscript that is not thought to be appropriate by one journal to be passed on to another, along with the reviewers' reports and their identities, we are happy to cooperate, subject to the permission of the reviewers. This can help to speed up the review process.
As we roll out the new community journals, we are already planning the third phase of our life as a publisher—the creation of an online repository for all technically sound research reports in both biology and medicine, including clinical trial reports. It is an ambitious plan, but one that we believe will provide authors with more choices and in the end, an open-access venue for the widest range of research.
| 17526080 | PMC1140962 | CC0 | 2021-01-05 11:13:39 | no | PLoS Med. 2005 May 31; 2(5):e158 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020158 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1752608010.1371/journal.pmed.0020158EditorialOtherMedical JournalsEditorial Policies (Including Conflicts of Interest)Meeting the Needs of Different Research Communities EditorialThe PLoS Medicine Editors 5 2005 31 5 2005 2 5 e158Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Public Domain declaration, which stipulates that, once placed in the public domain, this work may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose.The Public Library of Science is launching three new journals in 2005, led by academics and aimed at the needs of different research communities.
PLoS launches three new community journals in 2005
==== Body
PLoS was founded in October 2000 not as a publisher, but as a grassroots movement of researchers who believed that the results of the global research enterprise should be a freely available public resource. So why did PLoS become a publisher, and what are we doing in this role to serve the world's varying biomedical researchers and clinicians?
At its inception, the first action by PLoS was to encourage scientific and medical journal publishers to make the archival research literature freely available. An open letter, signed by almost 34,000 scientists from 180 countries, urged publishers to deposit copies of their research articles in a full text public repository, such as PubMed Central, within six months of publication. Sadly, the vast majority of publishers declined to deposit their works—a depressing situation that continues to this day.
We concluded that the only way forward was to publish our own journals. These would provide an alternative, open-access venue for important discoveries in science and medicine and would serve as a model for showing that open-access publication is viable.
The first phase of our life as a publisher involved launching our two flagship journals—PLoS Biology in October 2003 and PLoS Medicine in October 2004. These two journals provide an open-access alternative to the best subscription journals in the life sciences and medicine, respectively.
These first journals have helped to put open access on the map. In the 18 months since we started as a publisher not only have many researchers embraced these fledgling journals as a top tier “home” for their work, but also there has been growing international support for open-access initiatives.
For example, 55 institutions worldwide have so far signed the Berlin Declaration on Open Access to Knowledge in the Science and Humanities. Beginning this month, researchers who are funded by the United States National Institutes of Health are being asked by the agency to deposit a copy of their accepted research papers into PubMed Central. And in the United Kingdom, the Wellcome Trust is making it a requirement of its grant conditions that Wellcome Trust–funded researchers deposit an electronic version of their manuscripts in a UK portal of PubMed Central within six months of publication.
PLoS is now entering its second phase as a publisher, in which we launch the first three PLoS community journals. The case for launching these journals was compelling—we wanted to serve three research communities that had few open-access alternatives to the subscription journals in their field. As a result, we are launching PLoS Computational Biology (www.ploscompbiol.org), a collaboration between PLoS and the International Society for Computational Biology, scheduled to start publishing in June 2005; PLoS Genetics (www.plosgenetics.org), scheduled for July 2005; and PLoS Pathogens (www.plospathogens.org), scheduled for September 2005.
With the arrival of the community journals, we are providing a greater range of open-access venues for researchers who wish to ensure that anyone can read, use, and build on their work. Unlike PLoS Biology and PLoS Medicine, which are run by PLoS editorial staff, each community journal is run by the community itself—that is, by an academic editor-in-chief and editorial board, with production support from PLoS staff. The community-led nature of the new journals, coupled with a business model in which publication costs are borne largely by publication charges, provides an example for other journals that wish to transition to open access.
When submitting their work to PLoS, how do researchers and clinicians distinguish between PLoS Medicine and the PLoS community journals? PLoS Medicine remains committed to publishing the best medical research that is relevant to a broad international community of clinicians and researchers. The community journals are aimed at a more specialized audience. PLoS Pathogens, for example, will publish basic scientific research that “significantly advances the understanding of pathogens and how they interact with their host organisms.” But we also think that PLoS Medicine readers will find much of interest in the community journals, and vice versa, and we will cross-link between articles in the different journals.
The different PLoS journals will be editorially independent and submissions will remain confidential to each journal. But if an author would like a manuscript that is not thought to be appropriate by one journal to be passed on to another, along with the reviewers' reports and their identities, we are happy to cooperate, subject to the permission of the reviewers. This can help to speed up the review process.
As we roll out the new community journals, we are already planning the third phase of our life as a publisher—the creation of an online repository for all technically sound research reports in both biology and medicine, including clinical trial reports. It is an ambitious plan, but one that we believe will provide authors with more choices and in the end, an open-access venue for the widest range of research.
| 0 | PMC1140963 | CC0 | 2021-01-05 10:39:51 | no | PLoS Med. 2005 May 31; 2(5):e173 | latin-1 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020173 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030244SynopsisBioinformatics/Computational BiologyCell BiologyDevelopmentGenetics/Genomics/Gene TherapyMolecular Biology/Structural BiologyBiochemistryDrosophilamiRNA Processing: Dicer-1 Meets Its Match Synopsis7 2005 24 5 2005 24 5 2005 3 7 e244Copyright: © 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.
Processing of Pre-microRNAs by the Dicer-1-Loquacious Complex in Drosophila Cells
Normal microRNA Maturation and Germ-Line Stem Cell Maintenance Requires Loquacious, a Double-Stranded RNA-Binding Domain Protein
==== Body
In recent years, the control of gene expression by small RNA molecules has emerged as a major new mechanism for gene regulation. The small RNAs interfere with the expression of their target gene by reducing its transcription, triggering the destruction of the gene transcript, or inhibiting its translation into a protein. This discovery has not only altered views of gene regulation, but also provided molecular geneticists with powerful new tools with which to study and manipulate the function of any gene. The biology of these small RNAs is, therefore, under intense scrutiny.
Small RNAs are generated by specific pathways, the elements of which are being rapidly discovered. In this issue of PLoS Biology, two groups have identified a missing piece in one such pathway—in the fruitfly Drosophila. The pathway under investigation leads to the production of a type of small RNA called a microRNA (miRNA). These are 21–23 nucleotides in length, and are involved in regulating the expression of many genes. miRNAs start life as a much bigger transcript called a pri-miRNA, which is processed in two steps. First, it is converted into a shorter pre-miRNA, by the action of two proteins: Drosha, an RNAse III enzyme; and Pasha, which contains double-stranded RNA binding domains (dsRBDs). The pre-miRNA is then transported to the cytoplasm and is trimmed again into a double-stranded miRNA by a different RNAse III enzyme called Dicer-1.
In a separate pathway, RNAs called small interfering RNAs (siRNAs) depend on the Dicer-2 RNAse III and a dsRBD protein called R2D2 for their function. These pathways are also conserved in other organisms. Thus, a pattern emerges: the functions of small RNAs tend to require the combined actions of an RNAse III and a dsRBD protein. But why doesn't Dicer-1 have a partner? The answer, provided by the two studies from the labs of Phil Zamore and Haruhiko and Mikiko Siomi, is that we just hadn't found it yet.
The two groups took different approaches to finding Dicer-1's partner. Zamore's group looked for genes resembling other dsRBD-encoding genes, while the Siomi lab did a functional screen for new genes specifically implicated in miRNA processing. They both homed in on a new gene with great similarity to R2D2, and showed that loss of function of the gene results in the accumulation of pre-miRNAs—very similar to loss of Dicer-1 function, which suggests that the two genes act together in the same pathway. The new potential partner of Dicer-1 was given the name loquacious (loqs), because failure to process the miRNAs in turn causes increased levels of expression of the target genes for the miRNAs.
Both groups also show that Loqs and Dicer-1 exist in a complex within the cell, and that the complex is able to process pre-miRNA into its mature form. The Siomis' lab went on to show that the complex contains a protein called Ago-1, which hints that the complex might also be involved in the action of miRNAs on their target genes, as well as in miRNA processing itself. Both groups also point out the similarity between Loqs and a human dsRBD protein called TRBP, which has been implicated in the response to infection by HIV.
There seems little doubt, then, that Dicer-1's partner has been found, and that the combined action of an RNAse III and a dsRBD protein is a consistent theme in the function of miRNAs and siRNAs. The identification of Loqs will help to refine our views of how miRNAs are processed, as well as how they can be manipulated. The connections made with processes such as stem cell maintenance (identified by the Zamore lab) and viral infection in these new studies also emphasize that gene regulation by small RNAs is relevant to a broad range of cellular physiology.
| 0 | PMC1141266 | CC BY | 2021-01-05 08:28:14 | no | PLoS Biol. 2005 Jul 24; 3(7):e244 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030244 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1591877010.1371/journal.pbio.0030236Research ArticleBioinformatics/Computational BiologyCell BiologyDevelopmentGenetics/Genomics/Gene TherapyMolecular Biology/Structural BiologyBiochemistryDrosophilaNormal microRNA Maturation and Germ-Line Stem Cell Maintenance Requires Loquacious, a Double-Stranded RNA-Binding Domain Protein Loquacious, Partner of Drosophila Dicer-1Förstemann Klaus
1
Tomari Yukihide
1
Du Tingting
1
Vagin Vasily V
2
Denli Ahmet M
3
Bratu Diana P
4
Klattenhoff Carla
4
Theurkauf William E
4
Zamore Phillip D [email protected]
1
1 Department of Biochemistry and Molecular Pharmacology, University of Massachusetts Medical School, Worcester, Massachusetts, United States of America,2 Institute of Molecular Genetics of RAS, Moscow, Russia,3 Watson School of Biological Sciences, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, United States of America,4 Program in Molecular Medicine, University of Massachusetts Medical School, Worcester, Massachusetts, United States of America,Carrington James C. Academic EditorOregon State UniversityUnited States of America7 2005 24 5 2005 24 5 2005 3 7 e23614 3 2005 30 4 2005 Copyright: © 2005 Förstemann 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.
Processing of Pre-microRNAs by the Dicer-1-Loquacious Complex in Drosophila Cells
miRNA processing: dicer-1 meets its match
microRNAs (miRNAs) are single-stranded, 21- to 23-nucleotide cellular RNAs that control the expression of cognate target genes. Primary miRNA (pri-miRNA) transcripts are transformed to mature miRNA by the successive actions of two RNase III endonucleases. Drosha converts pri-miRNA transcripts to precursor miRNA (pre-miRNA); Dicer, in turn, converts pre-miRNA to mature miRNA. Here, we show that normal processing of Drosophila pre-miRNAs by Dicer-1 requires the double-stranded RNA-binding domain (dsRBD) protein Loquacious (Loqs), a homolog of human TRBP, a protein first identified as binding the HIV trans-activator RNA (TAR). Efficient miRNA-directed silencing of a reporter transgene, complete repression of white by a dsRNA trigger, and silencing of the endogenous Stellate locus by Suppressor of Stellate, all require Loqs. In loqs
f00791 mutant ovaries, germ-line stem cells are not appropriately maintained. Loqs associates with Dcr-1, the Drosophila RNase III enzyme that processes pre-miRNA into mature miRNA. Thus, every known Drosophila RNase-III endonuclease is paired with a dsRBD protein that facilitates its function in small RNA biogenesis.
This and an accompanying paper by Saito et al. identify Loquacious, which encodes a double-stranded RNA binding domain protein, and partners with Dicer-1 in the processing of microRNAs.
==== Body
Introduction
MicroRNAs (miRNAs) are 21- to 23-nucleotide single-stranded RNAs that are encoded in the chromosomal DNA and repress cognate mRNA targets [1,2]. They are transcribed as long, hairpin-containing precursors [3] by RNA polymerase II [4–8] and processed in the nucleus by the multidomain RNase III endonuclease Drosha [9]. Drosha is assisted by its double-stranded RNA-binding domain (dsRBD) protein partner, known as Pasha in Drosophila
melanogaster [10] and DGCR8 in humans [11–13]. Exportin-5 (Ranbp21 in Drosophila) binds the resulting precursor miRNA (pre-miRNA)—likely recognizing the approximately two-nucleotide 3′ overhanging ends characteristic of these approximately 70-nucleotide hairpin structures—and transports them to the cytoplasm via the Ran-GDP–Ran-GTP transport system [14–16]. In the cytoplasm, a second RNase III endonuclease, Dicer, converts pre-miRNA into mature miRNA [17–20].
In Drosophila, two Dicer paralogs define parallel pathways for small RNA biogenesis. Dicer-1 (Dcr-1) liberates miRNA from pre-miRNA, whereas Dicer-2 (Dcr-2) excises small interfering RNA (siRNA) from long double-stranded RNA (dsRNA) [21–23]. Like Drosha, Drosophila Dcr-2 requires a dsRBD partner protein, R2D2, for its function in siRNA biogenesis. Unlike Drosha, Dcr-2 suffices to process its substrate. However, without R2D2, Dcr-2 cannot load the siRNAs it produces into the RNA-induced silencing complex (RISC), the RNA interference (RNAi) effector complex [21,24,25]. Although born as small RNA duplexes, both siRNA and miRNA function in RISC as single-stranded RNA guides for members of the Argonaute family of proteins [26–28]. Among the five Drosophila Argonaute proteins, two—Ago1 and Ago2—are required for small RNA-directed cleavage of target RNAs [29]. In fact, the Piwi domain of Argonaute proteins is a structural homolog of the endoribonuclease RNase H, an enzyme that cleaves the RNA strand of DNA–RNA hybrids [30–33]. Both human and Drosophila Ago2 and Drosophila Ago1 provide the Mg2+-dependent catalytic subunit of RISC [29,31,34,35,36].
The Drosophila genome encodes three RNase III endonucleases, all of which act in miRNA or siRNA biogenesis. Whereas both Drosha and Dcr-2 associate with dsRBD proteins that facilitate their functions, no dsRBD protein partner has been assigned to Dcr-1. We asked if Dcr-1 might also partner with a dsRBD protein. Here, we identify the dsRBD protein Loquacious (Loqs), a paralog of R2D2, as the partner of Dcr-1. Mutation of loqs in flies and depletion of loqs in Schneider-2 (S2) cells by dsRNA-triggered RNAi disrupt normal pre-miRNA processing. In vivo, loqs is required for robust miRNA-directed silencing and complete target gene repression directed by a transgene expressing dsRNA. Moreover, loss of Loqs function in the ovary disrupts germ-line stem cell maintenance, rendering loqs mutant females sterile.
Results
To identify a dsRBD protein partner for Dcr-1, we searched the conserved domain database [37] for all Drosophila proteins that contain dsRBDs. The protein encoded by the gene CG6866 has two dsRBDs, which are most closely related to dsRBD 1 and 2 of R2D2, suggesting that the two genes are paralogs (Figure 1A). CG6866 and R2D2 are 37% similar and 25% identical in the region of the two dsRBDs. A third dsRBD at the C-terminus of CG6866 was detected using the PFam collection of protein sequence motifs. This truncated domain deviates from the canonical dsRBD sequence. Because loss of CG6866 function de-silences both endogenous silencing and reporter expression in vivo (below), we named the gene loquacious (loqs).
loqs is located on the left arm of Chromosome 2 at polytene band 34B9. loqs produces at least three different mRNA isoforms through alternative splicing (Figure 1B). The shortest transcript, loqs RNA splice variant A (RA), encodes a 419-amino-acid protein, Loqs protein isoform A (PA), with a predicted molecular mass of 45 kDa. The transcript loqs RNA splice variant B (RB) contains one additional exon and encodes a protein of 465 amino acids, Loqs protein isoform B (PB), with a predicted molecular mass of 50 kDa. These two mRNA species were identified as cDNAs in the Drosophila genome sequencing project and annotated in FlyBase [38] among the Drosophila proteins that contain dsRBDs. Using non-quantitative RT-PCR, we detected a third splice variant, loqs RNA splice variant C (RC), in which an alternative splice acceptor site for exon 4 is used (Figure 1B, C, and D). Use of the alternative splice site creates a 5′-extended fourth exon and changes the reading frame, resulting in a truncated protein, Loqs protein isoform C (PC), 383 amino acids long (Figure 1C). Loqs PC has a predicted molecular mass of 41 kDa and lacks the entire third dsRBD of Loqs PA and PB (Figure 1B). loqs RA is the predominant mRNA species in dissected testes, whereas loqs RB is the most abundant species in ovaries. Both isoforms are expressed in the carcasses of males and females after removal of the gonads (Figure 1D and data not shown). Using two independent antibodies raised against an N-terminal Loqs peptide, but not using pre-immune sera, we detected a candidate protein for Loqs PC in S2 cells (see below), suggesting that the three loqs transcripts give rise to distinct Loqs protein isoforms.
Figure 1 Loqs, a dsRBD Partner Protein for Drosophila Dcr-1
(A) Each of the three D.
melanogaster RNase III endonucleases pairs with a different dsRBD protein, which assists in its function in RNA silencing.
(B) Differential splicing creates three loqs mRNA variants, loqs RA, RB, and RC. loqs RA and RB are reported in FlyBase. The RC splice variant is reported here. Arrows mark the position of the PCR primers used in (D); green lines, start codons; red lines, stop codons. The resulting protein isoforms are diagrammed to the right.
(C) Use of an alternative splice acceptor site extends the 5′ end of exon 4. The mRNA sequence surrounding the new exon–exon junction is shown, with the loqs RC-specific sequence in bold; the arrow marks the position of the last nucleotide of exon 3 relative to the putative transcription start site. When translated into protein, the exon 4 extension inserts 43 new amino acids (indicated below the mRNA sequence) and shifts the Loqs PC reading frame, truncating the protein.
(D) RT-PCR analysis of loqs mRNA species in males, female carcasses remaining after ovary dissection, dissected ovaries, and S2 cells. Males express more loqs RA than loqs RB, female somatic tissue expresses both loqs RA and loqs RB, while ovaries express predominantly loqs RB. loqs RC was observed only in S2 cells, together with loqs RA and loqs RB.
(E) The piggyBac transposon insertion f00791 lies 57 bp upstream of the reported transcription start site for loqs.
Thibault and co-workers reported a mutant allele of CG6866, loqs
f00791, recovered in a large-scale piggyBac transposon mutagenesis screen of Drosophila [39]. The f00791 piggyBac inserted 57 nucleotides upstream of the loqs transcription start site (Figure 1E); although annotated as lethal, homozygous mutant loqs
f00791 flies are viable but completely female sterile. Precise excision of the f00791 piggyBac transposon fully reverted the female sterility (data not shown). Analysis by quantitative RT-PCR using primers that amplify all three loqs mRNA splice variants (see Materials and Methods) showed that somatic female loqs
f00791 tissues express approximately 5-fold (4.76 ± 0.24; n = 3) less loqs mRNA than wild-type, while loqs
f00791 mutant ovaries express approximately 40-fold (42 ± 0.33; n = 3) less loqs mRNA than wild-type ovaries. Testes express approximately 3-fold (2.9 ± 0.5; n = 3) less loqs mRNA in the loqs
f00791 mutant than in wild type. These data suggest that the mutant phenotype should be strongest in ovaries, consistent with the mutation causing female sterility as its most obvious defect.
In Vivo, Normal Pre-miRNA Processing Requires Loqs
To assess the function of loqs in miRNA biogenesis, we isolated total RNA from loqs
f00791 males and determined the steady-state levels of mature and pre-miRNA for miR-277 and bantam (Figure 2A), which are both expressed in adult tissues. We detected a 100-fold increase in pre-miR-277 and a 12-fold increase in pre-bantam RNAs in homozygous mutant loqs
f00791 males, but not in heterozygous loqs
f00791 or heterozygous or homozygous r2d2 mutant males. In contrast, the amount of mature miR-277 or bantam was only slightly reduced in the loqs
f00791 homozygotes.
Figure 2 Loss of Loqs Function Increases the Steady-State Concentration of Pre-miRNA
(A) Northern analysis of total RNA from wild-type, loqs
f00791 heterozygotes and homozygotes, and r2d2 heterozygotes and homozygotes for whole males, probed for miR-277 and bantam. The membrane was first hybridized with the miR-277 probe, stripped and probed for 2S rRNA as a loading control, then stripped again and probed for bantam miRNA. Asterisk: the 2S probe was not completely removed before the hybridization with the bantam probe, resulting in an additional band above the mature bantam RNA.
(B) Total RNA from whole males, female carcasses remaining after ovary dissection, and dissected ovaries was probed for miR-7. As a control for successful dissection, the blot was also probed for miR-277, which is not expressed in ovaries (KF and PDZ, unpublished results). 2S rRNA again served as a loading control.
(C) Depletion of dcr-1 or loqs in S2 cells by RNAi leads to pre-miRNA accumulation. Total RNA was isolated after dsRNA-triggered RNAi of the indicated genes. The control sample was treated with dsRNA corresponding to the polylinker sequence of pLitmus28i.
(D) Depletion of Dcr-1, Dcr-2, Loqs, and Drosha was confirmed by Western blotting.
(E) Western blotting analysis demonstrates that Dcr-1 levels are not significantly reduced by depletion of Loqs by RNAi in S2 cells, but are lower in loqs
f00791 mutant ovaries.
Since loqs mRNA expression is lowest in the ovaries of loqs
f00791 mutant flies, we analyzed the levels of pre-miR-7 and mature miR-7, a miRNA that is expressed in whole males, manually dissected ovaries, and the female carcasses remaining after removing the ovaries (Figure 2B). While pre-miR-7 increased in all loqs
f00791 homozygous mutant tissues, relative to wild-type or loqs heterozygotes, the disruption of miR-7 production in ovaries was striking: not only did pre-miR-7 accumulate, but also mature miR-7 was dramatically reduced. These data suggest that Loqs protein function is required for the maturation of miRNA and demonstrate a direct correlation between loqs mutant allele strength and disruption of miRNA processing.
Loqs Is Required for Pre-miRNA Processing in Drosophila S2 Cells
To confirm the function of loqs in pre-miRNA processing, we depleted cultured Drosophila S2 cells of loqs mRNA by RNAi (Figure 2C). Eight days after incubating S2 cells with dsRNA corresponding to the first 300 nucleotides of the loqs coding sequence, we determined the steady-state levels of pre-miRNA and mature miRNA for miR-277 and bantam. Relative to an unrelated dsRNA control, dsRNA corresponding to dcr-1 caused a approximately 9-fold and approximately 23-fold increase in steady-state pre-miR-277 and bantam levels, respectively, and dsRNA corresponding to loqs caused a approximately 2-fold and approximately 6-fold increase in steady-state pre-miR-277 and bantam levels, respectively. In these experiments, RNAi of dcr-1 more completely depleted Dcr-1 protein than RNAi of loqs reduced Loqs protein (Figure 2D). RNAi of dcr-2,
r2d2, or drosha did not alter pre-miRNA levels for either miR-277 or bantam, nor did it alter Dcr-1 or Loqs levels. The Drosha/Pasha protein complex functions before pre-miRNA processing, converting primary miRNA (pri-miRNA) to pre-miRNA. Consistent with the idea that Loqs functions with Dcr-1 to convert pre-miRNA to mature miRNA, RNAi of drosha together with loqs alleviated the high pre-miRNA levels observed for RNAi of loqs alone, demonstrating that Loqs acts after Drosha.
Next, we examined processing of 20 nM exogenous pre-let-7 into mature let-7 in lysates from ovaries or S2 cells (Figure 3). Initial velocities were calculated for each reaction to permit comparison of processing rates (see Materials and Methods). Lysate from homozygous loqs
f00791 mutant ovaries processed pre-let-7 RNA to mature let-7 approximately 19-fold more slowly than wild-type ovary lysate (Figure 3A). Moreover, lysate prepared from S2 cells soaked with a green fluoresecent protein (GFP) control dsRNA (GFP[RNAi]) or drosha dsRNA (drosha[RNAi]) accurately and efficiently converted exogenous pre-let-7 RNA into mature let-7. In contrast, bothdcr-1(RNAi) and loqs(RNAi) S2 cell lysates converted pre-miRNA to mature miRNA approximately 5- and approximately 4-fold, respectively, more slowly than the control lysate (Figure 3B). Thus, Loqs is required for production in vivo of normal levels of miR-7, miR-277, and bantam, and the efficient conversion of pre-let-7 to mature let-7 in vitro. Together, these four miRNAs include both miRNAs found on the 5′ and on the 3′ side of the pre-miRNA stem, suggesting a general role for Loqs in pre-miRNA processing.
Figure 3 Loqs Is Required for Efficient pre-let-7 Processing In Vitro
(A) loqs
f00791 mutant ovary lysates processed pre-let-7 into mature let-7 miRNA ∼19-fold more slowly than wild-type. The data were fit to a first-order exponential equation, and initial velocities calculated from the fitted curve.
(B) Analysis of pre-let-7 processing in extracts from S2 cells. The cells were treated twice with dsRNA corresponding to the indicated genes.
Reduction of R2D2 protein by RNAi destabilizes Dcr-2; conversely, RNAi of Dcr-2 renders R2D2 unstable [21]. In contrast, RNAi of loqs in S2 cells reduced Dcr-1 protein levels by no more than 15% (Figure 2D and E), suggesting that Loqs functions together with Dcr-1 in pre-miRNA processing, rather than that Loqs is simply needed to stabilize Dcr-1 protein. However, loqs
f00791 mutant ovaries, which lack detectable Loqs protein, contain 70% less Dcr-1 than wild-type (Figure 2E). A role for Loqs in both Dcr-1 function and in Dcr-1 stability suggests that the two proteins physically interact, like R2D2 and Dcr-2. Therefore, we tested if Dcr-1 and Loqs are components of a common complex.
A Dcr-1 Protein Complex Contains Loqs
We expressed in S2 cells myc-tagged versions for two protein isoforms of Loqs, Loqs PA and Loqs PB, and immunoprecipitated the tagged proteins with anti-myc monoclonal antibodies. We analyzed the immunoprecipitated protein by Western blotting using a polyclonal anti-Dcr-1 antibody. Figure 4A shows that Dcr-1 protein co-immunoprecipitated with myc-tagged Loqs. When myc-tagged GFP was expressed in place of myc-tagged Loqs, no Dcr-1 protein was recovered in the anti-myc immunoprecipitate. Similarly, an affinity purified, polyclonal antibody directed against the N-terminus of endogenous Loqs protein also co-immunoprecipitated Dcr-1 protein (Figure 4A). This interaction was resistant to treatment with RNase A (data not shown). We could not detect co-immunoprecipitation of Dcr-2 with myc-tagged Loqs PB under conditions where Dcr-1 was readily detected (Figure 4B), but we cannot exclude that Dcr-2 is a substoichiometric component of a complex that contains both Dcr-1 and Loqs (see below).
Figure 4 Loqs and Dcr-1 Are Present in a Common Protein Complex in S2-Cells
(A) Dcr-1 associates with myc-tagged Loqs PA or PB, and with endogenous Loqs protein. Immunoprecipitation with anti-myc or anti-Loqs antibody was performed using lysates from S2 cells transfected with the indicated expression plasmid. Dcr-1 was detected by Western blotting.
(B) myc-tagged Loqs PB stably associates with Dcr-1 but not Dcr-2. S2 cells were transfected with plasmid expressing myc-tagged Loqs PB, then lysed and immunoprecipitated with anti-myc antibody. The immunoprecipitates were analyzed by Western blotting using anti-Dcr-1 or anti-Dcr-2 antibodies.
(C) S2 cells were transfected with plasmid expressing myc-tagged GFP, Loqs PA, or Loqs PB, then extracted and immunoprecipitated with anti-Dcr-1 antibody. The immunoprecipitates were analyzed by Western blotting using anti-myc antibody.
(D) Anti-Dcr-1 antibody was used to immunoprecipitate Dcr-1 and associated proteins from S2 cell lysates, and the immunoprecipitates were analyzed by Western blotting using anti-Loqs antibody to detect endogenous Loqs protein. The major Loqs protein isoform recovered was Loqs PB. In a longer exposure (bottom panel), a band corresponding in size to Loqs PA is visible. The most abundant Loqs isoform the input sample, Loqs PC, which lacks the third dsRBD, did not immunoprecipitate with Dcr-1, suggesting that the third dsRBD is required for the association of Loqs with Dcr-1.
When immunoprecipitated with anti-Dcr-1 antibody, both myc-tagged Loqs protein isoforms—PA and PB—associated with Dcr-1 (Figure 4C). Moreover, the antibody against endogenous Loqs protein detected two bands corresponding in size to Loqs PA and Loqs PB in the proteins immunoprecipitated with the anti-Dcr-1 antibody (Figure 4D). Loqs PB comprises only approximately 22% of the total Loqs protein in S2 cells, but corresponded to approximately 95% of the Loqs associated with Dcr-1. Loqs PA, which is expressed at comparable levels in S2 cells, accounts for most of the remaining Loqs associated with Dcr-1. In contrast, the putative Loqs PC protein comprises the majority of S2 cell Loqs, but was not recovered in the Dcr-1 immunoprecipitate. Intriguingly, Loqs PA and PB contain a third dsRBD that Loqs PC lacks; perhaps this third dsRBD is required for the association of Loqs with Dcr-1.
The immunoprecipitated Dcr-1–Loqs complexes accurately converted pre-miRNA to mature miRNA (Figure 5). Pre-miRNA processing by the immunoprecipitates was efficient and accurate when we used the anti-Dcr-1 antibody (Figure 5A), and when we used anti-myc antibody and expressed myc-tagged Loqs, but not when we used the anti-myc antibody and expressed myc-tagged GFP (Figure 5A and 5B). Thus, Dcr-1 and Loqs co-associate in a complex capable of converting pre-miRNA into mature miRNA. Our data also demonstrate that an N-terminal tandem myc tag does not perturb Loqs function in pre-miRNA cleavage.
Figure 5 Loqs Is Associated with Pre-miRNA Processing Activity in S2 Cells
(A) Pre-miRNA processing activity co-immunoprecipitates with myc-tagged Loqs PB and with endogenous Dcr-1 or endogenous Loqs, but not with myc-tagged GFP.
(B) Pre-miRNA processing activity co-purifies by immunoprecipitation with both Loqs protein isoforms that interact with Dcr-1, Loqs PA, and Loqs PB. The extracts used in (A) and (B) were independently prepared.
Next, we estimated the size of the pre-miRNA processing complex by gel filtration chromatography. Pre-miRNA processing activity chromatographed as a broad approximately 525-kDa peak that overlapped the peaks of both Dcr-1 and Loqs proteins (Figure 6A and 6B). Dcr-1 protein chromatographed as an approximately 480-kDa complex that overlapped the peak of Loqs PB, which chromatographed as an approximately 630-kDa complex. The Loqs PB isoform accounts for most of the Dcr-1-associated Loqs in S2 cells (see Figure 4D). The apparent size of the Dcr-1 complex suggests that it is either associated with proteins in addition to Loqs or that the complex has an elongated shape that increases its apparent molecular weight. Pre-miRNA processing activity, Loqs, and Dcr-1 were all well resolved from the approximately 230-kDa peak of Dcr-2 (theoretical mass = 197.7 kDa), which corresponds to the Dcr-2/R2D2 heterodimer (theoretical mass = 232.7 kDa). Although the peaks of Loqs and Dcr-1 do not co-migrate, Dcr-1 was stably associated with Loqs after gel filtration: Dcr-1 and Loqs reciprocally co-immunoprecipitated from the pooled peak Dcr-1 fractions (Figure 6C). Loqs was not detected in the Dcr-2 peak by this method (data not shown). Loqs PC, which did not associate with Dcr-1 in immunoprecipitation, chromatographed as a 58-kDa protein, suggesting that it is a free monomeric protein (data not shown).
Figure 6 Analysis of Complexes Containing Pre-miRNA Processing Activity, Dcr-1, and Loqs
(A) S2 cell lysate was fractionated by gel filtration chromatography and analyzed for pre-let-7 processing activity, and Dcr-1, Dcr-2, and Loqs proteins.
(B) The sizes of the distinct complexes containing Loqs (∼630 kDa), Dcr-1 (∼480 kDa), and Dcr-2 (∼230 kDa) and the broad complex containing pre-miRNA processing activity (∼525 kDa) were estimated using molecular weight standards (thyroglobulin, 669 kDa; ferritin, 440 kDa; catalase, 232 kDa; aldolase, 158 kDa; bovine serum albumin, 67 kDa; ovalbumin, 43 kDa; chymotrypsinogen A, 25 kDa) and recombinant Dcr-2 and R2D2 proteins (rDcr-2 and rR2D2). The blue asterisk denotes the peak of pre-let-7 processing activity detected in (A).
(C) Fractions containing the Dcr-1 peak were pooled and immunoprecipitated with either anti-Dcr-1 or anti-Loqs antibodies. Western blotting with anti-Dcr-1 and anti-Loqs antibodies demonstrated that Dcr-1 and Loqs remained associated through gel filtration chromatography.
A Loqs Mutation Reduces Silencing of a miRNA-Controlled Reporter Transgene In Vivo
The loqs
f00791 mutation caused pre-miRNAs to accumulate in the soma and the germ line and strongly reduced mature miR-7 levels in the female germ line, suggesting that Loqs function is required for miRNA-directed silencing in vivo. We introduced a miRNA-regulated yellow fluorescent protein (YFP) reporter into loqs
f00791 homozygous mutant flies. This transgenic reporter expresses in the eye a YFP mRNA bearing four miR-277 binding sites in its 3′ UTR. The four miRNA-binding sites pair with all but the central three nucleotides of miR-277 and are, therefore, predicted to repress reporter mRNA translation rather than trigger mRNA cleavage (Figure 7A). YFP fluorescence was readily detected in the eye and antennae in control flies in which the 3′ UTR of the YFP transgene lacked the four miR-277 binding sites (Figure 7A). When the reporter contained the miR-277 binding sites, YFP expression was repressed in the eye but readily visible in the antennae, indicating that miR-277 is expressed in the eye (loqs/CyO, Figure 7A and B). This expression was verified independently by Northern blots of RNA isolated from eyes dissected away from other tissues of the head (data not shown). Silencing of the miR-277-responsive YFP reporter in the eye was reduced in loqs
f00791 homozygous mutant flies (loqs/loqs, Figure 7A, B and C). As a control, we examined the effect of a strong r2d2 mutation on YFP reporter expression (Figure 7A and C). We measured the maximum fluorescence intensity in each eye for all four genotypes. Figure 7C shows that there was a significant (P < 1.9 × 10−7) increase in YFP fluorescence in eyes homozygous for the weak hypomorphic allele loqs
f00791. This allele reduced miR-277 levels in the soma approximately 2-fold (see Figure 2B); fluorescence in the eye of homozygous mutant loqs flies was 1.8 ± 0.17 (average maximum intensity ± standard deviation; n = 13) times greater than in the eyes of their age-matched heterozygous siblings. In contrast, flies homozygous for a strong hypomorphic r2d2 mutation show only a modest change in fluorescence (1.1 ± 0.09; n = 13; P < 0.025). The Dcr-2 partner protein R2D2 is required for RNAi triggered by exogenous dsRNA [21] or transgenes expressing long dsRNA hairpins (see below and Figure 8). We conclude that the reduced levels of Loqs protein in the loqs
f00791 mutant lead to a statistically significant reduction in miRNA-directed silencing and that the Loqs paralog R2D2 plays little, if any, role in miRNA function.
Figure 7 Silencing of a miRNA-Responsive YFP Reporter Requires loqs but Not r2d2
(A) A YFP transgene expressed from the Pax6-promoter showed strong fluorescence in the eye and weaker fluorescence in the antennae. Due to the underlying normal red eye pigment, the YFP fluorescence was observed in only those ommatidia that are aligned with the optical axis of the stereomicroscope. In heterozygous loqs
f00791/CyO flies bearing a miR-277-responsive, Pax6-promotor-driven, YFP transgene, YFP fluorescence was visible in the antennae but was repressed in the eye. In contrast, in homozygous mutant loqs
f00791 flies, YFP fluorescence was readily detected in the eye. A strong mutation in r2d2 did not comparably alter repression of the miR-277-regulated YFP reporter. The exposure time for the unregulated YFP reporter strain was one-fourth that used for the miR-277-responsive YFP strain. The exposure times were identical for the heterozygous and homozygous loqs and r2d2 flies.
(B) Additional images of eyes from loqs
f00791 heterozygous and homozygous flies bearing the miR-277-responsive YFP reporter transgene diagrammed in (A).
(C) Quantification of fluorescence of the miR-277-responsive YFP transgene in eyes heterozygous or homozygous for loqs or r2d2. The maximum pixel intensity was measured for each eye (excluding antennae and other tissues where miR-277 does not appear to function). The graph displays the average (n = 13) maximum pixel intensity ± standard deviation for each homozygous genotype, normalized to the average value for the corresponding heterozygotes. Statistical significance was estimated using a two-sample Student's t-test assuming unequal variance.
The images in (A) were acquired using a sensitive, GFP long-pass filter set that transmits yellow and red autofluorescence. Images in (B) and for quantitative analysis were acquired using a YFP-specific band-pass filter set that reduced the autofluorescence recorded.
Figure 8 Silencing of white by an IR Partially Depends on loqs
(A) The red eye color of wild-type flies (left) changes to orange (center) and white (right) in response to one or two copies, respectively, of a white IR transgene, which silences the endogenous white gene.
(B) Homozygous mutant r2d2 flies fail to silence white, even in the presence of two copies of the white-IR transgene; heterozygous r2d2/CyO flies repress white expression.
(C) In flies homozygous for loqs
f00791, silencing of white by the white-IR is less efficient; two copies of the white-IR do not produce completely white eyes, whereas they do in heterozygous loqs
f00791/CyO.
(D) The eye color change in loqs
f00791 flies is not caused by the increased white+ gene dose resulting from the mini-white marker in the piggyBac transposon that causes the loqs
f00791 mutation. Flies trans-heterozygous for loqs
f00791 and a mini-white-marked P-element have more red eye pigment than loqs
f00791 homozygous flies, but show more efficient silencing by the white-IR than loqs
f00791 homozygous animals.
(E) The eye pigment of the indicated genotypes was extracted and quantified by green light (480 nm) absorbance, relative to wild-type flies bearing no white-IR transgenes. The graph shows the mean and standard deviation of five independent measurements per genotype.
Loqs Participates in Silencing Triggered by Long dsRNAIn Vivo
dsRNA transcribed as an inverted repeat (IR) triggers silencing of corresponding mRNAs in flies [23,40]. For IR-silencing of the white gene, whose gene product is required to produce the red pigment that colors fly eyes, the extent of silencing is proportionate to the number of copies of the IR-white transgene [23,40] (Figure 8A), but is relatively insensitive to the number of copies of white present (TD and PDZ, unpublished). RNAi in Drosophila requires both Dcr-2, which transforms long dsRNA into siRNA, and R2D2, which collaborates with Dcr-2 to load siRNA into RISC. Thus, IR-silencing of white mRNA is lost in both dcr-2 [23] and r2d2 mutant flies (Figure 8B). We quantified the extent of white silencing by extracting the eye pigment in acidic ethanol and measuring its absorbance at 480 nm (Figures 8E and S1). Loss of R2D2 function in flies expressing one (or two) copies of the white IR transgene and two copies of the endogenous white locus restored red pigment levels to 74 ± 13 (or 73 ± 15 for two copies of IR-white) percent of wild-type flies lacking the white-IR. loqs
f00791 mutant flies were also defective in IR-triggered white silencing, but to a much smaller extent (Figure 8C and 8E). The loqs
f00791 mutation restored pigment levels in flies carrying one copy of the white IR-expressing transgene to 12 ± 2% of wild-type and to 8 ± 0.6% for flies carrying two copies of the white-IR (n = 5; Figures 8C and E). loqs
f00791 heterozygotes were statistically indistinguishable from wild-type flies bearing one copy of IR-white, whose eye pigment concentration was 4 ± 0.5 (or 2 ± 0.6 for two copies of IR-white) percent of wild-type in the absence of the IR-white transgene.
Insertion of a mini-white-expressing piggyBac transposon causes the loqs
f00791 allele. Thus, loqs
f00791 heterozygotes have two copies of the endogenous white locus and one copy of mini-white; loqs
f00791 homozygotes have two copies of endogenous white and two copies of mini-white. The presence of this additional copy of mini-white does not account for the darker red color of white-silenced loqs
f00791 flies, because loqs
f00791 heterozygotes bearing two copies of white, one copy of mini-white (in the piggyBac transposon inserted at loqs), and one copy of a P-element expressing mini-white are effectively silenced by IR-white (Figure 8D). In the absence of the IR-white transgene, the total amount of white expression in these flies is higher than in loqs
f00791 homozygotes (Figure 8D). Thus, reduction of Loqs function accounts for the partial desilencing of white in this system. The modest loss of silencing in the loqs
f00791 mutant flies may reflect the incomplete loss of Loqs protein in this allele. However, Carthew and co-workers previously reported that a dcr-1 null mutation leads to a similar, partial loss of white IR-silencing [23]. The small eye phenotype of dcr-1 null mutants unfortunately renders a quantitative comparison to loqs
f00791 impossible. We propose that—as for pre-miRNA processing—Dcr-1 and Loqs act together to enhance silencing by siRNAs.
Silencing of the Endogenous Stellate Locus Requires Loqs
loqs
f00791 males are incompletely fertile. When Oregon R females were mated to loqs
f00791 homozygous mutant males, only 17% of embryos hatched (n = 479); for loqs
f00791 heterozygous males, 47% of embryos hatched (n = 466). Ninety percent of embryos hatched (n = 753) for wild-type Oregon R males. Genes required for RNA silencing often reduce male fertility, because the X-linked geneSte is epigenetically silenced in testes by dsRNA derived from the bi-directionally transcribed Suppressor of Stellate (Su(Ste)) locus [41]. Ste silencing is genetically similar, but not identical, to RNAi, in that like RNAi it requires the function of the gene armitage
(armi) [24], but unlike RNAi does not require r2d2 (VVV and PDZ, unpublished data). In the absence of Ste silencing, Stellate protein accumulates as protein crystals in the testes. loqs
f00791 mutants contain Stellate crystals in their testes (Figure 9), much like armi72.1 mutants, identifying a second role for loqs in silencing by endogenous RNA triggers, distinct from its function in miRNA biogenesis.
Figure 9 Silencing of Stellate by the dsRNA-Generator Su(Ste) Requires loqs
Testes were stained for DNA (red) and Stellate protein (green). Defects in RNA silencing often lead to accumulation of Stellate protein crystals in testes. For example, the testes from the strong allele armi72.1, but not wild-type Oregon R testes, show Stellate protein staining. Testes from loqs
f00791 males show strong accumulation of Stellate protein, consistent with their significantly impaired fertility.
A Germ-Line Stem Cell Defect in loqs
f00791 Mutant Females
The loqs gene has a critical function in oogenesis, as loqs
f00791 females have small ovaries (Figure 10A) and are completely sterile. Drosophila ovaries comprise ovarioles that contain developmentally ordered egg chambers, which are produced continuously in the adult by germ-line stem cell division. As a result, mutations that block stem cell division or maintenance lead to ovarioles containing few egg chambers. loqs
f00791 mutant females lay no eggs. Whereas wild-type females contain 7 ± 0.8 (n = 15) previtellogenic egg chambers per ovariole, loqs
f00791 contain only 3 ± 0.8 (n = 20). Excision of the piggybac transposon in loqs
f00791 restores fertility, demonstrating that these defects reflect loss of Loqs function. The mature oocytes in loqs
f00791 ovarioles have normal dorsal appendages, indicating that dorsoventral patterning is normal. In contrast, mutations in armi, spnE, and aub disrupt both dorsoventral and anteroposterior patterning [42–44]. These mutations all disrupt RNAi andSte silencing, but display no global defects in miRNA biogenesis or function, unlike loqs [24,41,45,46].
Figure 10 loqsf00791 Fail to Maintain Germ-Line Stem Cells
(A) Wild-type ovarioles contain a germarium and a developmentally ordered array of six to eight egg chambers, whereas loqs
f00791 mutant ovarioles contain a smaller than normal germarium, two or three pre-vitellogenic egg chambers, and a late-stage egg chamber. Wild-type and loqs ovarioles are shown at the same magnification.
(B) In wild-type ovarioles, the germarium contains several newly formed germ-line cysts surrounded by somatic follicle cells. In contrast, loqs
f00791 mutant germaria contain few germ-line cells, which are not organized into distinct cysts. The follicle cell layer is also significantly reduced inloqs
f00791 germaria.
(C) Wild-type and loqs mutant germaria labeled for α-Spectrin (green) and filamentous Actin (red). In wild type, anti-α-Spectrin labels the spectrosome (ss), a structure characteristic of germ-line stem cells, which are normally found at the anterior of the germarium, apposed to the somatic terminal cells (tc). The cystoblasts, the daughters of the stem cells, also contain a spectrosome, but are located posterior to the stem cells. In loqs mutant ovaries, spectrosome-containing cells were not detected, indicating that normal germ-line stem cells are not present. These observations indicate that stem cells are not maintained.
In (A) and (B), ovaries were labeled for filamentous actin (red) using rhodamine phalloidin, DNA (blue) using TOTO3 (Molecular Probes), and the germ-line marker Vasa (green) using rabbit anti-Vasa antibody detected with fluorescein-conjugated anti-rabbit secondary antibody. In (B) and (C), wild-type and loqs germaria are shown at the same magnification.
Oogenesis is initiated in the germarium, which contains the germ-line stem cells as well as the early germ-line cysts that will form egg chambers. In loqs
f00791 mutant ovarioles, the germaria generally contain a limited number of cells that stain for Vasa, indicating that they are of germ-line origin (Figure 10B). No mitotic figures were observed, nor were separate cysts. Germ-line stem cells and their daughter cells, the cystoblasts, are characterized by the presence of a spherical structure, the spectrosome, that stains intensely with anti-Spectrin antibodies [47–49]. We stained wild-type and loqs
f00791 germaria with anti-α-Spectrin antibodies (Figures 10C and S2). We could not detect spectrosomes in the loqs mutant germaria, suggesting that in these germaria, dissected from flies 3–4 d old, no stem cells remained. Stem cells must have originally been present, because loqs mutant ovaries produce some late-stage oocytes. Thus, most of the original stem cells may have died or differentiated into cystoblasts without renewing the stem cell pool. At present, we cannot distinguish between these alternatives. We conclude that loqs
f00791 mutants, which are defective in three distinct types of RNA silencing, fail to maintain germ-line stem cells.
Discussion
RNase III Endonucleases Act with dsRNA-Binding Partner Proteins in RNA Silencing
Collectively, Dcr-1 and Loqs, Drosha and Pasha, and Dcr-2 and R2D2 comprise six of the 12 dsRBD proteins predicted to be encoded by the Drosophila genome [50]. Thus, at least half of all dsRBD proteins in flies participate in RNA silencing. In Caenorhabditis elegans, the R2D2-like dsRBD protein RDE-4 is required for RNA interference and interacts with DCR-1, the sole worm Dicer gene [51]. RDE-4 is equally similar to Loqs (E-value = 0.03) and R2D2 (E-value = 0.026; search restricted to C. elegans proteins). The Drosha/Pasha complex is also present in C. elegans [10] as well as cultured human cells [11–13]. Similarly, the Arabidopsis thaliana dsRBD protein HYL1 is required for the production of mature miRNAs, and hyl1 mutant plants have a phenotype similar to that of dicer-like 1 (dcl1) [52,53]. Hiraguri and co-workers [54] recently demonstrated that HYL1 is a dsRNA-binding protein that binds DCL1 and that the HYL1 paralog DRB4 binds the Dicer protein DCL4. Pairing of RNase III endonucleases with dsRBD proteins is thus a recurring theme in RNA silencing.
A dsRBD Partner for Human Dicer?
The human genome encodes one Dicer protein, which is more closely related to Drosophila Dcr-1 than Dcr-2. Sequence analysis of human proteins for similarity to either C. elegans RDE-4 or Drosophila R2D2 does not identify a reasonable candidate for a dsRBD partner protein for human Dicer. In contrast, the human TRBP is highly similar to Drosophila Loqs (E-value = 5 × 10−36). For comparison, the human proteins most similar to R2D2 or RDE-4 give E-values of 8 × 10−8 and 0.42, respectively, when the search is restricted to human proteins. Human TRBP was first identified [55] because it binds HIV trans-activator RNA (TAR), a stem-loop structure required for active HIV transcription [56–58]. Remarkably, the secondary structure of TAR resembles a miRNA precursor, and the recent discovery of Epstein-Barr virus-encoded miRNAs [59] has fueled speculation that TAR may be a viral pre-miRNA [60].
Deletion of PRBP, the mouse homolog of TRBP, yields viable mice that often die at the age of weaning. Surviving homozygous mutant males show defects in spermatogenesis attributed to abnormal sperm maturation rather than proliferation [61]. In contrast, Dicer knockout mice show very early embryonic lethality [62]. If mouse Dicer and PRBP collaborate to produce mature miRNA, the essential function of Dicer during mouse development must either be independent of miRNA function, or a redundant factor must replace PRBP during embryonic development but not spermatogenesis.
Loqs and Dcr-1 Protein Complexes
Together with dcr-1, the gene loqs is required in flies for normal pre-miRNA processing. Loqs and Dcr-1 reciprocally co-immunoprecipate. Pre-miRNA processing activity also co-immunoprecipitates with Dcr-1 and Loqs. However, in gel filtration chromatography, the two proteins overlap but do not precisely co-purify. Loqs and Dcr-1 may form a protein complex analogous to the Dcr-2/R2D2 and Drosha/Pasha complexes [10,21], but this complex may be transient, with Loqs also associating with other components of the RNA silencing machinery, perhaps even escorting the mature miRNA to Ago1, an approximately 110-kDa Argonaute protein associated with mature miRNAs in flies. In fact, the predominant Loqs-containing complex in S2 cell lysate is about 150 kDa larger than the peak of Dcr-1, so it could contain Dcr-1, Loqs, and Ago1. The data of Siomi and co-workers demonstrating that Ago1 associates with both Dcr-1 [28] and Loqs (Saito K, et al. DOI: 10.1371/journal.pbio.0030235) support such a view.
Cross-Talk between the Dcr-1 and Dcr-2 Pathways in Drosophila
In humans and C. elegans, a single Dicer gene is responsible for generating both siRNAs and miRNAs. Drosophila has apparently duplicated both its ancestral Dicer RNase III endonuclease and its dsRBD partner protein, dedicating Dcr-1/Loqs to miRNA processing and Dcr-2/R2D2 to RNAi. Nonetheless, these two pathways are not completely separate, because cells lackingdcr-1 are not fully competent for IR-triggered silencing [23]. Dcr-1 is not required for siRNA production, yet embryo extracts lacking Dcr-1 fail to assemble RISC [23]. Dcr-1 has been proposed to be a component of “holo-RISC,” an 80S complex containing many, but not all, components of the RNAi pathway in flies [63]. The loqs
f00791 mutation also reduced the efficiency of IR-triggered silencing in vivo. Therefore, we propose that Dcr-1 must partner with Loqs not only during the processing of pre-miRNA to mature miRNA, but also to ensure Dcr-1 function in the Dcr-2-dependent RNAi pathway.
Carthew and colleagues found no function for Dcr-2 in miRNA biogenesis [23]. Consistent with their results, we found little if any requirement for R2D2 in miRNA-directed silencing (see Figure 7C). Moreover, null or strong hypomorphic alleles of either dcr-2 or r2d2 show no overt phenotype, whereas the dcr-1Q1147X null mutation is embryonic lethal [23].
Stellate Silencing Requires Loqs
Endogenous silencing of the Stellate locus in testes is genetically distinct from miRNA-directed silencing, because it requires armitage, a gene that plays no general role in miRNA biogenesis or function [24]. Stellate silencing resembles RNAi in that Stellate expression is repressed by a dsRNA trigger transcribed from the Su(Ste) gene. Su(Ste) dsRNA produces siRNAs, called repeat-associated siRNAs, that are longer than the siRNAs produced in the RNAi pathway in Drosophila [41]. Even the weak allele described in this study, loqs
f00791
, which reduces loqs mRNA levels only approximately 3-fold in testes, dramatically de-silences Stellate. Given the intimate association of Dcr-1 with Loqs, our data raise the possibility that Loqs acts to silence Stellate in collaboration with Dcr-1, which may generate the Su(Ste) repeat-associated siRNAs.
Germ-Line Stem Cells and miRNAs
The loqs
f00791 mutation is the first viable allele in Drosophila with a generalized defect in miRNA production. The allele may therefore be useful for future phenotypic analysis of miRNA-dependent pathways during the life cycle of Drosophila. The most obvious phenotype of loqs
f00791 is female sterility. loqs
f00791 homozygotes produce few egg chambers, indicating a defect in germ-line stem cell maintenance or division. The loqs
f00791 phenotype is similar to mutants in piwi [64], which encodes a member of the Argonaute protein family of core RISC components. In piwi mutant ovaries, germ-line stem cells fail to divide and instead differentiate directly into cystoblasts, depleting the germarium of germ-line stem cells. loqs mutants display a similar phenotype: we did not detect germ-line stem cells (i.e., spectrosome-containing cells) in loqs
f00791 homozygous germaria, suggesting that Loqs is required to maintain stem cells. Piwi is required in terminal filament cells, somatic cells surrounding the tip of the germarium, to send a signal that prevents germ-line stem cells from differentiating [64,65]. Piwi is also required in germ-line stem cells themselves to stimulate their proliferation [65]. Perhaps Piwi is at the core of an effector complex loaded with small RNA produced by Dcr-1 and Loqs. Intriguingly, dcr-1 knockout mice die at embryonic day 7.5, apparently devoid of stem cells [62].
Materials and Methods
PiggyBac excision
To establish that insertion of the f00791 piggyBac transposon in the loqs gene caused the female sterility of loqs
f00791 mutants, we excised the transposon by introducing into loqs
f00791 heterozygotes a transgene expressing the piggyBac transposase from a Hermes element inserted on Chromosome 3 [66]. F1 male progeny of these flies were mated to yw; Sp/CyO virgins, and the resulting F2 progeny screened for loss of white expression (i.e., white eyes). Of 100 F2 progeny examined, one white
- male was recovered. A line established from this fly was homozygous female fertile.
Real-time RT-PCR analysis
Two μg of total RNA was reverse transcribed using 5′-GCG AAT TCT TTT TTT TTT TTT TTT TTT TTT-3′ oligonucleotide as primer and Superscript II reverse transcriptase (Invitrogen, Carlsbad, California, United States). After extension, cDNA samples were diluted 3-fold with water. One μl of diluted cDNA was used for quantitative PCR using the Quantitect SyBr-green kit (Qiagen, Valencia, California, United States) in a DNA Engine Opticon 2 (MJ Research [Bio-Rad, Hercules, California, United States]). Oligonucleotide primers were 5′-ATG GAC CAG GAG AAT TTC CAC GGC-3′ and 5′-GGC CTC GTC GCT GGG CAA TAT TAC-3′ for loqs and 5′-AAG TTG CTG CTC TGG TTG TCG-3′ and 5′-GCC ACA CGC AGC TCA TTG TAG-3′ for actin5C. Amplification efficiencies were identical for both oligonucleotide pairs.
RNA isolation and detection by Northern blot
RNA was isolated from whole flies, dissected organs or S2 cells using Trizol (Invitrogen) according to the manufacturer's instructions. The RNA was quantified by absorbance at 260 nm, and 2–10 μg of total RNA was resolved by electrophoresis through a 20% denaturing acrylamide/urea gel (National Diagnostics, Atlanta, Georgia, United States). As a positive control for miR-277 hybridization, 10 fmol of phosphorylated miR-277 synthetic oligonucleotide (Dharmacon, Lafayette, Colorado, United States) was included on the gel. After electrophoresis, the gel was transferred to Hybond N+ (Amersham-Pharmacia, Little Chalfont, United Kingdom) in 0.5x TBE in a semi-dry transfer system (Transblot SD, Bio-Rad) at 20 V for 60 min. The RNA was UV cross-linked to the membrane (Stratalinker, Stratagene, La Jolla, California, United States) and pre-hybridized in 10 ml Church buffer [67] for 60 min at 37 °C.
RNA (Dharmacon) or DNA (IDT, Coralville, Iowa, United States) probes (25 pmol per reaction) were 5′-radiolabeled with polynucleotide kinase (New England Biolabs, Beverly, Massachusetts, United States) and γ-32P-ATP (New England Nuclear, Boston, Massachusetts, United States) (330 μCi per reaction; specific activity 7,000 Ci/mmol). After labeling, unincorporated radioactivity was separated from the labeled probe using a Sephadex G-25 spin column (Roche, Basel, Switzerland). The labeled probe oligonucleotide was added to 10 ml of Church buffer and used for hybridization. For RNA probes, hybridization was carried out at 65 °C; DNA probes were hybridized at 37 °C. For both, hybridization was overnight followed by five 30-min washes with 2× SSC/0.1% (w/v) SDS. Membranes were exposed to phosphorimaging screens (Fuji, Tokyo, Japan). To strip probes, the membranes were boiled twice in 0.1% SDS for 1 min in a microwave oven. The following probes were used for detection: 5′-UCG UAC CAG AUA GUG CAU UUU CA-3′ for miR-277; 5′-CAG CTT TCA AAA TGA TCT CAC T-3′ for bantam; 5′-ACA ACA AAA UCA CUA GUC UUC CA-3′ for miR-7; 5′-TAC AAC CCT CAA CCA TAT GTA GTC CAA GCA-3′ for 2S rRNA.
Molecular cloning and generation of transgenic flies
Plasmids for the expression of myc-Loqs PA (pKF111) and myc-Loqs PB (pKF109) were created by PCR amplifying loqs mRNA with oligonucleotides 5′-AGC GGA TCC ATG GAA CAA AAA CTT ATT TCT GAA GAA GAC TTG GAA CAA AAA CTT ATT TCT GAA GAA GAC TTG GCC ATG GAC CAG GAG AAT TTC CAC GGC-3′ (appending two myc-tags to the N-terminus of Loqs) and 5′-TTA TGC GGC CGC CTA CTT CTT GGT CAT GAT CTT CAA GTA CTC-3′ from male and ovary cDNA, respectively. The reaction products were cloned into pUbi-Casper-SV40, which was created by inserting the SV-40 polyadenylation signal from pEGFP-N1 (Clontech, Palo Alto, California, United States) into pUbi-Casper2 (kind gift from Dr. Inge The). The vector for myc-tagged GFP expression (pKF63) was constructed similarly.
The vector for the expression of miR-277-responsive myc-YFP was constructed by first inserting the annealed oligonucleotides 5′-CAT GGA ACA AAA ACT TAT TTC TGA AGA AGA CTT GGG-3′ and 5′-CAT GCC CAA GTC TTC TTC AGA AAT AAG TTT TTG TTC-3′ into NcoI-cut pBSII-ITR1.1k-EYFP (a kind gift from Dr. Malcom Fraser) to add an N-terminal myc-tag. Then the vector was digested with NotI/XbaI and the annealed oligonucleotides 5′-GGC CTG TCG TAC CAG AGG ATG CAT TTA CAG TGT CGT ACC AGA GGA TGC ATT TAT GTC GTA CCA GAG GAT GCA TTT ACA GTG TCG TAC CAG AGG ATG CAT TTA-3′ and 5′-CTA GTA AAT GCA TCC TCT GGT ACG ACA CTG TAA ATG CAT CCT CTG GTA CGA CAT AAA TGC ATC CTC TGG TAC GAC ACT GTA AAT GCA TCC TCT GGT ACG ACA-3′ inserted, appending four miR-277 target sites to the 3′ UTR. Subsequently, the Pax6/EYFP/miR-277-target/SV-40-polyA cassette [68–70] was cloned into pPCar20.1 [71] creating pKF77. All of the described constructs were sequence verified. Transgenic flies were obtained by injection of pKF77 with Δ2–3 helper plasmid into ry506 embryos using standard methods [72].
S2 cell culture and RNAi
Drosophila S2 cells were the kind gift of Dr. Neal Silverman. The cells were cultured in Schneider's Drosophila medium (Life Technologies, Carlsbad, California, United States) supplemented with 10% FBS, penicillin-streptomycin mix (Life Technologies), and 0.2% of conditioned Schneider's medium. Transfection of plasmids was performed using siLentFect (Bio-Rad).
Gene fragments for the preparation of dsRNA were cloned into a Litmus28i vector (NEB) that was modified into a T/A cloning vector [73]. The following oligonucleotide pairs were used to obtain gene fragments: 5′-TTG GGC GAC GTT TTC GAG TCG ATC-3′ and 5′-TTT GGC CGC CGT GCA CTT GGC AAT-3′ for dcr-1; 5′-CTG CCC ATT TGC TCG ACA TCC CTC C-3′ and 5′-TTA CAG AGG TCA AAT CCA AGC TTG-3′ for dcr-2; 5′-ATG GAC CAG GAG AAT TTC CAC GGC-3′ and 5′-GGC CTC GTC GCT GGG CAA TAT TAC-3′ for loqs; 5′-ATA CAA TCT CCA CCA ATT TGT AGG-3′ and 5′-CGT CAA ATT ATT TAA AAT ATT TGT TTC-3′ for r2d2; 5′-AGC AGC AGC AGT GAT AGC GAT GGC-3′ and 5′-TCG GTT ATT TTA TTT GTT GCT TTA ATG-3′ for Drosha; 5′-GAT CAC ATG GTC CTG CTG GAG TTC GTG-3′ and 5′-CAG GTT CAG GGG GAG GTG TG-3′ for GFP. Gene fragments were amplified from the plasmid templates with both flanking T7 RNA polymerase promoters using oligonucleotides 5′-CTA TGA CCA TGA TTA CGC CAA GC-3′ and 5′-CAC GAC GTT GTA AAA CGA CGG CCA-3′. RNA synthesis from the PCR products was performed as described [74], and the phenol-extracted RNA products were denatured for 5 min at 95 °C and then re-annealed for 30 min at 65 °C. The concentration of dsRNA was estimated by native agarose gel electrophoresis and comparison to a DNA standard. S2 cells were seeded at 106 cells/ml, and dsRNA was added directly to the growth medium at a final concentration of 10 μg/ml. After three days, additional dsRNA was added, and the cells were diluted 5-fold on the following day to permit further growth. Eight days after the initial dsRNA treatment, the cells were harvested by centrifugation, washed three times in phosphate-buffered saline, and re-suspended per ml of original culture in 15 μl of lysis buffer (30 mM HEPES-KOH [pH 7.4], 100 mM KOAc, 2 mM Mg(OAc)2), supplemented with protease inhibitors (Complete, Roche). The cells were disrupted either with 50 strokes of a Dounce homogenizer using a “B” pestle or by freeze/thawing. The extract was separated from debris by centrifugation at 18,000 × g for 30 min and aliquots frozen at −80 °C.
In vitro pre-miRNA processing
Synthetic Drosophila pre-let-7 bearing the characteristic end structure created by Drosha processing of pri-miRNA (5′-UGA GGU AGU AGG UUG UAU AGU AGU AAU UAC ACA UCA UAC UAU ACA AUG UGC UAG CUU UCU-3′) was 5′-32P radiolabeled with PNK, gel purified, and re-folded by heating at 95 °C for 2 min, then incubating at 37 °C for 1 h. Pre-let-7 (20 nM) was incubated in a standard RNAi reaction with 50% (v/v) S2 cell lysate for 30 min. The reaction was deproteinized with Proteinase K and Phenol [74], then resolved by electrophoresis in a 15% denaturing polyacrylamide gel. Pre-let-7 and let-7 were quantified by phosphorimagery (BAS-5000; Fuji). The fraction processed (y) and time (t) was analyzed using Igor Pro 4.09A (Wavemetrics, Portland, Oregon, United States) by fitting the data to y = k1−k1e(−k2
t), where k1k2 corresponds to the initial velocity.
Immunoprecipitation and immunoblotting
For immunoprecipitation, 100 μl of S2 cell extract were incubated with 2 μl affinity-purified antibody or 2 μl monoclonal anti-myc antibodies (clone 9E10, Sigma, St. Louis, Missouri, United States) for 30 min at 4 °C. Subsequently, protein A/G agarose (Calbiochem, San Diego, California, United States) or anti-rabbit IgG agarose (eBioscience, San Diego, California, United States) was added and the samples agitated at 4 °C for 90 min. For RNase treatment, RNase A was added to a final concentration of 50 μg/ml, and the samples incubated for 15 min at 4 °C prior to immunoprecipitation. Beads were washed four times with 1 ml of lysis buffer containing 1% (v/v) Triton X-100 (Sigma).
For Western blotting, the proteins were separated on 8% polyacrylamide/SDS gels and transferred to PVDF-membrane. All incubations and washes were in TBS containing 0.02% (v/v) Tween-20. For the rabbit primary antibodies, we used a secondary antibody that does not recognize the reduced form of rabbit IgG (Trueblot, eBioscience), permitting detection of Loqs, which migrates near the heavy antibody chain present in the immunoprecipitates.
To generate anti-Loqs antibody, two rabbits were immunized with the KLH-conjugated peptide MDQENFHGSSC. The specificity of the antibody was verified by Western blotting using extracts prepared from S2 cells transfected with the myc-Loqs PB expression vector, using untransfected S2 cell extract for comparison. Both rabbit anti-sera reacted with the over-expressed protein and against three small endogenous proteins. The antibody was affinity purified using the peptide antigen immobilized on agarose beads. Anti-Dcr-2 antibody was raised in chicken using the KLH-conjugated peptide CNKADKSKDRTYKTE. IgY was affinity-purified from egg yolk using peptide antigen immobilized on agarose beads. Anti-Drosha antibody was kindly provided by Greg Hannon [10].
Gel-filtration chromatography
200 μl of S2 cell extract was separated by chromatography on a Superdex-200 HR 10/300 GL column (Amersham-Pharmacia) using a BioCad Sprint (PerSeptive Biosystems, Framingham, Massachusetts, United States) as described [75]. Protein from three-quarters of every other fraction was precipitated with 10% (v/v) trichloroacetic acid and 0.001% (w/v) deoxycholate and analyzed by Western blotting. The remainder of each fraction was analyzed for pre-miRNA processing activity.
Analysis of YFP reporter fluorescence and eye color using the white-IR transgene
Fluorescence and normal light images were taken with a Leica MZ-FLIII stereomicroscope equipped with a cooled color CCD-camera (Firecam, Leica, Wetzlar, Germany). The control animals expressing YFP without the miR-277 target sites contained a pBAC3xP3-EYFP, p-Gal4Δ-K10 insertion on the X chromosome [66]. Maximal pixel intensity was determined using ImageGuage 4.2 (Fuji). The average intrinsic background fluorescence present in Oregon R eyes (n = 4) was subtracted from the value determined for each YFP-expressing eye. Eye pigment was measured as described [76]. The heads of 10 males (3–4 d post eclosion) of each genotype were manually dissected. For each genotype, five samples of two heads each were homogenized in 0.1 ml of 0.01 M HCl in ethanol. The homogenates were placed at 4 °C overnight, warmed to 50 °C for 5 min, clarified by centrifugation, and the optical density at 480 nm of the supernatant measured relative to the value for the Oregon R stock.
Analysis of stellate expression in testes and determination of hatch rates and immunofluorescence microscopy
Stellate expression and hatch rates were analyzed as described previously [24]. Immunofluorescence microscopy was as described previously [77]. Spectrosome and fusome were labeled with monoclonal antibody 1B1 (Developmental Studies Hybridoma Bank, Iowa City, Iowa, United States), as described by Lin and Spradling [78].
Supporting Information
Figure S1 A Concentration Series Generated by Dilution of the Eye Pigment Extract from Oregon R Flies
The concentration of each sample, relative to the undiluted sample, was plotted versus its absorbance at 480 nm. The data were fit to a line using Igor Pro 5.01.
(622 KB EPS).
Click here for additional data file.
Figure S2 Loqs Disrupts Germ-Line Stem Cell Maintenance
Wild type and loqs mutants were labeled for α-Spectrin and Actin. In the merged images, α-Spectrin is green; Actin is red. Two examples of wild-type and loqs
f00791 mutant germaria are shown. In wild type, anti-α-Spectrin labeled both the spectrosome, a spherical structure unique to the germ-line stem cells and their daughters, the cystoblasts, and the highly branched fusome found in the cystocytes. The stem cells are located at the anterior of the germarium. In germaria isolated from 3 to 4 day-old loqs mutant females, none of the cells showed a prominent spectrosome, although fusome was detected. Thus, stem cells were originally present but were not maintained. The germ-line cells that remain appear to be cystocytes. The muscle sheath surrounding the ovarioles stains intensely for Actin. α-Spectrin was labeled with a monoclonal antibody; filamentous Actin was labeled with rhodamine-phalloidin.
Scale bar in the upper right panel = 10 μm.
(4.8 MB EPS).
Click here for additional data file.
Accession Numbers
The Arabidopsis Information Resource (http://www.arabidopsis.org) accession numbers for the genes and gene products discussed in this paper are: DCL1 (AT1G01040) and Hyl1 (AT1G09700).
The Ensembl (http://www.ensembl.org/Homo_sapiens) accession numbers for the genes and gene products discussed in this paper are: C. elegans rde4 (T20G5.11) and dcr1 (K12H4.8), human DGCR8 (NSG00000128191), Ago2 (ENSG00000123908), Exportin5 (ENSG00000124571) and TRBP (ENSG00000139546), and mouse PRBP (ENSMUSG00000023051).
The FlyBase (http://flybase.bio.indiana.edu) accession numbers for the genes and gene products discussed in this paper are: ago1 (CG6671, FBgn0026611), ago2 (CG7439, FBgn0046812), armi (CG11513, FBgn0041164), aub (CG6137, FBgn0000146), dcr-1 (CG4792, FBgn0039016), dcr-2 (CG6493, FBgn0034246), drosha (G8730, FBgn0031051), loqs (CG6866, FBgn0032515), pasha (CG1800, FBgn0039861), piwi (CG6122, FBgn0004872), r2d2 (CG7138, FBgn0031951), spnE (CG3158, FBgn0003483), Stellate (FBgn0003523), Su(Ste) (FBgn0003582), vasa (CG3506, FBgn0003970), and white (CG2759, FBgn0003996).
The Rfam (http://www.sanger.ac.uk/Software/Rfam/mirna/index.shtml) accession numbers for the genes and gene products discussed in this paper are: bantam (MI0000387), let-7 (MI0000416), miR-277 (MI0000360), miR-7 (MI0000127), and TAR (RF00250).
We thank members of the Zamore lab for encouragement, helpful discussions, and comments on the manuscript, and Birgit Koppetsch for help with confocal microscopy. We thank Richard Carthew and Dean P. Smith for kindly sharing fly stocks and Greg Hannon (anti-Dcr-1, anti-Drosha), Maria Pia Bozzetti (anti-Stellate) and Paul Lasko (anti-Vasa) for antibodies, Malcom Fraser for EYFP plasmids, and Inge The and Vivian Su for S2 cell expression vectors. PDZ is a W.M. Keck Foundation Young Scholar in Medical Research. This work was supported in part by grants from the National Institutes of Health to PDZ (GM62862–01 and GM65236–01) and WET (HD049116), and to post-doctoral fellowships from the Human Frontier Science Program to KF and YT.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. KF, YT, TD, VVV, WET, and PDZ conceived and designed the experiments. KF, YT, TD, VVV, DPB, CK, and WET performed the experiments. KF, YT, TD, VVV, DPB, CK, WET, and PDZ analyzed the data. KF, YT, TD, VVV, AMD, DPB, CK, WET, and PDZ contributed reagents/materials/analysis tools. KF, WET, YT, and PDZ wrote the paper.
Citation: Förstemann K, Tomari Y, Du T, Vagin VV, Denli AM, et al. (2005) Normal microRNA Maturation and Germ-Line Stem Cell Maintenance Requires Loquacious, a Double-Stranded RNA-Binding Domain Protein. PLoS Biol 3(7): e236.
Abbreviations
dsRBDdouble-stranded RNA binding domain
dsRNAdouble-stranded RNA
GFPgreen fluorescent protein
IRinverted repeat
miRNAmicroRNA
PAprotein isoform A
PBprotein isoform B
PCprotein isoform C
pre-miRNAprecursor miRNA
pri-miRNAprimary miRNA
RARNA splice variant A
RBRNA splice variant B
RCRNA splice variant C
RISCRNA-induced silencing complex
RNAiRNA interference
S2Schneider-2
siRNAsmall interfering RNA
TARtrans-activator RNA
YFPyellow fluorescent protein
==== Refs
References
Bartel DP MicroRNAs: Genomics, biogenesis, mechanism, and function Cell 2004 116 281 297 14744438
He L Hannon GJ MicroRNAs: Small RNAs with a big role in gene regulation Nat Rev Genet 2004 5 522 531 15211354
Lee Y Jeon K Lee JT Kim S Kim VN MicroRNA maturation: Stepwise processing and subcellular localization EMBO J 2002 21 4663 4670 12198168
Bracht J Hunter S Eachus R Weeks P Pasquinelli A Trans-splicing and polyadenylation of let-7 microRNA primary transcripts RNA 2004 10 1586 1594 15337850
Cai X Hagedorn C Cullen B Human microRNAs are processed from capped, polyadenylated transcripts that can also function as mRNAs RNA 2004 10 1957 1966 15525708
Lee Y Kim M Han J Yeom KH Lee S MicroRNA genes are transcribed by RNA polymerase II EMBO J 2004 23 4051 4060 15372072
Parizotto E Dunoyer P Rahm N Himber C Voinnet O In vivo investigation of the transcription, processing, endonucleolytic activity, and functional relevance of the spatial distribution of a plant miRNA Genes Dev 2004 18 2237 2242 15371337
Baskerville S Bartel DP Microarray profiling of microRNAs reveals frequent coexpression with neighboring miRNAs and host genes RNA 2005 11 241 247 15701730
Lee Y Ahn C Han J Choi H Kim J The nuclear RNase III Drosha initiates microRNA processing Nature 2003 425 415 419 14508493
Denli AM Tops BB Plasterk RH Ketting RF Hannon GJ Processing of primary microRNAs by the Microprocessor complex Nature 2004 432 231 235 15531879
Gregory RI Yan KP Amuthan G Chendrimada T Doratotaj B The Microprocessor complex mediates the genesis of microRNAs Nature 2004 432 235 240 15531877
Han J Lee Y Yeom KH Kim YK Jin H The Drosha-DGCR8 complex in primary microRNA processing Genes Dev 2004 18 3016 3027 15574589
Landthaler M Yalcin A Tuschl T The human DiGeorge syndrome critical region gene 8 and its D. melanogaster homolog are required for miRNA biogenesis Curr Biol 2004 14 2162 2167 15589161
Yi R Qin Y Macara IG Cullen BR Exportin-5 mediates the nuclear export of pre-microRNAs and short hairpin RNAs Genes Dev 2003 17 3011 3016 14681208
Lund E Guttinger S Calado A Dahlberg JE Kutay U Nuclear export of microRNA precursors Science 2004 303 95 98 14631048
Zeng Y Cullen BR Structural requirements for pre-microRNA binding and nuclear export by Exportin 5 Nucleic Acids Res 2004 32 4776 4785 15356295
Grishok A Pasquinelli AE Conte D Li N Parrish S Genes and mechanisms related to RNA interference regulate expression of the small temporal RNAs that control C. elegans developmental timing Cell 2001 106 23 34 11461699
Hutvágner G McLachlan J Pasquinelli AE Balint Tuschl T A cellular function for the RNA-interference enzyme Dicer in the maturation of the let-7 small temporal RNA Science 2001 293 834 838 11452083
Ketting RF Fischer SE Bernstein E Sijen T Hannon GJ Dicer functions in RNA interference and in synthesis of small RNA involved in developmental timing in C. elegans
Genes Dev 2001 15 2654 2659 11641272
Park W Li J Song R Messing J Chen X CARPEL FACTORY, a Dicer homolog, and HEN1, a novel protein, act in microRNA metabolism in Arabidopsis thaliana
Curr Biol 2002 12 1484 1495 12225663
Liu Q Rand TA Kalidas S Du F Kim HE R2D2, a bridge between the initiation and effector steps of the Drosophila RNAi pathway Science 2003 301 1921 1925 14512631
Carmell MA Hannon GJ RNase III enzymes and the initiation of gene silencing Nat Struct Mol Biol 2004 11 214 218 14983173
Lee YS Nakahara K Pham JW Kim K He Z Distinct roles for Drosophila Dicer-1 and Dicer-2 in the siRNA/miRNA silencing pathways Cell 2004 117 69 81 15066283
Tomari Y Du T Haley B Schwarz DS Bennett R RISC assembly defects in the Drosophila RNAi mutant armitage
Cell 2004 116 831 841 15035985
Tomari Y Matranga C Haley B Martinez N Zamore PD A protein sensor for siRNA asymmetry Science 2004 306 1377 1380 15550672
Martinez J Patkaniowska A Urlaub H Lührmann R Tuschl T Single-stranded antisense siRNAs guide target RNA cleavage in RNAi Cell 2002 110 563 574 12230974
Hutvágner G Zamore PD A microRNA in a multiple-turnover RNAi enzyme complex Science 2002 297 2056 2060 12154197
Schwarz DS Hutvágner G Haley B Zamore PD Evidence that siRNAs function as guides, not primers, in the Drosophila and human RNAi pathways Mol Cell 2002 10 537 548 12408822
Okamura K Ishizuka A Siomi H Siomi MC Distinct roles for Argonaute proteins in small RNA-directed RNA cleavage pathways Genes Dev 2004 18 1655 1666 15231716
Parker JS Roe SM Barford D Crystal structure of a PIWI protein suggests mechanisms for siRNA recognition and slicer activity EMBO J 2004 23 4727 4737 15565169
Song JJ Smith SK Hannon GJ Joshua-Tor L Crystal structure of Argonaute and its implications for RISC slicer activity Science 2004 305 1434 1437 15284453
Ma JB Yuan YR Meister G Pei Y Tuschl T Structural basis for 5′-end-specific recognition of guide RNA by the A. fulgidus Piwi protein Nature 2005 434 666 670 15800629
Parker JS Roe SM Barford D Structural insights into mRNA recognition from a PIWI domain-siRNA guide complex Nature 2005 434 663 666 15800628
Liu J Carmell MA Rivas FV Marsden CG Thomson JM Argonaute2 is the catalytic engine of mammalian RNAi Science 2004 305 1437 1441 15284456
Meister G Landthaler M Patkaniowska A Dorsett Y Teng G Human Argonaute2 mediates RNA cleavage targeted by miRNAs and siRNAs Mol Cell 2004 15 185 197 15260970
Rand TA Ginalski K Grishin NV Wang X Biochemical identification of Argonaute 2 as the sole protein required for RNA-induced silencing complex activity Proc Natl Acad Sci U S A 2004 101 14385 14389 15452342
Marchler-Bauer A Bryant SH CD-Search: Protein domain annotations on the fly Nucleic Acids Res 2004 32 W327 331 15215404
Drysdale RA Crosby MA Gelbart W Campbell K Emmert D FlyBase: Genes and gene models Nucleic Acids Res 2005 33 D390 395 Database Issue 15608223
Thibault ST Singer MA Miyazaki WY Milash B Dompe NA A complementary transposon tool kit for Drosophila melanogaster using P and piggyBac Nat Genet 2004 36 283 287 14981521
Kennerdell JR Carthew RW Heritable gene silencing in Drosophila using double-stranded RNA Nat Biotechnol 2000 18 896 898 10932163
Aravin AA Naumova NM Tulin AV Vagin VV Rozovsky YM Double-stranded RNA-mediated silencing of genomic tandem repeats and transposable elements in the D. melanogaster germline Curr Biol 2001 11 1017 1027 11470406
Schupbach T Wieschaus E Female sterile mutations on the second chromosome of Drosophila melanogaster . II. Mutations blocking oogenesis or altering egg morphology Genetics 1991 129 1119 1136 1783295
Gonzalez-Reyes A Elliott H St Johnston D Oocyte determination and the origin of polarity in Drosophila : The role of the spindle genes Development 1997 124 4927 4937 9362456
Cook HA Koppetsch BS Wu J Theurkauf WE The Drosophila SDE3 homolog armitage is required for oskar mRNA silencing and embryonic axis specification Cell 2004 116 817 829 15035984
Kennerdell JR Yamaguchi S Carthew RW RNAi is activated during Drosophila oocyte maturation in a manner dependent on aubergine and spindle-E Genes Dev 2002 16 1884 1889 12154120
Aravin AA Klenov MS Vagin VV Bantignies F Cavalli G Dissection of a natural RNA silencing process in the Drosophila melanogaster germ line Mol Cell Biol 2004 24 6742 6750 15254241
Lin H Yue L Spradling AC The Drosophila fusome, a germline-specific organelle, contains membrane skeletal proteins and functions in cyst formation Development 1994 120 947 956 7600970
Lin H Spradling AC Fusome asymmetry and oocyte determination in Drosophila
Dev Genet 1995 16 6 12 7758245
Deng W Lin H Spectrosomes and fusomes anchor mitotic spindles during asymmetric germ cell divisions and facilitate the formation of a polarized microtubule array for oocyte specification in Drosophila
Dev Biol 1997 189 79 94 9281339
Lasko P The Drosophila melanogaster genome: Translation factors and RNA binding proteins J Cell Biol 2000 150 F51 56 10908586
Tabara H Yigit E Siomi H Mello CC The dsRNA binding protein RDE-4 interacts with RDE-1, DCR-1, and a DexH-Box helicase to direct RNAi in C. elegans
Cell 2002 109 861 871 12110183
Han MH Goud S Song L Fedoroff N The Arabidopsis double-stranded RNA-binding protein HYL1 plays a role in microRNA-mediated gene regulation Proc Natl Acad Sci U S A 2004 101 1093 1098 14722360
Vazquez F Gasciolli V Crete P Vaucheret H The nuclear dsRNA binding protein HYL1 is required for microRNA accumulation and plant development, but not posttranscriptional transgene silencing Curr Biol 2004 14 346 351 14972688
Hiraguri A Itoh R Kondo N Nomura Y Aizawa D Specific interactions between Dicer-like proteins and HYL1/DRBfamily dsRNA-binding proteins in Arabidopsis thaliana
Plant Mol Biol 2005 57 173 188 15821876
Gatignol A Buckler-White A Berkhout B Jeang KT Characterization of a human TAR RNA-binding protein that activates the HIV-1 LTR Science 1991 251 1597 1600 2011739
Berkhout B Silverman RH Jeang KT Tat trans -activates the human immunodeficiency virus through a nascent RNA target Cell 1989 59 273 282 2478293
Dingwall C Ernberg I Gait MJ Green SM Heaphy S HIV-1 tat protein stimulates transcription by binding to a U-rich bulge in the stem of the TAR RNA structure EMBO J 1990 9 4145 4153 2249668
Marciniak RA Calnan BJ Frankel AD Sharp PA HIV-1 Tat protein trans -activates transcription in vitro Cell 1990 63 791 802 2225077
Pfeffer S Zavolan M Grasser FA Chien M Russo JJ Identification of virus-encoded microRNAs Science 2004 304 734 736 15118162
Bennasser Y Le SY Yeung ML Jeang KT HIV-1 encoded candidate micro-RNAs and their cellular targets Retrovirology 2004 1 43 15601472
Zhong J Peters AH Lee K Braun RE A double-stranded RNA binding protein required for activation of repressed messages in mammalian germ cells Nat Genet 1999 22 171 174 10369260
Bernstein E Kim SY Carmell MA Murchison EP Alcorn H Dicer is essential for mouse development Nat Genet 2003 35 215 217 14528307
Pham JW Pellino JL Lee YS Carthew RW Sontheimer EJ A Dicer-2-dependent 80s complex cleaves targeted mRNAs during RNAi in Drosophila
Cell 2004 117 83 94 15066284
Cox DN Chao A Baker J Chang L Qiao D A novel class of evolutionarily conserved genes defined by piwi are essential for stem cell self-renewal Genes Dev 1998 12 3715 3727 9851978
Cox DN Chao A Lin H piwi encodes a nucleoplasmic factor whose activity modulates the number and division rate of germline stem cells Development 2000 127 503 514 10631171
Horn C Offen N Nystedt S Hacker U Wimmer EA piggyBac-based insertional mutagenesis and enhancer detection as a tool for functional insect genomics Genetics 2003 163 647 661 12618403
Church G Gilbert W Genomic sequencing Proc Natl Acad Sci U S A 1984 81 1991 1995 6326095
Berghammer A Klingler M Wimmer E A universal marker for transgenic insects Nature 1999 402 370 371 10586872
Horn C Jaunich B Wimmer E Highly sensitive, fluorescent transformation marker for Drosophila transgenesis Dev Genes Evol 2000 210 623 629 11151299
Horn C Wimmer E A versatile vector set for animal transgenesis Dev Genes Evol 2000 210 630 637 11151300
Simon JA Sutton CA Lobell RB Glaser RL Lis JT Determinants of heat shock-induced chromosome puffing Cell 1985 40 805 817 3986904
Rubin GM Spradling AC Genetic transformation of Drosophila with transposable element vectors Science 1982 218 348 353 6289436
Kovalic D Kwak JH Weisblum B General method for direct cloning of DNA fragments generated by the polymerase chain reaction Nucleic Acids Res 1991 19 4560 1886782
Haley B Tang G Zamore PD In vitro analysis of RNA interference in Drosophila melanogaster
Methods 2003 30 330 336 12828947
Nykänen A Haley B Zamore PD ATP requirements and small interfering RNA structure in the RNA interference pathway Cell 2001 107 309 321 11701122
Pal-Bhadra M Leibovitch BA Gandhi SG Rao M Bhadra U Heterochromatic silencing and HP1 localization in Drosophila are dependent on the RNAi machinery Science 2004 303 669 672 14752161
Theurkauf WE Immunofluorescence analysis of the cytoskeleton during oogenesis and early embryogenesis Methods Cell Biol 1994 44 489 505 7707968
Lin H Spradling A A novel group of pumilio mutations affects the asymmetric division of germline stem cells in the Drosophila ovary Development 1997 124 2463 2476 9199372
| 15918770 | PMC1141267 | CC BY | 2021-01-05 08:21:24 | no | PLoS Biol. 2005 Jul 24; 3(7):e236 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030236 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1591876910.1371/journal.pbio.0030235Research ArticleBioinformatics/Computational BiologyCell BiologyDevelopmentGenetics/Genomics/Gene TherapyMolecular Biology/Structural BiologyBiochemistryDrosophilaProcessing of Pre-microRNAs by the Dicer-1–Loquacious Complex in Drosophila Cells Pre-miRNA Processing by Dicer-1-Loqs ComplexSaito Kuniaki
1
Ishizuka Akira
1
Siomi Haruhiko [email protected]
1
Siomi Mikiko C [email protected]
1
1Institute for Genome Research, University of Tokushima, Kuramoto, Tokushima, JapanCarrington James C. Academic EditorOregon State UniversityUnited States of America7 2005 24 5 2005 24 5 2005 3 7 e23516 3 2005 30 4 2005 Copyright: © 2005 Saito 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.
Normal microRNA Maturation and Germ-Line Stem Cell Maintenance Requires Loquacious, a Double-Stranded RNA-Binding Domain Protein
miRNA processing: dicer-1 meets its match
microRNAs (miRNAs) are a large family of 21- to 22-nucleotide non-coding RNAs that interact with target mRNAs at specific sites to induce cleavage of the message or inhibit translation. miRNAs are excised in a stepwise process from primary miRNA (pri-miRNA) transcripts. The Drosha-Pasha/DGCR8 complex in the nucleus cleaves pri-miRNAs to release hairpin-shaped precursor miRNAs (pre-miRNAs). These pre-miRNAs are then exported to the cytoplasm and further processed by Dicer to mature miRNAs. Here we show that Drosophila Dicer-1 interacts with Loquacious, a double-stranded RNA-binding domain protein. Depletion of Loquacious results in pre-miRNA accumulation in Drosophila S2 cells, as is the case for depletion of Dicer-1. Immuno-affinity purification experiments revealed that along with Dicer-1, Loquacious resides in a functional pre-miRNA processing complex, and stimulates and directs the specific pre-miRNA processing activity. These results support a model in which Loquacious mediates miRNA biogenesis and, thereby, the expression of genes regulated by miRNAs.
This and an accompanying paper by Förstemann et al. identify Loquacious, which encodes a double-stranded RNA binding domain protein, and partners with Dicer-1 in the processing of microRNAs.
==== Body
Introduction
microRNAs (miRNAs) act as RNA guides by binding to complementary sites on target mRNAs to regulate gene expression at the post-transcriptional level in plants and animals [1−12], much as small interfering RNAs (siRNAs) do in the RNA interference (RNAi) pathway [13−15]. The expression of miRNAs is often developmentally regulated in a tissue-specific manner, suggesting an important role for miRNAs in the regulation of endogenous gene expression [16–30]. The importance of miRNAs for development is also highlighted by a recent computer-based analysis that predicted nearly a thousand miRNA genes in the human genome [31]. Furthermore, recent studies have revealed that miRNAs regulate a large fraction of the protein-coding genes [32–34].
miRNAs are transcribed as long primary miRNA (pri-miRNA) transcripts by RNA polymerase II [35]. miRNA maturation begins with cleavage of the pri-miRNAs by the nuclear RNase III Drosha [36–38] to release approximately 70-nucleotide hairpin-shaped structures, called precursor miRNAs (pre-miRNAs). Pre-miRNAs are then exported to the cytoplasm by the protein Exportin 5, which recognizes the two-nucleotide 3′ overhang that is a signature of RNase III-mediated cleavage [39–41]. In the cytoplasm, pre-miRNAs are subsequently cleaved by a second RNase III enzyme, Dicer, into approximately 22-nucleotide miRNA duplexes, with an end structure characteristic of RNase III cleavage [42–44]. Only one of the two strands is predominantly transferred to the RNA-induced silencing complex (RISC) [45], which mediates either cleavage of the target mRNA or translation silencing, depending on the complementarity of the target [46] by a mechanism that remains unclear [47].
There is a growing list of double-stranded RNA (dsRNA)-binding proteins that play important yet distinct roles in the RNAi pathway [48]. Both Drosha and Dicer contain dsRNA-binding domains (dsRBDs). Drosha requires a dsRNA-binding protein partner known as Pasha in flies and Caenorhabditis elegans, and its ortholog DGCR8 in mammals to convert pri-miRNAs to pre-miRNAs [49–52]. In plants, the predominantly nuclear Dicer-like-1, equipped with two dsRBDs, is thought to catalyze both pri-miRNA and pre-miRNA processing [53,54]. The HYL1 protein, which also contains a tandem dsRBD, is required for miRNA accumulation and may play the same molecular role as Pasha/DGCR8 for Dicer-like-1 in plants [55,56]. In Drosophila, Dicer-2 is required for production of siRNAs [57,58], and forms a heterodimeric complex with the dsRNA-binding protein R2D2, which is required for its function in RISC assembly, although Dicer-2 alone suffices to convert long dsRNA into siRNAs [59]. Drosophila Dicer-1 is associated with the processing of pre-miRNAs [58,60]. However, if there is a dsRNA-binding protein partner for Dicer-1, it has not been identified.
Here, we show that Drosophila Dicer-1 interacts with the dsRBD protein Loquacious (Loqs). Depletion of Loqs results in accumulation of pre-miRNAs in Drosophila S2 cells. Loqs is predominantly cytoplasmic and is conserved in mammals. Immuno-affinity purification experiments, together with the use of recombinant Loqs, reveal that along with Dicer-1, Loqs resides in a functional pre-miRNA processing complex, and stimulates and directs specific pre-miRNA processing activity. These results support a model in which Loqs mediates miRNA biogenesis and, thereby, the expression of genes regulated by miRNAs.
Results
We have used RNAi-based reverse-genetic methods [61] to screen a list of Drosophila dsRBD proteins [62] for a protein(s) that has an effect on miRNA biogenesis in Drosophila S2 cells and found a novel protein equipped with three dsRBDs (two canonical dsRBDs at the N-terminal half, and one non-canonical dsRBD at the C-terminal), originally dubbed CG6866 (candidate gene 6866), which has a role in pre-miRNA processing (data presented below). This protein bears high similarity to R2D2 and to the C. elegans RNAi protein RDE-4 (Figure 1), both of which contain dsRBDs and interact with Dicer [59,63]. Thus the sequence data show that CG6866 is a paralog of R2D2. A parallel study presents genetic evidence that several types of silencing are lost in CG6866 mutant flies (Förstemann K, et al. DOI: 10.1371/journal.pbio.0030236). Therefore, CG6866 was designated as Loquacious (“very talkative”).
Figure 1 Loqs/CG6866 Is a Paralog of R2D2
A protein sequence alignment of Loqs, R2D2, and RDE-4 (C. elegans). The two canonical dsRBDs are boxed. Conserved residues are shaded in gray. It is noted that the C-terminal region of Loqs contains a non-canonical dsRBD (see Figure 8).
Depletion of Loqs and Dicer-1 by RNAi Resultsin Pre-miRNA Accumulation
Dicer-1 has been shown to be the pre-miRNA processing factor in Drosophila [58]. We have previously shown that depletion of Dicer-1 by RNAi resulted in a marked accumulation of pre-miR-bantam (pre-miR-ban) [60]. Depletion of Loqs by RNAi resulted in a similar effect to Dicer-1 depletion for miR-ban (Figure 2A). Loqs dsRNAs caused the suppression of Loqs mRNA (Figure 2B). RNAi against Loqs does not appear to affect Dicer-1 protein levels (lower panel in Figure 2B), suggesting that the observed pre-miRNA accumulation in Loqs-depleted cells is not simply due to destabilizing Dicer-1. Similar effects on miR-8 were seen in Dicer-1- and Loqs-depleted S2 cells (Figure 2C). Depletion of Dicer-2 and R2D2, which form the enzyme complex predominantly responsible for generating siRNAs from long dsRNA [59], had no significant effect on pre-miRNA processing (Figure 2A and 2C). These results show that along with Dicer-1, Loqs is essential for efficient pre-miRNA processing in vivo.
Figure 2 Depletion of Loqs and Dicer-1 Causes Pre-miRNA Accumulation
(A) dsRNAs of Loqs, Dicer-1 (Dcr-1), Dicer-2 (Dcr-2), R2D2, or EGFP were introduced to S2 cells by soaking. Total RNA was isolated 4 d after initial exposure of S2 cells to the indicated target genes. The RNA was separated on a 12% denaturing polyacrylamide gel, transferred to a nylon membrane, and probed with 5′-radiolabeled miR-bantam
(miR-ban) antisense oligodeoxynucleotide (top panel) and re-probed for U6 snRNA (bottom panel). After normalization for loading, the relative ratios for pre-miRNA (pre-miR-ban) (left panel) and mature miRNA (miR-ban) (right panel) were calculated and normalized to EGFP dsRNA experiments. Depletion of Dicer-1 (Dcr-1) by RNAi resulted in a marked accumulation of pre-miR-ban and a modest reduction in levels of mature miR-ban.
(B) Upper panel: Analysis of transcript levels using RT-PCR for Loqs and Dicer-1 (Dcr-1) after treatment of Drosophila S2 cells with dsRNAs against each protein. AGO2 was used as control. Lower panel: Loqs dsRNA did not affect Dicer-1 protein levels. dsRNAs of Loqs, Dicer-1, or EGFP were introduced to S2 cells by soaking. After 4 d, cell lysates were prepared and the levels of Dicer-1 protein were measured by Western blotting.
(C) As in (A), total RNA was probed with 5′-radiolabeled miR-8 antisense oligodeoxynucleotide (top panel) and re-probed for U6 snRNA (bottom panel).
Loqs Associates with Dicer-1 In Vivo and In Vitro
This observation prompted us to ask if Loqs forms a complex in vivo with Dicer-1. For these studies, we simultaneously expressed Dicer-1 tagged with the Flag epitope and Loqs tagged with the myc epitope in S2 cells. We then immunoprecipitated Dicer-1 with anti-Flag antibodies, and Loqs with anti-myc antibody and then analyzed the precipitates by immunoblotting (Figure 3A). In reciprocal assays, Dicer-1 and Loqs were found to co-precipitate. Consistent with these findings that Dicer-1 and Loqs form a complex in vivo, both proteins are localized predominantly in the cytoplasm of S2 cells (Figure 3B).
Figure 3 Loqs Associates with Dicer-1
(A) Loqs and Dicer-1 form a complex in vivo. Left panel: Protein extract was prepared from S2 cells that expressed both full-length myc-tagged Loqs (myc-Loqs) and Flag-tagged Dicer-1 (Flag-Dcr-1). Protein extract containing only myc-Loqs was also prepared (control). Total extracts (input) and the materials obtained after immunoprecipitation from the extracts with anti-Flag were run on an SDS-polyacrylamide gel. Western blots were prepared and immunostained with anti-myc (to detect myc-Loqs) or anti-Flag (to detect Flag–Dicer-1) antibodies. The protein band shown by an asterisk is an antibody used. Right panel: Protein extracts were prepared from S2 cells that expressed both myc-Loqs and Flag–Dicer-1, and immunoprecipitation was performed with anti-myc antibody (α-myc). Non-specific antisera were also employed as a negative control (n.i.). Flag–Dicer-1 was specifically co-immunoprecipitated with myc-Loqs.
(B) Immunofluorescence using anti-Flag antibodies show that both Loqs and Dicer-1 are predominantly localized in the cytoplasm in Drosophila S2 cells (α-Flag). Flag–Dicer-1 (Dcr-1) and Flag–Loqs were transiently expressed in the cells by transfection. The nuclear DNA was stained with propidium iodide. A DIC image of the same field is also shown (DIC).
(C) Loqs interacts with Dicer-1 in an RNase-resistant manner in vitro. 35S-labeled Dicer-1 was produced by an in vitro transcription and translation system in the presence of [35S]methionine, treated with RNaseA, and incubated with either GST–Loqs or GST itself immobilized on glutathione-Sepharose resins. After extensive washing, the bound fractions were resolved on an SDS-polyacrylamide gel and the protein labeled with 35S visualized by autoradiography. The Coomassie Blue stainings of GST and GST–Loqs used in this experiment are shown on the left.
We further investigated whether Loqs can bind to Dicer-1 in vitro. Dicer-1 was produced by an in vitro translation system and used in binding assays with recombinant Loqs fused to glutathione S-transferase (GST). GST–Loqs interacted with Dicer-1 even in the presence of RNase A, whereas GST itself showed no detectable binding (Figure 3C). These results demonstrate that the association of Loqs with Dicer-1 occurs both in vivo and in vitro, and that RNA molecules do not appear to mediate the association.
Dicer-1 and Loqs Are Present in a Functional Complex That Mediates Pre-miRNA Processing
To examine the functional connection between the Dicer-1–Loqs complex and pre-miRNA processing, we investigated if depletion of Dicer-1 or Loqs had any effect on the production of mature miRNA from the precursor. We first tested if cytoplasmic lysates of S2 cells were capable of processing synthetic Drosophila melanogaster
let-7 precursor RNA into functional mature let-7. In this experiment, the synthetic let-7 precursor RNA was converted to mature let-7 in S2 cytoplasmic lysates (Figure 4A), as is the case in embryo lysates [60]. In in vitro RNAi assay, target RNA harboring a sequence perfectly complementary to mature let-7 was cleaved efficiently within the let-7 complementary sequence (Figure 4B), thus showing production of functional let-7 in S2 cell lysates. Cytoplasmic lysates from Dicer-1- or Loqs-depleted cells were then subjected to the pre-let-7 processing assay. Both Dicer-1 and Loqs depletion led to reductions of mature let-7 compared with controls (Figure 4C), showing that both Dicer-1 and Loqs function in pre-miRNA processing.
Figure 4 Synthetic D. melanogaster
let-7 Precursor Processing in the Cytoplasmic Lysate of S2 Cells
(A) S2 cytoplasmic lysate is capable of processing pre-let-7 into mature let-7. Pre-let-7 (the sequence shown on the top) was transcribed with T7 RNA polymerase in vitro (T7-pre-let-7). After gel purification, it was incubated with S2 cytoplasmic lysate, and resultant RNA was subjected to Northern blotting using an oligodeoxynucleotide recognizing both pre-let-7 and mature let-7. Total RNA prepared from pupa was also applied (pupa), which shows where endogenous pre-let-7 (pre-let-7; ∼70 nucleotides) and mature let-7 (let-7; 21 nucleotides) migrate on the blot. The cytoplasmic lysate (+) lane, but not (−) lane, shows a band corresponding to maturelet-7, meaning that the synthetic T7-pre-let-7 was processed to the matured form in the lysate.
(B) Functional analysis of mature let-7 produced from synthetic T7-pre-let-7 in (A). Cap-labeled target RNA with let-7 target site (∼500 nucleotides) was incubated in S2 cytoplasmic lysate with or without T7-pre-let-7 for 3 h.let-7-directed cleavage product (∼150 mucleotides) was observed only when T7-pre-let-7 was included, indicating that let-7 produced in the lysate is functional.
(C) Dicer-1- or Loqs-depleted cytoplasmic lysate has less activity to produce mature let-7 from the precursor. Dicer-1 (Dcr-1) or Loqs was depleted from S2 cells by RNAi, and the cytoplasmic lysate was assayed for pre-let-7 processing as in (A). "-" indicates no lysate.
We next used pre-miR-ban as a substrate for pre-miRNA processing assays. It was shown recently that S2 cell extracts contained pri-miRNA processing activity that cleaved pri-miRNA into an approximately 60- to 70-bp pre-miRNA precursor [49]. This processing is known to occur in the nucleus; thus pre-miR-ban was prepared by in vitro processing of pri-miR-ban incubated with S2 nuclear extracts (Figure 5A). Uniformly labeled pre-miR-ban was then gel-purified and used as a substrate for analysis of pre-miRNA processing. Incubation of the pre-miRNA with S2 cytoplasmic extracts resulted in the appearance of a mature 21-nucleotide miR-ban (Figure 5B). We then examined the requirement of Dicer-1 and Loqs in pre-miR-ban processing. Incubation of pre-miRNA with Dicer-1- and Loqs-depleted S2 cytoplasmic extracts resulted in a marked reduction in mature miRNA levels (Figure 5B). In contrast, depletion of Dicer-2 or R2D2 showed no measurable reduction of mature miRNA levels (Figure 5B). We then assayed the pre-miRNA processing activity of the purified complexes (both Flag–Dicer-1 and Flag–Loqs complexes). That the Flag–Loqs complex contains Dicer-1 was confirmed by immunoblotting (data not shown). Both Dicer-1 and Loqs complexes were capable of generating maturemiR-ban from pre-miR-ban (Figure 5C). Several steps in the RNAi and miRNA pathways are known to require a divalent metal ion [64]. In addition, it is well known that RNase III-type enzymes require divalent metals for cleavage [65]. Flag–Dicer-1 complex was employed and the processing was performed in the presence of magnesium ions or EDTA in a buffer. As shown in Figure 5D, no pre-miRNA processing activity was detected at 10 mM EDTA. These results demonstrated that the Dicer-1–Loqs complex converts pre-miRNAs into mature miRNAs in a divalent metal ion–dependent manner.
Figure 5 In Vitro Processing Activities of Loqs and Dicer-1
(A) Preparation of pre-miR-ban. Uniformly labeled pri-miR-ban was incubated with S2 nuclear lysate for the processing. The resultant pre-miR-ban fragment (pre-miR-ban) was gel-purified and used as a substrate in the pre-miRNA processing assays.
(B) In vitro processing of pre-miR-ban using S2 cytoplasmic lysates. Cytoplasmic lysates were prepared 4 d after the initial exposure of S2 cells to the indicated target genes (as in Figure 2A) and used for in vitro processing. The gel-purified pre-miR-ban was incubated in cytoplasmic lysates for 1 h. “cyto. lysate” indicates parental S2 cytoplasmic lysate that shows activity for generating mature miR-ban from the gel-purified pre-miR-ban in (A).
(C) In vitro processing of pre-miR-ban using immunopurified Dicer-1 and Loqs complexes. Purified Flag–Dicer-1 (Flag-Dcr-1) and Flag–Loqs complexes were incubated with pre-miR-ban for 2 h and tested for processing activity. “−” shows the activity of a negative control prepared from parental S2 cells.
(D) The pre-miRNA processing activity of Flag–Dicer-1 complex in the presence and absence of magnesium ions. Purified Flag–Dicer-1 complex was incubated with pre-miR-ban with or without magnesium ions in buffers. Addition of EDTA caused the abolition of the activity.
Loqs Stimulates and Confers upon Dicer-1 the Specific Processing of Pre-miRNAs
To further examine the requirement for Loqs in pre-miRNA processing, we purified Flag–Dicer-1 complex under a harsher condition (high salt), where Dicer-1 was stripped of most Loqs protein (Figure 6A), and used this Dicer-1 complex in pre-miRNA processing assays with or without supplement of recombinant GST–Loqs (see left panel in Figure 3C). Without any supplement, the Flag–Dicer-1 complex purified under the harsh condition showed less activity than that under mild condition (Figure 6B). Then we added GST–Loqs in the assay mixture. The addition of GST–Loqs to the Dicer-1 complex stimulated the processing of pre-miRNA (Figure 6C). GST–Loqs alone did not show any significant pre-miRNA processing activity (Figure 6C). These results show that Loqs is required for stimulating the processing of pre-miRNAs. Interestingly, we found that the Dicer-1 complex purified under the harsh condition displayed considerable siRNA-generating activity on the long dsRNA substrate in vitro (Figure 6D), although previous genetic studies have shown that Dicer-1 is not required for siRNA production [58]. The addition of GST–Loqs inhibited this effect (Figure 6D). Western blot analysis failed to show that the Dicer-1 complex used in this experiment contains Dicer-2 (right panel inFigure 6D). GST–Loqs alone showed no activity for generating siRNAs from long dsRNAs. These results suggested that Dicer-1 stripped of much of its bound Loqs processes both dsRNA and pre-miRNA substrates, but re-addition of recombinant Loqs suppresses dsRNA processing activity and enhances pre-miRNA processing activity. Our findings thus imply that much of the apparent substrate specificity of Dicer-1 in vivo results from its association with Loqs. Although very unlikely (Figure 6D), it is, however, formally possible that the Dicer-1 immunoprecipitates may contain very small amounts of Dicer-2 protein that can catalyze long dsRNA cleavage, and that addition of a large amount of dsRBD-containing Loqs may block the activity of Dicer-2 in this experiment.
Figure 6 Loqs Stimulates the Specific Processing of Pre-miRNA by Dicer-1
(A) Flag–Dicer-1 complex was purified under a harsh condition. Protein extract was prepared from S2 cells that expressed both full-length myc-tagged Loqs (myc-Loqs) and Flag-tagged Dicer-1 (Flag-Dcr-1). The amounts of Loqs in Flag–Dicer-1 complexes prepared under high-salt condition and low-salt condition were examined by Western blotting using anti-myc antibody, which show that less Loqs was co-purified with Dicer-1 in the high-salt condition.
(B) The miRNA processing activities of Flag–Dicer-1 (Flag-Dcr-1) complexes in (A). Flag–Dicer-1 complex containing less Loqs showed a lower activity for the processing.
(C) Recombinant Loqs stimulates the in vitro processing of pre-miR-ban by Flag–Dicer-1 complex purified in high-salt condition. 100 ng of purified GST or GST–Loqs (see Figure 3C) were supplemented for the processing activity by Flag–Dicer-1 complex. GST–Loqs by itself does not show any pre-miRNA processing activity.
(D) GST–Loqs inhibits the siRNA-generating activity of Dicer-1. Uniformly labeled long dsRNA was incubated with Flag–Dicer-1 (Flag-Dcr-1) complex purified in high-salt condition in (A) with or without GST–Loqs. The Flag–Dicer-1 complex by itself showed a considerable activity of generating siRNA from long dsRNA. Addition of GST–Loqs, but not GST, inhibited the processing. Note that the Flag–Dicer-1 (Flag-Dcr-1) complex does not contain Dicer-2 (Dcr-2), judged by Western blot analysis using anti-Dicer-2 antibodies (right panel).
Dicer-1–Loqs Complexes Associate with Pre- and Mature miRNAs In Vivo
We examined the presence of endogenous miRNA in RNA preparations from Flag–Dicer-1 and Flag–Loqs complexes obtained from S2 cells using anti-Flag antibodies. Complexes were prepared as in Figure 3A, and RNA preparations from each complex were subjected to Northern blotting using an oligo probe recognizing both pre-miR-ban and mature miR-ban. The Dicer-1 complex contained both the pre- and mature form of miR-ban, and the complex seems to preferentially bind the precursor form of miR-ban (Figure 7A). In contrast, the precursor form of miR-ban was barely detectable in the Loqs complex, though it contained mature miR-ban. However, EDTA treatment, which inhibits pre-miRNA processing activity (see Figure 5D), resulted in an accumulation of pre-miR-ban in the Loqs complex (Figure 7A). This may suggest that part of Flag-tagged Loqs protein interacts with Dicer-1 or pre-miRNAs or both. Alternatively, Flag–Loqs complexes may rapidly process pre-miRNAs into mature miRNAs and, therefore, may only transiently interact with them. Nonetheless, these results suggest that Dicer-1–Loqs complexes associate with both pre- and mature miRNAs in vivo.
Figure 7 Dicer-1–Loqs Complexes Are Associated with Pre-miRNA and Mature miRNA In Vivo
(A) Northern blot analyses show that Flag–Dicer-1 (Flag-Dcr-1) complex contains both pre- and mature form of miR-ban. Notably, the precursor form of miR-ban was accumulated in the Dicer-1 complex. In the case of the Loqs complex, the precursor apparently accumulated within the complex when the Flag–Loqs complex was prepared in the presence of EDTA that was shown to inhibit the pre-miRNA processing activity in Figure 5D.
(B) AGO1 associates with Dicer-1 and Loqs. IgG-bound fractions prepared from S2 cells expressing AGO1–TAP, EGFP–TAP, or the parental S2 cells (−), were subjected to Western blotting using antibodies against Dicer-1, AGO1, and Flag (for Loqs). Dicer-1 is not present in AGO2-associated complex.
(C) In vitro processing of pre-miR-ban using affinity-purified AGO1 complexes. Purified AGO1–TAP or AGO2–TAP complexes were incubated with pre-miR-ban (precursor) and tested for processing activity. “−” shows the activity of a negative control prepared from parental S2 cells.
An AGO1-Associated Complex Contains Dicer-1 and Loqs, and Is Capable of Pre-miRNA Processing
We have previously shown that Argonaute protein AGO1 is required for stable production of mature miRNAs and associates with Dicer-1 [60]. Thus, we sought to ascertain if Loqs was also present in an AGO1-associated complex, and if so, if the AGO1 complex was capable of processing pre-miRNA in vitro. We simultaneously expressed Flag–Loqs and AGO1 tagged with TAP in S2 cells and purified the AGO1–TAP complex through immunoglobulin G (IgG) bead-binding. The IgG bound was then subjected to Western blot analysis using anti-Dicer-1, anti-AGO1, or anti-Flag (for Loqs detection) antibodies. Not only Dicer-1 but also Loqs was detected in the AGO1 complex (Figure 7B). These results indicated that all three proteins are present in the same complex, although they cannot exclude the possibility that there is one complex that contains AGO1 and Dicer-1 but not Loqs, and another complex that contains AGO1 and Loqs but not Dicer-1. The pre-miRNA processing activity of the AGO1 complex was then examined. As in Figure 5, pre-miR-ban was utilized as a substrate. The AGO1 complex was able to efficiently process pre-miR-ban into the mature form (Figure 7C). In contrast, another Argonaute protein AGO2-associated complex showed no such activity, which is consistent with our previous finding that the AGO2-associated complex does not contain Dicer-1 [60]. Considered together, these results showed that Dicer-1 and Loqs form a functional complex that mediates the genesis of mature miRNAs from pre-miRNAs, and suggested that the resultant mature miRNAs are loaded onto an AGO1-associated complex, which probably is miRNA-associated RISC [60], through specific interaction of AGO1 with Dicer-1 and Loqs.
Discussion
Our results indicate that Loqs and Dicer-1 form a complex that converts pre-miRNAs into mature miRNAs; so how do they act together in pre-miRNA processing? Sequence comparison reveals that Loqs is a paralog of R2D2 (see Figure 1). Therefore, Loqs may play the molecular role of R2D2 for Dicer-1. R2D2 forms a stable heterodimeric complex with Dicer-2, while either protein alone seems to be unstable in vivo [59]. In the absence of R2D2, Dicer-2 is still capable of efficiently processing long dsRNA into siRNAs. Therefore, the siRNA generating activity of Dicer-2 is not dependent upon R2D2. However, the resultant siRNAs are not effectively channeled into RISC in the absence of R2D2. The Dicer-2–R2D2 complex, but not Dicer-2 alone, binds to siRNA, which indicates that siRNA binding by the heterodimer is important for RISC entry [59,66]. In the case of Loqs, this protein alone is not capable of converting pre-miRNAs into mature miRNAs, but it clearly stimulates and directs the specific pre-miRNA processing activity of Dicer-1. Furthermore, knocking down Loqs markedly reduced the pre-miRNA processing activity in cytoplasmic lysates in vitro (see Figures 4C and 5B), but did not cause a significant reduction of the level of Dicer-1 protein (see Figure 2B); implying that Dicer-1 may largely depend on Loqs for its pre-miRNA processing activity. Thus, the molecular role of Loqs for Dicer-1 is not simply similar to that of R2D2 for Dicer-2.
It can be envisioned that Loqs may have one of several roles in pre-miRNA processing. Dicer-1 contains only one dsRBD, which may not be sufficient for strong interaction with and/or specific recognition of the pre-miRNA substrate (see Figure 6C and 6D). Loqs, containing three dsRBDs with no other identifiable domains being apparent, could provide the additional RNA-binding modules required for specific recognition of the pre-miRNA, and thereby stabilize pre-miRNA binding for Dicer-1. Loqs could also organize binding of Dicer-1 on the pre-miRNA, contributing to the specific positioning of the Dicer-1 cleavage site. Alternatively, since dsRBDs are known to not only bind dsRNAs but also mediate protein–protein interactions [67], Loqs may directly bind Dicer-1 through its dsRBDs. This protein–protein interaction may trigger a conformational change of Dicer-1 that facilitates either the formation of an intramolecular dimer of its two RNase III domains [50,68], which creates a pair of catalytic sites, or the handover of the Dicer-1 cleaved mature miRNAs to the RISC.
Sequence analysis revealed that protein activator of protein kinase dsRNA dependent (PKR) (PACT) [69] and HIV TAR RNA binding protein (TRBP) [70] in mammals bear 34% identity to Loqs, and share a highly similar domain structure with it (Figure 8). Both PACT and TRBP are thought to play a role in the regulation of translation through modulating PKR that also contains two dsRBDs [71–73]. PACT interacts with PKR and enhances the autophosphorylation of PKR [67], which in turn, phosphorylates the α subunit of eukaryotic translation initiation factor 2 (eIF2α) and leads to an inhibition of mRNA translation in response to viral infection and other stimuli. TRBP prevents PKR-mediated inhibition of protein synthesis through binding to PKR [74]. Considered together, it will be important to find out Loqs' partners other than Dicer-1 for possible involvement of Loqs in miRNA-mediated translational regulation in Drosophila.
Figure 8 Domain Structure of Loqs and ItsHomo sapiens and Xenopus Homologs
The dsRNA-binding motif (dsRBD) is indicated as a black box. These proteins contain three putative dsRBDs. Loqs shares ∼34% amino acid identity with TRBP and PACT and ∼31% identity with Xlrbpa (Xenopus laevis RNA-binding protein A). Sequence comparison between Loqs and its human and Xenopus homologs also showed a higher degree of amino acid conservation in dsRBDs including C-terminal non-canonical dsRBDs. TheXenopus homolog, Xlrbpa, of TRBP/PACT has been found to associate with ribosomes in the cytoplasm [77], as is the case for many RNAi factors including miRNAs [47,78−82].
Materials and Methods
RNAi
dsRNAs were introduced to S2 cells by soaking essentially as described [75]. Briefly, approximately 5 × 106 cells were soaked in 1 ml of serum-free medium containing 15 μg of dsRNA for 30 min at room temperature followed by addition of 2 ml of the medium containing 15% serum, 3 mM glutamine, and penicillin-streptomycin. After 4 d, cells were harvested and subjected to total RNA preparation for Northern blot analysis, or cytoplasmic lysate preparation for in vitro processing assays. dsRNAs used in RNAi were: double-stranded RNA for enhanced green fluorescent protein (EGFP), homologous to nucleotides 11–717 of the EGFP coding sequence; dsDcr-2, 4091–4888 of the Dicer-2 coding; dsR2D2, 1–936; dsDcr-1, 10–950 of the Dicer-2 coding; dsLoqs, 330-1342.
Northern blot analysis
Total RNA was isolated from S2 cells with ISOGEN (Nippon Gene, Toyama, Japan). 20 μg of total RNA was separated on 12% acrylamide-denaturing gel and transferred onto Hybond-N+ membrane (Amersham Bioscience, Little Chalfont, United Kingdom). After UV-crosslinking, the hybridization was performed at 42 °C in 0.2 M sodium phosphate (pH 7.2), 7 % SDS, and 1 mM EDTA with end-labeled antisense oligodeoxynucleotide, and washed at 42 °C in 2× saline sodium citrate and 0.1% SDS. Oligodeoxynucleotides used as probes were: bantam, 5′-
CAGCTTTCAAAATGATCTCAC-3′; miR-8, 5′-
GACATCTTTACCTGACAGTATTA-3′; U6 snRNA, 5′-
GGGCCATGCTAATCTTCTCTGTA-3′; and let-7, 5′-
AACTATACAACCTACTACCTCA-3.′ The blots were exposed on BAS-MS2040 imaging plates, and signals were quantified using BAS-2500 (Fuji, Tokyo, Japan).
RT-PCR analysis
One μg of total RNA was used for the first-strand cDNA synthesis with Stratascript RT and random primers (Stratagene, La Jolla, California, United States). Sequences of the oligonucleotide primers for RT-PCR were: Dicer-1, 5′-
ACCAATGTACTGCGTTTGCA and 5′-
GTTTGCTGATCACAGAACTTAACGTT; Loqs, 5′-
ATGGACCAGGAGAATTTCCACGG-3′ and 5′-
CTACTTCTTGGTCATGATCTTCAAGTAC-3′; and AGO2, 5′-
GCACAAGTGTGCGGTCTTGTATT-3′ and 5′-
GTGAACTGCTTAATGCATTG-3′.
Immunofluorescence analysis
Immunofluorescence analysis was performed by fixing S2 cells with 2% formaldehyde for 15 min. Cells were permeabilized using 0.1% Triton X-100. Flag-tagged proteins were stained for 30 min with anti-Flag M2 (1:1,000 dilution) antibody (Sigma, St. Louis, Missouri, United States). After extensive wash in PBS, cells were treated with 100 μg/ml RNaseA for 30 min and then stained with 0.4 μg/ml propidium iodide. Alexa-488 anti-mouse IgG was used as secondary antibody. All images were collected using a Zeiss (Oberkochen, Germany) LSM510 laser scanning microscope.
Immunoprecipitation of Flag–Dicer-1, Flag–Loqs, and myc-Loqs
S2 cell lines, stably expressed 3× Flag-tagged Dicer-1 or Loqs, or myc-tagged-Loqs under the control of metallothionein promoter (originally from pRmHa-3 vector), were established. The expression of each protein was induced by adding copper ions into the medium. After overnight incubation, the whole cell extract was prepared in Buffer A (30 mM HEPES pH 7.4, 150 mM KOAc, 2 mM MgOAc, 5 mM DTT, 2 μg/ml Leupeptin, 2 μg/ml Pepstatin, 0.5% Aprotinin) containing 0.1% NP-40 by sonication, and followed by centrifugation. Flag–Dicer-1 and Flag–Loqs were bound to anti-Flag M2 agarose beads at 4 °C for 1 h. Immunoprecipitated proteins were then recovered with 2× SDS sample buffer or Elution buffer (Buffer A containing 400 μg/ml of 3× Flag peptides, 10% glycerol, and 100 mM KOAc). For Northern blot analysis of RNAs co-purified with Flag–Dicer-1 or Flag–Loqs, immunoprecipitates on beads were treated with ISOGEN and subjected to RNA purification.
Protein–protein interaction assays
To produce [35S]methionine-labeled proteins by a TNT in vitro transcription and translation kit (Promega, Madison, Wisconsin, United States), the Dicer-1 cDNA was inserted into an expression vector, pET-28 (Novagen, Madison, Wisconsin, United States). GST pull-down assays were carried out using GST–Loqs and GST itself that were bound to glutathione Sepharose 4B resins (Amersham Biosciences) in Buffer A containing 0.1% NP-40. After incubation with TNT products and extensive washing, the bound proteins were separated by SDS-PAGE. RNaseA treatment was carried out by adding the enzyme to the binding mixture. To produce GST fusion protein, Loqs cDNA was subcloned into a pGEX-5X expression vector (Amersham Biosciences). The fusion proteins, as well as GST itself, were induced and purified as described by the manufacturer.
Preparation of nuclear and cytoplasmic lysate for in vitro processing assays
S2 cells were suspended at approximately 1 × 108 cells/ml into Hypotonic buffer (Buffer A without KOAc) and lysed by passing through a 30G needle. After centrifugation at 500 × g for 20 min, the supernatant and the precipitate were separated. The supernatant was centrifuged to obtain the supernatant as a cytoplasmic lysate. The pellet was washed twice with Hypotonic buffer and lysed by sonication in Buffer A containing 100 mM KOAc and 20% glycerol, followed by centrifugation to obtain the supernatant as a nuclear lysate. Total protein concentration in each lysate was determined with Protein assay (Bio-Rad, Hercules, California, United States) and adjusted to be equal.
Preparation of pre-miR-ban
A DNA fragment coding pri-miR-ban was obtained from PCR reaction (primers used are: 5′-
CGCTCAGATGCAGATGTTGTTGAT-3′ and 5′-
GATCGGTCGGCATAAG
TTCAAAGC-3′) and cloned into the SmaI site of pBluescriptSK vector in the same direction with the T3 promoter. The plasmid was digested with ClaI, gel-isolated, and used as a template for in vitro transcription reaction with MEGAscript T3 Kit (Ambion, Austin, Texas, United States) in the presence of [α32P]GTP. In vitro processing reaction of pri-miRNAs was performed with some modifications to previously reported method [36]. Briefly, in 1 ml reaction, 10 mM creatine phosphate, 0.5 mM ATP, 30 μg/ml creatine kinase, 0.1 U/ul RNasin, 0.1 μg yeast RNA, and 500 μl nuclear lysate were added, and pri-miR-bantam in 0.5× Buffer A with 100 mM KOAc was further added to the mixture. After 2 h incubation at 26 °C, RNAs were purified with ISOGEN LS (Nippon Gene) and separated on 7.5% acrylamide denaturing gel, from which pre-miR-ban (about 60 nucleotides in length) was recovered.
In Vitro pre-miRNA processing assays
The condition used for in vitro pre-miRNA processing with cytoplasmic lysates was the same as that for the in vitro pri-miR-bantam processing. Cytoplasmic lysate used in this assay was 5 μl in a 10-μl reaction. For processing assays with purified complexes, immuno-purified Flag–Dicer-1 or Flag–Loqs was used instead of crude cytoplasmic lysate and the final concentration of buffer adjusted. For Mg++-depletion assay, 10 mM EDTA was added instead of Mg++. For the processing by Flag–Dicer-1, high-salt purified (800 mM KOAc) Flag–Dicer-1 was added in the presence or absence of bacterially produced GST–Loqs and the final concentration of buffer adjusted.
In Vitro cleavage assay
Preparation of cap-labeled ftz RNA with a let-7 target site and RNAi reaction were carried out essentially as described [60]. In brief, 104 cpm of cap-labeled let-7 target RNA was incubated with 200 nM in vitro transcribed pre-let-7 RNA in Buffer A containing 100 mM KOAc, 10 mM creatine phosphate, 0.5 mM ATP, 30 μg/ml creatine kinase, and 0.1 U/ul RNasin. Reactions were allowed to proceed for 3 h at 26 °C. Cleavage products of the RNAi reaction were analyzed by electrophoresis on 4% denaturing polyacrylamide gels.
TAP purification
The expression of AGO1–TAP or AGO2–TAP in S2 cells was induced by adding copper ion into the medium [60]. After overnight incubation, the cytoplasmic lysate was prepared in a Buffer A containing 150 mM KOAc. AGO1–TAP or AGO2–TAP and associated materials to the TAP-tagged fusion protein were bound to IgG Sepharose (Amersham Biosciences). Bound proteins on IgG beads were directly used for in vitro pre-miRNA processing assay, or eluted with SDS sample buffer for Western blotting analysis. The polyclonal antibodies against AGO1 were a kind gift from T. Uemura (Kyoto University) [76]. The anti-Dicer-1 antibody (AB4735) was purchased from Abcam (Cambridge, United Kingdom).
Supporting Information
Accession Numbers
The GenBank (http://www.ncbi.nih.gov/Genbank/) accession numbers for the genes and gene products discussed in this paper are: ago-1 (NM_079010), ago-2 (NM_140518), dicer-1 (NM_079729), dicer-2 (NM_079054), loqs/cg6866 (NM_135802), and r2d2 (NM_135308).
The Rfam (http://www.sanger.ac.uk/Software/Rfam/mirna/index.shtml) accession numbers for the genes and gene products discussed in this paper are: bantam (MI0000387), let-7 (MI0000416), and miR-8 (MI0000128).
We thank members of the Siomi laboratory for discussions and comments on the manuscript. KS was supported in part by a fellowship from the Fragile X Research Foundation (FRAXA). AI is a predoctoral fellow (DC1) of the Japan Society for the Promotion of Science (JSPS). This work was supported by grants to MCS and HS from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (MEXT) and the JSPS.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. KS, AI, HS, and MCS conceived and designed the experiments. KS, AI, and MCS performed the experiments. KS, AI, HS, and MCS analyzed the data. HS, KS, AI, and MCS wrote the paper.
Citation: Saito K, Ishizuka A, Siomi H, Siomi MC (2005) Processing of pre-microRNAs by the Dicer-1-loquacious complex in Drosophila cells. PLoS Biol 3(7): e235.
Abbreviations
dsRBDdouble-stranded RNA-binding domain
dsRNAdouble-stranded RNA
EGFPenhanced green fluorescent protein
GSTglutathione S-transferase
IgGimmunoglobulin G
miRNAmicroRNA
PACTprotein activator of PKR
PKRprotein kinase dsRNA dependent
pre-miRNAprecursor miRNA
pri-miRNAprimary miRNA
RISCRNA-induced silencing complex
RNAiRNA interference
siRNAsmall interfering RNA
TRBPTAR RNA binding protein.
==== Refs
References
Lee RC Feinbaum RL Ambros V The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14 Cell 1993 75 843 854 8252621
Reinhart BJ Slack FJ Basson M Pasquinelli AE Bettinger JC The 21-nucleotide let-7 RNA regulates developmental timing in Caenorhabditis elegans
Nature 2000 403 901 906 10706289
Llave C Xie Z Kasschau KD Carrington JC Cleavage of Scarecrow-like mRNA targets directed by a class of Arabidopsis miRNA Science 2002 297 2053 2056 12242443
Brennecke J Hipfner DR Stark A Russell RB Cohen SM
bantam encodes a developmentally regulated microRNA that controls cell proliferation and regulates the proapoptotic gene hid in Drosophila
Cell 2003 113 25 36 12679032
Palatnik JF Allen E Wu X Schommer C Schwab R Control of leaf morphogenesis by microRNAs Nature 2003 425 257 263 12931144
Xu P Vernooy SY Guo M Hay BA The Drosophila microRNA Mir-14 suppresses cell death and is required for normal fat metabolism Curr Biol 2003 13 790 795 12725740
Johnston RJ Hobert O A microRNA controlling left/right neuronal asymmetry inCaenorhabditis elegans
Nature 2003 426 845 849 14685240
Chen CZ Li L Lodish HF Bartel DP MicroRNAs modulate hematopoietic lineage differentiation Science 2004 303 83 86 14657504
Yekta S Shih IH Bartel DP MicroRNA-directed cleavage of HOXB8 mRNA Science 2004 304 594 596 15105502
Carrington JC Ambros V Role of microRNAs in plant and animal development Science 2003 301 336 338 12869753
Bartel DP MicroRNAs: Genomics, biogenesis, mechanism, and function Cell 2004 116 281 297 14744438
Baulcombe D RNA silencing in plants Nature 2004 431 356 363 15372043
Meister G Tuschl T Mechanisms of gene silencing by double-stranded RNA Nature 2004 431 343 349 15372041
Sontheimer EJ Assembly and function of RNA silencing complexes Nat Rev Mol Cell Biol 2005 6 127 138 15654322
Tomari Y Zamore PD Perspective: Machines for RNAi Genes Dev 2005 19 517 529 15741316
Pasquinelli AE Reinhart BJ Slack F Martindale MQ Kuroda MI Conservation of the sequence and temporal expression of let-7 heterochronic regulatory RNA Nature 2000 408 86 89 11081512
Lagos-Quintana M Rauhut R Lendeckel W Tuschl T Identification of novel genes coding for small expressed RNAs Science 2001 294 853 858 11679670
Lagos-Quintana M Rauhut R Yalcin A Meyer J Lendeckel W Tuschl T Identification of tissue-specific microRNAs from mouse Curr Biol 2002 12 735 739 12007417
Lagos-Quintana M Rauhut R Meyer J Borkhardt A Tuschl T New microRNAs from mouse and human RNA 2003 9 175 179 12554859
Lau NC Lim LP Weinstein EG Bartel DP An abundant class of tiny RNAs with probable regulatory roles in Caenorhabditis elegans
Science 2001 294 858 862 11679671
Lee RC Ambros V An extensive class of small RNAs in Caenorhabditis elegans
Science 2001 294 862 864 11679672
Mourelatos Z Dostie J Paushkin S Sharma A Charroux B miRNPs: A novel class of ribonucleoproteins containing numerous microRNAs Genes Dev 2002 16 720 728 11914277
Reinhart BJ Weinstein EG Rhoades MW Bartel B Bartel DP MicroRNAs in plants Genes Dev 2002 16 1616 1626 12101121
Ambros V Lee RC Lavanway A Williams PT Jewell D MicroRNAs and other tiny endogenous RNAs in C. elegans
Curr Biol 2003 13 807 818 12747828
Aravin AA Lagos-Quintana M Yalcin A Zavolan M Marks D The small RNA profile during Drosophila melanogaster development Dev Cell 2003 5 337 350 12919683
Dostie J Mourelatos Z Yang M Sharma A Dreyfuss G Numerous microRNPs in neuronal cells containing novel microRNAs RNA 2003 9 180 186 12554860
Grad Y Aach J Hayes GD Reinhart BJ Church GM Computational and experimental identification of C. elegans microRNAs Mol Cell 2003 11 1253 1263 12769849
Houbaviy HB Murray MF Sharp PA Embryonic stem cell-specific MicroRNAs Dev Cell 2003 5 351 358 12919684
Lai EC Tomancak P Williams RW Rubin GM Computational identification of Drosophila microRNA genes Genome Biol 2003 4 R42 12844358
Lim LP Lau NC Weinstein EG Abdelhakim A Yekta S The microRNAs of Caenorhabditis elegans
Genes Dev 2003 17 991 1008 12672692
Berezikov E Guryev V van de Belt J Wienholds E Plasterk RHA Phylogenetic shadowing and computational identification of human microRNA genes Cell 2005 120 21 24 15652478
John B Enright AJ Aravin A Tuschl T Sander C Human microRNAs targets PLoS Biol 2004 2 e363 15502875
Brennecke J Stark A Russell RB Cohen SM Principles of microRNA-target recognition PLoS Biol 2005 3 e85 15723116
Lim LP Lau NC Garrett-Engele P Grimson A Schelter JM Microarray analysis shows that some microRNAs downregulate large numbers of target mRNAs Nature 2005 433 769 773 15685193
Lee Y Kim M Han J Yeom KH Lee S MicroRNA genes are transcribed by RNA polymerase II EMBO J 2004 23 4051 4060 15372072
Lee Y Jeon K Lee JT Kim S Kim VN MicroRNA maturation: Stepwise processing and subcellular localization EMBO J 2002 21 4663 4670 12198168
Lee Y Ahn C Han J Choi H Kim J The nuclear RNase III Drosha initiates microRNA processing Nature 2003 425 415 419 14508493
Basyuk E Suavet F Doglio A Bordonne R Bertrand E Human let-7 stem-loop precursors harbor features of RNase III cleavage products Nucleic Acids Res 2003 31 6593 6597 14602919
Yi R Qin Y Macara IG Cullen BR Exportin-5 mediates the nuclear export of pre-microRNAs and short hairpin RNAs Genes Dev 2003 17 3011 3016 14681208
Bohnsack MT Czaplinski K Gorlich D Exportin 5 is a RanGTP-dependent dsRNA-binding protein that mediates nuclear export of pre-miRNAs RNA 2004 10 185 191 14730017
Lund E Guttinger S Calado A Dahlberg JE Kutay U Nuclear export of microRNA precursors Science 2004 303 95 98 14631048
Hutvagner G McLachlan J Pasquinelli AE Balint E Tuschl T A cellular function for the RNA-interference enzyme Dicer in the maturation of the let-7 small temporal RNA Science 2001 293 834 838 11452083
Grishok A Pasquinelli AE Conte D Li N Parrish S Genes and mechanisms related to RNA interference regulate expression of the small temporal RNAs that control C. elegans developmental timing Cell 2001 106 23 34 11461699
Ketting RF Fischer SE Bernstein E Sijen T Hannon GJ Dicer functions in RNA interference and in synthesis of small RNA involved in developmental timing in C. elegans
Genes Dev 2001 15 2654 2659 11641272
Schwarz DS Hutvagner G Du T Xu Z Aronin N Asymmetry in the assembly of the RNAi enzyme complex Cell 2003 115 199 208 14567917
Hutvagner G Zamore PD A microRNA in a multiple-turnover RNAi enzyme complex Science 2002 297 2056 2060 12154197
Olsen PH Ambros V The lin-4 regulatory RNA controls developmental timing in Caenorhabditis elegans by blocking LIN-14 protein synthesis after the initiation of translation Dev Biol 1999 216 671 680 10642801
Tomari Y Zamore PD MicroRNA biogenesis: Drosha can't cut it without a partner Curr Biol 2005 15 R61 64 15668159
Denli AM Tops BB Plasterk RH Ketting RF Hannon GJ Processing of primary microRNAs by the Microprocessor complex Nature 2004 432 231 235 15531879
Gregory RI Yan KP Amuthan G Chendrimada T Doratotaj B The Microprocessor complex mediates the genesis of microRNAs Nature 2004 432 235 240 15531877
Han J Lee Y Yeom KH Kim YK Jin H The Drosh-DGCR8 complex in primary microRNA processing Genes Dev 2005 18 3016 3027
Landthaler M Yalcin A Tuschl T The human DiGeorge syndrome critical region gene 8 and its D. melanogaster homolog are required for miRNA biogenesis Curr Biol 2004 15 2162 2167
Papp I Mette MF Aufsatz W Daxinger L Schauer SE Evidence for nuclear processing of plant micro RNA and short interfering RNA precursors Plant Physiol 2003 132 1382 1390 12857820
Kurihara Y Watanabe Y
Arabidopsis micro-RNA biogenesis through Dicer-like 1 protein functions Proc Natl Acad Sci U S A 2004 101 12753 12758 15314213
Vazquez F Gasciolli V Crete P Vaucheret H The nuclear dsRNA binding protein HYL1 is required for microRNA accumulation and plant development, but not posttranscriptional transgene silencing Curr Biol 2004 14 346 351 14972688
Han MH Goud S Song L Fedoroff N The Arabidopsis double-stranded RNA-binding protein HYL1 plays a role in microRNA-mediated gene regulation Proc Natl Acad Sci U S A 2004 101 1093 1098 14722360
Bernstein E Caudy AA Hammond SM Hannon GJ Role for a bidentate ribonuclease in the initiation step of RNA interference Nature 2001 409 363 366 11201747
Lee YS Nakahara K Pham JW Kim K He Z Distinct roles for Drosophila Dicer-1 and Dicer-2 in the siRNA/miRNA silencing pathways Cell 2004 117 69 81 15066283
Liu Q Rand TA Kalidas S Du F Kim HE R2D2, a bridge between the initiation and effector steps of the Drosophila RNAi pathway Science 2003 301 1921 1925 14512631
Okamura K Ishizuka A Siomi H Siomi MC Distinct roles for Argonaute proteins in small RNA-directed cleavage pathways Genes Dev 2004 18 1655 1666 15231716
Boutros M Kiger AA Armknecht S Kerr K Hild M Genome-wide RNAi analysis of growth and viability in Drosophila cells Science 2004 303 832 835 14764878
Lasko P The Drosophila melanogaster genome: Translation factors and RNA binding proteins J Cell Biol 2000 150 F51 F56 10908586
Tabara H Yigit E Siomi H Mello CC The dsRNA binding protein RDE-4 interacts with RDE-1, DCR-1, and a DExH-box helicase to direct RNAi in C. elegans
Cell 2002 109 861 871 12110183
Schwarz DS Tomari Y Zamore PD The RNA-induced silencing complex is a Mg2+-dependent endonuclease Curr Biol 2004 14 787 791 15120070
Nicholson AW Hannon GJ The ribonuclease III superfamily: Forms and functions in RNA, maturation, decay, and gene silencing RNAi: A guide to gene silencing 2003 Cold Spring Harbor Laboratory Press, Plainview (New York) 149 174
Tomari Y Matranga C Haley B Martinez N Zamore PD A protein sensor for siRNA asymmetry Science 2004 306 1377 1380 15550672
Peters GA Hartmann R Qin J Sen GC Modular structure of PACT: Distinct domains for binding and activating PKR Mol Cell Biol 2001 21 1908 1920 11238927
Zhang H Kolb F Jaskiewicz L Westhof E Filipowicz W Single processing center models for human Dicer and bacterial RNase III Cell 2004 118 57 68 15242644
Patel RC Sen GC PACT, a protein activator for the interferon-induced protein kinase, PKR EMBO J 1998 17 4379 4390 9687506
Gatignol A Buckler-White A Berkhout B Jeang KT Characterization of a human TAR RNA-binding protein that activates the HIV-1 LTR Science 1991 251 1597 1600 2011739
Fierro-Monti I Mathews MB Proteins binding to duplexed RNA: One motif, multiple functions Trends Biochem Sci 2000 25 241 246 10782096
Kaempfer R RNA sensors: Novel regulators of gene expression EMBO Rep 2003 4 1043 1047 14593443
Saunders LR Barber GN The dsRNA binding protein family: Critical roles, diverse cellular functions FASEB J 2003 17 961 983 12773480
Park H Davies MV Langland JO Chang HW Nam YS TAR RNA-binding protein is an inhibitor of the interferon-induced protein kinase PKR Proc Natl Acad Sci U S A 1994 91 4713 4717 7515177
Clemens JC Worby CA Simonson-Leff N Muda M Maehama T Use of double-stranded RNA interference in Drosophila cell lines to dissect signal transduction pathways Proc Natl Acad Sci U S A 2000 97 6499 6503 10823906
Kataoka Y Takeichi M Uemura T Developmental roles and molecular characterization of a Drosophila homologue of Arabidopsis Argonaute1 , the founder of a novel gene superfamily Genes Cells 2001 6 313 325 11318874
Eckmann CR Jantsch MF Xlrbpa, a double-stranded RNA-binding protein associated with ribosomes and heterogeneous nuclear RNPs J Cell Biol 1997 138 239 253 9230068
Hammond SM Boettcher S Caudy AA Kobayashi R Hannon GJ Argonaute2, a link between genetic and biochemical analyses of RNAi Science 2001 293 1146 1150 11498593
Ishizuka A Siomi MC Siomi H A Drosophila fragile X protein interacts with components of RNAi and ribosomal proteins Genes Dev 2002 16 2497 2508 12368261
Djikeng A Shi H Tschudi C Shen S Ullu E An siRNA ribonucleoprotein is found associated with polyribosomes in Trypanosoma brucei
RNA 2003 9 802 808 12810914
Kim J Krichevsky A Grad Y Hayes GD Kosik KS Identification of many microRNAs that copurify with polyribosomes in mammalian neurons Proc Natl Acad Sci U S A 2004 101 360 365 14691248
Pham JW Pellino JL Lee YS Carthew RW Sontheimer EJ A Dicer-2-dependent 80s complex cleaves targeted mRNAs during RNAi in Drosophila
Cell 2004 117 83 94 15066284
| 15918769 | PMC1141268 | CC BY | 2021-01-05 08:28:15 | no | PLoS Biol. 2005 Jul 24; 3(7):e235 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030235 | oa_comm |
==== Front
BMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 1472-6882-5-101588246810.1186/1472-6882-5-10Research ArticleElectrical impedance along connective tissue planes associated with acupuncture meridians Ahn Andrew C [email protected] Junru [email protected] Gary J [email protected] Richard [email protected] Helene M [email protected] Division for Research and Education in Complementary and Integrative Medical Therapies, Harvard Medical School, Boston, MA, USA2 Departments of Physics, University of Vermont, Burlington, VT, USA3 Department of Medical Biostatistics, University of Vermont, Burlington, VT, USA4 Research Department, Oregon College of Oriental Medicine, Portland OR, USA5 Department of Neurology, University of Vermont, Burlington, VT, USA2005 9 5 2005 5 10 10 14 10 2004 9 5 2005 Copyright © 2005 Ahn et al; licensee BioMed Central Ltd.2005Ahn 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
Acupuncture points and meridians are commonly believed to possess unique electrical properties. The experimental support for this claim is limited given the technical and methodological shortcomings of prior studies. Recent studies indicate a correspondence between acupuncture meridians and connective tissue planes. We hypothesized that segments of acupuncture meridians that are associated with loose connective tissue planes (between muscles or between muscle and bone) visible by ultrasound have greater electrical conductance (less electrical impedance) than non-meridian, parallel control segments.
Methods
We used a four-electrode method to measure the electrical impedance along segments of the Pericardium and Spleen meridians and corresponding parallel control segments in 23 human subjects. Meridian segments were determined by palpation and proportional measurements. Connective tissue planes underlying those segments were imaged with an ultrasound scanner. Along each meridian segment, four gold-plated needles were inserted along a straight line and used as electrodes. A parallel series of four control needles were placed 0.8 cm medial to the meridian needles. For each set of four needles, a 3.3 kHz alternating (AC) constant amplitude current was introduced at three different amplitudes (20, 40, and 80 μAmps) to the outer two needles, while the voltage was measured between the inner two needles. Tissue impedance between the two inner needles was calculated based on Ohm's law (ratio of voltage to current intensity).
Results
At the Pericardium location, mean tissue impedance was significantly lower at meridian segments (70.4 ± 5.7 Ω) compared with control segments (75.0 ± 5.9 Ω) (p = 0.0003). At the Spleen location, mean impedance for meridian (67.8 ± 6.8 Ω) and control segments (68.5 ± 7.5 Ω) were not significantly different (p = 0.70).
Conclusion
Tissue impedance was on average lower along the Pericardium meridian, but not along the Spleen meridian, compared with their respective controls. Ultrasound imaging of meridian and control segments suggested that contact of the needle with connective tissue may explain the decrease in electrical impedance noted at the Pericardium meridian. Further studies are needed to determine whether tissue impedance is lower in (1) connective tissue in general compared with muscle and (2) meridian-associated vs. non meridian-associated connective tissue.
==== Body
Background
In classic Chinese medicine theory, acupuncture meridians represent channels through which energy or "meridian qi" flows. Acupuncture points, traditionally located along these meridians, embody needling sites where the flow of qi may be affected. Acupuncture points and meridians are at the core of traditional acupuncture practice, yet anatomical and physiological explanations for these concepts remain elusive.
One widespread, yet controversial, explanation for acupuncture meridians involves electrical activity. In the acupuncture community, it is widely believed that acupuncture points and meridians are endowed with unique electrical properties. This constitutes the underlying assumption behind the use of electrical point locators commonly used in clinical practice and research. The experimental evidence in support of this practice, however, has been equivocal to date. Prior studies have reported that acupuncture points possess increased electrical conductivity compared to non-acupuncture points, and proposed that acupuncture meridians act as conduits for electrical current [1-16]. These studies, in general, were limited by small sample sizes, poor research design and procedural descriptions, and/or lack of rigorous statistical analyses. In addition, most studies used surface electrodes which may cause confounding by various factors including pressure, skin moisture, electrode contact and abrasion of the stratum corneum. For these reasons, the associations between acupuncture points or meridians and certain electrical properties have remained controversial.
Another proposed explanation for acupuncture points and meridians involves connective tissue. Recent studies have shown that acupuncture points exhibit a different biomechanical response to needling compared with non-acupuncture points [17], that this biomechanical response involves connective tissue [18,19], and that the network formed by acupuncture meridians may correspond to the body-wide network formed by connective tissue [20]. While the physiologic significance of this anatomical association is at present unclear, some researchers have proposed that the collagen content within connective tissue imparts electrical conductive properties [21,22]. These authors also suggested that connective tissue may act as the medium through which electrical communications travel within the acupuncture meridian network.
The goal of this study was to combine ultrasound evaluation and tissue impedance measurements to examine the electrical properties of connective tissue planes associated with meridians. We hypothesized that electrical impedance (which is inversely proportional to electrical conductivity) is lower along two acupuncture meridians associated with loose connective tissue planes (between muscles or between muscle and bone) visible by ultrasound, compared with non-meridians. In order to overcome the limitations associated with prior electrodermal studies, we used gold-plated acupuncture needles inserted into the tissues instead of surface electrodes and used a four-electrode technique with digital data acquisition to measure tissue electrical impedance. In this technique, four electrodes are placed in a straight line, a constant amplitude alternating (AC) current is passed between the two outer electrodes while voltage is measured between the two inner electrodes, and electrical impedance is calculated as the ratio of current to voltage amplitudes.
Methods
Setting
Twenty four subjects (18 female, 6 male) were recruited to participate in the study. Participants were recruited via flyers placed throughout the University of Vermont campus area. Subjects were excluded if they were under 18 years old, pregnant, used anticoagulation medications or had a history of bleeding disorder. One female subject withdrew part way through testing, leaving 23 subjects for analysis. Subjects' age was 39 ± 11.8 (mean ± SD) years. Demographic representation was: 18 non-Hispanic White, 2 Hispanic, 2 Native American and 1 Asian. Each subject was compensated for participation. The testing was performed in the General Clinical Research Center at the University of Vermont Medical Center from April to June 2004. The Institutional Review Board at the University of Vermont Medical Center approved the study methods and procedures.
Instruments
An impedance meter based on the four-electrode technique was designed and fabricated specifically for this study. Prior to initiation of the study, rigorous evaluation of the instrument was performed in vitro and in vivo. The impedance meter outputs a regulated AC 3.3 kHz sinusoidal current of variable amplitude (20, 40, and 80 μA) between two outer electrodes (a) and (d) as shown in Figure 1. Voltage is measured by a voltmeter located between two inner electrodes (b) and (c) (Figure 1). The current used is alternating (AC) as opposed to direct (DC) current to avoid saturating of the electrode surfaces as a result of mobile ion build up (electrolysis). The four-electrode method is superior to a two-electrode method because it minimizes error caused by the contact impedance between electrodes and tissue. The instrument was calibrated by measuring known precision resistors with a commercial Wheatstone bridge; errors were less than 1%. Furthermore, preliminary testing of voltage measurements using gold-plated acupuncture needles in 0.9% NaCl solution indicated stable readings over a 15 minute period and did not change when the polarity of the electrodes was reversed. Standard stainless steel acupuncture needles also were tested but demonstrated substantial voltage fluctuations due to reaction of the metal with the ionic solution. An AC frequency of 3.3 kHz was low enough to minimize ambient electrical interference occurring at higher frequencies and high enough to avoid electrode saturation occurring at lower frequencies.
Figure 1 Block diagram of impedance meter. Through a rechargeable battery, a sine-wave alternating current is delivered to the outer two electrodes (a) and (d). A current sensor registers the amount of current delivered to the electrodes. The inner electrodes (b) and (c) are attached to the voltmeter which registers the electrical potential difference between them. The current and voltage readings may be recorded as time series through connections to a computer (BNC outputs).
Procedure
Two different segments on the skin surface located along acupuncture meridians were chosen for measurement. These segments corresponded to portions of the Pericardium and Spleen meridians [23,24] and will herein be referred to as "Pericardium meridian segment" and "Spleen meridian segment" respectively. The segments were chosen because they were straight, located on flat skin surfaces and easily accessible for ultrasound scanning. The Pericardium segment was located between the flexor carpi radialis and flexor digitorum superficialis muscles, whereas the Spleen segment was located between the medial crest of the tibia and the flexor digitorum longus muscle. The proximal end of the Pericardium meridian segment was one cun (a unit of proportional measurements used in acupuncture practice [24]) distal to the elbow crease and the distal end of the Spleen meridian segment was four cun proximal to the medial malleolus. The location of each meridian and control segment was drawn on the surface of the skin with a marker. Control segments were located 0.8 cm medial and parallel to each meridian segment. Representative ultrasound images for the meridian and control segments are displayed in Figure 2. Ultrasound images were obtained using a GE (Vingmed) System Five (Oslo, Norway) scanner with a 10 MHz linear array probe and an imaging depth of 45 mm. Images were obtained transversely to the meridian and control segments. The right arm and right leg were used for testing. In certain cases where scarring, cysts, or skin inflammation precluded testing on the right side, the left side was used instead for both ultrasound imaging and needling.
Figure 2 Depiction of the Spleen and Pericardium meridians in relation to surface and sub-surface anatomical landmarks. 2 A, B: Cross sectional images of the right forearm as shown through ultrasound and gross anatomical cross section (obtained from the Visual Human Database). The Pericardium segment is located between the flexor carpi radialis and flexor digitorum superficialis muscles. 2 C, D: Cross sectional images of the right leg as shown through ultrasound and cadaveric cross section. The Spleen segment is located between the medial crest of the tibia and the flexor digitorum longus muscle. Blue arrows point to approximate sites where meridian and control needles were inserted.
After cleaning the skin with alcohol, a sterile adhesive holder (Suture aid booties, Sterion Incorporated, Ham Lake, MN) was placed on the skin at each location. Each holder held sterile acupuncture needle guide tubes (four for the meridian needles and four for the control needles) that each had been cut to a length of 33 mm (Figure 3A). The four meridian guide tubes were placed over the meridian segment, with the proximal guide tube located at the proximal end of the segment. Within meridian and control segments, the guide tubes were 13 mm apart, with the exception of the two inner guide tubes where the interval distance was 25 mm. This created a scaffold for the needles to maintain consistent (1) inter-needle distance within a given segment, (2) inter-segment distance between meridian and control segments and (3) needle insertion depth.
Figure 3 Experimental setup. 3A- Holder placed on Pericardium meridian and control skin segments. Guide tubes indicate the location of needles on both segments. In this image, needles (a), (b), (c), and (d) are inserted along the control segment. Current is passed between electrodes (a) and (d) and voltage is measured between (b) and (c). 3B- Tissue impedance meter connected to laptop computer and to needle electrodes.
For each segment, the order of testing (meridian vs. control) was randomized. After randomization, four gold-plated needles (Viva 0.25 mm diameter, 40 mm length, Helio Medical Supplies, Inc., San Jose, CA) were connected to the impedance meter device via electrical cables (Teca, Oxford Instruments Medical, Hawthorne, NY). The handle of each acupuncture needle was inserted 10 mm into the female end of the connector cable, while the male end was inserted into the impedance meter device (Figure 3). The gold plated acupuncture needles thus acted as electrodes. Four needles were inserted through each of the four guide tubes overlying either a meridian or control segment and then into the skin and underlying tissues. Each needle was inserted until the hub of the connector contacted the end of the guide tube. This ensured that all needles were inserted to the same depth of 10 mm. In all cases, electrodes (a) and (d) were connected to the two output electrodes of the constant current source and electrode (a) was inserted into the most proximal needling site. A 3.3 kHz AC current was then introduced at three different amplitudes (20, 40, and 80 μA) to the outer two needles while the voltage was measured between the inner two needles. Measurements were made twice for two to five seconds at each current amplitude. The voltage was read on an analogue voltmeter as well as stored as a computer digital time series using LabView 5.1 software and data acquisition board (National Instruments Corporation, Austin, TX). During the testing, subjects laid supine on a standard hospital bed with the head elevated at approximately 60 degrees above horizontal. Tissue impedance was determined by taking the slope of the linear regression of voltage as a function of current intensity using least squares methodology.
Statistical analysis
Two-factor repeated measures analysis of variance was used to test for differences in mean tissue impedance across experimental conditions. The repeated factors represented segment (meridian vs. control), and location (Pericardium vs. Spleen). If there was evidence of an interaction between the two factors (i.e. segment differences were determined to be location dependent), simple effects within each location were examined based on an F-test corresponding to the appropriate contrast. Statistical analyses were performed using SAS statistical software (SAS Version 8, SAS Institute, Cary, NC).
Results
Differences in tissue impedance between control and meridian segments for individual subjects are shown in Figure 4. Overall, there was a significant difference in mean tissue impedance between meridian and control segments (Repeated Measures ANOVA, main effect p = 0.004). There was also near-significant evidence suggesting that the difference was location-specific (Repeated Measures ANOVA, interaction p = 0.07). When comparisons were performed within each location, difference in mean tissue impedance at the Pericardium location were highly significant (p = 0.0003) (Figure 5). Tissue impedance was 70.4 ± 5.7 Ω (mean ± SE) for the Pericardium meridian segment and 75.0 ± 5.9 Ω for the Pericardium control segment. Analyses performed within the Spleen location did not result in significant differences between the two segments (p = 0.70) (Figure 5). Tissue impedance was 67.8 ± 6.8 Ω and 68.5 ± 7.5 Ω for Spleen meridian and control segments respectively.
Figure 4 Tissue impedance measurements for individual subjects. The difference in impedance between control and meridian segments (Impedance difference) is shown for the Pericardium (A) and Spleen (B) meridian location.
Figure 5 Mean tissue impedance at Pericardium and Spleen meridian segments and at corresponding control segments. Bar graphs represent mean ± SE. * indicates p < 0.01.
Discussion
We found that tissue impedance was lower along the Pericardium meridian (compared with control) but not along the Spleen meridian. Several factors may have contributed to this observed difference. First, although both Pericardium and Spleen meridian segments were associated with loose connective tissue planes (between muscles or between muscle and bone), the Pericardium meridian-associated connective tissue was in general more clearly defined in the ultrasound images than the Spleen meridian-associated connective tissue. Second, in clinical settings, the Spleen channel may need to be needled deeper than the Pericardium channel [25-27] suggesting that the 1 cm needle penetration at the Spleen segment may not have been sufficient. Third, the proximity of control segments to other (non-meridian/meridian) connective tissue planes may also have influenced our results. The Pericardium control needles were inserted into the flexor digitorum superficialis muscle which has a relatively wide transverse width (Figure 2A) and therefore the Pericardium control needles were inserted into the belly of that muscle in nearly all subjects. In contrast, because of the variable width of the flexor digitorum longus and its orientation perpendicular to the skin surface, the Spleen control segment was often close to the connective tissue plane separating this muscle from the soleus (Figure 2C). Therefore in some subjects, the Spleen control needle may have penetrated as much connective tissue as the Spleen meridian needles. Indeed in the ten subjects for which both medial and lateral edges of the flexor digitorum longus could be clearly delineated on the ultrasound images, greater flexor digitorum longus muscle width was positively correlated with greater tissue impedance in Spleen control relative to meridian segments (r = 0.60). In this regard, it is interesting to note that there is some variability among major acupuncture texts about the respective locations of the Spleen, Kidney and Liver meridians in this portion of the leg [23,25,27], all of these meridians running longitudinally medial to the medial edge of the tibia. Thus the connective tissue plane separating the flexor digitorum longus and soleus may in fact represent the Kidney or Liver meridian. On the other hand, this connective tissue plane may not represent any acupuncture meridian. Further studies are needed to (1) test segments located over inter-muscular connective tissue planes (whether meridian-associated or not), compared with segments located over muscle to determine whether connective tissue in general has decreased impedance compared with muscle and (2) test segments located over meridian-associated connective tissue vs. non-meridian associated connective tissue to determine whether meridians have decreased impedance compared with non meridian-associated connective tissue. Selection of areas where connective tissue planes are spaced as far apart as possible (such as the thigh and the upper arm) would be advantageous to test these hypotheses.
Compared to previous studies examining the electrical properties of acupuncture points and meridians, our measurement methods offer several advantages. First, we used needles to overcome the variability associated with surface electrodes. By directly accessing the tissue, the needles bypass potential confounders such as pressure, sweat, skin abrasions, and variable surface topography, among other factors. Second, we utilized a four-electrode method to perform our tissue impedance measurements. To date, most investigations of electrical impedance of acupuncture points or meridians have used a two-electrode method, i.e., the two electrodes used to introduce an electrical current were also connected to a voltmeter to measure the voltage between the two electrodes [1-4,6-8,10,11,14,15]. This method can cause significant fluctuation of voltage between the two electrodes due to variable contact impedance between electrodes and tissue. The four electrode technique used in this study is considered the standard in biophysical sciences and is widely employed to measure electrical conductance/impedance of biological tissue [28-37]. This method minimizes error due to fluctuation in voltage and electrode contact impedance. Three prior studies have evaluated the bioelectrical properties of acupuncture meridians using a four-electrode technique in vivo [38,39] and using an in vitro gel model [40]. Both in vivo studies reported finding lower electrical impedance along acupuncture meridians. However, in one study [39] the four electrodes were not all placed along a straight line, making it difficult to relate that study's results to our findings, and in the other the results were not analyzed statistically [38].
Since the investigator was not blinded to which segment corresponded to the meridian and which to the control, guarding against potential sources of experimental bias during needle placement and data collection was an important aspect of this study. Use of (1) adhesive scaffolding and needle guide tubes (ensuring that all needles were equidistant and inserted to the same depth) and (2) computerized data collection (analogue data was only collected as a backup for the digital files) ensured that minimal opportunity was present for bias to influence the experimental results.
This study has several limitations. Because we used uninsulated needles, our measurements reflected the total effective impedance between the two inner electrodes and therefore we could not resolve components of this tissue impedance contributed by individual tissue layers. In addition, the study of electrical properties of biological tissues is extremely complex, and the presence of tissue anisotropy, possible biological semiconductor properties (lending to internal charge storage between adjacent biological structures), and tissue heterogeneity precludes attributing the observed differences in electrical impedance to any particular molecule or structure. These complexities will present challenges for future studies in delineating exact causes for any possible electrical communication within the acupuncture network. Given these complexities, however, it is remarkable that the impedance values of the Pericardium segment were consistently lower than that of its control. This study will provide initial groundwork for future, more extensive investigations.
Conclusion
In summary, tissue impedance was lower along the Pericardium meridian (compared with control) but not along the Spleen meridian. Ultrasound imaging of meridian and control segments suggest that (1) lack of difference at the Spleen location may have been due to both control and meridian needles penetrating connective tissue and (2) tissue impedance may be influenced by needle penetration of connective tissue, whether meridian-associated or not.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AA – Participated in testing of human subjects and wrote the manuscript.
JW – Designed and supervised in vitro testing and calibration of the tissue impedance instrument.
GJB – Conducted statistical analysis and assisted in writing of the manuscript.
RH – Participated in designing the study and in vitro testing of tissue impedance instrument and edited the manuscript.
HML – Supervised the design and coordination of the study, testing, and manuscript preparation.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Debbie Stevens-Tuttle, Zoe Amos, Mete Erturk, Nicole A. Bouffard and James R. Fox for technical assistance. This work was supported by NIH National Research Service Award, Grant No. T32-AT0051-03 from the National Center for Complementary and Alternative Medicine. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Center for Complementary and Alternative Medicine, National Institutes of Health. This study was performed at the University of Vermont General Clinical Research Center at Fletcher Allen Health Care supported by NIH Grant M01RR00109. For images used in the figures, written consent was obtained from the subjects for publication of study.
==== Refs
Chen KG Electrical properties of meridians IEEE Eng Med Biol Mag 1996 15 58 63 10.1109/51.499759
Hyvarinen J Karlsson M Low-resistance skin points that may coincide with acupuncture loci Med Biol 1977 55 88 94 865155
Nakatani Y Yamashita K Ryodoraku Acupuncture 1977 Tokyo, Ryodoraku Research Institute
Nakatani Y Skin electric resistance and ryodoraku J Autonomic Nerve 1956 6
Niboyet J Nouvelle constatations sur les proprietes electriques des ponts Chinois Bull Soc Acup 1958 30
Reichmanis M Marino AA Becker RO DC skin conductance variation at acupuncture loci Am J Chin Med 1976 4 69 72 1266803 10.1142/S0192415X7600010X
Reichmanis M Marino AA Becker RO Electrical correlates of acupuncture points IEEE Trans Biomed Eng 1975 22 533 535 1184029
Becker RO Reichmanis M Marino AA Spadaro JA Electrophysiological correlates of acupuncture points and meridians. Psychoenergetic Systems 1976 1 105 112
Becker RO Selden G The body electric : electromagnetism and the foundation of life 1985 1st New York, Morrow 364 p.
Terral C Rabischong P A scientific basis for acupuncture? The Journal of Alternative and Complementary Medicine 1997 3 S 55-S-65
Voll R Nosodenanwendung in Diagnostik und therapie 1977 Uelzen, Germany, ML-Verlage
Zhu Z Research advances in the electrical specificity of meridians and acupuncture points American Journal of Acupuncture 1981 9 203 216
Brown ML Ulett GA Stern JA Acupuncture loci: techniques for location Am J Chin Med 1974 2 67 74 4812065 10.1142/S0192415X74000080
Reichmanis M Marino AA Becker RO Laplace plane analysis of transient impedance between acupuncture points Li-4 and Li-12 IEEE Trans Biomed Eng 1977 24 402 405 881215
Nakatani Y An aspect of the study of Ryodoraku Clinic of Chinese Medicine 1956 3 54 56
Niboyet JEH Bourdiol RJ Regard PG Traite d'acupuncture Traité d'acupuncture 1970 Sainte-Ruffine,, Maissonneuve 140 351
Langevin HM Churchill DL Fox JR Badger GJ Garra BS Biomechanical response to acupuncture needling in humans J Appl Physiol 2001 91 2471 2478 11717207
Langevin HM Churchill DL Cipolla MJ Mechanical signaling through connective tissue: A mechanism for the therapeutic effect of acupuncture FASEB J 2001 15 2275 2282 11641255 10.1096/fj.01-0015hyp
Langevin HM Churchill DL Wu J Badger GJ Yandow JA Fox JR Krag MH Evidence of connective tissue involvement in acupuncture FASEB J 2002 16 872 874 11967233
Langevin HM Yandow JA Relationship of acupuncture points and meridians to connective tissue planes Anat Rec (New Anat) 2002 269 257 265 10.1002/ar.10185
Ho MW Knight DP The acupuncture system and the liquid crystalline collagen fibers of the connective tissues Am J Chin Med 1998 26 251 263 9862013 10.1142/S0192415X98000294
Oschman JL Energy medicine : the scientific basis 2000 Edinburgh ; New York, Churchill Livingstone xv, 275 p.
Deadman P Al-Khafaji M Baker K A Manual of Acupuncture 1998 East Sussex, Journal of Chinese Medicine Publications
Cheng XM Acupuncture Reference Book 1995 , GuiZhou Publishing
Cheng H Teng L Cheng H and Teng L Chinese acupuncture and moxibustion Chapter 14: Acupuncture Techniques 1987 1st Beijing, Foreign Language Press 322 329
Beijing, Shanghai and Nanjing Colleges of Traditional Chinese Medicine Essentials of Chinese Acupuncture. 1980 Beijing, Foreign Language Press
O'Connor J Bensky D Shanghai Zhong yi xue yuan. Acupuncture : a comprehensive text 1981 Chicago, Eastland Press xvi, 741 p., [3] folded leaves of plates
Lukaski HC Regional bioelectrical impedance analysis: applications in health and medicine Acta Diabetol 2003 40 Suppl 1 S196 9 14618471 10.1007/s00592-003-0064-4
Salazar Y Bragos R Casas O Cinca J Rosell J Transmural versus nontransmural in situ electrical impedance spectrum for healthy, ischemic, and healed myocardium IEEE Trans Biomed Eng 2004 51 1421 1427 15311828 10.1109/TBME.2004.828030
Cinca J Warren M Carreno A Tresanchez M Armadans L Gomez P Soler-Soler J Changes in myocardial electrical impedance induced by coronary artery occlusion in pigs with and without preconditioning: correlation with local ST-segment potential and ventricular arrhythmias Circulation 1997 96 3079 3086 9386178
Ellenby MI Small KW Wells RM Hoyt DJ Lowe JE On-line detection of reversible myocardial ischemic injury by measurement of myocardial electrical impedance Ann Thorac Surg 1987 44 587 597 3689045
Fallert MA Mirotznik MS Downing SW Savage EB Foster KR Josephson ME Bogen DK Myocardial electrical impedance mapping of ischemic sheep hearts and healing aneurysms Circulation 1993 87 199 207 8419008
Steendijk P van der Velde ET Baan J Dependence of anisotropic myocardial electrical resistivity on cardiac phase and excitation frequency Basic Res Cardiol 1994 89 411 426 7702534 10.1007/BF00788279
Tsai JZ Cao H Tungjitkusolmun S Woo EJ Vorperian VR Webster JG Dependence of apparent resistance of four-electrode probes on insertion depth IEEE Trans Biomed Eng 2000 47 41 48 10646278 10.1109/10.817618
Tsai JZ Will JA Hubbard-Van Stelle S Cao H Tungjitkusolmun S Choy YB Haemmerich D Vorperian VR Webster JG Error analysis of tissue resistivity measurement IEEE Trans Biomed Eng 2002 49 484 494 12002180 10.1109/10.995687
Tsai JZ Will JA Hubbard-Van Stelle S Cao H Tungjitkusolmun S Choy YB Haemmerich D Vorperian VR Webster JG In-vivo measurement of swine myocardial resistivity IEEE Trans Biomed Eng 2002 49 472 483 12002179 10.1109/10.995686
Haemmerich D Staelin ST Tsai JZ Tungjitkusolmun S Mahvi DM Webster JG In vivo electrical conductivity of hepatic tumours Physiol Meas 2003 24 251 260 12812412 10.1088/0967-3334/24/2/302
Yang W Chang R Investigation of lower resistance meridian
I. Method of investigation Peking University Academic Journal 128 136
Zhang W Xu R Zhu Z The influence of acupuncture on the impedance measured by four electrodes on meridians Acupunct Electrother Res 1999 24 181 188 10768415
Zhang W Zhuang F Tian Y Li H [A simulating study of biophysical features along meridians on a gel model] Sheng Wu Yi Xue Gong Cheng Xue Za Zhi 2001 18 357 361 11605488
| 15882468 | PMC1142259 | CC BY | 2021-01-04 16:31:46 | no | BMC Complement Altern Med. 2005 May 9; 5:10 | utf-8 | BMC Complement Altern Med | 2,005 | 10.1186/1472-6882-5-10 | oa_comm |
==== Front
BMC BiochemBMC Biochemistry1471-2091BioMed Central London 1471-2091-6-81587635710.1186/1471-2091-6-8Research ArticleRational polynomial representation of ribonucleotide reductase activity Radivoyevitch Tomas [email protected] Ossama B [email protected] Barry S [email protected] Epidemiology and Biostatistics, Case Western Reserve University, Cleveland, OH 44106, USA2 Medicine, University of Pittsburgh, Pittsburgh, PA 15261, USA3 Chemistry, University of Pennsylvania, Philadelphia, PA 19104, USA2005 6 5 2005 6 8 8 29 12 2004 6 5 2005 Copyright © 2005 Radivoyevitch et al; licensee BioMed Central Ltd.2005Radivoyevitch 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
Recent data suggest that ribonucleotide reductase (RNR) exists not only as a heterodimer R12R22 of R12 and R22 homodimers, but also as tetramers R14R24 and hexamers R16R26. Recent data also suggest that ATP binds the R1 subunit at a previously undescribed hexamerization site, in addition to its binding to previously described dimerization and tetramerization sites. Thus, the current view is that R1 has four NDP substrate binding possibilities, four dimerization site binding possibilities (dATP, ATP, dGTP, or dTTP), two tetramerization site binding possibilities (dATP or ATP), and one hexamerization site binding possibility (ATP), in addition to possibilities of unbound site states. This large number of internal R1 states implies an even larger number of quaternary states. A mathematical model of RNR activity which explicitly represents the states of R1 currently exists, but it is complicated in several ways: (1) it includes up to six-fold nested sums; (2) it uses different mathematical structures under different substrate-modulator conditions; and (3) it requires root solutions of high order polynomials to determine R1 proportions in mono-, di-, tetra- and hexamer states and thus RNR activity as a function of modulator and total R1 concentrations.
Results
We present four (one for each NDP) rational polynomial models of RNR activity as a function of substrate and reaction rate modifier concentrations. The new models avoid the complications of the earlier model without compromising curve fits to recent data.
Conclusion
Compared to the earlier model of recent data, the new rational polynomial models are simpler, adequately fitting, and likely better suited for biochemical network simulations.
==== Body
Background
Ribonucleotide reductase (RNR) is a key component of de novo deoxynucleotide (dNTP) metabolism and an important target of cancer therapies [1]. This enzyme, which reduces ribonucleoside diphosphates into corresponding deoxyribonucleoside diphosphates, is exquisitely controlled to properly balance dNTP fluxes in the face of changing scheduled (S phase) and unscheduled (DNA damage/repair) dNTP synthesis demands [2].
Recent data [3-6] suggest that ribonucleotide reductase (RNR) exists not only as a heterodimer R12R22 of R12 and R22 homodimers [2], but also as a R14R24 tetramer and as a R16R26 hexamer, where hexamer formation is driven by ATP binding to a previously undescribed hexamerization site. Thus, in addition to its four substrate binding possibilities in ADP, GDP, CDP, or UDP, and four dimerization/specificity site binding possibilities in dATP, ATP, dGTP, or dTTP, the current view (Figure 1) is that R1 has two tetramerization/inhibitory site binding possibilities in dATP or ATP, and one hexamerization/activation site binding possibility in ATP, in addition to possibilities of unbound site states. The resulting large number of possible R1 states implies an even larger number of quaternary states, and this leads to a complicated mathematical model of RNR activity [3-6]. This model, although useful for explaining RNR activity data, is not useful for biochemical network simulations because: a) it is unwieldy (including up to six-fold nested sums), b) it uses different mathematical structures under different substrate-modulator conditions, and c) it requires root solutions of high order polynomials to determine R1 proportions in mono-, di-, tetra- and hexamer quaternary states, and thus RNR activity, as a function of modulator and total R1 concentrations. Simpler mathematical reaction rate models of RNR are needed if deoxynucleotide metabolism [7] is to be represented using Systems Biology Markup Language (SBML) [8-10], a standard which requires single algebraic expression reaction rate laws in some applications [11,12]. Based on recent data from Cooperman's group [4-6], such expressions are provided here for RNR.
Figure 1 Quaternary states of R1. Modulators of RNR activity listed in this figure bind R1 to create higher order quaternary states. Tetramers exist in an equilibrium between low activity states (see k4 in Table 1) and inactive states (kcat = 0). Adapted from Scheme 1 in [4,5].
Results
Reaction activities are viewed here as weighted sums of enzyme state specific activities multiplied by probabilities of enzymes being in specific states. For example, a Michaelis-Menten reaction rate law is viewed as
where the probability that the enzyme is in a loaded/reactive state (with activity kcat) is P(EA) and the probability that the enzyme is in an empty/unreactive state (with no activity) is P(E).
The RNR models presented here are based on the following four enzyme state probability assumptions:
1. The probability that a particular R1 subunit is bound to NDP is assumed to be
2. The probability that L ∈ {ATP, dATP, dTTP, dGTP} is also bound to the dimerization/specificity site, conditional on NDP binding, is assumed to be
where, if L = dTTP and NDP = GDP for example,
is the probability that dTTP is bound to the dimerization/specificity site given that GDP is bound to the substrate site; the probability of an empty dimerization/specificity site is thus
3. The probability that the tetramerization site is either empty, occupied by ATP, or occupied by dATP, is assumed to be, respectively,
4. Finally, the probability that the hexamerization site is occupied by ATP is assumed to be
The subscripts a (activation), i (inactivation), and s (specificity) on the binding constants correspond to h (hexamerization), t (tetramerization), and d (dimerization) subscripts on P, respectively. That parameter values differ depending upon which NDP substrate is bound to the active site (see Table 1) is indicated by the conditional probabilities.
Table 1 Parameter estimates of the reductase models. Fits to data are as shown in Figure 2.
substrate KsdATP KsATP KsdTTP KsdGTP KidATP KiATP KaATP k2 k2dA k2A k2e k4 k6
ADP 2 200 2.4b 0.5 1.25 300 2000d 0.21 0.03 0.16
GDP 1 100 0.5 2 2 190 2400 0.28a 0.04 0.19
CDP 2 70 1.55b 2c 1.5 600 1400 0.25a 0.29a 0.08 0.32
UDP 1 100 0.7b 2c 0.5 200 800 0.26 0.26 0.26a
Binding constants in μM, rate constants in 1/s.
afixed values taken directly from table 5 of [5].
busing Eq. 10, these were adjusted to yield fluxes of 12.5, 12.5, 20 and 5 (uM/min) for ADP, GDP, CDP and UDP, respectively, under assumptions of E0 = 16 μM, ADP = 430 μM, GDP = 110 μM, CDP = 55 μM, UDP = 170 μM, KmADP = 12 μM, KmGDP = 4.9 μM, KmCDP = 2 μM, KmUDP = 6.4 μM, ATP = 1450 μM, dATP = 10.5 μM, dGTP = 7.3 μM, dTTP = 30 μM and the remaining parameter values in Table 1, see [7].
cno data, thus, these can be assumed to have any value between .5 and 2; a default value of 2 was carried down from GDP.
dno data, 2000 is based on the other rows.
Previous work [3-6] has shown that the dimer and hexamer states are active, that the tetramer state is slightly active for ADP and GDP and is otherwise inactive, that dimer state activity for CDP and UDP exists when ATP or dATP is bound to the dimerization/specificity site, and that an empty dimerization/specificity site still permits the formation of some dimer with CDP reductase activity, see Table 5 of [5]. Thus, based on the enzyme state probabilities given above, for kcat implicitly defined through
we propose the following expressions:
ADP reduction
GDP reduction
CDP reduction
UDP reduction
In these equations, for ADP and GDP, the first factor is the probability that the dimer site is occupied, and for CDP and UDP, the first factor is the expectation of kcat conditional on R1 being in a dimer state (i.e. having an empty tetramerization site). In the ADP and GDP models, the second factor is the conditional expectation of kcat given that the dimerization site is occupied: the first term of this second factor has in its numerator the statement that kcat = k2 if the tetramerization site is empty, or kcat = k4 if it is occupied by either dATP or ATP, and the second term states that kcat = k6 if the hexamerization site is occupied by ATP. For the CDP and UDP models, the first term of the second factor is the probability of an empty tetramerization site (the event that the corresponding first factor was conditioned on), and the second term states that if the ATP concentration is high enough that the hexamerization/second term dominates the tetramerization/first term whilst the first factor approaches k2A, k6 is the overall kcat. This rationale served as our model selection guide. Importantly, the models fitted recent data [3-6] very well, see Figure 2 and Table 1.
Figure 2 Data from [4-6] and corresponding curve fits (Table 1) of the RNR activity models. In these plots, from left to right, for ADP reduction dGTP was 2.1 uM or variable, for GDP reduction dTTP was 100 uM, 300 uM, variable, or 85 uM, and for pyrimidines specificity site binding concentrations were as shown. In all cases NDP and R2 were at saturating levels.
Discussion
In general, when an integrated system is engineered from component subsystems, the behavior of the overall system depends on component input-output specifications more so than the details of component implementations. By analogy, when enzymological data are applied to biochemical network modeling, rather than the elucidation of reaction mechanisms, it can be expected that the reaction surfaces themselves (i.e. the enzyme's input-output characteristics) determine network behavior more so than the details of how such surfaces are represented. Thus, for applications to systems biology, large confidence intervals (CI) in the model parameter estimates of Table 1 (not shown) are not a problem because only goodness-of-fit (Fig. 2) really matters; this claim assumes an operating range within the data range, since similarly fitting models often veer apart when used in extrapolations. If reaction mechanism inferences were instead being sought, the large CI in the model parameter estimates would have been a problem, e.g. the squared terms in the model suggest cooperative binding, but this choice provides only slightly better curve fits compared to linear terms (not shown), so cooperative binding cannot be inferred from this model.
In the RNR model presented here, the proportion of R1 units existing in monomer, dimer, tetramer, or hexamer states, and thus the RNR activity per unit enzyme, depends on site binding occupancies but does not depend on the total R1 concentration. In the more complicated previous model [3-6], higher total R1 concentrations favor higher order quaternary states. The degree to which this is so is illustrated by plots of predicted GDP reductase activity as a function of ATP concentration at various R1 concentrations (Figure 3). Consistent with the formation of higher order quaternary R1 states, these plots contract to the left as the total R1 concentrations increase from 1 μM to 100 μM. In future work, the model given here will be altered to capture such trends without losing its simplicity; the total R1 concentration will enter such a model not only as a linear modulator of the reaction surface amplitude (i.e. E0 in Eq. 10), but also as a modifier of reaction surface shape parameters, e.g. KaATP will be replaced by a decreasing function of R1.
Figure 3 GDP reductase as a function of ATP concentration at various R1 concentrations, predicted using the earlier model [3-6].
Conclusion
We identified a rational polynomial model of RNR activity that has single algebraic expressions for each reductase reaction rate law. The expressions provide reasonably good fits (Fig. 2) to recent data [3-6]. Compared to previous reaction rate expressions for this data [3-6], the new expressions are simpler and thus better suited for biochemical network simulations, particularly those constrained to use enzyme reaction rate laws defined as single algebraic expressions [11,12].
Methods
The parameter estimates shown in Table 1 were obtained through a trial-and-error iterative process of nonlinear least squares curve fitting under various, convergence enabling, parameter fixations (i.e. profile searches). In the end, the curve fits were those of Figure 2 with corresponding parameter estimates in Table 1; large 95% confidence intervals (not shown) allowed rounding of the parameter estimates to somewhat arbitrary choices. Non-linear least squares parameter estimations were performed using the optimization method of Nelder and Mead [13] and the statistical computing environment R [14]. All parameters were estimated as ec to assure positive values. For additional details, R scripts are available with the data as supplementary material [15].
List of abbreviations
RNR = ribonucleotide reductase; dNTP = deoxynucleoside tripshospate; dNDP = deoxynucleoside dipshospate; a = activation; i = inactivation; s = specificity; h = hexamerization; t = tetramerization; d = dimerization; CI = confidence interval; SBML = Systems Biology Markup Language.
Authors' contributions
TR performed the curve fits to the data and explored various model choices.
OBK and BSC provided the original model, its simulations (Figure 3) and the data.
Acknowledgements
We thank Dr. Charles Scott for sharing his data. TR was supported by the Biostatistics Core Facility of the Comprehensive Cancer Center of Case Western Reserve University and University Hospitals of Cleveland (P30 CA43703), the American Cancer Society (IRG-91-022-09), the National Cancer Institute's Integrative Cancer Biology Program (P20 CA112963-01) and NIH grant 1K25 CA104791-01A1. BSC was supported by NIH grant CA 58567. OBK was a recipient of a postdoctoral fellowship award from the Pennsylvania-Delaware Affiliate of the AHA, and was supported by grants DK061296 and DK066883 from the NIH.
==== Refs
Tsimberidou AM Alvarado Y Giles FJ Evolving role of ribonucleoside reductase inhibitors in hematologic malignancies Expert Rev Anticancer Ther 2002 2 437 448 12647987 10.1586/14737140.2.4.437
Kolberg M Strand KR Graff P Andersson KK Structure, function, and mechanism of ribonucleotide reductases Biochim Biophys Acta 2004 1699 1 34 15158709
Cooperman BS Kashlan OB A comprehensive model for the allosteric regulation of Class Ia ribonucleotide reductases Adv Enzyme Regul 2003 43 167 182 12791390 10.1016/S0065-2571(02)00035-3
Kashlan OB Cooperman BS Comprehensive model for allosteric regulation of mammalian ribonucleotide reductase: refinements and consequences Biochemistry 2003 42 1696 1706 12578384 10.1021/bi020634d
Kashlan OB Scott CP Lear JD Cooperman BS A comprehensive model for the allosteric regulation of mammalian ribonucleotide reductase. Functional consequences of ATP- and dATP-induced oligomerization of the large subunit Biochemistry 2002 41 462 474 11781084 10.1021/bi011653a
Scott CP Kashlan OB Lear JD Cooperman BS A quantitative model for allosteric control of purine reduction by murine ribonucleotide reductase Biochemistry 2001 40 1651 1661 11327824 10.1021/bi002335z
Jackson RC The Theoretical Foundations of Cancer Chemotherapy Introduced by Computer Models 1992 New York , Academic Press 447
Hucka M Finney A Sauro HM Bolouri H Doyle JC Kitano H Arkin AP Bornstein BJ Bray D Cornish-Bowden A Cuellar AA Dronov S Gilles ED Ginkel M Gor V Goryanin II Hedley WJ Hodgman TC Hofmeyr JH Hunter PJ Juty NS Kasberger JL Kremling A Kummer U Le Novere N Loew LM Lucio D Mendes P Minch E Mjolsness ED Nakayama Y Nelson MR Nielsen PF Sakurada T Schaff JC Shapiro BE Shimizu TS Spence HD Stelling J Takahashi K Tomita M Wagner J Wang J The systems biology markup language (SBML): a medium for representation and exchange of biochemical network models Bioinformatics 2003 19 524 531 12611808 10.1093/bioinformatics/btg015
Finney A Hucka M Systems biology markup language: Level 2 and beyond Biochem Soc Trans 2003 31 1472 1473 14641091
Systems Biology Markup Language
Radivoyevitch T A two-way interface between limited Systems Biology Markup Language and R BMC Bioinformatics 2004 5 190 15585059 10.1186/1471-2105-5-190
Radivoyevitch T SBMLR
Nelder JA Mead R A simplex algorithm for function minimization Computer Journal 1965 7 308 313
Ihaka R Gentleman R R:a language for data analysis and graphics Journal of Computational and graphical statistics 1996 5 299 314
Radivoyevitch T Kashlan OB Cooperman BS Rational Polynomial Representation of Ribonucleotide Reductase Activity
| 15876357 | PMC1142302 | CC BY | 2021-01-04 16:36:49 | no | BMC Biochem. 2005 May 6; 6:8 | utf-8 | BMC Biochem | 2,005 | 10.1186/1471-2091-6-8 | oa_comm |
==== Front
Aust New Zealand Health PolicyAustralia and New Zealand Health Policy1743-8462BioMed Central London 1743-8462-2-91587782010.1186/1743-8462-2-9CommentaryRecent developments in the funding and organisation of the New Zealand health system Ashton Toni [email protected] Centre for Health Services Research and Policy, School of Population Health. University of Auckland, Private Bag 92019, Auckland, New Zealand2005 7 5 2005 2 9 9 1 2 2005 7 5 2005 Copyright © 2005 Ashton; licensee BioMed Central Ltd.2005Ashton; 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.
During the 1990s, the New Zealand health sector went through a decade of turbulence with a series of major structural changes being introduced in a relatively short period of time. The new millennium brought further change, with the establishment of 21 district health boards and the restoration of a less commercially-oriented system. The sector now appears to be more stable. However many incremental changes are in train and there has been considerable turbulence below the surface as key players jostle for position. This paper reports on some of the recent changes that have occurred in the restructuring of the New Zealand health system. Three issues are discussed: the devolution of funds and decision-making to district health boards, developments in primary health care, and the position of the private health insurance industry.
==== Body
Introduction
The New Zealand health system has gone through a series of fairly radical structural changes over the past decade or so. Most notable – and most radical – were the 1993 so-called "health reforms" in which an attempt was made to introduce market-like incentives into the system by requiring public and private providers to compete for service contracts from public purchasers. Although it was probably too early to expect any major improvements in health sector performance, the general consensus that emerged from policy analysts was that the new system was unlikely to achieve any significant efficiency gains [1-3]. The competitive arrangements were also rather alien to many of those working within the health system who were more accustomed to a public sector style of philosophy [4].
The new millennium brought yet another round of restructuring following the election of a Labour-led coalition government. The general direction of change this time around was towards a more planned and community-oriented system, with 21 district health boards (DHBs) being responsible for meeting the health and disability service needs of the people living within their district. The aims of the restructuring into DHBs are set out in the New Zealand Public Health and Disability Act 2000. In addition to providing appropriate health and disability services for all New Zealanders, the objectives are to reduce health disparities (especially by improving the health outcomes of Mäori); to foster community participation (via elected representatives, openness of decision-making and community consultation on strategic planning); and to facilitate access to and dissemination of information pertaining to service delivery.
A strong primary care system is seen as central to improving the health status of the population generally and, more particularly, to reducing health inequalities. The government has therefore given priority to implementing the Primary Health Care Strategy [5]. The strategy is complex and multi-pronged, and includes changes to both the organisation of primary health services and to the level and method of subsidising these services.
As in 1993, the fundamental method of financing health care through general taxation has remained unchanged during this latest round of reform. While the share of public spending in total health expenditure declined from a high of 88.1% in 1981/82 to 76.6% in 1992/93, it has since remained fairly stable at around 77% of total health expenditure [6,7]. However, after increasing steadily from 2.8% of total health expenditure in 1989/90 to 6.8% in 1996/97, the share of private health insurance in total funding subsequently declined to 5.7% in 2001/02 following a series of rises in premiums [6,8]. The lack of growth in membership of private health insurance, together with the subsidisation of private insurance in Australia, has encouraged the health insurance industry to lobby the government for a greater role for private insurance in financing the health system.
We are now three years into the new structure. The DHBs are well-established, with the second round of elections for board membership having taken place in October 2004. DHBs have been active in working with primary health care providers to establish networks of providers called Primary Health Organisations (PHOs) and by April 2005, 77 PHOs had been established covering more than 90% of the population [9]. While significant progress has been made away from the commercially-oriented environment that prevailed during the 1990s towards a more community-focussed health system, the path has not always been smooth. This paper describes and discusses three particular issues: the process of devolving funds and decision-making to the DHBs, developments in the primary health sector, and the private health insurance lobby. To place the issues into a historical and political context, the paper begins with a brief overview of the New Zealand health system and its development. Each of the three issues is then described and discussed in more detail.
Overview of the New Zealand health system
The roots of the New Zealand health system as it is today were first formed through the Social Security Act 1938 when the (Labour) government of the day outlined its vision of free health services for all New Zealanders, regardless of ability to pay. However widespread opposition from the medical profession meant that the Act was never implemented in full [2,10]. Instead, a dual system of funding emerged in which mental health, maternity, and hospital services were fully funded by the government, while GPs retained the right to charge a fee over and above any subsidy for general practice consultations. A dual system of provision also emerged, with most primary services being provided in the private sector but most secondary and tertiary services being provided by public hospitals. As public provision of hospital services expanded, the number of private hospital beds initially declined. However, the introduction of private medical insurance in the early 1960s, together with subsidies for the maintenance of private hospitals, eventually reversed this trend. [10,11].
By the 1970s, concern was growing that "the fragmented pattern of health care delivery means that New Zealand lacks a national health service" [12]. After a series of reviews and proposals, 30 local hospital boards were eventually gradually replaced by 14 area health boards between 1983 and 1989. The funding and provision of public health services and public hospital services were amalgamated under area health boards. However, primary medical services remained separately funded and provided. Thus the vision of a "national health service" remained illusive.
Area health boards had a number of features that were subsequently reintroduced in 2001 as part of the district health board structure. These features include:
• governance by a locally-based and (mostly) locally-elected board;
• funding by means of a population-based formula;
• a reorientation away from curative services towards prevention;
• planning of services in consultation with key stakeholders;
• a more strategic approach to health service delivery, including the use of national goals and targets.
Establishment of the 14 area health boards had not long been completed before they were abolished in 1991 and the so-called 'purchaser-provider split' was introduced in 1993 following two-years of preparation. In reality the split only applied to those services that had previously been provided by the area health boards: that is, public hospital services, public health services, and a limited range of community-based services. The roles had always been separated for most primary health services and no attempt was made to apply the principle to privately-funded hospital services. The main objective of separating the roles of purchaser and provider was to secure efficiency gains by introducing market-like incentives into the health system and requiring a more commercial approach to health service delivery. The separation also effectively allowed, for the first time, funding for all personal health services (i.e. including primary health services) to be amalgamated into a single funding stream. Funding for public health services was initially "unbundled" and given to a separate purchasing agent so as to ensure that these funds could not be spent instead on treatment services. However, these arrangements were short-lived and from 1996 the same purchasing authorities became responsible for both personal and public health services (although funding for public health services remained ear-marked specifically for this purpose). Between 1991 and 1997, the formal locus of responsibility for purchasing health services shifted from local purchasing (under 14 area health boards), to regional purchasing (under 4 regional health authorities) to central purchasing (under a single health funding authority) [13].
The nature and impact of these changes have been discussed in detail elsewhere [1,3,10,13]. Of relevance for this paper is the fact that, while funding for primary health care had effectively been merged with the funding of other services, New Zealand still lacked a coordinated national health system. The system was unplanned and often uncoordinated, service delivery was still fragmented, and problems of access to primary medical care due to high patient copayments remained. In addition, the Labour party, which won the right to lead a coalition government following the 1999 election, was ideologically opposed to "a model which promotes competitive tendering for contracts" [[14], p4].
Restructuring commenced once again early in 2000, the key aim being "the restoration of a non commercial system, with the focus on the provision of quality services" [[14], p.4]. Responsibility for purchasing services was transferred from the central purchasing agent (the Health Funding Authority) to the Ministry of Health as a temporary measure until the 21 DHBs could be established. As noted above, DHBs have a number of parallels with area health boards. However the range of services covered by DHBs is wider than those covered by area health boards. Most significantly, their responsibilities include primary health services as well as secondary and tertiary care. Some of the issues that have arisen during the process of devolving funds and decision-making to the DHBs are discussed in the next section.
Devolution of funds and decision-making to district health boards
Between 1993 and 1999 when the purchaser-provider split arrangements were in place, health services were purchased from providers largely on a cost-and-volume basis. A return to population-based funding at the district level therefore requires some redistribution of funds across the 21 districts. The formula that is now being used for determining a district's share of public funding takes into account the demographic make-up of each district, plus additional adjustments for unmet need, overseas visitors and the degree of rurality. Perceived problems with the population-based funding formula mean that many DHBs remain unhappy with their resultant quotas. These perceived problems include: systematically inaccurate population forecasts (primarily as a result of rapid internal migration); inadequate adjustment for people who have a high need for services but who historically may have under-utilised these services; no adjustment for new immigrants who often have special health service needs; and the possibility of 'medical migration' of people with on-going special service needs to the larger boards which are able to provide a more comprehensive service. There has also been much debate over the speed with which funds should be reallocated away from those boards which are over-funded to those that are under-funded. These issues have created some tensions amongst the DHBs. In the longer term, the aim of achieving equitable access to services for all New Zealanders may continue to be compromised if there are inherent inequities within the funding formula.
The initial establishment of the 21 DHBs reportedly went relatively smoothly [15]. In part, this was because, in an effort to minimise disruption, the number and size of the DHBs were configured to match the services and implicit boundaries of the publicly-owned organisations (previously called Hospital and Health Services). This pragmatic approach facilitated implementation of the new model and kept the associated costs down. However it also created a rather large number of boards for a population of only four million. A large number of DHBs increases transaction costs (for example, of the associated bureaucracy and of tracking the flows of patients across district boundaries) and may result in losses of potential economies of scale. It has also meant that scarce expertise is spread very thinly, especially in areas such as Mäori health and public health [[15], p.100]. The size of the 21 DHBs varies between about 33,000 and over 430,000 people [16]. The dynamics, management and issues facing these 21 organisations therefore vary significantly. The aim of the government to achieve equity of access to services for all New Zealanders in all regions may be difficult where there is such diversity amongst those organisations which are responsible for allocating the funds to providers within their districts.
While DHB establishment was itself relatively smooth, the process of devolving funds and decision-making to the districts has not been trouble-free. Devolution of funds involved, amongst other things, the transfer of responsibility for numerous contracts for services from the Ministry of Health to the DHBs. The DHBs found that many of the contracts were either inaccurate or incomplete and in some cases there were long delays in getting copies of the contracts [15]. The DHBs therefore did not have the information that they required to monitor the services being provided under the contracts.
Devolution of the funds for some services has also not occurred as soon as had originally been envisaged. By early 2005, the Ministry of Health still retained responsibility for the funding of all public health services, and for disability support services for people aged below 65 years. The ministry has also retained control over much discretionary spending, so that new money transferred to DHBs is sometimes already tagged for spending on specific services. Reasons for this lack of full devolution of funds are unclear. It may be associated with a limited capacity in some DHBs to manage more contracts. More fundamentally, it may reflect a perception on the part of the ministry that, notwithstanding the desire to encourage responsiveness to local needs and preferences, the purchasing of some services – particularly some public health services – may be better managed at the national level.
Even when funds have been fully devolved to the DHBs, this does not imply full autonomy by DHBs in decision-making. A key feature of the reforms this time around has been the development by the Ministry of Health of two overarching national strategies to provide strategic direction for the health and disability sectors as a whole and to ensure a degree of national consistency where decisions are decentralised [17,18]. Strategies have also been drawn up to guide the development of services for sub-groups of the population (such as Mäori and older people) [19,20] and for various sectors of the health system (such as primary health care) [5]. The 21 DHBs are required to adhere to the framework and priorities outlined in these national strategies when drawing up their annual plans and strategic plans, all of which must be signed off by the Minister of Health. However the boundaries between the responsibilities of the Ministry and of the DHBs are, as yet, by no means clear. While the DHB model shifts the locus of decision-making for the funding and provision or purchase of services to the district level, control over strategic direction by the central government constrains local decision-making. Moreover local and national priorities may sometimes be in conflict.
The minister has sometimes shown some reluctance to allow DHBs to make decisions about the provision of particular services in their areas, especially where this involves some disinvestment in services. For example, on one occasion the minister reversed a DHB's decision to stop providing after-hours surgical services in a rural hospital because of concerns about patient safety. This is seen by some DHBs as the minister interfering with local preferences [[15], p.98]. As the DHBs become more established and more experienced, the minister and the ministry may be more willing to allow the DHBs a greater degree of autonomy in decision-making. However, these tensions between the centre and the regions are by no means unique (or new) to New Zealand. Rather they reflect the hierarchical nature of a tax-funded system in which one organisation (the central government) is responsible for financing the system while other organisations (the DHBs) are responsible for spending the money.
Developments in the primary health sector
As noted above, recent developments in the primary health sector include changes both in the way that services are organised as well as in the method and level of subsidy for these services. The restructuring involves the grouping of general practitioners (GPs), primary care nurses and other primary health care providers under umbrella groups called Primary Health Organisations (PHOs). These PHOs are non-profit organisations which contract with DHBs to provide a comprehensive set of preventive and treatment services for their enrolled populations. PHOs are required to involve their communities in their governing processes, and to work with their enrolees to develop services that reflect their particular priorities and needs. In some districts, PHOs have been established on a geographic basis so that membership is determined by area of residence. In other districts, people can choose between two or more PHOs so that PHOs effectively compete both for GPs and for individual patients. PHOs – like DHBs – vary in size, from around 3,000 patients to over 330,000 [9].
With respect to the funding of primary health care, a key goal of the government's Primary Health Care Strategy is to remove the cost barrier that currently deters some people from seeking care. Government subsidies for general practice services (and also for pharmaceuticals) have historically been paid on a fee-for-service basis in New Zealand, with subsidies being targeted to low income and high risk people. Because the subsidy levels have not been regularly increased with inflation, and because GPs have retained the right to set their own levels of copayments, this has resulted in a significant cost barrier for some people to GP services [21,22]. In a national survey undertaken in 2002/03, around 6% of adults reported that they had not visited a GP within the last 12 months because of cost [22].
In an effort to remove or reduce this cost barrier, the move to PHOs is being accompanied by three changes to the way that government subsidies for primary care are paid:
• a change in the way that the general medical service subsidy is paid from fee-for-service payments to GPs to capitation funding of PHOs;
• the phased introduction of higher government subsidies for general practice services and pharmaceuticals for all New Zealanders [23];
• a shift from subsidies that are targeted towards high need individuals towards subsidies that are paid on a universal basis.
The bulk of government funding paid to PHOs is determined by two alternative capitation formulas, one of which – the 'Access' formula – provides a higher rate of subsidy. This formula applies only to those PHOs in which 50% or more of their enrolled population is either Mäori or Pacific, or living in a deprived area (as defined by the NZDep2001 Index which combines 8 census variables that reflect aspects of social and material deprivation.) All other PHOs are paid under the 'Interim' formula at a lower capitation rate. As the name suggests, payment to providers at this lower level is intended only as an interim measure, the intention being to gradually extend the higher rate of subsidy to all New Zealanders over the next few years. Higher rates of subsidy are now being paid to all PHOs for children up to the age of 17 years (since October 2003) and for those aged 65 years or over (since July 2004). Subsidies will be gradually increased for other age groups through until July 2007, at which time the higher capitation rates will apply to the whole population [23].
Payment to PHOs by the two different formulas has introduced some inequities into the system and caused considerable concern amongst many people working within the sector. Because the capitation payment covers all enrolees in a PHO, wealthier people who belong to PHOs that are paid under the Access formula will be paying less for GP consultations than poorer people who belong to other PHOs.
The differential in the subsidy levels has also stimulated aggressive – and sometimes acrimonious – competition between providers in some districts. It has encouraged PHOs to compete to enlist general practices that have a high proportion of deprived people on their registers. It has also encouraged individual practices to actively enrol particular patients as a means of obtaining eligibility for higher subsidies through the Access formula [24]. In those areas where there are both access and interim PHOs, patients have an incentive to shop around amongst GPs on the basis of price. From an administrative perspective, the move from fee-for-service reimbursements to capitation payment has meant that PHOs have had to tackle many technical difficulties and set up new management systems during the establishment phase [25]. All of these pressures have created a rather unstable environment which does not align with the government's vision of a primary health sector in which services are specifically tailored to meet the needs of a stable and identifiable population.
Preliminary evaluation of the impact of the subsidy increases on patient fees indicates that the fees being charged to patients by GPs have not always fallen by as much as might have been expected, had the subsidy increases been passed on in full to patients [26]. A survey of GP fees in February 2004 showed that GPs belonging to PHOs funded by the Access formula were generally charging all of their patients fees that are significantly lower than other PHOs [27]. However fees in other PHOs (i.e. those paid under the "Interim" formula) were generally higher than in GP practices which did not belong to a PHO at all. This indicates that the higher subsidy paid to PHOs has not always been passed on to patients as the government had hoped. A later survey found that the fees charged to people aged 65 and over fell by an average of 24% following the introduction of a patient subsidy for this group on 1 July 2004 [26]. However fees charged to these patients had increased by an average of 12% in the months prior to the introduction of the subsidy.
As a result of the information from this evaluation, the government is now working more closely with PHOs and District Health Boards in an effort to ensure that GP fees are set at reasonable levels. However, subsidy levels are not automatically adjusted in line with inflation. As long as GPs retain the right to set their own fee levels, universal low cost access to primary health care could prove difficult to achieve and to sustain in the longer term.
Private health insurance
Unlike Australia, private health insurance in New Zealand is unregulated and, since the abolition of the tax deductibility of premiums in the late 1980s, does not receive any direct financial assistance from the government. It has also not been a topic of any significant public debate. However, the issue is of interest, first, because government policy towards private insurance industry in New Zealand contrasts sharply with that in Australia, and second, because the industry is currently lobbying for change based upon much the same arguments as those that were used to support the introduction of the Private Health Insurance Incentive Scheme in Australia.
In October 2004, the insurance industry published two reports within two days, both of which lobbied for direct government assistance to private insurance as a means of enhancing efficiency, equity and choice within the health sector [28,29]. The first report was commissioned by the Southern Cross Medical Care Society, which provides health insurance to around two thirds of the people who have private health insurance cover in New Zealand. Written in collaboration with some Australians, the report claimed that the health insurance industry in New Zealand faces "serious decline" and that, without government assistance, health insurance coverage "may halve over the next 10 years" [[28], p.i]. The proposed solution was a 30% rebate on insurance premiums akin to that in Australia. The authors argued that the Australian experience ".....shows that a rebate on health insurance premiums has boosted coverage to a healthy level, reduced pressure on the public health system, improved the fairness of the health system (by the government paying some health costs of both insured and uninsured people) and generally secured the future of the health insurance industry" [[28], p.i].
The second report, published by the Health Funds Association of New Zealand (i.e. the body that represents the interests of the health insurance industry), claimed that "public health inflation is at record levels" [[29], p.1] and that such rates of increase "will quickly become unsustainable" [[29], p.13]. Based upon an estimated public health inflation factor of 8% per annum over the last 3 years, the report projected that public health expenditure would reach 63% of GDP by 2050! The report went on to suggest that costs could be contained if contestable funding was to be introduced into the DHB system. It proposed that people earning in excess of NZ$38,000 should be required to purchase their own health insurance, with the government refunding the amount of their tax that would otherwise have been used to purchase health services. Any contributions made by employers should be exempt from the fringe benefit tax. DHBs would then sell their services to those people who are privately insured at a price equal to the true cost of the service. The report claimed that such a system would encourage both DHBs and individuals (or their employers) to focus on value for money, thereby providing the necessary incentives to keep health inflation down.
Interpretation of the data which form the basis for the claims made in each of these reports is highly questionable. This comment applies both to the evidence presented to illustrate that there is a problem with current financing arrangements in New Zealand, as well as to the impact of the proposed solutions. Even the basic premise that private health insurance is in serious decline is not well supported. While the proportion of the population covered by private insurance has indeed declined from a peak of around 45% in the late 1980s [30] it has remained fairly stable at around one third of the population over the past five years [28]. And although the proportion of insured with comprehensive cover (as opposed to cover mainly for hospital services) has declined from 20% in 2000 to 14% in 2004 [28], this could equally reflect improvements in access to publicly-funded primary health services as much as a response to increases in insurance premiums.
It is also difficult to accept that New Zealand should follow Australia in introducing a 30% tax rebate on premiums. In Australia, the rebate was one part of a package of subsidies and regulations. Separating out the precise effects of the rebate from the effects of the other components of the package is problematic and requires the adoption of a number of assumptions. Even so, there appears to be some consensus amongst analysts that, while the rebate does appear to have stimulated an increase in insurance coverage in the short term, the size of the effect may have been less than the government expected, and may not have been large enough to justify the expenditure [31,32]. Moreover, recent figures suggest that health insurance coverage in Australia is now declining [33].
It is even more difficult to justify a subsidy on the grounds that it will reduce pressure on the public health system. As Richardson recently pointed out in this journal, in Australia, changes in public and private bed numbers indicate that problems of access to the public health system are determined primarily by constraints on the supply side, rather than by an excess demand caused by an inability to afford private health insurance [34]. Vaithianathan, too, has shown that the demand for public hospital beds is unlikely to decline because an insurance subsidy is most likely to increase insurance coverage of people who previously paid directly for the use of private hospital beds, rather than of people who currently use public hospitals [35]. Even if a subsidy does actually encourage a shift from the public sector to the private sector, Frech III and Hopkins have suggested that, from a theoretical perspective, the optimal subsidy may actually be negative (i.e private health insurance should be taxed) [32].
In the second report [29], while the main justification for a greater role for private health insurance was escalating costs in the public sector, the meaning of the term "public health inflation" was unclear. In some instances [e.g. [29], p.12], the term seems to apply to changes in public health expenditure, while in other cases it apparently refers to increases in public hospital costs adjusted for hospital throughput [[29], p.9]. Neither of these are good indicators of cost increases across all of the services that are publicly funded, but either way, the estimated figure appears to have been simply extrapolated to the year 2050, thus producing an estimate of public health expenditure that is both excessive and, more importantly, newsworthy. Even if such a figure could be substantiated, there is little evidence from the international literature in support of the claim that contestable insurance funding is likely to assist in controlling costs. If anything, total health expenditure tends to be higher in insurance-funded systems than in systems that are predominantly funded by general taxation [36,37]. Reasons for this higher expenditure include the difficulty of containing costs in a system where there are multiple purchasers, and where reimbursements are usually on a fee-for-service basis.
While the government did not respond publicly to the claims and proposals made by the health insurance industry in these two reports, it did take two decisive actions. First, it requested briefings from the Treasury on both of the reports immediately prior to their public release [38,39]. Second, on the same day that the second report was published the Ministry of Health released its own report on the future funding of health services in New Zealand [40]. One of the conclusions from this report – which had in fact been written for the Ministry two years earlier but which had not been released – was that "there should be no public subsidies of private health insurance in New Zealand" [[40], p.xiv]. The main reasons behind this conclusion were (a) inequalities are likely to be exacerbated, because expenditure on private insurance increases with income; (b) control over health expenditure is more difficult under private insurance than under direct public funding; (c) greater value is likely to be achieved by increasing expenditure in the public sector because service provision tends to be more expensive and administration costs tend to be higher in the private sector; and (d) because demand for private insurance is relatively insensitive to price changes, the cost of a health insurance subsidy will be greater than the value of any health services that are stimulated by that subsidy.
In summary, while a rebate on private insurance may well improve the health of the private insurance industry, there appears to be little evidence that it would make any useful contribution towards improving the health of New Zealanders. More fundamentally, the current government would require any changes to financing arrangements to align with the principles which underpin the New Zealand Health Strategy. One of these principles is: "timely and equitable access for all New Zealanders to a comprehensive range of health and disability services, regardless of ability to pay" [[17], p.vi]. As Richardson has noted, the egalitarian desire of equalising access to health care regardless of ability to pay, and reducing inequalities in health are "more easily achieved through a compulsory public health system" [[34], p.5]. In contrast, contestable funding and subsidisation of private insurance are more appropriate for a health system aimed at maximising individual choice.
Conclusion
This paper has described and discussed just some of the issues that currently face the New Zealand health system. Many of these issues are not new but rather are renewed manifestations of debates which have been recurrent features in various waves of health sector reform. In particular, the division of responsibilities between the centre and the regions, uncontrolled copayments for GP consultations, and the tax treatment of private health insurance are all issues which have repeatedly challenged decision-makers.
During any period of change, there are inevitably conflicts, tensions and disagreements as new boundaries are drawn and the various players jostle for position. This certainly has been, and continues to be, the case in this latest round of reforms in New Zealand, most particularly in the primary health care sector. As the system matures, the boundaries of responsibility across the sector should become clearer. If the vision of a community-oriented system is to become a reality, full devolution of responsibilities from the centre to the districts will be essential as the DHBs continue to build their capacity, capability and experience. The Ministry of Health can then focus on providing strategic direction to the sector and on monitoring performance through appropriate accountability mechanisms.
In spite of the tensions and difficulties associated with implementation of the new structure, the sector already appears to be developing a much stronger sense of direction and purpose. This is in sharp contrast to the 1990s when the change towards a more market-oriented system resulted in a high degree of uncertainty for, and alienation of, many service providers [41]. Another difference is that the government has invested in a number of evaluation projects by health service researchers this time around. These evaluations should highlight the main strengths and weaknesses of the new institutional arrangements, and, where necessary, indicate areas where further adjustments are required.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
The useful comments of two referees on an earlier draft are acknowledged with appreciation.
==== Refs
Cumming J Salmond G Ranade W Reforming New Zealand health care Markets and Health Care 1998 London and New York, Longman 122 146
Easton B The New Zealand health reforms of the 1990s in context Applied Health Economics and Health Policy 2002 1 107 112 14619257
Ashton T Running on the spot: lessons from a decade of health reform in New Zealand Applied Health Economics and Health Policy 2002 1 97 106 14619256
Hornblow A New Zealand's health reforms: a clash of cultures British Medical Journal 1997 314 1892 1894 9224137
King A The Primary Health Care Strategy 2001 Wellington, Ministry of Health
Ministry of Health Health Expenditure Trends in New Zealand 1990 - 2002 2004 Wellington, Ministry of Health
Ministry of Health Health Expenditure Trends in New Zealand 1980-2000 2002 Wellington, Ministry of Health
Johnston M Threat of more health cover rises New Zealand Herald 2000 Auckland, A5
Ministry of Health Established Primary Health Organisations
Gauld R Revolving Doors: New Zealand's Health Reforms 2001 Wellington, Institute of Policy Studies 254
Smith PA The Private Prescription, the Story of Southern Cross Health Care 1994 1st Auckland, Southern Cross 251
McGuigan T A Health Service for New Zealand 1975 Wellington, Government Printer
Ashton T Mays N Devlin N Continuity through change; the rhetoric and reality of health reform in New Zealand Social Science and Medicine
New Zealand Labour Party Focus on Patients: Labour on Health 1999 Wellington, New Zealand Labour Party
Health Reforms Research Team Interim Report on Health Reforms 2001 Research Project
Ministry of Health Frequently Asked Questions About District Health Boards
King A The New Zealand National Health Strategy 2000 Wellington, Ministry of Health
Ministry of Health The New Zealand Disability Strategy: Making a World of Difference: Whakanui Oranga 2001 Wellington, Ministry of Health
King A Turia T He Korowai Oranga: Maori Health Strategy 2002 Wellington, Ministry of Health
King A Health of Older People Strategy: Health Sector Action to 2001 to Support Positive Ageing 2002 Wellington, Ministry of Health
Schoen C Osborn R Trang Huynh P Doty M Davis K Zapert K Peugh J Primary Care and Health System Performance: Adults' Experiences in Five Countries Health Affairs Web Exclusive
Ministry of Health A Portrait of Health 2004 Wellington, Ministry of Health
King A Cabinet Paper: Primary Health Care Strategy - Achieving Low Cost Access
CBG Health Research Ltd. Reducing Inequalities Contingency Funded Projects 2003 Wellington, Ministry of Health
Perera R MacDonald J Cumming J Goodhead A Primary Health Organisations: The First Year (June 2002 - June 2003) from the PHO Perspective 2003 Wellington, Health Services Research Centre, Victoria University
Ministry of Health General Practitioner Fees Information: A Summary of Key Findings from Five Reports
CBG Health Research Ltd. National GP Fee Survey: A report prepared for the Ministry of Health 2004 Wellington, Ministry of Health
Econtech Pty Ltd in association with Harper Associates Choice and Affordability - for Healthcare in New Zealand. A report prepared for Southern Cross Medical Care Society
von Lanthen K The Long Term Funding of New Zealand Healthcare 2004 Wellington, Health Funds Association of New Zealand
Department of Statistics New Zealand Official 1990 Yearbook 1990 Wellington, Department of Statistics
Vaithianathan R A critique of the private health insurance regulations Australian Economic Review 2004 37 257 270 10.1111/j.1467-8462.2004.00328.x
Frech III HE Hopkins S Why subsidize private health insurance? Australian Economic Review 2004 37 243 256 10.1111/j.1467-8462.2004.00327.x
Hall J Election 2004: Paying for Health Care Australian Review of Public Affairs
Richardson J Priorities of health policy: cost shifting or population health Aust New Zealand Health Policy 2005 2 1 15679895 10.1186/1743-8462-2-1
Vaithianathan R Will subsidizing private health insurance help the public health system? Economic Record 2002 78 277 283 10.1111/1475-4932.00057
OECD OECD Health Data 2000: A Comparative Analysis of 29 Countries 2000 Paris, Organisation for Economic Cooperation and Development
Mossialos E Dixon A Mossialos E, Dixon A, Figueras A and Kutzin J Funding health care in Europe: weighing up the options Funding Health Care: Options for Europe 2002 Buckingham, Open University Press
New Zealand Treasury Ministerial Briefing - Choice and Affordability: a Proposed Rebate for Health Insurance (Southern Cross) 2004 Wellington, New Zealand Treasury
New Zealand Treasury Treasury Report on Proposals for Funding of NZ Health Care 2004 Wellington, New Zealand Treasury
Ministry of Health Future Funding of Health and Disability Services in New Zealand: Report to the Director General of Health, April 2002 2002 Wellington, Ministry of Health
Ashton T The rocky road to health reform: some lessons from New Zealand Australian Health Review 2001 24 151 156 11357730
| 15877820 | PMC1142303 | CC BY | 2021-01-04 16:38:28 | no | Aust New Zealand Health Policy. 2005 May 7; 2:9 | utf-8 | Aust New Zealand Health Policy | 2,005 | 10.1186/1743-8462-2-9 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1161589008210.1186/1471-2105-6-116SoftwarePredicting functional sites with an automated algorithm suitable for heterogeneous datasets La David [email protected] Dennis R [email protected] Department of Biological Sciences, California State Polytechnic University, Pomona, California 91768 USA2 Department of Chemistry and Center for Macromolecular Modeling & Materials Design, California State Polytechnic University, Pomona, California 91768, USA2005 13 5 2005 6 116 116 13 1 2005 13 5 2005 Copyright © 2005 La and Livesay; 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
In a previous report (La et al., Proteins, 2005), we have demonstrated that the identification of phylogenetic motifs, protein sequence fragments conserving the overall familial phylogeny, represent a promising approach for sequence/function annotation. Across a structurally and functionally heterogeneous dataset, phylogenetic motifs have been demonstrated to correspond to a wide variety of functional site archetypes, including those defined by surface loops, active site clefts, and less exposed regions. However, in our original demonstration of the technique, phylogenetic motif identification is dependent upon a manually determined similarity threshold, prohibiting large-scale application of the technique.
Results
In this report, we present an algorithmic approach that determines thresholds without human subjectivity. The approach relies on significant raw data preprocessing to improve signal detection. Subsequently, Partition Around Medoids Clustering (PAMC) of the similarity scores assesses sequence fragments where functional annotation remains in question. The accuracy of the approach is confirmed through comparisons to our previous (manual) results and structural analyses. Triosephosphate isomerase and arginyl-tRNA synthetase are discussed as exemplar cases. A quantitative functional site prediction assessment algorithm indicates that the phylogenetic motif predictions, which require sequence information only, are nearly as good as those from evolutionary trace methods that do incorporate structure.
Conclusion
The automated threshold detection algorithm has been incorporated into MINER, our web-based phylogenetic motif identification server. MINER is freely available on the web at . Pre-calculated functional site predictions of the COG database and an implementation of the threshold detection algorithm, in the R statistical language, can also be accessed at the website.
==== Body
Background
Due to the exponential growth of genomic and protein sequence data, development of automated strategies for large scale functional site identification is an important post-genomic challenge. Many recent efforts predict functional sites from sequence alone. Strong candidates for functional sites include individual highly conserved positions within a sequence alignment and highly conserved sequence motifs [1-5]. Although attractive due to their relative simplicity, conservation-based approaches frequently result in too many false positives to be satisfactory [3]. Further, sequence regions with significant variability can also be functionally important [6], especially when their composition may define sub-family functional specificity. The Evolutionary Trace (ET) procedure [7], and similar approaches [6,8,9], address this problem by using evolutionary information to identify subfamily conserved, yet overall variable, positions. It has been demonstrated that the ET methods identify statistically significant structural clusters [10] and has subsequently been established in large scale analyses [11].
Recently, we demonstrated that sequence fragments conserving the overall phylogeny, termed phylogenetic motifs (PMs), are very likely to correspond to protein functional sites [12]. We briefly highlight the key results of our previous report here (see Implementation for a technical description of the approach). Despite little overall sequence proximity, PMs are structurally clustered around a wide variety of protein functional regions, including sites defined by surface loops, active site clefts, substrate binding epitopes, and protein-protein interfaces. Ostensibly, PMs identify sequence clusters of ET positions, and, as expected, the results of the two approaches are similar. However, compared to raw ET predictions, PMs seem to be more structurally focused on active site regions. Lastly, we have demonstrated appreciable tree significance of the fragment windows, especially in PM regions, using bootstrap analysis.
In a recent review, Jones and Thornton [13] classify protein functional site prediction strategies into one of two groups: (1) those based on sequence conservation and (2) those based on "feature" (i.e. phylogenetic) conservation. Congruence between phylogenetic and traditional motifs has been clearly established [12]. As a consequence, PMs bridge the two normally disparate groups. For example, in the case of triosephosphate isomerase, all seven traditional motifs are also identified as PMs, including the PROSITE [14] definition of the family. However, due to the large number of clades within some families (e.g. cytochrome P450) there are instances when PM sequences are not overall conserved. This point is in enticing because it suggests that PMs can sometimes functionally annotate regions that traditional methods would ignore. Furthermore, it has been reported [12] that traditional motifs not conserving phylogeny are less likely to be structurally clustered around known functional sites.
PMs are identified using a sliding sequence window algorithm that exhaustively compares the phylogenetic similarity of each fragment window to the complete familial tree. Phylogenetic similarity z-scores (PSZs), which are defined below, quantify the similarity between the window and familial tree. (Smaller PSZs indicate greater phylogenetic similarity.) In our current approach, all overlapping windows scoring past some PSZ threshold are identified as a PM. The PSZ threshold is manually adjusted to define what constitutes a "hit". Automated threshold determination is a difficult problem because no single threshold value is sufficient for every instance. Each protein family requires a unique value to correctly identify a signal (a PM window) from noise. In our previous work, we manually set threshold values where PSZs strongly deviate from all other values. Structural analyses indicate that ideal PSZ threshold ranges between -1.5 and -2.0. In other words, the ideal phylogenetic similarity cutoff for accurate functional site predictions generally falls between 1.5 and 2 standard deviations away from the mean of the PSZ distribution.
In this report, we describe an automated algorithm for determining proper PSZ thresholds. Structural analyses and comparisons to our previous manual thresholds indicate that the approach retains functional site prediction accuracy. The method utilizes significant raw data preprocessing that eliminates unnecessary (redundant) data points. Subsequently, the robust Partition Around Medoids Clustering (PAMC) algorithm is employed to automatically determine the appropriate PSZ thresholds. The implemented approach is computationally efficient and demonstrated to be suitable for large, heterogeneous datasets, which remains a difficult bioinformatic problem [15].
Implementation
Phylogenetic motif identification
During PM identification, we mask the multiple sequence alignment by purging all highly (more than 50%) gapped positions. The masked alignment is parsed into a series of sequence fragment windows of fixed width. In this report, a window width of five, which has previously been shown to be most sensitive for identifying functional sites [12], is used throughout. Except for the copper, zinc superoxide dismutase and myoglobin families, which use the same datasets as before [12], all sequences are taken from the Clusters of Orthologous Groups (COGs) database [16]. Only COGs with more than 25 sequences are investigated to ensure proper and significant tree construction. Pairwise tree similarity is calculated using a modified partition metric algorithm [17], which counts the number topological differences between the fragment window and familial trees. Thus, smaller partition metric scores correspond to greater tree similarity. Phylogenetic similarity is measured using z-scores calculated from the raw partition metric distribution. Although not the best alignment method for distantly related sequences, ClustalW (v1.83) alignments are used throughout [18]. Given the similarity within the COG families, ClustalW alignments are satisfactory.
Phylogenetic trees are constructed using the neighbor-joining implementation within ClustalW. Neighbor-joining is a distance-based approach for constructing phylogenetic trees commonly used for bootstrap analysis that requires massive tree sampling [19]. Similarly, due to the large number of window trees required here, the algorithmic efficiency of distance-based methods is necessary. For example, in the medium-sized triosephosphate isomerase protein family, over 250 trees must be calculated. Furthermore, as Kuhner and Felsenstein point out [20], distance-based approaches actually outperform maximum-likelihood methods when applied to short sequences. MINER, our web-based implementation of the PM identification algorithm, is available online at [21]. A standalone version of MINER, implemented in PERL, is freely available to the Academic community upon request.
Raw data preprocessing
Empirically, our manual assessment of functional site prediction accuracy indicates that all PSZs below χ ≈ -2 should be identified as PMs, whereas PSZ's above -1 should never be considered. In the subsequent clustering step, only scores between -1 and χ are used to define the PSZ. This simplification is taken because the objective of this work is to automatically classify windows whose determination remain in question. The significance of the PSZs outside this range is known a priori, thus they can be eliminated from further consideration. Clustering of the data points between -1 and χ (termed the "gap") was originally expected to automatically determine the appropriate PSZ threshold. However, several different clustering techniques (hierarchical, k-means, PAMC, and expectation-maximization) have failed to provide satisfactory results.
In order to accentuate differences within the PSZ distribution, and thus simplifying the clustering problem, the following preprocessing procedure is employed. As stated above, all overlapping windows scoring past some predetermined threshold are defined as PMs. We use the same rationale in detecting PSZ thresholds. The process begins by identifying all overlapping windows scoring past -1. For the purpose of threshold detection only, we "sharpen" these regions by selecting the lowest window score as a reference; all other scores are eliminated. This process has the effect of reducing the number of contiguous and related PSZ scores (corresponding to overlapping windows) into a single value (see Figure 1). After accentuating the high phylogenetic similarity regions, PAMC can robustly identify the ideal PSZ threshold. Several different upper bounds have been considered, but our empirical results indicate that -1 is best.
Automated PSZ threshold determination
PAMC is a partitioning algorithm based on the k-means approach of clustering [22]. In k-means clustering [23], the center of a cluster is represented by its arithmetic average. In PAMC, each cluster is represented by the median value, making PAMC a k-medoids approach. The basic concept of PAMC is to partition a dataset containing a number of points into k number clusters. PAMC starts from an initial random set of medoids and iteratively swaps medoids with non-medoids to evaluate if the total distances between clusters are improved. PAMC is more effective and robust than k-means for small datasets because medians are less biased by outliers deviating from the mean. Our PSZ dataset is small, especially after preprocessing, making PAMC an appropriate clustering choice. In addition, because the goal is to separate signals (cluster one) from noise (cluster two), k-based approaches are ideal. The preprocessed gap is differentiated into k = 2 number of groups. Clustering is performed by the PAMC implementation within the cluster package of the R statistical language [24]. The PAM algorithm implemented in R simply uses the Euclidean measure by default and the Manhattan as a defined alternative. We use the Euclidean measure throughout.
In determining the ideal PSZ threshold value, the number of data points in the signal cluster is counted. If the signal cluster contains five or less data points, the threshold is set to the most accommodating (least negative) value in that cluster. However, an algorithmic override that defines the PSZ threshold at the first (rank ordered) PSZ above χ if any of the following three situations occur: (1) if the signal cluster contains more than five data points, (2) if less than three points reside in the gap – it does not make sense to cluster so few data points into two groups, or (3) no PSZs lower than -2 are present within the distribution. The algorithmic override prevents normalizes the number of putative functional sites, preventing both too many and too few predictions. Empirical results investigating the accuracy of the method's predictions vis-à-vis structure indicate that the algorithmic override maximizes accuracy. For example, in examples where the override reduces the number of predictions, frequently the excluded sites are structurally removed from the active site region. The ideal value of χ is established in the next section.
Quantitative assessment of functional site predictions
The accuracy of the functional site predictions herein is quantitatively determined using the method put forth by Aloy et al. [25]. In the scheme, a known functional site sphere is defined by the location of SITE and ACTSITE records within a PDB file. In line with our previous report [12], we also include residues directly interacting with substrates and catalytically important metal ions, which are identified using LIGPLOT [26]. Prediction spheres are similarly constructed for each PM. The accuracy of each PM prediction is based on the relative location of the known functional site and prediction spheres. If the prediction sphere is completely engulfed within the known functional site sphere, then the prediction is deemed correct. If the two spheres partially overlap, then the prediction provides useful information. And, if there is no overlap between the two spheres, then the prediction is wrong. In the case of the known functional site, the sphere is centered on the geometric center of the CB atoms (CA for glycine) of all functional residues. Similarly, the PM sphere is centered on the geometric center of the corresponding CB atoms (CA for glycine). In both cases, the sphere is made just large enough to include all functional or PM residues.
Results and discussion
Establishing algorithm parameters
As described below, the automatic threshold determination problem is simplified when considering only the most extreme of several contiguous and related, PSZ values. We call this process data sharpening. As a consequence, identification of the true phylogenetic signals is greatly simplified. In contrast, contiguous windows with similar values are the result of a single PM, making it difficult to properly count the number of true signals in an unsharpened dataset. Because PMs are defined as all overlapping windows scoring past the PSZ threshold, reducing the complexity of the problem to be in line with the number of PMs, versus number of windows, makes intuitive sense. For example, if two PMs are considered, the first consisting of three overlapping windows and the second with five, only two unique signals, compared to the eight constituent windows, are considered. Comparisons of sharpened and unsharpened datasets are demonstrated in Figure 1.
Thresholds are determined by first evaluating the optimal range using the PAM clustering algorithm. Determining whether a threshold can be placed within the range of -1.0 and -2.0 allows thresholds to sensitively accommodate more functional sites, widening this range results in more stringent thresholds. Figure 1C illustrates different thresholds determined when considering three different PSZ ranges. By broadening the range of the triosephosphate isomerase (TIM) dataset, two distinct thresholds are found. Ranges of {-1.0: -2.0} and {-1.0: -2.5} identify the same threshold (PSZ = -1.65). However, expanding the gap to {-1.0: -3.0} results in a significantly more stringent threshold (PSZ = -2.86). The former PSZ threshold is more similar to our manual determination of PSZ = -1.5 [12].
Similarly, the arginyl-tRNA synthetase family is evaluated using the same three gap ranges. Like TIM, the determined threshold becomes more stringent as the gap broadens. However, the three gap ranges result in three distinct PSZ threshold values. The arginyl-tRNA synthetase example is noteworthy because the determined threshold in all three instances is algorithmically set below χ. This occurs because too many points exist in the PAM identified signal cluster. Since we assume functional sites cover only a fraction of the protein sequence space, when the signal cluster is larger than five, it is disregarded and the PSZ threshold is set at the first (rank ordered) PSZ past χ.
Comparison of all three gap ranges on the determined PSZ threshold for the 15 functionally and structurally diverse proteins used previously [12] is partially used to determine the ideal gap range. In all cases, the functional significance of the manually determined threshold has been demonstrated using structural analysis. Additionally, the exact catalytic role of many of the identified PMs (especially TIM [27], enolase [28], inorganic pyrophosphatase [29], copper, zince superoxide dismutase [30], and TATA-box binding protein [31]) has also been delineated. For example, eight PMs are identified in the case of TIM, which cover all eight LIGPLOT [26] identified electrostatic interactions between enzyme and substrate. Furthermore, the flexible "lid" region, which covers the active site during catalysis [32-35], is also identified as a functional site. We have also recently demonstrated that PMs within TIM (and two other TIM-barrel families) also correspond to evolutionarily conserved electrostatic networks that fine-tune the pKa values of catalytic residues [36].
The frequency of threshold values (displayed as kernel density estimates) determined using the three different gap ranges is illustrated in Figure 2 for the entire COG database [16]. Of the three PSZ ranges tested, the threshold distribution resulting from the narrowest range is most similar to the distribution of our structurally verified dataset. Using a two-sample t-test, the statistical significance between the PSZ threshold distributions can be assessed. The t-test results (t = -0.41, p = 0.69) indicate that the manual and {-1.0: -2.0} distributions are not statistically different. However, when evaluating the manual threshold distribution with the other two gap ranges, (t = -6.08, p = 2.14 × 10-05) and (t = -9.49, p = 6.85 × 10-08) for {-1.0: -2.5} and {-1.0: -3.0}, respectively, we find that the differences are highly significant. Furthermore, the distribution of thresholds from the {-1.0: -2.0} show the most frequent PSZ thresholds are set around -1.5 and -2.0, which is in line with our original conclusions. A gap range of {-1.0: -2.0} is used throughout the remainder of this report. Table 1 compares the manually and automatically determined PSZ thresholds.
Functional annotation of the COG database
Using the procedure established above, we exhaustively functionally annotated the most recent update [16] of the COG database. After parsing out COGs with less than 25 sequences, our dataset is composed of 1571 protein families. The number of PMs identified resembles a bell curve centered on 6.1 motifs per COG (Figure 3A). The standard deviation is 2.9. 24 PMs, the most of any COG investigated, are identified within the cobalamin biosynthesis protein family. Due to the extreme size of this protein, the number of identified PMs is within the expected range – the cobalamin biosynthesis protein family alignment is the second longest in our dataset. Consistent with our earlier qualitative observations [36], Figure 4 reveals a roughly linear correlation between alignment length and the number of phylogenetic motifs identified per COG.
In total, 9558 PMs are identified. Compared to the number of PMs per COG, there is much more heterogeneity within the motif width distribution (Figure 3B). The theoretical lower bound on PM width is five (one fragment window); whereas there is no limit on their maximum size. A motif width of five is by far the most common, occurring 51% of the time. The maximum width observed, which occurs in the methyl-accepting chemotaxis protein family, is 42 (occurring once). The large motif corresponds to the chemotaxis transduction 2 domain. As a stark contrast, only one other PM (width = 5) is identified within this family. The second, and much smaller, motif coincides with the PROSITE [14] definition (R-T-E- [EQ]-Q) of the family. The [EQ] position is a site of reversible methylation.
The large-scale nature of this analysis provides an opportunity to assess the dependence of several factors on the automatically determined PSZ thresholds. Correlations between the determined PSZ thresholds and number of PMs identified, number of sequences in the dataset, and alignment length are calculated (see Table 2). As discussed, a roughly linear (R = 0.68) correlation between number of PMs identified and alignment length is identified (Figure 4). However, no other strong correlations are identified between any of the probed characteristics. While more-or-less uninteresting, this result is actually encouraging because it indicates that PMs, in addition to being accurate, represent a robust functional site prediction algorithm suitable for large, heterogeneous datasets.
Molecular examples
Clustering of the three different TIM gap ranges uncovers two putative PSZ thresholds (see Figure 1C). We demonstrate above that the narrowest gap range (and as a consequence, the most lenient PSZ threshold) to be appropriate. In this case, however, both the {-1.0: -2.0} and the {-1.0: -2.5} gap ranges set the threshold at -1.83. The determined threshold sensitively accommodates the complete substrate binding epitope, including all eight enzyme-substrate electrostatic interactions. However, the {-1.0: -3.0} range identifies a more stringent threshold (-2.75), which misses one enzyme-substrate salt bridge and one hydrogen bond. Several other less drastic differences also occur. A structural analysis of these automatically set thresholds is shown in Figure 5. Despite the differences between the two thresholds, both identify PMs that correspond to the PROSITE [14] definition of the family, the flexible "lid", and most of the enzyme-substrate contacts. In both cases, all identified PMs are structurally clustered at the C-terminal end of the barrel.
We also structurally verify functional site prediction accuracy within the arginyl-tRNA synthetase family, which is a previously unreported example. As with TIM, we evaluate the same three gap ranges. Structural verification and comparison of these three thresholds is illustrated in Figure 6A–C. The more accommodating PSZ threshold identifies two structurally unique PM clusters. The first is composed of four PMs, and corresponds to the enzyme active site. Several stabilizing enzyme-tRNA and enzyme-Arg interactions are included in this region (Figure 6D). The second PM structural cluster is composed of a single PM, and corresponds to three enzyme-tRNA H-bonds at the tRNA anticodon arm [37]. Making the PSZ threshold more stringent, by widening to the gap range to {-1.0: -2.5}, eliminates two PMs, including the anticodon arm PM. Only one PM is identified at the most stringent level. Like with TIM, the ends of the identified PMs are trimmed at increasingly stringent PSZ thresholds.
Assessment of functional site predictions
The quantitative assessment of functional site predictions from computational predictions remains an open problem in bioinformatics. Much of the difficulty arises from the fact that function and more specifically, functional sites, are ill defined concepts [38]. Aloy et al. [25] have attempted to standardize assessment of functional site predictions through comparisons to catalytically important residues (see Methods for technical details). Predicting functional sites with an automated evolutionary trace method [7] utilizing structural information on 86 protein families, Aloy et al. demonstrate impressive results: 79% correct, 15% useful information, and 6% Wrong. We use the same assessment strategy on 30 sampled protein families from the COG database (see Table 3). Fourteen of the analyzed families correspond to our earlier manual analysis [12] and the remaining are arbitrarily picked from examples with at least one solved structure. Figure 7 demonstrates that the PM functional site predictions are of similar quality to the overall accuracy reported by Aloy et al. Note that the dataset analyzed in Aloy et al. is not the same as the dataset analyzed here. This result is particularly encouraging due to the lack of structural details in the PM technique.
As Aloy et al. [25] point out, while the quality and robustness of SITE and ACTSITE records within PDB files is of varying quality, their approach does provide an automated and unbiased method for assessing functional site predictions. However, in automated efforts, examples of known "functional sites" will always be missed. For example, assessment of the TIM PM predictions indicates that 3 are correct, 4 provide useful information, and 1 is wrong. The "wrong" prediction actually corresponds to an evolutionarily conserved dimer interface epitope (see Figure 5) that includes several stabilizing monomer-monomer interactions [39]. Despite being far removed from the active site, binding of a small molecule at the dimer interface can inactivate the enzyme [40]. One of the three enzyme-inhibitor contacts occurs from Phe75, which is a residue within the "wrong" PM prediction. As a consequence, it could be argued that this PM is functional. This discussion is included here to encapsulate the ambiguity involved in functional site definitions and the difficulty in assessing their predictions.
Conclusion
In this report, we present an automated algorithm which determines appropriate PSZ thresholds appropriate to functional site predictions. We demonstrate that our methodology is robust enough for large-scale analyses, while sensitive enough to identify known functional sites. For example, the method predicts all structural contacts, including the catalytic residue, between triosephosphate isomerase and its substrate. Additionally, the functionally important flexible "lid" is also identified. In the case of arginyl-tRNA synthetase, PMs correspond to regions surrounding both the amino acid/tRNA acceptor stem and enzyme-anticodon interactions. Using a quantitative structure-based functional site assessment algorithm, we demonstrate that the sequence-only PM predictions compare favorably to those from evolutionary trace approaches that are dependent upon solved structures.
Availability and requirements
• Project name: MINER
• Project home page:
• Operating system(s): Platform independent
• Programming language: PERL
• Other requirements: Chime
• License: GNU GPL
• Any restrictions to use by non-academics: License needed
List of abbreviations
Evolutionary trace (ET); Phylogenetic motif (PM); Phylogenetic similarity z-score (PSZ); Partition around medoids clustering (PAMC); Clusters of orthologous groups (COG); Triosephosphate isomerase (TIM).
Authors' contributions
D. La was primarily responsible for the development and testing of the described methodology. D.R. Livesay oversaw the research. Both authors contributed equally to the writing of this manuscript.
Acknowledgements
The authors would like to thank the reviewers for making us aware of the structure-based functional site assessment algorithm described in Aloy et al., J. Mol. Biol., 2001. Shankar Subramaniam is acknowledged for a number of valuable suggestions and proof-reading the manuscript. This work was partially supported by an American Chemical Society Petroleum Research Fund grant (36848-GB4), a National Science Foundation MRI-grant (0321333), and a supercomputer allocation (MCB00018N) from the National Center for Supercomputing Applications to D.R.L.
Figures and Tables
Figure 1 Partitioning Around Medoids Clustering (PAMC) is used to partition PSZs between -1.0 and χ = -2.0 into two groups (signal and noise clusters). The results from the (A) raw and (B) sharpened PSZ datasets are shown. Red indicated sharpened data points. We demonstrate two common scenarios (left) triosephosphate isomerase and (right) arginyl-tRNA synthetase. In all cases, it is clear that sharpening the PSZ dataset allows one to more easily discern the number of distinctive potential signals under the partition boundary (gap) in question. (C) The effect of different sharpening ranges is demonstrated. PSZ ranges tested include -1.0 to -2.0 (red), -2.5 (dark orange), and -3.0 (light orange). In the case of triosephosphate isomerase the first two ranges give identical results. The ideal threshold is found to be -1.65 and -2.20 for triosephosphate isomerase and arginyl-tRNA synthetase, respectively.
Figure 2 The frequencies of the three different PSZ thresholds are found on 1571 protein families (all COGs with more than 25 sequences). The normalized frequencies are represented as kernel density estimates. All optimal ranges used consist of two maximum densities from -1.0 to -3.0 (PSZ = -3.04 and -1.82), from -1.0 to -2.5 (PSZ = -2.60 and -1.74), and from -1.0 to -2.0 (PSZ = -2.01 and -1.58). Kernel densities from fifteen manually set thresholds that have been structurally verified also contain two maximum densities (PSZ = -2.12 and -1.53). Of the three ranges tested, the {-1.0: -2.0} gap is the only one that is statistically similar to the structurally verified results. The coloring of the three ranges is the same as in Figure 2C.
Figure 3 (A) Histogram tabulating the number of phylogenetic motifs per COG. (B) Histogram tabulating phylogenetic motif width (note logarithmic scale). The theoretical lower bound on motif width (five, which equals one fragment window) occurs 51% of the time. The {-1.0: -2.0} gap range is used in both cases.
Figure 4 Number of phylogenetic motifs per COG, using the {-1.0: -2.0} gap range, is plotted against alignment length. As expected, there is a direct correlation between the two quantities. At least one phylogenetic motif is identified in each COG. Conversely, in large COGs 10+ phylogenetic motifs are routinely identified.
Figure 5 Triosephosphate isomerase phylogenetic motifs identified at the two different PSZ thresholds. Colored spheres indicate phylogenetic motif α-carbons identified past the (A) -1.83 and (B) -2.75 thresholds, which correspond to gap ranges of {-1.0: -2.0} and {-1.0, -3.0}, respectively. The {-1.0: -2.5} gap range identifies the same threshold as (A). The substrate analog is colored white. Making the threshold more stringent fails to identify residues (the red phylogenetic motif) involved in two stabilizing enzyme-substrate interactions (a salt bridge and an H-bond).
Figure 6 Arginyl-tRNA synthetase phylogenetic motifs at the three PSZ thresholds are highlighted. Residues colored red indicate phylogenetic motifs identified using the (A) -2.04, (B) -2.93, and (C) -3.26 PSZ thresholds, which correspond to gap ranges of {-1.0: -2.0}, {-1.0, -2.5}, and {-1.0, -3.0}, respectively. The arginine substrate is colored yellow; the tRNA is colored blue; and the anticodon is color cyan. Using (A), two structurally unique clusters of phylogenetic motifs are identified. One corresponds to the enzyme active site, whereas the other corresponds to three H-bonds between the enzyme and the anti-codon arm of tRNA. (D) is the same (A), but with the rest of the enzyme removed for clarity.
Figure 7 Pie chart describing the relative accuracy of the automated phylogenetic motif (PM) method using the {-1.0: -2.0} gap range. Assessment is determined using the structure-based approach described in [25]. The relative accuracy of the automated PM method is encouraging as it does not require any a priori structural information.
Table 1 Manual verses automatically determined thresholdsa
Protein family PDB ID Manual PSZ Auto PSZ
Acetyglucosamine-6-phoshate deacyetylase 1O12 -2.20 -2.25
Alcohol dehydrogenase 1JVB -2.20 -2.15
CuZnSOD 1SPD -1.50 -1.50
Cytochrome P450 1N4G -2.20 -2.10
Enolase 2ONE -2.00 -2.18
Glyceroladhyde-3-phosphate dehydrogenase 1DC4 -1.80 -1.90
Glycerol kinase 1B05 -2.00 -2.03
Glutamate dehydrogenase 1HWZ -1.80 -2.12
Inorganic pyrophosphatase 1I6T -1.80 -1.67
Myoglobin 1MBA -1.50 -1.51
Succinate dehydrogenase – FAD 1NEK -2.00 -2.20
Succinate dehydrogenase – Fe/S 1NEK -2.20 -1.63
Succinate dehydrogenase – Heme 1NEK -2.20 -1.88
TATA box binding protein 1TBP -1.50 -2.16
Triosephosphate isomerase 7TIM -1.50 -1.85
a A sequence window width of 5 is used for each of the above examples and gap range of {-1: -2}.
Table 2 Correlation matrix of various phylogenetic motif parametersa
PSZ threshold # phylogenetic motifs # sequences
# phylogenetic motifs 0.15
# sequences 0.08 -0.07
MSA length -0.32 0.68 0.00
a A sequence window width of 5 is used for each of the above examples and gap range of {-1: -2}.
Table 3 Structural assessment of the phylogenetic motif functional site predictionsa
Protein family PDB ID Correct Useful info Wrong
A. Proteins from Table 1
Acetyglucosamine-6-phoshate deacyetylaseb 1O12 6 1 0
Alcohol dehydrogenase 1JVB 2 1 1
CuZnSOD 1SPD 3 2 1
Cytochrome P450 1N4G 6 4 0
Enolase 2ONE 6 2 0
Glyceroladhyde-3-phosphate dehydrogenase 1DC4 2 2 3
Glycerol kinase 1B05 7 1 0
Glutamate dehydrogenase 1HWZ 7 0 0
Inorganic pyrophosphatase 1I6T 3 2 0
Myoglobin 1MBA 4 1 0
Succinate dehydrogenase – FAD 1NEK 4 1 2
Succinate dehydrogenase – Fe/S 1NEK 4 0 0
Succinate dehydrogenase – Heme 1NEK 5 0 0
TATA box binding protein 1TBP n/ac
Triosephosphate isomerase 7TIM 3 4 1
B. Additional examples
Acetate kinase 1G99 4 4 0
Aconitase A 7ACN 8 3 5
Alanine racemase 1L6F 4 1 1
Arginyl-tRNA synthetase 1F7U 2 2 1
Biotin carboxylaseb 1DV2 3 3 2
Catalase 1GGF 9 0 0
Citrate synthase 1NXG 8 1 0
Isocitrate dehydrogenases 1IA2 3 0 0
Malate/lactate dehydrogenases 1EMD 6 1 0
Malate synthase 1P7T 6 0 0
Phosphomannomutase 1P5D 5 2 0
Selenocysteine lyase 1ECX 5 2 1
Threonine aldolase 1LW5 6 2 0
Thymidylate synthase 1AIQ 4 0 1
Transaldolase 1ONR 1 1 1
Trehalose-6-phosphate synthase 1UQU 5 3 3
a A sequence window width of 5 is used for each of the above examples and gap range of {-1: -2}.
b The known functional site defined in Acetyglucosamine-6-phoshate deacyetylase and biotin carboxylase is prohibitively incomplete; as such, functional sites indicated in [41] and [42], respectively, are also included. c Because so much of the TATA-box binding protein surface area directly interacts with its DNA substrate, the known functional site sphere encompasses nearly the entire protein, making the assessment unsporting because all predictions will trivially be correct.
==== Refs
Blom N Gammeltoft S Brunak S Sequence and structure-based prediction of eukaryotic protein phosphorylation sites J Mol Biol 1999 294 1351 1362 10600390 10.1006/jmbi.1999.3310
Oliveira L Paiva PB Paiva AC Vriend G Sequence analysis reveals how G protein-coupled receptors transduce the signal to the G protein Proteins 2003 52 553 560 12910455 10.1002/prot.10489
Puntervoll P Linding R Gemund C Chabanis-Davidson S Mattingsdal M Cameron S Martin DM Ausiello G Brannetti B Costantini A Ferre F Maselli V Via A Cesareni G Diella F Superti-Furga G Wyrwicz L Ramu C McGuigan C Gudavalli R Letunic I Bork P Rychlewski L Kuster B Helmer-Citterich M Hunter WN Aasland R Gibson TJ ELM server: A new resource for investigating short functional sites in modular eukaryotic proteins Nucleic Acids Res 2003 31 3625 3630 12824381 10.1093/nar/gkg545
Qu K Martin DL Lawrence CE Motifs and structural fold of the cofactor binding site of human glutamate decarboxylase Protein Sci 1998 7 1092 1105 9605314
Lu X Zhai C Gopalakrishnan V Buchanan BG Automatic annotation of protein motif function with Gene Ontology terms BMC Bioinformatics 2004 5 122 15345032 10.1186/1471-2105-5-122
del Sol MA Pazos F Valencia A Automatic methods for predicting functionally important residues J Mol Biol 2003 326 1289 1302 12589769 10.1016/S0022-2836(02)01451-1
Lichtarge O Bourne HR Cohen FE An evolutionary trace method defines binding surfaces common to protein families J Mol Biol 1996 257 342 358 8609628 10.1006/jmbi.1996.0167
Bickel PJ Kechris KJ Spector PC Wedemayer GJ Glazer AN Finding important sites in protein sequences Proc Natl Acad Sci U S A 2002 99 14764 14771 12417758 10.1073/pnas.222508899
Armon A Graur D Ben Tal N ConSurf: an algorithmic tool for the identification of functional regions in proteins by surface mapping of phylogenetic information J Mol Biol 2001 307 447 463 11243830 10.1006/jmbi.2000.4474
Madabushi S Yao H Marsh M Kristensen DM Philippi A Sowa ME Lichtarge O Structural clusters of evolutionary trace residues are statistically significant and common in proteins J Mol Biol 2002 316 139 154 11829509 10.1006/jmbi.2001.5327
Yao H Kristensen DM Mihalek I Sowa ME Shaw C Kimmel M Kavraki L Lichtarge O An accurate, sensitive, and scalable method to identify functional sites in protein structures J Mol Biol 2003 326 255 261 12547207 10.1016/S0022-2836(02)01336-0
La D Sutch B Livesay DR Predicting protein functional sites with phylogenetic motifs Proteins 2005 58 309 320 15573397 10.1002/prot.20321
Jones S Thornton JM Searching for functional sites in protein structures Curr Opin Chem Biol 2004 8 3 7 15036149 10.1016/j.cbpa.2003.11.001
Hulo N Sigrist CJ Le SV Langendijk-Genevaux PS Bordoli L Gattiker A De Castro E Bucher P Bairoch A Recent improvements to the PROSITE database Nucleic Acids Res 2004 32 D134 D137 14681377 10.1093/nar/gkh044
La D Silver M Edgar RC Livesay DR Using motif-based methods in multiple genome analyses: a case study comparing orthologous mesophilic and thermophilic proteins Biochemistry 2003 42 8988 8998 12885231 10.1021/bi027435e
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
Penny D Hendy M The use of tree comparison metrics Systematic Zoology 1985 34 75 82
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
Felsenstein J Inferring Phylogenies 2004 1 Sunderland, MA: Sinauer Associates, Inc
Kuhner MK Felsenstein J A simulation comparison of phylogeny algorithms under equal and unequal evolutionary rates Mol Biol Evol 1994 11 459 468 8015439
La D Livesay DR MINER: software for phylogenetic motif identification Nucleic Acids Res 2005
Kaufman L Rousseeuw PJ Finding Groups in Data: An Introduction to Cluster Analysis 1990 New York: Wiley
MacQueen JB Some Methods for classification and Analysis of Multivariate Observations 1967 Berkeley: University of California Press 281 297
Ihaka R Gentleman R R: a language for data analysis and graphics J Comp Grap Stat 1996 5 299 314
Aloy P Querol E Aviles FX Sternberg MJ Automated structure-based prediction of functional sites in proteins: applications to assessing the validity of inheriting protein function from homology in genome annotation and to protein docking J Mol Biol 2001 311 395 408 11478868 10.1006/jmbi.2001.4870
Wallace AC Laskowski RA Thornton JM LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions Protein Eng 1995 8 127 134 7630882
Kursula I Partanen S Lambeir AM Antonov DM Augustyns K Wierenga RK Structural determinants for ligand binding and catalysis of triosephosphate isomerase Eur J Biochem 2001 268 5189 5196 11589711 10.1046/j.0014-2956.2001.02452.x
Babbitt PC Mrachko GT Hasson MS Huisman GW Kolter R Ringe D Petsko GA Kenyon GL Gerlt JA A functionally diverse enzyme superfamily that abstracts the alpha protons of carboxylic acids Science 1995 267 1159 1161 7855594
Cooperman BS Baykov AA Lahti R Evolutionary conservation of the active site of soluble inorganic pyrophosphatase Trends Biochem Sci 1992 17 262 266 1323891 10.1016/0968-0004(92)90406-Y
Livesay DR Jambeck P Rojnuckarin A Subramaniam S Conservation of electrostatic properties within enzyme families and superfamilies Biochemistry 2003 42 3464 3473 12653550 10.1021/bi026918f
DeDecker BS O'Brien R Fleming PJ Geiger JH Jackson SP Sigler PB The crystal structure of a hyperthermophilic archaeal TATA-box binding protein J Mol Biol 1996 264 1072 1084 9000631 10.1006/jmbi.1996.0697
Joseph D Petsko GA Karplus M Anatomy of a conformational change: hinged "lid" motion of the triosephosphate isomerase loop Science 1990 249 1425 1428 2402636
Rozovsky S Jogl G Tong L McDermott AE Solution-state NMR investigations of triosephosphate isomerase active site loop motion: ligand release in relation to active site loop dynamics J Mol Biol 2001 310 271 280 11419952 10.1006/jmbi.2001.4673
Rozovsky S McDermott AE The time scale of the catalytic loop motion in triosephosphate isomerase J Mol Biol 2001 310 259 270 11419951 10.1006/jmbi.2001.4672
Wierenga RK Borchert TV Noble ME Crystallographic binding studies with triosephosphate isomerases: conformational changes induced by substrate and substrate-analogues FEBS Lett 1992 307 34 39 1639191 10.1016/0014-5793(92)80897-P
Livesay DR La D The evolutionary origins and catalytic importance of conserved electrostatic networks within TIM-barrel proteins Protein Sci 2005 14 1158 1170 15840824 10.1110/ps.041221105
Delagoutte B Moras D Cavarelli J tRNA aminoacylation by arginyl-tRNA synthetase: induced conformations during substrates binding EMBO J 2000 19 5599 5610 11060012 10.1093/emboj/19.21.5599
Whisstock JC Lesk AM Prediction of protein function from protein sequence and structure Q Rev Biophys 2003 36 307 340 15029827 10.1017/S0033583503003901
Kursula I Wierenga RK Crystal structure of triosephosphate isomerase complexed with 2-phosphoglycolate at 0.83-A resolution J Biol Chem 2003 278 9544 9551 12522213 10.1074/jbc.M211389200
Tellez-Valencia A Olivares-Illana V Hernandez-Santoyo A Perez-Montfort R Costas M Rodriguez-Romero A Lopez-Calahorra F Tuena DG-P Gomez-Puyou A Inactivation of triosephosphate isomerase from Trypanosoma cruzi by an agent that perturbs its dimer interface J Mol Biol 2004 341 1355 1365 15321726 10.1016/j.jmb.2004.06.056
Schorken U Thorell S Schurmann M Jia J Sprenger GA Schneider G Identification of catalytically important residues in the active site of Escherichia coli transaldolase Eur J Biochem 2001 268 2408 2415 11298760 10.1046/j.1432-1327.2001.02128.x
Blanchard CZ Lee YM Frantom PA Waldrop GL Mutations at four active site residues of biotin carboxylase abolish substrate-induced synergism by biotin Biochemistry 1999 38 3393 3400 10079084 10.1021/bi982660a
| 15890082 | PMC1142304 | CC BY | 2021-01-04 16:02:48 | no | BMC Bioinformatics. 2005 May 13; 6:116 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-116 | oa_comm |
==== Front
BMC BiotechnolBMC Biotechnology1472-6750BioMed Central London 1472-6750-5-111588514010.1186/1472-6750-5-11Methodology ArticleLiposome retention in size exclusion chromatography Ruysschaert Tristan [email protected] Audrey [email protected] Jean-Luc [email protected] Sylviane [email protected] Mathias [email protected] Didier [email protected] Groupe de biotechnologie des protéines, Institut de Pharmacologie et de Biologie Structurale, F-31077, Toulouse, France2 Electron Microscopy Department, Rangueil Hospital Medical School, University of Toulouse, 31062 Toulouse, France3 Groupe Physico-chimie des systèmes polyphasés, Université Paris-sud, F-92296, Châtenay-Malabry, France4 International University Bremen, D-28726 Bremen, Germany2005 10 5 2005 5 11 11 24 11 2004 10 5 2005 Copyright © 2005 Ruysschaert 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
Size exclusion chromatography is the method of choice for separating free from liposome-encapsulated molecules. However, if the column is not presaturated with lipids this type of chromatography causes a significant loss of lipid material. To date, the mechanism of lipid retention is poorly understood. It has been speculated that lipid binds to the column material or the entire liposome is entrapped inside the void.
Results
Here we show that intact liposomes and their contents are retained in the exclusion gel. Retention depends on the pore size, the smaller the pores, the higher the retention. Retained liposomes are not tightly fixed to the beads and are slowly released from the gels upon direct or inverted eluent flow, long washing steps or column repacking. Further addition of free liposomes leads to the elution of part of the gel-trapped liposomes, showing that the retention is transitory. Trapping reversibility should be related to a mechanism of partitioning of the liposomes between the stationary phase, water-swelled polymeric gel, and the mobile aqueous phase.
Conclusion
Retention of liposomes by size exclusion gels is a dynamic and reversible process, which should be accounted for to control lipid loss and sample contamination during chromatography.
==== Body
Background
Liposomes are self-assembled phospholipids enclosing a droplet of the aqueous medium in which they are formed [1]. Liposomes have numerous applications namely as in vivo drug delivery vehicles [2]. Drugs interact with liposomes in several different ways depending on their solubility and polarity characteristics. They can be inserted in the lipid chain bilayer region, intercalated in the polar head group region, adsorbed on the membrane surface, anchored by a hydrophobic tail or entrapped in the inner aqueous compartment. A prerequisite for the use of drug-loaded liposomes is to be able to separate encapsulated from free drugs. Recently, we encapsulated enzymes in liposomes to enhance the enzyme stability with respect to dilution and proteases [3,4]. Size exclusion chromatography (SEC) is an old and widely used tool to separate small solutes from liposomes or to narrow the size distribution [5,6]. For example, Sepharose 2B, 4B Sephacryl S-1000 and high performance exclusion gels of the TSK-PW series are suitable for separating small unilamellar vesicles from larger ones [6-10]. In 1977, Sorensen et al. [11] showed loss of lipid material during SEC of liposomes. In 1980, Van Renswoude et al. [12] observed by microscopy the retention of fluorescent liposomes between and on the polymer beads. The loss of lipids could reach 20 to 40% of the deposited material [8,13]. These findings suggested a step of gel pre-saturation with lipids to avoid loss during SEC [14]. To achieve high quality separation, the column pre-treatment is preferentially carried out with sonicated liposomes as their small sizes ensure efficient penetration of the lipids within the gel pores [15].
Here we show that when liposomes containing a hydrophilic protein pass through a SEC column, lipids and protein are simultaneously retained suggesting that the liposomes retained remain whole. Retention is, however, transitory and reversible: there is an equilibrium between retained liposomes and eluting ones. A pre-saturation step should thus be performed with liposomes of the same composition as that of the liposomes to be chromatographied in accordance with previous recommendations [10].
Results
SEC elution pattern of liposomes
In order to quantify the amount of lipid retained on a freshly prepared column (not previously saturated with lipids) we injected 400 μL liposomes incorporating 1 mol % Rhod-PE as fluorescent label and loaded with 5 to 10 acetylcholinesterase molecules as internal material. The sample was passed through an 8 mL G25 column. The majority of lipid and enzyme was excluded from the pores of the gel beads and eluted together at the void volume of the column (from 6 to10 mL, Fig. 1). To determine the amount of lipid and enzyme retained in the gel, TX-100 was added to the elution buffer such that the final eluent concentration was 0.5% (w/v). This caused co-elution of a significant amount of lipid and enzyme. Control measurements performed either with free enzyme, or with a mixture of free enzyme and empty liposomes, did not show any retention of enzyme under similar conditions. A further experiment was performed with calcein-loaded liposomes. Identical results were obtained: TX-100 treatment led to the concomitant elution of lipids and calcein under a broader peak than in the case of enzyme-loaded liposomes, due to the smaller molecular mass of calcein (see below). Retention of lipids and of hydrophilic encapsulated materials were identical, suggesting that intact liposomes were retained during SEC.
Column saturation
We observed lipid retention when the chromatography gel was incubated with liposomes prior to being poured into the column. For convenience, we used this method to characterize the relationship between the amount of lipid added to the gel and the amount retained. 2 mL of swollen G25 were incubated in 5 mL buffer containing various concentrations of Rhod-PE labeled liposomes to evaluate the parameters underlying liposome retention. Apparently lipid retention onto the gel beads shows a saturation limit (Fig. 2). It appears that saturation is obtained by passing at least 5 μmole of lipids per ml gel.
Relationship between retention of lipids and encapsulated enzyme
Liposomes loaded with AChE were incubated with 5 mL of fresh G25 in 9 mL buffer. Different concentrations of liposome were used to obtain different retention efficiencies. After 2 h, the gel was poured into a column, washed and eluted with TX-100. The lipid and enzyme content of the fractions were quantified. We observed a linear correlation between lipid and protein retention (Fig. 3). In a control experiment, when liposomes and free enzyme were loaded together on the column, no enzyme retention was observed. This suggests that retention of membrane and intravesicular content are linked, consistent with the hypothesis of non-damaged liposome retention. Furthermore, we observed greater retention of AChE than of lipids suggesting liposome reorganization during the washing step.
Retention depends on gel exclusion limits
Labeled liposomes were incubated with different gels and retention was estimated by elution with TX-100. Retention appeared to depend on the pore size of the beads responsible for the size exclusion (Fig. 4). The smaller the pore sizes, the greater the amount of liposomes retained. The exclusion limit of the gels in our study was significantly smaller than the liposomes size. Exception was noted for Sepharose 4B, this gel reached the range of 60 nm diameters exclusion limit, which is enough to allow small liposomes to penetrate the pores of the SEC gel and to obey an effective permeation process. This chromatography may explain the apparent higher retention than expected. On the other hand, retention was independent of the size of the gel beads since G25 beads of different sizes, fine (20–80 μm diameter), medium (50–150 μm) and coarse (100–300 μm) exhibited the same retention efficiency (data not shown).
Elution by extensive washing step, flow inversion or column repacking
In order to distinguish binding from kinetic trapping, we modified the washing volume. 400 μL of liposome solution were loaded onto 8 mL G25 gels and washed with different amounts of buffer. The remaining retained liposomes were quantified with TX-100 elution. Fig. 5 shows that retention decreased as the washing step was increased.
To test if liposomes were trapped between the beads due to column packing, elution was applied backwards. First 400 μL of Rhod-PE labeled liposomes were loaded on 8 mL G25 and washed. After exclusion of non-retained liposomes, the buffer flow was inverted. The elution profile (Fig. 6) showed that some lipids were eluted by the flow inversion. However, only a small proportion of the retained liposomes were eluted as evidence by the amount of lipids eluted by adding TX-100. Identically, depacking a gel and pouring it into a column once more resulted to partial elution of the retained liposomes.
Direct observation by scanning electron microscopy (SEM)
SEM photographs of sephadex beads were performed to observe how lipids are retained in SEC columns and to investigate how lipids bind to the beads, as individual molecules, as membranes or as liposomes at the surface or inside the beads. Liposomes containing 40% of PE and extruded at 200 nm were passed onto a Sephadex G-25 fine column. After intensive washing, samples diluted in Sorensen buffer were fixed with 2% glutaraldehyde. Without liposomes, the beads appeared homogeneous (Fig 7 A). In contrast, beads originating from the column were covered by lipid aggregates resembling liposomes. Thus, it seems that entire liposomes are retained on the beads. Some liposome aggregation appeared on the beads which may result from the glutaraldehyde cross-linking. Liposomes appeared to be larger than expected with a 500 nm diameter compared to the 250 nm estimated by dynamic light scattering before loading. This difference suggests that liposomes fused either during the chromatography process or during fixation. It also appeared that some beads retained few liposomes (Fig 7 B) while others were completely covered (Fig 7 C). This heterogeneity might result from heterogeneity of column saturation from the top to the bottom.
Elution of retained lipids by liposomes in the mobile phase
We tested if liposomes could be dissociated from the column by other liposomes. Liposomes labeled with fluorescent lipids and containing AChE, were loaded on a G25 column. Following exclusion of non-retained liposomes, unlabeled liposomes were loaded on the column. It appeared that some retained lipids and enzymes were co-eluted by the unlabeled liposomes (Fig. 8). Enzymes eluted by liposome were encapsulated because their activity was detectable only by using TX-100 in the solution. TX-100 disrupted liposome bilayers and allows the enzyme substrate to reach AChE which is unable to cross the lipid membrane [16]. The same experiment was performed with an intravesicular tracer of smaller size: calcein. This fluorescent probe was loaded in rhod-PE labeled liposomes. The elution profile (Fig. 9) shows one peak for excluded liposomes. Passing unlabeled liposomes resulted in an elution of both retained labeled lipids and calcein.
In order to test lipid exchanges between liposomes in the column, the previous experiment was repeated replacing the unlabeled passing liposomes by NBD-PE labeled liposome on Rhod-PE ones retained in the gel. As previously, passage of NBD-PE labeled liposomes, led to the elution of Rhod-PE liposomes. In addition, we observed elution of NBD-PE liposomes by adding TX-100. Thus, it appears that binding and release of liposomes on Sephadex columns is a dynamic process.
Discussion
What is retained, lipids or liposomes?
Lipid retention has been observed since the beginning of the 80's [12] and usually pre-saturation steps are used to avoid this phenomenon. However, a question remained: are liposomes or lipids retained? To address this, we compared the retention of lipids and the retention of encapsulated materials, enzyme or calcein. It appeared that intravesicular material is also retained suggesting that lipids are retained as intact liposomes without leakage (Fig. 1, 3 and 9). Furthermore SEM observations of the beads clearly showed the presence of lipid aggregates resembling liposomes, attached to the beads (Fig. 7). The next question should be: why are liposomes retained in SEC gels, whereas their size and shape would predict their exclusion? A first hypothesis is a direct interaction of lipids with the beads at the liquid/solid interface of water and the poly-dextran beads, this interaction may promote lipid reorganization as in liquid/air monolayer experiments. As full liposomes are retained in columns, this hypothesis seems less probable because such an interaction would destroy the liposome integrity and its contents would be released. A second hypothesis arises from liposomes' flexibility: they may be deformed and thereby pass through the polymer net and remain stuck [17]. Several experiments favor this hypothesis. i) Retention depends on the bead pore size (Fig. 4) ii) Partial elution can be performed by inversion of the buffer flow (Fig. 6), depacking and repacking the gel in the column or extensive washing (Fig 5).
Is it possible to saturate a column to avoid liposome retention?
Figure 2 shows a saturation plateau at 60 nmoles lipids per mL of G25 suggesting that once the column has been saturated, SEC can be performed without liposome retention. However, if all the pores are filled with liposomes, like in the picture in fig 7C, the efficiency of chromatography may be affected. Furthermore, saturation depends on the washing step, the longer it is, the less the column remains saturated. Saturation is not static but seems to be the result of a dynamic process resulting from exchanges between free and entrapped liposomes. In conclusion, saturation of a column should be used with caution bearing in mind the reversibility of the saturation process, which can cause contamination of the samples to be analyzed.
Interaction of liposome with beads and leakage
Two observations suggest that retention of liposomes inside the beads is associated with liposome reorganization. First, retention of enzyme appeared to be higher than lipid retention suggesting that entrapped liposomes lose their lipids (Fig. 3) and second, free liposomes passing through the column are able to elute fixed liposomes (Fig. 8 and 9), most probably by fusion and fission. Is there some leakage of encapsulated molecules during SEC? Andersson and Lundahl [18] hypothesized leakage of glucose from the liposome due to interaction between the liposomes and the gel bead surfaces, and disruption of liposomes by shear forces from the liquid flow. However, simple permeation of glucose across the lipid bilayer is significant and does not allow permeation to be distinguished from leakage during the several minutes needed to perform SEC. However, our enzymatic test did not provide evidence of any liposome leakage or internal content release with time. This strongly supports that the transitory retention of lipids does not induce dramatic vesicle leakage or bilayer permeability enhancement and rather agrees with intact liposome trapping.
Relation with immobilized-liposome chromatography
Immobilized-liposome chromatography is used for studying the partitioning of compounds into phospholipid bilayers. Several methods are used to bind liposomes to the matrix. Specific ligands such as hydrophobic ligands [19] can be coupled to the gel beads. Liposomes can be formed in the presence of beads using the detergent dialysis method. In this method, lipids, detergent and beads are mixed and dialyzed. Elimination of detergent molecules during the dialysis results in the formation of liposomes entrapped in the beads [19]. A third method comes close to liposome chromatography: beads are mixed with liposomes and binding is achieved by several cycles of freezing and thawing [20]. Entrapped liposomes are thought to be suspended in micro cavities of gel beads [20]. Our results suggest that part of the liposomes spontaneously rearrange in gel beads without outer stress. Buffer flow would be sufficient to push the liposomes inside the bead net and to immobilize them.
Conclusion
How should SEC be performed with liposomes? As retention is inversely related to the bead exclusion limit, the larger-pore gels should be preferred for liposome separation from small molecules as suggested by Grabielle-Madelmont et al [15]. As retention is proportional to the gel volume, the bed volume of the column has to be calculated at minima for optimal separation, which also avoids sample dilution. The pre-saturation step diminishes liposome retention but as there are liposome exchanges, the purity of the sample may be affected if the retained liposomes are different from the free liposomes. Two ways are possible: (1) for analytic purposes, it is thus better to inject the liposomes to be sized several times until gel saturation and then perform size analysis (2) for preparative purposes like separation of loaded liposomes from non-encapsulated material, the problem is trickier and the experimental conditions should be adapted to minimize loss of lipid and contamination.
Methods
Materials
The lipids, egg-phosphatidylethanolamine (PE), brain Lα-phophatidylserine (PS), Lα-phosphatidylethanolamine-N-(lissamine-rhodamine-B-sulfonyl) (Rhod-PE), Lα-phosphatidylethanolamine N-(4-nitrobenzo-2-oxa-1,3-diazole) (NBD-PE), were purchased from Avanti Polar Lipids, Birmingham, AL, USA. Free rhodamine or free NBD which may contaminate labelled lipids were eliminated from liposome solution by chromatogaphy or dialysis. Egg-phosphatidylcholine (PC), was from Lipoid, Ludwigshafen, Germany. Calcein (high purity) was from molecular probes, Leiden, the Netherlands. Sephadex G-25 Medium, G-75 Medium and G-100 Medium, Sephacryl S-100 HR and Sepharose 4B were from Amersham Biosciences, Biogel P2 (45–90 μm) was from Bio-Rad.
Unless otherwise stated, all experiments were performed using 145 mM NaCl, 2.5 mM HEPES, pH 7.4 buffer as aqueous phase. Chromatography gels were washed before use by 0.5% Triton X-100 (TX-100) (Merck), water and equilibrated with buffer.
Liposome preparation
1 μmole egg-PC dissolved in CHCl3 was placed in a 10 mL glass tube. Then this was dried under a stream of N2 and under vacuum for three hours to form a dry lipid film. 200 μL of buffer were added to the lipid film and vortexed to peel off the lipid. The liposome suspension obtained was frozen in liquid N2 and thawed in a 25°C water-bath 10 times. Buffer was then added to obtain a 1 mM lipid dispersion, which was passed 10 times through a 0.2 μm polycarbonate filter (Schleicher & Schuell, Dassel, Germany). Labeled liposomes were prepared according to the same procedure except that the initial lipid film was obtained from a mixed chloroform solution of egg-PC and 1 mol% Rhod-PE or NBD-PE. Liposome size was estimated to be 230 nm (diameter) by dynamic light scattering.
Fluorescence
To follow lipid elution in SEC, we measured the lipid content in each fraction by two methods: a phosphate assay to measure all lipids and a fluorescent assay to measure labelled lipids. As we found a good correlation between the two methods, we used only the second one allowing studies with lower amount of lipids to analyze retention of liposomes. Rhod-PE (excitation wavelength 550/emission wavelength 590 nm) and NBD-PE (excitation wavelength 470/ emission wavelength 530 nm) were used as membrane fluorescent markers. Calcein was used at 500 μM at an excitation wavelength of 492 nm and emission wavelength of 517 nm. Fluorescence was measured using a lamp LPS220 and a photomultiplier detection system PDS 810 from photon Technology International.
Enzyme encapsulation
To obtain liposomes loaded with 5–10 enzyme molecules, 1 μmole of dried eggPC was dissolved with 200 μL of enzyme solution containing 4 nanomoles of Drosophila melanogaster acetylcholinesterase (AChE) in buffer. The liposome dispersion obtained was frozen in liquid N2 and thawed in a 25°C water-bath 25 times to allow the entrance of the protein [3,4]. Non-encapsulated AChE was removed by inverse affinity chromatography, passing the liposome dispersion through a column filled with procainamide gel that retains free AChE. Elution contained at least 90% encapsulated enzyme and less than 10% free enzyme. Liposome containing fractions were pooled and lipid concentration was adjusted to 1 mM. Extrusion was performed 10 times through a 0.2 μm filter. AChE activity was measured with the sensitive method of Ellman [21] with 1 mM acetylthiocholine after dissolution of liposomes with 0.1% (w/v) TX-100.
Calcein encapsulation
To obtain liposomes loaded with calcein, 1 μmole of dried eggPC was dissolved with 1 mL of buffer containing 500 μM of calcein. The liposome dispersion obtained was frozen in liquid N2 and thawed in a 25°C water-bath 25 times to allow the entrance of the calcein and extrusion was performed 10 times through a 0.2 μm filter. Non-encapsulated probe was removed by passing the liposomes through a Sephadex G-100 column. The fluorescence of the fractions was measured in the presence of TX-100 to avoid calcein self-quenching.
SEC preparation
1 cm diameter columns were used with two different methods. In the first one, the gel was freshly packed and the liposomes were deposited when the eluent level reached the top of the gel. Buffer flow was driven by simple gravity. The sample was allowed to penetrate the gel entirely before adding repetitive small buffer volume to proceed to the elution. In the second one, swollen gel was incubated in a buffer containing liposomes with gentle agitation for two hours. The gel was then packed into the column and non retained liposomes were eluted by successive additions of small amounts of buffer.
Sample preparation for scanning electron microscopy (SEM)
For SEM analysis, liposomes containing 60% PC and 40% PE were prepared as previously described. PE was chosen to allow cross-linking by glutaraldehyde. After passing and washing the liposomes on G-25 fine beads were fixed with 2% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) (Sorensen buffer) containing 6 mM CaCl2, for 4 h, at 4°C. After an extensive wash in the same buffer, samples were removed, post-fixed for 1 h at room temperature with 1% osmium tetroxide in 0.1 M cacodylate buffer (pH 7.4), dehydrated in a graded ethanol series, dried by critical point drier with an EMSCOPE CPD 750 and coated with gold-palladium for 3 minutes at 100 Ångströms / minute, and observed with a S-450 scanning electron microscope (Hitachi) at an accelerating voltage of 10 kV.
List of abbreviations
AChE: Acetylcholinesterase
NBD-PE: Lα-phosphatidylethanolamine N-(4-nitrobenzo-2-oxa-1,3-diazole)
PC: Egg-phosphatidylcholine
PE: Phosphatidylethanolamine
Rhod-PE: Lα-phosphatidylethanolamine-N-(lissamine-rhodamine-B-sulfonyl)
SEC: Size Exclusion Chromatography
SEM : Scanning Electron Microscopy
TX-100: Triton X-100
Authors' contributions
TR and AM carried out the SEC studies. JLD carried out the preparation and the observation of samples by SEM. SL and MW participated in the design of the study and in the data analysis. DF conceived coordinated the study.. All authors read and approved the final manuscript.
Acknowledgements
We acknowledge financial support through RTN Network Nanocapsules with functionalised surfaces and walls (HPRN CT-2000 00159) and Lipoids for the gift of egg-PC lipids.
Figures and Tables
Figure 1 Elution pattern of lipids and encapsulated enzymes following liposome chromatography on G25 column. Liposomes labeled with rhod-PE and containing AChE were chromatographied on G25M. Fractions of 1.5 mL were collected. Fluorescence (in red) and enzyme activity (in blue) were measured as a percentage of the total fluorescence or enzyme activity collected. Arrows indicate the injection of TX-100 on the top of the columns. The general elution profile is represented in the inset.
Figure 2 Liposome retention depends on the amount of lipids loaded. Liposomes labeled with Rhod-PE were incubated with 2 mL G25M gel two hours with gentle agitation and poured in columns. Non-retained liposomes collected by washing the column with 40 mL buffer and retained liposomes were eluted by adding 0.5% TX-100.
Figure 3 Relation between lipid and intravesicular marker retention. Liposomes labeled with rhod-PE and containing AChE were chromatographied on G25M. Fluorescence and activity were collected before and after addition of TX-100.
Figure 4 Liposome retention depends on bead pore size. Liposomes labeled with Rhod-PE were incubated with different gels and poured into columns. Fluorescence was collected before and after TX-100 addition to estimate the proportion of liposomes retained.
Figure 5 Elution of retained liposomes by the buffer flow. Liposomes labeled with rhod-PE were chromatographied on 8 mL G25M. TX-100 was added after different extensive washing volumes to estimate the slow elution due to shearing forces of the buffer flow.
Figure 6 Flow inversion eluted a part of the retained liposomes. Liposomes were chromatographied on G25M. Fractions of 1.5 mL were collected. First arrows indicate the inversion of buffer flow and the second one the injection of 0.5% TX-100 on the top of the column.
Figure 7 SEM observation of retained liposomes on G-25 fine beads. A : Sephadex beads non incubated with liposomes, fixed with glutaraldehyde, washed, stained and prepared for SEM. B, C, D: Liposomes, containing PC and PE (6/4), retained by a 2 mL sephadex G-25 fine columns were fixed by glutaraldehyde, washed, stained and prepared for SEM observations.
Figure 8 Elution of retained labeled liposomes with Rhod-PE and AChE by passing non-labeled liposomes. Liposomes, labeled with rhod-PE (fluorescence represented with red triangles, ) and containing AChE (activity represented with blue squares, ) were chromatographied on G25M. Fractions of 1.5 mL were collected. The first arrow indicates the injection of unlabeled liposomes on the column and the second arrow indicates the injection of 0.5% TX-100 on the top of the columns.
Figure 9 Elution of retained labeled liposomes with Rhod-PE and calcein by passing non-labeled liposomes. Liposomes, labeled with rhod-PE (fluorescence represented with red triangles, ) and containing calcein, 500 μM (fluorescence represented with blue circles, ) were chromatographied on G25M. Fractions of 1.5 mL were collected. The first arrow indicates the injection of unlabeled liposomes on the column and the second arrow indicates the injection of 0.5% TX-100 on the top of the columns.
==== Refs
Bangham AD Standish MM Watkins JC Diffusion of univalent ions across the lamellae of swollen phospholipids J Mol Biol 1965 13 238 252 5859039
Lasic DD Vallner JJ Working PK Sterically stabilized liposomes in cancer therapy and gene delivery Curr Opin Mol Ther 1999 1 177 185 11715941
Colletier JP Chaize B Winterhalter M Fournier D Protein encapsulation in liposomes: efficiency depends on interactions between protein and phospholipid bilayer BMC Biotechnol 2002 2 9 12003642 10.1186/1472-6750-2-9
Chaize B Winterhalter M Fournier D Encapsulation of acetylcholinesterase in preformed liposomes Biotechniques 2003 34 1158 60, 1162 12813881
Finkelstein MC Weissmann G Enzyme replacement via liposomes. Variations in lipid compositions determine liposomal integrity in biological fluids Biochim Biophys Acta 1979 587 202 216 486550
Huang C Studies on phosphatidylcholine vesicles. Formation and physical characteristics Biochemistry 1969 8 344 352 5777332 10.1021/bi00829a048
Lesieur S Grabielle-Madelmont C Paternostre MT Ollivon M Size analysis and stability study of lipid vesicles by high-performance gel exclusion chromatography, turbidity, and dynamic light scattering Anal Biochem 1991 192 334 343 2035833 10.1016/0003-2697(91)90545-5
Barenholz Y Gibbes D Litman BJ Goll J Thompson TE Carlson RD A simple method for the preparation of homogeneous phospholipid vesicles Biochemistry 1977 16 2806 2810 889789 10.1021/bi00631a035
Nozaki Y Lasic DD Tanford JA Size analysis of phospholipid vesicle preparations Science 1982 217 366 367 7089571
Lesieur S Grabielle-Madelmont C Paternostre MT Ollivon M Study of size distribution and stability of liposomes by high performance gel exclusion chromatography Chem Phys Lipids 1993 64 57 82 10.1016/0009-3084(93)90058-B
Sorensen EN Weisman G Vidaver GA A sephadex column procedure for measuring uptake and loss of low molecular weight solutes from small, lipid-rich vesicles Anal Biochem 1977 82 376 384 562078 10.1016/0003-2697(77)90175-0
Van Renswoude AJ Blumenthal R Weinstein JN Thin-layer chromatography with agarose gels. A quick, simple method for evaluating liposome size Biochim Biophys Acta 1980 595 151 156 7349878
Schurtenberger PHH Characterization of the size distribution of unilamellar vesicles by gel filtration, quasi-elastic light scattering and electron microscopy. Biochim Biophys Acta 1984 778 470 480
Reynolds JA Nozaki Y Tanford C Gel-exclusion chromatography on S1000 Sephacryl: application to phospholipid vesicles Anal Biochem 1983 130 471 474 6869835 10.1016/0003-2697(83)90618-8
Grabielle-Madelmont C Lesieur S Ollivon M Characterization of loaded liposomes by size exclusion chromatography J Biochem Biophys Methods 2003 56 189 217 12834977 10.1016/S0165-022X(03)00059-9
Chaize B Fournier D Sorting out molecules reacting with acetylcholinesterase by enzyme encapsulation in liposome. Biosens Bioelect 2004 in press
Lundahl P Zeng CM Lagerquist Hagglund C Gottschalk I Greijer E Chromatographic approaches to liposomes, proteoliposomes and biomembrane vesicles J Chromatogr B Biomed Sci Appl 1999 722 103 120 10068136 10.1016/S0378-4347(98)00370-3
Andersson L Lundahl P Substantial glucose leakage from liposomes on filters and upon molecular-sieve chromatography in determinations of reconstituted glucose-transport activity and liposome volumes Biochim Biophys Acta 1990 1030 258 268 2261488
Lundahl P Yang Q Liposome chromatography: liposomes immobilized in gel beads as a stationary phase for aqueous column chromatography J Chromatogr 1991 544 283 304 1885691 10.1016/S0021-9673(01)83992-3
Lagerquist C Beigi F Karlen A Lennernas H Lundahl P Effects of cholesterol and model transmembrane proteins on drug partitioning into lipid bilayers as analysed by immobilized-liposome chromatography J Pharm Pharmacol 2001 53 1477 1487 11732750 10.1211/0022357011778016
Ellman GL Courtney KD Andres VJ Feather-Stone RM A new and rapid colorimetric determination of acetylcholinesterase activity Biochem Pharmacol 1961 7 88 95 13726518 10.1016/0006-2952(61)90145-9
| 15885140 | PMC1142305 | CC BY | 2021-01-04 16:02:57 | no | BMC Biotechnol. 2005 May 10; 5:11 | utf-8 | BMC Biotechnol | 2,005 | 10.1186/1472-6750-5-11 | oa_comm |
==== Front
BMC BiotechnolBMC Biotechnology1472-6750BioMed Central London 1472-6750-5-121588514410.1186/1472-6750-5-12Methodology ArticleA simple technique for quantifying apoptosis in 96-well plates Ribble Deborah [email protected] Nathaniel B [email protected] David A [email protected] Yiqun G [email protected] UCHSC at Fitzsimons, Dermatology Department, Mail Stop #8127, PO Box 6511, Aurora, CO 80045, USA2005 10 5 2005 5 12 12 24 1 2005 10 5 2005 Copyright © 2005 Ribble 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
Analyzing apoptosis has been an integral component of many biological studies. However, currently available methods for quantifying apoptosis have various limitations including multiple, sometimes cell-damaging steps, the inability to quantify live, necrotic and apoptotic cells at the same time, and non-specific detection (i.e. "false positive"). To overcome the shortcomings of current methods that quantify apoptosis in vitro and to take advantage of the 96-well plate format, we present here a modified ethidium bromide and acridine orange (EB/AO) staining assay, which may be performed entirely in a 96-well plate. Our method combines the advantages of the 96-well format and the conventional EB/AO method for apoptotic quantification.
Results
We compared our method and the conventional EB/AO method for quantifying apoptosis of suspension cells (Jurkat) and adherent cells (A375) under normal growth and apoptosis-inducing conditions. We found that our new EB/AO method achieved quantification results comparable to those produced using the conventional EB/AO method for both suspension and adherent cells.
Conclusion
By eliminating the detaching and washing steps, our method drastically reduces the time needed to perform the test, minimizes damage to adherent cells, and decreases the possibility of losing floating cells. Overall, our method is an improvement over the currently available techniques especially for adherent cells.
==== Body
Background
Apoptosis, a type of programmed cell death, is an active process. It is a normal component of the development and health of multicellular organisms. The study of apoptosis is an important field of biological inquiry since a deficiency or an excess of apoptosis is one of the causes for cancers, autoimmune disorders, diabetes, Alzheimer's, organ and bone marrow transplant rejection, and many other diseases. Accordingly, a quick and easy assay for quantification of apoptosis would be very useful for many biological researchers.
Currently, methods available to help detect apoptosis in vitro include several morphological staining methods (such as ethidium bromide and acridine orange (EB/AO) [1,2], DAPI (4; 6-diamidino-2phenylidole) [2], Hoechst staining [2], and etc), Annexin V staining [3-6], DNA ladder [7,8], TUNEL (Terminal deoxynucleotidyl transferase mediated dUTP Nick End Labeling) [9-11], Caspase-3/7 activity [12-16], and ssDNA staining [17-22]. However, these methods have at least one of the following limitations:
1) Involvement of multiple steps
All current staining methods, Annexin V, and DNA laddering assays require detaching, washing and transferring the cells. These procedures might damage the cell membranes and change the cell population distribution of live, apoptotic and/or necrotic cells. Procedures with multiple steps also require more time to perform the assay, and more materials allowing for loss of the cells through the procedures.
2) Lack of the ability to quantify live, apoptotic and necrotic cells at the same time
DAPI staining, caspase-3/7 activity, DNA laddering, and ssDNA staining methods only detect increase of apoptotic signals, and can not easily quantify percentage of live, apoptotic, and necrotic cells.
3) Non-specific detection
The TUNEL assay is widely used for detecting apoptotic cells. However, it has been shown to provide false positive signals in some necrotic cells [18,21,22].
Despite the many characteristics of apoptotic cells analyzed by current methods, chromatin condensation and nuclear fragmentation remain the hallmarks of apoptotic cells [1,23-25]. It has been suggested that as a rule, classification of cell death in a given model should always include morphological examination coupled with at least one other assay [26]. Fluorescence light microscopy with differential uptake of fluorescent DNA binding dyes (such as EB/AO staining) is a method of choice for its simplicity, rapidity, and accuracy. In such an assay, apoptotic index and cell membrane integrity can be assessed simultaneously and there is no cell fixation step, thus avoiding a number of potential artifacts [26]. Acridine orange (AO) permeates all cells and makes the nuclei appear green. Ethidium bromide (EB) is only taken up by cells when cytoplasmic membrane integrity is lost, and stains the nucleus red. EB also dominates over AO. Thus live cells have a normal green nucleus; early apoptotic cells have bright green nucleus with condensed or fragmented chromatin; late apoptotic cells display condensed and fragmented orange chromatin; cells that have died from direct necrosis have a structurally normal orange nucleus [26].
A 96-well plate format is ideal for examining multiple cell types and performing multiple assays while using very small amounts of materials. It is also a suitable format for high throughput drug screening. To overcome the shortcomings of current methods that quantify apoptosis in vitro, and to take advantage of the 96-well plate format, we present here an improved EB/AO staining assay performed entirely in a 96-well plate.
Results and Discussion
We modified the conventional EB/AO method by centrifuging the cells in a 96-well plate to bring down all the cells, including floaters, to the bottom of the plate. Using this method, we obtained quantification of live, necrotic, and apoptotic cells for both suspension cells (Jurkat, Figure 1), and adherent cells (A375, Figure 2), under normal growth and apoptosis-inducing conditions that was comparable to the conventional EB/AO method.
Figure 1 compares the results from our method and the conventional EB/AO methods for both control and treated Jurkat cells. Figure 1A exemplifies cell morphology obtained using both methods. We note that the definition is sharper by eye through the microscope than in the photos. In addition, the pictures obtained using the conventional method, which uses microscope slides, were sharper than those produced using the 96-well based method. This might be due to the relatively poor optical quality of the plastic bottom of 96-well plates. However, this did not encumber the quantification of live, apoptotic, and necrotic cells using the 96-well based method. Further, if high optical clarity is desirable, one may use glass-bottom 96-well plates to eliminate this possible limitation, or use a microscope with Hoffman modulation contrast to increase the optical quality of the plastic [27].
Determining the live, apoptotic and necrotic cells was comparable between these two methods for Jurkat cells: A) live cells have normal nuclei staining which present green chromatin with organized structures; B) apoptotic cells contain condensed or fragmented chromatin (green or orange); C) necrotic cells have similar normal nuclei staining as live cells except the chromatin is orange instead of green.
Overall, quantification of live, necrotic, and apoptotic control and treated Jurkat cells was quite comparable using both methods (Figure 1B and 1C). The chi-square tests showed that the cell counts were independent of the method used for each treatment condition (For Jurkat Control: Conventional vs. Modified, d.f. = 2, chi-square = 0.34, p > 0.8; For Jurkat Treated: Conventional vs. Modified, d.f. = 2, chi-square = 0.0015, p > 0.995). The largest differences in cell counts between the methods were only a 2% lower live cell count and a 2% higher apoptotic cell count for the treated condition using the conventional method (not statistically significant, Figure 1C). However, these small variations would be within the standard deviations of these assays. Thus, for suspension cells, the modified method was quite similar to the conventional method in terms of quantification and detection of the live, apoptotic, and necrotic cells.
Figure 2 shows the comparison between the modified and conventional methods of staining for A375 adherent cells. Figure 2A presents the morphology of A375 cells from different conditions using the conventional and our 96-well based method. Our method illustrated the morphology of live adherent cells attached to the plate in the control well, and it showed rounded up, detached cells in the treated well. This allowed us to observe more features for distinguishing live cells from apoptotic cells (Figure 2A). In contrast, all the cells stained from the conventional method were rounded up since it requires one to detach the cells. Thus, the modified method allows better morphological assessment of apoptotic condition when compared to the conventional staining technique.
As for Jurkat cells, the quantification of live, necrotic, and apoptotic cell populations in both control and treated A375 cells was quite comparable using both methods (Figure 2B and 2C). The chi-square tests showed that the cell counts were again statistically independent of the staining method used for both treatment conditions (For A375 Control: Conventional vs. Modified, d.f. = 2, chi-square = 0.31, p > 0.8; For A375 Treated: Conventional vs. Modified, d.f. = 2, chi-square = 0.86, p > 0.6). In this case, the largest differences in quantification obtained using these two methods were only 3% necrotic and apoptotic cell counts for treated A375 cells (not statistically significant, Figure 2C). Thus, in accordance with the suspension cell results, the two staining methods were quite comparable for adherent cells. Further, we also found similar results with A375 cells treated with other apoptotic triggers (data not shown).
We note that there was a lot more debris in the adherent cell samples produced using the conventional method. This suggests that the conventional method, with multiple step preparation of adherent cells, could cause some cell damage and break apart some highly fragile cells such as late apoptotic or necrotic cells. This undesired artifact might change the distribution of live, apoptotic, and necrotic cells in the cell population under some conditions.
These results suggest that our method is especially well suited for analyzing adherent cells, since it provides further characteristics to differentiate live cells from apoptotic cells. Additionally, it is much easier to collect all the cells, including floaters, than using the conventional method. Furthermore, by eliminating the detaching and washing steps, this method drastically reduces the time needed to perform the test, decreases the possibility of losing floating cells as in the conventional method, minimizes damage to adherent cells, and may allow more accurate quantification of apoptotic status under certain conditions.
Table 1 summarizes the advantages and disadvantages from various apoptotic detection methods including our 96-well-based EB/AO assay. Our method is the only one with the advantages of having one-step, high specificity, quantification of all three types of cells at the same time, and available in a 96-well format.
Conclusion
We present here an improved EB/AO staining method by substituting the detaching and washing steps with a simpler centrifugation step with 96-well plates. This modified method combines the advantage of the 96-well format and the conventional EB/AO staining method for apoptotic quantification. Our new technique is easy to perform, time efficient, and especially suitable for adherent cells. In addition, inverted fluorescence microscopes and 96-well plate holders for centrifuges are readily available in most research departments. Our new method only requires small amounts of inexpensive EB and AO for each experiment.
Since this method is straightforward to perform with only one step of gentle centrifugation using 96-well plates, it can also be easily combined with other 96-well-plate-based assays within one experiment, such as cell viability assay (MTS), cell death assay (LDH), or certain caspase activity assays. Therefore, multiple endpoints of cell death and apoptosis can be measured in a single experiment with very small amounts of cells. This could be very valuable for the cells difficult to grow in large amounts (e.g., the short term cultures of patient or animal samples).
In summary, our apoptotic quantification method represents an improvement over the currently available techniques especially for adherent cells. It is highly specific, simple, time efficient, and available in a 96-well format. Further, it can quantify live, apoptotic and necrotic cells at the same time.
Methods
Cell lines and normal culture conditions
Both A375 (a human melanoma cell line) and Jurkat (a human leukemia T cell line) were obtained from ATCC (American Type Culture Collection, Manassas, VA, USA). Cells were maintained in RPMI Medium 1640 (Invitrogen, Grand Island, New York, USA) with 10% fetal bovine serum (Gemini Bio-Products, Inc. CA, USA) and grown in an incubator at 37°C with 5% CO2. All tissue culture plates and other plastic wares were from Costar (Corning, New York, USA).
Reagents
Camptothecin was purchased from Biovision (Mountain View, CA, USA). Acridine orange (AO), ethidium bromide (EB), and ethylenediamine-tetraacetic acid (EDTA) were purchased from Sigma (St. Louis, MO, USA).
The dye mix for the EB/AO staining was 100 μg/ml acridine orange and 100 μg/ml ethidium bromide in PBS [2].
Induction of apoptosis in Jurkat cells with camptothecin treatment
Four ml of 5 × 105 cells/well were seeded in a 6-well plate the night before the treatment. Cells were treated with camptothecin at a final concentration of 6 μM for 4 hr in the 37°C incubator before the cells were subjected to EB/AO staining. Then, 1 ml of cell suspension was used for conventional EB/AO staining, and 100 μl of cell suspension was transferred to a 96-well plate for modified EB/AO staining.
Induction of apoptosis in A375 cells with camptothecin treatment
For the conventional EB/AO method, 2.5 × 105 cells/well were seeded in a 6-well plate the night before the treatment. For our modified EB/AO method, 1.6 × 104 cells/well were seeded in a 96-well plate. Cells were treated with camptothecin at a final concentration of 6 μM for 48 hr in the incubator before the cells were subjected to EB/AO staining procedure.
Conventional EB/AO staining
Procedures were followed as described previously in Current Protocols of Immunology [1]. Briefly, cells were harvested as the following: 1) For suspension cells, 1 ml of Jurkat cell suspensions was transferred to a 15 ml tube. 2) For adherent cells, supernatant (medium and floating A375 cells) were transferred to 15 ml tubes. The rest of the adherent cells were detached with PBS-EDTA, Dulbecco's phosphate buffered saline (Invitrogen, Grand Island, New York, USA) containing 1 mM EDTA. The supernatant and the detached cells from the same sample were pooled together in the 15 ml tubes.
Both Jurkat and A375 cells were pelleted by centrifuged at 1,000 RPM (129 g) for 5 minutes using a Beckman Model TJ-6 centrifuge, and washed with 1 ml of cold PBS once. Cell pellets were then re-suspended in 25 μl cold PBS and 2 μl EB/AO dye mix was added. Stained cell suspension (10 μl) were placed on a clean microscope slide and covered with a coverslip. Cells were viewed and counted using a Nikon eclipse TS100 inverted microscope at 400× magnification with excitation filter 480/30 nm; dichromatic mirror cut-on 505 nm LP; and barrier filter 535/40 nm (Melville, NY, USA). Pictures were taken with a Nikon COLPIX digital camera. Tests were done in triplicate, counting a minimum of 100 total cells each.
96-well-based EB/AO staining
For both suspension and adherent cells, 96-well plates were centrifuged at 1,000 RPM (129 g) for 5 minutes using a Beckman Model TJ-6 centrifuge with inserts for 96-well plates. EB/AO dye mix (8 μl) was added to each well, and cells were viewed under the same microscope as above. Tests were done in triplicate, counting a minimum of 100 total cells each.
Statistical analysis
To test the hypothesis that frequencies of observed live, necrotic, and apoptotic cells were independent of the staining method used, we employed chi-square tests for independence in contingency tables using Microsoft Excel. For each contingency table, triplicate counts of live, necrotic, and apoptotic cells for both the traditional and modified staining methods were averaged.
Authors' contributions
YGS conceived the study, participated in its design and coordination, and drafted the manuscript. DR participated in experiment design, carried out the experiments, and helped with the first draft of the manuscript. NBG performed the statistical analysis and final editing of the manuscript. DAN provided the general guidance and research funding for the study. All authors read and approved the final manuscript.
Acknowledgements
This work was supported in part by Dermatology Foundation career development award (YGS), NIAMS grant R01AR26427-18 (DAN) and Veterans Administrations merit grant PN: 0012 (DAN).
Figures and Tables
Figure 1 Comparison of a 96-well-based and the conventional EB/AO method for Jurkat suspension cells. Jurkat cells were treated with 6 μM camptothecin for induction of apoptosis. (A) -Images of the cells. Yellow arrows next to "L" point to live cells; white arrows next to "A" indicate apoptotic cells; and red arrows next to "N" indicate necrotic cells. (B) Quantification of untreated live, necrotic, and apoptotic cells. (C) Quantification of live, necrotic, and apoptotic cells treated with camptothecin. p-values from chi-square tests were marked on the bottom of each graph.
Figure 2 Comparison of a 96-well-based and the conventional EB/AO method for A375 adherent cells. A375 cells were treated with 6 μM camptothecin for induction of apoptosis. Symbols and labels are used in the same way as in Figure 1.
Table 1 Comparison of the advantages of various apoptotic detection methods.
ASSAY ONE-STEP QUANTIFICATION OF LIVE, NECROTIC AND APOPTOTIC CELLS SPECIFICITY 96-WELL FORMAT
Our 96-well-based EB/AO + + + +
Caspase 3/7 Activity + - + +
Conventional EB/AO - + + -
Annexin V/PI - + + -
ssDNA - - + +
DNA Laddering - - + -
TUNEL - - - +
DAPI stain - - + -
==== Refs
Cohen JJ Apoptosis Immunol Today 1993 14 126 130 8466628 10.1016/0167-5699(93)90214-6
Coligan JE Kruisbeek AM Margulies DH Shevach EM Strober W Coico R Current Protocols in Immunology Related isolation procedures and functional assays 1995 1 John Wiley & Sons, Inc. 3.17.1
Boersma AW Nooter K Oostrum RG Stoter G Quantification of apoptotic cells with fluorescein isothiocyanate-labeled annexin V in chinese hamster ovary cell cultures treated with cisplatin Cytometry 1996 24 123 130 8725661 10.1002/(SICI)1097-0320(19960601)24:2<123::AID-CYTO4>3.0.CO;2-K
Martin SJ Reutelingsperger CP McGahon AJ Rader JA van Schie RC LaFace DM Green DR Early redistribution of plasma membrane phosphatidylserine is a general feature of apoptosis regardless of the initiating stimulus: inhibition by overexpression of Bcl-2 and Abl J Exp Med 1995 182 1545 1556 7595224 10.1084/jem.182.5.1545
Homburg CH de Haas M von dem Borne AE Verhoeven AJ Reutelingsperger CP Roos D Human neutrophils lose their surface Fc gamma RIII and acquire Annexin V binding sites during apoptosis in vitro Blood 1995 85 532 540 7812008
Vermes I Haanen C Steffens-Nakken H Reutelingsperger C A novel assay for apoptosis. Flow cytometric detection of phosphatidylserine expression on early apoptotic cells using fluorescein labelled Annexin V J Immunol Methods 1995 184 39 51 7622868 10.1016/0022-1759(95)00072-I
Wyllie AH Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation Nature 1980 284 555 556 6245367 10.1038/284555a0
Park DJ Patek PQ Detergent and enzyme treatment of apoptotic cells for the observation of DNA fragmentation Biotechniques 1998 24 558 560 9564520
Li X Darzynkiewicz Z Labelling DNA strand breaks with BrdUTP. Detection of apoptosis and cell proliferation Cell Prolif 1995 28 571 579 8555370
Li X Traganos F Melamed MR Darzynkiewicz Z Single-step procedure for labeling DNA strand breaks with fluorescein- or BODIPY-conjugated deoxynucleotides: detection of apoptosis and bromodeoxyuridine incorporation Cytometry 1995 20 172 180 7664627
Goldman AS Baker MK Piddington R Herold R Inhibition of programmed cell death in mouse embryonic palate in vitro by cortisol and phenytoin: receptor involvement and requirement of protein synthesis Proc Soc Exp Biol Med 1983 174 239 243 6634717
Thornberry NA Rano TA Peterson EP Rasper DM Timkey T Garcia-Calvo M Houtzager VM Nordstrom PA Roy S Vaillancourt JP Chapman KT Nicholson DW A combinatorial approach defines specificities of members of the caspase family and granzyme B. Functional relationships established for key mediators of apoptosis J Biol Chem 1997 272 17907 17911 9218414 10.1074/jbc.272.29.17907
Thornberry NA Lazebnik Y Caspases: enemies within Science 1998 281 1312 1316 9721091 10.1126/science.281.5381.1312
Karvinen J Hurskainen P Gopalakrishnan S Burns D Warrior U Hemmila I Homogeneous time-resolved fluorescence quenching assay (LANCE) for caspase-3 J Biomol Screen 2002 7 223 231 12097185 10.1089/108705702760047727
Gopalakrishnan SM Karvinen J Kofron JL Burns DJ Warrior U Application of Micro Arrayed Compound Screening (microARCS) to identify inhibitors of caspase-3 J Biomol Screen 2002 7 317 323 12230885 10.1089/108705702320351169
Preaudat M Ouled-Diaf J Alpha-Bazin B Mathis G Mitsugi T Aono Y Takahashi K Takemoto H A homogeneous caspase-3 activity assay using HTRF technology J Biomol Screen 2002 7 267 274 12097189 10.1089/108705702760047763
Frankfurt OS Immunoassay for single-stranded DNA in apoptotic cells Methods Mol Biol 2004 282 85 101 15105558
Frankfurt OS Krishan A Enzyme-linked immunosorbent assay (ELISA) for the specific detection of apoptotic cells and its application to rapid drug screening J Immunol Methods 2001 253 133 144 11384675 10.1016/S0022-1759(01)00387-8
Frankfurt OS Immunoassay for single-stranded DNA in apoptotic cells Methods Mol Biol 1999 113 621 631 10443457
Frankfurt OS Krishan A Microplate screening for apoptosis with antibody to single-stranded DNA distinguishes anticancer drugs from toxic chemicals J Biomol Screen 2003 8 185 190 12844439 10.1177/1087057103253326
Frankfurt OS Krishan A Identification of apoptotic cells by formamide-induced dna denaturation in condensed chromatin J Histochem Cytochem 2001 49 369 378 11181740
Frankfurt OS Robb JA Sugarbaker EV Villa L Monoclonal antibody to single-stranded DNA is a specific and sensitive cellular marker of apoptosis Exp Cell Res 1996 226 387 397 8806443 10.1006/excr.1996.0240
Cohen GM Sun XM Snowden RT Ormerod MG Dinsdale D Identification of a transitional preapoptotic population of thymocytes J Immunol 1993 151 566 574 8335900
Stewart BW Mechanisms of apoptosis: integration of genetic, biochemical, and cellular indicators J Natl Cancer Inst 1994 86 1286 1296 8064887
Walker NI Harmon BV Gobe GC Kerr JF Patterns of cell death Methods Achiev Exp Pathol 1988 13 18 54 3045494
Renvoize C Biola A Pallardy M Breard J Apoptosis: identification of dying cells Cell Biol Toxicol 1998 14 111 120 9553722 10.1023/A:1007429904664
Davidson MW Abramowitz M Molecular Expressions Microscopy Primer Specialized Microscopy Techniques 1998
| 15885144 | PMC1142306 | CC BY | 2021-01-04 16:02:57 | no | BMC Biotechnol. 2005 May 10; 5:12 | utf-8 | BMC Biotechnol | 2,005 | 10.1186/1472-6750-5-12 | oa_comm |
==== Front
BMC BiotechnolBMC Biotechnology1472-6750BioMed Central London 1472-6750-5-131588515010.1186/1472-6750-5-13Research ArticleCharacterisation and application of a bovine U6 promoter for expression of short hairpin RNAs Lambeth Luke S [email protected] Robert J [email protected] Morley [email protected] Brian P [email protected] Sean [email protected] Timothy J [email protected] CSIRO Livestock Industries, Australian Animal Health Laboratory, Geelong, VIC 3220, Australia2 School of Biological and Chemical Sciences, Deakin University, Geelong, VIC 3217, Australia3 CSIRO Livestock Industries, Queensland Bioscience Precinct, St Lucia, QLD 4067, Australia2005 11 5 2005 5 13 13 7 2 2005 11 5 2005 Copyright © 2005 Lambeth 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 use of small interfering RNA (siRNA) molecules in animals to achieve double-stranded RNA-mediated interference (RNAi) has recently emerged as a powerful method of sequence-specific gene knockdown. As DNA-based expression of short hairpin RNA (shRNA) for RNAi may offer some advantages over chemical and in vitro synthesised siRNA, a number of vectors for expression of shRNA have been developed. These often feature polymerase III (pol. III) promoters of either mouse or human origin.
Results
To develop a shRNA expression vector specifically for bovine RNAi applications, we identified and characterised a novel bovine U6 small nuclear RNA (snRNA) promoter from bovine sequence data. This promoter is the putative bovine homologue of the human U6-8 snRNA promoter, and features a number of functional sequence elements that are characteristic of these types of pol. III promoters. A PCR based cloning strategy was used to incorporate this promoter sequence into plasmid vectors along with shRNA sequences for RNAi. The promoter was then used to express shRNAs, which resulted in the efficient knockdown of an exogenous reporter gene and an endogenous bovine gene.
Conclusion
We have mined data from the bovine genome sequencing project to identify a functional bovine U6 promoter and used the promoter sequence to construct a shRNA expression vector. The use of this native bovine promoter in shRNA expression is an important component of our future development of RNAi therapeutic and transgenic applications in bovine species.
==== Body
Background
RNA interference (RNAi), a method of sequence specific gene knockdown, has been used to analyse gene function in plants, invertebrates, and more recently mammalian cells [1-3]. The conserved RNAi pathway involves the processing of double stranded RNA (dsRNA) duplexes into 21–23 nucleotide (nt) molecules known as small interfering RNAs (siRNA) to initiate gene knockdown [4-6]. Since the discovery of RNAi in animals [7] the use of long dsRNA in lower eukaryotes, especially in the model organism Caenorhabditis elegans, has been used to determine gene function [8,9]. However, in mammalian systems the cellular uptake of long dsRNA induces an antiviral defence mechanism initiated by interferon (IFN), leading to non-specific translational shutdown and apoptosis [10-12].
This non-specific cellular activity can be circumvented by the direct transfection of either chemically synthesised or in vitro transcribed siRNAs of approximately 21 nt in length into mammalian cells [1,13]. These short molecules do not activate the IFN response, but can induce reliable and efficient transient knockdown of target genes [14,15]. As a consequence, the development of DNA-based vectors for expression of short hairpin RNA (shRNA) molecules that are processed within the cell to produce active siRNA molecules has progressed rapidly [16-18]. Such DNA expression constructs have achieved highly efficient gene knockdown without induction of the IFN response.
DNA-based vectors offer some additional advantages over chemical and in vitro synthesised siRNA. Vector construction is much less expensive compared to the chemical synthesis of siRNA, selection of transfected cells is possible via antibiotic selection and the option of inducible shRNA transcription is also available. shRNA expression vectors have been engineered using both viral (including retroviral [19], adenoviral [20] and lentiviral [21] vectors), and plasmid systems [16-18]. These vectors often utilize promoters from a small class of pol. III promoters [22,23] to drive the expression of shRNA. Promoters of this type are preferred because they naturally direct the synthesis of small, highly abundant non-coding RNA transcripts, with defined termination sequences consisting of 4–5 thymidines (Ts) and have no requirement for downstream promoter elements [22-24].
The human U6 snRNA promoter is the best studied type III pol. III promoter. It has characteristic promoter elements known as the enhancer and core regions [25,26], and is frequently used in RNAi expression vectors. A total of nine full-length U6 loci have been identified from the human genome [27]. These genes are dispersed throughout the genome and five of these are potentially active, including the previously described human U6 promoter, now denoted U6-1. In the current study we describe the characterisation of the bovine U6 snRNA promoter and its expression of shRNA molecules in bovine cells. We used a PCR based cloning strategy to construct a plasmid vector that features the bovine U6 promoter to drive expression of shRNA molecules directed at the exogenously expressed Enhanced Green Fluorescent Protein (EGFP) and the endogenous bovine gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH).
Results and discussion
Characterisation of a bovine U6 promoter
A bovine BAC clone was identified from GenBank (Accession no. CC528275) that contained a 107 nt region that shared 100% identity with the human U6 snRNA sequence. A region directly upstream of this sequence also shared significant homology with previously identified human U6 promoter sequences. BLAT analysis revealed that this BAC-clone shared the greatest sequence homology with the human U6-8 promoter located on chromosome 14 [27]. The bovine BAC sequence contained upstream promoter elements consistent with the human U6 promoters and other human pol. III promoters [27-30] (Fig. 1). These elements are also present in the mouse U6 promoter used in the pSilencer 1.0-U6 siRNA Expression Vector (Ambion). The location and spacing of these elements is similar for all human U6 promoters and their requirement for pol. III activity is well documented [28-30]. The presence of these elements in the bovine BAC-clone sequence directly upstream of the full-length U6 snRNA suggested that this region may be an active bovine promoter and could be used effectively to promote shRNA expression.
shRNA expression vector construction
Using the predicted bovine U6 snRNA promoter sequence as template, PCR based cloning strategies were used to generate shRNA expression vectors targeting EGFP and bovine GAPDH. A two-step PCR reaction was used to produce a vector targeting EGFP (pBovineU6-shEGFP) (Fig. 2A). One-step PCR reactions were used to produce an shRNA expression vector targeting GAPDH (pBovineU6-shGapdh), and a non-specific control shRNA (pBovineU6-shScrambled) (Fig. 2B). In addition, using the mouse U6 snRNA promoter sequence from pSilencer-1.0 as template in one-step PCR reactions, shRNA expression vectors targeting EGFP (pMouseU6-shEGFP) and bovine GAPDH (pMouseU6-shGapdh) were produced (Fig. 2B).
In all constructs, the first nucleotide of the predicted shRNA was a guanine (G) residue, corresponding to the first nucleotide of the native U6 snRNA. An XbaI restriction enzyme site was engineered downstream of the termination signal to allow screening for full-length shRNA products inserted into pGEM-T Easy (Promega) which lacks an XbaI site. All final shRNA expression constructs consisted of either the full-length bovine or mouse U6 promoter, a shRNA sense sequence, a loop sequence, a shRNA antisense sequence, a termination sequence and an XbaI site.
Activity of the bovine U6 promoter measured by EGFP knockdown
To analyse the function of the bovine U6 promoter, the level of EGFP expression in cells cotransfected with pBovineU6-shEGFP and pEGFP-N1 (Clonetech) were directly compared with EGFP expression in cells cotransfected with pBovineU6-shScrambled and pEGFP-N1. An indication of the bovine promoter efficiency was provided by direct comparison to the level of EGFP knockdown by the same shRNA molecule expressed from pMouseU6-shEGFP. Prior to validation in bovine cells, both mouse and bovine U6 promoter driven shRNA expression vectors were first validated in Vero cells. As this cell line lacks the interferon α, β and ω genes [31,32], a significant reduction in EGFP expression could be attributed to RNAi and not the result of non-specific inhibition of protein translation characteristic of the IFN response triggered by expressed exogenous dsRNA.
For each transfection condition, knockdown of EGFP in Vero and MDBK cell lines was visualised by fluorescence microscopy (Fig. 3A). Flow cytometry was used to determine the Mean Fluorescence Intensity (MFI) (Fig. 3B). Results in both cell lines showed that cells transfected with either pMouseU6-shEGFP or pBovineU6-shEGFP exhibited large reductions in EGFP expression when compared with cells transfected with pBovineU6-shScrambled.
To ensure that the observed reduction in EGFP expression could be directly attributed to RNAi induced by expressed shRNA, the transcription of these molecules was detected in transfected cells. A radiolabelled RNA probe complementary in sequence to the EGFP shRNA sequence was used in an RNAse protection assay to visualise these molecules. The mouse microRNA (miRNA) miR-16 probe used as a loading control for each condition produced a strong signal at the expected size for all samples (Fig. 3C). The EGFP shRNA was detected only in those samples that were transfected with either of the EGFP shRNA vectors (Fig. 3C). There were no obvious differences in the amount of shRNA produced by either the mouse or bovine U6 promoters.
Activity of the bovine U6 promoter measured by GAPDH knockdown
To further validate the function of the bovine U6 promoter for RNAi, shRNA expression vectors targeting the endogenous bovine gene GAPDH were produced. An active shRNA sequence was identified by testing three siRNA sequences directed at this gene. Real-time PCR was used to determine the relative amount of GAPDH in MDBK cells transfected with the three siRNAs compared to cells transfected with the scrambled sequence siRNA control. A similar level of knockdown was achieved for each of the bovine GAPDH siRNAs (Fig. 4A). The most appropriate siRNA sequence, GAPDH #1, was selected for use in construction of an shRNA vector as this sequence did not contain any runs of T or A residues that may cause premature termination of expressed shRNAs. This sequence was used for both mouse and bovine U6 shRNA vectors (pMouse-shGapdh and pBovine-shGapdh). Using Real-time PCR, the relative amount of GAPDH in MDBK cells transfected with pMouse-shGapdh and pBovine-shGapdh was determined by comparison to cells transfected with the irrelevant control plasmid pBovine-shScrambled (Fig. 4B). Both vectors produced a very similar level of GAPDH knockdown, although not as effective as the GAPDH siRNAs. This is probably because the transfection efficiency for siRNAs is generally much greater than that of large plasmid vectors.
Conclusion
We mined data from the bovine sequencing project and identified a bovine U6 snRNA promoter that is the putative bovine homologue of the human U6-8 snRNA promoter. This pol. III promoter sequence contains a number of functional sequence elements that are characteristic of this type of promoter and are essential for promoter function. The bovine U6 promoter sequence was used in the construction of plasmid based shRNA expression vectors pBovineU6-shEGFP and pBovineU6-shGapdh. These vectors efficiently induced RNAi in MDBK and Vero cells through production of shRNA molecules targeted at the exogenously expressed reporter gene EGFP, and the endogenous bovine gene GAPDH. The use of this promoter sequence and the shRNA vector cloning strategy described here will be advantageous in RNAi functional genomic experiments in bovine cells. The characterisation of this U6 promoter is an important step in the development of novel bovine species specific RNAi based therapeutics. For ethical reasons it is essential to minimise the introduction of non-bovine DNA sequences, consequently this research could be of significance in the transgenic delivery of shRNA molecules in bovine species.
Methods
Bovine U6 promoter isolation
Oligonucleotides synthesized in this study are indicated in Table 1. All the oligonucleotides were obtained from GeneWorks Pty Ltd, except for LL91 which was obtained from Proligo, and bovineGpdhPROBE from Applied Biosystems. The bovine BAC-end sequences deposited in the Genome Survey Sequence (GSS) section of GenBank, were compared with the U6 >95% identity sequence set for Rfam models RF00026 using BLAST with default parameters and a threshold e-value of 0.01 [33]. All hits detected were scored against the appropriate Rfam covariance model using the cmsearch function of INFERNAL [34]. Flanking repeat elements were identified using Repeatmasker with the minus cow option . Flanking sequences highly conserved between Bos taurus and the human genome sequence were identified using BLAT search (BLAST-like Alignment tool; UC-Santa Cruz genome server [35]) from the UC-Santa Cruz genome bioinformatics site .
A region of the bovine genome homologous to a bovine BAC-end sequence (GenBank Accession no. CC528275) predicted to contain a U6 snRNA promoter was amplified from Bos taurus genomic DNA isolated from whole blood using Wizard Genomic DNA Purification kit (Promega). First round PCR amplification used forward primer LL16 with a reverse primer TD66 designed from the highly conserved human U6 snRNA sequence. This PCR product was then used as template for semi-nested PCR with LL16 and antisense primer TD72 located 11 nt upstream of TD66 in the U6 snRNA sequence. The second round PCR produced a clean band of expected size (507 nt) that was gel purified using QIAquick gel extraction kit (Qiagen), ligated into pGEM-T Easy (Promega) as per the manufacturers instructions and sequenced.
Expression vector construction and shRNA target sites
A bovine U6 EGFP shRNA construct (pBovineU6-shEGFP) was produced using a two-step PCR approach. The 1st PCR paired LL16 with reverse primer LL23, comprising the last 20 nt of the promoter sequence, EGFP shRNA sense, loop, and 3 nt EGFP shRNA antisense sequence. This PCR product was used as template for semi-nested PCR to produce the full-length shRNA template, using LL16 and reverse primer LL13 with a short overlapping region, comprising the 4nt EGFP shRNA sense, loop, EGFP shRNA antisense, termination and XbaI.
A mouse U6 EGFP shRNA construct (pMouseU6-shEGFP) was produced using one-step PCR with pSilencer 1.0-U6 siRNA Expression Vector (Ambion) as template. Universal primer M13 Forward was paired with reverse primer TD134, comprising the last 20 nt of the mouse promoter, and all other EGFP shRNA components.
The EGFP shRNA sequence used for both bovine and mouse vectors, had been shown previously to be effective in silencing gene expression [36]. A bovine U6 EGFP scrambled shRNA sequence (pBovineU6-shScrambled) was produced using a sequence that showed no significant homology to available monkey or bovine sequence data. Forward primer LL16 was paired with reverse primer LL31 comprising the last 20 nt of the bovine U6 promoter, and all other EGFP scrambled shRNA components.
Three siRNAs were designed based on the bovine GAPDH sequence obtained from GeneBank (Accession no. U85042) using the criteria described by Elbashir et al., 2002 (37). Target sites for bovine GAPDH were; GAPDH #1 (5'-AAGTTCAACGGCACAGTCAAG-3'); GAPDH #2 (5'-AACTTGACTGTGCCGTTGAAC-3'); GAPDH #3 (5'-AAGGTCATCCATGACCACTTT-3'). To generate siRNAs, oligonucleotides for sense and antisense-strands of each siRNA together with T7 promoter sites were synthesized and siRNAs were produced using the Ambion Silencer siRNA construction kit (Ambion). Oligonucleotides used for each were; GAPDH #1: LBG-as1 and LBG-as2; GAPDH #2: LBG-as3 and LBG-s4; GAPDH #3: LBG-as5 and LBG-s6 (Table 1).
A bovine U6 GAPDH shRNA expression construct (pBovineU6-shGapdh) was generated using a one-step PCR approach. LL16 was paired with LL05 which comprised the last 20 nt of the bovine U6 promoter and all other GAPDH shRNA sequence components including the GAPDH #1 siRNA sequence. Similarly, a mouse U6 GAPDH shRNA expression construct (pMouseU6-shGapdh) was also generated. M13 universal primer was paired with LL06, which also comprised the last 20 nt of the mouse U6 promoter and all other GAPDH shRNA components. All PCR products for shRNA expression constructs were ligated into pGEM-T Easy (Promega) and sequenced.
Cell culture and transfection
MDBK (Madin Darby Bovine Kidney) and Vero (African Green monkey kidney) cell lines were cultured in Eagle's minimal essential medium (EMEM) medium containing 10% fetal calf serum (FCS), 2 mM glutamine, 10 mM HEPES, supplemented with penicillin (100 U/ml) and streptomycin (100 μg/ml). All cells were cultured in humidified atmosphere containing 5% CO2 at 37°C. Vero cells were grown to approximately 80% confluence on either 24-well plates (Nunc) for Flow Cytometry, on 8-well chamber slides (Lab-Tek) for fluorescence microscopy, or 6-well plates for shRNA detection. Cotransfection with 500 ng for 24-well plates and chamber slides, or 2.5 μg for 6-well plates of plasmid DNA (shRNA plasmids and/or pEGFP-N1 (Clonetech)) was carried out using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. MDBK cells were grown to 80–90% confluence in 75 cm2 flasks (Corning), harvested and divided into aliquots of 1 × 106 cells per transfection. Approximately 2.5 μg of each plasmid was transfected per aliquot by electroporation using a Nucleofector (Amaxa) according to the manufacturer's instructions. Electroporated cells were dispensed into 8-well chamber slides for fluorescence microscopy, 6-well plates (Nunc) for flow cytometry or 24-well plates for Real-time PCR. Transfection of GAPDH siRNAs in MDBK cells grown to approximately 80% confluence on 24-well plates (Nunc) was achieved using Lipofectamine 2000 with approximately 10 nM of each siRNA.
EGFP and GAPDH knockdown assays
EGFP expression was monitored at 72-hour post-transfection using fluorescence microscopy (Leica DMLB). Vero and MDBK cells cultured in 8-well chamber slides were washed with PBSA and mounting solution was applied (9:1 glycerol: PBSA). Cells to be analysed by flow cytometry were trypsinized and washed in PBSA. Cells were then resuspended in 200 μL 0.01% sodium azide and 2% FCS in PBSA and analysed using a FACScalibur (Becton Dickinson) flow cytometer. Data analysis was performed using CELLQuest software (Becton Dickinson).
Real-time RT-PCR reactions were carried out 48-hours post-transfection. MDBK cells transfected with GAPDH siRNAs and GAPDH shRNA vectors were harvest and cDNA was produced using the Cells-to-cDNA II kit (Ambion) using random hexamers. cDNA (2.5 μL) was added to a 22.5 μL Real-time RT -PCR master mix containing 12.5 μL 2X TaqMan Universal PCR Master Mix (Applied Biosystems), 2.5 μL (9 μM) each of sense primer BovineGpdhF and antisense primer bovineGpdhR, 2.5 μL (2.5 μM) of TaqMan MGB Probe (bovineGpdhPROBE) (Applied Biosystems) and 2.5 μL RNase-free water (Table 1). For RNA normalization, an 18s rRNA PCR was performed for each cDNA using the same PCR reagents except for the primers and probes which were the Eukaryotic 18srRNA Endogenous Control (Applied Biosystems). Real-time PCR was carried out in a 7700 Sequence Detection Instrument (Applied Biosystems) using the following thermal cycling profile: 95°C 1 min, followed by 35 cycles of amplification (95°C 15s, 61°C 30s, 68°C 30s).
shRNA detection
Detection of EGFP shRNAs was carried by out using an RNAse protection assay on extracts enriched for small RNAs isolated from transfected Vero cells using a mirVana miRNA Isolation Kit (Ambion). The RNA oligonucleotide LL91 was end labelled with [gamma-32P] ATP using a mirVana Probe & Marker Kit (Ambion) and hybridised to the enriched RNAs using the mirVana miRNA Detection Kit (Ambion). The RNA fragments were then separated by electrophoresis on a 15% denaturing polyacrylamide/ 8 M Urea gel and detected by autoradiography.
Authors' contributions
LSL carried out all experiments and drafted the manuscript. TJD, RJM and MM participated in design and coordination of the research and revision of the manuscript. BPD and SM helped with bioinformatics and revision of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Dr. Jef Hammond and Dr. Sandra Sapats for critically reading the manuscript. The technical assistance provided by Scott Tyack, Terry Wise, Matthew Bruce, Daniel Layton and Tony Pye was also greatly appreciated.
Figures and Tables
Figure 1 Promoter element sequences of the bovine U6, human U6-8 and mouse U6 promoters. The distal promoter regions containing the SPH and OCT sequences and proximal promoter regions containing the PSE and TATA sequence elements are shown for each promoter. Matches to the consensus sequence delineated at the top of the SPH, OCT and PSE sequences are shown in upper case (SPH consensus is from Schaub et al., 1999 [28], PSE consensus is from Dahlberg et al., 1988 [29] and OCT consensus is from Sturm et al. 1988 [30]). Note that the relative positions of the SPH and OCT elements are switched in the mouse promoter.
Figure 2 Schematic representation of the PCR strategies used to produce shRNA expression vectors. (A): The two-step PCR method used to generate pBovineU6-shEGFP. The 1st PCR amplified the bovine U6 promoter, EGFP shRNA sense, loop, and 3 nt of EGFP shRNA antisense using primers LL16 and LL23. The 2nd PCR amplified the bovine U6 promoter and the remaining EGFP shRNA components including EGFP shRNA antisense, terminator and XbaI using primers LL16 and LL13. (B): The one-step PCR method used to generate pMouseU6-shEGFP, pBovineU6-shScrambled, pMouseU6-shGapdh and pBovineU6-shGapdh. PCR reactions used forward primers paired with single reverse primers comprising all shRNA components. All final PCR products consisted of a mouse or bovine U6 promoter, shRNA sense, loop, shRNA antisense, termination sequence and XbaI site.
Figure 3 Inhibition of EGFP expression and detection of expressed shRNA from bovine and mouse U6 promoters. (1) transfection free negative control (cells only); (2) transfection of pEGFP-N1; (3) cotransfection of pEGFP-N1 and pBovineU6-shScrambled; (4) cotransfection of pEGFP-N1 and mouseU6-shEGFP; (5) cotransfection of pEGFP-N1 and pBovineU6-shEGFP. All transfections were performed in triplicate. (A): Typical fluorescence images of Vero and MDBK cells (Magnifications ×100). (B): Mean Fluorescence Intensity (MFI) of Vero and MDBK cells 72 hours post-transfection determined by flow cytometry. MFI of EGFP expressed as a percentage of the MFI of pEGFP-N1 + pBovineU6-shScrambled. (C): Detection of EGFP shRNAs in Vero cell extracts. Small RNAs were isolated from Vero cells and hybridised to a radiolabelled RNA probe identical in sequence to the EGFP shRNA (shEGFP). shEGFP was detected in lanes 4 and 5 only, and miR-16 was used as a loading control.
Figure 4 Inhibition of bovine GAPDH expression in MDBK cells. The amount of GAPDH in MDBK cells was determined by Real-time PCR 48 hours post-transfection, expressed as a relative percentage compared to the amount of GAPDH in cells transfected with a scramble sequence siRNA. (A): Cells transfected with either siRNAs directed at bovine GAPDH or a scrambled sequence control siRNA. (B): MDBK cells transfected with pBovine-shScrambled, pMouse-shGapdh and pBovine-shGapdh.
Table 1 Synthesized oligonucleotides used in this study
Name Sequence
LL16 5'-CCAGCAAAGCAACCGATTC-3'
TD66 5'-GGCCATGCTAATCTTCTCTG-3'
TD72 5'-TTTTAGTATATGTGCTGCCG-3'
LL23 5'-ATCTCTCTTGAAGATGAACTTCAGGGTCAGCGGTTTGCCTTTCGTCCACGG-3'
LL13 5'-TCTAGATTCCAAAAAAGCTGACCCTGAAGTTCATCTCTCTTGAAGATG-3'
TD134 5'-TCTAGATTCCAAAAAAGCTGACCCTGAAGTTCATCTCTCTTGAAGATGAACTTCAGGG
TCAGCCAAACAAGGCTTTCTCCA-3'
LL31 5'-TCTAGATTCCAAAAAAAGCGCAGTGTTACTCCACTTCTCTTGAAAGTGGAGTAACACT
GCGCTGGTTTGCCTTTCGTCCACGG-3'
LBG-as1 5'-AAGTTCAACGGCACAGTCAAGCCTGTCTC-3'
LBG-s2 5'-AACTTGACTGTGCCGTTGAACCCTGTCTC-3'
LBG-as3 5'-AAGGTCATCCATGACCACTTTCCTGTCTC-3'
LBG-s4 5'-AAAAAGTGGTCATGGATGACCCCTGTCTC-3'
LBG-as5 5'-AAGCTCATTTCCTGGTACGACCCTGTCTC-3'
LBG-s6 5'-AAGTCGTACCAGGAAATGAGCCCTGTCTC-3'
LL05 5'- TCTAGATTCCAAAAAAGTTCAACGGCACAGTCAAGTCTCTTGAACTTGACTGTGCCGT
TGAACGGTTTGCCTTTCGTCCACGG-3'
LL06 5'-TCTAGATTCCAAAAAAGTTCAACGGCACAGTCAAGTCTCTTGAACTTGACTGTGCCGTT
GAACCAAACAAGGCTTTCTCCA-3'
BovineGpdhF 5'-AACGACCACTTTGTCAAGCTCAT-3'
BovineGpdhR 5'-TCCACCACCCTGTTGCTGTA-3'
BovineGpdhPROBE 5'-TCCTGGTACGACAATGA-3'
LL91 5'-RGRAURGRARARCUURCRARGRGRGURCRARGRC-3'
'r' prefix to G, A and C represents ribonucleotide, and U indicates ribo-uridine.
==== Refs
Napoli C Lemieux C Jorgensen R Introduction of a Chimeric Chalcone Synthase Gene into Petunia Results in Reversible Co-Suppression of Homologous Genes in trans Plant Cell 1990 2 279 289 12354959 10.1105/tpc.2.4.279
Sharp PA RNAi and double-stranded RNA Genes Dev 1999 13 139 141 9925636 10.1101/gad.13.22.3015
Elbashir SM Harborth J Lendeckel W Yalcin A Weber K Tuschl T Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells Nature 2001 411 494 498 11373684 10.1038/35078107
Fire A RNA-triggered gene silencing Trends Genet 1999 15 358 363 10461204 10.1016/S0168-9525(99)01818-1
Bosher JM Labouesse M RNA interference: genetic wand and genetic watchdog Nat Cell Biol 2000 2 E31 E36 10655601 10.1038/35000102
Hannon GJ RNA interference Nature 2002 418 244 251 12110901 10.1038/418244a
Fire A Xu S Montgomery MK Kostas SA Driver SE Mello CC Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans Nature 1998 391 806 811 9486653 10.1038/35888
Barstead R Genome-wide RNAi Curr Opin Chem Biol 2001 5 63 66 11166650 10.1016/S1367-5931(00)00173-3
Ashrafi K Chang FY Watts JL Fraser AG Kamath RS Ahringer J Ruvkun G Genome-wide RNAi analysis of Caenorhabditis elegans fat regulatory genes Nature 2003 421 268 272 12529643 10.1038/nature01279
Williams BR Role of the double-stranded RNA-activated protein kinase (PKR) in cell regulation Biochem Soc Trans 1997 25 509 13 9191145
Stark GR Kerr IM Williams BR Silverman RH Schreiber RD How cells respond to interferons Annu Rev Biochem 1998 67 227 264 9759489 10.1146/annurev.biochem.67.1.227
Gil J Esteban M Induction of apoptosis by the dsRNA-dependent protein kinase (PKR): mechanism of action Apoptosis 2000 5 107 114 11232238 10.1023/A:1009664109241
Tuschl T Expanding small RNA interference Nat Biotechnol 2002 20 446 448 11981553 10.1038/nbt0502-446
Dykxhoorn DM Novina CD Sharp PA Killing the messenger: short RNAs that silence gene expression Nat Rev Mol Cell Biol 2003 4 457 467 12778125 10.1038/nrm1129
Duxbury MS Whang EE RNA interference: a practical approach J Surg Res 2004 117 339 344 15047140 10.1016/j.jss.2003.09.010
Brummelkamp TR Bernards R Agami R A system for stable expression of short interfering RNAs in mammalian cells: Science 2002 296 550 553 11910072 10.1126/science.1068999
Yu JY DeRuiter SL Turner DL RNA interference by expression of short-interfering RNAs and hairpin RNAs in mammalian cells Proc Natl Acad Sci U S A 2002 99 6047 6052 11972060 10.1073/pnas.092143499
Paddison PJ Caudy AA Bernstein E Hannon GJ Conklin DS Short hairpin RNAs (shRNAs) induce sequence-specific silencing in mammalian cells Genes Dev 2002 16 948 958 11959843 10.1101/gad.981002
Devroe E Silver PA Retrovirus-delivered siRNA BMC Biotechnol 2002 2 15 12199908 10.1186/1472-6750-2-15
Huang A Chen Y Wang X Zhao S Su N White DW Functional silencing of hepatic microsomal glucose-6-phosphatase gene expression in vivo by adenovirus-mediated delivery of short hairpin RNA FEBS Lett 2004 558 69 73 14759518 10.1016/S0014-5793(03)01499-6
Abbas-Terki T Blanco-Bose W Deglon N Pralong W Aebischer P Lentiviral-mediated RNA interference Hum Gene Ther 2002 13 2197 2201 12542850 10.1089/104303402320987888
Paule MR White RJ Transcription by RNA polymerase I and III Nucleic Acids Res 2000 28 1283 1298 10684922 10.1093/nar/28.6.1283
Schramm L Hernandez N Recruitment of RNA polymerase III to its target promoters Genes Dev 2002 16 2593 2620 12381659 10.1101/gad.1018902
Geiduschek EP Kassavetis GA The RNA polymerase III Transcription Apparatus J Mol Biol 2001 310 1 26 11419933 10.1006/jmbi.2001.4732
Kunkel GR Pederson T Upstream elements required for efficient transcription of a human U6 RNA gene resemble those of U1 and U2 genes even though a different polymerase is used Genes Dev 1988 2 196 204 3360322
Ma B Hernandez N Redundant cooperative interactions for assembly of a human U6 transcription initiation complex Mol Cell Biol 2002 22 8067 8078 12391172 10.1128/MCB.22.22.8067-8078.2002
Domitrovich AM Kunkel GR Multiple, dispersed human U6 small nuclear RNA genes with varied transcriptional efficiencies Nucleic Acids Res 2003 31 2344 2352 12711679 10.1093/nar/gkg331
Schaub M Krol A Carbon P Flexible Zinc Finger Requirement for Binding of the Transcriptional Activator Staf to U6 Small Nuclear RNA and tRNA(Sec) Promoters J Biol Chem 1999 274 24241 24249 10446199 10.1074/jbc.274.34.24241
Dahlberg JE Lund E Birnstiel ML The genes and transcription of the major small nuclear RNAs Structure and Function of Major and Minor Small Nuclear Ribonucleoprotein Particles 1988 Berlin Germany: Springer-Verlag 38 70
Sturm RA Herr W The ubiquitous octamer-binding protein Oct-1 contains a POU domain with a homeo box subdomain Genes Dev 1988 2 1582 1599 2905684
Diaz MO Ziemin S Le Beau MM Pitha P Smith SD Chilcote RR Rowley JD Homozygous deletion of the alpha- and beta 1-interferon genes in human leukemia and derived cell lines Proc Natl Acad Sci U S A 1988 85 5259 5263 3134658
Diaz MO Pomykala HM Bohlander SK Maltepe E Malik K Brownstein B Olopade OI Structure of the human type-I interferon gene cluster determined from a YAC clone contig Genomics 1994 22 540 552 8001965 10.1006/geno.1994.1427
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
Eddy SR 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
Kent WJ BLAT-the BLAST-like alignment tool Genome Res 2002 12 656 664 11932250 10.1101/gr.229202. Article published online before March 2002
Kim DH Rossi JJ Coupling RNAi-mediated target downregulation with gene replacement Antisense Nucleic Acid Drug Dev 2003 13 151 155 12954115 10.1089/108729003768247619
Elbashir SM Harborth J Weber K Tuschl T Analysis of gene function in somatic mammalian cells using small interfering RNAs Methods 2002 26 199 213 12054897 10.1016/S1046-2023(02)00023-3
| 15885150 | PMC1142307 | CC BY | 2021-01-04 16:02:57 | no | BMC Biotechnol. 2005 May 11; 5:13 | utf-8 | BMC Biotechnol | 2,005 | 10.1186/1472-6750-5-13 | oa_comm |
==== Front
BMC Evol BiolBMC Evolutionary Biology1471-2148BioMed Central London 1471-2148-5-301587635110.1186/1471-2148-5-30Research ArticleDiaspora, a large family of Ty3-gypsy retrotransposons in Glycine max, is an envelope-less member of an endogenous plant retrovirus lineage Yano Sho T [email protected] Bahman [email protected] Arpita [email protected] Howard M [email protected] Department of Molecular Genetics and Cell Biology, University of Chicago, Chicago, IL 60637 USA2 Department of Biomolecular Chemistry, University of Wisconsin, Madison, WI 53706 USA3 Neuronautics, Inc., Evanston, IL 60201 USA4 Department of Biology, Loyola University Chicago, Chicago, IL 60626 USA2005 5 5 2005 5 30 30 23 12 2004 5 5 2005 Copyright © 2005 Yano et al; licensee BioMed Central Ltd.2005Yano 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 chromosomes of higher plants are littered with retrotransposons that, in many cases, constitute as much as 80% of plant genomes. Long terminal repeat retrotransposons have been especially successful colonizers of the chromosomes of higher plants and examinations of their function, evolution, and dispersal are essential to understanding the evolution of eukaryotic genomes. In soybean, several families of retrotransposons have been identified, including at least two that, by virtue of the presence of an envelope-like gene, may constitute endogenous retroviruses. However, most elements are highly degenerate and are often sequestered in regions of the genome that sequencing projects initially shun. In addition, finding potentially functional copies from genomic DNA is rare. This study provides a mechanism to surmount these issues to generate a consensus sequence that can then be functionally and phylogenetically evaluated.
Results
Diaspora is a multicopy member of the Ty3-gypsy-like family of LTR retrotransposons and comprises at least 0.5% of the soybean genome. Although the Diaspora family is highly degenerate, and with the exception of this report, is not represented in the Genbank nr database, a full-length consensus sequence was generated from short overlapping sequences using a combination of experimental and in silico methods. Diaspora is 11,737 bp in length and contains a single 1892-codon ORF that encodes a gag-pol polyprotein. Phylogenetic analysis indicates that it is closely related to Athila and Calypso retroelements from Arabidopsis and soybean, respectively. These in turn form the framework of an endogenous retrovirus lineage whose members possess an envelope-like gene. Diaspora appears to lack any trace of this coding region.
Conclusion
A combination of empirical sequencing and retrieval of unannotated Genome Survey Sequence database entries was successfully used to construct a full-length representative of the Diaspora family in Glycine max. Diaspora is presently the only fully characterized member of a lineage of putative plant endogenous retroviruses that contains virtually no trace of an extra coding region. The loss of an envelope-like coding domain suggests that non-infectious retrotransposons could swiftly evolve from infectious retroviruses, possibly by anomalous splicing of genomic RNA.
==== Body
Background
Eukaryotic genomes are littered with dozens to tens of thousands of copies of reverse transcriptase (RT)-based retroelements [1-3]. Among these are a diverse collection of elements characterized by long terminal repeats (LTR) that include the Ty1-copia-like and Ty3-gypsy-like retrotransposon families, endogenous retroviruses, and mammalian lentiviruses [4]. LTR retrotransposons have been especially successful colonizers of the chromosomes of higher plants where they constitute as much as 80% of these genomes [3,5-7]. In soybean, several families of LTR retrotransposons have been identified [8-10], including at least two that possess an env-like ORF and resemble mammalian endogenous retroviruses [10,11].
The evolutionary relationship between retrotransposons and retroviruses has been well established by phylogenetic tree constructions. However, the branches linking these groups are, not unexpectedly, long ones [4,10,12-15]. The major structural difference between retrotransposon and retrovirus genomes is the presence of an envelope gene (env) in the latter. Retroviral envelope proteins sponsor receptor binding, cell fusion, and particle budding, and contain transmembrane and coiled-coil domains[16]. While the de novo acquisition of an env-like coding region by transduction could conceivably occur in a single step, the functional evolution of such a coding domain might be expected to occur over considerable stretches of evolutionary time [15,17]. But could the loss of such a coding domain occur in a single step? This question is far from implausible, considering that all retroelement genomes are RNA transcripts and many are substrates for splicing reactions. A single event of anomalous packaging of an improperly spliced subgenomic RNA, followed by reverse transcription could lead to an env-less element in an evolutionary blink of an eye.
In the present study, the characterization of the soybean retrotransposon, Diaspora, provides evidence for a relatively rapid transition between enveloped retroelements and non-enveloped retrotransposons. Our phylogenetic analysis suggests that the Diaspora retrotransposon emerged from a lineage of plant endogenous retroviruses that possesses an env-like gene [10].
Diaspora was initially encountered in a genomic clone as a 5'and 3'-truncated copy nested between copies of another LTR retroelement (Laten, unpublished). Using both direct sequencing and in silico analysis, we generated a full-length consensus copy of Diaspora and confirmed 1) its membership in the Ty3-gypsy-like family of LTR retrotransposons and 2) its status as the only member of an endogenous retrovirus lineage lacking an env-like gene. The in silico procedure can be extended to construct consensus sequences for other repetitive DNA families from degenerate elements and from single-pass-read genome survey sequences, provided the copy numbers are sufficiently high and constitute a robust collection of overlapping sequences.
Results
AF095730 is related to gypsy group LTR retrotransposons
Sequencing of subclone pAMH3C [GenBank: U96295] initially led to the characterization of an env-like gene and the 3' LTR of the SIRE1 endogenous retrovirus belonging to the Ty1-copia group of retroelements [11]. The DNA adjacent to the SIRE1 LTR constituted an initially unidentified 544 bp ORF that gave no hits in BLASTn or BLASTx searches. However, when the sequence of the adjacent subclones, pAMH3G and pAMH3D, were addended to pAMH3C and assembled into a contig [Genbank: AF095730], a single, 1383-codon ORF with a nonsense mutation at position 2604, and frameshifts at 2813 and 3139 was generated (Laten and Das, unpublished). The frameshifts occurred in runs of six thymidines and five adenosines, respectively. When the frameshifts were adjusted and the conceptual translation was used to query the Genbank protein database, numerous high scoring hits to retrotransposon reverse transcriptases and integrases were obtained (Laten and Das, unpublished).
The large collection of sequences of reverse transcriptases and integrases that were retrieved, most as contiguous polyproteins, all belonged to the Ty3-gypsy group of LTR retrotransposons. While the BLASTp search identified AF095730 homology to numerous accessions from residue 135 to the carboxyl terminal, no sequences with similarity to the first 134 amino acids were found. The highest scoring hits were identified as Athila-like, related to the Ty3-gypsy group element from A. thaliana [18] that has subsequently been shown to be present in a wide range of plant genomes, including soybean, other dicots, and monocots [10]. We have named the new soybean element family Diaspora.
Diaspora is a multi-copy family
Because the Diaspora DNA in the original genomic clone was truncated at both ends, we initially probed the λFIXII genomic library for additional Diaspora copies. Hybridization detected a few thousand positive plaques, confirming the moderately high copy number of this family. DNAs from a random sample of twenty positive clones were amplified using primers derived from the ends of AF095730 (PDIA01F/PDIA02R and PDIA03F/PDIA04R). No clone produced amplicons with both primer pairs, suggesting that all copies of Diaspora in these clones were either 5'- and/or 3'- truncated, or polymorphic at the primer sites (data not shown). We inferred that the library would not readily yield full-length elements. Retrospectively, this finding would have been anticipated had the unusual length of Diaspora, approaching that of the average insert size in λFIXII, been known (see below). With the availability of BAC soybean genomic libraries with inserts in excess of 100 kb [19,20], the possibility of isolating full-length Diaspora copies became a virtual certainty and the λ clones were abandoned in favor of BACs. Filters containing microarrays of BAC clones derived from G. max cv. Forest [20] were probed for the presence of Diaspora. Hundreds of clones hybridized to a prot-rt probe (pAMH3D) and based on the number of hybridizing clones in the library, we estimated that Diaspora represents at least 0.5% of the G. max genome.
Diaspora family members are truncated and heterogeneous
DNA was recovered from twenty, randomly chosen BAC clones that hybridized to pAMH3D, and PCR-amplified using the primer pairs derived from the ends of AF095730 (PDIA01F/PDIA02R and PDIA03F/PDIA04R). Surprisingly, only five clones were amplified by both primer pairs, suggesting that many copies were either 5' or 3' truncated or markedly polymorphic. Truncation would be consistent with the characterization of disrupted and nested retrotransposons first reported in maize [21]. Of those containing both termini, none successfully served as templates for more than two additional PCR-amplifications using the complete set of AF095730-based primer pairs (PDIA5F through 13R). These findings suggested that Diaspora is a relatively heterogeneous family. This was confirmed by limited sequencing of λFIXII [GenBank: AF095730 and AY656632-AY656653] and BAC [GenBank: AY656654-AY656662] clones using the primers listed in Table 1. A total of 15,433 nucleotides were sequenced, of which 7293 were non-overlapping. It appeared that sequencing individual members of the Diaspora family directly from genomic clones would not lead to satisfactory descriptions of functional coding or regulatory regions, so an in silico strategy for these characteristics was pursued.
Table 1 Primers used for PCR amplification and DNA sequencing.
Oligomer Sequence Oligomer Sequence
PDIA01F AACCTCAACAGCAAAATCAACCA PDIA12R CACTTTGCGAGCTGTCCTTTGA
PDIA02R GAGGGCTGGACCATCTGAGGT PDIA13F TGCGGATTCACCCATTC
PDIA03F TGGGCACATCGGACTGCTTAC PDIA14R CCAAAGACAACCCGATAAGGAG
PDIA04R GACATGCCTTTCCAAAGACAACC PDIA15F TTCCTATCTCCTTCTTTGCTTT
PDIA05F GGCCCAAGCAGACCATACA PDIA16F TTGCCCCATTGATTGCTTG
PDIA06R TAAAAATCAACAGGGAAAATCAGT PDIA17R TTTCAAATCACAAAATGTCAAG
PDIA07F TGTCTCCGCATTGATTGGTAAA PDIA18R TGTAAGTCAGATGGATTGCCA
PDIA08R ATTGGCTGTCGGAGATAGGATAAA PDIA19R GCTCCAAGGTCCATCACGA
PDIA09F AAACCAGTAAGACAGCCACAGAGA PDIA20R GGACATCCTCATCAGGGTATTG
PDIA10R CAAGGACAGCCCCCAATG PDIA21F CATGGGTGCTTTGAGGGTAA
PDIA11F GAGGTGCGATCTTTTCTTGGTC
Diaspora sequences recovered by BLASTn queries
Prior to the initiation of plant genome sequencing projects, AF095730 was used to search Genbank for related sequences. At that time, BLASTp searches returned a sizeable collection of previously characterized pol polyproteins from Ty3-gypsy-like retrotransposons (Laten, unpublished). As more and more soybean BAC-end sequences were deposited in the GSS database [22]; J. Shultz, K. Meksem, J. Shetty, C. Town, H. Koo, J. Potter, K. Wakefield, H. Zhang, C. Wu and D. Lightfoot, unpublished] the growing robustness of our BLASTn results made it clear that Diaspora was a high copy-number retrotransposon and that the database hits derived exclusively from BAC ends might be assembled into a contiguous, full-length, consensus Diaspora sequence.
Genbank sequences retrieved using sequentially selected segments of AF095730 as queries were assembled into an expanding contig. The BAC-end sequences ranged from 400 to 900 nucleotides in length. Hits with bit scores ≥200 were added to the contig and this cutoff value generated a manageable collection of sequences. Sequences anchored to the ends of AF095730 were used to extend the consensus beyond the 5' and 3' ends of AF095730 and the assembly of an expanding contig was launched. Primers generated from these flanking regions (PDIA15F through 21F) were also used to amplify and sequence additional regions from two of the full-length Diaspora candidates in BAC clones.
Two hundred seven Genbank accessions, including thirty submissions from the present study, totaling 141,423 nucleotides were collected to generate the contig. To avoid bias, duplicate sequences from different accessions were purged from the alignment. There were only three positions for which a strict consensus nucleotide could not be assigned.
Fig. 1 is a histogram of the density distribution of sequences used to generate the contig. The average coverage across the length of the contig was 14.8 accessions, although inclusion of many additional sequences that met the scoring criterion in regions of high sequence conservation was not pursued. Because the initial soybean BAC libraries were created by EcoRI, HindIII, or BamHI digestion, the local robustness of the assembly was dependent on the density of these sites in Diaspora. Sub-families that lacked a particular cleavage site would be under-represented. In contrast, the assembly of regions far from these sites was made possible by sub-families with additional recognition sites for one of these enzymes.
Figure 1 Histogram of local densities of Genbank Accessions used to construct a Diaspora consensus contig. Recognition sites of restriction enzymes used to generate BAC libraries are indicated. Restriction sites in () are found in < 50% of sequences. The right LTR is not shown. The contig was assembled from the following Genbank Accessions: AF095730 (this study), AY656632-AY656656 (this study), AY656659-AY656662 (this study), AQ989187, AQ989208, AQ989232, AQ989271, AQ989295, AZ044709, AZ045083, AZ221405, AZ301361, AZ302029, AZ536637, AZ933330, AZ936131, BE611677, BH000863, BH000924, BH001187, BH023628, BH023632, BH023632, BH173556, BH405523, BH405626, BH405659, BH405669, BH610143, BH610157, BH610193, BH840834, BH854486, BH888573, BH897988, BH912698, BI974271, BU546431, CC062189, CC062259, CC062269, CC062279, CC062321, CC062333, CC062399, CC062412, CC062425, CC062501, CC062524, CC062576, CC062745, CC062865, CG811196, CG812831, CG813036, CG813244, CG813336, CG813336, CG813447, CG813495, CG813591, CG813669, CG813710, CG813854, CG813944, CG814001, CG814027, CG814297, CG814428, CG814537, CG814691, CG814705, CG814739, CG814773, CG814814, CG814837, CG814944, CG814960, CG815296, CG815349, CG815376, CG815566, CG815593, CG815931, CG815990, CG816077, CG816195, CG816437, CG816499, CG816820, CG816902, CG816924, CG816965, CG817175, CG817237, CG817248, CG817294, CG817426, CG817444, CG817647, CG817665, CG817749, CG817754, CG817777, CG817807, CG817873, CG817996, CG818405, CG818428, CG818443, CG818626, CG818673, CG818711, CG819087, CG819204, CG819222, CG819552, CG819604, CG819672, CG819766, CG819790, CG819813, CG819936, CG819977, CG820067, CG820103, CG820158, CG820299, CG820411, CG820560, CG820627, CG820654, CG820656, CG820670, CG820673, CG820702, CG820718, CG820816, CG820848, CG820850, CG820868, CG821026, CG821085, CG821093, CG821150, CG821179, CG821206, CG821219, CG821294, CG821311, CG821532, CG821597, CG821693, CG821710, CG821772, CG821963, CG822140, CG822195, CG822264, CG822361, CG822369, CG822426, CG822466, CG822466, CG822582, CG823113, CG823202, CG823294, CG823320, CG823499, CG823505, CG823511, CG823713, CG824266, CG824332, CG824372, CG824380, CG824407, CG824533, CG825062, CG825163, CG825591, CG825777, CG825811, CG825933, CG826013, CL867862, CL8811208, CL881708, CL882298, CL886562, CL891285, CL899081
Structure of Diaspora
The length of the Diaspora consensus is 11,737 bp (Fig. 2), far longer than all but a handful of retrotransposons. The exceptional length of Diaspora is due primarily to the unusual length of its LTRs and the long gap between the upstream LTR and the gag start codon (Fig. 2). Like nearly all other retroelements, the LTRs terminate in TG...CA. The element is characterized by a contiguous 1892-codon ORF whose conceptual translation yields a single gag-pol polyprotein (Fig. 3) characteristic of Ty3-gypsy-like retrotransposons [23]. Not surprisingly, the consensus contains neither nonsense codons nor frameshifts. This translated ORF possesses core domains for gag (CDD17379), reverse transcriptase (CDD16610) and integrase (CDD25582). There is also a CX2CX4HX4C zinc finger motif in gag and a conserved protease catalytic domain motif, AMLDLGAS (Fig. 3). Interestingly, the first thirty amino acids of the translated ORF are not similar to the amino termini of any gag proteins in Genbank. Similarities to several gag proteins begin at position 31. Translation of the other five reading frames yielded no lengthy ORFs nor any similarities to any sequences in BLASTP searches.
Figure 2 A. Consensus nucleotide sequence of Diaspora. LTR in red, PBS in green, ORF in blue, PPT in maroon. B. Structural organization of Diaspora. PBS: tRNA primer binding site; Gag: Gag core domain (CDD17379); Z: CCHC Zn finger domain; P: protease catalytic core; RT: reverse transcriptase core domain (CDD16610); Int: integrase core domain (CDD25582); PPT: polypurine tract. (⇨) ORF. Consensus restriction sites as in Fig. 2 H: HindIII; E: EcoRI; B: BamHI.
Figure 3 Conceptual translation of the Diaspora ORF. Teal: Gag core domain; blue: Zn finger domain; red: protease catalytic core; green: RT core domain; violet: integrase core domain
As in the Calypso group [10], the tRNA primer binding site (PBS) begins 5 bp beyond the 3' end of the LTR and is perfectly complementary to the 3' terminal 18 bases of tRNAAsp from Glycine max [24] (Fig. 4). At 873 bp, the distance between the LTR and the putative gag start codon is unusually long and not shared by related elements. This region contains no extended ORFs and neither BLASTn nor tBLASTn searches of the nr database retrieved significant hits.
Figure 4 Base pairing of PBS (upper case) from Diaspora and Calypso with the 3' end of tRNAAsp. LTR terminus underlined.
Five potential splice donor sites, all in the LTR between 1400 and 2200 bp upstream of the gag-pol ORF, were predicted with medium confidence and two potential acceptor sites flanked the start codon. Although without splicing, the 5'UTR of any Diaspora transcript would be exceptionally long, the biological relevance of these sites is not known, and there are no reported examples of introns upstream of gag for any retrovirus or LTR retrotransposon.
The pol stop codon is 128 bp upstream of the polypurine tract (PPT) that abuts the 3' LTR. Thus Diaspora contains no envelope-like coding sequence beyond pol unlike those reported for its closest relatives, including members of the Athila and Calypso families (Wright and Voytas, 2002), BAGY-2 (Vicient et al., 2001), and Cyclops-1 [8,10,13]. When this short region was used in BLASTn and tBLASTx searches, no additional sequences with significant probabilities were recovered. Interestingly, translation of this short region yields a strongly predicted transmembrane domain, although it is interrupted by two stop codons (data not shown).
The Diaspora LTR is 2524 bp in length (Fig. 2), making it one of the longest among retrotransposons and contributing to its unusual length. By comparison, the RIRE3 LTR is 2316 [25], BARE-1 is 1829 [26], Athila1-1 is 1539, and Cyclops-1 is 1504 [8]. Only the LTRs from Ogre and BAGY-1, at over 5,000 and 4200 bp, respectively [27,28], are longer.
The length of the Diaspora LTR made it impossible to construct unique 5' or 3' LTRs by the in silico method employed. In addition, the absence of a contiguous element prohibited characterization of target site duplications. However, we identified eight accessions from the database that contained the tRNA PBS and part of the adjacent upstream LTR. The longest of these extended 491 bp upstream of the PBS. Twenty-two sequences contained the PPT and part of the adjacent downstream LTR. The longest of these extended 596 bp into the LTR. Thus, the central 1437 bp could not be uniquely assigned to either LTR. As a consequence, the available LTR sequences were merged to generate a single, consensus LTR that was affixed to both Diaspora ends.
Thirteen LTR sequences were 5' junctions and sixteen were 3', based on the complete absence of sequence similarity beyond the 5' or 3' ends, respectively, of the aligned LTR sequences. When the flanking DNAs of these 29 sequences, were used in BLASTn searches to query the GSS database, all but three generated dozens of hits (data not shown), and thus constituted repetitive elements themselves. Of the repetitive flanking DNAs, 75% represented Diaspora insertions into the coding regions of other retrotransposons. In addition, insertions into the coding regions of transposons related to En/Spm and Tam3 were also found. The identity of the three low- or single-copy sequences could not be ascertained. The Diaspora family therefore appears to be embedded in retrotransposon and transposon-rich regions. We have made similar observations for the SIRE1 retroelement (unpublished). Searches focused on the region upstream of the PPT failed to uncover any Diaspora copies with additional DNA between pol and the PPT.
Among the sequences used to assemble the contig, non-coding regions contained a variety of short indels, especially in homonucleotide runs and dinucleotide repeats, presumably from replication slippage. The sequences in the GSS collection from which the consensus was built represented unedited submissions, and excluding single base indels that might have been the result of unedited miscalls, most of the indels in the ORF retained the correct reading frame. Eight accessions: BH023632, CG813336, CG820702, CG821179, CG822466, AY656639, AY656648, and AY656656, were chimeric and probably represented truncated copies. All but two of these sequences were within gag or the putative 5'UTR. Since chimeric sequences in GSS would invariably produce lower bit scores, they were generally excluded form the contig. Consequently, many of the slightly lower scoring sequences initially retrieved in our BLASTn search were also chimeric, but were not retained for the assembly and were not further characterized.
Diaspora is phylogenetically related to plant endogenous retroviruses
The conserved region of RT was translated and the region representing peptide domains 2 though 7 [29] was used in a BLASTp search to retrieve closely related accessions from Genbank. All of the sequences retrieved were from higher plants and their distribution among species reflected, to a large extent, the current progress of genome sequencing projects. The sequences were aligned (see Additional file 1) and a neighbor joining tree was generated (Fig. 5) and was rooted to the RT from gypsy.
Figure 5 Neighbor-joining phylogenetic tree using p-distances based on conserved RT domains 2 through 7 [29] of gypsy-like LTR retroelements from higher plants. The tree is rooted to gypsy. Bootstrap values from 1000 pseudo-replicates shown as percentages only at nodes with > 50% support. Vertical line indicates genus; key below. Named elements followed by Genbank Accession numbers; unnamed elements designated by Accession Number and, for translated nucleotide sequences, first nucleotide position. Wilma 634M12: AY494981 [51]; Wilma 426K20: AY146588 [51]; Wilma 107M9: AY368673 [51]; BAGY-2: AJ279072 [13]; Tmt1-1: AC146683 (115510-125622); Athila4-1: AC007209 [10]; Athila6-1: AF104920 [10]; Athila1-1: AB005248 [52]; Athila5-1: AF147260 [10]; AP005726: 133249; AC136972: 155124; Calypso2-1: AF186183 [10]; Calypso3-1: AF186185 [10]; Calypso5-1: AF186186 [10]; Calypso4-1: AF186185 [10]; Cyclops-1: AJ000639 [8]; AP004896: 78828; Tlc1-1: AP006432 (23839-35862); Tlc1-2: AP006350 (29612-19200); BBRE1: T12085; TfcII sr1: AF219199; TfcII sr25: AF219208; TfcII sr18: AF219207; cot8-6: AF378037 [10]; cot5-3: AF378037 [10]; cot8-7: AAL06412 [10]; Tpb1-1: AC149297 (90224-102138); syc2-3: AF378052 [10]; syc4-2: AF378053 [10]; DiasporaLc: AP007806: 43868; Tat4-1: AB005247 [44]; Cinful-1: AF049110 [45]; Grande1-4: X97604 [46]; RIRE2: AB030283 [47]; Reina: U69258 [48]; Cereba: AY040832 [49]; RIRE7: BAA89466 [50]; RIRE7-2: AL731604 (96205-102279); Dea1: T07863 [51]; del1-46: X13886 [52]; BAGY-1: Y14573 [27]; Tekay: AAL59229 [53]; RIRE3-2: AC123974 (48149-59938); RIRE3: AB014738 [50]; Retrosat2: AAM74400; Retrosat2-2: AL662955 (58578-70224); Gypsy: P10401 [54]. aTriticum; bHordeum; cMedicago; dArabidopsis; eOryza; fGlycine; gPisum; hLotus; iVicia; jFritillaria; kGossypium; lPopulus; mPlatanus; nZea; oAnanas; pLilium; qSorghum; rDrosophila
The tree resolves two major clades, designated A and B (Fig. 5). With respect to coding potential beyond pol, clade B members have none, and in all cases, the pol stop codon is closely followed by a PPT and the LTR. In contrast, with the exception of Diaspora, all members of clade A for which sequences downstream of pol are available contain a putative env-like pseudogene. The major structural difference between Diaspora and other members of this group is illustrated in Fig. 6. Clade A is further partitioned into sister clades AI and AII with 94% bootstrap support. Clade AI is further divided, with 100% bootstrap support, into AIa/b and AIc. The bifurcation of AIa and AIb in Fig. 5 is only weakly supported (40%) and may not be significant. Clades AIa and AIb are populated exclusively with env-containing members, including the Athila and Calypso families. The only full length members of AIc are Diaspora, DiasporaLc, and Tpb1-1. The other members of the AIc lineage are sequences from PCR-amplified rt fragments from sycamore and cotton [10], and three representative rt-containing genomic clones from Fritillaria. DNA downstream of the Fritillaria rt has not been characterized (C. Baysdorfer, personal communication). In contrast to Diaspora in G. max and its close relative in L. corniculatus, 1479 bp separate the pol stop codon from the LTR in Tpb1-1. While this region contains no identifiable or extended ORFs, there is a proximal 19-codon ORF whose conceptual translation is predicted with very high confidence to be a transmembrane domain (data not shown). There is also a 47-base polyA segment in the middle of this region, suggesting the interval contains an integrated cDNA. Thus, the AIIc lineage is not monophyletic for the absence of a long pol-LTR interval, but whether this region in Tpb1-1 represents a degenerate env cannot be determined.
Figure 6 Structural organization of Athila, Calypso, and Diaspora consensus elements.
In conclusion, the nesting of clade AIc within a much larger group of elements with an env-like gene suggests that at least Diaspora suffered a complete and nearly precise loss of a coding region, rather than failed to acquire one. To more exhaustively search for other members of this group, the RT from members of lineage AIc were used in tBLASTn searches. All hits, however, were already in the tree.
Discussion
The DNA sequence of the previously unreported Diaspora retrotransposon was created by a combination of experimental and in silico methods utilizing Glycine max sequences currently available as single-pass-read accessions in public databases. To date, the publicly available G. max sequence collections, including the NR and HTGS databases, contain no full-length copies of this element. Consensus sequences for transposons and retroelements have frequently been generated from alignments of multiple family members [30-35], but the construction of a full-length consensus sequence of a new element from large numbers of short overlapping fragments has not. While appropriate for early stage genome projects like that of soybean, resorting to such a strategy is neither required nor efficacious in genomes that have been extensively sequenced, like those of Arabidopsis, rice, Drosophila and humans.
Diaspora has a single uninterrupted ORF encoding gag, protease, RT, and integrase as a single polyprotein. Consensus assemblies for Athila and Calypso elements also contain a single ORF for these proteins [10]. While we cannot infer that functional copies of Diaspora still exist in the G. max genome, the density of the contig assembly and the presence of a strongly conserved consensus nucleotide at virtually every position of the assembly supports the argument that a reasonable facsimile of a past functional element is depicted.
The Diaspora family is also present in the Lotus corniculatus genome, where we discovered an apparently 5'truncated copy on a Phase I HTGS clone, AP007806. Excluding indels, the lotus sequence shares approximately 80% nucleotide identity with the Diaspora consensus sequence over a length of 7 kb. Most of the indels in the coding region are in-frame. Like Diaspora, the lotus element lacks an env-like region. With the exception of two 7-bp and one 15-bp indels, the short intervals between the pol termination codon and the LTR are 88% identical between the two. Additional truncated copies of Diaspora family members are present on ten other Phase I HTGS clones from L. corniculatus.
Diaspora is unusual in several respects. 1) It has unusually long LTRs. 2) At 873 bp, the distance between the LTR, which should contain the promoter and transcriptional start sites, and the gag start codon is far longer than every other characterized retroelement except one. And 3) Diaspora is the only characterized envelope-less member of a lineage of plant gypsy-like endogenous retroviruses.
The significance of the extended length of the LTRs found in BAGY-1 [28], Ogre [27], RIRE3 [25], and Diaspora are difficult to ascertain since virtually nothing is known about the biology of these elements. The same is true for the unusually long regions between the LTR and the gag start codon in Diaspora and Ogre, although in the case of Ogre, this region contains a 550-codon ORF whose conceptual translation yields a polypeptide of unknown function. Since transcriptional start sites are always found at the U3-R junction of the LTR, an exceptionally long 5'UTR would result unless splicing occurred. However, there have been no introns reported upstream of gag in any LTR retroelement, including Ogre [27], for which transcripts have been characterized,.
Unlike all other characterized members of an apparent plant endogenous retrovirus lineage [10], Diaspora lacks an envelope-like coding domain downstream of pol. Few members of this lineage contain functional gag-pol genes, based on the presence of nonsense and frameshift mutations, and none contain a functional env-like gene based on these same criteria. The only other fully sequenced member of clade AIIc, Tpb1-1, contains a strongly predicted TM just downstream of the pol stop codon, but is contaminated by an apparent retrogene, and the region could not be characterized as env-like based on amino acid similarity.
The hypothetical env-like proteins exhibit little primary sequence similarity, and only those found in Calypso and Cyclops-2, which share 29% amino acid identity, appear to be homologous [10,12]. Without significant sequence similarity, we are reluctant to speculate whether the predicted transmembrane domain for the translated 128 bp fragment between pol and the PPT in Diaspora reflects a vestige of an env-like gene.
The tree in Fig. 5 is similar to that generated by Wright and Voytas [10]. In their study, many members of the endogenous retrovirus lineage were derived from rt-delimited PCR amplicons, few of which are included in on our tree because the presence of an env-like region was not empirically determined. Our analysis, however, includes several additional full-length elements whose env-like status has been determined. We infer from this analysis that an ancestral Diaspora element suffered a deletion of this region. Whether Tpb1-1 suffered a similar fate but subsequently acquired a retrogene is open to speculation. Using this region to query Genbank in BLASTn and tBLASTx searches yielded no hits. While envelope capture by LTR retrotransposons has been credited with the creation of infectious retroviruses [15,17], only the env genes of invertebrate elements have been phylogenetically linked to unrelated viruses [14]. The failure to uncover an analogous linkage in retroviruses has been attributed, in part, to accelerated divergence promoted by host-induced immune responses that fuel positive selection for envelope variants [14].
Studies focused on envelope loss have not been reported, although phylogenetic relatedness between mammalian retroviruses and endogenous retroviruses with env pseudogenes is recognized [36]. Although the env genes of most human endogenous retrovirus (HERV) families are marked by frameshifts, nonsense mutations, and deletions [37], in only one family, HERV-L, are all vestiges of the env gene lost [37]. In the case of HERV-L, the region between pol and the LTR is occupied by a dUTPase coding domain [37]. Other members of the class III HERV clade that contains HERV-L, including HERV-S, contain env pseudogenes [36]. Interestingly, with a copy number of 575, the HERV-L family is second only to the HERV-H family in abundance [36].
Diaspora and DiasporaLc possess no trace of the env-like genes that are present in all members of clades AIa, AIb, and AII (Fig. 5). Whether other members of clade AIc, from cotton (Gossypium), sycamore (Platanus) and lily (Fritillaria), also lack an env region is not known, and the precise node within this clade that represents envelope loss cannot be assessed.
One explanation for an abrupt and complete loss of env is anomalous splicing of a genomic transcript containing gag-pol-env. Retroelement genomes are packaged as genomic RNA transcripts and retroviral transcripts destined for translation are often substrates for a complex pattern of splicing [38]. Alternatively, illegitimate recombination could lead to DNA loss and has been proposed as a major component of element elimination from plant genomes [39-41]. Many of the individual sequences that made up the contig contained short deletions not shared by others. This was especially true in non-coding regions (data not shown). Whatever the explanation, Diaspora appears to be an example of a retrotransposon that evolved from an endogenous retrovirus.
The nature of selective forces, if any, that might drive the loss of an env gene is open to speculation. Env genes, required for retroviral infectivity, are not thought to be required for retrotransposition, and it is possible that for some retroelements the gene or its protein product might attenuate the process, promoting selection for their inactivation, but with concomitant loss of infectivity. As noted above, in one of the largest families of HERV, the env gene has been replaced with a dUTPase. In plant genomes, however, the copy numbers of both putative endogenous retroviruses like SIRE1 [11], Calypso[10], and BAGY-2 [13] and retrotransposons like BARE-1 [42], Opie-2 [21], and Diaspora reach into the thousands. On the other hand, env genes in both mammalian and plant endogenous retroviruses are far more degenerate than those in pol, suggesting they are far less sensitive to purifying selection. The proliferation of one retroelement form or the other may be the result of random mutation and genetic drift. Nonetheless, retrotransposition, with or without an env gene, has been a far more successful long term reproductive strategy than retroviral infection.
Methods
DNA isolation, amplification, and sequencing
DNA containing the SIRE1 endogenous retrovirus was recovered from a λFIXII soybean genomic library (Stratagene) by standard plaque hybridization [43], and HindIII-digested fragments were sub-cloned into pSPORT1 (Life Technologies) as described [11]. DNA from three contiguous subclones, pAMH3C, pAMH3G, and pAMH3D were isolated and sequenced as described [11]. The junctions and contiguity of these subclones were confirmed by direct sequencing of the intact λFIXII genomic clone across the HindIII junctions. These sequences were previously deposited [Genbank: U96295 and AF095730]. A BLASTp search with the conceptual translation of AF095730 (see below) indicated that this accession contained the pol region of an uncharacterized retrotransposon. Several additional positive clones from this library were recovered and segments of the isolated DNAs were sequenced directly or amplified using Taq DNA Polymerase (Promega). For amplifications, reactions were preheated for 3 min. at 94°C, then 30 cycles were run at 94°C for 30 sec., 54°C for 30 sec., and 72°C for 1 to 2 min. Amplicons were spin column-purified (Qiagen) and sequenced as described [11]. Sequences were deposited [GenBank: AY656632-AY656653].
pAMH3D, compromising the protease and RT coding domains, was used to probe a soybean BAC library [20] (generously provided by K. Meksem) under moderate stringency [43] for the presence of sequences related to AF095730. Ten clones were chosen arbitrarily for amplification and sequencing. DNAs from BAC clones were recovered using Procipitate (Ligochem) and selected regions were amplified using Taq DNA Polymerase (Promega). After preheating reactions for 3 min. at 94°C, 30 cycles were run at 94°C for 30 sec., 54°C for 30 sec., and 72°C for 1 min. Regions within pAMH3D were first PCR-amplified using primer pairs PDIA01F-02R, PDIA03F-04R, PDIA05F-06R, PDIA07F-08R, PDIA09F-10R, PDIA11F-12R, and PDIA13F-14R (see Table 1). The amplicons were purified on Qiagen spin columns and sequenced directly without cloning as described [11]. Sequences for regions beyond the ends of AF095730 were generated directly from BAC DNA using outward facing primers (PDIA02R and 03F) derived from the termini of AF095730, followed by additional outward extensions with primers PDIA16F, 17R, 18R, 19R, 20R, 21F (Table 1). BAC clone sequences have been deposited [GenBank: AY656654-AY656662].
In silico methods
Selected regions of AF095730 and their conceptual translations were used to query all relevant Genbank databases, including nr, Genome Sequence Survey (GSS) and Expressed Sequence Tag (EST), with BLASTn and BLASTp searches [44,45]. Conserved protein domains were identified with CDD [46].
For assembly of the consensus nucleotide sequence, Glycine max accessions from the GSS and EST databases with bit scores greater than 200 (E values < 10-52) were added to the consensus construct. These criteria generally reflect >90% DNA sequence identity over at least 200 bp of overlap. The nr nucleotide database contained no significant hits. New additions to the ends of the expanding consensus were used to re-query the databases until the LTR redundancy was recognized and the contig formed a circle. Contigs were assembled using the Seqman program from Lasergene 5 (DNAStar). To locate the LTR, direct repeats greater than 25 bp were first identified using Lasergene GeneQuest (DNAStar) and the termini of the LTRs were confirmed by manual inspection, as were other non-coding features of the sequence. Because LTR junction sequences at both termini contained either internal element DNA or external flanking DNAs, these were carefully examined for consensus DNA (internal) or unique DNAs (external). External DNAs were trimmed from the contig. Potential splice junctions were evaluated using GeneSplicer [47] and NetGene2 [48]. Transmembrane domains were predicted using TMPred [49]
The pol region of the conceptually translated consensus sequence was used to query the Genbank protein database for related sequences. A ClustalW alignment (see Additional file 1) was generated from a contiguous region of RT representing conserved domains two through seven [29] using Lasergene 5 (DNAStar), and a neighbor joining tree using p distances with 1000 bootstrap pseudo-replicates was constructed using MEGA2 [50].
Abbreviations
LTR: long terminal repeat; RT: reverse transcriptase; prot: protease; env: envelope; PBS: tRNA primer binding site; PPT: polypurine tract; GSS: genome survey sequence; HTGS: high throughput genomic sequence
Authors' contributions
SY identified, recovered, amplified, and sequenced DNA from BAC library clones and sequenced DNA from the λ library clones. PB identified and recovered DNA from the λ library clones. HL isolated and sequenced DNA from λ library sub-clones and performed all the in silico and phylogenetic analyses. HL prepared the manuscript for review by the authors and all authors approved the final draft.
Supplementary Material
Additional File 1
ClustalW alignment of conserved reverse transcriptase domains for selected plant Ty3-gypsy family retroelements. Amino acid sequence alignments generated by ClustalW using the Megalign program from Lasergene 5 were imported into MEGA2 for phylogenetic analysis.
Click here for file
Acknowledgements
Supported in part with Loyola Mulcahy Undergraduate Research Fellowships to STY and PB. Thanks to H. Mears, J. Smith, and J. Damergis for providing extra hands and spiritual support, and to K. Meksem, A. Jamai, and J. Shultz for BAC filters. This work was supported in part by U.S. Dept. of Defense Advanced Research Projects Grant N66001-03-1-8941.
==== Refs
Kazazian HH Mobile elements: Drivers of genome evolution Science 2004 303 1626 1632 15016989 10.1126/science.1089670
Deininger PL Roy-Engel AM Craig NL, Craigie R, Gellert M and Lambowitz AM Mobile Elements in Animal and Plant Genomes Mobile DNA II 2002 Washington, D.C., ASM Press 1074 1092
Feschotte C Jiang N Wessler SR Plant transposable elements: Where genetics meets genomics Nat Rev Genet 2002 3 329 341 11988759 10.1038/nrg793
Eickbush TH Malik HS Craig NL, Craigie R, Gellert M and Lambowitz AM Origins and Evolution of Retrotransposons Mobile DNA II 2002 Washington, D.C., ASM Press 1111 1144
Gao L McCarthy EM Ganko EW McDonald JF Evolutionary history of Oryza sativa LTR retrotransposons: a preliminary survey of the rice genome sequences BMC Genomics 2004 5 18 15040813 10.1186/1471-2164-5-18
Young ND Mudge J Ellis TH Legume genomes: more than peas in a pod. Curr Opin Plant Biol 2003 6 199 204 12667879 10.1016/S1369-5266(03)00006-2
SanMiguel PJ Ramakrishna W Bennetzen JL Busso CS Dubcovsky J Transposable elements, genes and recombination in a 215-kb contig from wheat chromosome 5A(m) Funct Integr Genomics 2002 2 70 80 12021852 10.1007/s10142-002-0056-4
Chavanne F Zhang DX Liaud MF Cerff R Structure and evolution of Cyclops: a novel giant retrotransposon of the Ty3/Gypsy family highly amplified in pea and other legume species Plant Mol Biol 1998 37 363 375 9617807 10.1023/A:1005969626142
Bhattacharyya MK Gonzales RA Kraft M Buzzell RI A copia-like retrotransposon Tgmr closely linked to the Rps1-k allele that confers race-specific resistance of soybean to Phytophthora sojae Plant Mol Biol 1997 34 255 264 9207841 10.1023/A:1005851623493
Wright DA Voytas DF Athila4 of Arabidopsis and Calypso of soybean define a lineage of endogenous plant retroviruses Genome Res 2002 12 122 131 11779837 10.1101/gr.196001
Laten HM Majumdar A Gaucher EA SIRE-1, a copia/Ty1-like retroelement from soybean, encodes a retroviral envelope-like protein Proc Natl Acad Sci U S A 1998 95 6897 6902 9618510 10.1073/pnas.95.12.6897
Peterson-Burch BD Wright DA Laten HM Voytas DF Retroviruses in plants? Trends Genet 2000 16 151 152 10729827 10.1016/S0168-9525(00)01981-8
Vicient CM Kalendar R Schulman AH Envelope-class retrovirus-like elements are widespread, transcribed and spliced, and insertionally polymorphic in plants Genome Res 2001 11 2041 2049 11731494 10.1101/gr.193301
Laten HM Phylogenetic evidence for Ty1-copia-like endogenous retroviruses in plant genomes Genetica 1999 107 87 93 10952201 10.1023/A:1003901009861
Malik HS Henikoff S Eickbush TH Poised for contagion: Evolutionary origins of the infectious abilities of invertebrate retroviruses Genome Res 2000 10 1307 1318 10984449 10.1101/gr.145000
Gallo SA Finnegan CM Viard M Raviv Y Dimitrov A Rawat SS Puri A Durell S Blumenthal R The HIV Env-mediated fusion reaction Biochim Biophys Acta - Biomembranes 2003 1614 36 50 10.1016/S0005-2736(03)00161-5
Kim FJ Battini JL Manel N Sitbon M Emergence of vertebrate retroviruses and envelope capture Virology 2004 318 183 191 14972546 10.1016/j.virol.2003.09.026
Pelissier T Tutois S Deragon JM Tourmente S Genestier S Picard G Athila, a new retroelement from Arabidopsis thaliana Plant Mol Biol 1995 29 441 452 8534844 10.1007/BF00020976
Marek LF Shoemaker RC Bac contig development by fingerprint analysis in soybean Genome 1997 40 420 427
Meksem K Zobrist K Ruben E Hyten D Quanzhou T Zhang HB Lightfoot DA Two large-insert soybean genomic libraries constructed in a binary vector: applications in chromosome walking and genome wide physical mapping Theor Appl Genet 2000 101 747 755 10.1007/s001220051540
SanMiguel P Tikhonov A Jin YK Motchoulskaia N Zakharov D Melake-Berhan A Springer PS Edwards KJ Lee M Avramova Z Bennetzen JL Nested retrotransposons in the intergenic regions of the maize genome Science 1996 274 765 768 8864112 10.1126/science.274.5288.765
Marek LF Mudge J Darnielle L Grant D Hanson N Paz M Yan HH Denny R Larson K Foster-Hartnett D Cooper A Danesh D Larsen D Schmidt T Staggs R Crow JA Retzel E Young ND Shoemaker RC Soybean genomic survey: BAC-end sequences near RFLP and SSR markers Genome 2001 44 572 581 11550890 10.1139/gen-44-4-572
Kumar A Bennetzen JL Plant retrotransposons Annu Rev Genet 1999 33 479 532 10690416 10.1146/annurev.genet.33.1.479
Waldron C Wills N Gesteland RF Plant tRNA genes: putative soybean genes for tRNAasp and tRNAmet J Mol Appl Genet 1985 3 7 17 4040149
Kumekawa N Ohtsubo H Horiuchi T Ohtsubo E Identification and characterization of novel retrotransposons of the gypsy type in rice Mol Gen Genet 1999 260 593 602 9928939 10.1007/s004380050933
Manninen I Schulman AH BARE-1, a copia-like retroelement in barley (Hordeum vulgare L.) Plant Mol Biol 1993 22 829 846 7689350 10.1007/BF00027369
Neumann P Pozarkova D Macas J Highly abundant pea LTR retrotransposon Ogre is constitutively transcribed and partially spliced Plant Mol Biol 2003 53 399 410 14750527 10.1023/B:PLAN.0000006945.77043.ce
Panstruga R Buschges R Piffanelli P Schulze-Lefert P A contiguous 60 kb genomic stretch from barley reveals molecular evidence for gene islands in a monocot genome Nucleic Acids Res 1998 26 1056 1062 9461468 10.1093/nar/26.4.1056
Xiong Y Eickbush TH Origin and evolution of retroelements based upon their reverse transcriptase sequences EMBO J 1990 9 3353 3362 1698615
Adey NB Tollefsbol TO Sparks AB Edgell MH Hutchison CAIII Molecular resurrection of an extinct ancestral promoter for mouse L1 Proc Natl Acad Sci U S A 1994 91 1569 1573 8108446
Ivics Z Hackett PB Plasterk RH Izsvak Z Molecular reconstruction of Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in human cells Cell 1997 91 501 510 9390559 10.1016/S0092-8674(00)80436-5
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
Kapitonov VV Jurka J MER53, a non-autonomous DNA transposon associated with a variety of functionally related defense genes in the human genome DNA Seq 1998 8 277 288 10993599
Robertson HM Walden KK Bmmar6, a second mori subfamily mariner transposon from the silkworm moth Bombyx mori Insect Mol Biol 2003 12 167 171 12653938 10.1046/j.1365-2583.2003.00398.x
Robertson HM Members of the pogo superfamily of DNA-mediated transposons in the human genome Mol Gen Genet 1996 252 761 766 8917322 10.1007/s004380050288
Gifford R Tristem M The evolution, distribution and diversity of endogenous retroviruses Virus Genes 2003 26 291 316 12876457 10.1023/A:1024455415443
Tristem M Identification and characterization of novel human endogenous retrovirus families by phylogenetic screening of the Human Genome Mapping Project database J Virology 2000 74 3715 3730 10729147 10.1128/JVI.74.8.3715-3730.2000
Jacquenet S Mereau A Bilodeau PS Damier L Stoltzfus CM Branlant C A second exon splicing silencer within human immunodeficiency virus type 1 tat exon 2 represses splicing of Tat mRNA and binds protein hnRNP H J Biol Chem 2001 276 40464 40475 11526107 10.1074/jbc.M104070200
Wicker T Yahiaoui N Guyot R Schlagenhauf E Liu ZD Dubcovsky J Keller B Rapid genome divergence at orthologous low molecular weight glutenin loci of the A and A(m) genomes of wheat Plant Cell 2003 15 1186 1197 12724543 10.1105/tpc.011023
Devos KM Brown JK Bennetzen JL Genome size reduction through illegitimate recombination counteracts genome expansion in Arabidopsis Genome Res 2002 12 1075 1079 12097344 10.1101/gr.132102
Ma JX Devos KM Bennetzen JL Analyses of LTR-retrotransposon structures reveal recent and rapid genomic DNA loss in rice Genome Research 2004 14 860 869 15078861 10.1101/gr.1466204
Suoniemi A Anamthawat-Jonsson K Arna T Schulman AH Retrotransposon BARE-1 is a major, dispersed component of the barley (Hordeum vulgare L.) genome Plant Mol Biol 1996 30 1321 1329 8704140 10.1007/BF00019563
Sambrook J Fritsch EF Maniatis T Molecular Cloning 1989 Cold Spring Harbor, NY, Cold Spring Harbor Laboratory Press
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
Altschul SF Madden TL Schaffer AA Zhang JH 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
Marchler-Bauer A Anderson JB DeWeese-Scott C Fedorova ND Geer LY He SQ Hurwitz DI Jackson JD Jacobs AR Lanczycki CJ Liebert CA Liu CL Madej T Marchler GH Mazumder R Nikolskaya AN Panchenko AR Rao BS Shoemaker BA Simonyan V Song JS Thiessen PA Vasudevan S Wang YL Yamashita RA CDD: a curated Entrez database of conserved domain alignments Nucleic Acids Res 2003 31 383 387 12520028 10.1093/nar/gkg087
Pertea M Lin X Salzberg SL GeneSplicer: a new computational method for splice site prediction Nucleic Acids Res 2001 29 1185 1190 11222768 10.1093/nar/29.5.1185
Hebsgaard SM Korning PG Tolstrup N Engelbrecht J Rouze P Brunak S Splice site prediction in Arabidopsis thaliana DNA by combining local and global sequence information Nucleic Acids Res 1996 24 3439 3452 8811101 10.1093/nar/24.17.3439
Hofman K Stoffel W TMbase - A database of membrane spanning protein segments Biol Chem Hoppe Seyler 1993 374 166
Kumar S Tamura K Jakobsen IB Nei M MEGA2: molecular evolutionary genetics analysis software Bioinformatics 2001 17 1244 1245 11751241 10.1093/bioinformatics/17.12.1244
Gu YQ Coleman-Derr D Kong X Anderson OD Rapid Genome Evolution Revealed by Comparative Sequence Analysis of Orthologous Regions from Four Triticeae Genomes Plant Physiol 2004 135 459 470 15122014 10.1104/pp.103.038083
Wright DA Voytas DF Potential retroviruses in plants: Tat1 is related to a group of Arabidopsis thaliana Ty3/gypsy retrotransposons that encode envelope- like proteins Genetics 1998 149 703 715 9611185
| 15876351 | PMC1142308 | CC BY | 2021-01-04 16:37:17 | no | BMC Evol Biol. 2005 May 5; 5:30 | utf-8 | BMC Evol Biol | 2,005 | 10.1186/1471-2148-5-30 | oa_comm |
==== Front
BMC Emerg MedBMC Emergency Medicine1471-227XBioMed Central London 1471-227X-5-21588246210.1186/1471-227X-5-2Research ArticleFailure of a patient-centered intervention to substantially increase the identification and referral for-treatment of ambulatory emergency department patients with occult psychiatric conditions: a randomized trial [ISRCTN61514736] Schriger David L [email protected] Patrick S [email protected] Wais A [email protected] Carol A [email protected] University of California, Los Angeles Emergency Medicine Center, Los Angeles, CA, USA2 University of California, Los Angeles School of Medicine, Los Angeles, CA, USA3 Department of Clinical Social Work, University of California, Los Angeles Medical Center, Los Angeles, CA, USA2005 9 5 2005 5 2 2 29 12 2004 9 5 2005 Copyright © 2005 Schriger et al; licensee BioMed Central Ltd.2005Schriger 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 previously demonstrated that a computerized psychiatric screening interview (the PRIME-MD) can be used in the Emergency Department (ED) waiting room to identify patients with mental illness. In that trial, however, informing the ED physician of the PRIME-MD results did not increase the frequency of psychiatric diagnosis, consultation or referral. We conducted this study to determine whether telling the patient and physician the PRIME-MD result would result in the majority of PRIME-MD-diagnosed patients being directed toward treatment for their mental illness.
Methods
In this single-site RCT, consenting patients with non-specific somatic chief complaints (e.g., fatigue, back pain, etc.) completed the computerized PRIME-MD in the waiting room and were randomly assigned to one of three groups: patient and physician told PRIME-MD results, patient told PRIME-MD results, and neither told PRIME-MD results.
The main outcome measure was the percentage of patients with a PRIME-MD diagnosis who received a psychiatric consultation or referral from the ED.
Results
183 (5% of all ED patients) were approached. 123 eligible patients consented to participate, completed the PRIME-MD and were randomized. 95 patients had outcomes recorded. 51 (54%) had a PRIME-MD diagnosis and 8 (16%) of them were given a psychiatric consultation or referral in the ED. While the frequency of consultation or referral increased as the intervention's intensity increased (tell neither = 11% (1/9), tell patient 15% (3/20), tell patient and physician 18% (4/22)), no group came close to the 50% threshold we sought. For this reason, we stopped the trial after an interim analysis.
Conclusion
Patients willingly completed the PRIME-MD and 54% had a PRIME-MD diagnosis. Unfortunately, at our institution, informing the patient (and physician) of the PRIME-MD results infrequently led to the patient being directed toward care for their psychiatric condition.
==== Body
Background
There is a higher prevalence of psychiatric conditions in patients presenting to emergency departments (ED)s with non-emergent complaints than in the general population [1-4]. On average, patients with untreated psychiatric illness have more frequent ED visits and use more health care services than those in the general public [5,6]. Studies in a variety of ED settings have documented that over 40% of ambulatory patients have underlying psychiatric conditions [1-4]. Since there are highly effective treatments for many psychiatric conditions, these patients are generally expected to achieve better health outcomes if their psychiatric condition is diagnosed and treated, especially when their somatic complaints (weak and dizzy, back pain, etc.) are not amenable to effective treatment. A collateral benefit would be a reduction in societal health care costs [7].
PRIME-MD is a screening tool for psychiatric conditions that uses closed ended questions to make DSM-IV diagnoses [8]. Its diagnostic validity has been established in a number of ambulatory care settings [9-12]. We have demonstrated that the computer version of the PRIME-MD can be used in the ED waiting room to identify patients with underlying psychiatric conditions that might be causing or exacerbating their somatic presenting complaints [4]. Unfortunately, in that study, the emergency physicians ignored the patient's PRIME-MD diagnoses and neither diagnosed nor treated the patient's mental health disorder. The same phenomenon has been observed in primary care [13].
In this study, we conducted focus groups to determine why our first trial failed and then conducted a trial using the stronger intervention of informing both the patient and the physician of the computer's findings. Our goal was to determine whether this intervention would result in the majority of patients with a PRIME-MD diagnosis being referred for evaluation and treatment of their psychiatric condition. By empowering the patient to act as his or her own advocate, we hoped to overcome whatever factors deter physicians from exploring these diagnostic possibilities with the patient. We designed a 3-limb randomized trial that included a control group, a group in which only the patient was informed of the PRIME-MD results and a group in which both patient and physician were informed of the results.
Methods
Study design and setting
Focus groups – An experienced facilitator used a set of open ended probes to conduct two 90 minute focus groups to explore why physicians might be reluctant to pursue psychiatric diagnosis, consultation or referral in patients given a psychiatric diagnosis by PRIME-MD. One group included 6 randomly selected EM residents, the other 6 EM faculty. Two observers took notes and identified main themes and points of disagreement. The results informed the development of the intervention and study materials.
Trial – This randomized, controlled clinical trial was conducted at the University of California Los Angeles Emergency Department, a teaching hospital and Level I trauma center. The annual census is 44,000. Study subjects included emergency and internal medicine house staff, emergency medicine faculty, and enrolled patients, all of whom were consented. The study was approved by the UCLA IRB.
Selection of participants
Patients age 18 or older presenting to the ED between the hours of 10 a.m. and 9 p.m. on most (84%) weekdays from March to September 2002 were recruited for study participation. A trained research assistant, stationed at the triage desk, listened to each intake interview and identified adults with diffuse somatic complaints (e.g., vague head, abdominal, back or body pain of non-acute onset; generalized weakness; "don't feel well") that did not seem to mandate emergency care or did not coincide with physical findings (e.g. complaining of rash but no rash visible). Patients arriving via ambulance, patients in extremis, and patients who indicated they were not comfortable reading English were excluded; as were patients whose symptoms suggested psychiatric illness (e.g. hearing voices, suicidal ideation, anxiety...), patients with recent substance abuse, and patients who already participated in the trial.
Intervention
Consenting patients were asked to complete the self-administered, PRIME-MD computerized psychiatric questionnaire (version 1.2, Pfizer, Inc, New York, NY) in the ED waiting room prior to seeing the physician. PRIME-MD poses a series of questions to screen patients in 7 psychiatric domains (mood, anxiety, alcohol abuse, eating disorder, obsessive-compulsive disorder, phobia, and somatization). Positive responses to these screening questions trigger additional questions to confirm or reject particular diagnoses within each domain. When the session is complete, the program prints a grid that indicates the presence or absence of 20 specific diagnoses within the 7 domains.
Patients were seated at a computer secluded from the rest of the waiting room. They answered questions using a Fastpoint light pen (Fastpoint Technologies, Stanton, CA). A research associate was present to assist with technical issues, time the session, and record any difficulties with the hardware or software. Upon conclusion of the PRIME-MD interview, but before results were known, the randomization software, using the random number function in STATA 6.0, assigned the patient to one of three groups: results given to patient and doctor (40%), results given only to patient (40%), or results given to neither (20%). Patients in the first two groups were given a packet with their PRIME-MD results, an explanation of these results, a glossary of terms, and a cover letter encouraging them to share their PRIME-MD results with the treating ED physician. All patients, regardless of randomization group, were provided a written invitation to speak with a psychiatric social worker at the conclusion of their ED visit All patients with PRIME-MD diagnoses "major depression" or "r/o bipolar disorder" were assessed for suicidality by the psychiatric social worker prior to discharge.
The physicians caring for patients in the "tell both" group were provided PRIME-MD results through identical pre-printed Post-it™ notes affixed over the parts of the medical record where the resident and the attending write their notes. Each Post-it™ note indicated which PRIME-MD domains were positive (or that the patient had no PRIME-MD diagnoses) and referred the physician to additional materials attached to the chart. These materials included a one-page cover sheet that introduced the physician to PRIME-MD, reviewed the evidence of its validity, and listed the patient's PRIME-MD diagnoses. As part of the study we produced a list of low-cost and no-cost psychiatric care options that could be offered to patients as part of their after care instructions. Unfortunately, many of the agencies on the list have many-month-long waiting lists for patients who do not have an acute psychiatric issue. Apart from these interventions, there was no attempt to alter usual care.
Methods of measurement
The primary outcome was whether each patient diagnosed by the PRIME-MD software left the ED having had a psychiatric consultation or with a referral for further evaluation and treatment of their psychiatric condition. Any indication that the patient was referred to a health care provider for help with a mental health issue, including a statement that the low-cost no-cost sheet had been provided was considered evidence of referral. Secondary outcomes were: whether the ED physician made a psychiatric diagnosis and whether the patient actually received follow-up for the psychiatric condition. This information was gleaned from standardized review of the medical record and after care instructions, and follow-up patient telephone interview which took place 2 to 4 weeks after the visit. We questioned the treating resident just after the patient had left the ED to ascertain whether she was aware of the PRIME-MD results and how she had acted upon them. Patient and physician interview procedures and results can be found in the Appendix.
Data collection and processing
Two investigators independently reviewed the chief complaints of all enrolled patients (presented independently of all other data) and excluded patients who had been inappropriately entered into the study. Trained research associates abstracted the patient's demographics, chief complaint, diagnoses, psychiatric consultations and psychiatric referrals from the ED medical record to a standardized form. Interrater reliability was assessed on a 10% sample of charts. Abstractor's were blinded to the patients' randomization status and PRIME-MD results. STATA 8.0 was used for data verification, database management, and statistical analysis. To maintain anonymity and confidentiality, physician and patient identifiers were dropped during database creation and replaced with randomly assigned numbers.
Primary data analysis
We designed this study to estimate the treatment effect with reasonable precision, not to perform formal hypothesis tests [14]. We took a Bayesian approach to this trial and incorporated data from our previous study and other reasonable priors in the analysis [15]. We used beta distributions to model all priors and likelihoods. For the control limb we set a prior distribution of beta (1.2, 15.8) which has a mean of 7% and a 95% credible interval of 0% to 23%. This prior is based on the control limb or our previous trial (7% (3 of 45) successes) but was widened to account for potential differences between the trials. It's information content is equivalent to a 17 person study. For the two other limbs we used beta (0.3, 1.2) distributions which have a mean of 20% and 95% credible interval of 0% to 87%. This wide interval reflected our uncertainty regarding the effect of the intervention and has information content equivalent to a study with a sample size of 1.5 subjects. All Bayesian calculations were performed in FAST*PRO and STATA 8.0 [16].
Our determination of sample size was guided by Bayesian estimations and traditional frequentist calculations. We decided a priori that the intervention was only worth doing if it resulted in more than 50% of subjects with a PRIME-MD diagnosis being offered referral. Our reasoning was that given the myriad of competing demands in the ED setting this particular intervention had to have a substantial impact (not just an incremental improvement over the status quo) in order to justify the cost and effort expended. As mentioned above, the control group of our previous trial had 3 of 45 patients with a PRIME-MD diagnosis receive psychiatric consultation or referral. A frequentist two limb trial would require 21 patients per group (80% power for difference between 50% and 7%, alpha .05). We sought to behav 80 patients in each active limb since we conservatively expected that at least 1/3rd of enrolled subjects would have a PRIME-MD diagnosis (42% did in our first study). Since the prior distribution for the control limb was considerably narrower than for the other limbs, we decided to randomize half that number of patients to this limb so that the posterior distributions of all limbs would be of similar precision.[15]. To conserve resources, we planned an interim analysis once we enrolled 40 patients in each active limb to determine whether the study had any chance of producing a clinically important result. Data analysts were blinded to the identity of the 3 experimental groups and the randomization proportions.
Results
Focus groups
There was general uniformity of opinion on the following themes: residents do not feel adequately trained to detect and treat occult psychiatric illness, attending physicians were concerned that focusing on psychiatric issues may cause the housestaff (but not them) to overlook somatic illness, and screening and diagnosis was pointless in a system that lacked any viable means for providing follow-up or treatment. No one questioned the validity of the PRIME-MD. There was heterogeneity of opinion regarding whether the ED was the proper place to screen and diagnose psychiatric conditions even if adequate follow-up was available. Some ED residents and attendings candidly revealed that they "did not go into emergency medicine to make non-emergent psychiatric diagnoses."
Randomized trial
At the time of the interim analysis and the stopping of the study, 4,054 patients had been triaged. 183 were eligible and 127 were randomized (Figure 1). 14 patients left prior to being seen, 4 did not complete their PRIME-MD session, and 11 patients were excluded after randomization when it was determined they had an unappreciated psychiatric complaint (e.g. "I haven't slept in 3 nights"). 95 patients were included in the analysis: 32 in the tell-both limb, 40 in the tell-patient limb and 23 in the tell-neither control limb (Table 1). Frequent chief complaints were musculoskeletal pain or minor trauma (23%), abdominal pain +/- nausea +/- vomiting (23%), weak +/- dizzy (12%), and headache (9%). No resident or attending saw a disproportionate number of patients (Table 1).
Figure 1 This graphic shows how the 95 subjects tallied in the final analysis were culled from the 4,054 patients seen in the ED during study hours. Patients were randomized to three groups: tell neither (PRIME-MD results not shared with the patient or physician – the control group), tell patient (PRIME-MD results shared with the patient), and tell both (PRIME-MD results shared with patient and physician). In the lowest level of boxes, the entry "ER MD Dx or Tx" indicates how many patients diagnosed by PRIME-MD were diagnosed, treated or referred by the emergency physician.
Table 1 Physician and patient information by study limb
Study Limb
Tell Neither Tell Patient Tell Both Total
Patient Characteristics
No. of patients 23 40 32 95
Age (y) mean (SD) 48(15) 46(14) 48(15) 47(15)
Female, No. (%) 12(53) 17(43) 19(60) 48(51)
Physician Characteristics
No. of Attendings 12 17 15 20
No. of Residents 19 30 30 51
No. of residents seeing:
1 patient 18 22 25 31
2 patients 1 4 3 9
3 patients 0 2 0 7
≥ 4 patients 0 0 0 4
The PRIME-MD session took between 4 and 8 minutes for half the patients (range 2–35 minutes). PRIME-MD made one or more diagnoses in 51 of the 95 (54%) patients (Table 2). Mood and anxiety disorders predominated. 24 patients were diagnosed with major depression and 4 with r/o bipolar disorder. The psychiatric social worker identified 2 patients in this group who were suicidal. The physicians of both of these patients were informed and appropriate assessment and treatment was instituted.
Table 2 PRIME-MD diagnoses by study limb
Study Limb
Tell Neither Tell Patient Tell both Total
No. of patients 23 40 32 95
Patients with
Any PRIME-MD diagnostic domain (%) 9(39) 20(50) 22(69) 51(54)
1 PRIME-MD diagnostic domain 3(13) 8(20) 10(31) 21(22)
2 PRIME-MD diagnostic domains 2(9) 7(18) 5(16) 14(14)
3 PRIME-MD diagnostic domains 4(17) 1(3) 5(16) 10(11)
≥ 4 PRIME-MD diagnostic domains 0(0) 4(10) 2(6) 6(6)
Any mood diagnosis 7(30) 12(30) 17(53) 36(37)
Any anxiety diagnosis 4(17) 14(35) 10(31) 28(29)
Any OCD diagnosis 5(22) 6(15) 7(22) 18(19)
Any alcohol/dependence 1(4) 5(13) 6(19) 12(13)
Any phobia diagnosis 2(9) 5(13) 2(6) 9(9)
Any eating disorder 0(0) 0(0) 2(6) 2(2)
Interrater agreement for the chart abstraction exceeded 90% for all items. 16% of patients diagnosed by the PRIME-MD were given a psychiatric consultation or referral in the ED (Table 3). While there was a suggestion that the intervention increased the likelihood that patients would be identified and offered treatment (control 11%, tell-patient 15%, tell-both 18%), no group approached the 50% level that we sought (Table 3 and Figure 2).
Table 3 Psychiatric diagnosis, consultation and referral in patients with a PRIME-MD diagnosis, by study limb (N (%))
Study Limb
Tell Neither Tell Patient Tell Both
Patients with a PRIME-MD diagnosis and a: 9 20 22
Physician psychiatric diagnosis 0 (0%) 2 (10%) 4 (18%)
Physician psychiatric consultation or referral 1 (11%) 3 (15%) 4 (18%)
Physician psychiatric diagnosis, consultation, or referral 1 (11%) 3 (15%) 4 (18%)
Frequentist 95% CI 0%, 48% 3%, 38% 5%, 40%
Bayesian Analysis of these data* mean
95% Credible Interval
Prior distribution for limb 7%
0%, 23% 20%
0%, 87% 20%
0%, 87%
Likelihood for limb (from observed data) 11%
0%, 37% 15%
3%, 33% 18%
5%, 36%
Posterior distribution for limb 9%
1%, 22% 15%
4%,33% 18%
6%, 36%
*All probabilities are modelled as beta distributions. The traditional, frequentist 95% confidence intervals for the trial can be compared to the 95% credible intervals of the posterior distributions produced by the Bayesian analysis which combines prior beliefs about the results of each limb with the observed data (see text and [14]).
Figure 2 Title: PRIME-MD diagnosis versus whether the ED physician made a psychiatric diagnosis, consultation or referral, by study limb. For each study limb a 2 × 2 table depicts how each patient's PRIME-MD status (psychiatric diagnosis versus none) compares to emergency physician's decision to make a psychiatric diagnosis, obtain psychiatric consultation, or refer the patient for evaluation of the psychiatric condition.
From the physician and patient questionnaires (response rates 41% and 55% respectively, see Appendix for details) we learned that: most physicians with a patient in the tell-both group learned that patients' PRIME-MD status from the stickers, not from the patient. Few patients said they told their ED physician about the PRIME-MD results and physicians confirmed that they were seldom told such information. While physicians generally agreed with the PRIME-MD diagnoses, they did not choose to act on them, most commonly because "the visit wasn't about that issue." Only 2 of 28 patients contacted at least 2 weeks after the ED visit reported seeing a physician or mental health care professional about their psychiatric problem.
Discussion
This small trial, when interpreted in the context of our previous trial [4], is sufficient to make the following points: 1) it is feasible to use the PRIME-MD in the ED, 2) there is a high prevalence of psychiatric illness in the ambulatory ED population, 3) in our ED, when physicians or patients or both are informed of the presence of a PRIME-MD psychiatric diagnosis, few psychiatric diagnoses are made and few patients receive consultation or referral to address the problem, 4) our strategy of using the patient to initiate a discussion of the patient's mental health did not work.
While 18% (95% Credible Interval 6%, 36%) of patients in the tell-both group were given consultation or referral, 10% (95% Credible Interval -9%, 28%) more than in the control group, we do not believe that the improvement was of sufficient magnitude to justify using PRIME-MD in the waiting room of our ED.
These findings are a bit disconcerting since the treatments for many of these psychiatric conditions are highly effective and the morbidity of the conditions is considerable. Furthermore, treatments for conditions related to the patient's chief complaints (low back pain, general weakness, etc.) have limited efficacy. It would therefore be highly desirable to develop a method for identifying and treating these patients' psychiatric conditions.
The focus groups and this trial suggest that emergency physicians are reluctant to consider psychiatric illness in patients presenting with somatic complaints for reasons such as: discomfort with the content and length of the psychiatric interview, lack of knowledge of diagnostic criteria, lack of interest in the activity, and belief that the activity is futile since the health care system provides few resources for the non-psychotic, underinsured patients who are the most likely target of this intervention. It might be possible to have the computer refer patients directly to providers, thereby bypassing the physicians' resistance, but this could only occur in a system where viable referral options exist. Until such time, it is unlikely that a successful intervention can be developed until readily available follow-up becomes available.
This study took place in a U.S. academic ED within a health care system that provides extremely limited mental health follow-up for underinsured and uninsured patients who are not homicidal, suicidal or floridly psychotic, the majority of patients in our study. Results might be different in a system (such as the U.K.'s) that provides reasonable treatment options for such patients. The inclusion criteria are somewhat subjective. It is quite likely that eligible patients were missed. Furthermore, it is possible that PRIME-MD produced false positive and false negative results. Nevertheless, it is highly unlikely that these selection and misclassification biases would affect our conclusion. The small size of our study makes our estimates imprecise, but not so much so that our conclusions are jeopardized. The poor response rates for physician and patient questionnaires seriously compromises their utility.
We powered the study to look for large effects and cannot comment on whether the use of PRIME-MD produces a modest increase in psychiatric diagnosis and referral in patients with occult psychiatric illness. In retrospect, we might have listened to the focus group participants more carefully. They candidly told us why they were not keen to diagnose and treat these patients. We ignored this and designed an intervention that attempted to use the patients to force their hand. We might have had more success had we organized a follow-up clinic for such patients and directly addressed the physicians' attitudes. We also learned that while the trained undergraduate student research assistants did quite a good job identifying, consenting, and enrolling patients, they were less effective with the physician questionnaire (40% completion rate). We were unrealistic to expect that a junior person could capture the attention of a busy ED resident to get the questionnaire completed. Finally, we might have used block randomization to avoid the imbalance in assignments that occurred in this study.
Conclusion
In summary, in our ED, telling PRIME-MD diagnoses to the patient (and physician) did not substantially increase the proportion of patients whose were diagnosed and referred for treatment of their psychiatric condition in a health care system that offered extremely limited follow-up possibilities for such patients.
Competing interests
The study was funded in part by an unrestricted gift from the Pfizer Corporation. Apart from providing financial support, Pfizer had no involvement in the planning, conduct, analysis or reporting of the trial.
Authors' contributions
DS and PG conceived and designed the study. PG and WN coordinated the study. CL conducted the patient interviews and performed many of the psychiatric interviews. All authors participated in the drafting and editing of the manuscript. DS takes responsibility for the conduct and reporting of the research.
Appendix
Methods
Physician questionnaire
Participating residents were approached as soon as each patient was discharged from the ED, regardless of randomization status. A research assistant administered one of two versions of a structured questionnaire. Physicians whose patient had a PRIME-MD diagnosis were asked whether they knew the PRIME-MD diagnoses, how they came to know them, whether they agreed with them (and why), and whether they took any actions (and why). Physicians whose patient did not have a PRIME-MD diagnosis were asked if they knew the patient's PRIME-MD status.
Patient questionnaire
The psychiatric social worker attempted to call each patient during a 2 week interval beginning 2 weeks after the ED visit. Each patient had been asked to provide a current phone number at the time of behaviour. Patients were asked 15 closed-ended questions about their current physical and mental health in comparison to the day of the ED visit, their knowledge of their PRIME-MD diagnoses, whether they discussed PRIME-MD with their ED physician, whether their ED physician addressed psychiatric issues during their visit, whether they had been given a referral regarding psychiatric issues, and what types of health care they had received subsequent to the ED visit. At the end of the interview, the social worker attempted to help patients who desired intervention but had not received it find an appropriate source of care.
Results
Physician questionnaire
19 of 44 (43%) physicians whose patient did not have a PRIME-MD completed the short questionnaire. 3 of the 4 physicians in the tell-both group knew that the patient did not have a PRIME-MD psychiatric diagnosis. In none of the other 15 cases did the physician know the patient's PRIME-MD status.
20 of 51 (39%) physicians who treated patients with a PRIME-MD diagnosis completed the long-form questionnaire. In 10 of the 20 cases, the physician knew the patient's PRIME-MD diagnosis, including 7 of 8 cases in the tell-both group, 3 of 9 physicians in the tell-patient group, and 0 of 3 patients in the tell-neither group. The physician agreed with all of the PRIME-MD diagnoses in 5 of these cases, agreed with some in 4, and was unsure in 1.
6 of the 10 physicians unaware of the patient's PRIME-MD diagnoses agreed with all (2) or some (4) of them when informed of them. 1 disagreed, 2 had no opinion, and 1 did not answer. Only 2 of these 10 physicians wished that they had known the PRIME-MD diagnoses prior to beginning the patient interview.
6 physicians stated that they informed the patient about a psychiatric diagnosis (4) or wrote one in the medical record (4). The physicians who did not cited reasons such as "I was not confident of my diagnosis," "I did not want to antagonize my patient," "I did not want to stigmatize the patient," and "that was not what this visit was about." 2 physicians stated that they referred patients for further evaluation and treatment of their psychiatric condition. Those who did not cited "Did not believe patient would benefit," "ED not the place for this activity," and "patient did not meet DSM criteria for major depression," in addition to the aforementioned reasons.
Patient questionnaire
Despite confirming phone numbers at the time of behaviour, and making up to 10 attempts at contact, we reached only 28 of the 51 patients with a PRIME-MD diagnosis. Most patients reported that their physical health and mental health were somewhat or greatly improved. 1 patient reported that her physical health had worsened. 1 patient reported that his mental health had deteriorated. Both of these patients had been provided with psychiatric referrals but neither had received care.
14 of 21 patients who were told their PRIME-MD status in the ED remembered it at the time of the call. Of these 14 patients, all but one correctly stated whether or not they had PRIME-MD diagnoses, although they were somewhat confused about which diagnoses they had, particularly with regards to the presence of mood or anxiety disorders. 5 patients (none from the control group) reported discussing their PRIME-MD status with their ED physician. 2 reported seeing a mental health specialist subsequent to the ED visit.
12 patients found the PRIME-MD evaluation "somewhat" or "very" helpful, 2 found it "somewhat harmful" and 14 were neutral. 22 were glad that they took it. None of the 6 who were not glad had a PRIME-MD diagnosis. Patients were generally grateful for the call and were generally not that concerned that their mental health needs (as identified by PRIME-MD) had not been met.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors thank Ian Coulter for conducting the focus groups, UCLA Public Health Student Matthew Wise for coordinating the study, UCLA undergraduates Sarah Spencer, Henry Hoang, Kim Trinh, Bridgette Rivas, and medical students Sarah Lewis (nee Kaufman), Sripha Ouk, and Edward Lin for help with patient recruitment and study execution.
==== Refs
Marchesi C Brusamonti E Borghi C Giannini A Di Ruvo R Minneo F Quarantelli C Maggini C Anxiety and depressive disorders in an emergency department ward of a general hospital: a control study Emerg Med J 2004 21 175 9 14988342 10.1136/emj.2003.006957
Gold I Baraff LJ Psychiatric screening in the emergency department: its effect on physician behaviour Ann Emerg Med 1989 18 875 880 2757286
Wulsin LR Arnold LM Hillard JR Axis I disorders in ER patients with atypical chest pain Int J Psychiatry Med 1991 21 37 46 2066256
Schriger DL Gibbons PS Langone CA Lee S Altshuler LL Enabling the diagnosis of occult psychiatric illness in the emergency department: a randomized, controlled trial of the computerized, self-administered PRIME-MD diagnostic system Ann Emerg Med 2001 37 132 40 11174229 10.1067/mem.2001.112255
Zane RD McAfee AT Sherburne S Billeter G Barsky A Panic disorder and emergency services utilization Acad Emerg Med 2003 10 1065 9 14525739 10.1197/S1069-6563(03)00349-X
Byrne M Murphy AW Plunkett PK McGee HM Murray A Bury G Frequent attenders to an emergency department: a study of primary health care use, medical profile, and psychosocial characteristics Ann Emerg Med 2003 41 309 18 12605196 10.1067/mem.2003.68
Katon WJ Lin E Russo J Unutzer J Increased medical costs of a population-based sample of depressed elderly patients Arch Gen Psychiatry 2003 60 897 903 12963671 10.1001/archpsyc.60.9.897
Pfizer Incorporated PRIME-MD, version 12 1997 New York, NY: Pfizer
Spitzer RL Williams JB Kroenke K Linzer M deGruy FV 3rdHahn SR Brody D Johnson JG Utility of a new procedure for diagnosing mental disorders in primary care. The PRIME-MD 1000 study JAMA 1994 272 1749 1756 7966923 10.1001/jama.272.22.1749
Spitzer RL Kroenke K Williams JB Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. Primary Care Evaluation of Mental Disorders. Patient Health Questionnaire JAMA 1999 282 1737 44 10568646 10.1001/jama.282.18.1737
Persoons P Luyckx K Desloovere C Vandenberghe J Fischler B Persoons P Luyckx K Desloovere C Vandenberghe J Fischler B Anxiety and mood disorders in otorhinolaryngology outpatients presenting with dizziness: validation of the self-administered PRIME-MD Patient Health Questionnaire and epidemiology Gen Hosp Psychiatry 2003 25 316 23 12972222 10.1016/S0163-8343(03)00072-0
Leopold KA Ahles TA Walch S Amdur RJ Mott LA Wiegand-Packard L Oxman TE Prevalence of mood disorders and utility of the PRIME-MD in patients undergoing radiation therapy Int J Radiat Oncol Biol Phys 1998 42 1105 12 9869236 10.1016/S0360-3016(98)00346-0
Rollman BL Hanusa BH Lowe HJ Gilbert T Kapoor WN Schulberg HC A randomized trial using computerized decision support to improve treatment of major depression in primary care J Gen Intern Med 2002 17 493 503 12133139 10.1046/j.1525-1497.2002.10421.x
Schriger DL Problems with current methods of data analysis and reporting, and suggestions for moving beyond incorrect ritual Eur J Emerg Med 2002 9 203 7 12131650 10.1097/00063110-200206000-00021
Spiegelhalter DJ Abrams KR Myles JP Bayesian Approaches to Clinical Trials and Health-Care Evaluation 2004 John Wiley & Sons, Chichester England
Eddy DM Hasselblad V Fast*Pro. Software for Meta-Analysis by the Confidence Profile Method 1992 Academic Press Inc., Harcourt Brace Jovanovich, San Diego, CA
| 15882462 | PMC1142309 | CC BY | 2021-01-04 16:31:03 | no | BMC Emerg Med. 2005 May 9; 5:2 | utf-8 | BMC Emerg Med | 2,005 | 10.1186/1471-227X-5-2 | oa_comm |
==== Front
BMC Ear Nose Throat DisordBMC Ear, Nose and Throat Disorders1472-6815BioMed Central London 1472-6815-5-31589006710.1186/1472-6815-5-3Research ArticleQuality of life, functional outcome, and voice handicap index in partial laryngectomy patients for early glottic cancer Kandogan Tolga [email protected] Aylin [email protected] Department of Otolaryngology & Head-Neck Surgery, SSK Izmir Hospital, Izmir, Turkey2 Selen Ses Merkezi, Ali Cetinkaya Bulvari No:31/1 Daire 24 Alsancak Izmir 35220 Turkey2005 12 5 2005 5 3 3 28 9 2004 12 5 2005 Copyright © 2005 Tolga and Aylin; 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
In this study, we aim to gather information about the quality of life issues, functional outcomes and voice problems facing early glottic cancer patients treated with the surgical techniques such as laryngofissure cordectomy, fronto-lateral laryngectomy, or cricohyoidopexi. In particular, consistency of life and voice quality issues with the laryngeal tissue excised during surgery is examined. In addition, the effects of arytenoidectomy to the life and voice quality are also studied.
Methods
29 male patients were enrolled voluntarily in the study. The average age was 53.9 years. Three out of 10 patients with laryngofissure cordectomy also had arytenoidectomy. 11 patients had fronto-lateral laryngectomy with Tucker reconstruction, two of which also had arytenoidectomy. There were eight patients with cricohyoidopexi and bilateral functional neck dissection. Three of these patients also had arytenoidectomy. In bilateral functional neck dissection cases, spinal accessory nerve was preserved and level V of the neck was not dissected. None of the patients had neither radiotherapy nor voice therapy. Cordectomy patients never had a temporary tracheotomy or were connected to a feeding tube. Data was collected for 13 months for the cordectomy group, 14 months for fronto-lateral laryngectomy and cricohyoidopexi groups on average post-operatively. Statistical analysis in this study was carried out using the one-way analysis of variance, and the Post-Hoc group comparisons were made after Bonferroni and Scheffé-procedures.
In order to determine the effects of arytenoidectomy, a regression analysis is carried out to see if there are statistical differences in answers given to the survey questions among patients who were arytenoidectomized during their surgeries.
Results
There was a statistically significant difference between cordectomy and cricohyoidopexi group in answers to the University of Washington- Quality of Life- Revised survey part 1. (p = 0). A statistically significant difference was also established between cordectomy and fronto-lateral laryngectomy groups, as well as between cordectomy and cricohyoidopexi groups in answers to the University of Washington- Quality of Life- Revised survey part 2. (p = 0,036 and p = 0.009, respectively). Cricohyoidopexi group has given the lowest scores and the cordectomy group has given the highest scores in three survey questions representing the quality of life, performances and new voices. These ranges are also consistent with the laryngeal tissue excised during surgery (cricohyoidopexi > fronto-lateral laryngectomy > cordectomy). There was no statistically significant difference between groups in Performance Status Scale for Head and Neck cancer patients instrument. The difference between the Voice Handicap Index and Voice Handicap Index (functional); Voice Handicap Index (physical) and Voice Handicap Index (emotional) scores in three patient groups was not significant either. All of the patients evaluated that their new voices have similar functional, physical and emotional impact on their life. Decanulation and oral feeding times of cricohyoidopexi and fronto-lateral laryngectomy patients are found to be significantly longer than cordectomy patients. Lastly, the removal of arytenoid does not have any significant adverse effects on the quality of life, the functional outcomes, or the quality of voice.
Conclusion
In the present study, all patients with early glottic cancer, treated with different surgical technics reported fairly good quality of life outcomes, functional results and voice qualities. This study also finds that the removal of arytenoid does not have any adverse effects on the quality of life and voice from the patients' point of view.
==== Body
Background
In clinical research, QOL is recognized as an important endpoint in addition to the traditional endpoints such as response rate, disease-free survival, and overall survival [1]. It is particularly relevant for patients with head and neck cancer because social interaction and emotional expression depend to a great extent on the structural and functional integrity of the head and neck region [2]. Today, treatment policies are aimed to improve and maintain QOL during and after treatment.
Although early-stage glottic carcinomas (T1 and T2) are currently treated with radiation therapy or endoscopic laser resection with success, conservation laryngeal surgeries such as laryngofissure [C], [FLL], or [CHP] are also available alternatives for treatment [3]. In the surgical management of laryngeal malignancies, surgical alteration or removal of the larynx affect some of the most fundamental life functions including airway, digestion, and speech.
The measurement of voice quality outcomes following treatment for laryngeal cancer is a relatively new concept. Voice quality encompasses social, psychosocial, mental and physical components. Given patients with same diagnosis and same treatment, the outcome data may be totally different for them. Evaluation of voice quality following laryngeal surgery can be one of the first attempts to assess the functional outcome following successful treatment.
Although health-related quality of life issues in early stages of laryngeal cancer are more important when comparing different therapy modalities such as surgery versus radiation therapy, in this study we aim to gather information about the QOL issues, functional outcomes and voice problems facing early glottic cancer patients that were treated with the surgical techniques mentioned above.
Results of answers provided to survey questions are compared to determine whether life and voice quality issues are consistent with the laryngeal tissue excised during surgery. In addition, the effects of arytenoidectomy to the life and voice quality are also studied.
Methods
29 male patients were enrolled voluntarily in the study. The average age was 53.9 years. There were 10 patients with laryngofissure [C], three of which also had (A). Two out of 11 patients with [FLL] with Tucker reconstruction also had (A). There were eight patients with [CHP] and (BFND). Three of these patients also had (A). In BFND cases, spinal accessory nerve was preserved and level V of the neck was not dissected. Margins were checked with frozen sections per-operatively. None of the patients had neither radiotherapy nor voice therapy. [C] patients never had a temporary tracheotomy nor connected to a feeding tube. Data was collected in [C] group over 13 months, [FLL] group over 14 months and in [CHP] group over 14 months on average post-operatively. The data on the stage of the laryngeal cancer, the decanulation and the oral feeding times, the surgeries applied, and the scores of the questionnaires are listed in additional file 1. [see Additional file 1]
UW-QOL-R
The UW-QOL-R was given to all patients treated for early glottic cancer with the above-mentioned surgeries. The patient selects an answer from three to five choices, depending on the question. The UW-QOL-R data were scored on a scale of 0 to 100, with a score of 100 being totally functional and 0 being completely incapacitated [4]. Two small modifications have been made to increase the sensitivity to the UW-QOL-R. First, "My speech is not normal, but I can say all words" was added to the speech domain. Second modification was the addition of a question regarding the use of saliva substitutes under the saliva domain [4].
In part 2 of UW-QOL-R, the patient answers to global quality of life questions, selects an answer from five to six choices, ranging from very poor to outstanding, or from much worse to much better, depending on the question.
PSS-HN
In PSS-HN, eating in public (EIP), understandability of speech (UOS), and normalcy of diet (NOD) are each individually assessed on a zero to 100-point scale [4].
VHI
The patients were instructed that these statements are how many people describe their voices and the effects of their voices on their lives. The patients marked the response that indicates how frequently they have the same experience. 0 = Never 1 = Almost Never 2 = Sometimes 3 = Almost Always 4 = Always.
0 to 30 = These are low scores, and indicate that most likely there is a minimal amount of handicap associated with the voice disorder. 31 to 60= Denotes a moderate amount of handicap due to the voice problem. 60 to 120= These scores represent a significant and serious amount of handicap due to a voice problem[5].
Statistical Analysis
Statistical analysis in this study includes one-way analysis of variance, ANOVA, and the Post-Hoc group comparisons after Bonferroni and Scheffé-procedures.
In order to determine the effects of arytenoidectomy, the following regression model is adopted to see if there are statistical differences in answers given to survey questions among different patient groups that were arytenoidectomized during their surgeries.
Y = α + β1.CHP + β2.FLL + β3.A + β4.CHP.A + β5.FLL.A
where Y is the answer given by a patient to a question, α is the mean of answers given by patients who had the [C] surgery technique. [CHP] and [FLL] are dummy variables for surgery techniques [CHP] and [FLL], respectively. The coefficients of these variables will show the amount of difference in answers given by patients who had these techniques, and those who had the [C] technique. The t statistic of these coefficients will be used in testing whether these differences are statistically significant or not. Similarly, [A] is a dummy variable for patients who had the A procedure applied after these above-mentioned surgery techniques. Its coefficient quantifies the difference in answers given by patients who had A procedure applied after a surgery technique, and those who did not have. The last two variables are interaction variables between the [CHP] and [FLL] dummy variables, and the [A] dummy variable. Their purpose in the model is to test whether [A] applied after [CHP] or [FLL] leads to statistically different answers by patients than [A] applied after [C].
Results
The regression results are given in Table 1. The effects of the [A] procedure were different depending on the previous surgical technique applied. If applied after [CHP], the answers of the patients to first questionnaire were generally higher, but insignificant than those of patients who had [A] after [C]. The measures of the second and third questions were higher too, if [A] is applied after [CHP], rather than after [C]. These measures are also statistically higher if [A] is applied after [FLL], rather than after [C], as well. However, the answers to first questionnaire are sometimes higher, sometimes lower if [A] is applied after [FLL], rather than after [C].
The mean values for each analyzed parameter in questionnaires for three patient groups are shown on table 2. There was a statistically significant difference between [C] and [CHP] group in answers to the UW-QOL-R questionnaire part 1. (p = 0). There was also a statistically significant difference between [C] and [FLL] group and [C] and [CHP] group in answers to the UW-QOL-R questionnaire part 2. (p = 0,036 and p = 0.009). [CHP] group have given the lowest scores and the [C] group has given the highest scores in three questionnaires to represent their quality of life, performances and new voices. These ranges are also consistent with the laryngeal tissue excised during surgery ([CHP]> [FLL]> [C]).
There was not a statistically significant difference between groups in PSS-HN instrument. The difference between the VHI and VHI-F, VHI-P, VHI-E scores in three patient groups were not statistically different either (Table 3). All of the patients evaluated that their new voices had similar functional, physical and emotional impact on their life. Decanulation and oral feeding times of [CHP] and [FLL] patients has been found to be significantly longer than [C] patients (Table 4).
Discussion
Patients generally differ in their response to cancer and to the therapeutic interventions used in its treatment. Even patients with a similar oncologic site and stage who receive identical treatment can differ in their own assessment of quality of life [6].
Not surprisingly, [CHP] group have given the lowest scores and the [C] group has given the highest scores in three questionnaires, representing their quality of life, performances and new voices (Table 2). These ranges are also consistent with the laryngeal tissue excised during surgery ([CHP]> [FLL]> [C]). A possible reason for this finding could be that these [C] patients were never tracheotomized, so they were more comfortable in the postoperative period. Furthermore, they were hospitalized for fewer days, and they never had a feeding problem after the operation. Lastly, the most important two factors, swallowing and communication have never been a problem for this group. The highest score may reflect this general well-being after the operation.
Another interesting outcome of our study is that although the arytenoidectomized patients had longer oral feeding and decanulation times, there are no statistically significant effects on the quality of life, the functional outcomes, or the quality of voice. This suggests that the more laryngeal tissue is removed with the exception of arytenoid, the worse the quality of life, the functional outcomes and the quality of voice will be. This finding further implies that the arytenoid removal does not have an adverse effect on the quality of life, the functional outcomes, or the quality of voice from the patients' point of view. Removing the arytenoid only makes the oral feeding and the decanulation times longer in the first weeks after the operation, but after the tissue heals completely and laryngeal reflexes return to normal, larynx begins compensating arytenoid excision, and functions satisfactorily.
UW-QOL-R
Since appearance, taste, saliva, chewing and pain are not major issues for most early-stage glottic cancer patients, their discussion was excluded. Swallowing is considered important by the patients, if especially one arytenoid was resected during surgery. Difficulty in swallowing may exist by these patients because of aspiration risk, until the larynx adapts to this new post-surgical situation, and laryngeal reflexes returned to normal. The effect of arytenoidectomy on the swallowing parameter was also analyzed in our study, and interestingly, it is found that the arytenoidectomy has no adverse effect on swallowing after the patients have been discharged.
Since the improvement in swallowing, and the normalcy of voice communication, the two major criteria for quality of life after the laryngeal surgery [7], reflect the level of quality of life in patients treated for laryngeal cancer, it can be suggested again that the [CHP] group had the least favorable quality of life after the operation among the study groups. The significantly longer decanulation and oral feeding times, i.e. longer hospitalization, could be a sign for a relatively worse quality of life.
PSS-HN
The measures of PSS-HN scored lower for patients who had the [A] procedure. It could be suggested that adding arytenoidectomy would result in lowering of the functional capabilities of the patients, by statistically insignificant amount though.
UOS is strongly related to voice production. The amount of laryngeal tissue excised is highest in patients treated with [CHP] and as expected, the VHI is lowest in these patients. The lowest scores for UOS were given to the [CHP] patients, and as expected, the UOS score is significantly lower in the [CHP] group.
VHI
Vocal results after [C] and [FLL] could be unsatisfactory, since there is a gap in the creation of a new glottis due to removal of a considerable mass of tissue. After [CHP], the results could be much worse, since both the vocal folds were excised and a neoglottis formation is anticipated. Although statistically insignificant, the [CHP] patient group had given the lowest VHI scores in assessing their new voices in our study. In all of the patient groups, the quality of voice was found to be sufficient to hold a normal individual conversation. However, the voice was defined by the patients as hoarse and dull. It was rated to be insufficient to make a conversation in a noisy atmosphere, since it can not be raised satisfactorily. It also should be remembered that all patients with laryngeal malignancies had voice and communication problems at diagnosis.
Many authors have also studied the quality of voice following radiation therapy for T1 and T2 glottic cancers. Traditionally, radiation therapy has been advocated over surgery by many authors because of the belief that voice quality is superior in the radiation therapy cohort. A review of the literature shows a wide variation in findings [8-15].
The measures VHI questionnaire scored lower for patients who had the [A] procedure. It could be hypothesized that the addition of arytenoidectomy would result in increases in voice handicap of the patients, though statistically insignificant.
Conclusion
In the present study, all patients with early glottic cancer, treated with different surgical techniques reported fairly good QOL outcomes, functional results and voice qualities. The results of this study also imply that the removal of arytenoid does not have any statistically significant adverse effects on the quality of life, the functional outcomes, or the quality of voice from the patients' point of view.
Abbreviations
[C] = Cordectomy
[FLL] = Fronto-lateral laryngectomy
[CHP] = Cricohyoidopexi
[A] = Arytenoidectomy
[BFND] = Bilateral functional neck dissection
[QOL] = Quality of life
[UW-QOL-R] = University of Washington- Quality of Life- Revised
[PSS-HN] = Performance Status Scale for Head and Neck cancer patients
[EIP] = Eating in public
[UOS] = Understandability of speech
[NOD] = Normalcy of diet
[VHI] = Voice Handicap Index
[VHI-F] = Voice Handicap Index, functional
[VHI-P] = Voice Handicap Index, physical
[VHI-E] = Voice Handicap Index, emotional
Competing Interests
The author(s) declare that they have no competing interests.
Authors' Contributions
TK applied the survey to the patients, assisted in statistical analyses and drafted the manuscript. AS participated in the design 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
Click here for file
Acknowledgements
The authors would like to thank Yener Kandogan, Ph. D. Assistant Professor of Business Economics in University of Michigan for his assistance in statistical analyses.
Figures and Tables
Table 1 Regression Results. Numbers in parentheses below the coefficients are the t statistics.
VARIABLE Constant CHP FLL A CHP.A FLL.A
Pain 82.14 -2.14 6.75 9.52 -14.52 1.59
(8.67)** (-0.15) (0.53) (0.55) (-0.58) (0.06)
Appearance 89.29 -19.29 -3.17 10.71 10.95 -9.33
(8.29)** (-1.16) (-0.22) (0.54) (0.38) (-0.31)
Activity 85.71 -15.71 -2.38 14.29 7.38 -22.62
(9.04)** (-1.07) (-0.19) (0.82) (0.29) (-0.86)
Recreation 89.29 -19.29 5.16 10.71 2.62 -55.16
(10.47)** (-1.46) (0.45) (0.69) (0.12) (-2.34) **
Swallowing 100.00 -26.60 -14.78 -11.00 26.60 9.28
(12.80)** (-2.20)** (-1.42) (-0.77) (1.28) (0.43)
Chewing 100.00 -10.00 -11.11 0.00 10.00 11.11
(16.53)** (-1.07) (-1.38) (0.00) (0.62) (0.67)
Speech 89.29 -19.29 -8.73 10.71 2.62 -3.77
(11.61)** (-1.62)* (-0.85) (0.76) (0.13) (-0.18)
Shoulder 85.71 -5.51 6.95 3.29 -5.49 -12.45
(9.34)** (-0.39) (0.57) (0.20) (-0.22) (-0.49)
Taste 95.29 -8.49 -6.40 -17.62 19.82 28.73
(11.84)** (-0.68) (-0.60) (-1.20) (0.93) (1.29)
Saliva 95.29 -8.49 -10.06 -6.29 19.49 21.06
(13.81)** (-0.79) (-1.09) (-0.50) (1.06) (1.11)
EIP 100.00 -20.00 -25.00 -33.33 53.33 58.33
(8.52)** (-1.10) (-1.60) (-1.56) (1.71)* (1.80)*
UOS 82.14 -27.14 -1.59 -32.14 68.81 39.09
(11.58)** (-2.47)** (-0.17) (-2.48)** (3.65)** (2.00)**
NOD 100.00 -4.00 -7.78 -0.00 4.00 7.78
(21.52)** (-0.56) (-1.26) (-0.00) (0.32) (0.61)
VHI 69.43 -12.43 -10.54 -6.43 8.10 15.04
(12.15)** (-1.40) (-1.38) (-0.62) (0.53) (0.95)
* denotes significance at 90% confidence level, and ** denotes significance at 95% confidence level.
Table 2 The mean values for each analyzed parameter in questionaires, decanulation and oral feeding times for 3 patient group.
C (n = 10) FLL (n = 11) CHP (n = 8)
UW-QOL-R part 1
Pain 85 91 78
Appearance 92,5 86,3 78,1
Activity 90 81,8 78,1
Recreation 92,5 86,3 75
Swallowing 96,7 84,9 79,3
Chewing 100 90,9 93,8
Speech 92,5 81,8 75
Shoulder 86,7 91 79,4
Taste 90 91 87,6
Saliva 93,4 87,9 91,8
Total score 919,3 872,9 816,1
UW-QOL-R part 2
Question 1 40 13,6 12,5
Question 2 51,7 26,8 22,8
Question 3 55,1 34,5 14
Total score 146,8 74,9 49,3
PSS-HN
Eip 90 79,5 87,5
Uos 72,5 81,8 68,8
Nod 100 93,6 97,5
Total score 262,5 254,9 253,8
VHI
VHI 67,5 60,4 57,6
VHI-f 21,8 24 20
VHI-p 22 19,7 17,1
VHI-e 23,7 16,7 20,6
Decanulation (day) 0 3,36 19,25
Oral feeding (day) 1 7,63 22,38
Table 3 Number of patient groups and VHI, VHI-F, VHI-P and VHI-E mean values and standart deviations (SD).
GROUP VHI VHI-F VHI-P VHI-E
C (n = 10) Mean (SD) 67,50 (17,34) 21,80 (5,49) 22,00 (7,45) 23,70 (9,04)
FLL (n = 11) Mean (SD) 60,45 (13,97) 24,00 (7,95) 19,73 (6,66) 16,73 (8,58)
CHP (n = 8) Mean (SD) 57,63 (10,78) 20,00 (4,14) 17,13 (5,46) 20,50 (3,74)
Table 4 Decanulation and oral feeding times (day) of the patients and standart deviations (SD).
Decanulation (day) (SD) Oral feeding (day) (SD)
C (n = 7) 0 0
C + A (n = 3) 0 1
FLL (n = 9) 2,4 (0,8) 6,2 (1,3)
FLL + A (n = 2) 6,5 (0,7) 13 (1,4)
CHP (n = 5) 18,2 (1,9) 20,6 (2,0)
CHP + A (n = 3) 21,3 (0,5) 24,6 (1,1)
==== Refs
Quality of life and clinical trials [editorial] Lancet 1995 1 1 2
Breitbart W Holland J Psychosocial aspects of head and neck cancer Semin Oncol 1988 15 61 69 3344444
Marks SC Conservation laryngeal surgery Curr Oncol Rep 1999 1 99 104 11122805
Smith JC Johnson JT Cognetti DM Landsittel DP Gooding WE Cano ER Myers EN Quality of Life, Functional Outcome, and Costs of Early Glottic Cancer Laryngoscope 2003 68 76 12514385 10.1097/00005537-200301000-00013
Benninger MS Ahuja AS Gardner G Grywalski C Assessing outcomes for dysphonic patients J Voice 1998 12 540 50 9988041
Hassan SJ Weymuller EA Assessment of quality of life in head and neck cancer patients Head Neck 1993 15 485 496 8253555
Dejonckere PH Occupational Voice: Care and Cure 2001 Kugler Publications, The Hague 113 128
Morgan DA Robinson MarshHF Bradley PJ Vocal quality 10 years after radiotherapy for early glottic cancer Clin Radiol 1988 39 295 296 3396281
Aref A Dworkin J Devi S Denton L Fontanesi J Objective evaluation of the quality of voice following radiation therapy for T1 glottic cancer Radiother Oncol 1997 45 149 153 9424005 10.1016/S0167-8140(97)00154-0
Verdonck-de Leeuw IM Hilgers FJ Keus RB Koopmans-van Beinum FJ Greven AJ de Jong JM Vreeburg G Bartelink H Multidimensional assessment of voice characteristics after radiotherapy for early glottic cancer Laryngoscope 1999 109 241 8 10890774 10.1097/00005537-199902000-00014
Casiano RR Cooper JD Lundy DS Chandler JR Laser cordectomy for T1 glottic carcinoma: a 10-year experience and videostroboscopic findings Otolaryngol Head Neck Surg 1991 104 831 837 1908976
Delsupehe KG Zink I Lejaegere M Bastian RW Voice quality after narrow-margin laser cordectomy compared with laryngeal irradiation Otolaryngol Head Neck Surg 1999 121 528 533 10547464
McGuirt WF Blalock D Koufman JA Feehs RS Hilliard AJ Greven K Randall M Comparative voice results after laser resection or irradiation of T1 vocal cord carcinoma Arch Otolaryngol Head Neck Surg 1994 120 951 955 8074822
Stoeckli SJ Guidicelli M Schneider A Huber A Schmid S Quality of life after treatment for early laryngeal carcinoma Eur Arch Otorhinolaryngol 2001 258 96 9 11307614 10.1007/s004050000307
Schneider A Guidicelli M Stockli SJ Quality of life after treatment of laryngeal carcinoma: surgery versus radiotherapy Schweiz Med Wochenschr Suppl 2000 116 31S 34S 10780067
| 15890067 | PMC1142310 | CC BY | 2021-01-04 16:03:26 | no | BMC Ear Nose Throat Disord. 2005 May 12; 5:3 | utf-8 | BMC Ear Nose Throat Disord | 2,005 | 10.1186/1472-6815-5-3 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-661587782310.1186/1471-2164-6-66Methodology ArticleAn analysis of the use of genomic DNA as a universal reference in two channel DNA microarrays Gadgil Mugdha [email protected] Wei [email protected] Chetan [email protected] Vivek [email protected] Wei-Shou [email protected] Department of Chemical Engineering and Materials Science, University of Minnesota, 421 Washnigton Ave. S.E., Minneapolis, MN 55455 USA2 School of Mathematics, University of Minnesota, 270A Vincent Hall, Minneapolis, MN 55455 USA3 Scientific Computing and Mathematical Modeling, GlaxoSmithKline, Research Triangle Park, NC 27709 USA4 Department of Microbiology, University of Minnesota, MMC 196, 420 Delaware Street, S.E., Minneapolis, MN 55455 USA2005 8 5 2005 6 66 66 9 9 2004 8 5 2005 Copyright © 2005 Gadgil et al; licensee BioMed Central Ltd.2005Gadgil 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
DNA microarray is an invaluable tool for gene expression explorations. In the two-dye microarray, fluorescence intensities of two samples, each labeled with a different dye, are compared after hybridization. To compare a large number of samples, the 'reference design' is widely used, in which all RNA samples are hybridized to a common reference. Genomic DNA is an attractive candidate for use as a universal reference, especially for bacterial systems with a low percentage of non-coding sequences. However, genomic DNA, comprising of both the sense and anti-sense strands, is unlike the single stranded cDNA usually used in microarray hybridizations. The presence of the antisense strand in the 'reference' leads to reactions between complementary labeled strands in solution and may cause the assay result to deviate from true values.
Results
We have developed a mathematical model to predict the validity of using genomic DNA as a reference in the microarray assay. The model predicts that the assay can accurately estimate relative concentrations for a wide range of initial cDNA concentrations. Experimental results of DNA microarray assay using genomic DNA as a reference correlated well to those obtained by a direct hybridization between two cDNA samples. The model predicts that the initial concentrations of labeled genomic DNA strands and immobilized strands, and the hybridization time do not significantly affect the assay performance. At low values of the rate constant for hybridization between immobilized and mobile strands, the assay performance varies with the hybridization time and initial cDNA concentrations. For the case where a microarray with immobilized single strands is used, results from hybridizations using genomic DNA as a reference will correspond to true ratios under all conditions.
Conclusion
Simulation using the mathematical model, and the experimental study presented here show the potential utility of microarray assays using genomic DNA as a reference. We conclude that the use of genomic DNA as reference DNA should greatly facilitate comparative transcriptome analysis.
==== Body
Background
The rapid increase in the number of completely sequenced genomes in the past few years has generated much effort in functional genomics, particularly studies seeking to assign biological functions to DNA sequences. Comparative gene expression profiling is widely used to study the functional role of genes. The DNA microarray assay provides an invaluable technique for large scale expression analysis. In the two-channel DNA microarray assay, RNA from two samples is reverse transcribed to cDNA and labeled with two distinct fluorescent dyes before being co-hybridized to immobilized DNA strands on a microarray slide. Spotted arrays currently being used can be divided into two groups based on the nature of immobilized DNA used: one in which the immobilized DNA is comprised of both sense and antisense strands (usually PCR product) and the other where the immobilized DNA is single stranded consisting of only the sense strands (usually, chemically synthesized oligonucleotides). During hybridization, the two fluorescently labeled cDNA samples compete for hybridization to the immobilized strands. Hybridization reactions between complementary strands occur only between the labeled antisense strand and immobilized sense strand. The ratio of the intensities of the two fluorescently labeled cDNAs is used to quantify the relative levels of transcripts in the two samples [1,2]. This method serves well for pair-wise comparison of transcript levels in two samples. With over ten thousand different DNA species immobilized on the microarray, the relative transcription level of all the corresponding genes in the two samples can be obtained in a single assay.
DNA microarrays have found applications in gene discovery, disease diagnosis, pharmacogenomics and toxicology research. They are increasingly used for a series of related samples, for which a comparison across all samples and all genes is desirable. When a large number of samples are to be compared, a combinatorial approach pairing all possible pairs (or at least a number of combinations of pairings of the sample) is often taken. This results in a large number of microarrays, requiring a large amount of each RNA sample. A 'loop design', where every sample is directly compared to two other samples to form a closed loop, has been proposed to overcome this problem [3,4]. The ratios calculated using a loop design have variable levels of precision since some samples are more directly related than others [5]. When a new sample is to be inserted into the earlier 'loop', RNA for at least two of the previous samples is needed to pair with the new sample to form a new node in the closed loop.
Another approach to tackle the issue of a large combinatorial pair-wise comparison is the 'reference design' [3] in which a common reference sample is introduced with which all RNA samples are hybridized. Two possible universal references are RNA pooled from various samples and genomic DNA [6]. For a given set of samples, pooled RNA provides an excellent reference. However, if the experimental conditions change, the possibility arises that some new transcripts may not be represented in the initially-pooled RNA.
Genomic DNA is an attractive candidate for use as a common reference. It is isolated from cells or tissue and sheared to fragments in a narrow range of length. It is easier to prepare, maintain and reproduce, as compared to RNA. It is especially useful for microorganisms, which lack repetitive sequences in their genome, and microarrays using genomic DNA as a reference have been used to identify genes differentially expressed in various growth stages of Mycobacterium tuberculosis [7]. It has also been recently reported that the data obtained using genomic DNA as a reference in microarray experiments with Arabidopsis thaliana employing 70-mer oligonucleotide microarrays was in agreement with ratios obtained from direct hybridizations [8]. Genomic DNA samples, isolated from stationary phase cultures where the chromosome is not being replicated, have the same representation of all genes as in the genome. Since the transcript level of each gene is being referenced to its own representation on the genome (for most genes, it is a single copy), the relative expression can be compared across different genes in the sample (i.e., from the same hybridization) as well as across different samples. The use of microarrays using genomic DNA in a range of applications including genomic diversity studies [9,10] and aneuploidy detection using comparative genomic hybridization [11] has been demonstrated.
In the conventional two-channel cDNA hybridization [1], both the cDNA samples are antisense strands. The probability of hybridization between strands in solution is very low. On the other hand, using sheared genomic DNA as a reference, hybridization between complementary sense and anti-sense strands can occur in solution between the complementary genomic DNA strands, and between genomic DNA sense strands and their cDNA counterparts. The number of strands lost to hybridization in the solution phase may differ for different RNA species as well as for the two complementary strands of the same species. This may result in decreased fidelity in the ratio of cDNA to genomic DNA as a representation of gene expression level. With the complexity of hybridization in both the solution phase and the immobilized surface phase, and between double strands of genomic DNA and single strands of cDNA, it is difficult to assess the effect of using genomic DNA as a common reference. Adapting a mathematical model we developed previously to assess diffusional constraints on DNA microarray assay [12], we have constructed a kinetic model for microarray hybridization and predicted the effectiveness as well as potential pitfalls of using genomic DNA as a reference. We examined the effect of the various parameters that may affect surface hybridization. Here we report the framework of the model and our findings.
Mathematical model for microarray hybridization
A DNA microarray consists of thousands of spots, each spot containing DNA strands of known sequence immobilized on an impermeable surface. We simulate hybridization reactions for one spot on the microarray, which is considered to be a sample chamber with two compartments; a spot phase and a solution phase (Figure 1). The spot phase is the small volume in which the bound strands are assumed to be present at a uniform concentration. The solution phase, which comprises a vast majority of the volume of the microarray chamber, contains only fluorescently labeled single strands at the beginning of the hybridization and double strands formed by hybridization of complementary mobile single strands during the course of the hybridization.
Figure 1 Schematic of the system for simulating hybridization for a two-color assay. A: Solution phase, B: spot phase
Five kinds of single-stranded DNA molecules are present in the system: labeled single stranded cDNA (Sample) reverse transcribed from RNA, denoted as S; the two identically labeled complementary (anti-sense and sense) strands from genomic DNA (Reference), denoted as R and R'; and the non-labeled anti-sense and sense strands immobilized on the array surface (Bound strands) denoted by B and B'. The nomenclature used in this study is summarised in Table 1. The. The double-stranded species are denoted by combining the constituent single strand symbols, for example RB' denotes the labeled complex formed by hybridization of genomic DNA antisense strands to the sense bound strands. We assume that each phase is well-mixed, and hence all mobile species are present at uniform concentrations within each phase.
Table 1 Nomenclature
Symbol Description
S cDNA sample anti-sense strand
R Genomic DNA anti-sense strand
R' Genomic DNA sense strand
B Bound (immobilized) anti-sense strand
B' Bound (immobilized) sense strand
Subscript 0 Initial concentration
SR' Double strand formed in solution by hybridization of anti-sense cDNA strand and sense genomic DNA strand
RR' Double strand formed in solution by hybridization of anti-sense and sense genomic DNA strands
SB' Double strand formed on surface by hybridization of anti-sense cDNA strand and sense bound strands
RB' Double strand formed on surface by hybridization of anti-sense genomic DNA strand and sense bound strands
R'B Double strand formed on surface by hybridization of sense genomic DNA strand and anti-sense bound strands
h Height of spot phase
r Radius of spot phase
kb Rate constant of backward reaction of hybridization
kf Rate constant of forward reaction of hybridization between mobile species
kf-bound Rate constant of forward reaction of hybridization between mobile and bound species
kt Rate constant for transport between the two phases
γ Ratio obtained from a hybridization assay using genomic DNA as a reference { = [SB']/([RB']+[R'B])}
α Assay efficiency = (γ1/γ2)/(S10/S20), where 1 and 2 denote samples 1 and 2
εS Amount of S reacted with R'
εR Amount of R reacted with R'
We consider the case that PCR products are used for immobilization; thus, both sense and antisense strands of probe DNA are immobilized. We assume that the two complementary bound strands do not hybridize to each other [13] due to the immobilization procedure used and neglect the formation of BB'. The mobile species in the two compartments are considered to move across the phase boundary at a rate proportional to the difference in the concentrations of identical species present in the two phases. The proportionality constant is the effective mass transfer coefficient for the transport of mobile DNA strands and is estimated from the diffusivity as discussed below.
The model equations take the form of a mass balance on each component in each phase that accounts for the change in concentration due to reactions within that phase and transport between the two phases. Non-specific hybridization in both phases is neglected. All these equations are of the form:
where = R', = B', A1 = R, A2 = B, A3 = S, subscript p denotes either solution or spot phase and p' is the other phase; kf is the forward rate of hybridization; kb is the backward rate of hybridization; kt is the rate of transport across the spot-phase solution-phase boundary; Vt is the total volume of the sample chamber and Vp is the volume of phase p.
Model parameters
Model geometry
The diameter of the spot phase is set to 0.01 cm as seen in a typical microarray spotted on poly-lysine slides [14], with a height of 2 × 10-4 cm (Figure 1). This gives a spot volume of 0.2 × 10-12 l. Typically, labeled strands are resuspended in a volume of 4.5 × 10-5 l, and this volume is applied under a cover slip of area 6.5 cm × 2.4 cm. For this geometry, the height of the microarray chamber is about 2.9 × 10-3 cm. The solution phase volume is the difference between the sample chamber volume (4.5 × 10-5 l) and the spot volume (0.2 × 10-12 l). It has been shown before that only the mobile strands within a radius of 0.15 cm from the spot are available for hybridization due to transport effects [12]. Hence, we have assumed the solution phase to have a diameter of 0.15 cm and a height of 2.9 × 10-3 cm corresponding to the height of the microarray chamber.
Hybridization rate constants
The hybridization rate constants are assumed to be identical for DNA strands from both genomic DNA and cDNA. However, the rate constants for reaction between two mobile strands may differ from the rate constant for the reaction of a mobile strand and a bound strand. A forward rate constant (kf) for DNA hybridizations in solution of 106 M-1 s-1 [15,16] was used for simulations. The rate constant of hybridization could be slower for hybridization of mobile strands to immobilized strands (kf-bound) on a solid surface and values in a range up to 100-fold lower than kf have been used in simulations. We also discuss the effect of this reduced forward rate constant of hybridization between mobile and bound strands for a range between 106 M-1s-1 and 104 M-1s-1. The backward rate constant for dissociation of DNA double strands as calculated from equilibrium constants reported in literature ranges from 10-3 s-1 to 10-1 s-1 [17,18]. Simulations were carried out using a backward rate constant ranging from 10-1 s-1 to 0 s-1 (irreversible). In all simulations performed, the deviation from the true value was greatest when the hybridization reaction is set to be irreversible. The results discussed below are for the case with the backward rate constant set to 0 s-1. The results obtained when the reaction is set to be reversible are very close to the true value under all conditions tested.
Rate constant for transport
The rate constant for transport (kt) is estimated from the diffusion coefficient as , where D is the diffusion coefficient, h is the height of the spot phase and r is the radius of the spot phase. The diffusivity of DNA single strands in solution has been estimated to be 10-7 cm2/s [19-21] which leads to an estimate of 1 s-1 for kt. The simulations reported in the next section were carried out under both very fast transport (kt = 10-3 s-1) and very slow transport (kt = 10-3 s-1) conditions. The transport rate has no effect on the performance of the assay in the range tested (data not shown).
Initial single-strand concentrations
From the yeast transcriptome data published by Velculescu et al. [22], the mass percentages of mRNA that belong to the rare, intermediate and abundant classes are estimated to be 65.2, 32.4 and 2.4% respectively. For an mRNA sample of 0.2 μg used in each microarray assay, the total amount of rare, intermediate and abundant species are 0.13 μg, 0.07 μg and 4.8 × 10-3 μg respectively. To convert those numbers to molar concentration we calculate the number of genes within each abundance class using the intensity data from microarray experiments of S. coelicolor. The S. coelicolor transcripts were divided into three abundance classes using the following intensity cut-offs: Intensity < 2000 as rare, 2000 < Intensity < 20000 as intermediate and >20000 as abundant sequences. Normalized intensity values were used for this estimate. 72.8%, 26% and 1.2% of all genes were classified rare, intermediate and abundant respectively. The corresponding number of rare, intermediate and abundant species are 5697, 2034 and 94 respectively. Assuming a sample volume of 4.5 × 10-5 l, 100% reverse transcription efficiency, and an average strand length of 1000 we calculate the rare species concentration as ~0.1 pM, intermediate species concentration ~1 pM and abundant species concentration ~20 pM.
Previous experimental reports of the use of genomic DNA as a reference for microarray hybridizations used genomic DNA concentrations ranging from 0.1 μg to 4 μg for M. tuberculosis [7]. Since M. tuberculosis has ~4000 genes, this translates to a concentration of each gene from ~1 pM to 34 pM, which is the range of genomic DNA concentrations used for simulations.
To calculate the concentration of the immobilized species, we assume the concentration of DNA in the spotting solution to be 0.1 g/l and that 10-9 l of the solution is spotted on the microarray. We also assume that 75% of the DNA thus spotted is washed away in the microarray post-processing steps. 2.5 × 10-11 g DNA remaining on the microarray is uniformly distributed in the 0.2 × 10-10 l spot phase volume. Assuming an average DNA strand length of 1000 bp, this is approximately equal to an immobilized strand concentration in the spot phase of 10-6 M.
Results and Discussion
The mathematical model described above takes into account DNA hybridization between single stranded cDNA and double stranded genomic DNA (gDNA) in solution and immobilized double strands on a microarray surface. This model considers hybridization only on one spot on the microarray. The immobilized strands are distributed uniformly in the spot phase and the mobile strands are present both in the solution phase and spot phase. Hybridization between mobile and bound species occurs in the spot phase. All concentrations described in the following sections are the concentrations in the spot phase. The fluorescence intensity corresponding to hybridized cDNA sample is expressed as IS = [SB'] and the channel corresponding to the hybridized genomic DNA reference as IR = [RB'] + [R'B]. The result obtained for hybridization with genomic DNA used as a reference is a hybridization ratio γ = Is/IR = [SB'] / ([RB'] + [R'B]). Since the concentration of all genes in a genomic DNA sample is equal, the ratio (γ) for different genes is an indication of the relative abundance of the transcript for those genes.
Typically when genomic DNA is used as the reference, the ratio (γ1) from one hybridization of cDNA derived from sample one, is compared to another ratio (γ2) from sample two, to obtain the relative expression level of the transcripts in samples 1 and 2. Ideally this "ratio of ratios", i.e. γ1/γ2, should be equal to the ratio of the initial concentrations of the transcript in those samples. We simulate this experimental process with two different initial cDNA concentrations (S10 and S20) and the same genomic DNA concentration to obtain the ratios γ1 and γ2. The accuracy of the microarray assay is quantified by the accuracy index α,
A value of one for α corresponds to a perfect assay, where the measured relative concentration of the transcript in the two samples (γ1/γ2) is exactly equal to the true relative concentration (S10/S20). Any deviation in α from unity is a measure of the error of the assay.
These results are also applicable to the comparison of expression levels of two genes in one cDNA sample, as the model makes no distinction between hybridization on two spots on one microarray and hybridization to a spot corresponding to the same sequence in two different microarray experiments. We systematically vary model parameters to investigate the effect of different hybridization conditions, transcript abundance levels, and degree of differential expression on the performance of the microarray assay.
Effectiveness of using genomic DNA as a reference
The effectiveness of using genomic DNA as a reference in the microarray assay was predicted by simulations using the model and parameters described above. Figure 2a shows the variation in α with hybridization time for different RNA abundance levels and differential expression ratios corresponding to biologically realistic scenarios as listed in Table 2. These cases represent a wide range of possible combinations of the three abundance classes and differential expression ratios (2, 10 and 100). A differential expression ratio of 2 is used as an example of a small change in expression level and 100 as an example of a large change in expression level. The parameter values used are kf = 106 M-1s-1, kf-bound = 106 M-1s-1, kb = 0 s-1, initial genomic DNA concentration = 1 pM, initial bound strand concentration = 10-6 M, kt = 1 s-1. For a wide range of initial cDNA concentrations and differential expression ratios, the accuracy index α is within 5% from unity (Figure 2a), indicating that the assay performance is robust to the initial concentration of single strands. The only condition where α is significantly different from 1 (~1.17) is when intermediate species are upregulated 100 fold, a situation not likely to happen frequently in cells. Furthermore, in most microarray assays, a 17% deviation from the true value is not considered large. The model simulation predicts that for abundant species, the accuracy decreases with hybridization time, indicating short hybridization time will lead to better results. However, the concentrations of the double strands formed as a result of hybridization of single strands in solution to the immobilized strands (shown in Figure 2b for [SB'] as a function of hybridization time) increase with time for approximately the first 15 hours. The fluorescent intensities detected when the microarray slide is scanned is proportional to the concentration of these double stranded species. Hence the hybridization time has to be long enough to obtain intensities sufficiently above background levels for accurate measurement.
Figure 2 Effect of hybridization time on α and [SB'] for different levels of RNA abundance and differential expression. 2a) The variation in α with the hybridization time is shown for 8 different cases of abundance level and differential expression listed in Table 2. The parameters used are kf = 106 M-1s-1, kf-bound = 106 M-1s-1, kb = 0 s-1, initial genomic DNA concentration = 1 pM, bound strand concentration = 106 pM, transport rate = 1 s-1 2b) Change in concentration of SB', the intensity corresponding to the cDNA channel, with time for rare and intermediate species. □ S0 = 0.1 pM, ○ S0 = 1 pM, * S0 = 2 pM
Table 2 List of the combinations of abundance levels and differential expression ratios corresponding to biologically realistic scenarios used in simulations
C1 (pM) C2 (pM) Differential expression Comment
0.1 0.2 2 Rare species upregulated 2 fold
1 2 2 Intermediate species upregulated 2 fold
20 10 2 Abundant species downregulated 2 fold
0.1 1 10 Rare species upregulated 10 fold
1 10 10 Intermediate species upregulated 10 fold
20 200 10 Abundant species upregulated 10 fold
0.1 10 100 Rare species upregulated 100 fold
1 100 100 Intermediate species upregulated 100 fold
The model prediction that using genomic DNA as a reference can provide an accurate measurement in a microarray assay was verified experimentally. The transcript levels of two samples were assayed using both direct cDNA: cDNA hybridization (cDNA1:cDNA2) as well as by hybridizing to genomic DNA (cDNA:gDNA). This ratio of the two samples (γ1/γ2) was obtained by dividing the ratios obtained from those two cDNA samples ([cDNA1/gDNA])/ ([cDNA2/gDNA]) (Additional file 1) Microarray data can be found in the Supplementary material. Figure 3 shows a scatter plot of the relative transcript level obtained by these two methods. The ratio obtained from indirect comparison is within 1.5 fold of that obtained from direct comparison for 91% of genes. Out of the remaining 9% genes, 81% have a ratio obtained from direct comparison within 1.5-fold. For 99.3% of all the genes, the ratio obtained from indirect comparison is within 2 fold of that obtained from direct comparison. 70% of the remaining 0.7% have a ratio obtained from direct comparison within 1.5-fold. Thus, in general, the ratio obtained from hybridizations using genomic DNA as a reference is consistent with those obtained from direct cDNA: cDNA hybridization. As can be seen in Figure 3, this is true over a large range of differential expression (128-fold). Also, since total RNA samples containing a wide range of transcript abundance levels were used in this experiment, the dataset demonstrates that the accuracy of the assay is maintained over all mRNA expression levels.
Figure 3 Scatter plot showing log2 transformed ratios obtained from direct cDNA: cDNA hybridization and indirect comparison using genomic DNA as a reference. cDNA: cDNA hybridization was carried out using two RNA samples isolated from S. coelicolor mycelia obtained from liquid culture at early (Sample 1) and late (Sample 2) growth stages. For the indirect comparison using genomic DNA, each of the two samples was hybridized with genomic DNA.
Effect of genomic DNA concentration
In previous investigations several concentrations of the genomic DNA ranging from 1 to 35 pM have been used [7,23]. To examine the effect of genomic DNA concentrations on the microarray assay, we carried out simulations for a range of genomic DNA (1 to 35 pM) and cDNA concentrations (0.1 to 20 pM) with different degrees of downregulation (2, 10 and 100-fold). α is within 6% of unity for all values of genomic DNA concentrations examined (data not shown). Thus, the assessment of the ratio between the transcript levels of a gene in two samples does not vary significantly with the genomic DNA concentration used in the experiment.
However, in a microarray experiment, accurate assessment of the ratio of expression levels can be attained only if the fluorescent intensities can be accurately measured. Spots with lower intensities and closer to the background level are prone to increased noise interference and decreased accuracy. At higher concentration of genomic DNA, the intensity of the sample channel decreases. This effect is more profound for rare species, compounding the problem of their low intensity. This leads to the need to carefully select the genomic DNA concentration for achieving intensities which are significantly above background levels.
Effect of bound strand concentration
In the preparation of microarray slides the amount of DNA immobilized on the slide varies. Many factors, including differences in the amount of DNA deposited, spot morphology, DNA retention in the spotting and post-processing procedure contribute to such variation. In cDNA:cDNA hybridization, the ratio of transcript levels from two different samples is not affected by the amount of DNA immobilized on the slides, since the same transcript species from both samples are affected to the same extent. In contrast, when cDNA:gDNA is used, variation in the amount of immobilized DNA may exert a different effect on different samples. We simulated the effect on α of this changing bound strand concentration by using bound strand concentrations of 10-4 and 10-8 M for all cases listed in Table 2. This range spans values from 100-fold higher to 100-fold lower than the estimated bound strand concentration. All other parameters are same as those used for the simulations in Figure 2. The findings are similar to the plot shown in Figure 2a (results not shown). Thus, the variation in the amount of DNA immobilized on slides does not significantly affect the accuracy of the assay. However, a lower bound strand concentration does result in reduced concentrations of the double strands formed, and hence lower intensities on the microarray. As an illustration, when the bound strand concentration is lowered from 10-6 M to 10-8 M, the intensity of the sample channel (corresponding to [SB']) decreases 24-fold for rare species, 25-fold for intermediate species and 38-fold for abundant species.
Accuracy affected by transcript abundance at low hybridization rate
An uncertainty in the simulation is the effect of reaction rate constants; especially, the effect of change in the relative magnitude of forward rate constants for hybridization between mobile and mobile, and mobile and bound strands, is further investigated. To evaluate this, kf-bound was varied 100-fold from 106 M-1s-1 to 104 M-1s-1 and the results are shown in Figure 4. For all values of kf-bound, α is within 21% of unity for all the RNA abundance classes and all differential expression ratios examined. In general, the assay is more accurate for rare and intermediate species (error in α within 3% over the entire range of kf-bound tested) as compared to abundant species and for lower differential expression ratios compared to higher ratios. For the abundant species, as the forward rate constant for hybridization between mobile and bound species (kf-bound) is reduced, the accuracy of the assay decreases. The model predicts that the highest error (a ~1.21) will be observed for abundant species with high differential expression ratios.
Figure 4 Effect of a decreased rate of hybridization between mobile strands and immobilized strands for different differential expression ratios and RNA abundance levels on α. □ Differential expression = 2 fold, × Differential expression = 10 fold, Δ Differential expression = 100 fold. All other parameters are same as in Figure 2.
The simulation results presented above illustrate that the accuracy index α is most sensitive to the rate constant of the forward reaction for hybridization between mobile and bound species (kf-bound) and the deviation of α from unity is highest for abundant species in the sample. The ratio determined in the hybridization assay is . Ideally, this should reflect the true ratio . The ratio [SB']/ [RB'] is always close to S0/R0 since both S and R hybridize with the same species: R' in solution and B' on the surface. This is because in our analysis, we assume that S and R have identical reaction kinetics, they are thus both stoichiometrically and kinetically indistinguishable from each other. Therefore, the deviation in [RB']/ [R'B] from R0/R0' (= 1) governs the deviation in γ from the true ratio. Since the reaction is irreversible, for [R'B] to be equal to [RB'], the product [R'] [B] should be equal to [R] [B']. Let εS be the amount of S reacted with R' and εR be the amount of R reacted with R'. Since the stoichiometric ratios for reactions of S with R' and R with R' are both 1, R' = -εR - εS and R = R0 - εR. Under conditions where S0 is low (for rare and intermediate species), εS is relatively small and R ≈ R'. However, for abundant species, εS is significant and as a result, R/R' is greater than R0/. Also, since
S+B' → SB'
R+B' → RB'
R'+B → R'B
B' gets consumed more than B. Again, for rare and intermediate abundance species, this difference in B and B' is insignificant since the bound strands are in excess and hence the assay works very well for rare and intermediate species. The difference between B and B' is significant for abundant species (S0 is large) for which B/B' is greater than B0/B0'. The exact deviation in B/B' and R/R' depends on the rate constants for hybridization and hence it is not surprising that the rate constants have a major effect on hybridization results. This is illustrated by values for B, B', R and R' for S0 = 20 pM, R0 = = 1 pM and kf-bound = 106 M-1 s-1 and 104 M-1 s-1, at the end of 24 hours of hybridization time, presented in Table 3. The products [R] [B'], [R'] [B] and ratios [B]/ [B'] and [R]/ [R'] are also calculated. Thus, for the larger value of kf-bound, [R'] [B] ≈ [R] [B'], which is not the case for the lower values of kf-bound. This means that factors, such as mixing conditions in the sample chamber, which affect reaction kinetics, will affect the accuracy of the assay.
Table 3 An illustration of concentrations of different species at a hybridization time of 24 hours. S0 = 10 pM, R0 = = 1 pM
kf-bound M-1 s-1 [B] pM [B'] pM [R] pM [R'] pM [R] [B'] pM2 [R'] [B] pM2 [B]/ [B'] [R]/ [R']
106 9.6 × 105 4.6 × 105 0.63 0.27 2.9 × 105 2.5 × 105 2.09 2.33
104 1.0 × 106 9.9 × 105 0.94 0.40 9.3 × 105 4.0 × 105 1.01 2.35
One implication of the above discussion is that if single stranded species are used for immobilization (only B' immobilized, [B] = 0), as are used in spotted oligo-arrays, the assay will be robust to a wide range of hybridization conditions. This is also seen from our simulations where for all the conditions discussed in this paper, α is equal to one if only sense strands are immobilized on the microarray slide, a situation encountered during the use of oligonucleotide spotted arrays.
Conclusion
The use of genomic DNA as a reference is useful to assess the expression levels of a large array of genes among different samples. We have developed a kinetic model to predict the effect of using genomic DNA in the microarray assay under a wide range of conditions. The model predicts that the assay can accurately estimate the relative concentrations for a wide range of initial cDNA concentrations and ratios from hybridizations using genomic DNA as a reference will correspond to true ratios under all conditions if single stranded oligonucleotide microarrays are used.
The model also serves as a useful tool to predict the performance of such assays under varying conditions that are otherwise difficult to carry out experimentally. Despite a number of publications on its application, the use of genomic DNA as a reference for microarray assay is still not wide spread. We carried out this study to verify on a theoretical basis the validity of this approach and the results are indeed reassuring. We expect that the use of genomic DNA as reference will accelerate especially for comparative transcriptome analysis involving a wide range of samples from different sources.
Methods
Genomic DNA extraction
Streptomyces coelicolor A3(2) M145 spores were inoculated into complex media YEME. On reaching stationary phase, mycelia were harvested by centrifugation at 4000 × g and were used to isolate genomic DNA using the Kirby Mix method [24].
RNA extraction
Two RNA samples were isolated from S. coelicolor mycelia obtained from liquid culture at early (Sample 1) and late (Sample 2) growth stages. Mycelia samples were ground in a mortar in the presence of liquid nitrogen and then lysed by addition of RLT buffer from the RNeasy Mini Kit (Qiagen Inc., Valencia, CA). Total RNA was then isolated from the lysate using the RNeasy Mini Kit according to the manufacturer's protocol.
Microarray hybridization
S. coelicolor microarray containing duplicate spots representing 90% of the genes in the genome was used for hybridizations to compare RNA from Samples 1 and 2. The construction of the microarray is described on our website at .
cDNA: cDNA hybridization
10 μg of total RNA was used for each sample as starting material. Total RNA was reverse transcribed into cDNA incorporating aa-dUTP (Ambion, Austin, TX) and then labeled with Cy3 (Amersham Biosciences, Piscataway, NJ) or Alexa647 (Invitrogen, Carlsbad, CA).
Data presented here is an average from four replicate hybridizations. For two hybridizations, Sample 1 was labeled with Cy3 and Sample 2 with Alexa647, while for the other two hybridizations, the dyes were reversed with Sample 1 labeled with Alexa647 and Sample 2 with Cy3.
Genomic DNA hybridization
Hybridization with genomic DNA (gDNA) was carried out using 10 μg total RNA and 200 ng genomic DNA. Genomic DNA was nebulized to the length range of 500 bp to 1 kb. A nebulizer containing 2 ml of buffered genomic DNA solution (approximately 1 mg) containing 40% glycerol was placed in an ice-bath and was subjected to nitrogen gas at a pressure of 25 psi for 3 minutes. The resulted DNA fragments were purified by ethanol precipitation and were resuspended to a concentration of about 1 μg/μl. The fragmented genomic DNA was then labeled with Cy3 dye using Label IT® Cy™3 Labeling Kit (Mirus, Madison, WI). The labeling reaction consisted of 20% Label IT Reagent and 1 μg genomic DNA in 7 μl reaction volume. The reaction was incubated at 37°C for 3 hours and the labeled genomic DNA was purified with MinElute PCR purification kit (Qiagen Inc., Valencia, CA) as per the manufacturer's instructions. For hybridization with genomic DNA, cDNA was labeled with Alexa647 (Invitrogen, Carlsbad, CA). Four replicate hybridizations were performed and ratios obtained from the four hybridizations were averaged as described below.
All hybridizations were carried out at 50°C for 16 hours. Details of all protocols are available on our website at or are available as supplimentary material (see Additional file 2). Microarray slides were scanned after hybridization and washing using ScanArray (PerkinElmer, Boston, MA) and the images were quantified using GenePix Pro 5.1 (Axon Instruments, Union City, CA). The median intensity for each spot was used for further analysis.
Data analysis
Four replicate hybridizations were performed for each experiment with the following pairs of fluorescently labeled samples: cDNA1-gDNA, cDNA2-gDNA and cDNA1-cDNA2. Thus, for each experiment 8 replicate data points were obtained for each gene. The median intensity of pixels within a spot was used for analysis. Bad spots were filtered out based on the following criteria: 1) Spots flagged based on visual inspection during image analysis; 2) Spots with diameter was less than 70 μm; 3) Spots where the intensity of both channels was less than 200. The remaining 'good' spots were then normalized. LOWESS, a non-linear normalization algorithm from the commercial software GeneSpring (Silicon Genetics, Redwood City, CA), was used for cDNA:cDNA hybridization and linear normalization was used for cDNA:gDNA hybridization. For the linear normalization, we linearly scale the intensities of all the spots within each channel so that the sum of intensities of all spots in one channel is equal to 40,000,000. For all three experiments, after normalization, the average and standard deviation (SD) of the log2 transformed ratio for all the replicate spots for each gene was calculated. Spots outside the range [mean - 1.5 SD, mean + 1.5 SD] were considered outliers and therefore discarded. The average of the log2 transformed ratios for the remaining spots was calculated.
Authors' contributions
MG developed the model, carried out simulations, analyzed results and drafted the manuscript. WL carried out the microarray hybridization experiments. CG participated in model development and analysis of results. VK supervised the project. WSH conceived of the study, participated in its design and analysis and supervised the project. All authors participated in writing the manuscript and approved the final manuscript.
Supplementary Material
Additional File 1
Microarray data in a tab-delimited text format with arbitrary gene identifiers for the S. coelicolor genes.
Click here for file
Additional File 2
Details of the microarray hybridization protocols used in PDF format.
Click here for file
Acknowledgements
This work was supported in part by a grant from NIH (GM55850). The bioinformatics support was provided by the University of Minnesota Supercomputing Institute.
==== Refs
Schena M Shalon D Davis RW Brown PO Quantitative monitoring of gene expression patterns with a complementary DNA microarray Science 1995 270 467 470 7569999
Brown PO Botstein D Exploring the new world of the genome with DNA microarrays Nat Genet 1999 21 33 37 9915498 10.1038/4462
Kerr MK Churchill GA Experimental design for gene expression microarrays Biostatistics 2001 2 183 201 12933549 10.1093/biostatistics/2.2.183
Townsend JP Multifactorial experimental design and the transitivity of ratios with spotted DNA microarrays BMC Genomics 2003 4 41 14525623 10.1186/1471-2164-4-41
Yang YH Speed T Design issues for cDNA microarray experiments Nat Rev Genet 2002 3 579 588 12154381
Kim H Zhao B Snesrud EC Haas BJ Town CD Quackenbush J Use of RNA and genomic DNA references for inferred comparisons in DNA microarray analyses Biotechniques 2002 33 924 930 12398202
Talaat AM Howard ST Hale Wt Lyons R Garner H Johnston SA Genomic DNA standards for gene expression profiling in Mycobacterium tuberculosis Nucleic Acids Res 2002 30 e104 12384606 10.1093/nar/gnf103
Williams BA Gwirtz RM Wold BJ Genomic DNA as a cohybridization standard for mammalian microarray measurements Nucleic Acids Res 2004 32 e81 15190126 10.1093/nar/gnh078
Pearson BM Pin C Wright J I'Anson K Humphrey T Wells JM Comparative genome analysis of Campylobacter jejuni using whole genome DNA microarrays FEBS Lett 2003 554 224 230 14596944 10.1016/S0014-5793(03)01164-5
Kato-Maeda M Rhee JT Gingeras TR Salamon H Drenkow J Smittipat N Small PM Comparing genomes within the species Mycobacterium tuberculosis Genome Res 2001 11 547 554 11282970 10.1101/gr.166401
Hu DG Webb G Hussey N Aneuploidy detection in single cells using DNA array-based comparative genomic hybridization Mol Hum Reprod 2004 10 283 289 14985479 10.1093/humrep/gah038
Gadgil C Yeckel A Derby JJ Hu WS A diffusion-reaction model for DNA microarray assays J Biotechnol 2004 114 31 45 15464596 10.1016/j.jbiotec.2004.05.008
Gillespie D Spiegelman S A quantitative assay for DNA-RNA hybrids with DNA immobilized on a membrane J Mol Biol 1965 12 829 842 4955314
Gerhold D Rushmore T Caskey CT DNA chips: promising toys have become powerful tools Trends Biochem Sci 1999 24 168 173 10322428 10.1016/S0968-0004(99)01382-1
Ermolaeva OD Wagner MC SUBTRACT: a computer program for modeling the process of subtractive hydridization Comput Appl Biosci 1995 11 457 462 8521056
Craig ME Crothers DM Doty P Relaxation kinetics of dimer formation by self complementary oligonucleotides J Mol Biol 1971 62 383 401 5138338 10.1016/0022-2836(71)90434-7
Wilkins Stevens P Henry MR Kelso DM DNA hybridization on microparticles: determining capture-probe density and equilibrium dissociation constants Nucleic Acids Res 1999 27 1719 1727 10076004 10.1093/nar/27.7.1719
Stillman BA Tonkinson JL Expression microarray hybridization kinetics depend on length of the immobilized DNA but are independent of immobilization substrate Anal Biochem 2001 295 149 157 11488616 10.1006/abio.2001.5212
Ferrari ME Bloomfield VA Scattering and diffusion of mononucleosomal DNA: effect of counterion valence and salt and DNA concentration Macromolecules 1992 25 5266 5276 10.1021/ma00046a025
Goinga HT Pecora R Dynamics of low molecular weight DNA fragments in dilute and semidilute solutions Macromolecules 1991 24 6128 6138 10.1021/ma00023a013
Tinland B Pluen A Strum J Weill G Persistence length of single-stranded DNA Macromolecules 1997 30 5763 5765 10.1021/ma970381+
Velculescu VE Zhang L Zhou W Vogelstein J Basrai MA Bassett DE JrHieter P Vogelstein B Kinzler KW Characterization of the yeast transcriptome Cell 1997 88 243 251 9008165 10.1016/S0092-8674(00)81845-0
Wei Y Lee JM Richmond C Blattner FR Rafalski JA LaRossa RA High-density microarray-mediated gene expression profiling of Escherichia coli J Bacteriol 2001 183 545 556 11133948 10.1128/JB.183.2.545-556.2001
Kieser T Bibb MJ Buttner MJ Chater KF Hopwood DA Practical Streptomyces Genetics 2000 Norwich: The John Innes Foundation
| 15877823 | PMC1142311 | CC BY | 2021-01-04 16:32:46 | no | BMC Genomics. 2005 May 8; 6:66 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-66 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-701588514610.1186/1471-2164-6-70Research ArticlePigs in sequence space: A 0.66X coverage pig genome survey based on shotgun sequencing Wernersson Rasmus [email protected] Mikkel H [email protected]ørgensen Frank G [email protected] Jan [email protected] Frank [email protected]ærfeldt Hans-Henrik [email protected] Ole F [email protected] Thomas [email protected]øj Henrik [email protected] Ami [email protected] Jun [email protected] Bin [email protected] Songnian [email protected] Wei [email protected] Wei [email protected] Gane KS [email protected] Jun [email protected] Jian [email protected] Christian [email protected] Merete [email protected] Søren [email protected] Huanming [email protected] Lars [email protected] Center for Biological Sequence Analysis, Technical University of Denmark, Lyngby, Denmark2 Bioinformatics Research Center, University of Aarhus, Aarhus, Denmark3 Division of Genetics, The Royal Veterinary and Agricultural University, Copenhagen, Denmark4 Department of Animal Breeding and Genetics, Danish Institute of Agricultural Sciences, Foulum, Denmark5 Institute of Human Genetics, University of Aarhus, Aarhus, Denmark6 Beijing Genomics Institute, Beijing, China2005 10 5 2005 6 70 70 13 12 2004 10 5 2005 Copyright © 2005 Wernersson et al; licensee BioMed Central Ltd.2005Wernersson 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
Comparative whole genome analysis of Mammalia can benefit from the addition of more species. The pig is an obvious choice due to its economic and medical importance as well as its evolutionary position in the artiodactyls.
Results
We have generated ~3.84 million shotgun sequences (0.66X coverage) from the pig genome. The data are hereby released (NCBI Trace repository with center name "SDJVP", and project name "Sino-Danish Pig Genome Project") together with an initial evolutionary analysis.
The non-repetitive fraction of the sequences was aligned to the UCSC human-mouse alignment and the resulting three-species alignments were annotated using the human genome annotation. Ultra-conserved elements and miRNAs were identified. The results show that for each of these types of orthologous data, pig is much closer to human than mouse is. Purifying selection has been more efficient in pig compared to human, but not as efficient as in mouse, and pig seems to have an isochore structure most similar to the structure in human.
Conclusion
The addition of the pig to the set of species sequenced at low coverage adds to the understanding of selective pressures that have acted on the human genome by bisecting the evolutionary branch between human and mouse with the mouse branch being approximately 3 times as long as the human branch. Additionally, the joint alignment of the shot-gun sequences to the human-mouse alignment offers the investigator a rapid way to defining specific regions for analysis and resequencing.
==== Body
Background
The domesticated pig (Sus scrofa) is an obvious choice for genome sequencing, because of its important economic value for meat production and its relevance to biomedical research. The evolutionary position of the pig as an artiodactyl, where no other large scale sequencing efforts have so far been published, makes it valuable for comparative genomics.
Comparative vertebrate genome analysis – e.g. with the aim of understanding evolutionary pressures on the human sequence – is most cost-effective with relatively low coverage, genome wide sequencing of species at different evolutionary distances [1,2]. The mouse and rat genomes [3,4] and the chimpanzee genome [5] have offered sequences, which are evolutionary quite distant and very close to the human genome sequence, respectively. The evolutionary close chimpanzee sequence can pinpoint significant recent changes in genes, but is not efficient for identification of important regions by comparative approaches because many regions are extremely conserved by chance effects and differences in mutation rates over the genome. The more distantly related rodent sequences are useful in the search for conserved regions of biological importance. However, more species are needed, and an artiodactyl like pig is an obvious choice for the following reasons. 1), By comparing rodents and human, it is not possible to determine whether observable differences, e.g. difference in isochore structure, are mainly due to changes in the rodent or primate lineage since their divergence (approximately 90 million years ago, see 6); 2) Rodents have only a subset of the biological functions important to humans. By including pig additional functions will be covered; 3) Even though rodents, artiodactyls and primates diverged at approximately the same time [6,7], molecular evolution has been faster in the rodent branch, thus the pig is expected to be closer in sequence to human than mouse is. The importance of these points is generally appreciated and comparative genomics sequencing initiatives focusing on restricted regions of the genomes have recently provided much insight [e.g. [1,2]].
The Chinese-Danish pig sequencing consortium has generated about 3.84 million high quality sequences from 5 pig breeds. The present study releases these data and reports an initial evolutionary analysis which confirms that pig and human are closer in sequence space and quantifies the rates of evolution in the pig, rodent and human lineages for various categories of the genome sequences.
Results
Table 1 shows the amount of high quality sequence obtained from 5 pig breeds (NCBI Trace repository under center name "SDJVP", and project name "Sino-Danish Pig Genome Project"). The average trimmed length of the ~3.84 million sequences was 543 base pairs, yielding a total of 2.1 billion base pairs, equivalent to 0.66X coverage of redundancy of the 3.15 billion base pair pig genome. It is expected that 1-(1-543/3.15 × 109)^3.84 × 106 = 48% of the pig genome sequence has been hit at least once by this sequencing project. The low coverage prevents making a real assembly of the pig sequences and, thus, the contig coverage is not estimated. The analyses are therefore based on a very large number of short alignments. Repeatmasking (supplementary Table 1) masked 36% of all base pairs. The distribution of repeat types is overall very similar to what is observed in human, except for the expected absence of Alu-elements (Additional file 1). Overall, 38% of the coding fraction of the human-mouse alignment, 38% of the 5' UTR, 33 % of the 3' UTR, 23% of the intron region and 24% of the intergenic region could be expanded to a three-species alignment with the addition of the pig reads. This coverage of the human-mouse alignment by the pig genome sequences was close to our prior expectation. Since only 48% of the base pairs in the pig genome are expected to have been hit, we would only expect to hit at most 48% of the human-mouse alignment, assuming perfect conservation. However, in practice there is some lack of power in BLAST due to the fragmented nature of the pig shotgun reads (being fragmented even more by the repeatmasking), and we expect that some of the human-mouse alignment has no longer an orthologues region in the pig genome. For the non-coding regions, the coverage of the human-mouse alignment by the pig genome sequences is lower than for the coding regions, but this may be explained by lower selective constraints and a much higher rate of insertions-deletions in these regions.
Table 1 Overview of the number of raw reads generated from each breed.
Breed Number of high quality reads Number of bases
Hampshire 707,281 363,550,668
Yorkshire 1,204,666 652,086,833
Landrace 650,609 342,562,503
Duroc 1,015,722 574,663,060
ErHuaLian 256,993 150,835,661
Total 3,835,271 2,083,698,725
The alignments were used to generate the phylogenetic trees in Figure 1. As the pig, mouse and human lineages are believed to have diverged at approximately the same time, the trees allow for separate studies of evolution on the human and mouse branches since the divergence of the two species (the root). Due to a generally lower rate of nucleotide substitutions in the pig and human lineages, the porcine sequences are more similar to the human than to the mouse sequences. Overall, the exonic sequences show the slowest evolution, followed by 5' UTR, 3'UTR, intergenic and intronic regions, reflecting different levels of selective constraint on these domains.
Figure 1 Evolutionary distances between mouse, pig and human for conserved sequences divided into functional classes using the annotation of the human genome. Branch lengths are estimated using the HKY substitution model with gamma correction [12].
Ultra-conserved sequences
By aligning the set of ultra-conserved regions against the pig genome reads using BLAST, we were able to find 239 of the 481 known regions reported in Bejerano et al. (2004) with a significant hit of at least 150 bp. Only 12 of these regions were less than 98% conserved (85–97% identity). This result agrees very well with the expected 48% of the pig genome being covered and the assumption that these regions are very well conserved within Mammalia.
By aligning the pig shotgun data against all human transcripts (NCBI build 34) we found 758 completely conserved sequences exceeding 200 bp in length. Of these, 41 were also found to be completely conserved in the mouse genome, while 590 were less conserved (more than 95% identity over at least 80% of the length). BLASTing human transcripts vs. the fully assembled mouse genome (NCBI build 32), we found 2709 ultra-conserved regions. When aligning this set of sequences against the artificially fragmented mouse genomic dataset using BLAST it was only possible to classify 664 (24.5%) as ultra-conserved – less than the 758 elements found in the human-pig comparison.
miRNA
The set of pig miRNAs (Additional file 2) was compared to human and mouse and it was possible to obtain 50 three-way alignments. The evolutionary tree in Figure 1 was constructed using the HKY+gamma model from these alignments with gap positions removed. By construction, the miRNAs are more conserved than even the protein coding sequences, but with pig and human being phylogenetically closest. For the 50 triple-alignments, we obtained 25 cases where pig is closer to human than to mouse, 2 cases where pig is closer to mouse than to human, and 23 cases where pig is equally distant to human and mouse.
GC content
The intra-genomic variation in GC content among the individual alignments reflects the isochore structure of the genome. Thus, from the three species alignments, we calculated the GC content for each functional sequence class for each aligned fragment. For a given type of sequence, only alignments having more than 40 nucleotides of the specific type were used. Table 2 shows that the mean GC content is similar among the three species. The variance among alignments in GC content is generally lower in mouse than in pig and human, but mostly so for coding sequences, followed by the UTR and intron regions (Table 2). Figure 2 shows the distribution of GC% for the coding alignments. The reduced variability in GC content in mouse compared to human has been shown previously, e.g. Figure 8a in [4]. The results presented here suggest a very similar pattern in human and pig.
Table 2 Average GC content and the variance among alignments exceeding 40 bp for each species and each functional category. Variance is standardized to the variance observed in the human sequence.
Mean GC content Variance GC content
Type of sequence Human Mouse Pig Human Mouse Pig
Intron 0.390 0.413 0.407 1 0.82 1.02
Coding 0.487 0.500 0.496 1 0.69 1.01
3' UTR 0.404 0.426 0.418 1 0.77 1.03
5' UTR 0.595 0.593 0.592 1 0.81 0.92
Intergenic 0.384 0.399 0.396 1 0.91 1.01
Figure 2 The distribution of GC content in exons for human, pig and mouse. Only alignments with more than 40 base pairs of exon sequence were used.
Discussion
Even though divergence between pig and human occurred approximately at the same time as the divergence between human and mouse, the pig sequence is much more similar to the human sequence. Thus, the availability of the pig sequence effectively subdivides the human-mouse evolutionary branch at a position closest to human. This implies that one can determine which changes occurred on the human and mouse branches, respectively, since their divergence. The results of the phylogenetic analysis show that the relative length of the mouse, human and pig branches are different for the different types of data. Intronic sites and intergenic sites show a similar pattern, which also closely resembles that of synonymous sites [see [7]], reflecting that purifying selection is probably weak and similar for these regions. It is noteworthy that intergenic branch lengths are slightly shorter than intronic ones (and both are shorter than synonymous sites, see [7]). This may reflect either 1) more selective constraints on intergenic sequences than intronic, i.e. parts of the annotated intergenic sequences are indeed genic, or 2) a bias in the construction of the human-mouse alignment that make it easier to align sequence close to the conserved exons so that somewhat more divergent intronic sequence can be detected. The differences in evolutionary rates among the three species are most likely attributable to differences in generation times since they diverged. Non-synonymous sites show shorter branches that are much more similar in lengths among the species [7]. This reflects purifying selection, which has been strongest in the mouse lineage, followed by pig and then human [20,21]. The reason for this may be the larger average population size in the mouse since divergence. The UTR regions shows more selective constraint than introns.
The similarity between human and pig adds to the recently reported ultra-conserved regions [14]. Pig and human share more ultra-conserved regions than human and mouse, and (correcting for the coverage in the data presented) virtually all of the ultra-conserved elements defined by Bejerano et al. [14] are also found in pig and therefore most likely in the artiodactyls lineage.
Variation in GC content along the genome (isochore structure) is more pronounced in primates than in rodents [4,5]. The present results extend these findings and put artiodactyls in line with primates – lending further support to the suggestion that isochore evolution in rodents deviates from all other lineages, possibly because of extensive genome rearrangements [22].
Conclusion
A 0.66X coverage pig genome survey is hereby released. Even though it is only a beginning, the data offer many analytical possibilities and should also stimulate the international initiatives to generate a complete draft of the pig genome. The initial analysis of the data adds to our understanding of the evolutionary relationships of humans, mice and pigs. Further comparative genomic studies and more detailed genetic analyses will greatly improve our ability to elucidate pig as well as human biology and medicine.
Methods
Generation of pig shot-gun sequences
Genomic DNA was extracted from blood samples from 5 different breeds of Sus scrofa (domestic pig): ErHuaLian, Duroc, Landrace, Yorkshire and Hampshire. Following mechanical shearing, DNA fragments (1–3 kb) were isolated by gel electrophoresis and cloned into the SmaI restriction site of the pUC18 plasmid using blunt-end ligation. After transformation into E. coli (strain DH5a) and selection on LB-plates, individual clones were picked for the library. The quality of a library was checked by sequencing a small number of plasmids, which were assembled by phrap and aligned by BLAST to validate the randomness of the library and the proportion of contaminations – including mitochondrial DNA, human fragments and vector sequences. If the library was eligible, plasmid DNA was extracted from the individual clones using the membrane-filter method (Millipore). The pig genomic inserts were sequenced using the M13 bidirectional sequencing primers on the MegaBACE1000 platform using ET dye terminator. The chromatograms were registered in a relational database tracking all generated data to eliminate duplicated work and check for errors. Using phred v. 0.020425.c with a quality cut-off of 0.05 and the -trim_alt parameter the chromatograms were traced and the resulting traces were masked for vector sequences using CROSSMATCH (0.990329). Resulting sequences were resubmitted to the relational database and prepared for submission. Following removal of contamination from vector and bacterial host, we retained 3,835,271 reads of at least 150 bp.
All sequences with trace files have been submitted to the Ensembl/NCBI Trace repository under the center name "SDJVP", and project name "Sino-Danish Pig Genome Project".
Repetitive elements
Standard masking of repetitive sequences was performed using Repeatmasker version 2004/03/06 with RepBase Update 8.12 with Sus scrofa as query species, using default settings.
Construction of three species alignments
The newest build of the human-mouse pairwise alignment (hg17/mm5) based on the improved blastZ algoritm [8] was downloaded from the UCSC genome browser [9] and the repeatmasked shotgun-sequences were BLAST'ed up against this alignment using Megablast [10] with the following settings: (w) word size = 12, (e) minimum e-value = 0.0001, (x) extension parameter = 50, (u) Repeatmasked query sequences = True. Resulting hits were then used as tags to build the alignment around (see below).
All query sequences with multiple blast hits on different parts of the human-mouse alignment (approximately 10%) were removed at this point and only the remaining pig sequences were used in the further analyses. This was done to conservatively eliminate paralogous hits. The individual blast hits were used to position the pig query sequences on a specific part of the human-mouse alignment. For each BLAST hit a region extending 300 base pairs in both directions was then realigned using DIALIGN version 2.2.1 with default parameters [11]. Regions that afterwards were not considered by DIALIGN to be aligned were removed at this point. Subsequently an annotation file (refGene.txt) of the human genome (hg17) was downloaded from the UCSC genome server and used to annotate each position in the alignment according to the following classes 1) protein-coding exon (with reading frame position), 2) intron, 3) 5' UTR, 4) 3' UTR, and 5) intergenic. Phylogenetic analyses were performed on each of these classes separately.
Estimation of phylogenetic trees
A combined alignment for each of the five functional classes was constructed by concatenating the many small three species alignments, and gap positions were removed. For each class of data (exon, 5'UTR, 3'UTR, intron and intergenic), the implementation of the HKY85+GAMMA model [12] in PAML v. 3.14 [13] was used to estimate the number of substitutions in each evolutionary branch.
Analysis of ultra-conserved regions
The data set containing the 481 ultra-conserved regions defined by Bejerano et al. [14] was downloaded and used for the first part of this analysis. We estimated how many of these regions that can be found in the pig genome shotgun sequences by a simple BLAST (blastn) approach retaining only hits of a length of 150 bp or more. Since these ultra-conserved regions are based on external data, artefacts due to the relatively low coverage of the pig genome sequences are not important here.
In the second part of the analysis (searching for novel ultra-conserved regions), it was important to address the artefacts that arise from searching for ultra conserved regions in a fragmented dataset. In order to do a comparable search for ultra-conserved regions against the mouse and pig genomes, we artificially created a mouse data set resembling the fragmented pig shotgun data set. Since the identification criteria we use is 100% conservation over 200 bp or more, the length of the individual read as well as the quality become a major factor. The fragmented mouse data set was generated by downloading all ~79 million reads of the NCBI trace database, shuffle these randomly, and pick single reads until the same amount of nucleotides was reached as in the pig genomic reads dataset. Only pig reads with length > 200 bp were used and the data set consisted of 2,034,999,640 bp from pig and 2,665,153 reads (2,034,999,649 bp) from mouse.
Construction of the miRNA data set
The reads were BLAST searched [15] (default options) against the miRNA hairpin database [16] and matches of at least 60 nucleotides clustered. The obtained set of 219 clusters contained redundancy as the same pig reads hit the same miRNAs from different species (human, mouse, etc.). This could be converted into a unique set of 68 clusters (each containing 2–5 reads), i.e. each type of miRNA is now only represented once. Each cluster was assembled with Cap3 [17] from TGICL [18]. As a result 84 contigs and singletons were obtained and BLAST searched back against the miRNA database. Matches with more than 95% identity in an alignment length larger than 95% of the miRNA hairpin length were selected. This resulted in 54 hairpins complying with the criteria defined in Ambros et al. [19]. In total 51 different mature sequences are covered in the 54 hairpins.
Authors' contributions
LB, HY, CB, MF, SB initiated and coordinated the project, JUW, BL, SH, WD, WL, JY, JIW, HY organized the sequencing work, FP, HHS, HH, AK cleaned up the data and created the database, RW, MHS, FGJ, JG, OFC, TM performed the analyses, RW, MHS, FG, OFC, JUW, GW, LB drafted the paper. All authors have read and approved the paper.
Supplementary Material
Additional File 1
Distribution of repetitive elements in the pig genome survey sequences. Data obtained from Repeatmasking of shotgun sequences
Click here for file
Additional File 2
Details of investigated miRNAs.
Click here for file
Acknowledgements
This study was supported by grants from: Danish National Committee for Pig Production, Danish Trans-Disciplinary Research Programme: "Comparative Genomics Focusing on Man and Pig", Danish National Research Foundation: "Danish Platform for Integrative Biology", The Danish Technical Research Council, Danish Center for Scientific Computing, Chinese Academy of Sciences, Chinese Ministry of Science and Technology, Natural Science Foundation of China. We acknowledge the Sino-Danish Pig Genome Consortium consisting of The Danish Veterinary and Agricultural University (KVL), Denmark, the Danish Institute of Agricultural Sciences (DIAS), Denmark, and the Beijing Genomics Institute/James D. Watson Institute of Genome Sciences (BGI/WIGS), China.
==== Refs
Blakesley RW Hansen NF Mullikin JC Thomas PJ McDowell JC Maskeri B Young AC Benjamin B Brooks SY Coleman BI Gupta J Ho SL Karlins EM Maduro QL Stantripop S Tsurgeon C Vogt JL Walker MA Masiello CA Guan X Bouffard GG Green ED An intermediate grade of finished genomic sequence suitable for comparative analyses Genome Research 2004 14 2235 2244 15479945 10.1101/gr.2648404
Thomas JW Touchman JW Blakesley RW Bouffard GG Beckstrom-Sternberg SM Margulies EH Blanchette M Siepel AC Thomas PJ McDowell JC Maskeri B Hansen NF Schwartz MS Weber RJ Kent WJ Karolchik D Bruen TC Bevan R Cutler DJ Schwartz S Elnitski L Idol JR Prasad AB Lee-Lin SQ Maduro VV Summers TJ Portnoy ME Dietrich NL Akhter N Ayele K Benjamin B Cariaga K Brinkley CP Brooks SY Granite S Guan X Gupta J Haghighi P Ho SL Huang MC Karlins E Laric PL Legaspi R Lim MJ Maduro QL Masiello CA Mastrian SD McCloskey JC Pearson R Stantripop S Tiongson EE Tran JT Tsurgeon C Vogt JL Walker MA Wetherby KD Wiggins LS Young AC Zhang LH Osoegawa K Zhu B Zhao B Shu CL De Jong PJ Lawrence CE Smit AF Chakravarti A Haussler D Green P Miller W Green Comparative analyses of multi-species sequences from targeted genomic regions Nature 2003 424 788 793 12917688 10.1038/nature01858
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 Mouse Genome Sequencing Consortium Initial sequencing and comparative analysis of the mouse genome Nature 2002 420 520 562 12466850 10.1038/nature01262
Rat Genome Sequencing project consortium Genome sequence of the Brown Norway rat yields insight into mammalian evolution Nature 2004 428 493 521 15057822 10.1038/nature02426
The international Chimpanzee Chromosome 22 Consortium DNA sequence and comparative analysis of chimpanzee chromosome 22 Nature 2004 429 382 388 15164055 10.1038/nature02564
Springer MS Murphy WJ Eizirik E O'Brien SJ Placental mammal diversification and the Cretaceous-Tertiary boundary Proc Natl Acad Sci USA 2003 100 1056 1061 12552136 10.1073/pnas.0334222100
Jørgensen FG Hobolth A Jensen HH Bendixen C Fredholm M Schierup MH Comparative analysis of protein coding sequences from Human, Mouse and the domesticated Pig BMC Biology 2005 3 2 15679890 10.1186/1741-7007-3-2
Schwartz SW Kent J Smit A Zhang Z Baertsch R Hardison RC Haussler D Miller W Human-Mouse Alignments with BLASTZ Genome Res 2003 13 103 107 12529312 10.1101/gr.809403
Kent WJ Sugnet C Furey T Roskin K Pringle T Zahler A Haussler D The human genome browser at UCSC Genome Res 2002 12 996 1006 12045153 10.1101/gr.229102. Article published online before print in May 2002
Zhang Z Schwartz S Wagner L Miller W A greedy algorithm for aligning DNA sequences J Comput Biol 2000 7 203 214 10890397 10.1089/10665270050081478
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
Hasegawa M Kishino H Yano T Dating of the human-ape splitting by a molecular clock of mitochondrial DNA J Mol Evol 1985 22 160 174 3934395
Yang Z PAML: A program package for phylogenetic analysis by maximum likelihood Comput Appl Biosci 1997 13 555 556 9367129
Bejerano G Pheasant M Makunin I Stephen S Kent WJ Mattick JS Haussler D Ultraconserved elementsin the human genome Science 2004 304 1321 1325 15131266 10.1126/science.1098119
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
Griffiths-Jones S The miRNA Registry Nucleic Acids Res 2004 D109 D111 14681370 10.1093/nar/gkh023
Huang X Madan A CAP3: A DNA sequence assembly program Genome Res 1999 9 868 877 10508846 10.1101/gr.9.9.868
Pertea G Huang X Liang F Antonescu V Sultana R Karamycheva S Lee Y White J Cheung F Parvizi B Tsai J Quackenbush J TIGR Gene Indices clustering tools (TGICL): a software system for fast clustering of large EST datasets Bioinformatics 2003 19 651 652 12651724 10.1093/bioinformatics/btg034
Ambros V Bartel B Bartel DP Burge CB Carrington JC Chen X Dreyfuss G Eddy SR Griffiths-Jones S Marshall M Matzke M Ruvkun G Tuschl T A uniform system for microRNA annotation RNA 2003 9 277 279 12592000 10.1261/rna.2183803
Ohta T An examination of the generation-time effect on molecular evolution Proc Natl Acad Sci USA 1993 90 10676 10680 8248159
Ohta T Synonymous and nonsynonymous substitutions in mammalian genes and the nearly neutral theory J Mol Evol 1995 40 56 63 7714912 10.1007/BF00166595
Belle EM Duret L Galtier N Eyre-Walker A The decline of isochores in mammals: an assessment of the GC content variation along the mammalian phylogeny J Mol Evol 2004 58 653 660 15461422 10.1007/s00239-004-2587-x
| 15885146 | PMC1142312 | CC BY | 2021-01-04 16:39:53 | no | BMC Genomics. 2005 May 10; 6:70 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-70 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-711588820010.1186/1471-2164-6-71Research ArticleA study of inter-lab and inter-platform agreement of DNA microarray data Wang Huixia [email protected] Xuming [email protected] Mark [email protected] Carole [email protected] Lei [email protected] Department of Statistics, University of Illinois at Urbana-Champaign, 101 Illini Hall, 725 South Wright Street, Champaign, Illinois 61820, USA2 W. M. Keck Center for Comparative and Functional Genomics, University of Illinois at Urbana-Champaign, 1201 W. Gregory Drive, Urbana, Illinois 61801, USA2005 11 5 2005 6 71 71 30 7 2004 11 5 2005 Copyright © 2005 Wang et al; licensee BioMed Central Ltd.2005Wang 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.
As gene expression profile data from DNA microarrays accumulate rapidly, there is a natural need to compare data across labs and platforms. Comparisons of microarray data can be quite challenging due to data complexity and variability. Different labs may adopt different technology platforms. One may ask about the degree of agreement we can expect from different labs and different platforms. To address this question, we conducted a study of inter-lab and inter-platform agreement of microarray data across three platforms and three labs. The statistical measures of consistency and agreement used in this paper are the Pearson correlation, intraclass correlation, kappa coefficients, and a measure of intra-transcript correlation. The three platforms used in the present paper were Affymetrix GeneChip, custom cDNA arrays, and custom oligo arrays. Using the within-platform variability as a benchmark, we found that these technology platforms exhibited an acceptable level of agreement, but the agreement between two technologies within the same lab was greater than that between two labs using the same technology. The consistency of replicates in each experiment varies from lab to lab. When there is high consistency among replicates, different technologies show good agreement within and across labs using the same RNA samples. On the other hand, the lab effect, especially when confounded with the RNA sample effect, plays a bigger role than the platform effect on data agreement.
==== Body
Background
Diversity of microarray data poses some unique and interesting questions on cross-experiment comparisons and the analysis tools needed for such comparisons. Since the invention of the microarray technology in 1995 [1], statistical methods and data mining techniques specific for microarray data have mushroomed [2], many of which have been packaged into commercial software such as GeneSpring and Spotfire. Such tools are useful for handling individual experiments, including quality control, significance testing, and clustering. However, researchers have questioned whether studies across different labs and technology platforms will have an acceptable level of agreement.
Possible incompatibility of results between similar microarray experiments is a major challenge that needs to be addressed, even though the data produced within a single experiment may be consistent and easy to analyze. Different labs produce microarray data in different ways using different technology platforms, such as Affymetrix GeneChip, spotted cDNA array, and spotted oligo array. Affymetrix GeneChip uses one fluorescent dye while the spotted array uses two fluorescent dyes in the experiments. Direct comparison of raw data obtained from different technologies may not be meaningful. Instead, the final form of the data is often presented as relative expression levels, mostly ratios of intensities, after some statistical treatments including filtering, normalization and model-based estimation. Experiments using different technologies require different protocols for analyzing the raw data to derive the ratios. Scientists have published microarray data in a variety of formats including raw intensities and ratios of intensities. Does it make a difference which technology platform is chosen? Can we make use of the studies from different platforms and labs? To answer these questions and provide some guidance for platform comparisons, we report on a comparative study of three different platforms. The experiment is a simple two-tissue comparison between mouse liver and spleen. We used previously published data sets from two different sources as well as new data sets produced in house. There are several similar studies published in recent years [3-8]. A noticeable difference of this study from earlier ones is that we considered lab as a major factor in the comparison. In addition, we compared three major types of technology platforms, namely Affymetrix GeneChip, spotted cDNA array and spotted long oligo array. This study aims to provide a basis for further development of methodologies for comparing microarray data across different experiments and for the integration of microarray data from different labs.
Results
Data collection
As summarized in Table 1, a total of five data sets were either collected from a public source or generated in house. The samples for the experiments were normal mouse liver and spleen RNA, which were purchased from Clonetech (Catalog No. 64042-1 liver; Catalog No. 64044-1 spleen) except for the data set GNF generated by Su et al. [9] at the Genomics Institute of the Novatis Research Foundation. Detailed sample descriptions for the GNF data can be found at . Two data sets were downloaded from the NCBI Gene Expression Omnibus , which were generated by Choi et al. at California Institute of Technology (Cal Tech) using Agilent oligo (GEO accession: GSE334) and cDNA arrays (GEO accession: GSE330), respectively. Two other data sets were generated at the Functional Genomics Unit at the W. M. Keck Center for Comparative and Functional Genomics at the University of Illinois using an in-house printed cDNA mouse array and Affymetrix mouse expression set 430A, and the data sets are available at . Another data set was downloaded from and it was generated using Affymetrix Murine Genome set U74Av2. The data set names (e.g., KC for the cDNA data set generated at the Keck Center at the University of Illinois) given in Table 1 will be used throughout the paper.
Table 1 Summary of data collection
Data Set Array Genes Sample Replicates Data type Platform Lab Data Source
KC CI 15K cDNA 15K Clonetech 4 Raw intensity cDNA Keck In house
KAV Affymetrix 430A 23K Clonetech 2 AV(Bioconductor) Affymetrix Keck In house
KLW Affymetrix 430A 23K Clonetech 2 Li and Wong Affymetrix Keck In house
KRMA Affymetrix 430A 23K Clonetech 2 RMA Affymetrix Keck In house
CC Riken16K cDNA by Agilent 16K Clonetech 3 Raw intensity cDNA Cal Tech NCBI GEO
CO Riken16K Oligo by Agilent 16K Clonetech 3 Raw intensity Oligo Cal Tech NCBI GEO
GNF Affymetrix U74Av2 12K In house 2 AV(MAS4.0) Affymetrix GNF expression.gnf.org
Consistency of replicates
One indication of data reliability is the consistency of replicates in a particular data set. We used kappa coefficients as well as the Pearson correlation coefficients and intraclass correlation coefficients on the replicates within each data set. Those measures set a benchmark against which the reliability of different platforms can be assessed; see Figure 1. The data set KC has four replicates from double spots of each gene on the array and from the dye swap. Therefore, we can do six pairwise comparisons of replicates. The data sets CC and CO have three replicates each; therefore, there are three pairwise comparisons. All Affymetrix data sets have two replicates and thus only one comparison. From Figure 1, we see that the replicates were quite consistent within each technology. The replicates in all the data sets showed pairwise Pearson correlation coefficients of 0.80 or higher, intraclass correlation coefficients of 0.77 or higher, and kappa coefficients of 0.43 or higher. The data from the Cal Tech (CC and CO) showed the highest agreement among the replicates, and the data from GNF and KAV showed a low level of agreement between replicates. These results suggest a lab effect in microarray data experiments.
Figure 1 Consistency of replicates.
For the KC data, we can see that two pairwise comparisons gave slightly higher agreement than the other four. It is, we believe, due to the double spots of each gene on the array. The two comparisons with higher agreement are the comparisons between the replicates within the slides.
Pairwise comparisons among data sets
Using the matched genes by common UniGene IDs, we compared different data sets in this study. Table 2 shows the Pearson correlation coefficients (PCC), intraclass correlation coefficients (ICC) and intra-transcript correlation coefficients (ITC) on log ratios, and the kappa coefficients (Kappa) for two-fold changes. The detailed descriptions of these measurements can be found in the "Methods" section. Figure 2 shows the histograms of the pairwise PCC, Kappa, ICC and ITC, respectively. With the exception of the GNF, most of the kappa coefficients are between 0.4 and 0.6, as compared to 0.6 for the replicates within the KC data set.
Table 2 Correlation coefficients for pairwise comparisons between data sets. Pearson correlation coefficients (PCC), kappa coefficients (Kappa), intraclass correlation coefficients (ICC) and intra-transcript correlation coefficients (ITC) for pairwise comparisons.
Comparisons No. of Matched Unigene IDs PCC Kappa ICC ITC
GNF vs. KC 1,838 0.590 0.327 0.693 0.748
GNF vs. CC 1,374 0.513 0.312 0.678 0.774
GNF vs. CO 1,914 0.633 0.365 0.707 0.729
GNF vs. KAV 2,058 0.727 0.452 0.686 0.724
GNF vs. KLW 3,295 0.640 0.374 0.681 0.690
GNF vs. KRMA 3,452 0.686 0.400 0.706 0.705
KC vs. CC 2,730 0.597 0.363 0.681 0.830
KC vs. CO 3,043 0.641 0.423 0.714 0.812
KC vs. KAV 2,964 0.747 0.523 0.726 0.908
KC vs. KLW 4,362 0.680 0.461 0.714 0.868
KC vs. KRMA 4,516 0.725 0.493 0.736 0.893
CC vs. CO 3,262 0.688 0.429 0.770 0.836
CC vs. KAV 2,285 0.708 0.461 0.746 0.859
CC vs. KLW 3,658 0.650 0.407 0.739 0.837
CC vs. KRMA 3,843 0.707 0.472 0.772 0.862
CO vs. KAV 3,001 0.806 0.555 0.781 0.865
CO vs. KLW 4,725 0.759 0.503 0.782 0.847
CO vs. KRMA 5,018 0.805 0.580 0.813 0.854
KAV vs. KLW 7,181 0.923 0.666 0.832 0.917
KAV vs. KLW 7,237 0.955 0.734 0.848 0.938
KLW vs. KRMA 14,130 0.921 0.732 0.765 0.971
Figure 2 Correlation coefficients for pairwise comparisons between data sets.
From Figure 2 and Table 2, we see that the rankings of the pairwise comparisons are almost the same across four different measures of agreement. All the measures involving GNF are at the low end of the comparisons. The same is observed from the sensitivity check, which was done by leaving out one data set at a time and recording the changes of ICC as shown in Table 3. Excluding GNF resulted in the largest increase in ICC.
Table 3 Sensitivity check of the overall comparison among all the data sets
ICC
Before leaving out 0.662
Leave out GNF 0.703
leave out KLW 0.650
leave out KC 0.663
leave out CC 0.684
leave out CO 0.670
In a recent study, Jarvinen et al. [4] compared Affymetrix GeneChip, commercial cDNA array and a custom cDNA array using the same RNA samples from human cancer cell lines. They found that the data were more consistent between two commercial platforms and less consistent between custom arrays and commercial arrays. Their conclusion is consistent with our findings. In our study, KC is a custom cDNA array, whereas CC and CO are commercial cDNA and oligo arrays, respectively. If we do not consider comparisons involving GNF, we found that KC_CC (shorthand for KC versus CC) and KC_CO comparisons were ranked at the bottom. The samples used in KC, CC, and CO were identical, so biological variability is not an issue here. The variability among those data sets was mainly due to technical factors, such as platforms and labs conducting the experiments. Jarvinen et al. [4] analyzed the experiments conducted in one lab, and therefore their study was mostly concerned with the platform difference. Another study by Culhane et al. [3] used the co-inertia analysis to compare overall expression profiles across different platforms. Their analysis could be used on matched genes, as well as on all the data from different platforms. While they considered the agreement for the overall expression profiles, we focused on the agreement in expressions at the gene level. Using the within-platform variability as a benchmark, we found that these technology platforms exhibited an acceptable level of agreement. For example, the overall comparison of all five data sets using KLW for the in-house Affymetrix gives ICC = 0.662, as compared to the ICC around 0.8 for replicates within each data set. These results indicate that the agreement of different technologies is decent.
Since we only used a subset of genes in the study due to the gene matching problem, we asked whether the gene-to-gene variation was impacted by the choice of the subset. A comparison of the box plots for the full data set and for the subset used in the overall comparison showed that the variation in the subset of genes is similar to that of the full data set. For example, Figure 3 and Figure 4 give the box plots for the full data set and the subset of CO, respectively. Therefore, we believe that a reasonable conclusion can be made based on the subset being used.
Figure 3 Boxplot of the full data set of CO, with 7,282 Unigene IDs.
Figure 4 Boxplot of the subset of CO, overlapped with the other 4 datasets, with 551 Unigene IDs.
Discussion
UniGene has been widely used to match genes on different microarrays. Using UniGene and sequence similarity, Thompson, et al. [10] reported a number of gene markers that showed platform-independent expression profile. When we matched genes from different arrays using the mouse UniGene IDs, we found that there were multiple gene IDs in an array corresponding to one UniGene ID. Those genes were considered as "duplicate" genes, which made the cross matching of genes more complicated. A common approach is to average the expressions of those "duplicate" genes; however, we considered these "duplicate" genes as replicates in the technology and lab comparisons. One observation we should make is that the variability among these "duplicate" genes can be large. For example, in the CC arrays, there are 11,301 genes (upon filtering), corresponding to 8,318 UniGene IDs. Among the 8,318 UniGene IDs, 1,708 of them have "duplicate" genes. For the 12 Unigenes that have 10 or more "duplicate" genes, the ICC of the replicates decreased from 0.99 to 0.86 due to gene matching by Unigene IDs. Note that gene matching was performed only for the comparisons across data sets, and that we did not use the UniGene IDs in measuring the consistency within the same data set. This means that the agreement measures we obtained across different data sets were expected to be slightly lower than those from the replicates, even if the actual agreement was the same within or between data sets. In earlier studies, such as the one reported by Jarvinen et al., it has been shown that using different clones on different arrays is a major factor for the discrepancies between platforms. Using UniGene IDs to match the clones on different platforms can be problematic and result in biased comparisons. Sequence validation of the clones in different arrays may help resolve some of the problems and make the data from different platforms and labs more comparable.
From the present study, we showed that the GNF data set had the lowest agreement with the other data sets. This difference is compounded with the facts that the consistency of replicates within the GNF data set is the lowest and the sample used to generate the data is different from those used by other labs. We believe that the technology platform plays a relatively minor role in the disagreement, but the variation introduced by sample differences is one of the major factors. It has been shown that the expression level can vary significantly between genetically identical mice [11]. Variation among different individuals can be a significant factor for sample differences. Our analysis also indicates that data generated from different labs may have different quality even among the replicates, and thus quality control is important.
In this study, we also found that the lab effect can be greater than the platform effect. As shown in Figure 5, the comparisons between two different technology platforms in the same lab (KC_KRMA and CC_CO) showed better agreement than between two labs using the same technology (KC_CC). We also showed that the different summarization methods for Affymetrix exhibited good agreement.
Figure 5 Comparisons between the same technology but different labs (KC_CC) and comparisons between different technologies in the same lab (KRMA_KC and CO_CC).
Obviously, the present study has limitations. The results were generated from a very limited number of data sets. Using UniGene IDs for gene matching across data sets can also be questioned. Further research is clearly needed to address these limitations.
Conclusion
In this paper, we aim to address several issues in comparing microarray data across different platforms and different labs. We demonstrated that the consistency of replicates in each experiment varied from lab to lab. With high consistency among replicates, different technologies seemed to show good agreement within and across labs using the same RNA samples. A closer look at the results indicated that the variability between two labs using the same technology was higher than that between two technologies within the same lab. The source of RNA samples can make a difference in microarray data, however in our present study we do not show conclusive results pertaining to possible sample or lab effects, because we did not have data collected from two different samples within one lab.
Methods
Data processing
For the spotted arrays (KC, CC, and CO), we used raw intensity data from both Cy5 and Cy3. We filtered out the non-expressive data points using median plus three times median absolute deviation (MAD, [12]) of the negative control genes as a criterion. We then performed global lowess normalization on each slide. For the KC data, we also performed paired-slide normalization following the method in Yang et al. [13] because of the dye swap in the experiment.
For the in-house Affymetrix data, we used three summarization methods to generate the probe set level signals. KAV and KRMA are based on the R package affy of Bioconductor using the average difference (AV) between Perfect Match (PM) and Mis-Match (MM) probe pairs, and the Robust Multi-Array Average (RMA) expression measure developed by Irizarry et al. [14], and KLW is the model-based expression indexes developed by Li and Wong [15]. The GNF Affymetrix data were available to us only in the format of average differences. For both GNF and in-house Affymetrix probe set signals, we performed global lowess normalization for the three pairwise combinations of the liver and spleen slides, and used the averaged lowess adjustment for the normalization. The genes with probe set signals lower than 20 in either liver or spleen tissue were filtered out.
To stabilize the variances of data across the full range of gene expressions, we also performed the generalized log transformation for all the data sets following Durbin et al. [16].
Gene matching across arrays
There are five different data sets in this study. The origins of the genes vary in those datasets. In order to study inter-lab agreement, we have to first identify common genes represented in different arrays. Based on the annotation of each data set, we found that we could maximize the number of cross-matched genes using the mouse UniGene IDs (Build 107). Based on the common UniGene IDs, we found 551 common genes across all five different data sets. But in the pairwise comparisons, the number of common genes ranged from 1,374 to 5,018 (see Table 2). All comparisons between data sets were made from the matched genes.
Statistical procedures for inter-platform and inter-lab comparisons
In the analyses, the ratio is defined as normalized and transformed intensity from liver samples versus that from spleen samples.
Agreement of two-fold changes using kappa coefficients
An intuitive measurement of agreement is to count the percentage of genes falling in the same categories (two-fold up-regulated, no change, and two-fold down-regulated). However, this percentage can be high even if the data obtained from different platforms are not so compatible. Usually the ratios for the great majority of genes do not show a two-fold change, and the percentage of agreement can be high just due to chance. To adjust for this excess agreement expected by chance, we prefer to use the kappa coefficient, which is a popular measure of inter-rater agreement in many other areas of science. The kappa coefficient was first proposed by Cohen [17] for analysing dichotomous responses, and was extended later to more than two categories of responses. We applied this measure to three categories (two-fold up-regulated, no change, and two-fold down-regulated), and computed the kappa coefficients between two data sets from 3 by 3 frequency tables. For a study of q categories, the kappa coefficient is calculated by: , where is the overall agreement probability, is the measure of the likelihood of agreement by chance, and nij is the number of subjects in the (i, j) cell, ni+ is the sum of the i th row, n+ j is the sum of the j th column, and n is the total number of subjects.
For example, the kappa coefficient between KC and KAV is 0.523. Table 4 gives the two-fold gene regulation frequency table between KC and KAV. Except for the 8 genes that showed two-fold up-regulation in one data set but two-fold down-regulation in another, KC and KAV agreed very well.
Table 4 Frequency table for KAV and KC
KAV KC
Frequency -2 0 2 Total
-2 173 136 5 314
0 157 1,972 146 2,275
2 3 112 260 375
Total 333 220 411 2,964
Kappa coefficient = 0.523
Correlation coefficients of the ratios
We used three measures of correlation to compare the ratios from different data sets: Pearson correlation coefficient, intraclass correlation coefficient (ICC) and intra-transcript correlation coefficient (ITC). ICC measures the inter-rater reliability relative to the total variability of the ratios. Here, a rater could be a replicate or a technology platform. ICC is the variance of different ratios between UniGene IDs,, divided by the total variance. A high ICC (close to 1) means that the inter-rater ratios vary little relative to the overall variability in the data. In computing the ICC for the replicates, equals , where is the variance within UniGene IDs. If we consider lab as a random effect in the overall comparison, the total variance will equal , where is the variance between labs. The ICC incorporates both the association between raters and the rater differences, while the Pearson correlation is insensitive to the latter.
We introduced the ITC for pairwise comparisons as described below. For each gene i, we defined ρi to be the square root of the ratio of within dataset sum of squares (SSW) and the total sum of squares (TSS). A common SSW was used in comparing lab pairs to avoid the problem of having seemingly higher ITC's due to unusually large within-lab variability at some lab. We applied logit transformation to each ρi to get γi, and then calculated the average γ. Converting γ back to the correlation scale, we obtained .
Authors' contributions
HW and LL collected data from public resources. MB and CW generated in-house data. HW and LL processed the collected data. HW and XH did statistical analyses. All authors read and approved the manuscript.
Acknowledgements
This study was supported in part by the NIH Grant No. 2 P30 AR41940-10. We would like to acknowledge Al Bari for his work on printing the mouse cDNA array at the Keck Center, and thank the referees for helpful comments that led to improvements of the present paper.
==== Refs
Schena M Shalon D Davis RW Brown PO Quantitative monitoring of gene expression patterns with complementary DNA microarray Science 1995 270 467 470 7569999
Quackenbush J Computational analysis of microarray data Nat Rev Genet 2001 2 418 427 11389458 10.1038/35076576
Culhane AC Perriere G Higgins DG Cross-platform comparison and visualisation of gene expression data using co-inertia analysis BMC Bioinformatics 2003 4 59 14633289 10.1186/1471-2105-4-59
Jarvinen A Hautaniemi S Edgren H Auvinen P Saarela J Kallioniemi O Monni O Are data from different gene expression microarray platforms comparable? Genomics 2004 83 1164 1168 15177569 10.1016/j.ygeno.2004.01.004
Kuo WP Jenssen T Butte AJ Ohno-Machado L Kohane IS Analysis of matched mRNA measurements from two different microarray technologies Bioinformatics 2002 18 405 412 11934739 10.1093/bioinformatics/18.3.405
Yuen T Wurmbach E Pfeffer RL Ebersole BJ Sealfon SC Accuracy and calibration of commercial oligonucleotide and custom cDNA arrays Nucleic Acids Res 2002 30 e48 12000853 10.1093/nar/30.10.e48
Kothapalli R Yoder SJ Mane S Loughran TP Microarray results: how accurate are they? BMC Bioinformatics 2002 3 22 12194703 10.1186/1471-2105-3-22
Li J Pankratz M Johnson JA Differential gene expression patterns revealed by oligonucleotide versus long cDNA arrays Toxicol Sci 2002 69 383 390 12377987 10.1093/toxsci/69.2.383
Su AI Cooke M Ching KA Hakak Y Walker JR Wiltshire T Orth AP Vega RG Sapinoso LM Moqrich A Patapoutian A Hampton GM Schultz PG Hogenesch JB Large-scale analysis of the human and mouse transcriptomes Proc Natl Acad Sci 2002 99 4465 4470 11904358 10.1073/pnas.012025199
Thompson KL Afshari CA Amin RP Bertram TA Car B Cunningham M Kind C Kramer JA Lawton M Mirsky M Naciff JM Oreffo V Pine PS Sistare FD Identification of platform-independent gene expression markers of cisplatin nephrotoxicity Environmental Health Perspectives 2004 112 488 494 15033599
Pritchard CC Hsu L Delrow J Nelson PJ Project normal: Defining normal variance in mouse gene expression Proc Natl Acad Sci 2001 98 13266 13271 11698685 10.1073/pnas.221465998
Huber PJ Robust Statistics 1981 Cold Spring Harbor, John Wiley & Sons
Yang YH Dudoit S Luu P Speed TP Normalization for cDNA microarray Data Microarrays: Optical Technologies and Informatics 2001 SPIE BIOS San Jose, CA
Irizarry RA Hobbs B Collin F Beazer-Barclay YD Antonellis KJ Scherf U Speed TP Exploration, Normalization, and Summaries of High Density Oligonucleotide Array Probe Level Data Biostatistics 2003 4 249 264 12925520 10.1093/biostatistics/4.2.249
Li C Wong WH Model-based analysis of oligonucleotide arrays: Expression index computation and outlier detection Proc Natl Acad Sci 2001 98 31 36 11134512 10.1073/pnas.011404098
Durbin BP Hardin JS Hawkins DM Rocke DM A Variance-Stabilizing Transformation for Gene-Expression Microarray Data Bioinformatics 2002 18 S105 S110 12169537
Cohen J A coefficient of agreement for nominal scales Educational and Psychological Measurement 1960 20 37 46
| 15888200 | PMC1142313 | CC BY | 2021-01-04 16:39:53 | no | BMC Genomics. 2005 May 11; 6:71 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-71 | oa_comm |
==== Front
BMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 1472-6963-5-321586212210.1186/1472-6963-5-32Research ArticleAdapting developing country epidemiological assessment techniques to improve the quality of health needs assessments in developed countries Smith Susan M [email protected] Jean [email protected] Jillian [email protected]'Keeffe Frances [email protected] Deirdre [email protected]'Dowd Tom [email protected] Department of Public Health and Primary Care, University of Dublin, Trinity College Centre for Health Sciences, Adelaide and Meath Hospital Dublin, Incorporating the National Children's Hospital, Tallaght, Dublin, Ireland2 Drug Misuse Research Division, Health Research Board, Holbrook House, Holles Street, Dublin, Ireland3 HSE (NA), Westward House, Russell Street, Dublin, Ireland4 Summer Hill Health Centre, Dublin, Ireland2005 29 4 2005 5 32 32 8 9 2004 29 4 2005 Copyright © 2005 Smith et al; licensee BioMed Central Ltd.2005Smith 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 were commissioned to carry out three health assessments in urban areas of Dublin in Ireland. We required an epidemiologically robust method that could collect data rapidly and inexpensively. We were dealing with inadequate health information systems, weak planning data and a history of inadequate recipient involvement in health service planning. These problems had also been identified by researchers carrying out health assessments in developing countries. This paper reports our experience of adapting a cluster survey model originally developed by international organisations to assess community health needs and service coverage in developing countries and applying our adapted model to three urban areas in Dublin, Ireland
Methods
We adapted the model to control for socio-economic heterogeneity, to take account of the inadequate population list, to ensure a representative sample and to account for a higher prevalence of degenerative and chronic diseases. We employed formal as well as informal communication methods and adjusted data collection times to maximise participation.
Results
The model we adapted had the capacity to ascertain both health needs and health care delivery needs. The community participated throughout the process and members were trained and employed as data collectors. The assessments have been used by local health boards and non-governmental agencies to plan and deliver better or additional services.
Conclusion
We were able to carry out high quality health needs assessments in urban areas by adapting and applying a developing country health assessment method. Issues arose relating to health needs assessment as part of the planning cycle and the role of participants in the process.
==== Body
Background
In 2001, the Department of Public Health and Primary Care in Trinity College Dublin was commissioned to carry out three health assessments in urban areas in Dublin. We aimed to carry out health assessments that would represent the needs of the whole population and not just more vocal minorities.
We were aware that we were dealing with inadequate health information systems, weak planning data and a history of inadequate recipient involvement in health service planning. International health planners have reported similar issues in developing countries and as a result, have developed appropriate health assessment survey methods. The methods for such surveys must be robust, epidemiologically sound and have the potential to collect data rapidly, inexpensively and simply [1,2]. The survey design must provide information at local level and overcome unreliable population lists. If local health care providers and communities can participate in data collection this has the potential advantage of facilitating the meeting of such needs in the local context in due course [3].
To address these issues, a thirty-cluster survey method was designed to carry out health assessments in communities in developing countries [1,4,5]. We adapted the thirty-cluster survey method and applied it in an urban setting in a developed country. The results of these three health needs assessments have been previously published [6-8] and a summary is provided in Table 1 with reference to online links to the full reports. This paper reports the adaptations made to the thirty-cluster survey method and its implementation to establish the health needs of three urban communities in Dublin, Ireland.
Table 1 Summary of aims and main results of the Health Needs Assessments carried out. Full reports can be downloaded from: [see reference list for individual report links]
Tallaght Finglas Docklands
% (95% CI*)
Aims to assess the health needs of households and their individual members residing in each of the three areas surveyed
Sample size 420 households
1313 individuals 420 households
963 individuals 360 households
699 individuals
Response rate 80% 77% 75%
Proportion with a chronic illness (cardiovascular and respiratory disease and arthritis most common) at the time of the survey 22% (19%–24%) 31%(28%–34%) 27% (22%–31%)
Proportion with disability at the time of the survey 3% (2%–4%) 4% (3%–6%) 3% (2%–4%)
Psychosocial issues:
Experienced 'stress' in the 12 months prior to the survey 59% (54%–65%) 63% (57%–69%) 53% (47%–59%)
Experienced violence in the 12 months prior to the survey 10% (7%–13%) 13% (9%–17%) 11% (7%–15%)
Anxiety re: teenagers in household 60% (53%–66%) 61% (49%–73%) 64% (47%–81%)
Problem with drugs/ alcohol 2% (1%–3%) 1% (0.4%–2%) 1% (0.1%–3%)
Current smokers at the time of the survey 40% (36%–45%) 28% (24%–31%) 32% (27%–38%)
Womens' Health:
Using family planning method at the time of the survey 56% (48%–64%) 51% (41%–60%) 46% (36%–56%)
Cervical smear in the 5 years prior to the survey 58 (52%–64%) 52% (43%–61%) 54% (46%–62%)
Breast examination in the 5 years prior to the survey not available 47% (39%–56%) 43% (35%–51%)
Service use:
Used hospital service in the 12 months prior to the survey 25% (22%–28%) 33% (30%–66%) 24% (20%–29%)
Attended their GP in the 12 months prior to the survey 38% (33%–43%) 57% (54%–60%) 47% (42%–51%)
Visited dentist in the 12 months prior to the survey 15% (11%–18%) 12% (10%–14%) 12% (9%–15%)
On waiting list for health care at the time of the survey 4% (3%–5%) 6% (5%–8%) 4% (2%–5%)
Identification of services needed†
Improve out of hours GP care 52% 35% 52%
Improve services for elderly Not identified 36% 37%
Social work services Not identified Not identified 25%
Services for teenagers 19% 21% 14%
Local maternity service 47% Not identified Not identified
Health promotion clinics 24% 31% 12%
*95% confidence intervals were calculated for the main outcome measures adjusting for strata (deprivation level), primary sampling unit (clusters) and weight (for Finglas only).
† 95% confidence intervals were not calculated for the respondents' reported health needs.
Methods
Design and setting
The first two health assessments were commissioned by charitable foundations seeking to support the development of health services in two separate urban areas in Dublin [6,7]. The third assessment [8] was commissioned by a local health board in another area, following the publication of the preliminary results of the first assessment. Each area contained a diverse population though there were higher deprivation levels in the area covered by the first assessment and more elderly people living in the second and third areas studied [6-8].
Adaptations to the original model
The model was originally designed for use in homogenous populations in developing countries (Figure 1). However, the lack of socioeconomic homogeneity in urban areas in developed countries had to be considered in the selection of the study population. Detailed descriptions of the areas covered, including maps, and the slight variation in demographic characteristics in participants in each area are presented in the individual reports [6-8]. The Small Area Health Research Unit has calculated deprivation scores for all of the electoral divisions in Ireland, including the three areas surveyed (full details provided in section 2.1 of the Tallaght report [6]). These deprivation scores, based on data from the 1996 census, range from one to five depending on the level of deprivation. For each of the three areas surveyed, the deprivation scores were aggregated to form less deprived and more deprived groupings and separate samples were selected from each grouping [9]. The adaptations we made are summarized in Figure 2.
Figure 1 Features of model developed by the World Health Organisation, the Center for Disease Control and the Primary Health Care Management Advancement Programme to assess community health needs and service coverage [1,4,5].
Figure 2 Adaptation made to model for Dublin health assessments [6,7,8].
In developing countries, people are informed about surveys through community leaders and by word of mouth. We also used more formal methods, such as mailshots, posters at health and social care facilities, and advertisements in the local paper and on radio. Early on in the first survey, when many of the survey letters were returned unopened due to the high turnover of households in urban areas, we changed the addressee on survey envelopes to 'householder', giving us a response rate of at least 76% for each survey. In the second assessment, in a docklands regeneration area, access to households was difficult due to high security gates in apartment complexes, which was not considered prior to designing the survey.
Results
Health needs versus health service needs
The model we adapted had the capacity to incorporate health needs in the broadest sense as well as identify health service needs, and demonstrated the importance of community-based approaches to public health. The international shift to evidence based medicine has been criticised as removing the focus from the social and environmental determinants of health [10]. Some of our survey findings related to the broad determinants of health, such as the need for social outlets for teenagers and security systems for the elderly. This emphasizes the need for a multi-sectoral approach to the health of communities and stresses the importance of collecting truly representative views that reflect the social components of health. The fears expressed about personal safety led to an increased police presence on the street in one of the areas. The second assessment [7], was used by community activists to lobby the funding charity involved, to provide a bus service to transport older residents to the day centre. This focus on the broader determinants of health has also been reported in other health assessments with the development of new bus routes following such a survey in Edinburgh [11].
The reports highlight the potential discrepancy between communities' expectations of services and actual service provided as there was in fact quite high prevalence of use of hospital services. These discrepancies could be further explored using a qualitative approach. The addition of a parallel qualitative analysis of the health service providers' perspective in the second two surveys [7,8] provided more in-depth analysis of the health care services and indicated that the health service providers have a vast knowledge of the areas in which they work. Additional qualitative research involving community representatives would provide a greater depth of understanding of the inter-relationship between health needs and health care delivery needs and enable an exploration of the relationship between expectations and actual service provision.
Community participation
The communities were consulted and encouraged to participate throughout the process. For two to three months prior to each survey, the lead researcher informed key individuals, community groups and service providers about the proposed survey and elicited their perceived needs. In each area, several health and social services needs were identified and included in the final survey questionnaires. In addition, local community members were trained and used as data collectors.
It is argued that community participation is an essential component of health needs assessment and that it should be a cyclical and iterative process in order to ensure sustainability [10]. However, sustainability is difficult as participating community data collectors are upskilled and may move on into regular paid employment. Those remaining in community work reported improved insights into local problems thus benefiting their communities.
The importance of feeding results back to the community has been previously emphasized [12]. The assessment reports were launched in public venues. We invited elected public and health board representatives, local service providers, community members and representatives of the charitable foundations involved. The community data collectors played an important role in ensuring that the wider community was aware of the launch. They actively participated in the launch by presenting their perspective on the assessments. Following the second and third assessments, summary posters were designed and placed in public areas throughout each community.
Cost analysis
The time required to conduct each study and the cost in Euro (2001) are presented in Table 2. The cost shown for the first survey is an under-estimate as the funding did not cover the full salary costs of the lead researcher. The subsequent surveys were fully funded and allowed for a more comprehensive needs assessment including qualitative interviews with local service providers. The additional resources also allowed for wider dissemination of results, with the production of the summary posters. When compared with the cost of surveillance systems, these surveys represent good value for money. However, the addition of a qualitative component to the second and third surveys incurred an additional time and monetary cost.
Table 2 Time required to conduct surveys, and costs in Euro
First survey
(Tallaght) Second survey
(Docklands) Third survey
(Finglas)
Time from design to publication April 2001 to March
2002
(11 months) June 2001 to September
2002
(15 months) Nov 2001 to February
2003
(15 months)
Total expenditure in Euro 32,771 96,026 91,068
Impact of reports on services
The reports have been used by local healthcare providers, to lobby the health boards for more services, such as community antenatal clinics and a community paediatrician. The charitable foundation that commissioned the second assessment is now working with the local health board to provide a new primary healthcare centre. While it is not possible to prove conclusively that these service developments occurred solely as a result of the assessments, in each area the reports formed part of a process that directed changes in service delivery.
Discussion
Health assessment must be practical [13] and the model we have adapted was feasible and had the advantage of encouraging community involvement. The sampling method ensured genuine representation of all community members and minimized the potential bias inherent in focusing on the priorities of more vocal minorities.
Health needs assessment as part of the planning cycle
A favourable political environment has been described as a key component of successful health needs assessment [14]. For the two initial assessments [6,7], while the local health board approved the exercise, they had no direct role and consequently no ownership of the process. However, the first report led to another health board in Dublin commissioning the third assessment [8], which is being used to develop new services in this area. While local politicians and health board managers were briefly interested at the launch of the first report, they have not engaged further on the issues identified, which included lack of services and high levels of morbidity. By the time the third assessment was underway the environment had developed sufficiently to permit local health board involvement. As a result, the final assessment is genuinely part of the planning cycle [15] within that health board area, which has the statutory responsibility and capacity to deliver on the needs identified.
Roles of participants in the process
Local general practitioners and other healthcare providers were involved in all three assessments although previous reports have indicated that they do not see it as part of their core activities [16].
The role of charities in identifying need is worthy, but can raise expectations which they may have no remit to address. However, the commitment of the charity involved in the second report to building a new health centre in partnership with the local health board indicates that addressing needs identified can be part of the process even for non-governmental agencies without statutory responsibility to deliver services.
The role of academic departments in the health needs assessment process might also be questioned. The charities involved perceived the department in Trinity College to be objective and unaligned with service planners and providers. Academic departments possess the expertise and capacity to carry out assessments, but might better serve the process and the communities themselves by passing on these skills. Although this would lengthen the process and increase costs initially, it would be cost efficient in the future. Issues relating to sustainability and ownership of the process arise in health services research in general and the need for linkage and exchange between planners and researchers has been highlighted [17].
Conclusion
Our experience of using and adapting this health assessment methodology indicates the significant benefits for those seeking to carry out high quality health needs assessments in developed countries if they are prepared to learn from the experiences of developing country researchers. The adaptation of this methodology has provided more robust assessments that can be used by community members and service providers.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SS, JL, TOD, JD and DH conceived of this paper. SS wrote the initial draft and co-ordinated the production of the paper with the other authors. JL conceived the methods for the needs assessment and performed the additional statistical analysis for this paper. JL, JD, FO'K, DH and TO'D contributed to the three reports on which this paper is based. All authors contributed to the drafting of this manuscript and read and approved the final paper.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Lemeshow S Robinson D Surveys to measure programme coverage and impact: a review of the methodology used by the expanded programme on immunization World Health Stat Q 1985 38 65 75 4002731
Anker M Epidemiological and statistical methods for rapid health assessment: introduction World Health Stat Q 1991 44 94 106 1949886
Murray SA Experiences with rapid appraisal in primary care: involving the public in assessing health needs, orienting staff and educating medical students BMJ 1999 318 440 444 9974462
Rothenberg RB Lobanov A Singh KB Stroh G Observations on the application of EPI cluster survey methods for estimating disease incidence Bull World Health Organ 1985 63 93 99 3872739
Primary Healthcare Management Advancement Programme Assessing Community Health Needs and Coverage 1993 Geneva: Aga Khan Foundation
Department of Community Health and General Practice* People living in Tallaght and their Health, A community based cross-sectional survey 2002 Dublin: Trinity College [*The Department of Community Health and General Practice in Trinity College Dublin has since been renamed as the Department of Public Health and Primary Care]
Department of Community Health and General Practice* People living in the Dublin Docklands and their Health 2002 Dublin: Trinity College [*The Department of Community Health and General Practice in Trinity College Dublin has since been renamed as the Department of Public Health and Primary Care]
Department of Community Health and General Practice* People living in Finglas and their Health 2003 Dublin: Trinity College [*The Department of Community Health and General Practice in Trinity College Dublin has since been renamed as the Department of Public Health and Primary Care]
Lwanga SK Lemeshow S Sample Size Determination in Health Studies: A Practical Manual 1991 Geneva: World Health Organization
Israel B Schulz A Parker E Becker B Review of community-based research: Assessing partnership approaches to improve public health Annu Rev Public Health 1998 19 173 202 9611617 10.1146/annurev.publhealth.19.1.173
Stevens A Gillam S Needs assessment: from theory to practice BMJ 1998 316 1448 1452 9572762
McCaulay A Commanda L Freeman W Gibson N McCabe M Robbins C Twohig P Participatory research maximises community and lay involvement BMJ 1999 319 774 778 10488012
Wilkinson J Murray SA Assessment in primary care: practical issues and possible approaches BMJ 1998 316 1524 1528 9582151
Jordan J Wright J Ayres P Hawkings M Thomson R Wilkinson J Williams R Health needs assessment and needs-led health service change: a survey of projects involving public health doctors J Health Serv Res Policy 2002 7 71 80 11934371 10.1258/1355819021927719
Wright J Williams R Wilkinson JR Development and importance of health needs assessment BMJ 1998 316 1310 3 9554906
Hanlon P Murie J Gregan J McEwen J Moir D Russell E A study to determine how needs assessment is being used to improve health Public Health 1998 112 343 6 9807933 10.1016/S0033-3506(98)00267-4
Dash P Gowman N Traynor M Increasing the impact of health services research BMJ 2003 327 1339 1341 14656847 10.1136/bmj.327.7427.1339
| 15862122 | PMC1142314 | CC BY | 2021-01-04 16:31:50 | no | BMC Health Serv Res. 2005 Apr 29; 5:32 | utf-8 | BMC Health Serv Res | 2,005 | 10.1186/1472-6963-5-32 | oa_comm |
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-161585049210.1186/1471-2350-6-16Research ArticleXLMR in MRX families 29, 32, 33 and 38 results from the dup24 mutation in the ARX (Aristaless related homeobox) gene Stepp Monica L [email protected] A Lauren [email protected] Merran [email protected] Marie [email protected] Elke [email protected] David [email protected] Andrée [email protected] Jeanette JA [email protected] Jozef [email protected] Roger E [email protected] Charles E [email protected] J.C. Self Research Institute, Genetic Center, Greenwood, S.C., USA2 Department of Genetic Medicine, Women's and Children's Hospital, Adelaide, Australia3 Department of Pediatrics, The University of Adelaide, Adelaide, Australia Ottawa Health Research Institute, Ottawa, Ontario, Canada4 Medizinisch Genetisches Zentrum, Bayerstrasse 53, D-80335, Munchen, Munich, Germany5 Provincial Medical Genetics Program, St. John, Newfoundland, Canada6 Provincial Medical Genetics Program, Health Science Centre, St. John, Newfoundland, Canada7 Department Psychiatry & Physiology, Queen's University, Kingston, Ontario, Canada2005 25 4 2005 6 16 16 29 11 2004 25 4 2005 Copyright © 2005 Stepp 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
X-linked mental retardation (XLMR) is the leading cause of mental retardation in males. Mutations in the ARX gene in Xp22.1 have been found in numerous families with both nonsyndromic and syndromic XLMR. The most frequent mutation in this gene is a 24 bp duplication in exon 2. Based on this fact, a panel of XLMR families linked to Xp22 was tested for this particular ARX mutation.
Methods
Genomic DNA from XLMR families linked to Xp22.1 was amplified for exon 2 in ARX using a Cy5 labeled primer pair. The resulting amplicons were sized using the ALFexpress automated sequencer.
Results
A panel of 11 families with X-linked mental retardation was screened for the ARX 24dup mutation. Four nonsyndromic XLMR families – MRX29, MRX32, MRX33 and MRX38 – were found to have this particular gene mutation.
Conclusion
We have identified 4 additional XLMR families with the ARX dup24 mutation from a panel of 11 XLMR families linked to Xp22.1. This finding makes the ARX dup24 mutation the most common mutation in nonsyndromic XLMR families linked to Xp22.1. As this mutation can be readily tested for using an automated sequencer, screening should be considered for any male with nonsyndromic MR of unknown etiology.
==== Body
Background
The causes of mental retardation (MR) are exceptionally heterogeneous. More than 50 genes found on the human X chromosome have been reported to cause MR with another 150 loci being associated with linked syndromes, nonsyndromic families and unlinked syndromes [1,2]. The aristaless-related homeobox gene, ARX, maps to Xp22.1-p21.3, encompasses a genomic region of approximately 12.5 kb and is composed of 5 encoding exons [3]. The most frequent mutation reported in ARX is an in-frame 24 bp duplication in exon 2, which causes an expansion of the polyalanine tract at amino acid positions 144–155, from 12 to 20 alanines. This duplication has been reported in a Norwegian family with West Syndrome [3,4] as well as four published nonsyndromic XLMR families, MRX36 [5], MRX43, MRX54 and MRX76 [6]. We have recently identified the same 24 bp duplication in 4 other nonsyndromic XLMR families linked to Xp22.1: MRX29 [7], MRX32 [8], MRX33 [9], and MRX38 [10].
Methods
Genomic DNA from members of eleven XLMR families linked to Xp22.1, a West Syndrome family with a 21 bp insertion [3], two positive controls for the 24 bp duplication (dup24) [3], and a negative control was amplified by PCR using locus specific primer pairs designed for exon 2 of ARX. The Cy5 labeled forward primer sequence, ARXex2P1, was 5'ACG CCT GGG CCT AGG CAC TG 3' and the reverse primer sequence, ARXex2P1, was 5' CTC GGT GCC GGT GCC ACC AC 3'. These primers flank both the (GCG)10+7 and dup24 mutations. The size of the normal control PCR product is 584 bp, while the (GCG)10+7 mutation gives a 605 bp product and the dup24 mutation a 608 bp product. The parameters of the reaction consisted of an initial denaturation at 95°C followed by 30 cycles consisting of 95°C for 30 seconds, 62°C for 30 seconds, and 72°C for 30 seconds and completing with a final extension at 72°C for 5 minutes. PCR amplification was verified using a 1.5 % agarose gel, run at 120 volts, and stained with ethidium bromide. PCR products were then analyzed using the Amersham Biosciences ALFexpress1.
Results
The analysis revealed an altered pattern in a female carrier of MRX38 and one affected male from each of three other families: MRX29, MRX32, MRX33 (Figure 1). The pattern was identical to that observed for two samples known to have the dup24 mutation in exon 2 of ARX and clearly distinct from the 21 bp insertion in exon 2 that is present in the patient with West Syndrome (Figure 1). DNA sequencing of the PCR products confirmed the presence of the dup24 mutation. It is important to note the discrepancy in the size of the dup24 mutation and the sizing standard in lanes 1 and 4 (Figure 1). We hypothesize that the slower migration pattern observed for the dup24 mutation (~645 bp instead of the expected 608 bp) is due to its secondary (hairpin) structure.
Discussion
We have identified the dup24 mutation in exon 2 of ARX in four additional nonsyndromic XLMR families (MRX29, 32, 33, and 38) linked to Xp22.1. This same alteration has been observed in four other published nonsyndromic families, MRX36 [5], MRX43, MRX56 and MRX76 [6] linked to Xp22.1. As a result, this single mutation accounts for 8/11 (73%) of MRX families linked to Xp22.1 [11]. This 24 bp duplication is not restricted to nonsyndromic XLMR families. A family with West syndrome linked to Xp22.1 had this alteration [4], as well as 2 families with Partington syndrome, also linked to Xp22.1 [3]. This association of nonsyndromic and syndromic XLMR with the 24 bp duplication has also been noted by Kato et al [12].
The dup24 also accounts for 2/10 (20%) syndromic XLMR conditions linked to Xp22.1 [13,14]. This is quite striking and noteworthy since for most XLMR genes described thus far, mutations are found in less than 20% of families linked to any particular region [15]. Overall, it appears that when all mutations in ARX are considered, alterations in the gene accounts for 4/58 (7%) syndromic and 8/24 (25%) nonsyndromic XLMR conditions with known gene defects. Again, this is uncharacteristic of other XLMR genes, except for ATRX/XNP (mutations in 6 syndromes) or PQBP1 (mutations 5 syndromes) [14].
This being said, there is a "disconnect" when screening for ARX mutations is expanded to the general male MR population. Bienvenu et al. [6] were only able to identify two mutations in a panel of 148 XLMR families and only one dup24 mutation in 40 sporadic males with MR. Additionally, Gronskov et al. [16] identified only one dup24 mutation out of 682 patients. Other studies experienced a similar failure to find this mutation in 577 males with non-fragile X MR (Stepp and Schwartz, unpublished) and a cohort of 188 male patients [15]. This translates to a positive screen rate of 2/1501 male patients with MR, significantly different from the 6.6% rate observed in proven XLMR families [15].
At present, there is no clear explanation for this disparity in the frequency of ARX mutations between families with XLMR linked to Xp22.1 and sporadic cases of males with MR. It may simply mean that the assumption that the 25–30% excess of males with MR is due to genes on the X chromosomes is incorrect. Therefore, the male excess in the MR population might have to be accounted for by other mechanisms including yet to be revealed X-chromosome linked factors.
Conclusion
In conclusion, we have shown that 4 additional MRX families (MRX29, 32, 33 and 38), mapped to Xp22.1, have the dup24 mutation in exon 2 of the ARX gene. This mutation accounts for about 70% of MRX families linked to Xp22.1. However, it has been difficult to find this same mutation in large cohorts of males with mental retardation. Nonetheless, as the 24dup mutation in ARX can easily be tested using an automated DNA sequencer, screening of males with MR of unknown etiology for this particular mutation should be given consideration.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MLS, ALC, MF and MM carried out the molecular genetics studies. EH-F, DM, AM, JJAH and RES provided clinical material. JG and CES drafted the manuscript. CES conceived the study and participated in its design 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 has been supported by grant HD26202 from NICHD (CES), the National Health and Medical Research Council of Australia (JG) and in part, by the South Carolina Department of Disabilities and Special Needs (RES). This paper is dedicated to the memory of Ethan Francis Schwartz (1996–1998).
Figures and Tables
Figure 1 Chromatograph of samples with ARX 24dup mutation. Lane 17 is the 50–500 bp ladder lanes. Lane 14 is a normal control. Lane 3 is an affected male from MRX29. Lane 4 is an affected male from MRX32. Lane 5 is an affected male from MRX33. Lane 6 is a female carrier from MRX38. Lane 7 is a male with (GCG)10+7 ARX mutation. Lane 8 is an affected male with a known dup24 mutation and Lane 9 is a female carrier with the dup24 mutation.
==== Refs
Chelly J Mandel JL Monogenic causes of X-linked mental retardation Nat Rev Genet 2001 2 669 680 11533716 10.1038/35088558
Stevenson RE Schwartz CE Clinical and molecular contributions to the understanding of X-linked mental retardation Cytogenet Genome Res 2002 99 265 275 12900574 10.1159/000071603
Stromme P Mangelsdorf ME Shaw MA Lower KM Lewis SM Bruyere H Lutcherath V Gedeon AK Wallace RH Scheffer IE Turner G Partington M Frints SG Fryns JP Sutherland GR Mulley JC Gecz J Mutations in the human ortholog of Aristaless cause X-linked mental retardation and epilepsy Nat Genet 2002 30 441 445 11889467 10.1038/ng862
Stromme P Sundet K Mork C Cassiman JJ Fryns JP Claes S X linked mental retardation and infantile spasms in a family: new clinical data and linkage to Xp11.4-Xp22.11 J Med Genet 1999 36 374 378 10353782
Frints SG Froyen G Marynen P Willekens D Legius E Fryns JP Re-evaluation of MRX36 family after discovery of an ARX gene mutation reveals mild neurological features of Partington syndrome Am J Med Genet 2002 112 427 428 12376949 10.1002/ajmg.10628
Bienvenu T Poirier K Friocourt G Bahi N Beaumont D Fauchereau F Ben Jeema L Zemni R Vinet MC Francis F Couvert P Gomot M Moraine C van Bokhoven H Kalscheuer V Frints S Gecz J Ohzaki K Chaabouni H Fryns JP Desportes V Beldjord C Chelly J ARX, a novel Prd-class-homeobox gene highly expressed in the telencephalon, is mutated in X-linked mental retardation Hum Mol Genet 2002 11 981 991 11971879 10.1093/hmg/11.8.981
Hane B Schroer RJ Arena JF Lubs HA Schwartz CE Stevenson RE Nonsyndromic X-linked mental retardation: review and mapping of MRX29 to Xp21 Clin Genet 1996 50 176 183 9001795
Hane B Stevenson RE Arena JF Lubs HA Simensen RJ Schwartz CE Gene for apparently nonsyndromic X-linked mental retardation (MRX32) maps to an 18-Mb region of Xp21.2-p22 Am J Med Genet 1999 85 271 275 10398242 10.1002/(SICI)1096-8628(19990730)85:3<271::AID-AJMG17>3.0.CO;2-#
Holinski-Feder E Golla A Rost I Seidel H Rittinger O Meindl A Regional localization of two MRX genes to Xq28 (MRX28) and to Xp11.4-Xp22.12 (MRX33) Am J Med Genet 1996 64 125 130 8826462 10.1002/(SICI)1096-8628(19960712)64:1<125::AID-AJMG21>3.0.CO;2-O
Schutz CK Ives EJ Chalifoux M MacLaren L Farrell S Robinson PD White BN Holden JJ Regional localization of an X-linked mental retardation gene to Xp21.1-Xp22.13 (MRX38) Am J Med Genet 1996 64 89 96 8826457 10.1002/(SICI)1096-8628(19960712)64:1<89::AID-AJMG16>3.0.CO;2-O
Linkage limits for 48 XLMR syndromes which have been mapped (lod score >2), but the genes not yet cloned 2004
Kato M Das S Petras K Kitamura K Morohashi K Abuelo DN Barr M Bonneau D Brady AF Carpenter NJ Cipero KL Frisone F Fukuda T Guerrini R Iida E Itoh M Lewanda AF Nanba Y Oka A Proud VK Saugier-Veber P Schelley SL Selicorni A Shaner R Silengo M Stewart F Sugiyama N Toyama J Toutain A Vargas AL Yanazawa M Zackai EH Dobyns WB Mutations of ARX are associated with striking pleiotropy and consistent genotype-phenotype correlation Hum Mutat 2004 23 147 159 14722918 10.1002/humu.10310
Linkage limits for 55 MRX families which have been mapped (lod score >2). The locations of the 9 MRX genes which have been cloned are indicated by the solid arrows, 6 genes that cause MRX and MRXS are shown with open arrows 2004
Location of the 34 genes for XLMR syndromes which have been cloned and mutations demonstrated 2004
Mandel JL Chelly J Monogenic X-linked mental retardation: is it as frequent as currently estimated? The paradox of the ARX (Aristaless X) mutations Eur J Hum Genet 2004 12 689 693 15319782 10.1038/sj.ejhg.5201247
Gronskov K Hjalgrim H Nielsen IM Brondum-Nielsen K Screening of the ARX gene in 682 retarded males Eur J Hum Genet 2004 12 701 705 15199382 10.1038/sj.ejhg.5201222
| 15850492 | PMC1142315 | CC BY | 2021-01-04 16:03:33 | no | BMC Med Genet. 2005 Apr 25; 6:16 | utf-8 | BMC Med Genet | 2,005 | 10.1186/1471-2350-6-16 | oa_comm |
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-181587781310.1186/1471-2350-6-18Research ArticleMolecular breakpoint cloning and gene expression studies of a novel translocation t(4;15)(q27;q11.2) associated with Prader-Willi syndrome Schüle Birgitt [email protected] Mohammed [email protected] Emma [email protected] David I [email protected] Margaret [email protected] Howard R [email protected] RJ McKinlay [email protected] Uta [email protected] Department of Genetics, Stanford University School of Medicine, Stanford CA 94305, USA2 Murdoch Children's Research Institute and Paediatrics Department, University of Melbourne, Royal Children's Hospital, Parkville 3052, Victoria, Australia3 Department of Child Development and Rehabilitation, Royal Children's Hospital, Parkville 3052, Victoria, Australia4 Department of Pathology, King Fahad National Guard Hospital, Riyadh 11426, Saudi Arabia2005 6 5 2005 6 18 18 20 1 2005 6 5 2005 Copyright © 2005 Schüle 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
Prader-Willi syndrome (MIM #176270; PWS) is caused by lack of the paternally-derived copies, or their expression, of multiple genes in a 4 Mb region on chromosome 15q11.2. Known mechanisms include large deletions, maternal uniparental disomy or mutations involving the imprinting center. De novo balanced reciprocal translocations in 5 reported individuals had breakpoints clustering in SNRPN intron 2 or exon 20/intron 20. To further dissect the PWS phenotype and define the minimal critical region for PWS features, we have studied a 22 year old male with a milder PWS phenotype and a de novo translocation t(4;15)(q27;q11.2).
Methods
We used metaphase FISH to narrow the breakpoint region and molecular analyses to map the breakpoints on both chromosomes at the nucleotide level. The expression of genes on chromosome 15 on both sides of the breakpoint was determined by RT-PCR analyses.
Results
Pertinent clinical features include neonatal hypotonia with feeding difficulties, hypogonadism, short stature, late-onset obesity, learning difficulties, abnormal social behavior and marked tolerance to pain, as well as sticky saliva and narcolepsy. Relative macrocephaly and facial features are not typical for PWS. The translocation breakpoints were identified within SNRPN intron 17 and intron 10 of a spliced non-coding transcript in band 4q27. LINE and SINE sequences at the exchange points may have contributed to the translocation event. By RT-PCR of lymphoblasts and fibroblasts, we find that upstream SNURF/SNRPN exons and snoRNAs HBII-437 and HBII-13 are expressed, but the downstream snoRNAs PWCR1/HBII-85 and HBII-438A/B snoRNAs are not.
Conclusion
As part of the PWCR1/HBII-85 snoRNA cluster is highly conserved between human and mice, while no copy of HBII-438 has been found in mouse, we conclude that PWCR1/HBII-85 snoRNAs is likely to play a major role in the PWS- phenotype.
==== Body
Background
Prader-Willi syndrome (PWS) is a complex neurodevelopmental disorder and a classic example for genomic imprinting in humans. The incidence is about 1 in 10–20,000, and the clinical manifestations include decreased fetal activity, neonatal hypotonia, neonatal feeding difficulties, hyperphagia with obesity, hypogonadism, short stature, small hands and feet, characteristic facial features, and mild to moderate mental retardation. Diagnostic criteria have been proposed [1] and revised recently [2].
About 70% of individuals clinically diagnosed with PWS have a ~4 Mb interstitial deletion at 15q11-13 of paternal origin, with clustered breakpoints (BP) at either of two proximal sites (BP1 or BP2) and one on the distal site (BP3) (Fig. 1a). The majority of the remainder have maternal uniparental disomy 15. In about 1 % of the cases, the disease is due to aberrant imprinting and gene silencing. Of these, 14% have small deletions in the imprinting center (IC) region of the paternal allele that abolish the expression of all imprinted paternally-expressed genes in cis. In the remainder no demonstrable DNA sequence changes have been observed [3-7]. In Angelman syndrome (AS, MIM#105830), which is usually caused by the same mechanisms affecting the maternal chromosome 15, mutations in the maternal copy of a single gene, UBE3A (MIM#601623), encoding a ubiquitin ligase, are detected in about 5% of cases, whereas in PWS, no disease-causing mutations in a single imprinted gene have yet been reported.
Three paternally expressed genes have been identified between BP2 and SNRPN. These include MKRN3/ ZNF127 (MIM# 603856; Makorin 3 or Zinc finger protein 127) [8,9], MAGEL2/NDNL1 (MIM# 605283; MAGE-like 2 or Necdin-like 1) [10,11], and NDN (MIM# 602117; Necdin) [12,13] (Figure 1a). The small nuclear ribonucleoprotein polypeptide N (MIM# 182279; SNRPN) gene was the first gene with a known function to be mapped to the PWS/AS deletion region, and is expressed from the paternal chromosome only [14-17]. Multiple alternatively spliced transcripts originate at the SNPRN promoter [18-20]. The major SNRPN transcript is bi-cistronic encoding two mRNA species. Exons 1–3 encode a protein product of unknown function called SNURF (SNRPN upstream reading frame). Exons 4–10 encode SmN, a homolog of the SmB/B' protein that binds small nuclear RNAs involved in pre-mRNA splicing. The largest transcripts extend over a ~460 kb genomic region and include a large 3'UTR comprising up to 148 exons [19].
Multiple introns downstream of the SNURF-SNRPN coding region contain C/D box small nucleolar RNA (snoRNA) genes. There are two multi-copy snoRNA clusters (HBII-52 and PWCR1/HBII-85) [21,22], three single copy snoRNA genes (HBII-436, HBII-13, and HBII-437), and one snoRNA gene (HBII-438) present in two copies that are 240 kb apart [19]. Since the snoRNAs are derived from processed spliced-out introns, their expression is controlled by the SNRPN promoter and is highest in brain. The known function of other C/D box snoRNAs is to guide 2'- O – ribose methylation of ribosomal RNA or small nuclear RNA. This post-transcriptional modification is conserved throughout evolution and is thought to confer increased stability to the small RNA molecules [23]. The modification targets of the imprinted C/D box snoRNAs in the PWS/AS region are still unknown.
Spontaneous chromosome translocations can be extremely valuable for assessing the contributions of individual loci to the phenotype of microdeletion syndromes. Five individuals with features of PWS have been reported who have balanced reciprocal translocations with breakpoints in the PWS/AS deletion region. All of them involve the SNRPN locus. The breakpoints are located in intron 2 (proximal, n = 2), disrupting the SNURF/SNRPN coding region, or in exon 20a/intron 20 (distal, n = 3) within the 3'-untranslated region of the long SNRPN transcript. One individual with a proximal and two of three patients with a distal breakpoint meet the diagnostic criteria for PWS (score of 8 or more points) [20,24-28].
Here we report the clinical, cytogenetic and molecular characterization of a 22 year old male with features of PWS who has a different de novo balanced reciprocal translocation t(4;15)(q27;q11.2). We mapped the breakpoint to SNRPN intron 17 (position on chr 15: 22803227, UCSC Genome browser May 2004) and determined the expression of snoRNAs on both sides of the breakpoint in cultured fibroblasts and lymphoblasts.
Methods
Cytogenetic and FISH analysis
Metaphase spreads obtained from short-term blood lymphocyte cultures and Epstein-Barr virus (EBV)-transformed lymphoblastoid cells (LCL) were processed for high-resolution GTG- banding by standard methods. For FISH studies, Bacterial Artificial Chromosomes (BACs) were sourced from the RPCI-11 library and selected using the UCSC Genome Browser, Assemblies: July 2003 and May 2004). Fluorescence labelling, hybridization procedures and imaging were performed as previously described [29].
DNA methylation study
Genomic DNA was purified by phenol-chloroform extraction from LCLs from the study subject, a normal control, and a PWS individual (Patient E in [6], Coriell Human Mutant Cell Repository # GM12134). To investigate methylation at exon 1 of SNRPN, 50 μg DNA were used for the bisulfite reaction and PCR with primers according to standard protocols [30,31]. PCR products were separated on a 3% agarose gel, stained with ethidium bromide, and visualized under UV illumination.
Expression studies by RT-PCR and quantitative RT-PCR
Total RNA was extracted from LCL and primary fibroblast cultures (FB) using RNA Stat 60. The RNA was treated with DNaseI (Roche) and RT-PCR was performed using Superscript II (Invitrogen). Primers were designed for exon-to-exon amplification in an overlapping fashion – where possible – for SNRPN, MKRN3, MAGEL2, NDN, snoRNAs, and two ESTs within Intron 20 of SNRPN (Table 1).
For a subset of exons in the SNPRN gene and the snoRNAs HBII-13, HBII-437, and PWCR1/HBII-85, quantitative RT-PCR assays were performed with SYBR Green I™ dye in an ABI 7700 cycler (Applied Biosystems) by using standard protocols [32,33]. Primers were designed to amplify products of 50 bp in length. GAPDH expression was used as a reference. Each sample was run at least in triplicate. The results were interpreted as described previously [28].
LCL RNA samples from a PWS individual with a microdeletion of the imprinting center (GM12134), a normal individual, an individual with an intrachromosomal triplication of the PWS region on the paternally-derived chromosome 15 (Patient 1 in [34], Coriell Human Mutant Cell Repository # GM12135), and fibroblast RNA from another t(4;15) PWS individual with the breakpoint in intron 2 of SNRPN [27,28] served as controls.
Southern blot analysis
Southern blot analysis was performed according to standard methods with ExpressHyb™ solution (BD Biosciences). Genomic DNA from a normal individual and the t(4;15) carrier was cleaved in a double digestion with restriction enzymes NheI and BsaWI to release a 6.4 kb fragment, and with NheI and ApaI to release a 10 kb fragment in the normal chromosome. The DNA probes were synthesized by PCR from genomic DNA and cloned into a pCRII T/A-vector (Invitrogen). The probes were designed to hybridize within intron 16 (SB-1) and upstream of the ApaI restriction site (SB-3) (Table 1).
Breakpoint cloning with a PCR-based method
Genomic DNA from a normal individual and the t(4;15) carrier was cleaved in a double digestion with restriction enzymes EcoRV and ApaI, followed by adapter ligation according to the manufacturer's instructions (BD Genome Walker Universal Kit) [35]. A nested PCR reaction with adapter primers and sequence-specific primers was performed and the amplification products were cloned into the pC2.1 T/A-vector (Invitrogen) after gel purification. The clones were sequenced from both directions with universal primers from the vector (M13) and sequence specific primers.
Results
Clinical case report
The patient (Fig. 2) was born at 41 weeks of gestation with a birth weight of 8 lb. Pregnancy was uneventful, but fetal movements were somewhat reduced. In the newborn, poor muscle tone, weak cry, excessive sleepiness, and undescended testes were observed. During infancy, he had poor suck and prolonged feeding times, but his weight gain was satisfactory and he did not require tube feeding. He was suspected to have absence seizures of about 20 seconds duration, along with proneness to giggling, sometimes with eye-rolling. These episodes resolved by four years of age, and an EEG was normal.
He had a left esotropia that was surgically corrected. During childhood, sticky saliva, dry mouth, skin picking and a marked tolerance to pain were noted and have persisted.
Excessive daytime somnolence continued beyond infancy and treatment with amphetamine was started at 9 years of age. A sleep study, at 13 years of age, was normal. In 2002, a further sleep study and a multiple sleep latency tests confirmed the diagnosis of narcolepsy. His daytime sleepiness has continued to respond to dexamphetamine.
Regarding his body weight, there was no rapid weight gain between 1 and 6 years. Around 8 years of age, his interest in food increased, and now he would keep eating if he had unrestricted access to favorite sweet foods. He lives with his parents who help to control his food intake. At 14.5 years, he had small hands and feet, at the 20th percentile and 5th percentile, respectively, and showed mild truncal obesity. His head circumference of 56.7 cm was at the 98th percentile. Brain MRI scan was normal. At age 16 years, his height was 155.7 cm and weight 65 kg. At the age of 22 years, his height is approximately 164 cm and his weight has increased to 90 kg (BMI = 33.5).
At 13 years of age, he was found to have delayed puberty and reduced linear growth velocity with his height falling below the 3rd centile. Treatment with testosterone resulted in improved height gain and genital development. At 15 years of age, he had a left orchidopexy and removal of a dysplastic intra-abdominal right testis. He remains on 6 monthly testosterone implants because of reduced hypothalamic function. He has never been on growth hormone treatment.
Developmentally he had a mild delay in comparison to his older siblings. He attended normal school but had some difficulties due to rigid behaviours and poor peer interactions. Psychological testing (WISC 111, Wide Range Achievement test and BASC self report) revealed an overall normal intellect. However, he had some involuntary fluctuation in attention and significant visual perceptual difficulties, e.g. deficits in visual organization, in making sense of his visual world and transcribing visual material. These perceptual problems have had a significant effect on his learning and social life. At the age of 22, he is attending a mainstream high school requiring extra time and assistance in completing a diploma in information technology. He is good at dismantling computers and installing hardware, and prefers working on his computer to socializing. Hyperphagia and skin picking are still a challenge for him.
Cytogenetic analysis
High-resolution chromosome analysis showed an apparently balanced reciprocal translocation between the long arm of chromosome 4 and the proximal long arm of chromosome 15. The breakpoints were assigned to chromosome bands 4q27 and 15q11: 46, XY, t(4;15)(q27;q11) (Fig. 3a). Parental chromosomes were normal, indicating that the patient's translocation was de novo.
DNA methylation analysis
To exclude alternative explanations for the phenotype, such as an imprinting defect, DNA methylation analysis was performed. Methylation-specific PCR of the SNURF-SNRPN exon 1 region revealed a normal bi-parental methylation pattern (Fig. 3b).
Mapping of the translocation breakpoint by FISH
We performed cytogenetic and molecular studies to characterize the breakpoint at 15q11 in detail. Preliminary FISH analysis showed that the breakpoint in 15q11 was located between D15S11 and GABRB3, which flank the SNRPN locus (data not shown). On this basis, a chromosome walking strategy was used across this region to narrow down the breakpoint region. We identified two BACs, RP11-160D9 (current position 22577151-22735621 on UCSC Genome Browser, May 2004 release) and RP11-876N20 (current position 22857334-23036552), that flanked the breakpoint and, thus, mapped it to a ~122 kb interval (Fig. 1b).
Fine mapping of the breakpoint by SNRPN expression and Southern blot analysis
To further refine the breakpoint, we carried out quantitative RT-PCR and RT-PCR experiments using RNA from an LCL and skin fibroblasts (FB) for expression of SNRPN transcripts. As shown in Figure 4 and Table 2, we found expression of SNPRN exons 2, 3, and 14 to 17, but no expression of exons 18 to 20, and concluded that the breakpoint falls within intron 17. For mapping intron 17, we designed a Southern blot using unique restriction sites. DNA cleaved with NheI and BsaWI showed a 6.4 kb band for the t(4;15) carrier and the normal control (Fig. 5b, lanes 1 and 2), indicating that the breakpoint is located downstream of the BsaWI site. Samples doubly digested with NheI and ApaI (Fig. 5b, lanes 3–6), revealed additional bands for the translocation carrier. Besides the expected 10 kb band derived from the normal chromosome 15, there was a ~11.5 kb band in lane 4, detected with the SB-1 probe, and a ~7 kb band in lane 6, detected with probe SB-3 (Fig. 5b). The novel ~11.5 kb band arose from the der(15) chromosome, with an NheI site on the chromosome 15 portion and an ApaI site on the chromosome 4 portion (Fig. 5c, upper panel). The novel band of ~7 kb arose from the der(4) chromosome, with an ApaI site on the chromosome 15-part and an NheI site on the chromosome 4-part. Taken together, these results delimit the breakpoint region to ~3.6 kb between the BsaWI and ApaI sites (Fig. 5a).
Breakpoint mapping at the nucleotide level
By DNA sequencing, we mapped the breakpoint to SNRPN intron 17 (position chr 15: 22803227) and to chromosome 4 at position chr. 4:123965881 (UCSC Genome Browser, May 2004) (Fig. 6). On chromosome 4, a long terminal repeat (LTR) retrotransposon, LTR1B, spans the breakpoint. On chromosome 15, we found a short interspersed element (SINE), AluY, and a long interspersed element (LINE), L1M4, surrounding the breakpoint (Fig. 6a). Thirty-nine bp upstream of the breakpoint on chromosome 15 starts a common 26 bp core sequence of Alu elements (Alu-DEIN) in an inverted orientation. This sequence is known to be involved in gene rearrangements [36]. While the sequence across the breakpoint is contiguous on the der(15), an extra A is inserted on the der(4) chromosome (Fig. 6b). Furthermore, the breakpoint on chromosome 4 falls in a large intron between exons 10 and 11 of a spliced transcript (BC045668). By RT-PCR, we found that this transcript is expressed in fibroblasts, but not in LCLs (data not shown).
Expression of upstream genes MKRN3, MAGEL2, and NDN
Expression of the three imprinted genes MKRN3, MAGEL2, and NDN upstream of SNRPN was tested by RT-PCR in t(4;15) fibroblasts and found to be indistinguishable from expression in normal control fibroblasts (data not shown).
Expression of C/D box snoRNAs and intron-encoded ESTs
When testing for the intron-encoded C/D box snoRNAs, we were able to document expression of HBII-13 and HBII-437 and lack of expression for HBII-438A/B and HBII-85/PWCR1 (Fig. 7). By use of a more sensitive method, quantitative real-time RT-PCR, we obtained similar results for the SNRPN exons and snoRNAs tested (Table 2). Two ESTs, AK094315 and AB061718 (= HBT8) located in the 30 kb SNRPN intron 20 were not expressed in the PWS control [6] and t(4;15) LCLs, but were expressed in the normal control LCL (Fig. 7).
Discussion
Breakpoint mapping and mechanism of the translocation event
Dissecting the PWS deletion region and identifying individual genes as responsible for parts of the phenotype represent a challenge because all reported smaller deletions inactivate all imprinted genes on the paternally- derived chromosome 15. Rare reciprocal translocations, therefore, provide unique insights. We here report our studies of a 22 year old male with features of PWS who has a de novo balanced reciprocal translocation t(4;15)(q27;q11.2). This is the first such case where the translocation breakpoints have been identified at the DNA sequence level. The cytogenetic breakpoint designations in this individual are identical to those in another male PWS-like case with t(4;15)(q27;q11.2), previously reported by Kuslich and colleagues [27] and restudied by Gallagher and colleagues [28], which raised the intriguing possibility of a recurrent translocation that may be facilitated by genomic repeats or other distinct molecular features.
In the present case, however, we mapped the breakpoint to SNRPN intron 17 (position on chr. 15: 22803227, UCSC Genome Browser, May 2004) that differs from that in the previous case (SNRPN intron 2). Furthermore, the breakpoint in our case is novel as it does not fall into one of the two previously described "breakpoint clusters" in intron 2 and exon 20a/intron 20 (Table 3 and Table 4). On chromosome 4 (chr. 4 123965881), the LTR retrotransposon LTR1B is spanning the breakpoint, and a short interspersed element (SINE), AluY, and a LINE element, L1M4, surround the breakpoint on chromosome 15 (Fig. 6). Interestingly, 39 bp upstream of the breakpoint on chromosome 15 starts a common 26-bp core sequence of Alu elements (Alu-DEIN) that has been shown to be involved in gene rearrangements and has homology with prokaryotic χ, an 8-bp sequence motif known to stimulate recBC mediated recombination in E. coli[37]. The core sequence is identical to sequences in the left arm of the consensus Alu element [38]. Sequence analyses of regions directly adjacent to translocation breakpoints has shown presence of the 26-bp Alu core sequence at or close (within 20–50 bp downstream or upstream) to the sites of recombination [36]. Therefore, this sequence might stimulate homologous and non-homologous recombination within the core or at nearby sites and could be the mechanism of recombination in the t(4;15) case reported here.
The translocation is de novo, as is true for all the previously described cases with translocation breakpoints involving the SNPRN gene. Given the PWS-like phenotype, the translocation was assumed to be of paternal origin. This assumption was confirmed by the expression studies. Paternal origin of the translocation was formally proven in 2 of the 5 previously reported cases [26,27].
Karyotype – phenotype correlations
Two individuals with SNRPN intron 2 breakpoints were described as having classical PWS, meeting the major clinical criteria by age 3.5 years and additional minor clinical criteria [25,27]. The individuals with a breakpoint in SNRPN Exon 20/Intron 20 were described as having a milder or atypical form of PWS (Table 3 and Table 4). The weight gain started later than in classical PWS, at 7 and 5 years, respectively, for the patients described by Schulze et al. 1996 and Wirth et al. 2001, and at 8 years in our case. The characteristic facial features were absent in the case of Wirth et al. 2001, and also in the present case. But this is not a consistent feature in classical PWS, as in a retrospective evaluation of 90 molecularly-proven PWS cases, only 49% had the characteristic facial gestalt [39].
It is apparent from the review of the previously reported cases and the individual reported here (Table 3 and Table 4) that some of these translocation cases tend to have a milder, 'atypical' clinical picture, in comparison with classical PWS. There is not a complete absence of any of the major phenotypic features (neonatal hypotonia and feeding difficulty, hyperphagia from early childhood, obesity, cognitive compromise, hypogenitalism), but the degree of affection may be lower. None of the reported translocation cases had any additional features that might possibly be attributed to disruption of a gene on the reciprocal chromosome, and in no prior case had an attempt been made to identify a gene at this location.
Our sequence data mapped the breakpoint on chromosome 4 within intron 10 of a spliced polyadenylated transcript (BC045668). This unique cDNA clone represents a 3764 bp mRNA from a human testis library that does not appear to encode a protein. Its 5' end overlaps the interleukin 21 (IL21) transcript by 511 bp in the opposite direction (UCSC Genome Browser, May 2004). It appears unlikely that heterozygous disruption of this gene contributes to the phenotype in our patient.
Translocation has no effect on imprinting center methylation and upstream genes
To assess whether the translocation event had affected the allele-specific methylation pattern at the imprinting center (IC), and/or to exclude a coincident imprinting defect, we carried out methylation studies of the SNRPN exon 1 region that revealed a normal bi-parental methylation pattern. Similar results were reported for each of the other five PWS individuals who had translocation breakpoints within the SNRPN gene. These results predict that expression of the genes located centromeric to the SNRPN exon 1/ IC region, NDN, MAGEL2 and MKRN3, should not be affected in these individuals. By studying t(4;15) fibroblasts by RT-PCR, we indeed found expression of all three genes. Previously, only MKRN3 was reported to be expressed in the three PWS translocation cases in which it was studied [20,25,27].
In t(4;15) lymphoblasts, the SNRPN transcript was detectable by RT-PCR and quantitative RT-PCR and found to extend all the way to exon 17. The major transcript that encodes the SNURF/SNRPN proteins terminates in exon 10 [20] and, therefore, should be unaffected by this t(4;15) translocation. With the caveat that studies on peripheral tissues, fibroblasts and lymphoblasts, may not accurately reflect gene expression in the brain, our results indicate that SNURF/SNRPN and the centromeric genes MKRN3, NDN and MAGEL2 are unlikely to play a prime role in the causation of PWS-associated features, although it remains an open question whether their loss or non-functioning might contribute to the more marked phenotypic expression that is seen in typical PWS.
Genes downstream of the breakpoint are not expressed
With respect to expression of downstream transcripts, the reported results on LCLs with breakpoints in exon 20/intron 20 were consistent, whereas for the two patients with breakpoints in intron 2, the reported results were conflicting for expression of downstream transcripts IPW and PAR-1. In a re-evaluation of the t(4;15) case reported by Kuslich and colleagues [26], no expression of these transcripts and of the PWCR1/HBII-85 snoRNA cluster was detected by real-time quantitative RT-PCR [28].
Therefore, we focused our analysis on the snoRNAs and two ESTs in intron 20. As for the intron-encoded C/D box snoRNAs, HBII-13 and HBII-437 were expressed, and HBII-438A/B and HBII-85/PWCR1 were not. HBII-52 snoRNAs were not studied, as they are not expressed in the available tissues and have previously been excluded from contributing to the PWS phenotype [40,41]. The two ESTs in the large intron 20 that are highly expressed in brain tissues [42] were found to be expressed in a normal control LCL, but not in the t(4;15) LCL. This result suggests that these ESTs do not have their own promoter but are dependent on transcription from the SNRPN promoter that is located on the other translocation derivative in these cells. Therefore, these ESTs most likely represent stable derivatives of large alternatively spliced non-coding SNRPN transcripts.
Conclusion
(1) Expression of the ESTs and/or C/D box snoRNAs that are located downstream of the translocation breakpoint is not necessary for establishing and maintaining the paternal-specific pattern of gene expression pattern that is controlled by the imprinting center upstream of the translocation breakpoint.
(2) The C/D box snoRNAs HBII-438A and PWCR1/HBII-85 are the only stable transcripts in this region that are disrupted in this t(4;15) PWS individual. As PWCR1/HBII-85 sequences are highly conserved between human and mice, while no copy of HBII-438A has been found in mouse, we conclude that the basis of PWS pathogenesis resides, in whole or in part, in the absence of PWCR1/HBII-85 snoRNA. SNURF/SNRPN and the centromeric genes MKRN3, NDN and MAGEL2 are unlikely to play a major role in the causation of PWS-associated features. While the function of known C/D box snoRNAs is to guide 2'- O -ribose methylation of mainly ribosomal RNA, these novel imprinted snoRNAs have no known target. They might be involved in a posttranscriptional regulation process of a gene or genes that – if non-functional – gives rise to the PWS phenotype.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
BS carried out the molecular genetic studies (RT-PCR, methylation assay, Southern blot analysis, and breakpoint analysis) and drafted the manuscript. MA carried out quantitative RT-PCR assays. EN performed the FISH analysis with BAC clones. DIF carried out the initial cytogenetic analysis. MR revised the clinical data and re-examined the patient.
HRS supervised the cell culturing, cytogenetic and FISH studies. RJMG diagnosed the patient, collected the clinical data and obtained skin and blood samples. UF conceived the study design, and coordinated its progress, supervised the work of BS and MA and prepared the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We are indebted to the family participating in this study and to Prof. George A. Werther who referred the patient to us, and in his letter wrote "I wonder whether this translocation may involve the Prader-Willi gene". The work in the laboratory of UF was supported by grants from the NIH (HD41623) and the Deutsche Forschungsgemeinschaft (BS – SCHU 1567/1-1).
Figures and Tables
Figure 1 Mapping the t(4;15) breakpoint and expression patterns of SNRPN exons and intronic genes . a . Schematic map of human chromosome region 15q11-q13. Black and gray circles represent imprinted genes, expressed from the paternal or maternal allele, respectively. White circles designate bi-allelically expressed genes. BP1, 2, and 3 indicate the locations of the deletion breakpoint hotspots [43]. b . FISH results placed BAC RP11-160D9 highlighted in green (nucleotide position 22577151-22735621) proximal to the translocation breakpoint and RP11-876N20 highlighted in blue (position 22857334-23036552) distal to the breakpoint. Intron 17, comprising nucleotides 22795282 to 22811656, thus is located ~ 63.4 kb downstream of RP11-160D9 and ~ 42 kb upstream of RP11-876N20. c . On representation of the SNRPN region (not drawn to scale) boxes represent exons and ESTs, lines represent snoRNA copies. Orange boxes and lines indicate exons, ESTs or snoRNAs tested for expression either by RT-PCR or quantitative RT-PCR. Black flash indicates the breakpoint in intron 17 of the SNRPN locus.
Figure 2 t(4;15) carrier at 15 years of age . Note absence of typical PWS facial features and presence of mild truncal obesity.
Figure 3 a. High resolution G-banded ideograms and prometaphase chromosomes of the translocation derivatives and their normal homologs . An apparently balanced translocation t(4;15)(q27;q11) was identified with arrows indicating band location of breakpoints. b. DNA methylation analysis of CpG island of SNRPN promoter and exon 1. 1. The 174 bp PCR product is derived from the methylated maternal chromosome. 2. The 100 bp product is derived from the paternal chromosome. PWS: PWS control, Normal: normal control, and t(4;15) carrier; H2O: no template control. The t(4;15) carrier shows the normal bi-parental methylation pattern.
Figure 4 SNRPN expression analysis by RT-PCR of RNA from LCLs . On the left, the sizes of the PCR products are shown, and on the right, the location of the primers in SNRPN exons is listed. +RT: with reverse transcriptase; -RT: without reverse transcriptase; H2O: no template control. All SNRPN +RT products tested were absent in the PWS control, and present in the normal control. The t(4;15) cells were positive for SNURF/ SNRPN exons 2–3, 15–16 and 16–17 and negative for exons 18 through 20a. GAPDH primers were used as control for the integrity of the cDNA.
Figure 5 Southern blot analysis identifies breakpoint in SNRPN intron 17 . a . Restriction map of the intron 17 region of the SNRPN gene on the normal chromosome 15. Black arrowheads indicate the boundaries of intron 17. The positions of the two hybridization probes (SB-1 and SB-3) are indicated by green lines. b . Lanes 1 and 2 contain double digests with NheI and BsaWI to release a fragment of 6.4 kb, lanes 3 and 4 contain double digests with NheI and ApaI to release a fragment of 10 kb. The membrane was probed with probe SB-1. The arrow indicates an additional band above the 10 kb fragment ~11.5 kb in length. The two bands are not well resolved on the rendition of this blot. This novel fragment is represented in c, upper panel. Lanes 5 and 6 contain double digests with NheI and ApaI to release a 10 kb fragment. The membrane is probed with SB-3. The arrow indicates an additional band of ~ 7 kb. This novel fragment is represented schematically in c, lower panel. c . Schematic representation of the junction fragments identified on the Southern blot in b. The upper panel represents the der(15) and the lower panel represents the der(4). Chromosome 15 material is indicated as a black line and material from chromosome 4 as a blue line. Location of restriction sites and of hybridization probes (green lines) are indicated.
Figure 6 Repeat sequences surrounding the breakpoint . a . One hundred nucleotides on either side of the breakpoints on chromosome 4 and 15 contain repetitive sequences (grey lines). The Alu-DEIN sequence is located 13–39 bp upstream of the breakpoint on chromosome 15. b . Sequence across the breakpoint on the der(4) chromosome reveals an additional A inserted at the breakpoint. Arrows indicate the direction centromere to telomere.
Figure 7 Expression analysis in LCLs of snoRNAs and two ESTs in intron 20 . RT-PCR analysis of the C/D box snoRNAs reveals expression of HBII-13, but not of HBII-438A/B, PWCR1/HBII-85 and the two ESTs in intron 20 in the t(4;15) translocation carrier. +RT: with reverse transcriptase; -RT: without reverse transcriptase; H2O: no template control.
Table 1 Primers and conditions for PCR
RT-PCR
Gene/Exon fwd (5'-3') rev (5'-3'); complement strand size T ann.
ZNF127 GGG TTG CGG TTT TGC TAT TA TTT CTC GTG TGC TTC AAT GC 168 bp 59C
MAGEL2 CTG AAG CCT GGG ACT TTC TG GGA CCT TGG CCA CAA ACT TA 225 bp 59C
Necdin GAA GAA GCA CTC CAC CTT CG CCA TGA TTT GCA TCT TGG TG 164 bp 59C
SNRPN Ex 1–3 ATG GAG CGG GCA AGG GAT CGC GGT ACA ACT GAC ACT CTT GG 124 bp 53C
SNRPN Ex 14/16 CTG CAA ACA TAG GAG ATG ATA GTT CC CTT ATG AAA GCA CTG AGA TGA AGC C 459 bp 53C
SNRPN Ex 16/17 GAA AGT GAC CTA AAG AGT GTC ATT G CTT GCA GTT GGA CAG CCG ACT C 515 bp 53C
SNRPN Ex 17/18 AGA TAT CTT TAA AAT TGA GTC TTC TGT CCA TGA AGA TGC AGC ACT TTT GAA GAA 218 bp 53C
SNRPN Ex 19/20a CAT TGT GCT TAT TTA CTA TTT TTG TAG ACG CTG CAG GTG GTG ACC ATG TG 150 bp 53C
AK094315 TCT TCT CTA CCC TCA TTC CCA GC TCG CTA CAC CCC TTT GCT TAT G 222 bp 53C
AB061718 AGG AGG GGT TCA AAG ATG C CTG GTA AAC AAA CTG GTA AAG GTG 204 bp 50C
HBII-85/PWCR1 CGA TGA TGA GTC CCC CAT AAA AAC CAG TTC CGA TGA GAA CGA CG 79 bp 53C
HBII-13 GGA TTT GTG ATG AGC TGT GTT TAC GGA CTT CAG AGT AAT CAC GTT G 67 bp 54C
HBII-438A/B GGA TCG ATG ATG AGA ATA ATT ATT G GGA CCT CAG ATT GAC ATC TG 67 bp 53C
GABRB3 TCA GGC GGC ATT GGC GAT ACC ATA AAA ACT TGA CAG GCA GAG 352 bp 52C
GABRA5 AAT ATT GCC TTA ATG TTT CTA GCC TAT TCT ATT TCT TCG TGT 425 bp 48C
GABRG3 GCG TAT TCA CAT AGA CAT CTT G GAT TGG TCA CTA CTG GTC TGG 188 bp 52C
GAPDH TGG GCT ACA CTG AGC ACC AG GGG TGT CGC TGT TGA AGT CA 50 bp 53C
Quantitative RT-PCR
Gene/Exon fwd (5'-3') rev (5'-3');complement strand size T ann.
SNURF Ex2 ACG AAC TAC AGA ACA GCA CGT ACC CTG CGT TTG ACT TGG ACT TCC 50 bp 60C
SNURF Ex3 TTC TCA GCA GCA GCA AGT ACC T TGC CTC AGT TCA GCC TGG A 50 bp 60C
HBII-437 ATC ATT ATT TCT TGA ATT GG CCC TCA CGC TCC CTT TGC 50 bp 60C
SNRPN Ex 14/15 CTG CAA ACA TAG GAG ATG ATA GTT CC CAA AGA CGA TAA AAT GTT CCT TCT TG 50 bp 60C
SNRPN Ex 19/20a GGA ACC ACC ATT TGT CTA TGA TCC CTG CAG GTG GTG ACC ATG TG 50 bp 60C
HBII-438 ATA ATT GTC TGA GGA TGC T GAT TGA CAT CTG GAA TGA GTC 50 bp 60C
HBII-85/PWCR1 TCG ATG ATG AGT CCC CCA TAA CAT TTT GTT CAG CTT TTC CAA GG 50 bp 60C
PCR to generate Southern probes in intron 17
Gene/Exon fwd (5'-3') rev (5'-3');complement strand size T ann.
SB 1 ACC ATC AGT GAA TGA CCT GTT GC CCC AGC CTC TTT CCT ATG TCT TG 565 bp 53C
SB 3 TGG TAA ACT GAT GAG AGC ACA GCC GCC TGG GAG ACA GAA TGA GAA AC 416 bp 53C
Table 2 SNRPN and snoRNA expression analysis with quantitative RT-PCR
Amplification product PWS Normal control t-PWS (4;15) PWS triplication t-PWS intron 2 t-PWS (4;15)
LCL LCL LCL LCL FB FB
SNURF Ex 2 0.0001 0.53 0.54 2 0.002 0.56
SNURF Ex 3 0.0002 1.1 1.43 6.4 0.0004 0.56
HBII-437 0.00003 0.86 1.13 4.9 0.00008 0.1
SNRPN 14/15 0.0013 4.7 4.91 17.23 - -
SNRPN 19/20a 0.005 1.34 0.003 4.7 0.007 0.007
HBII-438 0.03 1.5 0.02 6.2 0.07 0.07
PWCR1/HBII-85 0.03 3.7 0.02 16.8 0.04 0.06
Sample identification: PWS: PWS with an IC microdeletion (patient E in [5]); t(4;15) PWS: the PWS case reported here; PWS-triplication: intrachromosomal triplication of the PWS region [33]; t-PWS intron 2: previously reported PWS case with t(4;15)(q27;q11.2) and breakpoint in SNRPN intron 2 [26, 27]; LCL, lymphoblastoid cell line, FB, fibroblast strain. The numbers represent the ratio of target product to GAPDH control product.
Table 3 Clinical findings associated with paternally-derived de novo reciprocal translocations involving SNRPN
Breakpoint in SNRPN Intron 2 Breakpoint in SNRPN Exon 20/ Intron 20 Breakpoint in SNRPN Intron 17
Sun et al. 1996 Kuslich et al. 1999 Schulze et al. 1996 Conroy et al. 1997 Wirth et al. 2001 Present case
Karyotype designations 46, XY, t(15;19) (q12;q13.41) 46, XY, t(4;15) (q27;q11.2) 46, XY, t(9;15) (q21;q12–q13) 46, XY, t(2;15) (q37.2;q11.2) 46, X, t(X;15) (q28;q12) 46 XY, t(4;15)(q27;q11.2)
Age of examination 3.5 years 3 years 3 months 29 years 4.5 years 18 years 22 years
Major criteria (each scores one point) from [1] as revised in [2].
1. Neonatal central hypotonia Floppy and lethargic in the first 6 months with poor suck (1pt.) Hypotonicity, poor sucking reflex during infancy (1pt.) Neonatal hypotonia (weak cry, poor suck) (1 pt.) Neonatal hypotonia, lethargy, poor suck (1 pt.) - Reduced tone with poor head control, poor suck (1 pt.)
2. Infantile feeding problems/ failure to thrive Failure to thrive (1pt.) Feeding problems in infancy, failure to thrive (1 pt.) Special feeding techniques, but no failure to thrive - Feeding problems, but no failure to thrive (1pt.)
3. Rapid weight gain between 1–6 years Obesity starting at 6 months, hyperphagia (1 pt.) Eating behavior leading to increased weight gain at age 2 yr (1 pt.) Periodic excessive weight gain from age 7 yr Onset of obesity at 1.5–2 yr with excessive appetite and food foraging (1 pt.) Obesity began at 4–5 yr with hyperphagia and food foraging (1 pt.) Late onset obesity (at approx. 8 yr)
4. Characteristic facial features Narrow bifrontal diameter, almond-shaped eyes, down-turned mouth (1pt.) Narrow bifrontal diameter, almond-shaped eyes, upslanted palpebral fissures (1 pt.) Narrow bifrontal diameter, narrow face, small mouth, poor facial mimic (1pt.) Narrow bifrontal diameter, squared nasal tip, downturned mouth (1 pt.) - -
5. Hypogonadism: genital hypoplasia, pubertal deficiency Undescended testes (1 pt.) Undescended small testes, hypogonadism (1 pt.) Hypoplastic genitalia, incomplete gonadal maturation with delayed pubertal signs after age 16 yr (1 pt.) Scrotum normal, penile length at 10th %ile Primary amenorrhea, hypoplastic uterus (1 pt.) Undescended small testes, hypogonadism, delayed pubertal signs (1 pt.)
6. Mental retardation, developmental delay Developmental delay (1 pt.) Developmental delay (1 pt.) Mental retardation, developmental delay/ learning problems (1 pt.) Developmental delay, special school setting (1 pt.) Slight developmental delay, school for mentally retarded children (1 pt.) Developmental delay, special school setting (1 pt.)
Score 5 points 6 points 5 points 4 points 3 points 4 points
Blank cell = no information
- = absent
Table 4 Clinical findings associated with paternally-derived de novo reciprocal translocations involving SNRPN (continued)
Sun et al. 1996 Kuslich et al. 1999 Schulze et al. 1996 Conroy et al. 1997 Wirth et al. 2001 Present case
Minor criteria (1/2 point each)
1. Decreased fetal movement and infantile lethargy Decreased fetal activity (0.5 pt.) Decreased fetal movements (0.5pt.) - - Slightly reduced fetal movements (0.5pt.)
2. Typical behaviour problems Behavior problems (0.5pt.) Temper tantrums, violent outbursts, obsessive-compulsive (0.5 pt.) Aggressive outbursts, rigid personality, perseveration (0.5pt.) Behavior problems with temper tantrums and severe aggressiveness (0.5 pt.) Temper tantrums, violent outbursts after food restrictions (0.5pt.) Temper tantrums, abnormal social behavior (0.5pt.)
3. Sleep disturbance, sleep apnea Sleep disturbance, sleep apnea (0.5pt.) Sleep disturbance (0.5pt.) Sleep disturbance, amphetamine treatment from age 9 ys. (0.5pt.)
4. Short stature for the family by age 15 years Short stature at the age of 15 (0.5pt.) 50–75th percentile (0.5pt.) 151 cm (3rd%tile) (0.5pt.) Height 155.7 cm at 16 years < 3rd %tile (0.5 pt.)
5. Hypopigmentation - - Hypopigmentation (0.5 pt.) - -
6. Small hands and /or feet for height age Hand length 25th percentile, finger length 10th%ile (0.5pt.) - Normal hands, but small feet (< 10th%tile) (0.5 pt.) Short 3rd finger bilaterally Hands 20th %ile, feet 5th %ile (0.5pt.)
7. Narrow hands with straight ulnar border - -
8. Eye abnormalities: esotropia, myopia - Esotropia (0.5 pt.) Alternating esotropia in infancy (0.5 pt.) Left esotropia (0.5 pt.) Esotropia (0.5pt.)
9. Thick viscous saliva Viscous saliva (0.5pt.) Thick viscous saliva (0.5 pt.) -
10. Speech articulation defect Articulation difficulty (0.5 pt.) Poor articulation (0.5pt.) -
11. Skin picking Skin picking (0.5pt.) Skin picking (0.5pt.) Skin picking (0.5pt.)
Score (minor only) 1.5 points 3 points 3.5 points 2.5 points 1.5 points 3.5 points
Total Score 6.5 points 9 points 8.5 points 8.5 points 4.5 points 7.5 points
Blank cell = no information
- = absent
==== Refs
Holm VA Cassidy SB Butler MG Hanchett JM Greenswag LR Whitman BY Greenberg F Prader-Willi syndrome: consensus diagnostic criteria Pediatrics 1993 91 398 402 8424017
Cassidy SB Prader-Willi syndrome 2001 301 322
Ledbetter DH Riccardi VM Airhart SD Strobel RJ Keenan BS Crawford JD Deletions of chromosome 15 as a cause of the Prader-Willi syndrome N Engl J Med 1981 304 325 329 7442771
Buiting K Gross S Lich C Gillessen-Kaesbach G el-Maarri O Horsthemke B Epimutations in Prader-Willi and Angelman syndromes: a molecular study of 136 patients with an imprinting defect Am J Hum Genet 2003 72 571 577 12545427 10.1086/367926
Mascari MJ Gottlieb W Rogan PK Butler MG Waller DA Armour JA Jeffreys AJ Ladda RL Nicholls RD The frequency of uniparental disomy in Prader-Willi syndrome. Implications for molecular diagnosis N Engl J Med 1992 326 1599 1607 1584261
Sutcliffe JS Nakao M Christian S Orstavik KH Tommerup N Ledbetter DH Beaudet AL Deletions of a differentially methylated CpG island at the SNRPN gene define a putative imprinting control region [see comments] Nat Genet 1994 8 52 58 7987392 10.1038/ng0994-52
Buiting K Saitoh S Gross S Dittrich B Schwartz S Nicholls RD Horsthemke B Inherited microdeletions in the Angelman and Prader-Willi syndromes define an imprinting centre on human chromosome 15 [published erratum appears in Nat Genet 1995 10:249] Nat Genet 1995 9 395 400 7795645 10.1038/ng0495-395
Jong MT Carey AH Caldwell KA Lau MH Handel MA Driscoll DJ Stewart CL Rinchik EM Nicholls RD Imprinting of a RING zinc-finger encoding gene in the mouse chromosome region homologous to the Prader-Willi syndrome genetic region Hum Mol Genet 1999 8 795 803 10196368 10.1093/hmg/8.5.795
Jong MT Gray TA Ji Y Glenn CC Saitoh S Driscoll DJ Nicholls RD A novel imprinted gene, encoding a RING zinc-finger protein, and overlapping antisense transcript in the Prader-Willi syndrome critical region Hum Mol Genet 1999 8 783 793 10196367 10.1093/hmg/8.5.783
Boccaccio I Glatt-Deeley H Watrin F Roeckel N Lalande M Muscatelli F The human MAGEL2 gene and its mouse homologue are paternally expressed and mapped to the Prader-Willi region Hum Mol Genet 1999 8 2497 2505 10556298 10.1093/hmg/8.13.2497
Lee S Kozlov S Hernandez L Chamberlain SJ Brannan CI Stewart CL Wevrick R Expression and imprinting of MAGEL2 suggest a role in prader-willi syndrome and the homologous murine imprinting phenotype Hum Mol Genet 2000 9 1813 1819 10915770 10.1093/hmg/9.12.1813
Jay P Rougeulle C Massacrier A Moncla A Mattei MG Malzac P Roeckel N Taviaux S Lefranc JL Cau P Berta P Lalande M Muscatelli F The human necdin gene, NDN, is maternally imprinted and located in the Prader-Willi syndrome chromosomal region Nat Genet 1997 17 357 361 9354807
MacDonald HR Wevrick R The necdin gene is deleted in Prader-Willi syndrome and is imprinted in human and mouse Hum Mol Genet 1997 6 1873 1878 9302265 10.1093/hmg/6.11.1873
Özçelik T Leff SE Robinson WP Donlon T Lalande M Sanjines E Schinzel A Francke U Small nuclear ribonucleoprotein polypeptide N (SNRPN), an expressed gene in the Prader-Willi syndrome critical region Nature Genet 1992 2 265 269 1303277 10.1038/ng1292-265
Nakao M Sutcliffe JS Durtschi B Mutirangura A Ledbetter DH Beaudet AL Imprinting analysis of three genes in the Prader-Willi/Angelman region: SNRPN, E6-associated protein, and PAR-2 (D15S225E) Hum Mol Genet 1994 3 309 315 8004100
Reed ML Leff SE Maternal imprinting of human SNRPN, a gene deleted in Prader-Willi syndrome Nat Genet 1994 6 163 167 7512861 10.1038/ng0294-163
Glenn CC Saitoh S Jong MT Filbrandt MM Surti U Driscoll DJ Nicholls RD Gene structure, DNA methylation, and imprinted expression of the human SNRPN gene Am J Hum Genet 1996 58 335 346 8571960
Gray TA Saitoh S Nicholls RD An imprinted, mammalian bicistronic transcript encodes two independent proteins Proc Natl Acad Sci U S A 1999 96 5616 5621 10318933 10.1073/pnas.96.10.5616
Runte M Huttenhofer A Gross S Kiefmann M Horsthemke B Buiting K The IC-SNURF-SNRPN transcript serves as a host for multiple small nucleolar RNA species and as an antisense RNA for UBE3A Hum Mol Genet 2001 10 2687 2700 11726556 10.1093/hmg/10.23.2687
Wirth J Back E Huttenhofer A Nothwang HG Lich C Gross S Menzel C Schinzel A Kioschis P Tommerup N Ropers HH Horsthemke B Buiting K A translocation breakpoint cluster disrupts the newly defined 3' end of the SNURF-SNRPN transcription unit on chromosome 15 Hum Mol Genet 2001 10 201 210 11159938 10.1093/hmg/10.3.201
de Los Santos T Schweizer J Rees CA Francke U Small evolutionarily conserved RNA, resembling C/D box small nucleolar RNA, is transcribed from PWCR1, a novel imprinted gene in the Prader-Willi deletion region, which is highly expressed in brain Am J Hum Genet 2000 67 1067 1082 11007541 10.1086/303106
Cavaille J Buiting K Kiefmann M Lalande M Brannan CI Horsthemke B Bachellerie JP Brosius J Huttenhofer A Identification of brain-specific and imprinted small nucleolar RNA genes exhibiting an unusual genomic organization Proc Natl Acad Sci U S A 2000 97 14311 14316 11106375 10.1073/pnas.250426397
Kiss T Small Nucleolar RNAs: An Abundant Group of Noncoding RNAs with Diverse Cellular Functions Cell 2002 109 145 148 12007400 10.1016/S0092-8674(02)00718-3
Schulze A Hansen C Skakkebaek NE Brondum-Nielsen K Ledbeter DH Tommerup N Exclusion of SNRPN as a major determinant of Prader-Willi syndrome by a translocation breakpoint Nat Genet 1996 12 452 454 8630505 10.1038/ng0496-452
Sun Y Nicholls RD Butler MG Saitoh S Hainline BE Palmer CG Breakage in the SNRPN locus in a balanced 46,XY,t(15;19) Prader-Willi syndrome patient Hum Mol Genet 1996 5 517 524 8845846 10.1093/hmg/5.4.517
Conroy JM Grebe TA Becker LA Tsuchiya K Nicholls RD Buiting K Horsthemke B Cassidy SB Schwartz S Balanced translocation 46,XY,t(2;15)(q37.2;q11.2) associated with atypical Prader-Willi syndrome Am J Hum Genet 1997 61 388 394 9311744
Kuslich CD Kobori JA Mohapatra G Gregorio-King C Donlon TA Prader-Willi syndrome is caused by disruption of the SNRPN gene Am J Hum Genet 1999 64 70 76 9915945 10.1086/302177
Gallagher RC Pils B Albalwi M Francke U Evidence for the role of PWCR1/HBII-85 C/D box small nucleolar RNAs in Prader-Willi syndrome Am J Hum Genet 2002 71 669 678 12154412 10.1086/342408
Li L Moore P Ngo C Petrovic V White SM Northrop E Ioannou PA McKinlay Gardner RJ Slater HR Identification of a haplosufficient 3.6-Mb region in human chromosome 11q14.3-->q21 Cytogenet Genome Res 2002 97 158 162 12438706 10.1159/000066612
Frommer M McDonald LE Millar DS Collis CM Watt F Grigg GW Molloy PL Paul CL A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands Proc Natl Acad Sci U S A 1992 89 1827 1831 1542678
Kubota T Das S Christian SL Baylin SB Herman JG Ledbetter DH Methylation-specific PCR simplifies imprinting analysis Nat Genet 1997 16 16 17 9140389
Bustin SA Absolute quantification of mRNA using real-time reverse transcription polymerase chain reaction assays J Mol Endocrinol 2000 25 169 193 11013345 10.1677/jme.0.0250169
Ginzinger DG Gene quantification using real-time quantitative PCR: an emerging technology hits the mainstream Exp Hematol 2002 30 503 512 12063017 10.1016/S0301-472X(02)00806-8
Ungaro P Christian SL Fantes JA Mutirangura A Black S Reynolds J Malcolm S Dobyns WB Ledbetter DH Molecular characterisation of four cases of intrachromosomal triplication of chromosome 15q11-q14 J Med Genet 2001 38 26 34 11134237 10.1136/jmg.38.1.26
Siebert PD Chenchik A Kellogg DE Lukyanov KA Lukyanov SA An improved PCR method for walking in uncloned genomic DNA Nucleic Acids Res 1995 23 1087 1088 7731798
Rüdiger NS Gregersen N Kielland-Brandt MC One short well conserved region of Alu-sequences is involved in human gene rearrangements and has homology with prokaryotic chi Nucleic Acids Res 1995 23 256 260 7862530
Dower NA Stahl FW Chi activity during transduction-associated recombination Proc Natl Acad Sci U S A 1981 78 7033 7037 7031667
Deininger PL Jolly DJ Rubin CM Friedmann T Schmid CW Base sequence studies of 300 nucleotide renatured repeated human DNA clones J Mol Biol 1981 151 17 33 6276559 10.1016/0022-2836(81)90219-9
Gunay-Aygun M Schwartz S Heeger S O'Riordan MA Cassidy SB The changing purpose of Prader-Willi syndrome clinical diagnostic criteria and proposed revised criteria Pediatrics 2001 108 E92 11694676 10.1542/peds.108.5.e92
Hamabe J Kuroki Y Imaizumi K Sugimoto T Fukushima Y Yamaguchi A Izumikawa Y Niikawa N DNA deletion and its parental origin in Angelman syndrome patients Am J Med Genet 1991 41 64 68 1683160
Runte M Varon R Horn D Horsthemke B Buiting K Exclusion of the C/D box snoRNA gene cluster HBII-52 from a major role in Prader-Willi syndrome Hum Genet 2005 116 228 230 15565282 10.1007/s00439-004-1219-2
Ota T Suzuki Y Nishikawa T Otsuki T Sugiyama T Irie R Wakamatsu A Complete sequencing and characterization of 21,243 full-length human cDNAs Nature Genetics 2004 36 40 45 14702039 10.1038/ng1285
Chai JH Locke DP Greally JM Knoll JH Ohta T Dunai J Yavor A Eichler EE Nicholls RD Identification of four highly conserved genes between breakpoint hotspots BP1 and BP2 of the Prader-Willi/Angelman syndromes deletion region that have undergone evolutionary transposition mediated by flanking duplicons Am J Hum Genet 2003 73 898 925 14508708 10.1086/378816
| 15877813 | PMC1142316 | CC BY | 2021-01-04 16:03:32 | no | BMC Med Genet. 2005 May 6; 6:18 | utf-8 | BMC Med Genet | 2,005 | 10.1186/1471-2350-6-18 | oa_comm |
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-191588246910.1186/1471-2350-6-19Technical AdvanceCYP3A4 and CYP3A5 genotyping by Pyrosequencing Garsa Adam A [email protected] Howard L [email protected] Sharon [email protected] Washington University School of Medicine, Department of Medicine, Division of Oncology, St. Louis, MO 63110, USA2 The Siteman Cancer Center, St. Louis, MO 63110, USA2005 9 5 2005 6 19 19 28 9 2004 9 5 2005 Copyright © 2005 Garsa 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
Human cytochrome P450 3A enzymes, particularly CYP3A4 and CYP3A5, play an important role in drug metabolism. CYP3A expression exhibits substantial interindividual variation, much of which may result from genetic variation. This study describes Pyrosequencing assays for key SNPs in CYP3A4 (CYP3A4*1B, CYP3A4*2, and CYP3A4*3) and CYP3A5 (CYP3A5*3C and CYP3A5*6).
Methods
Genotyping of 95 healthy European and 95 healthy African volunteers was performed using Pyrosequencing. Linkage disequilibrium, haplotype inference, Hardy-Weinberg equilibrium, and tag SNPs were also determined for these samples.
Results
CYP3A4*1B allele frequencies were 4% in Europeans and 82% in Africans. The CYP3A4*2 allele was found in neither population sample. CYP3A4*3 had an allele frequency of 2% in Europeans and 0% in Africans. The frequency of CYP3A5*3C was 94% in Europeans and 12% in Africans. No CYP3A5*6 variants were found in the European samples, but this allele had a frequency of 16% in the African samples. Allele frequencies and haplotypes show interethnic variation, highlighting the need to analyze clinically relevant SNPs and haplotypes in a variety of ethnic groups.
Conclusion
Pyrosequencing is a versatile technique that could improve the efficiency of SNP analysis for pharmacogenomic research with the ultimate goal of pre-screening patients for individual therapy selection.
==== Body
Background
The human cytochrome P450 3A (CYP3A) subfamily of enzymes plays an important role in drug metabolism. The four CYP3A genes lie within a 218 kb region of chromosome 7q22.1 in the following order: CYP3A5, CYP3A7, CYP3A4, and CYP3A43. CYP3A enzymes, primarily CYP3A4 and CYP3A5, catalyze the metabolism of a multitude of exogenous and endogenous compounds. As the most abundant group of CYPs in the liver and small intestine, CYP3A enzymes strongly affect the oral bioavailability and clearance of many drugs, and it is estimated that CYP3A enzymes are involved in the metabolism of over half of the drugs currently approved by the Food and Drug Administration [1-4].
Interindividual variation in CYP3A expression is substantial. Protein expression in liver and small intestine varies up to 40-fold, leading to variation in drug metabolism [2,5]. Genetic variation within the CYP3A genes may contribute to interindividual variability in drug metabolism. It has been suggested that approximately 90% of inter-individual differences in hepatic CYP3A activity are due to genetic variation [6]. Single nucleotide polymorphisms (SNPs) are the most common form of genetic variation in the CYP3A genes.
CYP3A4*1B is a 5' untranslated region -392A>G transition in CYP3A4 [7]. A number of associations between CYP3A4*1B and clinical phenotypes have been found. Rebbeck et al. have shown that prostate cancer patients are more likely to have the CYP3A4*1B allele than healthy controls, and this has been confirmed in other studies [7-9]. Additionally, homozygous wild-type (CYP3A4*1A/*1A) individuals have an increased risk for developing leukemia after epipodophyllotoxin therapy [10].
Relatively little is known about the effects of the other commonly studied CYP3A4 SNPs, CYP3A4*2 and CYP3A4*3. CYP3A4*2 is a SNP in exon 7 (15713T>C) that results in a Ser222Pro change. In vitro kinetic studies have shown that CYP3A4*2 has a 6-to 9-fold reduced intrinsic clearance for nifedipine compared to wild-type [11]. CYP3A4*3 is a 1334T>C transition causing a Met445Thr change. Although this SNP occurs within a conserved region, no difference in testosterone, progesterone, or 7-benzyloxy-4(trifluoromethyl)coumarin metabolism was found [12].
CYP3A5*3C is an IVS3-237A>G (6986A>G) transition within intron 3 of CYP3A5 [13]. This transition creates an alternative splice site in the pre-mRNA, leading to the production of aberrant mRNA with a premature stop codon [13]. This SNP leads to polymorphic expression of CYP3A5. CYP3A5*3C homozygotes lack CYP3A5 expression, while individuals with at least one CYP3A5*1 wild-type allele express CYP3A5 [13]. Polymorphic expression of CYP3A5 may account for some of the interindividual variation in clearance of CYP3A substrates. Indeed, CYP3A5 genotype is predictive of tacrolimus doses for lung and kidney transplant recipients [14,15].
CYP3A5*6 is a 14690G>A synonymous mutation that causes the formation of a splice variant mRNA. Exon 7 is deleted, resulting in a frameshift and a truncated protein [13]. Very little is known about the effects of this SNP, although CYP3A5*6 was found to have no effect on midazolam clearance in a small sample size [16].
For further analysis of these SNPs and their relations to clinical outcomes, an accurate, rapid, and cost efficient method of SNP evaluation is needed. This study describes the use of Pyrosequencing to assay key CYP3A4 and CYP3A5 SNPs.
Methods
Genotyping
PCR was performed on DNA from 95 healthy European volunteers and 95 healthy African volunteers, after IRB approval and written informed consent [17,18]. PCR primers were designed using Primer Express Version 2.0 (ABI, Foster City, CA, USA) and Pyrosequencing Primer SNP Design Version 1.01 software [19]. Primer sequences and PCR conditions are described in Table 1. PCR was carried out using 1–5 ng genomic DNA, 0.6 nmol each of forward and reverse oligonucleotide PCR primers (one of which is biotinylated) (Integrated DNA Technologies, Coralville, Iowa, USA) and 1X AmpliTaq Gold PCR Master Mix (Applied Biosystems, CA, USA), containing 255U (0.05 U/ml) AmpliTaq Gold DNA polymerase, Gene Amp PCR Gold Buffer (30 mmol/L Tris-HCL, 100 mmol/L KCl, pH 8.05), 400 mM dNTP and 5 mmol/L MgCl2. Pyrosequencing was carried out as described [20] using internal primer diluted in 1X Annealing Buffer (20 mmol/L Tris-Acetate, 2 mmol/L MgAc2), 2X Binding Wash Buffer II pH 7.6 (10 mmol/L Tris-HCL, 2M NaCl, 1 mmol/L EDTA, 0.1% Tween20), Streptavidin Sepharose Beads (Amersham Biosciences, Uppsala, Sweden), 0.2 M NaOH,70% Ethanol, and a PSQ HS96 SNP reagent kit (Pyrosequencing AB, Uppsala, Sweden). Samples were analyzed on a PSQ HS96A instrument with pyrosequencing software (Biotage, Uppsala, Sweden). A Tecan pipetting robot (Tecan, Research Triangle Park, NC, USA) was used for all of the steps apart from the addition and transfer of the sepharose beads.
Statistics
Pairwise linkage disequilibrium (|D'|), haplotype inference, and Hardy-Weinberg equilibrium were determined using the Polymorphism and Haplotype Analysis Suite [21,22]. Tag SNPs were determined using SNPtagger [23,24].
Results and discussion
Genotyping data from the European and African samples are shown in Table 2. All results are in Hardy-Weinberg equilibrium. CYP3A4*1B allele frequencies were 4% for Europeans and 82% for Africans. No CYP3A4*2 alleles were found in either the European or African population samples. CYP3A4*3 had an allele frequency of 2% in Europeans and 0% in Africans. The frequency of CYP3A5*3C was 94% in Europeans and 12% in Africans. No CYP3A5*6 variants were found in the European samples, but this allele had a frequency of 16% in the African samples. There were no individuals homozygous for both CYP3A5*3C and CYP3A5*6 in either population.
The CYP3A4 and CYP3A5 genes lie in close proximity (136 kb) to one another on chromosome 7q22.1, so haplotypes were determined across both genes for each population (Table 3). In Europeans, haplotype 1 was the predominant haplotype, with a 90% frequency. Haplotypes 2, 3, and 5 were also observed, with frequencies of 5.5%, 1.8%, and 1%, respectively. Haplotype 4 was not observed, but it has an inferred frequency of 1.8%. The African population had five observed haplotypes: Haplotype 4 (57%), Haplotype 2 (15%), Haplotype 5 (12%), Haplotype 6 (11%), and Haplotype 7 (3%). Haplotype 1 was not observed, but it has an inferred frequency of 0.6%. No loci are significantly linked in either population (data not shown). In addition, no haplotype tag SNPs could be identified in either population. However, genotyping Europeans for CYP3A5*3C could be used to identify the haplotype of 95.5% of the population. Similarly, genotyping Africans for CYP3A4*1B and CYP3A5*3C could be used to identify the haplotype of 84% of the population.
The frequency of CYP3A4*1B in Europeans (4%) is consistent with other studies [25]. The CYP3A4*1B frequency in Africans (82%) is much higher than in Europeans, and it is also higher than the 35–67% frequency seen in African Americans [25]. The rare CYP3A4*2 allele was not found in either of our population samples. To date CYP3A4*2 has only been described in a Finnish Caucasian population, with an allele frequency of 2.7% [11]. The CYP3A4*3 allele frequency in Europeans (2%) is consistent with frequencies reported in other studies [25].
CYP3A5*3C frequency shows dramatic interethnic variation. In Europeans, the CYP3A5*3C variant is the predominant allele (94% frequency), but this allele has a much lower frequency in the African population (12%). CYP3A5*6 frequency also shows interethnic variation. The CYP3A5*6 allele was not found in Europeans, but it was found in the African population at a frequency of 16%. Other studies have also failed to find CYP3A5*6 in Europeans, but this has been found in African Americans at a frequency of 13–16% [16,26].
Haplotype also shows interethnic variation. Haplotype 1 is the predominant haplotype in Europeans, with haplotypes 2–5 occurring at low frequencies. In contrast, haplotype 4 is the most common haplotype in Africans, and haplotypes 2, 5, and 6 all occur at frequencies greater than 10%. However, the presence of homozygous variants was rare for CYP3A4 in Europeans and for CYP3A5 in Africans. Consequently, the limits of in silico determining haplotype frequencies in these populations should be taken into account. Larger population studies are necessary for a more accurate understanding of CYP3A4-CYP3A5 haplotypes. Interethnic variation highlights the need to analyze clinically relevant SNPs and haplotypes in a variety of ethnic groups. An understanding of the genetic variation that exists in various populations will aid in tailoring health care to different populations.
Conclusion
Restriction fragment length polymorphism (RFLP) is the predominant method of SNP analysis used to assay CYP3A4 and CYP3A5 SNPs in previous studies [14,27]. Pyrosequencing offers several advantages over RFLP. For RFLP analysis, a SNP must alter a restriction enzyme cutting site. This limitation precludes many SNPs from RFLP analysis. Pyrosequencing assays can be designed for the vast majority of SNPs, making it a versatile alternative. Pyrosequencing also requires less time than RFLP. Post-PCR, Pyrosequencing steps take approximately 30 minutes for 96 samples, whereas enzyme digestion (1–2 h) and gel electrophoresis for RFLP take significantly longer. Additionally, Pyrosequencing assays are readily transferable to any lab with the appropriate equipment and require no on-site optimization. This procedure could improve the efficiency of SNP analysis for pharmacogenomic research with the ultimate goal of pre-screening patients for individual therapy selection.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AG performed the experimental analysis and drafted the manuscript. AAG, HLM and SM interpreted the results. HLM and SM conceived of the study design and implementation 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
This work was supported by the NIH Pharmacogenetics Research Network (U01 GM63340); , the General Clinical Research Center (P30 CA091842-01), R21 CA102461-01, and P01 CA101937-01. The authors thank Hilary Kannall, Christi Ralph and Chris Rose for technical assistance, and Derek Van Booven for computational expertise. This data has been deposited in the Phamacogenetics Knowledge Base .
Figures and Tables
Table 1 Primer sequence, PCR and Pyrosequencing conditions for CYP3A4 and CYP3A5 SNP analysis.
SNP Forward Primera(5'-3') Reverse Primera (5'-3') Number of Cycles Annealing Temperature (° C) Internal Primerbc(5'-3') Sequence to Analyzed
CYP3A4*1B aggacagcccatagagacaagg *atcaatgttactggggagtcc 55 55 F-ccatagagacaagggca A/GGAGA
CYP3A4*2 aacaatccacaagacccctt *atcttcaaatgtactacaaatcactga 55 55 F-tttggatccattctttc TCTC/TAAT
CYP3A4*3 cgtggaaccagattcagcaa *gaaggagaagttctgaaggactctg 55 65 F-ccagaaactgcattgg CAT/CGAGG
CYP3A5*3C *cccacgtatgtaccacccagc attagggtgtgacacagcaaga 55 65 R-ccaaacagggaagaga TAC/TTGe
CYP3A5*6 *tctttggggcctacagcatg aaagaaataatagcccacatacttattgagag 55 62 R-agaaaccaaattttaggaa CTTC/TTTAGe
a* = biotin molecule attached
bF = forward primer
cR = reverse primer
d Simplex entry nucleotide information for Pyrosequencing
eAssays on reverse complement strand
Table 2 Genotype and allele frequencies for CYP3A4 and CYP3A5 SNPs in European and African populations. Figures in brackets are 95% confidence intervals.
SNP European African
na Wild Type Heterozygous Variant p q na Wild Type Heterozygous Variant p q
CYP3A4*1B 93 86 7 0 0.96 (0.92–0.98) 0.04 (0.02–0.08) 88 3 25 60 0.18 (0.11–0.27) 0.82 (0.73–0.89)
CYP3A4*2 56 56 0 0 1 0 87 87 0 0 1 0
CYP3A4*3 94 90 4 0 0.98 (0.95–0.99) 0.02 (0.01–0.05) 93 93 0 0 1 0
CYP3A5*3C 95 0 12 83 0.06 (0.03–0.10) 0.94 (0.90–0.97) 95 72 23 0 0.88 (0.83–0.92) 0.12 (0.08–0.17)
CYP3A5*6 92 92 0 0 1 0 91 64 25 2 0.84 (0.78–0.89) 0.16 (0.11–0.22)
a n = number of passed samples out of 95 samples tested
Table 3 Haplotype analyses for CYP3A4 and CYP3A5.
CYP3A4*1B CYP3A4*2 CYP3A4*3 CYP3A5*3C CYP3A5*6 % European % African
Hap1 A C T G G 90 0.6a
Hap2 A C T A G 5.5 15
Hap3 A C C G G 1.8 0
Hap4 G C T A G 1.8a 57
Hap5 G C T G G 1 12
Hap6 G C T A A 0 11
Hap7 A C T A A 0 3
a inferred frequency
==== Refs
Shimada T Yamazaki H Mimura M Inui Y Guengerich FP Interindividual variations in human liver cytochrome P-450 enzymes involved in the oxidation of drugs, carcinogens and toxic chemicals: studies with liver microsomes of 30 Japanese and 30 Caucasians J Pharmacol Exp Ther 1994 270 414 423 8035341
Paine MF Khalighi M Fisher JM Shen DD Kunze KL Marsh CL Perkins JD Thummel KE Characterization of interintestinal and intraintestinal variations in human CYP3A-dependent metabolism J Pharmacol Exp Ther 1997 283 1552 1562 9400033
de Waziers I Cugnenc PH Yang CS Leroux JP Beaune PH Cytochrome P 450 isoenzymes, epoxide hydrolase and glutathione transferases in rat and human hepatic and extrahepatic tissues J Pharmacol Exp Ther 1990 253 387 394 2329521
Wrighton SA VandenBranden M Ring BJ The human drug metabolizing cytochromes P450 J Pharmacokinet Biopharm 1996 24 461 473 9131485
Watkins PB Cyclosporine and liver transplantation: will the midazolam test make blood level monitoring obsolete? Hepatology 1995 22 994 996 7657308 10.1016/0270-9139(95)90325-9
Ozdemir V Kalowa W Tang BK Paterson AD Walker SE Endrenyi L Kashuba AD Evaluation of the genetic component of variability in CYP3A4 activity: a repeated drug administration method Pharmacogenetics 2000 10 373 388 10898107 10.1097/00008571-200007000-00001
Rebbeck TR Jaffe JM Walker AH Wein AJ Malkowicz SB Modification of clinical presentation of prostate tumors by a novel genetic variant in CYP3A4 J Natl Cancer Inst 1998 90 1225 1229 9719084 10.1093/jnci/90.16.1225
Paris PL Kupelian PA Hall JM Williams TL Levin H Klein EA Casey G Witte JS Association between a CYP3A4 genetic variant and clinical presentation in African-American prostate cancer patients Cancer Epidemiol Biomarkers Prev 1999 8 901 905 10548319
Tayeb MT Clark C Sharp L Haites NE Rooney PH Murray GI Payne SN McLeod HL CYP3A4 promoter variant is associated with prostate cancer risk in men with benign prostate hyperplasia Oncol Rep 2002 9 653 655 11956645
Felix CA Walker AH Lange BJ Williams TM Winick NJ Cheung NK Lovett BD Nowell PC Blair IA Rebbeck TR Association of CYP3A4 genotype with treatment-related leukemia Proc Natl Acad Sci U S A 1998 95 13176 13181 9789061 10.1073/pnas.95.22.13176
Sata F Sapone A Elizondo G Stocker P Miller VP Zheng W Raunio H Crespi CL Gonzalez FJ CYP3A4 allelic variants with amino acid substitutions in exons 7 and 12: evidence for an allelic variant with altered catalytic activity Clin Pharmacol Ther 2000 67 48 56 10668853 10.1067/mcp.2000.104391
Eiselt R Domanski TL Zibat A Mueller R Presecan-Siedel E Hustert E Zanger UM Brockmoller J Klenk HP Meyer UA Khan KK He YA Halpert JR Wojnowski L Identification and functional characterization of eight CYP3A4 protein variants Pharmacogenetics 2001 11 447 458 11470997 10.1097/00008571-200107000-00008
Kuehl P Zhang J Lin Y Lamba J Assem M Schuetz J Watkins PB Daly A Wrighton SA Hall SD Maurel P Relling M Brimer C Yasuda K Venkataramanan R Strom S Thummel K Boguski MS Schuetz E Sequence diversity in CYP3A promoters and characterization of the genetic basis of polymorphic CYP3A5 expression Nat Genet 2001 27 383 391 11279519 10.1038/86882
Thervet E Anglicheau D King B Schlageter MH Cassinat B Beaune P Legendre C Daly AK Impact of cytochrome p450 3A5 genetic polymorphism on tacrolimus doses and concentration-to-dose ratio in renal transplant recipients Transplantation 2003 76 1233 1235 14578760 10.1097/01.TP.0000090753.99170.89
Zheng H Webber S Zeevi A Schuetz E Zhang J Bowman P Boyle G Law Y Miller S Lamba J Burckart GJ Tacrolimus dosing in pediatric heart transplant patients is related to CYP3A5 and MDR1 gene polymorphisms Am J Transplant 2003 3 477 483 12694072 10.1034/j.1600-6143.2003.00077.x
Floyd MD Gervasini G Masica AL Mayo G George ALJ Bhat K Kim RB Wilkinson GR Genotype-phenotype associations for common CYP3A4 and CYP3A5 variants in the basal and induced metabolism of midazolam in European- and African-American men and women Pharmacogenetics 2003 13 595 606 14515058 10.1097/00008571-200310000-00003
McLeod HL Syvanen AC Githang'a J Indalo A Ismail D Dewar K Ulmanen I Sludden J Ethnic differences in catechol O-methyltransferase pharmacogenetics: frequency of the codon 108/158 low activity allele is lower in Kenyan than Caucasian or South-west Asian individuals Pharmacogenetics 1998 8 195 199 9682265
Ameyaw MM Collie-Duguid ES Powrie RH Ofori-Adjei D McLeod HL Thiopurine methyltransferase alleles in British and Ghanaian populations Hum Mol Genet 1999 8 367 370 9931345 10.1093/hmg/8.2.367
Pyrosequencing primer design
Rose CM Marsh S Ameyaw MM McLeod HL Pharmacogenetic analysis of clinically relevant genetic polymorphisms Methods Mol Med 2003 85 225 237 12710211
Schaid DJ McDonnell SK Wang L Cunningham JM Thibodeau SN Caution on pedigree haplotype inference with software that assumes linkage equilibrium Am J Hum Genet 2002 71 992 995 12387273 10.1086/342666
Polymorphism and Haplotype Analysis Suite
Ke X Cardon LR Efficient selective screening of haplotype tag SNPs Bioinformatics 2003 19 287 288 12538253 10.1093/bioinformatics/19.2.287
SNPtagger
Lamba JK Lin YS Schuetz EG Thummel KE Genetic contribution to variable human CYP3A-mediated metabolism Adv Drug Deliv Rev 2002 54 1271 1294 12406645 10.1016/S0169-409X(02)00066-2
Hustert E Haberl M Burk O Wolbold R He YQ Klein K Nuessler AC Neuhaus P Klattig J Eiselt R Koch I Zibat A Brockmoller J Halpert JR Zanger UM Wojnowski L The genetic determinants of the CYP3A5 polymorphism Pharmacogenetics 2001 11 773 779 11740341 10.1097/00008571-200112000-00005
Kadlubar FF Berkowitz GS Delongchamp RR Wang C Green BL Tang G Lamba J Schuetz E Wolff MS The CYP3A4*1B variant is related to the onset of puberty, a known risk factor for the development of breast cancer Cancer Epidemiol Biomarkers Prev 2003 12 327 331 12692107
| 15882469 | PMC1142317 | CC BY | 2021-01-04 16:03:33 | no | BMC Med Genet. 2005 May 9; 6:19 | utf-8 | BMC Med Genet | 2,005 | 10.1186/1471-2350-6-19 | oa_comm |
==== Front
BMC Med EducBMC Medical Education1472-6920BioMed Central London 1472-6920-5-131588246410.1186/1472-6920-5-13Research ArticleClinical undergraduate training and assessment in primary health care: Experiences gained from Crete, Greece Belos George [email protected] Christos [email protected] Michael [email protected] John [email protected] Anastas [email protected] Koropi Health Centre, Athens, Greece2 Health Planning Unit, School of Medicine, University of Crete, Heraklion, Greece3 Clinic of Social and Family Medicine, School of Medicine, University of Crete, Heraklion, Greece4 Laboratory of Biostatistics, School of Medicine, University of Crete, Heraklion, Greece2005 9 5 2005 5 13 13 28 1 2005 9 5 2005 Copyright © 2005 Belos et al; licensee BioMed Central Ltd.2005Belos 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
Primary Health Care (PHC) is increasingly being introduced into undergraduate medical education. In Greece, the Faculty of Medicine of the University of Crete was the first to introduce a 4-week long training in primary health care. This paper presents the experiences gained from the initial implementation of the teaching of practice-based primary care in rural Crete and reports on the assessment scale that was developed.
Methods
284 students' case write-ups from the 6 primary care units (PCUs) where they were allocated for the period 1990 to 1994 were analysed. The demographic data of the students and patients and the number of home visits were studied. Content analysis of the students' write-ups was carried out, using an assessment scale consisting of 10 dichotomous variables, in order to quantify eight (8) primary qualitative criteria.
Results
Internal reliability was estimated by the index KR20 = 0.67. Face and content validity was found to conform to the standards set for the course, while logistic linear regression analysis showed that the quality criteria could be used as an assessment scale.
The number of home visits carried out varied between the various different PCUs (p < 0.001) and more were reported in the write-ups that fulfilled criteria related to the biopsychosocial approach (p < 0.05). Nine quantitative criteria were fulfilled in more than 90% of case reports, but laboratory investigations were reported only in 69.0% of case reports. Statistically significant differences between the PCUs were observed in the fulfilment of criteria related to the community approach, patient assessment and information related to the patient's perception of the illness, but not to those related to aspects of clinical patient management. Differences in reporting laboratory investigations (p < 0.001) are explained by the lack of such facilities in some PCUs. Demographic characteristics of the patients or the students' do not affect the criteria.
Conclusion
The primary health care course achieved the objectives of introducing students to comprehensive, community oriented care, although there was variation between the PCUs. The assessment scale that was developed to analyse the case-write ups of the students provided data that can be used to evaluate the course.
==== Body
Background
Over the past few decades, the necessity for community orientation of medical undergraduate training and for improving its integration with the health care system have been recognised by international and national bodies [1-3]. As a result, the majority of medical schools in the USA and Europe have embarked upon curricular initiatives to enhance practice-based primary care [4]. Although there is a variance in medical curricula across Europe, there is clear trend towards an increasing focus on primary health care (PHC), favouring a more generalist approach and setting educational objectives related to providing comprehensive care to ambulatory patients, taking account of the family circumstances of the patients, providing home care and integrating with the community.
The Faculty of Medicine of the University of Crete that opened in 1984 was the first medical school in Greece to include, since it was inaugurated, a four-week course in primary health care in the final year of the curriculum [5,6]. Students were allocated to one of 6 primary care units (PCUs) that collaborated with the Department of Social Medicine for this purpose. At that time, PHC was still in its early stages of implementation in the rural areas of Greece, with limited experience of a community-based approach and often lacking in facilities and staff. The collaboration has since developed into a network of PCUs with the medical faculty [7]. It was decided to carry out a retrospective study of the specific characteristics of this training during the earliest four years of the implementation of this course. This initial period was chosen because the essential features of PHC and of the biopsychosocial approach were still not incorporated into everyday practice, while the experiences gained during that phase influenced the content of the training in subsequent years. Drawing on an approach that was used in McGill University [8], an assessment scale was developed that measured qualitatively and quantitatively the content of the case write-ups of the students, for the purpose of evaluating the degree to which the training achieved the objectives of comprehensive PHC, based on the principles of the biopsychosocial model. The aim of this paper is to present the experiences gained from the initial implementation of the teaching of practice-based primary care in rural Crete and to report on the results of using the assessment scale that was developed for evaluating the course.
Methods
Setting
The six PCUs where the field work training in PHC took place: 5 on the Island of Crete and one on Santorini Island, from 1990 to 1994, starting when the first group of students reached the sixth year of their studies.
Fieldwork training and students' assessment
The course syllabus included following the PCU's daily schedule of work under the supervision of the medical staff of the PCU, who acted as clinical tutors. Students were involved in the management of all ambulatory patients that attend, including acute and emergency cases, follow-up visits, preventive care activities, home visits and community based projects. They maintained a logbook of their daily tasks. The students were thus exposed to the knowledge, skills and attitudes required in PHC, including health promotion activities, addressing the living conditions of the population and becoming familiar with the effect of social circumstances on the health of the individual, the family and the community [5,6].
One of the tasks that the students had to carry out was to follow patients with a health condition that affected their social and psychological life. These patients were allocated to the students by the clinical tutors and the students visited the patients at home. Students were invited to choose one of these patients to present as a case study. This method gave them the opportunity to present cases that stimulated their personal interest, a practice in line with the principles of case-based learning and clinical competence based on medical records [9]. Students filled in a standardised case write-up that was prepared by the academic staff of the Department of Social Medicine and originated from the biopsychosocial model [5,6]. Irrespective of the primary care unit where the course took place, students chose cases of the same severity level. These cases were actually representative of the type and kind of cases served in the primary care settings in Greece and they correspond to the common diagnoses made in PHC in rural Greece [10].
The students' assessment was carried out by rating their performance during the course and the quality of their record keeping, and was performed on location by their clinical tutors. Also, the academic staff who led the course performed an oral final exam. A standardised, Visual Analogue Scale (VAS) – based instrument was used additionally for this purpose [6].
Data Analysis
Ex post facto data from the case write-ups were analysed and reviewed. Two hundred eighty four students' standard case report write-ups (284) were collected. Three of these were excluded due to missing data. Most were of these write-ups were prepared by individual students, and 50 (17.8%) were prepared by student groups.
The criteria of quality assessment of the medical records were defined after content analysis of the requirements set by the Department of Social Medicine [5,6] and using standards set in the international literature [12-14]. In particular, 8 primary qualitative criteria were selected that addressed specific aspects of the training: (a) The community approach or relevance to PCU/ PHC, (b) Family record / structure or family pedigree, (c) Patient assessment / differential diagnosis or priority order, (d) Management strategy or plan for primary care, (e) Utility of PHC services / PCUs, (f) "Who" or person (information about the subject), (g) "What" or condition (information about the disease-illness-sickness of the subject), (h) "Where" or environment (information about the environment of the subject).
In order to quantify the quality of the records, ten variables, shown in table 7, were constructed, by slightly changing the order of the criteria and with the addition of the therapeutic approach and the availability of laboratory data. These variables were operationalised using a dichotomous approach, being assigned a value of 0 if absent and 1 if present and so made amenable to numerical analysis [15-17].
Table 7 Variables used in quantifying quality of records
Variable 1 Therapeutic approach of the issue patient, in the context of the given PCU
Variable 2 Laboratory data, in relation to the laboratory facilities of the given PCU
Variable 3 Community Approach: interrelation between patient's disease and primary care services
Variable 4 Family record: complete genogram or family pedigree, as well as record of existing dynamics within the family
Variable 5 Patient assessment: Prioritising of diagnostic problems and differential diagnosis (organisation of data), as well as the proposed steps and further measures needed for disease management within the primary care services
Variable 6 Management Strategy or Plan: Services suggested by the medical personnel of the PCU for the best possible management of the patient within the primary care services
Variable 7 Utility or Usefulness of primary care: Information related to the parts of the management plan that were actually implemented at the PCU and to the way this implementation offered positive or negative feedback to the PCU
9 and 10 All criteria covering the biopsychosocial point of view, as defined by Prof. Howard F. Stein [14], and in particular:
Variable 8 Information on "Who", i.e. about the subject-person (the way the patient perceives his/her disease or medical condition, and the effects on the patient's relationship with the other members of the family and the community) [3]
Variable 9 Information on "What", i.e. about the disease – illness – sickness perception of the subject, and the effects it has on the biological, functional and social level respectively
Variable 10 Information on "Where", i.e. data on the environment of the patient, and on how this environment affected the patient's medical condition or health problem [24-26]
Internal reliability and validity were assessed for this instrument, using 231 write-ups filled by individual students. They were considered satisfactory by split-half method of Kuder-Richardson-index, where KR20 = 0.67, on the ten quality criteria in table 7, since the concrete case write-ups were completed once during the course. Face and content validity were found to conform to the requirements set by the Department of Social Medicine. Structure validity and prognostic validity were proved using the linear regression through the origin method. Logistic linear regression analysis through the origin assessed the magnitude of the net effect of each independent variable on the sum-score (dependant variable: resulted as the sum of independent variables) [18,19].
Data from these case reports were classified according to content meaning and to the results of the survey, and were founded on the principles of content analysis [8]. From each case report, the following variables and demographic data were encoded: record date, PCU (1 to 6) where the course took place, gender of the participant, gender and age of the patient, type of the case (urgent, prescription visit, chronic disease), co-existing disease, number of student visits at the patient's home after first contact with the PCU for treatment and follow up (0 to 3 visits). The assessment was performed by two independent qualified generalists / reviewers in Koropi Health Centre in the Athens area. Their agreement was estimated, based on the following formula [20,21]:
, where:
x = the number of quality criteria common between the two (2) reviewers
ψ = the number of criteria used by reviewer A' (10, in total)
ω = the number of criteria used by reviewer B' (10, in total)
In this study, this equation formula is represented by a level of agreement β = 1.
Demographic data from the records were analysed with the usage of descriptive statistics. The level of statistic significance for quality parameters has been controlled with the x2 test, and wherever it was appropriate, Yates correction or Fisher's exact correction was applied [[20,22], and [23]].
Results
Descriptive data
42.3% of case reports was processed by male students, 39.8% was processed by female students, while 17.7% was processed by student groups (Table 1). Patients' demographics are presented in table 2. Patients over 66 years old dominated the picture, representing 44.8% of male patients and 50.0% of female.
Table 1 Distribution of students per gender and PCU
Primary Care Unit Male Students N (%) Female Students N (%) Teamwork N (%) Total
PCU 1 19 (33.3) 23 (40.3) 15 (26.3) 57
PCU 2 17 (43.5) 13 (33.3) 9 (23.0) 39
PCU 3 32 (47.7) 26 (38.8) 9 (13.4) 67
PCU 4 20 (47.6) 14 (33.3) 8 (19.0) 42
PCU 5 28 (43.7) 30 (46.8) 6 (9.3) 64
PCU 6 3 (25.0) 6 (50.0) 3 (25.0) 12
Total 119 (42.3) 112 (39.8) 50 (17.7) 281
Table 2 Distribution of patients per gender and age group
Age group Male N (%) Female N (%) Total N (%)
0–15 years 10 (8.0) 6 (3.8) 16 (5.7)
16–45 years 19 (15.2) 26 (16.7) 45 (16,0)
46–65 years 40 (32.0) 46 (29.5) 86 (30.6)
66+ years 56 (44.8) 78 (50.0) 134 (47.7)
Total 125 156 281
As for the number of home visits for health care and follow-up, 64.8% of the students performed a single visit and 31.0% performed two visits, while 2.8% of the students did not make any home visits at all (Table 3). The difference in the number of home visits between the PCUs reached statistical significance (p < 0.001). Whenever the write-ups fulfilled both the criteria of the family history and the biopsychosocial approach (variables 8, 9 and 10 in table 7), the numbers of home visits was larger, and vice versa. This difference reached statistical significance (p < 0. 05) (Table 4).
Table 3 Students' home visits per PCU
Primary care unit Number of home visits
0 N (%) 1 N (%) 2 N (%) 3 N (%)
PCU 1 -(-) 48 (48.2) 9 (15.8) - (-)
PCU 2 2 (5.1) 34 (87.2) 3 (7.7) -(-)
PCU 3 3 (4.5) 37 (55.2) 25 (37.3) 2 (3.0)
PCU 4 2 (4.8) 19(45.2) 21 (50.0) - (-)
PCU 5 - (-) 35 (54.7) 27 (42.2) 2 (3.1)
PCU 6 1 (8.3) 9 (75.0) 2 (16.7) - (-)
Total 8(2.8) 182(64.8) 87(31.0) 4 (1.4)
Table 4 Co-existence of family structure and total biopsychosocial approach (N %) by average of home visits (HV).
Primary care unit Male N (%) / HV Female N (%) / HV Team N (%) / HV
PCU 1 95/1.158 91/1.087 100/1.267
PCU 2 59/1.00 85/1.077 89/1.00
PCU 3 78/1.375 96/1.308 89/1.667
PCU 4 79/1.250 77/1.571 80/1.750
PCU 5 67/0.667 100/1.167 100/1.333
PCU 6 95/1.357 86/1.581 100/1.600
Data on assessment
The complete data on the distribution of the assessment variables of the case write-ups by PCUs are presented in table 5. More than 95% of the case write-ups fulfilled criteria "1" (therapeutic approach), "7" (usefulness of primary care) and "10" ("Where" issues), while more than nine out of ten fulfilled criteria "4" (family history), "6" (management plan) and "9" ("What" issues). In all these cases, differences between the various PCUs were not significant or showed only weak statistical significance. More than 90% of the write-ups also reported on criteria "3" (community approach), "5" (patient assessment and differential diagnosis) and "8" ("Who" issues: the patient's perception of the illness), and in these cases the differences between PCUs were statistically significant. Only 69.0% of the write-ups fulfilled criterion "2" (appropriate laboratory investigation), the difference between PCUs being statistically significant (p < 0.001).
Table 5 Distribution of criteria for the assessment of the case write-ups per PCU
Primary care unit PCU 1 N (%) PCU 2 N (%) PCU 3 N (%) PCU 4 N (%) PCU 5 N (%) PCU 6 N (%) TOTAL
1. Therapeutic Approach 98.2 92.3 98.5 95.2 98.4 100.0 97.2
2. Laboratory data (1) 66.7 38.5 62.7 100.0 70.3 100.0 69.0
3. Community Approach (2) 98.2 82.1 92.5 95.2 98.4 100.0 94.3
4. Family Record 98.2 84.6 91.0 92.9 93.8 100.0 92.9
5. Patient Assessment (3) 98.2 82.1 89.6 100.0 96.9 100.0 94.0
6. Management Strategy 94.7 92.3 88.1 90.5 98.4 100.0 93.2
7. Usefulness of primary care (4) 98.2 89.7 95.5 90.5 100.0 100.0 95.7
8. Who issues: Person (5) 100.0 97.4 93.9 100.0 82.8 100.0 94.3
9. What issues: disease – illness – sickness 96.5 92.3 95.5 95.2 85.9 91.7 92.9
10. Where issues: Environment (6) 100.0 100.0 95.5 97.6 90.6 100.0 96.4
(1) p < 0.001
(2) p < 0.01
(3) p < 0.01
(4) p = 0.06
(5) p < 0.001
(6) p = 0.06
The distribution of the criteria/variables of the case reports' assessment per gender and age group (Table 6) reveals that the individual demographic characteristics of the patients (age and gender) do not affect the quality of the training. Neither do the characteristics of the students (gender, individual or team work).
Table 6 Distribution of the criteria of the assessment of the case write-ups per gender and age group
Age 0–15 16–45 46–65 66+ TOTAL
Gender Male N (%) Female N (%) Male N (%) Female N (%) Male N (%) Female N (%) Male N (%) Female N (%) Male N (%) Female N (%)
1. Therapeutic Approach 100.0 100.0 94.7 96.2 97.5 100.0 98.2 94.9 97.6 96.8
2. Laboratory data 70.0 83.8 73.7 65.4 55.0 84.5 69.6 65.4 65.6 71.8
3. Community Approach (1) 100.0 100.0 94.7 100.0 80.0 97.8 94.6 96.2 90.4 97.4
4. Family Record 90.0 100.0 89.5 92.3 87.5 95.7 91.1 94.9 90.4 94.9
5. Patient Assessment 90.0 100.0 89.5 96.2 90.0 95.7 96.4 93.6 92.8 94.9
6. Management Strategy 100.0 100.0 94.7 100.0 87.5 95.7 94.6 91.0 92.0 94.2
7. Usefulness of primary care 100.0 100.0 94.7 92.3 87.5 95.7 96.4 100.0 93.6 97.4
8. Who issues: Person 100.0 83.3 94.7 100.0 85.0 95.7 94.6 96.1 92.0 96.1
9. What issues: disease-illness-sickness 100.0 100.0 100.0 96.2 85.0 95.7 91.1 92.3 91.2 94.2
10. Where issues: Environment 100.0 100.0 94.9 100.0 97.5 97.8 94.6 94.9 96.0 96.8
(1) p = 0.06
After performing the linear logistic regression through the origin analysis (the no-intercept model), data relative to the estimation of predictive validity suggest all independent variables as predictors on the dependent one (overall sum-score) since all the above criteria-variables make a significant contribution to the cumulative score. According to these findings, the descending order of the independent variables-criteria is: 8 ("Who" issues), 4 (family record), 2 (laboratory data), 3 (community approach), 7 (usefulness of primary care), 1 (therapeutic approach), 6 (management strategy), and 5 (patient assessment) (Unstandardised Coefficients: R Square = 1.000, B = 1.000, Std. Error = .000, Standardised Coefficients: Beta = 0.126–0.129, level of significance p < 0.001).
Discussion
This study was the first one of its kind ever to be performed in Greece. Although it is hampered by the lack of a control group, it is methodologically sound and provides useful information on the assessment of the whole training process. Teaching practice-based primary care requires medical students to understand the keys components of primary care, physicians to mobilise them and assessment tools for evaluating the undergraduate training.
Our study reports the first data available from the assessment of case write-ups. Case write-ups as an assessment method have an inherent weakness that could account for the fact that, despite its wide application, there are few studies that evaluate it as an assessment method. This weakness is that content analysis, method of choice for the assessment of the case write-ups, is inevitably subjective. To avoid this drawback, we carried out content analysis of our material by using a dichotomous (yes/no) approach of the qualitative variables. The level of agreement between the two reviewers about these quality criteria was found to be quite high (β = 1) which means full homogeneity and resemblance in the encoding between the two reviewers, underscoring the coexistence of objectiveness and content validity. Therefore, our results can be viewed only in the context of this methodological limitation. Further studies need to address the issue of concurrent validity of such an approach, complementing it with other methodological approaches that assess the educational process in a more detailed way, such as questionnaires with visual analogue scales or 5-point Likert scale and interviews with open-ended questions.
It is worth to note on the differences that were observed between the various PCUs. Thus, the number of home visits differed significantly among the PCUs (p < 0.001). This may be because time of our study some PCUs had developed home visits more than others. Further, the number of home visits proved to be an important factor for reporting on disease progress, as well as a means to evaluate factors related to the family and the social environment of the patient [12].
Regarding the criteria that were used to assess the case write-ups, it is interesting that those criteria that reflect the more clinical aspects of patient care were reported from almost all the PCUs, without significant differences. It is reasonable to assume that the clinical tutors in all the PCUs show the same interest in the treatment approach, the family history and the management plan, while they would encourage the students to report on the usefulness of PHC.
Reference to the patient assessment (the prioritisation of the elements referring to differential diagnosis) and to the community approach and its relation with PCU, was variably represented among the PCUs, reaching statistical significance (p < 0.01). The biopsychosocial approach ("who" issues) also received varying emphasis between the different PCUs (p < 0.001), although the overall rate was high. These differences may be attributed to a combination of factors: the actual contribution of the unit to the patient's management, the specific priorities adopted by each unit, the degree of familiarity of the clinical tutors with these aspects and the lack of interest in these parameters by the students.
Information obtained from the results of laboratory investigations was also found to differ significantly among the PCUs (p < 0.001). This can be explained by the fact that some of the units did not, at the time of the study, operate microbiology/biochemistry or radiology laboratories, and therefore it was not possible to carry out these investigations.
Since logistic linear regression through the origin analysis indicated that all of the proposed quality criteria function as predictors of the total assessment (in the aforementioned order), our findings are suggestive of a new total (trainee, trainer and the training site and program) assessment scale. This remains to be further investigated in the future as a useful methodological tool in ex post facto studies that are based on standardised medical records [24,25].
This study shows that the students' training achieved to a large extent the objectives that were set. However, achieving the objectives was affected by the orientation and the philosophy of the PCUs' personnel, and the attitude of the clinical tutors. On the other hand, individual demographic characteristics of the patients (age and gender), and the students (gender, individual or team work), as well as the disease / functional state of the patients, do not seem to affect the training.
Our study has several implications in undergraduate training in primary care. It is apparent that with the application of appropriate training courses, medical education can gradually shift from a model of illness and cure to a model of wellness and care [26,27]. However, emphasis should be given to support and/or select the PCUs that undertake to train students. The characteristics required to be fulfilled relate to the availability of basic medico-technological equipment as well as to the orientation towards primary care and the community. The fact that health care delivery in Greece is still highly fragmented and discontinuous makes it important to emphasise the holistic PHC approach and the biopsychosocial perspective. According to many authors, students are apt to immediately discern such characteristics [[9,12], and [13]].
At the same time, the assessment scale that we propose in this study proved to be a useful instrument in evaluating the entire training process (trainee, trainer, training site). Therefore, before participating in any training process, clinical tutors in PCUs must receive specific guidance on the quality criteria (table 7) used in the assessment of the students' case reports.
Almost ten years after this study was carried out, the Faculty of Medicine of the University of Crete remains to be the only one among seven medical schools in Greece to include primary care in its undergraduate curriculum. However, we know from our discussions with colleagues in other medical schools that they have recently started exploring the idea of introducing an elective course in PHC. In the meanwhile, the collaboration between the Department of Social Medicine and the Primary Care Units in Crete, which started as a means to train students in PHC, has developed into a network for health needs assessment [9] as well as for organisational and institutional development of PHC [28].
Conclusion
This study demonstrates that the course in primary health care in the curriculum of the Medical Faculty of the University of Crete achieved to a large extent the objectives of introducing the students to a more biopsychosocial approach, although the degree of success varied between the PCUs where students were allocated. It also shows that the objectives of the training can be evaluated by using an analysis of the case write-ups that students prepare and present. This study also prompts us to re-examine the training that is taking place today, so as to ascertain whether the gradual establishment of a climate that favours PHC has contributed to improvements in the content of the course and to the fulfilment of the objectives that were defined during the early period of the course. Research in primary medical care education may benefit from the development of a consensus on assessment scales but further discussion and innovative methodology are required.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GB collected the data, carried out the analysis and wrote the first draft of the manuscript. CL corrected the first daft, contributed to the analysis and interpretation of data, formed the layout of the manuscript and contributed to the manuscript's re-drafting. MF co-designed the contents of the training, initiated the collaboration with the PCUs and advised on the study design. The late JV carried out the statistical analysis and wrote the relevant sections. AP conceived the study design, co-designed the contents of the training and rewrote the manuscript. All the authors approved the final version of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Garcia-Barbero M Medical education in the light of the W.H.O., Health for All Strategy and the European Union Med Educ 1995 29 3 12 7623682
World Federation for Medical Education- A World Conference The changing medical profession Med Educ 1993 27 291 296
General Medical Council Tomorrow's doctors Recommendations on undergraduate medical education 2002 London GMC
Kurth RJ Irigoyen MM Schmidt HJ Structuring student learning in the primary care setting where is the evidence? J Eval Clin Pract 2001 7 325 333 11555090 10.1046/j.1365-2753.2001.00288.x
Philalithis A Pre-graduate medical education in PHC at the Department of Social/Family Medicine, University of Crete Proceedings of the national conference on PHC, Ioannina 1991 (in Greek)
Philalithis A Fioretos M The Clinical training in PHC: annual guide for the training 1991 Social/Family Medicine Department, University of Crete, Heraklion; Crete (in Greek)
Lionis C Trell E Health needs assessment in general practice: The Cretan approach Eur J Gen Pract 1999 5 75 77
McLeod PJ Faculty assessment of case reports of medical students J Med Educ 1987 62 673 678 3612729
Bullimore WD Study skills and tomorrow's doctors 1998 W.B. Saunders
Koutis AD Isacsson A Lindholm LH Lionis CD Svenninger K Fioretos M Use of Primary Health Care in Spili, Crete and in Dalby, Sweden Scand J Prim Health Care 1991 9 297 302 1792458
Maguire P Communication skills for doctors 2000 London Arnold
Billings JA Stoeckle JD The Clinical encounter, a guide to the medical interview and case presentation 1999 St. Louis Mosby
Deutsch SL Community-based teaching: A guide to developing education programs for medical students and residents in the practitioner's office 1997 Philadelphia, American College of Physicians
Crouch MA Roberts L Stein H The family in medical practice, a family systems primer, and Chapter 2: A systems view of the clinical relationship 1987 Heidelberg Springer-Verlag
Walton HJ WFME Primary health care in European medical education: A survey Med Educ 1985 19 167 188 4010563
Foldevi M Svedin CG "Phase examination" an assessment of consultation skills and integrative knowledge based in general practice Med Educ 1996 30 326 332 8949470
Horwitz RI Yu EC Assessing the reliability of epidemiological data obtained from medical records J Chronic Dis 1984 37 825 831 6542107 10.1016/0021-9681(84)90015-8
Josefowitz N Moss J Pike B Fainstat P Reliability of faculty assessments of student case histories: a problem in chiropractic education J Manipulative Physiol Ther 1983 6 33 35 6854157
Green SB Salkind NJ Using SPSS for Windows, analysing and understanding data 2000 2 New Jersey Prentice Hall
Abramson JA Abramson ZH Survey methods in community medicine, epidemiological research-programme evaluation-clinical trials 1999 5 New York, Churchil Livingstone
Ghiclione R Manuel d'analyse de contenu, Paris Edition Colin A 1980
Mainland D Elementary medical statistics 1999 2 Philadelphia W.B. Saunders
Wood M Mayo F Marsland D Practice-based recording as an epidemiological tool Annu Rev Public Health 1986 7 357 363 3521649 10.1146/annurev.pu.07.050186.002041
Weitzman S Bar-Ziv G Pilpel D Sachs E Naggan L Validation study on medical recording practices in primary care clinics Isr J Med Sci 1981 17 213 7228647
Jansen JJ Tan LH van der Vleuten CP van Luijk SJ Rethans JJ Grol RP Assessment of competence in technical clinical skills of general practitioners Med Educ 1995 29 247 253 7623721
The Society of Academic Primary Care, New Century-New Challenges A report from the heads of departments of general practice and primary care in the medical schools of the UK 2002 Liverpool: The Society of Academic Primary Care
Morrison J Watt G New century, new challenges for community based medical education Med Educ 2003 37 2 3 12535108 10.1046/j.1365-2923.2003.01428.x
Lionis C Tsitaki M Bardis V Philalithis A Seeking quality improvement in primary care in Crete: the first actions Croat Med J 2004 45 599 603 15495288
| 15882464 | PMC1142318 | CC BY | 2021-01-04 16:30:57 | no | BMC Med Educ. 2005 May 9; 5:13 | utf-8 | BMC Med Educ | 2,005 | 10.1186/1472-6920-5-13 | oa_comm |
==== Front
BMC Med EducBMC Medical Education1472-6920BioMed Central London 1472-6920-5-161588513710.1186/1472-6920-5-16DebateEvolutionary explanations in medical and health profession courses: are you answering your students' "why" questions? Harris Eugene E [email protected] Avelin A [email protected] Department of Biological Sciences and Geology, Queensborough Community College, City University of New York, New York City, USA2 Department of Cell Biology, New York University School of Medicine, New York City, USA2005 10 5 2005 5 16 16 22 9 2004 10 5 2005 Copyright © 2005 Harris and Malyango; licensee BioMed Central Ltd.2005Harris and Malyango; 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
Medical and pre-professional health students ask questions about human health that can be answered in two ways, by giving proximate and evolutionary explanations. Proximate explanations, most common in textbooks and classes, describe the immediate scientifically known biological mechanisms of anatomical characteristics or physiological processes. These explanations are necessary but insufficient. They can be complemented with evolutionary explanations that describe the evolutionary processes and principles that have resulted in human biology we study today. The main goal of the science of Darwinian Medicine is to investigate human disease, disorders, and medical complications from an evolutionary perspective.
Discussion
This paper contrasts the differences between these two types of explanations by describing principles of natural selection that underlie medical questions. Thus, why is human birth complicated? Why does sickle cell anemia exist? Why do we show symptoms like fever, diarrhea, and coughing when we have infection? Why do we suffer from ubiquitous age-related diseases like arteriosclerosis, Alzheimer's and others? Why are chronic diseases like type II diabetes and obesity so prevalent in modern society? Why hasn't natural selection eliminated the genes that cause common genetic diseases like hemochromatosis, cystic fibrosis, Tay sachs, PKU and others?
Summary
In giving students evolutionary explanations professors should underscore principles of natural selection, since these can be generalized for the analysis of many medical questions. From a research perspective, natural selection seems central to leading hypotheses of obesity and type II diabetes and might very well explain the occurrence of certain common genetic diseases like cystic fibrosis, hemochromatosis, Tay sachs, Fragile X syndrome, G6PD and others because of their compensating advantages. Furthermore, armed with evolutionary explanations, health care professionals can bring practical benefits to patients by treating their symptoms of infection more specifically and judiciously. They might also help curtail the evolutionary arms race between pathogens and antibiotic defenses.
==== Body
Background
The science of Darwinian or Evolutionary Medicine was formalized in the early 1990s most notably by the eminent evolutionary biologist George C. Williams and the physician and professor Randolph Nesse [1,2]. Its methodologies and aims are directed at the scientific investigation of evolutionary causes for human disease, disorders, malfunctions and apparent design failures in order to improve health care and to stimulate fruitful research directions.
To clarify what we mean by evolutionary causes it is useful to make a distinction from a different level of causality in biology, that of proximate causes. Proximate causes are at the level of the immediate mechanisms that give rise to disease, disorders, and malfunctions. These include detailed accounts of physiological processes, embryological development, and anatomical relationships and functions.
Evolutionary explanations, in contrast, are obtained via researchers taking an historical perspective (see Mayr [3]). They look back in time, through lenses of scientifically established evolutionary principles, in order to investigate medical problems and ailments. An evolutionary approach would ask such questions as: "Why hasn't natural selection eliminated a particular disease?" "Why do our bodies have certain flaws of engineering that make us susceptible to particular medical complications?" "Might certain modern diseases and ailments be the result of a mismatch between our biological heritage and our modern-day lifestyles?" and "Are characteristic physiological responses to disease in fact adaptive defenses that have developed over evolutionary time?"
Neither approach, proximate or evolutionary, can be ignored. In fact, such explanations should in theory feedback positively to stimulate novel research approaches [2]. Nonetheless, we realize that evolutionary explanations may seem foreign to many persons in the medical field. After all, most of modern day medicine is based on scientifically discovered proximate explanations, and these explanations dominate in medical textbooks. Thus, to illustrate the evolutionary perspective, in this paper we give a variety of examples for which evolutionary explanations are known or hypothesized. We also describe the theoretical evolutionary principles that underpin these examples and that can be generalized to the investigation of many medical questions.
Complications of human child-birth: proximate causes
Many of our students are perplexed when they hear about the pain accompanying child delivery. They wonder "Why is there so much pain?" and "Why is the birthing process such a complication, so much so that many doctors deliver by Caesarean section?" The proximate answer describes that the diameter of the human fetus's head is very large (the widest part of the fetus) and the diameter of pelvic outlet of the mother is only slightly bigger. This makes it a very tight passage for the fetus, a very painful process for the mother, and a complicated process for the obstetrician to manage. The release of the hormone relaxin during pregnancy (a proximate mechanism) helps to loosen the pubic symphysis and dilate the uterine cervix easing the complication to a degree. It is easier all around, even if not always necessary, if the fetus can detour the natural route and be delivered through an alternate passageway created in the mother's lower abdomen.
Being accustomed to such explanations, many students appear satisfied, though this is likely a cover up. Curious students naturally want further explanation, "But why is a human infant's head so big?" "Why isn't the mother's birth canal larger? And finally "Why is the human birthing process so complicated? These questions have evolutionary explanations – explanations that may only rarely emerge in medical and health profession's classes.
Complications of human child-birth: evolutionary causes
Following is a classic evolutionary explanation backed by considerable paleoanthropological evidence. Beginning about 2.5 million years ago, the evolutionary lineage leading to modern humans shows a trend towards increasing brain size so that by around 500,000 years ago our ancestors had brains (and crania) about as large as they are today [4]. However, despite the advantages, evolving such large brains posed serious problems for mothers who need to pass their infant's head through a relatively small pelvic canal [4,5].
"But if brain size evolved to its modern size so long ago, why hasn't evolution had enough time to fix the imperfect-fit... that is, why isn't the birth canal larger?" Well, females have indeed evolved a rounder pelvic opening to facilitate the uneasy passage of the newborn's head. This explains the anatomical differences we observe between male and female pelves. There is, however, a limit to the degree to which the female pelvic opening can enlarge. This constraint arises because of another unique human adaptation namely bipedalism – walking on two legs. Exceptional increase of the birth canal would require repositioning of the gluteal pelvic-abductors in such a way as to make bipedalism mechanically and energetically inefficient. Thus, the overall size of the birth canal can be viewed as an evolutionary compromise between two advantageous features of human evolution – increased brain size as well as the new style of locomotion. These two different evolutionary pressures have led to the difficult and imperfect process of human child birth [4,5]. (For further discussion of evolutionary obstetrics see Travathan's book Evolutionary Medicine [6]).
The example of sickle cell disease: a proximate perspective
Take as another example, sickle cell anemia. A student asks, "Why does sickle cell disease exist?" The proximate (or mechanistic) explanation is necessary but not sufficient. A point mutation at the sixth amino acid position of the β hemoglobin chain replaces glutamic acid with valine and produces the so-called abnormal hemoglobin type HbS (the normal type is HbA). When a cell having only the HbS type gives up oxygen to body tissues it loses the normal red blood cell shape (symmetrical biconcave) and becomes crescent or sickle-shaped. These homozygous individuals (HbS/HbS) develop sickle cell disease in which sickle-shaped RBCs block small peripheral blood vessels leading to very serious secondary effects. Without regular medical intervention few persons live to adulthood [7]. Promisingly, with the recent development of modern treatment regimens (e.g. stem cell transplantations, serial blood transfusions, and prophylactic treatments with penicillin) patients show increased survivability, living into their 40s and 50s and in some cases even older [8,9]. Unfortunately, many children with sickle cell disease do not have access to intensive treatment programs, either because it is not locally available or their health insurance plans limit access to it [10].
Many of our students are older students and have worked in hospitals or clinics in the New York City area. Most are aware that sickle cell disease is commonly encountered. On average, 1 in every 600 births in the African-American population is to a baby with sickle cell disease [11]. Therefore, these students will wonder, "If sickle cell disease is so lethal, then why hasn't it been eliminated?"
Sickle cell disease: an evolutionary perspective
An evolutionary explanation is that sickle cell anemia is an example of a disease caused by a gene that can also be beneficial. The HbS type is found at high frequencies (up to 20% and above) in much of tropical Africa. In most other parts of the world the frequency is very low (<1% or absent), excepting for regions such as the Mediterranean, the Arabian Peninsula, and the Americas where a high frequency is due to historic migration and admixture [7]. Importantly, within Africa the distribution of the (HbS) almost exactly matches the distribution of the mosquito-transmitted disease known as malaria, caused by the parasite Plasmodium falciparum. It is known that heterozygotes (HbS/HbA), having both types of hemoglobin (and who have only minor health problems), show resistance to malarial infection because the body targets the P. falciparum infected cells for destruction. In contrast, individuals homozygous for normal hemoglobin (HbA/HbA) suffer high mortality rates in early childhood due to malarial infection. Thus, the allele for sickle cell has been maintained because heterozygotes have a higher reproductive success than either of the two possible homozygotes. The evolutionary process explaining the phenomenon is known as balancing selection, a special form of natural selection that operates to keep two or more beneficial alleles at relatively high frequencies [7].
This classic example is known since the 1950s [12] and the relationships between mosquito ecology, agriculture, malaria, and frequencies of the sickle cell gene have been carefully studied [13]. Notwithstanding our knowledge, clear explanations of the evolutionary processes that underlie sickle-cell disease, balancing selection and heterozygote advantage, are still not found in many medical and pre-professional health texts leaving students needlessly wondering 'Why?' Indeed, understanding these evolutionary processes would help students understand why genes underlying other hemotological disorders like alpha and beta thalassemias, Hemoglobin E (Hb E) syndrome, and Glucose-6-phosphate dehydrogenase (G6PD) deficiency are found at high frequencies in certain populations from the Mediterranean, African and/or South East Asia [14,15]. Like with sickle cell, evidence suggests that natural selection has maintained these disease genes because they too confer resistance to malaria [14,15]. While these diseases were once geographically confined, they are now, due to forced, historic, or recent emigration, seen commonly in US or European hospitals and clinics. For example, Glader and Look [16] report that 41% of the South East Asian refugee population in the United States are carriers of, or have, one of the myriad hematological disorders. While in past decades it may have been rare for our students to come into contact with these diseases; today the likelihood is much greater.
Explaining symptoms
Evolutionary causes can also explain classic symptoms of illness. Traditionally, we teach that symptoms like fever, inflammation, diarrhea, coughing, vomiting are problems that need to be alleviated by medication. The proximate mechanisms of these symptoms are described in most textbooks. Not sufficiently discussed, though, are evolutionary explanations. These explanations derive from the ongoing evolutionary arms race between pathogens and us, each one trying to out-perform the other. Therefore, to what extent are so-called "symptoms" the body's way of defending itself, and to what extent are they manipulations by the pathogen, "intended" to increase its reproduction and spread?
Studies suggest that fever is an evolved defense mechanism against pathogens that if reduced by medicinal drugs can, in some cases, prolong illness and cause secondary infections [1,2,17]. Diarrhea has been hypothesized to be either a defense mechanism or a manipulation by the pathogen, or even both simultaneously [17]. Persons with Shigella bacteria develop severe bloody diarrhea. To test the diarrhea-as-defense hypothesis persons were treated with and without the anti-diarrhea medication Lomotil™. Results showed that illness was prolonged in the treated group as compared with the untreated group [18]. In contrast, Vibrio cholerae, the bacterium that causes cholera, carries a toxin on its cell membrane that induces diarrhea, presumably to help its spread. Infected people are severely affected and at risk for death due to loss of body fluids. In this case, treatment with anti-diarrhea medication may counter the pathogens strategy and aid against dehydration and against spread of the bacteria [17]. Ideally, the physician would want to know when defenses are useful to the patient and when they are manipulations by pathogens [1,17]. If each pathogen-host combination were to be analyzed separately, realizing that the balance of evolutionary forces acting on the host and pathogen may vary in each case, then treatment would likely become more specific and effective.
Discussion
Evolutionary explanations & natural selection
Can our students benefit from such evolutionary explanations? Yes, we believe they can. In fact, this is the central tenet of the science of Evolutionary Medicine: that principles of evolutionary theory can provide a unifying principle by which human biology, disease, and disorders can be understood [1,2]. The explanatory power lies in fully understanding how natural selection operates.
Following we explain several important principles of natural selection that are generally applicable to understanding human biology in health and disease (Table 1, and see [1,2]).
Table 1 Principles of natural selection generally applicable to understanding human biology in health and disease.
Natural selection:
• cannot build perfect designs because it compromises between different adaptations – bipedal walking vs. large brain size.
• will maintain a disease gene if it confers an advantage in a particular environment – sickle cell disease, alpha- & beta-thalassemia, Hb E syndrome, G6PD deficiency, Tay-sachs, cystic fibrosis.
• has shaped many human genes to ancient lifestyles (i.e. a hunter-gathering versus modern life-way) explaining chronic diseases like obesity, and type II diabetes.
• favors genes maximizing reproduction even if they compromise health (PKU, hemochromatosis, fragile X syndrome etc.) or longevity (Alzheimer's, atherosclerosis, prostate hyperplasia)
• explains the "arms race" between pathogens and us. We have evolved defenses both natural – fever, diarrhea, vomiting, inflammation – and manufactured – antibiotics. Pathogens evolve counterstrategies like antibiotic resistance, and manipulation of our defenses for their spread.
• Natural selection cannot build perfect designs because of compromises that exist between different adaptations (e.g. a pelvic girdle designed for bipedal locomotion versus one designed to facilitate passage of a large brained infant). Almost every human structure and physiological process can be analyzed in terms of compromises between costs and benefits. For example, building bones that will resist fracture would require further increasing their thickness and/or calcium content, though doing so would make us slower and take away calcium crucial for normal cellular functions. Indeed, a homeostatic mechanism has evolved that draws calcium from bones when cellular functions need it. However, that same mechanism can sometimes overdraw the mineral producing problems such as fracture, osteomalacia, and rickets.
• Natural selection works within the limitations imposed by the human body's long evolutionary legacy. George C. Williams has rhetorically questioned: why hasn't the birth canal evolved to open through the lower abdomen rather than through the pelvis, thus releasing the severe evolutionary constraint against further increase in brain size? The answer is simply that vertebrates set upon routing the female reproductive tract through the pelvic opening hundreds of millions of years ago and there has been no way for evolutionary processes to readjust the route [19].
• Natural selection may often adapt our genes to the context of a particular environment. Therefore, certain genotypic traits offer benefits only with respect to the environment in which they evolved (e.g. genes causing sickle cell disease and other hemoglobinopathies). When these genotypes are expressed in other environments, like sickle cell is in the United States, there is essentially no benefit because malaria is largely absent. As expected, the frequency of the sickle cell gene is declining among American Blacks [7]. As another example, in 1987 Rotter and Diamond [20] hypothesized that the high frequency of the genes causing Tay sachs (homozygote frequency of 1 in 6000 births) and cystic fibrosis (homozygote frequency of 1 in 1700 births), in individuals descended from Ashkenazi Jews and European populations respectively, results from the survival advantage the genes conferred to heterozygote individuals who endured past epidemics of tuberculosis and/or typhus.
Such evolutionary hypotheses can be tested scientifically, for example, witness studies of cystic fibrosis (CFTR) [21,22]. The gene was recently found to code for chloride ion protein-channels on somatic epithelial cell membranes, and the bacteria causing typhoid fever, Salmonella enterica serovar Typhi, was found to target the CFTR protein in GI cells to gain entry to them. Subsequent tests showed that serovar Typhi invasion rates in mice are reduced by over 80% in heterozygotes for the mutant allele (ΔF508 CFTR) compared to homozygotes (for the normal CFTR allele) due to reduced CFTR expressed on the cell surface. Moreover, when European populations were analyzed that reportedly suffered Typhoid scourges in the past, it was found that they showed higher frequencies for the mutated allele one to two generations after the scourge [22]. Together, the proximate and historical evidence strongly support Typhoid fever as an important selective agent (though perhaps not the only agent) that drove cystic fibrosis frequencies upwards due to heterozygote advantage. In contrast, it is unfortunate that the Tay sachs selective hypothesis remains largely untested almost two decades later.
• As an extension of the previous principle, natural selection has likely optimized most of our genes to the environment and lifestyle by which we have lived the longest – a hunter-gatherer lifestyle – likely explaining the presence of certain chronic diseases (type II diabetes, obesity and others). Our genes and homeostatic mechanisms evolved under vastly different environmental and nutritional conditions than those under which we live today [23-25]. For example, with respect to obesity, much recent research has focused on leptin, a hormone released by adipose tissue that regulates appetite by acting on the hypothalamus. Although further research is required, evidence suggests that under increasingly obese body conditions the brain becomes resistant to leptin throwing the leptin-hypothalamic homeostatic system into an unhealthy positive feedback loop that further stimulates appetite [26,27]. With respect to modern diets and lifestyles – high in calories and low in physical activity – the advantage of such a system makes little sense. However, under presumed ancient lifestyles, in which caloric intake almost balanced caloric consumption, the advantage of such a "thrifty genotype" becomes clearer especially if it helped our ancestors survive intermittent episodes when food was sparse [26-28].
• Natural selection operates to maximize the number and survival of one's offspring (and thereby one's genes) in the next generation and not to maximize health or longevity. Regarding longevity, dramatic examples of this principle are seen in many insects in which aging and death is swift or almost immediate after mating. In humans, it seems to account for the process of human senescence – the general deterioration of the body with age and the rather sharp up-turn in chronic diseases in post-reproductive years. Natural selection selects genes on the basis of the benefits they provide early in life even if they may cause adverse effects later on, a concept known as antagonistic pleiotropy [1,29](i.e. a gene having multiple but antagonistic effects). For example, an efficient mechanism of calcium deposition evolved to repair bone fractures early in life may lead to arteriosclerosis, excessive calcium build-up in arterial walls, later in life [1,29]. Unfortunately, the concept of antagonistic pleiotropy (and this possible example of it) advanced by George C. Williams in 1957 [29], has never been rigorously tested. Nonetheless, the concept has been recently applied in attempts to explain Alzheimer's disease, prostate hyperplasia, and certain types of cancer [30]. Regarding Alzheimer's disease, a recently advanced evolutionary hypothesis [31,32] is built on findings of neuroprotective functions for estrogen and apolipoproteins E2 and E3. As estrogen declines in post-menopausal women, these apolipoproteins (uniquely evolved in humans) help delay the onset of Alzheimer's-like pathologies (and interestingly cardiovascular disease). With longevity increases over time in the human evolutionary lineage, older women having the neuroprotective apolipoproteins would have enjoyed a selective advantage since their ability to competently care for their grandchildren (with whom they share 1/4 of their DNA) was prolonged [31,32]. This special form of natural selection, in which a person gains a benefit by helping close relatives, is termed kin selection [1,33]. This hypothesis of the evolution of Alzheimer's has generated much commentary and many scientifically testable predictions (see [31] and associated commentaries).
Furthermore, the principle that natural selection promotes reproductive success and not necessarily health seems to explain the maintenance of certain common disease genes because of possible compensating advantages. Following are several examples. Homozygotes for phenylketonuria (PKU; frequency of 1 in 10,000 Caucasian births) develop mental retardation because they cannot metabolize the amino acid phenylalanine, but heterozygote infants seem to show decreased probability of being miscarried. Hemochromatosis is a very common autosomal recessive disease in individuals of Northern European descent (homozygote frequency of 1 in 200 births) characterized by increased intestinal iron absorption. It has been hypothesized that the gene may be advantageous to women by compensating for high iron losses (through menstruation, pregnancy, and milk production.) [20] On the other hand, men with the gene may suffer from excessive iron stores late in life. Supporting evidence may be seen in the recent finding of increased longevity of female carriers over non-carriers [34] and recent population genetic analyses of SNPs that have revealed positive natural selection in the gene's history [35]. More recently, a different hypothesis has been advanced based on the finding that individuals bearing the hemochromatosis mutation (C282Y) show partial resistance to Yersinia, the bacterium causing plague [36]. In normal individuals, Yersinia initially multiply within iron-rich macrophages, however this does not occur in persons with C282Y because their macrophages lack iron, a mineral essential for the pathogen's survival. This evolutionary hypothesis, which requires further testing, suggests that natural selection favored persons bearing the C282Y allele during the devastating plague that ravaged Europe during the Middle-ages [36].
Compensating advantages may also maintain disease genes associated with diabetes Type I (reduced miscarriage when linked to the beneficial HLA DR3 gene [1,20]), Fragile X syndrome (increased fertility among heterozygous females [37] or prenatal advantages to offspring [38]), gout and hyperuricemia (beneficial antioxidant effects [39,40]) among other examples [1,2,6,20]. Surprisingly, many such evolutionary hypotheses initially advanced a decade or more ago on the bases of observed biased inheritance patterns, or because of the commonness of the disorder and other circumstantial evidence, have not received rigorous scientific testing. In fact, for each disease or disorder, predictions could be made with respect to its association with a hypothesized compensating advantage, and experiments designed to test the predictions. It would then be possible to determine the validity of the evolutionary hypotheses, and to clarify the proximate genetic and physiological mechanisms involved.
Summary
We have described different aspects of human biology of medical significance. All have evolutionary explanations rooted in evolutionary theory. There are, of course, many more such explanations that can be incorporated into our texts and into medical curricula [1,2,6]. We recommend Nesse and William's thought-provoking book on this subject as a good and comprehensive introduction to the field [1]. Indeed, summaries of topics exist that could serve as syllabi for medical courses [41-43]. If we can help answer our student's questions by providing them with evolutionary answers (to the extent possible given the field's current knowledge) and by sufficiently describing evolutionary principles, we will do them a large favor. They will gain a powerful set of organizing principles around which they can arrange the enormous amounts of proximate information they must learn.
Furthermore, those of our students who in the future will do basic and clinical research may indeed find it fruitful to apply evolutionary theory and methods in order to understand problems of medical interest. Evolutionary medicine can have enormous potential. For example, one recent evolutionary analysis identified genes that in the past seem to have enabled viruses in the smallpox and vaccinia family to evade our defenses, genes that could be targeted for drug design today [44]. Furthermore, as described above leading hypotheses of the origins of type II diabetes [27,45], and obesity [26,27] have evolutionary explanations at their cores.
As a parting thought, imagine that the questions our students ask are like the endless series of "why" questions children ask, that go on (and on and on)! Such questions are also like their answers. They cycle on in a positive feedback loop yielding greater illumination. Although it is crucial we study the proximate mechanisms of disease, evolutionary theory offers a scientific methodology that can lead us to examine these mechanisms in novel ways. If we do not give our students support in asking such questions, might they stop asking them? What would be the consequences of this? Finally, as food for thought, we ask you to consider the following question: are you answering your students' "why" questions?
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Both authors contributed equally to this work.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank two reviewers who generously gave thoughtful suggestions that helped us improve the manuscript.
==== Refs
Nesse RM Williams GC Why We Get Sick: The New Science of Darwinian Medicine 1994 New York: Vantage Books
Williams GC Nesse RM The dawn of Darwinian Medicine Q Rev Biol 1991 66 1 22 2052670 10.1086/417048
Mayr E The Growth of Evolutionary Thought 1982 Cambridge: Belknap Press
Pilbeam DR Harrison RA, Tanner JM, Pilbeam, DR and Baker PT Human evolution Human Biology: An Introduction to Human Evolution, Variation, Growth, and Adaptability 1988 3 Oxford: Oxford University Press
Lovejoy CO Evolution of human walking Sci Am 1988 259 82 89 3072675
Trevathan W McKenna JJ Smith EO Evolutionary Medicine 1999 Oxford: Oxford University Press
Molnar S Human Variation: Races Types and Ethnic Groups 2002 5 New Jersey: Prentice Hall
Quinn T Rogers ZR Buchanan GR Survival of children with sickle cell disease Blood 2004 103 4023 4027 14764527 10.1182/blood-2003-11-3758
McKerill TDH Cohen HW Billett HH The older sickle cell patient Am J Hematol 2004 76 101 106 15164373 10.1002/ajh.20075
Lane PA Sickle cell disease Pediatr Clin North Am 1996 43 639 663 8649903
Fixler J Styles L Sickle cell disease Pediatr Clin North Am 2002 49 1193 1210 12580362 10.1016/S0031-3955(02)00089-5
Allison AC Two lessons from the interface of genetics and medicine Genetics 2004 166 1591 1599 15126382 10.1534/genetics.166.4.1591
Livingston FB Anthropological implications of sickle cell gene distribution in West Africa Am Anthropol 1958 60 533 562 10.1525/aa.1958.60.3.02a00110
Weatherall D The thalassemias: the role of molecular genetics in an evolving global health problem Am J Hum Genet 2004 74 385 92 15053011 10.1086/381402
Verrelli BC McDonald JH Argyropoulos G Destro-Bisol G Froment A Drousiotou A Lefranc G Helal AN Loiselet J Tishkoff SA Evidence for balancing selection from nucleotide sequencing of human G6PD Am J Hum Genet 2002 71 1112 28 12378426 10.1086/344345
Glader BE Look KA Hemotologic disorders in children from Southeast Asia Pediatr Clin North Am 1996 43 665 681 8649904
Ewald PW Evolution of Infectious Disease 1984 Oxford: Oxford University Press
DuPont HL Hornick RB Adverse effect of Lomotil therapy in shigellosis JAMA 1973 226 1525 1528 4587313 10.1001/jama.226.13.1525
Williams GC The Raymond Pearl memorial lecture 1997: The quest for medical normalcy – who needs it? Am J Hum Biol 2000 12 10 16 11533999 10.1002/(SICI)1520-6300(200001/02)12:1<10::AID-AJHB2>3.0.CO;2-N
Rotter JI Diamond J What maintains the frequencies of human genetic diseases? Nature 1987 329 289 290 3114647 10.1038/329289a0
Pier GB Grout M Zaidi T Meluleni G Mueschenborn SS Banting G Ratcliff R Evans MJ Colledge WH Salmonella typhi uses CFTR to enter intestinal epithelial cells Nature 1998 393 79 82 9590693 10.1038/30006
Lyczak JB Cannon CL Pier GB Lung infections associated with cystic fibrosis Clin Microbiol Rev 2002 15 194 222 11932230 10.1128/CMR.15.2.194-222.2002
Eaton SB Strassman BI Nesse RM Neel JV Ewald PW Williams GC Weder. AB Eaton SB IIILindeberg S Konner MJ Mysterud I Cordain L Evolutionary health promotion Prev Med 2002 34 109 118 11817903 10.1006/pmed.2001.0876
Eaton SB Cordain L Lindeberg S Evolutionary health promotion: a consideration of common counterarguments Prev Med 2002 34 119 123 11817904 10.1006/pmed.2001.0966
Neel JV Weder AB Julius S Type II diabetes, essential hypertension, and obesity as "syndromes of impaired genetic homeostasis": the "thrifty genotype" hypothesis enters the 21st century Perspect Biol Med 1998 42 44 74 9894356
Banks WA Coon AB Robinson SM Moinuddin A Shultz JM Nakaoke R Morley JE Triglycerides induce leptin resistance at the blood brain barrier Diabetes 2004 53 1253 1260 15111494
Banks WA The many lives of leptin Peptides 2004 25 331 338 15134858 10.1016/j.peptides.2004.02.014
Neel JV Diabetes Mellitus: a thrifty genotype rendered detrimental by "progress"? Am J Hum Genet 1962 14 353 362 13937884
Williams GC Pleiotropy, natural selection, and the evolution of senescence Evolution Int J Org Evolution 1957 11 398 411
Wick G Berger P Jansen-Dürr P Grubeck-Loebenstein B A Darwinian evolutionary concept of age-related diseases Exp Gerontol 2003 38 13 25 12543257 10.1016/S0531-5565(02)00161-4
Finch CE Sapolsky RM The evolution of Alzheimer disease, the reproductive schedule, and apoE isoforms Neurobiol Aging 1999 20 407 428 10604433 10.1016/S0197-4580(99)00053-6
Sapolsky RM Finch CB Alzheimer's disease and some speculations about the evolution of its modifiers Ann N Y Acad Sci 2000 924 99 103 11193810
Maynard Smith J Group selection and kin selection Nature 1964 201 1145 1147
Lio D Balistreri CR Colonna-Romano G Motta M Franceschi C Malaguarnera M Candore G Caruso C Association between the MHC class I gene HFE polymorphism and longevity: a study in Sicilian Populations Genes Immun 2002 3 20 24 11857056 10.1038/sj.gene.6363823
Toomajian C Ajioka RS Jorde LB Kushner JP Kreitman M A method for detecting recent selection in the human genome from allele estimates Genetics 2003 165 287 97 14504236
Moalem S Percy ME Kruck TP Gelbart RR Epidemic pathogenic selection: an explanation for hereditary hemochromatosis? Med Hypotheses 2002 59 325 329 12208162 10.1016/S0306-9877(02)00179-2
Vogel F Crusio WE Kovac C Fryns JF Freund M Selective advantage of fra (X) heterozygotes Hum Genet 1990 86 25 32 2147668 10.1007/BF00205167
Drasinover V Ehrlich S Magel N Taub E Libman V Shohat T Halpern GJ Shohat M Increased transmission of intermediate alleles of the FMR1 gene compared with normal alleles among female heterozygotes Am J Med Genet 2000 93 155 157 10869119 10.1002/1096-8628(20000717)93:2<155::AID-AJMG13>3.0.CO;2-G
Ames BN Cathcart R Schwiers E Hockstein P Uric acid provides an antioxidant defense in humans against oxidant- and radical-caused aging and cancer: a hypothesis Proc Natl Acad Sci 1981 78 6858 62 6947260
Benzie FF Evolution of antioxidant defense mechanism Eur J Nutr 2000 39 53 61 10918985 10.1007/s003940070030
Nesse RM Williams GC Evolutionary biology in the medical curriculum – what every physician should know BioScience 1997 47 664 68
Downie JR Evolutionary biology Lancet 2004 363 1168 15064054 10.1016/S0140-6736(04)15922-9
Stearns SC Ebert D Evolution in health and disease: work in progress Q Rev Biol 2001 76 417 432 11783396 10.1086/420539
McLysaght A Baldi PF Gaut BS Extensive gene gain associated with adaptive evolution of poxviruses Proc Natl Acad Sci USA 2003 100 15655 15660 14660798 10.1073/pnas.2136653100
Diamond J The double puzzle of diabetes Nature 2003 423 599 602 12789325 10.1038/423599a
| 15885137 | PMC1142319 | CC BY | 2021-01-04 16:30:56 | no | BMC Med Educ. 2005 May 10; 5:16 | utf-8 | BMC Med Educ | 2,005 | 10.1186/1472-6920-5-16 | oa_comm |
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-231588245910.1186/1471-2180-5-23Research ArticleSerotypes, intimin variants and other virulence factors of eae positive Escherichia coli strains isolated from healthy cattle in Switzerland. Identification of a new intimin variant gene (eae-η2) Blanco Miguel [email protected] Sandra [email protected] Taurai [email protected] Claudio [email protected] Jesús E [email protected] Ghizlane [email protected] Jorge [email protected] Roger [email protected] Laboratorio de Referencia de E. coli (LREC), Departamento de Microbioloxía e Parasitoloxía, Facultade de Veterinaria, Universidade de Santiago de Compostela (USC), Lugo, Spain2 Institute for Food Safety and Hygiene, Vetsuisse Faculty, University of Zurich, CH-8057 Zurich, Switzerland2005 9 5 2005 5 23 23 21 2 2005 9 5 2005 Copyright © 2005 Blanco 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
Enteropathogenic Escherichia coli (EPEC) and Shigatoxin-producing Escherichia coli (STEC) share the ability to introduce attaching-and-effacing (A/E) lesions on intestinal cells. The genetic determinants for the production of A/E lesions are located on the locus of enterocyte effacement (LEE), a pathogenicity island that also contains the genes encoding intimin (eae). This study reports information on the occurrence of eae positive E. coli carried by healthy cattle at the point of slaughter, and on serotypes, intimin variants, and further virulence factors of isolated EPEC and STEC strains.
Results
Of 51 eae positive bovine E. coli strains, 59% were classified as EPEC and 41% as STEC. EPEC strains belonged to 18 O:H serotypes, six strains to typical EPEC serogroups. EPEC strains harbored a variety of intimin variants with eae-β1 being most frequently found. Moreover, nine EPEC strains harbored astA (EAST1), seven bfpA (bundlin), and only one strain was positive for the EAF plasmid. We have identified a new intimin gene (η2) in three bovine bfpA and astA-positive EPEC strains of serotype ONT:H45. STEC strains belonged to seven O:H serotypes with one serotype (O103:H2) accounting for 48% of the strains. The majority of bovine STEC strains (90%) belonged to five serotypes previously reported in association with hemolytic uremic syndrom (HUS), including one O157:H7 STEC strain. STEC strains harbored four intimin variants with eae-ε1 and eae-γ1 being most frequently found. Moreover, the majority of STEC strains carried only stx1 genes (13 strains), and was positive for ehxA (18 strains) encoding for Enterohemolysin. Four STEC strains showed a virulence pattern characteristic of highly virulent human strains (stx2 and eae positive).
Conclusion
Our data confirm that ruminants are an important source of serologically and genetically diverse intimin-harboring E. coli strains. Moreover, cattle have not only to be considered as important asymptomatic carriers of O157 STEC but can also be a reservoir of EPEC and eae positive non-O157 STEC, which are described in association with human diseases.
==== Body
Background
Enteropathogenic Escherichia coli (EPEC) and Shiga toxin-producing Escherichia coli (STEC) represent two of the at least six different categories of diarrheagenic E. coli recognized at present [1]. Unlike other diarrheagenic E. coli, EPEC and STEC share the ability to introduce attaching-and-effacing (A/E) lesions on intestinal epithelial cells. A/E lesions are characterized by destruction of the microvillus brush border through restructuring of the underlying cytoskeleton by signal transduction between bacterial and host cells, intimate adherence of strains to the intestinal epithelium, pedestal formation and aggregation of polymerized actin at the sites of bacterial attachment [1,2]. The genetic determinants for the production of A/E lesions are encoded in a pathogenicity island called the locus of enterocyte effacement (LEE) [3]. The LEE encodes the outer membrane protein intimin, which is encoded by the eae gene (for E. coli attachment effacement) localized in the central region of LEE, a type III secretion system, a number of secreted proteins (ESP), and Tir (for translocated intimin receptor), a protein encoded upstream of the eae gene, which is translocated into host cells [2]. Characterization of eae genes revealed the existence of different variants. At present, 17 genetic variants (α1, α2, β1, ξR/β2B, δ/κ/β2O, γ1, θ/γ2, ε1, νR/ε2, ζ, η, ι1, μR/ι2, λ, μB, νB, ξB) have been identified (Table 4) [4-9]. The heterogeneous C-terminal (3') end of intimin is responsible for receptor binding, and it is discussed that different intimin variants may be responsible for different host and tissue cell tropism [6,10-12]. Adu-Bobie et al. [4] found that antigenic variation exists within the cell-binding domain of intimins expressed by different clinical human EPEC and STEC isolates and defined four intimin types (α, β, γ, δ) based on type-specific PCR assays that used oligonucleotide primers complementary to the 3'end of specific eae genes. Oswald et al. [5] described a type-specific PCR assay which identifies a fifth intimin variant (intimin ε; referred in the present study as ε1). They divided the intimin alleles α, β, and γ based on restriction fragment length polymorphism (RFLP) analysis of PCR products into α1, α2, β1, β2, γ1 and γ2 subtypes. Furthermore, Oswald et al. [5] reclassified the δ type of Adu-Bobie et al. [4] as a β2 subtype (referred in the present study as d/β2O). Tarr and Whitam [13] presented a paper on the molecular evolution of intimin genes in human STEC and EPEC O111 clones and found two new types (ζ and θ). The new ζ intimin type also was identified by Jores et al [11] in bovine STEC and Blanco et al. [7] in ovine STEC. Zhang et al. [6] observed that the sequences of eae-γ2 and eae-θ were almost identical (99%), and they believe that these two sequences should be considered one eae variant (γ2/θ), as is referred in the present study. Recently, Zhang et al. [6] determined the sequences of three new intimin variant genes (κ, η, and ι; intimin ι referred in the present study as ι1) found in human STEC strains. The sequence of the 3'variable region of eae gene of a new intimin (λ) has been submitted to GenBank by China (unpublished data) (Accession no. AF439538) and the whole intimin λ gene was sequenced by Blanco et al. (unpublished data) from a human strain (Accession no. AJ715409) and Ramachandran et al. [14] (Accession no. AF530557) from a bovine strain. Ramachandran et al. [14] identified three new intimin genes in ruminant E. coli strains belonging to serotypes ONT:H- (intimin μ, referred in the present study as μR/ι2), O2-related:H19 (intimin ν, referred in the present study as νR/ε2) and ONT:HNT (intimin ξ, referred in the present study as ξR/β2B) and they have also developed an intimin typing PCR-RFLP scheme that reliably differentiates 14 intimin variants. When Ramachandran et al. [14] submitted to GenBank the nucleotide sequences of μ, ν, and ξ genes these intimin types were designated ι2, ε2 and β2, respectively. Blanco et al. [8] have identified four new intimin variant genes that they originally designated as β2, μ, ν, and ξ when the sequences were submitted to EMBL Nucleotide Sequence Database, and before knowing the results obtained by Ramachandran et al. [14]. The intimin β2 found by Blanco et al. (unpublished data) in human typical EPEC strains of classical EPEC serotype O119:H6 is identical to intimin ξ described by Ramachandran et al. [14] in one bovine strain of serotype ONT:HNT. Thus, in this study our β2 intimin is referred as ξR/β2B. The other three intimins (μ, ν, and ξ) discovered by Blanco et al. [[8], unpublished data] are different to the existing intimin types and are referred to as μB, νB, and ξB in this study, respectively.
EPEC strains are defined as eae-haboring diarrheagenic E. coli that possess the ability to form A/E lesions on intestinal cells and that do not possess Shiga toxin genes [15]. Most EPEC strains belong to a series of O antigenic groups known as EPEC serogroups: O26, O55, O86, O111, O114, O119, O125, O126, O127, O128, O142, and O158. EPEC are further classified as typical, when possessing the EAF (for EPEC adherence factor) plasmid that encodes localized adherence (LA) on cultured epithelial cells mediated by the Bundle Forming Pilus (BFP); whereas atypical EPEC strains do not possess the EAF plasmid [15,16]. Typical EPEC, a major cause of infant diarrhea in developing countries, are rare in industrialized countries, where atypical EPEC seem to be a more important cause of diarrhea [1,16]. Typical and atypical EPEC strains usually belong to certain serotype clusters, differ in their adherence patterns on cultured epithelial cells (typical: LA; atypical: diffuse adherence (DA); aggregative adherence (AA); localized adherence-like adherence (LAL)), are typically found in different hosts (typical EPEC strains have only been recovered from humans), and display differences in their intimin variants [16]. Atypical EPEC appear to be more closely related to STEC and as such are considered emerging pathogens [16,17].
STEC are responsible for a number of human gastrointestinal diseases, including diarrhea, bloody diarrhea and hemorrhagic colitis (HC). In a proportion of individuals, particularly in children, these conditions may be complicated by neurological and renal sequelae, including hemolytic-uremic syndrome (HUS) [18-20]. Most outbreaks and sporadic cases of HC and HUS have been attributed to O157:H7 STEC strains [20-22]. In contrast, especially in continental Europe, infections by non-O157 STEC, such as O26:H11/H-, O91:H21/H-, O103:H2, O111:H-, O113:H21, O117:H7, O118:H16, O121:H19, O128:H2/H-, O145:H28/H-, and O146:H21, are more common and frequently associated with severe illness in humans [1,23-27]. Probably, non-O157 STEC also play a more important role in disease in Argentina [28], Australia [29], Chile [1], and South Africa [1]. STEC are characterized by elaborating two groups of potent phage-encoded cytotoxins called Shiga toxins [18,19]. Pathogenic STEC very often possess other putative virulence factors as intimin (eae) and Enterohemolysin (ehxA) [19,30,31].
Although a number of studies have determined the eae subtypes of STEC strains isolated from humans, only a limited number of studies have been undertaken to determine the intimin variants from a diverse collection of STEC and EPEC isolated from cattle. Ruminants, especially cattle, represent an important reservoir of atypical EPEC and STEC [16,20,32,33]. This study reports information on the occurrence of eae positive E. coli carried by healthy cattle at the point of slaughter, and on serotypes, intimin variants, and further virulence factors of isolated EPEC and STEC strains.
Results
Detection and isolation of eae positive E. coli strains
Of the 330 fecal samples collected from cattle at slaughter, 132 (40%) tested positive for eae genes. From 51 randomly selected eae PCR-positive samples, 51 eae positive E. coli strains were isolated by colony dot blot hybridization, confirmed as E. coli by biochemical properties, and further characterized by phenotypic and genotypic traits. Thirty strains carried no stx genes and were therefore classified as EPEC, whereas the remaining 21 strains possessed at least one Shiga toxin gene, and were therefore considered as STEC.
Serotypes of bovine eae positive E. coli strains
The 51 eae positive E. coli strains belonged to 18 O serogroups, 9 H types, and 23 O:H serotypes. However, 46% of strains were of three serogroups, namely O26 (6 strains), O103 (12 strains), and O145 (5 strains), and 43% belonged to two H types, namely H2 (17 strains), and H45 (5 strains). Thirty-two percent of strains were of only three serotypes: O26:H- (3 strains), O103:H2 (11 strains), and O145:H- (5 strains). EPEC strains belonged to 18 O:H serotypes and eight strains were nontypeable (Table 1). STEC strains comprised seven O:H serotypes with one serotype (O103:H2) accounting for 48% of STEC strains (Table 2). Furthermore, all serotypes found in this study have been previously reported in human STEC. Five (O5:H-, O26:H-, O103:H2, O145:H-, and O157:H7) of the STEC serotypes comprising 19 (90%) strains have previously been associated with human STEC causing HUS. Serotypes O26:H- and O103:H2 were found in both EPEC and STEC strains.
Typing of intimin (eae) genes
Overall, the 51 isolated eae positive bovine E. coli strains comprised a variety of 10 intimin variants, namely α1(1 strain), β1 (17 strains), ξR/β2B (2 strains), γ1 (6 strains), γ2/θ (2 strains), ε1 (13 strains), ζ (3 strains), η2 (3 strains), ι1 (2 strains), and κ (2 strains). Intimins α2, δ/β2O, γ1, θ/γ2, ν R/ε2, η1, μR/ι2, λ, μB, νB, and ξB were neither detected in bovine EPEC nor in bovine STEC. EPEC strains harbored a variety of nine intimin variants (α1, β1, ξR/β2B, κ, θ/γ2, ε1, ζ, η2, ι1). In 15 strains (52%) eae-β (twelve times β1, twice β2), which was associated with nine serotypes, in three strains eae-ε1, in three strains eae-ζ, in three strains eae-η2, in two strains eae-ι1, in two strains eae-κ, in one strain eae-α1, and in one strain eae-γ2/θ were detected (Table 1). In contrast, STEC strains harbored only four intimin variants (β1,γ1,γ2/θ, ε1) with eae-ε1 (10 strains, all of serotype O103:H2) and eae-γ1 (5 strains of serotype O145:H-, and one strain of serotype O157:H7) being most frequently found (Table 2). Intimins β1, ε1 and γ2/θ were present in both EPEC and STEC strains.
Identification of a new intimin variant gene (eae-η2). Sequence comparison
The complete nucleotide sequence of the new η2 variant gene in three bovine bfpA and astA-positive EPEC strains of serotype ONT:H45 was determined. Furthermore, a fragment (566 bp strain FV5126 to 1728 bp strain FV5114/2) of the 3'variable region of the eae gene of seven representative strains was amplified and sequenced. The eae sequences were deposited in the European Bioinformatics Institute (EMBL Nucleotide Sequence Database) and the accession numbers assigned are indicated in table 3.
By using CLUSTAL W for optimal sequence alignment, we determined the genetic relationship of the new intimin gene (η2) and the remaining eae variants (Table 4). Thus, the eae-η2 sequence is very similar to eae-η1 (identy of 98%). Phylogenetic analysis revealed six groups of the closely related intimin genes: (i) α1, α2, ζ, and νB; (ii) β1, δ/β2O, κ, ξR/β2B; (iii) ε1, ξB, η1, η2 and νR/ε2; (iv) γ1, μB, and γ2/θ; (v) λ ; and (vi) ι1 and μR/ι2.
Further characterization of EPEC and STEC strains
Nine EPEC strains harbored the astA gene encoding for EAST1 (Table 1). Additionally, three astA positive strains of serotype ONT:H45 harbored the bfpA gene encoding bundlin, whereas only one strain was positive for the EAF plasmid. Overall, 23 different associations of serotypes with virulence factors were identified in EPEC strains.
Of the 21 eae positive E. coli harboring genes encoding for Shiga toxins, the majority (13 strains) carried only stx1 genes, 4 strains only stx2 genes, and 4 strains both toxin genes (Table 2). The only stx1 positive strains belonged to serotypes O5:H- (2 strains), O103:H2 (10 strains) and O111:H21 (1 strain), the only stx2 positive strains to serotypes O26:H-, O26:H2, O145:H-, and O157:H7, and all the strains positive for both toxin genes to serotype O145:H-. Moreover, the majority of STEC strains (86%) tested positive for the ehxA gene encoding for Enterohemolysin. Overall, ten different associations of serotypes with virulence factors were identified.
Discussion
Of 51 eae positive E. coli strains, 59% were classified as EPEC and 41% as STEC. The serogroups O26, O103, and O145 and serotypes O26:H-, O103:H2, O145:H- were most frequently found. Comparable distributions of EPEC (57%) and STEC (43%) strains and serogroups (44% of strains serotyped as O26, O103, O145, or O156) were reported in strains isolated from healthy cattle in Japan [34]. Of the typical EPEC serogroups (O26, O55, O86, O111, O114, O119, O125, O126, O127, O128, O142, and O156), only six strains were found in the present study: four O26 strains, one O119 strain, and one O128 strain. In bovine STEC strains, only a restricted number of serotypes have been most commonly found [8,20]: the predominant serotype in the majority of surveys realized in Europe was O113:H21, in America and Australia O26:H11, and in Japan O45:H8/H- and O145:H-. As in the present study, authors in Argentina, Canada, France, Germany, Spain, and the United States have found that many STEC recovered from cattle belonged to serotypes previously associated with human disease [8,35-38].
Intimin mediates the intimate bacterial attachment to the host cell surface of EPEC and STEC, and is required for the formation of the characteristic A/E lesions. In EPEC and STEC strains isolated from human patients and from healthy infants from Germany and Australia, intimin variants α1, β1, γ1 and γ2/θ were most frequently detected [5,6,8,17,26]. Moreover, STEC serotypes commonly recovered from outbreaks of HUS and hemorrhagic colitis (O157:H7/H-; O111:H-; O26:H11/H-) typically possess intimin γ1, γ2/θ, and β1 [5,6,8,14].
Overall, the 51 isolated eae positive bovine E. coli strains comprised a variety of 10 intimin variants, and most of them harbored eae-β1(33%), eae-ε1 (26%) and eae-γ1(16%). As it is known that different intimin variants are associated with distinctive phylogenetic lineages of LEE-positive E. coli [4,39,40], the identified variety of bovine eae positive strains further substantiates the reservoir function of cattle for infections of humans with LEE positive strains.
Intimin β1 appears to be the most widespread variant [7,8,14,27,39,41,42]. In that it has been found in both EPEC and STEC strains from humans and several animal species, and its presence is associated with multiple E. coli serotypes. Strains harboring eae-β1 include important human diarrheagenic serotypes such as EPEC O26:H11, O111:H2, O114:H2, O126:H2, O128:H2 and STEC O26:H11/H-, O118:H16 and O177:H- [5,26,27], serotypes partly also detected in this study.
Intimin ε1 was described for the first time in human and bovine STEC of serogroup O8, O11, O45, O103, O121, and O165 [5]. In this study, eae-ε1, was detected in three bovine EPEC strains including one O103:H2 strain, and in all 9 bovine STEC O103:H2 strains. Furthermore, in a recent study performed on sheep in Switzerland, intimin ε1 was detected in all STEC O103:H2 [43]. STEC O103:H2 strains have frequently been associated with HUS and hemorrhagic colitis [26,27,44,45].
Intimin γ1 was restricted to STEC and associated with serotypes considered as highly pathogenic to humans: O145:H- (five strains) and, O157:H7 (one strain). In addition to these serotypes, eae-γ1 was previously described in association with STEC O145:H28, and with EPEC O55:H7/H-, from which STEC O157:H7 are believed to have evolved [8,26,27,46].
Interestingly, one bovine EPEC O157:H45 strain harboring eae-α1 was detected. Moreover, eae-α1 was also recently detected in bovine EPEC O157:H45 [33], in two EPEC O142:H6 isolated from monkeys [47] and in a human STEC O177:H7 strain [25]. These findings disagree with the suggested restriction of intimin α1 to human EPEC strains of serotypes O55:H6, O127:H6, H34/H-, O142:H6 [4,5].
We have identified a new intimin gene (η2) in three bovine bfpA and astA-positive EPEC strains of serotype ONT:H45. The complete nucleotide sequences of the new eae-η2 (AJ879898, AJ879899, AJ879900) variant genes of the three strains were determined. The new intimin η2 also was found in one EPEC ONT:H45 strain isolated from a patient with diarrhea in Spain by Blanco et al. (unpublished data, EMBL nucleotide sequence AJ876652). Recently, a food-borne outbreak of diarrhea involving 41 students (ages 12 to 15) was reported in Japan by Yatsuyanagi et al. [48]. The implicated organism was a EPEC ONT:H45, which hybridized with the probes for eaeA, astA and bfpA genes.
In our study, about one third of EPEC strains harbored astA as further putative virulence factor; whereas only seven EPEC strains harbored bfpA encoding bundling, the structural subunit of the bundle-forming pilus in typical EPEC strains, and only one EPEC of serotype O157:H45 was positive for the EAF plasmid. Overall, bfpA seems to be isolated very rarely from bovine EPEC strains, since there are scarcely any reports of E. coli strains with bfpA isolated from cattle. However, in a recent study, we characterized 11 O157:H45 EPEC strains isolated from cattle in Switzerland and found 10 bfpA positive strains among them [33].
To assess the pathogenicity of STEC, further evaluation of virulence factors in addition to serotype is necessary. O157 and non-O157 STEC strains isolated from patients with severe symptoms such as bloody diarrhea, HC, and HUS frequently show a typical virulence spectrum, with such strains tending to be stx2 and eae positive [49-51]. Moreover, it was previously shown that bovine and human strains harboring the eae gene were statistically more likely to be positive for the ehxA gene encoding for Enterohemolysin [7,26,52]. However, the impact of Enterohemolysin, which was found in the majority of STEC strains, is controversially discussed. Additionally, STEC commonly recovered from outbreaks of HUS and HC typically possessed eae-β1, eae-γ1 and γ2/θ [5-7,39]. In consequence, apart of the stx2 gene, which was previously reported to correlate with severe disease in humans, intimin subtyping may facilitate further understanding of associations among serotype, eae and stx subtype. In consequence, seven STEC strains (14% of all eae positive strains and 33% of STEC strains) showed a virulence pattern characteristic of highly virulent human strains: four O145:H- strains harboring stx1, stx2, eae-γ1, and ehxA, one O26:H- strain harboring stx2, eae-β1, and ehxA, one O145:H- strain harboring stx2, eae-γ1, and ehxA, and one O157:H7 strain harboring stx2, eae-γ1, and ehxA.
Conclusion
Our data confirm that ruminants are an important source of serologically and genetically diverse intimin-harboring E. coli strains. Moreover, cattle have not only to be considered as important asymptomatic carriers of O157 STEC but can also be a reservoir of EPEC and eae positive non-O157 STEC, which are described in association with human diseases. The fecal carriage of foodborne pathogens among livestock animals at slaughter is strongly correlated with the hazard of carcasses contamination. In order to reduce the risk represented by STEC and EPEC, the maintenance of slaughter hygiene is consequently of central importance in meat production.
Methods
E. coli strains
From 330 fecal samples, collected from Swiss cattle at slaughter, 10 g were each enriched in 100 ml brilliant green bile broth (BBL, Cockeysville, Md.) at 37°C for 24 h. The enriched samples were streaked onto sheep blood agar (Difco Laboratories, Detroit, Mich.; 5% sheep blood Oxoid, Hampshire, UK), and after incubation at 37°C for another 24 h, the colonies were washed off with 2 ml of 0.85% saline solution. Two μl of each plate eluate were then evaluated by PCR with primers EAE-1 and EAE-2 (Table 5) targeting sequences at the 5' eae conserved region detecting all types of eae described at the moment. From the 132 eae PCR-positive samples, 51 eae PCR-positive samples were then randomly selected for strain isolation with an eae DNA probe and colony dot-blot hybridization.
The eae probes were prepared by labeling eae-PCR amplicons from E. coli O157:H7 strain 857/03 with DIG High Prime kit (Roche, Mannheim, Germany). Briefly, for colony hybridization, the 51 eae positive samples were plated onto sheep blood agar and incubated overnight at 37°C. Colonies were transferred to a nylon membrane (Roche), and lysed following standard methods. After washing, crosslinking, and prehybridization in DIG-Easy-Hyb buffer (Roche) at 37°C for about 30 min, hybridization of membranes with eae DNA probes was performed overnight at 42°C. After washing in pre-heated primary and secondary wash buffers, the presence of labeled probe was detected with in alkaline phosphatase-conjugated antibody detection kit and NBT/BCIP stock solution according to the instructions of the manufacturer (Roche). Positive colonies were picked from the original sheep blood agar and confirmed as E. coli by biochemical properties (acid production from mannitol, the o-nitrophenyl-β-D-galactopyranoside (ONPG) test, H2S and indole production, and proof of urease and lysine decarboxylase activities) and by PCR to be eae positive [7]. One randomly chosen colony per sample was used for further strain characterization.
Further strain characterization
Determination of O and H antigens was performed by the method described by Guinée et al. [53] with all available O (O1 to O185) and H (H1 to H56) antisera. E. coli isolates that are nonmotile are described as having an H- flagellum type. Antisera were obtained and absorbed with corresponding cross-reaction antigens to remove nonspecific agglutinins. O antisera were produced in the Laboratorio de Referencia E. coli (LREC) , and H antisera were obtained from the Statens Serum Institut (Copenhagen, Denmark). E. coli strains representing the new O groups O182 to O185 [unpublished data] were kindly provided by Flemming Scheutz (International Escherichia Centre, Statens Serum Institut, Copenhagen, Denmark).
All PCR assays applied in this study for characterization of eae positive E. coli strains were performed in a T3 thermocycler (Biometra, Göttingen, Germany). PCR reagents were purchased from PROMEGA (Madison, Wis.), and primers were synthesized by MICROSYNTH (Balgach, Switzerland). The 50-μl PCR mixtures normally consisted of 2 μl of bacterial suspension boiled at 100°C for 5 min in 42 μl of double-distilled water, 5 μl of 10-fold-concentrated polymerase synthesis buffer containing 2.0 mM MgCl2, 200 μM (each) desoxynucleosid triphosphate (dNTP), 30 pmol of each primer, and 2.5 U of Taq DNA polymerase. The PCR primers, target sequences, product sizes and references are listed in Tables 5 and 6. Differences in the normal PCR mixture and cycling conditions for each PCR were previously described in the cited literature. To identify eae variants, intimin type specific PCR assays using primers complementary to the heterogeneous 3' end of intimin genes were performed [8,9]. For detection of further putative virulence genes all strains were examined for the presence of stx genes with primers VT1 and VT2 [54]. Stx positive strains were examined for the presence of stx1 and stx2 genes [55,56]. Stx negative strains were further examined for the presence of the EAF plasmid [57], the bfpA gene on the EAF plasmid [58], and the astA gene encoding EAST1 [59].
Sequencing of the intimin (eae) genes
The nucleotide sequence of the amplification products purified with a QIAquick DNA purification kit (Qiagen) was determined by the dideoxynucleotide triphosphate chain termination method of Sanger, with the BigDye Terminator v3.1 Cycle Sequencing Kit and an ABI 3100 Genetic Analyzer (Applied Bio-Systems).
Phylogenetic analyses
Genetic distances and phylogenetic trees of eae sequences were calculated and constructed with the CLUSTAL W program [60]included in the EMBL sofware .
Nucleotide sequence accession numbers
The eae sequences of strains analyzed were deposited in the European Bioinformatics Institute (EMBL Nucleotide Sequence Database) and the accession numbers assigned are indicated in Table 3.
Authors' contributions
RS and JB designed the study and drafted the manuscript. SS and TT isolated the strains, JEB has done serotyping of the strains, MB designed the primers used for PCR typing of intimins, performed the PCR typing of eae genes and sequenced the eae genes, GD also participated in the PCR typing of intimins, and CZ was responsible for further strain characterization. All authors read, commented on and approved of the final manuscript.
Acknowledgements
This work was partly supported by grant from the Fondo de Investigación Sanitaria (FIS G03-025-COLIRED-O157), and the Xunta de Galicia (grants PGIDIT02BTF26101PR and PGIDIT04RAG261014PR).
Figures and Tables
Table 1 Serotypes and virulence factors of EPEC strains isolated from cattle in Switzerland (n = 30)
Serotype No. of strains eae astA bfpA
O2:H45 1 κ - -
O8:H19 1 β1 + -
O10:H- 1 γ2/θ - -
O15:H2 3 β1 - -
O26:H- 1 β1 + -
O26:H- 1 β1 - -
O26:H11 2 β1 - -
035:H2 1 β1 - -
O64:H25 1 β1 - -
O77:H19 1 β1 - -
O103:H2 1 ε1 - -
O103:H8 1 ι1 + +
O113:H6 2 ξR/β2B - -
O119:H8 1 ι1 - +
O128:H2 1 β1 - -
O156:H25 1 ζ - -
O157:H45 1 α1 + +
O177:H11 1 β1 + +
ONT:H- 1 ε1 - -
ONT:H- 1 κ - -
ONT:H19 1 ε1 + -
ONT:H25 2 ζ - -
ONT:H45 3 η2 + +
ONT: O antigen nontypeable with O1 to O185 antisera.
Table 2 Serotypes and virulence factors of eae positive STEC strains isolated from cattle in Switzerland (n = 21)
Serotype No. of strains stx eae ehxA
O5:H- b 1 stx1 β1 +
O5:H- b 1 stx1 β1 -
O26:H- b 1 stx2 β1 +
O26:H2 a,c 1 stx2 β1 -
O103:H2 b 9 stx1 ε1 +
O103:H2 b 1 stx1 ε1 -
O111:H21 a 1 stx1 γ2/θ +
O145:H- b 4 stx1, stx2 γ1 +
O145:H- b 1 stx2 γ1 +
O157:H7 b 1 stx2 γ1 +
a Serotypes previously found as human STEC strains
b Serotypes previously associated with human STEC strains that caused HUS
c Serotypes not yet reported within bovine STEC strains
Serotypes that caused human outbreaks are in bold
Table 3 Accession numbers of sequenced eae genes of some EPEC strains
Strain Origin Serotype Intimin EMBL nucleotide sequence bfpA EAF plasmid stx1 stx2
FV5109-4113/1 healthy cattle1 ONT:H45 η2 AJ879898 + - -
FV5113-4115/2 healthy cattle1 ONT:H45 η2 AJ879899 + - -
FV5114/1-3933/51 healthy cattle1 ONT:H45 η2 AJ879900 + - -
H03-53199a patient with diarrhea2 ONT:H45 η2 AJ876652 nd nd -
FV5114/2-3933/52 healthy cattle1 O156:H25 ζ AJ879901 - - -
FV5125-4125/1 healthy cattle1 O157:H45 α1 AJ879902 + - -
FV5123-4286/1 healthy cattle1 O2:H45 κ AJ879903 - - -
FV5133-3988/1 healthy cattle1 ONT:H- κ AJ879904 - - -
FV5126-3951/1 healthy cattle1 O119:H8 ι1 AJ879905 + - -
FV5120-4075/1 healthy cattle1 O103:H8 ι1 AJ879906 + - -
FV3854-1070/1 healthy cattle3 O157:H45 α1 AJ879907 + - -
1from Switzerland; 2from Spain; 3Stephan et al. [33]; nd: not done
Table 4 Genetic relationship of the new intimin gene (η2) (AJ879898, AJ879899, AJ879900) detected in bovine EPEC strains of serotype ONT:H45 and the remaining eae variants: Pairwise alignments calculated with CLUSTAL W
Designation of intimins in this study Designation of intimins in previous studies (References) ORF length (bp) Reference strain Serotype Origin Accession number Genetic relationship identities
α1 α (4) and α1 (5) 2820 E2348/69 O127:H6 Human M58154 87%
α2 α (4) and α2 (5) 2820 E. coli O125:H6 Human AF530555 87%
β1 β (4) and β1 (8) 2820 RDEC1 O15:H- Rabbit AF200363 84%
ξR/β2B ξ (14) and β2 (8) 2820 KB411 ONT:HNT Bovine AF530556 88%
ξR/β2B ξ (14) and β2 (8) 2820 FV359 O119:H6 Human AJ715407 88%
δ/β2O δ (4) and β2 (5) 2820 BL152.1 O86:H34 Human AJ875027 87%
κ κ (6) 2820 6044/95 O118:H5 Human AJ308552 87%
γ1 γ (4) and γ1(5) 2805 EDL933 O157:H7 Human AF071034 86%
γ2/θ γ2 (5) and θ (13) 2808 CL-37 O111:H8 Human AF449418 84%
ε1 ε(5) 2847 PMK5 O103:H2 Human AF116899 92%
νR/ε2 ν (14) and ε2 (GenBank AF530554) 2847 VR64/4 O2related:H19 Ovine AF530554 94%
ζ ζ (13) 2817 4795/95 O84:H4 Human AJ271407 86%
η1 η (6) 2847 CF11201 O125:H- Human AJ308550 98%
η2 new (this study) 2847 H03/53199a ONT:H45 Human AJ876652 100%
ι1 ι (6) 2814 7476/96 O145:H4 Human AJ308551 88%
μR/ι2 μ (14) and ι2 (GenBank AF530553) 2814 VR45 OR:H- Ovine AF530553 83%
λ λ (8, 14) 2817 EPEC-68.4 O34:H- Human AJ715409 86%
μB μ (8) 2808 EPEC-373 O55:H51 Human AJ705049 87%
νB ν (8) 2823 IH1229a O10:H- Human AJ705050 82%
ξB ξ (8) 2847 STEC-B49 O80:H- Bovine AJ705051 91%
Table 5 Sequences of oligonucleotide primers used for typing of eae intimin gene
Gene Primer Oligonucleotide sequence (5'-3')c Fragment size (bp) Annealing temperature Primer coordinates Accession number Reference
eaea EAE-1
EAE-2 GGAACGGCAGAGGTTAATCTGCAG
GGCGCTCATCATAGTCTTTC 775 55°C 1441–1460
2193–2215 AF022236 7
eae-α1 EAE-FB
EAE-A AAAACCGCGGAGATGACTTC
CACTCTTCGCATCTTGAGCT 820 60°C 1909–1928
2709–2728 AF022236 8
eae-α2 IH2498aF
IH2498aR AGACCTTAGGTACATTAAGTAAGC
TCCTGAGAAGAGGGTAATC 517 60°C 2099–2122
2597–2615 AF530555 8
eae-β1 B1A
B1B ACTTCGCCACTTAATGCCAGC
TTGCAGCACCCCATGTTGAAT 730 66°C 1924–1944
2633–2653 AF453441 9
eae-ξR/β2Bb,e B2A
B2B AAGGGGGGAACCCCTGTGTCA
ATTTATTCGCAGCCCCCCACG 604 66°C 2056–2076
2639–2659 AF530556
AJ715407 9
eae-δ/κ/β2Oc EAE-FB
EAE-D AAAACCGCGGAGATGACTTC
CTTGATACACCCGATGGTAAC 833 60°C 1909–1928
2721–2741 U66102 8
eae-γ1 EAE-FB
EAE-C1 AAAACCGCGGAGATGACTTC
AGAACGCTGCTCACTAGATGTC 804 60°C 1909–1928
2691–2712 AF071034 8
eae-θ/γ2d EAE-FB
EAE-C2 AAAACCGCGGAGATGACTTC
CTGATATTTTATCAGCTTCA 808 58°C 1909–1928
2697–2716 AF025311 8
eae-ε1 EAE-FB
LP5 AAAACCGCGGAGATGACTTC
AGCTCACTCGTAGATGACGGCAAGCG 722 66°C 1909–1928
2605–2630 AF116899 8
eae-νR/ε2e EAE-E2F
EAE-E2R AATACAGAAGTTAAGGCAT
ACGACCACTATTCATTTC 378 58°C 2230–2248
2590–2607 AF530554 9
eae-ζ Z1
Z2 GGTAAGCCGTTATCTGCC
ATAGCAAGTGGGGTGAAG 206 62°C 2062–2079
2250–2267 AF449417 8
eae-η EAE-FB
ETA-B AAAACCGCGGAGATGACTTC
TAAGCGACCACTATTCGTG 702 60°C 1899–1918
2582–2600 AJ308550 9
eae-ι1 EAE-FB
IOTA-B AAAACCGCGGAGATGACTTC
GTCATATTTAACTTTTACACTA 651 55°C 1909–1928
2538–2559 AJ308551 9
eae-μR/ι2e Iota2-F
Iota2-R CTGGTAAAGCGATAGTCAAAC
GCGTTTTTGAAGAAACATTTTGC 936 58°C 1850–1870
2763–2785 AF530553 9
eae-λ 68.4F
68.4R CGGTCAGCCTGTGAAGGGC
ATAGATGCCTCTTCCGGTATT 466 64°C 2061–2079
2506–2526 AJ715409 8
eae-μBf EAE-FB
FV373R AAAACCGCGGAGATGACTTC
ACTCATCATAATAAGCTTTTTGG 665 60°C 1909–1928
2541–2563 AJ705049 9
eae-νBf IH1229aF
IH1229aR CACAGCTTACAATTGATAACA
CTCACTATAAGTCATACGACT 311 60°C 269–289
559–579 AJ705050 8
eae-ξBf EAE-FB
B49R AAAACCGCGGAGATGACTTC
ACCACCTTTAGCAGTCAATTTG 468 66°C 1909–1928
363–384 AF116899 AJ705051 8
aUniversal oligonucleotide primer pair EAE1 and EAE-2 wih homology to the 5'conserved region of eae gene (detects all types of eae variants described at the moment). Primers used to detect the eae gene.
bThe intimin β2 of Blanco et al. (8) is identical to intimin ξ described by Ramachandran et al. (14). In this study our β2 is referred as ξR/β2B.
cThe intimin δ of Adu-Bobie et al. (4) is identical to intimin κ of Zhan et al. (6). The intimin δ of Adu-Bobie et al. (4) was also termed β2 by Oswald et al. (5). Thus, in the present study this intimin is referred as δ/κ/β2O.
dThe intimin θ of Tarr and Whittam (13) is the same as intimin γ2 of Oswald et al. (5).
eIntimins μ, ν, and ξ described by Ramachandran et al. (14) in ruminant E. coli strains. When Ramachandran et al. (14) submitted the nucleotide sequences of μ, ν, and ξ genes these intimin types were designated ι2, ε2 and β2, respectively.
fIntimins μ, ν, and ξ described by Blanco et al. (8, unpublished results).
Table 6 Sequences of oligonucleotide primers used for detection other virulence genes
Target Primer Oligonucleotide sequence (5'-3') Sequence Product size Reference
stx VT1
VT2 ATTGAGCAAAATAATTTATAT GTG
TGATGATGGCAATTCAGTAT 523 bp
520 bp 54
stx1 KS7
KS8 CCCGGATCCATGAAAAAAACATTATTAATAGC
CCCGAATTCAGCTATTCTGAGTCAACG 285 bp 55
stx2 VT2-e
VT2-f AATACATTATGGGAAAGTAATA
TAAACTGCACTTCAGCAAAT 348 bp 56
ehxA HlyA1
HlyA2 GGTGCAGCAGAAAAAGTTGTAG
TCTCGCCTGATAGTGTTTGGTA 311 bp 31
astA EAST11a
EAST11b CCATCAACACAGTATATCCGA
GGTCGCGAGTGACGGCTTTGT 111 bp 59
EAF EAF1
EAF25 CAGGGTAAAAGAAAGATGATAA
TATGGGGACCATGTATTATCA 397 bp 57
bfpA EP1
EP2 AATGGTGCTTGCGCTTGCTGC
GCCGCTTTATCCAACCTGGTA 326 bp 58
==== Refs
Nataro JP Kaper JB Diarrheagenic Escherichia coli Clin Microbiol Rev 1998 11 142 201 9457432
Frankel G Phillips AD Rosenshine I Dougan G Kaper JB Knutton S Enteropathogenic and enterohemorrhagic Escherichia coli : more subversive elements Mol Microbiol 1998 30 911 921 9988469 10.1046/j.1365-2958.1998.01144.x
McDaniel TK Kaper JB A cloned pathogenicity island from enteropathogenic Escherichia coli confers the attaching and effacing phenotype on E. coli K-12 Mol Microbiol 1997 23 399 407 9044273 10.1046/j.1365-2958.1997.2311591.x
Adu-Bobie J Frankel G Bain C Goncalves AG Trabulsi LR Douce G Knutton S Dougan G Detection of intimins alpha, beta, gamma, and delta, four intimin derivatives expressed by attaching and effacing microbial pathogens J Clin Microbiol 1998 36 662 668 9508292
Oswald E Schmidt H Morabito S Karch H Marches O Caprioli A Typing of intimin genes in human and animal enterohemorrhagic and enteropathogenic Escherichia coli :characterization of a new intimin variant Infect Immun 2000 68 64 71 10603369
Zhang WL Köhler B Oswald E Beutin L Karch H Morabito S Caprioli A Suerbaum S Schmidt H Genetic diversity of intimin genes of attaching and effacing Escherichia coli strains J Clin Microbiol 2002 40 4486 4492 12454140 10.1128/JCM.40.12.4486-4492.2002
Blanco M Blanco JE Mora A Rey J Alonso JM Hermoso M Hermoso J Alonso MP Dahbi G González EA Bernárdez MI Blanco J Serotypes, virulence genes and intimin types of Shiga toxin (Verotoxin)-producing Escherichia coli isolates from healthy sheep in Spain J Clin Microbiol 2003 41 1351 1356 12682113 10.1128/JCM.41.4.1351-1356.2003
Blanco M Blanco JE Mora A Dahbi G Alonso MP González EA Bernárdez MI Blanco J Serotypes, virulence genes and intimin types of Shiga toxin (Verotoxin)-producing Escherichia coli isolates from cattle in Spain: identification of a new intimin variant gene (eae-ξ) J Clin Microbiol 2004 42 645 651 14766831 10.1128/JCM.42.2.645-651.2004
Blanco M Podola NL Krüger A Sanz ME Blanco JE González EA Dahbi G Mora A Bernárdez MI Etcheverría AI Arroyo GH Lucchesi PMA Parma AE Blanco J Virulence genes and intimin types of Shiga-toxin-producing Escherichia coli isolated from cattle and beef products in Argentina Int Microbiol 2004 7 269 276 15666247
Fitzhenry RJ Pickard DJ Hartland EL Reece S Dougan G Phillips AD Frankel G Intimin type influences the site of human intestinal mucosal colonisation by enterohemorrhagic Escherichia coli O157:H7 Gut 2002 50 180 185 11788556 10.1136/gut.50.2.180
Jores J Zehmke K Eichberg J Rumer L Wieler LH Description of a novel intimin variant (type zeta) in the bovine O84:NM verotoxin-producing Escherichia coli strain 537/89 and the diagnostic value of intimin typing Exp Biol Med 2003 228 370 376
Torres AG Zhou X Kaper JB Adherence of diarrheagenic Escherichia coli strains to epithelial cells Infect Immun 2005 73 18 29 15618137 10.1128/IAI.73.1.18-29.2005
Tarr CL Whittam S Molecular evolution of the intimin gene in O111 clones of pathogenic Escherichia coli. J Bacteriol J Bacteriol 2002 184 479 487 11751825 10.1128/JB.184.2.479-487.2002
Ramachandran V Brett K Hornitzky MA Dowton M Beltelheim KA Walker MJ Djordjevic SP Distribution of intimin subtypes among Eschericha coli isolates from ruminant and human sources J Clin Microbiol 2003 41 5022 5032 14605134 10.1128/JCM.41.11.5022-5032.2003
Kaper JB Defining EPEC Rev Microbiol 1996 27 130 133
Trabulsi LR Keller R Tardelli Gomes TA Typical and atypical enteropathogenic Escherichia coli Emerg Infect Dis 2002 8 508 513 11996687
Beutin L Marchés O Bettelheim KA Gleier K Zimmermann S Schmidt H Oswald E HEp-2 cell adherence, actin aggregation, and intimin types of attaching and effacing Escherichia coli strains isolated from healthy infants in Germany and Australia Infect Immun 2003 71 3995 4002 12819087 10.1128/IAI.71.7.3995-4002.2003
Karmali MA Infection by verocytotoxin-producing Escherichia coli Clin Microbiol Rev 1985 2 5 38
Paton JC Paton AW Pathogenesis and diagnosis of Shiga toxin-producing E. coli infections Clin Microbiol Rev 1998 11 450 479 9665978
Blanco J Blanco M Blanco JE Mora A Alonso MP González EA Bernárdez MI Duffy G, Garvey P, McDowell D Epidemiology of verocytotoxigenic Escherichia coli (VTEC) in ruminants 2001 Vercytotoxigenic Escherichia coli, Food and Nutrition Press, Inc., Trumbull, Conn 113 148
Banatvala N Griffin PM Greene KD Barrett TJ Bibb WF Green JH Wells JG (Hemolytic Uremic SyndromeStudy Collaborators) The United States National Prospective Hemolytic Uremic Syndrome Study: microbiologic, serologic, clinical, and epidemiologic findings J Infect Dis 2001 183 1063 1070 11237831 10.1086/319269
Chapman PA Cerdan Malo AT Ellin M Ashton R Harkin MA Escherichia coli O157 in cattle and sheep at slaughter, on beef and lamb carcasses and in raw beef and lamb products in South Yorkshire, UK Int J Food Microbiol 2001 64 139 150 11252496 10.1016/S0168-1605(00)00453-0
Eklund M Scheutz F Siitonen A Clinical isolates of non-O157 Shiga toxin-producing Escherichia coli : serotypes, virulence characteristics, and molecular profiles of strains of the same serotype J Clin Microbiol 2001 39 2829 2834 11473999 10.1128/JCM.39.8.2829-2834.2001
Park CH Kim HJ Dixon DL Importance of testing stool specimens for Shiga toxins J Clin Microbiol 2002 40 3542 3543 12202617 10.1128/JCM.40.9.3542-3543.2002
Welinder-Olsson C Badenfors M Cheasty T Kjellin E Kaijser B Genetic profiling of enterohemorrhagic Escherichia coli strains in relation to clonality and clinical signs of infection J Clin Microbiol 2002 40 959 964 11880423 10.1128/JCM.40.3.959-954.2002
Beutin L Krause G Zimmermann S Kaulfuss S Gleier K Characterization of Shiga toxin-producing Escherichia coli strains isolated from human patients in Germany over a 3-year period J Clin Microbiol 2004 42 1099 1108 15004060 10.1128/JCM.42.3.1099-1108.2004
Blanco JE Blanco M Alonso MP Mora A Dahbi G Coira MA Blanco J Serotypes, virulence genes, and intimin types of Shiga toxin (verotoxin-) producing Escherichia coli isolates from human patients: Prevalence in Lugo, Spain, from 1992 through 1999 J Clin Microbiol 2004 42 311 319 14715771 10.1128/JCM.42.1.311-319.2004
Lopez EI Contrini MM De Rosa MF Kaper JB, O'Brien AD Epidemiology of Shiga toxin-producing Escherichia coli in South America Escherichia coli O157:H7 and other Shiga toxin-producing E coli strains 1998 Asm Press, Washington D.C 30 37
Elliott EJ Robins-Browne RM O'Loughlin EV Bennett-Wood V Bourke J Henning P Hogg GG Knight J Powell H Redmond D (Contributors to the Australian Paediatric Surveillance Unit) Nationwide study of hemolytic uremic syndrome: clinical, microbiological, and epidemiological features Arch Dis Child 2001 85 125 131 11466187 10.1136/adc.85.2.125
Beutin L Geier D Zimmermann S Scheutz F Virulence markers of Shiga-like toxin-producing Escherichia coli strains originating from healthy domestic animals of different species J Clin Microbiol 1995 33 631 635 7538509
Schmidt H Beutin L Karch H Molecular analysis of the plasmid-encoded hemolysin of Escherichia coli O157:H7 strain EDL 933 Infect Immun 1995 63 1055 1061 7868227
Rey J Blanco JE Blanco M Mora A Dhabi G Alonso JM Hermoso M Hermoso J Alonso MP Usera MA González EA Bernárdez MI Blanco J Serotypes, phage types and virulence genes of Shiga-producing Escherichia coli isolated from sheep in Spain Vet Microbiol 2003 94 47 56 12742715 10.1016/S0378-1135(03)00064-6
Stephan R Borel N Zweifel C Blanco M Blanco JE First isolation and further characterization of enteropathogenic Escherichia coli (EPEC) O157:H45 strains from cattle BMC Microbiol 2004 4 10 15113396 10.1186/1471-2180-4-10
Kobayashi H Miura A Hayashi H Ogawa T Endo T Hata E Eguchi M Yamamoto K Prevalence and characteristics of eae -positive Escherichia coli from healthy cattle in Japan Appl Environ Microbiol 2003 69 5690 5692 12957963 10.1128/AEM.69.9.5690-5692.2003
Johnson RP Clarke RC Wilson JB Read S Rahn K Renwick SA Sandhu KA Alves D Karmali MA Lior H Mcewen SA Spika JS Gyles CL Growing concerns and recent outbreaks involving non-O157 serotypes of verotoxigenic Escherichia coli J Food Prot 1996 59 1112 1122
Parma AE Sanz ME Blanco JE Blanco J Vinas MR Blanco M Padola NL Etcheverria AI Virulence genotypes and serotypes of verotoxigenic Escherichia coli isolated from cattle and foods in Argentina. Importance in public health Eur J Epidemiol 2000 16 757 762 11142505 10.1023/A:1026746016896
Pradel N Livrelli V De Champs C Palcoux JB Reynaud A Scheutz F Sirot J Joly B Forestier C Prevalence and characterization of Shiga toxin-producing Escherichia coli isolated from cattle, food, and children during a one-year prospective study in France J Clin Microbiol 2000 38 1023 1031 10698990
Padola NL Sanz ME Blanco JE Blanco M Blanco J Etcheverría AI Arroyo GH Usera MA Parma AE Serotypes and virulence genes of bovine Shigatoxigenic Escherichia coli (STEC) isolated from a feedlot in Argentina Vet Microbiol 2004 100 3 9 15135507 10.1016/S0378-1135(03)00127-5
McGraw EA Li J Selander RK Whittam TS Molecular evolution and mosaic structure of alpha, beta, and gamma intimins of pathogenic Escherichia coli Mol Biol Evol 1999 16 12 22 10331248
Reid SD Betting DJ Whittam TS Molecular detection and identification of intimin alleles in pathogenic Escherichia coli by multiplex PCR J Clin Microbiol 1999 37 2719 2722 10405431
Orden JA Yuste M Cid D Piacesi T Martinez S Ruiz-Santa-Quiteria JA De la Fuente R Typing of the eae and espB genes of attaching and effacing Escherichia coli isolates from ruminants Vet Microbiol 2003 96 203 215 14519337 10.1016/S0378-1135(03)00238-4
Aktan I Sprigings KA La Ragione RM Faulkner LM Paiba GA Woodward MJ Characterization of attaching-effacing Escherichia coli isolated from animals at slaughter in England and Wales Vet Microbiol 2004 102 43 53 15288926 10.1016/j.vetmic.2004.04.013
Zweifel C Blanco JE Blanco M Blanco J Stephan R Serotypes and virulence genes of ovine non-O157 Shiga toxin-producing Escherichia coli in Switzerland Int J Food Microbiol 2004 95 19 27 15240071 10.1016/j.ijfoodmicro.2004.01.015
China B Jacquemin E Dervin A-C Pirson V Mainil J Heterogeneity of the eae genes in attaching and effacing Escherichia coli from cattle: comparison with human strains Res Microbiol 1999 150 232 332 10.1016/S0923-2508(99)80058-8
Tarr PI Fouser LS Stapleton AE Wilson RA Kim HH Vary JC JrClausen CR Hemolytic-uremic syndrome in a six-year-old girl after a urinary tract infection with Shiga-toxin-producing Escherichia coli O103:H2 N Engl J Med 1996 335 635 638 8687518 10.1056/NEJM199608293350905
Feng P Lampel KA Karch H Whittam TS Genotypic and phenotypic changes in the emergence of Escherichia coli O157:H7 J Infect Dis 1998 177 1750 1753 9607864
Blanco M Blanco JE Blanco J Carvalho VM Lopes Onuma D Pestana de Castro AF Typing of intimin (eae) genes in attaching and effacing Escherichia coli strains from Monkeys J Clin Microbiol 2004 42 1382 1383 15004128 10.1128/JCM.42.3.1382-1383.2004
Yatsuyanagi J Saito S Miyajima Y Amano KI Enomoto K Characterization of atypical enteropathogenic Escherichia coli strains harboring the gene astAthat were associated with a waterborne outbreak of diarrhes in Japan J Clin Microbiol 2003 41 2033 2039 12734245 10.1128/JCM.41.5.2033-2039.2003
Piérard D Stevens D Moriau L Lior H Lauwers S Isolation and virulence factors of verotoxin-producing E. coli in human stool samples Clin Microbiol Infect 1997 3 531 540 11864177
Boerlin P McEwen SA Boerlin-Petzold F Wilson JB Johnson RP Gyles CL Association between virulence factors of Shiga toxin-producing Escherichia coli and disease in humans J Clin Microbiol 1999 37 497 503 9986802
Friedrich AW Bielaszewska M Zhang WL Pulz M Kuczius T Ammon A Karch H Escherichia coli harboring Shiga toxin 2 gene variants: frequency and association with clinical symptoms J Infect Dis 2002 185 74 84 11756984 10.1086/338115
Zweifel C Schumacher S Blanco M Blanco JE Tasara T Blanco J Stephan R Phenotypic and genotypic characteristics of non-O157 Shiga toxin-producing Escherichia coli (STEC) from Swiss cattle Vet Microbiol 2005 105 37 45 15607082 10.1016/j.vetmic.2004.10.007
Guinée PA Jansen WH Wadström T Sellwood R Escherichia coli associated with neonatal diarrhea in piglets and calves Curr Top Vet Anim Sci 1981 13 126 162
Burnens AP Frey A Lior H Nicolet J Prevalence and clinical significance of vero-cytotoxin-producing Escherichia coli (VTEC) isolated from cattle in herds with and without calf diarrhea J Vet Med B 1995 42 311 318
Rüssmann H Kothe E Schmidt H Franke S Harmsen D Caprioli A Karch H Genotyping of Shiga-like toxin genes in non-O157 Escherichia coli strains associated with hemolytic uremic syndrome J Med Microbiol 1995 42 404 410 7791204
Piérard D Muyldermans G Moriau L Stevens D Lauwers S Identification of new verocytotoxin type 2 variant B-subunit genes in human and animal Escherichia coli isolates J Clin Microbiol 1998 36 3317 3322 9774585
Franke J Granke S Schmidt H Schwarzkopf A Wieler LH Baljer G Beutin L Karch H Nucleotide sequence analysis of enteropathogenic Escherichia coli (EPEC) adherence factor probe and development of PCR for rapid detection of EPEC harboring virulence plasmids J Clin Microbiol 1994 32 2460 2463 7814482
Gunzburg ST Tornieporth NG Riley LW Identification of enteropathogenic Escherichia coli by PCR-based detection of the bundle-forming pilus gene J Clin Microbiol 1995 33 1375 1379 7615759
Yamamoto T Nakazawa M Detection and sequences of the enteroaggregative Escherichia coli heat-stable enterotoxin 1 gene in enterotoxigenic E. coli strains isolated from piglets and calves with diarrhea J Clin Microbiol 1997 35 223 227 8968912
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
| 15882459 | PMC1142320 | CC BY | 2021-01-04 16:03:39 | no | BMC Microbiol. 2005 May 9; 5:23 | utf-8 | BMC Microbiol | 2,005 | 10.1186/1471-2180-5-23 | oa_comm |
==== Front
BMC Med Inform Decis MakBMC Medical Informatics and Decision Making1472-6947BioMed Central London 1472-6947-5-131587635210.1186/1472-6947-5-13Research ArticleA controlled trial of automated classification of negation from clinical notes Elkin Peter L [email protected] Steven H [email protected] Brent A [email protected] Casey S [email protected] William [email protected] Larry R [email protected] Dietlind L [email protected] Department of Medicine, Mayo Foundation, Rochester, MN, USA2 Department of Biomedical Informatics Vanderbilt University, Nashville, TN and the Veterans Health Administration (VHA), USA3 Department of Internal Medicine, Johns Hopkins School of Medicine, Baltimore, MD, USA2005 5 5 2005 5 13 13 18 8 2004 5 5 2005 Copyright © 2005 Elkin et al; licensee BioMed Central Ltd.2005Elkin 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
Identification of negation in electronic health records is essential if we are to understand the computable meaning of the records: Our objective is to compare the accuracy of an automated mechanism for assignment of Negation to clinical concepts within a compositional expression with Human Assigned Negation. Also to perform a failure analysis to identify the causes of poorly identified negation (i.e. Missed Conceptual Representation, Inaccurate Conceptual Representation, Missed Negation, Inaccurate identification of Negation).
Methods
41 Clinical Documents (Medical Evaluations; sometimes outside of Mayo these are referred to as History and Physical Examinations) were parsed using the Mayo Vocabulary Server Parsing Engine. SNOMED-CT™ was used to provide concept coverage for the clinical concepts in the record. These records resulted in identification of Concepts and textual clues to Negation. These records were reviewed by an independent medical terminologist, and the results were tallied in a spreadsheet. Where questions on the review arose Internal Medicine Faculty were employed to make a final determination.
Results
SNOMED-CT was used to provide concept coverage of the 14,792 Concepts in 41 Health Records from John's Hopkins University. Of these, 1,823 Concepts were identified as negative by Human review. The sensitivity (Recall) of the assignment of negation was 97.2% (p < 0.001, Pearson Chi-Square test; when compared to a coin flip). The specificity of assignment of negation was 98.8%. The positive likelihood ratio of the negation was 81. The positive predictive value (Precision) was 91.2%
Conclusion
Automated assignment of negation to concepts identified in health records based on review of the text is feasible and practical. Lexical assignment of negation is a good test of true Negativity as judged by the high sensitivity, specificity and positive likelihood ratio of the test. SNOMED-CT had overall coverage of 88.7% of the concepts being negated.
==== Body
Background
A great wealth of patient specific medical data is stored as transcribed free text. While this format is useful for individuals reading the medical record, information stored as free-text is difficult to use in decision support systems or automated cross population studies [1]. Efforts to extract computer-usable information from free text archives vary widely. Traditionally, teams of trained abstractors have manually reviewed patients' charts. String matching is a simple algorithmic approach. Identifying concepts is a much more complex process. Algorithmic natural language understanding holds great promise, but remains difficult to achieve [2,3]. Despite the challenges, a number of groups have applied natural language processing techniques with varying degrees of success [4-10]. Concept-based indexing is another approach that has been applied to a number of areas including literature retrieval, health related web sites, clinical diagnoses, and medical narratives [11-16].
Natural language processing is routed in a logical representation of discourse. Until the 1920s logic and mathematics was considered spiritual not scientific. Since the time of Pythagoras, mathematics was considered a revelation of the divine order. In Principia Mathematica (Russell and Whitehead), demonstrated that mathematics was logical. Logical positivism was then applied to science and psychology.
Noam Chomsky's seminal work "The Logic Structure of Linguistic Theory," was published in 1955 in mimeograph form and in press in 1975. This work expressed the view that language was a cognitive activity and required a meta-model of language to effectively communicate. He demonstrated that the stimulus response model could not account for human language. This idea that language is processed led to the application of computer science to free text (natural language) processing. Computational linguistics (CL) is a field of computer science which seeks to understand and represent language in an interoperable set of semantics. CL overlaps with the field of Artificial Intelligence and has been often applied to machine translation from one human language to another. Naomi Sager in 1994 published in JAMIA a paper entitled "Natural Language Processing and the Representation of Clinical Data." Here Dr. Sager showed that for a set of discharge letters a recall of 92.5% and a precision of 98.6% could be achieved for a limited set of pre-selected data using the parser produced by the Linguistic String Project at New York University [1-3].
In 2004, Friedman et al reported a method for encoding concepts from health records using the UMLS [4]. In this study Dr. Friedman and colleagues used MedLEE to abstract concepts from the record and reported a recall of 77% and a precision of 89%. In 2001, Nadkarni provided a description of the fundamental building blocks needed for NLP [5]. He discussed their method for lexical matching and part of speech tagging in discharge summaries and surgical notes. Henry Lowe developed MicroMeSH an early MUMPS based terminology browser which incorporated robust lexical matching routines. Dr. Lowe working with Bill Hersh reported the accuracy of parsing radiology reports using the Sapphire indexing system [6]. Here they reported good sensitivity and they were able to improve performance by limiting the UMLS source vocabularies by section of the report.
MetaMap has the capacity to be used to code free text (natural language) to a controlled representation which can be any subset of the UMLS knowledge sources [7]. MetaMap uses a five step process which begins by using the SPECIALIST minimal commitment parser which identifies noun phrases without modifiers. The next step involves the identification of phrase variants. These variants are then used to identify candidate phrases from within the source material [8]. Linguistic principals are used to calculate a score for each potential match. Brennan and Aronson used MetaMap to improve consumer health information retrieval for patient [9].
We have built and described systems for concept based indexing, automated term composition, and automated term decomposition. In its current version, the system uses the SNOMED-CT terminology. The accuracy of this automated technique has previously been evaluated [10]. Many individuals have evaluated the accuracy of manual term composition [11,12]. The clinical coding center of the NHS has reported limited success with their own algorithm for automated term dissection in the past [13,14].
As we move toward compositional terminologies, the need to organize the terms within a compositional expression becomes important for both the readability and understanding of these composite terms [15,17]. Identifying concepts that are explicitly asserted as not being the case and separating them from positive assertions becomes of critical importance if we are to understand the implications of medical text. Linguistic negation is a challenging problem [18]. This trial evaluates a mechanism for automated assignment of negation status to concepts parsed from the terminology using a negation ontology. The text is analyzed to identify expressions indicating negation and a model of negation is applied to assign values to concepts. We have named this system the automated negation assignment grammar [10]. We recognize the following semantic types: Kernel concepts, Modifiers, Qualifiers or Negative Qualifiers [19]. A rule base is then applied which organizes the Modifiers, Qualifiers and Negative Qualifiers around the Kernel concepts. These are represented in a hierarchical structure with the degree of indentation being representative of semantic dependency. The accuracy of this automated technique has previously been evaluated [10]. Many individuals have evaluated the accuracy of manual term composition [11,12].
Identifying concepts that are explicitly asserted negatively (e.g. "no evidence of pneumonia") and separating them from positive assertions becomes of critical importance if we are to understand the implications of medical text.
To illustrate the importance of concept negation, we reference a case of a 62 year old female who presents with erythema over the dorsum of the left foot with exquisite tenderness over a wound situated over the mid foot. After a comprehensive clinical work up, she was found to have a Cellulitis of the left foot without signs of lymphangitic spread of her infection. In this case, it is an important distinction that our patient did not have "Lymphangitis" associated with her "Cellulitis, left foot," as opposed to a distinct separate case where the diagnosis of "Lymphangitis was present." Epidemiologically, if one was studying Lymphangitis, it would be important to exclude this patient's record from the analysis.
A previous study of Negation by Mutalik et al, described the lexical assignment of negation using the UMLS to code free text documents. Their intervention had a sensitivity of 95.7% and a specificity of 91.8% [20]. They did not report the UMLS coverage of the concepts that appeared in the text. They also noted that the words "no", "not" "denied/denies" and "without" made up 92.5% of the negation in their study. Chapman et al looked to identify negation in discharge summaries and identified negative UMLS concepts with a sensitivity of 77.8% and a specificity of 94.5% using regular expressions [21].
Methods
Study Design
Forty-one unique clinical records (which were notes comprising a set of history and physical medical evaluations) of forty-one separate patients were randomly selected from the outpatient section of the Department of Internal Medicine at John's Hopkins Medical School. These were indexed using the Mayo Vocabulary Server. The records are presented to the system as free-text ASCII files. The text is parsed using the Mayo Health Record Parser, which parses the text into sections consistent with the usual health record as presented below:
History
History of Present Illness
By Problem
Past Medical History
Social History
Medications
Allergies
Review of Systems
Physical Examination
By Body Part
Diagnostic Testing
Assessment / Report / Plan
By Problem
Negation Assignment
Negation is part of a larger system that assigns to concepts a level of certainty as part of the generation of a two-phase dyadic parse tree. Each sentence within each section is parsed first by a preprocessor, which breaks the input into text and operators (i.e. And, Or, Not, Maybe). The text is parsed using the Mayo Vocabulary Server, which returns a set of concepts representing the best match from within SNOMED-CT to the sentence fragment parsed. A rule base is applied to the text that assigns to each concept an attribute stating that the concept is a positive assertion, a negative assertion or an uncertain assertion. The software is not yet developed to a point where it could be used by other users, and thus is not freely available, but the authors will provide access to the software to any readers interested in validating the results. In this manuscript, we are focusing on the evaluation of the assignment to concepts the attribute "negative assertion" (see Figure 1). Mixed assertions such as "probably not", were considered uncertain assertions for the purposes of this evaluation.
Figure 1 Fictitious example of the assignment of negation for the clinical statement "Mr. Jones has a family history of carcinoma of the prostate, but has no family history of colon cancer." The statement "No" family history of cancer of colon represents a negative assertion.
The system uses the SNOMED-CT terminology to index clinical documents. The entire terminology was employed so that any SNOMED-CT concept appearing in these histories, regardless of their location within the ontology, would be codified. The text is analyzed to identify expressions indicating negation and a model of negation is applied to assign values to concepts. We call this the automated negation assignment grammar.
Examples of the types of terms which imply negation are "no", "denies", and has "ruled out." An example of terms, which stop the propagation of the assignment of negation, is "other than" (e.g. The patient denied a history of previous cardiac disease other than palpitations which he experienced while giving a presentation resulting in syncope.). In the previous example, positive assertions are in italics and negative assertions are in bold and the operators are underlined. The larger ontology of negation terms their lexical variants and the associated rules for applying them are being made available within a larger vocabulary server named the Mayo Vocabulary Server.
The results of the parsing were reviewed by an expert medical terminologist (independent from the study team). We chose to employ a single reviewer in this study who was independent from the study team, based on the high inter-rater agreement found in our recent evaluation of SNOMED-CT for the coding of problem list data (94.3% agreement; Kappa Statistic = 0.79) and our belief that the assignment of negation was relatively straightforward when compared with most health terminology judgments. For each occurrence of negation in the text as judged by a human reviewer, the number of correctly mapped negative concepts and the total number of negative concepts were tallied by subsection of the record. If the assertions were not tagged correctly, a record was made to distinguish whether there was a failure to map to the terminology, or if the engine simply mapped the term incorrectly. The failure analysis also included whether the missing information was a Kernel concept (the main point of the expression), a Modifier (a concept that changes the meaning of a term in a clinical sense like "severity"), a Qualifier (a concept that changes the meaning of a term in an administrative or temporal sense like "recurrent").
The sensitivity, specificity and positive predictive value and positive likelihood ratios for the assignment of the negation status are reported.
For an example of an in context negation assertion as found in a medical record, see figure 1.
Statistical Analysis
In addition to descriptive comparisons of the accuracy rates for the assignment of negation, a formal statistical comparison was performed. To determine if the results we obtained could be accounted for by chance alone the following method was employed. Testing was employed for equality of accuracy rates between the negation assignment and a coin flip to determine if the effect, which we have seen, could have been present from chance alone by the Pearson Chi-square statistic for equality of proportions.
Results
Overall, we identified 14,792 health concepts in the text of 41 clinical records using SNOMED-CT the Mayo Vocabulary Server Parsing software and the manual review. No attempt was made to filter out duplicate concepts. 13,082 of these were positive or uncertain assertions of which 12,921 were recognized by the parser as such. Out of the 2,028 negative concepts, 205 Concepts were not mapped by SNOMED-CT, but were identified by the human reviewer. Of the remaining 1,823 concepts the engine identified 1662 correctly (p < 0.001, Pearson Chi-Square test; when compared with a coin flip). One-hundred and sixty-one concepts were incorrectly assigned as negatives and another 48 were assigned incorrectly as positives. This resulted in the two-by-two table shown in Table 1. The sensitivity (Recall) of the assignment of negation was 97.2% (range 50% to 100%) and the specificity of that same assignment was 98.8% (range 33.33% to 100%). The positive likelihood ratio for the effect was 81, which indicates that our method is a good test for identifying negation. The reliability of this test as judged by the positive likelihood ratio compares favorably with other acceptable medical diagnostic tests such as the dobutamine stress echo that has a positive likelihood ratio of 24. The negative predictive value of the assignment was 99.6%. The positive predictive value (Precision) was 91.2%.
Table 1 Two by Two table of the results of the negation mapping compared with the initial human review. The "C" prior to the operator signifies Computer identification (e.g. Cnegative). True negation is based on the human review that served as our gold standard.
True Negation True (Positive or Uncertain)
CNegative 1662 161
Cpositive Or Cuncertain 48 12921
In Table 1 we show the results of the human review as compared with the automated assignment of Negation. As all concepts were coded as Positive, Uncertain or Negative assertions, we combined the Positive and Uncertain assertions for the purposes of this analysis. True negation was the rate of negation identified by the human reviewer from within the text. True positives or uncertain were the other concepts, which were encoded by the automated engine and tagged as either being any of the three types of assertions. "Cneg" is the computer-generated rate of assignment of negation and "Cpos" and "C?" are the rates of assignment of positivity or uncertainty.
The failure analysis showed that many of the concepts which were assigned as positive which should have been negative, were words such as "nontender" and "colorless" which were missed by our algorithm. Another class of problems stemmed from operators which appeared to the engine to be double negatives such as "but not" or "but it never."
In Table 2, the sensitivity and specificity of the negation routine varied throughout the different sections of the health record. Incalculable values were created when the true positives plus false negatives (TP + FN) were zero or when the true negatives plus the false positives (TN + FP) were zero (otherwise the equation for sensitivity or specificity would require division of the numerator by zero)
Table 2 The sensitivity and specificity of the negation routine by area of the record.
section subsection avg_sens avg_spec
History -- 92.30 80.0
History Allergies 100.0 Div 0
History Current meds Div 0 Div 0
History Family history 89.19 57.14
History HPI 94.71 64.44
History Medications 50.0 100.0
History PMH 94.12 10.0
History Social history 98.91 66.67
History Systems rev 99.43 88.89
Exam abdomen 98.39 50.0
Exam breasts 100.0 Div 0
Exam ENT 96.52 33.33
Exam Extremities 98.67 Div 0
Exam eyes 95.0 50.0
Exam gait 100.0 Div 0
Exam genitalia 100.0 Div 0
Exam head Div 0 Div 0
Exam heart 100.0 100.0
Exam lungs 100.0 Div 0
Exam lymph 100.0 Div 0
Exam musculo 100.0 Div 0
Exam neuro 100.0 Div 0
Exam rectum 100.0 100.0
Exam skin 100.0 Div 0
Exam thyroid 100.0 Div 0
Exam vitals 91.11 Div 0
Diagnostic Testing -- 100.0 100.0
Assessment -- 87.78 69.01
Assessment Addendum 50.0 71.43
Assessment Plan 70.0 Div 0
Assessment Problem/diagnosis 64.71 66.67
Discussion
The assignment of negation to concepts from a controlled terminology such as SNOMED-CT can be automatically assigned reliably. The assignment had a high positive likelihood ratio indicating that it is overall an accurate test of the records for this condition. The most common reason for failure was the inability of SNOMED-CT to represent the negative concepts. This was verified by browsing the terminology for the concept as well as by inability of the mapping engine to identify a correct match. The predictive value of the assignment of negation was highly statistically significant when compared with a coin flip (chance alone) with a p < 0.001.
The failure analysis identified unexpected methods of negation, which we are dealing with in a second generation of the software that handles roots and stems. We have also created an ontology of terms that start negation and another set which stop the propagation of the assignment of negation. Clearly there was variability in the accuracy of the algorithm across the various sections of the clinical record. This provides the Informatics research community with an opportunity to identify areas of focus for future research efforts.
We extend the work of Mutalik et al and Chapman et al by performing this study using SNOMED-CT and by utilizing a second independently developed ontology for negation. Also we used full medical evaluations in our study, which had a higher per case percentage of negative concepts as compared with the surgical reports and discharge summaries used in studies by previous authors. Previous unpublished usability data in our laboratory noted that clinicians require 95% accuracy for acceptance of a system that provides conceptual coding of clinical content. Therefore the negation algorithms should be an acceptable starting point for clinical applications for the physical examination, the HPI, social history, allergies and review of systems sections of the clinical record. More work is needed to understand the negation requirements of medications, family history, the vital signs, and the assessment section of the clinical record.
Compositional terminologies are one promising answer to the problem of clinical content completeness [22]. High-quality controlled health vocabularies provide a gateway to improved clinical data availability for outcomes research, utilization review, and improved management of the electronic medical record [23]. This promise is contingent upon data entry mechanisms, which will not disrupt the flow of a busy practice [24].
Creating well formed compositional expressions using a controlled health vocabulary can be labor intensive and time consuming. Given the ever-increasing demands on clinicians' time, we must work to create mechanisms, which aid the busy clinician as we migrate toward, an electronic clinical environment. Missed negation can lead to excess testing which in turn can lead to an increased rate of medical error. Likewise erroneous assignment of negation can lead to missing allergies and other important health data that can negatively impact patient safety. Automated tools designed to assist clinicians with the formulation of compositional expressions are necessary if we are to make use of powerful compositional terminologies.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors worked on the design of the study and reviewed the manuscript. PE, BB and DWR supervised the actual running of the experiment. All authors participated in the analysis of the results.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors wish to thank Karen Lee and Diana Monty for administrative support, and Raju Kalidindi for programming support. This work is supported in part by an NIH NLM grant (LM06918-A103).
==== Refs
Sager N Syntactic analysis of natural language. Advances in Computers 1967 8 New York, Academic Press 153 188
Grishman R Sager N Raze C Bookchin B The linguistic string parser.: ; Montvail, NJ. 1973 42 AFIPS Press 427 434 4730560
Sager N Grishman R The restriction language for computer grammars of natural language. Communications of the ACM 1975 18 390 400 10.1145/360881.360910
Friedman C Shagina L Lussier Y Hripcsak G Automated encoding of clinical documents based on natural language processing J Am Med Inform Assoc 2004 11 392 402 15187068 10.1197/jamia.M1552
Nadkarni P Chen R Brandt C UMLS concept indexing for production databases: a feasibility study J Am Med Inform Assoc 2001 8 80 91 11141514
Huang Y Lowe HJ Hersh WR A pilot study of contextual UMLS indexing to improve the precision of concept-based representation in XML-structured clinical radiology reports J Am Med Inform Assoc 2003 10 580 587 12925544 10.1197/jamia.M1369
Aronson AR Bodenreider O Chang HF Humphrey SM Mork JG Nelson SJ The NLM indexing initiative 2000 17 21 11079836
Aronson AR Effective mapping of biomedical text to the UMLS metathesaurus: the MetaMap program. 2001 17 21 11825149
Brennan PF Aronson AR Towards linking patients and clinical information: detecting UMLS concepts in e-mail J Biomed Inform 2003 36 334 341 14643729 10.1016/j.jbi.2003.09.017
Elkin PL Bailey KR Chute CG A randomized controlled trial of automated term composition Proc AMIA Symp 1998 765 769 9929322
Rector AL Thesauri and formal classifications: terminologies for people and machines Methods Inf Med 1998 37 501 509 9865048
Bernauer J Franz M Schoop D Schoop M Pretschner DP The compositional approach for representing medical concept systems Medinfo 1995 8 Pt 1 70 74 8591303
Price C Bentley TE Brown PJ Schulz EB O'Neil M Anatomical characterisation of surgical procedures in the Read Thesaurus Proc AMIA Annu Fall Symp 1996 110 114 8947638
Schulz EB Barrett JW Price C Read Code quality assurance: from simple syntax to semantic stability J Am Med Inform Assoc 1998 5 337 346 9670131
Rassinoux AM Miller RA Baud RH Scherrer JR Compositional and enumerative designs for medical language representation Proc AMIA Annu Fall Symp 1997 620 624 9357700
Elkin PL Ruggieri A Bergstrom L Bauer BA Lee M Ogren PV Chute CG A randomized controlled trial of concept based indexing of Web page content Proc AMIA Symp 2000 220 224 11079877
Schulz EB Price C Brown PJ Symbolic anatomic knowledge representation in the Read Codes version 3: structure and application J Am Med Inform Assoc 1997 4 38 48 8988473
Pacholczyk D Quafafou M L. G Optimistic vs. pessimistic interpretation of linguistic negation 2002 2443 Lec Notes Artif Int, 132 141
Chute CG Elkin PL A clinically derived terminology: qualification to reduction Proc AMIA Annu Fall Symp 1997 570 574 9357690
Mutalik PG Deshpande A Nadkarni PM Use of general-purpose negation detection to augment concept indexing of medical documents: a quantitative study using the UMLS J Am Med Inform Assoc 2001 8 598 609 11687566
Chapman WW Bridewell W Hanbury P Cooper GF Buchanan BG A simple algorithm for identifying negated findings and diseases in discharge summaries J Biomed Inform 2001 34 301 310 12123149 10.1006/jbin.2001.1029
Rogers JE Rector AL Terminological systems: bridging the generation gap Proc AMIA Annu Fall Symp 1997 610 614 9357698
Musen MA Wieckert KE Miller ET Campbell KE Fagan LM Development of a controlled medical terminology: knowledge acquisition and knowledge representation Methods Inf Med 1995 34 85 95 9082143
Rector AL Glowinski AJ Nowlan WA Rossi-Mori A Medical-concept models and medical records: an approach based on GALEN and PEN&PAD J Am Med Inform Assoc 1995 2 19 35 7895133
| 15876352 | PMC1142321 | CC BY | 2021-01-04 23:52:12 | no | BMC Med Inform Decis Mak. 2005 May 5; 5:13 | utf-8 | BMC Med Inform Decis Mak | 2,005 | 10.1186/1472-6947-5-13 | oa_comm |
==== Front
BMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 1471-2288-5-161587634810.1186/1471-2288-5-16Research ArticleEarly life factors, childhood cognition and postal questionnaire response rate in middle age: the Aberdeen Children of the 1950s study Nishiwaki Yuji [email protected] Heather [email protected] Susan M [email protected] David A [email protected] Department of Epidemiology and Population Health, London School of Hygiene & Tropical Medicine, London, UK2 Dugald Baird Centre, University of Aberdeen, Aberdeen, UK3 Liggins Institute and School of Population Health, University of Auckland, Auckland, New Zealand2005 5 5 2005 5 16 16 7 1 2005 5 5 2005 Copyright © 2005 Nishiwaki et al; licensee BioMed Central Ltd.2005Nishiwaki 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
Little is known about the relationship between early life factors and survey response in epidemiological studies of adults.
Methods
The Children of the 1950s cohort is composed of 12,150 children (boys 51.7%) born in Aberdeen 1950–56 and in primary schools in the city in 1962. Information on birth weight, gestational age, growth, behaviour and socio-economic position at birth and in childhood were obtained from contemporaneous records. Cognitive test scores at ages 7,9 and 11 years were also available from school records. The outcome was response to a postal questionnaire sent (2001–2003) to surviving cohort members in middle age.
Results
Of 11,282 potentially mailed subjects, 7,183 (63.7%) returned questionnaires. Response rates were highest among females, and those whose parents were married at birth, were in a non-manual social class at birth or in childhood, had fewer siblings, were taller and heavier in childhood for their age and had lower Rutter B behavioural scores. Childhood cognitive test scores at every age were strongly and positively related to the response rate to a postal questionnaire independently of other early life factors monotonically across the entire range of test scores. Those in the bottom fifth at age 11 had a response rate of 49% while those in the top fifth 75%.
Conclusion
The strength and consistency of the association of childhood cognition with questionnaire response rate in middle age is surprisingly large. It suggests that childhood cognition across the entire normal range is a powerful influence on the complex set of later behaviours that comprise questionnaire response. The extent of possible response bias in epidemiological studies of the associations between childhood characteristics (particularly those related to cognition) and later health is probably larger than is generally realised, at least in situations where the survey instrument is a postal questionnaire.
==== Body
Background
The potential bias introduced by non-response or non-participation is an issue that affects much of observational epidemiology. Extending our understanding of the factors that influence non-response is thus an important area of research in its own right [1,2]. It is well known that many characteristics measured in adulthood such as gender, age, socio-economic position, education, health status and smoking habits are associated with response rate [2-10]. However, very little is known about the association of characteristics in infancy and childhood with response or recruitment rates in adulthood. This is partly because there are relatively few contexts in which unbiased data on early life characteristics are available for the total population irrespective of later response behaviour. The increasing interest in life-course epidemiology [11,12], in which adult disease is related to factors across an individual's life course, makes an investigation of these early-life influences on response rate particularly pertinent at this time.
In this paper we look at the influence of a range of infant and childhood characteristics on adult questionnaire response rates in a large cohort study. We pay particular attention to childhood cognition as we are also interested in the broader question of why adult health is related to childhood cognition [13-15]. A priori one might expect childhood cognition to be related to response rates, partly through its link with educational attainment, that is known to be strongly associated with response rates. However, to date there have been very few studies of the association of factors in early life, including cognition, and questionnaire or survey response in adult life. In an analysis of the UK 1946 national birth cohort, Wadsworth et al. reported that lower cognitive score at 8 years of age was associated with previous refusal or failure to contact by the study nurse at age 53 years [16]. However, the findings from this paper may not be readily generalizable, as the study is unusual in that it involved repeated contact with participants over their entire life time. More typically, life-course studies have information from infancy and childhood from historic sources, and then attempt a first survey follow-up well into adult life.
We have used the Aberdeen Children of the 1950s historical cohort study to investigate the association between early life factors, including childhood cognition, and response to a postal questionnaire conducted in middle age that represented the first attempt to make direct contact with the study subjects since childhood.
Methods
The Aberdeen Children of the 1950s study
The background to this historical cohort study has been previously described [17]. Briefly, this cohort is a large subset of the Aberdeen Child Development Study (ACDS) [18]. This was conducted in the early 1960s to estimate the population prevalence of mental subnormality in children and to investigate its etiology. Aberdeen was chosen as both obstetric and educational records were of a high standard and the population was relatively stable.
The ACDS was comprised of all 15 thousand children born 1950–56 who were in primary school in Aberdeen, Scotland in December 1962. In order for the ACDS to achieve its main objectives, particular care was taken by the original investigators to include all children of primary school age in Aberdeen regardless of their cognitive function. Thus children in schools for individuals with moderate or severe learning difficulties were included as well as those in ordinary primary schools. However, an unknown but small number of children with very severe learning difficulties who did not attend any school will not have been included in the original cohort.
In December 1962 these children (aged 7–12 years) were administered a range of age-appropriate reading tests and were requested to provide their own and parental demographic information. In March 1964 they were resurveyed and behavioral information was collected by class teachers using a detailed behavioural inventory for each child. Information was also obtained retrospectively from school test records (cognition at 7 and 9 years of age) and school medical records such as height and weight. Cognition scores at 11 years of age were obtained prospectively for most children who were aged 8 years or more in December 1962.
The Children of the 1950s cohort is comprised of the 12,150 individuals (males 51.7%) in the ACDS who were born in Aberdeen and had been successfully linked to information from obstetric records held in the Aberdeen Maternity and Neonatal Databank. The process of revitalising this cohort was begun in 1998. This included tracing of the vital status and area of residence and address of the study participants and attempting to mail a questionnaire to those subjects believed to be currently resident in the UK. This 21 page questionnaire covered a range of topics including living conditions, occupation, education, income, height and weight, health, health-related behaviours, and parental vital status. A copy of the questionnaire is available from the authors on request. This was the first time that a direct attempt had been made to contact any of the cohort members since childhood, with the exception of a small subgroup (less than 1 in 5) who had been followed up in various investigations of respiratory function since the mid-1980s, known collectively as the WHEASE study [19,20].
Tracing of cohort members
The tracing of the current vital status and area of residence of the cohort members was carried out by the National Health Service Central Register (NHSCR) for Scotland, England and Wales. Using name, date of birth and the fact that the subjects were known to have been resident in Aberdeen in 1962, the NHSCR was able to identify all but 136 (1.1%) of subjects in their records.
Questionnaire mailing
The mailing of questionnaires to cohort members was begun in May 2001. For confidentiality reasons, the primary mailing to subjects believed to be resident in Scotland (n = 9,261) was carried out on our behalf by the Information and Statistics Division (ISD) of the National Health Service in Scotland who had access to current addresses. Those not responding to this initial mailing were re-mailed by ISD in August 2001. For those resident in England and Wales (n = 1,062), local health authorities undertook the mailing exercise (February – August 2002). A small proportion of subjects (n = 328) were given the questionnaires as a part of the WHEASE study. Those not returning a questionnaire after a mailing from ISD or health authorities and those who were not mailed through these channels for other reasons (mainly that a precise address could not be determined by ISD) were sent to the Driver and Vehicle Licensing Agency (DVLA) (n = 4,355). They agreed to send out questionnaires to those people who they could identify in their database (October 2002). However, in order to protect confidentiality DVLA were not able to tell us which subjects they were successful in identifying and sending a questionnaire to. Finally, between December 2002 and February 2003 a small number of non-responders (n = 556) were sent questionnaires via their siblings who had already responded earlier on. However, we were unable to ascertain how many of these questionnaires were actually delivered to their intended recipients.
Figure 1 summarizes the final outcomes of the questionnaire survey responses. We attempted to contact all subjects regardless of their early life characteristics including cognition. However, of our original 12,150 subjects we did not mail those who were dead (n = 479), were known to have emigrated from the UK (n = 291) or were members of the armed forces (or family of those in the armed forces) or those who were in prison or in a long stay psychiatric hospital (n = 62). Of those we tried to mail through the national health service for a small number (n = 27) the Health Authority or Health Board or General Practitioner refused to forward the questionnaire to the subjects' address. Finally we did not attempt to mail the very small number subjects (n = 9) who were residents of the Western Isles of Scotland or North Ireland.
Figure 1 Profile of mailing questionnaire in the Aberdeen Children of the 1950s study. Potential receiver includes subjects who were not traced through National Health Service Central Registry or cancelled General Practice registration as well as surviving cohort members.
Cognition and other factors in early life
The Moray House Picture Intelligence Test, Schonell and Adams Essential Intelligence Test and Moray House Test were used as cognitive tests at 7, 9 and 11 years respectively. The Moray House Test at 11 consisted of two ability tests (verbal reasoning 1 & 2) and two attainment tests (Arithmetic, English). The scores from these tests were combined together with a teacher's estimate to yield a "total transfer score"[21] that played a crucial role in determining the type of secondary school the child went onto. All these cognitive test scores were standardised to give means of 100 and standard deviations (SDs) of 15 using national norms for cognition at 7 and 9 and local norms for cognition at 11.
The following additional factors were also considered in this study. Sex and year of birth of children were obtained from the survey in 1962. Birth weight was measured in pounds and categorized into 4 groups (<6.5, 6.5–7.5, 7.5–8.5 and ≥8.5). Gestational age was categorized into 5 groups (≤38, 39, 40, 41 and ≥42 completed weeks of gestation). Height for age was derived from height and age at first medical examination, which was at around 5 years of age. A sex-specific z-score was calculated for each child within 3 month age bands. Using this z-score, subjects were classified into 4 groups (≤-1, -1 to 0, 0 to 1, ≥1 SD). Weight for age z-score categories were derived using the same approach. As a behavioural factor, Rutter B score was assessed by the children's class teachers based on 26 questions. A score of 9 or over was defined as indicating psychological disturbance [22,23]. In this analysis, scores were divided into four (0, 1–2, 3–8, and ≥9). Father's occupational class at birth and in 1962, number of siblings in 1962 and mother's marital status at birth were also obtained.
Ethical approval
The revitalisation of the Children of the 1950s cohort including the questionnaire survey was approved by the Scottish Multi-Centre Research Ethics Committee, the Scottish Privacy Advisory Committee and various local ethics committees in Scotland and England and Wales.
Statistical analysis
Absolute response rates (or more properly response proportions) were examined in cross-tabulations according to explanatory factors of interest. The associations between response proportions and explanatory factor were quantified in terms of odds ratios. Subjects who returned a questionnaire were considered as responders even if the questionnaire included some degree of item non-response. However, subjects who returned a questionnaire that was completely blank were classified as non-responders. In the original data, children who were considered not capable of being meaningfully tested using the standard instruments because of their degree of cognitive impairment (n = 8) were arbitrarily allocated a score of 50. As exclusion of these subjects did not change results, we included them in the analyses.
Logistic regression was used to model the association of explanatory variables with response coded as a binary variable. The trend in odds ratios were assessed by the score test. As cognitive test scores for females were slightly higher than those for males, sex-specific quintiles of cognitive test scores were used for categorization. Sex-adjusted odds ratios for a one SD increase in cognition scores at 7, 9 and 11 years were compared with those additionally adjusted for other early life factors. In addition, sex-adjusted odds ratios for cognition scores were compared with those adjusted for cognition scores at different ages. Interactions among early life factors were tested using the likelihood ratio test. All analyses were conducted using STATA 8 [24].
Results
The average age (range) of the subjects at the start of the questionnaire survey in May 2001 was 48.1 years (45.1 to 51.3). Of 11,282 (5,742 males, 5,540 females) potentially mailed subjects, 7,183 (3,432 males, 3,751 females) returned questionnaires. This constituted an overall response rate of 63.7%. A total of 3,831 (34.0%) did not respond, 245 (2.2%) refused to participate and 23 (0.2%) were unable to complete or returned blank questionnaires (Figure 1). Females had an appreciably higher response rate (67.7%) than males (59.8%). Table 1 summarises the association of response rate with key explanatory variables for males and females separately. In males and females response rates were higher for subjects whose mothers were married and whose fathers were in a non-manual social class when they were born. Response rates were also higher among those who were had fewer siblings, were taller and heavier for their age in childhood and did not exhibit symptoms of behavioural disorders (indicated by a low Rutter B score). Tests for interaction of each explanatory variable with sex were all non-significant with the exception of mother's marital status at birth where the married category was associated with a greater response rate in females compared to males. There was no association of response rate with year of birth or birth weight or gestational age (not shown).
Table 1 Questionnaire response rates by early life factors
Male Female
Number Response rate (%) Odds ratio (95% CI) Number Response rate (%) Odds ratio (95% CI)
Sex
Male 5,742 59.8 1.00
Female 5,540 67.7 1.41 (1.31 to 1.52)
Birth weight (lbs)
<6.5 1,118 58.7 1.00 1,451 65.0 1.00*
6.5–7.5 1,874 588 1.00 (0.86 to 1.17) 2,061 68.3 1.16 (1.01 to 1.34)
7.5–8.5 1,809 60.7 1.09 (0.93 to 1.27) 1,480 69.6 1.23 (1.06 to 1.44)
≥8.5 929 61.0 1.10 (0.92 to 1.32) 539 68.3 1.16 (0.94 to 1.43)
Missing† 12 83.3 9 22.2
Mother's marital status at birth‡
Single/widowed/divorced 175 52.0 1.00 190 50.5 1.00
Married 5,564 60.0 1.39 (1.02 to 1.87) 5,349 68.3 2.11 (1.58 to 2.83)
Missing† 3 66.7 1 0.0
Father's social class at birth
Unemployed/disabled/deceased 297 52.9 1.00* 336 56.9 1.00*
IV/V 1,765 55.9 1.13 (0.88 to 1.45) 1,633 63.4 1.32 (1.04 to 1.67)
III Manual 2,486 59.7 1.32 (1.04 to 1.68) 2,442 68.4 1.64 (1.30 to 2.07)
III Non-manual 633 65.9 1.72 (1.30 to 2.29) 613 73.9 2.15 (1.62 to 2.86)
I/II 561 69.2 2.00 (1.49 to 2.68) 515 77.9 2.67 (1.96 to 3.63)
Missing† 0 1 0.0
Father's social class in 1962
Unemployed/disabled/deceased 341 46.6 1.00* 357 53.5 1.00*
IV/V 1,424 55.4 1.42 (1.12 to 1.80) 1,419 64.3 1.57 (1.24 to 1.98)
III Manual 2,419 60.0 1.72 (1.36 to 2.16) 2,291 68.0 1.84 (1.47 to 2.31)
III Non-manual 701 67.3 2.36 (1.80 to 3.09) 630 73.2 2.37 (1.79 to 3.13)
I/II 776 67.8 2.41 (1.85 to 3.14) 781 76.6 2.84 (2.16 to 3.73)
Missing 81 43.2 0.87 (0.54 to 1.42) 62 50.0 0.87 (0.51 to 1.48)
Number of siblings in 1962
5 or more 1,041 50.4 1.00* 946 59.2 1.00*
4 938 57.7 1.34 (1.12 to 1.60) 942 65.2 1.29 (1.07 to 1.56)
3 1,424 61.5 1.57 (1.34 to 1.85) 1,420 70.9 1.67 (1.41 to 1.99)
2 1,749 64.2 1.76 (1.50 to 2.06) 1,677 71.3 1.71 (1.45 to 2.03)
1 545 64.6 1.79 (1.45 to 2.22) 510 68.8 1.52 (1.21 to 1.91)
Missing 45 35.6 0.54 (0.29 to 1.00) 45 53.3 0.79 (0.44 to 1.42)
Height for age at 1st medical exam 1962
≤-1 734 54.4 1.00* 872 62.3 1.00*
-1 to 0 1,882 57.1 1.12 (0.94 to 1.33) 1,969 66.0 1.17 (1.00 to 1.39)
0 to 1 2,139 62.3 1.39 (1.17 to 1.65) 1,775 69.9 1.40 (1.18 to 1.67)
≥1 737 66.9 1.70 (1.37 to 2.10) 723 73.7 1.70 (1.37 to 2.11)
Missing 250 52.8 0.94 (0.70 to 1.25) 201 67.7 1.27 (0.92 to 1.75)
Weight for age at 1st medical exam (SD)
≤-1 870 56.7 1.00* 795 63.9 1.00*
-1 to 0 2,045 59.5 1.12 (0.96 to 1.32) 2,109 65.5 1.07 (0.91 to 1.27)
0 to 1 1,810 59.2 1.11 (0.94 to 1.31) 1,667 70.0 1.32 (1.10 to 1.58)
≥1 767 67.8 1.61 (1.31 to 1.97) 769 72.7 1.50 (1.21 to 1.87)
Missing 250 52.8 0.86 (0.65 to 1.13) 200 67.5 1.17 (0.84 to 1.63)
Rutter B behavioral score
9+ 595 47.4 1.00* 320 57.2 1.00*
3–8 1,736 56.2 1.42 (1.18 to 1.72) 1,448 61.5 1.19 (0.93 to 1.53)
1–2 1,517 63.8 1.96 (1.61 to 2.38) 1,637 69.6 1.72 (1.34 to 2.20)
0 1,720 65.7 2.13 (1.75 to 2.58) 1,990 73.3 2.06 (1.61 to 2.63)
Missing 174 44.3 0.88 (0.63 to 1.24) 145 54.5 0.90 (0.60 to 1.33)
* p < 0.05 for the score test for trend (missings were not included).
† We did not calculate odds ratio and 95% confidence interval due to small sample size.
‡ p < 0.05 for the interaction with sex tested by using likelihood ratio test in logistic regression.
The associations of response rate with cognitive score at ages 7, 9 and 11 years for both sexes combined are shown in Figure 2. The almost monotonic increase in questionnaire response rate with each 5 point increase in score at each age is very striking. Response rates by quintile of cognitive score are summarised separately by sex in Table 2. Response rates increased progressively at all ages from the bottom to the top fifth of cognitive test score in both sexes. Each of the four components of cognition measured at 11 years (verbal reasoning × 2, English and arithmetic) showed similar associations with response and were adequately summarised by association with the total score. There was no evidence of a statistically significant interaction between sex and cognitive scores at any age, and thus in the remaining analyses we present results for both sexes combined.
Table 2 Questionnaire response rates by sex-specific cognition score quintiles
Male (n = 5,742) Female (n = 5,540)
Test score Test score
Mean Range Number Response rate (%) Odds ratio (95% CI) Mean Range Number Response rate (%) Odds ratio (95% CI)
Cognition at 7
Lowest quintile 84 50, 93 1,111 48.2 1.00* 84 50, 93 1,006 52.9 1.00*
2nd lowest 98 94, 102 1,046 54.0 1.26 (1.07 to 1.50) 98 94, 103 1,058 60.3 1.35 (1.14 to 1.61)
Middle 106 103, 110 1,128 59.3 1.57 (1.33 to 1.86) 107 104, 111 1,023 72.7 2.38 (1.97 to 2.87)
2nd highest 115 111, 120 1,126 67.4 2.23 (1.87 to 2.65) 116 112, 121 1,099 73.6 2.49 (2.06 to 3.00)
Highest quintile 129 121, 153 1,173 70.3 2.55 (2.14 to 3.05) 130 122, 166 1,080 78.4 3.24 (2.66 to 3.95)
Missing 158 50.0 1.08 (0.77 to 1.50) 274 66.1 1.73 (1.31 to 2.29)
Cognition at 9
Lowest quintile 85 50, 95 1,047 45.7 1.00* 89 60, 98 1,002 54.3 1.00*
2nd lowest 102 96, 106 998 57.2 1.59 (1.34 to 1.90) 104 99, 107 921 62.4 1.40 (1.17 to 1.68)
Middle 111 107, 115 1,207 60.5 1.82 (1.54 to 2.16) 111 108, 115 1,078 68.9 1.87 (1.56 to 2.24)
2nd highest 119 116, 124 1,049 64.9 2.20 (1.84 to 2.63) 119 116, 123 1,047 74.5 2.46 (2.03 to 2.98)
Highest quintile 135 125, 174 1,139 72.2 3.09 (2.57 to 3.71) 133 124, 176 1,110 79.1 3.19 (2.61 to 3.88)
Missing 302 49.7 1.17 (0.91 to 1.52) 382 60.5 1.29 (1.01 to 1.64)
Cognition at 11
Lowest quintile 78 54, 86 920 45.5 1.00* 80 47, 87 843 52.4 1.00*
2nd lowest 91 86, 95 903 54.6 1.44 (1.19 to 1.73) 92 88, 97 890 63.3 1.56 (1.29 to 1.89)
Middle 99 96, 103 934 63.6 2.09 (1.73 to 2.52) 101 97, 104 845 74.1 2.59 (2.10 to 3.20)
2nd highest 108 104, 113 931 67.2 2.45 (2.02 to 2.98) 109 105, 114 881 74.6 2.66 (2.16 to 3.28)
Highest quintile 121 113, 144 922 71.4 2.98 (2.44 to 3.64) 121 114, 146 874 78.6 3.33 (2.68 to 4.15)
Missing 1,132 56.7 1.57 (1.32 to 1.87) 1,207 64.3 1.63 (1.37 to 1.95)
*p < 0.05 for the score test for trend (missings were not included).
Figure 2 Response rates by cognition test scores at 7, 9 and 11 for males and females combined. Five point bands were used for cognition scores.
Sex-adjusted odds ratios for one SD increase in cognition scores at 7, 9 and 11 were slightly attenuated (less than 6%) by adjustment for socio-economic position, height and weight for age and Rutter B score (Table 3) but nevertheless remained substantial. Cognitive test scores at ages 7, 9 and 11 years are highly correlated (Pearson correlation coefficients: 0.74 to 0.88). How far the associations of response rate with cognitive score at each age are independent is also examined in Table 3. Nearly all of the effects at ages 7 and 9 were removed by adjustment for cognitive test score at age 11 years. In contrast, the effect of cognition at 11 was only slightly attenuated by adjustment for cognitive test score at 7 or 9.
Table 3 Odds ratios for questionnaire response by cognition scores with adjustment for other early life factors and cognition at different ages
Cognition at 7 Cognition at 9 Cognition at 11
OR† (95% CI) OR† (95% CI) OR† (95% CI)
Adjusted for
Sex 1.42 (1.36 to 1.49) 1.50 (1.43 to 1.57) 1.55 (1.47 to 1.62)
Sex, Marital status at birth 1.42 (1.35 to 1.48) 1.49 (1.42 to 1.56) 1.54 (1.47 to 1.62)
Sex, Father's occupation at birth & in 1962, Number of children 1.34 (1.27 to 1.41) 1.42 (1.35 to 1.49) 1.48 (1.40 to 1.56)
Sex, Height & Weight at 1st medical exam 1.40 (1.34 to 1.47) 1.48 (1.41 to 1.55) 1.53 (1.46 to 1.61)
Sex, Rutter B behavioral score 1.36 (1.29 to 1.42) 1.43 (1.36 to 1.51) 1.50 (1.42 to 1.58)
Sex, All early life factors‡ 1.29 (1.22 to 1.36) 1.37 (1.30 to 1.45) 1.43 (1.36 to 1.52)
Sex, Cognition at 7 - 1.36 (1.27 to 1.46) 1.46 (1.37 to 1.57)
Sex, Cognition at 9 1.14 (1.06 to 1.22) - 1.44 (1.30 to 1.59)
Sex, Cognition at 11 1.08 (1.01 to 1.15) 1.09 (0.98 to 1.20) -
* Due to missing data for some variables, this analysis was restricted to a subset of 8,452 subjects.
† Odds ratio for one standardized deviation increase in each cognition score using logistic regression.
‡ Adjusted for all early life variables (marital status at birth, father's occupation at birth & in 1962, number of children, height & weight at 1st medical exam, and Rutter B behavioural score)
Discussion
We have found that a wide range of characteristics in infancy and childhood were associated with questionnaire response rate. Cognitive test score in childhood was particularly strongly related to the probability of responding to the questionnaire survey independently of other factors, with an almost monotonic increase in response rate, with no evidence of a threshold effect. At age 11 response rates among those in the top fifth of the cognitive test score distribution had a response rate that in absolute terms was 25% higher than in the bottom fifth.
Although the possibility of obtaining biased samples in postal questionnaires with incomplete response rates has been well documented in the survey and epidemiological literature [3,5,7,25,26], the emphasis has been mainly upon the determinants of response related to characteristics of subjects at the time they receive the survey instrument. Ours is the first study that has looked at early life influences on response rate to a postal questionnaire in middle age in a cohort not previously contacted in adult life. A similar analysis by Wadsworth and colleagues [16] of childhood influences on later continued participation in the British 1946 birth cohort, in which subjects had been repeatedly contacted throughout life, found that continued participation was least likely among those who had been in the most disadvantaged socio-economic circumstances in childhood and those with lowest cognitive test scores at 8 years of age. The sex-adjusted odds ratio for avoidable loss to follow-up at age 53 years in the top compared to the bottom quarter of childhood cognition was 0.52 (95% Confidence Interval; 0.42 to 0.64). However, these analyses did not involve any multi-variable adjustments, so it is unclear how far this association may be explained by confounding with socio-economic and other factors.
Finding an association between childhood cognition and questionnaire response rate in adult life was expected. However, the strength and consistency of this association, with almost monotonic increases across the entire range of test scores, are surprising. At least two explanations for this powerful effect should be considered. The first one would hypothesise a pathway via education. High cognitive test scores in childhood are highly predictive of higher levels of education. Subjects with higher level of education may have higher response rate because they may be more health conscious, more interested in research and not feel intimidated by a relatively substantial questionnaire. The second possibility is a more direct pathway, where cognitive ability such as a problem solving (akin to completing a 21 page questionnaire) played an important role. Our study was not able to definitively distinguish between these two pathways. However, it should be noted that the effect of cognition at 11 was independent of cognition at 7 and 9 years, whereas the effect of cognition at 7 was attenuated substantially by adjusting for cognition at 11. The test at age 11 explicitly included components that measured educational attainment (mathematics and English), and the overall score was used as one of the key criteria for determining secondary school. Thus cognitive score at 11 is more likely to reflect educational progress to that age than the score at 7 years, suggesting that our data are most consistent with the link between childhood cognition and response rate being via attained educational level.
Some limitations of this study should be noted. Firstly, bias by undelivered mail might be possible. We do not know which subjects actually received a questionnaire. If cognitive test score was positively related to the probability of actually receiving the questionnaire this would generate, or at least contribute to, the observed gradient. However, this seems unlikely. As we have shown elsewhere [17], the lower a person's cognitive test score the less likely it was that they moved away from the Aberdeen area, and thus the lower the probability that the NHS had an incorrect or out-of-date address for them. Secondly, our data set included an appreciable proportion of subjects who did not have a cognitive test score at 11 (n = 2,339, 20.7% of 11,282). However, 80% (n = 1,873) of those without score at 11 were missing it simply because they had not reached this age during the survey period. We repeated the multivariate analyses after excluding cognition at 11 and confirmed that odds ratios for cognition at 7 and 9 were essentially unchanged. Therefore, we do not believe that selection bias due to this missing data could explain the observed associations. Finally, it should be noted that our findings may not automatically apply to other survey methods such as telephone and face to face interview.
Any postal questionnaire survey should be designed and undertaken to achieve the maximum possible response rate. However, in many contexts there will remain major concerns about non-response bias. In these situations the powerful links that exist between early life characteristics and response needs to be taken into account when interpreting the data. This can be done in part by employing sensitivity analyses. Moreover, in some situations multiple imputation methods may also be useful[27,28].
Conclusion
Our results indicate that the interpretation of associations between childhood and later life factors in life-course studies using postal questionnaires need to take account of the fact that factors in childhood can be strongly related to response rate. Particular caution is needed if the outcomes are related to childhood cognitive function. Sensitivity analyses that explore the extent of such biases are strongly recommended.
Quite apart from these methodological conclusions, the study finding shows that this simple and routinely collected information on cognition is extraordinarily predictive of the complex collection of later-life behaviours and circumstances that jointly generate the likelihood of a completed questionnaire being returned.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DAL, SMM and HC designed the questionnaire survey. HC supervised the data collection. YN analyzed, interpreted the data and drafted the paper in consultation with the other authors. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We are very grateful to Raymond Illsley for providing us with the data from the Aberdeen Child Development Survey and for his advice about the study. Graeme Ford played a crucial role in identifying individual cohort members and in helping us initiate the process of revitalisation. Sally Macintyre, Doris Campbell, George Davey Smith, Marion Hall, Bianca de Stavola, Susan Morton, David Batty, David Godden, Di Kuh, Debbie Lawlor, Glyn Lewis and Viveca Östberg collaborated with David Leon to revitalise the cohort. Heather Clark managed the study at the Dugald Baird Centre, Aberdeen with the assistance of Margaret Beveridge. We would also like to thank staff at the ISD (Edinburgh), GRO (Scotland) and NHSCR (Southport) for their substantial contributions, John Lemon who undertook the linkage to the Aberdeen Maternity and Neonatal Databank and Valerie Mccormack for the statistical advice. The Aberdeen Children of the 1950s Study was funded as a component project (G0828205) of a Medical Research Council Co-operative Group Life-course and trans-generational influences on disease risk (G9819083). A project on cognition and adult health in the cohort has been funded by the Chief Scientists Office, Scottish Executive Health Department, which currently funds HC. YN was funded by the Uehara Memorial Foundation.
==== Refs
Sandler DP On revealing what we'd rather hide: the problem of describing study participation Epidemiology 2002 13 117 11880747 10.1097/00001648-200203000-00001
Stang A Nonresponse research--an underdeveloped field in epidemiology Eur J Epidemiol 2003 18 929 931 14598921 10.1023/A:1025877501423
Eaker S Bergstrom R Bergstrom A Adami HO Nyren O Response rate to mailed epidemiologic questionnaires: a population-based randomized trial of variations in design and mailing routines Am J Epidemiol 1998 147 74 82 9440402
Hoeymans N Feskens EJ Van Den Bos GA Kromhout D Non-response bias in a study of cardiovascular diseases, functional status and self-rated health among elderly men Age Ageing 1998 27 35 40 9504364
Macera CA Jackson KL Davis DR Kronenfeld JJ Blair SN Patterns of non-response to a mail survey J Clin Epidemiol 1990 43 1427 1430 2254781 10.1016/0895-4356(90)90112-3
Criqui MH Barrett-Connor E Austin M Differences between respondents and non-respondents in a population-based cardiovascular disease study Am J Epidemiol 1978 108 367 372 727205
Paganini-Hill A Hsu G Chao A Ross RK Comparison of early and late respondents to a postal health survey questionnaire Epidemiology 1993 4 375 379 8347749
Goldberg M Chastang JF Leclerc A Zins M Bonenfant S Bugel I Kaniewski N Schmaus A Niedhammer I Piciotti M Chevalier A Godard C Imbernon E Socioeconomic, demographic, occupational, and health factors associated with participation in a long-term epidemiologic survey: a prospective study of the French GAZEL cohort and its target population Am J Epidemiol 2001 154 373 384 11495861 10.1093/aje/154.4.373
Korkeila K Suominen S Ahvenainen J Ojanlatva A Rautava P Helenius H Koskenvuo M Non-response and related factors in a nation-wide health survey Eur J Epidemiol 2001 17 991 999 12380710 10.1023/A:1020016922473
Sonne-Holm S Sorensen TI Jensen G Schnohr P Influence of fatness, intelligence, education and sociodemographic factors on response rate in a health survey J Epidemiol Community Health 1989 43 369 374 2614328
Ben-Shlomo Y Kuh D A life course approach to chronic disease epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives Int J Epidemiol 2002 31 285 293 11980781 10.1093/ije/31.2.285
Kuh D Ben-Shlomo Y Lynch J Hallqvist J Power C Life course epidemiology J Epidemiol Community Health 2003 57 778 783 14573579 10.1136/jech.57.10.778
Taylor MD Hart CL Davey Smith G Starr JM Hole DJ Whalley LJ Wilson V Deary IJ Childhood mental ability and smoking cessation in adulthood: prospective observational study linking the Scottish Mental Survey 1932 and the Midspan studies J Epidemiol Community Health 2003 57 464 465 12775797 10.1136/jech.57.6.464
Whalley LJ Deary IJ Longitudinal cohort study of childhood IQ and survival up to age 76 British Medical Journal 2001 322 819 11290633
Osler M Andersen AM Due P Lund R Damsgaard MT Holstein BE Socioeconomic position in early life, birth weight, childhood cognitive function, and adult mortality. A longitudinal study of Danish men born in 1953 J Epidemiol Community Health 2003 57 681 686 12933773 10.1136/jech.57.9.681
Wadsworth ME Butterworth SL Hardy RJ Kuh DJ Richards M Langenberg C Hilder WS Connor M The life course prospective design: an example of benefits and problems associated with study longevity Soc Sci Med 2003 57 2193 2205 14512249 10.1016/S0277-9536(03)00083-2
Batty GD Morton SM Campbell D Clark H Smith GD Hall M Macintyre S Leon DA The Aberdeen Children of the 1950s cohort study: background, methods and follow-up information on a new resource for the study of life course and intergenerational influences on health Paediatr Perinat Epidemiol 2004 18 221 239 15130162 10.1111/j.1365-3016.2004.00552.x
Birch HG Richardson SA Baird D Horobin G Ilsley R Mental Subnormality in the Community: A Clinical and Epidemiological Study 1970 Baltimore, The Williams & Wilkins Co.
Dawson B Illsley R Horobin G Mitchell R A survey of childhood asthma in Aberdeen Lancet 1969 1 827 830 4180376 10.1016/S0140-6736(69)92082-0
Bodner C Ross S Douglas G Little J Legge J Friend J Godden D The prevalence of adult onset wheeze: longitudinal study British Medical Journal 1997 314 792 793 9080997
Nisbet JD Entwistle NJ The Age of Transfer to Secondary Education. Publication 53 of the Scottish Council for Research in Education. 1966 London, University of London Press Ltd
Rutter M A children's behaviour questionnaire for completion by teachers: preliminary findings J Child Psychol Psychiatry 1967 8 1 11 6033260
Rutter M Tizard J Education, Health and Behaviour. 1970 London, Longman
Stata Corporation Stata Base Reference Manual 2003 Texas, Stata Press
Roberts LM Wilson S Roalfe A Bridge P A randomised controlled trial to determine the effect on response of including a lottery incentive in health surveys [ISRCTN32203485] BMC Health Serv Res 2004 4 30 15533256 10.1186/1472-6963-4-30
Hebert R Bravo G Korner-Bitensky N Voyer L Refusal and information bias associated with postal questionnaires and face-to-face interviews in very elderly subjects J Clin Epidemiol 1996 49 373 381 8676188 10.1016/0895-4356(95)00527-7
Barzi F Woodward M Imputations of missing values in practice: results from imputations of serum cholesterol in 28 cohort studies Am J Epidemiol 2004 160 34 45 15229115 10.1093/aje/kwh175
Engels JM Diehr P Imputation of missing longitudinal data: a comparison of methods J Clin Epidemiol 2003 56 968 976 14568628 10.1016/S0895-4356(03)00170-7
| 15876348 | PMC1142322 | CC BY | 2021-01-04 16:32:51 | no | BMC Med Res Methodol. 2005 May 5; 5:16 | utf-8 | BMC Med Res Methodol | 2,005 | 10.1186/1471-2288-5-16 | oa_comm |
==== Front
BMC NephrolBMC Nephrology1471-2369BioMed Central London 1471-2369-6-41585423110.1186/1471-2369-6-4Research ArticleTime course study of oxidative and nitrosative stress and antioxidant enzymes in K2Cr2O7-induced nephrotoxicity Pedraza-Chaverrí José [email protected] Diana [email protected] Omar N [email protected] Raymundo C [email protected]ández-Pando Rogelio [email protected]ías-Ruvalcaba Norma A [email protected] Perla D [email protected] Marcos I [email protected] Edilia [email protected]ívar Liliana [email protected] María E [email protected] María E [email protected] Facultad de Química, Departamento de Biología, Edificio B, Segundo Piso, Laboratorio 209, Universidad Nacional Autónoma de México (UNAM), Ciudad Universitaria, 04510, México, D.F., México2 Facultad de Medicina, Departamento de Farmacología, Universidad Nacional Autónoma de México (UNAM), Ciudad Universitaria, 04510, México, D.F., México3 Instituto Nacional de Ciencias Médicas y Nutrición "Salvador Zubirán", Departamento de Patología, 14000, México, D.F., México4 Facultad de Química, Edificio B, Laboratorio 124, Departamento de Química Orgánica, Universidad Nacional Autónoma de México (UNAM), Ciudad Universitaria 04510, México, D.F., México5 Departamento de Nefrología, Instituto Nacional de Cardiología "Ignacio Chávez", Juan Badiano #1, Col Sección XVI, 14080 Tlalpan, México, D.F., México6 Facultad de Química, Edificio B, Departamento de Química Analítica, Universidad Nacional Autónoma de México (UNAM), Ciudad Universitaria 04510, México, D.F., México2005 26 4 2005 6 4 4 30 10 2004 26 4 2005 Copyright © 2005 Pedraza-Chaverrí et al; licensee BioMed Central Ltd.2005Pedraza-Chaverrí 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
Potassium dichromate (K2Cr2O7)-induced nephrotoxicity is associated with oxidative and nitrosative stress. In this study we investigated the relation between the time course of the oxidative and nitrosative stress with kidney damage and alterations in the following antioxidant enzymes: Cu, Zn superoxide dismutase (Cu, Zn-SOD), Mn-SOD, glutathione peroxidase (GPx), glutathione reductase (GR), and catalase (CAT).
Methods
Nephrotoxicity was induced in rats by a single injection of K2Cr2O7. Groups of animals were sacrificed on days 1,2,3,4,6,8,10, and 12. Nephrotoxicity was evaluated by histological studies and by measuring creatinine clearance, serum creatinine, blood urea nitrogen (BUN), and urinary excretion of N-acetyl-β-D-glucosaminidase (NAG) and total protein. Oxidative and nitrosative stress were measured by immunohistochemical localization of protein carbonyls and 3-nitrotyrosine, respectively. Cu, Zn-SOD, Mn-SOD, and CAT were studied by immunohistochemical localization. The activity of total SOD, CAT, GPx, and GR was also measured as well as serum and kidney content of chromium and urinary excretion of NO2 -/NO3-. Data were compared by two-way analysis of variance followed by a post hoc test.
Results
Serum and kidney chromium content increased reaching the highest value on day 1. Nephrotoxicity was made evident by the decrease in creatinine clearance (days 1–4) and by the increase in serum creatinine (days 1–4), BUN (days 1–6), urinary excretion of NAG (days 1–4), and total protein (day 1–6) and by the structural damage to the proximal tubules (days 1–6). Oxidative and nitrosative stress were clearly evident on days 1–8. Urinary excretion of NO2-/NO3- decreased on days 2–6. Mn-SOD and Cu, Zn-SOD, estimated by immunohistochemistry, and total SOD activity remained unchanged. Activity of GPx decreased on days 3–12 and those of GR and CAT on days 2–10. Similar findings were observed by immunohistochemistry of CAT.
Conclusion
These data show the association between oxidative and nitrosative stress with functional and structural renal damage induced by K2Cr2O7. Renal antioxidant enzymes were regulated differentially and were not closely associated with oxidative or nitrosative stress or with kidney damage. In addition, the decrease in the urinary excretion of NO2-/NO3- was associated with the renal nitrosative stress suggesting that nitric oxide was derived to the formation of reactive nitrogen species involved in protein nitration.
==== Body
Background
K2Cr2O7 is a chemical compound that is widely used in metallurgy, chrome plating, chemical industry, textile manufacture, wood preservation, photography and photoengraving, refractory and stainless steel industry, and cooling systems [1]. Occupational exposure to chromium has been associated with welders, chrome-plating workers, and chromium pigment factories workers [2]. Chromium is known to cause allergic dermatitis [3], carcinogenicity [4], and ARF in humans [2,5,6] and in animals [7-11].
Acute renal failure (ARF) induced by potassium dichromate (K2Cr2O7) has been used as a model to study the pathophysiology of this disease [7,9]. Experimental data show that chromate affects selectively the convoluted section of the proximal tubules [7,12-15] and induces acute necrosis of renal tubules [14,15]. Clinical and experimental renal damage induced by K2Cr2O7 has been associated with oxidative stress [15-20]. In fact, some antioxidants such as ascorbic acid, vitamin E, N-acetyl cysteine, and glutathione prevent the K2Cr2O7 induced renal damage [21-25] whereas the inhibition of glutathione biosynthesis enhances it [22,24]. Evidences suggest that reactive oxygen species (ROS) are involved in Cr(VI)-induced cell injury [16,17,19,26]. Chromium reduction intermediates [Cr(V), Cr(IV), and Cr(III)], which could be generated under physiological conditions, may be toxic as they involve ROS production [27-29]. In vitro chromate reduction via H2O2 has been shown to produce hydroxyl radicals via a Fenton-like reaction [27,30-33].
We recently showed that stannous chloride (SnCl2) pretreatment has a protective role in K2Cr2O7-induced nephrotoxicity [15,34]. SnCl2 is a potent inductor of heme oxygenase-1 [15,35] and the protective role of SnCl2 in this experimental model has been attributed, at least in part, to the heme oxygenase-1 preinduction [15,34]. We also showed that renal activity of Cu, Zn-SOD, Mn-SOD, and GR remained unchanged at 24 and 48 h whereas GPx and CAT activities remained unchanged at 24 h but decreased at 48 h in K2Cr2O7-injected rats. In the present work we performed a time course study of functional and structural renal damage, oxidative and nitrosative stress, and the behavior of the antioxidant enzymes Cu, Zn superoxide dismutase (Cu, Zn-SOD), Mn superoxide dismutase (Mn-SOD), glutathione reductase (GR), glutathione peroxidase (GPx), and catalase (CAT) in K2Cr2O7-induced nephrotoxicity to know if the antioxidant enzymes respond in a coordinate way and if there is an association between (a) oxidative/nitrosative stress and renal damage, (b) oxidative/nitrosative stress and the antioxidant enzymes, and (c) antioxidant enzymes and renal damage.
Methods
Reagents
Formaldehyde, anhydrous absolute ethanol, xylol, methanol, chloroform, and K2Cr2O7 were from J.T. Baker (México, D.F.), p-nitrophenyl-N-acetyl-β-D-glucosaminide, acrylamide, NADPH, N, N'-methylene-bis-acrylamide, xanthine, xanthine oxidase, nitroblue tetrazolium (NBT), β-nicotinamide adenine dinucleotide phosphate, reduced form (β-NADPH), aprotinin, leupeptin, pepstatin, and 2,4-dinitrophenylhydrazine (DNPH) were from Sigma Chemical Co. (St. Louis, MO, USA). Commercial kits to measure creatinine and blood urea nitrogen were from Spinreact (Girona, Spain). Rabbit anti-rat polyclonal antibodies against Mn-SOD (Cat. # SOD 111) and Cu, Zn-SOD (Cat. # SOD 101) were from Stressgen Biotechnologies Co. (Victoria, BC, Canada). Rabbit anti-human CAT polyclonal antibodies (Cat. # 219010) were from Calbiochem (San Diego, CA, USA). Goat anti-dinitrophenol (DNP) polyclonal antibodies (Cat. # J06) were from Biomeda Corporation (Foster City, CA, USA). Rabbit anti-3-nitrotyrosine (3-NT) polyclonal antibodies (Cat. #06-284) were from Upstate (Lake Placid, NY, USA). Anti-rabbit immunoglobulin horseradish peroxidase antibodies (Cat. # NA-934) were from Amersham (Buckinghamshire, UK). Donkey anti-goat horseradish peroxidase antibodies (Cat. # SC2020) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Enhanced chemiluminiscence (ECL) kit for Western blot was purchased from Amersham Life Sciences (Buckinghamshire, England). All other chemicals were reagent grade and commercially available.
Experimental design
Seven-week-old male Wistar rats with an initial body weight of 200–210 g were used. Experimental work was approved by CONACYT (#25441) and DGAPA (IN210201) and followed the guidelines of Norma Oficial Mexicana (NOM-ECOL-087-1995). Two groups of rats were studied: 1) CT, control injected subcutaneously with 0.5 ml isotonic saline solution (n = 40); and 2) K2Cr2O7, treated with a single subcutaneous injection of 15 mg/Kg K2Cr2O7 [7] in a volume of 0.5 ml (n = 43). The study was performed in two stages: rats from days 1,2,3,4, and 6 were studied in the first one (n = 5/per group) and rats from days 8, 10, and 12 were studied in the second one (n = 5 for control group and n = 6 for K2Cr2O7 group). Rats were sacrificed on days 1,2,3,4,6,8,10, and 12. There was no mortality. Rats had free access to water and CT group was pair-fed to match the food intake of K2Cr2O7 group. Animals were maintained in metabolic cages to collect 24-h urine to measure N-acetyl-β-D-glucosaminidase (NAG), total protein, creatinine, and NO2-/NO3-.
Animals were sacrificed by decapitation and trunk blood was collected at room temperature to obtain serum to measure chromium content and the markers of renal function, creatinine and blood urea nitrogen (BUN). Glomerular filtration rate was estimated by creatinine clearance. The kidneys were obtained to perform histological analysis, chromium determination, the activity of antioxidant enzymes, Western blot of CAT, Cu, Zn-SOD, and Mn-SOD, and the immunohistochemical localization of 3-NT, protein carbonyls, CAT, Cu, Zn-SOD, and Mn-SOD.
Chromium concentration
Chromium content in serum and in total kidney was measured by graphite-furnace atomic-absorption spectrometry on a Perkin Elmer 3110 device with furnace.
Markers of nephrotoxicity
Creatinine and BUN were measured using commercial kits according to the instructions of manufacturers and were expressed as mg/dL. Creatinine clearance was calculated with the serum and urine data according to the standard formula. Urinary NAG excretion was determined at 405 nm using p-nitrophenyl-N-acetyl-β-D-glucosaminide as substrate and the data were expressed as U/24 h [7,15]. One unit of NAG was defined as the amount of enzyme that releases 1 μmol of p-nitrophenol in the assay conditions. Total protein in urine was measured by a turbidimetric method with 12.5% trichloroacetic acid at 420 nm [7,15] and the data were expressed as mg/24 h.
Histological studies
Thin slices of kidney tissue with cortex and medulla were fixed by immersion in buffered formalin (pH 7.4), dehydrated and embedded in paraffin. Sections (4 μm) were stained with hematoxylin and eosin (H&E) [15]. A quantitative histological damage was determined by using a Leica Qwin Image Analyzer (Cambridge, UK). The histological profile of twenty proximal tubules randomly selected per rat (5 rats per group) was recorded (n = 100 tubuli/group at each time point). The number of tubules with histopathological alterations like swelling, cytoplasmic vacuolization, desquamation or necrosis was registered and the data were expressed as percentage of damaged tubules. The percentage of damaged tubules of K2Cr2O7 and control groups was compared.
Immunohistochemical localization of 3-NT, protein carbonyls, Cu, Zn-SOD, Mn-SOD, and CAT
For immunohistochemistry, 4 μm sections were deparaffined with xylol and rehydrated with ethanol. Endogenous peroxidase was quenched/inhibited with 4.5% H2O2 in methanol by incubation for 1.5 h at room temperature. The sections used for DNP immunohistochemistry were incubated with 0.2% DNPH in 2 N HCl for 60 min at room temperature in absence of light and then were extensively washed with phosphate buffer saline. Nonspecific adsorption was minimized by leaving the sections in 3% bovine serum albumin in phosphate buffer saline for 30 min. Sections were incubated overnight with a 1:700 dilution of anti-3-NT antibodies [15], or a 1:500 dilution of anti-DNP antibodies, or a 1:250 dilution of anti Cu, Zn-SOD antibodies, or a 1:500 dilution of anti Mn-SOD antibodies, or a 1:500 dilution of anti CAT antibodies. The sections were incubated with a 1:500 dilution of a peroxidase conjugated anti-rabbit immunoglobulin antibodies (for 3-NT, Cu, Zn-SOD, Mn-SOD, and CAT) or with a 1:500 dilution of a peroxidase conjugated anti-goat Ig (for DNP) for 1 h, and finally incubated with H2O2-diaminobenzidine for 1 min. Sections were counterstained with hematoxylin and observed under light microscopy. All the sections were incubated under the same conditions with the same concentration of antibodies, and in the same running, so the immunostaining was comparable among the different experimental groups. The stained area was quantified using a SigmaScan Pro (version 4.01.003) (Jandel Scientific, San Rafael, CA). The data are expressed as percent respect to day 0.
Urinary excretion of NO2-/NO3-
Nitric oxide (NO) is a labile substance with a short half-life and it decomposes rapidly to NO2- and NO3- in biological solutions [36], and these stable breakdown products have been measured as an index of NO production [37]. NO2- and NO3- were measured in the 24-hour urine samples at all time points. Urine samples were first incubated with E. coli nitrate reductase to convert the NO3- to NO2-, as described previously [38,39]. To prepare this enzyme, E. coli was grown for 18 hours under anaerobic conditions in nitrate rich medium, washed, resuspended in phosphate buffer and frozen at -70°C until use. The samples were incubated with the enzyme in phosphate ammonium formate buffer (pH 7.3) for one hour at 37°C. After incubation, total NO2- in the samples (representing both NO2- and reduced NO3-) was measured using the Griess reagent. Known concentrations of NaNO2 and NaNO3 were used as standards in each assay. Data were expressed as nmol/min.
Activity of antioxidant enzymes
Kidney was homogenized in a Polytron (Model PT 2000, Brinkmann, Westbury, NY, USA) for 10 seconds in cold 50 mM potassium phosphate, 0.1% Triton X-100, pH = 7.0. The homogenate was centrifuged at 19,000 × g and 4°C for 30 min and the supernatant was separated to measure total protein and the activities of CAT, GPx, GR, and total SOD.
Total SOD activity in renal cortex homogenates was assayed spectrophotometrically at 560 nm by a previously reported method using NBT as the indicator reagent [34]. The amount of protein that inhibited NBT reduction to 50% of maximum was defined as one unit of SOD activity. Results were expressed as U/mg protein. CAT activity in renal cortex, was assayed by a method based on the disappearance of 30 mM H2O2 at 240 nm [34] and the data were expressed as k/mg protein. GPx activity in renal cortex was measured at 340 nm using GR and NADPH in a coupled reaction [34]. One unit of GPx was defined as the amount of enzyme that oxidize 1 μmol of NADPH/min. Data were expressed as U/mg protein. GR activity in renal cortex was assayed using oxidized glutathione as substrate and measuring the disappearance of NADPH at 340 nm [34]. One unit of GR was defined as the amount of enzyme that oxidize 1 μmol of NADPH/min. Data were expressed as U/mg protein.
Western blot
Renal cortex (100 mg) was homogenized in 300 μl of phosphate buffer 50 mM, pH 7.4 containing the following cocktail of protease inhibitors: leupeptin 5 μg/ml, pepstatin 7 μg/ml, aprotinin 5 μg/ml, and EDTA 1 mM. The homogenates were centrifuged at 1,000 × g and 4°C for 10 min. Thirty μg of protein were fractionated by reducing 12.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis and electroblotted to a nitrocellulose membrane (Hybond™ ECL™, Amersham, Buckinghamshire, England). Immunodetection was performed using specific primary antibodies against CAT (1:500 dilution), Mn-SOD (1,5000 dilution), or Cu, Zn-SOD (1:5,000 dilution). Membranes were then probed with the appropriate secondary antibody-peroxidase conjugate (1:5,000 dilution). The hybrids were visualized using ECL detection system and quantified by densitometry (Sigma ScanPro, version 4.0, San Rafael, CA, USA). Data were expressed as % of the control.
Statistics
The data are expressed as the mean ± SEM and were analyzed by two-way analysis of variance followed by Bonferroni t-test using the software Prism 2.01 (GraphPad, San Diego, CA, USA). A P value less than 0.05 was considered statistically significant.
Results and discussion
ROS have been involved in the pathophysiology of ARF induced by K2Cr2O7 and some antioxidants are able to ameliorate renal damage induced by this compound [15,16,18,23,25]. In addition, it has been previously shown that renal heme oxygenase-1 preinduction by SnCl2 ameliorates ARF and prevents oxidative and nitrosative stress induced 24 h after K2Cr2O7 injection [15,34]. In the present work we performed a time-course analysis of the nephrotoxicity, oxidative and nitrosative stress, and changes in antioxidant enzymes induced by K2Cr2O7. Body weight of control and K2Cr2O7-treated rats was similar at all time points (see additional file 1). Urinary volume was increased in K2Cr2O7-treated animals on days 3 (1.9-fold) and 4 (1.6-fold) and returned to control values thereafter (see additional file 1). Polyuria observed in our rats is consistent with previous data [2]. We measured the chromium content in serum and kidney and it was found that it increased on days 1–6 and 1–12, respectively (Fig. 1). In both cases the peak value was reached on day 1 and then the chromium concentration decreased gradually, but in kidney remained still significantly high on day 12 and in serum reached values not different from control rats since day 6. Our data are consistent with previous pharmacokinetic studies which have shown that chromium is rapidly distributed [40] and that the half life of chromium is longer in kidney than in blood serum [41].
Figure 1 Chromium concentration in (A) serum and (B) kidney in control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. *P < 0.001 vs. control group. n = 3–6.
Markers of nephrotoxicity are shown in Figs. 2 and 3. Creatinine clearance decreased on days 1–4 (Fig. 2A), and serum creatinine and BUN increased on days 1–4 and 1–6, respectively (Figs. 2B and 2C). Urinary excretion of NAG and total protein increased on days 1–4 and 1–6, respectively (Figs. 3A and 3B). The major damage was observed on days 2–4 (Figs. 2 and 3). On day 8, all these markers returned to control values indicating that K2Cr2O7-induced ARF is reversible which is consistent with previous observations [10]. These functional findings are in a close agreement with the histological data. Significant structural abnormalities were seen in the kidney cortex since the first day after K2Cr2O7 administration. Specifically, 50% of the epithelium from the proximal convoluted tubules showed cellular swelling, necrosis and partial or complete detachment from the basal membrane. On day 2, when the peak of the tubular damage was observed, 70% of the proximal convoluted tubules was affected (Figs. 4B and 5). On days 3 and 4 proximal convoluted tubules showing the above mentioned abnormalities decreased to 40 and 25%, respectively (Fig. 5). On day 4, many tubules also showed numerous cellular debris in their lumen and epithelium regeneration was also seen but it was more evident on days 6 and 8 (data not shown). On days 10 and 12 the kidney histology was almost normal (Fig. 4C), only few tubules were revisted by active regenerative epithelium (small cuboidal cells with big nucleus and occasional mitotic figures). The high levels of serum and kidney chromium correlated with the structural and functional renal alterations on days 1–6. Interestingly, the renal damage disappeared in spite of the kidney still having high levels of chromium on days 8–12. This may be explained by the fact that chromium (VI), which is readily taken up into tissues, is reduced inside the cell to the final stable product chromium (III) [19]. The biological effects of chromium (VI) are generally attributed to cellular uptake, because chromium (VI), in contrast to chromium (III), is easily taken up by cells through the sulfate anion transport system [42,43]. However, once inside, chromium (VI) is reduced through reactive intermediates such as chromium (V) and chromium (IV) to the more stable chromium (III) by cellular reductants including glutathione, vitamins C and B2, and flavoenzymes [42]. Thus, the formation of chromium (III) or other intermediate oxidation states, in particular chromium (V), is believed to play a role in the biological effect of chromium (VI) compounds. In vitro studies have shown that this reduction process causes the generation of active oxygen species [44] which are involved in renal damage [15,34]. Interestingly, it has been shown that a low dose of K2Cr2O7 (10 mg/Kg) is unable to induce nephrotoxicity suggesting a threshold of this compound to induce renal damage [10]. In addition, it is know that chromium is located in vacuoles inside the proximal tubular cells which may delay the excretion of this metal [13]. In fact it has been shown that chromium remains for a long time in several tissues including kidney [13,40,41] which is consistent with our data.
Figure 2 (A) Creatinine clearance, (B) serum creatinine, and (C) blood urea nitrogen control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. *P at least <0.05 vs. control group. n = 4–6.
Figure 3 Urinary excretion of (A) NAG and (B) total protein in control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. *P at least <0.05 vs. control group. n = 4–18.
Figure 4 Representative images of histology (first row) and immunohistochemical detection of 3-NT (second row) and DNP (third row) during the evolution of the kidney damage produced by K2Cr2O7. The study was performed in control rats (day 0) and on days 2 and 12 after a single injection of K2Cr2O7 (15 mg/Kg). (A) Normal kidney histology from control rat (H/E). (B) On day 2, there is extensive tubular damage manifested by swollen and necrotic epithelial cells (H/E). (C) total kidney regeneration has been produced on day 12 post K2Cr2O7 administration (H/E). (D) Normal kidney from control animals shows scarce 3-NT immunoreactivity. (E) On day 2, there is strong 3-NT immunostaining in the epithelium from the convoluted tubules. (F) Slight 3-NT immunostaining is observed in the tubular epithelium on day 12. (G) Normal kidney from control rat shows scarce DNP immunostaining in the tubular epithelium. (H) In contrast, on day 2 there is clear DNP immunostaining in the cytoplasm and nucleus of the tubular epithelial cells and mesangium. (I) Slight DNP immunostaining is observed on day 12. 400 X.
Figure 5 Quantitative histological analysis in control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. *P < 0.05 vs. control group. n = 20 tubules/rat and 5 rats/group.
The histological abnormalities correlated with renal functional alterations (days 1–6) and with the immunohistochemical detection of 3-NT and protein carbonyls. DNPH reacts with free carbonyls forming DNP-derived proteins which can be detected using antibodies against DNP. Carbonyl derivatives are formed by ROS mediated oxidation of side-chains of some amino acid residues and are important detectable markers of oxidative damage to proteins [45]. Kidneys from control non-treated animals did not show or only had slight immunostaining to 3-NT or protein carbonyls (Figs. 4D and 4G), while kidneys from K2Cr2O7-treated animals on day 2 showed strong 3-NT immunostaining in necrotic and swollen tubular epithelial cells, as well as some mesangial cells (Fig. 4E). This increase was significative (P < 0.0001 vs. day 0) (Table 1). In fact, it was observed from day 1 to day 8 (data not shown). Protein carbonyls immunostaining was also clear in the cytoplasm and nuclei from tubular epithelium and mesangial glomerular cells at the same time points (Fig. 4H and data not shown) (P < 0.001 vs. day 0) (Table 1). Then, on days 10 to 12, a striking decrease of 3-NT and DNP immunostaining was seen, being similar to the control non-treated rats (Figs. 4F and 4I). In previous works, we also have found in K2Cr2O7-treated rats an intense 3-NT immunostaining [15] and an increase in protein carbonyl content [15,34]. Our data confirm that ROS are involved in K2Cr2O7-induced nephrotoxicity [16,17,19,26,44] which has additionally been supported by the protective effect of several antioxidants in this experimental model [21-25].
Table 1 Quantitative analysis of 3-NT, DNP, catalase, Cu, Zn-SOD, and Mn-SOD immunostaining.
Day 0 Day 2 Day 12
3-Nitrotyrosine 100 ± 11 (4) 392 ± 33 (4)* 113 ± 13 (4)
Dinitrophenol 100 ± 11 (4) 172.03 ± 9* (4) 106.81 ± 9 (4)
Catalase 100 ± 6 (5) 82 ± 6.7 (4) 107 ± 4.9 (4)
Cu, Zn-SOD 100 ± 3 (4) 90 ± 8 (4) 91 ± 7 (4)
Mn-SOD 100 ± 5 (3) 94 ± 6 (3) 110 ± 5 (3)
Data are expressed as percentage vs. day 0. *P < 0.001 vs. day 0.
Number of rats studied are in parentheses.
On the other hand, it was found that urinary excretion of NO2-/NO3- decreased significantly in K2Cr2O7-treated rats on days 2–6 (Fig. 6). Urinary excretion of NO2-/NO3- is considered as an index of NO production [37] and therefore these data may suggest that NO production is decreased in K2Cr2O7-treated rats. However, the strong nitrosative stress observed in our rats by 3-NT immunostaining, may suggest that the decrease in urinary NO2-/NO3- excretion could be secondary to the NO consumption by its reaction with superoxide anion to generate peroxynitrite and other reactive nitrogen species involved in protein tyrosine nitration [46]. This is supported by the association between the time course alterations in 3-NT immunostaining (days 1–8) and the decrease in urinary NO2-/NO3- excretion (days 2–6).
Figure 6 Urinary excretion of NO2-/NO3- in control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. *P < 0.05 vs. Control group. n = 4–10.
Further we investigated the time response of the renal antioxidant enzymes in K2Cr2O7-treated rats. In control rats, strong CAT immunostaining was seen in the epithelial cells from proximal and distal convoluted tubules (Fig. 7A). From day 2 to day 10 after K2Cr2O7 administration, CAT immunostaining showed striking decrease, particularly in the epithelium from tubules with evident cellular damage (Fig. 7B and data not shown). Then, at day 12, when almost normal kidney histology was seen, strong CAT immunostaining was observed in a similar pattern than in the kidney from control animals (Fig. 7C). CAT activity (Fig. 8A) and content (Fig. 8B) decreased on days 2–10 and 3–10, respectively. In control animals the epithelial cells from the proximal and convoluted tubules showed strong immunoreactivity to Cu, Zn-SOD and Mn-SOD (Figs. 7D and 7G). This pattern of immunostaining did not show evident changes during the time course study (days 1–12) (Figs. 7E, 7F, 7H, 7I, and data not shown); even the damaged swollen epithelial tubular cells exhibited strong immunostaining to both SOD enzymes. These data are in agreement with kidney total SOD activity (Fig. 9) and with the protein content of Mn-SOD and Cu, Zn-SOD measured by Western blot which remained unchanged at all time points in K2Cr2O7-treated rats (data not shown). Finally, the activity of GPx and GR decreased on days 3–12 and on days 2–10, respectively (Fig. 10). It is very clear from the above data that there was a differential response of the antioxidant enzymes to K2Cr2O7 injection. Interestingly, no enzyme (SOD, CAT, GPx, and GR) was altered on day 1 when the kidney damage was very severe. Surprisingly, SOD activity and content remained essentially unchanged at all time points. This may be a consequence of the fact that both SOD enzymes were immunolocalized even in the damaged epithelial tubular cells. In contrast, the other enzymes decreased to reach their lowest values on day 6 returning then to basal values by day 12 with the exception of GPx which remained low. On days 2–6, there was some degree correlation between CAT, GPx, and GR and the markers of nephrotoxicity. However, from day 8 these enzymes remained low in spite of renal function and structure returned to control. Furthermore, the time course study suggests that the decrease in GPx, GR, and CAT activities may be secondary to the oxidative and nitrosative stress which are evident since day 1. In fact, it has been demonstrated that peroxynitrite impairs GPx activity [47] and superoxide anion inactivates GPx [48] and CAT [49]. Interestingly, GR and CAT remained low until day 10 and GPx remained low until day 12 in absence of oxidative and nitrosative stress. The reason why these enzymes remained low at the end of the study is not clear, however we may speculate that factors other than oxidative/nitrosative stress are involved in the diminution in these enzymes and/or proximal tubules have not reached yet the full capacity to synthesize these enzymes. Therefore, additional studies are required to explain why some antioxidant enzymes remained low on days 10–12 in absence of oxidative/nitrosative stress.
Figure 7 Representative immunohistochemical detection of CAT (first row), Cu, Zn-SOD (second row), and Mn-SOD (third row) during acute renal failure induced by K2Cr2O7. The study was performed in control rats (day 0) and on days 2 and 12 after a single injection of K2Cr2O7 (15 mg/Kg). (A) Normal kidney from control rats shows intense CAT immunoreactivity in the proximal and distal convoluted tubules. (B) In contrast, kidney cortex on day 2 shows evident decrease of CAT immunostaining. (C) Strong CAT immunoreactivity in the tubular epithelium is observed on day 12. (D) Normal kidney from control animals also showed strong Cu, Zn-SOD immunostaining in the cortex convoluted tubules. This strong Cu, Zn-SOD immunoreactivity is also observed on day 2 (E) and on day 12 (F). A similar pattern of Mn-SOD immunostaining is observed in normal kidney from control rat (G), and after 2 (H) and 12 (I) days of K2Cr2O7 injection. A-C and G-I 100X, D-F 400X.
Figure 8 (A) Activity and (B) content of CAT in kidney from control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. *P at least <0.05 vs. control group. n = 4–6.
Figure 9 Total superoxide dismutase activity in kidney from control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. n = 3–6.
Figure 10 Activity of (A) glutathione peroxidase and (B) glutathione reductase in kidney from control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. *P at least <0.05 vs. control group. n = 5–6.
Conclusion
The data show an association between oxidative/nitrosative stress and functional and structural renal damage induced by K2Cr2O7 in rats at different time points, but not between this damage and antioxidant enzymes.
List of abbreviations used
ARF Acute renal failure
BUN Blood urea nitrogen
CAT Catalase
Cu, Zn-SOD Copper and zinc superoxide dismutase
DNP Dinitrophenol
DNPH 2,4-dinitrophenylhydrazine
ECL Enhanced chemiluminiscence
GPx Glutathione peroxidase
GR Glutathione reductase
H2O2 Hydrogen peroxide
K2Cr2O7 Potassium dichromate
Mn-SOD Manganese superoxide dismutase
NADPH Nicotinamide adenine dinucleotide phosphate
NAG N-acetyl-β-D-glucosaminidase
NBT Nitroblue tetrazolium
NO Nitric oxide
NO2- Nitrate
NO3- Nitrite
3-NT 3-Nitrotyrosine
ROS Reactive oxygen species
SEM Standard error of the mean
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JPCH conceived and coordinated the study and wrote and edited the manuscript. DB performed histological, immunohistochemical and statistical analyses. Coordinated the study and wrote the manuscript. ONMC performed Western blot analysis. RCC performed animal studies and collected samples. RHP performed the histological and immunohistochemical analyses. NAMR performed animal studies and collected samples. PDM measured antioxidant enzyme activity. MIS performed animal studies and collected samples. ET measured urinary excretion of NO2-/NO3-. LS and MEC measured chromium concentration. MEIR coordinated the study and edited the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
(A) Body weight, and (B) urinary volume in control (○) and K2Cr2O7 (●)-treated rats. Data are mean ± SEM. *P at least <0.01 vs. control group. n = 5–18.
Click here for file
Acknowledgements
This work was supported by CONACYT (#25441) given to MEIR and DGAPA (IN210201) given to JPCH.
==== Refs
Barceloux DG Chromium J Toxicol Clin Toxicol 1999 37 173 194 10382554 10.1081/CLT-100102418
Verschoor MA Braget PC Herber RFM Zielhuis RL Zwennis WC Renal function of chrome-plating workers and welders Int Arch Occup Environ Health 1988 60 67 70 3350606 10.1007/BF00409381
Ingber A Gammelgaard B David M Detergents and bleaches are sources of chromium contact dermatitis in Israel Contact Dermatitis 1998 38 101 104 9506223
Norseth T The carcinogenicity of chromium and its salts Br J Ind Med 1986 43 649 651 3778833
Sharma BK Singhal PC Chugh KS Intravascular haemolysis and renal failure following potassium dichromate poisoing Postgrad Med J 1978 54 414 415 683912
Picaud JC Cochat P Parchoux B Berthier JC Gilly J Chareyre S Larbre F Acute renal failure in a child after chewing of match heads Nephron 1991 57 225 226 2020351
Pedraza-Chaverri J Moreno-Muñiz SI Cruz C Hernandez-Pando R Larriva-Sahd J Tapia E Urinary Angiotensin I-converting enzyme activity is increased in experimental acute renal failure Clin Invest Med 1995 18 424 434 8714786
Michael UF Logan JL Meeks LA The beneficial effects of thyroxine on nephrotoxic acute renal failure in the rat J Am Soc Nephrol 1991 1 1236 1240 1932636
Seiken G Grillo FG Schaudies RP Johnson JP Modulation of renal EGF in dichromate-induced acute renal failure treated with thyroid hormone Kidney Int 1994 45 1622 1627 7933809
Gumbleton M Nicholls PJ Dose-response and time-response biochemical and histological study of potassium dichromate-induced nephrotoxicity in the rat Food Chem Toxicol 1988 26 37 44 2894338 10.1016/0278-6915(88)90039-7
Ngaha EO Renal effects of potassium dichromate in the rat: comparison of urinary enzyme excretion with corresponding tissue patterns Gen Pharmacol 1981 12 497 500 7308739
Evan AP Dail WG The effects of sodium chromate on the proximal tubules of the rat kidney. Fine structural damage and lysozymuria Lab Invest 1974 30 704 715 4365167
Berry JP Hourdry J Galle P Lagrue G Chromium concentration by proximal renal tubule cells: an ultrastructural, microanalytical and cytochemical study J Histochem Cytochem 1978 26 651 657 690404
Franchini I Mutti A Cavatorta A Corradi A Cosi A Olivetti G Borghetti A Nephrotoxicity of chromium. Remarks on an experimental and epidemiological investigation Contrib Nephrol 1978 10 98 110 668382
Barrera D Maldonado PD Medina-Campos ON Hernández-Pando R Ibarra-Rubio ME Pedraza-Chaverri J HO-1 induction attenuates renal damage and oxidative stress induced by K2Cr2O7 Free Radic Biol Med 2003 34 1390 1398 12757849 10.1016/S0891-5849(03)00068-6
Sengupta T Chattopadhyay D Ghosh N Maulik G Chatterjee GC Impact of chromium on lipoperoxidative processes and subsequent operation of the glutathione cycle in rat renal system Indian J Biochem Biophys 1992 29 287 290 1512015
Stohs SJ Bagchi D Oxidative mechanisms in the toxicity of metal ions Free Radic Biol Med 1995 18 321 336 7744317 10.1016/0891-5849(94)00159-H
Huang YL Chen CY Sheu JY Chuang IC Pan JH Lin TH Lipid peroxidation in workers exposed to hexavalent chromium J Toxicol Environ Health A 1999 56 235 247 10706242 10.1080/009841099158088
Liu KJ Shi X In vivo reduction of chromium (VI) and its related free radical generation Mol Cell Biochem 2001 222 41 47 11678610 10.1023/A:1017994720562
Bagchi D Hassoun EA Bagchi M Stohs SJ Chromium-induced excretion of urinary lipid metabolites, DNA damage, nitric oxide production, and generation of reactive oxygen species in Sprague-Dawley rats Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 1995 110 177 187 7599967 10.1016/0742-8413(94)00093-P
Hojo Y Satomi Y In vitro nephrotoxicity induced in mice by chromium (VI): involvement of glutathione and chromium (V) Biol Trace Elem Res 1991 31 21 31 1724173
Standeven AM Wetterhahn KE Possible role of glutathione in chromium(VI) metabolism and toxicity in rats Pharmacol Toxicol 1991 68 469 476 1891442
Na KJ Jeong SY Lim CH The role of glutathione in the acute nephrotoxicity of sodium dichromate Arch Toxicol 1992 66 646 651 1482288
Sugiyama M Role of physiological antioxidants in chromium(VI)-induced cellular injury Free Radic Biol Med 1992 12 397 407 1592274 10.1016/0891-5849(92)90089-Y
Appenroth D Winnefeld K Vitamin E and C in the prevention of metal nephrotoxicity in developing rats Exp Toxicol Pathol 1998 50 391 396 9784013
Bagchi D Stohs SJ Downs BW Bagchi M Preuss HG Cytotoxicity and oxidative mechanisms of different forms of chromium Toxicology 2002 180 5 22 12324196 10.1016/S0300-483X(02)00378-5
O'Brien P Kortenkamp A Chemical models important in understanding the way in which chromate can damage DNA Environ Health Perspect 1994 102 3 10 7843120
Shi X Dalal NS Generation of hydroxyl radical by chromate in biologically relevant systems: role of Cr(VI) complex versus tetraperochromate(V) Environ Health Perspect 1994 102 231 236 7843104
Stohs SJ The role of free radicals in toxicity and disease J Basic Clin Physiol Pharmacol 1995 6 205 228 8852268
Shi XL Dalal NS Evidence for a Fenton type mechanism for the generation of OH radicals in the reduction of Cr(VI) in cellular media Arch Biochem Biophys 1990 281 90 95 2166480 10.1016/0003-9861(90)90417-W
Aiyar J Berkovits HJ Floyd RA Wetterhahn KE Reaction of chromium (VI) with glutathione or with hydrogen peroxide: identification of reactive intermediates and their role in chromium (VI)-induced DNA damage Environ Health Perspect 1991 92 53 62 1657590
Tsou TC Chen CL Liu TY Yang JL Induction of 8-hydroxydeoxyguanosine in DNA by chromium (III) plus hydrogen peroxide and its prevention by scavengers Carcinogenesis 1996 17 103 108 8565117
Liu KJ Shi X Dalal NS Synthesis of Cr(VI)-GSH, its identification and its free hydroxyl radical generation: a model compound for Cr(VI) carcinogenicity Biochem Biophys Res Commun 1997 235 54 58 9196034 10.1006/bbrc.1997.6277
Barrera D Maldonado PD Medina-Campos ON Hernández-Pando R Ibarra-Rubio ME Pedraza-Chaverri J Protective effect of SnCl2 on K2Cr2O7-induced nephrotoxicity in rats: the indispensability of HO-1 preinduction and lack of association with some antioxidant enzymes Life Sci 2003 73 3027 3041 14519451 10.1016/j.lfs.2003.06.002
Toda N Takahashi T Mizobuchi Z Fujii H Nakahira K Takahashi S Yamashita M Morita K Hirakawa M Tin chloride pretreatment prevents renal injury in rats with ischemic acute renal failure Crit Care Med 2002 30 1512 1522 12130972 10.1097/00003246-200207000-00020
Marletta MA Yoon PS Iyengar R Leaf CD Wishnok JS Macrophage oxidation of L-arginine to nitrite and nitrate: nitric oxide is an intermediate Biochemistry 1988 27 8706 8711 3242600 10.1021/bi00424a003
Moncada S Palmer RM Higgs EA Nitric oxide: physiology, pathophysiology, and pharmacology Pharmacol Rev 1991 43 109 142 1852778
Bartholomew B A rapid method for the assay of nitrate in urine using the nitrate reductase enzyme of Escherichia coli Food Chem Toxicol 1984 22 541 543 6378739 10.1016/0278-6915(84)90224-2
Granger DL Hibbs JB JrPerfect JR Durack DT Metabolic fate of L-arginine in relation to microbiostatic capability of murine macrophages J Clin Invest 1990 85 264 273 2404026
O'Flaherty EJ A physiologically based model of chromium kinetics in the rat Toxicol Appl Pharmacol 1996 138 54 64 8658513 10.1006/taap.1996.0097
Weber H Long-term study of the distribution of soluble chromate-51 in the rat after a single intratracheal administration J Toxicol Environ Health 1983 11 749 764 6620410
De Flora S Wetterhahn KE Mechanism of chromium (VI) metabolism and genotoxicity Life Chem Rep 1989 7 169 244
Dartsch PC Hildenbrand S Kimmel R Schmahl FW Investigations on the nephrotoxicity and hepatotoxicity of trivalent and hexavalent chromium compounds Int Arch Occup Environ Health 1998 71 S40 S45 9827879
Sugiyama M Role of cellular antioxidants in metal-induced damage Cell Biol Toxicol 1994 10 1 22 8076219 10.1007/BF00757183
Smith MA Sayre LM Anderson VE Harris PL Beal MF Kowall N Perry G Cytochemical demonstration of oxidative damage in Alzheimer disease by immunochemical enhancement of the carbonyl reaction with 2,4-dinitrophenylhydrazine J Histochem Cytochem 1998 46 731 735 9603784
Radi R Nitric oxide, oxidants, and protein tyrosine nitration Proc Natl Acad Sci U S A 2004 101 4003 4008 15020765 10.1073/pnas.0307446101
Padmaja S Squadrito GL Pryor WA Inactivation of glutathione peroxidase by peroxynitrite Arch Biochem Biophys 1998 349 1 6 9439576 10.1006/abbi.1997.0407
Blum J Fridovich I Inactivation of glutathione peroxidase by superoxide radical Arch Biochem Biophys 1985 240 500 508 2992378 10.1016/0003-9861(85)90056-6
Rister M Baehner RL The alteration of superoxide dismutase, catalase, glutathione peroxidase, and NAD(P)H cytochrome c reductase in guinea pig polymorphonuclear leukocytes and alveolar macrophages during hyperoxia J Clin Invest 1976 58 1174 1184 825533
| 15854231 | PMC1142323 | CC BY | 2021-01-04 16:39:31 | no | BMC Nephrol. 2005 Apr 26; 6:4 | utf-8 | BMC Nephrol | 2,005 | 10.1186/1471-2369-6-4 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-331587635610.1186/1471-2202-6-33Research ArticleAtaxia and peripheral nerve hypomyelination in ADAM22-deficient mice Sagane Koji [email protected] Kazuhiro [email protected] Junko [email protected] Tomoko [email protected] Eiki [email protected] Norimasa [email protected] Mitsuhiro [email protected] Tohru [email protected] Kazuto [email protected] Takeshi [email protected] Tsukuba Research Laboratories, Eisai Co., Ltd., Tokodai 5-1-3, Tsukuba, Ibaraki, 300-2635, Japan2 Kawashima Research Laboratories, Eisai Co., Ltd., Kawashimatakehaya-machi 1, Kakamigahara, Gifu, 501-6195, Japan2005 6 5 2005 6 33 33 15 12 2004 6 5 2005 Copyright © 2005 Sagane 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
ADAM22 is a member of the ADAM gene family, but the fact that it is expressed only in the nervous systems makes it unique. ADAM22's sequence similarity to other ADAMs suggests it to be an integrin binder and thus to have a role in cell-cell or cell-matrix interactions. To elucidate the physiological functions of ADAM22, we employed gene targeting to generate ADAM22 knockout mice.
Results
ADAM22-deficient mice were produced in a good accordance with the Mendelian ratio and appeared normal at birth. After one week, severe ataxia was observed, and all homozygotes died before weaning, probably due to convulsions. No major histological abnormalities were detected in the cerebral cortex or cerebellum of the homozygous mutants; however, marked hypomyelination of the peripheral nerves was observed.
Conclusion
The results of our study demonstrate that ADAM22 is closely involved in the correct functioning of the nervous system. Further analysis of ADAM22 will provide clues to understanding the mechanisms of human diseases such as epileptic seizures and peripheral neuropathy.
==== Body
Background
ADAM (A Disintegrin And Metalloprotease) is a family of membrane-spanning multi-domain proteins containing a metalloproteinase-like domain and a disintegrin-like domain. Currently, more than 30 ADAMs have been identified in mammals. Their biological activities implicate ADAMs in fertilization, myogenesis and neurogenesis by proteolysis and adhesion. Some types of ADAM are catalytically active metalloproteases and shed the extracellular domains of membrane-bound growth factors or receptors [1,2]. For example, ADAM17 (TACE) has been shown to cleave several substrates, including tumour necrosis factor alpha[3,4], heparin-binding epidermal growth factor-like growth factor [6,7] and transforming growth factor alpha [8]. Studies of ADAM17-null mice have revealed that ADAM17 is critical in embryogenesis and plays an essential role in the supply of growth factors [6,8]. ADAMs are also involved in cell-cell or cell-matrix adhesion through their interaction with integrins or syndecans. More than 10 ADAMs have been shown to support integrin-mediated cell adhesion in vitro [9]. It has been reported that ADAM2-null and ADAM3-null male mutants are infertile and their spermatozoa lack egg-binding abilities [10-12]. Both ADAM2 and ADAM3 are not metalloproteases because they lack catalytic site sequences in their metalloprotease domain. These studies clearly showed that non-proteinase members of ADAMs also have significant roles in vivo.
We have reported the findings of ADAM11, ADAM22 and ADAM23 genes and their restricted expression in the human and murine nervous systems [13-15]. Sequence analysis suggests that they are not metalloproteases, since they all lack a catalytic motif. It has been reported that ADAM23 protein is localised to the cell surface [16], interacts with alpha-v beta-3 integrin heterodimer [17] and the disruption of ADAM23 gene in the mouse results in premature death associated with ataxia and tremor [18]. Although the cause of death in this mouse is unknown, for these phenotypes, impaired cell-cell or cell-matrix interactions in the nervous system caused by loss of ADAM23 may be responsible. ADAM22 and ADAM23 share highly homologous sequences in their extracellular domains. Especially, it is evident in their putative integrin binding loop sequences, CR(E/D)AVN(E/D)CD, which is located centre of the disintegrin domain. These findings led us to hypothesize that ADAM22 is an integrin binder and plays an important role in the nervous system, as does ADAM23. To determine the physiological functions of ADAM22, we generated and analysed Adam22 gene-targeted mice.
Results
Generation of ADAM22-deficient mice
Mice carrying a targeted mutation in their Adam22 gene were generated by homologous recombination (Fig. 1A). Correct targeting events were confirmed by Southern blot analysis (Fig. 1B). Since the termination codon was introduced in exon 8 in the pro-protein domain, only the truncated form of the ADAM22 protein would be synthesized from this targeted allele. Because a pro-protein domain is always removed in the mature functional ADAM-proteins by the Furin-like proteases and is thought to be non-functional itself, we considered that this truncated form of ADAM22 protein has no function. Absence of mature ADAM22 protein in homozygous mutants was confirmed by Western blot analysis using the specific antibody, which recognizes the cytoplasmic domain of the ADAM22 protein (Fig. 1C). Homozygous mutants showed no noticeable defects at birth and were indistinguishable from wild-type or heterozygous littermates during the first week. At postnatal day 10 (P10), most of the homozygous mutants were distinguishable by abnormalities such as reduced body weight and uncoordinated movements of their limbs. After P10, all homozygotes displayed severe ataxia (Fig. 2) and began to die. To measure the survival rate and body weight of each genotype precisely, we backcrossed heterozygous male mutants to C57BL/6 females more than 6 times. The resulting (N 6) heterozygous males and females were intercrossed and the produced offspring were analysed. The numbers of survivors of each genotype every 5 days are shown in Table 1. At P10, the ratio of each genotype was in close accordance with the Mendelian ratio (20.5% +/+, 55.1% +/-, 24.4% -/- ; n = 78). This result shows that ADAM22 is not essential for embryogenesis. At P10, the average body weight of homozygous mutants was approximately half that of wild-type and heterozygous littermates (Table 2). Homozygous mutants died one by one after P10, and all homozygotes died before P20. Of more than 100 homozygotes we have produced, none have survived more than 25 days after birth. Meanwhile, heterozygous mutants looked normal, were fertile, and survived for more than one year without obvious defects.
Histopathology of ADAM22-deficient mice
We have reported the predominant expression of ADAM22 mRNA in the human and mouse brain by Northern blot analysis [13,14]. To determine the distribution of ADAM22 transcript in the adult mouse CNS precisely, in situ hybridisation analysis was performed using 35S-labeled RNA probe. As shown in Fig. 3, ADAM22 mRNA was expressed throughout the adult mouse CNS. Strong signal was detected in the cerebellar granule cells and hippocampal formation. In the spinal cord, hybridisation signal was restricted to the grey matter. The distribution of ADAM22 transcripts in the CNS was quite similar with that of ADAM11, whose neuronal expression has been reported [19]. These results suggest that the ADAM22 mRNA expression is neuronal in the mouse CNS.
Based on the mRNA distribution pattern, we performed immunohistopathological analysis of the cerebellum and hippocampus extensively. Despite the high level of expression of ADAM22 mRNAs in the cerebellar granule cells, granule cell layer was normally formed and Purkinje cell morphology (calbindin-staining) of the mutant mouse looked intact (Figs. 4A–D). Hippocampal formation was also normally formed (Figs. 4I–J). These results suggest that neuronal cell migration was not impaired in the mutant mouse. Myelin-formation detected by MBP (myelin basic protein) staining of the mutant in the cerebellum (Fig. 4H) and spinal cord (Fig. 4L) was also indistinguishable with the wild-type littermate (Figs. 4G,K). In summary, we could not find any signs of abnormalities in the mutant mouse brain by the light microscopic examination.
Next, the spinal cord and peripheral nerves of each genotype were analysed by toluidine blue stain, which reveals myelin formation. Surprisingly, lack of myelin or thin myelin was observed in the sciatic nerves (Fig. 5B) and trigeminal nerves (Fig. 5D) in homozygous mutants. Because no lesions were observed in the spinal cord (Fig. 5F), it was suggested that Schwann cell specific myelination defect occurred in the homozygous mice. To analyse the state of myelinating Schwann cells and axons, electron microscopic (EM) analysis of the sciatic nerve was performed (Fig. 6). Schwann cells formed thin or no myelin in the homozygous mutant (Fig. 6B). In contrast, heterozygous mice showed complete myelination (Fig. 6A). Morphology of the axons looked normal in each genotype. Immunohistochemical analysis of the sciatic nerve showed that increased number of nuclei (Fig. 7B) and reduced staining of MBP (Fig. 7D) were remarkable in the homozygous mutant. The Schwann cell marker, S100 staining signal was intensely observed in the homozygote as well as wild-type (Figs. 7E, F). These results suggest that proliferation of the Schwann cells was not impaired but differentiation is severely delayed in the ADAM22-deficient mice.
Novel ADAM22 transcript variants isolated from the peripheral nervous system
In the case of human ADAM22, isolation of five splicing variants and existence of two terminating exons have been reported [13,20,21]. However, the latter terminating exon (exon 31) of the mouse species has not been identified yet. To isolate the long form of mouse ADAM22 transcripts, at first, we sought the mouse genome by BLAST homology search. Search results showed that the mouse genome contig NT_039297 contained most of the mouse Adam22 gene. In this contig sequence, we found putative exons 30 and 31, those sequences were quite similar to human ADAM22 isoform 1 (GenBank # NM_021723). To confirm existence of transcripts, we designed primers on the putative exon 31 and performed RT-PCR using mouse cerebellum mRNA as a template. TA-cloning of the amplified fragments and sequencing analysis showed that 5 of 8 clones were type 1 isoform (G01), 2 clones were type 3 isoform (G03) and 1 clone (G08) contained a novel exon between exons 29 and 30. Next, we performed RT-PCR analysis of mRNAs purified from several adult mouse tissues, such as the spinal cord, dorsal root ganglion (DRG), sciatic nerve and cultured primary Schwann cells. Interestingly, multiple DNA fragments were amplified from each tissue (Fig. 8), and length of major fragments in each lane was different. For example, length of major amplified PCR fragments from the sciatic nerve (Fig. 8A, lane 5) and cultured Schwann cells (lane 6) were almost identical, and were shorter than those from other tissues. In contrast, major fragments from the DRG (lane 4) were longer than others. We performed TA-cloning and sequencing to analyse exon organization of the amplified transcripts (Fig. 8C). Forty seven clones were isolated in total and were classified in 16 independent clones (G01 – G23). Comparison with the genomic sequence revealed the existence of 8 novel exons (exons 27S, 27L, 29.1, 29.3, 29.5, 29.7, 30 and 31). These novel exons were flanked by well-defined introns that obey the GT-AG rule. Exon organization of each clone was shown in Fig. 8C. Characteristic feature of ADAM22 transcripts is tissue specific insertion or skipping of exons. Because the number of nucleotides of exons 26, 27, 27L, 29.3, 29.5 and 29.7 is a multiple of 3, insertion or skipping of these exons will not change the downstream reading frame. Known peptide motives were not found in the newly isolated sequences. This RT-PCR analysis is summarized as follows: ADAM22 transcripts were roughly subdivided in 3 groups, CNS-form (containing exon 27), DRG-form (containing exon 27L) and Schwann-cell-form (skipping exon 26 and 27). From these results, we concluded that ADAM22 is expressed in a cell-type specific manner, and plays an essential role in myelinogenesis in the PNS.
Discussion
In the present study, we examined the function of ADAM22 by generating mice that lacked the Adam22 gene. ADAM22-deficient mice exhibited severe ataxia and premature death. In contrast, heterozygous mutants were healthy, fertile and survived more than 1 year without obvious abnormalities. The cause of death in the homozygous mutants is not clear; however, it is likely to be due to multiple seizures. This is because convulsive seizures were occasionally observed in homozygotes, after which the suffered mice appeared exhausted and, in most cases, died on the next day. It is known that cortical dysplasia resulting from aberrant brain development causes seizure syndrome [22]. For example, mice lacking the neuron-specific activator of cyclin 5, p35, exhibit seizures and a severe neuronal migration defect [23]. Because ADAM22 protein is expressed on the cell surface and is likely to be involved in cell-cell or cell-matrix interactions, we hypothesized that depletion of ADAM22 would generate aberrant neuronal migration. Histochemical analysis was performed to verify this hypothesis. However, no marked abnormalities were observed in the mutant mouse brain, ruling out a critical role for ADAM22 in neuronal migration. Another possible explanation is that the premature death in first 2 weeks of life is caused by the dysfunction of the autonomic nervous system. Especially, nerves which control the breathing would be very important, because the respiration system undergoes significant maturation in the first 2–3 weeks after birth. For example, the myelin-deficient (MD) rat, which carries a point mutation in the proteolipid protein (PLP) gene, exhibits ataxia, tremor, dysmyelination and dies at approximately postnatal day 21. In MD rat, early death is caused by the dysfunction of the brain stem which is essential for autonomic control of respiration during hypoxia [24]. Further study is needed to verify the cause of death in ADAM22-deficient mice.
Another interesting phenotype observed in ADAM22-deficient mice was hypomyelinogenesis. Histopathological analysis revealed marked hypomyelination in the sciatic and trigeminal nerves. However, myelin formation in the spinal cord and the brain was normal. It was therefore evident that hypomyelinogenesis occurred specifically in the peripheral nerves in the knockout mice, suggesting Schwann cell dysfunction. In the mutant sciatic nerve, the density of pro-myelinating Schwann cells (S100-positive, MBP-negative) was markedly higher than that seen in heterozygotes. Higher pro-myelinating Schwann cell density and marked delay in myelin formation indicate that ADAM22 plays an important role in Schwann cell differentiation, but not in proliferation.
Investigations on why hypomyelinogenesis occurred in ADAM22-deficient mice centre on integrin. This is because several ADAMs have been shown to interact with integrins via disintegrin domains [9], and it has been suggested that alpha-6 beta-1 integrin on the Schwann cell plays an important role in myelin formation [25]. In addition, laminin-2 is a key player in peripheral myelin formation, because laminin-2 deficiency causes dysmyelination in mice and humans, and its receptors, alpha-6 beta-1 integrin and dystroglycan, are also shown to be deeply involved in myelin formation [26,27]. The conditional knockout studies of beta-1 integrin and laminin gamma-1 in Schwann cell suggested that both laminin-2 and integrins are essential for segregation of large bundles of axons in the early stage of myelin formation [25,28]. In contrast, the phenotype seen in the ADAM22-deficient mice was quite different: for example, unsorted bundles of axons were not seen in nerves and roots. We assume that ADAM22 partially modulates integrin-mediated signals or is involved at a later stage of myelin formation.
Conclusion
In summary, we generated ADAM22-deficient mice and proved that ADAM22 plays an essential role in both the CNS and the PNS. The results of this study suggest that ADAM22 is involved in human diseases such as epilepsy and peripheral neuropathy. The human ADAM22 gene is relatively large, at 300 kb in length, and is comprised of more than 30 exons. It is assigned on 7q21, where several chromosomal aberrations that are accompanied by neurological disorders have been identified [29,30]. Further detailed molecular function analysis of ADAM22 may lead to progress in uncovering the mechanisms that underlie certain human neurological diseases.
Methods
Construction of an Adam22 gene-targeting vector
The 18-kb genomic DNA fragments covering exons 5, 6, 7, 8 and 9 of the Adam22 gene were amplified from C57BL/6 genomic DNA by PCR using Expand Hi-Fidelity enzyme mix (Roche Diagnostics) and primers designed for each exon. To generate a mutation in the mouse Adam22 gene, we inserted the PGKneo cassette into exon 8 of the Adam22 gene. The final targeting construct consisted of a 10.5-kb genomic DNA fragment that was interrupted by the insertion of the PGKneo cassette and contained a MC1/TK cassette as a negative selection marker against random integration [5].
Generation of Adam22 deficient mice
The linearised targeting vector was electroporated into TT2 embryonic stem (ES) cells [31]. Homologous recombinants were selected by PCR. The extracted genomic DNA from each clone was amplified using the forward primer AGN2: 5'-GCCTGCTTGCCGAATATCATGGTGGAAAAT-3' in the PGKneo cassette, and the reverse primer MFP065R: 5'-ACTATTTCTGTGATGAGGGCACAGCATC-3' outside the targeting vector. Homologous recombined DNA was efficiently amplified by 35 cycles of the following amplification steps (94°C-30 s and 68°C-5 min). The targeting efficiency of this construct was about 4%. Correctly targeted ES clones were injected into fertilized ICR mouse eggs at the eight-cell stage. The resulting male chimeras were mated with C57BL/6N females, and heterozygous male and female mice were interbred to generate homozygous mice. The ataxic phenotypes of homozygous mice were observed in two independent lines.
Southern blot analysis
Mouse genomic DNA used for Southern blot analysis was extracted from the liver of each genotype. BamHI-digested genomic DNA was probed with the [32P]-labelled 0.4-kb SpeI-BamHI genomic fragment, which is located between exons 8 and 9.
Antibody production
His-tagged and MBP-fused recombinant protein containing cytoplasmic domain (cp) of human ADAM22 isoform 1 (position: 756–906, 151 amino acids in length) were produced in E. coli. His-tagged ADAM22-cp protein was purified in denatured condition using Ni-NTA resin (Qiagen GmbH) and dialysed in PBS. Precipitated protein was recovered and was mixed with Freund's complete adjuvant (Invitrogen), then, the mixture was used for immunization of rabbits. The antiserum raised against the His-tagged ADAM22-cp protein was incubated with MBP-fused ADAM22-cp protein coupled to Affi-Gel 10 beads (Bio-Rad). Beads with bound antibodies were washed in PBS, and the bound antibodies were eluted with 100 mM glycine, pH 3.0, and neutralized immediately. Using the affinity purified antibody, both human and mouse ADAM22 proteins were detected by Western blot analysis. However, unfortunately the antibody was not suitable for immunohistochemical analysis of mouse tissues.
Western blot analysis
The absence of ADAM22 protein in the Adam22 -/- mice was confirmed by Western blot analysis. Briefly, the cerebellum was isolated from a P14.5 mouse of each genotype and homogenized with a Polytron homogenizer in cell lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1% NP-40, Complete protease inhibitors [Roche Diagnostics]). After removal of cell debris by centrifugation, the supernatant was separated on 10% SDS-PAGE, and transferred to a nitrocellulose membrane. The blot was then incubated with polyclonal antibody against the cytoplasmic domain of ADAM22 (at 1:1,000). Bound antibodies were visualized with horseradish peroxidase-labelled second antibody and a ECL-plus chemiluminescence detection system (Amersham Biosciences Corp.).
Primary Schwann cell culture
Primary Schwann cells were prepared from 4-month-old C57BL/6 mice according to Manent's protocol [32] with minor modifications. Briefly, sciatic nerves were removed and incubated for 7 days in the pre-treatment medium, which consisted of D-MEM (high glucose) supplemented with 10% FCS, 50 mg/ml gentamicin (Invitrogen Corp.), 2.5 μg/ml fungizone (Invitrogen Corp.), 2 μM forskolin (EMD Biosciences Inc.) and 10 ng/ml recombinant heregulin-beta1 (R&D Systems). For dissociation, cultured sciatic nerves were cut into pieces and incubated at 37°C for 3 hours in Opti-MEM medium (Invitrogen Corp.) containing 130 U/ml collagenase type I (Invitrogen Corp.) and 0.4 U/ml dispase II (Roche Diagnostics). Dissociated cells were resuspended in the pre-treatment medium and plated on Poly-D-Lysine/Laminin coated plate (BD Biosciences). The purity of the cultured Schwann cells, as determined by indirect immunofluorescence analysis, approached 90 %.
RT-PCR analysis
Adult C57BL/6 male mice were used in this study. Total RNAs purified from the cerebellum, spinal cord, sciatic nerve, DRG and cultured Schwann cells using TRIzol (Invitrogen Corp.) and RNeasy kit (Qiagen GmbH) were analysed by RT-PCR using SuperScript II and random primer (Invitrogen Corp.), and PCR amplification (40 cycles; 94°C-30 s, 60°C-30 s and 68°C-5 min) with Expand Hi-Fidelity DNA polymerase (Roche Diagnostics) and ADAM22-specific primers. To detect splicing variants in the cytoplasmic domain, a forward primer was placed just upstream of the transmembrane domain and a reverse primer was set on the terminating exon. The primers used in this study were as follows. ISH04-forward: 5'-AACAGGCACTGGACAGGGGCTGAC-3' and ISH04-reverse: 5'-AATGGATGTCTCCCATAGCCTGGC-3'.
Histopathology
Mice were anesthetized with ethyl ether. Whole-body perfusion by 2% paraformaldehyde-glutaraldehyde solution followed by heparin-included saline was performed. Sciatic nerves, trigeminal nerves, brain and spinal cord were removed and fixed in 10% neutral-buffered formalin. The spinal cord and other nerve blocks were washed and postfixed with 2% osmium tetroxide, and were dehydrated in ethanol and equilibrated in Epon. Epon embedded semithin sections were stained with toluidine blue and were subjected to light microscopic examination. For electron microscopic analysis, thin sections were cut using an ultramicrotome, stained with 1.5% uranylacetate in 50% ethanol and 0.8% lead citrate, and analysed using electron microscope. All procedures were conducted according to the Eisai Animal Care Committee's guidelines.
Immunohistochemistry
Frozen sections were rinsed in 0.1% Triton X-100/PBS at room temperature for 1 hour, and were blocked in BLOCKACE solution (Dai-nippon). The following antibodies were incubated overnight in 0.1% BLOCKACE at 4°C:
monoclonal mouse anti-calbindin 28 K (Sigma, 1:200); monoclonal mouse anti-MBP (SMI-99, Sternberger Monoclonals Inc., 1:50); monoclonal mouse anti-Neu N (CHEMICON, 1:100); rabbit anti-S100 polyclonal (DAKO, 1:200). Nuclei were counterstained with 1 μg/ml DAPI (Sigma). MBP staining was performed after microwave epitope retrieval in 10 mmol/L citrate buffer (pH 6). Sections were incubated for 1 hour with secondary antibodies: Cy3-labeled donkey anti-rabbit IgG (Jackson Immuno Research, 1:200), FITC-labelled donkey anti-mouse IgG (Jackson Immuno Research, 1:200). Sections were photographed with an Olympus microscope (Olympus IX71) equipped with a high-resolution CCD camera.
RNA in situ hybridisation
In situ hybridisation on frozen sections were carried out using 35S-labelled RNA probes. Briefly, 610 bp of mouse ADAM22 cDNA (position: 1351–1960 from initiating ATG) was cloned into the pBluescript II SK(+) or the pBluescript II KS(+) vector. Using these plasmids as templates, sense and antisense labelled RNA probes were generated by T7 RNA polymerase and [α35S]UTP. Frozen brain and spinal cord from 2 month-old C57BL/6 female mice were used in this analysis. Pretreatment, hybridisation, RNase treatment and washing was carried out following the protocol described in the literature [33]. Dehydrated slides were attached to imaging plates for 48 hours and the autoradiograms were analysed using a Bio-Image Analyser (BAS3000, Fuji Photo Film).
Authors' contributions
KS is leading this project, and performed cDNA cloning, plasmid construction, antibody production, immunohistochemical study, and wrote the manuscript. KH and JK conducted histopathological experiments. TH analysed mRNA distribution by in situ hybridisation. ET, NM, MI, TO and KY contributed to the generation of knockout mice. TN supervised the project. All authors read and approved the manuscript.
Acknowledgements
We are grateful to Isao Tanaka, Masayuki Okada, Yoshiharu Mizui, Kappei Tsukahara, Hiroyuki Kato, Hiroo Ogura and Junro Kuromitsu (Eisai Co., Ltd.), and Shin-ichi Murase (Virginia University) for the fruitful discussions and advice.
Figures and Tables
Figure 1 Targeted mutation of Adam22. (A) The genomic structure of the wild-type Adam22 allele (top), the targeting construct (middle) and the disrupted allele (bottom) are shown. ADAM22 expression was disrupted by the insertion of a termination codon and a PGKneo cassette into exon 8. An MC1/TK cassette at the end of the targeting vector allows for negative selection. The 3' probe represents the position of the external probe used for Southern blot analysis, and expected BamHI [B] fragments are indicated by arrows. (B) Southern blot analysis of mouse genomic DNA. The expected DNA fragments for the wild-type allele and disrupted allele are 7.5-kb and 2.5-kb, respectively. +/+, wild-type; +/-, heterozygote; -/-, homozygote. (C) Western blot analysis of ADAM22 expression in the mouse cerebellum. Absence of ADAM22 protein in the Adam22 (-/-) mutant cerebellum was shown using anti-ADAM22-cp (cytoplasmic domain) polyclonal antibody.
Figure 2 Uncoordinated movements in the Adam22 (-/-) mice at postnatal day 18 (P18). (A) Adam22 (-/-) mice were smaller than their wild-type (+/+) littermates. (B, C) Adam22 (-/-) mice at P18 were unable to support themselves on their hindlimbs.
Figure 3 Neuronal ADAM22 mRNA expression in the CNS. To determine the ADAM22 mRNA distribution, in situ hybridisation analysis using 35S-labeled probe was performed. Coronal (A, B) and sagittal (C, D) sections of the mouse brain and spinal cord (E) were shown. Using the antisense probe (A, C), strong signals were obtained, especially in the hippocampus and the cerebellum, while no signals was detected by the sense probe (B, D). In the spinal cord, autoradiograms of ADAM22 mRNA was detected in the grey matter (E).
Figure 4 Normal neurodevelopment in the CNS of ADAM22-deficient mice. Sagittal sections of the cerebellum from wild-type mice (A, C, E, G) and homozygous mutants (B, D, F, H) at postnatal day 13 were stained for calbindin (C and D; green) or MBP (G and H; green), and were counterstained with DAPI (A, B, E, F; blue). Significant abnormalities were not observed in the homozygotes. Hippocampal neurons were stained by anti-Neu N antibody (I and J). Spinal myelin was analysed by MBP staining (K and L). These analyses showed no obvious differences between homozygotes (J, L) and wild-type littermates (I, K). Bar: (A, B, E, F) 100 μm, (K, L) 250 μm.
Figure 5 Hypomyelination of peripheral nerves in ADAM22-deficient mice. Epon embedded semithin cross-sections of the sciatic nerves (A, B), the trigeminal nerves (C, D) and the spinal cord [lateral funiculus] (E, F) of the indicated genotypes at postnatal day 10 were stained with toluidine blue. Note that the ADAM22-deficient mouse shows thin myelin or lack of myelin in the peripheral nerve fibres (B, D), but the spinal myelinated fibres are intact (F).
Figure 6 Electron microscopic analysis of sciatic nerves. Electron micrographs of the sciatic nerves from Adam22 +/- (A) and Adam22 -/- (B) mice at postnatal day 10 are shown. In the heterozygote (A), thick myelin was formed, while no myelin was formed in the ADAM22-deficient mouse (B). The axons looked normal in each genotype.
Figure 7 Marked delay in differentiation of the mutant Schwann cells. Transverse sections of the sciatic nerves of the indicated genotypes at postnatal day 13 were stained for MBP (C and D; green) or S100 (E and F; red), and were counterstained with DAPI (A, B; blue). In the mutant sciatic nerve, density of DAPI-signals were greatly increased (B) compared with those of the wild-type (A). Bar: (A, B) = 50 μm.
Figure 8 Adam22 gene structure and tissue specific transcripts. (A) RT-PCR analysis. Amplified DNA fragments were analysed by 1 % agarose gel electrophoresis. Lanes 1. 100 bp DNA ladder; 2. cerebellum; 3. spinal cord; 4. dorsal root ganglion; 5. sciatic nerve; 6. cultured Schwann cells; 7. distilled water (B) Exon organization of the mouse Adam22 gene is illustrated. Boxes indicate exons. The G01 transcript (orthologous to the human ADAM22 isoform 1 transcript) is composed of grey boxes. Boxes in black indicate non-coding region. (C) Summary of the isolated clones. The nucleotide sequence data have been deposited with the DDBJ/EMBL/GenBank Data Libraries under the accession numbers described in the table. (D) Number of clones isolated from each tissue is summarized. Cb; cerebellum, Sp; spinal cord, DRG; dorsal root ganglion, SN; sciatic nerve, cSC; cultured Schwann cells.
Table 1 Genotyping of the progeny of Adam22 heterozygous intercrosses.
Adam22 genotype
+/+ +/- -/- total
Day 10 16 43 19 78
Day 15 16 43 8 67
Day 20 16 43 0 59
From heterozygous intercrosses, 78 mice were produced. Numbers of mice in each genotype were counted on postnatal day 10, 15 and 20. No homozygotes were survived more than postnatal day 20. +/+, wild-type; +/-, heterozygote; -/-, homozygote.
Table 2 Body weight of progeny at postnatal day 10.5
genotype body weight (g) no. of pups
+/+ 6.08 ± 0.61 9
+/– 5.84 ± 0.81 16
–/– 3.12 ± 0.48 7
+/+, wild-type; +/-, heterozygote; -/-, homozygote.
==== Refs
Becherer JD Blobel CP Biochemical properties and functions of membrane-anchored metalloprotease-disintegrin proteins (ADAMs) Curr Top Dev Biol 2003 54 101 123 12696747
Seals DF Courtneidge SA The ADAMs family of metalloproteases: multidomain proteins with multiple functions Genes Dev 2003 17 7 30 12514095
Black RA Rauch CT Kozlosky CJ Peschon JJ Slack JL Wolfson MF Castner BJ Stocking KL Reddy P Srinivasan S Nelson N Boiani N Schooley KA Gerhart M Davis R Fitzner JN Johnson RS Paxton RJ March CJ Cerretti DP A metalloproteinase disintegrin that releases tumour-necrosis factor-alpha from cells Nature 1997 385 729 733 9034190
Moss ML Jin SL Milla ME Bickett DM Burkhart W Carter HL Chen WJ Clay WC Didsbury JR Hassler D Hoffman CR Kost TA Lambert MH Leesnitzer MA McCauley P McGeehan G Mitchell J Moyer M Pahel G Rocque W Overton LK Schoenen F Seaton T Su JL Becherer JD Cloning of a disintegrin metalloproteinase that processes precursor tumour-necrosis factor-alpha Nature 1997 385 733 736 9034191
Ino M Yoshinaga T Wakamori M Miyamoto N Takahashi E Sonoda J Kagaya T Oki T Nagasu T Nishizawa Y Tanaka I Imoto K Aizawa S Koch S Schwartz A Niidome T Sawada K Mori Y Functional disorders of the sympathetic nervous system in mice lacking the alpha 1B subunit (Cav 2.2) of N-type calcium channels Proc Natl Acad Sci U S A 2001 98 5323 5328 11296258
Jackson LF Qiu TH Sunnarborg SW Chang A Zhang C Patterson C Lee DC Defective valvulogenesis in HB-EGF and TACE-null mice is associated with aberrant BMP signaling Embo J 2003 22 2704 2716 12773386
Sunnarborg SW Hinkle CL Stevenson M Russell WE Raska CS Peschon JJ Castner BJ Gerhart MJ Paxton RJ Black RA Lee DC Tumor necrosis factor-alpha converting enzyme (TACE) regulates epidermal growth factor receptor ligand availability J Biol Chem 2002 277 12838 12845 11823465
Peschon JJ Slack JL Reddy P Stocking KL Sunnarborg SW Lee DC Russell WE Castner BJ Johnson RS Fitzner JN Boyce RW Nelson N Kozlosky CJ Wolfson MF Rauch CT Cerretti DP Paxton RJ March CJ Black RA An essential role for ectodomain shedding in mammalian development Science 1998 282 1281 1284 9812885
White JM ADAMs: modulators of cell-cell and cell-matrix interactions Curr Opin Cell Biol 2003 15 598 606 14519395
Cho C Bunch DO Faure JE Goulding EH Eddy EM Primakoff P Myles DG Fertilization defects in sperm from mice lacking fertilin beta Science 1998 281 1857 1859 9743500
Nishimura H Cho C Branciforte DR Myles DG Primakoff P Analysis of loss of adhesive function in sperm lacking cyritestin or fertilin beta Dev Biol 2001 233 204 213 11319869
Shamsadin R Adham IM Nayernia K Heinlein UA Oberwinkler H Engel W Male mice deficient for germ-cell cyritestin are infertile Biol Reprod 1999 61 1445 1451 10569988
Sagane K Ohya Y Hasegawa Y Tanaka I Metalloproteinase-like, disintegrin-like, cysteine-rich proteins MDC2 and MDC3: novel human cellular disintegrins highly expressed in the brain Biochem J 1998 334 ( Pt 1) 93 98 9693107
Sagane K Yamazaki K Mizui Y Tanaka I Cloning and chromosomal mapping of mouse ADAM11, ADAM22 and ADAM23 Gene 1999 236 79 86 10433968
Novak U ADAM proteins in the brain J Clin Neurosci 2004 11 227 235 14975408
Goldsmith AP Gossage SJ ffrench-Constant C ADAM23 is a cell-surface glycoprotein expressed by central nervous system neurons J Neurosci Res 2004 78 647 658 15505805
Cal S Freije JM Lopez JM Takada Y Lopez-Otin C ADAM 23/MDC3, a human disintegrin that promotes cell adhesion via interaction with the alphavbeta3 integrin through an RGD-independent mechanism Mol Biol Cell 2000 11 1457 1469 10749942
Mitchell KJ Pinson KI Kelly OG Brennan J Zupicich J Scherz P Leighton PA Goodrich LV Lu X Avery BJ Tate P Dill K Pangilinan E Wakenight P Tessier-Lavigne M Skarnes WC Functional analysis of secreted and transmembrane proteins critical to mouse development Nat Genet 2001 28 241 249 11431694
Rybnikova E Karkkainen I Pelto-Huikko M Huovila AP Developmental regulation and neuronal expression of the cellular disintegrin ADAM11 gene in mouse nervous system Neuroscience 2002 112 921 934 12088751
Poindexter K Nelson N DuBose RF Black RA Cerretti DP The identification of seven metalloproteinase-disintegrin (ADAM) genes from genomic libraries Gene 1999 237 61 70 10524237
Harada T Nishie A Torigoe K Ikezaki K Shono T Maehara Y Kuwano M Wada M The specific expression of three novel splice variant forms of human metalloprotease-like disintegrin-like cysteine-rich protein 2 gene inBrain tissues and gliomas Jpn J Cancer Res 2000 91 1001 1006 11050470
Noebels JL The biology of epilepsy genes Annu Rev Neurosci 2003 26 599 625 14527270
Chae T Kwon YT Bronson R Dikkes P Li E Tsai LH Mice lacking p35, a neuronal specific activator of Cdk5, display cortical lamination defects, seizures, and adult lethality Neuron 1997 18 29 42 9010203
Miller MJ Haxhiu MA Georgiadis P Gudz TI Kangas CD Macklin WB Proteolipid protein gene mutation induces altered ventilatory response to hypoxia in the myelin-deficient rat J Neurosci 2003 23 2265 2273 12657685
Feltri ML Graus Porta D Previtali SC Nodari A Migliavacca B Cassetti A Littlewood-Evans A Reichardt LF Messing A Quattrini A Mueller U Wrabetz L Conditional disruption of beta 1 integrin in Schwann cells impedes interactions with axons J Cell Biol 2002 156 199 209 11777940
Scherer SS Myelination: some receptors required J Cell Biol 2002 156 13 15 11781330
Saito F Moore SA Barresi R Henry MD Messing A Ross-Barta SE Cohn RD Williamson RA Sluka KA Sherman DL Brophy PJ Schmelzer JD Low PA Wrabetz L Feltri ML Campbell KP Unique role of dystroglycan in peripheral nerve myelination, nodal structure, and sodium channel stabilization Neuron 2003 38 747 758 12797959
Chen ZL Strickland S Laminin {gamma}1 is critical for Schwann cell differentiation, axon myelination, and regeneration in the peripheral nerve J Cell Biol 2003 163 889 899 14638863
Vermeulen S Menten B Van Roy N Van Limbergen H De Paepe A Mortier G Speleman F Molecular cytogenetic analysis of complex chromosomal rearrangements in patients with mental retardation and congenital malformations: delineation of 7q21.11 breakpoints Am J Med Genet 2004 124A 10 18
Scherer SW Cheung J MacDonald JR Osborne LR Nakabayashi K Herbrick JA Carson AR Parker-Katiraee L Skaug J Khaja R Zhang J Hudek AK Li M Haddad M Duggan GE Fernandez BA Kanematsu E Gentles S Christopoulos CC Choufani S Kwasnicka D Zheng XH Lai Z Nusskern D Zhang Q Gu Z Lu F Zeesman S Nowaczyk MJ Teshima I Chitayat D Shuman C Weksberg R Zackai EH Grebe TA Cox SR Kirkpatrick SJ Rahman N Friedman JM Heng HH Pelicci PG Lo-Coco F Belloni E Shaffer LG Pober B Morton CC Gusella JF Bruns GA Korf BR Quade BJ Ligon AH Ferguson H Higgins AW Leach NT Herrick SR Lemyre E Farra CG Kim HG Summers AM Gripp KW Roberts W Szatmari P Winsor EJ Grzeschik KH Teebi A Minassian BA Kere J Armengol L Pujana MA Estivill X Wilson MD Koop BF Tosi S Moore GE Boright AP Zlotorynski E Kerem B Kroisel PM Petek E Oscier DG Mould SJ Dohner H Dohner K Rommens JM Vincent JB Venter JC Li PW Mural RJ Adams MD Tsui LC Human chromosome 7: DNA sequence and biology Science 2003 300 767 772 12690205
Yagi T Tokunaga T Furuta Y Nada S Yoshida M Tsukada T Saga Y Takeda N Ikawa Y Aizawa S A novel ES cell line, TT2, with high germline-differentiating potency Anal Biochem 1993 214 70 76 8250257
Manent J Oguievetskaia K Bayer J Ratner N Giovannini M Magnetic cell sorting for enriching Schwann cells from adult mouse peripheral nerves J Neurosci Methods 2003 123 167 173 12606065
Nakamoto H Soeda Y Takami S Minami M Satoh M Localization of calcitonin receptor mRNA in the mouse brain: coexistence with serotonin transporter mRNA Brain Res Mol Brain Res 2000 76 93 102 10719219
| 15876356 | PMC1142324 | CC BY | 2021-01-04 16:03:48 | no | BMC Neurosci. 2005 May 6; 6:33 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-33 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-341588247310.1186/1471-2202-6-34Research ArticleProtection of cortical cells by equine estrogens against glutamate-induced excitotoxicity is mediated through a calcium independent mechanism Perrella Joel [email protected] Bhagu R [email protected] Department of Obstetrics and Gynecology, University of Toronto, Toronto, Canada2 Institute of Medical Sciences, University of Toronto, Toronto, Canada3 Department of Obstetrics and Gynecology, St. Michael's Hospital, Toronto, Canada2005 10 5 2005 6 34 34 10 11 2004 10 5 2005 Copyright © 2005 Perrella and Bhavnani; 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
High concentrations of glutamate can accumulate in the brain and may be involved in the pathogenesis of neurodegenerative disorders such as Alzheimer's disease. This form of neurotoxicity involves changes in the regulation of cellular calcium (Ca2+) and generation of free radicals such as peroxynitrite (ONOO-). Estrogen may protect against glutamate-induced cell death by reducing the excitotoxic Ca2+ influx associated with glutamate excitotoxicity. In this study, the inhibition of N-methyl-D-aspartate (NMDA) receptor and nitric oxide synthase (NOS) along with the effect of 17β-estradiol (17β-E2) and a more potent antioxidant Δ8, 17β-estradiol (Δ8, 17β-E2) on cell viability and intracellular Ca2+ ([Ca2+]i), following treatment of rat cortical cells with glutamate, was investigated.
Results
Primary rat cortical cells were cultured for 7–12 days in Neurobasal medium containing B27 supplements. Addition of glutamate (200 μM) decreased cell viability to 51.3 ± 0.7% compared to control. Treatment with the noncompetitive NMDAR antagonist, MK-801, and the NOS inhibitor, L-NAME, completely prevented cell death. Pretreatment (24 hrs) with 17β-E2 and Δ8, 17β-E2 (0.01 to 10 μM) significantly reduced cell death. 17β-E2 was more potent than Δ8, 17β-E2. Glutamate caused a rapid 2.5 fold increase in [Ca2+]i. Treatment with 0.001 to 10 μM MK-801 reduced the initial Ca2+ influx by 14–41% and increased cell viability significantly. Pretreatment with 17β-E2 and Δ8, 17β-E2 had no effect on Ca2+ influx but protected the cortical cells against glutamate-induced cell death.
Conclusion
Glutamate-induced cell death in cortical cultures can occur through NMDAR and NOS-linked mechanisms by increasing nitric oxide and ONOO-. Equine estrogens: 17β-E2 and Δ8, 17β-E2, significantly protected cortical cells against glutamate-induced excitotoxicity by a mechanism that appears to be independent of Ca2+ influx. To our knowledge, this is a first such observation. Whether the decrease in NOS related products such as ONOO-, is a mechanism by which estrogens protect against glutamate toxicity, remains to be investigated. Estrogen replacement therapy in healthy and young postmenopausal women may protect against neurodegenerative diseases by these mechanisms.
==== Body
Background
A number of neurodegenerative diseases, including Alzheimer's disease (AD) have been proposed to involve a dysregulation in the brain's glutamatergic system [1-4]. In addition, high (mM) glutamate concentrations have been documented to cause neuronal degeneration in various in vivo and in vitro models [1]. Although glutamate-induced cell death is associated with both apoptotic and necrotic changes [5] the mechanism of cell death remains to be established.
Two distinct pathways for glutamate-induced cell death have been described: the excitotoxic pathway and the oxidative pathway. The excitotoxic pathway involves the overactivation of glutamate receptors that leads to both rapid and slowly triggered cytotoxic events. The rapid effects involve the activation of the N-Methyl-D-Aspartate receptor (NMDAR) that lead to a large Ca2+ influx that may be detrimental to cell viability [6]. The oxidative pathway involves the breakdown of the glutamate-cystine antiporter and a drop in glutathione levels that allows for aberrant formation of reactive oxygen species (ROS) that are neurotoxic [7,8].
Although intracellular Ca2+ ([Ca2+]i) is necessary for a number of physiological processes, excessive amounts may lead to neuronal dysfunction and cell death. Neuronal increases in Ca2+ can activate a number of enzymes, such as phospholipases, proteases, endonucleases and nitric oxide synthase (NOS). Increase in the activity of these enzymes is associated with neuronal cell death [9,10]. Excessive glutamatergic stimulation is also associated with an increase in [Ca2+]i required for neuronal NOS (nNOS) activation and nitric oxide (NO) production within the neuron and this can result in increased cell death [9,11,12]. Therefore, the maintenance of proper Ca2+ homeostasis may be effective in preventing the progression of glutamate associated neuronal degeneration.
Previous studies demonstrate that estrogens are neuroprotective against the oxidative pathway of glutamate-induced cell death in a mouse hippocampal cell line, HT22 [13,14]. This finding, along with other studies describing estrogen's neuroprotective and neurotrophic action [13,15,16] support results from observational studies that exogenous estrogen use by postmenopausal women is neuroprotective [17-20].
Although estrogens are potent antioxidants [21,22] that can prevent oxidative damage in cell culture systems, recent evidence also suggests that estrogen may inhibit glutamate-induced excitotoxicity [23]. Moreover, estrogen can lower cytotoxic Ca2+ influxes induced by glutamate in hippocampal cells [24,25]. Whether this Ca2+ lowering, or buffering effect occurs in other neuronal cell types, such as cortical cells, that are known to be estrogen sensitive, and whether this effect is involved in estrogen's neuroprotective effects remains to be established.
In this study, the effect of 17β-estradiol (17β-E2) and a novel ring-B unsaturated equine estrogen with less feminizing effects [26,27] and greater antioxidant potential, Δ8, 17β-estradiol (Δ8, 17β-E2) [21,22] on cell viability and [Ca2+]i, following treatment of rat cortical cells with glutamate was studied. The concentrations of glutamate used in the present study are relatively lower than those used in previous studies where cell death may have occurred mainly via the oxidative pathway. It is hypothesized that both estrogens will protect cortical cells from glutamate induced excitotoxicity by modifying the initial Ca2+ influx in addition to acting as antioxidants.
Results
Purity of neuronal cultures
Freshly isolated embryonic cortical cells obtained from day 17–19 pregnant rats were cultured and maintained in serum-free neurobasal medium containing B27 supplements for 7 days as described under "Materials and Methods". Phase contrast microscopy indicated that the cells had characteristic morphology of neurons and their cellular extensions (dendrites) were clearly visible (Figure 1B). Immunocytochemistry using anti-rat neuronal specific enolase (NSE), and anti-glial fibrillary acidic protein (GFAP) indicates, as can be seen in Figure 1A and Figure 1C, that the cells stained positively only for neuronal enolase (green) while GFAP staining (red) was not detectable. These data further indicate that almost all cells seen by phase-contrast microscopy (Figure 1B) were positive for neuronal enolase (Figure 1A). Thus, the cultures used in our studies are purely neuronal.
Effect of glutamate on cell viability
Cortical neuronal cells (4 × 104) cultured for 7 days in 96-well plates were treated with increasing concentrations (0.1 to1000 μM) of glutamate for 20 min and cell viability was measured after 24 hrs and compared with vehicle-treated samples. As depicted in Figure 2, 0.1 to 5 μM glutamate resulted in minimal cell death. As the amount of glutamate increased between 5 to 100 μM, the cell death increased gradually in a dose-dependent manner. Thus, treatment with 10 μM glutamate and 50 μM glutamate decreased cell viability by 35.6 ± 3.6% and 54 ± 2.5% respectively. Higher concentrations (>50 μM) of glutamate did not result in any further decrease in cell viability (Figure 2). Thus cell death induced by acute glutamate treatment reached a plateau in these rat primary cortical cells with concentrations of glutamate between 50 and 1000 μM. For all subsequent experiments 200 μM glutamate was used to induce cell death and this concentration resulted in a mean cell death of 48.7 ± 0.7%.
Kinetics of glutamate-induced cell death
In order to determine the kinetics of glutamate excitotoxicity, rat primary cortical neurons were cultured for 7 days and then treated with 200 μM glutamate for 20 min. Cell death was then assessed at 2, 4, 8, and 24 hrs as described under "Materials and Methods".
The results depicted in Figure 3 indicated that acute glutamate treatment resulted in a time-dependent increase in LDH release compared to vehicle treated cells. Measurement of LDH release at 2 hours resulted in no observable difference in LDH release compared to vehicle treated samples, however after 4 hours, an increase in LDH release of 13 ± 2.7% (p < 0.05) compared to control was observed. At 8 hours LDH release increased by 40 ± 6.8% (p < 0.05) over control. No significant increase in LDH release was observed thereafter (Figure 3). The change in LDH release ratio between glutamate and the control cells at 8 hours (40 ± 6.8%) compared to 24 hrs (42 ± 3.4%) was not statistically significant. Thus, it appears that the majority of cell death in rat primary cortical cells from acute glutamate exposure occurred between 2 to 8 hours post glutamate treatment.
Effect of estrogen on glutamate-induced cell death
To determine whether the neurotoxic effect of glutamate can be prevented by estrogen treatment, cells were pretreated for 24 hrs with various concentrations, 0.01 to 10 μM, of 17β-E2 or Δ8, 17β-E2, glutamate treatment was then initiated and cell viability was measured after 24 hrs. An acute (20 minute), 200 μM glutamate exposure resulted in a 50% decrease in cell viability. As depicted in Figure 4, pretreatment with 17β-E2 or Δ8, 17β-E2 for 24 hrs resulted in a dose-dependent increase in cell viability against glutamate-induced neurotoxicity. No significant protection was observed with 0.01 to 0.05 μM 17β-E2, however, 0.1 to 5 μM 17β-E2 resulted in a 8.16 ± 1.93% to 20.94 ± 2.13% increase in protection (p < 0.05) respectively. Similarly, pretreatment with 0.01 to 0.5 μM Δ8, 17β-E2 resulted in an increase in protection from 0.63 ± 1.74% to 6.85 ± 2.09% respectively, however, statistical significance at these concentrations was not reached. In contrast, pretreatment with 1 and 5 μM Δ8, 17β-E2 resulted in a significant (p < 0.05) increase in protection of 9.24 ± 2.23% and 12.12 ± 2.11% respectively (Figure 4). Higher concentration (10 μM) of both estrogens did not result in any further increase in protection. Although both 17β-E2 and Δ8, 17β-E2 were able to significantly reduce the glutamate-induced cell death complete protection was not achieved with the concentrations tested. However, the results do indicate that 17β-E2 was more effective in preventing glutamate-induced cell death than Δ8, 17β-E2 (Figure 4) and this was found to be significant with the 0.1 to 10 μM concentrations of estrogen tested (p < 0.05). Pretreatment of cells with estrogen for 15 minutes did not protect the cells against glutamate neurotoxicity, (data not shown).
Effect of NMDA and NOS inhibitors on glutamate-induced cell death
To determine whether the neurotoxic effect of acute glutamate treatment on rat primary cortical neurons is mediated through the NMDAR or NO, cell viability was determined following inhibition of NMDAR function or inhibition of NOS activity in cells treated acutely with 200 μM glutamate. As can be seen in Figure 5, a 20 minute exposure with 200 μM glutamate in rat primary cortical cells resulted in a 44.43 ± 1.34% decrease in cell viability compared to the vehicle treated control (p < 0.001). A 15 minute pretreatment with 0.1 μM MK-801, a non-specific antagonist of the NMDAR, completely prevented cell death (p < 0.001) compared to glutamate alone treated samples. Similarly, a 15 minute pretreatment with 10 mM NOS inhibitor L-NAME also completely prevented the cell death induced by glutamate (p < 0.001) compared to glutamate alone treated samples. These results clearly indicate that cell death induced by glutamate involves both NMDAR and NO.
Effect of glutamate on intracellular calcium levels
To evaluate the effect of glutamate on [Ca2+]i concentrations in cortical neurons various glutamate concentrations, (0.1 to 200 μM), were added to wells and changes in relative fluorescence units (RFU) were recorded. Baseline readings were taken every 15 seconds for a total of 60 seconds prior to an injection (10 μl) of glutamate into each well (indicated by glu Figure 6A). Readings were continued for another 2 minutes post-injection. The kinetics of glutamate induced Ca2+ influx are depicted in Figure 6A, and as can be seen, glutamate treatment (Figure 6A) results in a rapid increase in [Ca2+]i with all concentrations of glutamate (0.1 to 200 μM) tested. This initial rapid rise in [Ca2+]i occurs at 15 seconds after glutamate addition and is followed by a reduced rate of Ca2+ influx. The kinetics of glutamate-induced Ca2+ influx following the initial rise in [Ca2+]i remain relatively stable, but are elevated at glutamate concentration between 0.1 and 1 μM. However, at higher glutamate concentration (5 to 200 μM), [Ca2+]i continues to increase following the initial rise in [Ca2+]i. In addition, glutamate treatment results in a dose dependent increase in the initial rise in [Ca2+]i levels measured at 15 seconds, and reaches significance (p < 0.05) with 0.5 μM to 200 μM glutamate (Figure 6B). Maximal [Ca2+]i levels were observed with 10 μM glutamate and further increase in glutamate concentration did not result in any further increase in [Ca2+]i (Figure 6B).
Effect of NMDAR antagonist, MK-801, on glutamate-induced Ca2+ influx
In order to evaluate the contribution of glutamate-induced Ca2+ influx on cell viability the effect of NMDAR antagonist MK-801 on Ca2+ influx and cell viability was measured 24 hours later in the same cell population after glutamate treatment. The maximum initial Ca2+ influx was reached 15 seconds after glutamate (200 μM) injection and percent of maximum initial Ca2+ influx was calculated for each concentration of MK-801 (0.001 to 10 μM). As can be seen in Figure 7, a 15 minute pretreatment with 0.001 to 10 μM MK-801 resulted in a significant dose-dependent reduction in the initial (15 seconds) Ca2+ influx by 14.17 ± 2.56% to 40.30 ± 2.28% respectively. Thus, between 0 μM and 10 μM MK-801, the concentration of [Ca2+]i drops from an initial high level of 350 nM (glutamate alone) at 15 seconds to 200 nM (glutamate +10 μM MK-801). This drop in Ca2+ influx was associated with a dose-dependent increase in cell viability compared to glutamate alone (Figure 8). MK-801 mediated protection reached significance (p < 0.05) with 0.01 μM MK-801 and 0.1 μM - 10 μM MK-801, completely prevented glutamate-induced cell death (Figure 8).
Effect of estrogen on glutamate-induced calcium influx
The effect of estrogen on acute, glutamate (200 μM) induced Ca2+ influx was measured at various concentrations (0.01 to 10 μM) of 17β-E2 and Δ8, 17β-E2. Representative traces for the effect of 5 μM 17β-E2 and Δ8, 17β-E2 on the glutamate-induced Ca2+ influx are presented in Figure 9. Baseline [Ca2+]i measurements were made every 15 seconds for a total of 60 seconds and measurements were continued every 15 seconds following glutamate addition (indicated by glu in Figure 9). As shown in Figure 9, prior to glutamate addition [Ca2+]ilevels were below 20 nM and remained relatively constant. Addition of 200 μM glutamate resulted in an immediate increase in concentration of [Ca2+]i to levels above 300 nM (Figure 9A and 9B). Pretreatment with 5 μM 17β-E2 or Δ8, 17β-E2 was not associated with any significant decrease in the levels of [Ca2+]i induced by glutamate (Figures 9A and 9B). However, pretreatment with estrogens resulted in a significant (p < 0.05) but modest increase (up to 20%) in cell protection compared to glutamate alone. Similarly, no effect on the initial glutamate induced Ca2+ influx was observed at all other concentrations (0.01 to 1 μM) of estrogen. These results indicate that protection against glutamate-induced cell death in rat primary cortical neuronal cells by estrogen does not involve a reduction in the glutamate-induced, NMDAR associated, Ca2+ influx, and thus estrogen mediated protection appears to be mediated by a Ca2+ independent mechanism.
Discussion
In this study, we demonstrate that in primary cultures of pure rat embryonic cortical neurons, acute (20 min) treatment with glutamate results in a dose and time-dependent cell death. Over 50% of cells die at a glutamate concentration of 50 μM. Higher concentrations (> 50 μM) of glutamate did not result in any further increase in cell death. These findings are in keeping with previous observations with cultures of purified embryonic rat spinal cord motor neurons where NMDA and non-NMDA glutamate receptor agonists kill a maximum of 40% of the neurons [28]. In contrast to these findings some studies with immortalized hippocampal cells (HT22) cultured in media containing serum have reported glutamate-induced cell death to occur only at concentrations in the millimolar range [8,13,14]. Taken together, these findings suggests that the cell type and culture conditions may play an important role in the sensitivity of cells towards glutamate-induced cell death. In addition, a main difference between rat primary cortical neuronal cells and the HT22 cells used, is that the HT22 cells are devoid of functional NMDARs and thus the cell death induced in these cells at high concentrations of glutamate is thought to proceed mainly through the oxidative pathway rather than the excitotoxic pathway [7,8].
Although the primary cortical neurons used in our study are sensitive to small changes in glutamate concentrations (0.1 to 50 μM), the extent of neuronal death under our experimental conditions (50%) is lower than that reported in a number of studies where mixed cell cultures containing neuronal, glial and other cell types were used. In these mixed cultures over 85% of motor neurons died following treatment with glutamate [29,30]. It has further been demonstrated that in such mixed cell cultures, glutamate receptor agonists may induce cell death by activating receptors on the susceptible neurons or by activating receptors in non-neuronal cells. This results in the release of substances such as NO that are toxic to neurons [31,32]. Thus, the lower extent (50%) of cell death observed in cultures maintained in neurobasal medium containing B27 supplements is most likely due to the high purity (absence of glia) of the neuronal cultures used in our study. The absence of glia in our neuronal cultures (Figure 1) maintained in serum free neurobasal medium and B27 supplements has also been reported in a number of earlier studies [16,23,33,34].
Alternatively, it is also possible that the lower extent of cell death after a short (20 min) exposure to excitotoxic levels of glutamate (200 μM) may be due to the presence of glutamate sensitive and insensitive neurons in our cultures and this has also been reported by others [23,28,35]. It has further been demonstrated in pure motor neuronal cultures (7 day old) that survival (60%) of a subset of neurons treated with glutamate was unlikely to be due to the lack of differential glutamate receptor expression [28]. These authors further suggest that selective vulnerability may be down stream of the rise in intracellular Ca2+ that occurs after treatment with glutamate or agonists [28].
Previous studies, in particular with mixed cortical cultures or spinal cord neurons, have indicated that the age or maturation of neurons is important in the extent of neuronal death following treatment with glutamate. Thus, in mixed 18 day old cultures of neurons, glutamate treatment resulted in over 80% neuronal cell death compared to only 25% cell death in 4 day old cultures [36]. Since we have used cortical neurons that have been maintained in culture for only 7 days, it is possible that the extent of neuronal cell death induced by glutamate may further increase in longer-term cultures. Although we have been able to maintain some neurons for longer periods than one week, these neurons, even in the absence of glutamate, continue to die at variable rates. In keeping with our observations, Fryer et al [28] reported that 95% of pure motor neurons in culture die by 7 days. It appears that long-term cultures of pure neurons maintained in absence of serum cannot at present be used to address the question of the effect of glutamate on neurons maintained in culture for extended periods of time.
The kinetics of glutamate-induced cell death indicated that while an acute (20 minute) exposure to glutamate was sufficient to cause cell death, cell death did not occur until 4 to 8 hours following glutamate treatment. This delay in cell death is suggestive of apoptosis. Indeed, recent evidence indicates that apoptotic mechanisms can play an important role in glutamate-induced neuronal cell death [13,14,37-40]. Thus, we have demonstrated in HT22 mouse hippocampal cells that glutamate induced apoptosis was associated with DNA fragmentation, morphological changes and up-regulation of the pro-apoptotic protein Bax and down-regulation of the anti-apoptotic protein Bcl-2. Moreover, this process was reversed by estrogens [14]. We have further shown that in primary cortical neurons glutamate-induced apoptosis by caspase-3 dependent and independent mechanisms involves mitochondrial release of cytochrome C and apoptosis inducing factor (AIF) and these apoptotic changes were reversed by equine estrogens [37].
While the mechanism of glutamate-induced cell death has not been fully elucidated, results from the present study clearly indicate that glutamate-induced cell death in primary cortical cells is NMDAR and NOS linked. The relatively short time required for glutamate to initiate the cell death process and the requirement for NMDAR activation and NO production, suggest that this cell death is primarily excitotoxic. These findings are in keeping with previous observations that suggest excitotoxic cell death in cortical cultures is dependent on NMDAR and NOS activity [11,12]. Furthermore, it has been suggested that it is the production of ONOO- from increased NO formation that may be responsible for cell death and that this increase in NO is dependent on NMDA associated Ca2+ influx [9]. Therefore, reduction in Ca2+ influx or NO production are two potential mechanisms by which neuroprotective agents may act.
The data from the present study clearly indicates that estrogen can protect against glutamate-induced excitotoxicity in primary cortical cells and a 24 hour pretreatment prior to glutamate insult is sufficient to provide this protection. In a previous study [23] protective effects of 17β-E2 were observed at physiological concentrations, while the concentrations we used were pharmacological. The differences are most likely due to culture conditions used: their media [23] contained antioxidants, higher cell density and lower glutamate concentrations. Since estrogens are potent antioxidants [21,22], the presence of other antioxidants in the media may explain the differences between our findings and those reported previously [23]. Although the concentrations of estrogen used in our study are high, significantly high levels [70 ng/hr/mL] were found in postmenopausal women taking a single therapeutic dose of conjugated equine estrogen [41]. Whether the levels used in the present study are attained in women taking these estrogens daily for extended periods of time remains to be determined.
It has been previously suggested that the antioxidant property of estrogen may be one possible mechanism by which estrogens exert their neuroprotective activity. Therefore, if estrogen mediated neuroprotection occurs solely through an antioxidant mechanism, then Δ8, 17β-E2 (a more potent antioxidant estrogen) [21,22,42] would be expected to be much more effective in preventing glutamate-induced cell death than 17β-E2. However, in the present study 17β-E2 was significantly more potent than Δ8, 17β-E2. The difference in neuroprotective potency suggests that the protective effect under the experimental conditions used in the present study are not solely dependent on the estrogen's antioxidant potency. The difference may also be due to the experimental design used in the present study (ex in vivo versus whole cell). Although Δ8, 17β-E2 is less potent than 17β-E2, its use as a potential neuroprotective agent warrants further investigation. This estrogen is a weaker uterotropic estrogen than 17β-E2 and its effects appear to be mediated mainly via the estrogen receptor β (ERβ) [26,27,43]. Thus, Δ8, 17β-E2 may have therapeutic applications in both women and men for prevention of neurodegeneration. Whether this neuroprotection also occurs through the well characterized genomic mechanism involving ERs (ERα and ERβ) or through the putative membrane receptor pathways [44] remains to be investigated.
As we have shown, glutamate-induced excitotoxicity is dependent on NMDAR activity and therefore, changes in Ca2+ influx across the cell membrane may trigger mechanisms involved in cell death [6]. Indeed, it was observed in the present study that following glutamate stimulation, a rapid dose-dependent increase in [Ca2+]i occurs that is associated with a dose-dependent decrease in cell viability. However, stimulation with glutamate above 10 μM did not further increase [Ca2+]i again suggesting that glutamate insensitive and sensitive cells may have been present in these primary cortical cultures.
The observation that glutamate-induced Ca2+ influx decreased significantly by MK-801 suggests that this to some extent occurs via NMDAR. Increasing concentrations of MK-801 resulted in a gradual lowering of the [Ca2+]i concentration within 15 seconds after glutamate treatment. Thus MK-801 (10 μM) reduced the [Ca2+]i from 350 nM to 200 nM in 15 seconds (Figure 7). Although the fall in [Ca2+]i levels between 0.01 to 0.1 μM MK-801 was less than 5% (270 nM to 260 nM respectively; Figure 7), the cell viability increased from 58% to 100%, (Figure 8). These types of glutamate-induced small changes in [Ca2+]i that result in large (total) and significant cell death of cortical/glial cell cultures have been previously observed [45]. These investigators explained this by suggesting the presence of a critical Ca2+ "threshold". They further suggested that even a small reduction in NMDA receptor activation may be sufficient to reduce Ca2+ influx below this critical threshold such that neural cell death is significantly reduced. Our results (Figures 6,7,8) indicate that perhaps such a threshold may also be present in pure neuronal cortical cells.
In the present study the effect of estrogen on excitotoxic Ca2+ influx was investigated and although 17β-E2 and Δ8, 17β-E2 both provided significant protection (Figure 9), neither had any effect on the glutamate-induced Ca2+influx contrary to our working hypothesis. Thus it appears that the estrogen mediated protection of primary cortical cells from acute glutamate exposure does not involve a decrease in the glutamate-induced Ca2+ influx, and to our knowledge this is a first such observation. In contrast to these findings a previous study in hippocampal cells cultured in media containing antioxidants, has shown that estrogen can reduce the glutamate-induced Ca2+ influx by increasing mitochondrial sequestration of Ca2+ [25]. This increased sequestration was attributed to the ability of estrogen to increase Bcl-2, an antiapoptotic protein that is important in mitochondrial Ca2+ sequestration [46,47]. Whether estrogen modulates Bcl-2 expression in primary cortical cells in a manner similar to our previous study with HT22 cells [14] or whether mitochondrial sequestration of Ca2+ occurs remains to be established. Furthermore, these investigators [25] cultured the hippocampal cells in medium that contained antioxidants and this may also explain some of the differences observed. Our results do however, indicate that the mechanism by means of which MK-801 and estrogens protect against glutamate-induced cell death, appears to be different. Further studies are needed to identify these differences.
There are many other effects of estrogen on the central nervous system that have been described, for example, estrogen can alter NOS expression throughout the brain and potentially upregulate antioxidant proteins through cyclic guanosine monophosphate linked mechanisms [48,49]. Such alterations may play a role in glutamate excitotoxicity by affecting subsequent events in the cell death pathway, downstream of the initial Ca2+ influx. In addition, the rapid activation of mitogen-activated protein kinase (MAPK) by estrogen has been shown to be necessary for estrogenic protection although how and what the function of MAPK is in preventing cell death remains to be established [50]. Similarly, the phosphorylation of CREB (cAMP [cyclic adenosine monophospate] response element-binding protein) and upregulation of neurotrophic factors such as brain-derived neurotrophic factor [51], may play a role in neuronal protection [52]. Whether these factors are involved in prevention of glutamate-induced excitotoxicity by equine estrogens remains to be determined.
Conclusion
Glutamate-induced cell death in cortical cultures occurs through NMDAR and NOS-linked mechanisms, presumably through elevated [Ca2+]i levels that activate NOS and increase the levels of NO and subsequently ONOO-free radicals. Equine estrogens, 17β-E2 and Δ8, 17β-E2, significantly protect against glutamate-induced excitotoxicity through a mechanism that is independent of Ca2+. This is to our knowledge a first such observation. Whether a decrease in NOS related products, such as ONOO-, may be one mechanism by which estrogen provides neuroprotection against glutamate toxicity, remain to be investigated.
Methods
Isolation and culture of cortical cells
Day 17–19 pregnant Sprague Dawley rats were anesthetized and sacrificed by cervical dislocation. The fetuses were delivered, decapitated and the brains were dissected out. Each cortex was isolated and individual cells were cultured essentially according to Brewer et al [33,34] with the following modifications. Briefly, the tissue was maintained in Hank's balanced salt solution (HBSS) (Invitrogen, Life Technologies, Burlington, ON, Canada) until enough tissue to give sufficient cell numbers for plating was collected. The tissue was then dispersed by mechanical dissociation through a series of constricted fire polished, siliconized Pasteur pipettes to obtain a suspension of single cells. The suspension was then centrifuged and the cells were washed to remove red blood cells and other impurities. An aliquot was stained with 4% trypan blue, live cells were counted using a hemocytometer and the suspension was then diluted accordingly with neurobasal defined medium (Invitrogen, Life Technologies, Burlington, ON, Canada) supplemented with 2% of B27 containing antioxidants (Invitrogen, Life Technologies, Burlington, ON, Canada), 0.5 mM glutamine, penicillin G (10 U/mL) and streptomycin (13 μg/mL) purchased all from Sigma, Canada. This medium was labeled as neurobasal complete (NBC).
Cells, 40,000 per well, were seeded onto poly-D-lysine coated 96 well plates (BD Biosciences, Discovery Labware, Bedford, MA, USA) and were maintained in NBC. Under these culture conditions, it has been previously demonstrated (16,23,33,34) that only neurons survive while glia are essentially absent. The purity of our cortical neuronal cultures is described separately. After day 1 of plating the media was completely replaced and on subsequent days 50% of the media was replaced. Cell cultures were maintained for 7–12 days at which time glutamate experiments were initiated. Preliminary experiments carried out during the standardization of the methods indicated that the cell viability decreased and was variable after day 7 in culture. Although some neurons were viable up to day 12, all experiments described in the manuscript were carried out in cultures maintained for only 7 days.
Detection of neuronal specific enolase (NSE) and glia-specific glial fibrillary acid protein (GFAP) by immunocytochemistry
Cortical cells on poly-D-lysine coated plates were cultured in NBC medium as described above. After seven days, cells were washed twice with PBS and then fixed with 4% formaldehyde in PBS for 15 minutes. Cells were then washed three times with PBS and non-specific binding was blocked for 5 mins with PBS containing 1% BSA (Sigma, Canada), 5% normal goat serum (Sigma, Canada), 0.05% Triton-X100 (BDH, Canada). Cells were incubated for 2 hrs at room temperature with rabbit anti-rat neuron specific enolase (anti NSE; Polyscience, PA, USA) 1:1000 dilution in block solution. After rinsing 4 times with PBS, the cells were incubated at room temperature with Alexa Fluor 488 F(ab')2 fragment of goat anti-rabbit IgG (H+L) secondary antibody (Molecular Probes, Invitrogen, Canada) 1:400 dilution in block solution. Cells were then rinsed 4 times with PBS and the fluorescence labeling was visualized and photographed using Olympus IMT-2 microscope, equipped with a mercury light source and spot slider digital camera (Diagnosis Instruments, MI, USA). The neuronal cells displayed characteristic green fluorescence (Figure 10A).
The neuronal cells maintained in NBC are devoid of glial cells and to obtain cultures containing glia, the cortical cells were prepared as described above but were maintained in Dubelcco's Modified Eagle Medium (DMEM, Sigma Canada) containing 10% Fetal Bovine Serum (FBS; Invitrogen Canada Inc; On, Canada). After 7 days in culture, the mixture of cells were processed for immunocytology by incubating with the primary antibody mouse anti-GFAP (1:4000 dilution; Sigma Canada) and then with Texas Red dye conjugated AffiniPure F (ab')2 fragment goat anti-mouse IgG (H+L) (1:100 dilution; Jackson Immunoresearch Laboratories, PA, USA) as described above for NSE, (data not shown). The GFAP positive glial cells stain red as seen in Figure 10B. For double staining, cells maintained in NBC medium were first incubated with primary and secondary antibodies for GFAP and after rinsing four times with PBS, the cells were incubated with antibodies for NSE. Cells were then visualized and photographed. The dual staining indicated that the cells were only positive for neuronal NSE and negative for glial GFAP (Figure 1A and 1C).
Treatment of cells
For determination of the effect of glutamate on the viability of rat primary cortical cells, the cells were washed twice with a modified Locke's buffer (154 mM NaCl, 5.6 mM KCl, 3.6 mM NaHCO3, 2.3 mM CaCl2, 5.6 mM Glucose, 5 mM HEPES, 10 μM Glycine, pH 7.4) and incubated for 24 hours at 37°C and 5% CO2 in NBC medium, plus 0.5 mM glutamine and 2% B27 minus antioxidants (NBC AO-). Following this, the cells were washed 3 times with modified Locke's buffer and glutamate exposure was initiated. All glutamate treatments were conducted in modified Locke's buffer. Various concentrations of glutamate were added to a final concentration of 0.1, 0.5, 1, 5, 10, 50, 100, 200, 500 and 1000 μM for a total of 20 minutes and changes in [Ca2+]i were recorded. The cells were then washed with modified Locke's buffer and NBC AO- was added and cells were incubated at 37°C, 5% CO2 for 24 hours, following which cell death/viability was measured. Measurement for Ca2+ and cell death/viability were performed on the same culture of cells.
The kinetics of glutamate-induced cell death following a 20 minute exposure to glutamate was also measured using a concentration of glutamate found to be toxic to 50% of the cells. Cell death measurements were made at 2, 4, 8 and 24 hours post glutamate treatment and comparisons between glutamate treated samples and vehicle (modified Locke's buffer) treated samples were made.
For the determination of the effects of estrogen on glutamate induced changes, cells were washed twice with modified Locke's buffer and then fresh NBC AO- and estrogens were added to final concentrations of 0.01, 0.05, 0.1, 0.5, 1 and 10 μM and the cells were incubated for 24 hours at 37°C and 5% CO2. Estrogens were dissolved in ethanol and diluted to working concentrations with NBC AO- medium. The final concentration of ethanol was 0.2% for control and estrogen treated samples. Following incubation with estrogen and prior to glutamate exposure cells were prepared for [Ca2+]i measurements as outlined below. The cells were washed and incubated in Locke's buffer containing sufficient (200 μM) glutamate (pre-determined) to give 50% cell death. Incubations were carried out at 37°C for 20 minutes and changes in [Ca2+]i were recorded. Following glutamate treatment cells were washed with fresh Locke's buffer and NBC AO- was added. The cells were again incubated for 24 hours at 37°C and 5% CO2, following which cell viability was determined.
In some experiments various concentrations of MK-801, an NMDAR antagonist, and 10 mM of L-NAME [53], a NOS inhibitor, were added 15 minutes prior to glutamate exposure following which changes in [Ca2+]i and cell viability were measured.
Evaluation of cell death and viability
The Lactate dehydrogenase (LDH) cytotoxicity assay (Promega G1780, VWR, Toronto, Ontario, Canada) and the CellTiter 96® AQueous Non-Radioactive cell proliferation assay (MTS method) (Promega G5430) were used to assess cell death and viability respectively. The LDH assay measures the amount of lactate dehydrogenase that is released from dead cells while the MTS method measures the formation of a formazan product in live cells. Changes in optical density (O.D.) at 490 nm were measured using the SpectraMax 340 microplate reader and SOFTmax® PRO software (Molecular Devices, California, U.S.A.) and comparisons of treated to control values were made. Both of these methods provide a reliable measure of cell death/viability.
Measurement of intracellular calcium levels
Intracellular Calcium levels were assessed using the fluorescent probe Fluo-3 AM (Molecular Probes, Eugene, OR, USA). Changes in relative fluorescence units (RFU) were monitored with the Fluoroskan Ascent FL fluorescent plate reader equipped with a micro-injection syringe pump (Labsystems, Helsinki, Finland). The protocol for intracellular dye loading was based on the method provided by the manufacturer (Molecular Probes and Labsystems Fluo-3 AM Application Note) with minor modifications. Briefly, cells were loaded with 5 μM Fluo-3 AM for 2 hours and then washed three times with Locke's buffer to remove excess dye. Cell plates were then placed in the Fluoroskan and incubated for an additional 15 minutes at 37°C to allow for complete de-esterification of dye. Measurements of relative fluorescence units (RFU) were made using an excitation and emission wavelength of 485 and 527 nm respectively. Change in [Ca2+]i (nM) was calculated by multiplying the ratio of ΔF-Fmin and Fmax-ΔF by Kd. Where ΔF is the change in RFU, Fmax is the maximum observed RFU following NP40 treatment, Fmin is the minimum observed RFU following EGTA Ca2+ chelation and Kd is the dissociation constant for Fluo-3 AM, 390 nM, provided by supplier (Molecular Probes).
Data analysis
The dose-response curves for effects of glutamate and estrogen on cell viability and [Ca2+]i were developed using Slide Write Plus V6 Software (Advanced Graphics Software, Inc., Encinitas, CA) using the non-linear curve fitting functions. Each estrogen experiment was carried out a minimum of six times with eight replicates per treatment condition and all values shown in figures are the mean ± standard error of the mean (SEM). All treatment conditions were compared to the respective controls and an effect was determined significant if p < 0.05. All statistical analyses were performed using Instat® Software (Graph Pad, San Diego, CA). To establish the effect of estrogen against glutamate induced cell death and [Ca2+]i, nonparametric statistical analyses were performed using the Kruskal-Wallis nonparametric analysis of variance test with the Dunn post hoc test to compare each estrogen concentration against the glutamate alone control.
Authors' contributions
JP is a graduate student who participated in the development of the hypothesis, study design and carried out most of the experimental work and preparation of the manuscript.
BB conceived the study and participated in the development of the hypothesis, the study design, and overall direction of the study and preparation of the manuscript.
Both authors have read and approved the final preparation of the manuscript and its submission to BMC Neuroscience.
Acknowledgements
This work was supported by the Medical Research Council of Canada Grant MT-11929 and a basic research grant from Women's Health Care Research, Wyeth Pharmaceuticals, Philadelphia, PA, USA
Figures and Tables
Figure 1 Immunocytochemistry for neuron specific enolase (NSE) (green) and glial fibrillary acidic protein (GFAP) (red). Cortical cells were grown in serum-free Neurobasal medium supplemented with 2% B27 for 7 days prior to dual staining and visualization with fluorescence and phase contrast microscopy. Images shown were taken at 300× magnification. All images are taken from the same field of view. A. Fluorescence image of cortical cells stained for NSE. Arrowheads represent cells stained positive for NSE. B. Phase contrast image of cortical cells. C. Fluorescence image of cells stained for GFAP. Note almost all cells in (B) stain positive for NSE (A) and no cells stain positive for GFAP (C) indicating absence of glia.
Figure 2 Effect of glutamate on cell viability. Neuronal cells were maintained for 7 days in conditions similar to those described under Figure 1. Cells were treated with various concentrations of glutamate for 20 mins. Cell viability was determined 24 hrs later. Untreated control consisted of cells exposed to Lockes buffer only. Each treatment was repeated 8 times. Error bars represent SEM.
Figure 3 Kinetics of glutamate-induced cell death. Neuronal cells were maintained for 7 days in conditions similar to those described under Figure 1. Cells were treated with 200 μM glutamate for 20 mins. LDH release was measured at 2, 4, 8 and 24 hrs later. Results are expressed as percent LDH increase from control values. Each point is an average of 8 replicates. Error bars represent SEM.
Figure 4 Effect of estrogen on glutamate-induced cell death. Neuronal cells were maintained for 6 days under conditions similar to those described under Figure 1. Cells were treated with various concentrations of 17β-E2 and Δ8, 17β-E2. After 24 hrs cells were treated with 200 μM glutamate for 20 mins. Cell viability was measured 24 hrs later. Results are expressed as increase in percent protection compared to glutamate alone. Solid bars represent 17β-E2. Hatched bars represent Δ8, 17β-E2. a. p < 0.05, between glutamate alone and estrogen treatment. b. p < 0.05, between 17β-E2 and Δ8, 17β-E2. Each experiment was repeated a minimum of 6 times with 8 replicates per experiment. Error bars represent SEM.
Figure 5 Effect of MK-801 and L-NAME on glutamate-induced cell death. Neuronal cells were maintained for 7 days in conditions similar to those described under Figure 1. Cells were treated with either 0.1 μM MK-801 or 10 mM L-NAME. After 15 mins cells were then treated with 200 μM glutamate. Results are expressed as percent of control. a. p < 0.05 compared to control. b. p < 0.05 compared to glutamate alone treatment. Each bar represents a mean of 8 replicates. Error bars represent SEM.
Figure 6 Effect of time and concentration of glutamate on [Ca2+]i. Neuronal cells were maintained for 7 days in conditions similar to those described under Figure 1. Cells were loaded with 5 μM Fluo-3 AM and treated with various concentrations of glutamate. Changes in fluorescence were monitored using the Fluorskan Ascent FL and results were converted to changes in [Ca2+]i (nM) as described in the materials and methods. A. [Ca2+]i measurements were made every 15 secs. Baseline levels of [Ca2+]i were monitored for 60 secs prior to injection with various concentrations of glutamate (indicated by glu). Following injection changes in [Ca2+]i were monitored for an additional 120 secs. B. Initial peak [Ca2+]i change measured at 15 secs post glutamate injection. Each point/bar represents a mean of 8 replicates. Error bars represent SEM.
Figure 7 Effect of various concentrations of MK-801 on glutamate-induced [Ca2+]i influx. Neuronal cells were maintained for 7 days in conditions similar to those described under Figure 1, prior to loading with 5 μM Fluo-3 AM. After a 15 min treatment with various concentrations of MK-801 cells were treated with 200 μM glutamate. Changes in fluorescence were monitored using the Fluorskan Ascent FL and results were converted to changes in [Ca2+]i (nM) as described in the materials and methods. [Ca2+]i measurements were recorded every 15 secs. Baseline levels of [Ca2+]i were monitored for 60 secs prior to injection with 200 μM glutamate (indicated by glu). Following injection changes in [Ca2+]i were monitored for an additional 120 secs. MK-801 (0 μM) represents 200 μM glutamate alone treatment. Each point represents a mean of 8 replicates. Error bars represent SEM.
Figure 8 Effect of MK-801 on glutamate-induced cell death. Cell viability was determined in the same cells used for experiments described in Figure 7 and is expressed as percent of control. Cells were pretreated for 15 minutes with various concentrations of MK-801 (0.001 to 10 μM) and then treated for 20 minutes with 200 μM glutamate. a. p < 0.05 compared to glutamate alone. Each bar represents a mean of 8 replicates. Error bars represent SEM.
Figure 9 Effect of estrogen treatment on glutamate-induced [Ca2+]i and cell death. Neuronal cells were maintained for 6 days in conditions similar to those described under Figure 1. Cells were treated with various concentrations of 17β-E2 and Δ8, 17β-E2. After 24 hrs cells were loaded with 5 μM Fluo-3 AM and then treated with 200 μM glutamate. [Ca2+]i measurements were made every 15 secs. Baseline levels of [Ca2+]i were monitored for 60 secs prior to injection with 200 μM glutamate (indicated by glu). Following injection changes in [Ca2+]i were monitored for an additional 120 secs. A. Representative trace of the effect of 5 μM 17β-E2 on glutamate-induced [Ca2+]i influx. Inset. Increase in percent protection with 5μM 17β-E2 against 200 μM glutamate-induced cell death (same cells as those in which [Ca2+]i measurements were made). B. Representative trace of the effect of Δ8, 17β-E2 on glutamate-induced [Ca2+]i influx. Inset. Increase in percent protection with 5 μM Δ8, 17β-E2 against 200 μM glutamate-induced cell death (same cells as those in which [Ca2+]i measurements were made). Each experiment was repeated a minimum of 6 times with 8 replicates of each treatment condition per experiment. Error bars represent SEM.
Figure 10 Detection of NSE (green) and GFAP (red) by immunocytochemistry in neurons and glia. Cells were labelled with either anti-NSE or anti-GFAP and visualized by fluorescence microscopy. Neuronal cells were grown in Neurobasal medium supplemented with 2% B27 and maintained for 7 days in culture. Glial cells were grown in DMEM supplemented with 10% Fetal Bovine Serum and maintained for 7 days in culture. (A) Neurons labelled with anti-NSE. (B) Glial cells labelled with anti-GFAP. Images were taken at 300× magnification.
==== Refs
Choi DW Glutamate neurotoxicity and diseases of the nervous system Neuron 1988 1 623 34 2908446 10.1016/0896-6273(88)90162-6
Lipton SA Rosenberg PA Excitatory amino acids as a final common pathway for neurologic disorders N Engl J Med 1994 330 613 22 7905600 10.1056/NEJM199403033300907
Coyle JT Puttfarcken P Oxidative stress, glutamate, and neurodegenerative disorders Science 1993 262 689 95 7901908
Olney JW Wozniak DF Farber NB Glumate receptor dysfunction and Alzheimer's disease Restor Neurol Neurosci 1998 13 75 83 12671289
Ankarcrona M Dypbukt JM Bonfoco E Zhivotovsky B Orrenius S Lipton SA Nicotera P Glutamate-induced neuronal death: a succession of necrosis or apoptosis depending on mitochondrial function Neuron 1995 15 961 73 7576644 10.1016/0896-6273(95)90186-8
Choi DW Excitotoxic cell death J Neurobiol 1992 23 1261 76 1361523
Murphy TH Miyamoto M Sastre A Schnaar RL Coyle JT Glutamate toxicity in a neuronal cell line involves inhibition of cystine transport leading to oxidative stress Neuron 1989 2 1547 58 2576375 10.1016/0896-6273(89)90043-3
Li Y Maher P Schubert D Phosphatidylcholine-specific phospholipase C regulates glutamate-induced nerve cell death Proc Natl Acad Sci U S A 1998 95 7748 53 9636222 10.1073/pnas.95.13.7748
Sattler R Tymianski M Molecular mechanisms of calcium-dependent excitotoxicity J Mol Med 2000 78 3 13 10759025 10.1007/s001090000077
Siegel GJ Agranoff BW Albers RW Fisher SK Uhler MD Basic Neurochemistry: molecular, cellular and medical aspects 1999 6 Lippincott Williams and Wilkins, American Society for Neurochemistry
Yamauchi M Omote K Ninomiya T Direct evidence for the role of nitric oxide on the glutamate-induced neuronal death in cultured cortical neurons Brain Res 1998 780 253 9 9507156 10.1016/S0006-8993(97)01201-8
Dawson VL Dawson TM London ED Bredt DS Snyder SH Nitric oxide mediates glutamate neurotoxicity in primary cortical cultures Proc Natl Acad Sci U S A 1991 88 6368 71 1648740
Bhavnani BR Berco M Binkley J Equine estrogens differentially prevent neuronal cell death induced by glutamate J Soc Gynecol Investig 2003 10 302 8 12853093 10.1016/S1071-5576(03)00087-X
Zhang YM Lu XF Bhavnani BR Equine estrogens differentially inhibit DNA fragmentation induced by glutamate in neuronal cells by modulation of regulatory proteins involved in programmed cell death BMC Neurosci 2003 4 32 14693041 10.1186/1471-2202-4-32
Brinton RD Chen S Montoya M Hsieh D Minaya J The estrogen replacement therapy of the Women's Health Initiative promotes the cellular mechanisms of memory and neuronal survival in neurons vulnerable to Alzheimer's disease Maturitas 2000 34 S35 52 10915920 10.1016/S0378-5122(00)00107-9
Brinton RD Proffitt P Tran J Luu R Equilin, a principal component of the estrogen replacement therapy premarin, increases the growth of cortical neurons via an NMDA receptor-dependent mechanism Exp Neurol 1997 147 211 20 9344546 10.1006/exnr.1997.6619
Paganini-Hill A Henderson VW Estrogen deficiency and risk of Alzheimer's disease in women Am J Epidemiol 1994 140 256 61 8030628
Henderson VW The epidemiology of estrogen replacement therapy and Alzheimer's disease Neurology 1997 48 S27 35 9153164
Tang MX Jacobs D Stern Y Marder K Schofield P Gurland B Andrews H Mayeux R Effect of oestrogen during menopause on risk and age at onset of Alzheimer's disease Lancet 1996 348 429 32 8709781 10.1016/S0140-6736(96)03356-9
Kawas C Resnick S Morrison A Brookmeyer R Corrada M Zonderman A Bacal C Lingle DD Metter E A prospective study of estrogen replacement therapy and the risk of developing Alzheimer's disease: the Baltimore Longitudinal Study of Aging Neurology 1997 48 1517 21 9191758
Perrella J Berco M Cecutti A Gerulath A Bhavnani BR Potential role of the interaction between equine estrogens, low-density lipoprotein (LDL) and high-density lipoprotein (HDL) in the prevention of coronary heart and neurodegenerative diseases in postmenopausal women Lipids Health Dis 2003 2 4 12848896 10.1186/1476-511X-2-4
Bhavnani BR Cecutti A Gerulath A Woolever AC Berco M Comparison of the antioxidant effects of equine estrogens, red wine components, vitamin E, and probucol on low-density lipoprotein oxidation in postmenopausal women Menopause 2001 8 408 19 11723413 10.1097/00042192-200111000-00005
Singer CA Rogers KL Strickland TM Dorsa DM Estrogen protects primary cortical neurons from glutamate toxicity Neurosci Lett 1996 212 13 6 8823751 10.1016/0304-3940(96)12760-9
Nilsen J Chen S Brinton RD Dual action of estrogen on glutamate-induced calcium signaling: mechanisms requiring interaction between estrogen receptors and src/mitogen activated protein kinase pathway Brain Res 2002 930 216 34 11879813 10.1016/S0006-8993(02)02254-0
Nilsen J Diaz Brinton R Mechanism of estrogen-mediated neuroprotection: regulation of mitochondrial calcium and Bcl-2 expression Proc Natl Acad Sci U S A 2003 100 2842 7 12604781 10.1073/pnas.0438041100
Bhavnani BR Cecutti A Gerulath A Pharmacokinetics and pharmacodynamics of a novel estrogen Δ8-estrone in postmenopausal women and men J Steroid Biochem Mol Biol 1998 67 119 31 9877212 10.1016/S0960-0760(98)00082-X
Bhavnani BR Lu XF Differential interaction of equine estrogens with recombinant human estrogen receptors alpha (ERα) and beta (ERβ) [abstract] J Soc Gynecol Investig 2003 10 298A 12853092
Fryer HJL Knox RJ Strittmatter SM Kalb RG Excitotoxic death of a subset of embryonic rat motor neurons in vitro J Neurochem 1999 72 500 13 9930721 10.1046/j.1471-4159.1999.0720500.x
Carriedo SG Yin HZ Lamberta R Weiss JH In vitro kainate injury to large, SMI-32(+) spinal neurons is Ca2+ dependent Neuroreport 1995 6 945 8 7612889
Carriedo SG Yin HZ Weiss JH Motor neurons are selectively vulnerable to AMPA/Kainate receptor-mediated injury in vitro J Neurosci 1996 16 4069 79 8753869
Bonfoco E Krainc D Ankarcrona M Nicotera P Lipton SA Apoptosis and necrosis: two distinct events induced, respectively, by mild and intense insults with N-methyl-D-aspartate or nitric oxide/superoxide in cortical cell cultures Proc Natl Acad Sci USA 1995 92 7162 6 7638161
Samdani AF Newcamp C Resink A Facchinetti F Hoffman BE Dawson VL Dawson TM Differential susceptibility to neurotoxicity mediated by neurotrophins and neuronal nitric oxide synthase J Neurosci 1997 17 4633 41 9169524
Brewer GJ Torricelli JR Evege EK Price PJ Optimized survival of hippocampal neurons in B27-supplemented Neurobasal, a new serum-free medium combination J Neurosci Res 1993 35 567 76 8377226 10.1002/jnr.490350513
Brewer GJ Serum-free B27/neurobasal medium supports differentiated growth of neurons from the striatum, substantia nigra, septum, cerebral cortex, cerebellum, and dentate gyrus J Neurosci Res 1995 42 674 83 8600300 10.1002/jnr.490420510
Schubert D Piasecki D Oxidative glutamate toxicity can be a component of the excitotoxicity cascade J Neurosci 2001 21 7455 62 11567035
Regan RF Choi DW Glutamate neurotoxicity in spinal cord cell culture Neuroscience 1991 43 585 91 1681469 10.1016/0306-4522(91)90317-H
Zhang Y Bhavnani BR Equine estrogens inhibit glutamate-induced neuronal apoptosis by caspase-3 dependent and independent mechanisms involving mitochondrial release of cytochrome c (Cyto C) and apoptosis inducing factor (AIF) [abstract] J Soc Gynecol Investig 2004 11 240
Mattson MP Excitotoxic and excitoprotective mechanisms: abundant targets for the prevention and treatment of neurodegenerative disorders Neuromolecular Med 2003 3 65 94 12728191 10.1385/NMM:3:2:65
Lok J Martin LJ Rapid subcellular redistribution of Bax precedes caspase-3 and endonuclease activation during excitotoxic neuronal apoptosis in rat brain J Neurotrauma 2002 19 815 28 12184852 10.1089/08977150260190410
Brecht S Gelderblom M Srinivasan A Mielke K Dityateva G Herdegen T Caspase-3 activation and DNA fragmentation in primary hippocampal neurons following glutamate excitotoxicity Brain Res Mol Brain Res 2001 94 25 34 11597762 10.1016/S0006-8993(01)02767-6
Bhavnani BR Nisker JA Martin J Aletebi F Watson L Milne JK Comparison of pharmacokinetics of a conjugated equine estrogen preparation (premarin) and a synthetic mixture of estrogens (C.E.S.) in postmenopausal women J Soc Gynecol Investig 2000 7 175 83 10865186 10.1016/S1071-5576(00)00049-6
Berco M Bhavnani BR Differential neuroprotective effects of equine estrogens against oxidized low density lipoprotein-induced neuronal cell death J Soc Gynecol Investig 2001 8 245 54 11525901 10.1016/S1071-5576(01)00111-3
Shui-Pang T Bhavnani BR Differential effects of various equine estrogens mediated via estrogen receptor α and β in HepG2 cells [abstract] The Endocrine Society, 85th Annual Meeting 2003 P1-104 12
Levin ER Cellular functions of plasma membrane estrogen receptors Steroids 2002 67 471 5 11960623 10.1016/S0039-128X(01)00179-9
Hartley DM Kurth MC Bjerkness L Weiss JH Choi DW Glutamate receptor-induced 45Ca2+ accumulation in cortical cell culture correlates with subsequent neural degeneration J Neurosci 1993 13 1993 2000 7683048
Murphy AN Bredesen DE Cortopassi G Wang E Fiskum G Bcl-2 potentiates the maximal calcium uptake capacity of neural cell mitochondria Proc Natl Acad Sci U S A 1996 93 9893 8 8790427 10.1073/pnas.93.18.9893
Ichimiya M Chang SH Liu H Berezesky IK Trump BF Amstad PA Effect of Bcl-2 on oxidant-induced cell death and intracellular Ca2+ mobilization Am J Physiol 1998 275 C832 9 9730968
Lee SY Andoh T Murphy DL Chiueh CC 17β-estradiol activates ICI 182,780-sensitive estrogen receptors and cyclic GMP-dependent thioredoxin expression for neuroprotection FASEB J 2003 17 947 8 12626428
Scordalakes EM Shetty SJ Rissman EF Roles of estrogen receptor alpha and androgen receptor in the regulation of neuronal nitric oxide synthase J Comp Neurol 2002 453 336 44 12389206 10.1002/cne.10413
Singer CA Figueroa-Masot XA Batchelor RH Dorsa DM The mitogen-activated protein kinase pathway mediates estrogen neuroprotection after glutamate toxicity in primary cortical neurons J Neurosci 1999 19 2455 63 10087060
Singh M Meyer EM Simpkins JW The effect of ovariectomy and estradiol replacement on brain-derived neurotrophic factor messenger ribonucleic acid expression in cortical and hippocampal brain regions of female Sprague-Dawley rats Endocrinology 1995 136 2320 4 7720680 10.1210/en.136.5.2320
Panickar KS Guan G King MA Rajakumar G Simpkins JW 17β-estradiol attenuates CREB decline in the rat hippocampus following seizure J Neurobiol 1997 33 961 7 9407016 10.1002/(SICI)1097-4695(199712)33:7<961::AID-NEU7>3.0.CO;2-6
Willis D Tomlinson A Frederick R Paul-Clark MJ Willoughby DA Modulation of heme oxygenase activity in rat brain and spleen by inhibitors and donors of nitric oxide Biochem Biophys Res Commun 1995 214 1152 6 7575523 10.1006/bbrc.1995.2406
| 15882473 | PMC1142325 | CC BY | 2021-01-04 16:03:49 | no | BMC Neurosci. 2005 May 10; 6:34 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-34 | oa_comm |
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-71587634710.1186/1471-2431-5-7Research ArticleRacemic epinephrine compared to salbutamol in hospitalized young children with bronchiolitis; a randomized controlled clinical trial [ISRCTN46561076] Langley Joanne M [email protected] Michael B [email protected] John C [email protected] Heather [email protected] Cecil R [email protected] Paul [email protected] Clinical Trials Research Centre, Dalhousie University, Halifax, Canada2 Departments of Pediatrics, Dalhousie University, Halifax, Canada3 Departments of Community Health and Epidemiology, Dalhousie University, Halifax, Canada4 Department of Pediatrics, Queen's University Belfast and Craigavon Area Hospital, Craigavon, Northern Ireland5 Department of Psychiatry, Dalhousie University, Halifax, Canada6 Saint John Regional Hospital, Atlantic Health Sciences Corporation, Saint John, Canada7 Division of Allergy, Immunology and Pulmonology, Department of Pediatric and Adolescent Medicine, Mayo Clinic Rochester, MN, USA2005 5 5 2005 5 7 7 24 11 2004 5 5 2005 Copyright © 2005 Langley et al; licensee BioMed Central Ltd.2005Langley 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
Bronchiolitis is the most common cause of lower respiratory tract illness in infancy, and hospital admission rates appear to be increasing in Canada and the United States. Inhaled beta agonists offer only modest short-term improvement. Trials of racemic epinephrine have shown conflicting results. We sought to determine if administration of racemic epinephrine during hospital stay for bronchiolitis improved respiratory distress, was safe, and shortened length of stay.
Methods
The study was a randomized, double-blind controlled trial of aerosolized racemic epinephrine compared to salbutamol every one to 4 hours in previously well children aged 6 weeks to ≤ 2 years of age hospitalized with bronchiolitis. The primary outcome was symptom improvement as measured by the Respiratory Distress Assessment Instrument (RDAI); secondary outcomes were length of stay in hospital, adverse events, and report of symptoms by structured parental telephone interview one week after discharge.
Results
62 children with a mean age of 6.4 months were enrolled; 80% of children had Respiratory Syncytial Virus (RSV). Racemic epinephrine resulted in significant improvement in wheezing and the total RDAI score on day 2 and over the entire stay (p < 0.05). The mean LOS in the epinephrine arm was 2.6 days (95% CI 2, 3.2) v. 3.4 days in those in the salbutamol group (95% CI 2.6, 4.2) (p > 0.05). Adverse events were not significantly different in the two arms. At one week post-discharge, over half of parents reported that their child still had a respiratory symptom and 40% had less than normal feeding.
Conclusion
Racemic epinephrine relieves respiratory distress in hospitalized infants with bronchiolitis and is safe but does not abbreviate hospital stay. Morbidity associated with bronchiolitis as identified by parents persists for at least one week after hospital discharge in most infants.
==== Body
Background
Bronchiolitis accounts for up to 60% of all lower respiratory tract illness in the first year of life [1] and up to 32% of hospitalizations for lower respiratory tract illness in this age group [2]. The incidence of hospitalisation for bronchiolitis in infancy has increased in both the United States [3] and Canada [4] in the last two decades, and a 100% increase in first time hospitalisation for lower respiratory tract illness in children under two years has been noted in northern Europe [5].
Bronchiolitis is characterized by tachypnea and wheezing due to obstruction of small airways, and therefore treatment has often included use of beta and alpha agonists delivered by aerosol in addition to supportive care. A systematic review of randomised clinical trials of the efficacy of beta-agonist aerosols suggests they offer only modest short term improvement [6]. Alpha agonist stimulation of the sympathetic nervous system would be expected to reduce capillary leakage by constricting precapillary arterioles, reducing hydrostatic pressure and consequently bronchial mucosal edema [7]. Since Wohl and Chernick first suggested this intervention in 1978 [8], multiple studies and systematic reviews [9-11] have been published. While there is evidence that acute symptoms of bronchiolitis measured in the short term may improve with epinephrine, these reviews have called for more studies assessing longer-term outcomes such as duration of stay, and that are clinically relevant to parents, clinicians and the health care system.
We report a randomised controlled trial of aerosolised epinephrine compared to salbutamol throughout hospital stay in infants with bronchiolitis to assess daily clinical improvement (respiratory distress, feeding), length of hospital stay and adverse events, and outcomes by parental report one week after discharge to the community.
Methods
This study was a randomized, double blind controlled trial of racemic epinephrine (Vaponefrin solution 2.25%, Aventis Pharma Inc, Laval, PQ) compared to control (salbutamol, Ventolin, GlaxoSmithKline Inc., Mississauga, ON) in children requiring hospitalization for management of bronchiolitis.
Patient population
Eligible children were aged greater than 6 weeks to ≤ 2 years with a clinical diagnosis of bronchiolitis by the admitting physician. Wheezing had to be present on physical examination and was defined as a high-pitched, musical, continuous respiratory sound. Only patients admitted for management of bronchiolitis were eligible. The parent or guardian had to be able cooperate with study requirements (ability to speak, read and write English, have a telephone at home and not expected to move within the next month). The participating institutions were the IWK Health Centre in Halifax, Nova Scotia and the Saint John Regional Hospital (SJRH) in Saint John, New Brunswick. The IWK is a university-affiliated primary and tertiary-care pediatric hospital with an urban population of 300,000 and is a referral center for the Maritime provinces (population 2 million) of Canada. SJRH serves a rural-urban population of 200,000.
Children were not eligible for enrollment if they had had a previous diagnosis of asthma, were critically ill, or had chronic pulmonary or cardiac disease. Other exclusion criteria included: allergy to sodium metabisulfite, presence of tachycardia exceeding 200 beats per minute, or use of glucocorticoids, sympathomimetic amines or monoamine oxidase inhibitor therapy.
Informed consent was obtained from the parent or guardian prior to enrolment. The protocol was approved by the Ethics Review Board at both participating institutions.
Study procedures
Study enrolment occurred in sequential winter respiratory seasons (November to April) from 1999 to 2002. Families were approached regarding study participation in the emergency department or within 24 hours of admission. Research nurses were available to enroll patients between 8 am and 8 pm.
Treatment allocation was determined by randomization, performed in blocks of four by the pharmacy department using a computer-generated random numbers table. Study drug was packaged in identical multidose vials labeled "study drug" with a code number. Both salbutamol and racemic epinephrine are clear, colorless liquids that are indistinguishable [12]. Participants were allocated to racemic epinephrine, 0.5 mls of 2.25% (Vaponefrin solution, Aventis Pharma, Montreal Quebec) or salbutamol respirator solution (Pharmel Inc., Montreal, Canada) by aerosol. Study drug was administered every one to four hours or more frequently at the request of the attending physician. Study drug was delivered by a wall flowmeter-nebulizer with face mask (Hospitak Inc., Farmingdale, NC) with oxygen at 5 to 7 L/min. A standard order sheet was used to ensure consistency of trial methodology. Salbutamol was given in 3 ml normal saline at a dosage of 1.5 mg for children weighing more than 10 kg, 1.25 mg for children >6 kg and < 10 kg, and 0.75 mg for those weighing less than 6 kgs. The heart rate was measured continuously during each aerosol and for one hour after. The heart rate, vomiting, presence of tremors or pallor were recorded in the health record by the bedside nurse at the end of every aerosol and one hour post aerosol.
Data collection
Baseline demographic data collected at study entry included inclusion and exclusion criteria, age, gender, concomitant medications and other illnesses. The caregiver/parent was asked to describe the child's feeding pattern (normal, less than normal, unable to feed). At the time of study enrolment and then daily (every morning) thereafter the study nurse measured oxygen saturation and wheezing and retractions using the Respiratory Distress Assessment Instrument (Table 1) [13], which was the primary outcome measure of the study. Oxygen saturation was measured using a pulse oximeter (Nellcor Pulse Oximeter, Nellcor Puritan Bennet Inc., Pleasanton, CA) with the infant in a quiet state after breathing room air for at least 10 minutes. If the oxygen saturation went below 85%, the measurement was halted. At the daily assessment, the study nurse interviewed caregivers and reviewed the health record to determine if adverse events were present (vomiting, tremors, pallor), the feeding pattern and recorded the maximum daily heart rate for that 24-hour period.
Table 1 Respiratory Distress Assessment Instrument (From: Lowell DI, Lister G, Von Kloss H, McCarthy P. Wheezing in infants: the response to epinephrine. Pediatrics 1987; 87:939-45.)
0 1 2 3 4 Maximum Points
Wheezing
Expiration None End 1/2 3/4 All 4
Inspiration None Part All ... ... 2
Location None Segmental Diffuse ... ... 2
Retractions
Supraclavicular None Mild Moderate Marked ... 3
Intercostal None Mild Moderate Marked ... 3
Subcostal None Mild Moderate Marked 3
Total ... ... ... 17
* Within each variable (wheezing, retractions) the subscores are summed to give a total score. The maximum total points for wheezing is 8 and for retractions is 9.
During the first two study enrolments, it was noted that a bright red nasal discharge was observed in some study participants, and interpreted by bedside nurses as bloody nasal discharge. This discoloration of nasal mucous was found to be a known effect of administration of aerosolized epinephrine, which is caused by the oxidation of the sulphite stabilizer. This effect was not known to the investigators at the time of study design and is not in the drug monograph, but has been reported in a recent trial of epinephrine in the emergency department setting [14]. Because this could lead to unblinding of treatment allocation, an amendment to the study protocol was made for all subsequent patients whereby the bedside nurse wiped the nose of study participants after each study drug administration and immediately before the study nurse performed the daily respiratory assessment.
A nasopharyngeal aspirate for Respiratory Syncytial virus (RSV) antigen was routinely done in participating hospitals to determine appropriate placement for infection control purposes. At the discretion of the attending physician, some children had respiratory tract samples submitted for respiratory virus culture (RSV, influenza, parainfluenza, adenovirus).
A secondary outcome measure was duration of hospital stay, measured using a method previously validated by the Pediatric Investigators Collaborative Network on Infections in Canada studies of hospitalized children with RSV infection [15]. Each day the study nurse assessed which of four reasons accounted for ongoing hospitalization 1) patient receiving drug treatment for bronchiolitis 2) patient receiving oxygen supplementation or parenteral fluids because of bronchiolitis 3) patient hospitalized because of underlying (pre-existing) illness only or 4) awaiting transport home or uncertain home environment. Only those days on which the reason for hospitalization were one or more of receiving medication for bronchiolitis (1) or oxygen supplementation or parenteral fluids because of bronchiolitis (2), were recorded as valid hospital days. Discharge timing, counted as the time the decision was made to discharge home, was at the discretion of the attending physician. Study personnel had no involvement in discharge planning and did not impose any discharge criteria.
All parents/guardians were telephoned seven days after hospital discharge by a research assistant to collect data about the child's convalescence: respiratory symptoms (retractions, wheezing), feeding pattern (normal, less than normal, unable to feed), adverse events from medication (shakiness, tremors, pallor," other problems") and whether they had required a visit to a physician or to the emergency department or hospital. The interviewer read closed-ended questions from a standard script.
Adverse events were collected during the hospital stay and during the post-discharge telephone call. The event was described and categorized according to severity (mild, moderate, severe), outcome (recovered fully, recovered with sequelae, ongoing, death) and relationship to study drug (related, probably or possibly related, unrelated, unable to classify). Mild adverse events were defined as "awareness of signs and symptoms, easily tolerated and require no interventions", moderate as "discomfort sufficient enough to interfere with normal activities and/or result in some sort of intervention" and severe as " inability to perform normal activities, distressing and/or incapacitating and definitely require intervention and/or medical attention."
The sample size was calculated to detect a difference in the RDAI score between day one and day three. The estimated sample size for a two-sample comparison of proportions of each group that achieved the four-unit difference was with a probability of type one error of 0.05 and type 2 error of 0.8 was 33 infants per group. The standard deviations were based on previously reported changes in RDAI in bronchiolitis [13].
Analysis
All randomized children were considered in the analysis. All analyses were performed using SAS 8.02 software (SAS Institute Inc., Cary, USA). Proportions and exact binomial intervals were calculated for discrete variables and comparisons between treatment groups were made using the Fisher's exact test. Summary statistics (mean, median, standard deviation, minimum and maximum) were calculated for continuous variables and comparisons were made between treatment groups using the Wilcoxon rank-sum test. Comparisons of trend across time were made using repeated measures analysis of variances. The RDAI was treated as a continuous measure. P-values less than or equal to 0.05 were considered statistically significant.
Results
Sixty-two children were enrolled with 31 in each treatment arm. Ten were enrolled from the Saint John site and 52 from the Halifax site. All participants completed in-hospital follow-up but the parent of one child in the epinephrine group could not be contacted for the one-week post-discharge telephone call. The mean age of participants was 6.4 months. RSV was identified in nasopharyngeal samples of 81.5% of children randomized to epinephrine and 78.6% of those randomized on salbutamol. The two groups were similar at baseline; characteristics at enrolment are seen in Table 2. On admission, only 15% of infants had a normal feeding pattern; 83% had decreased feeding by parental/caregiver report and 7% were unable to feed. 245 children were screened who did not enroll: 46 did not meet inclusion criteria and in 34 parents refused consent. The rest were ineligible because of exclusion criteria. The most commons reason for exclusion was previous diagnosis of asthma (n = 58) and previous administration of systemic steroids (n = 44).
Table 2 Characteristics of children less than two years of age admitted to hospital with bronchiolitis by treatment group, at enrolment.
Outcome Racemic Epinephrine Salbutamol
Male 45.20 % 64.50 %
Age (median) 5.49 months 3.32 months
RDAI total 7.35 (95% CI 6.52, 8.19) 8.29 (95% CI 7.24, 9.34)
Wheezing 4.23 (95% CI 3.60, 4.85) 5.06 (95% CI 4.28, 5.85)
Retractions 3.13 (95% CI 2.64, 3.62) 3.23% (95% CI 2.71, 3.74)
Oxygen saturation (mean) 94.6 % 93.5 %
Abnormal feeding pattern (less than normal or not feeding) 90% ((95% CI 7.35, 97.9) 80% (95% CI 61.4, 92.3)
RSV positive 81.5% (61.9, 93.7) 78.6 (95% CI 59, 91.7)
RDAI = Respiratory Distress Assessment Instrument, CI = confidence interval
Racemic epinephrine resulted in a significant improvement in wheezing compared to salbutamol on day 2 (p = 0.01) and over the entire hospital stay (p = 0.01) (Table 3), but not on other days. The total RDAI (wheezing and retractions) in children receiving racemic epinephrine was also significantly better on the second hospital day and over the entire stay (p = 0.02). On the third hospital day a significant difference in oxygen saturation was observed in children receiving racemic epinephrine compared to those receiving salbutamol (96.20% v. 93.80%, 98.80 v. 92.00 p = 0.03) but this difference was not significant when the two groups were compared over the entire hospital stay or on other days.
Table 3 Respiratory Distress Assessment Instrument scores by hospital day in children with bronchiolitis randomized to aerosolized racemic epinephrine or salbutamol.
Hospital Day (Number of children in hospital for that duration) Total Wheezing score Retractions Total RDAI score
Racemic epinephrine Salbutamol Racemic epinephrine Salbutamol Racemic epinephrine Salbutamol
0 (n = 62) 4.23 5.06 3.13 3.23 7.35 8.29
1 (n = 58) 2.04 3.70 1.64 1.64 3.68 6
2 (n = 40) 1.59 2.65 1.06 1.06 2.65 4.13
3 (n = 22) 2.09 2.73 1.09 1.09 3.18 4.27
4 (n = 9) 2.75 4.00 1.75 1.75 4.5 6.4
5 (n = 9) 2.00 2.60 0.75 0.75 2.75 4.2
The mean length of stay for children randomized to the racemic epinephrine group was 2.60 days (95% CI 2.00, 3.20) and 3.40 days for those randomized to salbutamol (95% CI 2.60, 4.20) (p > 0.05). No significant differences in length of stay or RDAI scores were seen when children confirmed as having RSV infection were compared to those without RSV.
There was no difference in wheezing score, total retraction score, total respiratory score oxygen saturation or duration of stay by study site.
Any adverse event (mild, moderate, severe) was reported in 45.20% (14/31) of children who received epinephrine compared to 51.60% (16/31) of those who received salbutamol (p > 0.05). There was one severe adverse event (fever greater than 39°C rectal), judged unrelated to study medication, which occurred in a child on salbutamol. Tremors and pallor were more common in children receiving racemic epinephrine than in those on salbutamol, but these differences were not statistically significant (19.40% (6/31) v. 9.70% (3/31) and 19.40% (6/31) v. 6.50% (2/31) respectively). Vomiting occurred in 19.4% (6/31) of those receiving epinephrine and 25.8% (8/31) of those on salbutamol; this difference was not statistically significant.
At the follow-up telephone call, one week after discharge over 60% of infants had at least one ongoing respiratory symptom (Table 4) and most were still not considered to be feeding normally by their parents. No statistically significant differences in outcomes at the follow-up phone call were identified.
Table 4 Prevalence of respiratory symptoms and drug-related adverse events one week after discharge in children randomized to aerosolized racemic epinephrine or salbutamol during their hospital stay for bronchiolitis.
Clinical event as reported by parent or guardian Percentage of children with symptom at one week post-discharge phone call (absolute numbers)* P value for comparison of proportions
Racemic epinephrine Salbutamol
Breathing difficulty 20.00 (6/30) 19.4 (6/31) 1.00
Wheezing 56.7 (17/30) 67.7 (21/31) 0.43
Chest retractions 13.3 (4/30) 25.8 (8/31) 0.33
Feeding pattern:
Normal feeding 50 (15/30) 54.8 (17/31) 0.80
Less than normal 50 (15/30) 41.9 (13/31) 0.61
Unable to feed 0 (0/31) 3.2 (1/31) 1.00
Vomiting 20 (6/30) 35.5 (11/31) 0.25
Tremors 0 (0/31) 0 (0/31) N/A
Pallor 13.3 (4/30) 6.5 (2/31) 0
Has had a visit to a physician 23.3 (7/30) 19.4 (6/31) 0.42
Has visited an emergency department or been hospitalized 3.3 (1/30) 3.2 (1/31) 1.00
N/A = not applicable; * Not mutually exclusive
Discussion
In this study, treatment of bronchiolitis with aerosolised racemic epinephrine over the course of a child's hospital stay was associated with improvement in respiratory symptoms, but did not result in a statistically significant difference in hospital length of stay. Although previous reports have found that nebulized epinephrine results in short-term clinical improvement in bronchiolitis [12,16-21], its effect on duration of hospital stay or need for admission is less clear. Two of four randomised clinical trials in the emergency department setting using 1 to 3 doses of epinephrine have found a difference in admission rate [12,19] and two have not [22,23]. It is possible that small reductions in length of stay could be detected by larger trials than those conducted thus far. Our study is only the second [24] in which nebulized racemic epinephrine was provided throughout the hospital admission. Other trials have administered one to three doses spaced on one day only. One might expect that short-term improvement could lead to the clinician's judgement that admission was not necessary. Two [20,21] of four randomised trials [24,25] have suggested that length of stay in hospital is abbreviated in children receiving epinephrine. The subgroup of children that might benefit from this therapy is not clear and larger trials will be necessary to identify if and when administrations of nebulized epinephrine during hospitalisation results in patient benefit. Our study supports the thesis that while airway edema may be improved following administration of racemic epinephrine it is not sustained and does not alter the natural history of bronchiolitis in infants. The inflammatory process initiated by RSV or other respiratory viruses is unaffected and mucous secretion and edema recurs after the effect of epinephrine has dissipated.
We choose salbutamol as a control for epinephrine because it was the local standard of care at the time our trial was designed and it was considered unethical to withhold a potentially beneficial therapy, even though that benefit was likely minimal [6]. Other randomized trials of epinephrine in hospitalized children have used as the control normal saline [22,25], salbutamol [21] or both [24].
Interestingly we noted a risk to unblinding of treatment allocation early in our trial, by the oxidation of the sulphite preservatives, which turn mucous red or brown in recipients of racemic epinephrine. We implemented measures to avoid unblinding, but could find no mention of this event in previous studies of epinephrine in bronchiolitis. If future trials of racemic epinephrine are conducted, study design should carefully address this possibility and consider using another formulation of epinephrine. Patel et al questioned study personnel and ward staff after completion of their trial and found no difference in the proportion of correct guesses as to allocation by treatment group [24].
Hospital admission for bronchiolitis occurs when the infant has significant respiratory distress or is unable to feed because of the work of breathing. By this point in the evolution of bronchiolitis, lower respiratory tract inflammation is well established and may be difficult to alter. Perhaps for this reason, trials of steroid therapy in infants hospitalized for bronchiolitis have shown no benefit [11]. Length of hospital stay or avoidance of hospital admission is a salient outcome measure for intervention trials of epinephrine because institutional care represents the largest component of direct expenditures (over 60%) for bronchiolitis [26]. Morbidity associated with bronchiolitis persisted for at least one week after hospital discharge in our population. Reducing costs and morbidity due to bronchiolitis may require more than one intervention, each at specific times during the illness. For example Schuh et al demonstrated a reduction in hospitalization in infants with bronchiolitis treated with dexamethasone (1 mg/kg) in the emergency department [27]. Bisgaard et al demonstrated a reduction in post-RSV symptomatology, principally cough, in infants treated with montelukast within the first week of illness [28]. All of these treatments, if their efficacy were borne out in larger trials, would reduce health care costs associated with this ubiquitous infection. Such studies will need to be large to capture clinically significant outcomes.
Use of racemic epinephrine multiple times over several days was not associated with significant adverse events compared to salbutamol. Epinephrine is a potent adrenergic agonist with potential cardiovascular side effects including tachycardia or bradycardia and hypertension. In the doses used for bronchiolitis such adverse events have not been reported. Given that short-term improvement may occur and its favourable safety profile, it seems reasonable to use aerosolized racemic epinephrine selectively for infants with acute distress to decrease the work of breathing or to avoid assisted ventilation.
Conclusion
Racemic epinephrine relieves respiratory distress (wheezing, retractions) in infants hospitalized for management of bronchiolitis and is safe but does not abbreviate hospital stay. Morbidity associated with bronchiolitis as identified by parents persists for at least one week after hospital discharge in most infants.
Abbreviations
Respiratory Distress Assessment Instrument (RDAI)
Respiratory Syncytial Virus (RSV)
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JML and MBS conceived the study and prepared the protocol. JCL planned the statistical analysis. JCL and PP provided intellectual input to study design. JML drafted the manuscript. JML, MBS and CBO supervised acquisition of study data. HJ conducted the analysis. All authors contributed to interpretation of study results and critically reviewed the manuscript for important intellectual content.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was funded by the Nova Scotia Lung Foundation (Legacy Award) and the IWK Health Centre Research Office Research Award, Canada. Vaponephrine was generously supplied by Aventis Pharma Inc. (Montreal, Canada)
The authors would like to thank the research coordinators Heather Samson, Trish Pottie and Ardith Ambrose and the nurses of the general pediatric floors at the IWK Health Center and Saint John Regional.
==== Refs
Wright AL Taussig LM Ray CG Harrison HR Holberg CJ The Group Health Medical Associates The Tucson Children's Respiratory Study II. Lower respiratory tract illness in the first year of life. Am J Epidemiol 1989 129 1232 1246 2729259
McConnochie KM Roghmann KJ Liptak GS Hospitalization for lower respiratory tract illness in infants:variation in rates among counties in New York State and areas within Monroe County. J Pediatr 1995 126 220 229 7844668
Shay DK Holman RC Newman RD Liu LL Stout JW Anderson LJ Bronchiolitis-associated hospitalizations among US children, 1980-1996 Jama 1999 282 1440 1446 10535434 10.1001/jama.282.15.1440
Langley JM LeBlanc JC Smith B Wang EE Increasing incidence of hospitalization for bronchiolitis among Canadian children, 1980-2000 J Infect Dis 2003 188 1764 1767 14639549 10.1086/379740
Wickman M Farahmand BY Persson PG Pershagen G Hospitalization for lower respiratory disease during 20 yrs among under 5 yr old children in Stockholm County: a population based survey Eur Respir J 1998 11 366 370 9551740 10.1183/09031936.98.11020366
Kellner JD Ohlsson A Gadomski AM Wang EE Bronchodilators for bronchiolitis Cochrane Database Syst Rev 2000 CD001266 10796626
Barr FE Patel NR Newth CJL The pharmacologic mechanism by which inhaled epinephrine reduces airway obstruction in respiratory syncytial virus-associated bronchiolitis. J Pediatr 2000 16 699 700 10802508
Wohl ME Chernick V State of the art: bronchiolitis Am Rev Respir Dis 1978 118 759 781 212970
Hartling L Wiebe N Russell K Patel H Klassen TP A meta-analysis of randomized controlled trials evaluating the efficacy of epinephrine for the treatment of acute viral bronchiolitis Arch Pediatr Adolesc Med 2003 157 957 964 14557155 10.1001/archpedi.157.10.957
Hartling L Wiebe N Russell K Patel H Klassen T Epinephrine for bronchiolitis Cochrane Database Syst Rev 2004 1 CD003123 14974006
King VJ Viswanathan M Bordley WC Jackman AM Sutton SF Lohr KN Carey TS Pharmacologic treatment of bronchiolitis in infants and children: a systematic review Arch Pediatr Adolesc Med 2004 158 127 137 14757604 10.1001/archpedi.158.2.127
Menon K Sutcliffe T Klassen TP A randomized trial comparing the efficacy of epinephrine with salbutamol in the treatment of acute bronchiolitis J Pediatr 1995 126 1004 1007 7776075
Lowell DI Lister G Von Koss H McCarthy P Wheezing in infants: the response to epinephrine Pediatrics 1987 79 939 945 3295741
Plint AC Osmond MH Klassen TP The efficacy of nebulized racemic epinephrine in children with acute asthma: a randomized, double-blind trial Acad Emerg Med 2000 7 1097 1103 11015240
Wang EEL Law BJ Boucher FD Stephens D Robinson JL Dobson S Langley JM McDonald J MacDonald NE Mitchell I Pediatric Investigators Collaborative Network on Infections in Canada (PICNIC) study of admission and management variation in patients hospitalized with respiratory syncytial viral lower respiratory tract infection. J Pediatr 1996 129 390 395 8804328
Sanchez I De Koster J Powell RE Wolstein R Chernick V Effect of racemic epinephrine and salbutamol on clinical score and pulmonary mechanics in infants with bronchiolitis J Pediatr 1993 122 145 151 8419602
Reijonen T Korppi M Pitkakangas S Tenhola S Remes K The clinical efficacy of nebulized racemic epinephrine and albuterol in acute bronchiolitis Arch Pediatr Adolesc Med 1995 149 686 692 7767427
Kristjansson S Lodrup Carlsen KC Wennergren G Strannegard IL Carlsen KH Nebulised racemic adrenaline in the treatment of acute bronchiolitis in infants and toddlers Arch Dis Child 1993 69 650 654 8285776
Ray MS Singh V Comparison of nebulized adrenaline versus salbutamol in wheeze associated respiratory tract infection in infants Indian Pediatr 2002 39 12 22 11805349
Mandelberg A Tal G Witzling M Someck E Houri S Balin A Priel IE Nebulized 3% hypertonic saline solution treatment in hospitalized infants with viral bronchiolitis Chest 2003 123 481 487 12576370 10.1378/chest.123.2.481
Bertrand P Aranibar H Castro E Sanchez I Efficacy of nebulized epinephrine versus salbutamol in hospitalized infants with bronchiolitis Pediatr Pulmonol 2001 31 284 288 11288210 10.1002/ppul.1040
Hariprakash S Alexander J Carroll W Ramesh P Randell T Turnbull F Lenney W Randomized controlled trial of nebulized adrenaline in acute bronchiolitis Pediatr Allergy Immunol 2003 14 134 139 12675760 10.1034/j.1399-3038.2003.00014.x
Mull CC Scarfone RJ Ferri LR Carlin T Salvaggio C Bechtel KA Trephan MA Rissman RL Gracely EJ A randomized trial of nebulized epinephrine vs albuterol in the emergency department treatment of bronchiolitis Arch Pediatr Adolesc Med 2004 158 113 118 14757602 10.1001/archpedi.158.2.113
Patel H Platt RW Pekeles GS Ducharme FM A randomized, controlled trial of the effectiveness of nebulized therapy with epinephrine compared with albuterol and saline in infants hospitalized for acute viral bronchiolitis J Pediatr 2002 141 818 824 12461499 10.1067/mpd.2002.129844
Wainwright C Altamirano L Cheney M Cheney J Barber S Price D Moloney S Kimberley A Woolfield N Cadzow S Fiumara F Wilson P Mego S VandeVelde D Sanders S O'Rourke P Francis P A multicenter, randomized, double-blind, controlled trial of nebulized epinephrine in infants with acute bronchiolitis N Engl J Med 2003 349 27 35 12840089 10.1056/NEJMoa022226
Langley JM Wang EEL Law BJ Stephens D Boucher FD Dobson S McDoanlad J MacDonald NE Mitchell I Robinson JL Economic evaluation of respiratory syncytial virus infection in Canadian children: a Pediatric Investigators Collaborative Network on Infections in Canada (PICNIC) study J Pediatr 1997 131 113 117 9255201
Schuh S Coates AL Binnie R Allin T Goia C Corey M Dick PT Efficacy of oral dexamethasone in outpatients with acute bronchiolitis J Pediatr 2002 140 27 32 11815760 10.1067/mpd.2002.120271
Bisgaard H A randomized trial of montelukast in respiratory syncytial virus postbronchiolitis Am J Respir Crit Care Med 2003 167 379 383 12406832 10.1164/rccm.200207-747OC
| 15876347 | PMC1142326 | CC BY | 2021-01-04 16:31:07 | no | BMC Pediatr. 2005 May 5; 5:7 | utf-8 | BMC Pediatr | 2,005 | 10.1186/1471-2431-5-7 | oa_comm |
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-81588245210.1186/1471-2431-5-8Research ArticleTechnology-dependency among patients discharged from a children's hospital: a retrospective cohort study Feudtner Chris [email protected] Nanci Larter [email protected] Barbara [email protected] Virginia [email protected] Ross M [email protected] John M [email protected] Pediatric Advanced Care Team, Integrated Care Service, and Pediatric Generalist Research Group, The Children's Hospital of Philadelphia; Philadelphia, USA2 The Leonard Davis Institute and the Center for Bioethics, University of Pennsylvania, Philadelphia, USA3 Center for Children with Special Needs, Children's Hospital & Regional Medical Center, Seattle, USA4 Pediatric Palliative Care Program, Children's Hospital and Regional Medical Center, Seattle, USA5 Department of Rehabilitation Medicine, University of Washington, Seattle, USA6 Department of Pediatrics, University of Washington, Seattle, USA2005 9 5 2005 5 8 8 27 10 2004 9 5 2005 Copyright © 2005 Feudtner et al; licensee BioMed Central Ltd.2005Feudtner 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
Advances in medical technology may be increasing the population of children who are technology-dependent (TD). We assessed the proportion of children discharged from a children's hospital who are judged to be TD, and determined the most common devices and number of prescription medications at the time of discharge.
Methods
Chart review of 100 randomly selected patients from all services discharged from a children's hospital during the year 2000. Data were reviewed independently by 4 investigators who classified the cases as TD if the failure or withdrawal of the technology would likely have adverse health consequences sufficient to require hospitalization. Only those cases where 3 or 4 raters agreed were classified as TD.
Results
Among the 100 randomly sampled patients, the median age was 7 years (range: 1 day to 24 years old), 52% were male, 86% primarily spoke English, and 54% were privately insured. The median length of stay was 3 days (range: 1 to 103 days). No diagnosis accounted for more than 5% of cases. 41% were deemed to be technology dependent, with 20% dependent upon devices, 32% dependent upon medications, and 11% dependent upon both devices and medications. Devices at the time of discharge included gastrostomy and jejeunostomy tubes (10%), central venous catheters (7%), and tracheotomies (1%). The median number of prescription medications was 2 (range: 0–13), with 12% of cases having 5 or more medications. Home care services were planned for 7% of cases.
Conclusion
Technology-dependency is common among children discharged from a children's hospital.
==== Body
Background
What proportion of children discharged from children's hospitals are technology-dependent (TD)? Advances in medical technology over the past century have profoundly changed the landscape of pediatric ambulatory and hospital care [1-3]. Through the combined effects of pharmaceuticals and medical devices, parents and health care professionals now care for a population of children who depend on medical technology to live or remain in their current state of health.
TD children and the complexity and costs of their care gained widespread attention in the United States during the 1980s. In a November 1981 news conference President Reagan cited the case of Katie Beckett, a 3-year-old Iowa girl who (due to lack of any means to fund home care) had lived in a hospital since having viral encephalitis at 3 months of age [4]. Two days later, a special waiver was issued for Ms. Beckett's case [5], and by the following summer the Secretary of the US Department of Health and Human Services had established a waiver program that enabled an individual residing at home to continue to be covered by Medicaid and receive Medicaid-funded long term care services [6]. In December 1982, the US Surgeon General sponsored a conference that focused on the example of "the ventilator-dependent child" [7]. Then in 1984 the US Supreme Court ruled that a child with spina bifida who required clean intermittent catheterization had the right to receive such a service in the school setting in order to benefit from special education [8].
The well-being of TD children, and the effect that caring for these children has on the well-being of their families and of society, is an important concern for clinicians, ethicists, and healthcare policy analysts [9]. The goals of maximizing the well-being of TD children, their families, and society (Figure 1) are supported by information generated by various studies such as an evaluation of pediatric home-ventilator programs [10]; cross-sectional observational population-based epidemiologic studies documenting the prevalence of technology-dependent children [11-14]; and surveys focusing on the health, psychosocial, and financial impact on mothers, parents, and families [15-23]. Given that dependence on technology is a complex multidimensional construct (Figure 2), each of these studies has had to define "technology-dependent," a task that has fundamental implications for the ensuing research. The most commonly cited definition of TD originated in a 1987 report issued by the former federal Office of Technology Assessment (OTA), which proposed as a "working definition" the scheme outlined in Figure 3[11]. The OTA report underscored how their hierarchical definition reflected both consensus expert opinion that OTA groups I-III were technology-dependent and the government's desire to limit expenditures by excluding OTA group IV from the definition, since the prevalence of children meeting the group IV definition was so large.
Figure 1 Goal and means for the understanding of technology-dependency.
Figure 2 Dimensions of dependency on technology. Examples enclosed in parentheses are meant only to illustrate extremes of the dimensions. CSF, cerebrospinal fluid; MDI, metered dose inhaler.
Figure 3 Definitions of technology-dependent.
Our research group, in an effort to improve the quality of services offered to hospitalized technology-dependent children and their families, developed a definition of TD children that emphasizes the failure consequence dimension of the TD construct. As indicated in Figure 3, we specified children as TD if the failure or cessation of a device or drug that they were using would have the likely consequence of a required hospitalization. While our definition does not explicitly encompass other dimensions of the TD construct, hospitalization would also be indicative of greater hindrance, system-reliance, and cost. Of note, our definition did not specify a minimal duration of dependency, since from a quality-improvement perspective the duration of dependency is not necessarily the most important dimension of the underlying TD construct. Although most studies of TD have focused only on reliance on machines or devices, we also considered medications as a form of technology, in keeping with the definition of technology put forth in the OTA report [11], and in keeping with our underlying concern regarding the effectiveness, safety, and burden of medical interventions that persist beyond the period of hospitalization.
Guided by this definition, we sought to determine the proportion of children discharged from a tertiary care children's hospital who are TD, describe their characteristics, and examine aspects of the care provided to these children upon discharge.
Methods
After obtaining Human Subjects Research Committee approval, we conducted a retrospective cohort study of a random sample of 100 patients admitted to Children's Hospital and Regional Medical Center in Seattle, Washington, a regional tertiary care center with approximately 11,000 hospitalizations annually. Electronic administrative data regarding all discharges during calendar year 2000 was used to select the subjects using a random number table. Eligible subjects included all pediatric patients admitted to any hospital service and of any age; adults admitted for the purpose of tissue or organ donation to a pediatric patient were excluded. For subjects who were discharged more than once during 2000, we randomly selected a single hospitalization episode. No patient was randomly selected twice.
For all 100 randomly-selected eligible hospitalizations of 100 patients, two investigators conducted a structured hospitalization chart review and abstraction using a data collection instrument. Prior to the conduct of this study, we had developed this data collection instrument specifically for this project, and pilot tested the instrument by having both investigators independently abstracting the same series of 5 hospitalization records and refining the instrument to assure consistent agreement regarding the abstracted data. Data collected included demographic information, all documented diagnoses, procedures, technology use, medication orders, feeding orders, and nursing orders, as well as whether home nursing services were ordered for post-discharge care.
Four investigators, consisting of three pediatricians and one nurse with a range of 6 to 40 years experience caring for children with special health care needs, then independently reviewed all 100 chart abstractions and determined whether each subject at the time of discharge from the hospital met the definition specified in Figure 3. We also sought to situate the construct of TD within the construct of children with special health care needs (CSHCN), which has been defined by the Maternal and Child Health Bureau of the United States as: "All children who have, or are at increased risk for, chronic physical, developmental, behavioral, or emotional conditions and who also require health and related services of a type or amount beyond that required by children generally." The MCHB definition is further refined by the definition of "chronic" as conditions expected to last 6 months or longer.
Given limitations in the data available in the hospitalization record and the nature of the TD and CSHCN definitions, the 4 investigators had to employ their knowledge and judgment to make the determination of TD or CSHCN. To assess the degree of consensus in these judgments, we used a 4-rater kappa statistic to compare the four evaluators' independent dichotomous ("yes or no") classifications of each subject regarding TD (kappa = 0.56) and CSHCN (kappa = 0.59). These values for the 4-rater kappa indicate moderate verging on good agreement. For the subsequent analysis of the data, only those cases where 3 or 4 raters independently classified the subject as either TD or CSHCN were considered to be TD or CSHCN. Examples of cases where only 2 of the 4 investigators interpreted the chart abstract data as indicated TD (and thus were not considered TD for the purposes of this study) included: a patient with the diagnoses of a brain tumor and obstructive hydrocephalus but no documented evidence of a cerebrospinal fluid shunt; a patient with a diagnosis of a malignancy but no documentation of discharge medications; and several instances of patients diagnosed with apparent persistent asthma but no documentation of a controlled medication having been prescribed at the time of discharge.
Characteristics of the sample subjects were then examined by calculating medians, ranges, and proportions. Inferences drawn from this random sample regarding the composition of the complete source population (that is, all children discharged from this hospital) were expressed as 95% confidence intervals of the sample proportions, derived from the binomial distribution. All analyses were conducted using Stata 8.2 (StataCorp, College Station, TX).
Results
Within our random sample of 100 patients admitted to a regional pediatric tertiary care center (Table 1), the mean age was 7.6 years (median 7.1 years). Just over half (52%) of the subjects were male. The majority (86%) were identified as speaking English as their primary language, and 54% as covered by private insurance.
Table 1 Demographic characteristic of subjects
Percentage (%) (n = 100) 95% CI
Age
Less than 1 month 7 3 – 14
1 – 11 months 15 9 – 24
1 – 4 years 21 13 – 30
5 – 9 years 21 13 – 30
10 – 14 years 24 16 – 34
15 – 24 years 12 6 – 20
Gender
Male 52 42 – 62
Female 48 38 – 58
Language
English 86 78 – 92
Spanish 7 3 – 14
Other 3 0.1 – 9
Unknown 4 0.1 – 10
Principal Insurance
Private 54 44 – 64
Government 40 30 – 50
Self-insured or none 4 1 – 10
Unknown 2 0.2 – 7
The primary reasons for hospitalization (Table 2) were led by cancer and the treatment of cancer (15%), followed by respiratory infections (9%), asthma (5%), gastroenteritis (5%), appendicitis (4%), epilepsy or seizures (3%), and the malfunction of a device or graft (3%). Most patients stayed 3 days or less (59%), although 14% of subjects were hospitalized 8 days or more (maximum of 103 days) during this hospitalization. Most subjects were discharged to home (91%), one subject died during the hospitalization, and 3% were transferred to another facility. Home nursing care was arranged for 8% of the subjects, and 7% had other forms of home care ordered.
Table 2 Medical characteristics of subjects
Percentage (%) (n = 100) 95% CI
Most frequent diagnostic categories *
Neoplasm and chemotherapy 15 9 – 24
Respiratory infections 9 4 – 16
Asthma 5 2 – 11
Gastroenteritis 5 2 – 11
Appendicitis 4 1 – 10
Epilepsy or seizure 3 0.6 – 9
Malfunctioning device or graft 3 0.6 – 9
Length of stay
1 day 10 5 – 18
2 days 25 17 – 35
3 days 24 16 – 34
4–7 days 27 17 – 37
8 to 103 days 14 8 – 22
Devices *
Gastrostomy or jejuenostomy tubes 10 5 – 18
Central venous line 7 3 – 14
Medication nebulizer 7 3 – 14
Ventriculoperitoneal shunt 2 0.2 – 7
Tracheotomy 1 0.03 – 5
Number of medications §
0 18 11 – 27
1 20 13 – 29
2 26 18 – 36
3 16 10 – 26
4 7 3 – 14
5 or more 12 6 – 20
Number of daily medication administrations noted ‡
0 or PRN 33 24 – 44
1–4 30 21 – 40
5–9 24 16 – 34
10–22 12 6 – 20
Home nursing care 8 4 – 15
Other forms of home care 7 3 – 14
Disposition
Home 91 84 – 96
Transfer 3 0.6 – 9
Death 1 0.03 – 6
Unknown 5 2 – 11
* For the categories of diagnoses and devices, a given subject could be counted multiple times if noted to have had multiple diagnoses or devices. §Medications were counted as those medications ordered at the time of discharge from the hospital; accordingly, the one case that died is omitted so that n = 99 for this variable as well as "number of daily medication administrations". ‡ Number of daily medication administrations is a count of the number of times any medication dose is delivered in a day, so that a count of 10 could indicate 2 medications given 5 times a day, or 10 medications each given once.
A quarter of all subjects (26%) used some form of medical device (Table 2), including gastrostomy or jejenonostomy tubes (10%), central venous catheters (7%), medication nebulizer (7%), ventriculoperitoneal cerebrospinal fluid shunts (2%), or tracheotomies (1%). Only 18% of subjects were discharged from the hospital with no medications ordered, whereas the majority of subjects (61%) had 2 or more prescriptions ordered, and 12% had 5 or more medications ordered. Focusing on one parameter that contributes to the burden of care – namely, the number of times each day that any form of any medication was ordered to be administered – 12% of subjects were ordered at the time of discharge to have between 10 and 22 daily medication administrations.
On the basis of information contained in the hospitalization record, 41% of all subjects were judged as being dependent on some form of technology, specifically 20% being dependent upon medical devices, 32% being dependent upon medications, and 11% being dependent upon both medical devices and medications (Figure 4). Among the subjects deemed to be dependent upon medications, the most prevalent clinical circumstances were children receiving either anti-epileptic drugs, long-term anti-inflammatory therapy, or anti-neoplastic and related medications, with other children dependent upon insulin, long-term analgesic medications, or anti-gastrointestinal reflux drugs. Fifty-eight percent of all subjects were deemed to meet the MCHB criteria of children having special health care needs.
Figure 4 Proportion of subjects with technology dependency or special health care needs. The span of each bar represents the proportion of the sample that exhibited the specified characteristic. The overlapping position of the bars vis-à-vis each other represents the proportion of subjects who exhibited several of these characterizes. Medications, Devices, and Both (medications and devices) indicate sub-types of technology-dependence.
TD patients were older than non-TD patients (median age category 10–14 years versus 1–4 years, respectively, p = 0.01), and were less likely to be covered by private insurance than non-TD patients (43.9% versus 61%, p = 0.06). TD patients were discharged with a greater number of prescriptions than patients without TD (mean 4.1 versus 1.3, p < 000.1, with the respective ranges being 0 to 15 versus 0 to 3). Overall, TD patients were not more likely to receive home services than non-TD patients (14.6% versus 8.5%, p = 0.3), but those patients dependent upon devices were more likely to receive home care than patients without device dependency (25% versus 7.5%, p = 0.03).
Discussion
This detailed review of 100 randomly selected patients admitted to a regional pediatric tertiary care center in the year 2000 suggests that a sizable proportion of all admitted patients (41%) are substantially dependent upon some form of medical technology in order to avoid deterioration in health that would necessitate hospital admission.
Our findings need to be interpreted in the context of this study's limitations. First, the degree to which the results can be generalized beyond the year 2000 for this single institution, serving as a tertiary referral center for a multi-state region in the Pacific Northwest, is unclear. An analysis of pediatric hospitalization time-trends in Washington State revealed that the care of children with complex medical conditions increasingly is concentrated in pediatric referral centers [24], suggesting that TD children may compose an enlarging proportion of discharges from children's hospitals. Second, the study's random sample of 100 distinct child discharges, while providing representative estimates of characteristics of all patients discharged from this hospital, is not as precise as a larger sample would be. For example, the small sample likely explains why certain technologies such as oxygen therapy, ventilator support, or renal dialysis were not observed. Third, the classification of patients into the categories of TD or CSHCN required the use of individual raters' judgment. While agreement among the 4 raters was moderate to good, we only classified patients into these categories if 3 of 4 raters agreed, thereby improving specificity but at the loss of sensitivity, so that our data may underestimate the proportions of both TD and CSHCN children. Finally, our study measures only a few attributes of care that can make care burdensome; the study also has no information regarding post-discharge occurrences, such as adherence to discharge instructions, the adequacy of follow-up ambulatory care, or unintended hospital readmissions.
These limitations notwithstanding, the prevalence of TD in this population of hospitalized children suggests that the phenomenon of technology-dependency warrants further study. As suggested by Figure 2, research efforts could seek a) to advance the performance of the technology, b) to assist families and care providers in making better decisions about when to adopt specific technologic solutions to health problems, and c) to improve how the encompassing system of health and supportive care maximizes the benefits of the technology while minimizing the associated risks and burdens. Accomplishing these aims through risk/benefit, cost/effectiveness, and cost/utility analyses will require more precise quantification of technologies' benefits, risks, costs, burdens, and broader societal consequences.
Three specific challenges hinder the pursuit of such a research agenda. First, no consensus definition of "technology-dependency" currently exists, either as a yes/no or as gradations of a technology-dependency classification scheme, perhaps based on the multidimensional construct proposed in Figure 2. The merits of any particular definition used to classify children as TD should be viewed in light of whether the definition can be used readily, reliably, accurately, and consistently for research purposes; and perhaps most important, whether the definition when applied enables the subsequent analysis to advance a value-focused research agenda that ultimately maximizes patient, family, and public well-being. Second, methods to assess the burden of care, quantitatively as well as qualitatively, need to be developed. Lacking such measures, evaluations of technology are likely to underestimate the indirect costs borne by families and care providers, and broader consequences to society. Third, better assessment of the value that children, parents, and others place on TD quality of life – and how these evaluations change over time with increased exposure to the technology and the quality of life that it supports – is essential not only for cost-utility analyses, but also to enable children or parents to make better informed decisions when presented with TD options of care.
Addressing these challenges would enable 1) documentation of technology-dependency incidence and changes in prevalence over time; 2) assessment of the impact of technology-dependency on patients' and families' physical, psychosocial, and financial well-being throughout an illness trajectory; 3) identification of health-care providers and agencies that care for TD children and the economic ramifications of providing such care; 4) evaluation of how TD patients and the care that they require affects the health-care system as well as schools, other social services, and parental employers; and 5) testing of specific interventions and management techniques that conceivably influence the burden of TD care. With the information that such studies would provide, we would be better able to perform individual- and population-level needs assessment and planning, to develop techniques to lessen burdens and increase the safety and efficacy of TD, and to create policies to promote high-quality care for these vulnerable patients and their families.
Conclusion
A substantial proportion of children discharged from a regional children's hospital were dependent upon technology. Further study is required to ascertain the quality of care received by TD children and their families and then potentially to improve their outcomes.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors participated in the design of the study and interpretation of the data; NLV and BM conducted the chart reviews; CF and VS performed the data analysis; CF drafted the manuscript; all authors revised the manuscript for key intellectual content. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This project was supported in part by the Children's Hospital and Regional Medical Center Funds for Excellence. Dr. Feudtner was supported by grant number K08 HS00002 from the Agency for Healthcare Research and Quality. We thank Kathy Fennell for her assistance in completing this study.
==== Refs
Baker JP The machine in the nursery : incubator technology and the origins of newborn intensive care 1996 Baltimore , Johns Hopkins University Press x, 247 p.
Stern A Markel H Formative years : children's health in the United States, 1880-2000 2002 Ann Arbor , University of Michigan Press xvi, 304 p.
Howell JD Technology in the hospital : transforming patient care in the early twentieth century 1995 Baltimore , Johns Hopkins University Press xv, 341 p.
Treaster JB Cost President cited for girl's treatment is described as 'low' New York Times 1981 New York A25
Headliners Home is better New York Times 1981 New York E9
Social Security Administration Supplemental Security Income for the Aged, Blind, and Disabled; Deeming of Income and Resources (20 CFR Part 416) 1982 47 Federal Registry 24274
U.S. Department of Health and Human Services Report of the Surgeon General's Workshop on Children with Handicaps and Their Families. Case Example: The Ventilator-Dependent Child 1982 U.S. Department of Health and Human Services
Chief Justice Burger Irving Independent School District v. Tatro 1984 468 U.S. Supreme Court of the United States 883
Lantos JD Kohrman AF Ethical aspects of pediatric home care Pediatrics 1992 89 920 924 1579404
Aday LA Wegener DH Andersen RM Aitken MJ Home care for ventilator-assisted children Health Aff (Millwood) 1989 8 137 147 2744692 10.1377/hlthaff.8.2.137
U.S. Congress Office of Technology Assessment Technology-Dependent Children: Hospital v. Home Care OTA-TM-H-38 1987 Washington, DC , U. S. Government Printing Office
Palfrey JS Walker DK Haynie M Singer JD Porter S Bushey B Cooperman P Technology's children: report of a statewide census of children dependent on medical supports Pediatrics 1991 87 611 618 2020504
Palfrey JS Haynie M Porter S Fenton T Cooperman-Vincent P Shaw D Johnson B Bierle T Walker DK Prevalence of medical technology assistance among children in Massachusetts in 1987 and 1990 Public Health Rep 1994 109 226 233 8153274
Glendinning C Kirk S Guiffrida A Lawton D Technology-dependent children in the community: definitions, numbers and costs Child Care Health Dev 2001 27 321 334 11437836 10.1046/j.1365-2214.2001.00187.x
Quint RD Chesterman E Crain LS Winkleby M Boyce WT Home care for ventilator-dependent children. Psychosocial impact on the family Am J Dis Child 1990 144 1238 1241 2239864
Patterson JM Leonard BJ Titus JC Home care for medically fragile children: impact on family health and well-being J Dev Behav Pediatr 1992 13 248 255 1506462
Petr CG Murdock B Chapin R Home care for children dependent on medical technology: the family perspective Soc Work Health Care 1995 21 5 22 8553191 10.1300/J010v21n01_02
Fleming J Challela M Eland J Hornick R Johnson P Martinson I Nativio D Nokes K Riddle I Steele N Sudela K Thomas R Turner Q Wheeler B Young A Impact on the family of children who are technology dependent and cared for in the home Pediatr Nurs 1994 20 379 388 7885750
Thyen U Terres NM Yazdgerdi SR Perrin JM Impact of long-term care of children assisted by technology on maternal health J Dev Behav Pediatr 1998 19 273 282 9717137
Thyen U Kuhlthau K Perrin JM Employment, child care, and mental health of mothers caring for children assisted by technology Pediatrics 1999 103 1235 1242 10353935 10.1542/peds.103.6.1235
Curran AL Sharples PM White C Knapp M Time costs of caring for children with severe disabilities compared with caring for children without disabilities Dev Med Child Neurol 2001 43 529 533 11508918
Roberts K Lawton D Acknowledging the extra care parents give their disabled children Child Care Health Dev 2001 27 307 319 11437835 10.1046/j.1365-2214.2001.00178.x
Roberts G Supporting children with serious health care needs. Analyzing the costs and benefits Eval Health Prof 2001 24 72 83 11233587 10.1177/01632780122034803
Neff JM Valentine J Trends in pediatric hospitalization by hospital type and chronic condition status: Washington State 1987-1998 Academy Health Annual Research Conference
| 15882452 | PMC1142327 | CC BY | 2021-01-04 16:31:06 | no | BMC Pediatr. 2005 May 9; 5:8 | utf-8 | BMC Pediatr | 2,005 | 10.1186/1471-2431-5-8 | oa_comm |
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-91588245010.1186/1471-2431-5-9Case ReportAtypical microbial infections of digestive tract may contribute to diarrhea in mucopolysaccharidosis patients: a MPS I case study Węgrzyn Grzegorz [email protected] Julianna [email protected] Anna [email protected]ńska Anna [email protected] Barbara [email protected] Ewa [email protected]óbkiewicz-Banecka Joanna [email protected]ęgrzyn Alicja [email protected] Department of Molecular Biology, University of Gdańsk, Kładki 24, 80-822 Gdańsk, Poland2 Department of Bacteriology, Provincial Hospital, Nowe Ogrody 1-6, 80-803 Gdańsk, Poland3 Department of Pediatrics, Children's Gastroenterology and Oncology, Medical University of Gdańsk, Nowe Ogrody 1-6, 80-803 Gdańsk, Poland4 Department of Metabolic Diseases, The Children's Memorial Health Institute, Al. Dzieci Polskich 20, 04-736 Warsaw, Poland5 Department of Genetics, Institute of Psychiatry and Neurology, Sobieskiego 9, 01-957 Warsaw, Poland6 Department of Genetics and Marine Biotechnology, Institute of Oceanology, Polish Academy of Sciences, Św. Wojciecha 5, 81-378 Gdynia, Poland7 Laboratory of Molecular Biology (affiliated with the University of Gdańsk), Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Kładki 24, 80-822 Gdańsk, Poland2005 9 5 2005 5 9 9 8 8 2004 9 5 2005 Copyright © 2005 Węgrzyn et al; licensee BioMed Central Ltd.2005Węgrzyn 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
Mucopolysaccharidoses are heritable, metabolic diseases caused by deficiency in an activity of one of specific lysosomal enzymes involved in degradation of mucoplysaccharides (glycosaminoglycans). Among many medical problems of patients with mucopolysaccharidoses, there are frequent episodes of diarrhea of unknown etiology.
Case presentation
A girl, diagnosed enzymatically for mucopolysaccharidosis type I (deficiency of α-L-iduronidase) at the age of 3 years and 9 months, was investigated until the age of 5 years and 4 months. Frequent loose stools and episodes of diarrhea, often accompanied by vomiting, were encountered. Detailed microbiological analyses were performed and atypical microbial infections (most often enetropathogenic Escherichia coli, but also other species, like Pseudomonas aeruginosa or Staphylococcus aureus, as well as adenoviruses) of the digestive tract were found in most severe diarrhea episodes. Often, isolations of pathogenic bacterial strains from stools of the investigated patient suffering from diarrhea were not obvious during the first screening, and only detailed microbiological studies, including re-isolation of colonies, gave the results of isolation of particular pathogenic strains (especially in the case of enetropathogenic E. coli).
Conclusion
We conclude that atypical microbial infections of digestive tract may contribute significantly to diarrhea in mucopolysaccaridosis patients. Since isolated strains were not typical and their isolation was often possible only after detailed investigation (not during a standard screening), such atypical microbial infections of digestive tract of mucopolysaccharidosis patients could be usually overlooked to date. Importantly, these atypical infections could be effectively treated with antimicrobial agents.
==== Body
Background
Mucopolysaccharidoses (MPS) are rare genetic diseases, inherited in an autosomal recessive manner (except MPS type II, which is X-linked) [1]. Each MPS is caused by deficiency in an activity of one of specific lysosomal enzymes involved in degradation of mucoplysaccharides (glycosaminoglycans, GAGs). When activity of such an enzyme is impaired, GAGs are not destroyed but their accumulate in lysosomes of virtually every cell of the body as well as outside of cells [1,2]. Depending on the kind of a deficient enzyme, several types of mucopolysaccaridosis can be distinguished [3].
Cells of MPS patients do not perform properly, which leads to progressive damage throughout the body (mainly connective tissue), including the heart, respiratory system, bones, joints and, in some cases, central nervous system [1,3]. In most cases the disease is fatal, however, prediction of its severity and clinical progress is usually difficult, even when biochemical and genetic data are available [4]. Among many medical problems, MPS patients suffer from frequent episodes of loose stools and diarrhea. Although mechanisms of other symptoms of MPS are known to some extend, the cause of diarrhea remains obscure [1,3]. One might speculate that it could be a defect in the autonomic nervous system caused by storage of glycosaminoglycans, and resultant abnormal motility of the bowel would cause the diarrhea. However, few evidences can support such a hypothesis. Therefore, we have investigated factor(s) that might contribute to this problem of MPS patients. Here, a case report is presented, in which a MPS I patient was investigated, however, it is worth mentioning that diarrhea is also a common problem for MPS patients of other types. MPS I is caused by a defect in the gene coding for α-L-iduronidase [5]. We found that episodes of diarrhea were accompanied by atypical microbial infections that could be treated by antimicrobial agents. Recently, a specific treatment of MPS I became available, since recombinant human α-L-iduronidase (Aldurazyme®, laronidase) has been shown to be effective [6,7]. During three months of treatment of the patient with Aldurazyme, no diarrhea episodes occurred, strongly suggesting that previous episodes results indirectly from MPS rather than from putative, MPS-independent, enhanced susceptibility of the patient to microbial infections of intestine.
Case presentation
A girl, diagnosed enzymatically for MPS I (McKusick's OMIM (MPS I, Hurler syndrome, Scheie syndrome): 25280) at the age of 3 years and 9 months, was investigated until the age of 5 years and 4 months. Genetic analysis (performed by us) indicated that one allele of the α-L-iduronidase gene contains a mutation causing a common Q70X truncation in the enzyme amino acid sequence. Further studies (performed at Consorzio Genetica Moleculare Umana, Monza, Italy, supervised by Prof. Andrea Biondi and coordinated by Genzyme Co.) confirmed the presence of this allele and indicated that the second mutation is an in frame deletion of three base pairs (codon 349), resulting in the loss of aspartic acid residue (Δ349). The investigated MPS I patient suffered, between other symptoms characteristic for this disease, from frequent diarrhea or loose stools, often accompanied by vomiting. No obvious contacts with persons suffering from intestinal infections, including family members (who also had no documented chronic or often diarrhea), were noticed shortly before or during these episodes. During the most severe episodes of diarrhea, detailed microbiological analyses of stools were performed. The patient was never treated with antibiotics before stool collection, thus, results of microbiological assays were not affected by such a treatment. Bacterial, viral or fungal infections were detected in all these analyses (Table 1).
Table 1 History of the most severe diarrhea events caused by microbial infections of intestine of the MPS I patient
Age of the patienta Detected pathogenic microbial strain(s) Antimicrobial agent(s) used Special remarks
11 months E. coli (group C) Trimethoprim and sulphamethoxazole Severe diarrhea and vomiting, metabolic acidosis, hospitalization and i.v. rehydration required
21 months (1 y. 9 mo.) E. coli (group C) Cefuroxime Initially applied trimethoprim and sulphamethoxazole did not give a therapeutic effect
27 months (2 y. 3 mo.) E. coli (group A) Nifuroxazide None
28 months (2 y. 4 mo.) P. aeruginosa None Tablets containing Lactobacillus rhamnosus were used to normalize intestinal flora
29 months (2 y. 5 mo.) E. coli (group C) Furazidin None
45 months (3 y. 9 mo.) Adenoviruses, C. albicans (in addition, H. parainfluenzae, S. viridans and Neiserria sp. were isolated from oral cavity) Cefuroxime, fluconazole, trimethoprim and sulphamethoxazole Very severe diarrhea and vomiting, severe dehydration and electrolyte depletion, metabolic acidosis, severe stomatitis. Hospitalization and parenteral nutrition required
49 months (4 y. 1 mo.) Adenoviruses None None
50 months (4 y. 2 mo.) E. coli (group C) Furazidin, trimethoprim and sulphamethoxazole Severe diarrhea, hospitalization and i.v. rehydration required
54 months (4 y. 6 mo.) Adenoviruses None None
55 months (4 y. 7 mo.) E. coli (group C) Furazidin None
56 months (4 y. 8 mo.) E. coli (group C), S. aureus, C. albicans Furazidin, fluconazole None
a Age of the patient is provided in months as well as in years (y.) and months (mo.) in parentheses
Isolation and preliminary analyses for identification and characterization of bacteria, viruses and fungi were performed using S-F selenium, blood agar, CNA, McConkey, SS, XLD, Sabouraud, Kligler and Singer, and BHI media (purchased from Becton Dickinson, Biomed or Bimex). Detailed analyses were performed using biochemical assay kits ID GN 32 (BioMerieux), EN-COCCUStest (Lachema) and ATB ID 32 C (BioMerieux), and immunological tests (kits purchased from Immunomed and BioMarieux). The microbiological assays were performed according to the semi-quantitative method, described previously [8]. Non-enteropathogenic strains of Escherichia coli, strains of Enterococcus faecalis and small amounts of other bacteria belonging to Enterobacteriaceae were considered as normal aerobic bacterial flora. Identification of enteropathogenic E. coli strains was performed using a specific latex-immunological assay (purchased from Biomex). In this assay, 10 randomly chosen E. coli colonies (classified on the basis of microbiological and biochemical tests described above) were streak onto nutrient agar plates, and following 18 h incubation at 35°C bacteria were tested immunologically using the Biomex kit. The O antigens (characteristic for enteropathogenic E. coli strains but not for other strains of this bacterium), were detectable in this assay, and corresponding strains, bearing particular antigens, were divided into following groups: group A (antigens O26, O55, O111, O127 and O142), group B (antigens O86, O119, O124, O125, O126 and O128) and group C (antigens O18, O25, O44 and O114).
As summarized in Table 1, all viral infections were caused by adenoviruses, and tests were negative after diarrhea was finished. Most bacterial infections were atypical. Enteropathogenic Escherichia coli belonging to group A or, more frequently, to group C, was detected in over 50% of positive tests. These bacteria may cause diarrhea in infants, but they are very rare agents causing intestinal problems in older children and adults, thus the tests for detection of these strains are usually not performed in routine studies if the patient is not an infant [9,10]. Pseudomonas aeruginosa, Staphylococcus aureus and Candida albicans were also detected, though significantly less frequently than E. coli group C. These species can be detected also in stools of healthy persons. Thus, in our assays, they were considered potential etiological agents only when they dominated quantitatively over other microbes treated as physiological flora, and the dominance was significant. In fact, during the episode at the age of 28 months (Table 1), P. aeruginosa was the only bacterium isolated from the stool sample. Although no antibiotic therapy was applied in this case (due to problems with efficacy of treatment of P. aeruginosa infections with antibiotics), the symptoms of diarrhea disappeared and normalization of the intestinal flora was achieved after using tablets containing Lactobacillus rhamnosus. This strongly suggests a significant contribution of the P. aeruginosa infection to diarrhea. At the age of 45 months, when the patient suffered from a very severe diarrhea and vomiting (Table 1), apart from adenoviruses, C. albicans was detected in large quantities in stool samples, and large amounts of Haemophilus parainfluenzae, Streptococcus viridans and Neiserria sp. were isolated from oral cavity. Although S. viridans and Neiserria are considered as a part of normal bacterial flora of oral cavity in infants, their abundance, together with that of H. parainfluenzae was accompanied with significant depletion of the rest of physiological flora. During the episode of diarrhea at the age of 56 months (Table 1), apart from isolation of enteropathogenic E. coli, severe depletion of physiological flora and presence of large quantities of S. aureus and C. albicans was found. The enteropathogenic E. coli strain might be the major bacterial contributor to this infection as treatment with furazidin (which is generally not used for treatment of S. aureus infections) and fluconazole (an antifungal agent) was effective.
It is worth mentioning that isolations of pathogenic bacterial strains from stools of the investigated patient suffering from diarrhea were often not obvious during the first screening, especially in the case of enteropathogenic E. coli. Only detailed microbiological studies, including re-isolation of colonies, gave the results of isolation of particular pathogenic strains (summarized in Table 1). We suppose that this could be the explanation of the fact that atypical microbial infections of digestive tract of MPS patients were usually overlooked to date. On the other hand, these infections apparently caused diarrhea. In most episodes accompanied by isolation of bacterial infection, drug sensitivities were determined and antibiotics were used accordingly (Table 1). After this treatment, diarrhea was resolved and pathogenic bacteria were not detected in stools. Similar results were obtained after therapy with an antifungal drug, when C. albicans was detected (data not shown).
Intestinal flora was normal when tested in periods of normal stools (microbiological tests were performed 5 times during such periods as control experiments; data not shown). This provides an additional support for the hypothesis that isolated pathogenic microbial strains caused diarrhea, rather than the patient was a carrier of an atypical bacterial flora.
The patient was treated with Aldurazyme (laronidase; recombinant human α-L-iduronidase), by weekly infusions of 100 U per 1 kg of body weight, during last three months of the study, and no episodes of diarrhea occurred during that period. This indicates that episodes of severe diarrhea, listed in Table 1, resulted directly or indirectly from accumulation of GAGs, rather than from a putative, unrelated to MPS I, enhanced susceptibility of the patient to microbial infections of intestine.
Conclusion
On the basis of the results presented in this report, we suggest that microbial infections of intestine may contribute significantly to episodes of diarrhea in MPS patients. Is it clear that influence of other factors and agents, for example those resulting from dietary incompliance, cannot be excluded. Nevertheless, it seems clear that the most severe diarrhea events in the MPS I patient were caused by microbial infections, as when pathogenic bacteria or fungi were isolated, antibiotic treatment always resulted in normalization of stools. One may speculate that GAGs, which accumulate in MPS patients' digestive tract, could provide a favorable medium for microbial growth. Under such conditions, any disturbance in the content of intestinal flora would make an overgrowth of a pathogenic strain significantly more likely than in a healthy person. Finally, accumulated GAGs might also impair production and/or action of IgA, a class of antibodies predominating in an intestine [11,12], which would make a patient more sensitive to microbial infections. One might consider that in MPS patients, diarrhea arises directly from GAG accumulation and partial dysfunction of autonomous nervous system. Although we cannot exclude such a possibility, the arguments listed above indicate that atypical microbial infections of digestive tract may contribute significantly to diarrhea in these patients.
Abbreviations
MPS, mucopolysaccharidosis; GAG, glycosaminoglycan. y., years; mo., months
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GW was the main coordinator of this study and drafted the manuscript. JK performed all microbiological analyses. AL and AT-S participated in the design of the study and performed clinical observations. BC performed biochemical analyses and participated in coordination of the study. EP and JJ-B performed genetic analysis. AW participated in the design of the study and contributed to preparation of the manuscript draft. 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 Polish Ministry of Science and Information Technology (grant no. 2 P05A 103 26). Written consent was obtained from parents of the patient for publication of this study.
==== Refs
Neufeld EF Muenzer J Scriver CR, Beaudet AL, Sly WS, Valle D The mucopolysaccharidoses The metabolic and molecular bases of inherited disease 2001 New York: McGraw-Hill Co 3421 3452
Hinek A Wilson SE Impaired elastogenesis in Hurler disease: dermatan sulfate accumulation linked to deficiency in elastin-binding protein and elastic fiber assembly Am J Pathol 2000 156 925 938 10702409
Hopwood JJ Morris CP The mucopolysaccharidoses: diagnosis, molecular genetics and treatment Mol Biol Med 1990 7 381 404 2128891
Węgrzyn G Węgrzyn A Tylki-Szymańska A A general model for genetic regulation of turnover of glycosaminoglycans suggests a possible procedure for prediction of severity and clinical progress of mucopolysaccharidoses Med Hypotheses 2004 62 986 992 15142662 10.1016/j.mehy.2003.12.027
Roubicek M Gehler J Spranger J The clinical spectrum of α-L-iduronidase deficiency Am J Med Genet 1985 20 471 481 3922223
Kakkis ED Muenzer J Tiller GE Waber L Belmont J Passage M Izykowski B Philips J Doroshow R Walot I Hoft R Neufeld EF Enzyme-replacement therapy in mucopolysaccharidosis I New Eng J Med 2001 344 182 188 11172140 10.1056/NEJM200101183440304
Wraith JE Clarke LA Beck M Kolodny EH Pastores GM Muenzer J Rapoport DM Berger KI Swiedler SJ Kakkis ED Braakman T Chadbourne E Walton-Bowen K Cox GF Enzyme replacement therapy for mucopolysaccharidosis I: a randomized, double-blinded, placebo-controlled, multinational study of recombinant human α-L-iduronidase (Laronidase) J Pediatr 2004 144 581 588 15126990 10.1016/j.jpeds.2004.01.046
Isenberg HD Clinical microbiology procedures handbook 1992 Washington DC: American Society for Microbiology
Cravioto A Reyes RE Trujillo F Uribe F Navarro A De La Roca JM Hernandez JM Perez G Vazquez V Risk of diarrhea during the first year of life associated with initial and subsequent colonization by specific enteropathogens Am J Epidemiol 1990 131 886 904 2157338
Baldwin TJ Pathogenicity of enteropathogenic Escherichia coli J Med Microbiol 1998 47 283 293 9568993
Farstad IN Carlsen H Morton HC Brandtzaeg P Immunoglobulin A cell distribution in the human small intestine: phenotypic and functional characteristics Immunology 2000 101 354 363 11106939 10.1046/j.1365-2567.2000.00118.x
Holtmeier W Hennenann A Casprary WF IgA and IgM V(H) repertoires in human colon: evidence for clonally expanded B cells that are widely disseminated Gastroenterology 2000 119 1253 1266 11054383
| 15882450 | PMC1142328 | CC BY | 2021-01-04 16:31:06 | no | BMC Pediatr. 2005 May 9; 5:9 | utf-8 | BMC Pediatr | 2,005 | 10.1186/1471-2431-5-9 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-451588246710.1186/1471-2458-5-45Research ArticleOverweight and obesity at school entry among migrant and German children: a cross-sectional study Will Beata [email protected] Hajo [email protected] Bernhard T [email protected] Dep. of Epidemiology & International Public Health, University of Bielefeld, D-33501 Bielefeld, Germany2 Dep. of Public Health Medicine, University of Bielefeld, D-33501 Bielefeld. Germany3 Department of Psychiatry and Psychotherapy, University of Muenster, D-48149 Muenster, Germany2005 9 5 2005 5 45 45 17 11 2004 9 5 2005 Copyright © 2005 Will et al; licensee BioMed Central Ltd.2005Will 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
Overweight and obesity have become a global epidemic and are increasing rapidly in both childhood and adolescence. Obesity is linked both to socioeconomic status and to ethnicity among adults. It is unclear whether similar associations exist in childhood. The aim of the present study was to assess differences in overweight and obesity in migrant and German children at school entry.
Methods
The body mass index (BMI) was calculated for 525 children attending the 2002 compulsory pre-school medical examinations in 12 schools in Bielefeld, Germany. We applied international BMI cut off points for overweight and obesity by sex and age. The migration status of children was based on sociodemographic data obtained from parents who were interviewed separately.
Results
The overall prevalence of overweight in children aged 6–7 was 11.9% (overweight incl. obesity), the obesity prevalence was 2.5%. The prevalence of overweight and obesity was higher for migrant children (14.7% and 3.1%) than for German children (9.1% and 1.9%). When stratified by parental social status, migrant children had a significantly higher prevalence of overweight than German children in the highest social class. (27.6% vs. 10.0%, p = 0.032) Regression models including country/region and socioeconomic status as independent variables indicated similar results. The patterns of overweight among migrant children differed only slightly depending on duration of stay of their family in Germany.
Conclusion
Our data indicate that children from ethnic minorities in Germany are more frequently overweight or obese than German children. Social class as well as family duration of stay after immigration influence the pattern of overweight and obesity in children at school entry.
==== Body
Background
Overweight and obesity are rapidly increasing among children and adults. In 1998, the World Health Organisation recognized obesity as a major worldwide public health epidemic [1]. The most recent World Health conference at Geneva called for specific action to halt this epidemic [2].
Overweight and obesity can lead to a wide array of health and social consequences. Childhood overweight appears to be associated with cardiovascular risk factors such a high blood pressure, hyperlipidemia, elevated insulin levels [3] and non-insulin-dependent diabetes mellitus (type 2 diabetes). Other comorbidities include asthma and orthopedic problems as well as a varity of more rare disorders [3,4]. Ultimately overweight in childhood is associated with premature mortality especially if combined with intrauterine growth retardation (IUGR). The psychological well-being and the quality of life can also be affected [3,5,6].
Furthermore overweight and obesity in childhood, particulary in adolescence tends to persist into adulthood. The risk of adult overweight is about twofold increased for individuals who were overweight as children compared with individuals who were not overweight [7,8]. Adulthood overweight are at increased risk of dyslipidaemia, hypertension and type 2 diabetes even if the extra weight was lost during adulthood [6]. Psychological consequences include social and psychological stress, with increased risk of negative self- esteem, social isolation and negative influence on the career and family incomes [7].
There are wide geographical variations of overweight. Comparing reported prevalences of childhood overweight in Europe, Lobstein et al. [9] point out that children residing in central and Eastern Europe have a lower level of overweight than children from other parts of Europe, especially from Southern Europe. Average prevalences in Eastern Europe range from 10% to 18% among children aged 7 to 11 years, whereas values around 20% to 35 % have been reported from countries like Greece and Spain [9,10]. It is suggested that the low obesity prevalence in Eastern Europe is a consequence of the huge economic burden and the associated poverty following the political transition in the 1990s.
Studies in industrialized countries show that children from families with lower socio-economic status particularly suffer from excess weight [11,12]. In less industrialized societies excess weight is found predominantly among children from families with higher socio-economic status [9].
The process of migration, with its economic, social and environmental consequences may also affect health and body weight among migrants. Several US studies show differences in body mass index (BMI) among ethnic minorities [13,14] but empirical data on BMI in children among different ethnic groups in Europe are scarce. In Germany children from non-German ethnic groups nowadays make up a large percentage of children overall. In some cities up to 50% of all children entering school may be first, second or third generation migrants [15]. There are several studies or health reports on overweight and obesity among German and migrant children at school entry [16-19]. Migrant children were consistently found to be more frequently overweight as compared to German children. In Berlin, overweight and obesity was noted as a particular problem among children of Turkish origin [20]. However, since the assessment of migration status is not uniform, comparisons between studies are difficult. In addition, family socioeconomic status and duration of migration are not routinely evaluated. We therefore investigated the prevalence of overweight and obesity among migrant and German children at school entry and to assess the influence of duration of migration and socioeconomic status on overweight and obesity in childhood. It should be noted that the primary aim of the study was to assess differences in health status and health service usage among Germants and Migrants and the current analysis therefore has exploratory character.
Methods
In 2002, a cross-sectional study was conducted at 12 primary schools in Bielefeld, a town of some 320.000 inhabitants in North-West Germany. The study population comprised of all children registered at these schools for school entry in 2002. Through the city public health office these children and their parents were invited to attend the compulsory medical examination taking place at the local school. The 12 schools (out of a total of some fifty primary schools in the town) were pragmatically selected based on information available for the preceding year. The aim was to include schools with a relatively balanced mix of migrant and German children and fair variation of social status among the population in the respective school districts.
Social and demographic information relevant to the children were derived from a 36 item-questionnaire filled out by parent or relative who accompanied the children [21]. When two adults accompanied the child we collected interview data from only one parent or relative, usually the mother. The questionnaire was available in 4 languages (German, Turkish, Russian, Polish). Parents or relatives were interviewed on the day of the school health examination with the assistance of trained interviewers.
The medical examination of the children was carried out by medical staff from the public health office in Bielefeld using a standardized model of medical school health examination [22]. During the physical and psychological examination, height and weight of children were recorded. All anthropometric measurements of the children were performed in the morning. Body weight was determined to the nearest 500 g using electronic or mechanic non -calibrated scales. During the measurements the children were dressed in light indoor clothes without shoes. No adjustments were made for clothing. Height was measured to the nearest 10 mm using a mobile scale.
657 children aged 6 and 7 appeared in the schools for pre-school examination. Interview data were collected from 565 adult companions (response rate of 86%). The sample size was further reduced to 537 data sets due to merging errors and missing medical data for some children. We also excluded all 12 children who had non-parental companions at the examination.
The statistical analyses were performed on merged data from 525 children (274 boys and 251 girls) and their parents, corresponding to an overall response rate of 79.9% (525/657). 267 (50.9%) parents were of German descent and 258 (49.1%) of Non-German descent. There were 240 females and 27 males among the German parents and 191 females and 67 males among the Non-German parents. Migrant status of the children was based on the migrant status of the interviewed parent. Parents who themselves or whose parents were born in Germany were classified as German, all others were defined as migrants.
The study population consisted of three main migrant groups: 36% of migrants were of Turkish origin, 25% originated from Russia and 15% from Poland. A total of 26 different countries of origin was noted.
A social class index consisting of three social classes (low, medium, high) was constructed using four indicators (primary qualification, professional education, employment status and size of household). We divided these variables into 3 levels (variable professional education in 2 levels) and scored 1–3 points. The sum of the points was than divided in 3 social classes (high social class 10–11 points, medium social class 7–9 points, low social class 4–6 points) [23].
BMI was calculated as body weight divided by body height squared (kg/m2). Overweight and obesity were defined according to the recommendations of the International Obesity Task Force IOTF, using international reference values based on data from six countries [24]. These age and sex specific BMI cut-off points for overweight and obesity in children (between 2 and 18 years) were constructed using dataset specific centiles corresponding to the widely accepted adult cut-points of a BMI of 25 kg/m2 (overweight) and 30 kg/m2 (obesity).
For the calculation of overweight and obesity the exact age of the children was rounded to the nearest half year. Children with BMI at or above cut-off value corresponding to a BMI of 25 kg/m2 in late adolescence (e.g. BMI of 17.55/17.34 for 6 year old boys/girls) were classified as overweight (or obese when the respective cut-points were reached). For comparative purposes, we also used national BMI reference curves for Germany [25] in some instances. Here, the age-and sex-specific 90th and 97th BMI percentile are used as cut-point for overweight and obesity, respectively.
Statistical analysis
The analysis was performed on the data set containing merged data from both the parental survey and childrens' medical examination.
We used descriptive statistical methods to compare obesity and overweight between migrant and German children. Data were stratified according to migrant status, social status and duration of stay in Germany. Differences in proportions were tested by Fisher's exact or Chi-square test, trends in the prevalence of overweight were analysed with chi-square trend test.
Logistic regression models were used to simultaneously analyse the association between overweight and gender, social status, ethnic background and duration of stay in Germany among migrants. Adjusted odds ratios and 95% confidence intervals were calculated. The statistical analysis was performed with the software package SPSS (version 11.5; SPSS Inc, Chicago, IL).
Results
In 2002 the overall point prevalence of overweight (including obesity) in examined children starting primary school was 11.9%. 2.5% of children were obese according to the criteria mentioned above. More girls than boys were overweight (15.3% vs. 8.8%; p = 0.03) and obese (4.0% vs. 1.1%; p = 0.046). Comparing German and children from ethnic minorities the prevalence of overweight as well as of obesity was higher in migrant children in both genders. Table 1 summarizes the results.
Table 1 Prevalence of overweight and obesity# of children by gender and migrant status
Migrants N = 258 Germans N = 265 p – value*
overweight obesity overweight obesity overweight obesity
All n/nTotal
%(95% CI) 38/258
14.7 (10.4–19.1) 8/258
3.1 (1.0–5.2) 24/265
9.1 (5.6–12.5) 5/265
1.9 (0.2–3.5) 0.058 0.413
Boys n/nTotal
%(95% CI) 13/129
10.1 (4.9–15.3) 2/129
1.6 (-0.6-3.7) 11/145
7.6 (3.3–11.9) 1/145
0.7 (-0.7-2.0) 0.524 0.603
Girls n/nTotal
%(95% CI) 25/129
19.4 (12.6–26.2) 6/129
4.7 (1.0–8.3) 13/120
10.8 (5.3–16.4) 4/120
3.3 (0.1–6.5) 0.077 0.751
*Fisher's exact test comparing migrants and Germans
#Variable BMI contains 2 missing values
When applying national German BMI references [25], slightly lower prevalence estimates were calculated. Still, the point prevalence of overweight in migrant children was significantly higher compared to German children (12.0% vs. 5.7%; p = 0.01) while differences in obesity prevalence were not significant (3.5% vs. 2.3%). When stratified by gender, only migrant girls were significantly more overweight than their German peers (p = 0.04). Children from families with low socioeconomic status (SES) were significantly more often overweight (18.9%) than children with average (8.9%) or high SES (14.7%) (p = 0.02). Among Germans the prevalence was highest among children from low SES families (23.5%). Conversely the prevalence of overweight was highest among migrant children with high SES (27.6%) (Table 2).
Table 2 Overweight by migrant status and social class
Migrants N = 240 Germans N = 256 p – value*
Social class# Prevalence of overweight
Low n/nTotal
%(95% CI) 14/78
17.9 (9.4–26.5) 4/17
23.5 (3.4–43.7) 0.733
Medium n/nTotal
%(95% CI) 14/133
10.5 (5.3–15.7) 12/159
7.5 (3.4–11.7) 0.414
High n/nTotal
%(95% CI) 8/29
27.6 (11.3–43.9) 8/80
10.0 (3.4–16.6) 0.032
*Fisher's exact test comparing migrants and Germans
#social class index using primary educational qualification, professional education, employment status and size of household (27 missing values). Variable BMI contains 2 missing values
On average, children from migrant families had a lower socio-economic status when compared to German children (p < 0.001). Further analyses of socioeconomic status among migrants from different countries and regions of origin indicated that on average Turkish children were of lower social status than children from Eastern European families (Data not shown).
Nevertheless, there were no differences in overweight prevalence when comparing children according to country/region of origin (Turkey, Eastern Europe/Russia, others) (Table 3).
Table 3 Overweight by ethnic origin of family
Ethnic origin
Turkey Eastern Europe/Russia Other All
Prevalence of overweight n/N
%(95% CI) 13/89
14,6 (7,3–21,9) 19/126
15,1 (8,8–21,3) 6/43
14,0 (3,6–24,3) 38/258
14,7 (10,4–19,1)
In the bivariate analysis we found slight differences in the patterns of overweight among migrants depending on duration of stay in Germany. The results indicated that the prevalence of overweight among migrant children was highest in the category of highest family duration of stay in Germany (20.9% versus 12.9% in those staying 1–9 years). However, the trend across categories was not significant (p = 0.32).
In the multivariate analysis (Table 4) long duration of stay of the family in Germany (30+ years) was associated with overweight in migrant children (OR = 1.92; 95% CI = 0.97–3.79). Migrant girls had an increased risk of overweight as compared to boys. Odds ratios for migrant children with low and high SES were similar, whereas a medium SES was associated with a decreased risk of overweight in the study. In a second model including only socioeconomic status and country/region of origin, results for SES remained unchanged. None of the odds ratios for specific country/region of origin adjusted for SES were significantly different from 1, using Germany as reference.
Table 4 Results of the logistic regression model of risk factors for overweight in migrant children
OR 95%CI p-value
Boys* 1
Girls 2.19 0.98–4.92 0.06
social class
Low * 1
Medium 0.57 0.33–0.97 0.04
High 1.63 0.83–3.22 0.16
origin
Turkey* 1
Eastern Europe 1.15 0.65–2.04 0.63
other 0.92 0.45–1.87 0.81
Duration of stay (years)
1–9 * 1
10–19 0.89 0.48–1.67 0.74
20–29 0.95 0.47–1.92 0.88
30+ 1.92 0.97–3.79 0.06
*Reference category
Discussion
The present cross-sectional study shows that the prevalence of overweight and obesity was slightly higher among migrant children than among German children. Socioeconomic status seems to be associated with overweight and obesity in migrant children at school entry whereas duration of stay after immigration to Germany shows no clear pattern of association. Other studies in Germany have reported more marked differences between German and migrant children with regard to overweight and obesity. Among boys in Bavaria, Kalies [26] found that the frequency of overweight and obesity in non-German compared to German children was 1.9 times higher for overweight and 2.4 times higher for obesity. Similar findings have been reported from Berlin [20] and the federal state of Lower Saxony [27]. For Northrhine-Westfalia where the city of Bielefeld is located, only basic data are available showing similar patterns but an overall much lower overweight prevalence of 7% or less in all groups [28] while older data from Dortmund, a larger but structurally comparable town, show an overweight prevalence of about 12% among second generation migrant children [19]. Results of our study are consistent with these data, although the smaller sample size affects discriminatory statistical power. The choice of reference and the definition of migrant status (see below) may also contribute to the slightly different results.
In Europe, only a few studies on Body Mass Index (BMI) examined the potential impact of ethnicity or migration status on overweight and obesity. A recent UK survey [29] showed that Indian and Pakistani boys had a higher prevalence of overweight compared with boys in the general UK population while Bangladeshi and Chinese boys had a lower overweight prevalence. Among girls, Afro-Caribbean and Pakistani girls more frequently were overweight while Indian and Chinese girls had a lower overweight prevalence compared to girls from the general population. In France, the children of Maghrebian immigrants were more obese than French children in cross-sectional surveys conducted in the 1970ies and the 1990ies [30]. The overall prevalence of obesity increased from 8% to 13% over this period. Data from health surveys in 1992/93 and 1993/94 among children in the Netherlands showed that the average body mass index was higher among Turkish and Moroccan children than among Dutch children [31].
Geographical differences in overweight in Europe were demonstrated by Lobstein et al. [9]. Children from countries in Central and East Europe generally showed a low prevalence of overweight. On the other hand, a higher prevalence of overweight was found in children in Southern European countries. Recent data from Edirne in Turkey demonstrate comparatively low prevalences of overweight and obesity among adolescents [32].
The international variability of definitions concerning paediatric overweight and obesity implies methodological difficulties when comparing prevalences internationally. In adults, the BMI cut-points of 25 and 30 are widely used to define overweight and obesity. Unlike in adults, BMI in children varies substantially by age and gender during childhood and adolescence. Unfortunately no commonly accepted standard has yet emerged. Different reference systems have been proposed and these references vary considerably. International sex – and age -specific cut-points proposed by the IOTF [24] were used in the present study in order to allow national as well as international comparisons. The IOTF reference has attracted some criticsm, mainly based on the argument that data from only six countries form the basis of the international reference, thus putting representativity on a world scale in doubt [33]. It appears disputable whether the use of national reference data is preferable when assessing overweight and obesity among migrant children. The comparison to reference curves from the country of origin would disregard the environmental (e.g. nutritional) changes associated with migration in many instances. On the other hand, using a German national reference is also not fully satisfactory since differences in lifestyle or genetically determined factors related to body composition might persist for generations. An international standard appeared the most suitable option in this situation. Undoubtedly, the arbitrary choice of cut-off points and other problems associated with BMI-based classification systems in childhood will require further debate [34-36].
Our results do not clearly show an impact of family duration of stay after immigration (prior to 1980) into Germany on overweight. However, a long duration was associated with relatively higher prevalences of overweight and obesity. Adaptation and assimilation as well as (modestly) growing wealth of migrants in Germany after decades of life in Germany could partially explain these findings, in particular since the overweight was highest among well-off children. An association of overweight among migrant children with a long duration of stay in Germany has not been reported previously. A period effect due to social and political circumstances as a potential explanation needs to be considered. The possible impact of such changes, however in opposite direction, can be more clearly seen in data from Russia. The prevalence of overweight in Russia decreased from 26.4% in 1992 to 10.2% in 1998 in 6–9 y-old children [35] and the prevalence of underweight in children rose during the same time [9]. Economical recession during this time is a plausible explanation of this finding. In Poland, Koziel et al. found a slightly decreasing trend in overweight between 1987 and 1997 in 14y-old boys [37].
In general, studies on socioeconomic status and overweight suggest that overweight is more prevalent among children of low income families in developed countries. In contrast, the vast majority of studies in developing countries show a positive relationship between socioeconomic status and obesity among children [9,38]. Our study, carried out in a developed country, generally shows consistent results since overweight was more frequent among lower class children. However, among children from families with a non-German origin, the highest prevalence of overweight was observed in children in the high social class. It should be noted, however, that misclassification of socioeconomic status cannot be ruled out in our study since the classification was based on data supplied by one parent, mostly mothers, only.
Further limitations of our study need to be taken into consideration. We did not collect a representative sample but had to use a more pragmatic approach in selecting the study group. However, the schools included in the study represent a fair cross-section of different primary schools in Bielefeld. The assessment of body weight and height was subject to some variation since identical scales were not available at all school health examinations. All examinations were done by the same three teams of school health personnel. Finally the migrant status of children was based on rather differentiated parental data, different from the usual, but unsatisfactory approach of using nationality as indicator. We favour this more complex approach, but realise that comparisons with other studies may be difficult.
Conclusion
Our study results underline the importance of overweight and obesity as an important public health issue in young children and highlight patterns of overweight and obesity among native and migrant children in Germany. New studies such as the currently ongoing youth health survey [39] will help to clarify the association of overweight and obesity in children with social class and the duration of residence in the host country after migration. Public health interventions aiming at a change in nutrition and behaviour (sports, healthy life style) are likely to benefit from more detailed knowledge about specific risk groups.
List of abbreviations
BMI: Body mass index
OR: Odds ratio
CI: Confidence interval
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HZ and BB planned and co-ordinated the study. BW participated in field work, managed the data and performed the statistical analysis. All authors jointly wrote the paper.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank the interviewers, all the schools and the Public Health Office in Bielefeld for their input and cooperation. We are particularly grateful to all parents and children who took part in the study.
==== Refs
World Health Organization Obesity: Preventing and managing the global epidemic. WHO technical report series 2000 894 Geneva, WHO
World Health Organization: Fifty-seventh world health assembly. Global strategy on diet, physical activity and health. Geneva, 17-22 May 2004
Deckelbaum RJ Williams CL Childhood obesity: the health issue Obes Res 2001 9 Suppl 4 239S 243S 11707548
Reilly JJ Methven E McDowell ZC Hacking B Alexander D Stewart L Kelnar CJ Health consequences of obesity Arch Dis Child 2003 88 748 752 12937090 10.1136/adc.88.9.748
Must A Anderson SE Effects of obesity on morbidity in children and adolescents Nutr Clin Care 2003 6 4 12 12841425
Dietz WH Health consequences of obesity in youth: childhood predictors of adult disease Pediatrics 1998 101 518 525 12224658
Dietz WH Childhood weight affects adult morbidity and mortality J Nutr 1998 128 411S 414S 9478038
Must A Morbidity and mortality associated with elevated body weight in children and adolescents Am J Clin Nutr 1996 63 445S 447S 8615339
Lobstein T Frelut ML Prevalence of overweight among children in Europe Obes Rev 2003 4 195 200 14649370 10.1046/j.1467-789X.2003.00116.x
Krassas GE Tzotzas T Tsametis C Konstantinidis T Prevalence and trends in overweight and obesity among children and adolescents in Thessaloniki, Greece J Pediatr Endocrinol Metab 2001 14 Suppl 5 1319 1326 11964029
Kromeyer-Hauschild K Zellner K Jaeger U Hoyer H Prevalence of overweight and obesity among school children in Jena (Germany) Int J Obes Relat Metab Disord 1999 23 1143 1150 10578204 10.1038/sj.ijo.0801044
Langnäse K Mast M Muller MJ Social class differences in overweight of prepubertal children in northwest Germany Int J Obes Relat Metab Disord 2002 26 566 572 12075585 10.1038/sj.ijo.0801956
Haas JS Lee LB Kaplan CP Sonneborn D Phillips KA Liang SY The association of race, socioeconomic status, and health insurance status with the prevalence of overweight among children and adolescents Am J Public Health 2003 93 2105 2110 14652342
Nelson JA Chiasson MA Ford V Childhood overweight in a New York City WIC population Am J Public Health 2004 94 458 462 14998814
Ministerium für Schule Jugend und Kinder des Landes Nordrhein-Westfalen Ausländische und ausgesiedelte Schülerinnen und Schüler, Ausländische Lehrerinnen und Lehrer im Schuljahr 2003/04 Statistische Übersicht 2004 344
Gawrich S Wie gesund sind unsere Schulanfänger?- Zur Interpretation epidemiologischer Auswertungen der Schuleingangsuntersuchung. Hessisches Ärzteblatt 2004 2 73 76
Erb J Winkler G Role of nationality in overweight and obesity in preschool children in Germany Monatsschrift Kinderheilkunde 2004 152 291 298 10.1007/s00112-003-0738-6
Bauer C Rosemeier A A handicap for life - Overweight and obesity in pre-school children in Karlsruhe Gesundheitswesen 2004 66 246 250 15100941 10.1055/s-2004-812979
Gesundheitsamt der Stadt Dortmund Die Gesundheit der Schulanfängerinnen und Schulanfänger in Dortmund- Ergebnisse der schulärztlichen Untersuchungen von 1985-1996. Dortmunder Berichte 1997 5
Delekat D Zur Gesundheitlichen Lage von Kindern in Berlin- Ergebnisse und Handlungsempfehlungen auf Basis der Einschulungsuntersuchungen 2001.Spezialbericht 2003-2 Gesundheitsberichterstattung Berlin 2003 Berlin, Senatsverwaltung für Gesundheit, Soziales und Verbraucherschutz
Zeeb H Baune BT Vollmer W Cremer D Kramer A Health situation of and health service provided for adult migrants - a survey conducted during school admittance examinations Gesundheitswesen 2004 66 76 84 14994205 10.1055/s-2004-812825
Schirm H Untersuchungen und Untersuchungsmodelle zur Feststellung der Einschulungsfähigkeit Sozial Pädgogik 1989 7 11
Ahrens W Bellach BM Jöckel KH Messung soziodemographischer Merkmale in der Epidemiologie 1998 1/98 München, MMV Medizin Verlag
Cole TJ Bellizzi MC Flegal KM Dietz WH Establishing a standard definition for child overweight and obesity worldwide: international survey BMJ 2000 320 1240 1243 10797032 10.1136/bmj.320.7244.1240
Kromeyer-Hauschild K Wabitsch M Kunze D Geller F Gelß HC Hesse V Perzentile für den Body-mass-Index für das Kinder- und Jugendalter unter Heranziehung verschiedener deutscher Stichproben. Monatsschrift Kinderheilkunde 2001 149 807 818 10.1007/s001120170107
Kalies H Lenz J von Kries R Prevalence of overweight and obesity and trends in body mass index in German pre-school children, 1982-1997 Int J Obes Relat Metab Disord 2002 26 1211 1217 12187398 10.1038/sj.ijo.0802013
Bruns-Philipps E Dreesman J Adipositas-Bericht. Übergewicht bei Schulanfängern.Eine Auswertung von Schuleingangsuntersuchungen 1993-2003. Gesundheitsberichterstattung 2004 Report 2 Hannover, Niedersächsisches Landesgesundheitsamt
Ministerium für Frauen Jugend Familie und Gesundheit des Landes Nordrhein- Westfalen Gesundheit von Zuwanderern in Nordrhein-Westfalen Gesundheitsberichte NRW 2000 Düsseldorf
Saxena S Ambler G Cole TJ Majeed A Ethnic group differences in overweight and obese children and young people in England: cross sectional survey Arch Dis Child 2004 89 30 36 14709498
Roville-Sausse F [Increase during the last 20 years of body mass of children 0 to 4 years of age born to Maghrebian immigrants] Rev Epidemiol Sante Publique 1999 47 37 44 10214675
Brussaard JH Erp-Baart MA Brants HA Hulshof KF Lowik MR Nutrition and health among migrants in The Netherlands Public Health Nutr 2001 4 659 664 11683558
Oner N Vatansever U Sari A Ekuklu E Guzel A Karasalihoglu S Boris NW Prevalence of underweight, overweight and obesity in Turkish adolescents Swiss Med Wkly 2004 134 529 533 15517506
Neovius M Linne Y Barkeling B Rossner S Discrepancies between classification systems of childhood obesity Obes Rev 2004 5 105 114 15086864 10.1111/j.1467-789X.2004.00136.x
Wang Y Wang JQ A comparison of international references for the assessment of child and adolescent overweight and obesity in different populations Eur J Clin Nutr 2002 56 973 982 12373618 10.1038/sj.ejcn.1601415
Wang Y Monteiro C Popkin BM Trends of obesity and underweight in older children and adolescents in the United States, Brazil, China, and Russia Am J Clin Nutr 2002 75 971 977 12036801
Chinn S Rona RJ International definitions of overweight and obesity for children: a lasting solution? Ann Hum Biol 2002 29 306 313 12031139 10.1080/03014460110085340
Koziel S Kolodziej H Ulijaszek SJ Parental education, body mass index and prevalence of obesity among 14-year-old boys between 1987 and 1997 in Wroclaw, Poland Eur J Epidemiol 2000 16 1163 1167 11484807 10.1023/A:1010924511774
Wang Y Cross-national comparison of childhood obesity: the epidemic and the relationship between obesity and socioeconomic status Int J Epidemiol 2001 30 1129 1136 11689534 10.1093/ije/30.5.1129
Schenk L [Migrant-specific participation barriers and accessibility in The Child and Adolescent Health Survey] Gesundheitswesen 2002 64 Suppl 1 S59 S68 12870218 10.1055/s-2002-39007
| 15882467 | PMC1142329 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 May 9; 5:45 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-45 | oa_comm |
==== Front
BMC PhysiolBMC Physiology1472-6793BioMed Central London 1472-6793-5-61587634610.1186/1472-6793-5-6Research ArticlePKC translocation and ERK1/2 activation in compensated right ventricular hypertrophy secondary to chronic emphysema. Avelar Erick [email protected] Thunder [email protected] Li [email protected] Joel [email protected] Ping [email protected] Sheldon E [email protected] John P [email protected] Division of Cardiology, University of Utah, 30 N. 1900 E., Salt Lake City, UT 84132, USA2 Division of Nutrition, University of Utah, 250 S. 1850 E., #214, Salt Lake City, UT 84112, USA3 Department of Health & Exercise Science, Gustavus Adolphus College, 800 W. College Ave., St. Peter, MN 56082, USA2005 5 5 2005 5 6 6 11 9 2004 5 5 2005 Copyright © 2005 Avelar 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
Right ventricular hypertrophy (RVH) is an important complication of chronic lung disease. However, the signal transduction pathways involved as well as the physiological changes to the right ventricle have not been investigated. Emphysema was produced in male, Syrian Golden hamsters by intra-tracheal instillation of 250 IU/kg elastase (Emp, n = 17). Saline treated animals served as controls (Con, n = 15).
Results
Nine months later, Emp hamsters had 75% greater lung volume, and evidence of RVH at the gross and myocyte level (RV:tibia length Emp 6.84 ± 1.18 vs. Con 5.14 ± 1.11 mg/mm; myocyte cross sectional area Emp 3737 vs. Con 2695 μm2), but not left ventricular hypertrophy. Serial echocardiographic analysis from baseline to nine months after induction of emphysema revealed increasing right ventricular internal dimension and decreased pulmonary artery acceleration time only in Emp hamsters. There was an increase in translocation of PKC βI and PKC ε from cytosolic to membranous cell fractions in RV of Emp hamsters. Phosphorylation of PKC ε was unchanged. Translocation of PKC α and βII were unchanged. Emp animals had a 22% increase in phospho-ERK 1/2, but no change in levels of total ERK 1/2 compared to Con.
Conclusion
These data suggest that PKC βI, ε and ERK 1/2 may play a role in mediating compensated RVH secondary to emphysema and may have clinical relevance in the pathogenesis of RVH.
signal transductionechocardiographypulmonary hypertension
==== Body
Background
Emphysema affects nearly 3 million people and is responsible for nearly 16,000 deaths per year in the United States [1]. Pulmonary parenchymal disease may lead to increased pulmonary vascular resistance and elevated pulmonary arterial pressures. The development of pulmonary hypertension and RVH are markers of increased mortality in emphysema [2]. The problem of cor pulmonale is quite important since in the United States approximately 20% of hospital admissions for heart failure are caused by right heart failure. However, the precise changes in pulmonary hemodynamics, right ventricular structure and function in emphysema, as well as the underlying cell signaling pathways activated in this condition, remain unknown. We investigated these parameters using a hamster model in which emphysema is produced by intratracheal administration of elastase. This causes a reproducible and progressive form of bulbous emphysema characterized by increased lung volume and pulmonary parenchymal pathology similar to that of humans with emphysema [3].
Even though many studies have identified a critical role for many protein kinase C isoforms (PKC α, βI, βII, δ, ε) and extracellular regulated kinase (ERK) 1/2 in the pathogenesis of left ventricular hypertrophy, very few studies have examined these kinases in RVH. It has been previously reported that PKC α, βI, and δ activation is increased in RVH due to pulmonary artery constriction[4], and PKC α and δ activation increased during RVH induced by volume overload[5]. In contrast, there have been no studies to our knowledge regarding the role of ERK1/2 during RVH.
Given the numerous human and animal studies identifying PKC and ERK1/2 as critical mediators of left ventricular hypertrophy [6-12], and the aforementioned studies, we hypothesized that ERK1/2 and differential PKC isoforms would be activated in RVH secondary to emphysema. The progression of changes in right ventricular size and function over time using a noninvasive ultrasound approach, as well as pulmonary hemodynamics were also determined. We report here that hamsters developed moderate, well-compensated RVH characterized by an absence of fibrosis, normal contractility, and mildly increased pulmonary pressure after nine months with emphysema. Furthermore, hamsters with emphysema had marked translocation of PKC βI and PKC ε to membranous cell fractions, along with increased ERK 1/2 phosphorylation.
Results
Total heart weights and RV weights were greater in Emp animals than Con, with the increased total heart weight in Emp being entirely due to the increase in RV mass (Table 1, Figure 1A &1C). Myocyte cross sectional area was greater in Emp right ventricles compared to controls (3737 μM2 vs. 2695 μM2) (Figure 1D) as was myocyte perimeter length (Table 1). The presence of lung pathology and air trapping in the Emp animals was evidenced by a 75% increase in lung volume (Table 1).
Echocardiographic assessment documented progressive right ventricular enlargement over time in the Emp hamsters, with significantly greater right ventricular internal dimension (RVID) vs. Con after nine months of emphysema (Figure 2B). Histological examination of heart cross sections showed no differences in fibrosis in either right or left ventricles of Emp hamsters compared to Con (Figure 1B).
There was a significant correlation (r = 0.68) between the RVID and lung volume measured (Figure 2C). Only trivial amounts of tricuspid valve regurgitation were found in any hamster at any time point and there was no detectable pulmonary insufficiency. Thus, RV and pulmonary artery pressures could not be estimated from the noninvasive studies. Right ventricular outflow tract (RVOT) Doppler profiles were adequate for analysis in all animals. RVOT acceleration time progressively shortened in the Emp animals and was significantly shorter compared to the Con after nine months (Figure 3A,B). There was an inverse correlation between the RVOTAT and RVID (r = - 0.68) at nine months (Figure 3C).
There was a mild, but statistically significant increase in right ventricular systolic pressure (RVSP) after nine months of emphysema (Figure 4A, Table 1). There was a strong positive correlation (r = 0.65) between RVSP and RVID (Figure 4C). Emp hamsters also had a slightly greater +dP/dt and similar -dP/dt compared to the Con group, implying a mostly normal pattern of RV contraction & relaxation (Figure 4B, Table 1).
No differences were detected in translocation of PKC isoforms α and βII between Emp and Con hamsters, as evidenced by the unchanged membrane-to-cytosol ratio (Figure 5A,B). There was a shift in distribution of PKC βI to membrane fractions resulting in an increase in translocation (Figure 5A,B). There was also a 95% increase in translocation of PKC ε in the RV of Emp hamsters (Figure 5A,B). The increase in PKC ε translocation was characterized by unchanged cytosolic fractions in coordination with a 110% increase in PKC ε found in the membrane fraction (Figure 5A,B). The marked increase in membrane levels of PKC ε indicates that protein expression of ε was also increased, with a greater proportion of that isoform being distributed in the membrane fractions. The average translocation index (membrane:cytosol) was 1.14 ± 0.11 for Con and 2.22 ± 0.33 for Emp animals. Levels of phosphorylated PKC ε from whole RV homogenates were quite variable, but means were similar between Con and Emp animals (Figure 5C). Increased levels of PKC ε membrane translocation, however, suggest that more of the phosphorylated PKC ε may be in the membrane fraction of Emp hamsters rather than the cytosol. PKC βI phosphorylation was not determined in this study since antibodies specific for phosphorylated PKC βI are not commercially available.
RV homogenates from Emp hamsters had increased phospho-ERK1/2 levels, resulting in a 34% increase in the ratio of phospho ERK1/2 to ERK1/2 (Figure 6). Total JNK levels were similar between Con and Emp groups (data not shown). Phospho-JNK was not detected in RV homogenates from either Con or Emp hamsters. This result most likely reflects insensitivity of the antibody used against hamster phospho-JNK and should not be taken as evidence that JNK is unimportant in the development of RVH.
Discussion
Despite the morbidity and mortality, associated with right heart failure, the pathogenesis of right ventricular hypertrophy has been studied much less than left ventricular disease. Evidence from clinical observations and animal studies suggest that the right ventricle does not respond to stress in precisely the same fashion as the left ventricle. Most patients with L-transposition of the great arteries, a condition in which the anatomic right ventricle serves as the systemic ventricle from birth, eventually develop severe RV failure. This usually occurs before the age of 30 [13]. Thus, clinical evidence suggests that the genetic program of the right ventricle is not designed to face an increased work load even when the adaptation is very gradual. Several animal studies also show different responses of the right and left ventricles to pressure overload [14,15].
The most commonly studied small animal model of right heart failure is the monocrotaline treated rat. Monocrotaline injection produces severe pulmonary hypertension over a period of several weeks [16]. The elastase treated hamster may be more closely related to clinical features of RVH secondary to emphysema than other models using pulmonary hypertension/overload for two reasons. First, the lung injury develops much more slowly with different degrees of hypoxia leading to varying levels of pulmonary artery constriction. This is substantiated by only modest echocardiographic changes (increased RVID and shortened RVOTAT) found in this study in the first six months after elastase treatment (Figure 2B, 3B). Second, the degree of pulmonary hypertension is less severe than that which is usually seen after monocrotaline injection where pulmonary pressure often exceeds 65 mmHg within six weeks [16,17]. Clinically, most patients with chronic obstructive pulmonary disease such as emphysema have mild or moderate pulmonary hypertension at rest [18], perhaps more similar to the mild increases in RV systolic pressure found in Emp animals in the present study. There have been several hypotheses to explain the development of pulmonary hypertension in the setting of emphysema, with hypoxic vasoconstriction being the most supported mechanism [19]. Other factors such as in situ thrombosis and polycythemia may play a role as well [18].
In the present study RV from the EMP animals had slightly increased +dP/dt, normal -dP/dt, and lack of fibrosis in histological sections, together indicating presence of compensated RVH. Despite these changes, we were unable to detect anything more than trivial degrees of tricuspid regurgitation in the Emp hamsters. Interestingly, even in the absence of tricuspid regurgitation, the RVOT acceleration time showed an inverse correlation with RV size. Based on these data it seems likely RVOT Doppler spectra may be a useful surrogate marker for determining the presence of early or mild forms of pulmonary hypertension in the absence of tricuspic valve regurgitation.
This is the first study to demonstrate PKC βI, PKC ε and ERK1/2 activation in RVH secondary to emphysema. In total we evaluated PKC α, βI, βII, δ, and ε in this study as these isoforms have all been implicated in the pathogenesis of cardiac hypertrophy and ventricular dysfunction in cell culture, insult driven, and transgenic animal models[6,7,12,20]. The unchanged levels of cytosolic PKC ε coupled with a marked increase in membrane PCK ε indicate that expression levels were upregulated in Emp. Thus, it appears right ventricular PKC ε response to emphysema is more complex than a simple re-distribution of basal levels, and involves greater production of PKC ε protein. With specific regard to PKC ε, several lines of evidence support a role for PKC ε in the development of cardiac hypertrophy as reviewed by Sabri et al. [21]. Recent studies using transgenic mice with cardiac specific overexpression of PKC ε now suggest that low to moderate PKC ε activation, characterized by a 6 fold increase in translocation [7] or 228% increase in PKC activity [22] result in physiological cardiac hypertrophy with normal cardiac function. In contrast, a high degree of PKC ε activation, characterized by a 34 fold increase in translocation [7] or 452% increase in activity [22] result in pathological hypertrophy characterized by poor contractility and heart failure. Furthermore, it has also been shown that moderate activation of PKC ε in vivo by it's RACK (Receptor for Activated C Kinase) increases translocation of PKC ε by only 20% and results in mild compensated hypertrophy [23]. It is thought that in the case of excessive PKC ε activation, heart failure is mediated by PKC ε induced cross activation of PKC βII, which has been previously demonstrated to cause a heart failure phenotype in transgenic mice [9,12]. The present study found a 95% increase in PKC ε translocation, but no increase in PKC βII in the RV of Emp hamsters. This appears to fall in to the range of "moderate" levels of PKC ε translocation that have been associated with compensated/physiological hypertrophy in the aforementioned studies. In contrast with our results, another study using rats subjected to pulmonary artery constriction for 3 weeks found greater membrane levels of PKC α and δ, but no change in PKCε [4]. It is possible that greater RV systolic pressure after pulmonary constriction (44 mmHg in pulmonary constricted rats vs. 26 mmHg in Emp hamsters) resulted in differential activation of PKC isoforms. Other differences that may have contributed to these contrasting results include the duration of the study (3 weeks vs. 9 months), the degree of RVH developed (much less in Emp hamsters), and the species studied.
An analysis of total PKC ε phosphorylation in whole RV homogenates found similar levels in both Emp and Con hamsters, however, these samples also demonstrated significant variation in phosphorylation state among individual hamsters. This may be due to the phospho specific PKC ε antibody used, which was derived against a human epitope, and may have had reduced specificity against the hamster protein. We used an antibody that recognized PKC ε phosphorylated at Ser 729, an amino acid in the C-terminal hydrophobic motif. Ser 729 is autophosphorylated by PKC ε itself following PDK-1 dependent phosphorylation of Thr 566[24]. PKC ε phosphorylated at this site is recognized as mature and catalytically active[25].
Previous studies using neonatal myocytes, animals models, and transgenic mice indicate that PKC β isoforms, particularly βII, also are key mediators in the pathophysiology of cardiac hypertrophy and heart failure [12,26]. While there are no transgenic models that specifically implicate PKC βI in pathogenesis of hypertrophy, other studies have found activation to be present during pressure overload hypertrophy in vivo [27,28]. Translocation of PKC βI may also be clinically relevant since increased activity in PKC β (I and II) has been reported by Bowling et al. in failing human hearts [10]. Our data is partially consistent with these finding as we found a significant increase in translocation of PKC βI, but no changes in PKC βII in hypertrophied RV of Emp hamsters. Since previous studies using transgenic mice indicate that PKC βII appears to produce a heart failure phenotype [9,12], we speculate that the lack of PKC βII translocation in this study may due to the compensated nature of the hypertrophied right ventricle at the time of analysis. Taken together, our data on PKC translocation suggest that future studies using pharmacological agents that can selectively inhibit PKC isoforms will help confirm a specific role for PKC in this form of RVH, and give insight into possible therapeutic interventions to limit the progression of RVH secondary to emphysema.
Levels of phospho ERK1/2 were 22% greater, along with a 34% increase in phospho ERK:total ERK in Emp compared to Con. These are mild increases compared to other in vivo studies using pressure overload imposed upon the left ventricle [29-31]. The importance of ERK signaling has been established in the hypertrophic response with recent evidence supporting a role for adaptive hypertrophy and survival as opposed to heart failure [32,33]. Given that previous studies have reported more vigorous ERK activation in response to pressure overload, it is difficult to determine the true physiological relevance of our ERK data with regards to the pathogenesis of right ventricular hypertrophy. However, it is possible that the degree of ERK activation may fluctuate over time as hypertrophy develops and progresses. However, a potential limitation in our phospho PKC ε and ERK1/2 analyses is that whole tissue homogenates were used for experiments. Thus we cannot assess the contribution from the non muscle fraction where activation of EGF receptor and receptor tyrosine kinases may be operative. Another limitation is the analysis of signal kinases after the establishment of RVH. While these data correlate PKC βI and ε translocation, and EKR1/2 activation with RVH, they do not confirm these signal kinases in mediating development RVH.
Conclusion
This study underscores the integrated nature of the cardio-pulmonary system such that experimentally induced lung disease significantly impacted cardiac physiology. To the best of our knowledge, the present study is the first to have identified activation of PKC and ERK/12 in RVH secondary to emphysema, and to describe the progressive nature of alterations in the right ventricle during progression of emphysema. Taken in context of previous studies, these data support the hypothesis that PKC and ERK may play an important role in the pathophysiology of RVH secondary to emphysema. It is also evident that after nine months of emphysema in this model, the RV is in a compensated pattern of hypertrophy with preserved systolic function. Overall, we believe that our data regarding right ventricular structure and function in this model of chronic parenchymal lung disease with mild RV pressure overload complement data from other studies of acute severe pulmonary hypertension induced with monocrotaline. Additional studies determining status of PKC and ERK immediately prior to the development of RVH in this model would help further define their relevance as critical mediators of RVH secondary to emphysema. Data from the present study, as well as future studies, may be useful to identify new diagnostic and treatment paradigms for right ventricular overload.
Methods
Emphysema model
Male Syrian Golden hamsters (9 week old, 100–121 g body weight) were maintained on a 12:12-h light-dark cycle, and supplied with rodent chow and water ad libitum. After one-week habituation period, the animals were divided into control (Con, n = 15) and emphysema (Emp, n = 17) groups at random. Under deep anesthesia with ketamine/xylazine (150/7.5 mg/kg im), either saline (0.3 ml/100 g body wt) or porcine elastase [25 IU/100 g body wt (Sigma Chemical, St. Louis, MO) in 0.3 ml of normal saline] was instilled intratracheally using a 27-gauge hypodermic needle as previously described [34].
Echocardiographic measurements
Parameters of cardiac structure and function were determined with transthoracic echocardiography (n = 8 con and n = 10 Emp hamsters) as previously described [35,36]. Hamsters were anesthetized using chloral hydrate (300 mg/kg, intramuscularly) and the chest was shaved. Two-dimensional and Doppler imaging were performed using a 4.5 – 10 MHz transducer (Vivid Five, General Electric). Right ventricular internal diastolic dimension (RVID) was measured in the apical four chamber view just below the level of the tricuspid valve. Pulsed-wave Doppler of the pulmonary outflow tract was recorded in the parasternal short axis view at the level of the aortic valve. The sample volume was placed proximal (3 mm) to the pulmonary valve leaflets and aligned to maximize laminar flow and the right ventricular outflow tract acceleration time (RVOTAT), velocity-time integral and ejection time were measured. Acceleration time was measured from the onset of systolic flow to highest laminar velocity in the spectral signal (Figure 2A). The tricuspid valve was examined for the presence of tricuspid regurgitation (TR) with color and continuous-wave Doppler in the parasternal short axis and apical four-chamber views. Digital images were analyzed off-line (EchoPAC, General Electric). An echocardiographer who was blinded to the treatment group performed and analyzed all studies. All measurements represent the means of three cardiac cycles.
Right heart catheterization
Following the final echo (nine months after elastase treatment) while still under anesthesia with chloral hydrate the right jugular vein was exposed and a temperature-calibrated 1.4 F micromanometer-tipped catheter (Millar Instruments) was inserted into the vein and advanced in retrograde fashion into the right ventricle. Heart rate, right ventricular pressure and first derivative of pressure (dP/dt) were recorded.
Tissue harvesting
Nine months following elastase injection, hamsters were euthanized and the hearts and lungs were removed. Whole hearts were quickly weighed and then placed in 4°C saline where the RV was dissected and weighed separately. In a subgroup of animals (Con, n = 8; Emp n = 10) the RVs were immediately frozen in liquid N2 and stored separately at -80°C until further analysis. A saline displacement technique was used to measure excised lung volume at 0-cm H2O airway pressure as previously described [37]. In addition, for each hamster, the left tibia was dissected and the length was measured so that RV mass and total heart mass could be normalized to the tibial length as has been previously described and validated[38].
Histological Assessment and myocyte measurements
In another subgroup of hamsters (Con, n = 7; Emp n = 5), the hearts were rinsed in ice cold saline and then fixed by immersion in formalin. Short axis slices from the mid wall level of hearts were cut in 10 μM sections from paraffin embedded tissue. Masson's trichrome-stained sections were examined by a pathologist (blinded to the animal identifications) and evaluated for presence of fibrosis.
Sections of RV adjacent to the portion used for histology were subjected to a previously reported technique designed to dissociate myocytes from fixed tissues [39]. Briefly, small pieces of the formalin-fixed RV tissue (~2 × 2 mm) were minced and incubated for 2 h in a 50% KOH solution (w/v made in 100 mM PBS) at 37°C. The partially dissociated tissue was then placed in a 2% trypsin solution for 3 hours and mechanically teased apart to produce mostly single myocytes. Samples were then centrifuged at1000 RPM for 5 min and the pellet containing intact myocytes was recovered. Myocytes were placed on glass slides and multiple fields from each heart were photographed with a digital camera at 20×. Right ventricular cell surface area and perimeter length was measured using NIH image software (V. 1.63) from 555 control and 524 Emp myocytes.
Preparation of RV homogenates for PKC immunoblotting
Cardiac lysates containing cytosolic and membrane proteins were prepared from the hamster RV frozen at -80°C as previously detailed [6]. All extraction procedures were performed at 4°C. Protein concentration of cell fractions were determined using a Bio Rad Protein assay (Bio-Rad, Hercules, CA) with bovine serum albumin (BSA) as a standard.
Preparation of RV homogenates for ERK 1/2, JNK, phospho PKC ε, and phospho c-Raf immunoblotting
The RV was homogenized with a tissuemizer in 1 ml of ice-cold RIPA buffer [50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM sodium orthovanadate, 1 mM NaF, and 10 μl/mL Sigma protease inhibitor cocktail (Sigma, St. Louis, MO, cat. #P-8340)]. Samples were then sonicated twice on ice and centrifuged at 11,000 × g for 10 mins at 4°C. Supernatants were recovered and stored at -80°C for subsequent immunoblotting. Protein concentration of RV lysate was determined using a Bio Rad Protein assay (Bio-Rad, Hercules, CA) with bovine serum albumin (BSA) as a standard.
Western blotting, transfer and densitometry
Electrophoresis and transfer of proteins to PVDF membranes were done as previously described [6,29]. Gels were loaded with 10 μg per lane for PKC α, βII, and ε experiments, 20 μg per lane for PKC βII experiments, and 50 μg per lane for ERK and JNK experiments. All PKC antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) all other antibodies / reagents were purchased from Cell Signal Technology (Beverly, MA). Primary antibody directed against PKC α, βI, βII, and ε were incubated overnight at 4°C in a 1:1000 dilution. Antibodies directed against total and phosphorylated ERK 1/2 & JNK, phospho c-Raf, and phospho PKC ε (recognizing phosphorylated Ser 729) were incubated at a 1:1000 dilution for 48 h at 4°C in 5% BSA Tris buffer with 0.05% Tween-20. Secondary antibody conjugated to horseradish peroxidase was incubated for 1 h at 1:10,000 dilution. Signals were visualized by enhanced chemiluminescence. Relative band density on film was measured with a scanner using NIH 1.63 image software (National Institutes of Health, Rockville, MD).
Statistical analysis
For all data except echocardiographic data, an unpaired Student t-test was performed to detect the difference between Emp and Con animals using SPSS V. 10 for Macintosh (Chicago, IL). For changes in echocardiographic parameters over time, a general linear model with repeated measures was used. Pearson's correlation coefficient was used to describe the relationship among the echocardiographic variables, hemodynamic data, and gross anatomical measurements. Significance was accepted at p ≤ 0.05 for all tests. All values are expressed as mean ± SD.
Authors' contributions
TJ helped supervise the scientific direction of project, assisted with immunoblots, conducted histology studies, analyzed immublots, histology, myocyte, animal morphometrical data, and drafted the manuscript. EA conducted serial echocardiographs and analysis, and drafted the manuscript. LD did the immunoblots and assisted in data analysis. JA determined myocyte dimensions. SEL helped supervise the scientific direction of the project, assisted with the echocardiographs and analysis, and helped draft the manuscript. JPM conceived of the study, prepared the hamster model, assisted in echocardiographs, and helped draft the manuscript.
Acknowledgements
The authors wish to acknowledge Hui-feng Jin and Rockni Jalili, M.D. for technical assistance. This work was supported by grants from the Morrison Trust Foundation to TJ, Department of Veterans Affairs and the National Institute of Health (HL 52338-06, T32HL7576) to SEL, and American Lung Association (ALA RG-013-N) to JM.
Figures and Tables
Figure 1 A. Gross cross sections of hearts from hamsters with emphysema and age matched controls. Hamsters with emphysema demonstrate a range of RV free wall thickness and chamber dilation, however all are larger than controls. Reference scale on bottom of picture is in mm. B. Histological sections of hearts were stained with Masson's Trichrome to visualize collagen content. No significant evidence of fibrosis were seen in Emp hamsters compared to Con. C. Right ventricle:tibia length is greater in Emp animals compared to Con. D. Myocytes cross sectional area was greater in right ventricles from Emp animals compared to Con. Data collected from individual myocytes (524 control, 555 Emp) collected from 7 Con and 5 Emp hamsters. Con, control; Emp, emphysema.
Figure 2 A. An example of the right ventricular internal diastolic dimension (RVID) as measured from the apical four-chamber view (RA, right atrium; RV, right ventricle; LA, left atrium; LV, left ventricle). B. RVID was measured at baseline, 3, 6, 8 and 9 months in the Emp group and at baseline and nine months in the Con group. Note the significant RV enlargement over time in the Emp group. RVID in the Emp group was significantly larger than the Con group after nine months (*p ≤ 0.05). C. There is a significant positive linear correlation between RVID and lung volume for both the Emp and Con groups after nine months. Con, control; Emp, emphysema.
Figure 3 A. Representative examples of right ventricular outflow tract Doppler flow patterns in Con and Emp hamsters nine months after elastase or vehicle treatment. A more symmetric contour of the Doppler spectrum is seen in the Con group as compared to the Emp animals which show an earlier peak flow velocity. The points used for measurement of RVOT acceleration time are indicated. B. Right ventricular outflow tract Doppler spectra were measured at baseline, 3, 6, 8 and 9 months in the Emp group and at baseline and nine months in the Con group. Note the significant decrease in pulmonary artery acceleration time (PAAT) over time in the Emp group. Right ventricular outflow tract acceleration time in the Emp group was significantly shorter that the Con group after nine months (*p ≤ 0.05 vs. Con). C. Right ventricular internal diastolic dimension (RVID) vs. PAAT for both the Emp and Con groups after nine months. Note the significant inverse linear correlation between RVID and PAAT in the Emp hamsters. Con, control; Emp, emphysema.
Figure 4 A. Representative examples of right ventricular pressure tracing in Con (n = 7) and Emp (n = 7) hamsters. Right ventricular systolic pressure (RVSP) was significantly higher in the Emp group. B. First derivative of RV pressure (dP/dt) obtained in Con (n = 7) and Emp (n = 7) hamsters. There was a higher +dP/dt in the Emp group vs. Con. C. There is a significant positive linear correlation between RVID and RVSP in Con and Emp groups suggesting that the increased RVSP, although mild, is enough to produce RV enlargement. Con, control; Emp, emphysema.
Figure 5 A. Protein kinase C (PKC) immunoreactivity during right ventricular hypertrophy in hamsters with emphysema. Membranous (Mem) and cytosolic (Cyt) fractions were prepared from hamster right ventricle. Representative immunoblots for PKC ε and PKC βI demonstrate increased translocation in Emp, while PKC α and βII were unchanged. Translocation was quantified by taking ratio of membrane to cytosolic levels (M:C) of each isoform. Average PKC ε translocation index (of three separate trials) was 1.14 ± 0.11 for Con (n = 5) and 2.22 ± 0.33 for Emp (n = 5) animals. B. Graph of M:C illustrates fold change normalized to controls. Fold changes are averages of three separate experiments for each PKC isoform. C. Levels of total phosphorylated PKC ε from whole right ventricular homogenates were unchanged among Con and Emp hamsters. *p ≤ 0.05. Con, control; Emp, emphysema.
Figure 6 Immunoblots of whole RV homogenates demonstrating a 34% increase in phERK1/2 to ERK1/2 ratio. Con, control; Emp, emphysema.
Table 1 Organ weights, right ventricle analysis, and heart rates of hamsters with emphysema.
Control n Emphysema n
Initial Body Wt. (g) 113 ± 9 15 112 ± 6 17
Final Body Wt. (g) 133 ± 16 15 132 ± 20 17
Lung Volume (mL) 1.43 ± 0.20 15 2.51 ± 0.68* 17
Heart Wt. (mg) 544 ± 78 15 616 ± 90* 17
Heart:tibia length (mg:mm) 19.4 ± 2.6 15 21.2 ± 3.2 17
RV Wt. (mg) 145 ± 33 8 197 ± 34* 10
Heart Rate (beats/min) 192 ± 30 7 190 ± 27 7
RVSP 22 ± 2 7 26 ± 2* 7
RVEDP 0.4 ± 0.4 7 0.9 ± 0.9 7
+dP/dT 1016 ± 162 7 1214 ± 139* 7
-dP/dT -827 ± 142 7 -944 ± 114 7
Myocyte perimeter length (μM) 250 ± 11 5 302 ± 18* 5
*p ≤ 0.05 RV, right ventricle; RVSP, right ventricular systolic pressure (mmHg); RVEDP, right ventricular end-diastolic pressure (mmHg); dP/dt, first derivative of pressure (mmHg/sec).
==== Refs
Lucas JW SJSBV Summary health statistics for U.S. Adults: National Health Interview Survey, 2001. National Center for Health Statistics, Vital Health Statistics 2004 10 Hyattsville, MD , National Center for Health Statistics
Vizza CD Lynch JP Ochoa LL Richardson G Trulock EP Right and left ventricular dysfunction in patients with severe pulmonary disease Chest 1998 113 576 583 9515827
Hayes JA Christensen TG Snider GL The hamster as a model of chronic bronchitis and emphysema in man Lab Anim Sci 1977 27 762 770 592728
Braun MU Szalai P Strasser RH Borst MM Right ventricular hypertrophy and apoptosis after pulmonary artery banding: regulation of PKC isozymes Cardiovasc Res 2003 59 658 667 14499867 10.1016/S0008-6363(03)00470-X
Braun MU LaRosee P Simonis G Borst MM Strasser RH Regulation of protein kinase C isozymes in volume overload cardiac hypertrophy Mol Cell Biochem 2004 262 135 143 15532718 10.1023/B:MCBI.0000038229.23132.9f
Jalili T Takeishi Y Song G Ball NA Howles G Walsh RA PKC translocation without changes in Galphaq and PLC-beta protein abundance in cardiac hypertrophy and failure Am J Physiol 1999 277 H2298 304 10600849
Takeishi Y Ping P Bolli R Kirkpatrick DL Hoit BD Walsh RA Transgenic overexpression of constitutively active protein kinase C epsilon causes concentric cardiac hypertrophy Circ Res 2000 86 1218 1223 10864911
Bueno OF Molkentin JD Involvement of extracellular signal-regulated kinases 1/2 in cardiac hypertrophy and cell death Circ Res 2002 91 776 781 12411391 10.1161/01.RES.0000038488.38975.1A
Pass JM Gao J Jones WK Wead WB Wu X Zhang J Baines CP Bolli R Zheng YT Joshua IG Ping P Enhanced PKC beta II translocation and PKC beta II-RACK1 interactions in PKC epsilon-induced heart failure: a role for RACK1 Am J Physiol Heart Circ Physiol 2001 281 H2500 10 11709417
Bowling N Walsh RA Song G Estridge T Sandusky GE Fouts RL Mintze K Pickard T Roden R Bristow MR Sabbah HN Mizrahi JL Gromo G King GL Vlahos CJ Increased protein kinase C activity and expression of Ca2+-sensitive isoforms in the failing human heart Circulation 1999 99 384 391 9918525
Haq S Choukroun G Lim H Tymitz KM del Monte F Gwathmey J Grazette L Michael A Hajjar R Force T Molkentin JD Differential activation of signal transduction pathways in human hearts with hypertrophy versus advanced heart failure Circulation 2001 103 670 677 11156878
Wakasaki H Koya D Schoen FJ Jirousek MR Ways DK Hoit BD Walsh RA King GL Targeted overexpression of protein kinase C beta2 isoform in myocardium causes cardiomyopathy Proc Natl Acad Sci U S A 1997 94 9320 9325 9256480 10.1073/pnas.94.17.9320
Dyer K Graham TP Congenitally Corrected Transposition of the Great Arteries: Current Treatment Options Curr Treat Options Cardiovasc Med 2003 5 399 407 12941208
Perreault CL Bing OH Brooks WW Ransil BJ Morgan JP Differential effects of cardiac hypertrophy and failure on right versus left ventricular calcium activation Circ Res 1990 67 707 712 2144483
Ecarnot-Laubriet A Rochette L Vergely C Sicard P Teyssier JR The activation pattern of the antioxidant enzymes in the right ventricle of rat in response to pressure overload is of heart failure type Heart Dis 2003 5 308 312 14503927 10.1097/01.hdx.0000089836.03515.a9
Bruner LH Hilliker KS Roth RA Pulmonary hypertension and ECG changes from monocrotaline pyrrole in the rat Am J Physiol 1983 245 H300 6 6224428
Werchan PM Summer WR Gerdes AM McDonough KH Right ventricular performance after monocrotaline-induced pulmonary hypertension Am J Physiol 1989 256 H1328 36 2524170
Voelkel NF Cool CD Pulmonary vascular involvement in chronic obstructive pulmonary disease Eur Respir J Suppl 2003 46 28s 32s 14621104
Evans AM Dipp M Hypoxic pulmonary vasoconstriction: cyclic adenosine diphosphate-ribose, smooth muscle Ca(2+) stores and the endothelium Respir Physiol Neurobiol 2002 132 3 15 12126692 10.1016/S1569-9048(02)00046-0
Hahn HS Marreez Y Odley A Sterbling A Yussman MG Hilty KC Bodi I Liggett SB Schwartz A Dorn GW Protein kinase Calpha negatively regulates systolic and diastolic function in pathological hypertrophy Circ Res 2003 93 1111 1119 14605019 10.1161/01.RES.0000105087.79373.17
Sabri A Steinberg SF Protein kinase C isoform-selective signals that lead to cardiac hypertrophy and the progression of heart failure Mol Cell Biochem 2003 251 97 101 14575310 10.1023/A:1025490017780
Pass JM Zheng Y Wead WB Zhang J Li RC Bolli R Ping P PKCepsilon activation induces dichotomous cardiac phenotypes and modulates PKCepsilon-RACK interactions and RACK expression Am J Physiol Heart Circ Physiol 2001 280 H946 55 11179034
Mochly-Rosen D Wu G Hahn H Osinska H Liron T Lorenz JN Yatani A Robbins J Dorn GW Cardiotrophic effects of protein kinase C epsilon: analysis by in vivo modulation of PKCepsilon translocation Circ Res 2000 86 1173 1179 10850970
Cenni V Doppler H Sonnenburg ED Maraldi N Newton AC Toker A Regulation of novel protein kinase C epsilon by phosphorylation Biochem J 2002 363 537 545 11964154 10.1042/0264-6021:3630537
Newton AC Regulation of the ABC kinases by phosphorylation: protein kinase C as a paradigm Biochem J 2003 370 361 371 12495431 10.1042/BJ20021626
Bowman JC Steinberg SF Jiang T Geenen DL Fishman GI Buttrick PM Expression of protein kinase C beta in the heart causes hypertrophy in adult mice and sudden death in neonates J Clin Invest 1997 100 2189 2195 9410895
Inagaki K Iwanaga Y Sarai N Onozawa Y Takenaka H Mochly-Rosen D Kihara Y Tissue angiotensin II during progression or ventricular hypertrophy to heart failure in hypertensive rats; differential effects on PKC epsilon and PKC beta J Mol Cell Cardiol 2002 34 1377 1385 12392998 10.1006/jmcc.2002.2089
Gu X Bishop SP Increased protein kinase C and isozyme redistribution in pressure-overload cardiac hypertrophy in the rat Circ Res 1994 75 926 931 7923639
Takeishi Y Huang Q Abe J Glassman M Che W Lee JD Kawakatsu H Lawrence EG Hoit BD Berk BC Walsh RA Src and multiple MAP kinase activation in cardiac hypertrophy and congestive heart failure under chronic pressure-overload: comparison with acute mechanical stretch J Mol Cell Cardiol 2001 33 1637 1648 11549343 10.1006/jmcc.2001.1427
Kim S Iwao H Activation of mitogen-activated protein kinases in cardiovascular hypertrophy and remodeling Jpn J Pharmacol 1999 80 97 102 10440527 10.1254/jjp.80.97
Rapacciuolo A Esposito G Caron K Mao L Thomas SA Rockman HA Important role of endogenous norepinephrine and epinephrine in the development of in vivo pressure-overload cardiac hypertrophy J Am Coll Cardiol 2001 38 876 882 11527648 10.1016/S0735-1097(01)01433-4
Bueno OF De Windt LJ Tymitz KM Witt SA Kimball TR Klevitsky R Hewett TE Jones SP Lefer DJ Peng CF Kitsis RN Molkentin JD The MEK1-ERK1/2 signaling pathway promotes compensated cardiac hypertrophy in transgenic mice Embo J 2000 19 6341 6350 11101507 10.1093/emboj/19.23.6341
Harris IS Zhang S Treskov I Kovacs A Weinheimer C Muslin AJ Raf-1 kinase is required for cardiac hypertrophy and cardiomyocyte survival in response to pressure overload Circulation 2004 110 718 723 15289381 10.1161/01.CIR.0000138190.50127.6A
Mattson JP Poole DC Pulmonary emphysema decreases hamster skeletal muscle oxidative enzyme capacity J Appl Physiol 1998 85 210 214 9655777
Litwin SE Katz SE Weinberg EO Lorell BH Aurigemma GP Douglas PS Serial echocardiographic-Doppler assessment of left ventricular geometry and function in rats with pressure-overload hypertrophy. Chronic angiotensin-converting enzyme inhibition attenuates the transition to heart failure Circulation 1995 91 2642 2654 7743628
Bossone E Avelar E Bach DS Gillespie B Rubenfire M Armstrong WF Diagnostic value of resting tricuspid regurgitation velocity and right ventricular ejection flow parameters for the detection of exercise induced pulmonary arterial hypertension Int J Card Imaging 2000 16 429 436 11482708 10.1023/A:1010604913656
Scherle W A simple method for volumetry of organs in quantitative stereology Mikroskopie 1970 26 57 60 5530651
Yin FC Spurgeon HA Rakusan K Weisfeldt ML Lakatta EG Use of tibial length to quantify cardiac hypertrophy: application in the aging rat Am J Physiol 1982 243 H941 7 6216817
Gerdes AM Onodera T Tamura T Said S Bohlmeyer TJ Abraham WT Bristow MR New method to evaluate myocyte remodeling from formalin-fixed biopsy and autopsy material J Card Fail 1998 4 343 348 9924856 10.1016/S1071-9164(98)90240-8
| 15876346 | PMC1142330 | CC BY | 2021-01-04 16:03:51 | no | BMC Physiol. 2005 May 5; 5:6 | utf-8 | BMC Physiol | 2,005 | 10.1186/1472-6793-5-6 | oa_comm |
==== Front
BMC PsychiatryBMC Psychiatry1471-244XBioMed Central London 1471-244X-5-231589288410.1186/1471-244X-5-23Research ArticleNo evidence for association between polymorphisms in GRM3 and schizophrenia Norton Nadine [email protected] Hywel J [email protected] Sarah [email protected] Dobril [email protected] Anna C [email protected] Amy [email protected] Nigel M [email protected] Pamela [email protected] Stanley [email protected]'Donovan Michael C [email protected] Michael J [email protected] Department of Psychological Medicine, Wales School of Medicine, Henry Wellcome Building, Cardiff University, Heath Park, Cardiff, UK2005 13 5 2005 5 23 23 15 2 2005 13 5 2005 Copyright © 2005 Norton et al; licensee BioMed Central Ltd.2005Norton 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
Three studies have previously reported data that were interpreted by the authors as supportive of association between schizophrenia and polymorphisms in the gene encoding the metabotropic glutamate receptor GRM3.
Methods
In a bid to examine this hypothesis, we examined seven SNPs spanning GRM3 in a UK case-control sample (schizophrenic cases n = 674, controls n = 716). These included all SNPs previously reported to be associated, alone or in haplotypes, with schizophrenia in European or European American samples.
Results
Our data showed no evidence for association with single markers, or 2, 3, 4 and 5 marker haplotypes, nor did any specific haplotypes show evidence for association according to previously observed patterns.
Conclusion
Examination of our own data and those of other groups leads us to conclude that at present, GRM3 should not be viewed as a gene for which there is replicated evidence for association with schizophrenia.
==== Body
Background
Glutamate is the major excitatory neurotransmitter in the central nervous system (CNS) where its effects are mediated through a diverse group of ionotropic receptors [N-methyl D-aspartate receptors (NMDA), alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionate receptors (AMPA), kainate (KA) and delta] and eight metabotropic receptor isoforms. Abnormal glutamatergic transmission has been proposed in schizophrenia, initially from the observation that exposure to NMDA receptor antagonists such as phencyclidine (PCP) can cause symptoms of schizophrenia [1,2] and also relapse in patients with schizophrenia [3]. More recently, agonists of metabotropic glutamate receptors 2 and 3 (GRM2 and GRM3) have been shown to ameliorate the adverse behavioral effects induced by phencyclidine in the mouse [4].
The GRM3 gene (GRM3) is therefore an attractive candidate gene for schizophrenia and there have been several reports of allelic association studies examining polymorphisms in this gene [5-8]. These findings, however, do not provide compelling evidence for association. Marti et al., [5] screened the coding exons and flanking intronic sequence of GRM3 for polymorphisms in 46 schizophrenics and 46 bipolar patients. Three variants were identified, one of which (+1131C>T, from here on referred to as rs2228595) revealed significant (p = 0.002) evidence for association in a German sample of 265 schizophrenics and 227 controls, a finding that did not replicate in the same study either in a second case-control sample or in a family-based association sample, each if which was also of German origin [5]. The second study of this gene was based upon low density linkage disequilibrium (LD) analysis in a Japanese population [6]. Of six SNPs examined, one, rs1468412, was nominally significantly associated with schizophrenia (p = 0.01). A number of 2 and 3 marker haplotype combinations also yielded support for the association that was surprisingly strong (minimum global p ~0.001) given the sample size was only 100 schizophrenic cases and 100 controls. In the third study [7], rs2228595, which showed un-replicated evidence for increased frequency of the T allele in the earlier German study [5] showed a trend for association of the C allele, (p = 0.08), in a sample of white European American siblings, but not in a second sample of 67 white European American families with one or more affected offspring, or in the two samples combined (616 transmissions observed, 612 expected). Analysis of a further 6 SNPs revealed nominally significant evidence (p = 0.02, uncorrected) for association between an intronic SNP (hcV11245618, from here on referred to as rs6465084) in their first sample. This finding did not replicate in the second Caucasian sample, with any trend being towards excess transmission of the opposite allele, nor in an African American sample of 51 families. The latter sample however yielded evidence for association (p = 0.03) to the allele at rs1468412 that was significantly under-represented in the earlier Japanese study [6]. Haplotype analysis in their Caucasian sib sample revealed a number of related significant 3 and 5 marker haplotypes (minimum global p = 0.006) but no replication data were presented and it is difficult to determine from the presented data to what extent these findings could be expected by chance.
In the most recent study, 7 SNPs were examined in 752 Han Chinese cases and controls [8]. Of the SNPs previously reported to be associated, only rs1468412 was genotyped in this study but this did not reveal evidence for association. However, rs2299225 (which has not been genotyped in any other published datasets) yielded nominally significant evidence (p = 0.03, uncorrected) as did a 3 marker haplotype (global p = 0.0017). Examination of the data suggest that the latter analysis essentially reflects the association at the single locus (frequencies of rs2299225 risk allele = 6.3% in cases, 4.5% in controls; frequency of three marker risk haplotype in cases = 6.2% in cases, 4% in controls), and the extra 'global' significance may have resulted from the use of a statistical test that does not allow for any error in estimating haplotype frequencies.
The hypothesis that this gene may influence schizophrenia susceptibility has gained currency from the observation that homozygosity for one of the alleles of a SNP in GRM3 is associated with certain aspects of cognition [7] that have been proposed as intermediate phenotypes for the disorder. Moreover, in that study, homozygosity of the same SNP was also associated in vivo with lower indices of synaptic neurotransmission and glutamatatergic function as determined by neuroimaging, and also in post-mortem brain by lower expression of the glial glutamate transporter EAAT2.
Thus the genetic evidence in favour of GRM3 is at present ambiguous. What is, however, abundantly clear from two of the studies [6,7] is that it is necessary to invoke genetic heterogeneity (of patterns of LD and/or of the susceptibility alleles per se) with respect to ethnicity if the existing data are to be interpreted as supportive of GRM3 as a true susceptibility gene for schizophrenia. Therefore in a bid to attempt to provide support for GRM3 as a susceptibility gene for schizophrenia, we have focussed on those polymorphisms and haplotypes that have provided significant evidence in Caucasian samples. We typed the SNP showing significance in one of the original German samples [5] as well as the SNPs reported by Egan and colleagues [7], in a large UK based case control sample of 674 cases and 716 controls.
Methods
Subjects
All case-control subjects used in this study were unrelated Caucasians born in the UK or Ireland. All cases met DSM-IV criteria for schizophrenia. Consensus diagnoses were made by two raters from all available information following a semi-structured interview, SCAN or PSE [9,10], and examination of case notes. The cases consisted of 456 males and 218 females, average age at collection 44.5 years ± 14.6, whilst the controls consisted of 482 males and 234 females, average age at collection 41.5 years ± 11.5 years. Control individuals were group matched to cases for age, sex, and ethnicity from more than 1400 blood donors recruited from the National Blood Transfusion Service. Individuals on medication are not allowed to donate blood in the UK nor are they remunerated even for expenses. Thus unlike in some countries, donating blood in the UK is entirely an altruistic process that does not tend to enrich for indigents, or people with substance abuses or psychosis. Donors were not screened for the absence of psychiatric illness, as this does not affect the power when a disease has the population prevalence of schizophrenia [11]. Multicentre and Local Research Ethics Committee approval was obtained, and all subjects, both cases and controls, gave written informed consent to participate.
Genotyping
All SNPs were genotyped using the Sequenom MassARRAY™ system as per the manufacturer's instructions. Assay design and PCR conditions are available on request. All assays were optimised initially by genotyping DNA from 30 CEPH parent-offspring trios from 21 families (Utah residents with ancestry from northern and western Europe), as detailed in the international hapmap project [12]. 46 of these DNA samples were re-genotyped along with the case-control sample to provide a measure of genotyping accuracy. All genotypes were called blind to sample identity and affected status.
Statistical analysis
All p values are two tailed. Tests of genotypic and allelic association were performed using contingency tables. Haplotype analyses were performed using the EM algorithm and a permutation test as implemented in program EH plus [13]. Linkage disequilibrium values were calculated using the ldmax program within the GOLD software [14].
Results
Allele frequencies for all SNPs and haplotype frequencies were similar to those reported previously in a Caucasian sample [7]. All markers were in Hardy-Weinberg equilibrium in both cases and controls. Double genotyping of 46 CEPH DNAs for all markers gave 100% concordance (a total of 312 genotypes). Genotype data were available from the Hapmap database for rs917071 and rs1468412. The genotypes we obtained for these individuals were identical to those in the Hapmap database. No significant differences, allelic or genotypic, were observed between cases and controls for any marker even uncorrected for multiple testing (table 1). The only finding approaching nominal significance was for marker rs187993 which yielded a genotypic p-value of 0.06 (2df). No 2, 3, 4 or 5 marker haplotypes yielded evidence for association (data not shown), nor did any of the specific 3 and 5 marker haplotypes reported as significant by Egan et al., [7] under specific 1df tests (table 2).
Table 1 GRM3 single marker data
n 11 12 22 p-value 1 2 p-value
rs187993 case 590 303 228 59 0.06 834 (0.71) 346 (0.29) 0.11
T/G control 637 287 288 62 862 (0.68) 412 (0.32)
rs13242038 case 660 431 213 16 0.19 1075 (0.81) 245 (0.19) 0.60
C/T control 703 460 214 29 1134 (0.81) 272 (0.19)
rs917071 case 659 337 272 50 0.61 946 (0.72) 372 (0.28) 0.33
C/T control 697 373 278 46 1024 (0.73) 370 (0.27)
rs6465084 case 666 375 246 45 0.93 996 (0.75) 336 (0.25) 0.73
A/G control 708 406 255 47 1067 (0.75) 349 (0.25)
rs2228595 case 663 561 98 4 0.85 1220 (0.92) 106 (0.08) 0.64
C/T control 705 602 100 3 1304 (0.92) 106 (0.08)
rs1468412 case 663 345 264 54 0.68 954 (0.72) 372 (0.28) 0.41
A/T control 698 375 274 49 1024 (0.73) 372 (0.27)
rs7804100 case 652 363 250 39 0.63 976 (0.75) 328 (0.25) 0.46
G/A control 704 401 269 34 1071 (0.76) 337 (0.24)
Genotype and allele counts are shown for each SNP. Allele frequencies are shown in brackets. Minor alleles are coded as 2, using coding strand format as per Egan et al., [6]. For comparison between studies, rs2228595 was identified by Marti et al., [8] as +1131C>T and is referred to as snp5 in Egan et al., [6]. rs13242038, rs6465084 and rs7804100 are referred to as hCV2627921, hCV11245618 and hCV2536213 respectively in Egan et al., [6].
Table 2 GRM3 haplotype data
SNP Study Haplotype
rs187993 2
rs13242038 1 1 1
rs917071 1 1 1 1
rs6465084 1 1 1 1
rs1468412 1 1 1
rs7804100 1 1
Global P valuea Egan et al 2004 0.23 0.05 0.01 0.01 0.01
Cardiff 0.52 0.80 0.79 0.88 0.82
Best p value b Egan et al 2004 0.01 0.01 0.01 0.00 0.00
Cardiff 0.09 0.58 0.60 0.60 0.98
Frequencyc Egan et al 2004 0.29 0.57 0.62 0.61 0.53
Cardiff cases 0.29 0.62 0.66 0.64 0.58
Cardiff controls 0.32 0.63 0.67 0.65 0.58
Comparison of haplotypes for Cardiff case-control sample and Egan et al., [6] Caucasian sample. a global p-value, b best p-value in Egan et al., [6] compared with same haplotype in Cardiff sample, c Frequency of significant haplotype. In the Cardiff analysis, the no. of controls from 623 to 693 and the no. of cases ranges from 582 to 652.
Allele frequencies and linkage disequilibrium data for the CEPH families were similar to those obtained for the cases and the controls. The LD relationships between markers are given in table 3.
Table 3 GRM3 linkage disequilibrium data
rs187993 rs13242038 rs917071 rs6465084 rs2228595 rs1468412 rs7804100
rs187993 x 1 0.94 0.67 0.96 0.57 0.35
rs13242038 0.11 x 0.50 0.38 1 0.32 0.48
rs917071 0.16 0.16 x 0.78 0.91 0.71 0.46
rs6465084 0.07 0.10 0.54 x 0.96 0.86 0.45
rs2228595 0.04 0.18 0.18 0.23 x 0.97 0.60
rs1468412 0.57 0.32 0.50 0.66 0.21 x 0.69
rs7804100 0.35 0.16 0.18 0.20 0.01 0.41 x
LD data for case control sample. D' is shown above the diagonal and r2 is shown below the diagonal.
Discussion
We sought to replicate specific findings of association between GRM3 and schizophrenia in European and European-American samples using a large, UK based case-control sample. However, our data reveal no support for association between any single markers or haplotypes that would represent either replication or indeed novel findings of association. With respect to the specific markers of interest from previous work, marker rs2228595 [5] showed an almost identical allele frequency in cases and controls, with a slight excess (less than 0.5%) of the minor T allele in cases. Thus our data are congruent with the overall combined data reported in the work of Marti and colleagues [5] and also the follow up study of Egan and colleagues [7], and do not support the hypothesis that this allele is associated with schizophrenia. Like Egan and colleagues [7], and unsurprisingly given the postulated ethnic heterogeneity at this locus, we could also not replicate the association between schizophrenia and rs1468412 previously reported in the Japanese sample [6]. However, in contrast to the sibs sample of Egan and colleagues [7], but similar to their family sample, we found a slight excess (0.6%) of the major allele of rs6465084 in controls rather than in cases. Our haplotype analyses were no more supportive of the findings of Egan and colleagues than our single locus analyses. The only specific haplotype reported as associated in that study that approached significance in our sample (nominal p = 0.09, 1df) was the 211 haplotype constructed from rs187993, rs13242038 and rs917071. However, in contrast to the earlier work, we found this to be more common in controls than cases.
There are numerous reasons why in the face of a true association in different samples, different markers and haplotypes might be associated with a disease [15] and therefore it should not in itself be a major concern that the data concerning GRM3 follow such a pattern. However, we should nevertheless remain vigilant to the onus in science being rejection of the null hypothesis, and, in the face of confusing patterns of association which may be explicable by heterogeneity, should have explicit criteria for what constitutes replication. A commonsense approach is that exact replication of an allele or haplotype under conventional levels of significance (p < 0.05) can be regarded as replication, with the proviso that if multiple specific hypotheses are tested, this is allowed for. No two studies of GRM3 meet this particular criterion for replication. In our own study, we anticipated that rs6465084 would be the most likely marker to replicate in our sample. Although only associated with affected status in one of two Caucasian samples in the study of Egan and colleagues, [7] this marker independently showed evidence for association with measures of glutamate transporter gene expression in post-mortem samples and with neuro-imaging based indices of functionality in controls. This suggested to us that if the findings were not due to chance, they might generalize beyond the specific group of cases reported in that study. Unfortunately, our study provides no support for that prediction.
Alternatively, in the presence of heterogeneity, any study reporting association to novel markers or haplotypes can be considered to provide significant replication provided the analysis allows appropriately (which is not the same as Bonferroni correction) for multiple testing. Given the numbers of markers it is now possible to test in a gene, failure to require this will inevitably lead to accumulation of studies with nominally significant evidence in support of certain markers or haplotypes. It is unclear that any two of the above studies can be considered to provide convincing 'stand alone' evidence for association that would meet this criterion; ours certainly does not. Thus even in the absence of our own data, we do not concur with the suggestion [16] that the evidence for GRM3 meets the criteria for significant evidence for association proposed in a meta-analysis of association studies with common variants [17]. However, we do concur with sentiments of Lohmueller et al., [17] whose paper was published "in the light of the seemingly high proportion of false positives in the literature" and who urged the publication of large studies and negative results.
While the current data concerning GRM3 do not allow rejection of the null hypothesis, they are insufficient to rule out the possibility of GRM3 as a true susceptibility gene for schizophrenia. GRM3 spans 221 kb and across this region, the level of linkage disequilibrium across GRM3 is moderate (table 3). The current crop of data from the Hapmap includes 63 informative SNPs of which only 28 are redundant using one of the commonly used methods [18]. Thus it does not appear that any of the previous studies, including our own, extract a high proportion of the genetic information relevant to GRM3. Whether the existing data suggest that such a study is of sufficient priority to justify embarking upon what is still for most, including ourselves, a costly and time consuming endeavour is clearly a matter for individual researchers. In our view it does not. However, future replication attempts would be greatly facilitated if this challenge were to be taken up by one or more of the groups with data that they themselves consider to be convincing.
Conclusion
In a bid to attempt to provide support for GRM3 as a susceptibility gene for schizophrenia, we have genotyped those polymorphisms that have provided significant evidence in Caucasian samples in a large UK based case control sample. Our data do not support the hypothesis that variation in GRM3 is associated with schizophrenia. However, given the moderate level of linkage disequilibrium across this gene, both this study and other current data are insufficient to rule out the possibility that GRM3 is a true susceptibility gene for schizophrenia. This will require studies which extract a higher proportion of the genetic data from GRM3.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
NN, HW, SD, DI, ACP, AG performed laboratory assays. NN performed the data-analysis and drafted the manuscript. NW participated in the design of the study and its coordination. PY performed phenotypic diagnosis. SZ performed sample collection and phenotypic diagnosis. MOD and MJO participated in the design of the study, interpretation of the data, and drafting of the manuscript. All authors read and approved the final manuscript.
Figure 1 Schematic of GRM3 structure with locations of SNPs genotyped in this study. Exons are represented as black boxes. Larger boxing represents coding sequence, and smaller represents UTR. Connecting black lines represent intronic sequence. rs2299225, (significant in Chen et al., [8]) maps to intron 3, but was not typed in this study.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by the MRC (UK).
==== Refs
Kim JS Kornhuber HH Schmid-Burgk W Holzmuller B Low cerebrospinal fluid glutamate in schizophrenic patients and a new hypothesis on schizophrenia Neurosci Lett 1980 20 379 382 6108541 10.1016/0304-3940(80)90178-0
Javitt DC Zukin SR Recent advances in the phencyclidine model of schizophrenia Am J Psychiatry 1991 148 1301 1308 1654746
Goff DC Wine L Glutamate in schizophrenia: clinical and research implications Schizophr Res 1997 27 157 168 9416645 10.1016/S0920-9964(97)00079-0
Moghaddam B Adams BW Reversal of phencyclidine effects by a group II metabotropic glutamate receptor agonist in rats Science 1998 281 1349 1352 9721099 10.1126/science.281.5381.1349
Marti SB Cichon S Propping P Nothen M Metabotropic glutamate receptor 3 (GRM3) gene variation is not associated with schizophrenia or bipolar affective disorder in the German population Am J Med Genet 2002 114 46 50 11840505 10.1002/ajmg.1624
Fujii Y Shibata H Kikuta R Makino C Tani A Hirata N Shibata A Ninomiya H Tashiro N Fukumaki Y Positive associations of polymorphisms in the metabotropic glutamate receptor type 3 gene (GRM3) with schizophrenia Psychiatr Genet 2003 13 71 76 12782962 10.1097/00041444-200306000-00003
Egan MF Straub RE Goldberg TE Yakub I Callicott JH Hariri AR Mattay VS Bertolino A Hyde TM Shannon-Weickert C Akil M Crook J Vakkalanka RK Balkissoon R Gibbs RA Kleinman JE Weinberger DR Variation in GRM3 affects cognition, prefrontal glutamate, and risk for schizophrenia Proc Natl Acad Sci USA 2004 101 12604 12609 15310849 10.1073/pnas.0405077101
Chen Q He G Chen Q Wu S Xu Y Feng G Li Y Wang L He L A case-control study of the relationship between the metabotropic glutamate receptor 3 gene and schizophrenia in the Chinese population Schizophr Res 2005 73 21 26 15567072 10.1016/j.schres.2004.07.002
Wing JK Cooper JE Satorius N The Measurement and Classification of Psychiatric illness 1974 Cambridge, UK: Cambridge University Press
Wing JK Babor T Brugha T SCAN: Schedules for the clinical assessment in neuropsychiatry Archives of General Psychiatry 1990 47 589 593 2190539
Owen MJ Holmans P McGuffin P Association studies in psychiatric genetics Mol Psychiatry 1997 2 270 273 9246659 10.1038/sj.mp.4000292
International Hapmap Project
Zhao JH Curtis D Sham PC Model-free analysis and permutation tests for allelic associations Hum Hered 2000 50 133 139 10799972 10.1159/000022901
Abecasis GR Cookson WO GOLD-graphical overview of linkage disequilibrium Bioinformatics 2000 16 182 183 10842743 10.1093/bioinformatics/16.2.182
Cardon LR Bell JI Association study designs for complex diseases Nat Rev Genet 2001 2 91 99 11253062 10.1038/35052543
Harrison PJ Weinberger DR Schizophrenia genes, gene expression, and neuropathology: on the matter of their convergence Mol Psychiatry 2005 10 40 68 15263907 10.1038/sj.mp.4001558
Lohmueller KE Pearce CL Pike M Lander ES Hirschhorn JN Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease Nat Genet 2003 33 177 182 12524541 10.1038/ng1071
Gabriel SB Schaffner SF Nguyen H Moore JM Roy J Blumenstiel B Higgins J DeFelice M Lochner A Faggart M Liu-Cordero SN Rotimi C Adeyemo A Cooper R Ward R Lander ES Daly MJ Altshuler D The structure of haplotype blocks in the human genome Science 2002 296 2225 2229 12029063 10.1126/science.1069424
| 15892884 | PMC1142331 | CC BY | 2021-01-04 16:33:02 | no | BMC Psychiatry. 2005 May 13; 5:23 | utf-8 | BMC Psychiatry | 2,005 | 10.1186/1471-244X-5-23 | oa_comm |
==== Front
Cancer Cell IntCancer Cell International1475-2867BioMed Central London 1475-2867-5-131588513610.1186/1475-2867-5-13Primary ResearchMorphological characterization of a human glioma cell l ine Machado Camila ML [email protected] André [email protected] José [email protected] Wirla MSC [email protected]çalves Estela M [email protected] Selma C [email protected] Liana [email protected] Department of Microbiology and Immunology, Biology Institute, State University of Campinas, Campinas, São Paulo, Brazil2 Department of Pathology, School of Medicine, State University of Campinas, Campinas, São Paulo, Brazil3 Department of Cellular Biology, Biology Institute, State University of Campinas, Campinas, São Paulo, Brazil2005 10 5 2005 5 13 13 24 6 2004 10 5 2005 Copyright © 2005 Machado et al; licensee BioMed Central Ltd.2005Machado 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 human malignant continuous cell line, named NG97, was recently established in our laboratory. This cell line has been serially subcultured over 100 times in standard culture media presenting no sign of cell senescence. The NG97 cell line has a doubling time of about 24 h. Immunocytochemical analysis of glial markers demonstrated that cells are positive for glial fibrillary acidic protein (GFAP) and S-100 protein, and negative for vimentin. Under phase-contrast microscope, cultures of NG97 showed cells with variable morphological features, such as small rounded cells, fusiform cells (fibroblastic-like cells), and dendritic-like cells. However, at confluence just small rounded and fusiform cells can be observed. At scanning electron microscopy (SEM) small rounded cells showed heterogeneous microextentions, including blebs and filopodia. Dendritic-like cells were flat and presented extensive prolongations, making several contacts with small rounded cells, while fusiform cells presented their surfaces dominated by microvilli.
We believe that the knowledge about NG97 cell line may be useful for a deeper understanding of biological and immunological characteristics of gliomas.
==== Body
Background
Malignant gliomas are the most common type of brain tumor in adults. These tumors are highly invasive and despite multi-modality treatment the mean survival of patients is still less than 1 year.
Cultures of malignant cells represent an excellent and permanent material for studying the biology of these tumors as, for example, specific antigens characterization, bioactive factors produced, determination of cellular proliferation, and heterogeneity of genotypic and phenotypic characteristics (Pohl et al. 1999; Tsujino et al. 1997; Bodmer et al. 1989; Di Tomaso et al. 2000; Halfter et al. 1998; Bigner et al. 1981).
Recently, we have established a human glioma cell line from tissue obtained from a patient diagnosed with glioblastoma multiforme of the right temporal lobe. Histological examination revealed a grade III astrocytoma according to the WHO classification. This cell line, called NG97, has been sub-cultured in standard culture media without feeder layer or collagen coatings. The injection of NG97 cells into congenitally athymic mice induce the formation of solid tumor masses that can be retransplanted every 4 weeks. These tumors present features of malignant gliomas characterized by cell pleomorphism, necrosis and aggressive growth (Grippo et al. 2001).
The present work was undertaken to study growth kinetics, expression of marker proteins and morphological characteristics of early passaged cells present in the NG97 cell line.
Results
Markers
Immunocytochemical analysis of glial markers in the NG97 cells demonstrated that a large number of cells were positive for GFAP and S100 protein (Figure 1A and 1B, respectively). GFAP presents a diffuse perinuclear condensation, and S-100 protein is uniformly observed in the cytoplasm and irregularly observed in the nucleus of some cells. On the other hand, vimentin was undetectable in this cell line (Figure 1C). Figure 1D shows a representative control of all immunocytochemical experiments.
Figure 1 Immunocytochemical staining in NG97 cell line at passage 15. Note the immunopositive staining for GFAP (A) and S100 protein (B). On the other hand, cells were vimentin-negative stained (C). A negative control is also showed (D). Scale bar = 25 μm
Microscopy Studies
Initially, NG97 cells formed mainly floating aggregates in the culture flasks and only small, rounded cells were seen (Figure 2A). At the 13th passage dendritic-like cells appear in the culture (Figure 2B). These cells present extensive prolongations making several contacts with small rounded cells and showed extra numerary nucleous (Figure 2C). As the cultures became dense, a third cellular type appears presenting a fusiform morphology (fibroblastic-like cells). At confluence, just small and fusiform cells can be observed in the culture (Figure 2D).
Figure 2 Phase contrast micrographs of NG97 cells. (A) small rounded cells growing as floating aggregates; (B) dendritic-like cells appear in the culture (→); (C) a dendritic-like cell with an extensive cytoplasmatic prolongation (→) and extra numerary nucleous (▶); (D) confluent monolayer with small rounded and fibroblastic-like cells. Scale bar = 50 μm (A and D); 25 μm (B and C).
Scanning electron microscopy of small rounded cells showed heterogeneity of cytoplasmatic prolongation, including blebs and filopodia (Figure 3A and 3B). Dendritic-like cells are illustrated in Figures 3C through 3F. These cells presented high degree of cellular flattening, absence of blebs and, numerous and extensive cytoplasmatic prolongations. They were attached to the substrate making contact with small rounded cells. The third morphologically distinct cell type is presented in Figures 3G and 3H. These fusiform cells presented numerous microvilli on surfaces.
Figure 3 Scanning electron micrograph of NG97 cell line. A, B: small rounded cells presenting blebs (Bl) and filopodia (Fi) on their surfaces; C, D, E: dendritic-like cells with extensive cytoplasmatic prolongations. The area in the rectangle is shown at higher magnification in F; G: culture with two morphologic distinct cellular types; H: fibroblastic-like cells presenting microvilli (Mi) on the membrane surface.
Growth kinetics
Until the 13th passage, when just small rounded cells were seen in the culture, a slow growth rate was observed (data not shown). At 13th passage, when the two other cell types appeared in the culture, the cells entered into an exponential growth phase. The population doubling time of NG97 cell line was about 24 h at 37°C and the saturation cell density was reached at 10 × 105 cells/cm2 (Figure 4). The high growth rate was observed for the successive passages
Figure 4 Growth curve of NG97 cell line.
Discussion
In this study basic characteristics of NG97 cell line are described. The investigated cell line was within passage 13 to 15. Our results show that NG97 cell line retains the expression of GFAP, which is a reliable marker of astrocytic cells, and S100 protein that was originally identified as brain specific (Moore, 1965). Literature has shown, by and large, a negative correlation between the degree of malignancy and expression of GFAP and S100 protein in the majority of human gliomas (Jacques et al., 1981; Duffy et al., 1982). However, NG97 cells are tumorigenic in nude mice, indicating that cells are neoplastic and malignant. Besides, these two biochemical markers are present in the xenografts of NG97 cells in nude mice (Grippo et al., 2001). Interestingly, the vimentin that has been identified in some human glioma cell lines (Roessmann et al. 1983; Rutka et al. 1998) was not detectable in NG97 cells. Hedberg and Chen (1986) found that a human adrenal tumor cell line, named SW-13, expressed vimentin filaments and clones derived from these cells were characterized as lacking any detectable cytoplasmic intermediate filaments (vim-). Later, Sarria et al. (1994) demonstrated that the nuclei of the SW-13 vim- cells often appeared to be highly folded, forming prominent lobes and clefts. However, the authors also showed that the effect of vimentin filaments on the invaginations or folding in the nucleus is not an absolute, and raise the possibility that this nuclear configuration could be an indirect effect of a metabolic difference between cells that contain or lack organized vimentin filaments. To all appearances, our results indicate that in NG97 cell line the absence of an organized vimentin filament network does not affect the shape of the nucleus.
Heterogeneous cell types can be found in NG97 cultures. At early passages, cultures grew slower and presented only the small, rounded cells. At 13th passage, dendritic-like cells appear in the culture. It is not clearly for us the exact events that lead to the appearance of this cell type in the culture. We hypothesize that one small, rounded cell accumulates unbalanced divisions forming an extra numerary nucleous cell that secretes some products capable to induce alterations on the other cells. More elaborate experiments would test this possibility. In addition, dendritic-like cells present numerous and extensive cytoplasmatic prolongations, which may be associated with communication between this cell and the small ones. It seems also that dendritic-like cells provide an anchorage to the small rounded cells, which in turn present an increase in the filopodias to ameliorate the substrate connections. Of note, when dendritic-like cells appear in the culture we have noted an increase of the cellular growth rate. Future analyses should test if dendritic-like cells are able to modulate cell growth.
Fusiform cells appear when the culture becomes dense. These cells are majority in confluent monolayer cultures and present a large number of microvilli on the surface that propitiates an intimate contact with the environment. In the same way, further studies of this cell will help to unveil more NG97 cell line secrets.
Conclusion
NG97 cells grow in vitro as three sub populations with distinct morphological appearance and, undoubtedly, constitute a glial-committed cell line since are positive for GFAP and S-100 protein. Until 13th passage only small rounded cells were seen in culture and the growth kinetics was very slow. From this point, two other cell types presenting dendritic and fibroblastic characteristics could be observed and results were evident for overgrowth of cells. The possibility that these cells are able to modulate cell growth can not be discarded and are now under investigation in our laboratory.
This cell line may prove useful for cellular and molecular studies as well as in studies of gliomas treatment.
Methods
Glioma Culture
NG97 cells were grown in plastic flasks (25 cm2) with RPMI 1640 medium (Sigma Chemical Co., St Louis, MO), supplemented with 50 μM 2-ME, 2 mM L- glutamine, 100 μg/mL garamycin and 20% inactivated fetal bovine serum (complete medium). The cultures were incubated at 37°C in an atmosphere containing 95% air and 5% CO2. The medium was changed after intervals of 48 hs and when the culture reached confluence, the subculture was performed by treatment with 0.05% trypsin and 0.02% ethylenediaminetetraacetic acid (EDTA).
Immunocytochemistry
Immunocytochemical analysis of glial markers (GFAP, vimentin and S-100 protein) was performed by using specific antibodies purchased from Dako Envision+ Systems/HRP (Dako Corporation, Carpinteria, CA). Briefly, cultured NG97 cells were harvested (at passage 15), washed using low speed centrifugation (150 × g, 10 minutes) and ressuspended in complete medium. Then, cells were cyto-centrifugated on glass slides, dried at room temperature for 15 minutes and fixed in cold acetone for 15 minutes at -20°C. After a thorough wash with 0.5% BSA in PBS, the cells were treated with polyclonal rabbit anti-GFAP, monoclonal anti-vimentin and polyclonal rabbit anti-S100 antibodies according to the manufactures' instructions. The bound primary antibody was detected using peroxidase labeled polymer conjugated to either mouse or rabbit secondary antibodies. Subsequently, the slides were incubated with a substrate mixture of 3,3-diaminobenzidine (DAB) and 0.02% H2O2. Cells were then counterstained with haematoxilin and eosin (HE). Control slides that stain positively for the specific antigens were used to assure correct staining and stability of reagents used. Negative controls included the omission of the primary antibody.
Phase Contrast Microscopy
Growing cells on cover slips were observed with a phase-contrast microscope (Olympus IX50 with a PMC35Dx photo micrographic system).
Scanning Electron Microscopy
NG97 cells were grown to sub confluence on 13 mm round cover slip in complete medium. The cells were fixed with 2.5% glutaraldehyde and 4% paraformaldehyde in phosphate buffer (pH 7.4) for 1 hour at room temperature. Then, the cells were post-fixed in 1.0% osmium tetroxide (OsO4) for 10 minutes, washed in 0.1 M phosphate buffer (pH 7.2) and dehydrated in a grade series of ethanol. Cover slips were critically point dried using liquid CO2 as transition fluid. The specimens were cold sputter coated with gold and observed in a JEOL JMS 5800 LV scanning electron microscope (SEM) accelerating voltage of 10 kV.
Growth Curve
NG97 cells were collected from 13th passage for determination of growth curves. Briefly, semi confluent cultures were trypsinized and cells resuspended in complete medium for counting. Cells (1 × 104) were plated into each well of a 12-well plate and counts from triplicate wells were made daily for 10 days. Trypsinized cells were counted in hemacytometer chamber and numbers were averaged for each time interval. Cell population doubling time was calculated from the linear phase of the growth curve, and the saturation density was the plateau point on the growth curve after the linear growth phase.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
This work is part of a Master's Dissertation by Camila M.L. Machado
Acknowledgements
The authors wish to thank Mrs. Rosemeire F. O. de Paula and Ms. Dirce L. Gabriel for excellent technical assistance. This work was supported in part by grants from the Fundo de Apoio ao Ensino e Pesquisa da Unicamp (FAEP/UNICAMP; #210/03).
==== Refs
Pohl U Wick W Weissenberger J Steinbach JP Dichgans J Aguzzi A Weller M Characterization of Tu- a glioma cell line derived from a spontaneous tumor in GFAP-v-src-transgenic mice: comparison with established murine glioma cell lines Int J Oncol 1999 15 829 834 10493969
Tsujino K Yamate J Tsukamoto Y Kumagai D Kannan Y Jippo T Kuwamura M Kotani T Takeya M Sakuma S Establishment and Characterization of cell lines derived from a transplantable rat malignant meningioma: morphological heterogeneity and production of nerve growth factor Acta Neuropathol 1997 93 461 470 9144584 10.1007/s004010050640
Bodmer S Strommer K Frei K Siepl C de Tribolet N Heid I Fontana A Immunosuppression and transforming growth factor-β in Glioblastoma J Immunol 1989 143 3222 3229 2809198
Di Tomaso E Pang JCS Ng HK Lam PYP Tian XX Suen KW Hui ABY Hjelm NM Establishment and characterization of a human cell line from paedriatic cerebellar glioblastoma multiform Neuropathol Appl Neurobiol 2000 26 22 30 10736064 10.1046/j.1365-2990.2000.00214.x
Halfter H Kremerskothen J Webwe J Hacker-Klom U Barnekow A Ringlestein EB Stogbauer F Growth inhibition of newly established human glioma cell lines by leukemia inhibitory factor J Neuro-Oncol 1998 39 1 18 10.1023/A:1005901423332
Bigner DD Bigner SH Pontèn J Westermark B Mahaley MS Ruoslahti E Herschman H Eng LF Wikstrand CJ Heterogeneity of genotypic and phenotypic characteristics of fifteen permanent cell lines derived from human gliomas J Neuropathol Exp Neurol 1981 40 201 229 6260907
Grippo M Penteado PF Carelli EF Cruz-Höfling MA Verinaud L Establishment and Partial Characterization of a Continuous Human Malignant Glioma Cell Line: NG97 Cell Mol Neurobiol 2001 21 421 428 11775071 10.1023/A:1012662423863
Moore BW A soluble protein characteristic of the nervous system. Biochem Biophys Res Commun 1965 19 739 744 10.1016/0006-291X(65)90320-7
Jacques CM Turpin JC Poreau A GFAP and S-100 protein levels as an index for malignancy in human gliomas and neurinomas J Natl Cancer Inst 1981 62 479 483
Duffy PE Huang YY Rapport MM The relationship of GFAP to the shape, motility and differentiation of human astrocytoma cells Exp Cell Res 1982 139 145 157 7200901 10.1016/0014-4827(82)90328-7
Roessmann U Velasco ME Gambetti P Autilio-Gambetti L Neuronal and astrocytic differentiation in human neuroepithelial neoplasms. An immunohistochemical study J Neuropathol Exp Neurol 1983 42 113 21 6338157
Rutka JT Ackerley C Hubbard SL Tilup A Dirks PB Jung S Ivanchuk S Kurimoto M Tsugu A Becker LE Characterization of glial filament-cytoskeletal interactions in human astrocytomas: an immuno-ultrastructural analysis Eur J Cell Biol 1998 76 279 287 9765058
Hedberg KK Chen LB Absence of intermediate filaments in a human adrenal cortex carcinoma derived cell line Exp Cell Res 1986 163 509 517 3956586 10.1016/0014-4827(86)90081-9
Sarria AJ Lieber JG Nordeen SK Evans RM The presence or absence of a vimentin-type intermediate filament network affects the shape of the nucleus in human SW-13 cells J Cell Science 1994 107 1593 1607 7962200
| 15885136 | PMC1142332 | CC BY | 2021-01-04 16:24:17 | no | Cancer Cell Int. 2005 May 10; 5:13 | utf-8 | Cancer Cell Int | 2,005 | 10.1186/1475-2867-5-13 | oa_comm |
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-61588247210.1186/1476-069X-4-6ReviewThe Bhopal disaster and its aftermath: a review Broughton Edward [email protected] Columbia University, Mailman School of Public Health, 600 W 168th St. New York, NY 10032 USA2005 10 5 2005 4 6 6 21 12 2004 10 5 2005 Copyright © 2005 Broughton; licensee BioMed Central Ltd.2005Broughton; 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.
On December 3 1984, more than 40 tons of methyl isocyanate gas leaked from a pesticide plant in Bhopal, India, immediately killing at least 3,800 people and causing significant morbidity and premature death for many thousands more. The company involved in what became the worst industrial accident in history immediately tried to dissociate itself from legal responsibility. Eventually it reached a settlement with the Indian Government through mediation of that country's Supreme Court and accepted moral responsibility. It paid $470 million in compensation, a relatively small amount of based on significant underestimations of the long-term health consequences of exposure and the number of people exposed. The disaster indicated a need for enforceable international standards for environmental safety, preventative strategies to avoid similar accidents and industrial disaster preparedness.
Since the disaster, India has experienced rapid industrialization. While some positive changes in government policy and behavior of a few industries have taken place, major threats to the environment from rapid and poorly regulated industrial growth remain. Widespread environmental degradation with significant adverse human health consequences continues to occur throughout India.
==== Body
December 2004 marked the twentieth anniversary of the massive toxic gas leak from Union Carbide Corporation's chemical plant in Bhopal in the state of Madhya Pradesh, India that killed more than 3,800 people. This review examines the health effects of exposure to the disaster, the legal response, the lessons learned and whether or not these are put into practice in India in terms of industrial development, environmental management and public health.
History
In the 1970s, the Indian government initiated policies to encourage foreign companies to invest in local industry. Union Carbide Corporation (UCC) was asked to build a plant for the manufacture of Sevin, a pesticide commonly used throughout Asia. As part of the deal, India's government insisted that a significant percentage of the investment come from local shareholders. The government itself had a 22% stake in the company's subsidiary, Union Carbide India Limited (UCIL) [1]. The company built the plant in Bhopal because of its central location and access to transport infrastructure. The specific site within the city was zoned for light industrial and commercial use, not for hazardous industry. The plant was initially approved only for formulation of pesticides from component chemicals, such as MIC imported from the parent company, in relatively small quantities. However, pressure from competition in the chemical industry led UCIL to implement "backward integration" – the manufacture of raw materials and intermediate products for formulation of the final product within one facility. This was inherently a more sophisticated and hazardous process [2].
In 1984, the plant was manufacturing Sevin at one quarter of its production capacity due to decreased demand for pesticides. Widespread crop failures and famine on the subcontinent in the 1980s led to increased indebtedness and decreased capital for farmers to invest in pesticides. Local managers were directed to close the plant and prepare it for sale in July 1984 due to decreased profitability [3]. When no ready buyer was found, UCIL made plans to dismantle key production units of the facility for shipment to another developing country. In the meantime, the facility continued to operate with safety equipment and procedures far below the standards found in its sister plant in Institute, West Virginia. The local government was aware of safety problems but was reticent to place heavy industrial safety and pollution control burdens on the struggling industry because it feared the economic effects of the loss of such a large employer [3].
At 11.00 PM on December 2 1984, while most of the one million residents of Bhopal slept, an operator at the plant noticed a small leak of methyl isocyanate (MIC) gas and increasing pressure inside a storage tank. The vent-gas scrubber, a safety device designer to neutralize toxic discharge from the MIC system, had been turned off three weeks prior [3]. Apparently a faulty valve had allowed one ton of water for cleaning internal pipes to mix with forty tons of MIC [1]. A 30 ton refrigeration unit that normally served as a safety component to cool the MIC storage tank had been drained of its coolant for use in another part of the plant [3]. Pressure and heat from the vigorous exothermic reaction in the tank continued to build. The gas flare safety system was out of action and had been for three months. At around 1.00 AM, December 3, loud rumbling reverberated around the plant as a safety valve gave way sending a plume of MIC gas into the early morning air [4]. Within hours, the streets of Bhopal were littered with human corpses and the carcasses of buffaloes, cows, dogs and birds. An estimated 3,800 people died immediately, mostly in the poor slum colony adjacent to the UCC plant [1,5]. Local hospitals were soon overwhelmed with the injured, a crisis further compounded by a lack of knowledge of exactly what gas was involved and what its effects were [1]. It became one of the worst chemical disasters in history and the name Bhopal became synonymous with industrial catastrophe [5].
Estimates of the number of people killed in the first few days by the plume from the UCC plant run as high as 10,000, with 15,000 to 20,000 premature deaths reportedly occurring in the subsequent two decades [6]. The Indian government reported that more than half a million people were exposed to the gas [7]. Several epidemiological studies conducted soon after the accident showed significant morbidity and increased mortality in the exposed population. Table 1. summarizes early and late effects on health. These data are likely to under-represent the true extent of adverse health effects because many exposed individuals left Bhopal immediately following the disaster never to return and were therefore lost to follow-up [8].
Table 1 Health effects of the Bhopal methyl isocyanate gas leak exposure [8, 30-32].
Early effects (0–6 months)
Ocular Chemosis, redness, watering, ulcers, photophobia
Respiratory Distress, pulmonary edema, pneumonitis, pneumothorax.
Gastrointestinal Persistent diarrhea, anorexia, persistent abdominal pain.
Genetic Increased chromosomal abnormalities.
Psychological Neuroses, anxiety states, adjustment reactions
Neurobehavioral Impaired audio and visual memory, impaired vigilance attention and response time, Impaired reasoning and spatial ability, impaired psychomotor coordination.
Late effects (6 months onwards)
Ocular Persistent watering, corneal opacities, chronic conjunctivitis
Respiratory Obstructive and restrictive airway disease, decreased lung function.
Reproductive Increased pregnancy loss, increased infant mortality, decreased placental/fetal weight
Genetic Increased chromosomal abnormalities
Neurobehavioral Impaired associate learning, motor speed, precision
Aftermath
Immediately after the disaster, UCC began attempts to dissociate itself from responsibility for the gas leak. Its principal tactic was to shift culpability to UCIL, stating the plant was wholly built and operated by the Indian subsidiary. It also fabricated scenarios involving sabotage by previously unknown Sikh extremist groups and disgruntled employees but this theory was impugned by numerous independent sources [1].
The toxic plume had barely cleared when, on December 7, the first multi-billion dollar lawsuit was filed by an American attorney in a U.S. court. This was the beginning of years of legal machinations in which the ethical implications of the tragedy and its affect on Bhopal's people were largely ignored. In March 1985, the Indian government enacted the Bhopal Gas Leak Disaster Act as a way of ensuring that claims arising from the accident would be dealt with speedily and equitably. The Act made the government the sole representative of the victims in legal proceedings both within and outside India. Eventually all cases were taken out of the U.S. legal system under the ruling of the presiding American judge and placed entirely under Indian jurisdiction much to the detriment of the injured parties.
In a settlement mediated by the Indian Supreme Court, UCC accepted moral responsibility and agreed to pay $470 million to the Indian government to be distributed to claimants as a full and final settlement. The figure was partly based on the disputed claim that only 3000 people died and 102,000 suffered permanent disabilities [9]. Upon announcing this settlement, shares of UCC rose $2 per share or 7% in value [1]. Had compensation in Bhopal been paid at the same rate that asbestosis victims where being awarded in US courts by defendant including UCC – which mined asbestos from 1963 to 1985 – the liability would have been greater than the $10 billion the company was worth and insured for in 1984 [10]. By the end of October 2003, according to the Bhopal Gas Tragedy Relief and Rehabilitation Department, compensation had been awarded to 554,895 people for injuries received and 15,310 survivors of those killed. The average amount to families of the dead was $2,200 [9].
At every turn, UCC has attempted to manipulate, obfuscate and withhold scientific data to the detriment of victims. Even to this date, the company has not stated exactly what was in the toxic cloud that enveloped the city on that December night [8]. When MIC is exposed to 200° heat, it forms degraded MIC that contains the more deadly hydrogen cyanide (HCN). There was clear evidence that the storage tank temperature did reach this level in the disaster. The cherry-red color of blood and viscera of some victims were characteristic of acute cyanide poisoning [11]. Moreover, many responded well to administration of sodium thiosulfate, an effective therapy for cyanide poisoning but not MIC exposure [11]. UCC initially recommended use of sodium thiosulfate but withdrew the statement later prompting suggestions that it attempted to cover up evidence of HCN in the gas leak. The presence of HCN was vigorously denied by UCC and was a point of conjecture among researchers [8,11-13].
As further insult, UCC discontinued operation at its Bhopal plant following the disaster but failed to clean up the industrial site completely. The plant continues to leak several toxic chemicals and heavy metals that have found their way into local aquifers. Dangerously contaminated water has now been added to the legacy left by the company for the people of Bhopal [1,14].
Lessons learned
The events in Bhopal revealed that expanding industrialization in developing countries without concurrent evolution in safety regulations could have catastrophic consequences [4]. The disaster demonstrated that seemingly local problems of industrial hazards and toxic contamination are often tied to global market dynamics. UCC's Sevin production plant was built in Madhya Pradesh not to avoid environmental regulations in the U.S. but to exploit the large and growing Indian pesticide market. However the manner in which the project was executed suggests the existence of a double standard for multinational corporations operating in developing countries [1]. Enforceable uniform international operating regulations for hazardous industries would have provided a mechanism for significantly improved in safety in Bhopal. Even without enforcement, international standards could provide norms for measuring performance of individual companies engaged in hazardous activities such as the manufacture of pesticides and other toxic chemicals in India [15]. National governments and international agencies should focus on widely applicable techniques for corporate responsibility and accident prevention as much in the developing world context as in advanced industrial nations [16]. Specifically, prevention should include risk reduction in plant location and design and safety legislation [17].
Local governments clearly cannot allow industrial facilities to be situated within urban areas, regardless of the evolution of land use over time. Industry and government need to bring proper financial support to local communities so they can provide medical and other necessary services to reduce morbidity, mortality and material loss in the case of industrial accidents.
Public health infrastructure was very weak in Bhopal in 1984. Tap water was available for only a few hours a day and was of very poor quality. With no functioning sewage system, untreated human waste was dumped into two nearby lakes, one a source of drinking water. The city had four major hospitals but there was a shortage of physicians and hospital beds. There was also no mass casualty emergency response system in place in the city [3]. Existing public health infrastructure needs to be taken into account when hazardous industries choose sites for manufacturing plants. Future management of industrial development requires that appropriate resources be devoted to advance planning before any disaster occurs [18]. Communities that do not possess infrastructure and technical expertise to respond adequately to such industrial accidents should not be chosen as sites for hazardous industry.
Since 1984
Following the events of December 3 1984 environmental awareness and activism in India increased significantly. The Environment Protection Act was passed in 1986, creating the Ministry of Environment and Forests (MoEF) and strengthening India's commitment to the environment. Under the new act, the MoEF was given overall responsibility for administering and enforcing environmental laws and policies. It established the importance of integrating environmental strategies into all industrial development plans for the country. However, despite greater government commitment to protect public health, forests, and wildlife, policies geared to developing the country's economy have taken precedence in the last 20 years [19].
India has undergone tremendous economic growth in the two decades since the Bhopal disaster. Gross domestic product (GDP) per capita has increased from $1,000 in 1984 to $2,900 in 2004 and it continues to grow at a rate of over 8% per year [20]. Rapid industrial development has contributed greatly to economic growth but there has been significant cost in environmental degradation and increased public health risks. Since abatement efforts consume a large portion of India's GDP, MoEF faces an uphill battle as it tries to fulfill its mandate of reducing industrial pollution [19]. Heavy reliance on coal-fired power plants and poor enforcement of vehicle emission laws have result from economic concerns taking precedence over environmental protection [19].
With the industrial growth since 1984, there has been an increase in small scale industries (SSIs) that are clustered about major urban areas in India. There are generally less stringent rules for the treatment of waste produced by SSIs due to less waste generation within each individual industry. This has allowed SSIs to dispose of untreated wastewater into drainage systems that flow directly into rivers. New Delhi's Yamuna River is illustrative. Dangerously high levels of heavy metals such as lead, cobalt, cadmium, chrome, nickel and zinc have been detected in this river which is a major supply of potable water to India's capital thus posing a potential health risk to the people living there and areas downstream [21].
Land pollution due to uncontrolled disposal of industrial solid and hazardous waste is also a problem throughout India. With rapid industrialization, the generation of industrial solid and hazardous waste has increased appreciably and the environmental impact is significant [22].
India relaxed its controls on foreign investment in order to accede to WTO rules and thereby attract an increasing flow of capital. In the process, a number of environmental regulations are being rolled back as growing foreign investments continue to roll in. The Indian experience is comparable to that of a number of developing countries that are experiencing the environmental impacts of structural adjustment. Exploitation and export of natural resources has accelerated on the subcontinent. Prohibitions against locating industrial facilities in ecologically sensitive zones have been eliminated while conservation zones are being stripped of their status so that pesticide, cement and bauxite mines can be built [23]. Heavy reliance on coal-fired power plants and poor enforcement of vehicle emission laws are other consequences of economic concerns taking precedence over environmental protection [19].
In March 2001, residents of Kodaikanal in southern India caught the Anglo-Dutch company, Unilever, red-handed when they discovered a dumpsite with toxic mercury laced waste from a thermometer factory run by the company's Indian subsidiary, Hindustan Lever. The 7.4 ton stockpile of mercury-laden glass was found in torn stacks spilling onto the ground in a scrap metal yard located near a school. In the fall of 2001, steel from the ruins of the World Trade Center was exported to India apparently without first being tested for contamination from asbestos and heavy metals present in the twin tower debris. Other examples of poor environmental stewardship and economic considerations taking precedence over public health concerns abound [24].
The Bhopal disaster could have changed the nature of the chemical industry and caused a reexamination of the necessity to produce such potentially harmful products in the first place. However the lessons of acute and chronic effects of exposure to pesticides and their precursors in Bhopal has not changed agricultural practice patterns. An estimated 3 million people per year suffer the consequences of pesticide poisoning with most exposure occurring in the agricultural developing world. It is reported to be the cause of at least 22,000 deaths in India each year. In the state of Kerala, significant mortality and morbidity have been reported following exposure to Endosulfan, a toxic pesticide whose use continued for 15 years after the events of Bhopal [25].
Aggressive marketing of asbestos continues in developing countries as a result of restrictions being placed on its use in developed nations due to the well-established link between asbestos products and respiratory diseases. India has become a major consumer, using around 100,000 tons of asbestos per year, 80% of which is imported with Canada being the largest overseas supplier. Mining, production and use of asbestos in India is very loosely regulated despite the health hazards. Reports have shown morbidity and mortality from asbestos related disease will continue in India without enforcement of a ban or significantly tighter controls [26,27].
UCC has shrunk to one sixth of its size since the Bhopal disaster in an effort to restructure and divest itself. By doing so, the company avoided a hostile takeover, placed a significant portion of UCC's assets out of legal reach of the victims and gave its shareholder and top executives bountiful profits [1]. The company still operates under the ownership of Dow Chemicals and still states on its website that the Bhopal disaster was "cause by deliberate sabotage". [28].
Some positive changes were seen following the Bhopal disaster. The British chemical company, ICI, whose Indian subsidiary manufactured pesticides, increased attention to health, safety and environmental issues following the events of December 1984. The subsidiary now spends 30–40% of their capital expenditures on environmental-related projects. However, they still do not adhere to standards as strict as their parent company in the UK. [24].
The US chemical giant DuPont learned its lesson of Bhopal in a different way. The company attempted for a decade to export a nylon plant from Richmond, VA to Goa, India. In its early negotiations with the Indian government, DuPont had sought and won a remarkable clause in its investment agreement that absolved it from all liabilities in case of an accident. But the people of Goa were not willing to acquiesce while an important ecological site was cleared for a heavy polluting industry. After nearly a decade of protesting by Goa's residents, DuPont was forced to scuttle plans there. Chennai was the next proposed site for the plastics plant. The state government there made significantly greater demand on DuPont for concessions on public health and environmental protection. Eventually, these plans were also aborted due to what the company called "financial concerns". [29].
Conclusion
The tragedy of Bhopal continues to be a warning sign at once ignored and heeded. Bhopal and its aftermath were a warning that the path to industrialization, for developing countries in general and India in particular, is fraught with human, environmental and economic perils. Some moves by the Indian government, including the formation of the MoEF, have served to offer some protection of the public's health from the harmful practices of local and multinational heavy industry and grassroots organizations that have also played a part in opposing rampant development. The Indian economy is growing at a tremendous rate but at significant cost in environmental health and public safety as large and small companies throughout the subcontinent continue to pollute. Far more remains to be done for public health in the context of industrialization to show that the lessons of the countless thousands dead in Bhopal have truly been heeded.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
J. Barab, B. Castleman, R Dhara and U Misra reviewed the manuscript and provided useful suggestions.
==== Refs
Fortun K Advocacy after Bhopal 2001 Chicago , University of Chicago Press 259
Shrivastava P Managing Industrial Crisis 1987 New Delhi , Vision Books 196
Shrivastava P Bhopal: Anatomy of a Crisis 1987 Cambridge, MA , Ballinger Publishing 184
Accident Summary, Union Carbide India Ltd., Bhopal, India: December 3, 1984. Hazardous Installations Directorate 2004 Health and Safety Executive
MacKenzie D Fresh evidence on Bhopal disaster. New Scientist 2002 19
Sharma DC Bhopal: 20 Years On Lancet 2005 365 111 112 15643703 10.1016/S0140-6736(05)17722-8
Cassells J Sovereign immunity: Law in an unequal world. Social and legal studies 1996 5 431 436
Dhara VR Dhara R The Union Carbide disaster in Bhopal: a review of health effects Arch Environ Health 2002 57 391 404 12641179
Kumar S Victims of gas leak in Bhopal seek redress on compensation Bmj 2004 329 366 15310593 10.1136/bmj.329.7462.366-b
Castleman B PP Ives J Appendix: the Bhopal disaster as a case study in double standards The export of hazards: trans-national corporations and environmental control issues 1985 London , Routledge and Kegan Paul 213 222
Mangla B Long-term effects of methyl isocyanate Lancet 1989 2 103 2567843 10.1016/S0140-6736(89)90340-1
Varma DR Hydrogen cyanide and Bhopal Lancet 1989 2 567 568 2570270 10.1016/S0140-6736(89)90695-8
Anderson N Long-term effects of mthyl isocyanate Lancet 1989 2 1259 10.1016/S0140-6736(89)92347-7
Chander J Water contamination: a legacy of the union carbide disaster in Bhopal, India Int J Occup Environ Health 2001 7 72 73 11210017
Tyagi YK Rosencranz A Some international law aspects of the Bhopal disaster Soc Sci Med 1988 27 1105 1112 3059509 10.1016/0277-9536(88)90305-X
Carlsten C The Bhopal disaster: prevention should have priority now Int J Occup Environ Health 2003 9 93 94 12749644
Bertazzi PA Future prevention and handling of environmental accidents Scand J Work Environ Health 1999 25 580 588 10884157
Dhara VR What ails the Bhopal disaster investigations? (And is there a cure?) Int J Occup Environ Health 2002 8 371 379 12412856
EIA energy D India: environmental issues
CIA The world factbook: India
Rawat M Moturi MC Subramanian V Inventory compilation and distribution of heavy metals in wastewater from small-scale industrial areas of Delhi, India J Environ Monit 2003 5 906 912 14710931 10.1039/b306628b
Vijay R Sihorwala TA Identification and leaching characteristics of sludge generated from metal pickling and electroplating industries by Toxicity Characteristics Leaching Procedure (TCLP) Environ Monit Assess 2003 84 193 202 12807260 10.1023/A:1023363423345
Karliner J The corporate planet 1997 San Francisco , Sierra Club Books 247
Bruno KKJ Earthsummit,biz:The corporate takeover of sustainable development 2002 Oakland, Ca , First Food Books 237 12422966
Power M The poison stream: letter from Kerala Harper's 2004 August, 2004 51 61
Joshi TK Gupta RK Asbestos in developing countries: magnitude of risk and its practical implications Int J Occup Med Environ Health 2004 17 179 185 15212222
Joshi TK Gupta RK Asbestos-related morbidity in India Int J Occup Environ Health 2003 9 249 253 12967161
Union Carbide Bhopal Information Center wwwbhopalcom/ucshtm 2005
Corporate Watch UK DuPont: A corporate profile.
Beckett WS Persistent respiratory effects in survivors of the Bhopal disaster Thorax 1998 53 Suppl 2 S43 6 10193347
Misra UK Kalita J A study of cognitive functions in methyl-iso-cyanate victims one year after bhopal accident Neurotoxicology 1997 18 381 386 9291488
Irani SF Mahashur AA A survey of Bhopal children affected by methyl isocyanate gas J Postgrad Med 1986 32 195 198 3585792
| 15882472 | PMC1142333 | CC BY | 2021-01-04 16:36:32 | no | Environ Health. 2005 May 10; 4:6 | utf-8 | Environ Health | 2,005 | 10.1186/1476-069X-4-6 | oa_comm |
==== Front
Int J Equity HealthInternational Journal for Equity in Health1475-9276BioMed Central London 1475-9276-4-61587174410.1186/1475-9276-4-6ResearchUser fees in private non-for-profit hospitals in Uganda: a survey and intervention for equity Amone Joseph [email protected] Salome [email protected] Adriano [email protected] Annet Kakinda [email protected] Anna [email protected] Gavino [email protected] Maurice [email protected] Luca [email protected] Stefano [email protected] Uganda Catholic Medical Bureau, Kampala, Uganda2 Ugandan Martyrs University, Nkozi, Uganda3 Unit for Health Services Research and International Health, IRCCS Burlo Garofolo, Via dell'Istria 65/1, 34137 Trieste, Italy4 Cuamm – Doctors with Africa, Padova, Italy2005 4 5 2005 4 6 6 15 6 2004 4 5 2005 Copyright © 2005 Amone et al; licensee BioMed Central Ltd.2005Amone 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
In developing countries, user fees may represent an important source of revenues for private-non-for-profit hospitals, but they may also affect access, use and equity.
Methods
This survey was conducted in ten hospitals of the Uganda Catholic Medical Bureau to assess differences in user fees policies and to propose changes that would better fit with the social concern explicitly pursued by the Bureau. Through a review of relevant hospital documents and reports, and through interviews with key informants, health workers and users, hospital and non-hospital cost was calculated, as well as overall expenditure and revenues. Lower fees were applied in some pilot hospitals after the survey.
Results
The percentage of revenues from user fees varied between 6% and 89% (average 40%). Some hospitals were more successful than others in getting external aid and government subsidies. These hospitals were applying lower fees and flat rates, and were offering free essential services to encourage access, as opposed to the fee-for-service policies implemented in less successful hospitals. The wide variation in user fees among hospitals was not justified by differences in case mix. None of the hospitals had a policy for exemption of the poor; the few users that actually got exempted were not really poor. To pay hospital and non-hospital expenses, about one third of users had to borrow money or sell goods and property. The fee system applied after the survey, based on flat and lower rates, brought about an increase in access and use of hospital services.
Conclusion
Our results confirm that user fees represent an unfair mechanism of financing for health services because they exclude the poor and the sick. To mitigate this effect, flat rates and lower fees for the most vulnerable users were introduced to replace the fee-for-service system in some hospitals after the survey. The results are encouraging: hospital use, especially for pregnancy, childbirth and childhood illness, increased immediately, with no detrimental effect on overall revenues. A more equitable user fees system is possible.
==== Body
Background
Since their institution, user fees have been used in private non-for-profit (PNFP) hospitals as a way to finance provision of services. The level of cost recovery, as a percentage of recurrent expenditure, is low in government health services [1], but is thought to contribute an average of 50–60% of PNFP hospital revenues and about 90–100% of the revenues of lower level health facilities in Uganda [2]. Yet, very little is known about the structure of user fees, their predictability for users, their effects on the use of PNFP health services, and the levels of payers' compliance. In the absence of other options of community financing, such as pre-payment schemes, user fees are likely to remain the main source of financing for the provision of health services in the PNFP health sector in Uganda and other sub-Saharan countries. But the PNFP health sector operates out of social concern and explicitly pursues equity and accessibility for the poor [3]. It is therefore of absolute importance to assess current practices, their effect on patterns and trends of use, and the additional suffering all this may add on patients, in order to address and correct inequities. The aim of this study was to offer a tool for a more focused and rational structure and management of user fees in PFNP health facilities belonging to the Uganda Catholic Medical Bureau (UCMB), in an overall health system perspective (i.e. coherently with government policies) and with equity of access in mind.
Methods
The survey was conducted in a convenient sample of ten hospitals (Figure 1), chosen by UCMB using the following criteria: availability of relevant documents, geographical distribution, variety of user fee policies, willingness to support the study and to use its results for policy changes. Seven surveyors in two teams gathered data during site visits in five weeks (one week per hospital between 3 July and 13 August 2000). The surveyors gathered and reviewed hospital documents (annual and financial reports for the previous five years; logbooks from different departments, including accountants' books; planning, budgeting and other financial documents, including agreements with the Ministry of Health; price lists for drugs and other services), and interviewed managers (group interviews with medical superintendents, senior nursing officers, hospital administrators, and accountants), health workers (individual interviews with professionals in charge of outpatients, inpatients, pharmacy, laboratory, X-rays, and preventive services), and users (30 outpatients and 10 inpatients per hospital). Outpatients were supposed to be selected through systematic sampling and stratified by age and sex (ten children under 10 years, ten adult women, and ten adult men). In some hospitals, however, the number of outpatients was so low at the time of the visit that all those available were consecutively interviewed trying to maintain the proposed distribution by age and sex. Inpatients were selected among those being discharged during the visit from maternity, children and adult wards. The questionnaires had been developed in advance and were field tested and slightly modified during the first hospital visit. The interviews with users were conducted by the local surveyors in the local language.
Figure 1 The location (A to J) of the ten hospitals.
Cost is reported in US dollars ($). The exchange rate at the time of the survey was 1 $ = 1,600 Ugandan shillings. In the user survey, cost includes hospital fees and other cost that users had to meet to get health care (non-hospital cost), such as travel expenses, indirect cost (money lost because of lost work), and money already spent for the same episode of illness in other health services, shops or pharmacies. The total cost is the amount paid by users for the whole episode of illness. Total cost, however, might be underestimated due to lack of information, and because of exemptions and/or inability to pay. Hospital fees were calculated from hospital receipts, while users estimated non-hospital cost.
The preliminary results of the survey were used to identify a group of hospitals where an intervention aimed at ensuring an equitable access was considered a priority. The main intervention consisted in the progressive replacement of the fee-for-service system with a system based on flat fees covering all the services provided for a single episode of disease, at both inpatient and outpatient levels. The flat fees for maternal and child health care and for some long term conditions were on average lowered more than the fees for other conditions. The fees previously charged for preventive services (e.g., antenatal care) were abolished or reduced to a token initial payment. The public was informed of these changes in fees level and system. Simultaneously, the system for accounting and financial reporting was streamlined and made transparent; health workers got a higher and more regular salary, including all the social benefits envisaged by the Ugandan law, as an incentive to improved quality of care [4]; the acquisition, storage and use of drugs and consumables was strictly monitored and rationalised.
Results
The catchment population of the ten hospitals ranged between 53,000 and 172,000; the number of beds from 92 to 320. The number of doctors varied from 3 to 28, that of qualified nurses and midwives from 16 to 115. Small hospitals tended to have a higher rate of hospital staff per 100 beds compared to larger hospitals, except for the large hospital E in Kampala, which had by far the highest rate. Table 1 shows some measures of hospital activity in 1999.
Table 1 Some measures of hospital activity (1999).
Hospital Total admissions Deliveries Cesarean section (%) Length of stay (days) Bed occupancy (%) Outpatient visits Antenatal visits
A 4,234 378 28 17.0 85 8,303 3,061
B 7,809 1,232 18 10.5 93 14,793 6,527
C 8,825 1,824 6 11.4 86 44,791 10,883
D 7,730 548 15 9.9 95 34,509 1,509
E 8,193 2,031 16 6.7 50 76,303 9,508
F 2,531 544 26 5.1 23 6,139 2,071
G 2,143 378 12 5.9 38 9,896 2,674
H 1,498 131 29 5.2 21 17,379 1,551
I 4,419 943 26 8.8 61 11,189 3,388
J 5,722 326 36 NA NA 13,984 1,387
The annual hospital expenditure ranged between $ 72,000 (F) and 1.07 million (E) in the financial year 1998/99 (mean $ 288,750), the last for which reports were available from all hospitals; but excluding hospital E the upper limit dropped to about $ 451,000 (D). Employment cost represented the largest portion of expenditure: from 27% (H) to 47% (F) (mean 35%); followed by medical goods and services: 7% (C) to 26% (J) (mean 16%). Table 2 shows the revenues for the same financial year for each hospital, and its sources grouped into user fees (outpatients and inpatients), government subsidies (the financial support given by the Ministry of Health to each hospital based on a service level agreement), external aid (only financial), and other sources (including donation in kind and services, income generating activities, credits, savings and interest on savings). The percentage from user fees varied widely among hospitals, from 6% to 89%, but on average reached 40%, slightly less than previously thought [2]. The difference between northern and southern hospitals was clear-cut. It was also clear that some hospitals were more successful than others in getting external aid and government subsidies.
Table 2 Source of revenues in the financial year 1998/99.
Hospital Total revenues ($) Percentage from:
User fees (outpatients) User fees (inpatients) Government subsidies External aid Other sources
A 187,000 8 11 18 40 22
B 260,000 6 12 16 18 48
C 355,000 3 3 8 70 17
D 584,000 6 5 14 23 52
E 1,183,000 27 38 1 25 8
F 72,000 17 57 25 0 1
G 112,000 39 50 3 6 3
H 143,000 14 15 26 32 14
I 218,000 9 60 12 0 19
J 145,000 23 55 18 0 3
Average 349,000 15 25 10 28 22
From hospital records and reports, the average outpatient fee appeared to vary between $ 0.28 (C) and 4.38 (G) (mean $ 1.75); for inpatients, the average fee was between $ 1.56 (C) and 55.38 (E) (mean $ 16.02). Most hospitals adopted a fee-for-service system, and very often users had to pay even for preventive services such as antenatal care. The typical outpatient bill included consultation (with a nurse or a clinical officer), drugs, lab tests and X-rays, if any; consultation with a doctor meant an extra charge. Inpatients, in addition to the above, had to pay for hospital stay (general expenses, bed stay, doctor fee), for supplies consumed and for the surgery, if any. In some northern hospitals, a set of services was subsidised (free, low cost or flat rates): care for common childhood illness, pregnancy and childbirth, and some common medical and obstetric conditions. In all hospitals, drugs for tuberculosis, STI and HIV/AIDS were subsidised by the government and administered free to patients; but in most hospitals these patients had to pay for other services. Most patients with chronic conditions had to pay the full fee. Only in one hospital (D) was the fee system adjusted to favour a more rational use of the referral system in the district and sub-district.
Figure 2 shows hospital, i.e. the fees the 303 outpatients and the 102 inpatients interviewed actually paid, and non-hospital cost, i.e. the amount of money patients declared that they had paid for travel and for seeking care elsewhere for the same episode of illness, plus their estimated loss of income due to illness and careseeking. The patients were similar in age and sex distribution in all hospitals. They were coming from a mean distance of 12 (outpatients) and 23 km (inpatients); about 45% were living within 5 km. Education was overall higher for men than for women, with 16% of illiteracy among men and 31% among women. Women were also more unemployed than men: 79% vs. 45%. Non-hospital cost was much less variable among sites than hospital cost, and was proportionally higher among outpatients than among inpatients. Total cost can be calculated by the sum of hospital and non-hospital cost.
Figure 2 Hospital and non-hospital cost borne by outpatients and inpatients. The limits of each box represent the 25th and 75th percentile values; whiskers represent the highest and lowest values excluding outliers; extreme values are shown as dots (1.5 to 3 times the box length) and asterisks (3 times or more the box length); vertical lines represent the median value in each hospital and overall.
The wide variation in user fees among hospitals was not justified by differences in case mix: most patients were seen or admitted for a very small range of conditions (malaria, diarrhoea, respiratory infections, accidents, normal or complicated delivery) in all hospitals. The variation remained even when user fees were compared for the same diagnosis among hospitals. Figure 3 shows the example of malaria, the most common diagnosis among all patients interviewed, in children aged less than 15 years seen as outpatients or admitted as inpatients. A child with malaria paid 1 US $ for admission in hospitals C and D, and 20 times that amount in hospital E. A severity score would help interpreting these figures, but it is difficult to imagine a difference in severity that would completely explain this difference in cost. The same variation was observed for other common conditions.
Figure 3 Hospital cost borne by outpatient and inpatient children (less than 15 years) with malaria. The limits of each box represent the 25th and 75th percentile values; whiskers represent the highest and lowest values excluding outliers; extreme values are shown as dots (1.5 to 3 times the box length) and asterisks (3 times or more the box length); vertical lines represent the median value in each hospital and overall; absence of box and whiskers means less than three cases; the value for hospital F (inpatients) coincides with the overall median.
At the time of the interview, 220 outpatients (73%) had fully paid their fees, 36 (12%) had paid partially, and 47 (16%) had not paid; among these, 22 were treated on credit and 21 were exempted; the non-payers were only four. It was impossible to interview all the inpatients after payment; some were interviewed the day before discharge or as soon as discharge was decided. Some patients referred that in previous episodes of illness they were denied or reduced treatment, drugs in particular, if they could not afford to pay. Few hospitals held funds to help the poor; according to hospital managers, poor patients would be exempted after some form of subjective judgement by members of the hospital management team. For this reason, indigent patients tended to avoid these hospitals in favour of other facilities, mostly government services. Out of 303 outpatients, only 34 (11%) benefited from some kind of exemption; among them, 11 were part of the hospital staff and the others were teachers or pupils of the catholic school, parish clerks and soldiers. Only one outpatient was exempted because he was poor. The inpatients who benefited from some kind of exemption were 6 out of 102 (6%): five for specific health conditions and one for indigence.
To pay hospital and non-hospital expenses, 34% of outpatients and 42% of inpatients had to borrow money; 24% and 30%, respectively, had to sell goods and property. Moneylenders were usually members of the family; interestingly, 22 women said that they had to borrow money from their husband. Regarding the sale of goods, outpatients sold vegetables (19/67, 28%), cash crops (11/67, 16%), or animals (20/67, 30%); among inpatients, 50% (16/32) sold animals and 13% (4/32) part of their land. Patients found particularly difficult to afford hospital fees during the planting season and at the time of enrolling children at school. Lack of money was the most important reason for seeking care elsewhere for the same episode of illness (more than 50% of users) or for coming late for care (38%).
From March 2001, the intervention was progressively applied in hospital F and in nearby hospital X, while hospital Y served as control (modest reduction of fees only for outpatients). Hospitals X and Y were not included in the survey and belong with hospital F to the same diocesis. Figure 4 shows the number of admissions and average fees in the children and maternity wards in 2001 in hospital F; it shows also the number of admissions and outpatient visits per year in hospitals F, X and Y between 1993, when user fees were widely introduced, and 2002. The upward trend after lowering the fees is clear. In hospital F the number of deliveries increased from 456 in 2000 to 605 in 2001, with 136 and 116 caesarean sections, respectively. Bed occupancy went up from 23% to 46% in the same period (up to 63% in December 2001), mostly because of the 185% and 63% increase in admissions in children and maternity wards. The reduction of average user fees was compensated by increased use, and the monthly revenue from user fees actually increased, from $ 2693 in January to $ 3421 in December 2001. The revenue for user fees was $ 33,630 in 2000 and $ 35,050 in 2001; due to the simultaneous growth of government subsidies, from $ 12,690 to $ 37,280, granted because of better access and use, the proportion of revenue from user fees went down from 73% to 41%. Despite an increased expenditure for salaries, the financial statement for 2001 showed a positive balance.
Figure 4 Selected measures of hospital activity before and after the intervention. Number of admissions and average fees in the children and maternity wards of hospital F in 2001 (above), and number of admissions and outpatient visits per year between 1993 and 2002 in hospitals F, X and Y (below).
Discussion
The results of this study can not be generalised because it was not carried out on a random sample of hospitals and users. But the information on fees appears to be reliable because the average fee per patient calculated from annual reports and financial statements is similar to the median fee estimated through interviews. Our results show wide differences among hospitals; these differences do not seem to reflect social and economic differences of users, except perhaps for hospital E in Kampala, nor differences in terms of offered services. Our results also suggest that the level and structure of user fees may affect access and use of services, as shown by other studies [1,5-11].
User fees are generally considered an unfair mechanism of financing for health services; they represent a barrier to access for the poor and the powerless, and they discriminate the sickest [12]. Even more so when they are not predictable by the users as in the majority of the hospitals we surveyed. Combined with the unpredictability of illness, user fees expose people to the risk of severe economic crisis from which it may prove difficult to emerge [13,14]. Exemption mechanisms are unlikely to soften these negative effects of user fees [15]. For these reasons, user fees interfere with the mission of the UCMB, which is to serve the entire population but in particular the most vulnerable groups.
But user fees represented a large proportion of hospital revenues in five of the surveyed hospitals. More equitable systems of health services financing, such as social health insurance and progressive tax contributions, must be designed but are far from being implemented in sub-Saharan Africa [16]. External aid could play an important role but its amount and regularity can not be anticipated. Government subsidies are low and unlikely to increase dramatically; they would not compensate for reduced revenue from user fees. A reduction of the already low hospital expenditure is also unlikely, although an improvement in efficiency may lead to better services provided at the same cost. As regrettable as it may sound, user fees are likely to remain an important factor for the sustainability of PNFP hospitals in Uganda and other countries in sub-Saharan Africa for many years.
Conclusion
The fee system applied in the pilot hospitals, i.e. flat rates and lower overall fees with reduced charges for the more vulnerable groups, after the survey seems feasible and effective. It is hoped that its extension to other PNFP hospitals in Uganda and elsewhere may ensure that a proportion of annual revenues (20% to 30% on average) continue to derive from user fees without negative effects on access, use and equity. This fee system, combined with regular revenue from government subsidies and some external aid, could allow these hospitals to survive while more pro-poor public financing systems are developed and implemented.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AC, AM and GM conceived and designed the study. SS organised the field survey and was in charge of the post-survey intervention. JA, SA, AC, AKK, AM, GM and MM conducted the field survey and gathered all the data, except those gathered by SS during the post-survey intervention. LR, AM, AC, GM and SS analysed the data; the interpretation was shared with the whole team. GM and AM wrote the initial report based on which AC drafted and finalised the paper. All authors read and approved the final manuscript.
Acknowledgements
This research has been funded by the Italian Episcopal Conference, but would not have been conducted without the support of the Hospital Management Teams and the Hospital staff. We are particularly indebted to Professor Michel Lejeune, vice-chancellor of Uganda Martyrs University, Nkozi, and to Dr Daniele Giusti, Executive Secretary of UCMB for their advice and suggestions and for the difficult coordination among the ten hospitals involved in the research.
==== Refs
Gilson L The lessons of user fee experience in Africa Health Policy Plan 1997 12 273 285 10176263
Giusti D Cost recovery rates in PNFP health sector 1997 Kampala, Uganda Catholic Medical Bureau
Church TRC Health Services Mission Statement and Policy 1998 Kampala, Episcopal Conference of Uganda
Kipp W Kamugisha J Jacobs P Burnham G Rubaale T User fees, health staff incentives, and service utilization in Kabarole District, Uganda Bull World Health Organ 2001 79 1032 1037 11731810
Mbugua JK Bloom GH Segall MM Impact of user charges on vulnerable groups: the case of Kibwezi in rural Kenya Soc Sci Med 1995 41 829 835 8571154 10.1016/0277-9536(94)00400-N
Mwabu G Mwanzia J Liambila W User charges in government health facilities in Kenya: effect on attendance and revenue Health Policy Plan 1995 10 164 170 10143454
Ogunbekun I Adeyi O Wouters A Morrow RH Costs and financing of improvements in the quality of maternal health services through the Bamako Initiative in Nigeria Health Policy Plan 1996 11 369 384 10164194
Haddad S Fournier P Quality, cost and utilization of health services in developing countries. A longitudinal study in Zaire Soc Sci Med 1995 40 743 753 7747209 10.1016/0277-9536(94)00134-F
Hussein AK Mujinja PG Impact of user charges on government health facilities in Tanzania East Afr Med J 1997 74 751 757 9557416
Asenso-Okyere WK Anum A Osei-Akoto I Adukonu A Cost recovery in Ghana: are there any changes in health care seeking behaviour? Health Policy Plan 1998 13 181 188 10180407 10.1093/heapol/13.2.181
Blas E Limbambala M User-payment, decentralization and health service utilization in Zambia Health Policy Plan 2001 16 Suppl 2 19 28 11772987
Gilson L Mills A Health sector reforms in sub-Saharan Africa: lessons of the last 10 years Health Policy 1995 32 215 243 10156640 10.1016/0168-8510(95)00737-D
Whitehead M Dahlgren G Evans T Equity and health sector reforms: can low-income countries escape the medical poverty trap? Lancet 2001 358 833 836 11564510 10.1016/S0140-6736(01)05975-X
Xu K Evans DB Kawabata K Zeramdini R Klavus J Murray CJ Household catastrophic health expenditure: a multicountry analysis Lancet 2003 362 111 117 12867110 10.1016/S0140-6736(03)13861-5
Willis CY Leighton C Protecting the poor under cost recovery: the role of means testing Health Policy Plan 1995 10 241 256 10151842
Bennett S Gilson L Health financing: designing and implementing pro-poor policies 2001 London, DFID Health Systems Resource Centre
| 15871744 | PMC1142334 | CC BY | 2021-01-04 16:39:31 | no | Int J Equity Health. 2005 May 4; 4:6 | utf-8 | Int J Equity Health | 2,005 | 10.1186/1475-9276-4-6 | oa_comm |
==== Front
J Circadian RhythmsJournal of Circadian Rhythms1740-3391BioMed Central London 1740-3391-3-71586971410.1186/1740-3391-3-7ReviewEvolution of temporal order in living organisms Paranjpe Dhanashree A [email protected] Vijay Kumar [email protected] Chronobiology Laboratory, Evolutionary and Organismal Biology Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, PO Box 6436, Bangalore 560 064, Karnataka, India2005 4 5 2005 3 7 7 31 3 2005 4 5 2005 Copyright © 2005 Paranjpe and Sharma; licensee BioMed Central Ltd.2005Paranjpe and Sharma; 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.
Circadian clocks are believed to have evolved in parallel with the geological history of the earth, and have since been fine-tuned under selection pressures imposed by cyclic factors in the environment. These clocks regulate a wide variety of behavioral and metabolic processes in many life forms. They enhance the fitness of organisms by improving their ability to efficiently anticipate periodic events in their external environments, especially periodic changes in light, temperature and humidity. Circadian clocks provide fitness advantage even to organisms living under constant conditions, such as those prevailing in the depth of oceans or in subterranean caves, perhaps by coordinating several metabolic processes in the internal milieu. Although the issue of adaptive significance of circadian rhythms has always remained central to circadian biology research, it has never been subjected to systematic and rigorous empirical validation. A few studies carried out on free-living animals under field conditions and simulated periodic and aperiodic conditions of the laboratory suggest that circadian rhythms are of adaptive value to their owners. However, most of these studies suffer from a number of drawbacks such as lack of population-level replication, lack of true controls and lack of adequate control on the genetic composition of the populations, which in many ways limits the potential insights gained from the studies. The present review is an effort to critically discuss studies that directly or indirectly touch upon the issue of adaptive significance of circadian rhythms and highlight some shortcomings that should be avoided while designing future experiments.
circadianadaptationcyanobacteriaDrosophiladevelopment timelifespan
==== Body
Introduction
The earth's rotation around its axis causes predictable changes in the geophysical environment, thereby providing organisms with options to occupy appropriate spatio-temporal niches. Most organisms place themselves suitably in such niches using precise time-keeping mechanism(s) that can measure passage of time on an approximately 24 h scale (and hence are known as circadian clocks) [1]. Extensive studies over the past fifty years on a wide range of organisms have revealed some unique features of these timekeeping devices that distinguish them from other biological clocks. Some of them are summarized as follows: circadian (circa = approximately; dies = a day) clocks (i) have an inherent near-24 h periodicity, (ii) are protected from changes in temperature, nutrition and pH, within physiologically permissible limits, and (iii) can be tuned to oscillate with exactly 24 h period – a key property of circadian clocks known as entrainment, which enables living organisms to keep track of time in their local environment. These clocks are ubiquitous and are found at various levels of organization and complexity, which suggests that they must provide adaptive advantage to their owners. Circadian clocks enhance the innate ability of organisms to survive under ever-changing environments by enabling them to efficiently anticipate periodic events such as availability of food, light and mates [1-6]. It is therefore not too surprising that a wide variety of organisms such as bacteria, fungi, fish, amphibians, reptiles, insects, mammals including humans, as well as plants are able to measure time on a 24 h scale. It is believed that circadian clocks have evolved under selection pressures comprising of periodic biotic and abiotic cycles of the environment, which act on these clocks under the entrained state. As a result, precisely timed rhythmic activities confer greater adaptive advantage compared to randomly occurring activities, and in turn those clocks that enable organisms to maintain such phases (time of the day) are selected for [2,7]. Hence, the free-running phenotypes of circadian clocks are considered to be an evolutionary outcome of natural selection on entrained clocks [4]. Although circadian clocks are believed to have arisen as a result of adaptive evolution under periodic environments, there has been hardly any rigorous and conclusive empirical study to support this [2].
Timekeeping in fluctuating environments
Circadian clocks regulate a number of key behaviors in a wide variety of organisms. For example, most insects emerge as adults from their pupal case (an act known as eclosion) close to "dawn", when humidity is highest in the environment [8-10]. It is believed that by timing eclosion to the early hours of the day, insects prevent desiccation and thus enhance their ability to survive [11]. Circadian clocks also help organisms to restrict their activity to species-specific times of the day, which enables them to find food and mates, escape predators, and avoid undue competition between sympatric species. For example, in Drosophila parasitoids, activity peaks of different species occur at different times of the day, which significantly reduces intrinsic competitive disadvantage for the inferior competitor, and such temporal partitioning is achieved at least partly with the help of circadian clocks [12]. Proximal advantages of possessing circadian clocks have also been evaluated in a few studies in other animals. In a study on guillemots (Uria lomvia), a greater percentage of fledglings jumping out of their nests at non-species-specific times fell prey to gulls compared to those jumping during species-specific times of the day [13]. Thus, timing the jumping activity during evening hours, in synchrony with other juveniles resulted in greater chances of survival in the young ones [14]. In ground squirrels living in the wild, the hypothalamus-based circadian clock – the suprachiasmatic nucleus (SCN) – has been shown to play an important role in survival. Under laboratory conditions, SCN-ablated animals survived equally well as the controls [15], but the SCN-ablated animals quickly fell prey to feral cats when released into a semi-natural enclosure [16] (Figure 1A). This suggests that functional clocks may not be essential for survival under controlled conditions, but might become crucial under natural environments. In a subsequent marathon field study on free-living chipmunks, Tamias striatus, DeCoursey and coworkers [17] demonstrated that reduction in survival of the SCN-ablated animals (Figure 1B) was due to enhanced predation, perhaps due to increased nighttime restlessness.
Figure 1 Circadian clocks are essential for survival of organisms under natural conditions. (A) Average survivorship of white-tailed antelope ground squirrels under semi-natural habitats. The animals were released in semi-natural habitat after surgical removal of their supra-chiasmatic nucleus (SCN) based circadian clocks. During the study period three out of five SCN-lesioned (SCN-X) individuals were predated upon as compared to two out of seven control animals. (modified after DeCoursey et al, 1997 [16]) (B) Average survivorship of free-living eastern chipmunks under natural environment. Free-living animals were captured from the study area and released back after surgical ablation of their SCN. The control animals were handled similarly and released back to the study area. During the eighty days of the study, more than 80% of SCN ablated individuals fell prey to weasels while mortality was significantly less in the surgical (sham) and intact controls. (modified after DeCoursey et al, 2000 [17])
Circadian clocks are also important for social insects such as honeybees and ants. Social insect colonies are normally faced with challenges such as changing colony sizes, time of the year, food availability, predation pressure and changing climatic conditions. The survival of these colonies under such demanding conditions requires a number of tasks to be performed simultaneously. These insects seem to have evolved division of labor, an arrangement that, in addition to enhancing efficiency of task management, promotes biological evolution of complexity and diversity [18]. In a series of experiments, Robinson and coworkers quite convincingly demonstrated that social insects use circadian clocks to efficiently manage division of labor [19]. In the colony of the Asian honeybee, Apis mellifera, young workers (nurses) perform tasks that can be categorized as "nursing" practically around the clock without taking any rest [20], while older honeybees (foragers) visit flowers to collect pollen and nectar in a rhythmic manner timed by well-developed circadian clocks [21]. It appears that honeybees use circadian plasticity to match age-dependent behavioural development, a phenomenon commonly known as age-polytheism [22,23]. In certain species of ants, virgin queens and males mate during nuptial flights, which occur at a species-specific time of the day, during the mating season [24,25]. Virgin queens and males use circadian clocks to time mating flight in order to encounter mating partners from neighboring colonies [24,25]. In the ant species Camponotus compressus, virgin queens and males use circadian clocks to time their mating flights by maintaining appropriate phase relationship with the cyclic environments, perhaps to facilitate cross-breeding between colonies and to avoid inter-species mating [26-28]. The worker castes of this species, namely the majors and media, also use circadian clocks to time their day-to-day repertoire. Foragers have well-developed circadian clocks, while soldiers guard the nest around the clock showing no obvious sign of rhythmicity. The media workers are task generalists. They are found foraging most of the time or are restricted to their colonies taking care of the queen and her brood. The activity patterns of media workers switched from nocturnal to diurnal, and clock period changed from less than 24 h to greater than 24 h, and vice-versa, suggesting that they are involved in shift work in their colonies. At the same time, activity of minor workers neither entrained to LD cycles nor showed any sign of free-run in DD regime, which matches well with their role as nurses. Thus, activity patterns of different castes of the ant species C. compressus seem to correlate well with the tasks assigned to them in their colonies [26].
Migratory birds use circadian clocks to keep track of rapidly changing day lengths in order to navigate at a specific time of the year to a more favorable climate [1]. Similarly, hibernating mammals use circadian clocks in their preparations to enter hibernation [1]. In a recent study it was reported that the Monarch butterflies, Danaus plexippus, undertake migratory flights every fall from northeastern America to their over-wintering grounds in central Mexico [29]. The authors demonstrated that circadian clocks play a key role in time-compensated navigation of migratory flight in Monarch butterflies [30]. European starlings use circadian clocks to compensate for changing position of the sun on long-distance journeys [31]. Similarly, golden-mantled squirrels enter hibernation in autumn when day length begins to shorten and mean daily temperature starts to drop [1]. These animals use circadian clocks to measure day lengths in order to prepare themselves for hibernation at an appropriate time of the year. On average, hibernation lasts for about 7 months with periodic wake-up bouts for sustaining brain and kidney functions through long winters. These wake-up bouts are regulated in part by circadian clocks [1], as SCN-ablation caused marked changes in the duration of wake-up bouts and the duration of hibernation [15,32]. Therefore, regularity of wake-up bouts appears to be essential even under hibernating conditions for rationing limited fat supply to last for the entire winter, as wake-up bouts are associated with muscular shivering and are metabolically expensive [1]. It is therefore evident from the above studies that circadian clocks are essential for organisms in maintaining appropriate temporal niches in their ecological and temporal environments. Previous studies suggest that circadian clocks provide proximal advantage to their owners, but they by no means serve to emphasize that these timers have any ultimate selective advantage.
Clock fidelity
The issue of entrainment and its implications in temporal niche selection has remained central to circadian rhythm research since its inception. It is believed that natural selection acts on the phase-relationship between biological rhythm and environmental cycle, defined as the time interval between a given phase of the biological rhythm and a predictable phase of the environmental cycle. Therefore, maintenance of precise timing for behavioral and metabolic activities should be one of the most essential functions of circadian clocks, especially for organisms living in natural environments where light, temperature, humidity, food, predators and competitors fluctuate with time of the day. How do organisms living in seemingly timeless environments such as caves, burrows and cozy apartments know the time in their local environment? Although, no clear answer to this question exists as yet, it is believed that they do so by synchronizing their circadian clocks with the help of reliable time cues in their external environment [7,33-35]. Entrainment of circadian clocks largely depends upon two key features: phase response curve (PRC) and free-running period (τ) [7,34-36]. The free-running period is considered as an invariant property as it is assumed to remain unchanged throughout the entrainment process [35,37]. Yet, studies on a wide range of organisms have revealed that τ of circadian clocks is not an invariant property, but varies in response to different environmental conditions, often reflecting residual effects of prior environmental experience typically referred to as "after-effects" [38-42]. For example, mice exposed to LD cycles continue to exhibit rhythmic locomotor activity in DD with τ close to those of the LD cycles previously experienced, for about 100 days [33]. Such after-effects may be of some functional significance, as they could help organisms to maintain a stable phase relationship with the environmental cycles, even when environmental LD cycles are perturbed due to cloud cover or behavioural changes [33,36,43,44]. Therefore, it appears that circadian clocks have evolved a number of mechanisms to enhance their stability in ever-fluctuating environments, which in turn could increase the organism's chances of survival under natural environment [34,35].
Dating clocks
While the proximate as well as ultimate driving forces for the evolution of circadian clocks remain largely unknown, much has been speculated as to when biological clocks might have first appeared and about what could have been the initial selection pressures that might have acted on early biological clocks [45,46]. It was believed that circadian clocks were a feature of eukaryotic organization, and that 24-h clocks would be of no advantage to prokaryotes, whose numbers double every few hours [5]. It was also believed that cellular organization of prokaryotes was too simple to accommodate complex mechanisms that are required to regulate circadian rhythms. However, it is no more a hypothesis but a fact that even primitive unicellular organisms such as cyanobacteria house functional circadian clock machinery [5]. This finding, thus pushed back the origin of circadian clocks by several hundred million years, and it is now believed that circadian clocks may have appeared on earth along with primitive life forms [46].
Circadian clocks in cyanobacteria are regulated by a cluster of three Kai (clock) genes – KaiA, KaiB and KaiC [47]. Using sequence data of these genes from several prokaryotic genomes, Dvornyk and co-workers [48] demonstrated that Kai genes and their homologs have quite different evolutionary histories. The KaiC gene is also found in Archaea and Proteobacteria [48], and among the three Kai genes, KaiC is evolutionarily the oldest. The origin of the Kai gene cluster appears to be one of the key events in the evolutionary history of cyanobacteria – one of the most primitive life forms on the earth. Based on the genomic data, the authors argued that circadian clocks have evolved in parallel with the geological history of earth, and natural selection, multiple lateral transfers, gene duplications and gene losses were among the major factors that further refined their evolution [48]. It is also possible that several features of circadian clocks have evolved in different organisms independently of each other and any similarity between them could be a result of convergent evolution [49].
It is also possible that the genes now involved in clock machinery formerly performed entirely different functions and were later appropriately modified to be incorporated in the clock machinery due to the changing needs of organisms in the face of cyclic selection forces. For example, Cryptochrome – the blue light sensitive photopigment used for circadian photoreception in Drosophila and plants – has been shown to exhibit striking similarity to bacterial photolyase, an enzyme involved in light-dependent DNA repair [51,52], suggesting that the Cry gene initially served as a key player in other cellular function(s) and might have been incorporated as part of the clock machinery at a much later stage. Regardless of the views about the original purpose of circadian clocks, there is a general belief among circadian biologists that circadian clocks evolved under the influence of cyclic factors such as light, temperature and humidity as primary selection forces. At some later stage, rhythmic activities of prey, predators and competitors might have provided additional selection pressures for its fine-tuning [46,53,54].
Several hypotheses have been put forward to explain the appearance of circadian clocks on this planet. Pittendrigh believed that circadian rhythms had evolved under selection pressure presented by environmental LD cycles, wherein photophobic processes were confined to darkness and photophilic processes to light [4]. Thus, according to Pittendrigh's "escape-from-light" hypothesis, circadian rhythms evolved to protect organisms from deleterious photo-oxidative effects of the environment by helping them reschedule light-sensitive reactions during the night [4,46,50]. For example, in cyanobacteria, key metabolic processes such as oxygen-evolving photosynthesis and oxygen-sensitive nitrogen fixation needs to be segregated in space and/or in time. Some groups of cyanobacteria have evolved special structures called heterocysts for nitrogen fixation, thus, allowing spatial segregation of the two incompatible processes, while in nonheterocyst cyanobacteria such segregation is achieved by scheduling the two processes at different times of the day [55,56]. To test the validity of the "escape from light" hypothesis, Nikaido and co-workers [57] performed experiments on unicellular alga Chlamydomonas reinhardtii. The survival of cells of C. reinhardtii was measured after exposing them to UV radiation at different times of the day. The results suggest that the cells were most sensitive to UV radiation during evening hours, when the UV component of solar radiation is normally attenuated. This suggests that the circadian timing system in C. reinhardtii has evolved to time crucial light-sensitive processes such as cell division during the later part of the day or in the early part of night to avoid deleterious effects of UV radiation [57].
Conserved clocks
Extensive genetic and molecular studies during the last three decades on model organisms such as bacteria, fungi, fruit flies and mice have provided in-depth understanding of the molecular mechanisms underlying circadian clocks. Although the finer details of the molecular players in the clock machinery appear to be different in many organisms, their functions bear remarkable degree of similarity across taxa [58] (Figures 2, 3, 4, 5). The underlying molecular mechanisms involve multiple feedback loops comprising of genes whose transcripts and/or protein products oscillate with near 24 h periodicity [59-61]. The positive elements in the molecular clockwork are transcriptional activators of one or more clock genes with DNA binding bHLH (basic Helix-Loop-Helix) motifs. These activators enhance the transcription of clock genes by binding to specific E-box sequences in the promotor region of the clock genes. This results in abundance of transcripts, which then translate to yield clock proteins. The protein products form heterodimers by interacting via a PAS (PER-ARNT-SIM) domain, and are subsequently phosphorylated in the cytoplasm by specific kinases, after which they enter the nucleus. Some heterodimers act as negative elements in the feedback loops, as their binding brings about conformational changes in the protein structures of the transcriptional activators, in a manner that they can no longer bind to the promoter region of the clock genes, thereby inhibiting their transcription. The positive elements of the loop also activate the transcription of a few clock-controlled genes (ccgs), which control overt rhythmicity directly or indirectly through yet unknown mechanisms. The molecular feedback loops are interconnected such that the protein heterodimer acting as transcriptional activator in one loop can inhibit transcription of clock genes in the other loop. Such components of molecular loops, which play dual roles in the core clock mechanisms, are particularly important for self-sustained molecular oscillations. The DNA binding bHLH domain [59,62] and the protein-protein interacting PAS domain are highly conserved in organisms ranging from cyanobacteria to mammals [63]. The KaiA protein in cyanobacteria [47] (Figure 2), WCC in Neurospora (Figure 3), CLK/CYC in Drosophila (Figure 4), and CLOCK/BMAL1 in mouse (Figure 5) act as transcriptional activators, of which all except KaiA are heterodimeric transcriptional activators. The negative elements such as KaiC protein in Synechococcus, FRQ in Neurospora, PER/ TIM in Drosophila and PER1, PER2, CRY1, CRY2 in mouse block their own transcription by interacting with transcriptional activators. In addition, WCC in Neurospora, CLK/CYC in Drosophila and CLK/ BMAL1 in mouse are some of the key elements that play dual roles; as transcriptional activators in one loop and transcriptional inhibitors in the other (Figures 2, 3, 4, 5). In addition, in many organisms genes involved in light input pathways are also involved in the core clock mechanisms [51]. The basic function of the molecular clock bears remarkable similarity in a wide range of organisms. Besides high degree of functional similarity between the molecular clocks, there is also a considerable degree of structural similarity between the clock genes of insects and mammals. The clock (clk) and doubletime (dbt) genes in Drosophila and mammals have considerable sequence homologies and have similar functional roles in the respective organisms [64,65]. Homologs of per gene have been reported in several species of Drosophila [66], and a number of other insect species such as housefly (Musca domestica) [67] and honeybee (Apis mellifera) [22]. Orthologs of per have been identified in mammalian system and more recently in Zebrafish [68]. Furthermore, the photopigment cryptochrome involved in the light input pathways of circadian clocks in fruit flies has been found to have remarkable structural and functional similarities to those of mammals and plants [51].
Figure 2 Molecular feedback loops of cyanobacteria. A cluster of KaiABC genes controls circadian rhythms in cyanobacteria. KaiA gene product acts as a positive regulator for KaiBC transcription, while KaiBC products along with other proteins inhibit their own transcription.
Figure 3 Interlocked molecular feedback loops of Neurospora. White-collar complex (WCC) acts as the transcriptional activator (positive element) of Frequency gene (Frq). The protein product of Frq undergoes phosphorylation in the cytoplasm under the influence of specific kinases, and subsequently acts as inhibitor of its own transcription (negative element). WCC levels are regulated by another gene called Vivid (Vvd), which in turn is regulated by WCC complex. Thus, WCC acts as one of the key components of Neurospora clock that connects the two loops, and hence appear to be important for the persistence of molecular oscillations. In addition, WCC is light sensitive, and appears to be crucial for light entrainment for the Neurospora molecular clock.
Figure 4 Interlocked molecular feedback loops in Drosophila melanogaster. CLOCK/CYCLE heterodimer acts as transcriptional activator (positive element) for period (per) and timeless (tim) genes. The heterodimer of PER/TIM is phosphorylated in the cytoplasm in the presence of specific kinases, and the phosphorylated complex then acts as inhibitor for its own transcription (negative element). The VRI and PDP1 proteins regulate the levels of CLK/CYC complex, which in turn are regulated by CLK/CYC. Thus, CLK/CYC heterodimer appears to be an important component that connects the two loops and is important for sustaining molecular oscillations. The protein Cryptochrome (CRY) has been implicated in the light entrainment pathways of the Drosophila molecular clock.
Figure 5 Interlocked molecular feedback loops of mammals. CLOCK/ BMAL1 heterodimer acts as the transcriptional activator (positive element) for Period (Per) and Cryptochrome (Cry) genes. The PER/CRY protein complex is phosphorylated in the cytoplasm by specific kinases, which then acts as inhibitor for their own transcription (negative element). In addition, these heterodimers activate Bmal1 transcription. CLK/BMAL1 transcription is inhibited by REV-ERBα, which in turn is regulated by CLK/BMAL1. Thus, CLK/BMAL1 heterodimer appears to be one of the key components of mammalian molecular clock, which connects the two loops. The Period1 gene product (PER1) is light-sensitive and appears to be important for the light entrainment of mammalian molecular clock.
Although the overall molecular mechanisms underlying circadian clocks in various organisms are to a great extent conserved, there are also subtle differences. For example, in Drosophila, mRNA and protein levels of cycle (cyc), which forms an important part of the transcriptional activator CLK/CYC, do not oscillate, whereas, dClk mRNA as well as protein levels show robust oscillations [60]. In mammals, the level of CLK protein does not oscillate, but BMAL1 is prominently rhythmic [60]. Furthermore, in Drosophila, CRY acts as a photopigment and as an important component of the core clock mechanisms in the peripheral clocks [69]. In the mammalian circadian timing system, CRY is only a part of the core molecular mechanisms; the possibility of its role in light perception has been ruled out [58]. Further, in contrast to the Drosophila molecular clock, which consists of only one cry and one per gene, the mammalian molecular clock consists of two Cry genes (mCry1 and mCry2) and three Per genes (mPer1, mPer2 and mPer3) [58]. Molecular mechanisms regulating circadian clocks in Chlamydomonas reinhardtii have been reported to be entirely different. Extensive search for potential homologs to genes that are known to encode components of the circadian clock in other organisms has revealed that there are no obvious homologs in the C. reinhardtii genome, except for the kinases and phosphatases that are involved in the molecular clockwork [70]. The kinases and phosphatases in fungi, plants, flies, mammals and C. reinhardtii are highly conserved, and it appears that they play a key role in the clock mechanisms. One of the two CRY proteins found in C. reinhardtii is closely related to plant CRYs, while the other one is more similar to animal CRYs. Since there are no homologs of any known clock genes in C. reinhardtii, it is possible that the green alga might host novel clock mechanisms involving some novel core clock components [70]. Barring a few exceptions such as those in Chlamydomonas circadian timing system, the overall molecular mechanisms underlying light input pathways, rhythm generating core mechanisms and rhythm transduction mechanisms that send rhythmic signals to efferent organs bear striking structural and functional similarities in organisms ranging from cyanobacteria to mammals. Given that the behavioural and metabolic processes regulated by circadian clocks are so diverse, it is astonishing that the underlying molecular mechanisms giving rise to these varieties of rhythmic phenomena are so similar across a wide range of taxa.
Clock for all seasons
Organisms living in temperate regions are exposed to drastic changes in photoperiod and temperature. Circadian clocks are believed to play an important role under such demanding situations [71]. Studies on several strains of D. littoralis originating from a wide range of geographic locations at different latitudes revealed a mild latitudinal trend in the phase and period of eclosion rhythm [72]. The northern strains had shorter period and earlier phase of eclosion compared to the southern strains [72]. Similar latitudinal clines for phase and amplitude of eclosion rhythm were also reported in D. auraria [71]. Since amplitude of circadian rhythm responds to changes in photoperiod as well as temperature, it was concluded that these insects use circadian clocks to sense seasonal changes in their environment [73]. Furthermore, fifty-seven European populations of D. littoralis showed latitudinal cline for adult diapause, where the northern populations responded to longer critical day lengths than the southern populations [72]. Later, in a separate study, clinal patterns in threonine-glycine (Thr-Gly) repeats were reported at the period locus in European and north African strains of D. melanogaster [74] and D. simulans [75]. The northern strains showed higher frequency of (Thr-Gly)17 compared to the southern strains, while the frequency of (Thr-Gly)20 was higher in the southern strains than in northern strains [74,75]. Further studies on the locomotor activity rhythm in these populations at 18°C and 29°C revealed that circadian clocks of the (Thr-Gly)20 variants had the most efficient temperature compensation ability, while this was not the case for the (Thr-Gly)17 variants, as they showed period shortening at lower temperatures [76]. Since clinal variation in phase and period in these strains appear to have arisen as a result of natural selection, presence of such latitudinal clines can be taken as an indirect evidence for the adaptive evolution of circadian clocks [71,72].
Clocks for birth and death
The assumption that circadian clocks influence fitness traits has formed the basis of several studies aimed at addressing adaptive significance of circadian rhythms. It is generally believed that faster clocks speed up development and cause reduction in lifespan, while slower clocks slow down development and lengthen lifespan [77-80]. Several studies have been carried out in a variety of organisms to investigate possible links between circadian clocks and life history traits such as pre-adult development time and adult lifespan. In an extensive study on the per mutants of D. melanogaster, which display circadian rhythms with widely different periods, pre-adult development time was measured under continuous dim light (LL), very bright continuous light (VLL), continuous darkness (DD), light/dark (LD) cycles of 12:12 h, and LD 12:12 h superimposed with temperature cycles (LD 12:12 T). Under all light regimes, development time of per mutants was positively correlated with τ of their circadian clocks, i.e. perS mutants (τ = 19 h) developed faster than wild type flies (τ = 24 h), which in turn developed faster than the perL mutants (τ = 28 h) [77]. A positive correlation between development time and clock period was seen even in absence of the overt rhythmicity under LL regime and also under entrained conditions such as those prevailing under LD cycles, which suggests that the per mutation has pleiotropic effects on circadian phenotype and pre-adult development time. In a recent study in D. melanogaster, pre-designed to bypass such pleiotropic effects, clock period and developmental time were positively correlated (faster eclosion rhythm was associated with faster development and slower oscillations accompanied slower development), thus suggesting a possible role of the periodicity of LD cycles and/or of eclosion rhythm in determining the duration of pre-adult development [80]. In a separate study on the melon fly (Bactrocera cucurbitae) that involved selection for faster and slower pre-adult development, faster developing lines had faster circadian clocks, whereas slower developing lines had slower circadian clocks [81]. The timing of behaviors such as locomotion and preening was shifted significantly to earlier hours of the day in faster developing lines compared to the slower developing lines. The mating peaks in the faster developing lines occurred close to dusk while most of the flies from the slower developing lines mated during the night [82]. The period of locomotor activity rhythm was shorter (τ ~ 22.6 h) in faster developing lines and longer (τ ~ 30.9 h) in the slower developing lines [83]. Although, most studies suggest a role of circadian clocks in timing pre-adult development, the robustness of such conclusion is limited by the fact that association between development time and circadian clocks in some of the studies shows very little effect of light regime, suggesting pleiotropic effects of the per mutation.
In a study on the tau mutant hamsters, heterozygous (τ ~ 22 h) animals under laboratory LD (14:10 h) cycles lived shorter than the wild type animals (τ ~ 24 h), but the average lifespan of homozygous animals (τ ~ 20 h) did not differ from those of the wild type animals [78]. Contradictory results were obtained in a similar study performed under constant dark (DD) conditions, wherein homozygous animals lived significantly longer than the wild type controls, while the average lifespan of heterozygote animals did not differ from those of the wild type controls [79]. Such differences in outcome could be due to the fact that the two studies were performed under different environmental conditions, and environmental factors are known to modify the outcome of such studies [77,78,84]. In a separate study in fruit flies (Drosophila melanogaster), significance of circadian clocks in physiological well being has been investigated in some detail. The lifespan of perT (short period mutant, τ = 16 h), and perL (long period mutant, τ = 29 h) mutants was reduced considerably compared to per+ (wild type, τ = 24 h) flies, even when flies were maintained under LD cycles with periodicity closer to the endogenous periodicity of the mutant lines [85]. The studies discussed above serve to emphasize that lifespan of D. melanogaster is not regulated by the clock period; rather it is determined by the genotype of the flies, which suggests pleiotropic effect of per mutation on clock period and lifespan. The role of circadian clocks in determining life-history traits is likely to be important for the adaptive evolution of organisms, especially under periodic environments. Evidence at hand provides at least strongly suggestive, if not conclusive, evidence that circadian clocks control key life-history traits. They also raise a possibility that some evolutionary response of life-history traits to forces of natural selection may be partly mediated through changes in circadian clocks.
Clocks for reproduction
The role of circadian clocks in reproductive output of D. melanogaster was investigated in great depth in clock mutants. Studies on loss of function mutants of D. melanogaster such as per0, tim0, cyc0, Clkjrk revealed that single mating among clock-deficient phenotypes result in ~ 40% lesser progeny compared to the wild type flies [86]. In general, null mutants laid fewer eggs, out of which only a few were fertile [86]. Further experiments on per0 and tim0 flies showed that the amount of sperm released from the testes to seminal vesicles of males was significantly reduced in the null mutants compared to the wild type flies [86].
Although egg-laying is rhythmic in flies of a wide range of genotypes, transcripts of per, and protein levels of per and tim do not oscillate in the ovaries of Drosophila females [87]. A constitutively high level of PER and TIM proteins were found in the follicle cells of developing oocytes throughout the day. Previous studies have demonstrated that PER and TIM interact in these follicle cells but do not translocate into the nucleus, thus leaving clock mechanisms truncated [88]. Therefore, for a long time it remained unclear as to what is the functional role of the two clock genes per and tim in the fly ovary. In a recent study, Beaver and co-workers [88] quite convincingly demonstrated that lack of functional per and tim in virgin females resulted in significantly fewer mature oocytes in per0 and tim0 flies compared to the wild type flies. Rescue of clock function in per0 mutants by ectopically expressing per in the lateral neurons alone did not enhance the production of mature oocytes [88]. Thus, suggesting that per and tim may have non-clock functions in the ovaries [88]. Fitness components and circadian phenotypes are both multigenic traits, and the underlying genes may have pleiotropic effects. Therefore, it is fair to speculate that mutations affecting clock may simultaneously reduce reproductive fitness via mechanisms that are independent of clock-related processes. Alternatively, manipulations in certain genes or processes closely related to fitness traits may also alter clock phenotype, through clock independent processes, thus leaving the issues related to adaptive significance of circadian clocks via reproductive output unresolved.
Resonating clocks
The ubiquitous presence of circadian clocks in a wide variety of phenomena and organisms suggests that they confer adaptive advantage to their owners, perhaps by enabling the organisms to synchronize to LD cycles, and thereby to maintain appropriate phase relationships between their external and internal cycles. Based on this logic, it was speculated that at the advent of early life forms on this planet several temporal patterns were present in living organisms, but only those which matched environmental periodicity managed to survive. Motivated by this thought, Pittendrigh and Bruce [89] proposed a hypothesis known as the "circadian resonance hypothesis", which states that organisms with clocks having periodicities matching those of cyclic environment perform "better" compared to those whose periodicities do not match the period of the environmental cycles. If circadian resonance were the driving force behind the evolution of circadian clocks, one would expect organisms with near-24 h periodicity to have greater fitness advantage under a 24 h environment than any other periodic or aperiodic environment. Indeed, fruit flies (D. melanogaster) lived significantly longer under 24 h LD cycles than either in 21 h (LD 10.5:10.5 h), 27 h (LD 13 5:13.5 h) LD cycles or under LL [90]. In blowflies (Phormia terranovae), lifespan of flies that were reared under 24 h LD cycles, was significantly greater under 24 h LD cycles than under any other non-24 h LD environment [91]. In a separate study on the per mutants of D. melanogaster, lifespan of male perT (τ = 16 h), and perL (τ = 29 h) flies was significantly reduced compared to the wild type flies even under short and long LD cycles [85], thus contradicting the tenets of circadian resonance hypothesis. The reproductive output in many organisms bears an inverse relationship with lifespan. Inferences on fitness advantage based upon lifespan data alone could therefore be misleading, and hence multiple fitness components should be taken into account to assess adaptive significance of circadian clocks [92-96]. The most convincing and perhaps the only unequivocal demonstration of circadian resonance came from extensive and elegant studies on cyanobacteria Synechococcus elongatus [6]. In this study the growth rates of various strains of cyanobacteria having different circadian periodicities were assayed while competing against each other. Under pure culture conditions in LL, all strains showed similar growth rates. The wild type (τ = 25 h) and two strains having mutations in the KaiC gene (τ = 23 h and τ = 30 h) were competed against each other in various combinations. When two strains were mixed in approximately equal proportions and cultured under LD cycles of 11:11 h and 15:15 h, strains whose clock period matched closely that of the LD cycle always out-competed strains whose clock periods were far removed from those of the LD cycles [6] (Figure 6). These results were reexamined in cyanobacterial strains having mutations on any of the three Kai genes (Kai A, KaiB and KaiC). The mutant strains displayed circadian periods ranging between 22 h to 30 h, and in competition experiments, strains whose periodicity matched those of the LD cycles out-competed others whose periods were far removed. Thus, fitness advantages conferred to cyanobacteria via circadian resonance appear to be independent of the genotype but depend upon clock period alone [97]. On the other hand, when mutant strains with dampened bioluminescence rhythm (CLAb) or those showing arrhythmic bioluminescence (CLAc) were competed against the wild type strain under LD 12:12 h regime in mixed cultures, the CLAb and CLAc strains lost to wild-type strains, but out-competed them under LL regime [97] (Figure 7), suggesting that circadian clocks may not be beneficial under all environments and, in fact, may even be deleterious under constant conditions. The authors argued that rhythmic suppression of photosynthesis under LL in the wild type strain probably makes them maladaptive compared to the arrhythmic strains that can photosynthesize continuously under LL. It is quite unlikely that rhythmic photosynthesis in the wild type strain could be maladaptive under LL, as continuous photosynthesis in arrhythmic strains may adversely affect oxygen labile nitrogen-fixation reaction making them no better than the rhythmic strains. As we have seen from the above studies, the results on adaptive significance of circadian rhythm accrued via circadian resonance are mostly conflicting and suggestive, but occasionally conclusive.
Figure 6 Circadian resonance in cyanobacteria. Rhythmic strains having different free-running periods were competed under LD cycles of different lengths. Strains whose free-running period matched that of LD cycles out-competed those with deviant periods. Middle panels represent initial composition of the competing strains. Values in the parenthesis indicate the free-running period of the cyanobacterial strains. (Figure modified after Ouyang et al, 1998 [6])
Figure 7 Competition between rhythmic and arrhythmic strains of cyanobacteria. Mutant strains with arrhythmic (CLAc), or dampened (CLAb) bioluminescence rhythm, as well as the rescued strains were competed against wild type strain under periodic and constant environments (LD cycles and LL, respectively). Rhythmic strains out competed the wild type strain under LD cycles, but the arrhythmic strains out competed rhythmic strains under LL. Middle panels represent initial composition of the competing strains. Values in the parenthesis indicate the free-running period of the cyanobacterial strains. (Figure modified after Woelfle et al, 2004 [97])
Clocks in the dark
An obvious corollary of circadian resonance hypothesis is that circadian clocks would be of less advantage to organisms living in constant environments such as depth of oceans, underground caves and rivers, or any such aperiodic environments [2]. Therefore, it was believed that organisms living in such seemingly timeless environments would lose the ability to measure time on a circadian scale, and the ability to entrain to periodic environmental cycles. On the contrary, circadian rhythms were found to persist in cave-dwelling fishes [98], in cave-dwelling millipedes [99], and in populations of D. melanogaster that had been reared for more than 600 generations under constant laboratory conditions [100-102]. Furthermore, in one of our recent studies we found that eclosion [103] and locomotor activity (Paranjpe et al., unpublished data) rhythms of these flies entrain to a wide range of periodic LD cycles ranging from 20 h to 28 h. In addition, these flies responded to brief light pulses by shifting the phase of their locomotor activity rhythm in a phase-dependent manner, quite similar to the wild type flies maintained under LD cycles (Paranjpe et al., unpublished data). Thus, it appears that important clock features such as period, precision, phase-relationship; phase response properties and ability to entrain to a wide range of LD cycles remain intact in organisms living in constant environments. In absence of cycling environments, where there is no apparent need to synchronize behavioral and metabolic processes with the environmental cycles, persistence of functional clocks and photo-entrainment mechanisms suggests that circadian clocks confer some "intrinsic adaptive advantage" to their owners. The intrinsic advantage of having circadian clocks is probably accrued by facilitating coordination of various internal metabolic processes within the organism [2,84]. The main focus of studies on adaptive significance of circadian rhythms so far has been to investigate extrinsic advantages of possessing circadian clocks in periodic environments, while studies on intrinsic adaptive advantages have always occupied the back seat.
Concluding remarks
Regulation of behavioural and metabolic processes on a circadian scale has traditionally been thought to be a characteristic feature of eukaryotic organization until it was demonstrated that even prokaryotes such as cyanobacteria possess circadian timing devices. Analysis of sequence data of a large number of prokaryotic genomes revealed that prokaryotic circadian clocks evolved in parallel with the geophysical history of our planet. It is believed that natural selection, multiple lateral transfers, and gene duplications and losses were the major forces that shaped the evolution of early circadian clocks. Besides the periodic biotic and abiotic forces of geophysical environment, the need to segregate metabolic processes according to optimal phases of the environmental cycles also appears to have acted as a force of natural selection that shaped circadian clocks. Irrespective of the disagreements about the forces of natural selection that acted on early clocks, there is a general agreement among circadian biologists that circadian clocks, as they exist now, may have evolved as a tool primarily adaptive to daily cycles of the natural environment. Initially several geophysical cycles may have played crucial roles in exerting selection pressure, while later, daily and seasonal changes may have further fine-tuned them.
Most studies on adaptive significance of circadian rhythms suffer from a number of drawbacks such as the lack of population-level replication, true controls and of adequate control on the genetic composition of the populations, which in many ways limit the potential insights gained from such studies. Moreover, these studies were carried out on mutant and often highly inbred animals. Besides the fact that mutants and inbred lines are likely to yield spurious genetic correlations between fitness components [104] due to genetic drift, it is hard to imagine how evolution of circadian timing systems could have taken place in terms of large changes in one particular gene. On the other hand if we assume that many genes make small contributions to circadian phenotype, then it is far more likely that such genes will be involved in the evolutionary fine-tuning of circadian clocks. Indeed, a number of quantitative trait loci (QTLs) have been identified on the mouse chromosomes 1, 6, 9, 11, 17 and 19 that can potentially contribute to the determination of period of wheel running rhythm in laboratory mice [105]. Variation in the period, phase and amplitude of 150 Arabidopsis accessions has also been attributed to QTLs in ARABIDOPSIS PSEUDO-RESPONSE REGULATOR family [106]. Such latitudinal clines in period length, phase and amplitude have been taken as evidence for adaptive significance of circadian timing system [106]. Most studies aimed at resolving issues related to the adaptive significance of circadian rhythms used lifespan as the sole indicator of fitness, though it is well known that higher reproductive output often results in early death. This suggests that one should simultaneously use multiple components of fitness to assess adaptive advantage. Finally, most studies used replication at the level of individual rather than populations, while the unit of replication in any study addressing evolutionary questions needs to be populations, not individuals. Therefore, the overwhelming impression one gets from the studies on adaptive significance of circadian rhythms is one of suggestive, but only occasionally conclusive, results. Perhaps, rigorously designed laboratory selection studies under different environmental conditions might help us examine adaptation as it occurs and the development of circadian organization associated with such adaptation.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DAP and VKS contributed equally to this article.
Acknowledgements
We thank D. Anitha, Shailesh Kumar, C. R. Akarsh, Dhanya Kumar, Shahnaz R. Lone, N. Rajanna and M. Manjesh for assistance during some of our experiments discussed in this review. We thank Roberto Refinetti and two anonymous reviewers for critically reading our manuscript.
==== Refs
DeCoursey PJ Dunlap JC, Loros JJ, DeCoursey PJ The behavioral ecology and evolution of biological timing systems Chronobiology: Biological Timekeeping 2004 Sunderland, Massachusetts, USA:Sinauer Associates, Inc. Publishers 27 66
Sharma VK Adaptive significance of circadian clocks Chronobiol Intl 2003 20 901 919 10.1081/CBI-120026099
Sharma VK Chandrashekaran MK Zeitgebers (Time cues) for biological clocks Current Sci
Pittendrigh CS Temporal organization: Reflections of a Darwinian clock-watcher Annu Rev Physiol 1993 55 17 54 10.1146/annurev.ph.55.030193.000313
Johnson CH Golden SS Circadian programs in cyanobacteria: Adaptiveness and mechanism Annu Rev Microbiol 1999 53 389 409 10547696 10.1146/annurev.micro.53.1.389
Ouyang Y Andersson CR Kondo T Golden SS Johnson CH Resonating circadian clocks enhance fitness in cyanobacteria Proc Natl Acad Sci USA 1998 95 8660 8664 9671734 10.1073/pnas.95.15.8660
Roenneberg T Daan S Merrow M The Art of Entrainment J Biol Rhythms 2003 18 183 194 12828276 10.1177/0748730403018003001
Bünning E Zur Kenntniss der endogonen Tagesrhythmik bei insekten und pflanzen Ber Dt Bot Ges 1935 53 594 623
Kalmus H Periodizität und autochronie (ideochronie) als zeitregelnde eigenschaffen der organismen Biologia Generalis 1935 11 93 114
Pittendrigh CS On temperature independence in the clock controlling emergence time in Drosophila Proc Natl Acad Sci USA 1954 40 1018 1029 16589583
Pittendrigh CS Buzzati-Traverso AA Perspectives in the study of biological clocks Perspectives in marine biology 1958 University of California Press 239 268
Fleury F Allemand R Vavre F Fouillet P Boulétreau M Adaptive significance of a circadian clock: temporal segregation of activities reduces intrinsic competitive inferiority in Drosophila parasitoids Proc R Soc Lond B 2000 267 1005 1010 10.1098/rspb.2000.1103
Daan S Tinbergen JM Young guillemots (Uria lomvia) leaving their Arctic breeding cliffs: A daily rhythm in numbers and risk Ardea 1980 67 96 100
Daan S Aschoff J Adaptive daily strategies in behavior Handbook of Behavioral Neurobiology 1981 4 Plenum Publishing Corporation 275 298
Ruby NF Dark J Heller HC Zucker I Suprachiasmatic nucleus: role in circannual body mass and hibernation rhythms of ground squirrels Brain Res 1998 782 63 72 9519250 10.1016/S0006-8993(97)01263-8
DeCoursey PJ Krulas JR Mele G Holley DC Circadian performance of suprachiasmatic nuclei (SCN)-lesioned antelope ground squirrels in a desert enclosure Physiol Behav 1997 62 1099 1108 9333206 10.1016/S0031-9384(97)00263-1
DeCoursey PJ Walker JK Smith SA A circadian pacemaker in free-living chipmunks: essential for survival? J Comp Physiol A 2000 186 169 180 10707315 10.1007/s003590050017
Bourke AFG Franks NR Social Evolution in Ants 1995 Princeton, New Jersy: Princeton University Press
Robinson GE Regulation of division of labor in insect societies Annu Rev Entomol 1992 37 637 665 1539941 10.1146/annurev.en.37.010192.003225
Moore D Angel JE Cheeseman IM Fahrbach SE Robinson GE Time keeping in the honey bee colony: integration of circadian rhythms and division of labor Behav Ecol Sociol 1998 43 147 160 10.1007/s002650050476
von Frisch K The Dance Language and Orientation of Honeybees 1967 Harvard University Press, Cambridge, Massachusetts
Toma DP Bloch G Moore D Robinson GE Changes in period mRNA levels in the brain and division of labor in honeybee colonies Proc Natl Acad Sci USA 2000 97 6914 6919 10841583 10.1073/pnas.97.12.6914
Bloch G Toma DP Robinson GE Behavioral rhythmicity, age, division of labor and period expression in the honeybee brain J Biol Rhythms 2001 16 444 456 11669418 10.1177/074873001129002123
Hölldobler B Wilson EO The Ants 1990 Berlin Heidelberg, New York: Springer-Verlag
McCluskey ES Periodicity and diversity in ant mating flights Comp Biochem Physiol 1992 103A 241 243 10.1016/0300-9629(92)90574-A
Sharma VK Lone SR Goel A Chandrashekaran MK Circadian consequences of social organization in the ant species Camponotus compressus Naturwissenschaften 2004 91 386 390 15278218
Sharma VK Lone SR Goel A Clocks for sex: loss of circadian rhythms in ants after mating? Naturwissenschaften 2004 91 334 337 15257388
Sharma VK Lone SR Mathew D Goel A Chandrashekaran MK Possible evidence for shift work schedules in the media workers of the ant species Camponotus compressus Chronobiol Int 2004 21 297 308 15332348 10.1081/CBI-120037817
Brower LP Monarch butterfly orientation: missing pieces of a magnificent puzzle J Exp Biol 1996 199 93 103 9317405
Froy O Gotter AL Casselman AL Reppert SM Illuminating the circadian clock in Monarch butterfly migration Science 2003 300 1303 1305 12764200 10.1126/science.1084874
Hoffmann K Experimental manipulation of the orientational clock in birds Cold Spring Harb Symp Quant Biol 1960 25 379 387 13715052
Ruby NF Dark J Heller HC Zucker I Ablation of suprachaismatic nucleus alters timing of hibernation in ground squirrels Proc Natl Acad Sci USA 1996 93 9864 9868 8790422 10.1073/pnas.93.18.9864
Pittendrigh CS Daan S A functional analysis of circadian pacemaker in nocturnal rodents. IV. Entrainment: pacemaker as clock J Comp Physiol A 1976 106 291 331 10.1007/BF01417859
Daan S Colin Pittendrigh, Jürgen Aschoff and the natural entrainment of circadian systems J Biol Rhythms 2000 15 195 207 10885874 10.1177/074873000129001396
Sharma VK Period responses to Zeitgeber signals stabilize circadian clocks during entrainment Chronobiol Int 2003 20 389 404 12868536 10.1081/CBI-120020422
Beersma DGM Daan S Hut RA Accuracy of circadian entrainment under fluctuating light conditions: Contributions of phase and period responses J Biol Rhythms 1999 14 320 329 10447313 10.1177/074873099129000740
Sharma VK Daan S Circadian phase and period responses to light stimuli in two nocturnal rodents Chronobiol Int 2004 19 659 670 12182494 10.1081/CBI-120005389
Pittendrigh CS Circadian rhythms and the circadian organization of living systems Cold Spring Harb Symp Quant Biol 1960 25 159 184 13736116
Sokolove PG Locomotory and stridulatory circadian rhythms in the cricket Teleogryllus commodus J Insect Physiol 1975 21 537 558 1117155 10.1016/0022-1910(75)90159-6
Christensen ND The circadian clock of Hemediena thoracica as a system of coupled oscillators M Sc Thesis 1978 University of Auckland, New Zealand
Page TL Block GD Circadian rhythmicity in cockroaches: effects of early post-embryonic development and aging Physiol Entamol 1980 5 271 281
Sheeba V Chandrashekaran MK Joshi A Sharma VK Developmental plasticity of the locomotor activity rhythm of Drosophila melanogaster J Insect Physiol 2002 48 25 32 12770129 10.1016/S0022-1910(01)00139-1
Enright JT The Timing of Sleep and Wakefullness 1980 Berlin, Springer-Verlag
Kramm KR Kramm DA Photoperiodic control of circadian activity rhythms in diurnal rodents Int J Biometeorol 1980 24 65 76 7364519
Edmunds LN Jr Cellular and Molecular Bases of Biological Clocks 1988 Berlin and New York: Springer-Verlag
Hastings JW Rusak B Boulos Z Prosser CL Circadian rhythms: The physiology of biological timing Neural and Integrative Animal Physiology 1991 Wiley-Liss, Inc 435 545
Ishiura M Kutsuna S Aoki S Iwasaki H Andersson CR Tanabe A Golden SS Johnson CH Kondo T Expression of a gene cluster KaiABC as a circadian feedback process in cyanobacteria Science 1998 281 1519 1523 9727980 10.1126/science.281.5382.1519
Dvornyk V Vinogradova O Nevo E Origin and evolution of circadian clock genes in prokaryotes Proc Natl Acad Sci USA 2003 100 2495 2500 12604787 10.1073/pnas.0130099100
Pittendrigh CS Arm DL Biological clocks Science in the Sixties 1965 Albuquerque: University of New Mexico, Office of Publications 96 111
Pittendrigh CS Biological Clocks: the functions, ancient and modern, of circadian oscillations Science and the Sixties 1965 Proc. Cloudcraft Symposium Air Force Office of Scientific Research 96 111
Emery P So WV Kaneko M Hall JC Rosbash M CRY, a Drosophila clock and light-regulated cryptochrome, is a major contributor to circadian rhythm resetting and photosensitivity Cell 1998 95 669 679 9845369 10.1016/S0092-8674(00)81637-2
Cashmore AR Jarillo JA Wu YJ Liu D Cryptochromes: Blue light receptors for plants and animals Science 1999 284 760 765 10221900 10.1126/science.284.5415.760
Aschoff J Aschoff J The phase angle of entrainment Circadian Clocks 1965 Amsterdam: North-Holland Publications 262 276
Hoffmann K Hastings JW, Schweiger HG Adaptive significance of biological rhythms corresponding to geophysical cycles The Molecular Basis of Circadian Rhythms 1976 Berlin: Abakon Verlagsgesellschaft 63 75
Stal LJ Krumbein WE Nitrogenase activity in the non-heterocystous cyanobacterium Oscillatoria sp. Grown under alternating light-dark cycles Arch Microbiol 1985 143 76 71
Mitsui A Kumazawa S Takahashi A Illmoto H Arai T Strategy by which nitrogen-fixing unicellular cyanobacteria grow photoautotrophically Nature 1986 323 720 722 10.1038/323720a0
Nikaido SS Johnson CH Daily and circadian variation in survival from ultraviolate radiation in Chlamydomonas reinhardtii Photochem Photobiol 2000 71 758 765 10857373 10.1562/0031-8655(2000)071<0758:DACVIS>2.0.CO;2
Glossop NRJ Hardin PE Central and peripheral circadian oscillator mechanisms in flies and mammals J Cell Sci 2002 115 3369 3377 12154068
Dunlap JC Molecular bases for circadian clocks Cell 1999 96 271 290 9988221 10.1016/S0092-8674(00)80566-8
Stanewsky R Genetic analysis of the circadian system in Drosophila melanogaster and mammals J Neurobiol 2003 54 111 147 12486701 10.1002/neu.10164
Glossop NRJ Houl JH Zheng H Fanny SNG Dudek SM Hardin PE VRILLE feeds back to control circadian transcription of clock in Drosophila circadian oscillators Neuron 2003 37 249 261 12546820 10.1016/S0896-6273(03)00002-3
Hall JC Molecular neurogenetics of biological rhythms J Neurogenet 1998 12 115 81 10320863
King DP Zhao Y Sangoram AM Wilsbacher LD Tanaka M Antoch MP Steeves TDL Vitaterna MH Kornhauser JM Lowrey PL Turek FW Takahashi JS Positional cloning of mouse circadian Clock gene Cell 1997 89 641 653 9160755 10.1016/S0092-8674(00)80245-7
Allada R White NE So WV Hall JC Rosbash M A mutant Drosophila homolog of mammalian Clock disrupts circadian rhythms and transcription of period and timeless Cell 1998 93 791 804 9630223 10.1016/S0092-8674(00)81440-3
Kloss B Price JL Saez L Blau J Rothenfluh A Wesley C Young MW The Drosophila clock gene double-time encodes a protein closely related to human casein kinase I-epsilon Cell 1998 94 97 107 9674431 10.1016/S0092-8674(00)81225-8
Rosato E Kyriacou CP Flies, clocks and evolution Phil Trans R Soc Lond B 2001 356 1769 1778 11710984 10.1098/rstb.2001.0961
Piccin A Couchman M Clayton JD Chalmers D Costa R Kyriacou CP The clock gene period of house fly, Musca domestica, rescues behavioral rhythmicity in Drosophila melanogaster. Evidence for intermolecular coevolution? Genetics 2000 154 747 758 10655226
Pando MP Pinchak AB Cermakian N Sassone-Corsi P A cell-based system that recapitulates the dynamic light-dependent regulation of the vertebrate clock Proc Natl Acad Sci USA 2001 98 10178 83 11517315 10.1073/pnas.181228598
Krishnan B Levin JD Lynch MKS Dowse HB Funes P Hall JC Hardin PE Dryer SE A new role for cryptochrome in a Drosophila circadian oscillator Science 2001 411 313 317
Mittag M Kiaulehn S Johnson CH The circadian clock in Chlamydomonas reinhardtii : What is it for? What is it similar to? Plant Phyisol 2005 137 399 409 10.1104/pp.104.052415
Pittendrigh CS Takamura T Latitudinal clines in the circadian properties of a pacemaker J Biol Rhythms 1989 4 217 235 2519590
Lankinen P Genetic variation of circadian eclosion rhythm, and its relation to photoperiodism in Drosophila littoralis PhD Thesis 1985 University of Oulu, Department of Genetics
Pittendrigh CS Kyner WT Takamura T The amplitude of circadian oscillations: Temperature dependence, latitudinal clines and the photoperiodic time measurement J Biol rhythms 1991 6 299 313 1773097
Costa R Peixoto AA Barbujani G Kyriacou CP A latitudinal cline in a Drosophila clock gene Proc R Soc Lond B 1992 250 43 49
Rosato E Peixoto AA Barbujani G Costa R Kyriacou CP Molecular polymorphism in the period gene of Drosophila simulans Genetics 1994 138 693 707 7851767
Sawyer LA Hennessy JM Peixoto AA Rosato E Parkinson H Costa R Kyriacou CP Natural variation in a Drosophila clock gene and temperature compensation Science 1997 278 2117 2120 9405346 10.1126/science.278.5346.2117
Kyriacou CP Oldroyd M Wood J Sharp M Hill M Clock mutations alter developmental timing in Drosophila Heredity 1990 64 395 401 2113515
Hurd MW Ralph MR The significance of circadian organization for longevity in the golden hamster J Biol Rhythms 1998 13 430 436 9783234 10.1177/074873098129000255
Oklejewicz M Daan S Enhanced longevity in tau mutant Syrian hamsters, Mesocricetus auratus J Biol Rhythms 2002 17 210 216 12054192 10.1177/07430402017003004
Paranjpe DA Anitha D Chandrashekaran MK Joshi A Sharma VK Possible role of eclosion rhythm in mediating the effects of light-dark environments on pre-adult development in Drosophila melanogaster BMC Dev Biol 2005 5 5 15725348 10.1186/1471-213X-5-5
Miyatake T Pleiotropic effect, clock gene, and reproductive isolation Popul Ecol 2002 44 201 207 10.1007/s101440200023
Miyatake T Correlated responses to selection for developmental period in Bactrocera cucurbitae (Diptera: Tephritidae): time of mating and daily activity rhythms Behav Genet 1997 27 489 498 9336086 10.1023/A:1025682618895
Shimizu T Miyatake T Watari Y Arai T A gene pleiotropically controlling developmental and circadian periods in the melon fly Bactrocera cucurbitae (Diptera: Tephritidae) Heredity 1997 79 600 605 10.1038/sj.hdy.6882610
Sharma VK Joshi A Kumar V Clocks, genes and evolution: the evolution of circadian organization Biological Clocks 2002 New Delhi: Narosa Publishers and Berlin: Springer-Verlag 5 23
Klarsfeld A Rouyer F Effects of circadian mutations and LD periodicity on the life span of Drosophila melanogaster J Biol Rhythms 1998 13 471 478 9850008 10.1177/074873098129000309
Beaver LM Gvakharia BO Vollintine TS Hege DM Stanewsky R Geibultowiz JM Loss of circadian clock function decreases reproductive fitness in males of Drosophila melanogaster Proc Natl Acad Sci USA 2002 99 2134 2139 11854509 10.1073/pnas.032426699
Hardin PE Analysis of period mRNA cycling in Drosophila head and body tissues indicates that body oscillators behave differently from head oscillators Mol Cell Biol 1994 14 7211 7218 7935436
Beaver LM Rush BL Gvakharia BO Geibultowiz JM Noncircadian regulation and function of clock genes period and timeless in oogenesis of Drosophila melanogaster J Biol Rhythms 2003 18 463 472 14667147 10.1177/0748730403259108
Pittendrigh CS Bruce V Withrow RB Daily rhythms as coupled oscillator systems and their relation to thermoperiodism and photoperiodism Photoperiodism and Related Phenomenon in Plants and Animals 1959 Washington DC: American Association of Advancement of Sciences 475 505
Pittendrigh CS Minis DH Circadian systems: Longevity as a function of circadian resonance in Drosophila melanogaster Proc Natl Acad Sci 1972 69 1537 1539 4624759
Von Saint Paul U Aschoff J Longevity among blowflies Phormia terraenovae R.D. kept in non-24-hour light-dark cycles J Comp Physiol A 1978 127 191 195 10.1007/BF01350109
Bell G Evolutionary and non-evolutionary theories of senescence Am Nat 1984 124 600 603 10.1086/284300
Partridge L Harvey PH Cost of reproduction Nature 1985 316 20 21 10.1038/316020a0
Stearns SC The Evolution of Life Histories 1992 New York: Oxford University Press
Zwaan BJ The evolutionary genetics of aging and longevity Heredity 1999 82 589 597 10383679 10.1046/j.1365-2540.1999.00544.x
Sheeba V Sharma VK Shubha K Chandrashekaran MK Joshi A The effect of different light regimes on adult life span in Drosophila melanogaster is partly mediated through reproductive output J Biol Rhythms 2000 15 380 392 11039916 10.1177/074873000129001477
Woelfle MA Ouyang Y Phanvijhitsiri K Johnson CH The adaptive value of circadian clocks: An experimental assessment in cyanobacteria Curr Biol 2004 14 1481 1486 15324665 10.1016/j.cub.2004.08.023
Trajano E Manno-Baretto L Free running locomotor activity rhythms in cave dwelling catfishes Trichomycterus sp. From Brazil Biol Rhythms Res 1996 27 329 335 10.1076/brhm.27.3.329.12958
Koilraj AJ Sharma VK Marimuthu G Chandrashekaran MK Presence of circadian rhythms in the locomotor activity of a cave dwelling millipede Glyphiulus cavernicolus sulu (Cambalidae, Spirostreptida) Chronobiol Int 2000 17 757 765 11128292 10.1081/CBI-100102111
Sheeba V Sharma VK Chandrashekaran MK Joshi A Persistence of eclosion rhythms in populations of Drosophila melanogaster after 600 generations in an aperiodic environment Naturwissenschaften 1999 86 448 449 10501695 10.1007/s001140050651
Sheeba V Chandrashekaran MK Joshi A Sharma VK Persistence of oviposition rhythm in individuals of Drosophila melanogaster reared in an aperiodic environment for several hundred generations J Exp Zool 2001 47 1217 1225
Sheeba V Chandrashekaran MK Joshi A Sharma VK Locomotor activity rhythm in Drosophila melanogaster after 600 generations in an aperiodic environment Naturwissenschaften 2002 89 512 514 12451454 10.1007/s00114-002-0360-9
Paranjpe DA Anitha D Kumar S Kumar D Verkhedkar K Chandrashekaran MK Joshi A Sharma VK Entrainment of eclosion rhythm in Drosophila melanogaster populations reared for more than 700 generations in constant light environment Chronobiol Int 2003 20 977 987 14680138 10.1081/CBI-120025247
Mueller LD Ayala FJ Fitness and density-dependent population growth in Drosophila melanogaster Genetics 1981 97 667 677 6795084
Hofstetter JR Possidente Mayeda AR Provisional QTL for circadian period of wheel running in laboratory mice: Quantitative genetics of period in RI mice Chronobiol Int 1999 16 269 279 10373097
Michael TP Salome PA Yu HJ Spencer TR Sharp EL McPeek MA Alonso JM Ecker JR McClung CR Enhanced fitness conferred by naturally occurring variation in the circadian clock Science 2003 302 1049 1053 14605371 10.1126/science.1082971
| 15869714 | PMC1142335 | CC BY | 2021-01-04 16:39:14 | no | J Circadian Rhythms. 2005 May 4; 3:7 | utf-8 | J Circadian Rhythms | 2,005 | 10.1186/1740-3391-3-7 | oa_comm |
==== Front
Lipids Health DisLipids in Health and Disease1476-511XBioMed Central London 1476-511X-4-101588245710.1186/1476-511X-4-10ReviewRole of cholesterol in parasitic infections Bansal Devendra [email protected] Harinderpal Singh [email protected] Rakesh [email protected] Department of Parasitology, Post Graduate Institute of Medical Education & Research, Chandigarh, India2 Central Research Institute, Kasauli, Himachal Pradesh, India2005 9 5 2005 4 10 10 16 3 2005 9 5 2005 Copyright © 2005 Bansal et al; licensee BioMed Central Ltd.2005Bansal 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.
The requirement of cholesterol for internalization of eukaryotic pathogens like protozoa (Leishmaniasis, Malaria and Toxoplasmosis) and the exchange of cholesterol along with other metabolites during reproduction in Schistosomes (helminths) under variable circumstances are poorly understood. In patients infected with some other helminthes, alterations in the lipid profile have been observed. Also, the mechanisms involved in lipid changes especially in membrane proteins related to parasite infections remain uncertain. Present review of literature shows that parasites induce significant changes in lipid parameters, as has been shown in the in vitro study where substitution of serum by lipid/cholesterol in medium and in experimental models (in vivo). Thus changes in lipid profile occur in patients having active infections with most of the parasites. Membrane proteins are probably involved in such reactions. All parasites may be metabolising cholesterol, but the exact relationship with pathogenic mechanism is not clear. So far, studies suggest that there may be some factors or enzymes, which allow the parasite to breakup and consume lipid/cholesterol. Further studies are needed for better understanding of the mechanisms involved in vivo. The present review analysis the various studies till date and the role of cholesterol in pathogenesis of different parasitic infections.
ProtozoaHelminthsPathogenesisLipids/Cholesterol
==== Body
Introduction
Parasitic protozoa and helminthes are responsible for some of the most devastating and prevalent diseases of humans, threatening the lives of nearly one-third of the worldwide human population leading to more than 2 million deaths annually. Habitats of parasites are extremely varied and common parasites of man (protozoa, helminthes and arthropods) normally inhabit the intestine, blood, liver, lungs brain, muscles and lymphatic tissues [1]. Many species of parasites have complex life cycles involving developmental stages that live in soil or water, or use various kinds of intermediate hosts, including vertebrates and invertebrates and cold and warm-blooded animals. In such varied environments, parasites have become adapted to using/tolerating widely differing oxygen, carbon dioxide, hydrogen ion concentrations and temperatures [1]. Their nutritional requirements and their means of obtaining and utilizing the nutrients required for growth, motility and reproduction are also varied. The requirement of cholesterol for internalization of eukaryotic pathogens under such variable circumstances is poorly understood.
The present review highlights the role of lipids and their metabolical mechanisms in protozoan and helminthic infections.
Cholesterol is a major constituent of eukaryotic membranes and plays a crucial role in cellular membrane organization, dynamics, function and sorting. It is often found distributed non-randomly in domains in membranes [2]. Recent observations suggest that cholesterol exerts many of its actions by maintaining a specialized type of membrane domain, termed "lipid rafts" in a functional state. Lipid rafts are enriched in cholesterol and sphingolipids, and have been thought to act as platform through which signal transduction events are coordinated and pathogens gain entry to infect host cells [3].
Relationship of serum cholesterol levels in man infected with parasites has drawn the attention of various workers. Since it has been shown in-vitro studies that parasites like Giardia and Entamoeba can grown in lipid rich media in the absence of serum, it would be interesting to determinate the mechanism of lipid/cholesterol utilization. Recent studies have shown elevated levels of lipoproteins like high density lipoprotein (HDL), low density lipoprotein (LDL) and total cholesterol in patients suffering from parasitic infection (4). In human body cholesterol is synthesized in liver, which incidentally happens to be a major extraintestinal site of infection with Entamoeba histolytica. Keeping this in view, the following queries arise (i) is there any correlation between cases of amoebic liver abscess (ALA) with cholesterol synthesis in liver? (ii) in case of intestinal amoebiasis, what is the source of cholesterol? (iii) how do the amoebae utilize cholesterol in these two different areas (iv) what is the role of cholesterol in enhancing virulence and pathogenecity of E. histolytica (v) does cholesterol help in cyst formation?
Entamoeba, Giardia and trichomonads form what can be usually termed as 'mucosal parasites' which lack mitochondria, well developed golgi complexes, and other organelles typical of higher eukaryotes [1]. These parasites have developed unique metabolic pathways that allow them to survive and multiply by scavenging nutrients from the host, but are unable to synthesize the majority of their own lipids and cholesterol de novo [5]. Therefore, an understanding how they transport and utilize exogenous lipids for metabolic purposes is extremely important? There is evidence suggesting that these parasites can take up the lipids and cholesterol they need from lipoprotein particles present in the host gut and tissue in vivo and from the growth medium in vitro. Exogenous phospholipids have been shown to undergo fatty acid remodeling [by deacylation/reacylation reactions], which allow these protozoa to alter lipids, bypassing the synthesis of entirely new phospholipid molecules [5].
The endocytic process is essential for uptake of nutrients and other molecules by protozoan parasites. Membrane dynamics in endocytosis has been investigated in parasites such as Leishmania and Trypanosoma. The receptor-mediated endocytosis of LDL by Schistosoma mansoni has been reported [6]. A putative LDL-receptor has been identified on the tegument and gut lining from Schistosoma japonicum adult worms [7] and schistosomula of S. mansoni [8]. Recently, Coopens et al [9] demonstrated that the intracellular parasite, Toxoplasma gondii, acquires host cholesterol that is endocytosed by the LDL pathway, a process that is specifically increased in infected cells. Interference with LDL endocytosis or cholesterol translocation reduced the intracellular survival of T. gondii [9]. In a classic experiment, Aley et al [10] demonstrated the phenomenon of endocytosis through non-acidified vesicles in E. histolytica. They showed that these non-acidified vesicles (distinct from intracellular acidified vesicles) were derived from plasma membranes and played important roles in endocytosing nutrients and other biological molecules into cells.
Cholesterol – biosynthesis and metabolism
Cholesterol is an amphipathic lipid and as such is an essential structural component of all cell membranes and of the outer layer of plasma lipoproteins. It is present in tissues and in plasma lipoprotein either as free cholesterol or, combined with a long-chain fatty acid, as cholesteryl ester. It is synthesized in many tissues from acetyl-CoA and is ultimately eliminated from the body in the bile as cholesterol or bile salts [11]. Lipoprotein transports free cholesterol in the circulation, where it readily equilibrates cholesterol in other lipoproteins and in membranes. Cholesteryl ester is a stored form of cholesterol found in most tissues. It is transported as cargo in the hydrophobic core of lipoproteins [12]. So, the fact that cholesterol in a hydrophobic molecule, which resides in lipoproteins and cell membranes, raises two questions: (i) how does the cell sense the level of cholesterol? (ii) how in this cholesterol specific signaling transmitted to the nucleus for the regulation of various genes?
Recent studies directed to resolve these questions, led to the discovery of a novel cell surface cholesterol-sensor designated as receptor-Ck which was not only shown to be ubiquitously present in various human organs but also [through its signaling pathway] regulated various genes involved in cholesterol homeostasis [HMG CoA synthase; HMG CoA reductase; apo 'B'- specific LDL – receptor] cell growth [cyclin 'D'; C – fos; C-myc; p27 etc]; cell death [Bc1-2] through a 47 KDa transcription factor [derived from the cleavage of 125 KDa SREBP] having affinity for genomic sterol regulatory element [SRE] sequence as well as through other transcription factors [13-17].
Anaerobic parasites
Entamoeba
Entamoeba histolytica parasitizes the gastro-intestinal tract of humans and is a major cause of morbidity and mortality in tropical and subtropical countries [18]. Although several successive life cycle stages of E. histolytica have been documented, briefly, it can be classified into two main morphologic forms, i.e. trophozoites and cysts. In E. histolytica, excystation occur in the small intestine, where four amoebae are released from the mature quadri nucleate cyst. Trophozoites dwell in the colon, where they multiply and encyst typically producing four nucleated cyst [19]. Cholesterol has been reported to be a growth – promoting factor of E. histolytica [20]. The virulence of a virulent strain could be revived in vitro by adding cholesterol to medium or in vivo by feeding cholesterol to experimental animals or the host [21-23]. Cholesterol is thought to act as an irritant to mucus membrane and thereby helps the amoebae to establish and colonize at the injured site and this enhances the virulence [24]. In-vitro study has shown that substitution of lipid/cholesterol instead of serum, supported vigorous growth of E. histolytica in axenic culture [25].
Recently, Laughlin et al [26] have shown that the disruption of cholesterol rich raft-like membrane domains in Entamoeba with the cholesterol-binding agent filipin and methyl-β-cyclodextrin inhibit several important virulence functions, fluid phase pinocytosis and adhesion to host cell monolayers. However, disruption of raft-like domains did not inhibit constitutive secretion of cysteine proteases, another important virulent function of Entamoeba. Cysteine proteinases in Entamoeba are likely to be associated with tissue invasion and pathogenesis [27]. Andra et al [28], revealed that the lipid composition of amoebic membranes prevents binding of the cytolytic molecules and that both the phospholipid ingredients and the high content of cholesterol contributes to the protection of the toxin-producing cell.
More recently the authors have assessed the impact of lipid parameter in patients infected with E. histolytica and E. dispar. The result showed significant lower levels of lipid profile [total cholesterol, HDL & LDL] in E. histolytica and E. dispar cyst passers and ALA patients as compared to healthy controls. Also, cyst passers had lower levels of cholesterol than ALA cases [29].
Giardia
Giardiasis is the most common waterborne disease in human, which is caused by an enteric flagellated protozoan Giardia lamblia. Giardia is widespread with children being the most vulnerable [30]. Giardia exists in two morphologic forms: trophozoite and cysts. In G. lamblia, excystation is accomplished in two steps first by limiting the acidic conditions present in the stomach and secondly by the protease-rich and slightly alkaline small intestine. Encystation, in turn, can be induced in vitro by starving the trophozoites of cholesterol, either by using lipoprotein-deficient serum or augmenting the bile concentration in the culture medium [31,32]. Membrane biogenesis in Giardia requires cholesterol [33,34]. Because Giardia is unable to synthesize cholesterol [33] it obtains the same from upper small intestine, which is rich in biliary and dietary cholesterol [35,36]. Effects of bile salts on encystations are directly related to the uptake of cholesterol by the trophozoites [37].
In vitro, it was observed that replacing bovine serum with a lipoprotein cholesterol (LPC) solution and bovine serum albumin in pre-encystation and excystation media, stimulated G.1amblia encystations and encystation specific secretory vesicles (ESV) formation [38]. It is established that cholesterol-dependent down-regulation of encystations-specific CWP-1 gene expression has contributed to the inhibition of Giardia encystation process [31,37]. Previously, we have reported that receptor Ck dependent signaling is responsible for the regulation of Giardia encystations process by cholesterol [39]. Recently we have observed that patients infected with G. lamblia showed lower levels of lipid parameters (total cholesterol, HDLc and LDLc) as compared to control healthy group [29].
Trichomonas
Trichomonas vaginalis is a sexually transmitted protozoan parasite, adheres to the vaginal epithelium, causing vaginitis and other complications in women [40]. It is an anaerobic protozoan flagellate, which lacks mitochondria and peroxisomes, but has a specialized double-membrane-bounded organelle called the hydrogenosome, which is involved in metabolic processes that extend glycolysis [41]. In vitro study has shown that when serum was replaced by bovine serum albumin and cholesterol, it resulted in good growth [42].
Apicomplexan parasites
Plasmodium
Plasmodium falciparum is the most pathogenic species causing human malaria. Erythrocytes become infected following attachment and invasion by merozoite. Various phases of parasite development can be observed during the 48 h – erythrocytic cycle i.e. rings, trophozoites and schizonts. In vitro the parasite grows in 5–10% human serum in an atmosphere of low oxygen [43]. Very little is known about mechanism involved in lipid changes related to malaria. Hypercholesterolemia and hypertriglyceridemia was observed in both uncomplicated and complicated malaria [44-46], whereas Kittl et al [47], have shown no correlation between severity of malaria attacks and extent of HDL – cholesterol decrease. Human serum HDL is necessary for P. falciparum in in vitro culture. However, it has been reported that HDL can be toxic for the parasite at high concentrations [48].
Recently, Imrie et al [49], have also reported that in the absence of serum, HDL in low concentration (0.75 mg/ml) supported growth of P. falciparum in vitro, whereas at high concentration (3 mg/ml), it was toxic to the parasite. Recent findings, however, would suggest that the plasmodium genome contains genes encoding enzymes of phospholipids metabolism, allowing de novo synthesis of phosphatidyl choline via the knneddy pathway and necessitating only the uptake of the small choline molecule [50]. In addition, the genome of P. falciparum contains genes similar to those for type II fatty acid synthesis pathway. The protein products of these genes are located within the apicoplast and allow for the production of fatty acids, some of which are unique to the parasite [50]. Thus the parasite may be able to meet many of its lipid requirements from its own biosynthetic pathways, although some extracellular lipids are necessary for in vitro growth. It has also been seen that plasma membrane cholesterol plays a role in the pathogenesis of immune evasion and clinical manifestations of falciparum malaria [51].
There are few studies which suggest that there are changes in lipid plasma or serum levels in-vivo after infection. But, no significant changes were seen in the plasma cholesterol during and after infection of malaria [52]. However, low levels of the cholesterol in patients infected with malaria as compared to normal healthy controls have been reported [53]. In another study changes in plasma lipoprotein was seen in acute malaria resulting decreased levels of HDL and LDL and moderately increased triglycerides [54]. In malaria endemic areas, when plasma levels of cholesterol, triglcerides, HDLc and LDLc were analysed in children infected with P. falciparum, investigators have found significantly low levels of lipid profile [55]. Brotons et al [56] reviewed that population studies on common lipid parameters and observed that cholesterol values are lower in Africa, where, malaria is endemic, than in many other parts of the world.
Toxoplasma
Toxoplasma gondii, an apicomplexan protozoan parasite, is an important pathogen of humans and animals. It is widely distributed with a very high prevalence in many regions, and can cause serious infections in immuno-compromised patients (particularly AIDS patients) and in the developing fetus [57]. Host cell cholesterol is implicated in the entry and replication of an increasing number of intracellular microbial pathogens. However, recently new mechanism has been described by which host cholesterol specifically controls entry of an intracellular pathogen. Briefly, the parasitophorus vacuole membrane (PVM) surrounding T. gondii contains cholesterol. At the time of cell entry host plasma membrane cholesterol is incorporated into the forming PVM during invasion, through a caveolae independent mechanism. Depleting host cell plasma membrane cholesterol blocks parasite internalization by reducing the release of rhoptry proteins that are necessary for invasion [58].
In another study, Coppens et al [9] had shown that T. gondii exploits host low density lipoprotein receptor-mediated endocytosis for cholesterol acquisition. Whereas, acyl-CoA: cholesterol acyl transferase (ACAT) and cholesterol esters play a crucial role in the optimal replication of T. gondii [59]. These studies indicate the cholesterol does have a role in the pathogenesis of Toxoplasmosis.
Cryptosporidium
Heterogeneous distribution of membrane cholesterol at the attachment site of Cryptosporidium muris to host cells has already been investigated. Although many filipin-cholesterol complexes were observed on the plasma membrane of host cells and parasites, a line showing no complexes was evident at the above two membrane junctures. These observations indicate that parasitic infection of C. muris altered the organization of membrane cholesterol [60]. The exact role of cholesterol in pathogenesis/virulence of parasite needs to be determined.
Kinetoplastid parasites
Leishmania
Leishmania is an obligate intracellular parasite that infects macrophages of the vertebrate host, resulting in visceral, cutaneous and mucocutaneous leishmaniasis in humans. Recently Pucadyil [61] reported that plasma membrane cholesterol is required for efficient attachment and internalization of the parasite in macrophages, leading to Leishmania donovani infection. Rodrigues et al [62] have shown that when amastigotes and promastigotes forms of L. amzonensis are incubated with 22,26-azasterol, which is a delta (24(25))-sterol methyltransferase (SMT) inhibitor, it results in arrest of growth. They also observed that alteration occur in the lipid composition of the parasite membrane, resulting in loss of viability. The study suggests the use of 24-SMT inhibitor could be used as selective antileishmanial agent. Dietz et al [63] have reported successful treatment of Brazilian kala-azar using lipid encapsulated amphotericin B. Thus, a level of cholesterol in the patients infected with leishmaniasis has to be determined.
Trypanosoma
African trypanosomes are lipid auxotrophs that live in the bloodstream of their human and animal hosts and are unable to synthesize cholesterol but appear to bind and take up plasma low-density lipoproteins (LDL) from their host [64]. Whether cholesterol homeostasis of this unicellular parasite also requires interactions with host high-density lipoprotein (HDL) particles is unknown. Trypanosomes require lipoproteins to multiply under axenic culture conditions [65]. Recently, Green et al [66] reported that HDL, LDL, and trypanosome lytic factor (TLF1) were bound and taken up by a lipoprotein scavenger receptor, which may constitute the parasite's major pathway mediating the uptake of essential lipids. Frequent turnover of variable surface glycoproteins is a well established method of immune evasion by Trypanosomes. Does it involve the role of HDL, LDL, and TLF1 in getting eliminated by attaching to the surface receptors of Trypanosomes?
Helminthes
Schistosoma
It has been shown that Schistosome infection could be counteracting the effects of an atherogenic diet by modulating host lipid metabolism and inducing a reduction in blood total cholesterol concentration [67]. Little information is available on the lipid changes caused by S. mansoni reinfection. Popiel et al [68], studied the metabolism of cholesterol uptake by paired and unpaired worms of S. mansoni during pairing, the results showed labeled worms lost upto 65% of their cholesterol. This suggests that normal cholesterol transfer in worm pairs is bi-directional and that it is facilitated by physical contact between juxtaposed membranes. Cholesterol exchange in schistosome worm pairs may be partly or wholly consequences of normal tegumental turnover of the molecule.
In another study Silveira et al [69] investigated the transfer of cholesterol and its metabolites between adult male and female worms of S. mansoni. They found that the adult male and female worms of S. mansoni are able to incorporate cholesterol and convert it into several metabolites. On the other hand, Schistosomula cannot convert the incorporated cholesterol. However, a significant reduction in levels of serum lipid profile was observed in mice infected with S. mansoni [70]. These changes might be attributed to several metabolites released by S. mansoni, which affect the host hepatic tissue resulting in decreased synthesis of these parameters and their release into the circulation. The ability of Schistosomula to convert cholesterol into its metabolites, shows that it is a property acquired, used by adults only and is specifically active during pairing. What is the exact role of cholesterol during pairing needs to be investigated. What happens to cholesterol levels in serum of patients also needs to be studied.
Intestinal worms
In-vivo study has shown the decreased serum lipid levels in the Shipibo population (Peru), showed a significant inverse correlation between worm egg excretion and HDL levels in hookworm, Strongyloides and Trichuris infected patients but not in Ascaris infected cases [71]. The mechanisms underlying the observed association between intestinal worm load and HDL reduction are not completely understood and may include reduced HDL synthesis in the gut wall due to inflammatory toxic irritation. Hence, cholesterol may have a role in pathogenesis by helping the larvae to survive in the host tissue.
Ascaris
During larval ascariasis the metabolism of lipids is significantly disturbed. Decreased levels of total cholesterol, HDL cholesterol and triglycerides were observed in guinea pigs. The changes are due to the break in liver function and, presumably, changes in hormone secretion, which are provoked by the presence of the parasite [72]. It has also been seen that cholesterol enhanced larval survival and the yield and growth of L4 of A. summ larvae when added to RPMI-1640 culture medium [73].
Ancylostoma
In vitro A. duodenalis grown axenically in medium supplemented with cholesterol gives good growth [74]. Phospholipid/cholesterol ratio in liver plasma membrane of the infected golden hamsters with A. ceylanicum group was significantly reduced, which suggest that both the structural and functional organization of membrane may be a biochemical basis of the hepatotoxic effects [75].
Filaria
Decreased levels of lipid contents were observed in liver of Mastomys natalensis during the development of B. malayi infection [76]. The lipid peroxide formation was enhanced in liver during the development of filarial infection. The studies in human beings are lacking.
Hymenolepis
Balb/c mice infected with Hymenolepis microstoma significantly affected the lipid metabolism [77]. The changes were in relation to nutritional interactions between host and parasite and the possible effect on host hormone levels. Johnson et al [78] studied on the mechanisms of cholesterol uptake by the rat tapeworm H. diminuta. The results support the hypothesis that the tapeworm absorbs cholesterol by a specific carrier-mediated process. Cholesterol uptake is reduced when the capacity of the micellar phase of the medium is increased, suggesting that uptake involves the intermediate partitioning of sterol from micelles into the aqueous phase of the medium [79].
Conclusion
The mechanisms involved in lipid changes related to parasite infections remain uncertain. Cholesterol starvation initiates encystations. This indicates that cholesterol has a role in pathogenesis as it helps the parasite to remain in trophozoite stage. However, many workers have investigated that medium containing lipoproteins supported the growth of the parasites in vitro. It is possible that all these parasites use LDL-like receptors to endocytose various lipoprotein particles. However, typical receptor binding affinity, number of LDL-receptors exposed on cell surfaces, receptor internalizations etc. in mucous-dwelling parasites are not known. Few population based studies showed altered lipid profiles in patients infected with malaria compared to asymptomatic control group. So, to evaluate the impact of malaria on lipid parameters at the population levels in different age groups, studies are needed to explore the relevance of this finding in different patterns of hyperendemic regions. The association of parasites and the contents of their membranes before penetration/invasion and realignments seen after attachment need to be studied. Also, the role of PVM and their contents in protection of parasites within PVM also need to be studied.
So far, the studies suggest that there may be some factors or enzymes, which allow the parasite to breakup and consume lipid/cholesterol. Further studies are needed for better understanding of the mechanisms involved in vivo.
Future strategies
The evidence suggests that parasites are able to remodel/metabolize host lipids for their growth and to generate phospholipids membrane. So, the identification and characterization of more enzymes involved in these pathways will be more important for the proper understanding of these steps. Recent progress in molecular biology and parasite genome projects will assist researchers in the near future to identify the genes and enzymes of lipid metabolic pathways.
==== Refs
Beaver PC Jung RC Cupp EW Clinical Parasitology Lea & Febiger, 600 Washington Square Philadelphia, PA 19106 USA 1984
Maxfield FR Plasma membrane microdomains Curr Opin Cell Biol 2002 14 483 87 12383800 10.1016/S0955-0674(02)00351-4
Simons K Toomre D Lipid rafts and signal transduction Nat Rev Mol Cell Biol 2000 1 31 9 11413487 10.1038/35036052
Faucher JF Milama EN Missinou MA Ngomo R Kombila M Kremsner PG The impact of malaria on common lipid parameters Parasitol Res 2002 88 1040 43 12444452 10.1007/s00436-002-0712-6
Das S Stevens T Castillo C Villasenor A Arredondo H Reddy K Lipid metabolism in mucous-dwelling amitochondriate protozoa Int J Parasitol 2002 32 655 57 12062485 10.1016/S0020-7519(02)00006-1
Rumjanek FD Campos EG Afonso LC Evidence for the occurrence of LDL receptors in extracts of schistosomula of Schistosoma mansoni Mol Biochem Parasitol 1988 28 145 52 3367933 10.1016/0166-6851(88)90062-X
Rogers MV Henkle KJ Fidge NH Mitchell GF Identification of a multispecific lipoprotein receptor in adult Schistosoma japonicum by ligand blotting analyses Mol Biochem Parasitol 1989 35 79 88 2548094 10.1016/0166-6851(89)90145-X
Xu X Caulfield JP Characterization of human low density lipoprotein binding proteins on the surface of schistosomula of Schistosoma mansoni Eur J Cell Biol 1992 57 229 35 1324838
Coppens I Sinai AP Joiner KA Toxoplasma gondii exploits host low-density lipoprotein receptor-mediated endocytosis for cholesterol acquisition J Cell Biol 2000 149 167 80 10747095 10.1083/jcb.149.1.167
Aley SB Cohn ZA Scott WA Endocytosis in Entamoeba histolytica. Evidence for a unique non-acidified compartment J Exp Med 1984 160 724 37 6206186 10.1084/jem.160.3.724
Fredrickson DS Levy RI Stanbury JB, Wyngaarden JB, Fredrickson DS Familial hyperlipoproteinemia In the metabolic basis of inherited disease 1972 3 New York, McGraw-Hill Book co
Tietz NW Fundamentals of Clinical Chemistry 1994 W.B. Saunders Company Philadelphia, USA
Kaul D Do human platelets possess LDL-receptor specific for apoprotein 'B' or cholesterol? Curr Sci 1993 65 883 86
Kaul D Receptor-Ck and leukemogenesis Leuk Res 1998 22 389 94 9652724 10.1016/S0145-2126(98)00007-1
Kaul D Kaur M Receptor-Ck regulates the expression of cyclin-dependent kinase inhibitors Mol Cell Biochem 1999 200 183 86 10569199 10.1023/A:1006981614054
Kaur M Kaul D Sobti RC Receptor-Ck dependent regulation of genes involved in the cell cycle Mol Cell Biochem 1998 181 137 42 9562250 10.1023/A:1006890215502
Kaul D Khosla VK Molecular basis of 'cholesterol feedback lesion' in CNS tumors Nerol India 2000 48 174 77
Walsh JA Problems in recognition and diagnosis of amoebiasis: estimation of the global magnitude of mortality and morbidity Rev Infect Dis 1986 8 228 38 2871619
Martinez-Palomo A Martinez-Baez M Selective primary health care: strategies for control of disease in the developing world X amoebiasis Rev Infect Dis 1983 5 1093 02 6140740
Sharma R Effect of cholesterol on the growth and virulence of E. histolytica Trans Roy Soc Trop Med Hyg 1959 53 278 13669033 10.1016/0035-9203(59)90008-2
Vinayak VK Chakravarti RN Agarwal KC Naik SR Chuttani PN Pathogenicity of Entamoeba histolytica – effect of cholesterol on the virulence of strains of amoebae Indian J Med Res 1978 67 545 52 210121
Bhatti HS Bhushnurmath S Mahajan RC Ganguly NK Sehgal R An experimental model of ameboma in guinea pig Exp Parasitol 1992 74 283 289 1582480 10.1016/0014-4894(92)90151-Y
Singh BN Srivastava RVN Dutta GP Virulence of strain of E. histolytica to rats and the effect of cholesterol, rat caecal and hamster liver passage on the virulence of non-invasive strains Indian J Expt Biol 1971 9 21
Gargouri ML Utilisation du cholesterol bans I'amibiase experimental du cobaye Ann parasitol Hum Comp 1967 42 399 4294818
Mata-Cardenas BD Morales-Vallarta M Vargas-Villarreal J Said-Fernandez S PACSR*: a serum replacement for axenic cultivation of Entamoeba histolytica Trans Roy Soc Trop Med Hyg 1996 90 586 8944281 10.1016/S0035-9203(96)90334-X
Laughlin RC McGuan GC Powell RR Welter BH Temesvari LA Involvement of raft-like plasma memebrane domains of Entamoeba histolytica in pinocytosis and adhesion Infect Immun 2004 72 5349 57 15322032 10.1128/IAI.72.9.5349-5357.2004
Que X Reed SL Cysteine proteinases and the pathogenesis of amebiasis Clin Microbiol Rev 2000 13 196 206 10755997 10.1128/CMR.13.2.196-206.2000
Andra J Berninghausen O Leippe M Membrane lipid composition protects Entamoeba histolytica from self-destruction by its pore-forming toxins FEBS Lett 2004 23 109 15 15094050 10.1016/S0014-5793(04)00324-2
Bansal D Bhatti HS Sehgal R Altered lipid parameters in patients infected with Entamoeba histolytica, Entamoeba dispar and Giardia lamblia Br J Biomed Sci
Adam RD Biology of Giardia lamblia Clin Microbiol Rev 2001 14 447 75 11432808 10.1128/CMR.14.3.447-475.2001
Lujan HD Mowatt MR Nash TE The mechanisms of Giardia lamblia differentiation into cysts Microbiol Mol Biol Rev 1997 61 294 304 9293183
Gillin FD Reiner DS McCaffery M Cell biology of the primitive eukaryote Giardia lamblia Annu Rev Microbiol 1996 50 679 705 8905095 10.1146/annurev.micro.50.1.679
Jarrol EL Muller PJ Meyer EA Morse SA Lipid and carbohydrate metabolism of Giardia lamblia Mol Biochem Parasitol 1981 2 187 96 6783899 10.1016/0166-6851(81)90099-2
Kaneda Y Goutsu T Lipid analysis of Giardia lamblia and its culture medium Ann Trop Med Parasitol 1985 82 83 90 3401072
Field FJ Kam NTP Mathur SN Regulation of cholesterol metabolism in the intestine Gastroenterology 1990 99 539 51 2194897
Thomson ABR Schoeller C Keelan M Smith L Clandinin MT Lipid absorption: passing through the unstirred layers, brush border membrane and beyond Can J Physiol Pharmacol 1993 71 531 55 8306192
Lujan HD Mowatt MR Nash TE Lipid requirements and lipid uptake by Giardia lamblia trophozoites in culture J Euk Microbiol 1996 43 237 42 8640194
Reiner DS Hetsko ML Gillin FD A lipoprotein-cholesterol-albumin serum substitute stimulates Giardia lamblia encystations vesicle formation J Euk Microbiol 1995 42 622 27 7581338
Kaul D Rani R Sehgal R Receptor-Ck regulates giardia encystations process Mol and Cell Biochem 2001 225 167 69 11716359 10.1023/A:1012277001794
Fouts AC Kraus SJ Trichomonas vaginalis : reevaluation of its clinical presentation and laboratory diagnosis J Infect Dis 1980 141 137 143 6965976
Bui ET Johnson PJ Identification and characterization of [Fe]-hydrogenases in the hydrogenosome of Trichomonas vaginalis Mol Biochem Parasitol 1996 76 305 310 8920017 10.1016/0166-6851(96)02567-4
Linstead D New defined and semi-defined media for cultivation of the flagellate Trichomonas vaginalis Parasitology 1981 81 125 37 6973740
The biology of malaria parasites: Report of a WHO scientific group WHO Technical Report Series 1987
Das BS Thurnham DI Das DB Plasma α-tocopherol, retinol, and carotenoids in children with falciparum malaria Am J Clin Nutr 1996 64 94 100 8669421
Davis TME Sturm M Zhang YR Spancer JL Graham RM Li GQ Taylor RR Platelet-activating factor and lipid metabolism in acute malaria J Infect 1993 26 279 85 8505562
Mohanty S Mishra SK Das BS Satpathy SK Mohanty D Patnaik JK Bose TK Altered plasma lipid pattern in falciparum malaria Ann Trop Med Parasitol 1992 86 601 06 1304701
Kittl EM Diridl G Lenhart V Neuwald C Tomasits J Pichler H Bauer k HDL-cholesterol as a sensitive diagnostic criterion in malaria Wien Klin Wochenschr 1992 104 21 24 1546481
Grellier P Rigomier D Clavey V Fruchart JC Schrevel J Lipid traffic between high density lipoproteins and Plasmodium falciparum -infected red blood cells J Cell Biol 1991 112 267 77 1988461 10.1083/jcb.112.2.267
Imrie H Ferguson DJP Carter M Drain J Schiflett A Hajdur SL Day KP Light and electron microscopical observations of the effects of high-density lipoprotein on growth of Plasmodium falciparum in vitro Parasitol 2004 128 577 84 10.1017/S0031182004005025
Vial HJ Eldin P Tielens AG Vanhellmond JJ Phospholipids in parasitic protozoa Mol Biochem Parasitol 2003 126 143 54 12615313 10.1016/S0166-6851(02)00281-5
Sein KK Aikawa M The prime role of plasma membrane cholesterol in the pathogenesis of immune evasion and clinical manifestations of falciparum malaria Med Hypotheses 1998 51 105 10 9881815 10.1016/S0306-9877(98)90102-5
Onongbu IC Onyeneke EC Plasma lipid changes in human malaria Tropenmed Parasitol 1983 34 193 96 6356530
Djoumessi S Serum lipids and lipoproteins during malaria infection Pathol Biol 1989 37 909 11 2694079
Nilson EI Nilson EP Changes in plasma lipoproteins in acute malaria J Intern Med 1990 227 151 55 2179459
Baptisa JL Vervoort T Vander SP Wery M Changes in plasma lipid levels as a function of Plasmodium falciparum infection in Sao Tome Parasite 1996 3 335 40 9033909
Brotons C Ribera A Perich RM Abrodos D Magana P Pablo S Terradas D Fernadez F Permanyer G Worldwide distribution of blood lipids and lipoproteins in childhood and adolescence: a review study Atherosclerosis 1997 139 1 9 9699886 10.1016/S0021-9150(98)00056-2
Holliman RE Gordon Cook Toxoplasmosis: Manson's tropical diseases 1996 20 London: WB Saunders 1246 1254
Coppens I Joiner KA Host but not parasite cholesterol controls Toxoplasma cell entry by modulating organelle Mol Bio Cell 2003 14 3804 20 12972565 10.1091/mbc.E02-12-0830
Sonda S Ting LM Novak S Kim K Maher JJ Farese RV Ernst JD Cholesterol esterification by host and parasite is essential for optimal proliferation of Toxoplasma gondii J Biol Chem 2001 276 34434 40 11457847 10.1074/jbc.M105025200
Yoshikawa H Kimata I Iseki M Heterogenous distribution of membrane cholesterol at the attachment site of Cryptosporidium muris to host cells J Eukaryot Microbiol 1997 44 454 56 9304814
Pucadyil TJ Tewary P Madhubala R Chattopadhyay A Cholesterol is required for Leishmania donovani infection: implications in leishmaniasis Mol Biochem Parasitol 2004 133 145 52 14698427 10.1016/j.molbiopara.2003.10.002
Rodrigues JC Attias M Rodriguez C Urbina JA Souza W Ultrastructural and biochemical alterations induced by 22,26-azasterol, a delta(24(25))-sterol methyltransferase inhibitor, on promastigote and amastigote forms of Leishmania amazonensis Antimicrob Agents Chemother 2002 46 487 99 11796362 10.1128/AAC.46.2.487-499.2002
Dietze R Milan EP Berman JD Grogl M Falqueto A Feitosa TF Luz KG Suassuna FA Marinho LA Ksionski G Treatment of Brazilian kala-azar with a short cource of amhocil (amphotericin B cholesterol dispersion) Clin Infect Dis 1993 17 981 86 8110956
Gillett MP Owen JS Characteristics of the binding of human and bovine high-density lipoproteins by bloodstream forms of the African trypanosome, Trypanosoma brucei brucei Biochim Biophys Acta 1992 1123 239 48 1536861
Coppens I Baudhuin P Opperdoes FR Courtoy PJ Receptors for the Host Low Density Lipoproteins on the Hemoflagellate Trypanosoma brucei: Purification and Involvement in the Growth of the Parasite Proc Natl Acad Sci USA 1988 85 6753 57 3413122
Green HP Del Pilar Molina Portela M St Jean EN Lugli EB Raper J Evidence for a Trypanosoma brucei lipoprotein scavenger receptor J Biol Chem 2003 278 422 7 12401813 10.1074/jbc.M207215200
Doenhoff MJ Stanley RG Griffiths K Jackson CL An anti-atherogenic effect of Schistosoma mansoni infections in mice associated with a parasite-induced lowering of blood total cholesterol Parasitology 2002 125 415 21 12458825 10.1017/S0031182002002275
Popiel I Bash PF Schistosoma mansoni : cholesterol uptake by paired and unpaired worms Exp Parasitol 1986 61 343 47 3709750 10.1016/0014-4894(86)90189-X
Silveira AM Friche AA Rumjanek FD Transfer of (14C) cholesterol and its metabolites between adult male and female worms of Schistosoma mansoni Comp Biochem Physiol B 1986 85 851 57 3816158 10.1016/0305-0491(86)90186-0
el-Marzouki ZM Amin AM Changes in serum lipids of mice experimentally infected with Schistosoma mansoni Egypt Soc Parasitol 1997 27 419 29
Wiedermann U Stemberger H Unfried E Widhalm K Kundli M Altenriederer M Savedra M Wiedermann G Intestinal worm burden and serum cholesterol or lipid concentration in a Shipibo population (Peru) Zentrabl Bakteriol 1991 275 279 86
Biadun W Studies of serum lipids in guinea pigs with larval ascariasis Wiad Parazytol 1990 36 15 26 2092676
Urban JF Douvres FW Xu S Culture requirement of Ascaris suum larvae using a stationary multi-well system: increased survival, development and growth with cholesterol Vet Parasitol 1984 14 33 42 6538364 10.1016/0304-4017(84)90131-6
el-Naggar HM Abou-Zakham AA Romia SM Hegazi MM el-Khouly ES Axenic growth in vitro of Ancylostoma duodenale to the filariform (resting) larvae J Egypt Soc Parasitol 1991 21 127 34 2033286
Srivastava SC Liver plasma membrane-bound enzymes and lipids in golden hamsters infected with Ancylostoma ceylanicum Int J Parasitol 1994 24 249 51 7913086 10.1016/0020-7519(94)90032-9
Joshi A Saxena JK Murthy PK Sen AB Ghatak S Brugia malayi: status of host during different stages of infection Folia Parasitol 1989 36 169 75 2527791
Rath EA Walkey M Fatty acid and cholesterol synthesis in mice infected with the tapeworm Hymenolepis microstoma Parasitol 1987 95 79 92
Johnson WJ Cain GD Studies on the mechanism of cholesterol uptake by the rat tapeworm Hymenolepis diminuta (Cestoda) Comp Biochem Physiol B 1988 91 59 67 3197395 10.1016/0305-0491(88)90114-9
Johnson WJ Cain GD The selective uptake of cholesterol uptake by the rat tapeworm Hymenolepis diminuta (Cestoda) Comp Biochem Physiol B 1988 91 51 58 3197394 10.1016/0305-0491(88)90113-7
| 15882457 | PMC1142336 | CC BY | 2021-01-04 16:39:19 | no | Lipids Health Dis. 2005 May 9; 4:10 | utf-8 | Lipids Health Dis | 2,005 | 10.1186/1476-511X-4-10 | oa_comm |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.