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
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2631625964210.1186/1471-2105-6-263Methodology ArticleSelecting additional tag SNPs for tolerating missing data in genotyping Huang Yao-Ting [email protected] Kui [email protected] Ting [email protected] Kun-Mao [email protected] Department of Computer Science and Information Engineering, National Taiwan University, Taipei, Taiwan2 Graduate Institute of Networking and Multimedia, National Taiwan University, Taipei, Taiwan3 Section on Statistical Genetics, Department of Biostatistics, University of Alabama at Birmingham, USA4 Department of Biological Sciences, University of Southern California, Los Angeles, CA 90089, USA2005 1 11 2005 6 263 263 26 5 2005 1 11 2005 Copyright © 2005 Huang et al; licensee BioMed Central Ltd.2005Huang 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 studies have shown that the patterns of linkage disequilibrium observed in human populations have a block-like structure, and a small subset of SNPs (called tag SNPs) is sufficient to distinguish each pair of haplotype patterns in the block. In reality, some tag SNPs may be missing, and we may fail to distinguish two distinct haplotypes due to the ambiguity caused by missing data.
Results
We show there exists a subset of SNPs (referred to as robust tag SNPs) which can still distinguish all distinct haplotypes even when some SNPs are missing. The problem of finding minimum robust tag SNPs is shown to be NP-hard. To find robust tag SNPs efficiently, we propose two greedy algorithms and one linear programming relaxation algorithm. The experimental results indicate that (1) the solutions found by these algorithms are quite close to the optimal solution; (2) the genotyping cost saved by using tag SNPs can be as high as 80%; and (3) genotyping additional tag SNPs for tolerating missing data is still cost-effective.
Conclusion
Genotyping robust tag SNPs is more practical than just genotyping the minimum tag SNPs if we can not avoid the occurrence of missing data. Our theoretical analysis and experimental results show that the performance of our algorithms is not only efficient but the solution found is also close to the optimal solution.
==== Body
Background
In recent years, Single Nucleotide Polymorphisms (SNPs) have become the preferred marker for association studies of genetic diseases or traits. A set of linked SNPs on one chromosome is called a haplotype. Recent studies have shown that the patterns of Linkage Disequilibrium (LD) observed in human populations have a block-like structure [4,13]. The chromosome recombination only takes place at some low LD regions called recombination hotspots. The high LD region between these hotspots is often referred to as a "haplotype block." Within a haplotype block, there is little or even no recombination occurred, and the SNPs in the block tend to be inherited together. Due to the low haplotype diversity within a block, the information carried by these SNPs is highly redundant. Thus, a small subset of SNPs (called "tag SNPs") is sufficient to distinguish each pair of patterns in the block [7,13,17-19]. Haplotype blocks with corresponding tag SNPs are quite useful and cost-effective for association studies as it does not require genotyping all SNPs. Many studies have tried to find the minimum set of tag SNPs in a haplotype block. In a large-scale study of human Chromosome 21, Patil et al. [13] developed a greedy algorithm to partition the haplotypes into 4,135 blocks with 4,563 tag SNPs. Zhang et al. [17-19] used a dynamic programming approach to reduce the numbers of blocks and tag SNPs to 2,575 and 3,562, respectively. Bafna et al. [1] showed that the problem of minimizing tag SNPs is NP-hard and gave efficient algorithms for special cases of this problem.
Figure 1 The influence of missing data and auxiliary tag SNPs. (A) A haplotype block defined by 12 SNPs and 4 haplotype patterns. Each column represents a haplotype pattern and each row represents a SNP locus. The black and grey boxes stand for the major and minor alleles at each SNP locus, respectively. (B) Tag SNPs genotyped without missing data. (C) Tag SNPs genotyped with missing data. (D) The auxiliary tag SNP S5 for h2. (E) The auxiliary tag SNP S8 for h3.
In reality, a SNP may not be genotyped and considered to be missing data (i.e., we fail to obtain the allele configuration of the SNP) if it does not pass the threshold of data quality [13,16,19,20]. These missing data may cause ambiguity when using the minimum set of tag SNPs to distinguish an unknown haplotype sample. Figure 1 illustrates the influence of missing data when identifying haplotype samples. In this figure, a haplotype block (see Figure 1 (A)) defined by 12 SNPs and 4 haplotype patterns is presented (from the public haplotype data of human Chromosome 21 [13]). We follow the same assumption as previous studies that all SNPs are diallelic (i.e., taking on only two values) [1,13]. Suppose we select SNPs S1 and S12 as tag SNPs. The haplotype sample h1 is identified as haplotype pattern P3 unambiguously (see Figure 1 (B)). Consider haplotype samples h2 and h3 with one missing tag SNP (see Figure 1 (C)). h2 can be identified as haplotype patterns P2 or P3, and h3 can be identified as P1 or P3. As a result, these missing tag SNPs result in ambiguity when distinguishing unknown haplotype samples.
Although we can not avoid the occurrence of missing data, the remaining SNPs within the haplotype block may provide abundant information to resolve the ambiguity. For example, if we re-genotype an additional SNP S5 for h2 (see Figure 1 (D)), h2 is identified as haplotype pattern P3 unambiguously. On the other hand, if SNP S8 is re-genotyped (see Figure 1 (E)), h3 is also identified unambiguously. These additional SNPs are referred to as "auxiliary tag SNPs," which can be found from the remaining SNPs in the block and are able to resolve the ambiguity caused by missing data.
Alternatively, instead of re-genotyping auxiliary tag SNPs whenever encountering missing data, we work on a set of SNPs which is not affected by the occurrence of missing data. Figure 2 illustrates a set of SNPs which can tolerate one missing SNP. Suppose we select SNPs S1, S5, S8, and S12 to be genotyped. Note that no matter which SNP is missing, each of the 16 missing patterns can be distinguished by the remaining three SNPs. Therefore, all haplotype samples with one missing SNP can still be identified unambiguously. We refer to these SNPs as "robust tag SNPs," which are able to tolerate a number of missing data. The important feature of robust tag SNPs is that although they consume more SNPs than the "tag SNPs" defined in previous studies, they guarantee that all haplotype samples with a number of missing data can be distinguished unambiguously. When the occurrence of missing data is frequent, the cost of re-genotyping processes can be reduced by robust tag SNPs.
Figure 2 The robust tag SNPs. A set of robust tag SNPs for tolerating one missing tag SNP.
This paper focuses on the problem of finding robust tag SNPs to tolerate a number of missing data. Throughout this paper, we denote m as the maximum number of missing SNPs to be tolerated, which corresponds to different missing rates in different genotyping experiments. And we wish to find a minimum set of robust tag SNPs which can distinguish each pair of haplotypes even when up to m SNPs are missing. We assume that the haplotype phases and block partition are available as the input. Numerous methods have been developed to infer haplotypes from genotype data [12,14,15]. Several algorithms have also been proposed to find the block partition [4,13,17]. The problem of finding minimum robust tag SNPs is shown to be NP-hard (See Theorem 1). To find robust tag SNPs efficiently, we propose two greedy algorithms and one linear programming (LP) relaxation algorithm. The proposed algorithms have been implemented and tested on a variety of simulated and empirical data. We also analyze the efficiency and solutions of these algorithms. An algorithm for finding auxiliary tag SNPs is described assuming robust tag SNPs have been computed in advance.
Results
We propose two greedy algorithms which select the robust tag SNPs one by one in different greedy manners. In addition, we reformulate this problem as an integer programming problem and design an LP-relaxation algorithm to solve this problem. The greedy and LP-relaxation algorithms are able to find solutions within factors of (m + 1) lnK(K−1)2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGSbaBcqqGUbGBdaWcaaqaaiabdUealjabcIcaOiabdUealjabgkHiTiabigdaXiabcMcaPaqaaiabikdaYaaaaaa@363F@, ln((m+1)K(K−1)2)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGSbaBcqqGUbGBcqqGOaakcqqGOaakcqWGTbqBcqGHRaWkcqaIXaqmcqGGPaqkdaWcaaqaaiabdUealjabcIcaOiabdUealjabgkHiTiabigdaXiabcMcaPaqaaiabikdaYaaacqGGPaqkaaa@3CD6@, and O(m ln K) of the optimal solution respectively, where m is the maximum number of missing SNPs allowed and K is the number of haplotype patterns in the block.
We have implemented the first and second greedy algorithms in JAVA [see Additional files 1 and 2]. The LP-relaxation algorithm has been implemented in Perl [see Additional file 3], where the LP problem is solved via a program called "lp_solve" [11]. The LP-relaxation algorithm is a randomized method. Thus, this program is repeated for 10 times to explore different solutions and the best solution among them is chosen as the output.
In order to evaluate the solutions and efficiency of our algorithms, we also implement a program in JAVA (referred to as "OPT") which uses a brute force method to find the optimal solution. For a given data set of N SNPs, the OPT program examines all possible solutions (i.e., all subsets of (N1)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaqadaqaauaabeqaceaaaeaacqWGobGtaeaacqaIXaqmaaaacaGLOaGaayzkaaaaaa@3059@, (N2)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaqadaqaauaabeqaceaaaeaacqWGobGtaeaacqaIYaGmaaaacaGLOaGaayzkaaaaaa@305B@, ..., and (NN)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaqadaqaauaabeqaceaaaeaacqWGobGtaeaacqWGobGtaaaacaGLOaGaayzkaaaaaa@308E@). The minimum subset of SNPs that can tolerate m missing SNPs is chosen as the output. Due to the NP-hardness of this problem, the OPT program fails to output the optimal solution within a reasonable period of time in many data sets. As a consequence, we skip some impossible solution space to speed up this program by the following two observations: (1) the solutions with less than or equal to m SNPs are the impossible ones since m SNPs might be missing; and (2) for a data set containing K haplotype patterns, the minimum number of SNPs required to distinguish each of them is at least log K (see Lemma 2). As a result, we can examine the possible solutions only for subsets of (Nm+log K)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaqadaqaauaabeqaceaaaeaacqWGobGtaeaacqWGTbqBcqGHRaWkcyGGSbaBcqGGVbWBcqGGNbWzcaaMc8UaaGPaVlabdUealbaaaiaawIcacaGLPaaaaaa@3A01@, (Nm+log K+1)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaqadaqaauaabeqaceaaaeaacqWGobGtaeaacqWGTbqBcqGHRaWkcyGGSbaBcqGGVbWBcqGGNbWzcaaMc8UaaGPaVlabdUealjabgUcaRiabigdaXaaaaiaawIcacaGLPaaaaaa@3BD3@ ..., and (NN)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaqadaqaauaabeqaceaaaeaacqWGobGtaeaacqWGobGtaaaacaGLOaGaayzkaaaaaa@308E@. By searching possible solutions from small subsets to large ones, the OPT program can stop and output the optimal solution immediately when a subset that can tolerate m missing SNPs is found.
Results on simulated data
Theoretically, all SNPs will reach complete linkage equilibrium after sufficient chromosome recombination takes place. We first generate 100 data sets containing short haplotypes which simulate this bottleneck model [12,14,15]. Each data set consists of 10 haplotypes with 20 SNPs. These haplotypes are created by randomly assigning the major or minor alleles at each SNP locus. Let m be the number of missing SNPs allowed and Sa be the average number of robust tag SNPs over 100 data sets. Figure 3 (a) plots Sa with respect to m (roughly corresponding to SNP missing rates from 0% to 33%). When m = 0, all programs find the same number of SNPs as the optimal solution. The iterative LP-relaxation algorithm slightly outperforms the others as m increases. When m > 6, more than 20 SNPs are required to tolerate missing data. Thus, no data sets contain enough SNPs for solutions.
Figure 3 Experimental results on random data. (a) Results from data sets containing 10 haplotypes and 20 SNPs. (b) Results from data sets containing 10 haplotypes and 40 SNPs.
We then generate 100 data sets containing long haplotypes. Each data set is composed of 10 haplotypes with 40 SNPs. Figure 3 (b) illustrates the experimental results on these long data sets (corresponding to SNP missing rates from 0% to 37%). The optimal solutions for m > 2 can not be found by the OPT program within a reasonable period of time (after one week computation) and are not shown in this figure. It is because the possible solutions in long data sets are too large to enumerate. On the other hand, both greedy and iterative LP-relaxation algorithms run in polynomial time and always output a solution efficiently. In this experiment, both greedy algorithms slightly outperforms the iterative LP-relaxation algorithm. In addition, the number of missing SNPs allowed is larger than those in short data sets. For example, to tolerate 10 missing SNPs (i.e., m = 10), all programs output less than 28 SNPs. The remaining SNPs in each data set are still sufficient to tolerate more missing SNPs.
Hudson (2002) [10] provides a program which can simulate a set of haplotypes under the assumption of neutral evolution and uniformly distributed recombination rate using the coalescent model. We use Hudson's program to generate 100 short data sets with 10 haplotypes and 20 SNPs and 100 long data sets with 10 haplotypes and 40 SNPs. Figure 4 (a) shows the experimental results on Hudson's short data sets (corresponding to SNP missing rates from 0% to 23%). The number of missing SNPs allowed are less than that of random data. It is because Hudson's program generates coalescent haplotypes which are similar to each other. As a result, many SNPs can not be used to distinguish haplotypes and the amount of tag SNPs is inadequate to tolerate larger missing SNPs. In this experiment, we observe that the iterative LP-relaxation algorithm finds solutions quite close to the optimal solutions and slightly outperforms the other two algorithms.
Figure 4 Experimental results on Hudson's data. (a) Results from data sets containing 10 haplotypes and 20 SNPs. (b) Results from data sets containing 10 haplotypes and 40 SNPs.
Figure 4 (b) illustrates the experimental results on long data sets generated by Hudson's program (corresponding to SNP missing rates from 0% to 29%). The optimal solutions for m > 2 again can not be found by the OPT program within a reasonable period of time. In this experiment, the performance of the first greedy and iterative LP-relaxation algorithms are similar, and they slightly outperform the second greedy algorithm as m becomes large.
Results on real data
We also test these programs on two real data sets: (1) public haplotype data of human Chromosome 21 released by Patil et al. [13]; and (2) a 500 KB region on human Chromosome 5q31 which may contain a genetic variant related to the Crohn disease by Daly et al. [4]. Patil's data include 20 haplotypes of 24,047 SNPs spanning over about 32.4 MB, which are partitioned into 4,135 haplotype blocks. By genotyping 103 SNPs with minor allele frequency at least 5%, Daly et al. partition the 500 KB region into 11 haplotype blocks. Each haplotype block in these real data sets contains different numbers of SNPs and haplotypes (e.g., from several SNPs to hundreds of SNPs). When m increases, some short blocks may not contain enough SNPs for tolerating missing data (e.g., m > the number of SNPs in a block). As a consequence, Sa here stands for the average number of robust tag SNPs over those blocks still containing solutions.
Figure 5 (a) shows the experimental results on Patil's 4,135 blocks. Because there are many long blocks in Patil's data (e.g., more than one hundred SNPs), the optimal solution for m > 2 can not be found within a reasonable period of time. The experimental result indicates that all algorithms find similar number of robust tag SNPs when m is small. The LP-relaxation algorithm slightly outperforms the others as m increases.
Figure 5 Experimental results on real data. (a) Results from Patil's Chromosome 21 data, (b) Results from Daly's Chromosome 5q31 data.
Figure 5 (b) illustrates the experimental results on Daly's 11 blocks. Because the haplotype blocks partitioned by Daly et al. are very short (e.g., most blocks contain less than 12 SNPs), all optimal solutions still can be found. The solutions found by each algorithm is almost the same as optimal solutions. Theoretically, Sa should grow monotonically as m increases. But due to the small number of blocks in Daly's data set, Sa does not grow smoothly when m increases from 2 to 3. To explain this phenomenon, we report the detailed result of the first greedy algorithm in Table 1. For each of the 11 blocks, the number of robust tag SNPs found with respect to different values of m is listed in the table. Note that as mentioned before, some blocks may not contain enough SNPs for tolerating large missing data as m increases. When m increases from 2 to 3, Blocks 4 and 10 (which consumes 8 and 5 SNPs) do not contain enough SNPs for a solution and are discarded. As a result, Sa (for m = 3) is computed only using Blocks 1 and 2 and the value is lower than the previous one (i.e., from 4.75 to 4). This phenomenon is not shown in Figure 5 (a) because it is amortized by thousands of blocks in Patil's data set.
Table 1 The detailed result of first greedy algorithm on Daly's 11 blocks.
Block ID 1 2 3 4 5 6 7 8 9 10 11 Sa
m = 0 1 1 2 3 3 2 3 2 2 2 2 23/11 = 2.09
m = 1 2 2 f 5 f 3 5 4 f 3 3 27/8 = 3.375
m = 2 3 3 f 8 f f f f f 5 f 19/4 = 4.75
m = 3 4 4 f f f f f f f f f 8/2 = 4
m = 4 5 5 f f f f f f f f f 10/2 = 5
m = 5 6 f f f f f f f f f f 6/1 = 6
m = 6 7 f f f f f f f f f f 7/1 = 7
f: fail to contain enough SNPs for tolerating m missing SNPs
Discussion
In terms of efficiency, the first and second greedy algorithms are faster than the LP-relaxation algorithm. The greedy algorithms usually returns a solution in seconds and the LP-relaxation algorithm requires about half minute for a solution. It is because the running time of LP-relaxation algorithm is bounded by the time of solving the LP problem. Furthermore, this LP-relaxation algorithm is repeated for 10 times to explore 10 different solutions. The OPT program for searching the optimal solution is apparently slower than the others. The optimal solution usually can not be found within a reasonable period of time if the size of the block becomes large. ¿From our empirical study, the optimal solution can be found in reasonable time by the OPT program if the block contains less than 20 SNPs (e.g., the short random data sets). But for those large data sets with more than 40 SNPs, the OPT program is significantly outperformed by the approximation algorithms (e.g., fail to output a solution within one week computation).
Assuming no missing data (i.e., m = 0), we compare the solutions found by each algorithm with the optimal solution. Table 2 lists the numbers of total tag SNPs found by each algorithm in previous experiments. In the experiments on random and Daly's data, the solution found by each algorithm is as good as the optimal solution. In the experiments on Hudson's and Patil's data, these algorithms still find solutions quite close to the optimal solution. For example, the approximation ratios of these algorithms are only 472443≈1.07
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabisda0iabiEda3iabikdaYaqaaiabisda0iabisda0iabiodaZaaacqGHijYUcqaIXaqmcqGGUaGlcqaIWaamcqaI3aWnaaa@37F1@ and 46574595≈1.01
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabisda0iabiAda2iabiwda1iabiEda3aqaaiabisda0iabiwda1iabiMda5iabiwda1aaacqGHijYUcqaIXaqmcqGGUaGlcqaIWaamcqaIXaqmaaa@39EB@, respectively.
Table 2 The number of total tag SNPs found by each algorithm. The percentage of tag SNPs with respect to total SNPs is shown in parentheses.
Random data Hudson's data Patil's data Daly's data
Total blocks 100 100 100 100 4135 11
Total SNPs 2000 4000 2000 4000 24047 103
1st Greedy 400 (20%) 400 (10%) 509 (25.5%) 472 (11.8%) 4610 (19.2%) 23 (22.3%)
2nd Greedy 400 (20%) 400 (10%) 509 (25.5%) 472 (11.8%) 4610 (19.2%) 23 (22.3%)
LP-relaxation 400 (20%) 400 (10%) 509 (25.5%) 471 (11.8%) 4657 (19.4%) 23 (22.3%)
OPT 400 (20%) 400 (10%) 492 (24.6%) 443 (11.1%) 4595 (19.1%) 23 (22.3%)
We then analyze the genotyping cost that can be saved by using tag SNPs. In Table 2, the percentage of tag SNPs in each data set is shown in parentheses. The experimental results indicate that the cost of genotyping tag SNPs is significantly reduced in comparison with genotyping all SNPs in a block. For example, in Patil's data, we only need to genotype about 19% of tag SNPs in each block, which saves about 81% genotyping cost.
The tradeoffs between the number of additional tag SNPs required and the number of missing SNPs allowed are discussed in the following. In practice, missing data in the genotyping experiment are usually limited to certain missing rate. We transform the maximum number of missing SNPs allowed into maximum missing rates allowed by calculating the percentage of m with respect to the number of robust tag SNPs. Table 3 lists the results of the first greedy algorithm applied on random and Hudson's long data sets. The number of additional tag SNPs grows with respect to m linearly. However, we observe that the maximum missing rate allowed grows slowly as m becomes large. This is because more additional tag SNPs are required in order to tolerate more missing SNPs. But under the same SNP missing rate, genotyping these additional tag SNPs may also increase the number of missing SNPs, which reduces the power of robust tag SNPs. On the positive side, when m is small, the corresponding maximum missing rate allowed is sufficient for most genotyping experiments since their missing rates are usually less than 10%. For example, the robust tag SNPs with m = 1 are sufficient to tolerate 10% missing SNPs, and they only requires at most 3 additional SNPs. As a result, genotyping additional tag SNPs for tolerating missing data is cost-effective under the current genotyping environment.
Table 3 The tradeoffs between additional tag SNPs required and maximum missing rates allowed. These results come from the first greedy algorithm applied on random and Hudson's data sets with 40 SNPs.
m 0 1 2 3 4 5
Random data (40 SNPs) average number of robust tag SNPs 4 6 8.51 10.47 12.89 14.92
corresponding SNP missing rate 0 16.7% 23.5% 28.6% 31.0% 33.5%
average number of extra tag SNPs 0 2 4.51 6.47 8.89 10.92
Hudson's data (40 SNPs) average number of robust tag SNPs 4.72 7.71 11.28 14.67 18.23 21.67
corresponding SNP missing rate 0 13.0% 17.7% 20.4% 21.9% 23.1%
average number of extra tag SNPs 0 2.99 6.56 9.95 13.51 16.95
In reality, not all haplotypes are of equal importance or confidence. When selecting robust tag SNPs, it might be desirable to weight them according to their population frequency. To incorporate the frequency of haplotypes into this problem, there are two possible ways:
1. It can be easily done by discarding the rare haplotypes and retain the common haplotypes as the input of our algorithms. This approach would not require modification to our algorithms. But the retained common haplotypes will be processed as equally weighted.
2. Our algorithms try to find a set of SNPs such that each pair of haplotypes are distinguished by a threshold of at least (m + 1) SNPs. A simplest way to weight the haplotypes is choosing different thresholds for each pair of haplotypes according to their population frequency. The haplotype pairs with higher frequency can then be assigned with more tag SNPs than the lower ones by our algorithms.
Conclusion
In this paper, we show there exists a set of robust tag SNPs which is able to tolerate a number of missing data. Our study indicates that genotyping robust tag SNPs is more practical than genotyping minimum tag SNPs for association studies if we can not avoid the occurrence of missing data. We describe two greedy and one LP-relaxation approximation algorithms for finding robust tag SNPs. Our experimental results and theoretical analysis show that these algorithms are not only efficient but the solutions found are also close to the optimal solution. In terms of genotyping cost, we observe that the genotyping cost saved by using robust tag SNPs is significant, and genotyping additional tag SNPs to tolerate missing data is still cost-effective. One future direction is to assign weights to different types of SNPs (e.g., SNPs in coding or non-coding regions), and design algorithms for the selection of weighted tag SNPs.
Software availability
Project name: efficient algorithms for utilizing SNP information.
Project home page:
Operating system: the implemented greedy algorithms are platform independent, and the implemented LP-relaxation algorithm runs on the Windows operating system.
Programming language: the greedy algorithms are implemented in JAVA, and the LP-relaxation algorithm is implemented in Perl.
Methods
Assume we are given a haplotype block containing N SNPs and K haplotype patterns. This block is denoted by an N × K binary matrix Mh (see Figure 6 (A)). Define Mh[i,j] ∈ {1,2} for each i ∈ [1, N] and j ∈ [1, K], where 1 and 2 represent the major and minor alleles, respectively. In reality, the haplotype block may also contain missing data. This formulation can be easily extended to handle missing data by treating them as wild card symbols. To simplify the presentation of this paper, we will assume no missing data in the block. Let C be the set of all SNPs in Mh. The robust tag SNPs C' ⊆ C are a subset of SNPs which is able to distinguish each pair of haplotype patterns unambiguously when at most m SNPs are missing. Note that the missing data may occur at any SNP locus and thus create different missing patterns (see Figure 2). For any haplotype pattern with up to m missing SNPs, the set of robust tag SNPs C' is required to distinguish all of them unambiguously.
Figure 6 Reformulation of the MRTS problem. (A) The haplotype matrix Mh containing N SNPs and K haplotype patterns. (B) The bipartite graph corresponding to Mh.
To distinguish a haplotype pattern unambiguously, each pair of patterns must be distinguished by at least one SNP in C'. For example (see Figure 6 (A)), we say patterns P1 and P2 can be distinguished by SNP S2 since Mh[2,1] ≠ Mh[2,2]. A formal definition of this problem is given below.
Problem: Minimum Robust Tag SNPs (MRTS)
Input: An N × K matrix Mh and an integer m.
Output: The minimum subset of SNPs C' ⊆ C which satisfies:
(1) for each pair of patterns Pi and Pj, these is a SNP Sk ∈ C' such that Mh[k, i] ≠ Mh[k, j];
(2) when at most m SNPs are discarded from C' arbitrarily, (1) still holds.
We then reformulate MRTS to a variant of the set covering problem [6]. Each SNP Sk ∈ C (i.e., the k-th row in Mh) is reformulated to a set Sk'
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaei4jaCcaaaaa@303F@ = {(i, j) | M[k, i] ≠ M[k, j] and i <j}. For example, suppose the k-th row in Mh is {1,1,1,2}. The corresponding set Sk'
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaei4jaCcaaaaa@303F@ = {(1,4), (2,4), (3,4)}. In other words, Sk'
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaei4jaCcaaaaa@303F@ stores the pairs of patterns distinguished by SNP Sk. Define P as the set that contains all pairs of patterns (i.e., P = {(i,j) | 1 ≤ i <j ≤ K} = {(1,2), (1,3), ..., (K - l,K)}).
Consider each element in P and each reformulated set of C as nodes in an undirected bipartite graph (see Figure 6 (B). If SNP Sk can distinguish patterns Pi and Pj (i.e., (i,j) ∈ Sk'
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaei4jaCcaaaaa@303F@), there is an edge connecting the nodes (i, j) and Sk'
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaei4jaCcaaaaa@303F@. The following lemma implies that each pair of patterns must be distinguished by at least (m + 1) SNPs to tolerate m missing SNPs.
Lemma 1. C' ⊆ C is the set of robust tag SNPs which allows at most m missing SNPs iff each node in P has at least (m + 1) edges connecting to each node in C'.
Proof. Let C' be the set of robust tag SNPs which allows at most m missing SNPs. Suppose patterns Pi and Pj are distinguished by only m SNPs in C' (i.e., (i, j) has only m edges connecting to nodes in C'). However, if these m SNPs are all missing, no other SNPs in C' are able to distinguish patterns Pi and Pj, which is a contradiction. Thus, each pair of patterns must be distinguished by at least (m + 1) SNPs, which implies that each node in P must have at least (m + 1) edges connecting to nodes in C'. The proof of the other direction is similar.
In the following, we give a lower bound regarding the minimum number of robust tag SNPs required, which is used to skip some solution space by the OPT program.
Lemma 2. Given K haplotype patterns, the minimum number of robust tag SNPs required is at least log K.
Proof. Recall that the value of a SNP is binary. The maximum number of distinct haplotypes which can be distinguished by N SNPs is at most 2N. As a result, for a given data set containing K haplotype patterns, the minimum number of SNPs required is at least log K.
The following theorem shows the NP-hardness of the MRTS problem, which implies there is no polynomial time algorithm to find the optimal solution of MRTS.
Theorem 1. The MRTS problem is NP-hard.
Proof. When m = 0, MRTS is the same as the original problem of finding minimum tag SNPs, which is known as the minimum test set problem [6,17]. Since the minimum test set problem is NP-hard and can be reduced to a special case of MRTS, MRTS is NP-hard.
The first greedy algorithm
To solve MRTS efficiently, we propose a greedy algorithm which returns a solution not too larger than the optimal solution. By Lemma 1, to tolerate m missing tag SNPs, we need to find a subset of SNPs C' ⊆ C such that each pair of patterns in P is distinguished by at least (m + 1) SNPs in C'. Assume that the SNPs selected by this algorithm are stored in a (m + 1) × |P| table (see Figure 7 (A)). Initially, each grid in the table is empty. Once a SNP Sk, (that can distinguish patterns Pi and Pj) is selected, one grid of the column (i, j) is filled in with Sk, and we say that this grid is covered by Sk.
Figure 7 An example of the first greedy algorithm. The SNPs S1, S4, S2, and S3 are selected by the first greedy algorithm. (A) The table that stores each selected SNP.
This greedy algorithm works by covering the grids from the first row to the (m + 1)-th row, and greedily selects a SNP which covers most uncovered grids in the i-th row at each iteration. In other words, while working on the i-th row, a SNP is selected if its reformulated set S' maximizes |S' ∩ Ri |, where Ri is the set of uncovered grids at the i-th row.
Figure 7 illustrates an example for this algorithm to tolerate one missing tag SNP (i.e., m = 1). The SNPs S1, S4, S2, and S3 are selected in order. When all grids in this table are covered, each pair of patterns is distinguished by (m + 1) SNPs in the corresponding column. Thus, the SNPs in this table are the robust tag SNPs which can tolerate up to m missing SNPs. The pseudo code of this greedy algorithm is given below.
Algorithm: FlRST-GREEDY-ALGORITHM (C, P, m)
1 Ri ← P, ∀i ∈ [1, m + 1]
2 C' ← φ
3 for i = 1 to m + 1 do
4 while Ri ≠ φ do
5 select and remove a SNP S from C that maximizes |S' ∩ Ri|
6 C' ← C' ∪ S
7 j ← i
8 while S' ≠ φ and j ≤ m + 1 do
9 Stmp ← S' ∩ Rj //Stmp is a temporary variable for holding the result of S' ∩ Ri
10 Rj ← Rj - Stmp
11 S' ← S' - Stmp
12 j ← j + l
13 endwhile
14 endwhile
15 endfor
16 return C'
The time complexity of this algorithm is analyzed as follows. At Line 4, the number of iterations of the intermediate loop is bounded by |Ri| ≤ |P|. Within the loop body (Lines 5–13), Line 5 takes O(|C||P|) because we need to check all SNPs in C and examine the uncovered grids of Ri. The inner loop (Lines 8–13) takes only O(|S'|). Thus, the entire program runs in O(m|C||P|2).
We now show the solution C' returned by the first greedy algorithm is not too larger than the optimal solution C*. Suppose the algorithm selects the k-th SNP when working on the i-th row. Let |Skc
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaem4yamgaaaaa@30B8@| be the number of grids in the i-th row covered by the k-th selected SNP (i.e., |Skc
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaem4yamgaaaaa@30B8@| = |S' ∩ Ri|; see Line 5 in FIRST-GREEDY-ALGORITHM). For example (see Figure 7), S2c=2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaeGOmaidabaGaem4yamgaaOGaeyypa0JaeGOmaidaaa@324D@ since the second selected SNP (i.e., S4) covers two grids in the first row. We incur 1 unit of cost to each selected SNP, and spread this cost among the grids in Skc
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaem4yamgaaaaa@30B8@[3]. In other words, each grid at the i-th row and j-th column is assigned a cost Cji
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaemOAaOgabaGaemyAaKgaaaaa@30A2@ (see Figure 8), where
Figure 8 Analysis of the first greedy algorithm. This figure shows the cost Cji
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaemOAaOgabaGaemyAaKgaaaaa@30A2@ of each grid for the first greedy algorithm.
Cji={1|Skc|if the algorithm selects the k-th SNP when covering the i-th row; 0otherwise.
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@9D18@
Since each selected SNP is assigned 1 unit of cost, the sum of Cji
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaemOAaOgabaGaemyAaKgaaaaa@30A2@ for each grid in the table is equal to |C'|,
i.e.,
|C'| =∑i=1m+1∑j=1K(K−1)2Cji. (1)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqGG8baFcqWGdbWqcqGGNaWjcqGG8baFcaaMc8UaaGPaVlabg2da9maaqahabaWaaabCaeaacqWGdbWqdaqhaaWcbaGaemOAaOgabaGaemyAaKgaaaqaaiabdQgaQjabg2da9iabigdaXaqaamaaleaameaacqWGlbWscqGGOaakcqWGlbWscqGHsislcqaIXaqmcqGGPaqkaeaacqaIYaGmaaaaniabggHiLdaaleaacqWGPbqAcqGH9aqpcqaIXaqmaeaacqWGTbqBcqGHRaWkcqaIXaqma0GaeyyeIuoakiabc6caUiaaxMaacaWLjaWaaeWaaeaacqaIXaqmaiaawIcacaGLPaaaaaa@537C@
Let Rki
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGsbGudaqhaaWcbaGaem4AaSgabaGaemyAaKgaaaaa@30C2@ be the number of uncovered grids in the i-th row before the k-th iteration (i.e., (k - 1) SNPs have been selected by the algorithm). For example (see Figure 8), R21=2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGsbGudaqhaaWcbaGaeGOmaidabaGaeGymaedaaOGaeyypa0JaeGOmaidaaa@31EC@ since two grids in the first row are still uncovered before the second SNP is selected. Define Ci'
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaemyAaKgabaGaei4jaCcaaaaa@301B@ as the set of iterations used by the algorithm when working on the i-th row. For example (see Figure 8), C2'={3,4}
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaeGOmaidabaGaei4jaCcaaOGaeyypa0Jaei4EaSNaeG4mamJaeiilaWIaeGinaqJaeiyFa0haaa@368C@ since this algorithm works on the second row in the third and fourth iterations. We can rewrite (1) as
∑i=1m+1∑j=1K(K−1)2Cji=∑i=1m+1∑k∈Ci'(Rk−1i−Rki)1|Skc|. (2)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@7177@
Lemma 3. The k-th selected SNP has |Skc|≥Rk−1i|C*|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaabdaqaaiabdofatnaaDaaaleaacqWGRbWAaeaacqWGJbWyaaaakiaawEa7caGLiWoacqGHLjYSdaWcaaqaaiabdkfasnaaDaaaleaacqWGRbWAcqGHsislcqaIXaqmaeaacqWGPbqAaaaakeaacqGG8baFcqWGdbWqcqGGQaGkcqGG8baFaaaaaa@40A0@.
Proof. Suppose the algorithm is working on the i-th row at the beginning of the k-th iteration. Let Ck*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaem4AaSgabaGaeiOkaOcaaaaa@3025@ be the set of SNPs in C* (the optimal solution) that has been selected by the algorithm before the k-th iteration, and the set of remaining SNPs in C* be Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@. We claim that there exists a SNP in Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@ which can cover at least Rki|Ck¯*|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabdkfasnaaDaaaleaacqWGRbWAaeaacqWGPbqAaaaakeaadaabdaqaaiabdoeadnaaDaaaleaacuWGRbWAgaqeaaqaaiabcQcaQaaaaOGaay5bSlaawIa7aaaaaaa@3797@ grids in the i-th row. Otherwise (i.e., each SNP in Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@ covers less than Rki|Ck¯*|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabdkfasnaaDaaaleaacqWGRbWAaeaacqWGPbqAaaaakeaadaabdaqaaiabdoeadnaaDaaaleaacuWGRbWAgaqeaaqaaiabcQcaQaaaaOGaay5bSlaawIa7aaaaaaa@3797@ grids), all SNPs in Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@ will cover less than (Rki|Ck¯*|×|Ck¯*|=Rki)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqGGOaakdaWcaaqaaiabdkfasnaaDaaaleaacqWGRbWAaeaacqWGPbqAaaaakeaadaabdaqaaiabdoeadnaaDaaaleaacuWGRbWAgaqeaaqaaiabcQcaQaaaaOGaay5bSlaawIa7aaaacqGHxdaTdaabdaqaaiabdoeadnaaDaaaleaacuWGRbWAgaqeaaqaaiabcQcaQaaaaOGaay5bSlaawIa7aiabg2da9iabdkfasnaaDaaaleaacqWGRbWAaeaacqWGPbqAaaGccqGGPaqkaaa@473F@ grids in the i-th row. But since Ck*∪Ck¯*=C*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaem4AaSgabaGaeiOkaOcaaOGaeSOkIuLaem4qam0aa0baaSqaaiqbdUgaRzaaraaabaGaeiOkaOcaaOGaeyypa0Jaem4qamKaeiOkaOcaaa@37EB@, this implies that C* can not cover all grids in Rki
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGsbGudaqhaaWcbaGaem4AaSgabaGaemyAaKgaaaaa@30C2@, which is a contradiction. Because all SNPs in Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@ are candidates to the greedy algorithm, the k-th selected SNP must cover at least Rki|Ck¯*|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabdkfasnaaDaaaleaacqWGRbWAaeaacqWGPbqAaaaakeaadaabdaqaaiabdoeadnaaDaaaleaacuWGRbWAgaqeaaqaaiabcQcaQaaaaOGaay5bSlaawIa7aaaaaaa@3797@ grids in the i-th row, which implies |Skc|≥Rk−1i|C*|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaabdaqaaiabdofatnaaDaaaleaacqWGRbWAaeaacqWGJbWyaaaakiaawEa7caGLiWoacqGHLjYSdaWcaaqaaiabdkfasnaaDaaaleaacqWGRbWAcqGHsislcqaIXaqmaeaacqWGPbqAaaaakeaacqGG8baFcqWGdbWqcqGGQaGkcqGG8baFaaaaaa@40A0@ since |C*| ≥ |Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@| and |Rki|≤|Rk−1i|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaabdaqaaiabdkfasnaaDaaaleaacqWGRbWAaeaacqWGPbqAaaaakiaawEa7caGLiWoacqGHKjYOdaabdaqaaiabdkfasnaaDaaaleaacqWGRbWAcqGHsislcqaIXaqmaeaacqWGPbqAaaaakiaawEa7caGLiWoaaaa@3EC0@. □
Theorem 2. The first greedy algorithm gives a solution of (m + 1) lnK(K−1)2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGSbaBcqqGUbGBdaWcaaqaaiabdUealjabcIcaOiabdUealjabgkHiTiabigdaXiabcMcaPaqaaiabikdaYaaaaaa@363F@approximation.
Proof. Define the d-th harmonic number as H(d)=∑i=1d1i
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGibascqGGOaakcqWGKbazcqGGPaqkcqGH9aqpdaaeWaqaamaalaaabaGaeGymaedabaGaemyAaKgaaaWcbaGaemyAaKMaeyypa0JaeGymaedabaGaemizaqganiabggHiLdaaaa@3ACF@ and H(0) = 0. By (2) and Lemma 3,
∑i=1m+1∑j=1K(K−1)2Cji=∑i=1m+1∑k∈Ci'(Rk−1i−Rki)1|Skc|≤∑i=1m+1∑k∈Ci'(Rk−1i−Rki)|C*|Rk−1i=∑i=1m+1∑k∈Ci'(∑l=Rki+1Rk−1i|C*|Rk−1i)≤|C*|∑i=1m+1∑k∈Ci'∑l=Rki+1Rk−1i1l (l≤Rk−1i)=|C*|∑i=1m+1∑k∈Ci'(∑l=1Rk−1i1l−∑l=1Rki1l)≤|C*|∑i=1m+1∑k∈Ci'(H(Rk−1i)−H(Rki))≤|C*|∑i=1m+1(H(R0i)−H(R|Ci'|i))≤|C*|(m+1)max{H(R0i)} (R|Ci'|i=0 and H(0)=0)≤|C*|(m+1) ln|P|. (H(R0i)≤H(|P|)) (3)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakqaaeeqaamaaqahabaWaaabCaeaacqWGdbWqdaqhaaWcbaGaemOAaOgabaGaemyAaKgaaaqaaiabdQgaQjabg2da9iabigdaXaqaamaaleaameaacqWGlbWscqGGOaakcqWGlbWscqGHsislcqaIXaqmcqGGPaqkaeaacqaIYaGmaaaaniabggHiLdaaleaacqWGPbqAcqGH9aqpcqaIXaqmaeaacqWGTbqBcqGHRaWkcqaIXaqma0GaeyyeIuoakiabg2da9maaqahabaWaaabuaeaadaqadaqaaiabdkfasnaaDaaaleaacqWGRbWAcqGHsislcqaIXaqmaeaacqWGPbqAaaGccqGHsislcqWGsbGudaqhaaWcbaGaem4AaSgabaGaemyAaKgaaaGccaGLOaGaayzkaaaaleaacqWGRbWAcqGHiiIZcqWGdbWqdaqhaaadbaGaemyAaKgabaGaei4jaCcaaaWcbeqdcqGHris5aaWcbaGaemyAaKMaeyypa0JaeGymaedabaGaemyBa0Maey4kaSIaeGymaedaniabggHiLdGcdaWcbaWcbaGaeGymaedabaWaaqWaaeaacqWGtbWudaqhaaadbaGaem4AaSgabaGaem4yamgaaaWccaGLhWUaayjcSdaaaOGaeyizIm6aaabCaeaadaaeqbqaamaabmaabaGaemOuai1aa0baaSqaaiabdUgaRjabgkHiTiabigdaXaqaaiabdMgaPbaakiabgkHiTiabdkfasnaaDaaaleaacqWGRbWAaeaacqWGPbqAaaaakiaawIcacaGLPaaaaSqaaiabdUgaRjabgIGiolabdoeadnaaDaaameaacqWGPbqAaeaacqGGNaWjaaaaleqaniabggHiLdaaleaacqWGPbqAcqGH9aqpcqaIXaqmaeaacqWGTbqBcqGHRaWkcqaIXaqma0GaeyyeIuoakmaalaaabaWaaqWaaeaacqWGdbWqcqGGQaGkaiaawEa7caGLiWoaaeaacqWGsbGudaqhaaWcbaGaem4AaSMaeyOeI0IaeGymaedabaGaemyAaKgaaaaaaOqaaiabg2da9maaqahabaWaaabuaeaacqGGOaakdaaeWbqaamaalaaabaWaaqWaaeaacqWGdbWqcqGGQaGkaiaawEa7caGLiWoaaeaacqWGsbGudaqhaaWcbaGaem4AaSMaeyOeI0IaeGymaedabaGaemyAaKgaaaaaaeaacqWGSbaBcqGH9aqpcqWGsbGudaqhaaadbaGaem4AaSgabaGaemyAaKgaaSGaey4kaSIaeGymaedabaGaemOuai1aa0baaWqaaiabdUgaRjabgkHiTiabigdaXaqaaiabdMgaPbaaa0GaeyyeIuoakiabcMcaPaWcbaGaem4AaSMaeyicI4Saem4qam0aa0baaWqaaiabdMgaPbqaaiabcEcaNaaaaSqab0GaeyyeIuoaaSqaaiabdMgaPjabg2da9iabigdaXaqaaiabd2gaTjabgUcaRiabigdaXaqdcqGHris5aaGcbaGaeyizIm6aaqWaaeaacqWGdbWqcqGGQaGkaiaawEa7caGLiWoadaaeWbqaamaaqafabaWaaabCaeaadaWcaaqaaiabigdaXaqaaiabdYgaSbaaaSqaaiabdYgaSjabg2da9iabdkfasnaaDaaameaacqWGRbWAaeaacqWGPbqAaaWccqGHRaWkcqaIXaqmaeaacqWGsbGudaqhaaadbaGaem4AaSMaeyOeI0IaeGymaedabaGaemyAaKgaaaqdcqGHris5aaWcbaGaem4AaSMaeyicI4Saem4qam0aa0baaWqaaiabdMgaPbqaaiabcEcaNaaaaSqab0GaeyyeIuoaaSqaaiabdMgaPjabg2da9iabigdaXaqaaiabd2gaTjabgUcaRiabigdaXaqdcqGHris5aOGaaCzcaiaaxMaadaqadaqaaiabdYgaSjabgsMiJkabdkfasnaaDaaaleaacqWGRbWAcqGHsislcqaIXaqmaeaacqWGPbqAaaaakiaawIcacaGLPaaaaeaacqGH9aqpdaabdaqaaiabdoeadjabcQcaQaGaay5bSlaawIa7amaaqahabaWaaabuaeaacqGGOaakdaaeWbqaamaalaaabaGaeGymaedabaGaemiBaWgaaaWcbaGaemiBaWMaeyypa0JaeGymaedabaGaemOuai1aa0baaWqaaiabdUgaRjabgkHiTiabigdaXaqaaiabdMgaPbaaa0GaeyyeIuoakiabgkHiTmaaqahabaWaaSaaaeaacqaIXaqmaeaacqWGSbaBaaaaleaacqWGSbaBcqGH9aqpcqaIXaqmaeaacqWGsbGudaqhaaadbaGaem4AaSgabaGaemyAaKgaaaqdcqGHris5aaWcbaGaem4AaSMaeyicI4Saem4qam0aa0baaWqaaiabdMgaPbqaaiabcEcaNaaaaSqab0GaeyyeIuoaaSqaaiabdMgaPjabg2da9iabigdaXaqaaiabd2gaTjabgUcaRiabigdaXaqdcqGHris5aOGaeiykaKcabaGaeyizIm6aaqWaaeaacqWGdbWqcqGGQaGkaiaawEa7caGLiWoadaaeWbqaamaaqafabaGaeiikaGIaemisaGKaeiikaGIaemOuai1aa0baaSqaaiabdUgaRjabgkHiTiabigdaXaqaaiabdMgaPbaakiabcMcaPiabgkHiTiabdIeaijabcIcaOiabdkfasnaaDaaaleaacqWGRbWAaeaacqWGPbqAaaGccqGGPaqkcqGGPaqkaSqaaiabdUgaRjabgIGiolabdoeadnaaDaaameaacqWGPbqAaeaacqGGNaWjaaaaleqaniabggHiLdaaleaacqWGPbqAcqGH9aqpcqaIXaqmaeaacqWGTbqBcqGHRaWkcqaIXaqma0GaeyyeIuoaaOqaaiabgsMiJoaaemaabaGaem4qamKaeiOkaOcacaGLhWUaayjcSdWaaabCaeaacqGGOaakcqWGibascqGGOaakcqWGsbGudaqhaaWcbaGaeGimaadabaGaemyAaKgaaOGaeiykaKIaeyOeI0IaemisaGKaeiikaGIaemOuai1aa0baaSqaamaaemaabaGaem4qam0aa0baaWqaaiabdMgaPbqaaiabcEcaNaaaaSGaay5bSlaawIa7aaqaaiabdMgaPbaaaeaacqWGPbqAcqGH9aqpcqaIXaqmaeaacqWGTbqBcqGHRaWkcqaIXaqma0GaeyyeIuoakiabcMcaPiabcMcaPaqaaiabgsMiJoaaemaabaGaem4qamKaeiOkaOcacaGLhWUaayjcSdGaeiikaGIaemyBa0Maey4kaSIaeGymaeJaeiykaKIagiyBa0MaeiyyaeMaeiiEaGNaei4EaSNaemisaGKaeiikaGIaemOuai1aa0baaSqaaiabicdaWaqaaiabdMgaPbaakiabcMcaPiabc2ha9jaaxMaacaWLjaGaeiikaGIaemOuai1aa0baaSqaamaaemaabaGaem4qam0aa0baaWqaaiabdMgaPbqaaiabcEcaNaaaaSGaay5bSlaawIa7aaqaaiabdMgaPbaakiabg2da9iabicdaWiabbccaGiabbggaHjabb6gaUjabbsgaKjabbccaGiabdIeaijabcIcaOiabicdaWiabcMcaPiabg2da9iabicdaWiabcMcaPaqaaiabgsMiJoaaemaabaGaem4qamKaeiOkaOcacaGLhWUaayjcSdGaeiikaGIaemyBa0Maey4kaSIaeGymaeJaeiykaKIaeeiiaaIaeeiBaWMaeeOBa4MaeeiFaWNaemiuaaLaeiiFaWNaeiOla4IaaCzcaiaaxMaacaWLjaGaeiikaGIaemisaGKaeiikaGIaemOuai1aa0baaSqaaiabicdaWaqaaiabdMgaPbaakiabcMcaPiabgsMiJkabdIeaijabcIcaOiabcYha8jabdcfaqjabcYha8jabcMcaPiabcMcaPiaaxMaacaWLjaWaaeWaaeaacqaIZaWmaiaawIcacaGLPaaaaaaa@DEEA@
By (1) and (3), we get
|C'||C*|≤(m+1) ln |P| = (m+1) lnK(K−1)2.
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabcYha8jabdoeadjabcEcaNiabcYha8bqaaiabcYha8jabdoeadjabcQcaQiabcYha8baacqGHKjYOcqGGOaakcqWGTbqBcqGHRaWkcqaIXaqmcqGGPaqkcqqGGaaicqqGSbaBcqqGUbGBcqqGGaaicqqG8baFcqWGqbaucqGG8baFcqqGGaaicqqG9aqpcqqGGaaicqqGOaakcqWGTbqBcqGHRaWkcqaIXaqmcqGGPaqkcqqGGaaicqqGSbaBcqqGUbGBdaWcaaqaaiabdUealjabcIcaOiabdUealjabgkHiTiabigdaXiabcMcaPaqaaiabikdaYaaacqGGUaGlaaa@5853@
The second greedy algorithm
This section describes the second greedy algorithm which returns a solution of better approximation than that of the first greedy algorithm. Let Ri be the set of uncovered grids at the i-th row. Unlike the row-by-row manner of the first greedy algorithm, this algorithm greedily selects a SNP that covers most uncovered grids in the table (i.e., its reformulated set S' maximizing |S' ∩ (R1 ∪ ... ∪ Rm+1)|). Let T be the collection of Ri (i.e., T is the set of all uncovered grids in the table). If the grids in the i-th row are all covered (i.e., Ri = φ), Ri is removed from T. This algorithm runs until T = φ (i.e., all grids in the table are covered).
Figure 9 illustrates an example for this algorithm with m set to 1. The SNPs S1, S2, S4, and S5 are selected in order. Since this algorithm runs until all grids are covered, the set of SNPs in this table is able to tolerate m missing tag SNPs. The pseudo code of this algorithm is given below.
Figure 9 An example of the second greedy algorithm. The SNPs S1, S2, S4, and S5 are selected by the second greedy algorithm. (A) The table that stores each selected SNP.
Algorithm: SECOND-GREEDY-ALGORITHM (C, P, m)
1 Ri ← P, ∀ i ∈ [1, m + 1]
2 T ← {R1, R2,... ,Rm+1}
3 C' ← φ
4 while T ≠ φ do
5 select and remove a SNP S from C that maximizes |S' ∩ (R1 ∪ ... ∪ Rm+1)|
6 C' ← C' ∪ S
7 for each Ri ∈ T and S' ≠ ø do
8 Stmp ← S' ∩ Ri // Stmp is a temporary variable for holding the result of S' ∩ Ri
9 Ri ← Ri - Stmp
10 S' ← S' - Stmp
11 if Ri = φ then T ← T - Ri
12 endfor
13 end while
14 return C'
The time complexity of this algorithm is analyzed as follows. At Line 4, the number of iterations of the loop is bounded by O(|T|) = O(m|P|). Within the loop, Line 5 takes O(|C||P|) time because we need to check each SNP in C and examine if it can cover any uncovered grid in each column. The inner loop (Lines 7–12) is bounded by O(|S'|) <O(|P|). Thus, the running time of this program is O(m|C||P|2).
We now evaluate the solution returned by the second greedy algorithm. Let C' and C* be the set of SNPs selected by this algorithm and the optimal solution, respectively. Let |Skc
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaem4AaSgabaGaem4yamgaaaaa@30B8@| be the number of grids in the table covered by the k-th selected SNP. For example (see Figure 9), |S2c
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGtbWudaqhaaWcbaGaeGOmaidabaGaem4yamgaaaaa@304B@| = 4 since the second selected SNP (i.e., S2) covers four grids in the table. Define Tk as the number of uncovered grids in the table before the k-th iteration. We have the following lemma similar to Lemma 3.
Lemma 4. The k-th selected SNP has |Skc|≥Tk−1|C*|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaabdaqaaiabdofatnaaDaaaleaacqWGRbWAaeaacqWGJbWyaaaakiaawEa7caGLiWoacqGHLjYSdaWcaaqaaiabdsfaunaaBaaaleaacqWGRbWAcqGHsislcqaIXaqmaeqaaaGcbaGaeiiFaWNaem4qamKaeiOkaOIaeiiFaWhaaaaa@3F48@.
Proof. The proof is similar to that of Lemma 3. Let Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@ be the set of remaining SNPs in C* which has not been selected before the k-th iteration. We claim that there exists a SNP in Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@ which can cover at least Tk|Ck¯*|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabdsfaunaaBaaaleaacqWGRbWAaeqaaaGcbaWaaqWaaeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaakiaawEa7caGLiWoaaaaaaa@363F@ grids in the table. Otherwise, we can get the same contradiction (i.e., C* fails to cover all grids) as in Lemma 3. Since |C*| ≥ |Ck¯*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGafm4AaSMbaebaaeaacqGGQaGkaaaaaa@303D@| and Tk-1 ≤ Tk, we have |Skc|≥Tk−1|C*|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaabdaqaaiabdofatnaaDaaaleaacqWGRbWAaeaacqWGJbWyaaaakiaawEa7caGLiWoacqGHLjYSdaWcaaqaaiabdsfaunaaBaaaleaacqWGRbWAcqGHsislcqaIXaqmaeqaaaGcbaGaeiiFaWNaem4qamKaeiOkaOIaeiiFaWhaaaaa@3F48@. □
Theorem 3. The second greedy algorithm gives a solution of ln((m+1)K(K−1)2)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGSbaBcqqGUbGBcqqGOaakcqqGOaakcqWGTbqBcqGHRaWkcqaIXaqmcqGGPaqkdaWcaaqaaiabdUealjabcIcaOiabdUealjabgkHiTiabigdaXiabcMcaPaqaaiabikdaYaaacqGGPaqkaaa@3CD6@approximation.
Proof. Each grid at the i-th row and j-th column is assigned a cost Cji=1|Skc|
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaemOAaOgabaGaemyAaKgaaOGaeyypa0ZaaSaaaeaacqaIXaqmaeaadaabdaqaaiabdofatnaaDaaaleaacqWGRbWAaeaacqWGJbWyaaaakiaawEa7caGLiWoaaaaaaa@39E8@ (see Figure 10) if it is covered by the k-th selected SNP. The sum of Cji
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaemOAaOgabaGaemyAaKgaaaaa@30A2@ for each grid is
Figure 10 Analysis of the second greedy algorithm. This figure shows the cost Cji
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGdbWqdaqhaaWcbaGaemOAaOgabaGaemyAaKgaaaaa@30A2@ of each grid for the second greedy algorithm.
|C'| = ∑i=1m+1∑j=1K(K−1)2Cji=∑k=1|C'|(Tk−1−Tk)1|Skc|(see (1) and (2))≤∑k=1|C'|(Tk−1−Tk)|C*|TK−1(by Lemma 4)≤|C*|(H(T0)−H(T|C'|))(see the proof in Theorem 2)≤|C*|ln((m+1)|P|). (4)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@073E@
By (4), we have
|C'||C*|≤ln((m+1)|P|)=ln((m+1)K(K−1)2).
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabcYha8jabdoeadjabcEcaNiabcYha8bqaaiabcYha8jabdoeadjabcQcaQiabcYha8baacqGHKjYOcyGGSbaBcqGGUbGBcqGGOaakcqGGOaakcqWGTbqBcqGHRaWkcqaIXaqmcqGGPaqkcqGG8baFcqWGqbaucqGG8baFcqGGPaqkcqGH9aqpcyGGSbaBcqGGUbGBcqGGOaakcqGGOaakcqWGTbqBcqGHRaWkcqaIXaqmcqGGPaqkdaWcaaqaaiabdUealjabcIcaOiabdUealjabgkHiTiabigdaXiabcMcaPaqaaiabikdaYaaacqGGPaqkcqGGUaGlaaa@57E3@
The iterative LP-relaxation algorithm
In practice, a probabilistic approach is sometimes more useful since the randomization can explore different solutions. In this section, we reformulate the MRTS problem to an Integer Programming (IP) problem. Based on the IP problem, we propose an iterative Linear Programming (LP)-relaxation algorithm. The iterative LP-relaxation algorithm is described below.
Step 1. Given a haplotype block containing N SNPs and K haplotype patterns. Let {x1,x2, ...,xN} be the set of integer variables for the N SNPs, where xk = 1 if the SNP Sk is selected and xk = 0 otherwise. Define D(Pi, Pj) as the set of SNPs which are able to distinguish Pi and Pj patterns. By Lemma 1, to allow at most m missing SNPs, each pair of patterns must be distinguished by at least (m + 1) SNPs. Therefore, for each set D(Pi, Pj), at least (m + 1) SNPs have to be selected to distinguish Pi and Pj patterns. As a consequence, the MRTS problem can be formulated as the following IP problem:
Minimize∑k=1NxkSubjectto ∑k∈D(Pi,Pj)xk=0 or 1.xk≥m+1,for all 1≤i<j≤K, (5)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaafaqaaeGacaaabaacbeGae8xta0Kae8xAaKMae8NBa4Mae8xAaKMae8xBa0Mae8xAaKMae8NEaONae8xzaugabaWaaabCaeaacqWG4baEdaWgaaWcbaGaem4AaSgabeaaaeaacqWGRbWAcqGH9aqpcqaIXaqmaeaacqWGobGta0GaeyyeIuoaaOqaaiab=nfatjab=vha1jab=jgaIjab=PgaQjab=vgaLjab=ngaJjab=rha0jab=bcaGiab=rha0jab=9gaVbqaaiabbccaGmaaqafabaGaemiEaG3aaSbaaSqaaiabdUgaRbqabaGccqGHLjYScqWGTbqBcqGHRaWkcqaIXaqmcqGGSaalcaWLjaGaeeOzayMaee4Ba8MaeeOCaiNaeeiiaaIaeeyyaeMaeeiBaWMaeeiBaWMaeeiiaaIaeeymaeJaeyizImQaemyAaKMaeyipaWJaemOAaOMaeyizImQaem4saSKaeiilaWcaleaafaqabeGabaaabaGaem4AaSMaeyicI4SaemiraqKaeiikaGIaemiuaa1aaSbaaWqaaiabdMgaPbqabaWccqGGSaalcqWGqbaudaWgaaadbaGaemOAaOgabeaaliabcMcaPaqaaiabdIha4naaBaaameaacqWGRbWAaeqaaSGaeyypa0JaeGimaaJaeeiiaaIaee4Ba8MaeeOCaiNaeeiiaaIaeeymaeJaeeOla4caaaqab0GaeyyeIuoakiaaxMaacaWLjaWaaeWaaeaacqaI1aqnaiaawIcacaGLPaaaaaaaaa@894F@
Step 2. Since solving the IP problem is NP-hard [6], we relax the integer constraint of xk, and the IP problem becomes an LP problem defined as follows:
Minimize∑k=1NykSubjectto∑k∈D(Pi,Pj)0 ≤ yk ≤ 1.yk≥m+1, (6)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaafaqaaeGacaaabaWexLMBbXgBcf2CPn2qVrwzqf2zLnharyGvLjhzH5wyaGabbiaa=1eacaWFPbGaa8NBaiaa=LgacaWFTbGaa8xAaiaa=PhacaWFLbaabaWaaabCaeaacqWG5bqEdaWgaaWcbaGaem4AaSgabeaaaeaacqWGRbWAcqGH9aqpcqaIXaqmaeaacqWGobGta0GaeyyeIuoaaOqaaiaa=nfacaWF1bGaa8Nyaiaa=PgacaWFLbGaa83yaiaa=rhacaWFGaGaa8hDaiaa=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@961B@
The above LP problem can be solved in polynomial time by efficient algorithms such as the interior point method (Forsgren et al., 2002) [5].
Step 3. Let {y1, y2, ..., yN} be the set of linear solutions obtained from (6), where 0 ≤ yk ≤ 1. We construct the corresponding integer solutions {x1, x2, ..., xN} by the following randomized rounding method:
Assign{xk=1 with probability yk,xk=0 with probability 1-yk.
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@776C@
Note that the constructed integer solutions do not necessary satisfy all inequalities in (5). The randomized rounding method simply assigns xk to 1 or 0 using the value of yk as the likelihood, regardless of the inequalities in (5).
Step 4. We check whether the integer solutions constructed in Step 3 satisfy all inequalities in (5) or not.
Case 1. If some inequalities in (5) are still unsatisfied, we repeat Steps 1, 2, and 3 only for those unsatisfied inequalities until all of them are satisfied.
Case 2. If all inequalities in (5) are satisfied, we construct a final solution by setting xk = 1 if xk is assigned to 1 in any one of the iterations and setting xk = 0 otherwise.
We now evaluate the solution returned by the iterative LP-relaxation algorithm. The selection of each SNP is considered as a Bernoulli random variable xk taking values 1 (or 0) with probability yk (or 1 - yk). Let Xi,j be the sum of random variables in one inequality of (5), i.e.,
Xi,j=∑k∈D{Pi,Pj}xk.
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGybawdaWgaaWcbaGaemyAaKMaeiilaWIaemOAaOgabeaakiabg2da9maaqafabaGaemiEaG3aaSbaaSqaaiabdUgaRbqabaGccqGGUaGlaSqaaiabdUgaRjabgIGiolabdseaenaacmaabaGaemiuaa1aaSbaaWqaaiabdMgaPbqabaWccqGGSaalcqWGqbaudaWgaaadbaGaemOAaOgabeaaaSGaay5Eaiaaw2haaaqab0GaeyyeIuoaaaa@454F@
By (6), the expected value of Xi,j (after randomized rounding) is
E[Xi,j]=∑k∈D{Pi,Pj}E[xk]=∑k∈D{Pi,Pj}yk≥m+1. (7)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@6B7C@
Lemma 5. The probability that an inequality in (5) is not satisfied after randomized rounding is less than e−12(m+1)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGLbqzdaahaaWcbeqaaiabgkHiTmaaleaameaacqaIXaqmaeaacqaIYaGmcqGGOaakcqWGTbqBcqGHRaWkcqaIXaqmcqGGPaqkaaaaaaaa@3601@.
Proof. The probability that an inequality in (5) is not satisfied is P[Xi,j <m + 1] = P[Xi,j ≤ m]. By the Chernoff bound (i.e., P[X ≤ (1 - θ) E[X]] ≤ e−θ2E[X]2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGLbqzdaahaaWcbeqaaiabgkHiTmaaleaameaacqaH4oqCdaahaaqabeaacqaIYaGmaaGaemyrauKaei4waSLaemiwaGLaeiyxa0fabaGaeGOmaidaaaaaaaa@37C0@), we have
P[Xi,j≤m]≤e−(E[Xi,j]−m)22E[Xi,j]. (8)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGqbaucqGGBbWwcqWGybawdaWgaaWcbaGaemyAaKMaeiilaWIaemOAaOgabeaakiabgsMiJkabd2gaTjabc2faDjabgsMiJkabdwgaLnaaCaaaleqabaGaeyOeI0YaaSqaaWqaaiabcIcaOiabdweafjabcUfaBjabdIfaynaaBaaabaGaemyAaKMaeiilaWIaemOAaOgabeaacqGGDbqxcqGHsislcqWGTbqBcqGGPaqkdaahaaqabeaacqaIYaGmaaaabaGaeGOmaiJaemyrauKaei4waSLaemiwaG1aaSbaaeaacqWGPbqAcqGGSaalcqWGQbGAaeqaaiabc2faDbaaaaGccqGGUaGlcaWLjaGaaCzcamaabmaabaGaeGioaGdacaGLOaGaayzkaaaaaa@5880@
By (7), we know E[Xi,j] ≤ m + 1. Since the right-hand side of (8) decreases when E[Xi,j] > m, we can replace E[Xi,j] with (m + 1) to obtain an upper bound, i.e.,
P[Xi,j≤m]≤e−(E[Xi,j]−m)22E[Xi,j]≤e−12(m+1).≤e−(m+1−m)22(m+1)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@723D@
Theorem 4. The iterative LP-relaxation algorithm gives a solution of O(m ln K) approximation.
Proof. Suppose this algorithm runs for t iterations. The probability that all K(K−1)2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabdUealjabcIcaOiabdUealjabgkHiTiabigdaXiabcMcaPaqaaiabikdaYaaaaaa@337D@ inequalities in (5) are satisfied after t iterations is
(1−(e−1/2(m+1))t)K(K−1)2=(1−e−t/2(m+1))K(K−1)2≈e−K(K−1)2e−t/2(m+1).
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@71C3@
When t = 2(m + 1) lnK(K−1)2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGSbaBcqqGUbGBdaWcaaqaaiabdUealjabcIcaOiabdUealjabgkHiTiabigdaXiabcMcaPaqaaiabikdaYaaaaaa@363F@, the algorithm stops and returns a solution with probability e-1. Define OPT(IP) and OPT(LP) as the optimal solutions of the IP problem and the LP problem, respectively. Since the solution space of LP includes that of IP,
OPT(LP) ≤ OPT(IP).
Let the set of solutions returned in t iterations be {Z1, Z2,...,Zt}.
E[Z1]=E[∑k=1Nxk]=∑k=1Nyk=OPT(LP).
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGfbqrcqGGBbWwcqWGAbGwdaWgaaWcbaGaeGymaedabeaakiabc2faDjabg2da9iabdweafjabcUfaBnaaqahabaGaemiEaG3aaSbaaSqaaiabdUgaRbqabaGccqGGDbqxcqGH9aqpdaaeWbqaaiabdMha5naaBaaaleaacqWGRbWAaeqaaOGaeyypa0Jaem4ta8KaemiuaaLaemivaqLaeiikaGIaemitaWKaemiuaaLaeiykaKIaeiOla4caleaacqWGRbWAcqGH9aqpcqaIXaqmaeaacqWGobGta0GaeyyeIuoaaSqaaiabdUgaRjabg2da9iabigdaXaqaaiabd6eaobqdcqGHris5aaaa@5540@
Note that we repeat this algorithm only for those unsatisfied inequalities. Thus, E[Z1] ≥ E[Z2] ≥ ... ≥ E[Zt]. Let xp denote the final solution obtained in Step 4. The expected final solution is
E[∑p=1Nxp]≤E[∑p=1tZp]≤t×E[Z1]≤t×OPT(LP)≤2(m+1) ln K(K−1)2×OPT(IP)=O(m ln K)×OPT(IP).
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@92D2@
With a high probability, the iterative LP-relaxation algorithm stops after O(m ln K) iterations and finds a solution of O(m ln K) approximation. □
An algorithm for finding auxiliary tag SNPs
This section describes an algorithm for finding auxiliary tag SNPs assuming robust tag SNPs have been computed in advance. Given a haplotype block Mh containing N SNPs and K haplotypes, we define Ctag ⊆ C as the set of tag SNPs genotyped from a haplotype sample with some missing data. This haplotype sample may fail to be distinguished because of the ambiguity caused by missing data. We wish to find the minimum number of auxiliary tag SNPs from the remaining SNPs in the block to resolve the ambiguity. A formal definition of this problem is given below.
Problem: Minimum Auxiliary Tag SNPs (MATS)
Input: An N × K matrix Mh, and a set of SNPs Ctag genotyped from a sample with missing data.
Output: The minimum subset of SNPs Caux ⊆ C - Ctag such that each pair of ambiguous patterns can be distinguished by SNPs in Caux.
The following theorem shows the NP-hardness of the MATS problem.
Theorem 5. The MATS problem is NP-hard.
Proof. Consider that all SNPs in Ctag are missing. This special case of the MATS problem becomes finding the minimum tag SNPs from C - Ctag, which is already known to be NP-hard [17]. Therefore, MATS is also NP-hard. □
Although the MATS problem is NP-hard, we show that auxiliary tag SNPs can be found efficiently when robust tag SNPs have been computed in advance. Without loss of generality, assume that these robust tag SNPs are stored in an (m + 1) × |P| table Tr (see Figure 11 (A)).
Figure 11 An example of finding auxiliary tag SNPs. The SNP S1 is missing and SNP S4 is the auxiliary tag SNP for h2. (A) The table that stores the set of robust tag SNPs.
Step 1. The patterns that match the haplotype sample are stored into a set A. For example (see Figure 11), if we genotype SNPs S1, S2, and S3 for the sample h2 and the SNP S1 is missing, patterns P1 and P3 both match h2. Thus, A = {P1, P3}
Step 2. If |A| = 1, the sample is identified unambiguously and we are done (e.g., h1 in Figure 11). If |A| > 1 (e.g., h2), for each pair of ambiguous patterns in A (e.g., P1 and P3), traverse the corresponding column in Tr, find the next unused SNP (e.g., S4), and add the SNP to Caux. As a result, the SNPs in Caux can distinguish each pair of ambiguous patterns, which are the auxiliary tag SNPs for the haplotype sample.
The worst case of this algorithm is that all SNPs in Ctag are missing data, and we need to traverse each column in Tr. Thus, the running time of this algorithm is O(|Tr|) = O(m|P|).
Authors' contributions
YTH and KMC design and implement the greedy algorithms. KZ and TC design and implement the iterative LP relaxation algorithm. All authors write and approve the manuscript.
Supplementary Material
Additional File 1
The program for the first greedy algorithm. The Greedyl.zip file is compressed using WinZip and contains the JAVA source code for the first greedy algorithm.
Click here for file
Additional File 2
The program for the second greedy algorithm. The Greedy2.zip file is compressed using WinZip and contains the JAVA source code for the second greedy algorithm.
Click here for file
Additional File 3
The program for the iterative LP-relaxation algorithm. The ILP.zip file is compressed using WinZip and contains the Perl script for the iterative LP-relaxation algorithm.
Click here for file
Acknowledgements
We thank the referees for their valuable comments that resulted in numerous improvements in the presentation. Yao-Ting Huang and Kun-Mao Chao were supported in part by NSC grants 93-2213-E-002-029 and 94-2213-E-002-091 from the National Science Council, Taiwan. Ting Chen was supported in part by NIH CEGS: Implications of Haplotype Structure in the Human Genome, Grant No. P50 HG002790.
==== Refs
Bafna V Halldórsson BV Schwartz R Clark AG Istrail S Haplotypes and informative SNP selection algorithms: don't block out information Proc RECOMB'03 2003 19 27
Carlson CS Eberle MA Rieder MJ Yi Q Kruglyak L Nickerson DA Selecting a maximally informative set of single-nucleotide polymorphisms for association analyses using linkage disequilibrium Am J Hum Genet 2004 74 106 120 14681826
Cormen TH Leiserson CE Rivest RL Stein C Introduction to algorithms 2001 The MIT Press
Daly MJ Rioux JD Schaffner SF Hudson TJ Lander ES High-resolution haplotype structure in the human genome Nat Genet 2001 29 229 232 11586305
Forsgren A Gill PE Wright MH Interior methods for nonlinear optimization SIAM Rev 2002 44 525 597
Garey MR Johnson DS Computers and intractability 1979 Freeman, New York
Halldórsson BV Bafna V Lippert R Schwartz R Vega FM Clark AG Istrail S Optimal haplotype block-free selection of tagging SNPs for genome-wide association studies Genome Research 2004 1633 1640 15289481
Halperin E Eskin E Haplotype reconstruction from genotype data using imperfect phylogeny Bioinformatics 2004
Hinds DA Stuve LL Nilsen GB Halperin E Eskin E Ballinger DG Frazer KA Cox DR Whole-genome patterns of common DNA variation in three human populations Science 2005 307 1072 1079 15718463
Hudson RR Generating samples under a Wright-Fisher neutral model of genetic variation Bioinformatics 2002 18 337 338 11847089
LP Solve
Niu T Qin Z Xu X Liu JS Bayesian haplotype inference for multiple linked single-nucleotide polymorphisms Am J Hum Genet 2002 70 157 159 11741196
Patil N Berno AJ Hinds DA Barrett WA Doshi JM Hacker CR Kautzer CR Lee DH Marjoribanks C McDonough DP Nguyen BT Norris MC Sheehan JB Shen N Stern D Stokowski RP Thomas DJ Trulson MO Vyas KR Frazer KA Fodor SP Cox DR Blocks of limited haplotype diversity revealed by high-resolution scanning of human chromosome 21 Science 2001 294 1719 1723 11721056
Stephens M Donnelly P A comparison of bayesian methods for haplotype reconstruction from population genotype data Am J Hum Genet 2003 73 1162 1169 14574645
Wang L Xu Y Haplotype inference by maximum parsimony Bioinformatics 2003 19 1773 1780 14512348
Yang Y Zhang J Hoh J Matsuda F Xu P Lathrop M Ott J Efficiency of single-nucleotide polymorphism haplotype estimation from pooled DNA Proc Nat Acad Set 2003 100 7225 7230
Zhang K Deng M Chen T Waterman MS Sun F A dynamic programming algorithm for haplotype partitioning Proc Nat Acad Sci 2002 99 7335 7339 12032283
Zhang K Sun F Waterman MS Chen T Haplotype block partition with limited resources and applications to human chromosome 21 haplotype data Am J Hum Genet 2003 73 63 73 12802783
Zhang K Qin ZS Liu JS Chen T Waterman MS Sun F Haplotype block partition and tag SNP selection using genotype data and their applications to association studies Genome Research 2004 14 908 916 15078859
Zhao JH Lissarrague S Essioux L Sham PC GENECOUNTING: haplotype analysis with missing genotypes Bioinformatics 2002 18 1694 1695 12490459
|
16259642
|
PMC1316880
|
CC BY
|
2021-01-04 16:27:47
|
no
|
BMC Bioinformatics. 2005 Nov 1; 6:263
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-263
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2761629724310.1186/1471-2105-6-276Methodology ArticlePrincipal component analysis for predicting transcription-factor binding motifs from array-derived data Liu Yunlong [email protected] Matthew P [email protected] Hiroki [email protected] Department of Biomedical Engineering, Indiana University – Purdue University Indianapolis, Indianapolis, IN 46202, USA.2 Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN 47907, USA.3 Department of Anatomy and Cell Biology, Indiana University – Purdue University Indianapolis, Indianapolis, IN 46202, USA.4 Department of Veteran's Affairs, White River Jct, VT 05009, USA.5 Department of Medicine, Dartmouth Medical School, Hanover, NH 03755, USA.2005 18 11 2005 6 276 276 2 5 2005 18 11 2005 Copyright © 2005 Liu et al; licensee BioMed Central Ltd.2005Liu 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 responses to interleukin 1 (IL-1) in human chondrocytes constitute a complex regulatory mechanism, where multiple transcription factors interact combinatorially to transcription-factor binding motifs (TFBMs). In order to select a critical set of TFBMs from genomic DNA information and an array-derived data, an efficient algorithm to solve a combinatorial optimization problem is required. Although computational approaches based on evolutionary algorithms are commonly employed, an analytical algorithm would be useful to predict TFBMs at nearly no computational cost and evaluate varying modelling conditions. Singular value decomposition (SVD) is a powerful method to derive primary components of a given matrix. Applying SVD to a promoter matrix defined from regulatory DNA sequences, we derived a novel method to predict the critical set of TFBMs.
Results
The promoter matrix was defined to establish a quantitative relationship between the IL-1-driven mRNA alteration and genomic DNA sequences of the IL-1 responsive genes. The matrix was decomposed with SVD, and the effects of 8 potential TFBMs (5'-CAGGC-3', 5'-CGCCC-3', 5'-CCGCC-3', 5'-ATGGG-3', 5'-GGGAA-3', 5'-CGTCC-3', 5'-AAAGG-3', and 5'-ACCCA-3') were predicted from a pool of 512 random DNA sequences. The prediction included matches to the core binding motifs of biologically known TFBMs such as AP2, SP1, EGR1, KROX, GC-BOX, ABI4, ETF, E2F, SRF, STAT, IK-1, PPARγ, STAF, ROAZ, and NFκB, and their significance was evaluated numerically using Monte Carlo simulation and genetic algorithm.
Conclusion
The described SVD-based prediction is an analytical method to provide a set of potential TFBMs involved in transcriptional regulation. The results would be useful to evaluate analytically a contribution of individual DNA sequences.
==== Body
Background
The use of microarrays has led to a significant number of exciting discoveries establishing important links between mRNA expression patterns and cellular states [1,2]. Mathematical and computational models have been developed to understand and characterize the molecular mechanisms underlying expression patterns [3,4]. However, it remains difficult to discover and validate novel transcription-factor binding motifs (TFBMs) in the human genome. The popular approach to identify TFBMs utilizes sequence comparisons among co-expressed genes [5] or across multi-species [6]. Although any consensus motif can be searched among the co-regulated genes in hierarchical clusters [7,8], this approach is not aimed to build a global model with multiple binding motifs. TFBM can be inspected through phylogenetic footprinting [6,9,10], but identifying orthologous genes and their associated regulatory regions are not always possible. Model-based approaches, initially developed using yeast genome [3], encounter difficulty in evaluating the astronomical number of TFBM selections in the combinatorial problem [11,12]. Although multiple binding motifs were selected in the yeast dataset using a recursive formula, prediction of TFBMs would be affected depending on the order of selected motifs [3]. Some models lack statistical standards for determining the number of TFBMs having combinatorial roles that are critical in expression patterns. Thus, a predictive model that provides a comprehensive set of TFBMs still needs to be developed.
The specific aim of the current study was to devise a model for predicting known and de novo transcription factor binding motifs from array-derived mRNA expression levels by developing a unique principal component analysis. We employed the responses of human chondrocytes to interleukin-1 (IL-1) as a model system [13]. IL-1 is a pro-inflammatory cytokine, and it stimulates not only inflammatory responses but also tissue degeneration [5]. More than 100 microarray analyses have been conducted to analyze IL-1-driven responses in various cell types, including chondrocytes [14,15], and significant efforts have been made to understand transcriptional mechanisms of IL-1 response [16-18]. However, few of the previous studies have validated the global roles of multiple critical TFBMs in downregulation or upregulation of a cluster of genes.
In this principal component analysis, we introduced the Akaike information criterion (AIC) test, singular value decomposition (SVD), and a genetic algorithm (GA) to predict and evaluate TFBMs from a pool of random DNA sequences (Fig. 1). The predictive model was formulated using state vectors, which represented a contribution of potential TFBMs to the IL-1 responses. The promoter matrix was defined to build the quantitative relationship between the mRNA expression vector and the state vectors, and a unique SVD procedure was applied to the promoter matrix. Although one previous study defined the mRNA expression level as a state variable, dynamical correlations among the mRNA levels do not directly represent biological processes [19]. Here, a state variable was defined as an activation level of each TFBM, and SVD was used to link the primary components in the expression vector to the influential TFBM candidates in the state vector through the eigen gene vectors and the eigen TFBM vectors. The analytical prediction of TFBMs with SVD was evaluated numerically using Monte Carlo simulation and GA.
Figure 1 Flowchart of the model-based analysis of transcription factor binding motifs (TFBMs). The mRNA expression data and the human genome sequence information were used to formulate the mathematical model. The putative TFBMs were selected through the Akaike Information Criterion (AIC) analysis and the Singular Value Decomposition (SVD) eigen value analysis. The predicted TFBMs were evaluated with the genetic algorithm (GA) numerical analysis and the Monte-Carlo simulation, and the model-based TFBM network was linked to the known transcription factors and their binding motifs.
Results
Prediction and validation of novel and known TFBMs were conducted using logarithmic ratios of the IL-1-driven mRNA alterations in human chondrocytes (Fig. 1). First, AIC was used to determine a statistically meaningful number of TFBMs in the model. Second, the contribution of each of the 512 TFBM candidates to the IL-1 responses was evaluated by decomposing the promoter matrix with SVD. Third, the SVD-based priority of TFBMs was evaluated numerically by GA and Monte-Carlo simulation. Fourth, a linkage was established among the predicted and known TFBMs.
Messenger RNA ratios and AIC analysis
Using data obtained in primary cultures of human articular chondrocytes, 45 IL-1-responsive genes were selected and the ratios of mRNA levels from IL-1-treated cells against mRNA levels in untreated cells were calculated from the list of IL-1-responsive genes in primary chondrocytes published by Vincenti and Brinckerhoff [13]. As shown in Fig. 2A, the relative mRNA levels are represented in a greyscale, and the logarithmic ratios are illustrated in a green to red color code. The mRNA ratios for 33 genes were positive (upregulation; indicated by green), while the ratios for 12 genes were negative (downregulation; indicated by red). Using Eq. (1) and the SVD procedure, these logarithmic mRNA ratios were modelled against 1 to 32 TFBMs that were chosen from random DNA sequences of 5 bp in length (Fig. 2B). As expected, the model error decreased monotonically as the number of TFBMs increased from 1 to 32. In order to estimate the proper number of TFBMs in the model, AIC was calculated using Eq. (2) (Fig. 2C). The minimum AIC was obtained with 8 TFBMs, which were used as models for further analysis.
Figure 2 Selection of 45 IL-1-responsive genes and AIC analysis. (A) Ratios of mRNA expression in chondrocytes. The grayscale columns marked "-" and "+" represent the mRNA levels without and with the IL-1 treatment, respectively. The color-coded column displays the logarithmic mRNA expression ratio (the mRNA level in cells treated with IL-1 to the untreated control level). The darker color indicates the greater alteration, and "red" and "green" illustrate up- and down-regulation, respectively. (B) Modeled mRNA ratios based on the 300-bp upstream regulatory DNA region. As TFBM candidates, 512 DNA fragments, 5 bp in length, were considered. The mathematical models with (a) 1, (b) 2, (c) 4, (d) 8, (e) 16, and (f) 32 putative TFBMs are illustrated. (C) AIC analysis. The minimum AIC value was obtained when the number of TFBMs was 8.
SVD analysis
Using the SVD procedure, the promoter matrix H, built from the 300-bp upstream flanking sequences, was factorized into three matrices in Eq. (4). Using the eigen gene vectors in U (Fig. 3A) and the eigen values in Λ (Fig. 3B), the observed mRNA ratios were decomposed linearly with definition of the weighing factors, ki (Fig. 3C), in Eq. (5). Out of 45 eigen values, the primary and the secondary eigen values were 133.4 and 64.6. Shannon entropy was calculated as 0.65 [6], and the eigen values suggested a relatively even spread distribution among the 45 eigen gene vectors. Note that that Shannon entropy takes values between 0 and 1, and a smaller value suggests that expression data are dominated by influential eigen values. Using the weighing factors for each of the eigen TFBM vectors, the most influential 8 TFBMs, whose contribution to the expression levels of IL-1-responsive genes was predicted to be larger than the others, were selected. First, the eigen TFBM vectors (Fig. 4A) were derived as a complement of the eigen gene vectors. Then, each TFBM candidate in the eigen TFBM vectors was weighted by the same weighting factors defined in Eq. (5). This weighting process predicted the contributions of TFBM candidates to the observed value of z (Fig. 4B). Lastly, the overall significance to the selected 45 genes was estimated by adding the 45 row elements in the eigen TFBM vectors (Fig. 4C). The predicted TFBM candidates were 5'-CAGGC-3', 5'-CGCCC-3', 5'-CCGCC-3', 5'-CACCG-3', 5'-GCGCC-3', 5'-ATGGG-3', 5'-GGGAA-3', and 5'-CCGCG-3'.
Figure 3 SVD analysis for the 45 IL-1-responsive genes. (A) Forty-five eigen genes in the matrix U in H = UΛVT. (B) Eigen values, λ1, λ2,..., λ45, in the matrix Λ. (C) Weighting factors, ki, for the i-th eigen gene.
Figure 4 SVD-based selection of TFBMs. (A) Eigen TFBM vectors in the matrix VT in H = UΛVT. (B) Weighted eigen TFBM vectors with the weighting factor, ki. (C) Putative TFBMs predicted from the SVD analysis.
GA analysis, Monte-Carlo simulation, and leave-one-out test
In order to evaluate the selection of 8 TFBMs based on the above principal component analysis, the numerical search for TFBM candidates was conducted with the GA analysis. Starting with 200 digital chromosomes in Eq. (6), including the chromosome for the SVD solution, the population of chromosomes was evolved for 104 generations. During evolution, the model error was reduced through artificial chromosome recombinations and mutations (Fig. 5A). The sum square error for the mRNA ratios was 15.94 (SVD solution) and 7.55 (GA solution). These values were smaller than the Monte-Carlo results of 58.97 ± 8.61 (N = 10,000) using a random selection of TFBMs (Fig. 5B). The GA solution reduced the error of the SVD solution by 52.6% by retaining five SVD-driven TFBMs and introducing three new TFBMs, 5'-CGTCC-3', 5'-AAAGG-3', and 5'-ACCCA-3' (Fig. 5C).
Figure 5 GA analysis and Monte-Carlo simulation. (A) Evolution of the model error in the GA analysis during 10,000 generations. (B) Model error in Monte-Carlo simulation. The labels, a and b, indicate the error in the GA analysis and the SVD analysis, respectively. (C) Comparison between the GA-predicted TFBMs and the SVD-predicted TFBMs.
In order to further examine the SVD-based model, we conducted a leave-one-out test. In this test, (N - 1) genes were used to build a model and one gene was used to validate the model through any difference between the observed and the predicted expression levels. The process was repeated N times (N = 45) by removing one gene at a time. The model error for a complete set of leave-one-out tests was 33. To evaluate significance of the leave-one-out model error, Monte-Carlo analyses were conducted using two datasets. In the first dataset the elements in the promoter matrix was reshuffled, and in the second dataset the order of mRNA expression levels was randomized. The model error was 108 ± 31 (mean ± s.d.) and 93 ± 23 for the first and the second datasets, respectively (Fig. 6).
Figure 6 Leave-one-out test. (A) Model error with a randomized promoter matrix. The mean and s.d. of sum square error is 108 ± 31 (N = 1,000), and the label "a" indicates the error with the original promoter sequences. (B) Model error with randomized gene expression ratios. The mean and s.d. of sum square error is 93 ± 23 (N = 1,000), and the label "b" indicates the error with the original gene expression ratios.
Linkage to known TFBMs
The 8 TFBM candidates obtained from the GA analysis were graphically linked to the known TFBMs (Fig. 7). The GA-based TFBMs are represented by 8 boxes in the first column, and each box is linked to the biologically known TFBMs such as AP2, SP1, EGR1, etc. For instance, 5'-CGCCC-3', one of the TFBMs predicted by GA, is part of consensus sequences of SP1, EGR1, KROX, GC-BOX, and ABI4.
Figure 7 Linkage between the predicted TFBMs and the biologically known TFBMs. Eight TFBMs, derived from the GA analysis, were linked to the known biological motifs with the list of consensus sequences. The abbreviations are R (A, G), Y (C, T), K (G, T), W (A, T), S (C, G), B (C, G, T), D (A, G, T), H (A, C, T), V (A, C, G), and N (A, C, G, T). The binding factors (transcription factors) to the consensus sequences are included.
Discussion
In this report, we have presented a predictive model and its validation using the transcriptional responses to IL-1 in human chondrocytes as a model system. From a pool of 512 random DNA sequences of 5 bp in length as potential TFBM candidates, the SVD analysis and the GA simulation both identified 8 TFBMs. Five out of 8 TFBM candidates were identical in both analyses, and several of the known TFBMs, including AP2, EGR1, GC-BOX, SP1, NFκB, and LEF1, coincided with the predicted TFBMs.
Prior to application to the mammalian gene expression in the current study, the described approach was examined to build a model for a Ras/cAMP signal transduction pathway in yeast. This pathway is well characterized in yeast, and a cAMP responsive element (CRE; 5'-[A/G][A/C][T/C]GCAGT-3'), which is conserved in eukaryotes, is known to be involved. The SVD-based approach with 5-bp sequences predicted a part of CRE (5'-AATGC-3') together with two yeast-specific binding motifs such as 5'-AGGGG-3' (binding motif for MSN2/MSN4; stress responsive element) and 5'-ACCGG-3' (binding motif for LEU3). Since both MSN2 and LEU3 are differentially expressed in response to Ras activation [5], the results allowed us to apply this principal component approach to the current study on the human IL-1 responses (see additional file).
In the prediction phase of TFBM analysis, we demonstrated that the SVD analysis prioritized the contribution of individual TFBM candidates, and the GA algorithm was employed to evaluate independently the SVD solution. SVD is computationally inexpensive, and the results are reproducible since no random parameters are involved. It is straightforward to incorporate the effects of degenerate binding sequences by modifying a linear combination of the eigen TFBM vectors and adding contributions from redundant sequences in the final SVD procedure. More specifically, to any TFBM candidate there are 15 degenerate motifs with one base-pair mismatch and the contribution of these degenerate sequences can be included in the model with an appropriate weighting factor. The standard computational complexity of SVD procedure is estimated as O(m2n) or O(mn2) [20]. The complexity can be reduced to O(mn) by implementing the average algorithm or employing parallel computing [21]. GA is a heuristic solver suitable for searching efficiently the suboptimal solutions. There are 1.1 × 1017 combinations to predict 8 TFBMs from 512 candidates in this study. It is virtually impossible to evaluate all combinations, although either SVD- or GA-based TFBM prediction is not globally optimal in terms of minimization of the prediction error. A predicted 5-bp TFBM can represent more than one motif longer than 5 bp sequences.
The use of mathematical and computational procedures such as AIC, SVD, and GA have been used previously to analyze the behaviour of complex biological systems [22,23]. In prediction of TFBMs from the microarray data, however, the described usage here is unique in a novel state-variable representation. Since many genes are regulated by multiple TFBMs, a statistical standard such as AIC may be used appropriately to validate the number of TFBMs that are meaningful in array-derived data. The previous use of SVD has been limited to clustering expression patterns in the eigen gene space [22,24]. The unique feature of the described predictive model is to link the eigen gene space to the eigen TFBM space by applying SVD to the promoter matrix defined from TFBMs. Evolutionary algorithm such as GA has been used to estimate the values of parameters [25,26]. We employed GA to select the set of TFBMs from an artificial chromosome that is composed of on/off switches for 512 random DNA sequences.
The predictive model in this study generated many testable hypotheses on known TFBMs, as well as novel TFBM candidates, and led us to the analysis of transcription factors. Five out of the 8 TFBM candidates were linked to known transcription factors. Among them, AP2 is known to be involved in stress responses [27] and LEF1 is known to be involved in a wnt signalling pathway [28]. However, neither AP2 nor LEF1 is reported to be responsive to IL-1. EGR1 increases expression of inflammatory cytokines and is involved in IL-1-induced downregulation of the type II collagen promoter in chondrocytes [29], and the GC-box is a widely distributed promoter component. The binding site of SP1 is recognized by SP3, which may oppose positive effects of SP1 [30]. NFκB is a pivotal transcription factor that is both induced at the mRNA level, as shown here, and activated by proinflammatory cytokines [31-33]. However, the relatively long degenerate consensus sequence of its binding site 5'-GGG(A/G)(C/A/T)T(T/C)(T/C)CC-3'requires a further linkage analysis to the predicted TFBM of 5'-GGGAA-3'. In a separate study, the promoter competition assay was conducted to evaluate the role of the SVD-selected TFBMs using three IL-1-responsive genes, LIF, NFκB2, and IRF1 [34]. In the assay, the stimulatory effects of 5'-CAGGC-3' and 5'-CGCCC-3', as well as the inhibitory effects of 5'-CCGCC-3', 5'-CACCG-3', and 5'-GCGCC-3', were consistent to the SVD prediction. In order to further validate the stimulatory role of 5'-CAGGC-3', a gel shift assay was conducted. As predicted, incubation with the nuclear extracts isolated from the IL-1-treated cells retarded the mobility of the DNA fragment containing 5'-CAGGC-3' (see additional file).
The described state-variable formulation of the predictive model can be extended to include redundancy in TFBM consensus sequences, temporal mRNA profiles, and interactions of TFBMs with transcription factors and cofactors. Short motifs such as 5 bp TFBMs in this study may present less specificity, but the described SVD procedure can increase specificity easier than any combinatorial search algorithm such as GA. The model can be extended to predict a dynamical state of TFBMs associated with the regulation of the temporal mRNA expression profiles [23]. Interactions among TFBMs through transcription factors and cofactors can be implemented through the nonlinear version [35].
Conclusion
Identification of TFBMs in the human genome is critically important in the post Human Genome Project era [36]. Although experimental evaluation is mandatory to gain biological insights from the model-based predictive results, an analytical model at nearly no computational cost would be useful to provide initial conditions for numerical optimization or predict a set of potential targets for experimental verification. Although the prediction is dependent on definition of regulatory regions, the described model-based analysis allowed us to gain a new biochemical insight on the IL-1 responses by integrating the SVD procedure and Akaike information criterion. In conclusion, the current study on gene responses to IL-1 demonstrates that application of the primary component analysis would predict and validate the novel and known TFBMs from the microarray data using genomic DNA information.
Methods
Determination of mRNA ratios
The mRNA expression data for the IL-1-responsive genes in primary cultures of human articular chondrocytes were obtained from the lists published by Vincenti and Brinckerhoff [13]. The logarithmic ratios of mRNA levels in IL-1β (10 ng/ml)-treated chondrocytes to control mRNA levels were determined for 45 IL-1-responsive genes, whose transcription initiation site was identifiable in the GenBank sequences or by the PEG program [37,38]. The logarithmic ratio makes it easy to characterize both upregulation and downregulation to the control level, and it has been widely used to model array-derived expression data in yeast and human [3,39]. The described SVD-based approach is effective for modelling both upregulated and downregulated genes, and the positive and negative values represent the upregulated and downregulated genes, respectively (Fig. 2A).
Definition of promoter matrix
Prior to mathematical formulation, a promoter matrix HnxM was defined, where n was the number of the IL-1-responsive genes and M was the total number of TFBM candidates. The element hij in HnxM represented the number of appearance of the j-th TFBM candidate on the 5'-end flanking region, 300 bp in length in the current study, of the i-th IL-1-responsive gene. The length of 300 bp was determined to minimize the least-square model error from the upstream regions of 100 bp to 5000 bp with a 100-bp interval. In this study, 512 TFBM candidates (M = 512), 5-bp DNA sequences including 5'-AAAAA-3', 5'-AAAAC-3', etc., were initially screened without considering polarity of DNA strands, and the critical TFBMs were selected by the SVD-based procedures described below. Since the length of motifs varies from 5 to 30 bp in TRANSFAC database, the 5-bp sequences were chosen as a potential core binding motif and their linkage to known motifs with redundancy was considered using TRANSFAC database.
Formulation
Using the promoter matrix Hnxm, the mRNA level of each IL-1-responsive gene was modelled [40]:
z= Hnxmx (1)
where z was the mRNA expression vector representing the logarithmic mRNA ratios for the 45 IL-1-responsive genes, and x was the state vector representing the role of TFBM candidates in achieving the observed values in z. Note that we used M as the total number of TFBM candidates (M = 512 in this study), m as the number of TFBMs in the SVD-based model, and m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaieGacuWFTbqBgaqcaaaa@2E28@ as the estimate of m based on Akaike information criterion below.
Akaike information criterion
In order to avoid underfitting or overfitting the mRNA ratios with TFBM candidates, AIC was defined and used as an indicator of statistical measure [41]:
AIC(m)=−2logL(x¯,m)+2m(2)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGbbqqcqWGjbqscqWGdbWqcqGGOaakcqWGTbqBcqGGPaqkcqGH9aqpcqGHsislcqaIYaGmcyGGSbaBcqGGVbWBcqGGNbWzcqWGmbatcqGGOaakcuWG4baEgaqcgaqhaiabcYcaSiabd2gaTjabcMcaPiabgUcaRiabikdaYiabd2gaTjaaxMaacaWLjaWaaeWaaeaacqaIYaGmaiaawIcacaGLPaaaaaa@47B4@
where L(x^¯
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacuWG4baEgaqcgaqhaaaa@2E5B@
, m) was the likelihood function, and x^¯
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacuWG4baEgaqcgaqhaaaa@2E5B@
was the estimate of x. The value of x^¯
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacuWG4baEgaqcgaqhaaaa@2E5B@
was determined using the singular value decomposition procedure described below. The likelihood of the expression vector, z, with the estimate of the state vector, x^¯
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacuWG4baEgaqcgaqhaaaa@2E5B@
, was calculated:
L(x¯,m)=(2πσ2)−n2exp{−12σ2(z¯−Hnxmx¯)T(z¯−Hnxmx¯)}(3)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@686F@
where σ2 was a model error variance. Prior to constructing the final SVD-based model, a set of preliminary models for m = 1, 2, ..., M were built using the singular value decomposition procedure, and AIC(m) was minimized by treating m as a parameter. Note that AIC(m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaieGacuWFTbqBgaqcaaaa@2E28@) ≤ AIC(m) for m = 1, 2, ..., M, and m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaieGacuWFTbqBgaqcaaaa@2E28@ = 8 in this study.
Singular value decomposition (SVD)
SVD is a matrix decomposition technique which can be applied to any rectangular matrix. It decomposes a matrix into two orthogonal matrices and one eigenvalue matrix. Two orthogonal matrices represent the column and the row spaces in the original matrix, and the eigenvalue matrix relates these two spaces. In order to evaluate the contribution of 512 potential TFBMs to the IL-1 responses, the promoter matrix HnxM was factorized using SVD:
HnxM = UnxnΛnxMVMxMT (4)
where Unxn(u1, u2, ..., un) was defined as the eigen gene matrix, ΛnxM(λ1, λ2, ..., λn;OnXM-n) was a matrix containing n eigen values in the first n column vectors, and VMxM(v1, v2, ..., vM) was defined as an eigen TFBM matrix. Note that Unxn and VMxM are orthogonal and therefore UnxnT Unxn = Inxn and VMxMVMxMT = IMxM. In the Unxn space, the mRNA expression vector, z, can be expressed as a linear combination of the orthogonal vectors u1, u2,..., un and the eigen values λ1, λ2,...,λn with ki (i = 1, 2, ..., n):
z¯=∑i=1nkiλiu¯i(5)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacuWG6bGEgaqhaiabg2da9maaqahabaGaem4AaS2aaSbaaSqaaiabdMgaPbqabaGccqaH7oaBdaWgaaWcbaGaemyAaKgabeaakiqbdwha1zaaDaWaaSbaaSqaaiabdMgaPbqabaaabaGaemyAaKMaeyypa0JaeGymaedabaGaemOBa4ganiabggHiLdGccaWLjaGaaCzcamaabmaabaGaeGynaudacaGLOaGaayzkaaaaaa@4365@
Determination of ki was achieved by projecting the vector z to λiui direction. Therefore, taking an inner product between z and λiui gave ki.
Since ui and vi are the associated bases in the gene space and the TFBM space respectively, the factor ki (i = 1, 2,..., n) for describing the expression in gene space can be used to model the contribution of individual TFBMs in the TFBM space. For instance, let us consider one extreme case where z was parallel to u1. Then, a contribution of TFBMs to z would be proportional to v1 and not affected by the other vi (i ≠ 1) since the eigenvalue matrix ΛnxM does not have any non-diagonal components. Therefore, the elements in v1 would be used to indicate potential importance of M TFBM candidates. In a general case, the SVD procedure allowed us to evaluate n pairs of ui and vi through λi and ki without conducting any combinatorial search. In order to model the gene space using the observed mRNA expression of z, the orthogonal vectors (u1, u2,..., un) are linearly combined using ki (i = 1, 2,..., n). In order to model the TFBM space, a linear combination of the orthogonal vectors (v1, v1,..., vn) is made.
Based on the above rationale, we evaluated the linear combination of the eigen TFBM vectors in a form of a¯=∑i=1nkiv¯i
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacuWGHbqygaqhaiabg2da9maaqahabaGaem4AaS2aaSbaaSqaaiabdMgaPbqabaGccuWG2bGDgaqhamaaBaaaleaacqWGPbqAaeqaaaqaaiabdMgaPjabg2da9iabigdaXaqaaiabd6gaUbqdcqGHris5aaaa@3C20@. This vector a plays the similar role of zin Eq. (5). M elements in a indicates the role and the contribution of M TFBM candidates. The positive/negative value suggests a stimulatory/inhibitory role, and a larger absolute value implies a stronger contribution. Therefore, the procedure to select m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaieGacuWFTbqBgaqcaaaa@2E28@ TFBMs is to choose a set of top m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaieGacuWFTbqBgaqcaaaa@2E28@ TFBMs whose value in a is larger than other (512 - m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaieGacuWFTbqBgaqcaaaa@2E28@) TFBMs. To include redundancy in TFBM consensus sequences, a weighted linear combination of elements in a can be used. In summary, the principal component analysis allows us to identify the principal expression components using the eigen gene vectors and to predict the principal TFBM using the eigen TFBM vectors. With the weighting factors defined from the observed value of z, the vector a indicates the predicted contribution of individual TFBM candidates to the observed expression pattern.
Genetic Algorithm (GA) and Monte Carlo simulations
In order to evaluate the SVD-based prediction of TFBMs, the numerical simulations with GA were conducted using the procedure described previously [42]. In a chromosome-like bit map, 512 TFBM candidates were embedded:
C = [c1, c2,..., c512] (6)
where each chromosomal element took "1" and "0" for inclusion and non-inclusion in the model, respectively. Note that ∑i=1512ci=m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaaeWbqaaiabdogaJnaaBaaaleaacqWGPbqAaeqaaaqaaiabdMgaPjabg2da9iabigdaXaqaaiabiwda1iabigdaXiabikdaYaqdcqGHris5aOGaeyypa0JafmyBa0MbaKaaaaa@3A69@ for any chromosome, and the promoter matrix was constructed based on the value of each chromosomal element ci. Two hundred chromosomes represented the population, and one chromosome in the first generation corresponded to the SVD selection. In each generation, 100 chromosomes with smaller errors were recombined, and the other 100 chromosomes with larger errors were mutated. The model error was defined as |z¯-Hnxmx¯^|2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaabdaqaamXvP5wqSXMqHnxAJn0BKvguHDwzZbqegqvATv2CG4uz3bIuV1wyUbaceiGab8NEayaaDaGaa8xlaiaa=HeadaWgaaWcbaGaa8NBaiaa=HhacaWFTbaabeaakiqa=HhagaqhgaqcaaGaay5bSlaawIa7amaaCaaaleqabaGaa8Nmaaaaaaa@44B3@, and the state variable, x, was estimated using a least-square scheme:
x¯^=(HnxmT Hnxm)−1HnxmT z¯ (7)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacuWG4baEgaqhgaqcaiabg2da9iabcIcaOiabdIeainaaDaaaleaacqWGUbGBcqWG4baEcqWGTbqBaeaacqGGGcaOcqGGGcaOcqGGGcaOcqGGGcaOcqGGGcaOcqGGGcaOcqGGGcaOcqGGGcaOcqWGubavaaGccqWGibasdaWgaaWcbaGaemOBa4MaemiEaGNaemyBa0gabeaakiabcMcaPmaaCaaaleqabaGaeyOeI0IaeGymaedaaOGaemisaG0aa0baaSqaaiabd6gaUjabdIha4jabd2gaTbqaaiabcckaGkabcckaGkabcckaGkabcckaGkabcckaGkabcckaGkabcckaGkabcckaGkabdsfaubaakiqbdQha6zaaDaGaaCzcaiaaxMaadaqadaqaaiabiEda3aGaayjkaiaawMcaaaaa@6824@
Note that n = 45 and m = m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaieGacuWFTbqBgaqcaaaa@2E28@ = 8 in GA. Monte Carlo simulation was also performed to evaluate numerically the SVD- and GA-based selection of TFBMs [42]. A set of m^
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaieGacuWFTbqBgaqcaaaa@2E28@ TFBMs was randomly chosen from 512 TFBM candidates, and the error distribution associated with the randomly selected TFBMs was compared to the error in the model-based prediction. The simulation was conducted 10,000 times.
Linkage map among TFBMs
The 8 TFBM candidates, derived from the GA analysis, were linked to the biologically known TFBMs. We evaluated the 5-bp core consensus sequences identical to the known TFBMs using TRANSFAC database [43]. Since the motifs in the database ranges up to 30 bp, it is possible that a 5-bp TFBM candidate corresponds to multiple motifs in the database. Namely, the state vector could represent the combined role of binding motifs when the predicted motifs are shared among transcription factors.
Supplementary Material
Additional File 1
• Part I – Experimental evaluation of the SVD-based model for IL1 responses. • Part II – SVD analysis for yeast Ras/cAMP signaling pathway.
Click here for file
Acknowledgements
We thank Ying Bai, Sonsy Zachariah, Hui Sun, and Hui Zhao for the data collection and technical support. This study was supported by NIH RR17012, and Indiana 21st century research and technology fund (to H.Y.), and NIH AR46977, and a Veteran's Administration Merit Award (to M.P.V.).
==== Refs
de Jong H Modeling and simulation of genetic regulatory systems: a literature review J Comput Biol 2002 9 67 103 11911796 10.1089/10665270252833208
Lockhart DJ Winzeler EA Genomics, gene expression and DNA arrays Nature 2000 405 827 836 10866209 10.1038/35015701
Bussemaker HJ Li H Siggia ED Regulatory element detection using correlation with expression Nat Genet 2001 27 167 171 11175784 10.1038/84792
Conlon EM Liu XS Lieb JD Liu JS Integrating regulatory motif discovery and genome-wide expression analysis Proc Natl Acad Sci U S A 2003 100 3339 3344 12626739 10.1073/pnas.0630591100
Lander ES Linton LM Birren B Nusbaum C Zody MC Baldwin J Devon K Dewar K Doyle M FitzHugh W Funke R Gage D Harris K Heaford A Howland J Kann L Lehoczky J LeVine R McEwan P McKernan K Meldrim J Mesirov JP Miranda C Morris W Naylor J Raymond C Rosetti M Santos R Sheridan A Sougnez C Stange-Thomann N Stojanovic N Subramanian A Wyman D Rogers J Sulston J Ainscough R Beck S Bentley D Burton J Clee C Carter N Coulson A Deadman R Deloukas P Dunham A Dunham I Durbin R French L Grafham D Gregory S Hubbard T Humphray S Hunt A Jones M Lloyd C McMurray A Matthews L Mercer S Milne S Mullikin JC Mungall A Plumb R Ross M Shownkeen R Sims S Waterston RH Wilson RK Hillier LW McPherson JD Marra MA Mardis ER Fulton LA Chinwalla AT Pepin KH Gish WR Chissoe SL Wendl MC Delehaunty KD Miner TL Delehaunty A Kramer JB Cook LL Fulton RS Johnson DL Minx PJ Clifton SW Hawkins T Branscomb E Predki P Richardson P Wenning S Slezak T Doggett N Cheng JF Olsen A Lucas S Elkin C Uberbacher E Frazier M Gibbs RA Muzny DM Scherer SE Bouck JB Sodergren EJ Worley KC Rives CM Gorrell JH Metzker ML Naylor SL Kucherlapati RS Nelson DL Weinstock GM Sakaki Y Fujiyama A Hattori M Yada T Toyoda A Itoh T Kawagoe C Watanabe H Totoki Y Taylor T Weissenbach J Heilig R Saurin W Artiguenave F Brottier P Bruls T Pelletier E Robert C Wincker P Smith DR Doucette-Stamm L Rubenfield M Weinstock K Lee HM Dubois J Rosenthal A Platzer M Nyakatura G Taudien S Rump A Yang H Yu J Wang J Huang G Gu J Hood L Rowen L Madan A Qin S Davis RW Federspiel NA Abola AP Proctor MJ Myers RM Schmutz J Dickson M Grimwood J Cox DR Olson MV Kaul R Shimizu N Kawasaki K Minoshima S Evans GA Athanasiou M Schultz R Roe BA Chen F Pan H Ramser J Lehrach H Reinhardt R McCombie WR de la Bastide M Dedhia N Blocker H Hornischer K Nordsiek G Agarwala R Aravind L Bailey JA Bateman A Batzoglou S Birney E Bork P Brown DG Burge CB Cerutti L Chen HC Church D Clamp M Copley RR Doerks T Eddy SR Eichler EE Furey TS Galagan J Gilbert JG Harmon C Hayashizaki Y Haussler D Hermjakob H Hokamp K Jang W Johnson LS Jones TA Kasif S Kaspryzk A Kennedy S Kent WJ Kitts P Koonin EV Korf I Kulp D Lancet D Lowe TM McLysaght A Mikkelsen T Moran JV Mulder N Pollara VJ Ponting CP Schuler G Schultz J Slater G Smit AF Stupka E Szustakowski J Thierry-Mieg D Thierry-Mieg J Wagner L Wallis J Wheeler R Williams A Wolf YI Wolfe KH Yang SP Yeh RF Collins F Guyer MS Peterson J Felsenfeld A Wetterstrand KA Patrinos A Morgan MJ Szustakowki J de Jong P Catanese JJ Osoegawa K Shizuya H Choi S Chen YJ Initial sequencing and analysis of the human genome Nature 2001 409 860 921 11237011 10.1038/35057062
Thompson W Palumbo MJ Wasserman WW Liu JS Lawrence CE Decoding human regulatory circuits Genome Res 2004 14 1967 1974 15466295 10.1101/gr.2589004
Karolchik D Baertsch R Diekhans M Furey TS Hinrichs A Lu YT Roskin KM Schwartz M Sugnet CW Thomas DJ Weber RJ Haussler D Kent WJ The UCSC Genome Browser Database Nucleic Acids Res 2003 31 51 54 12519945 10.1093/nar/gkg129
Gupta M Liu JS Discovery of Conserved Sequence Patterns Using a Stochastic Dictionary Model Journal of the American Statistical Association 2003 461 98
Grad YH Roth FP Halfon MS Church GM Prediction of similarly acting cis-regulatory modules by subsequence profiling and comparative genomics in Drosophila melanogaster and D.pseudoobscura Bioinformatics 2004 20 2738 2750 15145800 10.1093/bioinformatics/bth320
Sharan R Ovcharenko I Ben-Hur A Karp RM CREME: a framework for identifying cis-regulatory modules in human-mouse conserved segments Bioinformatics 2003 19 Suppl 1 i283 91 12855471 10.1093/bioinformatics/btg1039
Keles S van der Laan M Eisen MB Identification of regulatory elements using a feature selection method Bioinformatics 2002 18 1167 1175 12217908 10.1093/bioinformatics/18.9.1167
Xu Y Selaru FM Yin J Zou TT Shustova V Mori Y Sato F Liu TC Olaru A Wang S Kimos MC Perry K Desai K Greenwald BD Krasna MJ Shibata D Abraham JM Meltzer SJ Artificial neural networks and gene filtering distinguish between global gene expression profiles of Barrett's esophagus and esophageal cancer Cancer Res 2002 62 3493 3497 12067993
Vincenti MP Brinckerhoff CE Early response genes induced in chondrocytes stimulated with the inflammatory cytokine interleukin-1beta Arthritis Res 2001 3 381 388 11714393 10.1186/ar331
Heller RA Schena M Chai A Shalon D Bedilion T Gilmore J Woolley DE Davis RW Discovery and analysis of inflammatory disease-related genes using cDNA microarrays Proc Natl Acad Sci U S A 1997 94 2150 2155 9122163 10.1073/pnas.94.6.2150
Elliott SF Coon CI Hays E Stadheim TA Vincenti MP Bcl-3 is an interleukin-1-responsive gene in chondrocytes and synovial fibroblasts that activates transcription of the matrix metalloproteinase 1 gene Arthritis Rheum 2002 46 3230 3239 12483727 10.1002/art.10675
Chadjichristos C Ghayor C Kypriotou M Martin G Renard E Ala-Kokko L Suske G de Crombrugghe B Pujol JP Galera P Sp1 and Sp3 transcription factors mediate interleukin-1 beta down-regulation of human type II collagen gene expression in articular chondrocytes J Biol Chem 2003 278 39762 39772 12888570 10.1074/jbc.M303541200
Francois M Richette P Tsagris L Raymondjean M Fulchignoni-Lataud MC Forest C Savouret JF Corvol MT Peroxisome proliferator-activated receptor-gamma down-regulates chondrocyte matrix metalloproteinase-1 via a novel composite element J Biol Chem 2004 279 28411 28418 15090544 10.1074/jbc.M312708200
Imamura T Imamura C Iwamoto Y Sandell LJ Transcriptional Co-activators CREB-binding protein/p300 increase chondrocyte Cd-rap gene expression by multiple mechanisms including sequestration of the repressor CCAAT/enhancer-binding protein J Biol Chem 2005 280 16625 16634 15722556 10.1074/jbc.M411469200
Gardner TS di Bernardo D Lorenz D Collins JJ Inferring genetic networks and identifying compound mode of action via expression profiling Science 2003 301 102 105 12843395 10.1126/science.1081900
Troyanskaya O Cantor M Sherlock G Brown P Hastie T Tibshirani R Botstein D Altman RB Missing value estimation methods for DNA microarrays Bioinformatics 2001 17 520 525 11395428 10.1093/bioinformatics/17.6.520
Chuang HYH Chen L Efficient Computation of the Singlular Value Decomposition on Cube Connected SIMD Machine: Reno.
1989 276 282
Alter O Brown PO Botstein D Singular value decomposition for genome-wide expression data processing and modeling Proc Natl Acad Sci U S A 2000 97 10101 10106 10963673 10.1073/pnas.97.18.10101
Liu Y Sun HB Yokota H Regulating gene expression using optimal control theory Proc 3rd IEEE Sym Bioinfo Bioeng 2003 1 3
Holter NS Mitra M Maritan A Cieplak M Banavar JR Fedoroff NV Fundamental patterns underlying gene expression profiles: simplicity from complexity Proc Natl Acad Sci U S A 2000 97 8409 8414 10890920 10.1073/pnas.150242097
Holland JH Adaptation in natural and artificial systems 1975 Ann Arbor , The University of Michigan Press
Li L Weinberg CR Darden TA Pedersen LG Gene selection for sample classification based on gene expression data: study of sensitivity to choice of parameters of the GA/KNN method Bioinformatics 2001 17 1131 1142 11751221 10.1093/bioinformatics/17.12.1131
Grether-Beck S Buettner R Krutmann J Ultraviolet A radiation-induced expression of human genes: molecular and photobiological mechanisms Biol Chem 1997 378 1231 1236 9426182
Eastman Q Grosschedl R Regulation of LEF-1/TCF transcription factors by Wnt and other signals Curr Opin Cell Biol 1999 11 233 240 10209158 10.1016/S0955-0674(99)80031-3
Tan L Peng H Osaki M Choy BK Auron PE Sandell LJ Goldring MB Egr-1 mediates transcriptional repression of COL2A1 promoter activity by interleukin-1beta J Biol Chem 2003
Philipsen S Suske G A tale of three fingers: the family of mammalian Sp/XKLF transcription factors Nucleic Acids Res 1999 27 2991 3000 10454592 10.1093/nar/27.15.2991
Vincenti MP Coon CI Brinckerhoff CE Nuclear factor kappaB/p50 activates an element in the distal matrix metalloproteinase 1 promoter in interleukin-1beta-stimulated synovial fibroblasts Arthritis Rheum 1998 41 1987 1994 9811054 10.1002/1529-0131(199811)41:11<1987::AID-ART14>3.0.CO;2-8
Ding GJ Fischer PA Boltz RC Schmidt JA Colaianne JJ Gough A Rubin RA Miller DK Characterization and quantitation of NF-kappaB nuclear translocation induced by interleukin-1 and tumor necrosis factor-alpha. Development and use of a high capacity fluorescence cytometric system J Biol Chem 1998 273 28897 28905 9786892 10.1074/jbc.273.44.28897
Barnes PJ Karin M Nuclear factor-kappaB: a pivotal transcription factor in chronic inflammatory diseases N Engl J Med 1997 336 1066 1071 9091804 10.1056/NEJM199704103361506
Sun HB Malacinski GM Yokota H Promoter competition assay for analyzing gene regulation in joint tissue engineering Front Biosci 2002 7 a169 74 12133815
Sun HB Liu Y Qian L Yokota H Model-based analysis of matrix metalloproteinase expression under mechanical shear Ann Biomed Eng 2003 31 171 180 12627825 10.1114/1.1540635
Collins FS Green ED Guttmacher AE Guyer MS A vision for the future of genomics research Nature 2003 422 835 847 12695777 10.1038/nature01626
Venter JC Adams MD Myers EW Li PW Mural RJ Sutton GG Smith HO Yandell M Evans CA Holt RA Gocayne JD Amanatides P Ballew RM Huson DH Wortman JR Zhang Q Kodira CD Zheng XH Chen L Skupski M Subramanian G Thomas PD Zhang J Gabor Miklos GL Nelson C Broder S Clark AG Nadeau J McKusick VA Zinder N Levine AJ Roberts RJ Simon M Slayman C Hunkapiller M Bolanos R Delcher A Dew I Fasulo D Flanigan M Florea L Halpern A Hannenhalli S Kravitz S Levy S Mobarry C Reinert K Remington K Abu-Threideh J Beasley E Biddick K Bonazzi V Brandon R Cargill M Chandramouliswaran I Charlab R Chaturvedi K Deng Z Di Francesco V Dunn P Eilbeck K Evangelista C Gabrielian AE Gan W Ge W Gong F Gu Z Guan P Heiman TJ Higgins ME Ji RR Ke Z Ketchum KA Lai Z Lei Y Li Z Li J Liang Y Lin X Lu F Merkulov GV Milshina N Moore HM Naik AK Narayan VA Neelam B Nusskern D Rusch DB Salzberg S Shao W Shue B Sun J Wang Z Wang A Wang X Wang J Wei M Wides R Xiao C Yan C Yao A Ye J Zhan M Zhang W Zhang H Zhao Q Zheng L Zhong F Zhong W Zhu S Zhao S Gilbert D Baumhueter S Spier G Carter C Cravchik A Woodage T Ali F An H Awe A Baldwin D Baden H Barnstead M Barrow I Beeson K Busam D Carver A Center A Cheng ML Curry L Danaher S Davenport L Desilets R Dietz S Dodson K Doup L Ferriera S Garg N Gluecksmann A Hart B Haynes J Haynes C Heiner C Hladun S Hostin D Houck J Howland T Ibegwam C Johnson J Kalush F Kline L Koduru S Love A Mann F May D McCawley S McIntosh T McMullen I Moy M Moy L Murphy B Nelson K Pfannkoch C Pratts E Puri V Qureshi H Reardon M Rodriguez R Rogers YH Romblad D Ruhfel B Scott R Sitter C Smallwood M Stewart E Strong R Suh E Thomas R Tint NN Tse S Vech C Wang G Wetter J Williams S Williams M Windsor S Winn-Deen E Wolfe K Zaveri J Zaveri K Abril JF Guigo R Campbell MJ Sjolander KV Karlak B Kejariwal A Mi H Lazareva B Hatton T Narechania A Diemer K Muruganujan A Guo N Sato S Bafna V Istrail S Lippert R Schwartz R Walenz B Yooseph S Allen D Basu A Baxendale J Blick L Caminha M Carnes-Stine J Caulk P Chiang YH Coyne M Dahlke C Mays A Dombroski M Donnelly M Ely D Esparham S Fosler C Gire H Glanowski S Glasser K Glodek A Gorokhov M Graham K Gropman B Harris M Heil J Henderson S Hoover J Jennings D Jordan C Jordan J Kasha J Kagan L Kraft C Levitsky A Lewis M Liu X Lopez J Ma D Majoros W McDaniel J Murphy S Newman M Nguyen T Nguyen N Nodell M Pan S Peck J Peterson M Rowe W Sanders R Scott J Simpson M Smith T Sprague A Stockwell T Turner R Venter E Wang M Wen M Wu D Wu M Xia A Zandieh A Zhu X The sequence of the human genome Science 2001 291 1304 1351 11181995 10.1126/science.1058040
Davuluri RV Grosse I Zhang MQ Computational identification of promoters and first exons in the human genome Nat Genet 2001 29 412 417 11726928 10.1038/ng780
Liu Y Yokota H Modelling and idenification of transcription-factor binding motifs in human chondrogenesis Systems Biology 2004 1 85 92 17052118 10.1049/sb:20045012
Qian L Liu Y Sun HB Yokota H Systems analysis of matrix metalloproteinase mRNA expression in skeletal tissues Front Biosci 2002 7 a126 34 12045007
Akaike H A new look at the statistical model identification IEEE Transactions on Automatic Control 1974 AC-19 716 723 10.1109/TAC.1974.1100705
Liu Y Yokota H Modelling and identification of transcription-factor binding motifs in human chondrogenesis Systems Biology 2004 1 85 92 17052118 10.1049/sb:20045012
Wingender E Dietze P Karas H Knuppel R TRANSFAC: a database on transcription factors and their DNA binding sites Nucleic Acids Res 1996 24 238 241 8594589 10.1093/nar/24.1.238
|
16297243
|
PMC1316881
|
CC BY
|
2021-01-04 16:27:47
|
no
|
BMC Bioinformatics. 2005 Nov 18; 6:276
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-276
|
oa_comm
|
==== Front
PLoS Comput BiolPLoS Comput. BiolpcbiplcbploscompPLoS Computational Biology1553-734X1553-7358Public Library of Science San Francisco, USA 1647732510.1371/journal.pcbi.001007405-PLCB-RA-0230R1plcb-01-07-08Research ArticleBioinformatics - Computational BiologySystems BiologyVirusesFunctional Alignment of Regulatory Networks: A Study of Temperate Phages Alignment of Regulatory NetworksTrusina Ala 12*Sneppen Kim 2Dodd Ian B 3Shearwin Keith E 3Egan J. Barry 31 Department of Theoretical Physics, Umeå University, Umeå, Sweden
2 Niels Bohr Institute, Copenhagen, Denmark
3 Discipline of Biochemistry, School of Molecular and Biomedical Science, University of Adelaide, South Australia, Australia
Holmes Eddie EditorPennsylvania State University, United States of America* To whom correspondence should be addressed. E-mail: [email protected] 2005 23 12 2005 1 7 e746 9 2005 11 11 2005 Copyright: © 2005 Trusina 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 relationship between the design and functionality of molecular networks is now a key issue in biology. Comparison of regulatory networks performing similar tasks can provide insights into how network architecture is constrained by the functions it directs. Here, we discuss methods of network comparison based on network architecture and signaling logic. Introducing local and global signaling scores for the difference between two networks, we quantify similarities between evolutionarily closely and distantly related bacteriophages. Despite the large evolutionary separation between phage λ and 186, their networks are found to be similar when difference is measured in terms of global signaling. We finally discuss how network alignment can be used to pinpoint protein similarities viewed from the network perspective.
Synopsis
Networks of interacting genes and proteins orchestrate the complex functions of every living cell. Decoding the logic of these biochemical circuits is a central challenge facing biology today. Trusina et al. describe a mathematical method for aligning two regulatory networks based on their signaling properties, and apply it to a case study of three bacteriophages, simple biological “computers” whose genetics are exceptionally well characterized. The comparison reveals a surprising similarity between regulatory networks of the creatures, even when they have distant evolutionary relationships. The method introduced here should be applicable to other networks, and thus help to illuminate the computational substructures of living systems.
Citation:Trusina A, Sneppen K, Dodd IB, Shearwin KE, Egan JB (2005) Functional alignment of regulatory networks: A study of temperate phages. PLoS Comput Biol 1(7): e74.
==== Body
Introduction
The functioning of living organisms is based on an intricate network of genes and proteins regulating each other. Various organisms differ due to not only differences in the constituting components (genes/proteins) but also the organization of these regulatory networks. It is, therefore, important to address similarities and differences in not only protein sequences but also the interaction patterns of the proteins. Thus, large-scale analysis of protein–protein and protein–DNA interactions have provided insight into the local design features of subcellular signaling [1–3]; network alignment based on sequence similarities permits alignment of related motifs [4,5].
Here we suggest comparison of networks through an alignment method that is based solely on network architecture and signaling logic, which thus does not rely on sequence similarity of the involved proteins.
As a case study, we considered the regulatory networks of two well-characterized temperate bacteriophages of E. coli, λ and 186 (Figure 1). These two phages represent two distinct classes of temperate bacteriophages: the lambdoid phages—which include λ, P22, 434, HK97, and HK022, and the P2 group—which includes P2, 186, HP1, K139, and PSP3. λ and 186 are not detectably related in sequence and have different genome organizations. Using tBLASTx [6] to compare all of the reading frames, there are only two clearly homologous protein pairs: the λ endolysin R/186 (E-score = 10−34) and a pair of early lytic proteins of unknown function (E-score = 2 × 10−4). No significant similarity was detectable at the nucleotide level (using BLASTn, [6]). On the genome level, the arrangement of genes, promoters, and operators is very different [7–10]. As a control of methodology, we also considered the P22 phage, which, as a member of the lambdoid family, allows us to compare topologies of evolutionarily related networks.
Figure 1 The Genetic Regulatory Networks for Phage 186, Phage λ, and Phage P22, All of Which Are Temperate and Infect E. coli
The proteins are colored according to their functions and expression mode in the lysis–lysogeny life cycle of the phages. We summarize the influence of one protein on another by either a green arrow (positive, e.g., transcriptional activation) or a red arrow (negative, e.g., repression). The dashed lines show relatively weak regulations.
As a temperate phage, 186 and λ each can be in two states: a lytic state where many proteins are active in the replication of the phage DNA and the construction and release of virus particles; and a lysogenic state where the phage genome is integrated into the bacterial chromosome and only a few proteins are active. For both phages, three core proteins (CI, Cro, and CII in λ, and CI, Apl, and CII in 186) do the main computations, with the switch into lysogeny being coordinated by CII and the reverse switch into the lytic mode initiated by activation of the host SOS response recombination (RecA) protein. The gene-regulatory networks of all temperate phages have evolved to provide lysogenic and lytic states, and, moreover, to switch from one state to another when particular signals have been received from bacterial proteins, and thus effectively perform the same function.
Given that 186 and λ are both temperate, i.e., performing a similar function, but evolutionarily separated, we asked whether we could detect structural similarities, and at what scale these similarities were detectable.
Results
Visual comparison of the 186 and λ networks (see Figure 1) suggests both strong similarities but also major differences. One way to quantify the similarity of two networks is by edit distance [11]. Assume that we know which nodes (here, proteins) in networks A and B should be paired. For networks of the same size, we define edit distance as the number of insertions or removals of edges (regulatory connections) one has to perform on network A to obtain B. This is quantified as
The elements Aij and Bij specify whether the direct regulation of i on protein j is positive, negative, or absent, and are constructed such that each element can keep both positive and negative links (for details, see equation 2 below).
In case we do not know which nodes in networks A and B should be paired, we find the optimal identification by minimizing DE as described in the Materials and Methods section. This yields the minimal distance between the networks, as well as an optimal alignment of the individual nodes. We call this distance the edit difference.
The minimal edit difference between related phages is small, DE (λ, P22) = 18, compared with the larger scores for evolutionarily separated phages (Table 1). DE = 18 means that the λ network of 62 proteins and 144 connections can be constructed by making 18 edits of the connections in a 62-protein subset of the 67-protein P22 network (adding or removing a link is a single edit; changing the sign of a connection is two edits). To get an idea of the significance of the obtained DE values, we compared them with optimal alignments of 500 randomized versions of the two networks. The randomization procedure was designed to conserve the local properties of the networks to try to keep their general biological features. First, the core-hub topology common in biological networks [3] was maintained by conserving for each protein the number of its regulators (inputs) and the number of proteins regulated by it (outputs). Second, the number of each sign (positive and negative) of the input and output connections was kept for each node.
Table 1 The Overall Difference Measures, DE, DS, between the Networks, with Respective P-Scores as Defined in Text
The constraint of preserving the local properties does not fix the network completely: while keeping the number of positively or negatively regulated proteins, one can still change which of them are being regulated. The structure of the resulting random networks is rather different, as seen in the examples shown in Figure 2.
Figure 2 Illustration of the Differences between the Real 186 and λ Networks (Top) and an Example of Their Randomized Counterparts (Bottom)
These examples of randomized networks show that it possible to preserve local properties, yet obtain different network structures.
Overall, we found that DE scores between any pair of randomized networks are similar. When comparing scores between real networks with those of their random counterparts in Table 1, one sees no clear trend. In particular, the differences between these randomized versions for λr and 186r were indistinguishable from those of the real networks: DE (186r, λr) = 32 ± 2.
We reasoned that the functional similarity of networks might be better reflected in a less local measure of functionality. We therefore introduced a signaling difference, DS, which aims to capture both direct (as in DE) and indirect regulation through a sequence of intermediate proteins. For each pair of proteins (i, j), we considered whether i sends a signal to j, and if so whether the signal along the shortest path is positive or negative. In this spirit, we define the sign of a signal as the product of the signs of all links on the shortest path from i to j. An example where this procedure nicely reflects the functionality in terms of its Boolean logic [12] is found in the pathway from RecA to CI in the two phages. In λ, active RecA directly catalyzes self-cleavage of CI [13]; whereas in 186, RecA acts through the degradation of a repressor protein (LexA), which in turn represses the protein Tum [9], which in the absence of repression binds CI and prevents it from performing its function. Thus, the simple −1 signal in λ is in 186 replaced by a signaling consisting of (−1) × (−1) × (−1) = −1. In other words, repressing a repressor is effectively an activation.
Because the regulation of one protein by another may be positive through one series of links and negative through another, two matrices were used for each network, one for positive signals (AS+ and BS+) and one for negative signals (AS− and BS−). If the effect of protein i on protein j is only positive, then one is placed into
and zero into
. If the effect is only negative, then zero is placed into
and one into
. If there are positive and negative signals along paths of equal length (e.g., from RecA to λ CII via LexA or CI), then one is placed into both matrices. Observe that when positive and negative signals come to the same node, they are not canceling each other. This is intentional, as often signals will arrive at different times or at different conditions. (An example of this is the two paths from RecA to CII over CI and LexA, respectively, of which only the RecA-LexA-CII path is activated during lysis.)
The signaling difference between two networks A and B is then defined as
which takes into account differences in both positive and negative signaling along the shortest paths between any pair of nodes. Like DE, the minimum difference DS is calculated by optimizing which proteins in 186 should be identified with which proteins in λ, and, in addition, which λ proteins should be excluded. Excluding a protein means that the signaling to and from that protein is not counted in DS, whereas signaling across the excluded protein is included.
Optimizing protein alignment based on signaling, we found that DS (186, λ) = 43. Again, the significance of this difference was determined by repeatedly performing randomization of the networks as described above, creating the AS+ and AS− matrices and obtaining the minimal DS. The differences between random networks, DS (186r, λr) = 109 ± 33, is much larger than between the real networks. This is further quantified by a P-score, P (DS > DS (random)) = 0.01, defined as the probability that two randomized networks will have a smaller difference than that between the real networks.
Thus, all three networks are similar in their signaling pattern. To confirm that this signaling similarity is not generally conserved among biological networks, we compared the phage networks with other networks that perform different functions (e.g., the Saccharomyces cerevisiae cell-cycle network [14], and the Bacillus subtilis competence networks [15]). We found that DS is much larger and the P-scores are close to one in these alignments, indicating that the low signaling difference between the phage networks is a special property of these functionally similar networks.
We also considered other variants of the difference measures, in particular including all non-repetitive paths between pairs of proteins, with all paths weighted equally. In that case, we also found that DS − all (λ, 186) = 390 between real networks is smaller than DS − all (λr, 186r) = 583 ± 122 between the randomized counterparts. Also, using the shortest paths, we investigated differences between networks where weak links (see the dashed ones in Figure 1) are weighted less (by a factor of 0.5 or removed altogether). DS scores between networks got smaller, but overall significance remained similar.
Discussion
The pathway-related DS score allowed us to identify significant similarity between two distantly related biological networks (see Table 1). In contrast, the edit difference measure, which looks only at the local wiring structure, is sometimes blind to this more global “homology.” Thus, although edit difference partially captures network similarities through a patchwork of local matchings, it is less sensitive to pathway disruptions.
It is not clear whether the functional similarity between the λ and 186 networks detected by the DS measure is a result of convergent evolution or is a remnant of a shared ancestral network. In either scenario, it is clear that the two network structures must be strongly constrained by functional requirements, given the evolutionary separation of the two phages. A potential bias should be noted here: knowledge of the three phage networks is not complete, even for λ, and it is thus possible that some of the observed similarity in the networks is due to knowledge of connections in one phage network having influenced the discovery of connections in the others.
The DS alignment allows us to address the role of various proteins in pathway disruptions. Figure 3 lines up the λ and 186 proteins on the basis of pre-existing knowledge of their function or mode of expression and indicates the optimal DS alignment and the contribution of each pair to the signaling difference. The two alignments show good matches for late lytic genes as well as for the regulators CI, CII, and B from 186 aligned with CI, CII, and Q in λ. Thus, in general, functions of proteins in one network teach us about protein properties in the other network. The lack of a good match between Apl (in 186) and Cro (in λ) is due to the weak links from Cro and reflects a different functional role of Cro and Apl in the late lytic development of phages. Insisting on alignment of Cro with Apl results in DS = 219, thus emphasizing the particular role of Cro as a repressor of late lysis in λ.
Figure 3 Alignment of Two Phage Networks
Placement of proteins is based on our knowledge [7–9,13,22], and the lines connecting them are associated with the minimal DS alignment. Proteins that perform similar functions or are regulated similarly are placed on the same level; thus, horizontal lines mark ideal matching. Blue lines correspond to meaningful alignments, and red lines are the misalignments. The numbers above the lines, di, reflect the differences in signaling between the aligned proteins and are the contributions to the minimal difference
. The numbers in the parentheses indicate multiple equivalent proteins, making the sum of all shown signaling differences equal to 2 × 43. The key regulators RecA, LexA, and CI are identified correctly, whereas the misidentification of CII with CIII is reasonable since both favor entry into lysogeny through the same pathway. The major discrepancy is associated with the different roles of Cro and Apl during lysis (the weak links from Cro to Q and N in λ).
Comparison of molecular networks is becoming an important element of modern systems biology, both with regard to predicting eventual missing links [16], and to increasing our understanding of functionality of information processing in the networks. The alignment methods presented here address the similarities on a local, respectively larger scale, associated with signaling across networks.
In this regard, we found that evolutionary relationships (λ − P22) imply similar local regulation, with a low DE score. For all temperate phages, evolved to do similar “computation,” their regulatory networks are found to be similar when viewed from a more global perspective where both direct and indirect signals are included (low DS score compared to random expectation). Thus the mechanistic and structural differences on the scale of genome and promoter organization disappear when considering the large scale of the protein regulatory networks. Going beyond immediate regulations allows us to capture functional similarity in the most robust way.
Materials and Methods
The present paper is based on the data on three bacteriophages: λ, P22, and 186. The regulatory networks were compiled from these database entries and various literature sources: for λ ([7,13,
17,18,] and references therein), for 186 ([8–10] and references therein), and for P22 ([19] and references therein).
In the Results section, we define two differences scores, DE and DS, between a pair of networks A and B. Provided that we know which proteins in A should be identified with which in B, the scores are calculated as in equation 1 and equation 2. In case we do not know which nodes in networks A and B should be paired, we need to find the optimal identification of nodes between them. To do so, we define an alignment procedure through the Metropolis Algorithm [20], designed to reach the minimal distance D between the networks: given two nodes and their corresponding partners in the other network, the elementary step is to switch partners and re-evaluate the distance. Iterating this procedure and using simulated annealing [21], the difference score between the two networks converges to a global minimum.
If the two networks are of different sizes, we count only the contribution from a number of nodes given by the smaller of the two networks. In the larger network, these nodes are selected to minimize the distance using the above algorithm.
We would like to note that the above method is not intended to reflect any evolutionary process, but is used to find the optimal mapping of pairs of proteins that look similar from the network perspective. The method is limited by the network size, and in practice it works for networks of fewer than 200 nodes.
The realization of the alignment algorithm in the form of the Java applet (Sun Microsystems, Santa Clara, California, United States) is available at http://www.cmol.nbi.dk/models/compar/compar.html.
Supporting Information
Accession Numbers
The GenBank (http://www.ncbi.nlm.nih.gov) accession numbers for the genes and gene products discussed in this paper are: P22 (NC002371) and λ (J02459).
The Swiss-Prot (http://ca.expasy.org/sprot) accession numbers for genes and gene products discussed in this paper are: 186 (PO80309), Apl (P21681), CI (P03034 and P08707), CII (P03042 and P21678), CIII (P03044), Cro (P03040), LexA (P03033), R (P03706 and PO80309), RecA (P03017), and Tum (P41063).
We warmly thank S. Brown, S. Krishna, and S. Strogatz for constructive comments on the manuscript. The work was supported by Swedish Research Council Grants 621 2003 6290 and 629 2002 6258, and by the Danish National Research Foundation through the center Models of Life at the Niels Bohr Institute. Work in the Egan lab is supported by the National Institutes of Health.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. AT, KS, IBD, KES, and JBE analyzed the data and wrote the paper.
==== Refs
References
Jeong H Tombor B Albert R Oltvai ZN Barabási AL 2000 The large-scale organization of metabolic networks Nature 407 651 654 11034217
Shen-Orr S Milo R Mangan S Alon U 2002 Network motifs in the transcriptional regulation network of Escherichia coli Nat Genet 31 64 68 11967538
Maslov S Sneppen K 2002 Specificity and stability in topology of protein networks Science 296 910 913 11988575
Kelley PB Yuan B Lewitter F Sharan R Stockwell BR 2004 PathBLAST: A tool for alignment of protein interaction networks Nucleic Acids Res 32 83 88
Berg J Lassig M 2004 Local graph alignment and motif search in biological networks Proc Natl Acad Sci U S A 101 14689 14694 15448202
Tatusova TA Maddenm TL 1999 Blast 2 sequences—A new tool for comparing protein and nucleotide sequences FEMS Microbiol Lett 174 247 250 10339815
Dodd IB Shearwin KE Egan JB 2005 Revisited gene regulation in bacteriophage lambda Curr Opin Genet Dev 15 145 152 15797197
Portelli R Dodd IB Xue Q Egan JB 1998 The late-expressed region of the temperate coliphage 186 genome Virology 248 117 130 9705261
Shearwin KE Brumby AM Egan JB 1998 The Tum protein of coliphage 186 is an antirepressor J Biol Chem 273 5708 5715 9488703
Dodd IB Egan JB 2002 Action at a distance in CI repressor regulation of the bacteriophage 186 genetic switch Mol Microbiol 45 697 710 12139616
Bunke H 1997 On a relation between graph edit distance and maximum common subgraph Pattern Recognit Lett 18 689 694
Kauffman SA 1993 The origins of order: Self organization and selection in evolution New York Oxford University Press 709 p.
Little JW 1984 Autodigestion of lexA and phage lambda repressors Proc Natl Acad Sci U S A 81 1375 1379 6231641
Li F Long T Lu Y Ouyang Q Tang C 2004 The yeast cell-cycle network is robustly designed Proc Natl Acad Sci U S A 101 4781 4786 15037758
Hamoen LW Venema G Kuipers OP 2003 Controlling competence in Bacillus subtilis: Shared use of regulators Microbiology 149 9 17 12576575
Albert I Albert R 2004 Conserved network motifs allow protein–protein interaction prediction Bioinformatics 20 3346 3352 15247093
Hendrix RW Roberts JW Stahl FW Weisberg RA 1983 Lambda II Cold Spring Harbor (New York) Cold Spring Harbor Laboratory Press 694 p.
Kobiler O Oppenheim AB Herman C 2004 Recruitment of host ATP-dependent proteases by bacteriophage lambda J Struct Biol 146 72 78 15037238
Pedulla ML Ford ME Karthikeyan T Houtz JM Hendrix RW 2003 Corrected sequence of the bacteriophage P22 genome J Bacteriol 18 1475 1477
Metropolis M Rosenbluth AW Rosenbluth MN Teller AH Teller E 1953 Equation of state calculations by fast computing machines J Chem Phys 21 1087 1092
Kirkpatrick S Gelatt CD Jr Vecchi MP 1983 Optimization by simulated annealing Science 220 671 680 17813860
Kobiler O Rokney A Friedman N Court DL Stavans J 2005 Quantitative kinetic analysis of the bacteriophage lambda genetic network Proc Natl Acad Sci U S A 102 4470 4475 15728384
|
16477325
|
PMC1317652
|
CC BY
|
2021-01-05 09:18:23
|
no
|
PLoS Comput Biol. 2005 Dec 23; 1(7):e74
|
utf-8
|
PLoS Comput Biol
| 2,005 |
10.1371/journal.pcbi.0010074
|
oa_comm
|
==== Front
PLoS PathogPLoS PathogppatplpaplospathPLoS Pathogens1553-73661553-7374Public Library of Science San Francisco, USA 1638929710.1371/journal.ppat.001004405-PLPA-RA-0119R2plpa-01-04-07Research ArticleEvolutionGenetics/Gene DiscoveryGenetics/Comparative GenomicsParasitologyPlasmodiumA Plasmodium Whole-Genome Synteny Map: Indels and Synteny Breakpoints as Foci for Species-Specific Genes A
Plasmodium Whole-Genome Synteny Map
Kooij Taco W. A 1¤Carlton Jane M 2Bidwell Shelby L 2Hall Neil 23Ramesar Jai 1Janse Chris J 1Waters Andrew P 1*
1 Department of Parasitology, Malaria Group, Leiden University Medical Centre, Leiden, The Netherlands
2 The Institute for Genomic Research, Rockville, Maryland, United States of America
3 Pathogen Sequencing Unit, The Wellcome Trust Sanger Institute, The Wellcome Trust Genome Campus, Hinxton, Cambridge, United Kingdom
Boothroyd John EditorStanford University, United States of America* To whom correspondence should be addressed. E-mail: [email protected]¤ Current address: Nuffield Department of Clinical Laboratory Sciences, University of Oxford and Blood Research Laboratory, National Blood Service, John Radcliffe Hospital, Headington, Oxford, United Kingdom
12 2005 23 12 2005 1 4 e448 8 2005 16 11 2005 Copyright: © 2005 Kooij 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.Whole-genome comparisons are highly informative regarding genome evolution and can reveal the conservation of genome organization and gene content, gene regulatory elements, and presence of species-specific genes. Initial comparative genome analyses of the human malaria parasite Plasmodium falciparum and rodent malaria parasites (RMPs) revealed a core set of 4,500 Plasmodium orthologs located in the highly syntenic central regions of the chromosomes that sharply defined the boundaries of the variable subtelomeric regions. We used composite RMP contigs, based on partial DNA sequences of three RMPs, to generate a whole-genome synteny map of P. falciparum and the RMPs. The core regions of the 14 chromosomes of P. falciparum and the RMPs are organized in 36 synteny blocks, representing groups of genes that have been stably inherited since these malaria species diverged, but whose relative organization has altered as a result of a predicted minimum of 15 recombination events. P. falciparum-specific genes and gene families are found in the variable subtelomeric regions (575 genes), at synteny breakpoints (42 genes), and as intrasyntenic indels (126 genes). Of the 168 non-subtelomeric P. falciparum genes, including two newly discovered gene families, 68% are predicted to be exported to the surface of the blood stage parasite or infected erythrocyte. Chromosomal rearrangements are implicated in the generation and dispersal of P. falciparum-specific gene families, including one encoding receptor-associated protein kinases. The data show that both synteny breakpoints and intrasyntenic indels can be foci for species-specific genes with a predicted role in host-parasite interactions and suggest that, besides rearrangements in the subtelomeric regions, chromosomal rearrangements may also be involved in the generation of species-specific gene families. A majority of these genes are expressed in blood stages, suggesting that the vertebrate host exerts a greater selective pressure than the mosquito vector, resulting in the acquisition of diversity.
Synopsis
Malaria, caused by the parasite Plasmodium falciparum, is one of the most devastating infectious diseases. Rodent malaria parasites (RMPs), such as P. berghei, P. chabaudi, and P. yoelii, are used as models for P. falciparum. For the use of these models in studies of human disease, insight into both the similarities and differences in the genomics and biology of these parasites is important. The availability of significant but partial genome data of the RMPs enabled the construction of a virtual composite RMP genome and its comparison with the P. falciparum genome, generating a so-called synteny map. Analysis of this map provided the desired comparative insights. A high level of conservation exists between roughly 85% of the genes at the level of content and order, but 168 P. falciparum-specific genes that disrupted the conserved genome segments were identified. The majority of these genes were predicted to play a role in host–parasite interactions. This study indicates that determination of the synteny breakpoints may help to rapidly identify the species-specific gene content of future Plasmodium genomes, providing the malaria research community with a powerful investigative tool. The findings may also be of interest to those studying chromosomal evolution.
Citation:Kooij TWA, Carlton JM, Bidwell SL, Hall N, Ramesar J, et al. (2005) A Plasmodium whole-genome synteny map: Indels and synteny breakpoints as foci for species-specific genes. PLoS Pathog 1(4): e44.
==== Body
Introduction
Comparative genomics enables inferences to be drawn concerning the coding potential of related genomes and the evolutionary forces that have influenced genome organization [1]. The resolving power of whole-genome comparisons to a large extent depends upon the proximity of the phylogenetic relationship between the species. Comparative eukaryotic genome studies of several species from a wide range of lineages and different times of divergence have revealed that the level of both the conservation of organization and the recombination rates are relatively variable. Human and mouse, which diverged ~75 million years (My) ago, have a predicted gene content that is 80% orthologous [2] arranged in 281 synteny blocks (SBs) larger than 1 Mb [3]. Three-way alignment of the human genome with that of mouse and rat confirmed the conservation of ~280 SBs between human and each of the rodent genomes, while the more closely related rat and mouse genomes are ~90% orthologous with a reduced number of 105 shared SBs of larger average size [4]. Subsequent publication of the chicken genome, which diverged from the mammalian genomes ~310 My ago, provided the first nonmammalian amniote genome sequence and allowed a four-way whole-genome comparison [5] revealing 586 smaller, conserved SBs. Here, roughly 50% of the human genes have a chicken ortholog reducing to 35% that have orthologs in both chicken and pufferfish (estimated time of divergence ~450 My). These data show that, in terms of the extent of organization and gene homology, the level of genomic conservation can generally be considered to be relatively proportional to the time of divergence, within these species. However, a more recent comparison of genome sequences from eight mammals demonstrated that the rates of chromosomal rearrangements can vary both between species and in time (about 0.2–2 breaks/My) [6].
In contrast with the relatively slow evolution of mammalian and chicken chromosome structure, gene order and linkage in Diptera species has altered at a much higher rate. Although 50% of the genes are orthologs, little conservation of synteny could be observed in comparisons of the genomes of the fruit fly with two different malaria mosquitoes, which diverged ~250 My ago [7,8]. Even in the more closely related Diptera [8,9], extensive reshuffling and inversion have altered the gene order and organization, although genes were found to be located on the same chromosome arms. Similarly, the genomes of the nematodes Caenorhabditis elegans and C. briggsae, which diverged ~100 My ago, share 60% gene orthology but are arranged as 4,837 microsyntenic clusters [10].
The continuing efforts to sequence a variety of unicellular parasites has resulted in the publication of a comparison of the genome sequences of three human protozoan pathogens, Trypanosoma brucei, T. cruzi, and Leishmania major [11], and two apicomplexan parasites infecting cattle, Theileria annulata and T. parva [12]. The two Theileria species are very closely related, with 81% (T. annulata) and 86% (T. parva) orthologous genes and no interchromosomal rearrangements [12], comparable to the well-conserved genomes of four yeast species that diverged only 5–20 My ago and show relatively few (1–5) translocations [13]. The trypanosomatid species T. brucei and L. major share 68% and 75% gene orthology, respectively, organized in 110 SBs, despite having diverged as long as 200–500 My ago (chromosomal recombination rate of ~0.2–0.5 breaks/My) [11]. In conclusion, these comparative genome studies indicate that effective recombination rates and levels of gene orthology can vary greatly between species but are relatively low in protozoa.
In both pathogenic bacteria and certain unicellular eukaryotes (e.g., the trypanosomatids listed above), including members of the genus Plasmodium that are the etiological agents of malaria, the organization and gene content of the subtelomeric regions of chromosomes are highly variable and typically contain large gene families encoding proteins that may be involved in host-pathogen interactions and antigenic variation [14]. The subtelomeric regions of P. falciparum, for example, harbor a repertoire of unique gene families, including 59 var [15–17], 149 rif, and 28 stevor [18,19]. The var family encodes the erythrocyte membrane protein 1 (PfEMP1), which is a variant antigen expressed at the erythrocyte surface. PfEMP1 is involved in the binding of parasite-infected erythrocytes to receptors of host endothelial cells, erythrocytes, lymphocytes, and blood platelets [14], is subject to antigenic variation, and is thought to play a role in virulence. Other Plasmodium species lack the P. falciparum-specific var, rif, and stevor families, but the subtelomeric regions of their chromosomes also harbor (species-specific) gene families. For example, the human parasite P. vivax; P. knowlesi, which infects primates; and three rodent malaria parasites (RMPs; P. berghei, P. chabaudi, and P. yoelii) share the pir superfamily [20,21]. Proteins encoded by the pir superfamily are also found on the surface of infected erythrocytes and may be implicated in antigenic variation [21]. It is generally believed that the subtelomeric location of gene families confers an enhanced capacity for gene diversification and amplification through mechanisms of ectopic recombination that may be between different chromosomes [22]. Such recombination may be facilitated through the clustering of telomeres at the nuclear periphery [23].
Genome sequence data for Plasmodium species are extensive and include a complete genome sequence for the major human pathogen P. falciparum [24] and 5× coverage of the genome of a RMP, P. yoelii [25]. The P. yoelii contigs, when aligned with the 14 P. falciparum chromosomes, demonstrated extensive similarity over the relatively short length of these contigs. However, similarity was evident only in the core regions of the chromosomes mainly containing conserved genes (4,500) that are present in all characterized Plasmodium species [20] and which are bounded by the variable subtelomeric regions that contain the different gene families. In addition to the genome sequence of P. yoelii, partial genome sequence and analysis have been published for two other RMPs, P. berghei and P. chabaudi, whose core genome sequence and organization are so similar [26–28] that it has proved possible to merge the sequenced DNA contigs of the three RMPs to form composite RMP (cRMP) contigs that cover 90% of the core RMP genomes [20,25]. In this study, the cRMP contigs and 138 sequence tagged site (STS) markers (Table S1) have been used to produce a whole-genome synteny map for the three RMPs that, when compared with the P. falciparum genome, identified 36 SBs describing the core genome. This synteny map shows that species-specific genes—including rapidly evolving P. falciparum gene families—are found not only in the subtelomeric regions but also at synteny breakpoints (SBPs) and as intrasyntenic indels. Our data suggest that chromosomal rearrangements in the core regions might be involved in the generation and subsequent dispersal of one such P. falciparum-specific gene family. These results show that not only recombination in the more frequently recombining subtelomeric regions but also chromosome-internal rearrangements may influence diversity and complexity of the Plasmodium genome, increasing the ability of the parasite to successfully interact with its vertebrate host.
Results
A Whole-Genome Synteny Map of Four Plasmodium Species
A total of 7,392 contigs of the three RMPs, aligned with the P. falciparum genome, were used to generate 910 cRMP contigs (see Materials and Methods, Figure 1, and Tables 1 and S2). The tiling paths of all cRMP contigs are shown for both the individual P. falciparum and RMP chromosomes (Tables S3–S30). The cRMP contigs that were syntenic with the P. falciparum genome totaled 17.2 Mb (75%) of the 22.9 Mb P. falciparum genome, equivalent to 90% of the predicted total region of synteny. After linkage of the aligned cRMP contigs 229 gaps remained. No synteny could be observed in the subtelomeric regions of chromosomes between RMPs and P. falciparum [25], largely due to divergence of subtelomeric repeat sequences and gene families, but also to the poor assembly of these regions in the RMP genome projects [20].
Figure 1 Deduced Organization of the cRMP sera Locus
(A) The combination of three P. berghei, six P. chabaudi, and two P. yoelii contigs (thick black lines) in a region of Pfchr2 containing eight sera copies demonstrates the strength of the “composite genome approach.” Syntenic genes (black, linked by dashed vertical lines; left, PFB0315w and PFB0320c; right, PFB0365w) flank the sera clusters and reveal the presence of five sera genes in the RMPs.
(B) Phylogenetic analysis revealed a close relation between pfsera8, pfsera7, and pfsera6 and their syntenic orthologs in the RMPs (shaded gray, linked by dashed vertical lines in [A]). Other sera copies (pfsera1–5, pbsera1–2, and pysera1–2) clustered in species-specific groups (linked by solid horizontal lines in [A]). Circles represent branch points with bootstrap values of 100% (white), 90%–99% (light gray), and 65%–89% (dark gray).
Table 1 Summary of the Characteristics of the cRMP Contigs, Scaffolds, SBs, and SBPs
When the alignment of the cRMP contigs with the P. falciparum genome was examined, 19 were identified with MUMmer hits to two different P. falciparum chromosomes, indicating that these contigs covered a SBP between the cRMP and the P. falciparum genomes. In addition, three SBPs were determined by chromosome mapping of STS markers and confirmed by PCR analysis, linking the cRMP contigs on either side of the SBP (unpublished data). In total, we found 22 SBPs in the core regions of the P. falciparum genome when compared to the core cRMP genome. Since the cRMP and P. falciparum genomes comprise 14 chromosomes, these 22 SBPs define a total of 36 SBs. Chromosome mapping of 138 P. berghei and P. yoelii STS markers (see Table S1) confirmed the 22 SBPs and the chromosomal location of the 36 SBs in the RMPs. The majority (23 of 28) of P. falciparum subtelomeric regions coincided with putative locations of cRMP subtelomeric regions, while the remaining five P. falciparum subtelomeric linked SBs were linked to SBPs in the cRMP genome. Conversely, five SBs that are linked to SBPs in P. falciparum were linked to subtelomeric regions in the cRMP genome. Figure 2 shows the reciprocal synteny maps of the P. falciparum and cRMP genomes.
Figure 2 A Whole-Genome Synteny Map of P. falciparum and Three RMPs
Synteny map of the core regions of all chromosomes of P. falciparum (left) and the RMPs (right), showing the 36 SBs, 22 SBPs, 14 CAT regions, P. falciparum-specific indels, and translocations in the RMP chromosomes. The 36 SBs, colored according to their chromosomal location in the cRMP genome, are named with a Roman and an Arabic number referring to the corresponding chromosome location in P. falciparum and the cRMP genome, respectively. Letters give the order in which the SBs are connected. Small arrows indicate the inverted orientation of a SB in P. falciparum relative to the cRMP genome. Indels containing P. falciparum-specific intrasyntenic genes are indicated through interruption of the colored SBs. P. falciparum telomeres are shown as white arrow heads (▹). SBs forming the cRMP chromosomes are linked by gray lines. In the cRMP genomes, the 23 coinciding subtelomeric linked ends are shown as white arrowheads (▹) and the five P. falciparum subtelomeric ends that are chromosomal internal in the cRMP chromosomes are indicated by small white arrowheads (▹). The 11 syntenic P. falciparum CAT regions [29] are shown as white circles (○), two inconsistent CAT regions as white circles with a cross (⊗), and three newly recognized CAT regions as white diamonds (⋄). Chromosome-internal var clusters are shown as white arrows (); stars and circles on sticks indicate rrna gene units () and tstk genes (); black stars and circles represent nonsyntenic genes; while syntenic genes (three rrna gene units and one tstk gene) are colored according to their chromosomal location in the RMPs. Bars under the cRMP chromosomes represent the differences in the organization of the SBs of P. yoelii, P. chabaudi, and P. vinckei as a result of translocations. Colors indicate the cRMP chromosome with which recombination has taken place, while color gradients represent the ill-defined regions of the translocation breakpoints.
Centrally located AT-rich (CAT) regions of 2–3 kb (average >97% AT) found on all P. falciparum chromosomes (with the exception of P. falciparum Chromosome 13 [Pfchr13]) have been predicted to be centromeres [29], and functional proof for their centromere function is accruing (S. Iwanaga, CJJ, and APW, unpublished data). While no CAT regions had been sequenced in the RMP genomes, genes immediately up- and downstream of 11 of the P. falciparum CAT regions were syntenic and located at 11 different cRMP chromosomes (Figure 2, Table S31). The predicted centromere of Pfchr7 is located in a SBP and therefore cannot be syntenic, and RMP sequences aligning with the predicted centromere of Pfchr6 did not show an elevated AT content in the cRMP chromosome. Assuming complete synteny of the CAT regions, we suggested new positions for the CAT regions of Pfchr6, 7, and 13 in the regions syntenic with cRMP Chromosome 1 (cRMPchr1), 6, and 13, respectively. Unpublished releases of the latest P. falciparum sequences confirmed these predictions (M. Berriman, personal communication). These results indicate that each of the 14 cRMP chromosomes contained one of the syntenic regions surrounding the P. falciparum CAT regions. Cloning and sequencing of two 1.5-kb regions of cRMPchr5 and 13 that aligned with the CAT regions of Pfchr10 and Pfchr13, respectively, revealed these were also extremely AT-rich (>97%) and consistent with the size and gene paucity of the P. falciparum CAT regions.
Comparison of the organization and location of common orthologous gene families of RMPs and P. falciparum allowed species-specific features of these families to be defined. For example, P. falciparum possesses a cluster of eight genes encoding putative serine proteases known as sera [30,31]. The P. berghei and P. yoelii databases both contain five sera, whose organization in the individual RMP genomes was unresolved, yet could be reconstructed using the cRMP contigs, demonstrating one utility of the cRMP contig construction (see Figure 1A). Combining the synteny analysis with standard phylogenetic analysis (see Figure 1B) indicated that all RMP sera cluster at a single locus on cRMPchr3, which aligns with the P. falciparum sera cluster on Pfchr2. Within these clusters, direct orthologs for three sera (RMP sera3–5 and P. falciparum sera6–8) were immediately adjacent and thus syntenic. The remaining RMP sera1–2 and the pfsera1–5 are also immediately adjacent to one another and each positioned similarly within the sera cluster in both genomes but form different phylogenetic clades and can be considered species-specific.
Inferring the Pathway of Synteny Rearrangements Between the cRMP and P. falciparum Genomes
The organization of the three RMP genomes is highly conserved, and only one or two chromosomal rearrangements were noted when the genomes of the individual RMP species were compared with the cRMP genome (Figure 2). The organization of the P. berghei genome is identical to that of the cRMP genome, suggesting it is also most similar to the genome structure of the most recent common ancestor of the RMPs.
The P. falciparum genome organization could be generated from the cRMP genome in a minimum of 15 recombination events when the following assumptions were made: (i) that the resulting genome always consists of 14 chromosomes; (ii) that all chromosomes always contain only one of the SBs containing a CAT region; and (iii) that a recombination event generating a subtelomeric from a chromosome-internal region (or vice versa, collectively termed telomere conversions) has happened only once. These 15 recombination events included eight single crossover events, five telomere conversions, one inversion of an entire SB, and one insertion involving an intersyntenic var cluster (Figure 3). This most parsimonious pattern of gross chromosomal rearrangements was supported by analysis using the GRIMM (genome rearrangements in man and mouse) algorithm [3] that identified one inversion and 15 translocations, counting the var cluster insertion as two single translocation events (unpublished data). The relatively low number of 15 rearrangements events suggests that gross chromosomal rearrangements resulting in the loss of or change in synteny is infrequent in Plasmodium. However, the same recombination events could be associated with the formation and dispersal of (members of) species-specific gene families (see below).
Figure 3 Schematic Representation of the 15 Recombination Events
Schematic representation of the 15 recombination events that would permit the 36 SBs to be rearranged to generate the P. falciparum genome from the cRMP genome. See Figure 2 for the numbering of the SBs and the symbols used in this figure. Gray lines between SBs represent links as present in the cRMP genome; gray dashed lines indicate intermediate links, and black arrows show links corresponding to the P. falciparum genome. Five subtelomeric regions of the cRMP genome must become chromosome-internal in the P. falciparum genome (A), thereby generating five subtelomeric regions in P. falciparum that are linked to SBPs in the cRMP genome. SB “XIVc:13b” is inverted (B), and SBs “VIIc:14b” and “VIIIb:14c” are inserted between SBs “VIIb:12c” and “VIIIc:12d,” a process likely to involve chromosome-internal clusters of var and rif genes possibly mediated by vicar genes (C). Eight single crossover events generate the remaining links between the remaining SBs (D).
P. falciparum-Specific Genes Are Found Both at SBPs and in Intrasyntenic Indels
The average size of species-specific DNA regions located between SBs (intersyntenic regions) is significantly smaller in the cRMP genome (~2.5 kb, range 0.4–15 kb) than in the P. falciparum genome (~16 kb, range 0.7–106 kb). Only four of the 19 intersyntenic regions in the cRMP genome for which sequence data are available contain a species-specific open reading frame, but only the nonsyntenic c-rrna gene unit on cRMPchr5 is known to be expressed (Tables 2 and S32). In contrast, eight of the 22 intersyntenic regions in P. falciparum contain clusters of one to 13 genes without RMP orthologs (Tables 2 and S33). These 42 intersyntenic genes include 14 var and six rif genes, as well as five other genes, which all encode proteins containing the Plasmodium export element/vacuolar transport signal motif (PEXEL/VTS) [32,33]—e.g., glycophorin-binding protein 130 precursor: GBP130 [34] and two receptor-associated protein kinases: PfTSTK7a, and PfTSTK10a (see also below). The PEXEL/VTS motif is one element that is associated with transport of the proteins to the surface of the infected erythrocyte. A further 12 genes encode proteins with a transmembrane domain at the N-terminal end (e.g., MAL7P1.58 of the pfmc-2tm family, which encodes proteins localized to the Maurer's clefts [35]), seven of which also have a signal peptide (e.g., PF10_0164 of the etramp family [36] and five var internal cluster associated repeat [vicar] genes; see also below). Figure 4A provides a detailed example of the SBP on Pfchr10 and alignment of the flanking syntenic regions with P. yoelii contigs. In conclusion, it seems that the majority of the intersyntenic, P. falciparum-specific, SBP-associated genes encode predicted exported proteins destined for the membrane surface of the cell-free parasite or the infected erythrocyte.
Table 2 Summary of Inter- and Intrasyntenic Gene Content of P. falciparum and Comparison to Intersyntenic Gene Content of the RMPs
Figure 4 Inter- and Intrasyntenic Indels Contain Clusters of P. falciparum-Specific Genes
(A) A detailed illustration of the SBP between the SBs “Xa:12a” and “Xb:5a” (Pfchr10) flanked by P. yoelii contigs MALPY00409 (gray) and MALPY00055 (purple). The last gene on MALPY00409 is located on Pfchr3 (“IIIc:12b”) and defines the SBP; MALPY00055 is the last syntenic contig flanking a subtelomeric region that contains a cRMP-specific gene encoding a hypothetical protein (red) and a nonsyntenic etramp (white with red outline). The P. falciparum intersyntenic region contains three annotated genes (white with red outline): gbp130, pftstk10a, and etramp; and three genes encoding hypothetical proteins (red). Interestingly, four of six genes encode putative secreted proteins with N-terminal transmembrane domains destined for the parasite surface or infected host cell membrane (asterisks; see Table S33).
(B) A detailed illustration of a ~22-kb indel within SB “Xb:5a” that contains P. falciparum-specific genes directly upstream of a region containing genes that are highly diverged in the RMPs. Only four of 12 genes annotated on MALPY00271 have a clear ortholog (purple) and the last gene (PY01020), which encodes a hypothetical protein, shows low similarity at the N-terminal end with PF10_0348 (horizontal purple lines). Comparison of the P. yoelii and P. falciparum annotations revealed the presence in both species of six genes with the same orientation and comparable size, including four genes that encode hypothetical proteins (black with purple outline) and two annotated genes (white with purple outline): the putative P. yoelii lsa1, and the putative msp paralog P. yoelii H103. MALPY01161 and MALPY00271 are physically linked as determined with Grouper software and are therefore between 500 and 2,000 bp apart, leaving no space for the remaining genes in the ~22 kb P. falciparum indel that include S-antigen, glurp, msp3, msp6, and H101 (all white with red outline) and one gene encoding a hypothetical protein (red). In the entire regions, 12 of 15 genes encode putative secreted proteins destined for the parasite surface or infected host cell membrane (asterisks; see Table S34).
In addition to the species-specific genes located at SBPs, P. falciparum-specific genes were also found clustered in small intrasyntenic regions that interrupt the SBs (i.e. indels, Tables 2 and S34). These 82 indels, including four var clusters, range in size from one to nine genes but are generally less gene-rich than the intersyntenic regions (1.5 genes/indel compared to 5.3 genes/SBP). Whereas only two of eight SBPs contain a single P. falciparum-specific gene, 65 of 82 of the intrasyntenic indels contain only one gene. The 126 intrasyntenic, P. falciparum-specific genes include nine var and four rif genes as well as an additional six genes with the PEXEL/VTS motif [32,33] including pftstk13 (MAL13P1.109, see also Discussion). Another 59 of these genes encode proteins with an N-terminal transmembrane domain, 40 of which also contain a signal peptide, giving a total of 78 genes encoding potential secreted or surface proteins. For example, a multigenic indel on Pfchr10 (Figure 4B) contains a cluster of six P. falciparum-specific genes that are all expressed in merozoites [37–39] and encode three known merozoite surface protein paralogs (MSP3, MSP6, and H101), glutamate-rich protein (GLURP), S-antigen, and a hypothetical protein containing a signal peptide sequence. The presence of a fourth msp paralog H103 in the neighboring syntenic region suggests that the gene content of this indel might have arisen in part through local gene duplication [40].
Evolution of Gene Families Associated with Recombination Events at SBPs
In order to analyze whether recombination events in the core regions that resulted in the loss of synteny are associated with the dispersal and formation of species-specific gene families, all intersyntenic genes of P. falciparum and the RMPs were analyzed for the presence and location of orthologous genes in their respective genomes. In addition to members of the var, rif, and rrna families, one intrasyntenic (pftstk13) and two intersyntenic (pftstk7a and pftstk10a) P. falciparum genes were identified that belong to a gene family encoding 21 transforming growth factor β receptor-like serine/threonine protein kinases (PfTSTK) [41–43]. In addition to these three genes, 17 members are located in the subtelomeric regions of 10 different chromosomes (Table S35), and one member is located adjacent to the Pfchr8 CAT region (M. Berriman, personal communication). In the RMP genome there is a single member of this family on cRMPchr12 syntenic to the copy near the Pfchr8 CAT region. Phylogenetic analysis groups these syntenic kinases in the same clade as the unique members of all other characterized Plasmodium species, with exception of the proteins encoded by the multiple tstk genes found in Plasmodium reichenowi, a very close relative of P. falciparum infecting chimpanzees [44]. These findings suggest that the syntenic pftstk on Pfchr8 could be the progenitor gene of this P. falciparum-specific gene family (Figure 5A).
Figure 5 Origin and Putative Mechanism of Expansion of the tstk Family in P. falciparum
(A) Analysis of P. falciparum-specific genes at the SBPs revealed a gene family encoding receptor-associated protein kinases (TSTK). Maximum likelihood distances were calculated for the C-terminal 400 amino acids of all TSTKs, including those found for other Plasmodium species, Toxoplasma gondii, Cryptosporidium parvum, and C. hominis. The tree was rooted using the clade with the three non-Plasmodium sequences as the outgroup (shaded dark gray). The syntenic progenitor genes clearly form one clade (shaded light gray), while the clustering of the other 20 mainly subtelomeric pftstk is more ambiguous (the three non-subtelomeric copies are shown in bold and include pftstk7a, which appears most closely related to the clade of progenitor genes). Circles represent branch points with bootstrap values of 100% (white), 90%–99% (light gray) and 65%–89% (dark gray).
(B) See Figure 2 for the numbering of the SBs and the symbols used in this figure. Based on the 15 recombination events described in Figure 3 and the phylogenetic analysis of the tstk family, we suggest the origin and putative evolution of the pftstk family as shown here. Phylogenetic analysis suggests that the intersyntenic pftstk7a is most closely related to the progenitor founder gene, pftstk0. Interestingly, this gene is the first nonsyntenic gene upstream of SB “VIIe:2b.” This SB is linked in the cRMP genome to SB “I:2a” that in P. falciparum is also flanked by a member of the tstk family, the subtelomeric pftstk1. Based on these observations we suggest that the founder gene pftstk0 was duplicated after the split of P. falciparum from the other Plasmodium species but before SBs “VIIe:2b” and “I:2a” were separated (1). This gene was then directly involved in the breakage of this link, creating Pfchr1 (“I:2a”) and destroying the telomere of “VIId:6d” by addition of “VIIe:2b” (2). During this recombination process, the gene was duplicated and is now present not only as two chromosome-internal copies on “VIIIc:12d” (pftstk0) and between “VIId:6d” and “VIIe:2b” (pftstk7a) but also as a first telomeric copy on the newly formed telomere of Pfchr1 (pftstk1). From here the gene could expand to the other subtelomeric regions (3). Local gene duplications resulted in the generation of seven copies on Pfchr9 and two copies on Pfchr4. After a copy of pftstk ended up at the left-hand cRMP subtelomeric end of SB “Xb:5a,” the telomere conversion linked SB “Xa:12a” to SB “Xb:5a,” which turned this telomeric copy into an intersyntenic gene (pftstk10a). The last non-subtelomeric copy, pftstk13, most likely resulted from a different process of mobility of P. falciparum-specific elements creating the intrasyntenic genes.
Two different recombination pathways that would generate the pftstk family are consistent with the data. (i) A copy of the syntenic, orthologous progenitor pftstk on Pfchr8 relocated to a subtelomeric region, where it underwent extensive gene duplication and redistribution. The centrally located pftstk genes could then have originated from telomere changes. (ii) Combining the information on the location and phylogeny of the pftstk family with the predicted 15 synteny rearrangements suggests that both chromosome-internal rearrangements resulting in the loss of synteny and subtelomeric recombination are associated with the evolution and distribution of this family (Figure 5B). P. falciparum-specific duplication/translocation of the ancestral tstk to an ancestral “cRMPchr2” followed by chromosome breakage and recombination may have led to the translocation of a tstk copy to a subtelomeric position (Pfchr1). Additional subtelomeric copies may be translocated to the nine additional subtelomeric locations by ectopic recombination events between different chromosomes similar to the events suggested to play a crucial role in the generation of var gene diversity [23]. The intersyntenic copy on Pfchr10 might be the result of a subsequent recombination event leading to the internalization of this gene. The intrasyntenic pftstk13 may have originated independently of the mechanism that generated this gene family in a similar (if obscure) mechanism to other intrasyntenic genes with apparent subtelomeric origin, including the var and rif genes. All the predicted duplication and translocation events required to distribute the pftstk family could be linked to the proposed rearrangement pathway that converts the RMP genome organization to that of P. falciparum. Since there are alternative pathways for the order of the suggested SBP recombination events (also indicated by the GRIMM algorithm analysis; Table S36), further elucidation of the pathway of recombination from the genome organization of the most recent common ancestor of Plasmodium awaits the availability of the genome of a third species [45].
Identification of a New Putative Gene Family Associated with Chromosome-Internal var Clusters
Since repetitive sequences might be associated with recombination events between SBs, the intergenic regions flanking SBPs were examined using the MEME algorithm. This analysis resulted in the identification of a highly conserved P. falciparum-specific gene family consisting of seven putative genes and eight pseudogenes termed var internal cluster associated repeat (vicar) genes. These genes were found to be associated with five of seven chromosome-internal var clusters. Of these seven genes, five have a signal peptide and five genes have one or two transmembrane domains; only one of these genes is identified in the current annotation (MAL7P1.39) and is supported by transcriptome data [38]. The sequences correspond to the previously described GC-rich elements that were suggested to serve as regulatory elements for var-related genetic processes [29]. No other repetitive sequences were identified that, in the light of current knowledge, could be associated with chromosomal recombination events.
Discussion
The generation of composite contigs from three closely related Plasmodium species infecting rodents greatly facilitated the construction of a synteny map between the RMPs and P. falciparum and significantly reduced the need for experimental data from PCR and STS mapping studies. Current contig assembly algorithms rely upon a minimum of 95% sequence identity between sequence reads [46], a criterion not met by the RMP sequences. The high degree of synteny and similarity of gene content of the core Plasmodium genome enabled the compilation of cRMP contigs using sequences of the three RMPs with a lower sequence identity by aligning them to the assembled P. falciparum sequence. With only 229 gaps remaining and the location of 138 STS markers identified, the synteny map is a comprehensive tool for identifying the location of most genes. Individually, cRMP contigs are not sufficient to build an entire composite genome, since coverage and linkage of the cRMP scaffolds are incomplete. An unknown proportion of small rearrangements such as single gene insertions, inversions, or deletions will have been missed. Thus the need for continued sequencing to completion of at least one RMP genome remains. Approximately 4,500 (85%) of the 5,300 predicted P. falciparum genes have an ortholog in at least one of the RMPs, and these likely represent the core set of Plasmodium genes [20]. A similar level of orthology is seen in the genome organization, since the 36 SBs cover 84% of both genomes.
The synteny maps of P. falciparum and cRMP demonstrated that only a minimum of 15 recombination events are needed to generate the P. falciparum genome from the 36 SBs of the RMPs, compared with 245 events needed to convert the human genome organization to that of the mouse [3]. This relatively low number of Plasmodium genome rearrangements suggests either that divergence of P. falciparum and the RMPs might be relatively recent or that chromosomal rearrangements in Plasmodium are infrequent, either as a result of unknown (intrinsic) features of the DNA or due to some higher order organization of the genome [26]. Because the evolutionary relationships and the time of divergence between P. falciparum and other Plasmodium species is unclear [44,47–52], it is not yet possible to draw conclusions on the rate of chromosomal rearrangements in Plasmodium. A rough estimate consistent with published data would be that P. falciparum diverged and developed separately between 50 and 200 My ago. Thus the effective chromosomal recombination rate would be between 0.08 and 0.3 breaks/My. In comparison, the recombination rate in yeast species appears to be ~0.2 breaks/My [13]. Both are at the lower end of the range of rates observed for different mammalian species [6]. The genomes of different trypanosomatid species were also suggested to have a low recombination rate [11].
In many species, centromeres have been associated with chromosomal rearrangements and have proven to be positionally dynamic, with transposable elements often found to function in centromere relocation [1]. Plasmodium centromeres have not been functionally characterized but based on previous predictions, preliminary functional evidence (S. Iwanaga, CJJ, and APW), and the distribution of the CAT regions as demonstrated by the Plasmodium synteny map, it is tempting to suggest that the predicted centromeres of Plasmodium are positionally static. One of the assumptions upon which the initial intuitive derivation of the minimum 15 recombination events was based was that each chromosome at any time always contains one CAT region and one only, in keeping with their still-hypothetical function as centromeres. The GRIMM analysis did not include such an assumption, yet it predicted the same number of rearrangements, while maintaining a single SB containing a CAT region in each newly formed chromosome, emphasizing their predicted lack of involvement in the recombination events identified in this study. Furthermore, these recombination events are also unlikely to involve transposable elements, since these were not found in a cross-species comparison of the sequences in the vicinity of SBPs, consistent with previous studies [24].
In contrast to the low number of chromosomal rearrangements in the Plasmodium genomes, a relatively large proportion (15%) of the P. falciparum genes have no readily identifiable ortholog in any of the RMPs. These genes (including the well known var, rif, and stevor families) are mainly located in the subtelomeric regions, which appear to have a higher rate of gene evolution in many organisms, including Plasmodium [1,22]. However, this study shows that a significant proportion of P. falciparum-specific genes and members of gene families are not restricted to the subtelomeric region of the chromosomes but can be found as intrasyntenic indels and at SBPs. The majority (115 genes [68%]) of these 168 genes encode predicted or known surface or secreted proteins that are predominantly expressed in asexual blood stage parasites (both infected erythrocytes and merozoites) and thus are involved in parasite interactions with the human host and possibly associated with immune selection/evasion. Interestingly, several of the larger clusters of genes, such as the indel containing msp3 and msp6, appear to be coordinately expressed and may even be transcribed in an operon-like manner [53], despite earlier analyses that did not find evidence for the existence of such clusters [37]. Perhaps surprisingly, indels containing RMP-specific genes were not readily found, and although this may be in part due to the incomplete RMP genome sequence data that are currently available, the depth of coverage of the cRMP genome suggests that RMP indels are not as frequent as in P. falciparum. However, indels are not absent from the RMP genomes, and evidence is accumulating for RMP indels that contain members of the pir superfamily normally found in the subtelomeric regions reminiscent of the organization of the var family in the P. falciparum genome (see Tables S3–S30) [20,21].
To test whether SBPs are significantly more associated with chromosome-internal P. falciparum-specific genes than what might be expected based on a random distribution of the SBPs, we used computer simulations to generate randomly distributed SBPs in the genome and compared these with the inter- and intrasyntenic gene content. Using a conservative and a more relaxed approach (see Materials and Methods), we showed that based on a random breakage model, between 1.9 and 3.0 of the 22 SBPs on average could be expected to be associated with P. falciparum-specific gene clusters. This is significantly different (p < 0.001) from the observed association of eight (36%) of the 22 SBPs with P. falciparum-specific genes. This result indicates a nonrandom distribution of P. falciparum-specific genes associating with a higher frequency to SBPs and, therefore, with chromosomal rearrangements that have led to loss of synteny. Interestingly, from comparisons of the human and mouse genomes, evidence has emerged for a similar nonrandom distribution of repeat sequences in the genome and their association with SBPs [54,55].
The presence of members of species-specific gene families at the SBPs suggests that recombination events resulting in loss of synteny helped shape species-specific gene content. SBPs and the intrasyntenic indels might therefore distinguish islands where variations in gene content occur (and then evolve) between the different Plasmodium species. The location and phylogeny of the pftstk family and the chromosomal rearrangements between SBs were consistent with different possible recombination pathways and mechanisms. Interestingly, the processes of gene duplication and translocation described for the tstk family could also be associated with the generation of two other gene families in P. falciparum encoding acyl-CoA binding proteins (ACP; four P. falciparum genes and one cRMP gene) and acyl-CoA synthetases (ACS; 11 P. falciparum genes and three cRMP genes). Both families have one syntenic copy in P. falciparum and the RMPs that are located in the P. falciparum genome next to an indel. The syntenic acp is located next to an indel on Pfchr8, and the syntenic acs next to an indel on Pfchr2 (PFB0685c). This latter gene appears to have undergone local gene duplication, followed by relocalization and expansion to seven subtelomeric copies in P. falciparum (unpublished data). In conclusion, our data show that both SBPs and intrasyntenic indels can be foci for species-specific genes with a predicted role in host-parasite interactions and indicate that not only rearrangements in the subtelomeric regions but also chromosomal rearrangements are involved in the generation of species-specific gene families. The majority are expressed in blood stages (complete list in Table S34), suggesting that the vertebrate host exerts a greater selective pressure than the mosquito vector, resulting in the acquisition of diversity.
It is already evident that a single recombinational mechanism underlying the origin of the inter- and intrasyntenic gene content or the generation of gene families in P. falciparum cannot be postulated. The 42 SBP-associated genes of P. falciparum can be classified into three groups: (i) two single genes that are associated with single crossover events; (ii) three clusters of genes (total 12 genes) that might have their origin in subtelomeric regions that became chromosome-internal after a telomere change (these include the SBPs containing pftstk genes); and (iii) three var clusters, two associated with the insertion of SBs “VIIc:14b” and “VIIb:14c” and one associated with a single crossover event (total 28 genes; see Table S33). Thus it is clear that different recombination mechanisms were involved in shaping the P. falciparum genome. Evidence from both the 15 SBP-associated recombination events and previous var gene classifications [56] cannot be reconciled with an origin of central var clusters associated with telomere recombination changes and subsequent internalization of subtelomeric var genes. Both SBP and intrasyntenic var clusters are associated with the vicar genes identified in this study and previously described as the GC-rich elements [29]. The position of vicar elements is consistent with an as yet unproven role in recombination.
The pairwise whole-genome comparison presented here, while indicating that 15 chromosomal rearrangements can create the P. falciparum genome organization from that of the RMP, does not resolve the organization of the most recent common ancestor, which requires more complete Plasmodium genomes. Genome-wide comparison of the location and distribution of SBPs between different Plasmodium species should provide a reliable dataset enabling construction of a definitive phylogeny of the genus and resolving issues of precise clade topology [45]. In addition, whole-genome comparisons and the identification of SBPs might prove to be an effective means of identifying species-specific genes and members of gene families that are involved in host-parasite interactions and immune evasion, including antigenic variation.
Materials and Methods
Creation of a cRMP genome.
7,215 contigs of three RMP genomes, P. yoelii yoelii (17XNL line) [25], P. berghei (ANKA strain), and P. chabaudi chabaudi (AS strain) [20] were previously aligned with the P. falciparum genome using MUMmer to identify annotation-independent protein similarities [57]. We manually aligned an additional 177 contigs using linkage data from the P. yoelii genome publication and by performing BLASTN analyses with ~500-bp sized sequences from the ends of the RMP contigs, thus closing gaps in the synteny map and “walking” toward the telomeric ends. Linking of these 7,392 contigs through identification of overlapping contigs resulted in the generation of 910 cRMP contigs (see Figure 1A for an example of the procedure to generate cRMP contigs). The high level of nucleotide identity between the genomes of the three RMPs (P. yoelii versus P. berghei, 91.3%; P. yoelii versus P. chabaudi, 88.1%; and P. berghei versus P. chabaudi, 87.1%) facilitated this process. The cRMP contigs that showed MUMmer hits to two different P. falciparum chromosomes revealed SBPs. Linkage between adjacent P. y. yoelii contigs had previously been established using Grouper [58], through the alignment of overlapping P. yoelii expressed sequence tags and by PCR amplification [25]. Combining these data with the 910 cRMP contigs resulted in the generation of 243 scaffolds of linked cRMP contigs. STS markers were used to determine chromosomal locations of the linked cRMP contigs. These markers included 79 previously described and 59 new markers strategically chosen based on the position of the SBPs (see Table S1). All markers were hybridized to chromosomes of P. yoelii, P. berghei, P. chabaudi, and P. vinckei that had been separated by pulsed field gel electrophoresis [27].
Analysis of the synteny map of the cRMP and P. falciparum genomes.
Intergenic sequences flanking the SBs at all 22 P. falciparum SBPs as well as the five subtelomere linked ends that are chromosome-internal in the RMPs (92 kb in total) were analyzed for repetitive motifs using MEME [59]. The intergenic sequences of the 20 RMP SBPs for which sequence was available were also analyzed. Nonsyntenic genes were compared with the genome data of the different Plasmodium species by TBLASTN analysis, and the expression profiles and putative functions of these genes were investigated using data available from PlasmoDB [30,31,38,39]. The predicted protein sequences of the tstk family members were analyzed for functional domains by SMART [60].
GRIMM [3] was used to confirm the suggested minimum 15 recombination events. To test the significance of the association between SBPs and P. falciparum-specific gene content, we used computer simulations to reassign the 22 chromosome-internal SBPs to random positions in the core genome of P. falciparum, thus excluding the subtelomeric regions. We used two different approaches: The first approach utilized the sizes of the entire SBP regions, including the species-specific gene content, while the second approach utilized fixed SBP sizes (5 kb, slightly larger than the largest noncoding intergenic, intersyntenic regions). For both approaches, we counted the number of associations of the virtual SBPs of 1,000 random distributions with the locations of all inter- and intrasyntenic genes.
Phylogenetic analyses of members of the TSTK and SERA families were performed using manually corrected ClustalW alignments [61]. Protein parsimonies, pairwise distances and maximum likelihood distances were calculated using different regions of alignment with algorithms and matrices from the phylogeny inference package (PHYLIP) [62] and gave comparable results. For the final tree construction, 100 bootstrap trees were generated (each with 10× jumbling) of a manually corrected alignment of roughly 400 amino acids of the C-terminal ends of all TSTKs containing the serine/threonine protein kinase domain using SEQBOOT [63]. Maximum likelihood distances [64] were calculated using default parameter settings and 10× jumbling. The 100 bootstrap trees thus constructed were combined using CONSENSE [65]. The tree was rooted using the clade of non-Plasmodium TSTKs as the outgroup with RETREE, and the final tree was drawn using DRAWTREE, both also available from PHYLIP [62].
Supporting Information
Table S1 STS Marker List
(111 KB XLS)
Click here for additional data file.
Table S2 Details of the cRMPchrs
(64 KB TXT)
Click here for additional data file.
Table S3 Pfchr1
(138 KB XLS)
Click here for additional data file.
Table S4 Pfchr2
(214 KB XLS)
Click here for additional data file.
Table S5 Pfchr3
(257 KB XLS)
Click here for additional data file.
Table S6 Pfchr4
(238 KB XLS)
Click here for additional data file.
Table S7 Pfchr5
(369 KB XLS)
Click here for additional data file.
Table S8 Pfchr6
(382 KB XLS)
Click here for additional data file.
Table S9 Pfchr7
(303 KB XLS)
Click here for additional data file.
Table S10 Pfchr8
(346 KB XLS)
Click here for additional data file.
Table S11 Pfchr9
(357 KB XLS)
Click here for additional data file.
Table S12 Pfchr10
(381 KB XLS)
Click here for additional data file.
Table S13 Pfchr11
(540 KB XLS)
Click here for additional data file.
Table S14 Pfchr12
(585 KB XLS)
Click here for additional data file.
Table S15 Pfchr13
(714 KB XLS)
Click here for additional data file.
Table S16 Pfchr14
(809 KB XLS)
Click here for additional data file.
Table S17 cRMPchr1
(215 KB DOC)
Click here for additional data file.
Table S18 cRMPchr2
(216 KB DOC)
Click here for additional data file.
Table S19 cRMPchr3
(235 KB DOC)
Click here for additional data file.
Table S20 cRMPchr4
(270 KB DOC)
Click here for additional data file.
Table S21 cRMPchr5
(296 KB DOC)
Click here for additional data file.
Table S22 cRMPchr6
(332 KB DOC)
Click here for additional data file.
Table S23 cRMPchr7
(285 KB DOC)
Click here for additional data file.
Table S24 cRMPchr8
(429 KB DOC)
Click here for additional data file.
Table S25 cRMPchr9
(598 KB DOC)
Click here for additional data file.
Table S26 cRMPchr10
(513 KB DOC)
Click here for additional data file.
Table S27 cRMPchr11
(592 KB DOC)
Click here for additional data file.
Table S28 cRMPchr12
(627 KB DOC)
Click here for additional data file.
Table S29 cRMPchr13
(842 KB DOC)
Click here for additional data file.
Table S30 cRMPchr14
(845 KB DOC)
Click here for additional data file.
Table S31 Centromere Predictions
(68 KB DOC)
Click here for additional data file.
Table S32 RMP Intersyntenic Genes
(38 KB DOC)
Click here for additional data file.
Table S33 Pf Intersyntenic Genes
(175 KB DOC)
Click here for additional data file.
Table S34 Pf Intrasyntenic Genes
(477 KB DOC)
Click here for additional data file.
Table S35 Pf Subtelomeric Genes
(687 KB TXT)
Click here for additional data file.
Table S36 GRIMM Analysis
(171 KB DOC)
Click here for additional data file.
Accession Numbers
The GenBank (http://www.ncbi.nlm.nih.gov) accession numbers for the sequences of two putative P. yoelii centromeres (Chromosomes 5 and 13) are DQ054838 and DQ054839, respectively.
All datasets will become available through the official Web site of the Plasmodium genome project, PlasmoDB (http://plasmodb.org) [30,31]. The PlasmoDB accession numbers for the P. falciparum cluster of eight genes encoding putative serine proteases known as sera are PFB0325c–PFB0360c. The PlasmoDB accession numbers for other genes and gene products discussed in this paper are, for P. falciparum: etramp (PF10_0164), gbp130 (PF10_0159), glurp (PF10_0344), H101 (PF10_0347), H103 (PF10_0352), hypothetical protein (PF10_0342), lsa1 (PF10_0356), msp3 (PF10_0345), msp6 (PF10_0346), pftstk1 (PFA0130c), pftstk7a (MAL7P1.144), pftstk10a (PF10_0160), pftstk13 (MAL13P1.109), and S-antigen (PF10_0343); for P. berghei: H103 (PB105993.00.0), lsa1 (PB101910.00.0+PB105996.00.0), and five sera (PB000649.01.0, PB000352.01.0, PB000107.03.0, PB107093.00.0, PB000108.03.0); for P. yoelii: etramp (PY00205), H103 (PY01016), lsa1 (PY01014), and five sera (PY02063, PY02062+PY00294, PY00293, PY00292, PY00291).
P. berghei and P. yoelii gene models referred to in the text are available from GeneDB (http://www.genedb.org), and GeneIndices (http://www.tigr.org/tdb/tgi/protist.shtml).
We would like to thank Matthew Berriman and The Wellcome Trust Sanger Institute for kindly providing prepublication P. falciparum sequences and Ross Coppel for constructive criticism. TWAK was supported by a Leiden University PhD fellowship. We would like to thank the anonymous reviewers for their constructive criticism that resulted in a significant reshaping of this manuscript.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. TWAK, JMC, NH, CJJ, and APW conceived and designed the experiments. TWAK, JMC, SLB, JR, and CJJ performed the experiments. TWAK, CJJ, and APW analyzed the data. TWAK, JMC, CJJ, and APW wrote the paper.
Abbreviations
CATcentrally located AT-rich
cRMPcomposite rodent malaria parasite
cRMPchrcomposite rodent malaria parasite chromosome
Mymillion years
PEXEL/VTS
Plasmodium export element/vacuolar transport signal
Pfchr
P. falciparum chromosome
RMProdent malaria parasite
SBsynteny block
SBPsynteny breakpoint
STSsequence tagged site
TSTKtransforming growth factor β receptor-like serine/threonine protein kinases
==== Refs
References
Eichler EE Sankoff D 2003 Structural dynamics of eukaryotic chromosome evolution Science 301 793 797 12907789
Waterston RH Lindblad-Toh K Birney E Rogers J Abril JF 2002 Initial sequencing and comparative analysis of the mouse genome Nature 420 520 562 12466850
Pevzner P Tesler G 2003 Genome rearrangements in mammalian evolution: Lessons from human and mouse genomes Genome Res 13 37 45 12529304
Gibbs RA Weinstock GM Metzker ML Muzny DM Sodergren EJ 2004 Genome sequence of the Brown Norway rat yields insights into mammalian evolution Nature 428 493 521 15057822
Hillier LW Miller W Birney E Warren W Hardison RC 2004 Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution Nature 432 695 716 15592404
Murphy WJ Larkin DM Everts-van der Wind A Bourque G Tesler G 2005 Dynamics of mammalian chromosome evolution inferred from multispecies comparative maps Science 309 613 617 16040707
Zdobnov EM von Mering C Letunic I Torrents D Suyama M 2002 Comparative genome and proteome analysis of Anopheles gambiae and Drosophila melanogaster
Science 298 149 159 12364792
Severson DW DeBruyn B Lovin DD Brown SE Knudson DL 2004 Comparative genome analysis of the yellow fever mosquito Aedes aegypti with Drosophila melanogaster and the malaria vector mosquito Anopheles gambiae
J Hered 95 103 113 15073225
Sharakhov IV Serazin AC Grushko OG Dana A Lobo N 2002 Inversions and gene order shuffling in Anopheles gambiae and A. funestus
Science 298 182 185 12364797
Stein LD Bao Z Blasiar D Blumenthal T Brent MR 2003 The genome sequence of Caenorhabditis briggsae: A platform for comparative genomics PLoS Biol 1 E45 14624247
El Sayed NM Myler PJ Blandin G Berriman M Crabtree J 2005 Comparative genomics of trypanosomatid parasitic protozoa Science 309 404 409 16020724
Pain A Renauld H Berriman M Murphy L Yeats CA 2005 Genome of the host-cell transforming parasite Theileria annulata compared with T. parva
Science 309 131 133 15994557
Kellis M Patterson N Endrizzi M Birren B Lander ES 2003 Sequencing and comparison of yeast species to identify genes and regulatory elements Nature 423 241 254 12748633
Kyes S Horrocks P Newbold C 2001 Antigenic variation at the infected red cell surface in malaria Annu Rev Microbiol 55 673 707 11544371
Su XZ Heatwole VM Wertheimer SP Guinet F Herrfeldt JA 1995 The large diverse gene family var encodes proteins involved in cytoadherence and antigenic variation of Plasmodium falciparum -infected erythrocytes Cell 82 89 100 7606788
Smith JD Chitnis CE Craig AG Roberts DJ Hudson-Taylor DE 1995 Switches in expression of Plasmodium falciparum var genes correlate with changes in antigenic and cytoadherent phenotypes of infected erythrocytes Cell 82 101 110 7606775
Baruch DI Pasloske BL Singh HB Bi X Ma XC 1995 Cloning the P. falciparum gene encoding PfEMP1, a malarial variant antigen and adherence receptor on the surface of parasitized human erythrocytes Cell 82 77 87 7541722
Cheng Q Cloonan N Fischer K Thompson J Waine G 1998
stevor and rif are Plasmodium falciparum multicopy gene families which potentially encode variant antigens Mol Biochem Parasitol 97 161 176 9879895
Kyes SA Rowe JA Kriek N Newbold CI 1999 Rifins: a second family of clonally variant proteins expressed on the surface of red cells infected with Plasmodium falciparum
Proc Natl Acad Sci U S A 96 9333 9338 10430943
Hall N Karras M Raine JD Carlton JM Kooij TW 2005 A comprehensive survey of the Plasmodium life cycle by genomic, transcriptomic, and proteomic analyses Science 307 82 86 15637271
Janssen CS Phillips RS Turner CM Barrett MP 2004
Plasmodium interspersed repeats: The major multigene superfamily of malaria parasites Nucleic Acids Res 32 5712 5720 15507685
Barry JD Ginger ML Burton P McCulloch R 2003 Why are parasite contingency genes often associated with telomeres? Int J Parasitol 33 29 45 12547344
Freitas-Junior LH Bottius E Pirrit LA Deitsch KW Scheidig C 2000 Frequent ectopic recombination of virulence factor genes in telomeric chromosome clusters of P. falciparum
Nature 407 1018 1022 11069183
Gardner MJ Hall N Fung E White O Berriman M 2002 Genome sequence of the human malaria parasite Plasmodium falciparum
Nature 419 498 511 12368864
Carlton JM Angiuoli SV Suh BB Kooij TW Pertea M 2002 Genome sequence and comparative analysis of the model rodent malaria parasite Plasmodium yoelii yoelii
Nature 419 512 519 12368865
van Lin LH Janse CJ Waters AP 2000 The conserved genome organisation of non-falciparum malaria species: The need to know more Int J Parasitol 30 357 370 10731560
Janse CJ Carlton JM Walliker D Waters AP 1994 Conserved location of genes on polymorphic chromosomes of four species of malaria parasites Mol Biochem Parasitol 68 285 296 7739674
Carlton JM Vinkenoog R Waters AP Walliker D 1998 Gene synteny in species of Plasmodium
Mol Biochem Parasitol 93 285 294 9662712
Hall N Pain A Berriman M Churcher C Harris B 2002 Sequence of Plasmodium falciparum chromosomes 1, 3–9 and 13 Nature 419 527 531 12368867
Kissinger JC Brunk BP Crabtree J Fraunholz MJ Gajria B 2002 The Plasmodium genome database Nature 419 490 492 12368860
Bahl A Brunk B Crabtree J Fraunholz MJ Gajria B 2003 PlasmoDB: The Plasmodium genome resource. A database integrating experimental and computational data Nucleic Acids Res 31 212 215 12519984
Hiller NL Bhattacharjee S van Ooij C Liolios K Harrison T 2004 A host-targeting signal in virulence proteins reveals a secretome in malarial infection Science 306 1934 1937 15591203
Marti M Good RT Rug M Knuepfer E Cowman AF 2004 Targeting malaria virulence and remodeling proteins to the host erythrocyte Science 306 1930 1933 15591202
Perkins ME 1984 Surface proteins of Plasmodium falciparum merozoites binding to the erythrocyte receptor, glycophorin J Exp Med 160 788 798 6206188
Sam-Yellowe TY Florens L Johnson JR Wang T Drazba JA 2004 A Plasmodium gene family encoding Maurer's cleft membrane proteins: Structural properties and expression profiling Genome Res 14 1052 1059 15140830
Spielmann T Fergusen DJ Beck HP 2003
etramps, a new Plasmodium falciparum gene family coding for developmentally regulated and highly charged membrane proteins located at the parasite-host cell interface Mol Biol Cell 14 1529 1544 12686607
Florens L Washburn MP Raine JD Anthony RM Grainger M 2002 A proteomic view of the Plasmodium falciparum life cycle Nature 419 520 526 12368866
Le Roch KG Zhou Y Blair PL Grainger M Moch JK 2003 Discovery of gene function by expression profiling of the malaria parasite life cycle Science 301 1503 1508 12893887
Bozdech Z Llinas M Pulliam BL Wong ED Zhu J 2003 The transcriptome of the intraerythrocytic developmental cycle of Plasmodium falciparum
PLoS Biol 1 e5. DOI: 10.1371/journal.pbio.0000005. 12929205
Pearce JA Mills K Triglia T Cowman AF Anders RF 2005 Characterisation of two novel proteins from the asexual stage of Plasmodium falciparum, H101 and H103 Mol Biochem Parasitol 139 141 151 15664649
Anamika Srinivasan N Krupa A 2005 A genomic perspective of protein kinases in Plasmodium falciparum
Proteins 58 180 189 15515182
Schneider AG Mercereau-Puijalon O 2005 A new Apicomplexa-specific protein kinase family: Multiple members in Plasmodium falciparum, all with an export signature BMC Genomics 6 30 15752424
Ward P Equinet L Packer J Doerig C 2004 Protein kinases of the human malaria parasite Plasmodium falciparum: The kinome of a divergent eukaryote BMC Genomics 5 79 15479470
Escalante AA Ayala FJ 1994 Phylogeny of the malarial genus Plasmodium, derived from rRNA gene sequences Proc Natl Acad Sci U S A 91 11373 11377 7972067
Nadeau JH Sankoff D 1997 Landmarks in the Rosetta Stone of mammalian comparative maps Nat Genet 15 6 7 8988158
Sutton GG White OR Adams MD Kerlavage AR 1995 TIGR Assembler: A new tool for assembling large shotgun sequencing projects Genome Sci Technol 1 9 19
Waters AP Higgins DG McCutchan TF 1991
Plasmodium falciparum appears to have arisen as a result of lateral transfer between avian and human hosts Proc Natl Acad Sci U S A 88 3140 3144 2014232
Escalante AA Barrio E Ayala FJ 1995 Evolutionary origin of human and primate malarias: Evidence from the circumsporozoite protein gene Mol Biol Evol 12 616 626 7659017
McCutchan TF, Kissinger JC, Touray MG, Rogers MJ, Li J et al. 1996 Comparison of circumsporozoite proteins from avian and mammalian malarias: Biological and phylogenetic implications Proc Natl Acad Sci U S A 93 11889 11894 8876233
Rathore D Wahl AM Sullivan M McCutchan TF 2001 A phylogenetic comparison of gene trees constructed from plastid, mitochondrial and genomic DNA of Plasmodium species Mol Biochem Parasitol 114 89 94 11356517
Kissinger JC Souza PC Soarest CO Paul R Wahl AM 2002 Molecular phylogenetic analysis of the avian malarial parasite Plasmodium (Novyella) juxtanucleare J Parasitol 88 769 773 12197128
Rich SM Ayala FJ 2003 Progress in malaria research: The case for phylogenetics Adv Parasitol 54 255 280 14711087
Carlton JM 1999 Gene synteny across Plasmodium spp: Could “operon-like” structures exist? Parasitol Today 15 178 179 10322349
Bailey JA Baertsch R Kent WJ Haussler D Eichler EE 2004 Hotspots of mammalian chromosomal evolution Genome Biol 5 R23 15059256
Armengol L Pujana MA Cheung J Scherer SW Estivill X 2003 Enrichment of segmental duplications in regions of breaks of synteny between the human and mouse genomes suggest their involvement in evolutionary rearrangements Hum Mol Genet 12 2201 2208 12915466
Kraemer SM Smith JD 2003 Evidence for the importance of genetic structuring to the structural and functional specialization of the Plasmodium falciparum var gene family Mol Microbiol 50 1527 1538 14651636
Delcher AL Phillippy A Carlton J Salzberg SL 2002 Fast algorithms for large-scale genome alignment and comparison Nucleic Acids Res 30 2478 2483 12034836
Gardner MJ Tettelin H Carucci DJ Cummings LM Aravind L 1998 Chromosome 2 sequence of the human malaria parasite Plasmodium falciparum
Science 282 1126 1132 9804551
Bailey TL Elkan C 1994 Fitting a mixture model by expectation maximization to discover motifs in biopolymers Proc Int Conf Intell Syst Mol Biol 2 28 36 7584402
Schultz J Milpetz F Bork P Ponting CP 1998 SMART, a simple modular architecture research tool: Identification of signaling domains Proc Natl Acad Sci U S A 95 5857 5864 9600884
Thompson JD Higgins DG Gibson TJ 1994 CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice Nucleic Acids Res 22 4673 4680 7984417
Felsenstein J 1996 Inferring phylogenies from protein sequences by parsimony, distance, and likelihood methods Methods Enzymol 266 418 427 8743697
Felsenstein J 1985 Confidence limits on phylogenies: An approach using the bootstrap Evolution 39 783 791 28561359
Strimmer K von Haeseler A 1996 Quartet puzzling: A quartet maximum likelihood method for reconstructing tree topologies Mol Biol Evol 13 964 969
Day WH McMorris FR 1993 A consensus program for molecular sequences Comput Appl Biosci 9 653 656 8143150
|
16389297
|
PMC1317653
|
CC BY
|
2021-01-05 12:11:25
|
no
|
PLoS Pathog. 2005 Dec 23; 1(4):e44
|
utf-8
|
PLoS Pathog
| 2,005 |
10.1371/journal.ppat.0010044
|
oa_comm
|
==== Front
PLoS GenetPLoS GenetpgenplgeplosgenPLoS Genetics1553-73901553-7404Public Library of Science San Francisco, USA 1638929810.1371/journal.pgen.001007905-PLGE-RA-0241R2plge-01-06-07Research ArticlePlant ScienceGenetics/EpigeneticsArabidopsis (Thale Cress)Differential Regulation of Strand-Specific Transcripts from Arabidopsis Centromeric Satellite Repeats Centromeric Transcripts in
ArabidopsisMay Bruce P ¤aLippman Zachary B ¤bFang Yuda Spector David L Martienssen Robert A *Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, United States of AmericaDoebley John EditorUniversity of Wisconsin, United States of America* To whom correspondence should be addressed. E-mail: [email protected]¤a Current address: NTL Ltd, Ho Chi Minh City, Vietnam
¤b Current address: Faculty of Agriculture, Institute of Plant Sciences, Hebrew University of Jerusalem, Rehovot, Israel
12 2005 23 12 2005 1 6 e7918 8 2005 16 11 2005 Copyright: © 2005 May 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.Centromeres interact with the spindle apparatus to enable chromosome disjunction and typically contain thousands of tandemly arranged satellite repeats interspersed with retrotransposons. While their role has been obscure, centromeric repeats are epigenetically modified and centromere specification has a strong epigenetic component. In the yeast Schizosaccharomyces pombe, long heterochromatic repeats are transcribed and contribute to centromere function via RNA interference (RNAi). In the higher plant Arabidopsis thaliana, as in mammalian cells, centromeric satellite repeats are short (180 base pairs), are found in thousands of tandem copies, and are methylated. We have found transcripts from both strands of canonical, bulk Arabidopsis repeats. At least one subfamily of 180–base pair repeats is transcribed from only one strand and regulated by RNAi and histone modification. A second subfamily of repeats is also silenced, but silencing is lost on both strands in mutants in the CpG DNA methyltransferase MET1, the histone deacetylase HDA6/SIL1, or the chromatin remodeling ATPase DDM1. This regulation is due to transcription from Athila2 retrotransposons, which integrate in both orientations relative to the repeats, and differs between strains of Arabidopsis. Silencing lost in met1 or hda6 is reestablished in backcrosses to wild-type, but silencing lost in RNAi mutants and ddm1 is not. Twenty-four–nucleotide small interfering RNAs from centromeric repeats are retained in met1 and hda6, but not in ddm1, and may have a role in this epigenetic inheritance. Histone H3 lysine-9 dimethylation is associated with both classes of repeats. We propose roles for transcribed repeats in the epigenetic inheritance and evolution of centromeres.
Synopsis
Centromeres are regions of the chromosome that pull the chromosomes to the correct daughter cell during division. They are surrounded by tens of thousands of short satellite repeats, commonly called “junk” DNA. The authors show that these repeats are transcribed into RNA, which is subject to RNA interference, giving rise to large amounts of small interfering RNA. Transcripts are associated with chromosomes during interphase, and mutants in heterochromatin formation have elevated transcript levels. At least two classes of transcripts are silenced by two different epigenetic mechanisms, in part because of transposons inserted into them. This pattern of insertion and regulation varies between natural accessions of Arabidopsis. The authors' results suggest a model for centromere evolution and speciation driven by mismatch between pericentromeric repeats and small interfering RNAs in wide crosses.
Citation:May BP, Lippman ZB, Fang Y, Spector DL, Martienssen RA (2005) Differential regulation of strand-specific transcripts from Arabidopsis centromeric satellite repeats. PLoS Genet 1(6): e79.
==== Body
Introduction
The centromeric regions of animal and plant chromosomes are large assemblies of thousands of short (approximately 151 to 340 base pairs [bp]) satellite repeats in head-to-tail orientation with interspersed retroelements. In Arabidopsis thaliana, these comprise 177- to 179-bp satellite repeats (cen180, also known as pAL1 and AtCon; Figure 1A), Athila LTR-retroelements, and 106B repeats, which are 398-bp internal portions of Athila2 LTRs [1–3]. The only noted common feature among eukaryotic alpha satellites is a binding site for CENP-B, an evolutionary relative of pogo transposase that is necessary for de novo centromere formation [4,5] but not for the centromeres of the human Y chromosome and some marker chromosomes that lack CENP-B [6,7]. In G2 of the cell cycle, a modified histone H3, CENP-A in humans and CenH3 (HTR12) in Arabidopsis, is incorporated into centromeric nucleosomes independently of DNA replication [8–10]. A complex of proteins, some of which are directly recruited by CENP-A, then assembles to form the kinetochore that moves the mitotic chromatid or meiotic univalent poleward along depolymerizing microtubules during anaphase.
Figure 1 Transcripts from Centromeric Repeats in Landsberg erecta
(A) cen180 repeats and oligonucleotide primers used in RT-PCR. C1, C2, and C3 are regions of conserved DNA sequence; V is a region of variable sequence. Fc and Rc primers were selected from the conserved regions.
(B) RT-PCR of cen180 and 106B repeat transcripts. RT was performed using the primers indicated, followed by PCR with the first primer and a corresponding primer on the other strand. Fc and Rc detect transcripts derived from both strands of bulk repeats, but the F primer detects strand-specific transcripts belonging to a subfamily of these repeats. Primers recognizing ACTIN2 transcripts served as positive controls. Negative controls lacking reverse transcriptase (“no RT”) tested for DNA contamination.
Although a specific binding site for CENP-A has been hypothesized, constitutive expression of CENP-A results in ectopic deposition at sites throughout the genome [11] and CENP-A has been observed to spread to formerly euchromatic regions lacking repeats during neocentromere creation [12,13]. Transgenic arrays of human centromeric repeats did not attract CENP-A until the cells were treated with an inhibitor of histone deacetylase [14] and yeast Cse4 can functionally substitute for CENP-A in human cells [15]. Thus, deposition of CENP-A may be guided by epigenetic features other than DNA sequence [16,17]. Centromeres appear cytologically as central constrictions flanked by conspicuous pericentromeric heterochromatin. Pericentromeric heterochromatin appears transcriptionally silent: it is depleted of histone H3 methylated at lysine 4 (H3K4me2) as well as histone acetylation, enriched for histone H3 methylated at lysine 9 (H3K9me2), and in mammals, filamentous fungi, and plants, it is enriched for 5-methylcytosine incorporated into the DNA [18–21]. In Schizosaccharomyces pombe, which lacks DNA methylation, the central region mediates attachment to the kinetochore while the heterochromatic outer regions attract H3K9me2, Swi6, and cohesin [22]. Unexpectedly, the heterochromatic repeats are transcribed and subject to RNA interference (RNAi) [23]. RNAi guides histone modification via an interaction between complexes containing small interfering RNAs (siRNAs) and nascent transcripts at the locus [24–26]. H3K9me2 in turn binds a repressor, Swi6 in yeast and heterochromatic protein 1 (HP1) in plants and animals, that propagates a more densely coiled, transcriptionally silenced chromatin structure [22]. In fission yeast, Swi6 interacts with the cohesin complex [27] so that the outer regions remain associated longer during mitosis than the central region or the chromosome arms [28]. Lagging chromosomes are observed at high frequency (20%) in mutants defective in heterochromatin formation such as clr4 (H3K9me2 methyltransferase), swi6 (HP1 homolog), rad21/cohesin, and RNAi [24,29].
This connection between RNAi and centromeric silencing has since been extended to mammalian, insect, and avian cells, but it is not known if the interaction with cohesin is common to other organisms [30–32]. Drosophila centromeres have subregions containing either the centromeric histone (Cid) or H3K4me2, and these cluster into higher order domains, with Cid clusters facing the kinetochore [33]. Mutants in Arabidopsis homologs of swi6 and rad21 have early flowering and meiotic phenotypes, respectively, but lagging mitotic chromosomes have not been reported [34,35].
DNA methylation is found in many eukaryotes (although not extensively, if at all, in S. pombe, Drosophila, or Caenorhabditis elegans) and interacts with histone deacetylation and H3K9me2, but the order of steps is unclear: In Arabidopsis, loss of H3K9me2 is accompanied by loss of DNA methylation [36–39] and loss of DNA methylation is accompanied by loss of H3K9me2 [39–41], although the latter effect may not be direct [20]. The DNA methyltransferase CHROMOMETHYLASE3 (CMT3), the histone H3 lysine-9 methyltransferase KRYPTONITE (KYP), and the RNAi component AGO4 maintain hypermethylation induced by transcription of long inverted repeats [37,42–45]. In contrast, DNA METHYLTRANSFERASE1 (MET1) and HISTONE DEACETYLASE6 (HDA6) are required to maintain silencing of promoters induced by transcription of short inverted repeats [46]. Despite the presence of siRNA, other components of the siRNA pathway have not been recovered in these screens, and AGO4 is not required for silencing mediated in trans by short inverted repeats [45]. But AGO4 is required, along with DICER-LIKE3 (DCL3) and RNA-DEPENDENT RNA POLYMERASE2 (RDR2), for DRM1- and DRM2-dependent DNA methylation and silencing of transgenes when they are introduced via agrobacterium-mediated transformation [47]. Similarly, two pathways have been identified for transposon silencing in Arabidopsis: One involves CMT3, KYP, and AGO1, and the other involves MET1, DDM1, and HDA6 [39]. In this second pathway, resilencing of some but not other transposons in backcrosses suggests a role for siRNA in the establishment of silencing [39]. For most transposons, a role for siRNA in the maintenance of silencing is not apparent in dcl3 mutants, although other dicers can compensate for the loss of DCL3 [48] so that such a role cannot be ruled out.
We have found that centromeric satellite repeats in Arabidopsis are transcribed strand-specifically and that the transcripts remain associated in the nucleus. Only a subset of 180-bp repeats is transcribed, but the repeats are partially silenced by an RNAi-based system including DCL1, AGO1, CMT3, KYP, and HDA6/SIL1. Loss of silencing is inherited epigenetically in backcrosses. Other 180-bp repeats are silent in wild-type (WT) but transcribed in met1, ddm1, and hda6. Athila2 LTR retroelements direct the transcription of these satellite repeats, and silencing is restored in met1 and hda6 backcrosses. Consistent with differing patterns of retroelement insertion, different repeats are transcribed in the Landsberg and Columbia strains of Arabidopsis, providing a possible mechanism for the rapid evolution of centromeric satellite repeats.
Results
Centromeric Repeats Are Transcribed and Remain in the Nucleus
Arabidopsis cen180 repeat families are 78% to 96% identical in sequence. Variant repeats are located on different chromosomes [49] and may, like human alpha satellites, form subdomains [50], but each repeat contains common conserved regions [2]. We designed cen180 Fc and Rc primers (Figure 1A) to recognize two conserved regions, and these primers amplified a large number of repeats from genomic DNA. Transcripts from centromeric repeats were detected by RT followed by PCR. “F-strand transcripts” are transcribed from the forward (Watson) DNA strand and reverse transcribed into cDNA using an F primer; “R-strand transcripts” are transcribed from the reverse (Crick) DNA strand and reverse transcribed into cDNA using an R primer. Using the highly conserved Fc and Rc primers, transcripts from both strands could be detected (Figure 1B). However, transcripts from each strand could be derived from different subfamilies of repeats. In order to test this possibility, specific primers were used (F primer and R primer, Figure 1A) to amplify only a subset of cen180 repeats [20]. In this case, we found that F-strand transcripts were much more abundant than R-strand transcripts in vegetative and floral tissues, suggesting strand-specific transcription of individual subfamilies of repeats (Figure 1B). Transcription from only one strand was consistent with transcripts initiating in tandem orientation from the repeats, while the amplification of 180-bp multimers indicated that transcripts did not terminate within each repeat. Transcript levels were highest in young seedlings and developing inflorescences, indicating expression in dividing cells.
The same subfamily of cen180 repeats were amplified from genomic DNA and used as strand-specific probes for in situ hybridization (see Materials and Methods). In longitudinal sections of WT inflorescences, F-strand transcripts were more abundant than R-strand transcripts (Figure 2), consistent with RT-PCR (Figure 1B). Young, proliferating tissues (Figure 2A and 2B) had more signal than mature tissue (Figure 2C and 2D). Strand and tissue specificity indicated that the signal came from cellular RNA and was not due to background hybridization with fortuitously denatured DNA. Under higher magnification, small punctate nuclear signals were seen with the F-strand probe, consistent with nascent transcripts remaining at the centromere (Figure 2E). In contrast, what little signal could be detected with the R-strand probe was nuclear but not punctate (Figure 2B and 2D).
Figure 2 In Situ Hybridization of Centromeric Probes with Longitudinal Sections of Inflorescences
(A) Inflorescence meristem and developing buds probed with digoxigenin-labeled F strand of cen180 F + R PCR products and (B) probed with R strand. Mature tissue about 5 mm below the meristem (C) probed with F strand of cen180 and (D) probed with R strand. The scale bars are 50 μm.
(E) More highly magnified view of nuclei from developing buds probed with F strand (scale bar is 10 μm).
Satellite Repeat Transcripts Are Regulated by DNA Methylation, Histone Modification, and RNAi
In S. pombe, H3K9me2 depends on RNA silencing of pericentromeric repeats [23], and vice versa [51]. Similarly, in Arabidopsis, H3K9me2, RNAi, and DNA methylations are mutually interdependent [39,45]. We examined the level of pericentromeric transcripts in Arabidopsis mutants defective in DNA and histone methylation as well as in RNAi (Figure 3A). Elevated transcript levels were detected in dcl1–9 (a weak allele of DICER-LIKE1), ago1–9 (a strong allele of ARGONAUTE1), rdr2–1 (a T-DNA insertion in RNA DEPENDENT RNA POLYMERASE2), dcl3–1 (a T-DNA insertion in DICER-LIKE3), kyp-2 (a splice-site mutation in histone H3 K9 methyltransferase), hda6/sil1 (an allele of HISTONE DEACETYLASE6), cmt3-m5662 (a null allele of the CNG DNA methyltransferase), met1–1 (a strong allele of the CG DNA methyltransferase), and ddm1–2 (a hypomorphic allele of the swi/snf chromatin remodeling ATPase). Strand-specific RT-PCR (Figure 3A) indicated that F-strand transcripts were elevated in abundance in cmt3, kyp, hda6, and the RNAi mutants, whereas both F- and R-strand transcripts accumulated in met1 and ddm1. In backcrosses to WT, elevated transcript levels were inherited epigenetically from cmt3, ddm1, and kyp. Reestablishment of silencing was observed for both strands in met1/+ backcrosses and on one strand in hda6/+ backcrosses. dcl1–9 and ago1–9 were not tested in this way due to sterility [39]. Northern analysis of met1 and dcl1 mutants revealed heterogeneous transcripts ranging from approximately 0.1 kb to more than 5 kb, indicating multiple, irregular transcription initiation and/or termination sites (Figure S1).
Figure 3 Regulation of Centromeric Transcripts
(A) Strand-specific RT-PCR of cen180 repeats. RT was performed with forward and reverse primers indicated to the left of each panel. Mutants in DNA methylation (cmt3, met1), chromatin remodeling (ddm1), histone modification (kyp, hda6), and RNAi (dcl1, ago1) were in the Landsberg erecta (Ler) ecotype. rdr2 and dcl3 were in the Columbia background. m indicates homozygous mutants and b indicates backcrosses of mutants to WT (wt). Positive controls were primers recognizing the ACTIN2 gene; negative controls lacked reverse transcriptase (no RT).
(B) Restriction site analysis of repeats. Repeats were amplified by cen180 F+R primers from genomic DNA (lanes 1), cDNA from met1 mutants (lanes 2), and cDNA from dcl1 mutants (lanes 3). Repeats in genomic DNA that have TaqI sites are not transcribed in either mutant, while those with Sau3AI sites are not transcribed in dcl1.
(C) Oligonucleotide probes specific for the subfamilies of F + R cen180 repeats expressed in dcl1 and met1 mutants.
(D) Strand-specific RT-PCR of cen180 repeats. RT was performed with F or R primers as in (A). Southern hybridization of these products with specific oligonucleotide probes distinguished subfamilies of repeats that were differentially regulated.
(E) Detection of read-through transcripts between cen180 and 106B repeats by strand-specific RT-PCR. 106B Fout (lane 1) and cen180 F (lane 2) primers were used to prime cDNA, followed by PCR amplification. The configuration of the transcripts is shown below each lane. Other combinations of primers did not give detectable products. RNA was prepared from aerial tissues from 28-day-old plants.
Satellite Repeat Subfamilies Are Silenced by Different Mechanisms
One explanation for this differential regulation was that further subfamilies of F + R repeats were differentially regulated on each strand. We therefore examined which particular repeats were expressed in met1 compared with dcl1, ago1, and WT using restriction and sequence analysis of cDNA. Digestion of PCR products from cDNA and genomic DNA with TaqI revealed that transcribed repeats, which were sensitive to TaqI digestion, were a minority of those amplified from genomic DNA, most of which were not sensitive (Figure 3B). Further, RT-PCR products from met1 were digested by Sau3AI, indicating they had Sau3AI sites, but cDNAs from dcl1 were not digested and must therefore differ in sequence. Transcribed repeats were sequenced and grouped by cDNA sequence similarity using CLUSTALW. Whereas those from WT, dcl1, and ago1 formed mixed clusters, transcripts from met1 were from a distinct subfamily of repeats (Figure S2).
Diagnostic 20-mers were designed from the most highly diverged satellite cDNAs (Figure 3C) and hybridized to strand-specific RT-PCR products from the various mutants (Figure 3D). Hybridization was exclusive and complementary. F-strand transcripts from the subfamily of repeats expressed in WT were elevated in cmt3, kyp, hda6, dcl1, and ago1, but transcripts from the R strand were not seen. Elevated transcript levels were inherited epigenetically when kyp, hda6, and cmt3 mutants were backcrossed to WT. The subfamily of repeats expressed in met1 was transcribed from both strands in met1 and hda6 but only from the F strand in ddm1. These transcripts could not be detected in WT and were epigenetically inherited in ddm1/+ but were substantially resilenced in hda6/+ or met1/+ backcross progeny. In this respect, they resemble ATGP1 gypsy-class LTR retrotransposons [39,41,52]. Additional R-strand transcripts, detected by RT-PCR in ddm1, did not hybridize with either probe, indicating a third subfamily of differentially regulated repeats.
The differentially regulated subfamilies of cen180 repeats are very similar, so sequence differences are unlikely to account for differences in regulation. A more likely explanation was that some cen180 transcripts might be driven by read-through from adjacent retroelements. The 106B LTR-like repeats and Athila-class LTR retroelements are integrated throughout pericentromeric regions [53] and are interspersed in random orientation with respect to cen180s (data not shown). F-strand 106B transcripts were detected in WT but were only weakly up-regulated in cmt3, kyp, and hda6 (Figure 3A). Transcripts from both strands were up-regulated in ddm1 and met1, and R-strand transcripts were up-regulated in hda6. In this respect, 106B repeats resembled both classes of cen180 repeats, and we examined the possibility they were co-transcribed using cDNA that was reverse transcribed with 106B or cen180 primers (Figure 3E). F-strand co-transcripts originating in cen180 repeat arrays could be detected in WT and dcl1. In contrast, reverse strand co-transcripts appeared to originate within 106B and could only be detected in met1. Subsequent RACE PCR showed that these co-transcripts originated in the LTR of Athila2 itself (not shown). Other co-transcripts were not seen. This indicated that differential genetic regulation of each class of cen180 repeat transcript could be explained by differential origin of the transcripts in 106B/Athila2 and cen180 repeats, respectively.
Polymorphic Regulation of Repeats
Most of the mutants described above were isolated in the Landsberg genetic background, but rdr2 and dcl3 were isolated in Columbia, and so this WT strain was also assayed for transcripts. Surprisingly, Columbia and Landsberg transcripts amplified with conserved primers Fc and Rc differed in their regulation (Figures 3A and 4). We therefore sequenced cen180 RT-PCR products from dcl3 and met1 mutants of Columbia and designed diagnostic primers for amplification of genomic DNA and cDNA (Figure 4). Primers corresponding to dcl3-and met1-specific transcripts amplified cDNA from each corresponding mutant in Columbia, as expected. However, no cDNA could be amplified from met1 or dcl1 mutants in Landsberg despite the presence of such repeats in the Landsberg genome (indicated by amplification of genomic DNA). Conversely, specific transcripts detected in dcl1 mutants in Landsberg were found to accumulate in met1 mutants of Columbia, indicating that the regulation of this class of repeats had changed during the divergence of these two ecotypes. Repeats transcribed specifically in met1 mutants of Landsberg do not appear to be present in the Columbia genome. The origin of this natural variation is discussed below.
Figure 4 Polymorphic Regulation of cen180 Transcripts
Primers were designed to recognize sample cen180 cDNAs from dcl1 and met1 mutants of Landsberg erecta (see Figure 3C) and dcl3 and met1 mutants of Columbia. PCR amplification using genomic DNA templates (DNA) tested for the presence of the particular repeats in the genome of each ecotype. RT-PCRs using RNA detected transcripts in WT (wt) and mutants of each ecotype as indicated. The cen180 Fc + Rc primers served as positive controls in the presence of reverse transcriptase and negative controls in its absence. RNA was prepared from aerial tissues from 28-day-old plants.
Histone Modifications Associated with Satellite Repeats
We attempted to examine histone methylation associated with centromeric repeats using semiquantitative chromatin immunoprecipitation, although this proved to be very difficult as previously reported [20,40] because of the very high copy numbers involved. Nonetheless, while quantitative changes were not reliable, qualitative changes could be detected (Figure S3). H3K9me2 was substantially reduced in kyp and ago1 over both classes of repeat, resembling transposons such as ATCOPIA4 [39] as well as pericentromeric repeats and retrotransposons in S. pombe and Drosophila [23,31]. In met1 and ddm1, H3K9me2 was largely lost from 106B but only slightly reduced at cen180. H3K4me2 underwent a modest but heritable increase in ddm1 and was found associated with each class of repeat in WT, perhaps reflecting the transcription of the repeats. The association of H3K4me2 with heterochromatic pericentromeric repeats has also been reported in S. pombe [23], while in Drosophila [33] it is interspersed with Cid in more central domains.
DNA methylation of centromeric repeats has been extensively investigated by gel blot analysis, and we have obtained similar results (Figure S4). CNG methylation is partial in WT and reduced in ddm1, cmt3, and kyp [21,37,42,43], but it is unaffected in hda6 [54]. CG methylation is more extensive than CNG methylation at the centromere in WT and is lost only in met1 and ddm1. However, met1 affects transcripts initiating in 106B/Athila2 LTRs and not those initiating within the class of cen180 repeats regulated by RNAi (Figure 3E). hda6 has a similar effect on 106B transcripts but has no effect on CG methylation, although H3K9me2 is reduced ([54] and data not shown). H3K9me2 was retained in rdr2 and dcl3 (data not shown). Thus, neither DNA methylation nor H3K9me2 correlates perfectly with the specific deregulation of 106B/Athila2- or cen180-driven satellite repeat transcripts, although complexes responsible for each modification play a major role.
siRNA from Centromeric Transcripts Depends on DDM1, DCL3, and RDR2
Centromeric repeat transcripts in S. pombe correspond to siRNAs [55] that depend on Dcr1, Rdr1, and Ago1 for their accumulation [56,57]. In Arabidopsis, 24-nucleotide siRNAs corresponding to 180-bp centromeric repeats were detected in WT and were unchanged in met1, hda6, cmt3, ago1, dcl1, and kyp but were reduced in ddm1 and almost entirely absent from dcl3 and rdr2 (Figure 5A and 5B). These blots were deliberately overexposed to show residual levels of 24 nucleotides and smaller classes of siRNA. It is not possible to tell if these are derived from specific subclasses of repeats. Small RNAs corresponding to both strands of 106B repeats were also unchanged in most mutants but were increased in ddm1, decreased in ago1 (Figure 5A), and could not be detected in dcl3 and rdr2 (Figure 5B), resembling siRNA derived from the gypsy class retrotransposons ATGP1 [39] and Athila2 (data not shown). Both dcl3 and rdr2 have been shown to be required for production of endogenous siRNAs, while dcl1 is required for processing of micro-RNAs [58] and trans-acting siRNAs [59,60].
Figure 5 siRNAs from Centromeric Repeats
Total RNA was purified from plants of the indicated genotypes, and small RNAs were enriched by solubility in PEG (see Materials and Methods). Mutants and WT (wt) were in the Landsberg erecta background (A) or the Columbia background (B). RNA was prepared from aerial tissues from 28-day-old plants.
Lagging Chromosomes Were Not Observed in Mutants Defective in Silencing
In the fission yeast S. pombe, mutants defective in centromeric silencing recruit, but fail to retain, cohesin in anaphase pericentromeric heterochromatin, because of the loss of H3K9me2 and associated Swi6 [61]. We therefore examined anaphase in centromeric silencing mutants in Arabidopsis. Developing floral or root tissue was fixed in paraformaldehyde and stained with DAPI to detect the presence of lagging or otherwise abnormal chromosomes during anaphase. In WT anaphase cells (N = 20), rdr2 mutant cells (N = 49), and ddm1 mutant cells (N = 23), no abnormality was seen (not shown). Further, these mutants were indistinguishable from WT in growth and fertility, unlike mutants in rad21/cohesin, which are semisterile due to defects in meiosis [34]. One explanation might be that H3K9me2 was largely retained in rdr2 and ddm1, so that any interaction with Swi6 homologs and cohesin was retained. However, mutants in the H3K9 methyltransferase kyp are also fully fertile, and lagging chromosomes were not observed in kyp anaphase cells either (N = 24). Differences between the yeast and plant systems are discussed below.
Discussion
Silencing of Centromeric Transcripts in Arabidopsis and Fission Yeast
Arabidopsis centromeric repeats are transcribed and, for at least one subfamily, transcripts from one strand accumulate predominantly and are regulated posttranscriptionally by RNAi, similar to the situation in S. pombe. This is consistent with the tandem orientation of these large blocks of repeats: putative promoter sequences found in each repeat would be expected to align in the same orientation as each other (Figure 6). Recently, each strand of the pericentromeric repeats in Arabidopsis was found to differ in DNA methylation [62]. While certainly an intriguing result, these differentially methylated sequences were located outside the satellite repeats and mostly composed of retrotransposons integrated in both orientations. Therefore, the significance of this methylation for strand-specific transcription of the satellite repeats is not yet clear.
Figure 6 A Model for the Regulation and Evolution of Centromeric Satellite Repeat Transcripts
(A) New tandemly arranged repeats are transcribed on one strand and processed by RNAi.
(B) Older repeats accumulate retrotransposon insertions that interrupt and silence the transcripts via MET1, HDA6, and DDM1.
(C) In met1, hda6, or ddm1 mutants, the silencing complex is lost and promoters from retrotransposon LTRs drive transcription from both strands. cen180 repeats are blue chevrons, and retrotransposons are green boxes (not to scale).
Even though they are transcribed from only one strand, tandem repeats can theoretically generate siRNA by reiterative rounds of RdRP replication and Dicer degradation [63]. The requirement for both DCL1/AGO1 and RDR2/DCL3 indicates both 21- and 24-nucleotide siRNAs may be involved. Although we could barely detect 21-nucleotide siRNA on blots, rare centromeric siRNA of this size has been detected by sequencing [64]. For other repeat subfamilies, both strands were silent, but transcription could be detected in mutants defective in the DNA methyltransferase MET1 and the type I histone deacetylase HDA6. These transcripts arose by read-out from LTR retrotransposons inserted among the repeats (see Figure 3E).
In fission yeast, RNAi is required to maintain normal levels of H3K9me2 at the outer pericentromeric repeats [23] and mutants have lagging chromosomes at anaphase [24,29]. A similar repeat is responsible for aspects of silencing at the mating type locus, but additional cis-acting elements are required for silencing in the absence of RNAi, requiring histone deacetylation instead [65]. In Arabidopsis, dcl3 and rdr2 mutants had little or no 24-nucleotide centromeric satellite siRNAs, yet had no detectable defects in H3K9me2 accumulation or in centromere function. This is also evidence of redundant mechanisms for maintaining H3K9me2, one RNAi based and the other depending on CG methylation and histone deacetylation of retrotransposons. However, kyp mutants lost most H3K9me2 without anaphase defects, although other H3K9 methyltransferases may be redundant with KYP [19], just as other Dicers are redundant with DCL3 [48]. Nevertheless, we cannot exclude the possibility that heterochromatic association of cohesin via H3K9me2 is not required for mitotic chromosome disjunction in plants [34], which lack centrosomes and may differ from animals in this respect.
If RNAi at the centromere is dispensable but transcripts are still found, this raises the possibility that the transcripts themselves may have some function. Centromeric transcripts from CRM retrotransposons and CentC satellite repeats are associated with immunoprecipitated kinetochores in maize, although neither full-length transcripts nor siRNAs were detected on Northern blots, so that the origin and fate of these transcripts are unknown [66]. Similarly, major and minor satellite repeats are transcribed from mouse centromeres, but in this case RNA interference is thought to play a role [30,67]. Our results indicate that centromeric transcripts can be long, persist in the nucleus, and show some correlation with mitotic activity and may associate with chromosomes. Whether the transcripts serve to recruit factors in addition to the RNAi apparatus remains to be seen.
The role of DDM1 in siRNA production is unclear, but it may stabilize siRNA in a complex [39], or even promote RNA-dependent RNA polymerase, in the same way as other swi/snf helicases promote DNA-dependent RNA polymerases [68]. This could account for the difference between 106B and cen180 siRNA accumulation in ddm1. If siRNA processing was reduced but 106B was transcriptionally up-regulated, this could lead to an overall increase in siRNA in ddm1. In contrast, if cen180 repeats were not transcriptionally up-regulated, siRNA would be decreased in ddm1. Combined losses of siRNA and H3K9me2 are most severe in ddm1, and chromocenters are severely disrupted [69], indicating they may play a role in heterochromatin association even though lagging chromosomes were not observed.
Maintenance and Reestablishment of Silencing
Silencing of centromeric repeats was restored in met1/+ following backcrosses to WT, as was methylation [70]. However, transgene silencing and methylation were not restored in similar backcrosses, although allelic differences could be responsible [52]. Transposon methylation could also be restored in met1/+ but only for those transposons that retained H3K9me2 and siRNA in met1 [39]. siRNA from cen180 repeats was also retained in met1 and hda6, but it was reduced in ddm1 and F-strand silencing was not restored in ddm1/+. These results are consistent with a silencing complex, composed of DDM1, MET1, and HDA6, while de novo histone and DNA modification is guided by siRNA [39]. HDA6 may act downstream of MET1 as CG methylation is unaffected in hda6 mutants [54] (Figure S4), consistent with the interaction of histone deacetylase with CG methyl-binding proteins in mammalian cells [71]. cen180 repeats transcribed in cmt3 and kyp are inherited epigenetically in an active state in backcrosses, indicating silencing cannot be reestablished in trans even in the presence of cen180 siRNA, which is retained in the mutants. H3K9me2 is lost in kyp and may play a role in reestablishment. In cmt3, H3K9me2 is retained but CNG methylation is lost so both marks may be required. However, silencing of some, but not all, of these transcripts is restored in hda6/+, indicating that epigenetic inheritance is not simply a matter of chromosomal modification.
Evolution of Centromeres
Centromeres are dynamic components of genome evolution. In sequenced arrays of human 171-bp satellites, the central repeats appear to be the youngest, with progressively older repeats on the outside. Formation and homogenization of new repeats are hypothesized to occur via unequal recombination, so that one repetitive sequence comes to dominate the centromere [50]. CENP-A has been proposed to co-evolve with such repeats, although binding to new, homogenous repeats has not been directly tested [72]. Transcribed repeats could also be replicated via RT and retrotransposition. New copies, being identical in sequence, would still be recognized by siRNA. As repeats age, retrotransposon insertions would recruit MET1 and DDM1, silencing the repeats transcriptionally and allowing them to diverge (Figure 6). This model predicts that younger RNAi-regulated repeats should be toward the center of the centromere and older MET1-regulated repeats should be on the flanks. In support, BLAST analysis showed that MET1-regulated repeats do indeed predominate in the sequenced portion of the genome, i.e., the exterior regions of the centromere (Figure S5). The regulation of subfamilies of centromeric repeats has diverged in the Columbia and Landsberg ecotypes, which are closely related [73]. Although the centromeric heterochromatin has not been completely sequenced in either strain, the pattern of retrotransposon insertions elsewhere in the genome is quite different [74], and this could account for the differential silencing of individual subfamilies of centromeric repeats in each strain (Figure 6). The transcription of centromeric repeats provides an attractive model for speciation. In wide crosses, paternal chromosomes might be destabilized if the pericentromeric repeats no longer match the sequence of maternal siRNA. This mechanism could contribute to hybrid incompatibility in polyploids as well as the loss of paternal chromosomes in wide crosses [75,76].
Materials and Methods
Plant material.
cmt3-m5662, met1–1 (E. Richards), ddm1–2 (E. Richards), kyp-2 (S. Jacobsen), hda6/sil1 (I. Furner), dcl1–9 (S. Jacobsen), ago1–9 mutations were introgressed into the Landsberg erecta ecotype. dcl3–1 (J. Carrington) and rdr2–1 (J. Carrington) are in the Columbia ecotype and were obtained from the Syngenta Arabidopsis Insertion Library (SAIL) collection of T-DNA insertions (Syngenta Biotechnology Inc. [SBI], Research Triangle Park, North Carolina, United States). Plants were grown 28 days (16 h light, 8 h dark) for RNA and chromatin preparation. Sources for seed stocks are indicated in parentheses.
Transcript analysis.
RNA was prepared, reverse transcribed into DNA, and amplified as described [39]. RT-PCRs were performed with 100 ng of RNA per reaction using the OneStep kit from Qiagen (Valencia, California, United States) according to the manufacturer's protocol. Negative controls to detect contaminating DNA were performed on the RNA preparations using cen180 F + R or cen180 Fc + Rc primers but no RT. The following reaction conditions were used in each cycle: 94 °C for 20 s, 60 °C for 30 s, and 72 °C for 1 min. Southern hybridization was performed using standard methods [77]. Primer sequences were selected from conserved and variable regions of cen180 repeats, also known as pAL1 and AtCon (see Supplementary Information). The cen180 F (forward) primer is conserved in Arabidopsis and related species, while the R (reverse) primer is specific to subsets of repeats [2,20]. The cen180 Fc and Rc primers match conserved regions within the repeat. For CLUSTALW analysis, sequences were determined for 25 to 30 cDNA clones from each mutant. A cen180 F primer with a T3 promoter and a cen180 R primer with a T7 promoter were used to amplify cen180 repeats, and these products were transcribed in vitro to prepare strand-specific probes for in situ hybridization, performed according to the protocol of Jeffrey Long (Salk Institute for Biological Studies, San Diego, California, United States) (for protocol, see http://www.its.caltech.edu/~plantlab/protocols/insitu.pdf).
Histone H3 methylation.
Chromatin immunoprecipitation (ChIP) was performed as previously described [39,40] using conserved F primers and R primers specific to each class of repeat. Semiquantification of the cen180 repeat ChIP data was performed by comparing PCR results from three different cycle numbers (13, 15, and 17 cycles), which were then analyzed by Southern blotting. In this way, the PCR of such highly repetitive sequences was maintained in the linear phase to avoid PCR saturation. DNA samples from each genotype were then normalized to each other by amplifying dilutions of total input DNA. Semiquantitative data were then obtained by comparing amplification with each set of primer pairs within the same ChIP extraction, which served as internal controls. In this way, control primers such as actin, whose association with lysine-9 is unclear, could be avoided. In all cases, mock precipitation with no antibody yielded little or no product. In Figure S3, PCR conditions were 17 cycles for F and R primers and 28 cycles for dcl- and met-specific primers. PCR products were blotted and probed with radiolabeled cen180 repeats.
siRNA hybridization.
Total RNA was purified, large RNAs were removed by precipitation in 5% PEG-8000/0.5 M NaCl, and 50 μg of the resulting enriched small RNA was loaded per lane in a 10% polyacrylamide/urea gel and separated by electrophoresis [39]. The gel was electroblotted and probed with in vitro transcribed, radiolabeled RNA. Imaging was performed on a PhosphorImager (Fuji, Tokyo, Japan).
Supporting Information
Figure S1 Northern Analysis of cen180 Transcripts
Total RNA (50 μg) from WT (lane 1), met1 mutants (lane2), and dcl1 mutants (lane 3) were separated on an agarose gel, blotted, and probed with radiolabeled transcripts corresponding to each strand (F and R) of cen180 repeats. Heterogeneous high-molecular-weight transcripts hybridize to both strands in met1 mutants. Prominent bands cross-hybridize with rDNA (left).
(171 KB JPG)
Click here for additional data file.
Figure S2 CLUSTALW Analysis of Sequences of Transcribed Repeats
An unrooted tree of cen180 cDNA sequences from different mutants in the Landsberg ecotype. Sequence identifiers beginning with w are from WT; d, from dcl1; a, from ago1; and m, from met1. The met1 branches are red. Sequences found in WT, ago1, and dcl1 were not found in met1.
(183 KB PDF)
Click here for additional data file.
Figure S3 Methylation of Lysine-9 (K9) and Lysine-4 (K4) of Histone H3 at Centromeric Repeats
Chromatin was immunoprecipitated with anti–dimethyl K9 antibody (K9), anti–dimethyl K4 antibody (K4), or no antibody (na). Precipitated samples and total input chromatin (tot) were then PCR-amplified with cen180 F + R primers, F + dcl–specific primers, F + met–specific primers, and 106B F + R primers.
(1.9 MB JPG)
Click here for additional data file.
Figure S4 Centromeric Repeat Methylation
Genomic DNAs were prepared from 3-week-old plants of the indicated genotypes. Methylation changes at (A) cen180 repeats and (B) 106B repeats were assayed by digestion with HpaII (left) and MspI (right) for each repeat followed by blotting and Southern hybridization.
(775 KB JPG)
Click here for additional data file.
Figure S5 Comparison of cen180 cDNAs from WT (wt), dcl3 Mutants, and met1 Mutants with the Sequenced Region of the Arabidopsis Genome
Sequences of 32 cDNAs from WT and each mutant in the Columbia ecotype were BLASTed against the Columbia genome (TIGR version 5, http://www.tigr.org) using a cutoff of E < 0.0001. The matches for each repeat were collected and sorted into bins according to score. Each bin is 10 score points.
(1.2 MB JPG)
Click here for additional data file.
Text S1 Primer Sequences
The same oligonucleotide primers were used for RT-PCR, ChIP, and probe amplification.
(24 KB DOC)
Click here for additional data file.
Accession Numbers
The dBEST accession numbers in GenBank (http://www.ncbi.nlm.nih.gov/Genbank) for the centromeric satellite cDNA sequences are DV671393 to DV671628.
We thank I. Furner, J. Carrington, S. Jacobsen, and E. Richards for Arabidopsis strains and E. Lam, K. Dawe, and G. Karpen for helpful discussions. ZBL was an Arnold and Mabel Beckman fellow of the Watson School of Biological Sciences. This work was supported by National Science Foundation Plant Genome Program grant DBI-077617 to DLS and RAM and grant DBI-077774 to RAM.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. BPM, ZBL, and RAM conceived and designed the experiments. BPM, ZBL, and YF performed the experiments. BPM, ZBL, YF, DLS, and RAM analyzed the data. BPM and ZBL contributed reagents/materials/analysis tools. BPM, ZBL, and RAM wrote the paper.
Abbreviations
bpbase pair
RNAiRNA interference
siRNAsmall interfering RNA
WTwild-type
==== Refs
References
Martinez-Zapater J Estelle MA Somerville CR 1986 A highly repeated DNA sequence in Arabidopsis thaliana
Mol Gen Genet 204 417 423
Hall SE Kettler G Preuss D 2003 Centromere satellites from Arabidopsis populations: Maintenance of conserved and variable domains Genome Res 13 195 205 12566397
Thompson HL Schmidt R Dean C 1996 Identification and distribution of seven classes of middle-repetitive DNA in the Arabidopsis thaliana genome Nucleic Acids Res 24 3017 3022 8760888
Ohzeki J Nakano M Okada T Masumoto H 2002 CENP-B box is required for de novo centromere chromatin assembly on human alphoid DNA J Cell Biol 159 765 775 12460987
Basu J Stromberg G Compitello G Willard HF Van Bokkelen G 2005 Rapid creation of BAC-based human artificial chromosome vectors by transposition with synthetic alpha-satellite arrays Nucleic Acids Res 33 587 596 15673719
Earnshaw WC 1991 When is a centromere not a kinetochore? J Cell Sci 99 Pt 1 1 4 1757494
Brown W Tyler-Smith C 1995 Centromere activation Trends Genet 11 337 339 7482782
Shelby RD Monier K Sullivan KF 2000 Chromatin assembly at kinetochores is uncoupled from DNA replication J Cell Biol 151 1113 1118 11086012
Ahmad K Henikoff S 2001 Centromeres are specialized replication domains in heterochromatin J Cell Biol 153 101 110 11285277
Smith MM 2002 Centromeres and variant histones: What, where, when and why? Curr Opin Cell Biol 14 279 285 12067649
Shelby RD Vafa O Sullivan KF 1997 Assembly of CENP-A into centromeric chromatin requires a cooperative array of nucleosomal DNA contact sites J Cell Biol 136 501 513 9024683
Maggert KA Karpen GH 2001 The activation of a neocentromere in Drosophila requires proximity to an endogenous centromere Genetics 158 1615 1628 11514450
Amor DJ Bentley K Ryan J Perry J Wong L 2004 Human centromere repositioning “in progress.” Proc Natl Acad Sci U S A 101 6542 6547 15084747
Nakano M Okamoto Y Ohzeki J Masumoto H 2003 Epigenetic assembly of centromeric chromatin at ectopic alpha-satellite sites on human chromosomes J Cell Sci 116 4021 4034 12953060
Wieland G Orthaus S Ohndorf S Diekmann S Hemmerich P 2004 Functional complementation of human centromere protein A (CENP-A) by Cse4p from Saccharomyces cerevisiae
Mol Cell Biol 24 6620 6630 15254229
Choo KH 2000 Centromerization Trends Cell Biol 10 182 188 10754560
Henikoff S Ahmad K Malik HS 2001 The centromere paradox: Stable inheritance with rapidly evolving DNA Science 293 1098 1102 11498581
Attwood JT Yung RL Richardson BC 2002 DNA methylation and the regulation of gene transcription Cell Mol Life Sci 59 241 257 11915942
Bender J 2004 Chromatin-based silencing mechanisms Curr Opin Plant Biol 7 521 526 15337094
Johnson L Cao X Jacobsen S 2002 Interplay between two epigenetic marks. DNA methylation and histone H3 lysine 9 methylation Curr Biol 12 1360 1367 12194816
Vongs A Kakutani T Martienssen RA Richards EJ 1993
Arabidopsis thaliana DNA methylation mutants Science 260 1926 1928 8316832
Bannister AJ Zegerman P Partridge JF Miska EA Thomas JO 2001 Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain Nature 410 120 124 11242054
Volpe TA Kidner C Hall IM Teng G Grewal SI 2002 Regulation of heterochromatic silencing and histone H3 lysine-9 methylation by RNAi Science 297 1833 1837 12193640
Volpe T Schramke V Hamilton GL White SA Teng G 2003 RNA interference is required for normal centromere function in fission yeast Chromosome Res 11 137 146 12733640
Verdel A Jia S Gerber S Sugiyama T Gygi S 2004 RNAi-mediated targeting of heterochromatin by the RITS complex Science 303 672 676 14704433
Motamedi MR Verdel A Colmenares SU Gerber SA Gygi SP 2004 Two RNAi complexes, RITS and RDRC, physically interact and localize to noncoding centromeric RNAs Cell 119 789 802 15607976
Bernard P Maure JF Partridge JF Genier S Javerzat JP 2001 Requirement of heterochromatin for cohesion at centromeres Science 294 2539 2542 11598266
Appelgren H Kniola B Ekwall K 2003 Distinct centromere domain structures with separate functions demonstrated in live fission yeast cells J Cell Sci 116 4035 4042 12928332
Hall IM Noma K Grewal SI 2003 RNA interference machinery regulates chromosome dynamics during mitosis and meiosis in fission yeast Proc Natl Acad Sci U S A 100 193 198 12509501
Kanellopoulou C Muljo SA Kung AL Ganesan S Drapkin R 2005 Dicer-deficient mouse embryonic stem cells are defective in differentiation and centromeric silencing Genes Dev 19 489 501 15713842
Pal-Bhadra M Leibovitch BA Gandhi SG Rao M Bhadra U 2004 Heterochromatic silencing and HP1 localization in Drosophila are dependent on the RNAi machinery Science 303 669 672 14752161
Fukagawa T Nogami M Yoshikawa M Ikeno M Okazaki T 2004 Dicer is essential for formation of the heterochromatin structure in vertebrate cells Nat Cell Biol 6 784 791 15247924
Sullivan BA Karpen GH 2004 Centromeric chromatin exhibits a histone modification pattern that is distinct from both euchromatin and heterochromatin Nat Struct Mol Biol 11 1076 1083 15475964
Bhatt AM Lister C Page T Fransz P Findlay K 1999 The DIF1 gene of Arabidopsis is required for meiotic chromosome segregation and belongs to the REC8/RAD21 cohesin gene family Plant J 19 463 472 10504568
Gaudin V Libault M Pouteau S Juul T Zhao G 2001 Mutations in LIKE HETEROCHROMATIN PROTEIN 1 affect flowering time and plant architecture in Arabidopsis Development 128 4847 4858 11731464
Tamaru H Selker EU 2001 A histone H3 methyltransferase controls DNA methylation in Neurospora crassa
Nature 414 277 283 11713521
Jackson JP Lindroth AM Cao X Jacobsen SE 2002 Control of CpNpG DNA methylation by the KRYPTONITE histone H3 methyltransferase Nature 416 556 560 11898023
Lehnertz B Ueda Y Derijck AA Braunschweig U Perez-Burgos L 2003 Suv39h-mediated histone H3 lysine 9 methylation directs DNA methylation to major satellite repeats at pericentric heterochromatin Curr Biol 13 1192 1200 12867029
Lippman Z May B Yordan C Singer T Martienssen R 2003 Distinct mechanisms determine transposon inheritance and methylation via small interfering RNA and histone modification PLoS Biol 1 e67. DOI: 10.1371/journal.pbio.0000067 14691539
Gendrel AV Lippman Z Yordan C Colot V Martienssen RA 2002 Dependence of heterochromatic histone H3 methylation patterns on the Arabidopsis gene DDM1 Science 297 1871 1873 12077425
Soppe WJ Jasencakova Z Houben A Kakutani T Meister A 2002 DNA methylation controls histone H3 lysine 9 methylation and heterochromatin assembly in Arabidopsis EMBO J 21 6549 6559 12456661
Lindroth AM Cao X Jackson JP Zilberman D McCallum CM 2001 Requirement of CHROMOMETHYLASE3 for maintenance of CpXpG methylation Science 292 2077 2080 11349138
Bartee L Malagnac F Bender J 2001 Arabidopsis cmt3 chromomethylase mutations block non-CG methylation and silencing of an endogenous gene Genes Dev 15 1753 1758 11459824
Malagnac F Bartee L Bender J 2002 An Arabidopsis SET domain protein required for maintenance but not establishment of DNA methylation EMBO J 21 6842 6852 12486005
Zilberman D Cao X Jacobsen SE 2003 ARGONAUTE4 control of locus-specific siRNA accumulation and DNA and histone methylation Science 299 716 719 12522258
Aufsatz W Mette MF Matzke AJ Matzke M 2004 The role of MET1 in RNA-directed de novo and maintenance methylation of CG dinucleotides Plant Mol Biol 54 793 804 15604652
Chan SW Zilberman D Xie Z Johansen LK Carrington JC 2004 RNA silencing genes control de novo DNA methylation Science 303 1336 14988555
Gasciolli V Mallory AC Bartel DP Vaucheret H 2005 Partially redundant functions of Arabidopsis DICER-like enzymes and a role for DCL4 in producing trans-acting siRNAs Curr Biol 15 1494 1500 16040244
Heslop-Harrison JS Murata M Ogura Y Schwarzacher T Motoyoshi F 1999 Polymorphisms and genomic organization of repetitive DNA from centromeric regions of Arabidopsis chromosomes Plant Cell 11 31 42 9878630
Schueler MG Higgins AW Rudd MK Gustashaw K Willard HF 2001 Genomic and genetic definition of a functional human centromere Science 294 109 115 11588252
Schramke V Allshire R 2003 Hairpin RNAs and retrotransposon LTRs effect RNAi and chromatin-based gene silencing Science 301 1069 1074 12869699
Saze H Scheid OM Paszkowski J 2003 Maintenance of CpG methylation is essential for epigenetic inheritance during plant gametogenesis Nat Genet 34 65 69 12669067
Fransz P Armstrong S Alonso-Blanco C Fischer TC Torres-Ruiz RA 1998 Cytogenetics for the model system Arabidopsis thaliana Plant J 13 867 876 9681023
Aufsatz W Mette MF van der Winden J Matzke M Matzke AJ 2002 HDA6, a putative histone deacetylase needed to enhance DNA methylation induced by double-stranded RNA EMBO J 21 6832 6841 12486004
Reinhart BJ Bartel DP 2002 Small RNAs correspond to centromere heterochromatic repeats Science 297 1831 12193644
Verdel A Moazed D 2005 RNAi-directed assembly of heterochromatin in fission yeast FEBS Lett 579 5872 5878 16223485
Li F Goto DB Zaratiegui M Tang X Martienssen R 2005 Two novel proteins, Dos1 and Dos2, interact with Rik1 to regulate heterochromatic RNA interference and histone modification Curr Biol 15 1448 1457 16040243
Xie Z Johansen LK Gustafson AM Kasschau KD Lellis AD 2004 Genetic and functional diversification of small RNA pathways in plants PLoS Biol 2 e104. DOI: 10.1371/journal.pbio.0020104 15024409
Vazquez F Vaucheret H Rajagopalan R Lepers C Gasciolli V 2004 Endogenous trans-acting siRNAs regulate the accumulation of Arabidopsis mRNAs Mol Cell 16 69 79 15469823
Allen E Xie Z Gustafson AM Carrington JC 2005 microRNA-directed phasing during trans-acting siRNA biogenesis in plants Cell 121 207 221 15851028
Pidoux A Mellone B Allshire R 2004 Analysis of chromatin in fission yeast Methods 33 252 259 15157893
Luo S Preuss D 2003 Strand-biased DNA methylation associated with centromeric regions in Arabidopsis Proc Natl Acad Sci U S A 100 11133 11138 12960391
Martienssen RA 2003 Maintenance of heterochromatin by RNA interference of tandem repeats Nat Genet 35 213 214 14593407
Gustafson AM Allen E Givan S Smith D Carrington JC 2005 ASRP: The Arabidopsis Small RNA Project Database Nucleic Acids Res 33 D637 D640 15608278
Martienssen RA Zaratiegui M Goto DB 2005 RNA interference and heterochromatin in the fission yeast Schizosaccharomyces pombe
Trends Genet 21 450 456 15979194
Topp CN Zhong CX Dawe RK 2004 Centromere-encoded RNAs are integral components of the maize kinetochore Proc Natl Acad Sci U S A 101 15986 15991 15514020
Rudert F Bronner S Garnier JM Dolle P 1995 Transcripts from opposite strands of gamma satellite DNA are differentially expressed during mouse development Mamm Genome 6 76 83 7767009
Schnitzler G Sif S Kingston RE 1998 Human SWI/SNF interconverts a nucleosome between its base state and a stable remodeled state Cell 94 17 27 9674423
Fransz P De Jong JH Lysak M Castiglione MR Schubert I 2002 Interphase chromosomes in Arabidopsis are organized as well defined chromocenters from which euchromatin loops emanate Proc Natl Acad Sci U S A 99 14584 14589 12384572
Kankel MW Ramsey DE Stokes TL Flowers SK Haag JR 2003 Arabidopsis MET1 cytosine methyltransferase mutants Genetics 163 1109 1122 12663548
Nan X Ng HH Johnson CA Laherty CD Turner BM 1998 Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex Nature 393 386 389 9620804
Malik HS Henikoff S 2002 Conflict begets complexity: The evolution of centromeres Curr Opin Genet Dev 12 711 718 12433586
Sharbel TF Haubold B Mitchell-Olds T 2000 Genetic isolation by distance in Arabidopsis thaliana: Biogeography and postglacial colonization of Europe Mol Ecol 9 2109 2118 11123622
Borevitz JO Liang D Plouffe D Chang HS Zhu T 2003 Large-scale identification of single-feature polymorphisms in complex genomes Genome Res 13 513 523 12618383
Osborn TC Pires JC Birchler JA Auger DL Chen ZJ 2003 Understanding mechanisms of novel gene expression in polyploids Trends Genet 19 141 147 12615008
Ananiev EV Riera-Lizarazu O Rines HW Phillips RL 1997 Oat-maize chromosome addition lines: A new system for mapping the maize genome Proc Natl Acad Sci U S A 94 3524 3529 9108009
Sambrook J Russell DW 2001 Molecular Cloning: A Laboratory Manual Cold Spring Harbor (New York) Cold Spring Harbor Laboratory Press 999 pp.
|
16389298
|
PMC1317654
|
CC BY
|
2021-01-05 07:59:38
|
no
|
PLoS Genet. 2005 Dec 23; 1(6):e79
|
utf-8
|
PLoS Genet
| 2,005 |
10.1371/journal.pgen.0010079
|
oa_comm
|
==== Front
Ann Gen PsychiatryAnnals of General Psychiatry1744-859XBioMed Central London 1744-859X-4-181631647310.1186/1744-859X-4-18ReviewPrimary care use of antipsychotic drugs: an audit and intervention study Mortimer Ann M [email protected] Charles J [email protected] Michael [email protected] Alison [email protected] Foundation Chair in Psychiatry/Head of Department, The Department of Psychiatry, The University of Hull, Cottingham Road, Hull, HU6 7RX, UK2 Research Nurse, The Department of Psychiatry, The University of Hull, Cottingham Road, Hull, HU6 7RX, UK3 Pharmaceutical Advisor, Eastern Hull Primary Care Team, Central Office, Netherhall, Wawne Road, Sutton, UK, Hull, HU7 4YG, UK4 Consultant Psychiatrist, Harrogate District Hospital, Lancaster Park Road, Harrogate, North Yorkshire, HG2 7SX, UK2005 29 11 2005 4 18 18 27 9 2005 29 11 2005 Copyright © 2005 Mortimer et al; licensee BioMed Central Ltd.2005Mortimer 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
Concerns regarding the use of antipsychotic medication in secondary care suggested an examination of primary care prescribing.
Aim
To audit and intervene in the suboptimal prescribing of antipsychotic drugs to primary care patients.
Design of study
Cross-sectional prevalence: subsequent open treatment intervention.
Setting
Seven of the 29 practices in the Eastern Hull Primary Care Trust.
Methods
Criteria for best practice were developed, against which prescribing standards were tested via audit. Patients identified as suboptimally prescribed for were invited to attend an expert review for intervention.
Results
1 in 100 of 53,000 patients was prescribed antipsychotic treatment. Diagnoses indicating this were impossible to ascertain reliably. Half the regimes failed one or more audit criteria, leaving diagnosis aside. Few practices agreed to patients being approached: of 179 invitations sent, only 40 patients attended. Of 32 still taking an antipsychotic drug, 26 required changes. Mean audit criteria failed were 3.4, lack of psychotic disorder diagnosis and problematic side effects being most frequent. Changes were fully implemented in only 16 patients: reasons for complete or partial failure to implement recommendations included the wishes or inaction of patients and professionals, and worsening of symptoms including two cases of antipsychotic withdrawal syndrome.
Conclusion
Primary care prescribing of antipsychotic drugs is infrequent, but most is unsatisfactory. Intervention is hampered by pluralistic reluctance: even with expert guidance, rationalisation is not without risk. Use of antipsychotic drugs in primary care patients whose diagnosis does not warrant this should be avoided.
How this fits in
This study adds to concerns regarding high levels of off-licence use of potentially harmful medication. It adds evidence of major difficulties in rationalizing suboptimal regimes despite expert input. Relevance to the clinician is that it is better to avoid such regimes in the first place especially if there is no clear 'exit strategy': if in doubt, seek a specialist opinion.
==== Body
Introduction
We have previously published on the utilization of high dose antipsychotic treatment and polypharmacy in secondary care, and lack of adherence to appropriate guidelines [25]. The publication of NICE guidance on antipsychotic treatment in schizophrenia [4] would, we assumed, result in positive changes in secondary care prescribing. This guidance recommended atypical antipsychotic drugs in many common clinical situations, including for new patients, relapsing patients and symptomatically well controlled patients if side effects were unacceptable: polypharmacy and high doses were advised against.
Given the unsatisfactory state of secondary care prescribing demonstrated by our first study, we considered that the situation in primary care may benefit from examination particularly in the context of NICE. We therefore set up a further audit to identify patients of general practitioners receiving potentially problematic antipsychotic regimes, with a subsequent optional intervention to be offered to these GPs and their patients to rationalize their medication. The overall aim was to improve the wellbeing of a large number of patients currently receiving antipsychotic treatment sub-optimally. Optimizing such medication regimes should, we anticipated, have the effect of minimizing symptoms and side effects while maximizing quality of life.
The Eastern Hull Primary Care Trust (PCT) agreed to support the audit. This PCT has a catchment population of 125,000, with a typical range of urban inner-city health & social problems. It comprises 29 practices including 57 GPs, 17 of them single handed. There were at the time of the audit 12 community pharmacists working with 23 of the practices, offering hands-on prescribing support. From 2000 to 2003, the total number of prescriptions for antipsychotic drugs in Eastern Hull PCT rose moderately from 12117 to 12703 per year: however their cost rose markedly, from £215,752 to £324,511. While the usage and cost of conventional and depot medications remained constant, the usage and cost of atypical antipsychotic drugs, particularly olanzapine and risperidone, increased substantially.
Method
The following audit criteria were adopted to define possible suboptimal prescribing in the patient group. They were derived from the literature, and a process of discussion and consensus finding between the four authors.
1. On thioridazine [22]
2. On more than one antipsychotic drug [4]
3. Psychotropic polypharmacy (increased risks of side effects and interactions: evidence in support of efficacy unclear in many diagnostic categories)
4. Greater than recommended maintenance dose [4]
5. Dose less than a quarter of recommended maintenance dose (therefore dubious efficacy)
6. No current diagnosis indicating an antipsychotic i.e. psychosis or short-term behavioural disturbance [3]
7. Long term anticholinergic treatment [27]
8. Not reviewed by GP or psychiatrist for 1 year
9. Unresolved problematic symptoms
10. Unresolved problematic side effects
Community pharmacists working in GP practices attended a training session about the project and the audit criteria, run by AM and a research nurse. They then audited all patients prescribed any antipsychotic medication in 7 of the 29 practices in Eastern Hull PCT. Patients were identified through electronic patient records systems at the surgeries. Audit criteria for identified patients were checked using electronic records, longhand records and personal enquiry of the GP if necessary.
For the subsequent intervention study, GPs were asked for permission to invite patients identified as failing any audit criteria for an appointment with AM and CS. Participating surgeries were provided with the text of an invitation letter, to be printed out on surgery notepaper and sent to eligible patients by practice staff: this preserved patient anonymity. GPs were offered advice regarding their patients who failed to respond or refused to be seen. Patients agreeing to a review were notified to CS, who subsequently attended the surgery to examine their notes and summarized their history prior to an appointment with AM and CS. Patients were seen at the surgery or, if they preferred, at their home. When seen, patients were asked to provide written consent for AM and CS to administer ratings of symptoms, side effects, general function and quality of life. The current medication regime and the patients' general mental health and well-being were then discussed. Proposed changes in medication, if any, were shared with the patient, and written advice on those agreed was given. Patients were informed that a follow-up appointment would be sent to assess progress once the changes had been implemented. The GP was informed in writing of the evaluation, and asked to implement the recommendations regarding medication changes.
Symptoms were rated with the Brief Psychiatric Rating Scale (BPRS), which identifies a broad spectrum of psychopathology across diagnostic groupings [20]. Antipsychotic side effects were measured with the Abnormal Involuntary Movements Scale (AIMS) [2] which assesses Parkinsonism, dyskinesia and akathisia. Side effects were also enquired about in general terms with each patient. General function was assessed using the Global Assessment of Function (GAF) [9] and the Clinical Global Impression (CGI) [1]. Quality of life was measured with the Quality of Life Self-Assessment Scale (QLSAS) [24]. Basic demographic and clinical data were collected: age, sex and clinical diagnosis from GP notes and the interview. At follow-up after a clinically appropriate period, patients' general mental health was reviewed and their medication noted: the rating scales were repeated. Non-parametric Wilcoxon signed ranks tests were carried out in order to ascertain whether changes in medication were associated with any significant changes in rating scale scores.
Results
Almost 53,000 general practice patients were screened by the community pharmacists: 1% were prescribed antipsychotic drugs. The most frequent reasons for audit criterion failure were psychotropic polypharmacy and chronic anticholinergic treatment. However, community pharmacists reported insurmountable difficulty in establishing the diagnosis of patients prescribed antipsychotic drugs by their GPs even when case notes were scrutinized and personal enquiries made of the GPs. This criterion therefore had to be abandoned as the majority of those prescribed antipsychotic treatment would have failed it. Similar caveats applied to the criteria regarding unresolved symptoms and side effects: no figures were returned, although all these criteria were examined in patients presenting for the subsequent intervention. Excluding unfeasible criteria, overall 280 i.e. just over half the patients were being prescribed regimes of medications which failed one or more audit criteria.
A minority of practices accepted the opportunity for review of their patients. 179 invitations to patients were sent: only 74 replies were received. We were informed later that 13 of the patients resided in a single nursing home: none replied. 54 patients accepted an appointment to be seen: 14 failed to attend, leaving 40 patients who underwent at least an initial evaluation. This represented only 23% of the number eligible for a review, whose GPs had agreed to their being approached.
The mean age of the patients was 59 years, with a range of 62 years: the oldest patient was 95 and the youngest patient was 33. 15 patients were men and 25 were women: there were no significant sex differences in age or any rating scale scores either initially or at follow-up. The diagnoses of these patients indicated that most were being prescribed antipsychotic medication off license. Clinically the diagnoses included 12 patients with uni-polar depression, 8 with learning disability, 6 with schizophrenia, 4 with anxiety or panic disorder and 3 with vertigo: 1 each dementia, personality disorder, bipolar disorder, alcohol dependence, obsessive-compulsive disorder and restless legs. In one patient no formal diagnosis could be arrived at even after careful scrutiny of her history and two personal interviews with both the patient and her mother.
32 patients were still taking antipsychotic treatment when seen. Figure 1 demonstrates the pattern of failure of audit criteria of these patients: our investigations revealed that 5 of the 29 with no formal psychotic disorder diagnosis did in fact have convincing evidence of psychotic symptoms either previously or currently. All the patients on more than one antipsychotic drug were diagnosed with schizophrenia. The mean number of criteria failed per patient was 3.4, with a range of 1–6: the standard deviation was 1.2.
Figure 1 Failure of Audit Criteria.
Only 8 (25%) of the patients were prescribed atypical drugs, the rest were prescribed conventional treatment. Clinical actions were recommended for 26 out of the 32 patients remaining on antipsychotic treatment at the time of interview. In half of the patients [15], stopping antipsychotic treatment altogether was advised. All 11 patients taking anticholinergic drugs on a chronic basis were advised to cease them. Other psychotropic drugs were suggested to be discontinued in 8 patients, some of whom had already stopped antipsychotic treatment before the first interview. In only 5 of the 32 patients was an atypical antipsychotic treatment recommended instead of existing conventional treatment.
Rating scale scores demonstrated that the 26 patients whose prescribing required amendment were minimally or mildly symptomatic for the most part. However they had a significant burden of motor side effects, and their function was far from optimal (see Table 3). Patients experienced great difficulty in filling in the QLSAS: this scale comprises a comprehensive list pertaining to life in general e.g. utilities, housing, access to leisure etc. Patients were asked to mark items with which they were not satisfied. Patients did not appear to relate well to the items as stated, and frequently tried to mark all which were satisfactory, becoming confused when directed not to. This difficulty was not compensated for by the QLSAS's freedom from mood and side effect items, and its use had to be dispensed with.
Table 3 Rating scale scores at each interview, changes and significance over time
Initial interview: n = 26 Follow-up interview: n = 24 p
Mean score range Sd Mean score range sd
BPRS 7 1–17 4 5 0–12 3 0.006
CGI 2.8 0–6 1.4 2.6 1–6 1.3 ns
AIMS 10 0–39 10 7 0–27 7 0.001
GAS 61 10–93 22 66 15–95 20 ns
BPRS – Brief Psychiatric Rating Scale
CGI – Clinical Global Impression
AIMS – Abnormal Involuntary Movements Scale
GAS – Global Assessment Schedule
BPRS symptoms scores at second interview had improved to a statistically but not clinically significant degree. AIMS side effects scores had reduced significantly: CGI and GAS scores were improved, but this was not statistically significant. Although all 24 patients who attended follow-up were included in the analysis, a third, i.e. 8 patients had not altered their medication as advised, either partially or at all. 4 patients unfortunately felt worse on their new regimes than previously, and had reverted to their former prescriptions. These included 2 patients with definite and unpleasant conventional antipsychotic withdrawal syndromes. One patient decided herself not to make the changes after considering what had been advised: the CPN of another patient appeared to be the deciding factor in the continuation of the patient's suboptimal treatment, citing the consequences of relapse. In two patients the GP and consultant failed to alter the prescription for reasons of oversight.
Illustrative cases
1 A 63 year old man with a 20 year history of chronic depression subsequent to a road accident (which caused several hours' loss of consciousness) and a one year history of epilepsy. He had mild depressive symptoms but no psychotic symptoms at any point. He was taking 75 mg chlorpromazine, 10 mg amitriptyline (originally for headache) and 5 mg nitrazepam daily. He had a marked tremor and complained of restlessness. He was advised to reduce and stop the chlorpromazine over six weeks on the grounds of tremor, probable akathisia, depressogenic and theoretical epileptogenic effects. When reviewed three months later, the patient described a severe exacerbation of restlessness, feeling hot, cold and sweating during his dosage reduction, to the point where his GP was obliged to reinstate the original dose. The patient's GP had substituted citalopram 20 mg for the 10 mg amitriptyline at our suggestion. The patient reported feeling more relaxed on this regime and furthermore had been able to stop using codeine for his headaches and laxatives for his previous constipation subsequent to codeine.
2 A 73 year old lady with diagnoses of mild learning difficulties and bipolar affective disorder, stable for the last three years and living in a nursing home. She was taking carbamazepine 300 mg bd, risperidone 2 mg bd, paroxetine 20 mg bd, and thyroxine. She was grossly obese with a BMI of 40, suffered from osteoarthritis and walked with a Zimmer frame. She also suffered from Parkinsonism and osteoporosis. We advised gradual alterations culminating in valproate semi-sodium as mood stabilising mono-therapy only. The grounds for this were the lack of tolerability and poor efficacy of carbamazepine compared to valproate semi-sodium, its induction of enzymes reducing antipsychotic levels, the mutually antagonistic effects of risperidone and paroxetine on mood, the side effects of Parkinsonism of both risperidone and paroxetine, and the side effects of hyperprolactinaemia, which can exacerbate osteoporosis, and weight gain of risperidone. At review two months later, no changes of any kind had been implemented. Following discussions amongst the treating team it was decided "the community nurse thinks there should be no changes to her medication as over the last 3–4 years she has been stable...she is 73 years old and not a young woman"
3 A 59 year old man with bipolar affective disorder and a recent TIA, taking 700 mg lithium daily (level 0.9), chlorpromazine 300 mg daily and 10 mg procyclidine daily. He complained of anxiety symptoms, restlessness and a tremor of several months' duration. He was advised to reduce and stop his chlorpromazine and procyclidine over a three month period, and reduce the dose of lithium to 600 mg daily. At review the patient had successfully stopped these medications and his tremor was much reduced. His GP had started a small dose of buspirone, and his anxiety and general mood were much improved. He was much more socially active and was attending further education.
4 A 55 year old lady, the wife of patient 3 above. Her GP referred her with addiction to sleeping tablets and mentioned that she stayed in bed all day. She was taking chlorpromazine 300 mg, stopped two weeks before being seen by ourselves, as she had begun to complain of worsening tremor, but when seen was still taking procyclidine 10 mg daily. At interview the patient gave a four year history of chronic anxiety and depression previous to which she had probably been dependent on alcohol for 11 years, consuming 70 units per week. Her depressive symptoms approached psychotic intensity and in addition she had orofacial dyskinesia. She was advised to stop procyclidine and to commence venlafaxine up to 225 mg daily: she had failed to respond to SSRIs previously. When seen four months later, the patient's husband said she was like a different woman: her depression had almost completely resolved, she was attending further education and had managed to give up smoking. Her GP had added a small dose of buspirone to her venlafaxine. She had successfully stopped her procyclidine and had no orofacial dyskinesia.
5 A 34 year old man with schizophrenia taking 10 mg risperidone daily i.e. greater than the recommended dose, and fluoxetine 20 mg daily: no indication for fluoxetine could be established. The patient had a BMI of 36 alongside poorly controlled positive symptoms, negative symptoms which had led to his losing his employment, and side effects of restlessness, gastrointestinal disturbance, blurred vision and abnormal involuntary movements alongside marked weight gain. The patient was advised to stop fluoxetine which was thought to be exacerbating his positive symptoms and abnormal movements, and responsible for gastro-intestinal disturbance. He was advised to substitute amisulpride at the low dose of 300 mg daily for the large dose of risperidone: this drug is associated with very little weight gain and is very effective for negative symptoms at such low doses, while maintaining efficacy for positive symptoms. At review the patient was taking 200 mg of amisulpride daily: his positive and negative symptoms were much improved, he was much more active and no longer complained of abnormal movements or gastro-intestinal disturbance. In addition, he reported much better memory and concentration.
6 A 73 year old lady taking venlafaxine 150 mg daily and 5 mg olanzapine at night. She had a history of recurrent depression but never any psychotic symptoms. Three years previously a consultant psychiatrist had advised reduction of the antipsychotic drug but this had not been implemented. The patient was not depressed at all but complained of having gained at least 7 lb weight on olanzapine: her BMI was 26. She was advised to stop this drug. At review four months later, the patient had stopped the olanzapine successfully: she had lost 7 lb in weight, and her BMI was 24. Furthermore the patient felt her energy levels were significantly improved, with less sedation and more capacity for physical activity.
Discussion
Antipsychotic prescription is not rare in primary care patients: furthermore in this study over half was potentially problematic in terms of accepted prescribing standards, leaving aside the lack of diagnostic justification available in GP records. The situation in secondary care has been investigated using suboptimal prescribing criteria not dissimilar to our own [19]. It was found that nearly 46% of regimes were suboptimal: greater consultant contact was associated with better prescribing practice. These authors concluded that prescribing practices in real-world settings frequently deviated from evidence-based guidelines. We would add that this deviation may be substantially more extensive in primary and general secondary care compared to specialist secondary care, and would tentatively assume that the lack of consultant psychiatrist input may be a factor here. For instance another primary care audit of 170 patients prescribed atypical antipsychotics drugs found nearly all were subject to psychotropic polypharmacy, over a third had no licensed indication, 30% were over 75 years old, only half were monitored six monthly or more: half had not seen a consultant [6]. A population based observational study in primary care demonstrated a 16% increase in the use of antipsychotic drugs over a decade [14]. More than half of all first-time use was for depression, panic and anxiety disorder with less than 10% for psychosis: thioridazine, which was virtually withdrawn shortly after this study ended in 2000, was most commonly prescribed throughout.
Further research on atypical antipsychotic drug prescribing trends in primary care found a six-fold increase in five years in the West Midlands: rates of use varied three-fold within the region even when local population need was accounted for. In generalist secondary care medicine in Germany, it has been shown that a minority of prescriptions for antipsychotic drugs were for indications of psychosis, over half were for patients age 65 or older, and only 40% were given by psychiatrists: the rates of prescription had risen in parallel with a decrease in prescribing benzodiazepines [16].
A recent study of one sixth of the population of Italy reported that one in 50 elderly people were prescribed antipsychotic drugs during a single year, two-thirds being conventional drugs [21]. In nursing homes in the USA, 27.6% of residents were given antipsychotic treatment in 2000–2001: less than half received treatment following appropriate guidelines, and its effectiveness did not differ whether guidelines were followed or not [5]. A further German study [11] demonstrated that 6% of a population of 25 million were prescribed antipsychotic drugs at least once within a two and a half year period: again, most prescriptions were for conventional antipsychotic drugs, written by non-specialists. These authors expressed concern regarding the high frequency of psychotropic polypharmacy, and co-prescription of cardiovascular and metabolic treatments. Some of the atypical antipsychotic drugs may be particularly associated with cardiac and metabolic side effects.
A French utilization study has confirmed high rates of psychotropic polypharmacy alongside antipsychotic treatment, despite lack of evidence for the efficacy of such combinations [17]. By contrast, a study of private psychiatric practice in Switzerland demonstrated strong adherence to international guidelines, with low use of antipsychotic polypharmacy and psychotropic comedication [23].
There is no shortage of material advising against the practices which we, and others in the field, have encountered. Patients without schizophrenia and the elderly may be particularly liable to serious side effects of antipsychotic drugs [8]. Antipsychotic polypharmacy cannot be generally recommended, even in schizophrenia, because of lack of efficacy [13]: furthermore, it is associated with greater use of anticholinergic and benzodiazepine drugs [12]. Unlicensed prescribing of antipsychotics in dementia is not recommended: their use is associated with a threefold increased risk of serious cerebral cerebrovascular events [7]. It has been known for many years that non-psychotic subjects acutely exposed to conventional antipsychotic drugs may suffer persistent adverse effects, including dysphoria (subjectively unpleasant mood) for several weeks [15].
The patients in our study were not particularly symptomatic but were middle aged/elderly, and had a significant burden of motor side effects. Our patients' experience of worsening of symptoms and antipsychotic withdrawal syndromes is of particular concern. Re-emergence of symptoms for which the drug was originally prescribed has been described previously in a learning disabled population who discontinued thioridazine [18]. Deterioration was associated with longer period of treatment, and occurred regardless of whether the thioridazine was replaced with another antipsychotic drug.
More recent work has highlighted the gap between guidelines and utilization in the real world [26]. Economic and social conditions, specifically rapidly increasing economic growth, may be associated with rapidly increasing drug consumption [10]. If psychotropic medications are being prescribed for symptoms such as depression, insomnia and anxiety, which can be attributed as much or more to social and personal problems rather than genuine illness, doctors are in effect providing a medical solution where none is indicated. This excessive reliance on pharmacotherapy may bring with it irrational combinations of drugs in inadequate doses for long periods: clearly contrary to the principles of rational evidence based therapy.
Our limited results suggest stopping redundant antipsychotics reduces side effect burden. However, getting these patients seen, and implementing change, is very difficult indeed and not entirely without risk to patients' wellbeing. The obvious conclusion to be drawn is that the prescription of antipsychotic drugs, particularly in the long term, should be avoided in patients in whom these drugs are not indicated, or in whom benefits are likely to be marginal.
Table 1 Patients prescribed antipsychotic drugs in 7 practices in East Hull: failure of individual audit criteria
Total number of patients audited 52885 %
Patients prescribed antipsychotic drugs 534 1.01
Prescribed droperidol 0 0
Prescribed thioridazine 18 3.8
>1 antipsychotic 32 5.9
Antipsychotic + other psychotropic 172 32.2
No positive diagnosis 46 8.6
Anticholinergic drugs >3 months 64 12.0
Not reviewed within 12 months 16 3.0
Table 2 Patients failing single or multiple audit criteria: 'no positive diagnosis' excluded
Failing one criterion 39%
Failing two criteria 11%
Failing three criteria 4%
Failing 4 or more criteria 0.2%
Total failing one or more 52.4%
Acknowledgements
We would like to thank Eastern Hull PCT, the pharmacists, practice staff and the GPs for their ongoing help and support. We would also like to thank the 40 patients who attended the intervention interview for their co-operation.
The following organizations gave financial support for this research;
• Hull & East Riding Community NHS Trust
• Sanofi Synthelabo
• Astra-Zeneca
• Janssen-Cilag
==== Refs
Tyson PJ Mortimer AM Wheeler JA High-dose antipsychotic treatment in clinical practice: a review, audit and survey of consultant psychiatrist opinions Psychiatric Bulletin 1999 23 661 664
Barnett D Guidance on the use of newer (atypical) antipsychotic drugs for the treatment of schizophrenia National Institute for Clinical Excellence Technology Appraisal Guidance 2002 London
Reilly JG Avis S Ferrier IN al QTc-interval abnormalities and psychotropic drug therapy in psychiatric patients Lancet 2000 355 1048 1052 10744090 10.1016/S0140-6736(00)02035-3
Thioridazine British National Formulary 2005 British Medical Association; Royal Phamaceutical Society of Great Britain 186
Organisation WH Prophylactic use of anticholinergics in patients on long-term neuroleptic treatment: a consensus statement British Journal of Psychiatry 1990 156 412 1971768
Overall JE Gorham DR The brief psychiatric rating scale Psychological Report 1962 10 799 812
Guy W ECDEU Assessment Manual for Psychopharmacology 1976 Revised Washington DC, US Department of Health, Education and Welfare
Frances A Pincus HA The global assessment of functional scale (GAF) Diagnostic and Statistical Manual of Mental Disorder, Version IV 1994 Washington DC, American Psychiatric Association
Guy W and Bonato RR National Institute of Mental Health, CGI: clinical global impressions Manual for the ECDEU assessment battery 1970 2 revised Chevy Chase 12.1 12.6
Skantze K Malm U A new approach to facilitation of working alliances based on patient's quality of life goals Nordic Journal of Psychiatry 1994 48 37 56
King DJ Burke M Lucas RA Antipsychotic drug-induced dysphoria British Journal of Psychiatry 1995 167 480 482 8829716
Meagher D Moran M Sub-optimal prescribing in an adult community mental health service: prevalence and determinants Psychiatric Bulletin 2003 27 266-270 10.1192/pb.27.7.266
Brzezicki A Amin R Use of atypical antipsychotics in Croydon PCGs Progress in Neurology and Psychiatry 2002 20 29
Kaye JA Bradbury BD Jick H Changes in antipsychotic drug prescribing by general practioners in the United Kingdom from 1991 to 2000: a population-based observational study British Journal of Clinical Pharmacology 2003 56 569 575 14651732 10.1046/j.1365-2125.2003.01905.x
Linden M Thiels C Epidemiology of prescriptions for neuroleptic drugs: tranquillizers rather than antipsychotics Pharmacopsychiatry 2001 34 150 154 11518477 10.1055/s-2001-15880
Percudani D Barbui C Fortino I Petrovich L Antipsychotic drug prescribing in the elderly is cause for concern International Clinical Psychopharmacology 2004 19 347 350 15486521 10.1097/00004850-200411000-00006
Briesacher BA Limcangco MR Simoni-Wastila L al The Quality of Antipsychotic Drug Prescribing in Nursing Homes Archives of International Medicine 2005 165 1280 1285 10.1001/archinte.165.11.1280
Hamann J Ruppert A Auby P al Antipsychotic prescribing patterns in Germany: a retrospective analysis using a large outpatient prescription database International Clinical Psychopharmacology 2003 18 237 242 12817158 10.1097/00004850-200307000-00006
Martin K Begaud B Verdoux H al Patterns of risperidone prescription: a utilization study in south-west France Acta Psychiatrica Scandinavica 2004 109 202 206 14984392 10.1046/j.0001-690X.2003.00238.x
Simon AE Peter M Hess L Valterio C Antipsychotic use in patients with schizophrenia treated in private psychiatry Swiss Medical Weekly 2005 135 109 115 15832227
Casey DE Side effect profiles of new antipsychotic agents Journal of Clinical Psychiatry 1996 57 40 45 8941170
Karow A Lambert M Polypharmacy in treatment with psychotropic drugs: the underestimated phenomemon Current Opinion in Psychiatry 2003 16 713 718 10.1097/00001504-200311000-00019
Johnsen E Svingen GF Jorgensen HA Practice regarding antipsychotic therapy: a cross-sectional survey in two Norwegian hospitals Nordic Journal of Psychiatry 2004 58 313 317 15370781 10.1080/08039480410005837
Bullock R Treatment of behavioural and psychiatric symptoms in dementia: implications of recent safety warnings Current Medical Research & Opinion 2005 21 1 10 15881470 10.1185/030079904X16777
Matthews T Weston SN Experience of thioridazine use before and after the Committee on Safety of Medicines warning Psychiatric Bulletin 2003 27 87 89 10.1192/pb.27.3.87
Verdoux H Begaud B Pharmaco-epidemiology: what do (and don't) we know about utilisation and impact of psychotropic medications in real-life conditons? British Journal of Psychiatry 2004 185 93 94 15286057 10.1192/bjp.185.2.93
Ghodse H Rational prescribing of psychotropic medicines International Psychiatry 2004 5 1 2
|
16316473
|
PMC1318453
|
CC BY
|
2021-01-04 16:39:07
|
no
|
Ann Gen Psychiatry. 2005 Nov 29; 4:18
|
utf-8
|
Ann Gen Psychiatry
| 2,005 |
10.1186/1744-859X-4-18
|
oa_comm
|
==== Front
Health Res Policy SystHealth Research Policy and Systems1478-4505BioMed Central London 1478-4505-3-81632116710.1186/1478-4505-3-8ResearchAnalysis of adequacy levels for human resources improvement within primary health care framework in Africa Parent Florence [email protected] Audrey [email protected] Yves [email protected] Colette [email protected] Dominique [email protected] Michèle [email protected] Danielle [email protected]êque Alain [email protected] Ketele Jean-Marie [email protected] Department of Epidemiology and Health Promotion, School of Public Health, Université Libre de Bruxelles (ULB), Brussels, Belgium2 UMR PRODIG, Paris, France3 Institut Supérieur d'Enseignement Infirmier (ISEI), Brussels, Belgium4 Centre de Pédagogie Universitaire, Université Catholique de Mons (FUCAM), Belgium5 Education Department, Université Catholique de Louvain (UCL), Belgium2005 2 12 2005 3 8 8 30 5 2005 2 12 2005 Copyright © 2005 Florence et al; licensee BioMed Central Ltd.2005Florence 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.
Human resources in health care system in sub-Saharan Africa are generally picturing a lack of adequacy between expected skills from the professionals and health care needs expressed by the populations. It is, however, possible to analyse these various lacks of adequacy related to human resource management and their determinants to enhance the effectiveness of the health care system. From two projects focused on nurse professionals within the health care system in Central Africa, we present an analytic grid for adequacy levels looking into the following aspects:
- adequacy between skills-based profiles for health system professionals, quality of care and service delivery (health care system /medical standards), needs and expectations from the populations,
- adequacy between allocation of health system professionals, quality of care and services delivered (health care system /medical standards), needs and expectations from the populations,
- adequacy between human resource management within health care system and medical standards,
- adequacy between human resource management within education/teaching/training and needs from health care system and education sectors,
- adequacy between basic and on-going education and realities of tasks expected and implemented by different categories of professionals within the health care system body,
- adequacy between intentions for initial and on-going trainings and teaching programs in health sciences for trainers (teachers/supervisors/health care system professionals/ directors (teaching managers) of schools...).
This tool is necessary for decision-makers as well as for health care system professionals who share common objectives for changes at each level of intervention within the health system. Setting this adequacy implies interdisciplinary and participative approaches for concerned actors in order to provide an overall vision of a more broaden system than health district, small island with self-rationality, and in which they operate.
==== Body
Introduction
The organization of health systems in sub-Saharan Africa is, more than elsewhere, born in close connection with the establishment of political and territorial structures, initially within the colonial framework, then within the building of independent States [1,2]. Since 1980s, the economic and financial crisis of several States is marked by their disinvestments in the development and planning of programs associated with the promotion decentralization models [3]. Inspired by WHO taken over by the World Bank in 1990s, the systems of care are organized at the base, within and around Health districts liased with "minimum and complementary packages for care" to provide better answers to populations' requests [4]. Within this framework, many questions persist concerning improvement of medical systems, focused mainly if not exclusively, on financial and organizational techniques. In view of these persistent inefficiencies, priorities for medical actions in sub-Saharan Africa usually reinforce mechanistic approaches where an overall vision of the whole situation is eluded in favour of an approach covering separately different operational sectors in the fields of planning, training, implementation and evaluation.
The management of human resources in health more often participates to the sustainability of an "inhospitable medicine" in Africa [5]. It is however a recent stake in the rich as well as in the poor countries [6]. Within health systems, it represents health care implementation. It questions practices, their findings and efficiencies in the heart of interactions between various actors concerned: professionals and populations. Since 1990s, it remains an object of increasing concern of works and thoughts on improvement of the effectiveness of health care structures [7]. The case studies are becoming more sensitive on human resources and their management in the health systems, especially in Africa [8,9] where, more than elsewhere, the quality of care seems to be lacking, in the actions as well as in the perceptions from the populations who have poorly recourse to it [10,11].
In sub-Saharan Africa, actors and observers agree in recognizing the discordances and inefficiency of health care practices while intensifying many national programs for building human resources' capacity. These projects, however, rarely adopt a global approach to needs and roles for health care professionals relocated in all sectors of interventions such as medical, educational or planning. In this article, we underline the need to consider health care professionals in their interactions with all the actors in the health environment in one area at a given time. These stakes are translated early in the 21st Century by a redefinition of the organization of health districts around their human resources. This public health objective is at the core basis of projects and action plans aiming particularly at a better adequacy between on the one hand the offer of on-going training (specific and on-the-job training) or initial (initial training curriculum), and, on the other hand, expectations from health care professionals as well as the needs for the populations.
The approach on organization of health care systems in Africa by its human resources management supposes the assertion of new frameworks of analysis and extended action to political, sociological, educational and motivational dimensions. Discussion on human resources management appears then in all the complexity of its multi-factorial dimensions using not only the health objects for which these resources exist, but also areas of training, planning and sociological determinants related to actors' behaviours. If approaches in more systemic terms of health districts and their actors have already been initiated [12], the consideration of links between these various dimensions still remains insufficient.
The article defines in an original way these multi-factorial and multi-level links corresponding to levels of adequacies or organizational inadequacies, determinants from the coherence and effectiveness of health care systems in sub-Saharan Africa. It recommends a new framework of analysis and understanding of these forms of (in)adequacy on human resources' management in relation to expected competences from professionals and the needs for the populations. Planning this conceptual framework based on field works in Democratic Republic of Congo (DRC) and in Rwanda plays a role in improving medical systems, in general, and structuring primary health care in particular.
Methods
Development of the conceptual framework
Thinking on human resources in the health sector in sub-Saharan Africa leads to develop adapted tools. The first step consists in context-oriented human resources and their management within the entire health system and its actors, internally as well as externally. As recent works remind it, for the comprehension of the social world, to extract any element from its context is eminently dangerous. In fact we face a health system in which it becomes essential to replace all interrogations about the place and role of the professionals [11]. The human resource management should be relocated within the entire aspects of the organizational modalities and improvement of medical systems (States' policies to staffs and populations of therapeutic practices and their initial trainings) to provide answers to recurring interrogations which raise concern and difficulties to obtain quantitative and qualitative adequacies for health care professionals with the needs of objective and perceived care for the populations. This questioning is not added to many explanations on dysfunctions of the African medical systems but opens way to the formulation of objectives for changes aiming at a better effectiveness in the health care system.
The search for a greater effectiveness regarding nursing practices and their recourse opens to the second step. This one is the opposite to break with actions burst in sectors of technical performances and scattered in distinct skills. In the approach to improve medical systems, there is not question any more to dissociate spatio-temporal dimensions in liaison with sectors of operation separating, for example, the local and immediate levels of operative functions for health care structures, on the one hand, and levels including organization and decision of the health care systems, on the other hand. These approaches by sectorial activities produce changes, which can only intervene within the medical system, without modifying its structure and its functioning, nor the links and their effects between its elements.
These two steps are found within a systemic approach of human resources in the health care sector. The systemic vision, supported by a research-action inter-sectorial approach, puts in perspective human resources in its interactions with all components of the medical systems. It opens a way towards a representation of changes, which suppose the overall progress of the system, to which they apply. The aim is the development of a new model of knowledge from the two human resources capacity building projects in central Africa concerning the field of health:
The first project, initiated three years ago in the Democratic Republic of Congo (DRC), recommends to support initial teaching in health sciences in the secondary level of education at the national level. In a first stage, the project develops coherence, relevance and understanding for a significant number of actors and stakeholders of strategic orientations, founders of the reform of the nurse sciences program required by the Department of Health Sciences Education within the Ministry of Health. In 2005, this reform is on the way. The whole process enables autonomy of teachers, as well as of learners, managers, department staff and supervisors.... It is by a methodological work calling upon concepts such as active pedagogies, skilled-based approach [13,14], organizational learning, thinking and self-assessment, built by partnership and interrelationships between all the actors together, giving greater importance to improving health care quality and their perceptions within the framework of Structures for primary health care.
The second project in progress is a national support to schools of nurse sciences in Rwanda. The steps and methods are similar to those launched in DRC. The interest carried to human resources passes by a second reading of the training package related to health sciences (professional, higher and academic levels). The search for a better adequacy between trainings and health professional expectations as well as those of the public regarding care, underlines the necessity to train nurses in technical secondary level on the skill-based approach. The project is also involved in an in-depth work with the various local and national actors: teachers, internship supervisors, directors of educational establishments, learners, and Ministries. It articulates, indeed, several organizational and institutional levels: local learning environments complying with medical standards, human resource requirements planning and training schemes.
When projects for general thinking are located and specific to a category of professionals (nurses), actors and fields are relocated in the entire medical environment including the populations, social and political supervisors. The stake is not just the detailed observation of actors and their relations with the health care systems, but to go in fine beyond traditional explanatory models of health care (dis)functioning in Africa focused on districts. The different sites contribute by developing an analysis framework on more complex realities than simple setting in linear equation between, on the one hand, the medical standards planned by national institutions, and, on the other hand, the assessment of local requirements in human resources without integration neither for their training modalities nor for the expectations expressed by the populations.
A grid of analysis is suggested where human resource management, including for nurses of primary health care structures, falls under the overall medical system and the diversity of its political actors, health care professionals or not. These components are considered within their dynamic interactions, as much undergone as built. It makes it possible to avoid separating artificially human resource management, perspectives for planning, training education, and evaluation. Persistent dichotomies between spheres of health and education are checked through penalizing field-based discordances between professionals' skill profiles and their needs while meeting populations-expressed expectations. The perspectives for efficient changes of a health system assumes improving different adequacy levels which are to avoid reduced searching " for oasis of rationality" [12], limited to dimensions of each health district, and to the implementation of sectorial projects launched together in space and time.
Results
Presentation of the conceptual framework
The diagram constitutes a grid of analysis of levels of (in)adequacies in human resources management in health system with different components and organization methods for health care systems taking part in its (dis)functioning. The structure obtained is prompted by the articulation between levels and system organization sectors: States producing heath care system standards, with planning programs and health care system management and its professionals, and local actors and their training for care practice.
The elements are illustrated by the six "boxes" which define the major levels of adequacy between the management of the health care professionals and ways of improvements, in terms of quality of care and health system effectiveness. The arrows show links, i.e. the interdependences between these forms and levels of adequacy. The structure is provided by the overall composition of the diagram.
Elements, links and structure account for a "construction of health":
- Field of action of national policies inspired from international standards;
- Societies' stake and their Community participation;
- Object of local implementation of care.
These various levels of organization of health care system appear through three "sectors" shown in the diagram: from macro level developing qualitative and quantitative health care standards (the "Health sector" of the national and international policies concerned), to the meso, spaces and actors of the medical training registered in projects of companies like the "sectors of planning and of the human stock management" and, to the microphone of the daily practices of health whose forms of application depend directly on the "sector basic training" of the professionals.
These sectors show the importance of levels for observation, stakes and human resources analysis, as thoughts of social sciences in the field are concerned of articulation modalities between macro and microanalyses levels of standards and health care practices [15,16]. The articulation between forms fixing-up micro and macro levels is made to better apprehend political and social stakes of the health system. This critical reading tackles modalities for training and allocation for human resources within the health system and the forms of distribution of care needs. This diagram is not a simple picture structuring health systems. It allows also dialectical approach between the analysis of (in)adequacy levels of the human resource management in health system and of the inequalities in the heart of these inadequacies, that is to say, their differentiated distribution in the populations and spaces.
Lastly, the reading of the diagram can and should be done with flexibility, without giving an advantage to a particular input except actors, observers, and readers' experiences. This detailed presentation of adequacy levels does not answer to a specific order. On the other hand, for each one, this article underlines initially specificities before discussing them and opening on its links with other constitutive levels of adequacy of the overall structure. This somewhat formal rigidity is meeting concern to facilitate the legibility and understanding of selected adequacies but not to reify adequacy forms and their categories of actors in the heart of the adjustments' management between offer and needs in terms of human resources for the health systems. This management in sectors of training, planning or yet of accessibility accounts for the necessarily evolutionary structure of health systems as well as included elements and links.
Adequacy levels for human resource management in health
Figure 1 Grid of analysis for adequacy levels improving Human Resource Management in the field of Health.
Adequacy between the skills-based profiles of health professionals, quality for health care offers, services (health system standards), needs and expectations expressed by the populations
Within the development of health system schemes at the international level [17,18] the States and their Ministry for Health define their own health system standards according to local contexts (economical situation and structural records). These standards establish, in particular, relevant conditions of assigning activities between health care centres and the population, within the decentralized framework of health districts, and first referral hospitals located within or near the district. Without targeting a strict correspondence between population expectations and its needs, the qualitative adequacy between these two realities, established by health professionals and local epidemiological priorities, is essential to ensure an effective reference to health care structures. It necessitates accessibility being facilitated according to perception and acceptance by actors of the qualitative normative frameworks.
The standards should highlight measurements and needs expected from the population and health system professionals inclusively. Experts should be confronted with problems of priority health defined at the national level for populations in a given territory and time. These priorities can be defined according to practical orientations, in technical acts and expected activities of health professionals. Acts and resources can be specified according to categories of professionals, by developing skills-based profiles. The management and training for human resources are considered in the overall organization of the health system. It thus appears essential to check from experts and even from the population for the adjustment of health standards and skills-based profiles towards the reality of situations experienced and perceived by actors, in particular the public and primary health care professionals.
To question the adequacy of standards within the health environment with the reality of health problems encountered and perceived by the population, requires to meet actors during thorough qualitative surveys while making it possible to better determine their needs and the expression of their expectancies in order to integrate them in the development process of health standards. In Africa, these standards are still more often built without taking into account elements such as mental health or the role of tradi-therapeutics. The evolutions of the health standards should be adapted to problems experienced and felt by the populations (or defined through the health system as the new vertical programs). The qualitative adequacy of standards to the needs of societies is not, in fact, a static question but engages a progressive and continuous research/action. If health standards are a qualitative framework of reference, their definition and performances, open towards adequacy levels in more quantitative terms and whose implementation depends on available resources and appropriate needs.
Adequacy between the distribution of health system professionals, quality for health care offers, services (health system standards), needs and expectations expressed by the populations
Health standards quantitatively define modalities for human resources allocation within health districts in accordance with minimum packages and complementary activities. If situations are generally optimised by national health authorities, they will be conditioned by country capacities not only for human resources, but also for their management.
It is necessary to question not only the qualitative adequacy of health standards towards the needs expressed by populations, but also their quantitative adequacy according to available human resources in the sector. If standards are not adjusted to this reality, then it would be preferable to refer to more specific health references rather separate from real conditions of health care practice.
To favour these adequacies, complementary mechanisms should be considered like the development of professionals' skills-based profiles of the management. These mechanisms can in return validate or revise health standards.
Human resource management adequacy in health system versus health standards
Concretely, the question of the adequacy of qualitative and quantitative health standards with human resource management operates the passage of their definition according to an ideal situation with that of their definition as reflection of the national, provincial, regional, health district capacity according to the considered degree of decentralization.
This approach, first testing of field-based realities, requires a detailed inventory of human resources in health system. This thorough knowledge should be a priority in central Africa. Failing this it is difficult to set planning for human resource management. In certain situations, the setting-up of a Medical Association and Nurses Council will improve this knowledge and with it the adequacy required, since one of the first work of these organizations is its manpower census.
It is only from these records that progress can be made in terms of objectives, operations and health structures performances, in particular in terms of articulations between educational and health fields regarding the staff management (diplomas, ethical issues).
Adequacy between human resource management in the field of education/teaching/training and the needs expressed by health and education sectors
The planning of health sectors often considers exclusively its work in terms of management concerning its own human resources without any perspectives of coordination regarding human resource management in health and education system. This coordination however appears necessary insofar as the health and education sectors provide their contribution to the improvement of health care and services quality.
The adequacy between staff managements concerned with health and education remains essential indeed. It relates to learning modalities to reach the qualitative objectives for training and health care recommended by health standards. Thus a learning/teaching ratio in liaison with the teaching mechanism of the initial education should be referred to the existing resources in the field of education in order to adapt teaching structures to the real possibilities of the country and to consider the close links between trainings and health care quality regarding the staff management.
Initial and on-going trainings adequacy versus realities of the expected tasks and implemented by different categories of health system professionals within the health system organization
For a category of health professionals, the adequacy raises the issue of similarity of technical and professional education of the associated or technical level of education. This priority by category of actions is justified by the strong implication of these nursing personnel within the primary health care where they are in charge in certain contexts of more than 85% of the health care offer and services. Vis-à-vis to their important role in care practices, methodology in educational engineering defines a rather high and full profile of skills, that is to say, at the same time general practitioning and technical. The training adequacy to the implemented tasks for a category of health professionals, underlines the importance of health district as a body belonging to a coherent organization, as a first level of achievement within the field-based Human Resource Management, their inter-articulations, in particular according to categories of training and/or vocational identities which constitute the diversity of a health system.
The relocation of these adequacies by and between professional categories in the global vision of changes for the improvement of care and the effectiveness of health systems prevents professional conflicts of identities while supporting the inter-sectorial approaches. The on-the-job training should be finally adjusted with the realities on the field. The difficulty is then the planning of trainings, which has to be coherent between different professionals, and be integrated in the action plans of the health districts. The on-going training is part of vertical management programs with basic skills required by different health system professionals. The difference between the needs for care and professionals' skills are probably due to the belief by the teachers of a spontaneous transfer of knowledge in situations.
Adequacy of initial and on-going trainings' intentions versus teaching programs in health sciences for the trainers (teachers/supervisors/health professionals/Heads of establishments...)
In practice, one notes a lack of on-going training for teachers in many vertical programs. An aggravation of the dichotomy, established from the beginning by the inadequacy of teaching programs in health sciences, is proven by on-the-job trainings. It can be a question as well of updating vertical programs as actualising them on the concepts of primary health care and organization of health system.
Improving programs assumes reinforcing links between training objectives for future health-system professionals and teaching mechanisms in classroom as well as in the field of operation. Assessments and training plans for various levels of trainers are preconditions to any modification of program.
Beyond individual investments in training future health-system professionals, the systemic approach of actors opens a way towards an overall perspective of places and levels for trainings: from the hospital to the community, through health centres and the numerous administrative bodies. The sustainability of a teaching innovation necessitates some knowledge of adequacies by the team of trainers who ensure the follow-up and extension for changes adapted to the contexts of their achievement.
Discussion
An overall vision remains essential for a qualitative improvement of the health care and services' system. A participative and representative process of the diversity sectors at various levels of organization plays a role in better adequacies considered not only in their specificity but also in comparing the ones to the others between and within micro, meso and macro levels [19]. Among those actors at three organizational levels (Health – Education/Planning – Human Resource Management), the question is the supporting of a common vision for change. This common horizon should come with change and even precede it. It is necessary that decision-makers and health experts from various adequacies share a same objective for change beyond specific modalities of its establishment depending on categories of actors and their level of intervention within the health system. For the decision-makers, it is a question of validating health standards for planning and programs in the contexts and conditions of their implementation. For the experts, it is a question of implementing more these planning and programs standards as effectively as possible. These are carried out and observed during trainings, practices and recourse to health care. The experts take part in the development of health-system standards while the decision-makers integrate practical field-based knowledge in the modalities for change.
Beyond field-based surveys included in its construction, the model is carrying a vision for change of the health system in all its dimensions and likely to be adopted by all the actors who share the standards and participate to its dynamic. Admittedly, in the programs presented, the attention focuses on actors at the peripheral level, nurses and health-system professionals, who are in daily contact with the health care demand and offer. However, these actors are also apprehended by their registration, their articulations with the selections, and inclusive standards of the health care system.
It is less a question of identifying a presumably adequate level of improvement relating to health systems similar to what seems to be limited to health district, than to release an overall and contextualized perspective of requirements and stakes regarding this improvement. By not taking into account the human resource management from its economist or technicist dimensions, the approach by its adequacy levels within the health systems does not take for an operational level, or even of privileged observation. The adequacies are included within the interactions between the different spatio-temporal levels of health-system organization in sub-Saharan Africa.
These spacio-temporal articulations, especially between daily and localized health practices and higher levels of development of health-system standards as educational for the organization of health care systems and their personnel, were already explored in Africa in particular by health-system anthropologists and sociologists [16]. Along with these works and the operational dimension given to the research, the identification and improvement of adequacy levels and modalities for human resource management adopt a framework of analysis dimension or level of operation. Being of multi inter-sectorial levels, this flexible model is distinguished from approaches focused on "robustness" of health districts to propose dynamic research/actions. It is no more a question of proceeding by fragmentation, neither by (de)limiting fields of operation and thinking field level in health districts, nor by artificial distinction of fields for human resource management (educational, technical, economic, sociological, political dimensions...).
The framework of analysis is not a conceptual abstract tool or "self-sufficient" recommending to improve health care structures. The health district remains an essential framework because it is a localized territory and a tool of implementation for the health system. It is in the heart of relations between actors, their aspirations and health organization, especially training for staff. The stakes for a greater effectiveness of health systems in Africa are located inside these interrelationships and their adjustments. The setting of adequacy is not conformed to modalities of implementation based on transfer of models and health standards developed international and national levels. For the development of health systems, the adequacy requires the advisory and participative approaches of all related actors at various space and time scales. The levels of decision, operation, development or evaluation are neither opposite nor juxtaposed between them. The level of health districts is, however, reinforced but not rigidified or isolated. It is approached not like a health system asset in sub-Saharan Africa but as a production space for health practices in permanent construction.
Conclusion
Health like education is frequently presented as "everybody's business" in the 21st century societies. However, these two fields are often treated separately. The sub-Saharan African societies are representative of this dichotomy while, more than elsewhere, the poverty of the greatest population results in a poor adequacy and effectiveness between supply of care and its demand expressed by the populations. Improving primary heath care by the improvement of human resource management focuses, less on questions of means, decreased since the 1980 decade within the framework of the States of sub-Saharan Africa, than on links between their modalities of management, training and planning [20]. For years it was thought that human resources were only one aspect of dysfunctions of health districts, it is now evident that their modes of organization, in their technical, economical, educational, political, and sociological dimensions are the levers of real changes [21]. The merit for the framework of analysis proposed is to make available to various actors an overall perspective of the system to facilitate the emergence of a common vision for a more effective, socially and politically more coherent model. It recalls that behaviours of the actors, at the origin of the majority of health system dysfunctions are determined by factors amongst others educational, which are not only environmental or institutional. To act on human resources is therefore, to act on health system as a whole and thus to consider a real and sustainable impact.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
We thank all of the teachers of the medical technical institutes of Kinshasa who participated in the development of this research framework.
==== Refs
Hours B L'Harmattan L'Etat sorcier Santé publique et société au Cameroun 1985 Paris
Arnold D Bynum WF, Porter R Medicine and colonialism Companion Encyclopedia of the History of medicine 1994 London/New York, Routledge 1393 1416
Gruenais ME Communauté et Etats dans les systèmes de santé en Afrique Hours B, dir, Systèmes et politiques de santé De la santé publique à l'anthropologie 2001 Paris, Karthala
World Bank Better Health in Africa 1994 Washington DONC, The World Bank
Jaffre Y Prual A Le « corps des sages-femmes » entre identités professionnelles et sociales Sciences Sociales et Santé 1993 2 63 80
Gazier B Les stratégies des ressources humaines 1993 Paris, La Découverte – Syros
Aiach P Fassin D dir, Les métiers de la santé Enjeux de pouvoir et quête de légitimité 1994 Paris, Anthropos-Economica
Jaffre Y Olivier de Sardan JP Une médecine inhospitalière 2003 Paris, Karthala
Metboul M Une anthropologie de la proximité Les professionnels de la santé en Algérie 1994 Paris, L'Harmattan
Ciardi P Poloni A Kinda F Qualitative Survey of Utilisation of health Services, Participation and Health needs in Burkina Faso's rural Communities Final Report 1993 World Bank
Gobatto I Les médecins acteurs dans les systèmes de santé Une étude de cas au BurkinaFaso Hours B, dir, Systèmes et politiques de santé De la santé publique à l'anthropologie 2001 Paris, Karthala
Grosdos D Le district sanitaire urbain en Afrique subsaharienne Enjeux, pratiques et politiques 2004 Paris, Karthala – UCL
Roegiers X Une pédagogie de l'intégration : compétences et intégration des acquis dans 2000 l'enseignement. Paris-Bruxelles: De Boeck Université
De Ketele JM Chastrette M Cros D Mettelin P Thomas J Guide du formateur 2001 De Boeck Université. Belgium
Reval J Jeux d'échelles La micro-analyse à l'expérience 1996 Paris, Gallimard, le Seuil
Gobatto I Les Pratiques de santé dans un monde globalisé 2003 Paris, Karthala
Falola T Ityavyar D The political economy of Health in Africa 1992 Athens, Ohio University
Fassin D L'Internationalisation de la santé, entre culturalisme et universalisme Esprit 1997 229 83 105
Parent F Kahombo G Bapitani J Garant M Coppieters Y Levêque A Piette D A model for analysis, systemic planning and strategic synthesis for health science teaching in the Democratic Republic of the Congo: a vision for action Human Resources for Health 2004 2 16 15585060 10.1186/1478-4491-2-16
Gonzalez-Block MA Health policy and systems research agendas in developing countries Health Research Policy and Systems 2004 2 6 15296509 10.1186/1478-4505-2-6
Bonani M Garant M De Boeck Systèmes scolaires et pilotage de l'innovation. Emergence et implantation du changement 1996 243 Paris
|
16321167
|
PMC1318454
|
CC BY
|
2021-01-04 16:37:16
|
no
|
Health Res Policy Syst. 2005 Dec 2; 3:8
|
utf-8
|
Health Res Policy Syst
| 2,005 |
10.1186/1478-4505-3-8
|
oa_comm
|
==== Front
BMC NephrolBMC Nephrology1471-2369BioMed Central London 1471-2369-6-141632975810.1186/1471-2369-6-14DatabaseThe IgA nephropathy Biobank. An important starting point for the genetic dissection of a complex trait Schena Francesco P [email protected] Giuseppina [email protected] Diletta D [email protected] Francesco [email protected] Marina [email protected] Antonio [email protected] Doroti [email protected] Jürgen [email protected] Peter R [email protected] Klaus [email protected] Efstathios [email protected] Dimitrios [email protected] Leopoldo [email protected] Luigi [email protected] Gian M [email protected]à Giovanni M [email protected] European IgA nephropathy [email protected] Renal Unit, University of Bari, Italy2 Renal Unit, University of Brescia, Italy3 Dipartimento di Genetica, Biologia e Biochimica, University of Torino, Italy4 Dipartimento di Biochimica, Biofisica e Chimica delle Macromolecole, University of Trieste, Italy5 Division of Nephrology and Immunology, University of Aachen, Germany6 Institute of Human Genetics, University of Aachen, Germany7 Renal Unit, Aristotelian University of Thessaloniki, Greece8 Genetic Unit, IRCCS-Casa Sollievo della Sofferenza, San Giovanni Rotondo, Foggia, Italy9 Renal Unit, Ospedale Gaslini, Genova, Italy10 Renal Unit, University of Bologna, Italy2005 5 12 2005 6 14 14 7 7 2005 5 12 2005 Copyright © 2005 Schena 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
IgA nephropathy (IgAN) or Berger's disease, is the most common glomerulonephritis in the world diagnosed in renal biopsied patients. The involvement of genetic factors in the pathogenesis of the IgAN is evidenced by ethnic and geographic variations in prevalence, familial clustering in isolated populations, familial aggregation and by the identification of a genetic linkage to locus IGAN1 mapped on 6q22–23. This study seems to imply a single major locus, but the hypothesis of multiple interacting loci or genetic heterogeneity cannot be ruled out. The organization of a multi-centre Biobank for the collection of biological samples and clinical data from IgAN patients and relatives is an important starting point for the identification of the disease susceptibility genes.
Description
The IgAN Consortium organized a Biobank, recruiting IgAN patients and relatives following a common protocol. A website was constructed to allow scientific information to be shared between partners and to divulge obtained data (URL: ). The electronic database, the core of the website includes data concerning the subjects enrolled. A search page gives open access to the database and allows groups of patients to be selected according to their clinical characteristics. DNA samples of IgAN patients and relatives belonging to 72 multiplex extended pedigrees were collected. Moreover, 159 trios (sons/daughters affected and healthy parents), 1068 patients with biopsy-proven IgAN and 1040 healthy subjects were included in the IgAN Consortium Biobank. Some valuable and statistically productive genetic studies have been launched within the 5th Framework Programme 1998–2002 of the European project No. QLG1-2000-00464 and preliminary data have been published in "Technology Marketplace" website: .
Conclusion
The first world IgAN Biobank with a readily accessible database has been constituted. The knowledge gained from the study of Mendelian diseases has shown that the genetic dissection of a complex trait is more powerful when combined linkage-based, association-based, and sequence-based approaches are performed. This Biobank continuously expanded contains a sample size of adequately matched IgAN patients and healthy subjects, extended multiplex pedigrees, parent-child trios, thus permitting the combined genetic approaches with collaborative studies.
==== Body
Background
IgA nephropathy (IgAN) or Berger's disease is the most common glomerulonephritis worldwide in patients undergoing renal biopsy. Diagnosis is based on the occurrence of mesangial deposits of IgA in the glomeruli in the presence of recurrent episodes of intra-infectious macroscopic hematuria or persistent microscopic hematuria and/or proteinuria. The frequency of this disease is renal biopsy policy-dependent [1]. In some countries, the majority of patients with recurrent macroscopic hematuria undergo renal biopsy only in the presence of proteinuria or mild renal insufficiency. In Eastern countries, however, many young subjects with persistent microscopic hematuria and mild proteinuria often receive renal biopsy. This is not the policy in most Western countries, in addition often asymptomatic subjects with persistent urinary abnormalities (microhematuria and/or proteinuria) decline renal biopsy. This large variability in symptoms induces different approaches in performing renal biopsy. Nevertheless, renal biopsy practice is not the sole cause of the different disease worldwide prevalence. Striking ethnic variation in prevalence [2-6] along with familial clustering [7-9] are suggestive of an important role of the genetic component in the pathogenesis of the disease. IgAN may occur in a sporadic or a familial form according to the clinical evidence that one or more subjects belonging to the same family are affected by biopsy-proven IgAN [10-12]. Moreover, sub-clinical renal abnormalities are often evidenced among relatives of IgAN patients [9,11,13].
Genetic analysis of large pedigrees of IgAN families is considered the most promising approach to identify IgAN susceptibility genes. The genome wide scanning in 30 large extended IgAN multiplex families (24 from Italy and 6 from United States) was performed by Gharavi et al [14]. This study identified the locus, called IGAN1, located on chromosome 6q22–23 in linkage with IgAN. It yielded a significant peak Lod-score of 5.6, with 60% of families linked, assuming an autosomal dominant mode of inheritance with reduced penetrance. This mode of inheritance of familial IgAN is more consistent with the involvement of a single gene with a large effect located in IGAN1. Nevertheless, multifactorial determination, with the interplay of many genes, each conferring a small effect cannot be excluded. The knowledge gained from the study of Mendelian diseases has shown that genetic dissection of a complex trait is more powerful when combined linkage-based, association-based, and sequence-based approaches are performed [15]. Association-based studies are possible when a large sample size of adequately matched IgAN patients and healthy subjects is collected. Prior insight into the pathogenesis of the disease allows to establish the candidate genes to be studied according to their known or suspected function. Over the past decade, a shift has occurred away from case-control association studies, towards family-based designs in which extended pedigrees, relative-pairs and parents-child trios are used to test for association. For this purpose, different powerful and sensitive methods of analysis have been developed, most of which are based on the trasmission/disequilibrium test (TDT) [16].
Considering the possibility of identifying IgAN susceptibility genes a group of European scientists, specialists in this disease, constituted an IgAN Consortium and organized a Biobank project recruiting biological samples and clinical data from IgAN patients and relatives. This report describes the creation of the first Biobank with a readily accessible database of the most common worldwide glomerulonephritis, a rare disease which requires a multi-centre project organization.
Construction and content
The IgAN Consortium was based on the constitution of a collaborative study group including expert nephrologists from Italy (FPS, FS, GMF), Germany (JF) and Greece (EA), geneticists from Italy (AA, LB, GC, GMG) and Germany (KZ). The European project No. QLG1-2000-00464 was funded by the 5th Framework Programme 1998–2002. Additional funds were obtained in Italy from MIUR (Ministero dell'Istruzione, Universita' e Ricerca PRIN 2001-067748; L.488/92 Cluster 03; FIRB 2001-RBNE013JYN).
The main aim of the project was to constitute a genomic DNA bank of well characterized IgAN patients and their relatives from three different European countries (Germany, Greece and Italy) to be used for genetic studies. The following objectives were reached:
1. a) Definition of a common protocol for the diagnosis of familial, suspected and sporadic IgAN. IgAN was diagnosed on the basis of either recurrent episodes of macroscopic hematuria concomitant with upper respiratory tract and other infections or permanent microscopic hematuria with/out proteinuria. Mesangial glomerular IgA deposits in the renal biopsy and absence of systemic or hepatic diseases confirmed the diagnosis. Relatives of at least three generations received urinalysis. All relatives with persistent microscopic hematuria received an Addis count under contrast phase microscopy (glomerular hematuria was confirmed at least three times) and renal ultrasound. Relatives with suspected IgAN were informed and received renal biopsy to confirm the diagnosis. All subjects were categorized as: (I) affected subjects with biopsy-proven disease; (II) subjects probably affected by IgAN when persistent microscopic hematuria without biopsy-proven glomerulonephritis, or chronic renal insufficiency or ESRD treated with hemo or peritoneal dialysis, or renal transplantation after biopsy-proven IgAN or unknown glomerulonephritis, or death caused by chronic uremia, occurred; (III) non-affected subjects in the absence of urinary abnormalities; (IV) subjects with unknown status, who refused participation or in the absence of clinical information and laboratory findings. Familial IgAN was diagnosed when at least 2 family members had biopsy-proven IgAN. Suspected IgAN families were considered those in which one subject was a biopsy-proven IgAN patient and others were probably affected, were awaiting or refused renal biopsy, or had urinary abnormalities that did not justify its execution. They were checked systematically once a year. Sporadic IgAN was diagnosed when the disease occurred only in the patient and relatives were negative by urinalysis. b) Layout of a booklet containing information, letter of informed consent and instructions for the collection of blood samples. c) Definition of a data sheet for the collection of personal data, clinical and laboratory findings.
2. Constitution of the Biobank. Blood samples from all the participants after informed consent (disclosing information about the research objectives, benefits and risks) were obtained. Thirty ml of peripheral blood was drawn from each subject aged more than 15 years and 15 ml from each younger subject. Plasma and serum samples were stored at -80°C in order to be used for the biochemical (inflammatory cytokines and chemokines) and immunological parameter measurements (complement components, IgA serum level, IgA galactosylation) as required. EDTA tubes were used to collect blood samples for DNA isolation. The EDTA anticoagulated blood sample from each subject was divided into three aliquots, two of which were processed for DNA extraction, in a separate manner, using commercial kits. The first DNA sample was stored in ethanol 75% at 4°C, in the form of a pellet. The second one was suspended in deionised nuclease free water and subdivided into two aliquots ready to use. They were independently stored at 4°C and at -20°C. Storage at 4°C prevents degradation due to the mechanical stress of freezing and thawing and storage at -20°C inhibits possible DNase activity. This method was adopted to guarantee an adequate quality DNA sample and long term storage. Moreover, the last aliquot of whole blood was stored at -20°C in order to isolate a DNA sample whenever needed. DNA was isolated in the Bari, Trieste, Genova and Aachen laboratories. F.P.S., F.S., A.A., J.F., E.A. oversaw the accuracy of the DNA sample collection. All the IgAN patients and relatives, from whom DNA samples were collected, were listed in the database and were available to the IgAN Consortium partners and to the investigators proposing collaborative studies. The protocol was approved by the local Ethical Committees. The documents regarding each local Ethical Committee who have given approval for the study, in their original language, have been disclosed as a window in the IgAN Consortium website [17], purposely constructed for this study.
3. Construction of a website to share scientific information between partners and to divulge obtained data to the community. The IgAN Consortium website was designed and realized by Apulia Biotech (Valenzano, Bari, Italy) and is managed by Altanet SRL (Altamura, Bari, Italy) internet provider. The website includes three sections: (a) Information, (b) Registry and (c) News (fig. 1) all readily accessible for public vision. Section (a) contains guidelines for the collection of blood samples, information about the disease and the informed consent form to be compiled. Section (b) encloses the electronic database for the collection of all personal and clinical data of each enrolled subject. The IgAN Consortium partners are given personal passwords to insert and modify information included in the database. This section is also provided with a search page (clinical finding search page) that gives free access to the database. Groups of IgAN patients can be selected according to their specific characteristics. The clinical history can be viewed observing the confidentiality of each subject. Important information about IgAN Consortium policy for collaborative studies are given. This section is periodically updated in all its parts. Section (c) is devoted to public dissemination of achieved data. It presently includes the report of three years of activities.
4. Construction of an electronic database, for the collection of personal data, clinical, histological and laboratory findings. The database was developed in Filemaker Pro 5.0 obtaining a powerful cross platform relational database. The website interface was realized by using Lasso by Omnipilot Software Inc as the local data markup language (LDML) that connected our HyperText Markup Language (HTML) pages to the Filemaker database. 4D Webstar was used as Internet server. It encloses the list of IgAN patients and relatives enrolled, and the familial IgAN pedigrees. Pedigrees were drawn using the software Cyrillic 2.1. Each collected family and subject was univocally identified by a sequence of three numeric codes: partner unit, family and subject code. Each partner identified a staff member, responsible for the attribution of the codes. Progressive numeric codes for families and subjects were assigned according to the time of enrolment in this study. All the data collected for each enrolled subject at the onset of the disease, at the time of the renal biopsy and at the available follow up are included in the database. Serum creatinine (sCR) and daily proteinuria were obtained from IgAN patients at different times. Daily proteinuria was defined as mild (<1 g), moderate (1–3 g) and severe (>3 g). Renal function was evaluated by creatinine clearance obtained using the Cockcroft formula [18]. According to the K-DOQI guidelines [19] the stages of chronic kidney disease were classified as kidney damage with normal renal function (creatinine clearance ≥ 90 ml/min), mild (60–89 ml/min), moderate (30–59 ml/min) and severe impaired renal function (15–29 ml/min) and kidney failure (<15 ml/min or dialysis). According to the 2003 ESH/ESC hypertension guidelines [20], subjects were defined hypertensives when blood pressure was above 130/85 mmHg or when they received anti-hypertensive drugs. Three outcome measures were considered in the follow up of the IgAN patients: chronic renal insufficiency, dialysis and renal transplant. Renal biopsy specimens were scored considering the severity of glomerular, tubulointerstitial and vascular lesions. According to the WHO classification 3 histological grades (G) were identified: a) G1 (mild disease); b) G2 (moderate disease); c) G3 (severe disease) [21,22]. In the search page of the website Registry-section the biopsy-proven IgAN patients (BP-IgAN) may be selected according to these clinical and laboratory characteristics at the time of the renal biopsy or according to the main outcome measures. The result page included three important links: the "pedigree" field allows the visualization of the familial IgAN cases; the "info" field introduces the complete clinical history of each patient and the "follow up" field links directly to the data at different follow up times for each patient.
A medical geneticist collaborating with each nephrology unit involved in the project offered genetic counselling to IgAN patients and relatives. The main purpose of the counselling is to give information on a complex genetic disease such as IgAN and to illustrate the Biobank objectives, underlining that the genetic tests will not give an immediate answer to the possibility of developing the disease or not. The clinicians explained to all patients the potential familial incidence of the disease and invited their relatives to the genetic counselling session in which a clinical screening was proposed for all relatives. In addition, they were informed on the possibility to be enrolled in this Biobank. For this purpose, the aims of the project were discussed in lay terms with respect to the risk for the involved subjects. It was clearly stated that the subjects tested would not benefit personally from the genetic studies. Financial rewards were not offered to encourage participation. Patients and relatives were informed that the DNA would be used only for the present project i.e. genetic dissection of IgAN. Moreover, a detailed description of the genetic studies to be carried out was given. Informed consent was signed after the genetic counselling.
Normal subjects were recruited among healthy blood donors with negative urinalysis and no history of renal disease, diabetes, hypertension or metabolic disorders. They were matched for age, gender and ethnicity to the IgAN patient population enrolled. They were scrupulously informed about the aim of the study. They gave a written informed consent for DNA collection and for the handling of their genetic and personal data for the present project. The collection of DNA samples is still underway. The IgAN Consortium collected DNA samples from IgAN patients and relatives belonging to at least three generations, with particular attention paid to the collection of all first degree relatives. DNA samples, collected from the research units are listed in table 1. All the subjects enrolled were of Caucasian origin and belonging to different European geographic area: North and South Italy (Lombardia, Piemonte, Friuli Venezia Giulia and Puglia region), Greece (Thessaloniki), Germany (Nordrhein, Westfalia region). DNA samples from 72 multiplex IgAN families, 63 suspected IgAN families, 1068 patients with biopsy-proven IgAN were obtained. The number of collected relatives for each family enrolled was different. Fifty-one IgAN families have less than 10 relatives, 17 include a number of relatives ranging from 10 to 20 subjects and finally 4 IgAN families have more than 20 relatives. Forty-nine suspected IgAN families have less than 10, 13 less than 20 and 1 family has more than 20 relatives. A large sample size was obtained in the form of trios (son/daughter affected and healthy parents). A large number of healthy blood donors adequately matched for age, gender and origin to the patient populations were recruited as controls in each area.
Seventy-two IgAN families included DNA samples from 94 pairs of biopsy proven IgAN patients and a total number of 2573 pairs of relatives. Table 2 shows the distribution of pairs of relatives collected from the 72 IgAN families according to genetic relationship and disease.
A gradually increasing number of DNA samples collected during the first three years of our project (from beginning of October 2000 to the end of December 2003) is shown in figure 2. Some valuable and statistically productive genetic studies have been launched within the 5th Framework Programme 1998–2002 of the European project No. QLG1-2000-00464. Case-control and family-based association studies investigating the possible role of some Th1/Th2/Th3/TR-type, and of monocyte/macrophage cytokines gene polymorphisms and Core1 β 1,3- Galactosyltransferase (β1.3 C1GALT1) gene are presently under way. Some preliminary data are publicly available in the "Technology Marketplace" website [23]. Further investigations needs to strengthen the data obtained.
Utility and discussion
We have described the constitution of the European IgAN Biobank which is the first DNA bank created in the world by the multi-centre collaborative project for this disease. The main aim of this project was to organize a genomic DNA bank of biopsy-proven IgAN patients and their family members belonging to three different geographic areas of Europe (Germany, Greece and Italy). The Biobank has taken care of ethical aspects since the collection and storage of personal and genetic information follow the indications of the Council of Europe, as stated in the Convention on human rights and biomedicine (Nov 19, 1996), and those of UNESCO, as reported in the Declaration on human genome (Nov 11, 1997). In addition, the handling of personal and genetic data collected follow the indications and norms existing in each participating country.
The IgAN Consortium has set up a website to allow the partners to share the scientific information and for public dissemination of all the information related to the disease and new achieved data [17]. The accessible database without restriction permits to share clinical data of IgAN patients with other participating centres. A search page supports the investigators in selecting IgAN patient groups with respect to specific clinical data to perform prospective and retrospective studies.
Studies that seek to identify disease genes can be divided into two categories: linkage and association studies. Linkage studies speculate on the co-segregation of marker alleles and disease of interest in multiplex families while association studies detect linkage disequilibrium between marker and disease loci using case-control and/or family-based designs. In the simplest design, genetic variant frequencies of exposure are compared to diseased cases and non diseased controls. This approach, highly effective in detecting genes of modest effect, is susceptible to "population stratification bias" arising from differences in the genetic background of cases and controls. The distribution of alleles in a population is related to the ethnic and social background and geographical origin of their parents. There is a great debate on the possibility of analysing isolated populations for genetic studies, which may be homogeneous such as Finnish, Icelanders, Sardinians, or heterogeneous populations such as in the case of the UK Biobank [24-28]. Researchers reason that in a population such as Iceland's (270,000 individuals), which expanded from a relatively small number of founders and that did not experience significant immigration, there should be fewer opportunities for particular markers and disease genes to have become separated down the generations. On the other hand some investigators recently provide results that isolated populations are unlikely to be very different from more mixed populations in terms of linkage disequilibrium [29]. Tom Meade of the Medical Research Council (MRC) in Britain, acknowledged that heterogeneous populations will have some advantage because the results will be representative of the population as a whole [25]. The family-based approach in which controls are represented by healthy family members has attracted considerable attention since it minimizes population stratification bias and it avoids making erroneous conclusion. In our Biobank we proposed to collect DNA samples from patients, their relatives and an adequate number of controls of different geographic areas to enable each research unit, independently, to perform case-control and/or family-based association studies in a population belonging to a restricted area. Whenever statistical evidence for a genetic association was obtained, the same data would be checked for the populations collected in other geographic areas. This organization could be highly representative of the IgAN genetic background and more powerful in the detection of weak genetic effects.
It is the policy of the IgAN Consortium to favour further collaborative genetic studies. Other investigators not belonging to the IgAN Consortium, who would like to have access to the DNA Bank for genetic studies, must submit the study design to the scientific committee represented by the principal investigators of the IgAN Consortium research units (F.P.S., F.S., A.A., J.F., E.A., K.Z., L.Z.) [see Additional file 1] and must agree to the terms for data publication. Adequate procedures are used to minimize inadvertent release of personal information and to guarantee confidentiality of genetic results, such as disease genetic predisposition or family relationships (non-paternity, adoption). All results published must not reveal the patient's identity.
The availability of genetic material from participants in large epidemiological studies represents an invaluable resource for exploring genetic and environment influence on disease risk. In fact, the world's largest study on cancer is the multi-centre European Prospective Investigation into Cancer and Nutrition (EPIC), which started recruiting in the early 1990's and has now available more than 500,000 individuals [30]. This has allowed the project to move onto the study of genetics since its sample includes 350 cases of colon cancer and nearly 2000 of breast cancer. Our data collection of more than 1000 individuals with IgAN is close to the above mentioned EPIC data for each disease. Robert Hoover, director of epidemiology and biostatistics at the National Cancer Institution at Bethesda, Maryland, defines a large epidemiological study as a study which generates 1000 to 1500 individuals with a particular cancer in a reasonable time [26]. Keeping this in mind we collected DNA samples from more than 1000 IgAN patients. However, our Biobank will continue to enrol additional patients in the future for potential stratification of our population with respect to specific phenotypes thus promoting correlations with discovered genotypes.
During these last ten years many biobanks have been organized but there is no system for comprehensive registration of the DNA banks. Surely many case-control and family-based association studies will originate from these DNA banks in the future. The IgAN Consortium Biobank has allowed us to carry out a linkage study and various case-control and family-based association studies whose investigation is currently under way. We think that the inclusion of all DNA banks in a public registry might create many advantages for scientists and patients as follows: I) people involved in the scientific and clinical activity of a certain diseases may have a point of reference; II) network of scientists involved in basic and clinical research may obtain many advantages by the public registry; III) patients may contribute by donating their DNA sample to the bank thus increasing the number of collected DNA samples for future studies; IV) the public registry would be available to contribute knowledge which can be generalised regardless results from genetic studies. (Table 3)
Conclusion
IgAN is a complex disorder with both genes and environment contributing in the pathogenesis. A Biobank, depository of a large number of DNA samples from IgAN patients and their relatives, together with a database including their personal and clinical data, has a crucial role in the genetic dissection of IgAN. Such an organization favours and stimulates collaborative studies. The constitution of a well organized and multi-centre biobank, an adequate choice with respect to design strategies and improved genotyping methods could greatly enhance the current understanding of the molecular genetic basis of this disease. Finding complex disease genes may allow us to determine which subjects are at risk of IgAN before that genetic susceptibility is converted into disease. In addition, the identification of such genes should reveal more about the molecular pathway causing the disease, thus suggesting new and relevant targets for more efficient drug treatment.
Availability and requirements
A password is not required to search the IgAN Consortium website database, the password is only necessary to insert and to modify data [17].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
FPS founded the European IgAN Consortium. FS, AA, JF, EA, GMG and GMF contributed in organizing the IgAN Consortium Biobank. GC, DDT, MF, DP, PRM and DK collected DNA samples, clinical and personal data of the enrolled subjects. Moreover they provided the management of the database. KZ, LZ and LB contributed in the molecular genetic study design. The material submitted has been approved by all authors, has not been previously reported, and is not under consideration for publication elsewhere.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Appendix 1: The scientific committee represented by the principal investigators of the IgAN Consortium research units
Click here for file
Acknowledgements
This study was supported by the 5th European Framework Programme (QLG1-CT-2000-00464), the Ministero dell'Istruzione, Universita' e Ricerca PRIN 2001-067748, L.488/92 Cluster 03 and FIRB 2001-RBNE013JYN, CEGBA (Centro di Eccellenza Genomica in Campo Biomedico ed Agrario) and Ministero della Salute "Programmi Speciali" – Art. 12 bis, comma 6, d.lgs. 229/99 – Destinatario Istituzionale Regione Puglia. We are indebted to Mariella Mastrolonardo for her assistance during the revision of the manuscript.
Figures and Tables
Figure 1 IgAN Consortium website structure.
Figure 2 DNA samples collected from the beginning of the IgAN Consortium European project.
Table 1 Collection of DNA samples from IgAN patients, relatives and apparently healthy controls obtained by the IgAN Consortium.
Partner No. Local site IgAN families No. Trios No. Biopsy-proven IgAN patients No. Controls No.
1 Bari, Italy 32 (43*) 91 456 215
2 Brescia, Italy 31(5*) 47 360 204
3 Trieste, Italy 4(3*) 6 116 221
4 Aachen, Germany 2(8*) 9 62 200
5 Thessaloniki, Greece 3(4*) 6 74 200
Total samples 72(63*) 159 1068 1040
*Number in brackets represents additional suspected IgAN families. These are families in which one member is affected by biopsy-proven IgAN and others -not biopsied have recurrent macroscopic hematuria episodes in concomitance of upper respiratory tract infections or persistent microscopic hematuria.
Table 2 Distribution of pairs of relatives in 72 IgAN families according to genetic relationship and IgAN.
IgA Nephropathy status
Relationship A/A A/P A/N N/N Total pair
Spousal 0 2 25 35 62
Parent-offspring 20 19 147 211 397
Sibling 44 32 168 185 429
Half-sibling 0 0 4 0 4
Granparental 0 4 5 89 98
Avuncular 8 22 170 382 582
Cousin 5 6 40 238 289
Other 17 9 141 545 712
Total Pairs 94 94 700 1685 2573
(A) affected subjects with biopsy-proven disease; (P) subjects probably affected by IgAN when persistent microscopic hematuria without biopsy-proven glomerulonephritis, or chronic renal insufficiency or ESRD treated with hemo or peritoneal dialysis, or renal transplantation after biopsy-proven IgAN or unknown glomerulonephritis, or death for chronic uremia, occurred; (N) non-affected subjects in the absence of urinary abnormalities.
Table 3 Information in the Data DNA Bank
• Descriptive information
Brief title
Brief summary
Condition or disease
Availability for single-patient or expanded-access use
• Recruitment information
Overall studies status
• Location and Contact information
• News
==== Refs
Schena FP A retrospective analysis of the natural history of primary IgA nephropathy worldwide Am J Med 1990 89 209 215 2200265 10.1016/0002-9343(90)90300-3
Maisonneuve P Agodoa L Gellert R Stewart JH Buccianti G Lowenfels AB Wolfe RA Jones E Disney AP Briggs D McCredie M Boyle P Distribution of primary renal diseases leading to end-stage renal failure in the United States, Europe, and Australia/New Zealand: results from an international comparative study Am J Kidney Dis 2000 35 157 165 10620560
D'Amico G The commonest glomerulonephritis in the world: IgA nephropathy Q J Med 1987 64 709 727 3329736
Julian BA Waldo FB Rifai A Mestecky J IgA nephropathy, the most common glomerulonephritis worldwide. A neglected disease in the United States? Am J Med 1988 84 129 132 3337116 10.1016/0002-9343(88)90019-8
Jennette JC Wall SD Wilkman AS Low incidence of IgA nephropathy in blacks Kidney Int 1985 28 944 950 4087701
Hoy WE Hughson MD Smith SM Megill DM Mesangial proliferative glomerulonephritis in southwestern American Indians Am J Kidney Dis 1993 21 486 496 8488816
Julian BA Quiggins PA Thompson JS Woodford SY Gleason K Wyatt RJ Familial IgA nephropathy. Evidence of an inherited mechanism of disease N Engl J Med 1985 312 202 208 3855328
Wyatt RJ Rivas ML Julian BA Quiggins PA Woodford SY McMorrow RG Baehler RW Regionalization in hereditary IgA nephropathy Am J Hum Genet 1987 41 36 50 3605095
Scolari F Amoroso A Savoldi S Mazzola G Prati E Valzorio B Viola BF Nicola B Movilli E Sandrini M Campanini M Maiorca R Familial clustering of IgA nephropathy: further evidence in an Italian population Am J Kidney Dis 1999 33 857 865 10213640
Schena FP Cerullo G Rossini M Lanzilotta SG D'Altri C Manno C Increased risk of end-stage renal disease in familial IgA nephropathy J Am Soc Nephrol 2002 13 453 460 11805175
Schena FP Immunogenetic aspects of primary IgA nephropathy Kidney Int 1995 48 1998 2013 8587264
Egido J Julian BA Wyatt RJ Genetic factors in primary IgA nephropathy Nephrol Dial Transplant 1987 2 134 142 3114673
Schena FP Scivittaro V Ranieri E IgA nephropathy: pros and cons for a familial disease Contrib Nephrol 1993 104 36 45 8325032
Gharavi AG Yan Y Scolari F Schena FP Frasca GM Ghiggeri GM Cooper K Amoroso A Viola BF Battini G Caridi G Canova C Farhi A Subramanian V Nelson-Williams C Woodford S Julian BA Wyatt RJ Lifton RP IgA nephropathy, the most common cause of glomerulonephritis, is linked to 6q22–23 Nature Genetics 2000 26 354 357 11062479 10.1038/81677
Dean M Approaches to identify genes for complex human diseases: Lessons from Mendelian disorders Hum Mutat 2003 22 261 274 12955713 10.1002/humu.10259
Spielman RS McGinnis RE Ewens WJ Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM) Am J Hum Genet 1993 52 506 516 8447318
IgAN Consortium website
Cockcroft DW Gault MH Prediction of creatinine clearance from serum creatinine Nephron 1976 16 31 41 1244564
National Kidney Foundation K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification Am J Kidney Dis 2002 39 S1 S266 11904577 10.1053/ajkd.2002.32799
Cifkova R Erdine S Fagard R Farsang C Heagerty AM Kiowski W Kjeldsen S Luscher T Mallion JM Mancia G Poulter N Rahn KH Rodicio JL Ruilope LM van Zwieten P Waeber B Williams B Zanchetti A ESH/ESC Hypertension Guidelines Committee Practice guidelines for primary care physicians: 2003 ESH/ESC hypertension guidelines J Hypertens 2003 21 1779 1786 14508180
Churg J Sobin LH Renal Disease Classification and atlas of glomerular disease 1982 Igaku-Shoin, Tokyo, New York
Schena FP Coppo R Davison AM, Cameron JS, Grünfeld JP, Ponticelli C, Ritz E, Winearls CG, van T persele C IgA nephropathies Oxford Textbook of Clinical Nephrology 2005 3 Oxford University Press 469 501
The "Technology Marketplace" website
Risch NJ Searching for genetic determinants in the new millennium Nature 2000 405 847 856 10866211 10.1038/35015718
Gulcher JR Stefánsson K The Icelandic healthcare database and informed consent New Eng J Med 2000 342 1827 1830 10853008 10.1056/NEJM200006153422411
Abbott A Manhattan versus Reykjavik Nature 2000 406 340 342 10935610 10.1038/35019167
Barbour V UK Biobank: a project in search of a protocol? Lancet 2003 361 1734 1738 12767753 10.1016/S0140-6736(03)13377-6
Annas GJ Rules for research on human genetic variation – Lessons from Iceland New Eng J Med 2000 342 1830 1833 10853009 10.1056/NEJM200006153422412
Eaves IA Merriman TR Barber RA Nutland S Tuomilehto-Wolf E Tuomilehto J Cucca F Todd JA The genetically isolated populations of Finland and Sardinia may not be a panacea for linkage disequilibrium mapping of common disease genes Nature Genet 2000 25 320 323 10888882 10.1038/77091
Riboli E Hunt KJ Slimani N Ferrari P Norat T Fahey M Charrondiere UR Hemon B Casagrande C Vignat J Overvad K Tjonneland A Clavel-Chapelon F Thiebaut A Wahrendorf J Boeing H Trichopoulos D Trichopoulou A Vineis P Palli D Bueno-De-Mesquita HB Peeters PH Lund E Engeset D Gonzalez CA Barricarte A Berglund G Hallmans G Day NE Key TJ Kaaks R Saracci R European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection Public Health Nutr 2002 5 1113 1124 12639222 10.1079/PHN2002394
|
16329758
|
PMC1318455
|
CC BY
|
2021-01-04 16:03:44
|
no
|
BMC Nephrol. 2005 Dec 5; 6:14
|
utf-8
|
BMC Nephrol
| 2,005 |
10.1186/1471-2369-6-14
|
oa_comm
|
==== Front
BMC Cell BiolBMC Cell Biology1471-2121BioMed Central London 1471-2121-6-451633666410.1186/1471-2121-6-45Research ArticleCysteine-rich protein 1 (CRP1) regulates actin filament bundling Tran Thuan C [email protected] CoreyAyne [email protected] Tamara S [email protected] Jeffrey A [email protected] Department of Biochemistry and Biophysics, Oregon State University, Corvallis, Oregon 97331, USA2005 8 12 2005 6 45 45 2 9 2005 8 12 2005 Copyright © 2005 Tran et al; licensee BioMed Central Ltd.2005Tran 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
Cysteine-rich protein 1 (CRP1) is a LIM domain containing protein localized to the nucleus and the actin cytoskeleton. CRP1 has been demonstrated to bind the actin-bundling protein α-actinin and proposed to modulate the actin cytoskeleton; however, specific regulatory mechanisms have not been identified.
Results
CRP1 expression increased actin bundling in rat embryonic fibroblasts. Although CRP1 did not affect the bundling activity of α-actinin, CRP1 was found to stabilize the interaction of α-actinin with actin bundles and to directly bundle actin microfilaments. Using confocal and photobleaching fluorescence resonance energy transfer (FRET) microscopy, we demonstrate that there are two populations of CRP1 localized along actin stress fibers, one associated through interaction with α-actinin and one that appears to bind the actin filaments directly. Consistent with a role in regulating actin filament cross-linking, CRP1 also localized to the membrane ruffles of spreading and PDGF treated fibroblasts.
Conclusion
CRP1 regulates actin filament bundling by directly cross-linking actin filaments and stabilizing the interaction of α-actinin with actin filament bundles.
==== Body
Background
Stress fibers are bundles of actin microfilaments formed in cells following integrin-mediated attachment and spreading [1]. Regulation of these contractile fibers is critical for cell adhesion and motility. The microfilaments within stress fibers are held together by specialized bundling proteins, such as myosin and α-actinin, which can interact simultaneously with two actin filaments. These classical actin-bundling proteins have been studied extensively leading to a basic understanding of their interaction with actin filaments and regulation of stress fibers. One important discovery was the periodic and alternating association of myosin and α-actinin which is clearly visualized as a beaded pattern along stress fibers in cells stained for immunofluoresence microscopy [2]. Although it is not understood how this alternating association of myosin and α-actinin with the microfilaments is regulated, it is critical for the contractility of the stress fiber. In addition to myosin and α-actinin, actin stress fibers are decorated with numerous other proteins, associated either directly or through interaction with other actin-binding proteins. Determining the function of these ancillary proteins is important for understanding the regulation of stress fibers.
Modulation of the actin cytoskeleton by LIM domain proteins is an active and growing field of research [3]. LIM domains are cysteine-rich sequences of 50–60 amino acid residues that contain two tandem zinc fingers [4]. Several LIM proteins have been demonstrated to interact with and/or regulate α-actinin. Vallenius et al. [5] have reported that reversion-induced LIM (RIL) protein associates with α-actinin increasing its binding to actin filaments in vitro and may alter actin stress fibers in various cell types. Four and a half LIM domain protein 3 (FHL3) has been demonstrated to disrupt actin stress fibers in C2C12 myoblasts presumably by binding to actin filaments and inhibiting α-actinin bundling [6]. In addition, ENH [7], ALP [8], Cypher [9], CLIM1 [10], CLP-36 [11,12], zyxin [13], and the cysteine-rich protein (CRP) family [14,15] have also been demonstrated to interact with α-actinin, although it is not clear if these proteins influence α-actinin function.
The CRP family, which includes CRP1, CRP2, CRP3, and the thymus LIM protein (TLP), is a subgroup of LIM domain proteins containing two LIM domains linked to short glycine-rich repeats [16]. CRPs are highly conserved between species with CRP1 from human, chicken, and quail having an amino acid sequence identity greater than 90% [17]. Although the CRPs have different patterns of expression, evidence suggests that they are functionally similar [14,16]. CRPs are important for cell differentiation presumably by modulating protein-protein interactions involved in transcriptional regulation [16,18]. Outside of the nucleus, CRPs clearly localize to focal adhesions and the actin cytoskeleton, and have been postulated to play a role in controlling these structures [16].
CRP1, CRP2, and CRP3, have all been shown to bind α-actinin [14,15]. Interaction between the two proteins was first reported for CRP1 using affinity chromatography, solution and solid-phase binding assays, and co-immunoprecipitation from cell lysates [15]. A follow up study demonstrated that α-actinin could interact equally with CRP1, CRP2, and CRP3 [14]. In addition, it was determined that the CRPs were interacting with the actin-binding domain of α-actinin [15] and the α-actinin-binding site was mapped to amino acid residues 62–79 of human CRP1 [19]. Based on the binding assays and co-localization within the cell, it was proposed that the interaction of CRP with α-actinin was responsible, in part, for its localization to the actin cytoskeleton [14,15]. Recently, Grubinger and Gimona [20] demonstrated that CRP2 binds directly to actin filaments and suggested that CRP2 does not interact with α-actinin. We have found that there are two populations of CRP1 associated with actin stress fibers, one interacting with α-actinin and one that appears to interact directly with the actin filaments. Although, CRPs have been shown to bind to α-actinin and actin filaments, it is not clear how these proteins regulate the actin cytoskeleton. In this study, we show that CRP1 regulates actin filament bundling by directly cross-linking actin filaments and stabilizing the interaction of α-actinin with actin filament bundles.
Results
Expression of CRP1 increases F-actin bundling in REFs
To determine the influence of CRP1 on the actin cytoskeleton, rat embryonic fibroblasts (REFs) were transfected with increasing concentrations of DNA encoding CFP-CRP1 resulting in increasing levels of expression of the fluorescent fusion protein (Fig. 1A). Expression of CFP-CRP1 was observed in approximately 15% of the cells with levels ranging from 2–4 fold that of endogenous CRP1 (data not shown). Consistent with immunostaining for endogenous CRP1 (see Additional files 1 and 2) and previous reports [14-16,21,22], CFP-CRP1 was localized diffusely in the cytoplasm, along the actin cytoskeleton, and in the nucleus of some cells (~10%). These results suggest that the CFP tag does not significantly affect the function of the CRP1 protein. Although the morphology of the REFs was unaffected, expression of CFP-CRP1 appeared to increase the bundling of actin microfilaments (Fig. 1, see arrows). If the expression of CFP-CRP1 was resulting in increased actin filament bundling, then CFP-CRP1 would be expected to redistribute to the Triton X-100 insoluble cytoskeletal fraction. As shown in Fig. 1A, insoluble CFP-CRP1 increased with expression. In addition, an immunoreactive band migrating below CFP-CRP1 was recognized, presumably a proteolytic fragment. The distribution of actin and α-actinin in the Triton X-100 soluble and insoluble fractions were also examined. However, with a transfection efficiency of only 15%, a significant change in the insolubility of actin and α-actinin was not detected (data not shown).
Figure 1 Expression of CRP1 increases the bundling of cellular actin filaments. REFs transfected with pECFP-CRP1 as described in Experimental Procedures were lysed and the Triton X-100 soluble and insoluble fractions separated for analysis. The transfection conditions were: 1) 6 μL FuGENE, no DNA; 2) 3 μL FuGENE, 0.5 μg DNA; 3) 3 μL FuGENE, 1.0 μg DNA; 4) 6 μL FuGENE, 1.0 μg DNA; and 5) 6 μL FuGENE, 2.0 μg DNA. Immunoblot shows the expression of CFP-CRP1 in each fraction (A). REFs transfected with pECFP-CRP1 using transfection condition 4 were fixed and prepared for fluorescence microscopy. Rhodamine-phalloidin staining of F-actin (A, B, C) shows that the cells expressing CFP-CRP1 (A', B', C') have enlarged stress fibers (see arrows). Results are representative of 4 separate experiments. Bar = 10 μm.
CRP1 bundles F-actin independently of α-actinin
Since CRP1 has been demonstrated to bind to α-actinin [15], we proposed that CRP1 was regulating actin filament bundling by increasing α-actinin bundling activity. In order to test this hypothesis, α-actinin bundling activity was examined in the absence and presence of CRP1. For these experiments (Fig. 2), actin filament bundling was determined using a sedimentation assay with 2.5 μM α-actinin alone, 2.5 μM CRP1 alone, 1.25 μM α-actinin + 1.25 μM CRP1 (2.5 μM total protein), and 2.5 μM α-actinin + 2.5 μM CRP1 (5.0 μM total protein). Although, CRP1 did not appear to influence the bundling activity of α-actinin, we did find that when CRP1 was added to α-actinin bundled actin filaments, a 2-fold increase in α-actinin was observed in the bundles (Fig. 3). These results suggest that CRP1 was stabilizing the interaction of α-actinin with the actin filaments in the bundles.
Figure 2 CRP1 does not influence the bundling activity of α-actinin. α-Actinin and CRP1 were preincubated for 15 min at room temperature. Actin filaments were then added, incubated for 30 min at room temperature, and centrifuged at 10,000 × g. Proteins from the supernatant (S) and the pellet (P) were separated by electrophoresis and detected by Gelcode Blue staining (A). The percentage of total actin (B) and α-actinin and CRP1 (C) in the pellet was quantified as described in Experimental Procedures. n = 4 ± SEM.
Figure 3 CRP1 stabilizes the interaction of α-actinin with actin bundles. Actin filaments were bundled by incubation with α-actinin for 30 min. at room temperature. CRP1 was then added, incubated for an addition 30 min., and centrifuged at 10,000 × g. Proteins from the supernatant (S) and the pellet (P) were separated by electrophoresis and detected by Gelcode Blue staining (A). The percentage of total α-actinin in the pellet was quantified as described in Experimental Procedures (B). n = 3 ± SEM.
We also found that CRP1 alone could bundle actin filaments in a concentration dependent manner (Figs. 2 and 4). Furthermore, CRP1 bound to the actin filaments more efficiently than α-actinin, 75–90% compared to 35–45%, under all of the experimental conditions (Fig. 2C). To confirm that CRP1 was bundling the actin filaments, the bundling assay was carried out on glass coverslips and the proteins fixed with 3% formaldehyde followed by staining with rhodamine-phalloidin (Fig. 4C–F). Images of the samples clearly show that CRP1 induced the formation of actin filament bundles similar to that of α-actinin.
Figure 4 CRP1 bundles F-actin independently. The indicated concentration of CRP1 was incubated with actin filaments for 30 min at room temperature and centrifuged at 10,000 × g. Proteins from the supernatant (S) and the pellet (P) were separated by electrophoresis and detected by Gelcode Blue staining (A). The percentage of total actin in the pellet was quantified as described in Experimental Procedures (B). n = 4 ± SEM. Bundling reactions were also carried out on glass coverslips, fixed, and stained with rhodamine-phalloidin to allow visualization of the actin filaments bundles: (C) actin filaments, (D) actin filaments + 2.5 μM α-actinin, (E) actin filaments + 5.0 μM CRP1, and (F) actin filaments + 2.5 μM α-actinin + 2.5 μM CRP1. Results are representative of 4 separate experiments. Bar = 10 μm.
Two populations of CRP1 are associated with actin stress fibers
Previous reports have suggested that CRP1 is localized to actin filaments through interaction with α-actinin [14,15]. Since the above results demonstrated that CRP1 could directly bind to and bundle actin filaments, we asked if CRP1 was actually binding to α-actinin along actin stress fibers within the cell. REFs were co-transfected with DNA encoding CFP-CRP1 and YFP-α-actinin. As in previous experiments, approximately 15% of the cells expressed the tagged proteins with greater than 90% of these cells expressing both proteins. In cells fixed with 3% formaldehyde in PBS using standard protocols, YFP-actinin was observed in its classical beaded pattern along actin stress fibers and within focal adhesions (Fig. 5A). CFP-CRP1 was also observed in focal adhesions and along actin stress fibers; however localization along the stress fibers was continuous and not beaded like that of α-actinin (Fig. 5A). Cells were also fixed with 3% formaldehyde in Triton X-100 buffer in order to remove the soluble cytoplasmic proteins and proteins weakly associated with the cytoskeleton. Using this fixation protocol, CFP-CRP1 was observed to co-localize explicitly with YFP-α-actinin along actin stress fibers and in focal adhesions (Fig. 5B). Scatter plots are shown in Fig. 5 to represent localization of the two proteins in the images. For the unextracted cell, the distribution of fluorescence intensity for the two tags is diffuse as a result of the areas in the cell where the two proteins are not co-localized, with a correlation coefficient of R = 0.4. For the extracted cell, the co-localized populations of the two proteins are represented by the characteristic comet shaped scatter plot with a correlation coefficient of R = 0.7.
Figure 5 Two populations of CRP1 are localized along stress fibers. REFs expressing CFP-CRP1 and YFP-α-actinin were fixed with 3% formaldehyde in PBS (A) or Triton X-100 buffer (B) and prepared for confocal microscopy as described in Experimental Procedures. Images of CFP-CRP1 (A and B) and YFP-α-actinin (A' and B') localization are shown. A zoomed region is shown in the lower left corner of each image. Merged images are shown in A" and B". Results are representative of 4 separate experiments. Bar = 10 μm.
Although we had clearly identified a population of CRP1 that co-localized with α-actinin, this did not prove that the two proteins were interacting directly with each other. An established method for determining the interaction between two localized proteins in a cell involves fluorescence resonance energy transfer (FRET) microscopy. In these experiments, we followed the method for photobleaching FRET (pbFRET) described by Karpova et al. [23] using REFs co-expressing CFP-CRP1 (donor) and YFP-α-actinin (acceptor). For this method, the FRET efficiency is determined by measuring the dequenching of the donor emission after selective photobleaching of the acceptor [23,24]. In cells fixed with 3% formaldehyde in Triton X-100 buffer, we consistently observed an increase in CFP emission along actin stress fibers following photobleaching of the YFP (Fig. 6A–D). As a control, the CFP emission on a similar region which was not photobleached was also measured. The histogram in Fig. 6E shows that the distributions of the FRET efficiencies for the photobleached regions (Ef) was shifted positively compared to the unbleached control regions (Ec) with the mean Ef significantly different from the control Ec (9.6 ± 0.74 vs. 2.9 ± 0.61, p < 0.05). For comparison, the FRET efficiency for a CFP-YFP fusion protein was reported as 7.96 ± 0.38 for the photobleached regions compared to 2.21 ± 0.26 for the unbleached control regions [23].
Figure 6 CRP1 binds α-actinin along actin stress fibers. REFs were co-transfected with pECFP-CRP1 and pEYFP-α-actinin and FRET microscopy carried out as described in Experimental Procedures. (A) Overlay image of cell expressing CFP-CRP1 and YFP-α-actinin showing the ROI that was photobleached during the FRET experiment. A color scaled image of the CFP fluorescence of the cell prior to (B) and immediately following (C) the photobleaching period. Red represents a high signal and blue, a low signal. Bar = 10 μm. (D) The intensity of the CFP fluorescence within the ROI was quantified for each image captured during the experiment. (E) The distribution of FRET efficiencies for the photobleached ROIs, Ef, and control non-photobleached ROIs, Ec, were plotted on the bar graph.
CRP1 localizes to dynamic actin structures
α-Actinin is not only localized to stress fibers and focal adhesions, but also to more dynamic actin structures such as the membrane ruffles of spreading and PDGF treated cells. Thus, if CRP1 cross-links actin filaments and stabilizes the interaction of α-actinin with actin filaments, localization to membrane ruffles would also be expected. Previously, we have shown that platelet-derived growth factor (PDGF) stimulates the relocation of α-actinin to membrane ruffles [25]. In order to determine if CRP1 would relocalize with α-actinin, REFs co-expressing YFP-α-actinin and CFP-CRP1 were stimulated with PDGF for 30 min followed by fixation in Triton X-100 buffer. Images of the two proteins clearly show that CFP-CRP1 is co-localized with α-actinin in the membrane ruffles of PDGF treated cells (Fig. 7A). Experiments were also carried out with co-expressing cells showing the co-localization of CRP1 and α-actinin to the membrane ruffles of REFs during cell spreading (Fig. 7B).
Figure 7 CRP1 localizes to dynamic actin structures. REFs expressing CFP-CRP1 (A and B) and YFP-α-actinin (A' and B') were stimulated with 30 ng/ml PDGF for 30 min or plated on fibronectin for 30 min as previously described [25], fixed with 3% formaldehyde in Triton X-100 buffer, and prepared for microscopy as described in Experimental Procedures. Results are representative of 3 separate experiments. Bar = 10 μm.
Discussion
CRP1 is a LIM domain containing protein which has been demonstrated to localize in the nucleus, the cytoplasm, along actin filaments, and in focal adhesions. Recent studies have shown that the nuclear population of CRPs mediates protein-protein interactions regulating transcription and differentiation[18]. Although CRPs have been found to bind to α-actinin, zyxin, and actin filaments, little is known about how CRPs regulate the actin cytoskeleton. In this study, we examined the regulation of actin filaments by CRP1 in vitro and in cultured cells.
Expression of CRP1 in fibroblasts resulted in increased actin filament bundling (Fig. 1). In addition, we found that CRP1 directly binds to and bundles actin filaments (Fig. 2 and 4). CRP2 has recently been reported to bind actin filaments [20]; however, this is the first study to show that any CRP can bundle actin filaments.
After observing an increase in actin filament bundling in the cells expressing CFP-CRP1, we expected to find that CRP1 was enhancing the bundling activity of α-actinin. However, in vitro bundling studies clearly demonstrated that CRP1 had no influence on the ability of α-actinin to bundle actin filaments (Fig. 3). Furthermore, these results showed that CRP1 and α-actinin do not compete for binding to actin filaments and therefore must bind at different sites. Interestingly, approximately twice as much CRP1 protein pellets with the actin filament bundles compared to α-actinin. The significance of these findings is not clear, but may reflect differences in the mechanisms by which CRP1 and α-actinin bind and bundle actin filaments. Further studies are necessary to determine how CRP1 is bundling actin filaments.
Although α-actinin had previously been demonstrated to bind and presumably localize CRP1 to actin filaments [14,15], the new evidence that CRP1 could directly bind and bundle actin filaments prompted further investigation. Detergent extraction of cells has long been used to improve examination of the cytoskeleton. Previously, we demonstrated that Triton X-100 extraction removes soluble cytoplasmic proteins leaving behind an intact cytoskeleton with associated adhesion and matrix proteins[25]. Extraction with Triton X-100 has also been used to separate and define adhesion and cytoskeletal proteins based on stable association with the actin cytoskeleton [26]. Triton X-100 extraction of fibroblasts co-expressing CFP-CRP1 and YFP-α-actinin allowed us to differentiate two populations of CRP1 along actin stress fibers (Fig. 5). The Triton X-100 resistant population of CRP1 co-localized with α-actinin along stress fibers, whereas the less stable Triton X-100 susceptible population appeared to represent CRP1 which was directly associated with actin filaments. Furthermore, confocal microscopy demonstrated that CFP-CRP1 and YFP-α-actinin were in close enough proximity for FRET between the CFP and YFP (Fig. 6). Since the tags need to be within ~50Å for efficient FRET to occur [24], the results indicate that CRP1 and α-actinin are bound to each other along actin stress fibers.
Conclusion
Evidence from this and other studies suggests that the different populations of CRP1 mediate protein-protein interactions unique to each cellular compartment. Nuclear CRP1 regulates interactions between transcription factors[18]. Cytoplasmic CRP1 modulates the actin cytoskeleton by two mechanisms: 1) stabilizing α-actinin interaction with actin bundles and 2) cross-linking actin filaments. Further studies are needed to determine how the different populations of CRP1 coordinate to modulate the function of the cell.
Methods
Proteins and DNA constructs
α-Actinin was purified from chicken gizzard as previously described [27]. Non-muscle actin (99% pure; 80% β-actin, 20% γ-actin) was polymerized following the manufacturer's protocol (Cytoskeleton, Inc., Denver, CO). The plasmid containing chicken CRP1 cloned into the EcoRI site of pBlueScript II KS (pBSIIKS, Stratagene) was generously provided by Mary C. Beckerle (Univ. of Utah) [15]. The EcoRI insert of CRP1-pBSIIKS was sub-cloned into the enhanced cyan fluorescent fusion protein vector pECFP-C1 (BD Biosciences) and the BamHI-HindIII insert sub-cloned into pProEx HTb (Invitrogen). Nucleotide sequences were confirmed by sequence analysis. His-tagged CRP1 protein was expressed in BL21 bacteria and purified using Ni-NTA resin (Qiagen) following procedures described by the manufacturer. The his tag was cleaved from CRP1 while still bound to the resin using recombinant TEV protease (Invitrogen) following the manufacturer's protocol. The untagged CRP1 protein was concentrated and buffer-exchanged (10 mM HEPES, pH 7.0, 50 mM NaCl, 1 mM EDTA) in a centrifugal filter device (Amicon). YFP-α-actinin was constructed by subcloning the α-actinin gene [28] into the HindIII restriction site of the enhanced yellow fluorescent fusion protein vector pEYFP-N1 (BD Biosciences).
F-actin bundling assays
The bundling of F-actin was determined by sedimentation assays as previously described [28,29]. F-actin (10.4 μM) was incubated with the indicated concentration of CRP1 or α-actinin in bundling buffer (10 mM HEPES, pH 7.0, 50 mM NaCl, 1 mM EDTA) for 30 min at room temperature and centrifuged at 10,000 × g for 30 min. The supernatant and pellet were separated and the proteins analyzed by electrophoresis. Proteins were detected by Gelcode Blue (Pierce) staining and quantified using a KODAK ImageStation 440CF.
F-actin bundles were visualized by fluorescence microscopy following modification of previously described procedures [30,31]. Briefly, 50 μl of assay solution was incubated on a glass coverslip inside a 12-well tissue culture dish. After 30 min, the proteins were fixed by adding 3% formaldehyde in phosphate buffered saline for an additional 30 min. Coverslips were then stained and processed for microscopy as described previously[25].
Cell culture and fluorescence microscopy
Rat embryonic fibroblasts (REFs) were cultured as described previously [25]. Cells were transfected with pECFP-CRP1 and pEYFP-α-actinin using FuGENE 6 (Roche) following manufacturer's protocols. The expression curve of CFP-CRP1 was carried out by varying the ratio of FuGENE to DNA in a final volume of 100 μL serum-free media. The transfection conditions were: 6 μL FuGENE, no DNA; 3 μL FuGENE, 0.5 μg DNA; 3 μL FuGENE, 1.0 μg DNA; 6 μL FuGENE, 1.0 μg DNA; and 6 μL FuGENE, 2.0 μg DNA. Twenty-four hours after transfection, cells were prepared for fluorescence microscopy or scraped into ice-cold lysis buffer (10 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EGTA, 1 mM EDTA, 2 mM Na3VO4, 1% Triton X-100, 0.5% NP-40, 30 mM sodium pyrophosphate, 50 mM NaF, 1 μg/ml leupeptin, and 1 μg/ml aprotinin) as described previously [25]. The lysates were centrifuged at 10,000 × g for 10 min at 4°C, protein from the supernatant and pellet separated by electrophoresis, and immunoblotted with anti-GFP (Santa Cruz), anti-α-actinin (Chemicon), or anti-actin (Sigma). Proteins were detected by enhanced chemiluminence (Pierce) and quantified using a KODAK ImageStation 440CF. For fluorescence microscopy, cells were fixed for 30 min at room temperature with 3% formaldehyde (Tousimis) in PBS or in Triton X-100 buffer (20 mM Tris, pH 7.4, 50 mM NaCl, 1 mM EGTA, 5 mM EDTA, 100 μM Na3VO4, 50 mM sodium pyrophosphate, 1 μg/mL leupeptin, 1 μg/mL aprotinin, and 0.5 % Triton X-100). Digital images were captured using a Zeiss axiovert 100S microscope equipped with a Photometrics CoolSNAP HQ CCD camera controlled by MetaMorph software. Co-localization studies were carried out using a Zeiss LSM 510 confocal microscope. The scatter plots and correlation coefficients were determined using Zeiss Physiology Software v3.2.
Fluorescence resonance energy transfer (FRET) microscopy
REFs were co-transfected with pECFP-CRP1 and pEYFP-α-actinin, cultured for an additional 24 hrs, fixed using 3% formaldehyde in Triton X-100 buffer, and prepared for confocal microscopy as described above. The FRET assays were carried out following the procedure described by Karpova et al. [23]. Briefly, cells were imaged with a Zeiss LSM 510 confocal microscope operated by Zeiss Physiology Software v3.2 using a 63 × 1.3 NA Zeiss oil immersion lens. The microscopy system was configured in multitracking mode to excite the CFP with a 458 nm and YFP with a 514 nm laser line. A region of interest (ROI) containing actin stress fibers was selected for photobleaching. Using the time series function, 5 images of the cell were collected followed by selective photobleaching of the YFP within the ROI with the 514 nm laser line (typically, 150 iterations at 100% laser power was sufficient), and then the collection of 5 additional images. The FRET efficiency was calculated as a percentage using the following formula E = 100 × (Ipostbleach – Iprebleach)/Iprebleach, where I is the intensity of CFP fluorescence within the ROI. As a control, ROIs were selected from non-bleached regions of the cell.
List of abbreviations
The abbreviations used are: RIL, reversion-induced LIM; FHL3, Four and a half LIM domain protein 3; ENH, enigma homologue protein; ALP, actinin-associated LIM protein; CLIM, 36kDa carboxyl terminal LIM domain protein; CRP, cysteine-rich protein; F-actin, filamentous actin; PDGF, platelet-derived growth factor; CFP, cyan fluorescent protein; YFP, yellow fluorescent protein; REFs, rat embryonic fibroblasts.
Authors' contributions
TCT carried out the cell and molecular studies. CS carried out the in vitro bundling assays and spreading experiments. TSF participated in the confocal microscopy analysis. JAG conceived of the study, and participated in its design and coordination and drafted the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Immunostaining for endogenous CRP1. Fluorescence microscopy of REFs stained with antibodies recognizing the C-terminal 17 amino acids residues (PKGFGFGQGAGALVHSE) of rat CRP1. Bar = 10 μm.
Click here for file
Additional File 2
Immunostaining for endogenous CRP1. Fluorescence microscopy of REFs stained with antibodies recognizing the C-terminal 17 amino acids residues (PKGFGFGQGAGALVHSE) of rat CRP1. Bar = 10 μm.
Click here for file
Acknowledgements
The authors thank Dr. Mary Beckerle (Univ. of Utah) for providing CRP1-pBSIIKS. Research was supported by a grant to J.A.G. from the National Institutes of Health (GM 63711). This publication was made possible in part by the Confocal Microscopy (grant number 1S10RR107903-01) and Cell Imaging and Culture Facility of the Environmental Health Sciences Center, Oregon State University, fromgrant number P30 ES00210, National Institute of EnvironmentalHealth Sciences, National Institutes of Health.
==== Refs
Chrzanowska-Wodnicka M Burridge K Rho-stimulated contractility drives the formation of stress fibers and focal adhesions J Cell Biol 1996 133 1403 1415 8682874 10.1083/jcb.133.6.1403
Langanger G Moeremans M Daneels G Sobieszek A De Brabander M De Mey J The molecular organization of myosin in stress fibers of cultured cells J Cell Biol 1986 102 200 209 3510218 10.1083/jcb.102.1.200
Khurana T Khurana B Noegel AA LIM proteins: association with the actin cytoskeleton Protoplasma 2002 219 1 12 11926060 10.1007/s007090200000
Bach I The LIM domain: regulation by association Mech Dev 2000 91 5 17 10704826 10.1016/S0925-4773(99)00314-7
Vallenius T Scharm B Vesikansa A Luukko K Schafer R Makela TP The PDZ-LIM protein RIL modulates actin stress fiber turnover and enhances the association of alpha-actinin with F-actin Exp Cell Res 2004 293 117 128 14729062 10.1016/j.yexcr.2003.09.004
Coghill ID Brown S Cottle DL McGrath MJ Robinson PA Nandurkar HH Dyson JM Mitchell CA FHL3 Is an Actin-binding Protein That Regulates {alpha}-Actinin-mediated Actin Bundling: FHL3 LOCALIZES TO ACTIN STRESS FIBERS AND ENHANCES CELL SPREADING AND STRESS FIBER DISASSEMBLY J Biol Chem 2003 278 24139 24152 12704194 10.1074/jbc.M213259200
Nakagawa N Hoshijima M Oyasu M Saito N Tanizawa K Kuroda S ENH, containing PDZ and LIM domains, heart/skeletal muscle-specific protein, associates with cytoskeletal proteins through the PDZ domain Biochem Biophys Res Commun 2000 272 505 512 10833443 10.1006/bbrc.2000.2787
Xia H Winokur ST Kuo WL Altherr MR Bredt DS Actinin-associated LIM protein: identification of a domain interaction between PDZ and spectrin-like repeat motifs J Cell Biol 1997 139 507 515 9334352 10.1083/jcb.139.2.507
Zhou Q Ruiz-Lozano P Martone ME Chen J Cypher, a striated muscle-restricted PDZ and LIM domain-containing protein, binds to alpha-actinin-2 and protein kinase C J Biol Chem 1999 274 19807 19813 10391924 10.1074/jbc.274.28.19807
Kotaka M Kostin S Ngai S Chan K Lau Y Lee SM Li H Ng EK Schaper J Tsui SK Fung K Lee C Waye MM Interaction of hCLIM1, an enigma family protein, with alpha-actinin 2 J Cell Biochem 2000 78 558 565 10861853 10.1002/1097-4644(20000915)78:4<558::AID-JCB5>3.0.CO;2-I
Vallenius T Luukko K Makela TP CLP-36 PDZ-LIM protein associates with nonmuscle alpha-actinin-1 and alpha-actinin-4 J Biol Chem 2000 275 11100 11105 10753915 10.1074/jbc.275.15.11100
Bauer K Kratzer M Otte M de Quintana KL Hagmann J Arnold GJ Eckerskorn C Lottspeich F Siess W Human CLP36, a PDZ-domain and LIM-domain protein, binds to alpha-actinin-1 and associates with actin filaments and stress fibers in activated platelets and endothelial cells Blood 2000 96 4236 4245 11110697
Crawford AW Michelsen JW Beckerle MC An interaction between zyxin and alpha-actinin J Cell Biol 1992 116 1381 1393 1541635 10.1083/jcb.116.6.1381
Louis HA Pino JD Schmeichel KL Pomies P Beckerle MC Comparison of Three Members of the Cysteine-rich Protein Family Reveals Functional Conservation and Divergent Patterns of Gene Expression J Biol Chem 1997 272 27484 27491 9341203 10.1074/jbc.272.43.27484
Pomies P Louis HA Beckerle MC CRP1, a LIM domain protein implicated in muscle differentiation, interacts with alpha-actinin J Cell Biol 1997 139 157 168 9314536 10.1083/jcb.139.1.157
Weiskirchen R Gunther K The CRP/MLP/TLP family of LIM domain proteins: acting by connecting Bioessays 2003 25 152 162 12539241 10.1002/bies.10226
Weiskirchen R Pino JD Macalma T Bister K Beckerle MC The cysteine-rich protein family of highly related LIM domain proteins J Biol Chem 1995 270 28946 28954 7499425 10.1074/jbc.270.48.28946
Chang DF Belaguli NS Iyer D Roberts WB Wu SP Dong XR Marx JG Moore MS Beckerle MC Majesky MW Schwartz RJ Cysteine-rich LIM-only proteins CRP1 and CRP2 are potent smooth muscle differentiation cofactors Dev Cell 2003 4 107 118 12530967 10.1016/S1534-5807(02)00396-9
Harper BD Beckerle MC Pomies P Fine mapping of the alpha-actinin binding site within cysteine-rich protein Biochem J 2000 350 Pt 1 269 274 10926853 10.1042/0264-6021:3500269
Grubinger M Gimona M CRP2 is an autonomous actin-binding protein FEBS Lett 2004 557 88 92 14741346 10.1016/S0014-5793(03)01451-0
Crawford AW Pino JD Beckerle MC Biochemical and molecular characterization of the chicken cysteine-rich protein, a developmentally regulated LIM-domain protein that is associated with the actin cytoskeleton J Cell Biol 1994 124 117 127 8294495 10.1083/jcb.124.1.117
Sadler I Crawford AW Michelsen JW Beckerle MC Zyxin and cCRP: two interactive LIM domain proteins associated with the cytoskeleton J Cell Biol 1992 119 1573 1587 1469049 10.1083/jcb.119.6.1573
Karpova TS Baumann CT He L Wu X Grammer A Lipsky P Hager GL McNally JG Fluorescence resonance energy transfer from cyan to yellow fluorescent protein detected by acceptor photobleaching using confocal microscopy and a single laser J Microsc 2003 209 56 70 12535185 10.1046/j.1365-2818.2003.01100.x
Siegel RM Chan FK Zacharias DA Swofford R Holmes KL Tsien RY Lenardo MJ Measurement of molecular interactions in living cells by fluorescence resonance energy transfer between variants of the green fluorescent protein Sci STKE 2000 2000 PL1 11752595
Greenwood JA Theibert AB Prestwich GD Murphy-Ullrich JE Restructuring of focal adhesion plaques by PI 3-kinase. Regulation by PtdIns (3,4,5)-p(3) binding to alpha-actinin J Cell Biol 2000 150 627 642 10931873 10.1083/jcb.150.3.627
Greene DK Tumova S Couchman JR Woods A Syndecan-4 Associates with alpha -Actinin J Biol Chem 2003 278 7617 7623 12493766 10.1074/jbc.M207123200
Feramisco JR Burridge K A rapid purification of alpha-actinin, filamin, and a 130,000-dalton protein from smooth muscle J Biol Chem 1980 255 1194 1199 7356657
Fraley TS Tran TC Corgan AM Nash CA Hao J Critchley DR Greenwood JA Phosphoinositide Binding Inhibits {alpha}-Actinin Bundling Activity J Biol Chem 2003 278 24039 24045 12716899 10.1074/jbc.M213288200
Corgan AM Singleton C Santoso CB Greenwood JA Phosphoinositides differentially regulate alpha-actinin flexibility and function Biochem J 2004 378 1067 1072 14670080 10.1042/BJ20031124
Shin JH Gardel ML Mahadevan L Matsudaira P Weitz DA Relating microstructure to rheology of a bundled and cross-linked F-actin network in vitro Proc Natl Acad Sci U S A 2004 101 9636 9641 15210969 10.1073/pnas.0308733101
Pelletier O Pokidysheva E Hirst LS Bouxsein N Li Y Safinya CR Structure of actin cross-linked with alpha-actinin: a network of bundles Phys Rev Lett 2003 91 148102 14611558 10.1103/PhysRevLett.91.148102
|
16336664
|
PMC1318456
|
CC BY
|
2021-01-04 16:31:30
|
no
|
BMC Cell Biol. 2005 Dec 8; 6:45
|
utf-8
|
BMC Cell Biol
| 2,005 |
10.1186/1471-2121-6-45
|
oa_comm
|
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-1261632975710.1186/1471-2458-5-126Research ArticlePatterns of alcohol drinking and its association with obesity: data from the third national health and nutrition examination survey, 1988–1994 Arif Ahmed A [email protected] James E [email protected] Texas Tech University Health Sciences Center, Department of Family & Community Medicine, Division of Health Services Research, Lubbock, TX, USA2 Mayo Clinic Family Medicine Program/Rochester, Department of Family Medicine, 406 West Main Street, Kasson MN, USA2005 5 12 2005 5 126 126 13 8 2005 5 12 2005 Copyright © 2005 Arif and Rohrer; licensee BioMed Central Ltd.2005Arif and Rohrer; 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 reports suggest that alcohol use may have a protective effect on obesity. This study explores association between obesity and alcohol consumption in the non-smoking U.S. adult population.
Methods
We analyzed data on a total of 8,236 respondents who participated in the Third National Health and Nutrition Examination Survey. Body mass index (weight-kg/height-m2) was derived from measured height and weight data and categorized into: normal weight, overweight, and obese. Alcohol consumption was measured using following measures: history of drinking, binge drinking, quantity of drinks/day, frequency of drinking, and average volume of drinks/week.
Results
Mean body mass index in this sample of non-smokers was 26.4 (95% CI: 26.1, 26.7). Approximately 46% of respondents were classified as current drinkers. Current drinkers had lower odds of obesity (Adjusted odds ratio = 0.73, 95% CI: 0.55, 0.97) as compared to non-drinkers. The odds of overweight and obesity were significantly greater among binge drinkers and those consuming four or more drinks/day. However, those who reported drinking one or two drinks per day had 0.46 (95% CI: 0.34, 0.62) and 0.59 (95% CI: 0.41, 0.86) times the odds of obesity, respectively. Similarly, the odds of obesity were significantly lower among those who reported drinking frequently and consuming less than five drinks per week. The association between overweight and other alcohol measures was less pronounced.
Conclusion
The results suggest further exploring the possible role of moderate alcohol drinking in controlling body weight in adults.
==== Body
Background
The proportion of the population that is overweight or obese has increased steadily in many developed nations including the U.S. [1,2]. The continual rise in the prevalence of obesity is of concern to public health agencies because of its potentially explosive effect on the health of individuals and the cost to the health care system [3,4]. Alcohol consumption is common in the U.S., with the past-month prevalence of consuming at least one drink is reported at fifty percent among 12 years and older [5]. Average intake of alcohol among adults in the U.S. is approximately 10% of the total daily energy intake [6]. Alcohol, which is the second most energy dense macronutrient consumed [7], is known to reduce oxidation of fat and favors fat storage which may result in weight gain [8]. However, previous studies have shown that despite the added calories, alcohol drinkers do not gain extra weight as compared to non-drinkers [6]. Other studies have reported a J-shaped relationship between alcohol consumption and body mass index and waist-to-hip ratio such that light to moderate drinking has beneficial effect in reducing weight whereas non-drinking and heavy or risky drinking having the opposite effect [9,10]. However, the inverse association between alcohol drinking and weight gain has been prominent among women with the results among men remaining inconclusive [9,11-14]. The aims of this study were to assess the cross-sectional relationship between alcohol drinking and obesity in a nationally representative sample of non-smoking adults, using different measures of alcohol consumption.
Methods
Sample selection
The National Health and Nutrition Examination Survey III is a large cross-sectional survey of non-institutionalized U.S. population conducted between 1988 and 1994 [15]. Data for this study were extracted from two data files: the adult questionnaire and the physical examination data file. Initially, a total of 19,618 respondents 18 years or older were selected. Those who were pregnant (n = 322) or whose information on lifetime drinking (n = 2850) or drinking in the previous 12 months (n = 2) were missing were excluded from the analysis. The analysis was limited only to lifetime non-smokers leaving a final sample size of 8,236.
Variables
The main dependent variable was body mass index (BMI = Weight in kg/height in m2). Weight and height were measured during physical examination of study participants. BMI was recorded from the physical examination data file and categorized into normal weight (BMI<25 kg/m2), overweight (BMI 25 kg/m2–29.9 kg/m2) and obese (BMI ≥30 kg/m2).
The main independent variable of interest was alcohol consumption. Alcohol consumption was measured using these questions: 1) History of drinking: in your entire life, have you had at least 12 drinks of any kind of alcoholic beverage (variable MAPE1)? In the past 12 months did you have at least 12 drinks of any kind of alcoholic beverage (variable MAPE2)?; 2) Quantity of drinking: on the average, on the days that you drank alcohol, how many drinks did you have a day (variable MAPE4)?, those consuming four or more drinks were classified as heavy drinkers; 3) Frequency of drinking (Number of drinking days): in the past 12 months, how many days of the year did you drink any alcoholic beverages (variable MAPE3s)?; 4) Average volume of drinking alcoholic beverages = frequency*quantity/52, and 5) Binge drinking: In the past 12 months, how many days of the year did you have 5 or more drinks on a single day (variable MAPE6S)? The variables MAPE1 and MAPE2 were used to classify respondents to non-drinker, ex-drinker, and current drinker categories. A drink was considered as a 12-oz beer, a 4-oz glass of wine, or an ounce of liquor. In NHANES III no distinction was made between different types of alcoholic drinks. The covariates included in the model were gender, age, race/ethnicity, marital status, poverty income ratio, education, rural/urban, self-rated health, and leisure time physical activity.
Statistical analysis
Descriptive statistics were used to describe the characteristics of the study sample by drinking status. The association between weight status and risky alcohol drinking, which included both heavy drinking and binge drinking, was modeled using logistic regression analysis. The association between BMI and different measures of alcohol consumption was assessed using chi-square statistics. Multinomial logistic regression analyses were used to assess relationship between the categorical BMI variable and alcohol consumption adjusted for all the covariates. In a multinomial logistic regression model the odds of associations for both the overweight and obese categories were compared simultaneously to the normal weight category which was used as a common referent. Odds ratios and 95 percent confidence intervals were computed and adjusted for the confounding effects of age, sex, race/ethnicity, poverty income ratio, education, marital status, rural/urban, self-rated health, and leisure time physical activities. Because of the complex survey design, STATA statistical software version 8.2 (College Station, TX), which allows incorporation of sampling weights, strata, and primary sampling units, was used to perform all the analyses.
Results
The overall prevalence of overweight (25 kg/m2<BMI< 30 kg/m2) and obesity (BMI≥30 kg/m2) in this sample of never smokers was 31.4 percent (95% CI: 29.5, 33.4) and 21.9 percent (95% CI: 20.0, 23.8), respectively. Although men had the highest prevalence of overweight 39.6 percent (95% CI: 36.1, 43.1), the prevalence of obesity was highest among women 24.9 percent (95% CI: 22.3, 27.4). Non-Hispanic Blacks had the highest prevalence of obesity (33.0 percent, 95% CI: 30.5, 35.5).
Approximately 46 percent of the sample was categorized as current drinkers; about ten percent were classified as heavy drinkers, consuming four or more drinks per day (Table 1). The prevalence of current drinking was almost twice among male non-smokers as compared to female non-smokers (63 percent vs 35 percent, p < 0.001). The mean (geometric mean) number of alcohol drinks consumed per day by the respondents was 2.3; 2.8 drinks/day among men and 1.9 drinks/day among women. Almost one-third of obese individuals were current drinkers.
Table 1 Characteristics of the study sample by drinking status- the third National Health and Nutrition Examination Survey, 1988–1994
Variables Non-Drinkers Ex-Drinkers Current Drinkers
(n = 2539) %a (n = 2755) %a (n = 2942) %a
All 24.1 30.0 45.9
Gender
Female 30.9 33.8 35.4
Male 13.6 23.9 62.5
Age
18–29 24.2 21.3 54.5
30–39 14.0 27.5 58.6
40–49 22.2 32.1 45.7
50–59 26.1 34.2 39.7
60–69 27.9 38.0 34.1
70+ 41.3 42.2 16.5
Race/Ethnicity
Non-Hispanic White 19.9 29.7 50.4
Non-Hispanic Black 31.7 35.0 33.3
Hispanic 28.8 29.6 41.6
Other Race/Ethnicity 59.2 19.6 21.2
Poverty Income Ratio
Above Poverty 22.5 29.8 47.7
Below Poverty 37.5 30.7 31.8
Education
High School or more 19.5 29.2 51.3
Less than High School 41.0 32.6 26.5
Marital Status
Married/Living as Married 22.5 32.0 45.5
Widowed/Divorced/Separated 28.0 38.2 33.8
Never Married 25.4 19.4 55.2
Location
Rural 29.1 32.3 38.6
Urban 19.6 27.7 52.7
Health
Excellent/Very Good/ Good 22.1 28.3 49.6
Fair/Poor 37.3 39.7 23.0
Leisure Time Physical Activityb
First Quartile (none) 40.2 34.2 25.6
Second Quartile (1–9) 24.7 32.5 42.8
Third Quartile (10–30) 19.4 30.2 50.4
Fourth Quartile (>30) 19.2 24.2 56.5
Body Mass Indexc
Normal Weight 23.5 25.8 50.7
Overweight 22.1 31.3 46.6
Obese 28.4 36.6 35.0
a Weighted percent
b Nine leisure time physical activities (walking, jogging/running, bicycle, swimming, aerobics, other dancing, calisthenics/exercises, garden/yard work, weight lift) were summed and categorized into four quartiles.
c Normal Weight (BMI <25 kg-m2), Overweight (25 kg-m2<BMI< 30 kg-m2), Obese (BMI ≥ 30 kg-m2)
Table 2 presents the cross-sectional relationship between obesity and drinking patterns. The odds of obesity among current drinkers were 0.73 times lower than the odds among non-drinkers (95%CI: 0.55, 0.97). Significantly greater odds of overweight and obesity were observed among those engaged in binge drinking. Similarly, those who reported drinking four or more drinks per day had 30 percent greater odds of being overweight (95%CI: 1.00, 1.68) and 46 percent greater odds of obesity (95%CI: 0.98, 2.17); however respondents who reported drinking one or two drinks/day had significantly lower odds of obesity. When examining the frequency of drinking days, significantly lower odds of obesity were observed among respondents in higher quartiles. Similarly, those who consumed less than five drinks/week had 0.62 times reduce odds of obesity (95%CI: 0.46, 0.82) as compared to non/ex-drinkers (Table 2). The results were less pronounced in the overweight category.
Table 2 Multiple logistic regression analysis of overweight and obesity among never smokers- the third National Health and Nutrition Examination Survey, 1988–1994
Overweight Obese
Odds Ratioa 95% CI Odds Ratioa 95% CI
History of Drinking
Non-Drinker 1.00 1.00
Ex-Drinker 1.14 (0.90–1.45) 1.10 (0.88–1.37)
Current Drinker 0.98 (0.71–1.35) 0.73 (0.55–0.97)
Binge Drinking
No Binge drinking 1.00 1.00
Yes Binge drinking 1.45 (1.02–2.05) 1.77 (1.18–2.65)
Non/Ex-Drinker 1.27 (0.95–1.71) 1.80 (1.30–2.50)
Quantity of Drinks/day
Non/Ex-Drinker 1.00 1.00
1 drink/day 0.71 (0.53–0.95) 0.46 (0.34–0.62)
2 drinks/day 0.83 (0.61–1.14) 0.59 (0.41–0.86)
3 drinks/day 1.40 (0.87–2.26) 1.07 (0.64–1.79)
4 or more drinks/day 1.30 (1.00–1.68) 1.46 (0.98–2.17)
Number of drinking days/year
Non/Ex-Drinker 1.00 1.00
1st quartile 1.06 (0.75–1.50) 1.02 (0.69–1.52)
2nd quartile 0.75 (0.59–0.96) 0.49 (0.35–0.68)
3rd quartile 0.99 (0.75–1.29) 0.64 (0.47–0.88)
4th quartile 0.84 (0.60–1.18) 0.61 (0.38–0.99)
Average Volume of Drinks/weekb
Non/Ex-Drinker 1.00 1.00
< 5 drinks/week 0.84 (0.65–1.08) 0.62 (0.46–0.82)
5–9 drinks/week 1.08 (0.73–1.58) 0.79 (0.51–1.23)
10–14 drinks/week 1.03 (0.60–1.76) 1.24 (0.66–2.36)
15 or more drinks/week 1.49 (0.98–2.26) 1.10 (0.63–1.91)
a Adjusted for age, sex, race/ethnicity, poverty income ratio, education, marital status, rural/urban, self-rated health, and leisure time physical activities. Nine leisure time physical activities (walking, jogging/running, bicycle, swimming, aerobics, other dancing, calisthenics/exercises, garden/yard work, weight lift) were summed and categorized into four quartiles.
b Average volume of drinking alcoholic beverages = frequency*quantity/52.
Discussion
This cross-sectional study found an inverse relationship between moderate consumption of alcohol and obesity in a large representative sample of non-smoking U.S. adults. Current drinkers had the lowest odds of obesity. A dose-response relationship was observed with increasing quantity of drinks and odds of obesity. Study participants reporting drinking one or two drinks per day had lower odds of obesity. The association was less pronounced in the overweight category.
The major limitation of our study lies in its cross-sectional nature which precludes establishing any cause and effect relationship. The strength of our study was that it includes anthropometric measurements of body mass index which makes our results more reliable and avoids potential bias that may result from self-reported height and weight data. However, we note that our findings confirm those reported in studies using self-reported height and weight. We used different measures of alcohol drinking habits to illuminate the role of alcohol drinking on obesity, which only few of the cross-sectional studies have done before. However, the accuracy of self-reported data on alcohol consumption is subject to debate. Assuming under-reporting of alcohol consumption by study respondents, it is likely that our study results are biased downwards. Measurement of alcohol use also has not been consistent in epidemiological studies. Differences in measurements can make cross-study comparisons difficult. Our study was limited to persons who never smoked; hence findings are not generalizable to smokers or past smokers. Use of odds ratio, instead of prevalence ratio, as an effect measure in a cross-sectional study of diseases of high prevalence has been shown to overestimate the true association [16]. Statistical methods to estimate prevalence ratio and its variance in a survey sample using a categorical outcome variable are not well developed; hence the odds ratio remain the effect measure of choice in such studies.
In this study, the odds of overweight and obesity were significantly higher among those who indulged in binge drinking and/or heavy drinking (consuming four or more drinks/day). In contrast, light-to-moderate drinking (consuming one or two drinks/day) was associated with lower odds of overweight and obesity. We observed a J-shape pattern, reported by others [9,10] among women binge drinkers and both men and women heavy drinkers (consuming four or more drinks/day). The odds of obesity were twice (Adjusted OR = 2.36, 95%CI: 1.58, 3.53) as likely among women binge drinkers and non/ex-drinkers (Adjusted OR = 2.09, 95%CI: 1.46, 2.99) as compared to non-bingers. No significant association was found among overweight women or overweight and/or obese men. For heavy drinkers, the adjusted odds of obesity were 2.40 (95%CI: 1.23, 4.58) for women and 2.25 (95%CI: 1.45, 3.48) for men heavy drinkers and 1.82 (95%CI: 1.32, 2.51) for women and 1.52 (95%CI: 1.00, 2.30) for men non/ex-drinkers, as compared to those who consumed less than fours drinks/day. In our earlier study of primary care patients we did not observe any association between obesity and binge drinking, possibly due to a small sample size [13]. Binge drinking is considered as high risk drinking and consumption of a high amount of alcohol has been associated with increased morbidity, mortality, and poor self-rated health [17-19]; approximately 8% of the U.S. adult population engages in risky drinking behavior [20]. In a prospective study of British adults, Wannamethee and Shaper [18], reported that heavy alcohol drinkers (≥ 30 g/d) had the highest prevalence of weight gain and obesity, irrespective of the type of alcohol consumed.
Other studies have shown an inverse association between increasing quantity of alcohol consumption and weight gain. In a study of Danish adults, Vadstrup et al [21] found an inverse association of waist circumference (measured ten years after the baseline) with total drinks of wine consumed per week. Those who consumed one to seven drinks per week had smallest waist circumference. No information on other measures of drinking was available. In a randomized controlled trial of German adults, Flechtner-Mors et al. [22] evaluated the effectiveness of energy restricted diet among moderate alcohol drinkers (consuming one/two drinks/day) and found a reduction in body weight among overweight and obese individuals. Breslow and Smothers [14] assessed the relationship between the continuous measure of BMI and four categories of quantity of drinking and also found a linear dose-response relationship between BMI and increasing quantity of drinking alcohol. Our results were similar but provided additional information that the beneficial effect of drinking disappears beyond consuming two drinks a day and may actually result in weight gain with heavy drinking.
The frequency of drinking was inversely related to obesity in this sample. Respondents consuming alcohol more frequently were less likely to be obese. In a recent study of primary care patients attending community clinics, we [13] found a significantly lower odds of obesity among those who consumed alcohol three or more days per month as compared to non-drinkers (Adjusted OR = 0.49, p = 0.037). Tolstrup et al [23] studied both the quantity and frequency of drinking in a Dannish population. The odds of obesity were lower among frequent drinkers, consuming alcohol seven days a week, as compared to those drinking less frequently. Similarly, Breslow and Smothers [14] also found the lowest BMI among persons who drank small quantities regularly. The effect was primarily observed among females. In our study the association was similar among males and females. The consistency of the inverse relationship observed between obesity and frequency of drinking suggest that the beneficial effect of drinking on obesity is present when the alcohol is consumed in moderate amounts on a regular basis.
We also explored association between overweight/obesity and average volume of drinking per week. Those drinking less than five drinks per week had lower odds of obesity. It has been argued that the association between BMI and alcohol consumption is obscured when average volume is used as a measure of alcohol consumption [14]. However, our results were consistent with the results obtained using other measures of alcohol consumption in the study.
Conclusion
The beneficial effect of moderate use of alcohol beverages on diabetes and other chronic diseases has been well established [24-27]. However, prospective epidemiological studies are needed to confirm if the same beneficial effects can be extended to the obesity. Actively promoting moderate use of alcohol as a strategy to combat obesity would be inappropriate at this early stage of our understanding about the underlying mechanisms that link alcohol use with weight control. Furthermore, it should be noted that the data give no evidence to advise non-drinkers to start drinking alcohol just for reducing body weight. However, the evidence reported here argues against a strategy of promoting complete abstention at-least among those who regularly consume alcohol.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AAA carried out the study, performed statistical analyses and drafted the manuscript. JER participated in the design of the study and drafting of the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Centers for Disease Control and Prevention: National Center for Chronic Disease Prevention and Health Promotion. Overweight and Obesity Accessed December 15, 2004
Kumanyika S Jeffery RW Morabia A Ritenbaugh C Antipatis VJ Obesity prevention: the case for action Int J Obes Relat Metab Disord 2002 26 425 436 11896500 10.1038/sj.ijo.0801938
Finkelstein EA Fiebelkorn IC Wang G State-level estimates of annual medical expenditures attributable to obesity Obes Res 2004 12 18 24 14742838
Elmer PJ Brown JB Nichols GA Oster G Effects of weight gain on medical care costs Int J Obes Relat Metab Disord 2004 28 1365 1373 15356665 10.1038/sj.ijo.0802774
Office of Applied Studies [OAS], 2004c, Table 2.45B Accessed August 3, 2000
Jequier E Alcohol intake and body weight: a paradox Am J Clin Nutr 1999 69 173 174 9989676
Prentice AM Alcohol and obesity Int J Obes Relat Metab Disord 1995 19 S44 50 8581112
Zurlo F Lillioja S Esposito-Del Puente A Nyomba BL Raz I Saad MF Swinburn BA Knowler WC Bogardus C Ravussin E Low ratio of fat to carbohydrate oxidation as predictor of weight gain: study of 24-h RQ Am J Physiol 1990 259 E650 657 2240203
Colditz GA Giovannucci E Rimm EB Stampfer MJ Rosner B Speizer FE Gordis E Willett WC Alcohol intake in relation to diet and obesity in women and men Am J Clin Nutr 1991 54 49 55 2058587
Lukasiewicz E Mennen LI Bertrais S Arnault N Preziosi P Galan P Hercberg S Alcohol intake in relation to body mass index and waist-to-hip ratio: the importance of type of alcoholic beverage Public Health Nutr 2005 8 315 320 15918929 10.1079/PHN2004680
Williamson DF Forman MR Binkin NJ Gentry EM Remington PL Trowbridge FL Alcohol and body weight in United States adults Am J Public Health 1987 77 1324 1330 3498373
Liu S Serdula MK Williamson DF Mokdad AH Byers T A prospective study of alcohol intake and change in body weight among US adults Am J Epidemiol 1994 140 912 920 7977278
Rohrer JE Rohland BM Denison A Way A Frequency of alcohol use and obesity in community medicine patients BMC Fam Pract 2005 6 17 15845150 10.1186/1471-2296-6-17
Breslow RA Smothers BA Drinking patterns and body mass index in never smokers: National Health Interview Survey, 1997–2001 Am J Epidemiol 2005 161 368 376 15692081 10.1093/aje/kwi061
National Center for Health Statistics Plan and operation of the Third National Health and Nutrition Examination Survey, 1988–94 Vital and health statistics, series 1: programs and collection procedures no 32 DHHS publication no (PHS) 94-1308 (GPO no 017-022-01260-0) 1994 Hyattsville, MD: National Center for Health Statistics
Behrens T Taeger D Wellmann J Keil U Different methods to calculate effect estimates in cross-sectional studies. A comparison between prevalence odds ratio and prevalence ratio Methods Inf Med 2004 43 505 509 15702210
Gronbaek M Deis A Sorensen TI Becker U Borch-Johnsen K Muller C Schnohr P Jensen G Influence of sex, age, body mass index, and smoking on alcohol intake and mortality Bmj 1994 308 302 306 8124118
Wannamethee SG Shaper AG Alcohol, body weight, and weight gain in middle-aged men Am J Clin Nutr 2003 77 1312 1317 12716687
Okosun IS Seale JP Daniel JB Eriksen MP Poor health is associated with episodic heavy alcohol use: evidence from a National Survey Public Health 2005 119 509 517 15826892 10.1016/j.puhe.2004.08.016
Coups EJ Gaba A Orleans CT Physician screening for multiple behavioral health risk factors Am J Prev Med 2004 27 34 41 15275672 10.1016/j.amepre.2004.04.021
Vadstrup ES Petersen L Sorensen TI Gronbaek M Waist circumference in relation to history of amount and type of alcohol: results from the Copenhagen City Heart Study Int J Obes Relat Metab Disord 2003 27 238 246 12587005 10.1038/sj.ijo.802203
Flechtner-Mors M Biesalski HK Jenkinson CP Adler G Ditschuneit HH Effects of moderate consumption of white wine on weight loss in overweight and obese subjects Int J Obes Relat Metab Disord 2004 28 1420 1426 15356671 10.1038/sj.ijo.0802786
Tolstrup JS Heitmann BL Tjonneland AM Overvad OK Sorensen TI Gronbaek MN The relation between drinking pattern and body mass index and waist and hip circumference Int J Obes Relat Metab Disord 2005 29 490 497
National Institute on Alcohol Abuse and Alcoholism State of the Science Report on the Effects of Moderate Drinking National Institutes of Health, Department of Health and Human Services December 19, 2003
Kaplan MS Huguet N Newsom JT McFarland BH Lindsay J Prevalence and correlates of overweight and obesity among older adults: findings from the Canadian National Population Health Survey J Gerontol A Biol Sci Med Sci 2003 58 1018 1030 14630884
Smothers B Bertolucci D Alcohol consumption and health-promoting behavior in a U.S. household sample: leisure-time physical activity J Stud Alcohol 2001 62 467 476 11513224
Fine LJ Philogene GS Gramling R Coups EJ Sinha S Prevalence of multiple chronic disease risk factors. 2001 National Health Interview Survey Am J Prev Med 2004 27 18 24 15275670 10.1016/j.amepre.2004.04.017
|
16329757
|
PMC1318457
|
CC BY
|
2021-01-04 16:28:59
|
no
|
BMC Public Health. 2005 Dec 5; 5:126
|
utf-8
|
BMC Public Health
| 2,005 |
10.1186/1471-2458-5-126
|
oa_comm
|
==== Front
BMC BiochemBMC Biochemistry1471-2091BioMed Central London 1471-2091-6-261630955510.1186/1471-2091-6-26Research ArticleReduced Flavin: NMR investigation of N(5)-H exchange mechanism, estimation of ionisation constants and assessment of properties as biological catalyst Macheroux Peter [email protected] Sandro [email protected] Christoph [email protected]üterjans Heinz [email protected]üller Franz [email protected] Graz University of Technology, Institute of Biochemistry, Petersgasse 12, A-8010 Graz, Austria2 Fachbereich Biologie, Universität Konstanz, D-78457 Konstanz, Germany3 Institut fur Biophysikalische Chemie, J.W. Goethe-Universität, Biozentrum N230, Marie-Curie-Strasse 9, D-60439 Frankfurt am Main, Germany4 Wylstrasse 13, CH-6052 Hergiswil, Switzerland2005 25 11 2005 6 26 26 27 6 2005 25 11 2005 Copyright © 2005 Macheroux et al; licensee BioMed Central Ltd.2005Macheroux 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 flavin in its FMN and FAD forms is a versatile cofactor that is involved in catalysis of most disparate types of biological reactions. These include redox reactions such as dehydrogenations, activation of dioxygen, electron transfer, bioluminescence, blue light reception, photobiochemistry (as in photolyases), redox signaling etc. Recently, hitherto unrecognized types of biological reactions have been uncovered that do not involve redox shuffles, and might involve the reduced form of the flavin as a catalyst. The present work addresses properties of reduced flavin relevant in this context.
Results
N(5)-H exchange reactions of the flavin reduced form and its pH dependence were studied using the 15N-NMR-signals of 15N-enriched, reduced flavin in the pH range from 5 to 12. The chemical shifts of the N(3) and N(5) resonances are not affected to a relevant extent in this pH range. This contrasts with the multiplicity of the N(5)-resonance, which strongly depends on pH. It is a doublet between pH 8.45 and 10.25 that coalesces into a singlet at lower and higher pH values. From the line width of the 15N(5) signal the pH-dependent rate of hydrogen exchange was deduced. The multiplicity of the 15N(5) signal and the proton exchange rates are little dependent on the buffer system used.
Conclusion
The exchange rates allow an estimation of the pKa value of N(5)-H deprotonation in reduced flavin to be ≥ 20. This value imposes specific constraints for mechanisms of flavoprotein catalysis based on this process. On the other hand the pK ≈ 4 for N(5)-H protonation (to form N(5)+-H2) would be consistent with a role of N(5)-H as a base.
==== Body
Background
The isoalloxazine ring system is the redox active moiety of the coenzyme forms (FMN or FAD) present in flavoenzymes. These are involved in a variety of biological processes, spanning a wide spectrum with regard to the underlying chemical reaction mechanisms. These range from the classical (de)hydrogenation, the uptake, release and transport of electrons, the production of light (bioluminescence), photochemistry of the reduced form (as in photolyases), light signal transduction (as in blue light receptors), activation of oxygen and redox sensing, to name only the most prominent ones [1]. In addition to these functions, others have emerged that appear to require hitherto unrecognized roles of reduced flavin in chemical catalysis, such as reactions that are redox-neutral (for a review see [2]). These will be addressed briefly below.
In reduced flavin, N(5) is crucial for the functioning of the isoalloxazine system as it is the locus involved in uptake/release of redox equivalents and is in general in contact with reacting ligands. In the intermediate pH range reduced flavin N(5) exists in its neutral, N(5)-H form, it can be protonated to yield N(5)H2+ at low pH and might exist in the anionic form N(5)- at very high pH (cf Fig. 5, below). Based on kinetic arguments [3] Bruice and co-workers have estimated the pKa for deprotonation of this group as being around 20 in free flavin. Urban and Lederer, on the other hand, imply a value around 15 for flavocytochrome b2 [4] and it has been postulated that this pK might even be lower than 7 [5]. A method useful for the assessment of the properties of reduced flavin, and specifically of N(5) is the nuclear magnetic resonance (NMR) spectroscopy. It has been used to study the interactions of apoprotein and flavin, and the perturbations of these interactions induced by binding of substrate/ligands[6]. In these studies, we have observed that the N(5)H group in most two-electron reduced flavoproteins appears as a doublet in the 15N-NMR spectrum due to the N(5)-H coupling, while free flavin exhibits a singlet in the pH range 5–8 due to fast proton exchange. Therefore, the doublets observed in reduced flavoproteins have been interpreted as resulting from the absence or from slow proton exchange on the NMR time scale. This can result from inaccessibility of the N(5)H group to bulk water. However, no systematic study on the basic mode of N(5)-H exchange in free reduced flavin in aqueous solution is available. The great variety of chemical reactions mentioned above raises the question about the physical interactions between the apoprotein and the coenzyme that are responsible for the tuning necessary to catalyze particular reactions.
A detailed knowledge of this process and of its mechanism would provide several insights into flavoenzyme structure and function: a) It would provide a basis for the interpretation of 15N-NMR spectra of reduced flavoproteins. b) It could help understand a variety of exchange processes of substrate/product-linked hydrogens in various dehydrogenation reactions involving flavoproteins. c) It might help clarify the possible role of reduced flavin in the β-elimination of halide from β-halogenated substrates catalyzed by several flavoproteins. X-ray structural information has shown that there are no basic amino acid residues at the active centers of e.g. (oxidized) D-amino acid oxidase [7-9]. The hypothesis has thus been put forward that N(5) of the reduced enzyme flavin is the base that is involved in enzyme catalyzed elimination [10]. In addition labeled hydrogen is abstracted from the α position and is partially incorporated into the β position of product [11]. These data would also be compatible with the proposed role of reduced flavin N(5)-H being involved in label transposition/exchange [7-9,12]. d) Chorismate synthase catalyzes a redox-neutral anti-1,4-elimination of a hydrogen and a phosphate group and has an absolute requirement for a reduced FMN cofactor. Again, it has been assumed that an amino acid functional group serves as a base in the elimination of the hydrogen but the recent elucidation of the 3D-structure has revealed that the only functional group that could be invoked in this process is the N(5) position of the flavin [12,13]. e) Based on recent studies it has been proposed that in the dehydrogenation reaction catalyzed by monoamine oxidase, the flavin N(5) of the postulated C(4a)-flavin substrate adduct (which is isoelectronic with the flavin hydroquinone) acts as a base in hydrogen abstraction [14]. Thus, a critical functional role of the flavin N(5) nitrogen in its reduced 1,5-dihydro, or its isoelectronic 4a,5-dihydro forms appears to emerge. The present study was undertaken to increase our understanding of processes involving this position and specifically the pH-dependent proton exchange and to gain information on the basicity/nucleophilicity of N(5) in reduced flavin, which provides basic information on possible mechanistic roles in flavoenzyme catalysis.
Results and Discussion
Chemical shifts
The pH-dependence of the 15N chemical shifts of reduced free flavin (N(l), N(3), N(5), N(10)) has been previously investigated at pH 5.2 – 8 [15], a pH range where most flavoproteins are active and stable. This study revealed that the chemical shift of the N(1) atom in reduced free flavin is strongly pH-dependent due to its ionization. A pKa value of 6.8 was calculated for this process [15]. Also the signal of the N(10) atom shows a similar pH-dependence; the chemical shift difference between that of the neutral and of the anionic species, however, is smaller than that of the N(1) atom. This observation was rationalized previously as resulting from a change of sp2 hybridization of the N(10) atom [15,16]. The chemical shifts of the N(3) and the N(5) atom are practically independent of pH in the range studied. In aqueous solution only a singlet was observed for the N(5)H and the N(3)H groups. The signal of the N(3)H group remains as a sharp, narrow line over the pH range studied, whereas that of the N(5)H group exhibits a sharp line at pH 5.4 and a broadened one at pH 7 and higher [15,16].
In the present study the pH-dependence of the 15N chemical shifts of reduced free flavin was extended to the pH range 4.0 to 12.3. The 15N chemical shifts at pH 4.0 are identical, or almost identical to those observed earlier at pH 5.0 [15,16], with the exception that the chemical shift of the N(1) atom is shifted upfield by 1.4 ppm. Also at pH 12.3 the 15N chemical shifts of the N(5) and N(10) atoms are identical with those reported previously for pH 8.0 [15,16]. The 15N chemical shift of the N(3) atom shows a slight downfield shift (+1.9 ppm) at pH >11 indicating the onset of deprotonation of the N(3)H group. Also a small downfield shift (+0.7 ppm) is observed for the N(1) atom in the high pH region, probably related to the (partial) ionization of the N(3)H group.
Signal structure of the N(5)-H resonance
The two nitrogen atoms in reduced free flavin bearing a hydrogen are N(3) and N(5). The signal of N(3) appears as a singlet in the pH range studied and from this rapid exchange can be assumed. As shown in Fig. 1 the shape of the N(5) resonance signal depends strongly on the pH. At low pH values there is a sharp singlet that broadens with increasing pH. Above pH 7.15 the broad singlet begins to show a doublet structure (first coalescence point), the resolution of which increases and finally gives rise to a doublet (pH 8.45 – 10.25). Further increase of the pH leads to broadening of the doublet lines again and eventually to a second coalescence point around pH 10.5 (see also below). At even higher pH values the signal becomes again a sharp singlet. These data demonstrate that below the first and above the second coalescence point we have two pH regions of fast hydrogen exchange. This is supported by the fact that the line widths of the decoupled and undecoupled resonance lines are identical (4 Hz, natural line width) at pH 12.3 and pH 5.0. In the pH range between the two coalescence points the exchange is comparatively slow. The asymmetric shape of the doublet (higher intensity of the high field line in Fig. 1, pH 8.45 – 10.25) can be a result of either the high magnetic field (11.4 Tesla), which causes a prevailing contribution of the CSA-relaxation mechanism [17] or could be due to a particular proton exchange mechanism [18]. Which of the two effects is prevailing, cannot be deduced from the present data.
Figure 1 Multiplicity of the 15N-NMR signal of the reduced FMN N(5)H group as a function of pH. The flavin concentration was 5 mM in 250 mM Tris + 100 mM NaCl. The obtained FID was processed further by exponential multiplication using a line broadening factor of 20 Hz in order to improve the signal to noise ratio.
The 1J15 N(5)-H coupling constant can be determined from the spectra obtained in the slow exchange region (Fig. 1, spectra 7 and 8) and was found to be 85 Hz. This value agrees well with that determined previously in a CHCl3 solution (87.5 Hz) [15]. The coupling constant of reduced flavin is similar to those reported for eneamines and aniline derivatives [19]. 1J15N-H coupling constants are mainly governed by the hybridization of the nitrogen atom. A low s-character gives rise to small coupling constant (the coupling constant for tetrahedral ammonia is 73 Hz) whereas a high s-character results in a high coupling constant (for linear nitriles coupling constants as high as 135 Hz have been found [20]). The observed coupling constant of 85 Hz, which is similar to the value found for pyrrole (96.5 Hz [20]), indicates that the N(5) atom possesses approximately 31 % s-character, i.e. it is highly sp2 hybridized. The pH-dependence of the shape of the N(5)-resonance was not affected by buffer systems like Tris, phosphate and borate.
Determination of N(5)-Hydrogen exchange rates
The exchange rate of the N(5) hydrogen was determined using equation (1) for fast exchange (below pH 7.8 and above pH 10.5) [21] and equation (2) for slow exchange (between pH 7.8 and 10.5) [22]:
k = 1/τe = 4 π pA pB Δν2/(Δν1/2 - Δν1/2°) (1)
k = 1/τe = π (Δν1/2 - Δν1/2°) (2)
with k as the exchange rate; τe the lifetime of each state, Δν1/2 the half line width of the resonance line, Δν1/2° the half line width of the proton decoupled signal which should be equal to the so-called natural line width (4 Hz, see above) under conditions of fast proton exchange. Δν is the difference of the resonance frequencies of the exchanging species under the condition of slow exchange, i.e. equal to the coupling constant 1JN-H, pA and pB are the molar fractions of the exchanging species in state A and B, respectively. The exchange rates have also been determined in phosphate and borate buffers at a few selected pH values; they are in the range of those determined in Tris buffer. The signal structure at pH 10.5 (transition from doublet to singlet) indicates that the coalescence point is close to this pH value. This is supported by the fact that the exchange rates at this pH, calculated according to eqs. 1 and 2, yield very similar values, i.e. 191 s-1 and 186 s-1, respectively.
Possible mechanisms of exchange
The values obtained for the exchange rate in Tris buffer are summarized in Table 1 and are plotted versus pH in Fig. 2. The semi-logarithmic representation (panel A) shows that the exchange rate first decreases with increasing pH, it reaches a minimum around pH 9.5, this being followed by a steeper increase. The double-logarithmic representation (panel B) shows that the two segments have different profiles at low and high pH: The curve in this panel was generated based on the rules of Dixon [23] for the pH dependence of kinetic parameters of a species that is present in a pH dependent equilibrium linked by specific pK's and includes reactions with linear dependences from either H+ or OH-, or no dependence. For the "Dixon analysis" several modi of exchange can be envisaged, 4 of which (2b, 3a, 3b, and 3c,) are relevant for the pH range investigated in this study (Fig. 3). These include the reduced flavin at its 4 possible ionisation states along with the structures involved. In the present case the species undergoing ionisation is reduced flavin with a pKa ≈ 7 for the N(1)-H group [15]. At very low pH (Fig. 3, 1) the reduced flavin will be protonated at N(5) (pK ≈ -1.2, [24]) and exchange will be with H2O. This process (Fig. 3, 1) is not expected to play a role in the present case. Similarly, the [H3O+] dependent mechanism (2a) is probably not relevant since the data suggest that a corresponding slope is absent at pH <6. The neutral reduced flavin can exchange by reaction (2b) with H2O (Fig. 2B, horizontal segment up to pH ≈ 7). At pH 7–9 the [H3O+] dependent process (3a) then appears to become prominent (linear segment with slope = 1). The profile in Fig. 2B clearly indicates that at higher pH [OH-] dependent processes become dominant. This change in mechanism would be initiated by the switch of mode (2b) to the pH independent mechanism (3b)(short horizontal section between pH 9–10, Fig. 2B). This phase is envisaged to be followed by the [OH-] dependent mode (3c) (section at pH >10, linear increase with [OH-]). In Fig. 3, the equilibrium between species (2) (3) that encompasses the structures included by the square brackets, reflects the pH dependent change in ionisation state of the reduced flavin with a pKa ≈ 7. The point should be stressed that in the representation of Fig. 2B and according to the Dixon's rules only the "break" at ≈ pH 7 can be attributed to a microscopic pK. That corresponding to pH ≈ 8.7 is taken to reflect the change in mechanisms as outlined above, i.e. it is an apparent pK and has no structural significance since it cannot be attributed to ionisations of the exchange partners or of the buffer. The same probably holds true also for the break corresponding to a pH ≈ 10 since the pKa for the second ionisation of reduced flavin (N(3)-H) is estimated as ≈ 14.
Table 1 Line widths of the 15N resonance signal of the reduced FMN N(5) atom as function of pH and estimation of the N(5)-H pKa.
pH Line width (Δν1/2-Δν1/2°) (Hz) Exchange rate of N(5)-H (s-1) pKa (calculated)
6 21 525
6.7 42 270
7.15 47 242
7.8 34 107
8.95 18 57
9.05 6 19
9.7 3 9 20.5
10.25 14 44 20.3
10.5 61 191 20.0
11 33.5 339 20.2
The flavin was 5 mM in 250 mM Tris + 100 mM NaCl. The estimation of the pKa value is based on the line width of the corresponding 15N signal as described in the text.
Figure 2 pH dependence of the rate of N(5)-H exchange. Conditions as detailed in the Legend of Fig. 1 and in the Experimental section. Panel (A): Semi logarithmic plot. The line through the data points was generated with a polynomial algorithm and has no specific meaning. Panel (B): Same data as in panel (A), however logarithmic representation of the exchange rate. The curve (—) through the data points was generated according to Dixon's criteria and using a pH independent rate = 102.7 (s-1), the pKa = 7.15 of reduced flavin (ionisation at N(1)-H), and apparent pK's = 8.7 and 9.8. The dashed lines (---) represent the single, unperturbed reactions that correspond to the processes described by equations 2–4 in Fig. 3. Their "breaks" occur at changes in ionisation state (pKa) or reflect changes in mechanism. See text for further details.
Estimation of ionisation constants for the reduced flavin N(5)
The pKa of the flavin N(5)-H corresponding to the following exchange process:
can be estimated based on the general rates of proton transfer (k1 and k-1, as indicated in equation 3) between two exchanging species connected to the equilibrium constants Kflavin-N(5) and KH2O as in equation 4:
Kflavin-N(5)=k1k−1⋅KH2O (4)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGlbWsdaWgaaWcbaGaeeOzayMaeeiBaWMaeeyyaeMaeeODayNaeeyAaKMaeeOBa4Maeeyla0IaeeOta4KaeeikaGIaeeynauJaeeykaKcabeaakiabg2da9maalaaabaGaee4AaS2aaSbaaSqaaiabigdaXaqabaaakeaacqqGRbWAdaWgaaWcbaGaeyOeI0IaeGymaedabeaaaaGccqGHflY1cqqGlbWsdaWgaaWcbaGaeeisaGKaeeOmaiJaee4ta8eabeaakiaaxMaacaWLjaWaaeWaaeaacqaI0aanaiaawIcacaGLPaaaaaa@4C60@
The exchange rate in the thermodynamically favorable direction can reasonably be assumed as being diffusion controlled (in this case k-1 = 1010 M-1·s-1). For an estimation the HO-catalyzed process was selected since it is assumed to involve direct interaction with HO- (3c, in Fig. 3) and appears to best approximate a linear dependence from pH. Thus the data between pH 9.7 and 11 were employed. The values summarized in Table 1 were obtained using the mentioned value for a diffusion controlled proton transfer rate and the equilibrium constant of water (= 10-15.74 [25]). The average pKa value obtained from Table 1 is 20.2 (± 0.3).
Figure 3 Species involved in H-exchange at the reduced flavin position N(5) at its various stages of ionisation. (1) represents the exchange of cationic N(5). Species (2) involves exchange of neutral reduced flavin with H+ or H2O. (3) represent the modi of exchange of (mono-) anionic reduced flavin with H+, H2O or HO-. (4) and (5) are di- and tri-anionic reduced flavins that are likely to exchange with H2O.
An estimation of the pKa of the flavin N(5)-H can also be derived from the chemical shift of the 15N(5) center. 1J15 N-H coupling constants are mainly governed by the hybridization of the nitrogen atom. Since the N(5)-H coupling constant is similar to that reported for eneamines and aniline derivatives [22] it is reasonable to assume that the N(5) hybridization is similar also to that of a series of substituted anilines that are used in the correlation diagram of Fig. 4, where their chemical shift is plotted against their reported pKa values. From the chemical shift for the reduced flavin N(5) of ≈ 60 ppm [15] a pK >25 is obtained. While this value is higher than that derived from the exchange experiments it correlates with the chemical entity of a phenylene diamine substituted with a moderately electron-deficient (neutral) pyrimidine moiety. Inspection of the chemical structures in the lowest row in Fig. 3 shows that in a chemical system the species that deprotonates at position N(5)-H is a phenylene diamine linked to a pyrimidine carrying (already) two negative charges. The latter would undoubtedly increase the pKa for formation of the N(5) anion compared to a species in which neutral reduced flavin would ionize at position N(5)-H. The latter might be the case in a protein environment where specific charges or H-bridges might affect microscopic pK's. Nevertheless protein induced pK shifts with values >10 units are unlikely within the same molecule since it would require a stabilisation energy >59 KJ/Mol. A pKa value of ≥ 20 is also in good agreement with the estimation by Venkataram & Bruice (pKa = 19 – 23 [3]) deduced from the pH dependence of the rates of decay of reduced flavin model compounds that have the 4a,5-dihydro structure.
Figure 4 Correlation of 15N-chemical shifts of aniline derivatives with their pK's and extrapolation of the pKa value for N(5)H of reduced FMN based on its 15N chemical shift. The compounds are: (1) aniline, 2: 2-azaaniline; (3) 4-cyanoaniline; (4) 4-azaaniline; (5) 2,5-diazaaniline; (6) 4-nitroaniline; (7) 2-nitroaniline; (8) 4-chloro-2-nitroaniline; (9) 2,3-dinitroaniline (data from [34]). The horizontal arrow indicates the experimentally determined 15N chemical shift of N(5)H of reduced flavin and vertical arrow extrapolates to an estimated pKa 26–30.
Figure 5 Selected ionisations and tautomeric forms of reduced flavin. (A) and (B) are the anionic, respectively the neutral forms of reduced flavin encountered at the active center of flavoproteins. The pK interconnecting (A) and (B) is ≈ 7 in the free state [26], but can vary strongly at the active center of proteins. Protonation of (A) at N(5) to yield (C) can not be observed in the free state but this chromophore is obtained by twofold alkylation at N(5) [26]. The pK's linking species (B), (C) and (D) have been estimated from the properties of appropriate model compounds [26]. See text for further details.
The second ionisation of the reduced flavin relevant in an enzymological context is the (de)protonation at position N(1) with a pK ≈ 7 (species (A) (B), in Fig. 5). However, the locus of catalytic action is not N(1), but N(5). Thus, while the ionisation at N(1) plays a very important role in modulating the properties of reduced flavin and the flavin redox potential, it is unlikely to participate directly in base catalysis. A third ionisation that most likely is relevant in the biochemical context is the protonation of anionic reduced flavin (species (A) in Fig. 5) at position N(5)-H to yield the zwitterionic species (C). This species is not observed in free solution since the tautomeric equilibrium between species (B) and (C) favor the former by ≥ 2 orders of magnitude [26]. The N(5)-dialkylated analog to species (C), however, has been described [26] and is converted to (D) with a pK ≈ 4. The combination of the equilibria shown in Fig. 5 thus allow the estimation of the equilibrium between species (A) and (C) (Fig. 5) to correspond to a putative pK >4. While this "pK" has no direct significance in the free system it is easily conceivable that the protein could stabilize a negative charge at position N(1) as is probably the case with flavodoxins [15,27,28], thereby facilitating formation of a putative species (C). In this context it should be noted that the most nucleophilic functional group in (A) is not N(1)-C(2) = O, the locus of the negative charge, but N(5) as reflected by the main position of alkylation [26].
Conclusion
Although a doublet for N(5)H of reduced free flavin can only be observed in the narrow pH range 8.5 – 10.5, it has been documented for reduced flavoproteins at pH values as low as 5 [15,27,28]. This suggests that in specific, reduced flavoproteins N(5)-H exchange is slow and that access of bulk solvent to this position is hindered. Occurrence of a doublet was observed in reduced thioredoxin reductase in which the N(1) atom is protonated [29]. Conversely, in reduced flavodoxins and many other flavoproteins the flavin N(1) position is not accessible for protonation [30] the flavin existing in the anionic form as long as the protein does not unfold (pH-dependent process). It thus appears that the rate of exchange at N(5)-H is not dependent on the ionisation state at N(1)-H but will be dictated by the environment and accessibility of solvent of the specific protein. However, no such conclusion can be drawn for the pH range 8.5 – 10.5 where the doublet of the N(5)-H is an inherent property of the reduced flavin. Unlike the N(5)H resonance in reduced free flavin the N(3)H resonance shows a narrow singlet over the whole pH range studied, indicating that proton exchange is fast. Therefore, the appearance of a doublet for this resonance in flavoproteins is compatible with hindered accessibility of this position for bulk water as previously interpreted [6].
The estimation of the pKa value of ≥ 20 for the N(5) group in reduced flavin might be an useful parameter for the formulation of mechanisms involving it. For instance Urban and Lederer [4,5] have postulated a deprotonated N(5) with a pKa of about 15 or even as low as 7 [5] as a catalytically relevant species in flavocytochrome b2. While our data cannot exclude such a pKa, it imposes specific energetic restraints for its formulation. On the other hand the reduced flavin N(5)-H with an estimated pKb, ≥ 4 for the free molecule possesses sufficient basicity to act as a base according to mechanism 3b (Fig. 3). Clearly, this property can be modulated by the protein environment to suit specific purposes. The conclusion is thus that the reduced flavin N(5)-H has properties that qualify it as a base catalyst for biochemical reactions. This could be realized in several cases: The recent elucidation of the structure of chorismate synthase [12,13] suggests that N(5)-H of the reduced flavin cofactor is involved in the abstraction of the hydrogen in the C(6) pro-R position of the substrate 5-enolpyruvylshikimate 3-phosphate. A similar mechanism, abstraction of a proton from the α-CH2 (pro-R) of an amine by monoamine oxidase [14], seems to be emerging, replacing the previously proposed single electron transfer mechanism. Likewise, N(5) appears to be involved in the elimination of halide from β-Cl-alanine catalyzed by D-amino acid oxidase [9]. At the active sites of these enzymes there is no amino acid functional group that could participate in the required base catalysis. The data presented in this paper thus sustain the notion that N(5) of the reduced flavin cofactor can play a direct role in the catalysis of some flavin-dependent enzymes, an involvement that has not yet been addressed in sufficient detail. Our studies towards a characterization of the hydrogen exchange processes at this position therefore lay the foundation for a critical assessment of this role.
Methods
Experimental details
The synthesis and purification of [1,3,5,10-15 N4]-7-methy1-10- ribitylisoalloxazine-5'-phosphate was described previously [31,32]. 15N-NMR measurements were performed at 15°C, if not otherwise stated, with a Bruker AM 500 NMR instrument equipped with an Aspect 3000 and a temperature control unit. NMR-spectra were recorded with 12 μs pulses (= 30° flip angle) and a relaxation delay of 2 sec. In a typical experiment 10 mm Wilmad precision NMR tubes contained 1.8 ml of a 3 to 8 mM solution of flavin (or otherwise indicated in the text) in 250 mM Tris/100 mM NaCl-buffer and 0.2 ml D2O for field frequency lock. Reduction of the flavin solution was achieved by flushing the sealed NMR tube with argon before a two to three fold excess of a concentrated dithionite solution was added with a syringe. The pH of the resulting reduced solution was measured after the NMR experiment with a pH meter from Radiometer (Copenhagen, Sweden) equipped with a glass electrode from Ingold (Frankfurt, Germany). All 15N chemical shifts are expressed relative to liquid ammonia at 25°C and are corrected for bulk volume susceptibility. Neat CH3 15NO3 (™(CH3NO3) - ™(NH3) = 381.9 ppm) was used as an external standard according to Witanowski et al [33]. The proton exchange rate (1/τe) was calculated according to Gutowsky et al [34] in the case of a fast exchange reaction, and according to Grunwald et al [35] in the case of a slow exchange reaction.
Authors' contributions
PM carried out most NMR experimental work with CS and evaluated the NMR data, he participated in drafting the manuscript.
SG conceived the study, participated in its design and coordination, as well as in the evaluation of the kinetic data and drafting of the manuscript.
FM synthesized the labeled flavins used in the cooperative studies with the group of HR, and was involved in the drafting and finalization of the paper.
==== Refs
Ghisla S Massey V Mechanisms of Flavoprotein-catalyzed Reactions Eur J Biochem 1989 181 1 17 2653819 10.1111/j.1432-1033.1989.tb14688.x
Bornemann S Flavoenzymes that catalyse reactions with no net redox change Nat Prod Rep 2002 19 761 772 12521268 10.1039/b108916c
Venkataram UV Bruice TC On the Mechanism of Flavin-Catalyzed Dehydrogenation a,β to an Acyl Function. The Mechanism of 1,5-Dihydroflavin Reduction of Maleimides J Am Chem Soc 1984 106 5703 5709 10.1021/ja00331a047
Urban P Lederer F Intermolecular hydrogen transfer catalyzed by a flavodehydrogenase, bakers' yeast flavocytochrome b2 J Biol Chem 1985 260 11115 11122 3897228
Lederer F Extreme pKa displacements at the active sites of FMN-dependent alpha-hydroxy acid-oxidizing enzymes Protein Sci 1992 1 540 548 1338973
Müller F Müller F Nuclear magnetic resonance studies on flavoproteins Chemistry and Biochemistry of Flavoproteins 1992 III Boca Raton, Florida: CRC Press 558 595
Umhau S Pollegioni L Molla G Diederichs K Welte W Pilone MS Ghisla S The x-ray structure of D-amino acid oxidase at very high resolution identifies the chemical mechanism of flavin-dependent substrate dehydrogenation Proc Natl Acad Sci USA 2000 97 12463 12468 11070076 10.1073/pnas.97.23.12463
Mattevi A Vanoni MA Curti B Structure of D-amino acid oxidase: new insights from an old enzyme Current Opinion in Structural Biology 1997 7 804 810 9434899 10.1016/S0959-440X(97)80150-9
Molla G Pilone MS Pollegioni L Ghisla S Chapman S, Perham R, Scrutton N Studies on the elimination reaction of Rhodotorula gracilis D-amino acid oxidase with β-chloro-D-alanine Flavins and Flavoproteins 2002: Proc 14th Int Symp, Cambridge, UK, 2002 2002 2002 Agency for Scientific Publ., Berlin 299 304
Ghisla S Pollegioni L D-Amino Acid Oxidase: Still New Lessons from a Seventy Year Old Enzyme Flavins and Flavoproteins, Proc 15th Int Symp 2005
Walsh CT Krodel E Massey V Abeles RH Studies on the elimination reaction of D-amino acid oxidase with α-amino-β-chlorobutyrate. Further evidence for abstraction of substrate α-hydrogen as a proton J Biol Chem 1973 248 1946 1955 4144104
Maclean J Ali S The structure of chorismate synthase reveals a novel flavin binding site fundamental to a unique chemical reaction Structure (Camb) 2003 11 1499 1511 14656434 10.1016/j.str.2003.11.005
Lauhon CT Bartlett PA Substrate analogs as mechanistic probes for the bifunctional chorismate synthase from Neurospora crassa Biochemistry 1994 33 14100 14108 7947820 10.1021/bi00251a019
Edmondson DE Mattevi A Binda C Li M Hubalek F Structure and mechanism of monoamine oxidase Curr Med Chem 2004 11 1983 1993 15279562
Franken HD Rüterjans H Müller F Nuclear-magnetic-resonance investigation of 15N- labeled flavins, free and bound to Megasphaera elsdenii apoflavodoxin Eur J Biochem 1984 138 481 489 6692831 10.1111/j.1432-1033.1984.tb07942.x
Moonen CT Vervoort J Müller F Reinvestigation of the structure of oxidized and reduced flavin: carbon-13 and nitrogen-15 nuclear magnetic resonance study Biochemistry 1984 23 4859 4867 6498164 10.1021/bi00316a007
Rüterjans H Kaun E Hull WE Limbach HH Evidence for tautomerism in nucleic acid base pairs. 1 H NMR study of 15N labeled tRNA Nucleic Acids Res 1982 10 7027 7039 7177856
Kaplan JI Fraenkel G NMR of chemically exchanging systems 1980 New York: Academic press
Martin G Martin ML Gouesnard JP Diel P, Fluck E, Korfeld R 15N-NMR Spectroscopy 1981 18 Berlin: Springer Verlag 187 326
Binsch G Lambert JB Roberts BW Roberts JD Nitrogen-15 magnetic resonance spectroscopy II. Coupling constants J Am Chem Soc 1964 86 5564 5570 10.1021/ja01078a033
Gunter H NMR-Spectroscopy 1983 Stuttgart: Thieme-Verlag
Blomberg F Maurer W Rüterjans H 15N nuclear magnetic resonance investigations on amino acids Proc Natl Acad Sci USA 1976 73 1409 1413 5722
Dixon M Webb EC Enzymes 1979 New York: Longman/Academic Press
Ghisla S Macheroux P Sanner C Rüterjans H Müller F Curti B, Ronchi S, Zanetti G Ionization properties of reduced 1,5-dihydroflavin, rates of N(5)-H exchange with solvent Flavins and Flavoproteins 1991 Berlin, New York: Walter de Gruyter 27 32
Adamson AW A Textbook of Physical Chemistry 1973 New York: Academic Press
Ghisla S Hartmann U Hemmerich P Müller F Synthese, Struktur und Reaktivität von Dihydroflavinen Liebig's Ann Chem 1973 1973 1388 1415
Yalloway GN Mayhew SG Boren SJ Vervoort J Ghisla S, Macheroux P, Sund H Effects of pH on the 13C and 15N NMR spectra of the hydroquinone of Desulfovibrio vulgaris flavodoxin and its G61A mutant Flavins and Flavoproteins 1999 Berlin: Agency for Scientific Publications 187 190
Vervoort J Müller F LeGall J Bacher A Sedlmaier H Carbon-13 and nitrogen-15 nuclear-magnetic-resonance investigation on Desulfovibrio vulgaris flavodoxin Eur J Biochem 1985 151 49 57 4029133 10.1111/j.1432-1033.1985.tb09067.x
Eisenreich W Kemter K Bacher A Mulrooney SB Williams CH JrMüller F 13C-, 15N- and 31P-NMR studies of oxidized and reduced low molecular mass thioredoxin reductase and some mutant proteins Eur J Biochem 2004 271 1437 1452 15066170 10.1111/j.1432-1033.2004.04043.x
Ludwig ML Schopfer LM Metzger AL Partridge KA Massey V Structure and oxidation-reduction behavior of 1-deaza-FMN flavodoxins: modulation of redox potentials in flavodoxins Biochemistry 1990 29 10364 10375 2261478 10.1021/bi00497a011
van Schagen CG Müller F A 13C nuclear-magnetic-resonance study on free flavins and Megasphaera elsdenii and Azotobacter vinelandii flavodoxin. 13C-enriched flavins as probes for the study of flavoprotein active sites Eur J Biochem 1981 120 33 39 7308219 10.1111/j.1432-1033.1981.tb05666.x
Müller F Vervoort J Lee J Horowitz M Carreira LA Coherent anti-Stokes Raman spectra of isoalloxazines J Ramn Spectrosc 1983 14 106 117 10.1002/jrs.1250140211
Witanowski M Stefaniak L Webb GA Annu Rep NMR Spectros 1981 11B 1 148
Gutowsky HS McCall DW Slichter CP Nuclear magnetic resonance multiplets in liquids J Chem Phys 1953 21 279 292 10.1063/1.1698874
Grunwald E Loewenstein A Meiboom S Rates and mechanisms of protolysis of methylammonium ion in aqueous solution studied by proton magnetic resonance J Chem Phys 1957 27 630 640 10.1063/1.1743802
|
16309555
|
PMC1318458
|
CC BY
|
2021-01-04 16:26:24
|
no
|
BMC Biochem. 2005 Nov 25; 6:26
|
utf-8
|
BMC Biochem
| 2,005 |
10.1186/1471-2091-6-26
|
oa_comm
|
==== Front
BMC Med Inform Decis MakBMC Medical Informatics and Decision Making1472-6947BioMed Central London 1472-6947-5-371632114510.1186/1472-6947-5-37SoftwareSLIM: an alternative Web interface for MEDLINE/PubMed searches – a preliminary study Muin Michael [email protected] Paul [email protected] Fang [email protected] Michael [email protected] Office of High Performance Computing and Communications, National Library of Medicine, 8600 Rockville Pike, Bethesda, Maryland 20894, USA2005 1 12 2005 5 37 37 5 8 2005 1 12 2005 Copyright © 2005 Muin et al; licensee BioMed Central Ltd.2005Muin et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
With the rapid growth of medical information and the pervasiveness of the Internet, online search and retrieval systems have become indispensable tools in medicine. The progress of Web technologies can provide expert searching capabilities to non-expert information seekers. The objective of the project is to create an alternative search interface for MEDLINE/PubMed searches using JavaScript slider bars. SLIM, or Slider Interface for MEDLINE/PubMed searches, was developed with PHP and JavaScript. Interactive slider bars in the search form controlled search parameters such as limits, filters and MeSH terminologies. Connections to PubMed were done using the Entrez Programming Utilities (E-Utilities). Custom scripts were created to mimic the automatic term mapping process of Entrez. Page generation times for both local and remote connections were recorded.
Results
Alpha testing by developers showed SLIM to be functionally stable. Page generation times to simulate loading times were recorded the first week of alpha and beta testing. Average page generation times for the index page, previews and searches were 2.94 milliseconds, 0.63 seconds and 3.84 seconds, respectively. Eighteen physicians from the US, Australia and the Philippines participated in the beta testing and provided feedback through an online survey. Most users found the search interface user-friendly and easy to use. Information on MeSH terms and the ability to instantly hide and display abstracts were identified as distinctive features.
Conclusion
SLIM can be an interactive time-saving tool for online medical literature research that improves user control and capability to instantly refine and refocus search strategies. With continued development and by integrating search limits, methodology filters, MeSH terms and levels of evidence, SLIM may be useful in the practice of evidence-based medicine.
==== Body
Background
There is unprecedented growth of medical information. PubMed, a service of the National Library of Medicine, includes over 15 million citations of biomedical articles [1]. MEDLINE, the largest component of PubMed, covers over 4,800 journals. Searching for relevant and updated information in MEDLINE/PubMed can be challenging. Keyword searches without search limits and filters may retrieve thousands of citations.
Several studies have demonstrated that the number of physicians using the Internet is increasing [2-4]. De Groote and Dorsch confirmed that a large percentage of users in academic health sciences prefer online resources over print, and many choose to access these online resources remotely [5]. Convenience and availability of full text journals were important factors in selecting online resources to use. These trends suggest that widely available Web technologies that enhance browsing capabilities may improve MEDLINE/PubMed search experiences.
Many MEDLINE/PubMed searches start with single-word terms or free-text phrases. Wildemuth and Moore [6] noted that this search technique is prone to syntactical or typographical errors but can be improved with the use of an online thesaurus and the inclusion of synonyms in the search concepts. Entrez, the integrated, text-based search and retrieval system used for PubMed [7], addressed this issue by integrating Medical Subject Headings (MeSH) in the search process. MeSH is a controlled vocabulary thesaurus of the National Library of Medicine available online through the MeSH Browser [8].
The practice of evidence-based medicine encourages the selection of appropriate current evidence through effective literature searches. The process includes the refinement of the search strategy and evaluation of search results. This may require familiarity with PubMed search limits, clinical study filters and MeSH terms. Gallagher, Allen and Wyer [9] stated that "MeSH headings make the searcher less dependent on the words the author of an article chooses to use in the abstract or title and allow the searcher to rely more on the actual content of the article." The use of MeSH headings, MeSH subheadings and text words to refine and redirect MEDLINE searches can make the search process more effective.
The goal of the project was to develop a Web-based application with an alternative search interface using slider controllers to implement search limits, methodology filters and MeSH terminologies. The intentions were to enhance user interaction with the MEDLINE/PubMed database, allow immediate refinement and redirection of search strategies and provide options to easily use the MeSH thesaurus.
Implementation
Interface development and design
SLIM, or Slider Interface for MEDLINE/PubMed searches, is a Web-based application accessible through the Internet with a Web browser. The interface has three main components: the search form, the information box and the search results.
The search form (Figure 1) is the constant component in all views of the application and the only component initially loaded. It contains a text box for the search terms along with five slider controls. The JavaScript slider bars move between arbitrary values of 0 to 100 with mouse clicks or keyboard strokes. The default setting for all slider controls is zero, i.e. no limits or filters. Checkboxes are available for limits not appropriate for slider controls, such as Human studies and English language limits.
Figure 1 SLIM interface showing its three main components: search form, information box and search results. Green box: Information box. Red ellipse: Text link to display abstract
SLIM allows a preview of the number of results through the information box (Figure 1.) The information box is displayed below the search form for both previews and full searches. It also provides additional search information such as mapped MeSH Terms, mapped Subheadings and unmapped terms.
When a full search is done, the results are presented below the information box. Each citation contains the article title, authors, journal name, volume, issue, pages and publication types. Abstracts, if available, are loaded on the page embedded with JavaScript functions that allow the abstracts to be hidden and displayed instantly with the click of a mouse (Figure 2 and 3). Abstracts are hidden by default. Links to related articles and PubMed are also provided.
Figure 2 Citation view with abstract hidden. Green ellipse: Text link to show abstract
Figure 3 Citation view with abstract displayed. Red ellipse: Text link to hide abstract
Limits and filters
SLIM follows the search and syntax rules of PubMed. Limits and filters are implemented by appending phrases and search tags to the search terms before querying the MEDLINE/PubMed database through the Entrez Programming Utilities (E-Utilities) [10]. The slider bars represent step implementations of common limits and filters in PubMed. Each slider bar holds values from 0 to 100, which are divided into subsets according to the number of options in that slider bar. When the slider is moved, the value where it stops is matched against the range of numbers assigned to a specific limit or filter. This scheme allows flexibility in designing the slider bar algorithms. Any changes (adding in or taking out filters) result in minor adjustments in the range of numbers in each subset.
The first three slider bars control search limits. These limits are Publication Date, Journal Subset and Age Groups. The three limits are modified versions of the Limits tab of PubMed. For the publication date, the slider has 11 steps and allows the user to choose different date ranges from 10 years back to the current year. The second slider contains 11 journal subset filters. Users have the option to search within the PubMed database, the MEDLINE subset [11] or the Core Clinical Journals [12]. Within each subset, users can limit results to availability of abstracts, full text or free full text. Full text articles provide links to publishers and may require subscription. Table 1 lists the journal subset descriptions and the corresponding PubMed limits. The age group slider is a reordered version of the age group drop-down menu in PubMed. It has 13 options starting with "Newborn" and ending with "80 and over". Table 2 lists the age group descriptions and the corresponding PubMed limits.
Table 1 Journal subset slider bar limit descriptions and corresponding PubMed filters
Slider Value Limit Description PubMed Filter
0 Default (All PubMed) No filter
1–5 PubMed: with abstracts only AND hasabstract
6–15 PubMed: full text AND full text[sb]
16–25 PubMed: free full text AND free full text[sb]
26–35 MEDLINE (4800+ journals) AND medline[sb]
36–45 MEDLINE: with abstracts only AND medline[sb] AND hasabstract
46–55 MEDLINE: full text AND medline[sb] AND full text[sb]
56–65 MEDLINE: free full text AND medline[sb] AND free full text[sb]
66–75 Core Clinical (120 journals) AND jsubsetaim
76–85 Core Clinical: with abstracts only AND jsubsetaim AND hasabstract
86–95 Core Clinical: full text AND jsubsetaim AND full text[sb]
95–100 Core Clinical: free full text AND jsubsetaim AND free full text[sb]
Table 2 Age group slider bar limit description and corresponding PubMed filters
Slider Value Limit Description PubMed Filter
0 Default No age limits
1–8 Newborn (Birth – 1 month) "infant, newborn"[MeSH Terms]
9–16 Infant (1–23 months) "infant"[MeSH Terms:noexp]
17–24 Infant Group (Birth-23 months) "infant"[MeSH Terms]
25–32 Preschool Child (2–5 years) "child, preschool"[MeSH Terms]
33–40 Child (6–12 years) "child"[MeSH Terms:noexp]
41–48 Adolescent (13–18 years) "adolescent"[MeSH Terms]
49–56 Child Group (0–18 years) ("infant"[MeSH Terms] OR "child"[MeSH Terms] OR "adolescent"[MeSH Terms])
57–64 Adult (19–44 years) "adult"[MeSH Terms:noexp]
65–72 Middle Aged (45–64 years) "middle aged"[MeSH Terms]
73–80 Middle Aged + Aged (45+ years) ("middle aged"[MeSH Terms] OR "aged"[MeSH Terms])
81–88 Adult Group (19+ years) "adult"[MeSH Terms]
89–96 Aged (65+ years) "aged"[MeSH Terms]
97–100 80 and over (80+ years) "aged, 80 and over"[MeSH Terms]
The fourth slider bar is a study design filter based on publication types, study designs or citation subsets. Although each filter within the slider functions independently, the slider values are listed according to the hierarchy of levels of evidence [13,14]. Three study designs, case-control studies, cohort studies and randomized controlled trials, are further divided into broad or narrow searches, giving a total of 9 slider levels. Research methodology filters for therapy [15] were adopted for randomized control trials. Other filters from the Clinical Studies Categories [15], e.g. diagnosis and prognosis, were not incorporated and will be evaluated for future versions of the application. The systematic reviews subset of PubMed was used unmodified for systematic review searches [16]. Table 3 lists the study design levels and the corresponding PubMed filters used for the slider control.
Table 3 Study design slider bar limit descriptions and corresponding PubMed filters
Slider Value Limit Description PubMed Filter
0 Default No filters
1–12 Case Reports AND ("case reports"[Publication Type] OR (case report[TIAB] OR case reported[TIAB] OR case reporting[TIAB] OR case reports[TIAB]))
13–24 Cross-sectional Surveys AND ("cross-sectional studies"[TIAB] NOT Medline[SB]) OR "cross-sectional studies"[MeSH Terms] OR (cross [TIAB] AND sectional [TIAB])
25–36 Case-control studies, broad search AND ((case control stud*[TIAB] NOT Medline[SB]) OR "case-control studies"[MeSH Terms] OR case-control stud* [TIAB])
37–48 Case-control studies, narrow search AND (case control stud*[TIAB] OR "case-control studies"[MeSH Terms:noexp] OR "case-control studies"[MAJR])
49–60 Cohort Studies, broad search AND ("cohort studies"[MeSH Terms] OR (cohort[TIAB] AND stud* [TIAB]))
61–72 Cohort Studies, narrow search AND ("cohort studies"[MeSH Terms:noexp] OR "cohort studies"[MAJR])
73–84 Randomized Controlled Trials, broad search [15] AND ((clinical[Title/Abstract] AND trial[Title/Abstract]) OR clinical trials[MeSH Terms] OR clinical trial[Publication Type] OR random*[Title/Abstract] OR random allocation[MeSH Terms] OR therapeutic use[MeSH Subheading])
85–96 Randomized Controlled Trials, narrow search [15] AND (randomized controlled trial[Publication Type] OR (randomized[Title/Abstract] AND controlled[Title/Abstract] AND trial[Title/Abstract]))
97–100 Systematic Reviews [16] AND systematic[sb]
Search mapping is a slider bar designed for intermediate to advanced users of PubMed. The filters use search tags and MeSH term operations to modify the search query. Search tags and MeSH term operations are short words or phrases enclosed in square brackets and appended to the keywords to refine search strategies. Initial modification substitutes the "[Text Word]" search tag with the "[TIAB]" search tag that redirects the search to keywords in the Title or Abstract. Subsequent levels in the slider involve adding mapped MeSH terms as filters, with options to search MeSH terms as major topics, or exclude MeSH terms below the current term in the MeSH tree. Table 4 describes the search mapping algorithm.
Table 4 Search mapping slider bar limit description and corresponding PubMed filter algorithm
Slider Value Limit Description PubMed Filter Algorithm
0 Default No change in details
1–14 Mapped keywords searched in Title and Abstract Keyword[Text Word] → Keyword[TIAB]
15–29 ANY Mapped MeSH Terms required TIABDetails + AND (MH OR MH)
30–44 ANY Mapped MeSH Terms required but with no tree explosion TIABDetails + AND (MH:noexp OR MH:noexp)
45–59 ALL Mapped MeSH Terms required TIABDetails + AND (MH AND MH)
60–74 ANY Mapped MeSH Terms should be MAJOR topic TIABDetails + AND (MH[MAJR] OR MH[MAJR])
75–89 ALL Mapped MeSH Terms required but with no tree explosion TIABDetails + AND (MH:noexp AND MH:noexp)
90–100 ALL Mapped MeSH Terms should be MAJOR topic TIABDetails + AND (MH[MAJR] AND MH[MAJR])
MeSH subheadings are topical qualifiers that describe a particular aspect of a subject such as etiology or therapeutic use. Users have the option to require these qualifiers through a checkbox. The MeSH subheadings are grouped and appended as a filter to the search query. This feature is closely linked with the search mapping algorithm. The subheading filter is only added if the search mapping algorithm requires MeSH terms, i.e. second level and above.
System architecture and development
SLIM was written in PHP and developed on an Apache 2.0.52 server running PHP 4.3.10. The PHP scripts generate a HyperText Markup Language (HTML) and JavaScript search form. JavaScript provides most of the functionality of the search form and search results. Free and open source JavaScript codes were downloaded from the Internet for the slider controls [17,18]. Customized JavaScript codes were written to enhance the usability of the form and slider controls by providing pre-defined limit combinations, tool tips and slider labels.
The application connects to the MEDLINE/PubMed database using tools from Entrez Programming Utilities (E-Utilities) [10]. The ESearch tool searches and retrieves primary IDs and term translations. The EFetch tool retrieves records from a list of one or more primary IDs. The E-Utilities server generates remote XML documents for both processes. Custom PHP scripts were written to parse the XML files. SLIM sends two successive passes through the ESearch tool and one final pass through the EFetch tool when the search form is submitted.
Automatic term mapping is the process where terms entered in the PubMed query box without a search tag are matched against the Medical Subject Headings translation table, the journals translation table, the full author translation table and an author index [19]. To optimize modifications done by SLIM on the search terms, it was essential to emulate the mapping and translation algorithms of Entrez PubMed. The first ESearch pass was designed to mimic this process. By sending unmodified search terms to the E-Utilities server and retrieving the translation stack from the XML document, a custom PHP function was able to build the detailed search query from the parsed XML elements. The goal was to capture the process found in the Details tab of PubMed. Using the translation stack, mapped MeSH terms and subheadings were identified and recorded. Terms tagged as "All Fields" were identified as unmapped terms.
Depending on slider bar and search form input, the detailed search query built on the first ESearch pass is modified by appending user-defined limits and study design filters, or by converting search tags. The modified query is processed once again through the ESearch tool to get the final list of PubMed IDs (PMIDs) from the second XML document. The ID list is sent to the EFetch tool to retrieve and display the details of the first 200 citations.
Performance and usability testing
To simulate performance testing, benchmarking timer functions were embedded in the PHP scripts to measure page loading and search times. All values generated during the alpha-testing with developers and beta-testing with users were recorded in a MySQL database for data analysis.
An online survey form was created to gather preliminary user opinion on stability and usability of the application. All seven questions used Likert scales to record answers. A call for participation in the usability testing was made on a mailing list of an international group of practicing physicians. Users were asked to use the system as a replacement for their regular PubMed search engine for two weeks. Comments and discussion on the application were encouraged.
Results
Initial qualitative testing by application developers demonstrated that SLIM is functional and stable. Occasional XML file connection problems occurred, but rarely. These were attributed to technical issues in the E-Utilities server and were quickly resolved by resubmitting the search form. The application was tested for compatibility with multiple browsers and worked in default installations of Internet Explorer 6.0, Mozilla Firefox 1.06, Opera 8.02 and Safari 1.2.
During performance testing, the index page was loaded 216 times with a mean page generation time of 2.94 milliseconds and standard deviation of 3.20 milliseconds. Search previews were processed 65 times with a mean page generation time of 0.63 seconds and standard deviation of 0.78 seconds. Complete searches with display of results were done 142 times with a mean page generation time of 3.84 seconds and standard deviation of 6.28 seconds. Table 5 gives a summary of the page generation time tests including standard deviations and quartiles.
Table 5 Summary of Page Generation Times
Index Page Preview Page Search Page
Number of Loading Tests 216 65 142
Mean 2.94 ms 0.63 s 3.84 s
Standard deviation 3.20 ms 0.78 s 6.28 s
Minimum value 2.38 ms 0.11 s 0.00 s
25th percentile 2.48 ms 0.30 s 0.52 s
Median 2.53 ms 0.41 s 1.12 s
75th percentile 2.61 ms 0.62 s 3.61 s
Maximum value 47.85 ms 5.54 s 35.04 s
ms – milliseconds
s – seconds
Eighteen physicians from the US, Australia and the Philippines participated in the beta-testing phase of the application and provided performance and usability feedback through an online survey. The mode was used as measure of central tendency as recommended by experts because Likert scales fall within the ordinal level of measurement [20,21]. Table 6 lists the tabulated answers and mode score for each statement in the user survey. Seven users agreed that the slider interface is more user-friendly than traditional search interfaces. Nine users found the Web-based application stable, while nine users strongly agreed that speed was acceptable. Eight users thought that the interface was easy to use. Seven physicians strongly agreed that similar interactive features are desirable in other MEDLINE searches. Six strongly believed that seeing mapped MeSH terms and unmapped keywords in their searches is a useful feature. Nine strongly preferred the option to hide and display the abstracts.
Table 6 Tabulation of answers and mode scores from user survey (n = 18)
Statements Answers/Total Respondents Mode Score
1 SD 2 D 3 N 4 A 5 SA
1 The new interface with slider controls is more user-friendly than traditional search interfaces. 1/18 3/18 4/18 7/18 3/18 4
2 The Web-based application is stable. 0/18 0/18 5/18 9/18 4/18 4
3 The speed of search and results display is acceptable. 0/18 0/18 1/18 8/18 9/18 5
4 The new interface is easy-to-use with minimal confusion on how controls work. 0/18 2/18 5/18 8/18 3/18 4
5 I'd like to see similar interactive features in other MEDLINE search engines. 1/18 1/18 5/18 4/18 7/18 5
6 Seeing mapped MeSH terms and unmapped keywords in the information box is a useful feature. 0/18 3/18 4/18 5/18 6/18 5
7 I like the feature where I can hide and display abstracts. 0/18 1/18 3/18 5/18 9/18 5
1 – Strongly disagree (SD)
2 – Disagree (D)
3 – Neutral (N)
4 – Agree (A)
5 – Strongly agree (SA)
Discussion
Search interface design
Research in data mining, natural language processing and methodology filters move towards developing backend algorithms and protocols for online information search and retrieval systems. Despite all these advances, users continue to use default settings in their PubMed searches. This is a preliminary study of an ongoing project to improve search interface usability by enhancing user interaction with advanced features of PubMed.
Limits and filters were chosen based on adaptability to structured series of implementations. Publication date and age groups were obvious choices. The journal subset slider narrowed down the search pool by combining subsets in PubMed with selected filters for availability of text. The hierarchy of levels of evidence provided a structured implementation for publication types and clinical study categories. The search mapping feature increased user involvement with MeSH operations. The project continues to study most of the limits and filters available in PubMed for integration in future versions of the application.
The information box was created as an educational tool for PubMed users and researchers by providing another level of feedback. It is nothing more than a short report on the term mapping and translation algorithms of Entrez PubMed. Users are informed of the MeSH term or MeSH subheading equivalents of their search terms. Unmapped terms can prompt users to modify specific keywords in their search.
Usability survey
Users who participated in the online survey were practicing physicians interested or involved in Medical Informatics. All were frequent users of PubMed and categorized themselves as intermediate to advanced users. The familiarity with the current Entrez PubMed interface often accounts for the negative responses in the online survey. Although most user comments were positive, one user stated that advanced users of PubMed might feel more comfortable with the consistent interface of Entrez PubMed, whereas slider controls might be more useful for novice or non-expert users. One user suggested that slider labels be more descriptive. In response, some label texts were expanded to include descriptions of the slider settings. A tutorial page was also suggested.
Limitations of the study
The study has several limitations. First, the search interface omits several limits existing in the PubMed limits page like subsets and non-English languages. The developers continue to look into integrating currently available filters in newer versions of the application. Second, only eighteen physicians interested in Medical Informatics projects participated in the beta-testing phase. A broad mix of PubMed researchers is a target for future usability studies. Third, the retrieval performance to measure sensitivity and specificity of the methodology filters was not tested. Review of Clinical Study Category filters not incorporated in the application, e.g. diagnosis and prognosis, is in process. Ongoing efforts to improve the application include evaluating search algorithms for accuracy and precision and adopting validated PubMed filters from previous published studies. The developers continue to monitor advancements in Web technology and add new interactive features in the application.
The goal of the project was to create a PubMed search application that allows users of all levels to easily go beyond basic keyword searches and move towards evidence-based principles. This is consistent with the practice of evidence-based medicine which advocates the formulation of effective search strategies to find current evidence. With increasing access to the Internet, online bibliographic databases have become important real-time resources for current evidence at the point-of-care. Although tools to control search parameters are available, these often require advanced familiarity with the search interface elusive to beginners. By exploring progressive Web technologies and creating an interactive search interface, the application may prove valuable in bridging the gap between expert and non-expert users of PubMed.
Conclusion
The Web-based application offers an alternative search interface to facilitate MEDLINE/PubMed searches. JavaScript slider bars control search limits, add filters or modify search terms with ease. Textual link controls can hide or display abstracts with a mouse. Initial qualitative testing and user feedback were positive which reinforced the approach of enhancing user interaction to improve online research. SLIM can be an interactive time-saving tool for online medical literature research that improves user control and capability to instantly refocus search strategies. With continued development and by integrating search limits, methodology filters, MeSH terms and levels of evidence, SLIM may be useful in the practice of evidence-based medicine.
Availability and requirements
Project name: SLIM (Slider Interface for MEDLINE/PubMed searches)
Project home page:
Operating systems: Platform independent
Programming language: PHP, JavaScript
Other requirements: JavaScript-enabled browsers, e.g. Fire Fox 1.0 or higher, IE 5 or higher
License: Free, anyone may use the service
Any restrictions to use by non-academics: None
List of abbreviations
SLIM: Slider Interface for MEDLINE/PubMed searches
MeSH: Medical Subject Headings
PHP: PHP: Hypertext Preprocessor
HTML: HyperText Markup Language
XML: Extensible Markup Language
PMID: PubMed ID
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MM conceived of the study, designed the interface, developed filters and algorithms, wrote the PHP codes, organized the usability testing and drafted the manuscript. PF assisted in the study and interface design, participated in usability testing, developed filters and algorithms and reviewed the initial drafts of the manuscript. FL wrote the custom JavaScript codes and assisted in design of the interface. MA participated in the study design, evaluation study and gave final approval of the version to be published. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This research was supported by the Intramural Research Program of the National Institutes of Health, the National Library of Medicine and Lister Hill National Center for Biomedical Communications.
The JavaScript slider controls were created by Erik Arvidsson and downloaded from WebFX 17 under the GNU GPL license. JavaScript codes for Cool DHTML Tooltips and Switch Content were downloaded free of charge from Dynamic Drive 18.
==== Refs
PubMed Overview
Carney PA Poor DA Schifferdecker KE Gephart DS Brooks WB Nierenberg DW Computer use among community-based primary care physician preceptors Acad Med 2004 79 580 90 15165980 10.1097/00001888-200406000-00017
Parekh SG Nazarian DG Lim CK Adoption of information technology by resident physicians Clin Orthop Relat Res 2004 107 11 15123934 10.1097/01.blo.0000126865.22310.59
Miller RH Hillman JM Given RS Physician use of IT: results from the Deloitte Research Survey J Healthc Inf Manag 2004 18 72 80 14971083
De Groote SL Dorsch JL Measuring use patterns of online journals and databases J Med Libr Assoc 2003 91 231 40 12883574
Wildemuth BM Moore ME End-user search behaviors and their relationship to search effectiveness Bull Med Libr Assoc 1995 83 294 304 7581185
NCBI Databases
MeSH Fact Sheet
Gallagher PE Allen TY Wyer PC How to find evidence when you need it, part 2: a clinician's guide to MEDLINE: the basics Ann Emerg Med 2002 39 436 40 11919532 10.1067/mem.2002.122774
Entrez Programming Utilities
List of Journals Indexed for MEDLINE
Abridged Index Medicus (AIM) Journal Titles
Brighton B Bhandari M Tornetta P 3rdFelson DT Hierarchy of evidence: from case reports to randomized controlled trials Clin Orthop Relat Res 2003 19 24 12897592 10.1097/01.blo.0000079323.41006.12
Centre for Evidence-Based Medicine, Levels of Evidence
PubMed Clinical Queries Filter Table
PubMed Systematic Reviews Subset
WebFX
Dynamic Drive
PubMed Help: Automatic Term Mapping
Jamieson S Likert scales: how to (ab)use them Medical Education 2004 38 1217 1218 15566531 10.1111/j.1365-2929.2004.02012.x
Milui HCI/Usability Articles: Likert scales and their uses
|
16321145
|
PMC1318459
|
CC BY
|
2021-01-04 23:52:11
|
no
|
BMC Med Inform Decis Mak. 2005 Dec 1; 5:37
|
utf-8
|
BMC Med Inform Decis Mak
| 2,005 |
10.1186/1472-6947-5-37
|
oa_comm
|
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-1051629319010.1186/1471-2334-5-105Research ArticleAcquired homotypic and heterotypic immunity against oculogenital Chlamydia trachomatis serovars following female genital tract infection in mice Lyons Joseph M [email protected]é Servaas A [email protected] Lucy P [email protected]ña A Salvador [email protected] James I [email protected] Department of Infectious Diseases, City of Hope National Medical Center and Beckman Research Institute, Duarte, California 91010, USA2 Laboratory of Immunogenetics, Section Immunogenetics of Infectious Diseases, VU University Medical Center, Amsterdam, The Netherlands3 Analytical Cytometry Laboratory, City of Hope National Medical Center and Beckman Research Institute, Duarte, California 91010, USA2005 17 11 2005 5 105 105 25 2 2005 17 11 2005 Copyright © 2005 Lyons et al; licensee BioMed Central Ltd.2005Lyons 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
Chlamydia trachomatis is the most common sexually transmitted bacterial pathogen causing female genital tract infection throughout the world. Reinfection with the same serovar, as well as multiple infections with different serovars, occurs in humans. Using a murine model of female C. trachomatis genital tract infection, we determined if homotypic and/or heterotypic protection against reinfection was induced following infection with human oculogenital strains of C. trachomatis belonging to two serovars (D and H) that have been shown to vary significantly in the course of infection in the murine model.
Methods
Groups of outbred CF-1 mice were reinfected intravaginally with a strain of either serovar D or H, two months after initial infection with these strains. Cellular immune and serologic status, both quantitative and qualitative, was assessed following initial infection, and the course of infection was monitored by culturing vaginal samples collected every 2–7 days following reinfection.
Results
Serovar D was both more virulent (longer duration of infection) and immunogenic (higher level of circulating and vaginal IgG and higher incidence of IgA in vaginal secretions) in the mouse genital tract. Although both serovars induced cross-reacting antibodies during the course of primary infection, prior infection with serovar H resulted in only a slight reduction in the median duration of infection against homotypic reinfection (p ~ 0.10), while prior infection with serovar D resulted in significant reduction in the median duration of infection against both homotypic (p < 0.01) and heterotypic reinfection (p < 0.01) when compared to primary infection in age and conditions matched controls.
Conclusion
Serovar D infection resulted in significant homotypic and heterotypic protection against reinfection, while primary infection with serovar H resulted in only slight homotypic protection. In addition to being the first demonstration of acquired heterotypic immunity between human oculogenital serovars, the differences in the level and extent of this immunity could in part explain the stable difference in serovar prevalence among human isolates.
==== Body
Background
Chlamydia trachomatis is the most common bacterial pathogen associated with sexually transmitted genital tract infections both in the United States [1] and worldwide [2]. It is generally accepted that most female genital tract infections with C. trachomatis are both asymptomatic and without severe sequelae [3]; and, that despite improved screening programs and the availability of highly effective antibiotics [4-6], there has been a significant increase in the incidence of C. trachomatis genital tract infection within the last half decade [2,7]. Although epidemiologic studies suggest that prior infection with C. trachomatis confers some short term protection against reinfection [8,9], the exact nature of this acquired immunity remains undetermined as do issues relating to the serovar specificity and possible involvement of this immunity in the more severe sequelae associated with multiple and/or chronic infection [10-12].
It has been advanced that serovar specific immune responses, particularly those made to the major outer membrane protein (MOMP), contribute to protection, whereas responses to broadly shared antigens, particularly those induced by chlamydial heat shock protein 60 (Hsp60), are associated with the immunopathologic injury that contributes to ectopic pregnancy and tubal infertility [13-19]. There is also increasing evidence from both human epidemiologic studies and animal model experiments to support the hypothesis that a protective immune response to reinfection is a complex, site of infection variable interaction between C. trachomatis and specific Th1 dominant cellular responses and interferon-gamma mediated Th1 augmenting humoral responses [20,21].
In attempts to understand the salient features and specific components of this interplay, a great deal of work has been performed using slightly modified versions of a murine model of female genital tract infection first described by Tuffrey and Taylor-Robinson [22]. In our laboratory, we routinely use strains belonging to the human oculogenital biovar of C. trachomatis and have previously reported significant differences in the course of infection among strains belonging to 7 oculogenital serovars, which loosely correlated with the prevalence of the serovars among human clinical isolates, especially for the most and least prevalent serovars, i.e. D and E versus H and I [23].
The purpose of the present study was to expand on these observations by assessing the degree of homotypic and heterotypic protection against reinfection that follows resolution of infection with human isolates of C. trachomatis. Two strains were selected from the previously studied collection of strains: the strain of serovar D which was shown to establish the longest duration of infection (median duration of 38 days) and greatest humoral response; and the serovar H strain which had the shortest duration (median duration of 7 days) and lowest humoral response. In addition, lower genital tract infection with serovar D was shown to both ascend into the uterine horns with a greater frequency than serovar H, as well as shed more infectious units during the acute phase of infection [23]. In a separate study we demonstrated a link between certain in vitro growth characteristics and differences in the level of cytotoxic activity associated with elementary bodies between these strains [24]. Thus, strain selection was made with the intention of reflecting the greatest diversity observed among the strains studied in the murine model, as well as to represent serovars that have significantly different prevalence rates among human isolates in the hope that the results of the study would provide some insight into the possible causes of these differences. Duration of genital tract infection was used to determine the extent of protection, and humoral and cellular immune response data were evaluated to identify factors that associated with any observed protection.
Methods
Animals
Outbred CF-1 female mice (Charles River Labs) were purchased at 7 weeks of age and were allowed to acclimate for one week prior to use. All experiments were conducted in a BL-2 containment facility in compliance with animal care regulations and under protocols approved by the institutional research animal care committee.
Bacteria and culture technique
Mycoplasma-free and pure PCR-typed strains of serovar D and H [23,24] were propagated, purified, titrated, and isolated in cycloheximide treated McCoy cell monolayers using standard techniques. Separate vials of the same -70°C stored stocks were used for both infections.
Murine model
Progesterone, in the form of medroxyprogesterone acetate (Depo-Provera, Upjohn), was given subcutaneously (sc) in 2.5 mg doses, 10 and 3 days prior to infection [22,23]. Prior to infection, vaginal specimens were obtained for culture and cytology, and 2 hours later mice were inoculated intravaginally by direct instillation of 25 μl of sucrose phosphate buffered transport medium (SP) containing 1–3 × 107 inclusion forming units (ifu).
Sample collection
The presence of Chlamydia in the lower genital tract was determined by swabbing the vaginal vault and ectocervix with a calcium alginate swab, which was cultured for the organism. Plasma and vaginal secretions were obtained prior to infection. Blood was taken from a small tail vein incision, diluted 1:10 in PBS, and the plasma (P) was separated by centrifugation. Vaginal secretions (VS) were collected over a two-hour period by absorption into a piece of surgical sponge, and eluted into 400 μl of PBS. All samples were frozen at -20°C until tested.
Serological analyses
Anti-chlamydial IgG and IgA titers in blood and vaginal secretions were determined by indirect solid phase enzyme immunoassay (EIA) using SDS solubilized Ct elementary bodies as antigen. Western immunoblot analysis were performed on a similar antigen preparation which had been SDS-PAGE separated and transferred to nitrocellulose paper. After incubation with sample (P at a final dilution1:200; VS at a final dilution of 1:20), reactive bands were visualized with standard EIA reagents.
Spleen cell analysis
Cellular immune responses were determined using a standard assay for 3H-thymidine (3H-Td) incorporation. Specific responses were measured using formalin preserved elementary bodies (EB) at a ratio of 100 EB:1 spleen cell, and non-specific responses measured using concavalin A (Con A) at 2 μg/ml, phytohemagglutinin (PHA) at 5 μg/ml, and lipopolysaccharide (LPS) at 100 μg/ml. Single-cell suspensions were prepared by gently pressing the spleen through a nylon sieve. Debris was allowed to settle for 2 min, and the supernatant containing single cells was spun down at 500 × g for 10 minutes. Erythrocytes were lysed with NH4Cl solution and cells were washed three times with RPMI 1640 medium containing 10% fetal bovine serum and plated in triplicate, in 96-well plates at a concentration of 5 × 105 cells per well. Proliferative responses were measured by uptake of 1 μCi of 3H-thymidine (3HTd) per well for the last 24 hours of a 72 hour incubation period. As a measure of response, a stimulation index (SI) was calculated (SI = 3H-Td incorporation with stimulation/3HTd incorporation without stimulation).
Statistical evaluation
Duration of infection data were analyzed by the Wilcoxon Rank Sum Test; meaned data by F-Test; and frequency data by Chi Square.
Results
Homotypic and heterotypic protection against reinfection
Consistent with our previous report [23], intravaginal inoculation with the serovar D strain resulted in a longer duration of infection in previously uninfected mice compared with the serovar H strain (22.5 days versus 15.5 days, p = 0.05) (Table 1). Analysis of the effect of prior C. trachomatis genital tract infection on the duration of infection following homotypic and heterotypic reinfection showed that infection with serovar H resulted in only slight homotypic protection (15.5 days versus 10 days duration of infection: p = 0.1), while serovar D infection resulted in significant homotypic (22.5 days versus 12 days duration of infection: p < 0.01) and heterotypic protection (15.5 days versus 5 days duration of infection: p < 0.01) against reinfection (Table 1).
Table 1 Acquired Homotypic and Heterotypic Immunity Against Oculogenital Chlamydia trachomatis Serovars Following Female Genital Tract Infection in Mice
Infection Serovar Number of Animals Number of Animals Culture Positive on Reinfection Day Median Duration of Infection Wilcoxon Rank Sum p Value
Initial1 Reinfection 2 4 6 8 10 14 17 21 24 28 31 35 38 42 45 48 52 55
D H 12 12 5 3 2 5 2 1 0 0 0 0 0 0 0 0 0 0 0 5 <0.01
H H 12 12 11 8 8 10 3 1 2 0 0 0 0 0 0 0 0 0 0 10 ~0.10
None H 12 12 12 12 12 11 6 5 2 2 1 0 0 1 0 0 0 0 0 15.5 --
D D 12 12 9 7 9 7 5 4 2 2 1 0 0 0 0 0 0 0 0 12 <0.01
H D 12 12 12 12 12 11 6 10 7 7 4 1 2 1 0 0 0 0 0 24 NS
None D 12 12 12 12 12 11 7 5 6 4 4 2 1 1 2 0 0 0 0 22.5 --
1Median duration of infections during the primary infection phase of this study were 10 days with serovar H and 33 days with serovar D.
Humoral immune responses
Quantitative
As in our previous study [23], genital tact infection with serovar D resulted in a significantly greater quantitative anti-chlamydial humoral response compared to the response to serovar H infection (Table 2). Serovar D induced significantly higher plasma (p < 0.05) and vaginal (p < 0.05) IgG levels as compared to serovar H. Although no quantitative differences between the IgA responses following infection with either serovar D and H were observed in the vaginal secretions of animals with detectable levels of antibody, a significant difference was observed in the frequency of IgA positive vaginal secretions from serovar H and D infected animals prior to reinfection (11/24 for H vs 21/24 for D, p < 0.01, data not shown). In all cases, similar homologous and heterologous antibody titers were detected against both antigen preparations used in the assay.
Table 2 Serologic Analysis of Plasma and Vaginal Secretions Following Infection and Immediately Prior to Reinfection with Chlamydia trachomatis Serovars D and H
Plasma and Vaginal Secretion Titers (Log 2) (p Value)
Infection Serovar Number of Animals Plasma IgG VS IgG VS IgA1
Primary Secondary H D H D H D
D H 12 13.0 13.6 8.3 8.7 7.2 8.3
(<0.05) (<0.05) (NS)
H H 12 11.0 11.2 5.3 6.1 6.7 6.8
None H 12 <6 <6 <2 <2 <2 <2
D D 12 12.5 13.1 7.8 8.3 7.3 7.6
(<0.05) (<0.05) (NS)
H D 12 11.4 11.4 5.3 5.5 6.0 6.4
None D 12 <6 <6 <2 <2 <2 <2
1Values listed are the mean titers of IgA positive animals only.
Qualitative
Tables 3 and 4 contain the Western immunoblot analysis of plasma and vaginal secretions from representative animals following resolution of initial infection and immediately prior to reinfection with serovar H (Table 3) or with serovar D (Table 4). Consistent with the quantitative findings, plasma and vaginal secretions from serovar D infected mice contained antibodies to a greater array of antigens than specimens from serovar H infected mice; and although often more intense when homologous, plasma from mice infected with either serovar gave similar immunoblot patterns against both antigen preparations, thus demonstrating the induction of a high level of cross-reacting IgG during infection.
Table 3 Serological Analysis of Plasma and Vaginal Secretions from Representative Animals Following Primary Infection with Either Chlamydia trachomatis Serovar H or D and Immediately Prior to Infection with Serovar H
H Primary Infection D Primary Infection
Animal J1 H3 J3 L1 H4 J2 R1 X1 X2 R4 T2 Y4
Infection Duration (Days)
Primary 24 24 8 24 6 24 38 42 24 42 28 42
Secondary 2 4 10 10 21 21 2 2 4 6 14 17
Plasma and Vaginal Secretion Titers (Log 2) and IgG Immunoblot Reactions Against serovar H and D
H D H D H D H D H D H D H D H D H D H D H D H D
P IgG 14 13 13 13 7 8 13 13 7 8 13 13 12 13 13 14 13 13 14 14 13 13 12 13
V IgG 8 8 8 8 2 3 9 9 <2 <2 8 9 7 8 8 7 8 9 8 8 8 8 8 8
V IgA 9 8 8 8 <2 <2 8 8 <2 <2 4 4 5 7 8 9 7 8 7 8 8 9 8 9
MW (kD) P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V
190 - - - - - - -
140
120 · - · -
115
110 · ·
108
106 • • • •
102
100 - ·
96
92 · - • ·
90
78 · -
72
66 · - · - • ·
64
62 · - · · • · · • • - • · · · • · • · • · · - · · • · • ·
61.5
60 ● · • • • · • • · - · - • • ● • • • • • • • • • • ● • • ● • • • ● • • •
58 · -
56 · -
50 · - · - · -
46.5 • · ● • · - • · · - ● · · - ● • • • • • • · • • • • • · · - ● ●
40
38
36 · - · - · -
33
29 · - · - · - • • • • · · • •
26
22.5 • -
19.5
15.5 · · · ·
- = No reaction
Plasma = P
· = Barely visible
Vaginal secretions = V
• = Weak
● = Strong
Table 4 Serological Analysis of Plasma and Vaginal Secretions from Representative Animals Following Primary Infection with Either Chlamydia trachomatis Serovar H or D and Immediately Prior to Infection with Serovar D
H Primary Infection D Primary Infection
Animal I1 G3 K2 G2 K3 I4 Z1 S4 Q4 Z2 Q2 S1
Infection Duration (Days)
Primary 6 10 6 31 14 6 45 21 14 45 52 10
Secondary 10 17 24 28 35 38 2 4 10 14 24 28
Plasma and Vaginal Secretion Titers (Log 2) and IgG Immunoblot Reactions Against serovar H and D
H D H D H D H D H D H D H D H D H D H D H D H D
P IgG 11 11 13 13 11 11 12 12 9 9 13 11 14 14 12 13 11 11 13 14 14 14 11 11
V IgG 5 5 7 6 2 4 6 7 4 4 5 5 11 11 7 7 7 8 9 10 9 9 4 5
V IgA <2 <2 <2 <2 <2 <2 7 8 <2 <2 <2 2 3 5 <2 <2 4 6 8 9 10 9 <2 <2
MW (kD) P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V P V
190 - - - - - - -
140
120 · -
115
110
108
106 · - · -
102
100
96
92 - ·
90
78 · -
72
66 · -
64
62 ● · • - • · • • · · • · · · · · · - ● · ● · ● ● ● ●
61.5
60 ● • ● • · - · - • · • · ● • • • • · • • · - · · • · • - • • • • · · • ·
58
56
50
46.5 · - · - ● • • · • · • · • · · - ● · • · · - ● • · - • • • · • · · - ● • · - • ·
40
38 - ·
36 • · • • · - · -
33
29 · - · - - · • · • • · - · - · · • ·
26
22.5 · - · -
19.5 · - · · · -
15.5 · · · ·
- = No reaction
Plasma = P
· = Barely visible
Vaginal secretions = V
• = Weak
● = Strong
With the exception of the link between quantitative humoral response and protection, statistical analysis of the serologic and duration of infection data did not detect a humoral factor(s) that correlated with the shorter duration of infection observed following reinfection.
Splenic lymphocyte responses
Table 5 summarizes the chlamydia-specific and mitogen non-specific splenic lymphocyte responses obtained from groups of four animals 55 days following genital tract inoculation with either serovar H or D. Mean responses to PHA and both elementary body preparations were greater following infection with either serovar when compared to non-infected controls (p 0.05); while responses to both Con-A and LPS were not different from controls. No association could be found between a particular pattern or magnitude of cellular responses and an individual animals duration of infection.
Table 5 Splenic Lymphocyte Responses 55 Days After Primary Infection with Serovars H and D
Mean Stimulation Index ± 1 SEM
EB Serovar Standard Mitogens
Group H D Con-A PHA LPS
Control 13.0 ± 1.7 15.9 ± 1.8 79.5 ± 12.0 6.2 ± 1.8 58.4 ± 5.3
H 19.2 ± 3.1* 23.4 ± 4.1* 96.2 ± 15.6 11.3 ± 6.4* 58.7 ± 15.8
D 19 ± 1.5* 21.7 ± 1.7* 78.8 ± 6.8 11.9 ± 6.4* 51.1 ± 6.3
*Significantly different from Control mean at p 0.05
Discussion
Using a murine model of C. trachomatis female genital tract infection, we have demonstrated that homotypic immunity against reinfection was induced following initial infection with either serovar D or H, but that a more significant level of protection was observed following infection with serovar D. However, heterotypic protection against reinfection was strain dependant and was seen only when the initial infection was with the more virulent and immunogenic strain of serovar D. These results are the first demonstration of heterotypic immunity between two oculogenital serovars following female genital tract infection in the mouse model, as well as being the first comparative study that suggests a possible strain dependent restriction on the process. The findings are consistent with other studies that have observed heterotypic immunity in the context of MoPn and serovar E. Also consistent with prior reports is the correlation between the level of protection observed and the virulence of the strains used to infect the mouse genital tract, with the more virulent MoPn inducing a more solid level of heterotypic immunity against serovar E than was observed in the reverse situation [25,26]. Finally, the observed differences in heterotypic and homotypic responses between serovars D and H could explain in part the relatively stable differences in the frequency of these two serovars among human isolates from different geographic locations [27-33], i.e. a prior infection with a more virulent/immunogenic serovar would confer a greater degree of protection against infection with a less virulent/immunogenic serovar, thus reducing the incidence of the latter by reducing the efficiency of serial transmission.
Consistent with human epidemiologic data [8,9], the protection observed in this model against reinfection with oculogenital serovars of C. trachomatis was not complete, but rather acted in a way that reduced the duration of infection and level of bacterial shedding (data not shown) during infection. With the notable exception that serovar H infection resulted in a lesser quantitative and qualitative humoral response, a thorough analysis of the individual and collective data was unable to identify any specific element(s) that correlated with protection. This finding is consistent with a report in which a similarly extensive analysis of cellular and humoral responses was performed in a comparison of the acquired immunity induced by infection with MoPn and serovar E [25]. Although not proof, this does support the current working hypothesis that acquired immunity to C. trachomatis female genital tract infection is a complex and integrated phenomena that relies on both Th1 and Th2 type responses made during the course of infection, which in turn enhance innate immune responses upon reinfection [20]. This complexity, which likely arises out of the ever changing physiologic and immunologic milieu within and between anatomically distinct but connected regions of the female genital tract, may account for the difficulty identifying specific components of what may be a flexible pattern of responses that lead to a given individual's level of protection against or risk of severe upper genital tract pathology [21].
Women with recurrent C. trachomatis infection are at increased risk of reproductive sequelae, including pelvic inflammatory disease, ectopic pregnancy and tubal infertility [10-12], which have been linked to both cellular and humoral immune responses induced during infection [13-19]. How the nature and level of homotypic and/or extent of heterotypic immunity in the murine model extrapolate to the risk of upper genital tract pathology and infertility was not assessed in this study, and is an area of investigation that has yet to be systematically addressed. Most of the experimental data relating to the severe sequelae associated with C. trachomatis female genital tract infection has been obtained in studies using C. muridarum, MoPn [34,35]. As a result, it has not been possible to clearly assess the immunologic features that are thought to contribute to severe sequelae within the human female genital tract, because the damage that occurs within the murine genital tract following infection with MoPn is a consequence of acute and not chronic infectious processes and/or recurrent infection [36,37]. Typically in the mouse and in most women, human oculogenital serovars are limited in their ability to ascend with any major pathologic consequence from the initial site of infection within the lower genital tract. However, infection of female C3H/HeN with a strain of serovar E has been shown to ascend and cause infertility without gross pathology at a low incidence following a single infection and with increased incidence upon reinfection [26]. Of particular interest is that none of these mice developed hydrosalpinx, which is the hallmark of upper tract infection with MoPn, indicating a possible alternative mechanism for the induction of infertility, one different from the tubal dilation, scarring and associated hydosalpinx that occurs as a sequelae to MoPn infection. Although no specific immunologic mechanism was identified as contributing to the development of infertility, it was noted that the C3H/HeN mouse is an Hsp60 responding strain and that Hsp60 responding strains of mice are more susceptible to MoPn induced infertility while non-responders can be infected but do not experience severe upper tract pathology. It will be worthwhile to determine if the two C. trachomatis strains used in the present study show a similar relationship to each other in C3H/HeN mice and whether homotypic and/or heterotypic elements of immunity play a role in the progression of events that lead to severe upper tract pathology, which is essentially the reason why C. trachomatis genital tract infections are significant and why intervention and prevention strategies are needed.
Conclusion
In conclusion, we demonstrate in the first study comparing phenotypically different strains representing two human oculogenital serovars that both the level of homotypic protection against reinfection and the ability to confer heterotypic protection correlated with the virulence/immunogenicity of the strain. Extrapolating the results to human epidemiologic data could explain in part the relatively stable differences in the frequency between the most and least prevalent serovars based on a serovars ability to induce or not induce heterotypic immunity. Although no specific cellular or humoral factor(s) could be identified that associated with the observed protection, it is clear that heterotypic immunity can be induced and that the systematic study of human oculogenital serovars in the mouse model of female genital tract infection could provide information that leads to an understanding of what distinguishes a protective from an immunopathic response to infection.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Joseph M. Lyons participated in the design, planning, execution and supervision of the study, and drafted original and final manuscripts. Servaas A. Morré analyzed the data and constructed the tables that appear in the manuscript. Lucy P. Airo-Brown performed the animal experiments and developed the serologic assays used in the study. A. Salvador Peña provided direction during the development of the integrated approach to the study of the Chlamydia trachomatis infections of the female genital tract of which this report is a part. James I. Ito planned and approved the study design, and participated in the preparation of the manuscript. All authors read and approved the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported in part by Public Health Service grant AI-23792 from the National Institute of Allergy and Infectious Diseases.
==== Refs
Weinstock H Berman S Cates W Sexually transmitted diseases among American youth: incidence and prevalence estimates, 2000 Perspect Sex Reprod Health 2004 36 6 10 14982671 10.1363/3600604
Gerbase AC Rowley JT Heymann DH Berkley SF Piot P Global prevalence and incidence estimates of selected curable STDs Sex Trans Infect 1998 74 S12 16
van Valkengoed IGM Morré SA van den Brule AJC Meijer CJLM Bouter LM Boeke AJP Overestimation of Complication-rates of Chlamydia trachomatis Screening programs: implications for cost-effectiveness analyses Int J Epidemiol 2004 33 416 425 15082651 10.1093/ije/dyh029
Workowski KA Lampe MF Wong KG Watts MB Stamm WE Long-term eradication of Chlamydia trachomatis genital infection after antimicrobial therapy: evidence against persistent infection JAMA 1993 270 2071 2075 8305018 10.1001/jama.270.17.2071
Scholes D Stergachis A Heidrich FE Andrilla H Holmes KK Stamm WE Prevention of pelvic inflammatory disease by screening for cervical chlamydial infection N Engl J Med 1996 334 1362 1366 8614421 10.1056/NEJM199605233342103
van Valkengoed IG Morre SA van den Brule AJ Meijer CJ Bouter LM van Eijk JT Boeke AJ Follow-up, treatment, and reinfection rates among asymptomatic Chlamydia trachomatis cases in general practice Br J Gen Pract 2002 52 623 627 12171219
Miller WC Ford CA Morris M Handcock MS Schmitz JL Hobbs MM Cohen MS Harris KM Udry JR Prevalence of chlamydial and gonococcal infections among young adults in the United States JAMA 2004 291 2229 2236 15138245 10.1001/jama.291.18.2229
Barnes RC Roddy RE Stamm WE Oriel D, Ridgeway G, Schachter J, Taylor-Robinson D, Ward M Serovars of Chlamydia trachomatis causing repeated genital infections Chlamydial Infections Proceedings of the Sixth International Symposium on Human Chlamydial Infections 1986 Cambridge, UK, Cambridge University Press 503 507
Rietmeijer CA van Bemmelen R Judson FN Douglas JM Incidence and repeat infection rates of Chlamydia trachomatis among male and female patients in an STD clinic: implications for screening and rescreening Sex Trans Dis 2002 29 65 72 10.1097/00007435-200202000-00001
Brunham RC Maclean IW Binns B Pelling RW Chlamydia trachomatis: its role in tubal infertility J Infect Dis 1985 152 1275 1282 3840827
Westrom LR Joesoef R Reynolds G Hagdu A Thompson SE Pelvic inflammatory disease and fertility. A cohort study of 1,844 women with laparoscopically verified disease and 657 control women with normal laparoscopic results Sex Trans Dis 1992 19 185 192
Hillis SD Owens LM Marchbanks PA Amsterdam LF Mac Kenzie WR Recurrent chlamydial infections increase the risks of hospitalization for ectopic pregnancy and pelvic inflammatory disease Am J Obstet Gynecol 1997 176 103 107 9024098
Ward ME The immunobiology and immunopathology of chlamydial infections APMIS 1995 103 769 796 8546843
Dean D Oudens E Bolan G Padian N Schachter J Major outer membrane protein variants of Chlamydia trachomatis are associated with severe upper genital tract infections and histopathology in San Francisco J Infect Dis 1995 172 1013 1022 7561174
Lampe MF Wong KG Stamm WE Sequence conservation in the major outer membrane protein gene among Chlamydia trachomatis strains isolated from the upper and lower urogenital tract J Infect Dis 1995 172 589 592 7622913
Brunham RC Peeling RW Chlamydia trachomatis antigens: role in immunity and pathogenesis Infect Agents Dis 1997 3 218 233 7866655
Eckert LO Hawes SE Wolner-Hanssen P Money DM Peeling RW Brunham RC Stevens CE Eschenbach DA Stamm WE Prevalence and correlates of antibody to chlamydial heat shock protein in women attending sexually transmitted disease clinics and women with confirmed pelvic inflammatory disease J Infect Dis 1997 175 1453 1458 9180186
Domeika M Domeika K Paavonen J Mardh PA Witkin SS Humoral immune response to conserved epitopes of Chlamydia trachomatis and human 60-kDa heat-shock protein in women with pelvic inflammatory disease J Infect Dis 1998 177 714 719 9498452
Kinnunen A Molander P Morrison R Lehtinen M Karttunen R Tiitinen A Paavonen J Surcel HM Chlamydial heat shock protein 60 – specific T cells in inflamed salpingeal tissue Fertil Steril 2002 77 162 166 11779608 10.1016/S0015-0282(01)02922-3
Stephens RS The cellular paradigm of chlamydial pathogenesis Trends Microbiol 3003 11 44 51 12526854 10.1016/S0966-842X(02)00011-2
Kelly KA Cellular immunity and Chlamydia genital infection: induction, recruitment, and effector mechanisms Int Rev Immunol 2003 22 3 41 12710502 10.1080/08830180305229
Tuffrey M Taylor-Robinson D Progesterone as a key factor in the development of a mouse model for genital-tract infection with Chlamydia trachomatis FEMS Micribiol Lett 1988 12 111 115 10.1016/0378-1097(81)90004-5
Ito JI Lyons JM Airo-Brown LP Variation in virulence among oculogenital serovars of Chlamydia trachomatis in experimental genital tract infection Infect Immun 1990 58 2021 3 2341189
Lyons JM Ito JI Peña AS Morré SA Differences in the growth characteristics and elementary body associated cytotoxin activity between Chlamydia trachomatis oculogenital serovars D and H and Chlamydia muridarum (MoPn) J Clin Pathol 2005 58 397 401 15790704 10.1136/jcp.2004.021543
Ramsey KH Cotter TW Salyer RD Miranpuri GS Yanez MA Poulsen CE DeWolfe JL Byrne GI Prior genital tract infection with a murine or human biovar of Chlamydia trachomatis protects mice against heterotypic challenge infection Infect Immun 1999 67 3019 3025 10338514
Ramsey KH DeWolfe JL Salyer RD Disease outcome subsequent to primary and secondary urogenital infection with murine or human biovars of Chlamydia trachomatis Infect Immun 2000 68 7186 7189 11083853 10.1128/IAI.68.12.7186-7189.2000
Barnes RC Rompalo AM Stamm WE Comparison of Chlamydia trachomatis serovars causing rectal and cervical infections J Infect Dis 1987 156 953 958 3680995
Kuo CC Wang SP Holmes KK Grayston TJ Immunotypes of Chlamydia trachomatis isolates in Seattle, Washington Infect Immun 1983 41 865 868 6874078
Mabey DC Forsey WT Treharne JD Serotypes of Chlamydia trachomatis in The Gambia Lancet 1987 2 452 2887750 10.1016/S0140-6736(87)90984-6
Wagenvoort JHT Suchland RJ Stamm WE Serovar distribution of urogenital Chlamydia trachomatis strains in The Netherlands Genitourin Med 1988 64 159 161 3410465
Morré SA Rozendaal L van Valkengoed IG Boeke AJ van Voorst Vader PC Schirm J de Blok S van Den Hoek JA van Doornum GJ Meijer CJ van Den Brule AJ Urogenital Chlamydia trachomatis serovars in men and women with a symptomatic or asymptomatic infection: an association with clinical manifestations? J Clin Microbiol 2000 38 2292 2296 10834991
Suchland RJ Eckert LO Hawes SE Stamm WE Longitudinal assessment of infecting serovars of Chlamydia trachomatis in Seattle public health clinics: 1988–1996 Sex Transm Dis 2003 30 357 361 12671559
Spaargaren J Verhaest I Mooij S Smit C Fennema HS Coutinho RA Peña AS Morré SA Analysis of Chlamydia trachomatis serovar distribution changes in the Netherlands (1986–2002) Sex Transm Infect 2004 80 151 152 15054183 10.1136/sti.2003.006395
de la Maza LM Pal S Khamesipour A Peterson EM Intravaginal inoculation of mice with the Chlamydia trachomatis mouse pneumonitis biovar results in infertility Infect Immun 1994 62 2094 2097 8168974
Darville T Andrews CW Laffoon KK Shymasani W Kishen LR Rank RG Mouse strain-dependent variation in the course and outcome of chlamydial genital tract infection is associated with differences in host response Infec Immun 1997 65 3065 3073 9234755
Morrison RP Caldwell HD Immunity to murine chlamydial genital infection Infect Immun 2002 70 2741 51 12010958 10.1128/IAI.70.6.2741-2751.2002
Morré SA Lyons JM Ito JI Jr Murine models of Chlamydia trachomatis genital tract infection: use of mouse pneumonitis strain versus human strains Infect Immun 2000 68 7209 7211 11203323 10.1128/IAI.68.12.7209-7211.2000
|
16293190
|
PMC1318460
|
CC BY
|
2021-01-04 16:28:17
|
no
|
BMC Infect Dis. 2005 Nov 17; 5:105
|
utf-8
|
BMC Infect Dis
| 2,005 |
10.1186/1471-2334-5-105
|
oa_comm
|
==== Front
BMC Fam PractBMC Family Practice1471-2296BioMed Central London 1471-2296-6-481629724510.1186/1471-2296-6-48Research ArticleHow and why community hospital clinicians document a positive screen for intimate partner violence: a cross-sectional study Gerber Megan R [email protected] Karen S [email protected] Richard C [email protected] David H [email protected] Harvard Medical School, Cambridge Health Alliance, Cambridge, MA, USA2 Tufts-New England Medical Center, Boston, MA, USA2005 19 11 2005 6 48 48 26 4 2005 19 11 2005 Copyright © 2005 Gerber et al; licensee BioMed Central Ltd.2005Gerber 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
This two-part study examines primary care clinicians' chart documentation and attitudes when confronted by a positive waiting room screen for intimate partner violence (IPV).
Methods
Patients at community hospital-affiliated health centers completed a screening questionnaire in waiting rooms that primary care providers (PCPs) were subsequently given at the time of the visit. We first reviewed the medical records of patients who screened positive for IPV, evaluating the presence and quality of documentation. Next we administered a survey to PCPs that measured their knowledge, attitudes and practice regarding IPV.
Results
Seventy-two percent of charts contained some documentation of IPV, however only 10% contained both a referral and safety plan. PCPs were more likely to refer patients (p < .05) who screened positively for mood or anxiety disorders, disclosed that they feared for their safety or were economically disadvantaged. Those that feared for their safety or endorsed mood or anxiety disorders were more likely to have notation of a safety plan in their records. When surveyed, 81.6% of clinicians strongly agreed that it is their role to inquire about IPV, but only 68% expressed confidence in their ability to manage it. In contrast, 93% expressed confidence in managing depression. Sixty-seven percent identified time constraints as a barrier to care. Predictors of PCP confidence in treating patients who have experienced IPV (p < .05) included hours of recent training and clinical experience with IPV.
Conclusion
Mandatory waiting room screening for IPV does not result in high levels of referral or safety planning by PCPs. Despite the implementation of a screening process, clinicians lack confidence and time to address IPV in their patient populations suggesting that alternative methods of training and supporting PCPs need to be developed.
==== Body
Background
Intimate Partner Violence (IPV) is a major public health problem in the U.S. Nearly one-quarter of all women and 7.4% of men have been physically and/or sexually abused by an intimate partner in their adult lives [1]. Women are more likely than men to experience IPV and to be injured during an assault [1]. IPV is prevalent across geographic settings, social classes and ethnic groups [2].
Medical patients frequently have a history of IPV. In primary care practices, 5.5 to 14% of patients present with a history of recent IPV (abuse occurring in the preceding 12 months), while reported lifetime prevalence in these same settings ranges from 21 to 51% [3-5]. Women who have experienced IPV have higher rates of chronic medical conditions [6-8], and utilize more health care services [7-9].
Despite the high prevalence of IPV and its well-established adverse health impact, clinicians often do not assess their patients for IPV exposure [5,10,11]. Previous work has addressed barriers to clinician assessment for IPV. The barriers reported in the literature vary based on practice type and location, but recurring common themes include lack of time [12], fear of offending patients [10,12], fear of retaliation by the partner [11], and lack of confidence, training or inadequate resources [10,12]. Busy clinicians may simply forget to ask about IPV [10]. Despite these obstacles, most patients are willing to discuss the subject of partner violence with their medical care providers and believe that clinicians can be of assistance [13]. Abused women are even more likely than non-abused women to endorse screening [14]. Notwithstanding, universal screening for IPV remains controversial [15,16].
While prior research has examined barriers to clinician-initiated screening for IPV [10,12,17], less is known about how clinicians respond when information about a patient's IPV status is elicited for them. A better understanding of what patient or clinician characteristics drive an observed response to a positive IPV screen could inform ongoing efforts to train providers or even redirect resources in clinical settings. To address these issues, we conducted a two-part study to assess primary care provider (PCP) response to a positive screen for IPV. Using data collected in a mental health screening program at a community hospital and its affiliated health centers, we first conducted a medical record review of patients who had screened positive for IPV. Following the chart review, we surveyed PCPs to explore their attitudes and practices around intervention for patients who have experienced IPV.
Methods
The study took place at an urban network of publicly owned community hospitals and health centers that serves an ethnically diverse population of poor and middle-income patients. In 1997, the Department of Medicine developed a program to train family practitioners, internists and primary care nurse practitioners to identify and treat patients who were victims of IPV. This program was linked to a citywide program administered by the local Department of Public Health. The training initiative included publication of a manual for PCPs and a series of three Grand Rounds presentations in 1997.
In the spring of 1998, the Department began routine screening for mental health (MH) and IPV at most of its primary care sites. A series of three IPV screening questions were appended to the MH screen which was adapted from the PRIME-MD [18]. These questions were answered in a variable time period, usually in the waiting room, prior to the encounter with the PCP. The questions were:
1. Is violence/abuse a concern in your personal relationships (past or present)?
2. Has anyone hit, pushed, slapped or threatened you in the past year?
3. Is anyone in your personal life causing you to be fearful for your safety?
Question 1 queries lifetime abuse while question 2 asks about incidents occurring in the preceding 12 months.
Clinical encounters were not delayed if a patient did not have time to complete the screen. The instrument was initially only available in English; in the second year of the study it was translated into Spanish, Portuguese and Haitian Creole. Interpreters assisted patients with completion of the questions when necessary. During the patient visit, the PCP, a nurse practitioner (NP) or physician was asked to review the mental health screen along with the responses to the IPV questions. After evaluating the patient, the provider was directed to document an assessment and plan on a separate response form, and the clinic staff faxed this form back to the Quality Management department. The response form contained a checklist intended to facilitate documentation (Figure 1). A database of screen-positive patients over the two-year period 1998–2000 was compiled.
Figure 1 Checklist of Clinician Responses.
Chart review
In the first segment of the study, we reviewed the charts of patients who had screened positive for IPV in order to determine the nature and extent of clinician documentation following the positive IPV screen. Study standards for ideal chart documentation were derived from national consensus and state medical association guidelines that recommend detailed documentation of the circumstances of the abuse and of subsequent assessment, intervention and referral [2,19]. Thorough medical record documentation may provide important support to victims in legal settings [19]. Documentation in medical records also facilitates continuity of care between visits and among clinicians.
Four chart reviewers (two internal medicine physicians, a medical student and a dentist) abstracted data. We checked for inter-observer variability by having reviewers independently abstract sample charts: 90% agreement was demonstrated. As part of the review, we collected demographic data for both patient and provider characteristics. In addition, charts were examined for the following: any written acknowledgement of the positive IPV screen (whether on the faxed back response form or in the progress note), documentation of referral, safety planning and detailed descriptive circumstances of the abuse. Data on co-morbid psychiatric conditions were collected at the time of IPV screening and then extracted for our analysis from the same database compiled by Quality Management.
We next performed basic summary statistics analyzing patient and clinician demographic variables, psychiatric co-morbidities and chart review outcomes by response on each of the three screening questions to examine how these characteristics differed by category of IPV or concern for safety. Logistic regression analysis was then carried out to estimate two separate models examining which variables predict referral and safety planning for both lifetime and 12-month IPV.
Clinician survey
In the second phase of the project, we designed a provider survey to characterize PCP beliefs and practices, in order to better understand the findings from the initial chart review (see Additional file 1). One of the goals of this portion of the study was to measure the impact our locally designed departmental training had on clinicians' practice, so we chose to develop our own instrument rather than use an existing one. We referenced previous work that had queried clinicians' knowledge, attitudes and practice regarding IPV in clinical practice [20-22]. Informed by prior studies, we measured our clinicians' perception of self-efficacy and sense of their role in addressing IPV [20,21]. We queried clinicians about the barriers to addressing IPV reported in previous studies [10-12,17]. Additionally drawing upon prior work, we inquired about clinicians' screening practices and their estimate of IPV prevalence among their patient panels [10,11,21]. Since we hypothesized that clinician confidence was an important factor in addressing IPV in practice, we included several items directly comparing management of IPV to other common medical and behavioral health conditions; a similar comparison has been made in other survey studies for adverse health behaviors [21]. Our survey also measured the frequency with which clinicians accessed existing hospital services when working with patients who have experienced IPV. We collected demographic data from clinicians including gender, age, and year of graduation from professional school. We also queried self-report of attendance at the original training Grand Rounds and the number of hours spent training since 1997. The survey contained scales measuring attitudes toward IPV diagnosis, perceived barriers to caring for victims, and types of intervention and resources accessed.
Between April and July of 2002, the survey was distributed and returned by interoffice mail and at a department meeting to clinicians practicing adult primary care at the same clinical sites that participated in the screening program. This sample was deemed representative of the clinicians caring for the screen-positive patients in 1998–2000. We used multivariate regression analyses to estimate predictors of PCP confidence with IPV.
The hospital institutional review board (IRB) approved the two-part study. Analyses were conducted using STATA 8.0 Statistical software (Stata Corp., College Station, TX).
Results
Chart review
During the two-year period 1998–2000, 4.9% (115/2341) of patients at 11 different neighborhood health centers (NHC) and primary care clinics screened positive for IPV. Of the screen-positive patients, charts of 95 (83%) were located for review. Of the 95 charts reviewed, five contained documentation of abuse that was not partner-inflicted (for example, political violence in foreign countries). These five patients were excluded from further analyses. The final number of charts eligible for review was 90.
Of the ninety IPV+ charts reviewed, 72% contained some clinician documentation of the positive screen; however, nearly one-third of the charts (28%) reviewed contained no acknowledgement of the positive IPV screen. Only nine charts (10%) contained a detailed history of the abuse along with documentation of referral and safety planning (Figure 2).
Figure 2 Observed Chart Documentation of IPV (N = 90).
Seventy-two percent of the IPV+ patients were female and the mean age was 34 years (Table 1). Female patients comprised the majority of patients screening positive for both lifetime and 12 month IPV. The majority of IPV+ patients were white but more non-white patients endorsed a 12-month history of IPV (Table 1). Of the encounters reviewed, 64% were charted by physicians. All of the positive screens were from questionnaires completed in English. The majority of patients who screened positive for IPV were on Medicaid or Free Care (by definition these patients are of low income). More than three-quarters of the patients who answered affirmatively on the question querying fear for safety were female and of low income. The majority of patients screening positive for lifetime and 12-month IPV also screened positive for co-morbid mood and anxiety disorders. While a minority (23%) endorsed an eating disorder, more patients with 12-month IPV reported this condition.
Table 1 Covariates and IPV.
Covariate Total N (%) Lifetime IPV
N = 77 12 month IPV
N = 30 Fear for Safety
N = 9
Ethnicity
White = REF 51/90 (57) 47/74 (64) 13/29 (45) 4/8 (50)
Patient Gender
Female = REF 65/90 (72) 56/77 (73) 21/30 (70) 7/9 (78)
Age (mean years) 34 34 33 33
Insurance
Medicaid/Free Care = REF 69/90 (77) 60/74 (81) 21/29 (72) 8/9 (89)
Clinician Characteristics
Gender
Female = REF 65/90 (72) 58/77 (75) 18/30 (60) 8/9 (89)
Clinician Type
Physician = REF 58/90 (64) 51/77 (66) 19/29 (66) 4/9 (44)
Psych Screen
Eating DO 21/90 (23) 17/74 (10) 10/29 (34) 4/9 (44)
Mood 47/90 (52) 39/74 (53) 18/29 (62) 6/9 (67)
Anxiety 59/90 (66) 50/74 (68) 21/29 (72) 8/9 (89)
Chart Review Outcomes
Referral 44/90 (49) 40/76 (53) 14/30 (47) 7/9 (78)
Safety Plan 13/90 (14) 12/77 (16) 7/30 (23) 5/9 (56)
Data on the unavailable charts (n = 20) was abstracted from the hospital electronic medical record (EMR) system. The patients whose charts were unavailable for review appeared comparable to those reviewed: their mean age was 35 years and 71% were female. However, 71% of the patients whose charts were missing were described as white by the hospital EMR. Medical record number transcription error and misfiling were the most common reasons for charts to be unavailable for review. All of the sites had unavailable records, no clustering of missing charts by site was observed.
The determinants of referral and of safety planning are shown in Tables 2 and 3, respectively. We found that clinicians were more likely to refer patients (p < .05) who screened positively for mood or anxiety disorders or who had low income (Medicaid or Free Care insurance types), although the latter was not significant when adjusted for potential confounders (Table 2). Female clinicians' charts exhibited a higher proportion of documented referrals than those of their male counterparts, however this finding did not reach statistical significance. While nurse practitioners' charts had higher rates of documented referral than those of physicians, this difference also failed to achieve statistical significance.
Table 2 Predictors of Referral Documentation in Medical Records.
Covariate Univariate O.R. (95% C.I.) Lifetime Multivariate O.R. (95% C.I.) 12-month Multivariate O.R. (95% C.I.)
Screen for IPV
Lifetime IPV "Is violence/abuse a concern in your personal relationships? (past or present)" 2.5 (0.70–8.8) 6.93 (1.09–43.84) ____
12-month IPV "Has anyone hit, pushed, slapped or threatened you in the past year?" 0.85 (0.35–2.0) ____ 0.35 (0.94–1.30)
Fear for Safety "Is anyone in your personal life causing you to be fearful for your safety?" 4.10 (0.80–21) 2.35 (0.33–16.73) 6.31 (0.72–55.49)
Patient Demographics
Age 0.980 (0.94–1.02) 0.99 (0.93–1.04) 0.99 (0.94–1.05)
Ethnicity 1.37 (0.57–2.29) 2.41 (0.76–7.66) 2.16 (0.71–6.60)
Gender
Female = REF 1.35 (0.53–3.43) 0.49 (0.12–2.06) 0.54 (0.13–2.27)
Insurance
Free Care/Medicaid = REF 4.70 (1.40–15.80) 2.88 (0.74–11.07) 2.83 (0.73–10.93)
Clinician Characteristics
Female 2.15 (0.83–5.57) 2.53 (0.59–10.87) 2.39 (0.54–10.61)
Type
Physician = REF 0.56 (.23–1.37) 0.43 (0.13–1.45) 0.51 (0.16–1.62)
Site 0.97 (.84–1.12) 0.87 (0.70–1.07) 0.89 (0.73–1.10)
Screen for Co-morbid Psychiatric Conditions
Eating Disorder .88 (.33–2.36) ________ _______
Mood 2.59 (1.08–6.20) 3.61 (1.20–10.88) 3.64 (1.24–10.7)
Anxiety 2.99 (1.15–7.73) 4.25 (1.20–15.10) 4.02 (1.18–13.73)
Table 3 Predictors of Safety Planning Documentation in Medical Records.
Covariate Univariate O.R. (95% C.I.) Lifetime Multivariate O.R. (95% C.I.) 12-month Multivariate O.R. (95% C.I.)
Screen for IPV
Lifetime IPV "Is violence/abuse a concern in your personal relationships? (past or present)" 2.22 (.26–18.66) 1.14 (.05–23.72) ____
12-month IPV "Has anyone hit, pushed, slapped or threatened you in the past year?" 2.74 (.83–9.05) ____ 5.66 (0.79–40.52)
Fear for Safety "Is anyone in your personal life causing you to be fearful for your safety?" 11.40 (2.54–51.31) ____ ____
Patient Demographics
Age 1.00 (.95–1.06) 0.98 (.92–1.05) 0.98 (.91–1.04)
Ethnicity 0.93 (.28–1.06) 1.30 (.24–7.18) 1.17 (.21–6.71)
Gender† ____ ____ ____
Insurance
Free Care/Medicaid = REF 0.85 (0.21–3.49) 2.61 (.23–30.20) 2.81 (.20–38.76)
Clinician Characteristics
Female 2.34 (.48–11.4) 0.25 (.01–5.90) 0.27 (.01–7.68)
Type
Physician = REF 1.24 (0.35–4.40) 4.65 (.74–29.30) 5.28 (.74–37.38)
Site 1.32 (1.07–1.63) 2.18 (1.23–3.86) 2.44 (1.18–5.08)
Screen for Co-morbid Psychiatric Conditions
Eating Disorder 0.93 (0.23–3.77) 0.20 (.02–2.47) .07 (.003–1.78)
Mood 0.69 (.21–2.25) 0.08 (.01–.88) .03 (.001–.71)
Anxiety 1.70 (.43–6.74) 15.03 (1.56–143.89) 16.45 (1.31–206.42)
†Only female patients' chart contained documentation of safety planning.
The results for the Safety Planning model showed a different pattern (Table 3). Patient endorsement of question 3, expressing fear for safety, was associated with safety planning documentation in charts. Only female patients' charts contained any documentation of safety planning. The site, or NHC, where screening took place was significant in both univariate and multivariate models, suggesting that safety planning practices may be site-specific. As with referral documentation, screening for the psychiatric co-morbidities of mood or anxiety appear to result in higher rates of safety planning.
Survey
Seventy clinicians identified as eligible were surveyed. The overall response rate was 84%, 88% (40/45) for physicians and 76% (19/25) for nurse practitioners. The majority of respondents were physicians (57%). By self-report, nurse practitioners saw a mean of 10 cases annually while physicians reported 6 (p-value .057, 2-sample T-test).
The majority (81.6%) of PCPs strongly agreed that it is their role to inquire about IPV in the primary care setting. However, only 68% agreed that they were confident in their ability to diagnose and manage patients who had experienced IPV. By contrast, these clinicians reported high degrees of confidence with other medical and behavioral health conditions: 90% expressed comfort treating Chronic Obstructive Pulmonary Disease, 93% were comfortable with depression and 83% with substance abuse. A time constraint was the practice barrier endorsed most frequently in our study (67%).
The only factors significantly associated with provider confidence in IPV management were the number of hours of IPV training reported since 1997 and self report of frequency of treating patients who have experienced IPV. These two variables remained significant when adjusted for demographics and year of professional school graduation (Table 4).
Table 4 Determinants of provider confidence in IPV Management (Survey Data), Multivariate Linear Regression.
Variable Unadjusted β (p-value) Adjusted β (p-value)
Clinician Age -0.24 (.124) -0.26 (.290)
Clinician Type REF = MD 0.45 (.102) 0.22 (.521)
Years since professional school graduation -0.02 (.241) .004 (.871)
Provider gender REF = female -0.21 (0.423) -0.38 (.279)
Number hours trained since 1997 REF =< 3 -0.37 (0.029) -0.43 (0.025)
Clinical experience with IPV 0.47 (<.001) 0.29 (0.042)
Discussion
While prior studies in the primary care setting have focused on practice and frequency of provider-generated screening for IPV, our study evaluated an intervention that bypassed potential clinician barriers to initiating screening with use of a waiting room questionnaire. The finding that 4.9% of patients screened positive for IPV is consistent with other reports from primary care ambulatory settings [3-5]. However, the high proportion of males screening positive (31%) exceeds estimates in other studies. It is possible that some of these represent false positives. Review of chart narratives suggested that some of these male patients might actually have been batterers. Further study is needed to understand if battering behavior can be reliably detected in the primary care setting and whether detection would be efficacious.
While the majority of charts (72%) contained some documentation of the positive screen, nearly one-third of charts contained no medical record notation of the positive screen for IPV. This finding suggests that mandated waiting room screening is not sufficient to insure proper documentation and assessment by clinicians. Of the charts that did contain written acknowledgement of the positive screen, only 10% demonstrated adherence to the most complete level of documentation. While prior research has shown that medical records accurately reflect clinical reasoning [23], clinician behavior, particularly in sensitive matters, may not be fully represented by the medical record. At best, documentation is likely a proxy measure of clinicians' responses to a positive IPV screen. It is possible that appropriate care was, nevertheless, being delivered. In another study of an educational intervention in urban neighborhood health centers, documentation in the medical record did not change despite increases in the rate of screening and referral [24].
Unmeasured patient factors, such as degree of readiness to change, remain a possible explanation for the observed low rates of clinician documentation. This may be particularly true for documentation of referral and safety planning. Research applying the transtheoretical model of behavior change to abused women demonstrates that acknowledging that a relationship is abusive and accepting help often occur through a process of change [25,26]. It is possible that some of the patients identified through the screening program were not yet ready to engage in such a discussion with their clinicians or accept referrals. This phenomenon may underlie the lack of chart documentation. Possibly patient reluctance to have IPV documented may also have contributed to the low rates. Still, it is reasonable to expect that clinicians would, at a minimum, document the positive screen, even if the patient did not accept a referral or engage in a conversation about safety planning. In the earliest stage of precontemplation, research supports documenting suspicion about IPV [25].
Given the potential pitfalls of relying on medical record documentation, some investigators have recommended use of patient exit interviews as a more accurate measure of interventions for partner violence [22]. Yet for victims of IPV, medical record documentation is critical to protecting the patient's interests and safety [19] and thorough medical record documentation of IPV is recommended in national consensus guidelines [2].
When clinicians do document, our data suggests that certain patient factors – anxiety, mood, poverty and fear for safety – predict documentation by clinicians. Based on these findings, further study could determine whether questions that communicate a patient's perception of risk to clinicians result in improved documentation. Poverty may be a facilitating factor, as these patients had better insurance coverage for counseling than some privately insured patients did. Only female patients received safety planning. This finding may reflect a true subset of patients at high risk for violence and may also support the possibility that a proportion of the male patients screened falsely positive and were not at-risk victims. Gender bias on the part of clinicians is another potential explanation for this finding.
While both recent and lifetime IPV have important implications for health status, detection of recent IPV brings with it potential implications for a patient's safety. In this study, the 12-month IPV question was not predictive of referral documentation while the lifetime query was. It is unclear whether this finding relates to provider discomfort with the topic or deficiencies in training, both of which should be addressed in future work.
The provider survey results suggest that lack of confidence and experience with IPV, along with time constraints, may drive the observed variations in clinician approach to patients who have experienced IPV. Conversely, the majority of clinicians expressed comfort in their ability to manage depression and substance abuse. This difference is notable since these conditions, like IPV, have been considered sensitive and of a private nature in the past and might still carry a stigma for some patients.
Limitations
Our study has a number of limitations. The screening questions were not validated and generated false positive screens for IPV. The questions were narrow in scope and did not address the dimensions of emotional or sexual abuse, possibly failing to identify an important segment of the target population, and likely resulting in an underestimate of the true prevalence of IPV in our population. This may also impact the external validity of the study. Our sample was a convenience sample and we do not know the total number of patients who received the questionnaires in waiting rooms; the only screens counted were those returned by the clinical sites. Distributing the screening questions in the waiting room may have resulted in false negative responses, as some of the patients may have been accompanied by a battering partner or may have had concerns for confidentiality. The lack of positive screens on any instrument translated into another language raises concerns about literacy and cultural barriers to written disclosure of sensitive information. Data on gender preference was often not available in charts rendering us unable to determine whether a percentage of the males screening positive were men who have sex with men. In addition, PCP survey responses may have been biased to comply with the department's strongly stated position on the importance of IPV recognition and intervention. These responses may not be generalizeable to other departments that lack such a clear mandate. The urban setting of the study may also limit the generalizeability of our results to institutions in non-urban areas. Finally, our sample size was small, limiting the ability to detect statistically significant relationships among the variables and producing wide confidence intervals for some of the estimates.
Conclusion
When prompted by a positive IPV screen, clinicians frequently address this issue; however their chart documentation (and possibly their approach) is inadequate. They express less confidence in their ability to address IPV than other medical conditions that carry a similar social stigma. Why are clinicians so much more confident in managing other potentially stigmatizing conditions, such as substance abuse? One possible explanation is a shift in professional mores: conditions such as this one have evolved from being considered private, personal afflictions to diseases with a clear health impact. Only after a shift in the way these conditions were regarded did strategies for effective training and proven intervention evolve. It is only in the last decade that IPV has been recognized as a significant medical concern [27]. Alternative approaches are necessary to assist clinicians in adequately addressing IPV in office practice; as the results of this study demonstrate, waiting room screening alone is not sufficient. Potential areas for future investigation include the development and evaluation of experiential educational methods for clinicians and identification of the optimum frequency for retraining PCPs in proper documentation, referral and safety planning. Electronic medical record prompts may also aid busy clinicians and provide seamless, legible documentation of IPV. It is incumbent upon primary care clinicians to improve the identification and quality of care for the large number of patients they routinely see who experience partner violence.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MRG designed the study, oversaw data collection, performed the statistical analyses and drafted the manuscript. KSL participated in the design of the study and helped draft the manuscript. RCH contributed to data interpretation and manuscript revisions. DHB conceived of the study, contributed to data interpretation and helped revise the manuscript. All authors have 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
Clinician survey uploaded as Additional File 1.
Click here for file
Acknowledgements
The authors wish to thank Drs. Ethelind Cheng, Susan Phillip and Peggy Timothé, for invaluable assistance with data collection and survey development, and Drs. Stephanie Woolhandler, Erika Lichter and William Bachman for their advice regarding manuscript revisions.
==== Refs
Tjaden P Thoennes N Extent, nature and consequences of intimate partner violence: findings from the national violence against women survey 2000
Family Violence Prevention Fund (FVPF) National consensus guidelines on identifying and responding to domestic violence victimization in health care 2002 San Francisco: FVPF
Bauer HM Rodriguez MA Perez-Stable EJ Prevalence and determinants of intimate partner abuse among public hospital primary care patients JGIM 2000 15 811 817 11119174 10.1046/j.1525-1497.2000.91217.x
Gin NE Rucker L Frayne S Cygan R Hubell FA Prevalence of domestic violence among patients in three ambulatory care internal medicine clinics JGIM 1991 6 317 22 1890502
McCauley J Kern DE Kolodner K Dill L Schroeder AF De Chant HK Ryden J Bass EB Derogatis LR "The battering syndrome": prevalence and clinical characteristics of domestic violence in primary care internal medicine practices Ann Int Med 1995 123 737 46 7574191
Coker AL Davis KE Arias I Desai S Sanderson M Brandt HM Smith PH Physical and mental health effects of intimate partner violence for men and women Am J Prev Med 2002 23 260 268 12406480 10.1016/S0749-3797(02)00514-7
Hathaway JE Mucci LA Silverman JG Brooks DR Mathews R Pavlos CA Health status and health care use of Massachusetts women reporting partner abuse Am J Prev Med 2000 19 302 307 11064235 10.1016/S0749-3797(00)00236-1
Ulrich YC Cain KC Sugg NK Rivara FP Rubanowice DM Thompson RS Medical care utilization patterns in women with diagnosed domestic violence Am J Prev Med 2003 24 9 15 12554018 10.1016/S0749-3797(02)00577-9
Sansone RA Wiederman MW Sansone LA Health care utilization and history of trauma among women in a primary care setting Violence and Victims 1997 12 165 72 9403986
Elliott L Nerney M Jones T Friedman PD Barriers to screening for domestic violence JGIM 2002 17 112 116 11841526 10.1046/j.1525-1497.2002.10233.x
Rodriguez MA Bauer HM McLoughlin E Grumbach K Screening and intervention for intimate partner abuse: practices and attitudes of primary care physicians JAMA 1999 282 468 474 10442663 10.1001/jama.282.5.468
Sugg NK Inui T Primary care physicians' response to domestic violence: opening Pandora's box JAMA 1992 276 3157 3160 1593735 10.1001/jama.267.23.3157
Friedman LS Samet JH Roberts MS Hudlin M Hans P Inquiry about victimization experiences. A survey of patient preferences and physician practices Arch Int Med 1992 153 1186 90 1599346 10.1001/archinte.152.6.1186
Gielen AC O'Campo PJ Campbell JC Scollenberger J Woods AB Jones AS Dienemann JA Kub J Wynne EC Women's opinions about domestic violence screening and mandatory reporting Am J Prev Med 2000 19 279 285 11064232 10.1016/S0749-3797(00)00234-8
U.S. Preventive Services Task Force Screening for family and intimate partner violence: recommendation statement Ann Fam Med 2004 2 156 60 15083857 10.1370/afm.128
Ramsay J Richardson J Carter YH Davidson LL Feder G Should health professionals screen women for domestic violence? Systematic review BMJ 2002 325 314 12169509 10.1136/bmj.325.7359.314
Waalen J Goodwin MM Spitz AM Petersen R Saltzman LE Screening for intimate partner violence by health care providers: barriers and interventions Am J Prev Med 2000 19 230 237 11064226 10.1016/S0749-3797(00)00229-4
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 56 7966923 10.1001/jama.272.22.1749
Alpert EJ editor Massachusetts Medical Society (MMS) Campaign against domestic violence. Partner Violence: How to Recognize and Treat Victims of Abuse. A Guide for Physicians and other Health Care Professionals 1999 3 Waltham: MMS
Maiuro RD Vitaliano PP Sugg NK Thompson DC Rivara FP Thompson RS Development of a health care provider survey for domestic violence Am J Prev Med 2000 19 245 252 11064228 10.1016/S0749-3797(00)00230-0
Sugg NK Thompson RS Thompson DC Maiuro R Rivara FP Domestic violence and primary care Arch Fam Med 1999 8 301 306 10418535 10.1001/archfami.8.4.301
Thompson RS Rivara FP Thompson DC Barlow WE Sugg NK Maiuro RD Rabanowice DM Identification and management of domestic violence: a randomized trial Am J Prev Med 2000 19 253 263 11064229 10.1016/S0749-3797(00)00231-2
Kosecoff J Fink A Brook RH Chassin MR The appropriateness of using a medical procedure. Is information in the medical record valid? Med Care 1987 25 196 201 3821224
Harwell TS Casten RJ Armstrong KA Dempsey S Coons HL Davis M Results of a domestic violence training program offered to the staff of urban community health centers. Evaluation Committee of the Philadelphia Family Violence Working Group Am J Prev Med 1998 15 235 42 9791642 10.1016/S0749-3797(98)00070-1
Zink T Medical management of intimate partner violence considering the stages of change: precontemplation and contemplation Ann Fam Med 2004 2 231 239 15209200 10.1370/afm.74
Burke JG Gielen AC McDonnell KA O'Campo P Maman S The Process of ending abuse in intimate relationships Violence against Women 2001 7 1144 1163 10.1177/10778010122183784
Alpert EJ Domestic violence and clinical medicine: learning from our patients and from our fears JGIM 2002 17 162 163 11841533 10.1046/j.1525-1497.2002.11229.x
|
16297245
|
PMC1318461
|
CC BY
|
2021-01-04 16:29:13
|
no
|
BMC Fam Pract. 2005 Nov 19; 6:48
|
utf-8
|
BMC Fam Pract
| 2,005 |
10.1186/1471-2296-6-48
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-1411632421210.1186/1465-9921-6-141ResearchTGF-β1 and serum both stimulate contraction but differentially affect apoptosis in 3D collagen gels Kobayashi Tetsu [email protected] Xiangde [email protected] Hui Jung [email protected] Tadashi [email protected] Fu-Qiang [email protected] Shinji [email protected] Qiuhong [email protected] Yun Kui [email protected] John R [email protected] Peter [email protected] Stephen I [email protected] Department of Internal Medicine, University of Nebraska Medical Center, Omaha, Nebraska, USA2 Seoul Adventist Hospital and WonKwang University Sanbon Medical Center, Seoul, Korea3 Department of Respiratory Medicine, Graduate School of Medicine, University of Tokyo, Tokyo, Japan4 Department of Respiratory Medicine, West China Hospital, West China Medical School Sichuan University, Chengdu, Sichuan P.R. China5 The 4th Department of Internal Medicine, Nippon Medical School, Tokyo, Japan6 Department of Pulmonary and Critical Care Medicine, The First Hospital of Tsinghua University, Beijing, P.R. China7 Department of Respiratory Diseases, Jincheng Hospital, Lanzhou, P.R. China8 University of Minnesota, Minneapolis, Minnesota, USA2005 2 12 2005 6 1 141 141 13 4 2005 2 12 2005 Copyright © 2005 Kobayashi et al; licensee BioMed Central Ltd.2005Kobayashi 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.
Apoptosis of fibroblasts may be key for the removal of cells following repair processes. Contraction of three-dimensional collagen gels is a model of wound healing and remodeling. Here two potent inducers of contraction, TGF-β1 and fetal calf serum (FCS) were evaluated for their effect on fibroblast apoptosis in contracting collagen gels. Human fetal lung fibroblasts were cultured in floating type I collagen gels, exposed to TGF-β1 or FCS, and allowed to contract for 5 days. Apoptosis was evaluated using TUNEL and confirmed by DNA content profiling. Both TGF-β1 and serum significantly augmented collagen gel contraction. TGF-β1 also increased apoptosis assessed by TUNEL positivity and DNA content analysis. In contrast, serum did not affect apoptosis. TGF-β1 induction of apoptosis was associated with augmented expression of Bax, a pro-apoptotic member of the Bax/Bcl-2 family, inhibition of Bcl-2, an anti-apoptotic member of the same family, and inhibition of both cIAP-1 and XIAP, two inhibitors of the caspase cascade. Serum was associated with an increase in cIAP-1 and Bcl-2, anti-apoptotic proteins. Interestingly, serum was also associated with an apparent increase in Bax, a pro-apoptotic protein. Blockade of Smad3 with either siRNA or by using murine fibroblasts deficient in Smad3 resulted in a lack of TGF-β induction of augmented contraction and apoptosis. Contraction induced by different factors, therefore, may be differentially associated with apoptosis, which may be related to the persistence or resolution of the fibroblasts that accumulate following injury.
transforming growth factor-betaapoptosisgel contractionfibrosiswound repair
==== Body
Background
The development of fibrosis is thought to share a number of important features with normal wound repair. Both fibrosis and wound repair are characterized by the recruitment and activation of fibroblasts that differentiate to myofibroblasts [1-3]. These cells accumulate within tissue, produce extracellular matrix and remodel the local environment. Both fibrotic tissues and normal healing wounds are also characterized by myofibroblast contraction of extracellular matrix. Fibrosis, however, differs from normal wound healing in a number of important respects. Prominent among these, normal wound healing is characterized by the eventual resorption of much, if not all, of the excess connective tissue matrix and mesenchymal cells that characterize the healing phase [4]. In fibrosis, in contrast, normal tissue structures are permanently disrupted by excessive fibrotic material.
The three transforming growth factor-beta (TGF-β) isoforms are members of a family of signaling molecules [5]. TGF-β1 is believed to be a key factor in mediating both mesenchymal cell participation in wound repair and in a number of pathologic settings in fibrosis [6]. TGF-β is a potent activator of fibroblasts, inducing their differentiation into myofibroblasts and stimulating their production of extracellular matrix [7,8]. In in vitro experiments, TGF-β has been reported to inhibit fibroblast/myofibroblast apoptosis [9,10]. These in vitro experiments, however, have evaluated fibroblasts in monolayer culture. Culture of fibroblasts in three-dimensional collagen gels has been used as a system that more closely resembles tissues undergoing repair. These observations, therefore, raise an interesting and potentially important question: What would be the effect of TGF-β on the apoptosis of fibroblasts in three-dimensional collagen gel culture? Augmentation of contraction and in addition to apoptosis might lead to the net accumulation of contracted connective tissue and hence be a mechanism for the development of fibrosis.
TGF-β1 stimulates fibroblast contraction of extracellular collagenous matrices [11,12]. Interestingly, fibroblasts in a contracting matrix have been reported to undergo apoptosis [13,14]. The degree of apoptosis, moreover, has been associated with the degree of contraction in several studies [13-15]. The current study, therefore, was designed to determine the effect of TGF-β1 on fibroblast apoptosis in contracting three-dimensional collagen gels. TGF-β1 was found to stimulate both contraction of collagen gels and the apoptosis of fibroblasts in contracting gels. This contrasted with a slight inhibition of apoptosis in fibroblasts in three-dimensional gels that were constrained from contracting. It also contrasted with the effect of serum and PDGF, which stimulated contraction without stimulating apoptosis. These results, therefore, suggest that TGF-β1 may stimulate contraction of fibroblasts which, in turn, may lead to fibroblast apoptosis. Such a coordinated action may be a key feature of normal tissue repair by preventing the persistent accumulation of fibroblasts within tissues. These findings suggest that growth factors other than TGF-β may contribute to the contraction with persistence of fibroblasts that is noted in fibrotic tissues.
Methods
Materials and cell culture
Type I Collagen (rat tail tendon collagen [RTTC]) was extracted from rat-tail tendons by a previously published method [16]. Protein concentration was determined by weighing a lyophilized aliquot from each batch of collagen. The RTTC was stored at 4°C until use. Dulbecco's modified Eagle's medium (DMEM), fetal calf serum (FCS), trypsin/EDTA, penicillin G sodium, and streptomycin were purchased from Invitrogen (Life Technologies, Grand Island, NY). Amphotericin B was purchased from Pharma-Tek (Elmira, NY). The terminal transferase dUTP nick end labeling (TUNEL) assay kit was purchased from Roche Diagnostic Corporation (Indianapolis, IN). Goat anti-caspase 3 antibody (CRP32), which reacts with both precursor and active forms of human caspase 3, and goat anti-PARP, which reacts with both intact and cleaved forms of human PARP, rabbit anti-cIAP-1 antibody, mouse anti-XIAP antibody, recombinant human TGF-β1, PDGF-BB and anti-TGF-β1 antibody were purchased from R&D Systems (Minneapolis, MN). Mouse anti-Bcl-2 antibody and mouse anti-Bax antibody were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Rabbit anti-goat and mouse IgG horseradish peroxidase were purchased from Rockland Immunochemicals (Gilbertsville, PA). Propidium iodide, staurosporine and anti-β-actin antibody were purchased from Sigma (St. Louis, MO).
Human fetal lung fibroblasts (HFL-1) were obtained from the American Type Culture Collection (Rockville, MD). Smad2 knockout and corresponding wildtype, and Smad3 knockout and corresponding wildtype were kind gifts from Dr. A. Roberts (NIH). The Smad2 knockout (S2KO) mouse fibroblasts were established from mouse embryo-derived fibroblasts harboring the null allele Smad2Δex2 in the homozygous state, as described [17,18]. Smad3 knockout (S3KO) mice were generated by targeted deletion of exon 8 in the Smad3 gene by homologous recombination, as described [18,19]. The cells were cultured in 100-mm tissue culture dishes (Falcon; Becton-Dickinson Labware, Lincoln Park, NJ) in Dulbecco's Modified Eagle's Medium (DMEM), supplemented with 10% fetal calf serum (FCS), 50 U/ml penicillin G sodium, 50 μg/ml streptomycin sulfate, and 1 μg/ml amphotericin B. The fibroblasts were refed three times weekly, and cells between passages 15 to 18 for human and 34 to 45 for murine were used.
Small interfering RNA (siRNA) for Smad3 was designed to target the coding sequence of human Smad3 and effectively inhibits Smad protein expression as described previously [20]. siRNA for Smad2 and non-specific siRNA for control were purchased from Dharmacon (SMARTpool). Transfection of siRNA was also performed as described previously [20]. After 24 hours transfection, HFL-1 cells were harvested and used for gel contraction assay.
Three-dimensional collagen gel culture
Prior to preparing collagen gels as described below, fibroblasts were detached by 0.05% trypsin in 0.53 mM EDTA and suspended in 10 ml serum-free DMEM containing soybean trypsin inhibitor. The cell number was then counted with Coulter Counter. Collagen gels were prepared, as previously described [16], by mixing RTTC, distilled water, 4 × DMEM and cells. The final concentration was 1 × DMEM, 0.75 mg/ml of collagen, and fibroblasts were present at 3 × 105 cells/ml for human and 4.5 × 105 cells/ml for murine. Following this, 500 μl of the mixture was cast into each well of a 24-well culture plate (Falcon). The solution was then allowed to polymerize at room temperature, generally completed in 20 min. After polymerization, the gels were either allowed to remain attached to the plates in which they were case or, for the gel contraction assay, the gels were gently released from the plates in which they were cast and transferred into 60-mm tissue culture dishes (three gels in each dish), which contained 5 ml of SF-DMEM with or without FCS, TGF-β1 and PDGF-BB, respectively. The concentrations of TGF-β1 used were based on previous studies [21,22]. The area of each gel was measured daily with an image analyzer (Optomax, Burlington, MA). Data are expressed as the percentage of area compared with the initial gel area. For attached gels, gels were left attached in the plates and 1 ml of SF-DMEM with or without FCS or TGF-β1 was added. The gels were then incubated at 37°C in a 5% CO2 atmosphere.
DNA quantification
To estimate cell number in three-dimensional collagen gels, DNA was assayed fluorometrically with Hoechst dye no. 33258 (Sigma) by a modification of a previously published method [23]. Collagen gels were solubilized by heating to 60°C for 10 min and cell suspensions were collected by centrifugation at 2,000 × g for 5 min and resuspended in 1 ml of distilled water. After sonication, the suspensions were mixed with 2 ml of TNE buffer (3 M NaCl, 10 mM Tris, and 1.5 mM EDTA, pH7.4) containing 2 μg/ml of Hoechst dye no. 33258. Fluorescence intensity was measured with a fluorescence spectrometer (LS-5; Perkin-Elmer, Boston, MA) with excitation at 356 nm and emission at 458 nm.
Determination of apoptosis (TUNEL assay)
For determination of apoptosis, TUNEL assay was performed following manufacturer's instructions. Briefly, collagen gels were transferred from medium or plates attached to Eppendorf tubes (Fisher, Pittsburgh, PA) and then solubilized with heating at 60°C for 10 min. This method effectively solubilized the collagen gels without resulting in further DNA damage, as assessed by TUNEL assay (data not shown). Cell suspensions were collected by centrifugation at 2,000 × g for 5 min and resuspended in 150 μl of 10% FCS-DMEM. The resuspended cells were then used to prepare cytospins, 0.5 × 105 cells/spot, 1,000 × g for 5 min. Cytospin preparations were fixed with freshly prepared paraformaldehyde (4% in phosphate-buffered saline [PBS]; pH 7.4) for 1 h at room temperature. The cells were permeabilized with 0.1% Triton X-100 (in 0.1% sodium citrate) for 2 min at 4°C and rinsed with PBS. The TUNEL reaction was then performed using the manufacturer's instructions (Roche). The number of cells stained by the TUNEL method was expressed as a percentage of the total number of cells stained with the counterstain propidium iodide. At least 500 nuclei were counted on each cytospin sample in 5–10 randomly selected viewing fields.
Profile of DNA content by flow cytometry
For three-dimensional collagen gel culture, DNA content was analysed as described [24]. Briefly, fibroblast-populated (2 ml of 3 × 105cells/ml) collagen gels were cast into 6-well tissue culture plates (Falcon). After polymerization, gels were gently released and incubated with 1 % FCS-DMEM for 24 h, 100 pM TGF-β1 or with 1 μM staurosporine for 6 h (positive control). Gels were then transferred into 15-ml conical tubes and incubated with 0.05% Trypsin/0.53 mM EDTA-4Na (Invitrogen) for 10 min (500 μl/gel) at 37°C in a 5% CO2 atmosphere. Collagenase (1 mg/ml in DMEM) was then added (1 ml/gel) and incubated while shaking at 37°C in a 5% CO2 atmosphere for 30 min or until the gels were completely dissolved. DMEM containing 10% FCS was then added to stop the enzymatic reaction, and cells were pelleted by centrifugation. Cells were then fixed with Telford method and flow cytometry was performed as described below.
Flow cytometric analysis of DNA content was performed as previously described [25]. Briefly, cells were fixed with cold 70% ethanol in PBS for 30 min at 4°C. Cells were then pelleted by centrifugation and resuspended in the staining solution (50 μg propidium iodide, 100 μg RNAse A in 1 ml PBS for 106 cells) at 4°C for 1 h followed by flow cytometric analysis without washing. Since harvesting cells from the gels at day 5 results in formation of considerable debris which made the DNA profiling assay problematic, we chose day 1 for DNA profiling.
Western blot analysis
Three-dimensional collagen gel culture was performed as described above. After collecting cells by centrifugation, cells were washed with sterile PBS twice, and then put 100 μl cell lysis buffer (35 mM Tris-HCl, pH 7.4, 0.4 mMEGTA, 10 mM MgCl2, 100 μg/ml aprotinin, 1 μM phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, and 0.1% Triton X-100). Lysates were briefly sonicated on ice and centrifuged at 10,000 g for 3 minutes. The protein concentration in the cell lysates was measured using the BIO-RAD Protein Assay Kit. 10% SDS-polyacrylamide gel electrophoresis was performed under reducing conditions. To accomplish this, cell lysate proteins were diluted with 2× concentrated sample buffer (250 mM Tris-HCl, pH 6.9, 4% SDS, 10% glycerol, 0.006% bromphenol blue, 2% β-mercaptoethanol) and heated at 95°C for 5 minutes before loading (10 μg/lane). After SDS-PAGE, proteins were transferred onto PVDF membrane (BIO-RAD). The membrane was blocked for 1 h at room temperature with 5% skim milk in PBS-Tween and incubated overnight at 4°C with proper each antibody concentrations, respectively. After incubation with HRP-conjugated anti-Rabbit or mouse-IgG, an ECL Western blot detection system was used according to the manufacture's instruction (Amersham Biosciences, Piscataway, NJ).
Statistical analysis
Results are presented as mean ± SEM. Statistical comparison of paired data was performed using Student's t test, whereas multigroup data were analyzed by ANOVA followed by the Tukey's or Bonferroni's post-test using Statview software (Abacus Concepts Inc., Cary, NC). P < 0.05 was considered significant.
Results
Effect of FCS and TGF-β1 on fibroblast-mediated collagen gel contraction
Both FCS and TGF-β1 increased the contraction of collagen gels in a concentration-dependent manner over the period of observation. After 5 days, control gels (SF-DMEM) were 50.0 ± 1.1% of their initial area (Figure 1). In contrast, gels exposed to FCS (0.1% or 1%) were 21.6 ± 1.0% and 13.1 ± 0.1% of their original size after 5 d, respectively (Figure 1). Gels exposed to TGF-β1 (10 pM or 100 pM) were 32.8 ± 0.5% and 28.8 ± 1.5% of their original size after 5 d, respectively (Figure 1). The effect of FCS and TGF-β1 were both concentration- and time-dependent. Addition of anti-TGF-β antibodies did not alter the effect of serum but did completely block the effect of TGF-β (data not shown).
Figure 1 Effect of TGF-β1 and FCS on collagen gel contraction mediated by HFL-1 cells. Fibroblast-populated collagen gels were released into 60 mm tissue culture dishes with or without FCS or TGF-β1. Gel size was measured daily with an image analyzer. Vertical axis: gel size expressed as % of initial size. Horizontal axis: Time (days of culture). Both serum and TGF-β1 significantly augmented collagen gel contraction in a concentration-dependent manner. *p < 0.05 as compared with control. Data are shown as means ± SEM. Data presented are from one representative experiment of three experiments performed on separate occasions.
Effect of FCS and TGF-β1 on apoptosis
To determine the effect of FCS and TGF-β1 on fibroblast apoptosis, two methods were used. First, cells in three-dimensional collagen gels were cultured in SF-DMEM, 0.1% or 1%FCS-DMEM, 10 pM or 100 pM TGF-β1, and as an additional comparator 100 pM PDGF-BB for 5 days, and then TUNEL staining which measures DNA strand breaks, a feature of apoptosis cells, was performed (Figure 2). After 5 days, 11.6 ± 0.3% of control cells were TUNEL positive (Figure 3). TGF-β treated cells had increased TUNEL positivity while FCS treated cells had decreased TUNEL positivity. To quantify this, 500 cells from each condition were counted. In the presence of 0.1% FCS or 1% FCS, 10.3 ±0.5% and 7.1 ± 0.9% of the cells were TUNEL positive, respectively (Figure 3). PDGF-BB (100 pM) stimulated gel contraction similarly to TGF-β1 (data not shown) but did not result in increased apoptosis above control, 10.8 ± 0.4% of the PDGF-BB treated cells were TUNEL positive. In contrast, in the presence of 10 pM or 100 pM TGF-β1, TUNEL positive cell numbers were significantly increased to 22.3 ± 0.4% and 31.4 ± 1.4%, respectively (Figure 3) (p < 0.05, compared with control).
Figure 2 TUNEL staining in HFL-1 cells. Fibroblast-populated collagen gels were released into 60 mm tissue culture dishes with or without FCS or TGF-β1. On day 5, collagen gels were digested, cells isolated, cytocentrifuge preparation made, and stained by TUNEL. A: positive control (DNAse-1 treated), B: negative control (without terminal transferase), C: FCS free, D: FCS 1%, E: TGF-β1(100 pM). Red: PI stained normal cells. Green: TUNEL positive cells. Data presented are from one representative experiment. Similar results were obtained in three experiments performed on separate occasions.
Figure 3 TUNEL positivity in HFL-1 with FCS, TGF-β1 and PDGF-BB. After staining, TUNEL positive cells as a % of total cells were counted under the microscope in 5 high-power fields. Vertical axis: TUNEL positivity expressed as % of positive control (DNAse treated). Horizontal axis: condition. TGF-β1 increased TUNEL positivity. In contrast, FCS or PDGF-BB did not affect TUNEL positivity. *p < 0.05, as compared with control. Data are shown as means ± SEM. Data presented are from one representative experiment of three experiments performed on separate occasions.
To confirm the presence of apoptosis, profiling of DNA content was performed by flow cytometry. As a positive control, a group of gels were treated with staurosporin. After 24-hours, 1% FCS had tendency to decrease the amount of hypodiploid DNA compared to control cultures. In contrast, the TGF-β1 group increased the amount of hypodiploid DNA compared to control, indicating TGF-β1 increased apoptosis while FCS did not (Figure 4).
Figure 4 Representative profile of DNA content in fibroblasts. Collagen gels with fibroblasts were floated in (A) Staurosporine 1 μM for 6 hours, (B) SF-DMEM, (C) 1% FCS-DMEM and (D) TGF-β1 100 pM for 24 hours. Cells were then isolated and analyzed by flow cytometry. Vertical axis: cell number; horizontal axis: DNA content. The percentage of cells with hypodiploid DNA taken as an index of apoptosis is shown in each panel. Figure presented is from one representative experiment of three experiments performed on separate occasions. *p < 0.01. Data are shown as means ± SEM. Comparison of the means were done by one-way ANOVA.
Figure 5 DNA amount in contracting collagen gels. Fibroblasts were embedded in collagen gels and cultured in floating media containing 1% FCS or 100 pM TGF-β1 or control. DNA content, as a surrogate for cell number, was determined at the time of plating and after 5 and 10 days. *P < 0.05, as compared with SF-DMEM. Data are shown as means ± SEM.
Time course of cell numbers in three-dimensional collagen gel
To further confirm that apoptosis was occurring, the DNA amount, which can be used as a surrogate for cell number, was assessed in floating collagen gels. After casting gels in the presence of either serum-free DMEM, 1% fetal calf serum or 100 pM TGF-β1, DNA amount was assessed after 5 and after 10 days without further refeeding. As expected, DNA content decreased over time in control cultures incubated in DMEM alone. In the presence of 1% FCS, DNA amount decreased, but the decrease was statistically significantly less than that which occurred under control conditions (p < 0.05). In contrast, in the presence of TGF-β1, the decrease in DNA amount was larger than that which occurred in control (p < 0.05).
Effect of FCS and TGF-β1 on apoptosis related protein expression
A large number of proteins can serve as positive or negative regulators of the apoptosis process. To further confirm the differential effect of fetal calf serum and TGF-β1 on apoptosis, several apoptosis-related proteins were evaluated by Western blot (Figure 6). Staurosporin, which is an active control and induced apoptosis, increased the expression of Bax and induced the cleavage of both PARP and caspase 3, three markers of active apoptosis while it simultaneously inhibited the expression of Bcl-2, cIAP-1 and XIAP, three inhibitors of apoptosis. In contrast to the effects of staurosporin, 1% FCS stimulated the expression of Bcl-2, cIAP-1 and XIAP, the inhibitors of apoptosis, while it resulted in no cleavage of PARP or caspase 3. TGF-β1, in contrast, resembled staurosporin by increasing the expression of Bax and initiating the cleavage of PARP and caspase 3, all markers of active apoptosis, while it simultaneously inhibited the expression of Bcl-2 and XIAP.
Figure 6 Western blots of selected pro-apoptotic and anti-apoptotic factors. Fibroblasts were embedded in collagen gel and cultured in floating media with 1% FCS, 100 pM TGF-β1, staurosporine or control. After a day, collagen gels were digested, cells were collected, lysed and the cell lysate evaluated by Western blot. Data presented are from one representative experiment. Similar results were obtained in three experiments performed on separate occasions.
Effect of FCS and TGF-β1 on apoptosis in the attached gels
To determine if the effect of FCS and TGF-β1 on fibroblast apoptosis in collagen gels was related to contraction, cells in three-dimensional collagen gels were cultured in SF-DMEM, 1% FCS-DMEM or 100 pM TGF-β1 for 5 days and the gels were left attached to the plates, which prevents contraction. After this, the cultures were harvested and TUNEL staining was performed (Figure 7A). In contrast to contracting gels, 100 pM TGF-β1 did not significantly increase the percentage of TUNEL positive cells in attached gels. Similarly, in contrast to the effect on floating gels, TGF-β exposure had no effect in activating caspase 3 in gels that were constrained from contracting (Figure 7B).
Figure 7 TUNEL positivity in HFL-1 cells cultured in attached gels and Western blots for selected pro-apoptotic and anti-apoptotic factors. TUNEL Positivity. (A) Fibroblasts embedded in collagen gels which were left attached to the plates preventing contraction. After 5 days, gels were digested and stained for TUNEL. TUNEL positive cells were counted in 5 high-power fields and expressed as % of total cells. Data are presented as % of positive control (DNAse treated). Data are shown as means ± SEM. Western blot for selected pro-apoptotic and anti-apoptotic factors. (B) Collagen gels were digested, cells were collected, lysed and the cell lysate were evaluated by Western blot. Data presented are from one representative experiment.
Role of Smad2 and Smad3 in TGF-β induced apoptosis of fibroblasts in floating collagen gels
To determine the role of Smad2 and Smad3 on fibroblasts apoptosis, two methods were used. Murine lung fibroblasts from S2KO and S3KO and the corresponding wildtype (S2WT and S3WT) and HFL-1 cells incubated with siRNA targeting Smad2 and Smad3 were cultured in 3-D collagen gels with or without TGF-β1. As expected, TGF-β1 did not induced augmented contraction in Smad3 KO cells as previously described [11] or in Smad3 siRNA treated HFL-1 cells (data not shown). In contrast, TGF-β1 significantly augmented contraction in Smad2 KO cells in both wildtype controls [11] and in Smad2 siRNA treated and control HFL-1 cells (data not shown). After 5 days, TUNEL staining was performed. S2KO cells and both types of wildtype control cells as well as Smad2 siRNA treated and control HFL-1 cells had increased TUNEL positivity after TGF-β1 treatment (Figure 8). In contrast, TGF-β1 had no effect on TUNEL positivity in either Smad3 knockout mouse or Smad3 siRNA treated HFL-1 cells. Similarly, TGF-β did not result in the activation of caspase 3 in Smad3 siRNA treated HFL-1 cells (Figure 9).
Figure 8 TUNEL positivity and Western blot of selected pro-apoptotic and anti-apoptotic factors in murine fibroblasts and HFL-1 cells with or without TGF-β1. After staining, TUNEL positive cells as a % of total cells were counted under the microscope in 5 high-power fields. Panel A: Murine Smad3 KO and control cells; Panel B: HFL-1 cells ± siRNAs. Vertical axis: TUNEL positivity expressed as % of positive control (DNAse treated). Horizontal axis: condition. TGF-β1 increased TUNEL positivity in all cell types except in S3KO cells (Panel A) and Smad3 siRNA cells (Panel B). *p < 0.05, as compared with control. Data are shown as means ± SEM. Data presented are from one representative experiment of three experiments performed on separate occasions.
Figure 9 Western blot for selected pro- and anti-apoptotic proteins in HFL-1 cells treated with Smad3. Collagen gels made from Smad siRNA cells and control siRNA cells were digested, cells were collected, lysed and the cell lysate were evaluated by Western blot. Data presented are from one representative experiment repeated twice.
Discussion
The current study evaluated the survival of fibroblasts in contracting three-dimensional collagen gels. As expected, TGF-β1, PDGF-BB and serum all stimulated fibroblast-mediated contraction of three-dimensional collagen gels. TGF-β1 also stimulated apoptosis in the fibroblasts as assessed by both TUNEL assay and confirmed by DNA profiling to quantify cells with hypodiploid DNA content. In contrast, neither fetal calf serum nor PDGF-BB altered fibroblast apoptosis in contracting collagen gels. The stimulatory effect of TGF-β1 on apoptosis was associated with an increase in pro-apoptotic markers, including cleaved caspase 3, Bax and cleaved PARP, as well as inhibition of anti-apoptotic factors, including Bcl-2, cIAP-1 and XIAP. The ability of TGF-β1 to stimulate apoptosis required contraction of the three-dimensional collagen gels as no induction of apoptosis was noted in gels that were constrained from contraction.
TGF-β1 is one of three TGF-β isoforms that are members of a family of signaling molecules [5] TGF-β1 is believed to be a key factor in a variety of physiological and disease processes mediating a diverse range of cellular responses, including down regulation of inflammation, stimulation or inhibition of various cells types and regulation of differentiation of many target cells. TGF-β1 is believed to play a particularly important role as a mediator of wound healing [6]. TGF-β1 is a potent activator of fibroblasts stimulating fibroblast proliferation, production of extracellular matrix and differentiation into myofibroblasts. Because of these actions, TGF-β1 driven fibroblast activation is believed to play a major role in wound repair, scar formation and tissue fibrosis [26,27].
Tissue fibrosis differs from normal wound repair in several important features. While both are characterized by proliferation and accumulation of fibroblasts together with the extracellular matrix produced by these cells, normal granulation tissue is characterized by a resolution phase [28]. Specifically, as granulation tissue contracts, fibroblast apoptosis together with resorption of some of the collagenous extracellular matrix characteristically takes place. In fibrotic tissues, the severity of scarring and fibrosis, therefore, is dependent not only on the degree of fibroblast activation, but also on the relative lack of resolution. While the mechanisms that regulate resolution are incompletely understood, the current study supports the concept that TGF-β1 can drive fibroblast apoptosis concurrent with tissue contraction and that TGF-β1 differs from other growth factors in this regard. These results, which were obtained with fibroblasts cultured in three-dimensional collagen gels, contrast markedly with previous studies that evaluated fibroblasts cultured in monolayer culture where TGF-β inhibits apoptosis.
The members of the TGF-β family signal through a family of receptors, the activin receptors, which in turn signal through a family of signal transduction molecules, the Smads [29]. TGF-β signals primarily through the TGF-β RII (activin IIB) which phosphorylates the TGF-β RI (activin I). The activin I receptor, in turn, phosphorylates two Smad proteins, Smad 2 and Smad 3, which subsequently bind Smad 4 and mediate TGF-β signaling. While these represent the best characterized mechanisms for TGF-β signaling, other signaling pathways independent of Smad 2 and 3 have been reported [30]. The concentrations of TGF-β used in the current study were based on previous in vitro studies and are in the range expected for TGF-β to be active on its receptor. In vivo concentrations of TGF-β have been measured and are generally many-fold greater than those used. In vitro measurements, however, have generally assessed total TGF-β rather than the active form. Thus, while measures of in vivo active TGF-β concentrations are unavailable, the concentrations used in the current study are likely to be biologically relevant.
The culture of fibroblasts in three-dimensional collagen gels has been used for several decades as a model of tissue contraction that characterizes wound healing [1]. When cultured in floating collagen gels, fibroblasts attach to the collagenous matrix through integrin-dependent mechanisms and exert mechanical tension, which can cause floating gels to contract. In addition, concurrent with contraction, fibroblasts undergo apoptosis [13-15]. Interestingly, the amount of apoptosis is related to the amount of contraction [13,14]. Gels prepared with smaller concentrations of collagen, for example, undergo greater degrees of contraction, and a higher percentage of fibroblasts undergo apoptosis [14]. While the mechanisms that regulate apoptosis under these conditions are not fully established, cell spreading may play a role [31]. Specifically, cells that are not effectively spread are susceptible to apoptosis. Contraction, therefore, may be related to apoptosis induction. An effect of mechanical tension may also play a role. Finally, although our results suggest that contraction, per se, is related to induction of apoptosis, it is possible that other effects of TGF-β that also depend on Smad3 signaling mediate this effect.
Fibroblasts cultured in collagen gels can also proliferate. However, their response to growth factors in gel culture can be attenuated. Under the conditions used in the current assay, we have previously shown that there is minimal stimulation of proliferation with serum concentration 1% or less [16]. Serum contains many factors that can inhibit apoptosis [32], although the factors involved remained to be defined. Whether serum stimulation of contraction results from the same factor(s) that block apoptosis remain to be determined, although PDGF can do both. The overall effect of serum, however, contrasts with that of TGF-β. The link between TGF-β induced contraction and apoptosis may be a mechanism to prevent the accumulation of fibroblasts in resolving wounds. In contrast, the persistence of fibroblasts induced by other factor(s) present in serum may be a mechanism that contributes to scar formation or fibrosis.
The key finding of the current study is that augmented contraction induced by TGF-β is associated with apoptosis. This contrasts with augmented contraction induced by either PDGF or serum that is not associated with augmented apoptosis. These results suggest that contraction that takes place in the presence of TGF-β can be associated with apoptosis of fibroblasts. While TGF-β has been suggested to be a ''pro-fibrotic'' mediator because of its frequent association with both tissue injury and repair and with fibrotic processes and with its ability to activate fibroblasts, the present study suggests that TGF-β may stimulate fibroblasts in such a way that ''resolution'' is possible. The failure of apoptosis to occur in the presence of augmented contraction induced by PDGF and serum, however, suggests that other growth factors, that could function in collaboration with TGF-β, may be responsible for the persistence of fibroblasts and, hence, the development of fibrosis.
In order to determine the mechanisms by which TGF-β signaling leads to apoptosis, two approaches were used. TGF-β signaling was suppressed using siRNAs for either Smad 2 or Smad 3 and fibroblasts cultured from Smad 2 or Smad 3 deficient mice were compared with appropriate controls. As previously described [11], the absence of Smad 2 signaling had no effect on TGF-β1 or PDGF-BB stimulation of collagen gel contraction, while the absence of Smad 3 signaling blocked the ability of TGF-β1 to augment contraction, but not the ability of PDGF-BB to augment contraction. Using both siRNA and genetically deficient mice, loss of Smad 2 signaling had no effect on TGF-β1 augmentation of apoptosis, while loss of Smad 3 signaling blocked the ability of TGF-β1 to augment apoptosis. Thus, inhibition of apoptosis was always associated with inhibition of contraction.
The effect of TGF-β contrasted with the effect of serum which augmented contraction but did not stimulate apoptosis. These differing effects on apoptosis were paralleled by effects on apoptosis-related proteins. The mechanisms that prevent apoptosis in the presence of serum (or PDGF-BB) are unclear. In the present study, neither PDGF-BB nor serum affected apoptosis in a statistically significant manner. However, a small inhibition of apoptosis that did not achieve statistical significance was observed. Thus, it is possible that PDGF-BB or other growth factors could actively suppress apoptosis. In this context, the presence of serum was associated with an increase in cIAP-1 and Bcl-2, anti-apoptotic proteins. Interestingly, serum was also associated with an apparent increase in Bax, a pro-apoptotic protein. It seems likely, therefore, that factors present in serum may be able to affect the balance between pro- and anti-apoptotic factors and through such mechanisms could stimulate contraction while inhibiting apoptosis.
Apoptosis, or programmed cell death, is a highly regulated intracellular process. It can be initiated through several signaling mechanisms, including both activation of specific receptors as well as through non-specific effects such as DNA damage [33-35]. Apoptosis is regulated at several levels. Important among these is the proteolytic caspase cascade [36]. The caspases form a series of enzymatic reactions that, through successive cleavage events, can lead to the activation of caspase 3 which functions as a "cellular executioner." Concurrently, proteolytic cleavage can degrade the enzyme PARP which serves to maintain DNA integrity. The cleavage of PARP, an enzyme that mediates DNA repair, is believed to be an early step that commits a cell to death rather than DNA repair [37,38]. Similarly, cleavage of caspase 3 to its active form is believed to be a step that commits a cell to apoptosis as caspase 3 subsequently degrades many key cellular proteins. The commitment of a cell to apoptosis, therefore, can be regulated by controlling the activity of caspases. Several mechanisms exist by which this can be accomplished, including the release of the co-factor cytochrome C from mitochondria [39], which is both positively and negatively regulated by members of the Bax/Bcl family and by regulation through a family of inhibitors of caspases [40]. In this context, TGF-β1 induction of apoptosis in contracting three-dimensional collagen gels was associated with augmented expression of Bax, a pro-apoptotic member of the Bax/Bcl-2 family together with inhibition of Bcl-2, an anti-apoptotic member of the same family. Similarly, TGF-β1 was associated with inhibition of both cIAP-1 and XIAP, two inhibitors of the caspase cascade. The mechanisms by which TGF-β induces these effects are beyond the scope of the current proposal, but appears to require contraction of the three-dimensional collagen gels. This raises the possibility that the effect of TGF-β is indirect and may be related to cell spreading, for example.
In conclusion, the current study demonstrates that TGF-β1 induction of three-dimensional collagen gel contraction is associated with apoptosis. This induction of apoptosis requires contraction of the three-dimensional collagen gels and differs from other factors, including serum and PDGF-BB that induce contraction but not apoptosis. The ability of TGF-β to induce apoptosis may play a key role during wound repair. Abnormal regulation of apoptosis during the resolution phase following tissue repair could contribute importantly to both hypertrophic scar formation as well as to tissue fibrosis. The ability of tissues to contract normally may be important in this regard, and processes that increase mechanical tension in tissues or constrain contraction by other mechanisms may contribute to fibrosis and tissue remodeling. This study, therefore, supports the concept that TGF-β induction of fibroblast apoptosis is one of its many functions related to tissue repair and remodeling. Alterations in this function could contribute to the formation of hypertrophic scar or tissue fibrosis.
Acknowledgements
We appreciate the gift of murine Smad2 and Smad3 deficient fibroblasts provided by Dr. A. Roberts and the secretarial support of Ms. Lillian Richards. This work was supported by NIH remodeling grant and the Larson Endowment, University of Nebraska Medical Center.
==== Refs
Grinnell F Fibroblasts, myofibroblasts and wound contraction. J Cell Biol 1994 124 401 404 8106541 10.1083/jcb.124.4.401
Desmouliere A Gabbiani G Modulation of fibroblastic cytoskeletal features during pathological situations: the role of extracellular matrix and cytokines Cell Motil Cytoskeleton 1994 29 195 203 7895283 10.1002/cm.970290302
Ronnov-Jessen L Petersen OW Induction of a-smooth muscle actin by transforming growth factor-b1 in quiescent human breast gland fibroblasts. Lab Investigation 1993 68 696
Mutsaers SE Bishop JE McGrouther G Laurent GJ Mechanisms of tissue repair: from wound healing to fibrosis Int J Biochem Cell Biol 1997 29 5 17 9076937 10.1016/S1357-2725(96)00115-X
Massague J TGF-beta signal transduction Ann Rev Biochem 1998 67 753 791 9759503 10.1146/annurev.biochem.67.1.753
Blobe GC Schiemann WP Lodish HF Role of transforming growth factor beta in human disease N Engl J Med 2000 342 1350 1358 10793168 10.1056/NEJM200005043421807
Leask A Abraham DJ TGF-beta signaling and the fibrotic response Faseb J 2004 18 816 827 15117886 10.1096/fj.03-1273rev
Grotendorst GR Connective tissue growth factor: a mediator of TGF-beta action on fibroblasts Cytokine Growth Factor Rev 1997 8 171 179 9462483 10.1016/S1359-6101(97)00010-5
Zhang HY Phan SH Inhibition of myofibroblast apoptosis by transforming growth factor beta1. Am J Respir Cell Mol Biol 1999 21 658 665 10572062
Chen HH Zhao S Song JG TGF-beta1 suppresses apoptosis via differential regulation of MAP kinases and ceramide production Cell Death Differ 2003 10 516 527 12728250 10.1038/sj.cdd.4401171
Liu X Wen FQ Kobayashi T Abe S Fang QH Piek E Bottinger EP Roberts AB Rennard SI Smad3 mediates the TGF-beta induced contraction of type 1 collagen gels by mouse embryo fibroblasts. Cell Motility and the Cytoskeleton 2003 54 248 253 12589683 10.1002/cm.10098
Liu XD Umino T Ertl R Veys T Skold CM Takigawa K Romberger DJ Spurzem JR Zhu YK Kohyama T Wang H Rennard SI Persistence of TGF-beta1 induction of increased fibroblast contractility In Vitro Cell Dev Biol Anim 2001 37 193 201 11370814 10.1290/1071-2690(2001)037<0193:POTIOI>2.0.CO;2
Grinnell F Zhu M Carlson MA Abrams JM Release of mechanical tension triggers apoptosis of human fibroblasts in a model of regressing granulation tissue Exp Cell Res 1999 248 608 619 10222153 10.1006/excr.1999.4440
Zhu YK Umino T Liu XD Wang HJ Romberger DJ Spurzem JR Rennard SI Contraction of fibroblast-containing collagen gels: Initial collagen concentration regulates the degree of contraction and cell survival. In Vitro 2001 37 10 16
Fluck J Querfeld C Cremer A Niland S Krieg T Sollberg S Normal human primary fibroblasts undergo apoptosis in three-dimensional contractile collagen gels J Invest Dermatol 1998 110 153 157 9457911 10.1046/j.1523-1747.1998.00095.x
Mio T Adachi Y Romberger DJ Ertl RF Rennard SI Regulation of fibroblast proliferation in three dimensional collagen gel matrix. In Vitro Cell Dev Biol 1996 32 427 433
Heyer J Escalante-Alcalde D Lia M Boettinger E Edelmann W Stewart CL Kucherlapati R Postgastrulation Smad2-deficient embryos show defects in embryo turning and anterior morphogenesis Proc Natl Acad Sci U S A 1999 96 12595 12600 10535967 10.1073/pnas.96.22.12595
Piek E Ju WJ Heyer J Escalante AD Stewart CL Weinstein M Deng C Kucheriapati R Bottinger P Roberts AB Functional characterization of transforming growth factor beta signaling in Smad2- and Smad3-deficient fibroblasts. J Biol Chem 2001 276 19945 19953 11262418 10.1074/jbc.M102382200
Yang X Letterio JJ Lechleider RJ Chen L Hayman R Gu H Roberts AB Deng C Targeted disruption of SMAD3 results in impaired mucosal immunity and diminished T cell responsiveness to TGF-beta Embo J 1999 18 1280 1291 10064594 10.1093/emboj/18.5.1280
Kobayashi T Liu X Wen FQ Fang Q Abe S Wang XQ Hashimoto M Shen L Kawasaki S Kim HJ Kohyama T Rennard S Smad3 mediates TGF-beta1 induction of VEGF production in lung fibroblasts Biochem Biophys Res Commun 2005 327(2) 393 398 15629128 10.1016/j.bbrc.2004.12.032
Aoki F Kurabayashi M Hasegawa Y Kojima I Attenuation of Bleomycin-induced Pulmonary Fibrosis by Follistatin Am J Respir Crit Care Med 2005 172 713 720 15976370 10.1164/rccm.200412-1620OC
Kamaraju AK Roberts AB Role of Rho/ROCK and p38 MAP kinase pathways in transforming growth factor-beta-mediated Smad-dependent growth inhibition of human breast carcinoma cells in vivo J Biol Chem 2005 280 1024 1036 15520018 10.1074/jbc.M403960200
Labarca C Paigen K A simple, rapid and sensitive DNA assay procedure. Anal Biochem 1980 102 344 352 6158890 10.1016/0003-2697(80)90165-7
Kim H Liu X Kobayashi T Conner H Kohyama T Wen FQ Fang Q Abe S Bitterman P Rennard SI Reversible cigarette smoke extract-induced DNA damage in human lung fibroblasts Am J Respir Cell Mol Biol 2004 31 483 490 15256382 10.1165/rcmb.2002-0300OC
Gruenert DC Finkbeiner WE Widdicombe JH Culture and transformation of human airway epithelial cells Lung Cell Mol Physiol 1995 268 L347 360
Phan SH Kunkel SL Lung cytokine production in bleomycin-induced pulmonary fibrosis Exp Lung Res 1992 18 29 43 1374023
Zhang K Flanders KC Phan SH Cellular localization of transforming growth factor-beta expression in bleomycin-induced pulmonary fibrosis Am J Pathol 1995 147 352 361 7543734
Lorena D Uchio K Costa AM Desmouliere A Normal scarring: importance of myofibroblasts Wound Repair Regen 2002 10 86 92 12028520 10.1046/j.1524-475X.2002.00201.x
Huse M Muir TW Xu L Chen YG Kuriyan J Massague J The TGF beta receptor activation process: an inhibitor- to substrate-binding switch Mol Cell 2001 8 671 682 11583628 10.1016/S1097-2765(01)00332-X
Roberts AB Russo A Felici A Flanders KC Smad3: a key player in pathogenetic mechanisms dependent on TGF-beta Ann N Y Acad Sci 2003 995 1 10 12814934
Chen CS Mrksich M Huang S Whitesides GM Ingber DE Geometric control of cell life and death. Science 1997 276 1425 1428 9162012 10.1126/science.276.5317.1425
Tung PS Fritz IB Transforming growth factor-beta and platelet-derived growth factor synergistically stimulate contraction by testicular peritubular cells in culture in serum-free medium. J Cell Physiol 1991 146 386 393 2022693 10.1002/jcp.1041460308
Grafstrom RC Dypbukt JM Willey JC Sundqvist K Edman C Atzori L Harris CC Pathobiological effects of acrolein in cultured human bronchial epithelial cells. Cancer Res 1988 48 1717 1721 3349453
Baumgartner KB Samet JM Coultas DB Stidley CA Hunt WC Colby TV Waldron JA Occupational and environmental risk factors for idiopathic pulmonary fibrosis: a multicenter case-control study. Collaborating Centers Am J Epidemiol 2000 152 307 315 10968375 10.1093/aje/152.4.307
Kuwano K Hara N Signal transduction pathways of apoptosis and inflammation induced by the tumor necrosis factor receptor family Am J Respir Cell Mol Biol 2000 22 147 149 10657934
Shi Y Mechanisms of caspase activation and inhibition during apoptosis Mol Cell 2002 9 459 470 11931755 10.1016/S1097-2765(02)00482-3
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 10.1016/S1383-5742(02)00009-1
Virag L Szabo C The therapeutic potential of poly(ADP-Ribose) polymerase inhibitors Pharmacol Rev 2002 54 375 429 12223530 10.1124/pr.54.3.375
Chen M Wang J Initiator caspases in apoptosis signaling pathways Apoptosis 2002 7 313 319 12101390 10.1023/A:1016167228059
Adams JM Cory S The Bcl-2 protein family: arbiters of cell survival Science 1998 281 1322 1326 9735050 10.1126/science.281.5381.1322
|
16324212
|
PMC1318462
|
CC BY
|
2021-01-04 16:36:25
|
no
|
Respir Res. 2005 Dec 2; 6(1):141
|
utf-8
|
Respir Res
| 2,005 |
10.1186/1465-9921-6-141
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1481629318510.1186/1471-2407-5-148Research ArticleMembrane testosterone binding sites in prostate carcinoma as a potential new marker and therapeutic target: Study in paraffin tissue sections Dambaki Constantina [email protected] Christina [email protected] Marilena [email protected] Katherine [email protected] Michael [email protected] Ploutarchos [email protected] Panayiotis A [email protected] Elias [email protected] Efstathios N [email protected] Department of Pathology, University of Crete, School of Medicine, P.O. Box 2208, Heraklion, GR-71003, Crete, Greece2 Department of Experimental Endocrinology, University of Crete, School of Medicine, P.O. Box 2208, Heraklion, GR-71003, Crete, Greece3 Department of Urology, University of Crete, School of Medicine, P.O. Box 2208, Heraklion, GR-71003, Crete, Greece4 Department of Biochemistry, University of Crete, School of Medicine, P.O. Box 2208, Heraklion, GR-71003, Crete, Greece2005 17 11 2005 5 148 148 22 1 2005 17 11 2005 Copyright © 2005 Dambaki 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
Steroid action is mediated, in addition to classical intracellular receptors, by recently identified membrane sites, that generate rapid non-genomic effects. We have recently identified a membrane androgen receptor site on prostate carcinoma cells, mediating testosterone rapid effects on the cytoskeleton and secretion within minutes.
Methods
The aim of this study was to investigate whether membrane androgen receptors are differentially expressed in prostate carcinomas, and their relationship to the tumor grade. We examined the expression of membrane androgen receptors in archival material of 109 prostate carcinomas and 103 benign prostate hyperplasias, using fluorescein-labeled BSA-coupled testosterone.
Results
We report that membrane androgen receptors are preferentially expressed in prostate carcinomas, and they correlate to their grade using the Gleason's microscopic grading score system.
Conclusion
We conclude that membrane androgen receptors may represent an index of tumor aggressiveness and possibly specific targets for new therapeutic regimens.
==== Body
Background
The biological activity of androgen occurs predominantly through binding to an intracellular androgen receptor (iAR) protein, a member of the nuclear receptor superfamily, functioning as a ligand-activated transcription factor [1]. However, in recent years, a number of reports indicate additional rapid androgen actions, including the rapid activation of kinase-signaling cascades, modification of actin cytoskeleton and modulation of intracellular calcium levels [see [2,3], for reviews]. These actions are considered to be non-genomic in nature because they occur in cells lacking functional iARs in the presence of inhibitors of transcription and translation, or are observed too rapidly to involve changes in gene transcription [4].
We have previously identified membrane androgen receptor (mAR) sites in prostate and breast cancer cell lines [5-7]. In a preliminary work performed on freshly prepared epithelial cells from prostate carcinoma, non-cancerous peritumoral tissue, and benign prostate hyperplasia (BPH), we have shown that mARs are expressed preferentially on carcinoma cells [8]. Activation of mARs induces actin cytoskeleton polymerization and redistribution [5,9], secretion of Prostate Specific Antigen (PSA) and apoptosis [6,7]. Membrane androgen receptor sites were responsible for the induction of testosterone-BSA-induced apoptosis in T47D breast cancer cells [7]. It is interesting to note that mARs may be different from iARs, as they are not recognized by antibodies against the latter [5], and are not inhibited by a series of commonly used antiandrogens in vitro [9] or in vivo [6].
In the present work we developed a method for the in situ detection of mARs in formalin-fixed and paraffin-embedded tissues. The aim of the present study was to: (1) Validate a method for the detection of mARs in formalin-fixed paraffin-embedded specimens of prostate tumors, and (2) explore the distribution of mARs in prostate carcinomas. More specifically, we have tried to reply to the following questions: (a) Are mARs equally distributed in cases of prostate carcinoma and BPH, and (b) are mARs expressed in prostate carcinoma related to the gravity of the disease, as expressed by the Gleason's score?
Methods
One hundred and nine (109) cases of prostate carcinoma (age 45–93 years, mean ± SD 71.17 ± 9.02, median 71 years, followed for 4 to 72 months) and one hundred and three (103) cases of BPH (age 37–92, mean ± SD 70.44 ± 8.38, median 70 years), retrieved from the archives of the Pathology Department of the University Hospital of Crete, were analyzed retrospectively. Prostate carcinoma specimens corresponded to 23 radical prostatectomies, 15 transurethral resections harboring incidentally encountered carcinoma, and 71 transrectal biopsies. Three cases of carcinoma diagnosed in transurethral resection specimens were rejected for technical reasons. Survival of patients ranged from 1–72 months (mean ± SD 31.4 ± 9.32, median 29.5 months). Benign prostate hyperplasia cases corresponded to 96 transurethral resection and 7 prostatectomy specimens. Hematoxylin and Eosin (H&E) stained slides were reviewed by two investigators independently and blindly to the patients' clinical data. The Gleason's sum (combined score, Gleason's score) [10,11] of carcinoma cases was reevaluated by two observers. The results obtained from the histological study of carcinomas using the Gleason's grading system are presented in Table 1. In case of discrepancy between the two observers, the final decision was reached by consensus.
A set of three serial, 3 μm thick, tissue sections, embedded in Paramat extra (BDH Lab Supplies, Poole, UK), were taken on negatively charged slides (SuperFrost Plus, Kindler O GmbH, Freiburg, Germany) from each representative paraffin block. Two of the slides were used for the needs of mAR detection, while the third was stained with H&E for morphological study, and for grading in the case of carcinomas.
In order to evaluate the presence of mAR in tissue sections, we had to (partially) regenerate membrane proteins. In this aim, in an initial set of experiments, we have assayed different combinations of melting temperatures and times of incubation. Our findings indicated that the optimal conditions for the identification of the highest specific binding on the target molecules (mARs) were achieved by mild melting of the embedding medium (Paramat extra) at 42.5°C for 20 min, followed by dewaxing in xylene (three times, 10 min. each) and rehydration in decreasing concentrations of ethanol. Then, specimens were washed in distilled water for 20 min. and incubated at 37°C in a citrate buffer 0.01 M (pH 6.2) for two hours. Finally, they were washed in Tris Buffered Saline (TBS, 10 mM, NaCl 150 mM, pH 7.4) and processed for the detection of mARs. As in the present study we used albumin-conjugated steroids, and in order to minimize the non-specific binding of BSA we pre-incubated slides with 3% BSA for 40 min. Then the slides were washed in TBS and one of them (test slide) was incubated for 1 hour with 10-6 M Testosterone-BSA-FITC while the other (control slide) with 10-6 M BSA-FITC (both reagents were obtained from Sigma, St Louis, MO) in TBS.
Another potential source of non-specific staining could be the association of the ligand (testosterone) with iARs, due to the fact that microscopic tissue sections contain broken (sectioned) cells. Since our previous experiments have shown that classical antiandrogens react with iARs but not with mARs [6,9], we routinely used cyproterone acetate (a specific androgen receptor antagonist, Sigma, St Louis, MO) at a concentration of 10-4 M, diluted in the testosterone-BSA-FITC or BSA-FITC solution in order to block iARs (see Results). After one-hour incubation, slides were rinsed with TBS, coverslipped using Polyvinyl Alcohol Mounting Medium with Dabco Antifading (Fluka Biochemika) and examined under a fluorescence microscope.
An introductory comparative study using both a confocal laser scanning microscope (Leica TCS SP) and a fluorescence microscope (Nikon Microphot-FXA fluorescence microscope, mounted with a Nikon DX-DB2 photographic camera) showed no significant advantage of the confocal over the conventional fluorescence microscope. The latter was therefore used for the routine detection of mARs. Results were expressed as a percentage of mAR positive cells in reference to total number of cells examined in 10 sequential high power fields (X400) in a representative area of each test slide. The localization and pattern of staining were recorded.
Clinical data, namely age, previous treatment, follow-up and survival, were retrieved from the patients' files. Statistical analysis was performed by the Systat V 10.0 (SPSS, Chicago, IL) program. In the case of positive/negative cases of BPH and prostate carcinoma, we have used significance limits of the four-fold table test, using the χ2 test, with 1 degree of freedom. The same test was used for the comparison of positive and negative cases in each group. For exploring the correlation between positive/negative cases of mAR expression and other parameters (age, treatment, survival, Gleason's sum) the Spearman rank correlation coefficient was calculated. The Mann-Whitney U test was used to compare the distribution of continuously-scaled outcomes by mAR expression. All statistical evaluations were double-sided.
This study was approved by the Ethics Committee of the University Hospital of Heraklion, and the Ethics Committee of the University of Crete.
Results
1. Validation of the method
Previous results indicated that mARs can be detected in freshly prepared prostate carcinoma tissue [8]. The method presented here for the in situ detection of mARs is based on the binding of a ligand (FITC-BSA-bound testosterone) on mARs. Membrane androgen receptors are membrane proteins [5,12], expected to be (partially) denatured in microscopic tissue sections by formalin fixation, paraffin embedding and high temperatures, conditions known to modify integral membrane proteins. Therefore, we attempted to renaturate these proteins by incubating the tissue sections at relatively low temperatures during deparaffinization and renaturation process (mild tissue dewaxing and protein renaturation). Different combinations of melting temperatures (from 40° to 60°C), incubation times, xylene treatment and acidification were tried. As indicated in the Methods section, the best results were obtained with a mild paraffin melting and protein renaturing temperature of 42.5°C, followed by mild acidification in citrate buffer to dissociate any ligands loosely-bound to mARs or iARs [13] and extensive washing with TBS pH 7.4.
When slides were incubated with testosterone-BSA-FITC without any treatment, they presented a high non-specific binding (Figure 1, upper panel), due probably to the non-specific absorption of BSA to membrane and/or intracellular structures or molecules. For this reason, slides were preincubated with a 3% BSA solution prior to ligand binding. As shown (Figure 2A and 2B) this treatment resulted in an attenuation of conjugated or unconjugated BSA-FITC to membranes.
Another matter of concern was that iARs, regularly present in prostate carcinoma cells, could bind the ligand since, in tissue sections, one expects an interaction of the conjugated ligand with membrane and (due to sectioning) intracellular structures and/or molecules, including iARs. Our previous results with prostate cancer cell lines have shown that interaction of testosterone-BSA with mARs is not influenced (at the level of binding or the resulting signaling cascade and secretion) by antiandrogens [6,9] Therefore, to eliminate a possible interaction of BSA conjugated ligand with iARs, we used 10-4M cyproterone acetate in the binding experiments.
This treatment decreased nuclear, membrane and intracellular staining (Figure 1). The treatment of slides with testosterone-BSA-FITC in the presence of 3% BSA and 10-4M cyproterone acetate decreased non-specific staining in nearly background levels, making more discernible the interaction of testosterone-BSA-FITC with specific membrane-binding sites (compare Figure 1 with Figure 2A right panel).
According to the above results, for the detection of mARs in tissue sections, we applied BSA (3%) and cyproterone acetate (10-4M) throughout the present study. These conditions were found to accurately discriminate between positive (>75% of cells labeled by testosterone-BSA-FITC) and negative (<10% of cells labeled by testosterone-BSA-FITC) cases.
2. Prostate carcinoma specimens preferentially express mARs
On examination of the specimens, positive cases of both carcinomas and BPHs were found to present peripheral, membrane type, fluorescence. In contrast, morphologically normal-looking prostate epithelial tissue showed no specific fluorescence. In Figure 2, representative cases of prostate carcinoma are presented, showing a positive (Figure 2A) or negative (Figure 2B) staining for mARs. Positive cases showed peripheral, continuous and/or stippled, membrane type, fluorescence. This extended over more than 75% of carcinoma cells of the tissue section. Negative cases, on the other hand, showed no more than 10% reacting cells. Intensity of staining in positive cells was not considerably different among cases. Thus, reading and translating mAR positivity seems to be a matter of nearly a "black and white" effect and membrane pattern recognition.
The tissue stroma sometimes showed a diffuse positive reaction, probably due to connective tissue non-specific binding. This non-specific staining did not interfere, though, with the estimation of mAR expression in epithelial cells of the tissue section. Ninety percent (90%) of BPHs were negative for mAR, while about 38% of carcinomas were positive (Figure 3) (χ2 = 23.24, p < 0.0001). No correlation was found between mAR expression and age of the patient, treatment before or after surgery and survival.
3. Membrane Androgen Receptor sites are related to the level of differentiation of prostate carcinoma
As mentioned above, 38% of prostate carcinomas were found to express mARs (χ2 6.68, p < 0.1, among all cases of carcinoma, n = 109). Thus, mAR expression cannot be used for the discrimination between malignant and benign prostate cases. To evaluate the possible role of mARs in the evolution of prostate carcinoma, we correlated the presence or absence of mARs to the differentiation of the prostate carcinoma, estimating the Gleason's sum of each case on the corresponding H&E stained slide. Using correlation analysis, our results showed that mAR expression is positively related to higher Gleason's sum (Mann-Whitney U = 991, p < 0.04, Spearman's rho = 0.236, p < 0.01, Figure 4). It is worthy to note that 50% of poorly differentiated carcinomas (Gleason's sum 8 and 9, in our series) expressed mARs while in the Gleason's sum 9 carcinomas, mARs expression predominated. As seen in Figure 4, there is an increased expression of mARs with increasing Gleason's sum (15%, 40%, 44%, 36%, 71%, for Gleason's sums from 5 to 9 respectively). No difference was found between Gleason's sum 4+3 and Gleason's sum 3+4 regarding mAR expression (11 out of 21 cases of Gleason's sum 3+4, and 9 out of 20 cases of Gleason's sum 4+3 expressed mAR).
Different percentages of mAR positivity were revealed when prostate carcinoma cases were stratified according to the type of tissue sample. The number and percentages of positive carcinoma cases in each type of specimen is presented in Table 2. Radical prostatectomies showed 13.04% positivity and transurethral resection specimens showed 26.22%, while almost half of the transrectal biopsies were found to express mARs (47.80%). The number of transrectal biopsies (68 studied) was higher correlated to the number of cases of the other two groups (radical prostatectomies 23 and transurethral resections 15). Statistical analysis showed that there is a significant difference in mAR expression between the three types of tissue samples (binomial distribution comparison, p < 0.01).
Discussion
The classical mode of action of steroid hormones includes their entrance into hormone-dependent cell and binding to their intracellular cognitive receptors, which subsequently dimerize and translocate to the nucleus, acting as specific transcription factors [14]. This sequence of events requires time for their completion (>2 hours). In recent years, however, steroids were found to exert a number of effects in seconds or minutes, a phenomenon not explained by their intracellular action. Ion mobilization, activation of specific signaling cascades or secretion are included among these rapid actions and are considered non-genomic, independent of transcription or translation [[15], see [16], for a review]. A number of possible explanations for the non-genomic steroid actions include the presence of membrane binding sites, different from intracellular receptors [17], the anchorage of intracellular receptors to the plasma membrane through post-translational modification of receptor molecules [18], the interaction of intracellular steroid receptors with other membrane-bound receptors, or the activation of signaling molecules through intracellular receptor-bound steroids.
Recently, a progesterone binding membrane site has been isolated and found to belong to the seven-loop G-protein coupled receptor superfamily [19]. Data on other steroids binding membrane sites have also been reported [2,4,20-22], although their molecular nature has not been elucidated until now. More recently, we have reported the identification of mARs in prostate cancer cell lines [5,6] and in a small series of human prostate carcinomas [8]. Membrane androgen receptor sites activation induces Ca2+ flux [23], actin reorganization [9] and PSA secretion [5], in an antiandrogen-independent manner, inducing apoptosis in vitro [6,7] and in vivo [6]. In addition, tritiated testosterone has been found to bind efficiently and specifically to acid-stripped membranes [5].
The above data indicate that the detection of mARs could be of value in the diagnosis; and possibly mARs themselves could be exploitable in a new therapeutic approach of the mAR positive prostate carcinomas, as prostate carcinoma xenografts in nude mice are significantly reduced in size by testosterone-BSA treatment of animals [6]. In the present study we investigated whether mARs could also be detected in formalin-fixed and paraffin-embedded specimens of prostate carcinoma. The main disadvantage in mARs detection is that, until the elucidation of their molecular structure, no antibodies interacting with them are available. In addition, antibodies to iAR do not cross-react with mAR [7]. Therefore, for the time being, the only appropriate method for mAR detection is that based on the use of labeled non permeable ligands, such as testosterone-BSA-FITC. Previous results of experiments of our team performed in formalin- or PFA-fixed cell lines (LNCaP, DU-145, and PC12) and circulating leukemic lymphocytes have shown that mAR could be effectively detected on the cell surface. In view of these data, in the present study we used mild deparaffinization, in combination with citrate buffer pretreatment, known to induce dissociation of ligands or iAR loosely-bound on membrane sites [7,13]. Our results indicate that, under such conditions, mARs could be detected in prostate carcinomas on formalin-fixed and paraffin embedded tissue sections.
A problem encountered in our attempt to reveal in situ mARs was that BSA-FITC bound non-specifically to/or it was absorbed by stromal structures or cell and/or intracellular components. In order to decrease this non-specific interaction, we pre-incubated tissue sections on slides with high concentrations of BSA, thus eliminating or efficiently decreasing this interaction. Intracellular ARs (iARs) do not interfere with the detection of mARs by the method presented here. Indeed, as found in our previous studies, these latter sites do not react with antiandrogens, which can block intracellular receptors [6,9]. We have therefore exploited this differential antiandrogen selectivity in order to block selectively iAR binding, by the use of high concentrations of the antiandrogen cyproterone acetate. In addition, we report that, although confocal microscopy could be considered more precise and accurate than the -cheaper in hardware- conventional fluorescence microscopy, once the researcher becomes accustomed to mAR expression pattern, conventional fluorescence microscopy can be a valuable alternative in revealing the presence of mAR in tissue sections.
Our findings indicate that intensity is not a factor to take into account in the evaluation of mAR expression, since, in individual positive cases, a rather uniform staining intensity was observed. In the case of advanced prostate carcinomas there is a heterogeneity of cell populations expressing iAR [24,25], suggesting iAR counting in hot spots. Such heterogeneity was not observed in our study of mARs, and thus positivity was estimated in 10 sequential high power fields in representative tissue section regions (X400).
The most appropriate samples for the study of mAR expression seem to be the transrectal biopsies (50% positivity), which constituted most of our specimens (71 out of 106 finally studied), followed by the TUR specimens (26.66% positivity), and the radical prostatectomy specimens (13.04% positivity). This signifies that the quality of fixation (as depended on time and room temperature) is of critical importance in the study of mARs.
We report that only 38% of prostate carcinoma cases are positive for mAR staining. In this respect, mAR positivity could not be a discriminant factor for prostate cancer diagnosis. Stratification of cases, according to the Gleason's sum revealed that mAR-positive staining correlates with less differentiated tumors. No significant difference in mAR expression between Gleason's sum 4+3 (9 out of 20 cases) and Gleason's sum 3+4 (11 out of 21 cases) prostate carcinomas was found. Nevertheless, no solid conclusions can be drawn, due to the relatively small number of cases studied.
Membrane androgen receptors are not only expressed in the prostate. They have been found in osteoblasts, lymphoid cells and Sertoli cells [reviewed in [12]] independently of the presence or not of iARs. Preliminary data from our group studies indicate that mARs are expressed in higher grade carcinomas of the breast, B-lymphomas, pheochromocytomas and in AC133+ cord blood stem cells (unpublished observations). In addition, in estrogen receptor negative (ER-negative) carcinomas of the breast we have found a strong staining for mAR [7]. In this respect, mAR might be considered rather as a marker of "immature" and behaviorally aggressive cells than a tissue specific marker. If this comes true, it could explain the finding that tumors with high Gleason's sum preferentially express mARs.
Conclusion
Our results indicate that mARs detection cannot discriminate between prostate carcinoma and BPH, not being exploitable in the differentiation between low-grade carcinomas and BPHs. Nevertheless mARs could be used for a biologic grading of prostate carcinomas, since the more aggressive ones-of higher histopathological grade- preferentially express mARs. Results of recent studies of our team indicate that agonists of mARs-being expressed in cases of both iAR-positive and iAR-negative prostate cancer cell lines-might act as specific therapeutics directing carcinoma cells towards apoptosis [6]. In addition, the activation of mAR-signaling cascade leads to apoptosis [6], while activation of iARs is antiapoptotic [26]. This, in combination with the finding that prostate carcinomas of low Gleason score have a significantly higher iAR content than those with high Gleason score [27], signifies a different biological role for mARs in relation to iARs. Hence, the possibility that combined examination of iARs and mARs in prostate carcinomas, in the context of a test clinically applicable and easy to perform, could lead to a more accurate biologic grading of predictive and/or prognostic importance. To test the prognostic power of mARs, further studies in prostate carcinomas are needed. Moreover, mARs' study in large series of prostate carcinomas is imperative in order to fully elucidate the pathway of their action and reveal whether they represent a possible target for new therapies of these carcinomas, especially in view of our results on regression of prostate tumor xenografts by testosterone-BSA [6].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CD, CK, MK and KD have performed the collection, staining and evaluation of cases; MN and PA collected the clinical data, PAT, EC and ENS supervised and controlled the whole study
Pre-publication history
The pre-publication history for this paper can be accessed here:
Figures and Tables
Figure 1 Prostate carcinoma staining with testosterone-BSA-FITC. Representative patterns of prostate carcinoma tissue section staining with testosterone-BSA-FITC in absence of BSA and cyproterone acetate preincubation (upper panel) or in presence of cyproterone acetate and in absence of BSA (lower panel). As explained in the text, BSA preincubation decreases or eliminates non-specific interactions of the tracer (testosterone-BSA-FITC) with prostate stroma or cell membranes, while the antiandrogen cyproterone acetate binds selectively to iAR.
Figure 2 Representative cases of mAR positive (A) and negative (B) tumor specimens. A representative case of mAR positive (A) and a negative case (B) are shown. Left panels present H&E staining, middle panels present BSA-FITC staining, while right panels present staining with testosterone-BSA-FITC. Note the membrane localization of fluorescence in positive cells. Lower panel presents a higher magnification of the square region shown. In all cases preincubation with 3% BSA and 10-4M cyproterone acetate was performed, prior to mAR detection. Compare Figure 2 with Figure 1 (no preincubation).
Figure 3 Distribution of positive and negative mARs in BPH and prostate carcinoma. Ninety % of benign prostate hyperplasia cases (n = 103) were negative for mAR receptor expression. On the other hand, prostate carcinoma cases (n = 106) showed mAR expression in a significant percentage (38% of these cases were positive) (χ2 = 16.7, p < 0.0001).
Figure 4 Correlation of mAR positivity with Gleason's sum in prostate carcinomas. Membrane androgen receptors are preferentially expressed in prostate carcinomas and their presence is correlated to Gleason's sum.
Table 1 Type of specimens and their corresponding Gleason's sum.
Type of specimen Gleason's sum
2 3 4 5 6 7 8 9 10 Total number
3+4 4+3
Radical prostatectomy 2 0 2 2 7 2 4 2 2 0 23
Transurethral resection 2 0 1 5 0 2 0 2 3 0 15
Transrectal biopsy 1 0 3 3 18 17 16 9 1 0 68
Total 5 0 6 10 25 21 20 13 6 0 106
Table 2 Type of specimens and results of the study for mAR expression
Type of specimen Positive Negative Rejected Total % of positivity
Radical prostatectomy 3 20 0 23 13,04
Transurethral resection 4 11 0 15 26,66
Transrectal biopsy 34 34 3 71 50,00
Total 41 65 3 109 38,60
==== Refs
Kumar MV Tindall DJ Transcriptional regulation of the steroid receptor genes. Prog Nucleic Acid Res Mol Biol 1998 59 289 306 9427846
Falkenstein E Wehling M Nongenomically initiated steroid actions Eur J Clin Invest 2000 30 Suppl 3 51 54 11281368 10.1046/j.1365-2362.2000.0300s3051.x
Schmidt BM Gerdes D Feuring M Falkenstein E Christ M Wehling M Rapid, nongenomic steroid actions: A new age? Front Neuroendocrinol 2000 21 57 94 10662536 10.1006/frne.1999.0189
Losel R Wehling M Nongenomic actions of steroid hormones Nat Rev Mol Cell Biol 2003 4 46 56 12511868 10.1038/nrm1009
Kampa M Papakonstanti EA Hatzoglou A Stathopoulos EN Stournaras C Castanas E The human prostate cancer cell line LNCaP bears functional membrane testosterone receptors that increase PSA secretion and modify actin cytoskeleton Faseb J 2002 16 1429 1431 12205037
Hatzoglou A Kampa M Kogia C Charalampopoulos I Theodoropoulos PA Anezinis P Dambaki C Papakonstanti EA Stathopoulos EN C S Gravanis A Castanas E Membrane androgen receptor activation induces apoptotic regression of human prostate cancer cells in vitro and in vivo. J Clin Endocrinol Metab 2005 90 893 903 15585562 10.1210/jc.2004-0801
Kampa M Nifli AP Charalampopoulos I Alexaki VI Theodoropoulos PA Stathopoulos EN Gravanis A Castanas E Opposing effects of estradiol- and testosterone-membrane binding sites on T47D breast cancer cell apoptosis. Exp Cell Res 2005 307 41 51 15922725 10.1016/j.yexcr.2005.02.027
Stathopoulos EN Dambaki C Kampa M Theodoropoulos PA Anezinis P Delakas D Delides GS Castanas E Membrane androgen binding sites are preferentially expressed in human prostate carcinoma cells BMC Clin Pathol 2003 3 1 12588669 10.1186/1472-6890-3-1
Papakonstanti EA Kampa M Castanas E Stournaras C A rapid, nongenomic, signaling pathway regulates the actin reorganization induced by activation of membrane testosterone receptors Mol Endocrinol 2003 17 870 881 12554777 10.1210/me.2002-0253
Gleason DF Tannenbaum M Histologic grading and clinical staging of prostatic carcinoma. Urologic pathology: the prostate 1977 Philadelphia, Lea & Febiger 171 197
Rosai Ackermans Surgical Pathology 2004 1 Edinburgh, Mosby
Heinlein CA Chang C The roles of androgen receptors and androgen-binding proteins in nongenomic androgen actions. Mol Endocrinol 2002 16 2181 2187 12351684 10.1210/me.2002-0070
Hatzoglou A Romain S Dalapascha H Castanas E Acidification reveals a greater number of epidermal growth factor receptors in human membrane preparations. Clin Chim Acta 1994 227 97 109 7955425 10.1016/0009-8981(94)90139-2
Zhou ZX Wong CI Sar M Wilson EM The androgen receptor: an overview Recent Prog Horm Res 1994 49 249 274 8146426
Simoncini T Genazzani AR Non-genomic actions of sex steroid hormones Eur J Endocrinol 2003 148 281 292 12611608 10.1530/eje.0.1480281
Kampa M Castanas E Membrane steroid receptor signaling in normal and neoplastic cells. Mol Cell Endocrinol 2005 (in press)
Benten WP Lieberherr M Sekeris CE Wunderlich F Testosterone induces Ca2+ influx via non-genomic surface receptors in activated T cells FEBS Lett 1997 407 211 214 9166901 10.1016/S0014-5793(97)00346-3
Razandi M Pedram A Merchenthaler I Greene GL Levin ER Plasma membrane estrogen receptors exist and function as dimers. Mol Endocrinol 2004 15231873
Dambaki Constantina Kogia Christina Kampa Marilena Darivianaki Katherine Nomikos Michael Anezinis Ploutarchos Theodoropoulos PanayiotisA Castanas Elias Stathopoulos EfstathiosN Membrane testosterone binding sites in prostate carcinoma as a potential new marker and therapeutic target: Study in paraffin tissue sections BMC Cancer 2005 5 148 16293185 10.1186/1471-2407-5-148
Levin ER Cell localization, physiology, and nongenomic actions of estrogen receptors J Appl Physiol 2001 91 1860 1867 11568173
Cato AC Nestl A Mink S Rapid actions of steroid receptors in cellular signaling pathways Sci STKE 2002 2002 RE9 12084906
Herve JC Non-genomic effects of steroid hormones on membrane channels Mini Rev Med Chem 2002 2 411 417 12370060 10.2174/1389557023405981
Kampa M Papakonstanti EA Alexaki VI Hatzoglou A Stournaras C Castanas E The opioid agonist ethylketocyclazocine reverts the rapid, non-genomic effects of membrane testosterone receptors in the human prostate LNCaP cell line Exp Cell Res 2004 294 434 445 15023532 10.1016/j.yexcr.2003.11.027
Pertschuk LP Macchia RJ Androgen binding in prostatic carcinoma by histochemistry: update 1985 Appl Pathol 1985 3 221 229 3842264
Magi-Galluzzi C Xu X Hlatky L Hahnfeldt P Kaplan I Hsiao P Chang C Loda M Heterogeneity of androgen receptor content in advanced prostate cancer Mod Pathol 1997 10 839 845 9267828
Gao J Isaacs JT Development of an androgen receptor-null model for identifying the initiation site for androgen stimulation of proliferation and suppression of programmed (apoptotic) death of PC-82 human prostate cancer cells Cancer Res 1998 58 3299 3306 9699659
Takeda H Akakura K Masai M Akimoto S Yatani R Shimazaki J Androgen receptor content of prostate carcinoma cells estimated by immunohistochemistry is related to prognosis of patients with stage D2 prostate carcinoma Cancer 1996 77 934 940 8608487 10.1002/(SICI)1097-0142(19960301)77:5<934::AID-CNCR20>3.0.CO;2-3
|
16293185
|
PMC1318463
|
CC BY
|
2021-01-04 16:03:08
|
no
|
BMC Cancer. 2005 Nov 17; 5:148
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-148
|
oa_comm
|
==== Front
J Neuroengineering RehabilJournal of NeuroEngineering and Rehabilitation1743-0003BioMed Central London 1743-0003-2-321631863310.1186/1743-0003-2-32ResearchHypoxia silences the neural activities in the early phase of the phrenic neurogram of eupnea in the piglet Akay Metin [email protected] Neural Engineering & Informatics Laboratory, Harrington Department of Bioengineering, Ira A. Fulton School of Engineering, Arizona State University, Tempe, AZ 85287-9709, USA2005 30 11 2005 2 32 32 17 5 2005 30 11 2005 Copyright © 2005 Akay; licensee BioMed Central Ltd.2005Akay; 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.
Objective
We investigated phrenic neurogram patterns during eupnea (normal breathing) and severe hypoxia (gasping) during early maturation in the piglet.
Methods
We used continuous wavelet transform and short time Fourier transform methods to examine the similarity of breathing patterns in both time and frequency domains during early maturation. The phrenic neurogram was recorded during eupnea, severe hypoxia, and recovery from severe hypoxia in piglets in three different age groups: 3–6 days, 10–15 days and 29–35 days.
Results
During the first week of postnatal age, respiratory patterns of phrenic activity were marked by frequency components between 30 and 300 Hz during both the early (first half) and late (second half) phases of the neurogram signals during eupnea. The results suggest that there is little difference between the respiratory patterns in both time and frequency domains during eupnea compared to gasping for the first week of postnatal age in piglets. After the first week of postnatal age, the duration of the phrenic neurogram burst significantly increases and the patterns during the early phase of the phrenic neurogram are different from those observed for gasping. However, the patterns that mark the late phase of the phrenic neurograms are still the same as those of gasping.
Conclusion
Our most significant finding is that hypoxia silences the neural activity in the early phase of phrenic neurogram regardless of maturation.
==== Body
Introduction
Production of progressive brain hypoxia in an anesthetized, vagotomized, peripherally-chemodenervated cat results in depression of respiratory output and a stereotypical progression of respiratory pattern changes, as hypoxia progresses [1-4]. Initially, the amplitude of the phrenic neurogram is depressed, with a fall in phrenic firing frequency only occurring as the hypoxia becomes more severe. As arterial O2 content falls, progressive respiratory depression continues until the phrenic output is completely silenced. If hypoxia is allowed to progress beyond this point, gasping will eventually ensue [6-8]. This form of respiration is characterized by brief, intense inspiratory efforts of the diaphragm and other respiratory muscles, and has been interpreted as an attempt at "autoresuscitation" [9-12]. This interpretation is based on the observation that animals asphyxiated to the point of apnea by airway occlusion, will restore arterial oxygenation quickly if the occlusion is removed and gasping ensues. If the animal fails to gasp, arterial oxygenation does not improve and death occurs due to cardiovascular collapse inevitably occurs.
The relationship of the medullary gasp to eupneic breathing has been a point of contention for a number of years. Lumsden originally conceived gasping as being the product of a primitive medullary pattern generator which does not contribute to eupneic breathing [6,13]. More recently, St. John and associates have, over the course of several studies, closely examined this question and have concluded that gasping is the result of a unique medullary pattern generator [14-16] in agreement with Lumsden's finding. This conclusion was based on studies of gasping produced by reversibly cooling the pontomedullary junction of decerebrate cats. Although the gasping produced by this procedure has timing characteristics which differ slightly from those seen during hypoxic gasping (e.g., shorter inspiratory time) [15], the qualitative changes seen in the phrenic neurogram and other respiratory outputs during gasping following cooling, were the same as those seen during hypoxic or asphyxic gasping [15,16]. With this model, it was first shown that gasping differs fundamentally from eupnea, both in the pattern of assembly of single phrenic motoneurons to produce a phrenic burst, and in timing characteristics of the phrenic neurogram. The central respiratory controller was also shown to be unresponsive to peripheral chemoreceptor stimulation during gasping. When gasping is produced in the decerebrate cat under conditions of carbon monoxide hypoxia, the discharge frequency of expiratory neurons falls sharply with some units becoming totally silent. The discharge frequency of inspiratory neurons is unchanged during gasping but, unlike during eupnea, all inspiratory neurons fire simultaneously at the beginning of the inspiratory period during gasping [17].
Respiratory control has been studied largely on the basis of phenomenology. There have also been attempts to apply empirical, analytical techniques to the study of central respiratory patterning. Cohen [18] was the first to use autospectral analysis of the phrenic neurogram to gain insight into the central respiratory pattern generation. Subsequently, numerous frequency domain analyses of the phrenic neurogram during eupnea, and during manipulations of various respiratory afferents, have been performed. Virtually all respiratory outputs studied during eupnea (e.g., phrenic and laryngeal neurograms; diaphragmatic electromyograms) have been shown to display two prominent peaks in their spectra: a medium-frequency oscillation (MFO) in the frequency range of 20–50 Hz, and a high-frequency oscillation (HFO) between 50–100 Hz [19-21]. A HFO spectral peak, which is correlated to the phrenic neurogram HFO, has also been noted in medullary inspiratory neuronal activity. Based on these observations, the HFO has been considered to be a characteristic of the central, respiratory pattern generator. The source of the MFO is more problematic [22]. Richardson and Mitchell [23] have proposed that the MFO arises from the interaction of two pattern generators, while Christakos et al. [24], interpret the MFO as a reflection of the rhythmic augmenting discharge of individual phrenic motorneurons resulting from an augmenting drive of supraspinal origin.
Richardson and Mitchell [24] compared the frequency spectra of the phrenic neurogram during eupnea and gasping in decerebrate cats. Hypoxic gasping in decerebrate cats was associated with a high-frequency peak in the phrenic neurogram at 120 Hz, as opposed to the 80 Hz peak seen during eupnea. Spectral analysis of occasional eupneic, phrenic bursts which showed gasp-like augmentation at the end of inspiration, revealed the presence of both eupneic and gasping high-frequency peaks. The presence of a unique spectral peak during gasping was presented as support for the idea that respiratory pattern generation differs during eupnea and gasping.
Preliminary studies of Akay et al. [25] used the modified Yule-Walker autoregression (AR) technique of spectral analysis to analyze 19 eupneic and 13 gasping, phrenic neurograms in anesthetized cats before and during CO-hypoxia and hypoxic-hypoxia, in two preliminary experiments. Our results suggested that eupnea is characterized by three peaks in the AR spectrum, with the lowest peak frequency between 30 and 60 Hz. During gasping a distinctive low-frequency peak was evident in the spectrum below 30 Hz. During eupnea the power spectra of the phrenic neurogram of both cats exhibited two prominent peaks, the first at 40–55 Hz and the second at approximately 100 Hz. The frequencies of these peaks correspond to those described in previous spectral analyses of the phrenic neurogram during eupnea where the lower-frequency peak has been described as medium-frequency oscillation (MFO) and the higher-frequency peak as high-frequency oscillation (HFO) [18,23]. In our results, the transition from eupnea to gasping was characterized by the loss of the MFO, and the appearance of a major peak in the 10–30 Hz range. This shift to a lower frequency during gasping contrasts with the finding of Richardson and Mitchell [22] where gasping resulted in a new spectral peak at a frequency higher than the eupneic HFO. The shift of power to a lower frequency during gasping, observed in our preliminary studies, suggests that there is a synchronization of neuronal firing at a frequency of 20–25 Hz during gasping. The maximal firing frequency of an individual neuron is presumably determined by the kinetics of the ion conductance changes associated with the action potential propagation, which require a finite time for activation and inactivation before a second action potential can be propagated. A frequency of 20–25 Hz is slower than the maximal frequency observed in individual phrenic motoneurons during eupnea (50 Hz), but may represent the maximum firing frequency of a respiratory neuron under the severe hypoxic conditions associated with gasping where channel conductance kinetics may be compromised [25].
When viewed in the time domain, the phrenic neurogram displays a characteristic "ramp" pattern during inspiration and decrementing activity during a short post-inspiratory period [6,12,13,15,26]. This pattern results from an orderly recruitment of phrenic premotor and motor units throughout the period of inspiration. Cohen et al. [27], observed both low- and high-frequency neurogram patterns in piglets at birth, but the high-frequency component was shown to increase with age [27]. They also claimed that high frequency oscillations arise from brain stem respiratory neurons in the medulla and the low-frequency component was not increased with age and was believed to originate from respiratory efferent systems. Later, Webber [28] showed in adult cats that both the early and late phases of the phrenic neurogram have a high frequency component, which is around 82 Hz. Only the late phase has a low frequency component, which is around 29 Hz.
We recently showed that the breathing activities for the young group are not periodic signals, and that the characteristics of phrenic neurograms rapidly change with respect to time [29]. Furthermore our results showed that the phrenic neurogram consists of several dominant burst type activities (circular structured components) corresponding to the early and late phases of the inspiratory activity. However, dominant burst type activities (circular structured components) were only present during the late phase of the phrenic neurogram when maturation proceeds. These results suggest that the phrenic neurogram is not a periodic signal and that its characteristics change rapidly during maturation. The dominant burst type activities disappeared during the early phase of the phrenic neurogram although the burst activity and the continuous activity remained, but both them appear at the late phase of the phrenic neurogram as maturation proceeds [29].
The objective of the study herein presented was to investigate the similarity on the time-frequency respiratory patters during eupnea and severe hypoxia (gasping) and to determine whether hypoxia results in changes in the time-frequency patterns of the respiratory motor output. We have examined the phrenic neurogram in both time and frequency domains during the first few weeks of postnatal life using time-frequency analysis methods to gain insight into the behavior of the respiratory neural network during eupnea and severe hypoxia.
Methodology
Experiment
Experiments were performed in decerebrate piglets of both sexes. Piglets were divided into three age groups: 3–6 days (n = 4), 10–15 days (n = 3) and 29–35 days (n = 3). The animals were anesthetized with 4% isoflurane in O2. The trachea was then cannulated for subsequent delivery of anesthesia (2–3% isoflurane in O2). Cannulation of the femoral artery and vein, peripheral chemodenervation, vagotomy, paralysis, and ventilation were performed. The scalp and underlying muscles were cut and the cerebral hemisphere and the diencephalon were removed. After exposing the mesencephalon, a mid-collicular cut was made and the remaining brain structures rostral to the incision were removed. After completion of the decerebration, anesthesia was removed. Piglets were chemically paralyzed for the rest of the experiment. A minimum of one hour was allowed to elapse between removal of anesthesia and data collection. Piglets were ventilated with 40% O2 in N2 during eupnea. Then, severe hypoxia was produced by inhalation of 3–5% O2 in N2 until gasping was observed in the phrenic neurogram. Phrenic neurogram activity was also recorded during 30 min of reoxygenation (40% O2 in N2).
Data was digitized on line by using a commercial data acquisition and analysis software program (ADI, Powerlab). The phrenic nerve was isolated in the neck at the level of C5 rootlet. The nerve was cut and placed on a bipolar electrode for neuronal recording. The raw phrenic neurogram was bandpass filtered (10 – 300 kHz) and sampled at 1 kHz [29].
Continuous Wavelet Transform (CWT)
The continuous wavelet transform was utilized to analyze the phrenic neurogram signals. This transformation can be viewed as an inner product operation that allows one to measure the similarity or cross-correlation between the signal, s(t), and the wavelet function. The continuous wavelet transform of s(t) is defined as:
where b is a translation (shift) in time and a is the scale factor which represents a translation (shift) in frequency. In the study, we used the Morlet based CWT transform since it shows better time-frequency resolution compared to other orthogonal wavelet transform methods. The details of the Morlet based CWT are described elsewhere [30].
Results
For each piglet, the time-frequency representations during eupnea and severe hypoxia were estimated and compared. Figures 1 and 2 show the raw and the corresponding time-frequency representation of the typical raw phrenic neurograms of a 3-day old piglet during eupnea and severe hypoxia, respectively. Although severe hypoxia (gasping) reduced the time duration of phrenic neurograms during inspiration and increased the expiratory duration, the time-frequency representations during early and late phases of phrenic neurogram during eupnea and gasping showed components between 30 and 300 Hz and demonstrated similarities. In addition, all 4 piglets in the young group exhibited gasping patterns when they were exposed to severe hypoxia.
Figure 1 The raw phrenic neurogram and the corresponding time-frequency representation of the phrenic neurogram of a 3-day old piglet during eupnea (a) and gasping (b).
Figure 2 The raw phrenic neurogram and the corresponding time-frequency representation of the phrenic neurogram of a 10-day old piglet during eupnea (a) and gasping (b).
For the mid-age group, only one of 3 animals had gasping patterns and recovered when animals were reoxygenated. Figures 3 and 4 shows the similar features for a 10 days old piglet. The time frequency patterns were dominant between 30 and 300 Hz at the late phase of the phrenic neurogram during eupnea and about the same as those of gasping.
Figure 3 The raw phrenic neurogram and the corresponding time-frequency representation of the phrenic neurogram of a 30-day old piglet during eupnea (a) and gasping (b).
Figure 4 The mean ratio of the total energies above and below 150 Hz for the early and late phases of the phrenic neurogram during eupnea as well as the phrenic burst during gasping for 3 different age groups.
Figure 4 show the time-frequency patters for a 30-day old piglet. For the 29–35 days old age groups, the time frequency patterns between 30 and 300 Hz are only present for the late phase of the phrenic neurogram during eupnea. The time frequency patterns during gasping and the late phase of phrenic neurogram during eupnea showed considerable similarities. However, the patterns during the early phase of phrenic neurogram was not dominant and the signal components below 150 Hz were different from those marking phrenic neurograms during eupnea.
To investigate the similarity between the patterns in the early and late phases of the phrenic neurogram during eupnea and the patterns during gasping, time-frequency patterns for each piglet over 10 consecutive phrenic bursts during eupnea and 2–3 phrenic bursts during gasping were estimated for each group. Then, we calculated the mean total energies for four time-frequency regions, divided first in time (first and second half of the phrenic neurogram) and then frequency (above and below 150 Hz) during eupnea and the mean total energies below and above 150 Hz during gasping. The mean ratio of the total energies above and below 150 Hz for the early and late phases of the phrenic neurogram during eupnea as well as the phrenic burst during gasping were estimated. The mean ratios for the early, late phase during eupnea and gasping were 0.73 ± 0.1, 0.11 ± 0.1, 0.83 ± 0.19, respectively for the young group. They were 0.52 ± 0.17, 0.67 ± 0.1, 0.73 ± 0.25, for the mid-group and finally they were 0.22 ± 0.09, 0.67 ± 0.12, 0.73 ± 0.18, for the old age groups. Figure 7 summarizes the results. The mean ratios for the early and late phases of the phrenic neurograms during eupnea when compared to those of gasping were not statistically significant for the young age group. As maturation proceeds, the mean ratios for the early phase of phrenic neurograms during eupnea and phrenic bursts during gasping were statistically different although those for the late phases of phrenic bursts during eupnea and phrenic bursts during gasping remained statistically not different. Statistical analysis was performed via an analysis of variance (ANOVA) test.
Discussion and conclusion
Our previous study based on time-frequency analysis methods showed that the time-frequency patterns at the early and late phases of the phrenic neurogram were the same for the 3–6 days old age group. As maturation proceeds, the early phase of the phrenic neurograms demonstrated patterns below 150 Hz that were not dominant, but the patterns for the last phase of phrenic neurograms remained the same and were not influenced by maturation. In this study, we estimated the time-frequency patterns during early and late phases of phrenic neurograms during eupnea and compared them with those of gasping in order to investigate the similarities between these patterns.
Our preliminary data indicated that the patterns during early and late phases of the phrenic neurogram during eupnea are similar to those during gasping for the 3–6 days old group.
The piglets in the young group were very resistive and showed strong responses during gasping in all 4 piglets in this study. However, the mid-group (10–15 days) failed to gasp in 2 of 3 animals. But, all three animals in the old group exhibited the gasping patterns like those in the 3–6 days old group. Therefore, we suggest that the animals in the mid-group could be more vulnerable compared to those in the young and old age groups. In addition, the patterns during early and late phases of phrenic neurogram were almost the same as those of gasping. As maturation proceeds, the similarity between the late phase of phrenic neurogram and gasping remained. Nevertheless, hypoxia significantly reduced the phrenic activities in the early phase of phrenic neurograms and caused a shift in the associated frequency components toward the lower frequency range (i.e., below 150 Hz). Hypoxia significantly increased the expiratory duration and reduced the inspiratory duration (especially, as maturation proceeds). Our most significant finding is that hypoxia silences the neural activity in the early phase of phrenic neurogram regardless of maturation.
Although we do not know the exact mechanism underlying these changes in the patterns of the phrenic neurograms from eupnea to gasping, we speculate that gasping silences phrenic neurons responsible for the neural activities in the early phase of the phrenic neurogram and does not influence phrenic neurons responsible for the neural activities in the late phase of the phrenic neurogram during inspiration. In addition, it also significantly increases the duration of the phrenic neurogram during expiration. We also noted that patterns observed during gasping did not change significantly as maturation proceeds. We speculate that severe hypoxia silences respiratory neurons responsible for both early and late phases of phrenic neurograms in 2 of 3 piglets in the mid-group. We suspect that a reduction in the number of dendrites per cell after 2 weeks of maturation could be responsible for the failure of gasping patterns in these piglets [31].
Acknowledgements
This work was supported by NIH grant (HL 65732). The authors thank K. Johnson and Drs. N. Sekine, J. Bardonova, A. Curran and K. Moodie for their technical support.
==== Refs
Melton JE Neubauer JA Edelman NH CO2 sensitivity of cat phrenic neurogram during hypoxic respiratory depression J Appl Physiol 1988 65 7536 7543
Melton JE Chae LO Neubauer JA Edelman NH Extracellular potassium homeostasis in the cat medulla during progressive brain hypoxia J Appl Physiol 1991 70 1477 1482 2055824
Melton JE Wasicko MJ Neubauer JA Edelman NH Patterns of phrenic depression during progressive brain hypoxia FASEB J 1988 2 A510
Neubauer JA Melton JE Edelman NH Modulation of respiration during brain hypoxia (review) J Appl Physiol 1990 68 441 451 2180894
Neubauer JA Simone A Edelman NH Role of brain lactic acidosis in hypoxic depression of respiration J Appl Physiol 1988 65 1324 1331 3182502
Guntheroth WG Kawabori I Hypoxic apnea and gasping J Clin Invest 1975 56 1371 1377 811688
Lawson EE Thatch BJ Respiratory patterns during progressive asphyxia in newborn rabbits J Appl Physiol 1977 43 468 474 914718
Lumsden T Observations on the respiratory centres J Physiol 1923 57 354 367 16993579
Macefield G Nail B Phrenic and external intercostal motoneuron activity during progressive asphyxia J Appl Physiol 1987 63 1413 1420 3693174
Davis PJ Macefield G Nail BS Respiratory muscle activity during asphyxic apnoea and opisthotonus in the rabbit Respir Physiol 1986 65 285 294 3786968 10.1016/0034-5687(86)90013-7
Macefield G Nail B Inspiratory augmentation during asphyxic hyperpnoea and gasping: proprioceptive influences Respir Physiol 1986 64 57 68 3704381 10.1016/0034-5687(86)90060-5
Breckenridge CG Hoff HE Pontine and medullary regulation of respiration in the cat Am J Physiol 1950 160 385 394 15406334
Lumsden T Observations on the respiratory centres in the cat J Physiol 1923 57 153 160 16993609
St John WM Knuth KV A characterization of the respiratory pattern of gasping J Appl Physiol: Resp, Environ & Exercise Physiol 1981 50 984 993
St John WM Zhou D Fregosi RF Expiratory neural activities in gasping J Appl Physiol 1989 66 223 231 2917925 10.1063/1.343910
Zhou D Wasicko MJ Hu J St John WM Differing activities of medullary respiratory neurons in eupnea and gasping J Appl Physiol 1991 70 1265 1270 2032992 10.1063/1.349582
St John WM Bledsoe TA Sokol HW Identification of medullary loci critical for neurogenesis of gasping J Appl Physiol: Resp, Environ & Exercise Physiol 1984 56 1008 1019
Cohen MI Synchronization of discharge, spontaneous and evoked, between inspiratory neurons Acta Neurobiol Exp 1973 33 189 218
Jacobi MS Thach BT Effect of maturation on spontaneous recovery from hypoxic apnea by gasping J Appl Physiol 1989 66 2384 2390 2745303
Mitchell RA Herbert DA Synchronized high frequency synaptic potentials in medullary respiratory neurons Brain Res 1974 75 350 355 4842842 10.1016/0006-8993(74)90760-4
Cohen MI Feldman JL Discharge properties of dorsal medullary inspiratory neurons: relation to pulmonary afferent and phrenic efferent discharge J Neurophysiol 1984 51 753 776 6716123
Cohen MI See WR Christakos CN Sica AL High-frequency and medium-frequency components of different inspiratory nerve and their modification by various inputs Brain Res 1987 417 148 152 3113671 10.1016/0006-8993(87)90190-9
Christakos CN Cohen MI See WR Barnhardt R Fast rhythms in the discharges of medullary inspiratory neurons Brain Res 1988 463 362 367 3196923 10.1016/0006-8993(88)90411-8
Richardson CA Mitchell RA Power spectral analysis of inspiratory nerve activity in the decerebrate cat Brain Res 1982 233 317 336 6800563 10.1016/0006-8993(82)91205-7
Akay M Melton JE Welkowitz W Edelman NH Neubauer JA Autoregressive spectral analysis of phrenic neurogram during eupnea and gasping J Appl Physiol 1996 81 530 540 8872615
Suthers GK Henderson-Smart DJ Read DJC Postnatal changes in the rate of high frequency bursts of inspiratory activity in cats and dogs Brain Res 1977 132 537 540 912403 10.1016/0006-8993(77)90200-1
Cohen HL Gootman PM Steele AM Eberle LP Rao PP Age-related changes in power spectra of efferent phrenic activity in the piglet Brain Res 1987 426 179 182 3690315 10.1016/0006-8993(87)90439-2
Webber CL High-frequency oscillations within early and late phases of the phrenic neurogram J Appl Physiol 1989 66 886 893 2708218
Akay M Sekine M The Effects of Maturation on Early and Late Phases of Phrenic Neurogram in Piglets during Maturation IEEE Trans on Biomedical Engineering, IEEE Trans on Biomedical Engineering 2004 51 1954 1959 10.1109/TBME.2004.834257
Akay M Time Frequency and Wavelets in Biomedical Signal Processing 1997 New York: Wiley-IEEE Press
Jacobi MS Gershan WM Thach BT Mechanism of failure of recovery from hypoxic apnea by gaping in 17- to 23 day-old mice J Appl Physiol 1991 71 1098 1105 1757305
|
16318633
|
PMC1318464
|
CC BY
|
2021-01-04 16:37:41
|
no
|
J Neuroengineering Rehabil. 2005 Nov 30; 2:32
|
utf-8
|
J Neuroeng Rehabil
| 2,005 |
10.1186/1743-0003-2-32
|
oa_comm
|
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-281631647110.1186/1476-069X-4-28ResearchA cross-sectional study of the association between persistent organochlorine pollutants and diabetes Rylander Lars [email protected] Anna [email protected] Lars [email protected] Division of Occupational and Environmental Medicine and Psychiatric Epidemiology, Department of Laboratory Medicine, University Hospital, SE-221 85 Lund, Sweden2005 29 11 2005 4 28 28 6 9 2005 29 11 2005 Copyright © 2005 Rylander et al; licensee BioMed Central Ltd.2005Rylander 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
Experimental evidence supports the hypothesis that persistent organochlorine pollutants (POPs) may cause type 2 diabetes mellitus, whereas there is no fully convincing epidemiological evidence for such an association. In Sweden the most important source of POP exposure is fatty fish. We have assessed the association between serum levels of POPs and prevalence of diabetes in Swedish fishermen and their wives, with high consumption of fatty fish from the Baltic Sea.
Methods
In 196 men (median age 60 years) and 184 women (median age 64 years), we analyzed 2,2',4,4',5,5'-hexachlorobiphenyl (CB-153) and 1,1-dichloro-2,2-bis(p-chlorophenyl)-ethylene (p,p'-DDE) in serum using gas chromatography-mass spectrometry. The participants were asked if they had diabetes and, if so, since which year and about medication and diet. The Odds Ratios (OR) for diabetes with respect to continuous exposure variables were analyzed with logistic regression, adjusting for potential confounders. Moreover trends of diabetes prevalence with respect to trichotomized exposure variables were tested with Jonckheere-Terpstra's test.
Results
Six percent of the men and 5% of the women had diabetes. After confounder adjustment CB-153 was significantly associated with diabetes prevalence using both categorized and continuous exposure data (an increase of 100 ng/g lipid corresponded to an OR of 1.16, 95% confidence interval [CI] 1.03, 1.32, p = 0.03). Similar associations were observed for p,p'-DDE (an increase of 100 ng/g lipid corresponded to an OR of 1.05, 95% CI 1.01, 1.09, p = 0.006). Gender stratified analyses showed among men consistent positive associations with CB-153, but a more ambiguous pattern with respect to DDE. In contrast, among the women the associations with p,p'-DDE were stronger than with CB-153.
Conclusion
The study provides support that POP exposure might contribute to type 2 diabetes mellitus.
==== Body
Introduction
It is widely believed that the increase in incidence of type 2 diabetes mellitus (T2DM) and obesity is the result of a complex interplay between genetic and environmental factors [1]. T2DM is due to resistance to insulin action and a relative deficiency of insulin. Age, obesity, central adiposity, lack of physical activity, dietary glycemic load, as well as certain genotypic variants are the main factors identified as responsible for the disease [2,3].
Exposure to dioxin has been linked to drastic reductions in glucose uptake in guinea pigs, mice and rats, in vivo as well as in vitro [4], leading to speculation that chronic low-level exposure to dioxins might be a risk factor for T2DM [5].
Previous epidemiological studies have recently been reviewed [6]. Several studies have linked high dioxin burdens to increased risks of T2DM or modified glucose metabolism [7-10]. A Belgian study published later showed highly significant elevation of serum levels of dioxins and polychlorinated biphenyls (PCBs) among patients with type T2DM [11]. Moreover, also chlorinated insecticides have been associated with T2DM. In a group of pesticide users and an unexposed group, Morgan et al [12] found that subjects with T2DM had higher blood levels of dichlorodiphenyl trichloroethane (DDT) and 1,1-dichloro-2,2-bis(p-chlorophenyl)-ethylene (p,p'-DDE), which is the major metabolite of DDT.
A major difficulty in these studies is, however, that exposure occurred sometimes many years before the epidemiologic study, which makes it difficult to determine whether the higher serum levels of persistent organochlorine pollutants (POPs) in diabetics truly reflect a higher exposure to these pollutants, which in turn may contribute to diabetogenesis, or whether they are merely the consequence of T2DM-induced metabolic perturbations facilitating the accumulation of these pollutants. Thus, the possibility of a reversed causality cannot be excluded.
In Sweden, consumption of fatty fish from the Baltic Sea, off the Swedish east coast, is the single major exposure source for POPs, and cohorts of professional fishermen and their families from the Swedish east coast have been found to constitute excellent study bases for epidemiological evaluations of human health effects of POPs [13,14]. We have chosen to use 2,2',4,4',5,5'-hexachlorobiphenyl (CB-153) as a biomarker for POP exposure, because it correlates very well (r ≥ 0.98) with both total PCB concentration in plasma and serum from Swedish subjects [15,16] and with the 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) toxic equivalent (TEQ) in plasma from PCB (r = 0.89) as well as the total POP derived TEQ (r = 0.74) in plasma in American Vietnam veterans [17]. Another relevant exposure biomarker is the anti-androgenic compound p,p'-DDE.
The aim of the present study was to assess the associations between biomarkers for POP and prevalence of diabetes in high exposure cohorts of middle-aged and elderly men and women. The results showed an association between the POP markers in serum and prevalence of diabetes.
Methods
Study population and interview
Previously established cohorts of professional fishermen and their wives from the Swedish east coast [14,18] were linked to the Swedish Population Register. A postal questionnaire was sent in year 2000 to 1500 fishermen and 1291 fishermen's wives that were born between 1920 and 1954, living in Sweden and still alive at 31 December 1999 [28]. There were 813 men (54%) and 779 women (77%) who responded to the questionnaire. Out of them 510 men and 596 women were positive to participate in future clinical studies. We invited a subset of them to a study mainly focused on bone mineral density [11], but also comprising other potential health effects of POP exposure. The aim was to include 200 men and 200 women in the study, and we consecutively contacted subjects by phone for agreements until enough subjects were recruited. Details of the recruitment process have been given elsewhere [19]. The final study groups comprised 196 men and 184 women.
The participants were interviewed, using a standardized questionnaire. The subjects were asked if they had diabetes and, if so, since which year. Moreover, they were asked if they had per oral antidiabetic drugs, insulin or were on diet. We measured current weight and height. In addition, they were asked about their weight at the age of 25 years. Descriptive data for the participants are shown in Table 1.
Table 1 Characteristics for the 196 men and 184 women from Sweden that participated in the study
Men Women
Diabetes Diabetes
No (n = 184) Yes (n = 12) No (n = 174) Yes (n = 10)
Mean, median
(5th, 95th%) Mean, median
(5th, 95th%) Mean, median
(5th, 95th%) Mean, median
(5th, 95th%)
Age (yr)
Current 60, 59
(49, 75) 60, 60
(53, 67) 63, 61
(51, 77) 64, 64
(51, 74)
At diagnosis - 52, 52
(30, 62) - 55, 56
(45, 65)
Time since diagnosis (yr) - 9, 5
(2, 26) - 10, 10
(0, 27)
Body Mass Index (kg/m2)
Current 28.6, 28.1
(23.5, 35.5) 29.4, 30.1
(22.4, 33.3) 27.9, 27.2
(21.3, 35.8) 30.4, 29.5
(26.6, 41.3)
At 25 years of age 23.4, 24.0
(20.0, 28.9) 26.0, 25.4
(21.3, 32.6) 22.0, 22.0
(17.9, 26.6) 23.0, 22.5
(19.7, 26.6)
Exposure (ng/g lipid)
CB-153 430, 360
(110, 950) 670, 560
(360, 1600) 280, 240
(94, 620) 300, 230
(110, 810)
p,p'-DDE 800, 570
(110, 2100) 1100, 1100
(390, 2400) 800, 590
(100, 2300) 1600, 990
(300, 5300)
Out of the group of 813 men that responded to the questionnaire there were 617 subjects that did not participate in the clinical examination. The non-participants had similar age distribution (median 62 years, range 49–84) as the 196 participants (median 60, range 49–84). In addition, the BMI distributions were also very similar among the non-participants (median 26.5 kg/m2, range 17.1–39.9) and the participants (median 27.2, range 20.5–38.5). Out of the group of 779 women that responded to the questionnaire there were 595 subjects that did not participate in the clinical examination. There was neither any difference between the non-participating women and the 184 participants with respect to age (median 65 years, range 49–84 versus median 64 years, range 49–83) or BMI (median 26.5 kg/m2, range 17.1–39.9 versus median 26.2 kg/m2, range 19.7–38.2).
The study was performed in accordance with the Declaration of Helsinki and approved by The Lund University Ethic's Committee. All participants provided written informed consents.
Blood sampling
Venous blood samples were drawn between 8.00 and 10.00 A.M, after 12 hr fasting, into sterile Vacutainer glass tubes (BD Vacutainer, Plymouth, UK). Serum was separated by centrifugation (4000 rpm, 10 minutes) and transferred to glass bottles and special tubes. All serum samples were stored at -80°C until analysis.
Determination of CB-153 and p,p'-DDE in serum
The analyses were performed applying solid phase extraction using on-column degradation of the lipids and analysis by gas chromatography mass spectrometry as previously described [20-22]. Levels of detection, coefficients of variation and participation in quality control programs have been described in detail elsewhere [22].
Determination of serum lipids by enzymatic methods
Serum concentrations of triglycerides and cholesterol were determined by enzymatic methods as described elsewhere [22]. The total lipid concentration in serum (g/L) was calculated by the following equations [23]:
Men: Total = 0.96 + 1.28*(triglycerides + cholesterol)
Women: Total = 1.13 + 1.31*(triglycerides + cholesterol).
Statistics
The effect estimations (odds ratios, OR) between the exposure variables CB-153 and p,p'-DDE, respectively, and diabetes were obtained from logistic regressions. The exposure variables were treated as continuous variables. Due to the high correlation between CB-153 and p,p'-DDE (women r = 0.68; men r = 0.64) these variables were not included in the models simultaneously. As potential confounders we considered gender, current age (as continuous variable) and BMI at 25 years of age (as continuous variable). In addition, the exposure variables were categorized into three equally sized groups. For evaluation whether there were trends in the data with respect to prevalence of diabetes, Jonckheere-Terpstra test (StatXact Statistical Software) was applied. Moreover, separate analyses were performed for men and women. We did also test whether time elapsed since diagnosis of diabetes were correlated (Spearman's correlation test) with the exposure variables and age.
Results
Twelve of the 196 men (6%) and 10 of the 184 women (5%) had diabetes. Five of the male diabetics had per oral antidiabetic drugs, two had a combination of per oral drugs and insulin, one had insulin only, and the remaining four were only on a diet. The corresponding figures for the female diabetics were four, two, one and three, respectively.
For the whole data set CB-153 was significantly associated with diabetes (an increase of 100 ng/g lipid corresponded to an OR of 1.16, 95% confidence interval [CI] 1.03, 1.32, p = 0.03). Among the men the corresponding OR was 1.20 (95% CI 1.04, 1.39, p = 0.01) and among the women 1.06 (95% CI 0.75, 1.50, p = 0.74). Regarding p,p'-DDE, there was for the whole data set a significant association between exposure and diabetes (an increase of 100 ng/g lipid corresponded to an OR of 1.05, 95% CI 1.01, 1.09, p = 0.006). The corresponding ORs were for the men 1.05 (95% CI 0.98, 1.11, p = 0.14) and for the women 1.05 (95% CI 1.01, 1.10, p = 0.02). None of the ORs changed more than marginally (<3%) by including the potential confounders. Moreover, excluding the two subjects with insulin only therapy changed the risk estimates <1%.
Using the exposure data categorized into tertiles there were for the whole data set significant positive trends between CB-153 and p,p'-DDE exposure, respectively, and diabetes (Table 2). Among the men significant positive trends between CB-153 exposure and diabetes (p = 0.005) and between p,p'-DDE exposure and diabetes (p = 0.04) were observed. Among the women the pattern was very similar regarding p,p'-DDE exposure and diabetes (p = 0.07), whereas no such association was observed for the CB-153 exposure.
Table 2 Prevalence of diabetes in relation to tertiles of CB-153 and p,p'-DDE in serum.
Gender Diabetes p for trend a
Exposure (ng/g lipid) Yes/No
Male
CB-153
-290 0/64
>290–475 4/61 0.005
>475 8/58
p,p'-DDE
-410 1/63
>410–850 4/61 0.04
>850 7/60
Female
CB-153
-180 3/57
>180–290 4/57 0.94
>290 3/60
p,p'-DDE
-375 1/59
>375–860 3/59 0.07
>860 6/56
a Jonckheere-Terpstra's test
Time elapsed since diagnosis of diabetes was among the men negatively correlated with CB-153 exposure (rs = -0.59, p = 0.04), whereas it tended to be positively correlated among the women (rs = 0.53, p = 0.12).
Discussion
The main finding of the present study was that diabetics had significantly higher serum levels of both CB-153 and p,p'-DDE than non-diabetic control subjects, using both continuous and categorized exposure data. Gender stratified analyses showed among men consistent positive associations with CB-153, but a more ambiguous pattern with respect to DDE. In contrast, among the women the associations with p,p'-DDE were stronger than with CB-153. We have no biological explanation for this gender difference, which might be a random finding considering the relatively small size of the study.
Our overall results are in concordance with a number of previous epidemiological studies showing associations between T2DM and dioxin exposure [7-10], but also with PCB [11] and DDT/DDE [12] exposures. The epidemiological findings have some biological plausibility as TCDD in experimental studies of guinea pigs, mice and rats decreases cellular glucose uptake [4]. Moreover, it has recently been hypothesized that dioxins and dioxin-like PCBs could promote T2DM by interaction with peroxisome proliferators-activated receptor-γ, a ligand-activated transcription factor controlling lipid metabolism and homeostasis that is linked with T2DM [5,24]. There are no experimental data supporting that di-ortho PCB congeners such as CB-153, will have a diabetogenic effect by themselves, but CB-153 serves as a good proxy marker also for TCDD TEQ and the total POP derived TEQ.
An obvious caveat interpreting the epidemiological cross-sectional studies is to know the direction of the causality between POP exposure and T2DM. A reversed causality cannot be excluded, meaning that the disease affects the serum levels of POP. T2DM is associated with a variety of metabolic changes, which quite conceivably could alter the metabolism of POPs. T2DM can alter the pharmacokinetics of some drugs due to e.g. glycosylation of plasma proteins or displacement by increased plasma levels of free fatty acids, or through deteriorated kidney function [25] and also the activity of cytochrome P450 [26]. T2DM is also known to cause a dysregulation of fat metabolism, which in turn might influence the distribution and elimination of lipophilic compounds such as PCBs and dioxins [11]. If diabetics have a slower rate of excretion of TCDD and other POPs, this could account for the observed associations with T2DM [4]. The possibility of a slower elimination of dioxins in T2DM was, however, not supported by a recent study on Vietnam veterans, in whom no difference in TCDD half-life was found between diabetic and non-diabetic patients [27]. Moreover, if T2DM would slow down the excretion of POPs from the body, time elapsed since diagnosis of T2DM should be expected to be positively correlated with CB-153 in serum. Such a correlation was indicated, however not significant, among the women in the present study. On the other hand among the men there was a significant negative correlation between time since T2DM diagnosis and CB-153 in serum, which speaks against that the T2DM would have slowed down their POP excretion.
In the present study we used self-reported diabetes, and had no access to medical records. However, considering the age distribution, time elapsed since diagnosis and that only one man and one woman had insulin as single therapy, we feel convinced that almost all of the patients had a T2DM. Moreover, the prevalence figures in the present study (6% for men and 5% for men), are well in concordance with was observed in a recent Swedish population based study on similar age groups (about 7% in men and about 5% in women) [28]. Moreover, a sensitivity analysis excluding the two diabetic subjects with insulin only therapy showed that the risk estimates were changed with <1%.
When calculating BMI, we may have slightly underestimated the height at 25 years of age by measuring the current height, but we think that this has only introduced a minor non-differential misclassification.
Conclusion
This cross-sectional study provides support for the hypothesis that POP exposure might contribute to type 2 diabetes mellitus. Even if we cannot exclude the possibility of a reversed causality, the presently observed negative correlation between time period elapsed since diabetes diagnosis and CB-153 level in serum, speaks for the hypothesis of POP as a risk factor.
List of Abbreviations
BMI – Body Mass Index
CB-153 – 2,2',4,4',5,5'-hexachlorobiphenyl
p,p'-DDE – 1,1-dichloro-2,2-bis(p-chlorophenyl)-ethylene
OR – Odds Ratio
POPs – Persistent Organochlorine Pollutants
TEQ – Toxic Equivalent
T2DM – Type 2 diabetes mellitus
TCDD – 2,3,7,8-tetrachlorodibenzo-p-dioxin
Competing interests
The author(s) declare that they have no competeting interests.
Authors' contributions
LH initiated the project. LR and ARH performed the statistical analyses. All authors participated in the design of the study and of writing the manuscript. All authors have read and approved the final manuscript.
Acknowledgements
Financial support was given by the Swedish Research Council for Medicine, the Swedish Research Council for Environment, Agriculture Sciences and Spatial Planning, the Medical Faculty of Lund University, and Region Skåne funds. We thank Lic med sci Ewa Wallin, Ms Hélène Åkesson and Ms Berit Holmskov for their skilful technical assistance.
==== Refs
Schulze MB Hu FB Primary prevention of diabetes: what can be done and how much can be prevented? Annu Rev Public Health 2005 26 445 467 15760297 10.1146/annurev.publhealth.26.021304.144532
O'Rahilly S Science, medicine, and the future. Non-insulin dependent diabetes mellitus: the gathering storm Br Med J 1997 314 955 959 9099123
Uusitupa M Gene-diet interaction in relation to the prevention of obesity and type 2 diabetes: Evidence from the Finnish Diabetes Prevention Study Nutrition Metab Cardiovasc Dis 2005 15 225 233 10.1016/j.numecd.2005.03.004
Enan E Matsumura F 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)-induced changes in glucose transporting activity in guinea pigs, mice and rats in vivo and in vitro J Biochem Toxicol 1994 9 97 106 8071950
Remillards RB Bunce NJ Linking dioxins to diabetes: epidemiology and biologic plausibility Environ Health Perspect 2002 110 853 858 12204817
Longnecker MP Daniels JL Environmental contaminants as etiologic factors for diabetes Environ Health Perspect 2001 109 871 6 11744505
Pazderova-Vejlupkova J Lukas E Nemcova M Pickova J Jirasek L The development and prognosis of chronic intoxication by tetrachlorodibenzo-p-dioxin in men Arch Environ Health 1981 36 5 11 7469493
Henriksen GL Ketchum NS Michalek JE Swaby JA Serum dioxin and diabetes mellitus in veterans of Operation Ranch hand Epidemiology 1997 8 252 258 9115019 10.1097/00001648-199705000-00005
Longnecker MP Michalek JE Serum dioxin level in relation to diabetes mellitus among Air Force Veterans with background levels of exposure Epidemiology 2000 11 44 48 10615842 10.1097/00001648-200001000-00010
Cramner M Louie S Kennedy RH Kern PA Fonseca VA Exposure to 2,3,7,8 tetrachlorodibenzo-p-dioxin (TCDD) is associated with hyperinsulinemia and insulin resistance Toxicol Sci 2000 56 431 436 10911003 10.1093/toxsci/56.2.431
Fierens S Mairesse H Heilier J-F de Burbure C Focant J-F Eppe G de Pauw E Bernard A Dioxin/polychlorinated biphenyl body burden, diabetes and endometriosis: findings in a population-based study in Belgium Biomarkers 2003 8 529 534 15195683 10.1080/1354750032000158420
Morgan DP Lin LI Saikally HH Morbidity and mortality in workers occupationally exposed to pesticides Arch Environ Contam Toxicol 1980 9 349 382 7396557 10.1007/BF01057414
Svensson BG Nilsson A Hansson M Rappe C Åkesson B Skerfving S Exposure to dioxins and dibenzofurans through the consumption of fish N Engl J Med 1991 324 8 12 1898530
Svensson B-G Nilsson A Jonsson E Schütz A Åkesson B Hagmar L Fish consumption and exposure to persistent organochlorine compounds, mercury, selenium and methylamines in Swedish fishermen Scand J Work Environ Health 1995 21 94 103
Grimvall E Rylander L Nilsson-Ehle P Nilsson U Strömberg U Hagmar L Östman C Monitoring of polychlorinated biphenyls in human blood plasma with respect to age, lactation and fish consumption; methodology developments Arch Environ Contam Toxicol 1997 32 329 336 9096084 10.1007/s002449900193
Glynn AW Wolk A Aune M Atuma S Zettermark S Maehle-Schmid M Darnerud PO Becker W Vessby B Adami HO Serum concentrations of organochlorines in men: a search for markers of exposure Sci Total Environ 2000 263 197 208 11194153 10.1016/S0048-9697(00)00703-8
Gladen BC Longnecker MP Schecter AJ Correlations among polychlorinated biphenyls, dioxins, and furans in humans Am J Ind Med 1999 35 15 20 9884741 10.1002/(SICI)1097-0274(199901)35:1<15::AID-AJIM3>3.0.CO;2-Z
Rylander L Hagmar L Mortality and cancer incidence among women with a high consumption of fatty fish contaminated with persistent organochlorine compounds Scand J Work Environ Health 1995 21 419 426 8824747
Wallin E Rylander L Jönssson BAG Lundh T Isaksson A Hagmar L Exposure to CB-153 and p,p'-DDE and bone mineral density in middle-aged and elderly men and women Osteoporosis Int 2005 [epub ahead of print]
Richthoff J Rylander L Jönsson B Mårtensson H Hagmar L Nilsson-Ehle P Stridsberg M Giwercman A Serum levels of 2,2',4,4',5,5'-hexachlorobiphenyl (CB-153) in relation to markers of reproductive function in young males from the general Swedish population Environ Health Perspect 2003 111 409 13 12676591
Rignell-Hydbom A Rylander L Giwercman A Jönsson BAG Nilsson-Ehle P Hagmar L Exposure to CB-153 and p,p'-DDE and male reproductive function Human Reproduction 2004 19 2066 2075 15284211 10.1093/humrep/deh362
Jönsson BAG Rylander L Rignell-Hydbom A Giwercman A Toft G Pedersen HS Ludwicki JK Zvezday V Spanò M Bizzaro D Bonefeld-Jörgensen EC Manicardi G-C Lindh C Bonde JP Hagmar L Determinants of Serum Concentrations of 2,2'4,4'5,5'-hexachlorobiphenyl (CB-153) and 1,1-dichloro-2,2-bis (p-chlorophenyl)-ethylene (p,p'-DDE) in a cross-sectional study of 3161 Men and Women from Greenland, Sweden, Poland and Ukraine Environmental Health: A Global Access Science Source 2003 4 27 10.1186/1476-069X-4-27
Rylander L Nilsson-Ehle P Hagmar L A simplified precise method for adjusting serum levels of persistent organohalogen pollutants to total serum lipids Chemosphere 2005 Jul 7
Smith SA Peroxisome proliferator-activated receptors and the regulation of mammalian lipid metabolism Biochem Soc Transaction 2001 30 1086 1090 10.1042/BST0301086
Gwilt PR Nahhas RR Tracewell WG The effects of diabetes mellitus on pharmacokinetics and pharmacodynamics in humans Clin Pharmacokinet 1991 6 477 90 2044331
Pass GJ Becker W Kluge R Linnartz K Plum L Giessen K Joost HG Effect of hyperinsulinemia and type 2 diabetes-like hyperglycemia on expression of hepatic cytochrome p450 and glutathione S transferase isoforms in a New Zewaland obese-derived mouse backcross population J Pharmacol Exper Therapeutics 2002 302 442 450 10.1124/jpet.102.033553
Michalek J Ketchum N Tripathi R Diabetes mellitus and 2,3,7,8-tetrachlorodibenzo-p-dioxin elimination in veterans of Operation Ranch Hand J Toxicol Environ Health A 2003 66 211 221 12521668 10.1080/15287390306373
Eliasson M Lindahl B Lundberg Stegmayr B No increase in the prevalence of known diabetes between 1986 and 1999 in subjects 25–64 years of age in northern Sweden Diabetic Medicine 2002 19 874 880 12358879 10.1046/j.1464-5491.2002.00789.x
|
16316471
|
PMC1318465
|
CC BY
|
2021-01-04 16:36:33
|
no
|
Environ Health. 2005 Nov 29; 4:28
|
utf-8
|
Environ Health
| 2,005 |
10.1186/1476-069X-4-28
|
oa_comm
|
==== Front
BMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 1471-2288-5-371632116110.1186/1471-2288-5-37Research ArticleStability of response characteristics of a Delphi panel: application of bootstrap data expansion Akins Ralitsa B [email protected] Homer [email protected] Bryan R [email protected] Quality and Patient Safety Initiatives, Rural and Community Health Institute, The Texas A&M University System Health Science Center, College Station, Texas, USA2 Department of Educational Administration and Human Resource Development, The Texas A&M University, College Station, Texas, USA2005 1 12 2005 5 37 37 22 7 2005 1 12 2005 Copyright © 2005 Akins et al; licensee BioMed Central Ltd.2005Akins 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
Delphi surveys with panels of experts in a particular area of interest have been widely utilized in the fields of clinical medicine, nursing practice, medical education and healthcare services. Despite this wide applicability of the Delphi methodology, there is no clear identification of what constitutes a sufficient number of Delphi survey participants to ensure stability of results.
Methods
The study analyzed the response characteristics from the first round of a Delphi survey conducted with 23 experts in healthcare quality and patient safety. The panel members had similar training and subject matter understanding of the Malcolm Baldrige Criteria for Performance Excellence in Healthcare. The raw data from the first round sampling, which usually contains the largest diversity of responses, were augmented via bootstrap sampling to obtain computer-generated results for two larger samples obtained by sampling with replacement. Response characteristics (mean, trimmed mean, standard deviation and 95% confidence intervals) for 54 survey items were compared for the responses of the 23 actual study participants and two computer-generated samples of 1000 and 2000 resampling iterations.
Results
The results from this study indicate that the response characteristics of a small expert panel in a well-defined knowledge area are stable in light of augmented sampling.
Conclusion
Panels of similarly trained experts (who possess a general understanding in the field of interest) provide effective and reliable utilization of a small sample from a limited number of experts in a field of study to develop reliable criteria that inform judgment and support effective decision-making.
==== Body
Background
Since its development in the 1950's, the Delphi method has been broadly utilized in various fields of study including clinical medicine, nursing practice, medical education and healthcare services [1,2]. Despite its wide application, however, many questions regarding this methodology still continue to intrigue researchers. One such question is whether small expert samples are sufficient to conduct a Delphi study; sufficient in this case refers to the stability of panel responses.
In this methodological paper, the adequacy of utilization of a small number of experts in a Delphi panel is discussed. Computer software (SPSS 12.5) was used to augment the data from the first round of a Delphi survey conducted with 23 healthcare quality and patient safety experts and to study the similarities and differences between the response characteristics of the original data and the computer-generated samples.
Sample size in delphi studies
There is no agreement on the panel size for Delphi studies, nor recommendation or unequivocal definition of "small" or "large" samples [1-3]. There is a lack of agreement around the expert sample size and no criteria against which a sample size choice could be judged. Studies have been conducted with virtually any panel size. Reid (1988) studied published articles on Delphi applications in healthcare and noted that there were from 10 to 1685 panellists utilized [4]. Delphi studies with fewer than 10 participants are rarely conducted. For example, a panel of only 5 experts was asked to identify serious drug interactions most likely to occur in the ambulatory pharmacy setting [5], and the responses of an international panel of 6 experts were used to explore competence training for primary care nurses [6]. Many published Delphi studies use panels consisting of 10 to 100 or more panellists, as demonstrated by the following examples. A panel of 10 experts evaluated stage-tailored health promoting interventions [7], and 13 experts were utilized in studying a variety of skills in young children [8]. Two expert panels, consisting of 18 regional and 52 national experts respectively, participated in evaluating an existing pain evaluation system [9]. A multidisciplinary group of 23 participants developed recommendations for the treatment of gastroesophageal reflux disease [10]. Thirty participants were utilized to examine the factors impacting the effectiveness of continuing education in long-term healthcare environment [11] and 32 experts identified the types of scientific misconduct most likely to influence the results of a clinical trial, such as selective reporting and opportunistic use of the play of chance [12]. Two separate studies used a panel of 60 individuals to explore issues of nurse leadership in primary care and priorities in cancer nursing research, respectively [13,14]. Sixty-four medical educators participated in a Delphi panel to develop guidelines for bioterrorism curricula for medical students [15]. One hundred and ten panellists participated in identification of health areas with consumer involvement in research [16], and another panel of 110 pharmacists identified challenges for pharmacy executives [17]. A Delphi panel of 199 nurses explored paediatric oncology nurses' perceptions of parent educational needs [18]. The University of Virginia used 421 respondents in one Delphi study [19]. In another Delphi study, 2,865 participants were invited to participate and 1,142 returned their questionnaires [20].
The sample size in Delphi studies has been researcher and situation specific, and more often than not, convenience samples have been chosen dependent on availability of experts and resources. Given the lack of any Delphi sample size standards, there is confusion regarding how small a "small" sample can be. For example, in one Delphi study, a sample of 37 participants was considered too small for a definite conclusion [21]. In general, the confusion around the Delphi sample arises from the fact that there are no standards established in any methodologically acceptable way. The current literature presents only empirical choices on Delphi expert sample sizes made by individual researchers, such as convenience, purposive or criterion sampling [22].
Malcolm Baldrige award criteria for performance excellence in healthcare
The Malcolm Baldrige National Quality Award (MBNQA) was established in 1987 and is the most prestigious national quality award in the U.S. It is given by the United States Department of Commerce under the authority of the Malcolm Baldrige National Quality Improvement Act of 1987. The MBNQA recognizes superior continuous improvement programs focused on achieving and sustaining performance excellence for the long term.
The MBNQA framework consists of core values and concepts embodied in seven criteria categories: (1) Leadership; (2) Strategic planning; (3) Focus on patients, other customers and markets; (4) Measurement, analysis and knowledge management; (5) Staff focus; (6) Process management; and (7) Organizational performance results [25]. Although the criteria are results-oriented, they are non-prescriptive and adaptable, so that organizational structure and quality approaches may differ widely from one organization to another. The flexibility and adaptability of the Baldrige framework allow improvement changes and objective assessment of an organization's quest for quality.
The MBNQA has five sector categories: Manufacturing, Service, Small Business, Education, and Healthcare. The category of healthcare was added in 1998. Eligible applicants in the category of healthcare include hospitals, health maintenance organizations, long-term healthcare facilities, healthcare practitioner offices, home health agencies, and dialysis and ambulatory surgery centers.
Healthcare experts familiar with the MBNQA approaches are best suited to offer a systems approach. Therefore, their similar training, knowledge and understanding were targeted influences for the outcomes of the Delphi survey. The Malcolm Baldrige criteria provided a structured approach for producing results within a proven framework. Knowledge of the intricate details of the framework ensured an advantage in its application to the field of patient safety.
Importance of the study findings
The results from the analysis of and the comparison between the original responses of the expert panel and the computer generated samples indicated that the number of selected experts utilized in this panel was sufficient to insure reliability for a Delphi study in the field of interest. This finding is important because it establishes the stability of the results from a Delphi survey conducted with a small number of experts from a defined field of study. We can hypothesize that Delphi surveys with a similar number of experts with similar training and knowledge in other fields of study would also yield stable results. Additionally, this finding is important for practitioners in the field of quality training, showing that individuals with similar training, knowledge and understanding of the systems approach based on the Malcolm Baldrige quality criteria could be utilized in a Delphi panel with a constricted number of experts. Given the fact that specialized experts in a given field may be limited, the results of this study suggest that utilization of a small expert sample from limited numbers of experts in a field of study may be used with confidence.
Methods
The Delphi method
The Delphi method facilitates communication between and among a panel of experts, so that the process is effective and the group as a whole can deal with a complex problem [26]. This method improves the generation of critical ideas by structured collection of information and processing of the collective input from a panel of geographically dispersed experts [27]. The methodology originated in the early 1950's, when an Air Force-sponsored Rand project, titled "Project Delphi" sought to reach consensus, through a series of questionnaires and controlled feedback, among military experts on possible U.S. industrial targets for attacks from Russia [26]. The Delphi methodology has applications in many fields, including healthcare, education and sociology.
The advantages of the method are numerous and include:
• The ability to conduct a study in geographically dispersed locations without physically bringing the respondents together;
• Time and cost-effectiveness;
• Discussion of broad and complex problems;
• The ability for a group of experts with no prior history of communication with one another to effectively discuss a problem as a group;
• Participants can have sufficient time to synthesize their ideas;
• Participants can respond at their convenience;
• There is a record of the group activity that can be further reviewed;
• The anonymity of participants provides them with the opportunity to freely express opinions and positions;
• The process has proven to be effective in a variety of fields, problems, and situations [28].
Researchers use the Delphi method to translate scientific knowledge and professional experience into informed judgment, and support effective decision-making [22]. For subject matters in which the best available information is the judgment of knowledgeable individuals, the Delphi method has demonstrated decision-making advantages over traditional conferences, group discussions, brainstorming, and other interactive group activities. The focus in a Delphi study is on the stability of the group opinion rather than on individuals' opinions, thus measuring the group result is superior to measuring individual rankings [27].
Healthcare Delphi survey
A Delphi survey with 23 experts from 18 US states was conducted to create a patient safety tool to guide patient safety improvement in US healthcare organizations. The MBNQA framework was used as a general matrix for the tool and was extended to the field of patient safety. The Delphi study was reviewed and approved by the Institutional Review Board – Human Subjects in Research at Texas A&M University (protocol # 2003–0071).
The MBNQA examiners are trained to have in-depth knowledge and extensive experience relevant to the seven Baldrige categories in at least one, and preferably more than one industry or service sector. Consequently, it was important that the Delphi panel members had expertise in the application of the Baldrige process, as well as in patient safety systems.
Study sample size selection
Given that the intent of the Delphi survey was to examine the patient safety systems in the context of a nationally accepted management framework (the Malcolm Baldrige National Quality Award Criteria for Performance Excellence in Healthcare), all study experts were selected using stringent criteria, including knowledge of and/or training in the Malcolm Baldrige Criteria for Performance Excellence in Healthcare, and knowledge and experience in patient safety. The number of experts with such qualifications was fairly limited (n ~ 100) and the sample of Delphi panel participants was small (n = 23).
The sample size for the study was based initially on an empirically selected small sample size (n = 15) and the expected response rate necessary to achieve this sample size. It was critical to consider what response rate was usually obtained in surveys in the particular study area (healthcare quality and patient safety). A survey on the quality of healthcare and the problem of medical errors administered to a large random sample of Colorado physicians, national physicians and Colorado households, revealed response rates of 66% for the Colorado physician sample, 36% for the national physician sample, and 82% for the Colorado household sample [23]. The psychometric validation process for the Safety Attitude Questionnaire conducted in 160 healthcare sites in the U.S., England and New Zealand obtained a response rate of 67% [24]. Sumsion (1998), as discussed by Hasson, Keeney and McKenna (2000), argued that in order to maintain the rigor of the Delphi technique, a response rate of 70% must be maintained [22]. Based on the healthcare study response rates as found in the literature, it was concluded that for this study a response rate of 70% could be expected. Thus, to obtain at least 15 respondents, the study should begin with 22–23 Delphi panellists, where a sample size of 15 to 23 respondents was considered to be small. Responses were obtained from all 23 experts that had made a commitment to serve on the Delphi panel.
Selection of Delphi experts
Delphi participants are not selected randomly; rather, they are purposefully selected to apply their knowledge and experience to a certain issue based on criteria, which are developed from the nature of the problem under investigation. The following criteria were utilized to qualify experts in healthcare quality improvement and patient safety for inclusion in the original Delphi panel:
(a) Judges, senior examiners or examiners for the Malcolm Baldrige National Quality Award in healthcare;
(b) Senior administrators in healthcare institutions that have won or have applied for the Malcolm Baldrige National Quality Award in healthcare;
(c) Senior administrators in healthcare institutions that have won a state quality award within the last five calendar years;
(d) Leaders in state or national organizations or programs that emphasize continuous quality improvement and/or patient safety;
(e) Experts possessing more than one of the aforementioned criteria.
Based on these criteria, only about 100 healthcare experts nationwide qualified for participation in the Delphi panel. Barriers to identification and inclusion of experts were the confidentiality of MBNQA applicant names and the scarcity of healthcare quality award winners at a state level. Approximately one quarter of the qualified experts were recruited for participation in the panel.
Since the names of the healthcare institutions, which have applied for the Malcolm Baldrige Award are kept confidential, obtaining information regarding the application status of a healthcare institution is a subject of individual contacts and institution's willingness to share such information. The reviewers for the category of healthcare available through the Malcolm Baldrige list of examiners were reached via phone and asked if they would consider sharing information on the applicant status of their institutions. Information was also solicited whether the examiners' organizations had won state quality awards within the last five years, and whether the examiners were senior administrators in their respective institutions. If the examiners and senior healthcare administrators qualified as experts in healthcare quality improvement and patient safety according to the study criteria described above, they were invited to participate in the study. In general, the study participants were recruited via telephone and/or letter contact and were selected from (1) the list of Malcolm Baldrige examiners, (2) senior administrators from healthcare institutions that had won national and/or state quality awards, and (3) referrals from (1) and (2). The recruitment of participants was discontinued after 23 qualified individuals confirmed their willingness to serve on the Delphi panel.
Importance rating scale
The Delphi panel utilized a four-choice Likert scale for assessing the importance of suggested critical processes for patient safety systems in healthcare institutions. The scale was modelled according to the original importance scale developed by Turoff [26]. The participants in the panel were asked to indicate the importance of the Delphi survey items from 1 to 4, where "4" represented processes very important to patient safety systems in healthcare institutions, and "1" represented unimportant (irrelevant). All survey items that were identified by the expert group as "very important" or "important" for patient safety in the third study round, when the experts reached consensus, were included in the final patient safety tool. The Delphi survey concluded in three rounds with creation of a process-centred tool for evaluating patient safety performance and guiding strategic improvement at the institutional level, extending the MBNQA criteria to the area of patient safety [29].
Bootstrap study design
After the Delphi panel created the patient safety tool, the concern about possible group bias with small expert numbers remained, because it has been argued that increased group size is beneficial in Delphi surveys [27]. To study possible differences in response characteristics and to explore the possibility for group bias in the study group of experts and, therefore, to assess the possible error in the creation of the patient safety tool, we generated via computer program (SPSS 12.5) two large samples of expert ratings. Since the variation in expert opinions was greatest in the first study round, encompassing the whole spectrum of possible ratings from 1 to 4, the results from the first survey round were utilized as the basis for computer generation of the expanded samples. The expert responses to the survey items were randomly selected with replacement by the computer program based on the raw data from the first round for the actual survey experts. This resampling technique is called bootstrap.
The bootstrap method was developed by Efron in 1979 and has found wide use in the field of applied statistics [30]. Bootstrap is a Monte Carlo-type data augmentation method utilizing resampling with replacement that can be used with observed data. While Monte Carlo techniques usually generate fictitious data, bootstrap resamples with replacement from the original observed values and generates multiple bootstrap samples as a proxy to the independent real sample. Each bootstrap sample is a random sub-sample (with the same size as the original sample) taken with replacement from the observed values. The original sample is treated as the "virtual population" and the sample is duplicated multiple times. The procedure can be repeated as many times as desired. Resampling has proven valid for any kind of data, including random and non-random data [31]. During the last three decades, the bootstrap resampling has been used widely in applied statistics [32].
Advantages and limitations of the bootstrap technique
Resampling (bootstrapping) of a random sample of an unknown population is considered to model the distribution of that population, where the vaguer the knowledge about the population distribution is, the more valuable the bootstrapping technique proves to be [33]. Since classical statistical techniques are primarily designed for parametric statistics with normal distributions, the bootstrap technique has an advantage in distributions with no convenient statistical formulae, overcoming the limitations of the classical approaches in working with small sample sizes and non-normal distributions [34]. Efron and Tibshirani proposed that the technique reduces the assumptions required to validate analysis and eliminates theoretical calculations required to assess accuracy; its major application is in determination of confidence intervals, where 1,000 or more iterations are necessary to estimate the confidence intervals [30]. The simplicity of the method allows its application in a wide variety of studies and is considered superior to standard statistical tests of significance because it reduces the threat of multiple comparisons bias and provides information on the distribution of scores (and not parametric distributions); the technique is not dependent on a specific nominal size such as 5% and therefore is more accurate [34,35]. The bootstrap technique may have limited accuracy in very small sample sizes (n < 20), in extremely skewed distributions, and if extreme outliers are present [34].
Data manipulation
In this study, statistics for each bootstrap resample were saved in memory and later used for estimation of sampling variance, confidence intervals and assessment of bias for the raw data [30]. The characteristics of the generated samples, when analyzed collectively are used to provide a more representative expression of the underlying population, in this study – the population of patient safety experts knowledgeable about the Malcolm Baldrige framework. The hypothesis was that strict expert inclusion criteria based on training in, and knowledge and understanding of the MBNQA framework in the original sample of 23 experts would provide stability of responses, even if the number of responses was increased by computer generated bootstrap samples.
The bootstrap samples were generated using SPSS 12.5 software. The characteristics of the SPSS 12.5 model were as follows:
NonLinear Regression
MODEL PROGRAM b0 = 2
COMPUTE PRED_ = b0
CNLR VAR00001
/OUTFILE='C:\DOCUME~1\User\LOCALS~1\Temp\spss452\SPSSFNLR.TMP'
/PRED PRED_
/BOOTSTRAP 1000 [2000]
/CRITERIA STEPLIMIT 2 ISTEP 1E+20.
More specifically, the regression routine and the subroutine of nonlinear regression were employed. Once these routines were selected, the options feature of nonlinear regression was invoked. The bootstrap option was selected and the "paste" option was used to indicate the number of bootstrap samples to be derived.
Results
The patient safety tool created by the Delphi panel was based on the seven Malcolm Baldrige categories and included 54 survey items (38 critical processes ranked directly, and one indirectly defined by 16 performance measures). For each survey item, the following inferential statistics were derived:
- Mean (average) – the measure that represents the arithmetic average for the group of experts
- 95% confidence interval – representing the upper and lower limits between which 95% of the sample expert scores will be expected to fall
- 5% trimmed mean – calculation of experts' average score with exclusion of the highest and lowest 5% of the scores; the difference between the mean and the trimmed mean shows whether there are many outliers in the rankings among the experts in the sample (real or computer generated)
- Standard deviation – describes the variability of the score distribution.
These descriptive statistics were calculated for each of the following three samples:
(a) original expert responses of 23 experts in healthcare quality and patient safety (n = 23)
(b) augmented-computer-generated sample of 1000 iterations of resampling and the original sample (n = 1001)
(c) augmented-computer-generated sample of 2000 iterations of resampling and the original sample (n = 2001).
The results for each of the calculated statistics (for each of the three groups of data) are presented in Table 1.
Table 1 Comparison between statistical data for the original and augmented samples
Survey Item Data Entries Mean 95% CI 5% Trim. Mean SD
1 Participants, n = 23 3.64 3.39–3.88 3.70 .5680
Iterations, n = 1001 3.64 3.63–3.65 3.64 .1157
Iterations, n = 2001 3.64 3.63–3.64 3.64 .1172
2 Participants, n = 23 3.64 3.39–3.88 3.70 .5680
Iterations, n = 1001 3.64 3.63–3.65 3.64 .1154
Iterations, n = 2001 3.64 3.63–3.64 3.64 .1151
3 Participants, n = 23 3.46 3.12–3.79 3.55 .7820
Iterations, n = 1001 3.45 3.44–3.46 3.46 .1582
Iterations, n = 2001 3.46 3.45–3.46 3.46 .1589
4 Participants, n = 23 3.26 2.99–3.53 3.29 .6190
Iterations, n = 1001 3.26 3.26–3.27 3.27 .1257
Iterations, n = 2001 3.26 3.25–3.26 3.26 .1276
5 Participants, n = 23 3.13 2.89–3.37 3.14 .5480
Iterations, n = 1001 3.13 3.13–3.14 3.13 .1093
Iterations, n = 2001 3.13 3.13–3.14 3.13 .1133
6 Participants, n = 23 3.22 2.93–3.51 3.24 .6710
Iterations, n = 1001 3.22 3.21–3.23 3.22 .1347
Iterations, n = 2001 3.22 3.21–3.22 3.22 .1371
7 Participants, n = 23 3.11 2.85–3.36 3.12 .5950
Iterations, n = 1001 3.10 3.09–3.11 3.10 .1235
Iterations, n = 2001 3.10 3.10–3.11 3.10 .1226
8 Participants, n = 23 3.14 2.81–3.46 3.20 .7570
Iterations, n = 1001 3.14 3.13–3.14 3.14 .1505
Iterations, n = 2001 3.14 3.13–3.14 3.14 .1569
9 Participants, n = 23 3.18 2.85–3.52 3.25 .7770
Iterations, n = 1001 3.18 3.17–3.19 3.18 .1576
Iterations, n = 2001 3.18 3.18–3.19 3.19 .1580
10 Participants, n = 23 3.17 2.93–3.40 3.18 .5436
Iterations, n = 1001 3.17 2.95–3.38 3.16 .1088
Iterations, n = 2001 3.17 2.95–3.38 3.15 .1098
11 Participants, n = 23 3.30 2.95–3.66 3.39 .8220
Iterations, n = 1001 3.30 3.29–3.31 3.30 .1685
Iterations, n = 2001 3.31 3.30–3.31 3.31 .1652
12 Participants, n = 23 3.09 2.74–3.43 3.14 .7930
Iterations, n = 1001 3.09 3.08–3.10 3.09 .1600
Iterations, n = 2001 3.09 3.08–3.10 3.09 .1583
13 Participants, n = 23 2.87 2.54–3.20 2.90 .7570
Iterations, n = 1001 2.86 2.86–2.87 2.87 .1489
Iterations, n = 2001 2.87 2.86–2.88 2.87 .1528
14 Participants, n = 23 2.53 2.23–2.82 2.59 .6821
Iterations, n = 1001 2.52 2.52–2.53 2.53 .1412
Iterations, n = 2001 2.83 2.82–2.83 2.83 .1578
15 Participants, n = 23 2.83 2.49–3.16 2.86 .7780
Iterations, n = 1001 2.82 2.81–2.83 2.83 .1594
Iterations, n = 2001 2.83 2.82–2.83 2.83 .1578
16 Participants, n = 23 3.23 2.91–3.54 3.25 .7340
Iterations, n = 1001 3.23 3.22–3.24 3.23 .1495
Iterations, n = 2001 3.23 3.22–3.23 3.23 .1490
17 Participants, n = 23 3.35 3.01–3.68 3.39 .7750
Iterations, n = 1001 3.35 3.34–3.36 3.35 .1569
Iterations, n = 2001 3.35 3.34–3.36 3.35 .1557
18 Participants, n = 23 3.39 3.08–3.70 3.43 .7220
Iterations, n = 1001 3.39 3.39–3.40 3.40 .1455
Iterations, n = 2001 3.39 3.38–3.40 3.39 .1483
19 Participants, n = 23 3.33 3.01–3.65 3.42 .7390
Iterations, n = 1001 3.33 3.32–3.33 3.33 .1536
Iterations, n = 2001 3.33 3.33–3.34 3.33 .1528
20 Participants, n = 23 3.57 3.28–3.85 3.63 .6620
Iterations, n = 1001 3.57 3.56–3.58 3.57 .1318
Iterations, n = 2001 3.56 3.56–3.57 3.56 .1355
21 Participants, n = 23 2.96 2.65–3.26 3.00 .7057
Iterations, n = 1001 2.96 2.67–3.24 2.93 .1427
Iterations, n = 2001 2.96 2.68–3.23 2.95 .1384
22 Participants, n = 23 3.39 3.11–3.68 3.43 .6560
Iterations, n = 1001 3.39 3.38–3.40 3.39 .1385
Iterations, n = 2001 3.39 3.38–3.39 3.39 .1323
23 Participants, n = 23 3.30 2.95–3.66 3.39 .8220
Iterations, n = 1001 3.30 3.29–3.31 3.30 .1706
Iterations, n = 2001 3.30 3.30–3.31 3.31 .1714
24 Participants, n = 23 3.39 3.18–3.61 3.38 .4990
Iterations, n = 1001 3.39 3.39–3.40 3.39 .1024
Iterations, n = 2001 3.39 3.39–3.40 3.39 .0986
25 Participants, n = 23 3.48 3.22–3.73 3.52 .5930
Iterations, n = 1001 3.48 3.47–3.49 3.48 .1243
Iterations, n = 2001 3.47 3.47–3.48 3.48 .1195
26 Participants, n = 23 2.83 2.42–3.23 2.86 .9370
Iterations, n = 1001 2.83 2.81–2.84 2.83 .1874
Iterations, n = 2001 2.82 2.82–2.83 2.82 .1915
27 Participants, n = 23 2.89 2.61–3.18 2.93 .6670
Iterations, n = 1001 2.89 2.89–2.90 2.90 .1387
Iterations, n = 2001 2.89 2.89–2.90 2.89 .1369
28 Participants, n = 23 3.26 2.91–3.61 3.34 .8100
Iterations, n = 1001 3.26 3.25–3.27 3.27 .1744
Iterations, n = 2001 3.26 3.25–3.26 3.26 .1675
29 Participants, n = 23 2.83 2.44–3.21 2.86 .8870
Iterations, n = 1001 2.82 2.81–2.84 2.83 .1845
Iterations, n = 2001 2.83 2.82–2.84 2.83 .1814
30 Participants, n = 23 3.43 3.12–3.75 3.48 .7280
Iterations, n = 1001 3.44 3.43–3.45 3.44 .1496
Iterations, n = 2001 3.44 3.43–3.44 3.44 .1473
31 Participants, n = 23 3.09 2.74–3.43 3.14 .7930
Iterations, n = 1001 3.08 3.07–3.09 3.09 .1576
Iterations, n = 2001 3.09 3.08–3.09 3.09 .1639
32 Participants, n = 23 3.06 2.75–3.36 3.11 .7050
Iterations, n = 1001 3.05 3.04–3.06 3.05 .1500
Iterations, n = 2001 3.05 3.05–3.06 3.06 .1428
33 Participants, n = 23 3.17 2.84–3.51 3.24 .7780
Iterations, n = 1001 3.17 3.16–3.18 3.17 .1602
Iterations, n = 2001 3.17 3.17–3.18 3.18 .1579
34 Participants, n = 23 3.26 2.94–3.59 3.34 .7520
Iterations, n = 1001 3.27 3.26–3.28 3.28 .1528
Iterations, n = 2001 3.26 3.26–3.27 3.27 .1512
35 Participants, n = 23 3.39 3.11–3.68 3.43 .6560
Iterations, n = 1001 3.39 3.39–3.40 3.40 .1333
Iterations, n = 2001 3.39 3.38–3.39 3.39 .1314
36 Participants, n = 23 3.06 2.85–3.26 3.06 .4740
Iterations, n = 1001 3.06 3.05–3.07 3.06 .0991
Iterations, n = 2001 3.06 3.05–3.06 3.06 .0972
37 Participants, n = 23 3.32 3.05–3.59 3.35 .6310
Iterations, n = 1001 3.31 3.30–3.32 3.31 .1282
Iterations, n = 2001 3.31 3.31–3.32 3.31 .1294
38 Participants, n = 23 2.89 2.60–3.18 2.93 .6670
Iterations, n = 1001 2.89 2.88–2.90 2.89 .1369
Iterations, n = 2001 2.89 2.89–2.90 2.89 .1346
39 Participants, n = 23 3.22 2.90–3.54 3.24 .7360
Iterations, n = 1001 3.22 3.21–3.23 3.22 .1543
Iterations, n = 2001 3.48 3.47–3.48 3.48 .1022
40 Participants, n = 23 3.48 3.16–3.79 3.53 .7300
Iterations, n = 1001 3.48 3.47–3.49 3.48 .1519
Iterations, n = 2001 3.48 3.47–3.48 3.48 .1465
41 Participants, n = 23 3.39 3.08–3.70 3.43 .7220
Iterations, n = 1001 3.40 3.39–3.41 3.40 .1492
Iterations, n = 2001 3.40 3.39–3.40 3.40 .1450
42 Participants, n = 23 3.65 3.44–3.86 3.67 .4870
Iterations, n = 1001 3.65 3.64–3.66 3.65 .1041
Iterations, n = 2001 3.65 3.65–3.66 3.65 .1003
43 Participants, n = 23 3.61 3.32–3.89 3.68 .6560
Iterations, n = 1001 3.61 3.60–3.62 3.61 .1338
Iterations, n = 2001 3.61 3.60–3.61 3.61 .1362
44 Participants, n = 23 3.61 3.39–3.82 3.62 .4990
Iterations, n = 1001 3.61 3.60–3.62 3.61 .1020
Iterations, n = 2001 3.61 3.60–3.61 3.61 .1036
45 Participants, n = 23 3.30 3.00–3.61 3.34 .7030
Iterations, n = 1001 3.31 3.30–3.32 3.31 .1461
Iterations, n = 2001 3.30 3.30–3.31 3.30 .1424
46 Participants, n = 23 3.43 3.07–3.80 3.53 .8430
Iterations, n = 1001 3.43 3.42–3.45 3.44 .1814
Iterations, n = 2001 3.44 3.43–3.44 3.44 .1735
47 Participants, n = 23 3.39 3.11–3.68 3.43 .6560
Iterations, n = 1001 3.39 3.38–3.40 3.39 .1306
Iterations, n = 2001 3.39 3.38–3.40 3.39 .1368
48 Participants, n = 23 3.48 3.26–3.70 3.48 .5110
Iterations, n = 1001 3.48 3.47–3.48 3.48 .1037
Iterations, n = 2001 3.48 3.47–3.48 3.48 .1022
49 Participants, n = 23 3.61 3.39–3.82 3.62 .4990
Iterations, n = 1001 3.61 3.60–3.62 3.61 .1018
Iterations, n = 2001 3.60 3.60–3.61 3.60 .1003
50 Participants, n = 23 3.26 2.94–3.59 3.29 .7520
Iterations, n = 1001 3.25 3.25–3.26 3.26 .1490
Iterations, n = 2001 3.27 3.26–3.27 3.27 .1527
51 Participants, n = 23 3.30 3.03–3.58 3.34 .6350
Iterations, n = 1001 3.31 3.30–3.31 3.31 .1313
Iterations, n = 2001 3.30 3.30–3.31 3.30 .1316
52 Participants, n = 23 3.39 3.11–3.68 3.43 .6560
Iterations, n = 1001 3.39 3.38–3.40 3.39 .1336
Iterations, n = 2001 3.39 3.39–3.40 3.39 .1325
53 Participants, n = 23 3.18 2.85–3.50 3.25 .7530
Iterations, n = 1001 3.18 3.17–3.19 3.18 .1542
Iterations, n = 2001 3.18 3.17–3.19 3.18 .1543
54 Participants, n = 23 3.18 2.85–3.50 3.25 .7530
Iterations, n = 1001 3.17 3.16–3.18 3.17 .1556
Iterations, n = 2001 3.18 3.18–3.19 3.18 .1550
In general, the means of the expert scores per critical process remained stable across the three data sets (one original and two augmented); the confidence intervals of the three samples were overlapping for each of the critical processes with the confidence intervals in bigger samples being more compact; the trimmed mean exhibited stability, and the standard deviation decreased with increasing number of experts. When the standard deviation decreases, this indicates that the typical deviation of expert opinions has not increased relative to the increased number of responses. Therefore, we can conclude that the original expert sample yielded results concerning the importance of the critical processes in the patient safety tool that were comparable to the results of the expanded samples.
The Delphi survey results showed stability after bootstrap resampling data expansion. Therefore, the stability of the results of the bootstrap data expansion validated the patient safety strategic planning tool developed by the Delphi study on patient safety [29].
Discussion
Application of Delphi studies in healthcare
In the field of healthcare, the Delphi method has been used in planning for the future and formulating policies and programs in biomedical research, behavioral research, mental health, reproductive health, pharmacology, services for the elderly, family planning services, accidents and injuries, development of core competencies for advanced nursing practitioners and development of clinical care protocols [27,36-41]. The Delphi method, as a useful way of identifying and measuring uncertainty, has been widely utilized in medical and health services research to explore issues in health services organizations, to support design of educational programs, to define professionals' roles, to define effects of medical staffing levels, to develop criteria for appropriateness of clinical treatment, and to make long-term projections of need for patient care [42]. The Delphi methodology also has been used to modify the National Board of Medical Examiners (NBME) Medicine Subject Exam (Shelf) in order to align the national exams with the internal medicine clerkship curriculum developed by the Society of General Internal Medicine (SGIM) and the Clerkship Directors in Internal Medicine (CDIM) [43]. Consensus building through Delphi survey technique can contribute significantly to broadening knowledge and effective decision making in health and social care [22]. Furthermore, the Delphi approach can be used as a senior management education tool, environmental planning tool, and for comparison with similar healthcare institutions. Putting together the structure of a model, developmental planning and exploration of policy options are among the explicit application areas identified since the early 1970's [26]. Additionally, the Delphi method has been used to delineate the barriers to performance in health services and identify three types of barriers to optimal healthcare performance: solution development barriers, problem selection barriers, and evaluation barriers. It has been argued that the forecasting accuracy of Delphi studies is strongly reliable; for example, a Delphi study with medical doctors evaluating the forecasting application of the method, revealed that in 75% of the cases the estimated values proved to be less than 10% different from the observed [26].
Challenges in selection of a Delphi panel
The questions arising around the formation of a Delphi panel are typical for selection and formation of any group – committee, task force, panel, or study group. Thus, while panel member selection is a problem that should be addressed, it is by no means unique to Delphi studies. It has been argued that the amount of bias expressed by study participants is offset by the fact that in answering the questions each participant exhibits a standard deviation which is comparable to, or greater than participant's individual mean (i.e., an optimistic panellist is pessimistic in some of his/her responses, and vice versa) [26].
The selection of criteria that would qualify an individual to participate on the Delphi panel depends on the context, scope and aims of the particular study. Some of the general criteria include:
• Knowledge and practical engagement with the issue under investigation
• Capacity and willingness to contribute to the exploration of a particular problem
• Assurance that sufficient time will be dedicated to the Delphi exercise,
• Good written communication skills
• Experts' skills and knowledge need not necessarily be accompanied by standard academic qualifications or degrees [44].
Delphi participants are purposefully selected to apply their knowledge and experience to a certain issue based on a criteria set. For example, the experts for a national two-round Delphi study on the effectiveness and risks of coronary angiography were chosen on the basis of their clinical expertise, community influence, and diversity of geographic location [45]. Since the Delphi method relies on repeated questionnaires to the same initially selected sample of participants, the method requires a continued commitment from the panellists and is heavily dependant on the time and continued involvement on the part of the study participants. The widespread employment of electronic communications calls for consideration of the computer literacy and skills of the target sample before utilizing electronic means of communication [22]. Since the sample size for Delphi panels has not been established, it is important to know whether the selected Delphi panel for a particular study would yield stable results.
Utilization of bootstrap in healthcare research
During the last decade, the use of bootstrap data expansion in healthcare research became more prominent. Some examples of utilization of bootstrap in contemporary healthcare research include: constructing confidence intervals for treatment differences, analysing cost-effectiveness in randomized controlled trials, assessing provider performance for providers with small numbers of observed events, and studying the genetic linkages in viral genome sequences [46-51]. There is a growing validation of the value of bootstrap in medical statistics and an increasing recognition that the bootstrap technique can supplement and extend the conventional statistical thinking. Bootstrap can be used for calculation of confidence intervals, hypothesis testing, linear regression and correlation in variable prediction, and non-linear regression in immunoassay techniques [33]. The bootstrap method, with its computational simplicity and performance similar to the fully Bayesian approaches, was found to be a very useful addition to healthcare researchers' statistical toolkit [52].
Conclusion
Although experiments carried out in the 1950's and 1960's suggested that group error is reduced with increased group size, the sample size for constructing a Delphi panel is not a statistically-bound decision and good results can be obtained by a comparatively small group of homogenous experts [44]. However, the size of a "small" Delphi sample has not been unequivocally established.
The findings of this study are important because:
1. It was found that reliable outcomes could be obtained with a Delphi panel consisting of a relatively small number of Delphi experts (23) selected via strict inclusion criteria. This finding is particularly important for conducting Delphi surveys in knowledge or practice fields where the population of experts (the total number of qualifying knowledgeable individuals) is limited. Experts who have similar training and general understanding in the field of interest allow for effective and reliable utilization of a small sample from a limited number of experts in the field of study.
2. The stability of the results from a Delphi survey conducted with a small number of experts from a defined field of study was established. It is hypothesized that Delphi surveys in other fields of study, conducted with a small number of experts with similar training and knowledge, would also yield reliable results. Given the fact that the number of specialized experts in a particular field may be limited, this study validated the stability of response characteristics of a small expert sample from limited numbers of experts. Therefore, consistency of expert training may allow utilization of small numbers of experts in fields where many experts may be available but participation of a limited number of experts on the Delphi panel may be more practical.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
RBA, HT and BRC developed the project idea and the study design. RBA collected the data. RBA and HT conducted the study. HT provided statistical advice and BRC provided conceptual advice. All authors contributed to the analysis of data. The manuscript was initially drafted by RBA. HT and BRC critically reviewed and edited the manuscript. All authors contributed to and approved the final version submitted for publication.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors would like to thank the participants in the Delphi study for their commitment and dedication. The authors would also like to express their gratitude to Ms. Xue Hu (Texas A&M University) for her skills and precision in handling survey data in support of this paper. Additionally, the authors would like to recognize one individual, who chose to remain anonymous, for her support in running the SPSS bootstrap model.
==== Refs
Williams PL Webb C The Delphi technique: a methodological discussion J Adv Nurs 1994 19 180 6 8138622
Fischer RG The Delphi method: a description, review and criticism J Acad Librariansh 1978 4 64 70
Wilhelm WJ Alchemy of the Oracle: the Delphi technique The Delta Pi Epsilon Journal 2001 43 6 26
Reid NG Ellis R The Delphi technique: its contribution to the evaluation of professional practice Professional competence and quality assurance in the caring professions 1988 New York, NY: Chapman and Hall
Malone DC Abarca J Hansten PD Grizzle AJ Armstrong EP Van Bergen RC Duncan-Edgar BS Solomon SL Lipton RB Identification of serious drug-drug interactions: results of the partnership to prevent drug-drug interactions Am J Geriatr Pharmacother 2005 3 65 76 10.1016/j.amjopharm.2005.05.001
Strasser S London L Kortenbout E Developing a competence framework and evaluation tool for primary care nurses in South Africa Educ Health 2005 18 133 144
De Vet E Brug J De Nooijer J Dijkstra A De Vries NK Determinants of forward stage transitions: a Delphi study Health Educ Res 2005 20 195 205 15328302 10.1093/her/cyg111
Howell S Kemp C Defining early number sense: a participatory Australian study J Educ Psychol 2005 25 555 571 10.1080/01443410500046838
Nekolaichuk CL Fainsinger RL Lawlor PG A validation study of a pain classification system for advanced cancer patients using content experts: the Edmonton Classification System for cancer pain Palliat Med 2005 19 466 476 16218159 10.1191/0269216305pm1055oa
Armstrong D Marshall JK Chiba N Enns R Fallone CA Fass R Hollingworth R Hunt RH Kahrilas PJ Mayrand S Moayyedi P Paterson WG Sadowski D van Zanten SJ Canadian consensus conference on the management of gastroesophageal reflux disease in adults – update 2004 Can J Gastroenterol 2005 19 15 35 15685294
Stolee P Esbaugh J Aylward S Cathers T Harvey DP Hillier LM Keat N Feightner JW Factors associated with the effectiveness of continuing education in long-term care Gerontologist 2005 45 399 405 15933280
Al-Marzouki S Roberts I Marshall T Evans S The effect of scientific misconduct on the results of clinical trials: a Delphi survey Contemp Clin Trials 2005 26 331 337 15911467 10.1016/j.cct.2005.01.011
McKenna H Keeney S Bradley M Nurse leadership within primary care: The perceptions of community nurses, GPs, policy makers and members of the public J Nurs Manag 2004 12 69 76 15101458
Mcilfatrick SJ Keeney S Identifying cancer nursing research priorities using the Delphi technique J Adv Nurs 2003 42 629 636 12787236 10.1046/j.1365-2648.2003.02666.x
Coico R Kachur E Lima V Lipper S Guidelines for preklership bioterrorism curricula Acad Med 2004 79 366 375 15044173 10.1097/00001888-200404000-00019
Telford R Boote JD Cooper CL What does it mean to involve consumers successfully in NHS research? A consensus study Health Expect 2004 7 209 220 15327460 10.1111/j.1369-7625.2004.00278.x
Meadows AB Maine LL Keyes EK Pearson K Finstuen K Pharmacy executive leadership issues and associated skills, knowledge and abilities J Am Pharm Assoc 2005 45 55 62 10.1331/1544345052843183
Kelly KP Porock D A survey of paediatric oncology nurses' perceptions of parent educational needs J Pediatr Oncol Nurs 2005 22 58 66 15574727 10.1177/1043454204272537
Cypert FR Gant WL The Delphi technique: A case study Phi Delta Kappan 1971 272 273
Barnette JJ Danielson LC Algozzine RF Delphi methodology: an empirical investigation Educational Research Quarterly 1978 3 67 73
Synnott G McKie D International issues in PR: Researching research and prioritising priorities Journal of Public Relations Research 1997 9 259 282 10.1207/s1532754xjprr0904_02
Hasson F Keeney S McKenna H Research guidelines for the Delphi survey technique J Adv Nurs 2000 32 1008 1015 11095242 10.1046/j.1365-2648.2000.01567.x
Robinson AR Hohmann KB Rifkin JI Topp D Gilroy CM Pickard JR Anderson RJ Physician and public opinions on quality of health care and the problem of medical errors Arch Intern Med 2002 162 2186 2190 12390060 10.1001/archinte.162.19.2186
Sexton JB Helmreih RL Rowan K Vella K Boyden J Neilands TB Roberts PR Thomas EJ The safety attitude questionnaire: a psychometric validation Technical report 2003 The University of Texas Center of Excellence for Patient Safety Research and Practice, Austin, Texas AHRQ Grants 1PO1HS1154401 and U18HS1116401
Baldrige National Quality Program 2005 Healthcare criteria for performance excellence 2005 Gaithersburg, MD: United States Department of Commerce, Technology Administration, National Institute of Standards and Technology
Linstone HA Turoff M The Delphi Method: techniques and applications 1975 Reading, MA: Addison-Wesley Publishing Company
Adler M Ziglio E Gazing into the Oracle: the Delphi method and its application to social policy and public health 1996 Bristol, PA: Jessica Kingsley Publishers, Ltd
Rotondi A Gustafson D Adler M, Ziglio E Theoretical, methodological and practical issues arising out of the Delphi method Gazing into the Oracle: the Delphi method and its application to social policy and public health 1996 Bristol, PA: Jessica Kingsley Publishers, Ltd 34 55
Akins RB A Process-centered tool for evaluating patient safety performance and guiding strategic improvement Advances in patient safety: from research to implementation 2005 4 Rockville, MD: Agency for Healthcare Research and Quality 109 125
Efron B Tibshirani R An introduction to the bootstrap Monographs on statistics and applied probability, No 57 1993 London: Chapman and Hall
Edgington ES Randomization tests 1995 New York: M. Dekker
Good P Permutation tests: a practical guide for resampling methods for testing hypotheses 2000 New York: Springer-Verlag
Henderson AR The bootstrap: a technique for data-driven statistics. Using computer intensive analyses to explore experimental data Clin Chim Acta 2005 359 1 26 15936746 10.1016/j.cccn.2005.04.002
Linden A Adams JL Roberts N Evaluating disease management program effectiveness: an introduction to the bootstrap technique Disease Management & Health Outcomes 2005 13 159 167
Lloyd CJ On comparing the accuracy of competing tests of the same hypotheses of simulation data J Stat Plan Inference 2005 128 497 508 10.1016/j.jspi.2003.12.002
Calzone KA Jenkins J Masny A Core competencies in cancer genetics for advanced practice oncology nurses Oncol Nurs Forum 2002 29 1327 1333 12370702
Wang CC Wang Y Zhang K Fang J Liu W Luo S Tang S Wang S Li VC Reproductive health indicators for China rural areas Soc Sci Med 2003 57 217 225 12765703 10.1016/S0277-9536(02)00341-6
Catlin AJ Carter BS Creation of a neonatal end-of-life palliative care protocol J Perinat Neonal Nurs 2002 21 37 49
Sharp D Liebenau A Stocks N Bennewith O Evans M Jones WB Peters T Goldberg D Gunnell D Locally developed guidelines for the aftercare of deliberate self-harm patients in general practice Primary Health Care Research and Development 2003 4 21 28 10.1191/1463423603pc119oa
Escobar A Qintana JM Arostegui I Azkarate J Guenaga JI Arenaza JC Garai I Development of explicit criteria for total knee replacement Int J Technol Assess Health Care 2003 19 57 70 12701939 10.1017/S0266462303000060
McBride AJ Pates R Ramadan R McGowan C Delphi survey of experts' opinions on strategies used by community pharmacists to reduce over-the-counter drug misuse Addiction 2003 98 487 497 12653818 10.1046/j.1360-0443.2003.00345.x
Jones J Hunter D Consensus methods for medical and health services research Br Med J 1995 311 376 380 7640549
Elnicki DM Lescisin DA Case S Improving the National Board of Medical Examiners Internal Medicine Subject Exam for use in clerkship evaluation J Gen Intern Med 2002 17 435 440 12133157 10.1046/j.1525-1497.2002.10673.x
Ziglio E Adler M, Ziglio E The Delphi method and its contribution to decision making Gazing into the Oracle: the Delphi method and its application to social policy and public health 1996 Bristol, PA: Jessica Kingsley Publishers, Ltd 3 33
Bernstein SJ Laouri M Hilborne LH Leape LL Kahan JP Park RE Kamberg CJ Brook RH Coronary Angiography: a literature review and ratings of appropriateness and necessity 1992 The New York City Rand Institute, JRA-03. Santa Monica, CA: The Rand Corporation
Maetzel A Strand V Tugwell P Wells G Bombardier C Economic comparison of Leflunomide and Methotrexate in patients with rheumatoid arthritis Pharmacoeconomics 2002 20 61 70 11817993
Manktelow B Abrams K Draper E A comparison of six methods of estimating confidence intervals for an indirectly standardized outcome ratio J Epidemiol Community Health A36
Peris F Martinez E Badia X Brosa M Iatrogenic cost factors incorporating mild and moderate adverse events in the economic comparison of Aceclofenac and other NSAIDs Pharmacoeconomics 2001 19 779 790 11548913
Neymark N Adriaenssen I Gorlia T Caleo S Bolla M Brochon D Cost-effectiveness of the addition of early hormonal therapy in locally advanced prostate cancer: results decisively determined by the cut-off time point chosen for the analysis Eur J Cancer 2001 37 1768 1774 11549430 10.1016/S0959-8049(01)00197-6
Chen M Kianifard F A nonparametric procedure associated with a clinically meaningful efficacy measure Biostatistics 2000 1 293 298 12933510 10.1093/biostatistics/1.3.293
Bredell H Crookes RL Heynes du PA Schoub BD Morris L Molecular investigation of two possible cases of accidental HIV-1 transmission in South Africa AIDS Res Hum Retroviruses 2003 19 613 617 12921093 10.1089/088922203322230987
MacNab YC Farrel PJ Gustafson P Wen S Estimation in Bayesian disease mapping Biometrics 2004 60 865 873 15606406 10.1111/j.0006-341X.2004.00241.x
|
16321161
|
PMC1318466
|
CC BY
|
2021-01-04 16:32:51
|
no
|
BMC Med Res Methodol. 2005 Dec 1; 5:37
|
utf-8
|
BMC Med Res Methodol
| 2,005 |
10.1186/1471-2288-5-37
|
oa_comm
|
==== Front
BMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 1471-2288-5-371632116110.1186/1471-2288-5-37Research ArticleStability of response characteristics of a Delphi panel: application of bootstrap data expansion Akins Ralitsa B [email protected] Homer [email protected] Bryan R [email protected] Quality and Patient Safety Initiatives, Rural and Community Health Institute, The Texas A&M University System Health Science Center, College Station, Texas, USA2 Department of Educational Administration and Human Resource Development, The Texas A&M University, College Station, Texas, USA2005 1 12 2005 5 37 37 22 7 2005 1 12 2005 Copyright © 2005 Akins et al; licensee BioMed Central Ltd.2005Akins 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
Delphi surveys with panels of experts in a particular area of interest have been widely utilized in the fields of clinical medicine, nursing practice, medical education and healthcare services. Despite this wide applicability of the Delphi methodology, there is no clear identification of what constitutes a sufficient number of Delphi survey participants to ensure stability of results.
Methods
The study analyzed the response characteristics from the first round of a Delphi survey conducted with 23 experts in healthcare quality and patient safety. The panel members had similar training and subject matter understanding of the Malcolm Baldrige Criteria for Performance Excellence in Healthcare. The raw data from the first round sampling, which usually contains the largest diversity of responses, were augmented via bootstrap sampling to obtain computer-generated results for two larger samples obtained by sampling with replacement. Response characteristics (mean, trimmed mean, standard deviation and 95% confidence intervals) for 54 survey items were compared for the responses of the 23 actual study participants and two computer-generated samples of 1000 and 2000 resampling iterations.
Results
The results from this study indicate that the response characteristics of a small expert panel in a well-defined knowledge area are stable in light of augmented sampling.
Conclusion
Panels of similarly trained experts (who possess a general understanding in the field of interest) provide effective and reliable utilization of a small sample from a limited number of experts in a field of study to develop reliable criteria that inform judgment and support effective decision-making.
==== Body
Background
Since its development in the 1950's, the Delphi method has been broadly utilized in various fields of study including clinical medicine, nursing practice, medical education and healthcare services [1,2]. Despite its wide application, however, many questions regarding this methodology still continue to intrigue researchers. One such question is whether small expert samples are sufficient to conduct a Delphi study; sufficient in this case refers to the stability of panel responses.
In this methodological paper, the adequacy of utilization of a small number of experts in a Delphi panel is discussed. Computer software (SPSS 12.5) was used to augment the data from the first round of a Delphi survey conducted with 23 healthcare quality and patient safety experts and to study the similarities and differences between the response characteristics of the original data and the computer-generated samples.
Sample size in delphi studies
There is no agreement on the panel size for Delphi studies, nor recommendation or unequivocal definition of "small" or "large" samples [1-3]. There is a lack of agreement around the expert sample size and no criteria against which a sample size choice could be judged. Studies have been conducted with virtually any panel size. Reid (1988) studied published articles on Delphi applications in healthcare and noted that there were from 10 to 1685 panellists utilized [4]. Delphi studies with fewer than 10 participants are rarely conducted. For example, a panel of only 5 experts was asked to identify serious drug interactions most likely to occur in the ambulatory pharmacy setting [5], and the responses of an international panel of 6 experts were used to explore competence training for primary care nurses [6]. Many published Delphi studies use panels consisting of 10 to 100 or more panellists, as demonstrated by the following examples. A panel of 10 experts evaluated stage-tailored health promoting interventions [7], and 13 experts were utilized in studying a variety of skills in young children [8]. Two expert panels, consisting of 18 regional and 52 national experts respectively, participated in evaluating an existing pain evaluation system [9]. A multidisciplinary group of 23 participants developed recommendations for the treatment of gastroesophageal reflux disease [10]. Thirty participants were utilized to examine the factors impacting the effectiveness of continuing education in long-term healthcare environment [11] and 32 experts identified the types of scientific misconduct most likely to influence the results of a clinical trial, such as selective reporting and opportunistic use of the play of chance [12]. Two separate studies used a panel of 60 individuals to explore issues of nurse leadership in primary care and priorities in cancer nursing research, respectively [13,14]. Sixty-four medical educators participated in a Delphi panel to develop guidelines for bioterrorism curricula for medical students [15]. One hundred and ten panellists participated in identification of health areas with consumer involvement in research [16], and another panel of 110 pharmacists identified challenges for pharmacy executives [17]. A Delphi panel of 199 nurses explored paediatric oncology nurses' perceptions of parent educational needs [18]. The University of Virginia used 421 respondents in one Delphi study [19]. In another Delphi study, 2,865 participants were invited to participate and 1,142 returned their questionnaires [20].
The sample size in Delphi studies has been researcher and situation specific, and more often than not, convenience samples have been chosen dependent on availability of experts and resources. Given the lack of any Delphi sample size standards, there is confusion regarding how small a "small" sample can be. For example, in one Delphi study, a sample of 37 participants was considered too small for a definite conclusion [21]. In general, the confusion around the Delphi sample arises from the fact that there are no standards established in any methodologically acceptable way. The current literature presents only empirical choices on Delphi expert sample sizes made by individual researchers, such as convenience, purposive or criterion sampling [22].
Malcolm Baldrige award criteria for performance excellence in healthcare
The Malcolm Baldrige National Quality Award (MBNQA) was established in 1987 and is the most prestigious national quality award in the U.S. It is given by the United States Department of Commerce under the authority of the Malcolm Baldrige National Quality Improvement Act of 1987. The MBNQA recognizes superior continuous improvement programs focused on achieving and sustaining performance excellence for the long term.
The MBNQA framework consists of core values and concepts embodied in seven criteria categories: (1) Leadership; (2) Strategic planning; (3) Focus on patients, other customers and markets; (4) Measurement, analysis and knowledge management; (5) Staff focus; (6) Process management; and (7) Organizational performance results [25]. Although the criteria are results-oriented, they are non-prescriptive and adaptable, so that organizational structure and quality approaches may differ widely from one organization to another. The flexibility and adaptability of the Baldrige framework allow improvement changes and objective assessment of an organization's quest for quality.
The MBNQA has five sector categories: Manufacturing, Service, Small Business, Education, and Healthcare. The category of healthcare was added in 1998. Eligible applicants in the category of healthcare include hospitals, health maintenance organizations, long-term healthcare facilities, healthcare practitioner offices, home health agencies, and dialysis and ambulatory surgery centers.
Healthcare experts familiar with the MBNQA approaches are best suited to offer a systems approach. Therefore, their similar training, knowledge and understanding were targeted influences for the outcomes of the Delphi survey. The Malcolm Baldrige criteria provided a structured approach for producing results within a proven framework. Knowledge of the intricate details of the framework ensured an advantage in its application to the field of patient safety.
Importance of the study findings
The results from the analysis of and the comparison between the original responses of the expert panel and the computer generated samples indicated that the number of selected experts utilized in this panel was sufficient to insure reliability for a Delphi study in the field of interest. This finding is important because it establishes the stability of the results from a Delphi survey conducted with a small number of experts from a defined field of study. We can hypothesize that Delphi surveys with a similar number of experts with similar training and knowledge in other fields of study would also yield stable results. Additionally, this finding is important for practitioners in the field of quality training, showing that individuals with similar training, knowledge and understanding of the systems approach based on the Malcolm Baldrige quality criteria could be utilized in a Delphi panel with a constricted number of experts. Given the fact that specialized experts in a given field may be limited, the results of this study suggest that utilization of a small expert sample from limited numbers of experts in a field of study may be used with confidence.
Methods
The Delphi method
The Delphi method facilitates communication between and among a panel of experts, so that the process is effective and the group as a whole can deal with a complex problem [26]. This method improves the generation of critical ideas by structured collection of information and processing of the collective input from a panel of geographically dispersed experts [27]. The methodology originated in the early 1950's, when an Air Force-sponsored Rand project, titled "Project Delphi" sought to reach consensus, through a series of questionnaires and controlled feedback, among military experts on possible U.S. industrial targets for attacks from Russia [26]. The Delphi methodology has applications in many fields, including healthcare, education and sociology.
The advantages of the method are numerous and include:
• The ability to conduct a study in geographically dispersed locations without physically bringing the respondents together;
• Time and cost-effectiveness;
• Discussion of broad and complex problems;
• The ability for a group of experts with no prior history of communication with one another to effectively discuss a problem as a group;
• Participants can have sufficient time to synthesize their ideas;
• Participants can respond at their convenience;
• There is a record of the group activity that can be further reviewed;
• The anonymity of participants provides them with the opportunity to freely express opinions and positions;
• The process has proven to be effective in a variety of fields, problems, and situations [28].
Researchers use the Delphi method to translate scientific knowledge and professional experience into informed judgment, and support effective decision-making [22]. For subject matters in which the best available information is the judgment of knowledgeable individuals, the Delphi method has demonstrated decision-making advantages over traditional conferences, group discussions, brainstorming, and other interactive group activities. The focus in a Delphi study is on the stability of the group opinion rather than on individuals' opinions, thus measuring the group result is superior to measuring individual rankings [27].
Healthcare Delphi survey
A Delphi survey with 23 experts from 18 US states was conducted to create a patient safety tool to guide patient safety improvement in US healthcare organizations. The MBNQA framework was used as a general matrix for the tool and was extended to the field of patient safety. The Delphi study was reviewed and approved by the Institutional Review Board – Human Subjects in Research at Texas A&M University (protocol # 2003–0071).
The MBNQA examiners are trained to have in-depth knowledge and extensive experience relevant to the seven Baldrige categories in at least one, and preferably more than one industry or service sector. Consequently, it was important that the Delphi panel members had expertise in the application of the Baldrige process, as well as in patient safety systems.
Study sample size selection
Given that the intent of the Delphi survey was to examine the patient safety systems in the context of a nationally accepted management framework (the Malcolm Baldrige National Quality Award Criteria for Performance Excellence in Healthcare), all study experts were selected using stringent criteria, including knowledge of and/or training in the Malcolm Baldrige Criteria for Performance Excellence in Healthcare, and knowledge and experience in patient safety. The number of experts with such qualifications was fairly limited (n ~ 100) and the sample of Delphi panel participants was small (n = 23).
The sample size for the study was based initially on an empirically selected small sample size (n = 15) and the expected response rate necessary to achieve this sample size. It was critical to consider what response rate was usually obtained in surveys in the particular study area (healthcare quality and patient safety). A survey on the quality of healthcare and the problem of medical errors administered to a large random sample of Colorado physicians, national physicians and Colorado households, revealed response rates of 66% for the Colorado physician sample, 36% for the national physician sample, and 82% for the Colorado household sample [23]. The psychometric validation process for the Safety Attitude Questionnaire conducted in 160 healthcare sites in the U.S., England and New Zealand obtained a response rate of 67% [24]. Sumsion (1998), as discussed by Hasson, Keeney and McKenna (2000), argued that in order to maintain the rigor of the Delphi technique, a response rate of 70% must be maintained [22]. Based on the healthcare study response rates as found in the literature, it was concluded that for this study a response rate of 70% could be expected. Thus, to obtain at least 15 respondents, the study should begin with 22–23 Delphi panellists, where a sample size of 15 to 23 respondents was considered to be small. Responses were obtained from all 23 experts that had made a commitment to serve on the Delphi panel.
Selection of Delphi experts
Delphi participants are not selected randomly; rather, they are purposefully selected to apply their knowledge and experience to a certain issue based on criteria, which are developed from the nature of the problem under investigation. The following criteria were utilized to qualify experts in healthcare quality improvement and patient safety for inclusion in the original Delphi panel:
(a) Judges, senior examiners or examiners for the Malcolm Baldrige National Quality Award in healthcare;
(b) Senior administrators in healthcare institutions that have won or have applied for the Malcolm Baldrige National Quality Award in healthcare;
(c) Senior administrators in healthcare institutions that have won a state quality award within the last five calendar years;
(d) Leaders in state or national organizations or programs that emphasize continuous quality improvement and/or patient safety;
(e) Experts possessing more than one of the aforementioned criteria.
Based on these criteria, only about 100 healthcare experts nationwide qualified for participation in the Delphi panel. Barriers to identification and inclusion of experts were the confidentiality of MBNQA applicant names and the scarcity of healthcare quality award winners at a state level. Approximately one quarter of the qualified experts were recruited for participation in the panel.
Since the names of the healthcare institutions, which have applied for the Malcolm Baldrige Award are kept confidential, obtaining information regarding the application status of a healthcare institution is a subject of individual contacts and institution's willingness to share such information. The reviewers for the category of healthcare available through the Malcolm Baldrige list of examiners were reached via phone and asked if they would consider sharing information on the applicant status of their institutions. Information was also solicited whether the examiners' organizations had won state quality awards within the last five years, and whether the examiners were senior administrators in their respective institutions. If the examiners and senior healthcare administrators qualified as experts in healthcare quality improvement and patient safety according to the study criteria described above, they were invited to participate in the study. In general, the study participants were recruited via telephone and/or letter contact and were selected from (1) the list of Malcolm Baldrige examiners, (2) senior administrators from healthcare institutions that had won national and/or state quality awards, and (3) referrals from (1) and (2). The recruitment of participants was discontinued after 23 qualified individuals confirmed their willingness to serve on the Delphi panel.
Importance rating scale
The Delphi panel utilized a four-choice Likert scale for assessing the importance of suggested critical processes for patient safety systems in healthcare institutions. The scale was modelled according to the original importance scale developed by Turoff [26]. The participants in the panel were asked to indicate the importance of the Delphi survey items from 1 to 4, where "4" represented processes very important to patient safety systems in healthcare institutions, and "1" represented unimportant (irrelevant). All survey items that were identified by the expert group as "very important" or "important" for patient safety in the third study round, when the experts reached consensus, were included in the final patient safety tool. The Delphi survey concluded in three rounds with creation of a process-centred tool for evaluating patient safety performance and guiding strategic improvement at the institutional level, extending the MBNQA criteria to the area of patient safety [29].
Bootstrap study design
After the Delphi panel created the patient safety tool, the concern about possible group bias with small expert numbers remained, because it has been argued that increased group size is beneficial in Delphi surveys [27]. To study possible differences in response characteristics and to explore the possibility for group bias in the study group of experts and, therefore, to assess the possible error in the creation of the patient safety tool, we generated via computer program (SPSS 12.5) two large samples of expert ratings. Since the variation in expert opinions was greatest in the first study round, encompassing the whole spectrum of possible ratings from 1 to 4, the results from the first survey round were utilized as the basis for computer generation of the expanded samples. The expert responses to the survey items were randomly selected with replacement by the computer program based on the raw data from the first round for the actual survey experts. This resampling technique is called bootstrap.
The bootstrap method was developed by Efron in 1979 and has found wide use in the field of applied statistics [30]. Bootstrap is a Monte Carlo-type data augmentation method utilizing resampling with replacement that can be used with observed data. While Monte Carlo techniques usually generate fictitious data, bootstrap resamples with replacement from the original observed values and generates multiple bootstrap samples as a proxy to the independent real sample. Each bootstrap sample is a random sub-sample (with the same size as the original sample) taken with replacement from the observed values. The original sample is treated as the "virtual population" and the sample is duplicated multiple times. The procedure can be repeated as many times as desired. Resampling has proven valid for any kind of data, including random and non-random data [31]. During the last three decades, the bootstrap resampling has been used widely in applied statistics [32].
Advantages and limitations of the bootstrap technique
Resampling (bootstrapping) of a random sample of an unknown population is considered to model the distribution of that population, where the vaguer the knowledge about the population distribution is, the more valuable the bootstrapping technique proves to be [33]. Since classical statistical techniques are primarily designed for parametric statistics with normal distributions, the bootstrap technique has an advantage in distributions with no convenient statistical formulae, overcoming the limitations of the classical approaches in working with small sample sizes and non-normal distributions [34]. Efron and Tibshirani proposed that the technique reduces the assumptions required to validate analysis and eliminates theoretical calculations required to assess accuracy; its major application is in determination of confidence intervals, where 1,000 or more iterations are necessary to estimate the confidence intervals [30]. The simplicity of the method allows its application in a wide variety of studies and is considered superior to standard statistical tests of significance because it reduces the threat of multiple comparisons bias and provides information on the distribution of scores (and not parametric distributions); the technique is not dependent on a specific nominal size such as 5% and therefore is more accurate [34,35]. The bootstrap technique may have limited accuracy in very small sample sizes (n < 20), in extremely skewed distributions, and if extreme outliers are present [34].
Data manipulation
In this study, statistics for each bootstrap resample were saved in memory and later used for estimation of sampling variance, confidence intervals and assessment of bias for the raw data [30]. The characteristics of the generated samples, when analyzed collectively are used to provide a more representative expression of the underlying population, in this study – the population of patient safety experts knowledgeable about the Malcolm Baldrige framework. The hypothesis was that strict expert inclusion criteria based on training in, and knowledge and understanding of the MBNQA framework in the original sample of 23 experts would provide stability of responses, even if the number of responses was increased by computer generated bootstrap samples.
The bootstrap samples were generated using SPSS 12.5 software. The characteristics of the SPSS 12.5 model were as follows:
NonLinear Regression
MODEL PROGRAM b0 = 2
COMPUTE PRED_ = b0
CNLR VAR00001
/OUTFILE='C:\DOCUME~1\User\LOCALS~1\Temp\spss452\SPSSFNLR.TMP'
/PRED PRED_
/BOOTSTRAP 1000 [2000]
/CRITERIA STEPLIMIT 2 ISTEP 1E+20.
More specifically, the regression routine and the subroutine of nonlinear regression were employed. Once these routines were selected, the options feature of nonlinear regression was invoked. The bootstrap option was selected and the "paste" option was used to indicate the number of bootstrap samples to be derived.
Results
The patient safety tool created by the Delphi panel was based on the seven Malcolm Baldrige categories and included 54 survey items (38 critical processes ranked directly, and one indirectly defined by 16 performance measures). For each survey item, the following inferential statistics were derived:
- Mean (average) – the measure that represents the arithmetic average for the group of experts
- 95% confidence interval – representing the upper and lower limits between which 95% of the sample expert scores will be expected to fall
- 5% trimmed mean – calculation of experts' average score with exclusion of the highest and lowest 5% of the scores; the difference between the mean and the trimmed mean shows whether there are many outliers in the rankings among the experts in the sample (real or computer generated)
- Standard deviation – describes the variability of the score distribution.
These descriptive statistics were calculated for each of the following three samples:
(a) original expert responses of 23 experts in healthcare quality and patient safety (n = 23)
(b) augmented-computer-generated sample of 1000 iterations of resampling and the original sample (n = 1001)
(c) augmented-computer-generated sample of 2000 iterations of resampling and the original sample (n = 2001).
The results for each of the calculated statistics (for each of the three groups of data) are presented in Table 1.
Table 1 Comparison between statistical data for the original and augmented samples
Survey Item Data Entries Mean 95% CI 5% Trim. Mean SD
1 Participants, n = 23 3.64 3.39–3.88 3.70 .5680
Iterations, n = 1001 3.64 3.63–3.65 3.64 .1157
Iterations, n = 2001 3.64 3.63–3.64 3.64 .1172
2 Participants, n = 23 3.64 3.39–3.88 3.70 .5680
Iterations, n = 1001 3.64 3.63–3.65 3.64 .1154
Iterations, n = 2001 3.64 3.63–3.64 3.64 .1151
3 Participants, n = 23 3.46 3.12–3.79 3.55 .7820
Iterations, n = 1001 3.45 3.44–3.46 3.46 .1582
Iterations, n = 2001 3.46 3.45–3.46 3.46 .1589
4 Participants, n = 23 3.26 2.99–3.53 3.29 .6190
Iterations, n = 1001 3.26 3.26–3.27 3.27 .1257
Iterations, n = 2001 3.26 3.25–3.26 3.26 .1276
5 Participants, n = 23 3.13 2.89–3.37 3.14 .5480
Iterations, n = 1001 3.13 3.13–3.14 3.13 .1093
Iterations, n = 2001 3.13 3.13–3.14 3.13 .1133
6 Participants, n = 23 3.22 2.93–3.51 3.24 .6710
Iterations, n = 1001 3.22 3.21–3.23 3.22 .1347
Iterations, n = 2001 3.22 3.21–3.22 3.22 .1371
7 Participants, n = 23 3.11 2.85–3.36 3.12 .5950
Iterations, n = 1001 3.10 3.09–3.11 3.10 .1235
Iterations, n = 2001 3.10 3.10–3.11 3.10 .1226
8 Participants, n = 23 3.14 2.81–3.46 3.20 .7570
Iterations, n = 1001 3.14 3.13–3.14 3.14 .1505
Iterations, n = 2001 3.14 3.13–3.14 3.14 .1569
9 Participants, n = 23 3.18 2.85–3.52 3.25 .7770
Iterations, n = 1001 3.18 3.17–3.19 3.18 .1576
Iterations, n = 2001 3.18 3.18–3.19 3.19 .1580
10 Participants, n = 23 3.17 2.93–3.40 3.18 .5436
Iterations, n = 1001 3.17 2.95–3.38 3.16 .1088
Iterations, n = 2001 3.17 2.95–3.38 3.15 .1098
11 Participants, n = 23 3.30 2.95–3.66 3.39 .8220
Iterations, n = 1001 3.30 3.29–3.31 3.30 .1685
Iterations, n = 2001 3.31 3.30–3.31 3.31 .1652
12 Participants, n = 23 3.09 2.74–3.43 3.14 .7930
Iterations, n = 1001 3.09 3.08–3.10 3.09 .1600
Iterations, n = 2001 3.09 3.08–3.10 3.09 .1583
13 Participants, n = 23 2.87 2.54–3.20 2.90 .7570
Iterations, n = 1001 2.86 2.86–2.87 2.87 .1489
Iterations, n = 2001 2.87 2.86–2.88 2.87 .1528
14 Participants, n = 23 2.53 2.23–2.82 2.59 .6821
Iterations, n = 1001 2.52 2.52–2.53 2.53 .1412
Iterations, n = 2001 2.83 2.82–2.83 2.83 .1578
15 Participants, n = 23 2.83 2.49–3.16 2.86 .7780
Iterations, n = 1001 2.82 2.81–2.83 2.83 .1594
Iterations, n = 2001 2.83 2.82–2.83 2.83 .1578
16 Participants, n = 23 3.23 2.91–3.54 3.25 .7340
Iterations, n = 1001 3.23 3.22–3.24 3.23 .1495
Iterations, n = 2001 3.23 3.22–3.23 3.23 .1490
17 Participants, n = 23 3.35 3.01–3.68 3.39 .7750
Iterations, n = 1001 3.35 3.34–3.36 3.35 .1569
Iterations, n = 2001 3.35 3.34–3.36 3.35 .1557
18 Participants, n = 23 3.39 3.08–3.70 3.43 .7220
Iterations, n = 1001 3.39 3.39–3.40 3.40 .1455
Iterations, n = 2001 3.39 3.38–3.40 3.39 .1483
19 Participants, n = 23 3.33 3.01–3.65 3.42 .7390
Iterations, n = 1001 3.33 3.32–3.33 3.33 .1536
Iterations, n = 2001 3.33 3.33–3.34 3.33 .1528
20 Participants, n = 23 3.57 3.28–3.85 3.63 .6620
Iterations, n = 1001 3.57 3.56–3.58 3.57 .1318
Iterations, n = 2001 3.56 3.56–3.57 3.56 .1355
21 Participants, n = 23 2.96 2.65–3.26 3.00 .7057
Iterations, n = 1001 2.96 2.67–3.24 2.93 .1427
Iterations, n = 2001 2.96 2.68–3.23 2.95 .1384
22 Participants, n = 23 3.39 3.11–3.68 3.43 .6560
Iterations, n = 1001 3.39 3.38–3.40 3.39 .1385
Iterations, n = 2001 3.39 3.38–3.39 3.39 .1323
23 Participants, n = 23 3.30 2.95–3.66 3.39 .8220
Iterations, n = 1001 3.30 3.29–3.31 3.30 .1706
Iterations, n = 2001 3.30 3.30–3.31 3.31 .1714
24 Participants, n = 23 3.39 3.18–3.61 3.38 .4990
Iterations, n = 1001 3.39 3.39–3.40 3.39 .1024
Iterations, n = 2001 3.39 3.39–3.40 3.39 .0986
25 Participants, n = 23 3.48 3.22–3.73 3.52 .5930
Iterations, n = 1001 3.48 3.47–3.49 3.48 .1243
Iterations, n = 2001 3.47 3.47–3.48 3.48 .1195
26 Participants, n = 23 2.83 2.42–3.23 2.86 .9370
Iterations, n = 1001 2.83 2.81–2.84 2.83 .1874
Iterations, n = 2001 2.82 2.82–2.83 2.82 .1915
27 Participants, n = 23 2.89 2.61–3.18 2.93 .6670
Iterations, n = 1001 2.89 2.89–2.90 2.90 .1387
Iterations, n = 2001 2.89 2.89–2.90 2.89 .1369
28 Participants, n = 23 3.26 2.91–3.61 3.34 .8100
Iterations, n = 1001 3.26 3.25–3.27 3.27 .1744
Iterations, n = 2001 3.26 3.25–3.26 3.26 .1675
29 Participants, n = 23 2.83 2.44–3.21 2.86 .8870
Iterations, n = 1001 2.82 2.81–2.84 2.83 .1845
Iterations, n = 2001 2.83 2.82–2.84 2.83 .1814
30 Participants, n = 23 3.43 3.12–3.75 3.48 .7280
Iterations, n = 1001 3.44 3.43–3.45 3.44 .1496
Iterations, n = 2001 3.44 3.43–3.44 3.44 .1473
31 Participants, n = 23 3.09 2.74–3.43 3.14 .7930
Iterations, n = 1001 3.08 3.07–3.09 3.09 .1576
Iterations, n = 2001 3.09 3.08–3.09 3.09 .1639
32 Participants, n = 23 3.06 2.75–3.36 3.11 .7050
Iterations, n = 1001 3.05 3.04–3.06 3.05 .1500
Iterations, n = 2001 3.05 3.05–3.06 3.06 .1428
33 Participants, n = 23 3.17 2.84–3.51 3.24 .7780
Iterations, n = 1001 3.17 3.16–3.18 3.17 .1602
Iterations, n = 2001 3.17 3.17–3.18 3.18 .1579
34 Participants, n = 23 3.26 2.94–3.59 3.34 .7520
Iterations, n = 1001 3.27 3.26–3.28 3.28 .1528
Iterations, n = 2001 3.26 3.26–3.27 3.27 .1512
35 Participants, n = 23 3.39 3.11–3.68 3.43 .6560
Iterations, n = 1001 3.39 3.39–3.40 3.40 .1333
Iterations, n = 2001 3.39 3.38–3.39 3.39 .1314
36 Participants, n = 23 3.06 2.85–3.26 3.06 .4740
Iterations, n = 1001 3.06 3.05–3.07 3.06 .0991
Iterations, n = 2001 3.06 3.05–3.06 3.06 .0972
37 Participants, n = 23 3.32 3.05–3.59 3.35 .6310
Iterations, n = 1001 3.31 3.30–3.32 3.31 .1282
Iterations, n = 2001 3.31 3.31–3.32 3.31 .1294
38 Participants, n = 23 2.89 2.60–3.18 2.93 .6670
Iterations, n = 1001 2.89 2.88–2.90 2.89 .1369
Iterations, n = 2001 2.89 2.89–2.90 2.89 .1346
39 Participants, n = 23 3.22 2.90–3.54 3.24 .7360
Iterations, n = 1001 3.22 3.21–3.23 3.22 .1543
Iterations, n = 2001 3.48 3.47–3.48 3.48 .1022
40 Participants, n = 23 3.48 3.16–3.79 3.53 .7300
Iterations, n = 1001 3.48 3.47–3.49 3.48 .1519
Iterations, n = 2001 3.48 3.47–3.48 3.48 .1465
41 Participants, n = 23 3.39 3.08–3.70 3.43 .7220
Iterations, n = 1001 3.40 3.39–3.41 3.40 .1492
Iterations, n = 2001 3.40 3.39–3.40 3.40 .1450
42 Participants, n = 23 3.65 3.44–3.86 3.67 .4870
Iterations, n = 1001 3.65 3.64–3.66 3.65 .1041
Iterations, n = 2001 3.65 3.65–3.66 3.65 .1003
43 Participants, n = 23 3.61 3.32–3.89 3.68 .6560
Iterations, n = 1001 3.61 3.60–3.62 3.61 .1338
Iterations, n = 2001 3.61 3.60–3.61 3.61 .1362
44 Participants, n = 23 3.61 3.39–3.82 3.62 .4990
Iterations, n = 1001 3.61 3.60–3.62 3.61 .1020
Iterations, n = 2001 3.61 3.60–3.61 3.61 .1036
45 Participants, n = 23 3.30 3.00–3.61 3.34 .7030
Iterations, n = 1001 3.31 3.30–3.32 3.31 .1461
Iterations, n = 2001 3.30 3.30–3.31 3.30 .1424
46 Participants, n = 23 3.43 3.07–3.80 3.53 .8430
Iterations, n = 1001 3.43 3.42–3.45 3.44 .1814
Iterations, n = 2001 3.44 3.43–3.44 3.44 .1735
47 Participants, n = 23 3.39 3.11–3.68 3.43 .6560
Iterations, n = 1001 3.39 3.38–3.40 3.39 .1306
Iterations, n = 2001 3.39 3.38–3.40 3.39 .1368
48 Participants, n = 23 3.48 3.26–3.70 3.48 .5110
Iterations, n = 1001 3.48 3.47–3.48 3.48 .1037
Iterations, n = 2001 3.48 3.47–3.48 3.48 .1022
49 Participants, n = 23 3.61 3.39–3.82 3.62 .4990
Iterations, n = 1001 3.61 3.60–3.62 3.61 .1018
Iterations, n = 2001 3.60 3.60–3.61 3.60 .1003
50 Participants, n = 23 3.26 2.94–3.59 3.29 .7520
Iterations, n = 1001 3.25 3.25–3.26 3.26 .1490
Iterations, n = 2001 3.27 3.26–3.27 3.27 .1527
51 Participants, n = 23 3.30 3.03–3.58 3.34 .6350
Iterations, n = 1001 3.31 3.30–3.31 3.31 .1313
Iterations, n = 2001 3.30 3.30–3.31 3.30 .1316
52 Participants, n = 23 3.39 3.11–3.68 3.43 .6560
Iterations, n = 1001 3.39 3.38–3.40 3.39 .1336
Iterations, n = 2001 3.39 3.39–3.40 3.39 .1325
53 Participants, n = 23 3.18 2.85–3.50 3.25 .7530
Iterations, n = 1001 3.18 3.17–3.19 3.18 .1542
Iterations, n = 2001 3.18 3.17–3.19 3.18 .1543
54 Participants, n = 23 3.18 2.85–3.50 3.25 .7530
Iterations, n = 1001 3.17 3.16–3.18 3.17 .1556
Iterations, n = 2001 3.18 3.18–3.19 3.18 .1550
In general, the means of the expert scores per critical process remained stable across the three data sets (one original and two augmented); the confidence intervals of the three samples were overlapping for each of the critical processes with the confidence intervals in bigger samples being more compact; the trimmed mean exhibited stability, and the standard deviation decreased with increasing number of experts. When the standard deviation decreases, this indicates that the typical deviation of expert opinions has not increased relative to the increased number of responses. Therefore, we can conclude that the original expert sample yielded results concerning the importance of the critical processes in the patient safety tool that were comparable to the results of the expanded samples.
The Delphi survey results showed stability after bootstrap resampling data expansion. Therefore, the stability of the results of the bootstrap data expansion validated the patient safety strategic planning tool developed by the Delphi study on patient safety [29].
Discussion
Application of Delphi studies in healthcare
In the field of healthcare, the Delphi method has been used in planning for the future and formulating policies and programs in biomedical research, behavioral research, mental health, reproductive health, pharmacology, services for the elderly, family planning services, accidents and injuries, development of core competencies for advanced nursing practitioners and development of clinical care protocols [27,36-41]. The Delphi method, as a useful way of identifying and measuring uncertainty, has been widely utilized in medical and health services research to explore issues in health services organizations, to support design of educational programs, to define professionals' roles, to define effects of medical staffing levels, to develop criteria for appropriateness of clinical treatment, and to make long-term projections of need for patient care [42]. The Delphi methodology also has been used to modify the National Board of Medical Examiners (NBME) Medicine Subject Exam (Shelf) in order to align the national exams with the internal medicine clerkship curriculum developed by the Society of General Internal Medicine (SGIM) and the Clerkship Directors in Internal Medicine (CDIM) [43]. Consensus building through Delphi survey technique can contribute significantly to broadening knowledge and effective decision making in health and social care [22]. Furthermore, the Delphi approach can be used as a senior management education tool, environmental planning tool, and for comparison with similar healthcare institutions. Putting together the structure of a model, developmental planning and exploration of policy options are among the explicit application areas identified since the early 1970's [26]. Additionally, the Delphi method has been used to delineate the barriers to performance in health services and identify three types of barriers to optimal healthcare performance: solution development barriers, problem selection barriers, and evaluation barriers. It has been argued that the forecasting accuracy of Delphi studies is strongly reliable; for example, a Delphi study with medical doctors evaluating the forecasting application of the method, revealed that in 75% of the cases the estimated values proved to be less than 10% different from the observed [26].
Challenges in selection of a Delphi panel
The questions arising around the formation of a Delphi panel are typical for selection and formation of any group – committee, task force, panel, or study group. Thus, while panel member selection is a problem that should be addressed, it is by no means unique to Delphi studies. It has been argued that the amount of bias expressed by study participants is offset by the fact that in answering the questions each participant exhibits a standard deviation which is comparable to, or greater than participant's individual mean (i.e., an optimistic panellist is pessimistic in some of his/her responses, and vice versa) [26].
The selection of criteria that would qualify an individual to participate on the Delphi panel depends on the context, scope and aims of the particular study. Some of the general criteria include:
• Knowledge and practical engagement with the issue under investigation
• Capacity and willingness to contribute to the exploration of a particular problem
• Assurance that sufficient time will be dedicated to the Delphi exercise,
• Good written communication skills
• Experts' skills and knowledge need not necessarily be accompanied by standard academic qualifications or degrees [44].
Delphi participants are purposefully selected to apply their knowledge and experience to a certain issue based on a criteria set. For example, the experts for a national two-round Delphi study on the effectiveness and risks of coronary angiography were chosen on the basis of their clinical expertise, community influence, and diversity of geographic location [45]. Since the Delphi method relies on repeated questionnaires to the same initially selected sample of participants, the method requires a continued commitment from the panellists and is heavily dependant on the time and continued involvement on the part of the study participants. The widespread employment of electronic communications calls for consideration of the computer literacy and skills of the target sample before utilizing electronic means of communication [22]. Since the sample size for Delphi panels has not been established, it is important to know whether the selected Delphi panel for a particular study would yield stable results.
Utilization of bootstrap in healthcare research
During the last decade, the use of bootstrap data expansion in healthcare research became more prominent. Some examples of utilization of bootstrap in contemporary healthcare research include: constructing confidence intervals for treatment differences, analysing cost-effectiveness in randomized controlled trials, assessing provider performance for providers with small numbers of observed events, and studying the genetic linkages in viral genome sequences [46-51]. There is a growing validation of the value of bootstrap in medical statistics and an increasing recognition that the bootstrap technique can supplement and extend the conventional statistical thinking. Bootstrap can be used for calculation of confidence intervals, hypothesis testing, linear regression and correlation in variable prediction, and non-linear regression in immunoassay techniques [33]. The bootstrap method, with its computational simplicity and performance similar to the fully Bayesian approaches, was found to be a very useful addition to healthcare researchers' statistical toolkit [52].
Conclusion
Although experiments carried out in the 1950's and 1960's suggested that group error is reduced with increased group size, the sample size for constructing a Delphi panel is not a statistically-bound decision and good results can be obtained by a comparatively small group of homogenous experts [44]. However, the size of a "small" Delphi sample has not been unequivocally established.
The findings of this study are important because:
1. It was found that reliable outcomes could be obtained with a Delphi panel consisting of a relatively small number of Delphi experts (23) selected via strict inclusion criteria. This finding is particularly important for conducting Delphi surveys in knowledge or practice fields where the population of experts (the total number of qualifying knowledgeable individuals) is limited. Experts who have similar training and general understanding in the field of interest allow for effective and reliable utilization of a small sample from a limited number of experts in the field of study.
2. The stability of the results from a Delphi survey conducted with a small number of experts from a defined field of study was established. It is hypothesized that Delphi surveys in other fields of study, conducted with a small number of experts with similar training and knowledge, would also yield reliable results. Given the fact that the number of specialized experts in a particular field may be limited, this study validated the stability of response characteristics of a small expert sample from limited numbers of experts. Therefore, consistency of expert training may allow utilization of small numbers of experts in fields where many experts may be available but participation of a limited number of experts on the Delphi panel may be more practical.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
RBA, HT and BRC developed the project idea and the study design. RBA collected the data. RBA and HT conducted the study. HT provided statistical advice and BRC provided conceptual advice. All authors contributed to the analysis of data. The manuscript was initially drafted by RBA. HT and BRC critically reviewed and edited the manuscript. All authors contributed to and approved the final version submitted for publication.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors would like to thank the participants in the Delphi study for their commitment and dedication. The authors would also like to express their gratitude to Ms. Xue Hu (Texas A&M University) for her skills and precision in handling survey data in support of this paper. Additionally, the authors would like to recognize one individual, who chose to remain anonymous, for her support in running the SPSS bootstrap model.
==== Refs
Williams PL Webb C The Delphi technique: a methodological discussion J Adv Nurs 1994 19 180 6 8138622
Fischer RG The Delphi method: a description, review and criticism J Acad Librariansh 1978 4 64 70
Wilhelm WJ Alchemy of the Oracle: the Delphi technique The Delta Pi Epsilon Journal 2001 43 6 26
Reid NG Ellis R The Delphi technique: its contribution to the evaluation of professional practice Professional competence and quality assurance in the caring professions 1988 New York, NY: Chapman and Hall
Malone DC Abarca J Hansten PD Grizzle AJ Armstrong EP Van Bergen RC Duncan-Edgar BS Solomon SL Lipton RB Identification of serious drug-drug interactions: results of the partnership to prevent drug-drug interactions Am J Geriatr Pharmacother 2005 3 65 76 10.1016/j.amjopharm.2005.05.001
Strasser S London L Kortenbout E Developing a competence framework and evaluation tool for primary care nurses in South Africa Educ Health 2005 18 133 144
De Vet E Brug J De Nooijer J Dijkstra A De Vries NK Determinants of forward stage transitions: a Delphi study Health Educ Res 2005 20 195 205 15328302 10.1093/her/cyg111
Howell S Kemp C Defining early number sense: a participatory Australian study J Educ Psychol 2005 25 555 571 10.1080/01443410500046838
Nekolaichuk CL Fainsinger RL Lawlor PG A validation study of a pain classification system for advanced cancer patients using content experts: the Edmonton Classification System for cancer pain Palliat Med 2005 19 466 476 16218159 10.1191/0269216305pm1055oa
Armstrong D Marshall JK Chiba N Enns R Fallone CA Fass R Hollingworth R Hunt RH Kahrilas PJ Mayrand S Moayyedi P Paterson WG Sadowski D van Zanten SJ Canadian consensus conference on the management of gastroesophageal reflux disease in adults – update 2004 Can J Gastroenterol 2005 19 15 35 15685294
Stolee P Esbaugh J Aylward S Cathers T Harvey DP Hillier LM Keat N Feightner JW Factors associated with the effectiveness of continuing education in long-term care Gerontologist 2005 45 399 405 15933280
Al-Marzouki S Roberts I Marshall T Evans S The effect of scientific misconduct on the results of clinical trials: a Delphi survey Contemp Clin Trials 2005 26 331 337 15911467 10.1016/j.cct.2005.01.011
McKenna H Keeney S Bradley M Nurse leadership within primary care: The perceptions of community nurses, GPs, policy makers and members of the public J Nurs Manag 2004 12 69 76 15101458
Mcilfatrick SJ Keeney S Identifying cancer nursing research priorities using the Delphi technique J Adv Nurs 2003 42 629 636 12787236 10.1046/j.1365-2648.2003.02666.x
Coico R Kachur E Lima V Lipper S Guidelines for preklership bioterrorism curricula Acad Med 2004 79 366 375 15044173 10.1097/00001888-200404000-00019
Telford R Boote JD Cooper CL What does it mean to involve consumers successfully in NHS research? A consensus study Health Expect 2004 7 209 220 15327460 10.1111/j.1369-7625.2004.00278.x
Meadows AB Maine LL Keyes EK Pearson K Finstuen K Pharmacy executive leadership issues and associated skills, knowledge and abilities J Am Pharm Assoc 2005 45 55 62 10.1331/1544345052843183
Kelly KP Porock D A survey of paediatric oncology nurses' perceptions of parent educational needs J Pediatr Oncol Nurs 2005 22 58 66 15574727 10.1177/1043454204272537
Cypert FR Gant WL The Delphi technique: A case study Phi Delta Kappan 1971 272 273
Barnette JJ Danielson LC Algozzine RF Delphi methodology: an empirical investigation Educational Research Quarterly 1978 3 67 73
Synnott G McKie D International issues in PR: Researching research and prioritising priorities Journal of Public Relations Research 1997 9 259 282 10.1207/s1532754xjprr0904_02
Hasson F Keeney S McKenna H Research guidelines for the Delphi survey technique J Adv Nurs 2000 32 1008 1015 11095242 10.1046/j.1365-2648.2000.01567.x
Robinson AR Hohmann KB Rifkin JI Topp D Gilroy CM Pickard JR Anderson RJ Physician and public opinions on quality of health care and the problem of medical errors Arch Intern Med 2002 162 2186 2190 12390060 10.1001/archinte.162.19.2186
Sexton JB Helmreih RL Rowan K Vella K Boyden J Neilands TB Roberts PR Thomas EJ The safety attitude questionnaire: a psychometric validation Technical report 2003 The University of Texas Center of Excellence for Patient Safety Research and Practice, Austin, Texas AHRQ Grants 1PO1HS1154401 and U18HS1116401
Baldrige National Quality Program 2005 Healthcare criteria for performance excellence 2005 Gaithersburg, MD: United States Department of Commerce, Technology Administration, National Institute of Standards and Technology
Linstone HA Turoff M The Delphi Method: techniques and applications 1975 Reading, MA: Addison-Wesley Publishing Company
Adler M Ziglio E Gazing into the Oracle: the Delphi method and its application to social policy and public health 1996 Bristol, PA: Jessica Kingsley Publishers, Ltd
Rotondi A Gustafson D Adler M, Ziglio E Theoretical, methodological and practical issues arising out of the Delphi method Gazing into the Oracle: the Delphi method and its application to social policy and public health 1996 Bristol, PA: Jessica Kingsley Publishers, Ltd 34 55
Akins RB A Process-centered tool for evaluating patient safety performance and guiding strategic improvement Advances in patient safety: from research to implementation 2005 4 Rockville, MD: Agency for Healthcare Research and Quality 109 125
Efron B Tibshirani R An introduction to the bootstrap Monographs on statistics and applied probability, No 57 1993 London: Chapman and Hall
Edgington ES Randomization tests 1995 New York: M. Dekker
Good P Permutation tests: a practical guide for resampling methods for testing hypotheses 2000 New York: Springer-Verlag
Henderson AR The bootstrap: a technique for data-driven statistics. Using computer intensive analyses to explore experimental data Clin Chim Acta 2005 359 1 26 15936746 10.1016/j.cccn.2005.04.002
Linden A Adams JL Roberts N Evaluating disease management program effectiveness: an introduction to the bootstrap technique Disease Management & Health Outcomes 2005 13 159 167
Lloyd CJ On comparing the accuracy of competing tests of the same hypotheses of simulation data J Stat Plan Inference 2005 128 497 508 10.1016/j.jspi.2003.12.002
Calzone KA Jenkins J Masny A Core competencies in cancer genetics for advanced practice oncology nurses Oncol Nurs Forum 2002 29 1327 1333 12370702
Wang CC Wang Y Zhang K Fang J Liu W Luo S Tang S Wang S Li VC Reproductive health indicators for China rural areas Soc Sci Med 2003 57 217 225 12765703 10.1016/S0277-9536(02)00341-6
Catlin AJ Carter BS Creation of a neonatal end-of-life palliative care protocol J Perinat Neonal Nurs 2002 21 37 49
Sharp D Liebenau A Stocks N Bennewith O Evans M Jones WB Peters T Goldberg D Gunnell D Locally developed guidelines for the aftercare of deliberate self-harm patients in general practice Primary Health Care Research and Development 2003 4 21 28 10.1191/1463423603pc119oa
Escobar A Qintana JM Arostegui I Azkarate J Guenaga JI Arenaza JC Garai I Development of explicit criteria for total knee replacement Int J Technol Assess Health Care 2003 19 57 70 12701939 10.1017/S0266462303000060
McBride AJ Pates R Ramadan R McGowan C Delphi survey of experts' opinions on strategies used by community pharmacists to reduce over-the-counter drug misuse Addiction 2003 98 487 497 12653818 10.1046/j.1360-0443.2003.00345.x
Jones J Hunter D Consensus methods for medical and health services research Br Med J 1995 311 376 380 7640549
Elnicki DM Lescisin DA Case S Improving the National Board of Medical Examiners Internal Medicine Subject Exam for use in clerkship evaluation J Gen Intern Med 2002 17 435 440 12133157 10.1046/j.1525-1497.2002.10673.x
Ziglio E Adler M, Ziglio E The Delphi method and its contribution to decision making Gazing into the Oracle: the Delphi method and its application to social policy and public health 1996 Bristol, PA: Jessica Kingsley Publishers, Ltd 3 33
Bernstein SJ Laouri M Hilborne LH Leape LL Kahan JP Park RE Kamberg CJ Brook RH Coronary Angiography: a literature review and ratings of appropriateness and necessity 1992 The New York City Rand Institute, JRA-03. Santa Monica, CA: The Rand Corporation
Maetzel A Strand V Tugwell P Wells G Bombardier C Economic comparison of Leflunomide and Methotrexate in patients with rheumatoid arthritis Pharmacoeconomics 2002 20 61 70 11817993
Manktelow B Abrams K Draper E A comparison of six methods of estimating confidence intervals for an indirectly standardized outcome ratio J Epidemiol Community Health A36
Peris F Martinez E Badia X Brosa M Iatrogenic cost factors incorporating mild and moderate adverse events in the economic comparison of Aceclofenac and other NSAIDs Pharmacoeconomics 2001 19 779 790 11548913
Neymark N Adriaenssen I Gorlia T Caleo S Bolla M Brochon D Cost-effectiveness of the addition of early hormonal therapy in locally advanced prostate cancer: results decisively determined by the cut-off time point chosen for the analysis Eur J Cancer 2001 37 1768 1774 11549430 10.1016/S0959-8049(01)00197-6
Chen M Kianifard F A nonparametric procedure associated with a clinically meaningful efficacy measure Biostatistics 2000 1 293 298 12933510 10.1093/biostatistics/1.3.293
Bredell H Crookes RL Heynes du PA Schoub BD Morris L Molecular investigation of two possible cases of accidental HIV-1 transmission in South Africa AIDS Res Hum Retroviruses 2003 19 613 617 12921093 10.1089/088922203322230987
MacNab YC Farrel PJ Gustafson P Wen S Estimation in Bayesian disease mapping Biometrics 2004 60 865 873 15606406 10.1111/j.0006-341X.2004.00241.x
|
16316460
|
PMC1318467
|
CC BY
|
2021-01-04 16:29:46
|
no
|
BMC Dermatol. 2005 Nov 29; 5:12
|
latin-1
|
BMC Dermatol
| 2,005 |
10.1186/1471-5945-5-12
|
oa_comm
|
==== Front
Biomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 1475-925X-4-661633664810.1186/1475-925X-4-66CorrectionCorrection: Mechanical properties of femoral trabecular bone in dogs Pressel Thomas [email protected] Anas [email protected] Ute [email protected] Andrea [email protected] Bernd-Arno [email protected] Ingo [email protected] Henning [email protected] Department of Orthopaedic Surgery, Hannover Medical School, Anna-von-Borries-Str. 1–7, 30625 Hannover, Germany2 Institute of Metal Forming and Metal Forming Machine Tools, University of Hannover, Schönebecker Allee 2, 30823 Garbsen, Germany3 Clinic for Small Domestic Animals, School of Veterinary Medicine Hannover, Bischofsholer Damm 15, 30173 Hannover, Germany2005 7 12 2005 4 66 66 2 12 2005 7 12 2005 Copyright © 2005 Pressel 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.
==== Body
After the publication of this work [1], we became aware of the fact that the frequency of the ultrasound transmitter that we used for determining the elastic moduli of the trabecular bone specimens was not correctly specified. The oscillation frequency of the ultrasound transmitter was 2 MHz (and not 100 MHz as stated in our work) while we used a sampling rate of 100 MHz. In our publication, the oscillation frequency and sampling rate were confounded. Therefore also the statement in the discussion that we might have determined elastic moduli of trabecular bone tissue rather than the elastic properties of whole specimens because we used an ultrasound frequency > 2 MHz is wrong and has to be omitted.
For measurement, the cubic bone specimens were not immersed in Ringer's solution but only were kept moist all the time.
Apart from these corrections concerning the methods and interpretation of the data, the results reported in our publication and the conclusions are absolutely correct.
We apologize for the inconvenience that this inaccuracy may have caused.
==== Refs
Pressel T Bouguecha A Vogt U Meyer-Lindenberg A Behrens BA Nolte I Windhagen H Mechanical properties of femoral trabecular bone in dogs Biomed Eng Online 2005 4 17 15774014 10.1186/1475-925X-4-17
|
0
|
PMC1318468
|
CC BY
|
2021-01-04 16:03:54
|
no
|
Biomed Eng Online. 2005 Dec 7; 4:66
|
utf-8
|
Biomed Eng Online
| 2,005 |
10.1186/1475-925X-4-66
|
oa_comm
|
==== Front
Biomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 1475-925X-4-661633664810.1186/1475-925X-4-66CorrectionCorrection: Mechanical properties of femoral trabecular bone in dogs Pressel Thomas [email protected] Anas [email protected] Ute [email protected] Andrea [email protected] Bernd-Arno [email protected] Ingo [email protected] Henning [email protected] Department of Orthopaedic Surgery, Hannover Medical School, Anna-von-Borries-Str. 1–7, 30625 Hannover, Germany2 Institute of Metal Forming and Metal Forming Machine Tools, University of Hannover, Schönebecker Allee 2, 30823 Garbsen, Germany3 Clinic for Small Domestic Animals, School of Veterinary Medicine Hannover, Bischofsholer Damm 15, 30173 Hannover, Germany2005 7 12 2005 4 66 66 2 12 2005 7 12 2005 Copyright © 2005 Pressel 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.
==== Body
After the publication of this work [1], we became aware of the fact that the frequency of the ultrasound transmitter that we used for determining the elastic moduli of the trabecular bone specimens was not correctly specified. The oscillation frequency of the ultrasound transmitter was 2 MHz (and not 100 MHz as stated in our work) while we used a sampling rate of 100 MHz. In our publication, the oscillation frequency and sampling rate were confounded. Therefore also the statement in the discussion that we might have determined elastic moduli of trabecular bone tissue rather than the elastic properties of whole specimens because we used an ultrasound frequency > 2 MHz is wrong and has to be omitted.
For measurement, the cubic bone specimens were not immersed in Ringer's solution but only were kept moist all the time.
Apart from these corrections concerning the methods and interpretation of the data, the results reported in our publication and the conclusions are absolutely correct.
We apologize for the inconvenience that this inaccuracy may have caused.
==== Refs
Pressel T Bouguecha A Vogt U Meyer-Lindenberg A Behrens BA Nolte I Windhagen H Mechanical properties of femoral trabecular bone in dogs Biomed Eng Online 2005 4 17 15774014 10.1186/1475-925X-4-17
|
16321153
|
PMC1318469
|
CC BY
|
2021-01-04 16:03:50
|
no
|
BMC Ophthalmol. 2005 Dec 1; 5:28
|
latin-1
|
BMC Ophthalmol
| 2,005 |
10.1186/1471-2415-5-28
|
oa_comm
|
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-741633665310.1186/1742-4690-2-74EditorialThe value of Institute of Human Virology meeting abstracts and beyond Jeang Kuan-Teh [email protected] The National Institutes of Health, Bethesda, Maryland, USA2005 7 12 2005 2 74 74 29 11 2005 7 12 2005 Copyright © 2005 Jeang; licensee BioMed Central Ltd.2005Jeang; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
This month Retrovirology publishes the meeting abstracts from the 10th annual Institute of Human Virology conference held August 29th to September 2nd, 2005 in Baltimore, Maryland, USA. In this editorial, the rationale for publishing meeting abstracts is discussed.
==== Body
To celebrate the 10th annual meeting of the Institute of Human Virology (IHV), Retrovirology publishes this month a supplement [1] which includes 315 abstracts of presentations that took place August 29th to September 2nd, 2005 in Baltimore, Maryland, USA. This compilation of meeting abstracts, as with all other items published in Retrovirology, will be listed in PubMed, indexed in MedLine, and permanently archived in PubMed Central. The IHV abstracts will be available for all to read in an unrestricted "Open Access" manner. This latter privilege is important because fully two-thirds of the users of free and open data bases such as PubMed are in fact not academics. The users may be patients, students, teachers, or healthcare professionals; and they would be barred from information-access by a fee-based format.
Is there value to publishing Meeting Abstracts? On several levels, the answer appears to be "yes". While the IHV abstracts are the first of its kind for Retrovirology, our experience with publishing meeting reports [2-5] tells of strong readership interest. Our statistics show that the published 2004 Cold Spring Harbor Retrovirus Meeting Report [2], a meeting attended by ~500 conferees, has been read in Retrovirology 3668 times over the past 14 months. A separate report of the 2005 Twelfth West Coast Retrovirus meeting [5], attended by ~125 scientists, was accessed 610 times in the first ten days after its publication. Independently, Scherer et al. [6] found in a study of 2,391 meeting abstracts that 51% of the abstracts later appeared as full articles in journals. In another survey, 84% of journals were found to permit the citation of meeting abstracts in bibliographies [7]. Because on average an entire year lags between the time that a paper/poster is presented at a meeting and its eventual publication in a journal, publishing meeting abstracts arguably serves to narrow a knowledge gap between those who attended a meeting and those who did not [7]. Moreover, extant data support that the "open access" approach to publishing scientific information promotes a higher rate of citation to the published work [8]. Thus, it stands to reason that there is value for both authors and readers of Retrovirology meeting abstracts.
Let me close this writing by telling you a personal anecdote which illustrated for me why archiving of meeting abstracts is important. In the early 1980's, I was a graduate student working in one of three laboratories worldwide which were competing on the cloning and the characterization of the cytomegalovirus (CMV) immediate-early (IE) promoter. This is the same CMV promoter that is resident in the many mammalian expression vectors which most of you purchase commercially. My memory tells me (although my memory has faded with age) that between 1981 to 1984, I made several presentations on CMV promoter research at the then annual Herpesvirus meeting held at Cold Spring Harbor. Later, in 1990, a patent for the use of the CMV IE promoter was filed by a competitor's group. Many years passed, until approximately five years ago when I unexpectedly received a telephone call from a patent attorney at a high-priced law firm in New York City. The attorney represented a biotech firm which was keen on contesting the issued CMV IE patent. The attorney wanted to know "What did I say publicly about my CMV IE promoter research at meetings?" "And when did I say them?". I recall at that moment when confronted to recall accurately minute historical details critical to a legal contest, I wished fervently for the existence of an open access, easily searchable, repository of meeting abstracts.
Retrovirology is committed to the goal of free public access to permanently archived digitally formatted scientific information. Meeting abstracts published in Retrovirology are initially viewable online in our journal, and are then permanently deposited into the PubMed Central archive. Retrovirology is currently accessed ~1,000 times daily. If you are a meeting organizer interested in the rapid and broad dissemination (with permanent archiving) of the presentations from your conference, it may be worth your while to consider publishing your meeting in Retrovirology.
==== Refs
Abstracts from the 2005 International Meeting of The Instituteof Human Virology Retrovirology 2005 2 S1 10.1186/1742-4690-2-1
Freed EO Ross SR Retroviruses 2004: Review of the 2004 Cold Spring Harbor Retroviruses conference Retrovirology 2004 1 25 15357866 10.1186/1742-4690-1-25
Menu E Müller-Trutwin MC Pancino G Saez-Cirion A Bain C Inchauspé G Gras GS Mabondzo AM Samri A Boutboul F Le Grand R First Dominique Dormont international conference on "Host-pathogen interactions in chronic infections – viral and host determinants of HCV, HCMV, and HIV infections" Retrovirology 2005 2 24 15813969 10.1186/1742-4690-2-24
Lairmore MD Fujii M 12th international conference on human retrovirology: HTLV and related retroviruses Retrovirology 2005 2 61 16202161 10.1186/1742-4690-2-61
Barry SM Melar M Gallay P Hope TJ Review of the twelfth West Coast retrovirus meeting Retrovirology 2005 2 72 16293194 10.1186/1742-4690-2-72
Scherer RW Dickersin K Langenberg P Full publication of results initially presented in abstracts: a meta-analysis JAMA 1994 272 158 162 8015133 10.1001/jama.272.2.158
Kelly JA Scientific meeting abstracts: significance, access, and trends Bull Med Libr Assoc 1998 86 68 76 9549015
Antelman K Do Open-access articles have a greater research impact? College Res Libr News 2004 65 372 382
|
0
|
PMC1318470
|
CC BY
|
2021-01-04 16:36:39
|
no
|
Retrovirology. 2005 Dec 7; 2:74
|
utf-8
|
Retrovirology
| 2,005 |
10.1186/1742-4690-2-74
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2811631646210.1186/1471-2105-6-281Methodology ArticleAn algorithm for the determination and quantification of components of nucleic acid mixtures based on single sequencing reactions Pozhitkov Alexander [email protected] Kathryn [email protected] Diethard [email protected] Institut für Genetik, der Universität zu Köln, Zülpicherstrasse 47, 50674 Köln, Germany2 Civil and Environmental Engineering, University of Washington, Seattle, 98195, WA, USA2005 29 11 2005 6 281 281 6 7 2005 29 11 2005 Copyright © 2005 Pozhitkov et al; licensee BioMed Central Ltd.2005Pozhitkov 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
Determination and quantification of nucleic acid components in a mixture is usually accomplished by microarray approaches, where the mixtures are hybridized against specific probes. As an alternative, we propose here that a single sequencing reaction from a mixture of nucleic acids holds enough information to potentially distinguish the different components, provided it is known which components can occur in the mixture.
Results
We describe an algorithm that is based on a set of linear equations which can be solved when the sequencing profiles of the individual components are known and when the number of sequenced nucleotides is larger than the number of components in the mixture. We have implemented the procedure for one type of sequencing approach, pyrosequencing, which produces a stepwise output of peaks that is particularly suitable for the procedure. As an example we use signature sequences from ribosomal RNA to distinguish and quantify several different species in a mixture. Using simulations, we show that the procedure may also be applicable for dideoxy sequencing on capillary sequencers, requiring only some instrument specific adaptations of protocols and software.
Conclusion
The parallel sequencing approach described here may become a simple and cheap alternative to microarray experiments which aim at routine re-determination and quantification of known nucleic acid components from environmental samples or tissue samples.
==== Body
Background
The identification of individual components from a mixture of nucleic acid sequences relies currently on molecular hybridization approaches, such as microarrays. Technically, these make use of the inherent combinatorial complexity of strings of nucleotides, as well as the base pairing reaction between single stranded molecules. These allow to potentially resolve vast numbers of different sequences, provided one knows which sequences can occur in the mixture. In contrast, DNA sequencing reactions do not require a priori knowledge of the sequence that is to be determined, but they can only identify one sequence at a time. However, the patterns produced in sequencing reactions have also an inherent combinatorial complexity with respect to the order and height of peaks. Accordingly, mixtures of sequencing reactions should also produce characteristic patterns which reflect the components of the mixture. Thus, it should be possible to deconvolute a mixture pattern into its individual components, provided the components that can potentially occur in the mixture are known. We propose here an algorithm that is based on solving a system of linear equations which describe the peak patterns that can theoretically be obtained. The number of sequences that can be resolved with this algorithm depends on the number of sequencing steps that can be recorded. Thus, the procedure can not resolve the very large number of sequences that can potentially be resolved by microarrays, but applications where this high resolution potential is not necessary could be done more efficiently with the parallel sequencing procedure described here.
One particular application, which we use here to exemplify the procedure, would be species identification through DNA taxonomy [1-5]. It has been proposed that the identification of organisms could be based on short characteristic strings of nucleotides from universally occurring genes, such as the mitochondrial cytochrome oxidase gene [3] or ribosomal RNAs [5]. This requires on the one hand to build a database of such sequences and on the other hand to re-sequence given samples and to compare them with the database. The re-sequencing could become more efficient, when several samples could be pooled and determined in a single sequencing reaction. Also, environmental samples of small organisms are usually obtained as mixtures and it would be highly advantageous, if the species within them could be determined without separation. We use here the example of ribosomal RNA signature sequences to show that this approach is indeed promising. For this we employed the pyrosequencing procedure [6], because of its inherent property to separate each sequencing step. However, using simulations, we suggest that dideoxy sequencing procedures [7] might work as well, provided some technical adjustments can be done.
Results
Conventional dideoxy sequencing procedures require single templates from which labeled DNA fragments are produced, which are then resolved by electrophoresis [7]. In the pyrosequencing procedure, each individual nucleotide in a template sequence is probed for incorporation of all possible nucleotides and a light signal is generated when the incorporation is successful [6]. In both procedures, the signal intensity depends on the concentration of the template and the sequence context of the nucleotide that is being determined. The latter depends on the enzyme that is used, as well as on the reaction conditions. However, for a given enzyme and condition, each template will produce a reproducible pattern of signals.
The system of linear equations that is used for the deconvolution algorithm (see methods) can be solved when the number of equations is at least equal to the number of components in the mix. Each equation reflects one sequencing step, i.e. the recorded sequence length determines the complexity of the mixture that can be analysed. In practice, one will require more steps than the theoretical minimum, to compensate for non-informative (i.e. conserved) positions and experimental noise. Since the pyrosequencing procedure produces a stepwise output for each nucleotide position probed, it can be directly employed to test whether the algorithm works under practical conditions.
Implementation for pyrosequencing
The ribosomal RNA sequences that were used for the test were derived from meiobenthos organisms. A special feature of rRNA sequences is the fact that they are composed of a patchwork of highly conserved and highly divergent regions. Hence it is possible to use a single primer for the sequencing reaction next to a region with high diversity which is particularly informative with respect to species discrimination. Figure 1 shows the relevant sequence alignments for the organisms used in this study. The respective fragments were previously cloned [5,8], allowing to determine the individual sequence profiles and to generate defined mixtures. Four different mixes were generated and sequenced in parallel to the single sequences to obtain the respective profiles. Each mix was done in four replicates and the solutions were calculated for each of the four libraries. The estimated standard deviations for each solution (see Methods), as well as for the replicates were low in most cases (data listed in suppl. Table 1). Figure 2 displays the observed and expected values for each mix (note that "Algae" was always used as a negative control and not added to the mix). This shows that the different components were indeed always identified, although not always at their expected concentration. The Nematode and the Tardigrade showed consistent downward biases, which are compensated by slight upward biases for the other components. The reasons for these biases are not yet clear. Still, the results demonstrate that the algorithm works in principle, although systematic under- or over-estimations of the relative concentration of the component might occur (see discussion).
Figure 1 Sequence alignments for the seven taxa used in this study covering the region that is probed by the pyrosequencing procedure with 60 dispensation steps (dispensation order: A-T-G-C). The underlined part represents the primer that was used for the sequencing reaction. The length of the sequence recorded by the pyrosequencing procedure depends on the exact order of the nucleotides and the order of the dispensation steps. Hence, it is slightly different for the different sequences.
Figure 2 Results from the mixture experiments. Observed and expected values are plotted for each mixture. The observed values are averages from four replicates, each evaluated with four replicates of the library of profiles. The actual values of the replicates, as well as the standard deviations are listed in supplementary Table 1.
Simulation of dideoxy sequencing
Because the pyrosequencing procedure is more restricted with respect to the number of nucleotides that can be sequenced than dideoxy sequencing, it could be useful to adapt the procedure to dideoxy sequencing as well. However, the capillary sequencers that are nowadays used for resolving the DNA fragments are highly tuned towards the specific task of sequence determination, which interferes with the requirement of peak synchronization for the parallel sequencing procedure. Since this is a technical problem that can not be easily solved without interfering with the basic function of the respective instruments, we have simulated the results of dideoxy sequencing, to test the parallel sequencing procedure and to assess at the same time the influence of noise in the system.
Sequencing profiles were simulated with variations of position specific peak heights (see Methods). From these profiles we generated mixtures of one hundred sequences, with concentration differences of three orders of magnitude. A typical abundance profile is shown in Figure 3 together with the simulated composite sequence profile. In the absence of noise, this profile can be unequivocally solved for each component, based on 200 sequenced positions. However, real experiments would include experimental noise. We have simulated this noise at three different levels, 1%, 5% and 10% noise in each of the peaks of the mixture. Note that the noise level is determined by the ratio of the noise to the highest peak of the electropherogram (see Methods). By calculating detection limits defined as three times standard deviation of the solution for the negative control, one can assess the level at which faithful determination is still possible. We found that for the 1% noise level, components present at approximately 0.06% or lower would not be faithfully recovered. This threshold would rise to about 0.3% at the 5% noise level and about 0.6% at the 10% noise level (Figure 3). In our experience, the actual noise levels in peak determinations for identical sequences are currently about 5% on an ABI capillary sequencer. Thus, with the currently available procedures, one would already be able to resolve a dynamic range of 30-fold concentration differences in hundreds of components.
Figure 3 (A) Distribution of the simulated abundance profile of the components in the mix used for assessing the influence of noise in Figure 3. The concentrations were randomly assigned to 99 samples. The 0th sample had concentration 0 as a negative control. (B) Section of a simulated peak profile of the mix with the abscissa depicted as a time line measured in scan numbers and the ordinate in arbitrary intensity units.
Discussion
Our results show that although the experimental procedures will have to be optimized, it is evident that parallel sequencing can in principle be applied to determine the components of mixtures of nucleic acids. The approach will have a particular power for applications where routine re-determination of a limited number of sequence components is required. In such cases it will be possible to experimentally determine correction factors for cases where systematic under- or over-representations of components are observed, as it was the case for two of the components in our experiments. Although we do not yet know the reasons for these deviations, it appears that they are highly reproducible, implying that correction factors would solve this problem.
The most immediate application for our procedure would be in the field of DNA taxonomy or DNA barcoding. There are currently ongoing efforts to obtain DNA barcodes for a large diversity of organisms, including those from soil [1] or benthos [5] samples. These samples include a mixture of small organisms that might be suitable indicators for ecological quality, i.e. a routine determination may become relevant for environmental monitoring. The use of ribosomal rRNA signature sequences would be particularly relevant for such samples, since the rRNA molecules occur in high concentrations in each cell, which could potentially allow using them directly as templates for the sequencing reaction. Furthermore, since the pyrosequencing procedure does not require resolving the reaction products by electrophoresis, it would seem feasible that relatively simple instruments can be constructed to determine the composite sequence profiles, even under outdoor conditions.
However, our procedure may also be useful for medical diagnostic approaches, which aim at routine determinations of limited numbers of components. It seems possible that certain diseases or cancers are characterized by the misexpression of a small set of genes. There are ongoing efforts to develop dedicated microarray assays for diagnosing these on a routine basis. Instead, one could envision using the parallel sequencing procedure to achieve the same goal. In this case one would have to preamplify the samples with a set of primers that are specific for the respective genes and which carry a universal priming site for the sequencing primer. This mixture could then be directly sequenced and the relative concentrations of the components would a measure of the level of expression in the original sample. The parallel sequencing procedure could thus at least partially substitute diagnostic microarrays in the future.
The pyrosequencing procedure that we have employed here can produce only a limited number of reliable sequencing steps, i.e. only a small number of components could be experimentally resolved. However, optimized procedures [9] can yield reading lengths of up to 200 nt, which would expand the applicability to more complex mixtures. Even longer sequence reads can be achieved with the dideoxy sequencing procedure on capillary sequencers. With these, one should be able to assess in the order of hundreds of components in parallel, depending on the level of experimental noise. The greatest problem for implementing the parallel sequencing procedure directly to currently available sequencers is the lack of synchronization between the peaks in different capillaries, i.e. the profile libraries can not be easily matched with the experimental runs which provide a complex picture of composite peaks. However, this problem may be solved by including size standards in each capillary and adjusting the recording software for peaks to these standards, a task that has to be solved in an instrument specific manner.
Conclusion
The parallel sequencing procedure has the capacity to substitute many applications where dedicated microarrays would currently be the only solution. Although further experimental optimizations will be required, these are expected to be solvable in principle. The immediate applicability lies in the field of DNA barcoding and DNA taxonomy, but applications in the field of medical diagnostics would also seem feasible.
Methods
Templates and pyrosequencing
The ribosomal rRNA templates that were used for the implementation of the pyrosequencing procedure were derived from a project where the D3–D5 expansion segment region of the LSU was cloned from organisms that are present in meiobenthos samples [5,8]. The seven sequences that were chosen represent an algae (A), a nematode (N), a tardigrade (T), three crustaceans (C = cyclops, H = harpacticoid, O = ostracod) and an insect (E = ephemeroptera). The fragments were amplified from the clones using primers designed with the software PROBE [10]: forward primer 5'-GAC-CCG-TCT-TGA-AAC-ACG-G-3' and a biotinylated reverse primer 5'-ATC-GAT-TTG-CAC-GTC-AGA-A-3'. Pyrosequencing was performed with 5'-GAA-ACA-CGG-ACC-AAG-GAG-T-3' as sequencing primer according to the instructions of the supplier (Biotage, Uppsala) and on the PSQ96 MA instrument (Biotage). The pyrograms for sixty dispensation cycles were recorded and exported into Excel (Microsoft Inc.) for the further calculation steps.
Deconvolution of the mixture for the pyrosequencing procedure
Each component of the mixture contributes to the final pattern recorded for this mixture. Determination of the pattern contributions, i.e. quantification of the components in the mix, can be achieved by solving a system of linear equations. In theory the following system of equations describes the pyrosequencing process:
{∑i=1Nkji(Z)nji(Z)xi=Sj(Z)...
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaGabaabaeqabaWaaabCaeaacqWGRbWAdaWgaaWcbaGaemOAaOMaemyAaKgabeaakiabcIcaOiabdQfaAjabcMcaPiabd6gaUnaaBaaaleaacqWGQbGAcqWGPbqAaeqaaOWaaeWaaeaacqWGAbGwaiaawIcacaGLPaaacqWG4baEdaWgaaWcbaGaemyAaKgabeaakiabg2da9iabdofatnaaBaaaleaacqWGQbGAaeqaaOWaaeWaaeaacqWGAbGwaiaawIcacaGLPaaaaSqaaiabdMgaPjabg2da9iabigdaXaqaaiabd6eaobqdcqGHris5aaGcbaGaeiOla4IaeiOla4IaeiOla4caaiaawUhaaaaa@4F2A@
Sj – is the peak intensity at j-th step of a specified nucleotide; kji(Z) – the linear coefficient between signal intensity and incorporation event of a specified nucleotide Z (A,T,G,C) at the j-th step of sequencing for the i-th organism; nji(Z) – number of available incorporation events for the nucleotide at the j-th step for the i-th organism (0,1,2,3...) since polynucleotide sequence can have successive repetitions of the same nucleotide several times; xi – the sought concentration of the i-th organism; N – total number of organisms.
In practice the coefficients kji(Z) are unknown, hence it is necessary to record pyrosequencing profiles for each expected component prior to solving the linear system. Pre-recorded profiles represent a set of kji(Z)nji(Z) which is then used to deconvolute a mix. Needless to say that the dispensation order of dNTPs determines pyrograms, therefore it must be consistent throughout individual components and the sample (in our case: A-T-G-C).
In a matrix form the system can be re-written as:
N·X = S
where N – matrix of nji multiplied by kji(Z), actually the pre-recorded profiles; X vector of xi; S – vector of peak intensities. This system can analytically be solved by the "least squares" solution which minimizes the square of the norm of the residual difference [11]:
X = (NT·N)-1·NT·S
To obtain an estimator for the standard deviations of the solutions one has to assume that the values in the matrices N and S, being physical measurements of light intensity, are distributed normally. This assumption allows to calculate errors associated with each solution using the following procedure (according to[11]): a non-scaled covariance matrix for the vector of solutions (X) can be computed as:
C = (NT·N)-1
This matrix needs to be scaled by a factor that can be determined as follows:
sf=(NX−S)T⋅(NX−S)r−p
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGZbWCcqWGMbGzcqGH9aqpdaWcaaqaamaabmaabaGaeCOta4KaeCiwaGLaeyOeI0IaeC4uamfacaGLOaGaayzkaaWaaWbaaSqabeaacqWHubavaaGccqGHflY1daqadaqaaiabh6eaojabhIfayjabgkHiTiabhofatbGaayjkaiaawMcaaaqaaiabdkhaYjabgkHiTiabdchaWbaaaaa@441F@
where X is the solution of the above equation, r – number of rows and p – number of columns in N respectively [12].
The diagonal elements of the scaled covariance matrix are variances (squared standard deviations) of each solution in the vector X. Therefore, these diagonal elements can be used as a measure of an error associated with each solution. Note that values in the covariance matrix as well as the solution are only meaningful if there is a good correlation between S and NX, where X is the computed solution. Correlating S and NX allows determining how well the pyrogram of the unknown sample can be explained in terms of the pyrograms of individual components. If the correlation is poor, the design matrix N is not adequate to the sample (i.e. there are too many unknown RNAs in the sample) and the solution as well as covariance matrix are meaningless.
The number of steps required for an unambiguous solution must be at least as many as the number of components to be identified plus possible additional steps to provide non-singularity of the matrix N. Any further additional steps provide an overdefinition of the system that makes its resolution more robust. For a given set of anticipated components it is possible to determine a minimal number of sequencing steps beforehand assuming all kji(Z) = 1 and nji(Z) taken from the sequences of these components. A simple algorithm simulating pyrosequencing evaluates singularity of the matrix N at each step until N is no more singular. The solution of the linear equations was carried out with the MathCad (Math Soft Inc.).
Implementation for dideoxy sequencing
A mixture electropherogram from a capillary sequencer can be represented as follows:
y(t)=∑i=0N−1aifi(t)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWG5bqEcqGGOaakcqWG0baDcqGGPaqkcqGH9aqpdaaeWbqaaiabdggaHnaaBaaaleaacqWGPbqAaeqaaOGaemOzay2aaSbaaSqaaiabdMgaPbqabaGccqGGOaakcqWG0baDcqGGPaqkaSqaaiabdMgaPjabg2da9iabicdaWaqaaiabd6eaojabgkHiTiabigdaXaqdcqGHris5aaaa@43CA@
where fi(t) is the electropherogram of i-th component, t – the scan number or time after the start of electrophoresis, ai – the quantity of the i-th component and N – number of components. Thus, the problem can be converged to finding ai, which is the best linear fit given that fi(t) is known beforehand – again from the set of pre-recorded electropherograms of expected components. One solves a system of linear equations:
FA = Y
where A is a {ai}T vector, F – matrix, which k-th row is {fi(tk)} and Y – vector {y(tk)}T. The index k runs from 0 to at least N-1. The tk is a subset of scans from the entire electropherogram. The solution is found as shown above
The functions {fi(t)} are determined by sequencing the individual components and storing the electropherograms in the profile library.
Simulation of Dideoxy sequencing results
Gauss-shaped peaks were simulated for random sequences, with unequal assignment of peak height at each position to mimic the known differential incorporation effects. A component library was build from these simulated electropherograms. Concentrations of each component of the library were chosen to be 1.
A mixture was composed from randomized amounts of components adding up to 1. This way of mixture simulation reflects a realistic case where the total amount of DNA or RNA of the mix would be the same as that used to record the library. The distribution of the components was chosen to represent abundant and rare components (compare Fig. 3). The system was solved as described above. Solutions represent fractions of each component of the library in the sample mixture. To estimate the influence of noise, normally distributed random numbers were used to change each peak in the simulated electropherogram of the sample and the library at a specified noise level. The noise level was determined as a fraction of the maximal peak of an electropherogram.
Authors' contributions
A.P. has developed the algorithm and conducted the simulation experiments, as well as initial experimental test. K.S. has done the pyrosequencing experiments. D.T. has initiated and devised the approach and has written the manuscript together with A.P.
Figure 4 Simulation of the effect of noise on the recovery of the correct concentrations of the components in a mix of 99 samples with different concentrations. The distribution of the concentrations is shown in Figure 3. From top to bottom: 1, 5 and 10 % noise level. The graphs represent a direct comparison of given (X axis) and found (Y axis) values. A solid line indicates detection limit calculated as 3 times standard deviation of the negative control. Solutions below this line are not reliable.
Supplementary Material
Additional File 1
Supplemetary Table 1 lists the individual values for the pyrosequencing deconvolution experiment
Click here for file
Acknowledgements
This work was supported by funds of the ministry of science and research of Nord-Rhein-Westfalen, as well as funds from the Verband der Chemischen Industrie.
==== Refs
Floyd R Abebe E Papert A Blaxter M Molecular barcodes for soil nematode identification Mol Ecol 2002 11 839 850 11972769 10.1046/j.1365-294X.2002.01485.x
Tautz D Arctander P Minelli A Thomas RH Vogler AP A plea for DNA taxonomy Trends in Ecol Evol 2003 18 70 74 10.1016/S0169-5347(02)00041-1
Hebert PD Cywinska A Ball SL deWaard JR Biological identifications through DNA barcodes Proc R Soc Lond B Biol Sci 2003 270 313 321 10.1098/rspb.2002.2218
Blaxter ML The promise of a DNA taxonomy Philos Trans R Soc Lond B Biol Sci 2004 359 669 679 15253352 10.1098/rstb.2003.1447
Markmann M Tautz D Reverse taxonomy: an approach towards determining the diversity of meiobenthic organisms based on ribosomal RNA signature sequences Phil Trans Roy Soc 2005 360 1917 1924 10.1098/rstb.2005.1723
Ronaghi M Uhlen M Nyrén P A sequencing method based on real-time pyrophosphate Science 1998 281 363 365 9705713 10.1126/science.281.5375.363
Sanger F Nicklen S Coulson AR DNA sequencing with chain terminating inhibitors Proc Natl Acad Sci USA 1977 74 5463 5467 271968
Markmann M Entwicklung und Anwendung einer 28S rDNA Sequenzdatenbank zur Aufschlüsselung der Artenvielfalt limnischer Meiobenthosfauna im Hinblick auf den Einsatz moderner Chiptechnologie PhD thesis 2000 University of Munich
Gharizadeh B Nordström T Ahmadian A Ronaghi M Nyrén P Long-Read Pyrosequencing Using Pure 2'-Deoxyadenosine-5'-O'-(1-thiotriphosphate) Sp-Isomer Anal Biochem 2002 301 82 90 11811970 10.1006/abio.2001.5494
Pozhitkov AE Tautz D An algorithm and program for finding sequence specific oligonucleotide probes for species identification BMC Bioinformatics 2002 3 9 11882251 10.1186/1471-2105-3-9
Teukolsky SA Vetterling WT Flannery BP Numerical Recipes in C. The art of Scientific Computing 1992 W. H. Press
Lawson CL Hanson RJ Solving Least Squares Problems 1974 Prentice-Hall, Inc.
|
16316462
|
PMC1318471
|
CC BY
|
2021-01-04 16:27:47
|
no
|
BMC Bioinformatics. 2005 Nov 29; 6:281
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-281
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1711632115610.1186/1471-2164-6-171Research ArticleAn EST-based approach for identifying genes expressed in the intestine and gills of pre-smolt Atlantic salmon (Salmo salar) Hagen-Larsen Heidi [email protected] Jon K [email protected] Frank [email protected] Alexei [email protected]øyheim Bjørn [email protected] Norwegian School of Veterinary Science, Department of Basic Sciences and Aquatic Medicine. PO Box 8146 Dep., NO-0033 Oslo, Norway2 Biotechnology Centre of Oslo (BIO), University of Oslo, PO Box 1125 Blindern, 0317 Oslo, Norway3 Danish Institute of Agricultural Sciences, Department of Animal Breeding and Genetics, PO Box 50, DK-8830 Tjele, Denmark4 Rikshospitalet University Hospital, Department of Dermatology, NO-0027 Oslo, Norway2005 1 12 2005 6 171 171 15 6 2005 1 12 2005 Copyright © 2005 Hagen-Larsen et al; licensee BioMed Central Ltd.2005Hagen-Larsen 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 Atlantic salmon is an important aquaculture species and a very interesting species biologically, since it spawns in fresh water and develops through several stages before becoming a smolt, the stage at which it migrates to the sea to feed. The dramatic change of habitat requires physiological, morphological and behavioural changes to prepare the salmon for its new environment. These changes are called the parr-smolt transformation or smoltification, and pre-adapt the salmon for survival and growth in the marine environment. The development of hypo-osmotic regulatory ability plays an important part in facilitating the transition from rivers to the sea. The physiological mechanisms behind the developmental changes are largely unknown. An understanding of the transformation process will be vital to the future of the aquaculture industry. A knowledge of which genes are expressed prior to the smoltification process is an important basis for further studies.
Results
In all, 2974 unique sequences, consisting of 779 contigs and 2195 singlets, were generated for Atlantic salmon from two cDNA libraries constructed from the gills and the intestine, accession numbers [Genbank: CK877169-CK879929, CK884015-CK886537 and CN181112-CN181464]. Nearly 50% of the sequences were assigned putative functions because they showed similarity to known genes, mostly from other species, in one or more of the databases used. The Swiss-Prot database returned significant hits for 1005 sequences. These could be assigned predicted gene products, and 967 were annotated using Gene Ontology (GO) terms for molecular function, biological process and/or cellular component, employing an annotation transfer procedure.
Conclusion
This paper describes the construction of two cDNA libraries from pre-smolt Atlantic salmon (Salmo salar) and the subsequent EST sequencing, clustering and assigning of putative function to 1005 genes expressed in the gills and/or intestine.
==== Body
Background
The number of known genes in Atlantic salmon (Salmo salar) is limited compared with the very large numbers known in mammals. However, the salmon is so important as an aquaculture species that a lot of effort has been put into acquiring knowledge of its genome compared to many of the other aquatic species. Although only 14 full-length genes are listed for Atlantic salmon at NCBI [1], there are 111 457-nucleotide sequences (September 2005), many containing complete cDNA sequences. The genome and biology of the Atlantic salmon are complex and the genome has not been well characterised. The salmonidae species underwent a genome duplication event about 25–100 million years ago [2,3]. This event resulted in a tetraploid genome, but since then the salmon has been gradually returning to the diploid condition. This process is still continuing, and the salmon now has a partially tetraploid genome. Although many genes have been lost after the duplication event [4], the fact that regions of the genome are duplicated and almost identical does complicate the hunt for genes in this species. It means that there may be four almost identical copies of a gene in the genome, but it is an open question whether all four copies are active. Furthermore, the Atlantic salmon goes through a process known as smoltification as a step in its maturation. The salmon spends the first part of its life in fresh water before migrating to the ocean, where it lives until returning to rivers to spawn. Smoltification [5] involves synchronised morphological, physiological and behavioural changes that enable the young salmon (parr) to survive in the ocean and to grow and migrate normally. Since smoltification occurs while the fish are still in fresh water, they are pre-adapted to the marine environment. Given this enormous change in both fish biology and environment, it seems reasonable to suggest that the pre-smolt expresses a different set of genes from the smolt. It is therefore important to identify the genes that are expressed both before and after the smoltification process and at each of the different developmental stages in order to obtain a complete set of genes expressed in the salmon genome. One strategy for identifying the genes expressed in specific life stages and tissues is to use expressed sequence tags or ESTs. These are short stretches of single pass sequences obtained from sequencing cDNA [6,7], and are widely used for gene discovery, mapping, polymorphism analysis, expression studies and gene prediction. Gene discovery methods using ESTs include hunting for new members of gene families in the same species (paralogues), for functionally equivalent genes in other species (orthologues), or even for alternatively spliced forms of known genes. ESTs are also used to predict or refine computational predictions of the location of genes in genomic DNA. Recently, a lot of effort has been put into acquiring ESTs from Atlantic salmon, and over 108,242 EST sequences (September 2005), mainly from smoltified fish, have been deposited in GenBank [8-11]The aim of the present study was to increase the number of ESTs available from pre-smolts in order to identify genes that are expressed in the early life stages of Atlantic salmon. We have focused on the gills and the intestine.
Results
One of the goals was to accumulate cDNA libraries with large insert sizes. The average insert size was determined from 192 clones from each library fraction after digesting the clones with EcoRI/XhoI. The approximate size distribution for the libraries is given in table 1.
Table 1 Fraction sizes and number of sequences for the gills and intestine cDNA libraries
Library fraction Mean insert sizeb Clones seq.c Good seq.c Unique seq.c Annotated seq.c
Gills 5 1,500 bp
Gills 6 900 bp
Gills 7 700 bp 4128 3015
Intestine 4 1,300 bp
Intestine 5 1,050 bp
Intestine 6 730 bp 4128 3247
Total 2974 1005
a The numbers after the tissue type indicate when the cDNA was collected from the column, the lowest numbers being eluted first. The mean insert sizes are given in the second column
b Shown in base pairs
c Seq. is an abbreviation of sequence
A total of 4128 clones from each library were sequenced from the 5'direction. After basecalling and trimming of vector, contaminants and poor quality sequences, the number of sequences was reduced to 3015 for the gills and 3247 for the intestine. The trimmed sequences were clustered and assembled. After the combined clustering of the gill and intestine sequences, a total of 2974 unique sequences were left, of which 2195 were singlets and 779 were consensus sequences from contigs. Of these, 1491 showed similarity to known sequences after they were run through a combined blastx and blastn pipeline. A summary of the BLAST results is given in Table 2. All the sequence alignment results are accessible at the Salmon Genome Project website [11] in the libraries section (Data and Results > cDNA libraries > Gills, Intestine). All trimmed EST sequences have been submitted to GenBank. The accession numbers are [Genbank: CK877169-CK879929, CK884015-CK886537 and CN181112-CN181464]. The annotation of the clustered ESTs using GO terms is accessible through links at the Salmon Genome Project website [11] in the annotation section (Data and Results > Annotations > SGP Intestine-Gills). A shorter version is given in Table 3 for molecular function.
Table 2 Number of hits observed in the different databases
Database Hitsc %d
PDBa 489 16
SwissProtb 1005 34
nrb 1276 43
ntb 961 32
no hits 1483 50
a Threshold for a significant hit for blastx against PDB was defined as expectation value <10-10
b Threshold for a significant hit for blastx against SwissProt and nr, and blastn against nr(= nt) was defined as expectation value <10–15
c The number of significant hits observed in each databases
d The number of significant hits observed given as a percentage
Table 3 Gene Ontologya(Molecular function) of the sequences with a significant blastx hit in SwissProt
Molecular Functionb Nr of Hits
%antioxidant activity 10
%glutathione-disulfide reductase activity 2
%peroxidase activity 5
%binding 505
%antigen binding 1
%carbohydrate binding 5
%cofactor binding 1
%drug binding 1
%glycosaminoglycan binding 12
%isoprenoid binding 1
%lipid binding 23
%metal ion binding 103
%nucleic acid binding 208
%nucleotide binding 132
%peptide binding 3
%protein binding 112
%pyridoxal phosphate binding 1
%receptor binding 17
%selenium binding 1
%steroid binding 2
%tetrapyrrole binding 2
%vitamin binding 1
%catalytic activity 358
%helicase activity 15
%hydrolase activity 162
%integrase activity 1
%isomerase activity 22
%kinase activity 28
%ligase activity 22
%lyase activity 11
%oxidoreductase activity 88
%small protein activating enzyme activity 2
%small protein conjugating enzyme activity 8
%transferase activity 59
%transposase activity 2
%chaperone activity 23
%heat shock protein activity 7
%enzyme regulator activity 30
%caspase regulator activity 1
%enzyme activator activity 10
%enzyme inhibitor activity 16
%GTPase regulator activity 9
%kinase regulator activity 3
%nitric-oxide synthase regulator activity 1
%molecular_function unknown 29
%motor activity 11
%microtubule motor activity 2
%signal transducer activity 54
%receptor activity 26
%receptor binding 17
%receptor signaling protein activity 4
%structural molecule activity 160
%extracellular matrix structural constituent 2
%structural constituent of bone 1
%structural constituent of cytoskeleton 10
%structural constituent of eye lens 2
%structural constituent of muscle 2
%structural constituent of ribosome 103
%transcription regulator activity 42
%RNA polymerase I transcription factor activity 1
%RNA polymerase II transcription factor activity 5
%transcription cofactor activity 11
%transcription factor activity 24
%transcriptional repressor activity 3
%translation regulator activity 21
%translation factor activity, nucleic acid binding 21
%transporter activity 156
%amine/polyamine transporter activity 2
%auxiliary transport protein activity 1
%carbohydrate transporter activity 1
%carrier activity 66
%channel/pore class transporter activity 8
%drug transporter activity 1
%electron transporter activity 34
%intracellular transporter activity 3
%ion transporter activity 58
%lipid transporter activity 5
%neurotransmitter transporter activity 2
%organic acid transporter activity 3
%oxygen transporter activity 6
%peptide transporter activity 2
%protein transporter activity 33
a A selection of GO categories is presented. All annotations are accessible at
b Indented terms are 'children' of the above 'parent' term.
Table 3 shows that the sequences can be classified in 12 of the major categories for molecular function, the two largest groups being catalytic activity (26%) and binding (36%).
Discussion
The total number of gill and intestine clones available was 8256. Table 1 shows the sampled means of the insert sizes after fractionation, which varied between 700 and 1500 base pairs. cDNA cloning rarely produces full-length gene products but rather products containing varying proportions of the whole sequence. 5'sequencing may overcome this problem since clustering can be used to cover larger parts of the transcripts. In this study, after trimming and quality control followed by clustering, approximately 65% of the trimmed sequences from both libraries ended up in contigs.
Sequence annotation was carried out using GO terms. A fairly strict criterion of E-value below 1.0 × 10-15 for a significant hit was used both for Blastx run against the Swiss-Prot and nr protein databases, and blastn against the nr nucleotide database. This rather low value was chosen to reduce the number of hits on very remote homologous genes with functions unrelated or only remotely related to the sequences in the current libraries. For blastx against the PDB database, the significance level chosen was an E-value below 1.0 × 10-10, and 489 or 16% of the sequences gave significant hits in this database. Almost all of these showed sequence similarity above 30%, indicating that they might be suitable targets for homology modelling of protein 3D structure if the hit produces a biologically meaningful alignment.
Approximately one third of the sequences gave hits when blastx was run against the Swiss-Prot database, and were annotated by annotation transfer from GO-annotated UniProt sequences as described in section 2.6. Since we used a fairly strict criterion for a significant hit, we would argue that the derived GO annotation is meaningful for a large fraction of these 1005 sequences. We plan to use a more elaborate annotation pipeline for future work, including estimates of accuracy and sensitivity and making it possible to detect more remote homologous sequences, for example by profile-based sequence similarity searches. Annotating the sequences and collecting the links to all annotated sequences in tables that correspond to molecular function, biological process, and cellular component GO terms makes it possible in future studies to extract genes of interest merely by looking at the GO term tables from the links at the Salmon Genome Project website [11] in the Annotation section (Data and Results > Annotations > SGP Intestine-Gills). Figure 1 illustrates this. A researcher interested in immunology would simply go to the biological process GO term table and search for 'immune response', and could then follow each of the 25 links to the sequences that appear to have a connection with immunology and their sequence alignment results. This will make it possible to find possibly relevant sequences much more directly, without having to browse through thousands of sequences and their associated sequence alignment data.
Figure 1 Gene Ontology Annotation. The annotated sequences can be reached through three documents corresponding to molecular function, biological process, and cellular component GO terms. There are links to all results for a given GO term -"immune response" is used in the figure – that can be followed to the sequence information and alignment summary for each sequence. The main tables give links to the fasta file containing the sequence itself and the BLAST results for each sequence. The most significant sequence alignment results are also given with various details such as expectation value and length of sequence alignment hit. See the web site for further details.
In the two libraries, we discovered several genes that may be involved in the smoltification process, genes involved in cell homeostasis and genes coding for hormones that may influence salmon maturation. Certain genes were of particular interest, for example the glucagon-family neuropeptide precursor, tyrosine 3/ tryptophan 5-monooxygenase activities and thyroid receptor interactin protein 6 (TRIP6). In addition, five different genes related to growth factor activity were also present: granulin, neuregulin 1, syntenin, Bmp2 and pleiotrohin 1. All of the above-mentioned genes are thought to play a part in the smoltification process in salmon, along with the osmoregulatory genes. However, it is possible that some of the more interesting genes involved in the smoltification process have not yet been annotated and characterised and are therefore listed as unknown. Approximately 50% of the sequences gave no significant hits in the blastx-blastn pipeline. A large proportion of these gave hits with higher E-values to more remote homologous sequences, but a significant number, at least 500–1000 sequences, are of completely unknown function at present. However, the 5' UTR of fish may differ significantly from that of mammals and terrestrial organisms. This means that the failure to match 500–1000 sequences does not confirm that the genes are unique to fish or markedly different from the coding region of transcripts of other organisms. A number of factors, including the 5'UTR, may help to explain why no similarity was found for these sequences, but the unusual biology of salmon and the fact that the sequences are from pre-smolt fish may be contributory factors. Pre-smolt salmon live in fresh water, and presumably different sets of genes are turned on and off in response to the two different habitats, the freshwater and the marine environment. In addition, the lack of similarity to already known sequences in other vertebrates could influence these results. Although large amounts of sequence data from other species have been generated, a large fraction of the genes in higher organisms is still uncharacterised. It therefore seems likely that the lack of similarity of these ESTs to known genes is partly explained by the fact that many genes are still uncharacterised.
Conclusion
This work has made available approximately 3000 sequences from pre-smolt salmon, one third of which have been annotated for function, biological process and cellular component. Several hundred sequences code for proteins that show enough sequence-similarity to proteins of known structure to be suitable candidates for homology modelling. The results of this study will be a valuable resource for future studies of Atlantic salmon biology since it is now possible to search for genes sequenced in salmon rather than having to use comparative data from other species.
Methods
1. Construction of cDNA libraries
Pre-smolt Atlantic salmon (Salmo salar) from a commercial aquaculture-bred stock (National Norwegian Breeding Programme for salmon, Norway), weighing 50–100 g, were collected in late September for the construction of two cDNA libraries from the intestine and gills. The entire intestine and the whole of the gills were used to construct the libraries. Tissue samples of 400 mg were used for total RNA extraction, which was performed using the guanidine thiocyanate-phenol-chloroform extraction method [12]. mRNA was further isolated using the Poly (A) Quick® mRNA isolation kit (Stratagene Cloning Systems, California, USA). Three μg of total RNA and 0.1 μg of mRNA were run on an agarose gel to investigate the integrity/quality of the samples. cDNA synthesis was performed using the pBluescript® II XR cDNA library construction kit (Stratagene Cloning Systems, California, USA), which generates directional cDNA libraries using an oligo (dT)18 primer. For each library the cDNA was size fractionated using a gel filtration column, and three fractions containing most of the cDNA were ligated into the plasmid vector and subsequently transformed into XL10 Gold ultra competent cells.
2. Pre-screening of the cDNA libraries
Pre-screening was performed to separate abundant genes from rare ones. From each of the three cDNA fractions from both tissues, approximately 50 000 colonies were plated out on agar plates (140 mm in diameter), with an average of 2000 colonies per plate. The colonies were transferred to 132 mm BIOTRANS nylon membranes, pore size 1.2 Micron, before hybridisation to total cDNA. Total cDNA from the intestine was used to screen the intestine filters and total cDNA from the gills to screen the gill filters. The cDNA was labelled using α32P-dCTP in a PCR assay. In order not to label the vector, the plasmids containing the cDNA were digested with XhoI/EcoRI to open up the circle. Then M13–20 primer was used in a 'one way' PCR to label the insert. The colonies that hybridised were presumed to be the more abundant ones in the libraries. A total of 3840 non-hybridising clones (the presumed rare genes) and 288 hybridising clones were picked from each of the gill and intestine libraries and gridded in duplicate into 384 well microtiterplates. The plates were grown overnight in LB, then glycerol was added and the plates were sealed and stored at -80°C.
3. Enrichment of pBluescript cDNA libraries
The remaining cDNA was amplified in semi-solid fashion, as described by the manufacturer (Stratagene Cloning Systems, California, USA). Amplification in suspension allowed three-dimensional, uniform growth, which reduced the likelihood of under-representation of particular clones. Up to 5 × 105 cfu/bottle of primary library was added to each bottle of the semi-solid suspension, before incubation on an ice-water bath for an hour to achieve the required semi-solid state of the medium, followed by incubation for 40–45 hours at 30°C. Afterwards the contents of the bottles were spun down and resuspended in 50 ml of 2xLB-glycerol (12.5%) for the gill library and 100 ml for the intestine library, ending up with 8 × 109 clones from the gills and 3.1 × 1010 clones from the intestine. Total volumes of 1 ml were aliquoted into tubes and subsequently frozen at -80°C for future use.
4. DNA isolation procedures
A few microlitres of the pre-screened cDNA library was cultured overnight in LB before the DNA was isolated using the QIAprep 96 Turbo Miniprep kit from QIAGEN in a QIAvac96 according to the manufacturer's manual.
5. Size determination of the library fractions
To determine the insert sizes of the clones from the different size fractions, a rough estimate was made for 192 randomly picked clones from each fraction. The insert size was determined by digesting the insert out of the vector using the restriction enzymes EcoRI and XhoI, followed by separation on an agarose gel. The gels were scanned on a Typhoon 9410 Imager and analysed using the ImageQuant TL software, both from Amersham Biosciences.
6. DNA sequencing
Sequencing was done from the 5'-end using T3 as sequencing primer. 5'-sequencing was chosen in order to assign functional annotation to as many transcripts as possible. The sequencing reactions were performed using the ABI PRISM® BigDye™ Terminators Cycle Sequencing Kit (Applied Biosystems), and run on the ABI 377 (Applied Biosystems) or on the Megabace 1000 (Amersham Pharmacia).
7. Clustering, sequence comparison and annotation transfer
Phred [13,14] and cross_match (Green P: unpublished) were used for basecalling and trimming of vector, respectively. The sequences were masked for repeats against a Danio rerio repeat library [15], and sequences containing contaminants or that were of poor quality were removed. The remaining high quality sequences were assembled and clustered with the Sequencher 4.1 package. In order to examine sequence similarity to known genes, NCBI blast sequence alignment [16] was performed against the NCBI databases [17] (on 08.03.2004). For each sequence, blastx was run against the PDB, Swiss-Prot and nr protein sequence databases, and blastn was run against the nr nucleotide sequence database. We defined a significant database hit as having an expectation value (E-value) below 1.0 × 10-15 for all sequence alignments, except for blastx against PDB where 1.0 × 10-10 was used. All sequences that gave a significant blastx hit in Swiss-Prot were annotated by annotation transfer, applying the Gene Ontology (GO) [18,19] assignments for the UniProt database produced by the GOA project of the European Bioinformatics Institute[20]. The gene_association.goa_uniprot database of 26.04.2004 was used [21] together with GO terms from the GO release of 11.05.2004 [22]. The sequences were annotated on the basis of the single best hit in the Swiss-Prot database.
Authors' contributions
HHL performed the experiments and drafted the manuscript. JKL participated in the sequence comparisons, carried out the annotation transfer, designed the web pages and drafted the bioinformatics section of the manuscript. FP carried out the sequence alignments on the entire SALGENE sequences. AA participated in the sequence comparisons and provided supervision on the bioinformatics section. BH conceived the study, participated in its design and coordination, provided supervision and helped to draft the manuscript. All authors have read and approved the final manuscript.
Acknowledgements
The authors would like to thank Agate Noer and Kristin Vekterud for their assistance with library construction and sequencing. This work was supported by the EU FAIR Programme SALGENE project (FAIR CT98-4314); and grant 130162/130 as part of a project entitled "Strategic QTL research plan for disease resistance in Atlantic salmon and cattle" and grant 139617/140 as part of a project entitled "Salmon Genome Project (SGP)", both supported by the Research Council of Norway.
==== Refs
NCBI - National Center for Biotechnology Information [http://www.ncbi.nlm.nih.gov/entrez]
Allendorf FWTGH Turner BJ Tetraploidy and the evolution of salmonid fishes Evolutionary genetcs of fishes 1984 New York , Plenum Press 1 53
Ohno S Wolf U Atkin NB Evolution from fish to mammals by gene duplication Hereditas 1968 59 169 187 5662632
Wolfe KH Yesterday's polyploids and the mystery of diploidization Nat Rev Genet 2001 2 333 341 11331899 10.1038/35072009
Boeuf G Rankin JCJFB Salmonid smolting: a pre-adaptation to the oceanic environment Fish Ecophysiology 1993 London , Chapman & Hall pp 105 135
Adams MD Kelley JM Gocayne JD Dubnick M Polymeropoulos MH Xiao H Merril CR Wu A Olde B Moreno RF Complementary DNA sequencing: expressed sequence tags and human genome project Science 1991 252 1651 1656 2047873
Marra MA Hillier L Waterston RH Expressed sequence tags--ESTablishing bridges between genomes Trends Genet 1998 14 4 7 9448457 10.1016/S0168-9525(97)01355-3
Davey GC Caplice NC Martin SA Powell R A survey of genes in the Atlantic salmon (Salmo salar) as identified by expressed sequence tags Gene 2001 263 121 130 11223250 10.1016/S0378-1119(00)00587-4
Martin SA Caplice NC Davey GC Powell R EST-based identification of genes expressed in the liver of adult Atlantic salmon (Salmo salar) Biochem Biophys Res Commun 2002 293 578 585 12054641 10.1016/S0006-291X(02)00263-2
Rise ML von Schalburg KR Brown GD Mawer MA Devlin RH Kuipers N Busby M Beetz-Sargent M Alberto R Gibbs AR Hunt P Shukin R Zeznik JA Nelson C Jones SR Smailus DE Jones SJ Schein JE Marra MA Butterfield YS Stott JM Ng SH Davidson WS Koop BF Development and application of a salmonid EST database and cDNA microarray: data mining and interspecific hybridization characteristics Genome Res 2004 14 478 490 14962987 10.1101/gr.1687304
SGP - Salmon genome project [http://www.salmongenome.no]
Chomczynski P Sacchi N Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction Anal Biochem 1987 162 156 159 2440339 10.1016/0003-2697(87)90021-2
Ewing B Hillier L Wendl MC Green P Base-calling of automated sequencer traces using phred. I. Accuracy assessment Genome Res 1998 8 175 185 9521921
Ewing B Green P Base-calling of automated sequencer traces using phred. II. Error probabilities Genome Res 1998 8 186 194 9521922
Repeat Masker [http://ftp.genomewashington.edu/RM/RepeatMasker.html]
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
NCBI - National Center for Biotechnology Information [ftp://ftp.ncbi.nih.gov/blast/db]
Ashburner M Ball CA Blake JA Botstein D Butler H Cherry JM Davis AP Dolinski K Dwight SS Eppig JT Harris MA Hill DP Issel-Tarver L Kasarskis A Lewis S Matese JC Richardson JE Ringwald M Rubin GM Sherlock G Gene ontology: tool for the unification of biology. The Gene Ontology Consortium Nat Genet 2000 25 25 29 10802651 10.1038/75556
Ashburner M Lewis S On ontologies for biologists: the Gene Ontology--untangling the web Novartis Found Symp 2002 247 66 80; discussion 80-3, 84-90, 244-52 12539950
European Bioinformatics Institute [http://www.ebi.ac.uk/GOA]
European Bioinformatics Institute [ftp://ftp.ebi.ac.uk/pub/databases/GO/goa/UNIPROT]
The Gene Ontology [http://www.geneontology.org]
|
16321156
|
PMC1318472
|
CC BY
|
2021-01-04 16:32:47
|
no
|
BMC Genomics. 2005 Dec 1; 6:171
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-171
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpmedplosmedPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1636391910.1371/journal.pbio.0040007Research ArticleAnimal BehaviorDevelopmentEcologyEvolutionZoologyOtherGeriatricsThe Evolution of Senescence and Post-Reproductive Lifespan in Guppies (Poecilia reticulata)
Evolution of Post-Reproductive LifespanReznick David
1
*Bryant Michael
1
2
Holmes Donna
3
1Department of Biology and Center for Conservation Biology, University of California Riverside, Riverside, California, United States of America2School of Critical Studies, California Institute of the Arts, Valencia, California, United States of America3Department of Biological Sciences, University of Idaho, Moscow, Idaho, United States of AmericaKirkwood Thomas Academic EditorUniversity of Newcastle upon TyneUnited Kingdom* To whom correspondence should be addressed. E-mail: [email protected] 2006 27 12 2005 27 12 2005 4 1 e728 7 2005 31 10 2005 Copyright: © 2006 Reznick et al.2006This 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.
Getting an Evolutionary Handle on Life after Reproduction
The study of post-reproductive lifespan has been of interest primarily with regard to the extended post-menopausal lifespan seen in humans. This unusual feature of human demography has been hypothesized to have evolved because of the “grandmother” effect, or the contributions that post-reproductive females make to the fitness of their children and grandchildren. While some correlative analyses of human populations support this hypothesis, few formal, experimental studies have addressed the evolution of post-reproductive lifespan. As part of an ongoing study of life history evolution in guppies, we compared lifespans of individual guppies derived from populations that differ in their extrinsic mortality rates. Some of these populations co-occur with predators that increase mortality rate, whereas other nearby populations above barrier waterfalls are relatively free from predation. Theory predicts that such differences in extrinsic mortality will select for differences in the age at maturity, allocation of resources to reproduction, and patterns of senescence, including reproductive declines. As part of our evaluation of these predictions, we quantified differences among populations in post-reproductive lifespan. We present here the first formal, comparative study of the evolution of post-reproductive lifespan as a component of the evolution of the entire life history.
Guppies that evolved with predators and that experienced high extrinsic mortality mature at an earlier age but also have longer lifespans. We divided the lifespan into three non-overlapping components: birth to age at first reproduction, age at first reproduction to age at last reproduction (reproductive lifespan), and age at last reproduction to age at death (post-reproductive lifespan). Guppies from high-predation environments live longer because they have a longer reproductive lifespan, which is the component of the life history that can make a direct contribution to individual fitness. We found no differences among populations in post-reproductive lifespan, which is as predicted since there can be no contribution of this segment of the life history to an individual's fitness.
Prior work on the evolution of post-reproductive lifespan has been dominated by speculation and correlative analyses. We show here that this component of the life history is accessible to formal study as part of experiments that quantify the different segments of an individual's life history. Populations of guppies subject to different mortality pressures from predation evolved differences in total lifespan, but not in post-reproductive lifespan. Rather than showing the direct effects of selection characterizing other life-history traits, post-reproductive lifespan in these fish appears to be a random add-on at the end of the life history. These findings support the hypothesis that differences in lifespan evolving in response to selection are confined to the reproductive lifespan, or those segments of the life history that make a direct contribution to fitness. We also show, for the first time, that fish can have reproductive senescence and extended post-reproductive lifespans despite the general observation that they are capable of producing new primary oocytes throughout their lives.
An analysis of the causes of variation in post-reproductive lifespan reveals that fish senesce and that the evolution of lifespan in guppies is due to selection during their reproductive stage.
==== Body
Introduction
The post-reproductive segment of the human lifespan is often considered to be an adaptive feature of aging, since post-reproductive women can make significant contributions to the fitness of their children and grandchildren [1,2]. Although there is published evidence for the adaptive significance of the post-reproductive life stage in humans [3], few formal empirical analyses have been conducted to conclusively demonstrate that this segment of the life history has been shaped by natural selection. Evolutionary theory predicts that an extended post-reproductive lifespan should evolve only when post-reproductive females can contribute significantly to the fitness of their offspring or relatives. Such fitness contributions can occur only when post-reproductive maternal care is required for the last young that mothers produce [4], or when animals have extended social networks allowing post-reproductive females to contribute significantly to the fitness of close relatives, as in humans [5,6]. Where such kin networks exist, any advantage of post-reproductive lifespan must also be sufficient to offset the waning force of natural selection associated with the low probability of females surviving to advanced ages [5,6]. If no such post-reproductive contribution to fitness exists, then any post-reproductive lifespan that we see should represent a variable and random segment of the life history incidental to differences in aging rates of reproductive and somatic tissues.
The phenomenon of post-reproductive lifespan is by no means limited to human females, or even to female primates. Midlife cessation of ovulation, followed by a post-reproductive lifespan of up to a third of the reported species maximum lifespan, has been reliably documented in a wide range of female vertebrates in captivity or under other favorable conditions. These species include guppies (Poecilia reticulata) [7], platyfish [8], Japanese quail [9,10,11], budgerigar “parakeets” [12], laboratory rats and mice [13,14,15], opossums[16], and primates. [17,18,19]. Most of these animals lack well-developed kin networks and engage in very limited, if any, maternal care. In some (e.g., platyfish and Japanese quail), males also exhibit midlife loss of reproductive capacity [8,20,11]. These observations suggest that, rather than being a result of kin selection, midlife fertility loss and post-reproductive lifespan may be correlates of extended lifespan under particularly favorable conditions characterized by good nutrition and unusually low rates of mortality from parasites, predators, disease, and accident.
The strongest evidence for the adaptive significance of post-reproductive lifespan in humans comes from historical, multigenerational demographic records for 18th and 19th century populations in Canada and Finland [3]. In analyses of data of this type, survival of grandmothers has been associated with improved reproductive success in their offspring. These analyses also suggest that this “grandmother effect” requires active behavioral interactions between older females and kin. The effect was detected only when grandmothers lived in the same town as kin receiving the benefit and accrued only by grandchildren that survived past weaning. The positive effect of grandmothering was present even after controlling statistically for potentially confounding factors, such as socioeconomic status, temporal trends in survival, or geography.
Comparable analyses have been performed for other species. Packer et al. [4] report on baboons and lions, both of which live in extended family groups in which grandmothers have extended interactions with kin. In both species, however, females generally live only long enough to care for their last-born offspring, and no positive effect of grandmothers' post-reproductive survival was detected. Female salmon, like lions and baboons, live long enough after spawning to enhance the fitness of their offspring. They prevent the nest site from being reused by late-arriving spawners; such reuse causes a substantial reduction in the viability of the eggs [21].
We have evaluated the pre-reproductive, reproductive, and post-reproductive lifespan in guppies in relation to selective regimes shaping the evolution of their early life histories and patterns of senescence. We compared life-history traits in guppies from high- versus low-predation environments in Trinidad. Guppies from high-predation environments co-occur with predators that frequently prey on guppies, and particularly on large, adult-size classes. Guppies from low-predation environments co-occur with just a single species of fish, Rivulus hartii, that feeds on guppies only occasionally, and tends to feed on small, immature-size classes. These two types of localities are often found in the same drainage in close proximity of one another, separated by barrier waterfalls that exclude larger predators but not guppies or Rivulus.
In nature, guppies from high-predation localities sustain higher mortality rates than their counterparts from low-predation localities [22]. Furthermore, guppies from high-predation localities attain maturity at an earlier age, produce more offspring per litter, and reproduce more often than guppies from low predation localities [23,24], which is the predicted evolutionary response to higher adult mortality rates [6]. In addition, we have manipulated the mortality rates that guppies experience in nature with guppy or predator introduction experiments and have shown that these life history patterns can evolve in relatively brief intervals of time [25,26,27]. We have also documented senescence in natural populations in the form of an acceleration of mortality rate with age [28].
We compared the patterns of aging and senescence in the second laboratory-born generation of guppies from high- versus low-predation environments. We evaluated guppies from the Yarra and Oropuche drainages and included a high- and low-predation environment from each drainage, for a total of four populations. Since genetic data show that guppies adapted to these regimes independently in each locality [29,30], this approach provides duplicate studies of adaptation of guppy populations to the presence or absence of predators. Classical evolutionary senescence theory predicts that animals from high-predation environments will experience earlier or more rapid senescence than those from low-predation environments, either as a byproduct of intense selection for increased investment in reproduction early in life or because of the accumulation of deleterious mutations affecting older individuals [31,32]. We reported previously that the early life histories of high-predation guppies in this experiment were different from those from low-predation localities in the same drainage in a fashion that is consistent with our earlier comparisons [33]: They matured earlier, produced more young per litter, and reproduced more frequently. Contrary to expectation, however, we found that when guppies were reared in the lab free from predation, those from high-predation environments have lower initial mortality rates, lower mortality rates throughout their lives, and longer median lifespans [34]. Furthermore, guppies from high-predation localities have higher fecundity throughout their lives.
Here we consider in more detail why guppies from high-predation environments have longer median lifespans and, specifically, whether the length of post-reproductive lifespan varies in guppies that have evolved under different mortality regimes. Life-history theory predicts that selection will shape only segments of the life history that contribute directly to fitness or are correlated with other life-history variables. Specifically, selection should favor the evolution of an extended post-reproductive lifespan only if post-reproductive individuals can contribute to their inclusive fitness in some way, either by caring for their own young, other kin, or grand-offspring. Since guppies are livebearers and provide no parental care after birth, we predicted that there should be no direct selection for an extended post-reproductive lifespan in this species. Any post-reproductive lifespan we observe, therefore, should represent a by-product of different aging rates of different parts of the soma or, alternatively, a correlate of selection for traits—including reproductive characteristics—that are adaptive earlier in the life history.
Unlike birds and mammals, which produce the vast majority of oocytes before birth, fish are generally believed to produce new oocytes throughout their lives. For this reason, fish have been suggested to show little or no reproductive senescence [35]. The empirical evidence for such unlimited reproductive capacity, however, is based on isolated observations of very old, fertile individuals and a few small-scale laboratory studies of senescence in fish; few comparative studies have directly addressed reproductive aging in fish. Here we formally evaluate, for the first time, whether or not there is reproductive senescence with an extended post-reproductive lifespan in a species of fish.
Results
Life-History Segments
We have previously shown that guppies from high-predation environments in this study are younger at maturity than those from low-predation environments [33], and that these high-predation guppies have lower rates of loss of reproductive function, lower mortality rates throughout their lives, and longer median lifespans [34].
We first present analyses of variance (ANOVA) of the three, non-overlapping segments of the lifespan: age at first reproduction, reproductive lifespan (age at last reproduction minus age at first reproduction), and post-reproductive lifespan (age at death minus age at last reproduction). As we found earlier, guppies from high-predation localities are younger at first reproduction. Their reproductive lifespan is also significantly longer, enough for them to be far older, on average, when they cease reproduction. Furthermore, reproductive lifespan is longer in guppies from the Oropuche drainage than in the Yarra drainage and in guppies reared on low as opposed to high levels of food availability (Tables 1 and 2). Low food caused a much larger increase in the reproductive lifespan of guppies from low- as opposed to high-predation localities, which in turn resulted in a marginally significant interaction between food availability and predation. There were no differences among any treatment groups for the duration of post-reproductive lifespan.
Table 1 Summary of ANOVAs for the Age at First Reproduction and the Duration of the Reproductive Lifespan
Table 2 Least-Square Means (Standard Error) from the ANOVAs in Table 1
Interbrood Interval
The inference that an individual has a significant post-reproductive lifespan is based on an analysis of each individual's frequency of reproduction. For each individual, we asked whether or not it had lived beyond the time interval in which it could be expected to produce another litter of young. Fish usually reproduced at regular intervals of 22–35 days and produced an average of 15 to 28 litters of young during their lives (Figure 1, Table 3). As they aged, some individuals appeared to skip litters or even cease reproducing for an interval of time, then resume reproduction. This irregular cessation of reproduction accounts for the unusually long interbrood intervals, sometimes in excess of 100 d, seen in the figure. Sample lifetime reproductive schedules for three individuals from one treatment group (Figure 1, bottom) illustrate some of the individual variation that we saw in this study. One individual reproduced at regular intervals throughout its life, then died before the age when it could be expected to produce a new litter of young. A second individual reproduced less regularly and appears to have skipped what would have been its fifth and eighth litters, then died a few days after its last litter was born. Neither of these fish showed evidence of having lived past the time when they might have produced an additional litter of young, so neither was scored as having a significant post-reproductive lifespan. A third individual produced 22 broods of young at regular intervals, temporarily ceased reproduction, then produced five more litters at regular intervals. It then ceased reproduction and died long after the birth of its last brood. This individual was likely to have lived beyond its capacity to reproduce.
Figure 1 Interbrood Intervals
Here we illustrate the frequency distribution of interbrood intervals for all individual intervals in the experiment. Vales are plotted for all litters less than 20 d, then per day for days 20–50. We then report the total number of litters for days 51–100, then in blocks per 100-d interval thereafter (e.g., values shown as “100” are 101–200). The bottom of the figure details the reproductive history of three individuals from the Oropuche high-predation treatment group to illustrate individual variation. Each individual's litter birth dates are represented by an open symbol on the x-axis, which represents the individuals age. The age at death is represented by a closed symbol. Details of these examples are described in the text.
Table 3 Values for Mean Growth Rate, Mean Asymptotic Body Size, Median Interbrood Interval, and Mean Number of Litters per Lifetime for Each Treatment Group
To accommodate this individual variation, we quantified the 90th percentile for the expected number of days between successive litters for each individual in the study, based on that individual's parturition records. We interpret living beyond that time interval as the equivalent of a one-tailed test of the probability that that individual lived beyond its capacity to produce an additional litter of young (see Materials and Methods). Individuals that died within that time interval might still have had the capacity to reproduce.
Post-Reproductive Lifespan
Approximately 60% of the individuals in this study had post-reproductive lifespans exceeding their 90th percentile (Figure 2). We interpret this to mean that these individuals are likely to have lived beyond the time required for them to produce one more litter of young or that they had outlived their ability to reproduce.
Figure 2 The Distribution of Post-Reproductive Lifespans
Each drainage × predation combination is displayed separately. Filled bars represent females who died within the 90th percentile for their interbrood interval, or the interval during which they are expected to produce another brood of offspring. Open bars represent females that lived beyond their 90th percentile and hence are interpreted as having a significant post-reproductive lifespan.
We used logistic regression to ask whether or not there were differences among treatment groups in the probability of having a significant post-reproductive lifespan. The full model, which included drainage, food availability and predation, did not account for a significant proportion of the variability in the probability of living beyond an individual's 90th percentile (likelihood ratio Chi Square = 6.19, 3 df [degrees of freedom], p = 0.1027; data in Table 4). None of the three main effects were significant on their own (p
drainage = 0.0810, p
food = 0.0800, p
predation = 0.7058); the drainage effect was marginal because guppies from the Oropuche drainage were more likely to live beyond their 90th percentile than those from the Yarra drainage. We then evaluated the effects of food and predation within each drainage in four separate analyses. The effect of food was not significant within the Oropuche drainage (p = 0.61) and was marginally significant in the Yarra drainage (p = 0.0559); low-food fish were more likely to live beyond their 90th percentile. The effect of predation was not significant in either drainage (Yarra: p = 0.7097; Oropuche: p = 0.9784). Therefore, in spite of the substantial and significant differences among drainages and predation communities for other components of the life history, there were no significant differences among drainages or predation communities in either the duration of post-reproductive lifespan or the probability that an individual was likely to have ceased reproduction before dying.
Table 4 Number of Females with and without a Significant Post-Reproductive Lifespan
Comfort's [39] evaluation of reproduction in old, domestic guppies presents the only other data set for fish that we could find that includes information on post-reproductive lifespan, but it does not include any analyses. He only reports on the age at last reproduction and age at death, so we were only able to evaluate the distribution of the duration of post-reproductive lifespan for his fish. This distribution is similar in shape to our data and shows that 28 out of 36 fish lived beyond the age when they could be expected to produce another litter of young (Figure 3) if their variance in time intervals between successive litters is the same as in our fish.
Figure 3 The Distribution of Values for Post-Reproductive Lifespan in Comfort (1961)
Comfort [7] reported the age at last reproduction and age at death in one of his studies of senescence in domestic guppies. We report here a summary distribution of their post-reproductive lifespan. For purposes of comparison, the dotted vertical line represents the mean 90th percentile for the number of days between successive litters of young in our data set for Trinidadian guppies. The dashed line represents the 95th percentile for this mean value. This figure shows that the shape of the distribution of post-reproductive lifespans for Comfort's guppies was similar to that observed in our study. It also shows that the average duration of the post-reproductive lifespan was longer, which correlates positively with their longer total lifespans.
Discussion
Comparisons of Guppy Aging to Other Organisms
Our analyses show that guppies have a pattern of reproductive senescence, including significant post-reproductive lifespans, that is similar to that of many birds and mammals. We previously showed that there is an age-specific acceleration in mortality rate and decline in reproductive performance of guppies with age [34]. Here we have shown that most individuals also have an extended post-reproductive lifespan. The similarity between guppies and mammals in mortality was shown long-ago by Comfort [36], but comparable lifetime data for female reproduction are not available for any species of fish. Our findings refute the prediction that, because many fish appear to generate primary oocytes throughout adult life, fish will show negligible reproductive senescence [35]. Such predictions have been based on the observation of the production of oocytes in old, wild-caught fish [35,37] or fertility and presence of viable oocytes in some of the oldest representatives in laboratory studies of senescence on guppies [38,39] and platyfish and annual killifish [35].
From a quantitative perspective, we found that the median post-reproductive lifespan of guppies was 6.4% and the mean was 13.6% of the total lifespan, with a range of 0%–76% of the total lifespan. Investigators have reported post-reproductive lifespans of “up to” or “over” 30% of the total lifespan in birds and mammals [12,16], but these percentages tend to be reported as a single figure, rather than as being derived from a formal statistical analysis, as done here. Our data are thus not directly comparable to those of others, and we are not able to make direct comparisons between the duration of post-reproductive lifespans in guppies, birds, and mammals.
Mammals and birds produce the vast majority of primary oocytes during embryonic development [40,41]; this finite pool of eggs declines steadily during the lifespan. The usual explanation for midlife loss of ovulatory capacity in women and laboratory rodents is the depletion of this finite population of oocyte stores [42,43] however, recent work on mice [44] and humans [45] suggests that they retain primordial germ cells that continue to contribute to the pool of primary oocytes after birth. If guppies, like other fish, retain the capacity for producing new primary oocytes throughout adulthood, midlife cessation of ovulation in female fish, birds, or mammals may not always be dictated by the extinction of a finite population of primary oocytes, as conventional wisdom holds.
The Evolution of Lifespan
High-predation guppy populations have longer total lifespans because they have significantly longer reproductive lifespans. The differences in longevity of these populations can be visualized with a timeline that is the sum of the mean age at first reproduction, mean duration of the reproductive lifespan, and median post-reproductive lifespan (Figure 4). This visual summary of the results shows graphically that the reason that guppies from high-predation localities live longer is solely because they have longer reproductive lifespans. It thus appears that evolution has shaped that component of the life history that makes a direct contribution to fitness and that, as predicted, the post-reproductive lifespan, which has no impact on fitness, is highly variable and that there are no significant differences among treatment groups in this variable.
Figure 4 Summary of the Total Lifespan of Guppies, Reported Separately for Each Drainage × Predation Combination
The timeline reports the mean age at first reproduction, mean age at last reproduction (mean age at first reproduction plus mean reproductive lifespan), and total lifespan (mean age at first reproduction plus mean reproductive lifespan plus median post-reproductive lifespan). The range of values is reported in parentheses next to each mean and median value. This summary illustrates the overall differences in total lifespan and the fact that these differences are attributable to the duration of the reproductive lifespan alone. Note that we averaged the results for high- and low-food availability to simplify the presentation.
The marginally significant results for drainage of origin and food availability in the logistic regression analysis of post-reproductive lifespan indicate that a larger study could reveal that guppies from the Oropuche drainage and guppies reared on high food availability are more likely to have significant differences in post-reproductive lifespans. In both cases, the higher probability of an extended post-reproductive lifespan is positively correlated with the duration of the reproductive lifespan, which suggests that it increased as a correlate of the increase in reproductive lifespan. This correlation is relevant to our second hypothesis explaining heterogeneity among treatment groups in post-reproductive lifespan, which is that the post-reproductive period might be positively correlated with other components of the life history, regardless of whether or not it contributes to individual fitness. Hendry et al.'s results [21] for anadromous salmon provide a telling contrast to our results. Salmon are a particularly dramatic example of the evolution of resource allocation to reproduction, somatic maintenance, and post-reproductive lifespan because they stop feeding as they migrate from the ocean into their natal streams. Their life history from that point on is like the flight of a ballistic missile, since all activities are fueled by the reserves that they obtained prior to the cessation of feeding. In the population that Hendry et al. studied, there is variation in when females arrive at the breeding ground. Early-arriving females have an advantage in attaining the best breeding sites, but they must also live long enough to protect their nests from late-arriving females. If the nest is not guarded, then the site can be reused, which drastically reduces the survival of her eggs. There is thus positive selection for post-reproductive survival in early-arriving females. These females do indeed live longer after laying their eggs than late-arriving females. They appear to be able to do so because they have an increased retention of fat reserves for somatic maintenance during the guarding period, but do so at the expense of reduced fecundity. It is this tradeoff between reproduction, fat storage, and post-reproductive lifespan that argues for the differences in allocation between early- and late-arriving females being an adaptive feature of the life history. Analyses of neutral genetic variation at microsatellite loci confirm that there is sufficient genetic isolation between the early- and late-arriving cohorts for there to be limited gene flow between them and some adaptive divergence between them.
More generally, it is fair to say that we know far too little about the extent to which natural selection can shape post-reproductive lifespan, largely because too little effort has been invested in comparative, evolutionary studies. Most of the arguments that we are familiar with pertain to humans which, by themselves, represent an unreplicated observation. The comparison between humans and closely related primates [1] is telling because it argues that an extended post-reproductive lifespan is a recent innovation and is not typical of other great apes or other hominids [46]. When combined with Lahdenpera et al.'s demonstration [3] that the post-reproductive lifespan of menopausal woman is strongly associated with their children's reproductive success, then there is a stronger argument that the post-reproductive lifespan of humans is in fact an adaptation. The aggregate of results thus far argues that it is possible to develop predictions about post-reproductive survival in such a comparative framework and thus perform adaptive analyses, as done for so many other aspects of the life history.
Materials and Methods
We compared the life histories of guppies from high- and low-predation localities in the Yarra and Oropuche drainages of the Northern Range Mountains of Trinidad. Genetic data suggest that the adaptation of guppies to the presence or absence of predators occurred independently in each drainage [29,30]. One high- and one low-predation locality was sampled from each drainage, hence the four localities can be thought of as two paired comparisons between high- and low-predation environments. The subjects of this experiment consisted of the second generation of laboratory-born offspring derived from wild-caught females. The breeding design used to produce these fish used an equal number of grandchildren descended from each wild-caught female. The design thus retained the genetic diversity of the original sample while avoiding inadvertent selection for adaptation to laboratory conditions (details in [33,34]).
Each locality was represented by 30 sibships, and each sibship by two sisters, one reared at high and one at low levels of resource availability; the experiment thus included a total of 240 individuals in eight treatment groups, but after accidental deaths, only 226 were included in the analyses. The fish were measured every other week, and food availability was increased to accommodate growth. All individuals in a given food treatment were given the same amount of food for each 2-wk time interval. Food availability became constant after 9 mo. The high-food treatment received an average of 2.5 times as much food as the low-food treatment.
Food was included as a factor in the design because low-predation localities tend to have lower levels of food availability and hence lower growth rates and smaller asymptotic body sizes in comparison to guppies from high-predation localities. These differences largely disappear when the fish are reared in the lab on similar levels of food availability. The food-availability trajectories were chosen to approximate the average growth rates and asymptotic body sizes of fish from high- versus low-food environments in nature [47]. Partridge and Barton [48] argued that the results of comparative studies of aging in a common environment may be biased if they differ in how well each is adapted to the chosen environment in the lab. Our alternative food levels match the alternative environments experienced in the field and enable us to assess the possibility of this type of adaptation, which would be seen as an interaction between food availability and predation regime. No such interactions were seen in the variables considered in this paper, but since food represented a significant main effect on many dependent variables, it is retained in the presentation of these results. Table 3 includes mean values for asymptotic body size and growth rate that illustrate the consequences of the different food levels for growth. Growth rates are estimated from a fitting of the Von Bertalanffy growth equation to each individual's growth history using Proc NLIN [49].
Each female was isolated in a 7.8-l aquarium when 25 d old. She received a measured volume of food, liver paste in the morning and newly hatched brine shrimp nauplii in the afternoon. Food availability was quantified volumetrically, to the nearest microliter, with a Hamilton micropipette. Females were mated once a week until they gave birth to their first litter of young, then were mated again after the birth of each litter, when they are particularly receptive to mating [50]. All females were maintained for the entirety of their lifespans. The dependent variables include: growth, asymptotic body size, age at maturity, age when each litter was produced, number and size of young in each litter, age at last reproduction, and age at death.
The focus of the current analyses is the division of the lifespan of each individual into the non-overlapping pre-reproductive, reproductive, and post-reproductive segments.
Post-reproductive lifespan
For each individual, we asked whether the time interval between when it last gave birth and when it died was significantly longer than the interval of time required for that individual to have given birth to its next litter of young. Because individuals vary in the regularity of the intervals between successive litters, we calculated the 90th percentile for the duration of time between successive litters for each individual based on its own reproductive schedule. We chose the 90th percentile as a threshold for classifying individuals as either dying within the time interval when they could have produced an additional litter or as having lived longer than the expected amount of time required to produce another litter of young. If the time interval between the birth of the last litter and death exceeded the 90th percentile interval, then it was judged to have lived beyond the time when it should have given birth to another litter of young and hence was likely to have had a significant post-reproductive lifespan. Dichotomizing the data into those that did or did not live beyond this confidence interval provides a criterion for comparing treatment groups for the probability of significant post-reproductive lifespan. The results were qualitatively the same if we used the 95th percentile as the threshold.
We first compared treatment groups with parametric analyses (SAS, Proc GLM [49]) by dividing each individual's lifespan in to three non-overlapping segments: the age at maturity, the reproductive lifespan (age at last reproduction minus age at maturity), and the post-reproductive lifespan (age at death minus age at last reproduction). The distribution of the first two segments of the life history approximated a normal distribution sufficiently well to justify the use of parametric statistics. The distribution of post-reproductive lifespan did not conform to a normal distribution when either untransformed or log-transformed. Although ANOVAs are robust to non-normality, we also performed a permutation test [51] on the post-reproductive lifespan results to confirm that there were not significant main effects or interactions.
We performed logistic regression analyses on the relationship between post-reproductive lifespan and the different treatment groups (drainage, predation, and food availability). We classified individuals according to whether they died within or outside of the 90th percentile for time between litters. We used the SAS Logistic Regression Procedure [49] to evaluate whether or not the inclusion of drainage, food availability, or predation accounted for significant differences in the probability of having a post-reproductive lifespan.
We extracted similar data on post-reproductive lifespan from Comfort's [7] results for domestic guppies. He reported the age at last reproduction and age at death for 36 fish kept singly or in groups of five. Since he did not report other details of their life history, we were not able to calculate 90th percentile for time between litters, as we did for wild guppies. We could, however, characterize the distribution of post-reproductive lifespans and compare them with data from our laboratory.
We gratefully acknowledge the support of the National Science Foundation (DEB-9707473). We also acknowledge the support of Heather Bryga, Bronson Bassir, and Dionna Elder for overseeing the execution of the experiment, and a host of students for their help in maintaining the experiment.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. DR conceived and designed the experiments and supervised the execution of the experiments. MB analyzed the data. DR, MB, and DH wrote the paper.
Citation: Reznick D, Bryant M, Holmes D (2006) The evolution of senescence and post-reproductive lifespan in guppies (Poecilia reticulata). PLoS Biol 4(1): e7.
Abbreviations
ANOVAanalysis of variance
==== Refs
References
Hawkes K O'Connell JF Blurton-Jones NG Alvarez H Charnov EL Grandmothering, menopause, and the evolution of human life histories Proc Natl Acad Sci U S A 1998 95 1336 1339 9448332
Shanley DP Kirkwood TBL Evolution of the human menopause Bioessays 2001 23 282 287 11223885
Lahdenpera M Lummaa V Helle S Tremblay M Russell AF Fitness benefits of prolonged post-reproductive lifespan in women Nature 2004 428 178 181 15014499
Packer C Tatar M Collins A Reproductive cessation in female mammals Nature 1998 392 807 811 9572138
Hamilton WD The moulding of senescence by natural selection J Theor Biol 1966 12 12 45 6015424
Charlesworth B Evolution in age-structured populations 1994 Cambridge (United Kingdom) Cambridge University Press 306
Comfort A Age and reproduction in female Lebistes
Gerontologia 1961 5 146 149 13880756
Schreibman MP Margolis-Nunno H Schreibman MP Scanes CG The brain-pituitary-gonad axis in poikilotherms Development, maturation, and senescence of neuroendocrine systems: A comparative approach 1989 San Diego (California) Academic Press 97 134
Woodard AE Abplanalp H Longevity and reproduction in Japanese quail maintained under stimulatory lighting Poult Sci 1971 50 688 692 5106128
Cherkin A Eckardt MJ Effects of dimethylaminoethanol upon life-span and behavior of aged Japanese quail J Gerontol 1977 32 38 45 830732
Ottinger MA Quail and other short-lived birds Exp Gerontol 2001 36 859 868 11295519
Holmes DJ Ottinger MA Birds as long-lived animal models for the study of aging Exp Gerontol 2003 38 1365 1375 14698817
Faddy MJ Gosden RG Edwards RG Ovarian follicular dynamics in mice: A comparative study of three inbred strains and an F1 hybrid J Endocrinol 1983 96 23 33 6822780
Gosden RG Laing SC Felicio LS Nelson JF Finch CE Imminent oocyte exhaustion and reduced follicular recruitment mark the transition to acyclicity in aging C57BL/6J mice Biol Reprod 1983 28 255 260 6838945
Nelson JF Felicio LS Reproductive aging in the female: An etiological perspective Rev Biol Res Aging 1985 2 251 314
Austad SN Retarded senescence in an insular population of Virginia opossums (Didelphis virginiana)
J Zool 1993 229 695 708
VomSaal FS Finch CE Nelson CE Knobil E Neill JD Natural history and mechanisms of reproductive aging in humans, laboratory rodents, and other selected vertebrates The physiology of reproduction 1994 New York Raven Press 1213 1314
Austad SN Menopause: An evolutionary perspective Exp Gerontol 1994 29 255 263 7925746
Gage TB The comparative demography of primates: With some comments on the evolution of life histories Ann Rev Anthropol 1998 27 197 221 12295432
Ottinger MA Balthazart J Altered endocrine and behavioral responses with reproductive aging in the male Japanese quail Horm Behav 1986 20 83 94 3514416
Hendry AP Morbey YE Berg OK Wenburg JK Adaptive variation in senescence: Reproductive lifespan in a wild salmon population Proc Roy Soc London B 2004 271 259 266
Reznick DN Butler MJ Rodd FH Ross P Life history evolution in guppies (Poecilia reticulata) . 6. Differential mortality as a mechanism for natural selection Evolution Int J Org Evolution 1996 50 1651 1660
Reznick D Genetic determination of offspring size in the guppy (Poecilia reticulata)
Am Nat 1982 120 181 188
Reznick DN Bryga H Life-history evolution in guppies (Poecilia reticulata Poeciliidae). V. Genetic basis of parallelism in life histories Am Nat 1996 147 339 359
Reznick DN Bryga H Life-history evolution in guppies. 1. Phenotypic and genotypic changes in an introduction experiment Evolution Int J Org Evolution 1987 41 1370 1385
Reznick DA Bryga H Endler JA Experimentally induced life-history evolution in a natural population Nature 1990 346 357 359
Reznick DN Shaw FH Rodd FH Shaw RG Evaluation of the rate of evolution in natural populations of guppies (Poecilia reticulata)
Science 1997 275 1934 1937 9072971
Bryant MJ Reznick D Comparative studies of senescence in natural populations of guppies Am Nat 2004 163 55 68 14767836
Carvalho GR Shaw PW Magurran AE Seghers BH Marked genetic divergence revealed by allozymes among populations of the guppy Poecilia reticulata (Poeciliidae), in Trinidad Biol J Linn Soc London 1991 42 389 405
Fajen A Breden F Mitochondrial DNA sequence variation among natural populations of the Trinidad guppy Poecilia reticulata
Evolution Int J Org Evolution 1992 46 1457 1465
Medawar PB An unsolved problem of biology 1952 London HK Lewis and Co 24 p
Kirkwood TBL Holliday R The evolution of ageing and longevity Proc R Soc Lond B Biol Sci 1979 205 531 546 42059
Reznick DN Buckwalter G Groff J Elder D The evolution of senescence in natural populations of guppies (Poecilia reticulata) : A comparative approach Exp Gerontol 2001 36 791 812 11295514
Reznick DN Bryant MJ Roff D Ghalambor CK Ghalambor DE Effect of extrinsic mortality on the evolution of senescence in guppies Nature 2004 431 1095 1099 15510147
Finch CE Longevity, senescence and the genome 1990 Chicago University of Chicago Press 922
Comfort A Effect of delayed and resumed growth on the longevity of a fish (Lebistes reticulatus Peters) in captivity Gerontologia 1963 8 150 155
de Bruin JP Gosden RG Finch CE Leaman BM Ovarian aging in two species of long-lived rockfish, Sebastes aleutianus and S. alutus
Biol Reprod 2004 71 1036 1042 15151924
Comfort A The effect of age on growth-resumption in fish (Lebistes) checked by food restriction Gerontologia 1960 4 177 186 13694915
Comfort A The longevity and mortality of a fish (Lebistes reticulatus Peters) in captivity Gerontologia 1961 5 209 222
Tokarz RR Autoradiographic study of effects of mammalian gonadotropins (follicle-stimulating-hormone and luteinizing-hormone) and estradiol-17beta on [3H]thymidine labeling of surface epithelial cells, pre-follicular cells, and oogonia in ovary of lizard Anolis carolinensis
Gen Comp Endocrinol 1978 35 179 188 566690
Guraya SS Ovarian follicles in reptiles and birds 1989 Berlin (Germany) Springer-Verlag 285
Gosden RG The biology of menopause: The causes and consequences of ovarian aging 1985 New York Academic Press 188
Nelson JF Karelus K Bergman MD Felicio LS Neuroendocrine involvement in aging: Evidence from studies of reproductive aging and caloric restriction Neurobiol Aging 1995 16 837 843 8532119
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
Bukovsky A Caudle MR Svetlikova M Upadhyaya NB Formation of primordial follicles in adult human ovaries Reprod Biol Endocrinol 2004 2 20 15115550
Caspari R Lee SH Older age becomes common late in human evolution Proc Natl Acad Sci U S A 2004 101 10895 10900 15252198
Reznick DN Butler MJI Rodd FH Life history evolution in guppies. 7. The comparative ecology of high and low predation environments Am Nat 2001 157 126 140 18707267
Partridge L Barton NH On measuring the rate of ageing Proc R Soc Lond B Biol Sci 1996 263 1365 1371
SAS Institute SAS/STAT user's guide, version 8 1999 Cary (North Carolina) SAS Institute 3,884 p
Crow RT Liley NR A sexual pheromone in the guppy, Poecilia reticulata (Peters) Can J Zool 1979 57 184 188
Potvin C Roff DA Distribution-free and robust statistical methods: Viable alternatives to parametric statistics Ecology 1993 74 1617 1628
|
16363919
|
PMC1318473
|
CC BY
|
2021-01-05 08:21:46
|
no
|
PLoS Biol. 2006 Jan 27; 4(1):e7
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0040007
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpmedplosmedPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1636673410.1371/journal.pbio.0040009Research ArticleCell BiologyImmunologyAllergy/ImmunologyObstetrics/GynecologyHomo (Human)PrimatesMammalsVertebratesEukaryotesActivation of NK Cells by an Endocytosed Receptor for Soluble HLA-G NK Cell Activation by Receptor EndocytosisRajagopalan Sumati
1
Bryceson Yenan T
1
Kuppusamy Shanmuga P
1
Geraghty Daniel E
2
van der Meer Arnold
3
Joosten Irma
3
Long Eric O
1
*1Laboratory of Immunogenetics, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Rockville, Maryland, United States of America2Fred Hutchinson Cancer Research Center, Seattle, Washington, United States of America3Department of Blood Transfusion and Transplantation Immunology, Radboud University Nijmegen Medical Centre, Nijmegen, the NetherlandsPloegh Hidde Academic EditorHarvard Medical SchoolUnited States of America* To whom correspondence should be addressed. E-mail: [email protected] 2006 27 12 2005 27 12 2005 4 1 e929 3 2005 26 10 2005 This 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.2006This 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.
Inside a Killer: Immune Signals May Promote Vascular Growth
Signaling from endosomes is emerging as a mechanism by which selected receptors provide sustained signals distinct from those generated at the plasma membrane. The activity of natural killer (NK) cells, which are important effectors of innate immunity and regulators of adaptive immunity, is controlled primarily by receptors that are at the cell surface. Here we show that cytokine secretion by resting human NK cells is induced by soluble, but not solid-phase, antibodies to the killer cell immunoglobulin-like receptor (KIR) 2DL4, a receptor for human leukocyte antigen (HLA)-G. KIR2DL4 was constitutively internalized into Rab5-positive compartments via a dynamin-dependent process. Soluble HLA-G was endocytosed into KIR2DL4–containing compartments in NK cells and in 293T cells transfected with KIR2DL4. Chemokine secretion induced by KIR2DL4 transfection into 293T cells occurred only with recombinant forms of KIR2DL4 that trafficked to endosomes. The profile of genes up-regulated by KIR2DL4 engagement on resting NK cells revealed a proinflammatory/proangiogenic response. Soluble HLA-G induced secretion of a similar set of cytokines and chemokines. This unique stimulation of resting NK cells by soluble HLA-G, which is endocytosed by KIR2DL4, implies that NK cells may provide useful functions at sites of HLA-G expression, such as promotion of vascularization in maternal decidua during early pregnancy.
KIR2DL4, a human killer cell immunoglobulin receptor expressed on natural killer cells, can be stimulated by soluble antibody or the soluble version of its natural ligand (HLA-G), and may signal from within endosomes.
==== Body
Introduction
Natural killer (NK) cells are a subset of lymphocytes that mediate innate immunity and regulate adaptive immunity via cytokine secretion and cytotoxic activity. The activation of NK cell function is a result of the integration of activating and inhibitory signals delivered by NK cell receptors [1,2]. NK cell receptors recognize ligands that are either up-regulated or expressed constitutively on target cells. Cells can become sensitive to NK cell–mediated cytotoxicity for various reasons, such as loss of major histocompatibility complex (MHC) class I expression or up-regulation of surface molecules in response to DNA damage or stress, which can occur as a result of infection or transformation [3,4]. Normal cells are resistant to NK cell-mediated cytotoxicity due to the presence of inhibitory receptors on the surface of NK cells that recognize major histocompatibility complex (MHC) class I molecules. The interaction between inhibitory receptors and their MHC ligands on target cells transduces a negative signal that blocks the lytic activity of NK cells. NK cells in peripheral blood represent 5%–10% of circulating lymphocytes. In contrast, NK cells are the predominant lymphocyte subset in uterus. In early pregnancy, uterine NK cells proliferate and remain in the decidua basalis, which consists of uterine tissue at the maternal–fetal interface [5]. The precise function of uterine NK cells is still unknown.
Both peripheral blood and uterine human NK cells express killer cell immunoglobulin-like receptors (KIR), a family of NK cell receptors that recognize MHC class I molecules. KIR are type I transmembrane glycoproteins with two or three Ig-like domains and cytoplasmic tails of varying lengths [6]. KIRs with long cytoplasmic domains (KIR2DL and KIR3DL) are inhibitory receptors that contain cytoplasmic immunoreceptor tyrosine-based inhibition motifs (ITIM). Those KIRs with short cytoplasmic domains (KIR2DS and KIR3DS) are activating receptors that associate with the adapter DAP12 via a lysine residue in their transmembrane region. With one exception, KIR genes are not expressed in all NK cells. Rather, each NK cell expresses its own repertoire of KIR genes.
KIR2DL4 is an evolutionarily conserved, framework member of the KIR gene family that is expressed by all KIR haplotypes and in all NK cells. In contrast to other activating or inhibitory KIR family members, which regulate NK cell cytotoxicity and cytokine production, KIR2DL4 activates cytokine production, but not cytotoxicity, in resting NK cells from peripheral blood [7]. It is unique in its genomic organization and regulation and in its protein structure and function [7–12]. KIR2DL4 is polymorphic with two reported genetic variants, designated 10A and 9A [13]. While the product of the 10A allele is detectable on the cell surface, the 9A allele, encoding a protein with a truncated cytoplasmic tail, is not stable at the cell surface [14]. KIR2DL4 has a charged arginine residue in its transmembrane region; however, unlike other activating KIR2DS that pair with the adapter DAP-12, KIR2DL4 can associate with the FcRI γ chain [15]. Engagement of KIR2DL4 results in activation despite the inhibitory potential conferred by the presence of an ITIM in its cytoplasmic tail [7,16,17]. In resting, peripheral blood NK cells, ligation of KIR2DL4 with monoclonal antibody (mAb) results in interferon (IFN)-γ production but not cytotoxicity [7]. The very low cell surface expression of KIR2DL4 [14,17] has been difficult to reconcile with the functional outcome associated with this receptor.
Data suggest that KIR2DL4 binds the nonclassical class I molecule HLA-G [8,9]. HLA-G, a nonclassical class I molecule of limited polymorphism, has a unique expression pattern restricted mainly to trophoblast cells that invade the maternal decidua during early pregnancy [18]. HLA-G expression may be inducible in other cell types, in response to inflammation, infection, and transformation [19]. Several isoforms of HLA-G are expressed in the placenta, including membrane-bound forms (HLA-G1, -G2, -G3, and -G4) and soluble forms (HLA-G5 and -G6) [20]. Whereas membrane HLA-G expression in trophoblast cells is restricted to extravillous trophoblast cells, which invade the maternal decidua, expression of soluble HLA-G was detected in all types of placental trophoblast cells [21]. To date, the precise role of HLA-G in the placenta remains unclear [22]. A common hypothesis proposes that expression of HLA-G on invading trophoblast cells is needed to prevent NK cell attack [9,23]. However, this might not be necessary as trophoblast cells are intrinsically resistant to NK cell–mediated lysis [24]. An alternative hypothesis proposes that trophoblast–NK cell interactions regulate expression of cytokines by NK cells to promote remodeling of the maternal vasculature, which is required to establish adequate blood supply to the fetus [5,25,26]. The benefit of NK cell activation during early pregnancy is supported by genetic studies on preeclampsia, a potentially fatal disease due to incomplete remodeling of spiral arteries by trophoblast cells. Resistance to preeclampsia correlated with combinations of fetal HLA genes and maternal KIR genes that seem to favor NK cell activation over NK cell inhibition [27].
In mice as well, NK cells are present in the uterine mucosa where they increase in number after implantation and become activated as decidualization occurs [28]. In both mice and humans, uterine NK cells have limited cytotoxic potential but produce cytokines including IFN-γ and tumor necrosis factor (TNF)-α [20,29]. In several different strains of genetically manipulated mice that lack NK cells, decidual integrity is compromised and spiral artery modification is minimal. In addition, IFN-γ produced by mouse uterine NK cells has been implicated in the remodeling of the vasculature of the maternal arteries necessary for a normal pregnancy [30]. The signals that induce activation of uterine NK cells are still unknown.
Results presented here support a role for KIR2DL4 in the activation of NK cells at sites of soluble HLA-G expression. Unlike other members of the KIR family, KIR2DL4 is localized predominantly in endosomes and mediates endocytosis of its ligand, soluble HLA-G. Furthermore, engagement of KIR2DL4 by soluble ligand activates a proinflammatory/proangiogenic response, consistent with a role in promoting vascularization during early pregnancy.
Results
Interferon-γ Secretion Is Induced by Soluble, but Not Plate-Coated, or Bead-Bound Antibodies to KIR2DL4
Resting NK cells typically express very low or undetectable levels of KIR2DL4 on the cell surface. Activation of NK cells with interleukin (IL)-2 and feeder cells resulted in a transient increase in the level of cell surface KIR2DL4 (Figure 1A). The increase usually peaked between 7 and 14 days, and by 3 weeks in culture, receptor levels were no longer detectable. However, despite low surface levels of KIR2DL4 on resting NK cells, secretion of IFN-γ was observed upon engagement by anti-KIR2DL4 mAbs [7]. Three different mAbs (IgG1 33 and IgM 36 and 64) in solution, and even purified Fab of mAb 33, could enhance IFN-γ secretion in the absence of cross-linking with secondary antibodies [7] (Figure 1B and unpublished data). In contrast, plate-coated or bead-bound mAb 33 did not enhance IFN-γ secretion (Figure 1B). The lack of stimulation by cross-linked antibodies was not due to induction of cell death, as determined with annexin-V–FITC and propidium iodide to detect apoptotic and necrotic cells (unpublished data). Activation by soluble antibodies to KIR2DL4 was in direct contrast to the response to other NK cell receptors such as CD16 or 2B4 (CD244), for which soluble antibodies were not sufficient to trigger a response, and cross-linking by plate-coating or by binding to beads was required to obtain comparable IFN-γ release (Figure 1B and unpublished data).
Figure 1 IFN-γ Secretion by Resting NK Cells Is Induced by Soluble Antibodies to KIR2DL4
(A) KIR2DL4 expression on the surface of NK cells immediately after isolation (day 0) or after 7, 14, or 21 d in culture with autologous feeder cells and rIL-2. Cells were stained with mAb 33.
(B) Resting NK cells were incubated with soluble (10 μg/ml), plate-coated (5 μg/0.1 ml/well), or bead-bound (4 beads/cell) antibodies to KIR2DL4 (mAb 33) or CD16 (mAb 3G8). After 24 h, culture supernatants were tested by ELISA for IFN-γ production.
Dynamin-Dependent Endocytosis of KIR2DL4
Activation by soluble antibodies to KIR2DL4, despite negligible receptor expression at the cell surface, could be explained if KIR2DL4 were actively endocytosed into resting NK cells. To examine the subcellular distribution of KIR2DL4, both resting and activated NK cells were fixed, permeabilized, stained with KIR2DL4-specific mAb 33, and analyzed by confocal microscopy. Most of the receptor resides in vesicular structures, in contrast to other KIR receptors such as KIR2DL1, which is detectable at the cell surface (Figure 2A). To test if active endocytosis was a mechanism for delivery of KIR2DL4 into intracellular compartments, resting NK cells were incubated with Cy3-conjugated mAb 33 at 37 °C for different times prior to analysis by confocal microscopy. As shown in Figure 2B, at 30 min, some of the KIR2DL4 had begun to internalize, and by 120 min, extensive endocytosis of KIR2DL4 was detected by the presence of the 33-Cy3 mAb in intracellular vesicles. The staining of internalized KIR2DL4 at 120 min was very similar to that detected by intracellular staining for total KIR2DL4 (Figure 2A). Endocytosis was also seen with the Fab portion of the KIR2DL4-specific mAb 33, ruling out Fc receptor–mediated internalization of mAb 33 (Figure 2B).
Figure 2 KIR2DL4 Is Endocytosed into Intracellular Vesicles
(A) The B cell line 721.221 (221), resting NK cells, and activated NK cells were fixed, permeabilized, and stained with Cy3-conjugated mAb 33. In the image on the right, resting NK cells were fixed, permeabilized, and stained with a polyclonal antibody specific for the tail of KIR2DL1, followed by Alexa-568–conjugated secondary antibodies.
(B) Resting NK cells were incubated at 37 °C for 30 min or 120 min with KIR2DL4-specific mAb 33-Cy3, with a Fab of control mAb 2A8 or with a Fab of anti-KIR2DL4 mAb 33, as indicated. Cells were then fixed and analyzed by confocal microscopy.
(C) The 293T cells stably transfected with KIR2DL4-gfp (293T-2DL4-gfp) were fixed, permeabilized, and stained with anti-KIR2DL4 mAb 33 followed by Alexa-568–conjugated secondary antibodies to show co-localization with KIR2DL4-gfp. Single confocal sections are shown.
(D) The 293T-2DL4-gfp cells were incubated at 37 °C for 120 min with anti-KIR2DL4 mAb 33, either intact (33 whole Ab) or as a Fab (33-Fab). Cells were fixed, stained with Alexa-568–conjugated secondary antibodies, and analyzed by confocal microscopy. Single confocal sections are shown.
An intracellular, vesicular pattern of KIR2DL4 distribution was also observed after stable transfection of gfp-tagged KIR2DL4 in the human embryonic kidney cell line 293T (Figure 2C, center). In addition, good co-localization of anti-2DL4 mAb 33 and gfp-tagged KIR2DL4 was observed, validating the specificity of the staining (Figure 2C). Uptake of mAb 33 was also detected in 293T-2DL4-gfp cells (Figure 2D). Endocytosis of KIR2DL4 was not induced by mAb-mediated cross-linking, because the purified Fab portion of mAb 33 was also endocytosed (Figure 2D). In both instances, mAb was internalized to the same compartments where endogenous KIR2DL4-gfp was localized (Figure 2D). These data show that KIR2DL4 can also be endocytosed in non-NK cells.
Next, we tested whether endocytosis of KIR2DL4 was dependent on the GTPase dynamin, a central player in clathrin-mediated endocytosis [31]. The endocytosis assay was performed in transfected 293T cells expressing HA-tagged KIR2DL4 and either wild-type gfp-tagged dynamin or a dominant-negative mutant (K44A) of gfp-tagged dynamin, which competes for dynamin function due to a defect in GTP binding and hydrolysis [32]. Cells were incubated with Cy3-conjugated anti-KIR2DL4 mAb 33 for 2 h and processed for confocal imaging. Cells expressing dynamin were identified by fluorescence of gfp. Whereas endocytosed KIR2DL4 could be seen in vesicular structures in the presence of wild-type dynamin, dynamin K44A hindered the transport of KIR2DL4 into vesicles and caused a striking redistribution of the receptor at the plasma membrane (Figure 3).
Figure 3 Internalization of KIR2DL4 Is Dynamin Dependent
The 293T cells transfected with HA-tagged KIR2DL4 together with either wild-type dynamin-gfp (Dynamin Egfp-WT) or a dominant-negative mutant of dynamin (dynamin Egfp-K44A) were loaded with KIR2DL4-specific Cy3-conjugated mAb 33 for 120 min and fixed. Individual confocal sections are shown.
To visualize endocytic compartments containing KIR2DL4 at higher resolution, the NK cell line NKL and the NK cell line YTS stably transfected with KIR2DL4-gfp (YTS-2DL4-gfp) were loaded with mAb 33 for 120 min at 37 °C. Cells were fixed, incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies, and reacted with 3,3′-diaminobenzidine substrate. Electron micrographs revealed HRP reaction product mainly in vesicles that ranged between 250 and 500 nm in size in both NKL and YTS-2DL4-gfp cells (Figure 4A). The majority of HRP-reactive vesicles contained a rim of HRP reaction product and an electron-lucent central region characteristic of early endosomes. Some of the vesicles also had more electron-dense content (Figure 4A). There was negligible vesicular staining of cells not loaded with mAb 33 (unpublished data). Staining of a few tubular structures and multivesicular bodies was also observed, consistent with localization in a subset of recycling endosomes and lysosomes, respectively (unpublished data). Thus, electron microscopy confirmed the vesicular localization of endocytosed KIR2DL4.
Figure 4 Immunolocalization of Endocytosed KIR2DL4 in NK Cells by Electron Microscopy
(A) The NK cell lines NKL and YTS-2DL4-gfp were loaded with anti-KIR2DL4 mAb 33 for 120 min. Endocytosed receptor was detected using HRP-conjugated sheep anti-F(ab′)2 mouse IgG. The HRP reaction product visible as a dark stain identifies the location of endocytosed KIR2DL4. Vesicular structures positive for KIR2DL4 ranged between 250 and 500 nm in size.
(B) The NK cell line YTS, stably transfected with KIR2DL4-gfp (YTS-2DL4-gfp), was fixed, permeabilized, and stained with antibody to perforin followed by Alexa-568–conjugated secondary antibodies. Single confocal sections are shown.
Confocal microscopy was performed with YTS-2DL4-gfp cells to test how KIR2DL4-containing vesicles relate to perforin-containing cytotoxic granules. Anti-perforin antibodies decorated vesicles in YTS-2DL4-gfp cells that were distinct from KIR2DL4-containing vesicles (Figure 4B). The lack of overlap was confirmed by Z-stack/three-dimensional reconstruction studies (unpublished data). Resting NK cells co-stained with mAb 33 and anti-perforin antibodies also revealed no overlap of KIR2DL4 with perforin (unpublished data). Lack of overlap was also seen with antibodies to granzyme A (unpublished data). The chemokine RANTES resides in specialized storage vesicles in T cells [33]. Confocal experiments revealed that RANTES was also in vesicular compartments in NK cells but that those compartments were distinct from vesicular compartments containing KIR2DL4 (unpublished data).
KIR2DL4 Resides in Rab5+ Early Endosomes
Co-localization of KIR2DL4 with a number of markers for endocytic and lysosomal compartments was tested in resting NK cells and in 293T-2DL4-gfp cells. Rab5 and its effector EEA-1 are markers for early endosomes; mannose 6-phosphate receptor (M6PR) is a marker for late endosomes; and lysosome-associated membrane protein 1 (LAMP-1) is a marker for lysosomes. Resting NK cells were fixed, permeabilized, and stained with anti-KIR2DL4 mAb 33 and with Alexa-488–conjugated secondary antibodies, followed by staining with antibodies against Rab5, EEA-1, or M6PR and with Alexa-568–conjugated secondary antibodies (Figure 5 and unpublished data). The 293T-2DL4-gfp cells were stained for the same endocytic markers and KIR2DL4 was visualized by gfp fluorescence. In both resting NK cells and 293T-2DL4-gfp cells, significant overlap of staining was obtained only in the case of Rab5 (Figure 5). Even though Rab5 and EEA-1 identify early endosomes, the overlap of staining of KIR2DL4 with EEA-1 was more limited. This was confirmed by three-color confocal analysis to detect KIR2DL4, Rab5, and EEA-1 in the same cells (unpublished data). Likewise, there was limited overlap of KIR2DL4 staining with M6PR (Figure 5) and LAMP-1 (unpublished data).
Figure 5 KIR2DL4 Resides in Endocytic Compartments
Resting NK cells and 293T-2DL4-gfp cells were fixed, permeabilized, and stained with antibodies against Rab5, EEA-1, and M6PR, followed by Alexa-568–conjugated secondary antibodies. NK cells were further stained with mAb 33 coupled to Alexa-488 to detect KIR2DL4.
Members of the Rab GTPase family, such as Rab4, Rab5, Rab7, and Rab11, occupy distinct yet overlapping membrane domains on early and recycling endosomes and serve as determinants of organelle identity [34]. Rab5 is a key regulator of transport to early endosomes; Rab4 and Rab11 are implicated in recycling, from endosomes back to the plasma membrane; and Rab7 has a role in the late endocytic pathway and in lysosome biogenesis. To show unequivocally that KIR2DL4 co-localizes with Rab5 and to address the specificity of this co-localization, gfp-tagged versions Rab4, Rab5, Rab7, and Rab11 were co-expressed with HA-tagged KIR2DL4 in 293T cells. The cells were stained with anti-HA antibodies followed by Alexa-568–conjugated secondary antibodies, and confocal images were obtained. As seen previously with endogenous Rab5 (Figure 5), all KIR2DL4-containing vesicles corresponded to Rab5-positive compartments (Figure 6). In contrast, overlap with compartments containing other Rab proteins was minimal (Figure 6). Thus, KIR2DL4 resides in Rab5-containing subcompartments of the early endocytic pathway.
Figure 6 KIR2DL4 Localizes to a Subset of Endosomes Containing Rab5
The 293T cells were transfected with HA-tagged KIR2DL4 and gfp-tagged versions of Rab4, Rab5, Rab7, and Rab11. Forty hours after transfection, cells were fixed and stained with anti-HA mAb, followed by Alexa-568–conjugated secondary antibodies to detect KIR2DL4. Single confocal sections are shown.
KIR2DL4 Binds Both Membrane and Soluble Forms of HLA-G
Soluble, recombinant Ig-fusion proteins of KIR2DL4 bound to cells expressing high levels of HLA-G [8,9]. However, one study could not reproduce these results [35], and soluble forms of HLA-G have not bound to NK cells [36] or to KIR2DL4 [35]. Failure to detect this interaction may reflect an intrinsic low affinity, as was previously noted for interactions between activating KIR and their HLA ligands [37,38]. The specificity of interaction between KIR2DL4 and HLA-G was explored further using mAbs to MHC class I and KIR2DL4 to block binding of KIR2DL4-Ig fusion proteins to cells expressing HLA-G. Background binding of KIR2DL4-Ig to 721.221 cells, which lack all classical HLA class I genes, was not inhibited by anti-class I mAbs but was reduced by anti-KIR2DL4 mAb 33 (Figure 7). In contrast, the enhanced binding of KIR2DL4-Ig to transfected 721.221 cells expressing membrane-bound HLA-G (221-G) was blocked by mAb DX17, a pan–class I antibody known to block KIR–HLA class I interactions [39], by the HLA-G–specific mAb G233, and by the KIR2DL4-specific mAb 33 (Figure 7). Thus, binding was blocked using specific mAbs to either KIR2DL4-Ig fusion protein or HLA-G expressed on 221-G cells.
Figure 7 Binding of KIR2DL4-Ig Fusion Proteins to HLA-G–Expressing Cells Is Blocked by Anti-KIR2DL4 and Anti-HLA Class I mAbs
(Top) 221-G cells were stained with mAb DX17 (pan–HLA class I mAb) and mAb G233 (HLA-G–specific mAb) (solid lines). Staining with secondary antibody alone is also shown (dotted lines).
(Bottom) The 221 and 221-G cells were incubated with 50 μg/ml KIR2DL4-Ig fusion protein in the presence of 20 μg/ml of either isotype-matched control Abs or mAbs specific for class I (DX17), HLA-G (G233), or KIR2DL4 (33). Cells were then stained with goat anti-human IgG1 secondary antibodies and assessed by flow cytometry. The data are expressed as mean fluorescence intensity (MFI).
We explored the possibility that soluble HLA-G may be a natural ligand of KIR2DL4. To date, evidence that HLA-G binds directly to KIR2DL4 is lacking. Soluble HLA-G produced in Escherichia coli and refolded with β2-microglobulin and the peptide KGPPAALTL was tested for endocytosis into KIR2DL4-containing compartments. HLA-Cw3 produced in E. coli and refolded with β2-microglobulin and peptide GAVDPLLAL was used as a specificity control. This refolded HLA-C preparation bound to KIR2DL2 by surface plasmon-resonance analysis [40] (and P. Sun, personal communication). Staining of 721.221 cells expressing HLA-Cw3 with F4/326 is also included (Figure 8A) to show that the mAb can detect HLA-Cw3 in fixed and permeabilized samples. Resting NK cells were loaded with either refolded HLA-G or refolded HLA-Cw3 for 2 h. Internalized soluble HLA-G, but not HLA-Cw3, was detected in intracellular vesicles (Figure 8A). It was difficult to show unambiguous co-localization of KIR2DL4 and HLA-G in resting NK cells, possibly because HLA-G bound to KIR2DL4 interfered with the staining of KIR2DL4 by mAb 33. Therefore, co-localization experiments were done with the NK cell line YTS that was stably transfected with gfp-tagged KIR2DL4. Endocytosed HLA-G was detected in intracellular vesicles that contained gfp-tagged KIR2DL4 (Figure 8B). Extensive co-localization of refolded HLA-G and KIR2DL4-gfp was also seen after incubation for 2 h with 293T-2DL4-gfp cells (Figure 8C). As seen with resting NK cells, refolded HLA-Cw3 was not detected in compartments containing KIR2DL4-gfp in 293T-2DL4-gfp cells.
Figure 8 Cell Surface Shed and Secreted, Soluble HLA-G Is Endocytosed into KIR2DL4-Containing Vesicles
(A) Endocytosis of soluble HLA-G in resting NK cells. The 221 cells and 221 cells transfected with HLA-Cw3 (221-Cw3) were fixed, permeabilized, and stained with mAb F4/326. Resting NK cells were incubated at 37 °C for 120 min with soluble, refolded HLA-C or HLA-G. Cells were then fixed, permeabilized, and stained with reagents to detect HLA-C (F4/326) or HLA-G (G233) as indicated.
(B) The NK cell line YTS-2DL4-gfp was loaded at 37 °C for 120 min with refolded HLA-G. Cells were fixed, permeabilized, and stained with mAb G233 to detect co-localization of soluble HLA-G with gfp-tagged KIR2DL4.
(C) Recombinant soluble molecules of HLA-G but not HLA-C are endocytosed into 293T-2DL4-gfp cells. Refolded HLA-G and HLA-C were incubated with 293T-2DL4-gfp cells for 2 h. Cells were then fixed, permeabilized, and stained with either mAb G233 (to detect endocytosed HLA-G; upper) or mAb F4/326 (to detect endocytosed HLA-C; middle).
(D) The 293T-2DL4-gfp cells were co-cultured with an equal number of 221 cells, 221 cells expressing transmembrane HLA-G (221-G), and 221 cells expressing a soluble isoform of HLA-G (221-sG) for 48 h. Adherent 293T-2DL4-gfp cells were fixed, permeabilized, and stained with mAb G233 followed by Alexa-568–conjugated secondary antibodies prior to acquisition of confocal images. Two 221-G cells are visible in the middle panel.
Soluble HLA-G can be produced by alternative splicing [41] and from cell surface–expressed HLA-G by proteolytic cleavage [42]. Transfected 721.221 cells expressing membrane-bound HLA-G (221-G) or secreted HLA-G (221-sG) were mixed with untransfected 293T cells and with 293T cells stably expressing KIR2DL4-gfp. After 48 h of co-culture, cells were gently washed and the adherent 293T-2DL4-gfp cells were fixed, permeabilized, and stained with the HLA-G–specific mAb G233. No internalization of HLA-G occurred after co-culture with 221 cells (Figure 8A). In contrast, co-culture with either 221-G or 221-sG resulted in internalization of HLA-G in KIR2DL4-containing vesicles. Note the vesicular staining of endocytosed HLA-G in 293T-2DL4-gfp cells in contrast to the cell surface staining of HLA-G on the 221-G cells included for comparison (Figure 8A, middle row, left panel). No HLA-G was detected in untransfected 293T cells co-cultured with 221-G and 221-sG cells (unpublished data). Therefore, soluble HLA-G produced by shedding from the cell surface or by secretion of an alternatively spliced form is internalized by KIR2DL4 into endosomes.
To further quantify the interaction of KIR2DL4 with soluble HLA-G, a recombinant, single-chain HLA-G (sHLA-G) expressed in Chinese hamster ovary (CHO) cells was used. This single-chain HLA-G contains an HLA-G–binding peptide (RLPKDFVDL), the extracellular domains (α1, α2, α3) of HLA-G, and β2-microglobulin all connected via linkers [43]. 293T-2DL4-gfp cells were incubated for 2 h with sHLA-G at 37 °C. Cells were fixed, permeabilized, and stained with mAb G233. Vesicular staining of sHLA-G, which co-localized with gfp-tagged KIR2DL4, was visible, similar to that seen after incubation with soluble anti-2DL4 mAb 33 (Figure 9A). The fluorescence intensity of endocytosed sHLA-G was quantified and compared to that of KIR2DL4-gfp in 293T cells. The expectation was that sHLA-G uptake should increase with higher expression levels of KIR2DL4-gfp. Red fluorescence in vesicles was compared to the green fluorescence of KIR2DL4-gfp in 38 individual cells (Figure 9B). The correlation between the level of expression of gfp-tagged KIR2DL4 and the extent of sHLA-G loading was highly significant (p ≤ 0.0001).
Figure 9 Endocytosis of Soluble HLA-G into 293T-2DL4-gfp Cells Is Blocked by Anti-KIR2DL4 mAb and by Soluble KIR2DL4
(A) Recombinant, soluble, sHLA-G at 50 μg/ml or mAb 33 (50 μg/ml) was incubated with 293T-2DL4-gfp cells. sHLA-G was also incubated together with 50 μg/ml mAb 33, 50 μg/ml KIR2DL1-Ig, or 50 μg/ml KIR2DL4-Ig, as indicated on the left. Cells were fixed, permeabilized, and stained with either Alexa-568–conjugated secondary antibodies to detect mAb 33 or anti-HLA-G mAb G233, as indicated. Individual confocal sections are shown.
(B) Uptake of sHLA-G into 293T-2DL4-gfp cells correlates with level of KIR2DL4 expression. Red fluorescence intensity of G233 staining and green fluorescence intensity of gfp were quantified in 38 individual cells and plotted on a log scale. A best-fit line was generated by linear regression analysis using EXCEL data analysis software.
(C) Ratio of red to green fluorescence was quantified for each loading condition as indicated. Average of 10 cells is shown, and standard deviation is shown as bars.
To test whether KIR2DL4 was involved in sHLA-G internalization, sHLA-G was incubated with 293T-2DL4-gfp cells in the presence of anti-KIR2DL4 mAb 33, which blocks KIR2DL4 binding to HLA-G on cells (Figure 7). Internalization of sHLA-G was blocked in the presence of mAb 33, whereas mAb 33 was still internalized in the presence of sHLA-G (Figure 9A and 9C). Similar results were obtained in YTS-2DL4-gfp cells (unpublished data). To test whether sHLA-G interacts directly with KIR2DL4, sHLA-G was loaded in the presence of soluble KIR2DL4-Ig fusion protein. Soluble KIR2DL1-Ig fusion protein with specificity for a subset of HLA-C alleles was used as a control. Soluble KIR2DL4-Ig prevented the uptake of sHLA-G into 293T-2DL4-gfp cells, whereas soluble KIR2DL1-Ig had no effect (Figure 9A and 9C). These results imply a direct interaction between soluble forms of HLA-G and KIR2DL4.
Engagement of KIR2DL4 by Soluble Ligand Induces a Unique Profile of Cytokine/Chemokine Secretion
We had shown previously that resting NK cells could be stimulated with anti-KIR2DL4 mAbs to secrete IFN-γ [7]. To get further information on the transcriptional response activated by this receptor, exploratory DNA microarray experiments were performed. Freshly isolated resting NK cells were stimulated either with control, isotype-matched, soluble antibodies or with KIR2DL4–specific soluble IgM mAbs for 5 h. Among the approximately 14,000 genes examined, all the genes that were up-regulated greater than 2-fold in resting NK cells from at least two of three independent donors are listed in Figure 10A. Interestingly, only a small subset of genes showed up-regulation. They include an array of proinflammatory/proangiogenic cytokines, such as IL-6, IL-1β, TNF-α, and IL-23 (comprised of the IL-23α and IL-12β subunits), and chemokines, such as IL-8, MIP-3α, MIP-1δ, MIP-1α, and MIP-2β.
Figure 10 Binding of Soluble mAb to KIR2DL4 Up-regulates Multiple Genes in Resting NK Cells
(A) All genes exhibiting greater than 2-fold up-regulation in microarray experiments with resting NK cells of at least two of three different individuals are listed.
(B) Semiquantitative RT-PCR was performed on total RNA isolated at different time points from resting NK cells stimulated with either control IgM mAbs or anti-KIR2DL4 IgM mAbs 36 and 64.
(C) Time course of TNF-α, IL-1β, and IFN-γ secretion by resting NK cells stimulated with anti-KIR2DL4 mAbs 36 and 64 (open symbols) or control IgM mAbs (closed symbols). Squares and triangles represent data obtained from two different donors. Protein secretion was detected by ELISA.
The KIR2DL4-induced expression of the majority of these genes was confirmed by semiquantitative RT-PCR analysis of RNA isolated at different timepoints after stimulation of resting NK cells by either control or anti-KIR2DL4 mAbs (Figure 10B). Such a time course analysis highlighted differences in the kinetics of responses. Transcription of the GRO3 and COX-2 genes was up-regulated transiently, whereas all of the other genes examined showed sustained expression up to the 16 h timepoint. Responses detectable within the first two hours included TNF-α, IL-1β, IL-8, MIP-3α, and GRO-3. In contrast, transcription of IL-6, IL-12β(p40), IL-23α(p19), COX-2, and MARCKS increased only after 2 h. Previous work [7] and Figure 1B show that stimulation by KIR2DL4 induces IFN-γ production. However, IFNG gene transcription was up-regulated only approximately 1.5-fold in the microarray studies and therefore did not score as positive. RT-PCR analysis reconciled this apparent discrepancy, as induction of IFNG transcription occurred only 8 h after stimulation. The time course of IFN-γ protein secretion supported the transcriptional data (Figure 10C). There is delayed secretion of IFN-γ compared to other cytokines such as TNF-α and IL-1β. IFNG gene transcription is a late response, which requires de novo protein synthesis, as it was inhibited by cycloheximide (Y. T. Bryceson, unpublished observation). Thus, exploratory microarray analysis using anti-KIR2DL4 mAbs to stimulate resting NK cells revealed the up-regulation of a select number of proinflammatory/proangiogenic cytokine and chemokine genes. However, it is possible that stimulation of KIR2DL4 by a natural ligand, such as soluble HLA-G, induces a different response.
To test whether a natural ligand of KIR2DL4 could confirm the observed transcriptional responses induced by anti-KIR2DL4 mAbs, protein expression was tested after incubation of resting NK cells with soluble HLA-G. Resting NK cells were incubated with sHLA-G produced in mammalian cells and tested for secretion of a number of cytokines and chemokines. As specificity control for activation by soluble HLA-G, mAb G233, which blocks the interaction between HLA-G and KIR2DL4 (Figure 7), was used. As shown in Figure 11, soluble KIR2DL4-specific mAb and sHLA-G activated the secretion of cytokines and chemokines, for which transcriptional activation had been observed in DNA microarrays. Responses induced by sHLA-G were inhibited by anti–HLA-G mAb G233 but not by an isotype-matched control IgG. Among the three donors tested, quantitative variation in protein secretion was observed. To take such variability into account, the amount of cytokine/chemokine induced by sHLA-G was compared to that induced by anti-KIR2DL4 mAb for each donor separately. For six of eight cytokines/chemokines tested, sHLA-G induced at least 50% of the amount induced by anti-KIR2DL4 mAb (Figure 11). To further evaluate the specificity of responses, secretion of IL-5, IL-12 (which shares the IL-12β p40 subunit with IL-23), and IL-13 was also determined. NK cells can produce IL-5 and IL-13 under certain conditions but have not been reported to produce IL-12. No secretion of any of these three cytokines was induced by anti-KIR2DL4 and by sHLA-G (unpublished data). Therefore, soluble HLA-G activates a qualitatively and quantitatively similar cytokine/chemokine response as KIR2DL4-specific mAb.
Figure 11 Cytokine/Chemokine Synthesis Induced by Soluble HLA-G in Resting NK Cells
Resting NK cells (5 × 105 cells/well) from three different donors were incubated separately for 48 h with either soluble, control IgM Ab (cIg), soluble anti-KIR2DL4 IgM mAb 36 (anti-2DL4), soluble HLA-G produced in CHO cells (sHLA-G), or beads coated with control anti-HA IgG1 mAb 16B12 (cIg) or with anti-CD16 IgG1 mAb 3G8 (3G8), as indicated. sHLA-G was used together with control IgG2a mAb (cIg) or with anti-HLA-G IgG2a mAb G233 (G233), as indicated. Secretion of the cytokines/chemokines listed on the left is given in pg/ml for each donor separately. Secretion induced by sHLA-G and by Ab-coated beads was compared to that induced by anti-KIR2DL4 mAb for each donor separately and is expressed as a percentage of the anti-KIR2DL4 response. The graphs represent the average ± standard deviation from three experiments.
To test whether the KIR2DL4-induced response was unique, responses elicited by CD16 cross-linking were evaluated in parallel. Anti-CD16 mAb attached to beads induced secretion of only a subset of the cytokines/chemokines induced by KIR2DL4 engagement (Figure 11). Very little or no secretion of IL-6, IL-8, and IL-23 was induced by CD16 cross-linking. Therefore, activation of KIR2DL4 by soluble ligand induces secretion of a unique profile of proinflammatory/proangiogenic mediators, which may be relevant to functional interaction between KIR2DL4 on uterine NK cells and HLA-G produced by trophoblast cells.
KIR2DL4 Induces IL-8 Secretion in 293T Cells
The 293T cells are capable of producing IL-8, as TLR-4 agonists induced IL-8 secretion in TLR-4–transfected 293T cells [44]. As KIR2DL4 engagement in resting NK cells resulted in secretion of IL-8, the ability of KIR2DL4 to induce IL-8 secretion in 293T cells was tested. Transfected 293T cells expressing receptors 2B4 and gp49B were used for comparison. Expression of KIR2DL4, but not 2B4 or gp49B, in 293T cells induced IL-8 secretion (Figure 12A). As 293T cells did not stain with anti-HLA-G mAb G233 (see Figure 8D and unpublished data), IL-8 secretion may be induced by KIR2DL4 expression alone, independent of ligand binding. IL-8 expression in KIR2DL4-transfected 293T cells was used as a convenient assay to test mutants of KIR2DL4 (Figure 12B) for their ability to signal in 293T cells. A KIR2DL4 molecule lacking the arginine residue in the transmembrane region [2DL4(RY-GT)] was tested. This arginine may contribute to KIR2DL4 pairing with the FcRI γ chain [15], similar to other activating receptors that carry a charge in the transmembrane region. Second, a truncated KIR2DL4 (2DL4-TR), lacking its cytoplasmic tail, was tested. Confocal microscopy showed that both of these mutants were targeted to endocytic vesicles in 293T cells, similar to wild-type KIR2DL4 (unpublished data), showing that the cytoplasmic tail of KIR2DL4 is not required for endosomal targeting. Secretion of IL-8 was obtained with the 2DL4(RY-GT) mutant (Figure 12C), showing that the charged amino acid in the transmembrane region is not required for signaling. In contrast, the 2DL4-TR mutant did not signal for IL-8 secretion (Figure 12C), despite its localization in endosomes, pointing to a role for the transmembrane region and/or cytoplasmic tail in signaling. These results suggested that the natural 9A allele of KIR2DL4, which encodes a receptor with a truncated cytoplasmic tail, would not signal for IL-8 secretion. A frameshift mutation was introduced in wild-type KIR2DL4 cDNA in order to produce a molecule identical to the KIR2DL4 9A allotype: consistent with the truncated mutant, expression of the 9A mutant in 293T cells did not induce IL-8 secretion (unpublished data).
Figure 12 IL-8 Secretion Induced by KIR2DL4 Is Independent of the Transmembrane Arginine Residue and Requires Trafficking to Intracellular Vesicles
(A) Expression of KIR2DL4 in 293T cells induces IL-8 secretion. The 293T cells were transfected with HA-tagged 2B4, gp49B, and KIR2DL4. After 48 h, culture supernatants were tested for IL-8 by ELISA.
(B) Schematic representation of KIR2DL4 variants used in this study. Receptor localization was determined by confocal analysis of immunofluorescence staining of 293T cells transfected with the indicated constructs for 48 h.
(C) IL-8 secretion induced by KIR2DL4 mutants in 293T cells. The 293T cells were transfected with HA-tagged KIR2DL4, KIR2DL4(RY-GT), and KIR2DL4-TR. After 48 h, culture supernatants were tested for IL-8 by ELISA. Transfection efficiency was verified by monitoring HA-positive cells by confocal microscopy.
(D) Cell surface–expressed gp49B/2DL4 chimera does not induce IL-8 secretion. The 293T cells were transfected with HA-tagged 2B4, KIR2DL4, and gp49B/2DL4 chimera, as indicated. After 48 h, beads coated with control IgG (solid bars) or anti-HA mAb (hatched bars) were added at four beads per cell. Twelve hours later, culture supernatants were tested for IL-8 secretion. The percentage of 293T cells expressing receptor at the cell surface was monitored by HA staining and flow cytometry, and was as follows: vector control, 2%; 2B4, 39%; KIR2DL4, 9%; gp49B, 31%; gp49B/2DL4, 26%.
Another mutant was informative with respect to the signaling function of KIR2DL4. A chimeric, HA-tagged receptor containing the extracellular regions of gp49B fused to the transmembrane and cytoplasmic domains of KIR2DL4 was expressed at the surface of 293T cells and did not traffic to endosomes (Figure 12B), indicating that the luminal domain of KIR2DL4 is required for proper targeting. No IL-8 secretion was induced by this chimeric gp49B/2DL4, even after stimulation for 12 h with beads coated with anti-HA antibodies (Figure 12D). Cross-linking HA-tagged 2B4 with the same beads on 2B4-transfected 293T cells resulted in a very modest increase in IL-8 secretion. These results suggest that endosomal localization of KIR2DL4 is required for its signaling function.
Discussion
The NK cell receptor KIR2DL4 emerges as an activation receptor specific for soluble HLA-G. Constitutive endocytosis of KIR2DL4 occurs in freshly isolated, resting NK cells, in IL-2–activated NK cells, in NK cell lines, and in 293T cells transfected with KIR2DL4. At steady state, most of KIR2DL4 is intracellular, which explains the difficulty in detecting cell surface receptor by flow cytometry. Endocytosis explains also the unusual property of KIR2DL4 to induce cytokine and chemokine secretion in response to soluble mAbs and to soluble HLA-G, in the absence of further cross-linking, but not in response to solid-phase mAbs (either plate-coated or bead-bound). Therefore, lack of detectable KIR2DL4 at the cell surface should not be taken as evidence for a lack of function. Endocytosed KIR2DL4 showed extensive co-localization with Rab5, a GTPase that marks a subset of early endosomes.
Endosomes have emerged recently as a specialized signaling compartment [45,46]. Some endocytosed receptors, such as the epidermal growth factor receptor and Toll-like receptor 9, signal from endosomes [45–47]. It is thought that endocytic vesicles carry signaling platforms where endocytosed receptors link with downstream signaling molecules and provide sustained signaling [45]. Our data strongly support signal transduction by KIR2DL4 in endosomes. First, KIR2DL4 is actively internalized into endosomes, where most of KIR2DL4 resides at steady state. Second, solid-phase Abs to KIR2DL4 did not induce IFN-γ secretion, indicating that cross-linking KIR2DL4 at the cell surface does not result in signaling. Third, expression of KIR2DL4 and of KIR2DL4 mutants in 293T cells resulted in secretion of IL-8 only for those receptors that were targeted to endosomes. A chimeric receptor with the transmembrane region and cytoplasmic tail of KIR2DL4 that remained at the cell surface did not activate IL-8 secretion, even after cross-linking at the cell surface. Extensive co-localization of endocytosed KIR2DL4 with a subset of early endosomes that carry the small GTPase Rab5 is interesting, as Rab5 was recently assigned a role in endosomal signaling by the epidermal growth factor receptor and by Toll-like receptor 9 [45,46]. The signaling pathway triggered by KIR2DL4 requires further study. Ligand-bound KIR2DL4 must somehow be distinguished from free, endogenous KIR2DL4 in NK cells, as resting NK cells do not respond to KIR2DL4 until soluble ligand is added. This distinction may not be made in transfected 293T cells, which appear to respond to KIR2DL4 in the absence of a ligand. Thus, KIR2DL4 signaling is subject to tight regulation in NK cells.
Signaling by KIR2DL4 in 293T cells also implies that immunoreceptor tyrosine-based activation motif (ITAM)-containing subunits, such as the FcRI γ chain, are unlikely to be necessary. This conclusion is further supported by the lack of requirement for the arginine residue in the transmembrane region of KIR2DL4 for signaling in 293T cells. Unlike other NK cell activation receptors with charged residues in the transmembrane region, such as NKG2D and NKp46, which require a partner chain for transport out of the endoplasmic reticulum, KIR2DL4 is readily transported to the cell surface and to endosomes, even in the absence of a known adapter. The position of the arginine at the fourth amino acid position from the N-terminus of the transmembrane region, rather than deeper into the lipid bilayer, suggests that membrane insertion may not be compromised [48]. However, induction of cytotoxicity by IL-2–activated NK cells through KIR2DL4 in a redirected lysis assay had been dependent on the arginine-tyrosine motif in the transmembrane region of KIR2DL4 [8]. It is therefore likely that KIR2DL4 can signal in different ways, for cytokine secretion independent of an arginine-mediated association (this study), and for cytotoxicity through an arginine-mediated association with the FcRI γ chain [8,15]. The biological context in which KIR2DL4-mediated cytotoxicity would occur is not clear. On the other hand, stimulation of a proinflammatory or proangiogenic response by soluble HLA-G could be a useful property in various physiological contexts. In particular, our study supports a new model for the role of NK cells in early pregnancy.
A ligand of KIR2DL4 is the nonclassical MHC class I molecule HLA-G, which is produced naturally in soluble form by alternative splicing [41,49] and by proteolytic cleavage of membrane-bound HLA-G [42]. In vitro studies have shown that HLA-G induces apoptosis of T cells, inhibition of alloproliferation of T cells and unfractionated mononuclear cells, and inhibition of NK cell cytotoxicity [19,42,50,51]. In contrast, HLA-G stimulated proliferation and cytokine secretion in uterine NK cells [29]. In this study, four different soluble forms of HLA-G have been internalized into KIR2DL4-containing endosomes. HLA-G shed from the cell membrane and the naturally secreted form of HLA-G, each produced in transfected 721.221 cells, were endocytosed into KIR2DL4-containing compartments of co-cultured cells. In addition, recombinant HLA-G, produced in E. coli and refolded with β2-microglobulin and peptide, was also internalized, as was an sHLA-G produced in transfected CHO cells. Competition for HLA-G uptake with a soluble, recombinant KIR2DL4-Ig fusion protein established, for the first time, a direct interaction between KIR2DL4 and HLA-G. So far, evidence for KIR2DL4 binding to HLA-G had been indirect. Binding of soluble KIR2DL4-Ig fusion proteins to cells expressing high levels of HLA-G [8,9] had been detected, but binding of soluble HLA-G tetramers to NK cells or soluble HLA-G to KIR2DL4 in surface plasmon resonance studies had not been detected [35,36]. The detection of KIR2DL4–HLA-G interaction may have been possible here due to accumulation of soluble HLA-G into KIR2DL4-containing endosomes.
In pregnancy, tolerance of the semiallogeneic fetus by the maternal immune system has long presented an immunological paradox [52,53]. To avoid rejection of the fetus, several factors must regulate maternal responsiveness to fetal alloantigens. Regulatory mechanisms may include inhibitory receptors on lymphocytes that bind ligands on fetal cells [5,54] and suppression of T lymphocytes by tryptophan catabolism via indoleamine-2,3-dioxygenase activity [55]. An alternate and nonexclusive view is that interactions of fetal cells with immune cells at the maternal–fetal interface play a role in proper implantation by shaping vascular adaptations of the surrounding maternal tissue via cytokine secretion [25,56]. Uterine tissue in early human pregnancy is characterized by extensive vascular remodeling, by invasion of HLA-G–expressing trophoblast cells, and by an abundance of maternal NK cells. Engagement of KIR2DL4 by soluble ligand induced a profile of gene transcription and protein secretion that is consistent with a proangiogenic response. The controversial role of HLA-G expression on extravillous cytotrophoblast cells—be it for inhibition or activation of maternal NK cells—may be settled in favor of activation of NK cells through soluble HLA-G for the purpose of vascular remodeling. Successful implantation and placentation require vascular endothelial growth factor, an angiogenic growth factor that is secreted by trophoblast cells throughout gestation [57]. The products of several genes up-regulated on NK cells by KIR2DL4 engagement, including TNF-α, IL-1β, and IFN-γ, induce vascular endothelial growth factor production in trophoblast cells, thereby influencing uterine angiogenesis [58]. Another gene induced by KIR2DL4 in NK cells, COX-2, is a rate-limiting enzyme in prostaglandin biosynthesis, which is required for uterine angiogenesis during implantation in mice [59]. In addition, IL-8 is known to induce neovascularization in tumor and viral models [60,61]. Secretion of IL-23, which is encoded by IL12B and IL23A genes, was also up-regulated by KIR2DL4 engagement with soluble ligands. Selective induction of these two genes occurs in mouse uterine NK cells at gestation day 6 [62]. In summary, the spectrum of proangiogenic mediators that are induced in resting NK cells by soluble HLA-G and by anti-KIR2DL4 mAb supports a role for KIR2DL4 in facilitating implantation and neovascularization during pregnancy. Our data, together with evidence that secretion of soluble HLA-G by in vitro fertilized human embryos is associated with a higher pregnancy rate [63,64] and that reduced levels of soluble HLA-G during early pregnancy correlate with a higher incidence of preeclampsia [65], suggest that KIR2DL4 has a positive role in reproductive success.
The description of soluble HLA-G expression in other cell types and tissues, such as erythroblasts during erythropoiesis [66], suggests the interesting possibility that a KIR2DL4-mediated signal in NK cells could serve a proinflammatory or proangiogenic function in other contexts as well. In this study, we show that constitutive internalization of KIR2DL4 results in uptake of soluble HLA-G and accumulation into endocytic compartments where the bulk of KIR2DL4 resides. Binding of soluble agonist by KIR2DL4 has the potential advantage of promoting specific NK cell responses without the complicating contribution of multiple receptor–ligand interactions that occur during cell–cell contacts, which engage a wide array of activating and inhibitory NK cell receptors.
Materials and Methods
Cell lines and culture
The 293T cells were obtained from
ATCC (American Type Culture Collection, Manassas, Virginia, United States); YTS cells expressing an ecotropic receptor were a gift from G. Cohen (Harvard Medical School, Boston, Massachusetts, United States); NKL cells were a gift from M. Robertson (Indiana University School of Medicine, Indianapolis, Indiana, United States); and 221-HLA-Cw3 cells were a gift from J. Gumperz and P. Parham (Stanford University, Palo Alto, California, United States). The 221-HLA-G cells that do not express HLA-E [54] (referred to here as 221-G) cells were a gift from M. Lopez-Botet (University Pompeu Fabra, Barcelona, Spain); 221 cells transfected with the soluble form of HLA-G (221-sG) have been described previously [67]. NKL cells were cultured in RPMI 1640 medium containing 10% fetal calf serum, 1% l-glutamine, 1% sodium pyruvate, and 200 U/ml of recombinant IL-2 (National Cancer Institute–FCRDC, Frederick, Maryland, United States). All other cell lines were cultured in Iscove's medium supplemented with 10% fetal calf serum and 1% l-glutamine.
Human polyclonal NK cells were isolated from peripheral blood lymphocytes (PBL) using the MACS NK cell negative isolation kit (Miltenyi Biotech, Auburn, California, United States). NK cells were greater than 95% CD3− and CD56+. These freshly isolated NK cells (“resting NK”) were cultured in Iscove's medium containing 10% human serum without added IL-2 or feeders. “Activated NK” refers to NK cell populations from a given donor that was expanded with autologous irradiated PBL as feeders and IL-2, as described previously [7].
The 293T cells were transiently transfected with LipofectAMINE 2000 according to the supplier's instructions. The 293T-2DL4-gfp cells are 293T cells that were stably transfected, using LipofectAMINE 2000, with a KIR2DL4 cDNA fused to gfp in the vector pBABE (puro)CMV+ using SalI and BamH1 sites. Transfected cells were selected and maintained in Iscove's medium containing puromycin at a concentration of 2 μg/ml. YTS-2DL4-gfp refers to YTS cells transduced as described [68] to stably express the KIR2DL4-gfp fusion protein.
Antibodies and reagents
mAbs to KIR2DL4, namely, 33 (IgG1), 36, and 64 (IgM), were generated in our laboratory [7]. To prepare Fab fragments, the 33 antibody was produced in mouse ascites fluid and the IgG1 was precipitated with 40% saturated ammonium sulfate and purified from the precipitate by ion-exchange chromatography on DEAE-Sepharose using an NaCl gradient. The IgG1 was then cleaved with papain at an enzyme-to-antibody ratio (wt/wt) of 1:20 in 100 mM Tris-HCl, 150 mM NaCl, 2 mM mercaptoethanol, 2 mM NaEDTA (pH 7.1) at 37 °C for 5 h. The reaction was quenched by addition of fresh iodoacetamide to a final concentration of 7.5 mM. Fab was purified from the digestion mixture by DEAE-Sepharose chromatography. The Fab fractions were pooled and concentrated in 20 mM Tris-HCl (pH 7.4). The integrity and purity of the Fab fragments were confirmed by SDS-PAGE analysis. The control Fab portion of IgG1 mAb 2A8, specific for malaria protein PV525, was obtained from K. Singh (National Institute of Allergy and Infectious Diseases [NIAID], National Institutes of Health [NIH], Rockville, Maryland, United States).
Antibodies to CD16 (3G8), Rab5 (610457), and EEA-1 (BD61045) were purchased from Becton Dickinson (Franklin Lakes, New Jersey, United States) Antibody to M6PR was purchased from Affinity Bioreagents (Golden, Colorado, United States), and antibody to perforin was purchased from Endogen Pierce (Rockford, Illinois, United States). An affinity-purified polyclonal rabbit antibody specific for the carboxy-terminal end of KIR2DL1 has been described [69]. MAb DX17 specific for MHC class I was a gift from L. Lanier (University of California San Francisco, California, United States), and mAb G233 specific for HLA-G was a gift from A. Moffett (Cambridge University, Cambridge, United Kingdom). MAb F4/326, specific for HLA-C alleles, was a gift from S.-Y. Yang (New York University, New York, New York, United States). Anti-HA mAb was obtained from Cell Signaling (Beverly, Massachusetts, United States), and all isotype control mAbs were obtained from Sigma (St. Louis, Missouri, United States). All secondary goat anti-mouse antibodies conjugated to either Alexa-Fluor 488 or Alexa-Fluor 564 were obtained from Molecular Probes (Eugene, Oregon, United States).
Rab5-gfp, Dynamin Egfp-WT, and Dynamin Egfp-K44A mutant constructs were obtained from J. Bonifacino (National Institute of Child Health and Human Development, NIH). Constructs encoding Rab4-gfp, Rab7-gfp, and Rab11-gfp cloned in pEgfp-C and C3 vectors (Clontech, Palo Alto, California, United States) were obtained from M. Zerial (Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, Germany). These constructs have been previously described [70]. Expression constructs for transfection into 293T cells consisted of cDNAs for 2B4 [71], mouse gp49B [72], KIR2DL4, and KIR2DL4 variants cloned into the pDisplay vector (Stratagene, La Jolla, California, United States) that introduces an HA-tag at the amino-terminus. The KIR2DL4(RY-GT) construct was generated by site-directed mutagenesis of R246G and Y247T. The truncated KIR2DL4-TR was engineered by introduction of a stop codon at position K269 and contains the amino acids CSKK at the carboxy-terminus. The gp49B/2DL4 chimera was engineered using PCR to fuse the extracellular domain of gp49B ending at S226 with the transmembrane and tail of KIR2DL4 beginning at position S224. The amino acids at the boundary are DQSS/SWPS.
KIR2DL1-Ig and KIR2DL4-Ig fusion proteins have been described previously [8]. The recombinant, single-chain soluble HLA-G product has been described [43] and was obtained from Organon NV (Oss, the Netherlands). Bacterially expressed and refolded HLA-G was produced as described for HLA-E [73]. Refolded HLA-C has been described [40] and was obtained from P. Sun (NIAID, NIH).
Immunostaining and confocal microscopy
Cells were allowed to settle on poly-L-lysine–coated two-well culture slides (BioCoat, BD, Bedford, Massachusetts, United States) for 30 min prior to fixation in PBS/4% paraformaldehyde. Nonspecific sites were saturated and cells were permeabilized for 30 min with PBS/10% normal donkey serum/0.5% Triton X-100. Cells were stained with the relevant primary antibody for 1 h and revealed with goat anti-mouse or goat anti-rabbit Alexa-Fluor–conjugated secondary antibodies in PBS/3% normal donkey serum/0.5% Triton X-100. In some cases, cells were stained with directly conjugated antibodies (e.g., 33-Cy3). Cells were washed and mounted in slides using the Prolong Anti-fade kit (Molecular Probes, Eugene, Oregon, United States).
Images were processed using a confocal laser-scanning microscope (Axiovert 200M LSM 510 META; Zeiss, Jena, Germany) fitted with a 1.3 Oil DIC Plan-Neofluar ×63 objective was used. Images were acquired using channel mode, multitrack acquisition with the main beam splitters HFT 488/543 for the excitation of gfp and Alexa-Fluor 568. Filter sets of BP 505–530 for gfp and LP560 for Alexa-Fluor 568 were used. Parameters were adjusted to yield scan control of fixed pixel density at 512 × 512 pixels, 12-bit pixel depth, and a pinhole size of 90 μm. To quantify the co-localization of KIR2DL4 with loaded soluble HLA-G, the LSM 510 imaging examiner software was used.
Electron microscopy
The NK cell lines NKL and YTS-2DL4-gfp were loaded with KIR2DL4-specific Ab 33 for 2 h. Cells were rinsed in HBSS before fixation with periodate-lysine-paraformaldehyde (PLP) fixative containing 0.25% glutaraldehyde for 2 h at room temperature. Cells were permeabilized with PBS containing 0.01% saponin for 5 min at room temperature and incubated for 1 h with peroxidase-conjugated F(ab′)2 sheep anti-mouse IgG (Jackson Immunoresearch Laboratories, West Grove, Pennsylvania, United States) in PBS containing 0.01% saponin. Cells were rinsed in PBS and fixed in 1.5% glutaraldehyde in 0.1 M sodium cacodylate (pH 7.0) plus 5% sucrose for 1 h. Cells were then rinsed with 50 mM Tris-HCl (pH 7.4) plus 7.5% sucrose before development with Immunopure Metal–enhanced 3,3′-diaminobenzidine substrate (Pierce, Rockford, Illinois, United States). Cells were then rinsed three times with 50 mM Tris-HCl, (pH 7.4) plus 7.5% sucrose before fixation in 4% paraformaldehyde/2.5%glutaraldehyde in 100 mM sodium cacodylate buffer (pH 7.4). Cells were postfixed in 1% OsO4/1%K3Fe(CN)6, washed with H2O, dehydrated in a graded ethanol series, and embedded in Spurr's resin. Thin sections were cut with an MT-7000 ultramicrotome (RMC, Tucson, Arizona, United States). Samples were examined on a Hitachi H7500 TEM at 80 kV and images were captured with a CCD camera (Advanced Microscopy Technologies, Danvers, Massachusetts, United States).
Cellular assays
For cytokine/chemokine assays, resting NK cells were incubated with soluble mAbs at 10 μg/ml, plate-coated antibodies (5 μg Ab/100 μl/well of a 96-well plate) or bead-bound antibodies (1 μg antibody/4 × 107 beads; four beads per cell). For soluble HLA-G stimulation of resting NK cells, recombinant single-chain HLA-G (30 μg/ml) was used together with either control IgG2a mAb or anti-HLA-G mAb G233 at 90 μg/ml. After the indicated timepoints, supernatants were removed and tested for the presence of cytokines by ELISA. ELISA kits for IL-1β, TNF-α, IL-6, IL-23, MIP-3α, IL-12, IL-5, and IL-13 (R & D Systems, Minneapolis, Minnesota, United States), IFN-γ (Pierce), and IL-8 (Biosource, Camarillo, California, United States) were used according to manufacturers' instructions.
For endocytosis assays, resting NK cells, 293T-2DL4-gfp, or YTS-2DL4-gfp cells were loaded with mAb 33 (10 μg/ml) or with 50 μg/ml of soluble HLA-G for the indicated time periods at 37 °C. Cells were then washed in PBS and plated onto chamber slides for immunofluorescence staining. For the blocking studies, mAb 33, KIR2DL4-Ig, or KIR2DL1-Ig (50 μg/ml) was added at the same time as soluble HLA-G.
Binding assays using Ig-fusion proteins were done essentially as described previously [8,74]. For blocking studies, mAbs to MHC molecules were added to the cells 20 min prior to the addition of KIR2DL4-Ig fusion proteins. KIR2DL4-specific mAb 33 was preincubated with KIR2DL4-Ig fusion proteins for 20 min prior to addition to the 221 cells.
For microarray analysis, NK cells were stimulated with either isotype control IgM antibodies (Sigma) or anti-2DL4 mAbs 36 and 64 for 5 h at 37 °C. Total RNA was purified from cell pellets by TriZOL/chloroform extraction followed by column purification (RNeasy Midi Kit; Qiagen, Valencia, California, United States). The integrity of isolated RNA was examined by denaturing TAE agarose gel electrophoresis. Direct labeling of first-strand cDNA with Cy3-dUTP or Cy5-dUTP (Amersham Pharmacia, Piscataway, New Jersey, United States), respectively, was performed in oligo(dT) primed reverse transcription reactions (SuperScript II Reverse Transcriptase; Invitrogen, Carlsbad, California, United States). The remaining RNA was degraded, and Cy3- and Cy5-labeled probes were combined and purified by buffer exchange on Vivaspin 500 columns (10K cutoff; Vivascience, Hannover, Germany). Labeled cDNA was hybridized to 14,000 oligonucleotide arrays (NIAID Microarray Facility, NIH). Arrays were scanned on an Axon GenePix 4000B scanner (Axon Instruments, Foster City, California, United States). Image analysis was carried out using GenePix (Axon Instruments) and mAdb (National Cancer Institute; http://madb.nci.nih.gov) software.
To investigate transcription of microarray candidate genes by RT-PCR, specific primers were designed to amplify regions of approximately 200 bp that spanned exon-intron boundaries. Total RNA was isolated by TriZOL/chloroform extraction followed by column purification (RNeasy Mini Kit; Qiagen). First-strand cDNA was synthesized with oligo(dT) primed SuperScript II reverse transcriptase (Invitrogen) from 600 ng of total RNA, according to the manufacturers' protocol. PCRs were performed with Taq polymerase according to the manufacturers' protocol in 25 μl reactions. First-strand cDNA template corresponding to 10 ng total RNA per reaction were used in touch-down PCRs with five cycles (annealing temperature 70 °C) followed by 20 to 35 cycles (annealing temperature 65 °C). Primers to CCL20 (forward 5′-
CCAATGAAGGCTGTGACATCAATG-3′ and reverse 5′-
ACCTCCAACCCCAGCAAGGTTC-3′), CXCL3 (forward 5′-
TGCAGACACTGCAGGGAATTCAC-3′ and reverse 5′-
CTTCTCTCCTGTCAGTTGGTGCT-3′), GAPD (forward 5′-
GGCATGGACTGTGGTCATGAG>-3′ and reverse 5′-
TGCACCACCAACTGCTTAGC-3′), IFNG (forward 5′-
AGCGGATAATGGAACTCTTTTCTTAG-3′ and reverse 5′-
AAGTTTGAAGTAAAAGGAGACAATTTGG-3′), IL1B (forward 5′-
TCCAGGGACAGGATATGGAGCAA-3′ and reverse 5′-
GCTTTTCCATCTTCTTCTTTGGGT-3′), IL6 (forward 5′-
TCGGTACATCCTCGACGGCATC-3′ and reverse 5′-
ATACCTCAAACTCCAAAAGACCAG-3′), IL8 (forward 5′-
CTGATTTCTGCAGCTCTGTGTGA-3′ and reverse 5′-
GGGTCCAGACAGAGCTCTCTTC-3′), IL12B (forward 5′-
ACATTCTGCGTTCAGGTCCAGG-3′ and reverse 5′-
CTCCAAATTTTCATCCTGGATCAG-3′), IL23A (forward 5′-
ACTCAGTGCCAGCAGCTTTCAC>-3′ and reverse 5′-
CAGACCCTGGTGGATCCTTTGC-3′), MARCKS (forward 5′-
GAGCAAGCTTTTGTGAGATAATCG-3′ and reverse 5′-
GAATGATTTGAGATGGGATCTGTG-3′), PTGS2 (forward 5′-
AATCATTCACCAGGCAAATTGCTG-3′ and reverse 5′-
CTTCCAACTCTGCAGACATTTCC-3′), and TNF (forward 5′-
CTCTTCTGCCTGCTGCACTTTGG-3′ and reverse 5′-
CCATTGGCCAGGAGGGCATTGG-3′) were used.
PCR products were separated by horizontal agarose gel electrophoresis, stained with ethidium bromide, visualized with UV light, and photographed. Black/white images were inverted with Adobe Photoshop software (Adobe Systems, San Jose, California, United States).
Supporting Information
Accession Numbers
The GenBank (http://www.ncbi.nlm.nih.gov/Genbank/) accession numbers of the cDNAs used in this paper are KIR2DL4 (DQ266438) and HLA-G1 (M90683).
We thank Mary Peterson for YTS-2DL4-gfp cells and for nucleotide sequencing; David Dorward for electron microscopy; M. Wilson for DNA microarrays; H. Young for discussions; J. Bonifacino, L. Lanier, M. Lopez-Botet, A. Moffett, K. Singh, P. Sun, and M. Zerial for reagents; and the Department of Transfusion Medicine at NIH for blood samples. YTB was supported by the NIH–Karolinska Institute Graduate Partnership Program. This research was supported in part by the Intramural Research Program of the NIAID, NIH.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. SR, YTB, and EOL conceived and designed the experiments. SR, YTB, SPK, and AV performed the experiments. SR, YTB, DEG, AV, IJ, and EOL analyzed the data. SR, YTB, SPK, DEG, AV, and IJ contributed reagents/materials/analysis tools. SR and EOL wrote the paper.
Citation: Rajagopalan S, Bryceson YT, Kuppusamy SP, Geraghty DE, van der Meer A, et al. (2006) Activation of NK cells by an endocytosed receptor for soluble HLA-G. PLoS Biol 4(1): e9.
Abbreviations
Abantibody
gfpgreen fluorescence protein
HLAhuman leukocyte antigen
HRPhorseradish peroxidase
IFNinterferon
ILinterleukin
KIRkiller cell immunoglobulin-like receptor
LAMP-1lysosome-associated membrane protein 1
M6PRmannose 6-phosphate receptor
mAbmonoclonal antibody
MHCmajor histocompatibility complex
MIPmacrophage inflammatory protein
NKnatural killer
TNFtumor necrosis factor
==== Refs
References
Moretta L Moretta A Unravelling natural killer cell function: Triggering and inhibitory human NK receptors EMBO J 2004 23 255 259 14685277
Lanier LL NK cell recognition Annu Rev Immunol 2005 23 225 274 15771571
Cerwenka A Lanier LL Natural killer cells, viruses, and cancer Nat Rev Immunol 2001 1 41 49 11905813
Gasser S Orsulic S Brown EJ Raulet DH The DNA damage pathway regulates innate immune system ligands of the NKG2D receptor Nature 2005 436 1186 1190 15995699
Moffett-King A Natural killer cells and pregnancy Nat Rev Immunol 2002 2 656 663 12209134
Wagtmann N Biassoni R Cantoni C Verdiani S Malnati M Molecular clones of the p58 natural killer cell receptor reveal Ig-related molecules with diversity in both the extra- and intracellular domains Immunity 1995 2 439 449 7749980
Rajagopalan S Fu J Long EO Cutting edge: Induction of IFN-gamma production but not cytotoxicity by the killer cell Ig-like receptor KIR2DL4 (CD158d) in resting NK cells J Immunol 2001 167 1877 1881 11489965
Rajagopalan S Long EO A human histocompatibility leukocyte antigen (HLA)-G-specific receptor expressed on all natural killer cells [published erratum appears in J Exp Med 2000;191:2027] J Exp Med 1999 189 1093 1100 10190900
Ponte M Cantoni C Biassoni R Tradori-Cappai A Bentivoglio G Inhibitory receptors sensing HLA-G1 molecules in pregnancy: Decidua-associated natural killer cells express LIR-1 and CD94/NKG2A and acquire p49, an HLA-G1-specific receptor Proc Natl Acad Sci U S A 1999 96 5674 5679 10318943
Vilches C Parham P KIR: Diverse, rapidly evolving receptors of innate and adaptive immunity Annu Rev Immunol 2002 20 217 251 11861603
Parham P MHC class I molecules and KIRs in human history, health and survival Nat Rev Immunol 2005 5 201 214 15719024
Stewart CA Van Bergen J Trowsdale J Different and divergent regulation of the KIR2DL4 and KIR3DL1 promoters J Immunol 2003 170 6073 6081 12794136
Witt CS Whiteway JM Warren HS Barden A Rogers M Alleles of the KIR2DL4 receptor and their lack of association with pre-eclampsia Eur J Immunol 2002 32 18 29 11754000
Goodridge JP Witt CS Christiansen FT Warren HS KIR2DL4 (CD158d) genotype influences expression and function in NK cells J Immunol 2003 171 1768 1774 12902476
Kikuchi-Maki A Catina TL Campbell KS Cutting edge: KIR2DL4 transduces signals into human NK cells through association with the Fc receptor gamma protein J Immunol 2005 174 3859 3863 15778339
Faure M Long EO KIR2DL4 (CD158d), an NK cell-activating receptor with inhibitory potential J Immunol 2002 168 6208 6214 12055234
Kikuchi-Maki A Yusa S Catina TL Campbell KS KIR2DL4 is an IL-2-regulated NK cell receptor that exhibits limited expression in humans but triggers strong IFN-gamma production J Immunol 2003 171 3415 3425 14500636
Kovats S Main EK Librach C Stubblebine M Fisher SJ A class I antigen, HLA-G, expressed in human trophoblasts Science 1990 248 220 223 2326636
LeMaoult J Le Discorde M Rouas-Freiss N Moreau P Menier C Biology and functions of human leukocyte antigen-G in health and sickness Tissue Antigens 2003 62 273 284 12974794
Trundley A Moffett A Human uterine leukocytes and pregnancy Tissue Antigens 2004 63 1 12 14651517
Ishitani A Sageshima N Lee N Dorofeeva N Hatake K Protein expression and peptide binding suggest unique and interacting functional roles for HLA-E, F, and G in maternal-placental immune recognition J Immunol 2003 171 1376 1384 12874228
Bainbridge D Ellis S Le Bouteiller P Sargent I HLA-G remains a mystery Immunol Today 2001 22 548 552
Carosella ED Dausset J Rouas-Freiss N Immunotolerant functions of HLA-G Cell Mol Life Sci 1999 55 327 333 10228553
Avril T Jarousseau AC Watier H Boucraut J Le Bouteiller P Trophoblast cell line resistance to NK lysis mainly involves an HLA class I-independent mechanism J Immunol 1999 162 5902 5909 10229826
Loke YW King A Decidual natural-killer-cell interaction with trophoblast: Cytolysis or cytokine production? Biochem Soc Trans 2000 28 196 198 10816126
Parham P NK cells and trophoblasts: Partners in pregnancy J Exp Med 2004 200 951 955 15492121
Hiby SE Walker JJ O'Shaughnessy KM Redman CWG Carrington M Combinations of maternal KIR and fetal HLA-C genes influence the risk of preeclampsia and reproductive success J Exp Med 2004 200 957 965 15477349
Croy BA Esadeg S Chantakru S Van den Heuvel M Paffaro VAJ Update on pathways regulating the activation of uterine natural killer cells, their interactions with decidual spiral arteries and homing of their precursors to the uterus J Reprod Immunol 2003 59 175 191 12896821
van der Meer A Lukassen HG van Lierop MJ Wijnands F Mosselman S Membrane-bound HLA-G activates proliferation and interferon-gamma production by uterine natural killer cells Mol Hum Reprod 2004 10 189 195 14981146
Croy BA He H Esadeg S Wei QX McCartney D Uterine natural killer cells: insights into their cellular and molecular biology from mouse modelling Reproduction 2003 126 149 160 12887272
Conner SD Schmid SL Regulated portals of entry into the cell Nature 2003 422 37 44 12621426
van der Bliek AM Redelmeier TE Damke H Tisdale EJ Meyerowitz EM Mutations in human dynamin block an intermediate stage in coated vesicle formation J Cell Biol 1993 122 553 563 8101525
Catalfamo M Karpova T McNally J Costes SV Lockett SJ Human CD8+ T cells store RANTES in a unique secretory compartment and release it rapidly after TcR stimulation Immunity 2004 20 219 230 14975243
de Renzis S Sonnichsen B Zerial M Divalent Rab effectors regulate the sub-compartmental organization and sorting of early endosomes Nat Cell Biol 2002 4 124 133 11788822
Boyson JE Erskine R Whitman MC Chiu M Lau JM Disulfide bond-mediated dimerization of HLA-G on the cell surface Proc Natl Acad Sci U S A 2002 99 16180 16185 12454284
Allan DSJ Colonna M Lanier LL Churakova TD Abrams JS Tetrameric complexes of human histocompatibility leukocyte antigen (HLA)-G bind to peripheral blood myelomonocytic cells J Exp Med 1999 189 1149 1155 10190906
Winter CC Gumperz JE Parham P Long EO Wagtmann N Direct binding and functional transfer of NK cell inhibitory receptors reveal novel patterns of HLA-C allotype recognition J Immunol 1998 161 571 577 9670929
Valés-Gómez M Erskine RA Deacon MP Strominger JL Reyburn HT The role of zinc in the binding of killer cell Ig-like receptors to class I MHC proteins Proc Natl Acad Sci U S A 2001 98 1734 1739 11172020
D'Andrea A Chang C Phillips JH Lanier LL Regulation of T cell lymphokine production by killer cell inhibitory receptor recognition of self HLA class I alleles J Exp Med 1996 184 789 794 8760835
Boyington JC Motyka SA Schuck P Brooks AG Sun PD Crystal structure of an NK cell immunoglobulin-like receptor in complex with its class I MHC ligand Nature 2000 405 537 543 10850706
Fujii T Ishitani A Geraghty DE A soluble form of the HLA-G antigen is encoded by a messenger ribonucleic acid containing intron 4 J Immunol 1994 153 5516 5524 7989753
Park GM Lee S Park B Kim E Shin J Soluble HLA-G generated by proteolytic shedding inhibits NK-mediated cell lysis Biochem Biophys Res Commun 2004 313 606 611 14697234
Van Lierop MJ Wijnands F Loke YW Emmer PM Lukassen HG Detection of HLA-G by a specific sandwich ELISA using monoclonal antibodies G233 and 56B Mol Hum Reprod 2002 8 776 784 12149411
Yang H Young DW Gusovsky F Chow JC Cellular events mediated by lipopolysaccharide-stimulated toll-like receptor 4. MD-2 is required for activation of mitogen-activated protein kinases and Elk-1 J Biol Chem 2000 275 20861 20866 10877845
Miaczynska M Pelkmans L Zerial M Not just a sink: Endosomes in control of signal transduction Curr Opin Cell Biol 2004 16 400 406 15261672
Takeshita F Gursel I Ishii KJ Suzuki K Gursel M Signal transduction pathways mediated by the interaction of CpG DNA with Toll-like receptor 9 Semin Immunol 2004 16 17 22 14751759
Vieira AV Lamaze C Schmid SL Control of EGF receptor signaling by clathrin-mediated endocytosis Science 1996 274 2086 2089 8953040
MacKinnon R Membrane protein insertion and stability Science 2005 307 1425 1426 15746418
Menier C Riteau B Dausset J Carosella ED Rouas-Freiss N HLA-G truncated isoforms can substitute for HLA-G1 in fetal survival Hum Immunol 2000 61 1118 1125 11137216
Fournel S Aguerre-Girr M Huc X Lenfant F Alam A Soluble HLA-G1 triggers CD95/CD95 ligand-mediated apoptosis in activated CD8+ cells by interacting with CD8 J Immunol 2000 164 6100 6104 10843658
Lila N Rouas-Freiss N Dausset J Carpentier A Carosella ED Soluble HLA-G protein secreted by allo-specific CD4+ T cells suppresses the allo-proliferative response: A CD4+ T cell regulatory mechanism Proc Natl Acad Sci U S A 2001 98 12150 12155 11572934
Mellor AL Munn DH Immunology at the maternal-fetal interface: Lessons for T cell tolerance and suppression Annu Rev Immunol 2000 18 367 391 10837063
Moffett A Loke YW The immunological paradox of pregnancy: A reappraisal Placenta 2004 25 1 8 15013633
Navarro F Llano M Bellon T Colonna M Geraghty DE The ILT2 (LIR1) and CD94/NKG2A NK cell receptors respectively recognize HLA-G1 and HLA-E molecules co-expressed on target cells Eur J Immunol 1999 29 277 283 9933109
Mellor AL Munn DH Tryptophan catabolism prevents maternal T cells from activating lethal anti-fetal immune responses J Reprod Immunol 2001 52 5 13 11600174
Golos TG Nonhuman primate placental MHC expression: A model for exploring mechanisms of human maternal-fetal immune tolerance Hum Immunol 2003 64 1102 1109 14602241
Jackson MR Carney EW Lye SJ Ritchie JW Localization of two angiogenic growth factors (PDECGF and VEGF) in human placentae throughout gestation Placenta 1994 15 341 353 7937592
Choi SJ Park JY Lee YK Choi HI Lee YS Effects of cytokines on VEGF expression and secretion by human first trimester trophoblast cell line Am J Reprod Immunol 2002 48 70 76 12389595
Matsumoto H Ma WG Daikoku T Zhao X Paria BC Cyclooxygenase-2 differentially directs uterine angiogenesis during implantation in mice J Biol Chem 2002 277 29260 29267 12034746
Lane BR Liu J Bock PJ Schols D Coffey MJ Interleukin-8 and growth-regulated oncogene alpha mediate angiogenesis in Kaposi's sarcoma J Virol 2002 76 11570 11583 12388718
Sparmann A Bar-Sagi D Ras-induced interleukin-8 expression plays a critical role in tumor growth and angiogenesis Cancer Cell 2004 6 447 458 15542429
Zhang JH He H Borzychowski AM Takeda K Akira S Analysis of cytokine regulators inducing interferon production by mouse uterine natural killer cells Biol Reprod 2003 69 404 411 12646495
Fuzzi B Rizzo R Criscuoli L Noci I Melchiorri L HLA-G expression in early embryos is a fundamental prerequisite for the obtainment of pregnancy Eur J Immunol 2002 32 311 315 11807769
Yie SM Balakier H Motamedi G Librach CL Secretion of human leukocyte antigen-G by human embryos is associated with a higher in vitro fertilization pregnancy rate Fertil Steril 2005 83 30 36 15652883
Yie SM Taylor RN Librach C Low plasma HLA-G protein concentrations in early gestation indicate the development of preeclampsia later in pregnancy Am J Obstet Gynecol 2005 193 204 208 16021080
Menier C Rabreau M Challier JC Le Discorde M Carosella ED Erythroblasts secrete the nonclassical HLA-G molecule from primitive to definitive hematopoeeisis Blood 2004 104 3153 3160 15284117
Lee N Malacko AR Ishitani A Chen MC Bajorath J The membrane-bound and soluble forms of HLA-G bind identical sets of endogenous peptides but differ with respect to TAP association Immunity 1995 3 591 600 7584149
Cohen GB Gandhi RT Davis DM Mandelboim O Chen BK The selective downregulation of class I major histocompatibility complex proteins by HIV-1 protects HIV-infected cells from NK cells Immunity 1999 10 661 671 10403641
Wagtmann N Rajagopalan S Winter CC Peruzzi M Long EO Killer cell inhibitory receptors specific for HLA-C and HLA-B identified by direct binding and by functional transfer Immunity 1995 3 801 809 8777725
Sonnichsen B De Renzis S Nielsen E Rietdorf J Zerial M Distinct membrane domains on endosomes in the recycling pathway visualized by multicolor imaging of Rab4, Rab5, and Rab11 J Cell Biol 2000 149 901 914 10811830
Eissmann P Beauchamp L Wooters J Tilton JC Long EO Molecular basis for positive and negative signaling by the natural killer cell receptor 2B4 (CD244) Blood 2005 105 4722 4729 15713798
Rojo S Burshtyn DN Long EO Wagtmann N Type I transmembrane receptor with inhibitory function in mouse mast cells and NK cells J Immunol 1997 158 9 12 8977169
Strong RK Holmes MA Li P Braun L Lee N HLA-E allelic variants. Correlating differential expression, peptide affinities, crystal structures, and thermal stabilities J Biol Chem 2003 278 5082 5090 12411439
Winter CC Long EO Campbell KS Colonna M Binding of soluble KIR-Fc fusion proteins to HLA class I Natural Killer Cell Protocols 2000 Totowa (New Jersey) Humana Press 239 250
|
16366734
|
PMC1318474
|
CC0
|
2021-01-05 08:21:47
|
no
|
PLoS Biol. 2006 Jan 27; 4(1):e9
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0040009
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpmedplosmedPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0040016SynopsisAnimal BehaviorDevelopmentEcologyEvolutionZoologyOtherGeriatricsGetting an Evolutionary Handle on Life after Reproduction Synopsis1 2006 27 12 2005 27 12 2005 4 1 e16Copyright: © 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 Evolution of Senescence and Post-Reproductive Lifespan in Guppies (Poecilia reticulata)
==== Body
When Richard Dawkins famously called organisms “throwaway survival machines” that exist solely to preserve the genes that made them, critics balked at the specter of genetic determinism. But Dawkins' selfish genes derive straight from classical evolutionary theory that says life exists to reproduce, and that natural selection should act on any traits that increase reproductive success. Since many animals live beyond their fertile years, biologists have searched for evolutionary clues to this extended lifespan. What role, if any, does natural selection play in the evolution of the postreproductive lifespan?
For natural selection to shape the twilight years, postreproductive females should contribute to the fitness of their offspring or relatives, a hypothesis called the “grandmother effect.” Such contributions require that organisms spawn helpless offspring or live in extended groups where postreproductive females help raise the young. Though many mammals, including lions and baboons, rear dependent young and operate within complex social groups, studies have found no evidence of a granny effect, and females mostly live just long enough to care for their last born. For nonsocial animals that spawn independent young, extended lifespan is associated with good nutrition and the absence of disease and predators. While historical records and demographic analyses offer support for an adaptive granny effect in humans—which some biologists have offered as a possible explanation for the existence of menopause—few studies have experimentally tested for signs of selection in the evolution of a postreproductive lifespan.
Analysis of the causes of variation in post-reproductive lifespan indicates that the evolution of lifespan in guppies is due to selection during their reproductive stage
In a new study, David Reznick, Michael Bryant, and Donna Holmes expand on their ongoing investigations of the life history of guppies confronting different predatory threats in Trinidad. Individuals facing different mortality threats should evolve different adaptations in their life histories, such as age at first reproduction, investment in reproduction, and patterns of senescence, including declines in reproduction. Since guppies are livebearers that provide no postnatal maternal care, Reznick et al. predicted the populations would show no differences in postreproductive lifespan—which is what they found. Though overall lifespan varied among the populations, these variations stemmed from differences in time allotted only to reproduction. Postreproductive lifespan, in contrast, showed no signs of being under selection, and appeared to be what the authors called a “random add-on at the end of the life history.” Random or not, this is the first demonstration of a postreproductive lifespan in fish.
Reznick et al. raised a second-generation laboratory brood of wild guppies taken from high- and low-predation streams at two locations in the mountains of Trinidad. (The high-predation sites harbor predators that frequently prey on guppies. Low-predation sites are found in the same streams, above waterfalls that exclude predators but not guppies.) A high- and low-predation site was sampled from each site, and feeding was manipulated to reflect food availability in the wild (fish in low-predation environments typically eat less and weigh less than fish in high-predation environments). Females were mated once a week until they produced offspring, and were mated again after each brood (when copulation is most likely).
The authors measured growth rate, body size, interbrood interval, and litters per lifetime for each population, and divided each individual's lifespan into age at first birth, reproductive phase, and postreproductive lifespan. Guppies from high-predation localities gave birth sooner than those from low-predation sites; they also reproduced over a longer period and were much older when they stopped reproducing. To estimate postreproductive lifespan, Reznick et al. determined whether the time between last birth and death significantly exceeded the time needed to spawn another litter (calculated as a threshold, since interbrood intervals varied for each individual). About 60% of individuals lived beyond the time they would have been expected to produce another brood. While the authors found no differences in the probability that any particular group would enjoy an extended postreproductive lifespan or that an individual would stop reproducing before dying, they did find that the probability of experiencing an extended postreproductive lifespan increased along with the length of reproductive lifespan. Thus, even though postreproductive lifespan has no direct effect on fitness, it is linked to a component of life history that does.
Altogether, these results provide the first experimental confirmation that evolution works selectively on those aspects of life history that directly affect fitness. These findings also refute the suggestion that fish may experience little or no reproductive senescence based on evidence that they continue to produce eggs as adults. It's an open question whether postreproductive lifespan can influence fitness enough to be under selection. But in a field dominated by investigations into the origins of human menopause and extended lifespan, the authors make a strong case for using experimental comparative analyses of other species to gain an evolutionary perspective on the human condition. —Liza Gross
|
0
|
PMC1318475
|
CC BY
|
2021-01-05 08:21:46
|
no
|
PLoS Biol. 2006 Jan 27; 4(1):e16
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0040016
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpmedplosmedPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0040030SynopsisCell BiologyImmunologyAllergy/ImmunologyObstetrics/GynecologyHomo (Human)PrimatesMammalsVertebratesEukaryotesInside a Killer: Immune Signals May Promote Vascular Growth Synopsis1 2006 27 12 2005 27 12 2005 4 1 e30Copyright: © 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.
Activation of NK Cells by an Endocytosed Receptor for Soluble HLA-G
==== Body
“Natural killer cells”—their very name chillingly evokes their primary function as the hit men of the body's immune system. But according to a new study by Sumati Rajagopalan, Eric Long, and their colleagues, these killers may also play a part in ensuring the success of a pregnancy. And the signal that provokes this unexpected action is itself unusual, acting not from the surface of the cell, but from within it.
Natural killer (NK) cells are a type of white blood cell. NK cells circulate in the blood stream and are also residents in a few tissues, including the uterus, where they are the primary type of white blood cell. On their surfaces, they carry a variety of receptors, which sense the exterior environment and trigger a host of internal responses. When activated, NK cells are cytotoxic—they kill target cells, such as virus-infected cells and tumor cells—and secrete soluble signaling molecules that help mount immune responses to infection by parasites, bacteria, and other invaders.
NK cell receptors bind to proteins on the surfaces of other cells. One type of NK cell receptor, called killer cell immunoglobulin-like receptor (KIR)2DL4, binds to the protein human leukocyte antigen (HLA)-G, whose precise role is unknown and which occurs in both membrane-bound and secreted (soluble) forms. Interestingly, HLA-G is primarily found on trophoblast cells, a type of embryonic tissue that invades the uterine lining to support the developing fetus.
A natural killer cell receptor called KIR2DL4 stimulates an immune signaling pathway from inside an intracellular vesicular structure (shown above)
The response of NK cells activated by KIR2DL4 is unusual: they release cytokines, immune chemicals that trigger actions in other cells. But this cytokine release, unlike that associated with other receptors, is not accompanied by cytotoxicity.
To understand the role of KIR2DL4-bearing NK cells in the uterus, and how they might respond to HLA-G signals, the authors examined the KIR2DL4/HLA-G interaction in detail. By activating the receptor with an antibody (to ensure that only KIR2DL4 would be triggered), they first showed that cytokine secretion could be triggered only when the antibody was soluble, not when bound to a solid surface. This behavior immediately suggested that the interaction between the two doesn't end at the membrane, since if it did, the immobilized antibodies would have sufficed to activate the cells.
They next showed that KIR2DL4 is brought into the cell by endocytosis, an energy-requiring process that cells use to carry membrane-bound molecules to the interior. They found that KIR2DL4 was present in endosomes, the vesicles formed by endocytosis. KIR2DL4 was found primarily in so-called early endosomes, but much less so in later stages, when the vesicle is being prepared for merging with a lysosome, the cell's disposal system.
What is KIR2DL4 doing in the early endosome? Rajagopalan et al. found that the receptor binds to the soluble, secreted form of HLA-G, and that HLA-G is carried with it into the endosome. Here, the complex triggered cytokine production and NK cell signaling. KIR2DL4 that could not reach the endosome could not trigger cytokine production. Therefore, the authors suggest that KIR2DL4 signaling happens from within the endosome, not at the cell surface. The cytokines produced are known to be active not only in inflammatory reactions but also in angiogenesis, or formation of new blood vessels.
Endosomes have emerged recently as important signaling compartments, and these results support and extend that understanding. On the cell surface, the receptor senses the environment for soluble ligand; once internalized into endosomes, the membrane-bound receptor interacts with signaling molecules within endosomes to set off a cascade of reactions, ultimately leading to protein production. The authors suggest that internal signaling may increase the fidelity of the desired response by avoiding the many possibly conflicting signals arising from other receptors at the surface during cell–cell interactions.
The results presented here also suggest some intriguing ideas regarding how the immune system responds to the developing embryo. Within the uterus, the embryo represents a special challenge, since it is composed of cells bearing partly foreign (that is, paternal) genetic material that must not only be tolerated but allowed to intimately intertwine with the mother's tissue to develop a new blood supply. This remodeling of the vascular system is partly under immune system control. The angiogenic cytokines triggered by KIR2DL4 activation may support this process.
Further support for this idea comes from recent observations that a higher level of soluble HLA-G promotes higher rates of successful pregnancy, while reduced HLA-G correlates with higher rates of preeclampsia, a condition caused by insufficient vascular remodeling during pregnancy. Additional study of the KIR2DL4/ HLA-G signaling pathway may lead to a better understanding of this potentially fatal complication of pregnancy, and better ways to prevent it. As the production of soluble HLA-G can also be induced in certain cell types, including tumor cells, the KIR2DL4/HLA-G signaling pathway may also serve additional functions, unrelated to pregnancy. —Richard Robinson
|
0
|
PMC1318476
|
CC BY
|
2021-01-05 08:21:47
|
no
|
PLoS Biol. 2006 Jan 27; 4(1):e30
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0040030
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpmedplosmedPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1636673510.1371/journal.pbio.0040029Research ArticleNeuroscienceMus (Mouse)Dynamic Remodeling of Dendritic Arbors in GABAergic Interneurons of Adult Visual Cortex Dendritic Remodeling in Adult CortexLee Wei-Chung Allen
1
2
Huang Hayden
3
Feng Guoping
4
Sanes Joshua R
5
Brown Emery N
2
6
7
So Peter T
3
Nedivi Elly
1
2
8
*1The Picower Institute for Learning and Memory, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of America2Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of America3Department of Mechanical Engineering and Division of Biological Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of America4Department of Neurobiology, Duke University Medical Center, Durham, North Carolina, United States of America5Department of Molecular and Cellular Biology, Harvard University, Cambridge, Massachusetts, United States of America6MIT-Harvard Division of Health Science and Technology, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of America7Neuroscience Statistics Research Laboratory, Department of Anaesthesia and Critical Care, Massachusetts General Hospital, Boston, United States of America8Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of AmericaStevens Charles F. Academic EditorHoward Hughes Medical InstituteUnited States of America* To whom correspondence should be addressed. E-mail: [email protected] 2006 27 12 2005 27 12 2005 4 2 e2927 9 2005 22 11 2005 Copyright: © 2006 Lee et al.2006This 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 New Window into Structural Plasticity in the Adult Visual Cortex
Despite decades of evidence for functional plasticity in the adult brain, the role of structural plasticity in its manifestation remains unclear. To examine the extent of neuronal remodeling that occurs in the brain on a day-to-day basis, we used a multiphoton-based microscopy system for chronic in vivo imaging and reconstruction of entire neurons in the superficial layers of the rodent cerebral cortex. Here we show the first unambiguous evidence (to our knowledge) of dendrite growth and remodeling in adult neurons. Over a period of months, neurons could be seen extending and retracting existing branches, and in rare cases adding new branch tips. Neurons exhibiting dynamic arbor rearrangements were GABA-positive non-pyramidal interneurons, while pyramidal cells remained stable. These results are consistent with the idea that dendritic structural remodeling is a substrate for adult plasticity and they suggest that circuit rearrangement in the adult cortex is restricted by cell type–specific rules.
Chronic in vivo imaging of fluorescent-labeled neurons in adult mice reveals extension and retraction of dendrites in GABAergic non-pyramidal interneurons of the cerebral cortex.
==== Body
Introduction
Hubel and Wiesel's groundbreaking work in the 1960s and 1970s defined a critical period in brain development when manipulating visual inputs causes dramatic functional and structural changes in layer 4 of the primary visual cortex [1–3]. Since their finding that large-scale rearrangement of thalamic afferents in the visual cortex is restricted to a developmental critical period, the adult brain has been considered relatively hard-wired and limited in its capacity for structural change. Functional reorganization of primary sensory maps in the adult brain (reviewed in [4]), even across long distances [5], was explained as unmasking of existing connections and was not considered to require outright growth [6]. Although there are indications that the adult cortex is capable of anatomical change in response to peripheral manipulation, particularly in the superficial layers [7,8], the scale of change is small compared with the critical period and is difficult to detect against the general variance in the size and shape of cortical neurons. Moreover, such changes are seen only in response to external perturbation, leaving it unclear whether arbor remodeling normally occurs in the adult cortex on a day-to-day basis, and to what extent.
With the advent of new technologies to time-lapse image neuronal morphology in vivo [9], the issue is now being revisited. Repeated in vivo imaging of apical dendrites extending from layer 5 pyramidal neurons into the superficial layers has been used to investigate dendritic spine dynamics in both the somatosensory and the visual cortex [10–13]. Less attention has been paid to potential changes in the overall structure of dendritic arbors. In fact, it has been suggested that little if any structural plasticity occurs in the apical dendrites of layer 5 pyramidal neurons in the adult somatosensory cortex [10], or in the apical dendrites of mitral and tufted cells in the adult olfactory bulb [14]. No study, to our knowledge, has directly addressed the potential for structural dynamics in a cross-section of neurons that reflects the diversity of neocortical cell types. More specifically, the non-pyramidal neurons of the neocortex have yet to be the focus of investigation by in vivo imaging studies, despite their important role in adult cortical plasticity and in reorganization of cortical maps [15,16].
Here we investigate the dendritic arbor dynamics of pyramidal and non-pyramidal neurons in the superficial layers of the adult visual cortex in vivo. Our results show that while dendritic branches of pyramidal cells remain stable, non-pyramidal interneurons in these layers are dynamic, exhibiting a range of structural changes on a week-to-week basis.
Results
Repeated In Vivo Imaging of Cortical Neurons
Studies suggest that intracortical connections in the supragranular layers of the neocortex, and in particular axonal sprouting, may be a locus of adult structural plasticity [7]. To directly test the prediction that neurons in layer 2/3 are capable of structural change and to assess the extent of dendritic structural dynamics in the adult cortex, we chronically imaged neuronal morphology in the intact rodent brain. To allow long-term visualization of neuronal structure in vivo, cranial windows were bilaterally implanted over the visual cortices of thy1-GFP-S mice [17] at 4–6 wk of age. These mice express green fluorescent protein (GFP) in a sparse pseudo-random subset of neocortical neurons. Imaging began at least 2 wk after surgery to allow for recovery and optical clarification of the cranial windows. Brains were screened for optically accessible GFP-positive neurons using wide-field fluorescence, and neuronal location was noted using local landmarks in the brain's surface vasculature. Individual GFP-labeled neurons in layer 2/3 of the visual cortex of anesthetized adult mice were then time-lapse imaged using a custom-made two-photon microscope [18]. To include as many neuronal branch tips as possible within the imaging volume, nine slightly overlapping volumes were imaged in a 3 × 3 array through z–x–y translation of an automated motorized stage. Individual image planes were stitched together to create a montage of adjoining x–y sections for a given depth from the pial surface. In an attempt to provide a comprehensive view of adult structural plasticity, data collection was initially not restricted to any particular cell type. Six pyramidal cells and eight non-pyramidal cells from 13 animals were time-lapse imaged for 3–10 wk (Figure 1), and four-dimensional morphometric analysis was carried out by quantitative comparison of dendritic branch tip length (BTL) as a function of time. Branch tips that were not imaged clearly for multiple imaging sessions or whose termination was unclear were excluded from analysis.
Figure 1 Summary of Imaging Sessions Displayed by Age
Each neuron was named by an arbitrary three-letter code. Empty circles represent pyramidal neuron-imaging sessions, and filled circles represent non-pyramidal neuron-imaging sessions.
Pyramidal Cells Are Stable Over Time
Imaged pyramidal cells exhibited typical small pyramidal morphologies with a spiny apical dendrite and a skirt of spiny basal dendrites emanating from the lower half of a pyramid-shaped cell body. An example is shown in Figure 2 and in Video S1. The cell body of this pyramidal neuron, “dow,” was located 180 μm below the pial surface. It had a total of 57 dendritic branch tips on its apical dendrite and five primary basal dendrites. We monitored 28 of the 57 branch tips over 9 wk. Examination of individual branch tips revealed no overt sign of structural change (Figure 2A–2C). Similar examination of all cells in the pyramidal population did not identify any change in apical and basal dendritic branches. These data suggest that under normal conditions the dendritic branches of layer 2/3 pyramidal neurons in the visual cortex are relatively stable in the adult, and are consistent with previous studies reporting dendritic branch stability in other cortical areas [10,14].
Figure 2 Dendritic Arbors of Pyramidal Neurons Are Stable
(A) MZPs near the cell body of the pyramidal cell “dow” acquired over 9 wk.
(B) Two-dimensional projections of three-dimensional skeletal reconstructions of “dow.”
(C) High-magnification view of branch tip (green arrow) in region outlined by green box in (B).
Scale bars: (A and B), 50 μm; (C), 10 μm.
Dynamic Remodeling of Non-Pyramidal Neurons
We next examined layer 2/3 non-pyramidal neurons. Figure 3 shows maximum-intensity z-projections (MZPs) of image planes close (± 15 μm) to the cell body of a non-pyramidal neuron “nmr” with a bitufted dendritic morphology revealing its highly complex local arborization (Figure 3A and Video S2). The cell body's center of mass was 118 μm below the pial surface. This neuron had four primary dendrites with a total of 49 branch tips. Twenty-eight of the 49 branch tips were monitored for 4 wk. Four of the 28 branch tips exhibited variations in length. Two examples are shown where branch tips visibly elongated in the x–y plane (Figure 3A–F). Branch tip #20 elongated by approximately 16 μm over 4 wk (Figure 3C, 3D, and 3G). Concurrently, branch tip #15 increased in length by approximately 10 μm (Figure 3E–3G). Both branch tips emanate from the same primary dendrite whose dendritic branch length accounts for 62% of the total monitored dendritic length of the neuron. These results demonstrate that dendritic arbors of neurons within the adult neocortex are capable of growth.
Figure 3 Dendritic Growth in Multiple Branches of a Non-Pyramidal Neuron
(A) MZPs near the cell body of the (∼118-μm deep) non-pyramidal cell “nmr” acquired over 4 wk. Two examples of dendritic branch growth are indicated by red arrowheads.
(B) Two-dimensional projections of three-dimensional skeletal reconstructions of the non-pyramidal neuron “nmr.”
(C) High-magnification view of one growing branch tip (#20) (red box in [A and B]). Red arrowhead marks the approximate distal end of the branch tip at 11 wk.
(D) Three-dimensional isosurface reconstructions of branch tips in (C).
(E) High-magnification view of branch tip #15 (green box in [A and B]).
(F) Three-dimensional isosurface reconstructions of branch tips in (E).
(G) Plot of change in BTL of dynamic branch tips as a function of age. Number to the right denotes branch tip number.
Scale bars: (A and B), 25 μm; (B–F), 5 μm.
A different non-pyramidal neuron, “paz,” residing 78 μm below the pial surface, is shown in Figure 4. Two-dimensional projections of the three-dimensional traces show a moderately branched interneuron with a bitufted dendritic morphology (Figure 4A). This cell had seven primary dendrites with 47 branch tips. Twenty-nine of the 47 dendritic branch tips were monitored for 7 wk. Virtually all the branches were stable. However, two branches exhibited remodeling, one of which was so large in scale that it exceeded the imaging volume. Time-lapse images revealed that within as little as 2 wk, this branch tip more than doubled its length and exited the imaging volume (Figure 4). Although at 13 wk postnatal we were unable to follow the process to its termination, we measured a net extension of >92 μm from the branch tip at 11 wk to its location at the edge of the imaging volume 2 wk later. The axon of this neuron was found to project from the cell body in the opposite direction of this changing dendritic branch tip. This dramatic increase in BTL indicates that neurons in adult visual cortex have the capacity for large-scale remodeling.
Figure 4 Large-Scale Dendritic Branch Growth in a Non-Pyramidal Neuron
(A) Three-dimensional skeletal reconstructions of the non-pyramidal neuron “paz” from images acquired over 7 wk. Note that a growing branch tip exceeded the imaging volume after 11 wk, and black arrowheads denote its postulated continuation.
(B) MZPs of region outlined by red box in (A). The branch tip (#15) elongates radially in the x–y-axis and away from the pial surface in the z-axis. Non-specific labeling in the MZP is exacerbated by the summation of additional z stacks due to elongation on the z-axis. The traced branch of interest is highlighted by a green overlay. Red arrowheads mark the approximate distal end of the branch tip at 11 wk.
(C) Three-dimensional isosurface reconstructions of the region of the branch tip outlined by a green box in (B).
(D) Plot of change in BTL of dynamic branch tips as a function of age. Triangles denote the minimum length of the branch tip as it exceeds the border of the imaging volume. Number to the right denotes branch tip number.
Scale bars: (A), 25 μm; (B and C), 10 μm.
Although many non-pyramidal cells are spine-free, sparsely spinous non-pyramidal neurons were also observed and imaged. These cells typically exhibited multipolar dendritic morphologies (Figure 5A). The interneuron “zen” shown in Figure 5 had seven primary dendrites with 61 dendritic branch tips and seven spines on 2,814 μm (averaged over 5 wk) of monitored dendrite. The cell body's center of mass was 100 μm below the pial surface. The few spines on this neuron exhibited motility (Figure 5B and 5C) as previously described for spines on pyramidal neurons [10–13]. Six branch tips showed changes in length (Figure 5I). For example, branch tip #50 shown in Figure 5B and 5D elongated by approximately 7 μm (Figure 5I). In this cell we also observed a couple of rare de novo branch-tip additions. The initial addition of one new branch was first seen at 13 wk postnatal and could be observed elongating to approximately 22 μm by 14 wk postnatal (Figure 5E, 5F, and 5I). Also in the proximity of the branch addition, we observed the retraction of a putative axon of unknown origin (Figure 5E and 5F). A second branch-tip addition was concurrently seen on an independent dendrite (Figure 5G and 5H), extending approximately 9 μm from 11 wk to 15 wk postnatal (Figure 5I).
Figure 5 Branch Extensions, Retractions, and De Novo Branch Tip Addition in a Non-Pyramidal Neuron
(A) Three-dimensional skeletal reconstructions of the sparsely spinous non-pyramidal neuron “zen” from images acquired over 5 wk.
(B) High-magnification MZP view of region outlined by purple box in (A). Red arrowheads indicate examples of structural remodeling. Three-dimensional isosurface reconstructions show
(C), the elongation and retraction of a spine toward an axon (yellow overlay).
(D) Structural change in a cluster of branch tips (#50, far right) (red box in [B]).
(E) Higher-magnification MZP view of region outlined by green box in (A). Examples of process retraction and branch-tip (#3) addition are labeled with yellow and green arrowheads, respectively.
(F) Three-dimensional isosurface reconstructions of (E) with axon in yellow overlay.
(G) MZP view of region outlined by cyan box in (A) shows branch-tip (#16) addition on a different dendrite.
(H) Three-dimensional isosurface reconstructions of (G).
(I) Plot of change in BTL of dynamic branch tips as a function of age. Number to the right denotes branch tip number.
Scale bars: (A), 50 μm; (B–H), 5 μm.
Every one of the non-pyramidal cells imaged from layer 2/3 showed at least one and as many as seven changing dendritic branch tips. On average, approximately 14% of the monitored branch tips on non-pyramidal interneurons showed structural rearrangement. Of these, 3% were new branch-tip additions, 2% were loss of existing branch tips, and the rest were approximately half elongations and half retractions. Remodeling in some cases was incremental but could also occur in short temporal bursts. Changes were never observed in primary, first-order, dendritic branches, but they were otherwise not restricted by branch order. These data demonstrate that even without peripheral perturbations, branch tips of non-pyramidal cells in the superficial layers of the adult neocortex exhibit elongation, retraction, and branch-tip addition.
Dynamically Remodeling Neurons Express GABA
Neocortical interneurons are a diverse population with distinct morphological, physiological, and molecular subtypes [19–23]. Three of the eight imaged non-pyramidal neurons were unequivocally identified in coronal sections after post-imaging immunohistochemistry based on morphology and location. All of these cells were immunopositive for GABA (gamma-aminobutyric acid) (Figure 6A–6C), while pyramidal cells were GABA-negative (Figure 6D and 6F).
Figure 6 Remodeling Adult Non-Pyramidal Neurons in the Superficial Layers of the Visual Cortex Are Inhibitory GABAergic Interneurons
(A) Confocal image stack of a coronal section containing the chronically imaged non-pyramidal neuron “ttr” is visualized by GFP staining (green, filled arrows) and is immunopositive for GABA (visualized in red) (B), overlay of GFP and GABA shown in (C). (D) Confocal image stack of a coronal section containing the chronically imaged pyramidal neuron “dow” is immunonegative for GABA (visualized in red) (E), overlay of GFP and GABA shown in (F). Scale bar: (A–F), 100 μm.
DAPI staining located the border between layer 1 and layer 2/3 at ∼80 μm below the pial surface (Figure S1A–S1D), consistent with previous findings [24], thus placing the imaged non-pyramidal cell bodies within layer 2/3 or at the layer 1–2/3 border (Figure S1). In an attempt to further classify their subtype, we also probed the sections for parvalbumin, somatostatin, and cholecystokinin, but found the imaged neurons negative for all three (WCAL and EN, unpublished data). This was not surprising given the low representation of these subtypes in layers 1 and 2/3 of the visual cortex in both GFP- and non-GFP–labeled GABA-positive interneurons (Figure S1E). From 158 GFP-positive non-pyramidal cells in seven animals, 92% (SEM = 2.1%) were GABA-positive. Since virtually all the non-pyramidal interneurons in the superficial layers of the visual cortex are GABA-positive, it is likely that the five imaged neurons that were not successfully identified in the sectioned brains were also GABAergic, strongly suggesting that dendritic arbor remodeling in adult neurons occurs predominantly in inhibitory GABAergic interneurons.
Discussion
Although the capacity for change in the adult brain is limited compared with development, the adult cerebral cortex does maintain a degree of plasticity (reviewed in [4]). Electrophysiological recordings and anatomical analysis both suggest that potential sites for this plasticity are the horizontal connections within the superficial cortical layers [7,25–27]. Additional evidence comes from molecular studies demonstrating that in the superficial layers of adult striate cortex, visual input transcriptionally regulates genes involved in process outgrowth [28–31]. Together these data led us to hypothesize that neurons in the superficial layers of the adult cortex can undergo arbor remodeling without extreme peripheral perturbation. To test this hypothesis, we imaged the dendritic arbors of neurons in the visual cortex of adult mice over several months. Our findings show that in layer 2/3 the dendritic structure of pyramidal neurons is stable, while inhibitory interneurons undergo dendritic arbor remodeling.
Included in the analysis were 62% of the non-pyramidal dendrites, while 42% of pyramidal dendrites were successfully monitored. Given that 35 of the 259 monitored non-pyramidal branch tips changed, if the two cell types were equivalently dynamic, then the probability of a branch tip change event can be estimated as 35/259 = 0.135. In 124 monitored pyramidal branch tips, we would expect to have observed at least one branch tip change of this type with probability 1–10−8 (Protocol S1). Therefore, the probability that we missed an event due to sampling issues is 10−8. The fact that we did not observe any changes in the pyramidal branch-tip group allows us to reject the null hypothesis that the change probabilities for the two groups are the same, in favor of the more plausible alternative that the two groups have significantly different dynamic properties. Arguing against the possibility that missed events could be accounted for by a sampling bias is the fact that when comparing the size distributions of sampled branch tips for the pyramidal and non-pyramidal population, there is a similar sampling of processes in the 40- to 120-μm range (Figure S2), the size range where most changes occur within the non-pyramidal population (Figure S3).
Although there are excellent studies in the field that promote the view that adult neocortical structure is stable, to date all of them focus on pyramidal cell morphology, and most studies focus on spine dynamics [11–13]. Our data do not contradict but rather complement these studies. While our results are consistent with previous reports on stability of the apical dendrites on layer 5 pyramidal cells [10], we do not exclude the possibility that pyramidal neurons undergo any arbor remodeling, especially in response to perturbations of the sensory periphery [32]. Rather, we suggest that under normal conditions their structural rearrangements are less pronounced than those seen in layer 2/3 interneurons.
In felines and primates, the physiological plasticity manifested by cortical cells during the critical period for development of eye-specific preference is accompanied by clear activity-driven segregation of geniculocortical afferents into ocular dominance columns. Yet, there are many instances where physiological plasticity can be seen in the superficial cortical layers of adult animals without such a clear anatomical “readout” [33–35]. The scale of structural change during circuit refinement in the adult may be below the threshold for visualization with the anatomical methods used to monitor layer 4 afferent segregation. It is understandable that the structural remodeling described here was previously undetected by classical anatomical methods relying on sample statistics, as the changes are on the order of tens of microns (small relative to the entire dendritic arbor); occur on a subset of dendritic branches in a subset of cell types; and can occur sporadically in bursts of remodeling. In addition, when dendritic branches of the same neuron both grow and retract, the total net change may be negligible (∼1%–5% of the total monitored dendritic branch length). A technology where the same neurons can be vitally imaged over days and weeks in the intact brain allows for discrimination of small-scale structural dynamics, opening to reinterpretation the apparent disconnect between functional plasticity measured electrophysiologically in the extragranular layers and the absence of measurable anatomical change. Our results using chronic in vivo imaging demonstrate that there is an intrinsic capacity for structural remodeling in the superficial layers of the adult neocortex. Unlike during development when changes in dendritic arbors are widespread, changes in the adult are localized to a small subset of processes. The change in these individual processes on the scale of 20–90 μm, however, may be large enough to influence receptive field properties. The changing tips are specific to a certain neuronal subtype, suggesting cell type–specific rules to this remodeling. It has yet to be shown that this minority of “plastic” dendritic processes are ones that underlie the functional reorganization of adult cortical maps measured electrophysiologically. As there is previous evidence for horizontal axonal sprouting in the adult striate cortex following retinal lesions [7], and in the somatosensory cortex following whisker trimming [36], another area for further investigation would be the structural plasticity of axonal arbors in the long-range pyramidal cells of layer 2/3, and how they relate to dendritic changes.
Unresolved questions in the field are whether and what kind of structural changes are expected in adult plasticity and how best to detect and analyze them. If averaged across the entire arbor, the changes in BTL of non-pyramidal neurons correspond to approximately 5% of the monitored dendritic branch length. However, individual branch tips changed by 16% to 456% (not including new branch-tip additions), with average elongations and retractions of 16 μm. Net changes ranged from −5% to +8% of the average monitored branch length. Since changes are localized, it may be preferable for both detection and analysis methods to be implemented on a process-by-process basis rather than by conventional approaches that rely on statistical averaging across all the processes of a cell or an entire cell population. In conventional analyses, averaging across the entire arbor would bias against events with potential functional importance if they occurred in a minority of dendrites. Even when viewed on a process-by-process basis, clearly changing branch tips can be difficult to identify when looking at the absolute change in BTL (Figure 7). For example, branches on cells such as “nmr” and “zen” (shown in Figures 3 and 5, respectively) that are unambiguously changing may be scored as false negatives. Thus, an alternative analysis that better represents the data scored by direct observation is clearly needed. We found that scaling the changes in length by the average length of the dendritic branch tip enhanced our ability to detect changes and represented salient changes in structure better than a non-scaled analysis. Advances in such quantitative analysis of morphometric measures over time will determine how to best represent structural change of neurons in the adult cortex.
Figure 7 The Change in BTL is Plotted for Each Individual Monitored Branch Tip of Every Imaged Cell
Three-letter code, top right.
(A–F) pyramidal cells; (G–N) non-pyramidal cells. Triangles and dashed lines denote the minimum length of the branch tip as it exceeds the border of the imaging volume.
Twenty percent to 30% of the neurons in the neocortex are non-pyramidal interneurons, and most adult neocortical interneurons are considered inhibitory, using GABA as a neurotransmitter [19–23]. A diverse population of non-pyramidal interneurons preferentially populate the superficial layers of the neocortex and are characterized by their morphological, electrophysiological, molecular, and targeting properties [21,23]. Most mature non-pyramidal interneurons lack dendritic spines, and most project locally, usually arborizing within a cortical column or projecting horizontally across columns, but rarely projecting to distant brain regions [37]. Inhibitory interneurons are thought to have an important role in modulating excitatory circuitry by depressing, blocking, or sculpting the temporal response properties of excitatory neurons [38,39]. During development, inhibitory circuitry is crucial for the onset of critical-period ocular-dominance plasticity [40–43] (reviewed in [44]), plasticity of the somatosensory cortex [45], and refinement of visual receptive fields [46,47]. By shortening stimulus-evoked spike trains in immature neurons, GABAergic activity can decrease the temporal asynchrony of uncorrelated inputs [48]. In addition, interneurons can coordinately synapse onto nearby excitatory pyramidal cells in a developing network, locally synchronizing their spike timing. Both shortening prolonged discharge and orchestrating spike timing could enhance the ability of target neurons to participate in spike timing–dependent plasticity [45,46]. Interestingly, in adult monkeys monocular deprivation modifies the expression of GABA and glutamic acid decarboxylase in the primary visual cortex in an eye-specific manner [49], suggesting that GABAergic transmission is sensitive to activity-dependent plasticity in the adult. Our data indicating that the structural plasticity of interneurons is continuous through adulthood raises the intriguing possibility that local remodeling of inhibitory connections may underlie adult cortical plasticity. This finding would have important implications for models of cortical functional circuitry and its activity-dependent modulation.
Materials and Methods
Animal surgery
thy1-GFP-S mice [17] were anesthetized with 2.5% Avertin (0.015 ml/g IP), and anaesthesia was monitored by breathing rate and foot-pinch reflex. The skull overlying both visual cortices [24] was carefully removed, leaving behind the dura, and 5-mm-diameter circular glass cover slips (No. 1) were positioned over the openings and sealed in place with Palacos R bone cement. Following surgery, mice were given lactated Ringers solution (0.015 ml/g SC) and Buprenix (0.3 mg/ml SC 2× daily for 5 d) as an analgesic and returned to individual cages for recovery under observation. Surgeries were performed at 4–6 wk postnatal to allow at least a 2- wk recovery before imaging.
In vivo two-photon imaging
In vivo two-photon imaging was achieved using a custom-built microscope and acquisition software [18] modified for in vivo imaging by including a custom-made stereotaxic restraint affixed to a stage insert for the motorized stage (Prior Scientific, Cambridge, United Kingdom). While designed to run at high acquisition rates, for these experiments a conventional scanning rate was used to increase signal intensity by locking the polygonal mirror and using both raster-scanning mirrors. The light source for two-photon excitation was a commercial Ti:Sapphire laser, Mira (Coherent, Santa Clara, California, United States), pumped by a 10-W solid-state laser delivering 150 fs pulses at a rate of 80 MHz with the power delivered to the objective (with a transmittance of 20% to 30%) ranging from ∼100–250 mW depending on imaging depth. The excitation wavelength was set to ∼890 nm, with the excitation signal passing through an Achroplan 40×/0.8 NA water-immersion objective (Zeiss, Oberkochen, Germany) and collected after a barrier filter by a photomultiplier tube. Due in part to the sparse labeling of cells in the superficial layers of the thy1-GFP-S neocortex, the same cells could be identified and re-imaged for up to 3 mo using local fiduciary landmarks of the brain's surface vasculature.
Image acquisition and analysis
Adult mice (8–19 wk postnatal) previously implanted with cranial windows were anesthetized with 2.5% Avertin (0.015 ml/g IP). Anaesthesia was monitored by breathing rate and foot-pinch reflex, and additional doses of anaesthetic were administered during the imaging session as needed. The head was positioned in a custom-made stereotaxic restraint affixed to a stage insert for a motorized stage (Prior Scientific). Nine slightly overlapping volumes in a 3 × 3 array were imaged through z–x–y translation of a motorized stage (z spacing ∼1.5 μm). Due to variations in head position across imaging sessions, cells of interest were not always centered in the imaging volume. These shifts in registration slightly affected image borders, so that a fraction of dendritic branch tips were excluded at any given imaging session. Since the direction of shift is random, there was no intentional bias in exclusion of tips for a particular neuron or for a particular imaging session and the same exclusion rules applied to all neurons. However, the longer process radius of the pyramidal cell dendrites potentially biased against their sampling (see the Discussion section regarding this point). Raw scanner data were processed in Matlab (Mathworks, Natick, Massachusetts, United States) and ImageJ (National Institutes of Health, Bethesda, Maryland, United States). Individual image planes were stitched together (VIAS version 2.1, http://www.mssm.edu/cnic/tools.vias.html) such that each is a 3 × 3 montage of adjoining x–y sections at a given depth from the pial surface. Four-dimensional (x, y, z, t) stacks were traced and analyzed blind to age using Object-Image (http://simon.bio.uva.nl/object-image.html) [50] and Neurolucida (MicroBrightField, Williston, Vermont, United States). Three-dimensional surface reconstructions were generated using Imaris (Bitplane AG, Zurich, Switzerland).
The analysis included 124 branch tips from six pyramidal cells in six animals and 259 branch tips from eight non-pyramidal cells in seven animals at least 65 μm from the pial surface (two non-pyramidal cells, “ttr” and “ttc,” were from the same animal) ranging in age (at the time of imaging) 8–19 wk postnatal (Figure 1). Cells were arbitrarily named with a three-letter code. BTL was measured as the linear arc length from well-defined distal ends to the first encountered branch point. Axons were not included in the skeletal tracings. Dendritic and axonal branches were distinguished by morphology. Axons were typified as tubular, thin (sometimes less than the point spread function of the microscope) processes, often studded with varicosities every few microns. Dendrites were distinguished by thicker diameters (generally >2 μm), smooth, gradually tapering processes, and characteristic branching patterns. Since we could not measure the same branch tip multiple times for a given time point, we used the time-lapse measures of the pyramidal and non-pyramidal branch tips that did not show change as an upper bound of the measurement error we would have observed if we had made multiple measurements at the same time. The average SEM was 1.9 (standard deviation = 1.4) μm for the monitored pyramidal cell branch tips and 1.1 (standard deviation = 0.8) μm for non-changing non-pyramidal branch tips. The larger measurement error in pyramidal branch tips can be attributed to the extension of long branch tips over multiple stitching boundaries.
Immunohistochemistry
Previously imaged mice were heavily anesthetized with 2.5% Avertin (0.030 ml/g IP) and their brains processed for immunohistochemistry essentially as described [51]. Sections were first incubated with GABA (rabbit polyclonal antibody; 1:5000; Sigma, St. Louis, Missouri, United States), followed with Alexa555 conjugated goat IgG secondary antibodies (1:400; Molecular Probes, Eugene, Oregon, United States). Alternatively, sections were first incubated with parvalbumin (monoclonal antibody; 1:1000; Sigma) and somatostatin (rabbit polyclonal antibody; 1:1000; Chemicon, Temecula, California, United States) followed with appropriate Alexa555 and 647-conjugated goat IgG secondary antibodies (1:400; Molecular Probes). After visualization, sections were unmounted in PBS and reprocessed using cholecystokinin (monoclonal antibody #9303; 1:1000; CURE/Digestive Disease Research Center, VAGLAHS, Los Angeles, California, United States) and GABA (rabbit polyclonal antibody; 1:500; Chemicon), followed with appropriate secondary antibodies. Imaged cells were identified by location, morphology, and local landmarks. Images were acquired with a Fluoview confocal (Olympus, Tokyo, Japan) or an upright epi-fluorescence scope (Nikon, Tokyo, Japan) using a 20×/N.A. 0.5 (Olympus), 20×/N.A. 0.75 (Nikon), or 40×/N.A.1.30 (Nikon) objective.
Supporting Information
Figure S1 Layer Localization of Imaged Neurons and Distribution of GABAergic Subtypes
(A and C) Neurons shown in Figure 6 were in layer 2/3 as shown by DAPI staining (B and D) (border of layer 1 and layer 2/3 delineated by empty arrowheads). (E) GFP labeling of non-pyramidal cells in layers 1 and 2/3 shows a representative distribution of GABAergic, parvalbumin, somatostatin, and cholecystokinin immunopositive cells. Solid bars represent percent of layers 1/2/3 GFP-labeled non-pyramidal cells immunopositive for a given marker, and empty bars represent percent of all GABA-positive cells immunopositive for the same markers. Error bars represent SEM. Scale bar (A–D): 50 μm.
(5.8 MB TIF).
Click here for additional data file.
Figure S2 Histogram of the Number of Monitored Branch Tips for Different Branch Lengths of Pyramidal and Non-Pyramidal Cells
(837 KB TIF).
Click here for additional data file.
Figure S3 Histogram of the Number of Remodeling Events in Non-Pyramidal Neurons as a Function of Distance from the Cell Soma
(899 KB TIF).
Click here for additional data file.
Protocol S1 Analysis of the Dendritic Change Propensities for Non-pyramidal and Pyramidal Neurons
(47 KB DOC).
Click here for additional data file.
Video S1
Z-Stack of Pyramidal Cell “dow” Descending at 1.5-μm Steps
(1.4 MB WMV).
Click here for additional data file.
Video S2
Z-Stack of Non-Pyramidal Cell “nmr” Descending at 1.5-μm Steps
(862 KB WMV).
Click here for additional data file.
We thank Y. Amitai, T. Fujino, J. Fortin, and J. Hoch for helpful comments on the manuscript, R. Marini for help developing the cranial window preparation, J. Evans and K. Berggren for help with reconstruction software, and G. Di Cristo for suggestions regarding immunohistochemistry. This work was sponsored by grants from the National Eye Institute of the National Institutes of Health to EN and by a Poitras Fellowship to WCAL.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. WCAL, HH, PTS, and EN conceived and designed the experiments. WCAL and HH performed the experiments. WCAL, ENB, and EN analyzed the data. GF, JRS, and PTS contributed reagents/materials/analysis tools. WCAL and EN wrote the paper.
Citation: Lee WCA, Huang H, Feng G, Sanes JR, Brown EN, et al. (2006) Dynamic remodeling of dendritic arbors in GABAergic interneurons of adult visual cortex. PLoS Biol 4(2): e29.
Abbreviations
BTLbranch tip length
GABAgamma-aminobutyric acid
GFPgreen fluorescent protein
MZPmaximum-intensity z-projection
==== Refs
References
Hubel DH Wiesel TN The period of susceptibility to the physiological effects of unilateral eye closure in kittens J Physiol 1970 206 419 436 5498493
Hubel DH Wiesel TN LeVay S Plasticity of ocular dominance columns in monkey striate cortex Phil Trans R Soc Lond B 1977 278 377 409 19791
LeVay S Wiesel TN Hubel DH The development of ocular dominance columns in normal and visually deprived monkeys J Comp Neurol 1980 191 1 51 6772696
Buonomano DV Merzenich MM Cortical plasticity: From synapses to maps Annu Rev Neurosci 1998 21 149 186 9530495
Pons TP Garraghty PE Ommaya AK Kaas JH Taub E Massive cortical reorganization after sensory deafferentation in adult macaques Science 1991 252 1857 1860 1843843
Ramachandran VS Rogers-Ramachandran D Stewart M Perceptual correlates of massive cortical reorganization Science 1992 258 1159 1160 1439826
Darian-Smith C Gilbert CD Axonal sprouting accompanies functional reorganization in adult cat striate cortex Nature 1994 368 737 740 8152484
Florence SL Taub HB Kaas JH Large-scale sprouting of cortical connections after peripheral injury in adult macaque monkeys Science 1998 282 1117 1120 9804549
Denk W Strickler JH Webb WW Two-photon laser scanning fluorescence microscopy Science 1990 248 73 76 2321027
Trachtenberg JT Chen BE Knott GW Feng G Sanes JR Long-term in vivo imaging of experience-dependent synaptic plasticity in adult cortex Nature 2002 420 788 794 12490942
Grutzendler J Kasthuri N Gan WB Long-term dendritic spine stability in the adult cortex Nature 2002 420 812 816 12490949
Holtmaat AJ Trachtenberg JT Wilbrecht L Shepherd GM Zhang X Transient and persistent dendritic spines in the neocortex in vivo Neuron 2005 45 279 291 15664179
Zuo Y Lin A Chang P Gan WB Development of long-term dendritic spine stability in diverse regions of cerebral cortex Neuron 2005 46 181 189 15848798
Mizrahi A Katz LC Dendritic stability in the adult olfactory bulb Nat Neurosci 2003 6 1201 1207 14528309
Jacobs KM Donoghue JP Reshaping the cortical motor map by unmasking latent intracortical connections Science 1991 251 944 947 2000496
Jones EG GABAergic neurons and their role in cortical plasticity in primates Cereb Cortex 1993 3 361 372 8260806
Feng G Mellor RH Bernstein M Keller-Peck C Nguyen QT Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP Neuron 2000 28 41 45 11086982
Kim KH Buehler C So PTC High-speed, two-photon scanning microscope Appl Opt 1999 38 6004 6009 18324120
Feldman ML Peters A The forms of non-pyramidal neurons in the visual cortex of the rat J Comp Neurol 1978 179 761 793 346619
White EL Keller A Cortical circuits: Synaptic organization of the cerebral cortex—Structure, function and theory 1989 Boston Birkhauser 223
Peters A Kara DA The neuronal composition of area 17 of rat visual cortex. II. The nonpyramidal cells J Comp Neurol 1985 234 242 263 3988984
Kawaguchi Y Kubota Y GABAergic cell subtypes and their synaptic connections in rat frontal cortex Cereb Cortex 1997 7 476 486 9276173
Markram H Toledo-Rodriguez M Wang Y Gupta A Silberberg G Interneurons of the neocortical inhibitory system Nat Rev Neurosci 2004 5 793 807 15378039
Paxinos G Franklin KBJ The mouse brain in stereotaxic coordinates 2001 London Academic Press
Kaas JH Krubitzer LA Chino YM Langston AL Polley EH Reorganization of retinotopic cortical maps in adult mammals after lesions of the retina Science 1990 248 229 231 2326637
Gilbert CD Wiesel TN Receptive field dynamics in adult primary visual cortex Nature 1992 356 150 152 1545866
Darian-Smith C Gilbert CD Topographic reorganization in the striate cortex of the adult cat and monkey is cortically mediated J Neurosci 1995 15 1631 1647 7891124
Tsui CC Copeland NG Gilbert DJ Jenkins NA Barnes CA
Narp , a novel member of the pentraxin family, promotes neurite outgrowth and is dynamically regulated by neuronal activity J Neurosci 1996 16 2463 2478 8786423
Corriveau R Shatz CJ Nedivi E Dynamic regulation of cpg15 during activity-dependent synaptic development in the mammalian visual system J Neurosci 1999 19 7999 8008 10479700
Lein ES Shatz CJ Rapid regulation of brain-derived neurotrophic factor mRNA within eye-specific circuits during ocular dominance column formation J Neurosci 2000 20 1470 1483 10662837
Lee WCA Nedivi E Extended plasticity of visual cortex in dark-reared animals may result from prolonged expression of genes like cpg15
J Neurosci 2002 22 1807 1815 11880509
Hickmott PW Steen PA Large-scale changes in dendritic structure during reorganization of adult somatosensory cortex Nat Neurosci 2005 8 140 142 15657598
Mower GD Caplan CJ Christen WG Duffy FH Dark rearing prolongs physiological but not anatomical plasticity of the cat visual cortex J Comp Neurol 1985 235 448 466 3998219
Crair MC Gillespie DC Stryker MP The role of visual experience in the development of columns in cat visual cortex Science 1998 279 566 570 9438851
Trachtenberg JT Trepel C Stryker MP Rapid extragranular plasticity in the absence of thalamocortical plasticity in the developing primary visual cortex Science 2000 287 2029 2032 10720332
DePaola V Song S Holmaat A Wilbrecht L Caroni P Long-term imaging of axons and their synaptic terminals in the adult somatosensory cortex in vivo [abstract]. Neuroscience 2004; 23−27 October; San Diego, California. Society for Neuroscience 2004 Available: http://sfn.scholarone.com/itin2004 . Accessed 5 December 2005
Fairen A DeFelipe J Regidor J Peters A Jones EG Cellular components of the cerebral cortex Cerebral cortex 1984 New York Plenum 206 241
McBain CJ Fisahn A Interneurons unbound Nat Rev Neurosci 2001 2 11 23 11253355
Singer W Neurophysiology: The changing face of inhibition Curr Biol 1996 6 395 397 8723342
Hensch TK Fagiolini M Mataga N Stryker MP Baekkeskov S Local GABA circuit control of experience-dependent plasticity in developing visual cortex Science 1998 282 504 508
Huang ZJ Kirkwood A Pizzorusso T Porciatti V Morales B BDNF regulates the maturation of inhibition and the critical period of plasticity in mouse visual cortex Cell 1999 98 739 755 10499792
Hanover JL Huang ZJ Tonegawa S Stryker MP Brain-derived neurotrophic factor overexpression induces precocious critical period in mouse visual cortex J Neurosci 1999 19 1 5 9870932
Fagiolini M Hensch TK Inhibitory threshold for critical-period activation in primary visual cortex Nature 2000 404 183 186 10724170
Hensch TK Critical period regulation Annu Rev Neurosci 2004 27 549 579 15217343
Foeller E Feldman DE Synaptic basis for developmental plasticity in somatosensory cortex Curr Opin Neurobiol 2004 14 89 95 15018943
Dan Y Poo MM Spike timing-dependent plasticity of neural circuits Neuron 2004 44 23 30 15450157
Tao HW Poo MM Activity-dependent matching of excitatory and inhibitory inputs during refinement of visual receptive fields Neuron 2005 45 829 836 15797545
Feldman DE Timing-based LTP and LTD at vertical inputs to layers II/III pyramidal cells in rat barrel cortex Neuron 2000 27 45 56 10939330
Hendry SH Jones EG Reduction in number of immunostained GABAergic neurones in deprived-eye dominance columns of monkey area 17 Nature 1986 320 750 753 3703001
Ruthazer ES Cline HT Multiphoton imaging of neurons in living tissue: Acquisition and analysis of time-lapse morphological data Real-Time Imag 2002 8 175 188
Chattopadhyaya B Di Cristo G Higashiyama H Knott GW Kuhlman SJ Experience and activity-dependent maturation of perisomatic GABAergic innervation in primary visual cortex during a postnatal critical period J Neurosci 2004 24 9598 9611 15509747
|
16366735
|
PMC1318477
|
CC BY
|
2021-01-05 08:21:47
|
no
|
PLoS Biol. 2006 Feb 27; 4(2):e29
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0040029
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpmedplosmedPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1636673610.1371/journal.pbio.0040031Research ArticleCell BiologyPhysiologyDiabetes/Endocrinology/MetabolismMus (Mouse)MammalsVertebratesAnimalsEukaryotesSirt1 Regulates Insulin Secretion by Repressing UCP2 in Pancreatic β Cells Sirt1 Regulates Insulin SecretionBordone Laura
1
*¤Motta Maria Carla
1
Picard Frederic
2
Robinson Ashley
1
Jhala Ulupi S
3
Apfeld Javier
4
McDonagh Thomas
4
Lemieux Madeleine
5
McBurney Michael
5
Szilvasi Akos
6
Easlon Erin J
7
Lin Su-Ju
7
Guarente Leonard
1
*1Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of America2Laval Hospital Research Center, Québec City, Québec, Canada3The Whittier Institute for Diabetes, University of California San Diego, La Jolla, California, United States of America4Elixir Pharmaceuticals, Cambridge, Massachusetts, United States of America5Department of Medicine and Department of Biochemistry, Microbiology, and Immunology, University of Ottawa, and the Ottawa Regional Cancer Centre, Ottawa, Ontario, Canada6Novartis Institutes for BioMedical Research, Cambridge, Massachusetts, United States of America7Center for Genetics and Development, and Section of Microbiology, University of California Davis, Davis, California, United States of AmericaDillin Andy Academic EditorThe Salk InstituteUnited States of America* To whom correspondence should be addressed. E-mail: [email protected] (LB); E-mail: [email protected] (LG)¤ Current address: Novartis Institutes for BioMedical Research, Cambridge, Massachusetts, United States of America
2 2006 27 12 2005 27 12 2005 4 2 e3122 8 2005 22 11 2005 Copyright: © 2006 Bordone et al.2006This 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 Sirt1 Gene Promotes Insulin Secretion in Accord with Diet
Sir2 and insulin/IGF-1 are the major pathways that impinge upon aging in lower organisms. In Caenorhabditis elegans a possible genetic link between Sir2 and the insulin/IGF-1 pathway has been reported. Here we investigate such a link in mammals. We show that Sirt1 positively regulates insulin secretion in pancreatic β cells. Sirt1 represses the uncoupling protein (UCP) gene UCP2 by binding directly to the UCP2 promoter. In β cell lines in which Sirt1 is reduced by SiRNA, UCP2 levels are elevated and insulin secretion is blunted. The up-regulation of UCP2 is associated with a failure of cells to increase ATP levels after glucose stimulation. Knockdown of UCP2 restores the ability to secrete insulin in cells with reduced Sirt1, showing that UCP2 causes the defect in glucose-stimulated insulin secretion. Food deprivation induces UCP2 in mouse pancreas, which may occur via a reduction in NAD (a derivative of niacin) levels in the pancreas and down-regulation of Sirt1. Sirt1 knockout mice display constitutively high UCP2 expression. Our findings show that Sirt1 regulates UCP2 in β cells to affect insulin secretion.
Sirt1 is shown to regulate the expression of the metabolic decoupling gene UCP2 in pancreatic β cells, highlighting a possible role for Sirt1 in coordinating insulin release in response to changing dietary conditions.
==== Body
Introduction
Glucose homeostasis is maintained, in part, by pancreatic β cells, which secrete insulin in a highly regulated sequence of dependent events [1]. β cells metabolize glucose, resulting in an increase in the ATP/ADP ratio, the closing of the ATP-dependent K+ channel, the activation of the voltage-gated Ca+ channel and Ca+ influx, and the fusion of secretory vesicles to the plasma membrane to release insulin. Insulin is part of an organismal physiological axis in which it stimulates glucose uptake in metabolic tissues, such as muscle, and stores energy in the form of fat in white adipose tissue (WAT). Short-term food limitation (i.e., overnight [O/N] fasting) will therefore elicit the mobilization of glycogen stores and then fat from WAT for metabolism, and the lower level of blood glucose during fasting will result in low levels of insulin production by β cells.
Long-term calorie restriction (CR) has been known for 70 years to extend the life span of mammals dramatically [2], and it can also work in a variety of organisms, including yeast, flies, and rodents [3–5], although the mechanism of this effect has remained obscure. In mammals a characteristic set of physiological changes takes place during long-term CR, which overlaps the rapid physiological adaptations to short-term food limitation. One such change is the use of dietary fat or fat mobilized from WAT for energy [4]. Another is a large reduction in blood insulin levels accompanied by an increase in insulin sensitivity, i.e., the ability of insulin to promote glucose utilization [4]. In addition, gluconeogenesis is activated in the liver. These changes keep glucose available for the brain, and are closely associated with the longevity elicited by CR. The paucity of fat in WAT appears to be sufficient per se to promote a degree of longevity, since mice engineered for leanness—for example, a WAT-specific knockout (KO) of the insulin receptor—live longer [6,7].
Findings in model organisms suggest a mechanism for the longevity engendered by CR that implicates the silent mating type information regulation 2 gene (Sir2). This gene regulates the life span in yeast [8] and Caenorhabditis elegans [9] as a longevity determinant. In yeast, CR works by up-regulating the activity of Sir2 [10,11], a NAD-dependent deacetylase [12–14] (NAD is a derivative of niacin), by increasing respiration, and by increasing the NAD/NADH ratio [15] (NADH is the reduced form of NAD). CR is also reported to activate the NAD salvage pathway, which would deplete a Sir2 inhibitor, nicotinamide [3,10]. The Drosophila melanogaster Sir2 gene was also shown to mediate life extension in response to dietary restriction [16,17].
Since Sir2 appears to mediate the effects of CR on life span in simple model organisms, it seemed possible that Sir2 proteins also regulate the effects of food limitation and CR in mammals. The homolog of the yeast silencing information regulator2 (Sirt1) has also been implicated in several aspects of food limitation and CR in mammals. In WAT, Sirt1 represses the key regulatory protein peroxisome proliferator-activated receptor gamma (PPARγ), resulting in fat mobilization in response to food limitation [18]. In addition, Sirt1 regulates the FOXO (forkhead Box O) set of forkhead transcription factors [19,20], providing another link to metabolism and diet. Also, gluconeogenesis in the liver is regulated by Sirt1 [19], which works in concert with the transcriptional co-activator, peroxisome proliferator-activated receptor coactivator, PGC-1α [21]. Finally, Sirt1 may play a role in the observed stress resistance of CR animals, since it down-regulates several pro-apoptotic factors, such as p53, FOXO, and Bax [19,20,22–25].
In addition to the classical paradigm for insulin regulation by glucose outlined above, reports suggest a role of an uncoupling protein (UCP) in insulin secretion. UCPs belong to a family of mitochondrial inner membrane proteins. They function to uncouple oxygen consumption during respiration from the production of ATP by allowing proton leakage down an electrochemical gradient from the cytoplasm to the mitochondria [26–29]. Several mammalian UCP homologues have been identified and characterized [28]. UCP2 has been shown to promote proton leakage across the mitochondrial membrane [30–33], and has been proposed to play a role in lipid metabolism, insulin resistance, glucose utilization, regulation of reactive oxygen species, and macrophage-mediated immunity [28,34,35]. A role for UCP2 in insulin secretion has been demonstrated, since UCP2 KO mice show higher ATP levels in islets and increased insulin secretion upon glucose stimulation [36]. Conversely, UCP2 overexpression in cultured β cell lines reduces ATP levels and glucose-stimulated insulin secretion [37,38].
In this study, we show that Sirt1 functions as a positive regulator of insulin secretion in response to glucose by directly repressing the UCP2 gene. Our findings suggest that UCP2 is thus used in β cells to modulate the insulin response pathway as a function of diet. The fact that Sirt1 is a positive regulator of insulin may seem surprising, in light of the fact that sir-2.1 in C. elegans appears to be a negative regulator of the insulin-like signaling pathway [9]. We discuss how this difference makes sense in light of the physiological adaptations that take place in mammals in response to dietary changes.
Results
Sirt1 Is Preferentially Expressed in Pancreatic β Cells
To determine whether Sirt1 could potentially play a role in β cells, immunofluorescence of whole murine pancreas with anti-Sirt1 antibody was carried out, and 20 islets were examined by fluorescence microscopy. A representative islet is shown in Figure 1A (see lower left panel for hematoxylin-eosin staining). All pancreatic sections examined showed intense staining concentrated in islets using Sirt1 antibodies (Figure 1A, top right) but not with secondary antibody alone (Figure 1A, lower right), and a lower level of expression in the surrounding exocrine cells. DAPI bright spots (Figure 1A, top left) mark nuclei and their corresponding cells in the islets and surrounding tissue. This enrichment of Sirt1 in the islets and not exocrine cells is noteworthy, in light of the ubiquitous expression of this sirtuin in most somatic and germ tissues [39,40].
Figure 1 Sirt1 Is Localized in the Islets of Langerhans
The pancreas of wild-type mice was sectioned and stained as described.
(A) Nuclear staining using DAPI (top left). Immunofluorescence using Sirt1 antibody (top right); hematoxylin and eosin staining of the same section of pancreas (bottom left); immunofluorescence control using a rabbit secondary antibody (bottom right).
(B) Pancreases of wild-type (WT), Sirt1+/− heterozygotes (HET), or Sirt1−/− homozygous KO mice were stained with antibodies against insulin (blue), glucagon (red), or somatostatin (green) (shown in left column). Representative islets of mice of all three genotypes are shown. Pancreases were also silver-stained for morphometry (right column). Islets appear as dark figures and their area was determined by scanning, using Image-Pro 4.1 Plus software.
(C) The areas are shown as percentage of area of the entire pancreas.
Sirt1 KO Mice Have Low Levels of Blood Insulin
Because Sirt1 is highly expressed in β cells, we investigated whether this sirtuin plays a functional role in insulin production. First, we determined whether Sirt1 KO mice [39,40] showed any defects in the pancreatic β cells or the islets. Pancreases from wild-type, Sirt1+/− heterozygotes, or Sirt1−/− KO mice (4−5 each) were sectioned and stained with antibodies against insulin (blue), glucagons (red), and somatostatin (green). These stains mark islets for β cells, α cells, and δ cells, respectively. As shown in a typical section (Figure 1B), no differences were observed in the staining pattern of these three markers between wild-type and mutant islets. We measured pancreatic islet areas using Image-Pro 4.1 Plus software, and expressed the islet areas as a percentage of the total pancreatic area (Figure 1C). There were no significant differences in islet areas comparing wild-type, heterozygous, and KO mice. The absolute islet size was also not appreciably different in the mutant mice.
Next, we determined whether insulin production was altered in Sirt1 KO mice. Blood insulin was measured in males 2–4 mo old fed ad libitum or after 18 h of starvation (Figure 2A). Sirt1 KO mice (black bars) had much lower blood levels of insulin compared with littermate control animals (open bars) when the samples were collected under ad libitum conditions. This difference was also observed when the animals were starved O/N, in which case insulin levels were very low in both wild-type and KO mice. To assess more precisely the ability of animals to produce insulin, mice were given an injection of glucose after O/N starvation and insulin was measured after 2, 10, and 20 min. Whereas insulin induction was clearly evident in the wild-type mice at 10 min (open bars), induction was not observed in Sirt1 KO animals (black bars in Figure 2B, n = 4 or 5 for each measurement). A similar trend was noted at 20 min.
Figure 2 Sirt1 KO Mice Have a Lower Level of Insulin
(A) Plasma insulin levels in wild-type (open bars) or Sirt1 KO mice (black bars) ad libitum or after O/N starvation (n = 12 wild-type, 11 KO, *p < 0.03 in ad libitum and O/N starvation mice, ANOVA).
(B) Plasma insulin levels in Sirt1 KO mice (black bar) compared with wild-type mice (open bars) 2, 10, or 20 min after injection with glucose (n = 4 or 5, *p < 0.05 compared with wild-type, ANOVA).
(C) Insulin secretion in islets isolated from wild-type (open bars) or Sirt1 KO mice (black bars) after induction by 20 mM glucose for 1 h (n = 4, *p < 0.005 in wild-type, ANOVA).
(D) Glucose levels in wild-type (open bars) and Sirt1 KO (black bars) mice (n = 12 wild-type, 11 KO, *p < 0.03 ad libitum, ANOVA).
(E) Glucose tolerance tests in wild-type (black) and Sirt1 KO (green) mice (n = 6, *p < 0.05 at 20, 40, 60, and 120 min).
To further investigate the defect in insulin production, islets from four wild-type or four Sirt1 KO animals were isolated and incubated in vitro with or without glucose for 1 h, and insulin secretion was determined. As shown in Figure 2C, the basal level of insulin secreted into the media from islets of Sirt1 KO mice (black bars) was significantly lower than wild-type controls (open bars). Moreover, the islets isolated from Sirt1 KO mice were not induced to secrete insulin by glucose, while control islets were inducible, as expected.
Levels of blood glucose were then determined and, surprisingly, were lower in Sirt1 KO mice (Figure 2D). Further, these mice appeared to be hypermetabolic, since they ate more food per body weight than the wild-type (unpublished data). These findings suggested that the KO mice had a better ability to use the lower levels of insulin for glucose uptake, i.e., were more insulin sensitive. To address this possibility, we performed glucose tolerance tests in both wild-type and KO mice by injecting glucose intraperitoneally and measuring the kinetics of glucose clearance from the blood. The KO mice cleared the glucose significantly faster than the wild-type (Figure 2E). This increased glucose tolerance explains why the KO mice maintained lower levels of blood glucose, even with reduced levels of insulin. The surprising effect of knocking out Sirt1 on glucose tolerance likely derives from tissues other than the pancreas and is currently being investigated in greater detail. In summary, we do not know whether the reduction in insulin in Sirt1 KO mice is due to a β cell defect or is an indirect consequence of increased glucose tolerance in these animals.
Sirt1 Drives Glucose-Induced Insulin Secretion in Cultured Cells
To address whether Sirt1 played a positive role in insulin production specifically in β-cells, the pancreatic β-cell lines, INS-1 from rat, and MIN6 from mouse were employed. INS-1 cells were grown in culture, and nuclear expression of Sirt1 was evident by immunofluorescence (Figure 3) and Western blot (Figure 3C and 3E). INS-1 or MIN6 cells were treated with 10 mM nicotinamide, an inhibitor of Sirt1 [41], for 48 h, and the levels of insulin secreted into the media were measured without and with induction by glucose (Figure 3B and 3F, respectively). Whereas control cells showed robust induction of insulin secretion by glucose, nicotinamide treatment blocked this induction in both cell lines.
Figure 3 Sirt1 Is a Positive Regulator of Insulin Secretion in INS-1 and MIN6 Cells
(A) Immunofluorescence in INS-1 cells using Sirt1 antibody (green) and DAPI staining (blue). Nuclear localization of Sirt1 is evident.
(B) Induction of insulin secretion in INS-1 cells with 16.7 mM glucose (+) compared with 4 mM glucose control (−). The left side shows no nicotinamide and the right side shows treatment with 10 mM nicotinamide for 48 h prior to induction (n = 3 experiments done in triplicate, *p < 0.05 in the no nicotinamide experiment, ANOVA).
(C) Western blot of Sirt1 in INS-1 cells with knockdown levels of the protein (SiRNA Sirt1) compared with control cells (pSUPER).
(D) INS-1 cells infected with the pSUPERretro SiRNA-GFP control (open bars) or pSUPER retro SiRNA-Sirt1 knockdown cells (black bars) were induced for insulin secretion as in (B) (n = 3 experiments done in triplicate, *p < 0.008 in the control experiment, ANOVA).
(E) Western blot of Sirt1 in MIN6 cells with knockdown levels of the protein (SiRNA Sirt1) compared with control cells (pSUPER).
(F) Glucose induction (20 mM versus 4 mM) of insulin secretion in MIN6 cells with the pSUPER control vector (open bars) or the SiRNA Sirt1 vector (black bars) in the absence or presence of nicotinamide (n = 3 experiments done in triplicate*p < 0.05 in the control without nicotinamide, ANOVA).
(G) Glucose uptake in INS-1 cells stably transfected with control or SiRNA Sirt1 vectors. 2-NBDG fluorescence was determined by flow cytometry 10 min after addition and expressed as arbitrary units (n = 2, *p < 0.0005 compared with no glucose).
To test if the effects of nicotinamide were due specifically to inhibition of Sirt1, levels of this sirtuin were lowered by RNA interference. INS-1 and MIN6 cells were infected with a retrovirus carrying a Sirt1-SiRNA construct [18] or a control vector (pSUPER-SiRNA GFP), and stable cell lines were created as puromycin-resistant infected pools. These cells displayed Sirt1 protein levels that were knocked down compared with drug-resistant control pools generated from the vector (Figure 3C [INS-1] and 3E [MIN]). When these stable Sirt1 knockdown cells (Figure 3D, 3F, and 3G) were treated with glucose, induction of insulin secretion was eliminated, whereas contemporaneous assays of vector control cells (open bars) showed normal induction (Figure 3D and 3F).
We then tested if the defect in glucose-induced insulin secretion was due to a different glucose uptake between control and knockdown cells by measuring transport of fluorescence analog 2-NBDG. Glucose uptake [42,43] was not reduced and was perhaps slightly elevated in this assay in the knockdown cells (Figure 3G). Further, no decrease in the glucose transporter Glut 2 was observed (unpublished data). Finally, cell growth measurements revealed no difference in the growth rate of INS-1 knockdown cells compared with controls (Figure S1). These results show that knockdown of Sirt1 suppresses glucose-stimulated insulin secretion in two different β cell lines.
ATP/ADP Ratio Is Lower in Sirt1 Knockdown Cells
We next sought to investigate the mechanism by which Sirt1 regulates insulin secretion in β cells. First, we assayed the RNA level of the INS1 gene, encoding insulin, by Northern blot and found no difference in control versus Sirt1 knockdown β cells (Figure 4A). Similarly, we found no difference by Western blot in expression level of c/EBPβ, a transcription factor that regulates INS1 (Figure 4B). Because the insulin receptor is part of an autocrine induction loop for insulin [44], we also determined by Western blot the levels of this receptor (α and β subunits) and observed no difference between control and knockdown cells (Figure 4B). Next, we investigated whether Sirt1 affected the levels of the K+ channel by Western blot, and again found no effect (Figure 4C). While this analysis suggests that Sirt1 does not function by altering levels of these factors, it does not rule out the possibility that Sirt1 regulates their activity.
Figure 4 UCP2 is Up-Regulated in Sirt1 Knockdown Cells and in Sirt1 KO Mice
(A) Northern blot for the insulin gene INS-1 in cells with a control vector (pSUPER) or a SiRNA-Sirt1 vector (SiRNA Sirt1).
(B and C) Western blot analyses of targets involved in insulin synthesis and secretion using specific antibodies: cEBP/β, insulin receptor α and β, and kir6.2, one of the K+ channel receptor subunits.
(D) Measurement of ATP/ADP levels in INS-1 control cells (open bars) or Sirt1 knockdown cells (black bars) treated with 16.7 mM glucose (+) or 4 mM glucose (−) (n = 3 experiments done in triplicate, *p < 0.005 in the pSUPER experiment; ANOVA).
(E) Western blot analysis for UCP2 in INS-1 control cells (pSUPER) or knockdown cells (SiRNA Sirt1).
(F) Northern blot analysis for UCP2 in INS-1 control cells (pSUPER) or knockdown cells (SiRNA Sirt1).
(G) NADH levels in INS-1 cells after glucose addition as determined by autofluorescence [46] and expressed as arbitrary units. Cells stably transfected with control or Sirt1 SiRNA vectors were used (n = 2, *p < 0.05 compared with no glucose).
(H) UCP2 protein levels in isolated pancreatic islets of two wild-type or two Sirt1 KO mice. Tubulin or actin was used as loading control in all Western and Northern blots.
Finally, because of the central role of ATP in insulin secretion, we surmised that Sirt1 could play a role in the energetics of glucose utilization in β cells. We thus measured the ATP/ADP ratio in control and Sirt1 knockdown INS-1 cells after glucose induction. Control cells (Figure 4D, open bars) responded to glucose by increasing the ratio of ATP to ADP, as expected (Figure 4D). In contrast, in cells with the knockdown levels of Sirt1 (Figure 4D, black bars), the ATP/ADP ratio did not increase upon glucose stimulation. These findings show that knocking down Sirt1 results in a defect in ATP production in response to glucose. Basal ATP levels are not significantly altered in knockdown cells, consistent with the fact that they grow as well as control cells.
UCP2 Levels Are Increased in Sirt1 KO Mice and Knockdown Cells
One possibility for the failure of the Sirt1 knockdown cells to make ATP is that respiration is more uncoupled than in control cells, which would square well with the known link between the UCP2 and insulin production in β cells (36). Thus, we determined by Western blot the levels of UCP2 in control and Sirt1 knockdown cells. Strikingly, there was a significant increase in the UCP2 protein level in the knockdown cells (Figure 4E). This increase in protein level was mirrored by an increase in the mRNA in the same cells (Figure 4F), indicating that Sirt1 regulates UCP2 transcription. It was previously shown that UCP2 expression reduced NADH levels [45]. We therefore determined NADH levels by autofluorescence [46] and found a significantly lower level of NADH in the knockdown cells (Figure 4G).
Western blot for UCP2 in Sirt1 KO mice showed a similar effect. We observed an increase in UCP2 protein in the whole pancreas (unpublished data), which is a measure of the islets, since UCP2 is expressed in only the endocrine cells of the pancreas [38]. Moreover, UCP2 protein was also up-regulated in isolated islets of Sirt1 KO mice (Figure 4H). In summary, our expression studies suggest that Sirt1 is a repressor of UCP2 transcription in β cells, and by repressing this UCP, this sirtuin may allow cells to secrete insulin in response to glucose.
Sirt1 Binds to the UCP2 Promoter
To study further the repression of UCP2 by Sirt1, we carried out reporter assays in 293T cells transfected with a chloramphenicol acetyl-transferase (CAT) gene, whose expression is driven by the UCP2 promoter. Cells were also transfected with a control vector or with an expression vector for PPARγ, which is known to bind to and activate the UCP2 promoter [47]. In a control experiment, PPARγ activated the reporter in this assay (Figure 5A, left bars). In a parallel experiment, cells were also co-transfected with a Sirt1 expression vector. Sirt1 clearly repressed the activation of the UCP2 promoter (Figure 5A, right bars). Repression of this reporter by endogenous or expressed Sirt1 was alleviated by nicotinamide, a known inhibitor of Sirt1 (unpublished data).
Figure 5 Sirt1 Binds at the UCP2 Promoter and Represses the Gene
(A) In vitro CAT assay. 293T cells were transfected with a CAT reporter driven by the UCP2 promoter. Cells were also co-transfected with Sirt1 or not and with PPARγ or not, as indicated. CAT activity was determined (n = 3 experiments done in triplicate, *p < 0.05 in the no Sirt1 transfection experiment, ANOVA).
(B) Schematic representation of the primer sets (arrows) in the UCP2 promoter (shown schematically and with excerpted DNA sequence).
(C) Chromatin-immunoprecipitation (IP) was carried out on INS-1 control cells (lanes 1–3) or Sirt1 knockdown cells (columns 4–6) using Sirt1 antibody or a Gal4 control antibody, as indicated. PCR was carried out with the indicated primers. INPUT (columns 7–10) refers to PCR carried out on samples prepared prior to immunoprecipitation. Negative controls for the PCR (minus DNA) are also indicated (columns 11 and 12).
To determine whether repression of UCP2 was due to the direct binding of Sirt1 at the promoter, INS-1 cells were subjected to chromatin-immunoprecipitation, using Sirt1 or control antibodies. Primers specifically designed to span a known regulatory region of the UCP2 promoter (43) (Figure 5B, primer set 5) were used to probe by PCR the DNA in the immunoprecipitate. Sirt1 bound to this region of the UCP2 promoter in the control cells (Figure 5C, column 3) and to a significantly lesser extent in the Sirt1 knockdown cells (Figure 5C, column 6). Control primers designed in a different region of UCP2 upstream DNA (Figure 5B, primer set 4) showed no amplification (Figure 5C, columns 2 and 5). An anti-gal4 antibody control using primer set 5 also showed no binding (Figure 5C, columns 1 and 4). These findings indicate that Sirt1 represses UCP2 transcription by binding directly at the UCP2 promoter.
Reduction of UCP2 Restores Insulin Secretion in Sirt1 Knockdown β cells
The above findings show that Sirt1 represses UCP2, and alleviation of this repression correlated with blunted insulin secretion in response to glucose. To address whether the increase in UCP2 in Sirt1 knockdown cells caused the failure to secrete insulin, UCP2 was also knocked down in INS-1 cells with reduced Sirt1. Stable cell lines with UCP2 knocked down by SiRNA were derived from two different lines in which Sirt1 was already knocked down, as well as from control INS-1 cells. Several different UCP2 sequences were inserted into a pSUPER hairpin vector with the neomycin-resistance drug marker. These constructs were transfected into the SiRNA Sirt1 puromycin resistant cells or control cells, and stable populations of NeoR cells were assayed for UCP2 RNA and protein by Northern and Western blots. In the case of two different UCP2 SiRNA constructs, the levels of UCP2 RNA (Figure 6A) and protein (unpublished data) were clearly reduced in transfected cells. The levels of Sirt1 remained low in the double knockdown cells (unpublished data).
Figure 6 Knockdown of UCP2 in Sirt1 Knockdown Cells Restores Glucose-Induced Insulin Secretion
(A) Northern blot for UCP2 RNA in control INS-1 cells, and cells knocked down for Sirt1 (SiRNA Sirt1), UCP2 (SiRNA UCP2), or both Sirt1 and UCP2 (SiRNA Sirt1-SiRNA UCP2). RNAs were quantitated by densitometry, setting the level of UCP2 in control cells at 1.0.
(B) Insulin secretion in INS-1 control cells and cells with knockdown levels of Sirt1, UCP2, or both Sirt1 and UCP2 after treatment with 16.7 mM glucose (+) or 4mM glucose (−) for 1 h (n = 3 experiments done in triplicate, *p < 0.05 in SiRNA Sirt1-SiRNA UCP2, ANOVA).
Next, glucose-stimulated insulin secretion was assayed in Sirt1 or UCP2 knockdown cells and in cells with both Sirt1 and UCP2 knocked down. The Sirt1 knockdown cells were again defective in induction of insulin secretion, as expected. However, the double knockdown cells or cells with only the UCP2 SiRNA construct displayed insulin secretion in response to glucose (Figure 6B). A similar effect was observed in the second Sirt1 knockdown line in which UCP2 was also knocked down (unpublished data). This result demonstrates that the failure of Sirt1 knockdown cells to secrete insulin is due to the elevated levels of UCP2 in these cells. We conclude that Sirt1 acts as a positive regulator of insulin secretion in wild-type cells by repressing UCP2, thereby allowing coupling of glucose metabolism to ATP synthesis.
UCP2 Levels Increase in Food-Deprived Mice
Does the regulation of UCP2 play any role in the normal secretion of insulin in β cells of wild-type mice in response to diet? To address this question, we starved wild-type mice O/N and compared the levels of UCP2 in whole pancreas and in islets to mice feeding ad libitum. Importantly, we found an increase in UCP2 mRNA levels in whole pancreas of starved mice compared with the fed mice (Figure 7A). Further, the levels of UCP2 protein were also increased in starved mice (Figure 7B).
Figure 7 UCP2 mRNA or Protein Levels in Fed or Starved Wild-Type Mice
(A) Northern blot for UCP2 in whole pancreas of two ad libitum mice and two mice starved for 18 h.
(B) Western blot for UCP2 in isolated islets in two ad libitum and two starved mice.
(C) Western blot for UCP2 in wild-type (WT) or Sirt1 KO littermates either fed ad libitum or starved for 18 h. The experiment shown is representative of four pairs of wild-type and KO littermates analyzed.
(D) RT-PCR for UCP2 in wild-type or Sirt1 KO mice fed or starved.
In order to determine whether Sirt1 regulated this induction of UCP2 in mice, we starved Sirt1 KO mice O/N and compared the effect of starvation on UCP2 protein levels to wild-type mice. Four pairs of wild-type and KO littermates were compared and gave comparable results. The levels of UCP2 in KO mice fed ad libitum were elevated compared with wild-type, as expected (Figure 7C). Most importantly, these elevated levels in the fed KO mice were not further induced by starvation. In contrast, starvation induced UCP2 in wild-type mice, as before. A similar pattern of UCP2 RNA induction by starvation in wild-type but not KO mice was observed by RT-PCR (Figure 7D). The above findings suggest that an increase in UCP2 in β cells is part of a normal mechanism to regulate the capacity of β cells to produce insulin. Moreover, this induction appears to be mediated by alleviation of Sirt1-mediated repression.
These findings suggest that starvation causes a decrease in Sirt1 activity in β cells. In yeast, Sir2p activity is regulated by the NAD/NADH ratio. We thus measured NAD and NADH in the pancreas of seven fed and seven starved wild-type mice. Strikingly, there was a significant decrease in the level of NAD but not NADH in the starved mice (Figure 8). The level of Sirt1 protein in starved pancreas is roughly comparable to fed controls (unpublished data). These findings suggest that changes in NAD levels in the pancreas regulate Sirt1 activity and insulin secretion in response to diet.
Figure 8 NAD and NADH Levels in Fed and Starved Mice
Measurements were made in pancreases of seven fed and seven starved wild-type mice, and levels are expressed as nmol per gram of tissue. The decrease in NAD in starved mice is significant with p < 0.0005, while the NADH levels in fed versus starved are not significantly different.
Discussion
In this study we show that Sirt1 regulates insulin secretion in pancreatic β cells. Both in Sirt1 KO mice and in cultured β cell lines, a reduction in Sirt1 levels reduces the capacity of these cells to secrete insulin in response to glucose. Classical studies show that an increase in the ATP/ADP ratio due to glucose metabolism is the critical trigger in the induction of insulin secretion. In cells with reduced Sirt1, the increase in this ratio is blunted due to elevated levels of UCP2, which reduces the synthesis of ATP during respiration. Sirt1 directly represses UCP2 by binding to the promoter. Thus, by controlling UCP2, Sirt1 regulates the amplitude of insulin induction by glucose. We do not know whether the low levels of insulin in the whole body of Sirt1 KO mice represents a β cell defect, an indirect consequence of the elevated glucose tolerance in these animals, or both.
The use of UCP2 to regulate insulin production is distinct from the canonical role of UCPs. By uncoupling respiration from ATP synthesis in brown fat, for example, UCP1 generates heat and is critical for the non-shivering thermogenesis in rodents (reviewed in [48]). Here UCP2 is used to modulate the levels of ATP made in β cells, which dictates the extent of insulin secretion. Our findings are consistent with the elevated basal levels of insulin observed in fasted UCP2 KO mice [36] and suggest that Sirt1 and UCP2 are physiologically relevant regulators of insulin production (see “UCPs and Longevity–Cellular Basis”).
Physiology of Insulin Regulation by Sirt1
The fact that Sirt1 represses UCP2 expression in β cells is intriguing in light of the role of this sirtuin as a regulator of CR in lower organisms. Our findings indicate that Sirt1 may link the amplitude of insulin secretion to the diet. We found that O/N starvation led to an increase in the levels of UCP2 protein and RNA in the pancreas and isolated islets. This increase did not occur in Sirt1 KO mice, which already displayed higher levels of UCP2 when fed ad libitum. These findings suggest that regulation of UCP2 by Sirt1 is a physiologically important mechanism to blunt the chronic levels of insulin in starved animals. This mechanism may explain why insulin levels are so low during fasting, even though β cells are actively metabolizing fat and respiring (see following section). It remains to be seen whether this mechanism is important during long-term CR.
Regulation of Pancreatic Sirt1 Activity by NAD/NADH and Diet
The action of mammalian Sirt1 appears to differ from that of lower organisms. In C. elegans, sir2.1 appears to repress the output of the insulin/IGF pathway [9], but the mechanism has not been described. Also, Sir2 activity is activated during CR to extend the life span in yeast and Drosophila [11,49]. However, in mammals Sirt1 functions as a positive regulator of insulin secretion, raising the possibility that its activity in β cells is actually reduced by O/N starvation..
A lowering of Sirt1 activity occurs concomitant with the shift from carbohydrate-based metabolism to utilization of fatty acids, which is known to follow food deprivation. Because fatty acids are more reduced than carbohydrates, their metabolism to CO2 converts more NAD to NADH. Indeed, we show that the NAD/NADH ratio falls in pancreas of starved mice compared with fed controls, although we cannot be certain this change occurs specifically in islets, which represent a small percentage of pancreatic mass. A decrease in the NAD/NADH ratio has also been shown to down-regulate Sir2 activity in other physiological contexts [15,21,50].
UCPs and Longevity–Cellular Basis
The question arises whether the up-regulation of UCPs we find in β cells occurs in other tissues during food limitation or long-term CR. In a horizontal study of mice, a good correlation was found between the life span and the level of uncoupling of respiration [51]. It has also been shown that overexpression of a UCP increased the life span in Drosophila [52]. UCPs may confer benefit by damping production of reactive oxygen species during respiration. In one example, mitochondria from UCP3 KO mice generated higher levels of superoxide [53,54], suggesting that some degree of uncoupling can be beneficial by controlling reactive oxygen species production [55].
There are several reports that levels of UCPs may increase as a consequence of CR [56,57], but these findings are still not conclusive [58]. It will be interesting to see whether Sirt1 regulates UCPs in tissues other than β cells, and whether there is a broader change in the expression of UCPs during acute food limitation or CR.
Summary and Perspective
The induction of insulin secretion by glucose in pancreatic β cells has been a paradigm in mammalian cellular physiology. Here we describe a new level of regulation in which Sirt1 represses UCP2 to modulate the amplitude of insulin induction by glucose in β cells. We suggest that this mechanism serves to regulate chronic levels of insulin in accord with levels of food intake. This mechanism may reinforce the low levels of insulin production during food limitation and coordinate insulin release from β cells to the insulin sensitivity of metabolic tissues set by the diet. Regulation of UCP2 by Sirt1 may also be an important axis that is dysregulated by excess fat to contribute to obesity-induced diabetes.
Materials and Methods
Cell culture
INS-1 (a gift from K. Olson, Michigan State University, Lansing, Michigan, United States) and MIN6 cells (a gift from S. Imai, Washington University in St. Louis, St. Louis, Missouri, United States) were grown at 5% CO2/95% air at 37 °C. INS-1 cells were cultured in RPMI-1640 medium containing 11.1 mM glucose supplemented with 10% fetal bovine serum (FBS), 1 mM pyruvate, 10 mM HEPES, 50 μM 2-mercaptoethanol, 100 U penicillin/ml, and 100 μg/ml streptomycin. Cells were passaged weekly after trypsin-EDTA treatment. All studies were performed in INS-1 passages between 70 and 84. MIN6 cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 4.5 g/l glucose and L-glutamine, 15% FBS, 3.4 g/l sodium bicarbonate, 5 μl/l 2-mercaptoethanol, 100 U penicillin/ml, and 100 μg/ml streptomycin. All studies were performed between passages 27 and 35.
Animal experimentation
The Sirt1 KO animals and their controls were obtained from M. McBurney [39]. The mice were housed under controlled conditions: temperature (25 ± 1 °C) and light cycle (7 a.m.–7 p.m.). Animals were fed with regular chow and cared for in accordance with the MIT Committee on Animal Care (Massachusetts Institute of Technology, Cambridge, Massachusetts, United States). Assays were performed on 4- to 8-mo-old males (129/Sv CD1 mixed background).
Retroviral infection of INS-1 and MIN6 cells
Phoenix cells were transfected with either pBABE, pBABE-T1 [18], pSUPERretro (Oligogene), pSUPERretro-SiRNA-T1 (5′-
GCTGCATCCAAGGGCCATG-3′), or pSUPER retro SiRNA-GFP (gift from T. Brummelkamp, Whitehead Institute for Biomedical Research, Cambridge, United States of America) using Lipofectamine 2000 (Invitrogen, Carlsbad, California, United States). 48 h post-transfection, the medium containing retroviruses was collected and added with 1 mg/ml polybrene to INS-1 or MIN6 cells. Infected cells were treated with 1 mg/ml of puromycin for selection.
In order to create double knockdown cell lines, INS-1 cells stably transfected with control vector and pSUPERretro-SiRNA-T1 were infected with retroviruses carrying pSUPER retro-SiRNA UCP2. These cells were selected with 200 μg/ml gentamicin (Mediatech, Herndon, Virginia, United States).
Cell-growth analysis
INS-1 cells were seeded at 1 × 105 cells/6 cm plate. Cells were collected, resuspended in the same volume of medium, and counted using the trypan blue exclusion method.
Insulin secretion assay
0.5 × 106 cells were placed in 12-well plates in growing medium. In the case of INS-1 cells, fresh medium containing 4.0 mM glucose was added after 2 d. Cells were cultured for an additional day. In all cases, the medium was removed on the day of the experiment, and the cells washed three times with warm KRB buffer (119 mM NaCl, 4.74 mM KCl, 2.54 mM CaCl2, 1.19 mM MgS04, 1.19 mM KH2P04, 25 mM NaHC03, 10 mM HEPES [pH 7.4], 0.1 g BSA). The cells were then incubated with 1 ml of KRB at 37 °C for 60 min. At the end of the incubation, cells were washed with KRB medium. The cells were then incubated with KRB containing different glucose concentrations (4–16.7 mM for INS-1 and 4–20 mM for MIN6) for 1 h at 37 °C. The supernatant was collected and insulin was measured by ELISA using mouse insulin as a standard (Alpco Diagnostics, Windham, New Hampshire, United States). Cells were then incubated O/N with acidified EtOH (75% EtOH, 1.5% HCl). Protein determination was performed using the Bio-Rad Dc Protein assay (Hercules, California, United States). The amounts of secreted insulin were corrected by the amounts of cell protein in each well.
Immunofluorescence
Pancreas was isolated from wild-type or Sirt1 KO mice and fixed in 4% paraformaldehyde. Consecutive 8-μm sections were immunostained with anti-Sirt1 antibody at 1:100 dilution (Upstate-Chemicon, Charlottesville, Virginia, United States). DAPI was used to stain the nuclei. Hematoxylin and eosin was used to visualize islets.
In vivo insulin measurement
Mice were subjected to O/N fast followed by intraperitoneal glucose injection (1 g/kg body weight). Blood samples were collected from the tail vein the night before the experiment (ad libitum), right before the glucose injection (time 0), and after the different time points indicated in Figure 2E. Insulin levels were measured using the Ultrasensitive Mouse Insulin EIA (Alpco Diagnostics) according to manufacturer's specifications.
Glucose tolerance test
Mice were fasted O/N (16 h) and injected intraperitoneally with a saline glucose solution at 1g/kg body weight. Plasma glucose levels were measured from tail blood before and 2, 5, 10, 25, 40, 80, and 120 min after the glucose injection. The test was repeated five times with results comparable to those in Figure 2D.
Western blot analysis
Cells were lysed in RIPA buffer (1% TritonX-100, 158 mM NaCl, 5 mM EDTA, 10 mM Tris [pH 7.0], protease inhibitors [Boehringer, Mannheim, Germany], 1 mM DTT, and 0.1% phenylmethylsulfonyl fluoride), sonicated, and centrifuged at 14,000 rpm for 10 min. Proteins were separated by SDS-PAGE, transferred to nitrocellulose membrane (Schleicher and Schuell Bioscience, Dassel, Germany), and were probed with a polyclonal antibody against Sirt1 or monoclonal anti-actin antibody (Chemicon, Temecula, California, United States). The UCP2 antibody (Calbiochem, San Diego, California, United States) was used in 1:1000 dilution. The cEBP/β, the insulin receptor alpha and beta, and the kir6.2 antibodies (Santa Cruz Biotechnology, Santa Cruz, California, United States) were used at 1:200, 1:200, 1:500, and 1:200, respectively.
Chromatin immunoprecipitation
Assays were performed on cells as previously described according to the Farnham protocol (http://mcardle.oncology.wisc.edu). Immunoprecipitations were performed with 1μg of Sirt1 antibody. PCR primers were designed to span a regulatory region in the UCP2 promoter (5′-
CGTCTGTTCAAAGCGTCTCA; 5′-
CCAGCTGGAGTCTTCTCCTT) (set 5) or a control region in the UCP2 promoter (5′-
AGGTTGTTTCTGGGCCATGTGCTCTAA; 5′-
TAGACCCTGGCCACCCTGAGCGCGAAAT) (set 4).
Islets isolation and morphometric analysis
Islets were isolated using a collagenase technique as described previously [59]. The pancreas was dissected and islets separated by incubation with collagenase at 37 °C. Islets were then hand-picked using a microscope. Islets were used for Western blot analyses or for glucose-induced insulin secretion assays. To measure insulin secretion, islets were incubated in KRB with 4 mM glucose for 1 h and then incubated with KRB containing different glucose concentrations (4–20 mM) for 1 h at 37 °C. The experiment was carried out twice with four mice per treatment, and the repeat experiment gave data comparable to Figure 2E.
Four to five pancreases each from WT, Sir2+/−, and Sir2−/− mice were fixed in 2% chilled paraformaldehyde, and paraffin embedded, and three sections (5 μm) separated by at least 100 μm were dewaxed using xylene, rehydrated through serial dilutions of ethyl-alcohol, and subjected to antigen retrieval using 10 mM citrate (pH 6.1), followed by DAKO high pH antigen retrieval solution (Dako, Glostrup, Denmark). The sections were washed and stained with the respective antibodies in staining buffer with 100 mM NaCl, 3% BSA, 0.5% Triton-X-100, and 50 mM Na-PO4 (pH 7.4). Primary antibodies included sheep anti-insulin (The Binding Site, Birmingham, United Kingdom), mouse anti-glucagon (Sigma, St. Louis, Missouri, United States), and rabbit anti-somatostatin (Dako). The entire pancreatic section of each stained slide was arbitrarily divided into non-overlapping contiguous portions (using a microscope inset measuring guide) and imaged at 10× magnification. The islet area (in square micrometers) and the total area of each section were determined by tracing the outline to assess the area of the islets, using image analyzer software (Image-Pro 4.1 Plus). The percentage of islet cell area in the pancreas was expressed as a percentage of the total pancreatic area. For 4× images, islets were stained using a cocktail of antibodies against insulin, glucagon, and somatostatin and HRP-linked secondary antibodies with DAB as a chromogen.
ATP/ADP measurements
1 × 106 cells were washed with 1× PBS. Ice-cold 6% (v/v) HClO4 was added, and immediately the cells were scraped from the plate. The solution was neutralized with 1 M KHCO3, centrifuged briefly, and the supernatant was passed through a 0.2-m filter (Nanosep, Lund, Sweden), and subjected to reversed phase chromatography using a Targa C18 250 × 4.6 mm 5-mm column as described in [61]. Nucleotides were detected at 260 nm with a Waters 486 tunable detector. Peak heights were measured. Nucleotide identities were confirmed by co-migration with known standards.
RNA analyses
Northern blot—Total RNA from INS-1 cells was extracted using Triazol reagent (Invitrogen). For Northern blot analysis, 10 μg of RNA samples was separated on a formaldehyde gel, transferred to nylon membrane, and then hybridized with gene-specific probes. Normalization of mRNA was done by using actin as probe (kind gift from K. Olson). For whole-pancreas RNA extraction, fresh pancreas was stored in the RNA later solution (Ambion, Austin, Texas, United States) before extracting total RNA with the Qiagen system (Qiagen, Valencia, California, United States). The UCP2 probe was a kind gift from D. Ricquier, CNRS, Paris, France. RT-PCR/cDNA was synthesized from 1 μg of total RNA from the pancreas. The reverse transcriptase reaction was performed at 25 °C for 10 min, 42 °C for 1 h, 95 °C for 5 min, for one cycle. The PCR reaction (95 °C for 30 s, 58 °C for 30 s, 72 °C for 30 s, for 25 cycles) was performed using the respective sense and antisense oligonucleotide primers: UCP2: 5′-
GCCCGGGCTGGTGGTGGTC and 5′-
CCCCGAAGGCAGAAGTGAAGTGG; β-actin: 5′-
GAACCCTAAGGCCAACCGTGAAAAGAT and 5′-
ACCGCTCGTTGCCAATAGTGATG.
CAT assay
Transfections for CAT assays were done as described [60]. 293T cells were transiently transfected with pSU2N2-mUCP2-CAT (which contains 7 kb of the mouse UCP2 promoter; gift from D. Ricquier), pCMV-betaGal (to correct for transfection efficiency), and either pCMV-mSirt1 and pSPORT6-mPPAR-gamma2 (gift from B. Spiegelman, Harvard Medical School, Massachusetts, United States of America) or their respective empty vectors. 24 h later, total CAT enzyme levels were measured by ELISA (Roche, Basel, Switzerland).
NAD and NADH determination
NAD and NADH nucleotides were measured as described [15]. About 10 mg of frozen pancreas tissues was homogenized in 300 μl of acid extraction buffer to obtain NAD concentration, or alkali buffer to obtain NADH concentration. 240 μl of supernatant was neutralized with 120 μl of buffer. The concentration of nucleotides was measured fluorimetrically after an enzymatic cycling reaction using 2 μl of sample. All values were detected within the linear range and are expressed as nmol per gram of tissue.
Quantitation of NADH autofluorescence and glucose uptake
1.5 × 106 INS-1 cells were seeded in six-well plates. Fresh medium containing 4.0 mM glucose was added the day before the experiment. The medium was removed on the day of the experiment, and the cells washed 3× with warm KRB buffer. The cells were then incubated with 1 ml of KRB at 37 °C for 1 h. At the end of the incubation, the media was removed and the cells were incubated for 10 min with KRB containing different glucose concentrations: 4–16.7 mM (for NADH determination), or 200 μM 2-NBDG (Molecular Probes) for glucose uptake determination. At the end of the incubation, cells were washed 3× with cold 1× PBS. We then used a four-laser LSR II digital flow cytometer (Becton-Dickinson, Mountain View, California, United States) with Diva software. The 488-mm laser excited the fluorescent glucose, and emission was detected with a 530/30BP filter and expressed as arbitrary units. NADH was measured according to Thorell [46] with a 355-mm laser and detected with a 440/40BP filter. Cell viability was assessed using PI. Data were analyzed with FlowJo (Tree Star, Ashland, Oregon, United States).
Supporting Information
Figure S1 Growth Curve Plot of Cell Number vs. Time of INS-1 Cells Transfected with pSUPER (black line) or SiRNA Sirt1 (red line)
(163 KB TIF).
Click here for additional data file.
We thank B. Lowell, D. Chen, M. Haigis, X. Li, J. E. van Veen, and P. DiStefano for comments on the manuscript. We would like to thank E. Horrigan and J. Goslin for technical assistance. We would also like to thank Xianshu Huang, who did the embedding and the sectioning of the tissues. We thank Brian Lavan and Francine M. Gregoire for helpful advice on glucose tolerance tests. SJL is an Ellison Medical Foundation scholar. This work was supported by grants from the National Institute on Aging to SJL, and the National Institutes of Health to LG.
Competing interests. LG is a founder, consultant, and board member for Elixir Pharmaceuticals.
Author contributions. LB and LG conceived and designed the experiments. LB, MCM, FP, USJ, AS, and EJE performed the experiments. LB, USJ, and LG analyzed the data. LB, JA, TM, ML, MM, and SJL contributed reagents/materials/analysis tools. LB and LG wrote the paper.
Citation: Bordone L, Motta MC, Picard F, Robinson A, Jhala US, et al. (2006) Sirt1 regulates insulin secretion by repressing UCP2 in pancreatic β cells. PLoS Biol 4(2): e31.
Abbreviations
CATchloramphenicol acetyl-transferase
CRcalorie restriction
KOknockout
NADa derivative of niacin
NADHthe reduced form of NAD
O/Novernight
Sir2silent mating type information regulation 2
Sirt1Homolog of the yeast silencing information regulator2
UCPuncoupling protein
WATwhite adipose tissue
==== Refs
References
Saltiel AR Kahn CR Insulin signalling and the regulation of glucose and lipid metabolism Nature 2001 414 799 806 11742412
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
Anderson RM Bitterman KJ Wood JG Medvedik O Cohen H Manipulation of a nuclear NAD+ salvage pathway delays aging without altering steady-state NAD+ levels J Biol Chem 2002 277 18881 18890 11884393
Weindruch R Walford RL The retardation of aging and disease by dietary restriction 1988 Springfield (Illinois) Charles C Thomas 436
Barrows CH Kokkonen GC Moment GB Dietary restriction and life extension, biological mechanisms Nutritional approaches to aging research 1982 Boca Raton (Florida) CRC Press 219 243
Bluher M Kahn BB Kahn CR Extended longevity in mice lacking the insulin receptor in adipose tissue Science 2003 299 572 574 12543978
Chiu CH Lin WD Huang SY Lee YH Effect of a C/EBP gene replacement on mitochondrial biogenesis in fat cells Genes Dev 2004 18 1970 1975 15289464
Kaeberlein M McVey M Guarente L The SIR2/3/4 complex and SIR2 alone promote longevity in Saccharomyces cerevisiae by two different mechanisms Genes Dev 1999 13 2570 2580 10521401
Tissenbaum HA Guarente L Increased dosage of a sir-2 gene extends lifespan in Caenorhabditis elegans
Nature 2001 410 227 230 11242085
Lin SJ Defossez PA Guarente L Requirement of NAD and SIR2 for life-span extension by calorie restriction in Saccharomyces cerevisiae
Science 2000 289 2126 2128 11000115
Lin SJ Kaeberlein M Andalis AA Sturtz LA Defossez PA Calorie restriction extends Saccharomyces cerevisiae lifespan by increasing respiration Nature 2002 418 344 348 12124627
Imai S Armstrong CM Kaeberlein M Guarente L Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase Nature 2000 403 795 800 10693811
Landry J Sutton A Tafrov ST Heller RC Stebbins J The silencing protein SIR2 and its homologs are NAD-dependent protein deacetylases Proc Natl Acad Sci U S A 2000 97 5807 5811 10811920
Smith JS Brachmann CB Celic I Kenna MA Muhammad S A phylogenetically conserved NAD+-dependent protein deacetylase activity in the Sir2 protein family Proc Natl Acad Sci U S A 2000 97 6658 6663 10841563
Lin SJ Ford E Haigis M Liszt G Guarente L Calorie restriction extends yeast life span by lowering the level of NADH Genes Dev 2004 18 12 16 14724176
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
Picard F Kurtev M Chung N Topark-Ngarm A Senawong T Sirt1 promotes fat mobilization in white adipocytes by repressing PPAR-gamma Nature 2004 429 771 776 15175761
Motta MC Divecha N Lemieux M Kamel C Chen D Mammalian SIRT1 represses forkhead transcription factors Cell 2004 116 551 563 14980222
Brunet A Sweeney LB Sturgill JF Chua KF Greer PL Stress-dependent regulation of FOXO transcription factors by the SIRT1 deacetylase Science 2004 303 2011 2015 14976264
Rodgers JT Lerin C Haas W Gygi SP Spiegelman BM Nutrient control of glucose homeostasis through a complex of PGC-1alpha and SIRT1 Nature 2005 434 113 118 15744310
Vaziri H Dessain SK Ng Eaton E Imai SI Frye RA hSIR2(SIRT1) functions as an NAD-dependent p53 deacetylase Cell 2001 107 149 159 11672523
Luo J Nikolaev AY Imai S Chen D Su F Negative control of p53 by Sir2alpha promotes cell survival under stress Cell 2001 107 137 148 11672522
Cohen HY Miller C Bitterman KJ Wall NR Hekking B Calorie restriction promotes mammalian cell survival by inducing the SIRT1 deacetylase Science 2004 305 390 392 15205477
van der Horst A Tertoolen LG de Vries-Smits LM Frye RA Medema RH FOXO4 is acetylated upon peroxide stress and deacetylated by the longevity protein hSir2(SIRT1) J Biol Chem 2004 279 28873 28879 15126506
Jacobsson A Stadler U Glotzer MA Kozak LP Mitochondrial uncoupling protein from mouse brown fat. Molecular cloning, genetic mapping, and mRNA expression J Biol Chem 1985 260 16250 16254 2999153
Bouillaud F Ricquier D Thibault J Weissenbach J Molecular approach to thermogenesis in brown adipose tissue: cDNA cloning of the mitochondrial uncoupling protein Proc Natl Acad Sci U S A 1985 82 445 448 3855564
Ricquier D Bouillaud F The uncoupling protein homologues: UCP1, UCP2, UCP3, StUCP and AtUCP Biochem J 2000 345 Pt 2 161 179 10620491
Ricquier D Bouillaud F Mitochondrial uncoupling proteins: From mitochondria to the regulation of energy balance J Physiol 2000 529 Pt 1 3 10 11080246
Echtay KS Winkler E Frischmuth K Klingenberg M Uncoupling proteins 2 and 3 are highly active H(+) transporters and highly nucleotide sensitive when activated by coenzyme Q (ubiquinone) Proc Natl Acad Sci U S A 2001 98 1416 1421 11171965
Fleury C Neverova M Collins S Raimbault S Champigny O Uncoupling protein-2: A novel gene linked to obesity and hyperinsulinemia Nat Genet 1997 15 269 272 9054939
Jaburek M Varecha M Gimeno RE Dembski M Jezek P Transport function and regulation of mitochondrial uncoupling proteins 2 and 3 J Biol Chem 1999 274 26003 26007 10473545
Rial E Gonzalez-Barroso M Fleury C Iturrizaga S Sanchis D Retinoids activate proton transport by the uncoupling proteins UCP1 and UCP2 Embo J 1999 18 5827 5833 10545094
Dulloo AG Samec S Uncoupling proteins: Their roles in adaptive thermogenesis and substrate metabolism reconsidered Br J Nutr 2001 86 123 139 11502224
Arsenijevic D Onuma H Pecqueur C Raimbault S Manning BS Disruption of the uncoupling protein-2 gene in mice reveals a role in immunity and reactive oxygen species production Nat Genet 2000 26 435 439 11101840
Zhang CY Baffy G Perret P Krauss S Peroni O Uncoupling protein-2 negatively regulates insulin secretion and is a major link between obesity, beta cell dysfunction, and type 2 diabetes Cell 2001 105 745 755 11440717
Chan CB De Leo D Joseph JW McQuaid TS Ha XF Increased uncoupling protein-2 levels in beta-cells are associated with impaired glucose-stimulated insulin secretion: Mechanism of action Diabetes 2001 50 1302 1310 11375330
Chan CB MacDonald PE Saleh MC Johns DC Marban E Overexpression of uncoupling protein 2 inhibits glucose-stimulated insulin secretion from rat islets Diabetes 1999 48 1482 1486 10389858
McBurney MW Yang X Jardine K Hixon M Boekelheide K The mammalian SIR2alpha protein has a role in embryogenesis and gametogenesis Mol Cell Biol 2003 23 38 54 12482959
Cheng HL Mostoslavsky R Saito S Manis JP Gu Y Developmental defects and p53 hyperacetylation in Sir2 homolog (SIRT1)-deficient mice Proc Natl Acad Sci U S A 2003 100 10794 10799 12960381
Bitterman KJ Anderson RM Cohen HY Latorre-Esteves M Sinclair DA Inhibition of silencing and accelerated aging by nicotinamide, a putative negative regulator of yeast sir2 and human SIRT1 J Biol Chem 2002 277 45099 45107 12297502
Yamada K Nakata M Horimoto N Saito M Matsuoka H Measurement of glucose uptake and intracellular calcium concentration in single, living pancreatic beta-cells J Biol Chem 2000 275 22278 22283 10748091
Zou C Wang Y Shen Z 2-NBDG as a fluorescent indicator for direct glucose uptake measurement J Biochem Biophys Methods 2005 64 207 215 16182371
Kulkarni RN Bruning JC Winnay JN Postic C Magnuson MA Tissue-specific knockout of the insulin receptor in pancreatic beta cells creates an insulin secretory defect similar to that in type 2 diabetes Cell 1999 96 329 339 10025399
Mills EM Xu D Fergusson MM Combs CA Xu Y Regulation of cellular oncosis by uncoupling protein 2 J Biol Chem 2002 277 27385 27392 12011039
Thorell B Flow-cytometric monitoring of intracellular flavins simultaneously with NAD(P)H levels Cytometry 1983 4 61 65 6617395
Medvedev AV Snedden SK Raimbault S Ricquier D Collins S Transcriptional regulation of the mouse uncoupling protein-2 gene. Double E-box motif is required for peroxisome proliferator-activated receptor-gamma-dependent activation J Biol Chem 2001 276 10817 10823 11150307
Rousset S Alves-Guerra MC Mozo J Miroux B Cassard-Doulcier AM The biology of mitochondrial uncoupling proteins Diabetes 2004 53 Suppl 1 S130 S135 14749278
Lamming DW Wood JG Sinclair DA Small molecules that regulate lifespan: Evidence for xenohormesis Mol Microbiol 2004 53 1003 1009 15306006
Fulco M Schiltz RL Iezzi S King MT Zhao P Sir2 regulates skeletal muscle differentiation as a potential sensor of the redox state Mol Cell 2003 12 51 62 12887892
Speakman JR Talbot DA Selman C Snart S McLaren JS Uncoupled and surviving: Individual mice with high metabolism have greater mitochondrial uncoupling and live longer Aging Cell 2004 3 87 95 15153176
Fridell YW Sanchez-Blanco A Silvia BA Helfand SL Targeted expression of the human uncoupling protein 2 (hUCP2) to adult neurons extends life span in the fly Cell Metab 2005 1 145 152 16054055
Vidal-Puig AJ Grujic D Zhang CY Hagen T Boss O Energy metabolism in uncoupling protein 3 gene knockout mice J Biol Chem 2000 275 16258 16266 10748196
Gong DW Monemdjou S Gavrilova O Leon LR Marcus-Samuels B Lack of obesity and normal response to fasting and thyroid hormone in mice lacking uncoupling protein-3 J Biol Chem 2000 275 16251 16257 10748195
Bordone L Guarente L Calorie restriction, SIRT1 and metabolism: Understanding longevity Nat Rev Mol Cell Biol 2005 6 298 305 15768047
Pecqueur C Alves-Guerra MC Gelly C Levi-Meyrueis C Couplan E Uncoupling protein 2, in vivo distribution, induction upon oxidative stress, and evidence for translational regulation J Biol Chem 2001 276 8705 8712 11098051
Xiao H Massaro D Massaro GD Clerch LB Expression of lung uncoupling protein-2 mRNA is modulated developmentally and by caloric intake Exp Biol Med (Maywood) 2004 229 479 485 15169966
Boss O Samec S Paoloni-Giacobino A Rossier C Dulloo A Uncoupling protein-3: A new member of the mitochondrial carrier family with tissue-specific expression FEBS Lett 1997 408 39 42 9180264
Picard F Wanatabe M Schoonjans K Lydon J O'Malley BW Progesterone receptor knockout mice have an improved glucose homeostasis secondary to beta-cell proliferation Proc Natl Acad Sci U S A 2002 99 15644 15648 12438645
Picard F Gehin M Annicotte J Rocchi S Champy MF SRC-1 and TIF2 control energy balance between white and brown adipose tissues Cell 2002 111 931 941 12507421
Stocchi V Cucchiarini L Magnani M Palma P Crescentini G Simultaneous extraction and reverse phase high performance liquid chromatographic determination of adenine and pyridine nucleotides in human red blood cells Anal Biochem 1985 146 118 124 3993925
|
16366736
|
PMC1318478
|
CC BY
|
2021-01-05 08:28:16
|
no
|
PLoS Biol. 2006 Feb 27; 4(2):e31
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0040031
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 2007652810.1371/journal.pbio.0040042SynopsisNeuroscienceMus (Mouse)A New Window into Structural Plasticity in the Adult Visual Cortex SynopsisGross Liza 2 2006 27 12 2005 27 12 2005 4 2 e42Copyright: © 2006 Public Library of Science.2006This 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.
Dynamic Remodelling of Dendritic Arbors in GABAergic Interneurons of Adult Visual Cortex
==== Body
The developing human brain is a hotbed of activity that continues well beyond the first year. During early postnatal development, we manufacture some 250,000 neurons per minute, then spend the next few years building the connections that underlie brain function. It has long been assumed that the neural plasticity of youth eventually settles down by adulthood. Though experimentally induced lesions in the adult cat and monkey cortex can produce anatomical changes, these findings are based on inferences from statistical evidence across different populations rather than on direct observation. And while neuroscientists have known for decades that the adult brain can reorganize neural pathways in response to new experiences—by changing the firing pattern and responses of neurons, for example—it has remained an open question whether structural changes accompany this functional plasticity.
In a new study, Wei-Chung Allen Lee and Elly Nedivi, along with Hayden Huang and Peter So, and their colleagues, take advantage of recent advances in imaging technology and single-cell genetic labeling techniques to investigate this question in mice. Continuous observations of the mouse adult visual cortex over the course of a few months revealed that the adult brain can indeed rewire its circuits under normal conditions. These rearrangements appear to follow neuron-specific rules, with one type of neuron undergoing a range of structural modifications while another maintains its original architecture.
One of the dendritic branch tips of this inhibitory neuron grew out of the microscope's imaging area in just two weeks
Many studies have focused on pyramidal neurons—excitatory neurons that promote neuron firing—but few have focused on the possible structural dynamics of a range of different neuron types. In this study, Lee et al. focused on a cross-section of neurons, imaging every neuron they saw. The only neurons they saw growing were the inhibitory, nonpyramidal neurons, which inhibit the activity of cortical neurons and lack the classic pyramid structure that so easily identifies their excitatory brethren. Since these neurons can help adjust the brain's internal maps by inhibiting signaling in response to new stimuli or learning, the authors wondered if they could be involved in structural changes as well.
The authors focused on the surface layers of the neocortex. (The neocortex consists of six cell layers, with layer 1 closest to the cortical surface; the authors focused on layers 2 and 3.) To allow direct observation of the area, they implanted a glass window over the two areas of the visual cortex in four- to six-week-old mice. These mice express fluorescent protein in neocortical neurons, allowing Lee et al. to track the location and morphology of these neurons using two-photon microscopy. Time-lapse images of six pyramidal neurons and eight nonpyramidal neurons in 13 mice were taken over the course of four to ten weeks. The length of dendritic branch tips were measured over time to evaluate physical changes in the neurons.
The pyramidal neurons showed no structural changes in individual branch tips, but the nonpyramidal neurons showed dynamic changes, with one branch tip undergoing dramatic remodeling. “Within as little as two weeks,” the authors note, “this branch tip more than doubled its length and exited the imaging volume.” One nonpyramidal neuron even showed a few new branch tip additions.
All of the nonpyramidal neurons showed at least one and up to seven dynamic branch tips, with an average of about 14% showing structural modifications. (The authors monitored up to 50 branch tips from a single neuron.) This remodeling occurred both incrementally and in short bursts, and involved both branch tip growth and shrinking. Lee et al. confirmed that these nonpyramidal neurons were in fact inhibitory interneurons by showing that they expressed gamma-aminobutyric acid (GABA)—a neurotransmitter that inhibits neuron firing—while the pyramidal neurons did not.
Since the laws of probability suggest that given the changes observed in the nonpyramidal neurons, at least one pyramidal branch tip of 124 monitored should change if all things are equal, the authors argue that the pyramidal and nonpyramidal neurons have different dynamic properties. The branch tips of nonpyramidal cells in the adult neocortex can grow, retract, and sprout new additions—without experimental manipulations. Many studies support the idea of a relatively stable adult neocortical structure, but as Lee et al. point out, they focused on pyramidal neurons, while this study focused on nonpyramidal neurons. Both may be right. Under normal conditions at least, pyramidal structural modifications are far less obvious than those seen in the nonpyramidal neurons.
It remains to be seen whether the structural plasticity seen here underlies observed functional reorganizations. Probing this question will depend on determining what kinds of structural changes might be expected, figuring out how to detect them, and then interpreting the changes. Studying the responses of axonal arbors connected to the nonpyramidal dendrites, for example, may prove instructive. Based on these results, direct observation of specific neurons in a local pathway should yield promising results.
|
20076528
|
PMC1318479
|
CC BY
|
2021-01-05 08:21:47
|
no
|
PLoS Biol. 2006 Feb 27; 4(2):e42
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0040042
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 2007653010.1371/journal.pbio.0040044SynopsisCell BiologyPhysiologyDiabetes/Endocrinology/MetabolismMus (Mouse)MammalsVertebratesAnimalsEukaryotesThe
Sirt1 Gene Promotes Insulin Secretion in Accord with Diet
SynopsisChanut Françoise 2 2006 27 12 2005 27 12 2005 4 2 e44Copyright: © 2006 Public Library of Science.2006This 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.
Sirt1 Regulates Insulin Secretion by Repressing UCP2 in Pancreatic β Cells
==== Body
Those seeking the fountain of youth would do well to watch their calories. Indeed, caloric restriction has been known for many years to increase longevity in laboratory animals. A simple explanation for this effect is that reducing food intake slows down the body's metabolism, which reduces the formation of toxic byproducts that cause tissue damage and aging.
Whatever the mechanisms, the benefits of caloric restriction hold from furry mammals to single-celled fungi. In yeast, nematodes, and fruit flies, caloric restriction increases the activity of the
Sir2 gene, which in turn changes the expression of genes related to metabolism. One of the key regulators of mammalian metabolism is the pancreatic hormone insulin. In a new study, Laura Bordone, Leonard Guarente, and their colleagues show that the mammalian homolog of
Sir2, called
Sirt1, modulates insulin production in response to diet.
In times of famine, the body taps into its own resources to provide energy for its working tissues. For instance, it mobilizes the lipid molecules stored in fat, and coaxes the liver into producing the simple sugar glucose. Cells take up glucose and lipids from the blood, and extract their chemical energy. In times of plenty, glucose and lipids come from food. As their levels rise in the blood, the pancreas secretes insulin, which stimulates the uptake of glucose by muscles and lipids by fat. An important function of insulin is to regulate glucose levels in the blood; its secretion is therefore tightly controlled by glucose concentration. But during fasting—and starvation—insulin secretion dips to very low levels, an adaptation that increases glucose availability for the brain.
Bordone et al. asked whether Sirt1 influenced insulin production. They disrupted the
Sirt1 gene of mice, and found that these mice produced very little insulin, regardless of whether they were well fed or starved. These results suggested that Sirt1 is necessary for glucose to induce insulin production.
The authors next asked at what step of insulin production Sirt1 acts. Insulin is made by specialized cells of the pancreas, called ß cells. ß cells can only secrete insulin when they accumulate enough ATP. This happens when glucose levels rise in the blood, after a meal for instance, because ß cells metabolize glucose into ATP. Bordone and her colleagues found that ß cells with an inactive
Sirt1 gene did not secrete as much insulin in response to glucose as normal ß cells. Nor did they convert glucose into ATP as efficiently as normal ß cells. This last observation led the authors to examine the activity of a type of protein known as uncoupling protein (UCP), which diverts glucose breakdown from ATP synthesis. In ß cells, the UCP2 protein is known to inhibit insulin secretion by routing glucose metabolism toward a molecule called NADH, rather than toward ATP.
The authors demonstrate that Sirt1 inhibits the production of UCP2 by directly preventing the expression of the
UCP2 gene. How does the interaction between Sirt1 and UCP2 relate to caloric restriction? The authors find that in starved mice, UCP2 levels increase in ß cells. This suggests that caloric restriction induces a decrease in Sirt1 activity in mice. This result is somewhat surprising since in yeast and other organisms, caloric restriction increases Sir2 expression.
Because Sirt1 and insulin have many roles in mammals, it is at present unclear how they mediate the effect of diet on lifespan. An intriguing hypothesis stems from the fact that UCP2 dampens the formation of toxic metabolic byproducts that precipitate aging. If the relationship between Sirt1 and UCP2 holds in more tissues than just ß cells, Sirt1 may open a simple path to a longer life.
|
20076530
|
PMC1318480
|
CC BY
|
2021-01-05 08:21:48
|
no
|
PLoS Biol. 2006 Feb 27; 4(2):e44
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0040044
|
oa_comm
|
==== Front
J Negat Results BiomedJournal of Negative Results in Biomedicine1477-5751BioMed Central London 1477-5751-4-91633665410.1186/1477-5751-4-9ResearchBeta-defensin genomic copy number is not a modifier locus for cystic fibrosis Hollox Edward J [email protected] Jane [email protected] Uta [email protected] Juliana [email protected] Eric WFW [email protected] John AL [email protected] Institute of Genetics, University of Nottingham, Nottingham, UK2 Department of Gene Therapy, Faculty of Medicine at the National Heart and Lung Institute, Imperial College, London, UK3 Paediatric Respiratory Disease, Royal Brompton Hospital, London, UK4 Adult Cystic Fibrosis, Royal Brompton Hospital, London, UK2005 7 12 2005 4 9 9 8 3 2005 7 12 2005 Copyright © 2005 Hollox et al; licensee BioMed Central Ltd.2005Hollox 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.
Human beta-defensin 2 (DEFB4, also known as DEFB2 or hBD-2) is a salt-sensitive antimicrobial protein that is expressed in lung epithelia. Previous work has shown that it is encoded in a cluster of beta-defensin genes at 8p23.1, which varies in copy number between 2 and 12 in different individuals. We determined the copy number of this locus in 355 patients with cystic fibrosis (CF), and tested for correlation between beta-defensin cluster genomic copy number and lung disease associated with CF. No significant association was found.
==== Body
Background
Cystic fibrosis (CF) is an autosomal recessive genetic disease in epithelia (Online Mendelian Inheritance in Man (OMIM) 219700, ), with chronic lung infection being the major cause of morbidity and mortality. It has a a carrier frequency of about 1 in 25 in the United Kingdom, and the disease is due to mutations in the cystic fibrosis transmembrane conductance regulator gene (CFTR) gene, which encodes a membrane chloride channel (OMIM 602421). There is a wide spectrum of different mutations known to cause CF, but the most common is the ΔF508 mutation, which exists in the European population as a polymorphic allele with a frequency of between 2 and 3%.
The course of lung infection can vary greatly between patients with the same CFTR mutation, and despite the wide spectrum of disease causing mutations, only a few correlate with lung disease severity. The search for genetic loci modifying lung disease in CF patients has produced several candidates, with the gene for mannose binding lectin (MBL) being the most notable. MBL is a part of the innate immune system, and low MBL in serum has been associated with infections in both adults and children [1,2]. Several studies have reported links with severity of CF lung disease, although the conclusion regarding the impact of the heterozygous deficient state differ amongst these [3-5].
Like MBL, beta-defensins are an important part of innate immunity. The peptide human beta-defensin 2 (encoded by the gene DEFB4, also known as DEFB2 or hBD-2) has been identified as a salt-sensitive antibacterial agent expressed in the lung airway [6]. Given the chronic bacterial colonisation of the lungs, this is a potential candidate for a CF modifier locus. It is possible that variation in DEFB4 levels could cause changes in the antibacterial effectiveness of airway surface fluid. Deterioration in lung function during clinical progress of CF is correlated with DEFB4 expression levels, suggesting again that DEFB4 has an important antibacterial role in airway surface fluid [7]. Additionally, DEFB4 expression is up-regulated by mucoid Pseudomonas aeruginosa, and has very effective bactericidal activity against it [8,9], but another important pathogen in CF lungs, Burkholderia cepacia, is not killed by airway surface fluid expressing DEFB4 [10].
Three beta-defensins are expressed in significant levels in airway epithelia: human beta-defensin 1 (DEFB1) which is expressed in the lung [11], human beta-defensin 3 (DEFB103, also known as DEFB3 or hBD-3) which is expressed in the trachea [12], and DEFB4, also expressed in the lung [6]. All three genes are part of a beta-defensin cluster at chromosomal locus 8p23.1 [13]. This cluster of beta defensins, with the exception of DEFB1, has been shown to be on a repeat unit of at least 260 kb which varies in copy number in different people. Individuals can have between 2 and 12 copies, with most individuals having 3, 4 or 5 copies, and copy number is correlated with basal unstimulated expression level of DEFB4 in lymphoblastoid cell lines [14]. There have been several genome-wide studies confirming this copy number variation [15-17], and they highlight this type of large scale genomic polymorphism as an underappreciated source of phenotypic variation. Indeed, there is evidence that variable CCL3L1 gene copy number is associated with susceptibility to HIV infection [18].
The hypothesis that we wished to test in this study was that copy number variation of this beta-defensin cluster (and, by implication, DEFB4) is associated with lung disease in CF patients.
Results
We determined beta-defensin cluster copy number for 355 CF patients, of whom 343 had full clinical data. 159 patients were ΔF508 CFTR homozygotes, 141 were ΔF508 CFTR heterozygotes and 43 had non-ΔF508 mutations on both chromosomes. The mean age was 29.1 ± 8.9 years and 57% were male. Figure 1 shows the distribution of beta-defensin copy number phenotypes in the full CF patient cohort. There was no difference in the copy number phenotype distribution between this cohort and a random sample of 167 normal UK individuals (E.J.Hollox and J.A.L.Armour unpublished data).
Figure 1 Beta-defensin diploid copy number distribution in the cohort of 355 cystic fibrosis patients analysed.
Each analysis was performed with the whole cohort, and with the ΔF508 CFTR homozygotes only. There were no significant correlations between beta-defensin copy number and each respiratory clinical parameter measured (mean and current FEV1, mean and current FVC). The patients were then arbitrarily divided into two groups based on copy number: one group comprising patients with 3 and fewer copies of the β-defensin region (n = 77), and another group with patients who have 4 copies and more (n = 266). Differences in lung function and inflammatory marker status were tested using two-tailed Mann-Whitney U-tests, differences in infection status were tested using χ2 tests. With this sample size, a difference of 9% in FEV1 median between the two groups could be detected with 80% power. No significant differences between any two groups were seen.
We also decided to determine the effectiveness of microsatellite dosage ratios in reinforcing, or possibly acting as a replacement for, MAPH in determining copy number. We analysed a subset of samples with copy numbers of three or four by determining the genotype of a microsatellite in the intron of DEFB4. The allelic distributions are shown in Figure 2. Out of 114 samples tested, 81 gave informative results, of which 71 (88%) agreed with the determination by MAPH. The samples that differed typically gave unrounded MAPH scores between the two values (range 3.42–3.54 or 4.5–4.54), and may be due either to error rate or, perhaps more likely given the small confidence intervals of each MAPH result (see Methods) and its reliability for other loci [19-23], genuine copy number heterogeneity between probes across the repeats. We should be able to distinguish between genuine copy number heterogeneity and MAPH error rate by analysing the copy number estimate given by the DEFB4 probe alone, which is 735 bp from the microsatellite in the intron of DEFB4, and should report the same copy number as the microsatellite. Analysis of this copy number estimate for the 10 discrepant results shows that it agrees with the microsatellite estimate for 5 samples. Therefore, the error rate in distinguishing 3 from 4 copies by MAPH is 5/81 = 6%. The result also suggests that copy number heterogeneity involving the DEFB4 probe (probe F) has a frequency of around 6%. In addition, copy number heterogeneity involving MAPH probe H has been found in several well-characterised individual DNAs (E.J.Hollox, J.A.L.Armour and J.C.K.Barber, unpublished data).
Figure 2 Allelic distribution of DEFB4 microsatellite (EPEV-2) alleles in individuals with three and four copies of the beta-defensin gene cluster.
Discussion
This is the first study to test the hypothesis, raised in a previous paper [14], that variable β-defensin genomic copy number affects immune system function; in this case lung infection in a susceptible CF patient cohort. From the data we have presented, it is clear that either β-defensin genomic copy number has no effect on lung function in CF patients, or that any effect is too small for it to be detected using this cohort size. We cannot formally rule out an effect of single nucleotide polymorphisms (SNPs) within or surrounding the DEFB4 gene, given the apparent lack of association between copy number alleles and microsatellite alleles, and by inference any single nucleotide polymorphism (SNP). Any further study on these SNPs will have to carefully consider the copy number variation at this locus, given that a single nucleotide change could not only be allelic but differ between paralogues on the same chromosome [16].
There could be two reasons why β-defensin genomic copy number has no effect on lung function in CF patients. Firstly, unlike in cultured lymphocytes [14], there may be no correlation between DEFB4 copy number and gene expression in lung epithelia. Ideally a positive correlation between DEFB4 copy number and DEFB4 expression levels in non-inflamed lungs should be established, but realistically such studies are difficult to perform for both technical and ethical reasons. Even if such a direct link were proved, the lung in CF patients is characterised by a strong inflammatory response, and, given the dramatic increase in DEFB4 peptide levels in inflamed tissue [24], this could hide any differences in basal expression levels due to genomic copy number variation. Secondly, differences in DEFB4 expression are masked by other factors in the CF lung: for example, elastolytic cathepsins which breakdown DEFB4 and DEFB103 are present in the CF lung, but not in the healthy lung [25]. A recent report studying single nucleotide variation in the DEFB1 gene and lung function in CF patients has found no link, providing further evidence against beta-defensins modifying lung function in CF patients [26].
The key to further progress is to understand role of antimicrobial peptides in the inflamed lung, and exactly how they contribute to airway surface defence. A recent paper shows that DEFB4 is more than an antimicrobial peptide, and attracts neutrophils to sites of inflammation and infection [27], so it is possible that its chemoattractant role is as important in the lung as its antimicrobial role. How DEFB4 interacts with other components of the inflammatory response, and the innate and adaptive immune response, is only just beginning to be understood, and it is clear that much more work is needed before we understand these processes in lung disease.
Methods
Patient samples
DNA was extracted from peripheral blood from cystic fibrosis patients using the QIAamp DNA Blood Midi kit (Qiagen) according to the manufacturer's recommendations as part of a larger modifier gene study. Samples were coded and anonymised. All participants gave informed consent and the study was approved by the Royal Brompton Hospital and Harefield NHS Trust Ethics Committee.
Clinical parameters
Clinical data were obtained from the clinical databases, patients' hospital notes and computerised microbiology reports. Forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) were obtained from annual lung function laboratory records, had been corrected for sex, age and height and are expressed as percent of predicted normal values. Annual values were recorded where available from 1990–2002 and annual rate of decline was calculated for all subjects with 2 or more values after linear regression of all available data. Transcutaneous oxygen saturation with the patient breathing room air was obtained at the annual visit, and a record of the requirement for supplementary oxygen was made. Additional data included the presence of infection with P. aeruginosa or B. cepacia and most recent total white blood cell count, C-reactive protein and liver function tests.
Multiplex amplifiable probe hybridisation (MAPH)
MAPH was carried out essentially as previously described [28,29], with fluorescent detection of amplification products by an Applied Biosystems 3100 Genetic Analyzer. A probe set including six probes mapping to the beta-defensin cluster was constructed (Table 1). The values for these six probes were normalised first for lane intensity against control subtelomeric probes in the same probe set which do not show copy number variation [20]. These ratios were then calibrated against a weighted average of two control genomic samples in the same experiment, usually N8 (3 copies) and N25 (7 copies) to get a value for the number of copies of each probe per diploid cell [14]. In two instances, two other known controls were used on duplicate samples (European Collection of Animal Cell Cultures (ECACC) samples BO0183 3 copies – AF0105 4 copies). Of the 355 samples tested, 205 were tested in duplicate. Copy number was calculated by averaging the values of each probe mapping to this region and presenting the data both as a mean value +/- 95% confidence intervals (CIs), and as a rounded integer, representing the most likely copy number. Initially all six probes were used in the calculation, but it was found that DEFB104 probe was unusually sensitive to DNA quantity and removing the values for that probe in many cases improved the confidence intervals. All data in this paper are from five probes only. The 95% CIs for each copy number reading in the full cohort ranged from 0.05 to 1.2 with a median value of 0.24.
Table 1 MAPH probes that detect beta-defensin copy number variation
Probe name Hollox et al (Reference 14) probe Genomic location Accession number Location in clone
DEFB106 - DEFB106 exon 2 AF252830.5 6703–7022
DEFB104 - DEFB104 exon 2 AF252830.5 18832–19143
DEFB4B G DEFB4 exon 2 NM_004942 117–237
G13705 H SPAG11 intron 2 G13705 121–326
DEFB103 - DEFB103 exon 1 AF252831.2 97176–97312
DEFB105 - DEFB105 exon 3 AF202031.5 57470–57612
Microsatellite analysis
A microsatellite (EPEV-2) in the intron of DEFB4, was amplified using the following primers (EPEV-2F 5'-GCACCAGAGACCTCATGTTTTC-3' and EPEV-2R 5'-GTAACTTACAGTTGAAAACCAC-3'), with EPEV-2F 5' fluorescently-labelled with 6-FAM. The PCR conditions were as follows: 0.2 mM each of dATP, dCTP, dGTP, dTTP, 1 mM MgCl2, 75 mM Tris-HCl (pH 8.8 at 25°C), 20 mM (NH4)2SO4, 0.01% (v/v) Tween 20 in a final volume of 10 μl, with 5 pmol of each primer, 10 ng genomic DNA, 0.5 units Taq DNA polymerase (ABgene), followed by cycling for 25 cycles at 95°C 1 minute, 60°C 1 minute 72°C 1 minute, and a final extension incubation of 72°C 20 minutes. 1.5 μl of the PCR product was mixed with 10 μl Hi-Di formamide and run on an ABI 3100 using standard conditions, with G500 ROX marker. Allelic sizes ranged from 240 bp to 272 bp, and samples were omitted from analysis if stutter peaks made the interpretation of allele peak areas or allele lengths ambiguous.
Authors' contributions
EJH carried out the molecular laboratory work, and participated in the design of the study. JD and EJH carried out the statistical analysis, and JD together with JB participated in clinical data collection and method design. UG, EA, and JALA participated in the design of the study. All authors helped to draft the manuscript, and have read and approved the final manuscript.
Acknowledgements
We wish to thank Terence Low for DNA extraction, Ning Shen for help with clinical data collection, and Jess Tyson for helpful advice and comments. EJH is supported by a Wellcome Trust Bioarchaeology Postdoctoral Fellowship (grant no. 071024).
==== Refs
Summerfield JA Ryder S Sumiya M Thursz M Gorchein A Monteil MA Turner MW Mannose binding protein gene mutations associated with unusual and severe infections in adults Lancet 1995 345 886 889 7707811 10.1016/S0140-6736(95)90009-8
Summerfield JA Sumiya M Levin M Turner MW Association of mutations in mannose binding protein gene with childhood infection in consecutive hospital series Bmj 1997 314 1229 1232 9154025
Gabolde M Guilloud-Bataille M Feingold J Besmond C Association of variant alleles of mannose binding lectin with severity of pulmonary disease in cystic fibrosis: cohort study Bmj 1999 319 1166 1167 10541506
Garred P Pressler T Madsen HO Frederiksen B Svejgaard A Hoiby N Schwartz M Koch C Association of mannose-binding lectin gene heterogeneity with severity of lung disease and survival in cystic fibrosis J Clin Invest 1999 104 431 437 10449435
Davies JC Turner MW Klein N Impaired pulmonary status in cystic fibrosis adults with two mutated MBL-2 alleles Eur Respir J 2004 24 798 804 15516675 10.1183/09031936.04.00055404
Bals R Wang X Wu Z Freeman T Bafna V Zasloff M Wilson JM Human beta-defensin 2 is a salt-sensitive peptide antibiotic expressed in human lung J Clin Invest 1998 102 874 880 9727055
Chen CI Schaller-Bals S Paul KP Wahn U Bals R Beta-defensins and LL-37 in bronchoalveolar lavage fluid of patients with cystic fibrosis J Cyst Fibros 2004 3 45 50 15463886 10.1016/j.jcf.2003.12.008
Harder J Meyer-Hoffert U Teran LM Schwichtenberg L Bartels J Maune S Schroder JM Mucoid Pseudomonas aeruginosa, TNF-alpha, and IL-1beta, but not IL-6, induce human beta-defensin-2 in respiratory epithelia Am J Respir Cell Mol Biol 2000 22 714 721 10837369
Harder J Bartels J Christophers E Schroder JM A peptide antibiotic from human skin Nature 1997 387 861 9202117 10.1038/43088
Baird RM Brown H Smith AW Watson ML Burkholderia cepacia is resistant to the antimicrobial activity of airway epithelial cells Immunopharmacology 1999 44 267 272 10598883 10.1016/S0162-3109(99)00122-8
Goldman MJ Anderson GM Stolzenberg ED Kari UP Zasloff M Wilson JM Human beta-defensin-1 is a salt-sensitive antibiotic in lung that is inactivated in cystic fibrosis Cell 1997 88 553 560 9038346 10.1016/S0092-8674(00)81895-4
Jia HP Schutte BC Schudy A Linzmeier R Guthmiller JM Johnson GK Tack BF Mitros JP Rosenthal A Ganz T McCray PBJ Discovery of new human beta-defensins using a genomics-based approach Gene 2001 263 211 218 11223260 10.1016/S0378-1119(00)00569-2
Semple CA Rolfe M Dorin JR Duplication and selection in the evolution of primate beta-defensin genes Genome Biol 2003 4 R31 12734011 10.1186/gb-2003-4-5-r31
Hollox EJ Armour JA Barber JC Extensive normal copy number variation of a beta-defensin antimicrobial-gene cluster Am J Hum Genet 2003 73 591 600 12916016 10.1086/378157
Iafrate AJ Feuk L Rivera MN Listewnik ML Donahoe PK Qi Y Scherer SW Lee C Detection of large-scale variation in the human genome Nat Genet 2004 36 949 951 15286789 10.1038/ng1416
Fredman D White SJ Potter S Eichler EE Den Dunnen JT Brookes AJ Complex SNP-related sequence variation in segmental genome duplications Nat Genet 2004 36 861 866 15247918 10.1038/ng1401
Sebat J Lakshmi B Troge J Alexander J Young J Lundin P Maner S Massa H Walker M Chi M Navin N Lucito R Healy J Hicks J Ye K Reiner A Gilliam TC Trask B Patterson N Zetterberg A Wigler M Large-scale copy number polymorphism in the human genome Science 2004 305 525 528 15273396 10.1126/science.1098918
Gonzalez E Kulkarni H Bolivar H Mangano A Sanchez R Catano G Nibbs RJ Freedman BI Quinones MP Bamshad MJ Murthy KK Rovin BH Bradley W Clark RA Anderson SA O'Connell R J Agan BK Ahuja SS Bologna R Sen L Dolan MJ Ahuja SK The influence of CCL3L1 gene-containing segmental duplications on HIV-1/AIDS susceptibility Science 2005 307 1434 1440 15637236 10.1126/science.1101160
White S Kalf M Liu Q Villerius M Engelsma D Kriek M Vollebregt E Bakker B Van Ommen GJ Breuning MH Den Dunnen JT Comprehensive Detection of Genomic Duplications and Deletions in the DMD Gene, by Use of Multiplex Amplifiable Probe Hybridization Am J Hum Genet 2002 71 365 374 12111668 10.1086/341942
Hollox EJ Atia T Cross G Parkin T Armour JAL High-throughput screening of human subtelomeric DNA for copy number changes using Multiplex Amplifiable Probe Hybridisation (MAPH) J Med Genet 2002 39 790 795 12414816 10.1136/jmg.39.11.790
Hollox EJ Armour JA No evidence for DNA copy number change associated with the DUP25 cytogenetic phenotype Eur J Hum Genet 2003 11 911 912 14508506 10.1038/sj.ejhg.5201074
Reid AG Tarpey PS Nacheva EP High-resolution analysis of acquired genomic imbalances in bone marrow samples from chronic myeloid leukemia patients by use of multiple short DNA probes Genes Chromosomes Cancer 2003 37 282 290 12759926 10.1002/gcc.10215
Aldred PM Hollox EJ Armour JA Copy number polymorphism and expression level variation of the human {alpha}-defensin genes DEFA1 and DEFA3 Hum Mol Genet 2005 14 2045 2052 15944200 10.1093/hmg/ddi209
Singh PK Jia HP Wiles K Hesselberth J Liu L Conway BA Greenberg EP Valore EV Welsh MJ Ganz T Tack BF McCray PBJ Production of beta-defensins by human airway epithelia Proc Natl Acad Sci U S A 1998 95 14961 14966 9843998 10.1073/pnas.95.25.14961
Taggart CC Greene CM Smith SG Levine RL McCray PBJ O'Neill S McElvaney NG Inactivation of human beta-defensins 2 and 3 by elastolytic cathepsins J Immunol 2003 171 931 937 12847264
Vankeerberghen A Scudiero O De Boeck K Macek MJ Pignatti PF Van Hul N Nuytten H Salvatore F Castaldo G Zemkova D Vavrova V Cassiman JJ Cuppens H Distribution of human beta-defensin polymorphisms in various control and cystic fibrosis populations Genomics 2005 85 574 581 15820309 10.1016/j.ygeno.2005.02.003
Niyonsaba F Ogawa H Nagaoka I Human beta-defensin-2 functions as a chemotactic agent for tumour necrosis factor-alpha-treated human neutrophils Immunology 2004 111 273 281 15009427 10.1111/j.0019-2805.2004.01816.x
Armour JA Sismani C Patsalis PC Cross G Measurement of locus copy number by hybridisation with amplifiable probes Nucleic Acids Res 2000 28 605 609 10606661 10.1093/nar/28.2.605
www.maph.info
|
16336654
|
PMC1318481
|
CC BY
|
2021-01-04 16:37:37
|
no
|
J Negat Results Biomed. 2005 Dec 7; 4:9
|
utf-8
|
J Negat Results Biomed
| 2,005 |
10.1186/1477-5751-4-9
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1531632421610.1186/1471-2407-5-153Research ArticleThe impact of age on post-operative outcomes of colorectal cancer patients undergoing surgical treatment Latkauskas Tadas [email protected]ė Giedrė [email protected] Juozas [email protected]čiauskienė Rasa [email protected] Algimantas [email protected]žinskas Žilvinas [email protected] Dainius [email protected] Unit of Coloproctology, Department of Surgery, Kaunas Medical University Clinics, Eivenių 2, Kaunas, Lithuania2 University of Vilnius, Institute of Oncology, Santariskiu 1, Vilnius, Lithuania3 Department of oncology, Kaunas Medical University, Eivenių 2, Kaunas, Lithuania2005 2 12 2005 5 153 153 25 7 2005 2 12 2005 Copyright © 2005 Latkauskas et al; licensee BioMed Central Ltd.2005Latkauskas 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 study was to evaluate the impact of age on outcomes in colorectal cancer surgery.
Methods
patients on hospital database treated for colorectal cancer during the period 1995 – 2002 were divided into two groups: Group 1 – patients of 75 years or older (n = 154), and Group 2 – those younger than 75 years (n = 532).
Results
In Group 1, for colon cancers, proximal tumors were significantly more common (23% vs. 13.5%, p < 0.05), complicated cases were more frequent (46 % vs. 33%, p = 0.002), bowel obstruction more common at presentation (40% vs. 26.5%, p = 0.001), and more frequent emergency surgery required (24% vs. 14%, p = 0.003). Postoperative overall morbidity was higher in the elderly group, but with no differences in surgical complications rate. Overall 5 year survival was 39% vs. 55% (p = 0.0006) and cancer related 5 year survival was 44% vs. 62% (p = 0.0006). Multivariate Cox analysis showed that age was not an independent risk factor for postoperative mortality.
Conclusion
Preoperative complications and co-morbidities, more advanced disease, and higher postoperative nonsurgical complication rates adversely affect postoperative outcomes after surgery for colorectal cancer in the elderly.
==== Body
Background
Colorectal cancer is a disease of elderly, with only 5% recorded in those younger than 40 years. Elderly patients form a highly heterogeneous group in respect of both general physical status, and number and types of co-morbidities [1,2]. This, to some degree, has resulted in different concepts in management, not only from surgical and anesthesiological perspectives, but also from the expectancy of uneventful recovery and long term survival, combined with acceptable quality of life. Historically, it was suggested that elderly patients do not fare well after surgery for colorectal cancer, with high rates of emergency presentations, inoperability and peri-operative mortality [3], although more recent publications have encouraged the same surgical approach as for younger patients [2,4,5]. The purpose of this study was to evaluate impact of age on colorectal cancer presentation, surgical management and outcomes from a single institution serving a population of approximately 2 million over a six year period.
Methods
Data of all patients treated in Kaunas Medical University Hospital for colorectal cancer during 1996 – 2002 were collected retrospectively and prospectively. Data included: age; gender; location of the tumor; TNM classification; operative risk factors; whether surgery was performed in the emergency or elective settings; incidence of radical or palliative resections; and the short-term and long-term outcomes. Postoperative mortality was defined as death occurring within the first 30 days after operation. All patients were followed up in the outpatient department and were also seen by a consultant oncologist for consideration of adjuvant therapy. A decision to give adjuvant therapy was based on tumour stage, biological age of the patient and co-morbid risk factors. The Lithuanian Cancer Register supplied the date and cause of death of those who died during the follow-up period. Statistical analysis was made using SPSS package. The chi-square or Fisher exact tests were used for comparison of categorical variables between the groups. Cox analysis was used to identify independent factors for postoperative mortality.
Results
Over the six year period, 686 patients with colorectal cancer were treated. They were divided on the basis of age into two groups: Group 1 included 154 patients (23%) with ages of 75 years and older, and Group 2 included 532 patients (77%), under 75 years of age.
Demographics (Table 1)
Table 1 Age, gender, location of the tumours and TNM classification;
Group1 (age ≥ 75 years; n = 154) Group2 (age <75 years; n = 532) p
Median age 80.7 ± 4.7 58.29 ± 9.9 0,02
Gender
Male 73 270 0.5
Female 81 262
Location
Rectum 65(42%) 212 (39.8%) 0.45
Colon 88 (58%) 319 (60%)
sigmoid 42 (27%) 170 (32%) 0.24
descendens 3 (2%) 17 (3%) 0.4
left angle 5 (3.2%) 12 (2.2%) 0.5
transversum 2 (1.2%) 48 (9%) 0.001
Right colon: 36 (23%) 72 (13.5%) 0.04
right angle 7 (4.5%) 18 (3%) 0.5
ascendens 19 (14%) 35 (6.5%) 0.02
caecum 10 (6.4%) 19 (4%) 0.1
TNM stage
T3–4 464 (87%) 129 (83.7%) 0.2
N+ 62 (40%) 193 (36%) 0.02
M+ 40 (25%) 127 (23%) 0.4
The median age in Group 1 was 80.7 ± 4.7 years, and 58.29 ± 9.9 in Group2. There was no significant difference in gender ratio. There were no differences in ratio of colonic to rectal tumors between the two groups (Group 1: 89 (58%) colonic, 65 (42%) rectal; Group 2: 320 (60%) colonic, 212 (39%) rectal, p > 0.05). However, amongst colonic tumors, proximal lesions (caecum, ascending colon, hepatic flexure and proximal transverse colon) were significantly commoner in Group 1 (36 (23%); Group 2: 72 (13.5%), (p = 0.04).
When compared in respect of tumor stage, the only significant difference lay in nodal status, which was more advanced in Group 1 (node positive disease: Group 1: 62 (40%); Group 2 :193 (36%), (p = 0.02).
Complications and surgery
Overall preoperative complications (Table 2) were diagnosed in 72 (46%) patients in Group 1, compared with 176 (33%) in Group 2 (p = 0.002), the commonest being bowel obstruction (Group 1: 62 (40%); Group 2: 141 (26.5%), p = 0.001). Co-morbidities were more frequent amongst the elderly patients 80% vs. 55% (p = 0.0001), with more patients in Group 1 classified as ASA 3–5 (p = 0.0001). Emergency surgery (mainly because of bowel obstruction, or perforation) was performed in 115 (17 %) patients overall (Table 2), but was more frequent within the elderly (Group 1: 38(24%); Group 2: 77(14%), p = 0.003).
Table 2 Preoperative complications, co-morbidity, ASA (American Society of Anesthesiologists) distribution and procedures performed;
Group 1 (age ≥ 75 years; n = 154) Group 2 (age <75 years; n = 532) p
Preoperative complications: 72 (46%) 176 (33%) 0.002
Obstruction 62 (40%) 141 (26.5%) 0.001
Perforation 8 28 0.47
Co-morbidity 123 (80%) 296 (55%) 0.0001
ASA 3–5 111 248 0.0001
Surgery
Emergency 38 (24%) 77 (14%) 0.003
Curative 104 (68%) 384 (82%) 0,2
Palliative 50 (32%) 148 (18%)
Resection rate 131 (85%) 481 (90%) 0,01
On the basis of preoperative staging and surgeon decision concerning radicality, 68% of patients underwent resection with curative intent in group 1, compared with 82% in group 2 (p = 0.2). However, the resection rate was lower in elderly group – 85% vs. 90% (p = 0.01), with palliative colostomy or by-pass surgery performed in 16 (10%) patients in Group 1, compared with 20 (3.7%) in Group 2 (p = 0.0001). There were no other differences between the two groups with regard to the type of operation.
Postoperative complications were recorded in 57(37%) patients in the elderly cohort, compared with 164 (30%) in the younger group (p > 0.05), although general complications (pneumonia, pulmonary failure, pulmonary embolism, arrhythmia, myocardial infarction, cardiovascular failure and urinary infection) were more frequent in the elderly (Group 1: 34 (22%); Group 2: 73 (13.7%), p = 0.02). There were no differences between the groups in surgical complication rate (Group 1: 24 (15.5 %); Group 2: 91 (17%), p = 0.6), either in the elective or emergency setting (Table 3), but the general non-surgical complication rate was higher after elective surgery in the elderly (Group 1: 22 (18.9 %); Group 2: 53 (11.6%), p = 0.04).
Table 3 Post-operative complications
age ≥ 75 years; n = 38; emergency age <75 years; n = 77; emergency p age ≥ 75 years; n = 116; elective age <75 years; n = 455; elective p
Overall morbidity 19 (50%) 44 (57.1%) 0,4 38 (32.7%) 120 (26.4%) 0,17
General compl. 12 (31.6%) 20 (25.9%) 0,5 22 (18.96%) 53 (11.6%) 0,04
Surgical compl. 7 (18.4%) 24 (31.2%) 0,14 17 (14.6%) 67 (14.7%) 0,9
Survival
The postoperative mortality rate was 11% (n = 18) in elderly group, compared with 5% (n = 26) in the younger cohort (p = 0.002). The influence of age, TNM stage, ASA distribution, emergency operation and etc. on postoperative mortality were evaluated using Multivariate Cox analysis. The age was not an independent risk factor for postoperative mortality in mentioned analysis.
Two year survival (Figure 1) in the elderly group was 55%, compared with 67% in group 2 (p = 0.004). Five year survival was respectively 39% and 55%, (p = 0.0006). Cancer related survival (Figure 2) at 2 years was 59% vs. 70% (p = 0.004), and at 5 years, 44% vs. 62% (p = 0.0006).
Figure 1 Overall survival of patients.
Figure 2 Cancer related survival. (Group 1 < 75 years).
Discussion
No standard definition of "elderly" exists, with different authors using thresholds of 65 [2,6], 70 [7,8,16], 75 [9,10], 80 [2,4] and 85 years [17]. Data from the Lithuanian Office of National Statistics [11] shows that average life expectancy in Lithuania is 71.66 years, 65.88 years for men, and 77.41 years for women. We selected 75 years as the threshold, because it is more than Lithuanian medium life expectancy, and such a division creates a so called elderly group which constitutes approximately one quarter of all colorectal cancer patients, as used in other studies [2]. "Biological" age in different nationalities and populations varies, and it is reasonable that a 75 year old Lithuanian is equivalent to an 80 year old Western European, or 85 year old Japanese, due to differing life expectancies of these populations, although more detailed demographic assessments would be necessary to validate such assumptions.
Previous studies have demonstrated an age-related right shift of colorectal cancer [12,13], this supported by the present study. Marush et al. reported ageing was associated with more locally advanced tumors, but not with metastatic dissemination [2]. This study revealed only differences in nodal status, but no differences between the groups in T stage. However, mechanical bowel obstruction (the most common preoperative complication in both groups), a clinical indicator of locally advanced disease was more frequent in the elderly group. Acute presentation was more frequent in the elderly group, with emergency surgery performed in 24% of such patients, compared to 14% in younger group (p = 0.003), similar to previous studies reported [2,17].
Our study has confirmed that the majority of complications arising in the surgical management of elderly patients with colorectal cancer are not truly surgical, but of a more general nature, both pre- and post-operatively, the latter perhaps compounded by less than optimal preoperative preparation. Menke et al. [14] reported a co-morbidity of 28.1% in patients older than 80 years, and Wolters et al. [8] incidences of 49% for hypertension, 18% for coronary heart disease, and 39% for pulmonary disease, equivalent to our total co-morbidity frequency of 80%, compared to 55% in the younger patient group (p = 0.0001).
Resection rates for elderly patients are usually slightly lower than in younger patients, due to preoperative complications and co-morbidities, and more extensive tumors [2,4,17]. Nevertheless, advances made in surgical technique, anesthetic and postoperative intensive care have resulted in an increase in possibility to perform surgery from 80% up to 95% of cases [16]. The resection rate in our study was 85% in the elderly, comparing with 90% in those younger than 75 years (p = 0.01), with a concomitant increase in the rates of palliative colostomy formation and by-pass procedures in the elderly (10% vs. 3.7%), (p = 0.0001). The recorded reasons for not performing resection included locally advanced disease, presence of preoperative complications and co-morbidities, and emergency nature of the surgery. Postoperative morbidity and mortality is a significant source of concern in the management of the elderly patient with colorectal cancer. Postoperative morbidity is governed by a higher incidence of general complications rate [2,17], with specific surgical postoperative complications occurring at no greater frequency than in younger patients [2]. The difference in rates of general complications following elective surgery, but not after emergency surgery is more difficult to explain.
Overall mortality was 11% in the elderly group, compared with 5% in those under 75 years old (p = 0.002). Emergency surgery for colorectal carcinoma in the elderly is associated with higher morbidity and mortality, reported rates varying between 6% and 38% for emergency operations and 0.9% and 18% for elective operations in those over 70 [7]. The risk of postoperative death in patients over 80 years rises to 11.9 – 38% after emergency surgery, and 7.4 – 11.4% in elective cases [7,8], making the postoperative mortality rates of the present study acceptable.
Overall survival is, not surprisingly, poorer in the elderly, but any differences are much less strong when expressed as cancer related survival rates [4,15]. The overall 2-years and 5-years survival rate in patients over 75 years of age was lower than that observed in younger patients in the present study, perhaps due to the lower frequency of curative operations, and the higher proportion of deaths from other causes in the elderly group. Barrier et al. [4] reported that in those patients operated with curative intent, the 5-year cancer-specific survival rate was not significantly different between the two age groups. Unfortunately the results of the present study do not support this; it is possible however, that a registered cause of death as cancer, without detailed evaluation or postmortem examination, was in fact incorrect.
Conclusion
In Lithuania, patients with colorectal cancer have similar demographic profiles as those in other countries. Based on a threshold of 75 years, preoperative complication and emergency surgery rates are more common in elderly patients, but postoperative surgical morbidity rates are similar to those observed in younger patients. Postoperative non-surgical morbidity was higher in the elderly group, which influenced postoperative mortality. Overall survival was better in younger patients and the same difference remained in cancer specific long term survival.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TL abstracted data, made data analysis, drafted and revised the manuscript.
GR collected data, performed statistical analysis.
JK participated in data extraction from Lithuanian cancer registry.
RJ collected and abstracted data, participated in the planning of the study.
AT participated in the planning of the study and coordinated the writing of the manuscript.
ZS participated in the planning of the study and coordinated the writing of the manuscript.
DP participated in the planning of the study and coordinated the writing and approved the final manuscript.
All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Ries LAG Eisner MP Kosary CL editors SEER cancer statistics review, 1973–1997 2000 Bethesda, MD: Nacional Cancer Institute, NIH Pub. No.00-2789
Marusch F Koch A Schmidt U Zippel R Gastmeier K Impact of age on the short- term postoperative outcome of patients undergoing surgery for colorectal carcinoma Colorectal Dis 2002 17 177 184 10.1007/s003840100344
Rankin FW Johnson CC Major operations in elderly patients Surgery 1939 5 763 774
Barrier A Ferro L Houry S Lacaine F Huguire M Rectal cancer surgery in patients more than 80 years of age The American J of Surg 2003 185 54 57 10.1016/S0002-9610(02)01120-0
Sunouchi K Namiki K Mori M How should patients 80 years of age or older with colorectal carcinoma be treated? Long-term and short-term outcome and postoperative cytokine levels Dis Colon Rectum 2000 43 233 241 10696898 10.1007/BF02236988
Paksoy M Ipek T Colak T Cebeci H Influence of age on prognosis and management of patients with colorectal carcinoma Eur J Surg 1999 165 55 59 10069635
Waldron R Donovan I Drumm J Mottram S Tedman S Emergency presentation and mortality from colorectal cancer in the elderly Br J of Surg 1986 73 214 216 3947921
Wolters U Isenberg J Stutzer H Colorectal carcinoma – aspects of surgery in the elderly Anticancer Res 1997 17 1273 1276 9137484
Tomoda H Tsujitani S Furusawa M Surgery for colorectal cancer in elderly patients- a comparison with younger adult patients Jpn J Surg 1998 18 397 402 3172581
Makela J Kiviniemi H Laitinen S Survival after operations for colorectal cancer in patients aged 75 years and over Eur J Surg 2000 166 472 479
Office of national Statistics Demografic and cancer statistics 1996–2000
Kempainen M Raiha I Rajala T Souranda L Characteristics of colorectal cancer in elderly patients Gerontology 1993 39 222 227 8244050
Arai T Takubo K Sawabe M Esaki Y Pathologig characteristics of colorectal cancer in the elderly; a retrospective study of 947 surgical cases Clin Gastroenterology 2000 31 67 72 10.1097/00004836-200007000-00016
Menke H Graf J Heintz A Klein A Junginger T Risk factors of perioerative morbidity and mortality with special reference to tumor stage, site and age Zentralbl Chir 1993 118 40 46 8451887
Edna TH Bjerkeset T Colorectal cancer in patients of over 80 years of age Hepatogastroenterology 1998 45 2142 2145 9951881
Poon R Law W Chu K Wong J Emergency resection and primary anastomosis for left-sided obstructing colorectal carcinoma in the elderly Br J Surg 1998 85 1539 1542 9823920 10.1046/j.1365-2168.1998.00903.x
Colorectal Cancer Collaborative Group Surgery for colorectal cancer in elderly patients; A systematic review. Colorectal Cancer Collaborative Group Lancet 2000 356 968 974 11041397 10.1016/S0140-6736(00)02713-6
|
16324216
|
PMC1318482
|
CC BY
|
2021-01-04 16:39:11
|
no
|
BMC Cancer. 2005 Dec 2; 5:153
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-153
|
oa_comm
|
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-401630768010.1186/1471-2350-6-40Research ArticleGenetic, household and spatial clustering of leprosy on an island in Indonesia: a population-based study Bakker Mirjam I [email protected] Linda [email protected] Mochammad [email protected] Agnes [email protected] Paul R [email protected] Linda [email protected] Jeanine J [email protected] KIT (Koninklijk Instituut voor de Tropen/Royal Tropical Institute), KIT Biomedical Research, Meibergdreef 39, 1105 AZ Amsterdam, The Netherlands2 Department of Microbiology, Faculty of Medicine, Hasanuddin University, Kampus Tamalanrea KM 10, Jln. Perintis Kemerdekaan, 90245 Makassar, Indonesia3 Department of Medical Statistics and Bioinformatics, Leiden University Medical Centre, Postbus 9600, 2300 RC Leiden, The Netherlands2005 24 11 2005 6 40 40 27 6 2005 24 11 2005 Copyright © 2005 Bakker 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
It is generally accepted that genetic factors play a role in susceptibility to both leprosy per se and leprosy type, but only few studies have tempted to quantify this. Estimating the contribution of genetic factors to clustering of leprosy within families is difficult since these persons often share the same environment. The first aim of this study was to test which correlation structure (genetic, household or spatial) gives the best explanation for the distribution of leprosy patients and seropositive persons and second to quantify the role of genetic factors in the occurrence of leprosy and seropositivity.
Methods
The three correlation structures were proposed for population data (n = 560), collected on a geographically isolated island highly endemic for leprosy, to explain the distribution of leprosy per se, leprosy type and persons harbouring Mycobacterium leprae-specific antibodies. Heritability estimates and risk ratios for siblings were calculated to quantify the genetic effect. Leprosy was clinically diagnosed and specific anti-M. leprae antibodies were measured using ELISA.
Results
For leprosy per se in the total population the genetic correlation structure fitted best. In the population with relative stable household status (persons under 21 years and above 39 years) all structures were significant. For multibacillary leprosy (MB) genetic factors seemed more important than for paucibacillary leprosy. Seropositivity could be explained best by the spatial model, but the genetic model was also significant. Heritability was 57% for leprosy per se and 31% for seropositivity.
Conclusion
Genetic factors seem to play an important role in the clustering of patients with a more advanced form of leprosy, and they could explain more than half of the total phenotypic variance.
==== Body
Background
Leprosy, which annually affects about 700,000 new patients world wide [1], is a chronic disease caused by Mycobacterium leprae. It is thought that in endemic areas many individuals are infected with M. leprae, but only a few of those infected actually develop the disease [2]. Leprosy manifests itself as a disease spectrum, which for treatment purposes has been divided into two forms: multibacillary (MB) and paucibacillary (PB) leprosy. Clustering of leprosy patients within households, neighbourhoods and families has been reported several times [among others: [3-6]].
The discussion of the role played by genetic factors in the susceptibility to leprosy is long-standing, but came to prominence again with the recent identification of certain susceptibility loci [7-9]. In the past many studies have been performed on the genetic susceptibility of leprosy, such as segregation studies to unravel the mode of inheritance [10,11] and association studies to identify susceptibility genes [12,13]. Several reviews on this topic have been written [14,15]. It has been generally accepted that genetic factors do play a role in susceptibility to both leprosy per se and leprosy type and that probably multiple genes are involved. Suggestions have been made for a strong genetic component [16,17], but only few studies have tempted to quantify this. Wallace, Clayton and Fine [18] estimated the relative recurrence risk ratio to measure the genetic effect in Northern Malawi. They suggested that both genetic and non-genetic factors may play a role in the susceptibility to leprosy, with host genetics playing a small but significant role (siblings risk ratio was 2).
Interpretation of leprosy clustering within families or households is difficult, since those persons may not only share the same genes but also have close contact and the same or similar socio-economic circumstances. The main research question of this study was to test which correlation structure (genetic, household or spatial) gives the best explanation for the distribution of leprosy patients and persons harbouring specific anti-M. leprae antibodies (seropositive persons) in our study population. Furthermore, heritability estimates and the siblings recurrence risk ratio were calculated to give an indication of the maximum contribution of genetic factors to the occurrence of leprosy and seropositivity. The presence of M. leprae specific antibodies was used as a marker for leprosy infection, and was recently identified as a risk factor for MB leprosy [19]. The study was performed in a geographically isolated island population, that has a strong founder effect as it was founded by only a few persons about 100 years ago. Of this unique island population, which was found to be highly endemic for leprosy during a population survey in 2000 [6], the family structure was unravelled and an extended pedigree prepared.
Methods
Prior to the study we received ethical clearance from the Ethical Research Committee of the Hasanuddin University and from the Ministry of Health of the Republic of Indonesia.
Study population
The study described here was performed on the island Kembanglemari, which has 634 inhabitants and is situated in the Flores Sea. The island is part of Pangkep District of South Sulawesi Province in Indonesia and located 268 km from Makassar. The inhabitants originate from South Sulawesi and belong to the Mandar ethnic group.
Data and sample collection
Clinical data were collected in June 2000. During an active door-to-door survey 88.3% of the population was examined for clinical symptoms of leprosy [6]. The diagnosis was based on the WHO classification. Patients with one lesion were classified as PB1 and with 2–5 lesions as PB2-5; patients with more than five lesions and/or with a positive bacterial index (BI) in at least one of three skin smears were classified as MB. Persons who reported to have completed a full course of multi-drug treatment, without active lesions and skin smear negative, were marked as patients released from treatment (RFT).
At the same time blood samples were collected of the population above 5 years: 68.1% of the population. Serum was separated by centrifugation on the same day and kept frozen until use.
During two subsequent population surveys in April 2002 and April 2003 the parent names of the majority of the inhabitants were asked. Furthermore, during the survey in April 2002 interviews were held with elderly people and leprosy patients about their family structure and ancestors. With these data an extended pedigree was prepared. To determine the occurrence of inbreeding the kinship coefficient (the probability that two alleles, at a randomly chosen locus, one chosen randomly from individual i and one from j are identical by descent) was computed for parents [20].
The longitudes and latitudes of every fifth house were measured using a hand-held Global Positioning System (GPS, Garmin, Kansas USA). In Arcview 3.2 (Esri, California USA) the remaining houses were situated between the geo-referenced houses using a detailed hand-drawn map. The resulting map was used to prepare a geographical distance matrix between all inhabitants.
IgM antibody detection
The presence of IgM antibodies to M. leprae phenolic glycolypid I (PGL-I) was measured by an enzyme-linked immunosorbent assay (ELISA) as described previously [21] using the natural trisaccharide moiety of PGL-I linked to bovine serum albumin (NT-P-BSA) as antigen. Serum was diluted 1:500 and tested in duplo. The optical density at 450 nm (OD) of each serum sample was calculated by subtracting the OD value of BSA coated wells from that of NT-P-BSA coated wells. A positive reference serum on each plate was used to minimize plate-to-plate variation. The cut-off value for seropositivity was set at 0.200 [21]; any criterion for setting a cut-off is arbitrary since the distribution of antibody concentration is unimodal [22].
Statistical analysis
Leprosy prevalence was defined as the proportion of the sum of leprosy patients and RFT patients over the population screened for leprosy in June 2000. Even though it is not common practice, for the purpose of this particular research question RFT patients were included in the prevalence. Seroprevalence was defined as the proportion of seropositive persons (including seropositive patients) over the population screened for antibodies.
A score statistic Q [23]
Q=∑i,j(zi−πi)Rij(zj−πj)∑i(zi−πi)2
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGrbqucqGH9aqpdaWcaaqaamaaqafabaGaeiikaGIaemOEaO3aaSbaaSqaaiabdMgaPbqabaGccqGHsisliiaacqWFapaCdaWgaaWcbaGaemyAaKgabeaakiabcMcaPiabdkfasnaaBaaaleaacqWGPbqAcqWGQbGAaeqaaOGaeiikaGIaemOEaO3aaSbaaSqaaiabdQgaQbqabaGccqGHsislcqWFapaCdaWgaaWcbaGaemOAaOgabeaakiabcMcaPaWcbaGaemyAaKMaeiilaWIaemOAaOgabeqdcqGHris5aaGcbaWaaabuaeaacqGGOaakcqWG6bGEdaWgaaWcbaGaemyAaKgabeaakiabgkHiTiab=b8aWnaaBaaaleaacqWGPbqAaeqaaOGaeiykaKYaaWbaaSqabeaacqaIYaGmaaaabaGaemyAaKgabeqdcqGHris5aaaaaaa@5851@
was used to test clustering of leprosy per se, MB leprosy, PB leprosy and seropositivity due to genetic, household and spatial effects. Here zi is the outcome for subject i (1 if affected and 0 otherwise), πi is the age and sex specific prevalence and Rij is the genetic, household or spatial correlation for subject i and j. The specific Rij are described below. In the simple case of πi = 0.5 for all i, the statistic Q reduces to the sum over concordant pairs (i,j) (for example leprosy patient - leprosy patient) of Rij minus the sum over disconcordant pairs (i,j) (leprosy patient - person with no leprosy) of Rij. In general the statistic Q tends to be large when concordant pairs have higher correlations Rij compared to discordant pairs. For the score test it is important to realise that healthy persons also provide information, although not as much as the patients. The distribution of Q under the null hypothesis of no correlation can be approximated by a chi-square distribution with scale parameter 0.5Var(Q)/E(Q) and degrees of freedom of 2E(Q)2/Var(Q). Formulae for the expectation and variance of Q can be found in Houwing-Duistermaat et al [23].
The correlation structures Rij corresponding to the genetic, household and spatial effects were based on distances between individuals. For the genetic model correlation between pairs is based on genetic distances (dg) in the pedigree; siblings have a higher correlation compared to cousin-pairs and unrelated persons have no correlation. Specifically Rij = 1/2dg which is equivalent to two times the kinship coefficient. In the household model the distances between individuals sharing the same household is zero which gives a correlation of 1, and distance infinite for inhabitants of different households (Rij = 0). The spatial model is an extension of the household model. The distance for the spatial model (de) equals the distance between 2 households in metres. We used the following formula: Rij = exp(-deij/44). In previous studies it was shown that apart from household contacts also first and second neighbours have an increased risk of developing leprosy [5]. The number 44 still gives a good correlation between a house and its second neighbour: for de = 11 (the mean distance between a house and its nearest (first) neighbour) Rij = 0.779, for de = 22 (the assumed distance between house and its second neighbour) Rij = 0.607 and for de = 33 Rij = 0.473. This last correlation is seen as a moderate correlation [24]. Thus, the correlation decreases when the distance between 2 households becomes larger. We performed a sensitivity analysis in which the number 44 was changed into 33 and 55. Spearman rank correlation coefficients were computed between the correlations Rij of the different random effects.
In the analysis for leprosy per se all patients detected in June 2000 and the RFT patients were included. These RFT patients were excluded from the separate analyses for MB and PB leprosy, because classification could not be confirmed. For leprosy per se the test was, apart from the total population, also performed on a subpopulation which was expected to have a relatively stable household status over the last 20 years, namely the population below 21 and above 39 years. From the data it was seen that up to the age of 20 84% (291/346) still lived in the same house as their mother, and that after the age of 20 this percentage was much smaller (12%; 25/214), indicating that most people moved when they were around 20 year of age. Interviews learned that most people move only once in their life, namely when they get married and move from their parental house into their own house. Persons aged 21–39 were excluded because it is expected that most of these persons had a change in household status within the last 20 years.
Heritability estimates were calculated for leprosy per se and for seropositivity using a random effects model with a logit link and assuming Gaussian random effects [25]. The heritability estimates are presented with two-sided 95% confidence intervals (95% CI). The confidence intervals were estimated using profile likelihood. Both the score statistic and the heritability estimates were adjusted for the covariates age (continuous) and sex.
Finally, the risk ratios for siblings (λs) for leprosy per se and seropositivity, defined as the ratio of the risk of leprosy/seropositivity for siblings of affected persons to the risk for the general population, were calculated separately for the group under 21 years and the group of 21 years and older according to the method described by Olson and Cordell [26]. Confidence intervals were calculated according to the method of Zou and Zhao [27]. Different age groups were used because the risk of leproys/seropositivity for the general population (i.e. prevalence) differed between age groups.
Results
In June 2000 634 persons were living in the 120 houses on Kembanglemari (average: 5.3 persons per house). Of the 560 persons screened, 28 were diagnosed with leprosy (12 were classified as MB, 3 as PB2-5 and 13 as PB1) and 3 persons were identified who had leprosy in the past, but were released from treatment (Table 1). Of the 432 persons who were tested for antibodies against M. leprae, 37 (8.6%) were seropositive. Two (17%) of the 12 MB patients, none of the PB patients and none of the RFT patients were seropositive. Table 2 shows leprosy and seroprevalence for the different age groups: leprosy was more prevalent among adults than among children while seroprevalence was higher among children. No difference was found between men and women with regard to leprosy (4.9% and 6.1%, respectively, p = 0.52) or seropositivity (7.4% and 9.6%, respectively, p = 0.41). Figures 1 and 2 show maps of the island with the patients and the seropositive persons per house, respectively.
Four families were identified on the island. The pedigrees included 791 persons in total, of which 589 were living on the island in June 2000 and 202 persons were parents who had already died, lived somewhere else or could not be traced. One pedigree included 568 (89.6%) inhabitants, all leprosy patients and 35 seropositive persons and consisted of six generations. Figure 3 shows part of this pedigree as an example. Of the 560 persons screened for leprosy 535 were included in the pedigrees and of the 432 persons who were tested for antibodies 411 were included. The unrelated single individuals (25 in leprosy analyses and 21 in seropositivity analyses) were retained in the analysis as independents.
In total 184 couples with children were counted. Of 79 pairs who were alive in 2000, both parents were known. Of these, 15 pairs (19%) were genetically related (kinship coefficient varied between 0.156–0.0156). They had 62 children (56 screened for leprosy, 34 tested for antibodies). None of the children had leprosy and 4 were seropositive (11.8%).
For the total population the test statistic of Houwing-Duistermaat et al. [23] (results in Table 3) showed that for leprosy per se and MB leprosy the genetic correlation structure fitted best (p < 0.001 and p = 0.002, respectively). For MB leprosy the household model also fitted (p = 0.005). For PB leprosy neither the genetic nor the household nor the spatial correlation structure were significant. For seropositivity the spatial correlation structure fitted best (p = 0.003), but the genetic model was significant too (p = 0.016). For leprosy per se in the population below 21 and above 39 years the genetic correlation structure fitted best, but all structures were significant (p < 0.001 for genetic, p = 0.002 for household and p = 0.017 for spatial). The sensitivity analysis showed no substantial differences in the results when using 33 or 55 instead of 44 in the spatial correlation structure. In the total population, the Spearman rank correlation coefficient between the spatial and genetic correlations was 0.10, between the household and genetic correlations 0.15, and between the spatial and household correlations 0.16.
The heritability estimate for leprosy was 0.57 (95% CI: 0.22–0.81) and for seropositivity 0.31 (95% CI: 0–0.71). The risk ratio for siblings (λS) for leprosy per se in the population below 21 years was 6.4 (95% CI: 0–17.0) and in the population above 20 years 2.9 (95% CI: 0–9.3). For seropositivity λS was 1.3 (95% CI: 0–3.0) in the population <21 years and 2.7 (95% CI: 0–7.0) in the population >20 years.
Discussion
Strengths of the study
This study describes a unique island population with a strong founder effect: 90% of the population belonged to the same pedigree in which consanguineous marriages took place and the leprosy prevalence was extremely high (5.5%). This makes the population very suitable to study whether genetic effects can explain the distribution of leprosy-related traits within the families.
Principal findings
In the total population clustering of leprosy per se could only be explained by genetic factors and not by contact status. In this particular population the heritability of leprosy per se was 57%. For PB leprosy no clustering could be detected, but for MB leprosy both the genetic and the household were significant. For seropositivity genetic factors seemed less important compared to leprosy: the heritability of seropositivity was lower, namely 31%, and although the genetic model was significant, the spatial model explained the clustering of seropositivity better.
Potential biases
This population-based study included 560 persons of which 31 were affected with leprosy. Although the total population is large, and has one of the highest prevalences of leprosy in the world, the number of leprosy patients is still rather small. Therefore the confidence intervals of the heritability and siblings risk ratios are rather large, which may limit the weight that should be given to the results.
Since the three effects give similar correlation structures in the data (family members are for example living in the same house), a significant effect may be a confounder for one of the other effects [28]. Therefore the heritability estimates may also partly reflect shared environmental effects. Fortunately the correlations between the structures appeared to be rather small (≤0.16) indicating only small overlap between the various effects. The score statistics indicate which correlation structure fits best to the data. Thus the heritability estimate of 0.57 for leprosy represents genetic effects while the heritability of seropositivy probably also measures spatial effects, since the spatial correlation structure fitted better to the data than the genetic correlation structure.
Information bias could have occurred since leprosy patients were interviewed as a separate group and they may have had a better insight into their genetic distance to other leprosy patients. However, since information was collected in multiple ways, i.e. also by interviewing elderly persons and through two population surveys, it is expected that this bias will be minimal.
Household status as well as the distance matrix (used for the spatial model) were determined at the time of the screening in June 2000. The household status in 2000 does not necessarily reflect the household status at time of transmission/infection due to the long incubation time (estimated to vary between 2 and 12 years [29]). During this time patients could have moved to different houses. To overcome this problem we decided to apply the test statistic for leprosy per se also for the population excluding those aged between 20 and 40 years. Young adults seem to move out of their parental house around their 20th birthday. The 20-year lag time is to take into account the incubation period and detection delay. We assumed that within the group younger than 20 and older than 40 years the household status for most patients would be similar at time of infection and diagnosis.
Interpretation
In contrast to the results of the general population, where susceptibility to leprosy per se could only be significantly explained by genetic factors, among the population below 21 and above 39 years all three correlation structures were significant. Especially the household and genetic effects were highly significant in this subgroup, making it more difficult to distinguish between the effects. Genetically closely related persons probably live in the same household for this age group. The smallest p-value for genetic effects suggests, however, that genetically related individuals living in different households tend to have similar outcomes and/or distantly related individuals living in the same household have different outcomes.
When looking separately for MB and PB leprosy in the total population, MB leprosy appeared to have a genetic as well as a household effect. However, for PB leprosy clustering could not be detected at all, meaning that it was randomly spread in the population. The PB group consisted mainly of PB patients with a single lesion (78%). Part of these PB1 patients may have spontaneously healed if there had been no active screening [16]. The fact that we do not have many PB patients with 2 to 5 lesions, could be due to over-diagnosis of MB patients. In that way it would be better to describe the MB patients as the patients with a more advanced form of leprosy. It seems that for actual progression to a more advanced state of disease genetic factors become more important.
Although not significant, probably due to small numbers, a relatively high recurrence risk ratio of 6.4 for siblings for leprosy per se was found in the population below 21 years, indicating that brothers/sisters of a patient had a more than 6 times higher risk of developing leprosy compared to the general population in that age group. Since in this young population most of the siblings still live together in the parental house, the recurrence risk ratio measures also household effects. In the population above 20 years the recurrence risk ratio for siblings was 2.9, which is comparable to other studies: in a population in South India a λs of 2.4 for tuberculoid leprosy [30] and in a population in Vietnam a λs of 2.21 for leprosy per se [17] has been estimated. The λs for infectious diseases usually lies between 1.5–5 which is much lower than for example for autoimmune diseases like type 1 diabetes and multiple sclerosis with λs between 15–20 [31].
The recurrence risk ratio for siblings measures the excess risk of siblings of affected persons compared to the population risk and could be used for developing rational screening procedures. However, it is not adjusted for age or sex and only uses the information of clustering within siblings. Moreover within siblings the effect of sharing households will be relatively large. In contrast, the heritability estimates are based on correlation between all genetically related subjects and thus also on relatives who do not share household for example cousin pairs. It estimates the proportion of the genetic variance explaining the total phenotypic variance in a defined population and is used to quantify the degree of genetic contribution to the development of a disease. We found that in the total population for leprosy per se 57% of the total variance could be explained by genetic factors. Probably MB leprosy is responsible for most of this, since the genetic effect was highly significant for MB leprosy but not significant for PB leprosy. Only one other study, performed in the Philippines, described a heritability of lepromatous leprosy among men of 80% [32]. However, this relatively high estimate is only based on siblings and thus the contribution of shared environment to the estimate may be relatively high.
The distribution of seropositive persons could be explained best by the spatial model, but the genetic model was also significant. The household model could not significantly explain clustering. A heritability of 31% was found and in the population below 21 years siblings of seropositive persons did not have an increased risk to be seropositive compared to the general young population. In the population above 20 years the recurrence risk ratio for siblings was 2.7, which is comparable with that for leprosy per se. It seems that especially in the young population genetic factors are less important for seropositivity than for leprosy, which could be reflected by the high seroprevalence in the young population. In a recent publication we showed that living in the vicinity of two seropositive patients increased the risk of harbouring antibodies against M. leprae [33]. It seems that having contact with an infectious patient is an important factor in harbouring antibodies, but to develop MB leprosy genetic factors become more important.
Leprosy and seroprevalence were not significantly increased among children of genetically related parents. However, the sample size was rather small (56 and 34 children, respectively), which makes it difficult to draw conclusions. In our study genetically related parents have a higher chance to appear in the younger generations of the pedigree, since in the older generations the correlation between parents was often unknown. The fact that we found 15 pairs of related parents, of which three persons who married a related person also had parents who were related, indicates that marrying a related person is customary on this island. This suggests that also in the older generations of the family tree inbreeding may have occurred, which makes this population interesting for studying recessive genes.
Use of study population for detection of genes
Until now genome scans that are published either studied the phenotypes leprosy per se [17] or PB leprosy [7]. Here we showed that in our population genetic factors appear to be important for advanced forms of leprosy. A logical next step would be to see which (candidate) genes, such as the PARK2 and PACRG genes [9], could explain the genetic effect found in this population. If none of the already known chromosome regions can explain the effect, a genome-wide scan could be carried out to detect new regions.
Conclusion
Since leprosy is thought to spread from person to person, contact with a leprosy patient is essential for the transmission of M. leprae. Among many factors that could influence the development of infection and disease, such as age, nutritional status and contact with other mycobacteria, genetic factors probably also play a role. In this highly endemic area for leprosy genetic factors could explain up to 57% of the total variance. This unique study population is very suitable to confirm the role of already known chromosome regions in controlling leprosy or to search for new susceptibility loci.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MB was involved in study design, data collection and statistical analyses and finalized the manuscript. LM was involved in study design, data collection and wrote the first draft of the paper. MH was involved in study design and coordinated the data collection. AK participated in implementation and data collection. PK was involved in study design and interpretation of results. LO was involved in study design and supervision. JH performed the statistical analysis and interpreted results. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We gratefully acknowledge the Netherlands Leprosy Relief for the financial support of this study. We appreciate the permission and active support given by the manager and personnel of Health District Pangkep. We want to thank all people who were part of the research team. We are grateful to all the inhabitants of the island Kembanglemari for their co-operation.
Figures and Tables
Figure 1 Map of Kembanglemari Island showing the new patients per house. = 0, ● = 1, = 2, + = 3 new patients. Not all houses without patients are shown.
Figure 2 Map of Kembanglemari Island showing the seropositive persons per house. = 0, ● = 1, = 2, + = 3 seropositive persons. Not all houses without seropositive persons are shown.
Figure 3 Part of the extended pedigree of family one, showing family relations between leprosy patients.
Table 1 Characteristics of study population in June 2000
n %
Persons living on Kembanglemari 634
Persons screened for leprosy 560 88.3
Persons screened for antibodies 432 68.1
Persons included in pedigrees 589 92.9
Leprosy status
No leprosy 529 94.5
PB1a leprosy 13 2.3
PB2-5b leprosy 3 0.5
MBc leprosy 12 1.9
RFTd (no classification available) 3 0.5
Serological status
Seronegative 395 91.4
Seropositive 37 8.6
a PB1 = single lesion paucibacillary leprosy
b PB2-5 = paucibacillary leprosy with 2–5 lesions
c MB = multibacillary leprosy
d RFT = released from treatment
Table 2 Leprosy and seroprevalence per agegroup
Age group (years) Screened for leprosy RFTa MB Leprosyb PB Leprosyc Leprosy N (%) Tested for IgM Seropositive N (%)
0–5 122 0 0 0 0 7 1 (14.3)
6–20 219 0 4 8 12 (5.5) 211 24 (11.4)
21–39 136 2 5 5 12 (8.8) 133 9 (6.8)
40–59 57 1 3 0 4 (7.0) 55 1 (1.8)
≥60 26 0 0 3 3 (11.5) 26 2 (7.7)
Total 560 3 12 16 31 (5.5) 432 37 (8.6)
a RFT = released from treatment
b MB = multibacillary leprosy
c PB = paucibacillary leprosy
Table 3 P-values for testing the null hypothesis of no clustering of leprosy per se, MB leprosy, PB leprosy and of seropositivity due to genetic, household and spatial effects.
Study population Trait N (sero)preve. Genetic Household Spatial
Total population
Leprosy per se 560 5.5%
<0.001
0.084 0.145
MBa leprosy 541c 2.2%
0.002
0.005 0.105
PBb leprosy 545d 2.9% 0.311 0.154 0.555
Seropositivity 432 8.6% 0.016 0.118
0.003
Population <21 and >39 years
Leprosy per se 424 4.5%
<0.001
0.002 0.017
Test-statistic is adjusted for age and sex.
Italic: p-value <0.05; Bold: best fitting model
a MB = multibacillary leprosy; b PB = paucibacillary leprosy
c 541 = 560 - 3 RFT patients - 16 PB patients
d 545 = 560 - 3 RFT patients - 12 MB patients
e (sero)prev = leprosy prevalence or seroprevalence
==== Refs
World Health Organization Leprosy – Global situation Wkly Epidemiol Rec 2002 1 1 8
Britton WJ Lockwood DNJ Leprosy Lancet 2004 363 1209 1219 15081655 10.1016/S0140-6736(04)15952-7
Bagshawe A Burgh de S Fung SC Chuah J Berry G The epidemiology of leprosy in a high prevalence village in Papua New Guinea Trans R Soc Trop Med Hyg 1989 89 121 127 2603189 10.1016/0035-9203(89)90735-9
Fine PE Sterne JA Ponnighaus JM Bliss L Saui J Chihana A Munthali M Warndorff DK Household and dwelling contact as risk factors for leprosy in Northern Malawi Am J Epidemiol 1997 146 91 102 9215227
Van Beers SM Hatta M Klatser PR Patient contact is the major determinant in incident leprosy: implication for future control Int J Lepr Other Mycobact Dis 1999 67 119 128 10472363
Bakker MI Hatta M Kwenang A Klatser PR Oskam L Epidemiology of leprosy of five isolated islands in the Flores Sea, Indonesia Trop Med Int Health 2002 7 780 787 12225510 10.1046/j.1365-3156.2002.00931.x
Siddiqui MR Meisner S Tosh K Balaksishnan K Ghei S Fisher SE Golding M Narayan NPS Sitaraman T Sengupta U Pitchappan R Hill AVS A major susceptibility locus for leprosy in India maps to chromosome 10p13 Nat genet 2001 27 439 441 11279529 10.1038/86958
Tosh K Meisner S Siddiqui MR Balakrishnan K Ghei S Golding M Sengupta U Pitchappan RM Hill AVS A region of chromosome 20 is linked to leprosy susceptibility in a South Indian population J Infect Dis 2002 186 1190 1193 12355375 10.1086/343806
Mira MT Alcais A Nguyen VT Moraes MO Di Flumeri C Vu HT Mai CP Nguyen TH Nguyen NB Pham XK Sarno EN Alter A Montpetit A Moraes ME Moraes JR Dore C Gallant CJ Lepage P Verner A Van De Vosse E Hudson TJ Abel L Schurr E Susceptibility to leprosy is associated with PARK2 and PACRG Nature 2004 427 636 640 14737177 10.1038/nature02326
Wagener DK Schauf V Nelson KE Scollard D Brown A Smith T Segregation analysis of leprosy in families of northern Thailand Genet Epidemiol 1988 5 95 105 3402727 10.1002/gepi.1370050205
Abel L Vu DL Oberti J Nguyen VT Van VC Guilloud-Bataille M Schurr E Lagrange PH Complex segregation analysis of leprosy in southern Vietnam Genet Epidemiol 1995 12 63 82 7713401 10.1002/gepi.1370120107
Roy S Frodsham A Saha B Hazra SK Mascie-Taylor CG Hill AV Association of vitamin D receptor genotype with leprosy type J Infect Dis 1999 179 187 191 9841838 10.1086/314536
Shaw MA Donaldson IJ Collins A Peacock CS Lins-Lainson Z Shaw JJ Ramos F Silveira F Blackwell JM Association and linkage of leprosy phenotypes with HLA class II and tumour necrosis factor genes Genes Immun 2001 2 196 204 11477474 10.1038/sj.gene.6363754
Casanova JL Abel L Genetic Dissection of Immunity to Mycobacteria: the human model Annu Rev Immunol 2002 20 581 620 11861613 10.1146/annurev.immunol.20.081501.125851
Fitness J Tosh H Hill AVS Genetics of susceptibility to leprosy Genes Immun 2002 3 441 453 12486602 10.1038/sj.gene.6363926
Shields ED Russell DA Pericak-Vance MA Genetic epidemiology of the susceptibility of leprosy J Clin Invest 1987 79 1139 1143 3549780
Mira MT Alcaïs A Thuc NV Tahi VH Huong NT Ba NN Verner A Hudson TJ Abel L Schurr E Chromosome 6q25 is linked to susceptibility to leprosy in a Vietnamese population Nat Genet 2003 33 412 415 12577057 10.1038/ng1096
Wallace C Clayton D Fine P Estimating the relative recurrence risk ratio for leprosy in Karonga District, Malawi Lepr Rev 2003 74 133 140 12862254
Douglas JT Cellona RV Fajardo TT Abalos RM Balagon MVF Klatser PR Prospective study of serological conversion as a risk factor for development of leprosy among household contacts Clin Diagn Lab Immunol 2004 11 897 900 15358649 10.1128/CDLI.11.5.897-900.2004
Thompson EA Pedigree Analysis in Human Genetics 1986 Baltimore: Johns Hopkins University Press
Bührer-Sékula S Sarno EN Oskam L Koop S Wichers I Nery JA Vieira LM de Matos HJ Faber WR Klatser PR Use of ML dipstick as a tool to classify leprosy patients Int J Lepr Other Mycobact Dis 2000 68 456 463 11332289
Fine PEM Ponnighaus JM Burgess P Clarkson JA Draper CC Seroepidemiological studies of leprosy in Northern Malawi on an enzyme-linked immunosorbent assay using synthetic glycoconjugate antigen Int J Lepr Other Mycobact Dis 1988 56 243 254 3045222
Houwing-Duistermaat JJ Derkx BH Rosendaal FR van Houwelingen HC Testing familial aggregation Biometrics 1995 51 1292 1301 8589223
Altman DG Practical statistics for medical research 1991 London: Chapman & Hall
Houwing-Duistermaat JJ van Houwelingen HC de Winter JP Estimation of individual genetic effects from binary observations on relatives applied to a family history of respiratory illnesses and chronic lung disease of newborns Biometrics 2000 56 808 814 10985220 10.1111/j.0006-341X.2000.00808.x
Olson JM Cordell HJ Ascertainment bias in the estimation of sibling genetic risk parameters Genet Epidemiol 2000 18 217 235 10723107 10.1002/(SICI)1098-2272(200003)18:3<217::AID-GEPI3>3.0.CO;2-8
Zou G Zhao H The estimation of sibling genetic risk parameters revisited Genet Epidemiol 2004 26 286 293 15095388 10.1002/gepi.10322
Houwing-Duistermaat JJ Van Houwelingen HC Terhell A Modelling the cause of dependency with application to filaria infection Stat Med 1998 17 2939 2954 9921611 10.1002/(SICI)1097-0258(19981230)17:24<2939::AID-SIM904>3.3.CO;2-L
Noordeen SK Hastings The epidemiology of leprosy Leprosy 1985 London: Churchill Livingstone 15 30
Risch N Assessing the role of HLA-linked and unlinked determinants of disease Am J Hum Genet 1987 40 1 14 3468804
Segal S Hill AV Genetic susceptibility to infectious disease Trends Microbiol 2003 11 445 448 13678861 10.1016/S0966-842X(03)00207-5
Smith DG The genetic hypothesis for susceptibility to lepromatous leprosy Hum Genet 1979 50 163 177 511131 10.1007/BF00390238
Bakker MI Hatta M Kwenang A Faber WR Van Beers SM Klatser PR Oskam L Population survey to determine risk factors for Mycobacterium leprae transmission and infection Int J Epidemiol 2004 33 1329 1336 15256520 10.1093/ije/dyh266
|
16307680
|
PMC1318483
|
CC BY
|
2021-01-04 16:03:33
|
no
|
BMC Med Genet. 2005 Nov 24; 6:40
|
utf-8
|
BMC Med Genet
| 2,005 |
10.1186/1471-2350-6-40
|
oa_comm
|
==== Front
BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central 1471-2458-5-1191628393210.1186/1471-2458-5-119DebateNeglected diseases of neglected populations: Thinking to reshape the determinants of health in Latin America and the Caribbean Ehrenberg John P [email protected] Steven K [email protected] Chief, Communicable Diseases Unit, Area of Disease Prevention and Control, Pan American Health Organization/World Health Organization (PAHO/WHO), 525 23rd Street NW, Washington, DC 20037, USA2 Regional Advisor, Communicable Diseases Unit, Area of Disease Prevention and Control, Pan American Health Organization/World Health Organization (PAHO/WHO), 525 23rd Street NW, Washington, DC 20037, USA2005 11 11 2005 5 119 119 24 11 2004 11 11 2005 Copyright ©2005 Ehrenberg and Ault; licensee BioMed Central Ltd.2005Ehrenberg and Ault; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Background
People living in poverty throughout the developing world are heavily burdened with neglected communicable diseases and often marginalized by the health sector. These diseases are currently referred to as Neglected Diseases of Neglected Populations. The neglected diseases create social and financial burdens to the individual, the family, the community, and the nation.
Discussion
Numerous studies of successful individual interventions to manage communicable disease determinants in various types of communities have been published, but few have applied multiple interventions in an integrated, coordinated manner. We have identified a series of successful interventions and developed three hypothetical scenarios where such interventions could be applied in an integrated, multi-disease, inter-programmatic, and/or inter-sectoral approach for prevention and control of neglected diseases in three different populations: a slum, an indigenous community, and a city with a mix of populations.
Summary
The objective of this paper is to identify new opportunities to address neglected diseases, improve community health and promote sustainable development in neglected populations by highlighting examples of key risk and protective factors for neglected diseases which can be managed and implemented through multi-disease-based, integrated, inter-programmatic, and/or inter-sectoral approaches. Based on a literature review, analysis and development of scenarios we visualize how multiple interventions could manage multiple disease problems and propose these as possible strategies to be tested. We seek to stimulate intra- and inter-sectoral dialogue which will help in the construction of new strategies for neglected diseases (particularly for the parasitic diseases) which could benefit the poor and marginalized based on the principle of sustainability and understanding of key determinants of health, and lead to the establishment of pilot projects and activities which can contribute to the achievement of the Millennium Development Goals.
==== Body
Background
People living in poverty throughout the developing world are heavily burdened from a series of communicable diseases, particularly parasitic diseases. They also tend to be marginalized by the health sector, as are many of the diseases that affect them. These diseases which are currently referred to as Neglected Diseases of Neglected Populations, pose a major challenge to the fulfillment of the Millennium Development Goals [1]. Some of these neglected diseases (NDs) of parasitic origin include lymphatic filariasis, soil-transmitted helminthiasis, schistosomiasis, onchocerciasis, leishmaniasis, African trypanosomiasis, Chagas disease, ectoparasitic skin infestations and parasitic zoonoses, among others [2].
These diseases are called neglected because they affect the poor, they are not subject to compulsory reporting in most countries, and are therefore not perceived as major public health burdens as compared to HIV/AIDS, tuberculosis, and malaria for instance. Most of them do not lead to epidemiologic emergencies, and consequently attract little attention from the media and the public sector. Furthermore, the private sector does not necessarily consider this group of diseases as a lucrative target, a phenomenon which severely hampers spending on research and development of specific drugs, vaccines and diagnostic tools [3]. However, new non-profit drug companies and private-public partnerships are beginning to address this gap at least for leishmaniasis, African trypanosomiasis and Chagas disease [4].
Burden of disease and vulnerable groups
The neglected parasitic diseases result in a high financial burden to the individual, the family, the community, the country and even the region – impairing its development [5]. For example, in the late 1970s, US$4.4 million were lost annually by people infected with Ascaris lumbricoides in Kenya, just in the form of unabsorbed food [6], and such costs can only rise as a population grows in the absence of specific prevention and control measures. A recent meta-analysis reassessing the disabilities created by schistosomiasis infections has concluded that this disease causes significantly more accumulated morbidity (anemia, chronic diarrhea and pain, undernutrition from protein loss, exercise intolerance, infertility, poor school performance) than previously thought, indicating it creates a very significant burden on the health of infected individuals [7]. Parasitic diseases, whether vector-borne, food-borne, water-borne or soil-transmitted tend to affect certain vulnerable groups, such as school-age children, the women of childbearing age or the breadwinners (male or female) in a household [8]. For example, poor working people with chronic lymphatic filariasis in Orissa, India lose an average of 68 workdays per year (19% of their work year), and spend an average of US$8.70 per year for treatment of their condition [9], much more than the per capita health expenditure by the national health services. Other vulnerable groups in society such as indigenous populations and minority ethnic groups, infants and pre-school children, the elderly, those with physical limitations, and immune-compromised people such as those with HIV/AIDS can be highly burdened with certain parasitic and other communicable diseases. Additional high-risk populations often include people living in slums [10], migrant workers (e.g., itinerant gold miners in Brazil [11]) and those living in agricultural labor camps or plantations (e.g., Guatemalan coffee pickers with onchocerciasis [12]).
Objective
The objective of this paper is to identify new opportunities to address NDs, improve community health and promote sustainable development in neglected populations by highlighting examples of key risk and protective factors of NDs which can be managed through multi-disease-based, integrated, inter-programmatic, and/or inter-sectoral approaches.
Many determinants of health lie outside the purview of the health sector. Furthermore, the policies of the sectors that exert influence on these negative health impacts are usually not established according to public health criteria (see Section A, below). Consequently, addressing comprehensive and sustainable solutions to these health problems cannot be the sole responsibility of the health sector. Soil-transmitted helminthiasis and schistosomiasis are good examples of the multi-sectoral and multi-factoral contextual determinants of NDs, where interventions via other sectors to improve water quality and quantity, provide safe excreta disposal, combined with periodic drug treatment and health education are the keys to sustainable control [13,14].
Partnerships with other sectors capable of effective action will be necessary. The reduction of NDs will ultimately contribute to the sustainable development of poverty stricken populations and may contribute to an increase in economic growth of the countries affected by these diseases.
Discussion
Rationale
Every day thousands of people living in poverty get sick and suffer or die of preventable, communicable diseases throughout the world. This accounts for the major difference in the magnitude of mortality and morbidity rates between developed and developing countries. Among the communicable diseases, the NDs such as lymphatic filariasis may be considered proxy indicators of the level of socioeconomic development [15], and communicable diseases are pervasive in countries or regions where gross national product is low or income distribution highly skewed [16]. Some of these diseases would cease to exist with an increase of the gross national product and a more balanced income distribution. However, this is a long-term solution to problems that demand immediate actions.
NDs inhibit the capacity of poor and neglected communities to achieve sustainable development, that is – the physical disability, stunting of child growth and intellect, and mortality created by NDs weakens the ability of both current and future generations to meet their basic human needs in a long-term manner, generate sufficient income, and may also constrain them from innovating and adopting ecologically sustainable practices at home and work.
This proposed strategy for the prevention and control of NDs in neglected populations is based on integrated, multi-disease, inter-programmatic and/or inter-sectoral approaches to manage multiple health risks and protective factors in the short and medium term, similar to that recently suggested by Molyneux and Nantulya [17] and others [18,19]. For example, in view of the American region's current demographics, provisions will need to be taken to tailor the approaches to both urban and rural populations (70% of the population in the Americas is urban). The expectation is that this new set of approaches can increase program sustainability of NDs control and elimination efforts. Furthermore, this set of approaches could strengthen existing health services and epidemiologic surveillance systems [20], as well as contribute to their integration into a multi-disease based surveillance and control system.
A. Multifactorial determinants of disease
In order to successfully develop an agenda for NDs in neglected populations, it is important to consider the contextual determinants of health. These are both intrinsic and extrinsic to human populations and their combination will determine the epidemiological pattern of these communicable diseases (disease spectrum).
Intrinsic determinants
Intrinsic determinants of disease are biological in nature (i.e., genetic makeup, immune response). Most of the intrinsic determinants can be manipulated only as a function of advances in medical research and technology (e.g., development of new vaccines, drugs and diagnostic tools). Significant progress has been made (private and academic sectors) in developing some of these tools, specifically those that target lucrative markets. However, their development for tropical diseases (e.g., the Tropical Diseases Research/World Health Organization drug research program and private sector initiatives) has been slow and it has been very difficult for neglected populations to access these tools given their high costs. This is clearly an issue of inequity in health services delivery to the poor that deserves further attention.
Extrinsic determinants
Extrinsic determinants of disease include poverty, vector ecology and behavior, and various human activities (sometimes combined with natural disasters) such as poorly planned agricultural and irrigation development, uncontrolled urbanization, and indiscriminant insecticide use, and improper self-treatment with medications. These are discussed below.
Poverty
Clearly, poverty is one of the most critical extrinsic determinants that impact the health of individuals and groups. It also increases the vulnerability to diseases by limiting their access to high quality health care, good housing and safe food [21]. It is also associated with social violence [22], drug addiction [23] and HIV/AIDS transmission [24]. Managing these determinants would require intense advocacy, improving living conditions, implementing health and environmental education, and social communication.
Vector ecology and behavior
Extrinsic determinants also include vector behavior, as well as characteristics of their habitat or environments [25]. The World Health Organization (WHO) has recently developed a strategic framework for Integrated Vector Management (IVM) to improve the control of malaria, dengue, and Chagas disease among other vector-borne diseases [26] with emphasis on interventions based on vector ecology and environmental determinants. Furthermore, some communities are currently experimenting with incorporating IVM into their on-going vector-borne disease control programs (e.g., in lymphatic filariasis elimination programs; and in malaria, leishmaniasis and schistosomiasis and Chagas disease control programs).
Human activities and the environment
Another group of extrinsic determinants of health include human activities and environmental determinants [27]. Human activities have an impact on the environment, and in doing so, they create conditions which have an impact on the epidemiological pattern of some communicable diseases including NDs [25]. They may also increase susceptibility to natural disasters (e.g., intense deforestation combined with heavy rains in Haiti led to disastrous mud slides in 2004). Some natural disasters such as flooding can lead to more breeding sites for disease vectors [28], and to an increased risk of disease transmission and outbreaks [29]. The indiscriminate use of insecticides in agriculture and public health interventions has led to resistance phenomena in some disease vector species [30]. The indiscriminate use of drugs and the pervasive practice of self-medication in developing countries have contributed to the widespread occurrence of drug resistance in some parasite populations, as is the case of malaria in parts of Africa [31] and Southeast Asia [32].
Combined health determinants and their health outcomes
Extrinsic and intrinsic determinants of communicable disease transmission will often synergize in a negative way when clustered together. Deficient diets leading to immune deficiencies and lack of nutrients in high-risk (neglected) populations will lead to malnutrition and to an increase in the susceptibility to human pathogens. Severe hunger in poor children is also associated with chronic illnesses, behavioral problems (high anxiety and stress) [33] and learning disabilities [34]. Lack of access to health services will result in the deterioration of a person's health status, thus further hampering his or her productivity as a member of the work force and within the family (added burden to the family). Malnutrition [35], diarrhea, anemia and other complications of soil-transmitted helminth infections will often lead to growth stunting [36], school absenteeism [37] and also affect a child's ability to learn [38], reducing his or her chances of a better-paid and safer job later in life. Untreated parasitic skin infestations or infections (e.g., tungiasis and scabies often with secondary bacterial infection; and cutaneous larva migrans) [39] can lead to school absenteeism and lost work days. Poverty, poor housing, high population densities and unsafe or inadequate living conditions, combined with environmental conditions favoring vector breeding will readily promote the spread of some communicable diseases and trigger outbreaks in poor communities [40].
Many NDs such as soil-transmitted helminthiasis, schistosomiasis, or trachoma tend to cluster, both geographically and socially, in poor communities, neighborhoods and families, as do other communicable diseases such as malaria. Furthermore, only a small number of persons in a community will have high intensity infections (i.e., soil-transmitted helminth and schistosome infections tend to aggregate, that is, worm burden is concentrated in a small proportion of individuals in any community) [41,42]. In principle, knowledge of these phenomena can facilitate their prevention and control at the community level.
B. Neglected populations and the need for an inter-sectoral approach
Beyond the health sector
Many determinants, especially the environmental determinants of disease, injury and death in developing countries lie outside the purview of the health sector. These determinants include poor living conditions such as unsafe drinking water, inadequate sanitation and excreta disposal, poor drainage, inadequate solid waste removal, poor housing, and indoor air pollution. The policies of the sectors that exert influence on these negative health impacts are usually not established according to public health criteria. Consequently, addressing comprehensive and sustainable solutions to these health problems cannot be the sole responsibility of the health sector.
Reducing risk factors needs to go hand in hand with the adoption of protective factors, including better access to services (health, environment, and education) and employment opportunities, backed by political commitment to guarantee their sustainability.
This situation calls for "thinking outside the box" (beyond the health sector) by incorporating an inter-sectoral approach [13,14,19,43], one that addresses the multiplicity of risk and protective factors and proposes strategies relying on synergies with other public health interventions (e.g., school deworming and nutritional programs), inter-programmatic synergies (e.g., IVM), articulations with sustainable development-based programs (e.g., schistosomiasis control and aquaculture micro-enterprise), and/or partnerships encompassing a wider set of stakeholders (e.g., Global Fund to Fight AIDS, Tuberculosis, and Malaria; Global Environmental Facility) [19].
Inter-sectoral approaches have been applied down to the community level for the control of infectious diseases and nutritional problems [44]. Furthermore, there are examples throughout the Americas that show that integrated, inter-sectoral work is being successfully conducted at the community level. For example, integrated risk factor analysis combining parasitological, environmental, social and economic risk factors in various communities to select foci for lymphatic filariasis mass treatment is used in the favelas (highly-impoverished urban communities) of Sanitary District II of the Lymphatic Filariasis Control Program called "Xo Filariase! [Shoo, Filariasis!] " in Recife, Brazil (personal communications, SA with T. Maciel Lyra, SMS/Recife, 2003–2004). Increasing the knowledge of protective and risk determinants and how they interact with each other in communities particularly exposed to NDs will help us tailor the interventions and find common ground with other sectors in order to achieve control or elimination of these infections.
Justification
A. Why pursue integrated, multi-disease, inter-programmatic, and/or inter-sectoral, approaches for neglected diseases control and elimination?
1 Integrated, multi-disease, inter-programmatic, and/or inter-sectoral approaches give added value to disease control and elimination interventions. For example, the use of albendazole in lymphatic filariasis-endemic areas provides the added value of controlling soil-transmitted helminths [45]. Promoting the use of insecticide-treated bed nets or curtains helps interrupt the transmission of malaria and other vector-borne diseases in lymphatic filariasis foci [46]. Ivermectin is used to treat lymphatic filariasis and onchocerciasis. It has additional benefits against soil-transmitted helminths and head lice [47].
2 Integrated, multi-disease interventions are cost-effective by articulating one disease control intervention into another one, for example soil-transmitted helminthiasis and schistosomiasis control using combined therapies with albendazole and praziquantel [38]; or soil-transmitted helminth control combined with lymphatic filariasis elimination using albendazole plus diethylcarbamazine (DEC) [48].
3 We can spread several benefits to the community with the same intervention; an initiative that would surely be welcome by the community. For example, improved housing can protect against contact with some Chagas disease vectors and simultaneously reduce the risk of developing acute respiratory infections (ARI), thus improving the family's quality of life and perhaps even the market value of the dwelling.
4 Inter-sectoral interventions have positive impacts on family health and economic security, environmental sanitation, and even income generation, all of which are important to families and the community at large. Such interventions, when targeted to the more vulnerable or neglected groups, also assist in reducing health inequalities, an important new issue for many health agencies.
5 This proposed strategy also supports the UN Millennium Development Goals (MDGs) including 10 out of the 18 Millennium Declaration targets. Deworming cost-effectively improves nutritional status of poor children and communities, contributing to the goal of Eradication of Hunger (MDG-1) [49]. Deworming can improve school attendance [37,50] and thus increases the chances of completing primary education successfully (MDG-2, Primary Education). Promoting income-generating activities (e.g. micro-enterprises [51] for poor women) and educating impoverished mothers in child care contribute to the Empowerment of Women (MDG-3). Reducing the burden of parasitic diseases contributes to the Reduction of Child Mortality (MDG-4). Controlling iron deficiency and anemia due to hookworm [52] results in Improvement of Maternal Health (MDG-5). Combat of NDs [53] such as the intestinal helminth infections, leishmaniasis, parasitic skin diseases, and Chagas disease contributes to the goal of Combat HIV, Malaria and other Diseases (MDG-6). Implementing environmental sanitation reduces fecal contamination of groundwater and surface waters, and thus contributes to Ensuring Environmental Sustainability (MDG-7). An inter-sectoral approach to NDs prevention and control with a sustainable development focus involves establishing extended partnerships, compatible with the goal of Global Partnerships for Development (MDG-8) [54].
B.What opportunities and entry points are there for integrated, multi-disease, inter-programmatic, and/or inter-sectoral approaches?
The strategy proposed here focuses on a set of micro-level interventions (community, family, individual) to manage selected extrinsic determinants of NDs through integrated, multi-disease, inter-programmatic, and/or inter-sectoral approaches including improved health services, environmental sanitation and improved housing, better access to foods and key micronutrients, educational access for children and women, community participation and micro-enterprise development (e.g., urban and rural household-level food production). However, there is no universal recipe or protocol to manage these determinants. A specific strategy will have to be tailored to the local conditions, partners and resources available in each community or area. General examples of the opportunities and entry points to manage the determinants are elaborated below. Macro-level (policy) interventions are beyond the scope of this analysis and are dealt with elsewhere.
Health services
Management of morbidity due to leprosy is well integrated in health services in most leprosy-endemic countries, and could easily accommodate a lymphatic filariasis morbidity component. Combined mass drug administration (MDA) for lymphatic filariasis elimination and soil-transmitted helminth control is recommended by WHO. Soil-transmitted helminth and schistosomiasis control may be articulated with integrated management of child and adolescent health and development strategies and their syndromic approach to disease control. Where reduction of maternal mortality and improvement of women's health is the core of a country's strategic health plan hookworm control can be an integral part of such a plan. Combined MDA for schistosomiasis and soil-transmitted helminth control has been recommended by WHO for school-age children at high risk in areas of their geographic overlap [55]. Soil-transmitted helminthiasis and schistosomiasis control can, in principle, be combined with lymphatic filariasis elimination [38] in areas where the three disease groups overlap as in some coastal areas of NE Brazil.
Public health interventions can also be integrated along other sectors such as:
Environment (environmental sanitation, environmental quality & environmental management)
Lack of potable water supply, sanitation (especially excreta disposal), lack of solid waste collection and disposal and household cleanliness, and lack of animal control have all been identified as important risk factors in the transmission of diarrheal diseases, intestinal parasites, and skin diseases [56], and lack of household and yard tidiness with the presence of dengue vectors [57]. Multi-sectoral interventions in these areas have resulted in major reductions in the incidence and prevalence of some NDs and mortality in infants and young children in poor communities [58]. A recent systematic review and meta-analysis presents the positive impacts of water supply and hygiene interventions in reducing diarrhea in less-developed countries [59]. Examples of innovative interventions which address the disease determinants include home-based water treatment systems with filtration, flocculation and safe sealed storage containers [60]; simplified or condominial sewerage for shantytowns and other areas of urban poverty [61] and ecological disposal of excreta by the separation of urine and feces [62]; promotion of manual sanitary landfills; improved household and neighborhood drainage systems [63]; education to promote household hygiene, cleanliness and tidiness [64]; and animal corralling [65]. Improvements in rural housing (floors, ceilings, walls, and windows) have reduced the transmission of Chagas disease [66], while the use of mesh screens (for eaves, windows and doors) and sealing eaves have reduced human exposure to malaria vectors [67]. Domestic fly and mosquito control can be implemented in the household by use of deltamethrine-impregnated curtains [68]. Urban drainage improvements and improved household water storage systems can reduce urban malaria [69]. Improved wood stoves (biofuel stoves), substitution of biofuels with gas or other improved rural fuel/energy sources, and improved kitchen ventilation all reduce exposure of women and young children to harmful indoor air pollutants arising from exposure to smoke from traditional woodstoves and open fires [70]. Solar stoves, solar water disinfection and solar energy panels are possible alternatives to traditional wood stoves, boiling water with biofuels and burning kerosene lamps in the home, though their higher costs must be addressed [71]. Encouraging urban reforestation with local fast-growing trees and planting bamboo on the steep slopes where shanty towns are located will reduce soil erosion and vulnerability to small landslides [72]. The presence or absence of several of the interventions mentioned here is proposed as useful indicators of exposure to environmental health risks in developing countries [73].
3 Education and School Health
School deworming programs (health education and treatment interventions) have been successfully incorporated into school health programs [74] and can be part of Healthy Schools initiatives. They can also be part of other community-based sustainable development initiatives which, for example, combine improved water supply, safe excreta disposal and hygiene education interventions for the school children [75]. The FRESH initiative (Focusing Resources on Effective School Health) is a model of an integrated approach to improving the health of school children; it focuses on four key components – health-related school policies, provision of safe water and sanitation, skills-based health education, and school-based health and nutrition services (including deworming and addressing micronutrient deficiencies) [76]. Lack of hygienic behavior by children clearly increases the risk of diarrheal diseases [77] and soil-transmitted helminthiasis [78]. Programs which educate women and children about the importance of proper handwashing with soap or ash and clean water can reduce diarrheal diseases incidence, acute lower respiratory infections, impetigo and soil-transmitted helminth infections in children of poor communities [59,79,80]. Nutrition education for women focused on basic nutritional concepts has a positive impact on the nutrition of their own children as well as the children of her neighbors [81]. Deworming programs have even been combined with family planning education programs in the Philippines [82].
4 Nutrition and Food Security
Improved nutritional status is one necessary element of food security for the individual and the family. Treatment of intestinal worms enhances the value of food and micronutrient delivery programs [83]; anti-helminth treatment helps control iron deficiency anemia and its sequelae (including cognitive impairment). Deworming activities can be combined with vitamin A supplementation and the trachoma SAFE intervention (Surgery, Antibiotic Therapy, Facial Cleanliness and Environmental Improvement). Addressing key micronutrient deficits [84] (e.g., zinc deficiency which is causally associated with diarrhea, pneumonia and malaria in children under age 5 [85]) can be done by adding micronutrients to key foods in the local diet or to condiments such as table salt [86]. In areas endemic for lymphatic filariasis, DEC is added to table salt for mass treatment of the disease and can eliminate transmission within one to two years. In principle, DEC-salt can be combined with iodine and fluoride; Haiti has conducted a successful pilot project fortifying salt with DEC and iodine. Nutrition, deworming and family planning efforts have been integrated in a project in Sri Lanka [87]. Where hunger and undernutrition are addressed by supplemental feeding, school feeding programs, and other nutritional interventions in poor communities [88], some pilot projects have successfully combined deworming and supplemental nutrition in Africa, Asia and the Americas through partnerships involving non-governmental organizations (NGOs) and international agencies such as WHO, UNICEF and the World Food Program [89,90].
5 Economic Development
Cooperatives can supply and manage insecticide-treated bednets or curtains [91] to the community and generate income. For example, in Leogane, Haiti, the NGO "KOLEMO" (Komite de Leyogan Pou Moustik) has developed a successful microlending program for local seamstresses affected by lymphatic filariasis to stitch mosquito bednets, which are then treated with the pyrethroid deltramethrin by the NGO and sold by the seamstresses to the community for a small profit which pays their labor. Use of the bednets reduces the number of bites by local mosquitoes that transmit lymphatic filariasis and malaria. Another cooperative program in Leogane, Haiti has provided nutritional rehabilitation of malnourished children [92], and improved employment and income distribution through innovative interventions such as local-level micro-financing and micro-credit schemes; these have also worked in Bangladesh and elsewhere [93]. Sustainable rural development projects (as promoted by the Heifer Project International, World Resources Institute, World Wildlife Fund, Nature Conservancy, International Union for the Conservation of Nature, etc.) which have integrated "better life programs" [94] and micro-credit projects [95] provide opportunities for rural and urban income generation and can also provide windows of opportunity for education in the prevention and control of NDs.
6 Urban Improvement and Renewal
Well-planned urban improvement and renewal projects can improve community health and safety [96,97] (e.g., improved housing and street lighting; reducing environmental lead exposure [98]; more parks and green areas for recreation [99], safe bikeways and pedestrian paths [100]) and other integrated urban planning approaches such as integrated water basin development in Indonesia [101]. Integrated planning of urban improvement programs where the public health agency participates fully with other city agencies (public works, environment, and social services) [102] have proven successful in Cuba (e.g., Movimiento de Municipios Productivos [103]). Local public health authorities can also promote the use of a new public health tool, health impact assessment (HIA) [104], for urban development and urban renewal projects [105], thus helping prevent disease problems.
7 Agriculture (Family and Community Agriculture, Horticulture and Agroforestry; Small Animal Husbandry)
Home gardens, community gardens and home-based small animal husbandry [106] and aquaculture [107] may be promoted to address income generation, nutrition and health needs simultaneously [108]. At home families can produce fruits, vegetables, nuts and medicinal plants in home gardens and even in small containers. Family food production also can help address the serious risks of food insecurity in poor communities though further study is needed [109]. Successful urban agriculture models exist in East Africa [110] and elsewhere. Local neighborhood cooperatives can supply safe compost for use in urban homes and community gardens. It can also be sold. Planting selected fruit trees at home can help augment the dietary intake of vitamin A [108], vitamin C and perhaps other micronutrients hopefully without need for pesticide application. As an example, the hearty shrub acerola or Barbados cherry (Malphighia glabra) is widely grown in home gardens in the Caribbean, Mexico, and Central and South America and contains high levels of vitamin C; it is propagated readily from seeds or cuttings. Selected local fruit and nut trees can also be planted in public places with free access as seen with avocado and jack fruit trees in Brasilia, Brazil and walnut trees in Yolo County, California. Planting fruit and nut trees could be targeted to areas in or near highly-impoverished communities. Small-scale animal husbandry (urban or rural) can increase women's and family income [111]. Small-scale animal husbandry is also an opportunity for veterinarians, public health specialists, educators and community members to work together to educate poor communities about preventing and controlling local parasitic and zoonotic diseases (e.g., taeniasis, toxocariasis, toxoplasmosis, tungiasis, rabies and leishmaniasis) through animal health care and vaccination, animal control and/or corralling.
8 Primary Environmental Care (PEC)
Primary Environmental Care is an integrated and participatory approach to analyzing and solving community environmental health problems which was developed in the 1990s by UNICEF and PAHO/WHO [112]. The PEC approach is based on cooperation and partnerships between different stakeholders to identify key problems and develop sustainable solutions, and relies on active community participation [113]. For example, PEC is a tool that can in principle promote safer and more hygienic local environments and community responsibility, such as the corralling and control of pigs [114], poultry [65] and livestock and stray dogs [115] in the community. PEC could also help address problems such as family and neighborhood exposure to indoor air pollutants from biofuel use that contribute to the risk of ARI especially in women who work all day at home and house-bound young children [73,116].
9 Promotion of Tourism
Tourism, both domestic and international, is a key part of the economy of many countries (e.g., Haiti's economy once depended strongly on tourism; Honduras' economy today continues to rely on tourism), cities and towns. However, the concerns of the tourist sector about communicable diseases and food safety [117] can stifle investment and growth of this sector. Integrated public health interventions focusing on improving environmental and sanitary conditions and hygiene behaviors in public markets and tourist hotels could help generate more tourism and income while reducing the occurrence of certain communicable and NDs (e.g., infestations of rodents associated with leptospirosis transmission, and insect vectors associated with dengue, leishmaniasis, malaria or lymphatic filariasis transmission). Food safety programs for tourist restaurants, hotels, street vendors and public markets are being promoted in some countries to address food-borne parasitic and microbial infections and travelers' diarrhea [118].
Three scenarios
We have developed three hypothetical scenarios to illustrate how an integrated, multi-disease, inter-programmatic, and/or inter-sectoral approach could be applied in three very different populations (an impoverished periurban community, an indigenous community, and a medium-sized city with a mix of populations).
The city of Jaboatão dos Guararapes on the coast of Pernambuco state, Brazil has a favela of about 40,000 residents including many children, mostly poor and undernourished migrants from the dry interior of the state seeking employment in the urban area (i.e., environmental and economic refugees). The favela, located beside a large shallow lake (Lago do Naútico) receiving untreated sewage from two cities, is endemic for lymphatic filariasis, schistosomaisis (Schistosoma mansoni) and soil- transmitted helminthiasis among other NDs and experiences domestic fly infestations. The favela has no potable water system or sewerage, no solid waste collection, and few latrines. Drinking water is trucked in and sold by private vendors. Shallow surface drainage canals cross the community, and are choked with weeds and trash and harbor the intermediate host snail of S. mansoni. The lake, used for fishing, recreation and washing clothes and dishes, floods the community periodically thus spreading excreta widely and it also harbors the snail intermediate hosts of S. mansoni.
In a multi-staged approach to prevention and control in this favela, the most critical interventions to improve the health of this community with respect to these NDs would be regular chemotherapy using a benzimidazole drug (albendazole or mebendazole) for soil-transmitted helminth infections and praziquantel against S. mansoni infections. If the community prevalence or individual worm burdens are high enough, MDA would be indicated to reduce worm burdens and accompanying morbidity in children, adolescents and adults. Community health education and appropriate social marketing of targeted health messages would be part of these first-stage interventions. In a second stage of interventions (medium- and long-term), the community would reap significant health benefits from a series of infrastructure interventions especially provision of safe excreta disposal and safe community drinking water supplies, improved surface-water drainage to free blocked canals, frequent solid waste collection and secure disposal, each accompanied by health and environmental education and social marketing. To aid in the sustainability of these interventions through community participation, improved opportunities for local employment could be encouraged through training and organization of the existing trash scavengers who already operate their own waste-separation (recycling) sites. Primary environmental care may play a useful role in promoting waste management and alleviating bad surface drainage, through community-based actions. Women's cooperatives and micro-credit programs could be promoted, with a health focus or with health improvement as an expected outcome. Community small-scale horticulture (home gardens) and animal husbandry (e.g., poultry) could help address chronic undernutrition especially in the community's children. In summary, the management actions to control the major extrinsic determinants of disease in this impoverished community in Jaboatão dos Guararapes must be undertaken not only by the health sector but other sectors as well [13,14].
The Yanomani communities of northern Brazil (states of Roraima and Amazonas) are a migratory indigenous population, who live in an onchocerciasis-endemic area and number about 10,000 people. Several families (often 30–100 people) and their dogs usually live in large communal houses or lodges called malocas without interior walls. Malnourished young children are not uncommon, while trachoma and tuberculosis are reported from the Yanomami. The communities also endure high burdens of soil-transmitted helminths and the ectoparasitic flea Tunga penetrans which causes severe disability of hands and feet. Tungiasis is also present in their dogs. An on-going indigenous health care intervention, in which MDA with ivermectin is provided to the Yanomami communities twice a year for onchocerciasis elimination, is expected to have a beneficial effect by also reducing the soil-transmitted helminths in children and adolescents, though this has not been measured as yet. Topical ivermectin application can be used to reduce the lesions caused by T. penetrans fleas [119]. Ivermectin (in a different dosage) is also used to treat tungiasis in dogs, and could be delivered to their pets during the twice-yearly mass treatment rounds for onchocerciasis elimination. Micronutrient supplements such as vitamin A and trachoma screening with antibiotic treatment could be provided to the most vulnerable groups, typically children and adolescent and pregnant women, during the course of the ivermectin mass treatments to eliminate onchocerciasis. Tuberculosis and leprosy screening could also be carried out during MDA. The communities seldom have a safe water source or safe excreta disposal systems. Water is usually taken directly from streams and open defecation is the norm. If accepted by the communities, health education on hygiene behavior (especially hand hygiene), simple home-based water treatment systems (discussed above) and/or improved environmental sanitation (e.g., mixing feces with wood-fire ash or lime and shallow burial, or perhaps dual-chamber ventilated improved pit (VIP) latrines made of local materials) could reduce the worm burden. Improved safe water with hand hygiene could reduce trachoma, though any increase in water use inside the maloca would need to be accompanied by wastewater management. Smoke from wood fires used in the malocas probably contributes to the burden of upper respiratory diseases seen in the Yanomami communities, especially children. Pilot projects for improved housing ventilation and perhaps the introduction of efficient biofuel ovens/stoves could be of interest to the communities, and incorporated into their health or environmental education programs. For communities which return to their villages each year, there may be a role for introducing the cultivation of perennial fruit trees if not home gardens.
In a perhaps typical example of the complexity of urban communicable diseases and NDs ecology, the city of Imperatriz in western Maranhão state, NE Brazil has a diverse population of over 230,000 residents including impoverished areas, squatter settlements, and river traders and fishing communities, and is located along the banks of a large river (River Tocantins) which brings annual flooding of parts of the city. The city experiences transmission of malaria along the riverside, and leishmaniasis, leptospirosis and dengue cases in the city, among other communicable diseases. Figure 1 is a model which illustrates sets of possible vector-borne disease interventions for cities (involving several municipal services sectors) to manage certain determinants of these diseases simultaneously and synergistically. This model of specific environmental interventions uses a simple systems diagram, which could be adapted for other specific communicable disease and NDs control contexts. The sets of interventions are shown in green with arrows from each intervention leading to two or more parasitic or vector-borne diseases (in red), indicating that the set of interventions can in principle be used in an integrated manner against several diseases (or their vectors) in the same geographic areas. (Not every intervention in the box can be used against every disease or vector linked by the arrows; similar interventions are clumped together to make the model more illustrative.) The model's true application in the field will depend on the array of local vectors, diseases, environmental conditions, health and environmental services, and epidemiologically-important human behaviors in specific populations.
Figure 1 Multi-disease interventions for the prevention and control of vectors in urban areas – conceptual model.
In reading the illustrative model, for example, the interventions by the sector responsible for improving municipal drinking water services (in terms of quality, quantity, frequency of availability, cost, coverage and/or control of leaks) can be protective against dengue vectors, urban and periurban schistosomiasis, and even domestic rodents and flies which often prefer moist environments (such as those created by leaking water pipes and sewerage pipes). In another example, the municipal sector responsible for solid waste management can improve solid waste collection and sanitary disposal services; promote recycling and waste reduction, and collection and control of special wastes (discarded tires etc.) which can have an impact on the vectors of dengue (when breeding in discarded tires or other discarded water-holding wastes), leptospirosis and domestic fly plagues (by timely removal and disposal of discarded foods in garbage which act as food sources for domestic rodents and flies), and lymphatic filariasis (where uncollected solid waste on the streets blocks and pools urban surface drainage thus creating better conditions for Culex quinquefasciatus vector breeding).
Summary
• Neglected populations living in poverty throughout the developing world tend to be extremely burdened with a series of neglected communicable diseases and often marginalized by the health sector.
• In order to successfully strengthen the public health agenda for NDs in neglected populations, it is important to consider the contextual determinants of health. These are both intrinsic and extrinsic to human populations. The combination of both determines the epidemiological pattern of communicable diseases (disease spectrum) at the local level and help identify appropriate interventions.
• Integrated, multi-disease, inter-programmatic, and/or inter-sectoral approaches give added value to disease control and elimination interventions, as have been demonstrated in various countries.
• We have argued for a new way to think about addressing the problems of NDs in neglected populations, using integrated, multi-disease, inter-programmatic, and/or inter-sectoral approaches. Numerous examples of specific successful interventions to manage disease determinants in various types of communities have been presented, but few studies exist where such measures were applied in an integrated and coordinated manner.
• Three hypothetical scenarios have been developed here to illustrate how multiple interventions could manage multiple disease problems through integrated, multi-disease, inter-programmatic, and/or inter-sectoral approaches in three very different populations (an impoverished periurban favela, an indigenous community, and a medium-sized city with a mix of populations and diseases).
• It is our aim that this review and analysis will stimulate intra- and inter-sectoral dialogue which will help in the construction of new NDs strategies (particularly for the parasitic diseases) which could benefit the poor and marginalized based on the principle of sustainability and understanding of some key determinants of health, and lead to the establishment of pilot projects and activities which can contribute to the MDGs.
List of Abbreviations
ARI. Acute respiratory infections.
BTI. Bacillus thuringiensis ssp. israelensis
DEC. Diethylcarbamazine.
FRESH. Focusing Resources on Effective School Health
HIA. Health impact assessment.
HIV/AIDS. Human Immunodeficiency Virus/Acquired Immunodeficiency Syndrome.
IVM. Integrated Vector Management.
MDA. Mass drug administration.
MDGs. Millennium Development Goals.
NDs. Neglected diseases.
NGO. Non-governmental Organization.
PAHO. Pan American Health Organization.
PEC. Primary Environmental Care.
SAFE. Surgery, Antibiotics, Face-washing and Environmental improvement.
UNICEF. United Nations International Children's Emergency Fund.
VIP. Ventilated improved pit
WHO. World Health Organization.
Competing interests
The author(s) declare that though they are employees of the Pan American Health Organization and the World Health Organization, the contents of this paper are the sole responsibility of its authors and should not be construed as speaking for the policies of the Pan American Health Organization and the World Health Organization. This paper is a tool towards further scientific dialogue and discussion among public health professionals.
Authors' contributions
JE conceived the idea of the paper, wrote the first draft, and is the principle conceptual author. SA was responsible for additional concepts, for gathering the majority of the supporting evidence, examples, and for creating the three scenarios. SA developed Figure 1. Both authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2458/5/119/prepub
Acknowledgements
We thank Dr. Stephen Corber, DPC/PAHO/WHO, and Dr. Mariane Claeson of the World Bank for helpful comments and suggestions. We thank reviewer Dr. Juerg Utzinger, Swiss Tropical Institute, for additional important suggestions particularly the section on the three scenarios.
==== Refs
The Millennium Development Goals for Health – Rising to the Challenges. January The World Bank 2004 http://www-wds.worldbank.org/servlet/WDSContentServer/WDSP/IB/2004/07/15/000009486_20040715130626/Rendered/PDF/296730PAPER0Mi1ent0goals0for0health.pdf
World Health Organization, International Workshop, Intensified control of neglected diseases, Summary Report, 10–12 December 2003 http://whqlibdoc.who.int/hq/2004/WHO_CDS_CPE_CEE_2004.45.pdf
Trouiller P Olliaro P Torreele E Orbinski J Laing R Ford N Drug development for neglected diseases: a deficient market and a public-health policy failure Lancet 2002 359 2188 2194 10.1016/S0140-6736(02)09096-7 12090998
Mandelbaum-Schmid J In focus: New generation of non-profit initiatives tackles world's "neglected" diseases Bull World Health Organ 2004 82 395 396 15298231
Savioli L Stansfield S Bundy DAP Mitchell A Bhatia R Engels D Montresor A Neira M Shein AM Schistosomiasis and soil-transmitted helminth infections: forging control efforts Trans R Soc Trop Med Hyg 2002 96 577 579 10.1016/S0035-9203(02)90316-0 12625126
Stephenson LS Methods to evaluate nutritional and economic implications of Ascaris infections Soc Sci Med 1984 19 1061 1065 10.1016/0277-9536(84)90309-5 6441259
King CH Dickman K Tisch DJ Reassessment of the cost of chronic helmintic infection: a meta-analysis of disability-related outcomes in endemic schistosomiasis Lancet 2005 365 1561 1569 10.1016/S0140-6736(05)66457-4 15866310
Ramaiah KD Radhamani MP John KR Evans DB Guyatt H Joseph A Datta M Vanamail P The impact of lymphatic filariasis on labour inputs in southern India: results of a multi-site study Ann Trop Med Parasitol 2000 94 353 364 10945045
Babu BV Nayak AN Dhal K Acharya AS Jangid PK Mallick G The economic loss due to treatment costs and work loss to individuals with chronic lymphatic filariasis in rural communities of Orissa, India Acta Trop 2002 82 31 38 10.1016/S0001-706X(02)00030-X 11904101
Ferrari AA Solymos GM Castillo RM Sigulem DM Risk factors for protein-energy malnutrition in pre-school shantytown children of São Paulo, Brazil Rev Paulista Med 1998 116 1654 1660
Vosti SA Malaria among gold miners in southern Para, Brazil: estimates of determinants and individual costs Soc Sci Med 1990 30 1097 1105 10.1016/0277-9536(90)90296-5 2363060
Fazen LE Anderson RI Marroquin HF Arthes FG Buck AA Onchocerciasis in Guatemala. I. Epidemiological studies of microfilaria Am J Trop Med Hyg 1975 24 52 57 1111355
Hunter JM Rey L Chu KY Adekolu-John E Mott KE Parasitic Diseases in Water Resources Development: The Need for Intersectoral Negotiation 1993 Geneva: World Health Organization
Utzinger J Bergquist R Xiao SH Singer BH Tanner M Sustainable schistosomiasis control--the way forward Lancet 2003 362 1932 1934 10.1016/S0140-6736(03)14968-9 14667754
Durrheim DN Wynd S Liese B Gyapong JO Lymphatic filariasis endemicity – an indicator of poverty? Trop Med Int Health 2004 9 843 845 10.1111/j.1365-3156.2004.01287.x 15303987
Ackers ML Quick RE Drasbek CJ Hutwagner L Tauxe RV Are there national risk factors for epidemic cholera? The correlation between socioeconomic and demographic indices and cholera incidence in Latin America Int J Epidemiol 1998 27 330 334 10.1093/ije/27.2.330 9602419
Molyneux DH Nantulya VM Linking disease control programmes in rural Africa: a pro-poor strategy to reach Abuja targets and millennium development goals BMJ 2004 328 1129 1132 10.1136/bmj.328.7448.1129 15130985
Molyneux DH "Neglected" diseases but unrecognised successes--challenges and opportunities for infectious disease control Lancet 2004 364 380 383 10.1016/S0140-6736(04)16728-7 15276399
Editors Thinking beyond deworming Lancet 2004 364 1993 1994 10.1016/S0140-6736(04)17521-1 15582039
Task Force on Health Systems Research Informed choices for attaining the Millennium Development Goals: towards an international cooperative agenda for health-systems research Lancet 2004 364 997 1003 10.1016/S0140-6736(04)17026-8 15364193
Larrea C Freire W Social inequality and child malnutrition in four Andean countries Rev Panam Salud Pública 2002 11 356 364 12162832
Macedo AC Paim JS Silva LM Costa Md Mda C Violence and social inequalities: mortality rates due to homicides and life conditions in Salvador, Brazil Rev Saúde Publica 2001 35 515 522 11799464
Marzuk PM Tardiff K Leon AC Hirsch CS Stajic M Portera L Hartwell N Poverty and fatal accidental drug overdoses of cocaine and opiates in New York City: an ecological study Am J Drug Alcohol Abuse 1997 23 221 228 9143635
Holtgrave DR Crosby RA Social capital, poverty, and income inequality as predictors of gonorrhoea, syphilis, chlamydia and AIDS case rates in the United States Sex Transm Infect 2003 79 62 64 10.1136/sti.79.1.62 12576618
Wilson ML Aron JL, Patz JA Ecology and infectious disease Ecosystem Change and Public Health: A Global Perspective 2001 Baltimore: Johns Hopkins University Press 283 324
World Health Organization WHO Global strategic framework for integrated vector management. WHO/CDS/CPE/PVC/2004.10 2004 WHO, Geneva
Ezzati M Lopez AD Rodgers A Vander Hoorn S Murray CJ Comparative Risk Assessment Collaborating Group Selected major risk factors and global and regional burden of disease Lancet 2002 360 1347 1360 10.1016/S0140-6736(02)11403-6 12423980
Saenz R Bissell RA Paniagua F Post-disaster malaria in Costa Rica Prehospital Disaster Med 1995 10 154 160 10155423
Sur D Dutta P Nair GB Bhattacharya SK Severe cholera outbreak following floods in a northern district of West Bengal Indian J Med Res 2000 112 178 182 12452126
Mouchet J Brengues J Agriculture-health interface in the field of epidemiology of vector-borne diseases and the control of vectors Bull Soc Pathol Exot 1990 83 376 393 2208469
Ezedinachi EN Ejezie GC Emeribe AO Problems of chloroquine-resistant P. falciparum in Nigeria: one antimalaria drug's utilization in metropolitan Calabar Cent Afr J Med 1991 37 16 20 2060003
D'Alessandro U Buttiens H History and importance of antimalarial drug resistance Trop Med Int Health 2001 6 845 848 10.1046/j.1365-3156.2001.00819.x 11703837
Weinreb L Wehler C Perloff J Scott R Hosmer D Sagor L Gundersen C Hunger: its impact on children's health and mental health Pediatrics 2002 110 41 10.1542/peds.110.4.e41
Nokes C Bundy DAP Does helminth infection affect mental processing and educational achievement? Parasitol Today 1994 10 14 18 10.1016/0169-4758(94)90348-4 15275558
Lunn RG Northrop-Clewes CA The impact of gastrointestinal parasites on protein-energy malnutrition in man Proc Nutr Soc 1993 52 101 111 10.1079/PNS19930042 8493254
Moore SR Lima AA Conaway MR Schorling JB Soares AM Guerrant RL Early childhood diarrhoea and helminthiases associate with long-term linear growth faltering Int J Epidemiol 2001 30 1457 1464 10.1093/ije/30.6.1457 11821364
Miguel E Kremer M Worms: Identifying impacts on education and health in the presence of treatment externalities Econometrica 2004 72 159 217 10.1111/j.1468-0262.2004.00481.x
WHO Prevention and control of schistosomiasis and soil-transmitted helminthiasis: report of a WHO expert committee WHO Tech Rep Ser No. 912 2002 World Health Organization, Geneva
Heukelbach J van Haeff E Rump B Wilcke T Moura RC Feldmeier H Parasitic skin diseases: health care-seeking in a slum in north-east Brazil Trop Med Int Health 2003 8 368 373 10.1046/j.1365-3156.2003.01038.x 12667157
Thakur JS Swami HM Dutt R Mehta M Gupta V Epidemiological investigation of cholera outbreak in a periurban slum colony in Chandigarh Indian J Med Sci 2001 55 429 433 12026505
Forrester JE Scott ME Bundy DAP Golden MHN Clustering of Ascaris lumbricoides and Trichuris trichiurainfections within households Trans R Soc Trop Med Hyg 1988 82 282 288 10.1016/0035-9203(88)90448-8 3188157
Bradley DJ May RM Consequences of helminth aggregation for the dynamics of schistosomiasis Trans R Soc Trop Med Hyg 1978 72 262 273 10.1016/0035-9203(78)90205-5 675786
Evans AC Stephenson LS Not by drugs alone: the fight against parasitic helminths World Health Forum 1995 16 258 261 7546164
Webb K Hawe P Noort M Collaborative intersectoral approaches to nutrition in a community on the urban fringe Health Educ Behav 2001 28 3 306 319 11380051
Fox LM Furness BW Haser JK Desire D Brissau JM Milord MD Lafontant J Lammie PJ Beach MJ Tolerance and efficacy of combined diethylcarbamazine and albendazole for treatment of Wuchereria bancrofti and intestinal helminth infections in Haitian children Am J Trop Med Hyg 2005 73 115 121 16014845
Prasittisuk C Vector-control synergies, between Roll Back Malaria and the Global Programme to Eliminate Lymphatic Filariasis Ann Trop Med Parasitol 2002 96 Suppl 2 S133 S137 10.1179/000349802125002482 12630409
Heukelbach J Winter B Wilcke T Muehlen M Albrecht S de Oliveira FAS Kerr-Pontes LRS Liesenfeld O Feldmeier H Selective mass treatment with ivermectin to control intestinal helminthiases and parasitic skin diseases in a severely affected population Bull World Health Organ 2004 82 563 571 15375445
Stephenson LS Holland CV Ottesen EA Chappell LH Eds Controlling Intestinal Helminths while Eliminating Lymphatic Filariasis 2002 Cambridge: Cambridge University Press
Awasthi S Pande VK Fletcher RH Effectiveness and cost-effectiveness of albendazole in improving nutritional status of pre-school children in urban slums Indian Pediatr 2000 37 19 29 10745385
Simeon DT Grantham-McGregor SM Callender JE Wong MS Treatment of Trichuris trichiura infection improves growth, spelling scores and school attendance in some children J Nutr 1995 125 1875 1883 7616304
Dunford C MkNelly B Best Practices:Using Microfinance to Improve Health and Nutrition Security Global Health Council Health Link 2002 118 9 22
Steketee RW Pregnancy, nutrition and parasitic diseases J Nutr 2003 133 Suppl 2 1661S 1667S 12730482
Yamey G The world's most neglected diseases BMJ 2002 325 176 177 10.1136/bmj.325.7357.176 12142292
Von Schirnding Y Health and sustainable development: can we rise to the challenge? Lancet 2002 360 632 637 10.1016/S0140-6736(02)09777-5 12241950
Savioli L Albonico M Engels D Montresor A Progress in the prevention and control of schistosomiasis and soil-transmitted helminthiasis Parasitol Int 2004 53 103 113 10.1016/j.parint.2004.01.001 15081942
Esrey SA Potash JB Roberts L Shiff C Effects of improved water supply and sanitation on ascariasis, diarrhoea, dracunculiasis, hookworm infection, schistosomiasis, and trachoma Bull World Health Organ 1991 69 609 621 1835675
Tun-Lin W Kay BH Barnes A The Premise Condition Index: a tool for streamlining surveys of Aedes aegypti Am J Trop Med Hyg 1995 53 591 594 8561259
Heller L Colosimo EA Antunes CM Environmental sanitation conditions and health impact: a case-control study Rev Soc Brasil Med Trop 2003 36 41 50
Fewtrell L Kaufmann RB Kay D Enanoria W Haller L Colford JM JrWater, sanitation, and hygiene interventions to reduce diarrhoea in less developed countries: a systematic review and meta-analysis Lancet Infect Dis 2005 5 42 52 10.1016/S1473-3099(04)01253-8 15620560
Clasen TF Brown J Collin S Suntura O Cairncross S Reducing diarrhea through the use of household-based ceramic water filters: A randomized, controlled trial in rural Bolivia Am J Trop Med Hyg 2004 70 651 657 15211008
III Congreso de las Americas de Municipios Saludables y Comunidades Saludables, Medellín, Antioquia, Colombia, 08 al 12 de marzo de 1999 http://www.disaster-info.net/col-ops/saludambiente/Medellin99.htm
Otterpohl R Braun U Oldenburg M Innovative technologies for decentralised water-, wastewater and biowaste management in urban and peri-urban areas Water Sci Technol 2003 48 23 32 14753515
Moraes LR Cancio JA Cairncross S Impact of drainage and sewerage on intestinal nematode infections in poor urban areas in Salvador, Brazil Trans R Soc Trop Med Hyg 2004 98 197 204 10.1016/S0035-9203(03)00043-9 15049458
Curtis V Biran A Deverell K Hughes C Bellamy K Drasar B Hygiene in the home: relating bugs and behaviour Soc Sci Med 2003 57 657 672 10.1016/S0277-9536(02)00409-4 12821014
Harvey SA Winch PJ Leontsini E Torres Gayoso C Lopez Romero S Gilman RH Oberhelman RA Domestic poultry-raising practices in a Peruvian shantytown: implications for control of Campylobacter jejuni -associated diarrhea Acta Trop 2003 86 41 54 10.1016/S0001-706X(03)00006-8 12711102
Rojas-De-Arias A Chagas disease prevention through improved housing using an ecosystem approach to health Cad Saude Publica 2001 Suppl 17 89 97 11426269
Lindsay SW Jawara M Paine K Pinder M Walraven GE Emerson PM Changes in house design reduce exposure to malaria mosquitoes Trop Med Int Health 2003 8 512 517 10.1046/j.1365-3156.2003.01059.x 12791056
Ansari MA Razdan RK Concurrent control of mosquitoes and domestic pests by use of deltamethrin-treated curtains in the New Delhi Municipal Committee, India J Am Mosq Control Assoc 2001 17 131 136 11480821
Hyma B Ramesh A Chakrapani KP Urban malaria control situation and environmental issues, Madras City, India Ecol Dis 1983 2 321 335 6681162
Bruce N McCracken J Albalak R Schei MA Smith KR Lopez V West C Impact of improved stoves, house construction and child location on levels of indoor air pollution exposure in young Guatemalan children J Expo Anal Environ Epidemiol 2004 14 Suppl 1 S26 33 10.1038/sj.jea.7500355 15118742
World Energy Council, The Challenge of Rural Energy Poverty in Developing Countries. 4. Promising Technology Developments http://www.worldenergy.org
INBAR technical Report 20 Designing and Building with Bamboo http://www.inbar.int/publication/txt/INBAR_Technical_Report_No20.htm
Ezzati M Utzinger J Cairncross S Cohen AJ Singer BH Environmental risks in the developing world: exposure indicators for evaluating interventions, programmes, and policies J Epidemiol Community Health 2005 59 15 22 10.1136/jech.2003.019471 15598721
Satoto Hendratno S Subagyo H Rafilluddin Z Mogensen M Hall A Bundy D Crompton DWT, Montresor A, Nesheim MC, Savioli L Partnership for Child Development: an international programme to improve the health of school-age children by school-based health services including deworming Controlling Diseases due to Helminth Infections 2003 Geneva: World Health Organization 93 97
Taylor M Coovadia HM Kvalsvig JD Jinabhai CC Reddy P Helminth control as an entry point for health-promoting schools in KwaZulu-Natal S Afr Med J 1999 89 273 279 10226673
Focusing Resources On Effective School Health: A FRESH Start to Improving the Quality and Equity of Education http://www.schoolsandhealth.org/FRESH.htm
Strina A Cairncross S Barreto ML Larrea C Prado MS Childhood diarrhea and observed hygiene behavior in Salvador, Brazil Am J Epidemiol 2003 157 1032 1038 10.1093/aje/kwg075 12777367
Olsen A Samuelsen H Onyango-Ouma W A study of risk factors for intestinal helminth infections using epidemiological and anthropological approaches J Biosoc Sci 2001 33 569 584 10.1017/S0021932001005697 11683225
Asaolu SO Ofoezie IE The role of health education and sanitation in the control of helminth infections Acta Trop 2003 86 283 294 10.1016/S0001-706X(03)00060-3 12745145
Luby SP Agboatwalla M Feikin DR Painter J Billhimer W Altaf A Hoekstra RM Effect of handwashing on child health: a randomised controlled trial Lancet 2005 366 225 233 10.1016/S0140-6736(05)66912-7 16023513
Alderman H Hentschel J Sabates R With the help of one's neighbors: externalities in the production of nutrition in Peru Soc Sci Med 2003 56 2019 2031 10.1016/S0277-9536(02)00183-1 12697194
Anigan G Integrating parasite control and family planning Initiatives Popul 1979 5 24 27 12338133
Tanumihardjo SA Permaesih D Muhilal Vitamin A status and hemoglobin concentrations are improved in Indonesian children with vitamin A and deworming interventions Eur J Clin Nutr 2004 58 1223 1230 10.1038/sj.ejcn.1601953 15054437
Black R Micronutrient deficiency – an underlying cause of morbidity and mortality Bull World Health Organ 2003 81 79 12751414
Black RE Zinc deficiency, infectious disease and mortality in the developing world J Nutr 2003 Suppl 133 1485S 1489S 12730449
Zimmermann MB Zeder C Chaouki N Saad A Torresani T Hurrell RF Dual fortification of salt with iodine and microencapsulated iron: a randomized, double-blind, controlled trial in Moroccan schoolchildren Am J Clin Nutr 2003 77 425 432 12540404
Devendra T Sri Lanka : An innovative approach to the implementation of integrated parasite control, nutrition and family planning projects Asian Pac Popul Programme News 1984 13 28 31 12313055
Rivera JA Habicht J-P Effect of supplementary feeding on the prevention of mild-to-moderate wasting in conditions of endemic malnutrition in Guatemala Bull World Health Organ 2002 80 926 932 12571719
Bordigon GP Shakya DR Crompton DWT, Montresor A, Nesheim MC, Savioli L A deworming program in Nepal supported by the World Food Program Controlling Diseases due to Helminth Infections 2003 Geneva: World Health Organization 87 92
Malekafzali H Abdollahi Z Mafi A Naghavi M Community-based nutritional intervention for reducing malnutrition among children under 5 years of age in the Islamic Republic of Iran East Mediterr Health J 2000 6 238 245 11556008
Kroeger A Avinna A Ordonnez-Gonzalez J Escandon C Community cooperatives and insecticide-treated materials for malaria control: a new experience in Latin America Malar J 2002 1 15 http://www.malariajournal.com/start.asp10.1186/1475-2875-1-15 12473181
Bolles K Speraw C Berggren G Lafontant JG Ti Foyer (Hearth) community-based nutrition activities informed by the positive deviance approach in Leogane, Haiti: a programmatic description Food Nutr Bull 2002 23 11 17 12503227
Nanda P Women's participation in rural credit programmes in Bangladesh and their demand for formal health care: is there a positive impact? Health Econ 1999 8 415 428 10.1002/(SICI)1099-1050(199908)8:5<415::AID-HEC450>3.0.CO;2-L 10470548
Atinmo T Oyewole D Finding solutions to the nutritional dilemmas in Africa for child health: HIV/AIDS orphans, poverty and hunger Asia Pac J Clin Nutr 2004 Suppl 13 S6 15294468
Molitor C Becoming successful entrepreneurs. Bangladesh. ADB supports pioneering family-based approach to provide micro-credit and skills training Integration 1996 47 38 40 12347308
Blackman T Harvey J Lawrence M Simon A Neighbourhood renewal and health: evidence from a local case study Health Place 2001 7 93 103 10.1016/S1353-8292(01)00003-X 11470223
Harpham T Burton S Blue I Healthy city projects in developing countries: the first evaluation Health Promot Internation 2001 16 111 125 10.1093/heapro/16.2.111
Fuzesi Z Levy BS Levenstein C Barbeau E Ori II Rest K Tistyan L A multisectoral approach to prevention of lead poisoning in Hungary: lessons learned and potential for replicability elsewhere Int J Occup Environ Health 1997 3 60 67 9891102
Takano T Nakamura K Watanabe M Urban residential environments and senior citizens' longevity in megacity areas: the importance of walkable green spaces J Epidemiol Community Health 2002 56 913 918 10.1136/jech.56.12.913 12461111
Saelens BE Sallis JF Frank LD Environmental correlates of walking and cycling: findings from the transportation, urban design, and planning literatures Ann Behav Med 2003 25 80 91 10.1207/S15324796ABM2502_03 12704009
Manan A Ibrahim M Community-based river management in Southeast Sulawesi, Indonesia: a case study of the Bau-Bau River Water Sci Technol 2003 48 181 190 14653649
Spiegel J Bonet M Yassi A Tate RB Concepcion M Canizares M Evaluating the effectiveness of a multi-component intervention to improve health in an inner-city Havana community Int J Occup Environ Health 2003 9 118 127 12848239
PAHO/ WHO Municipios productivos y seguridad alimentaría al nivel local http://www.paho.org/Spanish/ad/dpc/vp/rimsa12_19-s.pdf
Douglas MJ Conway L Gorman D Gavin S Hanlon P Achieving better health through health impact assessment Health Bull (Edinb) 2001 59 300 305 12664743
Cave B Curtis S Developing a practical guide to assess the potential health impact of urban regeneration schemes Promot Educ 2001 8 12 16 11379428
Paterson RT Joaquin N Chamon K Palomino E The productivity of small animal species in small-scale mixed farming systems in subtropical Bolivia Trop Anim Health Prod 2001 33 1 14 10.1023/A:1010369024658 11234187
Tontisirin K Nantel G Bhattacharjee L Food-based strategies to meet the challenges of micronutrient malnutrition in the developing world Proc Nutr Soc 2002 61 243 250 10.1079/PNS2002155 12133206
Faber M Venter SL Benade AJ Increased vitamin A intake in children aged 2–5 years through targeted home-gardens in a rural South African community Public Health Nutr 2002 5 11 16 10.1079/PHN2001239 12001973
Schipani S van der Haar F Sinawat S Maleevong K Dietary intake and nutritional status of young children in families practicing mixed home gardening in northeast Thailand Food Nutr Bull 2002 23 175 180 12094667
Egziabher AG Lee-Smith D Maxwell DG Memon PA Mougeot LJA Sawil CJ Cities Feeding People: An Examination of Urban Agriculture in East Africa 1994 Ottawa: IDRC http://www.idrc.ca
Dessie T Ogle B Village poultry production systems in the central highlands of Ethiopia Trop Anim Health Prod 2001 33 521 537 10.1023/A:1012740832558 11770206
Annon Environmental primary care for the 21 st century Rev Panam Salud Pública 1998 4 290 296 9924511
Bermejo A Bekui A Community participation in disease control Soc Sci Med 1993 36 1145 1150 10.1016/0277-9536(93)90234-U 8511643
Vazquez-Flores S Hygiene and restraint of pigs is associated with absence of Taenia solium cysticercosis in a rural community of Mexico Salud Pública Mex 2001 43 574 576 11816233
Jimenez S Perez A Gil H Schantz P Ramalle E Juste R Progress in control of cystic echinococcosis in La Rioja, Spain: decline in infection prevalences in human and animal hosts and economic costs and benefits Acta Trop 2002 83 213 221 10.1016/S0001-706X(02)00091-8 12204394
Ezzati M Indoor air pollution and health in developing countries Lancet 2005 366 104 106 10.1016/S0140-6736(05)66845-6 16005317
Rodriguez-Garcia R The health-development link: travel as a public health issue J Community Health 2001 26 93 112 10.1023/A:1005225129295 11322757
Cartwright RY Chahed M Foodborne diseases in travellers World Health Stat Q 1997 50 102 110 9282392
Heukelbach J Eisele M Jackson A Feldmeier H Topical treatment of tungiasis: a randomized, controlled trial Ann Trop Med Parasitol 2003 97 743 749 10.1179/000349803225002408 14613633
|
16283932
|
PMC1318484
|
CC BY
|
2021-01-04 22:25:40
|
no
|
BMC Public Health. 2005 Nov 11; 5:119
|
utf-8
|
BMC Public Health
| 2,005 |
10.1186/1471-2458-5-119
|
oa_comm
|
==== Front
Cardiovasc DiabetolCardiovascular Diabetology1475-2840BioMed Central London 1475-2840-4-181632116210.1186/1475-2840-4-18Original InvestigationPrevalence of antiplatelet therapy in patients with diabetes Miller Shaun R [email protected] Benjamin [email protected] Charles D [email protected] University of Vermont College of Medicine, Burlington, Vermont, USA2005 1 12 2005 4 18 18 4 10 2005 1 12 2005 Copyright © 2005 Miller et al; licensee BioMed Central Ltd.2005Miller 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.
Objective
To determine the prevalence of, and patient characteristics associated with, antiplatelet therapy in a cohort of primary care patients with Type 1 or Type2 diabetes.
Methods
Subjects participating in a randomized trial of a decision support system were interviewed at home and medication usage verified by a research assistant. Eligibility for antiplatelet therapy was determined by American Diabetes Association criteria and clinical contraindications. The association between antiplatelet use and patient characteristics was examined using bivariate and multivariate logistic regression.
Results
The mean age of subjects was 64 years (range 31–93). The prevalence of antiplatelet use was 54% overall; 45% for subjects without known CVD vs. 78% for those with CVD; 46% for women vs. 63% for men; and 45% for younger subjects (age< 65) vs. 62% for senior citizens. After controlling for race/ethnicity, income, education, marital status, insurance status and prescription coverage, the following were associated with the use of antiplatelet therapy: presence of known CVD (OR 3.4 [2.2, 5.1]), male sex (OR 2.0 [1.4, 2.8]), and age > = 65 (OR 1.9 [1.3, 2.7]). The prevalence of antiplatelet therapy for younger women without CVD was 32.8% compared to a prevalence of 90.3% for older men with CVD.
Conclusion
Despite clinical practice guidelines recommending antiplatelet therapy for patients with diabetes, there are still many eligible patients not receiving this beneficial therapy, particularly patients under 65, women, and patients without known CVD. Effective methods to increase antiplatelet use should be considered at the national, community, practice and provider level.
==== Body
Introduction
Cardiovascular disease (CVD) is the leading cause of morbidity and mortality in adults with diabetes [1-4]. Antiplatelet therapy, with either aspirin or the newer platelet aggregation inhibitors, has been shown to be safe and cost effective for reducing the risk of recurrent vascular events [5-8]. Consensus guidelines recommend the use of antiplatelet therapy for both primary and secondary prevention of CVD [9,10]. In 1997, the American Diabetes Association (ADA) recommended antiplatelet therapy for adults with diabetes and co-existing CVD, and for adults with diabetes over 30 years of age, even in the absence of CVD [11]. Prior to the publication of the ADA recommendations for antiplatelet prophylaxis, the national rate of aspirin use among patients with diabetes was estimated at 13% for individuals without CVD and at 37% for those with CVD [12]. By 2001 this latter prevalence, as determined by telephone survey, had increased to 48.7% [13]. Current estimates suggest that approximately 5% of adults cannot tolerate aspirin therapy. For these individuals, an alternative antiplatelet agent may be used [14].
Despite increasing evidence to support its effectiveness among patients with diabetes, antiplatelet therapy has been under-utilized [12,15,16], particularly in women [13]. While several observational studies have examined the prevalence of aspirin use both before and after the publication of the 1997 ADA recommendations, none have included the use of other antiplatelet agents and may therefore have underestimated the prevalence of antiplatelet therapy. The goal of this study is to determine the prevalence of antiplatelet therapy (aspirin and newer platelet aggregation inhibitors) for both primary and secondary prevention of CVD in diabetes and to examine the patient characteristics that are associated with failure to use this important therapy.
Methods
This study was part of a larger project, the Vermont Diabetes Information System (VDIS), a cluster-randomized trial of a laboratory-based diabetes decision support system in a region-wide sample of 7295 adults with diabetes from 55 community Primary Care practices [17]. We did not distinguish between Type 1 and Type 2 diabetes because this distinction is not clinically important when recommending antiplatelet therapy. A field survey targeted at a sub-sample of subjects was designed to provide a better understanding of the non-laboratory features of diabetes. Patients were selected at random from the subjects in each practice participating in the VDIS trial and invited by phone to participate in an in-home interview. Patient names were randomly sorted and patients contacted by telephone until a sample of approximately 15% of the patients from each practice agreed to an interview. We attempted to contact 4209 patients and reached 1576. Of these, 1006 agreed to be interviewed. Demographic information including age, sex, race, ethnicity, education, income, marital status and history of cardiovascular disease were obtained by questionnaire. A complete list of medications was obtained by a research assistant by direct observation of all of the medication containers and recording of the medication name, dose, frequency and route of administration. The interviews occurred between July 2003 and March 2005. The University of Vermont Institutional Review Board approved the study and all subjects gave written informed consent to participate in the interview.
For the purposes of this cross-sectional study, a subset of interviewed subjects was created using inclusion and exclusion criteria based on the current American Diabetes Association (ADA) recommendations for the use of antiplatelet therapy [18]. The subset of subjects who were eligible for antiplatelet therapy consisted of all subjects in the VDIS interview cohort 30 years or older, and those under 30 years with a self-reported history of either coronary heart disease, stroke or transient ischemic attack, or peripheral vascular disease. For the purposes of the study we defined cardiovascular disease (CVD) as any of the above manifestations of vascular disease. We excluded patients with specific contraindications to antiplatelet therapy: peptic ulcer disease (144), severe liver disease (13), and those on current warfarin therapy (75), for whom decisions about concomitant use of antiplatelet therapy and anticoagulation would be individualized. No information was available about side effects or previous discontinuation of therapy was available. Some subjects had more than one exclusion; a total of 221 subjects were excluded for a final sample of 785 subjects. Antiplatelet use was defined as daily use of aspirin (at least 75 mg/day); clopidogrel; ticlodipine; or cilostazol; or a combination of aspirin and clopidogrel, ticlodipine, or cilostazol daily. The specific indication for the anti-platelet agent in each subject was not known.
We used logistic regression to test the bivariate association of anti-platelet use with variables that were potentially important based on previous research and clinical judgment, including age, sex, race/ethnicity, income, education, marital status, insurance status and pharmacy benefits, years since diagnosis of diabetes, smoking, body mass index, frequency of visits to primary care physician, specialist involvement in care (endocrinologist visit in the last year, attendance at a diabetes education class within the last year), and the various categories of CVD. Variables that demonstrated an association in bivariate modeling at a significance level of p < 0.1 were further examined with multivariate regression modeling in which insignificant (p > 0.05) variables were eliminated in a backward stepwise fashion.
Results
The characteristics of the study population are presented in Table 1. The mean age was 64 years with half the population over age 65. Most graduated high school and fewer than 3% were uninsured. Most subjects were overweight or obese (89%), with 18% falling in the severely obese category (body mass index of 40 or greater). Twenty-six percent of the population had cardiovascular disease, with myocardial infarction being the most common manifestation in 16%.
Table 1 VDIS Subjects Eligible for Antiplatelet Therapy (N = 785)
Characteristic N Mean or Prevalence
Female 435 55%
Age, mean (SD) (range) 785 64 (11.8) (31–93)
Age > = 65 395 50%
Race/Ethnicity
White, non-Hispanic 764 97.6%
Education
High school graduate 599 77%
Marital Status
Married or living as married 503 64%
Current smoker 125 16%
Endocrinology consult visit in past year 122 16%
Diabetes education visit in past year 173 23%
Insurance Status *
None 21 2.7%
Private 472 60.5%
Medicare 447 57.6%
Medicaid 143 18.5%
Military 33 4.3%
Annual household income
< $30,000 407 56.4%
Body Mass Index
Normal (< 25) 85 10.9%
Overweight (25–29.9) 182 23.4%
Obese (> = 30) 510 65.6%
Very Obese (> = 40) 141 18.2%
Cardiovascular Disease §
Any CVD 206 26.2%
Myocardial Infarction 127 16.2%
Stroke 68 8.7%
Peripheral Vascular Disease 61 7.8%
Number of prescription medications, mean (SD) 785 6.2 (3.5)
Number of MD visits in previous year, mean (SD) 785 1.5 (2.2)
Years since diagnosis of DM, mean (SD) 785 9.9 (10.1)
Antiplatelet therapy (aspirin or other) 421 53.6%
Aspirin only 371 47.3%
Non-aspirin platelet aggregation inhibitor only 20 2.5%
Aspirin and platelet aggregation inhibitor 30 3.8%
Antiplatelet therapy if known CVD 206 78.2%
* Subjects may have more that one type of insurance coverage
§Subjects may have more than one type of CVD
The prevalence of antiplatelet use was 53.6% (47.3% aspirin alone, 2.5% newer platelet aggregation inhibitor and 3.8% both) for all eligible subjects and 78.2% for subjects with known CVD. The characteristics associated with antiplatelet medication use are noted in Table 2. Male sex and older age are both associated with a two-fold increase in antiplatelet use (p < 0.001). Cardiovascular disease was associated with a three-fold increase in antiplatelet use, with MI showing a six-fold increase (p < 0.001). Other factors that were associated with anti-platelet agent use were: an endocrinology visit in the previous year (p = 0.004), and Medicare insurance coverage (p < 0.001).
Table 2 Bivariate Associations with Anti-Platelet Use in Adults with Diabetes
Characteristic Odds Ratio [95% CI] P Value
Male sex 2.0 [1.5, 2.7] <0.001
Age > 65 years 2.0 [1.5, 2.6] <0.001
Race/ethnicity, White, non-Hispanic 2.0 [0.8, 5.2] 0.15
Education: > HS graduate 0.8 [0.6, 1.1] 0.25
Married or living as married 1.0 [0.8, 1.4] 0.84
Smoker 0.8 [.05, 1.1] 0.16
Endocrinology consult within 1 year 1.8 [1.2, 2.7] 0.004
Diabetes education class within 1 year 1.4 [1.0, 2.0] 0.06
Insurance category
None 0.28 [0.10, 0.78] 0.02
Private 1.15 [0.86, 1.53] 0.35
Medicare 1.88 [1.40, 2.51] <0.001
Medicaid 0.98 [0.68, 1.4] 0.91
Military 1.78 [0.85, 3.70] 0.13
Income < $30,000 1.1 [0.8, 1.5] 0.55
Prescription coverage 0.9 [0.7, 1.1] 0.19
Body Mass Index
BMI category (normal, overweight, obese) 0.8 [0.7, 1.0] 0.11
Obese (BMI > 30) 0.8 [0.6, 1.0] 0.07
Severe Obesity (BMI > 40) 0.6 [.04, 0.9] 0.01
Type of Cardiovascular Disease
Any CVD 3.3 [2.4, 4.3] <0.001
Myocardial Infarction 6.7 [4.0, 11.3] <0.001
Stroke 2.1 [1.2, 3.5] 0.008
Peripheral Vascular Disease 3.1 [1.7, 5.8] <0.001
Number of PCP visits in previous month 1.04 [0.97, 1.11] 0.26
Duration of diabetes (in years) 1.01 [1.00, 1.03] 0.13
In multivariable analysis, three characteristics remained independently associated with antiplatelet use while controlling for important covariates (see Table 3). Subjects with a history CVD were more likely to be on appropriate antiplatelet therapy (OR 3.4 [CI 2.2, 5.1]), as were subjects 65 or older (OR 1.9 [CI 1.3, 2.7]) and men (OR 2.0 [CI 1.4, 2.8]), (p < 0.001 for each). Table 4 indicates the prevalence of antiplatelet use in each of these patient subgroups. The lowest rates are among women under 65 without CVD (32.8%), and highest among older men with known CVD (90.3%).
Table 3 Multivariate Model of Characteristics Associated with Anti-Platelet Use in Adults with Diabetes*
Characteristic Odds Ratio [95% CI] P Value
CVD history 3.4 [2.2, 5.1] <0.001
Senior (age > 65) 1.9 [1.3, 2.7] <0.001
Male sex 2.0 [1.4, 2.8] <0.001
*Controlling for race, income, education, marital status, insurance, prescription coverage
Table 4 Prevalence of Antiplatelet Use by Patient Characteristic
Sex Age % on Antiplatelet therapy N
CVD Absent Female Less than 65 33% 180
65 or older 45% 151
Male Less than 65 49% 137
65 or older 60% 111
CVD Present Female Less than 65 58% 33
65 or older 77% 70
Male Less than 65 78% 41
65 or older 90% 62
Among the 206 subjects with known CVD we found similar associations with antiplatelet use with age > 65 (OR 3.0 CI [1.4, 6.3]), and male sex (OR 3.3 [1.5, 7.1]).
Discussion
We found a prevalence of antiplatelet therapy use among adults with diabetes of 53.6% (47.3% aspirin alone, 2.5% newer platelet aggregation inhibitor and 3.8% both), which is similar to the recent nationally representative telephone survey estimate of aspirin use of 48.7% by Persell [13]. Among patients with CVD we found a prevalence of antiplatelet therapy of 78.2%, compared to 74.2% by Persell [13].
We found the highest rates among subjects with CAD. Following the CAPRIE trial in 1996, which showed a slight advantage in secondary prevention of cardiovascular events for clopidogrel vs. aspirin, clopidogrel has been increasingly used both in addition to aspirin and as its replacement [19,20]. The newer platelet aggregation inhibitors are also increasingly used for acute coronary syndrome and after percutaneous coronary intervention [21]. The strong evidence for CAD indications is reflected in our findings that subjects with coronary artery disease were the most likely to be receiving antiplatelet therapy.
Our motivation in exploring the factors associated with antiplatelet agent use was to help identify subgroups that may be targeted for special efforts to increase antiplatelet therapy. We found that women, patients younger than 65, and those without CVD were less likely to be using antiplatelet therapy. On the other end of the spectrum, over 90% of men over 65 with CVD were taking antiplatelet therapy. This high level of use among those at the highest risk supports the achievability of the consensus guidelines. A recent meta-analysis including 287 studies and 135,000 patients at high vascular risk showed that antiplatelet therapy reduced serious vascular events (non-fatal MI, non-fatal stroke, vascular death) by 36 (SE 5) per 1000 patients treated for two years [22]. Assuming that we can move from our overall prevalence of antiplatelet therapy of 54% to our best rate of 90%, we estimate that another 13 serious vascular events per 1000 could be averted over two years. If this is projected to the 18.2 million adults with diabetes in the United States [23], we estimate that 238,000 serious vascular events could be averted.
Why are patients with diabetes not receiving antiplatelet therapy despite consensus guidelines? First of all, prescribers may feel there is some ambiguity regarding the role of aspirin in CVD primary prevention for patients with diabetes. For example, while the Primary Prevention Project, which randomized over 4000 diabetic and non-diabetic subjects with CVD risk factors to aspirin or no aspirin, was stopped early because of the beneficial effects of aspirin, in the subgroup with diabetes the benefits were smaller and not statistically significant. [24] This raises the question of potential differences in the role of antiplatelet therapy in diabetes. Secondly, even if prescribers agree with the guideline, there are other barriers to achieving perfect compliance. A qualitative study exploring reasons cited by physicians for not prescribing aspirin included: difficulties in applying generic guidelines to individuals, patient resistance to taking aspirin, prioritization of other issues in a time constrained visit, and communication problems in reviewing the medications of patients with stroke [25].
Why might women be less likely to be receiving antiplatelet therapy? Gender differences have been well documented in the diagnosis and treatment of heart disease [26] In addition, the effects of aspirin may be different in men and women; a recent study of primary prevention of CVD in almost 40,000 women over 45 years of age showed that, while stroke risk was lowered, myocardial infarction and overall cardiovascular mortality were not [27]. Physicians may be less enthusiastic about the evidence base supporting the use of antiplatelet therapy in women. For patients under age 65, physicians (and patients themselves) may not perceive the risks of CVD as high enough to warrant antiplatelet therapy.
We observed an association between the use of antiplatelet therapy and the type of CVD. Patients with a history of prior myocardial infarction were more likely to be on antiplatelet therapy (86%) than those with a history of peripheral vascular disease (77%) or cerebrovascular accident (CVA) (69%). Furthermore, only 54% of patients with CVA and no other CVD were using an antiplatelet agent. Antiplatelet therapy has been shown to reduce the risk of recurrent CVA by 11% to 15% in patients with prior ischemic stroke of non-cardiac origin and reduce the risk of stroke, MI, and vascular death, by 22% [28]. The extent to which stroke patients and their physicians avoid antiplatelet therapy due to risk of bleeding is not known. This lower use of antiplatelet therapy in stroke patients identifies an area for potential investigation and intervention to improve anti-platelet regimens in this patient population.
Health insurance coverage has been shown to be an important factor in the delivery of medical services. Increasing levels of health insurance have a positive correlation with the likelihood that an individual will receive appropriate preventive care [29]. In diabetes, poor insurance coverage has been associated with delayed or omitted preventive services [30]. We found that health insurance coverage was not an important predictor of anti-platelet therapy in patients of this cohort, but the level of health insurance coverage was high and subjects were under the care of a primary care provider suggesting good access to care. In the case of an expensive medicine like clopidogrel, lack of prescription drug coverage could contribute to lack of use. However, aspirin, which comprises the majority of the antiplatelet agents in our study, is a low cost, nonprescription medication.
This study has several limitations. Our population, while representative of patients receiving primary care in the rural Northeast may not be representative of all adults with diabetes in the U.S. We do not have information regarding allergies or side effects associated with antiplatelet medications. It is possible that eligible subjects were unable to tolerate therapy, though it is unlikely this would be the case in more than 5% of subjects. It is unlikely that medication intolerance would be correlated with age, sex or cardiovascular disease. We do not have information regarding the indication for aspirin use, though 98% of subjects reported low-dose aspirin use (< = 325 mg/d) suggesting prophylaxis. Our analysis does not indicate causality and the exact mechanisms promoting or deterring the use of recommended interventions is unknown.
There have been a variety of successful interventions directed at increasing the use of antiplatelet therapy for the prevention of CVD including: HMO-directed quality improvement efforts [31], intensive multifaceted case management [32], pharmacy-directed interventions [33,34], and electronic medical record reminder systems [35]. A VA study found that physician counseling was highly associated with antiplatelet therapy and suggested that this simple intervention could prevent many cardiovascular events and deaths [36]. There are many ways in which antiplatelet use can be increased; it is now a question of which approach can be most efficiently adapted in each clinical setting.
Conclusion
Despite clinical practice guidelines recommending antiplatelet therapy for patients with diabetes, there are still many eligible patients not receiving this beneficial therapy, particularly patients under 65, women, and patients without known CVD. Effective methods to increase antiplatelet use should be considered at the national, community, practice and provider level.
Acknowledgements
Funded by the National Institute of Diabetes and Digestive and Kidney Diseases (R01 DK61167 and K24 DK068380).
==== Refs
Moss SE Klein R Klein BE Cause-specific mortality in a population-based study of diabetes Am J Public Health 1991 81 1158 1162 1951827
Turner R Cull C Holman R United Kingdom Prospective Diabetes Study 17: a 9-year update of a randomized, controlled trial on the effect of improved metabolic control on complications in non-insulin-dependent diabetes mellitus Ann Intern Med 1996 124 136 145 8554206
Haffner SM Lehto S Ronnemaa T Pyorala K Laakso M Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction N Engl J Med 1998 339 229 234 9673301 10.1056/NEJM199807233390404
Kannel WB McGee DL Diabetes and cardiovascular disease. The Framingham study Jama 1979 241 2035 2038 430798 10.1001/jama.241.19.2035
ETDRS I Aspirin effects on mortality and morbidity in patients with diabetes mellitus. Early Treatment Diabetic Retinopathy Study report 14. ETDRS Investigators Jama 1992 268 1292 1300 1507375 10.1001/jama.268.10.1292
Gaspoz JM Coxson PG Goldman PA Williams LW Kuntz KM Hunink MG Goldman L Cost effectiveness of aspirin, clopidogrel, or both for secondary prevention of coronary heart disease N Engl J Med 2002 346 1800 1806 12050341 10.1056/NEJM200206063462309
Harpaz D Gottlieb S Graff E Boyko V Kishon Y Behar S Effects of aspirin treatment on survival in non-insulin-dependent diabetic patients with coronary artery disease. Israeli Bezafibrate Infarction Prevention Study Group Am J Med 1998 105 494 499 9870835 10.1016/S0002-9343(98)00328-3
Nobles-James C James EA Sowers JR Prevention of cardiovascular complications of diabetes mellitus by aspirin Cardiovasc Drug Rev 2004 22 215 226 15492769
Aspirin for the primary prevention of cardiovascular events: recommendation and rationale Ann Intern Med 2002 136 157 160 11790071
Hennekens CH Dyken ML Fuster V Aspirin as a therapeutic agent in cardiovascular disease: a statement for healthcare professionals from the American Heart Association Circulation 1997 96 2751 2753 9355934
ADA Aspirin therapy in diabetes. American Diabetes Association Diabetes Care 1997 20 1772 1773 9353621
Rolka DB Fagot-Campagna A Narayan KM Aspirin use among adults with diabetes: estimates from the Third National Health and Nutrition Examination Survey Diabetes Care 2001 24 197 201 11213865
Persell SD Baker DW Aspirin use among adults with diabetes: recent trends and emerging sex disparities Arch Intern Med 2004 164 2492 2499 15596642 10.1001/archinte.164.22.2492
Harrington RA Becker RC Ezekowitz M Meade TW O'Connor CM Vorchheimer DA Guyatt GH Antithrombotic therapy for coronary artery disease: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy Chest 2004 126 513S 548S 15383483 10.1378/chest.126.3_suppl.513S
McFarlane SI Jacober SJ Winer N Kaur J Castro JP Wui MA Gliwa A Von Gizycki H Sowers JR Control of cardiovascular risk factors in patients with diabetes and hypertension at urban academic medical centers Diabetes Care 2002 25 718 723 11919131
Stafford RS Aspirin use is low among United States outpatients with coronary artery disease Circulation 2000 101 1097 1101 10715254
MacLean CD Littenberg B Gagnon MS Reardon M Turner PD Jordan C The Vermont Diabetes Information System (VDIS): Study Design and Subject Recruitment for a Cluster Randomized Trial of a Decision Support System in a Regional Sample of Primary Care Practices Clinical Trials 2004 1 532 544 16279294 10.1191/1740774504cn051oa
ADA Aspirin Therapy in Diabetes Diabetes Care 2004 27 S72 S73 14693931
CAPRIE A randomised, blinded, trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). CAPRIE Steering Committee Lancet 1996 348 1329 1339 8918275 10.1016/S0140-6736(96)09457-3
Tran H Anand SS Oral antiplatelet therapy in cerebrovascular disease, coronary artery disease, and peripheral arterial disease Jama 2004 292 1867 1874 15494585 10.1001/jama.292.15.1867
Mehta SR Yusuf S Short- and long-term oral antiplatelet therapy in acute coronary syndromes and percutaneous coronary intervention J Am Coll Cardiol 2003 41 79S 88S 12644345 10.1016/S0735-1097(02)02831-0
Antithrombotic_Trialists'_Collaboration Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients BMJ 2002 324 71 86 11786451
Engelgau MM Geiss LS Saaddine JB Boyle JP Benjamin SM Gregg EW Tierney EF Rios-Burrows N Mokdad AH Ford ES Imperatore G Narayan KM The evolving diabetes burden in the United States Ann Intern Med 2004 140 945 950 15172919
Sacco M Pellegrini F Roncaglioni MC Avanzini F Tognoni G Nicolucci A Primary prevention of cardiovascular events with low-dose aspirin and vitamin E in type 2 diabetic patients: results of the Primary Prevention Project (PPP) trial Diabetes Care 2003 26 3264 3272 14633812
Short D Frischer M Bashford J Ashcroft D Why are eligible patients not prescribed aspirin in primary care? A qualitative study indicating measures for improvement BMC Fam Pract 2003 4 9 12871601 10.1186/1471-2296-4-9
Tecce MA Dasgupta I Doherty JU Heart disease in older women. Gender differences affect diagnosis and treatment Geriatrics 2003 58 33 9 14682096
Ridker PM Cook NR Lee IM Gordon D Gaziano JM Manson JE Hennekens CH Buring JE A randomized trial of low-dose aspirin in the primary prevention of cardiovascular disease in women N Engl J Med 2005 352 1293 1304 15753114 10.1056/NEJMoa050613
Diener HC Ringleb P Antithrombotic Secondary Prevention After Stroke Curr Treat Options Cardiovasc Med 2002 4 429 440 12194815
Faulkner LA Schauffler HH The effect of health insurance coverage on the appropriate use of recommended clinical preventive services Am J Prev Med 1997 13 453 458 9415792
Saaddine JB Engelgau MM Beckles GL Gregg EW Thompson TJ Narayan KM A diabetes report card for the United States: quality of care in the 1990s Ann Intern Med 2002 136 565 574 11955024
Improvement of Cardiac Outcomes in Kaiser Permanente of Ohio The Permanente Journal 2000 4 78 80
Gaede P Vedel P Larsen N Jensen GV Parving HH Pedersen O Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes N Engl J Med 2003 348 383 393 12556541 10.1056/NEJMoa021778
Faragon JJ Waite NM Hobson EH Seoldo N VanAmburgh JA Migden H Improving aspirin prophylaxis in a primary care diabetic population Pharmacotherapy 2003 23 73 79 12523462 10.1592/phco.23.1.73.31923
Haggerty SA Cerulli J Zeolla MM Cottrell JS Weck MB Faragon JJ Community pharmacy Target Intervention Program to improve aspirin use in persons with diabetes J Am Pharm Assoc (Wash DC) 2005 45 17 22 15730113
Filippi A Sabatini A Badioli L Samani F Mazzaglia G Catapano A Cricelli C Effects of an automated electronic reminder in changing the antiplatelet drug-prescribing behavior among Italian general practitioners in diabetic patients: an intervention trial Diabetes Care 2003 26 1497 1500 12716811
Krein SL Vijan S Pogach LM Hogan MM Kerr EA Aspirin use and counseling about aspirin among patients with diabetes Diabetes Care 2002 25 965 970 12032100
|
16321162
|
PMC1318485
|
CC BY
|
2021-01-04 16:38:10
|
no
|
Cardiovasc Diabetol. 2005 Dec 1; 4:18
|
utf-8
|
Cardiovasc Diabetol
| 2,005 |
10.1186/1475-2840-4-18
|
oa_comm
|
==== Front
Nucl ReceptNuclear Receptor1478-1336BioMed Central London 1478-1336-3-41630955710.1186/1478-1336-3-4ResearchAssociation of common variation in the PPARA gene with incident myocardial infarction in individuals with type 2 diabetes: A Go-DARTS study Doney Alex SF [email protected] Bettina [email protected] Simon P [email protected] Andrew D [email protected] Graham [email protected] Colin NA [email protected] The Institute of Cardiovascular Research, Ninewells Hospital and Medical School, Dundee, DD1 9SY, Scotland, UK2 Biomedical Research Centre, Ninewells Hospital and Medical School, Dundee, DD1 9SY, Scotland, UK3 Division of Medicine and Therapeutics, Ninewells Hospital and Medical School, Dundee, DD1 9SY, Scotland, UK2005 25 11 2005 3 4 4 10 10 2005 25 11 2005 Copyright © 2005 Doney et al; licensee BioMed Central Ltd.2005Doney 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 variants of the PPARA gene have been found to associate with ischaemic heart disease in non diabetic men. The L162V variant was found to be protective while the C2528G variant increased risk. L162V has also been associated with altered lipid measures. We therefore sought to determine the effect of PPARA gene variation on susceptibility to myocardial infarction in patients with type 2 diabetes. 1810 subjects with type 2 diabetes from the prospective Go-DARTS study were genotyped for the L162V and C2528G variants in the PPARA gene and the association of the variants with incident non-fatal myocardial infarction was examined. Cox's proportional hazards was used to interrogate time to event from recruitment, and linear regression for analysing association of genotype with quantitative clinical traits.
Results
The V162 allele was associated with decreased risk of non-fatal myocardial infarction (HR = 0.31, 95%CI 0.10–0.93 p = 0.037) whereas the C2528 allele was associated with increased risk (HR = 2.77 95%CI 1.34–5.75 p = 0.006). Similarly V162 was associated with a later mean age of diagnosis with type 2 diabetes and C2582 an earlier age of diagnosis. C2528 was also associated with increased total cholesterol and LDL cholesterol, which did not account for the observed increased risk. Haplotype analysis demonstrated that when both rare variants occurred on the same haplotype the effect of each was abrogated.
Conclusion
Genetic variation at the PPARA locus is important in determining cardiovascular risk in both male and female patients with diabetes. This genotype associated risk appears to be independent of the effect of these genotypes on lipid profiles and age of diagnosis with diabetes.
==== Body
Background
Dysregulation of fatty acid metabolism plays a pivotal role in the aetiology of type 2 diabetes [1], and explains, at least in part, the association between obesity, type 2 diabetes and cardiovascular disease (CVD). PPARα is a member of the nuclear receptor super-family of ligand-activated transcription factors. Ligands for PPARα include polyunsaturated fatty acids and the fibrate class of lipid-lowering drugs [2]. It is expressed at high levels in several cell types involved in the atherosclerotic process [3], and its activation has beneficial effects on plasma lipids, endothelial function and markers of inflammation [4]. Thus, the PPARA gene is a strong candidate for a genetic determinant of CVD risk in people with type 2 diabetes [5].
The PPARA gene has been screened for common variation [6-8]. The most studied variant is the leucine 162 valine (L162V) polymorphism, present at allele frequencies between 5 and 10%, and situated in the DNA binding domain. Functional studies have demonstrated that the V162 allele is more active in vitro [7,9], and the V162 allele has been associated with altered plasma lipid levels [6,8,10], delayed progression of angiographically determined CV disease in the Lopid Coronary Angiography Trial (LOCAT), and reduced risk of ischemic heart disease in the Second Northwick Park Heart Study (NPHS2) [11]. A second, more common, G→C variant situated in intron 7 (G2528C) is in partial allelic association with the L162V variant and shows opposing effects on cardiovascular risk and cardiac growth [9,11]. Recently it has been demonstrated that haplotypes of these variants in association with a further A→C variant in intron 1 influence age of onset of type 2 and time to requiring insulin [12].
PPARα activators improve the dyslipidemia associated with type 2 diabetes and may be particularly beneficial in lowering risk of CVD in subjects with type 2 diabetes or metabolic syndrome [13]. We therefore investigated the association between PPARA gene variation with risk of CVD and diabetes related traits in Caucasian subjects with type 2 diabetes participating in the prospective population-based Genetics of Diabetes Audit and Research in Tayside Scotland (Go-DARTS) study [14-16].
Results
The clinical characteristics of the genotyped cohort are shown in table 1. The allele frequencies of both polymorphisms were consistent with those previously published for European non-diabetic populations (table 2). The two polymorphisms were both in Hardy-Weinberg equilibrium and were in significant linkage disequilibrium D' = 0.204 p =< 0.00001. Estimated haplotype frequencies indicated very similar values with those previously published (table 3). There was little evidence that the genotypes either singly, or when included in the model together, were associated with blood pressure or lipid measurements (table 4). V162 was associated with a generally more cardioprotective profile with V/V homozygotes having lower systolic and diastolic blood pressure, lower LDL cholesterol and higher HDL cholesterol than L/L homozygotes, although none of these differences were significant. Conversely C2528 homozygotes had a small but significantly higher total cholesterol and calculated LDL cholesterol compared to G/G homozygotes. We also found an association of genotype with age diagnosed with type 2 diabetes, with the V162 allele being associated with a significantly later age of diagnosis and the C2528 allele with a significantly earlier age of diagnosis (Table 5). When we considered haplotypes we found that V162-G2528 was associated with almost a 4 year delay in diagnosis with diabetes compared to the common L162-G2582 haplotype (p = 0.004). This association was completely abrogated when C2528 occurred together with V162 as a haplotype.
Table 1 Clinical characteristics of the Go-DARTS cohort
No of individuals 1810 (54% male)
Age at recruitment (years) 63.1 (9.6)
Age at diagnosis 54.9 (9.0)
Body Mass Index (kg/m2) 30.5 (5.4)
Insulin treatment 839 (44.1%)
Smoking History 958 (50.4%)
Prevalent cerebrovascular disease 67 (3.5%)
Prevalent angina 178 (9.4%)
Previous myocardial infarction 323 (17.0%)
Data shown are mean (SD) for continuous variables and n (%) for categorical variables.
Table 2 PPARA genotype distribution and allele frequencies in the Go-DARTS cohort. The corresponding allele frequencies from the Second Northwick Park Heart Study (NPHS2)11 is shown for comparison.
n n
L162V L/L 1573 G2528C GG 1216
L/V 224 GC 529
V/V 13 CC 64
1810 1809
Go-DARTS allele freq. 0.069 (0.061–0.077) 0.182 (0.169–0.194)
NPHS allele freq. 0.063 0.174
Table 3 Estimated Haplotype frequencies in Go-DARTS. Frequencies in NPHS2 are given for comparison
Haplotype Go-DARTS NPHS2
L162-G2528 0.802 0.804
L162-C2528 0.130 0.132
V162-G2528 0.016 0.021
V162-C2528 0.052 0.041
Table 4 Biochemical parameters at genotyping. Mean and 95% confidence intervals of all readings taken within 2 years prior to enrolment in study
L162V
L/L L/V V/V
BMI 30.5 30.2–30.7 30.7 30.0–31.4 28.9 26.0–31.9
SBP mmHg 142.5 141.7-141.3 141.3 139.3–143.3 134.9 126.7–143.1
DBP mmHg 79.5 79.1–79.8 79.7 78.6–80.7 77.3 73.1–81.5
Cholrat† mmol/L 4.5 4.4–4.6 4.8 4.5–5.0 4.3 3.3–5.3
Chol mmol/L 5.2 5.2–5.3 5.3 5.1–5.4 5.3 4.8–5.8
Trigs mmol/L 2.7 2.6–2.8 2.8 2.6–3.1 2.4 1.3–3.5
HDL mmol/L 1.22 1.20–1.24 1.23 1.20–1.28 1.31 1.13–1.50
LDL mmol/L 2.89 2.84–2.92 2.89 2.78–3.00 2.92 2.42–3.41
G2528C
G/G G/C C/C
BMI 30.5 30.2–30.8 30.4 29.9–30.8 31.40 30.1–32.7
SBP mmHg 142.5 141.7–143.4 142 140.7–143.3 140.2 135.5–143.9
DBP mmHg 79.5 79.0–79.9 79.7 79.0–80.3 78.2 76.3–80.1
Cholrat† mmol/L 4.49 4.37–4.60 4.74 4.57–4.91 4.51 4.03–4.98
Chol mmol/L 5.20 5.15–5.25 5.26 5.18–5.34 5.56 5.33–5.79*
Trigs mmol/L 2.73 2.62–2.83 2.75 2.59–2.91 2.65 2.18–3.12
HDL mmol/L 1.22 1.20–1.24 1.23 1.20–1.25 1.25 1.17–1.34
LDL mmol/L 2.88 2.83–2.92 2.88 2.81–2.95 3.22 3.01–3.44*
*P < 0.05 ANOVA co-dominant model
† Cholrat = Total cholesterol/HDL Ratio
Table 5 Influence of genotype and inferred haplotypes on age diagnosed with type 2 diabetes.
Age diagnosed
Beta 95% CI p
V162* 2.6 0.2–5.1 0.034
C2528† -1.1 -2.0–0.2 0.022
Haplotype
L162-G2582 Ref
L162-C2528 -0.40 -1.25 – 0.45 0.36
V162-G2528 3.89 1.26 – 6.51 0.004
V162-C2528 -0.28 -1.65 – 1.10 0.69
*Co-dominant model
† Dominant model
Both variants included in the model
During a median follow up time 49.6 months there were 108 non-fatal myocardial infarction events and 355 deaths from all causes. In a fully adjusted Cox's proportional hazards model (table 6) we found that V162 was significantly protective against non-fatal myocardial infarction (HR 0.31, 95%CI 0.10–0.93, p = 0.037), while the C2528 variant was associated with a significantly higher risk of non-fatal myocardial infarction (HR 2.77, 95%CI 1.34–5.75, p = 0.006). This association was found to be similar in both sexes. Neither variant demonstrated any evidence of an association with risk of myocardial infarction when considered in isolation. Again, when we considered haplotypes, we found that compared to the haplotype with both common alleles, the haplotype L162-C2528 was associated with a significantly increased cardiovascular risk (HR 1.68 95%CI 1.16–2.43 p = 0.006) and the V162-G2528 a decreased risk although in this case this was not significant (HR 0.54, 95%CI 0.20–1.48, p = 0.23). Again the relative associations of each variant were completely abrogated when they occurred together on the same haplotype. The inclusion of total cholesterol in the model did not attenuate these observed associations but rather further strengthened them (V162: HR 0.28, 95% CI 0.09–0.89, p = 0.031 and C2528: HR 2.87, 95%CI 1.38–5.95, p = 0.005) demonstrating that the increased risk associated with the C2528 was not linked to raised cholesterol levels. When a combined endpoint of death from all cause, and non-fatal myocardial infarction was considered in the same model it was found that the V162 continued to demonstrate a reduced risk of an event although the association was attenuated (HR 0.52, 95%CI 0.28–0.98, p = 0.044). C2528 again demonstrated an increased risk although this was now weak and borderline non-significant (HR 1.52, 95% CI 0.99–2.31, p = 0.052).
Table 6 Prospective model of PPARA variants and non-fatal myocardial infarction risk in the Go-DARTs cohort. A full set of data was available on 1806 individuals, 108 recorded non fatal myocardial infarctions during the period of observation, with a total of 94497.6 months of observation. Both PPARA variants were analysed using a co-dominant model.
Hazard Ratio 95% CI P
V162 0.31 0.10 0.93 0.037
C2528 2.77 1.34 5.75 0.006
Smoking 1.39 0.93 2.10 0.112
Gender 0.72 0.48 1.08 0.107
Age at recruitment 1.05 1.02 1.07 <0.001
Insulin treatment 2.56 1.69 3.89 <0.001
Prevalent angina 5.64 3.80 8.40 <0.001
Prevalent cerebrovascular disease 1.29 0.67 2.51 0.445
Prevalent myocardial infarction 3.90 2.60 5.81 <0.001
Haplotypes
L162-G2582 Ref
L162-C2528 1.68 1.16–2.43 0.006
V162-G2528 0.54 0.20–1.48 0.23
V162-C2528 0.96 0.48–1.94 0.91
Discussion
It has been previously demonstrated that two common variants at the PPARA locus are associated with opposing risks of development of atherosclerotic vascular disease and myocardial infarction in two separate populations of non-diabetic male subjects taking part in the LOCAT and NPHS2 studies [11]. Individuals with type 2 diabetes are however particularly susceptible to atherosclerotic macrovascular disease, and PPARα activators such as the fibrate group of drugs appear to be particularly beneficial in reducing cardiovascular events in this group of patients [13]. In this study we have confirmed the observation that V162 is associated with a decreased risk and the C2528 variant is associated with an increased risk of cardiovascular disease and that this observation can be extended to individuals with type 2 diabetes. We also found that the association is similar in both male and female individuals. Finally we confirm a recent finding that these variants are associated with opposing influences on age of diagnosis with type 2 diabetes [12], and that the C2528 variant is also associated with significantly higher total cholesterol and calculated LDL cholesterol levels.
Several studies have considered the potential clinical importance of genetic variation at the PPARA locus although most have concentrated on lipid levels and have considered the L162V variant in isolation. These studies have been inconsistent indicating that L162V may influence levels of cholesterol or other lipoproteins, depending on the population analysed [6,8,10,12,17,18], while other studies have found no evidence for such an association [19,20]. These inconsistencies may be due in large part to differing environments, genetic background and diseased status (including medications prescribed) between the populations considered. For instance the relative concentration in the diet of saturated to polyunsaturated fat has recently been demonstrated to significantly affect association of L162Vgenotype with lipoprotein profile [21,22]. Furthermore, it is likely that there will be differential usage of fibrate (as well as other lipid modifying) drugs between individuals with type 2 diabetes and non-diabetic populations which may also influence the lipid levels differentially by PPARα haplotype [23,24]. Gene/gene interactions may also be important as evidenced by the observation that variants in the PPARD and APOE genes can influence the observed association [10,20]. The inevitability of gene/environment interactions, and the observed inconsistency between studies, illustrate the difficulties of considering single measures of quantitative traits. Such measures are likely to vary considerably over an individual's lifetime, depending on health status and diet. Importantly, the clinical measures in this study were a mean of multiple measures taken over up to a three year period and therefore represent a limited integration of such temporal fluctuations.
In this study we found that the G2582C variant, but not the L162V variant, influenced lipid levels and this association was not influenced by L162V. The difference between the mean values of LDL cholesterol for GG individuals compared to CC was rather small (0.36 mmol/L) and even in this high risk population did not account for the increased cardiovascular risk associated with C2582. This was not unexpected, as in keeping with the previous studies in non-diabetic men, inclusion of total cholesterol or LDL cholesterol in the model did not affect the association with cardiovascular outcomes, indicating that the increased risk associated with the C2528 variant is not likely to be through its influence on lipid levels.
Few studies have considered cardiovascular disease or considered variants other than the L162V. One recent study also demonstrated a non significant trend towards a protective effect of V162 in individuals with diabetes [19]. This study did not consider the G2825C variant. The present study however confirms that the V162 variant is protective against nonfatal myocardial infarction, while the C2528 variant is associated with an increased risk in this population with type 2 diabetes. These observations also appear to affect overall mortality in this population. This apparent consistency across studies with respect to cardiovascular events probably reflects the small, though global, modulation in phenotype acting across an individual's lifetime due to the slight changes in activity of PPARα associated with each variant. Unlike single measures of lipids or lipoproteins, this is less likely to be effected by temporal environmental changes. This is also likely to be true for clinical events such as age of diagnosis. In the previous study that considered age of diagnosis a further variant in intron 1 of PPARA was used to construct haplotypes with L162V and G2825C [12].
In this study, as in ours, the C2825 was associated with an earlier age of diagnosis, with the V162 allele being associated with protection from early diagnosis, in a manner consistent with the modulation of CVD risk. This is the first study to present the association with age-of-diagnosis of type 2 diabetes in the same population with the association with cardiovascular risk, and we can state that inclusion of age of diagnosis does not modulate the observed associations with cardiovascular risk and vice-versa, demonstrating that these are independent observations. This is not surprising, as PPARA variation is associated with CVD risk regardless of diabetic status.
The G2825C variant is in a non-coding region (intron 7) and therefore unlikely to be a directly causal variant, however several other studies have indicated its potential biological importance [9,11,12] and also a reduced response to fibrate therapy [25]. These observations together with that of PPARα activation through fibrate therapy have given rise to the suggestion that the C2528 variant is associated with reduced RNA transcription and hence lower PPARα levels. Although this suggestion together with a mechanism remains to be demonstrated, it is probably due to a further (as yet) unidentified variant in or near the PPARA gene.
We have previously demonstrated a similar situation in the PPARG locus in which the biological effect of a variant with probable functional consequences is consistently modified in an opposing manner by the presence of a variant for which a function is difficult to ascribe [14,15]. Given the role of the PPARs as master controllers of energy metabolism it is likely, as the accumulating evidence appears to suggest, that they will manifest epistasis and balanced polymorphism allowing for rapid adaptive evolutionary response to widely differing environmental challenges. Indeed, the existence of widespread, balanced variation has recently been suggested for genes involved in complex traits [26].
Conclusion
We have confirmed the potential importance of genetic variation at the PPARA locus in modulating susceptibility to cardiovascular disease, and have shown that this association is relevant to individuals with type 2 diabetes.
Methods
In the Tayside region of Scotland detailed clinical information on all individuals with diabetes mellitus is recorded on a continuously updated electronic clinical information system known as DARTS (Diabetes Audit and Research in Tayside Scotland) [16]. Validated, region-wide electronic record-linkage techniques facilitate the identification of individuals with diabetes in the Tayside population with a sensitivity of 97% [16]. Relevant clinical data is linked to databases containing all inpatient hospital admissions in Tayside from 1980 with diagnostic codes from ICD-9/10 (International Classification of Diseases, ninth and tenth revisions), and records of death certificates from the registrar general. This automated electronic follow-up is manually validated through a continuous cycle of review by dedicated study clinicians. Incident cardiovascular events in this population have been described previously [27].
Following written informed consent from individuals registered on DARTS, blood samples for genetic studies have been collected, thereby creating a genetic sub-study known as Go-DARTS. Rigorous compliance with NHS data protection and encryption standards is maintained and the study was approved by the local research ethics committee.
The PPARα L162V and G2528C genotypes were determined in 1,810 individuals all of whom were Caucasian with type 2 diabetes diagnosed between the age of 35 and 70 years. Taqman (Applied Biosystems) allelic discrimination assays were used. The primers and probes used for the allelic discrimination assays were as follows: L162V Forward primer-CAGAAACAAATGCCAGTATTGTCG, Reverse primer-GGCCACCTTACCTACCGTTGTG, L162 probe (FAM labelled) – TTCACAAGTGCCTTTCTGTCGGGATGT, V162 probe (TET labelled) – TTCACAAGTGCGTTTCTGTCGGGATGT.
G2528C Forward primer-TCCTTAAATATGGTGGAACACTTGAAG, Reverse primer-TCACAACCACCAGTTTTGCAT, G2528 probe (FAM labelled) – ATATCTAGTTTGGATTCAAAAGCTTCATTTCCCA, C2528 probe (TET labelled) – ATATCTAGTTTCGATTCAAAAGCTTCATTTCCCA.
Statistics
For each clinical measure the mean was determined from multiple measures obtained up to a maximum period of three years (up to two years prior to enrolment, and up to one year following enrolment). LDL cholesterol was estimated through the use of the Freidwald equation. Linear regression was used to determine the association of genotype with each measure corrected for age at genotyping. For determining the association of genotype with age of diagnosis, this was corrected for gender and presence of a history of smoking by determining residuals and adding these to the overall mean age of diagnosis. Cox's proportional hazards was used to model time to first event. All individuals were followed from the point of genotyping until a non-fatal myocardial infarction occurred, or a composite of non-fatal myocardial infarct or all cause death. Censoring occurred either at the end of the study or death from any cause. Both variants were included in the model and it was found that a co-dominant model for each variant produced the best fit. The following variables were also included in the model; age at genotyping, gender, history of smoking, treatment with insulin, a previous history of a myocardial infarction, prevalent angina and prevalent cerebrovascular disease. Haplotype frequency estimates together with haplotype effects were determined using the THESIAS Program [28,29]. In the case of survival analysis by haplotype using THESIAS only age at genotyping, prevalent angina and a previous history of a myocardial infarction were included in the model.
List of abbreviations
DARTS Diabetes Audit and Research in Tayside Study
Go-DARTS Genetics of DARTS
PPARα Peroxisome Proliferator Activated Receptor alpha
PPARA Gene for PPARα
PPARD Gene for PPARδ
CVD Cardiovascular disease
LOCAT Lopoid Coronary Angiography Trial
NPHS2 Second Northwick Park Heart Study
HR Hazard Ratio
CI Confidence Interval
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
ASFD and CNAP wrote the manuscript and performed the analysis, CNAP and ADM conceived the study and participated in its design and coordination. GL helped draft the manuscript and contributed to the data analysis. BF and SL performed the genotyping. All authors read and approved the final manuscript.
Acknowledgements
The Go-DARTS recruitment was funded by an anonymous trust donations to Tenovus Tayside. Colin Palmer and Andrew Morris are supported by the Scottish Higher Education Funding Council (SHEFC) and the Scottish Executive Genetic Health Initiative. Go-DARTS is currently supported by the Wellcome Trust.
==== Refs
McGarry JD Banting lecture 2001: dysregulation of fatty acid metabolism in the etiology of type 2 diabetes Diabetes 2002 51 7 18 11756317
Krey G Braissant O L'Horset F Kalkhoven E Perroud M Parker MG Wahli W Fatty acids, eicosanoids, and hypolipidemic agents identified as ligands of peroxisome proliferator-activated receptors by coactivator-dependent receptor ligand assay Mol Endocrinol 1997 11 779 791 9171241 10.1210/me.11.6.779
Braissant O Foufelle F Scotto C Dauca M Wahli W Differential expression of peroxisome proliferator-activated receptors (PPARs): tissue distribution of PPAR-alpha, -beta, and -gamma in the adult rat Endocrinology 1996 137 354 366 8536636 10.1210/en.137.1.354
Marx N Duez H Fruchart JC Staels B Peroxisome proliferator-activated receptors and atherogenesis: regulators of gene expression in vascular cells Circ Res 2004 94 1168 1178 15142970 10.1161/01.RES.0000127122.22685.0A
Vosper H Khoudoli GA Graham TL Palmer CN Peroxisome proliferator-activated receptor agonists, hyperlipidaemia, and atherosclerosis Pharmacol Ther 2002 95 47 62 12163127 10.1016/S0163-7258(02)00232-2
Flavell DM Pineda Torra I Jamshidi Y Evans D Diamond JR Elkeles RS Bujac SR Miller G Talmud PJ Staels B Humphries SE Variation in the PPARalpha gene is associated with altered function in vitro and plasma lipid concentrations in Type II diabetic subjects Diabetologia 2000 43 673 680 10855543 10.1007/s001250051357
Sapone A Peters JM Sakai S Tomita S Papiha SS Dai R Friedman FK Gonzalez FJ The human peroxisome proliferator-activated receptor alpha gene: identification and functional characterization of two natural allelic variants Pharmacogenetics 2000 10 321 333 10862523 10.1097/00008571-200006000-00006
Vohl MC Lepage P Gaudet D Brewer CG Betard C Perron P Houde G Cellier C Faith JM Despres JP Morgan K Hudson TJ Molecular scanning of the human PPARa gene: association of the L162v mutation with hyperapobetalipoproteinemia J Lipid Res 2000 41 945 952 10828087
Jamshidi Y Montgomery HE Hense HW Myerson SG Torra IP Staels B World MJ Doering A Erdmann J Hengstenberg C Humphries SE Schunkert H Flavell DM Peroxisome proliferator--activated receptor alpha gene regulates left ventricular growth in response to exercise and hypertension Circulation 2002 105 950 955 11864924 10.1161/hc0802.104535
Tai ES Demissie S Cupples LA Corella D Wilson PW Schaefer EJ Ordovas JM Association between the PPARA L162V polymorphism and plasma lipid levels: the Framingham Offspring Study Arterioscler Thromb Vasc Biol 2002 22 805 810 12006394 10.1161/01.ATV.0000012302.11991.42
Flavell DM Jamshidi Y Hawe E Pineda Torra I Taskinen MR Frick MH Nieminen MS Kesaniemi YA Pasternack A Staels B Miller G Humphries SE Talmud PJ Syvanne M Peroxisome proliferator-activated receptor alpha gene variants influence progression of coronary atherosclerosis and risk of coronary artery disease Circulation 2002 105 1440 1445 11914252 10.1161/01.CIR.0000012145.80593.25
Flavell DM Ireland H Stephens JW Hawe E Acharya J Mather H Hurel SJ Humphries SE Peroxisome proliferator-activated receptor alpha gene variation influences age of onset and progression of type 2 diabetes Diabetes 2005 54 582 586 15677519
Robins SJ Cardiovascular disease with diabetes or the metabolic syndrome: should statins or fibrates be first line lipid therapy? Curr Opin Lipidol 2003 14 575 583 14624134 10.1097/00041433-200312000-00005
Doney AS Fischer B Leese G Morris AD Palmer CN Cardiovascular risk in type 2 diabetes is associated with variation at the PPARG locus: a Go-DARTS study Arterioscler Thromb Vasc Biol 2004 24 2403 2407 15486307 10.1161/01.ATV.0000147897.57527.e4
Doney AS Fischer B Cecil JE Boylan K McGuigan FE Ralston SH Morris AD Palmer CN Association of the Pro12Ala and C1431T variants of PPARG and their haplotypes with susceptibility to Type 2 diabetes Diabetologia 2004 47 555 558 14730381 10.1007/s00125-003-1323-1
Morris AD Boyle DI MacAlpine R Emslie-Smith A Jung RT Newton RW MacDonald TM The diabetes audit and research in Tayside Scotland (DARTS) study: electronic record linkage to create a diabetes register. DARTS/MEMO Collaboration.[see comment] BMJ 1997 315 524 528 9329309
Lacquemant C Lepretre F Pineda Torra I Manraj M Charpentier G Ruiz J Staels B Froguel P Mutation screening of the PPARalpha gene in type 2 diabetes associated with coronary heart disease Diabetes & Metabolism 2000 26 393 401 11119019
Jamshidi Y Flavell DM Hawe E MacCallum PK Meade TW Humphries SE Genetic determinants of the response to bezafibrate treatment in the lower extremity arterial disease event reduction (LEADER) trial Atherosclerosis 2002 163 183 192 12048138 10.1016/S0021-9150(02)00002-3
Gouni-Berthold I Giannakidou E Muller-Wieland D Faust M Kotzka J Berthold HK Krone W Association between the PPARalpha L162V polymorphism, plasma lipoprotein levels, and atherosclerotic disease in patients with diabetes mellitus type 2 and in nondiabetic controls American Heart Journal 2004 147 1117 1124 15199365 10.1016/j.ahj.2003.12.005
Skogsberg J Kannisto K Cassel TN Hamsten A Eriksson P Ehrenborg E Evidence that peroxisome proliferator-activated receptor delta influences cholesterol metabolism in men Arterioscler Thromb Vasc Biol 2003 23 637 643 12615676 10.1161/01.ATV.0000064383.88696.24
Paradis AM Fontaine-Bisson B Bosse Y Robitaille J Lemieux S Jacques H Lamarche B Tchernof A Couture P Vohl MC The peroxisome proliferator-activated receptor alpha Leu162Val polymorphism influences the metabolic response to a dietary intervention altering fatty acid proportions in healthy men American Journal of Clinical Nutrition 2005 81 523 530 15699244
Tai ES Corella D Demissie S Cupples LA Coltell O Schaefer EJ Tucker KL Ordovas JM Polyunsaturated fatty acids interact with the PPARA-L162V polymorphism to affect plasma triglyceride and apolipoprotein C-III concentrations in the Framingham Heart Study J Nutr 2005 135 397 403 15735069
Bosse Y Pascot A Dumont M Brochu M Prud'homme D Bergeron J Despres JP Vohl MC Influences of the PPAR alpha-L162V polymorphism on plasma HDL(2)-cholesterol response of abdominally obese men treated with gemfibrozil Genetics in Medicine 2002 4 311 315 12172398 10.1097/00125817-200207000-00010
Brisson D Ledoux K Bosse Y St-Pierre J Julien P Perron P Hudson TJ Vohl MC Gaudet D Effect of apolipoprotein E, peroxisome proliferator-activated receptor alpha and lipoprotein lipase gene mutations on the ability of fenofibrate to improve lipid profiles and reach clinical guideline targets among hypertriglyceridemic patients Pharmacogenetics 2002 12 313 320 12042669 10.1097/00008571-200206000-00007
Foucher C Rattier S Flavell DM Talmud PJ Humphries SE Kastelein JJ Ayyobi A Pimstone S Frohlich J Ansquer JC Steiner G investigators D Response to micronized fenofibrate treatment is associated with the peroxisome-proliferator-activated receptors alpha G/C intron7 polymorphism in subjects with type 2 diabetes Pharmacogenetics 2004 14 823 829 15608561 10.1097/00008571-200412000-00005
Kroymann J Mitchell-Olds T Epistasis and balanced polymorphism influencing complex trait variation Nature 2005 435 95 98 15875023 10.1038/nature03480
Evans JM Wang J Morris AD Comparison of cardiovascular risk between patients with type 2 diabetes and those who had had a myocardial infarction: cross sectional and cohort studies.[erratum appears in BMJ 2002 Jun 8;324(7350):1357] BMJ 2002 324 939 942 11964337 10.1136/bmj.324.7343.939
Tregouet DA Escolano S Tiret L Mallet A Golmard JL A new algorithm for haplotype-based association analysis: the Stochastic-EM algorithm Ann Hum Genet 2004 68 165 177 15008795 10.1046/j.1529-8817.2003.00085.x
Tregouet DA Tiret L Cox proportional hazards survival regression in haplotype-based association analysis using the Stochastic-EM algorithm Eur J Hum Genet 2004 12 971 974 15241485 10.1038/sj.ejhg.5201238
|
16309557
|
PMC1318486
|
CC BY
|
2021-01-04 16:37:44
|
no
|
Nucl Recept. 2005 Nov 25; 3:4
|
utf-8
|
Nucl Recept
| 2,005 |
10.1186/1478-1336-3-4
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-1351628393310.1186/1465-9921-6-135ResearchProinflammatory cytokine responses induced by influenza A (H5N1) viruses in primary human alveolar and bronchial epithelial cells Chan MCW [email protected] CY [email protected] WH [email protected] SW [email protected] JM [email protected] YO [email protected] RWY [email protected] HT [email protected] LLM [email protected] Y [email protected] JSM [email protected] Department of Microbiology, The University of Hong Kong, Queen Mary Hospital, Hong Kong Special Administrative Region of China2 Department of Cardiothoracic Surgery, Grantham Hospital, Wong Chuk Hang, Aberdeen, Hong Kong Special Administrative Region of China3 Department of Anatomy, The University of Hong Kong, Pokfulam, Hong Kong Special Administrative Region of China4 Department of Pathology, The University of Hong Kong, Queen Mary Hospital, Hong Kong Special Administrative Region of China5 National Institute of Hygiene and Epidemiology, Hanoi, Vietnam2005 11 11 2005 6 1 135 135 16 6 2005 11 11 2005 Copyright © 2005 Chan et al; licensee BioMed Central Ltd.2005Chan 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
Fatal human respiratory disease associated with influenza A subtype H5N1 has been documented in Hong Kong, and more recently in Vietnam, Thailand and Cambodia. We previously demonstrated that patients with H5N1 disease had unusually high serum levels of IP-10 (interferon-gamma-inducible protein-10). Furthermore, when compared with human influenza virus subtype H1N1, the H5N1 viruses in 1997 (A/Hong Kong/483/97) (H5N1/97) were more potent inducers of pro-inflammatory cytokines (e.g. tumor necrosis factor-a) and chemokines (e.g. IP-10) from primary human macrophages in vitro, which suggests that cytokines dysregulation may play a role in pathogenesis of H5N1 disease. Since respiratory epithelial cells are the primary target cell for replication of influenza viruses, it is pertinent to investigate the cytokine induction profile of H5N1 viruses in these cells.
Methods
We used quantitative RT-PCR and ELISA to compare the profile of cytokine and chemokine gene expression induced by H5N1 viruses A/HK/483/97 (H5N1/97), A/Vietnam/1194/04 and A/Vietnam/3046/04 (both H5N1/04) with that of human H1N1 virus in human primary alveolar and bronchial epithelial cells in vitro.
Results
We demonstrated that in comparison to human H1N1 viruses, H5N1/97 and H5N1/04 viruses were more potent inducers of IP-10, interferon beta, RANTES (regulated on activation, normal T cell expressed and secreted) and interleukin 6 (IL-6) in primary human alveolar and bronchial epithelial cells in vitro. Recent H5N1 viruses from Vietnam (H5N1/04) appeared to be even more potent at inducing IP-10 than H5N1/97 virus.
Conclusion
The H5N1/97 and H5N1/04 subtype influenza A viruses are more potent inducers of proinflammatory cytokines and chemokines in primary human respiratory epithelial cells than subtype H1N1 virus. We suggest that this hyper-induction of cytokines may be relevant to the pathogenesis of human H5N1 disease.
avianchemokinesIP-10pathogenesis
==== Body
Background
Influenza pandemics arise from genetic reassortment between avian and human influenza viruses or alternatively by the direct adaptation of a avian influenza viruses to efficient human-to-human transmission [1]. Avian influenza A subtype H5N1 transmitted from poultry to humans in Hong Kong in 1997 (H5N1/97) causing fatal human respiratory disease [2,3]. The subsequent re-emergence of human H5N1 disease in southern China [4], Vietnam [5], Thailand and Cambodia [6] has raised the specter of a new influenza pandemic. While human-to-human transmission of the H5N1 subtype influenza virus appears to be inefficient so far, the disease has exceptional severity in those affected with reported mortality rates ranging from 33% in Hong Kong in 1997 to 55% in Thailand and Vietnam in 2004. The reasons for this unusual severity of human disease have remained unclear.
While dissemination outside the respiratory tract was not demonstrated in human H5N1 disease in 1997 and 2003 [4,7], there is some evidence that more recent H5N1 viruses may occasionally disseminate to multiple organs contributing to unusual disease manifestations such as meningo-encephalitis [8]. However, most patients with H5N1 disease had a primary viral pneumonia complicated by the syndromes of acute respiratory distress and multiple organ dysfunction [4-7,9] with lymphopenia and haemophagocytosis being notable findings. The syndromes of acute respiratory distress and multiple organ dysfunction as well as haemophagocytosis have previously been associated with cytokine dysregulation [10,11].
Influenza virus infection of blood-monocyte-derived murine and human [12,13] macrophages and porcine alveolar macrophages [14] have been shown to result in induction of pro-inflammatory cytokines. Furthermore, we have previously demonstrated that, when compared to human H1N1 and H3N2 influenza viruses, infection of H5N1/97-like viruses lead to the hyper-induction of proinflammatory cytokines in human primary macrophage cultures in vitro [12]. We also reported that patients with H5N1 disease have unusually high serum concentrations of chemokines IP-10 (interferon-gamma-inducible protein-10) and MIG (monokine induced by interferon γ) [4]. We have therefore hypothesized that this differential hyper-induction of cytokines and chemokines may contribute to the unusual severity of human H5N1 disease [4,12].
While macrophages are a key sentinel cell of the immune system and are permissive to influenza virus replication, the primary target cell for the virus are respiratory epithelial cells [15]. In primates experimentally infected with H5N1/97 virus, the type I and II pneumocytes and alveolar macrophages were found to contain viral antigen [16]. Virus infection of alveolar pneumocytes was also demonstrated in the lung of a patient with fatal H5N1 disease [17]. Human alveolar epithelial cells are vital for the maintenance of lung function and the pulmonary air-blood barrier. In addition, human respiratory epithelial cells respond to viral infections by mounting a cytokine response that contributes both to the innate and adaptive host defenses [18]. Furthermore, type II pneumocytes express class II major histocompatibility complex (MHC) molecules in vivo [19]. Expression of class II MHC is usually limited to specialized cells of the immune system whose role is to present foreign antigen to helper T cells [20,21]. The expression of these molecules on alveolar epithelial cells is likely to be of relevance to the adaptive immune response. Therefore it is important to study cytokine responses induced by infection of epithelial cells with influenza viruses including H5N1 viruses.
Human influenza A viruses have been previously reported to induce interleukin 6 (IL-6), interleukin 8 (IL-8) and RANTES (regulated on activation, normal T cell expressed and secreted) in vitro from the transformed bronchial epithelial cell line (NCI-H292) [18]. However, the physiological relevance of findings from transformed cell lines is uncertain and primary alveolar epithelial cell cultures would be a more relevant model [22]. Here, we have compared the cytokine profiles induced by H5N1/97 and H1N1 viruses in human primary type II pneumocytes and bronchial epithelial cells in vitro to test the hypothesis that H5N1/97 and H5N1/04 viruses differentially hyper-induce pro-inflammatory cytokines in respiratory epithelial cells.
Materials and methods
Viruses
An influenza virus isolated from a patient with fatal influenza A H5N1 disease in Hong Kong in 1997, A/Hong Kong/483/97 (H5N1/97), viruses from patients with H5N1 disease in Vietnam in 2004, A/Vietnam/1194/04 and A/Vietnam/3046/04 (both abbreviated as H5N1/04) and a human H1N1 virus A/Hong Kong/54/98 (H1N1) were studied. Viruses were initially isolated in Madin-Darby canine kidney (MDCK) cells. They were cloned by limiting dilution, and seed virus stocks were prepared in MDCK cells. Virus infectivity was assessed by titration of tissue culture infection dose 50% (TCID50) in MDCK cells. The H5N1 influenza viruses used in this study were handled in a BL3 biocontainment facility.
Cells
Primary human bronchial epithelial cells (NHBE) were obtained from Cambrex Bio Science (Walkersville, Inc., Maryland, USA). NHBE cells were grown according to the suppliers instructions in serum-free and hormone supplemented bronchial epithelial growth media (BEGM) which included supplements of 13 g/l bovine pituitary extract, 0.5 g/l hydrocortisone, 0.5 mg/l human recombinant epidermal growth factor, 0.5 g/l epinephrine, 10 g/l transferrin, 5 g/l insulin, 0.1 mg/l retinoic acid, 6.5 mg/l 3,3',5-triiodo-L-thryonine, 50 g/l gentamicin, and 50 mg/l amphotericin B (Cambrex Bio Science, Walkersville, Inc., Maryland, USA). Medium was changed daily starting from the day after seeding. Cells reached confluency in approximately 9 to 10 days, and nearly confluent cells were subcultured using trypsin/EDTA (Cambrex) at a ratio of 1:5. Experiments were carried out on the same batch of cells at passage 3 to 4. The cells were incubated in a humidified atmosphere (5% CO2, 37°C) under liquid-covered conditions.
Primary human alveolar epithelial cells (type II pneumocytes) were isolated from human non-tumor lung tissue obtained from 13 patients (mean age 65 yr [range, 46–77 yr], 10 males and 3 females) undergoing lung resection in Grantham Hospital, Hong Kong. The research protocol was approved by the ethics committee of the University of Hong Kong and Hospital Authority Hong Kong West Cluster. Human type II pneumocytes were isolated using a modification of the methods previously described [19,23]. Briefly, after removing visible bronchi, the lung tissue was chopped into pieces of >0.5 mm thickness using a tissue chopper, washed with balanced salt solution (BSS, 137 mM NaCl, 5 mM KCl, 0.7 mM Na2HPO4, 10 mM HEPES, 5.5 mM glucose, pH 7.4) for 30 min at 37°C three times to partially remove macrophages and blood cells. The tissue was digested using a combination of trypsin (0.5%, GIBCO BRL, Gaithersburg, MD, USA) and elastase (2 units/ml, Worthington Biochemical Corporation, Lakewood, NJ, USA) twice for 15 min at 37°C in a shaking water-bath. The partially digested tissue was minced in the presence of 40% fetal bovine serum (FBS) in DMEM/F12 medium and DNase I (350 units/ml) (GIBCO BRL, Gaithersburg, MD, USA), and cell clumps dispersed by repeatedly pipetting the cell suspension for 10 minutes. After filtration through gauze and a 40 μm cell strainer to ensure a single cell suspension, the cells were incubated with a 1:1 mixture of DMEM/F12 medium and small airway growth medium (SAGM, Cambrex Bio Science Walkersville, Inc., Maryland, USA) containing 5% FBS and 350 units/ml DNase I, on tissue-culture treated plastic Petri dishes in a humidified incubator (5% CO2, 37°C) for 2 hours in order to let macrophage attach on the plastic surface. The non-adherent cells were layered on a discontinuous Percoll density gradient (densities 1.089 and 1.040 g/ml) and centrifuged at 25 × g for 20 min. The cell layer at the interface of the two gradients was collected and washed four times with BSS to remove the Percoll. To remove remaining alveolar macrophages, the cell suspension was incubated with magnetic beads coated with anti-CD-14 antibodies at room temperature for 20 min under constant mixing. After the removal of the beads using a magnet and assessment of cell viability by trypan-blue exclusion, the purified type II pneumocyte suspension was suspended in SAGM supplemented with 1% FBS, 100 units/ml penicillin and 100 μg/ml streptomycin, and plated at a cell density of 300,000 cells/cm2. The cells were maintained in a humidified atmosphere (5% CO2, 37°C) under liquid-covered conditions, and growth medium was changed daily starting from 60 hours after plating the cells.
Characterization of human type II pneumocytes
Staining for alkaline phosphatase
Human type II pneumocytes were identified by staining for alkaline phosphatase. Freshly isolated cells were spun down on glass slides, air-dried, and stained for 20 min at room temperature. The stain was prepared by dissolving 10 mg naphthol AS bi-phosphate (Sigma) in 40 μl DMSO and was diluted in 10 ml of 0.125 M 2-amino-2-methyl propanol buffer (pH 8.9, Sigma) containing 10 mg fast red (Sigma). The slide was washed and counterstained in 1% methylene green (Sigma) for 30 seconds and was mounted in aqueous medium [19].
Transmission electron microscopy
Cells were fixed in 2% glutaraldehyde (Electron Microscopy Sciences, Washington, PA, USA), washed three times in phosphate buffered saline and serially dehydrated in acetone. The tissue was post-fixed in 1% osmium tetroxide and embedded in an Araldite resin (Polysciences, Inc., Washington, PS, USA). Semi-thin sections (1 μm) were cut using an ultra-microtome (Reichert Ultracut S, Leica Aktiengesellscharft, Wien, Australia) with a diamond knife and were stained with toluidine blue for light microscopic examination. Ultra-thin sections (80 nm) mounted on copper grids were electron contrasted with uranyl acetate (1.5 hours, 30°C, Electron Microscopy Sciences) and lead citrate (40 minutes, 20°C, Electron Microscopy Sciences, Washington, PA, USA), and were examined with a transmission electron microscope (EM 208S, FEI Company, Hillsboro, Oregon, USA).
Flow cytometry
The expression of cell surface antigen was measured by staining purified type II pneumocytes with optimal dilution of rabbit anti-human surfactant protein-C (SP-C) (Upstate, Lake Placid, NY, USA) monoclonal antibodies (24°C, 30 minutes) followed by a fluorescein isothiocyanate (FITC-conjugated goat anti-mouse IgG antibody; Sigma, F-0257, 24°C, 30 minutes). Each cell preparation was also stained with antibody specific for monocyte/macrophage surface antigen (CD14 conjugated with FITC, MCA2185F; Serotec. Oxford, UK). The cells were examined by the flow cytometry (FACSSCalibur; Becton Dickinson), and the FITC-stained cells were detected by measuring green light emitted at 530 nm (FL1 channel). The percentage of cells expressing the epithelial and macrophage makers were determined.
Influenza virus infection of type II pneumocytes and bronchial epithelial cells
Human type II pneumocytes and bronchial epithelial cells (seeded at 1 × 106 cells per well in 24-well tissue-culture plates) were infected at a multiplicity of infection (MOI) of two unless otherwise indicated. After 60 min of virus adsorption, the virus inoculum was removed, and the cells were washed with warm culture medium (SAGM for type II pneumocytes and BEBM for bronchial epithelial cells) and incubated in medium supplemented with 0.6 mg/L penicillin, 60 mg/L streptomycin, and 2 mg/L N-p-tosyl-L-phenylalanine chloromethyl ketone-treated-trypsin (Sigma, St Louis, MO, USA). Aliquots of culture supernatant were collected and frozen at -80°C for subsequent virus titration and cytokine analysis. The supernatants were titrated on MDCK cells and the viral titre was quantitated as log10TCID50/ml. RNA was extracted from cells for analysis of cytokine gene expression. Ten hours after infection, replicate cell monolayers were fixed and analyzed by immuno-fluorescent staining specific for influenza virus nucleoprotein (DAKO Imagen, Dako Diagnostics Ltd, Ely, UK) to determine the proportion of cells infected.
Quantification of cytokine mRNA by real-time quantitative RT-PCR
DNase-treated total RNA was isolated by means of RNeasy Mini kit (Qiagen, Hilden, Germany). The cDNA was synthesized from mRNA with poly(dT) primers and Superscript II reverse transcriptase (Life Technologies, Rockville, MD, USA) and quantified by real-time PCR analysis with a LightCycler (Roche, Mannheim, Germany). The mRNA for IP-10, interferon beta, IL-6, RANTES and tumor necrosis factor (TNF) alpha were quantitated using real-time RT-PCR. The oligonucleotide primers and methods used for real-time quantification of cytokines, viral matrix gene and the housekeeping gene product γ-actin mRNA have been described previously [12,24].
Quantification of cytokine proteins by ELISA
The concentrations of IP-10, RANTES, interleukin 6 and interferon beta proteins in the primary human bronchial and alveolar epithelial cell supernatants were measured by a specific ELISA assay (R&D Systems, Minneapolis, MN, USA). Samples of culture supernatant were irradiated with ultraviolet light (CL-100 Ultra Violet Cross linker) for 15 min to inactivate any infectious virus before the ELISA assays were done. Previous experiments had confirmed that the dose of ultraviolet light used did not affect cytokine concentration as measured by ELISA (data not shown).
Statistical analysis
The quantitative cytokine and chemokine mRNA and protein expression profile were compared using one-way ANOVA, followed by Bonferroni multiple-comparison test. Differences were considered significant at p < 0.05.
Results
In vitro infection of human type II pneumocytes
Primary human type II pneumocyte yields were 3.5 ± 0.9 × 106 cells/g lung tissue at 92 ± 5% cell purity as demonstrated by the expression of the type II pneumocyte specific marker surfactant protein C (SP-C), lack of the monocyte/macrophage cell surface antigen (CD14) (Fig. 1A and 1B), and by staining for alkaline phosphatase activity. The contaminating cells were predominantly fibroblasts with monocyte/macrophage cells being less than 2%. Cell viability was 91 ± 7% (n = 13). Differences in age and sex of the lung donor had no apparent effects on cell isolation yields and the performance of the cells in culture. The isolated cells spread to form a confluent monolayer, exhibiting protruding nuclei surrounded by thin cytoplasmic extensions. The identity of the cells in culture as human type II pneumocytes was confirmed by demonstrating the presence of lamellar bodies and microvilli by thin section electron microscopy (Figure 2).
Figure 1 (A) Primary human type II pneumocytes were stained with antibody surfactant protein-C (shaded curve) and control antibody (unshaded curve) to confirm their identity. (B) Human type II pneumocytes isolated were stained with anti-CD14 FITC-conjugated antibodies (shaded curve) specific for macrophage surface antigen to check for any contaminant macrophage.
Figure 2 Transmission electron micrographs of human type II pneumocytes cultured in vitro (A) and the lamellar bodies in the cytoplasm demonstrated using higher magnification (B) (Bars: 1 μm and 50 nm respectively). The cells were scraped off the culture flask, fixed in 2% glutaraldehyde and embedded in Araldite resin.
Previous studies have demonstrated that avian influenza viruses can infect human airway epithelial cells [25]. We first wanted to determine whether alveolar epithelial cells that constitutively reside in the lung can be infected with avian and human influenza viruses in vitro. The cells were infected with influenza A subtypes H5N1 (483/97, 1194/04 and 3046/04) and H1N1 (54/98) at a MOI of 2 and the proportion of cells expressing influenza A virus protein was analyzed at 10 hours post-infection by immunofluorescent staining using an antibody specific for the virus nucleoprotein (DAKO Imagen, Dako Diagnostics Ltd, Ely, UK). Similar proportions (93–100%) of type II pneumocytes infected with H5N1 and H1N1 virus had evidence of viral antigen (nucleoprotein) (Figure 4A). The quantification of influenza M-gene copies at 3 and 6 hours after infection in cells infected with H5N1 and H1N1 viruses showed comparable results at 3 and 6 hours post-infection (Figure 4B). Similarly, the infectious viral yield at 24 and 48 hours post-infection from alveolar epithelial cells infected with H5N1 and H1N1 viruses were not significantly different (Figure 4C).
Figure 4 Infection of human type II pneumocytes with human influenza viruses. (A) Purified alveolar epithelial cells were fixed and analyzed by immunofluorescent staining specific for influenza virus nucleoprotein (×150). (B) The influenza M-gene mRNA profiles were assayed after infection. The concentrations of M-gene mRNA were normalized to those of β-actin mRNA in the corresponding sample. Means of duplicate assays are shown. (C) Alveolar epithelial cells were infected with human influenza viruses and the infectious virus yield (log10TCID50/ml) was determined in aliquots of supernatant collected at various times. Data are the means and the standard errors of independent experiments from three separate donors.
Induction of pro-inflammatory cytokines and chemokines in type II pneumocytes
We investigated the cytokine induction profile induced by H1N1 and H5N1 viruses in primary human type II pneumocytes. Specifically, we also wanted to determine if the two viruses differed qualitatively or quantitatively in the profile of cytokines induced. The mRNA of several cytokines and chemokines were quantified using quantitative RT-PCR at 3 hr and 6 hr post-infection (Table I). The mRNA levels of IP-10, interferon beta, RANTES and IL-6 were significantly up-regulated by influenza virus when compared with the mock infected cells, the genes for IP-10 and interferon beta being the most highly induced. There was no detectable TNF alpha induction in these epithelial cells (data not shown). Inactivation of the virus by ultraviolet irradiation prior to infection of the alveolar epithelial cells abolished cytokine induction (data not shown) suggesting that virus replication was required for cytokine induction.
Table 1 mRNA profile of cytokine and chemokine gene expression of primary culture of human type II pneumocytes 3 h and 6 h after infection with A/Hong Kong/483/97 (H5N1/97), A/Vietnam/1194/04, A/Vietnam/3046/04 (both H5N1/04) and A/Hong Kong 54/98 (H1N1) influenza viruses denoted as fold increase compared to mock infected cells.
Gene products Ratio of expression over mock-infected cells
3 hours post infection 6 hours post infection
483/97 (H5N1/97) (-■-)c 1194/04 (H5N1/04) (-◆-)c 3046/04 (H5N1/04) (-×-)c 54/98 (H1N1) (-▲-)c 483/97 (H5N1/97) (-■-)c 1194/04 (H5N1/04) (-◆-)c 3046/04 (H5N1/04) (-×-)c 54/98 (H1N1) (-▲-)c
Interleukin 1 -1.2 0.3 0.9 1.1 1 0.9 0.8 -1.3
Interleukin 6 9.3*a 15.1*b 9.9*a 7.4*a 17.4*b 19.2*b 15.4*b 8.8*a
Interleukin 8 1 0.9 1.1 -1.2 -1.2 1.6 1.3 1.3
MCP-1 1.1 0.8 1.5 1 -1.2 1.7 1.3 1.3
TNF alpha Not detectable Not detectable
RANTES -1 9.5*a 2.2* 1.55 18.7*b 24.1*b 16.9*b 6.9*a
Interferon-alpha 1.1 0.8 0.6 1.2 -1.3 1.3 0.9 1.2
Interferon-beta 3.7*a 8.5*a 4.7*a -0.8 22.1*b 26.3*b 18.7*b 13.3*b
IP-10 3.9*a 7.9*a 6.3*a -3.5 37.9*b 46.8*b 29.7*b 8.1*a
Signals were normalized to the housekeeping gene, β-actin and expressed as a ratio over mock infected cells.
*Upregulation by two or more times over that of mock infection.
a p < 0.01 and b p < 0.001 (Bonferroni multiple-comparison test).
c Corresponding character symbols as shown in Figure 3 and 4.
When compared with human H1N1 influenza virus, the H5N1/97 and H5N1/04 viruses differentially up-regulated the transcription of IP-10, interferon beta, RANTES and IL-6 to significantly higher levels (p < 0.001) (Figure 5). These differences were not explainable by a difference in proportion of cells infected as indicated by immunofluorescence for viral antigen or differences in virus titre (Figure 4). Furthermore, an increase in the multiplicity of infection of 54/98 (H1N1) virus from 2 to 10 did not result in cytokine mRNA concentrations similar to those induced by H5N1/97 and H5N1/04 (data not shown).
Figure 5 Cytokine and chemokine gene expression profile of influenza-virus-infected human type II pneumocytes by quantitative RT-PCR. Cytokine and chemokine mRNA concentration were assayed 3 h and 6 h after infection with A/Hong Kong/483/97 (H5N1/97), A/Vietnam/1194/04, A/Vietnam/3046/04 (both H5N1/04) and A/Hong Kong 54/98 (H1N1) influenza viruses or in mock infected cells. H5N1/97 and both H5N1/04 influenza viruses induced significantly higher levels of IP-10, interferon-beta, RANTES and IL-6 when compared to H1N1 infected cells at 6 hours post-infection (p < 0.001, Bonferroni multiple comparison test). The mRNA concentrations of cytokine and chemokine mRNA were normalized to those β-actin mRNA in the corresponding samples. Means and standard deviation from experiments from five different donors are shown
Broadly, there were two patterns of kinetics of cytokine gene transcription. Cytokines up-regulated from 3 hr post-infection onwards included IP-10, interferon beta and IL-6 whereas RANTES mRNA was only up-regulated at 6 hr post-infection (Table 1). The observations remained valid whether the cytokine mRNA expression data were analyzed with or without normalization for γ-actin mRNA concentrations.
Infection and cytokine induction profile of primary human bronchial epithelial cells
The cytokine and chemokine profiles induced by H1N1, H5N1/97 and H5N1/04 viruses in primary human bronchial epithelial cells were similarly investigated. The identity of the cells in culture as human bronchial epithelial cells was confirmed by thin section electron microscopy (Figure 3). The overall gene expression profile was comparable to that seen with type II pneumocytes. The M-gene transcript copy numbers (Figure 6A) and infectious viral yields (Figure 6B) from bronchial epithelial cells infected with H5N1 and H1N1 viruses at an MOI of 2 were comparable. The H5N1/97 and H5N1/04 viruses differentially up-regulated the transcription of IP-10, interferon beta, RANTES and IL-6 to significantly higher levels than the human H1N1 virus (p < 0.001 for IP-10, RANTES and IL-6 and p < 0.01 for interferon beta) (Figure 7). In addition, the two H5N1/04 viruses (1194/04 and 3046/04) differentially up-regulated the transcription of monocyte chemotactic protein 1 (MCP-1) and IL-8 to significantly higher levels than the human H1N1 and H5N1/97 viruses (p < 0.05). None of the viruses induced TNF alpha in these cells.
Figure 3 Transmission electron micrographs of human bronchial epithelial cells in vitro at low (A) and high (B) magnification (Bars: 2 μm and 0.5 μm respectively). The cells were scraped off the culture flask, fixed in 2% glutaraldehyde and embedded in Araldite resin.
Figure 6 Infection of human bronchial epithelial cells with human influenza viruses. (A) The influenza M-gene mRNA profiles were assayed after infection. The concentrations of M-gene mRNA were normalized to those of β-actin mRNA in the corresponding sample. Means of duplicate assays are shown. (B) Virus yields (log10TCID50/ml) were determined in aliquots of supernatant collected from influenza-infected bronchial epithelial cells at various times. Data are the means and the standard errors of two independent experiments.
Figure 7 Cytokine and chemokine gene expression profile of influenza-virus-infected human bronchial epithelial cells by quantitative RT-PCR. Cytokine and chemokine mRNA concentration were assayed 3 h and 6 h after infection with A/Hong Kong/483/97 (H5N1/97), A/Vietnam/1194/04, A/Vietnam/3046/04 (both H5N1/04) and A/Hong Kong 54/98 (H1N1) influenza viruses or in mock infected cells. When compared with H1N1 infected cells, H5N1/97 and both H5N1/04 influenza viruses significantly up-regulated IP-10, RANTES and IL-6 (p < 0.001) and interferon beta (p < 0.01) at 6 hours post-infection (Bonferroni multiple comparison test). Both H5N1/04 viruses significantly up-regulated MCP-1 and IL-8 to levels higher than H1N1 and H5N1/97 infected cells (p < 0.05, Bonferroni multiple comparison test). The mRNA concentrations of cytokine and chemokine mRNA were normalized to those β-actin mRNA in the corresponding samples. Means and standard deviation of duplicate cultures and assays are shown.
Secretion of cytokine proteins from bronchial and alveolar epithelial cells
To confirm that the observed differences of mRNA are reflected in levels of cytokine and chemokine secreted, the concentrations of the IP-10, RANTES, interleukin 6 and interferon-beta proteins were measured by ELISA in culture supernatants of infected bronchial and alveolar epithelial cells. The amount of IP-10 and IL-6 secreted by bronchial and alveolar epithelial cells infected with all three H5N1 viruses at 24 hours post infection were significantly higher (p < 0.01) than that secreted by cells infected with H1N1 virus (Figure 8 and 9). At 24 hours post infection, levels of IP-10 induced by H5N1/97 and both H5N1/04 viruses were comparable. However, at 6 hours post-infection, the recent H5N1/04 viruses 1194/04 and 3046/04 appeared to be even more potent at inducing IP-10 than H5N1/97 virus (p < 0.05) (Figure 8). RANTES protein secreted from bronchial and alveolar epithelial cells in response to H5N1/97 and 1194/04 (H5N1/04) were significantly higher than that induced by H1N1 virus. Although the level RANTES mRNA in 3046/04 (H5N1/04) infected cells at 6 hours post infection was significantly higher than those H1N1 infected cells, the RANTES protein secreted by these cells at 24 hours post infection was only increased 4 fold (p = 0.062; not significant) (Figure 5 and 10). We failed to detect any interferon-beta proteins secreted from the supernatants of bronchial and alveolar epithelial cells after influenza viruses infection (data not shown) but it should be noted that the limit of detection of the interferon-beta ELISA was high (250 pg/ml).
Figure 8 IP-10, Interleukin-6 and RANTES production by primary human bronchial and alveolar epithelial cells infected with A/Hong Kong/483/97 (H5N1/97), A/Vietnam/1194/04, A/Vietnam/3046/04 (both H5N1/04) and A/Hong Kong 54/98 (H1N1) influenza viruses or in mock infected cells. Culture supernatants from influenza virus-infected human respiratory epithelial cells collected at 3 h, 6 h and 24 h after infection with H5N1 and H1N1 viruses were tested by ELISA for IP-10 (Figure 8), Interleukin-6 (Figure 9) and RANTES (Figure 10). The IP-10, Interleukin-6 and RANTES mRNA levels were assayed at 3 h and 6 h post infection (data not shown) with results comparable with that shown in figure 5 and 7. The results from bronchial epithelial cells represent the means and standard deviations of three independent experiments (from the same donor). The means and standard deviations of the results from alveolar epithelial cells are based on experiments from six separate donors. * indicates p < 0.01 compared with mock and ** indicates p < 0.05 compared with H5N1/97 and H1N1 infected cells using the Bonferroni multiple comparison test.
Figure 9 IP-10, Interleukin-6 and RANTES production by primary human bronchial and alveolar epithelial cells infected with A/Hong Kong/483/97 (H5N1/97), A/Vietnam/1194/04, A/Vietnam/3046/04 (both H5N1/04) and A/Hong Kong 54/98 (H1N1) influenza viruses or in mock infected cells. Culture supernatants from influenza virus-infected human respiratory epithelial cells collected at 3 h, 6 h and 24 h after infection with H5N1 and H1N1 viruses were tested by ELISA for IP-10 (Figure 8), Interleukin-6 (Figure 9) and RANTES (Figure 10). The IP-10, Interleukin-6 and RANTES mRNA levels were assayed at 3 h and 6 h post infection (data not shown) with results comparable with that shown in figure 5 and 7. The results from bronchial epithelial cells represent the means and standard deviations of three independent experiments (from the same donor). The means and standard deviations of the results from alveolar epithelial cells are based on experiments from six separate donors. * indicates p < 0.01 compared with mock and ** indicates p < 0.05 compared with H5N1/97 and H1N1 infected cells using the Bonferroni multiple comparison test.
Figure 10 IP-10, Interleukin-6 and RANTES production by primary human bronchial and alveolar epithelial cells infected with A/Hong Kong/483/97 (H5N1/97), A/Vietnam/1194/04, A/Vietnam/3046/04 (both H5N1/04) and A/Hong Kong 54/98 (H1N1) influenza viruses or in mock infected cells. Culture supernatants from influenza virus-infected human respiratory epithelial cells collected at 3 h, 6 h and 24 h after infection with H5N1 and H1N1 viruses were tested by ELISA for IP-10 (Figure 8), Interleukin-6 (Figure 9) and RANTES (Figure 10). The IP-10, Interleukin-6 and RANTES mRNA levels were assayed at 3 h and 6 h post infection (data not shown) with results comparable with that shown in figure 5 and 7. The results from bronchial epithelial cells represent the means and standard deviations of three independent experiments (from the same donor). The means and standard deviations of the results from alveolar epithelial cells are based on experiments from six separate donors. * indicates p < 0.01 compared with mock and ** indicates p < 0.05 compared with H5N1/97 and H1N1 infected cells using the Bonferroni multiple comparison test.
Discussion
We found that the replication efficiency of the H5N1 and H1N1 viruses was similar in both primary human alveolar (Figure 4) and bronchial epithelial cells (Figure 6). Both influenza virus subtypes induced an IP-10, interferon beta, RANTES, and IL-6 responses. The cytokine induction was dependent on viral replication since UV-inactivated virus did not induce any effect. Interestingly, we found that H5N1/97 and 1194/04 (H5N1/04) viruses were more potent inducers of IP-10, interferon-beta, RANTES and IL-6 mRNA and protein than the human H1N1 virus (Figure 5, 7, 8 to 10). Thus, the observed differences of mRNA are reflected in levels of cytokine and chemokine proteins secreted (Figure 8 to 10). The results with 3046/04 (H5N1/04) were generally similar to 1194/04 (H5N1/04) with the exception that the levels of RANTES protein in type II pneumocytes was not significantly elevated when compared with H1N1 virus infected cells (Figure 10) although the mRNA levels were (Figure 5). Our inability to detect any interferon-beta proteins in our experiments in spite of marked induction of mRNA is probably related to the limited sensitivity of the interferon beta ELISA. A more sensitive bioassay for interferon-beta may be required for this purpose. The type II pneumocytes used in these experiments were derived from a total of 13 donors and each set of experimental data is based on the results of at least three separate experiments from three donors therefore excluding a donor specific artifact. The bronchial epithelial cells were purchased from a commercial source and comes from one donor. However, since the results from these cells are broadly in line with those from the type II pneumocytes, again, we think that donor specific artifacts are unlikely to explain the results we have obtained. Finally, these results are also comparable to our previous observations from primary human monocyte derived macrophages [12] with the exception that in contrast to macrophages, no TNF alpha and IL-1 beta was induced in respiratory epithelial cells by any of the viruses tested.
This differential hyper-induction of cytokines was not explained by differences in the replication kinetics between the two virus subtypes. H5N1 viruses isolated from patients with H5N1 disease in Hong Kong in 1997, Vietnam in 2004 and human influenza viruses of the H1N1 subtype all replicate with similar efficiency. Increase in the MOI of the H1N1 virus did not result in an increase of cytokine responses to levels comparable to that of the H5N1 viruses. The cellular mechanisms underlying this differential cytokine hyper-induction by H5N1 viruses are presently poorly understood. Studies on the transformed bronchial epithelial cell line A549 previously demonstrated that toll-like receptor 3 (TLR-3) is involved in the influenza virus A initiated cytokine responses [27]. It remains to be determined whether H5N1 viruses also act via TLR-3 signaling in primary human epithelial cells.
Cytokine and chemokine responses in vivo result from autocrine and paracrine interactions involving many cell types. Chemokines such as IP-10 and MCP-1 are macrophage chemo-attractants and mediate the inflammatory response by further recruitment of circulating leukocytes into the inflamed tissue. We have previously demonstrated that IP-10 and MCP-1 are up-regulated in primary human macrophage by SARS-CoV [28]. The strong induction of chemokines in the lung micro-environment might explain the prominent macrophage infiltrate observed in the lungs of patients with fatal H5N1 [4] as well as SARS [29].
RANTES attracts monocytes, eosinophils, basophils and T cells, and selectively CD4+ T cells. Its production from the bronchial epithelial cells contributes to the infiltration of the inflammatory cells in airway viral infection [18]. IL-6 is a multifunctional cytokine that can regulate immune and inflammatory responses involved in the activation, growth and differentiation of T-cells [30] and can contribute to T cell mediated inflammatory reactions. In fact, autopsy examination showed an increased CD3+ T cells in the interstitium of the lung from patients with H5N1 diseases [4]. In addition, IL-6 has been shown to be released by macrophages and epithelial cells during lung injury [31] and the effects of IL-6 are synergistic with those of IL-1 and TNF-alpha [32]. We have previously demonstrated that other proinflammatory cytokines such as IL-1, TNF-alpha and IL-6 are hyper-induced in H5N1 infected macrophages [12]. Therefore, the differential up-regulation of IL-6 expression in human respiratory epithelial cells and the cytokines induced in macrophages by H5N1 viruses may contribute synergistically to the pathogenesis of human H5N1 disease.
The H5N1 viruses have continued to reassort, acquiring different internal genes from other influenza viruses of avian origin [33,34]. The H5N1/04 viruses, A/Vietnam/1194/04 and A/Vietnam/3046/04 represent the Z genotype viruses that emerged as the dominant virus genotype affecting poultry in south-east Asia [27,35]. Thus there appears to be an association between the property of hyper-inducing cytokines and high virulence. Additionally, in pig epithelial cells, H5N1/97 viruses were found to resist the antiviral effects of interferon [36] and this may also be relevant in pathogenesis. It is notable that patients with avian influenza (H5N1) disease appeared to have higher levels of IP-10 in their sera than those with infections with the human influenza viruses [4] providing in vivo data that parallels our present findings in vitro. Studies on recombinant viruses bearing the HA and NA of the 1918 "Spanish flu" pandemic virus showed that these viruses have enhanced virulence for mice and induce higher levels of macrophage-derived chemokines in vivo in mice [37]. However, such observations of hyper-induction of cytokines in vivo may simply reflect more extensive replication of the respective virus. The studies in vitro with H5N1 viruses exclude such potential confounding factors and it would be relevant to study the cytokine profiles of the 1918 recombinant viruses in in vitro models similar to those described here.
Conclusion
H5N1 subtype influenza A viruses associated with human disease are more potent than human H1N1 virus at inducing proinflammatory cytokines and chemokines, including IP-10, interferon beta, IL-6 and RANTES, from human primary alveolar and bronchial epithelial cells infected in vitro. Previous findings showed that H5N1/97 viruses also hyper-induce cytokines from macrophages and that patients with H5N1 disease have high levels of IP-10 and other chemokines in the serum. These findings may be relevant to the pathogenesis of H5N1 disease. The recent re-emergence of H5N1 disease in humans is a cause for renewed pandemic concern and highlights the need for a better understanding of the pathogenesis of human H5N1 disease. Such understanding will lead to new strategies for managing human H5N1 disease and enhance our preparedness to confront pandemic influenza, whether from H5N1 or other influenza A subtypes.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JSM Peiris, MCW Chan and CY Cheung conceived the study, planned the overall experimental design and wrote the manuscript. MCW Chan carried out the experiments; MCW Chan, CY Cheung and YO Chan carried out experiments in the BL3 laboratory and RWY Chan assisted in experiments in the BL2 laboratory. WH Chui provided the lung biopsy specimens, SW Tsao helped to develop the methods for primary culture of the human alveolar epithelial cells, JM Nicholls advised on morphogical studies, and LLM Poon and Y Guan advised in experimental design. All authors critically reviewed the manuscript.
Acknowledgements
This research was supported by a research grants to MCW Chan from the Research Fund for the Control of Infectious Diseases (RFCID 03040712) and the Small Project Funding, CRGC, The University of Hong Kong and research grants to JSM Peiris from the Research Fund for the Control of Infectious Diseases (RFCID 01030172), the Research Grants Councils of Hong Kong (HKU 7459/03M) and The University of Hong Kong Research Achievement Award, 2005.
==== Refs
Webster RG Bean WJ Gorman OT Chambers TM Kawaoka Y Evolution and ecology of influenza A viruses Microbiol Rev 1992 56 1 152 79 1579108
Claas EC Osterhaus AD van Beek R De Jong JC Rimmelzwaan GF Senne DA Krauss S Shortridge KF Webster RG Human influenza A H5N1 virus related to a highly pathogenic avian influenza virus Lancet 1998 351 472 77 9482438 10.1016/S0140-6736(97)11212-0
Subbarao K Klimov A Katz J Regenery H Lim W Hall H Perdue M Swayne D Bender C Huang J Hemphill M Rowe T Shaw M Xu X Fukuda K Cox N Characterization of an avian influenza A (H5N1) virus isolated from a child with a fatal respiratory illness Science 1998 279 393 396 9430591 10.1126/science.279.5349.393
Peiris JSM Yu WC Leung CW Cheung CY NG WF Nicholls JM NG TK Chan KH Lai ST Lim WL Yuen KY Guan Y Re-emergence of fatal human influenza A subtype H5N1 disease Lancet 2004 363 617 619 14987888 10.1016/S0140-6736(04)15595-5
Tran TH Nguyen TL Nguyen TD Luong TS Pham PM Nguyen VC Pham TS Vo CD Le TQ Ngo TT Dao BK Le PP Nguyen TT Hoang TL Cao VT Le TG Nguyen DT Le HN Nguyen KT Le HS Le VT Christiane D Tran TT Menno de J Schultsz C Cheng P Lim W Horby P Farrar J World Health Organization International Avian Influenza Investigative Team Avian influenza A (H5N1) in 10 patients in Vietnam N Eng J Med 2004 350 1179 88 10.1056/NEJMoa040419
World Health Organization Avian influenza – cumulative number of cases – update 18
To KF Chan PKS Chan KF Lee WK Lam WY Wong KF Tang NLS Tsang DNC Sung RYT Buckley TA Tam JS Cheng AF Pathology of fatal human infection associated with avian influenza A H5N1 Virus Journal of Medical Virology 2001 63 242 246 11170064 10.1002/1096-9071(200103)63:3<242::AID-JMV1007>3.0.CO;2-N
De Jong Menno D Bach Van Cam Phan Tu Qui vo Minh Hien Tran Tan Thanh Nguyen Bach Hue Marcel Beld Le Thi Phuong Truong Huu Khanh Nguyen Van Vinh chau Tran Tinh Hien Do Quang Ha Jeremy Farrar Fatal avian influenza A (H5N1) in a child presenting with diarrhea followed by coma New England Journal of Medicine 2005 352 686 691 15716562 10.1056/NEJMoa044307
Yuen KY Chan PK Peiris M Tsang DN Que TL Shortridge KF Cheung PT To WK Ho ET Sung R Cheng AF Clinical features and rapid viral diagnosis of human disease associated with avian influenza A H5N1 virus Lancet 1998 351 467 71 9482437 10.1016/S0140-6736(98)01182-9
Fisman DN Hemophagocytic syndrome and infection Emerg Infect Dis 2000 6 601 08 11076718
Headley AS Tolley E Meduri GU Infections and the inflammatory response in acute respiratory distress syndrome Chest 1997 111 1306 21 9149588
Cheung CY Poon LLM Lu AS Luk W Lau YL Shortridge KF Gordon S Guan Y Peiris JSM Induction of proinflammatory cytokines in human macrophages by influenza A (H5N1) viruses: a mechanism for the unusual severity of human disease? Lancet 2002 360 1831 1837 12480361 10.1016/S0140-6736(02)11772-7
Fesq H Bacher M Nain M Gemsa D Programmed cell death (apoptosis) in human monocytes infected by influenza A virus Immunobiology 1994 190 175 182 8082884
Seo SH Webby R Webster RG No apoptotic deaths and different levels of inductions of inflammatory cytokines in alveolar macrophages infected with influenza viruses Virology 2004 329 270 279 15518807
Ebisawa IT Kitamoto O Takeuchi Y Makino M Immunocytologic study of nasal epithelial cells in influenza Am Rev Respir Dis 1969 99 507 15 4889594
Kuiken T Rimmelzwaan GF Van Amerongen G Osterhaus AD Pathology of human influenza A (H5N1) virus infection in cynomolgus macaques (Macaca fascicularis) Vet Pathol 2003 40 304 10 12724572 10.1354/vp.40-3-304
Uiprasertkul M Puthavathana P Sangsiriwut K Pooruk P Srisook K Peiris M Nicholls JM Chokephaibulkit K Vanprapar N Auewarakul P Influenza A H5N1 replication sites in humans Emerg Infect Dis 2005 11 1036 41 16022777
Adachi M Matsukura S Tokunaga H Kokubu F Expression of cytokines on human bronchial epithelial cells induced by influenza virus A Int Arach Allergy Immunol 1997 113 307 311
Cunningham AC Milne DS Wilkes J Dark JH Tetley TD Kirby JA Constitutive expression of MHC and adhesion molecules by alveolar epithelial cells (type II pneumocytes) isolated from human lung and comparison with immunocytochemical findings Journal of Cell Sciences 1994 107 443 449
Glanville AR Tazelaar HD Theodore J Imoto E Rouse RV Baldwin JC Robin ED The distribution of MHC class I and II antigens on bronchial epithelium Amer Rev Respir Dis 1989 139 330 334 2464294
Peters U Papadopoulos T Muller-Hermelink HK MHC class II antigens on lung epithelial of human fetuses and neonates. Ontogeny and expression in lungs with histologic evidence of infection Lab Invest 1990 63 38 43 1695696
Cheek JM Kim KJ Crandall ED Tight monolayers of rat alveolar epithelial cells: bioelectric properties and active sodium transport Am J Physiol 1989 256 C688 C693 2923201
Elbert KJ Schafer UF Schafers HJ Kim KJ Lee VHL Lehr CM Monolayers of human alveolar epithelial cells in primary culture for pulmonary absorption and transport studies Pharmaceutical Research 1999 16 5 601 08 10349999 10.1023/A:1018887501927
Blaschke V Reich K Blaschke S Neumann C Rapid quantitation of proinflammatory and chemoattractant cytokine expression in small tissue samples and monocyte-derived dendrite cells: validation of a new real-time RT-PCR technology J Immunol Methods 2000 246 79 90 11121549 10.1016/S0022-1759(00)00304-5
Matrosovich MN Matrosovich TY Gray T Roberts NA Klent HD Human and avian influenza viruses target different cell types in culture of human airway epithelium Proc Natl Acad Sci USA 2004 101 13 4620 4624 15070767 10.1073/pnas.0308001101
Matrosovich MN Zhou N Kawaoka Y Webster R The surface glycoproteins of H5 influenza viruses isolated from humans, chickens, and wild aquatic birds have distinguishable properties J Virol 1999 73 1146 1155 9882316
Guillot L Le Goffie R Bloch S Escriou N Akira S Chignard M Si-Tahar M Involvement of Toll-like receptor 3 in the immune response of lung epithelial cells to double stranded RNA and influenza A virus J Biol Chem 2005 280 5571 5580 15579900 10.1074/jbc.M410592200
Cheung CY Poon LL Ng IH Luk W Sia SF Wu MH Chan KH Yuen KY Gordon S Guan Y Peiris JS Cytokine responses in severe acute respiratory syndrome coronavirus-infected macrophages in vitro: possible relevance to pathogenesis J Virol 2005 79 7819 26 15919935 10.1128/JVI.79.12.7819-7826.2005
Nicholls JM Poon LLM Lee KC Ng WF Lai ST Leung CY Chu CM Hui PK Mak KL Lim WL Yan KW Chan KH Tsang NC Guan Y Yuen KY Peiris JSM Lung pathology of fatal severe acute respiratory syndrome Lancet 2003 361 1173 78
Van Snick J Interleukin-6: an overview Annu Rev Immunol 1990 8 253 78 2188664 10.1146/annurev.immunol.8.1.253
Hierholzer C Kalff JC Omert L Tsukada K Loeffert JE Watkins SC Billiar TR Tweardy DJ Interleukin-6 production in hemorrhagic shock is accompanied by neutrophil recruitment and lung injury Am J Physiol 1998 275 611 21
Le JM Fredrickson G Reis LF Diamantstein T Hirano T Kishimoto T Vilcek J Interleukin 2-dependent and interleukin 2-independent pathways of regulation of thymocyte function by interleukin 6 Proc Natl Acad Sci USA 1988 85 22 8643 7 3263651
Guan Y Peiris JS Lipatov AS Ellis TM Dyrting KC Krauss S Zhang LJ Webster RG Shortridge KF Emergence of multiple genotypes of H5N1 avian influenza viruses in Hong Kong SAR Proc Natl Acad Sci USA 2002 99 13 8950 5 12077307 10.1073/pnas.132268999
Guan Y Poon LL Cheung CY Ellis TM Lim W Lipatov AS Chan KH Sturm-Ramirez KM Cheung CL Leung YH Yuen KY Webster RG Peiris JS H5N1 influenza: a protean pandemic threat Proc Natl Acad Sci USA 2004 101 21 8156 61 15148370 10.1073/pnas.0402443101
Li KS Guan Y Wang J Smith GJ Xu KM Duan L Rahardjo AP Puthavathana P Buranathai C Nguyen TD Estoepangestie AT Chaisingh A Auewarakul P Long HT Hanh NT Webby RJ Poon LL Chen H Shortridge KF Yuen KY Webster RG Peiris JS Genesis of a highly pathogenic and potentially pandemic H5N1 influenza virus in eastern Asia Nature 2004 430 209 13 15241415 10.1038/nature02746
Seo SH Hoffmann E Webster RG Lethal H5N1 influenza viruses escape host antiviral-cytokine responses Nat Med 2002 8 950 54 12195436 10.1038/nm757
Kobasa D Takada A Shinya K Hatta M Halfmann P Theriault S Suzuki H Nishimura H Mitamura K Sugaya N Usui T Murata T Maeda Y Watanabe S Suresh M Suzuki T Suzuki Y Feldmann H Kawaoka Y Enhanced virulence of influenza A viruses with the haemagglutinin of the 1918 pandemic virus Nature 2004 431 703 7 15470432 10.1038/nature02951
|
16283933
|
PMC1318487
|
CC BY
|
2021-01-04 16:36:26
|
no
|
Respir Res. 2005 Nov 11; 6(1):135
|
utf-8
|
Respir Res
| 2,005 |
10.1186/1465-9921-6-135
|
oa_comm
|
==== Front
BMC Med Inform Decis MakBMC Medical Informatics and Decision Making1472-6947BioMed Central London 1472-6947-5-361631367610.1186/1472-6947-5-36Research ArticleDevelopment and initial testing of a computer-based patient decision aid to promote colorectal cancer screening for primary care practice Kim Jane [email protected] Annie [email protected] Sarah [email protected] Carmen [email protected] Marci [email protected] Lisa [email protected] Beth [email protected] Sue [email protected] Regina [email protected] Michael [email protected] Preventive Medicine Residency Program, University of North Carolina, Chapel Hill, NC, USA2 Division of General Internal Medicine, Department of Medicine, University of North Carolina, Chapel Hill, NC, USA3 Department of Nutrition, University of North Carolina School of Public Health, Chapel Hill, NC, USA4 Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC, USA2005 28 11 2005 5 36 36 20 5 2005 28 11 2005 Copyright © 2005 Kim et al; licensee BioMed Central Ltd.2005Kim et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Although colorectal cancer screening is recommended by major policy-making organizations, rates of screening remain low. Our aim was to develop a patient-directed, computer-based decision aid about colorectal cancer screening and investigate whether it could increase patient interest in screening.
Methods
We used content from evidence-based literature reviews and our previous decision aid research to develop a prototype. We performed two rounds of usability testing with representative patients to revise the content and format. The final decision aid consisted of an introductory segment, four test-specific segments, and information to allow comparison of the tests across several key parameters. We then conducted a before-after uncontrolled trial of 80 patients 50–75 years old recruited from an academic internal medicine practice.
Results
Mean viewing time was 19 minutes. The decision aid improved patients' intent to ask providers for screening from a mean score of 2.8 (1 = not at all likely to ask, 4 = very likely to ask) before viewing the decision aid to 3.2 afterwards (difference, 0.4; p < 0.0001, paired t-test). Most found the aid useful and reported that it improved their knowledge about screening. Sixty percent said they were ready to be tested, 18% needed more information, and 22% were not ready to be screened. Within 6 months of viewing, 43% of patients had completed screening tests.
Conclusion
We conclude that a computer-based decision aid can increase patient intent to be screened and increase interest in screening. Practice Implications: This decision aid can be viewed by patients prior to provider appointments to increase motivation to be screened and to help them decide about which modality to use for screening. Further work is required to integrate the decision aid with other practice change strategies to raise screening rates to target levels.
==== Body
Background
Colorectal cancer (CRC) is the second leading cause of cancer death and third most diagnosed cancer in the United States [1,2]. Multiple policy-making organizations have published evidence-based CRC screening guidelines that recommend annual fecal occult blood testing (FOBT), sigmoidoscopy or barium enema every 5 years, or colonoscopy every 10 years for adults 50 and older [3-5]. Despite these guidelines, rates of colorectal cancer screening in the U.S. are low with only approximately 50% of adults 50 and older reporting a CRC screening test within recommended time intervals [6,7].
Given the many testing options, colorectal cancer screening can be a complex issue, and time limitations may prevent providers from adequately discussing all options with patients. Patient decision aids offer a means of circumventing these limitations. Decision aids are educational materials that help individuals understand their choices for screening or treatment [8,9]. They have been shown to improve patients' knowledge, reduce decisional conflict, and increase active patient participation in medical decision making [8].
We previously conducted a randomized controlled trial of an 11-minute videotape-based decision aid on CRC screening paired with a color-coded chart marker and brochure [10]. The decision aid was based on the transtheoretical stages of change model [11]. Patients indicated their stage of readiness to be screened by selecting a color-coded brochure, and a chart marker of the same color was attached to their charts. The decision aid increased patients' intent to ask their providers for screening, and 47% of intervention group participants had CRC screening tests ordered compared to 26% of controls. A chart review 3 months afterwards found increased screening test completion in the intervention group compared to controls (37% vs. 23%, respectively, p < 0.05).
Despite the videotape-based decision aid's success in raising screening rates, there were some limitations in its format and content. First, it only offered FOBT and sigmoidoscopy; colonoscopy was not an available option at the time of the original study. Because it was produced in videotape form, the decision aid could not be easily updated to include newly endorsed screening modalities or other new information. In addition, the videotape could not be tailored to meet different levels of patient interest in CRC screening. Recent advances in web-based technology made it possible to produce a computer-based decision aid that can be updated and customized to meet patients' individual knowledge needs.
We sought to develop, refine, and pilot test whether a computer-based, patient-directed decision aid can increase patient interest in screening, meet informational needs, and lead to the ordering and completion of colorectal cancer screening tests.
Methods
Computer-based decision aid development
We based the content of the computer-based decision aid, called CHOICE™ version 1.0, on systematic reviews of the literature and our videotape decision aid [10]. New segments on colonoscopy, barium enema, and comparative information about the tests were developed. We also developed patient vignettes by filming interviews with patients who had undergone screening and agreed to discuss their experiences. The decision aid, two self-administered questionnaires, and a data collection mechanism using Microsoft Access as a back-end database were programmed into a web-based format using Active Server Pages. The decision aid was developed for use on a local computer version, although an Internet-based version has subsequently been developed.
Decision aid format and content
The final version of the decision aid consists of a modular design with a 5-minute introduction and five additional 3–5 minute segments that describe individual screening tests or comparative information about the tests. An audio track accompanies the entire decision aid and explains all figures that are presented, making the content accessible to users with varying levels of literacy. Readability was not formally assessed given that the decision aid was not based on prose text. However, we designed the interface, including the audio track and figures, with the intent to make the information easy to understand. The patient uses a mouse to activate an introductory segment that explains what colorectal cancer is, describes its prevalence and lifetime risk, outlines the importance of colorectal cancer screening and the benefits of early detection, and gives a brief overview of each recommended screening modality. A physician provided narration and multiple interviews were included of patients describing their experiences and views of colorectal cancer screening. This portion of the decision aid was to be viewed by all users.
After the introduction, the decision aid directs patients to a menu of choices that allows them to choose one or more additional test-specific video segments by clicking on the name of the test with the mouse (Figure 1). Each segment contains footage that utilizes physician and patient narration and describes how each test is performed, the preparation required, and common patient concerns about the procedure. While the physician narrates, animated cartoons (Figure 2) and text visually depict the anatomic area examined by the test and the way the screening test is performed. Our aim was to present a balanced overview of the risks and benefits of each test, including preparation required, discomfort anticipated during the test, the risk of adverse effects, as well as the ability of each test to detect colorectal cancer or precancerous polyps. After the overview about the test, the decision aid presents multiple 5–15 second patient vignettes. These vignettes consist of interviews with white and African American men and women describing their reasons for choosing a particular screening modality, their experiences with the test, and the benefits and downsides of the test. Additional video footage shows patients preparing for the tests. Follow-up of abnormal results and comparisons of different screening modalities to one another are explained via text, narration, and graphs (Figure 3). At the end of each segment, patients are directed to the navigation screen that allows them to choose another test information segment, examine comparative information, or to go to the post-intervention questionnaire. After initial testing, the decision aid was modified to require that patients watch the introduction and at least one other video segment.
Figure 1 Menu of choices.
Figure 2 Illustration of endoscopy technique.
Figure 3 Comparison of screening tests.
Usability testing
We recruited twelve patients 50–75 years old to participate in each round of usability testing. These patients were a convenience sample recruited from the University of North Carolina-Chapel Hill (UNC) general internal medicine practice. We attempted to identify users with a wide range of previous computer experience and education levels. The computer used for the usability testing and the pilot study was an IBM PC with a Pentium processor and Windows operating system. Research team members experienced in usability testing observed and videotaped patients as they worked with the decision aid using the Think Aloud technique [12]. The study investigators modified the decision aid via an iterative process based on feedback from the usability testing that identified users' difficulties with the decision aid. In the first session, users, particularly those less familiar with computers, had several problems with navigation. These problems included confusion about how to use the mouse, difficulty in reading the type font, and inability to move between screens. Based on these results, we increased font sizes, changed the placement and color contrast of navigation elements and content, and simplified navigation options and operation of the video segments. After these changes, patients in the second round of testing with varying levels of computer experience were able to navigate through the decision aid with greater ease and less confusion than first-round users.
Study design
To pilot test the efficacy of the decision aid, we then conducted a before-after uncontrolled trial at the UNC general internal medicine practice. The practice consists of 75 medical residents, 20 attending physicians, and approximately 12,000 patients, of whom approximately 5,000 are between 50 and 75 years of age.
Population
We enrolled a convenience sample of patients from June 2003-April 2004. The participants were adults 50 to 75 years old who were cared for in the general internal medicine practice and presented to their provider for a scheduled appointment. Eligibility criteria were: 1) the absence of a personal or family history of colon cancer in a first degree relative; 2) sufficient general health to undergo screening as determined by the research assistant (RA) or primary care provider; and 3) the ability to communicate in English. The RA sought permission from providers and then approached eligible patients and asked them to enroll. Clinic providers also referred patients for participation. If the patient agreed to participate, the RA obtained written informed consent. Both patients who were up-to-date with screening and those who were due for screening were allowed to enroll with the rationale that individuals in both groups could learn more about screening options for their next decision about screening. We defined up-to-date status as having an FOBT in the past year, sigmoidoscopy or barium enema in the past 5 years, or colonoscopy in the past 10 years. If we found in patients' medical records or baseline surveys that they were up-to-date with screening, they were asked to respond as if they were deciding about their next screening opportunity.
Intervention
Participants viewed the decision aid on a computer in a private area in the clinic either before or after their scheduled appointment. The RA was present during their viewing session. The patients were encouraged to navigate independently through the decision aid and questionnaires, but were offered assistance from the RA if needed.
Data collection
Patients completed web-based, self-administered questionnaires before and after viewing the decision aid (see Additional file 1). We based the questionnaires' content and format on the questions used in our videotape decision aid trial. After viewing the decision aid, participants indicated their stage of readiness to be screened by choosing one of three color-coded stages: green indicated that they were ready to be screened, yellow that they needed more information, and red that they did not want screening. They completed an additional paper-based questionnaire based on their stage of readiness (Appendix A). We asked patients who were ready to be screened what criteria were most important in deciding on screening and which test they would prefer to have. The RA conducted an electronic chart review 6 months afterwards to determine if CRC screening tests were ordered and completed.
Outcome measures
The primary outcome measures were: 1) intent to ask providers about screening; and 2) interest in CRC screening. Intent to ask providers for screening was measured on a 4-point Likert scale (1 = not at all likely to ask, 4 = very likely to ask). Patients' interest in being screened for CRC in the next 6 months was also measured on a 4-point Likert scale (1 = not at all interested, 4 = definitely interested).
Secondary outcome measures included: subjective change in knowledge about screening, helpfulness of the information in making a decision about screening, and preferences for shared decision making with their provider.
Other outcome measures were the proportions of CRC screening tests ordered and completed after 6 months. Test ordering was defined as an FOBT order recorded in the clinic's database or a colonoscopy, sigmoidoscopy, or barium enema order in the electronic medical record. Test completion was defined as a completed FOBT recorded in the clinic database or a completed colonoscopy, sigmoidoscopy, or barium enema report in the electronic medical record. Patients choosing a test were then asked to give their main reason for their particular choice.
Statistical analysis
We examined the characteristics of the sample by using univariate analyses to determine the distribution of each variable. The mean, range, and standard deviation were calculated for continuous variables and frequencies and percentages were tabulated for categorical variables. To assess change in interest and intent to be screened, we used paired t-tests to compare the difference in Likert scores before and after viewing the decision aid. Frequencies and percentages were tabulated for categorical variables in the questionnaires and for test ordering and completion.
Pearson's chi-square test was used to compare the percentage ordering and completing tests by stage of readiness to be screened. We conducted another analysis after excluding patients who were already up-to-date with screening because these patients might be less likely to have tests ordered or completed after viewing the decision aid. We used Pearson's chi-square test to compare the proportion completing tests among patients who had previously been screened, were up-to-date with screening, or had previous conversations about screening with their provider vs. those who had not.
Two-sided p-values < 0.05 were considered statistically significant. Stata version 8.2 (College Station, TX) was used for all analyses. Prior approval for the study was obtained from the University of North Carolina-Chapel Hill Institutional Review Board, and the research was carried out in compliance with the Helsinki Declaration [13].
Results
Patient characteristics
Approximately 260 patients were approached and 80 agreed to participate (31% response rate). The main reason for non-participation was lack of time to complete the study during the index visit. The demographic characteristics of participants are shown in Table 1; demographic characteristics were similar for those who were up-to-date with screening and those who were not. The mean age was 60 years. Fifty-nine percent were male and 69% were white. Of the 81% with health insurance, 18% had Medicare and 46% had private insurance. Approximately half of the participants had ever been screened for colon cancer and 18% were up-to-date with screening. Almost two-thirds of patients said that a provider had discussed CRC screening with them in the past.
Table 1 Characteristics of the sample (n = 80).
Characteristic Mean (range) or percent
Mean age 60 (49–75)
% Male 59
% White 69
% African American 29
% Insured 81
% More than high school education 65
% Self-rated excellent-good health 67
% Screened for colorectal cancer in the past 48
% Up to date with screening 18
Change in interest in screening and intent to be screened
Interest in screening and intent to ask providers about screening rose significantly after viewing the decision aid. Patients' interest in being screened in the 6 months after viewing the decision aid increased from a mean score of 3.2 before viewing the decision aid to 3.5 afterwards (difference, 0.3; p = 0.01, paired t-test, Figure 4). Patients' intent to ask for screening increased from 2.8 to 3.2 (difference, 0.4; p < 0.0001, paired t-test).
Figure 4 Change in intent to be screened after viewing the decision aid. * p = 0.01, paired t-test. Based on 4-point Likert scale, 1 = not at all interested, 4 = very interested ** p < 0.0001, paired t-test. Based on 4-point Likert scale, 1 = not at all likely to ask, 4 = very likely to ask
Process outcomes
Eighty-nine percent said that the information increased their knowledge about colon cancer, 78% said that the information helped them decide whether to be screened, and 90% felt that the amount of information presented was just right. Even among those who were not ready to be screened, most found the aid helpful and reported that it increased their knowledge about colorectal cancer. Ninety percent preferred to make decisions about their health together with their physician. The mean amount of time spent viewing the decision aid was 19 minutes.
Test ordering and completion
Forty-eight percent had either an FOBT or endoscopy (colonoscopy or sigmoidoscopy) ordered within 6 months of viewing the decision aid (Table 2). Thirty-three percent had an FOBT ordered and 26% had an order for endoscopy. In terms of test completion, 43% of participants completed either FOBT or endoscopy. Twenty-three percent completed FOBT and 26% completed endoscopies, and 2.5% completed both tests. No barium enemas were ordered or performed.
Table 2 Six-month follow-up: screening test ordering and completion
Total FOBT Colonoscopy/sigmoidoscopy
% Ordered 48 33 26
% Completed 43 23 26
All percentages based on n = 80
Self-reported stage of change and relationship with test completion
When asked about stage of readiness to be screened, 60% chose green (ready to be tested), 18% chose yellow (needed more information) and 22% chose red (not ready to be screened). There was a greater percentage of test ordering and completion among patients choosing green (51% and 43%, respectively) compared to those choosing yellow or red, but these differences did not reach statistical significance (Table 3).
Table 3 Test ordering and completion by stage of readiness to be screened
Stage of Readiness n % with test ordered* % completing test**
Green: ready to be tested 47 51 43
Yellow: need more Information 18 47 33
Red: not ready for screening 15 39 33
*p = 0.59, Pearson's chi-square test
**p = 0.71, Pearson's chi-square test
After excluding patients who were up-to-date with screening, there was no significant change in the percentage of patients having tests ordered and completed by stage of readiness to be screened. Screening test completion rates did not differ significantly between those who were already up-to-date with screening and those who were not in compliance. Individuals who were up-to-date had lower baseline intent and interest scores and fewer were ready to be screened compared to those who were not up-to-date (Table 4). More FOBTs were ordered in those who were up-to-date, but the percentage that had a lower endoscopy was similar to the percentage that had endoscopies among those due for screening. Having a previous history of screening or having previous conversations about screening with a provider did not result in higher test completion rates.
Table 4 Intent* and interest** in screening, test ordering and completion, and readiness to be screened by up-to-date§ status
Up-to-date (n = 14) Not up-to-date (n = 66)
Intent to ask for screening
before viewing aid 2.4 2.9
after viewing aid 2.6 3.3
Interest in screening
before viewing aid 2.6 3.3
after viewing aid 3.0 3.6
Readiness to be screened (%)
Ready 29 66
Not ready 43 18
Need more information 29 15
% with lower endoscopy ordered 21 28
% with lower endoscopy completed 21 27
% with FOBT ordered 57 21
% with FOBT completed 29 17
*Intent to ask for screening was based on the question: "How likely are you, at this visit, to ask your doctor about being tested for colon cancer?" and used a 4-point Likert scale, 1 = not at all likely to ask, 4 = very likely to ask
**Interest in screening was based on the question:"How interested are you in having a test for colon cancer in the next 6 months?" and used a 4-point Likert scale, 1 = not at all interested, 4 = definitely interested
§up-to-date: FOBT in the past year, sigmoidoscopy or barium enema in the past 5 years, or colonoscopy in the past 10 years
Test preferences
We asked patients who were ready to be screened which screening modality they preferred to have. Forty-two percent chose colonoscopy, 20% chose FOBT alone, and 18% chose FOBT in combination with sigmoidoscopy. However, only 28% actually had the test they preferred ordered by their provider. We also asked these patients to choose the most important factor in deciding on a screening test (Appendix A). More than half (54.5%) said that the ability of tests to find cancers or polyps, or test accuracy, was the most important criteria in selecting a screening method.
Discussion
We developed a computer-based colon cancer screening decision aid and found that it could increase patient interest in screening and intent to be screened. Most participants were ready to be screened after viewing the decision aid, 48% had tests ordered and 43% completed screening tests. These results are similar in magnitude to those from our videotape decision aid trial in which patients' intent to ask providers for screening increased significantly after viewing the aid and 37% completed tests.
In our current study, the computer-based decision aid subjectively improved patients' knowledge about screening and was useful to most in making decisions about screening. Other studies have found that similar tools increased patients' level of knowledge about screening, but effects on screening rates have varied. Zapka et al. conducted a randomized controlled trial of a video on sigmoidoscopy that was mailed to patients in advance of a scheduled visit [14]. A decision aid developed by Meade et al. improved patient knowledge about CRC screening as determined by a change in score from expert-validated pre- and post-tests [15]. Dolan et al. found that patients subjectively reported improved knowledge about CRC screening after using a decision aid [16]. In these studies, there was no difference in screening test ordering and completion between those who viewed the decision aid compared to those who did not [16]. In contrast to these studies that looked at the effect of a decision aid alone, our previous study using a combined intervention of a videotape decision aid and chart marker was able to increase screening test ordering and completion compared to controls [10].
Our computer-based aid differs from other decision aids for CRC screening in that patients were able to interact with the aid via its modular format and choose to view information based on their knowledge needs. Patients in previous trials of decision aids on CRC screening all received the same educational content regardless of their knowledge about screening [10,14-16]. Our computer-based aid was not truly tailored in that the decision aid was not customized to fit individual patients' characteristics. Each patient, however, was able to select the amount and content of information they received. In this way, the information on CRC screening may have achieved greater relevance to patients.
Only 28% of those who were ready to be screened had the test they preferred ordered by their provider. There are a number of possible reasons for the lack of congruence between patient preferences and test ordering. First, some patients may have viewed the decision aid after seeing their provider and thus did not have an opportunity to discuss screening at that visit or another visit within the 6-month follow-up window. Another possibility is that providers may not have been aware of patients' preferences and had not been trained to provide stage-appropriate responses. Third, patients who were already up-to-date with screening may not have had tests ordered. Excluding patients who were up-to-date from the analysis, however, did not increase the proportion of tests ordered, so this explanation is unlikely to account for the low test ordering rates.
Given these results, a patient-oriented decision aid alone may be insufficient to ensure test ordering based on patient preferences or increase test ordering and completion to desired levels; multifaceted interventions that target a combination of providers, patients, or office systems may be more likely to increase screening rates [17]. Interventions such as physician prompts and standing orders can increase performance of preventive care, including cancer screening [18,19]. Standing orders are another potential component of a multifaceted intervention. In a standing orders protocol, a nurse initiates test ordering based on patient preferences and a practice-approved protocol. Implementing the decision aid with office-system interventions may help improve rates of screening test ordering.
The proportion of tests ordered and completed for patients who answered green was higher than for patients who answered yellow or red, but the differences were not statistically significant. The study with its small sample size may have lacked power to detect a significant difference between the groups. In addition, approximately 40% of patients choosing red had tests ordered and one-third completed screening tests. In our previous videotape decision aid study, only 7% of patients choosing red had tests ordered and 4% completed tests. The seeming disconnect between patient interest and provider ordering in the current study is concerning and may be due to poor patient-provider communication about preferences or patients who changed their mind about screening after completing the questionnaire. More research needs to be done to determine why patients who were uncertain or not ready for screening had tests ordered and completed.
Of the patients who were ready for screening, most rated the ability to find cancer, or accuracy, as the most important factor in deciding on a test. Ling et al. previously found that most patients rate accuracy as the most important feature of a CRC screening test, but that providers thought that discomfort in undergoing a test was most important to patients [20]. Providers should be aware that many individuals value the accuracy of screening methods and counsel their patients accordingly.
There are a number of limitations to this study. Foremost, it was an uncontrolled trial without a comparison group, so it is unclear whether the proportion of patients having tests ordered and completed over 6 months represents an increase compared to the usual care of patients who did not view the decision aid but were otherwise eligible for the study. Our results, however, are fairly comparable to those from our videotape decision aid randomized trial conducted in three central North Carolina private primary care practices: overall there was a net 0.6 unit increase in intent to be screened after the decision aid. Among patients viewing the videotape decision aid, 47% of individuals had screening ordered compared to 26% of controls, and 37% completed tests vs. 23% of controls. We chose to use an uncontrolled design as the first phase of testing to evaluate whether the aid could increase interest in screening and was useful to patients in choosing a screening modality. Whether patients' increased interest in screening after viewing the computer-based decision aid can lead to an improvement in screening rates cannot be determined from this pilot study; this question will be better addressed in a larger, multi-center randomized trial with screening test completion as the main outcome.
Other limitations were the use of a convenience sample and selection bias. Given the volunteer study population with some subjects referred by physicians and the low response rate, the responses of those who chose to participate may be different from those who did not participate. Our results also could have been affected by the fact that almost half of the participants were previously screened and 18% were up-to-date with screening. We performed additional analyses excluding those who were up-to-date and did not find a change in the percentage of tests ordered and completed. There were also some differences in outcomes between those who were up-to-date and those who were not. Individuals who were up-to-date had lower mean intent and interest scores at baseline and after watching the aid, and less were ready to be screened after watching the aid compared to those who were not up-to-date (Table 4). These results are based on small numbers, however, and should be interpreted with caution.
Although patients were able to choose different videos in the decision aid, we did not track which segments were viewed by patients. Tracking may have provided additional information on how individual use of the decision aid was related to change in interest, test preferences, or test completion. However, mean viewing time was 19 minutes, indirectly suggesting that patients were accessing a significant portion of the content.
This study did not objectively measure screening knowledge before and after the aid, its effect on decisional conflict, or changes in anxiety or satisfaction with decisions, other important measures of a decision aid's effectiveness [8]. In this pilot study conducted in a busy primary care practice, we chose to focus on whether the aid could increase interest in screening and was useful in deciding on a screening modality. Future studies should assess whether this decision aid can decrease decisional conflict and improve objective knowledge about screening.
Because our study was conducted at a single site, our study findings may not be generalizable to other populations. Those in our convenience sample had high levels of education, most had insurance, and many had prior experience with screening. Other patient populations, including individuals not currently receiving regular medical care, might respond differently to the decision aid.
Although Medicare, Medicaid, and most private insurers cover CRC screening [21], cost may be an important issue for patients. We did not collect information on which tests were covered by patients' insurance carriers, co-pays and deductibles, or the importance of cost in patients' decisions about screening. Evaluating the effect of different levels of co-payment on patient preferences is an important area for future research.
A final limitation is that the decision aid may be somewhat challenging for those with limited computer skills. Although we did not objectively measure how many patients needed assistance, we observed that most patients completed the aid independently and required limited, if any, computer assistance. Whether the additional benefits of the web-based format outweigh the greater requirements for computer skill requires further research. We have developed a DVD version of the decision aid that preserves the ability to self-navigate but may be easier for computer-inexperienced users or those without access to a computer.
Conclusion
This computer-based decision aid on colorectal cancer screening increased patient interest in screening and intent to be screened. Most patients could independently navigate through the menu of choices to select segments that met their knowledge needs. It is interactive and takes approximately 20 minutes of patient time.
There are many ways in which the decision aid can be incorporated into primary care practice. Because the decision aid explains the importance of CRC screening and describes each test with its risks and benefits, it can potentially save providers time in counseling patients about screening. Internet- and DVD versions of the decision aid are currently in production, and we plan to produce a Spanish language version that will allow for dissemination of the aid to Spanish-speaking patients. The decision aid could be watched by patients at home in preparation for a provider appointment, a process we are currently testing in another study. In the office, nurses can identify patients at triage who are due for screening and direct them to watch the decision aid prior to an upcoming visit or while waiting to see their provider.
In this pilot study, a computer-based, patient-directed decision aid increased patient interest in colorectal cancer screening and subjectively improved knowledge about screening options. Most patients were ready to be screened after viewing the decision aid but only half the patients who wanted to be screened had tests ordered. Future research needs to be done to determine whether implementation of the decision aid with other interventions can effectively raise screening rates in a primary care setting.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JK participated in data collection, performed the statistical analysis, and drafted the manuscript. AW and SH participated in data collection. CL helped conceive of the study, participated in the design of the study and participated in the drafting and editing of the decision aid. MC, LS, and BF participated in the drafting and editing of the decision aid. SG programmed the decision aid interface. RM participated in the design of the decision aid. MP conceived of the study, participated in the drafting and editing of the decision aid, designed and coordinated the study, and helped to 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:
Supplementary Material
Additional File 1
Decision aid questionnaires
Click here for file
Acknowledgements
Dr. Kim's fellowship support was from the Physician Training Award in Preventive Medicine #PTAPM-01-085-01 from the American Cancer Society. Support for this project was provided through a Career Development Award for Primary Care Physicians # 01-195-01 from the American Cancer Society to Dr. Pignone and from the National Institutes of Health grant # DK 56350 for the UNC Clinical Nutrition Research Center
==== Refs
Jemal A Tiwari R Murray T Ghafoor A Samuels A Ward E Feuer E Thun MJ Cancer Statistics, 2004 CA: A Cancer Journal for Clinicians 2004 41 8 29 14974761
Cancer Facts and Figures 2003 Atlanta, American Cancer Society
Smith RA Cokkinides V Eyre HJ American Cancer Society guidelines for the early detection of cancer, 2003 CA Cancer J Clin 2003 53 27 43 12568442
Screening for colorectal cancer: recommendation and rationale Ann Intern Med 2002 137 129 131 12118971
Winawer S Fletcher R Rex D Bond J Burt R Ferrucci J Ganiats T Levin T Woolf S Johnson D Kirk L Litin S Simmang C Colorectal cancer screening and surveillance: clinical guidelines and rationale-Update based on new evidence Gastroenterology 2003 124 544 560 12557158 10.1053/gast.2003.50044
Centers for Disease Control and Prevention. Behavioral Risk Factor Surveillance System Prevalence Data
Seeff LC Nadel MR Klabunde CN Thompson T Shapiro JA Vernon SW Coates RJ Patterns and predictors of colorectal cancer test use in the adult U.S. population Cancer 2004 100 2093 2103 15139050 10.1002/cncr.20276
O'Connor AM Rostom A Fiset V Tetroe J Entwistle V Llewellyn-Thomas H Holmes-Rovner M Barry M Jones J Decision aids for patients facing health treatment or screening decisions: systematic review Bmj 1999 319 731 734 10487995
Barry MJ Health decision aids to facilitate shared decision making in office practice Ann Intern Med 2002 136 127 135 11790064
Pignone M Harris R Kinsinger L Videotape-based decision aid for colon cancer screening. A randomized, controlled trial Ann Intern Med 2000 133 761 769 11085838
Prochaska JO Redding CA Evers KE Glanz K, Lewis KM and Rimer B The transtheoretical model and stages of change Health behavior and health education: Theory, research, and practice 2002 San Francisco, Jossey-Bass 99 120
Dumas J Redish J A practical guide to usability testing 1993 xiv Norwood, NJ, Ablex Publishing Corporation 412
Helsinki Declaration
Zapka JG Lemon SC Puleo E Estabrook B Luckmann R Erban S Patient education for colon cancer screening: a randomized trial of a video mailed before a physical examination Ann Intern Med 2004 141 683 692 15520425
Meade CD McKinney WP Barnas GP Educating patients with limited literacy skills: the effectiveness of printed and videotaped materials about colon cancer Am J Public Health 1994 84 119 121 8279598
Dolan JG Frisina S Randomized controlled trial of a patient decision aid for colorectal cancer screening Med Decis Making 2002 22 125 139 11958495 10.1177/02729890222063017
Hulscher ME Wensing M Grol RP van der Weijden T van Weel C Interventions to improve the delivery of preventive services in primary care Am J Public Health 1999 89 737 746 10224987
Balas EA Weingarten S Garb CT Blumenthal D Boren SA Brown GD Improving preventive care by prompting physicians Arch Intern Med 2000 160 301 308 10668831 10.1001/archinte.160.3.301
Snell JL Buck EL Increasing cancer screening: a meta-analysis Prev Med 1996 25 702 707 8936572 10.1006/pmed.1996.0109
Ling BS Moskowitz MA Wachs D Pearson B Schroy PC Attitudes toward colorectal cancer screening tests J Gen Intern Med 2001 16 822 830 11903761 10.1046/j.1525-1497.2001.10337.x
Medicare preventive services...to help keep you healthy.
|
16313676
|
PMC1318488
|
CC BY
|
2021-01-04 23:52:11
|
no
|
BMC Med Inform Decis Mak. 2005 Nov 28; 5:36
|
utf-8
|
BMC Med Inform Decis Mak
| 2,005 |
10.1186/1472-6947-5-36
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-891632421710.1186/1743-422X-2-89ResearchEffect of ethanol on innate antiviral pathways and HCV replication in human liver cells Plumlee Courtney R [email protected] Catherine A [email protected] Nelson [email protected] Stephen J [email protected] Department of Laboratory Medicine, University of Washington, Seattle, USA2 Department of Biological Sciences, Columbia University, New York, NY3 Department of Pathology, University of Washington, Seattle, USA4 Department of Pathology, University of Washington, Seattle, USA5 Departments of Laboratory Medicine, Microbiology and Pathobiology, University of Washington, Seattle, USA2005 2 12 2005 2 89 89 6 9 2005 2 12 2005 Copyright © 2005 Plumlee et al; licensee BioMed Central Ltd.2005Plumlee 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.
Alcohol abuse reduces response rates to IFN therapy in patients with chronic hepatitis C. To model the molecular mechanisms behind this phenotype, we characterized the effects of ethanol on Jak-Stat and MAPK pathways in Huh7 human hepatoma cells, in HCV replicon cell lines, and in primary human hepatocytes. High physiological concentrations of acute ethanol activated the Jak-Stat and p38 MAPK pathways and inhibited HCV replication in several independent replicon cell lines. Moreover, acute ethanol induced Stat1 serine phosphorylation, which was partially mediated by the p38 MAPK pathway. In contrast, when combined with exogenously applied IFN-α, ethanol inhibited the antiviral actions of IFN against HCV replication, involving inhibition of IFN-induced Stat1 tyrosine phosphorylation. These effects of alcohol occurred independently of i) alcohol metabolism via ADH and CYP2E1, and ii) cytotoxic or cytostatic effects of ethanol. In this model system, ethanol directly perturbs the Jak-Stat pathway, and HCV replication.
Infection with Hepatitis C virus is a significant cause of morbidity and mortality throughout the world. With a propensity to progress to chronic infection, approximately 70% of patients with chronic viremia develop histological evidence of chronic liver diseases including chronic hepatitis, cirrhosis, and hepatocellular carcinoma. The situation is even more dire for patients who abuse ethanol, where the risk of developing end stage liver disease is significantly higher as compared to HCV patients who do not drink [1,2].
Recombinant interferon alpha (IFN-α) therapy produces sustained responses (ie clearance of viremia) in 8–12% of patients with chronic hepatitis C [3]. Significant improvements in response rates can be achieved with IFN plus ribavirin combination [4-6] and pegylated IFN plus ribavirin [7,8] therapies. However, over 50% of chronically infected patients still do not clear viremia. Moreover, HCV-infected patients who abuse alcohol have extremely low response rates to IFN therapy [9], but the mechanisms involved have not been clarified.
MAPKs play essential roles in regulation of differentiation, cell growth, and responses to cytokines, chemokines and stress. The core element in MAPK signaling consists of a module of 3 kinases, named MKKK, MKK, and MAPK, which sequentially phosphorylate each other [10]. Currently, four MAPK modules have been characterized in mammalian cells: Extracellular Regulated Kinases (ERK1 and 2), Stress activated/c-Jun N terminal kinase (SAPK/JNK), p38 MAP kinases, and ERK5 [11]. Interestingly, ethanol modulates MAPKs [12]. However, information on how ethanol affects MAPKs in the context of innate antiviral pathways such as the Jak-Stat pathway in human cells is extremely limited.
When IFN-α binds its receptor, two receptor associated tyrosine kinases, Tyk2 and Jak1 become activated by phosphorylation, and phosphorylate Stat1 and Stat2 on conserved tyrosine residues [13]. Stat1 and Stat2 combine with the IRF-9 protein to form the transcription factor interferon stimulated gene factor 3 (ISGF-3), which binds to the interferon stimulated response element (ISRE), and induces transcription of IFN-α-induced genes (ISG). The ISGs mediate the antiviral effects of IFN. The transcriptional activities of Stats 1, 3, 4, 5a, and 5b are also regulated by serine phosphorylation [14]. Phosphorylation of Stat1 on a conserved serine amino acid at position 727 (S727), results in maximal transcriptional activity of the ISGF-3 transcription factor complex [15]. Although cross-talk between p38 MAPK and the Jak-Stat pathway is essential for IFN-induced ISRE transcription, p38 does not participate in IFN induction of Stat1 serine phosphorylation [14,16-19]. However, cellular stress responses induced by stimuli such as ultraviolet light do induce p38 MAPK mediated Stat1 S727 phosphorylation [18].
In the current report, we postulated that alcohol and HCV proteins modulate MAPK and Jak-Stat pathways in human liver cells. To begin to address these issues, we characterized the interaction of acute ethanol on Jak-Stat and MAPK pathways in Huh7 cells, HCV replicon cells lines, and primary human hepatocytes.
HCVIFNvirus-host interactionssignal transductionalcohol
==== Body
Materials and methods
Cells and chemicals
Human hepatoma Huh7 cells were grown in DMEM containing 10% FBS, 1× penicillin, streptomycin, fungizone, 10mM L-glutamine, and 1× non-essential amino acids (all reagents were from Invitrogen; Carlsbad, CA). BB7 cells are derived from Huh7 cells and support the replication of a subgenomic HCV replicon containing a S2204I adaptive mutation in the NS5A gene [20]. FL-Neo cells are a stable Huh7 derived cell line containing a genomic length HCV replicon with the S2204I mutation in NS5A and a P1496L mutation in NS3. BB7 and FL-Neo cells were obtained from Apath, LLC. Subgenomic replicon cell lines 9–13 and 5-15-9-2-3 (referred to as 5–15 in this paper) containing different adaptive mutations [21-23] were kindly provided by Dr. Ralf Bartenschalger. Replicon cell lines were maintained in Huh7 media containing 400 μg/ml of G418 (Calbiochem; San Diego, CA). Primary human fetal hepatocytes were isolated and grown in chemically defined serum free medium as described [24]. Primary hepatocyte cultures were analyzed within 2 days of isolation. Cells were maintained in humidified incubators at 37°C with 5% CO2. Ethanol (AAPER; Shelbyville, KY) at concentrations of 0–200 mM, was added to cells at the same time as IFN-α (Sigma, St. Louis, MO). Relative to untreated cells, ethanol did not induce any cytotoxic or growth inhibitory effects on any of the cell types at any of the doses tested (see Additional File 1). MAPK inhibitors UO126, PD98059, and SB203508, used to inhibit p42/44, MEK1, and p38 MAPK pathways, respectively, were solubilized in DMSO and obtained from Calbiochem. ADH and CYP2E1 inhibitors 4-methylpyrazole (4-MP) and diallylsulfide (DAS) [25], were obtained from Sigma and solubilized in DMSO. In all experiments, the final concentration of DMSO was below 0.2%, so as to prevent DMSO inhibition of CYP2E1 [26].
Transfection
The day prior to transfection, 2 × 105 cells were plated in 12-well tissue culture plates. Endotoxin free plasmid DNA was purified (Endofree kit, Qiagen; Valencia, CA), and was introduced into cells with lipofectamine 2000 according to manufacturer's recommendations (Invitrogen). Transfection efficiency was monitored by including 0.5 μg of plasmid pQ150 (provided by Dr. Jeffery Vieira), which expresses GFP under control of the constitutive EF-1α promoter. Prior to harvesting protein lysates, cells expressing GFP were visualized by fluorescence microscopy and the transfection efficiency calculated by determining the percentage of green cells to total cells. For reporter gene studies, 0.5 μg of the luciferase gene under control of the interferon stimulated response element (ISRE) in plasmid pISRE-luc (ISRE promoter; Stratagene; La Jolla, CA), was transfected into cells in duplicate or triplicate. In certain experiments a dominant negative p38 (p38 AGF) expressing plasmid [27], provided by Dr. Michael Kracht, was transfected into cells. Twenty-four hours post-transfection, ethanol, either alone or in combination with IFN was added directly to cells. Six hours later, luciferase activity was measured on cell lysates and normalized for transfection efficiency and total protein content.
Western blot analysis
Protein lysates were quantitated by BCA Protein Assay (Pierce; Rockford, IL) and equal amounts (typically 20–30 μg) of total protein was separated on 4–20% SDS-PAGE gels. For detection of phosphorylated Stat1 proteins, Stat1 phosphotyrosine (Y701) and phosphoserine (S727) specific antibodies were used (Zymed-Invitrogen). Total Stat1 was detected using a polyclonal antibody (Zymed or Santa Cruz Biotechnology; Santa Cruz, CA). Total and phosphorylated forms of p42/44 (ERK2/1), and p38 MAPK were detected with specific antisera (Cell Signaling; Beverly, MA). Cytochrome P4502E1 (CYP2E1) was detected using polyclonal rabbit antiserum (provided by Arthur Cederbaum), while alcohol dehydrogenase (ADH) was detected with a mouse monoclonal antibody (AbCam; Cambridge, MA). HCV proteins were detected using random, de-identified HCV infected patient serum, as described [28]. Infected serum was inactivated by adding triton X-100 to 1% prior to use.
Kinase assays
The activity of p38 MAPK in Huh7 cells was assessed via kinase assay using a kit (Cell Signaling). Briefly, cell lysates were immunoprecipitated with antibodies that recognize the phosphorylated form of p38 MAPK. After stringent washing, recombinant ATF-2 protein, a substrate for p38, was added to immunoprecipitates and incubated for 30 minutes according to manufacturer's specifications. Phosphorylated protein ATF-2 was detected by western blot.
HCV RNA quantitation
HCV RNA was quantitated by real time RT-PCR, using a modified version of a recent procedure [29]. Total cellular RNA was isolated from replicon cells using a commercial kit (Qiagen). Ten nanograms of RNA was added to wells of a 384 well plate containing the EZ RT-PCR master mix (Perkin Elmer; Wellesley, MA). Samples were run on an ABI HT7900 real time RT-PCR machine. The RT reaction consisted of 50°C for 2 minutes followed by 60°C for 30 minutes. The PCR consisted of an initial denaturation of 2 minutes at 95°C, then 45 cycles of 95°C for 15 seconds followed by simultaneous annealing/extension at 60°C for 1 minute. For each run, dilutions of BB7 plasmid DNA (precisely quantitated using the PicoGreen DNA quantitation kit (Invitrogen)) ranging from 0–107 copies per tube were run in triplicate to generate a standard curve, which served as a reference to calculate HCV RNA copy number based on the cycle threshold (Ct). The HCV RNA copy number is reported as copies per 10 ng total cellular RNA. Additional controls included reactions lacking template as well as RNA from Huh7 cells. For both negative controls, these samples were always negative for HCV RNA.
ADH enzyme assay
Cells were harvested in PBS and whole cell extracts prepared via sonication. Aliquots of protein extracts were mixed with 0.1 M glycine pH 10.0 buffer, 2.4 mM β-nicotinamide adenine dinucleotide, and 33 mM ethanol, and conversion of NAD to NADH+ was monitored with a spectrophotometer at a wavelength of 340 nm. All reagents for the assay were from Sigma. As a positive control, purified human ADH (provided by Dr. Carol Stone) was also run in the assay.
Statistics
Differences between means of luciferase readings were compared using a Student's T-test. A p-value of <0.05 was considered significant. For western blots, data were analyzed with Image J, a software version of NIH Image for the Macintosh OS × operating system. Changes in protein levels were normalized to control western blots and expressed as fold or percent change relative to controls.
Results
Effect of acute ethanol on Jak-Stat pathway
Figure 1A depicts the effects of acute ethanol on the ISRE promoter in Huh7 cells. Ethanol did not appear to have significant effects on ISRE activity at 25 and 50 mM concentrations. However, at concentrations of 100 and 200 mM, ethanol caused statistically significant 3.0 (p = 0.03) and 5.0 (p < 0.001) fold increases in ISRE reporter gene activity, as compared to cells not treated with ethanol. The data suggest that high physiological doses of acute ethanol activate the ISRE, an IFN responsive promoter.
Figure 1 Effect of acute ethanol on the Jak-Stat pathway. A, high physiological doses of ethanol activate the ISRE. Huh7 cells were transfected with 0.7 μg of pISRE-luc, and 24 hours later, cells were stimulated with ethanol at the indicated concentrations. Protein lysates were assayed for luciferase activity 6 hours later. Error bars represent standard deviations. The experiment was repeated 6 times with similar results. B, acute ethanol induces Stat1 serine phosphorylation. Huh7 cells were left as untreated controls (C) or treated with 1,000 U/ml of IFN-α (IFN), or 100 mM ethanol (EtOH). Twenty minutes later, equal amounts of whole cell protein extracts were separated by SDS-PAGE, and blotted for phosphorylated Stat1 S727 (toppanel), Stat1 Y701 (second panel), Stat2 Y690 (third panel), and total forms of the Stat1 protein (lower panel). The figure is representative of 3 independent experiments, which yielded similar results.
To investigate this regulation further, we analyzed levels of phosphorylated Stat1 and Stat2, which are obligatory steps for ISRE activation. Stat1 and Stat2 activation involves phosphorylation on conserved tyrosines at amino acid positions 701 and 690, respectively, while phosphorylation of Stat1 also occurs on conserved serine amino acid at position 727 and provides maximal transcriptional activation [15]. Figure 1B depicts the levels of Stat1 S727 (top panel), Stat1 Y701 (second panel), and Stat2 Y690 (third panel), and the total levels of Stat1 protein (fourth panel) in Huh7 cells. Phosphorylation of Stat1 on S727 was induced by IFN-α or 100 mM ethanol. Stat1 Y701 and Stat2 Y690 phosphorylation occurred with IFN treatment, whereas no effect was observed with 100 mM ethanol. The differences in Stat phosphorylation were not due to differences in the amount of Stat1 protein, since total Stat1 protein levels were equivalent (Figure 1B, lower panel). Similar results were also observed for primary human fetal hepatocytes (see Additional File 2) and HCV replicon cells (data not shown).
Effect of acute ethanol on the p38 MAPK pathway
Since MAPKs are modulated by ethanol [12] and p38 MAPK is important in ISRE transcription [16,19], we next examined the effect of acute alcohol on the p38 MAPK pathway. Figure 2 depicts the effects of acute ethanol on the p38 MAPK pathway in Huh7 cells and primary human fetal hepatocytes. In these experiments, p38 kinase assays were performed. As shown in the upper panel of Figure 2, acute exposure of Huh7 cells to 25, 50, and 100 mM ethanol resulted in 61, 27, and 150-fold activation of p38 kinase activity, respectively, detected as an increase in recombinant ATF-2 phosphorylation, a natural substrate for p38 MAPK. The middle panel depicts the amounts of total p38 protein added to each immunoprecipitate. Acute ethanol at 25, 50, and 100 mM doses also activated p38 MAPK to levels 2.1, 2.2, and 5.2-fold in primary fetal human hepatocytes, relative to untreated cells (Figure 2, lower panel), although basal levels of MAPK were higher in these cells. The data suggest that acute ethanol activates p38 MAPK pathways in primary human fetal hepatocytes and Huh7 cells. Acute ethanol also activated p42/44 MAPK and SAPK in Huh7 (see Additional File 3) and BB7 replicon cells (data not shown).
Figure 2 Ethanol activates p38 MAPK in human liver cell cultures. Huh7 cells or primary human fetal hepatocytes (HFH) were left as untreated controls (C) or were treated with anisomycin (A) as a positive control, or with 25, 50, or 100 mM ethanol for 30 minutes. Active p38 MAPK was immunoprecipitated from cell lysates and kinase activity measured by phosphorylation of ATF-2. The figure is representative of 2 experiments that produced similar results.
Acute alcohol stimulation of the Jak-Stat pathway involves MAPKs
Since the p38 MAPK pathway cross-talks to the Jak-Stat pathway [16,19], we investigated the effect of acute alcohol on ISRE transcriptional activity and Stat1 phosphorylation in the presence of MAPK inhibitors and dominant negative mutants. We performed ISRE reporter gene experiments with IFN-α and alcohol treatments in the presence of the p38 MAPK inhibitor, SB203508. As shown in Figure 3A, SB203508 inhibited ethanol stimulation of the ISRE by up to 40%. Figure 3B presents related experiments examining the effect of small molecule inhibitors on ethanol induction of Stat1 serine phosphorylation. Huh7 cells were treated for 2 hours in the presence of DMSO carrier, UO126 (a p42/44 MAPK inhibitor), PD98059 (a MEK1 inhibitor) and SB203508 (a p38 inhibitor). Cells were then stimulated with 100 mM ethanol for 20 minutes. Ethanol induction of Stat1 serine phosphorylation was 90% inhibited by SB203508. Huh7 cells were also transfected with a vector expressing a dominant negative p38 protein (p38 AGF) [27], and the effect on ethanol induction of Stat1 serine phosphorylation was investigated. As shown in Figure 3C, expression of the p38 AGF dominant negative mutant abrogated both basal and ethanol induced Stat1 serine phosphorylation. Together, the data suggest that acute ethanol activation of p38 MAPK is partially involved in induction of ISRE transcription and Stat1 serine phosphorylation.
Figure 3 Involvement of p38 MAPK in ethanol induction of ISRE transcription and Stat1 serine phosphorylation. Panel A, Huh7 cells were transfected with 0.7 μg pISRE-luc, and 22 hours later, cells were incubated for 2 hours at the indicated μM concentrations of SB203508 (a p38 inhibitor), followed by 100 mM ethanol. Cell lysates were harvested 6 hours later and luciferase results were normalized to amounts of total cellular protein. Error bars represent standard deviations. The experiment was repeated 4 times with identical results. B, Huh7 cells were treated for 2 hours in the presence of 50 μM of various MAPK inhibitors, and stimulated with 100 mM alcohol for 20 minutes. Whole cell protein extracts were blotted for the serine phosphorylated form (S727) or total form of Stat1. The experiment was repeated twice, yielding similar results. C, Huh7 cells were transfected with control vector plasmid (Vec) or a plasmid expressing a dominant negative mutant p38 protein (p38 AGF). Twenty-four hours later, cells were not treated or treated for 20 minutes with 100 mM ethanol. Levels of S727 and total Stat1 and transfected p38 proteins were determined by western blot. The figure is representative of 2 independent experiments that produced similar results.
Effect Of acute alcohol on HCV replicons
Figure 4 depicts the effects of acute ethanol on HCV replication. BB7 cells were treated once with 0, 25, 50, or 100 mM of alcohol, or 20 U/ml of IFN-α. HCV RNA was quantitated using real time RT-PCR on equal amounts (10 ηg) of total cellular RNA isolated 72 hours after drug treatment (Figure 4A; left panel). As expected, IFN induced a significant 66-fold inhibition of HCV RNA at this time point. A single administration of 25 mM ethanol had no significant effect on HCV RNA replication, although a slight increase was noted. In contrast, 50 mM and 100 mM ethanol doses induced statistically significant inhibition of HCV RNA synthesis (p = 0.02 and p = 0.001, respectively). The doses of alcohol used did not affect BB7 cell growth, viability, or morphology (data not shown). HCV NS3 and NS5A protein expression was also inhibited in a dose-dependent fashion (1.7–5.4 fold) by ethanol (Figure 4B, right panel).
Figure 4 Effect of acute ethanol on HCV replication. BB7 replicon cells were treated once with 0, 25, 50, or 100 mM ethanol or 20 U/ml of IFN, and RNA and protein was harvested 72 hours later. A, HCV RNA copy number was determined by quantitative real time RT-PCR. The HCV RNA copy number is reported as copies per 10 ng total cellular RNA. Error bars represent standard deviations. B, HCV protein expression in BB7 cells treated with 0, 25, 50, and 100 mM ethanol, and control Huh7 cells. The positions of Stat1, HCV NS3 and NS5A proteins are indicated. The experiment was repeated twice with similar results.
Since replicons acquire adaptive mutations [20,23] and it is also possible that the cells acquire genetic or epigenetic mutations during the G418 selection process and continuous culturing [30,31], we questioned whether the previous data derived from a single replicon cell line was typical of other replicons. We therefore examined the effect of acute ethanol on HCV RNA and protein synthesis in 2 additional replicon lines, 9–13 and 5–15, obtained independently from BB7 cells [21-23]. Cells were treated with 100 or 200 mM of ethanol, and HCV RNA and protein was assessed 72 hours later. As shown in Figure 5, although the basal level of HCV RNA differed considerably between 5–15 and 9–13 replicon cells, both doses of ethanol inhibited HCV RNA and protein production by up to 50%. Together the data suggest that acute ethanol inhibits the replication of several independent cell lines that support robust HCV replication. The replication of a genomic length replicon cell line, FL-Neo, was also inhibited by acute ethanol (see Additional File 4).
Figure 5 Acute ethanol inhibits the replication of other HCV replicon lines. 9–13, and 5–15-replicon cell lines were treated with 0, 100, or 200 mM of ethanol, and HCV RNA (panel A) and protein (panel B) was quantitated by real time RT-PCR and western blot analysis as described above. B., quantitation of changes in HCV NS3 and NS5A protein expression. Scanned blots were analyzed with Image J. For each lane, pixel intensities of NS3 and NS5A bands were normalized to the total Stat1 pixel intensity, and the percent change relative to untreated cells was calculated.
Acute alcohol inhibits the IFN-α induced antiviral response towards HCV
We examined the combined effects of alcohol and IFN-α treatment on the Jak-Stat pathway. Huh7 cells were left untreated, or treated with IFN, or IFN plus ethanol. Figure 6A demonstrates that ethanol treatment inhibited IFN-α induction of Stat1 tyrosine phosphorylation. To investigate the effect of ethanol on the IFN induced antiviral response, BB7 replicon cells were treated with or without 100 mM ethanol in the presence of varying doses of IFN-α. HCV protein levels were analyzed by western blot 48 hours later. As shown in Figure 6B, in the absence of ethanol, IFN-α inhibited HCV protein in a dose dependent fashion, and this coincided with a dose-dependent increase in total Stat1 protein, a known ISG. When cells were treated with a single dose of 100 mM ethanol, increases in HCV NS3 and NS5A proteins were detected at IFN doses of 10, 20 and 100 U/ml relative to cells treated with IFN alone. Alcohol also inhibited IFN induced up-regulation of Stat1 at these concentrations. However, at IFN concentrations of 0, 0.1 and 1 U/ml, ethanol appeared to decrease the amount of HCV NS3 and NS5A protein expression, consistent with ethanol's IFN stimulatory and anti-HCV effects presented above. Figure 6C presents a quantitative summary of the protein data based on pixel intensity, and clearly demonstrates that IFN dose-dependently inhibits NS3 and NS5A protein expression by 10–100 fold. In the absence of IFN, acute ethanol inhibits NS3 and NS5A protein expression by 10-fold. However, acute ethanol prevents IFN-α-mediated clearance of HCV proteins. Similar effects were observed for HCV RNA production (data not shown). The data indicate that ethanol inhibits the antiviral actions of exogenously added IFN.
Figure 6 Acute alcohol inhibits the antiviral actions of IFN. A, Huh7 cells were left untreated (Ctrl), or treated with 1,000 U/ml of IFN-α alone (IFN) or with IFN-α plus 100 mM ethanol (IFN+EtOH). Proteins were probed for Stat1 Y701 (top panel), and total Stat1 proteins (lower panel). B, BB7 replicon cells were treated with or without 100 mM ethanol, followed immediately by 0, 0.1, 1, 10, 20, 100 IU/ml of IFN-α, and whole cell protein extracts were harvested 48 hours later. Equal amounts of total cellular protein were analyzed for the presence of Stat1, HCV NS5A and NS3 proteins, and p42/44 MAPK by western blot analysis. C, quantitation of HCV protein expression shown in panel B. For each lane, pixel intensities of NS3 and NS5A bands were normalized to the total p42/44 pixel intensity, and the fold decrease relative to untreated cells was calculated. The figure is representative of 2 independent experiments that produced identical results.
Expression of alcohol metabolizing enzymes in human liver cell cultures
Since ethanol can exert differential effects on cells depending on whether it is metabolized or not [32], the expression and activity of ADH and CYP2E1, the major ethanol-metabolizing enzymes, was examined in Huh7 and replicon cells. Figure 7A depicts western blot analysis of ADH (top panel), CYP2E1 (middle panel), and Stat1 (lower panel) protein expression in Huh7, BB7, 9–13, 5–15, and FL-Neo cells. Immortalized human hepatocytes (HH2), primary human fetal hepatocytes (HFH), and purified human ADH served as positive controls for ADH, while lysate from cells that were infected with a baculovirus expressing human CYP2E1 [33], as well as purified CYP2E1, served as controls for CYP2E1. All replicon and Huh7 cultures expressed very low to undetectable levels of ADH and CYP2E1 protein. To determine if Huh7 cells expressed a functional ADH enzyme, ADH enzyme assays were performed using purified human ADH as a positive control. Figure 7B demonstrates that purified ADH showed a linear increase in absorbance over time, while buffer alone remained at background levels. In contrast, Huh7 cells expressed minimal ADH enzymatic activity, with only slight increases over background detected after 3 minutes. Figure 7C demonstrates that ethanol induction of ISRE transcription was not affected in the presence of the ADH and CYP2E1 inhibitors, 4-MP and DAS. Note that the concentrations of 4-MP (5 mM) and DAS (10 μM) used in this assay were derived from a previous study [25]. At these concentrations, 4-MP and DAS had no effects on cell viability or proliferation (data not shown). The data indicate that Huh7 and replicon cells express low to undetectable levels of ADH and CYP2E1 proteins, and further suggest that the effects of ethanol on innate antiviral pathways is not due to ethanol metabolism via ADH or CYP2E1 in this model system.
Figure 7 Characterization of ethanol metabolizing enzymes in human liver cell cultures. A, western blot analysis of ADH1 and CYP2E1 expression levels in Huh7, BB7, 9–13, 5–15, and FL-Neo cells. Positive controls for ADH included primary human fetal hepatocytes (HFH) [24], and a well differentiated immortalized human liver cell line, HH2 (developed in NF's lab), while controls for CYP2E1 expression included baculovirus expressed CYP2E1 and purified CYP2E1. Western blots were probed with a monoclonal antibody against human ADH, and polyclonal rabbit antiserum against CYP2E1 and Stat1. B, ADH enzyme activity. Huh7 cells were harvested in PBS and whole cell protein extracts prepared via sonication. Conversion of NAD to NADH+ was monitored at a wavelength of 340 nm as described in the Materials and Methods. Purified ADH served as a positive control for ADH activity. C, effect of CYP2E1 and ADH inhibition on ethanol activation of the ISRE. Huh7 cells in 96 well plates were transfected in triplicate with 50 ηg of ISRE-luc and 12 hours later, were treated with 5 mM of the ADH inhibitor 4-MP and 10 mM of the CYP2E1 inhibitor DAS for an additional 12 hours. Cells were also separately exposed to 0.1% DMSO, as an additional control for possible solvent effects. Cells were then treated with 0, 100, or 200 mM ethanol, before luciferase activity was measured by BriteLite assay. Error bars represent standard deviations. The experiments were repeated twice with identical results.
Discussion
In the current study, it was demonstrated that high physiological doses of acute ethanol induces Stat1 serine phosphorylation and ISRE transcription. Given alone, ethanol appears to inhibit HCV replication in several independent replicon cell lines, and this is in part mediated by a Jak-Stat transduced antiviral response. In contrast, in the presence of exogenously added IFN-α, ethanol partially inhibits the antiviral actions of IFN-α, involving inhibition of IFN-α induced Stat1 tyrosine phosphorylation. Analysis of the effects of chronic ethanol administration on basal and IFN-α induced signaling responses is currently in progress.
We also found that acute exposure of human liver cells to physiological doses of ethanol activates the IFN system via the MAPK pathway. The data suggest that ethanol induces cross talk between the p38 MAPK and Jak-Stat pathway (Figure 8). Additional evidence for cross talk between these pathways derives from a study indicating that ERK2 binds to the α-chain of the IFN α/β receptor and STAT1 [34], and JAK2 may be required for MAP kinase pathway activation [35]. Furthermore, HCV proteins such as NS5A interact with and modulate MAPK and related pathways such as Grb2, Ras-ERK, and phosphoinositol 3 kinase (PI3K) [36-40]. However, p38 kinase activity, which is important in IFN-α and IFN-γ induced transcription, is not involved in IFN induced Stat1 serine phosphorylation [16,19]. Thus, induction of Stat1 serine phosphorylation by ethanol described in the current report may be mechanistically similar to UV-stress induced activation of Stat1 by p38 MAPK [18].
Figure 8 Summary of effects of acute ethanol on HCV replication. Ethanol effects in this system are independent of ethanol metabolism and as such may involve ethanol-induced perturbations in cell membranes, such as membrane fluidity. Left side, acute ethanol activates p38 MAPK which leads to Stat1 serine phosphorylation, Jak-Stat signaling and inhibition of HCV replication. Activated Stat1 may be involved in ISRE transcription but it is possible that other ISRE binding transcription factors such as Stat3 are involved in this process. Right side, ethanol inhibits the antiviral actions of exogenously applied IFN and this involves inhibition of IFN-induced Stat1 tyrosine phosphorylation, decreased Jak-Stat signaling and increased HCV replication in the presence of IFN. Inhibition of Jak-Stat signaling may involve ethanol perturbation of IFN-α induced changes in membrane fluidity, inhibition of IFN binding to its receptor, direct inhibition of Jak kinases, and/or induction of negative regulators of the Jak-Statpathway such as SOCS proteins.
Recent studies have demonstrated that alcohol abuse may be associated with increased HCV RNA titers in patients [9]. This could be due to an increase in release of HCV RNA from alcohol-damaged hepatocytes, a direct stimulatory effect of alcohol on HCV replication, or modulation of innate and acquired immune responses to HCV. A single published report by Zhang and colleagues found that ethanol stimulates HCV replication in the replicon system [41], while our data indicate that acute ethanol inhibits HCV replication. There are several explanations for the divergent results. First, different stable replicon cell lines were used in our study as compared to the published study, so it is very likely that both the replicons [21] and Huh7 cells [30,31] are genetically different. Second, in Zhang's study, alcohol was added to replicon cells daily, so 48 and 72-hour time points actually received 2 and 3 daily doses of ethanol. This is in direct contrast to our experimental design where a single "shot" of alcohol was given. Nonetheless, chronic ethanol treatment of cells for 3 consecutive days further inhibited HCV replication in our system (data not shown). Third, in our study, the observed effects on the IFN system and HCV replication appeared to be due to the direct action of ethanol, rather than via ethanol metabolism, as reported in the Zhang study [41]. However, the dose of the ADH inhibitor 4-methypyrazole used in Zhang's study was 0.1 μM, 50,000 fold lower than the 5 mM dose used in our study, and the dose used in a seminal study demonstrating the effect of various inhibitors of alcohol metabolism [25]. Further evidence for a direct effect of ethanol for the observed results in our study stems from the observation that all replicon and Huh7 cells expressed low to undetectable levels of ADH and CYP2E1 protein, and ethanol still induced ISRE transcription in the presence of ADH and CYP2E1 inhibitors. The Zhang study did not measure ADH and CYP2E1 protein expression. Furthermore, in our studies, the effects of ethanol on the Jak-Stat pathway occurred at an ethanol concentration of 100 mM, well above that of the Km for ADH (1–5 mM) and CYP2E1 (16 mM) [42]. Moreover, high-dose ethanol has been previously shown to activate IFN-β-dependent antiviral activities [43], reminiscent of the data reported in the present study. Collectively, our data suggest that ethanol acts directly on cells to modulate hepatocyte signaling pathways.
Exactly how ethanol induces these signaling responses is currently under investigation. Ethanol is known to act on lipids in cell membranes as well as interact directly with membrane proteins [44-47], so it is possible that changes in membrane fluidity (defined as the physical state of the phospholipids in terms of rate and angular motion) induce downstream signal transduction events (Figure 8). In terms of the activation of the Jak-Stat pathway by acute ethanol, it is possible that besides Stat1, other proteins with the capacity to bind ISRE-like sequences are involved in ethanol induced ISRE transcription. A possible candidate is Stat3, since Stat3 is modulated by ethanol [48]. Indeed, preliminary data suggest that Stat3 is also modulated by acute ethanol in our system (data not shown). As for ethanol inhibition of IFN-induced Stat1 tyrosine phosphorylation and antiviral actions, several mechanisms might be operative (Figure 8). Since IFN-β has been shown to modulate plasma membrane fluidity [49], ethanol might inhibit IFN-α induced changes in membrane fluidity. Other possible mechanisms include ethanol inhibition of IFN-receptor interactions, or induction of negative regulators of the Jak-Stat pathway such as suppressors of cytokine signaling (SOCS) proteins. For example, SOCS-1 inhibits IFN signaling by binding Jaks to prevent Stat phosphorylation [50]. Also of note is the observation that ethanol doses of 1–20 mM did not affect HCV replication (Figure 4A), so it is possible that ethanol-induced blockade of IFN antiviral activity is more relevant in vivo. The data presented herein highlight the complexity, and emphasize the need for further study of the cellular response to acute and chronic alcohol, on innate antiviral signaling pathways and HCV replication.
In conclusion, acute ethanol treatment of Huh7 hepatoma, HCV subgenomic and genomic-length replicon cells, and primary human fetal hepatocytes has multiple effects on innate cellular defense pathways. In particular, high physiological doses of ethanol can activate antiviral responses and inhibit HCV replication, whereas it can also inhibit the IFN-α induced antiviral response against HCV replication. The data suggest that the effects of alcohol on the IFN system are not simply a stimulation or inhibition, but rather reflect highly complex processes involving cross-talk of a number of signaling pathways. The net effect of ethanol likely depends on whether ethanol is given acutely or chronically, the dose of ethanol, and whether alcohol is metabolized or not.
Abbreviations
ADH: alcohol dehydrogenase
CYP2E1: cytochrome P450 2E1
DAS: diallysulfide
ERK: extracellular regulated kinase
HCV: hepatitis C virus
IFN: interferon
IFN-α: interferon alpha
ISG: interferon-stimulated gene
ISGF-3: interferon stimulated gene factor 3
ISRE: interferon stimulated response element
Jak: janus associated kinase
MAPK: mitogen activated protein kinase
RLU: relative light units
Stat: signal transducer and activator of transcription
4-MP: 4-methypyrazole
Supplementary Material
Additional File 1
Effect of acute ethanol on Huh7 (panel A) FL-Neo (panel B) genomic length replicon cell viability and proliferation. Cells were treated with once 0, 50, or 100 mM ethanol, and incubated at 37°C humidified incubator with 5% CO2 for 72 hours. Cells were lysed and total cellular ATP content measured by luciferase assay using the ATPlite system (Perkin Elmer). Error bars represent standard deviations of quadruplicate cultures. The experiment was repeated three times with identical results.
Click here for file
Additional File 2
Ethanol activation of Stat1 serine phosphorylation in primary human fetal hepatocytes. Cells were treated with 0, 25, and 50 mM ethanol or separately with 100 U/ml of IFN-α for 30 minutes, and blots were probed for Stat1 S727, Y701 and total proteins. The experiment was repeated twice with identical results.
Click here for file
Additional File 3
Ethanol activates p42/44 MAPK in Huh7 cells. Huh7 cells were grown in 0.5% serum-containing media for 48 hours, and stimulated with ethanol alone at the indicated concentrations, or with ethanol and 20% serum-containing medium. Thirty minutes later, equal amounts of whole cell protein extracts were separated by SDS-PAGE and blotted for phosphorylated forms of p42/44 (panel 1) or JNK (panels 2), or total forms of p42/44 (panel 3).
Click here for file
Additional File 4
Acute ethanol inhibits HCV replication in a genomic length replicon cell line. FL-Neo replicon cells were treated with 0, 100, or 200 mM of ethanol, and HCV RNA was quantitated by real time RT-PCR. The HCV RNA copy number is reported as copies per 10 ng total cellular RNA. Error bars represent standard deviations.
Click here for file
Acknowledgements
We thank Michael Austin, John Gallegos, Jacob Glaspey, Jamison Green, Amanda Heitzke, Kristen Miller, Paula McPoland, and Jessica Wagoner for technical assistance, Jeffery Vieira for pQ150, Arthur Cederbaum for CYP2E1 antiserum and helpful advice, Dennis Rasmussen for advice, Carol Stone for recombinant human ADH, Sidney Nelson for recombinant CYP2E1, Apath LLC and Ralf Bartenschlager for HCV replicon cell lines, and Michael Kracht for p38 plasmids. CL and NF are partially supported by NIH grant AI048214. SJP is partially supported by NIH grants AA13301 and DK62187, and the University of Washington Royalty Research Fund.
==== Refs
Schiff ER Hepatitis C and alcohol Hepatology 1997 26 39S 42S 9305662 10.1002/hep.510260707
Thomas DL Astemborski J Rai RM Anania FA Schaeffer M Galai N Nolt K Nelson KE Strathdee SA Johnson L The natural history of hepatitis C virus infection - Host, viral, and environmental factors JAMA 2000 284 450 456 10904508 10.1001/jama.284.4.450
Lindsay KL Therapy of hepatitis C: overview Hepatology 1997 26 71S 77S 9305668 10.1002/hep.510260713
Davis GL Esteban-Mur R Rustgi V Hoefs J Gordon SC Trepo C Shiffman ML Zeuzem S Craxi A Ling MH Interferon alfa-2b alone or in combination with ribavirin for the treatment of relapse of chronic hepatitis C New Engl J Med 1998 339 1493 1499 9819447 10.1056/NEJM199811193392102
McHutchison JG Gordon SC Schiff ER Shiffman ML Lee WM Rustgi VK Goodman ZD Ling MH Cort S Albrecht JK Interferon alfa-2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C N Engl J Med 1998 339 1485 1492 9819446 10.1056/NEJM199811193392101
Poynard T Marcellin P Lee SS Niederau C Minuk GS Ideo G Bain V Heathcote J Zeuzem S Trepo C Randomised trial of interferon alpha2b plus ribavirin for 48 weeks or for 24 weeks versus interferon alpha2b plus placebo for 48 weeks for treatment of chronic infection with hepatitis C virus. International Hepatitis Interventional Therapy Group Lancet 1998 352 1426 1432 9807989 10.1016/S0140-6736(98)07124-4
Manns MP McHutchison JG Gordon S Beaumont W Rustgi V Shiffman ML Lee WM Ling ML Albrecht JK Peginterferon alfa-2b plus ribavirin compared to interferon alfa-2b plus ribavirin for the treatment of chronic hepatitis C: 24 week treatment analysis of a multicenter, multinational phase III randomized control trial Hepatology 2000 32 A552 10.1053/jhep.2000.19149
Fried MW Shiffman ML Reddy KR Smith C Marinos G Goncales FL JrHaussinger D Diago M Carosi G Dhumeaux D Peginterferon alfa-2a plus ribavirin for chronic hepatitis C virus infection N Engl J Med 2002 347 975 982 12324553 10.1056/NEJMoa020047
Bhattacharya R Shuhart MC Hepatitis C and alcohol: interactions, outcomes, and implications J Clin Gastroenterol 2003 36 242 252 12590237 10.1097/00004836-200303000-00012
Widmann C Gibson S Jarpe MB Johnson GL Mitogen-activate protein kinase: conservation of a three-kinase module from yeast to humans Physiological Reviews 1999 79 143 180 9922370
Yang SH Sharrocks AD Whitmarsh AJ Transcriptional regulation by the MAP kinase signaling cascades Gene 2003 320 3 21 14597384 10.1016/S0378-1119(03)00816-3
Aroor AR Shukla SD MAP kinase signaling in diverse effects of ethanol Life Sci 2004 74 2339 2364 15027449 10.1016/j.lfs.2003.11.001
Darnell JE Kerr IM Stark GR Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins Science 1994 264 1415 1421 8197455
Decker T Kovarik P Serine phosphorylation of STATs Oncogene 2000 19 2628 2637 10851062 10.1038/sj.onc.1203481
Wen Z Zhong Z Darnell JEJ Maximal activation of transcription by Stat1 and Stat3 requires both tyrosine and serine phosphorylation Cell 1995 241 250 7543024 10.1016/0092-8674(95)90311-9
Li Y Sassano A Majchrzak B Deb DK Levy DE Gaestel M Nebreda AR Fish EN Platanias LC Role of p38alpha Map kinase in Type I interferon signaling J Biol Chem 2004 279 970 979 14578350 10.1074/jbc.M309927200
Platanias LC The p38 mitogen-activated protein kinase pathway and its role in interferon signaling Pharmacol Ther 2003 98 129 142 12725866 10.1016/S0163-7258(03)00016-0
Kovarik P Mangold M Ramsauer K Heidari H Steinborn R Zotter A Levy DE Muller M Decker T Specificity of signaling by STAT1 depends on SH2 and C-terminal domains that regulate Ser727 phosphorylation, differentially affecting specific target gene expression Embo J 2001 20 91 100 11226159 10.1093/emboj/20.1.91
Ramsauer K Sadzak I Porras A Pilz A Nebreda AR Decker T Kovarik P p38 MAPK enhances STAT1-dependent transcription independently of Ser-727 phosphorylation Proc Natl Acad Sci U S A 2002 99 12859 12864 12232043 10.1073/pnas.192264999
Blight KJ Kolykhalov AA Rice CM Efficient initiation of HCV RNA replication in cell culture Science 2000 290 1972 1974 11110665 10.1126/science.290.5498.1972
Krieger N Lohmann V Bartenschlager R Enhancement of hepatitis c virus RNA replication by cell culture- adaptive mutations J Virol 2001 75 4614 4624 11312331 10.1128/JVI.75.10.4614-4624.2001
Lohmann V Korner F Koch J Herian U Theilmann L Bartenschlager R Replication of subgenomic hepatitis C virus RNAs in a hepatoma cell line Science 1999 285 110 113 10390360 10.1126/science.285.5424.110
Lohmann V Korner F Dobierzewska A Bartenschlager R Mutations in hepatitis C virus RNAs conferring cell culture adaptation Journal Of Virology 2001 75 1437 1449 11152517 10.1128/JVI.75.3.1437-1449.2001
Lazaro CA Croager EJ Mitchell C Campbell JS Yu C Foraker J Rhim JA Yeoh GC Fausto N Establishment, characterization, and long-term maintenance of cultures of human fetal hepatocytes Hepatology 2003 38 1095 1106 14578848 10.1053/jhep.2003.50448
Wu D Cederbaum AI Ethanol cytotoxicity to a transfected HepG2 cell line expressing human cytochrome P4502E1 J Biol Chem 1996 271 23914 23919 8798623 10.1074/jbc.271.39.23914
Easterbrook J Lu C Sakai Y Li AP Effects of organic solvents on the activities of cytochrome P450 isoforms, UDP-dependent glucuronyl transferase, and phenol sulfotransferase in human hepatocytes Drug Metab Dispos 2001 29 141 144 11159803
Winzen R Kracht M Ritter B Wilhelm A Chen CY Shyu AB Muller M Gaestel M Resch K Holtmann H The p38 MAP kinase pathway signals for cytokine-induced mRNA stabilization via MAP kinase-activated protein kinase 2 and an AU-rich region-targeted mechanism Embo J 1999 18 4969 4980 10487749 10.1093/emboj/18.18.4969
Polyak SJ Paschal D McArdle S Gale M Moradpour D Gretch DR Characterization of the Effects of Hepatitis C Virus Non-Structural 5A Protein Expression In Human Cell Lines and on Interferon-Sensitive Virus Replication Hepatology 1999 29 1262 1271 10094974 10.1002/hep.510290438
Cook L Ng K-W Bagabag A Corey L Jerome KR Ultrasensitive quantitation of hepatitis C virus RNA using the MagNA Pure automated nucleic acid extraction system followed by real-time RT-PCR Journal of Clinical Microbiology 2004
Blight KJ McKeating JA Marcotrigiano J Rice CM Efficient replication of hepatitis C virus genotype 1a RNAs in cell culture J Virol 2003 77 3181 3190 12584342 10.1128/JVI.77.5.3181-3190.2003
Sumpter R JrLoo YM Foy E Li K Yoneyama M Fujita T Lemon SM Gale M Jr Regulating intracellular antiviral defense and permissiveness to hepatitis C virus RNA replication through a cellular RNA helicase, RIG-I J Virol 2005 79 2689 2699 15708988 10.1128/JVI.79.5.2689-2699.2005
Nagy LE Molecular aspects of alcohol metabolism: transcription factors involved in early ethanol-induced liver injury Annu Rev Nutr 2004 24 55 78 15189113 10.1146/annurev.nutr.24.012003.132258
Chen W Peter RM McArdle S Thummel KE Sigle RO Nelson SD Baculovirus expression and purification of human and rat cytochrome P450 2E1 Arch Biochem Biophys 1996 335 123 130 8914842 10.1006/abbi.1996.0489
David M Petricoin E 3rdBenjamin C Pine R Weber MJ Larner AC Requirement for MAP kinase (ERK2) activity in interferon alpha- and interferon beta-stimulated gene expression through STAT proteins Science 1995 269 1721 1723 7569900
Winston LA Hunter T JAK2, Ras, and Raf are required for activation of extracellular signal- regulated kinase/mitogen-activated protein kinase by growth hormone J Biol Chem 1995 270 30837 30840 8537333 10.1074/jbc.270.52.30837
Macdonald A Crowder K Street A McCormick C Harris M The hepatitis C virus NS5A protein binds to members of the Src family of tyrosine kinases and regulates kinase activity J Gen Virol 2004 85 721 729 14993658 10.1099/vir.0.19691-0
Street A Macdonald A Crowder K Harris M The Hepatitis C virus NS5A protein activates a phosphoinositide 3-kinase-dependent survival signaling cascade J Biol Chem 2004 279 12232 12241 14709551 10.1074/jbc.M312245200
Macdonald A Crowder K Street A McCormick C Saksela K Harris M The hepatitis C virus non-structural NS5A protein inhibits activating protein-1 function by perturbing ras-ERK pathway signaling J Biol Chem 2003 278 17775 17784 12621033 10.1074/jbc.M210900200
Tan SL Nakao H He YP Vijaysri S Neddermann P Jacobs BL Mayer BJ Katze MG NS5A, a nonstructural protein of hepatitis C virus, binds growth factor receptor-bound protein 2 adaptor protein in a Src homology 3 domain/ligand-dependent manner and perturbs mitogenic signaling Proc Nat Acad Sci USA 1999 96 5533 5538 10318918 10.1073/pnas.96.10.5533
He Y Nakao H Tan SL Polyak SJ Neddermann P Vijaysri S Jacobs BL Katze MG Subversion of cell signaling pathways by hepatitis C virus nonstructural 5A protein via interaction with Grb2 and P85 phosphatidylinositol 3-kinase J Virol 2002 76 9207 9217 12186904 10.1128/JVI.76.18.9207-9217.2002
Zhang T Li Y Lai JP Douglas SD Metzger DS O'Brien CP Ho WZ Alcohol potentiates hepatitis C virus replicon expression Hepatology 2003 38 57 65 12829987 10.1053/jhep.2003.50295
Lands WE A review of alcohol clearance in humans Alcohol 1998 15 147 160 9476961 10.1016/S0741-8329(97)00110-9
Chelbi-Alix MK Chousterman S Ethanol induces 2',5'-oligoadenylate synthetase and antiviral activities through interferon-beta production J Biol Chem 1992 267 1741 1745 1309794
Franks NP Lieb WR Molecular and cellular mechanisms of general anaesthesia Nature 1994 367 607 614 7509043 10.1038/367607a0
Peoples RW Li C Weight FF Lipid vs protein theories of alcohol action in the nervous system Annu Rev Pharmacol Toxicol 1996 36 185 201 8725387 10.1146/annurev.pa.36.040196.001153
Wilkemeyer MF Pajerski M Charness ME Alcohol inhibition of cell adhesion in BMP-treated NG108-15 cells Alcohol Clin Exp Res 1999 23 1711 1720 10591586
Wilkemeyer MF Sebastian AB Smith SA Charness ME Antagonists of alcohol inhibition of cell adhesion Proc Natl Acad Sci U S A 2000 97 3690 3695 10725368 10.1073/pnas.050545697
Chen J Kunos G Gao B Ethanol rapidly inhibits IL-6-activated STAT3 and C/EBP mRNA expression in freshly isolated rat hepatocytes FEBS Lett 1999 457 162 168 10486586 10.1016/S0014-5793(99)01031-5
Pfeffer LM Landsberger FR Tamm I Beta-interferon-induced time-dependent changes in the plasma membrane lipid bilayer of cultured cells J Interferon Res 1981 1 613 620 6180094
Kubo M Hanada T Yoshimura A Suppressors of cytokine signaling and immunity Nat Immunol 2003 4 1169 1176 14639467 10.1038/ni1012
|
16324217
|
PMC1318489
|
CC BY
|
2021-01-04 16:38:57
|
no
|
Virol J. 2005 Dec 2; 2:89
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-89
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1701631646310.1186/1471-2164-6-170Research ArticleEfficient single nucleotide polymorphism discovery in laboratory rat strains using wild rat-derived SNP candidates Smits Bart MG [email protected] Victor [email protected] Dimphy [email protected] Dirk [email protected] Hans J [email protected] Edwin [email protected] Hubrecht Laboratory, Functional Genomics Group, Uppsalalaan 8, 3584 CT Utrecht, The Netherlands2 Institute for Laboratory Animal Science, Hannover Medical School, Carl-Neuberg-Str.1, 30625 Hannover, Germany2005 29 11 2005 6 170 170 23 9 2005 29 11 2005 Copyright © 2005 Smits et al; licensee BioMed Central Ltd.2005Smits 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 laboratory rat (Rattus norvegicus) is an important model for studying many aspects of human health and disease. Detailed knowledge on genetic variation between strains is important from a biomedical, particularly pharmacogenetic point of view and useful for marker selection for genetic cloning and association studies.
Results
We show that Single Nucleotide Polymorphisms (SNPs) in commonly used rat strains are surprisingly well represented in wild rat isolates. Shotgun sequencing of 814 Kbp in one wild rat resulted in the identification of 485 SNPs as compared with the Brown Norway genome sequence. Genotyping 36 commonly used inbred rat strains showed that 84% of these alleles are also polymorphic in a representative set of laboratory rat strains.
Conclusion
We postulate that shotgun sequencing in a wild rat sample and subsequent genotyping in multiple laboratory or domesticated strains rather than direct shotgun sequencing of multiple strains, could be the most efficient SNP discovery approach. For the rat, laboratory strains still harbor a large portion of the haplotypes present in wild isolates, suggesting a relatively recent common origin and supporting the idea that rat inbred strains, in contrast to mouse inbred strains, originate from a single species, R. norvegicus.
==== Body
Background
Genetic variation exists between individuals (or strains) of all organisms and it makes up the genetic basis for phenotypic differences between individuals. In addition, genetic variation functions as a valuable resource for mapping phenotypic traits in model organisms. Single Nucleotide Polymorphisms (SNPs) are the most abundant form of genetic variation and therefore dominate high-resolution genetic mapping strategies. Moreover, numerous well-performing high-throughout SNP detection technologies have been developed, like oligonucleotide array-based technology, mass-spectrometry-based technology (MALDI-TOF), and sequence-based technology (pyrosequencing, DHPLC) [1], which makes automated SNP detection favored above the more labor-intensive detection of microsatellite markers [2].
Since the availability of its genome, the laboratory rat is gaining influence as a genetic model organism [3]. In addition, over 200 well-characterized inbred strains that are models for a wide variety of human diseases are available [4,5]. However, the availability of genetic tools, like a dense genome-wide SNP marker set, is still subordinate compared to other commonly used model organisms. This is illustrated by the number of entries in dbSNP, the central SNP repository of NCBI [6]: the amount of human (>10,000,000), chicken (>3,000,000), and mouse (>500,000) entries surpass the amount of rat entries (>43,000) spectacularly. In search for rat SNPs, experimental [7,8] and computational [9] approaches have been employed, but these efforts primarily resulted in SNPs associated with coding regions. For genetic mapping purposes, a much denser marker set, preferentially equally distributed over the genome, is required.
Laboratory rat strains are thought to be established from a limited number of founder animals originating from a domesticated wild population [10,11]. The value of inbred strains emanates from the close genetic uniformity that facilitates phenotyping and genotyping. In principle, inbred strains are selectively bred for certain traits from a genetically diverse pool, comprising diverse genetic information about the trait. However, since many of the current rat strains were derived from common ancestral stocks and simply inbred to increase genetic uniformity, inbred strains clearly share alleles [12]. Although such simplified models are essential for biomedical research, modulating effects on the clinical manifestation of a trait resulting from genetic heterogeneity in a population can only be studied to a limited extent in F1 hybrids. The use of a carefully chosen selection of inbred strains may address this issue, but the choice depends on knowledge on the relationship between the strains and hence the degree of genetic variation. Alternatively, wild-derived strains may be good alternatives to introduce sufficient genetic variation in laboratory experiments [13,14].
Based on a preliminary observation that alleles from laboratory rat strains are frequently detected in wild-derived samples, we developed a wild rat-based SNP discovery approach. The method consists of shotgun sequencing of a wild rat-derived genomic library followed by comparison with the published rat genome (strain Brown Norway). Genotyping commonly used rat strains for newly identified SNPs revealed that 84% of SNP-alleles (and 87% of all genetic variation) occurring between BN and a single wild individual is also represented in one or more laboratory strains. A user-friendly webtool allows exploration of the genetic variation between any arbitrary combinations of two strains that were used in this study, making all information directly available for experimental use.
Results
Wild rat-based SNP discovery
It is generally believed that commonly used rat strains originate from a wild-derived founder population of limited size [10]. To examine whether polymorphisms found in laboratory strains are still represented in individuals of the wild population, we typed two wild-derived samples for confirmed SNPs of the CASCAD database [9]. Interestingly, about 53% of alleles (n = 147), which were confirmed to exist in laboratory strains, were also represented in wild 1, wild 2 or both (not shown). Hence, a preselection of highly likely candidate SNPs could potentially be made by genotyping wild individuals and comparing the sequences to the rat genome sequence (Brown Norway).
Accordingly, we performed random shotgun sequencing on a genomic library of a wild rat (wild 1). We generated shotgun traces (814 Kbp) by bidirectional sequencing of about 1,600 colonies (Table 1). 85.5% of the reads (2545/2975; Table 1) could be mapped to a unique location in the Brown Norway rat genome using BLAT [15], resulting in the automated identification of nearly 5,000 ambiguous nucleotide positions (potential polymorphisms). Manual inspection of the sequencing reads reduced this set of potential polymorphisms to a set of 746 real SNPs and 122 indels. The average SNP rate between BN (BN/SsNMcw; genome sequencing project) and this single wild rat is estimated to be about 1 per 900 bp and, hence, discovery of a novel SNP can be expected every second shotgun read. A subset of the discovered SNPs was verified and genotyped in 36 commonly used strains (including BN). To this end, we designed primers for 451 SNP-containing amplicons (about 300 bp) of which 416 (92.2%) were successfully read by unidirectional sequencing of the PCR products, resulting in roughly 119 Kbp high quality sequence per strain or individual (Table 1).
Table 1 Statistics on shotgun sequencing of the wild rat-derived genomic library
picked colonies 1632
readable sequence reads/sequenced bases 2975/814,440
uniquely mapped (BLAT) reads/bases 2545/768,683
ambiguous positions 4902
strong candidates after manual inspection 868 (746 SNPs + 122 indels)
successfully read/amplicons designed* 416/451 (~1.65 candidate SNP/amplicon) (92.2%)
amplified bases per strain or wild individual 118,971
* Amplicons are designed for the 746 SNP candidates.
Wild rat-derived SNP characteristics
The verification of 746 candidate SNPs by amplicon-based resequencing in 36 inbred rat strains and three wild-derived samples (wild 1, 2, and 3) revealed 960 polymorphisms, consisting of 90 indels, seven 2-bp substitutions, one 3-bp substitution, one 5-bp substitution, and 861 SNPs, of which only one was tri-allelic. The amplicons are randomly distributed over the genome (Fig. 1). We observed heterozygous positions in the outbred strains, but unexpectedly some were also found in the inbred strains (for detailed information: [see Additional file 1] or [6]). For our analysis, we considered these loci to be polymorphic as compared to the BN genome sequence.
Figure 1 Distribution of amplicons (451 loci) designed for verification and subsequent genotyping of candidate shotgun-based SNPs in 36 commonly used inbred strains.
From the 746 shotgun-based candidate SNPs, 685 were located in the 416 PCR amplicons that worked, and 485 (71%) were reconfirmed by resequencing (shotgun-based; Table 2). Strikingly, for 408 (84%) of the confirmed SNPs, the wild rat allele is also present in one or more commonly used strains, with only 36 (7.4%) being specific to BN (Table 2). Of the remaining 77 (16%) SNPs, wild rat alleles are not present in any of the 36 selected strains and could be considered wild rat-specific. These results illustrate that shotgun sequencing one wild individual efficiently identifies shared polymorphisms among commonly used rat strains.
Table 2 SNP discovery results
shotgun- based genotyping-based (only wild 1*) genotyping-based (wild 1, 2, 3*) total (only wild 1*) total (wild 1, 2, 3*)
BN specific 36 (7.4 %) 9 (2.5 %) 7 (1.9 %) 45 (5.3 %) 43 (5.0 %)
wild specific 77 (15.9 %) 12 (3.4 %) 30 (8.0 %) 89 (10.6 %) 107 (12.4 %)
in 35 strains, not in wild 0 204 (57.0 %) 156 (41.5 %) 204 (24.2 %) 156 (18.1 %)
in 35 strains, shared with wild 372 (76.7 %) 133 (37.1 %) 183 (48.7 %) 505 (59.9 %) 555 (64.5 %)
total 485 (100 %) 358 (100 %) 376 (100 %) 843 (100 %) 861 (100 %)
* By genotyping two other wild individuals (wild 2 and 3), additional polymorphisms were identified, which could not have been found by shotgun sequencing only wild 1.
While genotyping by resequencing, 358 novel SNPs were discovered that were not identified in the shotgun sequencing experiment (genotyping-based; Table 2). About 39% (139) of this set can be accounted for by differences in the sequence coverage between the shotgun reads and the resequencing genotyping reads (Table 2), whereas the remaining part of this set is strongly biased towards SNPs that are not polymorphic between BN and wild rat 1 and thus could not have been discovered in the shotgun experiment. Interestingly, about 37% of the newly discovered SNPs are polymorphic between the shotgun sequenced wild rat and any of the inbred strains (Table 2). When considering all SNPs that are polymorphic in the set of 36 commonly used laboratory strains, of the majority (66%) the wild rat allele is found back in one of the strains (total; Table 2) and this percentage increases only slightly (70%) when two other wild individuals (wild 2 and 3) are included in the analysis. This indicates that wild rat-based SNP discovery is already highly efficient using a single wild sample.
Based on the genotyping results, the SNP rate between BN and the shotgun sequenced wild rat (wild 1) is 1 SNP per 190 bp (626 SNPs/119 Kbp). The SNP rate within the 36 rat strains, including BN, is 1 in 158 (Table 2; 45+204+505 SNPs/119 Kbp) and the SNP rate in the entire experiment, including the wild rat (wild 1), BN, and the other strains is 1 in 141 bp (Table 2; 843 SNPs/119 Kbp). To compare wild rat inter-individual variation with the inter-strain variation for commonly used inbred strains, we calculated the number of SNPs that are polymorphic when comparing arbitrary combinations of 3 strains. Genotyping of 861 SNP positions in the three wild rats resulted in 438 polymorphic positions, whereas the most polymorphic combination of inbred strains in this experiment (BN, BH, and SHR) yielded 427 SNPs. This indicates that three random, but potentially related, Dutch wild rats are about equally polymorphic as three carefully selected inbred strains. Inclusion of wild isolates from other locations worldwide may increase the efficiency of the SNP discovery approach.
Intraspecific phylogenetic network
Relationships among different rat strains have been determined previously by phylogenetic tree reconstruction based on microsatellite markers [16,17]. However, intraspecific relationships for laboratory strains are often very challenging to determine, due to small genetic distances and complex gene flow. The resulting multitude of plausible trees is best expressed by a network, which displays alternative potential evolutionary paths in the form of cycles [18]. We used Network software (v4.111 Reduced-Joining, [19]) to construct a spatial network, based on 861 SNP markers in 36 rat strains and three wild rat individuals (Fig. 2). The three wild individuals are grouped together, possibly due to the geographic and possibly genetic relation between the samples, but in accordance with the last paragraph of the previous section, they appear relatively unrelated as compared to the set of inbred strains.
Figure 2 Strain relationships in a network structure. End nodes (yellow dots) represent strains. Some end nodes are double-size, meaning that they are supported by two samples. Interconnecting nodes where lines come together, represent a possible precursor.
The majority of the SNPs (485 of 861) was selected for being polymorphic between wild 1 and BN. As a result, different BN substrains (BN/Ztm, BN/Crl), depicted as a double-sized end node because of high similarity, and different wild rat individuals (wild 1, wild 2, and wild 3) are grouped together as the outliers. Several strains that are known to be closely related (source RGD-strains: [20]) are also grouped together, like DA and COP or SS and SR. Interestingly, WKY is also an outlier, indicating that besides BN, this strain can be utilized as an alternative mapping strain. WKY is already commonly used as a normotensive control strain in genetic mapping of blood pressure quantitative trait loci [21]. WKY is known to be closely related to SHR and these strains are indeed grouped together (Fig. 2). Additionally, BDII and BDIX are related and BDE is an RI strain from E3. These strain combinations are also grouped together. Wistar is contributing to a large subset of these strains, like WKY, WC, BDII, MWF, LEW, and WF, which contributes to the complexity of the network structure.
Data availability
The use of genetic markers for mapping traits in rat strains has been exploited for long time already. Current marker sets in rats are mostly limited to microsatellites [22,23], which are not abundantly available and are commonly detected in a more laborious way than SNPs. In this study, we have determined a total of about 35,000 genotypes (about 960 loci in 36 inbred strains), out of which the vast majority are SNPs. This data is accessible via a versatile webtool [24]. Pairs of strains of interest can be selected and explored on presence of verified genetic variation. Besides a graphical representation of the location of the SNPs on a genome map, primer sequences that were successfully used in our experiments are also provided. In a pairwise comparison matrix (Table 3), we plotted the absolute number of polymorphic positions for each of the (sub-)strains or individuals used. Interestingly, for some strains different alleles are observed in substrains (e.g. BN/Crl differs from BN/Ztm at 4 positions), in line with previous observations [8].
Table 3 Absolute number of polymorphic positions between strains in a pairwise comparison.
ACI AO AUG BDE BDII BDIX BDV BH BN BN2 BS BUF COP DA DA2 E3 F344 LEW LE LOU LUDW MWF MNS NAR OM PAR PVG R33 RP SD SD2 SHR SR SS WAG WC WF WIST WKY wild3 wild2 wild1
ACI x
AO 131 x
AUG 111 167 x
BDE 145 178 158 x
BDII 130 166 143 148 x
BDIX 129 196 191 159 115 x
BDV 109 131 145 137 78 142 x
BH 145 189 192 212 156 177 167 x
BN 225 263 279 266 206 258 244 274 x
BN2 227 270 285 268 222 271 251 274 4 x
BS 159 167 163 178 166 185 143 180 246 251 x
BUF 148 158 181 171 168 190 151 166 250 263 158 x
COP 84 194 187 166 155 147 155 191 262 275 203 192 x
DA 62 107 116 134 115 130 117 130 210 216 150 126 116 x
DA2 76 141 144 157 149 167 127 176 261 269 193 167 151 2 x
E3 136 189 170 86 154 180 144 204 244 253 172 182 171 148 180 x
F344 132 165 178 178 152 150 138 144 170 176 141 156 166 121 165 168 x
LEW 156 178 213 197 160 163 141 169 213 222 166 160 191 133 181 186 16 x
LE 131 144 136 155 148 152 122 142 221 224 157 127 147 110 140 167 142 149 x
LOU 145 146 165 191 153 203 120 178 242 250 138 171 212 111 142 192 126 150 149 x
LUDW 153 175 186 198 161 169 153 195 252 263 183 177 189 123 163 215 115 125 133 161 x
MWF 148 147 172 166 136 166 111 167 209 222 135 148 185 133 164 158 115 134 148 136 163 x
MNS 151 167 178 173 158 186 122 169 239 250 155 194 210 128 176 166 123 132 156 141 159 137 x
NAR 147 169 184 212 168 177 145 177 233 249 146 155 197 134 193 188 137 166 122 170 153 164 151 x
OM 127 161 153 170 120 158 143 150 216 222 154 138 176 125 147 183 125 156 127 156 139 143 144 147 x
PAR 140 182 168 166 149 158 133 175 225 227 155 138 167 136 159 169 128 136 127 151 159 133 160 150 162 x
PVG 95 164 152 153 153 181 129 184 252 263 170 151 164 110 150 148 142 175 145 143 196 146 161 172 160 147 x
R33 155 198 183 213 169 198 186 177 261 257 173 210 204 142 184 223 183 209 164 185 196 191 187 182 159 173 189 x
RP 146 159 171 171 132 164 109 186 216 230 113 161 176 138 175 153 119 141 134 141 166 108 157 136 147 132 139 166 x
SD 121 154 156 177 153 160 129 118 233 247 149 134 174 100 148 184 131 138 135 125 145 144 141 149 103 133 130 147 138 x
SD2 95 116 134 150 117 130 122 85 219 220 117 96 145 104 115 168 109 115 92 99 110 118 121 107 90 121 109 126 109 16 x
SHR 159 212 166 186 179 175 168 204 264 275 188 178 180 135 189 205 156 176 173 202 188 180 200 184 182 177 192 207 187 185 139 x
SR 129 171 172 170 160 161 147 138 235 244 163 131 170 120 172 174 146 169 156 157 175 150 160 164 136 117 134 166 160 60 64 184 x
SS 114 145 167 183 142 150 139 111 249 253 149 152 175 117 161 186 136 161 121 134 156 146 153 128 130 143 145 145 144 69 46 191 83 x
WAG 120 105 156 147 143 158 115 160 197 200 96 125 164 94 136 151 110 132 127 96 145 120 121 129 140 126 128 154 108 118 84 181 119 110 x
WC 140 164 158 155 129 180 92 188 214 232 151 177 177 125 158 157 131 160 156 145 171 87 157 179 152 158 138 213 126 153 122 195 170 157 126 x
WF 155 183 195 211 162 164 152 183 266 274 176 173 193 129 175 224 120 123 133 156 50 180 158 148 141 155 196 189 157 141 97 201 179 135 149 186 x
WIST 101 116 116 113 98 101 112 93 160 162 107 91 119 83 114 133 82 86 82 110 95 89 94 85 105 101 96 115 100 83 51 112 71 67 76 114 91 x
WKY 169 208 198 213 164 188 168 208 264 276 208 181 205 128 170 228 180 210 165 229 206 200 220 195 189 197 208 205 194 196 148 115 196 189 183 216 203 101 x
wild3 137 181 153 177 162 164 134 149 175 194 156 173 153 111 172 169 162 182 156 184 173 171 160 149 134 140 161 154 165 161 120 152 160 152 140 158 178 108 163 x
wild2 197 233 207 223 182 213 194 187 256 268 203 221 211 181 227 218 204 243 180 234 220 213 210 203 177 198 219 202 214 210 163 190 208 205 208 213 210 141 210 52 x
wild1 334 414 368 406 329 405 339 372 520 551 369 400 395 315 392 397 386 446 338 415 410 413 391 403 331 352 377 371 387 373 321 334 375 388 372 404 410 280 357 134 157 x
The matrix is built from genotyping data of 960 polymorphisms in 36 strains and three wild individuals. Two inbred strains are represented by two substrains (BN and DA) and outbred SD is represented by two individuals from different stocks. Sets of polymorphisms, including a graphical representation, can be retrieved from [24].
Simulation experiment wild rat-based SNP discovery
To get insight in the benefits of using wild rats in SNP discovery studies, we simulated larger scale experiments based on the results obtained in the experiments described above. Shotgun sequencing of 814 Kbp resulted in the identification of 485 SNPs. For 408 of those, the wild rat allele was also represented in laboratory rat strains and hence of interest for research purposes. The maximum amount of SNPs that can be discovered by fully sequencing this single rat is calculated by multiplying the SNP frequency (408/814,440) with the rat genome size (2,48 Gbp), which is 1,252,911 SNPs. Since none of our shotgun reads were overlapping, we can calculate the relation between shotgun sequencing reads of the wild rat and the amount of SNPs that will be found by scaling up this methodology, assuming random distribution of 400 bp shotgun reads over the genome (Fig. 3a). One million shotgun reads of a single wild rat would already result in the discovery of 200,000 novel SNPs that are polymorphic in commonly used rat strains. This simulation indicates that a relatively small sequencing effort could potentially result in a vast expansion of the amount of genetic variation for the rat.
Figure 3 a) Simulation of wild rat-based SNP discovery experiment. Simulation is based on the discovery of 485 SNPs between wild 1 and BN in 814 Kbp of shotgun sequence. For 408 of those, the wild rat alleles is found back in one or more inbred strains. The relation between generation of randomly distributed 400 bp shotgun reads and estimated number of newly discovered SNPs is plotted. b) Simulation of SNP discovery experiment, using carefully selected (most polymorphic compared to BN) rat strains (SHR, AUG, and WF) or all rat strains, in comparison with wild rat-based SNP discovery. Simulation is based on 539, 304, 292, 287, and 754 SNPs for wild 1, AUG, SHR, WF, and all strains respectively, in 119 Kbp of genotyped sequence.
Because shotgun sequencing was only done in the wild rat 1, we cannot make a direct comparison between wild rat-based SNP discovery and SNP discovery based on rat strains separately. However, a similar simulation experiment can be performed by treating the genotyping resequencing as shotgun reads. For wild 1, this would result in the identification of 577 SNPs as compared to the BN genome sequence. For 539 of those, the wild rat allele is found back in one of the inbred strains. For the combination of three strains most polymorphic as compared to BN in this experiment, the latter number would be 304, 292, and 287 for AUG, SHR, and WF, respectively. Simulations based on these numbers show that it requires nearly two times as much shotgun sequencing in different inbred strains separately to discover the same amount of SNPs that can be found using the wild rat shotgun sequencing approach. It should be mentioned that parallel shotgun sequencing of all 36 inbred strains until saturation has the potential to yield 1.6 times as many SNPs as compared to the wild-derived approach (Fig. 3b). An advantage of using inbred strains for SNP discovery is that the genotype of the strain is immediately known. Nevertheless, reconfirmation of the SNP or genotyping of other strains of interest may be necessary anyway, minimizing the relevance of this advantage.
Discussion
An increase in the amount of documented genetic variation for the rat will be essential to allow for high-resolution genetic mapping of the many inherited traits that have now been described for a wide variety of rat inbred strains. In addition, insight into genetic variation between rat strains provides valuable information on genetic relationships between strains, which can be instrumental to dissect the genetic basis of phenotypic differences. The wild rat-based shotgun sequencing method described here provides an efficient approach to generate such a dense map of genetic variation. To be able to benefit from haplotype-based mapping approaches [25-28] a high marker density is needed to first reliably define haplotype blocks in strains of interest [29]. For the mouse, it has recently been announced that 15 inbred strains will be fully resequenced to achieve this goal [30]. With extreme dense genotype maps, it may even become possible to clone traits by haplotype-based in silico mapping [25], but to achieve this, it is estimated that complete sequences of over 50 strains are needed [29]. Although densities needed for these approaches are not reached, we do show here that wild rat-based SNP discovery is potentially much more effective than shotgun sequencing different inbred strains. We propose that the most effective SNP discovery strategy for the rat would be one based on shotgun sequencing of a single wild-derived sample and subsequent low-cost high-throughput genotyping of the resulting candidates in the laboratory strains of interest. Many other model organisms are currently undergoing full coverage sequencing and SNP discovery in these organisms will become increasingly important, especially for those organisms that are selectively bred for specific traits, such as cow and pig. Pilot experiments using for example wild-derived swine samples could be performed to test whether it is eligible to efficiently transfer the wild isolate-based SNP discovery strategy to other organisms.
Our results do provide insight in the genetic descent of the laboratory rat. It is generally accepted that current rat strains underwent two major genetic bottlenecks. First, they originate from a small founder population of domesticated wild rats and second, they were selectively inbred to obtain homogeneity [11]. The three Dutch wild rats used in this study are potentially relatively closely related as compared to wild rats from different parts of the world, but the genetic variation between them is mostly larger than or sporadically equal to any combination of three inbred strains, indeed suggesting the existence a common genetic bottleneck for laboratory strains. In addition, the laboratory rat does not show an extensive polymorphism rate in the MHC (major histocompatibilty complex) as compared to other species [31], like human, cattle etc. Cramer et al. has analyzed the MHC of wild rats and compared the data with those from inbred strains [32]. In line with our observation, there were not many new haplotypes.
We observed that wild rat genetic variation is to a large extent represented in the inbred strains, which is in sharp contrast to genetic variation in wild-derived mouse strains that is mostly unique [33]. Contrary to classical mouse inbred strains, where multiple subspecies contribute to the genetic make-up [13,34] and recent mouse strains, derived from different Mus species [35], laboratory rat strains are most likely descending from a single rat species, Rattus norvegicus [10].
An independent study using 42 microsatellites in German and Japanese wild-derived samples showed that the genetic profiles were quite divergent, partially owing to different geographic locations [36]. Our study involved only Dutch wild rats, suggesting that the inclusion of wild rats from different parts of the world could result in even more efficient SNP discovery, although it also remains to be demonstrated what proportion of the additional discovered alleles is present in the inbred strains and if a geographic bias for this exists.
When multiple SNPs are present per locus/amplicon, independent haplotypes can be discerned. The genetic variation identified here is mostly organized in a limited amount of haplotypes per locus (Table 4). Theoretically, an amplicon containing two or three SNPs can be represented by four and eight haplotypes, respectively, but in our dataset the vast majority of amplicons harboring multiple SNPs is represented by only two or three haplotypes (Table 4). Again, these observations suggest the existence of a common and small founding population with very limited haplotype diversity and/or a very narrow genetic bottleneck before inbred strain selection. The observed small genetic basis in a wide selection of laboratory rat strains does not mimic genetic variation in the human population and as a result, studies and pharmacological tests in rat models neglect potential modulatory effects caused by genetic variation. Although the use of F1 crosses and mosaic populations [37] could address this issue, our data suggests that wild-derived rats may be very useful to this end, since a large amount of all genetic variation present in a large selection of inbred strains, is already represented in a limited number of individuals. Therefore, it would be very interesting to investigate genetic variation in recently domesticated inbred [38] and outbred rats such as wild-type Groningen rats (WTG) [39]. Alternatively, careful selection of inbred strains based on genotyping data and subsequent random breeding may also expose the wild side of laboratory rats.
Table 4 Haplotype analysis in 36 strains for all SNP-containing amplicons
number of haplotypes
2 3 4 5 6 7 8 9 10 11 12
2 SNPs 46 57 8
3 SNPs 11 26 8 3 0 0 0
4 SNPs 4 11 5 3 1 1 0 0 0 0 0
5 SNPs 1 3 3 2 0 0 0 0 0 0 0
6 SNPs 1 1 1 1 1 0 0 0 0 0 0
7 SNPs 1 0 1 0 2 0 0 0 0 0 0
8 SNPs 0 0 0 0 0 0 0 0 0 0 0
9 SNPs 0 0 0 0 0 0 0 0 0 0 1
10 SNPs 0 0 0 0 0 0 0 0 0 0 0
11 SNPs 0 0 0 0 1 0 0 0 0 0 0
total 64 97 27 9 5 1 0 0 0 0 1 204*
*) Total number of amplicons that contains at least two SNPs. Amplicons containing no SNPs or only indels were excluded from this analysis. Amplicons containing 1 SNP are also excluded, since two-state SNPs always give rise to two haplotypes.
Conclusion
We describe a SNP discovery platform for the rat that is based on two steps. First, candidate SNPs are discovered by shotgun sequencing a wild rat, followed by genotyping laboratory strains of interest. We show that 84% of alleles in wild rats as compared to the sequenced Brown Norway rat genome are also represented in a set of 36 laboratory strains. Hence, the approach described here would be an efficient strategy for the discovery of novel informative SNPs in the laboratory rat. Inclusion of other wild samples, preferably from different locations in the world could result in an even more effective SNP discovery platform, as the three wild rats in our study, caught in relative close vicinity to each other, were already more polymorphic than the most polymorphic combination of carefully selected inbred strains. Based on the more than 34,000 genotyping datapoints obtained in this study, we postulate two things. First, laboratory rats originate from a single rat species, and inbred stains are relatively closely related with a limited number of haplotypes, reflecting known genetic bottlenecks in strain establishment. Second, wild rats have the potential to represent the degrees of genetic variation as present in the human population much more efficiently than a random selection of inbred strains. This makes them or wild-derived strains potentially well-suited for studying modulatory effects of genetic background variation on specific phenotypes, such as behavior or responses to drug treatment.
Methods
Genomic DNA isolation, shotgun library construction
Wild rat 1 (Rattus norvegicus) was caught in the canals of Utrecht and kindly provided by the Pest Control Service of the City of Utrecht (Utrecht, The Netherlands). Wild rat 2 was trapped in Gassel, a village located approximately 100 km south-east of Utrecht and was kindly provided by Tien Derks (Gassel, The Netherlands). Wild rat 3 was caught in a basement in Amsterdam, located 50 km north of Utrecht and kindly provided Romke Koch (Amsterdam, The Netherlands). Rat strains BN/Crl and Crl:Wistar (outbred) were obtained from Charles River The Netherlands. Liver samples of commonly used rat strains ACI/Ztm, BDE/Ztm, BDII/Ztm, BDIX/Ztm, BDV/Ztm, BH/Ztm, BN/Ztm, BS/Ztm, DA/Ztm, E3/Ztm, F344/Ztm, LE/Ztm, LEW/Ztm, LOU/CZtm, MNS/Ztm, MWF/Ztm, NAR/Ztm, OM/Ztm, PAR/Ztm, R33/Ztm, WC/Ztm, WF/Ztm, WKY/Ztm were provided by D.W. (Hannover Medical School, Germany) and liver samples of strains AO/OlaHsd, AUG/OlaHsd, BUF/SimRijHsd, COP/Hsd, DA/OlaHsd, LUDW/OlaHsd, PVG/OlaHsd, RP/AEurRijHsd, SHR/NHsd, SR/JrHsd, SS/JrHsd, WAG/RijHsd and 2 individuals of Hsd:SD (outbred) were kindly provided by Harlan (Horst, The Netherlands). Samples were lysed overnight in 20 ml lysis buffer, containing 100 mM Tris (pH 8.5), 200 mM of NaCl, 0.2% of SDS, 5 mM of EDTA, and 100 μg/ml of freshly added Proteinase K at 55°C under continuous rotation. Tissue debris was spinned down for 20 min at 10,000 × g and supernatant was transferred to a fresh tube. DNA was purified by phenol-chloroform extraction and precipitated by adding an equal volume of isopropanol, mixing and centrifugation for 20 min, 10,000 × g at 4°C. The supernatant was removed by gently inverting the tube and the pellets were washed with 70% ethanol and dissolved in 1000 μl water. The concentration was measured by Optical Densitometry at 260 nm.
Wild rat-derived genomic library construction and shotgun sequencing
Sheared wild rat-derived genomic DNA of approximately 1–2 Kbp in size was cloned into the SmaI-site of pUC19. Fractions of the glycerol stock of the transformed library (E. coli DH10B) were plated on LB-plates containing 50 μg/ml ampicilin, 200 μg/ml IPTG, and 0.01% X-gal for standard blue/white screening on inserts. White colonies were picked in 20 μl water. Lysis occurred at 95°C for 10 min. 5 μl of 5× diluted lysate was used for the PCR reaction. For PCR, universal M13 primers were used, namely M13F: TGTAAAACGACGGCCAGT, M13R: AGGAAACAGCTATGACCAT. PCR, sequencing and cycling conditions were similar as for strain genotyping, described below. Sequencing was performed using universal M13 primers.
PCR conditions for strain genotyping
PCR was carried out using a touchdown thermocycling program (92°C for 60 sec; 12 cycles of 92°C for 20 sec, 65°C for 20 sec with a decrement of 0.6°C per cycle, 72°C for 30 sec; followed by 20 cycles of 92°C for 20 sec, 58°C for 20 sec and 72°C for 30 sec; 72°C for 180 sec; GeneAmp9700, Applied Biosystems) and contained 30–50 ng genomic DNA, 0.2 μM of each forward primer and 0.2 μM of each reverse primer, 400 μM of each dNTP, 25 mM Tricine, 7.0% Glycerol (w/v), 1.6% DMSO (w/v), 2 mM MgCl2, 85 mM Ammonium acetate pH 8.7 and 0.2 U Taq Polymerase in a total volume of 10 μl.
Sequencing reactions, purification, and analysis
PCR products were diluted with 25 μl water and 1 μl was directly used as template for the sequencing reactions. Sequencing reactions, containing 0.25 μl BigDYE (v3.1; Applied Biosystems, Foster City, CA, USA), 3.75 μl 2.5× dilution buffer (Applied Biosystems) and 0.4 μM universal M13 primer in a total volume of 10 μl, were performed using cycling conditions recommended by the manufacturer (40 cycles of 92°C for 10 sec, 50°C for 5 sec and 60°C for 120 sec). Of sequencing products, 5 μl was purified by ethanol precipitation in the presence of 40 mM sodium-acetate and analyzed on 96-capillary 3730XL DNA analyzers (Applied Biosystems), using the standard RapidSeq protocol. Sequences were analyzed for presence of heterozygous mutations using PolyPhred [40], followed by manual inspection of the polymorphic positions.
Automation
All PCR and sequencing reactions were set up on a Tecan Genesis RSP200 liquid handling workstation, with a robotic and an 8-channel pipetting arm, an integrated 96-channel pipetting head (TEMO96, Tecan), and four integrated dual-384 well PCR blocks (Applied Biosystems).
Mapping of shotgun reads and SNP discovery
Shotgun reads were assigned to positions in the RGSC 3.1 rat genome assembly using blat search [15]. Shotgun reads that complied with our mapping criteria, namely those having at least 80 identical bp for the best hit and no more than 60 identical bp for second blat hit were retained for further analysis. Blast nucleotide sequence alignments between shotgun read and corresponding genomic segment were used for discovery of single base variations (including single base indels). A site was treated as polymorphic only in the case when it has identical 5'- and 3'-flanks of at least 5 bp. A custom designed web-application was employed for manual chromatogram inspection and confirmation of a correct shotgun base-call for every polymorphic SNP locus. Primer design for resequencing was performed using a local web-interface [41] to the PRIMER3 program [42].
Simulation model for wild rat-based SNP discovery
To estimate the number of SNPs to be discovered by the wild rat resequencing approach we performed computer simulations using the observed sample-specific polymorphism frequencies and the rat genome size of 2.48 Gbp as an input. We used a Monte-Carlo method for the placement of N 400-bp shotgun reads to the genome and calculated the total size of genome covered by N shotgun reads. To obtain a conservative estimate by assuming low heterozygosity in wild-derived strain the estimate of number of SNPs is given by product of covered genome size and polymorphism rate.
Authors' contributions
BMGS contributed to the production of the results, supervised the ongoing of the study, and drafted the manuscript. VG contributed to the computational support of the results, and contributed to the writing of the manuscript. DZ contributed to the production of sequencing reads and initial analysis of the results. DW contributed to the preparation of samples for the study and revised the manuscript. HJH participated in the interpretation of the results and revision of the manuscript. EC outlined and supervised the study, and revised the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Genotyping details; Detailed genotyping information, including allele information for the inbred and wild rat strains
Click here for file
Acknowledgements
We thank Harlan (Horst – Netherlands), the Pest Control Service of the City of Utrecht (Utrecht, The Netherlands), Romke Koch (Amsterdam, The Netherlands), and Tien Derks (Gassel, The Netherlands) for kindly providing rat tissue samples. This work was supported by the Dutch Ministry of Economic Affairs through the Innovation Oriented Research Program on Genomics.
==== Refs
Kristensen VN Kelefiotis D Kristensen T Borresen-Dale AL High-throughput methods for detection of genetic variation Biotechniques 2001 30 318 322 11233601
Sellick GS Longman C Tolmie J Newbury-Ecob R Geenhalgh L Hughes S Whiteford M Garrett C Houlston RS Genomewide linkage searches for Mendelian disease loci can be efficiently conducted using high-density SNP genotyping arrays Nucleic Acids Res 2004 32 e164 15561999 10.1093/nar/gnh163
Gibbs RA Weinstock GM Metzker ML Muzny DM Sodergren EJ Scherer S Scott G Steffen D Worley KC Burch PE al. Genome sequence of the Brown Norway rat yields insights into mammalian evolution Nature 2004 428 493 521 15057822 10.1038/nature02426
Greenhouse DD Festing MFW Hasan S Cohen AL Hedrich HJ Catalogue of Inbred Strains of Rats Genetic Monitoring of Inbred Strains of Rats 1990 Stuttgard , Gustav Fisher 410 480
Hedrich HJ Taxonomy and stocks and strains The Laboratory Rat 2005 Academic Press 71 91, in press
NCBI dbSNP
Zimdahl H Nyakatura G Brandt P Schulz H Hummel O Fartmann B Brett D Droege M Monti J Lee YA Sun Y Zhao S Winter EE Ponting CP Chen Y Kasprzyk A Birney E Ganten D Hubner N A SNP map of the rat genome generated from cDNA sequences Science 2004 303 807 14764869 10.1126/science.1092427
Smits BM van Zutphen BF Plasterk RH Cuppen E Genetic variation in coding regions between and within commonly used inbred rat strains Genome Res 2004 14 1285 1290 15231744 10.1101/gr.2155004
Guryev V Berezikov E Malik R Plasterk RH Cuppen E Single nucleotide polymorphisms associated with rat expressed sequences Genome Res 2004 14 1438 1443 15231757 10.1101/gr.2154304
Lindsey JR Baker HJ, Lindsey JR, Weisbroth SH Historical Foundations The Laboratory Rat 1979 1 New York , Academic Press 1 36
Hedrich HJ Krinke GJ History, Strains and Models The laboratory Rat 2000 London , Academic Press
Koch LG Britton SL Whishaw IQ, Kolb B Strains The Behavior of the Laboratory Rat 2005 Oxford, UK , Oxford University Press
Ideraabdullah FY de la Casa-Esperon E Bell TA Detwiler DA Magnuson T Sapienza C de Villena FP Genetic and haplotype diversity among wild-derived mouse inbred strains Genome Res 2004 14 1880 1887 15466288 10.1101/gr.2519704
Yalcin B Fullerton J Miller S Keays DA Brady S Bhomra A Jefferson A Volpi E Copley RR Flint J Mott R Unexpected complexity in the haplotypes of commonly used inbred strains of laboratory mice Proc Natl Acad Sci U S A 2004 101 9734 9739 15210992 10.1073/pnas.0401189101
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
Canzian F Phylogenetics of the laboratory rat Rattus norvegicus Genome Res 1997 7 262 267 9074929
Thomas MA Chen CF Jensen-Seaman MI Tonellato PJ Twigger SN Phylogenetics of rat inbred strains Mamm Genome 2003 14 61 64 12532268 10.1007/s00335-002-2204-5
Bandelt HJ Forster P Rohl A Median-joining networks for inferring intraspecific phylogenies Mol Biol Evol 1999 16 37 48 10331250
Homepage Fluxus Engineering
Rat Genome Database (RGD) strain information
Rapp JP Genetic analysis of inherited hypertension in the rat Physiol Rev 2000 80 135 172 10617767
Kwitek AE Gullings-Handley J Yu J Carlos DC Orlebeke K Nie J Eckert J Lemke A Andrae JW Bromberg S al. High-density rat radiation hybrid maps containing over 24,000 SSLPs, genes, and ESTs provide a direct link to the rat genome sequence Genome Res 2004 14 750 757 15060019 10.1101/gr.1968704
Steen RG Kwitek-Black AE Glenn C Gullings-Handley J Van Etten W Atkinson OS Appel D Twigger S Muir M Mull T al. A high-density integrated genetic linkage and radiation hybrid map of the laboratory rat Genome Res 1999 9 AP1 8 10400928
CASCAD SNPview
Grupe A Germer S Usuka J Aud D Belknap JK Klein RF Ahluwalia MK Higuchi R Peltz G In silico mapping of complex disease-related traits in mice Science 2001 292 1915 1918 11397946 10.1126/science.1058889
Liao G Wang J Guo J Allard J Cheng J Ng A Shafer S Puech A McPherson JD Foernzler D Peltz G Usuka J In silico genetics: identification of a functional element regulating H2-Ealpha gene expression Science 2004 306 690 695 15499019 10.1126/science.1100636
Pletcher MT McClurg P Batalov S Su AI Barnes SW Lagler E Korstanje R Wang X Nusskern D Bogue MA Mural RJ Paigen B Wiltshire T Use of a dense single nucleotide polymorphism map for in silico mapping in the mouse PLoS Biol 2004 2 e393 15534693 10.1371/journal.pbio.0020393
Wang X Korstanje R Higgins D Paigen B Haplotype analysis in multiple crosses to identify a QTL gene Genome Res 2004 14 1767 1772 15310659 10.1101/gr.2668204
Flint J Valdar W Shifman S Mott R Strategies for mapping and cloning quantitative trait genes in rodents Nat Rev Genet 2005 6 271 286 15803197 10.1038/nrg1576
Pearson H Mouse sequencing plan aims to boost models Nature 2004 432 5 15525947 10.1038/432005a
Gunther E Walter L The major histocompatibility complex of the rat (Rattus norvegicus) Immunogenetics 2001 53 520 542 11685465 10.1007/s002510100361
Cramer DV Chakravarti A Arenas O Humprieres J Mowery PA Genetic diversity within and between natural populations of Rattus norvegicus J Hered 1988 79 319 324 3183358
Campino S Behrschmidt C Bagot S Guenet JL Cazenave PA Holmberg D Penha-Goncalves C Unique genetic variation revealed by a microsatellite polymorphism survey in ten wild-derived inbred strains Genomics 2002 79 618 620 11991709 10.1006/geno.2002.6570
Abe K Noguchi H Tagawa K Yuzuriha M Toyoda A Kojima T Ezawa K Saitou N Hattori M Sakaki Y Moriwaki K Shiroishi T Contribution of Asian mouse subspecies Mus musculus molossinus to genomic constitution of strain C57BL/6J, as defined by BAC-end sequence-SNP analysis Genome Res 2004 14 2439 2447 15574823 10.1101/gr.2899304
Guenet JL Bonhomme F Wild mice: an ever-increasing contribution to a popular mammalian model Trends Genet 2003 19 24 31 12493245 10.1016/S0168-9525(02)00007-0
Voigt B Kitada K Kloting I Serikawa T Genetic comparison between laboratory rats and Japanese and German wild rats Mamm Genome 2000 11 789 790 10967140 10.1007/s003350010137
Cholnoky E Fischer J Varga M Gyorffy G Aspects of genetically defined populations in toxicity testing. II. Genotypic differences in sensitivity to a toxic dextran preparation Z Versuchstierkd 1974 16 43 48 4824483
Ohno K Niwa Y Kato S Kondo K Oda S Inouye M Yamamura H Establishment of new inbred strains derived from Japanese wild rats (Rattus norvegicus) Jikken Dobutsu 1994 43 251 255 8174628
de Boer SF Lesourd M Mocaer E Koolhaas JM Selective antiaggressive effects of alnespirone in resident-intruder test are mediated via 5-hydroxytryptamine1A receptors: A comparative pharmacological study with 8-hydroxy-2-dipropylaminotetralin, ipsapirone, buspirone, eltoprazine, and WAY-100635 J Pharmacol Exp Ther 1999 288 1125 1133 10027850
Nickerson DA Tobe VO Taylor SL PolyPhred: automating the detection and genotyping of single nucleotide substitutions using fluorescence-based resequencing Nucleic Acids Res 1997 25 2745 2751 9207020 10.1093/nar/25.14.2745
Local PRIMER3 interface
Rozen S Skaletsky H Primer3 on the WWW for general users and for biologist programmers Methods Mol Biol 2000 132 365 386 10547847
|
16316463
|
PMC1318490
|
CC BY
|
2021-01-04 16:32:47
|
no
|
BMC Genomics. 2005 Nov 29; 6:170
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-170
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2781630575010.1186/1471-2105-6-278Research ArticleMultiple sequence alignment accuracy and evolutionary distance estimation Rosenberg Michael S [email protected] Center for Evolutionary Functional Genomics, The Biodesign Institute, and the School of Life Sciences, Arizona State University, Tempe, AZ 85287-4501, USA2005 23 11 2005 6 278 278 14 7 2005 23 11 2005 Copyright © 2005 Rosenberg; licensee BioMed Central Ltd.2005Rosenberg; 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
Sequence alignment is a common tool in bioinformatics and comparative genomics. It is generally assumed that multiple sequence alignment yields better results than pair wise sequence alignment, but this assumption has rarely been tested, and never with the control provided by simulation analysis. This study used sequence simulation to examine the gain in accuracy of adding a third sequence to a pair wise alignment, particularly concentrating on how the phylogenetic position of the additional sequence relative to the first pair changes the accuracy of the initial pair's alignment as well as their estimated evolutionary distance.
Results
The maximal gain in alignment accuracy was found not when the third sequence is directly intermediate between the initial two sequences, but rather when it perfectly subdivides the branch leading from the root of the tree to one of the original sequences (making it half as close to one sequence as the other). Evolutionary distance estimation in the multiple alignment framework, however, is largely unrelated to alignment accuracy and rather is dependent on the position of the third sequence; the closer the branch leading to the third sequence is to the root of the tree, the larger the estimated distance between the first two sequences.
Conclusion
The bias in distance estimation appears to be a direct result of the standard greedy progressive algorithm used by many multiple alignment methods. These results have implications for choosing new taxa and genomes to sequence when resources are limited.
==== Body
Background
DNA sequence alignment is a common step in molecular evolutionary analysis. Aligned sequences are used for many purposes, including estimation of patterns of divergence, selection, the tempo and mode of evolutionary change, identification of functional elements and constraints, and phylogenetic history, just to name a few. Alignments are a hypothesis of site homology; as evolutionary distance among sequences increases, alignments are known to become less accurate [1-7]. The effect of alignment accuracy on downstream analysis in comparative genomics and bioinformatics is largely an unexplored topic, although some empirical studies have attempted to examine this with respect to functional element identification [8,9] and phylogenetic analysis [10-16].
Multiple sequence alignment, the alignment of more than two sequences, is generally thought to lead to more accurate alignments than simple pair wise alignments [4]. There are numerous approaches to multiple alignment, although most are based in some way on a progressive alignment algorithm [17,18] where similar sequences are aligned first and additional sequences are progressively added based on their divergence from the initial pair. While empirical studies have demonstrated how multiple alignments perform better than pair wise alignments [19,20], simulation methodologies have not been employed to characterize the improvement. In the simplest case, one can ask how much a pair wise alignment is improved by the addition of a third sequence (of intermediate phylogenetic position relative to the initial pair). How does varying the position of the third sequence, relative to the first two, affect the alignment? Logically, one might hypothesize that the greatest improvement will come when the third sequence is exactly equidistant from the initial pair, splitting the branch separating them in half (Figure 1); from a rooted perspective, this would be a polytomy. On the other hand, one might expect that the greatest improvement would be found from a third sequence which evenly splits a branch on the rooted tree; in an unrooted perspective this would mean the third sequence is half as close to one of the initial sequences as to the other (Figure 1).
Figure 1 Possible optimal locations of intermediate sequences. Cartoon phylogenies indicating possible hypothesized optimal locations for the addition of an intermediate sequence (C) to improve the alignment of a pair of target sequences (A and B). The left column contains unrooted trees, the right column rooted trees. In the top row C is equidistant from A and B. In the bottom row, C is equidistant from A and the root of the tree.
Beyond simple accuracy, multiple sequence alignment may affect downstream sequence analysis in unexpected ways relative to pair wise sequence alignment. In a previous study [6], I show that evolutionary distance estimation from DNA sequences can be surprisingly robust to alignment error (evolutionary distance is the number of substitutions per site which have occurred since a pair of sequences diverged from a common ancestral sequence). This previous work was based on alignments of paired sequences; the relationship between accuracy of alignment and distance estimation might differ under multiple alignment conditions.
The primary goal of this study was to use simulation to examine the improvement in alignment accuracy when going from pair wise to multiple alignments, profiling the change versus the position of the additional sequences. How much is the accuracy of alignment of a pair of sequences improved by the addition of a third sequence with an intermediate evolutionary history (relative to the initial pair)? Where should the third sequence be in order to maximize the accuracy of the initial pair's alignment? In addition, the effects of these multiple sequence alignment on evolutionary distance estimation were also profiled. Does the position of the third sequence have an effect on the estimation of evolutionary distance of the initial pair, independent of the accuracy of the alignments?
Results and discussion
Accuracy of multiple alignments
Figure 2 shows the difference in accuracy of alignment of sequences A and B between the three-sequence multiple alignments (ABC alignment) and the pair wise alignments (AB alignment) for Clustal versus the relative position of sequence C. The pattern was consistent for the two shortest divergences (Figure 2A–B): the multiple alignment was maximally more accurate than the pair wise alignment when sequence C was half as divergent from one of the target sequences than they were from each other (bottom row of Figure 1). In the final case (Figure 2C), sequences A and B were divergent to the point of being indistinguishable from random data [6]. Addition of the third sequence had a small improvement when it was relatively close to one of the original sequences, but as it moved deeper into the phylogeny it in-and-of itself became too divergent from sequence A to add any benefit to the alignment and actually marginally decreased the accuracy.
Figure 2 Improvement in alignment accuracy in multiple versus pair wise alignment. Absolute improvement in accuracy of alignment of sequences A and B in the multiple (ABC) alignment versus the pair wise (AB) alignment. Improvement was measured as ABC – AB, where ABC indicates the accuracy of the alignment of sequences A and B in the multiple alignment and AB indicates the accuracy of the alignment of these sequences in the pair wise alignment. Expected distance of sequences A and B was (A) d = 0.5; (B) d = 1.0; (C) d = 2.0. Lengths of branches x and y are illustrated in Figure 8. All points are averages of 1,000 simulation replicates. Error bars represent 95% confidence limits.
Figure 3 shows the identical data as relative improvement in accuracy, with all three simulation conditions superimposed. The curves for the two moderate divergences (d = 0.5 and d = 1.0) were extremely congruent with both showing a peak reduction of about 15% error in the multiple alignment versus the pair wise alignment. The larger divergence (d = 2.0) is essentially flat since the changes in absolute error seen in Figure 2C are minimal relative to the total amount of error (~92%) in the alignments. With respect to the actual number of sites (rather than proportions), these figures correspond to a maximal increase of 12 correctly aligned sites for d = 0.5, 60 sites for d = 1.0, and 15 sites for d = 2.0.
Figure 3 Relative improvement in alignment accuracy in multiple versus pair wise alignment. Relative improvement in accuracy of alignment of sequences A and B in the multiple (ABC) alignment versus the pair wise (AB) alignment. Relative improvement was measured as (ABC – AB) / AB, where ABC indicates the accuracy of the alignment of sequences A and B in the multiple alignment and AB indicates the accuracy of the alignment of these sequences in the pair wise alignment. d indicates the expected distance between sequences A and B. Lengths of branches x and y are illustrated in Figure 8. All points are averages of 1,000 simulation replicates.
Additional simulations for 4-taxon trees were also performed (results not shown). As would be expected from the above results, the maximal improvement in the alignment of sequences A and B was found when the fourth sequence was added to the center of the branch leading from the root to sequence B. Simulations of 16-taxon model trees were also conducted to investigate how much improvement can be made with the addition of even more sequences. Two extreme 16-taxon trees were used as initial simulation models, a perfectly balanced tree and a perfectly pectinate tree (Figure 4). For both cases, simulations with expected AB distances of both 1.0 and 2.0 were conducted and analyzed under the same conditions as the three-taxon phylogenies (100 replicates). The expected accuracies of the AB sequence alignment from a pair wise alignment for these distances are 62% and 7%, respectively [6]. The maximal accuracies when a third sequence was added (Figure 2) were 68% and 9%. When all 16 sequences were used in a multiple alignment, the observed accuracies of the AB alignment were, respectively, 88% and 38% for the balanced tree and 77% and 19% for the pectinate tree. Unsurprisingly, adding additional sequences had a large effect on alignment accuracy when they were added in a balanced fashion and a smaller effect when they were added in a pectinate fashion. Also as would be expected, varying the internal branch lengths (including both with and without a molecular clock) or simulating across 16-taxon trees with more realistic branching patterns (e.g., non clock-like Yule trees) yielded intermediate accuracies depending on the exact shape and structure of the tree (results not shown).
Figure 4 Model trees for 16 sequence analyses. Perfect balanced tree and perfect pectinate tree. A and B show the phylogenetic positions of these target sequences.
These results add an additional spin to the debate over the importance of taxon sampling in phylogeny reconstruction. Although it is generally thought that increased taxon sampling yields more accurate phylogenies, simulation studies have proven to be equivocal and controversial [21-29]. However, all of these studies (and most of the empirical ones as well) tacitly assume perfect sequence alignment. Even if increased taxon sampling has no direct effect on phylogenetic accuracy (a debatable point; see cited literature above), it certain appears to have an effect on the accuracy of alignment prior to the phylogenetic reconstruction process. How alignment accuracy may directly affect phylogeny reconstruction is a topic in need of further study (although see [30]).
These results also imply a specific strategy for choosing species to sequence when resources are limited. For example, given current estimates of the mammalian phylogeny [31-33], these results suggest that error in human and mouse sequence alignments could be reduced by about 15% if aligned with a third species with an evolutionary position similar to that of the Capuchin (Cebus albifrons) or the blind mole rat (Spalax judaei). In contrast, inclusion of sequences from the completed rat genome [34], would only be expected to decrease error in human and mouse alignments by about 7%.
Multiple alignment and evolutionary distance estimation
One would expect that estimation of evolutionary distance in a multiple alignment setting would follow that of alignment accuracy (Figures 2, 3). For alignments in Clustal, this intuition turned out to be surprisingly incorrect. Figure 5 shows the results of evolutionary distance estimation between sequences A and B from the true alignment, the pair wise AB alignment, and the multiple ABC alignments. For the intermediate distances (d = 0.5 and d = 1.0) there was a striking pattern: the estimate of distance between sequences A and B increased linearly under the multiple alignment as the branch leading to sequence C moves closer to the root. This change in distance estimation was uncorrelated with that of alignment accuracy which peaked when sequence C bisected the branch leading to sequence A (Figure 2). The slight dip at the upper end of the curves is explained by the progressive alignment procedure: pair wise alignments and distance estimates are used to set the order of sequence addition in the multiple alignment. When the branch leading to sequence C was close to the root, sequences A and B will occasionally appear to be more closely related to each other than sequence C is to A. When this occurs, A and B are aligned directly (with C on the outside) and produce the pair wise distance estimate which was consistently lower than that found from the multiple alignment (Figure 5). This was proved by realigning a subset of the data with sequence C close to the root but specifying the (correct) guide tree rather than allowing Clustal to estimate its own. The X marked by an arrow in Figure 5B shows the average distance estimate from the ABC alignments when the correct guide tree was provided; not only is the average higher than that of the original alignments but it is directly in line with the projected linear increase seen from the alignments where C branches closer to the tips. The largest simulated distance (d = 2.0) showed the beginning of a similar pattern to the intermediate distances (Figure 5C) before the alignments collapses into random noise (Figure 2C). The phylogenetic position of an intermediate sequence appears to bias evolutionary distance estimation, independent of alignment accuracy.
Figure 5 Evolutionary distance estimates versus position of third sequence. Estimated evolutionary distance of sequences A and B from the true, pair wise and three-taxon multiple alignments. Expected distance of sequences A and B was (A) d = 0.5; (B) d = 1.0; (C) d = 2.0. Lengths of branches x and y are illustrated in Figure 8. All points are averages of 1,000 simulation replicates. The X in panel (B) marked by an arrow indicates the mean distance estimate obtained for the three-taxon multiple alignments when Clustal was forced to use the correct guide tree rather than an estimating its own from the data; see the text for more information.
The accuracy (or lack) of the pair wise distance estimates were exactly what one would expect from previous work [6]. To explain the observed bias in multiple alignment, I partitioned the data that goes into the calculation of evolutionary distance into its base components: observed numbers of transitional and transversional differences. Figure 6 shows the breakdown for d = 0.5 and d = 1.0. Each panel has five parts: the percent of observed sites (between sequences A and B) with the specific change (transition or transversion) in the true alignment; the percent of sites with the specific change in the AB alignment; the percent of sites with the specific change in the AB alignment that were correctly aligned (found in the true alignment); the percent of sites with the specific change found in the ABC alignment; and the percent of sites with the specific change in the ABC alignment that were correctly aligned. The true and pair wise alignment both showed the expected pattern of having no relationship between observed (or correct) percents of change with a change in the phylogenetic position of sequence C. For the multiple alignment, the proportion of sites correctly aligned (both transitions and transversions) followed the identical pattern seen in Figure 2, that is, the peak in alignment accuracy occurs when the branch leading to sequence C bisects the branch leading to sequence A. However, the observed number of transversional changes (and to a lesser extent, transitional changes) increased linearly as the branch leading to sequence C moved closer to the root (Figure 6A,C). The bias in evolutionary distance estimation in multiple sequence alignment appears to primarily be due to an overabundance of hypothesized transversional differences among the sequences.
Figure 6 Observed transitional and transversional sites in the alignments. Percent of observed and correct transitional and transversional sites in the true, pair wise, and multiple sequence alignments. (A) Transversions for d = 0.5; (B) Transitions for d = 0.5; (C) Transversions for d = 1.0; (D) Transitions for d = 1.0. Lengths of branches x and branch y are illustrated in Figure 8. All points are averages of 1,000 simulation replicates.
Careful examination of the strict (greedy) progressive alignment algorithm [17,18] used by Clustal explains this pattern. When sequences A and B are aligned directly in a pair wise alignment, the hypothesized transitions and transversions are based solely on the properties of these sequences. In the multiple alignment, Clustal begins by aligning the closest pair of sequences (A and C). The more distant sequence C is from sequence A, the more transversions will have occurred since they shared a common ancestor. More transversions are therefore identified (hypothesized) between this pair during alignment. When sequence B is added to the alignment, numerous potential transversions have already been "set" between sequences A and C. Thus, any potential transversional difference between sequences A and B have less cost (on average) than would be found in the corresponding pair wise alignment because a transversion may already have been identified between sequences A and C. The biasing effect of phylogenetic position on distance estimation from pair wise alignment appears to be a consequence of the greedy progressive algorithm implemented in Clustal.
This study used Clustal because it is one of the most widely used alignment programs, particularly for high-throughput genomic analysis, and tends to be among the most accurate [7,35]. While it is quite possible that the resulting alignments could be improved by changing alignment parameters (such as the mismatch and gap costs), the purpose of this study is not to optimize the alignment but rather to examine the difference multiple alignment makes under simple conditions and to see examine downstream effects of these errors on distance estimation.
To partially examine whether the observed results are specific to Clustal or may be a more general alignment problem, most of the alignments were repeated with T-Coffee 1.37 [36]. Using the default parameters, T-Coffee produced significantly worse alignments and distance estimates than Clustal (the purpose of this study was not to compare the relative accuracy of these alignment methods and the absolute differences may be due to default parameterization choices and not the overall quality of the methods themselves). However, the shape of the alignment accuracy curves were the same (i.e., the peak gain in accuracy during multiple alignment occurred when the third sequence bisected the branch leading from the root to sequence A). The biasing effect of the position of the third sequence appeared to be less severe, and in some cases, completely absent; unfortunately the differences in accuracy made it difficult to systematically compare these results. Unlike Clustal, T-Coffee uses both global and local pair wise alignments to guide the production of the final global alignment. Although it is also a progressive algorithm, the intermediate alignments it produces make use of more information at early stages, which may help prevent the distance estimate bias.
There are many additional multiple DNA sequence alignment algorithms and programs available, some of which use similar progressive alignment schemes as Clustal and T-Coffee but allow for revision of previously aligned sequences, and some of which use very different approaches, including statistical alignments based on maximum likelihood or Bayesian methods (e.g., [37-41]). Some of these methods simultaneously estimate evolutionary distance and alignment [42,43], while methods for simultaneously estimating phylogenies and alignments are also being developed [30,44-48]. Comparisons of the overall accuracies of some of these programs in pair wise DNA sequence analysis has recently been conducted [7]. How these programs and algorithms compare under multiple alignment conditions and whether the observed biasing effect is widespread or narrow across algorithms is a task for future investigation.
Conclusion
Multiple sequence alignments do improve upon pair wise sequence alignments. The optimal taxon sampling strategy for maximally improving alignments is to bisect long branches in a balanced framework. Independent of alignment accuracy, however, multiple alignment using a progressive algorithm can bias evolutionary distance estimates, with larger estimates consistently found as intermediate sequences appear deeper in the phylogeny.
Methods
In a previous study [6], it was found that the shapes of alignment accuracy profiles (e.g., Figure 7) were largely independent of substitution model complexity, sequence length, and many model parameter choices. This study generally follows the methods from the previous study. All simulations were performed using MySSP [49]. Simulations were conducted using the Hasegawa-Kishino-Yano (HKY) model of nucleotide substitution [50]. Initial sequences consisted of 1000 random nucleotides, with initial and expected nucleotide frequencies of πC = πG = 0.3, πT = πA = 0.2. The transition-transversion bias was set to that observed at neutral sites in mammals, κ = 3.6 [51].
Figure 7 Alignment accuracy versus true distance. Proportion of sites correctly aligned versus true distance for the HKY substitution model. Arrows indicate the chosen divergence of sequences A and B and expected accuracy from a pair wise alignment used for the base simulations in this study. Figure modified from Rosenberg [6].
Initial sequences were allowed to evolve along fixed trees representing different levels of expected divergence. The initial sets of simulation consisted of three-taxon trees (Figure 8), where the expected divergences of the target sequences (A & B) were 0.5, 1.0, or 2.0. Previous work [6] indicates that pair-wise alignments of sequences A & B at these divergences will be about 94%, 62%, and 7% accurate, respectively, under these simulation and alignment conditions (Figure 7). The location of the intermediate sequence (C) was set to a variety of evenly spaced positions ranging from close to sequence A to close to the root.
Figure 8 Model tree structure. Model tree used for the three-taxon simulations (x and y indicate branch lengths). Sequences A and B were set to a fixed divergence (d = 2y). The position of C was varied from close to A (x « y) to close to the root (x ≈ y).
In addition to point substitutions under the HKY model, insertions and deletions were allowed to occur, with the expected rate of deletion events being one occurrence every 40 substitutions and the expected rate of insertion events being one occurrence every 100 substitutions (as observed in primates and rodents) [52]. Realized number of insertions and deletions were drawn from a Poisson distribution with mean equal to the expected value. The lengths of individual insertion and deletion events were also chosen from a truncated (so as not to include zero) Poisson distribution with a mean of 4 bases (as observed from primate and rodent lineages) [52,53]. Variation in insertion/deletion rate and size can have a large affect on alignment accuracy [6]. However, it is likely that changing the values of these parameters in the present study would have similar effects across all conditions. Each simulation condition was replicated 1000 times.
For every simulated data set, the fate of each of the original sites was tracked and an alignment representing the true homology was constructed for each data set (that is, the simulation program produced gapped sequences in which all aligned sites were truly homologous). The gaps were removed from the sequences and data sets consisting of all three sequences and of just sequences A and B were constructed. Each data set was aligned using Clustal W version 1.83 [54] with the default parameters, as is common in high-throughput analysis and comparative studies of this sort [3,6,7,36,55-57]. This produced a hypothesized alignment, just as one would obtain from analysis of real data.
The hypothesized alignments were compared to the true alignment derived from the simulation. Evolutionary distances between sequences A and B were estimated for the correct alignment, the AB hypothesized alignment, and the ABC hypothesized alignment using the Tamura-Nei formula [58].
Authors' contributions
MR designed, programmed, executed, and analyzed all parts of this study.
Acknowledgements
Thanks to Heath Ogden and anonymous reviewers for comments on this manuscript. This work was partially supported by the NIH R03-LM008637 and Arizona State University.
==== Refs
Pevsner J Bioinformatics and Functional Genomics 2003 Hoboken, NJ , Wiley 753
Briffeuil P Baudoux G Lambert C De Bolle X Vinals C Feytmans E Depiereux E Comparative analysis of seven multiple protein sequence alignment servers: Clues to enhances reliability of predictions Bioinformatics 1998 14 357 366 9632831
Thompson JD Plewniak F Poch O A comprehensive comparison of multiple sequence alignment programs Nucleic Acids Research 1999 27 2682 2690 10373585
Duret L Abdeddaim S Higgins D, Taylor W Multiple alignments for structrual, functional, or phylogenetic analyses of homologous sequences Bioinformatics: Sequence, Structure, and Databanks 2000 Oxford , Oxford University Press 51 76
Altschul SF Gish W Doolittle RF Local alignment statistics Methods in Enzymology: Computer Methods for Macromolecular Sequence Analysis Methods in Enzymology 1996 266 San Diego , Academic Press 460 480
Rosenberg MS Evolutionary distance estimation and fidelity of pair wise sequence alignment BMC Bioinformatics 2005 6 102 15840174
Pollard DA Bergman CM Stoye J Celniker SE Eisen MB Benchmarking tools for the alignment of functional noncoding DNA BMC Bioinformatics 2004 5 6 14736341
Frith MC Hansen U Spouge JL Weng Z Finding functional sequence elements by multiple local alignment Nucleic Acids Research 2004 32 189 200 14704356
Margulies EH Blanchette M Haussler D Green ED Identification and characterization of multi-species conserved sequences Genome Research 2003 13 2507 2518 14656959
Xia XH Xie Z Kjer KM 18S ribosomal RNA and tetrapod phylogeny Syst Biol 2003 52 283 295 12775520
Cammarano P Creti R Sanangelantoni AM Palm P The Archaea monophyly issue: A phylogeny of translational elongation factor G(2) sequences inferred from an optimized selection of alignment positions Journal of Molecular Evolution 1999 49 524 537 10486009
Kjer KM Aligned 18S and insect phylogeny Systematic Biology 2004 53 506 514 15503677
Kjer KM Use of rRNA secondary structure in phylogenetic studies to identify homologous positions: An example of alignment and data presentation from the frogs Molecular Phylogenetics and Evolution 1995 4 314 330 8845967
Titus T Frost DR Molecular homology assessment and phylogeny in the lizard family Opluridae (Squamata: Iguania) Molecular Phylogenetics and Evolution 1996 6 49 62 8812305
Morrison DA Ellis JT Effects of nucleotide sequence alignment on phylogeny estimation: A case study of 18S rDNAs of Acpiocomplexa Molecular Biology and Evolution 1997 14 428 441 9100373
Hwang UW Kiim W Tautz D Friedrich M Molecular phylogenetics at the Felsenstein zone: Approaching the Strepsipera probelm using 5.8S and 28S rDNA sequences. Molecular Phylogenetics and Evolution 1998 9 470 480 9667995
Feng DF Doolittle RF Progressive alignment and phylogenetic tree constrution of protein sequences Methods in Enzymology 1990 183 375 387 2314283
Feng DF Doolittle RF Progressive sequence alignment as a prerequisite to correct phylogenetic trees Journal of Molecular Evolution 1987 25 351 360 3118049
Raghava GPS Searle SMJ Audley PC Barber JD Barton GJ OXBench: A benchmark for evaluation of protein multiple sequence alignment accuracy BMC Bioinformatics 2003 4 47 14552658
Thompson JD Plewniak F Poch O BaliBASE: A benchmarch alignment database for the evaluation of multiple sequence alignment programs Bioinformatics 1999 1 87 88 10068696
Rosenberg MS Kumar S Taxon sampling, bioinformatics, and phylogenomics Systematic Biology 2003 52 119 124 12554445
Rosenberg MS Kumar S Incomplete taxon sampling is not a problem for phylogenetic inference Proceedings of the National Academy of Sciences USA 2001 98 10751 10756
Pollock DD Zwickl DJ McGuire JA Hillis DM Increased taxon sampling is advantageous for phylogenetic inference Systematic Biology 2002 51 664 671 12228008
Zwickl DJ Hillis DM Increased taxon sampling greatly reduces phylogenetic error Systematic Biology 2002 51 588 598 12228001
Kim J General inconsistency conditions for maximum parsimony: Effects of branch lengths and increasing numbers of taxa Systematic Biology 1996 45 363 374
Kim J Large-scale phylogenies and measuring the performance of phylogenetic estimators Systematic Biology 1998 47 43 60 12064240
Hendy MD Penny D A framework for the quantitative study of evolutionary trees Systematic Zoology 1989 38 297 309
Graybeal A Is it better to add taxa or characters to a difficult phylogenetic problem? Systematic Biology 1998 47 9 17 12064243
Poe S Swofford DL Taxon sampling revisited Nature 1999 398 299 300 10192331
Fleißner R Sequence alignment and phylogenetic inference Mathematisch-Naturwissenschaftlichen Fakultät 2003 Düsseldorf , Heinrich-Heine-Universität Düsseldorf 132
Murphy WJ Eizirik E Johnson WE Zhang YP Ryder OA O'Brien SJ Molecular phylogenetics and the origins of placental mammals Nature 2001 409 614 618 11214319
Murphy WJ Eizirik E O'Brien SJ Madsen O Scally M Douady CJ Teeling E Ryder OA Stanhope MJ de Jong WW Springer MS Resolution of the early placental mammal radiation using Bayesian phylogenetics Science 2001 294 2348 2351 11743200
Reyes A Gissi C Catzeflis F Nevo E Pesole G Saccone C Congruent mammalian trees from mitochondrial and nuclear genes using Bayesian methods Molecular Biology and Evolution 2004 21 397 403 14660685
Gibbs RA Weinstock GM Metzker ML Muzny DM Sodergren EJ Scherer S Scott G Steffen D Worley KC Burch PE Okwuonu G Hines S Lewis L DeRamo C Delgado O Dugan-Rocha S Miner G Morgan M Hawes A Gill R Holt RA Adams MD Amanatides PG Baden-Tillson H Barnstead M Chin S Evans CA Ferriera S Fosler C Glodek A Gu ZP Jennings D Kraft CL Nguyen T Pfannkoch CM Sitter C Sutton GG Venter JC Woodage T Smith D Lee HM Gustafson E Cahill P Kana A Doucette-Stamm L Weinstock K Fechtel K Weiss RB Dunn DM Green ED Blakesley RW Bouffard GG de Jong J Osoegawa K Zhu BL Marra M Schein J Bosdet I Fjell C Jones S Krzywinski M Mathewson C Siddiqui A Wye N McPherson J Zhao SY Fraser CM Shetty J Shatsman S Geer K Chen YX Abramzon S Nierman WC Havlak PH Chen R Durbin KJ Egan A Ren YR Song XZ Li BS Liu Y Qin X Cawley S Cooney AJ D'Souza LM Martin K Wu JQ Gonzalez-Garay ML Jackson AR Kalafus KJ McLeod MP Milosavljevic A Virk D Volkov A Wheeler DA Zhang ZD Bailey JA Eichler EE Tuzun E Birney E Mongin E Ureta-Vidal A Woodwark C Zdobnov E Bork P Suyama M Torrents D Alexandersson M Trask BJ Young JM Huang H Wang HJ Xing HM Daniels S Gietzen D Schmidt J Stevens K Vitt U Wingrove J Camara F Alba MM Abril JF Guigo R Smit A Dubchak I Rubin EM Couronne O Poliakov A Hubner N Ganten D Goesele C Hummel O Kreitler T Lee YA Monti J Schulz H Zimdahl H Himmelbauer H Lehrach H Jacob HJ Bromberg S Gullings-Handley J Jensen-Seaman MI Kwitek AE Lazar J Pasko D Tonellato PJ Twigger S Ponting P Duarte JM Rice S Goodstadt L Beatson SA Emes RD Winter EE Webber C Brandt P Nyakatura G Adetobi M Chiaromonte F Elnitski L Eswara P Hardison RC Hou MM Kolbe D Makova K Miller W Nekrutenko A Riemer C Schwartz S Taylor J Yang S Zhang Y Lindpaintner K Andrews TD Caccamo M Clamp M Clarke L Curwen V Durbin R Eyras E Searle SM Cooper GM Batzoglou S Brudno M Sidow A Stone EA Payseur BA Bourque G Lopez-Otin C Puente XS Chakrabarti K Chatterji S Dewey C Pachter L Bray N Yap VB Caspi A Tesler G Pevzner PA Haussler D Roskin KM Baertsch R Clawson H Furey TS Hinrichs AS Karolchik D Kent WJ Rosenbloom KR Trumbower H Weirauch M Cooper DN Stenson PD Ma B Brent M Arumugam M Shteynberg D Copley RR Taylor MS Riethman H Mudunuri U Peterson J Guyer M Felsenfeld A Old S Mockrin S Collins F Genome sequence of the Brown Norway rat yields insights into mammalian evolution Nature 2004 428 493 521 15057822
Hickson RE Simon C Perrey SW The performance of several multiple-sequence alignment programs in relation to secondary-structure features for an rRNA sequence Molecular Biology and Evolution 2000 17 530 539 10742045
Notredame C Higgins DG Heringa J T-Coffee: A novel method for fast and accurate multiple sequence alignment Journal of Molecular Biology 2000 302 205 217 10964570
Keightley PD Johnson T MCALIGN: Stochastic alignment of noncoding DNA sequences based on an evolutionary model of sequence evolution Genome Research 2004 14 442 450 14993209
Holmes I Bruno WJ Evolutionary HMMs: A Bayesian approach to multiple alignment Bioinformatics 2001 17 803 820 11590097
Thorne JL Kishino H Felsenstein J Inching toward reality: An improved likelihood model of sequence evolution Journal of Molecular Evolution 1992 34 3 16 1556741
Thorne JL Kishino H Felsenstein J An evolutionary model for maximul likelihood alignment of DNA sequences Journal of Molecular Evolution 1991 33 114 124 1920447
Metzler D Fleißner R Wakolbinger A von Haeseler A Assessing variability by joint sampling of alignments and mutation rates Journal of Molecular Evolution 2001 53 660 669 11677626
Hein J Wiuf C Knudsen B Møller MB Wibling G Statistical alignment: Computational properties, homology testing and goodness-of-fit Journal of Molecular Biology 2000 302 265 279 10964574
Fleißner R Metzler D von Haeseler A Bornberg-Bauer E, Rost U, Stoye J, Vingron M Can one estimate distances from pairwise sequence alignments? Proceedings of the German Conference on Bioinformatics 2000 Berlin , Logos Verlag 89 95
Gladstein D Wheeler WC POY: The Optimization of Alignment Characters 1997 New York , American Museum of Natural History
Redelings BD Suchard MA Joint Bayesian estimation of alignment and phylogeny Systematic Biology In press 16012107
Lunter G Miklos I Drummond A Jensen JL Hein J Bayesian coestimation of phylogeny and sequence alignment BMC Bioinformatics 2005 6 83 15804354
Fleissner R Metzler D von Haeseler A Simultaneous statistical multiple alignment and phylogeny reconstruction Systematic Biology 2005 54 548 561 16085574
Rosenberg MS MySSP: Non-stationary evolutionary sequence simulation, including indels Evolutionary Bioinformatics Online 2005 1 51 53
Hasegawa M Kishino H Yano T Dating of the human-ape splitting by a molecular clock of mitochondrial DNA Journal of Molecular Evolution 1985 22 160 174 3934395
Rosenberg MS Subramanian S Kumar S Patterns of transitional mutation biases within and among mammalian genomes Molecular Biology and Evolution 2003 20 988 993 12716982
Ophir R Graur D Patterns and rates of indel evolution in processed pseudogenes from humans and murids Gene 1997 205 191 202 9461394
Sundström H Webster MT Ellegren H Is the rate of insertion and deletion mutation male baised?: Molecular evolutionary analysis of avian and primate sex chromosome sequences Genetics 2003 164 259 268 12750337
Thompson JD Higgins DG Gibson TJ CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice Nucleic Acids Research 1994 22 4673 4680 7984417
Morgenstern B DIALIGN 2: Improvement of the segment-to-segment approach to multiple sequence alignment Bioinformatics 1999 15 211 218 10222408
Bray N Dubchak I Pachter L AVID: A global alignment program Genome Research 2003 13 97 102 12529311
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 Research 2003 13 721 731 12654723
Tamura K Nei M Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees Molecular Biology and Evolution 1993 10 512 526 8336541
|
16305750
|
PMC1318491
|
CC BY
|
2021-01-04 16:27:47
|
no
|
BMC Bioinformatics. 2005 Nov 23; 6:278
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-278
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-871630305210.1186/1743-422X-2-87ReviewReview of the temporal and geographical distribution of measles virus genotypes in the prevaccine and postvaccine eras Riddell Michaela A [email protected] Jennifer S [email protected] Paul A [email protected] Scientist/PhD Scholar, Victorian Infectious Diseases Reference Laboratory/WHO Western Pacific Measles Regional Reference Laboratory and Department of Public Health, School of Population Health, University of Melbourne, Parkville 3010, Victoria, Australia2 Centers for Disease Control and Prevention, Atlanta, GA, 30333 USA3 Dept. Molecular Microbiology and Immunology, Johns Hopkins School of Public Health, Baltimore MD 21205 USA2005 22 11 2005 2 87 87 3 6 2005 22 11 2005 Copyright © 2005 Riddell et al; licensee BioMed Central Ltd.2005Riddell 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.
Molecular epidemiological investigation of measles outbreaks can document the interruption of endemic measles transmission and is useful for establishing and clarifying epidemiological links between cases in geographically distinct clusters. To determine the distribution of measles virus genotypes in the prevaccine and postvaccine eras, a literature search of biomedical databases, measles surveillance websites and other electronic sources was conducted for English language reports of measles outbreaks or genetic characterization of measles virus isolates. Genotype assignments based on classification systems other than the currently accepted WHO nomenclature were reassigned using the current criteria. This review gives a comprehensive overview of the distribution of MV genotypes in the prevaccine and postvaccine eras and describes the geographically diverse distribution of some measles virus genotypes and the localized distributions of other genotypes.
==== Body
Introduction
Although measles virus (MV) is serologically monotypic, the genetic characterization of wild-type viruses has identified eight clades (A – H), which have been divided into 22 genotypes and one proposed genotype. Clades B, C, D, G and H each contain multiple genotypes (B1 – 3, C1 – 2, D1 – 10, G1 – 3, H1 – 2) while clades A, E and F each contain a single genotype (A, E, F) [1,2]. The sequences of the vaccine strains indicate that the wild type viruses from which they were derived were all members of genotype A. All measles genotypes can be neutralized by serum from vaccinated persons in vitro, although with varying efficiency [3,4]. There are no known biological differences between viruses of different genotypes. Specific measles genotypes are not associated with differences in severity of disease, likelihood of developing severe sequela such as subacute sclerosing panencephalitis or inclusion body encephalitis, or variability in sensitivity of laboratory diagnosis.
Analysis of the variability in the nucleotide sequences of wild-type MVs has enabled the use of molecular epidemiologic techniques for measles surveillance. The molecular data, when used in conjunction with standard case reporting and investigation, can help to identify epidemiological links between geographically distinct cases and outbreaks as well as track importations of MV. [5-7]. Also, approximately 5% of vaccine recipients experience mild symptoms (rash and fever) after vaccination and these cases could be misclassified as wild-type measles [8]. Genetic characterization of viral isolates or RT-PCR products is the only laboratory test that can differentiate between vaccine-associated cases and wild-type infection [6,9,10].
In 1998, the World Health Organization (WHO) recommended a standard protocol for the designation of measles genotypes. These recommendations, updated in 2001 and 2003, also included a standard analysis protocol and designation of standard reference strains (see Additional file 2) against which all newly characterized isolates should be compared [2,11,12]. The minimum amount of sequence data required to assign a virus to a genotype are the 450 nucleotides encoding the carboxy terminus of the N protein. The entire sequence of the coding region of the H gene should be obtained from representative isolates [11]. New genotypes are designated if the nucleotide sequence differs from the closest reference sequence by more than 2.5% in N and 2.0% in H [2]. Additionally, phylogenetic analysis should produce similar tree topographies using at least two different analysis methods. Several isolates or clinical specimens should be sequenced and at least one viral isolate should be available as the reference strain. Finally, new genotype classifications should be useful for epidemiological studies, by providing a means to identify the source or transmission pathway of infection and by contributing to our understanding of the global distribution of MV genotypes [2].
The purpose of this summary is to collate all available reports of MV genotypes and to standardize the published genotype nomenclature, according to the current WHO criteria, with the aim of giving a comprehensive overview of the distribution of MV genotypes in the prevaccine and postvaccine eras.
Methods
An examination of the National Library of Medicine "PubMed" [13] search engine using the keyword "measles" combined with "genotypes" and "sequence" was performed to identify English language publications or abstracts describing measles genotyping.
Additional sources included the reference lists of articles identified by "PubMed" and electronic sources such as the CDC and PAHO measles network Internet pages and the NCBI Genbank website [14-16]. Measles outbreak alerts were received through the WHO network, which distributes outbreak notifications. In addition, subscription based electronic newsletters such as ProMED mail [17] and Immunization newsbrief [18] were scrutinized for information relating to measles outbreaks. Direct contact was made with the notifying laboratory or health unit requesting genotype information if available.
A table produced by participants at the 1998 WHO meeting listed older classification systems and the comparable genotype classifications under the universally accepted system. This table was used to reclassify genotypes cited in publications prior to 1998 [11]. In some cases, later publications from the same or other groups were used to assign current genotypes to viruses classified before 1998.
Results and Discussion
One hundred and twenty eight studies were identified through the PubMed search, 67 of which described the genotype of MV isolates. Four internet websites were identified (including Genbank) and a further 27 articles were identified from the reference lists of cited publications or from outbreak notification lists such as ProMED [17] and Immunization Newsbrief [18].
Figure 1 and Table 1 summarize the temporal and geographical distribution of MV genotypes from the early 1950s to 2004 but do not differentiate between cases of endemic or imported measles virus. Genotype and location specific references are not cited in the following results section but can be found in the relevant genotype specific section in the comprehensive table which accompanies this paper (see Additional file 1, also available from the website of the WHO Western Pacific Regional Reference Laboratory for measles, The Victorian Infectious Diseases Reference Laboratory, Melbourne, Australia ).
Figure 1 Temporal distribution of measles virus genotypes 1951 – 2004. Summary of distribution of MV genotypes from the prevaccine era to 2004. Refer to Additional file 1 for complete referencing of data shown in figure. Data reflects publications available as of August 2005.
Table 1 Distribution of MV genotypes by WHO geographical region 1950s – 2004. Countries in which MV virus has been detected. No distinction has been made between endemic transmission or instances of MV importation. (data reflects publications available as of August 2005). Refer to Additional file 1 for details of endemic transmission and imported measles cases and for complete referencing of data.
Geno-type AFRO1 EMRO2 SEARO3 WPRO4 Europe5 Americas6
A China, Japan Romania, UK Finland, Russia, Czech Republic, Slovakia Brazil, USA, Argentina,
B1 Cameroon
B2 Gabon South Africa, Angola
B3 Gambia, Nigeria, Kenya, Ghana, Algeria, Cameroon, Rep. of Congo, Dem. Rep. of Congo Burkina Faso, Equatorial Guinea Sudan, Tunisia, Libya France, Spain Germany, UK, USA
C1 Japan Nth Ireland, Spain, Germany USA, Canada, Argentina
C2 Zimbabwe Morocco Australia Austria, France, Belgium, Netherlands, Czech Republic, Slovakia, Spain, Italy, Germany, UK, Luxembourg, Denmark, USA, Brazil, Canada,
D1 Australia UK, Nth Ireland
D2 South Africa, Zambia Ireland, UK, Spain USA
D3 South Africa Micronesia, Philippines, PNG, Japan, Australia, Taiwan UK, Denmark USA, Canada
D4 South Africa, Namibia, Kenya, Ethiopia Pakistan, Lebanon, Afghanistan, Syria, Iran India, Nepal Japan, Australia UK, Denmark, Netherlands, Germany, Spain, Croatia, Russia, USA, Canada
D5 Namibia Thailand, Bangladesh Japan, Malaysia, Micronesia, Australia, New Zealand, Cambodia, Guam, Rep. of Korea UK, Germany South America, USA, Canada, Brazil
D6 UK, Ireland, Spain, Germany, Austria, Italy, Greece, Croatia, Turkey, Ukraine, Poland, Russia, Luxembourg, Bosnia, Israel, Norway, Denmark, Netherlands USA, Canada, Brazil, Bolivia, Argentina, Uruguay, Dominican Republic \Haiti
D7 Sri Lanka, Myanmar (Burma), India Australia UK, Germany, Sweden, Europe, France, Spain, Italy El Salvador, USA, Canada, Mexico
D8 Ethiopia Pakistan, Oman, India, Bangladesh, Nepal Australia UK, Spain, Yugoslavia, Albania, Italy, Lithuania USA, Canada
D9 Indonesia Australia, Japan Europe Venezuela, Colombia
D10 Uganda
E Germany, Denmark USA, Canada
F Spain
G1 USA
G2 South Africa Indonesia Australia, Malaysia Netherlands, UK, Germany Mexico, USA
G3 E. Timor, Indonesia Australia
H1 Thailand Australia, China, New Zealand, Mongolia, Singapore, Japan, Rep. of Korea, Rep. of Marshall Islands UK, Spain, Netherlands, Denmark, Germany USA, Canada, Chile, Mexico
H2 China, Vietnam, Australia USA
1AFRO – WHO African region, 2EMRO – WHO Eastern Mediterranean region, 3SEARO – WHO South East Asian region, 4WPRO – WHO Western Pacific region, 5Europe – WHO European region, 6Americas – WHO region of the Americas
Routine molecular characterization of wild-type measles viruses was initiated in response to a global resurgence of measles disease in the late 1980s and the concurrent availability of sensitive techniques (e.g. RT-PCR and automated sequencing) for the investigation of viral genomes. Prior to that date, only a few isolates of measles were available for molecular characterization and reliable epidemiologic information was not available for many of these isolates. In the era before the widespread use of measles vaccine, genotypes A, C1, and D1 were detected. Genotype A virus includes the prototype Edmonston strain, the progenitor for most of the current measles vaccines. Analysis of MV sequences obtained from SSPE cases, resulting from initial infections that occurred during the 1950s and 1960s, detected genotypes C1, D1, E and F, providing further evidence that genotype A was not the only genotype detected during the prevaccine era [19-24]. However, data from these earlier studies must be interpreted cautiously due to the large number of mutations in SSPE sequences and the lack of standardization. Of course, detection of various genotypes in SSPE cases reflects efforts to study this devastating illness and should not be taken as an indication that one genotype is more likely to cause SSPE than another [25]. Retrospective sequence analysis of viral isolates collected during the 1970s showed continued detection of genotypes C1 and D1 and the first detections of genotypes C2, D2, D4, E and F.
As virologic surveillance expanded in the late 1980s and 1990s, the number of genotypes detected in cases and outbreaks increased substantially to include the 23 genotypes now recognized by the WHO. However, some genotypes (B1, D1, E, F, G1) have not been detected in the last 15 years and are considered inactive.
Genotype A has been detected in acute cases of measles in South and North America, China, Japan, Eastern Europe, Finland and the UK over the last 40 years. Since, it is difficult to distinguish wild-type viruses in genotype A from vaccine strains, these reports must be interpreted with caution since some of the sequences may have been derived from vaccine associated cases or been the result of laboratory contamination [22,26,27]. In the future, detection of genotype A viruses in association with acute cases of measles will need to be thoroughly scrutinized and additional sequence data will need to be obtained from both clinical samples and corresponding viral isolates.
Genotype B2, previously considered inactive [12], has recently been detected in South Africa and Angola [28]. Genotype B3 was first detected in 1993 in Gambia but has subsequently been detected in cases from Cameroon, Nigeria, Ghana, Burkina Faso, DR Congo and the Sudan. This genotype is the endemic genotype of West and Central Africa and has been imported into numerous countries including France, Germany and the USA.
Outbreaks involving genotype C1 have occurred in Canada, Japan, Germany and most recently in the early 1990s in Argentina, which was the last reported outbreak involving genotype C1 circulation. Genotype C2 has circulated widely throughout the European continent and has been exported to the USA and Canada from France, Italy and Germany, where it was known to be an endemic genotype until 2001. This genotype was also identified in Australia from 1990 to 1991 and Morocco in 1998 & 1999. An importation of genotype C2 to the USA was linked with travel from Zimbabwe in 1998 although there are no reports to indicate that this strain was circulating in Southern Africa during this time [6].
Characterization of archived MV isolates in Australia from 1971, suggest that genotype D1 may have been the endemic strain in Australia during the pre-vaccine era. Sequences from SSPE cases in Northern Ireland and the UK indicate that genotype D1 was also detected in Britain before the widespread use of vaccine. Genotype D1 has not been detected since 1986 and is considered inactive. Genotype D2 appears to have been the endemic strain of Southern Africa from the late 1970s to 2000. This genotype was also responsible for the large outbreak in Ireland in 1999 – 2000, which resulted in importations to both the UK and USA. Genotype D3 is currently endemic in Papua New Guinea and possibly the Philippines, given that several measles cases in the USA have been linked with travel from the Philippines. Additionally this genotype has been associated with a case of SSPE in South Africa, and has been detected in Australia, USA and Canada, the UK and Denmark, in most cases with epidemiological links of importation from Japan or the Philippines. Genotype D4 is widely distributed and has been associated with multiple outbreaks on the Indian sub-continent, East and South Africa and a large outbreak in Quebec Province, Canada in 1989. Recently genotype D4 viruses, imported from the Indian sub-continent and East and South Africa, have been epidemiologically linked with cases in Canada, the USA, the UK, other European countries and Australia. Genotypes D4 and D2 appear to have been co-circulating in Southern Africa from the late 1970s to the late 1990s. Genotype D5 is endemic in Cambodia and has been associated with measles cases detected in the Americas, the UK, Germany and Australia. Epidemiological investigations have identified Japan and Thailand as the main sources for these importations. Until recently both genotype D3, and genotype D5 were endemic in Japan [26,29-31]. However, recent evidence suggests that these genotypes may no longer be predominant in Japan [32]. Genotype D6 has circulated widely throughout the European continent and may have been the endemic genotype of Europe, in conjunction with genotype C2, since the 1990s. This genotype is endemic in Turkey [33] and the Russian Federation [34]. Genotype D7 circulated in the UK and Australia during the 1980s. Chains of transmission of this genotype have been associated, through epidemiological investigations, with Sweden and other European countries, including Italy where it was identified in the large measles outbreak in 2002. Genotype D7 has been imported into the US from multiple European sources from 2001 to 2003. Recently this genotype replaced genotypes C2 and D6 as the most commonly isolated genotype in Germany [35]. Genotype D8 appears to be co-circulating with genotype D4 on the Indian sub-continent and Ethiopia, although the first known description of this genotype was in the UK, from where it has been regularly detected. However, investigations have linked UK cases with importations of virus not only from the Indian sub-continent but also from the Balkans and Oman [36]. Genotype D8 has been imported into Australia and the USA from India and Bangladesh. Genotype D9, first described after importation to Australia from Indonesia (Bali) in 1999, was isolated during the large outbreak in 2000 – 2001 in Colombia and Venezuela. D9 was associated with an outbreak in Japan in 2004. Analysis of wild-type viruses isolated in Uganda in 2000–2002 indicated the presence of a new genotype, which has been proposed as genotype d10 [Genbank accession numbers AY923185 through AY923212] [37].
A few genotype E viruses and related SSPE cases were reported in the early 1970s. Genotype F sequences have been identified on two occasions, both were SSPE cases wherein acute measles infection was documented in 1967 and 1968. Thus both genotypes E and F probably circulated in the pre-vaccine era [19,22].
Clade G, previously consisting of one genotype (G), has recently been expanded to contain three genotypes. The original genotype G (now G1) had not been detected since 1983 and was thought to have been extinct. However, recent investigations have identified two new genotypes (G2 & G3), both of which have been predominantly associated with chains of transmission within and importation from Indonesia and Malaysia [38-40].
Clade H viruses originally consisted of a single genotype but recently this clade has been expanded to contain two genotypes (H1 & H2). Both genotypes are predominant in the Asian and South East Asian regions. Genotype H1 has mainly been associated with transmission within or importations from China and was detected during the large measles epidemic in Korea in 2000 – 2001 [41]. Genotype H1 may now be dominant in Japan [32,42]. The WHO Western Pacific Regional Reference Laboratory for measles recently confirmed circulation of this genotype in Mongolia. Genotype H2, first described from samples recovered from China, has been more recently associated with importations from Vietnam [43].
Some genotypes of MV are associated with a particular geographical region, while other genotypes are more widely distributed. In particular, clade B is predominant in measles transmission in Sub Saharan and Central Africa, clade G in South East Asia and clade H in South East Asia and China. Clade D viruses, on the other hand, appear to be more widely distributed and are endemic in Eastern Africa, parts of Europe and the Indian sub-continent.
Determination of measles genotypes in countries that have not yet conducted molecular surveillance can be investigated, by proxy, from cases epidemiologically linked to imported cases. For example, the Philippines have not reported an endemic MV genotype but multiple importations to the USA associated with travel to, or contact with, the Philippines have resulted in the supposition that genotype D3 is the predominant circulating genotype in the Philippines [6,44]. However, caution must be taken when identifying genotypes by proxy as the genotype detected may not be the type that is endemic in the region. In some cases, genotypes have been epidemiologically linked to countries with no history of circulation of that genotype. For example, genotype G2 has been reportedly associated with importations to the UK from Mexico, South Africa and Australia, none of which have reported endemic circulation of genotype G2 [36]. In these cases infection may have occurred while the patient was in transit or at venues frequented by other travellers and might not reflect the circulating genotype.
Simultaneous circulation of multiple genotypes has been reported in several regions. Genotypes D3 and D5 co-circulated in Japan since the mid 1980s and the relative number of isolations changed over time. During the late 1980s genotype D3 was detected more frequently, but by 1990 D5 was more common [42,45]. Genotypes D2 and D4 appear to be co-circulating throughout Eastern and Southern Africa. Genotypes C2 and D6 continue to be detected in some parts of Europe and North Africa [35].
Rima et al described a shift from genotype C2 to genotype D6 in Spain in the early 1990s [24]. Santibanez et al recently demonstrated the shift from detection of mostly genotypes C2 and D6 in Germany to detection of mostly genotype D7 [46]. The shift of genotypes occurs in countries that have sub-optimal measles control programs, resulting in interruption of endemic transmission for short periods. However, failure to maintain high levels of population immunity results in the accumulation of susceptible individuals and creates conditions that favour the rapid transmission of a newly introduced genotype. Therefore, the apparent genotype switching is most likely due to changes in the distribution of susceptible individuals in the region.
Nine new MV genotypes have been identified since 1990 reflecting increased surveillance of measles cases and technological advances, rather than recent evolution. The designation of new genotypes, such as the newly proposed genotype d10, is likely to continue as the molecular analysis of viral isolates becomes routinely integrated into more countries within the global WHO measles laboratory network and more sequence data are added to the database. For example, genotype B3 may eventually be reclassified as two separate genotypes since this genotype contains viruses in two distinct clusters [47-49]. Characterization of viruses imported into Australia has detected three previously unrecognised genotypes (D7, D9 & G3) due partly to the frequency of travel between South East Asia and Australia and also to the comprehensive measles surveillance conducted by Australian laboratories [7,38,50].
The mutation rate amongst field isolates of MV is low and appears to be random rather than driven by vaccine pressure or immune responses [3,24,26]. Within a genotype, nucleotide differences (virus lineage) can assist in distinguishing separate episodes of transmission [24,51,52]. In countries or regions with endemic (ongoing and constant MV transmission) measles, many lineages of a single genotype may co-exist; however as countries begin to move from endemic to epidemic measles (MV transmission resulting in a higher number of cases than normally expected, typically against a background of little or no MV transmission)[53], the diversity of sequences within the circulating genotypes decreases [43,54-57]. In fact, the genotype D6 virus associated with a large measles outbreak that occurred in several South American countries between 1996 and 1997 had identical N gene sequences suggesting rapid spread of a single lineage [51]. Analysis of measles viruses circulating in Burkina Faso, before and after a mass vaccination campaign, showed that the number of circulating lineages was greatly reduced following the campaign. Sequence analysis of viruses isolated from outbreaks that occurred after the vaccination campaign suggested that virus was introduced from a single source [57].
Many recent measles outbreaks have been reported with no accompanying molecular genotyping investigations, for instance in Afghanistan [58], Niger [59] and the Philippines [60]. These outbreaks highlight the need to extend molecular surveillance capabilities to regions where measles remains endemic. Recent studies have described the recovery of MV RNA by RT-PCR from oral fluid, dried blood and dried oral fluid [61-63]. These samples, which are easy to collect, prepare and transport by post to laboratories capable of MV genotyping, have the potential to extend molecular surveillance for measles virus to remote settings and countries with limited infrastructure. However conventional samples such as nasopharyngeal swabs, urine and peripheral blood lymphocytes should continue to be collected, if logistically possible, because of the higher sensitivity of these sample types for detecting MV RNA.
Molecular surveillance undertaken in the early stages of measles control can facilitate identification of endemic genotypes. Over time or after intervention programs continued molecular surveillance, in conjunction with case based epidemiological investigations, can detect the interruption of endemic transmission [1]. Additionally, molecular analysis of specimens from cases facilitates both linkage to, and separation from, contemporaneous cases and clusters, assisting classical epidemiological investigations and the tracking of chains of transmission [7,64].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MAR initiated the review and drafted the preliminary manuscript. JSR and PAR provided additional data and contributed to manuscript revisions. All authors read and approved the final manuscript.
Supplementary Material
Additional File 2
Sequence and alignment of World Health Organisation designated measles virus reference strains. Aligned sequence of the 456 nucleotides at the COOH terminus of the N protein for each measles virus reference strain as designated by the World Health Organisation. These sequences should be used in phylogenetic analysis to determine measles virus genotype of newly derived sequence.
Click here for file
Additional File 1
Table: Temporal and geographical distribution of measles of measles virus genotypes 1950 – 2004 (data reflects publications available as of August 2005). Measles virus genotypes listed alphabetically, by year of circulation, location and associated publication.
Click here for file
Acknowledgements
Thanks to Doris Chibo, Graham Tipples, David Brown, Li Jin for helpful suggestions and clarification of genotypes included in the table. Thanks to Doris Chibo and Heath Kelly for critical review of the earlier drafts of the manuscript. MAR received funding through a National Health and Medical Research Council Public Health PhD Research Scholarship. The authors welcome amendments, additions and updates to the comprehensive table submitted as Additional file 1. Regularly updated versions of the additional file will be available from the measles Global Specialized Laboratory at the Centers for Disease Control and Prevention, Atlanta Georgia, USA
==== Refs
Rota PA Bellini WJ Update on the global distribution of genotypes of wild type measles viruses J Infect Dis 2003 187 (Suppl 1) S270 276 12721925 10.1086/368042
World Health Organization Nomenclature for describing the genetic characteristics of wild-type measles viruses (update). Part I Wkly Epidemiol Rec 2001 76 242 247 11515240
Tamin A Rota PA Wang ZD Heath JL Anderson LJ Bellini WJ Antigenic analysis of current wild type and vaccine strains of measles virus J Infect Dis 1994 170 795 801 7930720
Klingele M Hartter HK Adu F Ammerlaan W Ikusika W Muller CP Resistance of recent measles virus wild-type isolates to antibody-mediated neutralization by vaccinees with antibody J Med Virol 2000 62 91 98 10935994 10.1002/1096-9071(200009)62:1<91::AID-JMV14>3.0.CO;2-B
Mulders MN Truong AT Muller CP Monitoring of measles elimination using molecular epidemiology Vaccine 2001 19 2245 2249 11257341 10.1016/S0264-410X(00)00453-9
Rota PA Liffick SL Rota JS Katz RS Redd S Papania M Bellini WJ Molecular epidemiology of measles viruses in the United States, 1997-2001 Emerg Infect Dis 2002 8 902 908 12194764
Chibo D Riddell MA Catton MG Lyon M Lum G Birch CJ Studies of measles viruses circulating in Australia between 1999 and 2001 reveals a new genotype Virus Res 2003 91 213 221 12573500 10.1016/S0168-1702(02)00273-3
Redd SC Markowitz LE Katz SL Plotkin SA and Orenstein WA Measles Vaccine Vaccines 1999 3rd Philadelphia, Saunders, W. B. 222 2266
Jenkin GA Chibo D Kelly HA Lynch PA Catton MG What is the cause of a rash after measles-mumps-rubella vaccination? Med J Aust 1999 171 194 195 10494235
Rota PA Khan AS Durigon E Yuran T Villamarzo YS Bellini WJ Detection of measles virus RNA in urine specimens from vaccine recipients J Clin Microbiol 1995 33 2485 2488 7494055
World Health Organization Expanded Programme on Immunization (EPI). Standardization of the nomenclature for describing the genetic characteristics of wild-type measles viruses Wkly Epidemiol Rec 1998 73 265 269 9745371
World Health Organization Update of the nomenclature for describing the genetic characteristics of wild- type measles viruses: new genotypes and reference strains Wkly Epidemiol Rec 2003 78 229 232 12866261
Medline/PubMed
Centers for Disease Control and Prevention Global Laboratory Network for Measles Surveillance : USA Genotyping Results 2000 - 2004
Pan American Health Organization Measles main page
Benson DA Karsch-Mizrachi I Lipman DJ Ostell J Wheeler DL GenBank: update Nucleic Acids Res 2004 32 D23 6 14681350 10.1093/nar/gkh045
International Society for Infectious Diseases ProMED mail
Infectious Diseases Society of America Immunization Newsbrief
Rima BK Earle JA Yeo RP Herlihy L Baczko K ter Meulen V Carabana J Caballero M Celma ML Fernandez-Munoz R Temporal and geographical distribution of measles virus genotypes J Gen Virol 1995 76 1173 1180 7730801
Rota JS Bellini WJ Rota PA Thompson RCA Measles Molecular Epidemiology of Infectious Diseases 2000 London, Arnold 168 180
Rota PA Bloom AE Vanchiere JA Bellini WJ Evolution of the nucleoprotein and matrix genes of wild-type strains of measles virus isolated from recent epidemics Virology 1994 198 724 730 8291252 10.1006/viro.1994.1086
Christensen LS Scholler S Schierup MH Vestergaard BF Mordhorst CH Sequence analysis of measles virus strains collected during the pre- and early-vaccination era in Denmark reveals a considerable diversity of ancient strains APMIS 2002 110 113 122 12064866 10.1034/j.1600-0463.2002.100201.x
Jin L Beard S Hunjan R Brown DW Miller E Characterization of measles virus strains causing SSPE: a study of 11 cases J Neurovirol 2002 8 335 344 12161818 10.1080/13550280290100752
Rima BK Earle JA Baczko K ter Meulen V Liebert UG Carstens C Carabana J Caballero M Celma ML Fernandez-Munoz R Sequence divergence of measles virus haemagglutinin during natural evolution and adaptation to cell culture J Gen Virol 1997 78 97 106 9010291
Rima BK Kurstak E Molecular biological basis of measles virus strain differences Measles and Poliomyelitis Vaccines, Immunization and Control 1993 Vienna, Springer-Verlag 149 160
Rota PA Rota JS Bellini WJ Molecular epidemiology of measles virus Semin Virol 1995 6 379 386 10.1016/S1044-5773(05)80015-0
Bellini WJ Rota PA Genetic diversity of wild-type measles viruses: implications for global measles elimination programs Emerg Infect Dis 1998 4 29 35 9452396
Smit SB Hardie D Tiemessen CT Measles virus genotype B2 is not inactive: evidence of continued circulation in Africa J Med Virol 2005 77 550 557 16254960 10.1002/jmv.20491
Takahashi M Nakayama T Kashiwagi Y Takami T Sonoda S Yamanaka T Ochiai H Ihara T Tajima T Single genotype of measles virus is dominant whereas several genotypes of mumps virus are co-circulating J Med Virol 2000 62 278 285 11002259 10.1002/1096-9071(200010)62:2<278::AID-JMV21>3.0.CO;2-1
Nakayama T Mori T Yamaguchi S Sonoda S Asamura S Yamashita R Takeuchi Y Urano T Detection of measles virus genome directly from clinical samples by reverse transcriptase-polymerase chain reaction and genetic variability Virus Res 1995 35 1 16 7754670 10.1016/0168-1702(94)00074-M
Katayama Y Shibahara K Kohama T Homma M Hotta H Molecular epidemiology and changing distribution of genotypes of measles virus field strains in Japan J Clin Microbiol 1997 35 2651 2653 9316925
Zhou J Fujino M Inou Y Kumada A Aoki Y Iwata S Nakayama T H1 genotype of measles virus was detected in outbreaks in Japan after 2000 J Med Virol 2003 70 642 648 12794730 10.1002/jmv.10443
Guris D Bayazit Y Ozdemirer U Buyurgan V Yalniz C Toprak I Aycan S Measles epidemiology and elimination strategies in Turkey J Infect Dis 2003 187 Suppl 1 S230 4 12721918
Ignatyev G Agafonov AP Kameneva S Atrasheuskaya A Molecular epidemiological study of measles in Western Siberia (Russia) during 1995 - 2001: in XIIth International Congress of Virology 2002 Poster Abstract 143V- #V1302 IUMS Paris, France
Tischer A Santibanez S Siedler A Heider A Hengel H Laboratory investigations are indispensable to monitor the progress of measles elimination--results of the German Measles Sentinel 1999-2003 J Clin Virol 2004 31 165 178 15465408 10.1016/j.jcv.2004.05.007
Ramsay ME Jin L White J Litton P Cohen B Brown D The elimination of indigenous measles transmission in England and Wales J Infect Dis 2003 187 (Suppl 1) S198 207 12721914 10.1086/368024
Muwonge A Nanyunja M Rota PA Bwogi J Lowe L Liffick S Bellini WJ Sylvester S New genotype of measles virus detected in Uganda Emerg Infect Dis 2005 11 1522 1526 16318690
Chibo D Riddell M Catton M Birch C Novel measles virus genotype, East Timor and Australia Emerg Infect Dis 2002 8 735 737 12095446
de Swart RL Wertheim-van Dillen PM van Binnendijk RS Muller CP Frenkel J Osterhaus AD Measles in a Dutch hospital introduced by an immuno-compromised infant from Indonesia infected with a new virus genotype Lancet 2000 355 201 202 10675124 10.1016/S0140-6736(99)04652-8
Rota PA Liffick S Rosenthal S Heriyanto B Chua KB Measles genotype G2 in Indonesia and Malaysia Lancet 2000 355 1557 1558 10801203 10.1016/S0140-6736(05)74612-2
Na BK Lee JS Shin GC Mi Shin J Lee JY Chung JK Ha DR Lee JK Ma SH Cho HW Kang C Kim WJ Sequence analysis of hemagglutinin and nucleoprotein genes of measles viruses isolated in Korea during the 2000 epidemic Virus Res 2001 81 143 149 11682133 10.1016/S0168-1702(01)00346-X
Nakayama T Zhou J Fujino M Current status of measles in Japan J Infect Chemother 2003 9 1 7 12673398 10.1007/s10156-002-0206-6
Xu W Tamin A Rota JS Zhang L Bellini WJ Rota PA New genetic group of measles virus isolated in the People's Republic of China Virus Res 1998 54 147 156 9696123 10.1016/S0168-1702(98)00020-3
Rota JS Rota PA Redd SB Redd SC Pattamadilok S Bellini WJ Genetic analysis of measles viruses isolated in the United States, 1995-1996 J Infect Dis 1998 177 204 208 9419189
Kubo H Iritani N Seto Y Co-circulation of two genotypes of measles virus and mutual change of the prevailing genotypes every few years in Osaka, Japan J Med Virol 2003 69 273 278 12683418 10.1002/jmv.10278
Santibanez S Tischer A Heider A Siedler A Hengel H Rapid replacement of endemic measles virus genotypes J Gen Virol 2002 83 2699 2708 12388805
Truong AT Kreis S Ammerlaan W Hartter HK Adu F Omilabu SA Oyefolu AO Berbers GA Muller CP Genotypic and antigenic characterization of hemagglutinin proteins of African measles virus isolates Virus Res 1999 62 89 95 10513290 10.1016/S0168-1702(99)00072-6
El Mubarak HS van de Bildt MW Mustafa OA Vos HW Mukhtar MM Ibrahim SA Andeweg AC El Hassan AM Osterhaus AD de Swart RL Genetic characterization of wild-type measles viruses circulating in suburban Khartoum, 1997-2000 J Gen Virol 2002 83 1437 1443 12029159
Kouomou DW Nerrienet E Mfoupouendoun J Tene G Whittle H Wild TF Measles virus strains circulating in Central and West Africa: Geographical distribution of two B3 genotypes J Med Virol 2002 68 433 440 12226833 10.1002/jmv.10222
Chibo D Birch CJ Rota PA Catton MG Molecular characterization of measles viruses isolated in Victoria, Australia, between 1973 and 1998 J Gen Virol 2000 81 2511 2518 10993941
Oliveira MI Rota PA Curti SP Figueiredo CA Afonso AM Theobaldo M Souza LT Liffick SL Bellini WJ Moraes JC Stevien KE Durigon EL Genetic homogeneity of measles viruses associated with a measles outbreak, Sao Paulo, Brazil, 1997 Emerg Infect Dis 2002 8 808 813 12141966
Taylor MJ Godfrey E Baczko K ter Meulen V Wild TF Rima BK Identification of several different lineages of measles virus J Gen Virol 1991 72 83 88 1990069
Last JM A Dictionary of Epidemiology 2001 4th New York, Oxford University Press
Mbugua FM Okoth FA Gray M Kamau T Kalu A Eggers R Borus P Kombich J Langat A Maritim P Lesiamon J Tipples GA Molecular epidemiology of measles virus in Kenya J Med Virol 2003 71 599 604 14556275 10.1002/jmv.10515
Horm SV Dumas C Svay S Feldon K Reynes JM Genetic characterization of wild-type measles viruses in Cambodia Virus Res 2003 97 31 37 14550585 10.1016/S0168-1702(03)00219-3
Hanses F Truong AT Ammerlaan W Ikusika O Adu F Oyefolu AO Omilabu SA Muller CP Molecular epidemiology of Nigerian and Ghanaian measles virus isolates reveals a genotype circulating widely in western and central Africa J Gen Virol 1999 80 871 877 10211955
Mulders MN Nebie YK Fack F Kapitanyuk T Sanou O Valea DC Muyembe-Tamfum JJ Ammerlaan W Muller CP Limited diversity of measles field isolates after a national immunization day in Burkina Faso: progress from endemic to epidemic transmission? J Infect Dis 2003 187 S277 82 12721926 10.1086/368036
Ahmad K Measles epidemic sweeps through Afghanistan Lancet 2000 355 1439 10791537 10.1016/S0140-6736(05)74645-6
BBC News online Niger: About 700 measles cases reported in capital Niamey.
Deutsche Presse Agentur Seven Children Die from Measles In Southern Philippine Village
Nigatu W Jin L Cohen BJ Nokes DJ Etana M Cutts FT Brown DW Measles virus strains circulating in Ethiopia in 1998-1999: Molecular characterisation using oral fluid samples and identification of a new genotype J Med Virol 2001 65 373 380 11536247 10.1002/jmv.2044
De Swart RL Nur Y Abdallah A Kruining H El Mubarak HS Ibrahim SA Van Den Hoogen B Groen J Osterhaus AD Combination of Reverse Transcriptase PCR Analysis and Immunoglobulin M Detection on Filter Paper Blood Samples Allows Diagnostic and Epidemiological Studies of Measles J Clin Microbiol 2001 39 270 273 11136782 10.1128/JCM.39.1.270-273.2001
Chibo D Riddell MA Catton MG Birch CJ Applicability of Oral Fluid Collected onto Filter Paper for Detection and Genetic Characterization of Measles Virus Strains J Clin Microbiol 2005 43 3145 3149 16000427 10.1128/JCM.43.7.3145-3149.2005
Rota JS Heath JL Rota PA King GE Celma ML Carabana J Fernandez-Munoz R Brown D Jin L Bellini WJ Molecular epidemiology of measles virus: identification of pathways of transmission and implications for measles elimination J Infect Dis 1996 173 32 37 8537679
|
16303052
|
PMC1318492
|
CC BY
|
2021-01-04 16:38:57
|
no
|
Virol J. 2005 Nov 22; 2:87
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-87
|
oa_comm
|
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-1241630955310.1186/1471-2458-5-124Research ArticleCardiovascular health knowledge and behavior in patient attendants at four tertiary care hospitals in Pakistan – a cause for concern Jafary Fahim H [email protected] Fawad [email protected] Hussain [email protected] Abdul [email protected] Murtaza [email protected] Atif [email protected] Mohammad A [email protected] Javed [email protected] Iqbal S [email protected] Irshad U [email protected] Department of Medicine, Section of Cardiology, Aga Khan University Hospital, Karachi, Pakistan2 Allied General Hospital, Faisalabad, Pakistan3 Mayo Hospital, Lahore, Pakistan4 Pakistan Institute of Medical Sciences, Islamabad, Pakistan2005 25 11 2005 5 124 124 8 5 2005 25 11 2005 Copyright © 2005 Jafary et al; licensee BioMed Central Ltd.2005Jafary 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
Knowledge about coronary heart disease (CHD) and its risk factors is an important pre-requisite for an individual to implement behavioral changes leading towards CHD prevention. There is scant data on the status of knowledge about CHD in the general population of Pakistan. The objective of this study was to assess knowledge of CHD in a broad Pakistani population and identify the factors associated with knowledge.
Methods
Cross sectional study was carried out at four tertiary care hospitals in Pakistan using convenience sampling. Standard questionnaire was used to interview 792 patient attendants (persons accompanying patients). Knowledge was computed as a continuous variable based on correct answers to fifteen questions. Multivariable linear regression was conducted to determine the factors independently associated with knowledge.
Results
The mean age was 38.1 (±13) years. 27.1% had received no formal education. The median knowledge score was 3.0 out of a possible maximum of 15. Only 14% were able to correctly describe CHD as a condition involving limitation in blood flow to the heart. Majority of respondents could identify only up to two risk factors for CHD. Most commonly identified risk factors were stress (43.4%), dietary fat (39.1%), smoking (31.9%) and lack of exercise (17.4%). About 20% were not able to identify even a single risk factor for CHD. Factors significantly associated with knowledge included age (p = 0.023), income (p < 0.001), education level (p < 0.001), residence (p < 0.001), a family history of CHD (p < 0.001) and a past history of diabetes (p = 0.004). Preventive practices were significantly lacking; 35%, 65.3% and 84.6% had never undergone assessment of blood pressure, glucose or cholesterol respectively. Only a minority felt that they would modify their diet, stop smoking or start exercising if a family member was to develop CHD.
Conclusion
This is the first study assessing the state of CHD knowledge in a relatively diverse non-patient population in Pakistan. There are striking gaps in knowledge about CHD, its risk factors and symptoms. These translate to inadequate preventive behavior patterns. Educational programs are urgently required to improve the level of understanding of CHD in the Pakistani population.
==== Body
Background
Coronary heart disease (CHD) is a major public health concern and accounts for more deaths than any other disease [1] worldwide. Developing countries were once considered to be less affected by CHD. However, the burden of CHD in developing countries has increased to epidemic proportions [2]. Furthermore, it is anticipated that within the next fifteen years, CHD will be the leading cause of death in these countries [1]. The prevalence of cardiovascular risk factors in the Asian population is high [3]. According to data from the National Health Survey of Pakistan, the prevalence of hypertension and diabetes is approximately 33% and 25% respectively, in persons over the age of 45 years [4,5] while the overall prevalence of smoking is 28% in men, going as high as 41% amongst those aged 40 to 49 years [6]. Furthermore dietary habits in Pakistan are heavily weighed towards the consumption of saturated fats as well as ghee (hydrogenated vegetable oil) [7]. Knowledge about CVD and its risk factors is an important (albeit not the only) pre-requisite for an individual to implement behavioral changes leading towards CHD prevention. Estimating the knowledge base of the community regarding CHD has important public health applications as it assists in developing targeted educational programs.
Studies in South Asians (Indians, Pakistanis and Bangladeshis) suggest a very poor degree of knowledge regarding CHD and its risk factors. In the study by Rankin [8] only a minority (approximately 20%) identified lack of exercise and obesity as a risk factor, while about 30% identified smoking as a factor related to CHD (amongst Pakistanis, this estimate was 14%) and approximately 30% considered dietary fat as a risk factor. Interestingly, only 5% of our respondents mentioned family history of CHD as a risk factor. There is scant data on the level of knowledge about CHD in the population of Pakistan. One prior study [9] has reported on the very poor awareness about risk factors for CHD in a lower middle-class urban population in Karachi, a major metropolitan city of Pakistan. In this selected population, consisting predominantly of women with a variable degree of formal education, less than 20% were aware of most risk factors for CHD. However, there is paucity of data on the status of knowledge in the broader population of the country, which consists of a much higher proportion of people with lower incomes, rural residency and no formal education. The aim of our study was to assess knowledge of CHD in a more diverse Pakistani population and identify the factors associated with knowledge in this study group. We also aimed to identify attitudes and behaviors regarding CHD prevention.
Methods
Study design
This was a multi-center cross sectional study conducted at four tertiary care hospitals in Pakistan. A convenience sample of 810 subjects consisting of attendants of patients (defined as persons accompanying patients to the hospital, usually family members or close relatives) visiting various outpatient departments of these four hospitals were invited to undergo an interview regarding their understanding of what coronary heart disease (CHD) was, risk factors, symptoms and practices related to CHD by trained medical officers. The four hospitals were chosen to enable the study population to pertain to a wide demographic spectrum of Pakistan. These hospitals were as follows: the Aga Khan University Hospital (AKUH) in Karachi is a private hospital that caters to a mixture of middle and upper class patients who (mostly) pay for their care. The Mayo Hospital in Lahore is a public hospital that caters primarily to poor patients who receive free care. The Pakistan Institute of Medical Sciences (PIMS) in a public hospital in Islamabad that caters to patients from a wide area that includes rural as well as urban areas and provides subsidized care. Finally, the Allied General Hospital (AGH) in Faisalabad is a large public hospital in Faisalabad that cares for urban as well as rural patients from central Punjab at nominal costs. The study was approved by each hospital's ethics review body.
Survey instrument
Study subjects were surveyed using a structured questionnaire that was developed to contain questions on four basic themes: (1) understanding of what CHD was; (2) knowledge of risk factors for CHD – respondents were asked what factors are related with CHD and the direction of the relationship between these factors and CHD was specified; (3) knowledge of the symptoms of CHD; and (4) preventive practices relating to CHD – whether subjects thought CHD was preventable and what preventive practices had they undertaken. The questionnaire was initially developed in English, translated into Urdu, then back-translated into English to check for consistency. The questionnaire, consisting of open-ended questions only, was pre-tested in 20 attendants and revised accordingly. All respondents were surveyed in an identical fashion. All questions were asked in Urdu. Responses to open-ended questions were recorded in English either by selection from a list of predefined answer categories or in writing if the particular response was not listed. Subjects were asked, at each question, whether they wished to add anything to their response, but were not prompted or given hints.
Study subjects
Patient attendants were selected as somewhat reflective of the general population for convenience reasons. Subjects were approached in the outpatients departments of the respective hospitals. If over the age of 18 years and able to give verbal consent, respondents underwent a standard questionnaire based survey. Subjects who were patients, under the age of 18 years, unable to communicate in Urdu (the national language of Pakistan) or were members of the hospital staff were excluded from this study. We estimated that a sample size of 371 would be required to estimate an assumed prevalence of 0.5 of various aspects knowledge about CHD at a confidence level of 95% with an error bound of 0.05.
Definitions
Knowledge was computed as a continuous variable using the cumulative score for each subject based on correct responses to 15 questions (Table 1). Each correct response was assigned one point. Urban dwelling was defined as residence within the geographical bounds of a major metropolitan city (Karachi, Islamabad, Rawalpindi, Faisalabad and Lahore). Rural dwelling was defined as residing outside the major metropolitan cities (most of Pakistan is fairly rural outside the major cities). Formal education was defined as education received in school (but did not include those going to traditional educational institutions or "madrassa").
Table 1 Questions on which coronary heart disease knowledge score was computed.
Question Response Score
What is CHD? Blocked artery 1
Risk factors for CHD Smoking 1
No exercise 1
Dietary fat 1
Stress 1
High blood pressure 1
Diabetes 1
Family history 1
Age 1
Male gender 1
Obesity 1
Cholesterol 1
Symptoms of CHD Chest pain 1
Dyspnea 1
Sweatiness 1
Maximum Score 15
CHD = coronary heart disease.
Statistical analysis
All variables were entered into Statistical Package for Social Sciences (SPSS) version 10. Means and standard deviations were calculated for continuous variables and frequencies for categorical variables. Univariate analysis was performed by simple linear regression to determine the factors associated with knowledge. Univariate covariates with a p value of ≤0.25 were entered into the multivariable model. Multivariable regression using a stepwise technique was conducted to adjust for confounders and determine the factors independently associated with knowledge. The standardized residuals of the regression model were examined for outliers as well as for non-normality and nonlinearity using probability plots and residual plots. A significant level was defined as p < 0.05.
Results
Approximately 1000 potential respondents were approached. Eight hundred and ten subjects consented to participate in the study. Complete data were available for 792 subjects. The demographic and clinical characteristics of the study group are shown in Table 2.
Table 2 Demographics and Clinical Characteristics of Study Population.
Characteristic N (%)
Hospital
AKU 153 (19.3)
PIMS 249 (31.4)
Mayo 241 (30.4)
AGH 149 (18.8)
Gender
Male 558 (70.5)
Female 234 (29.5)
Marital Status
Single† 202 (25.5)
Married 590 (74.5)
Residence
Rural 335 (42.3)
Urban 457 (57.7)
Age‡ (mean ± SD) 38.1 ± 13
Age‡ (categories)
≤30 277 (35.0)
31 – 60 469 (59.2)
>60 46 (5.8)
Smokers 314 (39.6)
Monthly income*
<3000 225 (28.4)
>3000 567 (71.6)
Education
No formal 215 (27.1)
Formal 577 (72.9)
Family Hx of IHD 373 (47.1)
Prior Diabetes 47 (5.9)
Prior Hypertension 118 (14.9)
Prior IHD 42 (5.3)
N = number. † includes separated and divorced. ‡ in years.
AKU = Aga Khan University Hospital. PIMS = Pakistan Institute of Medical Sciences.
Mayo = Mayo Hospital. AGH = Allied General Hospital. SD = standard deviation.
* in Rupees (1 Rupee = US$ 0.017). Hx = history. IHD = ischemic heart disease.
Subjects were asked about what they thought CHD meant. Only 14% were able to correctly describe the latter as a condition involving limitation in blood flow to the heart. A variety of other descriptions were offered including a "malfunction of the heart" by 35%. Table 3 displays the distribution of responses to this question.
Table 3 Responses to questions assessing knowledge of coronary heart disease
Response count (% of cases)
What is Coronary Heart Disease?*
Arterial blockage 109 (13.8)
Do not know 232 (29.3)
Chest pain 152 (19.2)
"Malfunction" of the heart 277 (35.0)
Valve problem 83 (10.5)
Other 142 (17.9)
What are the risk factors for Coronary Heart Disease?*
Do not know 162 (20.5)
Smoking 253 (31.9)
No exercise 138 (17.4)
Dietary fat 310 (39.1)
Dietary salt 26 (3.30)
Stress 344 (43.4)
High blood pressure 56 (7.10)
Cholesterol 59 (7.40)
Diabetes 15 (1.90)
Family history 39 (4.90)
Age 33 (4.20)
Male gender 9 (1.10)
Overwork 36 (4.50)
Obesity 112 (14.1)
Other 62 (7.80)
What are the symptoms of Coronary Heart Disease?*
Do not know 230 (29.0)
Chest pain 287 (36.2)
"Attack" 179 (22.6)
Sweatiness 116 (14.6)
Palpitations 136 (17.2)
Anxiety 177 (22.3)
Headache 17 (2.10)
Dyspnea 193 (24.4)
Other 44 (5.60)
* multiple responses
Participants were questioned about risk factors for CHD. As seen in Table 3, the most commonly identified risk factors were stress (43.4%), dietary fat (39.1%), smoking (31.9%) and lack of exercise (17.4%). However about 20% were not able to identify even a single risk factor for CHD (Figure 1). When questioned about the symptoms of CHD, only a minority were able to correctly identify symptoms of chest pain, dyspnea and diaphoresis (Table 3).
Figure 1 Risk factors for coronary heart disease identified by study subjects.
Subjects were questioned regarding preventive practices against CHD. Although nearly three quarters of the study group felt that CHD was preventable (Table 4), preventive practices in this study group were largely lacking.
Table 4 Preventive Practices and Attitudes towards coronary heart disease
Response count (% of cases)
Is CHD preventable?
Yes 604 (74.3)
No 37 (4.70)
Do not know 151 (19.1)
Have you had your blood pressure checked?
Never have 275 (34.7)
Every 2 years 298 (37.6)
>2 year intervals 219 (27.7)
Have you had your blood sugar checked?
Never have 517 (65.3)
Have at least once 275 (34.7)
Have you had your cholesterol checked?
Never have 670 (84.6)
Have at least once 122 (15.4)
Have you ever undertaken any preventive practices for CHD*?
None 354 (44.7)
Exercise 185 (23.4)
Dietary salt restriction 39 (4.90)
Dietary fat restriction 265 (33.5)
Weight control 38 (4.80)
Stress reduction 58 (7.30)
Reduced smoking 89 (11.2)
Medications 16 (2.00)
Home remedies 25 (3.20)
Other 34 (4.30)
If your family member develops CHD, what would you do*?
Nothing 72 (11.4)
Do not know 104 (16.5)
See a doctor 332 (52.7)
Change my diet 132 (21.0)
Start exercising 50 (7.90)
Stop smoking 40 (6.30)
Reduce stress 48 (7.60)
Other 21 (2.30)
CHD = coronary heart disease.
* multiple responses
Knowledge was assessed in this study as a continuous variable. Responses to 15 questions were scored as correct or incorrect and each respondent received one point for each correct answer (Table 1). The median knowledge score of the study group was 3.0 (range 0–11) out of a possible maximum of 15. Figure 2 shows the distribution of knowledge scores in the study group. No participant was able to achieve a score above 11.
Figure 2 Knowledge score for coronary heart disease.
Table 5 shows the results of the univariate and multivariable analyses for factors associated with knowledge of CHD, with the respective coefficients (slope of the regression line). In the univariate analysis, factors significantly associated with knowledge included age, gender, education (formal vs. no formal), monthly income (<3000 vs. >3000 Rupees), residence (rural vs. urban), smoking status, family history of CHD and a prior history of diabetes, hypertension or CHD (Table 5). Interestingly, there were also significant differences in the knowledge scores of subjects recruited from different hospitals. In the multivariable analysis factors independently associated with knowledge included age, education, residence, monthly income, family history of CHD, prior history of diabetes and hospital attended. Not unexpectedly, education had the most robust association with knowledge. In the adjusted analysis, gender, smoking and prior history of CHD and hypertension were not associated with knowledge.
Table 5 Factors associated with knowledge of coronary heart disease – univariate and multivariable analysis
N Mean Knowledge Score Unadjusted Coefficient8 (95% CI) p value Adjusted Coefficient8 (95% CI) p value
Age 0.01 (0.001, 0.03) ‡ 0.035 0.01 (0.002, 0.03) ‡ 0.023
Gender 0.008 0.20
Female 234 3.54 ref ref
Male 558 3.04 -0.50 (-0.87, -0.13) -0.22 (-0.56, 0.12)
Marital Status 0.64
Single 202 3.25 ref
Married 590 3.16 -0.09 (-0.48, 0.30) NA*
Education <0.001 <0.001
No Formal 215 1.80 ref ref
Formal Education 577 3.70 1.89 (1.54, 2.25) 1.42 (1.06, 1.77)
Monthly income † <0.001 <0.001
<3000 225 1.88 ref ref
>3000 567 3.70 1.82 (1.47, 2.17) 0.76 (0.40, 1.13)
Residence <0.001 <0.001
Rural 335 2.21 ref ref
Urban 457 3.90 1.69 (1.37, 2.01) 1.06 (0.75, 1.37)
Smoking Status <0.001 0.09
Non smoker 478 3.47 ref ref
Smoker 314 2.75 -0.728 (-1.07, -0.39) -0.27 (-0.57, 0.43)
Family Hx IHD <0.001 <0.001
No 419 2.68 ref ref
Yes 373 3.75 1.07 (0.74, 1.40) 0.61 (0.31, 0.91)
PMH – Hypertension <0.001 0.86
No 674 3.05 ref ref
Yes 118 3.94 0.89 (0.42, 1.36) 0.04 (-0.41, 0.49)
PMH – Diabetes <0.001 0.004
No 745 3.08 ref ref
Yes 47 4.85 1.78 (1.07, 2.48) 0.96 (0.30, 1.61)
PMH – IHD 0.004 0.28
No 750 3.13 ref ref
Yes 42 4.21 1.09 (0.34, 1.84) 0.37 (-0.31, 1.05)
Hospital
Mayo 241 2.29 ref ref
AKU 153 3.55 1.25 (0.78, 1.73) <0.001 0.15 (-0.29, 0.59) 0.50
PIMS 249 3.45 1.16 (0.74, 1.57) <0.001 0.42 (0.03, 0.81) 0.03
AGH 149 3.81 1.51 (1.03, 1.99) <0.001 1.20 (0.76, 1.64) <0.001
8coefficient = slope of regression line. Values represent change in knowledge (per unit increase in independent variable or with category change compared to reference category)
CI = confidence interval ‡ for each year increase in age ref = reference category
Hx = history PMH = past medical history IHD = ischemic heart disease
* not in final model † in Rupees (1 Rupee = US$ 0.017)
AKU = Aga Khan University Hospital, Karachi. Mayo = Mayo Hospital, Lahore. PIMS = Pakistan Institute of Medical Sciences, Islamabad. AGH = Allied General Hospital, Faisalabad.
Discussion
Our study demonstrates a striking lack of knowledge about CHD amongst patient attendants at four hospitals catering to patients belonging to a wide demographic spectrum. Only 14% of the study group could correctly identify what coronary heart disease was. The majority of subjects were only able to identify up to 2 risk factors, the most commonly identified factors being stress, obesity, dietary fat and smoking. The median knowledge score of the study group was 3.0 out of a possible 15, suggesting a severe lack of awareness about CHD.
Our findings are consistent with the study by Rankin et al. [8]. The gaps in knowledge are considerably high compared to what has been reported in Western countries [10].
The relatively poor knowledge about modifiable risk factors for CHD, including smoking (31.9%), obesity (14.1%), lack of exercise (17.4%) and dietary fat (39.1%) in our study population has been seen in other studies on South Asians [8]. It is possible that the very poor awareness about obesity and lack of exercise are related to their under representation in mass media campaigns as opposed to smoking and dietary fat. Our study highlights a significant lack of knowledge about modifiable risk factors and suggests the need for urgent emphasis on education amongst Pakistanis.
Respondents in our study appeared to lack sufficient knowledge about the symptoms of CHD. Only 36% knew that chest pain was a symptom of heart disease and a substantial number of subjects gave vague descriptions about possible symptoms of CHD (Table 3). Failure to recognize the symptoms of acute myocardial infarction is associated with delay in seeking medical care [11] and leads to worse clinical outcomes. Our study suggests a need for educating the public about symptoms of CHD.
Age appears to have a linear relationship with knowledge in our study. When age was analyzed as a categorical variable, mean knowledge scores progressively increased from the ≤30 year age group (3.06) to the >60 year group (3.93; coefficient 0.76; p = 0.03). This is different to the findings of Potvin [10] in Canadians where, compared to the age group 18–24 years, the odds ratios for knowledge about risk factors decreased as aged increased, particularly in those aged 65–74 years. One possible explanation could be a very limited emphasis on health education in schools in Pakistan, as a result of which individuals finishing school do so with very little knowledge about CHD. As the years go by, they are likely to accrue knowledge (albeit in small increments). On the other hand, modern school education in the west may involve substantially more time spent on health education; as a consequence younger individuals are likely to be more knowledgeable than the elderly who may not have received such instruction during their schooling years. The finding that knowledge increases with the acquisition of formal education is consistent with other data [10,12]. Likewise, the association of income with knowledge has been reported in other studies [10]. In our study this association was independent of the level of education.
There were interesting differences in knowledge score between subjects recruited from the four different hospitals. The mean knowledge score was significantly higher in patients presenting to AGH, a hospital that caters to a substantially poorer, uneducated and rural population compared to AKU and PIMS (Table 6). AGH conducts a large number of public awareness seminars and it is possible that this may be contributing to the higher knowledge scores in patients recruited from that hospital. This potential association between educational seminars for the public and knowledge deserves further investigation.
Table 6 Key demographic characteristics of study subjects interviewed from four hospitals in Pakistan.
AKU N (%) Mayo N (%) PIMS N (%) AGH N (%)
Mean age* (SD) 40.0 (15.9) 36.1 (11.4) 40.4 (12.5) 35.7 (12.1)
Income†
<3000 23 (15.0) 123 (51.0) 33 (13.3) 46 (30.9)
>3000 130 (85.0) 118 (49.0) 216 (86.7) 103 (69.1)
Education
No formal 27 (17.6) 84 (34.9) 47 (18.9) 57 (38.3)
Formal Education 126 (82.4) 157 (65.1) 202 (81.1) 92 (61.7)
Residence
Rural 40 (26.1) 126 (52.3) 89 (35.7) 80 (53.7)
Urban 113 (73.9) 115 (47.7) 160 (64.3) 69 (46.3)
AKU = Aga Khan University Hospital, Karachi. Mayo = Mayo Hospital, Lahore. PIMS = Pakistan Institute of Medical Sciences, Islamabad. AGH = Allied General Hospital, Faisalabad.
* in years; SD = standard deviation
†Self reported figures (per month in Rupees); estimates may be biased (1 Rupee = US$ 0.017).
Consistent with the overall poor state of knowledge in our study group, preventive practices were found to be significantly lacking. Although nearly three-fourths of the study group felt that CHD was preventable, over two-thirds had never had their blood pressure checked, and an overwhelming majority had never undergone an assessment of their blood glucose or cholesterol levels. Furthermore, only a minority felt that they would modify their diet, stop smoking or start exercising if a family member was to develop CHD (table 4), consistent with the fact that the vast majority of our study subjects did not identify a family history of CHD as a risk factor for developing CHD in the future.
Our study is not without limitations. First, the non-probability convenience sampling method introduces selection bias. As samples were taken from four different hospitals in Pakistan representing a wide demographic spectrum, we feel that our study population may be somewhat reflective of the knowledge state of the general population of Pakistan. However, patient attendants may be either less knowledgeable or more informed than the general population. On the one hand, due to more exposure to a medical environment patient attendants may know more about CHD than the general public and therefore, the true knowledge state of the public may be lower. On the other hand, diseased individuals may, in part, be in that state due to the poor knowledge of their household members. Hence, selection of subjects from households of patients risks bias whereby the study subjects may be less knowledgeable than population controls. Second, the use of an open-ended questionnaire introduces the potential of recall bias on the part of the respondents and may underestimate the knowledge state of the study group. Third, computation of a knowledge score based on correct answers to a set of questions is somewhat arbitrary, does not incorporate differential weightage that may be placed on different questions and has not been widely validated. Nevertheless, we feel that this score provides a fair estimate of the degree of knowledge of an individual. Fourth, the majority of study subjects consisted of men. This is partly related to the fact that women in Pakistan are less likely to accompany patients to the hospital; furthermore, when women were approached, if they were accompanied by other male members of the family, the latter would take the lead in answering questions. Thus women are under-represented in this study. Finally, the specialty of the specific outpatient clinic from where respondents were recruited was not recorded and, therefore, not adjusted for in the analysis.
Conclusion
In conclusion, this is the first study that has attempted to assess the state of CHD knowledge in a relatively diverse non-patient population in Pakistan. We found striking gaps in knowledge, particularly about the understanding of the nature CHD and its risk factors (physiologic factors like high blood pressure and elevated blood cholesterol in particular). In addition there was an equally significant lack of knowledge about common symptoms of CHD. These deficiencies in knowledge appear to translate into inadequate preventive behavior patterns. Factors significantly associated with knowledge included age, income, education level, residence, a family history of CHD and a past history of diabetes. An interesting association between knowledge and site of recruitment was noted in this study group which may be related to the number of public awareness seminars conducted by the site. This association deserves further investigation. Our study, despite some limitations, should raise strong concerns about the lack of knowledge and awareness about CHD amongst lay persons in Pakistan and should serve as a stimulus for establishing health education programs in the country.
List of abbreviations
AKU = Aga Khan University Hospital
PIMS = Pakistan Institute of Medical Sciences
Mayo = Mayo Hospital
AGH = Allied General Hospital
SD = standard deviation.
CHD = coronary heart disease.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
FHJ, FA, HM, AW, MS, AA conceptualized this study and participated in the study design and manuscript writing and review. FA, HM, AW, MS, AA, MAQ were involved in the data collection process. FHJ performed the statistical analysis. JA, ISK and IH supervised the data collection process and also participated in manuscript review. All authors have read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Lenfant C Can we prevent cardiovascular diseases in low- and middle-income countries? Bull World Health Organ 2001 79 980 2; discussion 983-7 11693981
Reddy KS Yusuf S Emerging epidemic of cardiovascular disease in developing countries Circulation 1998 97 596 601 9494031
Pinto RJ Risk factors for coronary heart disease in Asian Indians: clinical implications for prevention of coronary heart disease Indian Journal of Medical Sciences 1998 52 49 54 9770863
Jafar TH Levey AS Jafary FH White F Gul A Rahbar MH Khan AQ Hattersley A Schmid CH Chaturvedi N Ethnic subgroup differences in hypertension in Pakistan J Hypertens 2003 21 905 912 12714864 10.1097/00004872-200305000-00014
Jafar TH Jafary FH S. J Chaturvedi N Heart Disease Epidemic in Pakistan: Women and men at equal risk. Am Heart J 2005 150. 221 226 16086922 10.1016/j.ahj.2004.09.025
Ahmad K Jafary FH Jehan I Hatcher J Qayum A Chaturvedi N Jafar TH Prevalence and predictors of smoking in Pakistan: Results of the National Health Survey of Pakistan Eur J Cardiovasc Prev Rehabil 2005 12 203 208 15942416 10.1097/00149831-200506000-00003
Ismail J Jafar TH Jafary FH White F Faruqui AM Chaturvedi N Risk factors for non-fatal myocardial infarction in young South Asian adults Heart 2004 90 259 263 14966040 10.1136/hrt.2003.013631
Rankin J Bhopal R Understanding of heart disease and diabetes in a South Asian community: cross-sectional study testing the 'snowball' sample method Public Health 2001 115 253 260 11464296
Dodani S Mistry R Farooqi M Khwaja A Qureshi R Kazmi K Prevalence and awareness of risk factors and behaviours of coronary heart disease in an urban population of Karachi, the largest city of Pakistan: a community survey. J Public Health (Oxf) 2004 26 245 249 15454591 10.1093/pubmed/fdh154
Potvin L Richard L Edwards AC Knowledge of cardiovascular disease risk factors among the Canadian population: relationships with indicators of socioeconomic status Cmaj 2000 162 S5 11 10813022
McKinley S Dracup K Moser DK Ball K Yamasaki K Kim CJ Barnett M International comparison of factors associated with delay in presentation for AMI treatment. Eur J Cardiovasc Nurs 2004 3 225 230 15350232 10.1016/j.ejcnurse.2004.06.004
Shea S Stein AD Basch CE Lantigua R Maylahn C Strogatz DS Novick L Independent associations of educational attainment and ethnicity with behavioral risk factors for cardiovascular disease Am J Epidemiol 1991 134 567 582 1951262
|
16309553
|
PMC1318493
|
CC BY
|
2021-01-04 16:28:58
|
no
|
BMC Public Health. 2005 Nov 25; 5:124
|
utf-8
|
BMC Public Health
| 2,005 |
10.1186/1471-2458-5-124
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1761633664410.1186/1471-2164-6-176Methodology ArticleASAP: Amplification, sequencing & annotation of plastomes Dhingra Amit [email protected] Kevin M [email protected] Horticultural Sciences Department, University of Florida, Gainesville, FL 32611, USA2005 7 12 2005 6 176 176 2 8 2005 7 12 2005 Copyright © 2005 Dhingra and Folta; licensee BioMed Central Ltd.2005Dhingra and Folta; 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
Availability of DNA sequence information is vital for pursuing structural, functional and comparative genomics studies in plastids. Traditionally, the first step in mining the valuable information within a chloroplast genome requires sequencing a chloroplast plasmid library or BAC clones. These activities involve complicated preparatory procedures like chloroplast DNA isolation or identification of the appropriate BAC clones to be sequenced. Rolling circle amplification (RCA) is being used currently to amplify the chloroplast genome from purified chloroplast DNA and the resulting products are sheared and cloned prior to sequencing. Herein we present a universal high-throughput, rapid PCR-based technique to amplify, sequence and assemble plastid genome sequence from diverse species in a short time and at reasonable cost from total plant DNA, using the large inverted repeat region from strawberry and peach as proof of concept. The method exploits the highly conserved coding regions or intergenic regions of plastid genes. Using an informatics approach, chloroplast DNA sequence information from 5 available eudicot plastomes was aligned to identify the most conserved regions. Cognate primer pairs were then designed to generate ~1 – 1.2 kb overlapping amplicons from the inverted repeat region in 14 diverse genera.
Results
100% coverage of the inverted repeat region was obtained from Arabidopsis, tobacco, orange, strawberry, peach, lettuce, tomato and Amaranthus. Over 80% coverage was obtained from distant species, including Ginkgo, loblolly pine and Equisetum. Sequence from the inverted repeat region of strawberry and peach plastome was obtained, annotated and analyzed. Additionally, a polymorphic region identified from gel electrophoresis was sequenced from tomato and Amaranthus. Sequence analysis revealed large deletions in these species relative to tobacco plastome thus exhibiting the utility of this method for structural and comparative genomics studies.
Conclusion
This simple, inexpensive method now allows immediate access to plastid sequence, increasing experimental throughput and serving generally as a universal platform for plastid genome characterization. The method applies well to whole genome studies and speeds assessment of variability across species, making it a useful tool in plastid structural genomics.
==== Body
Background
Chloroplast DNA (cpDNA) represents an extranuclear capsule of genetic information, encoding essential structural and enzymatic proteins of the organelle. This satellite genome contains a wealth of information that has been shaped by speciation, rendering it a rich resource to trace evolutionary relationships between photosynthetic taxa [1]. Genetic manipulation of the chloroplast genome can transform the chloroplast into a bioreactor, allowing large-scale production of proteins vital to agriculture or pharmacology [2,3]. Maternal inheritance of plastid genome in most species ensures gene containment in genetically modified plants making it an attractive alternative for the integration of foreign genes [4,5].
Genomic and phylogenetic studies and efficient genetic modification begin with the base material of plastid DNA sequence. Despite their relatively small size, few plastid genomes have been fully sequenced, thus limiting comparative genomics studies across the species. Complete sequence coverage has been resolved for only 13 species representing model and crop plants, namely, Arabidopsis thaliana [6], Atropa belladonna [7], Medicago truncatula (GenBank: NC_003119), Oryza sativa [8], Spinacea oleracea [9], Nicotiana tabacum [10], Triticum aestivum [11], Lotus corniculatus [12], Zea mays [13], Panax ginseng [14], Cucumis sativus (GenBank: NC_007144), Glycine max [15] and Saccharum officinarum [16]. Comparison of a small set of representative coding or intergenic sequences derived from a large number of species has been used to perform phylogenetic studies [17], but there are many unresolved phylogenetic questions [18]. Use of complete genome data is another emerging approach in plant phylogenetics [19]. There are at least three federally-sponsored, multi-institution endeavors underway to sequence about 200 plastid genomes from plants [20-22].
There are a number of challenges to rapid access to chloroplast DNA sequence from a given species. Typical sequencing efforts begin with construction of chloroplast plasmid or other genomic libraries. Construction of such resources requires isolation of pure plastid DNA, which may be troublesome in some species. Even shotgun sequencing to appreciable coverage is considerably expensive and time consuming. From the standpoint of method and cost, the generation of plastid sequence data is considerable and a potential hindrance to productive data mining and engineering efforts. One recent report describes a sophisticated methodology using FACS (fluorescence-assisted cell sorting) and RCA (rolling circle amplification) for sequencing a plastid genome [23]. The degree of technological sophistication is inversely proportional to its wider applicability due to the prohibitive costs associated with it, slowing chloroplast genomics studies. These large-scale studies are expensive and prohibit testing of additional related species or ecotypes that may be informative. Additionally, smaller research programs do not necessarily have access to these tools and techniques. Most importantly some rare and/or difficult-to-obtain taxa that are not amenable to large-scale chloroplast DNA extraction can not be analyzed by the existing methodology. While downstream processing of sequence information is highly streamlined due to the presence of freely available tools on the World Wide Web, such as DOGMA [24,25], the lack of cost-effective innovation in rapid sequence acquisition has restricted plastid informatics studies.
To address these issues, we exploited the fact that chloroplast genomes are extremely well conserved in size, gene arrangement, and coding sequence, at least within major subgroups of the plant kingdom [26,27]. We formulated the hypothesis that conserved islands of cpDNA sequence could serve as universal anchors to generate overlapping PCR products comprised of conserved coding regions, and adjacent polymorphic intergenic regions. The resulting amplicons could then be sequenced, assembled, annotated and analyzed. This approach exploits the high degree of sequence conservation and general synteny within discrete portions of chloroplast genome. In this report this powerful technique has been applied to the large inverted repeat (IR) region of strawberry (Fragaria × ananassa) and peach (Prunus persica). The entire ~30 kb region was amplified from total DNA, sequenced, annotated and submitted to public databases in several days for a fraction of the cost of traditional or other recently published approaches [28]. To further validate this application, corresponding regions were amplified from a series of other eudicots and a monocot of agricultural importance as well as two gymnosperms (Pinus and the distant vascular plant Gingko) and a pteridophyte (Equisetum). This universal method represents a rapid, inexpensive means to obtain complete coverage of many higher plant plastid genome regions, and even substantial coverage from distant genera. The sequence information generated form this method can hasten phylogenetic and genomics studies and also help in identification of regulatory elements necessary for design of transformation vectors for the manipulation of chloroplast genomes of new species.
Results
In order to expand ongoing Rosaceae genomics studies [29], the original goal of this work was to use existing informatics resources to devise a PCR-based strategy to obtain plastid DNA sequence for cultivated strawberry and peach. This information would assist in identifying indel (insertion/deletion) polymorphisms or SNPs (single nucleotide polymorphisms) that could serve as an additional tool for phylogenetic analysis [30] and also allow the design of vectors useful for strawberry plastid engineering. A schematic explanation of the technique is shown in Figure 1. If a given primer pair fails to generate an amplicon in initial PCR trials, the forward primer can then be paired with the reverse primer from the next primer pair to obtain coverage of that region. Figure 2 demonstrates proof-of-concept, as universal primer set derived from the IR of five sequenced eudicots is sufficient to amplify the corresponding ~30 Kb region in commercial strawberry. The 27 primer pairs generate amplicons spanning this region. The corresponding PCR products were sequenced, and the sequence was immediately deposited to public databases. Here we proceeded from computational analyses to finished strawberry and peach IR sequence in one week for <$500.
Figure 1 Schematic representation of the ASAP approach. Three rounds of PCR allow for 100% coverage of a given region. F and R suffix to the numbers represent forward and reverse universal primers.
Figure 2 ASAP profile from Fragaria × ananassa. A. Round 1 touchdown PCR. B. Round 2 touchdown PCR and C. Round 3 touchdown extension PCR.
Since the method proved useful in strawberry and peach, its applicability across plant species was assessed. Total genomic DNA was derived from 13 diverse plant species and subjected to the ASAP protocol using the primers listed in Table 1 and the conditions stated in Table 2. The ASAP primer set effectively generated expected amplicons from all eudicot species tested. Expected results were obtained in Nicotiana and Arabidopsis. Complete coverage was obtained with the first round of PCR and the amplicon sizes were consistent with predictions (Table 1). Comparison of agarose gel electrophoresis profiles from the IR region revealed clearly discernible amplified fragment length polymorphisms (AFLPs) in the regions amplified with primer pairs 11, 17 and 27 (Figure 3). The profiles for these two species are in complete agreement with the calculated sizes.
Table 1 ASAP PCR primers. Primer sequences, annealing site and the relative position in tobacco (Nt), Arabidopsis (At) and maize (Zm) are presented, along with the anticipated amplicon size. * represent the primers with low sequence similarity in maize.
Primer Gene (Nicotiana) Sequence 5' to 3' Position in Nt Size Position in At Size Position in Zm Size
IRB1F Upstream of JLB ggatttttttttagtgaacgtgtcac 86657–86682 804 84258–84283 812 82644–82669* 1001
IRB1R Intron rpl2 aagtatcgacgtaatttcatagagtc 87436–87461 85045–85070 83622–83647
IRB2F Intron rpl2 catctggcttatgttcttcatgtagc 87397–87422 1048 85006–85031 1049 83583–83608 1043
IRB2R rpl23 3'end caactaggacagaaataaagcattgg 88420–88445 86030–86055 84601–84626
IRB3F rpl23 3'end atacgtctgtaatgcattgtatgtcc 88310–88335 1001 85920–85945 980 84491–84516
IRB3R YCF2/ORF2280 5' gaagatacaggagcgaaacaatcaac 89285–89311 86875–86900 Deleted
IRB4F YCF2/ORF2280 aagaaaaaatctctatttgatagggc 89181–89206 1031 86770–86795 1037 Deleted
IRB4R YCF2/ORF2280 tttcgttccgtttgaagaaaggaagg 90212–90187 87782–87807 Deleted
IRB5F YCF2/ORF2280 ggattccattagtaatgaggattcgg 90096–90121 1162 87685–87710 1171 Deleted
IRB5R YCF2/ORF2280 gaggctcgaaccatttcttctgactc 91233–91258 88831–88856 Deleted
IRB6F YCF2/ORF2280 cttcgaatatggaattcaaagggatc 91131–91156 1042 88729–88754 1081 Deleted
IRB6R YCF2/ORF2280 tgaatatgttagatacctgtgactcg 92148–92173 89785–89810 Deleted
IRB7F YCF2/ORF2280 acaattcctcaatatcttgttcattc 92049–92074 1091 89680–89705 1088 Deleted
IRB7R YCF2/ORF2280 tcttctagagaatctcctaattgttc 93115–93140 90743–90768 Deleted
IRB8F YCF2/ORF2280 gaaaaggtcaaatctttgatgattcc 92995–93020 1041 90623–90648 1035 Deleted
IRB8R YCF2/ORF2280 tttccggcatcatatccatagttagc 94011–94036 91633–91658 Deleted
IRB9F YCF2/ORF2280 ctgaacaagttcctggataacaagcc 93853–93878 1169 91493–91518 1151 Deleted
IRB9R YCF2/ORF2280 aaatctctgatcaaggatagaacaag 94997–95022 92619–92644 Deleted
IRB10F YCF2/ORF2280 gatctagttcatggcctattagaagt 94849–94874 1025 92471–92496 1021 86077–86102 1138
IRB10R ORF 87/YCF 15 taacatattcttccatggagctaagg 93849–95874 93475–93492 87190–87215*
IRB11F YCF2/ORF2280 3' cggatgaaatgaaaattggattcatg 95669–95724 1094 93330–93355 1319 No similarity
IRB11R ORF 79 aatcggacctgctttttacatatctc 96739–96763 94624–94649 No similarity
IRB12F ORF 79 ccaattgcttcgatttgaattatccg 96626–96644 1061 94483–94508 1101 88794–88819 1079
IRB12R ndhB 3' exon tggaaatcctagctattcttagcatg 97662–97687 95559–95584 89848–89873
IRB13F ndhB 3' exon attccaataattacatatccgatttg 97567–97592 1043 95464–95489 1049 89753–89778 1068
IRB13R ndhB 5' exon cttatcaatacacaaatgtataactc 98585–98610 96488–96513 90796–90821
IRB14F ndhB 5' exon tacgtcaggagtccattgatgagaag 98494–98519 1171 96397–96422 1206 90705–90730 1192
IRB14R rps 7 5'end aatatggctttcaaattaagttccga 99640–99665 97578–97603 91872–91897
IRB15F rps 7 5'end gtgcaaaagctctatttgcctctgcc 99551–99576 1230 97489–97514 1231 91783–91808 1245
IRB15R rps12 exon 2 tcactgcttatatacccggtattggc 100754–100781 98695–98720 93003–93028
IRB16F rps12 exon 2 tcctcgaacaatgtgatatctcacac 100699–100694 968 98608–98633 1079 92916–92941 859
IRB16R Spacer caacataggtcatcgaaaggatctcg 101642–101667 99662–99687 93750–93775
IRB17F Spacer gtgtgagcttatccatgcggttatgc 101554–101581 1116 99576–99601 1345 93669–93694 1339
IRB17R rrn16 start gcttcatattcgcccggagttcgctc 102645–102670 100896–100921 95043–95068*
IRB18F trnV 3' end aagtcatcagttcgagcctgattatc 102504–102529 1117 100751–100776 1122 94901–94926 1121
IRB18R rrn16 start tgagtttcattcttgcgaacgtactc 103596–103621 101848–101873 95997–96022
IRB19F rrn16 cgacactgacactgagagacgaaagc 103452–103477 1343 101704–101729 1341 95853–95878 1347
IRB19R trnI Intron/OriA atcgaaagttggatctacattggatc 104770–104795 103020–103045 97175–97200
IRB20F trnI start gggctattagctcagtggtagagcgc 104551–104576 877 102801–102826 898 96953–96978 1119
IRB20R trnA start caagagcggagctctaccaactgagc 105403–105428 103674–103699 98047–98072
IRB21F trnI end gaggtctctggttcaagtccaggatg 105229–105324 941 103571–103596 962 97943–97968 968
IRB21R trnA end ataagcggactcgaaccgctgacatc 106145–106170 104508–104533 98886–98911
IRB22F trnA intron agattttgagaagagttgctctttgg 106003–106028 1190 104367–104392 1188 98743–98768 1193
IRB22R 23S tagatgtccagtcaactgctgcgcct 107168–107193 105530–105555 99911–99936
IRB23F 23S gaaactaagtggaggtccgaaccgac 107053–107079 1225 105416–105441 1224 99797–99822 1291
IRB23R 23S cgctaccttaggaccgttatagttac 108253–108278 106615–106640 101063–101088
IRB24F 23S ggtctccgcaaagtcgtaagaccatg 108131–108156 1172 106493–106518 1169 100941–100966 1156
IRB24R 4.5S acatcactgcacttccacttgacacc 109278–109303 107637–107662 102072–102097
IRB25F 23S end ctgctgaaagcatctaagtagtaagc 109089–109115 963 107451–107476 929 101897–101922 920
IRB25R trnR end ggttgtgggcgaggagggattcgaac 110027–110052 108355–108380 102792–102817
IRB26F Spacer aaatggctggggagaggaaaggttcc 109905–109930 1182 108231–108256 1232 No similarity
IRB26R ORF350 attatcttcatgcataaggatactag 111062–111087 109438–109463 No similarity
IRB27F Spacer tggctctatttcattatattccatcc 110844–110869 1036 109225–109250 986 No similarity
IRB27R ORF 350 agtggatccctcttgttcctgtttag 111855–111880 110186–110211 No similarity
Table 2 ASAP PCR conditions. The thermalcycler parameters used to generate ASAP amplicons in successive rounds of PCR are presented.
ASAP PCR conditions
1. Touchdown PCR Temperature (Centigrade) Duration (Minutes)
Denaturation 94 4
10 Cycles
Denaturation 94 0:40
Annealing 55 – 0.5/cycle 0:40
Extension 72 0:40
25 Cycles
Denaturation 94 0:40
Annealing 50 0:40
Extension 72 0:40
Final extension 72 7
2. Touchdown PCR
Denaturation 94 4
10 Cycles
Denaturation 94 0:40
Annealing 52 – 0.5/cycle 0:40
Extension 72 0:40
25 Cycles
Denaturation 94 0:40
Annealing 47 0:40
Extension 72 0:40
Final extension 72 7
3. Touchdown – extension PCR
Denaturation 94 4
10 Cycles
Denaturation 94 0:40
Annealing 52 – 0.5/cycle 0:40
Extension 72 1:00 + .05/cycle
25 Cycles
Denaturation 94 0:40
Annealing 47 0:40
Extension 72 1:30 + 0.05/cycle
Final extension 72 10
Figure 3 Composite ASAP PCR profiles from 8 plant species. At – Arabidopsis thaliana, Nt – Nicotiana tabacum, Cs – Citrus sinensis, Pp – Prunus persica, Le – Lycopersicon esculentum, Ah – Amaranthus hypochondriacus, Ch – Coleus hybrida and Ls – Lactuca sativa. Horizontal lines across each species indicate 1 kb size. Vertical columns indicate the amplicons generated from a given primer pair in the 8 plant species.
The maize plastid genome lacks the ycf2 open reading frame in the IR region, therefore primer pairs 3 to 9 failed to produce any amplicons, as anticipated (Figure 4A). Similarly, primer pairs 11, 26 and 27 did not produce any amplicons. Using bl2Seq program the maize IR region was compared with the associated primers and no significant sequence similarity was found between them. Interestingly, one primer each in primer pairs 1 and 10 had very low sequence similarity and yet the amplicons were obtained. Using three rounds of PCR (Table 2) 100% coverage was obtained even in the monocot plastome (Figure 4C), indicating the applicability of eudicot-based primer designs to this taxonomic group. Absence of amplicons from primer pair 26 and 27 could be due to the fact that the primers annealed in the spacer region which could be unique to the eudicots. The results of the reactions are presented in both Table 3 and Figure 3. Table 3 presents the conditions required to produce the amplified regions from individual species, whether the products were obtained from PCR round 1, round 2 or round 3. Figure 3 shows the complete array of amplified products corresponding to the amplification conditions presented in Table 2.
Figure 4 Zea mays ASAP PCR profiles. A. Round 1 and 2 touchdown PCR. B. Round 3 touchdown extension PCR and C. Composite profile from A and B.
Table 3 ASAP coverage. Percentage coverage of the IR B region using ASAP method in 10 different genera and unique features of 4 diverse genera used in the study.
Percentage coverage of the IR region using ASAP
Species Family Round 1 (%) Round 2 (%) Round 3 (%) Final
Arabidopsis thaliana Crucifereae 100 100
Nicotiana tabacum Solanaceae 100 100
Citrus sinensis Rutaceae 100 100
Lycopersicon esculentum Solanaceae 59 100 100
Prunus persica Rosaceae 7.4 100 100
Amaranthus Amaranthaceae 81 100 100
Lactuca sativa Asteraceae 63 100 100
Fragaria × ananassa Rosaceae 74 89 100 100
Coleus × hybrida Lamiaceae 85 92 N/A 92
Zea mays Poaceae 40 80 100 100
Unique sps
Pisum sativum Fabaceae No Inverted Repeat
Ginkgo biloba Ginkgoaceae Small inverted repeat of 17 kb
Pinus taeda Pinaceae Gymnosperm
Equisetum hyemale Equisetaceae Pteridophyte
Fragaria and Prunus (Rosaceae)
Complete coverage of the plastid IR region from Fragaria was obtained after proceeding through three rounds of ASAP PCR (Table 2) with the 27 pairs of primers (Figure 2). These amplicons were generated using Pfu Turbo DNA polymerase (Stratagene Inc., Carlsbad, CA) in order to minimize potential errors generated during PCR reactions. These amplicons were directly sequenced in a 96-well format. The sequence was assembled and annotated as described in Methods.
Interestingly, in Prunus complete coverage was obtained with Round 2 PCR. Sequence comparison with Fragaria revealed that Prunus and Fragaria share considerable sequence similarity in the IR region as expected being from the same phylogenetic group. This is another demonstration of the utility of this technique where two members of the same taxonomic group were sequenced and compared in a very short time frame and in a cost-effective manner.
Other eudicots and identification of a variable region
The ASAP protocol was attempted in other eudicot species for which plastid genome sequence has not been reported. In Citrus and Lycopersicon complete coverage was obtained after Round 1 PCR and for the remaining species almost 99 – 100% coverage was obtained using Round 2 PCR conditions. Electrophoresis profiles revealed highly discernible AFLPs amongst different plant species. The most consistently variable region was represented by primer pair 11. In tobacco this amplicon represents sequences for orf87/ycf15, orf92, orf115, trnL and orf79. Gel electrophoresis profiles of amplicons generated from this primer pair revealed a great range of variability across all species tested (Figure 5A). AFLPs were discernable by gel electrophoresis between two solanaceous species, Nicotiana and Lycopersicon. Sequencing and alignment of this region from two members of Solanaceae, tobacco and tomato revealed a 95% – 98% sequence similarity in the aligning sequences. Tomato had two deletions in the region coding for orf92 and ycf15 in tobacco, which reconciles the smaller amplicon size. On the other extreme is the representative member of Caryophyllaceae, Amaranthus, where sequencing and subsequent alignments with tobacco revealed absence of ORFs between ycf2 and orf92 – trnL-CAA region (Figure 5B). Thus the ASAP method provides the advantage of analyzing a large region from a number of species and identifying a highly variable region at the same time.
Figure 5 A. PCR amplicons generated from primer pair 11 in 9 plant species. Fa – Fragaria × ananassa/Rosaceae, Pp – Prunus persica/Rosaceae, At – Arabidopsis thaliana/Crucifereae, Cs – Citrus sinensis/Rutaceae, Ch – Coleus hybrida/Lamiaceae, Nt – Nicotiana tabacum/Solanaceae, Le – Lycopersicon esculentum/Solanaceae, Ls – Lactuca sativa/Asteraceae, Ah – Amaranthus hypochondriacus/Caryophyllaceae. B. Schematic representation of the polymorphism between the amplicons generated from primer pair 11 in tobacco, tomato and amaranthus.
Distant species
To test the limits of this methodology, the same 27 pairs of primers were used against total DNA from Pisum sativum, Ginkgo biloba, Pinus taeda and Equisetum hyemale. These species represent a unique member of Fabaceae (Pisum – largest deletion resulting in removal of the rRNA cluster; has only one IR), an ancient and contemporary gymnosperm, and a pteridophyte. The primer pairs designed for eudicot plastid genomes were able to amplify the regions corresponding to primer pairs 14 to 25 in Pisum. In tobacco these primer pairs amplify the 98494 – 110052 nt region of the IR that includes the rrn operon. The bl2Seq program was used to determine the sequence similarity between the 27 primer pairs and the Pisum chloroplast genome sequence (Kindly provided by John Gray, John Innes Institute, UK). The observed amplicon patterns are consistent with what is anticipated from the sequence data. Primer pairs that failed to generate an amplicon do not share a significant sequence similarity with the Pisum plastid genome sequence (Figure 6). In the two gymnosperms, similar amplicon patterns were generated from the rrn operon region. Again in the pteridophyte only the primer pairs corresponding to the rrn operon produced an amplicon. The Equisetum chloroplast genome does possess ycf2 gene but comparative sequence analysis with higher plant ycf2 revealed no significant sequence similarity. Interestingly the amino acid sequence similarity was almost 94% (data not shown).
Figure 6 Composite ASAP PCR profiles from 4 unique plant species. Ps – Pisum sativum, Gb – Gingko biloba, Pt – Pinus taeda, Eh – Equisetum hyemale.
Discussion
Chloroplast sequencing efforts of model photosynthetic organisms have provided a wealth of information detailing structural features and plant phylogeny, as well as a basis for manipulation of the plastid genome in the interest of bioengineering. Current molecular phylogenetic studies are carried out using large or complete plastid genome sequences or small coding or intergenic sequences and the Amplification, Sequencing, and Annotation of Plastomes (ASAP) method caters to both approaches. Rapid generation of plastid sequence information is necessary as it can facilitate better design of plastid transformation vectors [15]. One of the factors in successful engineering of cotton and carrot plastomes was the use of species-specific flanking sequences in the transformation vectors [31,32].
The simple yet powerful ASAP technique described herein expands the capacity for any laboratory to dissect the chloroplast genome at the informatics level with a basic set of available resources. The obvious limitation of ASAP method is that some plastid genomes have undergone extensive rearrangements [33]. Therefore, this method complements other strategies [28] for obtaining plastid genome sequences and provides a convenient platform for plastomes that share a considerable level of synteny. A long range PCR based approach was used earlier to sequence the Amborella trichopoda plastid genome [34], requiring nebulization of large amplicons and cloning prior to sequencing. In contrast, the ASAP method performs the dual task of generating small amplicons for direct sequencing and the electrophoresis profiles provide structural genomics information. Direct sequencing of the PCR products limits incorrect base calls resulting from amplification or sequencing errors, as large pools of products are sequenced and even early PCR errors will not be scored incorrectly. In a worst-case scenario a base call at a given position will be unclear, requiring re-amplification and re-sequencing of the amplicon. This approach allows the implementation of non-proofreading, highly processive polymerases that limit costs yet generate substantial quantities of template for downstream analyses.
ASAP also allows description and characterization of the frequent islands of high sequence identity present within the coding regions of sequenced genomes. Using local identities found within the cpDNA sequences of five sequenced eudicot plant species, primer pairs were designed to produce overlapping amplicons that bracket sequences of plastid genes and the rich sequence variability resident to their adjacent intergenic regions. The ASAP method was herein shown to generate representative regions of 10 diverse plant species to almost 100% coverage. Even the rrn region was also amplified from four divergent genomes studied, such as Pisum, Equisetum, Pinus and Ginkgo, with expected efficiency. The success of the method suggests it an excellent first step in the analysis of any novel plastid genome. The ASAP method may be the only practical approach for some rare and/or difficult-to-obtain taxa that are not amenable to chloroplast DNA extraction.
One caveat of this technique is that plastid sequences are not confined to the chloroplast. Plastid DNA sequences are represented in both mitochondrial and nuclear genomes, and may serve as templates for the amplicons generated with the 27 primer pairs. The high copy number of plastomes in green leaf derived total DNA and long primers (26 bp each) used in this approach should preferentially amplify chloroplast sequences. Additionally, nuclear integrated plastid sequences are continuously shuffled and eliminated [35], making them less likely to be incorrectly amplified via this approach. However, with the design of specific primers this technique could be extended to plastid genomes of distant phylogenetic groups or mitochondrial genomes in species where high level of synteny is present.
Conclusion
The ASAP method represents a rapid means to generate a large amount of plastid genome information from simple PCR steps to facilitate bioinformatic dissection and functional genomic studies. The products generated spotlight AFLPs that serve as low-resolution beacons to report regions of high diversity, such as the hypervariable Region 11 (Figure 5). These regions may be particularly meaningful for phylogenetic analyses. In this capacity the ASAP methods may facilitate studies of hypervariable specific regions of the plastid genome, helping to quickly identify areas of likely importance, bypassing the necessity to sequence an entire genome as a first step. The ASAP method eliminates cloning steps, thus negating the need to identify and trim plasmid sequences before assembly. Another powerful facet of the technique is the potential to generate plastid fragment fingerprint for any species. AFLPs produced may be immediately comparable to those produced by other genomes, revealing deviations in otherwise conserved sequence, thus informing of structural-genomic variation. Even the absence of products in related species may be extremely informative.
To summarize, the ASAP method has the following advantages: The method is ideal for laboratories or programs with limited resources interested in obtaining chloroplast DNA sequence information for their particular research interest. ASAP method may be the only practical approach to obtain chloroplast DNA sequence from rare or small plant samples. This method provides a fingerprint of a given chloroplast region, which can be readily compared amongst different genera and give information of structural variability even without the sequence information. Another practical application of this approach is the use of PCR amplicons generated by ASAP method for construction of chloroplast genome microarrays from a given plant species.
Specific to this report, the method was used to amplify and sequence the large IR region from octoploid strawberry and peach. This ~30 kbp region from strawberry was amplified in 33 PCR reactions and then sequenced. The PCR fragments were generated in three thermalcycler runs over two days, and the entire process, from leaf to data on the server, was performed in under a week for under $500. With this minimal time and capital investment roughly 25% of the chloroplast genome has been deciphered; bidirectionally and with complete coverage. The entire process is now being scaled up to sequence an entire plastid genome within the context of a 96-well plate. In this system 1.2 to 1.5 kb amplicons may be produced and sequenced from this standard format. The common format also lends itself to robotic manipulation, and it is exciting to speculate that a single set of 96 primer pairs may be matched to a high-throughput robotic amplification and sequencing system to generate plastid DNA sequence at the rate of a plastid genome per day. It is our hope that this methodology will hasten study of chloroplast sequences, especially those from unusual organisms or those not considered worthy of large investment.
Methods
Primer design
The cpDNA sequences of five eudicot plant species namely Nicotiana tabacum (NC_001879), Arabidopsis thaliana (NC_000932), Atropa belladonna (NC_004561), Spinacea oleracea (NC_002202) and Panax ginseng (NC_006290) were aligned using ClustalW [36]. Tobacco chloroplast genome sequence was used as a reference and its coding regions were delineated in the aligned sequences. Highly conserved, putative primer sites were derived by hand parsing the aligned sequences of the IR B region. Primer candidates satisfied several criteria. A candidate primer must be resident to the coding region or conserved intergenic region and primer pairs must be spaced by ~1.0 to 1.2 kb. The primers must share 95% sequence identity among the representative plastid genomes. For universal, large-format application in simultaneous PCR reactions the primers should maintain approximately 50% GC content and a Tm of approximately 50°C. Table 1 lists the primer sequences, annealing sites, respective position in tobacco, Arabidopsis and maize plastid genomes and the expected amplicon sizes. This set of primers, used to amplify the large IR region in this manuscript, will be supplied by the authors upon request.
DNA preparation and primary optimization
Total plant DNA was isolated from fresh leaf tissue using the Qiagen DNeasy Kit (Qiagen Inc., Valencia, CA) according to manufacturer's instructions, except that the homogenized plant material was mixed in Buffer AP1 for 10 min on a platform vortexer for 10 min and the sample was centrifuged for 10 min at 1000 rpm to remove any debris at this stage. The supernatant was then incubated at 65°C for 10 min and from this point on the manufacturer's protocol was followed. DNA was isolated from two plant species namely Arabidopsis thaliana (ecotype Col-0) and tobacco (Nicotiana tabacum). These two species served as a positive control and a basis for detection of computationally predicted AFLPs. Upon validation of the technique with known genomes, amplification was performed on several crops of agricultural importance namely, strawberry (Fragaria × ananassa; cv. Strawberry Festival), Sweet orange (Citrus sinensis), lettuce (Lactuca sativa), peach (Prunus persica; cv. UF Sun), Tomato (Lycopersicon esculentum), coleus (Coleus × hybrida), Amaranthus (Amaranthus hypochondricus). Maize (Zea mays; W22) was tested as a model monocot. Since it has been sequenced, amplification discrepancies related to insertion/deletion and/or rearrangement can be anticipated and circumvented. The technique was also performed on diverse plant species to test the range of the approach. These species included Pisum sativum L. (Little Marvel), which does not have an IR region and has only one copy of the rRNA genes, an ancient gymnosperm (Ginkgo biloba), a contemporary gymnosperm loblolly pine (Pinus taeda), along with horsetail, (Equisetum hyemale), a pteridophyte which represents an ancient plastid genome that would test and define the limit of the application of the eudicot species-based primer designs reported here.
Before performing reactions with the 27 primer pairs that define the large IR, it was important to establish the amount of total DNA to use in each reaction. Variability in amplification may be introduced from several sources, namely the relative amount of chloroplast DNA to total DNA ratio and the tendency for inhibitory compounds to co-purify with DNA templates in total DNA isolation [37]. For instance, the relative cpDNA: total DNA ratio will be significantly different between Arabidopsis (haploid genome size ~140 Mbp) and Pinus (haploid genome size ~21658 Mbp), two organisms with massively different nuclear genome sizes. DNA preparation from some species, such as strawberry, may introduce phenolics or polysaccharides that could inhibit efficient amplification during PCR.
To accommodate both of these variables an initial reaction set is performed using the most conserved primers, those representing the 16 S rrn locus (primer set 19; Table 1). These primers predictably generate a fragment in all species tested, yet the yield varies considerably based on the amount of template used in the reaction. For instance, while coleus was best amplified with 1 ng per reaction, Equisetum required 10 ng per reaction (data not shown). The fidelity of the reaction is extremely dependent upon total isolated DNA concentration and the inherent cpDNA: nuclear DNA ratio in total DNA. A pilot experiment testing for the production of PCR product over 4–5 orders of magnitude must be performed to optimize conditions for subsequent reactions.
PCR conditions for plastome amplification
Touchdown PCR was utilized to generate PCR amplicons [38]. The first set of reactions was performed using Round I conditions, conditions conducive to amplification in the species for which the primers were designed (Figure 1; Table 2). This reaction would routinely produce amplicons in 24 out of 27 reactions. Reactions that failed to produce a product were reconstituted from fresh reagents and template, and PCR is performed using Reaction II conditions (Table 2). Typically this amplification was sufficient to obtain complete coverage of the large IR in 7 of the 10 species studied (Table 3). Amplicons generated from this approach will be made feely available upon request.
If Round II conditions failed to produce a PCR product it could be assumed that sequence differences in the primer landing site are present or those sites are deleted or rearranged. In this case PCR is performed using each fragment-specific primer and primers from adjacent fragments that successfully amplified in Round I and/or Round II. The process is outlined in Figure 1.
Sequencing was performed directly from PCR products in a 96-well format at the University of Florida ICBR Core Facility using ET Terminator (Amersham Inc, Schaumburg, IL) as reported earlier [29]. A 5 μl aliquot of PCR product from a 50 μl reaction was analyzed by gel electrophoresis to verify purity and concentration. PCR products were treated with ExoSAP to remove primers and nucleotides. Each amplicon was sequenced bidirectionally using the primers used in initial amplification. Sequencing primers were added at 10 pmol/μl in a 10 μl final reaction volume. Sequences with Phred score of >20 were used for assembly. Both strands were sequenced for each fragment thus providing a 2X coverage and, 4X coverage in the overlapping regions.
Assembly, annotation and dissemination of the sequences from Fragaria and Prunus
Sequences generated from a primer pair were first aligned using Blast 2 sequences (bl2seq) available at NCBI website [39,40]. Individual amplicons derived from this process were further assembled using CAP3 [41,42]. The sequence was then annotated using DOGMA [24,25]. Chloroplast genome sequences from the IR B region of Fragaria and Prunus were submitted to the GenBank under accession numbers FACPINVREP (Fa) and PPCPINVREP (Pp).
List of abbreviations
ASAP: Amplification, sequencing, annotation of plastomes; ORF: Open reading frame; AFLP: Amplified fragment length polymorphism; SNP: Single nucleotide polymorphism; PCR: Polymerase chain reaction; FACS: Fluorescence-assisted cell sorting; RCA: Rolling circle amplification; DOGMA: Dual organeller genome annotator
Authors' contributions
AD conceived of the ASAP methodology, designed primers, prepared DNA, performed all PCR/sequencing reactions, and assembled/reported all sequences. AD also prepared all figures and tables for the manuscript. KMF assisted in shaping the ASAP concept and prepared the first draft of the manuscript. Both authors contributed to the development of the final manuscript.
Acknowledgements
The authors would like to thank Professor Pam Soltis and Professor Doug Soltis, University of Florida, for critical review of the work and for their helpful comments and suggestions. The authors would also like to thank Jeremy Ramdial and Gene Peir for their technical support, Dr. Angel Alpuche-Solis, IPICYT, Mexico for Amaranthus and Lactuca sativa DNA; Professor John Davis, University of Florida for Pinus taeda DNA; Professor Harry Klee, University of Florida for tomato DNA; Professor Karen Koch, University of Florida for maize DNA; Professor Ben Elmo Whitty, University of Florida for tobacco plants and Professor John Gray, John Innes Institute, UK for pea chloroplast genome sequence. The authors acknowledge Denise Tombolato and Philip Stewart for critical reading and evaluation of this manuscript. This work was performed with support from NSF grant #0416877 (KMF) and funding from IFAS at the University of Florida.
==== Refs
Palmer JD Chloroplast DNA and molecular phylogeny Bioessays 1985 2 263 267 10.1002/bies.950020607
Maliga P Plastid transformation in higher plants Annu Rev Plant Biol 2004 55 289 313 15377222 10.1146/annurev.arplant.55.031903.141633
Grevich JJ Daniell H Chloroplast genetic engineering: Recent advances and future perspectives Critical Reviews in Plant Sciences 2005 24 83 107 10.1080/07352680590935387
Daniell H Kumar S Dufourmantel N Breakthrough in chloroplast genetic engineering of agronomically important crops Trends Biotechnol 2005 23 238 245 15866001 10.1016/j.tibtech.2005.03.008
Dhingra A Daniell H Salinas J and Sanchez-Serrano JJ Chloroplast genetic engineering via organogenesis or somatic embryogenesis Arabidopsis Protocols Methods in Molecular Biology 2005 323 Second Totowa, New Jersey, Humana Press 525
Sato S Nakamura Y Kaneko T Asamizu E Tabata S Complete structure of the chloroplast genome of Arabidopsis thaliana DNA Res 1999 6 283 290 10574454 10.1093/dnares/6.5.283
Schmitz-Linneweber C Regel R Du TG Hupfer H Herrmann RG Maier RM The plastid chromosome of Atropa belladonna and its comparison with that of Nicotiana tabacum: the role of RNA editing in generating divergence in the process of plant speciation Mol Biol Evol 2002 19 1602 1612 12200487
Hiratsuka J Shimada H Whittier R Ishibashi T Sakamoto M Mori M Kondo C Honji Y Sun CR Meng BY The complete sequence of the rice (Oryza sativa) chloroplast genome: intermolecular recombination between distinct tRNA genes accounts for a major plastid DNA inversion during the evolution of the cereals Mol Gen Genet 1989 217 185 194 2770692
Schmitz-Linneweber C Maier RM Alcaraz JP Cottet A Herrmann RG Mache R The plastid chromosome of spinach (Spinacea oleracea): complete nucleotide sequence and gene organization Plant Mol Biol 2001 45 307 315 11292076 10.1023/A:1006478403810
Shinozaki K Ohme M Tanaka M Wakasugi T Hayashida N Matsubayashi T Zaita N Chunwongse J Obokata J Yamaguchi-Shinozaki K Ohto C Torazawa K Meng BY Sugita M Deno H Kamogashira T Yamada K Kusuda J Takaiwa F Kato A Tohdoh N Shimada H Sugiura M The complete nucleotide sequence of tobacco chloroplast genome:its gene organization and expression EMBO J 1986 5 2043 2049 16453699
Ogihara Y Isono K Kojima T Endo A Hanaoka M Shiina T Terachi T Utsugi S Murata M Mori N Takumi S Ikeo K Gojobori T Murai R Murai K Matsuoka Y Ohnishi Y Tajiri H Tsunewaki K Structural features of a wheat plastome as revealed by complete sequencing of chloroplast DNA Mol Genet Genomics 2002 266 740 746 11810247 10.1007/s00438-001-0606-9
Kato T Kaneko T Sato S Nakamura Y Tabata S Complete structure of the chloroplast genome of a legume, Lotus japonicus 2000 7 323 330 11214967
Maier RM Neckermann K Igloi GL Kossel H Complete sequence of the maize chloroplast genome: gene content, hotspots of divergence and fine tuning of genetic information by transcript editing J Mol Biol 1995 251 614 628 7666415 10.1006/jmbi.1995.0460
Kim KJ Lee HL Complete chloroplast genome sequences from Korean ginseng (Panax schinseng Nees) and comparative analysis of sequence evolution among 17 vascular plants DNA Res 2004 11 247 261 15500250 10.1093/dnares/11.4.247
Saski C Lee SB Daniell H Wood TC Tomkins J Kim HG Jansen RK Complete chloroplast genome sequence of Glycine max and comparative analyses with other legume genomes 2005 59 309 322 16247559
Asano T Tsudzuki T Takahashi S Shimada H Kadowaki K Complete nucleotide sequence of the sugarcane (Saccharum officinarum) chloroplast genome: a comparative analysis of four monocot chloroplast genomes DNA Res 2004 11 93 99 15449542 10.1093/dnares/11.2.93
Soltis DE Albert VA Savolainen V Hilu K Qiu YL Chase MW Farris JS Stefanovic S Rice DW Palmer JD Soltis PS Genome-scale data, angiosperm relationships, and 'ending incongruence': a cautionary tale in phylogenetics Trends in Plant Science 2004 9 477 483 15465682 10.1016/j.tplants.2004.08.008
Wolf PG Rowe CA Sinclair RB Hasebe M Complete nucleotide sequence of the chloroplast genome from a leptosporangiate fern, Adiantum capillus-veneris L DNA Research 2003 10 59 65 12755170 10.1093/dnares/10.2.59
Martin W Deusch O Stawski N Grunheit N Goremykin V Chloroplast genome phylogenetics: why we need independent approaches to plant molecular evolution Trends Plant Sci 2005 10 203 209 15882651 10.1016/j.tplants.2005.03.007
Project DT
Project CGS
project C
Wolf PG Karol KG Mandoli DF Kuehl J Arumuganathan K Ellis MW Mishler BD Kelch DG Olmstead RG Boore JL The first complete chloroplast genome sequence of a lycophyte, Huperzia lucidula (Lycopodiaceae) Gene 2005 350 117 128 15788152 10.1016/j.gene.2005.01.018
Website D
Wyman SK Jansen RK Boore JL Automatic annotation of organellar genomes with DOGMA Bioinformatics 2004 20 3252 3255 15180927 10.1093/bioinformatics/bth352
Palmer JD Comparative organization of chloroplast genomes Annu Rev Genet 1985 19 325 354 3936406 10.1146/annurev.ge.19.120185.001545
Odintsova MS Iurina NP Plastidic genome of higher plants and algae: structure and function Mol Biol (Mosk) 2003 37 768 783 14593913
Jansen RK Raubeson LA Boore JL dePamphilis CW Chumley TW Haberle RC Wyman SK Alverson AJ Peery R Herman SJ Methods for obtaining and analyzing whole chloroplast genome sequences Methods in Enzymology 2005 Volume 395 , Academic Press 348 384 15865976 10.1016/S0076-6879(05)95020-9
Folta KM Staton M Stewart PJ Jung S Bies DH Jesudurai C Main D Expressed sequence tags (ESTs) and simple sequence repeat (SSR) markers from octoploid strawberry (Fragaria x ananassa) BMC Plant Biol 2005 5 12 15985176 10.1186/1471-2229-5-12
Lin J Davis TM S1 analysis of long PCR heteroduplexes: detection of chloroplast indel polymorphisms in Fragaria Theoretical and Applied Genetics 2000 101 415 420 10.1007/s001220051498
Kumar S Dhingra A Daniell H Plastid-expressed betaine aldehyde dehydrogenase gene in carrot cultured cells, roots, and leaves confers enhanced salt tolerance Plant Physiology 2004 136 2843 2854 15347789 10.1104/pp.104.045187
Kumar S Dhingra A Daniell H Stable transformation of the cotton plastid genome and maternal inheritance of transgenes Plant Molecular Biology 2004 56 203 216 15604738 10.1007/s11103-004-2907-y
Cosner ME Jansen RK Palmer JD Downie SR The highly rearranged chloroplast genome of Trachelium caeruleum (Campanulaceae): multiple inversions, inverted repeat expansion and contraction, transposition, insertions/deletions, and several repeat families Curr Genet 1997 31 419 429 9162114 10.1007/s002940050225
Goremykin VV Hirsch-Ernst KI Wolfl S Hellwig FH Analysis of the Amborella trichopoda chloroplast genome sequence suggests that Amborella is not a basal angiosperm Mol Biol Evol 2003 20 1499 1505 12832641 10.1093/molbev/msg159
Matsuo M Ito Y Yamauchi R Obokata J The rice nuclear genome continuously integrates, shuffles, and eliminates the chloroplast genome to cause chloroplast-nuclear DNA flux Plant Cell 2005 17 665 675 15705954 10.1105/tpc.104.027706
website CW
Demeke T Adams RP The effects of plant polysaccharides and buffer additives on PCR Biotechniques 1992 12 332 334 1571138
Don RH Cox PT Wainwright BJ Baker K Mattick JS Touchdown PCR to circumvent spurious priming during gene amplification Nucleic Acids Research 1991 19 4008 4008 1861999
Tatusova TA Madden TL BLAST 2 SEQUENCES, a new tool for comparing protein and nucleotide sequences Fems Microbiology Letters 1999 174 247 250 10339815 10.1016/S0378-1097(99)00149-4
Huang XQ Madan A CAP3: A DNA sequence assembly program Genome Research 1999 9 868 877 10508846 10.1101/gr.9.9.868
Huang XQ Madan A
|
16336644
|
PMC1318494
|
CC BY
|
2021-01-04 16:32:47
|
no
|
BMC Genomics. 2005 Dec 7; 6:176
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-176
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-911633668610.1186/1743-422X-2-91Short ReportImportance of disulphide bonds for vaccinia virus L1R protein function Blouch Robert E [email protected] Chelsea M [email protected] Dennis E [email protected] Department of Microbiology, Oregon State University, 220 Nash Hall, Corvallis, Oregon, 97331, USA 2 Siga Technologies, 4575 SW Research Way, Suite 230, Corvallis, Oregon, 97333, USA 2005 9 12 2005 2 91 91 9 8 2005 9 12 2005 Copyright © 2005 Blouch et al; licensee BioMed Central Ltd.2005Blouch 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.
L1R, a myristylated late gene product of vaccinia virus, is essential for formation of infectious intracellular mature virions (IMV). In its absence, only viral particles arrested at an immature stage are detected and no infectious progeny virus is produced. Previous studies have shown that the L1R protein is exclusively associated with the IMV membrane and that myristylation is required for correct targeting. The L1R protein contains six cysteine amino acid residues that have all been shown to participate in intramolecular disulphide bonds. However, it was not clear what role, if any, the disulfide bonds play in the membrane topology of the L1R protein. To address this question, a comprehensive library of L1R mutants in which the cysteine residues have been mutated to serine (either individually or in combination) were tested for their ability to rescue a L1R conditional lethal mutant virus under non-permissive conditions. Much to our surprise, we determined that C57 was not essential for production of infectious IMV. These results suggest that protein disulphide isomerases may be involved in reorganization of disulfide bonds within the L1R protein.
==== Body
Findings
Vaccinia virus (VV) continues to be the model organism for the investigation of the Orthopoxviridae family and as a result is the most widely studied and best understood virus in this family. This being said, our understanding of this virus family is still limited due to the size and complexity of these DNA viruses which maintain a broad host range having members that infect insects (entomopoxviruses) and a large number of vertebrates (chordopoxviruses). Two poxviruses known to cause disease in human hosts are variola, the causative agent of smallpox and Molluscum contagiosum, which causes small tumors on the skin and is an opportunistic pathogen in AIDS patients. Largest of the DNA viruses, the poxvirus genome encodes more than 200 gene products. One reason for the sheer number of genes is the viruses' unique ability to replicate its genome, form complex macromolecular structures and assemble infectious viral particles solely within the cytoplasmic compartment of infected cells.
It has previously been shown that the product of the VV L1R open reading frame is essential for the formation of intracellular mature virions (IMV) and plays a role in virion morphogenesis [1-4]. In the absence of L1R, only immature virion particles are formed and proteolytic cleavage of core proteins does not occur [1]. This prevents core condensation and arrests virion morphogenesis at a non-infectious stage. L1R is the target of neutralizing antibodies to IMV [5], therefore making it a potential target for the development of antivirals. However, the biological function of L1R remains largely unknown. L1R contains six conserved cysteine residues that have been shown to be oxidized to form three intramolecular disulphide bonds [6]. These are believed to be essential for correct protein folding and proper function. In addition, they may serve as a membrane attachment factor, playing a role in trafficking of L1R to the endoplasmic reticulum-golgi intermediate compartment (ERGIC).
In this report conditional-lethal expression of L1R and complementation with a library of cysteine-to-serine L1R mutants was used to investigate the importance of disulphide bond formation and the presence of the contributing cysteine residues to protein function.
A recombinant virus was constructed in which the expression of the L1R gene could be regulated by the presence or absence of TET using the components of the bacterial tetracycline operon [7]. This system has previously been shown to be successful in the regulation of the vaccinia virus I7L [8], G1L [9,10] and A14L [11] genes. A plasmid containing the tetracycline operator (TetO) just upstream of the L1R open reading frame (ORF) and including flanking genomic DNA sequence (including the native promoter) to aid in homologous recombination was used to create the recombinant virus vvTetO:L1R. T-Rex-293 cells (Invitrogen) which express the tetracycline repressor (TetR) were used to regulate expression of the L1R gene from the inducible mutant virus.
To verify that expression of L1R is essential for viral replication and can be regulated by tetracycline (Tet), a growth curve in the presence and absence of Tet was performed (Figure 1A). Under permissive conditions, in the presence of 0.1 μg/ml Tet, vvTetO:L1R grew to the same yield and with the same kinetics as wild type virus. However, in the absence of Tet, there was over a 3-log decrease in viral titer. Transfection of plasmid borne L1R, driven off of either its native promoter (p(wtp)L1R or a synthethic early/late promoter (p(E/Lp)L1R), resulted in a greater than 100-fold increase in infectious progeny virus over the control with no transfected DNA. (Figure 1B). There was concern that L1R being expressed constitutively at all times during infection as opposed to only at late times might negatively impact viral yield or in some way interrupt or slow the viral life cycle. This did not occur, most likely because three proteins, essential for disulphide bond formation in L1R, are expressed as late proteins. Without G4L, A2.5L, and E10R present early in the infection, L1R was not in its active conformation. Its presence in non-disulphide bonded form does not appear to hinder virion morphogenesis or viral assembly.
Figure 1 (A) Growth-curve kinetics comparing vv:Western Reserve to vv:TetO:L1R under permissive and non-permissive conditions. Each infection was performed at an MOI of 0.1 pfu and harvested at various times from 0 to 48 hpi and the resulting cell lysates were titered using BSC40 cells. (B) Transiently expressed L1R is capable of phenotypic rescue of conditional-lethal viral infection under non-permissive conditions. Infections were performed at 0.1 MOI with either VV-WR (WR) or VV-TetO:L1R (TetO) in the absence of tetracycline unless noted. Transfections of plasmid DNA were performed using 2 μg of pUC19, p(E/Lp)L1R or p(wtp)L1R. All infections were harvested at 24 hpi and titered on BSC40 cells.
L1R contains six cysteine amino acids that bind through disulphide bridges to form three stable intramolecular bonds in the active form of the protein [12]. Figure 2A shows the location of the six cysteine residues involved in disulphide bonding. In order to determine if all three bonds are essential to protein function and elucidate possible partnering models, plasmid DNA containing the L1R ORF expressed from the synthetic early/late promoter and containing individual cysteine to serine mutants were expressed during infections with vvTetO:L1R under non-permissive conditions. Five of the six mutants (C34S, C49S, C116S, C136S, and C158S) were incapable of rescuing the infections. Interestingly, L1R lacking the third cysteine at amino acid 57 was capable of 52% rescue and suggests that participation of C57 appears to be non-essential for protein function (Figure 2B). Rescue experiments were also performed using double mutants of L1R containing every possible variation of two cysteine-to-serine tandem mutants. The results showed that none of the double mutants were capable of significant rescue (Figure 2C).
Figure 2 (A) Diagram of the location of the six cysteine residues in L1R. (B) Transient Expression of L1R cysteine-to-serine single mutants. Infections were performed by transfection of 2 μg of pL1R each containing a single Cysteine-to-Serine mutation in the coding sequence at time of infection with VV-TetO:L1R at an MOI of 0.1 under non-permissive conditions. Infections were harvested at 24 hpi and titered on BSC40 cells. (C) Transient Expression of L1R cysteine-to-serine double mutants. Infections were performed by transfection of 2 μg of pL1R containing double Cysteine-to-Serine mutations in the coding sequence at time of infection with VV-TetO:L1R at an MOI of 0.1 under non-permissive conditions. Infections were harvested at 24 hpi and titered on BSC40 cells.
The Tet operon conditional-lethal system has been used to study the effects of mutations introduced into L1R and transiently expressed during infections under non-permissive conditions. Utilizing this approach, it was shown that five of the six cysteine amino acids present in wild-type L1R are essential for proper L1R function and active conformation. The single cysteine at aa-57 was shown to be non-essential in this role. This presents a puzzle considering there is previous research suggesting that there are three intramolecular disulphide bonds utilizing all six cysteine residues [12]. Confocal microscopy comparing wild-type infections and infections with transiently expressed mutant L1R verified that the protein was being made in from all six constructs (data not shown). It appeared that trafficking of L1R to the proper membrane may be dependent upon proper disulphide bonding as localization was altered in Cys-49 and Cys 116 mutants. These findings suggest the possibility that a cellular or virally encoded protein disulphide isomerase is required for proper disulphide pairing in active L1R. It is conceivable that cyteine-57 forms an incorrect disulphide pairing as an intermediate. Protein disulphide isomerase is then necessary to resolve this mispairing and the disulphide bond that is formed by Cys-57 and its unknown partner is not necessary for functional L1R. This is further evidenced by the crystal structure of L1R [12]. The terminal protein in the disulphide forming redox pathway is G4L [6], a cytoplasmic protein. If disulphide bonding of L1R were to occur in this fashion in the cytoplasm, the trafficking effects of the N-terminal myristoyl group, which would be hidden within L1R, could be lost. This could be circumvented if G4L established an intermediate disulphide bond configuration that exposes the myristoyl group and allows trafficking to the ERGIC. Then, once incorporated into the membrane of the ERGIC, the bonds are isomerized, converting L1R to its active confirmation. The ERGIC associated A2.5L/E10R heterodimer contains two C-XXX-C motifs that have been established in DsbC homodimers in E. coli [13]. The C57S mutant was capable of better than 50% rescue of infection under non-permissive conditions when compared to rescue with wild-type L1R. This small decrease could be attributed to lack of the third bond, however its absence does not abolish function.
A series of tandem mutants utilizing every combination of two cysteine-to-serine mutations was also tested in the same manner. None of the mutants were capable of significant rescue under non-permissive conditions. This is not surprising based upon observation of the single mutants except in one respect: it suggests that if there is a second incorrect pairing that a protein disulphide isomerase is needed to correct, that this is an essential intermediate and without it the correct bond alignment cannot be achieved. The virally encoded redox pathway described by Senkevich et al (2002), contains disulphide linked E10R and A2.5L which they compared to E. coli DsbB and yeast ERO1p and ERV2p which contain two pairs of active cysteine residues. G4L is likened to the downstream thioredoxin-like proteins DsbA in E. coli, and PDI and its homologues in the ER of yeast. It is possible that isomerase activity during vaccinia infection is achieved by one of the known viral redox proteins or by another, yet unknown viral protein. This activity, if shown to exist is not likely to be attributed to a cellular protein, as disulphide bond formation and isomerase activity is believed to occur solely in the lumen of the ER.
This study has shown that only two of the three intramolecular disulphide bonds are essential for L1R to perform its function in formation of infectious IMV particles. Cysteine residues at positions 34, 49, 116, 136 and 158 are essential for protein function and viral propagation. The cysteine at position 57 is non-essential and its partnering capabilities are not necessary for proper function of L1R. Cys-49 and Cys-116 disrupted localization if L1R as evidenced by confocal microscopy. The results also suggest the presence of isomerase activity in L1R bond reshuffling and that it may be a required factor in promoting proper protein conformation and function. Here, we propose that an incorrectly disulphide bonded intermediate mediates trafficking of L1R to the membrane of the ERGIC, where isomerization of these bonds results in an active conformation with the myristoyl group hidden with the hydrophobic cavity of the active protein.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CMB constructed the recombinant virus. REB conducted the experiments. CMB and REB wrote the manuscript. DEH conceived the study, coordinated the research efforts and edited the paper. All authors read and approved the final manuscript.
Acknowledgements
We kindly thank Dina Alzhanova for the confocal microscopy work. This work was supported by National Institute of Health grant AI21335.
==== Refs
Ravanello MP Hruby DE Conditional lethal expression of the vaccinia virus L1R myristylated protein reveals a role in virion assembly J Virol 1994 68 6401 6410 8083978
Ravanello MP Hruby DE Characterization of the vaccinia virus L1R myristylprotein as a component of the intracellular virion envelope J Gen Virol 1994 75 1479 1483 8207414
Ravanello MP Franke CA Hruby DE An NH2-terminal peptide from the vaccinia virus L1R protein directs the myristylation and virion envelope localization of a heterologous fusion protein J Bio Chem 1993 268 7585 7593 8463289
Martin KH Grosenbach DW Franke CA Hruby DE Identification and analysis of three myristylated vaccinia virus late proteins J Virol 1997 71 5218 5226 9188589
Wolffe EJ Vijaya S Moss B A myristylated membrane protein encoded by the vaccinia virus L1R open reading frame is the target of potent neutralizing monoclonal antibodies Virology 1995 211 53 63 7645236 10.1006/viro.1995.1378
Senkevich TG White CL Koonin EV Moss B Complete pathway for protein disulfide bond formation encoded by poxviruses Proc Natl Acad Sci 2002 99 6667 6672 11983854 10.1073/pnas.062163799
Jorgensen RA Reznikoff WS Organization of structural and regulatory genes that mediate tetracycline resistance in transposon Tn10 J Bacteriol 1979 138 705 14 378932
Byrd CM Hruby DE A conditional-lethal vaccinia virus mutant demonstrates that the I7L gene product is required for virion morphogenesis Virology J 2005 2 4 15701171 10.1186/1743-422X-2-4
Hedengren-Olcott M Hruby DE Conditional expression of vaccinia virus genes in mammalian cell lines expressing the tetracycline repressor J Virol Methods 2004 120 9 12 15234804 10.1016/j.jviromet.2004.03.016
Hedengren-Olcott M Byrd CM Watson J Hruby DE The vaccinia virus G1L putative metalloproteinase is essential for viral replication in vivo J Virol 2004 78 9947 53 15331728 10.1128/JVI.78.18.9947-9953.2004
Traktman P Liu K DeMasi J Rollins R Jesty S Unger B Elucidating the essential role of the A14 phosphoprotein in vaccinia virus morphogenesis: construction and characterization of a tetracycline-inducible recombinant J Virol 2000 74 3682 95 10729144 10.1128/JVI.74.8.3682-3695.2000
Su HP Garman SC Allison TJ Fogg C Moss B Garboczi DN The 1.51-Angstrom Structure of the Poxvirus L1 Protein, a Target of Potent Neutralizing Antibodies Proc Natl Acad Sci 2005 102 4240 45 15761054 10.1073/pnas.0501103102
Zhou Z Peng Y Hao SF Zeng ZH Wang CC Dimerization by Domain Hybridization Bestows Chaperone and Isomerase Activities J Biol Chem 2003 278 43292 8 12933788 10.1074/jbc.M306945200
|
16336686
|
PMC1318495
|
CC BY
|
2021-01-04 16:38:58
|
no
|
Virol J. 2005 Dec 9; 2:91
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-91
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1751633663810.1186/1471-2164-6-175Research ArticleSequence comparisons of plasmids pBJS-O of Spiroplasma citri and pSKU146 of S. kunkelii: implications for plasmid evolution Joshi Bharat D [email protected] Michael [email protected] Janet [email protected] Jacqueline [email protected] Ulrich [email protected] Department of Biochemistry and Molecular Biology, Oklahoma State University, 246 NRC, Stillwater, OK 74078, USA2 Department of Entomology and Plant Pathology, Oklahoma State University, 127 NRC, Stillwater, OK 74078, USA3 P&K Microbiology Services, Inc. 1936 Olney Ave., Cherry Hill, NJ 08003, USA2005 7 12 2005 6 175 175 13 7 2005 7 12 2005 Copyright © 2005 Joshi et al; licensee BioMed Central Ltd.2005Joshi 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
Spiroplasma citri BR3-3X and S. kunkelii CR2-3X cause serious diseases worldwide on citrus and maize species, respectively. S. citri BR3-3X harbors a plasmid, pBJS-Original (pBJS-O), that encodes the spiroplasma adhesion related protein 1 (SARP1), a protein implicated in binding of the pathogen to cells of its leafhopper vector, Circulifer tenellus. The S. kunkelii CR2-3X plasmid, pSKU146, encodes a homolog of SARP1, Sk-ARP1. Due to the close phylogenetic relationship of the two pathogens, we hypothesized that the two plasmids are closely related as well.
Results
The nucleotide sequence of pBJS-O was determined and compared to the sequences of a plasmid from BR3-T (pBJS-T), which is a multiply passaged leafhopper transmissible derivative of BR3-3X, and to known plasmid sequences including that of pSKU146. In addition to arp1, the 13,374 bp pBJS-O sequence putatively contains nine genes, recognized as open reading frames (ORFs). Several pBJS-O ORFs have homologs on pSKU146. However, the sequences flanking soj-like genes on both plasmids were found to be more distant from one another than sequences in any other region. Further, unlike pSKU146, pBJS-O lacks the conserved oriT region characteristic of the IncP group of bacterial plasmids. We were unable to identify a region in pBJS-O resembling a known plasmid origin of transfer. In regions where sequence was available for the plasmid from both BR3-3X and BR3-T, the pBJS-T sequence had a 0.4 kb deletion relative to its progenitor, pBJS-O. Southern blot hybridization of extrachromosomal DNA from various S. citri strains and spiroplasma species to an arp-specific probe and a probe made from the entire plasmid DNA of BR3-3X revealed limited conservation of both sequences in the genus Spiroplasma. Finally, we also report the presence on the BR3-3X chromosome of arp2, an S. citri homolog of arp1 that encodes the predicted protein SARP2. The C-terminal domain of SARP2 is homologous to that of SARP1, but its N-terminal domain is distinct.
Conclusion
Our data suggest that pBJS is a novel S. citri plasmid that does not belong to any known plasmid incompatibility group. The differences between pBJS-O and pSKU146 suggest that one or more events of recombination have contributed to the divergence of the plasmids of the two sister Spiroplasma species; the plasmid from S. citri itself has diverged slightly during the derivation of S. citri BR3-T from BR3-3X. Our data also show that pBJS-O encodes the putative adhesin SARP1. The presence of traE and mob on pBJS-O suggests a role for the plasmid in spiroplasmal conjugation.
==== Body
Background
The phytopathogenic spiroplasmas and phytoplasmas, which cause serious diseases of economically important plant species worldwide [[1] and [2]], are wall-less prokaryotes phylogenetically related to Gram-positive eubacteria with low G+C content [3]. They are transmitted in nature by phloem-feeding insects, predominantly leafhoppers, in a propagative manner [4]. Even though there are close to forty recognized spiroplasma species, only three plant pathogenic spiroplasmas have been identified and characterized to date. S. citri [[5,6] and [7]] is the causative agent of stubborn disease of citrus and brittle root disease of horseradish; S. kunkelii [[8,9] and [10]] is the etiological agent of corn stunt; and S. phoeniceum [11] causes periwinkle yellows. Unlike phytoplasmas, spiroplasmas can be cultured in vitro. Therefore, the relationships between S. citri and its insect vectors, the beet leafhopper, Circulifer tenellus, and the related species, C. haematoceps [12], have been investigated extensively, serving as models for investigating the molecular aspects of mollicute-vector interactions.
Spiroplasma binding to insect host and non-host cells, both in tissue-culture and within the intact insect, has been reported [[13] and [14]]. The loss and restoration of the ability of S. citri to adhere to tissue-cultured C. tenellus cells was associated with degradation and restoration of P89 (designated SARP1), a spiroplasma membrane protein [14]. Due to the possible direct involvement of SARP1 in the spiroplasma-leafhopper interaction, it was hypothesized that SARP1 is an adhesin. Later, Berg et al [15] reported cloning and characterization of arp1, the gene encoding SARP1, from S. citri BR3-T. They also reported that mature SARP1 protein contains a novel domain at the N-terminus, called "sarpin", made of six repeats of 39–42 amino acids each.
S. citri harbors several extrachromosomal DNAs with unique restriction patterns [[16-18] and [19]]. S. citri lines, derived from a clone, and sister clones of the same lines showed differences in their extrachromosomal DNAs [20]. In addition to known plasmids, there are replicative forms (RFs) of several viruses and other uncharacterized circular extrachromosomal DNAs in S. citri [21].
Plasmids have also been noted in strains of S. kunkelii [22]. Recently, Davis and colleagues [23] reported the complete sequence of the S. kunkelii CR2-3X plasmid pSKU146, which encodes a homolog of SARP1, Sk-ARP1. In the present study, we isolated and characterized a related indigenous plasmid, designated pBJS-Original (pBJS-O), from S. citri BR3-3X. This is a report of the discovery, distribution and characterization of that plasmid. Among other genes, pBJS-O contains arp1. The significance of the discovery of pBJS-O in relation to our current understanding of the S. citri-leafhopper interactions and potential genetic manipulations in mollicutes is discussed. Implications for the evolution of both pBJS-O and pSKU146 are also presented.
Results
Detection and analysis of arp2
SARP1 has been characterized previously and the gene encoding it, arp1, has been cloned and sequenced [GenBank:AJ297706] from S. citri BR3-T [15]. In the process, an RsaI restriction fragment was cloned and sequenced from BR3-T genomic DNA; the alignment of this fragment with AJ297706 revealed 92% similarity in the 3' 660 nucleotides of the former sequence. However in the 5' 55 bases of the total 715 bp, upstream from position 2370 in AJ297706, the new fragment was not similar to the known sequence. We designated this gene, which resembles but is not identical to arp1, as arp2 and its putative protein product as SARP2. As also noted by Bai et al. [22], the S. kunkelii CR2-3X genome contains two sequences similar to those of S. citri BR3-T arp genes. The predicted protein, Sk-ARP1 (for S. kunkelii adhesion related protein 1), encoded by the first sequence, Sk-arp1, contains seven rather than six sarpin repeats and has C-terminal domains resembling those of SARP1 [15]. The second sequence encodes a putative protein whose C-terminus is homologous to that of SARP1, but has an unrelated N-terminus. This protein is designated Sk-ARP2 (S. kunkelii adhesion related protein 2) and the corresponding gene is named Sk-arp2. SARP1 has sequence similarity with known adhesins. Fleury et al. [24] have shown that the predicted amino acid sequence of P40, a Mycoplasma agalactiae cytadhesin, is similar not only to that of SARP1 but also to the one of P50, an adhesin of M. hominis.
Isolation and distribution of Spiroplasma extrachromosomal DNA
We isolated extrachromosomal DNA from S. citri BR3-3X to test the hypothesis that this DNA contains an arp-like gene as in S. kunkelii. Restriction of the DNA with single enzymes, including BglII and NdeI, converted a DNA migrating with 9 kb into a fragment migrating close to 7 kb (Fig. 1). These results were consistent with the presence of a single major plasmid. We designated the plasmid pBJS-O. By nucleotide sequencing, we determined that the actual size of the plasmid was 13,374 bp and deposited the sequence in the EMBL Nucleotide Sequence Database [EMBL:AJ972409]
Figure 1 Restriction digests of pBJS-O DNA with MboI, BglII and NdeI. The marker used was the High Mass Ladder (Invitrogen Corp., Carlsbad, CA, USA). Sizes of the fragments are denoted in kb.
To test the conservation of pBJS-O in S. citri strains derived from S. citri BR3, plasmid preparations from S. citri BR3-3X and from BR3-G, BR3-T, BR3-M and BR3-P, lines derived from BR3-3X, were probed with a DNA fragment derived from arp1 (Fig. 2 and Table 1). All hybridized with the probe, producing two or more bands. To test the conservation of pBJS-O in other S. citri strains, other plant-associated spiroplasmas and the closest relative of S. citri, S. melliferum [25], the plasmids of S. kunkelii CR2-3X, S. melliferum, S. citri strains R8A2, ASP-1 and Beni Mellal, S. floricola and S. phoeniceum also were probed with the arp1-derived probe (Fig. 3A). Only S. kunkelii CR2-3X and S. melliferum reacted in the hybridization. However, when the same plasmids were probed with the whole pBJS-O plasmid as a probe (Fig. 3B), all the sample preparations, except those from S. citri Beni Mellal, S. floricola and S. phoeniceum, hybridized with the probe. All of the above Southern hybridization experiments revealed multiple reactive species in the plasmid preparations and the hybridization patterns of EcoRI-digested and undigested plasmid samples were very similar to each other. For comparison, the blots included EcoRI-digested chromosomal DNA of S. citri BR3-3X. A single hybridization signal distinct from those of plasmid preparations was observed (Fig. 2).
Figure 2 Southern blotting hybridization of (A) S. citri BR3-3X, BR3-G, BR3-T, and (B) S. citri BR3-M and BR3-P plasmid preparations to an arp1-derived probe. EcoRI-digested S. citri BR3-3X chromosomal DNA and EcoRI-digested and undigested plasmid preparations from BR3-3X, BR3-G, BR3-T, BR3-M and BR3-P are shown. D, digested with EcoRI; U, undigested. Hybridization in the marker lane is due to presence of short pBluescript vector sequences in the probe.
Table 1 Results of the Southern hybridizations of undigested plasmid preparations from various spiroplasma species and S. citri strains to either an arp1-derived probe or whole pBJS-O probe. + and - denote positive and negative hybridizations, respectively.
Spiroplasma Probe Biological features
Species* Strain arp1 pBJS-O Transmissibility Pathogenicity
BR3-3X + + + +
BR3-G + + - -
BR3-T + + + +
S. citri BR3-M + + + +
BR3-P + + Very low -
ASP-1 - + Unknown Unknown
R8A2 - + Unknown Unknown
Beni Mellal - - Unknown Unknown
S. kunkelii CR2-3X + + + +
S. phoeniceum P40 - - +** +
S. melliferum TS2 + + Unknown -
S. floricola 23-6 - - Unknown -
*S. citri, S. kunkelii, S. phoeniceum and S. melliferum belong to serogroup I, whereas S. floricola belongs to serogroup III. The biological features of the spiroplasmas, except for S. phoeniceum, are taken from references 25 and 46.
**Only experimental transmission to the plant host is known for this spiroplasma. It is unclear whether it can be naturally transmitted by leafhoppers.
Figure 3 Southern blotting hybridization of undigested plasmid preparations from different S. citri strains and spiroplasma species to (A) an arp1-derived probe and (B) the whole pBJS-O probe. Plasmids from S. citri ASP-1, R8A2 and Beni Mellal, and from S. floricola, S. melliferum, S. phoeniceum and S. kunkelii CR2-3X were used. The blot shown in panel A was stripped and rehybridized using the whole pBJS-O probe, shown in panel B.
arp1 and arp2 locations in S. citri BR3-3X
The Southern blot hybridization results suggest that arp-related sequences are present on both a plasmid and the chromosome. arp1 and arp2 from BR3-T are nearly identical over a considerable portion of their nucleotide sequence. Hence, using a probe containing this conserved region should detect both genes. Nevertheless, arp1 and arp2 differ at several positions in those regions. To determine whether the BR3-3X plasmid and chromosomal sequences represented arp1 or arp2 genes, we determined parts of the sequences of BR3-3X plasmid and chromosomal DNAs by direct sequencing and by sequencing amplified PCR products. Comparison of the BR3-3X arp sequences with those of BR3-T revealed that the BR3-3X arp2 sequence had diverged more from the other three sequences than the latter had from each other (Figs. 4 and 5). At positions where the two BR3-T genes differed from one another, the chromosomal BR3-3X sequence had arp2 residues in 21 positions and arp1 residues in only 3 positions (Figs. 4A and 4B). Conversely, at arp1- and arp2-specific positions, the BR3-3X plasmid DNA had no arp2 residues and 28 arp1 residues. Further, at all 57 positions at which chromosomal and plasmid sequences differed, the BR3-3X plasmid and arp1 nucleotides were identical. Hence, we conclude that, in S. citri BR3-3X, the arp1 gene resides on a plasmid and that the arp2 gene most likely resides on the chromosome. The newly determined arp2 sequences from BR3-3X and BR3-T were deposited in the EMBL Nucleotide Sequence Database [EMBL:AM040506 and EMBL:AM040505, respectively].
Figure 4 Comparison of partial nucleotide sequences of S. citri BR3-3X chromosomal (BR3-3X_chr) and plasmid (BR3-3X_pl) sequences to those of available BR3-T arp1 and arp2 genes (BR3-T_arp1 and BR3-T_arp2, respectively). Only the regions containing polymorphic positions are shown. (A) region of arp1 positions 3572 to 3800 (AJ297706). (B) region of arp1 positions 4118 to 4177 and (C) region of arp1 positions 4658 to 4717). S. citri BR3-T arp2 is not available for the last sequence alignment. Gaps are denoted by dashed lines, whereas dots denote identical bases.
Figure 5 An unrooted phylogenetic tree representing the nucleotide sequence alignment shown in Fig. 4A. The tree was generated using algorithms ClustalW and PHYLIP from the Biology Workbench using the neighbour-joining method.
Complete pBJS-O sequencing and analysis
The 4273 bp sequence [GenBank:AJ297706] originally cloned and characterized from S. citri BR3-T [15] contains a partial ORF soj, followed by ORF2, P89 (arp1) and another partial ORF, ORF4. AJ297706 was used to design primers and initiate primer walking to determine the complete pBJS-O plasmid sequence and allow its characterization. During sequencing, a segment (from nucleotide 1–80) of the assembled sequence proved particularly difficult to sequence. It contained three of the six oligopurine/oligopyrimidine tracts of 12 or more bp in the entire plasmid sequence. That the sequence of the tracts was consistent with triple-helix formation suggests that this region of the plasmid may readily form triple-helical structures interfering with sequencing.
The total plasmid sequence is 13,374 bp in length and contains ten predicted ORFs (Fig. 6 and Table 2), of which orf2 (S. citri ORF2) has no homologs, and orf9 and orf10 appear to have distant relatives (E values 0.34 and 0.024, respectively). Of the ten, six putative pBJS-O-ORFs have homologs in pSKU146, the recently characterized S. kunkelii CR2-3X plasmid [16]: arp1 (adhesin protein; E value 0.0), orf4 (hypothetical protein pSKU146_11; E value 0.0), traE (conjugation ATPase; E value 0.0), orf6 (hypothetical protein pSKU146_13; E value 9 × 10-45), mob (mobilization protein; E value 0.0) and orf8 (hypothetical protein pSKU146_17; E value 1 × 10-103). Predicted products of traE and mob are similar to proteins involved in conjugative DNA transfer in other bacterial genera. In the regions where the plasmid sequence was available from both BR3-3X and BR3-T, pBJS-T (the plasmid from S. citri BR3-T) sequence had a 0.4 kb deletion relative to pBJS-O, bringing the orf4 gene close to arp1 and traE. In BR3-3X, however, arp1 and orf4 are separated by 281 bp. The nucleotide sequence variations between pBJS-O and pBJS-T were found to be clustered. Two regions of enhanced variation were observed over a 200 bp stretch in the ORF2-arp1 intergenic region (positions 2700 to 2900 in pBJS-O). In a comparable stretch from position 5262 to 5544 in the arp1-ORF4 intergenic region, a single stretch of dissimilarity was found.
Figure 6 The ORF and restriction map of pBJS-O.
Table 2 Descriptions of ORFs present on pBJS-O.
ORF # Map Position Length (bp) Closest homolog (from BLASTP search) E value
1 1114–1896 783 Soj-like protein [S. citri]* 1 × 10-116
2 1916–2434 519 hypothetical protein [S. citri]* 1 × 10-100
3 2859–5255 2397 putative adhesin P89 [S. citri]* 0
4 5536–7101 1566 hypothetical protein [S. kunkelii] 0
5 7091–9613 2523 conjugation ATPase [S. kunkelii] 0
6 9617–9955 339 hypothetical protein [S. kunkelii] 9 × 10-45
7 10047–11558 1512 mobilization protein [S. kunkelii] 0
8 12338–12988 651 hypothetical protein [S. kunkelii] 1 × 10-103
9 270–581 312 Unknown 0.34
10 831–1127 297 Unknown 0.024
* AJ297706, the sequence originally characterized from S. citri BR3-T, is the source of these hits.
Algorithm TMHMM v. 2.0 was used to predict the locations of transmembrane helices and intervening loops in the putative products of traE (Fig. 7), mob and orf4. Although the TraE polypeptide was predicted to contain three transmembrane helices, the third helix was predicted at a lower probability than were the other two. Assuming the presence of three transmembrane helices, the protein was predicted to have the N-terminal region (about 10% of the length of the polypeptide) in the cytosol and almost all of the rest of the protein extracellular.
Figure 7 Predicted locations of transmembrane helices and intervening loops in the putative protein encoded by ORF5 (traE) of pBJS-O. The sequential amino acid positions in the primary sequence of the polypeptide are on the X-axis, while the probability score of each residue for being in a transmembrane helix is on the Y-axis in red. The blue and pink curves denote the probability of each amino acid in the sequence to be cytosolic or extracellular, respectively. In the schematic representation of the protein domains at the top, blue lines show the cytosolic portions, purple ones denote the extracellular portions and the thick horizontal bars denote the predicted transmembrane portions of the polypeptide, respectively.
Plasmid pSKU146 from S. kunkelii CR2-3X encodes the S. kunkelii homolog of SARP1, Sk-ARP1. In addition to sk-arp1, pSKU146 contains 17 ORFs. The pSKU146-ORFs having counterparts on pBJS-O were listed above. However, although both plasmids contain genes encoding the ParA-like protein, Soj, sequences surrounding those genes are more distant from one another than are sequences in any other regions. Further, unlike pSKU146, pBJS-O lacks the conserved oriT region characteristic of the IncP group of bacterial plasmids. Also, we were unable to identify a region in pBJS-O resembling a known plasmid origin of transfer.
Discussion
In the present study we report isolation, distribution and structural characterization of pBJS-O, an indigenous S. citri BR3-3X plasmid. We also present evidence that pBJS-O harbors arp1, the gene encoding SARP1, and describe the presence on the BR3-3X chromosome of arp2, an S. citri homolog of arp1. Finally, the sequences of pBJS-O, pBJS-T and the S. kunkelii CR2-3X plasmid, pSKU146, in relation to plasmid evolution are discussed.
Conservation of arp and pBJS-O sequences in Spiroplasma
In Southern hybridizations, the similarity in the hybridization patterns of EcoRI digested versus undigested pBJS-O preparations, despite the presence of two GAATTC recognition sequences, may be due to an adenine methylation system in S. citri. Restriction site modification in S. citri has been reported elsewhere. Rascoe et al. [26] detected multiple bands of S. citri extrachromosomal DNA by Southern blotting, which they attributed to incomplete restriction due to variable restriction site modification in the DNA, and Ye et al. [27] reported protection of an EcoRI site in the S. citri 16S rDNA. Moreover, differential methylation of restriction sites in the RF of the spiroplasma virus, SVTS2, allowed Sha et al. [28] to clone the full-length DNA.
S. citri BR3-3X showed probe-reactive sequences in both the chromosomal and extrachromosomal DNA fractions. However, that the patterns of hybridization of the two fractions differed significantly from each other demonstrates that the two fractions of BR3-3X DNA were not appreciably cross-contaminated. Sequence analyses of DNA from the two fractions showed that, in BR3-3X, arp1 resides on pBJS-O and arp2 on the chromosome. Hybridization of S. citri ASP-1 and R8A2 plasmid preparations with the pBJS-O probe (Fig. 3B), but not with the arp1 probe (Fig. 3A), indicates that each of these two strains contained a plasmid related to pBJS-O, which differed from pBJS-O in lacking arp1. Although S. citri ASP-1 and R8A2 were originally derived from the same parent strain, both have undergone extensive cultivation in vitro since their first isolation, which may have contributed to the differences between their plasmids and pBJS-O. Moreover, the differences in the maintenance regimes of the various spiroplasmas tested may have contributed to the evolution of their plasmids. In this paper we could not correlate pBJS-O and pBJS-O like sequences with either transmissibility or phytopathogenicity of the spiroplasmas tested. However, it is still hypothesized that SARP1 is involved in S. citri transmission by the insect vector.
Frequent chromosomal rearrangements such as inversions and deletions, leading to genome instability, have been reported in spiroplasmas, such as in the lines derived from S. citri strain BR3 [[29] and [30]]. In the present study, we detected a 0.4 kbp deletion in pBJS-T relative to pBJS-O. Unlike BR3-3X, which was stored frozen, S. citri BR3-T was maintained for several years in turnip plants via transmission by the natural insect vector C. tenellus, possibly leading to the sequence differences between pBJS-T and pBJS-O. A recombinational chromosomal rearrangement is indicated by the 5'-sequence differences between arp1 and arp2 reported above.
Recombination likely also played a role in the generation of pBJS-O like plasmids. The gene organization on pBJS-O is similar to that of the recently characterized IncP-like S. kunkelii CR2-3X plasmid, pSKU146. Yet, the two plasmids have substantially different sequences in the region including the soj-like gene in both plasmids and the IncP oriT sequence in pSKU146. Highly similar sequences in the remainder of the two plasmids suggest that recombination events have occurred during the generation of one or both plasmids.
Phage sequences have been implicated in many recombination events in spiroplasmas. Only a short region with similarity to a phage gene was found in pBJS-O. However, the observation of strong stops to sequencing reactions in the region of nucleotides 1 to 80 is reminiscent of a strong stop encountered during the sequencing of the SVTS2 phage [31]. This strong stop was attributed to potential secondary structure putatively involved in phage packaging. It is, thus, possible that pBJS-O has some phage-like properties.
pBJS-O genes
As mentioned above, ORF3, encoding SARP1, and adjacent ORFs [GenBank:AJ297706], had been cloned and characterized from S. citri BR3-T [15]. ORF3 was flanked downstream by a partial ORF (ORF4) having no known homologs. Upstream, ORF3 was flanked by ORF2, encoding a hypothetical protein with no similarity to any known protein, and ORF1, a partial ORF encoding a putative homolog of a ParA-like protein, Soj, which oscillates from pole to pole [32] and is important for chromosome partitioning in Bacillus subtilis [33]. In this study, the putative protein product of orf4 was predicted to contain eight transmembrane helices. Due to a 0.4 kb deletion in the derivation of pBJS-T, orf4 is possibly a part of the same transcription unit as arp1 and traE in this strain. In BR3-3X, arp1 and orf4 are separated by 281 bp, suggesting that they are transcribed separately. Consistent with different translational constraints on this region in BR3-T and BR3-3X, this region contains a large proportion of the differences between the lines. The translation start site of traE was predicted to be ten nucleotides upstream of the orf4 translation stop site.
Consistent with the observation of Bai et al. [22], putative products of the other pBJS-O ORFs, traE [34] and mob [35], are homologous to proteins that are components of the bacterial type IV secretion system involved in conjugative DNA transfer. Members of the TraE family of proteins are thought to form pili that, in addition to conjugation, are involved in processes like virus infection and biofilm formation. Bai et al. [22] reported the presence of three conserved transmembrane helices in four TraE homologs that they characterized from S. kunkelii M2. Ozbek et al. [36], in their transmission electron micrographs, reported the presence of structures resembling fimbriae and pili in S. kunkelii and Bai et al. [22] considered whether the structures may be involved in conjugation. Bové [37] reported that rod-shaped spiroplasma viruses, approximately 230–280 by 10–15 μm in size, can also be surface-associated. Because they can attach perpendicularly to the host membrane at their tips [[38] and [39]], they might resemble the structures reported as pili/fimbriae. In the putative TraE homolog reported here, unlike its S. kunkelii counterpart, two transmembrane helices were predicted at high probability and a third one at moderate probability. Should the third not actually be a transmembrane helix, the ATP binding site would be located intracellularly rather than extracellularly.
pBJS-O gene organization and evolution
Unlike pSKU146, pBJS-O was found to lack the conserved oriT region characteristic of the IncP group of plasmids. We were also unable to identify a region in pBJS-O resembling any other known plasmid origins of transfer, suggesting that pBJS-O belongs to a hitherto unidentified group of plasmids. Horizontal transfer of a promiscuous plasmid, such as an IncP plasmid, between phylogenetically related and unrelated bacteria would help the hosts quickly adapt to different niches [40]. It is possible that an IncP-like plasmid was acquired by the ancestor of S. citri and S. kunkelii. The plasmid may have co-evolved with the host chromosomes after the divergence of the two species, leading to the emergence of pBJS-O and pSKU146, respectively, and to the adaptation of the pathogens to phylogenetically distinct leafhopper vectors and plant hosts.
Future directions
Molecular genetic tools such as cloning and transposon-mediated mutagenesis are available for the study of mollicutes [41]. Cloned genes were expressed in S. citri GII-3 using artificial plasmids based on the S. citri chromosomal oriC [[42,43] and [44]]. However, those plasmids tend to integrate into the S. citri chromosome. When pCJ32, a derivative of the oriC plasmid pBOT1, containing an internal fragment of the gene scm1 (a motility-related S. citri gene), was transformed into S. citri GII-3 cells it successfully integrated into the host chromosome by homologous recombination and disrupted scm1, resulting in non-motile S. citri GII-3 mutants [45]. However, attempts to use pBOT1 in S. citri BR3-3X have been unsuccessful (F. Ye, unpublished data), possibly due to the incompatibility of the plasmid with the host. The indigenous S. citri BR-3X plasmid, pBJS-O, will help us develop a better vector for genetic manipulation not only in S. citri BR3-3X but also in other spiroplasmas.
Conclusion
We have shown that the S. citri BR3-3X plasmid, pBJS-O, encodes the putative adhesin SARP1. This is the first report of an S. citri plasmid encoding a putative adhesin. We have further shown that the arp1-like gene, arp2, resides on the BR3-3X chromosome. The indigenous S. citri BR3-3X plasmid, pBJS-O, will be useful for the development of a better vector for genetic manipulation not only in S. citri BR3-3X but also in other spiroplasmas. Our data also suggest that pBJS-O is a novel S. citri plasmid that does not belong to any known plasmid incompatibility group. The differences between pBJS-O and pSKU146 suggest that recombination has contributed to the divergence of the two plasmids.
Methods
Spiroplasmas
S. citri BR3 was isolated from horseradish plants with brittle root disease [7]. S. citri BR3-T, derived from the triply cloned parental isolate (BR3-3X) by repeated transmission in turnips via its insect vector C. tenellus, is insect-transmissible. BR3-M, derived by passage in liquid medium 43 times, is also a transmissible line. The lines BR3-G (maintained in periwinkle plants by graft transmission) and BR3-P (passed in liquid medium more than 130 times) are insect non-transmissible [46]. S. citri R8A2, isolated from citrus in Morocco [47], and its non-helical derivative ASP-1 (both obtained from R.E. Davis, USDA/ARS, Beltsville, MD), are non-transmissible. Also provided by R.E. Davis were S. citri Beni Mellal, originally isolated from C. hematoceps collected in Morocco; S. melliferum TS2, isolated from honeybees and S. floricola 23-6, isolated from a flower surface [48]. S. phoeniceum P40, a gift from G. Gasparich (Towson University, Towson, MD), was originally isolated from periwinkle in Syria [11]. S. kunkelii CR2-3X was isolated by one of us (J. Fletcher) from stunt-diseased corn collected in Costa Rica [49]. All spiroplasmas, except S. kunkelii CR2-3X, were grown in LD8 broth medium [50] at 31°C. The latter was grown in LD8A3 broth medium [51] at 28°C.
Purification of chromosomal and extrachromosomal ds DNAs from spiroplasmas
For Southern blot hybridization and PCR, extrachromosomal double-stranded (ds) DNA of spiroplasma strains was isolated using the QIAprep Spin Miniprep and the QIAGEN Plasmid Mini Kits (Qiagen, Santa Clarita, CA), following the manufacturer's protocols. For primer walking, S. citri BR3-3X extrachromosomal DNA was isolated using a previously published procedure [28]. The isolation of chromosomal DNA from S. citri BR3-3X cells was performed according to Murray and Thompson [52] and using 1.4 M NaCl, 2.5% cetyltrimethylammonium bromide [CTAB], 100 mM Tris-HCl, pH 8.0, and 20 mM EDTA in the extraction buffer.
PCR and sequencing using S. citri BR3-3X plasmid and chromosomal DNAs
To amplify the 3'- and flanking regions of arp genes from S. citri BR3-3X plasmid and chromosomal DNAs, two oligonucleotides were designed, forward (#7686) 5'-AACACTATTTTCACTGCGG-3', from the S. citri BR3-T arp1 sequence (GenBank accession number AJ297706), and reverse (#7960) 5'-TTTTCCATTGTTTTTGTCTCC-3', from the sequence homologous to ORF4 from the plasmid pSKU146 (pSKU146_11; accession number NC_006400). The PCR was carried out in a DNA thermal cycler (MJ Research, Waltham, MA) performing 35 cycles, each of 30 sec at 94°C, 1 min at 42°C and 3 min at 72°C. Reactions were performed separately in a volume of 50 μl containing 2.5 Units Taq polymerase (Promega), 0.20 μM primers, 200 μM of each dNTP, 1.5 mM MgCl2, and 100–150 ng BR3-3X plasmid and ~ 3.5 μg chromosomal DNA. The amplicons were sequenced using ~ 100 ng of each of the PCR products, 10 μM of the same primers used in the PCR in separate reactions by the ABI PRISM BigDye Terminator Cycle Sequencing method (version 1.0, Applied Biosystems, Foster City, CA) with an ABI PRISM 3700 Automated DNA Analyzer (Perkin Elmer Biosystems, Foster City, CA).
Southern blotting
Extrachromosomal DNA of each spiroplasma strain was digested with EcoRI (Life Technologies, Inc.) for 4 h at 37°C. The fragments were separated by electrophoresis on a 0.75% (w/v) agarose gel in 1× TAE running buffer and transferred to Hybond-N+ nylon membranes (Amersham Biosciences, Uppsala, Sweden) according to standard procedures. The blots were subsequently hybridized to Dig-11-UTP-labeled arp1-derived and whole-plasmid probes, labeled using a DIG DNA Labeling Kit (Roche Molecular Biochemicals, Indianapolis, IN), following the manufacturer's instructions. The arp1-derived probe was obtained by PCR, using clone pP89B (an RsaI fragment of S. citri BR3-T genomic DNA; [15]) as template, and primer pair T7 and #7483 (5'-TTTAACATCAACCGAACCC-3'). The probe comprised 657 bp of a DNA segment from S. citri BR3-T (AJ297706; positions 2315–2989) and 72 bp derived from the cloning vector (pBluescript). PCR was carried out in a DNA thermal cycler performing 34 cycles, each of 30 sec at 94°C, 30 sec at 54°C and 1 min at 72°C. Reactions were performed in a volume of 50 μl containing 1 Unit Taq polymerase, 0.25 μM primers, 250 μM of each dNTP, 50 – 100 ng template DNA, and 2.5 mM MgCl2. Hybridizations were performed at 55°C in Church buffer (0.5 M sodium phosphate buffer, pH 7.2, 7% SDS, and 1 mM EDTA) overnight followed by four washes, each of 20 min, at 55°C in washing buffer (40 mM sodium phosphate buffer, pH 7.2, containing 0.1% SDS). Detection of the DIG-labeled probes was performed using a DIG Luminescent Detection Kit (Roche) following the manufacturer's protocol.
Complete nucleotide sequencing of pBJS-O
The sequence AJ297706 was used to design primers to initiate primer walking to completely sequence and characterize the unknown portion of pBJS-O. The sequencing reactions were performed using ~ 1.2 μg of pBJS-O DNA and 40 μM of primers with the ABI PRISM BigDye Terminator Cycle Sequencing method and the ABI PRISM 3700 Automated DNA Analyzer, as mentioned above. The total 134 sequence reads with an average length of 600 bases gave us about 6× coverage of the entire plasmid sequence. The fragments were assembled from the trace files using the software package PipeOnline 2.0 [53]. Physical gaps in the sequence were closed by PCR and cloning of the products into vector pGEM-T (Promega). The clones were sequenced using primers T7 and SP6. The consensus sequence of the final assembly was annotated using the BLASTX search program [54] and the ORF Finder tool at NCBI, in which a minimum length of 100 bases was used for the nucleotide sequence of a putative ORF. The nucleotide and amino acid sequence analysis tools offered by the Biology Workbench at the San Diego Supercomputer Center, such as ClustalW and PHYLIP for generating the unrooted phylogenetic tree of the S. citri arp sequences, were used to further analyze the plasmid and the polypeptide sequences. BLASTN and BLASTP searches were carried out to find out relationships with the closest homologs. S. kunkelii CR2-3X genome sequence data were accessed and BLAST searches were performed at the Spiroplasma Genome Sequencing Project Web site mentioned above.
Authors' contributions
BDJ performed isolation, distribution and sequence characterization of pBJS-O. JR carried out pBJS-O sequencing and assisted BDJ in primer design and sequence assembly. MB performed S. citri BR3-T arp2 gene cloning and sequencing, and also assisted BDJ in pBJS-O distribution experiments. BDJ, MB, UM and JF planned the research, BDJ and UM wrote the manuscript and MB and JF reviewed it.
Note added in proof
During review of this manuscript sequences of plasmids from a different strain of S. citri were released [GenBank:AJ969069, GenBank:AJ969070, GenBank:AJ969071, GenBank:AJ969072, GenBank:AJ969073, GenBank:AJ969074].
Acknowledgements
We are indebted to Dr. Robert E. Davis of USDA/ARS, Beltsville, MD for personally communicating his work on the plasmid related to pBJS-O from S. kunkelii CR2-3X and sharing with us some of the unpublished data. Staff members of the Recombinant DNA/Protein Resource Facility at OSU are thanked for oligonucleotide synthesis, DNA sequencing, and valuable technical assistance and advice. Dr. Samir Gunjan is thanked for providing technical help in cloning pBJS-O PCR products while filling the gaps in the sequence. The people involved in the S. kunkelii CR2-3X genome sequencing project [B.A. Roe, S.P. Lin, H.G. Jia, H.M. Wu, D. Kupfer, and R.E. Davis] are thanked for making the sequence data publicly available.
We are grateful to Drs. Moses N. Vijaykumar, Richard C. Essenberg and Astri Wayadande for critically reviewing the manuscript. This work was supported by grants from the United States Department of Agriculture, the Robert J. Sirny Professorship to UM, and the Oklahoma Agricultural Experiment Station, whose Director has approved the manuscript for publication.
==== Refs
McCoy RE Caudwell A Chang CJ Chen TA Chiykowski LN Cousin MT Dale JL deLeeuw GTN Golino DA Hackett KJ Kirkpatrick BC Marwitz R Petzold H Sinha RC Sugiura M Whitcomb RF Yang IL Zhu BM Seemûller E Whitcomb RF, Tully JG Plant diseases associated with mycoplasma-like organisms The Mycoplasmas 1989 5 New York: Academic Press 545 640
Garnier M Foissac X Gaurivaud P Laigret F Renaudin J Saillard C Bové JM Mycoplasmas, plants, insect vectors: a matrimonial triangle C R Acad Sci Paris 2001 324 923 928 11570280
Woese CR Bacterial evolution Microbiol Rev 1987 51 221 271 2439888
Whitcomb RF Tully JG Bové JM Saglio P Spiroplasmas and acholeplasmas: multiplication in insects Science 1973 182 1251 1253 4796196
Saglio P Lhospital M Lafleche D Dupont G Bové JM Tully JG Freundt EA Spiroplasma citri gen. and sp. n.: a Mycoplasma-like Organism Associated with "Stubborn" Disease of Citrus Int J Syst Bacteriol 1973 23 191 204
Calavan EC Bové JM Whitcomb RF, Tully JG Ecology of Spiroplasma citri The Mycoplasmas 1989 5 New York: Academic Press 425 485
Fletcher J Schultz GA Davis RE Eastman CE Goodman RE Brittle root disease of horseradish: Evidence for an etiological role of Spiroplasma citri Phytopathology 1981 71 1073 1080
Chen TA Liao CH Corn stunt spiroplasma: isolation, cultivation, and proof of pathogenicity Science 1975 188 1015 1017
Williamson DL Whitcomb RF Plant mycoplasmas: a cultivable spiroplasma causes corn stunt disease Science 1975 188 1018 1020
Whitcomb RF Chen TA Williamson DL Liao C Tully JG Bové JM Mouches C Rose DL Coan ME Clark TB Spiroplasma kunkelii sp. nov.: Characterization of the Etiological Agent of Corn Stunt Disease Int J Syst Bact 1986 36 170 178
Saillard C Vignault JC Gadeau A Carle P Garnier M Fos A Bové JM Tully JG Whitcomb RF Discovery of a new plant-pathogenic spiroplasma Isr J Med Sci 1984 20 1013 1015 6511310
Bové JM Renaudin J Saillard C Foissac X Garnier M Spiroplasma citri, a plant pathogenic mollicute: relationships with its two hosts, the plant and the leafhopper vector Annu Rev Phytopathol 2003 41 483 500 12730387 10.1146/annurev.phyto.41.052102.104034
Steiner T McGarrity GJ Bové JM Phillips DM Garnier M Insect cell cultures in the study of attachment and pathogenicity of spiroplasmas and mycoplasmas Ann Microbiol (Paris) 1984 135A 47 53 6424527
Yu J Wayadande AC Fletcher J Spiroplasma citri surface protein P89 implicated in adhesion to cells of the vector Circulifer tenellus Phytopathology 2000 90 716 722
Berg M Melcher U Fletcher J Characterization of Spiroplasma citri adhesion related protein SARP1, which contains a domain of a novel family designated sarpin Gene 2001 275 57 64 11574152 10.1016/S0378-1119(01)00655-2
Barber CE Archer DB Daniels MJ Molecular biology of spiroplasma plasmids Yale J Biol Med 1983 56 777 781 6382830
Gasparich GE Hackett KJ Clark EA Renaudin J Whitcomb RF Occurrence of extrachromosomal deoxyribonucleic acids in spiroplasmas associated with plants, insects, and ticks Plasmid 1993 29 81 93 8469723 10.1006/plas.1993.1011
Mouches C Barroso G Gadeau A Bové JM Characterization of two cryptic plasmids from Spiroplasma citri and occurrence of their DNA sequences among various spiroplasmas Ann Microbiol Paris 1984 135A 17 24 6712061
Archer DB Best J Barber C Isolation and Restriction Mapping of a Spiroplasma Plasmid J Gen Microbiol 1981 126 511 514
Fletcher J Shaw ME Baker GR Dugan KJ Ye F Sha Y Zuck PD Myers GD Molecular characterization of Spiroplasma citri BR3 lines that differ in transmissibility by the leafhopper Circulifer tenellus Can J Microbiol 1996 42 124 131
Melcher U Fletcher J Genetic variation in Spiroplasma citri Eur J Plant Pathol 1999 105 519 533 10.1023/A:1008757716803
Bai X Fazzolari T Hogenhout SA Identification and characterization of traE genes of Spiroplasma kunkelii Gene 2004 336 81 91 15225878 10.1016/j.gene.2004.03.022
Davis RE Dally EL Jomantiene R Zhao Y Roe B Lin SP Shao J Cryptic plasmid pSKU146 from the wall-less plant pathogen Spiroplasma kunkelii encodes an adhesin and components of a type IV translocation-related conjugation system Plasmid 2005 53 179 190 15737404 10.1016/j.plasmid.2004.09.002
Fleury B Bergonier D Berthelot X Peterhans E Frey J Vilei EM Characterization of P40, a cytadhesin of Mycoplasma agalactiae Infect Immun 2002 70 5612 5621 12228289 10.1128/IAI.70.10.5612-5621.2002
Gasparich GE Spiroplasmas: evolution, adaptation and diversity Front Biosci 2002 7 d619 640 11861210
Rascoe J Melcher U Fletcher J Evaluation of variations in gene presence and expression among lines of Spiroplasma citri using arbitrarily-primed PCR of cDNA Phytopathology 1996 86 S96
Ye F Laigret F Carle P Bové JM Chromosomal heterogeneity among various strains of Spiroplasma citri Int J Syst Bact 1995 45 729 734
Sha Y Melcher U Davis RE Fletcher J Resistance of Spiroplasma citri lines to the virus SVTS2 is associated with integration of viral DNA sequences into host chromosomal and extrachromosomal DNA Appl Environ Microbiol 1995 61 3950 3959 16535161
Melcher U Sha Y Ye F Fletcher J Mechanisms of spiroplasma genome variation associated with SpV1-like viral DNA inferred from sequence comparisons Microb Comp Genomics 1999 4 29 46 10518300
Ye F Melcher U Rascoe JE Fletcher J Extensive chromosome aberrations in Spiroplasma citri strain BR3 Biochem Genet 1996 34 269 286 8894049
Sha Y Melcher U Davis RE Fletcher J Common elements of spiroplasma plectroviruses revealed by nucleotide sequence of SVTS2 Virus Genes 2000 20 47 56 10766306 10.1023/A:1008108106916
Quisel JD Lin DC Grossman AD Control of development by altered localization of a transcription factor in B. subtilis Mol Cell 1999 4 665 672 10619014 10.1016/S1097-2765(00)80377-9
Ireton K Gunther IV NW Grossman AD spoOJ is required for normal chromosome segregation as well as the initiation of sporulation in Bacillus subtilis J Bacteriol 1994 176 5320 5329 8071208
Winans SC Walker GC Conjugal transfer system of the N incompatibility plasmid pKM101 J Bacteriol 1985 161 402 410 2857162
Clewell D Helinski D Supercoiled circular DNA-protein complex in Escherichia coli: purification and induced conversion to an open circular DNA form Proc Natl Acad Sci 1969 62 1159 1166 4894690
Ozbek E Miller SA Meulia T Hogenhout SA Infection and replication sites of Spiroplasma kunkelii (Class: Mollicutes) in midgut and Malpighian tubules of the leafhopper Dalbulus maidis J Invert Pathol 2003 82 167 175 10.1016/S0022-2011(03)00031-4
Bové JM Wall-less prokaryotes of plants Ann Rev Phytopathol 1984 22 361 396 10.1146/annurev.py.22.090184.002045
Townsend R Markham PG Plaskitt KA Multiplication and morphology of Spiroplasma citri in the leafhopper Euscelidius variegatus Ann Appl Biol 1977 87 307 313
Sha Y-S Molecular characterization of spiroplasma viruses and the mechanism of resistance of Spiroplasma citri lines to infection by the virus SVTS2 PhD thesis 1993 Oklahoma State University, Entomology and Plant Pathology Department
Nojiri H Shintani M Omori T Divergence of mobile genetic elements involved in the distribution of xenobiotic-catabolic capacity Appl Microbiol Biotechnol 2004 64 154 174 14689248 10.1007/s00253-003-1509-y
Renaudin J Razin S, Herrmann R Extrachromosomal elements and gene transfer Molecular Biology and Pathogenicity of Mycoplasmas 2002 New York: Kluwer Academic Publishers/Plenum Press 347 370
Renaudin J Bové JM Razin S, Tully JG Plasmid and viral vectors for gene cloning and expression in Spiroplasma citri Molecular and Diagnostic Procedures in Mycoplasmology 1995 New York: Academic Press 167 178
Renaudin J Marais A Verdin E Duret S Foissac X Laigret F Bové JM Integrative and free Spiroplasma citri oriC plasmids: expression of the Spiroplasma phoeniceum spiralin in Spiroplasma citri J Bacteriol 1995 177 2800 2877
Ye F Renaudin J Bové JM Laigret F Cloning and sequencing of the replication origin (oriC) of the Spiroplasma citri chromosome and construction of autonomously replicating artificial plasmids Curr Microbiol 1994 29 23 29 7764984 10.1007/BF01570187
Duret S Danet JL Garnier M Renaudin J Gene disruption through homologous recombination in Spiroplasma citri: an scm1-disrupted motility mutant is pathogenic J Bacteriol 1999 181 7449 7456 10601200
Wayadande AC Fletcher J Transmission of Spiroplasma citri lines and their ability to cross gut and salivary gland barriers within the leafhopper vector Circulifer tenellus Phytopathology 1995 85 1256 1259
Saglio P Lafleche D Bonissol C Bové JM Culture in vitro des mycoplasmes associés au stubborn des agrumes et leur observation au microscope électronique C R Acad Sci Paris 1971 272 1387 1390
Davis RE Lee IM Basciano LK Spiroplasmas: serological grouping of strains associated with plants and insects Can J Microbiol 1979 25 861 866 93504
Zhao Y Hammond RW Jomantiene R Dally EL Lee IM Jia H Wu H Lin S Zhang P Kenton S Najar FZ Hua A Roe BA Fletcher J Davis RE Gene content and organization of an 85-kbp DNA segment from the genome of the phytopathogenic mollicute Spiroplasma kunkelii Molec Genet Genom 2003 269 592 602 10.1007/s00438-003-0878-3
Chen TA Davis RE Whitcomb RF, Tully JG Cultivation of spiroplasmas The Mycoplasmas 1979 New York: Academic Press 65 79
Lee I-M Davis RE Serum-free media for cultivation of spiroplasmas Can J Microbiol 1989 35 1092 1099
Murray MG Thompson WF Rapid isolation of high molecular weight plant DNA Nucleic Acids Res 1980 8 4321 4325 7433111
Ayoubi P Jin X Leite S Liu X Martajaja J Abduraham A Wan Q Yan W Misawa E Prade RA PipeOnline 2.0: automated EST processing and functional data sorting Nucleic Acids Res 2002 30 4761 4769 12409467 10.1093/nar/gkf585
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
S. kunkelii Genome Project
ORF Finder
Biology Workbench
|
16336638
|
PMC1318496
|
CC BY
|
2021-01-04 16:32:47
|
no
|
BMC Genomics. 2005 Dec 7; 6:175
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-175
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1637058310.1371/journal.pmed.0030049PerspectivesBioethicsOtherScience PolicyEpidemiology/Public HealthHealth PolicyMedical EthicsPrimary CareSocioeconomic determinants of healthResearch MethodsHealth services researchEthicsHealth PolicyWhy Are Ethnic Minorities Under-Represented in US Research Studies? PerspectiveSheikh Aziz Aziz Sheikh is Professor of Primary Care Research and Development, Division of Community Health Sciences, General Practice Section, University of Edinburgh, Edinburgh, United Kingdom. E-mail: [email protected]
Competing Interests: AS is Principal Investigator on an Asthma UK project (05/025) investigating barriers and facilitators to recruiting British South Asians into asthma studies.
2 2006 27 12 2005 3 2 e49Copyright: © 2006 Aziz Sheikh.2006This 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.
Are Racial and Ethnic Minorities Less Willing to Participate in Health Research?
Sheikh discusses a new study that found that the main barrier to the participation of minority ethnic people in research lies in their reduced likelihood of being invited to participate.
==== Body
Most economically developed nations are now multiethnic, and, given current demographic trends, there is reason to believe that societies will continue to become more ethnically and culturally diverse. For example, the 1991 and 2001 UK censuses, which both included a mandatory question on ethnic identity, revealed that the proportion of the UK population classifying themselves as belonging to a non-white minority ethnic group increased by 53% over this 10-year period, from 3 million to 4.6 million (or 7.9% of the UK population) [1].
We have more than two decades of research highlighting ethnic inequalities for a range of long-term disorders [2], such as asthma (Table 1), but despite the policy imperative to improve health outcomes for marginalised populations, there has, unfortunately, been little progress toward this end [3,4]. Perversely, data indicate that for some conditions these health inequalities may actually be increasing.
Table 1 Pooled Risk of Admissions for Asthma for South Asian Children and Blacks and South Asians of All Ages, Compared with Whites
Gopalakrishnan Netuveli and colleagues systematically reviewed the literature to look for evidence of ethnic variations in the UK for asthma frequency, morbidity, and health-services use [18]. The table, derived from their study data, shows that South Asian children had an increased risk of admission, and that compared with whites, South Asians and blacks of all ages had a greater risk of admission.
aRaw data unavailable.
CI, confidence interval; OR, odds ratio.
Why this is the case is almost certainly dependent on an array of complex socio-economic factors [5]. Hampering efforts to reverse these trends is the lack of long-term investment into researching the health needs of minority ethnic communities and, as is increasingly being shown, evidence of their systematic under-representation in research studies in general. This lack of investment and under-representation are concerning as it may reasonably be argued that greater resources and effort should be directed toward researching those sections of society that have the greatest capacity to benefit from such research. A study in this month's PLoS Medicine by David Wendler and colleagues investigates one possible source of under-representation—the willingness of ethnic minorities to participate in health research [6].
Ethnic Minorities' Willingness to Participate in Research
Previous work has shown that reporting of the ethnic profile of research participants in trials and other studies has been poor in both the US and the UK [7]. This poor reporting almost certainly reflects an underlying under-representation of these communities in these studies [8–10]. What is notable, however, is that the blame for this under-representation has typically been placed firmly at the feet of the marginalised. It is argued, often without any strong supporting evidence, that the minority ethnic groups in question either fail to understand the importance of the research process or are unable to participate because of language barriers. According to this argument, even if minority groups can comprehend the nature of the research and are able to participate, they may distrust it to the extent that they decline to participate [11,12]. Framed in such a manner, the answer to under-recruitment is seen to lie in greater integration of minority groups to the values of the majority, and the policy and research imperative is therefore to find ways of influencing the attitudes and practices of the minority communities in question.
But Wendler and colleagues' study provides strong empirical evidence to challenge the assumptions that have to date dominated discussions in this area [6]. In their systematic review, the authors set out to address the question of whether individuals from minority groups who are invited to participate in health research are less likely to consent to participate than non-minority individuals. They identified 20 health research studies that reported consent rates by race or ethnicity, 18 of which were single-site studies conducted exclusively in the US or multi-site studies where most of the sites were in the US. They found that when approached to participate, minority ethnic communities in the US are on the whole no less likely, and possibly even more likely than non-Hispanic whites, to agree to participate in research studies. Their work carefully teases out that the main barrier to the participation of ethnic minorities lies in their reduced likelihood of being invited to participate. This work thus places the burden of responsibility not on the marginalised, but on the research community: funders, ethics committees, and researchers alike.
Strengths and Weaknesses of the Study
The key strength of this work is the rigorous systematic review methodology used to identify studies and extract and summarise data. Its main limitation relates to the failure to contact authors who collected, but did not publish, relevant data on ethnicity and consent rates. Contacting authors in this way to obtain additional data that might not be published, and also in an attempt to uncover additional unpublished material, is standard practice in most rigorously conducted systematic reviews. In addition, the fact that this work confines its focus to the US situation renders it difficult to know to what extent the findings may be generalised beyond the US experiences. As the authors point out in their discussion, the US does not guarantee universal access to health care, and perhaps individuals from ethnic minority groups may be more likely than non-Hispanic whites to use participation in research as a way to obtain access to physicians and health care.
Next Steps
The findings from this study clearly have important and wide-ranging implications for the US research community (and possibly elsewhere as well). Funders must, for example, appreciate that to meaningfully involve ethnic minority groups in health research carries financial costs. For example, inviting individuals from these groups to participate in a study, and ensuring that they fully understand what participation involves, requires the use of interpreters and the generation of translated materials about the study—both of which are costly. Also, if sub-group analyses by ethnic groups are considered important, this will typically require considerable inflation of the sample sizes needed, thereby also increasing costs [13]. Similarly, ethics committees need to appreciate that differences in ethical values and practices across different ethnic groups need to be understood and not ignored. For example, insisting on written consent from people originating from an oral culture may unnecessarily hinder recruitment to a study; voice recordings of the consent procedure in such cases should be deemed sufficient. And as for researchers, there is a need for a better appreciation of where minority ethnic populations are located and how they are structured to allow cost-efficient recruitment and retention strategies to be developed.
There are now sufficient examples of studies on marginalised communities that clearly show that it should really be possible to engage with people, irrespective of their ethnic background, and encourage them to participate in research that is ultimately in their and/or their community's best interests. What is now needed is less blame directed at already marginalised people. Instead, those with the power to change the way in which research is conducted should translate the important insights provided by Wendler and colleagues' study into significantly more invitations extended to minority ethnic and racial groups to participate in the research endeavour.
Citation: Sheikh A (2006) Why are ethnic minorities under-represented in US research studies? PLoS Med 3(2): e49.
==== Refs
References
National Statistics Online Population size 2005 Available: http://www.statistics.gov.uk/cci/nugget.asp?id=273 . Accessed 14 November 2005
Black D Morris J Smith C Townsend P Inequalities in health: Report of a research working group 1980 London Department of Health and Social Security
Acheson D Barker D Chambers J Graham H Marmot M Independent inquiry into inequalities in health report 1998 London The Stationery Office
Nazroo J Gordon D Shaw M Dorling D Smith GD Ethnic inequalities in health Inequalities in health 1999 Bristol The Policy Press 155 169
Modood T Berthoud R Lakey J Nazroo J Smith P Ethnic minorities in Britain: Diversity and disadvantage 1997 London Policy Studies Institute
Wendler D Kington R Madans J Van Wye G Christ-Schmidt H Are racial and ethnic minorities less willing to participate in health research? PLoS Med 2006 3 e19 10.1371/journal.pmed.0030019 16318411
Sheikh A Netuveli G Kai J Panesar SS Comparison of reporting of ethnicity in US and European randomised controlled trials BMJ 2004 329 87 88 15138158
Mason S Hussain-Gambles M Leese B Atkin K Brown J Representation of South Asian people in randomised clinical trials: Analysis of trials' data BMJ 2003 326 1244 1245 12791739
Svensson CK Representation of American Blacks in clinical trials of new drugs JAMA 1989 261 263 265 2909024
Prescott RJ Counsell CE Gillespie WJ Grant AM Russell IT Factors that limit the quality, number and progress of randomised controlled trials Health Technol Assess 1999 3 1 143 10683591
Harris Y Gorelick PB Samuels P Bempong I Why African Americans may not be participating in clinical trials J Natl Med Assoc 1996 88 630 634 8918067
Corbie-Smith G The continuing legacy of the Tuskegee Syphilis Study: Considerations for clinical investigation Am J Med Sci 1999 317 5 8 9892266
Sheikh A Panesar SS Lasserson T Netuveli G Recruitment of ethnic minorities to asthma studies Thorax 2004 59 634 15223878
Ayres JG Acute asthma in Asian patients: Hospital admissions and duration of stay in a district with a high immigrant population Br J Dis Chest 1986 80 242 248 3790414
Myers P Ormerod LP Increased asthma admission rates in Asian patients: Blackburn 1987 Respir Med 1992 86 297 300 1448584
Jackson SHD Bannan LT Beevers DG Ethnic differences in respiratory diseases Postgrad Med J 1981 57 777 778
Gilthorpe MS Lay-Yee R Wilson RC Walters S Griffiths RK Variations in hospitalization rates for asthma among black and minority ethnic communities Respir Med 1998 92 642 648 9659530
Netuveli G Hurwitz B Levy M Fletcher M Barnes G Ethnic variations in UK asthma frequency, morbidity, and health-service use: A systematic review and meta-analysis Lancet 2005 365 312 317 15664226
|
16370583
|
PMC1322037
|
CC BY
|
2021-01-05 11:14:05
|
no
|
PLoS Med. 2006 Feb 27; 3(2):e49
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0030049
|
oa_comm
|
==== Front
BMC BiochemBMC Biochemistry1471-2091BioMed Central London 1471-2091-6-271632116010.1186/1471-2091-6-27Research ArticleMultiple phosphorylation events control mitotic degradation of the muscle transcription factor Myf5 Doucet Christine [email protected] Gustavo J [email protected] Catherine [email protected] Thierry [email protected] Gwendaline [email protected] Christian [email protected] Olivier [email protected] Centre de Recherches de Biochimie Macromoléculaire (CRBM), CNRS FRE 2593, Montpellier, France2 Present address: Burnham Institute for Medical Research, La Jolla, CA, USA3 Wellcome Trust/Cancer Research UK, Gurdon Institute, Cambridge, UK4 Celogos/Institut Pasteur, Paris, France2005 1 12 2005 6 27 27 5 8 2005 1 12 2005 Copyright © 2005 Doucet et al; licensee BioMed Central Ltd.2005Doucet 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 two myogenic regulatory factors Myf5 and MyoD are basic helix-loop-helix muscle transcription factors undergoing differential cell cycle dependent proteolysis in proliferating myoblasts. This regulated degradation results in the striking expression of these two factors at distinct phases of the cell cycle, and suggests that their precise and alternated disappearance is an important feature of myoblasts, maybe connected to the maintenance of the proliferative status and/or commitment to the myogenic lineage of these cells. One way to understand the biological function(s) of the cyclic expression of these proteins is to specifically alter their degradation, and to analyze the effects of their stabilization on cells. To this aim, we undertook the biochemical analysis of the mechanisms governing Myf5 mitotic degradation, using heterologous systems.
Results
We show here that mitotic degradation of Myf5 is conserved in non-myogenic cells, and is thus strictly under the control of the cell cycle apparatus. Using Xenopus egg extracts as an in vitro system to dissect the main steps of Myf5 mitotic proteolysis, we show that (1) Myf5 stability is regulated by a complex interplay of phosphorylation/dephosphorylation, probably involving various kinases and phosphatases, (2) Myf5 is ubiquitylated in mitotic extracts, and this is a prerequisite to its degradation by the proteasome and (3) at least in the Xenopus system, the E3 responsible for its mitotic degradation is not the APC/C (the major E3 during mitosis).
Conclusion
Altogether, our data strongly suggest that the mitotic degradation of Myf5 by the ubiquitin-proteasome system is precisely controlled by multiple phosphorylation of the protein, and that the APC/C is not involved in this process.
==== Body
Background
Terminal differentiation of skeletal muscle is orchestrated by the family of myogenic regulatory factors (MRFs), which contains Myf5, MyoD, myogenin and MRF4 (for review see [1,2]). These factors activate muscle-specific gene transcription, by binding specific DNA sequences (E-boxes) as heterodimers with ubiquitous E2A proteins such as E12 and E47, in cooperation with MEF2 family of MADS-box proteins (reviewed in [3]). They have first been characterized for their ability to convert certain non muscle cells into myoblasts after ectopic expression, a process known as "myogenic conversion" [4,5]. Among MRFs, MyoD and Myf5 are usually considered as "determination factors" since they are required for formation of skeletal muscle [6], and expressed at the proliferating myoblast stage, in contrast to myogenin that is induced as cells undergo cell cycle arrest, and MRF4 that is involved in myotube maturation [7]. However, recent data have shown that MRF4 is also expressed at early stages of muscle development, and can act upstream of MyoD and Myf5 [8].
Interestingly, MyoD bears intimate functional relationships with the cell cycle apparatus (reviewed in [9]): its transcriptional activity and stability are regulated by cyclin/CDK complexes [10-13], and it can repress cell cycle activators by physical interaction [14] or by activating expression of cell cycle inhibitors [15]. Regarding Myf5, despite numerous observations about the regulation of its gene and expression pattern during embryonic development [7,16], little is understood about its functions, probably due in large part to its redundancy with MyoD [6].
However, very intriguingly, in cultured myoblasts (C2 cells or primary myoblasts), intracellular protein levels of Myf5 and MyoD exhibit opposite cell-cycle fluctuations at the proliferative stage [17], with the result that, during the early G1 phase of the cell cycle, MyoD levels are high and Myf5 levels low, whereas the opposite is true at G2. Since fusion of myoblasts into myotubes occurs during G1 phase, an attractive hypothesis is that the MyoD/Myf5 ratio is an important determinant in myoblasts for the decision process between proliferation or differentiation. In this model, MyoD and Myf5 are more than muscle determination factors, and act in myoblasts as regulators of the proliferation/differentiation interface. Possibly connected to this hypothesis, it is interesting to note that, in cell culture, myoblast differentiation results in two distinct populations: a majority of plurinucleated myotubes that contain high levels of MyoD, but no Myf5, and a minority of quiescent cells, assimilated to "reserve cells", that contain Myf5 but no MyoD [17,18].
Several studies have demonstrated that the cell cycle variations of MyoD and Myf5 levels involve their specific and regulated proteolysis by the proteasome. MyoD degradation at the end of the G1 phase is promoted by its phosphorylation by the CyclinE/CDK2 complex [11,12,19]. Regarding the mechanisms of Myf5 accelerated degradation at the G2/M transition and throughout mitosis, much less is known, although it coincides with the phosphorylation of the protein [20], and seems to depend on the integrity of a Destruction-box (D-box) domain [21].
To test the hypothesis that the cyclic and alternated degradation of MyoD and Myf5 is not a simple consequence of proliferation, but an important event in the control of the proliferation/differentiation interface, we decided to specifically interfere with the mitotic degradation of Myf5 in order to gain insights into the biological role of this process. To this aim, the purpose of the present work was to initiate the characterization of the mechanisms regulating Myf5 mitotic degradation. Since in muscle cells the differentiation process and the cell cycle are tightly connected, we chose to use non-muscle experimental systems in which Myf5 mitotic degradation is conserved and can be studied in a cell cycle-dependent but differentiation-independent manner. We show that in Xenopus egg extracts, which are widely used to study mitosis-specific processes, Myf5 mitotic degradation occurs through a ubiquitin- and proteasome-dependent mechanism, in a manner tightly controlled by its phosphorylation, but independent of APC/C activity. Myf5 phosphorylation is itself dependent on the activity of several kinases and phosphatases, and thus depends on a delicate balance between these activities that finely tunes Myf5 stability.
Results
Accelerated degradation of Myf5 during mitosis is cell cycle- but not muscle-dependent
In order to better define the parameters that control Myf5 cyclic degradation, we first asked whether the responsible machinery is uniquely dependent on the cell cycle apparatus, or whether it requires components restricted to myogenic cells. To this end, we derived from HeLa cells a stable line expressing Myf5 under an inducible promoter.
In this cell line, induced Myf5 is easily detected in non-synchronized cells, but is absent or very low in cells arrested in mitosis with nocodazole (figure 1, lanes 1 & 2). Treatment with the proteasome inhibitor MG132 significantly increases Myf5 content in non-synchronized cells (data not shown), but has a dramatic effect on Myf5 expression in metaphase cells (figure 1, lane 3), suggesting that Myf5 degradation, which is already rapid in the other phases of the cell-cycle, is strongly accelerated during mitosis. In addition, the forms of Myf5 that are stabilized upon proteasome inhibition in metaphase-arrested cells are phosphorylated, since treatment of the cell extract with lambda phosphatase abolishes the shift seen after electrophoresis without phosphatase treatment (figure 1, lane 4).
Figure 1 Myf5 mitotic degradation is conserved in non-muscle cells. HeLa-S3 cells expressing Myf5 under an inducible promotor (tetoff) were treated (lanes 2, 3, 4) or not (AS, lane 1) with nocodazole (200 ng/ml) for 16 h, as in [20]. In lanes 3 and 4, cells were treated with a proteasome inhibitor (MG132, 50 μM) for 2 h prior to lysis (as described in Methods). In lane 4, extract was incubated with Lambda phosphatase (New England Biolabs) for 30 min at 30°C. 25 μg of each extract were resolved by 10% SDS-PAGE and immunoblotted with anti-Myf5 antibodies. AS: asynchronous cells; * indicates a non specific band recognized by the antibodies that can be used as a loading control.
These results strictly reproduce the data obtained with similar experiments in myogenic cells, showing that Myf5 is strongly destabilized in mitosis, and is stabilized by proteasome inhibitors as a phosphorylated form [20]. Thus, the machinery responsible for Myf5 mitotic degradation is present in non-myogenic cells, and does not require muscle specific components but only the basal cell cycle machinery present in all dividing cells.
Myf5 degradation is also cell cycle regulated in Xenopus egg extracts
Since Myf5 mitotic degradation is conserved in non-muscle cells, we decided to test whether we could use Xenopus egg extracts as an in vitro system to dissect the mechanisms involved. Indeed, Xenopus egg extracts recapitulate cell cycle events and have been extensively used for biochemical analyses of mitotic processes, particularly mitosis-specific protein degradation [22], since it is easy to prepare large amounts of extracts that reproduce mitotic or interphase conditions [23]. Compared to extracts of synchronized cultured cells, this system presents in addition the important advantage of limiting dilution of intracellular components and thus preserving the relative activities of enzymes such as protein kinases and phosphatases [23].
When radiolabeled in vitro translated Myf5 was incubated in mitotic extract (CSF, see Methods section), it was readily degraded, in clear contrast with the stability of the same protein incubated in interphase extract (figures 2A, 2B). As seen in cells, proteasome inhibitors completely stabilize Myf5 (figure 2D, 2E), which does not occur with inhibitors of other proteases (data not shown). Thus the mechanisms governing mitotic-dependent proteasomal degradation of Myf5 are conserved in Xenopus egg extracts. It is striking to note that Myf5 appears stable in interphase Xenopus extracts, although it is highly unstable in non-synchronized (i.e. mostly interphase) cultured cells [21]. Since in cells, the pathway responsible for Myf5 degradation in interphase is seemingly different from the pathway operating in mitosis [21], one likely explanation is that only the mitotic pathway is active in Xenopus extracts. The fact that cell cycles in early embryos lack G1 phase and that Xenopus egg extracts do not form nuclei unless sperm heads are added could explain the absence of this interphase pathway, if it requires intact nuclei for example. Thus, Xenopus egg extracts appear to be a unique system in which Myf5 mitotic degradation can be analyzed without interference of other (i.e. non-mitotic) pathways, and we decided to pursue our analyses in this system to learn more about the machinery involved specifically in Myf5 mitotic degradation. We note that the mitotic degradation pathway of Myf5 appears distinct from that of cyclin B, which is stable in mitotic extracts (CSF), and degraded only after calcium activation of these extracts [22].
Figure 2 Myf5 is degraded in a cell-cycle dependent manner in Xenopus egg extracts, and its degradation is correlated with changes in its phosphorylation status. (A) Myf5 is phosphorylated and degraded in mitotic (CSF) extracts, but dephosphorylated and stable in interphase extracts. Degradation assays were performed as described in Methods; 2 μl samples were taken at the indicated times and resolved by 10% SDS-PAGE. (B) Degradation of Myf5 was quantified and normalized (start = 100%) : black squares represent the average degradation of Myf5 in interphase extracts (5 independent experiments), open circles the average degradation of Myf5 in CSF extracts (7 independent experiments), and black circles the average degradation of dephosphorylated (deP) Myf5 (see Methods section) in CSF extracts (4 independent experiments). Bars represent standard deviation. For each lane, a region corresponding to all non-phosphorylated and phosphorylated forms of Myf5 was quantified using the ImageQuant software (Molecular Dynamics). The apparent increase of Myf5 level in interphase extracts is probably due to the accumulation of all the initially phosphorylated forms of Myf5 (some of them being close to background) into the fastest migrating form of Myf5. (C) All the slower migrating forms of Myf5 are phosphorylated: lane 1: in vitro translated Myf5; lane 2: in vitro translated Myf5 after 1 h incubation with CSF extract in the presence of 1 μM microcystin LR (this treatment leads to hyperphosphorylation of Myf5 (Myf5PPP), see text and figure 3 for details); lanes 3–6: in vitro translated Myf5 (lanes 3 and 4), or Myf5PPP (lanes 5 and 6) were immunoprecipitated with anti-Myf5 antibodies and incubated in the absence or presence of Lambda phosphatase (λ PPase), as indicated. Samples were resolved by 10% SDS-PAGE and analyzed using a PhosphoImager. (D) A mitotic extract was first incubated with 200 μM MG132, or the same volume of pure DMSO as a control, for 10 min at 25°C; dephosphorylated (deP) Myf5 (see method) was then added to the treated extract. At each time point, 2 μl samples were resolved by 10% SDS-PAGE. (E) Relative amounts of Myf5 in the gel shown in D were quantified as above and reported on the graph.
Myf5 mitotic degradation is controlled by its phosphorylation status
When the fate of Myf5 is analyzed by electrophoresis during incubation in mitotic extracts (figure 2A), there is, concomitant with its gradual disappearance, a clear up-shift in mobility of the protein. By contrast, in interphase extracts, Myf5 is down-shifted and is stable. This strongly suggests that, consistent with the results obtained in mammalian cells, Myf5 mitotic degradation requires its phosphorylation. In view of the heterogeneity of the migration of Myf5 in the gel, we tested whether the different forms correspond to differentially phosphorylated molecules of Myf5. We found that indeed, all the slower migrating forms of Myf5 can be down-shifted to the fastest migrating form upon incubation with lambda phosphatase (figure 2C, lanes 4 and 6). Thus Myf5 degradation is strongly correlated with its extensive phosphorylation.
Interestingly, in extracts degrading Myf5, the protein is gradually shifted to an upper residual form (that we will call thereafter the hyperphosphorylated form of Myf5, see discussion for this terminology) that appears to be stable (figure 2A). By contrast, if Myf5 is dephosphorylated after translation (using alkaline phosphatase bound to agarose beads that can be subsequently easily removed from the reticulocyte lysate), then the protein (deP Myf5) is slower than untreated Myf5 to reach the hyperphosphorylated form when incubated in Xenopus egg extracts (compare figures 2A and 2D) and, as a consequence, can be degraded to a greater extent (figure 2B). These observations suggest that Myf5 degradation both requires phosphorylation of the protein, and is inhibited by hyperphosphorylation. To verify this hypothesis, we tested whether the form accumulating upon prolonged incubation of Myf5 in mitotic extracts (i.e. the hyperphosphorylated form of Myf5) is indeed stable. To this end, we first incubated Myf5 in mitotic extracts treated with the phosphatase inhibitor microcystin LR. In these conditions, Myf5 is rapidly accumulated under its hyperphosphorylated form (Myf5PPP), and is not degraded (figure 3A). We then incubated this hyperphosphorylated form of Myf5 in a fresh mitotic extract, in the presence or absence of microcystin. As seen in figure 3B, in the presence of microcystin, the hyperphosphorylated form of Myf5 remains unchanged and stable. By contrast, in the absence of microcystin, the hyperphosphorylated form of Myf5 is dephosphorylated and degraded, in a proteasome-dependent manner (figures 3C &3D).
Figure 3 Hyperphosphorylated Myf5 is stable. An equal volume of reticulocyte lysate containing in vitro translated Myf5 and of Xenopus mitotic egg extract (CSF) were incubated together for 60 min in the presence of 1 μM microcystin LR. At each time point, 0.5 μl samples were analyzed by 10% SDS-PAGE (panel A). After this initial incubation, 2 μl of the reaction mixture containing hyperphosphorylated Myf5 (Myf5PPP) were further incubated in 18 μl of fresh mitotic extract, either in the presence of 1 μM microcystin (panel B) or without inhibitor addition (panel C), or without microcystin and in the presence of the proteasome inhibitor MG132 (200 μM, panel D). At each time point, 2 μl samples were analyzed by 10% SDS-PAGE.
Altogether, our results strongly support the following model: degradation of Myf5 is triggered by its phosphorylation, that occurs in multiple sites, but there is a kinetic competition between phosphorylation and degradation such that, if Myf5 reaches its hyperphosphorylated form, it escapes the degradation machinery. In extracts where both kinases and phosphatases modifying Myf5 are active, such as Xenopus mitotic egg extracts, the equilibrium between both types of activity permits some dephosphorylation (and thus subsequent degradation) of Myf5 after it has reached its hyperphosphorylated state, explaining the gradual disappearance of the hyperphosphorylated form seen in figure 2A. By contrast, if phosphatases are inhibited, then Myf5 remains hyperphosphorylated and is strongly stabilized (figures 3A, B).
The differences observed in the phosphorylation pattern of Myf5 between figures 2A and 2D show that, in mitotic Xenopus egg extracts, hyperphosphorylation of Myf5 is accelerated by the intervention of active kinases present in the reticulocyte lysate in which Myf5 is translated. However, kinase(s) triggering hyperphosphorylation of Myf5 also exist(s) in mitotic Xenopus egg extracts, since dephosphorylated Myf5 rapidly shifts to its hyperphosphorylated form if the extract is treated with the phosphatase inhibitor microcystin (data not shown). This indicates that the stability of Myf5 is tightly dependent upon the kinase/phosphatase equilibrium of the extracts, and suggests that in cells, multiple mechanisms – including perhaps hyperphosphorylation – may be used to control Myf5 stability under different biological conditions.
Several signaling pathways are likely to be involved in the control of Myf5 stability
An important question is the nature of the kinase(s) responsible for Myf5 mitotic degradation. An evident candidate is the CDK1 kinase, which is the hallmark of mitotic status, since its activity is high during mitosis but abruptly decreases upon mitosis exit, due to the degradation of its activator cyclin B [24,25]. Indeed, in vitro translated Myf5 can be directly phosphorylated by CDK1/cyclin B, as seen by the shift of the protein after its incubation with the purified kinase (figure 4A, lanes 1&3). However, if the kinase CDK1 is reactivated in interphase egg extracts by addition of non-degradable cyclin B [26], although phosphorylation of Myf5 resumes in a pattern very similar to that seen in CSF extracts, in most experiments Myf5 is not significantly degraded (data not shown). This suggests that at least one other kinase is required for Myf5 degradation, in addition to CDK1. We observed that Myf5 can nevertheless be degraded in some mitotic extracts resulting from reactivation of CDK1 in interphase extracts (data not shown), pointing to the possible involvement of the MAP kinase family. Indeed, the activity of MAP kinases can vary in interphase Xenopus egg extracts, depending on the protocols used to prepare them, which may or may not entail the degradation and/or inactivation of the upstream MAPKKK c-Mos [27]. We found that the protein kinase Erk (MAPK) is also able to phosphorylate Myf5 in vitro, apparently at different sites from CDK1/cyclin B (figure 4A, lanes 4&5). These results suggest that these two kinases may cooperate in the regulation of Myf5 stability. However, other enzymes may also be involved. For example, we found that Ca2+ – at very low concentrations – is important for Myf5 degradation, suggesting that a calcium-dependent kinase or phosphatase could be involved in the regulation of Myf5 stability. Indeed, as shown in figure 4B, degradation of Myf5 is accelerated in mitotic extracts (that are prepared in the presence of EGTA) supplemented by a low concentration of Ca2+, as compared to degradation in the same extract without supplementation. This reproducible result shows that the concentration of free Ca2+ that exists in the presence of the strong calcium chelator EGTA, albeit extremely low (in the range of 10-7M [28]), is nevertheless significant enough to alter the rate of degradation of Myf5. Interestingly, the activity of the calcineurin phosphatase has been shown to be sensitive to such low concentrations of Ca2+ [29]. This range of calcium concentrations does not trigger cyclin B degradation, which requires a transient increase of the concentration of free calcium from 10-7 to 10-5 M for its activation [28].
Figure 4 Several signaling pathways are likely to control Myf5 mitotic degradation (A) 1 μl of untreated (upper panel) or dephosphorylated (deP, lower panel) Myf5 was incubated with buffer (lane 2), recombinant CDK1/cyclin B (New England Biolabs, 40U, lane 3), recombinant Erk2 (New England Biolabs, 200U, lane 4) or both kinases (lane 5) for 30 min at 30°C. The reaction mix was resolved by 10% SDS-PAGE. "Start" is a non-incubated control. (B) Myf5 degradation is sensitive to very low concentrations of calcium: a mitotic extract was prepared using XB without CaCl2 but supplemented with 6 mM EGTA (see Methods), then incubated with 0.1 mM CaCl2 and 6 mM EGTA, or an equivalent volume of water as a control, for 10 min at 25°C. 2 μl of Myf5 were incubated for 1 hour with 18 μl of treated extract. 2 μl samples were resolved by 10% SDS-PAGE at the beginning and the end of the reaction. Myf5 degradation rate was quantified as previously described; the graph represents mean values of 3 experiments. Bars represent standard deviations.
To define which kinases are required for Myf5 mitotic degradation, we tested a panel of specific inhibitors (see Table 1) targeting the MAP kinase family, the GSK3 kinase, PKC, CDK1/cyclin B, and the calcineurin phosphatase. Unfortunately, we were unable to reproducibly alter Myf5 degradation using these inhibitors, although most of them altered Myf5 phosphorylation status (see Table 1). Overall, the many attempts that we have made to clarify the question of the kinases involved in Myf5 degradation in Xenopus extracts have led us to the conclusion that Myf5 degradation probably involves several kinases, but is so sensitive to the kinase/phosphatase equilibrium that alteration of the activity of many signaling pathways induces complex and contradictory effects on Myf5 stability that cannot be clarified by simple pharmacological approaches.
Table 1 List of kinase and phosphatase inhibitors tested in Xenopus egg extracts for their potential effect on Myf5 degradation These inhibitors were purchased from Sigma-Aldrich, stock solutions were prepared in ethanol or DMSO, following the furnisher instructions.
Inhibitor Action Concentrations tested Effect on Myf5 phosphorylation status (*) References
Indirubin-3'-oxime GSK3β and CDK5 inhibitor 101 to 100 μM Partial inhibition [54, 55]
Roscovitine CDK/cyclin inhibitor 100 μM2 Partial inhibition [56–58]
Staurosporine Broad spectrum kinase inhibitor (among which: PKA, PKG, CaMK, MLCK, PKC) 11 to 10 μM Strong inhibition [59, 60]
U0126 MEK1/2 inhibitor 501 to 100 μM Partial inhibition [61, 62]
GF109203X PKC inhibitor 51 to 50 μM Partial inhibition [63, 64]
Cyclosporin A Calcineurin inhibitor 11 to 10 μg/ml No effect [65, 66]
Microcystin LR Broad spectrum phosphatase inhibitor 1 μM Hyperphosphorylation [42, 67]
1 Usual concentration for treatment of cell cultures
2 Efficiency of this concentration was tested by measuring H1 Kinase activity after treatment.
(*) as judged by electrophoretic migration pattern
In addition, it seems likely that although multiple phosphorylations of Myf5 are required for its degradation, none of the individual phosphorylation sites is absolutely necessary. We investigated this question by mutational analyses of Myf5 in order to identify phosphorylation sites important for its degradation. Since Myf5 possesses about 20% serine/threonine (S/T) residues and many potential phosphorylation sites in its primary sequence, making it difficult to systematically change them all into non-phosphorylable residues, we decided to concentrate our work on S/T-proline (S/TP) motifs, which are potential phosphorylation sites for both CDK1/cyclin B and MAPK family kinases. We mutated into alanines the five serines (S10, S23, S158, S172, S231), and the threonine (T242) that are present in Myf5 S/TP motifs, and analyzed the degradation of these mutants in mitotic egg extracts. Although most mutants showed an altered pattern of phosphorylation when compared to the wild type protein, indicating that the mutated residues are likely to be phosphorylated in extracts, none of them was clearly stabilized (not shown). The fact that the mutant S158A, which was found to be stabilized in interphase but not mitotic mammalian cells [21], is still efficiently degraded in mitotic egg extracts, underlines that only the mitotic degradation of Myf5 is active in these extracts. If it cannot be excluded at present that other phosphorylation sites are the key determinants of Myf5 stability, there is another possible interpretation, that we presently favor in view of the probable involvement of both (at least) CDK1/cyclin B and MAP kinases into Myf5 degradation. That is, in a process analogous to that reported for the CDK inhibitor Sic1 in budding yeast [30,31], it would be the number of phosphorylated residues rather than their identity that controls Myf5 stability. If this hypothesis holds true, then only the mutation of most of these sites would significantly alter the degradation of the protein.
Myf5 degradation requires its prior ubiquitylation
Regulated protein degradation by the proteasome most often requires the prior conjugation on the substrate of a poly-ubiquitin chain that acts as a signal targeting the substrate to the proteasome [32]. Incubation of Myf5 in Xenopus mitotic egg extracts in the presence of proteasome and de-ubiquitylating enzymes inhibitors indeed led to the rapid formation of ubiquitylated forms of Myf5 (figure 5A). To test whether these ubiquitylated forms are obligatory intermediate products in Myf5 degradation, we titrated wild type ubiquitin (Ub) with a mutant form of Ub in which all the lysines have been replaced by arginines (UbK0). Since this lysine-less mutant cannot be itself conjugated by Ub, it acts as a chain terminator if integrated in a Ub-chain [33]. When added to a mitotic extract, this mutant strongly inhibits both Myf5 ubiquitylation (figure 5B) and degradation (figure 5C) in a concentration-dependent manner. Thus efficient poly-ubiquitylation of Myf5 is required for its subsequent degradation, showing that Myf5 is degraded in a ubiquitin- and proteasome- dependent process.
Figure 5 Polyubiquitylation of Myf5 is required for its degradation. (A) Myf5 is polyubiquitylated in mitotic extracts: Myf5 was incubated in a mitotic extract in the presence of ubiquitin (1 mg/ml), MG132 (200 μM) and ubiquitin aldehyde (5 μM). At the times indicated, 2 μl samples were analyzed by 10% SDS-PAGE to visualize the formation of high-MW adducts ((Ub)n-Myf5, poly-ubiquitylated Myf5 molecules). (B) Lysine-less (K0) ubiquitin inhibits Myf5 polyubiquitylation: Myf5 ubiquitylation was performed in a mitotic extract in the presence of MG132 (200 μM), ubiquitin aldehyde (1 μM), ubiquitin (Ub) and lysine-less-ubiquitin (UbK0) at the indicated concentrations, for 30 min at 25°C. (C) Lysine-less ubiquitin stabilizes Myf5: Degradation assays were performed in 20 μl reaction mixtures containing 16 μl of mitotic extract, 2 μl of radiolabeled in vitro translated Myf5 and a mixture of wild type and K0 ubiquitin as indicated. The reaction was performed at 25°C for 40 min, 2 μl samples were resolved by 10% SDS-PAGE. "START" corresponds to a 2 μl sample taken at time 0 from the reaction containing 1 mg/ml Ub. Radioactivity was quantified as previously described (see legend of figure 2) and percentages of degradation were calculated for each Ub/UbK0 ratio from the difference between the times 0 (not shown) and 40 minutes.
The E3 responsible for Myf5 ubiquitylation is not the APC/C
Protein ubiquitylation requires an enzymatic cascade involving 3 types of proteins called E1 (Ub-activating protein), E2 (Ub-carrier protein) and E3 [34]. In this cascade, the E3 component recruits both an E2 and the substrate to favor ubiquitylation of the latter, and thus acts as the specificity factor for the reaction. As a consequence, eukaryotic cells contain a very high number of E3 proteins [35].
However, many substrates degraded by the ubiquitin-proteasome pathway in mitosis (among which the mitotic cyclins and securin) are targeted for ubiquitylation by the E3 ligase APC/C (Anaphase-Promoting Complex / Cyclosome) [36,37]. Ubiquitylation of some of these proteins depends on a loosely conserved Destruction-box (D-box) motif, which is thought to be recognized by the APC/C adaptor proteins Cdc20/Fizzy [25,38] and Cdh1 [39]. A D-box-like motif is present in Myf5 sequence, and substitution of residues into this 9-amino-acid sequence significantly stabilizes the protein during mitosis [21], suggesting that Myf5 could be an APC/C substrate. However, the degradation pathway of Myf5 in mitotic cells appears different with respect of both timing and mechanism from that of known substrates of the APC/C [21]. To elucidate this apparent dilemma, we decided to test in our system whether the APC/C is responsible for Myf5 ubiquitylation.
We first used Xenopus interphase extracts in which APC/C is specifically activated by translation in the extract of the APC/C regulator Cdh1 (Fizzy-related in Xenopus), as these extracts have been shown to be able to efficiently degrade most known APC/C substrates in a Cdh1-dependent manner [39]. As shown in figure 6A, Myf5 is rapidly dephosphorylated and is stable for 1 h in these extracts, although the APC/C substrate Xkid [40], used here as a control, is completely degraded after 30 minutes, as expected. This result strongly suggests that APC/C is not involved in Myf5 degradation. However, Myf5 could be an unusual APC/C substrate requiring to be phosphorylated to be recognized by this E3. If so, then its dephosphorylation in these extracts could explain its stability. To avoid dephosphorylation, we decided to test by immunodepletion approaches whether the presence of APC/C is required for Myf5 ubiquitylation in mitotic eggs extracts. In these extracts, APC/C is active against certain substrates such as cyclin A [41], but its activity towards other substrates such as cyclin B is inhibited by the cytostatic factor (CSF). In order to abolish CSF effect and study a fully active APC/C, we treated a CSF extract with 1 μM microcystin LR [42]. Since in these conditions, as described above, Myf5 degradation is inhibited due to the accumulation of the protein in its hyperphosphorylated form, we fractionated the activated extract on a DEAE column. Both APC/C and E3 acting on Myf5 were eluted with 0.25 M NaCl (figure 6B, lanes 2). Immunodepletion of APC/C, using antibodies directed against its constitutive subunit CDC27, fails to abolish Myf5 ubiquitylation in the 0.25 M NaCl fraction, whereas, in the same conditions, cyclin B is no longer ubiquitylated (figure 6B, lanes 5). Moreover, cyclin B, but not Myf5, is ubiquitylated in the presence of the proteins immunoprecipitated with anti-CDC27 antibodies (figure 6B, lanes 4). Altogether, these results show that, at least in Xenopus egg extracts, the E3 responsible for Myf5 ubiquitylation in mitosis is not the APC/C, but another E3 that remains to be identified.
Figure 6 The E3 responsible for Myf5 ubiquitylation is not the APC/C. (A) In vitro translated Xkid (upper panel) or Myf5 (lower panel) were incubated in 15 μl Xenopus interphase egg extract in which Cdh1 has been translated [39]. After the times indicated, 2.5 μl were analyzed by SDS-PAGE and fluorography. (B) 1 ml of mitotic extract was activated with 1 μM microcystin LR (in order to fully activate APC/C), then fractionated on a DEAE column (see Methods). Bound proteins were eluted in two steps with buffer A containing 0.25 M and 0.5 M NaCl, respectively. Each fraction (lanes 2 & 3) was compared to a control reaction containing buffer (lane 1) for its ability to mediate ubiquitylation of either cyclin B (upper panel) or Myf5 (lower panel). Lanes 4–7: the 0.25 M NaCl eluate that mediates both cyclin B and Myf5 ubiquitylation was then subjected to immunoprecipitations using either anti-Cdc27 or control antibodies. For each immunoprecipitation, both the material bound to the beads or the supernatant were analyzed for ubiquitylation activity using cyclin B (upper panel) or Myf5 (lower panel) as a substrate. (Ub)n-cycB and (Ub)n-Myf5 indicate poly-ubiquitylated forms of cyclin B and Myf5, respectively.
The E3 involved in Myf5 ubiquitylation preferentially recognizes phosphorylated forms of Myf5
To further characterize the E3 responsible for Myf5 ubiquitylation, we undertook fractionation of mitotic egg extracts by chromatography (figure 7A). In this type of experiments, the fractions containing the E3 activity of interest are identified by their ability to allow ubiquitylation of Myf5 when added to a reaction mixture containing purified E1, E2, Ub and radiolabeled Myf5. The E3 activity was retained on a DEAE column, as described above (figure 6B, lane 2), and quantitatively recovered within the proteins eluted from the column with 0.25 M NaCl. Further fractionation of the DEAE 0.25 M NaCl eluate on a UnoQ column reproducibly led to the ubiquitylation pattern shown in figure 7B: a weak ubiquitylation activity could be detected in several fractions, but the major activity resulted from proteins eluted in fraction 15 (about 0.3 M NaCl). We failed to observe a robust ubiquitylation at this stage, probably because we removed from the active fractions the kinases important for Myf5 ubiquitylation, but also because the E3 appeared to be very unstable. Indeed its activity was essentially lost a few hours after UnoQ purification (data not shown). However, we used freshly prepared E3 fractions to test whether this E3 discriminates between phosphorylated and non-phosphorylated forms of Myf5, as suggested by our previous analyses in crude extracts. We took advantage of the fact that Myf5 is already phosphorylated during its translation in reticulocyte lysate and tested the fraction from the UnoQ column containing the E3 activity, for its ability to mediate ubiquitylation of either untreated (phosphorylated) Myf5 or Myf5 dephosphorylated (deP Myf5) by treatment of the lysate with alkaline phosphatase. Figure 7C shows that ubiquitylation of Myf5 occurs when using the phosphorylated form of Myf5, but is strongly diminished (although not abolished) when using non-phosphorylated Myf5. This result confirms that phosphorylation of Myf5 favors its recognition by the E3 mediating its ubiquitylation.
Figure 7 Myf5 ubiquitylation is controlled by phosphorylation. (A) Scheme of Myf5 E3 purification : for details see Methods section. (B) 2 ml of mitotic extract were fractionated as described in A; the fractions obtained from the UnoQ column were screened for their ability to ubiquitylate radiolabeled Myf5 in the presence of a ubiquitylation mix containing GST-Ub (see Methods). The reactions were resolved by 10% SDS-PAGE. (Ub)n-Myf5 indicates poly-ubiquitylated forms of Myf5. (C) The fraction containing the E3 was concentrated about 10-fold using a Centricon 10 K (Millipore). Untreated Myf5 (lanes 1 to 3) or dephosphorylated (deP) Myf5 (lanes 4 to 6) was incubated with the ubiquitylation mix alone (lanes 2, 5) or together with 6 μl of concentrated fraction containing the E3 (lanes 3, 6) for 30 minutes at 25°C. Non incubated Myf5 was loaded in lanes 1 and 4 (START) as control. Similar amounts of radioactive Myf5 and dephosphorylated Myf5 were used.
Discussion
The striking differential cell cycle-regulated degradation of both MyoD and Myf5 in proliferating myoblasts suggests that these two proteins, albeit homologous and partially redundant, carry out specific functions that must be turned down at a specific cell cycle stage. Physical and functional interactions of MyoD with cell cycle regulators, that impinge on the proliferation/differentiation interface in myoblasts, show that certain of these functions are interfering with cell cycle progression and suggest that, at least for MyoD, its cyclic degradation is required for maintenance of the proliferative status of the cells. For Myf5, much less is known, but it is likely that the same rationale can be followed, i.e. that its accelerated degradation at the late G2/M phase of the cell cycle is required for progression through subsequent phases.
Nature of the E3 responsible for Myf5 mitotic ubiquitylation
A strong argument in favor of an involvement of APC/C in Myf5 mitotic degradation was that Myf5 contains in its sequence a putative D-box motif, the mutation of which partially stabilizes Myf5 in mitotic cells, but not in interphase cells [21]. However, this degradation did not depend on UbcH10, the E2 known to function with APC/C, and the timing of degradation of Myf5 appeared different from that of known substrates of the APC/C: when examined in individual cells, Myf5 was always degraded before cyclin B, but the timing of its degradation relative to that of cyclin A appeared highly variable [21]. In addition, it is noteworthy that ubiquitylation of known substrates of the APC/C is not regulated by phosphorylation of the substrate [43]; the fine temporal tuning of ubiquitylation is rather controlled by post-translational modifications of the APC/C itself, respectively the core complex [44] or its activators [45], which modulate the activity of the E3. Altogether, the involvement of APC/C in Myf5 mitotic degradation remained an open issue.
Since mitotic degradation events are usually conserved between cells and organisms, the Xenopus egg extract system has been widely used to dissect these events and particularly to study the functions of APC/C [36,46,47]. To resolve the dilemma relative to the possible involvement of APC/C in the mitotic degradation of Myf5, we tested whether this complex is important for Myf5 mitotic degradation in Xenopus extracts. We found by two different approaches that APC/C does not participate to the ubiquitylation of Myf5 in Xenopus egg extracts. This result raises two questions. The first is whether the Xenopus system is representative of other cells and particularly myoblasts as far as Myf5 mitotic degradation is concerned. At this point, we believe that there is no reason to doubt it, as it would be a surprise if Myf5 was degraded in this system in a mitotic-dependent but different manner to that in cultured cells. Since the mechanisms controlling mitosis are conserved in higher eukaryotes, the frog system has been central for the understanding of many mechanisms controlling mitosis (including the discovery of APC/C) and, up until now, most data obtained with Xenopus egg extracts for APC/C substrates have been confirmed in other systems. But obviously a definitive conclusion on the role of APC/C in Myf5 degradation will require further studies in myoblasts. The second question is why, if APC/C is not important for Myf5 mitotic degradation, the Destruction-box mutants of Myf5 were specifically stabilized in cells during mitosis [21]. We are presently not able to answer this question, but several observations on this issue may be important. First, the D-box is a loosely conserved motif that can be found in many proteins, and several examples show that it cannot be automatically assimilated to a signature for targeting to the APC/C (see for example [40]). Thus, Myf5 D-box could be important for Myf5 mitotic degradation without actually acting as a genuine D-box motif. Second, there are several degradation pathways acting on Myf5 in cells, and one cannot exclude that the Destruction-box mutants of Myf5 interfere with other systems than the strictly mitotic pathway. In support of this notion, it is important to note that these mutants are only partially stabilized in mitotic cells: their further stabilization by proteasome inhibitors indeed indicates that proteasomes are still actively degrading these mutants in mitotic cells [21]. Moreover, we found no stabilization of these mutants in Xenopus egg extracts (data not shown), which, as shown by the absence of degradation of Myf5 in interphase, seem to possess only the mitotic pathway acting on Myf5. Third, because the D-box motif of Myf5 is adjacent to the DNA binding domain, these mutants have a decreased affinity for DNA as compared to the wild type protein [48]. Since binding to DNA and presence of various partners have been shown to alter MyoD degradation [49,50], it is possible that conformational changes induced by mutations in the D-box motif of Myf5 impact on its mitotic degradation by indirect ways.
Altogether, we believe that our data in Xenopus are a solid, albeit non definitive, argument in favor of the non-involvement of APC/C in Myf5 mitotic degradation. Interestingly, there are up to now few substrates known to be degraded in mitosis by the ubiquitin proteasome pathway, and whose ubiquitylation is not due to the APC/C [51]. The high incidence of the phosphorylation status of Myf5 on its ubiquitylation and degradation suggests that an E3 from the SCF family of complexes could be involved [43,52]. Since few tools are available to study SCF complexes in Xenopus egg extracts, we are currently analyzing the potential involvement of these complexes in Myf5 ubiquitylation using mammalian systems.
Control of Myf5 degradation by phosphorylation
Based on the homology of Myf5 to MyoD, we expected a simple mechanism in which a unique phosphorylation would trigger Myf5 mitotic degradation. Indeed phosphorylation of MyoD on its serine 200 by the CDK2/cyclin E kinase has been shown to be critical to entail rapid degradation of this protein at the end of the G1 phase of the cell-cycle [12]. However, although this simple scenario cannot be excluded at the moment, our results on the mechanisms controlling Myf5 stability are drawing a much more complex picture.
We found a significant degree of variability from extract to extract in the phosphorylation status of Myf5, that prevented us from obtaining clear conclusions on the nature of the kinase(s) controlling Myf5 ubiquitylation and degradation. We believe that some of the problems were due to the origin of the substrate we used. Indeed, there is a clear interference in this assay of kinases present in the reticulocyte lysate in which Myf5 was translated, and we observed a variability between different lysates that was likely to contribute to the difficulty of obtaining solid conclusions. However, attempts to translate Myf5 in other systems derived from wheat-germ or E. coli, or to co-translate the Myf5 partner E12 did not fundamentally solve the problem of variability.
We thus think that the problem resides elsewhere, in the complexity of Myf5 phosphorylation that makes the degradation of this protein extremely sensitive to the kinase/phosphatase equilibrium of the extracts. The concordant picture arising from all our attempts to modify Myf5 stability in Xenopus egg extracts, either by inhibition of kinases or phosphatases or by single mutations of the S/TP sites of Myf5, is that several enzymes impinge on Myf5 phosphorylation status and thus Myf5 stability, and that Myf5 ubiquitylation apparently requires phosphorylation on multiple sites. However, although most S/TP sites of Myf5 seem phosphorylated in mitotic extracts, as their mutation affects Myf5 migration in gel, none of these sites seems to be absolutely necessary for efficient degradation of the protein. This could suggest that it is phosphorylation itself or, by analogy with data obtained for the CDK inhibitor Sic1 [30,31], the number of phosphorylated residues rather than their identity that is the important parameter for Myf5 ubiquitylation.
An unexpected result in our experiments was the observation that a phosphorylated form of Myf5 was completely resistant to degradation. Because this form was the latest phospshorylated form to appear and the slowest migrating in gel, we called it the hyperphosphorylated form of Myf5. This terminology is coherent with the observation that Myf5 apparently escapes degradation only when it is fully phosphorylated. However at this point we cannot exclude that a single phosphorylation event is responsible for both Myf5 stabilization and shift to the slowest migrating form. An important question is whether the stabilizing hyperphosphorylation of Myf5 exists in cells and particularly in myoblasts. A possibility is that this hyperphosphorylation is not a standard process seen at each cell cycle, but a regulated mechanism occurring only in certain physiological situations. Alternatively, as it has been suggested recently for MyoD [13], such a mechanism could allow a low amount of Myf5 to be preserved during passage through mitosis, in order for the cell to be able to mobilize it immediately after completion of cell division.
Conclusion
In this article, we describe the work we performed using the Xenopus egg extract system to better define the mechanisms that control Myf5 degradation in mitosis. Altogether, our data are in favor of the following model (figure 8): in mitotic extracts, Myf5 is phosphorylated on numerous residues. Some of these residues, once phosphorylated, are recognized by an E3, distinct from APC/C, that mediates Myf5 polyubiquitylation, targeting it to the proteasome for degradation. However, when Myf5 is fully phosphorylated, it somehow escapes degradation and is stabilized. In such a model, phosphorylation of Myf5 is necessary for its mitotic ubiquitylation, but can also be used to stabilize the protein. Thus the stability of the protein can be finely tuned by differential activation of various signaling pathways, providing a precise and rapid way to adapt Myf5 levels in cells in function of the cell cycle stage.
Figure 8 A model for Myf5 mitotic degradation pathway. Myf5 is subjected to multiple phosphorylations in Xenopus mitotic egg extracts, that tightly control its stability: phosphorylation of the protein leads to its ubiquitylation by an E3 distinct from APC/C, and its subsequent degradation by the proteasome. However, a hyperphosphorylated form of Myf5 remains stable (see text for details).
Methods
Cell culture and lysates
HeLa-S3 cells expressing mouse Myf5 (UN6 clone) under an inducible promoter (tet-off) are usually cultured in DMEM containing 10% FCS (Cambrex), 1000 U/ml penicillin/streptomycin, 2 mM L-Glutamine, 500 μg/ml G418, 170 U/ml hygromycin B and 1 μg/ml tetracycline. Expression of Myf5 is induced for 24 hours by washing the cells three times with PBS and adding fresh medium without tetracycline. To prepare lysates, cells are scraped in cold PBS and centrifuged for 10 min at 1000 rpm. The pellet is then resuspended in about 5 pellet volumes of 20 mM Tris pH 7.5, 137 mM NaCl, 10% glycerol, 1% NP40, 1 mM orthovanadate, 20 mM NaF, protease inhibitor mix (Complete, Roche-Boehringer), and incubated for 30 min on ice, while vortexing every 5 min. The lysate is then centrifuged for 8 min at 16000 × g. The supernatant is removed and protein concentration is measured using BSA as a standard (Bradford Reagent Assay, Pierce).
Xenopus egg extracts
Mitotic extracts are prepared by gently crushing oocytes arrested in metaphase of meiosis II, with CDK1/cyclin B kinase activity maintained at a high level by cytostatic factor (CSF); these extracts, called CSF extracts, thus reproduce a pseudo-mitotic state. When eggs are fertilized, penetration of sperms induces a calcium wave that inactivates CSF activity and triggers cyclin B degradation, and CDK1 activity rapidly declines to an interphase level (reviewed in [53]). A synchronous release from CSF arrest can be artificially induced by activating the eggs with calcium ionophore treatment. To prepare extracts, eggs are dejellied in a solution containing 2% L-Cystein (pH 7.5), then washed extensively in XB Buffer (100 mM KCl, 0.1 mM CaCl2, 1 mM MgCl2, 10 mM HEPES pH 7.7, 50 mM Sucrose) supplemented with 6 mM EGTA and lysed by centrifugation for 20 min at 16000 × g. The supernatant (except the floating lipidic phase) is removed and supplemented with an ATP regenerating system (1.25 mM ATP, 1.25 mM MgCl2, 1.9 mM Creatine Phosphate, 6.25 μg/ml Creatine Phosphokinase), leupeptin (25 μg/ml) and cytochalasin B (25 μg/ml). After centrifugation for 20 min at 16000 × g, the supernatant is removed, aliquoted and frozen in liquid nitrogen. Frozen aliquots are kept at -80°C for several weeks. For interphase extracts, eggs were dejellied as described above, then activated by calcium ionophore and incubated for 10 minutes in XB containing 10 μg/ml cycloheximide (SIGMA-ALDRICH). Eggs were then lysed as described above.
In vitro production of [35S] methionine-labeled proteins
Human cyclin B, mouse Myf5 and frog Xkid purified plasmids are transcribed and translated in vitro in the presence of L-[35S]-methionine (Amersham, Redivue mix) in rabbit reticulocyte lysates, following manufacturer's instructions (Promega TNT Quick coupled system), then loaded on pre-equilibrated (Tris-HCl 20 mM, pH7.5) Biospin6 columns (BioRad), in order to eliminate non-incorporated radioactive methionine. The dephosphorylated form of Myf5 is obtained by incubating freshly translated Myf5 with alkaline phosphatase linked to agarose beads (SIGMA-ALDRICH) for 30 min at 37°C. The mix is then loaded on BioSpin6 columns as above to remove both the beads and the excess radioactive methionine.
In vitro degradation and ubiquitylation assays
For degradation assays, 18 μl of mitotic or interphase extracts are usually mixed with 2 μl of reticulocyte lysate containing the radioactive (in vitro translated) substrate and are further incubated at 25°C. At appropriate time points, 2 μl samples are subjected to SDS-PAGE (10% gels) analysis. Gels are then dried, exposed and the radioactive bands are quantified using a PhosphoImager (Molecular Dynamics). Ubiquitylation assays are performed in the presence of 200 μM MG132 (BioMol), 5 μM Ubiquitin aldehyde (BioMol), 1 mg/ml Ubiquitin (SIGMA-ALDRICH), 5 mM MgCl2 and 1 mM ATPγS. Lysineless mutant ubiquitin (UbK0) is purchased to Boston Biochem.
Purification of Myf5 E3
A mitotic extract is diluted twice in buffer A (20 mM Tris pH7.5, 1 mM DTT) supplemented with 1% Igepal CA-630 (Sigma), then incubated for 15 min at 4°C with the same volume of AffigelBlue DEAE beads (BioRad) pre-equilibrated in buffer A + 0.5% Igepal. The beads are then extensively washed with a large excess of buffer A containing 0.5% Igepal, then with buffer A only. Myf5 E3 activity is eluted with buffer A containing 0.25 M NaCl. This fraction is directly loaded on a UnoQ1 or a UnoQ12 column (BioRad) pre-equilibrated with buffer A containing 0.2 M NaCl. A gradient from 0.2 M to 0.5 M of NaCl is applied to the column, and Myf5 E3 activity is eluted at a concentration of approximately 0.3 M NaCl. The fractions obtained from the different columns are screened for their ability to ubiquitylate radiolabeled Myf5 in the presence of the ubiquitylation enzymes E1 (Xenopus, 50 ng), E2 (recombinant UBCH5B, 0.5 μg), purified recombinant GST-Ub (1 mg/ml) or ubiquitin (1 mg/ml), 1 mM ATP, 5 mM MgCl2, 5 μM ubiquitin aldehyde, 200 μM MG132, 20 mM Tris pH7.5. Ubiquitylation reactions are performed with 1 μl of radiolabeled Myf5, 6 μl of each fraction, in a final volume of 10 μl, and incubated for 30 min at 25°C. After appropriate incubation time, the reaction is stopped by addition of sample buffer, and the samples are analyzed by electrophoresis and PhosphoImaging.
Antibodies
Anti-Myf5 (C-20, Santa Cruz) is diluted 1:1000 for Western Blotting. Secondary anti-rabbit-HRP (Amersham) is diluted 1:10000. APC/C immuno-precipitations are realized with a home-made affinity-purified anti-Cdc27 [46].
Authors' contributions
CD was in charge of this project, and conducted most of the experiments shown in this article. GJG has critically contributed to this work by establishing its experimental basis in Xenopus egg extracts, and by demonstrating stabilization of Myf5 by hyperphosphorylation. CL participated in the conception of the study and established the HeLa cell line expressing Myf5. TL has been a continuous interlocutor during the whole work, and provided his expertise and important tools to analyze Myf5 and Cyclin B degradation in Xenopus egg extracts. GL prepared and tested CSF extracts, and produced various ubiquitylation enzymes. CP participated in the conception of the study. OC conceived the study, participated in its design and coordination. CD and OC wrote the manuscript, with special help of GJG and CL. All authors read and approved the final manuscript.
Acknowledgements
We thank the members of our laboratory for their continuous support and their helpful discussions. We are especially grateful to G. Carnac, A. Fernandez and S. Leibovitch for their suggestions and criticisms. We acknowledge the undergraduate students (T. Petit, L. Americh, O. Cexus, C. Vernet and D. Latreille) who participated to this work at various stages. This work has been supported by grants from the "Association Française contre les Myopathies" (AFM) and the European contract n°QLG1-CT-2001-02026. CD is the beneficiary of fellowships from the French Ministry of Research and the Fondation pour la Recherche Médicale (FRM).
==== Refs
Buckingham M Skeletal muscle formation in vertebrates Curr Opin Genet Dev 2001 11 440 448 11448631 10.1016/S0959-437X(00)00215-X
Puri PL Sartorelli V Regulation of muscle regulatory factors by DNA-binding, interacting proteins, and post-transcriptional modifications J Cell Physiol 2000 185 155 173 11025438 10.1002/1097-4652(200011)185:2<155::AID-JCP1>3.0.CO;2-Z
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
Aurade F Pinset C Chafey P Gros F Montarras D Myf5, MyoD, myogenin and MRF4 myogenic derivatives of the embryonic mesenchymal cell line C3H10T1/2 exhibit the same adult muscle phenotype Differentiation 1994 55 185 192 8187980 10.1046/j.1432-0436.1994.5530185.x
Davis RL Weintraub H Lassar AB Expression of a single transfected cDNA converts fibroblasts to myoblasts Cell 1987 51 987 1000 3690668 10.1016/0092-8674(87)90585-X
Rudnicki MA Schnegelsberg PN Stead RH Braun T Arnold HH Jaenisch R MyoD or Myf-5 is required for the formation of skeletal muscle Cell 1993 75 1351 1359 8269513 10.1016/0092-8674(93)90621-V
Montarras D Chelly J Bober E Arnold H Ott MO Gros F Pinset C Developmental patterns in the expression of Myf5, MyoD, myogenin, and MRF4 during myogenesis New Biol 1991 3 592 600 1911647
Kassar-Duchossoy L Gayraud-Morel B Gomes D Rocancourt D Buckingham M Shinin V Tajbakhsh S Mrf4 determines skeletal muscle identity in Myf5:Myod double-mutant mice Nature 2004 431 466 471 15386014 10.1038/nature02876
Kitzmann M Fernandez A Crosstalk between cell cycle regulators and the myogenic factor MyoD in skeletal myoblasts Cell Mol Life Sci 2001 58 571 579 11361092
Skapek SX Rhee J Spicer DB Lassar AB Inhibition of myogenic differentiation in proliferating myoblasts by cyclin D1-dependent kinase Science 1995 267 1022 1024 7863328
Kitzmann M Vandromme M Schaeffer V Carnac G Labbe JC Lamb N Fernandez A cdk1- and cdk2-mediated phosphorylation of MyoD Ser200 in growing C2 myoblasts: role in modulating MyoD half-life and myogenic activity Mol Cell Biol 1999 19 3167 3176 10082583
Tintignac LA Leibovitch MP Kitzmann M Fernandez A Ducommun B Meijer L Leibovitch SA Cyclin E-cdk2 phosphorylation promotes late G1-phase degradation of MyoD in muscle cells Exp Cell Res 2000 259 300 307 10942602 10.1006/excr.2000.4973
Tintignac LA Sirri V Leibovitch MP Lecluse Y Castedo M Metivier D Kroemer G Leibovitch SA Mutant MyoD lacking Cdc2 phosphorylation sites delays M-phase entry Mol Cell Biol 2004 24 1809 1821 14749395 10.1128/MCB.24.4.1809-1821.2004
Zhang JM Zhao X Wei Q Paterson BM Direct inhibition of G(1) cdk kinase activity by MyoD promotes myoblast cell cycle withdrawal and terminal differentiation Embo J 1999 18 6983 6993 10601020 10.1093/emboj/18.24.6983
Halevy O Novitch BG Spicer DB Skapek SX Rhee J Hannon GJ Beach D Lassar AB Correlation of terminal cell cycle arrest of skeletal muscle with induction of p21 by MyoD Science 1995 267 1018 1021 7863327
Ott MO Bober E Lyons G Arnold H Buckingham M Early expression of the myogenic regulatory gene, myf-5, in precursor cells of skeletal muscle in the mouse embryo Development 1991 111 1097 1107 1652425
Kitzmann M Carnac G Vandromme M Primig M Lamb NJ Fernandez A The muscle regulatory factors MyoD and myf-5 undergo distinct cell cycle-specific expression in muscle cells J Cell Biol 1998 142 1447 1459 9744876 10.1083/jcb.142.6.1447
Yoshida N Yoshida S Koishi K Masuda K Nabeshima Y Cell heterogeneity upon myogenic differentiation: down-regulation of MyoD and Myf-5 generates 'reserve cells' J Cell Sci 1998 111 ( Pt 6) 769 779 9472005
Song A Wang Q Goebl MG Harrington MA Phosphorylation of nuclear MyoD is required for its rapid degradation Mol Cell Biol 1998 18 4994 4999 9710583
Lindon C Montarras D Pinset C Cell cycle-regulated expression of the muscle determination factor Myf5 in proliferating myoblasts J Cell Biol 1998 140 111 118 9425159 10.1083/jcb.140.1.111
Lindon C Albagli O Domeyne P Montarras D Pinset C Constitutive instability of muscle regulatory factor Myf5 is distinct from its mitosis-specific disappearance, which requires a D-box-like motif overlapping the basic domain Mol Cell Biol 2000 20 8923 8932 11073992 10.1128/MCB.20.23.8923-8932.2000
Murray AW Kirschner MW Cyclin synthesis drives the early embryonic cell cycle Nature 1989 339 275 280 2566917 10.1038/339275a0
Clarke PR Pagano M Preparation of Xenopus Egg Extracts and Their Utilization in Cell Cycle Studies Cell Cycle - Materials and Methods 1995 , Springer Lab Manual 103 1116
Nurse P Universal control mechanism regulating onset of M-phase Nature 1990 344 503 508 2138713 10.1038/344503a0
Glotzer M Murray AW Kirschner MW Cyclin is degraded by the ubiquitin pathway Nature 1991 349 132 138 1846030 10.1038/349132a0
Solomon MJ Glotzer M Lee TH Philippe M Kirschner MW Cyclin activation of p34cdc2 Cell 1990 63 1013 1024 2147872 10.1016/0092-8674(90)90504-8
Castro A Peter M Magnaghi-Jaulin L Vigneron S Galas S Lorca T Labbe JC Cyclin B/cdc2 induces c-Mos stability by direct phosphorylation in Xenopus oocytes Mol Biol Cell 2001 12 2660 2671 11553706
Lorca T Abrieu A Means A Doree M Ca2+ is involved through type II calmodulin-dependent protein kinase in cyclin degradation and exit from metaphase Biochim Biophys Acta 1994 1223 325 332 7918666 10.1016/0167-4889(94)90091-4
Semsarian C Wu MJ Ju YK Marciniec T Yeoh T Allen DG Harvey RP Graham RM Skeletal muscle hypertrophy is mediated by a Ca2+-dependent calcineurin signalling pathway Nature 1999 400 576 581 10448861 10.1038/23054
Nash P Tang X Orlicky S Chen Q Gertler FB Mendenhall MD Sicheri F Pawson T Tyers M Multisite phosphorylation of a CDK inhibitor sets a threshold for the onset of DNA replication Nature 2001 414 514 521 11734846 10.1038/35107009
Deshaies RJ Ferrell JEJ Multisite phosphorylation and the countdown to S phase Cell 2001 107 819 822 11779457 10.1016/S0092-8674(01)00620-1
Hershko A Ciechanover A The ubiquitin system Annu Rev Biochem 1998 67 425 479 9759494 10.1146/annurev.biochem.67.1.425
Holloway SL Glotzer M King RW Murray AW Anaphase is initiated by proteolysis rather than by the inactivation of maturation-promoting factor Cell 1993 73 1393 1402 8391932 10.1016/0092-8674(93)90364-V
Hershko A Ciechanover A Varshavsky A Basic Medical Research Award. The ubiquitin system Nat Med 2000 6 1073 1081 11017125 10.1038/80384
Deshaies RJ SCF and Cullin/Ring H2-based ubiquitin ligases Annu Rev Cell Dev Biol 1999 15 435 467 10611969 10.1146/annurev.cellbio.15.1.435
King RW Peters JM Tugendreich S Rolfe M Hieter P Kirschner MW A 20S complex containing CDC27 and CDC16 catalyzes the mitosis-specific conjugation of ubiquitin to cyclin B Cell 1995 81 279 288 7736580 10.1016/0092-8674(95)90338-0
Sudakin V Ganoth D Dahan A Heller H Hershko J Luca FC Ruderman JV Hershko A The cyclosome, a large complex containing cyclin-selective ubiquitin ligase activity, targets cyclins for destruction at the end of mitosis Mol Biol Cell 1995 6 185 197 7787245
King RW Glotzer M Kirschner MW Mutagenic analysis of the destruction signal of mitotic cyclins and structural characterization of ubiquitinated intermediates Mol Biol Cell 1996 7 1343 1357 8885231
Castro A Arlot-Bonnemains Y Vigneron S Labbe JC Prigent C Lorca T APC/Fizzy-Related targets Aurora-A kinase for proteolysis EMBO Rep 2002 3 457 462 11964384 10.1093/embo-reports/kvf095
Castro A Vigneron S Bernis C Labbe JC Lorca T Xkid is degraded in a D-box, KEN-box, and A-box-independent pathway Mol Cell Biol 2003 23 4126 4138 12773557 10.1128/MCB.23.12.4126-4138.2003
Geley S Kramer E Gieffers C Gannon J Peters JM Hunt T Anaphase-promoting complex/cyclosome-dependent proteolysis of human cyclin A starts at the beginning of mitosis and is not subject to the spindle assembly checkpoint J Cell Biol 2001 153 137 148 11285280 10.1083/jcb.153.1.137
Brassac T Castro A Lorca T Le Peuch C Doree M Labbe JC Galas S The polo-like kinase Plx1 prevents premature inactivation of the APC(Fizzy)-dependent pathway in the early Xenopus cell cycle Oncogene 2000 19 3782 3790 10949933 10.1038/sj.onc.1203724
Vodermaier HC APC/C and SCF: controlling each other and the cell cycle Curr Biol 2004 14 R787 96 15380093 10.1016/j.cub.2004.09.020
Shteinberg M Protopopov Y Listovsky T Brandeis M Hershko A Phosphorylation of the cyclosome is required for its stimulation by Fizzy/cdc20 Biochem Biophys Res Commun 1999 260 193 198 10381365 10.1006/bbrc.1999.0884
Kramer ER Scheuringer N Podtelejnikov AV Mann M Peters JM Mitotic regulation of the APC activator proteins CDC20 and CDH1 Mol Biol Cell 2000 11 1555 1569 10793135
Lorca T Castro A Martinez AM Vigneron S Morin N Sigrist S Lehner C Doree M Labbe JC Fizzy is required for activation of the APC/cyclosome in Xenopus egg extracts Embo J 1998 17 3565 3575 9649427 10.1093/emboj/17.13.3565
Schmidt A Duncan PI Rauh NR Sauer G Fry AM Nigg EA Mayer TU Xenopus polo-like kinase Plx1 regulates XErp1, a novel inhibitor of APC/C activity Genes Dev 2005 19 502 513 15713843 10.1101/gad.320705
Lindon C Albagli O Pinset C Montarras D Cell density-dependent induction of endogenous myogenin (myf4) gene expression by Myf5 Dev Biol 2001 240 574 584 11784084 10.1006/dbio.2001.0435
Abu Hatoum O Gross-Mesilaty S Breitschopf K Hoffman A Gonen H Ciechanover A Bengal E Degradation of myogenic transcription factor MyoD by the ubiquitin pathway in vivo and in vitro: regulation by specific DNA binding Mol Cell Biol 1998 18 5670 5677 9742084
Lingbeck JM Trausch-Azar JS Ciechanover A Schwartz AL E12 and E47 modulate cellular localization and proteasome-mediated degradation of MyoD and Id1 Oncogene 2005 24 6376 6384 16007194
Margottin-Goguet F Hsu JY Loktev A Hsieh HM Reimann JD Jackson PK Prophase destruction of Emi1 by the SCF(betaTrCP/Slimb) ubiquitin ligase activates the anaphase promoting complex to allow progression beyond prometaphase Dev Cell 2003 4 813 826 12791267 10.1016/S1534-5807(03)00153-9
Reed SI Ratchets and clocks: the cell cycle, ubiquitylation and protein turnover Nat Rev Mol Cell Biol 2003 4 855 864 14625536 10.1038/nrm1246
Tunquist BJ Maller JL Under arrest: cytostatic factor (CSF)-mediated metaphase arrest in vertebrate eggs Genes Dev 2003 17 683 710 12651887 10.1101/gad.1071303
Hoessel R Leclerc S Endicott JA Nobel ME Lawrie A Tunnah P Leost M Damiens E Marie D Marko D Niederberger E Tang W Eisenbrand G Meijer L Indirubin, the active constituent of a Chinese antileukaemia medicine, inhibits cyclin-dependent kinases Nat Cell Biol 1999 1 60 67 10559866 10.1038/9035
Leclerc S Garnier M Hoessel R Marko D Bibb JA Snyder GL Greengard P Biernat J Wu YZ Mandelkow EM Eisenbrand G Meijer L Indirubins inhibit glycogen synthase kinase-3 beta and CDK5/p25, two protein kinases involved in abnormal tau phosphorylation in Alzheimer's disease. A property common to most cyclin-dependent kinase inhibitors? J Biol Chem 2001 276 251 260 11013232 10.1074/jbc.M002466200
Meijer L Raymond E Roscovitine and other purines as kinase inhibitors. From starfish oocytes to clinical trials Acc Chem Res 2003 36 417 425 12809528 10.1021/ar0201198
Meijer L Borgne A Mulner O Chong JP Blow JJ Inagaki N Inagaki M Delcros JG Moulinoux JP Biochemical and cellular effects of roscovitine, a potent and selective inhibitor of the cyclin-dependent kinases cdc2, cdk2 and cdk5 Eur J Biochem 1997 243 527 536 9030781 10.1111/j.1432-1033.1997.t01-2-00527.x
De Azevedo WF Leclerc S Meijer L Havlicek L Strnad M Kim SH Inhibition of cyclin-dependent kinases by purine analogues: crystal structure of human cdk2 complexed with roscovitine Eur J Biochem 1997 243 518 526 9030780 10.1111/j.1432-1033.1997.0518a.x
Meksuriyen D Cordell GA Biosynthesis of staurosporine, 1. 1H- and 13C-NMR assignments J Nat Prod 1988 51 884 892 3204380 10.1021/np50059a012
Yagi Y Sotani T Nagao T Horio T Yamamoto I Gohda E Induction by staurosporine of hepatocyte growth factor production in human skin fibroblasts independent of protein kinase inhibition Biochem Pharmacol 2003 66 1797 1808 14563490 10.1016/S0006-2952(03)00547-1
Duncia JV Santella JB Higley CA Pitts WJ Wityak J Frietze WE Rankin FW Sun JH Earl RA Tabaka AC Teleha CA Blom KF Favata MF Manos EJ Daulerio AJ Stradley DA Horiuchi K Copeland RA Scherle PA Trzaskos JM Magolda RL Trainor GL Wexler RR Hobbs FW Olson RE MEK inhibitors: the chemistry and biological activity of U0126, its analogs, and cyclization products Bioorg Med Chem Lett 1998 8 2839 2844 9873633 10.1016/S0960-894X(98)00522-8
Favata MF Horiuchi KY Manos EJ Daulerio AJ Stradley DA Feeser WS Van Dyk DE Pitts WJ Earl RA Hobbs F Copeland RA Magolda RL Scherle PA Trzaskos JM Identification of a novel inhibitor of mitogen-activated protein kinase kinase J Biol Chem 1998 273 18623 18632 9660836 10.1074/jbc.273.29.18623
Toullec D Pianetti P Coste H Bellevergue P Grand-Perret T Ajakane M Baudet V Boissin P Boursier E Loriolle F Duhamel L Charon D Kirilovsky J The bisindolylmaleimide GF 109203X is a potent and selective inhibitor of protein kinase C J Biol Chem 1991 266 15771 15781 1874734
Kiss Z Phillips H Anderson WH The bisindolylmaleimide GF 109203X, a selective inhibitor of protein kinase C, does not inhibit the potentiating effect of phorbol ester on ethanol-induced phospholipase C-mediated hydrolysis of phosphatidylethanolamine Biochim Biophys Acta 1995 1265 93 95 7857990 10.1016/0167-4889(94)00242-7
Kronke M Leonard WJ Depper JM Arya SK Wong-Staal F Gallo RC Waldmann TA Greene WC Cyclosporin A inhibits T-cell growth factor gene expression at the level of mRNA transcription Proc Natl Acad Sci U S A 1984 81 5214 5218 6332315
Manger B Hardy KJ Weiss A Stobo JD Differential effect of cyclosporin A on activation signaling in human T cell lines J Clin Invest 1986 77 1501 1506 2422208
Izumi T Maller JL Phosphorylation and activation of the Xenopus Cdc25 phosphatase in the absence of Cdc2 and Cdk2 kinase activity Mol Biol Cell 1995 6 215 226 7787247
|
16321160
|
PMC1322219
|
CC BY
|
2021-01-04 16:26:24
|
no
|
BMC Biochem. 2005 Dec 1; 6:27
|
utf-8
|
BMC Biochem
| 2,005 |
10.1186/1471-2091-6-27
|
oa_comm
|
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-421630068310.1186/1471-2431-5-42Research ArticleOpportunistic screening for iron-deficiency in 6–36 month old children presenting to the paediatric emergency department Pusic Martin V [email protected] Brenda J [email protected] David [email protected] University of B.C., Dept of Pediatrics, Vancouver, BC Canada2 University of Manitoba, Winnipeg, Manitoba, Canada3 McGill University, Dept. of Pediatrics, Montreal, Quebec, Canada2005 22 11 2005 5 42 42 6 6 2005 22 11 2005 Copyright © 2005 Pusic et al; licensee BioMed Central Ltd.2005Pusic 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 Complete Blood Count (CBC) is a test frequently performed on children presenting to the Pediatric Emergency Department (PED), usually for the evaluation of an infectious illness. The CBC also allows for screening for Iron-deficiency Anemia. This study aims to determine the prevalence of a low Mean Cell Volume (MCV) in children having a CBC performed during a PED visit and whether physicians acted upon the abnormal value.
Methods
We present a retrospective cohort study. We reviewed the PED charts of all children 6–36 months of age who had a CBC performed during a 4-month period and the red blood cell mean cell volume was low. Our main outcome variable was whether or not the possible iron deficiency was addressed through documentation of either iron therapy or further investigation.
Results
938 children had a CBC performed during the two periods. Of these, 78 (8%) had an abnormal MCV or Hemoglobin with no previously identified explanation. Physicians documented either treatment or follow-up investigations in 27 cases (35%, 95% CI: 24–46%). Factors associated with the physician documenting either treatment or investigation plan were the following: hemoglobin level (OR 12.6; 95%CI: 4.0, 39) and age ≤ 18 months (OR 4.2; 95%CI: 1.4, 13).
Conclusion
Children who have had a CBC in the PED can be screened for iron deficiency at no additional cost. Physicians may be under-utilizing this information.
==== Body
Background
Iron deficiency anemia in infants and young children can have serious adverse effects on their development including irreversible cognitive impairment and behaviour problems [1-5]. The negative sequelae can be prevented with simple treatment with iron-fortified infant formula or oral iron supplementation. High risk groups include children from low socioeconomic status families, Chinese children, aboriginal children, infants of low birth weight, children whose mothers were iron-deficient during pregnancy, and children who consume cow's milk [6,7]. Despite an overall decline in the incidence over the past 20 to 30 years there is still a significant proportion of children at risk. The prevalence of iron-deficiency anemia among a random group of Canadian 9 month olds was 7%, and 24% had low iron stores, despite the introduction of iron-fortified cereals by six months of age [7]. Among African-American children aged 9 to 36 months the prevalence of iron deficiency anemia was 8% [8].
The Canadian Task Force on Preventive Health Care provides a grade B recommendation (fair evidence to support) for routine hemoglobin measurement for infants at high risk [6]. The U.S. Preventive Services Task Force recommends that screening should be offered once to all infants, and the American Academy of Pediatrics recommends at least one measurement of hemoglobin or hematocrit in infancy, and at least during the ages of 1 to 4 years and again in adolescence [9].
Basic screening for iron-deficiency anemia consists of a complete blood count (CBC) looking for hemoglobin less than the age-adjusted normal range, with a sensitivity of 86% and a specificity of 97% [6]. The diagnosis is supported by a low mean corpuscular volume (MCV), and increased red cell distribution width (RDW). Confirmation of the diagnosis is commonly obtained with a trial of iron supplementation and measurement of the response [10]. More complex measures of iron stores including serum ferritin, transferring saturation and erythrocyte protoporphyrin activity are used to more accurately determine iron status [11].
It is unclear what proportion of infants and children are currently being screened as recommended. Ideally every child would be screened but this likely does not happen for reasons of cost and the pain associated with the blood test.
A large number of infants and children who are assessed in emergency departments have a CBC done. This presents an opportunity to screen for iron-deficiency anemia. However, in this setting the physician is likely to be focused on the white cell count as most of these patients are being assessed for the possibility of infection. As a result, CBC indices that suggest iron-deficiency anemia may be overlooked. Our hypothesis was that the CBC information that pertains to iron-deficiency is underutilized given that this is not usually the primary indication for the test. If confirmed, we proposed to raise physician awareness of this method of secondary screening.
Methods
Setting
The Montreal Children's Hospital is a 181-bed tertiary-care referral center affiliated with McGill University. The Pediatric Emergency Department (PED) has an annual census of 85,000 visits and is staffed by 6 full-time Pediatric Emergency Medicine physicians. A large group of community-based pediatricians also contribute from 1 to 6 shifts per month. Medical students, paediatric, family medicine and emergency medicine residents and fellows all do training rotations in the PED.
Study design
This is a retrospective before-after cohort study. We identified all children who had a Complete Blood Count (CBC) performed in the PED during a 2-month period (February–March 1999), implemented an educational intervention and then collected data during the same two months a year later. The Institutional Review Board approved the study.
Subjects
Using our laboratory information system, we were able to identify all children who had a CBC done in the PED during the two study periods. We then selected children aged 6 to 36 months based on their age at the time the CBC was done. We excluded children with a known history of the following: anemia on treatment, sickle cell disease or trait, thalassaemia or trait, chronic illness likely to cause anemia. Patients who presented twice during a study period were handled as follows: if, on the first visit iron-deficiency was recognized, they were excluded on the next visit. If, on the first visit, iron-deficiency was not identified, then they were included on the second visit.
Laboratory techniques
CBCs were determined using a Bayer Advia 120 hematology analyzer (Bayer AG, Leverkusen, Germany). We defined the lower limit of normal for MCV as 70 fl while 110 g/L was the cutoff for normal Hemoglobin.
Definitions
MCV addressed
We identified children as having had their iron status addressed if either treatment or further investigations was documented on the PED encounter sheet. Treatment could include any iron preparation initiated by the PED physician. Further assessment could include specific referral to their family doctor for assessment of iron deficiency anemia or arrangement of confirmatory laboratory tests (any of: Hgb Electrophoresis, Serum Fe, Ferritin, Free Erythrocyte Protoporphyrin, Transferrin Level).
MCV not addressed
Children with a low MCV who did not have either treatment or arrangements for further assessment, as defined above, documented on either their PED encounter sheet or, if admitted, on their discharge summary.
Data collection
From the Laboratory Information System we collected the hemoglobin measurement, all standard CBC indices, ordering physician name, patient date of birth and hospital chart number. We performed chart reviews on all children with low MCV. We limited the chart review to the PED encounter sheet of the CBC date as well as the inpatient chart for those patients who were admitted at the time of the initial encounter. From the charts we collected information concerning the reason for the visit, any past history relevant to anemia, a list of medications, and whether any note had been made regarding possible iron-deficiency.
Educational intervention
We collected the data in two sessions. In April 1999 we identified the first cohort (Feb–Mar 1999). These initial data were analyzed and showed that a significant number of patients were not having possible iron deficiency addressed. We presented this information (that we pediatricians were missing up to 75% of children with abnormal MCV's in this age group) in a one-hour PED Grand Round in June 2000. To reinforce the message, we mailed a one-page abstract to them with their monthly work schedules in January 2000. This abstract stressed the importance of examining the MCV results on all CBC's done. We collected the second set of data after these interventions in Feb-March 2000. We hoped that this intervention would improve the rate of recognition of abnormal MCVs.
Data analysis
We analyzed the data using descriptive statistics and used chi-squared tests to look for factors that differed between groups where the possible iron-deficiency had been addressed. We calculated 95% confidence intervals around proportions using the binomial distribution. We also developed a univariate logistic regression model to see which factors predict proper identification and documentation of iron deficiency. Our sample size was determined based on the following assumptions: with 25 subjects in each group, we could detect an improvement from 10% identification to 50% with a power of 80% at alpha = 0.05.
Results
PED patients with CBCs
During the four months studied, a total of 7571 patients aged 6–36 months were seen in the PED. Of these, 934 (12%) had CBC's performed which is 29% of the 3206 CBCs performed for all children seen in the PED during the study periods.
The MCV was less than 70 fl in 94 patients (10%) while 184 (19.7%) of the children had a hemoglobin < 110 g/L. Fifty-seven had both low MCV and low hemoglobin. The relationship between MCV and HgB is shown graphically in Figure 1. Children 18–36 months of age did not have significantly different MCV (79.4 fl) or hemoglobin (122 g/L) values compared with children 6–17 months (78.5 fl, 124 g/L respectively; both p = NS).
Figure 1 MCV versus Hemoglobin for all children having a CBC in the PED.
Patients with possible iron deficiency
Of the patients with low MCV, 16 were excluded from the analysis for the following reasons: sickle cell disease (1), thalassemia (3), previously identified iron deficiency (8) and chronic disease (2 – autoimmune hepatitis and liver mass). Two charts were missing. Discharge diagnosis and some form of discharge plan had been documented on every chart. For the 78 included patients, 45 (57%) had a hemoglobin value less than 110 g/L.
Patients' discharge diagnoses were available for all included patients. The most frequent diagnoses were: Gastroenteritis (19; 24%), Respiratory Illnesses (18; 23%), Nonspecific Viral Illness (13; 17%), and Febrile Seizures (4; 5%). No other diagnosis made up more than three of the cases.
Patients with low MCV's had an iron medication prescribed in 21 cases (27%). In another 6 (8%) cases, explicit follow-up was arranged so that, in total 27 of 78 patients had their low MCV addressed (35%; 95% CI: 24–46%). Even markedly decreased MCV's (≤65 fl) were addressed only 12 of 28 times (43%; 95% CI: 24–63%). Iron was addressed in 18/25 patients whose Hgb < 100.
Regression model
In our univariate logistic regression model (Table 2), patient age and the value of the Hemoglobin measurement predicted that the iron status would be addressed. The following variables were not significantly associated: patient has a primary care physician, presence of fever, intravenous placement, admission to hospital, full-time versus part-time status of ordering Emergency Physician. In addition, there was no improvement in our rate of identifying children after our educational interventions.
Table 1 Assessment And Treatment Of Children Who Had A CBC Performed In The PED That Showed MCV < 70 fl.
N Iron Therapy Prescribed No Iron Therapy but Explicit Follow-Up Arranged Neither Iron Therapy nor Follow-up Documented
Hemoglobin Normal
MCV Low 33 2 1 30
Hemoglobin Low
MCV Low 45 19 5 21
All with low MCV 78 21 6 51
Hemoglobin cutoff 110 g/L; MCV (mean cell volume) cutoff 70 fL
Table 2 Factors Potentially Associated with Identification of Potential Iron Deficiency in Children with MCV <70
Total N Number in whom Iron Prescribed or Tests Arranged Odds Ratio 95% Confidence Interval
Age 6–18 months 48 22 4.2 1.4, 13
19–36 months 30 5
Febrile Yes 47 18 1.2 0.4, 3.4
No 20 7
IV Placed Yes 43 15 1.0 0.4, 2.6
No 35 12
Admitted Yes 14 8 3.2 1.0, 10
No 64 19
Hemoglobin >100 53 9 12.6 4.0, 39
≤99 25 18
7
Full-time PEM Yes 46 19 0.63 0.2, 1.7
Attending Physician No 26 8
Educational Pre 49 14 2.0 0.8, 5.3
Interventions Post 29 13
Discussion
We report the results of a cohort study examining the prevalence of an index of iron deficiency (MCV) in children who have had a blood test in the PED. We found that PED physicians did not reliably document either a diagnostic or therapeutic plan to deal with the possibility of iron deficiency suggesting that this opportunity for secondary screening is being missed.
The risk of iron deficiency in U.S. children aged 12–35 months is approximately 9% [11]. Major national organizations such the American Academy of Pediatrics suggest that all children be screened with a blood test at age 12 months [9]. This is not commonly done possibly due to the perception that the risk of problems due to iron deficiency, for any individual child, are low especially when the diet includes iron-fortified foods. In addition, the cost, both financial and in terms of pain and inconvenience, is a barrier to routine screening.
The population of patients that present to urban PEDs is likely to be of low socio-economic status and skewed towards the preschool age group [12]. Both of these factors (low socio-economic status, age 3–24 months) are also predictive of iron deficiency suggesting that a PED-based secondary screening program could be successful.
For the PED patient population, we were unable to identify a published estimate of the predictive value of hemoglobin value or CBC indices compared with gold standard measures of iron status. Both low MCV and low Hemoglobin are imperfect predictors of iron deficiency. Using the National Health and Nutrition Examination Survey III (1988–1994) data, White showed that the positive predictive value of a low Hemoglobin value (<110 g/L) for iron deficiency is only 29% (95%CI: 20–38%) in children 12–35 months [11]. Iron deficiency was defined as the presence of ≥2 of the following: low ferritin, low transferrin saturation or high erythrocyte protoporphyrin.
Oski suggested a staging classification of iron status in which a period of iron-deficient erythropoiesis, characterized by low MCV and high RDW but normal hemoglobin, preceded frank anemia [3]. We felt that by identifying children at that stage, potential iron-deficiency could be intercepted at an early stage with simple interventions such as administration of an oral iron preparation [10]. Hemoglobin levels may be low due to recent viral illnesses or they may be elevated in dehydrated patients [10]. Interestingly, in our population, we found that more children had low hemoglobin values than low MCV's. We can only speculate that this might be due to the effects of intercurrent illness since we had no laboratory studies of iron stores available to us.
The health impact of mild iron deficiency is controversial. Anemia due to iron deficiency may be associated with behavioural changes, impaired psychomotor development and impaired cognitive function [13]. A frequently cited study found that children who were frankly anemic scored significantly lower on a test of cognitive function than controls [14]. The difference would not be perceptible in an individual child but would have important implications at the population level. Children with abnormal indices but normal hemoglobin did not differ from controls. This study and one other randomized controlled trial suggest that iron therapy can reverse the developmental effects after at least two months of therapy [13-15].
Widespread screening for thalassemia traits has not been routine likely because it is clinically benign, not of immediate clinical importance for young pediatric patients, and because they are concentrated within certain genetic backgrounds; however, discovering a thalassemia trait in a patient provides an opportunity for genetic counseling.
This is opportunistic secondary screening. The cost has already been paid and there is no need to change existing workflow to take advantage of this information in contrast to screening at the time of the well-child visit [16]. While it will likely only benefit those patients who have had a CBC done for another reason, heightened awareness of this issue might predispose PED physicians to address this issue more frequently in their patients. For example, increased screening of dietary histories in children at risk might be effective [17].
Our retrospective study is susceptible to a documentation bias. Physicians may well have considered iron deficiency but not documented it on the chart. This is all the more likely given that a health maintenance issue would rarely have been the primary indication for the visit to the PED. However, in informal discussions, our staff related that they were likely to be missing this diagnosis. Also, the absence of a Hawthorne effect in the publicized second phase of data collection suggests that the physicians do not consistently consider this diagnosis. We note that even if the documentation bias caused us to under-report appropriate therapy by a factor to two, we would still be missing close to half of the opportunities for closer assessment of these children.
Why do our physicians, many of whom are primary care pediatricians, not use this flagged information? The answer may lie in cognitive psychology. We speculate that a normally adaptive heuristic may be working against the physician in this particular situation. Heuristics are short-cuts that humans use to process information more quickly [18]. The CBC report presents the time-pressured PED physician a list of 12 numbers in a densely-typed column, all in the same font size. Some of these numbers are clearly more important than others: specifically, the White Blood Count, Hemoglobin and Platelet Counts. It may be that the physicians have subconsciously made a decision not to process all the numbers. Instead they have trained themselves to scan the list for only the three most important values. In essence, they see the list of numbers through an opaque mask that has only three holes cut out of it. This idea is supported by better performance of the physicians in detecting abnormal hemoglobin levels than abnormal MCVs.
Another heuristic is also likely working against the physician in this case. In the search satisfying heuristic, the physician who ordered a lab test for one reason (e.g. fever) will likely most closely attend to the feature of the result (e.g. White Blood Count) that is directly related to the reason for ordering the test. This may lead them to miss unanticipated results (e.g. anemia). What is not clear is why the physicians did not attend to the asterisks used to flag abnormal results (another adaptive heuristic is to scan a long list of lab reports for asterisks and then specifically attend to those results). Careful attention to the display of the information could increase physician attention to abnormal results without decreasing their efficiency [19].
We were disappointed that our educational interventions did not improve the situation. From the PED charts it appears that the rate at which clinicians arranged assessment or treated patients with evidence of iron deficiency did not change after we had presented a round to our Division and sent a one-page memo to every physician in the PED. These standard interventions have been shown to be relatively ineffective in changing physician behaviour in many settings [20]. While we did observe a trend towards improvment, it is likely instead that we will need to reanalyze and change our process of care [21]. For example, we could establish a protocol whereby all CBC samples that meet certain criteria for iron deficiency would automatically have a Free Erythrocyte Protoporphyrin test run on the same sample. The result would then be reported to the ordering physician. This strategy would have the advantage that, besides confirming the diagnosis, the added report would draw the physician's attention to the potential problem.
Conclusion
In summary, we found that Pediatric Emergency Dept. physicians are not using all of the information available to them when they consider Complete Blood Count results. Careful attention to the Hemoglobin and MCV can suggest an opportunity for secondary prevention of iron-deficiency in children 6–36 months of age presenting to the ED. This information comes at no cost to the patient and can help avoid cognitive impairment in some children.
Abbreviations
CBC: Complete Blood Count
PED: Pediatric Emergency Department
MCV: Mean Cell Volume
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors participated in the design of the study. BD designed the data collection instrument, and with MP, did the chart reviews; DM and MP analyzed the data; MP drafted the manuscript with considerable revision by DM; All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors gratefully acknowledge the cheerful assistance of Nathalie Gauthier who queried the Laboratory Information System database to identify patients for us and Suzanne MacDonald who did several of the chart reviews. We are also grateful to Gudrun Aubertin who helped write the background section of this manuscript.
==== Refs
de Andraca I Castillo M Walter T Psychomotor development and behavior in iron-deficient anemic infants Nutrition Review 1997 55 125 132
Grantham-McGregor S Ani C A review of studies on the effect of iron deficiency on cognitive development in children Journal of Nutrition 2001 131 649S 666S 11160596
Oski FA Iron deficiency in infancy and childhood New England Journal of Medicine 1993 329 190 193 8515791 10.1056/NEJM199307153290308
Saloojee H Pettifor JM Iron deficiency and impaired child development British Medical Journal 2001 323 1377 1378 11744547
Walter T Effect of iron-deficiency anaemia on cognitive skills in infancy and childhood Baillieres Clinical Haematology 1994 7 815 827
Feightner JW Prevention of iron-deficiency anemia in infants Canadian Task Force on the Periodic Health Examination Canadian Guide to Clinical Preventive Health Care 1994 Health Canada. Oτταωα 244 255
Innis SM Nelson CM Wadsworth LD MacLaren IA Lwanga D Incidence of iron-deficiency anaemia and depleted iron stores among nine-month-old infants in Vancouver, Canada Canadian Journal of Public Health 1997 88 Revue Canadienne de Sante Publique 80 84
Bogen DL Krause JP Serwint JR Outcome of Children identified as Anemic by Routine Screening in an Inner-city Clinic Archives of Pediatrics & Adolescent Medicine 2001 155 366 371 11231803
American Academy of Pediatrics Kleinman R Iron deficiency Pediatric Nutrition Handbook 1998 233 246
Schnall SF Berliner N Duffy TP Benz EJ Jr Hoffman R, Benz EJ Jr, Shattil SJ, Furie B, Cohen HJ, Silberstein LE Approach to the Adult and Child with Anemia Hematology: Basic Principles and Practice 2003 Philadelphia: Churchill Livingstone 367 382
White KC Anemia is a poor predictor of iron deficiency among toddlers in the United States: For heme the bell tolls Pediatrics 2005 115 315 320 Ref Type: Journal (Full) 15687438 10.1542/peds.2004-1488
Krauss BS Harakal T Fleisher GR The spectrum and frequency of illness presenting to a pediatric emergency department Pediatr Emerg Care 1991 7 67 71 1904578
Logan S Martins S Gilbert R Iron therapy for improving psychomotor development and cognitive function in children under the age of three with iron deficiency anaemia Cochrane Database Syst Rev 2001 CD001444 11405989
Idjradinata P Pollitt E Reversal of developmental delays in iron-deficient anaemic infants treated with iron Lancet 1993 341 1 4 7678046 10.1016/0140-6736(93)92477-B
Aukett MA Parks YA Scott PH Wharton BA Treatment with iron increases weight gain and psychomotor development Arch Dis Child 1986 61 849 857 2429622
James JA Laing GJ Logan S Rossdale M Feasibility of screening toddlers for iron deficiency anaemia in general practice [see comments] BMJ 1997 315 102 103 9240051
Boutry M Needlman R Use of diet history in the screening of iron deficiency Pediatrics 1996 98 1138 1142 8951266
Croskerry P The cognitive imperative: thinking about how we think Acad Emerg Med 2000 7 1223 1231 11073470
Tufte E Visual Explanations 1997 Chesire Connecticut: Graphics Press
Davis DA Thomson MA Oxman AD Haynes RB Changing physician performance. A systematic review of the effect of continuing medical education strategies JAMA 1995 274 700 705 7650822 10.1001/jama.274.9.700
Lomas J Diffusion, dissemination, and implementation: who should do what? Ann N Y Acad Sci 1993 703 226 235 8192299
|
16300683
|
PMC1322220
|
CC BY
|
2021-01-04 16:31:06
|
no
|
BMC Pediatr. 2005 Nov 22; 5:42
|
utf-8
|
BMC Pediatr
| 2,005 |
10.1186/1471-2431-5-42
|
oa_comm
|
==== Front
BMC Cell BiolBMC Cell Biology1471-2121BioMed Central London 1471-2121-6-411631862510.1186/1471-2121-6-41Research ArticleComputational modeling reveals molecular details of epidermal growth factor binding Mayawala Kapil [email protected] Dionisios G [email protected] Jeremy S [email protected] Department of Chemical Engineering, University of Delaware, Newark, DE, USA2 Molecular Genetics and Microbiology, Cancer Research and Treatment Center, University of New Mexico Health Sciences Center, and Chemical and Nuclear Engineering, University of New Mexico, Albuquerque, NM, USA2005 30 11 2005 6 41 41 2 8 2005 30 11 2005 Copyright © 2005 Mayawala et al; licensee BioMed Central Ltd.2005Mayawala 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 ErbB family of receptors are dysregulated in a number of cancers, and the signaling pathway of this receptor family is a critical target for several anti-cancer drugs. Therefore a detailed understanding of the mechanisms of receptor activation is critical. However, despite a plethora of biochemical studies and recent single particle tracking experiments, the early molecular mechanisms involving epidermal growth factor (EGF) binding and EGF receptor (EGFR) dimerization are not as well understood. Herein, we describe a spatially distributed Monte Carlo based simulation framework to enable the simulation of in vivo receptor diffusion and dimerization.
Results
Our simulation results are in agreement with the data from single particle tracking and biochemical experiments on EGFR. Furthermore, the simulations reveal that the sequence of receptor-receptor and ligand-receptor reaction events depends on the ligand concentration, receptor density and receptor mobility.
Conclusion
Our computer simulations reveal the mechanism of EGF binding on EGFR. Overall, we show that spatial simulation of receptor dynamics can be used to gain a mechanistic understanding of receptor activation which may in turn enable improved cancer treatments in the future.
==== Body
Background
Amplification of genes for the ErbB family of receptors is associated with poor outcome in women's cancers, including breast, ovarian and endometrial cancer. Under non-pathological conditions, epidermal growth factor (EGF) receptor (EGFR) or ErbB1 is activated by ligand-induced receptor dimerization, resulting in autophosphorylation and phosphorylation of various cellular substrates [1]. However, while it is clear that overexpression is a factor leading to ligand-independent signaling via these receptors, the mechanism by which functional dimerization and activation occurs is unknown. Since EGF binding represents the initial step for activating EGFR, considerable work has been devoted to elucidating the mechanisms of ligand binding and dimerization [1-7]. However, molecular details of ligand-induced receptor dimerization are not as well understood.
Apart from in vitro biochemical experiments to study mechanisms of EGFR activation [1], recent developments in microscopy have made it possible to visualize protein dynamics in living cells [8]. The current imaging methods either have a high spatial resolution, such as electron microscopy experiments using immunogold labeling [9] and covalent linking to chemical conjugates like ferritin [10], or high temporal resolution, such as fluorescence confocal microscopy [11], single particle tracking [5] and more recently quantum dots based ligands [12]. However, with the currently available imaging technologies, combined high temporal and spatial resolution (of multiple receptors) has not been achieved.
Computational efforts devoted to understanding the extracellular mechanisms leading to EGFR activation are mostly equilibrium studies [7,13-15] or continuum reaction-diffusion models; see references in [3,16]. Continuum partial differential equation based models have also been used to represent signaling processes in the plasma membrane assuming a continuum distribution of receptors [17].
While such studies have provided useful insights, they are not ideally suited for describing cell surface heterogeneities, such as microdomains and anomalous diffusion of surface receptors [18], which are important to capture the spatiotemporal receptor dynamics, and lack spatial correlations, known to arise from bimolecular reaction events [19], such as dimerization. Monte Carlo (MC) techniques have proven powerful for systems biology modeling [20-22]. In the past, spatial MC approaches have provided mechanistic understanding in other biological systems; see for example [20,23-33].
In this study, we have used a spatial MC framework which not only enables a realistic representation of the plasma membrane, but also facilitates integration of different types of biological data produced from biochemical and microscopy studies to gain insight into the mechanistic details of the underlying biological process. We have developed a general kinetic, lattice MC modeling framework to model the ligand (EGF) binding and dimerization of the EGFR. We compare our simulation results with single particle tracking experiments and analyzed the dominant mechanism of ligand binding and dimerization.
Results
Comparison of stochastic and deterministic models
Microscopic events modeled in this work are shown in Figs. 1 and 2. In order to test the MC algorithm and explore possible differences between stochastic and deterministic models, we have performed a number of simulations for various parameters. Results from the hybrid null-event MC algorithm were compared with an ordinary differential equations (ODEs) model with a set of parameters for which the process is reaction limited, i.e., diffusion is fast compared to reaction, in order to test the validity of the MC algorithm. Specifically, we simulated a dimerization reaction in the absence of ligand considering a high receptor number density of 11,000 per μm2 and assuming no dimers initially. Fig. 3(a) compares the concentration trajectories of dimerized EGFR. This comparison confirms that the hybrid null-event MC algorithm captures the time scales of the system resulting in the correct transient concentration profile. Additional validation carried out under diffusion control has again demonstrated the accuracy of our MC method (Mayawala et al., in preparation).
Figure 1 Schematic of simulated microscopic events. Each receptor can diffuse to an empty neighboring site, react with a neighboring receptor to form a dimer, and bind ligand. All events are reversible.
Figure 2 Reactions events considered in our model as given in [14].
Figure 3 Comparison of hybrid null-event MC and ODE models in terms of (a) dimerized EGFR in the absence of ligand at a high receptor density and diffusivity (11,000 per μm2, D = 2 × 10-14 m2sec-1) and assuming no initial dimers, and (b) EGF bound EGFR in the presence of ligand (160 nM) at low receptor density (125 per per μm2) and D = 2 × 10-15 m2sec-1. The reactions on the figure indicate the dominant processes responsible for the concentration trajectories. Error bars indicate 2 standard deviations obtained from 10 independent MC simulations.
Next in Fig. 3(b) we compared the MC and ODE concentration profiles of EGF bound EGFR monomer in the presence of ligand (160 nM), with a receptor number density of 125 receptors per μm2 and a low diffusivity of 2 × 10-15 m2s-1. The low values of receptor density and diffusivity result in a diffusion controlled case. Corresponding to these parameters, the receptor dimerization rate in the spatial MC model was slower compared to that of the ODE model. The diffusion limited dimerization of EGF bound EGFR monomer leads to a higher concentration of unbound receptor in the spatial MC model than in the ODE model. Thus, spatiotemporal MC simulations are required to capture the transient concentration profiles of the signaling species under diffusion limited conditions. Overall, low receptor densities and low diffusivities may render the system diffusion limited. Under such conditions, well-mixed simulations do not provide accurate dynamics. Use of spatial MC bypasses the question whether the system is diffusion or reaction limited. In a forthcoming communication, we will address quantitatively the conditions for which spatial MC simulations are needed.
Partial differential equations (PDEs) have traditionally been used to model diffusion-reaction processes when spatial effects become important. However, accurate representation of receptor-receptor reactions typically requires MC simulation due to the spatial inhomogeneous distribution of receptors stemming from spatial correlations [19,28]. Aside from spatial correlations, realistic representation of the plasma membrane microdomains and anomalous diffusion make MC simulation indispensable [18]. Due to these limitations, PDE models have not been employed here.
Comparison of hybrid null-event MC simulations with single particle tracking experiments
The dynamics of the ligand binding events were compared with the single particle tracking experiment of Sako et al. [5] at an EGF concentration of 0.16 nM in the 0–60 sec time interval. To compare simulation results with experimental data, EGF was assumed to be associated with Cy3 dye. A dimerized receptor with two EGF molecules was taken to fluoresce twice as intensely as a receptor (single or dimerized) bound to one EGF molecule. The predicted initial increase of low intensity spots (monomers plus dimers having one EGF bound) followed by a slower increase in high intensity spots (dimers with 2 molecules of EGF) is qualitatively consistent with the experimental data (see Fig. 4(a)). The initial increase in the low intensity signal was due to the rapid binding of EGF on predimerized EGFR. Furthermore, the increase in the total density of Cy3-EGF spots (total bound EGF on all receptors), shown in Fig. 4(b), is also consistent with experimental data.
Figure 4 (a) Evolution of intensity of dimerized receptors with two ligands (high intensity spots) and of monomer plus dimerized receptors with a single ligand bound (low intensity spots) along with the data of single particle tracking experiments by Sako et al. over time intervals of 20 sec. The simulations were performed for a receptor number density of 5500 per μm2, a diffusivity of D = 2 × 10-14 m2sec-1, and 18% dimers initially. The simulation intensity has been normalized with the experimental data. (b) Comparison of predicted density of Cy3-EGF spots with experimental data of Sako et al. The densities are normalized with the value at 60 sec. Good agreement of simulations with experimental data is found. In both panels, error bars indicate 2 standard deviations obtained from 10 independent MC simulations.
The possible sequences of events leading to the formation of EGF bound dimerized EGFR at 60 sec are shown in Fig. 5. Sako et al. [5] suggested sequence 1 as being dominant. However, the experimental study alone cannot unambiguously determine the sequence due to its limited spatial resolution and the fact that only ligand bound receptors can be tracked. Our simulations showed that 95–100% of the receptors follow sequence 1, 0–4.9% sequence 2, and the remaining receptors follow sequence 3. Our results are consistent with the hypothesis of Sako et al. [5]. This comparison serves as a model validation step. Small adjustments (20–30%) in the equilibrium and kinetic parameters tabulated in Table 1, which are well within the margins of error, lead to nearly proportional changes in intensity, i.e., no dramatic differences in the simulation profiles are seen (see appendix for details).
Figure 5 Sequence of reactions resulting in dimerized receptors with both receptors bound to ligand for simulations of Fig. 4. All reactions are reversible.
Table 1 Kinetic (reaction events given in Fig. 2) and transport (monomer and dimer diffusion) parameters used in hybrid null-event MC model (factors of 1/2 and 1/4 discussed in the Methods section have to be considered).
Equilibrium constants
K1 13.3 (molecule/site)-1 In the range to be consistent with ~18% of the monomer EGFR as dimers in the absence of EGF [9, 53–56]
K2 4.0 × 103 (molecule/site)-1 Calculated based on equilibrium relations given in [14]
K3 1.2 × 106 (molecule/site)-1 Calculated based on equilibrium relations given in [14]
K4 4 × 108 M-1 In the range suggested by [1, 4, 58–60]
K5, K6 1.2 × 1011 M-1 In the range suggested by [9, 53–56]
Kinetic parameters
k1b, k2b 0.17 sec-1 [61]
k3b 1.7 × 10-3 sec-1 [61]
k4b, k5b 2.9 × 10-3 sec-1 [34]
k6b 5.8 × 10-3 sec-1 [34]
Transport parameters
Dmonomer 2 × 10-14–2 × 10-15 m2sec-1 [49, 50]
Effect of ligand concentration on signaling reaction mechanism in A-431 cells (high receptor density)
Single particle tracking experiments [5] are typically limited to low ligand concentrations. High concentration of ligand would lead to fluorescence of a large number of EGFRs making it impossible to visualize individual particles. However, simulations can be used to elucidate the influence of extracellular EGF concentration on EGFR dimerization. Our simulations indicated that the relative contributions of sequences 1–3 at 60 sec change with ligand concentration (Fig. 6(a)). At low ligand concentration, sequence 1 dominates, whereas at higher ligand concentration, a significant fraction of dimers form via sequence 2. Furthermore, sequence 3 also occurs to appreciable extent at high concentration of EGF. At low ligand concentration, most of the ligand gets bound to dimerized receptors, which have a higher ligand affinity; however, the extent to which free EGFR dimerization can occur is limited. At higher ligand concentration, when a significant fraction of ligand is attached to monomers, the coupling between ligand attached monomer and free or ligand attached monomer gives rise to dimers. The relative contribution of the sequences also changes with time. Specifically, initial ligand binding occurs on predimerized receptors, and hence, the relative contribution of sequence 1 is higher at short times. At longer times, after binding of ligand on monomers, sequences 2 and 3 start contributing. With an increase in ligand concentration, the contributions of sequences 2 and 3 increase at a faster rate. The contribution of sequence 3 is higher at longer times after accumulation of ligand bound monomers. As a final note, the time needed to reach equilibrium substantially decreases as the concentration of ligand increases (not shown), e.g., to a total of a few sec at 160 nM. As a result, high ligand concentrations may challenge single particle tracking experiments also in terms of temporal resolution.
Figure 6 Contributions of the different reaction mechanisms at 60 sec for different concentrations of EGF with (a) a receptor number density of 5500 receptors per μm2 and D = 2 × 10-14 m2sec-1, (b) a receptor number density of 125 receptors per μm2 and D = 2 × 10-14 m2sec-1, and (c) a receptor number density of 125 receptors per μm2 and D = 2 × 10-15 m2sec-1.
Support for the suggested mechanisms also comes from biochemical studies. The experimental study of [34] reported that at low doses of EGF, inhibition of high affinity binding by mAb108 can kill almost 50–100% of EGF binding, indicating that most of the early binding takes place by sequence 1 at low EGF concentration. However, this inhibition is overcome at higher concentration (~20–50 times) of EGF, which is indicative of substantial formation of EGF bound dimerized EGFR via sequence 2, consistent with the results of our simulations. A larger scale simulation with variable receptor densities in different regions of the plasma membrane will be developed in the future for quantitative comparison with such biochemical experiments. A recent equilibrium based study [13] has shown that such spatial heterogeneities have strong influence on the amount of EGF binding on EGFR, motivating a more detailed analysis of EGFR on the plasma membrane.
Effect of ligand concentration and receptor mobility on signaling reaction mechanism in cells with normal receptor density
Two important factors influencing ligand binding and dimerization are the receptor density and receptor mobility. The receptor density can significantly influence the mechanism of EGF binding as shown in Fig. 6(b). At lower receptor density (125 receptors per μm2) sequence 1 occurs to a much lower extent as compared to the A-431 cells. For this lower receptor density, at lower EGF concentration sequence 2 is dominant, whereas at higher EGF concentrations, sequence 3 is dominant. Sequence 1 is not important at low receptor density, because of the low amount of EGF free dimers (negligible at the low receptor density considered in this work).
A tenfold decrease in receptor mobility (from 2 × 10-14 m2/s to 2 × 10-15 m2/s) leads to a very small increase in the extent of sequence 3, at the expense of sequences 1 and 2 (compare Figs. 6(b) and 6(c)). This small increase is observed only at low EGF concentration. At higher EGF concentration this increase is even smaller. Sequence 3 occurs to a larger extent at slower diffusion because dimerization is slowing down and so more monomers associate with ligand. At higher EGF concentration, this effect is not as prominent because EGF binding is faster leading to more EGF bound EGFRs, thereby increased dimerization occurs among EGF bound monomer EGFRs even with a higher receptor diffusivity.
Several studies have indicated inhomogeneities in the plasma membrane and excellent reviews have been published on this topic including [18,35-38]. These studies have suggested localization of receptors within small regions, called microdomains, in the plasma membrane. An implication of the containment of receptors in the microdomains is the observation of lower macroscopic diffusivity as has been discussed in [39]. As a result, the microscopic diffusivity can potentially be at least 1–2 orders of magnitude faster than the diffusivity reported in literature. Therefore, we have also studied the effect of a higher diffusivity. In contrast to decreasing diffusivity from 2 × 10-15 m2/s to 2 × 10-14 m2/s mentioned above, larger changes are observed at high ligand concentration (e.g., 1600 nM) and a receptor density of 125 receptors per μm2 for a change in diffusivity from 2 × 10-14 m2/s to 2 × 10-13 m2/s. Specifically, the contribution of sequence 2 increases from ~15% to ~30% at the expense of sequence 3 which decreases from ~85% to ~70%. An increase in receptor diffusivity leads to an increased rate of dimerization between an occupied and a free receptor in comparison to ligand binding on a free receptor. Overall, a faster diffusivity can lead to an overall increase in the dimerization rate but this effect is not dramatic under our simulation conditions.
Discussion
Our simulation results suggest future single particle tracking experiments or related microscopy experiments. It may be difficult to perform the single particle tracking experiments of [5] at higher ligand concentration in A-431 cells due to the difficulty in visualization of single EGFR and possibly to the short time scales over which transients are over. However, such experiments can potentially be performed in cells with a lower average receptor density. On such cells, the increased contributions of sequences 2 and 3 should be observed to further validate our model. Possible discrepancies between experiments and model could provide new insights to enhance our current understanding of the underlying signaling processes.
The variation in receptor density and receptor mobility can stem from different cell types as well as different spatial features/locations in the plasma membrane (see Methods section for references). Future microscopy experiments should be designed to observe the reaction events and transients of low and high intensity spots, as reported by [5], in different domains of the plasma membrane in the same cell. Such data can then be used to estimate the local density of the receptors which in turn can help in understanding the receptor distribution in the plasma membrane.
This work shows the influence of receptor density and receptor mobility as a biophysical control of signaling processes over the inflexible thermodynamic and biochemical properties. A key suggestion from this work is that it is not adequate to treat the receptor-receptor interactions based only on their kinetic and thermodynamic parameters. Instead, their spatial properties, such as local receptor densities and local lateral mobility, can play significant roles in determining the intracellular signaling. Herein, we have used a simplistic representation of the plasma membrane. This model can be extended to include experimentally reported anomalous hop-like diffusion [18] and other spatial features, such as clathrin coated pits, lipid rafts, caveolae and other microdomains [35,40]. Such features not only facilitate an enhanced control at the level of plasma membrane but can also be important for the wide diversity of signaling outcomes from limited varieties of ligands and receptors.
Conclusion
We have developed a computational framework for studying cell surface receptor dynamics that can bridge biochemical data on one hand with various microscopy experiments on the other, which currently lack simultaneous high spatial and temporal resolution. This work provides an important step forward in the era of in vivo imaging based modeling approaches [8,41]. For example, in this work comparison of MC simulations with single particle tracking experiments reveals how the sequence of receptor-receptor and ligand-receptor reaction events depends on the ligand concentration, receptor density and receptor mobility. Our computer simulations reveal the underlying mechanism on the plasma membrane leading to dimerized and ligand (EGF) bound receptors.
Considering the interest in targeted antibodies for cancer therapeutics [42], a detailed understanding of the biochemical mechanisms involved in signal sensing at the plasma membrane is desired. Future advancements in medicine will ultimately also include the mathematical analysis and modeling of ErbB receptor diffusion, dimerization and clustering, which will increase our understanding of tumorigenesis and lead to medical advances, such as individualized therapy for heterogeneous cancers.
Methods
A hybrid null-event algorithm
A coarse, molecular level based computational framework that leaves atomistic details out, e.g., conformations and vibrations of proteins into potential energy surface minima, but still provides the sequence of molecular events at the receptor length scale was employed. Herein, we have utilized a kinetic, lattice MC method for simulating the EGFR dynamics. Microscopic events modeled include receptor dimerization and decomposition, ligand-receptor association and dissociation, and Brownian diffusion of receptors (see Figs. 1 and 2). Formation of high-mers that happens at longer times is not considered in this work.
The existence of multiple timescales in the system and low surface density of receptors make lattice MC simulations computationally prohibitive. We have devised a hybrid between the continuous time MC method [43] and the null-event MC method (see [44] for an overview of spatial MC algorithms) to increase the speed of the null-event algorithm but maintain its flexibility. In our hybrid null-event algorithm, only lattice sites filled with receptors were randomly selected, resulting in two to four orders of magnitude speedup, depending on receptor density, relative to a traditional null-event MC algorithm where all sites are randomly chosen. Additionally, operations that are often involved in a continuous time MC method, such as summation of transition probabilities and searches over the entire lattice, are avoided, resulting in further acceleration of simulations.
The spatial domain was represented using a two-dimensional square lattice that was initially randomly populated with a given density of receptors. Periodic boundary conditions were employed [45]. After initialization, an occupied site was randomly picked and one of the microscopic events is possibly selected to occur based on probabilities described below.
Transition probability of diffusion
The probability of diffusion per unit time in all four directions on a square lattice was calculated using random walk theory from the diffusivity
Γd=4Da2, (1)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqHtoWrdaqhaaqaaaqaaiabdsgaKbaacqGH9aqpdaWcaaqaaiabisda0iabdseaebqaaiabdggaHnaaDaaabaaabaGaeGOmaidaaaaacqGGSaalcaWLjaGaaCzcamaabmaabaGaeGymaedacaGLOaGaayzkaaaaaa@39A2@
where a is the microscopic lattice pixel dimension (taken here to be 2 nm) and D is the corresponding diffusivity of a receptor or a dimer. The transition probability of diffusion per unit time in moving from site i to site j is
Γi→jd=14Γdσi(1−σj) j∈Bi (2)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqHtoWrdaqhaaqaaiabdMgaPjabgkziUkabdQgaQbqaaiabdsgaKbaacqGH9aqpdaWcaaqaaiabigdaXaqaaiabisda0aaacqqHtoWrdaqhaaqaaaqaaiabdsgaKbaacqaHdpWCdaWgaaqaaiabdMgaPbqabaGaeiikaGIaeGymaeJaeyOeI0Iaeq4Wdm3aaSbaaeaacqWGQbGAaeqaaiabcMcaPiaaykW7caaMc8UaaGPaVlaaykW7caaMc8UaemOAaOMaeyicI4SaemOqai0aaSbaaSqaaiabdMgaPbqabaGccaWLjaGaaCzcamaabmaabaGaeGOmaidacaGLOaGaayzkaaaaaa@5509@
where Bi denotes the set of sites to which diffusion from site i can occur. In our model, this set includes all 4 first-nearest neighboring sites. σi is the occupancy (discrete) function that is 1, if site i is filled, or 0, if site i is empty (a single index indicating the site is herein used to simplify notation). According to Eq. (2), the transition probability of diffusion per unit time along any direction on the square lattice can be 0 or 14Γd
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabigdaXaqaaiabisda0aaacqqHtoWrdaqhaaqaaaqaaiabdsgaKbaaaaa@317F@, depending on the occupancy of the first-nearest neighboring site in the corresponding direction.
Transition probability of reactions
The transition probability of a reaction was obtained in terms of the macroscopic reaction rate constants, k. For a first-order (e.g., the decomposition of an EGFR dimer) or pseudo-first order (e.g., the EGF binding onto a receptor because EGF is assumed at a constant concentration) reaction (see Fig. 2), one has
A→C, Γir=kσi (3)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGbbqqcqGHsgIRcqqGdbWqcqqGSaalcaaMc8UaaGPaVlaaykW7cqqHtoWrdaqhaaqaaiabdMgaPbqaaiabdkhaYbaacqGH9aqpcqWGRbWAiiaacqWFdpWCdaWgaaqaaiabdMgaPbqabaGaaCzcaiaaxMaadaqadaqaaiabiodaZaGaayjkaiaawMcaaaaa@43ED@
where Γir
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqHtoWrdaqhaaqaaiabdMgaPbqaaiabdkhaYbaaaaa@3100@ is the transition probability of reaction at site i. For a bimolecular reaction on a square lattice (e.g., receptor-receptor dimerization), the transition probability per unit time at selected site i was modeled as:
for the reaction: A + B→C, Γir=k4σiσj, (4)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGMbGzcqqGVbWBcqqGYbGCcqqGGaaicqqG0baDcqqGObaAcqqGLbqzcqqGGaaicqqGYbGCcqqGLbqzcqqGHbqycqqGJbWycqqG0baDcqqGPbqAcqqGVbWBcqqGUbGBcqqG6aGocqqGGaaicqqGbbqqcqqGGaaicqqGRaWkcqqGGaaicqqGcbGqcqGHsgIRieaacqWFdbWqcqGGSaalcqqGGaaicqqHtoWrdaqhaaqaaiab=LgaPbqaaiab=jhaYbaacqGH9aqpdaWcaaqaaiab=TgaRbqaaiabisda0aaaiiaacqGFdpWCdaWgaaqaaiab=LgaPbqabaGae43Wdm3aaSbaaeaacqWFQbGAaeqaaiabcYcaSiaaxMaacaWLjaWaaeWaaeaacqaI0aanaiaawIcacaGLPaaaaaa@5F0B@
and for the reaction: 2A→C, Γir=k2σiσj. (5)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGHbqycqqGUbGBcqqGKbazcqqGGaaicqqGMbGzcqqGVbWBcqqGYbGCcqqGGaaicqqG0baDcqqGObaAcqqGLbqzcqqGGaaicqqGYbGCcqqGLbqzcqqGHbqycqqGJbWycqqG0baDcqqGPbqAcqqGVbWBcqqGUbGBcqqG6aGocqqGGaaicqqGYaGmcqqGbbqqcqGHsgIRcqqGdbWqcqqGSaalcqqGGaaicqqHtoWrdaqhaaqaaGqaaiab=LgaPbqaaiab=jhaYbaacqGH9aqpdaWcaaqaaiab=TgaRbqaaiabikdaYaaaiiaacqGFdpWCdaWgaaqaaiab=LgaPbqabaGae43Wdm3aaSbaaeaacqWFQbGAaeqaaiabc6caUiaaxMaacaWLjaWaaeWaaeaacqaI1aqnaiaawIcacaGLPaaaaaa@6145@
Here the reacting species (A and B or A and A) occupy adjacent sites i and j, and the units of k are (molecules/site)-1sec-1. The factor of four in k/4 in Eq. (4) accounts for the fact that all four neighboring sites of site i were randomly chosen to search for the existence of species B. Similarly, the factor of 2 in Eq. 5 was due to the degeneracy of species participating in the homodimerization.
Event selection and time advancement
After an occupied site, say i, was selected, the transition probabilities per unit time of all possible events were computed. The probability for a certain event 'x' at site i was calculated as
pix=ΓixΓmax, (6)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGWbaCdaqhaaqaaiabdMgaPbqaaiabdIha4baacqGH9aqpdaWcaaqaaiabfo5ahnaaDaaabaGaemyAaKgabaGaemiEaGhaaaqaaiabfo5ahnaaBaaabaGagiyBa0MaeiyyaeMaeiiEaGhabeaaaaGaeiilaWIaaCzcaiaaxMaadaqadaqaaiabiAda2aGaayjkaiaawMcaaaaa@40D7@
where Γmax is a normalization constant to ensure that the selection probability of each event is always less than or equal to 1 and Γix
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqHtoWrdaqhaaqaaiabdMgaPbqaaiabdIha4baaaaa@310C@ is the transition probability of event x (reaction or diffusion) at site i, as defined above. The maximum values of transition probabilities per unit time of events described by Eqs. (2)-(4) are Γd/4, k, k/4, and k/2, respectively. We defined the normalization constant as
Γmax=4(Γd4+max{∑all forward reaction eventsΓr}),+max{∑all backward reaction eventsΓr}
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@9C41@
where multiplication by a factor of 4 accounts for the microscopic process in each of the four directions on the square lattice. For the reaction events given in Fig. 2
Γmax=Γd+4(k1f2+k2f4+k3f2)+∑i=46kif+∑i=16kib, (7)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@6577@
where the subscripts f and b refer to the forward and backward reaction events, respectively, and i denotes the corresponding reaction according to Fig. 2. A random number 'r' was finally chosen from a uniform distribution between 0 and 1. The events were randomly ranked. The smallest value of m satisfying ∑x=1mpix>r
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaaeWbqaaiabdchaWnaaDaaabaGaemyAaKgabaGaemiEaGhaaaqaaiabdIha4jabg2da9iabigdaXaqaaiabd2gaTbGaeyyeIuoacqGH+aGpcqWGYbGCaaa@3A7E@ criterion was chosen. If no value of m satisfied the above criterion, then no event was selected to occur (a null-event) and a new occupied site was again randomly picked. Otherwise, the mth event was selected, the populations on the lattice were updated accordingly to reflect the stoichiometry of the reaction or diffusion process, and the real time was advanced based on the most frequently selected event as suggested in [44]. In our simulations, real time was advanced based on the diffusion of receptors according to which the average time step after each successful diffusion event was calculated as
Δt=1∑i=1No.ofsites(σi∑j∈BiΓi→jd(1-σj))=114Γd∑occupiedsites(∑j∈Bi(1-σj)). (8)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaiiaacqWFuoarieaacqGF0baDcqGF9aqpdaWcaaqaaiab+fdaXaqaamaaqahabaWaaeWaaeaacqWFdpWCdaWgaaqaaiab+LgaPbqabaWaaabCaeaacqWFtoWrdaqhaaqaaiab+LgaPjabgkziUkab+PgaQbqaaiab+rgaKbaadaqadaqaaiab+fdaXiab+1caTiab=n8aZnaaBaaaleaacqGFQbGAaeqaaaGccaGLOaGaayzkaaaabaGae4NAaOMae8hcI4Sae4Nqai0aaSbaaSqaaiab+LgaPbqabaaakeaaaiabggHiLdaacaGLOaGaayzkaaaabaGae4xAaKMae4xpa0Jae4xmaedabaGae4Nta4Kae43Ba8Mae4Nla4Iae4hiaaIae43Ba8Mae4NzayMae4hiaaIae43CamNae4xAaKMae4hDaqNae4xzauMae43CamhacqGHris5aaaacqGF9aqpdaWcaaqaaiab+fdaXaqaamaalaaabaGae4xmaedabaGae4hnaqdaaiab=n5ahnaaCaaabeqaaiab+rgaKbaadaaeWbqaamaabmaabaWaaabCaeaadaqadaqaaiab+fdaXiab+1caTiab=n8aZnaaBaaaleaacqGFQbGAaeqaaaGccaGLOaGaayzkaaaabaGae4NAaOMae8hcI4Sae4Nqai0aaSbaaSqaaiab+LgaPbqabaaakeaaaiabggHiLdaacaGLOaGaayzkaaaabaGae43Ba8Mae43yamMae43yamMae4xDauNae4hCaaNae4xAaKMae4xzauMae4hzaqMae4hiaaIae43CamNae4xAaKMae4hDaqNae4xzauMae43CamhabaaacqGHris5aaaacqGFGaaicqGFUaGlcaWLjaGaaCzcamaabmaabaGae4hoaGdacaGLOaGaayzkaaaaaa@9004@
The summation in the denominator was updated each time a successful event happens by subtracting the previous occupancy values affected by the selected event and adding the new ones. In this way, the summation was carried out only once (at the beginning of a simulation) over all occupied sites. Finally, a new site is picked randomly until the desired real time is reached.
The hybrid null-event MC algorithm explained above was implemented (for details on the null-event algorithm see [44]) using Fortran 90. For each data point, 10 simulations with different seeds of the random number generator were used to collect statistics.
Simulation size and model parameters
In this work, the cell surface was represented using a 2-dimensional square lattice with each pixel being 2 nm × 2 nm in size. The number density of receptors ranges from ~102 receptors per μm2 on normal cells [46] to ~103 receptors per μm2 on human epithelioid carcinoma cells (A-431 cells) which overexpress EGFR [2]. However, the local density of receptors can be much higher in-vivo because of the localization of receptors in certain regions of the plasma membrane, such as in lipid rafts [35,47,48]. We simulated a low density (to represent normal cells) of 31 receptors on 500 nm × 500 nm mesh and a high density (to represent A-431 cells) of 55 receptors on 100 nm × 100 nm mesh, which are equivalent to receptor number densities of 125 and 5500 per μm2, respectively. The diffusivity of monomer EGFR has been reported to be around 2 × 10-14 m2s-1 [49,50]. Lower macroscopic diffusivities for EGFR have also been observed [49], which may be due to containment within cytoskeletal elements [51] or lipid rafts [52]. Based on these suggestions, the effect of a slower diffusivity is also analyzed by considering a diffusivity of 2 × 10-15 m2s-1. Simulations have been performed in the 0–60 sec time interval to capture the initial transients.
The model parameters are summarized in Table 1. We assumed two types of EGF binding on the cell surface: low affinity binding (on monomer EGFR) and high affinity binding (on dimerized EGFR). Experimental information suggests the existence of a high affinity (for receptor-ligand association) receptor population, most of which, if not all, is present in the form of dimers; see review by [1]. A recent equilibrium study has also shown that this interpretation is consistent with the experimentally reported concave-up shape of the Scatchard plot [13].
Experimental studies have provided evidence of predimerized receptors on A-431 cells to different extents [9,53-56]. Consequently, a fraction (~82%) of receptors was initially placed at random locations as monomers and the remaining as dimers on simulated A-431 cells for comparison with single particle tracking data. Corresponding to the dimerization equilibrium constant for this data, we found that at the lower receptor number density of 125 per μm2, there is negligible number of dimers in the absence of ligand. The receptor dimerization constants vary with ligand occupancy. Several experimental studies have shown that dimerization between unbounded receptors occurs with lower affinity than that between one bounded and one unbounded receptor. Finally, dimerization between two ligand bounded receptors occurs with the highest affinity [6,57].
Appendix
A sensitivity analysis was performed in which each kinetic parameter (ki, i = 1f, 4f, 5f, 6f, 1b, 4b, 5b, 6b) was increased by 20%, and the change in the high intensity spots was observed at three different times (20, 40 and 60 sec) from the mean of 10 independent MC simulations. The normalized sensitivity coefficient, reported in Fig. 7, is defined as
Figure 7 Normalized sensitivity coefficients at 3 different times (20, 40 and 60 sec) calculated by introducing a 20% increase in the kinetic parameter indicated on the x-axis.
(I−Io)/Io(ki−kio)/kio=I−Io0.2Io, (A1)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@5543@
where I is the % of high intensity spots (y axis in Fig. 4a) upon perturbing the kinetic parameter, ki, and Io is the nominal value corresponding to the original set of kinetic parameters, kio. Only 8 of the 12 kinetic parameters were independently perturbed because of the two equilibrium constrains reported in [14], i.e.,
K2=K1K5K4, and (A2)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGlbWsdaWgaaWcbaGaeGOmaidabeaakiabg2da9maalaaabaGaem4saS0aaSbaaSqaaiabigdaXaqabaGccqWGlbWsdaWgaaWcbaGaeGynaudabeaaaOqaaiabdUealnaaBaaaleaacqaI0aanaeqaaaaakiabcYcaSiabbccaGiabbggaHjabb6gaUjabbsgaKjabbccaGiaaxMaacaWLjaWaaeWaaeaacqqGbbqqcqqGYaGmaiaawIcacaGLPaaaaaa@4213@
K3=K1K5K6(K4)2. (A3)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGlbWsdaWgaaWcbaGaeG4mamdabeaakiabg2da9maalaaabaGaem4saS0aaSbaaSqaaiabigdaXaqabaGccqWGlbWsdaWgaaWcbaGaeGynaudabeaakiabdUealnaaBaaaleaacqaI2aGnaeqaaaGcbaWaaeWaaeaacqWGlbWsdaWgaaWcbaGaeGinaqdabeaaaOGaayjkaiaawMcaamaaCaaaleqabaGaeGOmaidaaaaakiabc6caUiaaxMaacaWLjaWaaeWaaeaacqqGbbqqcqqGZaWmaiaawIcacaGLPaaaaaa@4193@
The equilibrium relations determine the changes in the kinetic parameters of the dependent reactions 2 and 3 (see Fig. 2) upon perturbing those of the linearly independent reactions. A change in an equilibrium constant can be associated with a change in the forward, backward, or both rate constants. For simplicity a change in the rate constant of a forward (backward) linearly independent reaction is taken to cause a change in the forward (backward) rate constant of the linearly dependent reactions. Specifically, one has
k2f=(k2f)ofk1ffk5ffk4f, (A4)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGRbWAdaWgaaWcbaGaeGOmaiJaemOzaygabeaakiabg2da9maabmaabaGaem4AaS2aaSbaaSqaaiabikdaYiabdAgaMbqabaaakiaawIcacaGLPaaadaWgaaWcbaGaem4Ba8gabeaakmaalaaabaGaemOzay2aaSbaaSqaaiabdUgaRnaaBaaameaacqaIXaqmcqWGMbGzaeqaaaWcbeaakiabdAgaMnaaBaaaleaacqWGRbWAdaWgaaadbaGaeGynauJaemOzaygabeaaaSqabaaakeaacqWGMbGzdaWgaaWcbaGaem4AaS2aaSbaaWqaaiabisda0iabdAgaMbqabaaaleqaaaaakiabcYcaSiaaxMaacaWLjaWaaeWaaeaacqqGbbqqcqqG0aanaiaawIcacaGLPaaaaaa@4E8B@
k3f=(k3f)ofk1ffk5ffk6f(fk4f)2, (A5)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@56B4@
k2b=(k2b)ofk1bfk5bfk4b, and (A6)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@53F0@
k3b=(k3b)ofk1bfk5bfk6b(fk4b)2, (A7)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@5688@
where subscript 'o' denotes the nominal value, and fki
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGMbGzdaWgaaWcbaGaem4AaS2aaSbaaWqaaiabdMgaPbqabaaaleqaaaaa@3122@ is 1, if the kinetic parameter (ki) is not perturbed, and 1.2, if the parameter is increased by 20%.
Authors' contributions
KM carried out the simulations and drafted the manuscript. DGV and JSE edited the manuscript. All authors participated in the analysis of the data. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by grants from the US Department of Energy (DE-FG02-05ER25702) and the National Science Foundation (CTS-0312117). KM thanks Abhijit Chatterjee for useful discussions.
==== Refs
Jorissen RN Walker F Pouliot N Garrett TPJ Ward CW Burgess AW Epidermal growth factor receptor: mechanisms of activation and signalling Experimental Cell Research 2003 284 31 53 12648464 10.1016/S0014-4827(02)00098-8
Wiley HS Anomalous binding of epidermal growth factor to A431 cells is due to the effect of high receptor densities and a saturable endocytic system J Cell Biol 1988 107 801 810 3262110 10.1083/jcb.107.2.801
Wiley HS Shvartsman SY Lauffenburger DA Computational modeling of the EGF-receptor system: a paradigm for systems biology Trends in Cell Biology 2003 13 43 50 12480339 10.1016/S0962-8924(02)00009-0
Schlessinger J Allosteric regulation of the epidermal growth factor receptor kinase J Cell Biol 1986 103 2067 2072 3023396 10.1083/jcb.103.6.2067
Sako Y Minoghchi S Yanagida T Single-molecule imaging of EGFR signalling on the surface of living cells Nature Cell Biology 2000 2 168 172 10707088 10.1038/35004044
Lemmon MA Bu Z Ladbury JE Zhou M Pinchasi D Lax I Engelman DM Schlessinger J Two EGF molecules contribute additively to stabilization of the EGFR dimer The EMBO Journal 1997 16 281 294 9029149 10.1093/emboj/16.2.281
Klein P Mattoon D Lemmon MA Schlessinger J A structure-based model for ligand binding and dimerization of EGF receptors PNAS 2004 101 929 934 14732694 10.1073/pnas.0307285101
Weijer CJ Visualizing signals moving in cells Science 2003 300 96 100 12677060 10.1126/science.1082830
Vanbelzen N Rijken PJ Hage WJ Delaat SW Verkleij AJ Boonstra J Direct visualization and quantitative-analysis of epidermal growth factor-induced receptor clustering Journal Of Cellular Physiology 1988 134 413 420 3258312 10.1002/jcp.1041340312
Haigler HT McKanna JA Cohen S Direct visualization of the binding and internalization of a ferritin conjugate of epidermal growth factor in human carcinoma cells A-431 J Cell Biol 1979 81 382 395 313931 10.1083/jcb.81.2.382
Verveer PJ Wouters FS Reynolds AR Bastiaens PIH Quantitative Imaging of Lateral ErbB1 Receptor Signal Propagation in the Plasma Membrane Science 2000 290 1567 1570 11090353 10.1126/science.290.5496.1567
Lidke DS Nagy P Heintzmann R Arndt-Jovin DJ Post JN Grecco HE Jares-Erijman EA Jovin TM Quantum dot ligands provide new insights into erbB/HER receptor–mediated signal transduction Nature Biotechnology 2004 22 198 1203 14704683 10.1038/nbt929
Mayawala K Vlachos DG Edwards JS Heterogeneities in EGF receptor density at the cell surface can lead to concave up scatchard plot of EGF binding FEBS Letters 2005 579 3043 3047 15896781 10.1016/j.febslet.2005.04.059
Wofsy C Goldstein B Lund K Wiley HS Implications of epidermal growth factor (EGF) induced EGF receptor aggregation Biophys J 1992 63 98 110 1420877
Chamberlin SG Davies DE A unified model of c-erbB receptor homo- and heterodimerisation Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology 1998 1384 223 232 10.1016/S0167-4838(97)00203-3
Lauffenburger DA Linderman JJ Receptors models for binding, trafficking, and signaling 1993 New York, Oxford University Press
Haugh JM A unified model for signal transduction reactions in cellular membranes Biophysical Journal 2002 82 591 604 11806904
Kusumi A Nakada C Ritchie K Murase K Suzuki K Murakoshi H Kasai RS Kondo J Fujiwara T Paradigm Shift of the Plasma Membrane Concept from the Two-Dimensional Continuum Fluid to the Partitioned Fluid: High-Speed Single-Molecule Tracking of Membrane Molecules Annual Review of Biophysics and Biomolecular Structure 2005 34 351 378 10.1146/annurev.biophys.34.040204.144637
Ziff RM Gulari E Barshad Y Kinetic phase transitions in an irreversible surface-reaction model Physical Review Letters 1986 56 2553 2556 10033028 10.1103/PhysRevLett.56.2553
Slepchenko BM Schaff JC Carson JH Loew LM Computational Cell Biology: Spatiotemporal Simulation of Cellular Events Annual Review of Biophysics and Biomolecular Structure 2002 31 423 441 10.1146/annurev.biophys.31.101101.140930
Takahashi K Arjunan SNV Tomita M Space in systems biology of signaling pathways – towards intracellular molecular crowding in silico FEBS Letters 2005 579 1783 1788 15763552 10.1016/j.febslet.2005.01.072
Lemerle C Ventura BD Serrano L Space as the final frontier in stochastic simulations of biological systems FEBS Letters 2005 579 1789 1794 15763553 10.1016/j.febslet.2005.02.009
Saxton MJ Single-particle tracking: effects of corrals Biophys J 1995 69 389 398 8527652
Andrews SS Bray D Stochastic simulation of chemical reactions with spatial resolution and single molecule detail Physical Biology 2004 1 137 151 16204833 10.1088/1478-3967/1/3/001
Goldman J Andrews S Bray D Size and composition of membrane protein clusters predicted by Monte Carlo analysis European Biophysics Journal 2004 33 506 512 14997356 10.1007/s00249-004-0391-6
Levin MD Shimizu TS Bray D Binding and Diffusion of CheR Molecules Within a Cluster of Membrane Receptors Biophys J 2002 82 1809 1817 11916840
Mahama PA Linderman JJ A Monte Carlo study of the dynamics of G-protein activation Biophys J 1994 67 1345 1357 7811949
Shea LD Omann GM Linderman JJ Calculation of diffusion-limited kinetics for the reactions in collision coupling and receptor cross-linking Biophys J 1997 73 2949 2959 9414209
Shimizu TS Aksenov SV Bray D A Spatially Extended Stochastic Model of the Bacterial Chemotaxis Signalling Pathway Journal of Molecular Biology 2003 329 291 309 12758077 10.1016/S0022-2836(03)00437-6
Stiles JR Jr TMB Salpeter EE Salpeter MM Bower JM Monte Carlo simulation of neurotransmitter release using MCell, a general simulator of cellular physiological processes Computational Neuroscience 1998 New York, Plenum 279 284
Woolf PJ Linderman JJ Untangling Ligand Induced Activation and Desensitization of G-Protein-Coupled Receptors Biophys J 2003 84 3 13 12524261
Woolf PJ Linderman JJ An algebra of dimerization and its implications for G-protein coupled receptor signaling Journal of Theoretical Biology 2004 229 157 168 15207471 10.1016/j.jtbi.2004.03.012
Zhong H Wade SM Woolf PJ Linderman JJ Traynor JR Neubig RR A Spatial Focusing Model for G Protein Signals. Regulator Of G Protein Signaling (Rgs) Protein-Mediated Kinetic Scaffolding J Biol Chem 2003 278 7278 7284 12446706 10.1074/jbc.M208819200
Bellot F Moolenaar W Kris R Mirakhur B Verlaan I Ullrich A Schlessinger J Felder S High-affinity epidermal growth factor binding is specifically reduced by a monoclonal antibody, and appears necessary for early responses J Cell Biol 1990 110 491 502 2298813 10.1083/jcb.110.2.491
Laude AJ Prior IA Plasma membrane microdomains: organization, function and trafficking (Review) Molecular Membrane Biology 2004 21 193 205 15204627 10.1080/09687680410001700517
Parton RG Hancock JF Lipid rafts and plasma membrane microorganization: insights from Ras Trends in Cell Biology 2004 14 141 147 15003623 10.1016/j.tcb.2004.02.001
Kusumi A Sako Y Cell surface organization by the membrane skeleton Current Opinion in Cell Biology 1996 8 566 574 8791449 10.1016/S0955-0674(96)80036-6
Vereb G Szollosi J Matko J Nagy P Farkas T Vigh L Matyus L Waldmann TA Damjanovich S Dynamic, yet structured: The cell membrane three decades after the Singer-Nicolson model PNAS 2003 100 8053 8058 12832616 10.1073/pnas.1332550100
Ritchie K Shan XY Kondo J Iwasawa K Fujiwara T Kusumi A Detection of Non-Brownian Diffusion in the Cell Membrane in Single Molecule Tracking Biophys J 2005 88 2266 2277 15613635 10.1529/biophysj.104.054106
Roy CL Wrana JL Clathrin- and non-clathrin-mediated endocytic regulation of cell signalling Nature Reviews Molecular Cell Biology 2005 6 112 126 15687999 10.1038/nrm1571
Phair RD Misteli T Kinetic Modeling Approaches to In Vivo Imaging Nature Reviews Molecular Cell Biology 2001 2 898 907 11733769 10.1038/35103000
Roskoski JR The ErbB/HER receptor protein-tyrosine kinases and cancer Biochemical and Biophysical Research Communications 2004 319 1 11 15158434 10.1016/j.bbrc.2004.04.150
Gillespie DT Exact stochastic simulation of coupled chemical reactions Journal of Chemical Physics 1977 81 2340 2361 10.1021/j100540a008
Reese JS Raimondeau S Vlachos DG Monte Carlo algorithms for complex surface reaction mechanisms: Efficiency and accuracy Journal of Computational Physics 2001 173 302 321 10.1006/jcph.2001.6877
Jakobsson E Computer simulation studies of biological membranes: progress, promise and pitfalls Trends in Biochemical Sciences 1997 22 339 344 9301334 10.1016/S0968-0004(97)01096-7
Benveniste M Livneh E Schlessinger J Kam Z Overexpression of epidermal growth factor receptor in NIH-3T3- transfected cells slows its lateral diffusion and rate of endocytosis J Cell Biol 1988 106 1903 1909 3133376 10.1083/jcb.106.6.1903
Simons K Toomre D Lipid rafts and signal transduction Nature Reviews Molecular Cell Biology 2000 1 31 39 11413487 10.1038/35036052
Pike LJ Lipid rafts: bringing order to chaos J Lipid Res 2003 44 655 667 12562849 10.1194/jlr.R200021-JLR200
Kusumi A Sako Y Yamamoto M Confined lateral diffusion of membrane receptors as studied by single particle tracking (nanovid microscopy). Effects of calcium-induced differentiation in cultured epithelial cells Biophys J 1993 65 2021 2040 8298032
Rees AR Gregoriou M Johnson P Garland PB High affinity epidermal growth factor receptors on the surface of A431 cells have restricted lateral diffusion The EMBO Journal 1984 3 1843 1847 6090128
Wiegant FA Blok FJ Defize LH Linnemans WA Verkley AJ Boonstra J Epidermal growth factor receptors associated to cytoskeletal elements of epidermoid carcinoma (A431) cells J Cell Biol 1986 103 87 94 3013901 10.1083/jcb.103.1.87
Pralle A Keller P Florin EL Simons K Horber JKH Sphingolipid-Cholesterol Rafts Diffuse as Small Entities in the Plasma Membrane of Mammalian Cells J Cell Biol 2000 148 997 1008 10704449 10.1083/jcb.148.5.997
Martin-Fernandez M Clarke DT Tobin MJ Jones SV Jones GR Preformed Oligomeric Epidermal Growth Factor Receptors Undergo an Ectodomain Structure Change during Signaling Biophys J 2002 82 2415 2427 11964230
Yu X Sharma KD Takahashi T Iwamoto R Mekada E Ligand-independent Dimer Formation of Epidermal Growth Factor Receptor (EGFR) Is a Step Separable from Ligand-induced EGFR Signaling Mol Biol Cell 2002 13 2547 2557 12134089 10.1091/mbc.01-08-0411
Moriki T Maruyama H Maruyama IN Activation of preformed EGF receptor dimers by ligand-induced rotation of the transmembrane domain1 Journal of Molecular Biology 2001 311 1011 1026 11531336 10.1006/jmbi.2001.4923
Gadella TWJ Jovin TM Oligomerization of epidermal growth factor receptors on A431 cells studied by time-resolved fluorescence imaging microscopy. A stereochemical model for tyrosine kinase receptor activation J Cell Biol 1995 129 1543 1558 7790353 10.1083/jcb.129.6.1543
Sherrill JM Kyte J Activation of epidermal growth factor receptor by epidermal growth factor Biochemistry 1996 35 5705 5718 8639530 10.1021/bi9602268
Schlessinger J The epidermal growth factor receptor as a multifunctional allosteric protein Biochemistry 1988 27 3119 3123 3291943 10.1021/bi00409a002
Kawamoto T Sato JD Le A Polikoff J Sato GH Mendelsohn J Growth stimulation of A431 cells by epidermal growth factor: identification of high-affinity receptors for epidermal growth factor by an anti-receptor monoclonal antibody PNAS 1983 80 1337 1341 6298788
Lin CR Chen WS Lazar CS Carpenter CD Gill GN Evans RM Rosenfeld MG Protein kinase C phosphorylation at Thr 654 of the unoccupied EGF receptor and EGF binding regulate functional receptor loss by independent mechanisms Cell 1986 44 839 848 3006920 10.1016/0092-8674(86)90006-1
Hendriks BS Opresko LK Wiley HS Lauffenburger D Quantitative Analysis of HER2-mediated Effects on HER2 and Epidermal Growth Factor Receptor Endocytosis: Distribution Of Homo- And Heterodimers Depends On Relative Her2 Levels J Biol Chem 2003 278 23343 23351 12686539 10.1074/jbc.M300477200
|
16318625
|
PMC1322221
|
CC BY
|
2021-01-04 16:31:30
|
no
|
BMC Cell Biol. 2005 Nov 30; 6:41
|
utf-8
|
BMC Cell Biol
| 2,005 |
10.1186/1471-2121-6-41
|
oa_comm
|
==== Front
BMC Cell BiolBMC Cell Biology1471-2121BioMed Central London 1471-2121-6-411631862510.1186/1471-2121-6-41Research ArticleComputational modeling reveals molecular details of epidermal growth factor binding Mayawala Kapil [email protected] Dionisios G [email protected] Jeremy S [email protected] Department of Chemical Engineering, University of Delaware, Newark, DE, USA2 Molecular Genetics and Microbiology, Cancer Research and Treatment Center, University of New Mexico Health Sciences Center, and Chemical and Nuclear Engineering, University of New Mexico, Albuquerque, NM, USA2005 30 11 2005 6 41 41 2 8 2005 30 11 2005 Copyright © 2005 Mayawala et al; licensee BioMed Central Ltd.2005Mayawala 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 ErbB family of receptors are dysregulated in a number of cancers, and the signaling pathway of this receptor family is a critical target for several anti-cancer drugs. Therefore a detailed understanding of the mechanisms of receptor activation is critical. However, despite a plethora of biochemical studies and recent single particle tracking experiments, the early molecular mechanisms involving epidermal growth factor (EGF) binding and EGF receptor (EGFR) dimerization are not as well understood. Herein, we describe a spatially distributed Monte Carlo based simulation framework to enable the simulation of in vivo receptor diffusion and dimerization.
Results
Our simulation results are in agreement with the data from single particle tracking and biochemical experiments on EGFR. Furthermore, the simulations reveal that the sequence of receptor-receptor and ligand-receptor reaction events depends on the ligand concentration, receptor density and receptor mobility.
Conclusion
Our computer simulations reveal the mechanism of EGF binding on EGFR. Overall, we show that spatial simulation of receptor dynamics can be used to gain a mechanistic understanding of receptor activation which may in turn enable improved cancer treatments in the future.
==== Body
Background
Amplification of genes for the ErbB family of receptors is associated with poor outcome in women's cancers, including breast, ovarian and endometrial cancer. Under non-pathological conditions, epidermal growth factor (EGF) receptor (EGFR) or ErbB1 is activated by ligand-induced receptor dimerization, resulting in autophosphorylation and phosphorylation of various cellular substrates [1]. However, while it is clear that overexpression is a factor leading to ligand-independent signaling via these receptors, the mechanism by which functional dimerization and activation occurs is unknown. Since EGF binding represents the initial step for activating EGFR, considerable work has been devoted to elucidating the mechanisms of ligand binding and dimerization [1-7]. However, molecular details of ligand-induced receptor dimerization are not as well understood.
Apart from in vitro biochemical experiments to study mechanisms of EGFR activation [1], recent developments in microscopy have made it possible to visualize protein dynamics in living cells [8]. The current imaging methods either have a high spatial resolution, such as electron microscopy experiments using immunogold labeling [9] and covalent linking to chemical conjugates like ferritin [10], or high temporal resolution, such as fluorescence confocal microscopy [11], single particle tracking [5] and more recently quantum dots based ligands [12]. However, with the currently available imaging technologies, combined high temporal and spatial resolution (of multiple receptors) has not been achieved.
Computational efforts devoted to understanding the extracellular mechanisms leading to EGFR activation are mostly equilibrium studies [7,13-15] or continuum reaction-diffusion models; see references in [3,16]. Continuum partial differential equation based models have also been used to represent signaling processes in the plasma membrane assuming a continuum distribution of receptors [17].
While such studies have provided useful insights, they are not ideally suited for describing cell surface heterogeneities, such as microdomains and anomalous diffusion of surface receptors [18], which are important to capture the spatiotemporal receptor dynamics, and lack spatial correlations, known to arise from bimolecular reaction events [19], such as dimerization. Monte Carlo (MC) techniques have proven powerful for systems biology modeling [20-22]. In the past, spatial MC approaches have provided mechanistic understanding in other biological systems; see for example [20,23-33].
In this study, we have used a spatial MC framework which not only enables a realistic representation of the plasma membrane, but also facilitates integration of different types of biological data produced from biochemical and microscopy studies to gain insight into the mechanistic details of the underlying biological process. We have developed a general kinetic, lattice MC modeling framework to model the ligand (EGF) binding and dimerization of the EGFR. We compare our simulation results with single particle tracking experiments and analyzed the dominant mechanism of ligand binding and dimerization.
Results
Comparison of stochastic and deterministic models
Microscopic events modeled in this work are shown in Figs. 1 and 2. In order to test the MC algorithm and explore possible differences between stochastic and deterministic models, we have performed a number of simulations for various parameters. Results from the hybrid null-event MC algorithm were compared with an ordinary differential equations (ODEs) model with a set of parameters for which the process is reaction limited, i.e., diffusion is fast compared to reaction, in order to test the validity of the MC algorithm. Specifically, we simulated a dimerization reaction in the absence of ligand considering a high receptor number density of 11,000 per μm2 and assuming no dimers initially. Fig. 3(a) compares the concentration trajectories of dimerized EGFR. This comparison confirms that the hybrid null-event MC algorithm captures the time scales of the system resulting in the correct transient concentration profile. Additional validation carried out under diffusion control has again demonstrated the accuracy of our MC method (Mayawala et al., in preparation).
Figure 1 Schematic of simulated microscopic events. Each receptor can diffuse to an empty neighboring site, react with a neighboring receptor to form a dimer, and bind ligand. All events are reversible.
Figure 2 Reactions events considered in our model as given in [14].
Figure 3 Comparison of hybrid null-event MC and ODE models in terms of (a) dimerized EGFR in the absence of ligand at a high receptor density and diffusivity (11,000 per μm2, D = 2 × 10-14 m2sec-1) and assuming no initial dimers, and (b) EGF bound EGFR in the presence of ligand (160 nM) at low receptor density (125 per per μm2) and D = 2 × 10-15 m2sec-1. The reactions on the figure indicate the dominant processes responsible for the concentration trajectories. Error bars indicate 2 standard deviations obtained from 10 independent MC simulations.
Next in Fig. 3(b) we compared the MC and ODE concentration profiles of EGF bound EGFR monomer in the presence of ligand (160 nM), with a receptor number density of 125 receptors per μm2 and a low diffusivity of 2 × 10-15 m2s-1. The low values of receptor density and diffusivity result in a diffusion controlled case. Corresponding to these parameters, the receptor dimerization rate in the spatial MC model was slower compared to that of the ODE model. The diffusion limited dimerization of EGF bound EGFR monomer leads to a higher concentration of unbound receptor in the spatial MC model than in the ODE model. Thus, spatiotemporal MC simulations are required to capture the transient concentration profiles of the signaling species under diffusion limited conditions. Overall, low receptor densities and low diffusivities may render the system diffusion limited. Under such conditions, well-mixed simulations do not provide accurate dynamics. Use of spatial MC bypasses the question whether the system is diffusion or reaction limited. In a forthcoming communication, we will address quantitatively the conditions for which spatial MC simulations are needed.
Partial differential equations (PDEs) have traditionally been used to model diffusion-reaction processes when spatial effects become important. However, accurate representation of receptor-receptor reactions typically requires MC simulation due to the spatial inhomogeneous distribution of receptors stemming from spatial correlations [19,28]. Aside from spatial correlations, realistic representation of the plasma membrane microdomains and anomalous diffusion make MC simulation indispensable [18]. Due to these limitations, PDE models have not been employed here.
Comparison of hybrid null-event MC simulations with single particle tracking experiments
The dynamics of the ligand binding events were compared with the single particle tracking experiment of Sako et al. [5] at an EGF concentration of 0.16 nM in the 0–60 sec time interval. To compare simulation results with experimental data, EGF was assumed to be associated with Cy3 dye. A dimerized receptor with two EGF molecules was taken to fluoresce twice as intensely as a receptor (single or dimerized) bound to one EGF molecule. The predicted initial increase of low intensity spots (monomers plus dimers having one EGF bound) followed by a slower increase in high intensity spots (dimers with 2 molecules of EGF) is qualitatively consistent with the experimental data (see Fig. 4(a)). The initial increase in the low intensity signal was due to the rapid binding of EGF on predimerized EGFR. Furthermore, the increase in the total density of Cy3-EGF spots (total bound EGF on all receptors), shown in Fig. 4(b), is also consistent with experimental data.
Figure 4 (a) Evolution of intensity of dimerized receptors with two ligands (high intensity spots) and of monomer plus dimerized receptors with a single ligand bound (low intensity spots) along with the data of single particle tracking experiments by Sako et al. over time intervals of 20 sec. The simulations were performed for a receptor number density of 5500 per μm2, a diffusivity of D = 2 × 10-14 m2sec-1, and 18% dimers initially. The simulation intensity has been normalized with the experimental data. (b) Comparison of predicted density of Cy3-EGF spots with experimental data of Sako et al. The densities are normalized with the value at 60 sec. Good agreement of simulations with experimental data is found. In both panels, error bars indicate 2 standard deviations obtained from 10 independent MC simulations.
The possible sequences of events leading to the formation of EGF bound dimerized EGFR at 60 sec are shown in Fig. 5. Sako et al. [5] suggested sequence 1 as being dominant. However, the experimental study alone cannot unambiguously determine the sequence due to its limited spatial resolution and the fact that only ligand bound receptors can be tracked. Our simulations showed that 95–100% of the receptors follow sequence 1, 0–4.9% sequence 2, and the remaining receptors follow sequence 3. Our results are consistent with the hypothesis of Sako et al. [5]. This comparison serves as a model validation step. Small adjustments (20–30%) in the equilibrium and kinetic parameters tabulated in Table 1, which are well within the margins of error, lead to nearly proportional changes in intensity, i.e., no dramatic differences in the simulation profiles are seen (see appendix for details).
Figure 5 Sequence of reactions resulting in dimerized receptors with both receptors bound to ligand for simulations of Fig. 4. All reactions are reversible.
Table 1 Kinetic (reaction events given in Fig. 2) and transport (monomer and dimer diffusion) parameters used in hybrid null-event MC model (factors of 1/2 and 1/4 discussed in the Methods section have to be considered).
Equilibrium constants
K1 13.3 (molecule/site)-1 In the range to be consistent with ~18% of the monomer EGFR as dimers in the absence of EGF [9, 53–56]
K2 4.0 × 103 (molecule/site)-1 Calculated based on equilibrium relations given in [14]
K3 1.2 × 106 (molecule/site)-1 Calculated based on equilibrium relations given in [14]
K4 4 × 108 M-1 In the range suggested by [1, 4, 58–60]
K5, K6 1.2 × 1011 M-1 In the range suggested by [9, 53–56]
Kinetic parameters
k1b, k2b 0.17 sec-1 [61]
k3b 1.7 × 10-3 sec-1 [61]
k4b, k5b 2.9 × 10-3 sec-1 [34]
k6b 5.8 × 10-3 sec-1 [34]
Transport parameters
Dmonomer 2 × 10-14–2 × 10-15 m2sec-1 [49, 50]
Effect of ligand concentration on signaling reaction mechanism in A-431 cells (high receptor density)
Single particle tracking experiments [5] are typically limited to low ligand concentrations. High concentration of ligand would lead to fluorescence of a large number of EGFRs making it impossible to visualize individual particles. However, simulations can be used to elucidate the influence of extracellular EGF concentration on EGFR dimerization. Our simulations indicated that the relative contributions of sequences 1–3 at 60 sec change with ligand concentration (Fig. 6(a)). At low ligand concentration, sequence 1 dominates, whereas at higher ligand concentration, a significant fraction of dimers form via sequence 2. Furthermore, sequence 3 also occurs to appreciable extent at high concentration of EGF. At low ligand concentration, most of the ligand gets bound to dimerized receptors, which have a higher ligand affinity; however, the extent to which free EGFR dimerization can occur is limited. At higher ligand concentration, when a significant fraction of ligand is attached to monomers, the coupling between ligand attached monomer and free or ligand attached monomer gives rise to dimers. The relative contribution of the sequences also changes with time. Specifically, initial ligand binding occurs on predimerized receptors, and hence, the relative contribution of sequence 1 is higher at short times. At longer times, after binding of ligand on monomers, sequences 2 and 3 start contributing. With an increase in ligand concentration, the contributions of sequences 2 and 3 increase at a faster rate. The contribution of sequence 3 is higher at longer times after accumulation of ligand bound monomers. As a final note, the time needed to reach equilibrium substantially decreases as the concentration of ligand increases (not shown), e.g., to a total of a few sec at 160 nM. As a result, high ligand concentrations may challenge single particle tracking experiments also in terms of temporal resolution.
Figure 6 Contributions of the different reaction mechanisms at 60 sec for different concentrations of EGF with (a) a receptor number density of 5500 receptors per μm2 and D = 2 × 10-14 m2sec-1, (b) a receptor number density of 125 receptors per μm2 and D = 2 × 10-14 m2sec-1, and (c) a receptor number density of 125 receptors per μm2 and D = 2 × 10-15 m2sec-1.
Support for the suggested mechanisms also comes from biochemical studies. The experimental study of [34] reported that at low doses of EGF, inhibition of high affinity binding by mAb108 can kill almost 50–100% of EGF binding, indicating that most of the early binding takes place by sequence 1 at low EGF concentration. However, this inhibition is overcome at higher concentration (~20–50 times) of EGF, which is indicative of substantial formation of EGF bound dimerized EGFR via sequence 2, consistent with the results of our simulations. A larger scale simulation with variable receptor densities in different regions of the plasma membrane will be developed in the future for quantitative comparison with such biochemical experiments. A recent equilibrium based study [13] has shown that such spatial heterogeneities have strong influence on the amount of EGF binding on EGFR, motivating a more detailed analysis of EGFR on the plasma membrane.
Effect of ligand concentration and receptor mobility on signaling reaction mechanism in cells with normal receptor density
Two important factors influencing ligand binding and dimerization are the receptor density and receptor mobility. The receptor density can significantly influence the mechanism of EGF binding as shown in Fig. 6(b). At lower receptor density (125 receptors per μm2) sequence 1 occurs to a much lower extent as compared to the A-431 cells. For this lower receptor density, at lower EGF concentration sequence 2 is dominant, whereas at higher EGF concentrations, sequence 3 is dominant. Sequence 1 is not important at low receptor density, because of the low amount of EGF free dimers (negligible at the low receptor density considered in this work).
A tenfold decrease in receptor mobility (from 2 × 10-14 m2/s to 2 × 10-15 m2/s) leads to a very small increase in the extent of sequence 3, at the expense of sequences 1 and 2 (compare Figs. 6(b) and 6(c)). This small increase is observed only at low EGF concentration. At higher EGF concentration this increase is even smaller. Sequence 3 occurs to a larger extent at slower diffusion because dimerization is slowing down and so more monomers associate with ligand. At higher EGF concentration, this effect is not as prominent because EGF binding is faster leading to more EGF bound EGFRs, thereby increased dimerization occurs among EGF bound monomer EGFRs even with a higher receptor diffusivity.
Several studies have indicated inhomogeneities in the plasma membrane and excellent reviews have been published on this topic including [18,35-38]. These studies have suggested localization of receptors within small regions, called microdomains, in the plasma membrane. An implication of the containment of receptors in the microdomains is the observation of lower macroscopic diffusivity as has been discussed in [39]. As a result, the microscopic diffusivity can potentially be at least 1–2 orders of magnitude faster than the diffusivity reported in literature. Therefore, we have also studied the effect of a higher diffusivity. In contrast to decreasing diffusivity from 2 × 10-15 m2/s to 2 × 10-14 m2/s mentioned above, larger changes are observed at high ligand concentration (e.g., 1600 nM) and a receptor density of 125 receptors per μm2 for a change in diffusivity from 2 × 10-14 m2/s to 2 × 10-13 m2/s. Specifically, the contribution of sequence 2 increases from ~15% to ~30% at the expense of sequence 3 which decreases from ~85% to ~70%. An increase in receptor diffusivity leads to an increased rate of dimerization between an occupied and a free receptor in comparison to ligand binding on a free receptor. Overall, a faster diffusivity can lead to an overall increase in the dimerization rate but this effect is not dramatic under our simulation conditions.
Discussion
Our simulation results suggest future single particle tracking experiments or related microscopy experiments. It may be difficult to perform the single particle tracking experiments of [5] at higher ligand concentration in A-431 cells due to the difficulty in visualization of single EGFR and possibly to the short time scales over which transients are over. However, such experiments can potentially be performed in cells with a lower average receptor density. On such cells, the increased contributions of sequences 2 and 3 should be observed to further validate our model. Possible discrepancies between experiments and model could provide new insights to enhance our current understanding of the underlying signaling processes.
The variation in receptor density and receptor mobility can stem from different cell types as well as different spatial features/locations in the plasma membrane (see Methods section for references). Future microscopy experiments should be designed to observe the reaction events and transients of low and high intensity spots, as reported by [5], in different domains of the plasma membrane in the same cell. Such data can then be used to estimate the local density of the receptors which in turn can help in understanding the receptor distribution in the plasma membrane.
This work shows the influence of receptor density and receptor mobility as a biophysical control of signaling processes over the inflexible thermodynamic and biochemical properties. A key suggestion from this work is that it is not adequate to treat the receptor-receptor interactions based only on their kinetic and thermodynamic parameters. Instead, their spatial properties, such as local receptor densities and local lateral mobility, can play significant roles in determining the intracellular signaling. Herein, we have used a simplistic representation of the plasma membrane. This model can be extended to include experimentally reported anomalous hop-like diffusion [18] and other spatial features, such as clathrin coated pits, lipid rafts, caveolae and other microdomains [35,40]. Such features not only facilitate an enhanced control at the level of plasma membrane but can also be important for the wide diversity of signaling outcomes from limited varieties of ligands and receptors.
Conclusion
We have developed a computational framework for studying cell surface receptor dynamics that can bridge biochemical data on one hand with various microscopy experiments on the other, which currently lack simultaneous high spatial and temporal resolution. This work provides an important step forward in the era of in vivo imaging based modeling approaches [8,41]. For example, in this work comparison of MC simulations with single particle tracking experiments reveals how the sequence of receptor-receptor and ligand-receptor reaction events depends on the ligand concentration, receptor density and receptor mobility. Our computer simulations reveal the underlying mechanism on the plasma membrane leading to dimerized and ligand (EGF) bound receptors.
Considering the interest in targeted antibodies for cancer therapeutics [42], a detailed understanding of the biochemical mechanisms involved in signal sensing at the plasma membrane is desired. Future advancements in medicine will ultimately also include the mathematical analysis and modeling of ErbB receptor diffusion, dimerization and clustering, which will increase our understanding of tumorigenesis and lead to medical advances, such as individualized therapy for heterogeneous cancers.
Methods
A hybrid null-event algorithm
A coarse, molecular level based computational framework that leaves atomistic details out, e.g., conformations and vibrations of proteins into potential energy surface minima, but still provides the sequence of molecular events at the receptor length scale was employed. Herein, we have utilized a kinetic, lattice MC method for simulating the EGFR dynamics. Microscopic events modeled include receptor dimerization and decomposition, ligand-receptor association and dissociation, and Brownian diffusion of receptors (see Figs. 1 and 2). Formation of high-mers that happens at longer times is not considered in this work.
The existence of multiple timescales in the system and low surface density of receptors make lattice MC simulations computationally prohibitive. We have devised a hybrid between the continuous time MC method [43] and the null-event MC method (see [44] for an overview of spatial MC algorithms) to increase the speed of the null-event algorithm but maintain its flexibility. In our hybrid null-event algorithm, only lattice sites filled with receptors were randomly selected, resulting in two to four orders of magnitude speedup, depending on receptor density, relative to a traditional null-event MC algorithm where all sites are randomly chosen. Additionally, operations that are often involved in a continuous time MC method, such as summation of transition probabilities and searches over the entire lattice, are avoided, resulting in further acceleration of simulations.
The spatial domain was represented using a two-dimensional square lattice that was initially randomly populated with a given density of receptors. Periodic boundary conditions were employed [45]. After initialization, an occupied site was randomly picked and one of the microscopic events is possibly selected to occur based on probabilities described below.
Transition probability of diffusion
The probability of diffusion per unit time in all four directions on a square lattice was calculated using random walk theory from the diffusivity
Γd=4Da2, (1)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqHtoWrdaqhaaqaaaqaaiabdsgaKbaacqGH9aqpdaWcaaqaaiabisda0iabdseaebqaaiabdggaHnaaDaaabaaabaGaeGOmaidaaaaacqGGSaalcaWLjaGaaCzcamaabmaabaGaeGymaedacaGLOaGaayzkaaaaaa@39A2@
where a is the microscopic lattice pixel dimension (taken here to be 2 nm) and D is the corresponding diffusivity of a receptor or a dimer. The transition probability of diffusion per unit time in moving from site i to site j is
Γi→jd=14Γdσi(1−σj) j∈Bi (2)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqHtoWrdaqhaaqaaiabdMgaPjabgkziUkabdQgaQbqaaiabdsgaKbaacqGH9aqpdaWcaaqaaiabigdaXaqaaiabisda0aaacqqHtoWrdaqhaaqaaaqaaiabdsgaKbaacqaHdpWCdaWgaaqaaiabdMgaPbqabaGaeiikaGIaeGymaeJaeyOeI0Iaeq4Wdm3aaSbaaeaacqWGQbGAaeqaaiabcMcaPiaaykW7caaMc8UaaGPaVlaaykW7caaMc8UaemOAaOMaeyicI4SaemOqai0aaSbaaSqaaiabdMgaPbqabaGccaWLjaGaaCzcamaabmaabaGaeGOmaidacaGLOaGaayzkaaaaaa@5509@
where Bi denotes the set of sites to which diffusion from site i can occur. In our model, this set includes all 4 first-nearest neighboring sites. σi is the occupancy (discrete) function that is 1, if site i is filled, or 0, if site i is empty (a single index indicating the site is herein used to simplify notation). According to Eq. (2), the transition probability of diffusion per unit time along any direction on the square lattice can be 0 or 14Γd
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabigdaXaqaaiabisda0aaacqqHtoWrdaqhaaqaaaqaaiabdsgaKbaaaaa@317F@, depending on the occupancy of the first-nearest neighboring site in the corresponding direction.
Transition probability of reactions
The transition probability of a reaction was obtained in terms of the macroscopic reaction rate constants, k. For a first-order (e.g., the decomposition of an EGFR dimer) or pseudo-first order (e.g., the EGF binding onto a receptor because EGF is assumed at a constant concentration) reaction (see Fig. 2), one has
A→C, Γir=kσi (3)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGbbqqcqGHsgIRcqqGdbWqcqqGSaalcaaMc8UaaGPaVlaaykW7cqqHtoWrdaqhaaqaaiabdMgaPbqaaiabdkhaYbaacqGH9aqpcqWGRbWAiiaacqWFdpWCdaWgaaqaaiabdMgaPbqabaGaaCzcaiaaxMaadaqadaqaaiabiodaZaGaayjkaiaawMcaaaaa@43ED@
where Γir
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqHtoWrdaqhaaqaaiabdMgaPbqaaiabdkhaYbaaaaa@3100@ is the transition probability of reaction at site i. For a bimolecular reaction on a square lattice (e.g., receptor-receptor dimerization), the transition probability per unit time at selected site i was modeled as:
for the reaction: A + B→C, Γir=k4σiσj, (4)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGMbGzcqqGVbWBcqqGYbGCcqqGGaaicqqG0baDcqqGObaAcqqGLbqzcqqGGaaicqqGYbGCcqqGLbqzcqqGHbqycqqGJbWycqqG0baDcqqGPbqAcqqGVbWBcqqGUbGBcqqG6aGocqqGGaaicqqGbbqqcqqGGaaicqqGRaWkcqqGGaaicqqGcbGqcqGHsgIRieaacqWFdbWqcqGGSaalcqqGGaaicqqHtoWrdaqhaaqaaiab=LgaPbqaaiab=jhaYbaacqGH9aqpdaWcaaqaaiab=TgaRbqaaiabisda0aaaiiaacqGFdpWCdaWgaaqaaiab=LgaPbqabaGae43Wdm3aaSbaaeaacqWFQbGAaeqaaiabcYcaSiaaxMaacaWLjaWaaeWaaeaacqaI0aanaiaawIcacaGLPaaaaaa@5F0B@
and for the reaction: 2A→C, Γir=k2σiσj. (5)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqGHbqycqqGUbGBcqqGKbazcqqGGaaicqqGMbGzcqqGVbWBcqqGYbGCcqqGGaaicqqG0baDcqqGObaAcqqGLbqzcqqGGaaicqqGYbGCcqqGLbqzcqqGHbqycqqGJbWycqqG0baDcqqGPbqAcqqGVbWBcqqGUbGBcqqG6aGocqqGGaaicqqGYaGmcqqGbbqqcqGHsgIRcqqGdbWqcqqGSaalcqqGGaaicqqHtoWrdaqhaaqaaGqaaiab=LgaPbqaaiab=jhaYbaacqGH9aqpdaWcaaqaaiab=TgaRbqaaiabikdaYaaaiiaacqGFdpWCdaWgaaqaaiab=LgaPbqabaGae43Wdm3aaSbaaeaacqWFQbGAaeqaaiabc6caUiaaxMaacaWLjaWaaeWaaeaacqaI1aqnaiaawIcacaGLPaaaaaa@6145@
Here the reacting species (A and B or A and A) occupy adjacent sites i and j, and the units of k are (molecules/site)-1sec-1. The factor of four in k/4 in Eq. (4) accounts for the fact that all four neighboring sites of site i were randomly chosen to search for the existence of species B. Similarly, the factor of 2 in Eq. 5 was due to the degeneracy of species participating in the homodimerization.
Event selection and time advancement
After an occupied site, say i, was selected, the transition probabilities per unit time of all possible events were computed. The probability for a certain event 'x' at site i was calculated as
pix=ΓixΓmax, (6)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGWbaCdaqhaaqaaiabdMgaPbqaaiabdIha4baacqGH9aqpdaWcaaqaaiabfo5ahnaaDaaabaGaemyAaKgabaGaemiEaGhaaaqaaiabfo5ahnaaBaaabaGagiyBa0MaeiyyaeMaeiiEaGhabeaaaaGaeiilaWIaaCzcaiaaxMaadaqadaqaaiabiAda2aGaayjkaiaawMcaaaaa@40D7@
where Γmax is a normalization constant to ensure that the selection probability of each event is always less than or equal to 1 and Γix
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqqHtoWrdaqhaaqaaiabdMgaPbqaaiabdIha4baaaaa@310C@ is the transition probability of event x (reaction or diffusion) at site i, as defined above. The maximum values of transition probabilities per unit time of events described by Eqs. (2)-(4) are Γd/4, k, k/4, and k/2, respectively. We defined the normalization constant as
Γmax=4(Γd4+max{∑all forward reaction eventsΓr}),+max{∑all backward reaction eventsΓr}
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@9C41@
where multiplication by a factor of 4 accounts for the microscopic process in each of the four directions on the square lattice. For the reaction events given in Fig. 2
Γmax=Γd+4(k1f2+k2f4+k3f2)+∑i=46kif+∑i=16kib, (7)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@6577@
where the subscripts f and b refer to the forward and backward reaction events, respectively, and i denotes the corresponding reaction according to Fig. 2. A random number 'r' was finally chosen from a uniform distribution between 0 and 1. The events were randomly ranked. The smallest value of m satisfying ∑x=1mpix>r
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaadaaeWbqaaiabdchaWnaaDaaabaGaemyAaKgabaGaemiEaGhaaaqaaiabdIha4jabg2da9iabigdaXaqaaiabd2gaTbGaeyyeIuoacqGH+aGpcqWGYbGCaaa@3A7E@ criterion was chosen. If no value of m satisfied the above criterion, then no event was selected to occur (a null-event) and a new occupied site was again randomly picked. Otherwise, the mth event was selected, the populations on the lattice were updated accordingly to reflect the stoichiometry of the reaction or diffusion process, and the real time was advanced based on the most frequently selected event as suggested in [44]. In our simulations, real time was advanced based on the diffusion of receptors according to which the average time step after each successful diffusion event was calculated as
Δt=1∑i=1No.ofsites(σi∑j∈BiΓi→jd(1-σj))=114Γd∑occupiedsites(∑j∈Bi(1-σj)). (8)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaaiiaacqWFuoarieaacqGF0baDcqGF9aqpdaWcaaqaaiab+fdaXaqaamaaqahabaWaaeWaaeaacqWFdpWCdaWgaaqaaiab+LgaPbqabaWaaabCaeaacqWFtoWrdaqhaaqaaiab+LgaPjabgkziUkab+PgaQbqaaiab+rgaKbaadaqadaqaaiab+fdaXiab+1caTiab=n8aZnaaBaaaleaacqGFQbGAaeqaaaGccaGLOaGaayzkaaaabaGae4NAaOMae8hcI4Sae4Nqai0aaSbaaSqaaiab+LgaPbqabaaakeaaaiabggHiLdaacaGLOaGaayzkaaaabaGae4xAaKMae4xpa0Jae4xmaedabaGae4Nta4Kae43Ba8Mae4Nla4Iae4hiaaIae43Ba8Mae4NzayMae4hiaaIae43CamNae4xAaKMae4hDaqNae4xzauMae43CamhacqGHris5aaaacqGF9aqpdaWcaaqaaiab+fdaXaqaamaalaaabaGae4xmaedabaGae4hnaqdaaiab=n5ahnaaCaaabeqaaiab+rgaKbaadaaeWbqaamaabmaabaWaaabCaeaadaqadaqaaiab+fdaXiab+1caTiab=n8aZnaaBaaaleaacqGFQbGAaeqaaaGccaGLOaGaayzkaaaabaGae4NAaOMae8hcI4Sae4Nqai0aaSbaaSqaaiab+LgaPbqabaaakeaaaiabggHiLdaacaGLOaGaayzkaaaabaGae43Ba8Mae43yamMae43yamMae4xDauNae4hCaaNae4xAaKMae4xzauMae4hzaqMae4hiaaIae43CamNae4xAaKMae4hDaqNae4xzauMae43CamhabaaacqGHris5aaaacqGFGaaicqGFUaGlcaWLjaGaaCzcamaabmaabaGae4hoaGdacaGLOaGaayzkaaaaaa@9004@
The summation in the denominator was updated each time a successful event happens by subtracting the previous occupancy values affected by the selected event and adding the new ones. In this way, the summation was carried out only once (at the beginning of a simulation) over all occupied sites. Finally, a new site is picked randomly until the desired real time is reached.
The hybrid null-event MC algorithm explained above was implemented (for details on the null-event algorithm see [44]) using Fortran 90. For each data point, 10 simulations with different seeds of the random number generator were used to collect statistics.
Simulation size and model parameters
In this work, the cell surface was represented using a 2-dimensional square lattice with each pixel being 2 nm × 2 nm in size. The number density of receptors ranges from ~102 receptors per μm2 on normal cells [46] to ~103 receptors per μm2 on human epithelioid carcinoma cells (A-431 cells) which overexpress EGFR [2]. However, the local density of receptors can be much higher in-vivo because of the localization of receptors in certain regions of the plasma membrane, such as in lipid rafts [35,47,48]. We simulated a low density (to represent normal cells) of 31 receptors on 500 nm × 500 nm mesh and a high density (to represent A-431 cells) of 55 receptors on 100 nm × 100 nm mesh, which are equivalent to receptor number densities of 125 and 5500 per μm2, respectively. The diffusivity of monomer EGFR has been reported to be around 2 × 10-14 m2s-1 [49,50]. Lower macroscopic diffusivities for EGFR have also been observed [49], which may be due to containment within cytoskeletal elements [51] or lipid rafts [52]. Based on these suggestions, the effect of a slower diffusivity is also analyzed by considering a diffusivity of 2 × 10-15 m2s-1. Simulations have been performed in the 0–60 sec time interval to capture the initial transients.
The model parameters are summarized in Table 1. We assumed two types of EGF binding on the cell surface: low affinity binding (on monomer EGFR) and high affinity binding (on dimerized EGFR). Experimental information suggests the existence of a high affinity (for receptor-ligand association) receptor population, most of which, if not all, is present in the form of dimers; see review by [1]. A recent equilibrium study has also shown that this interpretation is consistent with the experimentally reported concave-up shape of the Scatchard plot [13].
Experimental studies have provided evidence of predimerized receptors on A-431 cells to different extents [9,53-56]. Consequently, a fraction (~82%) of receptors was initially placed at random locations as monomers and the remaining as dimers on simulated A-431 cells for comparison with single particle tracking data. Corresponding to the dimerization equilibrium constant for this data, we found that at the lower receptor number density of 125 per μm2, there is negligible number of dimers in the absence of ligand. The receptor dimerization constants vary with ligand occupancy. Several experimental studies have shown that dimerization between unbounded receptors occurs with lower affinity than that between one bounded and one unbounded receptor. Finally, dimerization between two ligand bounded receptors occurs with the highest affinity [6,57].
Appendix
A sensitivity analysis was performed in which each kinetic parameter (ki, i = 1f, 4f, 5f, 6f, 1b, 4b, 5b, 6b) was increased by 20%, and the change in the high intensity spots was observed at three different times (20, 40 and 60 sec) from the mean of 10 independent MC simulations. The normalized sensitivity coefficient, reported in Fig. 7, is defined as
Figure 7 Normalized sensitivity coefficients at 3 different times (20, 40 and 60 sec) calculated by introducing a 20% increase in the kinetic parameter indicated on the x-axis.
(I−Io)/Io(ki−kio)/kio=I−Io0.2Io, (A1)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@5543@
where I is the % of high intensity spots (y axis in Fig. 4a) upon perturbing the kinetic parameter, ki, and Io is the nominal value corresponding to the original set of kinetic parameters, kio. Only 8 of the 12 kinetic parameters were independently perturbed because of the two equilibrium constrains reported in [14], i.e.,
K2=K1K5K4, and (A2)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGlbWsdaWgaaWcbaGaeGOmaidabeaakiabg2da9maalaaabaGaem4saS0aaSbaaSqaaiabigdaXaqabaGccqWGlbWsdaWgaaWcbaGaeGynaudabeaaaOqaaiabdUealnaaBaaaleaacqaI0aanaeqaaaaakiabcYcaSiabbccaGiabbggaHjabb6gaUjabbsgaKjabbccaGiaaxMaacaWLjaWaaeWaaeaacqqGbbqqcqqGYaGmaiaawIcacaGLPaaaaaa@4213@
K3=K1K5K6(K4)2. (A3)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGlbWsdaWgaaWcbaGaeG4mamdabeaakiabg2da9maalaaabaGaem4saS0aaSbaaSqaaiabigdaXaqabaGccqWGlbWsdaWgaaWcbaGaeGynaudabeaakiabdUealnaaBaaaleaacqaI2aGnaeqaaaGcbaWaaeWaaeaacqWGlbWsdaWgaaWcbaGaeGinaqdabeaaaOGaayjkaiaawMcaamaaCaaaleqabaGaeGOmaidaaaaakiabc6caUiaaxMaacaWLjaWaaeWaaeaacqqGbbqqcqqGZaWmaiaawIcacaGLPaaaaaa@4193@
The equilibrium relations determine the changes in the kinetic parameters of the dependent reactions 2 and 3 (see Fig. 2) upon perturbing those of the linearly independent reactions. A change in an equilibrium constant can be associated with a change in the forward, backward, or both rate constants. For simplicity a change in the rate constant of a forward (backward) linearly independent reaction is taken to cause a change in the forward (backward) rate constant of the linearly dependent reactions. Specifically, one has
k2f=(k2f)ofk1ffk5ffk4f, (A4)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGRbWAdaWgaaWcbaGaeGOmaiJaemOzaygabeaakiabg2da9maabmaabaGaem4AaS2aaSbaaSqaaiabikdaYiabdAgaMbqabaaakiaawIcacaGLPaaadaWgaaWcbaGaem4Ba8gabeaakmaalaaabaGaemOzay2aaSbaaSqaaiabdUgaRnaaBaaameaacqaIXaqmcqWGMbGzaeqaaaWcbeaakiabdAgaMnaaBaaaleaacqWGRbWAdaWgaaadbaGaeGynauJaemOzaygabeaaaSqabaaakeaacqWGMbGzdaWgaaWcbaGaem4AaS2aaSbaaWqaaiabisda0iabdAgaMbqabaaaleqaaaaakiabcYcaSiaaxMaacaWLjaWaaeWaaeaacqqGbbqqcqqG0aanaiaawIcacaGLPaaaaaa@4E8B@
k3f=(k3f)ofk1ffk5ffk6f(fk4f)2, (A5)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@56B4@
k2b=(k2b)ofk1bfk5bfk4b, and (A6)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@53F0@
k3b=(k3b)ofk1bfk5bfk6b(fk4b)2, (A7)
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@5688@
where subscript 'o' denotes the nominal value, and fki
MathType@MTEF@5@5@+=feaafiart1ev1aqatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGMbGzdaWgaaWcbaGaem4AaS2aaSbaaWqaaiabdMgaPbqabaaaleqaaaaa@3122@ is 1, if the kinetic parameter (ki) is not perturbed, and 1.2, if the parameter is increased by 20%.
Authors' contributions
KM carried out the simulations and drafted the manuscript. DGV and JSE edited the manuscript. All authors participated in the analysis of the data. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by grants from the US Department of Energy (DE-FG02-05ER25702) and the National Science Foundation (CTS-0312117). KM thanks Abhijit Chatterjee for useful discussions.
==== Refs
Jorissen RN Walker F Pouliot N Garrett TPJ Ward CW Burgess AW Epidermal growth factor receptor: mechanisms of activation and signalling Experimental Cell Research 2003 284 31 53 12648464 10.1016/S0014-4827(02)00098-8
Wiley HS Anomalous binding of epidermal growth factor to A431 cells is due to the effect of high receptor densities and a saturable endocytic system J Cell Biol 1988 107 801 810 3262110 10.1083/jcb.107.2.801
Wiley HS Shvartsman SY Lauffenburger DA Computational modeling of the EGF-receptor system: a paradigm for systems biology Trends in Cell Biology 2003 13 43 50 12480339 10.1016/S0962-8924(02)00009-0
Schlessinger J Allosteric regulation of the epidermal growth factor receptor kinase J Cell Biol 1986 103 2067 2072 3023396 10.1083/jcb.103.6.2067
Sako Y Minoghchi S Yanagida T Single-molecule imaging of EGFR signalling on the surface of living cells Nature Cell Biology 2000 2 168 172 10707088 10.1038/35004044
Lemmon MA Bu Z Ladbury JE Zhou M Pinchasi D Lax I Engelman DM Schlessinger J Two EGF molecules contribute additively to stabilization of the EGFR dimer The EMBO Journal 1997 16 281 294 9029149 10.1093/emboj/16.2.281
Klein P Mattoon D Lemmon MA Schlessinger J A structure-based model for ligand binding and dimerization of EGF receptors PNAS 2004 101 929 934 14732694 10.1073/pnas.0307285101
Weijer CJ Visualizing signals moving in cells Science 2003 300 96 100 12677060 10.1126/science.1082830
Vanbelzen N Rijken PJ Hage WJ Delaat SW Verkleij AJ Boonstra J Direct visualization and quantitative-analysis of epidermal growth factor-induced receptor clustering Journal Of Cellular Physiology 1988 134 413 420 3258312 10.1002/jcp.1041340312
Haigler HT McKanna JA Cohen S Direct visualization of the binding and internalization of a ferritin conjugate of epidermal growth factor in human carcinoma cells A-431 J Cell Biol 1979 81 382 395 313931 10.1083/jcb.81.2.382
Verveer PJ Wouters FS Reynolds AR Bastiaens PIH Quantitative Imaging of Lateral ErbB1 Receptor Signal Propagation in the Plasma Membrane Science 2000 290 1567 1570 11090353 10.1126/science.290.5496.1567
Lidke DS Nagy P Heintzmann R Arndt-Jovin DJ Post JN Grecco HE Jares-Erijman EA Jovin TM Quantum dot ligands provide new insights into erbB/HER receptor–mediated signal transduction Nature Biotechnology 2004 22 198 1203 14704683 10.1038/nbt929
Mayawala K Vlachos DG Edwards JS Heterogeneities in EGF receptor density at the cell surface can lead to concave up scatchard plot of EGF binding FEBS Letters 2005 579 3043 3047 15896781 10.1016/j.febslet.2005.04.059
Wofsy C Goldstein B Lund K Wiley HS Implications of epidermal growth factor (EGF) induced EGF receptor aggregation Biophys J 1992 63 98 110 1420877
Chamberlin SG Davies DE A unified model of c-erbB receptor homo- and heterodimerisation Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology 1998 1384 223 232 10.1016/S0167-4838(97)00203-3
Lauffenburger DA Linderman JJ Receptors models for binding, trafficking, and signaling 1993 New York, Oxford University Press
Haugh JM A unified model for signal transduction reactions in cellular membranes Biophysical Journal 2002 82 591 604 11806904
Kusumi A Nakada C Ritchie K Murase K Suzuki K Murakoshi H Kasai RS Kondo J Fujiwara T Paradigm Shift of the Plasma Membrane Concept from the Two-Dimensional Continuum Fluid to the Partitioned Fluid: High-Speed Single-Molecule Tracking of Membrane Molecules Annual Review of Biophysics and Biomolecular Structure 2005 34 351 378 10.1146/annurev.biophys.34.040204.144637
Ziff RM Gulari E Barshad Y Kinetic phase transitions in an irreversible surface-reaction model Physical Review Letters 1986 56 2553 2556 10033028 10.1103/PhysRevLett.56.2553
Slepchenko BM Schaff JC Carson JH Loew LM Computational Cell Biology: Spatiotemporal Simulation of Cellular Events Annual Review of Biophysics and Biomolecular Structure 2002 31 423 441 10.1146/annurev.biophys.31.101101.140930
Takahashi K Arjunan SNV Tomita M Space in systems biology of signaling pathways – towards intracellular molecular crowding in silico FEBS Letters 2005 579 1783 1788 15763552 10.1016/j.febslet.2005.01.072
Lemerle C Ventura BD Serrano L Space as the final frontier in stochastic simulations of biological systems FEBS Letters 2005 579 1789 1794 15763553 10.1016/j.febslet.2005.02.009
Saxton MJ Single-particle tracking: effects of corrals Biophys J 1995 69 389 398 8527652
Andrews SS Bray D Stochastic simulation of chemical reactions with spatial resolution and single molecule detail Physical Biology 2004 1 137 151 16204833 10.1088/1478-3967/1/3/001
Goldman J Andrews S Bray D Size and composition of membrane protein clusters predicted by Monte Carlo analysis European Biophysics Journal 2004 33 506 512 14997356 10.1007/s00249-004-0391-6
Levin MD Shimizu TS Bray D Binding and Diffusion of CheR Molecules Within a Cluster of Membrane Receptors Biophys J 2002 82 1809 1817 11916840
Mahama PA Linderman JJ A Monte Carlo study of the dynamics of G-protein activation Biophys J 1994 67 1345 1357 7811949
Shea LD Omann GM Linderman JJ Calculation of diffusion-limited kinetics for the reactions in collision coupling and receptor cross-linking Biophys J 1997 73 2949 2959 9414209
Shimizu TS Aksenov SV Bray D A Spatially Extended Stochastic Model of the Bacterial Chemotaxis Signalling Pathway Journal of Molecular Biology 2003 329 291 309 12758077 10.1016/S0022-2836(03)00437-6
Stiles JR Jr TMB Salpeter EE Salpeter MM Bower JM Monte Carlo simulation of neurotransmitter release using MCell, a general simulator of cellular physiological processes Computational Neuroscience 1998 New York, Plenum 279 284
Woolf PJ Linderman JJ Untangling Ligand Induced Activation and Desensitization of G-Protein-Coupled Receptors Biophys J 2003 84 3 13 12524261
Woolf PJ Linderman JJ An algebra of dimerization and its implications for G-protein coupled receptor signaling Journal of Theoretical Biology 2004 229 157 168 15207471 10.1016/j.jtbi.2004.03.012
Zhong H Wade SM Woolf PJ Linderman JJ Traynor JR Neubig RR A Spatial Focusing Model for G Protein Signals. Regulator Of G Protein Signaling (Rgs) Protein-Mediated Kinetic Scaffolding J Biol Chem 2003 278 7278 7284 12446706 10.1074/jbc.M208819200
Bellot F Moolenaar W Kris R Mirakhur B Verlaan I Ullrich A Schlessinger J Felder S High-affinity epidermal growth factor binding is specifically reduced by a monoclonal antibody, and appears necessary for early responses J Cell Biol 1990 110 491 502 2298813 10.1083/jcb.110.2.491
Laude AJ Prior IA Plasma membrane microdomains: organization, function and trafficking (Review) Molecular Membrane Biology 2004 21 193 205 15204627 10.1080/09687680410001700517
Parton RG Hancock JF Lipid rafts and plasma membrane microorganization: insights from Ras Trends in Cell Biology 2004 14 141 147 15003623 10.1016/j.tcb.2004.02.001
Kusumi A Sako Y Cell surface organization by the membrane skeleton Current Opinion in Cell Biology 1996 8 566 574 8791449 10.1016/S0955-0674(96)80036-6
Vereb G Szollosi J Matko J Nagy P Farkas T Vigh L Matyus L Waldmann TA Damjanovich S Dynamic, yet structured: The cell membrane three decades after the Singer-Nicolson model PNAS 2003 100 8053 8058 12832616 10.1073/pnas.1332550100
Ritchie K Shan XY Kondo J Iwasawa K Fujiwara T Kusumi A Detection of Non-Brownian Diffusion in the Cell Membrane in Single Molecule Tracking Biophys J 2005 88 2266 2277 15613635 10.1529/biophysj.104.054106
Roy CL Wrana JL Clathrin- and non-clathrin-mediated endocytic regulation of cell signalling Nature Reviews Molecular Cell Biology 2005 6 112 126 15687999 10.1038/nrm1571
Phair RD Misteli T Kinetic Modeling Approaches to In Vivo Imaging Nature Reviews Molecular Cell Biology 2001 2 898 907 11733769 10.1038/35103000
Roskoski JR The ErbB/HER receptor protein-tyrosine kinases and cancer Biochemical and Biophysical Research Communications 2004 319 1 11 15158434 10.1016/j.bbrc.2004.04.150
Gillespie DT Exact stochastic simulation of coupled chemical reactions Journal of Chemical Physics 1977 81 2340 2361 10.1021/j100540a008
Reese JS Raimondeau S Vlachos DG Monte Carlo algorithms for complex surface reaction mechanisms: Efficiency and accuracy Journal of Computational Physics 2001 173 302 321 10.1006/jcph.2001.6877
Jakobsson E Computer simulation studies of biological membranes: progress, promise and pitfalls Trends in Biochemical Sciences 1997 22 339 344 9301334 10.1016/S0968-0004(97)01096-7
Benveniste M Livneh E Schlessinger J Kam Z Overexpression of epidermal growth factor receptor in NIH-3T3- transfected cells slows its lateral diffusion and rate of endocytosis J Cell Biol 1988 106 1903 1909 3133376 10.1083/jcb.106.6.1903
Simons K Toomre D Lipid rafts and signal transduction Nature Reviews Molecular Cell Biology 2000 1 31 39 11413487 10.1038/35036052
Pike LJ Lipid rafts: bringing order to chaos J Lipid Res 2003 44 655 667 12562849 10.1194/jlr.R200021-JLR200
Kusumi A Sako Y Yamamoto M Confined lateral diffusion of membrane receptors as studied by single particle tracking (nanovid microscopy). Effects of calcium-induced differentiation in cultured epithelial cells Biophys J 1993 65 2021 2040 8298032
Rees AR Gregoriou M Johnson P Garland PB High affinity epidermal growth factor receptors on the surface of A431 cells have restricted lateral diffusion The EMBO Journal 1984 3 1843 1847 6090128
Wiegant FA Blok FJ Defize LH Linnemans WA Verkley AJ Boonstra J Epidermal growth factor receptors associated to cytoskeletal elements of epidermoid carcinoma (A431) cells J Cell Biol 1986 103 87 94 3013901 10.1083/jcb.103.1.87
Pralle A Keller P Florin EL Simons K Horber JKH Sphingolipid-Cholesterol Rafts Diffuse as Small Entities in the Plasma Membrane of Mammalian Cells J Cell Biol 2000 148 997 1008 10704449 10.1083/jcb.148.5.997
Martin-Fernandez M Clarke DT Tobin MJ Jones SV Jones GR Preformed Oligomeric Epidermal Growth Factor Receptors Undergo an Ectodomain Structure Change during Signaling Biophys J 2002 82 2415 2427 11964230
Yu X Sharma KD Takahashi T Iwamoto R Mekada E Ligand-independent Dimer Formation of Epidermal Growth Factor Receptor (EGFR) Is a Step Separable from Ligand-induced EGFR Signaling Mol Biol Cell 2002 13 2547 2557 12134089 10.1091/mbc.01-08-0411
Moriki T Maruyama H Maruyama IN Activation of preformed EGF receptor dimers by ligand-induced rotation of the transmembrane domain1 Journal of Molecular Biology 2001 311 1011 1026 11531336 10.1006/jmbi.2001.4923
Gadella TWJ Jovin TM Oligomerization of epidermal growth factor receptors on A431 cells studied by time-resolved fluorescence imaging microscopy. A stereochemical model for tyrosine kinase receptor activation J Cell Biol 1995 129 1543 1558 7790353 10.1083/jcb.129.6.1543
Sherrill JM Kyte J Activation of epidermal growth factor receptor by epidermal growth factor Biochemistry 1996 35 5705 5718 8639530 10.1021/bi9602268
Schlessinger J The epidermal growth factor receptor as a multifunctional allosteric protein Biochemistry 1988 27 3119 3123 3291943 10.1021/bi00409a002
Kawamoto T Sato JD Le A Polikoff J Sato GH Mendelsohn J Growth stimulation of A431 cells by epidermal growth factor: identification of high-affinity receptors for epidermal growth factor by an anti-receptor monoclonal antibody PNAS 1983 80 1337 1341 6298788
Lin CR Chen WS Lazar CS Carpenter CD Gill GN Evans RM Rosenfeld MG Protein kinase C phosphorylation at Thr 654 of the unoccupied EGF receptor and EGF binding regulate functional receptor loss by independent mechanisms Cell 1986 44 839 848 3006920 10.1016/0092-8674(86)90006-1
Hendriks BS Opresko LK Wiley HS Lauffenburger D Quantitative Analysis of HER2-mediated Effects on HER2 and Epidermal Growth Factor Receptor Endocytosis: Distribution Of Homo- And Heterodimers Depends On Relative Her2 Levels J Biol Chem 2003 278 23343 23351 12686539 10.1074/jbc.M300477200
|
16318627
|
PMC1322222
|
CC BY
|
2021-01-04 16:30:02
|
no
|
BMC Urol. 2005 Nov 30; 5:16
|
latin-1
|
BMC Urol
| 2,005 |
10.1186/1471-2490-5-16
|
oa_comm
|
==== Front
Behav Brain FunctBehavioral and brain functions : BBF1744-9081BioMed Central London 1744-9081-1-211632116310.1186/1744-9081-1-21ResearchAuditory processing in individuals with auditory neuropathy Kumar Ajith U [email protected] M [email protected] Junior Research Fellow, Department of Audiology, All India Institute of Speech and Hearing, Manasagangothri, Mysore, Karnataka, 570006, India2 Director, All India Institute of Speech and Hearing, Manasagangothri, Mysore, Karnataka, 570006, India2005 1 12 2005 1 21 21 6 6 2005 1 12 2005 Copyright © 2005 Kumar and Jayaram; licensee BioMed Central Ltd.2005Kumar and Jayaram; 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
Auditory neuropathy is a disorder characterized by no or severely impaired auditory brainstem responses in presence of normal otoacoustic emissions and/or cochlear microphonics. Speech perception abilities in these individuals are disproportionate to their hearing sensitivity and reported to be dependent on cortical evoked potentials and temporal processing abilities. The disproportionate loss of auditory percept in presence of normal cochlear function is suggestive of impairment of auditory neural synchrony.
Methods
We studied the auditory evoked potentials and psychophysical abilities in 14 adults with auditory neuropathy to characterize their perceptual capabilities. Psychophysical tests included measurement of open set speech identification scores, just noticeable difference for transition duration of syllable /da/ and temporal modulation transfer function. Auditory evoked potentials measures were, recording of P1/N1, P2/N2 complex and mismatch negativity (MMN).
Results
Results revealed a significant correlation between temporal processing deficits and speech perception abilities. In majority of individuals with auditory neuropathy P1/N1, P2/N2 complex and mismatch negativity could be elicited with normal amplitude and latency. None of the measured evoked potential parameters correlated with the speech perception scores. Many of the subjects with auditory neuropathy showed normal MMN even though they could not discriminate the stimulus contrast behaviorally.
Conclusion
Conclusions drawn from the study are
1. Individuals with auditory neuropathy have severely affected temporal processing.
2. The presence of MMN may not be directly linked to presence of behavioral discrimination and to speech perception capabilities at least in adults with auditory neuropathy.
Auditory neuropathyspeech perceptiontemporal processingLLRs and MMN
==== Body
Background
Auditory neuropathy (AN) is recently described hearing disorder characterized by abnormal auditory nerve functioning in presence of normal cochlear receptor hair cell activity [1]. The clinical findings that define auditory neuropathy are
a) Presence of outer hair cell integrity in evoked otoacoustic emission or cochlear microphonics.
b) Absence of synchronized neural activity at the level of 8th nerve and brainstem.
Though the audiometric and electrophysiological findings are consistent with the 'retro outer hair cell dysfunction' exact site(s) of the pathology is yet to be determined. Some possible sites of lesion that could produce the audiometric and electrophysiological profile of AN include: inner hair cells, synaptic junction between inner hair cell and type I afferent nerve fibers, spiral ganglion cells, specific damage or demyelinization of type I auditory nerve fibers [1-3]. Therefore, AN consists of many varieties depending on the sites of lesion [4]. Speech perception ability in these patients also varies considerably. Some patients perform at the levels expected for patients with comparable degrees of sensory hearing loss and others show speech understanding which is disproportionate to their degree of hearing loss [5,6].
Speech perception abilities in these patients appear to depend on the extent of suprathreshold temporal distortions of cues rather than access to speech spectrum, unlike the patients with sensory hearing loss [7,6]. Zeng et al [8] reported the abnormal results on two measures of temporal perception in their group of children with AN: (i) gap detection threshold (identification of silence embedded in within the bursts of noise) and (ii) temporal modulation transfer function (measure of sensitivity to slow and fast amplitude fluctuation). They also found a correlation between temporal modulation transfer function (TMTF) and speech perception abilities in their patients. Rance et al [6] also reported poor performance on the task involving timing cues (TMTF, temporal aspects of frequency discrimination) in a group of 14 children with AN. These temporal processing abnormalities had significant correlation with speech perception abilities. They attributed the speech perception scores that are disproportionate to pure tone hearing loss to these suprathreshold temporal processing deficits.
Another factor that is reported to be related to speech perception abilities in these individuals is cortical evoked event related potentials. Rance et al [5] reported that a subgroup of children with AN, who had recordable cortical evoked potential performed well on open set speech perception task and derived significant benefit from amplification. In contrast, subjects who had no recordable cortical evoked potential performed poorly on the same tasks. From this observation they concluded that presence of cortical auditory evoked potential reflects some amount of preserved synchrony in central auditory system which contributes to better speech understanding despite the distortion that occurs at 8th nerve and auditory brainstem in these individuals.
Speech perception process can be investigated in neurophysiological as well as psychophysical perspective. An important aspect of this study is use of a combined neurophysiological and psychophysical approach. With this multidisciplinary technique we hope to gain insight into both stimulus representation and processing in individuals with AN. This study is sought to explore the relation between their psychoacoustic abilities and evoked potential parameters, in a group of adults with auditory neuropathy. Psychophysical experiments included were measurement of open set speech identification scores, just noticeable difference (JND) for transition duration of the syllable /da/ and temporal modulation transfer function. Auditory evoked potentials measures included recording of N1/P1, N2/P2 and Mismatch negativity (MMN) potentials.
Methods
Study was carried out in two phases, first phase involved psychophysical experiments and auditory evoked potentials were measured in the second phase.
Subjects
Two groups of subjects participated in the study. The first group consisted of 14 individuals with AN (16 to 30 years with the mean age of 23 years) and second group consisted of age and gender matched 30 normally hearing subjects. All AN subjects were recruited from Department of Audiology, All India Institute of Speech and Hearing, Mysore. No subject complained about any middle ear disease (assessed using otoscopy, tympanometry and clinical history), noise exposure or ototoxic drug usage. Results of different audiological measurements of AN subjects are shown in Table 1. As all the subjects had symmetrical hearing loss, (symmetrical hearing loss was operationally defined as the difference in thresholds between two ears at corresponding frequencies within 15 dB), pure tone thresholds were measured again with loudspeakers and these measurements were considered for all future purpose. Furthermore, subjects in the normally hearing group had their hearing thresholds within 15 dB HL at octave frequencies between 250 Hz to 8 kHz and normal results on immittance evaluation. All the subjects were native speakers of Kannada, a South Indian Dravidian language.
Table 1 Audiometric and electrophysiological details of auditor neuropathy subjects.
SN Age/sex PTA (.5, 1 and 2 KHz) Speech identification scores OAE ABR Acoustic reflex Efferent suppression Configuration
1 16/F 45.00 45.00 Present Absent Absent 0.2 Raising
2 16/M 13.00 84.00 Present absent Absent 0.0 Peaked
3 30/M 23.00 80.00 Present absent Absent 0.4 Peaked
4 24/F 25.00 38.00 Present absent Absent 0.0 Peaked
5 16/M 40.00 .00 Present absent Absent 0.1 Raising
6 26/M 45.00 4.00 Present absent Absent 0.0 Raising
7 23/F 45.00 5.00 Present absent Absent 0.0 Raising
8 23/M 75.00 .00 Present absent Absent 0.1 Flat
9 27/M 20.00 50.00 Present absent Absent 0.0 Peaked
10 23/M 40.00 86.00 Present absent Absent 0.3 Peaked
11 24/M 23.00 8.00 Present absent Absent 0.0 Peaked
12 25/F 45.00 .00 Present absent Absent 0.0 Raising
13 28/M 5.00 90.00 Present absent Absent 0.3 Peaked
14 25/F 10.00 95.00 Present absent Absent 0.2 Peaked
PTA = Pure tone average
OAE = Otoacoustic emissions
ABR = Auditory brainstem responses
Psychophysical tests
The experiment protocol consisted of speech identification score testing, measurement of JND for transition duration of /da/ and TMTF.
(a) Speech identification testing
Only AN subjects participated in this experiment. Vandana's speech identification test in Kannada was used to assess the open set speech perception abilities in the subjects. This test consists of 50 bisyllabic meaningful words in Kannada. Validity and reliability of this test on native speakers of Kannada have already been established by Vandana, [9]. Recorded material was presented at 'comfortable level' which ranged between 30 to 40 dB SL ref: Average thresholds at 500 Hz, 1 kHz, and 2 kHz, using MA-53 clinical audiometer through a loudspeaker kept at 1 m distance and 0° azimuth. Output of the loudspeaker was calibrated using Quest 1800 sound level meter and Quest 4180 free field microphone. A calibration tone recorded before the test material was used to adjust the Vu meter deflection to zero. The test was carried out in a quiet listening condition and each stimulus was presented in isolation without being embedded in a carrier phrase. The subjects were required to repeat each stimulus and a percentage of correct identification was determined. All the subjects were screened for misarticulations using Kannada Articulation Test [10]
(b) JND measurements
Both AN and normal listeners participated in this experiment. Stimulus was derived from retroflex /da/ uttered in isolation, by a 25 year old male native speaker of Kannada. The spoken was digitally recorded on a data acquisition system at 44 kHz sampling frequency. The transition duration was identified using both spectral and wave form view of the stimulus. Transition duration was lengthened up to 'original transition duration +100 ms' in 10 ms steps by means of Pitch Synchronized Overlap and Add (PSOLA) technique. PSOLA performs the lengthening of the stimulus in time domain and preserves most of physical characteristics of the stimulus such as spectral shape, amplitude distribution, and periodicity [11].
Subjects were tested individually in a sound attenuated room. Signals were played via a PC, at a sampling frequency of 44 kHz and were subsequently fed to a MA-53 audiometer. Subjects received the signals through audiometer's loudspeaker kept at a distance of 1 m and 0° azimuth. Presentation level of the stimulus was fixed at 30 dB SL ref: Average thresholds at 500 Hz, 1 kHz, and 2 kHz. Stimuli were presented at equal presentation level to compensate for the audibility in individuals with auditory neuropathy. JND was determined using an adaptive tracking technique (PEST) with AX same difference discrimination paradigm (in this A = anchor stimulus, X = Variable stimulus and subjects task is to indicate whether A is same as X or not). Inter stimulus interval between anchor and variable stimulus was 500 ms. Step size and the direction of variable stimulus were changed according to rules of PEST [12]. The subject's JND was determined by calculating the difference in transition duration between anchor and variable stimuli that is required to achieve a performance level of 69% correct responses. Test trials also included equal number of catch trials. Catch trial consisted of either two identical anchor or two identical non anchor stimuli.
(c) Temporal modulation transformer function
Both AN and normal listeners participated in this experiment. Modulation detection thresholds were measured by determining the sensitivity to sinusoidal amplitude modulation as a function of modulation frequency. Presentation level of the stimulus was kept at 30 dB SL ref: Average thresholds at 500 Hz, 1 kHz, and 2 kHz. Stimulus was presented through a loud speaker kept at a distance of 1 m and 0° azimuth. Stimuli were presented at equal sensation level to compensate for the audibility in patients with auditory neuropathy. A broad band noise was generated and controlled digitally to measure TMTF. Broad band noise had a duration of 500 ms and ramp of 2.5 ms. The modulated signal was derived by multiplying the 500 ms white noise by a dc shifted sine wave. The depth of the modulation was controlled by varying the amplitude of modulating sine wave. Modulation depth for the various stimuli varied between 0 to -30 dB and step size was 3 dB. Modulation detection thresholds were measured for 5 frequencies; 4 Hz, 16 Hz, 32 Hz, 64 Hz, 128 Hz, and 200 Hz,. Procedure was same as that described for the measurement of JNDs. In all the subjects at least at one modulation frequency the presentation level was changed and modulation detection threshold was rechecked to ensure that subjects are not using the loudness judgments.
Auditory evoked potential measurements
In this experiment both normal hearing subjects and individuals with AN participated. The cortical evoked potentials were obtained in one session lasting less than 15 min. The subjects were seated in a comfortable position to ensure relaxed posture to minimize muscular artifacts. They were instructed not to pay attention to the stimuli. A silent cartoon movie was played to minimize the possibility of active attention. The stimuli was unmodified /da/ and synthesized /da/ in which transition duration was lengthened by 100 ms. This was decided on the basis of a pilot study measuring the behavioral JND in AN subjects. Synthesis technique was same as the one used for psychoacoustic testing. These two contrasts were presented in an odd ball paradigm. Stimuli were presented at 'comfortable level' to both ears (usually 30 to 40 dB SL, Ref: Average thresholds at 500 Hz, 1 kHz, and 2 kHz) through EAR-3A insert receiver. IHS smart EP module was used to control the stimulus presentation and acquisition of evoked potential. Conventional recording techniques were used. After skin preparation at electrode site, silver-chloride disc electrodes were placed at Cz, with ipsilateral mastoid as reference, using conductive electrode paste and adhesive tape. Ground electrode was placed at Fz Data was acquired after ensuring that the impedance at all electrode sites was within permissible limits. The protocol used for recording is shown in Table 2.
Table 2 Protocol used for evoked potential testing
Stimulus Standard – unmodified /da/
Deviant – Synthesized /da/
Intensity 30 to 40 dB SL
Probability 5:1
Repetition rate 1.1/s
Analysis time 500 ms
Gain 75000
Band Pass Filter 1 to 30 Hz
Transducer EAR-3A insert ear phones
In order to probe the representation of these two stimulus contrasts at pre-attentive neural level MMN responses were derived from recorded cortical evoked potentials. MMN is a passively elicited cortical evoked potential that is known to reflect the brain's response to an acoustic change [13]. The MMN is seen as a negative deflection around 200 ms after stimulus presentation. MMN was identified in the difference wave between frequent and infrequent recordings. Grand average waveform was also constructed by utilizing the individual waveform which had MMN.
Results
Psychophysical tests
(a) Open set speech identification test
Open set speech identification scores in individuals with AN varied considerably. The mean speech identification score was 41.7% (SD: 38.8%), but scores ranged from 0% to 95%. Speech identification scores correlated with low frequency (250 Hz, 500 Hz, 1000 Hz) hearing thresholds (r = 0.67, p = 0.001) but not with the high frequency hearing thresholds (2000 Hz, 4000 Hz and 8000 Hz r = 0.3, p = .234).
(b) JND measurements
Figure 1 shows the mean and SD values of JND in transition duration for stimuli /da/. Independent sample 't' test showed a significant difference between two groups at .001 level. Of 14 subjects 10 could not differentiate the stimuli that differed in transition duration by as much as 100 ms. Four subjects whose JND were less than 100 ms also had their open set speech identification scores more than 80%.
Figure 1 Mean and SD (error bars show 1 SD) of JND in transition duration for the auditory neuropathy (AN) group and normally hearing subjects.
Temporal modulation transfer function
Figure 2 shows the TMTF for subjects with normal hearing and auditory neuropathy. Normal hearing listeners were most sensitive to slow temporal fluctuation and became less sensitive as the fluctuation rate was increased. Similar trend was noticed in individuals with AN. Average peak sensitivity of normal hearing listeners was -17.36 dB. In contrast, average peak sensitivity for auditory neuropathy group was -6.6 dB (SD: 5.4 dB). At higher modulation frequencies many of the AN (12 subjects) subjects did not even detect a modulation of depth of 0 dB (100%). Peak sensitivity of AN group tended to fall in two distinct categories. Eight individuals had peak sensitivity of more than -10.4 dB and 7 of these patients had open set speech identification scores more than 50%. Six subjects had peak sensitivity less than -5.6 dB and 5 of them had speech identification scores of less than 20%. One subjects in each category had paradoxical results on speech perception and TMTF results. When data from individual subjects were examined speech identification scores and temporal modulation transfer function in these two subjects were in extreme. Hence these two subjects were treated as outliers and when data from these two subjects were excluded a significant correlation was observed between peak sensitivity and speech identification scores. No relation could be established between JND measurements and TMTF.
Figure 2 TMTF for the auditory neuropathy (AN) group and normally hearing subjects. AN20 = TMTF for auditory neuropathy subjects with speech identification scores less than 20%. AN50+ = TMTF for auditory neuropathy subjects with speech identification scores more than 50% Normal = TMTF for normally hearing subjects.
Auditory evoked potential measurements
Before doing the analysis all the wave forms were corrected for baseline EEG activity by subtracting the pre-stimulus electrical activity (for 50 ms before the presentation of stimulus). Table 3 shows, the latencies and amplitudes of peaks P1, N1, P2 and N2 for AN and normal hearing group. P2/N2 complex was present is all 14 individuals whereas P1/N1 complex was not present in 4 subjects. Whenever LLRs were present, latency and amplitudes were within normal range. Presence or absence of LLR peaks did not bear any relation to the speech identification scores. Pearson's product moment correlation failed to evidence any significant correlation between evoked potential parameters and other psychophysical test results. Table 4 shows latency, amplitude and area of MMN parameters. Area of the MMN was determined by calculating the area between wave and baseline and took into account both the duration and amplitude of MMN response. In 5 of 14 subjects, MMN could not be elicited. Pearson's product moment correlation was performed between MMN parameters and other psychophysical measures. Only peak latency of MMN evidenced a significant correlation with speech identification scores. As the number of subjects with MMN present was less, to interpret the results of correlation, a scatter plot was drawn between MMN peak latency and speech identification scores. As seen from the scatter plot (Figure 3), no trend could be observed between MMN peak latency and speech identification scores. Figure 4 shows the grand average of MMN waveform in AN subjects and normal hearing listeners. Whenever the MMN was present in individuals with AN, wave form was indistinguishable from normal listeners.
Table 3 Mean and SD (values in parenthesis) of amplitude and latencies of LLR components in both groups
P1 N1 P2 N2
AN subjects Amplitude (in μV) 2.8 (0.9) 0.9 (0.8) 2.8 (2.08) -1.1 (2.3)
Latency (in ms) 81 (16.2) 125.4 (23.04) 154.1 (27.1) 205 (23)
Normal subjects Amplitude (in μV) 2.5 (0.6) -0.5 (0.5) 2.8 (1.5) -1.6 (1.5)
Latency (in ms) 69 (15.2) 120.5 (23.5) 145.3 (25.6) 200.2 (26.3)
Table 4 Mean and SD (values in parenthesis) of amplitude and latencies of mismatch nagativity components in the auditory neuropathy group
AN individuals Normal subjects
MMN (On set) MMN (Peak) MMN (Off set) MMN (On set) MMN (Peak) MMN (Off set)
Amplitude (in μV) -0.068 (0.6) -4.6 (2.1) 1.9 (3.2) -0.071 (0.2) -4.8 (1.5) 1.5 (2.5)
Latency (in ms) 117.3 (23.6) 186.4 (19.04) 209 (25.5) 120.5 (20.5) 180.6 (20.8) 204 (25.5)
Figure 3 Scatter plot between speech identification scores and peak latency of MMN in auditory neuropathy subjects.
Figure 4 Grand averaged MMN wave form in the auditory neuropathy group and normally hearing subjects.
Of the 9 subjects who had MMN, 5 of them could not behaviorally discriminate the two stimulus contrast (i.e. JND was more than 100 ms). Other 5 subjects who had no MMN also could not behaviorally discriminate the two-stimulus contrast. The MMN wave forms of those subjects who could behaviorally discriminate the stimulus contrast were virtually indistinguishable from those who did not behaviorally discriminate the contrast. This data indicate that presence of MMN does not necessarily indicate the presence of behavioral discrimination.
Discussion
The major findings of this research were:
i) Open set speech identification scores varied considerably in individuals with AN and speech identification scores had a significant correlation with the low frequency hearing sensitivity.
ii) All subjects with AN had severe temporal processing deficits as shown by JND measurements and TMTF.
iii) In majority of AN patients cortical evoked potentials could be recorded but none of the measured evoked potential parameters had any relation with psychophysical measurements.
Psychophysical measurements
Speech identification scores in AN individuals had good correlation with low frequency hearing sensitivity but not with the high frequency hearing sensitivity. This frequency specific correlation between hearing thresholds and speech identification scores, may be related to differential physiology between high frequency and low frequency coding. Low frequencies are usually coded by phase locked responses in type I auditory nerve fibers. Individuals with AN cannot use phase locking cues to the same extent as normal hearing listeners due to dyssynchronous firing of auditory nerve fibers. However, detection of the high frequency depends on place of excitation on basilar membrane and does not depend on the phase locking cues as much as low frequencies. We propose that, low frequency hearing sensitivity in these individuals may indicate the extent of temporal disruption in the auditory system. Its relation with speech identification scores is suggestive of importance of neural synchrony in understanding speech. This is also supported by other two observations:
i) A retrospective inspection of the data reveled, all 8 individuals who obtained speech identification scores more than 50% had their low frequency hearing sensitivity (average of 250 Hz, 500 Hz, and 1000 Hz) better than 25 dB HL and 6 individuals who had speech identification scores less than 20% had low frequency hearing sensitivity more than 40 dB HL.
ii) There was significant correlation between low frequency hearing sensitivity and peak modulation detection thresholds. Based on the above observations, we propose that low frequency hearing sensitivity in AN individuals may be the indicator of suprathreshold temporal processing deficits.
All AN individuals experienced severe difficulties in discriminating the speech stimuli that differed in time domain. As stimulus was presented at equal sensation levels to both the groups this resulted in difference in presentation levels (SPL) for each of the subjects. However, the difference in the JNDs for transition duration of syllable /da/ between two groups cannot be attributed to difference in presentation level (SPLs). It is shown that when the stimuli are sufficiently loud or at comfortable level auditory duration discrimination is independent of the intensity [14]. Individuals who had better discrimination abilities also possessed better open set speech identification scores. These findings stress the importance of perception of temporal variation in understanding speech information. Temporal processing deficits in individuals with AN are also demonstrated by poor performance on TMTF. Average peak sensitivity of individuals with AN was threefold more than the normals. Poor sensitivity to temporal modulations in these individuals is also reported by other investigators [6,8]. A significant correlation was observed between modulation detection thresholds and speech identification scores (when data from two subjects with paradoxical results were removed). This finding agrees with the results obtained from Rance et al [6], Zeng et al [7,8]
Difference between normal listeners and AN subjects in detection of modulation was more at higher modulation frequencies. The extent of temporal processing deficits were more than what is been reported for cochlear hearing loss of comparable degree [15]. This difference between two groups cannot be because of different presentation levels (SPA) used because modulation detection thresholds are reported to be stable over a wide range of intensities. In the auditory system, higher modulation frequencies are processed at auditory nerve and brainstem, whereas lower modulation frequencies are processed mainly in the thalamus and auditory cortex. As one ascends the auditory system, a neural encoding shift occurs. An emphasis on synchronous response for temporal coding exists at auditory nerve and brainstem (codes low frequencies) and less reliance on synchrony occurs as one move centrally (codes high frequencies) [16-18]. Hence, it can be expected that individuals with AN will have more problems in processing high rates of modulations which require synchronous firing of auditory nerve fibers. Inability of many the subjects to perceive amplitude modulation of 0 dB (100%) at higher modulation rates indicates the importance of temporal synchrony in auditory perception. Effects of reduced temporal fluctuations on speech perception in normal listeners have been reported previously [19]. Elevated modulation detection thresholds at slower modulation rates in combination with virtually no perception of modulations at high modulation rates are sufficient to disrupt the perception of amplitude envelop cues in normal speech. As this study measured only peak sensitivity, reduced peak sensitivity may also be due to reduced ability to perceive the amplitude changes in patients with auditory neuropathy.
Electrophysiological measures
P1/N1 and P2/N2 complex amplitude and latency did not appear to be related to degree of hearing loss or speech identification scores. This result is in contrast to Rance et al [5] who evidenced a strong relation between presence of event related potential and speech perception scores. This difference in the results may be due to difference in subjects and the stimuli. Rance et al [5] primarily studied children younger than 92 months and were fitted with the amplification devices before 28 months of age. This may have prevented the retrograde loss of speech perception abilities. In our subjects, average age at which amplification provided was 18 years. Many of the subjects were not identified in childhood as they had near normal hearing sensitivity and were grouped as slow learners in the class. This huge gap in the auditory experience between two groups might have adversely affected the speech perception abilities of the later. Presence of LLR components with normal latency and amplitude represent the stimulus registration in the primary auditory cortex, which do not involve complex decoding and representation of the signal as it is required for the speech perception.
Large numbers of studies in last decade have established MMN as an objective electrophysiological measure of auditory discrimination (e.g. [13]). Our results of MMN and behavioral discrimination are paradoxical. Significant MMN was seen in the majority of subjects with auditory neuropathy, even though stimulus contrast could not be behaviorally discriminated. Fried et al [20] have provided evidence for the existence of preconscious perception in the visual system. Preconscious perception describes the physiological or neurological process that occurs without behavioral or conscious perception. Some evidence of preconscious perception is also reported in auditory system using MMN. Allen et al. [16] reported the presence of MMN in normal listeners for the stimulus contrast that they could not behaviorally discriminate. Presence of MMN in AN subjects who could not behaviorally discriminate the stimulus contrast supports the hypothesis that neural generators responsible for the MMN are not necessarily linked to conscious perception [21]. But all the individuals who had no MMN could not behaviorally discriminate the stimulus contrast. These two results in combination support the notion that MMN is necessary, but not a sufficient component for conscious perception of stimulus change.
Another possible explanation for the discrepancy between behavioral discrimination and MMN in some AN subjects may be related to perception of stimulus onset cues. We hypothesis that cues in the stimulus onset play a major role in the behavioral discrimination between the stimulus contrasts that differ in transition duration. Kraus et al [20] reported that perception of any change in the stimulus onset was extremely difficult in a subject with AN who had normal hearing. Hence the individuals with AN had larger JND's. We propose that MMN, which was present in some AN individuals, was elicited by the difference in the later part of the stimulus. However, it is unclear that why AN individuals could not discriminate the stimulus contrasts by using the information in the later part of the stimulus that elicited the MMN.
Conclusion
Findings of this study indicate that individuals with AN have severely affected temporal processing abilities. These temporal processing deficits correlate significantly with the speech identification scores and hearing sensitivity in the low frequency region. Psychophysical measures including speech perception did not correlate with the electrophysiological measurements used at least in adults with AN.
Authors' contributions
AKU was involved in designing the study, data collection, analysis, interpretation and preparing the manuscript. JM was involved in designing the study, interpretation and preparing the manuscript.
==== Refs
Starr A Picton TW Sininger Y Hood L Berlin C Auditory neuropathy Brain 1996 119 741 753 8673487
Salvi RJ Wang J Ding D Stecker N Arnold S Auditory deprivation in the central auditory system resulting from selective inner hair cell loss: Animal model of auditory neuropathy Scand Audiol 1999 51 1 12 10803909
Starr A Michalewski HJ Zeng FG Brooks SF Linthicum F Kim CS Winnier D Keats B Pathology and physiology of auditory neuropathy with a novel mutation in the MPZ gene Brain 2003 126 1604 1619 12805115 10.1093/brain/awg156
Starr A Sininger YS Praat H The varieties of auditory neuropathy J Basic Clin Physiol Pharmacol 2000 11 215 229 11041385
Rance G Cone-Wesson B Wunderlich J Dowell R Speech perception and cortical event related potentials in children with auditory neuropathy Ear Hear 2002 23 239 253 12072616 10.1097/00003446-200206000-00008
Rance G McKay C Grayden D Perceptual characterization of children with auditory neuropathy Ear Hear 2004 25 34 46 14770016 10.1097/01.AUD.0000111259.59690.B8
Zeng FG Kong YY Michalewsk HJ Starr A Perceptual consequences of disrupted auditory nerve activity J Neurophysiol 2005 93 3050 3063 15615831 10.1152/jn.00985.2004
Zeng FG Oba S Garde S Sininger Y Starr A Temporal and speech processing deficits in auditory neuropathy Neuroreport 1999 10 3429 3435 10599857
Vandana Speech identification test in Kannada Independent project submitted to University of Mysore, Mysore 1998
Babu RM Ratna N Bettagiri R Test of articulation in Kannada The JAIISH 1972 3 7 19
Moulines E Laroche J Non parametric techniques for pitch scale and time scale modification of speech Speech Commun 1995 16 175 205 10.1016/0167-6393(94)00054-E
Taylor MM Creelman CD PEST: Efficient estimates on probability functions J Acoust Soc Am 1967 41 782 787 10.1121/1.1910407
Naatanen R The mismatch negativity: A power full tool for cognitive neuroscience Era Hear 1995 16 6 18 7774770
Moor BCJ Shailer MJ Schooneveldt GP Temporal modulation transfer function for band noise in subjects with cochlear hearing loss Br J Audiol 1992 26 229 237 1446186
Creelman DC Human discrimination of auditory duration J Acoust Soc Am 1962 34 582 593 10.1121/1.1918172
Wang X Sachs MB Neural encoding of the single formant stimuli in cat. I Responses of anteroventral cochlear neurons J Neurophysiol 1993 71 59 78 8158242
Allen J Kruas Bradlow Wang X Sachs MB Neural representation of consciously imperceptible speech sound differences Percept Psychophys 2000 62 1383 1393 11143450
Frisina RD Subcortical neural coding mechanisms for auditory temporal processing Hear Res 2001 158 1 27 11506933 10.1016/S0378-5955(01)00296-9
Drullman R Festen JM Plomp R Effects of temporal smearing on speech perception J Acoust Soc Am 1994 95 1053 1064 8132899 10.1121/1.408467
Fried I MacDonald KA Wilson CL Single neuron activity in human hippocampus and amygdala during the recognition of faces and objects Neuron 1997 18 753 765 9182800 10.1016/S0896-6273(00)80315-3
Kraus N Bradlow AR Cheatham MA Cunningham J King CD Koch DB Nicol TG McGee TJ Stein LK Wright BA Consequences of neural asynchrony: A case of auditory neuropathy J Assoc Res Otolaryngol 2000 1 33 45 11548236
Naatanen R Phoneme representations of the human brain as reflected by event-related potentials Electroencephalogr Clin Neurophysiol Suppl 1999 49 170 173 10533104
Vandana Speech identification test in Kannada Independent project submitted to University of Mysore, Mysore 1998
|
16321163
|
PMC1322223
|
CC BY
|
2021-01-04 16:39:20
|
no
|
Behav Brain Funct. 2005 Dec 1; 1:21
|
utf-8
|
Behav Brain Funct
| 2,005 |
10.1186/1744-9081-1-21
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1571633666610.1186/1471-2407-5-157Case ReportSynchronous adenocarcinoma and carcinoid tumor of the terminal ileum in a Crohn's disease patient Cioffi Ugo [email protected] Simone Matilde [email protected] Stefano [email protected] Michele M [email protected] Alessandro [email protected] Ettore Contessini [email protected] Department of Surgery, University of Milan, Fondazione Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena, IRCCS, Milano, Italy2 Department of Pathology, Azienda Ospedaliera San Paolo, Milano, Italy3 Istituto di Medicina Cardiovascolare, University of Milan, Fondazione Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena, IRCCS, Milano, Italy4 Department of Radiology, Fondazione Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena, IRCCS, Milano, Italy2005 8 12 2005 5 157 157 12 7 2005 8 12 2005 Copyright © 2005 Cioffi et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Several malignancies have been described in association with inflammatory bowel diseases, the most common being adenocarcinoma. Carcinoid tumor and Crohn disease has also been previously reported, however the coexistence of both neoplasms is quite rare and the clinical diagnosis is very difficult. Here we report what we believe to be the fourth case of a mixed adenocarcinoid tumor coexisting with Crohn's disease.
Case report
The patient presented with clinical and radiological features of intestinal obstruction. Laparotomy showed a stricturing lesion in the last 6 cm of the terminal ileum with proximal dilation. Only the histology of the resected surgical specimen proved the presence of a mixed adenocarcinoid tumor involving the terminal ileum.
Conclusion
Carcinoid tumor should be suspected in elderly patients with Crohn's disease presenting with intestinal obstruction and laparotomy should be considered to exclude malignancy.
==== Body
Background
It has been recognized that patients with Crohn's disease (CD) are at increased risk of developing malignant lesions [1,2]. Adenocarcinoma is one of the most common malignant tumors of the small intestine complicating CD, even if it remains quite rare compared with the large bowel malignancies [3-5]. Carcinoid tumors have also been described in association with CD [6,7].
We report here a case of ileal Crohn's disease associated with both adenocarcinoma and carcinoid tumor.
Case presentation
A 64-year old woman was admitted to the Gastroenterology Unit of another Hospital for abdominal pain, nausea, weight loss, and recurrent episodes of constipation. She had suffered from intermittent pain located in the right lower quadrant for 2 years, but denied chronic diarrhea. She was diagnosed as having Crohn's disease of the terminal ileum by ileocolonoscopy and histology. An upper gastrointestinal series excluded CD in the upper gastrointestinal tract.
After inductive therapy with azathioprine and corticosteroids, the patient was maintained with mesalazine 500 mg × 4 daily. Prednisone 50 mg/day was added only during acute phases; in the six years following the discharge, 2 episodes of abdominal pain and dyarrhea requiring steroid therapy were recorded. Seven years later the patient was admitted to our Institution for abdominal pain, nausea, iron deficiency anemia, hypoalbuminemia, and intestinal obstruction. On physical examination no masses in the abdomen were noted. Small bowel double-contrast enteroclysis (Fig. 1) showed a stricture of the terminal ileum with a thickening of the wall without extravasation of the contrast material. CT scan performed by using intravenous and oral contrast material revealed thickening of the terminal ileum with stranding of the mesenteric fat. There was no clinical evidence of carcinoid tumor so biochemical tests for carcinoid syndrome were not carried out.
At operation, there was a stricturing lesion in the last 6 cm of the terminal ileum with proximal dilation. The wall of the affected ileal loop and caecum was thick and fibrous. Resection of the terminal ileum and caecum was performed. The intestinal continuity was obtained with an end-to-end ileocolic hand-sewn anastomosis.
The surgical specimen included 28 cm of the terminal ileum and caecum. Grossly, all the length of surgical resection showed diffuse thickness of the intestinal wall. There was no evidence of malignancy (Fig. 2). Histologically, all resected surgical specimen was characterized by deep ulcers, marked proliferation of small lymph nodules involving all layers of the intestinal wall, sometimes with sarcoid-type granulomas and serosal inflammation. Moreover, two malignancies growing together were found in the ileal loop. The first showed glandular differentiation of large cells with round to oval nuclei sometimes with large nucleoli and abundant pale cytoplasm infiltrating the entire wall (pT2N0M0). The latter consisted of quite uniform medium size cells with round nuclei and inconspicuous nucleoli growing in a trabecular or acinar pattern (Fig. 3). Only the immunohistochemistry of this latter tumoral population showed strong and diffuse immunoreactivity to neuroendocrine markers such as Chromogranin A and Synaptophisin. All excised mesenteric lymph nodes were negative for malignant cells. The postoperative recovery was uneventful and the patient was discharged in general good condition. At 18-months follow-up the patient, checked by colonoscopy and ileoscopy was well, free from malignancy relapse or recurrence of CD, and with normal 24-hour urinary excretion of 5-hydroxyindoleacetic acid (5-HIAA).
Discussion
Malignant tumors of the small intestine are very rare in the general population, so even the few cases reported in the Literature in association with CD are enough to consider CD patients at 4–20 fold increased risk for small bowel malignant neoplasm [3-5].
About 2% of patients affected by CD will develop cancer in the course of their disease, and, in contrast to the patients with ulcerative colitis (UC), those with CD are at risk for developing malignancy even in the first decade of their disease [5]. The coexistence of CD with adenocarcinoma is predominantly seen in men, in the patients with excluded loop, and most frequently in the distal ileum in an area of active disease [2,5]. Most patients presenting with adenocarcinoma complicating CD have a high-grade malignancy with lymph node involvement or distant metastases, because the similarity of the presenting symptoms and of the radiography of these pathologies creates diagnostic problems for the physician and make an early diagnosis impossible [3].
Until now, there is no epidemiological evidence that patients with CD are at increased risk for carcinoid tumor [8]. However, when associated with CD, carcinoid tumors tend to be malignant and have a worse prognosis [6]. The majority of carcinoid tumors are asymptomatic and discovered at histology. When symptomatic, they may cause pain due to the intestinal obstruction or perforation mimicking CD clinically and radiologically and creating difficulty in the differential diagnosis.
Only three cases of a coexistent CD and mixed adenocarcinoid tumor have been reported in the English Literature [8-10]. Ours is the fourth case in the Literature, and the third in which the CD and the mixed adenocarcinoid tumor were located in the terminal ileum. As in the three previous reported cases, the definitive diagnosis of mixed adenocarcinoid tumor associated with CD was possible only on histology of the resected surgical specimen. Despite the unfavourable prognosis, our patient is still alive and free from malignancy recurrence at 18-months follow-up.
Reviewing the Literature [1,6,7], we believe that an ileal carcinoid tumor should be suspected in elderly CD patients presenting with obstructive symptoms as in our case. Moreover, when technically feasible, an ileoscopy with biopsy may play a role in the differential diagnosis between CD and small bowel malignancy favoring a precise early diagnosis and a more appropriate treatment. However, laparotomy should immediately be considered in CD patients presenting a diagnostic dilemma.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
UC, MDS enrolled the patient for clinical and surgical aspects, drafting the article.
SF histological examinations.
MMC setting up images, critical revision of the article.
AL radiological imaging.
ECA performed surgical operations, critical revision of the article, final approval of the version
All Authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Figures and Tables
Figure 1 Small bowel double-contrast enteroclysis (detail) showing a stricture (arrow) of the terminal ileum with a thickening of the wall.
Figure 2 The ileal loop shows an important thickness of the wall with disappearance of the normal mucosal plicae.
Figure 3 Strong immunoreactivity for Chromogranin A of carcinoid tumor infiltrating the muscular layer of the intestinal wall is clearly evident. The deep ulcer of CD is also present. (Magnification: ×25; Chromogen diaminobenzidine).
==== Refs
Kortbeek J Kelly JK Preshaw RM Carcinoid tumors and inflammatory bowel disease J Surg Oncol 1992 49 122 126 1738234
Baisse B Fontolliet C Bian YS Vuilleumier H Benhattar J Synchronous ileal and colonic adenocarcinomas associated with Crohn's disease: report of a case with focus on genetic alterations and carcinogenesis J Clin Pathol 2004 57 885 87 15280414 10.1136/jcp.2003.014811
Solem CA Harmsen WS Zinsmeister AR Loftus EV Jr Small intestinal adenocarcinoma in Crohn's disease. A case control study Inflamm Bowel Dis 2004 10 32 35 15058524 10.1097/00054725-200401000-00005
Mellemkjaer L Johansen C Gridley G Linet MS Kjaer SK Olsen JH Crohn's disease and cancer risk (Denmark) Cancer Causes and Control 2000 11 145 50 10710198 10.1023/A:1008988215904
Savoca PE Ballantyne GH Cahow CE Gastrointestinal malignancies in Crohn's disease. A 20-year experience Dis Colon Rectum 1990 33 7 11 2295280 10.1007/BF02053192
Bassi A Loughran C Foster P Carcinoid tumour of the terminal ileum simulating Crohn Disease Scand J Gastroenterol 2003 38 1004 1006 14531542 10.1080/00365520310003930
Hsu EY Feldman JM Lichtenstein GR Ileal carcinoid tumors simulating Crohn's Disease: Incidence among 176 consecutive cases of ileal carcinoid Am J Gastroenterol 1997 92 2062 64 9362193
Auber F Gambiez L Desreumaux P Mudry J Lecomte-Houcke M Cortot A Quandalle P Colombel JF Mixed adenocarcinoid tumor and Crohn's disease J Clin Gastroenterol 1998 26 353 54 9649033 10.1097/00004836-199806000-00035
Van Landingham SB Kluppel S Symmonds R JrSnyder SK Coexisting carcinoid tumor and Crohn's disease J Surg Oncol 1983 24 310 4 6361389
Hock YL Scott KWM Grace RH Mixed adenocarcinoma/carcinoid tumour of large bowel in a patient with Crohn's disease J Clin Pathol 1993 46 183 5 8459042
|
16336666
|
PMC1322224
|
CC BY
|
2021-01-04 16:03:05
|
no
|
BMC Cancer. 2005 Dec 8; 5:157
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-157
|
oa_comm
|
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-651630574010.1186/1471-2202-6-65Research ArticleA truncated Kv1.1 protein in the brain of the megencephaly mouse: expression and interaction Persson Ann-Sophie [email protected] Göran [email protected] Malin [email protected] Kristoffer [email protected] Johanna [email protected] Susanna [email protected]Århem Peter [email protected] Martin [email protected] Catharina [email protected] Neurogenetic Unit, Department of Molecular Medicine and Surgery, CMM, Karolinska Hospital, Karolinska Institutet, 171 76 Stockholm, Sweden2 Nobel Institute for Neurophysiology, Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden3 The Ludwig Institute for Cancer Research, Stockholm Branch, Stockholm, Sweden2005 23 11 2005 6 65 65 15 7 2005 23 11 2005 Copyright © 2005 Persson 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 megencephaly mouse, mceph/mceph, is epileptic and displays a dramatically increased brain volume and neuronal count. The responsible mutation was recently revealed to be an eleven base pair deletion, leading to a frame shift, in the gene encoding the potassium channel Kv1.1. The predicted MCEPH protein is truncated at amino acid 230 out of 495. Truncated proteins are usually not expressed since nonsense mRNAs are most often degraded. However, high Kv1.1 mRNA levels in mceph/mceph brain indicated that it escaped this control mechanism. Therefore, we hypothesized that the truncated Kv1.1 would be expressed and dysregulate other Kv1 subunits in the mceph/mceph mice.
Results
We found that the MCEPH protein is expressed in the brain of mceph/mceph mice. MCEPH was found to lack mature (Golgi) glycosylation, but to be core glycosylated and trapped in the endoplasmic reticulum (ER). Interactions between MCEPH and other Kv1 subunits were studied in cell culture, Xenopus oocytes and the brain. MCEPH can form tetramers with Kv1.1 in cell culture and has a dominant negative effect on Kv1.2 and Kv1.3 currents in oocytes. However, it does not retain Kv1.2 in the ER of neurons.
Conclusion
The megencephaly mice express a truncated Kv1.1 in the brain, and constitute a unique tool to study Kv1.1 trafficking relevant for understanding epilepsy, ataxia and pathologic brain overgrowth.
==== Body
Background
The megencephaly mice (BALB/cByJ-Kv1.1mceph/mceph, here denoted mceph/mceph) [1] have an 11 base pair deletion in the Shaker-like voltage-gated potassium channel subunit Kv1.1 [2]. This mutation causes progressive postnatal complex partial seizures and a unique pathologic brain overgrowth [2]. The enlargement is not uniform but restricted to the hippocampus and ventral cortex, with 28% and 72% larger area compared to wildtype at 12 weeks of age [3]. The enlargement is in part due to that the numbers of both neurons and glia cells are dramatically increased in the hippocampus which is caused by increased proliferation and/or reduced apoptosis (Almgren et al, unpublished).
The 11 base pair deletion in Kv1.1 leads to a frame shift and a premature stop codon. The predicted truncated Kv1.1 protein (MCEPH) will retain only the N-terminal (T1) domain, the first transmembrane domain (S1) and the first extracellular loop. Voltage gated potassium channels form a diverse group of membrane proteins, regulating membrane potential, neuronal excitability and nerve signaling [4]. The channels are hetero- or homotetramers, formed by a great variety of subunits, classified in 12 subfamilies [5]. Kv1.1 belongs to the Kv1 subfamily, consisting of eight members (Kv1.1 to Kv1.8). Kv1.1, Kv1.2 and Kv1.4 are the most abundant Kv1 subunits expressed in the brain [6]. Although the heteromeric structure means that a very large number of K channels can be formed in theory [7,8], the composition of the heteromultimers in the mammalian brain seems to be precisely regulated [9,10]. A direct link between potassium channel dysfunction and apoptosis is that reduced intracellular potassium levels appear to promote critical events early in the suicide program. Treatment with potassium channel blockers have been shown to block apoptosis in various cell types [11]. However, neither Kv1.1 mutations, the Kv1.1 null mouse [12] nor other epileptic models have previously been associated with pathologic brain overgrowth. Therefore, we hypothesized that the truncated Kv1.1 would be expressed and dysregulate other Kv1 subunits in the mceph/mceph mice.
Truncated proteins are usually not expressed. This is because mRNAs with a premature stop codon are degraded through nonsense-mediated mRNA decay (NMD) [13]. One exception to this is genes with no introns such as the Kv1 genes [14]. In mceph/mceph mice in situ hybridization have shown that there is no decay but instead an increased expression of Kv1.1 mRNA in the hippocampus, cortex and ventral cortex [2]. Thus, it is possible that MCEPH is expressed.
In humans Kv1.1 point mutations are reported in patients with the autosomal dominant disorder episodic ataxia type1 (EA1), [15]. Most are missense mutations but there is one report of a premature stop codon, resulting in a Kv1.1 protein that is truncated in the C-terminal (Kv1.1ΔC79). The patient carrying this mutation suffers from a drug resistant form of EA1 [16].
Truncated Kv1.1 channels have previously been studied in cell culture and Xenopus oocytes. When the Kv1.1ΔC79 protein is expressed in cell culture it is trapped in the endoplasmic reticulum (ER) and degraded [17]. However, full length Kv1.1 in cell culture is also retained in the ER [18,19]. Another experimental Kv1.1 protein truncated in the extracellular loop between S1 and S2 is able to assemble with full length Kv1.5 subunits in cell culture and the resulting complexes are trapped in the ER [20]. In Xenopus oocytes both the N- and C-terminal truncated Kv1.1 variants have a dominant negative effect on currents when coinjected with full length Kv1.1 and Kv1.2 [21,22]. Knowing if the MCEPH protein is expressed and how it interacts with other Kv1 subunits would provide further understanding of Kv1 trafficking and clues to the downstream effects seen in the mceph/mceph mice.
In this paper we show that a truncated Kv1.1 protein, MCEPH is expressed in the brain of the megencephaly mouse. The truncated protein is trapped in the ER and does not reach the plasma membrane of neurons. MCEPH has the ability to form multimers with Kv1.1. It has a dominant negative effect on Kv1.2 and Kv1.3 currents in Xenopus oocytes, but in the brain it does not appear to retain Kv1.2 in the ER.
Results
Antibody characterization
A polyclonal antiserum was produced by immunization of rabbits with a synthetic peptide corresponding to amino acids 4 to 27 of Kv1.1. The N-terminal epitope was necessary to be able to detect the truncated MCEPH protein. Western blot on brain lysate from wild type mice using the antiserum showed a strong band at 86 kDa but also several weaker bands. The 86 kDa band had the same size as that detected by the previously characterized monoclonal antibody against a C-terminal epitope of Kv1.1 [23]. Preimmune serum did not recognize the 86 kDa band but showed the same weaker bands as the antiserum. The antiserum was affinity purified against the peptide used for immunization which reduced the number of unspecific bands significantly (Figure 1 panel A). Preincubation of the affinity purified antibody with the peptide used for immunization absorbed the signal (Figure 1 panel B). The purified Kv1.1 N-terminal antibody was used for all further experiments.
In immunblotting of lysate from Kv1.1 null mice an 86 kDa band was present but at a lower intensity than in wild type (Figure 1, panel A). Since the null mouse does not have any Kv1.1 protein this means that the antibody cross reacts with another protein. Kv1 subunits have a high degree of similarity so it is likely that the antibody recognizes another Kv1 subunit. The peptide used for immunization share a stretch of six amino acids with Kv1.2. Both Kv1.1 and Kv1.2 are expressed in the brain as mature glycoproteins that are reported to be 86 and 88 kDa, respectively [23]. The shared amino acids and the size of the band makes Kv1.2 the likely candidate for the cross reactivity.
Acetone post fixed wild type and Kv1.1-null brains were used for immunohistochemistry. In wild type the staining resembled Kv1.1 but the background was high. The pattern in Kv1.1 null brain was very similar to wild type, indicating cross reactivity (data not shown). Immunohistochemistry on formalin fixed wild type brain sections showed the same staining pattern that previously has been reported for Kv1.1 [2]. In the hippocampus the immunoreactivity was mainly localized in the fiber networks in the middle part of the molecular layer, in the hilus and the pyramidal cell layer of CA3 (Figure 2, panel A). When formalin fixed tissue was used the pattern in Kv1.1 null was changed and clearly distinguishable from wild type (Figure 2, panel B).
The conclusion is that the antibody recognizes Kv1.1. It is cross reactive when used for immunoblotting but this is greatly reduced when used for immunohistochemistry on formalin fixed tissue.
MCEPH expression in the brain
The 11 bp deletion in Kv1.1 in mceph/mceph mice leads to a premature stop codon. The predicted truncated protein, MCEPH, has a calculated weight of 27 kDa. To determine if MCEPH is expressed in the brain the polyclonal Kv1.1 N-terminal antibody was used for immunoblotting and immunohistochemistry. Immunoblotting showed a band at approximately 30 kDa in brain lysate from mceph/mceph mice (Figure 1, panel C). The intensity of the band was much lower than that of the Kv1.1 protein in wild type mice suggesting a low expression of the truncated protein. The mceph/mceph mice do not have any full length Kv1.1 but, like the Kv1.1 null mice, they have the 86 kDa cross reactivity-caused band. However, the 30 kDa band was only present in mceph/mceph mice. No other band specific for mceph/mceph was detected.
Immunohistochemistry was performed on formalin fixed mceph/mceph mice brain sections. Interestingly, in mceph/mceph there was staining only around the nuclei of cells, and no staining in fibers (Figure 2, panel C and D). There was some background staining resembling that of the Kv1.1 null mouse but the overall pattern was significantly different. The main areas with MCEPH immunoreactivity were the hippocampus and ventral cortex. In the other areas examined there were only a few stained cell bodies. In the hippocampus MCEPH was expressed in neurons in the dentate gyrus, CA1 and in CA3. This pattern is in agreement with the previously reported mRNA expression [2]. Taken together, the immunoblotting and immunohistochemistry results show that MCEPH is expressed in the brain of the megencephaly mouse.
Trafficking of MCEPH
Kv1.1 has a single N-glycosylation site on the extracellular loop between S1 and S2 and is expressed on the plasma membrane as a mature glycoprotein. This glycosylation site is preserved in MCEPH which makes it possible to study the trafficking of MCEPH by analyzing its glycosylation pattern. Glycosylation starts in the ER where a high mannose carbohydrate is transferred to an asparagine residue. This carbohydrate can be cleaved off by the glycosidases EndoH and PNGaseF. Next, the protein is transported to the Golgi where the carbohydrate is modified and becomes resistant to EndoH but not to PNGaseF. This is used as a marker for ER to Golgi transport.
To investigate the glycosylation status, brain lysates from wild type and mceph/mceph mice were treated with the enzymes EndoH and PNGaseF and analyzed with immunoblotting. For Kv1.1 in wild type mice EndoH did not affect the size of the 86 kDa band but PNGaseF reduced it with approximately 20 kDa. This is in agreement with other studies that have shown that the 86 kDa band is the mature glycosylated form of Kv1.1 expressed in the plasma membrane [18]. Both EndoH and PNGaseF reduced the apparent molecular weight of MCEPH with approximately 3 kDa (Figure 3). This strongly indicates that MCEPH is core glycosylated in the ER but not transported to the Golgi. Neither mature nor unglycosylated MCEPH was detected in the brain.
Interactions of MCEPH
Kv1 channels are tetramers and the assembly of the Kv1 subunits takes place in the ER [24]. MCEPH contains the N-terminal tetramerization domain and might thus bind other Kv1 subunits.
Interactions with Kv1.2 in the brain
Kv1.1 and Kv1.2 are often associated in the same channels in the brain [10,25]. We have previously reported a disturbance in the expression of Kv1.2 in the mceph/mceph hippocampus [2]. To investigate if this disturbance could be caused by an interaction between MCEPH and Kv1.2, immunoprecipitation followed by immunoblotting was performed on brain lysate from wild type, Kv1.1 null and mceph/mceph mice. The monoclonal Kv1.1 and Kv1.2 antibodies each precipitated both Kv1.1 and Kv1.2 in wild type brain, verifying that correct experimental procedures were used. As a control the same antibodies were used for immunoprecipitation of lysate from Kv1.1 null brain. As expected, in the Kv1.1 null lysate no Kv1.1 coprecipitated with Kv1.2. Also, when the Kv1.1 antibody was used for immunoprecipitation neither Kv1.1 nor Kv1.2 was detected (Figure 4 panel A). The Kv1.1-N-terminal antibody precipitated Kv1.1 and Kv1.2 in wild type but also Kv1.2 in Kv1.1 null lysate. This means that it is not specific when used for immunoprecipitation. For mceph/mceph, immunoprecipitation was performed with anti-Kv1.2 whereas the Kv1.1-N-terminal antibody was used for immunoblotting. With this approach it was not possible to detect any MCEPH co-precipitating with Kv1.2 (Figure 4 panel B). This suggests that there is no persistent interaction between Kv1.2 and MCEPH in the brain.
On Western blots, Kv1.2 in brain appears as two bands at 60 and 88 kDa representing core and mature glycosylated form, respectively [26]. If MCEPH binds and traps Kv1.2 in the ER, the amount of core glycosylated Kv1.2 would be expected to increase and the mature glycosylated form decrease. Immunoblotting was performed on lysates from both whole brains and isolated hippocampi. The hippocampus was investigated since MCEPH is expressed primarily in this brain region. Most of the Kv1.2 protein in the hippocampus had mature glycosylation and there was only a very small amount of core glycosylated protein. No difference in Kv1.2 core glycosylation was seen between wild type and mceph/mceph in neither whole brain nor hippocampus lysate (Figure 4 panel C). When the immunoprecipitation and glycosylation data is combined it appears that MCEPH does not retain Kv1.2 in the ER in brain.
Interactions with Kv1.1 in HEK293 cells
To test if MCEPH can form multimers its ability to interact with wild type Kv1.1 was investigated by overexpression in cell culture. This system was chosen to allow higher expression levels and easier detection. For this experiment Kv1.1 and MCEPH were cloned into the vectors pZsGreen and pDsRed2 to construct fluorescent fusion proteins. HEK293A cells were transfected with the MCEPH-ZsGreen and the Kv1.1-DsRed constructs together or separately and analyzed 48 hours after transfection. The green and the red fluorescent signals were colocalized around the cell nucleus (data not shown). No staining was observed on the plasma membrane. Western blot with the polyclonal Kv1.1 N-terminal antibody on lysate from cotransfected HEK293 cells showed bands at 55 and 85 kDa corresponding to the MCEPH and Kv1.1 fusion proteins (Figure 4 panel D). Corresponding bands were found in cells transfected with single constructs which confirmed the identity of the bands. To investigate the interaction between MCEPH and Kv1.1, immunoprecipitation was performed on cell lysate from cotransfected cells using the monoclonal Kv1.1 C-terminal antibody. Immunoblotting with the polyclonal Kv1.1 N-terminal antibody detected the MCEPH-ZsGreen fusion protein as well as the Kv1.1-DsRed fusion protein (Figure 3 panel D). Since the C-terminal Kv1.1 antibody is specific for wild type Kv1.1 the detected MCEPH protein was associated with the Kv1.1 protein.
Interactions with Kv1.2 and Kv1.3 in Xenopus oocytes
To further investigate the ability of MCEPH to interact with Kv1 subunits we chose to analyze the effects of expressing MCEPH in combination with Kv1.2 or Kv1.3 in Xenopus laevis oocytes. Figure 5a shows the Kv1.2 and Kv1.3 currents associated with a series of voltage steps, followed by a step to +30 mV. Both channels show delayed rectifier behaviour and lack fast inactivation. Kv1.3, however, shows a marked slow inactivation.
MCEPH was found to reduce the expression of both Kv1.2 and Kv1.3 in the plasma membrane. Figure 5b shows the mean results of injecting mceph mRNA together with Kv1.2 or Kv1.3 mRNA (in equal amounts) given as the effect on peak conductance versus voltage curves (i.e activation curves). As seen, the curves for Kv1.2 and Kv1.3 are down-scaled by about 50 and 60%, respectively, without marked changes in slope or midpoint value. The MCEPH shift of the Kv1.3 activation curve was slightly bigger than that of Kv1.2 (-6 mV vs. +2 mV; Table 1). The activation time course of the currents associated with voltage steps was negligibly affected by MCEPH, reflected in the unaffected time to peak current at +50 mV for both Kv1.2 and Kv1.3 (Table 1). Similar effects were observed for the MCEPH effect on the slow inactivation kinetics. Mean inactivation versus voltage curves, measured from the peak current associated with the second pulse (Figure 5a) are shown in Figure 5c. The MCEPH shift of the Kv1.3 inactivation curve was slightly bigger than that of Kv1.2 (-5 mV vs. -1 mV; Table 1).
In summary, the measurements show that MCEPH has a dominant negative effect on both Kv1.2 and Kv1.3 channel expression in Xenopus oocytes. Furthermore, the small effects on activation and inactivation kinetics suggest that MCEPH does not form functional channel complexes with Kv1.2 or Kv1.3 subunits in the plasma membrane. Heteromeric complexes involving MCEPH subunits are expected to drastically affect channel kinetics.
Discussion
In this study we show that mceph/mceph mice express a truncated Kv1.1, MCEPH. The MCEPH protein is expressed in neurons throughout the brain but is trapped in the ER. MCEPH is able to bind full length Kv1.1 in cell culture and has a dominant negative effect when coexpressed with Kv1.2 and Kv1.3 in Xenopus oocytes. In contrast, in the brain MCEPH does not appear to retain Kv1.2 in the ER.
To be able to detect the truncated MCEPH protein an antibody against the N-terminal part of Kv1.1 was generated. Immunoblotting on brain lysate from mceph/mceph mice showed a band that has the same size as the predicted MCEPH protein, approximately 30 kDa. This band is present in all mceph/mceph brains studied but not in any wild type or Kv1.1 null mice. Based on this we conclude that MCEPH is expressed in the brain of the megencephaly mouse.
Kv1.1 in the brain is a mature glycoprotein that is mostly localized to axons and nerve terminals [6]. Truncated Kv1.1 variants have previously been shown to be retained in the ER in cell culture [17,20]. The draw back of these studies is that full length Kv1.1 homomers in cell culture is also largely retained in the ER since they are undetectable at the cell surface by immunohistochemical methods [19] but detectable at low levels with electrophysiological methods [27]. Lack of surface expression was seen by us both in HEK293 cells and in the neuron-like cell line PC12 (data not shown). The reason is a strong ER retention motif in the pore region [28]. The mceph/mceph mouse provides an opportunity to study the trafficking of a truncated Kv1.1 in neurons. Both EndoH and PNGaseF treatment reduced the molecular weight of MCEPH by approximately 3 kDa, suggesting that the protein is core glycosylated [18]. We detected neither maturely glycosylated nor unglycosylated MCEPH. The glycosylation data suggest that MCEPH is trapped in the ER and not further transported. This is supported by immunohistochemistry experiments were MCEPH immunoreactivity surrounds the nuclei in an ER-like pattern. There is no immunoreactivity in fibers consistent with the lack of mature glycosylated MCEPH. The low amount of MCEPH protein detected with immunoblotting compared to the high levels of mRNA suggests that the MCEPH protein is rapidly degraded. This is most likely due to ER-associated degradation (ERAD), a quality control mechanism in the ER that recognizes misfolded proteins [29]. These proteins are retro-translocated to the cytosol and degraded by proteasomes.
We have previously seen disturbances of staining pattern of Kv1.2 and Kv1.3 in the mceph/mceph hippocampus, which was not due to transcriptional regulation [2]. Previous studies of other truncated K1.1 variants have shown that they have a dominant negative effect in oocytes and form oligomers with full length Kv1 subunits in cell culture. We hypothesized that MCEPH interacts with Kv1 subunits and causes the expression disturbance possibly by retaining these subunits in the ER. Kv1.1 often forms heterotetramers with Kv1.2 subunits in the brain [10,25]. Nevertheless, we found no evidence for a persistent interaction between MCEPH and Kv1.2 in the brain, suggesting that there is no retention of Kv1.2 in the ER. However, we found MCEPH to be able to oligomerize when expressed in HEK293 cells. Moreover, electrophysiological measurements on Xenopus oocytes showed that MCEPH reduces both Kv1.2 and Kv1.3 currents, suggesting interaction between MCEPH and Kv1 subunits in this cell type. This indicates that the lack of ER retention seen for Kv1.2 in the brain is not caused by an inability of MCEPH to form multimers
It can be assumed that lack of Kv1.1 could lead to defect trafficking of Kv1 channels. This is not the case since the Kv1.1 null mouse has normal expression of Kv1.2 in the hippocampus (data not shown) and also the cerebellum and sciatic nerve [12,30]. One possibility is that the Kv1.2 and Kv1.3 expression disturbances in mceph/mceph hippocampus are related to the growth. The mceph/mceph hippocampus is significantly enlarged and has an increased amount of neurons (unpublished data). This could lead to disturbed organization of the dendritic trees and give rise to the abnormal expression patterns. Also, since MCEPH contains the N-terminal T1 domain, which has been shown to mediate beta subunit association [31,32], it is possible that MCEPH sequesters these auxiliary subunits within the ER and prevents them from forming complexes with Kv1.2 and Kv1.3 in mceph/mceph mice. Beta subunit association has been shown to promote Kv1 trafficking to the cell surface as well as to cause axonal targeting of Kv1.2 [33].
The megencephaly mouse is not a model for any known human disease. Still, it provides a unique tool to study Kv1 channels in the brain. The EA1 mutation R417Stop gives rise to a Kv1.1 protein that lacks the final 79 amino acids in the C-terminal domain (Kv1.1ΔC79). The Kv1.1ΔC79 protein and MCEPH have some similarities but there are major differences. Both truncated proteins have a dominant negative effect in oocytes and can assemble with full length subunits in cell culture. For the Kv1.1ΔC79, change in current kinetics when coinjected with Kv1.1 in Xenopus oocytes indicates that some channels containing truncated subunits can reach the plasma membrane [16]. This is not seen for MCEPH which is most likely due to the difference in length. Kv1.1ΔC79 might evade the rapid degradation by cellular control mechanisms since most of the protein remains. The truncated protein can then cause more damage by retaining subunits or by forming defect channels. This is supported by the inheritance patterns. The R417Stop mutation is dominant which suggests a new function for the truncated protein, while mceph is recessive, which indicates a loss of function.
The mechanism behind the brain growth in mceph/mceph is still unknown. We now know that MCEPH is expressed in the brain, but we do not know if it is MCEPH or lack of functional Kv1.1 that is responsible for the overgrowth. The Kv1.1 null mouse has not been reported to have an enlarged brain which suggests that MCEPH has an effect. However, the low levels of protein and recessive inheritance indicates that the mceph mutation causes a loss of function of Kv1.1. In line with both scenarios, it was previously shown that potassium channel dysfunction appears to reduce apoptosis [11], which might contribute to the overgrowth of the mceph/mceph brain.
Conclusion
For the first time we report the expression of a truncated Kv1.1 protein in brain. This protein is trapped within ER and has the ability to interact with Kv1 subunits. The megencephaly mice, where the protein is expressed, constitute a tool to study Kv1.1 trafficking relevant for understanding epilepsy, ataxia and pathologic brain overgrowth.
Methods
Antibodies
Primary antibodies
A rabbit polyclonal antibody against the N-terminal of Kv1.1 was generated using a synthetic peptide (CMSGENADEASTAPGHPQDGSYPRQ) corresponding to amino acids 4–27. Kv1 subunits have a high degree of homology. The peptide for immunization was selected to have as low similarity as possible to other Kv1 subunits. The peptide was conjugated to KLH and injected into rabbits for production of antiserum. The antiserum was affinity purified against the peptide using standard procedures. The affinity purified antibody was used at a 1:500 dilution for immunoblotting and at 1:400 for immunohistochemistry. Monoclonal anti-Kv1.1 and anti-Kv1.2 (Upstate Biotechnologies Inc, Waltham, MA, USA) were used for immunoblotting at dilutions of 1:200 and 1:1000, respectively.
Secondary antibodies
For immunoblotting, horseradish peroxidase coupled swine anti-rabbit (DAKO, Glostrup, Denmark 1:2000) and goat anti-mouse (Upstate Biotechnologies Inc 1:5000) were used. For immunohistochemistry a FITC labeled goat anti-rabbit antibody (Sigma-Aldrich, St. Louis, MO, USA 1:300) was used.
Experimental animals and preparation of tissues
All studies were performed in accordance with guidelines from the Swedish National Board for Laboratory Animals. The BALB/cByJ-mceph/mceph, BALB/cByJ-+/+ (wild type) and C3HeB/FeJ-Kcna1tm1Tem (Kv1.1 null) [12] mouse strains were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). Animals were euthanized by CO2, and the brains were removed and immersed in ice-cold phosphate-buffered saline (PBS) and immediately thereafter frozen on dry ice. Serial 14 μm sections were cut in a cryostat (Jung CM 3000, Leica Instruments GmbH, Nussloch, Germay) at -15° and thaw-mounted on microscope glass slides (Probe On, Fischer Scientific, Pittsburgh, PA, USA).
For immunohistochemistry, mice were anesthetized with isoflurane and perfused via the ascending aorta with Ca2+-free Tyrode's solution followed by a fixative containing 4.0% paraformaldehyde and 0.4% picric acid in 0.16 M phosphate buffer. Brains were dissected out, immersed in ice-cold fixative for 90 min and then rinsed in 0.1 M phosphate buffer (pH 7.4) containing 10% sucrose, 0.01% sodium azide and 0.02% bacitracin (Sigma-Aldrich). Serial 14 μm sections were cut in a cryostat and mounted on gelatin/chrome-alum coated slides.
For immunoprecipitation and immunoblotting, brains were homogenized in RIPA buffer (150 mM NaCl, 50 mM Tris pH 8,0, 1% NP-40, 0,1% SDS, 0,5% sodium deoxycholate) with a protease inhibitor cocktail (Sigma-Aldrich) and centrifuged at 1000 × g to pellet debris.
Plasmids
The complete Kv1.1 mRNA sequence [NCBI: NM_010595] was PCR amplified from genomic DNA from mceph/mceph and +/+ mice using High Fidelity Taq polymerase (Roche Diagnostics, Basel, Switzerland). The PCR products were TA-cloned into pCR2.1 (Invitrogen, Carlsbad, CA, USA). The coding sequence was then subcloned into the oocyte expression vector pGEM-HE (Protinac GmbH, Hamburg, Germany) using the HindIII sites at position 823 and 2712. Correct orientation of the insert was confirmed by restriction analysis and sequencing. Rat cDNA for Kv1.2 and Kv1.3 in the pGEM-HE vector was obtained from Protinac.
For expression in mammalian cells the coding sequence of Kv1.1 was PCR amplified from genomic DNA from mceph/mceph and +/+ mice using two different reverse primers that removed the stop codons. The PCR products were cloned into the vectors pDsRed2-N1 and pZsGreen-N1 (BD Biosciences Clontech Palo Alto, CA, USA) to construct fluorescent fusion proteins.
Cell culture and transfection
Human embryonal kidney cells HEK293A were cultivated in Dulbecco's modified Eagle's media (DMEM) supplemented with L-glutamine, 10% fetal bovine serum (Gibco, Rockville, MD, USA) and 100 μg/ml penicillin/streptomycin (Gibco) in a 37°C incubator containing 5% CO2 humidified air. For transfection, HEK293 cells were plated on 100 mm dishes. At 80% confluency the cells were transfected using Lipofectamine and Plus reagent (Invitrogen) according to the manufacturers protocol and analyzed 48 hours after transfection. The HEK293A cells were lysed in RIPA buffer with a protease inhibitor cocktail (Sigma-Aldrich). The lysate was spun 1000 × g to pellet debris and used for immunoprecipitation and immunoblotting.
Immunoblotting, immunoprecipitation and glycosylation analysis
For immunoblotting, brain or cell lysate was diluted with 2× Laemmli loading buffer with β-mercaptoethanol and loaded on 10% SDS-PAGE with 4% stacking gel. Proteins were transferred to nitrocellulose membranes (Hybond-ECL, Amersham Biosciences, Little Chalfon, UK). The membranes were blocked in 5% milk and incubated with the primary antibody. This was followed by incubation with a horseradish peroxidase coupled secondary antibody. ECL Western blotting reagent (Amersham Biosciences) was used for detection according to the manufacturer's protocol.
For immunoprecipitation, brain lysate was incubated with 4 μg of the indicated antibody for 2 hours on ice. Protein-A coupled sepharose beads (Amersham Biosciences) were added and the mixture was incubated for 1 hour. The beads were spun down and the supernatant was removed. The beads were then washed three times with RIPA buffer and once with 50 mM Tris pH 8.0. They were then resuspended in Laemmli loading buffer and boiled for 5 min. The supernatant was collected and analyzed by immunoblotting.
For glycosylation analysis, brain tissue was homogenized in a lysis buffer containing 50 mM Tris pH 8.0, 150 mM NaCl, 1 mM EDTA, 1% Triton X and protease inhibitors. The homogenate was spun at 1000 × g to pellet debris. The lysate was denatured with 0.1% SDS and 1% β-mercaptoethanol at 100° for 5 min. Samples were incubated with EndoH or PNGaseF according to the manufacturers protocol (New England Biolabs, Beverly, MA, USA) at 37° for 2 hours and analyzed with immunoblotting.
Immunohistochemistry
Brain sections were thawed quickly. Fresh frozen sections were fixed in ice cold acetone. All slides were blocked in PBS with 5% goat serum and 0.23% Triton-X. They were then incubated with the primary antibody diluted in blocking solution followed by incubation with a FITC labeled secondary antibody. The slides were mounted with Vectashield antifade mounting media (Vector laboratories, Burlingame, CA, USA) and examined under an Axioskop 2 microscope (Zeiss, Oberkochen, Germany).
Electrophysiology
The electrophysiological experiments comprised voltage clamp measurements on oocytes from Xenopus laevis, injected with mceph, Kcna2 and Kcna3 mRNA. Plasmids with the studied genes were linearized with restriction enzymes HindIII (Kv1.2 and Kv1.3) and Nhe1 (mceph). The linearized DNA was purified and transcribed with the Message Machine Kit (Ambion, Austin, TX, USA). The mRNA product was purified and dissolved in 10 μl ddH2O and stored at -70°C until injection in oocytes. The mRNA injected in oocytes was diluted to one tenth of the original concentration (2.4 mg/ml and 0.8 mg/ml, respectively). The oocytes (in stage V and VI) were surgically removed from the frog under anaesthesia, and treated with Liberase (0.25 mg/ml) for three hours. After being carefully rinsed the oocytes were incubated overnight in modified Barth's solution (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3, 0.015 mM HEPES, 0.33 mM Ca(NO3)2, 0.41 mM CaCl2, 0.82 mM MgSO4, pH adjusted to 7.5) with added 10 μg/ml pyruvate and 10 μg/ml penicillin-streptomycin at 10°C. They were injected with cRNA (50 nl/cell) using a Nanoject injector; Drummond Scientific, Broomall, PA) and incubated at room temperature (20–21°C) 16–24 hours before starting the electrophysiology experiments (day 3 to 4 after injection).
The electrophysiological experiments were performed with a two-electrode voltage-clamp (CA-1 amplifier, Dagan, Minneapolis, MN). Microelectrodes were made from borosilicate glass capillaries with a mechanical puller and filled with 3 M KCl solution, resulting in a resistance between 0.5 to 1.0 MΩ. The extracellular solution was composed of 88 mM NaCl, 1 mM KCl, 0.8 mM MgCl2, 0.4 mM CaCl2, 15 mM HEPES with pH is adjusted to 7.4. All measurements were carried out at room temperature (20–22°C). The holding potential was set to -80 mV and the interval between the test steps was 2 s (Kv1.2) and 30 s (Kv1.3). The recorded current was filtered by a low-pass Bessel filter (5 kHz) and sampled with intervals of 2 ms (Kv1.2) and 4 ms (Kv1.3). The software used for data collection and -analysis were pClamp5 and Clampfit 8.2 (Axon Instruments Inc., Union City, CA) and Origin 6.0 (Microcal Software Inc., Northampton, MA). To avoid different channel density levels due to trafficking during the electrophysiological experiments we restricted the measurements to a four-hour period.
The conductance (G) was calculated from the current (IK) and the associated voltage step (V) by
G = IK/(V + 80 mV). (1)
The conductance normalized to the maximal conductance under control conditions (Gmax CTRL) were fitted to the Boltzmann equation
G/Gmax CTRL = 1/[1 + exp((V - V1/2)/s)], (2)
where V1/2 is the potential at half-maximal conductance, i.e. midpoint potential, and s is the slope. Significance of differences was tested using the t-test.
Authors' contributions
SP, ASP, MS and CL initiated the study. ASP planned and performed the cloning, cell culture work, immunohistochemistry, immunoprecipitation, western blot and drafted the manuscript. MA performed and evaluated the immunohistochemistry. SP assisted in the cell culture work. GK performed most of the oocyte work, comprising expression and electrophysiology and drafted the electrophysiological section of the manuscript. KS, JN and PÅ participated in the design, performance and analysis of the oocyte experiments, and developed the oocyte section of the manuscript. CL designed and coordinated the study.
All authors edited the draft and approved the final version of the manuscript.
Acknowledgements
We thank Violeta Vladuljevic for help with cloning and Helena Källström for confocal microscopy instructions.
This work was supported by the Swedish Medical Research Council, Karolinska institutet foundations, Magnus Bergwall Foundation, Swedish Society of Medicine, the Swedish Society for Medical Research, and Åke Wibergs Foundation.
Figures and Tables
Figure 1 Immunoblot using the Kv1.1 N-terminal antibody on brain lysate from wild type, Kv1.1-null and mceph/mceph. A. Lysate from wild type (+/+), Kv1.1 null (-/-) and mceph/mceph (m/m) brains were loaded on SDS-PAGE and immunoblotted with the polyclonal Kv1.1 N-terminal antibody. In wild type lysate a strong band was detected at 86 kDa, corresponding to Kv1.1. The same band was seen in the Kv1.1 null and mceph/mceph lysates but at a lower intensity. The 86 kDa bands in Kv1.1 null and mceph/mceph lysates are due to antibody cross reactivity since neither Kv1.1 null nor mceph/mceph mice have any full-length Kv1.1 protein. B. The polyclonal Kv1.1 N-terminal antibody was preincubated with the peptide used for immunization. Lysate from wild type (+/+) and mceph/mceph (m/m) brains were loaded on SDS-PAGE and immunoblotted with the Kv1.1 N-terminal antibody without or after preincubation. The preincubation completely blocked the signal. C. A longer exposure of the immunoblot in panel A. In mceph/mceph brain lysate there was a unique band at approximately 30 kDa, which corresponds to the calculated weight of MCEPH (arrow).
Figure 2 Kv1.1 immunohistochemistry in wild type, Kv1.1-null and mceph/mceph hippocampus. Immunohistochemistry on formalin fixed brain sections. Exposure parameters were adjusted to obtain a strong and clear signal. Note exposure times given to relate between panels A. Wild type hippocampus showed the same staining pattern as that previously reported for Kv1.1 [2] (exposure 1100 ms) B. Kv1.1 null mouse hippocampus: some cross reactivity of the antibody was seen. (exposure 2700 ms) C. In the mceph/mceph hippocampus the immunoreactivity surrounded the nuclei of neurons especially in the dentate gyrus hilus (h) (exposure 2500 ms) D. Parietal neocortex in mceph/mceph (left) and wild type (right) brain: the staining of fibers in wild type was absent in mceph/mceph. Scale bar 200 μm; h, dentate gyrus hilus; Gr, dentate gyrus granular cell layer; wt, wild type; -/-, Kv1.1-null; m/m, mceph/mceph
Figure 3 Trafficking of MCEPH. Trafficking of MCEPH in the brain was determined by analyzing glycosylation pattern. Both EndoH and PNGaseF reduced the molecular weight with approximately 3 kDa (arrow). This corresponds to core glycosylation. No unglycosylated MCEPH was detected.
Figure 4 Analysis of interactions between MCEPH and Kv1 subunits. A. Immunoprecipitation was performed using a monoclonal Kv1.1 C-terminal antibody on brain lysate from wild type (+/+) and Kv1.1 null (-/-) mice. In wild type brain both Kv1.1 and Kv1.2 are detected. In Kv1.1 null brain neither Kv1.1 nor Kv1.2 was detected. B. Immunoprecipitation was performed with the anti-Kv1.2 monoclonal antibody on brain lysate from wild type (+/+) and mceph/mceph (m/m) mice. The immunoprecipitation reaction and corresponding brain lysate was loaded on SDS-PAGE. In brain lysate from mceph/mceph (BL) MCEPH was detected using the polyclonal Kv1.1 N-terminal antibody (arrow). However, no MCEPH band was detected in the immunoprecipitate from mceph/mceph (IP). C. Hippocampi were dissected and an equal amount of lysate from wild type and mceph/mceph was loaded on SDS-PAGE and immunoblotted with anti-Kv1.2. Only a very small fraction of the Kv1.2 was core glycosylated (60 kDa). There appeared to be no increase in the amount of core glycosylated Kv1.2 in mceph/mceph hippocampus compared to wild type. D. HEK293 cells were cotransfected with Kv1.1-DsRed and MCEPH-ZsGreen constructs. Both the 85 kDa Kv1.1-DsRed and the 55 kDa MCEPH-ZsGreen fusion proteins were detected with immunoblotting on cell lysate using the polyclonal Kv1.1 N-terminal antibody (lysate). The relative levels of the two fusion proteins in this overepressing cell system cannot be used to quantitate MCEPH expression or stability in brain since regulation and trafficking is known to be different between these two systems. The lysate was immunoprecipitated with the Kv1.1 C-terminal antibody. In the precipitate, both fusion proteins were detected with the Kv1.1 N-terminal antibody (IP). When the Kv1.1 C-terminal antibody was used for immunoblotting only the Kv1.1-DsRed fusion protein was detected.
Figure 5 Electrophysiological recordings demonstrating the interaction between MCEPH and Kv1.2 and Kv1.3 in Xenopus oocytes. A. Currents of Kv1.2 and Kv1.3, when expressed separately and when coexpressed with MCEPH in Xenopus oocytes. Pulse steps from -80 to +50 mV (increment between the steps 10 mV), followed by a step to +30 to enable measurements of the inactivation. B. Activation curves (peak conductance versus voltage) for Kv1.2 and Kv1.3, when expressed separately and when coexpressed with MCEPH. Same stimulation protocols as in A. Error bars mark standard error of mean. C. Inactivation curves (normalized peak conductance, associated with the second pulse, versus voltage of preceding step; see panel A) for Kv1.2 and Kv1.3, when expressed separately and when coexpressed with MCEPH. Same stimulation protocols as in A. Error bars mark standard deviation.
Table 1 Effects of MCEPH on Kv1.2 and Kv1.3 kinetics in Xenopus oocytes. Mean midpoints of activation and inactivation curves (V1/2) and mean half time values (t1/2) for Kv1.2 and Kv1.3 (±SD), when expressed separately and when coexpressed with MCEPH. n is the number of oocytes. The activation and inactivation curves were constructed from measurements based on the stimulation protocols in Fig. 5. Half time value (t1/2) was measured as time to half peak current at a 50 mV pulse step (±SD).
Steady-state activation Steady-state inactivation n
V1/2 (mV) t1/2 (ms) V1/2 (mV)
Kv1.2 -24 ± 4 3 ± 1 -17 ± 7 9
Kv1.2 + MCEPH -22 ± 3 3 ± 1 -18 ± 6 10
Kv1.3 -25 ± 4 6 ± 3 -29 ± 3 7
Kv1.3 + MCEPH -31 ± 4 6 ± 3 -34 ± 4 10
==== Refs
Donahue LR Cook SA Johnson KR Bronson RT Davisson MT Megencephaly: a new mouse mutation on chromosome 6 that causes hypertrophy of the brain Mamm Genome 1996 7 871 876 8995755 10.1007/s003359900259
Petersson S Persson AS Johansen JE Ingvar M Nilsson J Klement G Arhem P Schalling M Lavebratt C Truncation of the Shaker-like voltage-gated potassium channel, Kv1.1, causes megencephaly Eur J Neurosci 2003 18 3231 3240 14686897 10.1111/j.1460-9568.2003.03044.x
Diez M Schweinhardt P Petersson S Wang FH Lavebratt C Schalling M Hokfelt T Spenger C MRI and in situ hybridization reveal early disturbances in brain size and gene expression in the megencephalic (mceph/mceph) mouse Eur J Neurosci 2003 18 3218 3230 14686896 10.1111/j.1460-9568.2003.02994.x
Hille B Ion channels of excitable membranes 2001 3rd Sunderland, Mass. , Sinauer xviii, 814 p., [8] p. of plates
Yu FH Catterall WA The VGL-chanome: a protein superfamily specialized for electrical signaling and ionic homeostasis Sci STKE 2004 2004 re15 15467096 10.1126/stke.2532004re15
Misonou H Trimmer JS Determinants of voltage-gated potassium channel surface expression and localization in Mammalian neurons Crit Rev Biochem Mol Biol 2004 39 125 145 15596548 10.1080/10409230490475417
Isacoff EY Jan YN Jan LY Evidence for the formation of heteromultimeric potassium channels in Xenopus oocytes Nature 1990 345 530 534 2112229 10.1038/350232a0
Ruppersberg JP Schroter KH Sakmann B Stocker M Sewing S Pongs O Heteromultimeric channels formed by rat brain potassium-channel proteins Nature 1990 345 535 537 2348860 10.1038/345535a0
Coleman SK Newcombe J Pryke J Dolly JO Subunit composition of Kv1 channels in human CNS J Neurochem 1999 73 849 858 10428084 10.1046/j.1471-4159.1999.0730849.x
Wang FC Parcej DN Dolly JO alpha subunit compositions of Kv1.1-containing K+ channel subtypes fractionated from rat brain using dendrotoxins Eur J Biochem 1999 263 230 237 10429208 10.1046/j.1432-1327.1999.00493.x
O'Grady SM Lee SY Molecular diversity and function of voltage-gated (Kv) potassium channels in epithelial cells Int J Biochem Cell Biol 2005 37 1578 1594 15882958 10.1016/j.biocel.2005.04.002
Smart SL Lopantsev V Zhang CL Robbins CA Wang H Chiu SY Schwartzkroin PA Messing A Tempel BL Deletion of the K(V)1.1 potassium channel causes epilepsy in mice Neuron 1998 20 809 819 9581771 10.1016/S0896-6273(00)81018-1
Hentze MW Kulozik AE A perfect message: RNA surveillance and nonsense-mediated decay Cell 1999 96 307 310 10025395 10.1016/S0092-8674(00)80542-5
Brocke KS Neu-Yilik G Gehring NH Hentze MW Kulozik AE The human intronless melanocortin 4-receptor gene is NMD insensitive Hum Mol Genet 2002 11 331 335 11823452 10.1093/hmg/11.3.331
Browne DL Gancher ST Nutt JG Brunt ER Smith EA Kramer P Litt M Episodic ataxia/myokymia syndrome is associated with point mutations in the human potassium channel gene, KCNA1 Nat Genet 1994 8 136 140 7842011 10.1038/ng1094-136
Eunson LH Rea R Zuberi SM Youroukos S Panayiotopoulos CP Liguori R Avoni P McWilliam RC Stephenson JB Hanna MG Kullmann DM Spauschus A Clinical, genetic, and expression studies of mutations in the potassium channel gene KCNA1 reveal new phenotypic variability Ann Neurol 2000 48 647 656 11026449 10.1002/1531-8249(200010)48:4<647::AID-ANA12>3.0.CO;2-Q
Manganas LN Akhtar S Antonucci DE Campomanes CR Dolly JO Trimmer JS Episodic ataxia type-1 mutations in the Kv1.1 potassium channel display distinct folding and intracellular trafficking properties J Biol Chem 2001 276 49427 49434 11679591 10.1074/jbc.M109325200
Shi G Trimmer JS Differential asparagine-linked glycosylation of voltage-gated K+ channels in mammalian brain and in transfected cells J Membr Biol 1999 168 265 273 10191360 10.1007/s002329900515
Manganas LN Trimmer JS Subunit composition determines Kv1 potassium channel surface expression J Biol Chem 2000 275 29685 29693 10896669 10.1074/jbc.M005010200
Folco E Mathur R Mori Y Buckett P Koren G A cellular model for long QT syndrome. Trapping of heteromultimeric complexes consisting of truncated Kv1.1 potassium channel polypeptides and native Kv1.4 and Kv1.5 channels in the endoplasmic reticulum J Biol Chem 1997 272 26505 26510 9334228 10.1074/jbc.272.42.26505
Babila T Moscucci A Wang H Weaver FE Koren G Assembly of mammalian voltage-gated potassium channels: evidence for an important role of the first transmembrane segment Neuron 1994 12 615 626 8155323 10.1016/0896-6273(94)90217-8
Rea R Spauschus A Eunson LH Hanna MG Kullmann DM Variable K(+) channel subunit dysfunction in inherited mutations of KCNA1 J Physiol 2002 538 5 23 11773313 10.1113/jphysiol.2001.013242
Bekele-Arcuri Z Matos MF Manganas L Strassle BW Monaghan MM Rhodes KJ Trimmer JS Generation and characterization of subtype-specific monoclonal antibodies to K+ channel alpha- and beta-subunit polypeptides Neuropharmacology 1996 35 851 865 8938716 10.1016/0028-3908(96)00128-1
Rosenberg RL East JE Cell-free expression of functional Shaker potassium channels Nature 1992 360 166 169 1436093 10.1038/360166a0
Koschak A Bugianesi RM Mitterdorfer J Kaczorowski GJ Garcia ML Knaus HG Subunit composition of brain voltage-gated potassium channels determined by hongotoxin-1, a novel peptide derived from Centruroides limbatus venom J Biol Chem 1998 273 2639 2644 9446567 10.1074/jbc.273.5.2639
McCormack K Connor JX Zhou L Ho LL Ganetzky B Chiu SY Messing A Genetic analysis of the mammalian K+ channel beta subunit Kvbeta 2 (Kcnab2) J Biol Chem 2002 277 13219 13228 11825900 10.1074/jbc.M111465200
Trouet D Nilius B Voets T Droogmans G Eggermont J Use of a bicistronic GFP-expression vector to characterise ion channels after transfection in mammalian cells Pflugers Arch 1997 434 632 638 9242728 10.1007/s004240050445
Manganas LN Wang Q Scannevin RH Antonucci DE Rhodes KJ Trimmer JS Identification of a trafficking determinant localized to the Kv1 potassium channel pore Proc Natl Acad Sci U S A 2001 98 14055 14059 11698661 10.1073/pnas.241403898
Ellgaard L Helenius A Quality control in the endoplasmic reticulum Nat Rev Mol Cell Biol 2003 4 181 191 12612637 10.1038/nrm1052
Arroyo EJ Xu YT Zhou L Messing A Peles E Chiu SY Scherer SS Myelinating Schwann cells determine the internodal localization of Kv1.1, Kv1.2, Kvbeta2, and Caspr J Neurocytol 1999 28 333 347 10739575 10.1023/A:1007009613484
Yu W Xu J Li M NAB domain is essential for the subunit assembly of both alpha-alpha and alpha-beta complexes of shaker-like potassium channels Neuron 1996 16 441 453 8789959 10.1016/S0896-6273(00)80062-8
Sewing S Roeper J Pongs O Kv beta 1 subunit binding specific for shaker-related potassium channel alpha subunits Neuron 1996 16 455 463 8789960 10.1016/S0896-6273(00)80063-X
Gu C Jan YN Jan LY A conserved domain in axonal targeting of Kv1 (Shaker) voltage-gated potassium channels Science 2003 301 646 649 12893943 10.1126/science.1086998
|
16305740
|
PMC1322225
|
CC BY
|
2021-01-04 16:03:49
|
no
|
BMC Neurosci. 2005 Nov 23; 6:65
|
utf-8
|
BMC Neurosci
| 2,005 |
10.1186/1471-2202-6-65
|
oa_comm
|
==== Front
BMC Evol BiolBMC Evolutionary Biology1471-2148BioMed Central London 1471-2148-5-661628866210.1186/1471-2148-5-66Research ArticleCoiled-coil protein composition of 22 proteomes – differences and common themes in subcellular infrastructure and traffic control Rose Annkatrin [email protected] Shannon J [email protected] Eric A [email protected] Iris [email protected] Department of Plant Cellular and Molecular Biology, Plant Biotechnology Center, Ohio State University, 1060 Carmack Road, Columbus, OH 43210, USA2 Ohio Super Computer Center, 1224 Kinnear Road, Columbus, OH 43212, USA2005 16 11 2005 5 66 66 9 6 2005 16 11 2005 Copyright © 2005 Rose et al; licensee BioMed Central Ltd.2005Rose 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
Long alpha-helical coiled-coil proteins are involved in diverse organizational and regulatory processes in eukaryotic cells. They provide cables and networks in the cyto- and nucleoskeleton, molecular scaffolds that organize membrane systems and tissues, motors, levers, rotating arms, and possibly springs. Mutations in long coiled-coil proteins have been implemented in a growing number of human diseases. Using the coiled-coil prediction program MultiCoil, we have previously identified all long coiled-coil proteins from the model plant Arabidopsis thaliana and have established a searchable Arabidopsis coiled-coil protein database.
Results
Here, we have identified all proteins with long coiled-coil domains from 21 additional fully sequenced genomes. Because regions predicted to form coiled-coils interfere with sequence homology determination, we have developed a sequence comparison and clustering strategy based on masking predicted coiled-coil domains. Comparing and grouping all long coiled-coil proteins from 22 genomes, the kingdom-specificity of coiled-coil protein families was determined. At the same time, a number of proteins with unknown function could be grouped with already characterized proteins from other organisms.
Conclusion
MultiCoil predicts proteins with extended coiled-coil domains (more than 250 amino acids) to be largely absent from bacterial genomes, but present in archaea and eukaryotes. The structural maintenance of chromosomes proteins and their relatives are the only long coiled-coil protein family clearly conserved throughout all kingdoms, indicating their ancient nature. Motor proteins, membrane tethering and vesicle transport proteins are the dominant eukaryote-specific long coiled-coil proteins, suggesting that coiled-coil proteins have gained functions in the increasingly complex processes of subcellular infrastructure maintenance and trafficking control of the eukaryotic cell.
==== Body
Background
The coiled-coil was one of the earliest protein structures described and first discovered in the two-stranded coiled-coil protein alpha-keratin [1]. Coiled-coils consist of two or more alpha-helices winding around each other in a supercoil, a simple yet versatile protein fold [2]. Mutations in coiled-coil proteins have been implicated in a large variety of human diseases such as severe skin fragility, muscular dystrophies, neurodegenerative diseases, progeria, and cancer [3-10]. Spurred by medical interest, the number of investigated long coiled-coil proteins in yeast and animals has rapidly grown in recent years. Recently, a database of all long coiled-coil proteins in the model plant Arabidopsis was established to facilitate the identification and characterization of long coiled-coil proteins in plants [11]. In contrast to eukaryotic organisms, only few long coiled-coil proteins have been characterized in prokaryotes. Examples include chaperonins and nucleases, secretion proteins, and cytadherence factors [12-15].
The foremost feature of coiled-coil domains appears to be their ability to act as "cellular velcro" to hold together molecules, subcellular structures, and even tissues. They can act as protein-protein interaction motifs, for examples as dimerization domains in transcription factors and receptor kinases [16-18]. They function as "zippers" in membrane fusion proteins [19], and as adapters between molecules and solid state cellular structures, such as in microtubule organizing centers, the nuclear pores and lamina, actin- and microtubule-associated proteins and cytoskeleton-associated E3 ubiquitin ligases [20-24]. Extracellular coiled-coil proteins include cell adherence factors and surface receptors, vertebrate blood components such as apolipoproteins and fibrinogen-like clotting factors, and extracellular matrix components such as laminins and cartilage matrix proteins forming tissue scaffolds in metazoa [25,26].
Besides associating with and interconnecting other molecules and macromolecular structures, long coiled-coil domains exhibit a number of structural and mechanical functions [27]. Typically, long coiled-coil domains form rod-like tertiary structures [2] and assemble to dynamic fibers, meshworks and scaffolds. Examples are the intermediate filaments of the cytoskeleton and nuclear lamina [28]. Recent evidence suggests an important role for the dynamic properties of cytoplasmic intermediate filaments in neurodegenerative diseases [29]. Other coiled-coils act as spacers, for example in the yeast spindle pole body where the distance between the plaques is determined by the length of the coiled-coil domain in the connecting proteins [30,31]. Membrane-bound coiled-coil proteins such as the spectrins and golgins form scaffolds for membrane structures within the cell [32,33]. In combination with other functional domains, coiled-coil domains are an integral part of molecular motors, such as the actin motor myosin and the microtubule motors kinesin and dynein [34]. Other coiled-coil proteins with ATPase and GTPase domains often function in folding and repair, e.g. as chaperonins in protein folding, and topoisomerases and nucleases in DNA remodeling [35-37].
On a primary structure level, amino acid sequences with the capacity to form left-handed alpha-helical coiled-coils are characterized by a heptad repeat pattern in which residues in the first and fourth position are hydrophobic, and residues in the fifth and seventh position are predominantly charged or polar [38]. This pattern of hydrophobic and polar residues interferes with sequence comparison algorithms, which often lead to false predictions of homology between long coiled-coil proteins based on the low complexity and repeat nature of the underlying sequence motif. On the other hand, this repeat pattern can also be used to predict coiled-coil domains in amino acid sequences by computational means [39-42].
In the post-genomics era, such structure-prediction algorithms can now be applied to whole proteomes. Based on the prediction algorithm COILS, roughly 10% of all proteins encoded by eukaryotic genomes contain coiled-coil domains whereas prokaryotic genomes contain only 4–5% [43]. Using the MultiCoil program, one in every 11 proteins in yeast was predicted to contain a coiled-coil sequence [44]. However, these studies did not use a cut-off for domain length to determine coiled-coils. A minimum length of three to four heptad repeats is required for the formation of a stable coiled-coil using synthetic peptides [45-47]. Using this minimum domain length of 20 amino acids (or about three heptad repeats), 5.6% of the predicted ORFs in the fully sequenced Arabidopsis genome were found to encode coiled-coil proteins [11].
In a comparative genomics approach, we determined the coiled-coil content of 22 predicted whole proteomes using the prediction pipeline and processing software developed to create the ARABI-COIL database [11]. The 22 genomes analyzed included four archaeal genomes, ten bacterial genomes (three gram-positive and seven gram-negative species), and eight eukaryotic genomes (two each for yeasts, invertebrates, mammals, and plants).
Results
Prediction and selection of coiled-coil proteins was performed using the MultiCoil algorithm [42] and the ExtractProp processing software [11]. For the purpose of this study, "long coiled-coil" proteins were defined according to the parameters used to establish the ARABI-COIL database and included all sequences with at least one coiled-coil domain and minimum domain length of 70, two domains and minimum domain length of 50, and three or more domains and minimum domain length of 30 [11].
Eukaryotic genomes contain higher percentages of long coiled-coil proteins than prokaryotic genomes
Proteins predicted to form coiled-coil domains were present in all genomes analyzed (Table 1, Figure 1) and comprised between 2% and 8% of the total proteomes. The most pronounced difference between prokaryotic and eukaryotic genomes was in the percentage of genes per genome predicted to encode long or multiple coiled-coil domains. With increasing coiled-coil domain length cut-off, lower percentages of proteins were identified in bacterial genomes. With the exception of Bacillus subtilis, MultiCoil predicted no coiled-coil proteins with domains longer than 250 amino acids in the bacterial genomes analyzed. However, archaeal and eukaryotic genomes contain proteins predicted to form coiled-coils of this length. Strikingly, prediction of coiled-coil domains over 400 amino acids in length was completely absent in bacterial genomes, but present in eukaryotes as well as two archaea, Sulfolobus solfataricus and Archeoglobus fulgidus. These numbers however do not take discontinuous coiled-coil prediction into account, as evident in the case of prokaryotic SMC proteins (Figure 2).
Table 1 Proteome sequence data sets downloaded for MultiCoil analysis
Organism Number of proteins Date of download Comments (info as provided by EBI and TIGR)
Archaea (extremophiles):
Archaeoglobus fulgidus (A.f.) 2400 10-Jun-2004 hyperthermophilic, organoheterotrophic-lithoautotrophic, sulfur-metabolizing; glycoprotein envelope, flagellated
Methanococcus jannaschii (M.j.) 1782 10-Jun-2004 thermophilic, methanogenic, autotrophic, strict anaerobic; grows under high pressure in deep sea, flagellated
Sulfolobus solfataricus (S.s.) 2939 28-Jan-2004 thermophilic, sulfuric acid-producing, aerobic; no flagella, but pilus-like and pseudopodium-like structures
Thermoplasma acidophilum (T.a.) 1479 28-Jan-2004 thermoacidophilic; flagellated, no cell wall
Gram-positive bacteria:
Actinobacteria: Mycobacterium tuberculosis (M.t.) 3995 10-Jun-2004 animal pathogen (tuberculosis); no flagella
Bacilli: Bacillus subtilis (B.s.) 4167 28-Jan-2004 capable of producing endospores, flagellated
Mollicutes: Mycoplasma genitalium (M.g.) 486 10-Jun-2004 animal pathogen (surface parasite), smallest known self-replication cell & genome; no flagella
Gram-negative bacteria:
Alphaproteobacteria: Agrobacterium tumefaciens (A.tu.) 5393 10-Jun-2004 plant pathogen (crown gall); flagellated
Betaproteobacteria: Chromobacterium violaceum (C.v.) 4400 10-Jun-2004 subtropical/tropical; produces antimicrobial violacein, flagellated
Gammaproteobacteria: Escherichia coli K12 (E.c.) 4356 28-Jan-2004 enterobacterium, laboratory strain, flagellated
Epsilonproteobacteria: Heliobacter pylori (H.p.) 1556 10-Jun-2004 animal pathogen; micro-aerophilic, spiral-shaped, flagellated
Chlamydiae: Chlamydia pneumoniae (C.p.) 1110 10-Jun-2004 animal pathogen (obligate intracellular parasite), no flagella
Spirochaetes: Borrelia burgdorferi (B.b.) 1558 10-Jun-2004 animal pathogen (Lime disease); spiral-shaped, flagellated
Cyanobacteria: Synechocystis sp. PCC6803 (S.sp.) 3164 28-Jan-2004 photosynthetic (oxygenic), no flagella
Yeast:
Saccharomyces cerevisiae (S.c.) 6191 28-Jan-2004 baker's yeast
Schizosaccaromyces pombe (S.p.) 5037 28-Jan-2004 fission yeast
Metazoa:
Caenorhabditis elegans (C.e.) 22873 28-Jan-2004 nematode
Drosophila melanogaster (D.m.) 16196 28-Jan-2004 insect (fruitfly)
Mus musculus (M.m.) 27577 28-Jan-2004 mammal (mouse)
Homo sapiens (H.s.) 29024 28-Jan-2004 mammal (human)
Plants:
Arabidopsis thaliana (A.t.) 26945 28-Jan-2004 dicot
Oryza sativa ssp. Japonica (O.s.) 56056 9-Jan-2004 monocot (rice)
Proteome sequence sets were downloaded from the European Bioinformatics Institute (EBI) [106] or The Institute for Genome Research (TIGR) [107]. The number of protein sequence entries reflects the annotation of ORFs at the time of download.
Figure 1 Percentages of long coiled-coil proteins per genome. CC, coiled-coil length in amino acids, "CC total" includes all sequences predicted to contain a minimum stretch of 20 amino acids predicted to form a coiled-coil, "Long CC" includes all sequences with at least one coiled-coil domain and minimum domain length of 70, two domains and minimum domain length of 50, and three or more domains and minimum domain length of 30. A, archaea; B, Gram+ bacteria; C, Gram- bacteria; D, yeasts; E, metazoa; F, plants. 1, Thermoplasma acidophilum; 2, Methanococcus jannaschii; 3, Archaeoglobus fulgidus; 4, Sulfolobus solfataricus; 5, Mycoplasma genitalium; 6, Mycobacterium tuberculosis; 7, Bacillus subtilis; 8, Clamydia pneumoniae; 9, Heliobacter pylori; 10, Borrelia burgdorferi; 11, Synechocystis sp. PCC6803; 12, Escherichia coli; 13, Chromobacterium violaceum; 14, Agrobacterium tumefaciens; 15, Schizosaccharomyces pombe; 16, Saccharomyces cerevisiae; 17, Drosophila melanogaster; 18, Caenorhabditis elegans; 19, Mus musculus; 20, Homo sapiens; 21, Arabidopsis thaliana; 22, Oryza sativa.
Figure 2 ABC-ATPases in archaea and bacteria. Phylogenetic tree and schematic representation of domain structures of ABC-ATPases and related sequences found in the prokaryotic genomes analyzed. Conserved domains shown as identified in CDD [49]. aa, amino acids. For species name abbreviations, see Table 1.
Prokaryotic long coiled-coil proteins
Archaea
Four archaeal genomes were included in this study and tables with coiled-coil protein details are available in additional file 1 (Archeoglobus fulgidus, Table S1; Methanococcus jannaschii, Table S2; Sulfolobus solfataricus, Table S3; and Thermoplasma acidophilum, Table S4). 2–3% of the genes in these archaea were found to code for coiled-coil proteins. In contrast to eubacteria, all of the coiled-coil size-classes analyzed are represented in this group, with proteins predicted to form coiled-coils longer than 400 residues present in Methanococcus jannaschii and Archeoglobus fulgidus proteomes (see Figure 1).
Eubacteria
Bacterial genomes for this study were chosen from different families to represent a wide range of prokaryotic species. Three gram-positive bacterial genomes (additional file 1; Mycobacterium tuberculosis, Table S5; Bacillus subtilis, Table S6; and Mycoplasma genitalium, Table S7), and seven gram-negative bacterial genomes (Agrobacterium tumefaciens, Table S8; Chromobacterium violaceum, Table S9; Escherichia coli, Table S10; Heliobacter pylori, Table S11; Chlamydia pneumoniae, Table S12; Borrelia burgdorferi, Table S13; and the cyanobacterium Synechocystis, Table S14) were analyzed.
The largest prokaryotic coiled-coil domains were identified in proteins of the SMC, Rad50, SbcC and MukB families. These proteins contain globular head and tail domains separated by a coiled-coil rod with a hinge [48]. Figure 2 summarizes schematic diagrams of the domain structures of the prokaryotic SMC and SMC-like proteins identified in this study based on our coiled-coil prediction data and conserved domains as identified through Conserved Domain Database (CDD) searches [49]. Figure 3 shows a summary of additional long coiled-coil proteins with domains of at least 150 amino acids in length present in prokaryotic genomes. A number of these proteins are involved in membrane events, such as chemosensing via methyl-accepting chemotaxis proteins [50] and membrane fusion and vesicle formation mediated by AcrA, TolA, and incA proteins [51-53]. Others function as adhesion proteins, for example the lambda phage side tail fiber protein [54] and the hmw2 protein of the attachment organelle of Mycoplasma pneumoniae [15], or as enzymes of the cell wall such as the NlpC/P60 proteins [55].
Figure 3 Prokaryotic long coiled-coil proteins. Schematic representation of prokaryotic long coiled-coil proteins not belonging to the ABC-ATPase family. Only proteins with at least 150 amino acids predicted to be in a coiled-coil are shown. Blue, coiled-coil domain; green, signal peptide; yellow, transmembrane domain. Functional domains as identified in the CDD [49] are circled in red. tlpC-1, tlpC-2, methyl-accepting chemotaxis proteins homologous to B.s. tlpC [111]; hmw2, cytadherence protein [15], CHLPS incA, incA, inclusion membrane proteins [53]; TolA, [52]; OspD, outer surface protein D [112], [113]. For species name abbreviations, see Table 1.
Long coiled-coil domains cause clustering of unrelated coiled-coil sequences
Sequences predicted to form long coiled-coil domains were analyzed for family relationships and conservation across species in an all-against-all approach using the Smith-Waterman sequence comparison algorithm followed by clustering based on an adaptation of Kruskal's minimum cost spanning tree algorithm [56,57].
In a pilot analysis to test the feasibility of the clustering approach, all prokaryotic sequences meeting the aforementioned criteria for "long coiled-coil" proteins were included in the clustering. Due to the larger number of qualified sequences in the eukaryotic species, only the longest domains (at least 250 residues in length) or sequences largely covered by coiled-coil (at least 60% of the sequence) were included in the combined pilot sequence set comprising 527 unique sequences. A maximum P-score of 1.0e-20 was used as the critical threshold when selecting only the most prominent sequence similarities in this test group. In all, 12,013 pair-wise P-score values were selected, defining as many unique relationships from the 277,729 possible pair-wise relationships. Sequences were then grouped using Kruskal's minimum cost spanning tree algorithm using the P-score value as the edge weight for the selected P-score values. 166 independent non-overlapping sequence subsets (subtrees) were defined in this manner. The largest grouping consisted of 270 sequences, representing over half of the sequences in the pilot sequence set and including functionally distinct families such as for example myosins, golgins, and SMC proteins. Distinct clusters of long coiled-coil proteins besides this large, heterogeneous group were formed by the animal and yeast tropomyosins (two separate clusters), the laminins, the CASP/CDP-family and the nuclear lamins.
Masking of coiled-coil domains before clustering
To prevent clustering based on the inherent coiled-coil repeat similarities, amino acids predicted to form coiled-coil domains were computationally masked out before being subjected to sequence similarity comparison (Figure 4). The clustering of the sequences with masked coiled-coil domains yielded a much more accurate grouping of known long coiled-coil protein families such as the myosins, golgins, and SMC proteins (Table 2). The largest group of long coiled-coil proteins with 58 sequences comprised the myosin motor proteins. The laminins, CASP/CDP, and nuclear lamins still exhibited the prior cluster profile, however the tropomyosin clusters did not appear after masking the coiled-coil domains. The coiled-coil coverage for many of the tropomyosins was predicted as 100% in our analysis, effectively excluding this protein family from the sequence comparison after masking.
Figure 4 Flowchart of sequence comparison and clustering. Coiled-coil prediction data was generated using the program MultiCoil [42] and output processing and coiled-coil domain selection were performed as described for the ARABI-COIL database [11]. Coiled-coil prediction data was used to generate a set of sequences with coiled-coil domains masked out. The masked sequences were used as a query against unmasked sequences in an all-against-all Smith-Waterman sequence comparison (SW Search). The P-scores from this comparison were used for clustering of the output.
Table 2 Clustering results
Annotation # sequences Species
Myosins 56 A.t., C.e., D.m., H.s., M.m., O.s., S.c., S.p.
SMCs 13 A.f., A.tu., B.b., B.s., C.v., H.s., M.g., M.j., M.m., M.t., S.c., S.sp., T.a.
Laminins 10 D.m., H.s., M.m.
ROCK 9 C.e., D.m., H.s., M.m.
ELKS/ERC1 7 H.s., M.m.
SLAP 5 H.s., M.m.
Kinectin 5 H.s., M.m.
Periplakin 5 H.s., M.m.
DOC1/FILIP 5 H.s.
C-Nap 5 D.m., H.s., M.m.
CASP/CDP 4 H.s., M.m.
CENP-F 4 H.s., M.m.
Lamins 4 H.s., M.m.
Hypothetical 4 C.e.
Unknown 4 A.t., O.s., S.p.
Unknown 4 A.t., O.s.
Largest clusters identified in a pilot analysis using all prokaryotic long coiled-coil proteins and eukaryotic proteins with coiled-coil domains longer than 250 amino acids or more than 60% coverage. Only clusters with four or more members are listed. ROCK, Rho-associated coiled-coil containing kinase [114]; ELKS/ERC1, Rab6-interacting protein [115], SLAP, sarcolemmal-associated protein [116]; DOC1, downregulated in ovarian cancer 1 [117]; CENP-F, centromer protein F [118]. For species name abbreviations, see Table 1.
Clustering analysis with masked coiled-coil domains
After determining the consistency of clusters formed after masking coiled-coil domains with well-known coiled-coil protein families such as the SMC proteins, myosins and kinesins, we proceeded to cluster all 3576 predicted long coiled-coil sequences from the 22 genomes. The clustering algorithm was further improved to first preclude transitively similar sequences by requiring all sequences in each cluster to satisfy the P-score threshold for all pair-wise relationships within the cluster and secondly to identify "bridge" sequences meeting these criteria for multiple clusters (see Material and Methods for details). A P-score threshold of 10e-06 was selected as the appropriate balance of sequence coverage and cluster discrimination. Table 3 gives an overview of the sequences from each species contributing to the clustering analysis using the 1.0e-06 P-score cut-off. The high number of species-specific sequences found in rice is caused by retrotransposon repeats in the rice genome containing predicted coiled-coil domains within a putative transposase ORF. Figure 5 shows the distribution of clusters among the different kingdoms. Sequence annotation including species origin provided further insight into functions and relationships among sequences in each cluster. Additional information was obtained using Conserved Domain Database searches, multiple sequence alignments, and phylogenetic tree analysis of selected clusters (see Materials and Methods).
Table 3 Contribution to clusters
species ORFs total CCs total long CCs species-specific in cross-species clusters
Archaea
T.a. 1479 29 6 1 5
M.j. 1782 54 6 2 4
A.f. 2400 65 14 6 8
S.s. 2939 77 10 8 2
gram + bacteria
M.g. 486 21 5 3 2
M.t. 3995 70 10 7 3
B.s. 4167 144 17 10 7
gram - bacteria
C.p. 1110 43 6 5 1
H.p. 1556 87 9 6 3
B.b. 1558 76 14 13 1
S. sp. 3164 133 25 16 9
E.c. 4356 111 16 6 10
C.v. 4400 161 19 9 10
A.tu. 5393 161 18 10 8
yeast
S.p. 5037 303 62 25 37
S.c. 6191 344 73 35 38
plants
A.t. 26945 1518 284 59 225
O.s. 56056 3740 997 795 202
animals
D.m. 16196 1174 317 117 200
C.e. 22873 1234 304 144 160
M.m. 27577 1709 512 56 456
H.s. 29024 2400 855 189 666
Contribution of genomes to cross-species clusters (based on clustering using a P-score cut-off of 1.0e-06). For species name abbreviations, see Table 1.
Figure 5 Cluster distribution. Clustering after Smith-Waterman comparison of sequences with coiled-coil domains masked. Numbers within the circles and overlapping sections represent numbers of clusters containing sequences from the respective kingdoms. For kingdom-specific clusters, only clusters with sequences from at least two different species were counted.
Coiled-coil proteins conserved between prokaryotes and eukaryotes
The SMC proteins were identified as the single major cluster of long coiled-coil proteins containing sequences from eukaryotic as well as prokaryotic genomes (see Table 4). Another group of conserved proteins with long coiled-coils comprised a number of eukaryotic Ser/Thr-kinases and a homolog from the cyanobacterium Synechocystis (sll0776 in Figure S1, additional file 2). However, proteins belonging to this cluster could not be found in any other prokaryotic genome.
Table 4 Clusters with sequences from prokaryotes and eukaryotes
Cluster size
(# of sequences) Max. edge
(P-score) protein family putative function/site of action organisms represented
45 4.6E-7 Structural maintenance of chromosomes 1–4 condensin, cohesin (chromatin) A.f., A.t., A.tu., B.b., B.s., C.e., C.v., D.m., H.s., M.g., M.j., M.m., M.t., O.s., S.c., S.p, S.sp., T.a.
26 1.6E-10 Ser/Thr-kinases (DAP, DMK, GIN4, ROCK) signal transduction C.e., D.m., H.s., M.m., O.s., S.c., S.sp.
Numbers include "bridge" sequences qualifying for more than one cluster. Only clusters represented by at least 10 sequences are listed. DAP, death-associated protein kinase [119]; DMK, myotonic dystrophy kinase [120]; GIN4, growth inhibitory gene 4 [121]; ROCK, Rho-associated coiled-coil containing kinase [114]. For species name abbreviations, see Table 1.
A number of smaller cluster were formed containing proteins with shorter coiled-coil domains close to the cut-off for our analysis. One cluster comprised the translation initiation factor IF-2, containing the respective sequences from Drosophila, E. coli, mouse, rice and yeast. Another cluster with sequences conserved in prokaryotes as well as eukaryotes contained the AAA+ family ATPase ClpB/Hsp104 represented by plant, yeast and bacterial sequences. This protein functions as a protease/chaperonin in eubacteria, plants and mitochondria [35]. Two small clusters combined sequences from prokaryotes and plant genomes. One cluster comprised mitochondrial seryl-tRNA synthetases conserved in plant mitochondria as well as archaea while the second cluster comprised the PspA-like VIPP1 protein from plastids and the cyanobacterium Synechocystis. VIPP1 is involved in thylakoid biosynthesis in both chloroplasts as well as cyanobacteria, possibly acting in thylakoid membrane trafficking [58,59].
Prokaryotic coiled-coil protein clusters
Prokaryotic clusters comprised membrane-bound proteins and signal transducers, as well as membrane-spanning transporters and secretion proteins such as the HlyD family [60]. The only cluster specific to prokaryotes represented by more than ten sequences in this study comprised the methyl-accepting chemotaxis proteins (MCPs; Table 5; [50]). Smaller prokaryotic clusters contained the aforementioned ABC-ATPases RAD50 and SbcC involved in DNA repair and a highly conserved group of archaeal proteins of unknown function (COG1340, represented by NP_394939 in Figure 3).
Table 5 Prokaryotic clusters
Cluster size
(# of sequences) Max. edge
(P-score) protein family putative function/site of action organism
12 <E-40 Methyl-accepting chemotaxis proteins chemotactic sensor/signal transducer (bacterial envelope membrane) A.f., A.tu., B.s., C.v., E.c., S.sp.
Numbers include "bridge" sequences qualifying for more than one cluster. Only clusters represented by at least 10 sequences are listed. For species name abbreviations, see Table 1.
Eukaryotic coiled-coil protein clusters
The main clusters formed by eukaryotic sequences only (Table 6) were the eukaryotic motor proteins: the actin motor myosin and the microtubule motor kinesin and the related kinesin-like calmodulin-binding protein KCBP [34,61,62]. The proteins of the SMC5 and SMC6 families formed a eukaryotic cluster instead of clustering together with the condensin/cohesin SMCs 1–4 and the prokaryotic SMC proteins in our analysis (Figure 6B). Eukaryotic RAD50 proteins clustered separately from prokaryotic RAD50s as well, indicating a higher convergence of the non-coiled-coil RAD50 ATPase domains as compared to the SMC 1–4 head and tail domains. Additional larger clusters included eukaryotic Ser/Thr-kinases and a family comprised of the Retinoblastoma-associated protein RBP95, Ring Finger Proteins 20 and 40, and yeast Bre1p [63,64,23] (Figure S2, additional file 2, and Table S15, additional file 3). Formin-related proteins associated with growing actin fibers [65,66] were found in animal/yeast and animal/plant cluster combinations. Smaller conserved eukaryotic clusters included a number of proteins involved in vesicle transport, such as a Rab6 GTPase-activating protein involved in retrograde transport [67], the golgin CASP [68] and the vesicular transport proteins P115 (see Figure S3, additional file 2), autophagy protein APG6 [69,70], and early endosome antigen (EEA1, [71]) homologs (see Figure S4, additional file 2).
Table 6 Eukaryotic clusters
Cluster size
(# of sequences) Max. edge
(P-score) protein family putative function/site of action organisms represented
94 5.4E-10 Myosin heavy chain actin motor protein (muscle, cytoskeleton) A.t., C.e., D.m., H.s., M.m., O.s., S.c., S.p.
27 1.5E-36 Kinesin heavy chain (KCBP, KIFC) MT motor protein (cytoskeleton) A.t., C.e., D.m., H.s., M.m., O.s., S.c., S.p.
21 1.1E-35 Kinesin heavy chain (Cmet/Cana, MKRPs, NACK/HINKEL) MT motor protein (cytoskeleton) A.t., D.m., H.s., M.m., O.s., S.p.
17 1.2E-7 Structural maintenance of chromosomes 5–6/RAD18 DNA repair (chromatin) A.t., C.e., D.m., H.s., M.m., O.s., S.c., S.p.
12 5.0E-7 Kinases (GIN4, MET) signal transduction A.t., D.m., H.s., O.s., S.c.
11 <E-40 RAD50 (eukaryotic) DNA repair A.t., C.e., D.m., H.s., M.m., O.s., S.c., S.p.
11 1.2E-7 Retinoblastoma-associated protein, RING finger protein 20 E3 Ubi. ligase for H2B histone modification (nuclear)? A.t., C.e., D.m., H.s., M.m., O.s., S.c, S.p.
Numbers include "bridge" sequences qualifying for more than one cluster. Only clusters represented by at least 10 sequences are listed. MKRPs, mitochondrial kinesin-related proteins [122]; NACK/HINKEL, NPK1-activating kinesin [123]; MET, [124]. For species name abbreviations, see Table 1.
Figure 6 Contribution to clusters. Contribution of the different kingdoms to the complete sequence pool analyzed (A) and to SMC, myosin and kinesin clusters (B). Y-axis, number of sequences; X-axis, cluster IDs. Examples for characteristic protein families represented in clusters are as follows: clusters 99, 172, 717, SMC 1–4; cluster 165, SMC 5–6; clusters 39, 1220, type II myosins; cluster 125, type X myosins; cluster 1223, non-muscle myosins; clusters 1, 272, KIFCs; clusters 157, 220, PAKRPs; cluster 244, NACK, MKRPs. Proteins may qualify for two or more overlapping cluster, e.g. resulting in prokaryotic SMCs clustering with different types of diverged eukaryotic SMC proteins.
Yeast, yeast-plant, and yeast-animal coiled-coil protein clusters
Eukaryotic genomes included the baker's yeast (Saccharomyces cerevisiae) and fission yeast (Schizosaccharomyces pombe) as eukaryotic, unicellular organisms. Protein clusters found to be specific for yeast were typically small (one sequence from each yeast genome, see additional file 4) and comprised proteins involved in RNA export, such as Gle1, [72] and Mlp1 [73], the spindle assembly checkpoint protein Mad1 [74], and GRIP-domain golgins [75,76]. These proteins have known homologs in other eukaryotic proteomes, which did not cluster together with the yeast proteins, likely due to a high overall coverage with coiled-coil sequences (e.g. up to 70% coiled-coil coverage for Mlp1/Tpr, up to 74% for MAD1, and up to 75% for GRIP-golgins). Another functional group of yeast proteins were cell polarity proteins such as Spa2 and Tea1 [77,78]. Tea1 clustered together with a number of plant sequences of unknown function containing Kelch repeats [79] in combination with coiled-coil domains. Proteins that were found in clusters specific to yeasts and animals (Table 7) included the microtubule motor dynein as well as proteins involved in endocytosis and microtubule dynamics, such as intersectin, restin and cytoplasmic linker proteins (CLIP) [80]. A number of myosin subclusters, for example myosin type II, was represented only by yeast and animal but not plant sequences, consistent with previous findings [81] (see Table 7 and Figure 6B).
Table 7 Animal and yeast clusters
Cluster size
(# of sequences) Max. edge
(P-score) protein family putative function/site of action organism
57 3.3E-7 Myosin heavy chain XVIII non-muscle and smooth muscle myosins C.e., D.m., H.s., M.m., S.p.
50 6.5E-8 Myosin heavy chain type 2 actin motor protein (cytoskeleton) C.e., D.m., H.s., M.m., S.c., S.p.
27 <E-40 Dynein heavy chain MT motor (cytoskeleton, flagella) C.e., D.m., H.s., M.m., S.c.
17 3.1E-9 Intersectins, Eps15 endocytosis C.e., D.m., H.s., M.m., S.c., S.p.
16 9.5E-7 Png-1, IF-2, Neurofilament triplet L, Troponin T D.m., H.s., M.m., O.s., S.c.
14 <E-40 Restin, Dynactin, CLIP proteins linking endocytic vesicles to MTs (IF cytoskeleton), dynein activator (MTs in neurons), MT/IF associated (cytoskeleton) C.e., D.m., H.s., M.m., S.c., S.p.
14 1.6E-15 Myosin heavy chain V unusual myosin C.e., D.m., H.s., M.m., S.c.
14 <E-40 DRFs binds Rho-GTP and profilin, promotes actin polymerization (membrane cytoskeleton) D.m., H.s., M.m., S.c.
Numbers include "bridge" sequences qualifying for more than one cluster. Only clusters represented by at least 10 sequences are listed. CLIP, cytoplasmic linker protein [80]; DRFs, diaphanous-related formins [66]; Eps15, epidermal growth factor receptor substrate 15 [125]; Png-1, postmitotic neural gene-1 [126]. For species name abbreviations, see Table 1.
Animal coiled-coil protein clusters
From the metazoan kingdom, genomes from nematodes (Caenorhabditis elegans), flies (Drosophila melanogaster), and mammals (Mus musculus and Homo sapiens) were analyzed. Clusters that appeared to be specific to animals (Table 8) comprised a variety of proteins crosslinking cytoskeletal components with membranes, such as spectrin- and periplakin-like membrane-actin and membrane-IF crosslinkers [32,82], the plasmamembrane-scaffolding Liprins [83], the family of Merlin and Ezrin/Radixin/Moesin (ERM) proteins [84,85], and a number of Golgi- and vesicle-associated proteins. Other groups comprised centrosome-associated and mitotic spindle checkpoint proteins. Type X myosins grouped together in a metazoan cluster without plant or yeast sequences. Another animal-specific group contained coiled-coil proteins involved in structural integrity such as the extracellular scaffolding protein Laminin [26] and intermediate filament proteins including the nuclear lamins and neurofilaments [86,87]. Smaller animal-specific clusters contained protein sequences involved in cell attachment and motility, embryogenesis, spermatogenesis, and immune cell movement.
Table 8 Animal-specific clusters
Cluster size
(# of sequences) Max. edge
(P-score) protein family putative function/site of action organism
39 9.9E-7 Spectrins, Dystrophin, Nesprins membrane/actin/MT crosslinkers (cytoskeleton) C.e., D.m., H.s., M.m.
32 <E-40 Laminins scaffold protein (extracellular matrix) C.e., D.m., H.s., M.m.
25 6.7E-8 Plectin/Desmoplakin actin/MT crosslinkers (cytoskeleton) C.e., D.m., H.s., M.m.
22 6.5E-8 Myosin heavy chain (muscle) actin motor protein, muscle fibers C.e., D.m., H.s., M.m.
18 9.3E-9 Lamins nuclear IFs C.e., D.m., H.s., M.m.
17 5.8E-7 Neurofilament triplet L, M, Death inducer, Troponin T C.e., D.m., M.m., H.s.
14 7.5E-7 Neurofilament triplet H, M C.e., D.m., H.s., M.m.
14 7.5E-7 Neurofilaments (Desmin, Vimentin) IFs C.e., H.s., M.m.
13 2.6E-7 PP1, ASPP apoptosis stimulating D.m., H.s., M.m.
12 <E-40 Moesin, Ezrin, Radixin membrane organization and stabilization (membrane cytoskeleton, cytovilli) C.e., D.m, H.s., M.m.
11 1.7E-43 RUFY possible role in vesicle trafficking (endosomes?) D.m., H.s., M.m.
11 2.2E-11 Lamins nuclear IFs D.m., H.s., M.m.
11 3.4E-7 Restin, Dynactin, CLIP proteins linking endocytic vesicles to MTs (IF cytoskeleton), dynein activator (MTs in neurons), MT/IF associated (cytoskeleton) C.e., D.m., H.s., M.m.
11 3.9E-7 Png-1, Neurofilament triplet M, Troponin T C.e., D.m., H.s., M.m.
10 <E-40 Dystrophins C.e., D.m., H.s., M.m.
10 2.7E-8 prion-like protein C.e., D.m.
Numbers include "bridge" sequences qualifying for more than one cluster. Only clusters represented by at least 10 sequences are listed. ASPP, apoptosis stimulating of p53 protein [127]; PP1, protein phosphatase 1 [128]; RUFY, RUN and FYVE domain containing proteins [129]. For species name abbreviations, see Table 1.
A number of the clusters containing animal sequences were limited to mammalian sequences only (Table 9). The hair fiber protein keratin was found to form the largest group of proteins specific to mammals. Other mammlian clusters comprised neurofilament proteins and crosslinkers of the actin cytoskeleton and were found to overlap with clusters containing invertebrate sequences as well. A number of smaller mammalian clusters (see additional file 5, Table S17) contained sequences of unknown function which have so far only been characterized as autoantigens or cancer antigens. Smaller clusters included the centrosomal protein Ninein, which is involved in anchoring microtubule minus ends [88], and a number of other centrosomal proteins including TACCs, C-NAP1, and Centriolin [89-91]. Other clusters included mammalian reproductive organ-specific proteins, such as sperm tail-associated proteins and mammary gland-specific proteins, vertebrate-specific transcription factors and coactivators such as the SOX proteins [92], and regulators of endothelial cell motility and clotting factors in blood vessels.
Table 9 Clusters with mammalian sequences only
Cluster size
(# of sequences) Max. edge
(P-score) protein family putative function/site of action organism
51 <E-40 Keratin type II IF, cytoskeletal (cytokeratin), hair H.s., M.m.
45 5.0E-7 Keratin type I (hair keratin) IF, cuticular/hair H.s., M.m.
32 4.2E-7 Keratin type I IF, cytoskeletal (cytokeratin), root sheeth H.s., M.m.
31 9.2E-7 Keratin type II IF, cytoskeletal (cytokeratin), hair H.s., M.m.
24 3.9E-7 Keratin type I (hair keratin) IF, cuticular/hair H.s., M.m.
18 1.4E-14 Neurofilaments (Desmin, Internexin, Peripherin, Vimentin) IFs H.s., M.m.
13 1.2E-25 Interferon-induced guanylate-binding proteins H.s., M.m.
10 1.9E-8 Plectin, Desmoplakin, Periplakin, Envoplakin H.s., M.m.
Numbers include "bridge" sequences qualifying for more than one cluster. Only clusters represented by at least 10 sequences are listed. See Table S16 for protein details on clusters smaller than 10. For species name abbreviations, see Table 1.
Plant coiled-coil protein clusters
As representatives for the plant kingdom, a dicot (Arabidopsis thaliana) and a monocot (Oryza sativa) plant genome were analyzed. Clusters of long coiled-coil proteins specific to Arabidopsis and rice contained mostly sequences of so far unknown function (Table 10). The rice genome contains a large number of transposon-derived ORFs which are predicted to contain coiled-coil domains, therefore a large number of plant-specific clusters was represented by rice sequences only. These have been omitted from Table 10. Plant-specific clusters represented by both plant species analyzed included kinase interacting protein 1 (KIP1) and its relatives [93], the family of filament-like plant proteins, FPPs [94], and a cluster of putative Zinc finger transcription factors homologous to the x1 gene of maize [95]. Smaller clusters (see additional file 6, Table S18) included nuclear matrix constituent protein 1 (NMPC1) and relatives [96], and the chloroplast unusual positioning 1 (CHUP1) actin-interacting protein [97]. Several clusters showed overlap between the plant and animal kingdoms (Table 11). These included a number of kinesin subclusters, vesicle trafficking proteins, and Guanylate-binding proteins (Figure S5).
Table 10 Plant-specific clusters
Cluster size
(# of sequences) Max. edge
(P-score) protein family putative function/site of action organism
21 1.2E-7 Kinase-interacting protein 1-like signal transduction A.t., O.s.
13 7.6E-12 expressed proteins unknown A.t., O.s.
12 2.4E-9 FPPs unknown A.t., O.s.
11 2.9E-7 putative receptor kinases signal transduction A.t., O.s.
10 5.9E-18 Transcription factor X1-like proteins transcription A.t., O.s.
Numbers include "bridge" sequences qualifying for more than one cluster. Only clusters represented by at least 10 sequences are listed. See Table S17 for protein details on clusters smaller than 10. FPPs, filament-like plant proteins [94]. For species name abbreviations, see Table 1.
Table 11 Animal and plant clusters
Cluster size
(# of sequences) Max. edge
(P-score) protein family putative function/site of action organism
83 1.6E-38 Kinesin heavy chain (Chromokinesin, KIF3, 4) MT motor protein (cytoskeleton), nuclear A.t., C.e., D.m., H.s., M.m., O.s.
47 4.6E-39 Kinesin heavy chain (KIF2-4, NACK, FRA) MT motor protein (cytoskeleton) A.t., C.e., D.m., H.s., M.m., O.s.
29 3.0E-32 Kinesin heavy chain (KIF1, 13, 14, 16, 17) MT motor protein (cytoskeleton), axonal transporter of synaptic vesicles A.t., C.e., D.m., H.s., M.m., O.s.
16 5.9E-38 Kinesin heavy chain (PAKRP) MT motor protein (cytoskeleton) A.t., H.s., O.s.
14 2.1E-9 Plexin, Rab6 GTPase activating protein vesicle trafficking A.t., C.e., D.m., H.s., M.m., O.s.
13 9.4E-7 Guanylate-binding protein A.t., H.s., M.m., O.s.
12 4.3E-27 Kinesin heavy chain (KIFC1, TH65) MT motor protein (cytoskeleton) A.t., H.s., O.s.
12 9.6E-35 Kinesin heavy chain (PAKRP, MKRP) MT motor protein (cytoskeleton) A.t., H.s., O.s.
11 3.4E-7 DRFs binds Rho-GTP and profilin, promotes actin polymerization (membrane cytoskeleton) A.t., C.e., H.s., M.m., O.s.
Numbers include "bridge" sequences qualifying for more than one cluster. Only clusters represented by at least 10 sequences are listed. FRA, fragile fiber [130]; PAKRP, phragmoplast-associated kinesin-related protein [131], [132]. For species name abbreviations, see Table 1.
Discussion
The SMC proteins are the most widely conserved coiled-coil proteins
The most widely conserved family of long coiled-coil proteins found in our study comprised the SMC proteins. Representatives from almost all species analyzed were found in this cluster, with a few exceptions such as the gram-negative bacterium E. coli. This is consistent with previous findings that SMC proteins are present in eukaryotes as well as all gram-positive bacteria and nearly all archaea, but only less than half of the gram-negative bacteria. It has been proposed that eukaryotic smc genes evolved from archaeal precursors by two consecutive gene duplications [48]. Bacteria without SMC proteins often contain an SMC-related long coiled-coil protein involved in chromosome segregation or DNA repair, such as MukB or SbcC [98,13].
Prokaryotic coiled-coil filament proteins
While prokaryotic genomes contained less long coiled-coil proteins than eukaryotes, we found a number of so far uncharacterized long coiled-coil proteins as candidates for filament-forming prokaryotic coiled-coils. These included Heliobacter pylori proteins previously suggested as candidates for bacterial filament proteins [99].
Metazoan mitotic motor proteins lack homologs in plants
The presence of a nucleus in eukaryotic cells is closely linked with the presence of a motile cytoskeleton, in particular the mitotic structures necessary to orchestrate nuclear division, and the endocytic pathway. Dolan et al. [100] proposed a list of motility proteins involved in mitotic processes as candidates for homology searches in prokaryotes to determine their evolutionary origin. We found 70% of the suggested proteins (Astrin, CENP-E, Centrin, Dynein, Dynactin, Kinesin, Kinectin, MAD, NuMA, Pericentrin) among the long coiled-coil proteins identified in our analysis, however none of them clustered together with sequences from archaea or bacteria. Interestingly, with the exception of the kinesins, we also could not find any of these proteins clustering with plant sequences. With the exception of dynein, kinesin and MAD proteins, we could not find clustering of these mitotic motility proteins with yeast sequences either.
The organization of mitotic microtubule nucleation and the composition of the nuclear envelope in plant cells differ significantly from metazoan cells [101]. One hypothesis to explain these differences is the separate development of specialized mechanisms to orchestrate open mitosis in metazoan and plant lineages, leading to the evolution of different nuclear envelope compositions, targeting mechanisms, and mitotic spindle nucleation in the plant and animal kingdoms. This model explains the absence of many metazoan mitotic motility proteins in plants as well as yeast, which undergoes closed mitosis, and suggests that this group of proteins evolved after the occurrence of open mitosis.
We could not find any plant-specific classes of coiled-coil motor proteins, but noted kinesin subclusters largely represented by plant sequences only, indicating an expansion of this group of motor proteins during plant evolution (see Figure 6B). It has been noted before that Arabidopsis contains a surprisingly large number of kinesins [102], and it has been suggested that plant-specific kinesin subfamilies might be involved in stress responses or pathogen defenses [103].
Differences and similarities in cytoskeletal and membrane infrastructure between plants and animals
Besides the motor proteins (myosins, kinesins, dyneins), membrane tethering and vesicle transport proteins appear to be specific for eukaryotes in our clustering analysis, indicating another major class of specialized coiled-coil proteins that evolved after the formation of eukaryotic cells. It has been previously suggested that the higher content of long coiled-coil domains in metazoa compared to plants and protists indicates the presence of extensive coiled-coil matrices in animal cells and tissues [25]. One of the groups of coiled-coil proteins apparently absent in plants and yeasts are the nuclear matrix and intermediate filament proteins. No lamin sequences could be identified from the plant genomes. Other differences we noted between the plant and animal kingdoms are the lack of membrane-cytoskeleton crosslinkers and scaffolding proteins, such as spectrin-like proteins and many actin- and microtubule-associated proteins, in plant proteomes. This might indicate differences in the overall organization and networking of membrane systems and the actin and microtubule cytoskeleton in plant and animal cells.
Differences in coiled-coil content between genomes
Earlier surveys of coiled-coil sequences in GenBank had suggested that invertebrate genomes contain more coiled-coils than vertebrates, and that animal genomes contain four times more "extended" coiled-coils (>75 amino acids) than plant genomes [25]. While we could not find such a difference for the overall coiled-coil content or the group of proteins defined as "long" coiled-coils in this study, we did note a significantly lower percentage of coiled-coils longer than 250 amino acids in yeast as well as plants compared to the animal genomes (see Figure 1). On average, the yeasts contained one third of the percentage of coiled-coils present in vertebrate genomes with domains longer than 100 and longer than 250 residues (37% and 35%, respectively), whereas invertebrates contained about two thirds (60% and 73%, respectively). The plant genomes, however, contained on average 57% of the percentage of proteins with coiled-coil domains longer than 100 amino acids, but only 22% of the coiled-coils with 250 amino acids and longer when compared to vertebrates. An interesting observation is that the human genome appears to contain more extended coiled-coil proteins than the mouse genome. Our data suggests that this is caused by the human proteome sequence set containing more unique long coiled-coil proteins without homologs in other species (see Table 3), as well as more redundant sequences in clusters (e.g. comparing counts of human versus mouse sequences in clusters listed in additional file 5, Table S17).
Comparison with other genome-wide coiled-coil predictions
Comparable with the Arabidopsis coiled-coil protein database ARABI-COIL, this study takes a more restrictive approach to identifying coiled-coil proteins than previous genome-wide approaches to predict coiled-coil proteins [44,43]. In contrast to the older studies, our prediction criteria included a minimum coiled-coil domain length corresponding to about three heptad repeats to eliminate sequences with short stretches of predicted coiled-coils unlikely to form stable structures [11]. Using these parameters, on average about 6.4% of all proteins in the eukaryotic proteomes and about 3.5% in the prokaryotic proteomes (2.6% in archaea, 3.7% in bacteria) contained coiled-coil domains. Our results were consistent with the study of Liu and Rost [43] in that most eukaryotic genomes contained more coiled-coil proteins than prokaryotic genomes, and most bacterial genomes more than archaea. The more restrictive parameters used here resulted in predicting on average about 65–70% of the number of proteins found in those previous studies. Liu and Rost [43] further found an exceptionally high coiled-coil content in Heliobacter pylori with a higher percentage than C. elegans, and an exceptionally low coiled-coil content in Mycobacterium tuberculosis. Our analysis was consistent with these previous observations and resulted in 5.6% coiled-coil for Heliobacter pylori versus 5.4% in C. elegans and only 1.8% in Mycobacterium tuberculosis, the lowest percentage for all 22 genomes analyzed here.
Limitations of the prediction and clustering analysis
Discontinuous coiled-coil domain predictions
MultiCoil provides a more stringent coiled-coil prediction than other programs such as COILS, resulting in less false positive predictions. In tests on the PDB database of solved protein structures, two-thirds of the sequences predicted by COILS did not contain coiled-coils [104]. By comparison, the programs PAIRCOIL and MultiCoil perform significantly better [42]. Occasionally, however, the increased stringency might lead to prediction of fragmented domains where continuous domains have been experimentally verified, as evident in the case of the SMC proteins (see Figure 2).
Selection of long coiled-coil proteins only
In this study, we focused on proteins potentially involved in structural functions. As the emphasis was placed on proteins with long or multiple coiled-coil domains, it is possible that our selection criteria resulted in the exclusion of homologs of proteins with short stretches of coiled-coil that barely qualified for the analysis. The selection criteria applied in this study have been shown to exclude 97% of the known bZIP proteins from Arabidopsis [11]. Other examples we noted are the translation initiation factor IF-2, mitochondrial and prokaryotic seryl-tRNA synthetases, and the ClpB/HSP104 family of heatshock proteins. Members of these protein families failed to meet the selection criteria for long coiled-coil domains, making it difficult to draw conclusions for these protein families from our clustering analysis. We therefore focused our attention on clusters with mainly proteins containing longer coiled-coils (>150 amino acids).
Effect of coiled-coil masking in the clustering analysis
When clustering sequences with long coiled-coil domain in the pilot analysis, the majority of proteins with long coiled-coil domains was grouped together in one large cluster. Many of the proteins with unknown functions in this group were annotated as "myosin-like", however only about 20% of the proteins in the cluster actually contained a myosin motor domain. In the other cases, the only similarity to myosin was the presence of a long coiled-coil domain similar to the myosin coiled-coil tail. This illustrates the ease with which long coiled-coil domains can lead to misannotations in databases with annotations based on sequence similarity searches.
Masking the coiled-coil domains before sequence comparison and clustering significantly increased the specificity of the clustering analysis, however protein sequences with high coiled-coil coverage were lost in the subsequent clustering as the masking left little to no sequence for comparison. Examples are the animal and yeast tropomyosins, many of which were predicted to contain 100% coiled-coil coverage, paramyosin, and the plant cytoskeletal protein CIP1 with more than 80% coiled-coil coverage [105].
Conclusion
Our genome-wide identification of coiled-coil proteins and subsequent clustering provides data suggesting evolutionary conservation or uniqueness of coiled-coil proteins among 22 fully sequenced genomes. We found SMC, MukB, SbcC and Rad50 proteins to be the proteins with the longest coiled-coil domains occurring in prokaryotes, whereas eukaryotic proteomes also contained proteins with stretches of coiled-coil longer than the SMC rod domains. The high conservation of the SMC proteins and their structural relatives involved in chromosome maintenance and repair demonstrates the universal importance and conservation of DNA housekeeping mechanisms.
Long coiled-coil proteins specific to eukaryotes are predominantly involved in subcellular infrastructure maintenance and trafficking control. Table 12 gives an overview of the functional classes of long coiled-coil proteins found in our analysis and their representation in different kingdoms. The genomes of higher plants lack sequences coding for intermediate filament proteins. Many of the known mitotic spindle associated coiled-coil motor proteins in animals lack homologs in plants, consistent with the absence of a centrosomal microtubule organization center in plant cells. However, the kinesin family of microtubule motor proteins appears to have expanded during the evolution of higher plants.
Table 12 Summary of coiled-coil protein functions
Functional groups of coiled-coil proteins Examples Species represented
Chromatin organization and maintenance, chromosome segregation and DNA repair SMCs, Rad50, SbcC, MukB, MutS all kingdoms
Transcription and translation Transcription and translation initiation factors, reverse transcriptase all kingdoms
Protein trafficking and quality control Chaperonins, secretion proteins prokaryotes and organelles
Membrane sensors, channels and regulation of influx/export MCPs, ion channels prokaryotes
Sensor mechanisms and signal transduction Receptor kinases, GTPases eukaryotes – conserved, as well as plant and animal specific
Compartmentalization, stabilization and dynamics of membrane systems Golgins, SNAREs, endocytic proteins eukaryotes
Adherence Cell adherence, extracellular matrix, intracellular adapters eukaryotes and parasitic prokaryotes
Mechanical fiber and meshwork formation Keratin, intermediate filaments, flagellar (e.g. sperm tail) proteins eukaryotes, keratin only in mammals
Motility Muscle fibers, cell motility, actin and microtubule motors eukaryotes
Organization, stabilization and dynamics of the cytoskeleton Actin and microtubule crosslinkers eukaryotes, predominantly metazoa
Mitotic spindle assembly and checkpoint control Centrosome, kinetochore and spindle pole body proteins metazoa and yeast
The repeat nature of the coiled-coil motif makes it difficult to clearly determine sequence homology relationships between long coiled-coil proteins. Functional studies will have to reveal whether so far uncharacterized prokaryotic and plant coiled-coil proteins fulfill similar functions to metazoan counterparts.
Methods
Sequence data and pre-processing
Proteome sequence sets of fully sequenced genomes were downloaded from the European Bioinformatics Institute (EBI) [106] for organisms listed in Table 1, with the exception of rice. The rice proteome set was downloaded from The Institute for Genome Research (TIGR) [107]. An initial preprocessing of the FASTA files was conducted to standardize identifiers for the sequences for easier incorporation into a MySQL database.
Coiled-coil prediction and post-processing
Prediction and selection of coiled-coil proteins was performed using the underlying schema and software systems developed to create the ARABI-COIL database [11]. In summary, the modified FASTA files were used as input for the MultiCoil application installed on the Linux Cluster of the Ohio Supercomputer Center (OSC, Columbus, OH). The MultiCoil output was post-processed using the previously described Java-based ExtractProp Suite [11] and used to establish a database of coiled-coil prediction data for each organism. The same coiled-coil selectivity criteria applied to ARABI-COIL were used to select sequences predicted to contain long or multiple coiled-coil domains. These criteria impose a minimum coiled-coil domain of 30 residues if at least three domains are present in the translated reading frame, a minimum of 50 residues if at least two domains are present, and a minimum coiled-coil length of 70 residues if only a single domain is present. Intra-domain gaps of less than 20 residues were considered contiguous for purposes of establishing domain length. The resulting data was converted to XML and used to populate MySQL databases for each genome.
Masking of coiled-coil domains
To eliminate interference of the coiled-coil repeat motif with sequence homology analysis, coiled-coil domains were "masked" before subjecting the sequences to Smith-Waterman sequence similarity searches. Mask information was created based on the processed MultiCoil prediction data generated to populate the MySQL databases for each genome. A Java-based program was applied to the FASTA sequences selected for Smith-Waterman comparison to replace all amino acids predicted to be contained in coiled-coil domains with the letter X, effectively masking coiled-coil domains.
Sequence similarity comparison
Smith-Waterman comparison was conducted using the TimeLogic Smith-Waterman implementation at OSC and the Blosum62 scoring matrix on all unique sequences in the combined sequences set. Sequences with masked coiled-coil domains were used as query on unmasked sequence sets as target. A P-score cut-off of 1.0e-03 was used as a threshold for selecting sequence similarity relationships. For sequences to be characterized as pair-wise similar and recovered for use in the clustering analysis, the P-score value must be less than this threshold based on the query-target Smith-Waterman comparison.
Clustering analysis
After completing the pair-wise similarity calculation using the Smith-Waterman algorithm and extracting sequence pairs and associated P-scores, sequences were grouped using a modified version of Kruskal's minimum cost spanning tree algorithm [57]. The algorithm creates and progressively merges sub-trees of a graph in building a minimum cost spanning tree. In the algorithm, the weights of edges in the directed graph were determined by the pair-wise P-score similarity value for the sequence as a query relative to the related sequence as a target. An effective clustering can be achieved by using only P-score similarity values which are below a specified threshold, effectively creating a disconnected series of related sequences.
The clustering was tested in a pilot analysis on a combined sequence set including 527 prokaryotic long coiled-coil proteins and eukaryotic proteins containing extended coiled-coil domains of at least 250 amino acids in length or at least 60% of the protein sequence in a coiled-coil. Edges with P-scores greater than 1.0e-03 to 1.0e-15 were ignored when combining sub-trees in the algorithm. The success of the clustering was estimated by observing the clustering behavior of well-known coiled-coil protein families, such as SMC proteins and myosins. After testing the effects of masking the coiled-coil domains and optimizing cut-offs for P-scores during clustering, the complete coiled-coil sequence set containing 3576 long coiled-coil proteins from the 22 genomes was processed similarly. Different P-score thresholds were explored in efforts to increase specificity in the multi-genome sequence set while preserving comprehensive coverage. Employing Kruskal's algorithm, the 3576 sequence set resulted in 156 clusters covering 3567 sequences using a threshold of 1.0e-03, 467 clusters covering 3551 sequences using a threshold of 1.0e-6 and 850 clusters covering 3520 sequences using a threshold of 1.0e-15. (For comparison, the same algorithm yielded 490 clusters for the unmasked sequence set).
Even with the improved selectivity of the clustering demonstrated in the pilot investigation using masked coiled-coil sequences, the overall effectiveness of the resulting clustering still required refinement to achieve sufficient specificity. The use of Kruskal's algorithm for subset selection enabled transitively similar sequences to be included in specific clusters. (Transitively similar sequences are sequences in which sequence A is similar to sequence B and sequence B is similar to sequence C thereby clustering sequence A and C which would otherwise not belong to the same cluster.) One drawback of this simplified clustering is that a given sequence need only be similar to at least one other sequence in the cluster. This limitation resulted in clusters containing sequences which, while closely related to at least one other sequence in a cluster, were not closely related to every sequence within the cluster.
The algorithm was consequently improved to specifically preclude transitively similar sequences by requiring all sequences in a given cluster to satisfy the P-score threshold for all pair-wise relationships in the cluster. The new algorithm dramatically improved specificity, with the same 3576 masked sequence set generating 1213 non-overlapping clusters covering 3567 sequences, 1263 non-overlapping clusters covering 3551 sequences, and 1384 non-overlapping clusters covering 3520 sequences with the improved algorithm for the same corresponding P-score threshold values. The P-score threshold of 1.0e-06 was selected as the appropriate balance of sequence coverage and cluster discrimination required.
The interest in identifying sequences which qualified for more than one cluster and bridged multiple clusters of protein families drove a second modification of the clustering algorithm. By design, the modified Kruskal's algorithm created mutually orthogonal, non-overlapping clusters while precluding transitively similar sequences from populating the same cluster. The 'greedy' algorithm was modified to specifically identify transitively similar sequences between clusters, enabling a unique ability to identify "bridge" sequences which satisfy participation criteria in multiple clusters or protein families. The modification amounted to simply validating each sequence's individual ability to satisfy participation criteria for a cluster based on the non-overlapping cluster partitioning.
The software used to conduct the actual cluster analysis in the study is available for download at the Ohio Bioscience Library [108].
Cluster alignments and phylogenetic tree generation
Multiple sequence alignments and phylogenetic trees were generated for clusters of interest using sequences with masked coiled-coil domains and ClustalW version 1.82 incorporating the Blossum scoring matrix [109]. Phylogenetic trees were generated using the ClustalW program with a bootstrap parameter of 10,000 and displayed using the program TreeView v.1.6.6 [110].
List of abbreviations
CASP, CDP/cut alternatively spliced product
CC, coiled-coil
CDD, conserved domain database
CDP, CCAAT displacement protein
CENP, centromer protein
CIP1, COP1-interactive protein 1
CHUP1, chloroplast unusual positioning 1
CLIP, cytoplasmic linker protein
DAM, disheveled associated activator of morphogenesis
DIA1, Diaphanous-related formin 1
DOC1, downregulated in ovarian cancer 1
EBI, European Bioinformatics Institute
ERM, ezrin/radixin/moesin
FPPs, filament-like plant proteins
Hsp, heat shock protein
IF, intermediate filament
KCBP, kinesin-like calmodulin-binding protein
KIP1, kinase interacting protein 1
KLP, kinesin-like protein
MCP, methyl-accepting chemotaxis protein
MKRP, mitochondrial kinesin-related protein
MLP, myosin-like protein
NuMA, nuclear mitotic apparatus
ORF, open reading frame
OSC, Ohio Supercomputer Center
PAKRP, phragmoplast-associated kinesin-related protein
PP1, protein phosphatase 1
RBP, Retinoblastoma-binding protein
ROCK, Rho-associated coiled-coil containing kinase
SLAP, sarcolemmal-associated protein
SMC, structural maintenance of chromosomes
S/W, Smith-Waterman sequence comparison
TACC, transforming acidic coiled-coil
Tpr, translocated promoter region
VIPP1, vesicle-inducing plastid protein 1
XML, extensible markup language
Authors' contributions
AR coordinated this study, analyzed the data, and prepared the manuscript. SJS participated in MultiCoil and Smith-Waterman output processing and ClustalW analysis. EAS generated MultiCoil and Smith-Waterman outputs, developed software for pre- and post-processing and coiled-coil masking, and wrote the code for the clustering algorithm. IM proposed and supervised the study and edited the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional file 1
Prokaryotic coiled-coil proteins Tables S1-S14: Protein details of all long coiled-coil proteins predicted in the prokaryotic genomes analyzed in this study. Open file with Acrobat Reader.
Click here for file
Additional file 2
Eukaryotic clusters of interest Figures S1-S6: Phylogenetic trees based on ClustalW alignments of the sequences, displayed using TreeView v.1.6.6. Open file with Acrobat Reader.
Click here for file
Additional file 3
Sequence details for Figures S1-S6, supplement to additional file 2. Table S15: Protein information and prediction data for sequences contained in Figures S1-S6. AGI locus numbers from TAIR are used as sequence IDs for Arabidopsis, TIGR sequence IDs are used for rice and Synechocystis. All other sequence IDs correspond to the EBI identifiers in the downloaded FASTA files. Max. Coil Length, longest coiled-coil domain in the protein sequence; Coil Coverage, percent of sequence predicted to be in a coiled-coil. Open file with Microsoft Excel.
Click here for file
Additional file 4
Yeast clusters Table S18: Protein information and prediction data for sequences in yeast clusters with two species (Saccharomyces cerevisiae and Schizosaccharomyces pombe) represented. Sequence IDs correspond to the EBI identifiers in the downloaded FASTA files. Max. Coil Length, longest coiled-coil domain in the protein sequence; Coil Coverage, percent of sequence predicted to be in a coiled-coil. Open file with Microsoft Excel.
Click here for file
Additional file 5
Small mammalian clusters; supplement to Table 9. Table S16: Protein information and prediction data for sequences in mammalian clusters with two species (mouse, human) represented and less than 10 sequences per cluster. Sequence IDs correspond to the EBI identifiers in the downloaded FASTA files. Max. Coil Length, longest coiled-coil domain in the protein sequence; Coil Coverage, percent of sequence predicted to be in a coiled-coil. Open file with Microsoft Excel.
Click here for file
Additional file 6
Small plant clusters; supplement to Table 10. Table S17: Protein information and prediction data for sequences in plant clusters with two species (Arabidopsis, rice) represents and less than 10 sequences per cluster. AGI locus numbers from TAIR or NCBI RefSeq numbers are used as sequence IDs for Arabidopsis, TIGR sequence IDs are used for rice. Max. Coil Length, longest coiled-coil domain in the protein sequence; Coil Coverage, percent of sequence predicted to be in a coiled-coil. Open file with Microsoft Excel.
Click here for file
Acknowledgements
We thank the Ohio Supercomputer Center for providing usage time for this analysis. This work was supported by the National Science Foundation 2010 Project (grant no. NSF 0209339 to I.M.).
==== Refs
Crick FH The packing of alpha-helices: Simple coiled-coils Acta Cryst 1953 6 689 697 10.1107/S0365110X53001964
Burkhard P Stetefeld J Strelkov SV Coiled coils: a highly versatile protein folding motif Trends Cell Biol 2001 11 82 88 11166216 10.1016/S0962-8924(00)01898-5
Magin TM Reichelt J Hatzfeld M Emerging functions: diseases and animal models reshape our view of the cytoskeleton Exp Cell Res 2004 301 91 102 15501450 10.1016/j.yexcr.2004.08.018
Mounkes L Kozlov S Burke B Stewart CL The laminopathies: nuclear structure meets disease Curr Opin Genet Dev 2003 13 223 230 12787783 10.1016/S0959-437X(03)00058-3
Puls I Jonnakuty C LaMonte BH Holzbaur ELF Tokito M Mann E Floeter MK Bidus K Drayna D Oh SJ Brown RH JrLudlow CL Fischbeck KH Mutant dynactin in motor neuron disease Nat Genet 2003 33 455 456 12627231 10.1038/ng1123
Hirokawa N Takemura R Molecular motors in neuronal development, intracellular transport and diseases Curr Opin Neurobiol 2004 14 564 573 15464889 10.1016/j.conb.2004.08.011
Chigira S Sugita K Kita K Sugaya S Arase Y Ichinose M Shirasawa H Suzuki N Increased expression of the Huntingtin interacting protein-1 gene in cells from Hutchinson Gilford Syndrome (progeria) patients and aged donors J Gerontol A Biol Sci Med Sci 2003 58 B873 878 14570852
Mounkes LC Stewart CL Aging and nuclear organization: lamins and progeria Curr Opin Cell Biol 2004 16 322 327 15145358 10.1016/j.ceb.2004.03.009
Raff JW Centrosomes and cancer: lessons from a TACC Trends Cell Biol 2002 12 222 225 12062169 10.1016/S0962-8924(02)02268-7
McClatchey AI Merlin and ERM proteins: unappreciated roles in cancer development? Nat Rev Cancer 2003 3 877 883 14668818 10.1038/nrc1213
Rose A Manikantan S Schraegle SJ Maloy MA Stahlberg EA Meier I Genome-wide identification of Arabidopsis coiled-coil proteins and establishment of the ARABI-COIL database Plant Physiol 2004 134 927 939 15020757 10.1104/pp.103.035626
Lundin VF Stirling PC Gomez-Reino J Mwenifumbo JC Obst JM Valpuesta JM Leroux MR Molecular clamp mechanism of substrate binding by hydrophobic coiled-coil residues of the archaeal chaperone prefoldin Proc Natl Acad Sci USA 2004 101 4367 4372 15070724 10.1073/pnas.0306276101
Connelly JC Kirkham LA Leach DRF The SbcCD nuclease of Escherichia coli is a structural maintenance of chromosomes (SMC) family protein that cleaves hairpin DNA Proc Natl Acad Sci USA 1998 95 7969 7974 9653124 10.1073/pnas.95.14.7969
Delahay RM Frankel G Coiled-coil proteins associated with type III secretion systems: a versatile domain revisited Mol Microbiol 2002 45 905 916 12180912 10.1046/j.1365-2958.2002.03083.x
Balish MF Ross SM Fisseha M Krause DC Deletion analysis identifies key functional domains of the cytadherence-associated protein HMW2 of Mycoplasma pneumoniae Mol Microbiol 2003 50 1507 1516 14651634 10.1046/j.1365-2958.2003.03807.x
Jakoby M Weisshaar B Droge-Laser W Vincente-Carbajosa J Tiedemann J Kroi T Parcy F bZIP transcription factors in Arabidopsis Trends Plant Sci 2002 7 106 111 11906833 10.1016/S1360-1385(01)02223-3
Vinson C Myakishev M Acharya A Mir AA Moll JR Bonovich M Classification of human B-ZIP proteins based on dimerization properties Mol Cell Biol 2002 22 6321 6335 12192032 10.1128/MCB.22.18.6321-6335.2002
Wang Y Gao R Lynn DG Racheting up vir gene expression in Agrobacterium tumefaciens: coiled coils in histidine kinase signal transduction ChemBioChem 2002 3 311 317 11933231 10.1002/1439-7633(20020402)3:4<311::AID-CBIC311>3.0.CO;2-N
Blumenthal R Clague MJ Durell SR Epand RM Membrane fusion Chem Rev 2003 103 53 69 12517181 10.1021/cr000036+
Sillibourne JE Milne DM Takahashi M Ono Y Meek DW Centrosomal anchoring of the protein kinase CK1delta mediated by attachment to the large, coiled-coil scaffolding protein CG-NAP/AKAP450 J Mol Biol 2002 322 785 797 12270714 10.1016/S0022-2836(02)00857-4
Zhao X Wu CY Blobel G Mlp-dependent anchorage and stabilization of a desumoylating enzyme is required to prevent clonal lethality J Cell Biol 2004 167 605 11 15557117 10.1083/jcb.200405168
Zhen YY Libotte T Munck M Noegel AA Korenbaum E NUANCE, a giant protein connecting the nucleus and actin cytoskeleton J Cell Sci 2002 115 3207 3222 12118075
Wood A Krogan NJ Dover J Schneider J Heidt J Boateng MA Dean K Golshani A Zhang Y Greenblatt JF Johnston M Shilatifard A Bre1, an E3 ubiquitin ligase required for recruitment and substrate selection of Rad6 at a promoter Mol Cell 2003 11 267 274 12535539 10.1016/S1097-2765(02)00802-X
Hardtke CS Deng XW The cell biology of the COP/DET/FUS proteins. Regulating proteolysis in photomorphogenesis and beyond? Plant Physiol 2000 124 1548 1557 11115873 10.1104/pp.124.4.1548
Odgren PR Harvie LW JrFey EG Phylogenetic occurrence of coiled coil proteins: implications for tissue structure in metazoa via a coiled coil tissue matrix Proteins 1996 24 467 484 9162947 10.1002/(SICI)1097-0134(199604)24:4<467::AID-PROT6>3.0.CO;2-B
Kammerer RA alpha-Helical coiled-coil oligomerization domains in extracellular proteins Matrix Biol 1997 15 555 565 9138288 10.1016/S0945-053X(97)90031-7
Rose A Meier I Scaffolds, levers, rods and springs: diverse cellular functions of long coiled-coil proteins Cell Mol Life Sci 2004 61 1996 2009 15316650 10.1007/s00018-004-4039-6
Strelkov SV Herrmann H Aebi U Molecular architecture of intermediate filaments BioEssays 2003 25 243 251 12596228 10.1002/bies.10246
Helfand BT Chang L Goldman RD Intermediate filaments are dynamic and motile elements of cellular architecture J Cell Sci 2004 117 133 141 14676269 10.1242/jcs.00936
Schaerer F Morgan G Winey M Philippsen P Cnm67p is a spacer protein of the Saccharomyces cerevisiae spindle pole body outer plaque Mol Biol Cell 2001 12 2519 2533 11514632
Kilmartin JV Dyos SL Kershaw S Finch JT A spacer protein in the Saccharomyces cerevisiae spindle pole body whose transcript is cell cycle-regulated J Cell Biol 1993 123 1175 1184 7503995 10.1083/jcb.123.5.1175
De Matteis MA Morrow JS Spectrin tethers and mesh in the biosynthetic pathway J Cell Sci 2000 113 2331 2343 10852813
Barr FA Short B Golgins in the structure and dynamics of the Golgi apparatus Curr Opin Cell Biol 2003 15 405 413 12892780 10.1016/S0955-0674(03)00054-1
Schliwa M Woehlke G Molecular motors Nature 2003 422 759 765 12700770 10.1038/nature01601
Mogk A Bukau B Molecular chaperones: structure of a protein disaggregase Curr Biol 2004 14 R78 R80 14738756 10.1016/j.cub.2003.12.051
Graumann PL SMC proteins in bacteria: condensation motors for chromosome segregation? Biochimie 2001 83 53 59 11254975 10.1016/S0300-9084(00)01218-9
Cromie GA Leach DRF Recombinational repair of chromosomal DNA double-strand breaks generated by a restriction endonuclease Mol Microbiol 2001 41 873 883 11532150 10.1046/j.1365-2958.2001.02553.x
Mason JM Arndt KM Coiled coil domains: stability, specificity, and biological implications ChemBioChem 2004 5 170 176 14760737 10.1002/cbic.200300781
Berger B Wilson DB Wolf E Tonchev T Milla M Kim PS Predicting coiled coils by use of pairwise residue correlations Proc Natl Acad Sci USA 1995 92 8259 8263 7667278
Parry DA Coiled-coils in alpha-helix-containing proteins: analysis of the residue types within the heptad repeat and the use of these data in the prediction of coiled-coils in other proteins Biosci Rep 1982 2 1017 1024 7165792 10.1007/BF01122170
Lupas A Predicting coiled-coil regions in proteins Curr Opin Struct Biol 1997 7 388 393 9204281 10.1016/S0959-440X(97)80056-5
Wolf E Kim PS Berger B MultiCoil: a program for predicting two- and three-stranded coiled coils Protein Sci 1997 6 1179 1189 9194178
Liu J Rost B Comparing function and structure between entire genomes Protein Sci 2001 10 1970 1979 11567088 10.1110/ps.10101
Newman JRS Wolf E Kim PS A computationally directed screen identifying interacting coiled coils from Saccharomyces cerevisiae Proc Natl Acad Sci USA 2000 97 13203 13208 11087867 10.1073/pnas.97.24.13203
Lumb KJ Carr CM Kim PS Subdomain folding of the coiled coil leucine zipper from the bZIP transcriptional activator GCN4 Biochemistry 1994 33 7361 7367 8003501 10.1021/bi00189a042
Su JY Hodges RS Kay CM Effect of chain length on the formation and stability of synthetic alpha-helical coiled-coils Biochemistry 1994 33 15501 15510 7803412 10.1021/bi00255a032
Litowski JR Hodges RS Designing heterodimeric two-stranded alpha-helical coiled-coils: the effect of chain length on protein folding, stability and specificity J Pept Res 2001 58 477 492 12005418 10.1034/j.1399-3011.2001.10972.x
Soppa J Prokaryotic structural maintenance of chromosomes (SMC) proteins: distribution, phylogeny, and comparison with MukBs and additional prokaryotic and eukaryotic coiled-coil proteins Gene 2001 278 253 264 11707343 10.1016/S0378-1119(01)00733-8
Marchler-Bauer A Anderson JB Cherukuri PF DeWeese-Scott C Geer LY Gwadz M He S Hurwitz DI Jackson JD Ke Z Lanczycki CJ Liebert CA Liu C Lu F Marchler GH Mullokandov M Shoemaker BA Simonyan V Song JS Thiessen PA Yamashita RA Yin JJ Zhang D Bryant SH CDD: a conserved domain database for protein classification Nucl Acids Res 2005 33 D192 D196 15608175 10.1093/nar/gki069
Pereira M Parente JA Bataus LA Cardoso DD Soares RB Soares CM Chemotaxis and flagellar genes of Chromobacterium violaceum Genet Mol Res 2004 3 92 101 15100991
Ip H Stratton K Zgurskaya H Liu J pH-induced conformational changes of AcrA, the membrane fusion protein of Escherichia coli multidrug efflux system J Biol Chem 2003 278 50474 50482 14523004 10.1074/jbc.M305152200
Henry T Pommier S Journet L Bernadac A Gorvel JP Lloubes R Improved methods for producing outer membrane vesicles in Gram-negative bacteria Res Microbiol 2004 155 437 446 15249060 10.1016/j.resmic.2004.04.007
Hackstadt T Scidmore-Carlson MA Shaw EI Fischer ER The Chlamydia trachomatis IncA protein is required for homotypic vesicle fusion Cell Microbiol 1999 1 119 130 11207546 10.1046/j.1462-5822.1999.00012.x
Montag D Schwarz H Henning U A component of the side tail fiber of Escherichia coli bacteriophage lambda can functionally replace the receptor-recognizing part of a long tail fiber protein of the unrelated bacteriophage T4 J Bacteriol 1989 171 4378 4384 2526805
Anantharaman V Aravind L Evolutionary history, structural features and biochemical diversity of the NlpC/P60 superfamily of enzymes Genome Biol 2003 4 R11 12620121 10.1186/gb-2003-4-2-r11
Smith TF Waterman MS Identification of common molecular subsequences J Mol Biol 1981 147 195 197 7265238 10.1016/0022-2836(81)90087-5
Kruskal JB On the shortest spanning subtree of a graph and the traveling salesman problem Proc Amer Math Soc 1956 7 48 50
Kroll D Meierhoff K Bechtold N Kinoshita M Westphal S Vothknecht UC Soll J Westhoff P VIPP1, a nuclear gene of Arabidopsis thaliana essential for thylakoid membrane formation Proc Natl Acad Sci USA 2001 98 4238 4242 11274447 10.1073/pnas.061500998
Westphal S Heins L Soll J Vothknecht UC Vipp1 deletion mutant of Synechocystis: a connection between bacterial phage shock and thylakoid biogenesis? Proc Natl Acad Sci USA 2001 98 4243 4248 11274448 10.1073/pnas.061501198
Gentschev I Dietrich G Goebel W The E. coli alpha-hemolysin secretion system and its use in vaccine development Trends Microbiol 2002 10 39 45 11755084 10.1016/S0966-842X(01)02259-4
Vos JW Safida F Reddy ASN Hepler PK The kinesin-like calmodulin binding protein is differentially involved in cell division Plant Cell 2000 12 979 990 10852941 10.1105/tpc.12.6.979
Hirokawa N Takemura R Kinesin superfamily and their various functions and dynamics Exp Cell Res 2004 301 50 59 15501445 10.1016/j.yexcr.2004.08.010
Wen H Ao S RBP95, a novel leucine zipper protein, binds to retinoblastoma protein Biochem Biophys Res Comm 2000 275 141 148 10944455 10.1006/bbrc.2000.3242
Hwang WW Venkatasubrahmanyam S Ianculescu AG Tong A Boone A Madhani HD A conserved RING finger protein required for histone H2B monoubiquitination and cell size control Mol Cell 2003 11 261 266 12535538 10.1016/S1097-2765(02)00826-2
Watanabe N Higashida C Formins: processive cappers of growing actin filaments Exp Cell Res 2004 301 16 22 15501440 10.1016/j.yexcr.2004.08.020
Madrid R Gasteier JE Bouchet J Schröder S Geyer M Benichou S Fackler OT Oligomerization of the diaphanous-related formin FHOD1 requires a coiled-coil motif critical for its cytoskeletal and transcriptional activities FEBS Letters 2005 579 441 448 15642356 10.1016/j.febslet.2004.12.009
Cuif MH Possmayer F Zander H Bordes N Jollivet F Couedel-Courteille A Janoueix-Lerosey I Langsley G Bornens M Goud B Characterization of GAPCenA, a GTPase activating protein for Rab6, part of which associates with the centrosome EMBO J 1999 18 1772 1782 10202141 10.1093/emboj/18.7.1772
Gillingham AK Pfeifer AC Munro S CASP, the alternatively spliced product of the gene encoding the CCAAT-displacement protein transcription factor, is a Golgi membrane protein related to Giantin Mol Biol Cell 2002 13 3761 3774 12429822 10.1091/mbc.E02-06-0349
Puthenveedu M Linstedt AD Gene replacement reveals that p115/SNARE interactions are essential for Golgi biogenesis Proc Natl Acad Sci USA 2004 101 1253 1256 14736916 10.1073/pnas.0306373101
Kametaka S Okano T Ohsumi M Ohsumi Y Apg14p and Apg6/Vps30p form a protein complex essential for autophagy in the yeast, Saccharomyces cerevisiae J Biol Chem 1998 273 22284 22291 9712845 10.1074/jbc.273.35.22284
Mu FT Callaghan JM Steele-Mortimer O Stenmark H Parton RG Campbell PL McCluskey J Yeo JP Tock EPC Toh BH EEA1, and early endosome-associated protein. EEA1 is a conserved alpha-helical peripheral membrane protein flanked by cysteine "fingers" and contains a calmodulin-binding IQ motif J Biol Chem 1995 270 13503 13511 7768953 10.1074/jbc.270.22.13503
Suntharalingam M Alcazar-Roman AR Wente SR Nuclear export of the yeast mRNA-binding protein Nab2 is linked to a direct interaction with Gfd1 and to Gle1 function J Biol Chem 2004 279 35384 35391 15208322 10.1074/jbc.M402044200
Green DM Johnson CP Hagan H Corbett AH The C-terminal domain of myosin-like protein 1 (Mlp1p) is a docking site for heterogenous nuclear ribonucleoproteins that are required for mRNA export Proc Natl Acad Sci USA 2003 100 1010 1015 12531921 10.1073/pnas.0336594100
Gillet ES Espelin CW Sorger PK Spindle checkpoint proteins and chromosome-microtubule attachment in budding yeast J Cell Biol 2004 164 535 546 14769859 10.1083/jcb.200308100
Barr FA A novel Rab6-interacting domain defines a family of Golgi-targeted coiled-coil proteins Curr Biol 1999 9 381 384 10209123 10.1016/S0960-9822(99)80167-5
Munro S Nichols BJ The GRIP domain – a novel Golgi-targeting domain found in several coiled-coil proteins Curr Biol 1999 9 377 380 10209120 10.1016/S0960-9822(99)80166-3
van Drogen F Peter M Spa2p functions as a scaffold-like protein to recruit the Mpk1p MAP kinase module to sites of polarized growth Curr Biol 2002 12 1698 1703 12361575 10.1016/S0960-9822(02)01186-7
Behrens R Nurse P Roles of fission yeast tea1p in the localization of polarity factors and in organizing the microtubular cytoskeleton J Cell Biol 2002 157 783 793 12034771 10.1083/jcb.200112027
Adams J Kelso R Cooley L The kelch repeat superfamily of proteins: propellers of cell function Trends Cell Biol 2000 10 17 24 10603472 10.1016/S0962-8924(99)01673-6
Arnal I Heichette C Diamantopoulos GS Chretien D CLIP-170/tubilin-curved oligomers coassemble at microtubule ends and promote rescues Curr Biol 2004 14 2086 2095 15589150 10.1016/j.cub.2004.11.055
Reddy ASN Day IS Analysis of the myosins encoded in the recently completed Arabidopsis thaliana genome sequence Genome Biol 2001 2 RESEARCH0024 11516337 10.1186/gb-2001-2-7-research0024
Karashima T Watt FM Interaction of periplakin and envoplakin with intermediate filaments J Cell Sci 2002 115 5027 5037 12432088 10.1242/jcs.00191
Serra-Pages C Medley QG Tang M Hart A Streuli M Liprins, a family of LAR transmembrane protein-tyrosine phosphatase-interacting proteins J Biol Chem 1998 273 15611 15620 9624153 10.1074/jbc.273.25.15611
Bretscher A Edwards K Fehon RG ERM proteins and merlin: integrators at the cell cortex Nat Rev Mol Cell Biol 2002 3 586 599 12154370 10.1038/nrm882
Ivetic A Ridley AJ Ezrin/radixin/moesin proteins and Rho GTPase signalling in leucocytes Immunology 2004 112 165 176 15147559 10.1111/j.1365-2567.2004.01882.x
DePianto D Coulombe PA Intermediate filaments and tissue repair Exp Cell Res 2004 301 68 76 15501447 10.1016/j.yexcr.2004.08.007
Goldman RD Gruenbaum Y Moir RD Shumaker DK Spann TP Nuclear lamins: building blocks of nuclear architecture Genes Dev 2002 16 533 547 11877373 10.1101/gad.960502
Baird DH Myers KA Mogensen M Moss D Baas PW Distribution of the microtubule-related protein ninein in developing neurons Neuropharmacology 2004 47 677 683 15458839 10.1016/j.neuropharm.2004.07.016
Gergely F Centrosomal TACCtics Bioessays 2002 24 915 925 12325124 10.1002/bies.10162
Mayor T Stierhof YD Tanaka K Fry AM Nigg EA The centrosomal protein C-Nap1 is required for cell cycle-regulated centrosome cohesion J Cell Biol 2000 151 837 846 11076968 10.1083/jcb.151.4.837
Gromley A Jurczyk A Sillibourne J Halilovic E Mogensen M Groisman I Blomberg M Doxsey S A novel human protein of the maternal centriole is required for the final stages of cytokinesis and entry into S phase J Cell Biol 2003 161 535 545 12732615 10.1083/jcb.200301105
Wilson M Koopman P Matching SOX: partner proteins and co-factors of the SOX family of transcriptional regulators Curr Opin Genet Dev 2002 12 441 446 12100890 10.1016/S0959-437X(02)00323-4
Skirpan AL McCubbin AG Ishimizu T Wang X Hu Y Dowd PE Ma H Kao TH Isolation and characterization of kinase interacting protein 1, a pollen protein that interacts with the kinase domain of PRK1, a receptor-like kinase of petunia Plant Physiol 2001 126 1480 1492 11500547 10.1104/pp.126.4.1480
Gindullis F Rose A Patel S Meier I Four signature motifs define the first class of structurally related large coiled-coil proteins in plants BMC Genomics 2002 3 9 11972898 10.1186/1471-2164-3-9
Yao H Zhou Q Li J Smith H Yandeau M Nikolau BJ Schnable PS Molecular characterization of meiotic recombination across the 140-kb multigenic a1-sh2 interval of maize Proc Natl Acad Sci USA 2002 99 6157 6162 11959909 10.1073/pnas.082562199
Masuda K Xu ZJ Takahashi S Ito A Ono M Nomura K Inoue M Peripheral framework of carrot cell nucleus contains a novel protein predicted to exhibit a long alpha-helical domain Exp Cell Res 1997 232 173 181 9141634 10.1006/excr.1997.3531
Oikawa K Kasahara M Kiyosue T Kagawa T Suetsugu N Takahashi F Kanegae T Niwa Y Kadota A Wada M CHLOROPLAST UNUSUAL POSITIONING1 is essential for proper chloroplast positioning Plant Cell 2003 15 2805 2815 14615600 10.1105/tpc.016428
Dasgupta S Maisnier-Patin S Nordström K New genes with old modus operandi EMBO Rep 2000 1 323 327 11269497 10.1093/embo-reports/kvd077
Ausmess N Kuhn JR Jacobs-Wagner C The bacterial cytoskeleton: an intermediate filament-like function in cell shape Cell 2003 115 705 713 14675535 10.1016/S0092-8674(03)00935-8
Dolan MF Melnitsky H Margulis L Kolnicki R Motility proteins and the origin of the nucleus Anat Rec 2002 268 290 301 12382325 10.1002/ar.10161
Rose A Patel S Meier I The plant nuclear envelope Planta 2004 218 327 336 14610677 10.1007/s00425-003-1132-2
Reddy ASN Day IS Kinesins in the Arabidopsis genome: a comparative analysis among eukaryotes BMC Genomics 2001 2 2 11472632 10.1186/1471-2164-2-2
Lee YRJ Liu B Cytoskeletal motors in Arabidopsis. Sixty-one kinesins and seventeen myosins Plant Physiol 2004 136 3877 3883 15591445 10.1104/pp.104.052621
Berger B Singh M An iterative method for improved protein structural motif recognition J Comput Biol 1997 4 261 273 9278059
Matsui M Stoop CD von Armin AG Wei N Deng XW Arabidopsis COP1 protein specifically interacts in vitro with a cytoskeleton-associated protein, CIP1 Proc Natl Acad Sci USA 1995 92 4239 4243 7753789
European Bioinformatics Institute (EBI)
The Institute for Genome Research (TIGR)
Ohio Bioscience Library
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
Page RDM, TreeView, 2001
Hanlon DW Rosario MM Ordal GW Venema G Van Sinderen D Identification of TlpC, a novel 62 kDa MCP-like protein from Bacillus subtilis Microbiology 1994 140 1847 1854 7921238
Marconi RT Samuels DS Landry RK Garon CF Analysis of the distribution and molecular heterogeneity of the ospD gene among the lyme disease spirochetes: evidence for lateral gene exchange J Bacteriol 1994 176 4572 4582 7913928
Norris SJ Carter CJ Howell JK Barbour AG Low-passage-associated proteins of Borrelia burgdorferi B31: characterization and molecular cloning of OspD, a surface-exposed, plasmid-encoded lipoprotein Infect Immun 1992 60 4662 4672 1398980
Amano M Fukata Y Kaibuchi K Regulation and functions of Rho-associated kinase Exp Cell Res 2000 261 44 51 11082274 10.1006/excr.2000.5046
Ko J Na M Kim S Lee JR Kim E Interaction of the ERC family of RIM-binding proteins with the liprin-alpha family of multidomain proteins J Biol Chem 2003 278 42377 42385 12923177 10.1074/jbc.M307561200
Wielowieyski PA Sevinc S Guzzo R Salih M Wigle JT Tuana BS Alternative splicing, expression, and genomic structure of the 3' region of the gene encoding the sarcolemmal-associated proteins (SLAPs) defines a novel class of coiled-coil tail-anchored membrane proteins J Biol Chem 2000 275 38474 38481 10986292 10.1074/jbc.M007682200
Mok SC Wong KK Chan RK Lau CC Tsao SW Knapp RC Berkowitz RS Molecular cloning of differentially expressed genes in human epithelian ovarian cancer Gynecol Oncol 1994 52 247 252 8314147 10.1006/gyno.1994.1040
Liao H Winkfein RJ Mack G Rattner JB Yen TJ CENP-F is a protein of the nuclear matrix that assembles onto kinetochores at late G2 and is rapidly degraded after mitosis J Cell Biol 1995 130 507 518 7542657 10.1083/jcb.130.3.507
Ng MH Death associated protein kinase: from regulation of apoptosis to tumor suppressive functions and B cell malignancies Apoptosis 2002 7 261 270 11997670 10.1023/A:1015364104672
Zhang R Epstein HF Homodimerization through coiled-coil regions enhance activity of the myotonic dystrophy protein kinase FEBS Lett 2003 546 281 287 12832055 10.1016/S0014-5793(03)00601-X
Altman R Kellogg D Control of mitotic events by Nap1 and he Gin4 kinase J Cell Biol 1997 138 119 130 9214386 10.1083/jcb.138.1.119
Itoh R Fujiwara M Yoshida S Kinesin-related proteins with a mitochondrial targeting signal Plant Physiol 2001 127 724 726 11706156 10.1104/pp.127.3.724
Tanaka H Ishikawa M Kitamura S Takahashi Y Soyano T Machida C Machida Y The AtNACK1/HINKEL and STUD/TETRASPORE/AtNACK2 genes, which encode functionally redundant kinesins, are essential for cytokinesis in Arabidopsis Genes Cells 2004 9 1199 1211 15569152 10.1111/j.1365-2443.2004.00798.x
Hays JL Watowich SJ Oligomerization-dependent changes in the thermodynamic properties of the TPR-MET receptor tyrosine kinase Biochemistry 2004 43 10570 10578 15301554 10.1021/bi0363275
Salcini AE Chen H Iannolo G De Camilli P Di Fiore PP Epidermal growth factor pathway substrate 15, Esp15 Int J Biochem Cell Biol 1999 31 805 809 10481267 10.1016/S1357-2725(99)00042-4
Weiner JA Chun J Png-1, a nervous system-specific zinc finger gene, identifies regions containing postmitotic neurons during mammalian embryonic development J Comp Neurol 1997 381 130 142 9130664 10.1002/(SICI)1096-9861(19970505)381:2<130::AID-CNE2>3.0.CO;2-4
Samuels-Lev Y O'Connor DJ Bergamaschi D Trigante G Hsieh JK Zhong S Campargue I Naumovski L Crook T Lu X ASPP proteins specifically stimulate the apoptotic function of p53 Mol Cell 2001 8 781 794 11684014 10.1016/S1097-2765(01)00367-7
Munton RP Vizi S Mansuy IM The role of protein phosphatase-1 in the modulation of synaptic and structural plasticity FEBS Lett 2004 567 121 128 15165904 10.1016/j.febslet.2004.03.121
Yang J Kim O Wu J Qiu Y Interaction between tyrosine kinase Etk and a RUN domain- and FYVE domain-containing protein RUFY1 J Biol Chem 2002 277 30219 30226 11877430 10.1074/jbc.M111933200
Zhong R Burk DH Morrison H IIIYe ZH A kinesin-like protein is essential for oriented deposition of cellulose microfibrils and cell wall strength Plant Cell 2002 14 3101 3117 12468730 10.1105/tpc.005801
Lee YRJ Giang HM Liu B A novel plant kinesin-related protein specifically associates with the phragmoplast organelles Plant Cell 2001 13 2427 2439 11701879 10.1105/tpc.13.11.2427
Pan R Lee YRJ Liu B Localization of two homologous Arabidopsis kinesin-related proteins in the phragmoplast Planta 2004 220 156 164 15258761 10.1007/s00425-004-1324-4
|
16288662
|
PMC1322226
|
CC BY
|
2021-01-04 16:29:17
|
no
|
BMC Evol Biol. 2005 Nov 16; 5:66
|
utf-8
|
BMC Evol Biol
| 2,005 |
10.1186/1471-2148-5-66
|
oa_comm
|
==== Front
J Transl MedJournal of Translational Medicine1479-5876BioMed Central London 1479-5876-3-431632421310.1186/1479-5876-3-43MethodologyAcquired factor XII deficiency in a woman with recurrent pregnancy loss: working on a differential diagnosis in a single case D'Uva Maristella [email protected] Ida [email protected] Antonio [email protected] Antonio [email protected] Placido Giuseppe [email protected] Micco Pierpaolo [email protected] Dipartimento Universitario di Scienze Ostetriche Ginecologiche e Medicina della Riproduzione, Area Funzionale di Medicina della Riproduzione ed Endoscopia Ginecologica, Università degli Studi di Napoli Federico II, via Pansini 5 Building 9, 80131, Naples, Italy; *Thrombosis Center, Istituto Clinico Humanitas "IRCCS", via Manzoni 56, 20089, Rozzano – Milano – Italy2005 2 12 2005 3 43 43 5 10 2005 2 12 2005 Copyright © 2005 D'Uva et al; licensee BioMed Central Ltd.2005D'Uva 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
Antiphospholipid syndrome (APS) has been often associated to RPL since 1980 and some reports in the Literature rarely described antibodies to factor XII in patients with APS.
Case history
We report the case history of 34-year-old caucasian women with recurrent fetal loss and persistent prolonged activated partial thromboplastin time. Haemostatic tests revealed persistent light decrease of clotting factor XII with normal values of IgG and IgM anticardiolipin antibodies and transient positivity for lupus anticoagulant (LA). Few reports in the Literature described antibodies to factor XII in patient with antiphospholipid syndrome (APS) and transient LA. So, once other causes of RPL were excluded, the patient was diagnosed an unusual form of APS associated to antibodies to factor XII, reduced factor XII plasma levels, transient LA and prolonged activated partial thromboplastin time.
Discussion
We suggest to consider also antibodies directed to clotting factors (e.g. factor XII in our case) as second step of thrombophilia screening in RPL, in particular if a persistent prolonged aPTT is present without an apparent cause.
==== Body
Background
Recurrent pregnancy loss (RPL) is one of the most common cause of sterility. An our recent study underlined the relevant role of d-dimer, as marker of hypercoagulable state, to identify thrombophilia in women affected by primary or secondary sterility [1]
In 1999 a study written by Brenner et al. identified thrombophilia as principal cause of more than 40% of women affected by RPL [2]. Further studies underlined a pathogenetic role of inherited thrombophilia in women affected by RPL. Sanson et al., in fact, reported an increased frequency of antithrombin III, protein C and protein S deficiency in women with RPL [3], while Grandone et al. found an increased incidence of factor V Leiden in women with unexplained RPL [4]. Also prothrombin A20210G gene polymorphism has been reported as possible cause of RPL in several studies [2,5,6].
Yet, also acquired thrombophilia has been associated to RPL. A study by Dossenbach et al., in fact, revealed that elevated maternal plasma levels of clotting factor VIII tend to be associated to an increased risk of RPL [7]. Moreover, several studies in the Literature are available for the association of RPL and primary or secondary antiphospholipid syndrome [8,9]. On this topic also a rare condition as acquired deficiency of clotting factor XII has been described. Braulke et al. identified for the first time a factor XII deficiency in RPL [10], but subsequently Jones et al. reported acquired factor XII deficiency in a subpopulation of women with antiphospholipid antibodies and RPL [11-13]. We here report a really interesting case report of woman affected by unexplained RPL, prolonged activated partial thrombplastin time and mild/moderate reduction of clotting factor XII.
Case presentation
A 34-year-old Caucasian non smoking woman was referred to our Sterility Center. Her personal anamnesis revealed three early pregnancy loss within 8 and 12 week of gestation and one extrauterine pregnancy. The patient did not revealed previous thromboembolic disease (arterial or venous) nor haemorragic disorders; moreover patient was not ongoing any type of pharmacological treatment. A thorough familial anamnesis did not show a trend toward thromboembolic and/or haemorragic disease.
To understand pathophysiology of her RPL the patient performed several laboratory and instrumental tests.
A normal ovarian function and ovulation were detected by normal values of Follicle-stimulating Hormone, Luteinising Hormone, oestradiol and progesterone and by ovarian ultrasound scan (data not shown). Uterine and salpinxes malformation was excluded by hysterosalpingography and hysteroscopy (data not shown). Endocrinological diseases such as diabetes and dysthyroidism were evaluated and excluded by normal values of glycaemia, thriiodothyronine (i.e. FT3), thyroxine (i.e. FT4) and Thyroid-stimulating Hormone (data not shown). Inflammatory chronic diseases were excluded by normal values of erythro-sedimentation rate and acute phase C reactive protein and immunopathological chronic disease, such as erytematosus systemic lupus, by normal level of antinuclear antibodies (ANA), anitmithocondrial antibodies (AMA) and smooth muscle antibodies (SMA) too (data not shown); moreover patient did not suffer of chronic joint paint or fever or other related symptoms.
Yet, routine haemostatic tests showed normal value of prothrombin time, measured as International Normalised Ratio (PT INR, 1.15) and a prolonged activated thromboplastin time, measured as ratio (aPTT, 1.45) (table 1). To confirm this laboratory alteration, after 15 days a second step of haemostatic parameters were tested and confirmed normal value of PT INR and prolonged aPTT (1.46, table 1 and 2), associated to normal levels of anticardiolipin antibodies (tested by an ELISA method; IgM 1.7 U/MPL and IgG 3.9 U/GPL) (table 1 and 2), fibrinogen and clotting factors V, VIII, IX, X, XI (table1), while a reduced plasmatic level of clotting factor XII was detected (65%, table 1 and 2). So, we tested for aPTT and factor XII first degree relatives (i.e. one sister and parents) but did not find alterations (data not shown). Yet, lupus anticoagulant and antibodies to β-2-glycoprotein I were absent (table 1 and 2). Furthermore, because RPL was associated to reduced clotting inhibitors also protein C, protein S and antithrombin III were analysed and resulted in normal range (table 1) such as inherited thrombophilia associated to Factor V Leiden and prothrombin A20210G gene polymorphisms (table 1). Moreover, to confirm factor XII deficiency, after one month the patient was evaluated again and results showed again prolonged aPTT (1.48), reduced factor XII (55%). So, we tested aPTT adding to the plasma patient's a normal plasma sample with a ratio of plasma patient's to pool plasma 1/1 v/v; results revealed a partial correction of aPTT (1.30) suggesting a possible presence of an acquired clotting inhibitor and/or antiphospholipid antibodies. A new evaluation of common antiphospholipid antibodies began with normal levels of anticardiolipin antibodies (tested again with an ELISA method; IgM 1.5 and IgG 3.8) and absence of antibodies to β-2-Glycoprotein I, but a positivity for lupus anticoagulant was detected (table 2). Lupus anticoagulant was assayed according to recommendations of the International Society of Thrombosis and Haemostasis. Furthermore, at this time we tested again ANA that resulted again negative. Finally, because a transient positivity of lupus anticoagulant has been associated with factor XII deficiency in patients with antiphospholipid syndrome in rare cases [14,15], after one month the patient was evaluated one more time and in this time also antibodies to clotting factor XII were evaluated by an ELISA method. Results showed prolonged aPTT (1.44), reduced factor XII plasma levels (43%), normal range of anticardiolipin antibodies (IgM and IgG), absence of antibodies to β-2-Glycoprotein I, negativity for lupus anticoagulant, while antibodies to factor XII have been detected (table 2). So, according to our data and data available in the Literature the patient was diagnosed an unusual form of primary antiphospholipid syndrome and an antithrombotic treatment was suggested.
Discussion
Inherited factor XII deficiency seems to be not associated to bleeding tendency if referred to major surgery [16]. However, some individual with reduction of factor XII seems to have a trend toward thrombosis [16], but pathophysiological mechanisms underlying should be better understood. Therefore, deficiency of clotting factor XII has been reported as also thrombotic risk factor and Halbmayer et al. suggested to consider factor XII deficiency in patients with recurrent thromboembolism [17]. One of the possible mechanisms could be associated to the presence of acquired clotting inhibitors eventually associated to the presence of an antiphospholipid syndrome [18]. Acquired clotting inhibitors, in fact, have been identified in several disease and frequently are associated to a thrombophilic trend [19].
Furthermore, from a clinical point of view, alteration of haemostasis with a trend toward thrombophilia has been frequently associated to RPL [2,8]. According to a multivariate analysis on the etiology of thrombophilia by Yamada et al., clotting factor XII deficiency has been found in 4.2% of women affected by RPL [20]. Another interesting study by Iinuma et al. underlined clotting factor XII activity and not its 46C/T gene polymorphism as cause of RPL in a selected population [21]. However, factor XII deficiency may be due to inherited deficiency or acquired deficiency during an acquired disease such as antiphospholipid syndrome. In this last condition, in fact, we may find an acquired factor XII deficiency because the presence of antibodies to factor XII; moreover antibodies to factor XII may be present during antiphospholipid syndrome alone or together to lupus anticoagulant [11]: antibodies to factor XII have been described, in fact, also in patients with antiphospholipid syndrome and transient lupus anticoagulant [14,15]. Moreover, Jones et al. described that antibodies to factor XII can be present in women with RPL and antiphospholipid syndrome more than anticardiolipin antibodies and antibodies to β-2-glycoprotein I [11]. So, according to these data, antibodies to factor XII may be implicated in the pathophysiology of the hypercoagulable state in women with antiphospholipid syndrome showing RPL and their incidence in this clinical setting could be underestimated according to the data available from the Literature [18]. This hypothesis is really intriguing also in the case we described.
The reported patient affected by RPL, in fact, did not show an apparent cause of RPL other than reduction of clotting factor XII. Also laboratory tests, concerning thrombophilia, did not show common alteration of haemostasis associated to RPL such as factor V Leiden gene polymorphism or reduction of clotting inhibitors (i.e. protein C, protein S, antithrombin III). Yet, the clinical presentation and the presence of persistent prolonged aPTT suggested periodical tests concerning possible clotting factors deficiency and/or the presence of antiphospholipid syndrome. Results of this screening confirmed us the partial and acquired factor XII deficiency, due to the presence of antibodies to factor XII, in presence of transient lupus anticoagulant.
So, in conclusion we suggest to search in such cases also antibodies directed to clotting factors (e.g. factor XII in our case) as second step of thrombophilia screening in RPL, in particular if a persistent prolonged aPTT is present (figure 1) without an apparent cause.
Acknowledgements
Authors thank dr. Paola Ferrazzi, Centro Tormbosi, Istituto Clinico Humanitas, Rozzano, Milan, Italy, for her helpful suggestions in writing the case; further thanks are due to Prof. Biagio Di Micco, Biochimica Clinica, Università del Sannio, Benevento, Italy for his help to perform haemostatic assays.
Figures and Tables
Figure 1 text
Table 1 Screening for disorder of haemostasis in the patient with RPL.
Parameters (unit of measurement) Results Normal range
Prothrombin time (INR) 1.15 0.8–1.2
Activated partial thromboplastin time (ratio) 1.45 0.8–1.2
Fibrinogen (mg/dL) 275 220–420
Protein C (%) 93 60–125
Protein S (%) 92 60–125
Antithrombin III (%) 104 80–120
Anticardiolipin Ab IgM (U/GPL) 1.7 < 2.0
Anticardiolipin Ab IgG (U/MPL) 3.9 < 7.0
β-2-GP I Ab Absent Absent
Lupus anticoagulant Absent Absent
Factor XII (%) 65 80–120
Factor XI (%) 113 80–120
Factor X (%) 112 80–120
Factor IX (%) 99 80–120
Factor VIII (%) 88 65–155
Factor V (%) 110 80–120
MTHFRC677T gene polymorphism Wild type Wild type
PTHRA20210 gene polymorphism Wild Type Wild type
FVL gene polymorphism Wild Type Wild type
INR: international normalised ratio
MTHFRC677T: methylene-tetra-hydrp-folate reductase C677T gene polymorphism
PTHRA20210G: prothrombin A20210G gene polymorphism
FVL: factor V Leiden gene polymorphism
β-2-GP I Ab: Antibodies to β-2-glycoprotein I
Table 2 Monitor of the alteration of haemostasis in the patient with RPL.
Parameters (Unit of measurement) First screening Second screening Third screening Fourth screening Normal value
aPTT (ratio) 1.45 1.36 1.28 1.44 0.8–1.2
Factor XII (%) 65 50 55 43 80–120
Anticardiolipin Ab IgM (U/MPL) 1.7 1.3 1.5 1.5 < 2.0
Anticardiolipin Ab IgG (U/GPL) 3.9 5.4 3.8 2.5 < 7.0
β-2-GP I Ab Absent Absent Absent Absent Absent
Lupus anticoagulant Absent Absent Present Absent Absent
Anti-factor XII Ab Not tested Not tested Not tested Present Absent
Antinuclear antibodies (ANA) Absent Not tested Not tested Absent Not tested
aPTT: activated partial thromboplastin time
ab: antibodies
β-2-GP I Ab: Antibodies to β-2-glycoprotein I
==== Refs
Di Micco P D'Uva M Strina I Mollo A Amato V Niglio A De Placido G The role of d-dimer as first marker of thrombophilia in women affected by sterility: implications in pathophysiology and diagnosis of thrombophilia-induced sterility J Transl Med 2004 2 38 15535889 10.1186/1479-5876-2-38
Brenner B Sarig G Weiner Z Younis J Blumenfeld Z Lanir N Thrombophilic polymorphisms are common in women with fetal loss without apparent cause Thromb Haemost 1999 82 6 9 10456445
Sanson BJ Fierich PW Simioni P Zanardi S Hilsman MV Girolami A ten Cate JW Prins MH The risk of abortion and stillbirth in antithrombin-, protein C, and protein S deficient women Thromb Haemost 1996 75 387 388 8701393
Grandone E Margaglione M Colaizzo D d'Addedda M Cappucci G Vecchione G Factor V Leiden is associated with repeated and recurrent unexplained fetal losses Thromb Haemost 1997 77 822 824 9184385
Martinelli I Taioli E Cetin I Marinoni A Gerosa S Villa MV Bozzo M Mannucci PM Mutations in coagulation factors in women with unexplained late fetal loss N Engl J Med 2000 343 1015 1018 11018168 10.1056/NEJM200010053431405
Brenner B Inherited thrombophilia and pregnancy loss Thromb Haemost 1994 82 634 640 10605761
Dossenbach-Glaninger A von Trotsenburg M Krugluger W Dossenbach MR Oberkanins C Huber J Hompeier P Elevated coagulation factor VIII and the risk for maternal early pregnancy loss Thromb Haemost 2004 91 694 699 15045130
Asherson RA Cervera R "Primary", "secondary" and other variants of antiphospholipid syndrome Lupus 1994 3 293 298 7804319
Noble LS Kutteh WH Lashey N Franklin RD Herrada J Antiphospholipid antibodies associated with recurrent pregnancy loss: prospective, multicenter, controlled pilot study comparing treatment with low molecular weight heparin versus unfractioned heparin Fertil Steril 2005 83 684 690 15749498 10.1016/j.fertnstert.2004.11.002
Braulke I Priggmayer M Melloh P Hinney B Kostering H Gunther H Factor XII (Hageman) deficiency in women with abitual abortion: new subpopulation of recurrent aborters? Fertil Steril 1993 53 98 101 8419231
Jones DW Mackie IJ Gallimore MJ Winter M Antibodies to factor XII and recurrent fetal loss in patients with the anti-antiphospholid syndrome Br J Haematol 2001 113 550 552 11380430 10.1046/j.1365-2141.2001.02776.x
Jones DW Gallimore MJ Winter M Recurrent abortion-antibodies to factor XII or decrease in factor XII levels? Fertil Steril 2001 76 1288 1289 11730776 10.1016/S0015-0282(01)02917-X
Jones DW Gallimore MJ Winter M Antibodies to factor XII: a possible predictive marker for recurrent fetal loss Immunobiology 2003 207 43 46 12638902 10.1078/0171-2985-00207
Jones DW Gallimore MJ Harris SL Winter M Antibodies to factor XII associated to lupus anticoagulant Thromb Haemost 1999 81 387 390 10102466
Jaeger U Kapiotis S Pabinger I Puchhammer E Kyrle PA Lechner K Transient lupus anticoagulant associated with hypoprothrombinaemia and factor XII deficiency following adenovirus infection Ann Hematol 1993 67 95 99 8394145 10.1007/BF01788133
Hathaway WE Goodnight HS Jr Hathaway WE, Goodnight HS Jr Contact factor deficiencies Disorders of hemostasis and thrombosis A clinical guide 1993 New York: McGraw-Hill 147 154
Halbmayer WM Mannhalter C Feichtinger C Rubi K Fisher M The prevalence of factor XII deficiency in 103 orally anticoagulated outpatients suffering from recurrent venous and/or arterial thromboembolism Thromb Haemost 1992 68 285 290 1440493
Jones DW Gallimore MJ MacKie IJ Harris SL Winter M Reduced factor XII levels in patients with antiphospholipid syndrome are associated with antibodies to factor XII Br J Haematol 2000 110 721 726 10997986 10.1046/j.1365-2141.2000.02251.x
Chalkiadakis G Kyriakou D Oekonomaki D Tsiaoussis J Alexandrakis M Vasilakis S Kouroumalis E Acquired inhibitors to the coagulation factor XII associated with liver disease Am J Gatroenterol 1999 94 2551 2553 10.1111/j.1572-0241.1999.01317.x
Yamada H Kato EH Kobashi G Ebina Y Shimada S Morikawa M Yamada T Sakuragi N Fujimoto S Recurrent pregnancy loss: etiology of thrombophilia Semin Thromb Haemost 2001 27 121 129 10.1055/s-2001-14070
Iinuma Y Sugiura-Ogasawara M Makino A Ozaki Y Suzumori N Suzumori K Coagulation factor XII activity, but not an associated genetic polymorphism (46C/T), is linked to recurrent miscarriage Fertil Steril 2002 77 353 356 11821096 10.1016/S0015-0282(01)02989-2
|
16324213
|
PMC1322227
|
CC BY
|
2021-01-04 16:19:32
|
no
|
J Transl Med. 2005 Dec 2; 3:43
|
utf-8
|
J Transl Med
| 2,005 |
10.1186/1479-5876-3-43
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-901633665810.1186/1743-422X-2-90ResearchPhylogenetic evidence for the distinction of Saaremaa and Dobrava hantaviruses Sironen Tarja [email protected] Antti [email protected] Alexander [email protected] Department of Virology, Haartman Institute, Haartmaninkatu 3, FIN-00014 University of Helsinki, Finland2005 8 12 2005 2 90 90 27 6 2005 8 12 2005 Copyright © 2005 Sironen et al; licensee BioMed Central Ltd.2005Sironen 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.
Dobrava virus (DOBV) and Saaremaa virus (SAAV) are two closely related hantaviruses carried by different rodent species. The distinction of these two viruses has been a matter of debate. While the phylogenies based on the viral M segment sequences were repeatedly showing monophyly of SAAV strains, some trees based on the S segment sequences were not, thus causing questions on the demarcation between these two viruses. In order to clarify this issue, the current collection of the virus S segment sequences was subjected to extensive phylogenetic analysis using maximum likelihood, maximum parsimony and distant matrix methods. In all inferred phylogenies, the SAAV sequences were monophyletic and separated from DOBV sequences, thus supporting the view that SAAV and DOBV are distinct hantavirus species. Since collection of the S segment sequences used in this study "obeyed" the molecular clock, calculations of the split of DOBV and SAAV were now repeated resulting in an estimation of 3.0–3.7 MYA that is very close to the values obtained earlier.
==== Body
Background
Hantaviruses (genus Hantavirus, family Bunyaviridae) are enveloped viruses with a segmented, single-stranded RNA genome of negative polarity [1]. The large (L) segment encodes the viral RNA polymerase, the medium (M) segment the two surface glycoproteins, and the small (S) segment the nucleocapsid protein (N). Hantaviruses cause two human zoonoses, hemorrhagic fever with renal syndrome (HFRS) in Eurasia and hantavirus pulmonary syndrome (HPS) in the Americas [reviewed in [2]]. DOBV is carried by yellow-necked mouse (Apodemus flavicollis) and is associated with severe HFRS in the Balkans (Slovenia, Albania and Greece). SAAV is carried by striped field mouse (A. agrarius) [3]. So far, the virus has been found in Estonia, the European part of Russia, Slovakia, Slovenia, Hungary, Denmark and Germany [2].
SAAV was initially called an A. agrarius-carried variant of Dobrava virus [3], but the accumulating data suggest that the virus should be regarded as a distinct hantavirus species. It is carried by a specific rodent host [3], there is a four-fold difference in two-way cross-neutralization tests [4], and the coexistence of SAAV and DOBV in the same geographic region [5,6] indicates reproductive isolation. They also exhibit 6.1–6.3% difference in the glycoprotein precursor amino acid sequences. This level is a fraction lower than the officially accepted 7% cut-off value [1]. It should be mentioned that some of the officially approved, distinct hantavirus species show lower than 7% diversity in their N or GnGc-sequences: Sin Nombre and New York viruses, Topografov and Khabarovsk viruses, Rio Mamore and Laguna Negra viruses, and Blood Land Lake and Prospect Hill viruses [7].
SAAV and DOBV also exhibit only 3% diversity on their N protein sequences. This unusually low level of diversity is most probably a reflection of host switching in their evolution [8,9]; this event seems to be historically recent (2.7–3.4 MYA) and these two viruses are still diverging [8]. There is another important feature differentiating DOBV and SAAV, and that is the apparently different pathogenicity in humans: while DOBV causes severe HFRS in humans, SAAV causes a milder form of the disease, similar to nephropathia epidemica [2]. This difference is also reflected in different pathogenicity in suckling mice: DOBV is lethal to suckling mice, while SAAV is not [10].
The phylogenetic distinction of SAAV and DOBV was recently a matter of debate [11,12]. While the phylogenies based on the M segment/GnGc protein sequences were repeatedly showing monophyly of SAAV strains, some trees based on the S segment/N protein sequences were not [[11,13], and our unpublished observations], thus causing questions on the demarcation between these two viruses. In order to clarify this issue, the current collection of DOBV and SAAV S segment sequences was subjected to extensive phylogenetic analysis. Especially important additions to the dataset include an A. agrarius -derived SAAV strain from Denmark, Saaremaa/Lolland/Aa1403/2000 [AJ616854), and two DOBV sequences from southern Russia, P-s1223/Krasnodar-2000 (AF442623) and As-1/Goryachiy Klyuch-2000 (AF442622). Our earlier data indicated that these sequences could be helpful for resolving the S phylogeny [14].
Results and discussion
Our analysis was restricted to nt 37–1232 of the S segment available for all the strains. This part of the S segment includes almost complete coding region for the N protein. Accession numbers for the sequences are given in Table 1.
Table 1 Sequences used in the analysis
Strain Accession number
Saaremaa virus (SAAV) Saaremaa/160 V AJ009773
90Aa/97 AJ009775
Lolland/Aa1403/2000 AJ616854
Kurkino/44Aa/98 AJ131672
Kurkino/53Aa/98 AJ131673
East Slovakia/856/Aa AJ269549
East Slovakia/862/Aa AJ269550
Dobrava virus (DOBV) Slovenia L41916
East Slovakia/400Af/98 AY168576
Ano-Poroia/9Af/1999 AJ410615
Ano-Poroia/13Af/99 AJ410619
As-1/Goryachiy Klyuch-2000 AF442622
P-s1223/Krasnodar-2000 AF442623
Seoul virus (SEOV) Gou3 AB027522
L99 AF288299
Z37 AF187082
SR11 M34881
Hantaan virus (HTNV) Ah09 AF285264
84Fli AY017064
76–118 M14626
Lr1 AF288294
Andes virus (ANDV) AH-1 AF324902
Topografov virus (TOPV) Ls136V AJ011646
Sin Nombre virus (SNV) NM H10 L25784
El Moro Canyon virus (ELMCV) RM-97 U11427
Puumala virus (PUUV) Sotkamo X61035
Tula virus (TULV) Moravia/5302v/95 Z69991
Table 2 Bootstrap and puzzle support values for DOBV and SAAVclades in phylogenetic trees calculated using different methods.
method outgroup support for: DOBV support for: SAAV
maximum likelihood SEOV 100 70
maximum likelihood collection* 100 49
maximum likelihood no outgroup 100 100
maximum parsimony SEOV 100 75
maximum parsimony collection* 100 75
distance matrix: Neighbor-joining SEOV 100 84
distance matrix: Neighbor-joining collection* 100 91
distance matrix: Fitch-Margoliash SEOV 79 58
distance matrix: Fitch-Margoliash collection* 100 79
distance matrix: Fitch-Margoliash no outgroup 100 99
TreePuzzle** SEOV 99 87
TreePuzzle collection* 99 75
*A collection of hantavirus sequences including SNV, ANDV, ELMCV, TULV, TOPV, PUUV, SEOV strains SR11 and Gou3, HTNV strains 76–118 and 84Fli **Tamura-Nei was used as the nucleotide (nt) substitution model in TreePuzzle, as suggested by Modeltest.
Since recombinant sequences might influence phylogenetic reconstructions (e.g. by "breaking" the molecular clock [15]), we wanted to check whether the sequences used in this study included any recombinants ones. A similarity plot (Stuart Ray's SIMPLOT2.5) was created in order to visualize the pattern of similarity between the DOBV and SAAV S segment nucleotide sequences, and phylogenetic trees were created on partial sequences, that were possibly of recombinant origin. Although we have found some indications on a recombinant origin of the strain Lolland (in particular, nt 200–460 were most similar to the Estonian SAAV strains, while other regions, especially nt 1150–1450, were more similar to SAAV strains from Russia and Slovakia), they were not unequivocal. For instance, the SIMPLOT data were not mirrowed by a mosaic-like pattern of the N protein sequence of Lolland strain. Moreover, the presence of this sequence did not "break" the molecular clock (see below). The Lolland sequence was, therefore, not excluded from our data set.
Next, we wanted to study whether the new additional sequences would have any effect on the clustering of DOBV and SAAV. A phylogenetic tree was re-calculated with the same collection of sequences and same parameters as has been done by Klempa et al. [11] (Fig. 1). The additional DOBV and SAAV sequences were then included to this set, a new phylogenetic tree was created, and indeed, a change in the topology was seen. The SAAV sequences turned monophyletic with a puzzle support of 71% (Fig. 2).
Figure 1 Phylogenetic tree created with TreePuzzle for a smaller data set. The tree is based on the nt 37–1232 of the S segment sequences.
Figure 2 Phylogenetic tree created with TreePuzzle for a more representative data set. The tree is based on the nt 37–1232 of the S segment sequences. Two SAAV sequences that are placed differently on the trees shown on Fig. 1 and Fig 2 are underlined.
In order to confirm the phylogeny, trees were calculated using different algorithms listed earlier (Table 2). All methods agreed on placing DOBV and SAAV sequences into their own clusters. Placing of the two above mentioned DOBV sequences derived from southern Russia was more variable, but in most cases they were sharing a common ancestor with the other DOBV strains. The puzzle support values and bootstrap support for the DOBV cluster were in most cases very high (79–100%). For SAAV, the support was more variable, but only in two out of 12 phylogenies below the widely accepted confidentiality limit (70%) [16]. The support values were also varying depending on the phylogenetic algorithm, on the parameters used, and on the sequences chosen as outgroup. In the case of maximum likelihood trees, the use of additional hantavirus sequences as outgroup resulted in a lower bootstrap support for SAAV. In fact, a 100% support for SAAV monophyly was reached, when no outgroup sequences were used at all. This algorithm goes through an exhaustive search of all the possible trees, and it is possible that additional information creates an interfering noise to the phylogenetic signal. The opposite was happening with Fitch-Margoliash distance-matrix method. As more sequences were added, the bootstrap support for SAAV was increasing, most probably due to more accurate distance estimations. Nevertheless, in every tree, all the SAAV sequences were monophyletic and separated from DOBV. It should be stressed that bootstrap or puzzle support values do not estimate accuracy of a tree (i.e. right topology), but precision (how many trees had to be rejected) [17]. Phylogenies inferred here with different algorithms, and by varying the parameters used in the analyses (Table 2), gave a consensus answer on the monophyly of all SAAV strains, thus suggesting that this tree topology is most accurate.
Earlier it has been estimated, that the split of DOBV and SAAV happened 2,7–3.4 million years ago (MYA) (10). Since the larger collection of the S segment sequences used in this study "obeyed" the molecular clock, these calculations were now repeated resulting in an estimation of 3.0–3.7 MYA.
Conclusion
In all phylogenies inferred in this study using different approaches such as maximum likelihood, maximum parsimony and distant matrices, the SAAV sequences were monophyletic and separated from DOBV sequences, thus supporting the view that SAAVand DOBV are distinct hantavirus species.
Methods
Sequences were handled with BIOEDIT [18], and alignments were created using CLUSTALX [19]. The various methods used for phylogenetic analysis included maximum likelihood ("classic" maximum likelihood from PHYLIP [20] and TreePuzzle [21], maximum parsimony (PHYLIP) and distance matrix methods Neighbor joining and Fitch-Margoliash (PHYLIP). 500 boostrap replicates were used in PHYLIP programs and 10000 puzzling steps in TreePuzzle. MODELTEST and PAUP were used to check, which DNA substitution model would fit best to this data set [22,23]. The test for molecular clock and estimation of the time of split of these two viruses was done with TreePuzzle [21].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TS carried out experiments, participated in the analysis of the results and drafted the manuscript. AV participated in the analysis of the results and helped to draft the manuscript. AP designed the study, participated in the analysis of the results and helped to draft the manuscript.
==== Refs
Elliott RM Bouloy M Calisher CH Goldbach R Moyer JT Nichol ST Pettersson R Plyusnin A Schmaljohn CS van Regenmortel MHV, Fauquet CM, Bishop DHL, Carsten EB, Estes MK, Lemon SM, Maniloff J, Mayo MA, McGeoch DJ, Pringle CR, Wickner RB Family Bunyaviridae Virus taxonomy VIIth report of the International Committee on Taxonomy of Viruses 2000 San Diego: Academic Press 599 621
Vapalahti O Mustonen J Lundkvist Å Henttonen H Plyusnin A Vaheri A Hantavirus infections in Europe Lancet 2003 3 653 661
Nemirov K Vapalahti O Lundkvist Å Vasilenko V Golovljova I Plyusnina A Niemimaa J Laakkonen J Vaheri A Plyusnin A Isolation and characterization of Dobrava hantavirus carried by the striped field mouse (Apodemus agrarius) in Estonia J Gen Virol 1999 80 371 379 10073697
Brus-Sjölander K Golovljova I Plyusnin A Lundkvist Å Serological divergence of Dobrava and Saaremaa hantaviruses: evidence for two distinct serotypes J Epidemiol Infect 2002 128 99 103 10.1017/S095026880100632X
Avsic-Zupanc T Nemirov K Petrovec M Trilar T Poljak M Vaheri A Plyusnin A Genetic analysis of wild-type Dobrava hantavirus in Slovenia: co-existence of two distinct genetic lineages within the same natural focus J Gen Virol 2002 81 1747 1755 10859380
Sibold C Ulrich R Labuda M Lundkvist Å Martens H Schutt M Gerke P Leitmeyer K Meisel H Krüger DH Dobrava hantavirus causes hemorrhagic fever with renal syndrome in central Europe and is carried by two different Apodemus mice species J Med Virol 2001 63 158 167 11170053 10.1002/1096-9071(20000201)63:2<158::AID-JMV1011>3.0.CO;2-#
Plyusnin A Genetics of hantaviruses: implications to taxonomy (review) Arch Virol 2002 147 665 682 12038679 10.1007/s007050200017
Nemirov K Henttonen H Vaheri A Plyusnin A Phylogenetic evidence for host switching in the evolution of hantaviruses carried by Apodemus mice Virus Res 2002 90 207 215 Erratum 2003, 92:125–126 12457975 10.1016/S0168-1702(02)00179-X
Wang H Yoshimatsu K Ebihara H Ogino M Araki K Kariwa H Wang Z Luo Z Li D Hang C Arikawa J Genetic diversity of hantaviruses isolated in china and characterization of novel hantaviruses isolated from Niviventer confucianus and Rattus rattus Virology 2000 278 332 345 11118357 10.1006/viro.2000.0630
Klingström J Hardestam J Lundkvist Å Dobrava, but not Saaremaa, hantavirus is lethal and induces nitric oxide production in suckling mice Microbes and Infection 2005
Klempa B Schmidt HA Ulrich R Kaluz S Labuda M Meisel H Hjelle B Krüger DH Genetic interaction between distinct Dobrava hantavirus subtypes in Apodemus agrarius and A. flavicollis in nature J Virol 2003 77 804 809 12477889 10.1128/JVI.77.1.804-809.2003
Plyusnin A Vaheri A Lundkvist Å Genetic interaction between Dobrava and Saaremaa hantaviruses: now or millions of years ago? J Virol 2003 77 7156 7157 12768038 10.1128/JVI.77.12.7156-7158.2003
Plyusnin A Krüger DH Lundkvist Å Hantavirus infections in Europe. (Review) Adv Vir Res 2001 57 105 136
Nemirov K Andersen HK Leirs H Henttonen H Vaheri A Lundkvist Å Plyusnin A Saaremaa hantavirus in Denmark J Clin Virol 2004 30 254 257 15135745
Schierup MH Hein J Recombination and the molecular clock Mol Biol Evol 2000 17 1578 1579 11018163
Hillis DM Bull JJ An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis Syst Biol 1993 42 182 192
Page RDM Holmes EC Inferring molecular phylogeny Molecular evolution: a phylogenetic approach 1998 UK: Blackwell Science Ltd 216 225
Hall T BioEdit. Biological sequence alignment editor for Windows 1998 North Carolina State University, NC, USA
Thompson JD Gibson TJ Plewniak F Jeanmougin F Higgins DG The CLUSTAL X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools Nucl Acids Res 1997 25 4876 4882 9396791 10.1093/nar/25.24.4876
Felsenstein J PHYLIP – Phylogeny Inference Package (Version 3.2) 1989
Strimmer K von Haeseler A Quartet puzzling: A quartet maximum likelihood method for reconstructing tree topologies Mol Biol Evol 1996 13 964 969
Posada D Crandall KA MODELTEST: testing the model of DNA substitution Bioninformatics 1998 14 817 818 10.1093/bioinformatics/14.9.817
Swofford DL PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4 2003 Sinauer Associates, Sunderland, Massachusetts
|
16336658
|
PMC1322228
|
CC BY
|
2021-01-04 16:38:57
|
no
|
Virol J. 2005 Dec 8; 2:90
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-90
|
oa_comm
|
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-791632975610.1186/1477-7525-3-79ResearchThe reliability, validity and responsiveness of the Restless Legs Syndrome Quality of Life questionnaire (RLSQoL) in a trial population Abetz Linda [email protected] Robert [email protected] Richard P [email protected] Elena [email protected] Jeffrey [email protected] Mapi Values, Adelphi Mill, Bollington, Macclesfield, SK10 5JB, UK2 Johns Hopkins Bayview Medical Center, Neurology and Sleep Medicine, Asthma and Allergy Building 1B46b, 5501 Hopkins Bayview Circle, Baltimore, MD 21224, USA3 GlaxoSmithKline, Greenford Road, Greenford, Middlesex, UB6 0HE, UK2005 5 12 2005 3 79 79 24 6 2005 5 12 2005 Copyright © 2005 Abetz 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 aim of this study was to determine the reliability, validity and responsiveness of the Restless Legs Syndrome Quality of Life questionnaire (RLSQoL) in a clinical trial setting.
Methods
Two matching, placebo-controlled, multinational studies assessing the effectiveness and safety of ropinirole for treating moderate-to-severe Restless Legs Syndrome (RLS) formed the basis of this psychometric assessment. Validity and reliability were assessed using baseline data. Responsiveness was determined using longitudinal data collected at baseline and 12 weeks.
Results
A total of 547 subjects formed the baseline validation population; 519 were used for assessing responsiveness (n = 284/263 and 271/248 for both studies, respectively). Construct validity assessment confirmed that an overall life impact score could be calculated. All item-scale correlations were = 0.4, except items 1 (r = 0.36) and 5 (r = 0.35) in one study. Floor and ceiling effects were minimal. Cronbach's alpha values were 0.82 and 0.87, respectively, confirming internal consistency reliability. Correlations with the International Restless Legs Syndrome Study Group's severity rating scale (International Restless Legs Scale; IRLS) were moderate (r = -0.68 and -0.67, respectively; p < 0.0001). The RLSQoL was able to discriminate between levels of sleep problems (p < 0.0001) and between levels of global health status determined by a Clinical Global Impression of severity (CGI-S) (p < 0.0001). Responsiveness was demonstrated by significant differences in overall life impact change scores between CGI improvement levels after 12 weeks (p < 0.0001).
Conclusion
The RLSQoL is a valid, reliable and responsive measure of quality of life for patients with RLS, in a clinical trial setting where group comparisons are anticipated.
==== Body
Background
Restless Legs Syndrome (RLS) is a debilitating neurological disorder characterized by a strong urge to move the limbs, usually accompanied by unpleasant sensations, such as creeping, crawling, tingling or pain, particularly when the person lies down or sits for prolonged periods. In most patients, these sensations are felt in the legs, but they may also occur in the arms or trunk [1]. Both the urge to move and these sensations represent the primary symptoms of RLS. Movement brings almost immediate but variable relief from the symptoms, and this relief is maintained as long as the movement continues. If patients ignore the urge to move their legs, there may be intensification of the symptoms until the urge is satisfied, either voluntarily or involuntarily [2]. As a result, RLS can have serious adverse effects on sleep, leaving patients with reduced sleep time and daytime fatigue, reduced concentration and decreased motivation, which in some cases can lead to depression and anxiety [3,4].
The prevalence of RLS in the general population has been reported to range from 5% to 10% [5] with an increasing prevalence with age, and a somewhat higher rate in women than in men [2,6]. There is often a family history of RLS, particularly among patients who present with RLS before 45 years of age [7]. However, primary RLS is a recognized condition that tends to co-occur in the same family. Recently, possible genetic linkages have been established [8-10]. Secondary RLS has been reported to occur in association with a number of conditions, including pregnancy, end-stage renal failure and iron-deficiency anemia [2]. There are also a number of differential diagnoses to be excluded, including leg cramps or paraesthesias, and hypotensive akathisia.
A number of agents are used for the treatment of patients with RLS, including dopaminergic agents and opiates, and treatment varies according to symptom severity and frequency, and the presence or absence of painful symptoms. Although treatments may relieve some or all of the symptoms of RLS, Trenkwalder and colleagues [11] reported that they can also be associated with side effects, including increased symptom severity (augmentation) in the long term. Both the disorder and its treatment may, therefore, have an impact on patients' quality of life. For this reason, the Restless Legs Syndrome Quality of Life questionnaire (RLSQoL), a patient-reported measure of quality of life specific to patients with RLS, was developed.
The RLSQoL has been previously validated in an independent study [12]. Results from that study indicated the reliability and validity of the RLSQoL. However, the study was limited by its relatively small sample size and by the lack of a full assessment of the responsiveness of the RLSQoL to change over time (all subjects maintained their normal treatment regimen throughout the 2-week study, yielding few changes). The aim of the present study, therefore, was to assess the reliability, validity and responsiveness of the RLSQoL in a larger sample size, over a longer period of time (12 weeks) and with treatment intervention. Thus, this study is based on the patient populations of two recently completed, matching, phase-III, multicentre, randomized, double-blind, placebo-controlled studies assessing the efficacy and tolerability of ropinirole, a dopamine agonist, for the treatment of patients with RLS (Therapy with Ropinirole: Efficacy And Tolerability in RLS [TREAT RLS] 1 and 2). Findings from both studies have been published in full elsewhere [11,13].
The primary endpoint in both studies was change in total score of the International Restless Legs Syndrome Study Group's severity rating scale (International Restless Legs Scale, IRLS) [14,15], a clinician-administered report of patient symptom severity. However, the RLSQoL was included as a secondary endpoint. Although the findings of the two separate trials are being presented together in this paper, it should be noted that the analysis for each study was conducted separately.
Methods
Patient populations
Patients were eligible for inclusion in TREAT RLS 1 and 2 if they were at least 18 years of age, had a baseline IRLS total score of ≥ 15 (on a scale of 0 to 40, indicating moderate-to-severe RLS) and had moderate-to-severe frequency of RLS (experienced at least 15 nights with symptoms of RLS in the previous month or, if receiving treatment, had symptoms of this frequency prior to treatment). Patients were excluded from the study if they had any other movement or primary sleep disorder, if they required daytime treatment for RLS, if they were experiencing augmentation or end-of-dose rebound, or if they had secondary RLS associated with end-stage renal disease, iron-deficiency anemia or pregnancy. Patients were also excluded if they had a history of alcohol or drug abuse, previous intolerance to dopamine agonists, or were suffering from other clinically relevant conditions affecting assessments.
All patients gave written, informed consent before entering the studies, which was done according to the principles of the 1996 amendment of the Declaration of Helsinki and approved by local ethics committees.
Study design
Both studies were conducted in a matching double-blind, randomized, placebo-controlled fashion. Patients were recruited from hospitals, sleep centres and neurology clinics in 10 European countries in TREAT RLS 1 (Austria, Belgium, France, Germany, Italy, Netherlands, Norway, Spain, Sweden and the UK) and in six countries around the world in TREAT RLS 2 (Australia, Canada, Germany, Norway, the UK and the USA). After a wash-out phase of generally 5 half-lives or 7 consecutive nights medication-free, whichever was the longer period, patients were randomized to receive once-daily treatment with either ropinirole or placebo for 12 weeks. Other published articles report the study design in greater detail [11,13].
The primary endpoint in both studies was the change in the IRLS total score; secondary endpoints included the RLSQoL overall life impact score, the Medical Outcomes Study Sleep Problems Index II (MOS Sleep Scale) score and Clinical Global Impression (CGI) 'Improvement' (CGI-I) and 'Severity of Illness' (CGI-S) scores.
Outcome measures used in psychometric analysis
RLSQoL
The RLSQoL is a validated questionnaire consisting of 18 items, 13 of which are scored on a 5-point scale, the remainder being recorded as either a numerical value or a dichotomous response [12]. Ten of the items contribute to a single summary score, the overall life impact score, while the remaining eight items concern employment (one question), sexual interest (two questions) and work (five questions), and are summarized individually. Details of the questionnaire and scoring can be found in the appendix (see additional file 1). Higher scores on the RLSQoL overall life impact score indicate a better quality of life. Patients were asked to complete the RLSQoL at baseline and at weeks 8 and 12 of the treatment phase, or at time of withdrawal for patients who discontinued the studies prematurely. A full listing of the items of the RLSQoL is provided elsewhere [12]. The RLSQoL is available on request from Mapi Values.
MOS Sleep Scale
The MOS Sleep Scale is a self-administered scale measuring specific aspects of sleep (problems with sleep disturbance [initiation and maintenance], adequacy, somnolence, quantity, respiratory impairments and snoring) and is reliable and valid in the general US population [16]. It was designed for use in patients who may have varying co-morbidities, and hence is appropriate for a medically diverse patient population. The frequency with which each problem has been experienced during the previous 4 weeks is rated on a 6-point scale ranging from 'none of the time' to 'all of the time', except sleep quantity, which is reported in hours. All scores are transformed linearly to range from 0 to 100, again with the exception of the sleep quantity subscale, which is scored in hours. Higher scores indicate more of the attribute implied by the scale name (e.g. more sleep disturbance, more adequate sleep, greater sleep quantity). Patients were asked to complete the MOS Sleep Scale at baseline and at weeks 8 and 12 of the treatment phase, or at the time of withdrawal for patients who discontinued the study prematurely. The psychometric properties of the MOS Sleep Scale have been found to be satisfactory, both by Hayes and colleagues [17], and within each of the two clinical trial populations used in this study, as recently reported at the 16th Annual Scientific Meeting of the British Sleep Society, Cambridge, UK, 19–21 September 2004.
CGI
The CGI consists of three modules, the CGI 'Improvement' (CGI-I), the CGI 'Severity of Illness' (CGI-S), and the CGI 'Efficacy Index', and has been in use for nearly 3 decades [18]. In the present two studies, only the first two modules were used as outcome measures, although the CGI 'Efficacy Index' was used by the investigators to guide titration of the study medication. The CGI-I and CGI-S modules were assessed by the investigator, based on all information available at the time of rating. Both modules were rated on a scale of 0–7, where 0 refers to patients who were not assessed, 1 indicates 'very much improved' and 7 indicates 'very much worse'. Changes in the proportions of patients with scores of 'much improved' or 'very much improved' were identified as two key secondary endpoints. Both the CGI-I and CGI-S were assessed by the investigators at day 2 and weeks 1, 2, 3, 4, 5, 6, 7, 8 and 12 of the treatment phase, or at the time of withdrawal in patients who discontinued the study prematurely. The CGI-S was also assessed at baseline.
IRLS
The IRLS was developed and validated by the International Restless Legs Syndrome Study Group [14,15]. Subsequent validation studies were conducted for the IRLS using TREAT RLS 1 and TREAT RLS 2 data, and results confirmed the reliability, validity and responsiveness of the IRLS [19]. The IRLS consists of 10 questions concerning the patient's symptoms and the impact of these symptoms on daily activities and mood. Responses range from 0 to 4, with 0 representing the absence of a problem and 4 representing a very severe problem.
The IRLS was completed at the baseline visit, at day 2 and weeks 1, 2, 3, 4, 5, 6, 7, 8 and 12 of the treatment phase and at the follow-up assessment, or at the time of withdrawal for patients who withdrew prematurely.
Analysis
Study populations
In both studies, the intention-to-treat (ITT) population included all randomized patients who received at least one dose of study medication and who had at least one post-baseline efficacy measurement. Patients from the ITT population who had an evaluable RLSQoL (i.e. at least eight non-missing items among items 1–5, 7–10 and 13, as specified by the developer of the questionnaire) were included in the RLSQoL baseline validation population, which was used for all psychometric analyses of the questionnaire, except responsiveness. Patients included in the baseline validation population who also had an evaluable RLSQoL at the 12-week post-baseline visit were included in the longitudinal validation population, which was used for analysis of the responsiveness to change over time of the RLSQoL. All tests were performed on the total population samples, blinded to treatment status.
Psychometric validation analyses
The RLSQoL was assessed for the following psychometric properties: item convergent validity (item-scale correlations of ≥ 0.4) [20], floor and ceiling effects (the percentage scoring the lowest and highest possible scores), internal consistency reliability (Cronbach's alpha ≥ 0.7), concurrent validity, known groups validity, clinical validity and responsiveness.
Assessment of concurrent validity consists of examining the association between the measure being validated, and other well-validated measures, assessing similar constructs. In this instance, concurrent validity was evaluated by assessing correlations between the RLSQoL overall life impact score and the IRLS total score, a clinician-administered patient report of RLS symptom severity. As the RLSQoL has some items related to sleep and somnolence, the correlations of the RLSQoL overall life impact score with the MOS Sleep Scale Sleep Problems Index II was also assessed. Correlations of ≥ 0.40 were considered sufficient evidence of concurrent validity.
The known groups validity of the RLSQoL was assessed by describing and comparing RLSQoL overall life impact scores at baseline among groups of patients with mild, moderate and severe sleep problems, as defined by taking tertile scores for the Sleep Problems Index II of the MOS Sleep Scale. Tertile scores were used because no clinical cut-offs are available for the MOS Sleep Scale. Taking tertile scores involves dividing a normally distributed population into three groups of as close as possible to 33% of patients in each group [21]. Scores of 0–41, 42–56 and 57–100 were considered mild, moderate and severe, respectively. The hypothesis was that patients with more severe sleep problems would have worse quality of life, indicated by lower RLSQoL overall life impact scores.
The clinical validity of the RLSQoL was assessed in two ways. First, the correlation between the RLSQoL overall life impact score and the CGI-S score at baseline was assessed, with a correlation of ≥ 0.40 considered sufficient to confirm validity. Second, RLSQoL overall life impact scores at baseline were compared between severity subgroups defined by dividing the patients into three groups on the basis of their CGI-S scores. The groups were comprised as follows: (1) normal, not at all ill or borderline ill (CGI-S scores 1–2); (2) mild, moderately or markedly ill (CGI-S scores 3–5); (3) severely ill, or among the most extremely ill patients (CGI-S scores 6–7). Patients with a CGI-S score of 0 ('not assessed') were excluded from this analysis. The hypothesis was that for worse clinician-rated overall health status, RLSQoL overall life impact scores would also be worse.
The responsiveness of the RLSQoL to change over time was assessed in two ways. First by examining the correlations between the change in RLSQoL overall life impact scores (between baseline and weeks 8 and 12) and CGI-I scores (1–7) at weeks 8 and 12. Patients with a CGI-I score of 0 ('not assessed') were excluded from the analysis. Second, RLSQoL overall life impact change scores (subtracting baseline from week 12 assessments) were compared among patients defined as improved, unchanged and worsened at week 12, on the basis of their CGI-I scores. Patients with a CGI-I score of 0 ('not assessed') were excluded from the analysis. The effect size (ES) was used as a measure of the change in RLSQoL scores within each CGI-I group. ESs were calculated by dividing the change in mean RLSQoL overall life impact scores (from baseline to week 12) by the standard deviation of mean scores at baseline. The ES has been recommended in the literature as an appropriate benchmark for evaluating the magnitude and meaning of change in health status measures [22].
Cohen and colleagues defined effect sizes of 0.2, 0.5 and 0.8 as small, moderate and large, respectively [23]. We adopted Guyatt et al's guidance that size effects can be described as small, moderate or large when results are in the range of these parameters [24].
Statistics
No adjustments for multiplicity were performed. The Type 1 criterion was 0.05, and all hypothesis tests were two-sided. As tests of homeoscedasticity (equality of dispersion) and normality did not find the analysis data to be normally distributed, non-parametric tests were used for all comparisons. Therefore, Pearson's correlation coefficient was used for all correlations evaluated, the Kruskall-Wallis test was used for comparisons between more than two groups, the Mann-Whitney-Wilcoxon test was used for comparisons between pairs of groups and the Wilcoxon signed rank test was used for comparing two points in time within groups.
Results
Patient populations
TREAT RLS 1 included 284 patients in the baseline validation population, and 229 patients in the week 12 longitudinal validation (responsiveness) population. TREAT RLS 2 included 263 patients in the baseline population and 207 patients in the week 12 longitudinal validation population. Baseline patient characteristics for each study are shown in Table 1. In both studies, the majority of patients were categorized as moderately, markedly or severely ill. The mean age of patients in each study was 55 years and approximately two-thirds of patients in each study were women (63.0% and 59.7%, respectively). There was a slight difference in the mean age of onset in TREAT RLS 1 and TREAT RLS 2: 38 and 35 years, respectively.
Psychometric validation
Missing data
Missing data for TREAT RLS 1 and 2 are summarized in Table 2. In TREAT RLS 1, missing data for each of the items of the RLSQoL ranged from 0% missing data for items 1, 2, 4, 7, 8 and 10 to 3.52% (n = 10) for item 5 ('In the past 4 weeks how often were you late for work or your first appointment due to RLS?'). In TREAT RLS 2, missing data levels ranged from 0% missing data for items 1, 4, 7, 8, 9 and 13 to 1.52% for item 5. Therefore, missing data at the item level were not problematic in either study.
Factor analysis
The current scoring of the RLSQoL, devised by the developers of the questionnaire, suggests calculating a summary score, i.e. the overall life impact score.
A commonly used criterion for calculating a summary score is that the cumulative variance of the first factor in a principal component analysis of ≥ 0.40, although lower values are sometimes also used (e.g. 0.3). In the present two studies, principal component analysis resulted in a cumulative variance of 0.39 in TREAT RLS 1, and 0.46 in TREAT RLS 2. Given that the criterion was surpassed in TREAT RLS 2, and very narrowly missed in TREAT RLS 1, it was considered acceptable for the overall life impact score to be calculated.
The results of the construct validity analysis demonstrated excellent reliability and construct validity for the RLSQoL, as summarized in Table 3.
Item convergent validity
The criterion for item convergent validity (item-scale correlations ≥ 0.40) was satisfied by all items in TREAT RLS 2. In TREAT RLS 1, all except two items met the criterion for item convergent validity. However, both items only narrowly missed the 0.40 threshold with correlations of 0.36 (item 1: 'In the past 4 weeks how distressing to you were your restless legs?') and 0.35 (item 5: 'In the past 4 weeks how often were you late for work or your first appointments of the day due to restless legs?'), respectively.
Internal consistency reliability
Cronbach's alpha coefficients for the RLSQoL overall life impact score were 0.82 and 0.87 in TREAT RLS 1 and 2, respectively, indicating satisfactory internal consistency reliability for the RLSQoL in both trials.
Floor and ceiling effects for the RLSQoL overall life impact score
For the RLSQoL overall life impact score, in TREAT RLS 1, 0.35% (n = 1) of patients scored at floor, 0% scored at ceiling. In TREAT RLS 2, 0% of patients scored at floor and 0.38% (n = 1) scored at ceiling. Therefore, there were no significant floor or ceiling effects for the RLSQoL overall impact score in either study.
Concurrent validity
Correlations of the RLSQoL overall life impact score with the concurrent measures are provided in Table 4. All correlations were above the 0.40 standard set for concurrent validity. In both TREAT RLS 1 and 2, the RLSQoL overall life impact score was moderately correlated at a statistically significant level with both the IRLS total score (r = -0.68 and r = -0.67, respectively) and the Sleep Problems Index II (r = 0.59 and r = 0.60, respectively). These results confirm the concurrent validity of the RLSQoL.
Known groups validity
In both TREAT RLS 1 and 2, the results indicate that the RLSQoL overall life impact scores distinguished between groups of mild, moderate and severe sleep problems at a statistically significant level (p < 0.0001); patients with more severe sleep problems had lower RLSQoL overall life impact scores (poorer quality of life) (Figure 1). Mean overall life impact scores for mild, moderate and severe groups were 73.04, 63.68 and 49.26, respectively, in TREAT RLS 1, and 74.93, 67.44, and 50.54, respectively, in TREAT RLS 2. These results indicate the 'known groups' or discriminative validity of the RLSQoL.
Clinical validity
First, clinical validity was assessed by examining the correlation between RLSQoL overall life impact scores and CGI-S scores. The correlation between RLSQoL overall life impact scores and CGI-S scores was moderate (r = -0.42, p < 0.0001) in TREAT RLS 1, and low but statistically significant (r = -0.33, p < 0.0001) in TREAT RLS 2. In addition, statistically significant differences in RLSQoL overall life impact scores were observed between the three CGI-S subgroups (p < 0.0003) in both TREAT RLS 1 and 2 (Figure 2). Impairment in quality of life due to RLS was greater for groups with worse clinician-rated RLS severity. The differences between pairs of adjacent groups were assessed further using the Mann-Whitney-Wilcoxon test. However, in both studies, the subgroup of patients with CGI-S scores of 1–2 (very mild) was very small, most likely as a consequence of the inclusion criteria for the studies (moderate/severe). Comparisons with this mild group should, therefore, be interpreted with caution. Nevertheless, statistically significant differences in RLSQoL overall life impact scores between the two larger CGI-S subgroups, 3–5 and 6–7, were observed in both studies (p < 0.0001 in both). These findings provide evidence that the RLSQoL is clinically valid in this clinical trial population.
Responsiveness to change over time
When assessing responsiveness, trends for the correlations, change scores and effect sizes were similar at week 8 and week 12. (For brevity, only week 12 results are reported here; week 8 results are available on request.) First, the responsiveness of the RLSQoL to change over time was suggested by moderate and statistically significant correlations of RLSQoL overall life impact change scores with CGI-I scores (r = -0.51, p < 0.0001 in both studies). Investigating this relationship further, statistically significant differences in RLSQoL overall life impact change scores (from baseline to week 12) were observed between groups stratified according to CGI-I scores, in both studies (Tables 5 and 6; p < 0.0001). In both studies, there was a step-wise increase in effect sizes for the no change, and minimally, much and very much improved groups, indicating greater improvements in RLSQoL scores for the more improved CGI-I groups compared with the less improved and no change groups. Effect sizes indicated large improvements in overall life impact scores in the 'very much improved' (ES = 1.51, in both studies), and 'much improved' groups (ES = 1.15 and 1.00 in TREAT RLS 1 and 2, respectively), large and moderate improvements in the 'minimally improved' groups (ES = 0.74 and 0.54, respectively) and moderate or small improvements in the 'no change' group (ES = 0.37 and 0.31, respectively). In TREAT RLS 1 and 2, patient numbers were very small in the 'minimally worse' (5 and 6, respectively), 'much worse' (9 and 0, respectively) and 'very much worse' (0 and 1, respectively) CGI-I groups; therefore, comparisons with these groups cannot be evaluated.
These findings indicate that the RLSQoL is responsive to clinician-rated changes in health status.
Discussion
Based on the results of this psychometric evaluation, the RLSQoL has been found to be reliable, valid and responsive to change in both TREAT RLS 1 and 2. Although the psychometric validity of the RLSQoL has been demonstrated previously [12], due to the evolutionary nature of validation it is important to confirm the psychometrics of a questionnaire when used in different settings and with different populations. The results provide evidence of the psychometric integrity of the RLSQoL within the RLS populations studied, and support its use in patients with RLS, particularly in a clinical trial setting.
The factor analysis results support the calculation of a summary score, the overall life impact score, based on 10 of the 18 items. Although the factor analysis results in TREAT RLS 1 gave a cumulative variance of only 0.39, narrowly missing the 0.40 criterion for calculating a summary score, published research suggests that the cut-off point of 0.40 is 'arbitrary' [25]. Therefore, given that the criterion was surpassed in TREAT RLS 2 (with cumulative variance of 0.46), and only narrowly missed in TREAT RLS 1, it is still considered appropriate to calculate a summary score. This position is further supported by the satisfactory psychometric validation results, and evidence from a previous independent validation study which also concluded that calculating the overall life impact score was valid and appropriate [12].
Both the original validation study and the present research have focused on evaluating the psychometric validity of the overall life impact score: the validity of potential subscales could be investigated in future studies [12]. In addition, sample sizes did not permit factor analyses to be conducted for each country separately; further research could examine differences in scale structure by country or within different age groups.
Taking the findings of both studies together, item convergent validity results were satisfactory. Two items did narrowly miss the criterion (item-scale correlations ≥ 0.40) in TREAT RLS 1. However, given that they were only slightly below the threshold in TREAT RLS 1, and met the criterion in TREAT RLS 2, this was of little concern. The internal consistency reliability of items in the RLSQoL overall life impact score was acceptable in both studies, with Cronbach's coefficients exceeding the accepted standard (≥ 0.70). There were no significant floor or ceiling effects in either study.
In the assessment of concurrent validity, the moderate correlations between the RLSQoL overall life impact score and the IRLS total score indicate that there is some overlap between the RLSQoL and the IRLS, but not so much overlap as to suggest redundancy (> 0.9). The IRLS and RLSQoL assess different concepts (severity and quality of life, respectively) and therefore it is not surprising that the overlap is not greater.
Correlation between the RLSQoL overall life impact scale and the MOS Sleep Scale Sleep Problems Index II was also moderate in both studies, providing evidence that sleep problems do have an impact on quality of life. These results confirm the concurrent validity of the RLSQoL.
In both studies, the RLSQoL overall life impact score was able to distinguish between patients with mild, moderate and severe sleep problems. The results indicate that patients with more severe sleep problems also had lower RLSQoL overall life impact scores (poorer quality of life), thus demonstrating the known groups validity of the RLSQoL overall life impact score.
The findings indicated that as the CGI-S scores increased, RLSQoL overall life impact scores worsened. The fact that the correlation between CGI-S and the RLSQoL overall life impact scores was only low to moderate is unsurprising, given that correlations between doctors' ratings of severity and patient reports of severity are often low to moderate [26,27].
In addition, the RLSQoL overall impact scores were able to distinguish between the three clinician-rated severity groups. The sample sizes in the 'mild' groups were very small, and the results for these groups should, therefore, be interpreted with caution. However, RLSQoL overall life impact scores for the 'moderate' CGI-S groups were statistically significantly different from those for the 'severe' CGI-S groups in both studies. Of note, although these studies did not assess 'mild' RLS patients, the original validation did assess this group (in addition to patients with moderate and severe RLS). The combination of these results indicates the clinical validity of the RLSQoL in mild, moderate and severe groups [12].
The responsiveness of the RLSQoL to change over time was confirmed by comparing change scores from baseline to week 12 with clinicians' perceived changes (CGI-I). Correlation of changes in RLSQoL overall life impact score with CGI-I scores was moderate and statistically significant in both studies. RLSQoL overall life impact change scores were able to distinguish between CGI-I subgroups at a statistically significant level in both studies. RLSQoL scores were improved in patients rated by their clinicians as 'improved', as well as in those patients rated by their clinicians as 'unchanged'. However, effect sizes indicated the improvements were consistently 'large' in those patients rated as 'improved' but only small or moderate in those rated as 'unchanged'. Sample sizes for the worsened groups were small and results for these groups should be interpreted with caution. Improvements in the 'no change' CGI group suggest the potential presence of the 'Hawthorne effect'; that is, a response shift most probably due to positive psychosocial effects of participating in a clinical trial, regardless of the specific nature of any intervention [26].
The sensitivity of the RLSQoL to worsening RLS severity could not be fully evaluated in this study owing to the small sample size of worsening patients, and should be investigated further in a larger sample of worsening patients at the first opportunity. Therefore, it cannot be concluded that the RLSQoL is responsive to worsening until further research is conducted. Furthermore, due to the inclusion criteria for the trials, there were very few patients who were rated by their clinician as being 'normal, not at all ill', 'borderline ill' or 'mildly ill' at baseline (1.27%, 1.27% and 8.18%, respectively, from the CGI-S at baseline). Therefore, further study of the performance of the responsiveness of the RLSQoL in patients with 'mild' RLS is also warranted.
Finally, the focus of this research was on group comparisons; as a result, additional research is warranted to evaluate the usefulness of the RLSQoL in clinical practice to assess patients individually.
Conclusion
In conclusion, the RLSQoL is reliable, valid and responsive to improvements in patients with RLS, in a clinical trial setting. The RLSQoL is short, takes 10 minutes to complete, and covers those aspects of life most impacted by RLS. Other, more generic, questionnaires will not be as relevant to patients and therefore will not be as clinically relevant.
Authors' contributions
LA, EM, JK, and RPA all participated in the design of the study. LA and RA wrote the analysis plan and supervised the analysis. All authors were involved in the interpretation of the data. LA, RA, and JK were involved in drafting the article, which was then revised following critical review by EM and RPA. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Appendix: RLS Quality of Life Questionnaire
Click here for file
Acknowledgements
Both the clinical trials and the psychometric analysis were supported by GlaxoSmithKline Research and Development.
Figures and Tables
Figure 1 Known groups validity. RLSQoL overall life impact scores by mild, moderate and severe sleep problems. p < 0.0001 for comparisons of RLSQoL overall life impact scores among sleep problems severity groups defined by taking tertile scores for the Sleep Problems Index II in TREAT RLS 1 and 2 (Kruskall-Wallis test). RLSQoL = RLS Quality of Life questionnaire.
Figure 2 Clinical validity. Comparison of RLSQoL overall life impact scores at baseline among CGI-S groups. CGI-S subgroups 1–2 include normal, not at all ill and borderline ill patients; subgroups 3–5 include mild, moderate, and markedly ill patients; and subgroups 6–7 include severely ill, and the most extremely ill patients. p < 0.0003 (Kruskall-Wallis test comparing all three subgroups) and p < 0.000.1 (Mann-Whitney-Wilcoxon test comparing subgroups 3–5 with 6–7) for comparisons of RLSQoL overall life impact scores among collapsed CGI-S subgroups in both TREAT RLS 1 and 2; p = not significant (Mann-Whitney-Wilcoxon test comparing subgroups 1–2 with 3–5). RLSQoL = RLS Quality of Life questionnaire. CGI-S = Clinical Global Impression 'Severity of Illness'.
Figure 3 Responsiveness. Effect sizes as a measure of the change in RLSQoL overall life impact change scores between baseline and week 12, by CGI-I scores at week 12. p < 0.0001 for comparisons of the change in RLSQoL overall life impact scores among CGI-I groups in both TREAT RLS 1 and 2 (Kruskall-Wallis test). The sample size of n = 1 in the 'worsened' group in TREAT RLS 2 meant that effect sizes could not be calculated. RLSQoL = RLS Quality of Life questionnaire. CGI-I = Clinical Global Impression 'Improvement'.
Table 1 Patient characteristics at baseline for TREAT RLS 1 (n = 284) and TREAT RLS 2 (n = 263)
TREAT RLS 1 TREAT RLS 2
Age, years
Mean (SD) 55.1 (11.2) 55.4 (11.1)
Range 28.0–78.0 29.0–79.0
Sex, % (n)
Men 37.0 (105) 40.3 (106)
Women 63.0 (179) 59.7 (157)
Work status, % (n)
Full-time employment 36.3 (103) 43.4 (114)
Part-time employment 15.1 (43) 12.6 (33)
Volunteer/unpaid work 2.5 (7) 1.9 (5)
Unemployed due to RLS 1.1 (3) 0.8 (2)
Unemployed due to other (non-RLS) symptoms 3.2 (9) 0.4 (1)
Unemployed for other reasons 8.5 (24) 4.2 (11)
Retired 22.2 (63) 29.3 (77)
Homemaker 11.3 (32) 7.6 (20)
CGI severity of illness, % (n)
Missing data 0 (0) 0.4 (1)
Normal, not at all ill 2.1 (6) 0.4 (1)
Borderline ill 1.1 (3) 1.5 (4)
Mildly ill 9.2 (26) 7.2 (19)
Moderately ill 34.9 (99) 36.5 (96)
Markedly ill 32.0 (91) 31.2 (82)
Severely ill 19.0 (54) 20.2 (53)
Among the most extremely ill patients 1.8 (5) 2.7 (7)
Item 6 of the IRLS: 'How severe was your RLS as a whole?', % (n)
None 0 (0) 0 (0)
Mild 2.1 (6) 2.3 (6)
Moderate 34.9 (99) 36.9 (97)
Severe 43.7 (124) 44.9 (118)
Very severe 19.4 (55) 16.0 (42)
Item 7 of the IRLS: 'How often did you get RLS symptoms?', % (n)
Rarely (< 1 day a week) 0 (0) 0 (0)
Occasionally (1 day a week) 1.4 (4) 0.4 (1)
Sometimes (2–3 days a week) 10.6(30) 9.1 (24)
Often (4–5 days a week) 25.4 (72) 32.7 (86)
Very often (6–7 days a week) 62.7 (178) 57.8 (152)
Age of onset of symptoms, years Mean (SD) 38.2 (16.8) 34.6 (17.3)
CGI = Clinical Global Impression. IRLS = International Restless Legs Scale. SD = standard deviation.
Table 2 Missing data for the RLSQoL items in TREAT RLS 1 and 2
RLSQoL item TREAT RLS 1 % (n) TREAT RLS 2 % (n)
1 In the past 4 weeks how distressing to you were your restless legs? 0 (0) 0 (0)
2 How often in the past 4 weeks did RLS disrupt your routine evening activities? 0 (0) 0.38 (1)
3 How often in the past 4 weeks did RLS keep you from attending your evening social activities? 1.06 (3) 0.76 (2)
4 In the past 4 weeks how much trouble did you have getting up in the morning due to restless legs? 0 (0) 0 (0)
5 In the past 4 weeks how often were you late for work or your first appointments of the day due to RLS? 3.52 (10) 1.52 (4)
7 How often in the past 4 weeks did you have trouble concentrating in the afternoon? 0 (0) 0 (0)
8 How often in the past 4 weeks did you have trouble concentrating in the evening? 0 (0) 0 (0)
9 In the past 4 weeks how much was your ability to make good decisions affected by sleep problems? 0.35 (1) 0 (0)
10 How often in the past 4 weeks would you have avoided travelling when the trip would have lasted more than 2 hours? 0 (0) 0.38 (1)
13 In the past 4 weeks how much did your restless legs disturb your ability to carry out daily activities, for example carrying out a satisfactory family, home, social, school or work life? 0.35 (1) 0 (0)
Table 3 Construct validity and reliability of the RLSQoL overall life impact score
TREAT RLS 1 TREAT RLS 2
Item-convergent validity, % items with item-scale correlation ≥ 0.40 (coefficient range) 85% (0.35–0.66) 100% (0.46–0.70)
Floor/ceiling effects (% of respondents with minimum/maximum scale scores) 0.35%/0.00% 0.00%/0.38%
Internal consistency reliability, Cronbach's alpha coefficient (satisfactory if ≥ 0.70) 0.82 0.87
RLSQoL = Restless Legs Syndrome Quality of Life questionnaire.
Table 4 Concurrent validity: correlations between the RLSQoL overall life impact score and the scores for the IRLS total score and the MOS Sleep Problems Index II (TREAT RLS 1 and 2)a
RLSQoL overall life impact score IRLS total score MOS Sleep Scale Sleep Problems Index II
TREAT RLS 1 Spearman correlation coefficient -0.68 -0.59
p value < 0.0001 < 0.0001
Number of patients 284 284
TREAT RLS 2 Spearman correlation coefficient -0.67 -0.60
p value < 0.0001 < 0.0001
Number of patients 263 263
IRLS = International Restless Legs Scale. RLSQoL = Restless Legs Syndrome Quality of Life questionnaire. MOS = Medical Outcomes Study.
a Satisfactory if correlation ≥ 0.40.
Table 5 Changes in RLSQoL overall life impact scores from baseline to week 12: effect sizes and a comparison between CGI-I levels 1–7 (TREAT RLS 1)
n Baseline mean Baseline SD (total) Week 12 mean Mean change Effect size Kruskal-Wallis test (p value)
Very much improved 68 64.33 18.25 91.95 27.62 1.51 0.0001
Much improved 59 62.41 18.25 83.47 21.06 1.15
Minimally improved 36 57.84 18.25 71.38 13.54 0.74
No change 51 58.65 18.25 65.41 6.76 0.37
Minimally worse 5 76.13 18.25 82.61 6.48 0.36
Much worse 9 58.33 18.25 57.53 -0.8 -0.04
p < 0.0001 for comparisons of the change in RLSQoL overall life impact scores among CGI-I groups in TREAT RLS 1 (Kruskall-Wallis test). There were no patients included in the CGI-I 'very much worse' category. Effect sizeswere calculated by dividing the change in mean score (from baseline to week 12) by the standard deviation of the mean score at baseline. RLSQoL = RLS Quality of Life questionnaire. CGI-I = Clinical Global Impression 'Improvement'.
Table 6 Changes in RLSQoL overall life impact scores from baseline to week 12: effect sizes and a comparison between CGI-I levels 1–7 (TREAT RLS 2)
n Baseline mean Baseline SD (total) Week 12 mean Mean change Effect size Kruskal-Wallis test (p value)
Very much improved 74 63.78 18.83 92.15 28.47 1.51 0.0001
Much improved 43 59.19 18.83 77.94 18.75 1.00
Minimally improved 45 66.72 18.83 76.94 10.22 0.54
No change 37 65.48 18.83 71.39 5.91 0.31
Minimally worse 6 60.83 18.83 70.83 10.00 0.53
Very much worse 1 80.00 18.83 85.00 5.00
p < 0.0001 for comparisons of the change in RLSQoL overall life impact scores among CGI-I groups in TREAT RLS 2 (Kruskall-Wallis test). There were no patients included in the CGI-I 'much worse' category. The 'very much worse' group consisted of a single patient and therefore further statistical tests were not performed for this group. Effect sizes were calculated by dividing the change in mean score (from baseline to week 12) by the standard deviation of the mean score at baseline.
RLSQoL = RLS Quality of Life questionnaire. CGI-I = Clinical Global Impression 'Improvement'.
==== Refs
Allen RP Picchietti D Hening WA Trenkwalder C Walters AS Montplaisir J Restless legs syndrome: diagnostic criteria, special considerations, and epidemiology. A report from the restless legs syndrome diagnosis and epidemiology workshop at the national institutes of health Sleep Med 2003 4 101 119 14592341 10.1016/S1389-9457(03)00010-8
Earley CP Restless legs syndrome N Engl J Med 2003 348 2103 2109 12761367 10.1056/NEJMcp021288
Rothdach AJ Trenkwalder C Haberstock J Keil U Berger K Prevalence and risk factors of RLS in an elderly population: the MEMO study. Memory and Morbidity in Augsburg Elderly Neurology 2000 54 1064 1068 10720275
Allen RP Earley CJ Validation of the Johns Hopkins Restless Legs Severity Scale (JHRLSS) Sleep Med 2001 2 239 242 11311687 10.1016/S1389-9457(00)00080-0
Hening W Walters AS Allen RP Montplaisir J Myers A Ferini-Strambi L Impact, diagnosis and treatment of restless legs syndrome (RLS) in a primary care population: the REST (RLS Epidemiology, Symptoms, and Treatment) Primary Care study Sleep Med 2004 5 237 246 15165529 10.1016/j.sleep.2004.03.006
Allen RP Earley CJ Restless legs syndrome: a review of clinical and pathophysiologic features J Clin Neurophysiol 2001 18 128 147 11435804 10.1097/00004691-200103000-00004
Allen RP La Buda MC Becker P Earley CJ Family history study of the restless legs syndrome Sleep Med 2002 3 S3 7 14592159 10.1016/S1389-9457(02)00140-5
Desautels A Turecki G Montplaisir J Sequeira A Verner A Rouleau GA Identification of a major susceptibility locus for restless legs syndrome on chromosome 12q Am J Hum Genet 2001 69 1266 1270 11704926 10.1086/324649
Chen JT Garcia PA Alldredge BK Zonisamide-induced restless legs syndrome Neurology 2003 60 147 12525743
Bonati MT Ferini-Strambi L Aridon P Oldani A Zucconi M Casari G Autosomal dominant restless legs syndrome maps on chromosome 14q Brain 2003 126 1485 1492 12764067 10.1093/brain/awg137
Trenkwalder C Garcia-Borreguero D Montagna P Lainey E de Weerd AW Tidswell P Saletu-Zyhlarz G Telstad W Ferini-Strambi L Ropinirole in the treatment of restless legs syndrome: results from the TREAT RLS 1 study, a 12-week, randomised, placebo controlled study in 10 European countries J Neurol Neurosurg Psychiatry 2004 75 92 97 14707315
Abetz L Vallow S Kirsch J Allen R Washburn T Earley C Validation of the Restless Legs Syndrome Quality of Life Questionnaire Value Health 2005 8 157 167 15804324 10.1111/j.1524-4733.2005.03010.x
Walters AS Ondo WG Dreykluft T Grunstein R Lee D Sethi K on behalf of the TREAT RLS 2 Study Group Ropinirole is effective in the treatment of restless legs syndrome: TREAT RLS 2: a 12-week, double-blind, randomized, parallel-group, placebo-controlled study Mov Disord 2004 19 1414 1423 15390050 10.1002/mds.20257
Allen RP Kushida CA Atkinson MJ RLS QoL Consortium. Factor analysis of the International Restless Legs Syndrome Study Group's scale for restless legs severity Sleep Med 2003 4 133 135 14592343 10.1016/S1389-9457(02)00193-4
Walters AS LeBrocq C Dhar A Hening W Rosen R Allen RP Trenkwalder C on behalf of the International Restless Legs Syndrome Study Group Validation of the International Restless Legs Syndrome Study Group Rating Scale for restless legs syndrome Sleep Med 2003 4 121 132 14592342 10.1016/S1389-9457(02)00258-7
Hays RD Stewart AL Stewart AL, Ware JE Jr Sleep measures Measuring Functioning and Well-Being The Medical Outcomes Study Approach 1992 Durham and London: Duke University Press 235 259
Hays RD Martin SA Sesti AM Spritzer KL Psychometric properties of the medical outcomes study sleep measure Sleep Med 2005 6 41 44 15680294 10.1016/j.sleep.2004.07.006
Guy W Ed Clinical Global Impression (CGI) ECDEU Assessment Manual for Psychopharmacology 1976 . Rockville, MD: US Department of Health and Human Services, Public Health Service, Alcohol Drug Abuse and Mental Health Administration, NIMH Psychopharmacology Research Branch 218 222
Abetz L Arbuckle R Allen R Garcia-Borreguero D Hening W Walters A Mavraki E Kirsch J Validating the International Restless Legs Scale (IRLS) in a trial patient population 2004 Poster presented at the 16th Annual Meeting of the British Sleep Society, Cambridge, UK 19 21
Campbell DT Fiske DW Convergent and discriminant validation by the multitrait-multimethod matrix Psychol Bull 1959 56 81 105 13634291
Bennett RM Schein J Kosinski MR Hewitt DJ Jordan DM Rosenthal NR Impact of fibromyalgia pain on health-related quality of life before and after treatment with tramadol/acetaminophen Arthritis Rheum 2005 53 519 527 16082646 10.1002/art.21319
Kazis LE Anderson JJ Meenan RF Effect Sizes for interpreting changes in health Status Med Care 1989 27 S178 S189 2646488
Cohen J Statistical Power Analysis for the Behavioral Sciences 1988 2 Hillsdale, NJ: Lawrence Earlbaum Associates
Guyatt GH Osaba D Wu AW Wyrwich KW Norman GR the Clinical Significance Consensus Meeting Group Methods to explain the clinical significance of health status measures Mayo Clin Proc 2002 77 371 383 11936935
Dunteman GH Principal Components Analysis Sage University Paper series on Quantitative Applications in the Social Sciences, Series No 69 1989 Newbury Park, CA: Sage
Scheitel SM Boland BJ Wollan PC Silverstein MD Patient-physician agreement about medical diagnosis and cardiovascular risk factors in the ambulatory general medication examination Mayo Clin Proc 1996 71 1131 1137 8945482
Abetz L Shumaker S Marciniak A Thorsen H Niero M Kohlmann T Agreement levels between clinicians' and patients' reports of health: results from an international study of postmenopausal women 2003 Poster presented at The North American Menopause Society Meeting, Miami Beach, Florida, USA 19 21
Bouchet C Guillemin F Briancon S Non-specific effects in longitudinal studies: impact on quality of life measures J Clin Epidemiol 1996 49 15 20 8598506 10.1016/0895-4356(95)00540-4
|
16329756
|
PMC1322229
|
CC BY
|
2021-01-04 16:05:42
|
no
|
Health Qual Life Outcomes. 2005 Dec 5; 3:79
|
utf-8
|
Health Qual Life Outcomes
| 2,005 |
10.1186/1477-7525-3-79
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1571628008510.1186/1471-2164-6-157Research ArticleGenetic mapping of putative Chrna7 and Luzp2 neuronal transcriptional enhancers due to impact of a transgene-insertion and 6.8 Mb deletion in a mouse model of Prader-Willi and Angelman syndromes Stefan Mihaela [email protected] Kathryn C [email protected] Edyta [email protected] Jing-Hua [email protected] Tohru [email protected] Richard [email protected] John M [email protected] Robert D [email protected] Center for Neurobiology and Behavior, Department of Psychiatry, University of Pennsylvania, Philadelphia, PA 19104, USA2 Department of Genetics, University of Pennsylvania, Philadelphia, PA 19104, USA3 Division of Hematology, Department of Medicine, Albert Einstein College of Medicine, The Bronx, USA4 Department of Microbiology and Immunology, Feinberg School of Medicine, Northwestern University, Ward 6-231, 303 East Chicago Ave, Chicago, IL 60611, USA5 Birth Defects Laboratories, Children's Hospital of Pittsburgh, Room 2109 Rangos Research Center, 3460 Fifth Avenue, Pittsburgh, PA 15213, USA6 Health Science University of Hokkaido, Hokkaido, Japan7 Department of Pediatrics, Children's Hospital of Pittsburgh, 3460 Fifth Avenue, Pittsburgh, PA 152132005 9 11 2005 6 157 157 14 7 2005 9 11 2005 Copyright © 2005 Stefan et al; licensee BioMed Central Ltd.2005Stefan 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 and Angelman syndrome (PWS and AS) patients typically have an ~5 Mb deletion of human chromosome 15q11-q13, of opposite parental origin. A mouse model of PWS and AS has a transgenic insertion-deletion (TgPWS/TgAS) of chromosome 7B/C subsequent to paternal or maternal inheritance, respectively. In this study, we define the deletion endpoints and examine the impact on expression of flanking genes.
Results
Using molecular and cytological methods we demonstrate that 13 imprinted and 11 non-imprinted genes are included in the TgPWS/TgAS deletion. Normal expression levels were found in TgPWS brain for genes extending 9.1- or 5.6-Mb centromeric or telomeric of the deletion, respectively. Our molecular cytological studies map the proximal deletion breakpoint between the Luzp2 and Siglec-H loci, and we show that overall mRNA levels of Luzp2 in TgPWS and TgAS brain are significantly reduced by 17%. Intriguingly, 5' Chrna7 shows 1.7-fold decreased levels in TgPWS and TgAS brain whereas there is a ≥15-fold increase in expression in neonatal liver and spleen of these mouse models. By isolating a Chrna7-Tg fusion transcript from TgAS mice, we mapped the telomeric deletion breakpoint in Chrna7 intron 4.
Conclusion
Based on the extent of the deletion, TgPWS/TgAS mice are models for PWS/AS class I deletions. Other than for the first gene promoters immediately outside the deletion, since genes extending 5.6–9.1 Mb away from each end of the deletion show normal expression levels in TgPWS brain, this indicates that the transgene array does not induce silencing and there are no additional linked rearrangements. Using gene expression, non-coding conserved sequence (NCCS) and synteny data, we have genetically mapped a putative Luzp2 neuronal enhancer responsible for ~33% of allelic transcriptional activity. The Chrna7 results are explained by hypothesizing loss of an essential neuronal transcriptional enhancer required for ~80% of allelic Chrna7 promoter activity, while the Chrna7 promoter is upregulated in B lymphocytes by the transgene immunoglobulin enhancer. The mapping of a putative Chrna7 neuronal enhancer inside the deletion has significant implications for understanding the transcriptional regulation of this schizophrenia-susceptibility candidate gene.
==== Body
Background
Prader-Willi and Angelman syndrome (PWS and AS) are complex neurobehavioral disorders associated with loss of function of a cluster of differentially expressed imprinted genes in chromosome 15q11-q13 [1]. PWS is characterized by a neonatal stage of failure to thrive, hypotonia and respiratory distress followed by hyperphagia in early childhood with development of severe obesity, as well as short stature, hypogonadism, small hands and feet, mild to moderate mental retardation, and obsessive-compulsive behavior [2,3]. In contrast, AS patients have a more pronounced neurological disease including developmental delay, severe mental retardation with lack of speech, hyperactivity, seizures, aggressive behavior and excessive inappropriate laughter [2]. Most PWS and AS cases (~70%) are due to ~5 Mb de novo deletions spanning a 2 Mb imprinted domain and several adjacent non-imprinted genes [1]. There are two classes of deletions in PWS/AS patients, one from breakpoint 1 (BP1) to BP3 and the other from BP2 to BP3 [4]. Additionally, paternal or maternal uniparental disomy (pat or matUPD) explain 25% of PWS and 5% of AS cases, respectively, while 2–5% of PWS and AS cases result from imprinting defects (ID). In each mechanism, PWS arises from loss of ten paternally expressed loci, while AS arises from loss of function of the maternally expressed UBE3A gene [1].
Mouse models of PWS with either matUPD [5], an ID [6] or a paternally-inherited chromosome deletion [7] share a similar phenotype with failure to thrive, hypotonia and early postnatal lethality, modeling the first stage of the human syndrome [9,10]. Similarly, mouse models of AS have a patUPD [10], maternally-inherited chromosome deletion [7], or a maternal mutation of Ube3a [11,12]. In the transgenic (Tg) deletion mouse model, an Epstein Barr Virus LMP2A transgene integrated with ~80 copies into mouse chromosome 7B/C and created an ~5 Mb deletion of the mouse region homologous to the human PWS/AS genes (see Fig. 1A,B) [7]. As in human, the phenotype of the deletion mouse model depends on the parental origin: paternal or maternal inheritance of the Tg-deletion, respectively, results in the TgPWS mouse model characterized by severe neonatal hypoglycemia and early lethality [9] or in TgAS mice with a mild neurobehavioral phenotype and late onset obesity [7].
Figure 1 Genetic and physical maps of mouse chromosome 7B/C. (A) The mouse PWS/AS-homologous region and flanking genes. Symbols are: circles, protein-coding genes; ovals, RNA-coding genes; black, paternally-expressed; grey, maternally-expressed; white, biparentally-expressed; line arrows, transcriptional orientation of genes; IC; imprinting center; plo1, p-locus-associated obesity; cp1, cleft palate 1; oca2, oculocutaneous albinism type II (p, pink-eyed dilution); jdf2, juvenile development and fertility 2; cen, centromere; tel, telomere; *, genes represented on the Gene chip MG_U74Av2 (see Table 1); #, genes analyzed by QRT-PCR; horizontal grey bar, extent of the TgPWS/TgAS ~5 Mb deletion. The first 3 genes and one more distant gene extending out from each of the centromeric and telomeric deletion breakpoints are also shown. (B) Human synteny for genes from the mouse chromosome 7B/C domain. The human chromosome locations for each orthologous mouse gene shown in (A) are given. Arrowheads represent rodent-specific gene duplications for which no ortholog is found in human. (C) BAC contig across the TgPWS/TgAS centromeric deletion breakpoint region. Symbols are: white boxes, exons; line arrows, transcriptional orientation of genes; white bars, BACs, and one "wgs" (whole genome shotgun) contig; grey bars, BACs used as FISH probes; black boxes, Non-Coding Conserved Sequences (NCCSs) in RP24-246A19; *, D7Mit70. The interval from Luzp2 through Siglec-H and Tubgcp5 is shown. (D) BAC contig across the TgPWS/TgAS telomeric deletion breakpoint region. Symbols are as for Fig 1C, except: black boxes, four intronless imprinted genes. The interval from the telomeric end of the 2 Mb imprinted gene domain through Chrna7 to Cezanne2 and Klf13 is shown.
Previous imprinted gene expression studies and fluorescence in situ hybridization (FISH) showed that the Tg insertion-deletion comprised all of the orthologs of PWS and AS imprinted genes and of several flanking non-imprinted genes [Fig. 1A; [7,13,14]]. In addition, brain microarray and quantitative gene expression analyses confirmed the loss of expression of several imprinted genes (Snurf-Snrpn, Ndn, Magel2, Mkrn3) and 50% reduced expression of two non-imprinted loci (Herc2, Cyfip1) in TgPWS mouse brain, and demonstrated that an unidentified non-imprinted locus within the deletion acted in trans to regulate a chromosome 18B3 gene expression domain [15]. To define the exact characteristics of the transgene insertion-deletion TgPWS/TgAS mouse model, we have now determined the extent of the deletion and the effect of the Tg-insertion and deletion on expression of flanking genes. We delineated the deletion breakpoint positions between Siglec-H and Luzp2 at the centromeric end and within intron 4 of Chrna7 at the telomeric end. Most importantly, we describe tissue-specific positional effects of the Tg insertion and/or deletion on Luzp2 and Chrna7 expression which are likely due to the presence or absence of specific enhancer elements.
Results
FISH refines the TgPWS/TgAS deletion extent and identifies a centromeric breakpoint
We used FISH with BAC probes flanking the mouse PWS/AS-homologous domain in chromosome 7B/C (Fig. 1C,D) to define the centromeric and telomeric extent of the TgPWS/TgAS deletion. To identify both chromosome 7 homologues in our FISH experiments using splenocytes from TgAS mice or fibroblasts from TgPWS mice, we co-hybridized BAC RP22-434N7 from the Tyr gene locus at position 74.6 Mb (chromosome 7E1), which is intact in TgPWS and TgAS mice (Fig. 2, and data not shown) [7].
Figure 2 Fluorescence in situ hybridization (FISH) maps the TgAS centromeric and telomeric deletion extent. In each case, BAC RP22-434N7 from the chromosome 7 Tyrosinase (Tyr) locus (74.6 Mb; 44.0 cM) was hybridized as a control and is shown as a red signal. All chromosome 7B/C BACs used as probes are shown as green signals while chromosomes are stained with DAPI (blue). (A) BAC RP24-354P8 spanning the region including Siglec-H and Tubgcp5 shows a single-chromosome 7 signal indicating the deletion of this locus in TgAS splenocytes. (B) RP23-256L9 spans the p locus (exons 10–24) and is also deleted in TgAS mice. (C) RP24-426A19 spans the region 3' of Luzp2 and detects a weak signal on one chromosome 7 homologue, suggesting partial deletion and detection of the centromeric breakpoint in the mutant mice. The two insets show the images for individual probes. (D) RP23-199N11 covers part of the region between Frat3 and Chrna7 and is deleted in TgAS mice. (E) RP23-506 for the Klf13 locus is intact in the deletion mice. (F) RP23-16A4 spans all 10 exons of Chrna7 and shows an apparently intact signal in TgAS mice.
At the centromeric end of the deletion, BAC RP24-354P8 spanning the Siglec-H-D7Mit70-Tubgcp5 region (43.3 Mb, Fig. 1C) is deleted from one chromosome 7 in TgAS mice (Fig. 2A). Additionally, BACs RP23-256L9 (p gene exons 10–24; 44 Mb) and RP23-195C6 (Herc2-Nipa1; 43.6 Mb) are within the TgPWS/TgAS deletion (Fig. 1A, Fig. 2B, and data not shown) as expected based on their map position. In contrast, BAC RP24-426A19 from ~50 kb 3' of Luzp2 (42.9 Mb; Fig. 1C) shows a weak positive signal on one chromosome 7 homologue in TgAS mice (Fig. 2C), suggesting a partial deletion and detection of the centromeric breakpoint in TgAS mice.
At the telomeric end of the deletion, BAC RP23-266F22 (49.4 Mb; Fig. 1D) spanning four mouse PWS-region imprinted genes, including Frat3, Mkrn3, Magel2 and Ndn, was previously found to be deleted in TgPWS mice [13]. Extending telomeric from Frat3 towards Chrna7 (Fig. 1D), BACs RP23-100P8, RP23-199N11 and RP23-140J4 were all deleted in TgPWS mice (Fig. 2D, and data not shown). In contrast, BACs RP23-5O6 from the Klf13 region (50.9 Mb; Fig. 1D, Fig. 2E) and RP24-215K5 from 5' Cezanne2 (Fig. 1D; data not shown) were intact in TgPWS/TgAS mice. Two other BACs mapping more telomeric, RP23-76F18 (Rlbp1 locus, 66.4 Mb) and RP23-441D12 (Il-16 locus, 70.8 Mb), were also intact in TgPWS mice (data not shown). Intriguingly, BAC RP23-16A4 spanning Chrna7 (50.1 Mb; Fig. 1D) appears based on hybridization signal intensity to be fully intact in TgPWS mice (Fig. 2F). Nevertheless, molecular data described below indicate that about half of 16A4 is deleted in TgPWS/TgAS mice; it is likely that these ostensibly contradictory data are explained by the relative concentration of unique and repetitive DNA sequences at the telomeric and centromeric ends of BAC 16A4 (5' and 3' ends of Chrna7, respectively) as well as the difficulty in quantifying FISH signals.
Microarray and QRT-PCR analyses further define the TgPWS/TgAS deletion extent
Previous analysis of brain global gene expression in 5 TgPWS and 5 wildtype (WT) mice at P1 using a MG-U74Av2 gene chip array (Affymetrix) which assayed 12,000 genes and ESTs, demonstrated that all 4 paternally-expressed mouse PWS-region imprinted genes present on the array (Snurf-Snrpn, Ndn, Magel2 and Mkrn3) have dramatically reduced expression [15]. Additionally, mRNA levels for non-imprinted genes Herc2 and Cyfip1 were 50% in TgPWS compared with WT [15]. To assess the extent of the TgPWS/TgAS deletion and the potential effect of the transgene-insertion deletion on expression of flanking genes, we used data mining from several genome databases and analysis of several megabases of genome sequence to identify all genes in these regions, determined which of these were on the MG-U74Av2 gene chip, and then reanalyzed the brain microarray data from reference [15]. Although a 3–4 Mb "gene desert" lies just proximal of Tubgcp5, we identified 9 genes on the chip in the 9.1 Mb region centromeric of Tubgcp5, a locus which FISH data above had shown was deleted, of which four genes were not detectable in either TgPWS or WT brain at P1 (Table 1). Five genes (Gas2, Hrmt1l3, E2F8, Ptpn5, and Tsg101) were detectable and showed no change in TgPWS compared to WT mouse brain (Table 1), indicating that the TgPWS/TgAS deletion does not affect any of these genes (Fig. 1A). Telomeric of Frat3, an imprinted gene deleted in TgPWS mice [13], nine (Klf13, BB128963, Mcee, Apba2, Tjp1, Snrp2a, H47, Mef2a, and Igf1r) of the fourteen genes extending out a further 5.6 Mb and on the microarray were detectable and all had comparable levels in TgPWS and WT mice (Table 1). Therefore, the Tg insertion-induced deletion does not affect gene expression over large regions outside the boundaries of the rearrangement.
Table 1 Brain microarray data for mouse chromosome 7B2-C genesa in TgPWS vs. WT mice.
GenBank Affy ID Gene mRNA levelb Chr. 7 location (Mb and/or cM)
TgPWS WT
I. Genes centromeric of the TgPWS/TgAS deletion
NM_008087 94337_at Gas2 4.3 (3/5) 5.0 (2/5) 39.3 Mb; 26.8 cM
94338_g_at Gas2 43.1 (5/5) 42.92 (5/5) as above
AK049836 97539_at Hrmt1l3 50.2 (5/5) 44.5 (4/5) 37.25 Mb
NM_016865 103671_at Htatip2 23.8 (0/5) 21.0 (0/5) 37.2 Mb
AY957576 103202_f_at E2F8 35.3 (4/5) 30.4 (4/5) 36.3 Mb
103204_r_at E2F8 27.4 (1/5) 21.6 (2/5) as above
NM_013808 103084_at Csrp3 4.6 (0/5) 3.2 (0/5) 36.3 Mb
NM_013643 100406_at Ptpn5 399.8 (5/5) 406.1 (5/5) 34.5 Mb
NM_016855 95304_at Attp 8.5 (0/5) 8.7 (0/5) 34.3 Mb
NM_021884 94809_at Tsg101 56.5 (5/5) 56.6 (5/5) 34.3 Mb
NM_013580 93103_at Ldh3 1.7 (0/5) 1.3 (0/5) 34.3 Mb; 23.5 cM
II. Genes telomeric of the TgPWS/TgAS deletion
NM_007390 101131_at Chrna7 86.8 (0/5) 87.1 (0/5) 50.1 Mb; 30.0 cM
NM_021366 160617_at Klf13 145.2 (5/5) 155.6 (5/5) 50.9 Mb
NM_018752 102251_at Trpm1 77.7 (1/5) 56.7 (0/5) 51.2 Mb; 27.0 cM
BC055074 93235_at BB128963 51.3 (5/5) 46.0 (5/5) 51.3 Mb
NM_026483 104022_at Mphosph10 35.9 (0/5) 34.0 (0/5) 51.4 Mb
NM_028626 102022_at Mcee 79.6 (5/5) 81.1 (5/5) 51.4 Mb
NM_007461 92727_at Apba2 433.9 (5/5) 455.8 (5/5) 51.5 Mb; 25.5 cM
NM_009386 99935_at Tjp1 185.3 (5/5) 142.5 (5/5) 52.3 Mb; 28.5 cM
D50060 101196_at Pace4 78.3 (0/5) 71.2 (0/5) 52.9 Mb; 28.5 cM
NM_021336 101506_at Snrpa1 131.2 (5/5) 126 (5/5) 53.1 Mb
NM_024439 94245_at H47 141.8 (5/5) 144.2 (5/5) 53.1 Mb; 28.5 cM
NM_053080 98372_at Aldh1a3 13.3 (0/5) 9.8 (0/5) 53.4 Mb
NM_194070 93852_at Mef2a 399.9 (5/5) 417.4 (5/5) 54.3 Mb; 33.0 cM
AF056187 102224_at Igf1r 381.6 (5/5) 339.6 (5/5) 55.0 Mb; 33.0 cM
a The primary microarray data from this experiment were reported in reference 15. As summarized in Fig. 1A, reference 15 included the data for the imprinted genes within the TgPWS/TgAS deletion that are on the MG-U74Av2 chip (Snrpn, Ndn, Magel2, Mkrn3 and Ube3a) and for non-imprinted genes within the deletion (Cyfip1, Herc2, and Gabrb3, although the latter transcript was not detectable; the p gene is also represented on the chip although not detectable in TgPWS or WT mice, and not previously reported). Additionally, a cluster of chromosome 18B3 genes were found to be abnormally expressed in TgPWS mice [15]. b mRNA levels are presented as the medians of Signal (relative abundance of a transcript), for TgPWS (n = 5) and WT (n = 5), with the number of calls as "detected" out of 5 comparison experiments in parentheses.
We next assessed quantitative (Q) expression of genes that map in the proximity of the known deleted loci at the centromeric [14,15] and telomeric [14] ends of the deletion (Fig. 1A). At the centromeric end of the deletion, the Siglec-H gene maps just proximal of Tubgcp5 and is predicted to have five exons (Fig. 1C), with the rare feature of having the stop codon in the penultimate exon 4 [19]. Since there is no information regarding the pattern of expression for Siglec-H [20,21], we first assessed this using RT-PCR (Fig. 3A,B). In agreement with the general signaling function of Siglec genes in the haemopoietic, immune and nervous systems [22], Siglec-H was predominantly expressed in brain and spleen (Fig. 3A). Less robust expression was seen in lung and skeletal muscle, while there was little or no amplification in heart, liver, kidney and testis (Fig. 3A). A postnatal role for Siglec-H expression was suggested by the lack of amplification at early embryonic stages with increased expression after birth (Fig. 3B), as also found for ESTs in Unigene (data not shown).
Figure 3 Gene expression analyses for loci at or near the TgPWS/TgAS deletion breakpoints. (A) Tissue expression of Siglec-H. RT-PCR using a cDNA mouse panel shows that Siglec-H (143-bp band) is strongly expressed in adult brain and spleen, with moderate expression in lung and skeletal muscle tissues, but is low or absent in heart, liver, kidney and testis. RNApolII (167-bp) gene expression was used as a control for both (A) and (B). (B) Developmental Siglec-H expression. Very low or absent Siglec-H expression is detected by RT-PCR at E7-E15 with weak expression at late gestational stages (E17). After birth, SiglecH expression is robustly detected in brain at P1 and P4. (C-F) Relative quantification of brain mRNA levels for (C) Siglec-H, (D) Luzp2, (E) Cezanne2 and (F) Chrna7 was performed by QRT-PCR in two groups of 5 TgPWS (black bars) or 5 TgAS (gray bars) and 5 WT (white bars) mice at P1 and in 4 TgPWS or 4 TgAS and 4 WT mice at P4. All values are presented as the means ± SE: * P ≤ 0.05; ** P ≤ 0.001; *** P ≤ 0.0001, significant differences between WT and TgPWS or TgAS at the indicated time points (Independent samples t-test). (C) Siglec-H mRNA levels are decreased by ~2-fold in TgPWS and TgAS compared with WT at P1 and P4. SiglecH mRNA levels in TgPWS vs. WT were 0.52 ± 0.04 vs. 1.29 ± 0.15 (P = 0.002) and 0.56 ± 0.12 vs. 1.31 ± 0.19 (P = 0.01) at P1 and P4, respectively. Siglec-H mRNA levels were 0.45 ± 0.03 in TgAS compared with 1.18 ± 0.08 in WT at P1 (P = 0.0002) and 0.39 ± 0.04 compared with 0.92 ± 0.07 at P4 (P = 0.0008). (D) Luzp2 mRNA levels in TgPWS vs. WT were 0.98 ± 0.02 vs. 1.25 ± 0.07 (P = 0.009) and 0.91 ± 0.02 vs. 1.11 ± 0.05 (P = 0.01) at P1 and P4, respectively. In TgAS vs. WT mice, mRNA levels were 0.97 ± 0.07 vs. 1.06 ± 0.05 (P = 0.3) at P1 and 0.81 ± 0.05 vs. 0.98 ± 0.02 (P = 0.01) at P4. (E) Cezanne2 shows no difference in expression in TgPWS or TgAS and WT mice. At P1, mRNA levels were 1.01 ± 0.03 for TgPWS vs. 1.04 ± 0.1 for WT (P = 0.8) and 1.03 ± 0.09 for TgAS vs. 1.06 ± 0.03 for WT (P = 0.7). At P4, Cezanne2 mRNA levels were 0.92 ± 0.08 for TgPWS vs. 1.04 ± 0.03 for WT (P = 0.2) and 0.97 ± 0.23 for TgAS vs. 1.09 ± 0.06 for WT (P = 0.6). (F) Chrna7 mRNA levels are ~0.6-fold in TgPWS and TgAS compared with WT mouse brain. At P1 the mRNA levels were 0.64 ± 0.04 in TgPWS vs. 0.99 ± 0.08 in WT (P = 0.005) and 0.57 ± 0.02 in TgAS vs. 0.89 ± 0.1 in WT (P = 0.01), while at P4 mRNA levels for Chrna7 were 0.58 ± 0.01 for TgPWS vs. 1.05 ± 0.03 for WT (P < 0.0001) and 0.63 ± 0.04 for TgAS vs. 1.08 ± 0.12 for WT (P = 0.005).
By QRT-PCR, Siglec-H showed 2-fold decreased mRNA levels in the brain of both TgPWS and TgAS mouse models at P1 and P4 (P < 0.01; Fig. 3C), indicating that this gene appears to be within the TgPWS/TgAS deletion (Fig. 1A). It may also be noted that the 0.5-fold expression level in TgPWS and TgAS mice compared with WT indicates that Siglec-H is not an imprinted gene. The next gene centromeric of this, Luzp2, showed ~0.8-fold decreased mRNA levels in TgPWS brain at P1 and P4 and in TgAS brain at P4 (P < 0.05) (Fig. 3D). At P1 in TgAS, Luzp2 brain mRNA levels were decreased to 0.91-fold vs. WT (P = 0.3). Combined, the presence of the TgPWS/TgAS deletion resulted in a Luzp2 mRNA level in neonatal brain that is significantly (0.83-fold, P = 0.0002) less than in WT. We conclude that the Luzp2 gene is intact at the centromeric end of the TgPWS/TgAS deletion but suggest a model in which a neuronal regulatory element is deleted that accounts for ~33% of Luzp2 expression from the deleted allele (see Discussion for a description of this and alternative models).
At the telomeric end of the deletion, expression of Cezanne2, which maps just centromeric of Klf13 (Fig. 1D), was unchanged by QRT-PCR in TgPWS and TgAS brain compared with WT mice (Fig. 3E), and thus this gene is intact in the TgPWS/TgAS deletion mouse model. The only remaining gene between Cezanne2 (intact in TgPWS/TgAS mice) and Frat3 (deleted [13]) is Chrna7 (Fig. 1D), suggesting that the deletion breakpoint might lie in the vicinity of this gene. Chrna7 is orientated with a telomeric 5' end (Fig. 1D). Using QRT-PCR for Chrna7 and primers spanning exon 2 to exon 3, we found ~1.7-fold decreased Chrna7 mRNA levels in both TgPWS and TgAS brain at P1 and P4 (P < 0.05) (Fig. 3F). Since the 5' end of Chrna7 is not deleted in the TgPWS/TgAS mouse model (see below, which is also consistent with the normal FISH signal for BAC RP23-16A4 reported above), the finding of ~60% total mRNA levels suggests that only ~20% of the usual level of Chrna7 mRNA transcripts are generated from the deleted chromosome in TgPWS and TgAS mouse brain. The 80% reduction in expression in cis suggests that the deletion removes essential regulatory sequences (see Discussion).
Fine-mapping of the centromeric TgPWS/TgAS deletion breakpoint
The gene expression data described above suggest that Siglec-H is deleted in TgPWS and TgAS mice while Luzp2 is intact but shows slightly yet significantly reduced expression levels. In order to further delineate the centromeric breakpoint of the deletion and a potential mechanism for the ~33% reduction in Luzp2 expression from the deletion allele, we examined the genome sequence of BAC RP24-426A19 that lies between Siglec-H and Luzp2 (Fig. 1C) and that was identified above by FISH as partially deleted in TgPWS/TgAS mice. There are no exons for any gene encoded in this BAC. Nevertheless, using BLAST of RepeatMasked RP24-426A19 DNA sequence against the non-redundant GenBank database, extending from the centromeric end of BAC RP24-426A19 we identified five non-coding conserved sequences (NCCS1-5) at positions ~30.1, 39.9, 68.8, 107.3, and 134.5 kb, respectively (Fig. 1C). NCCS1-5 are 250-bp, 68-bp, 188-bp, 104-bp, and 157-bp in length and show 82%, 91%, 81%, 82% and 83% identity with human, respectively. As in mouse, the five NCCS elements in human are located adjacent to LUZP2 and map in chromosome 11p14.3 (Fig. 1B). All five NCCS elements are likely to be involved in the regulation of LUZP2, while other flanking genes are unlikely to be regulated by or to relate to NCCS1-5, as the next most centromeric gene (Gas2) maps 3.2 Mb from Luzp2 while the genes more telomeric are either non-syntenic in human (Fig. 1B), mapping to human chromosome 15q, or represent a rodent-specific acquisition (Siglec-H; see Discussion).
PCR primers were designed from four of the NCCS elements and Q-PCR of DNA from TgPWS and WT mice was used to examine the relative amount of each NCCS with respect to intron 1 of Gapdh, which is intact in all mice studied. As a deletion control, we similarly examined by Q-PCR the Snurf-Snrpn promoter, which as expected showed a 0.4 to 0.44-fold relative level in TgPWS vs. WT mice (Fig. 4A,4B). Whereas there was no difference between TgPWS and WT mice for NCCS1 and NCCS3 (Fig. 4A), indicating that NCCS1, NCCS2 and NCCS3 are intact in TgPWS mice, Q-PCR for NCCS4 and NCCS5 identified 1.8-fold and 2.1-fold lower levels for TgPWS vs. WT, respectively (Fig. 4B), indicating that these are deleted from one allele in TgPWS mice. Therefore, the centromeric breakpoint of the TgPWS/TgAS ~5 Mb deletion lies in the 38.5 kb region between NCCS3 and NCCS4.
Figure 4 Mapping of centromeric and telomeric TgPWS/TgAS deletion breakpoints and tissue specific up-regulation of Chrna7. (A) Q-PCR for the Snurf-Snrpn promoter, NCCS1 and NCCS3 relative to Gapdh intron 1. The amount of Snurf-Snrpn DNA was 1.01 ± 0.01 for the WT compared with 0.40 ± 0.05 for the TgPWS samples, P = 0.006. For NCCS1, DNA fold was 0.98 ± 0.02 in WT vs. 1.12 ± 0.30 in TgPWS, P = 0.68. For NCCS3, DNA fold was 0.90 ± 0.09 in WT compared with 1.09 ± 0.03 in TgPWS mice, P = 0.20. (B) Q-PCR for NCCS4 and NCCS5 relative to Gapdh intron 1. Snurf-Snrpn DNA-fold was 0.40 ± 0.07 in TgPWS vs. 0.90 ± 0.1 in WT mice, P = 0.02. The amount of DNA for NCCS4 was 0.50 ± 0.05 in TgPWS and 0.90 ± 0.1 in WT mice, P = 0.05. For NCCS5, DNA-fold was 0.45 ± 0.1 in TgPWS and 0.94 ± 0.06 in WT mice, P = 0.05. (C) Chrna7 expression (207-bp) is dramatically upregulated in liver of TgPWS and TgAS at P1 compared with WT mice by RT-PCR. RNApolII expression (167-bp) was used as a control in (C), (E) and (F). (D) Relative quantification by QRT-PCR of Chrna7 liver mRNA levels shows a ≥15-fold increased expression in TgPWS (black bars) vs. WT (open bars) mice. Individual values for Chrna7 mRNA expression of 4 TgPWS and 4 WT mice at P1 are shown. The average mRNA expression was 40.76 ± 14.30 for TgPWS and 2.70 ± 3.21 for WT, P = 0.002. (E) Up-regulation of Chrna7 expression in TgAS spleen compared with WT at P1 by RT-PCR. (F) Chrna7 is normally expressed in P1 liver from EμLMP2A transgenic mice compared with WT liver. (G) A Chrna7-LMP2A fusion transcript is identified by RT-PCR in TgPWS brain (lanes 1,2) but not in WT (lanes 3,4). (H) DNA sequence of the Chrna7-LMP2A fusion cDNA. Vertical lines mark the limit between exons 2, 3 and 4 of Chrna7 (bold font) and the 5' end of LMP2A sequence, while the 22-nt in italics is a transgene-specific sequence at the 5' end of the LMP2A exon.
Up-regulation of Chrna7 allows mapping the telomeric TgPWS/TgAS deletion breakpoint
In contrast to the results shown above for Chrna7 expression in brain, regular RT-PCR for Chrna7 in newborn liver showed a dramatic increased expression in TgPWS and TgAS compared with WT (Fig. 4C). Indeed, quantification of liver Chrna7 expression by QRT-PCR revealed ~15-fold increased mRNA levels in TgPWS mice (P = 0.002) (Fig. 4D). A similar up-regulation of Chrna7 expression was found by RT-PCR in spleen tissues of TgAS mice (Fig. 4E). However, EμLMP2A control mice with B-cell lineage expression of LMP2A showed normal expression of Chrna7 in newborn liver (Fig. 4F). Combined, these data on two different LMP2A transgene models indicates that LMP2A expression in B cells does not upregulate Chrna7 in trans, and suggests that the increased expression of Chrna7 observed in liver and spleen from TgPWS and TgAS mice results from a cis effect of the transgene insertion. Since B-cell transcription of the LMP2A transgene in these models is driven from an immunoglobulin heavy chain gene promoter and enhancer [7,16], we hypothesized that the latter enhancer was also driving transcription of the heterologous, endogenous Chrna7 promoter.
To test this hypothesis, we used an exon 2 forward primer for Chrna7 and a reverse primer (OL-105) from inside the Tg (Fig. 5A) to perform RT-PCR on cDNA from TgPWS vs. WT brain, which detected a 1.1 kb product specifically in TgPWS samples (Fig. 4G). The 1.1 kb band was cloned and sequenced, identifying it as a fusion transcript encoding Chrna7 exons 2–4 fused to the sense strand of the LMP2A transgene (Fig. 4H). The latter result indicates that at least the telomeric copy of the Tg array is orientated in a 5'-3' telomere to centromere fashion. Furthermore, we can conclude that the 5' end of the Chrna7 locus is intact and correctly splices through exon 4 but then splices into the LMP2A transgene. These data strongly suggest that the telomeric breakpoint of the TgPWS/TgAS ~5 Mb deletion is in the 38.1 kb Chrna7 intron 4.
Figure 5 Orientation of the LMP2A transgene array. (A) LMP2A transgene structure. Restriction enzyme sites and primer location and orientation are shown. (B) LMP2A transgene amplification in TgPWS brain (Br) and liver (Li) (lanes 2,4) using primers RN1316 and RN1317 compared with the WT controls (lanes 3,5). (C-D) The use of (C) RN1316 or (D) RN1317 alone fails to generate PCR products in TgPWS and WT brain and liver.
Orientation of the LMP2A transgene array
To fully understand how the transgene array might affect flanking gene expression, it is necessary to define the integrity of the array and the orientation of each transgene copy and to flanking genes. To characterize the transgene array, we amplified genomic DNA from TgPWS mice and WT controls using primers from within the transgene (Fig. 5A). Amplification with either single primer (RN1316 or RN1317) did not lead to any product (Fig. 5C,D), indicating that there are no head-to-head or tail-to-tail copies of adjacent transgene inserts. In contrast, PCR using primers from adjacent copies of the transgene results in amplification of the LMP2A array in TgPWS DNA (Fig. 5B). These data indicate that the ~80 copies of the transgene [7] are all in the same tail-to-head orientation. Combined with the data described above on the sense strand fusion of Chrna7 and LMP2A cDNA sequences, we conclude that the entire Tg array is orientated in a 5'-3' manner from the telomeric to centromeric end.
Discussion
A mouse model of PWS/AS class I deletions
The TgPWS and TgAS mouse models were created by insertion of an LMP2A transgene array of ~80 copies [7] that fortuitously generated a deletion equivalent to those that occur in PWS and AS in the human. Although the mouse deletion was previously estimated as being ~5 Mb in length, genome sequence analysis indicates that the deleted segment could be as large as 6.8 Mb (Siglec-H to Chrna7 distance). However, there are a few gaps in the sequence, particularly in several duplicated genomic regions throughout the imprinted domain, and so the exact size of this interval remains unknown. Nevertheless, our previous [7,13-15] and present studies allow the conclusion that 13 imprinted and 11 non-imprinted genes are included within the mouse PWS/AS-deletion region (see Fig. 1A). In addition to spanning all the PWS- and AS-homologous paternally and maternally expressed genes, respectively, the TgPWS/TgAS deletion in mice also includes the homologs of typically deleted non-imprinted genes (ie., Gabrb3, Gabra5, Gabrg3, p/Oca2, and Herc2) as well as those of the PWS/AS genes (Nipa1, Nipa2, Cyfip1 and Tubgcp5) that define the deletion as equivalent to human class I deletions [1,4]. Although the classic PWS or AS clinical phenotype is similar in both class I and class II deletions, respectively, a more severe neurobehavioral phenotype has been described for class I compared with class II deletion patients as well as for deletion vs. UPD patients in each syndrome [23,24]. The TgPWS and TgAS deletion mouse models may therefore prove useful to compare the neurobehavioral phenotype to the UPD or ID PWS [5,6] or Ube3a gene mutation AS [11,12] mouse models.
Although the TgPWS/TgAS mouse models show reduced expression of several other genes, including Siglec-H, Luzp2 and Chrna7, as compared to the human PWS and AS class I deletions, none of these additional genes are likely to contribute to the TgPWS and TgAS phenotypes. For example, since Siglec-H is a recent evolutionary addition in the rodent genome [20], its 50% reduction in expression levels due to hemizygosity at the centromeric end of the deletion is unlikely to have a phenotypic effect. Likewise, the ~17% reduction in overall expression of Luzp2 observed in brain of TgPWS/TgAS mice is unlikely to be functionally important, since complete loss of Luzp2 does not lead to any specific phenotype [25]. Although Chrna7 is another gene partially deleted in TgPWS/TgAS mice and the human ortholog maps in 15q13, the latter is outside the PWS/AS deletions and the TgPWS/TgAS phenotype is unlikely to be affected by Chrna7 hemizygosity, since Chrna7 heterozygous knockout mice show no abnormal phenotypes and even Chrna7-/- null mice have only mild skin and reproductive phenotypes [26-29].
Implications for evolutionary acquisition of new genes and breakpoint mechanisms
Our study also characterized expression of a recently identified gene, Siglec-H, which belongs to the Siglec gene family encoding immunoglobulin-like lectins that act as extracellular receptors for sialic acid residues of glycan chains [20,21]. The majority of Siglec genes map in a single CD33 (Siglec-3)-related cluster in human 19q13 and the syntenic mouse chromosome 7B2 region [20,21]. In contrast, we show here that Siglec-H, which is rodent-specific and has no primate ortholog [21], lies adjacent to the mouse PWS/AS-homologous region at a position 12.3 Mb from its ancestral location. It seems likely that Siglec-H has similar postnatal immunological functions as other Siglecs, given its robust expression in spleen. However, in contrast with the restricted expression of most other Siglec genes to the haematopoietic and immune systems, Siglec-H is also transcribed at high levels in postnatal brain. Although conjectural, it is tempting to speculate that neuronal expression of Siglec-H may have arisen due to an evolutionary positioning adjacent to single (Luzp2) [25] or clustered (ie., Tubgcp5, Cyfip1, Nipa2, Nipa1) [14] genes expressed at high levels in the nervous system, such that Siglec-H transcription may be under control of a neuronal enhancer for one or more of these genes.
Intriguingly, both the centromeric and telomeric TgPWS/TgAS deletion breakpoints map at or close to the positions of chromosomal evolutionary breakpoints, and at each of the latter two positions there are also rodent-specific gene duplication-insertions. At the centromeric end, the TgPWS/TgAS deletion breakpoint lies immediately adjacent to Siglec-H, which as discussed above arose in rodents by a genomic duplication from a precursor gene located ~12 Mb away. The evolutionary insertion of Siglec-H between Tubgcp5 and Luzp2 occurs right at the boundary of synteny with human chromosome 15q11.2-q13 and 11p14.3-p15.1, respectively (Fig. 1B). Moreover, the TUBGCP5 ortholog also lies within a few kilobases of a primate evolutionary breakpoint that translocated the TUBGCP5 – CYFIP1-NIPA2-NIPA1-HERC2-duplicon cluster to 15q11.2 from an ancestral 15q13 location [14]. Although the telomeric TgPWS/TgAS deletion breakpoint within Chrna7 is not right at an evolutionary breakpoint, this gene also lies adjacent to an evolutionary breakpoint separating genes whose human orthologs map to the PWS imprinted domain in 15q11.2 and a point 8.7 Mb away in 15q13.3 (Fig. 1B). At this position also, there are species-specific duplications in mouse and human, with the retrotransposed Frat3 limited to rodents [13] or the HERC2- and flanking duplicons in primates [5,14]. Other authors have also noted the apparent congruity of chromosome rearrangement and evolutionary breakpoints [14,30-32], suggesting that currently unknown chromosomal structural features in these regions may play recombinogenic roles.
Effect of the transgene or deletion on gene expression: mapping of putative neuronal enhancers
Transgene insertions may be accompanied by other chromosomal rearrangements such as deletions (this paper; [33]), duplications [34], inversions [35], translocations [36], or combinations of rearrangements [36,37]. These events can induce alterations of gene expression over large chromosomal regions [33], or the transgene itself can do so [38]. Extensive analysis of genes flanking the TgPWS/TgAS transgene insertion demonstrated no other rearrangements other than the PWS/AS-region deletion and that gene expression over large regions outside the immediate flanking genes was not affected. In contrast, expression from the first gene promoter outside either end of the transgene insertion-deletion was affected. For example, the 5' promoter and 3' end of Luzp2 are ~600 kb and 110–140 kb centromeric to the LMP2A transgene array-deletion breakpoint, respectively, and this gene shows a significantly reduced expression in TgPWS and TgAS mice. Although at present we cannot fully exclude three models that might lower the Luzp2 transcriptional level in cis in brain, which invoke either (i) a non-specific disruption of chromatin, (ii) a weak silencing effect from the transgene tandem array [45], or (iii) extended antisense transcripts from the transgene array given that the 5'-3' orientation of the LMP2A array is opposite to the Luzp2 direction of transcription, we feel that these are less likely than a fourth model that has additional experimental support. We propose that the 33% reduction in expression of Luzp2 from the deletion allele is most likely explained by deletion of a neuronal enhancer (Fig. 6A). Consistent with this model, we identified two NCCS elements (NCCS4 and NCCS5) conserved in all eutherian mammals sequenced to date and deleted in TgPWS and TgAS mice, that represent strong candidates to be the neuronal enhancer. As noted above, these elements map to the homologous chromosome location in human but are almost 4 Mb from the next most centromeric gene (Gas2) while other flanking genes in mouse are non-syntenic (Tubgcp5) or not present (Siglec-H) in human (Fig. 1B). Numerous NCCSs occur within mammalian genomes [39,40], and while the majority of these have no known function, many do act as enhancer elements [41-43]. Future in vitro and in vivo studies [44] will be able to examine this enhancer model for Luzp2.
Figure 6 Enhancer models for tissue-specific expression of Luzp2 and Chrna7. (A) Deletion of a putative neuronal enhancer 3' of Luzp2 lowers expression levels of the gene. Symbols are: white boxes, exons; grey vertical box, 3'-untranslated region (UTR); arrow, transcriptional orientation; zig-zag line, transgene insertion-deletion breakpoint; grey horizontal box, enhancer; dashed arrow, neuronal enhancer contributing ~33% to the Luzp2 allelic promoter activity; X, block to enhancer function. Only the first and last exons of Luzp2 are shown. (B) Activation in TgPWS and TgAS mice of the Chrna7 promoter in cis by the immunoglobulin (Ig) enhancer from the LMP2A transgene active in B lymphocytes. Symbols are as for Fig. 6A, except: grey vertical box, 5'-untranslated region (UTR); dashed arrow, Ig enhancer upregulation of Chrna7 transcription. (C) Activation of the Chrna7 promoter in WT mice by a putative neuronal enhancer. The location of the enhancer is drawn arbitrarily but must map 3' to the TgPWS/TgAS deletion breakpoint in intron 4. Symbols are as for Fig. 6B, except: dashed arrow, neuronal enhancer for upregulation (~80%) of normal Chrna7 transcription. (D) Deletion of a neuronal enhancer in TgPWS and TgAS mice virtually silences Chrna7 expression in cis. Symbols are as for Fig. 6A and 6C.
At the telomeric end of the deletion, expression of Chrna7 was dramatically upregulated in liver and spleen, but not brain, due to a tissue-specific positional effect of the transgene. We propose that the transgene immunogobulin Mu enhancer acts in B cells to promote transcription in cis from the heterologous Chrna7 promoter (Fig. 6B). Similarly, transgene regulatory elements 1 Mb away have been shown to activate Sox9 [38]. Moreover, our data combined with cloning of a 5'-Chrna7-LMP2A fusion transcript clearly demonstrates that the Chrna7 promoter is intact and fully capable of high levels of enhancer driven activity on the deletion allele in TgPWS/TgAS mice.
Perhaps the most intriguing finding of this study was our serendipitous discovery that a putative neuronal enhancer for the Chrna7 gene maps within the telomeric end of the TgPWS and TgAS deletion. In TgPWS and TgAS mice, the Chrna7 mRNA level in brain from the deletion chromosome 7 falls to 20% of normal levels despite the fact that the 5' promoter is not deleted. Non-enhancer models appear unlikely to explain this data: for example, models invoking (i) non-specific disruption of chromatin or (ii) a silencing effect from the transgene tandem array [45] are unlikely to account for the specific ~80% reduction in cis expression levels in TgPWS and TgAS mouse brain, while allowing a dramatic activation of the same promoter in immune tissues, and, finally (iii) Chrna7 is transcribed from upstream and in the same orientation as the transgene array, so that an antisense mechanism is not valid. Therefore, we propose that sequences that map 3' of the deletion breakpoint in Chrna7 intron 4 are critical for Chrna7 expression in neurons (Fig. 6C), and it is the deletion of this putative neuronal enhancer that accounts for the reduced Chrna7 expression in brain of TgPWS and TgAS mice (Fig. 6D). Similar analyses of Chrna7 gene expression in other tissues of TgPWS and TgAS mice, such as the neuromuscular junction, heart, and skin, will allow further assessment of the tissue specificity of the proposed enhancer. Interestingly, our data in brain and other tissues also imply that the Chrna7 promoter in vivo only has basal transcriptional activity whereas for high levels of transcription, Chrna7 requires activation by a tissue-specific enhancer(s). In support of this model, we recently identified as candidate enhancer elements two NCCSs within Chrna7 intron 4, whereas no other conserved non-coding sequences other than the minimal promoter occur in a 412 kb domain starting 18 kb 5' of Chrna7 and extending 3' of the gene towards the mouse PWS-region imprinted domain (R.D.N., K.C., and M.S., unpublished data). We are currently mapping the Chrna7 NCCS elements against the TgPWS/TgAS deletion endpoints, while in vitro and in vivo approaches will be needed for analysis of function as candidate neuronal transcriptional enhancers and/or for other cell types.
Conclusion
In this study, we used a variety of molecular cytological and genetic technologies to map the transgene insertion-chromosome deletion breakpoints of a mouse model of PWS and AS to ~38 kb regions between the Luzp2 and Siglec-H genes at the centromeric end and within Chrna7 intron 4 at the telomeric end, respectively. Gene expression analyses then allowed us to demonstrate that genes extending out a further 9.1- or 5.6-Mb centromeric or telomeric of the deletion, respectively, are not affected by either the deletion or insertion of the transgene tandem array. In contrast, transcription of genes at (Chrna7) or flanking (Luzp2) the transgene insertion-deletion breakpoints are affected by the disruption of normal chromosome architecture, with positional effects leading to up- or down-regulation dependent on the tissue-specificity and locations of enhancers within the transgene or putatively removed by the TgPWS/TgAS chromosome deletion. Using Luzp2 as an example, we demonstrate that analysis of phylogenetically conserved sequences and consideration of synteny allows the fine mapping of a putative neuronal enhancer element(s), while a similar model likely also applies for Chrna7. In human, CHRNA7 is of significant interest as a schizophrenia candidate gene with both genetic and functional support [46-48]. Nevertheless, the molecular basis of its candidacy remains uncertain since rare coding and in vitro characterized promoter variants in CHRNA7 [49,50] may be insufficient to account for deficient function in vivo [51], particularly in light of our observations that expression of Chrna7 in vivo requires a major participation of enhancer function. Identification of a neuronal enhancer and further studies in the human to examine for genetic variation and potential mutations might provide an explanation for a role of CHRNA7 as a schizophrenia-susceptibility gene.
Methods
Animals
We used three mouse models all of which have an LMP2A transgene insertion. In the TgPWS and TgAS models, the insertion of an LMP2A transgene array replaced the PWS/AS-homologous region in chromosome 7B/C [7,9,15] while the EμLMP2A mice have B-cell lineage expression of LMP2A [16]. WT littermates were used as controls in each experiment. The University of Pennsylvania Institutional Animal Care and Use Committee approved all animal experiments. All animals were bred and genotyped as described [7,16].
BAC contig assembly and fluorescence in situ hybridization (FISH) studies
Splenocytes from TgAS mice were cultured for 48 hr in the presence of concanavalin A for stimulation of cell proliferation. After harvesting, cells were treated with hypotonic buffer (0.075 M KCl) for 15 min, chromosomes were fixed with methanol-acetic acid (3:1) and spread onto slides [17]. For some probes, fibroblast cell lines from TgPWS mice were used. BAC DNA was labeled with biotin or digoxigenin and hybridized as previously described [18]. Image acquisition was performed using fluorescence microscopy, acquiring grayscale images using separate filters for each fluorophore followed by digital merging and pseudocoloring. BAC RP22-434N7, spanning the mouse chromosome 7 tyrosinase gene, was used as a control probe to identify both chromosome 7 homologues.
The chromosome 7B/C BAC contigs were generated by standard bioinformatics analyses using Ensembl [52] and BLAST [53]. The position and overlap of each BAC clone was determined by BLAST using gene cDNA/EST sequences, BAC end sequences, and STS sequences. Previously described BACs are RP23-266F22 [13], RP23-256L9, RP23-195C6, RP24-396N13 and RP24-354P8 [14]. It may be noted that we previously incorrectly assumed the position of Trpm1 (formerly Mlsn1) [7], whereas analysis of genomic sequence data now clearly place this locus telomeric to the TgPWS/TgAS deletion (see Fig. 1A).
Gene expression analyses
Microarray analysis
Brain global gene expression was assessed in five TgPWS and five WT male CD-1 sibs at P1 using the Affymetrix Murine Genome (MG) U74Av2 Array. The microarray experiments and data processing, with identification of genes significantly altered in TgPWS brain, were performed previously (see also Table 1) [15]. In the present study, we reanalyzed the microarray data for genes identified from bioinformatics analyses (see above) as mapping adjacent to the TgPWS/TgAS deletion.
QRT-PCR
Brain tissues from four or five each of TgPWS, TgAS and littermate WT mice at P1 and P4 were used for most QRT-PCR experiments. Liver Chrna7 expression was determined in four TgPWS and four WT mice at P1. QRT-PCR analysis was performed as described [15]. Briefly, total RNA was extracted using TRIzol (Invitrogen) and reverse-transcribed using Superscript First-Strand Synthesis System (Invitrogen). For all reactions, each sample was loaded in triplicate and SYBR green was used as the fluorescent dye. Gapdh was used as an internal control. Relative quantification of gene expression was carried out using a PRISM 7000 Sequence Detection System (Applied Biosystems Inc.) and data was processed using a comparative CT method [15]. A t-test for independent samples (Analyze-it) was used to generate two-tailed P statistics for each experiment. PCR primers were: Cezanne2 (exons 11–12): RN2262: 5'-TGATTCACAAGCTCCCCTAGCT-3' (F), RN2263: 5'-GGAGTGGACCTGGGTTCATCT-3' (R); Chrna7 (exons 2–3): RN2357: 5'-GCCGCTCACCGTGTACTTCT-3' (F), RN2358: 5'-GGTGGTTAAAACTTGGTTCTTCTCA-3' (R); Luzp2 (exons 7–8): RN2264: 5'-AAATCCAAGCCCAGCTGAAA-3' (F), RN2265: 5'-TGTTGGGCCTTAAATAACAAATCTT-3' (R); and Siglec-H (exons 1–2): RN2464: 5'-AGGATCTCTGTGCATGTGACAGA-3' (F), RN2465: 5'-AGGACGACCAAGCTCCAGTGT-3' (R).
Regular RT-PCR
Liver expression of Chrna7 was assessed in 2 TgPWS, 2 TgAS and 2 EμLMP2A mice and WT littermates at P1. In addition, Chrna7 expression in spleen was determined in 2 TgAS vs. 2 WT mice at P1. Extraction and reverse transcription of total RNA was performed as for QRT-PCR experiments. Chrna7 primers (exons 1–3) were: RN2368: 5'-GGAGGCATCTGGCTGGCTCTG-3' (F), and RN2358 (R; see above). Siglec-H expression was performed using a Mouse Multiple Tissue cDNA Panel I (BD Bioscience Clontech) as well as WT brain at P1 and P4. Siglec-H primers (exons 1–2) were RN2511: 5'-GTGACAACGGTTCTTACT-3' (F) and RN2465 (R). For each gene analysis, RNApolII expression was used as a control: RN2318: 5'-ACTCCTTCACTCACTGTCTTCCTGTT-3' (F), RN2319: 5'-TCCTGATCTTCTGCCACCACTGT-3'(R). PCR conditions for Chrna7, Siglec-H, and RNApolII were initial denaturation at 94°C, 10 min, followed by 32 cycles of denaturation at 94°C for 30 sec, annealing at 55°C, 30 sec, and extension at 72°C, 30 sec, with a final extension at 72°C, 10 min, using Gold Taq DNA polymerase (ABI).
NCCS analysis at the centromeric TgPWS/TgAS deletion breakpoint
We identified non-coding conserved sequences (NCCS) by taking BAC genomic sequence, masking repetitive sequences with RepeatMasker [54], then using BLAST analysis of the non-redundant (NR) and whole genome shotgun (wgs) databases. From BAC RP24-426A19, we identified five NCCS elements (see Results), and PCR primers for NCCS1, NCCS3, NCCS4 and NCCS5 were designed and used to quantify the DNA amount in WT and TgPWS mice by Q-PCR. Brain and liver DNA were extracted using a phenol-chloroform method. 500 ng of each DNA sample was used for Q-PCR and SYBR green was used as the fluorescent dye, with each Q-PCR reaction performed in triplicate. As an internal control for comparison of test DNA sequence, we used PCR for Gapdh intron 1: RN2518: 5'-GGCCGCCGCCATGT-3' (F) and RN2519: 5'-GGAAGGCCTAAGCAAGATTTCA-3' (R). In addition, primers from the Snurf-Snrpn promoter were used as a positive control for the deletion status in TgPWS DNA: RN2142: 5'-GCAAAAATGTGCGCATGTG-3' (F) and RN2143: 5'-CTCTCCTCTCTGCGCTAGTCTTG-3' (R). The CT method was used to process the data [15] and all the samples were normalized to the same WT sample. Primers for NCCS1 were: RN2515: 5'-AAATCATGAGCCAAGCCAAAA-3' (F) and RN2516: 5'-TGGCTTCCCTTATCACTTTCACA-3' (R); for NCCS3: RN2513: 5'-TTGGAACATGCAGAACAATGAAT-3' (F) and RN2514: 5'-AGGCTGCCAACCTGCAAA-3' (R); for NCCS4: RN2522: 5'-GCCACTAAATTGGATCCTTAGACATAT-3' (F) and RN2523: 5'-AAACCTGTTCCTACCCATGATAATCT-3' (R); and for NCCS5: RN2524: 5'-TCTCCCCCTAGGTCTTCTGTTTAA-3' (F) and RN2525: 5'-TGGCCAGTGATC ATGTACAGATC-3' (R).
Fine mapping of the telomeric TgPWS/TgAS deletion breakpoint
To fine map the telomeric TgPWS/TgAS breakpoint, total RNA from TgPWS brain tissue was reverse-transcribed using the Superscript First-Strand Synthesis System (Invitrogen) with a reverse primer from inside the transgene (Fig. 5A): OL-105: 5'-CGTGTGGCTTAC CTGCTGCCAATG-3'. cDNA was amplified by PCR using the RN2357 forward primer for Chrna7 (exon 2) and OL-105. The amplification conditions were initial denaturation at 94°C for 10 min, followed by 37 cycles of denaturation at 94°C for 4 min, annealing at 61°C for 1 min, and extension at 72°C for 6 min, with a final extension at 72°C for 10 min, using Platinum Pfx DNA polymerase (Invitrogen). PCR products were cloned into the pCR2.1-TA vector using a TOPO TA cloning kit (Invitrogen) and sequenced using standard techniques (University of Pennsylvania DNA sequencing Core facility).
Orientation of the LMP2A transgene
To determine the orientation of the LMP2A transgene array, three types of PCR amplifications were performed, first using forward and reverse primers located at the 3' and 5' ends of the transgene sequence, respectively (Fig. 5A): RN1316: 5'-GTTGTTTCCGCCATCGTACC-3' (R) and RN1317: 5'-TATGAGCTACTCTCTGACCC-3' (R). The other two reactions were performed using either forward or reverse primers alone. The amplification conditions were 95°C for 10 min followed by 32 cycles of denaturation at 95°C for 30 sec, annealing at 59°C for 30 sec, and extension at 72°C for 30 sec, with a final extension at 72°C for 10 min, using Platinum Taq Gold polymerase (ABI).
Authors' contributions
MS prepared tissue samples, performed PCR, RT-PCR, QRT-PCR and QPCR studies, microarray and genome sequence analyses and drafted the manuscript. KCC carried out some RT-PCR experiments and performed the identification of the telomeric deletion breakpoint and DNA sequence analyses. ES performed FISH studies. JHC contributed to BAC contig assembly. TO performed transgene array analyses and designed these experiments. RL established the transgenic mouse lines and provided the EμLMP2A mouse tissues. JMG designed, supervised and interpreted the FISH experiments. RDN designed the study, coordinated the project, performed genome sequence analysis, BAC contig assembly, participated in drafting and edited the manuscript.
Acknowledgements
We thank Brande Latney for technical assistance and Dr. Toni Portis for provision of tissues from EμLMP2A transgenic mice. This work was funded by grants to R.D.N. from the National Institutes of Health (HD31491 and ES10631) and from the Foundation for Prader-Willi Research (FPWR). R.L. is supported by Public Health Service grants CA62234, CA73507, and CA93444 from the National Cancer Institute and DE13127 from the National Institute of Dental and Craniofacial Research. M.S. was supported by a fellowship from the American Heart Association.
==== Refs
Nicholls RD Knepper JL Genome organization, function, and imprinting in Prader-Willi and Angelman syndromes Annu Rev Genomics Hum Genet 2001 2 153 175 11701647 10.1146/annurev.genom.2.1.153
Cassidy SB Schwartz S Prader-Willi and Angelman syndromes. Disorders of genomic imprinting Medicine (Baltimore) 1998 77 140 151 9556704 10.1097/00005792-199803000-00005
Pärtsch CJ Lammer C Gillessen-Kaesbach G Pankau R Adult patients with Prader-Willi syndrome: clinical characteristics, life circumstances and growth hormone secretion Growth Horm IGF Res 2000 10 S81 S85 10984259 10.1016/S1096-6374(00)80015-5
Amos-Landgraf JM Ji Y Gottlieb W Depinet T Wandstrat AE Cassidy SB Driscoll DJ Rogan PK Schwartz S Nicholls RD Chromosome breakage in the Prader-Willi and Angelman syndromes involves recombination between large, transcribed repeats at proximal and distal breakpoints Am J Hum Genet 1999 65 370 386 10417280 10.1086/302510
Cattanach BM Barr JA Evans EP Burtenshaw M Beechey CV Leff SE Brannan CI Copeland NG Jenkins NA Jones J A candidate mouse model for Prader-Willi syndrome which shows an absence of Snrpn expression Nat Genet 1992 2 270 274 1303278 10.1038/ng1292-270
Yang T Adamson TE Resnick JL Leff S Wevrick R Francke U Jenkins NA Copeland NG Brannan CI A mouse model for Prader-Willi syndrome imprinting-centre mutations Nat Genet 1998 19 25 31 9590284
Gabriel JM Merchant M Ohta T Ji Y Caldwell RG Ramsey MJ Tucker JD Longnecker R Nicholls RD A transgene insertion creating a heritable chromosome deletion mouse model of Prader-Willi and Angelman syndromes Proc Natl Acad Sci USA 1999 96 9258 9263 10430930 10.1073/pnas.96.16.9258
Nicholls RD Stefan M Ji H Qi Y Frayo RS Wharton RH Dhar MS Cummings DE Friedman MI Ahima RS Medeiros-Neto G, Halpern A, Bouchard C Mouse models for Prader-Willi and Angelman syndromes offer insights into novel obesity mechanisms Prog Obesity Res 2003 313 319
Stefan M Ji H Simmons RA Cummings DE Ahima RS Friedman MI Nicholls RD Hormonal and metabolic defects in a Prader-Willi syndrome mouse model with neonatal failure to thrive Endocrinology 2005 146 4377 4385 16002520 10.1210/en.2005-0371
Cattanach BM Barr JA Beechey CV Martin J Noebels J Jones J A candidate model for Angelman syndrome in the mouse Mamm Genome 1997 8 472 478 9195990 10.1007/s003359900479
Jiang YH Armstrong D Albrecht U Atkins CM Noebels JL Eichele G Sweatt JD Beaudet AL Mutation of the Angelman ubiquitin ligase in mice causes increased cytoplasmic p53 and deficits of contextual learning and long-term potentiation Neuron 1998 21 799 811 9808466 10.1016/S0896-6273(00)80596-6
Miura K Kishino T Li E Webber H Dikkes P Holmes GL Wagstaff J Neurobehavioral and electroencephalographic abnormalities in Ube3a maternal-deficient mice Neurobiol Dis 2002 9 149 159 11895368 10.1006/nbdi.2001.0463
Chai JH Locke DP Ohta T Greally JM Nicholls RD Retrotransposed genes such as Frat3 in the mouse Chromosome 7C Prader-Willi syndrome region acquire the imprinted status of their insertion site Mamm Genome 2001 12 813 821 11845283 10.1007/s00335-001-2083-1
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
Stefan M Portis T Longnecker R Nicholls RD A nonimprinted Prader-Willi Syndrome (PWS)-region gene regulates a different chromosomal domain in trans but the imprinted PWS loci do not alter genome-wide mRNA levels Genomics 2005 85 630 640 15820315 10.1016/j.ygeno.2005.02.004
Caldwell RG Wilson JB Anderson SJ Longnecker R Epstein-Barr virus LMP2A drives B cell development and survival in the absence of normal B cell receptor signals Immunity 1998 9 405 411 9768760 10.1016/S1074-7613(00)80623-8
Henegariu O Heerema NA Lowe Wright L Bray-Ward P Ward DC Vance GH Improvements in cytogenetic slide preparation: controlled chromosome spreading, chemical aging and gradual denaturing Cytometry 2001 43 101 109 11169574 10.1002/1097-0320(20010201)43:2<101::AID-CYTO1024>3.0.CO;2-8
Greally JM Gray TA Gabriel JM Song L Zemel S Nicholls RD Conserved characteristics of heterochromatin-forming DNA at the 15q11-q13 imprinting center Proc Nat Acad Sci USA 1999 96 14430 14435 10588722 10.1073/pnas.96.25.14430
Mendell JT Sharifi NA Meyers JL Martinez-Murillo F Dietz HC Nonsense surveillance regulates expression of diverse classes of mammalian transcripts and mutes genomic noise Nat Genet 2004 36 1073 1078 15448691 10.1038/ng1429
Angata T Hingorani R Varki NM Varki A Cloning and characterization of a novel mouse Siglec, mSiglec-F: differential evolution of the mouse and human (CD33) Siglec-3-related gene clusters J Biol Chem 2001 276 45128 45136 11579105 10.1074/jbc.M108573200
Angata T Margulies EH Green ED Varki A Large-scale sequencing of the CD33-related Siglec gene cluster in five mammalian species reveals rapid evolution by multiple mechanisms Proc Natl Acad Sci USA 2004 101 13251 13256 15331780 10.1073/pnas.0404833101
Crocker PR Siglecs: sialic-acid-binding immunoglobulin-like lectins in cell-cell interactions and signalling Curr Opin Struct Biol 2002 12 609 615 12464312 10.1016/S0959-440X(02)00375-5
Butler MG Bittel DC Kibiryeva N Talebizadeh Z Thompson T Behavioral differences among subjects with Prader-Willi syndrome and type I or type II deletion and maternal disomy Pediatrics 2004 113 565 5673 14993551 10.1542/peds.113.3.565
Varela MC Kok F Otto PA Koiffmann CP Phenotypic variability in Angelman syndrome: comparison among different deletion classes and between deletion and UPD subjects Eur J Hum Genet 2004 12 987 992 15470370 10.1038/sj.ejhg.5201264
Wu M Michaud EJ Johnson DK Cloning, functional study and comparative mapping of Luzp2 to mouse chromosome 7 and human chromosome 11p13-11p14 Mamm Genome 2003 14 323 334 12856284 10.1007/s00335-002-2248-6
Orr-Urtreger A Goldner FM Saeki M Lorenzo I Goldberg L De Biasi M Dani JA Patrick JW Beaudet AL Mice deficient in the alpha7 neuronal nicotinic acetylcholine receptor lack alpha-bungarotoxin binding sites and hippocampal fast nicotinic currents J Neurosci 1997 17 9165 9171 9364063
Arredondo J Nguyen VT Chernyavsky AI Bercovich D Orr-Urtreger A Kummer W Lips K Vetter DE Grando SA Central role of alpha7 nicotinic receptor in differentiation of the stratified squamous epithelium J Cell Biol 2002 159 325 336 12391028 10.1083/jcb.200206096
Morley BJ Rodriguez-Sierra JF A phenotype for the alpha 7 nicotinic acetylcholine receptor null mutant Brain Res 2004 1023 41 47 15364017 10.1016/j.brainres.2004.07.006
Bray C Son JH Kumar P Meizel S Mice deficient in CHRNA7, a subunit of the nicotinic acetylcholine receptor, produce sperm with impaired motility Biol Reprod 2005 73 807 814 15944242 10.1095/biolreprod.105.042184
Stankiewicz P Park SS Inoue K Lupski JR The evolutionary chromosome translocation 4;19 in Gorilla gorilla is associated with microduplication of the chromosome fragment syntenic to sequences surrounding the human proximal CMT1A-REP Genome Res 2001 11 1205 1210 11435402 10.1101/gr.181101
Bailey JA Baertsch R Kent WJ Haussler D Eichler EE Hotspots of mammalian chromosomal evolution Genome Biol 2004 5 R23 15059256 10.1186/gb-2004-5-4-r23
Mortlock DP Portnoy ME Chandler RL Green ED Comparative sequence analysis of the Gdf6 locus reveals a duplicon-mediated chromosomal rearrangement in rodents and rapidly diverging coding and regulatory sequences Genomics 2004 84 814 823 15475260 10.1016/j.ygeno.2004.07.009
MacDonald BT Adamska M Meisler MH Hypomorphic expression of Dkk1 in the doubleridge mouse: dose dependence and compensatory interactions with Lrp6 Development 2004 131 2543 2552 15115753 10.1242/dev.01126
Cunningham D Xiao Q Chatterjee A Sulik K Juriloff D Elder F Harrison W Schuster G Overbeek PA Herman GE Exma: an X-linked insertional mutation that disrupts forebrain and eye development Mamm Genome 2002 13 179 185 11956759 10.1007/s00335-001-2121-z
Smiraglia DJ Wu C Ellsworth MK Ratty AK Chapman VM Gross KW Genetic characterization of the chromosomal rearrangements that accompany the transgene insertion in the chakragati mouse mutant Genomics 1997 45 572 579 9367682 10.1006/geno.1997.4976
Kai M Irie M Okutsu T Inoue K Ogonuki N Miki H Yokoyama M Migishima R Muguruma K Fujimura H Kohda T Ogura A Kaneko-Ishino T Ishino F The novel dominant mutation Dspd leads to a severe spermiogenesis defect in mice Biol Reprod 2004 70 1213 1221 14695912 10.1095/biolreprod.103.024802
Morgan D Turnpenny L Goodship J Dai W Majumder K Matthews L Gardner A Schuster G Vien L Harrison W Elder FF Penman-Splitt M Overbeek P Strachan T Inversin, a novel gene in the vertebrate left-right axis pathway, is partially deleted in the inv mouse Nat Genet 1998 20 149 156 (Erratum in: 20, 312) 9771707 10.1038/2450
Qin Y Kong LK Poirier C Truong C Overbeek PA Bishop CE Long-range activation of Sox9 in Odd Sex (Ods) mice Hum Mol Genet 2004 13 1213 1218 15115764 10.1093/hmg/ddh141
Bejerano G Pheasant M Makunin I Stephen S Kent WJ Mattick JS Haussler D Ultraconserved elements in the human genome Science 2004 304 1321 1325 15131266 10.1126/science.1098119
Dermitzakis ET Reymond A Antonarakis SE Conserved non-genic sequences – an unexpected feature of mammalian genomes Nat Rev Genet 2005 6 151 157 15716910 10.1038/nrg1527
Loots GG Locksley RM Blankespoor CM Wang ZE Miller W Rubin EM Frazer KA Identification of a coordinate regulator of interleukins 4, 13, and 5 by cross-species sequence comparisons Science 2000 288 136 140 10753117 10.1126/science.288.5463.136
Nobrega MA Ovcharenko I Afzal V Rubin EM Scanning human gene deserts for long-range enhancers Science 2003 302 413 14563999 10.1126/science.1088328
Nobrega MA Pennacchio LA Comparative genomic analysis as a tool for biological discovery J Physiol 2004 554 31 39 14678488 10.1113/jphysiol.2003.050948
Gottgens B Barton LM Gilbert JG Bench AJ Sanchez MJ Bahn S Mistry S Grafham D McMurray A Vaudin M Amaya E Bentley DR Green AR Sinclair AM Analysis of vertebrate SCL loci identifies conserved enhancers Nat Biotechnol 2000 18 181 186 10657125 10.1038/72635
Henikoff S Conspiracy of silence among repeated transgenes Bioessays 1998 20 532 535 9723001 10.1002/(SICI)1521-1878(199807)20:7<532::AID-BIES3>3.0.CO;2-M
Freedman R Coon H MylesWorsley M OrrUrtreger A Olincy A Davis A Polymeropoulos M Holik J Hopkins J Hoff M Rosenthal J Waldo MC Reimherr F Wender P Yaw J Young DA Breese CR Adams C Patterson D Adler LE Kruglyak L Leonard S Byerley W Linkage of a neurophysiological deficit in schizophrenia to a chromosome 15 locus Proc Natl Acad Sci USA 1997 94 587 592 9012828 10.1073/pnas.94.2.587
Freedman R Leonard S Genetic linkage to schizophrenia at chromosome 15q14 Am J Med Genet 2001 105 655 657 11803510 10.1002/ajmg.1548
Perl O Ilani T Strous RD Lapidus R Fuchs S The alpha7 nicotinic acetylcholine receptor in schizophrenia: decreased mRNA levels in peripheral blood lymphocytes FASEB J 2003 17 1948 1950 12897067
Leonard S Gault J Hopkins J Logel J Vianzon R Short M Drebing C Berger R Venn D Sirota P Zerbe G Olincy A Ross RG Adler LE Freedman R Association of promoter variants in the alpha7 nicotinic acetylcholine receptor subunit gene with an inhibitory deficit found in schizophrenia Arch Gen Psychiatry 2002 59 1085 1096 12470124 10.1001/archpsyc.59.12.1085
Gault J Hopkins J Berger R Drebing C Logel J Walton C Short M Vianzon R Olincy A Ross RG Adler LE Freedman R Leonard S Comparison of polymorphisms in the alpha7 nicotinic receptor gene and its partial duplication in schizophrenic and control subjects Am J Med Genet B Neuropsychiatr Genet 2003 123 39 49 14582144 10.1002/ajmg.b.20061
Houy E Raux G Thibaut F Belmont A Demily C Allio G Haouzir S Fouldrin G Petit M Frebourg T Campion D The promoter-194 C polymorphism of the nicotinic alpha 7 receptor gene has a protective effect against the P50 sensory gating deficit Mol Psychiatry 2004 9 320 322 14569275 10.1038/sj.mp.4001443
Website title
Website title
Website title
|
16280085
|
PMC1322230
|
CC BY
|
2021-01-04 16:32:47
|
no
|
BMC Genomics. 2005 Nov 9; 6:157
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-157
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1541632421910.1186/1471-2407-5-154Research ArticleNeogenin expression may be inversely correlated to the tumorigenicity of human breast cancer Lee Jeong Eon [email protected] Hee Joung [email protected] Ji Yeon [email protected] Seok Won [email protected] Joon-Suk [email protected] Hyuk Jai [email protected] Wonshik [email protected] Sung-Won [email protected] Kyung-Sun [email protected] Dong-Young [email protected] Department of Surgery, Seoul National University College of Medicine, 28 Yongon-dong, Chongno-gu, Seoul 110-744, Korea2 Cancer Research Institute, Seoul National University College of Medicine, 28 Yongon-dong, Chongno-gu, Seoul 110-744, Korea3 Department of Surgery, Eulji University Hospital, 1306 Dunsan-dong, Seo-gu, Daejeon 302-799, Korea4 Laboratory of Stem Cell and Tumor Biology, Department of Veterinary Public Health, College of Veterinary Medicine, Seoul National University, San 56-1, Shilim-dong, Kwanak-gu, Seoul 151-742, Korea5 Department of Surgery Seoul National University Bundang Hospital, 300 Gumi-dong, Bundang-gu, Seongnam-Si, Gyeonggi-do 463-707, Korea2005 3 12 2005 5 154 154 22 1 2005 3 12 2005 Copyright © 2005 Lee 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
Neogenin is expressed in cap cells that have been suggested to be mammary stem or precursor cells. Neogenin is known to play an important role in mammary morphogenesis; however its relationship to tumorigenesis remains to be elucidated.
Methods
To compare the expression levels of neogenin in cells with different tumorigenicity, the expression levels in M13SV1, M13SV1R2 and M13SV1R2N1 cells, which are immortalized derivatives of type I human breast epithelial cells, were evaluated. Then we measured the expression level of neogenin in paired normal and cancer tissues from eight breast cancer patients. Tissue array analysis was performed for 54 human breast tissue samples with different histology, and the results were divided into four categories (none, weak, moderate, strong) by a single well-trained blinded pathologist and statistically analyzed.
Results
The nontumorigenic M13SV1 cells and normal tissues showed stronger expression of neogenin than the M13SV1R2N1 cells and the paired cancer tissues. In the tissue array, all (8/8) of the normal breast tissues showed strong neogenin expression, while 93.5% (43/46) of breast cancer tissues had either no expression or only moderate levels of neogenin expression. There was a significant difference, in the expression level of neogenin, in comparisons between normal and infiltrating ductal carcinoma (p < 0.001).
Conclusion
Neogenin may play a role in mammary carcinogenesis as well as morphogenesis, and the expression may be inversely correlated with mammary carcinogenicity. The value of neogenin as a potential prognostic factor needs further evaluation.
==== Body
Background
Neogenin, a member of the DCC (Deleted in Colorectal Cancer) receptor group, encodes a 1461 amino acid protein and has similar structure to DCC with 50.2% amino acid identity [1]. Neogenin is one of two groups of netrin-1 receptors [1,2]. Netrin-1 and its receptor groups were initially known to induce neurons and axons to target certain areas during neurogenesis [3]. Netrin-1 and neogenin do not only regulate axonal guidance but also play an important role in epithelial morphogenesis [4]. Neogenin is a cell adhesion molecule, and is expressed in a variety of developing tissues, such as the gastrointestinal tract of chickens [5], as well as on the surface of growing neurons [6].
Neogenin has been found to be expressed in cap cells at the terminal end buds of breast tissue. Cap cells have been suggested to be mammary stem or precursor cells located in the terminal ductal lobular units of the breast [7,8]. Recent data show that netrin-1 and its receptor neogenin provide an adhesive function during mammary gland morphogenesis; the binding of netrin-1 and neogenin stabilize the cap cell layer and prevent the influx of cap cells into the prelumenal compartment; this keeps the cap cells in their initial location [9]. The relationship between the expression of neogenin and the tumorigenicity of breast cancer has not been studied to date.
Kao et al. reported on obtaining two types of human breast epithelial cells (HBECs) in a breast tissue specimen from reduction mammoplasty [10]. Type I HBECs are known as potent breast stem cell candidates; they can differentiate into type II HBECs, and form budding/ductal organoids in Matrigel. It is also reported that type I HBECs have luminal epithelial cell markers, and stem cell characteristics, and that type II HBECs show basal epithelial phenotypes [11,12]. Nontumorigenic M13SV1 cells are immortalized cells obtained from the transfection of type I HBECs with simian virus 40 large T antigens. M13SV1R2 cells, obtained after x-ray irradiation (two doses of 2 Gy 8 days apart) of M13SV1 cells, express weak tumorigenicity when they are injected into athymic nude mice. The highly tumorigenic M13SV1R2N1 cells result from transduction of the rat neu oncogene into M13SV1R2 cells [11,13].
We hypothesized that neogenin might be related to mammary carcinogenesis. Therefore, we evaluated expression levels of neogenin in the M13SV1 cell line series (M13SV1, M13SV1R2, and M13SV1R2N1), the immortalized tumorigenic derivatives from type I HBECs, to compare the expression of neogenin in cells with different tumorigenicity. Then, we investigated neogenin expression in paired normal and cancer tissues from eight patients with infiltrating ductal cancer. Finally, to identify the pattern of neogenin expression on the basis of different histology, we performed tissue array analysis of 54 different human breast tissue samples.
Methods
Cells, cell culture and tissues
The M13SV1, M13SV1R2 and M13SV1R2N1 cells were cultured in MSU-1 medium, with 10% fetal bovine serum, containing 100 units of penicillin and 100 μg of streptomycin/ml, in a humidified incubator at 37°C with 5% CO2 [13].
To determine the expression levels of neogenin, in paired normal and cancer tissues, from the same patients, we selected paired specimens from eight patients and stored the samples in a liquid nitrogen tank. All of the patients had the diagnosis of an infiltrating ductal carcinoma. Normal tissues were sampled and were obtained from sites more than 5 cm away from the tumor margin of the cancer. All of the tissues were obtained and stored with the consent of each patient. The tissues to be used for the tissue array were stored in paraffin-embedded blocks.
Western blot analysis
The M13SV1, M13SV1R2 and M13SV1R2N1 cells were lysed in 20 mM HEPES (pH 7.2), 150 mM NaCl, 1% Triton X-100, 2% cholic acid, 1 mM EDTA, 1 mM EGTA, 0.1 mM DTT, 10 μg/ml leupetin, 10 μg/ml aprotinine and 1 mM phenylmetylsulfonyl fluoride (PMSF). Protein content (25 μg) was determined by the bicinchoninic acid (BCA) protein assay with bovine serum albumin as a standard. Cell lysates were released by heating samples at 95°C for 10 minutes in Laemmli cooking buffer and then separated by SDS (sodium dodecylsuldate)-polyacrylamide gel electrophoresis. Samples were transferred onto nitrocellulose filters and incubated separately with primary antibodies for two hours. Immunocomplexes were visualized using the peroxidase conjugated goat anti-rabbit IgG antibody. Detection was performed with the ECL Plex fluorescent Western blotting system (Amersham, Buckinghamshire, UK). The paired cancer and normal tissues from eight patients were treated in the same way. We have examined the expression of neogenin with 1:200 diluted anti-neogenin rabbit IgG polyclonal antibody (#sc-15337, Santa Cruz Biotechnology, California) as a primary antibody, and 1:5000 diluted horseradish peroxidase conjugated goat anti-rabbit IgG antibody (#sc-2004, Santa Cruz Biotechnology, California) as a secondary antibody.
Tissue array
Tissue array analysis was performed to examine the expression of neogenin in a variety of tissue types under the same conditions. We tested the tissues from the following sources: 8 normal, 31 infiltrating ductal carcinomas (IDC), 1 ductal carcinoma in situ (DCIS), 9 metastatic lymph nodes from breast cancer, and 5 other types of breast cancer cases (3 cases of invasive lobular carcinoma (ILC), one medullary carcinoma and one signet ring cell carcinoma). Tissues were embedded in paraffin; 4-μm sections, stained with hematoxylin and eosin (H&E), were obtained to identify morphologically representative areas of the specimen from which the core biopsies were taken. We marked the sampling site, on the corresponding blocks, with H&E stained slides. The blank paraffin blocks were then punched with a precision instrument (SuperBioChips Laboratory, South Korea), to make 2 mm × 60 holes; we then punched out the marked area of donor blocks, and filled in the holes of the recipient block. In this way, triplicate tissue cores, with diameters of 2 mm, were punched and arrayed on a recipient paraffin block. A layer of paraffin was coated to complete an array block with tightening of the paraffin. Then arrayed paraffin blocks were cut into 4-μm sections and placed on silane coated slides (MUTO pure chemicals, Tokyo, Japan). Sections from the paraffin-embedded tissue were deparaffinized, treated with 1% H2O2 in phosphate-buffered saline, and submitted for antigen retrieval by microwave oven treatment for 15 minutes in 0.01 mmol/L citrate buffers at pH 6.0. The slides were subsequently incubated in 10% normal horse serum for 30 minutes, followed by overnight incubation, at 4°C, with appropriately diluted rabbit polyclonal IgG antibodies to neogenin as a primary antibody and the horse radish peroxidase conjugated goat anti-rabbit IgG antibody as a secondary antibody. Fixation control was not used. Antibodies were the same as those used for the Western blot analysis, and diamino-benzidine (DAB) was used as a chromogen.
Statistical analysis
Fisher's exact test from the SAS statistical software Version 8 program (SAS Institute Inc., Cary, NC) was used for statistical calculation. All statistical analyses were two-tailed. We confirmed that all statistical results were reproducible with SPSS Version 11.5 (SPSS Inc., Chicago, IL).
Results
As a result of the Western blot analysis from the three cell lines, the data showed that the expression of neogenin was strongest in the nontumorigenic M13SV1 cells. Conversely, in the highly tumorigenic M13SV1R2N1 cells, neogenin was found to be only weakly expressed (Fig. 1). Western blot analysis of paired normal and IDC tissues from eight patients with breast cancer showed that the strength of expression of neogenin in normal tissue varied in each patient. However, there was a tendency toward higher expression of neogenin in normal tissues compared to the IDC tissues; this was clearly observed in 6 of 8 paired samples (Fig. 2).
Tissue array analysis was performed on 54 diverse human breast tissue samples with different histologic diagnoses (Table 1). The expression levels were measured, by a single blinded senior pathologist, based on the number of positive cells and the staining intensity; there were four categories: none, weak, moderate and strong, used for classification (Fig. 3). In concordance with the results of Western blot analysis, all of the normal breast tissues showed strong neogenin expression, while 93.5% (43/46) of breast cancer tissues showed no neogenin expression or only moderate levels. In 31 IDC cases, five (16.1%) did not have any neogenin expression, fourteen (45.2%) showed weak expression, ten (32.3%) showed moderate expression, and only two (6.5%) had strong levels of neogenin expression. There was a significant difference (p < 0.001) in the expression level of neogenin in normal compared to IDC cases.
Discussion
The results of the present study suggest that the expression of neogenin may be inversely related to the tumorigenicity of human breast cancer. In the Western blot analysis, neogenin expression is noted to be strong in nontumorigenic M13SV1 cells and decreased as the tumorigenicity increased; neogenin expression was weakest in the highly tumorigenic M13SV1R2N1 cells. The paired Western blot analysis from normal and IDC tissues in eight patients also resulted in an inverse pattern of neogenin expression with cancer. In the tissue array, all of the normal breast tissues showed a strong neogenin expression, while most cancer tissues showed weaker neogenin expression compared to normal tissues. These results suggest that the loss of neogenin expression might be important in mammary carcinogenesis.
Human mammary glands begin to develop with the primitive mammary epithelial cells growing into epithelium during the sixth week of fetal life. During the first three weeks of human life, the rudimentary mammary duct consists of a single duct gradually extending from the nipple. By the fifth week of age, the ducts fill approximately two-thirds of the fat pad, with a highly branched system that exhibits second and third order branching [14]. The terminal end bud (TEB) is the enlarged termini of ducts; a single layer of cap cells covers the tip of the end bud. TEBs are responsible for prodigious pubertal growth of cap cells and underlying preluminal epithelium. Cap cells have been suggested as mammary stem cells for ductal morphogenesis [7] or as cells serving as precursors to myoepithelial cells [8]. During puberty, the cap cells begin to grow into the adjacent fatty tissue and the parenchyma gains more fatty tissue and vessels. The accurate approximation of cap cells and preluminal epithelial cell layers is thought to be important for mammary morphogenesis; netrin-1 and neogenin play an important role in maintaining this approximation. The loss of neogenin during murine mammary morphogenesis results in widening of the subcapsular space; this anatomical detachment causes dissociation of the cap cells, and increased spillage of cap cells into the preluminal compartment. This may contribute to susceptibility for the development of human breast cancer [9,15].
Although a previous study reported that neogenin was expressed in breast cancer cell lines, and suggested that neogenin is not likely to be frequently affected in cancer, this study lacked important data at the translational level for the expression of neogenin in normal and breast cancer tissues [1]. Consistent with previous reports, neogenin was expressed in breast cancer cell lines in the present study. Moreover, we identified that there was a difference in the expression level of neogenin based on the tumorigenicity; this was observed as a result of experiments with three cell lines that have the same origin as the type I HBECs but different tumorigenicity. In the Western blot analysis for paired specimens, neogenin showed stronger expression in normal tissues than in each of the paired cancer samples.
It was difficult to determine the exact intensity of neogenin expression in tissue arrays with anti-neogenin polyclonal antibody. We could not find any commercial monoclonal anti-neogenin antibody, only polyclonal antibody. Future experiments with more specific monoclonal antibodies will be required for the precise demonstration of neogenin expression in tissues by immunohistochemical stain.
Conclusion
In summary, neogenin may play a role in mammary carcinogenesis as well as in mammary morphogenesis, and the expression of neogenin may be inversely correlated to mammary carcinogenicity. Further investigation is needed to identify the value of neogenin as a potential prognostic factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All the authors contributed to the conception during the initial stages and study design, and in the analysis and interpretation of the data, as well as to the drafting and critical revision of the important intellectual content. All the authors agreed to the final approval of the version to be published. JEL and HJK were equally involved in this study. JYB performed the Western analyses. DYN was in charge of the general supervision of this research. The coauthors declare the order of authorship was based on a joint decision.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by the grant No. 21-2003-006-0 from Seoul National University Hospital, grant No. 03-2005-008-0 from the Clinical Research Institute, Seoul National University Hospital and the grant no. SC-3180 from the Stem Cell Research Center, 21st Century Frontier R&D Program funded by the Ministry of Science & Technology of Korea.
Figures and Tables
Figure 1 The Western blot analyses of M13SV1, M13SV1R2 and M13SV1R2N1 cells. The expression of neogenin is strong in nontumorigenic M13SV1 cells while it is weak in highly tumorigenic M13SV1R2N1 cells. Detection is performed with the ECL Plex fluorescent Western blotting system (Amersham, Buckinghamshire, UK).
Figure 2 Western blot analyses of paired normal and cancer tissues from eight patients with infiltrating ductal carcinoma. There is a tendency toward higher neogenin expression in normal tissues compared to the IDC tissues; this is clear in samples 1–4, 6 and 8. Detection is performed with the ECL Plex fluorescent Western blotting system (Amersham, Buckinghamshire, UK).
Figure 3 The tissue array evaluated by a single pathologist showed different levels of neogenin expression in the various tissue samples. (A) There is no expression of neogenin observed in an IDC sample. (B) Another IDC sample shows a weak level of expression. (C) There is a moderate level of neogenin expression in an infiltrating lobular carcinoma sample. (D) A strong expression level of neogenin is observed in normal tissue. All pictures were magnified as × 200.
Table 1 The distribution of neogenin expression according to the histology in the tissue array
Histology Expression level of neogenin Total
None Weak Moderate Strong
Normal 0 0 0 8 8
Ductal carcinoma in situ 0 0 1 0 1
Infiltrating ductal carcinoma 5 14 10 2 31
Metastatic lymph node 2 3 4 0 9
Others* 0 1 3 1 5
Total 7 18 18 11 54
*Others consist of 3 cases of invasive lobular carcinoma, 1 medullary carcinoma and 1 signet ring cell carcinoma.
==== Refs
Meyerhardt JA Look AT Bigner SH Fearon ER Identification and characterization of neogenin, a DCC-related gene Oncogene 1997 14 1129 1136 9121761 10.1038/sj.onc.1200935
Keino-Masu K Masu M Hinck L Leonardo ED Chan SS Culotti JG Tessier-Lavigne M Deleted in Colorectal Cancer (DCC) encodes a netrin receptor Cell 1996 87 175 185 8861902 10.1016/S0092-8674(00)81336-7
Tessier-Lavigne M Goodman CS The molecular biology of axon guidance Science 1996 274 1123 1133 8895455 10.1126/science.274.5290.1123
Slorach EM Werb Z Epithelial morphogenesis: Netrin comes to a sticky and terminal end Curr Biol 2003 13 R491 3 12814570 10.1016/S0960-9822(03)00411-1
Vielmetter J Kayyem JF Roman JM Dreyer WJ Neogenin, an avian cell surface protein expressed during terminal neuronal differentiation, is closely related to the human tumor suppressor molecule deleted in colorectal cancer J Cell Biol 1994 127 2009 2020 7806578 10.1083/jcb.127.6.2009
Keeling SL Gad JM Cooper HM Mouse Neogenin, a DCC-like molecule, has four splice variants and is expressed widely in the adult mouse and during embryogenesis Oncogene 1997 15 691 700 9264410 10.1038/sj.onc.1201225
Williams JM Daniel CW Mammary ductal elongation: differentiation of myoepithelium and basal lamina during branching morphogenesis Dev Biol 1983 97 274 290 6852366 10.1016/0012-1606(83)90086-6
Sapino A Macri L Gugliotta P Pacchioni D Liu YJ Medina D Bussolati G Immunophenotypic properties and estrogen dependency of budding cell structures in the developing mouse mammary gland Differentiation 1993 55 13 18 7507867 10.1111/j.1432-0436.1993.tb00028.x
Srinivasan K Strickland P Valdes A Shin GC Hinck L Netrin-1/neogenin interaction stabilizes multipotent progenitor cap cells during mammary gland morphogenesis Dev Cell 2003 4 371 382 12636918 10.1016/S1534-5807(03)00054-6
Kao CY Nomata K Oakley CS Welsch CW Chang CC Two types of normal human breast epithelial cells derived from reduction mammoplasty: phenotypic characterization and response to SV40 transfection Carcinogenesis 1995 16 531 538 7697810
Chang CC Sun W Cruz A Saitoh M Tai MH Trosko JE A human breast epithelial cell type with stem cell characteristics as target cells for carcinogenesis Radiat Res 2001 155 201 207 11121235
Park JS Noh DY Kim SH Kim SH Kong G Chang CC Lee YS Trosko JE Kang KS Gene expression analysis in SV40-immortalized human breast luminal epithelial cells with stem cell characteristics using a cDNA microarray Int J Oncol 2004 24 1545 1558 15138599
Kang KS Sun W Nomata K Morita I Cruz A Liu CJ Trosko JE Chang CC Involvement of tyrosine phosphorylation of p185(c-erbB2/neu) in tumorigenicity induced by X-rays and the neu oncogene in human breast epithelial cells Mol Carcinog 1998 21 225 233 9585252 10.1002/(SICI)1098-2744(199804)21:4<225::AID-MC1>3.0.CO;2-J
Silberstein GB Postnatal mammary gland morphogenesis Microsc Res Tech 2001 52 155 162 11169863 10.1002/1097-0029(20010115)52:2<155::AID-JEMT1001>3.0.CO;2-P
Schubert C Breaking away from the breast Nat Med 2003 9 392 12669055 10.1038/nm0403-377b
|
16324219
|
PMC1322231
|
CC BY
|
2021-01-04 16:03:08
|
no
|
BMC Cancer. 2005 Dec 3; 5:154
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-154
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636390810.1371/journal.pmed.0020341Neglected DiseasesInfectious DiseasesMicrobiologyInfectious DiseasesMedicine in Developing CountriesMicrobiologyThe Global Campaign to Eliminate Leprosy Neglected DiseasesRinaldi Andrea Andrea Rinaldi is a science writer based in Cagliari, Italy. E-mail: [email protected]
Competing Interests: The author declares that no competing interests exist.
12 2005 27 12 2005 2 12 e341Copyright: © 2005 Andrea Rinaldi.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.While effective drug treatments have reduced the global disease burden, there remain important challenges to fighting and controlling the disease.
==== Body
Although leprosy is no longer a health problem in developed countries, it continues to affect millions of people in large parts of Asia, Africa, and Latin America. Effective chemotherapeutic treatments are available that have reduced the global disease burden dramatically, but there remain important challenges to fighting and controlling the disease.
Clinical Features
Leprosy is a chronic infection of the skin and peripheral nerves, caused by the obligate intracellular bacterium Mycobacterium leprae, the “Hansen's Bacillus” (Figure 1) [1]. Mainly transmitted by the aerosol spread of nasal secretions, the first symptoms may appear after an incubation period (usually five to ten years) following infection, and the onset is intermittent and gradual.
Figure 1 Photomicrograph of Mycobacterium leprae Taken from a Leprosy Skin Lesion
(Photo: Centers for Disease Control and Prevention)
Skin lesions and nerve damage are the main clinical features of the disease. Blindness may develop, resulting either from destruction of peripheral nerves within ocular tissues or from direct bacillary corneal invasion. Some patients—particularly those with lepromatous leprosy (see below)—may be affected by bacillary infiltration into the mucosa of the upper respiratory tract, bones, and testes [2]. Diagnosis of leprosy is mostly clinical and symptomatic, based on the presence of a few cardinal signs: hypopigmented or reddish-copper patches with definite sensory loss, with or without thickened nerves, and positive skin smears [3].
The social consequences for those affected with leprosy and for their families can be devastating.
The disease occurs in a wide spectrum of forms, which stem from the varying host immune response to the pathogen. Individuals with tuberculoid leprosy display a strong cell-based immune response that controls bacterial proliferation and lesions, whereas patients with lepromatous leprosy lack specific cellular immunity, ending up with high mycobacterial loads and severe clinical manifestations. Most patients have borderline forms. After the age of puberty, leprosy has a male to female ratio of 1.5–2.0 to 1. Britton and Lockwood have pointed out that this male preponderance is real—it is not related to underdiagnosis in women, although in some countries it is accentuated by delayed presentation by female patients, which results in higher rates of deformity [2].
The impairment of nerve function is due both to involvement of nerves by the primary infection, and to the acute immunological phenomena known as reversal reactions or type-1 leprosy reactions. These reactions occur in a third of patients with borderline forms of disease, are caused by spontaneous increases in T-cell reactivity to mycobacterial antigens, and are associated with the infiltration of interferon γ and TNFα-secreting CD4-positive lymphocytes in skin lesions and nerves, resulting in oedema and painful inflammation [2]. Sensory loss makes affected patients prone to inadvertent injury, leading to severe disabilities and visible deformities (Figure 2).
Figure 2 Typical Deformity in a Patient with Leprosy
(Photo: copyright WHO/P.Virot. This photo may not be reproduced for commercial purposes; see http://www.who.int/about/copyright/en/index.html)
Burden of Disease and Disability
Given the lifelong effects of nerve damage and its consequent disabilities—which often affect very young people—and the pressure this burden of disease and disability poses on fragile medical systems, the prevention, detection, and management of nerve function impairment are pivotal to all leprosy-control programmes [2].
Foot ulcers alone, for example, which are common in anaesthetic feet, can pose a huge burden on medical services. The social consequences for those affected with leprosy and for their families can be devastating. Stigma, community rejection, loss of employment, and sometimes forced isolation are still prevalent in both endemic and non-endemic countries [4].
The Biology of M. Leprae
One of the oldest recorded diseases [5], leprosy was also the first human pathogenic condition of bacterial origin for which the causative agent was identified. Despite these historical records, our knowledge of the biology of M. leprae has lagged behind, in good part because it cannot be grown in culture.
The full sequencing of the M. leprae genome, completed in 2001, has created possibilities for the development of new diagnostic tests and treatments for leprosy [6]. Analysis of the M. leprae genome has revealed that it contains fewer than half the functional genes of its closest relative, the tubercule bacillus M. tuberculosis [6,7]. This “minimal gene set”, the result of extensive gene deletion and decay that have eliminated many key metabolic pathways, renders the leprosy mycobacterium extremely slow in replicating and forces it to an intracellular existence.
A massive international effort was launched to eradicate leprosy worldwide.
In some areas such as the Middle East and Europe, leprosy declined after the late medieval period. One theory for the decline is that it was related to the increasing prevalence of tuberculosis—cross-immunity may have protected patients with tuberculosis from developing leprosy, or the compromised immunological status of patients with leprosy may have rendered them more susceptible to underlying latent tuberculosis infection, which resulted in increased mortality [8].
The Global Elimination Campaign
Leprosy now occurs mainly in resource-poor countries in tropical and warm temperate regions. Contrary to a widely believed myth, nowadays leprosy is a fully curable disease. A multidrug therapy (MDT) based on the combination of the antibiotics dapsone, rifampicin, and clofazimine was introduced in 1982 after dapsone-resistant strains appeared and spread. MDT proved highly efficacious in killing the bacteria without inducing resistance, although the optimal length of treatment and associated relapse rates are still controversial [2].
With such a powerful weapon at hand, a massive international effort was launched to eradicate leprosy worldwide. In 1991, the World Health Assembly adopted the target of “elimination of leprosy as a public health problem by the year 2000”. Elimination was defined as a reduction in the prevalence of patients with leprosy receiving antimicrobial therapy at a given time to less than 1 per 10,000 population. It was expected that by reducing the prevalence to this level, the transmission of M. leprae would be interrupted, leading to the gradual extinction of the disease. Since its introduction, some 13 to 14 million people have been cured with MDT (made available free of cost by the Sasakawa Foundation and then by Novartis), and full control of the disease (as assessed by prevalence rate) has been officially achieved in 112 of the 122 countries where leprosy was endemic in 1985.
The Final Push
The World Health Organization (WHO) dubbed the ambitious project “the final push to eliminate leprosy”. The strategy behind the slogan involves expanding MDT services to all health facilities and making leprosy diagnosis available, training health workers to diagnose and treat leprosy, promoting leprosy awareness and encouraging people to seek and continue treatment [9]. However, despite the impressive results obtained so far by the elimination campaign, this is still a work in progress.
According to the last WHO report (for 2003), ten countries in Africa, Asia, and Latin America still show prevalence rates above the selected threshold [10]. Topping this short list is a group of six endemic countries that together account for 83% of the leprosy cases registered worldwide: India, Brazil, Madagascar, Mozambique, Nepal, and Tanzania. According to WHO, in 2004 the number of patients with leprosy worldwide was 457,792 [11].
“We shall not be able to eliminate leprosy until we have a better understanding of its natural reservoir.”
Worryingly, whereas prevalence figures have fallen steadily in the last two decades, the annual rate of new cases did not follow a comparable trend—this rate has remained essentially unchanged over the past ten years. Indeed, the number of new cases detected during 1994 was 560,646, increasing to 804,357 in 1998, then falling again to 513,798 in 2003 [10,11]. On the basis of available information, WHO considers the “global target of leprosy elimination” as reached, and has shifted the strategy to the national level, for which elimination has been rescheduled for the end of 2005 [10]. In its plans, WHO estimates that eight out of the remaining ten countries will reach the new target, while India and Brazil will probably need additional time.
The key constraints to eliminating leprosy in those countries that lag behind the elimination campaign vary greatly from country to country. In some leprosy-endemic countries (such as Madagascar, Mozambique, Nepal, and Tanzania), access to many health facilities is extremely poor because of difficult terrain, displacement of populations in remote areas, or for security reasons [10]. In other countries, such as Brazil, important problems arise from the very centralised structure of the leprosy programme, and from its poor integration with general health services [10]. To deal with these very different scenarios, the strategies identified by WHO vary accordingly, proposing in some cases the complete restructuring of the national leprosy programmes [10].
A New Strategy for an Uncertain Future
So what will happen after 2005? Leprologists and people involved in disease control fear that once leprosy is declared “eliminated as a public health problem”, the future of anti-leprosy services and of leprosy workers and researchers will be at high risk [12,13]. Elimination is not eradication, many warn, and it must be clear to everyone that leprosy will continue to exist even in areas where the “elimination goal” has officially been reached. The term elimination itself makes people think the problem is over, say critics of the WHO policy, which can have detrimental effects on the future commitment of governments to sustain control activities, making it at the same time difficult for leprosy NGOs and scientists to raise funds for field and lab work.
Others believe that the concept of elimination itself, and the choice of prevalence as an indicator to measure the progress of the WHO-orchestrated campaign, are scientifically devoid of significance—as is the 2005 deadline. “As a matter of fact, the wrong indicator has been selected to reflect the progress toward elimination of leprosy,” says Piet Feenstra at the Royal Tropical Institute of Amsterdam (Amsterdam, The Netherlands), remarking that new-case detection and the proportion of children among new cases would serve much better to monitor the real disease status.
The “elimination” strategy must be swiftly converted to a “post-elimination” strategy.
The International Leprosy Association's Technical Forum has also noted that the expectation that reduction of prevalence to very low levels would lead to a reduction of the incidence within a few years was overoptimistic, as there was little evidence to support this hypothesis [14]. Since patients are only registered while they are on medication, prevalence figures by WHO standards vary depending on how long treatment lasts. “The decrease of prevalence is attributable primarily to the cleaning of the registers (discharge of cured or defaulting patients), to shortening the duration of treatment and, in some countries, to improved diagnostic accuracy, and is not a consequence of reduction of the transmission of Mycobacterium leprae,” Feenstra says.
“I believe there is probably a lot more leprosy in the world than the World Health Organization currently accepts,” agrees Helen Donoghue, a leprosy researcher at the Windeyer Institute of Medical Sciences (London, United Kingdom). The political implications of the “elimination goal”, and the way it was enforced by WHO, have also been questioned. “Over the last years, the elimination target has more and more become a political target [rather] than an epidemiological or program quality target,” says Feenstra. “For many, the indicator—the prevalence of patients registered for treatment—has become the goal in itself, and the actual goal—reduction of the leprosy transmission and incidence—has practically got out of sight”. Furthermore, the fixing of numerical targets may put excessive pressure on national leprosy programme managers, discouraging them from actively working to detect new cases, which in turn could jeopardise the country's elimination status.
Cracks in the Coalition
Another cause for serious concern is that the coalition that stands against leprosy is not as solid as it should be. In 1999, the Global Alliance for the Elimination of Leprosy (GAEL) was formed to inject new energy into the elimination campaign, bringing together WHO, the governments of the major endemic countries, the Japanese Nippon Foundation, the Novartis Foundation for Sustainable Development, the Danish Development Cooperation Agency (Danida), and the International Federation of Anti-Leprosy Associations (ILEP). Quite soon, major contrasts emerged between some of the GAEL partners, namely between WHO and ILEP, who always remained critical of the “elimination”-focussed strategy. The clash was so strong that ILEP was expelled from the alliance at the end of 2001.
Later on, probably in response to the increasing pressure to achieve leprosy control, WHO invited an independent team of experts led by Richard Skolnik, former Director of the Center for Global Health at George Washington University (Washington, District of Columbia, United States), to evaluate the GAEL. The evaluation report, published in 2003, recommends that WHO should take leprosy activities beyond 2005, dropping the “elimination” goal in favour of “an explicitly broad-based approach to the control of leprosy, the avoidance of nerve damage, and the rehabilitation of those in need” [15]. The team also explicitly called for the reconstitution of a refined alliance, where “collaborators will have to work more openly, collegially, and inclusively” [15].
There are signs that this new alliance is emerging. “The process of dialogue and collaboration with WHO headquarters in Geneva has already been reestablished and is improving constantly,” says Sunil Deepak, president of ILEP and medical director of the Italian leprosy NGO Associazione Italiana Amici di Raoul Follerau (AIFO). Deepak adds, “We are very optimistic about further strengthening of this collaboration”. Feenstra confirms that stakeholders are exploring new ways of dialogue. “WHO is currently, in collaboration with its partners, ILEP, The Nippon Foundation and Novartis, developing a new strategy for the period 2006–2010 for sustaining quality leprosy control activities,” he says (also see [16]).
A “Post-Elimination” Strategy
John Porter from the London School of Hygiene and Tropical Medicine (London, United Kingdom) recently argued that in order to make sure “the disease does not go underground”, the “elimination” strategy must be swiftly converted to a “post-elimination” strategy [12]. As recommended by many leprosy experts, this post-elimination strategy should focus on integrating leprosy control activities into primary health care services, assuring early case detection, adequate chemotherapy, prevention of disability for all patients with nerve damage, and physical rehabilitation of those already disabled [12,17,18].
Work to dispel the stigma of leprosy and to introduce patients back into their communities must also be strengthened, experts note, in order to end social discrimination toward people with leprosy [18]. Leprosy remains a disease of the poor, although the exact social factors that put people at risk have not been identified [19]. To break this link between leprosy and poverty, “leprosy should…now be included in the portfolio of diseases associated with poverty, and leprosy work incorporated into poverty reduction programmes,” points out Diana Lockwood of the London School of Hygiene and Tropical Medicine [20].
Lockwood and Suneetha have also suggested that the routine use of vaccination could benefit the outcome of WHO's anti-leprosy strategy [18]. Although the development of a specific and highly effective vaccine against leprosy is not yet a reality, the tuberculosis vaccine bacillus Calmette-Guérin, made with live attenuated Mycobacterium bovis with the eventual addition of heat-killed M. leprae, has been proven to offer some immunity to leprosy. Its reported efficacy ranges from 34% to 80% in different countries [2]. Despite this variable efficacy, bacillus Calmette-Guérin vaccination is already widely used in leprosy-endemic countries—but who should be vaccinated, when, and how often in order to achieve maximal protection among the population are all a matter for debate [21]. Another promising intervention for leprosy prevention comes from a recent study, conducted on five Indonesian islands, that found that giving people who are in close contact with patients with leprosy a short course of rifampicin can reduce their risk of developing the disease [22].
Finally, important research issues remain to be addressed. They include developing improved diagnostic tests and better ways to monitor and treat nerve damage, and understanding why MDT has not interrupted transmission [18]. “We shall not be able to eliminate leprosy until we have a better understanding of its natural reservoir,” says Donoghue. “There are several interesting reports that indicate that there may be an environmental reservoir for M. leprae—perhaps even in the soil in endemic regions.” Healthy human carriers also do exist, says Donoghue, who points out that by using sensitive molecular methods it is possible to detect the DNA from M. leprae in people who showed non-specific skin lesions and who had not been thought to have leprosy. Even after a ten-year MDT programme, more than 5% of healthy individuals in a leprosy-endemic area were positive for M. leprae DNA in their nasal passages, a recent study found, suggesting a high level of environmental contamination [23]. “Before anyone can talk about eliminating the disease we have to understand where the organisms are found, and the circumstances that result in an active infection,” Donoghue says.
Conclusion
It is widely believed that the leprosy elimination campaign has been a positive one. A comparably large consensus must now come together around the post-2005 leprosy agenda, to make sure that we do not lose the gains achieved to date or miss this unique opportunity to reach complete control of leprosy. Political, medical, or scientific disputes may have previously impeded unity among WHO and other stakeholders, but a reinforced partnership is now needed to continue the struggle to control and eradicate the disease.
Citation: Rinaldi A (2005) The global campaign to eliminate leprosy. PLoS Med 2(12): e341.
Abbreviations
GAELGlobal Alliance for the Elimination of Leprosy
ILEPInternational Federation of Anti-Leprosy Associations
MDTmultidrug therapy
WHOWorld Health Organization
==== Refs
References
Ustianowski AP Lockwood DNJ Leprosy: Current diagnostic and treatment approaches Curr Opin Infect Dis 2003 16 421 427 14501994
Britton WJ Lockwood DNJ Leprosy Lancet 2004 363 1209 1219 15081655
World Health Organization Leprosy 2005 Available: http://www.who.int/lep/disease/disease.htm . Accessed 5 May 2005
Withington SG Joha S Baird D Brink M Brink J Assessing socio-economic factors in relation to stigmatization, impairment status, and selection for socio-economic rehabilitation: A 1-year cohort of new leprosy cases in north Bangladesh Lepr Rev 2003 74 120 132 12862253
Ostrer BS Leprosy: Medical views of Leviticus Rabba Early Sci Med 2002 7 138 154 15045992
Cole ST Eiglmeier K Parkhill J James KD Thomson NR Massive gene decay in the leprosy bacillus Nature 2001 409 1007 1111 11234002
Vissa VD Brennan PJ The genome of Mycobacterium leprae : A minimal mycobacterial gene set. Genome Biol 2 2001 Available: http://genomebiology.com/2001/2/8/reviews/1023 . Accessed 24 October 2005
Donoghue HD Marcsik A Matheson C Vernon K Nuorala E Co-infection of Mycobacterium tuberculosis and Mycobacterium leprae in human archaeological samples: A possible explanation for the historical decline of leprosy Proc R Soc Lond B Biol Sci 2005 272 389 394
World Health Organization Guide to eliminate leprosy as a public health problem 2000 Available: http://www.who.int/lep/disease/Eliminate_Leprosy_V8.pdf . Accessed 28 April 2005
World Health Organization Leprosy elimination project Status report 2003 2004 Available: http://www.who.int/lep/Reports/s20042.pdf . Accessed 28 April 2005
World Health Organization Global leprosy situation 2004 2005 Available: http://www.who.int/lep/stat2002/global02.htm . Accessed 28 April 2005
Porter JDH Supporting ‘the individual’ with leprosy: The need for a ‘post-elimination strategy’ Lepr Rev 2004 75 2 6 15072121
Faber WR Klatser PR Elimination of leprosy and its consequences for research Adv Exp Med Biol 2003 531 295 297 12916801
International Leprosy Association Report of the International Leprosy Association's technical forum Lepr Rev 2002 73 S3 S61 12192965
Skolnik R Agueh F Justice J Lechat M Independent evaluation of the Global Alliance for the Elimination of Leprosy 2003 Available: http://gujhealth.gov.in/health_programmes/pdf/leprosy/GAEL%20Evaluation%20Report.pdf . Accessed 27 October 2005
World Health Organization Global strategy for further reducing the leprosy burden and sustaining leprosy control activities 2005 Available: http://www.who.int/lep/Reports/GlobalStrategy-PDF-verison.pdf . Accessed 13 July 2005
Visschedijk J Engelhard A Lever P Grossi MA Feenstra P Leprosy control strategies and the integration of health services: An international perspective Cad Saude Publica 2003 19 1567 1581 14999324
Lockwood DNJ Suneetha S Leprosy: Too complex a disease for a simple elimination paradigm Bull World Health Organ 2005 83 230 235 15798849
Kerr-Pontes LRS Montenegro ACD Barreto ML Werneck GL Feldmeier H Inequality and leprosy in northeast Brazil: An ecological study Int J Epidemiol 2004 33 262 269 15082624
Lockwood DNJ Leprosy and poverty Int J Epidemiol 2004 33 269 270 15082625
Smith WCS What is the best way to use BCG to protect against leprosy: When, for whom, and how often? Int J Lepr Other Mycobact Dis 2004 72 48 49 15217315
Bakker MI Hatta M Kwenang A Van Benthem BH Van Beers SM Prevention of leprosy using rifampicin as chemoprophylaxis Am J Trop Med Hyg 2005 72 443 448 15827283
Beyene D Aseffa A Harboe M Kidane D Macdonald M Nasal carriage of Mycobacterium leprae DNA in healthy individuals in Lega Robi village, Ethiopia Epidemiol Infect 2003 131 841 848 14596524
|
16363908
|
PMC1322279
|
CC BY
|
2021-01-05 10:40:28
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e341
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020341
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636390910.1371/journal.pmed.0020370Learning ForumDiabetes/Endocrinology/MetabolismObstetrics/GynecologyWomen's HealthEndocrinologyObstetricsWomen's HealthThyrotoxicosis and Pregnancy Learning ForumPerros Petros Petros Perros is in the Endocrine Unit, Freeman Hospital, Newcastle upon Tyne, United Kingdom. E-mail: [email protected]
Competing Interests: The author declares that no competing interests exist.
12 2005 27 12 2005 2 12 e370Copyright: © 2005 Petros Perros.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 35-year-old woman presented with a neck swelling after a missed abortion. Her thyroid function tests were in the thyrotoxic range. Perros discusses the further investigation and management of this patient.
==== Body
DESCRIPTION of CASE
A 35-year-old woman became aware of a swelling in her neck about ten weeks before referral. She had been trying to conceive for ten months. She had had a missed abortion five months earlier. Because of the neck swelling, her family doctor arranged thyroid function tests, which were in the thyrotoxic range on two occasions five weeks apart: serum free thyroxine, 26 and 28 pmol/l (normal range, 11–23); serum free tri-iodothyronine, 10.9 and 11 pmol/l (normal range, 3.5–6.5); serum thyroid-stimulating hormone (TSH), <0.05 mU/l (normal range, 0.3–4.1).
Her previous medical history included a partial thyroidectomy for thyrotoxicosis at the age of 24. Other than a goitre, she had no symptoms except increased appetite and a slight tremor, which she had been aware of for about eight weeks. Following the missed abortion, she had two normal menstrual periods. Her only medication was folic acid supplements. She worked part-time and had a two-and-a-half-year-old child.
On examination she was of average weight. Her hands were warm and moist. There was a fine tremor. A previous thyroidectomy scar was noted. The right lobe of the thyroid was palpable and felt smooth. There was a bruit over the right thyroid lobe on auscultation. She had lid retraction and lid lag but no other signs suggestive of thyroid-associated ophthalmopathy (TAO) (Figure 1). Her pulse rate was 100 beats per minute and regular. Her blood pressure was 150/70 mm Hg. The rest of the examination was normal.
Figure 1 The Patient's Appearance at the Time of Presentation with Recurrent Thyrotoxicosis
Note that there are no signs of TAO, but the patient has minimal upper lid retraction (the upper lid should normally be halfway between the limbus and the pupil).
What Is the Cause of Her Thyrotoxicosis?
Thyrotoxicosis is no more than a descriptor for a pattern of biochemical abnormalities. Before considering treatment, it is the clinician's task to define the underlying cause, as an accurate diagnosis is an essential guide to the most appropriate treatment (Box 1).
Box 1. Causes of Thyrotoxicosis
Common causes of thyrotoxicosis in a young female
Graves disease
Thyroiditis
Toxic multinodular goitre
Toxic adenoma
Iodine excess
Other rare causes of thyrotoxicosis
Hyperemesis gravidarum
Choriocarcinoma
TSH-producing pituitary adenoma
Iatrogenic thyrotoxicosis
Factitious thyrotoxicosis
Struma ovarii
Metastatic follicular thyroid cancer
Thyroid hormone resistance syndrome
The most likely causes in this case were Graves disease, thyroiditis, toxic multinodular goitre (TMNG), and toxic adenoma. The hallmark of TMNG or toxic adenoma is the presence of one or more palpable thyroid nodules. In this case the patient had previously undergone a partial thyroidectomy and a vascular thyroid remnant was palpable on the right thyroid lobe. Post-partum thyroiditis occurs within 12 months of childbirth; a variant of this condition occurs after miscarriage. In this patient's case post-partum thyroiditis was unlikely because her previous pregnancy was 2.5 years earlier; however, the miscarriage five months earlier may have been relevant. Viral thyroiditis is usually preceded by an upper respiratory tract infection and the thyroid gland is tender to touch; the absence of these features makes viral thyroiditis unlikely. “Silent” thyroiditis may present in this way and was a possibility here.
Laboratory tests that may help differentiate between the different causes of thyrotoxicosis include a radiolabelled technetium or iodide thyroid scan (Figure 2), and measurement of anti–thyroid peroxidase (TPO) antibodies, TSH receptor antibodies, and inflammatory markers (Table 1). The thyrotoxic phase of thyroiditis is usually followed by spontaneous euthyroidism and in some cases hypothyroidism. Repeating thyroid function tests within a few weeks of the first set may identify cases of thyroiditis.
Figure 2 Technetium 99 Thyroid Uptake Scans
(A) Normal.
(B) Graves disease: diffuse increased uptake in both thyroid lobes.
(C) TMNG: “hot” and “cold” areas of uneven uptake.
(D) Toxic adenoma: increased uptake in a single nodule with suppression of the surrounding thyroid.
(E) Thyroiditis: decreased or absent uptake.
(Image: Giovanni Maki)
Table 1 Diagnostic Tests for Identifying the Cause of Thyrotoxicosis
CRP, C-reactive protein; ESR, erythrocyte sedimentation rate.
In this case the prolonged time course of thyrotoxicosis, the presence of a vascular thyroid remnant, the persistently thyrotoxic thyroid function tests, and the elevated serum levels of TSH receptor antibodies (62 U/l; reference range, 0–10) were in favour of a diagnosis of recurrent Graves disease.
What Are the Effects of Thyrotoxicosis on Fertility and Risk of Abortion?
Menstrual irregularities occur in about 20% of thyrotoxic women [1]. Infertility is common in women with thyrotoxicosis even when they maintain ovulatory cycles [1]. Thyrotoxicosis also increases the risk of miscarriage to 26% [2].
How Should This Patient Be Treated?
There are three treatment options for thyrotoxicosis due to Graves disease: radioiodine (131I) therapy, thyroidectomy, and anti-thyroid drugs [3]. 131I therapy is safe and effective, but pregnancy should be deferred for 4–6 months after treatment as there are theoretical risks of fetal abnormalities. Most national regulatory authorities recommend avoidance of close contact with adults for a few days and with children and pregnant women for 2–3 weeks. 131I therapy was not appropriate for this patient because she wished to proceed with pregnancy as soon as possible and she had a two-and-a-half-year-old child, who would be difficult to care for after 131I therapy.
A second thyroidectomy is worthy of consideration, but involves general anaesthesia and a period of recuperation of a few weeks and therefore disruption of family and professional life. The risks of damage to the recurrent laryngeal nerves and parathyroid glands after a second thyroidectomy are considerably greater than after a first operation and are of the order of 5%–10%. Because of these considerations, thyroidectomy was not felt to be a suitable option.
Anti-thyroid drugs (carbimazole, methimazole, and propylthiouracil) restore euthyroidism within a few weeks of initiation of treatment [4]. Minor side effects (such as skin rashes) occur in about 5% of cases. Agranulocytosis is rare (∼0.4%), but the consequences are life threatening and all patients on anti-thyroid drugs must be made aware of this complication (Box 2). All anti-thyroid drugs have been used and are acceptable in pregnancy.
Box 2. Patient Information Leaflet Used by the Author to Remind Patients Receiving Anti-Thyroid Drugs of the Potential Complication of Agranulocytosis
“You have been started on a drug called Carbimazole/Methimazole/Propylthiouracil to control the activity of your thyroid gland. This is important treatment and Carbimazole/Methimazole/Propylthiouracil is a well established drug that has been used for many years. The great majority of people treated with Carbimazole/Methimazole/Propylthiouracil have no problems whatsoever.
“Some people occasionally develop a rash—if this happens please consult your doctor as soon as possible; you need not discontinue the drug unless he/she tells you to do so.
“More rarely, Carbimazole/Methimazole/Propylthiouracil affects white cells in the blood, in which case you would be likely to develop a very severe sore throat and to feel ill with a fever. If this happens while you are on Carbimazole/Methimazole/Propylthiouracil treatment you must attend either your family doctor or the hospital on the same day, to have your blood count checked. Take no more tablets until the blood count has been checked. If your white blood count is normal you can carry on with the Carbimazole/Methimazole/Propylthiouracil. If your white blood count is abnormal your family doctor or the hospital will need to deal with this problem urgently.
“Please keep this with you in case you need to show it to your doctor.”
Congenital anomalies have been reported in association with anti-thyroid drugs, but the increase in risk above background is very marginal. The risks of aplasia cutis and choanal and oesophageal atresia may be slightly lower with propylthiouracil than with other anti-thyroid drugs (choanal and oesophageal atresia, scalp defects, minor facial anomalies, and psychomotor delay compose an embryopathy implicated with methimazole use). But because the evidence is inconclusive and the additional risk minimal, all three drugs are widely used in pregnancy. The lowest dose of anti-thyroid drug that maintains euthyroidism should be used in women who wish to become or are already pregnant, in order to avoid fetal hypothyroidism and fetal goitre formation.
In this case propylthiouracil was used initially, at a dose of 50 mg four times per day. The patient was advised to take contraceptive measures until euthyroidism. Four weeks later her thyroid function tests had improved: serum free thyroxine, 13 pmol/l; serum total tri-iodothyronine, 2.5 nmol/l (normal range, 1.34–2.73); serum TSH, <0.05 mU/l. The dose of propylthiouracil was reduced to 25 mg four times per day, and the patient was advised that she could start trying to conceive.
What Are the Risks of TAO?
TAO is a complication that many patients fear. It can be disfiguring and difficult to treat [3]. If there are no clinical features of TAO at presentation, the risk of developing it in future is approximately 15%. Smoking is an important predisposing factor. As this patient was a non-smoker the probability of developing TAO is less than 10%.
What Monitoring Is Required during Pregnancy?
The dose of anti-thyroid drug usually needs to be decreased during pregnancy, and often Graves disease remits completely and the medication can be withdrawn. This is probably due to the overall immunosuppressive effect of pregnancy.
Monitoring of free thyroid hormone concentrations is of paramount importance during pregnancy and should be performed every 4–6 weeks, or more frequently if thyroid status is changing. The biochemical target is to achieve and maintain maternal serum free thyroxine levels at or slightly above the upper limit of normal, using the lowest dose of anti-thyroid drug possible. TSH receptor antibodies should be measured in the third trimester because positivity is predictive of neonatal thyrotoxicosis [5].
When the mother (as in this case) has a functioning thyroid gland or remnant in situ, maternal thyroid function mirrors that of the fetus. If there are concerns about fetal thyrotoxicosis (e.g., because maternal hyperthyroidism proves difficult to control), fetal heart rate monitoring should be undertaken. A persistent fetal tachycardia greater than 160 beats per minute is suggestive of fetal thyrotoxicosis. In cases where fetal thyrotoxicosis is diagnosed, monitoring of fetal growth and fetal goitre by ultrasound is imperative. In most cases the fetus can be treated satisfactorily by adjusting the dose of anti-thyroid drug in the mother and by following the fetal response clinically and by ultrasound.
What Are the Risks to the Fetus in a Woman with Graves Disease?
Poor control of maternal hyperthyroidism is associated with significant obstetric complications including miscarriage (26%), low birth weight, prematurity, (pre-)eclampsia, and possibly congenital malformations [6]. After the fetal thyroid matures (from 20 weeks of gestation onwards), maternal TSH receptor antibodies may act on the fetal thyroid to cause fetal thyrotoxicosis and goitre. The risk of fetal thyrotoxicosis is about 1% of all pregnancies in women with Graves disease, and if untreated, fetal mortality may be as high as 24%. Overtreatment may lead to hypothyroidism in the fetus, which is associated with subtle neurocognitive deficits later on in life, particularly if the hypothyroidism occurs in the first trimester [7]. Fetal goitre can develop as a result of fetal thyrotoxicosis or fetal hypothyroidism and in severe cases can obstruct labour.
What Are the Risks of Recurrence of Thyrotoxicosis after Delivery?
The risk of relapse of maternal thyrotoxicosis is high in the post-partum period (up to 80%), and close monitoring is required. Anti-thyroid drugs can be used safely during breastfeeding [8].
Prenatal Counselling of Women with Graves Disease
Pregnancy is a common concern among women of childbearing age who are receiving treatment for Graves disease. Some women may elect to have definitive treatment before pregnancy, which can be either a thyroidectomy or 131I therapy. The advantage of these treatment options is that the risk of maternal thyrotoxicosis during pregnancy is reduced, if not eliminated. Fertility is not affected by 131I therapy for thyrotoxicosis, but pregnancy should be deferred for 4–6 months after 131I therapy, although the basis of this recommendation is largely empirical. The risk of fetal and neonatal thyrotoxicosis is not eliminated by previous thyroidectomy or 131I therapy. The most important advice to women who have a previous history of thyroid dysfunction is to work with their practitioner to ensure that thyroid function tests are normal at the time of conception and throughout pregnancy.
DISCUSSION
Aetiology of Graves Disease
Graves disease is an autoimmune condition and is mediated by stimulatory autoantibodies to the TSH receptor. There is a significant genetic component to the aetiology of Graves disease, although environmental factors and stress also seem to confer risk [3]. Typically, the thyroid gland of patients with Graves disease is diffusely enlarged and vascular.
Toxic Multinodular Goitre
TMNG may also run in families. The pathogenesis is unknown. The disease begins with the formation of a single or few colloid nodules, and over a period of several years these become larger and more numerous. Some nodules are functioning and gradually acquire autonomy. With the passage of time serum TSH declines and may become undetectable until at a later stage serum free thyroid hormones rise. The hyperthyroidism of TMNG is usually mild and tends to occur in middle life or later. Toxic adenomas are benign neoplasms of the thyroid that are autonomous. In some cases they arise because of somatic mutations that lead to constitutive activation of the TSH receptor. As with TMNG, the hyperthyroidism tends to be mild.
Thyroiditis
Thyroiditis is due to an inflammatory process affecting the thyroid epithelium. Unlike other causes of thyrotoxicosis, there is no increased synthesis of thyroid hormones; instead, stored thyroid hormones in colloid are released into the circulation because of the leaky epithelium. The thyrotoxic phase of thyroiditis may be followed by a hypothyroid phase a few weeks later, but as a rule the patient recovers and euthyroidism ensues without any intervention.
Thyroiditis may occur after a viral infection (referred to as subacute or De Quervain thyroiditis), in which case the patient typically has a viral sore throat, the thyroid is tender, and inflammatory markers (erythrocyte sedimentation rate and C-reactive protein) are raised. “Silent” thyroiditis is autoimmune and characterised by positive anti-TPO antibodies.
Investigating the Cause of Thyrotoxicosis
In many cases of thyrotoxicosis the aetiology will be apparent from information that can be obtained from the history and clinical examination. In cases where there is doubt, additional investigations are indicated. Direct measurement of TSH receptor antibody levels is not widely available, but can be very valuable as modern assays are highly sensitive and specific. TSH receptor antibodies can occasionally be positive in post-partum thyroiditis (this seems to be particularly rare in Europe, though reported in North America and Japan), and in cases of doubt a thyroid scan showing no uptake of radioisotope is diagnostic of thyroiditis [9]. TSH receptor antibody measurement is indicated in pregnancy to assess the risks of fetal and neonatal thyrotoxicosis. Anti-TPO antibodies occur in a significant proportion of the normal population, and this limits the use of this test. High concentrations of anti-TPO antibodies are present in silent and post-partum thyroiditis. Radioisotope scans are useful in identifying the cause of thyrotoxicosis (Figure 2), but should be avoided in pregnancy.
Treatment of Thyrotoxicosis
The treatment of thyrotoxicosis depends on the underlying cause. Anti-thyroid drugs are effective in Graves disease, TMNG, and toxic adenoma (Table 2), but not in thyroiditis because the latter is not associated with increased de novo synthesis of thyroid hormones.
Table 2 Response of Common Causes of Thyrotoxicosis to 131I Therapy and Anti-Thyroid Drugs
After a course of anti-thyroid drug treatment, remission may be expected in Graves disease as a result of the immunosuppressive effect of anti-thyroid drugs on synthesis of TSH receptor antibodies, but relapse is the rule in cases of TMNG or toxic adenoma.
131I therapy is effective for Graves disease, TMNG, and toxic adenoma. 131I therapy is ineffective in thyroiditis because iodine uptake is reduced or absent in this condition (Figure 2). Most patients with Graves disease develop permanent hypothyroidism after 131I therapy, whereas most patients with TMNG and toxic adenoma do not. 131I therapy is associated with a small risk of exacerbation of new development of TAO, particularly in smokers.
Thyroiditis may require symptomatic treatment with beta blockers during the thyrotoxic phase.
The type of thyroidectomy (subtotal versus total) for Graves disease as primary treatment has been the subject of controversy for some years. The argument in favour of total thyroidectomy is that the risk of recurrence of the thyrotoxicosis is eliminated, and that if performed by skilled thyroid surgeons the probability of hypoparathyroidism and vocal cord palsy is no greater than for a subtotal thyroidectomy [10]. A subtotal thyroidectomy, on the other hand, provides the best chance of any treatment for Graves disease for long-term euthyroidism without the need for thyroxine or other treatments for thyrotoxicosis.
The choice of treatment for Graves disease should be tailored to the needs of the individual patient, but also depends on local facilities, surgical expertise, and patient choice.
Learning Points
Thyrotoxicosis is not a diagnosis, merely a biochemical result. An accurate clinical diagnosis encompassing the aetiology is imperative for optimal management.
The most common cause of thyrotoxicosis in women of childbearing age is Graves disease.
Thyrotoxicosis impairs fertility, and thyroid status should be assessed in women with secondary infertility or recurrent abortions.
Three treatments are available for thyrotoxicosis due to Graves disease: anti-thyroid drugs, 131I therapy, and thyroidectomy. The right treatment is that which suits the patient's individual circumstances best.
131I therapy is an absolute contraindication in pregnancy. Anti-thyroid drugs may be used safely, and the dose should be titrated to the minimum dose that maintains normal maternal thyroid hormone levels.
The hyperthyroidism of Graves disease usually remits after the first trimester, and anti-thyroid drugs can be withdrawn; however, relapse of maternal thyrotoxicosis in the post-partum period is common.
Uncontrolled maternal hyperthyroidism can lead to fetal thyrotoxicosis with devastating effects on the fetus. Fetal thyrotoxicosis can be treated satisfactorily by appropriate manipulation of the maternal dose of anti-thyroid drug and careful fetal monitoring.
Citation: Perros P (2005) Thyrotoxicosis and pregnancy. PLoS Med 2(12): e370.
The Learning Forum section editors are Susan Lightman and William Lynn.
Abbreviations
131Iradioiodine
TAOthyroid-associated ophthalmopathy
TMNGtoxic multinodular goitre
TPOthyroid peroxidase
TSHthyroid-stimulating hormone
==== Refs
References
Krassas GE Perros P Reproductive function in patients with thyroid diseases Hot Thyroidology 2 2002 Available: http://www.hotthyroidology.com/editorial_83.html . Accessed 12 October 2005
Abramson J Stagnaro-Green A Thyroid antibodies and fetal loss: An evolving story Thyroid 2001 11 57 63 11272098
Weetman AP Graves' disease N Engl J Med 26 2000 343 1236 1248
Cooper DS Antithyroid drugs N Engl J Med 2005 352 905 917 15745981
Laurberg P Nygaard B Glinoer D Grussendorf M Orgiazzi J Guidelines for TSH-receptor antibody measurements in pregnancy: Results of an evidence-based symposium organized by the European Thyroid Association Eur J Endocrinol 1998 139 584 586 9916861
Davis LE Lucas MJ Hankins GD Roark ML Cunningham FG Thyrotoxicosis complicating pregnancy Am J Obstet Gynecol 1989 160 63 70 2912104
Pop VJ Vulsma T Maternal hypothyroxinaemia during (early) gestation Lancet 2005 365 1604 1606 15885283
American Academy of Pediatrics Committee on Drugs Transfer of drugs and other chemicals into human milk Pediatrics 2001 108 776 789 11533352
Muller AF Drexhage HA Berghout A Postpartum thyroiditis and autoimmune thyroiditis in women of childbearing age: Recent insights and consequences for antenatal and postnatal care Endocr Rev 2001 22 605 630 11588143
Palit TK Miller CC Miltenburg DM The efficacy of thyroidectomy for Graves' disease: A meta-analysis J Surg Res 2000 90 161 165 10792958
|
16363909
|
PMC1322287
|
CC BY
|
2021-01-05 10:39:28
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e370
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020370
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636391010.1371/journal.pmed.0020400EssayInfectious DiseasesEpidemiology/Public HealthHealth PolicyHIV/AIDSMedical EthicsInternational healthMedicine in Developing CountriesInfectious DiseasesHIV Infection/AIDSMalariaMoral Imagination: The Missing Component in Global Health EssayBenatar Solomon R. Solomon R. Benatar is Professor of Medicine and Director of the Bioethics Centre, University of Cape Town, Cape Town, South Africa, and Visiting Professor in Medicine and Public Health Sciences at the University of Toronto, Toronto, Ontario, Canada. E-mail: [email protected]
Competing Interests: BSR is on the editorial board of PLoS Medicine. He declares that he has no competing interests.
12 2005 27 12 2005 2 12 e400Copyright: © 2005 Solomon R. Benatar.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.Benatar explores the underlying reasons for our failure to make adequate progress in improving global health.
==== Body
The deplorable state of global health and the failure to improve this state have been debated extensively. Recent editorials in the Lancet in relation to the failure of Roll Back Malaria and the potential failure of the 3 by 5 programme [1,2] illustrate how disappointment, surprise, and admonitions about such failures are usually followed by optimism about the success envisaged from future efforts [1,3].
There are several possible reasons for our failure to make adequate progress in improving global health. First, it seems that there is generally more interest in doing research to acquire new knowledge than in using existing knowledge, unless it is commercially profitable—illustrating how market forces are a more powerful influence on the practice of medicine than health needs [4]. Second, concern for those who are most severely affected by ill health seems to be generally transient, perhaps because they are anonymous and out of sight, but maybe also because their lives are less highly valued [5,6]. Third, there is a tendency to focus on new technologies through “silo” (narrowly contained) approaches to improving global health [7–9]. Fourth, there is insufficient attention to the social determinants of health [10,11].
Finally, while many are concerned about the plight of others, collective action through nongovernmental organisations can only achieve limited results, and there is reluctance to acknowledge and more explicitly address the indirect, causal, complex global system forces that underlie poverty and many fatal diseases [5,11–15]. Fortunately, there is now growing recognition that new infectious diseases pose a major threat to human health and security worldwide [16,17], and that imaginative new solutions are needed to improve global health [18,19].
Global distribution of wealth
(Figure adapted from [68])
While it is entirely appropriate to consider scientific and technological advances and economic growth as necessary for social progress, it is arguable that these will not be sufficient to ensure movement towards a more just world in which the health of whole populations could be improved. The controversy about globalisation versus antiglobalisation will not be revisited here, except to say that the debate should rather be about how globalisation can be modified to extend the benefits of progress more widely [20,21].
In this essay, I begin by suggesting that achieving substantial improvements in global health will depend on acknowledging that poor health at the level of whole populations reflects systemic dysfunction in a complex world. I then address why development aid is a necessary but not a sufficient solution for improving global health. I conclude with the idea that greater moral imagination (the ability of individuals and communities to empathise with others) and innovative 21st century approaches are required to break the impasse we currently face in improving global health.
An Unstable and Dysfunctional World
In the domain of economics, there is a disjunction between massive economic growth over the past 50 years and fair distribution of new wealth [22]. The global economy has increased 7-fold since 1950, yet the disparity in per capita gross domestic product between the 20 richest and the 20 poorest nations has more than doubled between 1960 and 1995 [23]. As a result, there are ever-widening disparities between rich and poor (Figures 1 and 2), and almost half the world's population lives on less than US$2 per day [24]. Disproportionate pursuit of short-term self-interest, fostered by market fundamentalism, emphasises production of goods for consumption by individuals, corporations, and governments, while long-term interests and the production of public goods for whole populations are undervalued [25].
Relationship between life expectancy and income per capita
(Figure adapted from [69])
Economic stability is threatened when aggregate economic growth is valued as an end in itself rather than as a means to improving human lives, and consequently, there is a failure to achieve a more just distribution of economic and social benefits [26]. Economic dysfunction persists when conventional economic theory continues to be revered and applied despite its many failures [26–29], and thus reduces the potential for improving global health and increasing human security worldwide [30].
In the domain of political and social life, instability is revealed by ongoing wars, ethnic conflict, fundamentalist attitudes, failed responses to genocide in many countries, large-scale disruption of communities, refugeeism, terrorism, fragmentation of health services, and attrition of public health-care services—all reflecting a lack of global leadership and a failure to achieve basic human rights for more people in the world [5,13,14,26,30,31]. Moreover, the full potential of the human rights approach is greatly diminished by a predominant focus on civil and political rights. Insufficient attention is paid to the social, cultural, and economic rights that are essential for human flourishing, and which are part of the “indivisible human rights” package described in the Universal Declaration of Human Rights, which most human rights activists use as their source of authority. It seems that higher value is placed on the rights and lives of those with resources than on the common good and the lives of the poor, and inadequate attention is given to identifying and motivating those who have duties to uphold a broad spectrum of rights [32,33].
These shortcomings, together with ecological instability from environmental degradation, global warming, and ongoing loss of biodiversity, arguably facilitate the creation of niches for the emergence and propagation of new infectious diseases, promote the development of multidrug resistance [34], and make it more difficult to maintain the social structures required to provide care and support for so many in need [35].
Development Aid: A Necessary but Insufficient Solution
Greenwood's call for increased development aid to provide the US$2–US$5 needed for each year of life that could be saved through an effective worldwide malaria control programme [3] resembles the approaches taken for tuberculosis and HIV/AIDS (http://www.theglobalfund.org). It must be gratefully acknowledged that generous philanthropy from concerned individuals and many foundations, organisations, and new global initiatives can, and do, make valued contributions to improving the health and health care of marginalised people in the world. Development aid from many countries should also be welcomed, and recent endeavours to increase aid from the current average of 0.23% gross domestic product to the recommended 0.7% are admirable [36].
However, development aid has been progressively reduced in recent years, and is increasingly being directed towards emergency humanitarian aid and the perceived security needs of wealthy nations, rather than towards sustainable development [37,38]. Therefore, the main problem is not merely lack of philanthropy and development aid. More poignantly, the problem is how the high profile given to relatively small amounts of aid eclipses recognition of the fact that financial, human, and other material resources are continuously being extracted from developing countries by wealthy nations striving for their own ongoing economic growth [19,24].
Modern trade rules [39], bribery and other means of controlling national economies and the lives of millions of poor people [40], and recruitment of health professionals trained at the expense of developing countries to sustain health care in wealthy countries [41] all reflect new forms of exploitation that result in much more being extracted from developing countries than is given to them in aid or in any other form. For example, annual farming subsidies of US$350 billion in industrialised countries [42] and trade protectionism cost developing countries US$50 billion annually in export earnings [43]. Allowing farmers in developing countries to sell their products at a fair price and not in competition with massive subsidies could eliminate the need for development aid [39,44].
Debt is another major problem. Poor countries' debt (US$2.2 trillion in 1997) has been associated with, and perpetuated through, arms trading (often coercively linked to aid) [45–47]. Such debt, particularly sub-Saharan Africa's debt of US$275 billion, fostered by both eager lenders and often corrupt borrowers can never be repaid. Sustaining debt perpetuates economic dependence and human misery. Resulting annual interest payments, of greater magnitude than the US$21 billion annual aid donated to Africa, cripple health and other social services and stultify development [48,49].
While some countries have achieved economic development, this has been generally less than desired and, sadly, lacking in most of sub-Saharan Africa. Moreover, much done in the name of development has been counterproductive, with adverse effects on the potential for globally improved human security [50–53]. The meaning of development and its evaluation needs to be reconsidered. Development means more than overall economic growth, and must include social progress, for example, in basic living conditions, education, and access to health care, so that all can have the opportunity to reach their achievable human capacities [50,54,55].
The unpalatable facts about how development is stultified are not being adequately confronted, and little attempt is made to acknowledge and address the complex systemic forces that sustain poverty and poor health [19,24]. Instead, obfuscation by politicians and indomitable optimism focused on economics, science, and human rights all promote continued hope for improving health in the developing world through market forces and new technologies [7–9].
Inadequate Moral Imagination
Some critical questions about world poverty have been asked and need to be answered [24]. Why does extreme poverty of almost half of humankind (income of less that US$2 per day) continue despite scientific, technological, economic, and moral progress? How do we explain why affluent individuals and wealthy nations are not morally embarrassed that so many people can be relegated to lives of poor quality with such limited opportunities to reach their full human potential? What does support (by individuals and nations) for processes that aggravate and sustain poverty tell us about ourselves and about the values we hold deeply? How can the rich remain secure in a world in which so many are so desperately poor that they may be provoked to rise up and rebel? Widening disparities within wealthy nations add another troubling dimension [56].
Many privileged people believe that poverty is not the fault of wealthy countries, but rather the result of bad government elsewhere. This is, indeed, partially true, and the prominent exposure of the extent of corruption and poor governance, for example in Africa [24,36,44,57], should be followed by sustained condemnation, retribution, and prevention. However, much less openly discussed is the complicity of powerful nations in supporting leaders who are despots and kleptocrats—by legitimising their right to sell their countries' natural resources, spend profligately on themselves, and incur debts that their impoverished citizens must repay [24,52]. Because wealthy nations, and by association their citizens, are deeply implicated in the generation and maintenance of forces that perpetuate social injustice and poverty, they need to face their responsibilities to alleviate the lives of those most adversely affected [24,52]. Reliance solely on perpetual philanthropy is clearly not the long-term solution to global health problems.
While we talk increasingly about disparities in wealth and health in an unjust world, most privileged people remain complacent about the suffering of the poor—both distant and within our midst [58,59]. In considering the many genocides across the world during the 20th century, Jonathan Glover has suggested that it is only moral imagination (our ability to imagine ourselves in the shoes of others) that can enable us to alter our outlook and actions significantly [60]. Our moral imagination is dulled, and insight into our global interdependence is diminished by insufficient public acknowledgment of how the quest of wealthy nations for endless economic growth, and luxuries that their citizens expect, has profoundly adverse effects on access to basic necessities of life for millions of others [24,26,49]. The ability to empathise with others requires the critical examination of our individual lives and of our nations' actions, the capacity to see ourselves as bound to all other human beings, and the sensitivity to imagine what it might be like to be a person living a very deprived and threatened life [24,61,62].
Making a diagnosis of social ills, like making diagnoses in medical practice, is much easier than providing effective remedies [63]. The magnitude and importance of achieving solidarity and cooperation in an interdependent world calls for a major research programme and considerable scholarship from many disciplines. Some pointers have been provided [18–21,24,64–66].
If lack of moral imagination were to be seen as one of the grand challenges for global health, resources and scholarly energy would surely be applied to promoting such imagination and to seeking innovative new approaches to improving global health. The quest for improved global health will be elusive if we continue to neglect the upstream forces that cause, sustain, and aggravate the poverty and misery that characterise the lives of almost half the world's population. The writing is on the wall [67].
I am grateful to Renee Fox (and to Amir Attaran, who reviewed the manuscript for PLoS Medicine) for their constructive comments. This work was supported in part by a grant from the United States National Institutes of Health's Fogarty International Center to the University of Cape Town's capacity building programme in International Research Ethics in Southern Africa (IRENSA Programme Director S. R. Benatar), and by the University of Toronto.
Citation: Benatar SR (2005) Moral imagination: The missing component in global health. PLoS Med 2(12): e400.
==== Refs
References
Reversing the failures of Roll Back Malaria Lancet 2005 365 1439 15856540
Predicting the failure of 3 by 5 Lancet 2005 365 1597 15889458
Greenwood BM Bojang K Whitty CJM Target GAT Malaria Lancet 2005 365 1487 1498 15850634
Trouiller P Olliaro P Torreele E Orbinski J Laing R Drug development for neglected diseases: A deficient market and a public health failure Lancet 2002 359 2188 2194 12090998
Benatar SR Global disparities in health and human rights Am J Public Health 1998 88 295 300 9491027
Gandy M Zumla A The return of the white plague: Global poverty and the ‘new’ tuberculosis 2003 London Verso 320
Singer PA Daar AS Harnessing genomics and biotechnology to improve global health equity Science 2001 294 87 89 11588248
Acharya T Daar AS Thorsteinsdóttir H Dowdeswell E Singer PA Strengthening the role of genomics in global health PLoS Med 2004 1 e40 10.1371/journal.pmed.0010040 15630462
World Health Organization Genomics and world health: Report of the Advisory Committee on Health Research 2002 Geneva World Health Organization Available: http://www.who.int/gb/ebwha/pdf_files/EB112/eeb1124.pdf . Accessed 26 October 2005
Marmot M Social determinants of health inequalities Lancet 2005 365 1099 15781105
Birn AE Gates' grandest challenge: Transcending technology as public health ideology Lancet 2005 366 514 519 Available: http://image.thelancet.com/extras/04art6429web.pdf . Accessed 26 October 2005 16084261
Benatar SR Van Niekerk AA Kopelman L The HIV/AIDS pandemic: A sign of instability in a complex global system Ethics and AIDS in Africa: The challenge to our thinking 2005 Cape Town David Philip Press 71 83
Beaglehole R Global public health: A new era 2003 Oxford Oxford University Press 284
Farmer P Pathologies of power 2003 Berkeley University of California Press 402
Green M Hulme D From correlates and characteristics to causes: Thinking about poverty from a chronic poverty perspective World Dev 2005 33 867 879
Garrett L The next pandemic? Foreign Aff 2005 84 3 23
Garrett L The lessons of HIV/AIDS Foreign Aff 2005 84 51 65
Benatar SR Daar AS Singer PA Global health ethics: The rationale for mutual caring Int Aff 2003 79 107 138
McKee M Gilmore AB Schwalbe N International cooperation and health. Part 1 J Epidemiol Community Health 2005 59 628 631 737 739 16020637
Sandbrook R Civilizing globalization: A survival guide 2003 Albany State University of New York Press 320
Royal Danish Ministry for Foreign Affairs Building a global community: Globalisation and the common good 2000 Copenhagen Royal Danish Ministry for Foreign Affairs 272
Coburn D Sandbrook R Globalization, neoliberalism and health Civilizing globalization 2003 Albany State University of New York Press 27 38
World Watch Institute Rich–poor gap widening 2003 Washington (D.C) Vital Signs 88 89 Available: http://www.worldwatch.org/features/vsow/2003/11/12/ . Accessed 26 October 2005
Pogge T World poverty and human rights 2002 Cambridge Polity Press 265
Kaul I Grunberg I Stern MA Global public goods: International cooperation in the 21st century 1999 New York Oxford University Press 546
Amin S Capitalism in the age of globalisation 1997 London Zed Books 158
Galbraith JK The affluent society 1960 Boston Houghton Mifflin 336
Galbraith JK The economics of innocent fraud: Truth for our time 2004 Boston Houghton Mifflin 64
Heilbroner R Milberg W The crisis of vision in modern economic thought 1995 Cambridge Cambridge University Press 125
Hong E Globalisation and the impact on health: A third world view 2000 Savar (Bangladesh) The Peoples' Health Assembly Available: http://www.phmovement.org/pubs/issuepapers/hong.html . Accessed 26 October 2005
Pugwash Conferences on Science and World Affairs Rotblat J World citizenship: Allegiance to humanity 1997 London Macmillan 234
Falk R Human rights horizons: The pursuit of justice in a globalizing world 2000 New York Routledge 288
Felice WF Taking suffering seriously: The importance of collective human rights 1996 Albany State University of New York Press 253
Garret L The coming plague: Newly emerging diseases in a world out of balance 1994 New York Farrar, Strauss and Giroux 768
Bakker I Gill S Power, production and social reproduction. Human in/security in the global political economy 2003 New York Palgrave Macmillan 256
Birdsall N How to help poor countries Foreign Aff 2005 84 136 152
Lancaster C Transforming foreign aid: United States assistance in the 21st century 2000 Washington (D.C.) Institute for International Economics 108
Woods N The shifting politics of foreign aid Int Aff 2005 81 393 409
Koivusalo M Deacon B Ollila E Koivusalo M Stubbs P The impact of WTO agreements on health and development policies Global social governances: Themes and prospects 2003 Helsinki Ministry for Foreign Affairs 77 129
Werlin H Corruption and foreign aid in Africa Orbis 2005 49 517 527
Dovlo D Taking more than a fair share? The migration of health professionals from poor to rich countries PLoS Med 2005 2 e109 10.1371/journal.pmed.0020109 15916458
Elliott L Don't let the rich cry poor over aid Guardian Weekly 2003 May 22 12
Kristof ND What did you do during the African holocaust? New York Times 2003 May 27 25 Sect A
Guest R The shackled continent: Power, corruption and foreign aid 2004 Washington (D.C.) Smithsonian Books 288
Sivard RL World military and social expenditures 1996, 16th edition 1996 Washington (D.C.) World Priorities Press 56
Sidel VW The international arms trade and its impact on health BMJ 1995 311 1677 1680 8541755
Pettifor A Debt, the most potent form of slavery: A discussion of the role of western lending policies in supporting the economies of poor countries 1996 London Debt Crisis Network 37
Anonymous Into Africa: Special report Toronto Star 2003 May 25 1 5 Sect F
Logie DE Benatar SR Africa in the 21st century: Can despair be turned to hope? BMJ 1997 315 1144 1146
Labonte R Schrecker T Sanders D Meeus W Fatal indifference: The G8 and global health 2004 Cape Town University of Cape Town Press 378
Hancock G Lords of poverty 1989 London MacMillan 234
Rist G The history of development: From western origins to global faith 1997 London Zed Books 320
Purdy M Countering terrorism: The missing pillar Int J 2005 60 3 24
Sen A Development as freedom 1999 New York Anchor Books 384
Bensimon C Benatar SR Developing sustainability: A new metaphor for progress Theor Med Bioeth 2005 In press
Raphael D Health inequalities in Canada: Current discourses and implications for public health action Crit Public Health 2000 10 193 216
Anonymous Erasing the scar. The Economist 2005 March 11 Available: http://www.economist.com/agenda/displayStory.cfm?story_id=3758983 . Accessed 26 October 2005
Galbraith JK The culture of contentment 1992 Boston Houghton Mifflin 195
Power S A problem from hell: America and the age of genocide 2002 New York Basic Books 656
Glover J Humanity: A moral history of the twentieth century 2001 New Haven Yale University Press 480
Nussbaum MC Cultivating humanity: A classical defense of reform in liberal education 1997 Cambridge (Massachusetts) Harvard University Press 328
Benatar SR Towards progress in resolving dilemmas in international research ethics J Law Med Ethics 2004 32 574 582 15807346
Benatar SR The coming catastrophe in international health: An analogy with lung cancer Int J 2001 56 611 631
Benatar SR Bankowski Z Bryant JH Gallagher J Streams of global change Ethics, Equity and Health for All 1997 Geneva Council for International Organizations of Medical Sciences 75 85
Benatar SR Cook G Zumla A Ethics and tropical diseases: A global perspective Manson's tropical diseases, 21st edition 2003 Edinburgh Elsevier Science 85 93
Benatar SR Fox RC Meeting threats to global health: A call for American leadership Perspect Biol Med 2005 48 344 361 16085992
Blair T Our common interest: Report of the Commission for Africa 2005 London Commission for Africa Available: http://www.commissionforafrica.org/english/report/thereport/english/11-03-05_cr_report.pdf . Accessed 26 October 2005
Anonymous The champagne glass of world poverty BMJ 1999 318 Available: http://bmj.bmjjournals.com/cgi/content/full/318/7189/0 . Accessed 31 October 2005
The World Bank The world development report 1993: Investing in health 1993 New York Oxford University Press 352
|
16363910
|
PMC1322290
|
CC BY
|
2021-01-05 10:39:26
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e400
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020400
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636391210.1371/journal.pmed.0020415PerspectivesRheumatologyRheumatologyConnective Tissue DiseaseTweaking Microtubules to Treat Scleroderma Perspectivevan Laar Jacob M *Huizinga Tom W. J Jacob M. van Laar is Associate Professor and Tom W.J. Huizinga is Professor of Experimental Rheumatology in the Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands.
Competing Interests: The authors declare that no competing interests exist.
*To whom correspondence should be addressed. E-mail: [email protected] 2005 27 12 2005 2 12 e415Copyright: © 2005 van Laar and Huizinga.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.
Paclitaxel Modulates TGFβ Signaling in Scleroderma Skin Grafts in Immunodeficient Mice
van Laar and Huizinga discuss a new study of a mouse model of scleroderma, which showed that stabilizing microtubules with paclitaxel led to reduced fibrosis.
==== Body
Systemic sclerosis (SSc; also referred to as “scleroderma”) is a rare but debilitating autoimmune disease clinically characterized by skin thickening and signs and symptoms of vasculopathy, which can involve the heart, lungs, kidneys, and gut. The disease spectrum can range from limited to diffuse disease, depending on the distribution of skin involvement, specificity of autoantibodies, and type of organ involvement. Patients with extensive skin thickening and organ dysfunction in particular are at risk of premature mortality [1].
The disease poses a challenge for the treating clinician, as no proven therapy exists that improves outcome, although recent data indicate that cyclophosphamide-based regimens may be effective in a subset of patients with early disease [2]. The etiology of SSc remains enigmatic, and few genetic and environmental predisposing factors have been identified.
Pathogenesis of SSc
Nevertheless, important aspects of its pathogenesis have been elucidated, particularly those related to progressive fibrosis, which is one of the hallmarks of the disease. Transforming growth factor β (TGFβ) is a pivotal cytokine in this process; it is a pleiotropic cytokine that induces matrix accumulation, regulates lymphocyte function and promotes endothelial cell apoptosis. Binding of TGFβ to the type II TGFβ receptor triggers its heterodimerization with, and activation of, type I TGFβ receptor. This activation results in a downstream signaling cascade with phosphorylation of specific receptor-regulated Smad (R-Smad) proteins (Smad2/3), which partner with Smad4 after dissociation from the TGFβ receptor (Figure 1). Smad2/3–Smad4 oligomers migrate to the nucleus, recruit other gene regulatory proteins, and activate transcription of specific target genes. In the absence of ligand stimulation, Smads reside predominantly in the cytoplasm; translocation of the activated R-Smad–Smad4 complex into the nucleus is a key step in signal transduction.
Figure 1 Simplified Model for Smad-Dependent Signaling Pathway Activated by TGFβ Showing the Consecutive Steps following TGFβ Binding to the Type II TGFβ receptor
Step 1: TGFβ binding to a type II receptor causes the receptor to recruit and phosphorylate a type I receptor. Step 2: phosphorylated type I receptor recruits and phosphorylates Smad2 or Smad3, upon which the Smads open up and expose a dimerization surface. Step 3: phosphorylated Smad2 or Smad3 dissociates from the receptor and oligomerizes with inhibitory Smad4. Step 4: the Smad2/3–Smad4 complex migrates to the nucleus, recruits other gene regulatory proteins (blue), and activates transcription of specific target genes.
Skin fibroblasts from patients with SSc express relatively high levels of TGFβ receptor, and contain high concentrations of R-Smad3 in the nucleus, while inhibitory Smad7 is functionally defective [3–5]. These and other data suggest that TGFβ signaling is constitutively activated in SSc fibroblasts, thus contributing to aberrant extracellular matrix synthesis. The important role of Smads in fibrosis is illustrated by the finding that Smad3-deficient mice are resistant to different forms of fibrosis. Not surprisingly, the TGFβ/Smad axis has been identified as a therapeutic target in fibrotic conditions such as SSc.
A New Study in a Mouse Model of SSc
A study published in this issue of PLoS Medicine by Liu et al. [6] shows that, in a hybrid human SSc skin–severe combined immunodeficient mouse xenotransplant model, stabilizing microtubules using paclitaxel (Taxol; a powerful anticancer agent and angiogenic inhibitor isolated from the bark of the Pacific yew tree) reduces production of phosphorylated Smad2/3 and expression of COL1A2 (one of the genes involved in production of collagen, whose promoter contains multiple Smad-binding elements). The end result is to lessen fibrosis histologically. The study takes advantage of an important animal model for scleroderma, the engraftment of SSc skin samples in immunodeficient mice. These samples have previously been shown to retain their phenotype and abnormal Smad expression [7]. The study also builds on previous work that has shown that microtubules provide a negative feedback loop in TGFβ signaling in cell lines by forming a complex with endogenous Smad2, Smad3, and Smad4, sequestering R-Smads away from the TGFβ receptor [8]. Taken together, these studies suggest that modulating TGFβ/Smad signaling with paclitaxel may be an effective means to treat skin fibrosis.
The Role of Other Signaling Cascades
However, recent data indicate that other signaling cascades are also perturbed [9], and it is, therefore, conceivable that the beneficial effects of paclitaxel on scleroderma skin thickening are not solely due to changes in TGFβ/Smad signaling. One of the read-outs of fibrogenesis in the study of Liu et al. is reduced expression of COL1A2, an essential gene involved in the biosynthesis of collagen. However, this process is complex: extensive posttranslational modification of the COL1A2 gene product occurs during the fibrotic process in which many key enzymes such as telopeptide lysyl hydroxylase are involved [10]. Future studies should address the effect of paclitaxel on the expression of the wide array of enzymes involved in fibrosis by genome-wide expression studies in patients treated with paclitaxel or ex vivo on scleroderma skin samples.
Next Steps
By contrast, scleroderma-like changes in patients with cancer have been ascribed to the use of taxanes, including paclitaxel [11]. Whether, as suggested by Liu et al., this paradoxical effect on skin relates to the use of low doses in the mouse model described by them rather than the high doses used in patients with cancer remains to be determined, but the point underscores the need for further studies. Further work is also needed on the in vivo effects of paclitaxel on the vasculature and immune abnormalities in SSc patients, which are difficult to evaluate using scleroderma skin grafts in immunodeficient mice. At the low doses used in the studies by Liu et al. no antiangiogenic effect was found.
Clearly, there is a delicate balance between microtubule stabilizing and destabilizing forces in scleroderma, which paclitaxel may alter. These findings suggest, however, that a small pilot study of such therapy in selected patients with diffuse SSc, though a daring endeavor, may be worth the risk.
Citation: van Laar JM, Huizinga TWJ (2005) Tweaking microtubules to treat scleroderma. PLoS Med 2(12): e415.
Abbreviations
R-Smadreceptor-regulated Smad
SScsystemic sclerosis
TGFβtransforming growth factor β
==== Refs
References
Medsger TA Clements PJ Furst DE Classification. Prognosis Systemic sclerosis, 2nd ed 2004 London Lippincott Williams and Wilkins 17 28
Clements Ph Furst DE Silver RM Tashkin DP Roth MD The Scleroderma Lung Study shows the beneficial effects of cyclophosphamide over placebo in systemic sclerosis patients with active alveolitis Arthritis Rheum 2005 52 S257
Kawakami T Ihn H Xu W Smith E LeRoy C Increased expression of TGF-beta receptors by scleroderma fibroblasts: Evidence for contribution of autocrine TGF-beta signaling to scleroderma phenotype J Invest Dermatol 1998 110 47 51 9424086
Mori Y Chen SJ Varga J Expression and regulation of intracellular SMAD signaling in scleroderma skin fibroblasts Arthritis Rheum 2003 48 1964 1978 12847691
Asano Y Ihn H Yamane K Kubo M Tamaki K Impaired Smad7-Smurf-mediated negative regulation of TGF-beta signaling in scleroderma fibroblasts J Clin Invest 2004 113 253 264 14722617
Liu X Zhu S Wang T Hummers L Wigley FM Paclitaxel modulates TGFβ signaling in scleroderma skin grafts in immunodeficient mice PLoS Med 2005 2 e354 10.1371/journal.pmed.0020354 16250671
Lakos G Takagawa S Chen SJ Ferreira AM Han G Targeted disruption of TGF-β/Smad3 signaling modulates skin fibrosis in a mouse model of scleroderma Am J Pathol 2004 165 203 217 15215176
Zhu S Goldschmidt-Clermont PJ Dong C Transforming growth factor-β-induced inhibition of myogenesis is mediated through Smad pathway and is modulated by microtubule dynamic stability Circ Res 2004 94 617 625 14739161
Pockwinse SM Rajgopal A Young DW Mujeeb KA Nickerson J Microtubule-dependent nuclear-cytoplasmic shuttling of Runx2 J Cell Physiol 2005 E-pub ahead of print
Van der Slot AJ Zuurmond AM Bardoel AF Wijmenga C Pruijs HE Identification of PLOD2 as telopeptide lysyl hydroxylase, an important enzyme in fibrosis J Biol Chem 2003 278 40967 40972 12881513
Farrant PBJ Mortimer PS Gore M Scleroderma and the taxanes. Is there really a link? Clin Dermatol 2005 29 360 362
|
16363912
|
PMC1322299
|
CC BY
|
2021-01-05 10:39:31
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e415
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020415
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636391310.1371/journal.pmed.0020417PerspectivesNeuroscienceOtherGeriatricsNeurology/NeurosurgeryDementiaBreaking Up (Amyloid) Is Hard to Do PerspectiveGandy Sam *Heppner Frank L Sam Gandy is at the Farber Institute for Neurosciences, Thomas Jefferson University, Philadelphia, Pennsylvania, United States of America. Frank L. Heppner is at the Institute of Neuropathology, University Hospital Zurich, University of Zurich, Zurich, Switzerland.
Competing Interests: SG is an advisor to Elan Pharmaceuticals, and receives grant support from the Robert Atkins Foundation. He is Chair of the National Medical and Scientific Advisory Council of the United States Alzheimer's Association. FLH declares that he has no competing interests.
*To whom correspondence should be addressed. E-mail: [email protected] 2005 27 12 2005 2 12 e417Copyright: © 2005 Gandy and Heppner.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.Gandy and Heppner discuss the implications of a new animal study on our understanding of the pathogenesis and treatment of Alzheimer disease.
==== Body
Standard “Alzheimerology” lore holds that the insolubility of amyloid plaques and neurofibrillary tangles was a great impediment to elucidating the molecular composition of each of these structures. Roughly 77 years passed between Alois Alzheimer's description of the clinical and pathological features of the illness suffered by Auguste D. and the now classical reports from George Glenner, Colin Masters, and Konrad Beyreuther describing partial solubilization, Edman degradation, and primary amino acid sequencing of the Aβ peptide (the protein that accumulates into amyloid plaques) [1–3]. Peter Davies (who discovered the cholinergic deficiency in Alzheimer disease during that same interval [4]) has long joked that one easy way to purify plaques and tangles is to allow the brain from an affected person to autolyze and liquefy completely (“on a summer sidewalk,” according to one colorful variation), whereupon the macro structure of the organ would disintegrate, leaving behind only fibrous clumps and twists.
There is little wonder, then, that the strategy of treating Alzheimer disease with “plaque busting” drugs was relatively slowly embraced: 20 years passed between the availability of amyloid aggregation assays and the clinical trials of the first specific antiamyloid-aggregation compounds. (“Aggregation” is essentially equivalent to clumping, and amyloid clumping can be monitored in a test tube since floating clumps cause light to disperse in a measurable fashion). Conventional wisdom was nihilistic, holding that therapeutic dissolution of amyloid deposits was probably too slow to be approachable. People love to see a dogma challenged, and the availability of mouse models of Alzheimer disease enabled Brad Hyman's group to show, in 2001, that deposits of human Aβ in the transgenic mouse brain were surprisingly dynamic, forming and, unexpectedly, dissolving over a timescale of days [5]. These data injected new optimism into the pursuit of antiamyloid strategies, and by 2005, over 30 discrete compounds or combinations were in development [6].
Old Worries Resurface
Now, with publication of a study by David Borchelt's group in this issue of PLoS Medicine [7], old worries about the efficiency of plaque clearance resurface. David Borchelt, Joanna Jankowsky, and their colleagues report the development of Aβ plaque-forming transgenic mice in which pathology is driven by brain-specific overexpression of a mutated form of the human amyloid precursor protein (APP). The innovation here is that human APP expression due to a genetic trick is extinguishable by adding the antibiotic tetracycline to the mouse food (“tet-off” APP mice). As rightly contended by these authors, the ability to abolish human APP gene expression—instantly and completely—can be conceptually envisioned as equivalent to the most effective antiamyloid strategy imaginable: in other words, a best-case scenario from the point of view of drug efficacy.
The rationale was to see how long human amyloid deposits would persist in a plaque-laden mouse brain once new accretion ceased (i.e., once tetracycline switched off new mutated human APP production). The results are arguably applicable to every antiamyloid strategy delivered to patients impaired by Alzheimer disease, since all are believed to enter therapy with at least some existing brain plaque burden. In this tet-off paradigm, unlike the paradigm used earlier by Hyman and colleagues, amyloid pathology was allowed to accumulate, and then human APP expression was completely shut off. Borchelt and colleagues were unable to detect any change in brain amyloid load for at least six months after complete cessation of Aβ biogenesis. The Borchelt data dovetail well with recent biophysical data, proving that the thermodynamic barrier to redissolution of amyloid fibrils is very high indeed (Figure 1) [8].
Figure 1 A Large Energy Barrier Prevents Rapid Redissolution of Fibrillar Amyloid
In this issue of PLoS Medicine, Borchelt and colleagues demonstrate in the living amyloid-laden mouse brain that Aβ plaques are cleared very slowly, even if synthesis of new Aβ precursor molecules is extinguished using a tet-off system [7]. In the recent, relevant, but independent, X-ray crystallography study [8], Nelson et al. envisioned the free-energy plot shown above as a graphic description of the kinetics of transition from monomeric Aβ to fibrillar Aβ (ΔGformation). For the reverse reaction, Nelson et al. envision the ΔGdissolution as the large free-energy barrier to spontaneous solubilization of amyloid fibrils. Presumably, it is this ΔGdissolution that underlies the slow disappearance of brain plaques in the Borchelt study in transgenic mice.
(Illustration: Sapna Khandwala, adapted from [8])
Implications for Therapy and Prevention
If this mouse model represents the best-case scenario, what are the realistic hopes of success for anti-Aβ therapies in the treatment of human Alzheimer disease? Several points come to mind. First, the recent proposal that oligomeric/dodecameric forms of Aβ (also known as Aβ-derived diffusible ligands [ADDLs]) are the real culprits in mediating Aβ neurotoxicity [9] provides some hope that ADDLs, not Aβ plaques, represent the most important Aβ biophysical form that must be purged in order to cause a therapeutic benefit. Borchelt and colleagues have not yet determined what happens to ADDL levels in the brains of their tet-off APP mice, but such data are eagerly anticipated.
Second, Borchelt and colleagues' data are consistent with the evidence that all model antiamyloid strategies are superior when initiated long before onset of amyloidosis. Such evidence fuels a new initiative of the United States Alzheimer's Association aimed not at treating but at preventing Alzheimer disease (http://www.alz.org/maintainyourbrain). Interventional strategies early in life should not only prevent amyloid pathology, but also take advantage of the greater regenerative capacity of the human brain at younger ages. The wisdom of preventing Alzheimer disease is further galvanized by the appreciation over the last five years of evidence that many modifiable diet and lifestyle factors (body mass index, cholesterol levels, control of diabetes and blood pressure, and mental and physical exercise) may modulate risk for late-life degenerative dementia [10].
The Challenges Ahead
Significant challenges lie ahead in understanding how these risks and pathologies are intertwined, but the aging of the baby boom population promises an unprecedented epidemic of dementia if effective interventions are not discovered soon. In the US, the population of individuals age 65 years and over is the fastest growing segment in society, and one-half of individuals over 85 experience dementia [11]. On a practical level, this means that almost everyone will be either a patient or a caregiver in the near future. Disasters such as tsunamis and hurricanes—as painful, dramatic, and expensive as they are—pale in comparison to the cataclysm our aging societies will face if current trends in the epidemiology of this dementing illness continue on unchecked.
This work was supported by United States National Institutes of Health grants NS41017 (to SG) and NS046006 (to FLH).
Citation: Gandy S, Heppner FL (2005) Breaking up (amyloid) is hard to do. PLoS Med 2(12): e417.
Abbreviations
ADDLAβ-derived diffusible ligand
APPamyloid precursor protein
==== Refs
References
Alzheimer A Stelzmann RA Schnitzlein HN Murtagh FR An English translation of Alzheimer's 1907 paper, “Uber eine eigenartige Erkankung der Hirnrinde.” Clin Anat 1995 8 429 431 8713166
Glenner GG Wong CW Alzheimer's disease: Initial report of the purification and characterization of a novel cerebrovascular amyloid protein Biochem Biophys Res Commun 1984 120 885 890 6375662
Masters CL Simms G Weinman NA Multhaup G McDonald BL Amyloid plaque core protein in Alzheimer disease and Down syndrome Proc Natl Acad Sci U S A 1985 82 4245 4249 3159021
Davies P Biochemical changes in Alzheimer's disease-senile dementia: Neurotransmitters in senile dementia of the Alzheimer's type Res Publ Assoc Res Nerv Ment Dis 1979 57 153 166 33427
Bacskai BJ Kajdasz ST Christie RH Carter C Games D Imaging of amyloid-beta deposits in brains of living mice permits direct observation of clearance of plaques with immunotherapy Nat Med 2001 7 369 372 11231639
Kwon MO Fischer F Matthisson M Herrling P List of drugs in development for neurodegenerative diseases Neurodegenerative Dis 2004 1 113 152
Jankowsky JL Slunt HH Gonzales V Savonenko AV Wen J Persistent amyloidosis following suppression of Aβ production in a transgenic model of Alzheimer disease PLoS Med 2005 2 e355 10.1371/journal.pmed.0020355 16279840
Nelson R Sawaya MR Balbirnie M Madsen Aø Riekel C Structure of the cross-beta spine of amyloid-like fibrils Nature 2005 435 773 778 15944695
De Felice FG Vieira MN Saraiva LM Figueroa-Villar JD Garcia-Abreu J Targeting the neurotoxic species in Alzheimer's disease: Inhibitors of Abeta oligomerization FASEB J 2004 18 1366 1372 15333579
Marx J Neuroscience. Preventing Alzheimer's: A lifelong commitment? Science 2005 309 864 866 16081709
Hebert LE Scherr PA Beckett LA Albert MS Pilgrim DM Age-specific incidence of Alzheimer's disease in a community population JAMA 1995 273 1354 1359 7715060
|
16363913
|
PMC1322301
|
CC BY
|
2021-01-05 10:39:54
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e417
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020417
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636391410.1371/journal.pmed.0020420Correspondence and Other CommunicationsPathologyControlling the Spread of HIV/AIDS in the Indian Subcontinent CorrespondenceAgoramoorthy Govindasamy
1
Hsu Minna J
2
1Tajen UniversityYanpu, TaiwanRepublic of China2National Sun Yat-sen UniversityKaohsiung, TaiwanRepublic of ChinaE-mail: [email protected]: [email protected]
Competing Interests: The authors have declared that no competing interests exist.
12 2005 27 12 2005 2 12 e420Copyright: © 2005 Agoramoorthy and Hsu.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.
==== Body
The article on HIV/AIDS infection by Singh and colleagues outlines an alarming fact about the spread of this deadly virus in Nepal [1]. We would like to add that more assertive campaigns are necessary to curb the spread of infection in the Indian subcontinent before it's too late. In the year 2000 alone, a total of 5.3 million people were infected with HIV worldwide [2]. Since the epidemic started two decades ago, HIV/AIDS has killed 22 million people globally. India, Indochina, and the former Soviet republics have seen the most rapid rise of HIV incidence in recent years. AIDS experts have raised alarm bells over its spread in the Asia-Pacific region, and called for a united effort to control it. The Joint United Nations Programme on HIV/AIDS (UNAIDS) estimates about 5 million people in India alone are infected.
The first report of HIV/AIDS infection in India was in 1986, and since then the virus has spread rapidly throughout the country. Both HIV serotypes 1 and 2 exist in India, and HIV-1 C is the most common subtype reported. Sexual transmission of HIV is the predominant route of transmission in India [3]. According to the Ministry of Health in New Delhi, only 3% of Indians use condoms for birth control, since the tradition and culture dictate that women undergo sterilization or take birth control pills. Prostitution plays a major role in spreading the disease among the heterosexuals in urban areas. Although Mumbai appears to be the main locus for AIDS, rapid spread has occurred through other major cities as well. Migration of people from cities to rural areas is so rapid that the disease may already be out of control in many areas. The blood screening tests conducted at most hospitals in rural India are not adequate to confirm presence of the virus, making blood transfusion unsafe. The National AIDS Control Organization (NACO), the apex body for controlling AIDS in India, has reported a high incidence (8.2%) of blood donors who are HIV-positive among healthy blood donors in urban areas [4].
AIDS is a sexually transmitted disease, and as long as people are educated thoroughly and warned about the dangerous consequences of unsafe sex, there is less to fear. Unfortunately, the intervention program launched by the NACO had very little impact in controlling the spread of the epidemic in India [4]. The current educational programs are often restricted to the passive dissemination of information through posters, media, and display of safe-sex billboards behind automobiles. More aggressive efforts are, therefore, needed to reach out to each and every rural/urban community throughout India to combat the spread of the disease. The state and central government agencies must build specialized shelters for people with HIV/AIDS. More funds must be spent for effective AIDS awareness campaigns, research, routine screening tests, and treatment.
According to the Asia Pacific Network of People Living with AIDS, a considerable number of people were refused treatment or delayed provision of treatment or health services after being diagnosed with HIV/AIDS. Breaches of confidentiality by health workers were common in Asian countries. Within families and communities, women were discriminated against more than men—including ridicule, harassment, and physical assault—and they were often forced to change their place of residence because of their HIV status [5].
Although politicians and policymakers are increasingly committed to AIDS prevention and control efforts in countries such as India, a multidisciplinary approach such as early identification and treatment of sexually transmitted diseases, promotion of condom usage, rapid blood screening to test for HIV in rural areas, public awareness campaigns, poverty eradication, and development of prevention interventions have to be considered for effective control of the spread of this virus in the Indian subcontinent.
Moreover, people from all walks of life must take an active role to promote AIDS awreness and prevention across the Indian subcontinent. It is time for the local and regional celebrities, such as political leaders, movie stars, and beauty pageant winners, in the Indian subcontinent to get involved in helping people with HIV and in educating the public, which would certainly raise awareness among the rural public more quickly than current efforts. It is time to remember how the late Princess of Wales reached out to people with AIDS, shook hands to console them, and also raised millions of dollars for their welfare. Countries in the Indian subcontinent have experienced and handled the outbreak of deadly epidemics in the past [6], and we hope that AIDS can also be controlled and eradicated eventually in the near future.
Citation: Agoramoorthy G, Hsu MJ (2005) Controlling the spread of HIV/AIDS in the Indian subcontinent. PLoS Med 2(12): e420.
==== Refs
References
Singh S Mills E Honeyman S Suvedi BK Pant NP HIV in Nepal: Is the violent conflict fuelling the epidemic? PLoS Med 2005 2 e216 10.1371/journal.pmed.0020216 16008505
Joint United Nations Programme on HIV/AIDS AIDS epidemic update 2002 Geneva Joint United Nations Programme on HIV/AIDS Available: http://www.unaids.org/html/pub/Publications/IRC-pub03/epiupdate2002_en_pdf.pdf . Accessed 9 November 2005
Godbole S Mehendale S HIV/AIDS epidemic in India: Risk factors, risk behaviour and strategies for prevention and control Indian J Med Res 2005 121 356 368 15817949
Choudhury N Ayagiri A Ray VL True HIV seroprevalence in Indian blood donors Transfus Med 2000 10 1 4
Paxton S Gonzales G Uppakaew K Abraham KK Okta S AIDS-related discrimination in Asia AIDS Care 2005 17 413 424 16036226
Karlen A Man and microbes Disease and plagues in history and modern times 1995 New York Simon and Schuster 266
|
16363914
|
PMC1322304
|
CC BY
|
2021-01-05 10:39:27
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e420
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020420
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636391510.1371/journal.pmed.0020421Correspondence and Other CommunicationsOtherOncologyWomen's HealthCancer: BreastClinical TrialsOncologyThe Statistical Significance of Suffering CorrespondenceSuthers Kristen
1
1GDNF 4 Parkinson'sWashington, District of ColumbiaUnited States of AmericaE-mail: [email protected]
Competing Interests: The author has declared that no competing interests exist.
12 2005 27 12 2005 2 12 e421Copyright: © 2005 Kristen Suthers.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.
When Clinical Trials Are Compromised: A Perspective from a Patient Advocate
==== Body
Musa Mayer makes several good points about the importance of enrolling people with life-threatening conditions in clinical trials in order to identify new treatments and speed the pipeline along for the greater good [1]. However, the idea that clinical trial enrollment suffers when seriously ill individuals are provided compassionate use of treatments is myopic; one does not negate the other. In many cases, persons who seek compassionate use of medications are ineligible for the clinical trials Mayer would want them to enroll in, and will likely die or suffer considerably before the experimental treatment they are seeking is approved for the public. In a world of limited resources, we need to ask, how do we encourage enrollment in clinical trials to develop treatments and cures that will benefit people in the future, while humanely treating those who are ineligible for these trials and suffer right now? The first step is to understand that clinical trial enrollment and compassionate-use programs are not competing interests today, as they perhaps were in the 1980s and 1990s. The next step is to educate the public, not only about the importance of enrollment in clinical trials, but about their rights as informed participants in the noble process of science. Mayer's perspective [1] fails to consider the ultimate goal of clinical trials: to relieve human suffering. It serves no one's interest to demand an all-or-nothing approach to scientific progress. As Einstein said, “Not everything that can be counted counts, and not everything that counts can be counted.”
Citation: Suthers K (2005) The statistical significance of suffering. PLoS Med 2(12): e421.
==== Refs
Reference
Mayer M When clinical trials are compromised: A perspective from a patient advocate PLoS Med 2005 2 e358 10.1371/journal.pmed.0020358 16220998
|
16363915
|
PMC1322305
|
CC BY
|
2021-01-05 10:39:29
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e421
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020421
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636391610.1371/journal.pmed.0020422Correspondence and Other CommunicationsHealth PolicyClinical trialsMedical InformaticsResearch MethodsCorrection/Clarification about FDA Review Documents CorrespondenceTurner Erick
1
1Portland VA Medical Center and Oregon Health and Science UniversityPortland, OregonUnited States of AmericaE-mail: [email protected]
Competing Interests: ET is author of the PLoS Medicine Essay cited in this Editorial and discussed in the present response. Also, ET is a former reviewer (medical officer) with the FDA.
12 2005 27 12 2005 2 12 e422Copyright: © 2005 Erick Turner.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.
Tackling Publication Bias in Clinical Trial Reporting
Editor's Reply
==== Body
Emma Veitch cites my PLoS Medicine Essay [1] about how the Food and Drug Administration's (FDA's) review of documents can serve as a source of clinical trials data, but she follows it up with the statement, “However, it is difficult to have confidence in data released by sponsors when the data have not been subjected to external, independent peer review. Furthermore, this information is not integrated with other data, or indexed” [2].
While I agree with the second assertion, the first assertion—that the data are not subjected to external, independent peer review—is off the mark. FDA reviews are indeed external and independent to the sponsor. These reviews are conducted not by the sponsors but by physicians and scientists employed by the United States government. True, the data originate with the sponsor. However, once the sponsor submits data to the FDA, a level of rigor and scrutiny is applied to them that is arguably higher than what occurs in the typical journal manuscript review process.
First, FDA reviewers typically revisit the original protocol submitted before the study was conducted in order to verify that the sponsor has not engaged in hypothesizing after the results are known (“HARKing”) [3]. By contrast, journal reviewers typically do not have access to the original protocol. As a result, they must trust that HARKing has not occurred, a dubious assumption in view of recent data [4].
Second, FDA statistical reviewers obtain the raw data from the sponsor, and determine whether the sponsor's findings can be replicated. By contrast, journal reviewers typically have access to only the summary statistics reported (perhaps selectively) to them by the authors or the sponsors. Consequently, reviewers can only speculate whether they could replicate the findings.
As a result, I believe that the FDA review process warrants a higher level of confidence than the conventional journal manuscript review process.
Citation: Turner E (2005) Correction/clarification about FDA review documents. PLoS Med 2(12): e422.
==== Refs
References
Turner EH A taxpayer-funded clinical trials registry and results database PLoS Med 2004 1 e60 10.1371/journal.pmed.0010060 15562322
Veitch E PLoS Medicine Editors Tackling publication bias in clinical trial reporting PLoS Med 2005 2 e367 10.1371/journal.pmed.0020367 17523250
Kerr NL HARKing: Hypothesizing after the results are known Pers Soc Psychol Rev 1998 2 196 217 15647155
Chan AW Hrobjartsson A Haahr MT Gotzsche PC Altman DG Empirical evidence for selective reporting of outcomes in randomized trials: Comparison of protocols to published articles JAMA 2004 291 2457 2465 15161896
|
16363916
|
PMC1322306
|
CC BY
|
2021-01-05 10:39:32
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e422
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020422
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 10.1371/journal.pmed.0020423Correspondence and Other CommunicationsHealth PolicyClinical trialsMedical InformaticsResearch MethodsEditor's Reply CorrespondenceVeitch Emma
1
1PLoS Clinical TrialsCambridgeUnited KingdomE-mail: [email protected]
Competing Interests: EV is Publications Manager for PLoS Clinical Trials.
12 2005 27 12 2005 2 12 e423Copyright: © 2005 Emma Veitch.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.
Tackling Publication Bias in Clinical Trial Reporting
Correction/Clarification about FDA Review Documents
==== Body
Erick Turner appropriately points out the high levels of rigor applied during regulatory authorities' review of clinical trial data [1]. However, the statement beginning “However, it is difficult to have confidence in data released by sponsors…” [2] was not intended to highlight the release of review documents by the Food and Drug Administration (FDA), but rather the publication of summary clinical trial data on sponsors' own Web sites, which does seem to lack an integral peer-review mechanism. I support efforts to make Drugs@FDA more systematic and comprehensive, an initiative which can sit comfortably alongside peer-reviewed journal publication.
Citation: Veitch E (2005) Editor's reply. PLoS Med 2(12): e423.
==== Refs
References
Turner E Correction/clarification about FDA review documents PLoS Med 2005 2 12 e422 16363916
Veitch E the PLoS Medicine Editors Tackling publication bias in clinical trial reporting PLoS Med 2005 2 e367 10.1371/journal.pmed.0020367 17523250
|
0
|
PMC1322307
|
CC BY
|
2021-01-05 10:39:29
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e423
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020423
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1636391810.1371/journal.pmed.0020424EditorialMedical EducationMedical EducationPublic HealthMedicine in Developing CountriesImproving Health by Investing in Medical Education EditorialThe PLoS Medicine Editors E-mail: [email protected]
Competing Interests: At the Negombo meeting, Health Action International Asia Pacific covered Gavin Yamey's accommodation, meals, and airport transfers, but PLoS paid for his flight.
12 2005 27 12 2005 2 12 e424Copyright: © 2005 The PLoS Medicine Editors. 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 author and source are credited.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 education in South Asia is reinventing itself, to create doctors who can meet the needs of the local community.
==== Body
One of the common criticisms of medical education is that there is often a mismatch between what is taught at medical school and the actual skills that are needed by doctors to provide locally relevant health care. This mismatch is particularly striking in South Asia, according to the 60 educators who gathered for the Second International Consultation on Undergraduate Medical and Pharmacy Education, held in Negombo, Sri Lanka, in September 2005. The aim of the consultation, organized by the health-activist group Health Action International Asia Pacific and the World Health Organization, was to consider how best to prepare medical students to meet the health needs of the region.
The needs seem overwhelming. South Asia faces high rates of communicable and noncommunicable diseases, road traffic injuries, maternal and child mortality and morbidity, rising tobacco use, violent conflicts, and the devastating effects of recent floods and earthquakes. And yet there is hope. Zulfiqar Bhutta and colleagues have argued that “the answers to the region's problems may already be with us” (BMJ 328: 777–778). They point out that despite a civil war, Sri Lanka has the best health indicators in the region—better than those of most other countries with comparable incomes. Sri Lanka's average life expectancy is 73 years, infant mortality is 16 per 1,000, and maternal mortality is 30 per 100,000 live births. India's Kerala state has similarly impressive health indicators, which are better than the national average. These two examples show what can be achieved when governments spend their limited resources on education (leading to high literacy rates) and on providing community-based primary care, rather than building expensive specialist hospitals.
If all countries in the region are to emulate the success of Sri Lanka and Kerala, they will need, among other things, to reorient their medical schools away from teaching students in acute hospital settings and toward community-based education. At the Negombo consultation, Qasem Chowdhury, Vice Chancellor of the Peoples' University-Gono Bishwabiddyaloy, Institute of Health Sciences, Bangladesh, laid out the challenge. Meeting the health needs of South Asia, he said, “requires a new type of educational program for health personnel that will make them responsive to the needs of the majority population of individual countries. Such training is most effective if it is carried out in close relation to the actual community in which health personnel are later to work.”
The educators at the consultation were united in calling for curricular reform. The traditional curriculum—overloaded with basic sciences teaching, focused on curative rather than preventive medicine, emphasizing factual learning rather than acquisition of skills, knowledge, and attitudes—is inappropriate for South Asia. And the neglect of social, economic, cultural, and political perspectives in undergraduate training means that doctors can never conduct their practice with an understanding of the fundamentally social nature of disease in the region.
Several inspiring examples of curricular reform were presented, such as that of the University of Sri Jayewardenepura, Sri Lanka. With support from the World Bank and Sri Lanka's Ministry of Higher Education, the university is adopting a new medical curriculum that integrates basic and clinical sciences, and emphasizes community-based learning and capacity building for research. The university is establishing “laboratories” for learning clinical skills, communicating with patients, and using information technology for self-directed learning and clinical practice. Clearly, different medical schools in South Asia will have different resources at their disposal for overhauling their curricula (the Afghani delegates at the meeting, for example, were desperately short of textbooks), but, nevertheless, all can take at least tentative steps toward problem-based, community-oriented, integrated teaching.
What other steps should schools take to produce doctors who can appropriately serve their communities? The same answers recurred throughout the consultation. Schools should assess students not just on knowledge but on broader skills that are essential for promoting community health. Students should be taught about rational drug prescribing, medical ethics and human rights, and the traditional systems of medicine that are hugely popular among their patients. And faculty development (training the teachers in educational skills) must be at the heart of the reforms.
When a consultation ends, the real work begins. Although there was a consensus about what kind of change is needed, there are many unanswered questions about the best way to bring about these reforms. One contentious question, for example, is whether the private sector has a role to play in providing medical education, particularly in parts of South Asia where public education is failing. How can such a question be answered? One suggestion at the consultation was to establish a regional network of community-based educators, to allow educators to share their experiences of—and research on—their educational reforms. We look forward to providing an update on these concerns in a future issue of PLoS Medicine.
Citation:
PLoS Medicine Editors (2005) Improving health by investing in medical education. PLoS Med 2(12): e424.
|
16363918
|
PMC1322308
|
CC BY
|
2021-01-05 10:39:31
|
no
|
PLoS Med. 2005 Dec 27; 2(12):e424
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020424
|
oa_comm
|
==== Front
Nutr Metab (Lond)Nutr Metab (Lond)Nutrition & Metabolism1743-7075BioMed Central 1743-7075-2-311628865510.1186/1743-7075-2-31ReviewCarbohydrate restriction improves the features of Metabolic Syndrome. Metabolic Syndrome may be defined by the response to carbohydrate restriction Volek Jeff S [email protected] Richard D [email protected] Human Performance Laboratory, Department of Kinesiology, University of Connecticut, Storrs, CT 06269-1110 USA2 Department of Biochemistry, SUNY Downstate Medical Center, Brooklyn, NY 11203 USA2005 16 11 2005 2 31 31 25 10 2005 16 11 2005 Copyright ©2005 Volek and Feinman; licensee BioMed Central Ltd.2005Volek and Feinman; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Metabolic Syndrome (MetS) represents a constellation of markers that indicates a predisposition to diabetes, cardiovascular disease and other pathologic states. The definition and treatment are a matter of current debate and there is not general agreement on a precise definition or, to some extent, whether the designation provides more information than the individual components. We consider here five indicators that are central to most definitions and we provide evidence from the literature that these are precisely the symptoms that respond to reduction in dietary carbohydrate (CHO). Carbohydrate restriction is one of several strategies for reducing body mass but even in the absence of weight loss or in comparison with low fat alternatives, CHO restriction is effective at ameliorating high fasting glucose and insulin, high plasma triglycerides (TAG), low HDL and high blood pressure. In addition, low fat, high CHO diets have long been known to raise TAG, lower HDL and, in the absence of weight loss, may worsen glycemic control. Thus, whereas there are numerous strategies for weight loss, a patient with high BMI and high TAG is likely to benefit most from a regimen that reduces CHO intake. Reviewing the literature, benefits of CHO restriction are seen in normal or overweight individuals, in normal patients who meet the criteria for MetS or in patients with frank diabetes. Moreover, in low fat studies that ameliorate LDL and total cholesterol, controls may do better on the symptoms of MetS. On this basis, we feel that MetS is a meaningful, useful phenomenon and may, in fact, be operationally defined as the set of markers that responds to CHO restriction. Insofar as this is an accurate characterization it is likely the result of the effect of dietary CHO on insulin metabolism. Glucose is the major insulin secretagogue and insulin resistance has been tied to the hyperinsulinemic state or the effect of such a state on lipid metabolism. The conclusion is probably not surprising but has not been explicitly stated before. The known effects of CHO-induced hypertriglyceridemia, the HDL-lowering effect of low fat, high CHO interventions and the obvious improvement in glucose and insulin from CHO restriction should have made this evident. In addition, recent studies suggest that a subset of MetS, the ratio of TAG/HDL, is a good marker for insulin resistance and risk of CVD, and this indicator is reliably reduced by CHO restriction and exacerbated by high CHO intake. Inability to make this connection in the past has probably been due to the fact that individual responses have been studied in isolation as well as to the emphasis of traditional therapeutic approaches on low fat rather than low CHO.
We emphasize that MetS is not a disease but a collection of markers. Individual physicians must decide whether high LDL, or other risk factors are more important than the features of MetS in any individual case but if MetS is to be considered it should be recognized that reducing CHO will bring improvement. Response of symptoms to CHO restriction might thus provide a new experimental criterion for MetS in the face of on-going controversy about a useful definition. As a guide to future research, the idea that control of insulin metabolism by CHO intake is, to a first approximation, the underlying mechanism in MetS is a testable hypothesis.
==== Body
Introduction
An association between obesity, diabetes, cardiovascular disease and hypertension has been recognized for some time. Reaven's 1988 Banting lecture is generally considered a turning point in codifying a unifying principle under the name of MetS or Syndrome X (Reviews: [1-7]. Although there is no universally accepted definition or mechanism, (Table 1) a rough common denominator is the set of five features: obesity (high body weight, BMI and/or waist circumference), high glucose and insulin levels, low HDL, high TAG and high blood pressure. Involvement of insulin resistance is generally a common feature and a likely causative agent for at least some of the symptoms. A subset of these metabolic markers, the TAG:HDL ratio, has been proposed as a simple marker for identifying insulin resistance [8]. It has recently been questioned whether the risk attributed to MetS is greater than the sum of the individual symptoms [9] and, ironically, Reaven has taken the "Con" side on a debate on the viability and diagnostic usefulness of the concept [10,11]. Nonetheless, there seems little controversy on the inherent potential for risk in the individual components.
Table 1 NCEP-ATP III and WHO Definitions of Metabolic Syndrome. Other definitions in references [1], [4] and [90].
ATP III Definition
Any three or more of the following criteria:
a) Waist circumference: >102 cm in men, >88 cm in women
b) Serum triglycerides: ≥150 mg/dL
c) HDL-cholesterol: <40 mg/dL in men, < 50 mg/dL in women
d) Blood pressure: ≥130/85 mm Hg
e) Serum glucose: >110 mg/dL
WHO Definition
Diabetes or IFG or IGT or insulin resistance, plus at least two of the following criteria
a) Waist-to-hip ratio: >0.90 in men, >0.85 in women
b) Serum triglycerides: >150 mg/dL or HDL-cholesterol: <35 mg/dL in men and <40 mg/dL in women
c) Blood pressure: >140/90 mmHg
d) Urinary albumin excretion rate > 20 μg/min or albumin/creatinine ratio >30 mg/g
In reading recent reviews of low CHO diets [12,13], we were struck by the fact that the symptoms of MetS are precisely the ones targeted by diets that restrict CHO. This effect is not entirely surprising since it has been known that dietary CHO tends to raise glucose, insulin, and TAG and lower HDL and conversely, replacing CHO with monounsaturated fat or with fat and protein improves glycemic control and dyslipidemia expressed as elevated TAG and lowered HDL. Nevertheless, most formal guidelines and clinical papers have not emphasized CHO restriction as a viable approach to treating MetS or the individual components [14,15] and although several authors have indicated an association between MetS and CHO restriction in passing [5,13,16,17] the explicit connection has not been made.
In this study we have isolated five features that are common to almost all definitions. Waist circumference is probably the currently preferred measure for obesity, but most of the literature provides data on body mass and we have used that measure. We have collected information in the literature supporting the notion that these symptoms are specifically ameliorated by reduction in dietary CHO, and to the extent that they have been directly compared, low CHO strategies appear to have an advantage over low fat diets or simple calorie reduction. We conclude that response to CHO restriction may be an operational definition for MetS and that a likely mechanism is the control of insulin metabolism. Finally, we side with those who maintain that MetS is a real thing in the sense that the concomitant appearance of several symptoms may provide different recommended strategies than the isolated factors. It is important to point out, however, that MetS is not a disease but a complex of markers and practitioners may decide that LDL or other factors are more important for individual patients and in these cases other therapies will be appropriate. On the other hand, reliance on LDL as a prime indicator must be tempered by the importance of LDL particle phenotypes which, in turn, correlate with TAG and HDL levels, a subset of MetS markers.
CHO restriction for weight loss
It is sometimes stated that MetS is caused by obesity [1]. In our view, this is only one of several possible theories and would assume that we know the causes of obesity. It is at least plausible that obesity and the features of MetS arise in parallel from disruptions of insulin metabolism (possibly a consequence of high insulin due to chronic high dietary CHO). Also a high prevalence of so called metabolically obese-normal-weight individuals with MetS has long been known [18]. In any case, it is generally agreed that the first line of attack against MetS or frank diabetes should be reduction in body mass. The method for attaining this weight loss, however, is more controversial. Studies in the literature imply that a low fat diet is a kind of standard although much recent evidence has indicated the value of strategies based on carbohydrate restriction. Whereas low fat diets for calorie reduction can undoubtedly be effective for many people, we feel that they cannot be taken as an established standard; to our knowledge, there has never been a long term study where a low fat diet was instituted in the absence of confounding features such as cessation of smoking and exercise. Also, fat restriction per se does not enhance long-term (one year or longer) weight loss or prevent regain of weight [19] and the record of compliance is modest at best [20]. Most important, the fact that, in the obesity and diabetes epidemic, fat consumption went down (for men, the absolute amount) and carbohydrate consumption went up [21,22], means that other approaches should be considered. Insofar as isocaloric comparisons have been made, low CHO diets do at least as well, and usually better, than low fat diets (see below). Most striking, in ad lib. trials, subjects on low CHO diets show a spontaneous reduction in calories without any dissatisfaction [23-25], a goal that is universally considered desirable but generally recognized as difficult to impose by cognitive admonitions on calorie restriction per se [26]. In general, published data support the idea that low CHO diets are at least as effective as other weight reduction methods. Further, experimental results show an improvement in lipid outcomes (discussed below), no damage to normal kidneys [27], the potentially beneficial rather than deleterious effects of ketone bodies [28-30] and the prevalence of strategies based on low glycemic index [31,32] or reduction of refined CHO or sweets, all approximations of low CHO diets.
Despite increased acceptance of low carbohydrate regimens, it is important to point out that there is a tendency to equate any kind of carbohydrate restriction with the popular Atkins diet [33] and to equate the Atkins diet with a recommendation for high fat and with high saturated fat, in particular. There are, however, many strategies for reducing carbohydrate intake both clinically and in popular diets [34,35]. Whereas high fat is permitted on the Atkins diet and other low carbohydrate diets it is not specifically recommended; as noted above, at least three published studies [23-25] and much anecdotal evidence suggests that, in practice, the major effect is reduction in carbohydrate intake with limited replacement with either fat or protein. In addition, although a deleterious effect of saturated fat, at least in the absence of CHO control, is established [36,37], it has been known since the Seven Countries study [38] that total dietary fat does not correlate with cardiovascular risk and the two effects should not be confused [36,37]. In any case, no particular diet is recommended here; the studies cited include all kinds of interventions, and the underlying rationale is the effect of carbohydrate on insulin. The principle espoused here should be evaluated on this basis.
It seems that a prudent statement of the state of affairs would be that, at this point, dieters have many strategies, none perfect, for weight loss and CHO restriction of some kind is one of them.
CHO restriction and MetS
Our argument in the following exposition is that a patient presenting with a high BMI or large waist line has several options for weight loss. Many factors, including physician experience, ethnic background, personal taste and genetic profile, will determine the first one to be tried. A patient presenting with high BMI and high TAG may have a clear best strategy because of the known benefit of CHO reduction and the accepted deleterious effect of high CHO intake. Data from the literature suggests that a patient with more than two of the symptoms of MetS or the particular combination of high TAG/HDL ratio should clearly try CHO restriction as a first strategy. Conversely, the patient with high BMI and high LDL might sensibly try a low fat strategy first. We provide a summary of cases in which CHO restriction is beneficial in the treatment of MetS or its individual symptoms. The review is meant to be representative rather than comprehensive but we think that the wide variety of cases studied and the range of conditions against a background of accepted effects of carbohydrate on the relevant parameters, provides a strong case for our thesis. In addition, the generally consistent benefit of CHO restriction allows a possible further basis for identifying the common thread in MetS if such truly exists.
CHO restriction improves symptoms of MetS
Table 2 (Table 2) shows the results of single arm studies in which CHO-restricted diets of various compositions and duration were used (for summary of details see, e.g., [39-41]). The regimens include very low CHO ketogenic diets (< 50 g/d) and encompasses subjects that were overweight, presented with symptoms of MetS, or were diabetic. It is clear that CHO restriction is effective in relieving these symptoms. Noteworthy is the recent study of Boden [23] which, while short in duration, carefully measured relevant parameters in patients with diabetes. Patients in this study spontaneously decreased food intake to a substantial degree, were satisfied with the diet, did not show substantial water loss and several were able to reduce or terminate medication.
Table 2 CHANGE
Reference # Subjects Duration CHO (g/d) weight (%) HDL (%) TAG (%) TAG/HDL (%) Glucose (%) Insulin (%)
Rickman et al. 1974 1 Normal Weight Men/Women 3–17 d 7 -4.9 -11.4
LaRosa et al. 1980 2 Obese Men/Women 8 wk 6 -8.3 -5.7 -32.6 -29.9
Phinney et al. 1980 3 Obese Men/Women 6 wk <20 -11.8 -24 -16.3 -57.3
Phinney et al. 1983 4 Normal Weight Men 4 wk <20 0.2 0 -26 -7.7 -23.3
Newbold, 1988 5 Men 3–12 mo 9.6 -34.8 -40.5
Volek et al. 2000 6 Normal Weight Men 8 wk 39 -5.4 10.0 -54.9 -56.9 -3.4 -28.0
Sharman et al. 2002 7 Normal Weight Men 6 wk 46 -2.8 11.5 -33 -39.9 -0.2 -34.2
Meckling et al. 2002 8 Obese Women 8 wk 71 -6.1 4.3 -40.3 -4.1 0.0
Westman et al. 2002 9 Obese Men/Women 6 mo Ad Lib -10.3 19.2 -43.1 -53.6
Dashti et al. 2003 10 Obese Men/Women 12 wk 20 to 30 -13.4 8.3 -50 -53.9 -37.1
Hays et al. 2003 11 Obese Men/Women w/ CVD 6 wk Ad Lib -5.2 -2.9 -39.9 -38.1 -7.4 -30.5
Obese Women PCOS 24 wk Ad Lib -14.3 0.4 -18.5 -18.8 5.7 -49.6
Obese Women Reactive Hypoglycemia 52 wk Ad Lib -19.9 3.4 -13.3 -33.2
Dashti et al. 2004 12 Obese Men/Women 24 wk 40 -14.2 20.4 -60.4 -67.1 -22.6
Boden et al. 2005 13 Obese/diabetic men/women 14 days 21 -1.8 -2 -35 -33.8 -16.0
Improvement is seen in the absence of weight loss
Since it is known that weight loss generally improves MetS, it is important to ask whether beneficial metabolic responses to low CHO diets are dependent on weight loss. The question was specifically addressed by Volek's group [42,43] in normal-weight men and women encouraged to maintain their weight and by Allick and colleagues in patients with diabetes mellitus type 2 [44] (Table 3). The studies in normal weight women [43] and type 2 diabetics [44], in particular, used a cross-over experimental design removing the confounding effect of group differences. These studies were also well-controlled. In the case of Allick, formula was used and, in the studies by Volek, compliance was documented by measuring elevation of serum and urine ketones, thereby eliminating dietary reporting errors as a confounding factor. Improvement in the TAG/HDL ratio ranged from 40 to 55%. In summary, a low CHO regimen clearly improves MetS relative to low fat diets even in the absence of weight loss.
Table 3 CHANGE
Reference # Subjects Duration Diet CHO (g/d) weight (%) HDL (%) TAG (%) TAG/HDL (%) Glucose (%) Insulin (%)
Sharman et al.2002 1 Normal Weight Men 6 wk Ketogenic 46 -2.8 11.5 -33.0 -39.9 -0.2 -34.2
6 wk Low-Fat 271 0.5 0.0 -5.3 -5.3 1.8 13.0
Volek et al. 2003 2 Normal Weight Women 4 wk Ketogenic 43 -2.0 32.0 -30.2 -47.2 -1.9 11.6
4 wk Low-Fat 249 -1.3 -7.7 3.8 0.5 -5.3 18.7
Allick et al. 2004 3 Type 2 Diabetics 2 wk Ketogenic 0 0 23.5 -43.9 -55 -16.9 -16.7
2 wk Low-Fat 775
In addition to studies in which weight maintenance was a feature of experimental design it is important to consider data reported by Foster [45] (Figure 1, (Table 4), who compared low CHO and LF diets. It is widely quoted that the low CHO diet is better at 6 months but that there is no difference in the diets at 12 months. However, it has been pointed out [12] that the particular form of low CHO diet used (Atkins diet) allowed increases in CHO consumption as the trial progressed indicating that it is likely this reintroduction of CHO that predisposes to long-term regain in weight. Most notable in this study, is that the improvement in lipid profile persisted (Figure 1, (Table 4) even after the effect on weight loss disappeared.
Figure 1 Comparison of features of Metabolic Syndrome on low carbohydrate vs. high carbohydrate diets. Data from reference [45].
Table 4 CHANGE
Reference # Subjects Duration Diet CHO (g/d) weight (%) HDL (%) TAG (%) TAG/HDL (%) Glucose (%) Insulin (%) DBP (mm Hg)
Brehm et al. 2003 1 Obese Women 6 mo Low-CHO 41–97 -9.3 13.4 -23.4 -32.4 -9.1 -14.8 -5
LF 163–169 -4.2 8.4 1.6 -6.3 -4.0 -23.0 -1
Sondike et al. 2003 2 Overweight Adolescents 12 wk Low-CHO 37 -10.7 8.7 -40.5 -45.2
LF 154 -4.1 4.2 -5.4 -9.2
Samaha et al. 2003 3 Obese Men/Women 6 mo Low-CHO 150 -4.5 0.0 -20.2 -20.2 -8.6 -27.3
LF 201 -1.4 -2.4 -4 -1.6 -1.6 5.6
Foster et al. 2003 4 Obese Men/Women 1 yr Low-CHO ad lib -7.3 18.2 -28.1 -29.5
LF ad lib -4.5 1.4 0.7 -2.6
Volek et al. 2004 5 Overweight Women 4 wk Low-CHO 29 -3.9 1.3 -23 -28.3 -3.8 -8.8
LF 186 -1.4 -8.6 -11.2 -4.2 1.3 23.2
Sharman et al. 2004 6 Overweight Men 6 wk Low-CHO 36 -5.6 -3.3 -44.1 -42.3 -5.8 -41.5
LF 224 -3.6 -6.6 -15 -8.3 -5.2 -28.1
Brehm et al. 2004 7 Obese Women 4 mo Low-CHO 69 -10.8 16.3 -37.3 -46.1 -9
LF 174 -6.8 4.5 -10.3 -14.2 -3
Meckling et al. 2004 8 Obese Men/Women 10 wk Low-CHO 59 -7.7 12.2 -29.4 -37.1 -8.0 -28.7 -6.1
LF 225 -7.4 -15.4 -25.4 -11.8 -10.2 -3.3 -5
Stern et al. 2004 9 Obese Men/Women 1 yr Low-CHO 120 -3.9 -2.8 -28.6 -26.8
LF 230 -2.3 -12.3 2.7 29.6
Yancy et al. 2004 10 Obese Men/Women 24 wk Low-CHO 30 -12.3 9.8 -47.2 -51.8 -6
LF 198 -6.7 -2.9 -14.4 -12.1 -5.2
Aude et al. 2004 11 Obese Men/Women 12 wk Low-CHO ad lib -6.2 -2.6 -23.2 -21.1
LF ad lib -3.4 -7 -10.5 -3.8
Seshadri et al. 2004 12 Obese Men/Women 6 mo Low-CHO 113 (-8.5) -2.4 -7.4 -40
LF 198 (-3.5 kg) -2.4 -2.3 11.2
McAuley et al. 2004 13 Obese Women 8 wk Low-CHO 41 -6.9 0.9 -38.8 -44.2 -5.9 -39.3
LF 172 -4.4 -6 -17.5 -15.1 -0.1 -28.4
Mod-PRO 130 -5.8 -4.1 -33.9 -31.8 -3.9 -24.4
Dansinger et al. 2004 14 Obese Men/Women 2 mo Low-CHO 103 -4.7 8.8 -27.6 -26.20 -10 -29.5
Mod-PRO 158 -4.6 4.6 -34 -30.50 -9.3 -27.7
LF 183 -4.3 -0.6 -7.1 -5.60 -5.7 -11
UltraLF 230 -4.9 -10.9 -0.6 8.40 -3.5 -7.7
In ad lib. comparisons low CHO diets do better than low fat diets for weight loss and MetS
At this point, we have established that CHO restriction improves MetS and that this can be independent of weight loss. Weight loss can, of course, occur with low fat diets and we next consider the extent to which one or the other strategy is more effective. (Table 4) summarizes results of several studies in the literature demonstrating that low CHO diets generally do better than low fat diets in ad lib. comparisons. Although there is great variability, a pattern of better responses on very low CHO is evident. It is notable that Samaha et al studied a population in which 39 % had diabetes and 43% had MetS [46].
Figure 1 shows data from the study of Foster, et al. [45] and, as noted above, despite the relative similarity in weight loss, the markers of MetS were more favorable in the low CHO arm than the LF arm.
In isocaloric comparisons, low CHO diets do better than LF diets for weight loss
Because weight reduction is considered the first line of attack in MetS or frank diabetes it is worth considering the record of low CHO diets on this parameter alone. It is generally agreed that the major effect of a low CHO diet is a spontaneous reduction in calories. In studies mentioned above, subjects did not significantly increase fat or protein intake but merely removed CHO from their diets [23-25]. Foster and Samaha also attributed the better performance of low CHO arm to decreased caloric intake, although this was not actually measured.
Beyond spontaneous caloric reduction, however, it has been shown that the macronutrient composition of the diet can affect the efficiency of energy utilization and greater efficacy, the so-called metabolic advantage, of low CHO diets compared to LF diets has been the subject of several reports (Reviews: [39,41]. It has long been argued that there must be some mistake because it is physically impossible and would violate the laws of thermodynamics. We have shown this argument is based on misunderstanding of the laws of thermodynamics [39,47-49] and the effect of variable efficiency is now better accepted [50,51]. The precise conditions that allow the so-called metabolic advantage to occur are not known although Cornier, et al. [51] have suggested that those subjects with insulin resistance will show a metabolic advantage on a low CHO diet whereas those who are insulin sensitive do better on low fat. This is consistent with the proposal here, namely that MetS, where insulin-resistance is generally considered a major component, can be defined by the response to CHO restriction. The study of Cornier, et al. [51] had only a small number of subjects and the low CHO arm was not particularly low (40%) but their theory follows from the general rationale of the effect of CHO on energy efficiency. The factors that determine whether a metabolic advantage can play a role in a CHO restricted diet is unknown but given that the insulin resistance association is reasonable, it would seem that some form of CHO restriction is one of the standard, if not preferred attacks on obesity where MetS is suspected.
Figure 2 shows data from Golay, et al. [52] This study is widely quoted as an example of how weight loss is independent of macronutrient composition; although the low CHO arm did better in weight loss, this was judged not significant. This may well be an experiment in which metabolic advantage does not occur – the effect is only possible, not required [39]. It is clear, however, from the figure that there is improvement in TAG and insulin and Golay's conclusion was that "...considering the greater improvement of fasting blood insulin, the glucose/insulin ratio and blood triglyceride, the low carbohydrate diet (25%) could be more favourable in the long-term [52]."
Figure 2 Per cent change in response to diet. Low carbohydrate (dark blue 25 % CHO) and low fat diets(light blue: 45 % CHO). Data from Golay, et al. [52].
Figure 3 Effect of carbohydrate on parameters of Metabolic Syndrome. Comparison of 40% CHO (blue) and 55% (red) CHO diets. Data from [54].
Is this new?
In Edgar Allan Poe's detective story The Purloined Letter, the police search the apartment for a missing blackmail note [53]. In the end, Poe's detective, Auguste Dupin reveals that it had been in plain view on the fireplace all along. The effect of CHO reduction on the symptoms of MetS has, in fact, been visible for some time. In a classic review in 1986, Reaven demonstrated the relative effects of 40% and 55% CHO [54]. Figure 3 shows data from that study: day long glucose, insulin and TAG levels were improved by the low CHO diet. Of interest, is that fasting glucose is not different on the two diets but there is a clear difference in the time course, common to several low CHO interventions (Table 3)). Reaven's experiment suggests that a nutritional approach to MetS is possible by lowering dietary CHO. The experiment should sensibly have spurred research to see whether still lower CHO had further beneficial effect. It took many years, however, before this was done.
Figure 4 Effect of diet on plasma glucose. Mean plasma glucose concentration before (triangles) and after 5 weeks on control diet (yellow circles: (CHO:fat:protein = 55:30:15)) or 5 weeks on lower carbohydrate diet (blue circles: (20:50:30)). Meal points are Breakfast (B), lunch (L) and dinner(D) plus 2 snacks (S1, S2). Data from reference [57].
Substitution of protein for CHO improves MetS
Despite the evidence from Reaven's experiment, a barrier to progress in understanding the role of CHO restriction was the accepted idea that high fat was unhealthy. At the same time, it was thought that an increase in protein would be deleterious for type 2 diabetics because of the increase in glucose due to gluconeogenesis. Nuttall and Gannon have summarized the history of this problem and work in their lab showed that, in fact, glycemic control was enhanced by a diet that was 40 % CHO, with protein replacing part of the carbohydrate [55,56]. Most recently, this group has shown the benefit of a 20 % CHO diet with higher protein [57]. Results in Figure 4 show a pattern similar to Reaven's but much more dramatic. Similar striking differences in the control of insulin and TAG were also demonstrated.
Figure 5 Changes in cholesterol/HDL for substitution for fat. Fat in the average American diet was substituted with the indicated substances at 10 % of energy. Data from reference [62].
Explicit low-fat/high CHO interventions exacerbate MetS
In the approach taken here, we see dietary fat as playing a largely passive role (not withstanding differences between different fats) and the disposition of dietary fat is controlled by insulin and other hormones that, in turn, depend on dietary CHO which we take as the controlling variable. Thus, characterizing a low CHO diet as high fat [58,59] ignores the question about the underlying mechanism; we have recently raised the question of whether "high fat" is a meaningful description in the absence of information about CHO [60]. In the cases we discuss next, the focus is interventions described or designed as low fat. Our point here, however, is that although low fat diets exacerbate the markers of MetS, it is likely the high CHO rather than the fat level that is important.
A very influential paper by Garg, et al. [61] describes a four center randomized study of patients with type 2 diabetes receiving glipizide treatment. The study compared diets where monounsaturated fat was substituted for CHO or vice-versa. Because of its importance, we quote from this article. The rationale for the study, according to the authors is that:
"Compared with diets rich in saturated fats, low-fat, high-carbohydrate diets are reported to reduce serum low-density lipoprotein (LDL) cholesterol levels. Recent studies, however, suggest the high-carbohydrate diets may accentuate hypertriglyceridemia, reduce serum high-density lipoprotein (HDL) cholesterol concentration, and may even worsen hyperglycemia and/or raise plasma insulin levels."
In short, the question is whether high CHO diets worsen MetS. The conclusions of the paper state:
"The study confirms that HC (high carbohydrate diets) increase plasma TAG levels and increase VLDL-C concentrations in NIDDM patients. In this study the HC diet raised fasting plasma triglyceride levels and VLDL-C concentrations by 24 % and 23 % respectively compared with the HMUF diet. Furthermore, daylong levels of plasma triglycerides were also elevated on the high-carbohydrate diet. Consistent with the results of previous studies, plasma levels of total cholesterol and LDL cholesterol were not different on the two diets in this study. The study, therefore substantiates the fact that high carbohydrate diets offer no advantage in lowering LDL levels in NIDDM patients compared with high-fat diets that are low in saturated fats."
The general case: substitution of fat for CHO improves MetS
The substitution of fat for CHO is, in fact, generally beneficial for MetS. In a recent meta-analysis, Mensink, et al. [62] showed the effect of substitution of different fat sources or carbohydrate for the fat in the average US diet at 10% of energy. The conclusion was that substitution of carbohydrate had the most unfavorable response on the total cholesterol to HDL ratio, significantly worse than butter or palm oil (Figure 5).
Summary of review and hypothesis to this point
We have summarized work in the literature showing that low CHO interventions improve the markers of MetS in normal subjects, patients with MetS and diabetics. In comparative studies, they are at least as effective as low fat diets for weight loss and, tend to show better improvement in the other markers of MetS. Isocaloric studies similarly support the idea that the markers of MetS respond preferentially to low CHO diets.
The state of accepted scientific thinking for the world at large is unknown but we have made the case that it is an acknowledged principle that low CHO diets tend to reduce TAG, raise HDL and improve glycemic control whereas LF/high CHO diets tend to have the opposite effect. Perhaps the strongest indication that such an idea is generally accepted is the paper by Rock, et al [63] where the effect of low fat diets in cancer patients was studied. To demonstrate compliance with the low fat recommendations, the authors showed increased TAG and reduced HDL levels. These effects were judged not significant enough to cause a risk for CVD but demonstrate that low fat (higher CHO) diets point in that direction.
We reiterate that this article is not meant to make recommendations – for which many factors must be considered – but rather to show the association between CHO restriction and improvement in symptoms of MetS. Many low fat interventions have successfully reduced LDL, an established risk factor for CVD. If our hypothesis is correct, however, these same interventions should worsen the features of MetS. To some extent this is established from the principles noted above, but we provide an example from a well done experiment in the literature to support this corollary. Finally, in considering the dialectic of treating MetS with CHO restriction vs. high LDL, with low fat diets, it is important to consider both individual variation and the role of LDL particle size. We consider that last.
Studies of low-fat diets
The most salient feature of the obesity epidemic from the standpoint of food consumption is the dramatic increase in CHO intake and the reduction in fat intake (for men, the absolute amount). To our knowledge, the decreased fat intake has not been accompanied by reduction in the incidence of CVD in unmedicated population. These data suggest that low fat diet recommendations per se are not likely to help MetS. If the fundamental idea proposed here is correct, then experimental interventions targeting lower fat that concomitantly raise CHO should, in fact, have a deleterious effect on the markers of MetS. Again, one will have to decide if the symptoms of MetS are more important than LDL or total cholesterol which are typically reduced on LF diets in the unmedicated population.
Delta-1 Study
The Delta-1 study is one of the very well done trials involving a large number of participants [64]. The goal was to determine "the effects of reducing total fat and saturated fat" although this is slightly misleading in that only saturated fat was reduced and any reduction in total fat was a consequence of this. The randomized and balanced diets that were compared all contained approximately 15 % of calories as protein. Other macronutrients were as shown in Table 5. (Table 6) shows the outcomes for all groups on lipids in the Delta-1 study. The results indicate that LDL is significantly reduced although only SEM is given so that it is not possible to know the range of responses of subjects. As a group, however, there were step-wise reductions in LDL and HDL going from the average American diet (AAD) to Step 1 to the low saturated fat diet. There is, as well, a corresponding increase in TAG and hence the TAG/HDL ratio. The authors concluded that the reduction in LDL should be associated with 10% to 20% reductions in cardiovascular disease in the population. Other authors have argued, however, that the effect of the 13% higher HDL seen on the AAD might be associated with a 36% reduction in the risk of death from coronary disease or of myocardial infarction [65]. Again, the purpose here is not to decide on the relative risk attached to different markers but only to point out that the markers for MetS provide another side to the story. An important follow-up in the Delta study to determine HDL subpopulations [66], showed that the more anti-atherogenic HDL2 particles were, in fact, decreased by reductions in saturated fat.
Table 5 Macronutrient composition of diets in the Delta-1 study. Data from reference [64].
Diet CHO (%) total fat (%) SFA (%) MUFA (%) PUFA (%)
Average American diet (AAD) 48 34 15 13 7
Step 1 55 29 9 13 7
Low Saturated Fat 59 25 6 12 7
Table 6 Ave American Diet vs. Step 1 Ave American Diet vs. Low Sat Fat
Ginsberg, et al.,1998 reference [64] AAD Step 1 delta % change pre post delta % change
Total cholesterol 202.1 191 -11.1 -5.5 202.1 183.4 -18.7 -9.3
LDL 131.4 122.2 -9.2 -7.0 131.4 116.9 -14.5 -11.0
TAG (mmol/l) antiln (log) 85.1 92.4 7.3 8.6 85.1 93 7.9 9.3
HDL (mmol/l) 52.2 48.5 -3.7 -7.1 52.2 46.2 -6.0 -11.5
Total/HDL 4.1 4.16 0.1 2.2 4.1 4.21 0.1 3.4
TAG/HDL (arbitrary units) 1.6 1.9 0.3 16.9 1.6 2.0 0.4 23.5
The overall conclusion is that "dietary changes suggested to be prudent for a large segment of the population will primarily affect the concentrations of the most prominent antiatherogenic HDL subpopulation. However, the simultaneous reduction in the atherogenic LDL subpopulation will most likely offset any potential negative effect on cardiovascular risk." As noted above, the decision as to "most likely" outcome must rest with individual patients and physicians.
Role of individual responses and LDL heterogeneity
As noted in Garg's study, changes in LDL may not be as reliable as changes in other markers. Volek, et al. for example showed that whereas TAG was reduced in almost every subject on a low CHO diet, responses in LDL were highly variable [67]. The importance of LDL subpopulations has recently been appreciated and, unlike total LDL, changes in specific LDL particles show a consistent pattern with respect to dietary change.
Greater atherogenic potential is associated with small, dense LDL particles [68]. Krauss and coworkers have carried out impressive work in defining the variability in different individuals. They identified a genetically influenced pattern (B) in people whose plasma contains small LDL particles. This subpopulation, typically 30 % of the American population, responded to low-fat diets by lowering LDL but the pattern B persisted [69]. The remaining subpopulation with larger buoyant particles (pattern A) responded to reduction in fat intake by a shift to the more atherogenic pattern B. Thus, for most of the populations studied, replacing dietary fat with CHO leads to a worsening of the LDL size distribution [70]. In a study described in reference [70], similar effects were seen when protein was substituted for carbohydrate without significant change in the fat content or composition. As summarized by Krauss, "This indicates that carbohydrate rather than fat is a major dietary determinant of expression or phenotype B in susceptible individuals." Although probably a semantic point, "susceptible" is redundant and describing pattern A and B as phenotypes may not be precise: Krauss has summarized how the relative amounts of CHO and fat affect the prevalence of pattern B [71] and the conclusion is that a strong relation exists between CHO intake (ranging from 40 to 75%) and the prevalence of the pattern B phenotype (Fig 6). In other words, there appears to be a continuous variability in phenotype characterized by sensitivity to CHO and everyone may be susceptible to conversion to pattern B at some CHO/fat ratio. The extrapolated line in Figure 6 suggests that a truly low CHO diet might reduce the level of atherogenic subtype to zero. Thus, whereas we have described the dialectic in practical applications as balancing the improvement in MetS with CHO restriction and the improvement in LDL from low-fat diets, focusing on LDL may have some caveats. In general, a growing body of work has shown improvement in LDL pattern switching from high CHO to low CHO diets [42,43,72-75].
Figure 6 Prevalence of pattern B phenotype as a function of the percentage dietary CHO in men. Data from reference [71].
The pattern B phenotype rarely occurs in isolation and, our major concern here is that it is metabolically linked to and co-expressed with other characteristics of MetS, particularly elevated TAG and low HDL. Krauss and colleagues reported that switching from a low CHO/high-fat diet (46% fat) to a high CHO/low-fat diet (26% fat) resulted in lowering of LDL, but also a worsening of TAG and HDL when switching to the low fat diet [69]. In men that were pattern A at a fat intake of 20 to 24%, a further reduction in dietary fat to 10% and CHO to 76% of energy resulted in conversion to B, with continued worsening of TAG and HDL, and no additional LDL-lowering [76].
The TAG/HDL connection
In the search for markers for both insulin resistance and predisposition to CVD, recent research has focused on the value of the ratio of TAG:HDL. McLaughlin, et al. [8] have shown a correlation between insulin resistance as measured by steady-state glucose levels after infusion of glucose, insulin and octreotide (to suppress endogenous insulin secretion). The conclusions of their study were that TAG and HDL were independently related to insulin resistance and the TAG/HDL ratio was the best predictor of insulin resistance. Of importance here is that the results showed that this ratio is comparable to the ATP III criteria for MetS in predicting insulin resistance and "even better in prediction the LDL phenotype B in two separate populations who were on different diets." This is, in fact, only the most recent of several studies (references in [8]) that have shown a correlation between TAG/HDL and insulin resistance and CVD risk as measured by LDL particle size. Table 2-4 as well as Figures 2 and 5 indicate that low CHO diets reliably reduce this marker. Inability to recognize this is again due to the separate conditions in which they are measured and the continued emphasis on reducing dietary fat above all else.
McLaughlin's analysis [8] identified a TAG/HDL ratio of ≥ 3.5 as a cutoff for identifying the insulin-resistant patient most at risk for CVD. It is of interest that in Foster's study [45] described above (Table 3; Figure 1), the average beginning values were 4.6 and 4.3 in the low CHO and low-fat arms, respectively, substantially above this cutoff value. After six months, the low CHO arm had reduced this marker to 3.7 while the LF group showed little change at 4.2. Similarly, in the Delta-1 study, neither the Step 1 diet nor the low fat diet were able to improve the TAG/HDL ratio which was above the threshold value of 3.5 (Table 6)).
Mechanism
A recent review by Ginsberg [77] has provided an excellent description of the possible mechanisms and central role of insulin resistance in mediating the dyslipidemia of MetS. In combination with a proposal by Volek [13] on the mechanism for the reversal of this process, a reasonable understanding of the connection between MetS and CHO restriction is possible.
A primary target of insulin is hormone-sensitive lipase. Adipocyte insulin resistance, plausibly a down regulation of insulin response due to continued stimulation (from higher dietary CHO), leads to increased lipolysis [78,79]. This will lead to greater delivery of fatty acids and an increase in hepatic esterification, and subsequent over production of VLDL, particularly the TAG-rich VLDL1. In combination with impaired plasma TAG clearance, a constant state of hypertriglyceridemia in the postabsorptive and postprandial period occurs. This leads to the exchange of TAG in VLDL for cholesteryl ester in LDL. The resulting TAG-rich LDL particle is a preferred substrate for hepatic lipase and lipoprotein lipase and thereby for generation of small, dense LDL. A similar neutral lipid exchange likely occurs with HDL whereby TAG-rich HDL is hydrolyzed by lipoprotein lipase resulting in the generation of smaller HDL particles that are rapidly removed from the circulation. In this way, elevated TAG resulting from disruption in insulin function, plays a central role in regulating the atherogenic dyslipidemia of MetS.
As noted above, Volek, et al. [13] have reviewed many studies showing that CHO restriction results in significant reductions in postprandial lipemia, and beneficial effects on HDL and intravascular processing of lipoproteins. A key component is what might be called the fatty acid paradox. Whereas insulin resistance is frequently characterized by high fatty acid levels, CHO restriction can improve insulin resistance while raising fatty acids. The latter effect is presumed to be due to lower insulin levels and disinhibition of hormone sensitive lipase. This is accompanied by enhanced cellular fatty acid uptake, mitochondrial transport and increased oxidation. The bias toward fat oxidation over storage reduces hepatic TAG and reduces synthesis and secretion of VLDL.
Discussion
A joint position statement by several organizations recommended "current dietary guidelines from the ADA, AHA and the NCEP-ATP III.... These recommendations may require modification, however, as new information is generated from additional diet intervention studies [80]." Rather than additional studies, however, we provide new information from an evaluation of papers already in the literature that may provide a basis for modification. Data compiled in Figures 2, 3, 4 show that low CHO diets improve the symptoms of MetS as defined by five common criteria. We propose that in addition to potential value as therapy, the response to CHO restriction might be considered an operational definition of MetS. Beyond formal categorizing, the idea is consistent with the generally held belief that MetS is intimately involved with some form of insulin resistance. (The importance of other factors such as inflammation are not mutually exclusive). Cornier et al. showed differential benefit of low CHO vs. low fat diets for people with or without insulin resistance [51]; this study might be thought of as a model for future work to follow this line of thinking. We emphasize that our main point is that an intimate connection between CHO restriction and the complex of symptoms of MetS is seen in the literature of both low CHO studies and low fat (higher CHO) studies. Individual judgment as to the importance of MetS compared to other specific factors or more global assessments such as the Framingham criteria will determine how use is made of this connection. We do believe, however, that ignoring studies on CHO restriction would be unscientific and unproductive.
How low is low carbohydrate?
Goneril. What need you five-and-twenty, ten, or five...?
Regan. What need one?
- William Shakespeare, King Lear.
The data summarized here suggest that some degree of CHO restriction would provide a first line of attack against the symptoms of MetS. The principle of CHO restriction is that by keeping insulin low, metabolism is biased towards lipid oxidation rather than storage, or the effects of fatty acids on peripheral tissues. Most studies that reported deleterious effects of saturated fat have been carried out in the presence of high CHO and there is a real question whether such effects carry over into hypocaloric conditions or those where insulin is better controlled [28,60,81,82].
In general, whereas current thinking in MetS emphasizes the consequences of insulin resistance, we feel that the role of CHO-induced hyperinsulinemia as a causative factor in generating the initial insulin resistance, dyslipidemia or obesity has been under-appreciated. In any case, there are now many ways to implement CHO restriction ranging from ketogenic diets (less than 50 g/d) to diets based on glycemic index, an indirect method of reducing insulin excursions. The question is how low is low? It is clear that 40% CHO is better than 55% for MetS and there has been some reluctance to go lower even though the studies that have done so show continued improvement. Perusal of Table 2-4 suggests that the lower, the better. Insofar as pattern B is associated with insulin resistance and MetS, examination of Figure 4 supports this idea. The barriers to exploring lower CHO diets appears to be continued emphasis on low fat intake although it has been known since Keys's Seven Country study that total fat in the diet does not correlate with cardiovascular risk [38]. At least as indicated in its popular diet book, the No-Fad Diet [83], the American Heart Association has removed its limitation on total fat which should open the door to more flexible diet interventions. Since many studies have shown that there is frequently a spontaneous reduction in total caloric intake in very low CHO diets and that CHO removed is not replaced by fat or protein, very low CHO now appears as a far more prudent choice than judged in the past.
Is this new?
The phenomenon of CHO-induced hypertriglyceridemia is long established [84-88]. In addition, low fat diets are known to reduce, not only LDL, but also HDL levels. For example, the 2004 recommendations of the American Diabetic Association (ADA) state that "Low-saturated fat (i.e., 10% of energy) high carbohydrate diets increase postprandial levels of plasma glucose, insulin, triglycerides and, in some studies, decrease plasma HDL cholesterol when compared in metabolic studies to isocaloric high monounsaturated fat diets." This is our conclusion (our italics). We think the ADA statement could have been more clearly worded: "Substitution of CHO for monounsaturated fat "increase(s) postprandial levels of plasma glucose, insulin..." or could have been more comprehensive: "substitution of CHO or protein for fat increases..." In other words, it has been known for some time that low fat reduces HDL as well as LDL concentrations as described clearly in the Delta-1 study. Again, Rock's study that used the increase in TAG and decrease in HDL as a marker for compliance to a low fat diet supports the idea as an accepted principle.
In combination with the experimentally observed and intuitively obvious reduction in fasting blood glucose and insulin, our proposal for the importance of CHO replacement in the diet hardly seems new. Yet such an idea, to our knowledge has never been made explicit. The primacy of the low fat paradigm in traditional thinking may have played a role in ignoring this obvious correlation. However, we have, in various places, presented or summarized evidence that low CHO diets have a beneficial effect on MetS [13,16] but the tight connection proposed here conceptually eluded us. Having raised the question, however, it is now clear that it is perfectly consistent with established knowledge.
Is MetS useful?
The intricacies of the debate among official agencies on the clinical importance of MetS [9] are beyond the scope of this article. As a first order approximation, however, we are inclined to follow Reaven's strategy in the Point/Counterpoint paper [10,89] for assessing the need for the concept of MetS. He describes a patient with a BMI of 27.8 kg/m2 who would not conform to the WHO definition of MetS because of acceptable values of glucose, TAG and an HDL level of 37 mg/dL and he points out that if the HDL level fell to 33 mg/dL such a patent would fit a new criteria but would not sensibly be treated in a different way. Under the approach considered here, the patient with the higher HDL would be presented with a number of options for weight loss whereas the patient with lower HDL might reasonably be counseled to CHO restriction as a first strategy. We therefore think that MetS or some combination of markers by any other name would have much more virility than the description of MetS as Methuselah in his amusing Counterpoint [87]. We also side with the "Pro" position on the value for basic research. Reaven raises the critical question: "How can there be a common etiology for a diagnostic category based on satisfying 3 of 5 arbitrarily defined criteria when any combination of the 3 will define the same phenotype as any other trio of abnormalities.?" A plausible answer to this question is that if all the markers in MetS are related to hyperinsulinemia and/or insulin resistance, then the relative Km's for insulin for the different target proteins and different tissues are likely to lead to a variable time course for specific individuals and markers are likely to exceed cutoff at different times for each patient. Appearance of one marker may then be indicative of other still silent conditions.
Recommendations for the American population
The data summarized here suggest that there is value in the definition of MetS and that a nutritional strategy based on CHO restriction might sensibly be the "default" diet, the first to be tried, for patients with MetS. In the case of normal weight individuals with MetS, CHO restriction may be the only effective non-pharmacological approach for treating the diversity of symptoms. The choice of any intervention, however, depends on individual assessment of the relative importance of different risk factors and our goal here is the establishment of the close link between CHO restriction and MetS rather than any recommendation.
The recent AHA/NHLBI Scientific Statement on Metabolic Syndrome [90,91] as well as the ATP III emphasize as the primary target, LDL, a marker that is not considered a feature of MetS and that may not even be high in many patients with MetS. Whereas nutritional recommendations are quite general (reduce weight, increase exercise), these reports emphasize a low fat diet (although limiting simple sugars). We think this is inconsistent and we have made the case that a low fat diet, if high in carbohydrate, seems to be widely accepted as raising TAG, lowering HDL and worsening glycemic control, seemingly the wrong thing for MetS. We also disagree with the assertion in the AHA/NHLBI statement that most low CHO diets are high in saturated fat. This statement is undocumented and, as noted in the introduction, essentially equates all reduced carbohydrate approaches and does not seem to distinguish between total and saturated fat. Even if it were true the statement avoids the question of the effect of saturated fat in the presence of low CHO or hypocaloric diets [60,81,82,92]. In practical terms, a recommendation to reduce saturated fat on low CHO diets might be more helpful than blanket prohibition. In addition, the AHA/NHLBI report [90] presents the rationale for low CHO as the effect on appetite. Whereas this may be a component, it has been stated many times, and is part of basic biochemical education [16,93-96], that the rationale of CHO restriction is the control of metabolism by insulin regulation. The effects described above clearly support this. Although we think that, within the framework of MetS, some recommendations can be made, in the end we are probably in agreement with Reaven's judgment that "What is required is less advice and more information [54]."
Questions raised
The data summarized here leave little room for doubt that the generally accepted deterioration in HDL and TAG levels with low fat diets and the established improvement in glycemic control with CHO restriction are part of a more general picture. The hypothesis that response to CHO restriction (because of the effect on insulin) is the defining feature of MetS is the proposed generalization. This idea raises several questions.
Is there a threshold level of CHO restriction that is necessary to elicit improvements in MetS? What is the effect of replacing the calories lost from CHO with protein or with fat, and in what proportion? How does this compare to not replacing them at all?
Does excessive CHO consumption cause MetS in susceptible individuals?
What is the relative risk in addressing MetS with CHO restriction compared to low fat diets for reduction in LDL? That is, what is the relative risk of high LDL vs. the symptoms that comprise MetS?
What is the role of genotype in determining the response to CHO restriction?
Summary
Five symptoms common to most definitions of MetS are those that are reliably improved by CHO restriction. Carbohydrate restriction is one strategy for weight loss but, in addition, improves glycemic control, insulin levels, TAG and HDL levels even in the absence of weight loss. We suggest that response to CHO restriction may, in fact, be an operational definition of MetS. Its underlying basis would rest on the idea that the features of MetS are associated with a disruption in insulin metabolism which is strongly influenced by dietary CHO. The extent to which this definition is useful may depend on its application by individual practitioners. Experimental studies that follow its lead or conversely disprove its fundamental premise should advance our understanding of obesity, diabetes and CVD. Dismissing CHO restriction without evidence, or expressing "concerns" rather than offering data will probably be less productive.
Abbreviations
Adult Treatment Panel (ATP III), American Diabetes Association (ADA), American Heart Association (AHA), Average American diet (AAD), carbohydrate (CHO), low fat (LF), Metabolic Syndrome (MetS), National Cholesterol Education Program (NCEP), triacylglycerol (TAG), World Health Organization (WHO)
Competing interests
The author(s) declare that they have no competing interests.
Note
Table 2 - Effect of carbohydrate restriction on markers for Metabolic Syndrome (See Table 2)
Data shown in bold indicate improvement in marker, plain, worsening. References:1. [97]; 2. [24]; 3. [98]; 4. [99]; 5. [100]; 6. [101]; 7. [42]; 8. [102]; 9. [103]; 10. [104]; 11. [74]; 12. [105]; 13. [23].
Table 3 - Effect of carbohydrate restriction on markers for Metabolic Syndrome under conditions of constant body mass (See Table 3)
Data shown in bold indicate low CHO shows greater improvement in markers for MetS than LF; plain, LF is better. Table reference 3 shows the ratio of low CHO to LF. References: 1. [42]; 2. [43]; 3. [44].
Table 4 - Comparison of low CHO vs. LF diets on markers for Metabolic Syndrome (See Table 4)
Data shown in bold indicate low CHO (or mod-PROT) shows greater improvement in marker than LF; plain, LF is better. Experiment in reference [106] was carried out for a longer time period but diets became very similar. References:1. [107]; 2. [113]; 3. [46]; 4. [45]; 5. [108]; 6. [72]; 7. [109]; 8. [110]; 9. [111]; 10. [112]; 11. [73]; 12. [75]; 13. [59]; 14. [106].
Table 6 - Outcomes of the Delta-1 study (See Table 6)
Data from reference [64]. bold indicates improvement in the parameter from Step 1 or low Saturated Fat diet compared to AAD; plain indicates worsening of parameter compared to AAD.
==== Refs
Eckel RH Grundy SM Zimmet PZ The metabolic syndrome Lancet 2005 365 9468 1415 1428 10.1016/S0140-6736(05)66378-7 15836891
Grundy SM Brewer HBJ Cleeman JI Smith SCJ Lenfant C Definition of metabolic syndrome: Report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition Circulation 2004 109 3 433 438 10.1161/01.CIR.0000111245.75752.C6 14744958
Grundy SM Hansen B Smith SCJ Cleeman JI Kahn RA Clinical management of metabolic syndrome: report of the American Heart Association/National Heart, Lung, and Blood Institute/American Diabetes Association conference on scientific issues related to management Arterioscler Thromb Vasc Biol 2004 24 2 e19 24 10.1161/01.ATV.0000112379.88385.67 14766740
Tonkin A The metabolic syndrome -- a growing problem European Heart Journal Supplements 2004 6 A37 A42
Aude YW Mego P Mehta JL Metabolic syndrome: dietary interventions Curr Opin Cardiol 2004 19 5 473 479 10.1097/01.hco.0000134610.68815.05 15316456
Hanley AJ Wagenknecht LE D'Agostino RBJ Zinman B Haffner SM Identification of subjects with insulin resistance and beta-cell dysfunction using alternative definitions of the metabolic syndrome Diabetes 2003 52 11 2740 2747 14578292
Hunt KJ Resendez RG Williams K Haffner SM Stern MP National Cholesterol Education Program versus World Health Organization metabolic syndrome in relation to all-cause and cardiovascular mortality in the San Antonio Heart Study Circulation 2004 110 10 1251 1257 10.1161/01.CIR.0000140762.04598.F9 15326061
McLaughlin T Reaven G Abbasi F Lamendola C Saad M Waters D Simon J Krauss RM Is there a simple way to identify insulin-resistant individuals at increased risk of cardiovascular disease? Am J Cardiol 2005 96 3 399 404 10.1016/j.amjcard.2005.03.085 16054467
Kahn R Buse J Ferrannini E Stern M The metabolic syndrome: time for a critical appraisal: joint statement from the American Diabetes Association and the European Association for the Study of Diabetes Diabetes Care 2005 28 9 2289 2304 16123508
Grundy SM Point: the metabolic syndrome still lives Clin Chem 2005 51 8 1352 1354 10.1373/clinchem.2005.050989 16040840
Reaven GM The metabolic syndrome: requiescat in pace Clin Chem 2005 51 6 931 938 10.1373/clinchem.2005.048611 15746300
Arora SK McFarlane SI The case for low carbohydrate diets in diabetes management Nutr Metab (Lond) 2005 2 16 10.1186/1743-7075-2-16 16018812
Volek JS Sharman MJ Forsythe CE Modification of lipoproteins by very low-carbohydrate diets J Nutr 2005 135 6 1339 1342 15930434
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) final report Circulation 2002 106 25 3143 3421 12485966
Ashen MD Blumenthal RS Clinical practice. Low HDL cholesterol levels N Engl J Med 2005 353 12 1252 1260 10.1056/NEJMcp044370 16177251
Feinman RD Makowske M Metabolic Syndrome and Low-Carbohydrate Ketogenic Diets in the Medical School Biochemistry Curriculum. Metabolic Syndrome and Related Disorders 2003 1 189 198 10.1089/154041903322716660 18370662
Westman EC Yancy Jr. WS Haub MD Volek JS Insulin Resistance from a Low-Carbohydrate, High Fat Diet Perspective Metabolic Syndrome and Related Disorders 2005 3 3 7 10.1089/met.2005.3.14 18370703
Ruderman N Chisholm D Pi-Sunyer X Schneider S The metabolically obese, normal-weight individual revisited Diabetes 1998 47 5 699 713 9588440
Willet W Increasing prevalence of overweight among US adults. The National Health and Nutrition Examination Surveys, 1960 to 1991 JAMA 1994 272 205 211 10.1001/jama.272.3.205 8022039
Poppitt SD Keogh GF Prentice AM Williams DE Sonnemans HM Valk EE Robinson E Wareham NJ Long-term effects of ad libitum low-fat, high-carbohydrate diets on body weight and serum lipids in overweight subjects with metabolic syndrome Am J Clin Nutr 2002 75 1 11 20 11756055
Enns CK Goldman JD Cook A Trends in Food and Nutrient Intakes by Adults: NFCS 1977-78, CSFII 1989-91, and CSFII 1994-95 Family Economics and Nutrition Review 1997 10 2 15
Kennedy ET Bowman SA Powell R Dietary-fat intake in the US population J Am Coll Nutr 1999 18 3 207 212 10376775
Boden G Sargrad K Homko C Mozzoli M Stein TP Effect of a low-carbohydrate diet on appetite, blood glucose levels, and insulin resistance in obese patients with type 2 diabetes Ann Intern Med 2005 142 6 403 411 15767618
Larosa JC Fry AG Muesing R Rosing DR Effects of high-protein, low-carbohydrate dieting on plasma lipoproteins and body weight J Am Diet Assoc 1980 77 3 264 270 7410754
Miller BV Bertino J Reed TG Burrington C Davidson LK Green A Gartung A Nafziger A An Evaluation of the Atkins' Diet Metabolic Syndrome and Related Disorders 2003 1 299 309 10.1089/1540419031361426 18370655
Bray GA The epidemic of obesity and changes in food intake: the Fluoride Hypothesis Physiol Behav 2004 82 1 115 121 10.1016/j.physbeh.2004.04.033 15234599
Martin WF Armstrong LE Rodriguez NR Dietary Protein Intake and Renal Function Nutr Metab (Lond) 2005 2 1 25 10.1186/1743-7075-2-25 16174292
Van der Auwera I Wera S Van Leuven F Henderson ST A ketogenic diet reduces amyloid beta 40 and 42 in a mouse model of Alzheimer's disease Nutr Metab (Lond) 2005
Veech RL The therapeutic implications of ketone bodies: the effects of ketone bodies in pathological conditions: ketosis, ketogenic diet, redox states, insulin resistance, and mitochondrial metabolism Prostaglandins Leukot Essent Fatty Acids 2004 70 3 309 319 10.1016/j.plefa.2003.09.007 14769489
Veech RL Chance B Kashiwaya Y Lardy HA Cahill GFJ Ketone bodies, potential therapeutic uses IUBMB Life 2001 51 4 241 247 11569918
Brand-Miller JC Glycemic index in relation to coronary disease Asia Pac J Clin Nutr 2004 13 Suppl S3 15294465
Brand-Miller JC Postprandial glycemia, glycemic index, and the prevention of type 2 diabetes Am J Clin Nutr 2004 80 2 243 244 15277141
Atkins RC Dr. Atkins' New Diet Revolution 2002 New York , Avon Books
Agatston A The South Beach Diet 2003 New York , Random House
Eades MR Eades MD Protein Power 1996 New York , Bantam Books
Hu FB Manson JE Willett WC Types of dietary fat and risk of coronary heart disease: a critical review J Am Coll Nutr 2001 20 1 5 19 11293467
Hu FB Stampfer MJ Manson JE Rimm E Colditz GA Rosner BA Hennekens CH Willett WC Dietary fat intake and the risk of coronary heart disease in women N Engl J Med 1997 337 21 1491 1499 10.1056/NEJM199711203372102 9366580
Keys A Coronary heart disease in seven countries 1970 41 (Suppl) 1 211
Feinman RD Fine EJ Thermodynamics and Metabolic Advantage of Weight Loss Diets. Metabolic Syndrome and Related Disorders 2003 1 209 219 10.1089/154041903322716688 18370664
Volek JS Westman EC Very-low-carbohydrate weight-loss diets revisited Cleve Clin J Med 2002 69 11 849, 853, 856 8 passim 12430970
Westman EC Mavropoulos J Yancy WS Volek JS A Review of Low-carbohydrate Ketogenic Diets Curr Atheroscler Rep 2003 5 6 476 483 14525681
Sharman MJ Kraemer WJ Love DM Avery NG Gomez AL Scheett TP Volek JS A ketogenic diet favorably affects serum biomarkers for cardiovascular disease in normal-weight men J Nutr 2002 132 7 1879 1885 12097663
Volek JS Sharman MJ Gomez AL Scheett TP Kraemer WJ An Isoenergetic Very Low-Carbohydrate Diet Is Associated With Improved Serum High-Density Lipoprotein Cholesterol (HDL-C), Total Cholesterol to HDL-C Ratio, Triacylglycerols, and Postprandial Lipemic Responses Compared to a Low-Fat Diet in Normal Weight, Normolipidemic Women J Nutr 2003 133 9 2756 2761 12949361
Allick G Bisschop PH Ackermans MT Endert E Meijer AJ Kuipers F Sauerwein HP Romijn JA A low-carbohydrate/high-fat diet improves glucoregulation in type 2 diabetes mellitus by reducing postabsorptive glycogenolysis J Clin Endocrinol Metab 2004 89 12 6193 6197 10.1210/jc.2004-1041 15579777
Foster GD Wyatt HR Hill JO McGuckin BG Brill C Mohammed BS Szapary PO Rader DJ Edman JS Klein S A randomized trial of a low-carbohydrate diet for obesity N Engl J Med 2003 348 21 2082 2090 10.1056/NEJMoa022207 12761365
Samaha FF Iqbal N Seshadri P Chicano KL Daily DA McGrory J Williams T Williams M Gracely EJ Stern L A low-carbohydrate as compared with a low-fat diet in severe obesity N Engl J Med 2003 348 21 2074 2081 10.1056/NEJMoa022637 12761364
Feinman RD Fine EJ "A Calorie is a calorie" violates the second law of thermodynamics. Nutrition Journal 2004 3 9
Feinman RD Fine EJ Whatever happened to the second law of thermodynamics? Am J Clin Nutr 2004 80 5 1445 6; author reply 1446 15531699
Fine EJ Feinman RD Thermodynamics of weight loss diets Nutr Metab (Lond) 2004 1 1 15 10.1186/1743-7075-1-15 15588283
Bray GA Champagne CM Beyond energy balance: there is more to obesity than kilocalories J Am Diet Assoc 2005 105 5 Suppl 1 S17 23 10.1016/j.jada.2005.02.018 15867891
Cornier MA Donahoo WT Pereira R Gurevich I Westergren R Enerback S Eckel PJ Goalstone ML Hill JO Eckel RH Draznin B Insulin sensitivity determines the effectiveness of dietary macronutrient composition on weight loss in obese women Obes Res 2005 13 4 703 709 15897479
Golay A Eigenheer C Morel Y Kujawski P Lehmann T de Tonnac N Weight-loss with low or high carbohydrate diet? Int J Obes Relat Metab Disord 1996 20 12 1067 1072 8968851
Poe EA The Purloined Letter http://xroadsvirginiaedu/~HYPER/POE/purloinehtml 1845
Reaven GM Effect of dietary carbohydrate on the metabolism of patients with non-insulin dependent diabetes mellitus Nutr Rev 1986 44 2 65 73 3703391
Gannon MC Nuttall FQ Saeed A Jordan K Hoover H An increase in dietary protein improves the blood glucose response in persons with type 2 diabetes Am J Clin Nutr 2003 78 4 734 741 14522731
Nuttall FQ Gannon MC Metabolic response of people with type 2 diabetes to a high protein diet Nutr Metab (Lond) 2004 1 1 6 10.1186/1743-7075-1-6 15507157
Gannon MC Nuttall FQ Effect of a high-protein, low-carbohydrate diet on blood glucose control in people with type 2 diabetes Diabetes 2004 53 9 2375 2382 15331548
Ludwig DS Jenkins DJ Carbohydrates and the postprandial state: have our cake and eat it too? Am J Clin Nutr 2004 80 4 797 798 15447882
McAuley KA Hopkins CM Smith KJ McLay RT Williams SM Taylor RW Mann JI Comparison of high-fat and high-protein diets with a high-carbohydrate diet in insulin-resistant obese women Diabetologia 2005 48 1 8 16 10.1007/s00125-004-1603-4 15616799
Feinman R When is a high fat diet not a high fat diet? Nutrition & Metabolism 2005 2 27
Garg A Bantle JP Henry RR Coulston AM Griver KA Raatz SK Brinkley L Chen YD Grundy SM Huet BA Effects of varying carbohydrate content of diet in patients with non-insulin-dependent diabetes mellitus Jama 1994 271 18 1421 1428 10.1001/jama.271.18.1421 7848401
Mensink RP Zock PL Kester AD Katan MB Effects of dietary fatty acids and carbohydrates on the ratio of serum total to HDL cholesterol and on serum lipids and apolipoproteins: a meta-analysis of 60 controlled trials Am J Clin Nutr 2003 77 5 1146 1155 12716665
Rock CL Flatt SW Thomson CA Stefanick ML Newman VA Jones L Natarajan L Pierce JP Chang RJ Witztum JL Plasma triacylglycerol and HDL cholesterol concentrations confirm self-reported changes in carbohydrate and fat intakes in women in a diet intervention trial J Nutr 2004 134 2 342 347 14747670
Ginsberg HN Kris-Etherton P Dennis B Elmer PJ Ershow A Lefevre M Pearson T Roheim P Ramakrishnan R Reed R Stewart K Stewart P Phillips K Anderson N Effects of reducing dietary saturated fatty acids on plasma lipids and lipoproteins in healthy subjects: the DELTA Study, protocol 1 Arterioscler Thromb Vasc Biol 1998 18 3 441 449 9514413
Gordon DJ Knoke J Probstfield JL Superko R Tyroler HA High-density lipoprotein cholesterol and coronary heart disease in hypercholesterolemic men: the Lipid Research Clinics Coronary Primary Prevention Trial Circulation 1986 74 1217 1225 3536151
Berglund L Oliver EH Fontanez N Holleran S Matthews K Roheim PS Ginsberg HN Ramakrishnan R Lefevre M HDL-subpopulation patterns in response to reductions in dietary total and saturated fat intakes in healthy subjects Am J Clin Nutr 1999 70 6 992 1000 10584043
Volek JS Sharman MJ Cardiovascular and hormonal aspects of very-low-carbohydrate ketogenic diets Obes Res 2004 12 Suppl 2 115S 23S 15601959
Lamarche B Tchernof A Moorjani S Cantin B Dagenais GR Lupien PJ Despres JP Small, dense low-density lipoprotein particles as a predictor of the risk of ischemic heart disease in men. Prospective results from the Quebec Cardiovascular Study Circulation 1997 95 1 69 75 8994419
Dreon DM Fernstrom HA Miller B Krauss RM Low-density lipoprotein subclass patterns and lipoprotein response to a reduced-fat diet in men Faseb J 1994 8 1 121 126 8299884
Krauss RM Dietary and Genetic Probes of Atherogenic Dyslipidemia Arterioscler Thromb Vasc Biol 2005
Krauss RM Atherogenic lipoprotein phenotype and diet-gene interactions J Nutr 2001 131 2 340S 3S 11160558
Sharman MJ Gomez AL Kraemer WJ Volek JS Very low-carbohydrate and low-fat diets affect fasting lipids and postprandial lipemia differently in overweight men J Nutr 2004 134 4 880 885 15051841
Aude YW Agatston AS Lopez-Jimenez F Lieberman EH Marie A Hansen M Rojas G Lamas GA Hennekens CH The national cholesterol education program diet vs a diet lower in carbohydrates and higher in protein and monounsaturated fat: a randomized trial Arch Intern Med 2004 164 19 2141 2146 10.1001/archinte.164.19.2141 15505128
Hays JH DiSabatino A Gorman RT Vincent S Stillabower ME Effect of a high saturated fat and no-starch diet on serum lipid subfractions in patients with documented atherosclerotic cardiovascular disease Mayo Clin Proc 2003 78 11 1331 1336 14601690
Seshadri P Iqbal N Stern L Williams M Chicano KL Daily DA McGrory J Gracely EJ Rader DJ Samaha FF A randomized study comparing the effects of a low-carbohydrate diet and a conventional diet on lipoprotein subfractions and C-reactive protein levels in patients with severe obesity Am J Med 2004 117 6 398 405 10.1016/j.amjmed.2004.04.009 15380496
Dreon DM Fernstrom HA Williams PT Krauss RM A very low-fat diet is not associated with improved lipoprotein profiles in men with a predominance of large, low-density lipoproteins Am J Clin Nutr 1999 69 3 411 418 10075324
Ginsberg HN Zhang YL Hernandez-Ono A Regulation of plasma triglycerides in insulin resistance and diabetes Arch Med Res 2005 36 3 232 240 10.1016/j.arcmed.2005.01.005 15925013
Boden G Shulman GI Free fatty acids in obesity and type 2 diabetes: defining their role in the development of insulin resistance and beta-cell dysfunction Eur J Clin Invest 2002 32 Suppl 3 14 23 10.1046/j.1365-2362.32.s3.3.x 12028371
Zammit VA Insulin stimulation of hepatic triacylglycerol secretion in the insulin-replete state: implications for the etiology of peripheral insulin resistance Ann N Y Acad Sci 2002 967 52 65 12079835
Klein S Burke LE Bray GA Blair S Allison DB Pi-Sunyer X Hong Y Eckel RH Clinical implications of obesity with specific focus on cardiovascular disease: a statement for professionals from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism: endorsed by the American College of Cardiology Foundation Circulation 2004 110 18 2952 2967 10.1161/01.CIR.0000145546.97738.1E 15509809
Mozaffarian D Rimm EB Herrington DM Dietary fats, carbohydrate, and progression of coronary atherosclerosis in postmenopausal women Am J Clin Nutr 2004 80 5 1175 1184 15531663
Volek JS Forsythe CE The case for not restricting saturated fat on a low carbohydrate diet Nutr Metab (Lond) 2005 2 21 10.1186/1743-7075-2-21 16135250
American Heart Association No-Fad Diet. A Personal Plan fo Healthy Weight Loss 2005 New York , Clarkson Potter
Hellerstein MK Carbohydrate-induced hypertriglyceridemia: modifying factors and implications for cardiovascular risk Curr Opin Lipidol 2002 13 1 33 40 10.1097/00041433-200202000-00006 11790961
Hudgins LC Effect of high-carbohydrate feeding on triglyceride and saturated fatty acid synthesis Proc Soc Exp Biol Med 2000 225 3 178 183 10.1046/j.1525-1373.2000.22521.x 11082210
Hudgins LC Hellerstein M Seidman C Neese R Diakun J Hirsch J Human fatty acid synthesis is stimulated by a eucaloric low fat, high carbohydrate diet J Clin Invest 1996 97 9 2081 2091 8621798
Parks EJ Hellerstein MK Carbohydrate-induced hypertriacylglycerolemia: historical perspective and review of biological mechanisms Am J Clin Nutr 2000 71 2 412 433 10648253
Parks EJ Krauss RM Christiansen MP Neese RA Hellerstein MK Effects of a low-fat, high-carbohydrate diet on VLDL-triglyceride assembly, production, and clearance J Clin Invest 1999 104 8 1087 1096 10525047
Reaven G Counterpoint: just being alive is not good enough Clin Chem 2005 51 8 1354 1357 10.1373/clinchem.2005.053587 16040841
Grundy SM Cleeman JI Daniels SR Donato KA Eckel RH Franklin BA Gordon DJ Krauss RM Savage PJ Smith SCJ Spertus JA Costa F Diagnosis and Management of the Metabolic Syndrome. An American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement Circulation 2005
Grundy SM Cleeman JI Daniels SR Donato KA Eckel RH Franklin BA Gordon DJ Krauss RM Savage PJ Smith SCJ Spertus JA Costa F Diagnosis and Management of the Metabolic Syndrome. An American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Executive Summary Circulation 2005
Knopp RH Retzlaff BM Saturated fat prevents coronary artery disease? An American paradox Am J Clin Nutr 2004 80 5 1102 1103 15531654
Pogozelski W Arpaia N Priore S The Metabolic Effects of Low-carbohydrate Diets and Incorporation into a Biochemistry Course Biochemistry and Molecular Biology Education 2005 33 91 100 21638552
Harris RA Crabb DW Devlin TM Chapter 22. Metabolic Interrelationships Textbook of Biochemistry With Clinical Correlations 2006 Sixth New York , John Wiley & Sons, Inc.
Makowske M Feinman RD Nutrition education: a questionnaire for assessment and teaching Nutr J 2005 4 1 2 10.1186/1475-2891-4-2 15649324
Smith C Marks AD Lieberman M Basic Medical Biochemistry: A Clinical Approach 2005 2nd Philadelphia , Lippincott Williams & Wilkins
Rickman F Mitchell N Dingman J Dalen JE Changes in serum cholesterol during the Stillman diet Jama 1974 228 1 54 58 10.1001/jama.228.1.54 4406145
Phinney SD Horton ES Sims EA Hanson JS Danforth EJ LaGrange BM Capacity for moderate exercise in obese subjects after adaptation to a hypocaloric, ketogenic diet J Clin Invest 1980 66 5 1152 1161 7000826
Phinney SD Bistrian BR Evans WJ Gervino E Blackburn GL The human metabolic response to chronic ketosis without caloric restriction: preservation of submaximal exercise capability with reduced carbohydrate oxidation Metabolism 1983 32 8 769 776 10.1016/0026-0495(83)90106-3 6865776
Newbold HL Reducing the serum cholesterol level with a diet high in animal fat South Med J 1988 81 1 61 63 3336803
Volek JS Gomez AL Kraemer WJ Fasting lipoprotein and postprandial triacylglycerol responses to a low-carbohydrate diet supplemented with n-3 fatty acids J Am Coll Nutr 2000 19 3 383 391 10872901
Meckling KA Gauthier M Grubb R Sanford J Effects of a hypocaloric, low-carbohydrate diet on weight loss, blood lipids, blood pressure, glucose tolerance, and body composition in free-living overweight women Can J Physiol Pharmacol 2002 80 11 1095 1105 10.1139/y02-140 12489929
Westman EC Yancy WS Edman JS Tomlin KF Perkins CE Effect of 6-month adherence to a very low carbohydrate diet program Am J Med 2002 113 1 30 36 10.1016/S0002-9343(02)01129-4 12106620
Dashti HM Bo-Abbas YY Asfar SK Mathew TC Hussein T Behbahani A Khoursheed MA Al-Sayer HM Al-Zaid NS Ketogenic diet modifies the risk factors of heart disease in obese patients Nutrition 2003 19 10 901 902 10.1016/S0899-9007(03)00161-8 14559328
Dashti HM Mathew TC Hussein T Asfar SK Behbahani A Khoursheed MA Al-Sayer HM Bo-Abbas YY Al-Zaid NS Long-term effects of a ketogenic diet in obese patients Exp Clin Cardiol 2004 9 3 200 205 19641727
Dansinger ML Gleason JA Griffith JL Selker HP Schaefer EJ Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss and heart disease risk reduction: a randomized trial Jama 2005 293 1 43 53 10.1001/jama.293.1.43 15632335
Brehm BJ Seeley RJ Daniels SR D'Alessio DA A randomized trial comparing a very low carbohydrate diet and a calorie-restricted low fat diet on body weight and cardiovascular risk factors in healthy women J Clin Endocrinol Metab 2003 88 4 1617 1623 10.1210/jc.2002-021480 12679447
Volek JS Sharman MJ Gomez AL Judelson DA Rubin MR Watson G Sokmen B Silvestre R French DN Kraemer WJ Comparison of energy-restricted very low-carbohydrate and low-fat diets on weight loss and body composition in overweight men and women Nutr Metab (Lond) 2004 1 1 13 10.1186/1743-7075-1-13 15533250
Brehm BJ Spang SE Lattin BL Seeley RJ Daniels SR D'Alessio DA The role of energy expenditure in the differential weight loss in obese women on low-fat and low-carbohydrate diets J Clin Endocrinol Metab 2004
Meckling KA O'Sullivan C Saari D Comparison of a low-fat diet to a low-carbohydrate diet on weight loss, body composition, and risk factors for diabetes and cardiovascular disease in free-living, overweight men and women J Clin Endocrinol Metab 2004 89 6 2717 2723 10.1210/jc.2003-031606 15181047
Stern L Iqbal N Seshadri P Chicano KL Daily DA McGrory J Williams M Gracely EJ Samaha FF The effects of low-carbohydrate versus conventional weight loss diets in severely obese adults: one-year follow-up of a randomized trial Ann Intern Med 2004 140 10 778 785 15148064
Yancy WSJ Olsen MK Guyton JR Bakst RP Westman EC A low-carbohydrate, ketogenic diet versus a low-fat diet to treat obesity and hyperlipidemia: a randomized, controlled trial Ann Intern Med 2004 140 10 769 777 15148063
|
16288655
|
PMC1323303
|
CC BY
|
2021-01-04 22:25:39
|
no
|
Nutr Metab (Lond). 2005 Nov 16; 2:31
|
utf-8
|
Nutr Metab (Lond)
| 2,005 |
10.1186/1743-7075-2-31
|
oa_comm
|
==== Front
Nutr Metab (Lond)Nutr Metab (Lond)Nutrition & Metabolism1743-7075BioMed Central 1743-7075-2-311628865510.1186/1743-7075-2-31ReviewCarbohydrate restriction improves the features of Metabolic Syndrome. Metabolic Syndrome may be defined by the response to carbohydrate restriction Volek Jeff S [email protected] Richard D [email protected] Human Performance Laboratory, Department of Kinesiology, University of Connecticut, Storrs, CT 06269-1110 USA2 Department of Biochemistry, SUNY Downstate Medical Center, Brooklyn, NY 11203 USA2005 16 11 2005 2 31 31 25 10 2005 16 11 2005 Copyright ©2005 Volek and Feinman; licensee BioMed Central Ltd.2005Volek and Feinman; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Metabolic Syndrome (MetS) represents a constellation of markers that indicates a predisposition to diabetes, cardiovascular disease and other pathologic states. The definition and treatment are a matter of current debate and there is not general agreement on a precise definition or, to some extent, whether the designation provides more information than the individual components. We consider here five indicators that are central to most definitions and we provide evidence from the literature that these are precisely the symptoms that respond to reduction in dietary carbohydrate (CHO). Carbohydrate restriction is one of several strategies for reducing body mass but even in the absence of weight loss or in comparison with low fat alternatives, CHO restriction is effective at ameliorating high fasting glucose and insulin, high plasma triglycerides (TAG), low HDL and high blood pressure. In addition, low fat, high CHO diets have long been known to raise TAG, lower HDL and, in the absence of weight loss, may worsen glycemic control. Thus, whereas there are numerous strategies for weight loss, a patient with high BMI and high TAG is likely to benefit most from a regimen that reduces CHO intake. Reviewing the literature, benefits of CHO restriction are seen in normal or overweight individuals, in normal patients who meet the criteria for MetS or in patients with frank diabetes. Moreover, in low fat studies that ameliorate LDL and total cholesterol, controls may do better on the symptoms of MetS. On this basis, we feel that MetS is a meaningful, useful phenomenon and may, in fact, be operationally defined as the set of markers that responds to CHO restriction. Insofar as this is an accurate characterization it is likely the result of the effect of dietary CHO on insulin metabolism. Glucose is the major insulin secretagogue and insulin resistance has been tied to the hyperinsulinemic state or the effect of such a state on lipid metabolism. The conclusion is probably not surprising but has not been explicitly stated before. The known effects of CHO-induced hypertriglyceridemia, the HDL-lowering effect of low fat, high CHO interventions and the obvious improvement in glucose and insulin from CHO restriction should have made this evident. In addition, recent studies suggest that a subset of MetS, the ratio of TAG/HDL, is a good marker for insulin resistance and risk of CVD, and this indicator is reliably reduced by CHO restriction and exacerbated by high CHO intake. Inability to make this connection in the past has probably been due to the fact that individual responses have been studied in isolation as well as to the emphasis of traditional therapeutic approaches on low fat rather than low CHO.
We emphasize that MetS is not a disease but a collection of markers. Individual physicians must decide whether high LDL, or other risk factors are more important than the features of MetS in any individual case but if MetS is to be considered it should be recognized that reducing CHO will bring improvement. Response of symptoms to CHO restriction might thus provide a new experimental criterion for MetS in the face of on-going controversy about a useful definition. As a guide to future research, the idea that control of insulin metabolism by CHO intake is, to a first approximation, the underlying mechanism in MetS is a testable hypothesis.
==== Body
Introduction
An association between obesity, diabetes, cardiovascular disease and hypertension has been recognized for some time. Reaven's 1988 Banting lecture is generally considered a turning point in codifying a unifying principle under the name of MetS or Syndrome X (Reviews: [1-7]. Although there is no universally accepted definition or mechanism, (Table 1) a rough common denominator is the set of five features: obesity (high body weight, BMI and/or waist circumference), high glucose and insulin levels, low HDL, high TAG and high blood pressure. Involvement of insulin resistance is generally a common feature and a likely causative agent for at least some of the symptoms. A subset of these metabolic markers, the TAG:HDL ratio, has been proposed as a simple marker for identifying insulin resistance [8]. It has recently been questioned whether the risk attributed to MetS is greater than the sum of the individual symptoms [9] and, ironically, Reaven has taken the "Con" side on a debate on the viability and diagnostic usefulness of the concept [10,11]. Nonetheless, there seems little controversy on the inherent potential for risk in the individual components.
Table 1 NCEP-ATP III and WHO Definitions of Metabolic Syndrome. Other definitions in references [1], [4] and [90].
ATP III Definition
Any three or more of the following criteria:
a) Waist circumference: >102 cm in men, >88 cm in women
b) Serum triglycerides: ≥150 mg/dL
c) HDL-cholesterol: <40 mg/dL in men, < 50 mg/dL in women
d) Blood pressure: ≥130/85 mm Hg
e) Serum glucose: >110 mg/dL
WHO Definition
Diabetes or IFG or IGT or insulin resistance, plus at least two of the following criteria
a) Waist-to-hip ratio: >0.90 in men, >0.85 in women
b) Serum triglycerides: >150 mg/dL or HDL-cholesterol: <35 mg/dL in men and <40 mg/dL in women
c) Blood pressure: >140/90 mmHg
d) Urinary albumin excretion rate > 20 μg/min or albumin/creatinine ratio >30 mg/g
In reading recent reviews of low CHO diets [12,13], we were struck by the fact that the symptoms of MetS are precisely the ones targeted by diets that restrict CHO. This effect is not entirely surprising since it has been known that dietary CHO tends to raise glucose, insulin, and TAG and lower HDL and conversely, replacing CHO with monounsaturated fat or with fat and protein improves glycemic control and dyslipidemia expressed as elevated TAG and lowered HDL. Nevertheless, most formal guidelines and clinical papers have not emphasized CHO restriction as a viable approach to treating MetS or the individual components [14,15] and although several authors have indicated an association between MetS and CHO restriction in passing [5,13,16,17] the explicit connection has not been made.
In this study we have isolated five features that are common to almost all definitions. Waist circumference is probably the currently preferred measure for obesity, but most of the literature provides data on body mass and we have used that measure. We have collected information in the literature supporting the notion that these symptoms are specifically ameliorated by reduction in dietary CHO, and to the extent that they have been directly compared, low CHO strategies appear to have an advantage over low fat diets or simple calorie reduction. We conclude that response to CHO restriction may be an operational definition for MetS and that a likely mechanism is the control of insulin metabolism. Finally, we side with those who maintain that MetS is a real thing in the sense that the concomitant appearance of several symptoms may provide different recommended strategies than the isolated factors. It is important to point out, however, that MetS is not a disease but a complex of markers and practitioners may decide that LDL or other factors are more important for individual patients and in these cases other therapies will be appropriate. On the other hand, reliance on LDL as a prime indicator must be tempered by the importance of LDL particle phenotypes which, in turn, correlate with TAG and HDL levels, a subset of MetS markers.
CHO restriction for weight loss
It is sometimes stated that MetS is caused by obesity [1]. In our view, this is only one of several possible theories and would assume that we know the causes of obesity. It is at least plausible that obesity and the features of MetS arise in parallel from disruptions of insulin metabolism (possibly a consequence of high insulin due to chronic high dietary CHO). Also a high prevalence of so called metabolically obese-normal-weight individuals with MetS has long been known [18]. In any case, it is generally agreed that the first line of attack against MetS or frank diabetes should be reduction in body mass. The method for attaining this weight loss, however, is more controversial. Studies in the literature imply that a low fat diet is a kind of standard although much recent evidence has indicated the value of strategies based on carbohydrate restriction. Whereas low fat diets for calorie reduction can undoubtedly be effective for many people, we feel that they cannot be taken as an established standard; to our knowledge, there has never been a long term study where a low fat diet was instituted in the absence of confounding features such as cessation of smoking and exercise. Also, fat restriction per se does not enhance long-term (one year or longer) weight loss or prevent regain of weight [19] and the record of compliance is modest at best [20]. Most important, the fact that, in the obesity and diabetes epidemic, fat consumption went down (for men, the absolute amount) and carbohydrate consumption went up [21,22], means that other approaches should be considered. Insofar as isocaloric comparisons have been made, low CHO diets do at least as well, and usually better, than low fat diets (see below). Most striking, in ad lib. trials, subjects on low CHO diets show a spontaneous reduction in calories without any dissatisfaction [23-25], a goal that is universally considered desirable but generally recognized as difficult to impose by cognitive admonitions on calorie restriction per se [26]. In general, published data support the idea that low CHO diets are at least as effective as other weight reduction methods. Further, experimental results show an improvement in lipid outcomes (discussed below), no damage to normal kidneys [27], the potentially beneficial rather than deleterious effects of ketone bodies [28-30] and the prevalence of strategies based on low glycemic index [31,32] or reduction of refined CHO or sweets, all approximations of low CHO diets.
Despite increased acceptance of low carbohydrate regimens, it is important to point out that there is a tendency to equate any kind of carbohydrate restriction with the popular Atkins diet [33] and to equate the Atkins diet with a recommendation for high fat and with high saturated fat, in particular. There are, however, many strategies for reducing carbohydrate intake both clinically and in popular diets [34,35]. Whereas high fat is permitted on the Atkins diet and other low carbohydrate diets it is not specifically recommended; as noted above, at least three published studies [23-25] and much anecdotal evidence suggests that, in practice, the major effect is reduction in carbohydrate intake with limited replacement with either fat or protein. In addition, although a deleterious effect of saturated fat, at least in the absence of CHO control, is established [36,37], it has been known since the Seven Countries study [38] that total dietary fat does not correlate with cardiovascular risk and the two effects should not be confused [36,37]. In any case, no particular diet is recommended here; the studies cited include all kinds of interventions, and the underlying rationale is the effect of carbohydrate on insulin. The principle espoused here should be evaluated on this basis.
It seems that a prudent statement of the state of affairs would be that, at this point, dieters have many strategies, none perfect, for weight loss and CHO restriction of some kind is one of them.
CHO restriction and MetS
Our argument in the following exposition is that a patient presenting with a high BMI or large waist line has several options for weight loss. Many factors, including physician experience, ethnic background, personal taste and genetic profile, will determine the first one to be tried. A patient presenting with high BMI and high TAG may have a clear best strategy because of the known benefit of CHO reduction and the accepted deleterious effect of high CHO intake. Data from the literature suggests that a patient with more than two of the symptoms of MetS or the particular combination of high TAG/HDL ratio should clearly try CHO restriction as a first strategy. Conversely, the patient with high BMI and high LDL might sensibly try a low fat strategy first. We provide a summary of cases in which CHO restriction is beneficial in the treatment of MetS or its individual symptoms. The review is meant to be representative rather than comprehensive but we think that the wide variety of cases studied and the range of conditions against a background of accepted effects of carbohydrate on the relevant parameters, provides a strong case for our thesis. In addition, the generally consistent benefit of CHO restriction allows a possible further basis for identifying the common thread in MetS if such truly exists.
CHO restriction improves symptoms of MetS
Table 2 (Table 2) shows the results of single arm studies in which CHO-restricted diets of various compositions and duration were used (for summary of details see, e.g., [39-41]). The regimens include very low CHO ketogenic diets (< 50 g/d) and encompasses subjects that were overweight, presented with symptoms of MetS, or were diabetic. It is clear that CHO restriction is effective in relieving these symptoms. Noteworthy is the recent study of Boden [23] which, while short in duration, carefully measured relevant parameters in patients with diabetes. Patients in this study spontaneously decreased food intake to a substantial degree, were satisfied with the diet, did not show substantial water loss and several were able to reduce or terminate medication.
Table 2 CHANGE
Reference # Subjects Duration CHO (g/d) weight (%) HDL (%) TAG (%) TAG/HDL (%) Glucose (%) Insulin (%)
Rickman et al. 1974 1 Normal Weight Men/Women 3–17 d 7 -4.9 -11.4
LaRosa et al. 1980 2 Obese Men/Women 8 wk 6 -8.3 -5.7 -32.6 -29.9
Phinney et al. 1980 3 Obese Men/Women 6 wk <20 -11.8 -24 -16.3 -57.3
Phinney et al. 1983 4 Normal Weight Men 4 wk <20 0.2 0 -26 -7.7 -23.3
Newbold, 1988 5 Men 3–12 mo 9.6 -34.8 -40.5
Volek et al. 2000 6 Normal Weight Men 8 wk 39 -5.4 10.0 -54.9 -56.9 -3.4 -28.0
Sharman et al. 2002 7 Normal Weight Men 6 wk 46 -2.8 11.5 -33 -39.9 -0.2 -34.2
Meckling et al. 2002 8 Obese Women 8 wk 71 -6.1 4.3 -40.3 -4.1 0.0
Westman et al. 2002 9 Obese Men/Women 6 mo Ad Lib -10.3 19.2 -43.1 -53.6
Dashti et al. 2003 10 Obese Men/Women 12 wk 20 to 30 -13.4 8.3 -50 -53.9 -37.1
Hays et al. 2003 11 Obese Men/Women w/ CVD 6 wk Ad Lib -5.2 -2.9 -39.9 -38.1 -7.4 -30.5
Obese Women PCOS 24 wk Ad Lib -14.3 0.4 -18.5 -18.8 5.7 -49.6
Obese Women Reactive Hypoglycemia 52 wk Ad Lib -19.9 3.4 -13.3 -33.2
Dashti et al. 2004 12 Obese Men/Women 24 wk 40 -14.2 20.4 -60.4 -67.1 -22.6
Boden et al. 2005 13 Obese/diabetic men/women 14 days 21 -1.8 -2 -35 -33.8 -16.0
Improvement is seen in the absence of weight loss
Since it is known that weight loss generally improves MetS, it is important to ask whether beneficial metabolic responses to low CHO diets are dependent on weight loss. The question was specifically addressed by Volek's group [42,43] in normal-weight men and women encouraged to maintain their weight and by Allick and colleagues in patients with diabetes mellitus type 2 [44] (Table 3). The studies in normal weight women [43] and type 2 diabetics [44], in particular, used a cross-over experimental design removing the confounding effect of group differences. These studies were also well-controlled. In the case of Allick, formula was used and, in the studies by Volek, compliance was documented by measuring elevation of serum and urine ketones, thereby eliminating dietary reporting errors as a confounding factor. Improvement in the TAG/HDL ratio ranged from 40 to 55%. In summary, a low CHO regimen clearly improves MetS relative to low fat diets even in the absence of weight loss.
Table 3 CHANGE
Reference # Subjects Duration Diet CHO (g/d) weight (%) HDL (%) TAG (%) TAG/HDL (%) Glucose (%) Insulin (%)
Sharman et al.2002 1 Normal Weight Men 6 wk Ketogenic 46 -2.8 11.5 -33.0 -39.9 -0.2 -34.2
6 wk Low-Fat 271 0.5 0.0 -5.3 -5.3 1.8 13.0
Volek et al. 2003 2 Normal Weight Women 4 wk Ketogenic 43 -2.0 32.0 -30.2 -47.2 -1.9 11.6
4 wk Low-Fat 249 -1.3 -7.7 3.8 0.5 -5.3 18.7
Allick et al. 2004 3 Type 2 Diabetics 2 wk Ketogenic 0 0 23.5 -43.9 -55 -16.9 -16.7
2 wk Low-Fat 775
In addition to studies in which weight maintenance was a feature of experimental design it is important to consider data reported by Foster [45] (Figure 1, (Table 4), who compared low CHO and LF diets. It is widely quoted that the low CHO diet is better at 6 months but that there is no difference in the diets at 12 months. However, it has been pointed out [12] that the particular form of low CHO diet used (Atkins diet) allowed increases in CHO consumption as the trial progressed indicating that it is likely this reintroduction of CHO that predisposes to long-term regain in weight. Most notable in this study, is that the improvement in lipid profile persisted (Figure 1, (Table 4) even after the effect on weight loss disappeared.
Figure 1 Comparison of features of Metabolic Syndrome on low carbohydrate vs. high carbohydrate diets. Data from reference [45].
Table 4 CHANGE
Reference # Subjects Duration Diet CHO (g/d) weight (%) HDL (%) TAG (%) TAG/HDL (%) Glucose (%) Insulin (%) DBP (mm Hg)
Brehm et al. 2003 1 Obese Women 6 mo Low-CHO 41–97 -9.3 13.4 -23.4 -32.4 -9.1 -14.8 -5
LF 163–169 -4.2 8.4 1.6 -6.3 -4.0 -23.0 -1
Sondike et al. 2003 2 Overweight Adolescents 12 wk Low-CHO 37 -10.7 8.7 -40.5 -45.2
LF 154 -4.1 4.2 -5.4 -9.2
Samaha et al. 2003 3 Obese Men/Women 6 mo Low-CHO 150 -4.5 0.0 -20.2 -20.2 -8.6 -27.3
LF 201 -1.4 -2.4 -4 -1.6 -1.6 5.6
Foster et al. 2003 4 Obese Men/Women 1 yr Low-CHO ad lib -7.3 18.2 -28.1 -29.5
LF ad lib -4.5 1.4 0.7 -2.6
Volek et al. 2004 5 Overweight Women 4 wk Low-CHO 29 -3.9 1.3 -23 -28.3 -3.8 -8.8
LF 186 -1.4 -8.6 -11.2 -4.2 1.3 23.2
Sharman et al. 2004 6 Overweight Men 6 wk Low-CHO 36 -5.6 -3.3 -44.1 -42.3 -5.8 -41.5
LF 224 -3.6 -6.6 -15 -8.3 -5.2 -28.1
Brehm et al. 2004 7 Obese Women 4 mo Low-CHO 69 -10.8 16.3 -37.3 -46.1 -9
LF 174 -6.8 4.5 -10.3 -14.2 -3
Meckling et al. 2004 8 Obese Men/Women 10 wk Low-CHO 59 -7.7 12.2 -29.4 -37.1 -8.0 -28.7 -6.1
LF 225 -7.4 -15.4 -25.4 -11.8 -10.2 -3.3 -5
Stern et al. 2004 9 Obese Men/Women 1 yr Low-CHO 120 -3.9 -2.8 -28.6 -26.8
LF 230 -2.3 -12.3 2.7 29.6
Yancy et al. 2004 10 Obese Men/Women 24 wk Low-CHO 30 -12.3 9.8 -47.2 -51.8 -6
LF 198 -6.7 -2.9 -14.4 -12.1 -5.2
Aude et al. 2004 11 Obese Men/Women 12 wk Low-CHO ad lib -6.2 -2.6 -23.2 -21.1
LF ad lib -3.4 -7 -10.5 -3.8
Seshadri et al. 2004 12 Obese Men/Women 6 mo Low-CHO 113 (-8.5) -2.4 -7.4 -40
LF 198 (-3.5 kg) -2.4 -2.3 11.2
McAuley et al. 2004 13 Obese Women 8 wk Low-CHO 41 -6.9 0.9 -38.8 -44.2 -5.9 -39.3
LF 172 -4.4 -6 -17.5 -15.1 -0.1 -28.4
Mod-PRO 130 -5.8 -4.1 -33.9 -31.8 -3.9 -24.4
Dansinger et al. 2004 14 Obese Men/Women 2 mo Low-CHO 103 -4.7 8.8 -27.6 -26.20 -10 -29.5
Mod-PRO 158 -4.6 4.6 -34 -30.50 -9.3 -27.7
LF 183 -4.3 -0.6 -7.1 -5.60 -5.7 -11
UltraLF 230 -4.9 -10.9 -0.6 8.40 -3.5 -7.7
In ad lib. comparisons low CHO diets do better than low fat diets for weight loss and MetS
At this point, we have established that CHO restriction improves MetS and that this can be independent of weight loss. Weight loss can, of course, occur with low fat diets and we next consider the extent to which one or the other strategy is more effective. (Table 4) summarizes results of several studies in the literature demonstrating that low CHO diets generally do better than low fat diets in ad lib. comparisons. Although there is great variability, a pattern of better responses on very low CHO is evident. It is notable that Samaha et al studied a population in which 39 % had diabetes and 43% had MetS [46].
Figure 1 shows data from the study of Foster, et al. [45] and, as noted above, despite the relative similarity in weight loss, the markers of MetS were more favorable in the low CHO arm than the LF arm.
In isocaloric comparisons, low CHO diets do better than LF diets for weight loss
Because weight reduction is considered the first line of attack in MetS or frank diabetes it is worth considering the record of low CHO diets on this parameter alone. It is generally agreed that the major effect of a low CHO diet is a spontaneous reduction in calories. In studies mentioned above, subjects did not significantly increase fat or protein intake but merely removed CHO from their diets [23-25]. Foster and Samaha also attributed the better performance of low CHO arm to decreased caloric intake, although this was not actually measured.
Beyond spontaneous caloric reduction, however, it has been shown that the macronutrient composition of the diet can affect the efficiency of energy utilization and greater efficacy, the so-called metabolic advantage, of low CHO diets compared to LF diets has been the subject of several reports (Reviews: [39,41]. It has long been argued that there must be some mistake because it is physically impossible and would violate the laws of thermodynamics. We have shown this argument is based on misunderstanding of the laws of thermodynamics [39,47-49] and the effect of variable efficiency is now better accepted [50,51]. The precise conditions that allow the so-called metabolic advantage to occur are not known although Cornier, et al. [51] have suggested that those subjects with insulin resistance will show a metabolic advantage on a low CHO diet whereas those who are insulin sensitive do better on low fat. This is consistent with the proposal here, namely that MetS, where insulin-resistance is generally considered a major component, can be defined by the response to CHO restriction. The study of Cornier, et al. [51] had only a small number of subjects and the low CHO arm was not particularly low (40%) but their theory follows from the general rationale of the effect of CHO on energy efficiency. The factors that determine whether a metabolic advantage can play a role in a CHO restricted diet is unknown but given that the insulin resistance association is reasonable, it would seem that some form of CHO restriction is one of the standard, if not preferred attacks on obesity where MetS is suspected.
Figure 2 shows data from Golay, et al. [52] This study is widely quoted as an example of how weight loss is independent of macronutrient composition; although the low CHO arm did better in weight loss, this was judged not significant. This may well be an experiment in which metabolic advantage does not occur – the effect is only possible, not required [39]. It is clear, however, from the figure that there is improvement in TAG and insulin and Golay's conclusion was that "...considering the greater improvement of fasting blood insulin, the glucose/insulin ratio and blood triglyceride, the low carbohydrate diet (25%) could be more favourable in the long-term [52]."
Figure 2 Per cent change in response to diet. Low carbohydrate (dark blue 25 % CHO) and low fat diets(light blue: 45 % CHO). Data from Golay, et al. [52].
Figure 3 Effect of carbohydrate on parameters of Metabolic Syndrome. Comparison of 40% CHO (blue) and 55% (red) CHO diets. Data from [54].
Is this new?
In Edgar Allan Poe's detective story The Purloined Letter, the police search the apartment for a missing blackmail note [53]. In the end, Poe's detective, Auguste Dupin reveals that it had been in plain view on the fireplace all along. The effect of CHO reduction on the symptoms of MetS has, in fact, been visible for some time. In a classic review in 1986, Reaven demonstrated the relative effects of 40% and 55% CHO [54]. Figure 3 shows data from that study: day long glucose, insulin and TAG levels were improved by the low CHO diet. Of interest, is that fasting glucose is not different on the two diets but there is a clear difference in the time course, common to several low CHO interventions (Table 3)). Reaven's experiment suggests that a nutritional approach to MetS is possible by lowering dietary CHO. The experiment should sensibly have spurred research to see whether still lower CHO had further beneficial effect. It took many years, however, before this was done.
Figure 4 Effect of diet on plasma glucose. Mean plasma glucose concentration before (triangles) and after 5 weeks on control diet (yellow circles: (CHO:fat:protein = 55:30:15)) or 5 weeks on lower carbohydrate diet (blue circles: (20:50:30)). Meal points are Breakfast (B), lunch (L) and dinner(D) plus 2 snacks (S1, S2). Data from reference [57].
Substitution of protein for CHO improves MetS
Despite the evidence from Reaven's experiment, a barrier to progress in understanding the role of CHO restriction was the accepted idea that high fat was unhealthy. At the same time, it was thought that an increase in protein would be deleterious for type 2 diabetics because of the increase in glucose due to gluconeogenesis. Nuttall and Gannon have summarized the history of this problem and work in their lab showed that, in fact, glycemic control was enhanced by a diet that was 40 % CHO, with protein replacing part of the carbohydrate [55,56]. Most recently, this group has shown the benefit of a 20 % CHO diet with higher protein [57]. Results in Figure 4 show a pattern similar to Reaven's but much more dramatic. Similar striking differences in the control of insulin and TAG were also demonstrated.
Figure 5 Changes in cholesterol/HDL for substitution for fat. Fat in the average American diet was substituted with the indicated substances at 10 % of energy. Data from reference [62].
Explicit low-fat/high CHO interventions exacerbate MetS
In the approach taken here, we see dietary fat as playing a largely passive role (not withstanding differences between different fats) and the disposition of dietary fat is controlled by insulin and other hormones that, in turn, depend on dietary CHO which we take as the controlling variable. Thus, characterizing a low CHO diet as high fat [58,59] ignores the question about the underlying mechanism; we have recently raised the question of whether "high fat" is a meaningful description in the absence of information about CHO [60]. In the cases we discuss next, the focus is interventions described or designed as low fat. Our point here, however, is that although low fat diets exacerbate the markers of MetS, it is likely the high CHO rather than the fat level that is important.
A very influential paper by Garg, et al. [61] describes a four center randomized study of patients with type 2 diabetes receiving glipizide treatment. The study compared diets where monounsaturated fat was substituted for CHO or vice-versa. Because of its importance, we quote from this article. The rationale for the study, according to the authors is that:
"Compared with diets rich in saturated fats, low-fat, high-carbohydrate diets are reported to reduce serum low-density lipoprotein (LDL) cholesterol levels. Recent studies, however, suggest the high-carbohydrate diets may accentuate hypertriglyceridemia, reduce serum high-density lipoprotein (HDL) cholesterol concentration, and may even worsen hyperglycemia and/or raise plasma insulin levels."
In short, the question is whether high CHO diets worsen MetS. The conclusions of the paper state:
"The study confirms that HC (high carbohydrate diets) increase plasma TAG levels and increase VLDL-C concentrations in NIDDM patients. In this study the HC diet raised fasting plasma triglyceride levels and VLDL-C concentrations by 24 % and 23 % respectively compared with the HMUF diet. Furthermore, daylong levels of plasma triglycerides were also elevated on the high-carbohydrate diet. Consistent with the results of previous studies, plasma levels of total cholesterol and LDL cholesterol were not different on the two diets in this study. The study, therefore substantiates the fact that high carbohydrate diets offer no advantage in lowering LDL levels in NIDDM patients compared with high-fat diets that are low in saturated fats."
The general case: substitution of fat for CHO improves MetS
The substitution of fat for CHO is, in fact, generally beneficial for MetS. In a recent meta-analysis, Mensink, et al. [62] showed the effect of substitution of different fat sources or carbohydrate for the fat in the average US diet at 10% of energy. The conclusion was that substitution of carbohydrate had the most unfavorable response on the total cholesterol to HDL ratio, significantly worse than butter or palm oil (Figure 5).
Summary of review and hypothesis to this point
We have summarized work in the literature showing that low CHO interventions improve the markers of MetS in normal subjects, patients with MetS and diabetics. In comparative studies, they are at least as effective as low fat diets for weight loss and, tend to show better improvement in the other markers of MetS. Isocaloric studies similarly support the idea that the markers of MetS respond preferentially to low CHO diets.
The state of accepted scientific thinking for the world at large is unknown but we have made the case that it is an acknowledged principle that low CHO diets tend to reduce TAG, raise HDL and improve glycemic control whereas LF/high CHO diets tend to have the opposite effect. Perhaps the strongest indication that such an idea is generally accepted is the paper by Rock, et al [63] where the effect of low fat diets in cancer patients was studied. To demonstrate compliance with the low fat recommendations, the authors showed increased TAG and reduced HDL levels. These effects were judged not significant enough to cause a risk for CVD but demonstrate that low fat (higher CHO) diets point in that direction.
We reiterate that this article is not meant to make recommendations – for which many factors must be considered – but rather to show the association between CHO restriction and improvement in symptoms of MetS. Many low fat interventions have successfully reduced LDL, an established risk factor for CVD. If our hypothesis is correct, however, these same interventions should worsen the features of MetS. To some extent this is established from the principles noted above, but we provide an example from a well done experiment in the literature to support this corollary. Finally, in considering the dialectic of treating MetS with CHO restriction vs. high LDL, with low fat diets, it is important to consider both individual variation and the role of LDL particle size. We consider that last.
Studies of low-fat diets
The most salient feature of the obesity epidemic from the standpoint of food consumption is the dramatic increase in CHO intake and the reduction in fat intake (for men, the absolute amount). To our knowledge, the decreased fat intake has not been accompanied by reduction in the incidence of CVD in unmedicated population. These data suggest that low fat diet recommendations per se are not likely to help MetS. If the fundamental idea proposed here is correct, then experimental interventions targeting lower fat that concomitantly raise CHO should, in fact, have a deleterious effect on the markers of MetS. Again, one will have to decide if the symptoms of MetS are more important than LDL or total cholesterol which are typically reduced on LF diets in the unmedicated population.
Delta-1 Study
The Delta-1 study is one of the very well done trials involving a large number of participants [64]. The goal was to determine "the effects of reducing total fat and saturated fat" although this is slightly misleading in that only saturated fat was reduced and any reduction in total fat was a consequence of this. The randomized and balanced diets that were compared all contained approximately 15 % of calories as protein. Other macronutrients were as shown in Table 5. (Table 6) shows the outcomes for all groups on lipids in the Delta-1 study. The results indicate that LDL is significantly reduced although only SEM is given so that it is not possible to know the range of responses of subjects. As a group, however, there were step-wise reductions in LDL and HDL going from the average American diet (AAD) to Step 1 to the low saturated fat diet. There is, as well, a corresponding increase in TAG and hence the TAG/HDL ratio. The authors concluded that the reduction in LDL should be associated with 10% to 20% reductions in cardiovascular disease in the population. Other authors have argued, however, that the effect of the 13% higher HDL seen on the AAD might be associated with a 36% reduction in the risk of death from coronary disease or of myocardial infarction [65]. Again, the purpose here is not to decide on the relative risk attached to different markers but only to point out that the markers for MetS provide another side to the story. An important follow-up in the Delta study to determine HDL subpopulations [66], showed that the more anti-atherogenic HDL2 particles were, in fact, decreased by reductions in saturated fat.
Table 5 Macronutrient composition of diets in the Delta-1 study. Data from reference [64].
Diet CHO (%) total fat (%) SFA (%) MUFA (%) PUFA (%)
Average American diet (AAD) 48 34 15 13 7
Step 1 55 29 9 13 7
Low Saturated Fat 59 25 6 12 7
Table 6 Ave American Diet vs. Step 1 Ave American Diet vs. Low Sat Fat
Ginsberg, et al.,1998 reference [64] AAD Step 1 delta % change pre post delta % change
Total cholesterol 202.1 191 -11.1 -5.5 202.1 183.4 -18.7 -9.3
LDL 131.4 122.2 -9.2 -7.0 131.4 116.9 -14.5 -11.0
TAG (mmol/l) antiln (log) 85.1 92.4 7.3 8.6 85.1 93 7.9 9.3
HDL (mmol/l) 52.2 48.5 -3.7 -7.1 52.2 46.2 -6.0 -11.5
Total/HDL 4.1 4.16 0.1 2.2 4.1 4.21 0.1 3.4
TAG/HDL (arbitrary units) 1.6 1.9 0.3 16.9 1.6 2.0 0.4 23.5
The overall conclusion is that "dietary changes suggested to be prudent for a large segment of the population will primarily affect the concentrations of the most prominent antiatherogenic HDL subpopulation. However, the simultaneous reduction in the atherogenic LDL subpopulation will most likely offset any potential negative effect on cardiovascular risk." As noted above, the decision as to "most likely" outcome must rest with individual patients and physicians.
Role of individual responses and LDL heterogeneity
As noted in Garg's study, changes in LDL may not be as reliable as changes in other markers. Volek, et al. for example showed that whereas TAG was reduced in almost every subject on a low CHO diet, responses in LDL were highly variable [67]. The importance of LDL subpopulations has recently been appreciated and, unlike total LDL, changes in specific LDL particles show a consistent pattern with respect to dietary change.
Greater atherogenic potential is associated with small, dense LDL particles [68]. Krauss and coworkers have carried out impressive work in defining the variability in different individuals. They identified a genetically influenced pattern (B) in people whose plasma contains small LDL particles. This subpopulation, typically 30 % of the American population, responded to low-fat diets by lowering LDL but the pattern B persisted [69]. The remaining subpopulation with larger buoyant particles (pattern A) responded to reduction in fat intake by a shift to the more atherogenic pattern B. Thus, for most of the populations studied, replacing dietary fat with CHO leads to a worsening of the LDL size distribution [70]. In a study described in reference [70], similar effects were seen when protein was substituted for carbohydrate without significant change in the fat content or composition. As summarized by Krauss, "This indicates that carbohydrate rather than fat is a major dietary determinant of expression or phenotype B in susceptible individuals." Although probably a semantic point, "susceptible" is redundant and describing pattern A and B as phenotypes may not be precise: Krauss has summarized how the relative amounts of CHO and fat affect the prevalence of pattern B [71] and the conclusion is that a strong relation exists between CHO intake (ranging from 40 to 75%) and the prevalence of the pattern B phenotype (Fig 6). In other words, there appears to be a continuous variability in phenotype characterized by sensitivity to CHO and everyone may be susceptible to conversion to pattern B at some CHO/fat ratio. The extrapolated line in Figure 6 suggests that a truly low CHO diet might reduce the level of atherogenic subtype to zero. Thus, whereas we have described the dialectic in practical applications as balancing the improvement in MetS with CHO restriction and the improvement in LDL from low-fat diets, focusing on LDL may have some caveats. In general, a growing body of work has shown improvement in LDL pattern switching from high CHO to low CHO diets [42,43,72-75].
Figure 6 Prevalence of pattern B phenotype as a function of the percentage dietary CHO in men. Data from reference [71].
The pattern B phenotype rarely occurs in isolation and, our major concern here is that it is metabolically linked to and co-expressed with other characteristics of MetS, particularly elevated TAG and low HDL. Krauss and colleagues reported that switching from a low CHO/high-fat diet (46% fat) to a high CHO/low-fat diet (26% fat) resulted in lowering of LDL, but also a worsening of TAG and HDL when switching to the low fat diet [69]. In men that were pattern A at a fat intake of 20 to 24%, a further reduction in dietary fat to 10% and CHO to 76% of energy resulted in conversion to B, with continued worsening of TAG and HDL, and no additional LDL-lowering [76].
The TAG/HDL connection
In the search for markers for both insulin resistance and predisposition to CVD, recent research has focused on the value of the ratio of TAG:HDL. McLaughlin, et al. [8] have shown a correlation between insulin resistance as measured by steady-state glucose levels after infusion of glucose, insulin and octreotide (to suppress endogenous insulin secretion). The conclusions of their study were that TAG and HDL were independently related to insulin resistance and the TAG/HDL ratio was the best predictor of insulin resistance. Of importance here is that the results showed that this ratio is comparable to the ATP III criteria for MetS in predicting insulin resistance and "even better in prediction the LDL phenotype B in two separate populations who were on different diets." This is, in fact, only the most recent of several studies (references in [8]) that have shown a correlation between TAG/HDL and insulin resistance and CVD risk as measured by LDL particle size. Table 2-4 as well as Figures 2 and 5 indicate that low CHO diets reliably reduce this marker. Inability to recognize this is again due to the separate conditions in which they are measured and the continued emphasis on reducing dietary fat above all else.
McLaughlin's analysis [8] identified a TAG/HDL ratio of ≥ 3.5 as a cutoff for identifying the insulin-resistant patient most at risk for CVD. It is of interest that in Foster's study [45] described above (Table 3; Figure 1), the average beginning values were 4.6 and 4.3 in the low CHO and low-fat arms, respectively, substantially above this cutoff value. After six months, the low CHO arm had reduced this marker to 3.7 while the LF group showed little change at 4.2. Similarly, in the Delta-1 study, neither the Step 1 diet nor the low fat diet were able to improve the TAG/HDL ratio which was above the threshold value of 3.5 (Table 6)).
Mechanism
A recent review by Ginsberg [77] has provided an excellent description of the possible mechanisms and central role of insulin resistance in mediating the dyslipidemia of MetS. In combination with a proposal by Volek [13] on the mechanism for the reversal of this process, a reasonable understanding of the connection between MetS and CHO restriction is possible.
A primary target of insulin is hormone-sensitive lipase. Adipocyte insulin resistance, plausibly a down regulation of insulin response due to continued stimulation (from higher dietary CHO), leads to increased lipolysis [78,79]. This will lead to greater delivery of fatty acids and an increase in hepatic esterification, and subsequent over production of VLDL, particularly the TAG-rich VLDL1. In combination with impaired plasma TAG clearance, a constant state of hypertriglyceridemia in the postabsorptive and postprandial period occurs. This leads to the exchange of TAG in VLDL for cholesteryl ester in LDL. The resulting TAG-rich LDL particle is a preferred substrate for hepatic lipase and lipoprotein lipase and thereby for generation of small, dense LDL. A similar neutral lipid exchange likely occurs with HDL whereby TAG-rich HDL is hydrolyzed by lipoprotein lipase resulting in the generation of smaller HDL particles that are rapidly removed from the circulation. In this way, elevated TAG resulting from disruption in insulin function, plays a central role in regulating the atherogenic dyslipidemia of MetS.
As noted above, Volek, et al. [13] have reviewed many studies showing that CHO restriction results in significant reductions in postprandial lipemia, and beneficial effects on HDL and intravascular processing of lipoproteins. A key component is what might be called the fatty acid paradox. Whereas insulin resistance is frequently characterized by high fatty acid levels, CHO restriction can improve insulin resistance while raising fatty acids. The latter effect is presumed to be due to lower insulin levels and disinhibition of hormone sensitive lipase. This is accompanied by enhanced cellular fatty acid uptake, mitochondrial transport and increased oxidation. The bias toward fat oxidation over storage reduces hepatic TAG and reduces synthesis and secretion of VLDL.
Discussion
A joint position statement by several organizations recommended "current dietary guidelines from the ADA, AHA and the NCEP-ATP III.... These recommendations may require modification, however, as new information is generated from additional diet intervention studies [80]." Rather than additional studies, however, we provide new information from an evaluation of papers already in the literature that may provide a basis for modification. Data compiled in Figures 2, 3, 4 show that low CHO diets improve the symptoms of MetS as defined by five common criteria. We propose that in addition to potential value as therapy, the response to CHO restriction might be considered an operational definition of MetS. Beyond formal categorizing, the idea is consistent with the generally held belief that MetS is intimately involved with some form of insulin resistance. (The importance of other factors such as inflammation are not mutually exclusive). Cornier et al. showed differential benefit of low CHO vs. low fat diets for people with or without insulin resistance [51]; this study might be thought of as a model for future work to follow this line of thinking. We emphasize that our main point is that an intimate connection between CHO restriction and the complex of symptoms of MetS is seen in the literature of both low CHO studies and low fat (higher CHO) studies. Individual judgment as to the importance of MetS compared to other specific factors or more global assessments such as the Framingham criteria will determine how use is made of this connection. We do believe, however, that ignoring studies on CHO restriction would be unscientific and unproductive.
How low is low carbohydrate?
Goneril. What need you five-and-twenty, ten, or five...?
Regan. What need one?
- William Shakespeare, King Lear.
The data summarized here suggest that some degree of CHO restriction would provide a first line of attack against the symptoms of MetS. The principle of CHO restriction is that by keeping insulin low, metabolism is biased towards lipid oxidation rather than storage, or the effects of fatty acids on peripheral tissues. Most studies that reported deleterious effects of saturated fat have been carried out in the presence of high CHO and there is a real question whether such effects carry over into hypocaloric conditions or those where insulin is better controlled [28,60,81,82].
In general, whereas current thinking in MetS emphasizes the consequences of insulin resistance, we feel that the role of CHO-induced hyperinsulinemia as a causative factor in generating the initial insulin resistance, dyslipidemia or obesity has been under-appreciated. In any case, there are now many ways to implement CHO restriction ranging from ketogenic diets (less than 50 g/d) to diets based on glycemic index, an indirect method of reducing insulin excursions. The question is how low is low? It is clear that 40% CHO is better than 55% for MetS and there has been some reluctance to go lower even though the studies that have done so show continued improvement. Perusal of Table 2-4 suggests that the lower, the better. Insofar as pattern B is associated with insulin resistance and MetS, examination of Figure 4 supports this idea. The barriers to exploring lower CHO diets appears to be continued emphasis on low fat intake although it has been known since Keys's Seven Country study that total fat in the diet does not correlate with cardiovascular risk [38]. At least as indicated in its popular diet book, the No-Fad Diet [83], the American Heart Association has removed its limitation on total fat which should open the door to more flexible diet interventions. Since many studies have shown that there is frequently a spontaneous reduction in total caloric intake in very low CHO diets and that CHO removed is not replaced by fat or protein, very low CHO now appears as a far more prudent choice than judged in the past.
Is this new?
The phenomenon of CHO-induced hypertriglyceridemia is long established [84-88]. In addition, low fat diets are known to reduce, not only LDL, but also HDL levels. For example, the 2004 recommendations of the American Diabetic Association (ADA) state that "Low-saturated fat (i.e., 10% of energy) high carbohydrate diets increase postprandial levels of plasma glucose, insulin, triglycerides and, in some studies, decrease plasma HDL cholesterol when compared in metabolic studies to isocaloric high monounsaturated fat diets." This is our conclusion (our italics). We think the ADA statement could have been more clearly worded: "Substitution of CHO for monounsaturated fat "increase(s) postprandial levels of plasma glucose, insulin..." or could have been more comprehensive: "substitution of CHO or protein for fat increases..." In other words, it has been known for some time that low fat reduces HDL as well as LDL concentrations as described clearly in the Delta-1 study. Again, Rock's study that used the increase in TAG and decrease in HDL as a marker for compliance to a low fat diet supports the idea as an accepted principle.
In combination with the experimentally observed and intuitively obvious reduction in fasting blood glucose and insulin, our proposal for the importance of CHO replacement in the diet hardly seems new. Yet such an idea, to our knowledge has never been made explicit. The primacy of the low fat paradigm in traditional thinking may have played a role in ignoring this obvious correlation. However, we have, in various places, presented or summarized evidence that low CHO diets have a beneficial effect on MetS [13,16] but the tight connection proposed here conceptually eluded us. Having raised the question, however, it is now clear that it is perfectly consistent with established knowledge.
Is MetS useful?
The intricacies of the debate among official agencies on the clinical importance of MetS [9] are beyond the scope of this article. As a first order approximation, however, we are inclined to follow Reaven's strategy in the Point/Counterpoint paper [10,89] for assessing the need for the concept of MetS. He describes a patient with a BMI of 27.8 kg/m2 who would not conform to the WHO definition of MetS because of acceptable values of glucose, TAG and an HDL level of 37 mg/dL and he points out that if the HDL level fell to 33 mg/dL such a patent would fit a new criteria but would not sensibly be treated in a different way. Under the approach considered here, the patient with the higher HDL would be presented with a number of options for weight loss whereas the patient with lower HDL might reasonably be counseled to CHO restriction as a first strategy. We therefore think that MetS or some combination of markers by any other name would have much more virility than the description of MetS as Methuselah in his amusing Counterpoint [87]. We also side with the "Pro" position on the value for basic research. Reaven raises the critical question: "How can there be a common etiology for a diagnostic category based on satisfying 3 of 5 arbitrarily defined criteria when any combination of the 3 will define the same phenotype as any other trio of abnormalities.?" A plausible answer to this question is that if all the markers in MetS are related to hyperinsulinemia and/or insulin resistance, then the relative Km's for insulin for the different target proteins and different tissues are likely to lead to a variable time course for specific individuals and markers are likely to exceed cutoff at different times for each patient. Appearance of one marker may then be indicative of other still silent conditions.
Recommendations for the American population
The data summarized here suggest that there is value in the definition of MetS and that a nutritional strategy based on CHO restriction might sensibly be the "default" diet, the first to be tried, for patients with MetS. In the case of normal weight individuals with MetS, CHO restriction may be the only effective non-pharmacological approach for treating the diversity of symptoms. The choice of any intervention, however, depends on individual assessment of the relative importance of different risk factors and our goal here is the establishment of the close link between CHO restriction and MetS rather than any recommendation.
The recent AHA/NHLBI Scientific Statement on Metabolic Syndrome [90,91] as well as the ATP III emphasize as the primary target, LDL, a marker that is not considered a feature of MetS and that may not even be high in many patients with MetS. Whereas nutritional recommendations are quite general (reduce weight, increase exercise), these reports emphasize a low fat diet (although limiting simple sugars). We think this is inconsistent and we have made the case that a low fat diet, if high in carbohydrate, seems to be widely accepted as raising TAG, lowering HDL and worsening glycemic control, seemingly the wrong thing for MetS. We also disagree with the assertion in the AHA/NHLBI statement that most low CHO diets are high in saturated fat. This statement is undocumented and, as noted in the introduction, essentially equates all reduced carbohydrate approaches and does not seem to distinguish between total and saturated fat. Even if it were true the statement avoids the question of the effect of saturated fat in the presence of low CHO or hypocaloric diets [60,81,82,92]. In practical terms, a recommendation to reduce saturated fat on low CHO diets might be more helpful than blanket prohibition. In addition, the AHA/NHLBI report [90] presents the rationale for low CHO as the effect on appetite. Whereas this may be a component, it has been stated many times, and is part of basic biochemical education [16,93-96], that the rationale of CHO restriction is the control of metabolism by insulin regulation. The effects described above clearly support this. Although we think that, within the framework of MetS, some recommendations can be made, in the end we are probably in agreement with Reaven's judgment that "What is required is less advice and more information [54]."
Questions raised
The data summarized here leave little room for doubt that the generally accepted deterioration in HDL and TAG levels with low fat diets and the established improvement in glycemic control with CHO restriction are part of a more general picture. The hypothesis that response to CHO restriction (because of the effect on insulin) is the defining feature of MetS is the proposed generalization. This idea raises several questions.
Is there a threshold level of CHO restriction that is necessary to elicit improvements in MetS? What is the effect of replacing the calories lost from CHO with protein or with fat, and in what proportion? How does this compare to not replacing them at all?
Does excessive CHO consumption cause MetS in susceptible individuals?
What is the relative risk in addressing MetS with CHO restriction compared to low fat diets for reduction in LDL? That is, what is the relative risk of high LDL vs. the symptoms that comprise MetS?
What is the role of genotype in determining the response to CHO restriction?
Summary
Five symptoms common to most definitions of MetS are those that are reliably improved by CHO restriction. Carbohydrate restriction is one strategy for weight loss but, in addition, improves glycemic control, insulin levels, TAG and HDL levels even in the absence of weight loss. We suggest that response to CHO restriction may, in fact, be an operational definition of MetS. Its underlying basis would rest on the idea that the features of MetS are associated with a disruption in insulin metabolism which is strongly influenced by dietary CHO. The extent to which this definition is useful may depend on its application by individual practitioners. Experimental studies that follow its lead or conversely disprove its fundamental premise should advance our understanding of obesity, diabetes and CVD. Dismissing CHO restriction without evidence, or expressing "concerns" rather than offering data will probably be less productive.
Abbreviations
Adult Treatment Panel (ATP III), American Diabetes Association (ADA), American Heart Association (AHA), Average American diet (AAD), carbohydrate (CHO), low fat (LF), Metabolic Syndrome (MetS), National Cholesterol Education Program (NCEP), triacylglycerol (TAG), World Health Organization (WHO)
Competing interests
The author(s) declare that they have no competing interests.
Note
Table 2 - Effect of carbohydrate restriction on markers for Metabolic Syndrome (See Table 2)
Data shown in bold indicate improvement in marker, plain, worsening. References:1. [97]; 2. [24]; 3. [98]; 4. [99]; 5. [100]; 6. [101]; 7. [42]; 8. [102]; 9. [103]; 10. [104]; 11. [74]; 12. [105]; 13. [23].
Table 3 - Effect of carbohydrate restriction on markers for Metabolic Syndrome under conditions of constant body mass (See Table 3)
Data shown in bold indicate low CHO shows greater improvement in markers for MetS than LF; plain, LF is better. Table reference 3 shows the ratio of low CHO to LF. References: 1. [42]; 2. [43]; 3. [44].
Table 4 - Comparison of low CHO vs. LF diets on markers for Metabolic Syndrome (See Table 4)
Data shown in bold indicate low CHO (or mod-PROT) shows greater improvement in marker than LF; plain, LF is better. Experiment in reference [106] was carried out for a longer time period but diets became very similar. References:1. [107]; 2. [113]; 3. [46]; 4. [45]; 5. [108]; 6. [72]; 7. [109]; 8. [110]; 9. [111]; 10. [112]; 11. [73]; 12. [75]; 13. [59]; 14. [106].
Table 6 - Outcomes of the Delta-1 study (See Table 6)
Data from reference [64]. bold indicates improvement in the parameter from Step 1 or low Saturated Fat diet compared to AAD; plain indicates worsening of parameter compared to AAD.
==== Refs
Eckel RH Grundy SM Zimmet PZ The metabolic syndrome Lancet 2005 365 9468 1415 1428 10.1016/S0140-6736(05)66378-7 15836891
Grundy SM Brewer HBJ Cleeman JI Smith SCJ Lenfant C Definition of metabolic syndrome: Report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition Circulation 2004 109 3 433 438 10.1161/01.CIR.0000111245.75752.C6 14744958
Grundy SM Hansen B Smith SCJ Cleeman JI Kahn RA Clinical management of metabolic syndrome: report of the American Heart Association/National Heart, Lung, and Blood Institute/American Diabetes Association conference on scientific issues related to management Arterioscler Thromb Vasc Biol 2004 24 2 e19 24 10.1161/01.ATV.0000112379.88385.67 14766740
Tonkin A The metabolic syndrome -- a growing problem European Heart Journal Supplements 2004 6 A37 A42
Aude YW Mego P Mehta JL Metabolic syndrome: dietary interventions Curr Opin Cardiol 2004 19 5 473 479 10.1097/01.hco.0000134610.68815.05 15316456
Hanley AJ Wagenknecht LE D'Agostino RBJ Zinman B Haffner SM Identification of subjects with insulin resistance and beta-cell dysfunction using alternative definitions of the metabolic syndrome Diabetes 2003 52 11 2740 2747 14578292
Hunt KJ Resendez RG Williams K Haffner SM Stern MP National Cholesterol Education Program versus World Health Organization metabolic syndrome in relation to all-cause and cardiovascular mortality in the San Antonio Heart Study Circulation 2004 110 10 1251 1257 10.1161/01.CIR.0000140762.04598.F9 15326061
McLaughlin T Reaven G Abbasi F Lamendola C Saad M Waters D Simon J Krauss RM Is there a simple way to identify insulin-resistant individuals at increased risk of cardiovascular disease? Am J Cardiol 2005 96 3 399 404 10.1016/j.amjcard.2005.03.085 16054467
Kahn R Buse J Ferrannini E Stern M The metabolic syndrome: time for a critical appraisal: joint statement from the American Diabetes Association and the European Association for the Study of Diabetes Diabetes Care 2005 28 9 2289 2304 16123508
Grundy SM Point: the metabolic syndrome still lives Clin Chem 2005 51 8 1352 1354 10.1373/clinchem.2005.050989 16040840
Reaven GM The metabolic syndrome: requiescat in pace Clin Chem 2005 51 6 931 938 10.1373/clinchem.2005.048611 15746300
Arora SK McFarlane SI The case for low carbohydrate diets in diabetes management Nutr Metab (Lond) 2005 2 16 10.1186/1743-7075-2-16 16018812
Volek JS Sharman MJ Forsythe CE Modification of lipoproteins by very low-carbohydrate diets J Nutr 2005 135 6 1339 1342 15930434
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) final report Circulation 2002 106 25 3143 3421 12485966
Ashen MD Blumenthal RS Clinical practice. Low HDL cholesterol levels N Engl J Med 2005 353 12 1252 1260 10.1056/NEJMcp044370 16177251
Feinman RD Makowske M Metabolic Syndrome and Low-Carbohydrate Ketogenic Diets in the Medical School Biochemistry Curriculum. Metabolic Syndrome and Related Disorders 2003 1 189 198 10.1089/154041903322716660 18370662
Westman EC Yancy Jr. WS Haub MD Volek JS Insulin Resistance from a Low-Carbohydrate, High Fat Diet Perspective Metabolic Syndrome and Related Disorders 2005 3 3 7 10.1089/met.2005.3.14 18370703
Ruderman N Chisholm D Pi-Sunyer X Schneider S The metabolically obese, normal-weight individual revisited Diabetes 1998 47 5 699 713 9588440
Willet W Increasing prevalence of overweight among US adults. The National Health and Nutrition Examination Surveys, 1960 to 1991 JAMA 1994 272 205 211 10.1001/jama.272.3.205 8022039
Poppitt SD Keogh GF Prentice AM Williams DE Sonnemans HM Valk EE Robinson E Wareham NJ Long-term effects of ad libitum low-fat, high-carbohydrate diets on body weight and serum lipids in overweight subjects with metabolic syndrome Am J Clin Nutr 2002 75 1 11 20 11756055
Enns CK Goldman JD Cook A Trends in Food and Nutrient Intakes by Adults: NFCS 1977-78, CSFII 1989-91, and CSFII 1994-95 Family Economics and Nutrition Review 1997 10 2 15
Kennedy ET Bowman SA Powell R Dietary-fat intake in the US population J Am Coll Nutr 1999 18 3 207 212 10376775
Boden G Sargrad K Homko C Mozzoli M Stein TP Effect of a low-carbohydrate diet on appetite, blood glucose levels, and insulin resistance in obese patients with type 2 diabetes Ann Intern Med 2005 142 6 403 411 15767618
Larosa JC Fry AG Muesing R Rosing DR Effects of high-protein, low-carbohydrate dieting on plasma lipoproteins and body weight J Am Diet Assoc 1980 77 3 264 270 7410754
Miller BV Bertino J Reed TG Burrington C Davidson LK Green A Gartung A Nafziger A An Evaluation of the Atkins' Diet Metabolic Syndrome and Related Disorders 2003 1 299 309 10.1089/1540419031361426 18370655
Bray GA The epidemic of obesity and changes in food intake: the Fluoride Hypothesis Physiol Behav 2004 82 1 115 121 10.1016/j.physbeh.2004.04.033 15234599
Martin WF Armstrong LE Rodriguez NR Dietary Protein Intake and Renal Function Nutr Metab (Lond) 2005 2 1 25 10.1186/1743-7075-2-25 16174292
Van der Auwera I Wera S Van Leuven F Henderson ST A ketogenic diet reduces amyloid beta 40 and 42 in a mouse model of Alzheimer's disease Nutr Metab (Lond) 2005
Veech RL The therapeutic implications of ketone bodies: the effects of ketone bodies in pathological conditions: ketosis, ketogenic diet, redox states, insulin resistance, and mitochondrial metabolism Prostaglandins Leukot Essent Fatty Acids 2004 70 3 309 319 10.1016/j.plefa.2003.09.007 14769489
Veech RL Chance B Kashiwaya Y Lardy HA Cahill GFJ Ketone bodies, potential therapeutic uses IUBMB Life 2001 51 4 241 247 11569918
Brand-Miller JC Glycemic index in relation to coronary disease Asia Pac J Clin Nutr 2004 13 Suppl S3 15294465
Brand-Miller JC Postprandial glycemia, glycemic index, and the prevention of type 2 diabetes Am J Clin Nutr 2004 80 2 243 244 15277141
Atkins RC Dr. Atkins' New Diet Revolution 2002 New York , Avon Books
Agatston A The South Beach Diet 2003 New York , Random House
Eades MR Eades MD Protein Power 1996 New York , Bantam Books
Hu FB Manson JE Willett WC Types of dietary fat and risk of coronary heart disease: a critical review J Am Coll Nutr 2001 20 1 5 19 11293467
Hu FB Stampfer MJ Manson JE Rimm E Colditz GA Rosner BA Hennekens CH Willett WC Dietary fat intake and the risk of coronary heart disease in women N Engl J Med 1997 337 21 1491 1499 10.1056/NEJM199711203372102 9366580
Keys A Coronary heart disease in seven countries 1970 41 (Suppl) 1 211
Feinman RD Fine EJ Thermodynamics and Metabolic Advantage of Weight Loss Diets. Metabolic Syndrome and Related Disorders 2003 1 209 219 10.1089/154041903322716688 18370664
Volek JS Westman EC Very-low-carbohydrate weight-loss diets revisited Cleve Clin J Med 2002 69 11 849, 853, 856 8 passim 12430970
Westman EC Mavropoulos J Yancy WS Volek JS A Review of Low-carbohydrate Ketogenic Diets Curr Atheroscler Rep 2003 5 6 476 483 14525681
Sharman MJ Kraemer WJ Love DM Avery NG Gomez AL Scheett TP Volek JS A ketogenic diet favorably affects serum biomarkers for cardiovascular disease in normal-weight men J Nutr 2002 132 7 1879 1885 12097663
Volek JS Sharman MJ Gomez AL Scheett TP Kraemer WJ An Isoenergetic Very Low-Carbohydrate Diet Is Associated With Improved Serum High-Density Lipoprotein Cholesterol (HDL-C), Total Cholesterol to HDL-C Ratio, Triacylglycerols, and Postprandial Lipemic Responses Compared to a Low-Fat Diet in Normal Weight, Normolipidemic Women J Nutr 2003 133 9 2756 2761 12949361
Allick G Bisschop PH Ackermans MT Endert E Meijer AJ Kuipers F Sauerwein HP Romijn JA A low-carbohydrate/high-fat diet improves glucoregulation in type 2 diabetes mellitus by reducing postabsorptive glycogenolysis J Clin Endocrinol Metab 2004 89 12 6193 6197 10.1210/jc.2004-1041 15579777
Foster GD Wyatt HR Hill JO McGuckin BG Brill C Mohammed BS Szapary PO Rader DJ Edman JS Klein S A randomized trial of a low-carbohydrate diet for obesity N Engl J Med 2003 348 21 2082 2090 10.1056/NEJMoa022207 12761365
Samaha FF Iqbal N Seshadri P Chicano KL Daily DA McGrory J Williams T Williams M Gracely EJ Stern L A low-carbohydrate as compared with a low-fat diet in severe obesity N Engl J Med 2003 348 21 2074 2081 10.1056/NEJMoa022637 12761364
Feinman RD Fine EJ "A Calorie is a calorie" violates the second law of thermodynamics. Nutrition Journal 2004 3 9
Feinman RD Fine EJ Whatever happened to the second law of thermodynamics? Am J Clin Nutr 2004 80 5 1445 6; author reply 1446 15531699
Fine EJ Feinman RD Thermodynamics of weight loss diets Nutr Metab (Lond) 2004 1 1 15 10.1186/1743-7075-1-15 15588283
Bray GA Champagne CM Beyond energy balance: there is more to obesity than kilocalories J Am Diet Assoc 2005 105 5 Suppl 1 S17 23 10.1016/j.jada.2005.02.018 15867891
Cornier MA Donahoo WT Pereira R Gurevich I Westergren R Enerback S Eckel PJ Goalstone ML Hill JO Eckel RH Draznin B Insulin sensitivity determines the effectiveness of dietary macronutrient composition on weight loss in obese women Obes Res 2005 13 4 703 709 15897479
Golay A Eigenheer C Morel Y Kujawski P Lehmann T de Tonnac N Weight-loss with low or high carbohydrate diet? Int J Obes Relat Metab Disord 1996 20 12 1067 1072 8968851
Poe EA The Purloined Letter http://xroadsvirginiaedu/~HYPER/POE/purloinehtml 1845
Reaven GM Effect of dietary carbohydrate on the metabolism of patients with non-insulin dependent diabetes mellitus Nutr Rev 1986 44 2 65 73 3703391
Gannon MC Nuttall FQ Saeed A Jordan K Hoover H An increase in dietary protein improves the blood glucose response in persons with type 2 diabetes Am J Clin Nutr 2003 78 4 734 741 14522731
Nuttall FQ Gannon MC Metabolic response of people with type 2 diabetes to a high protein diet Nutr Metab (Lond) 2004 1 1 6 10.1186/1743-7075-1-6 15507157
Gannon MC Nuttall FQ Effect of a high-protein, low-carbohydrate diet on blood glucose control in people with type 2 diabetes Diabetes 2004 53 9 2375 2382 15331548
Ludwig DS Jenkins DJ Carbohydrates and the postprandial state: have our cake and eat it too? Am J Clin Nutr 2004 80 4 797 798 15447882
McAuley KA Hopkins CM Smith KJ McLay RT Williams SM Taylor RW Mann JI Comparison of high-fat and high-protein diets with a high-carbohydrate diet in insulin-resistant obese women Diabetologia 2005 48 1 8 16 10.1007/s00125-004-1603-4 15616799
Feinman R When is a high fat diet not a high fat diet? Nutrition & Metabolism 2005 2 27
Garg A Bantle JP Henry RR Coulston AM Griver KA Raatz SK Brinkley L Chen YD Grundy SM Huet BA Effects of varying carbohydrate content of diet in patients with non-insulin-dependent diabetes mellitus Jama 1994 271 18 1421 1428 10.1001/jama.271.18.1421 7848401
Mensink RP Zock PL Kester AD Katan MB Effects of dietary fatty acids and carbohydrates on the ratio of serum total to HDL cholesterol and on serum lipids and apolipoproteins: a meta-analysis of 60 controlled trials Am J Clin Nutr 2003 77 5 1146 1155 12716665
Rock CL Flatt SW Thomson CA Stefanick ML Newman VA Jones L Natarajan L Pierce JP Chang RJ Witztum JL Plasma triacylglycerol and HDL cholesterol concentrations confirm self-reported changes in carbohydrate and fat intakes in women in a diet intervention trial J Nutr 2004 134 2 342 347 14747670
Ginsberg HN Kris-Etherton P Dennis B Elmer PJ Ershow A Lefevre M Pearson T Roheim P Ramakrishnan R Reed R Stewart K Stewart P Phillips K Anderson N Effects of reducing dietary saturated fatty acids on plasma lipids and lipoproteins in healthy subjects: the DELTA Study, protocol 1 Arterioscler Thromb Vasc Biol 1998 18 3 441 449 9514413
Gordon DJ Knoke J Probstfield JL Superko R Tyroler HA High-density lipoprotein cholesterol and coronary heart disease in hypercholesterolemic men: the Lipid Research Clinics Coronary Primary Prevention Trial Circulation 1986 74 1217 1225 3536151
Berglund L Oliver EH Fontanez N Holleran S Matthews K Roheim PS Ginsberg HN Ramakrishnan R Lefevre M HDL-subpopulation patterns in response to reductions in dietary total and saturated fat intakes in healthy subjects Am J Clin Nutr 1999 70 6 992 1000 10584043
Volek JS Sharman MJ Cardiovascular and hormonal aspects of very-low-carbohydrate ketogenic diets Obes Res 2004 12 Suppl 2 115S 23S 15601959
Lamarche B Tchernof A Moorjani S Cantin B Dagenais GR Lupien PJ Despres JP Small, dense low-density lipoprotein particles as a predictor of the risk of ischemic heart disease in men. Prospective results from the Quebec Cardiovascular Study Circulation 1997 95 1 69 75 8994419
Dreon DM Fernstrom HA Miller B Krauss RM Low-density lipoprotein subclass patterns and lipoprotein response to a reduced-fat diet in men Faseb J 1994 8 1 121 126 8299884
Krauss RM Dietary and Genetic Probes of Atherogenic Dyslipidemia Arterioscler Thromb Vasc Biol 2005
Krauss RM Atherogenic lipoprotein phenotype and diet-gene interactions J Nutr 2001 131 2 340S 3S 11160558
Sharman MJ Gomez AL Kraemer WJ Volek JS Very low-carbohydrate and low-fat diets affect fasting lipids and postprandial lipemia differently in overweight men J Nutr 2004 134 4 880 885 15051841
Aude YW Agatston AS Lopez-Jimenez F Lieberman EH Marie A Hansen M Rojas G Lamas GA Hennekens CH The national cholesterol education program diet vs a diet lower in carbohydrates and higher in protein and monounsaturated fat: a randomized trial Arch Intern Med 2004 164 19 2141 2146 10.1001/archinte.164.19.2141 15505128
Hays JH DiSabatino A Gorman RT Vincent S Stillabower ME Effect of a high saturated fat and no-starch diet on serum lipid subfractions in patients with documented atherosclerotic cardiovascular disease Mayo Clin Proc 2003 78 11 1331 1336 14601690
Seshadri P Iqbal N Stern L Williams M Chicano KL Daily DA McGrory J Gracely EJ Rader DJ Samaha FF A randomized study comparing the effects of a low-carbohydrate diet and a conventional diet on lipoprotein subfractions and C-reactive protein levels in patients with severe obesity Am J Med 2004 117 6 398 405 10.1016/j.amjmed.2004.04.009 15380496
Dreon DM Fernstrom HA Williams PT Krauss RM A very low-fat diet is not associated with improved lipoprotein profiles in men with a predominance of large, low-density lipoproteins Am J Clin Nutr 1999 69 3 411 418 10075324
Ginsberg HN Zhang YL Hernandez-Ono A Regulation of plasma triglycerides in insulin resistance and diabetes Arch Med Res 2005 36 3 232 240 10.1016/j.arcmed.2005.01.005 15925013
Boden G Shulman GI Free fatty acids in obesity and type 2 diabetes: defining their role in the development of insulin resistance and beta-cell dysfunction Eur J Clin Invest 2002 32 Suppl 3 14 23 10.1046/j.1365-2362.32.s3.3.x 12028371
Zammit VA Insulin stimulation of hepatic triacylglycerol secretion in the insulin-replete state: implications for the etiology of peripheral insulin resistance Ann N Y Acad Sci 2002 967 52 65 12079835
Klein S Burke LE Bray GA Blair S Allison DB Pi-Sunyer X Hong Y Eckel RH Clinical implications of obesity with specific focus on cardiovascular disease: a statement for professionals from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism: endorsed by the American College of Cardiology Foundation Circulation 2004 110 18 2952 2967 10.1161/01.CIR.0000145546.97738.1E 15509809
Mozaffarian D Rimm EB Herrington DM Dietary fats, carbohydrate, and progression of coronary atherosclerosis in postmenopausal women Am J Clin Nutr 2004 80 5 1175 1184 15531663
Volek JS Forsythe CE The case for not restricting saturated fat on a low carbohydrate diet Nutr Metab (Lond) 2005 2 21 10.1186/1743-7075-2-21 16135250
American Heart Association No-Fad Diet. A Personal Plan fo Healthy Weight Loss 2005 New York , Clarkson Potter
Hellerstein MK Carbohydrate-induced hypertriglyceridemia: modifying factors and implications for cardiovascular risk Curr Opin Lipidol 2002 13 1 33 40 10.1097/00041433-200202000-00006 11790961
Hudgins LC Effect of high-carbohydrate feeding on triglyceride and saturated fatty acid synthesis Proc Soc Exp Biol Med 2000 225 3 178 183 10.1046/j.1525-1373.2000.22521.x 11082210
Hudgins LC Hellerstein M Seidman C Neese R Diakun J Hirsch J Human fatty acid synthesis is stimulated by a eucaloric low fat, high carbohydrate diet J Clin Invest 1996 97 9 2081 2091 8621798
Parks EJ Hellerstein MK Carbohydrate-induced hypertriacylglycerolemia: historical perspective and review of biological mechanisms Am J Clin Nutr 2000 71 2 412 433 10648253
Parks EJ Krauss RM Christiansen MP Neese RA Hellerstein MK Effects of a low-fat, high-carbohydrate diet on VLDL-triglyceride assembly, production, and clearance J Clin Invest 1999 104 8 1087 1096 10525047
Reaven G Counterpoint: just being alive is not good enough Clin Chem 2005 51 8 1354 1357 10.1373/clinchem.2005.053587 16040841
Grundy SM Cleeman JI Daniels SR Donato KA Eckel RH Franklin BA Gordon DJ Krauss RM Savage PJ Smith SCJ Spertus JA Costa F Diagnosis and Management of the Metabolic Syndrome. An American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement Circulation 2005
Grundy SM Cleeman JI Daniels SR Donato KA Eckel RH Franklin BA Gordon DJ Krauss RM Savage PJ Smith SCJ Spertus JA Costa F Diagnosis and Management of the Metabolic Syndrome. An American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Executive Summary Circulation 2005
Knopp RH Retzlaff BM Saturated fat prevents coronary artery disease? An American paradox Am J Clin Nutr 2004 80 5 1102 1103 15531654
Pogozelski W Arpaia N Priore S The Metabolic Effects of Low-carbohydrate Diets and Incorporation into a Biochemistry Course Biochemistry and Molecular Biology Education 2005 33 91 100 21638552
Harris RA Crabb DW Devlin TM Chapter 22. Metabolic Interrelationships Textbook of Biochemistry With Clinical Correlations 2006 Sixth New York , John Wiley & Sons, Inc.
Makowske M Feinman RD Nutrition education: a questionnaire for assessment and teaching Nutr J 2005 4 1 2 10.1186/1475-2891-4-2 15649324
Smith C Marks AD Lieberman M Basic Medical Biochemistry: A Clinical Approach 2005 2nd Philadelphia , Lippincott Williams & Wilkins
Rickman F Mitchell N Dingman J Dalen JE Changes in serum cholesterol during the Stillman diet Jama 1974 228 1 54 58 10.1001/jama.228.1.54 4406145
Phinney SD Horton ES Sims EA Hanson JS Danforth EJ LaGrange BM Capacity for moderate exercise in obese subjects after adaptation to a hypocaloric, ketogenic diet J Clin Invest 1980 66 5 1152 1161 7000826
Phinney SD Bistrian BR Evans WJ Gervino E Blackburn GL The human metabolic response to chronic ketosis without caloric restriction: preservation of submaximal exercise capability with reduced carbohydrate oxidation Metabolism 1983 32 8 769 776 10.1016/0026-0495(83)90106-3 6865776
Newbold HL Reducing the serum cholesterol level with a diet high in animal fat South Med J 1988 81 1 61 63 3336803
Volek JS Gomez AL Kraemer WJ Fasting lipoprotein and postprandial triacylglycerol responses to a low-carbohydrate diet supplemented with n-3 fatty acids J Am Coll Nutr 2000 19 3 383 391 10872901
Meckling KA Gauthier M Grubb R Sanford J Effects of a hypocaloric, low-carbohydrate diet on weight loss, blood lipids, blood pressure, glucose tolerance, and body composition in free-living overweight women Can J Physiol Pharmacol 2002 80 11 1095 1105 10.1139/y02-140 12489929
Westman EC Yancy WS Edman JS Tomlin KF Perkins CE Effect of 6-month adherence to a very low carbohydrate diet program Am J Med 2002 113 1 30 36 10.1016/S0002-9343(02)01129-4 12106620
Dashti HM Bo-Abbas YY Asfar SK Mathew TC Hussein T Behbahani A Khoursheed MA Al-Sayer HM Al-Zaid NS Ketogenic diet modifies the risk factors of heart disease in obese patients Nutrition 2003 19 10 901 902 10.1016/S0899-9007(03)00161-8 14559328
Dashti HM Mathew TC Hussein T Asfar SK Behbahani A Khoursheed MA Al-Sayer HM Bo-Abbas YY Al-Zaid NS Long-term effects of a ketogenic diet in obese patients Exp Clin Cardiol 2004 9 3 200 205 19641727
Dansinger ML Gleason JA Griffith JL Selker HP Schaefer EJ Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss and heart disease risk reduction: a randomized trial Jama 2005 293 1 43 53 10.1001/jama.293.1.43 15632335
Brehm BJ Seeley RJ Daniels SR D'Alessio DA A randomized trial comparing a very low carbohydrate diet and a calorie-restricted low fat diet on body weight and cardiovascular risk factors in healthy women J Clin Endocrinol Metab 2003 88 4 1617 1623 10.1210/jc.2002-021480 12679447
Volek JS Sharman MJ Gomez AL Judelson DA Rubin MR Watson G Sokmen B Silvestre R French DN Kraemer WJ Comparison of energy-restricted very low-carbohydrate and low-fat diets on weight loss and body composition in overweight men and women Nutr Metab (Lond) 2004 1 1 13 10.1186/1743-7075-1-13 15533250
Brehm BJ Spang SE Lattin BL Seeley RJ Daniels SR D'Alessio DA The role of energy expenditure in the differential weight loss in obese women on low-fat and low-carbohydrate diets J Clin Endocrinol Metab 2004
Meckling KA O'Sullivan C Saari D Comparison of a low-fat diet to a low-carbohydrate diet on weight loss, body composition, and risk factors for diabetes and cardiovascular disease in free-living, overweight men and women J Clin Endocrinol Metab 2004 89 6 2717 2723 10.1210/jc.2003-031606 15181047
Stern L Iqbal N Seshadri P Chicano KL Daily DA McGrory J Williams M Gracely EJ Samaha FF The effects of low-carbohydrate versus conventional weight loss diets in severely obese adults: one-year follow-up of a randomized trial Ann Intern Med 2004 140 10 778 785 15148064
Yancy WSJ Olsen MK Guyton JR Bakst RP Westman EC A low-carbohydrate, ketogenic diet versus a low-fat diet to treat obesity and hyperlipidemia: a randomized, controlled trial Ann Intern Med 2004 140 10 769 777 15148063
|
15670454
|
PMC1323304
|
CC BY
|
2022-01-24 23:45:17
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A01
|
latin-1
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0119
Editorial
Toward Prevention and Control of Type 2 Diabetes: Challenges at the U.S.-Mexico Border and Beyond1
Bowman Barbara A PhD National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention
Mail Stop K-40, 4770 Buford Highway NE, Atlanta, GA 30341 770-488-5414 [email protected]
Vinicor Frank MD, MPH National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, Ga
1 2005
15 12 2004
2 1 A022005
==== Body
Type 2 diabetes makes a compelling case study for public health action (1). The disease respects no boundaries. It is increasingly common — occurring in both developed and developing countries (2), in men and women, at earlier ages than in past decades, and in persons of every race and ethnic group, with a high prevalence in Hispanic/Latino Americans and in other minority groups, including non-Hispanic blacks, American Indians, Alaska Natives, Asian Americans, and Native Hawaiian and other Pacific Islanders (3). As noted by Martorell (4) and Saldaña(5), family history and genetic factors appear to further increase the risk for type 2 diabetes in Hispanic/Latino Americans. In the United States, the prevalence of diabetes was estimated to be 18.2 million people (6.3% of the population) in 2002 (3), with dramatic increases predicted in the future (6).
The determinants of type 2 diabetes are largely understood. Two of the most important risk factors, obesity and physical inactivity, are modifiable. The natural history involves progression from prediabetes, a condition in which blood glucose metabolism is abnormal (although not yet in the diabetes range), to the development of type 2 diabetes. The rate of progression from prediabetes to type 2 diabetes is between 3% and 10% per year (7). However, progression from prediabetes to diabetes can be prevented or delayed with sustained weight loss and increased physical activity (8,9). The magnitude of the change needed for primary prevention of type 2 diabetes is relatively modest: a 7% to 10% weight loss and sustained moderate physical activity, at least 30 minutes per day (10). Today, the number of adults with prediabetes in the United States is estimated to be at least 41 million (3).
Type 2 diabetes leads to devastating health and economic consequences for individuals, their families, and society. The most serious complications include blindness, kidney disease, lower-limb amputations, and acceleration of coronary heart disease and stroke (3). After type 2 diabetes is diagnosed, treatment requires an increasingly intensive and complex regimen to control glucose, blood pressure, and lipids, in addition to ongoing preventive care for the eyes, kidneys, and feet (11). Health care and complications attributed to diabetes are costly: in 2002, the total cost of diabetes was estimated to be $132 billion, $92 billion of which was spent on direct medical costs and $40 billion of which was spent on indirect costs, including disability, work loss, and premature mortality (12). Clearly, ongoing access to high-quality health care is a paramount concern for preventing complications and death from diabetes. Such care is expensive, and much of the cost of drugs and supplies is not reimbursed, even for those with insurance coverage (13). While it is improving, the quality of clinical care for people with diabetes still falls short of established guidelines (14). Because of continued increases in the prevalence of obesity, the outlook for the future is ominous — the health system will likely be overwhelmed by type 2 diabetes (15).
The population groups at increased risk for diabetes, including Hispanic/Latino Americans, suffer a disproportionate burden of disease, further exacerbated by poverty and lack of access to health care (3,16). What public health responses are likely to be effective in reducing the present and future consequences of type 2 diabetes in population groups, such as people living along the U.S.-Mexico border? And, how long will it take to begin to turn the tide?
As detailed by Cohen et al in the series of articles from the Border Health Strategic Initiative, the solution to type 2 diabetes control must begin in the community (17). Extensive dialogue is a first step in engaging communities and identifying the priorities for community action. The papers by Cohen and associates demonstrate how communities and researchers can — and must — collaborate to assess targets for intervention and develop sustainable solutions to control type 2 diabetes. Insights gained from these interventions also can guide the development of effective community-based approaches for primary prevention of type 2 diabetes. Community-based participatory research and mobilization are critical to create the evidence base for elimination of health disparities, as shown in a recent compendium of papers describing the experience of Racial and Ethnic Approaches to Community Health (REACH) 2010 communities (18).
But having evidence is not enough. Improving the public's health will require rapid translation and dissemination of effective, community-based strategies for diabetes prevention and control and the commitment to sustain and reinforce these interventions (19). As shown by this promising initiative (17), collaboration across and within national and state borders and communities will be essential and must involve the entire community: where people live, work, play, and go to school. Improved clinical care alone will not be sufficient. One strategy now being implemented uses the essential public health services as strategic levers to strengthen the public health response to diabetes (20). Development, implementation, and evaluation of such strategies are needed urgently. We anticipate that publication of the papers by Cohen et al, which describe many challenges and some successes, will inspire readers of Preventing Chronic Disease to share their own lessons learned and promising approaches for public health action to prevent and control type 2 diabetes.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Bowman BA, Vinicor F. Toward prevention and control of type 2 diabetes: challenges at the U.S.-Mexico border and beyond. Prev Chronic Dis [serial online]. 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0119.htm.
==== Refs
1 Vinicor F 17 S1 1994 22 27 Diabetes Care Is diabetes a public health disorder? 8088219
2 Zimmet P Shaw J Alberti KG 2003 20 693 702 Diabetic Med Preventing type 2 diabetes and dysmetabolic syndrome in the real world: a realistic view 12925046
3 Centers for Disease Control and Prevention National diabetes fact sheet: general information and national estimates on diabetes in the United States, 2003 Rev. ed Department of Health and Human Services, Centers for Disease Control and Prevention Atlanta, GA 2004
4 Martorell R Prev Chronic Dis [serial online] Diabetes and Mexicans: why the two are linked 2005 1
5 Rodriguez-Saldaña J Prev Chronic Dis [serial online] Challenges and opportunities in border health 2005 1
6 Boyle JP Honeycutt AA Narayan KM Hoerger TJ Geiss LS Chen H 2001 1936 1940 24 Diabetes Care Projection of diabetes burden through 2050: impact of changing demography and disease prevalence in the U.S. 11679460
7 Unwin N Shaw J Zimmet P Alberti KG 2002 708 723 19 Diabetic Med Impaired glucose tolerance and impaired fasting glycaemia: the current status on definition and intervention 12207806
8 Knowler WC Barrett-Connor E Fowler SE Hamman RF Lachin JM Walker EA The Diabetes Prevention Program Research Group 2002 393 403 346 N Engl J Med Reduction in the incidence of type 2 diabetes with lifestyle modification or metformin 11832527
9 Tuomilehto J Lindstrom J Eriksson JG Valle TT Hamalainen H Ilanne-Parikka P 2001 1343 1350 344 N Engl J Med Prevention of type 2 diabetes mellitus by changes in lifestyle among participants with impaired glucose tolerance 11333990
10 Sherwin RS Anderson RM Buse JB Chin MH Eddy D Fradkin J 27 Suppl1 2004 S47 S54 Diabetes Care Prevention or delay of type 2 diabetes 14693925
11 American Diabetes Association 27 Suppl1 2004 S15 S35 Diabetes Care Standards of medical care in diabetes 14693923
12 Hogan P Dall T Nikolov P American Diabetes Association 2003 26 917 932 Diabetes Care Economic costs of diabetes in the U.S. in 2002 12610059
13 Schroeder S 2001 344 847 852 N Engl J Med Prospects for expanding health insurance coverage 11248165
14 Saaddine JB Engelgau MM Beckles GL Gregg EW Thompson TJ Narayan KM 2002 136 565 574 Ann Int Med A diabetes report card for the United States: quality of care in the 1990s 11955024
15 Alberti G 79 10 2001 907 Bull World Health Organ Noncommunicable disease: tomorrow's pandemics 11693971
16 Engelgau MM Geiss LS Saaddine JB Boyle JP Benjamin SM Gregg EW 140 11 6 1 2004 945 950 Ann Intern Med The evolving diabetes burden in the United States 15172919
17 Cohen SJ Ingram M Prev Chronic Dis [serial online] Border Health Strategic Initiative: overview and introduction to a community-based model for diabetes prevention and control 2005 1
18 14 3Suppl1 2004 S1-1 S1-141 Ethn Dis Community-based interventions to eliminate disparities in health: Lessons learned from the Racial and Ethnic Approaches to Community Health (REACH 2010) Program
19 Glasgow RE Klesges LM Dzewaltowski DA Bull SS Estabrooks P 2004 27 3 12 Ann Behav Med The future of health behavior change research: what is needed to improve translation of research into health promotion practice? 14979858
20 Satterfield DW Murphy D Essien JD Hosey G Stankus M Hoffman P 119 3 2004 311 321 Public Health Rep Using the Essential Public Health Services as strategic leverage to strengthen the public health response to diabetes 15158110
|
15670455
|
PMC1323305
|
CC BY
|
2022-01-27 23:22:37
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A02
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv12_04_0099
Essay
Challenges and Opportunities in Border Health1
Rodríguez-Saldaña Joel MD Research Center, Servicios de Salud de Hidalgo, Avenida México 300, Pachuca Hidalgo 42039 México 011(52)771-71-80770 [email protected]
1 2005
15 12 2004
2 1 A032005
==== Body
Approximately 11.5 million people reside in the 42 counties and 39 Mexican municipalities located along the U.S.-Mexico border, and 86% of those people reside in 14 pairs of sister cities, metropolitan areas divided by the international border (1). Border residents share similar resources and environmental problems: issues of great concern include air quality, water quantity and quality, and animal control. The communities along the border are economically and socially interdependent, with more than 1 million legal northbound crossings every day. The need to establish cooperation between the United States and Mexico for improving health has led to collaborative initiatives between the public and private sectors (1). The principal health problems at the U.S.-Mexico border are characterized by disparities in health systems (2), which result from the lower health standards and socioeconomic conditions of Mexican border communities compared with U.S. border communities.
Health-system disparities produce differences in and barriers to health care access and use(3,4). Documented cases that demand the creation of programs across the U.S.-Mexico border show different rates in the prevalence of infectious disease, including hepatitis A, salmonella, tuberculosis, dengue fever, and Helicobacter pylori infection (5,6). The magnitude and relevance of infectious disease as a major concern along the U.S.-Mexico border have prompted the establishment of binational agreements, such as the U.S.-Mexico Border Infectious Disease Surveillance Project, with the purpose of enhancing the effectiveness of infectious disease prevention (7). On the other hand, populations on both sides of the border share the impact of diseases — such as obesity and diabetes — resulting from similar lifestyle changes. The prevalence rate for diabetes along the U.S.-Mexico border is nearly 50% higher than the rate for the rest of the United States, and Hispanics are more vulnerable to suffering the burden of chronic complications because of genetic, economic, social, behavioral, and psychological factors.
This issue of Preventing Chronic Disease includes an introduction and overview (8) as well as additional articles on the Border Health Strategic Initiative (Border Health ¡SI!), a comprehensive community approach to diabetes prevention and control primarily concentrated in Yuma and Santa Cruz counties in Arizona. Border Health ¡SI! is based on models of community capacity building and community change and was established through a partnership between several border community groups and the University of Arizona. In addition to being comprehensive and community oriented, Border Health ¡SI! was designed to be acceptable to stakeholders, effective in fostering and sustaining change, adaptable to other communities, sustainable after funding, and process and outcome focused.
To reduce the incidence of diabetes among individuals with impaired glucose tolerance, Border Health ¡SI! has emphasized the management of risk factors such as obesity through lifestyle changes (e.g., nutritional counseling, increased physical activity, modest weight loss). The program has also focused on community-based diabetes care provided by a multidisciplinary team that targets patients with diabetes, their families, and their health care providers. Community-health outreach workers called promotores de salud have been instrumental in implementing interventions designed to change personal health risk factors.
The introductory article also describes the formation of community-based coalitions called Special Action Groups (SAGs), whose primary goal is to identify and implement plans for policy and environmental change. Meister et al (9) provide details on how the SAGs in two communities were formed and how they promoted activities to support physical activity and nutrition, and Steinfelt (10) reports on her experience as the community coordinator responsible for orchestrating SAG activities. Other articles in this issue, described below, provide examples of target populations.
Ingram et al report on the effectiveness of a series of diabetes education classes to assist participants in gaining knowledge and skills necessary to be physically active, control diet, monitor blood sugar, take medications, and be aware of complications (11). Promotores de salud play a key role in conducting outreach, participating in patient education, and providing educational support in an overall framework in which individual ability to manage diabetes is not separated from community context and support for diabetes care. Community health centers administered the program and provided a coordinator. Academic partners provided technical assistance and conducted evaluations. The culturally competent curriculum employed a variety of teaching methods to educate participants on how diabetes affects the body. In addition, program staff measured blood glucose, weight, and blood pressure at each of five weekly classes. Improvements in self-management behaviors, HbA1c, random blood glucose, and blood pressure were documented after five weeks. The authors conclude that successful implementation of a program like Border Health ¡SI! includes five essential elements: basic diabetes education, peer outreach and support, integration of diabetes and clinical care, access to medical care and medication, and sustainability.
Teufel-Shone et al (12) describe how the University of Arizona and two community health agencies collaborated to design, pilot, and assess the feasibility of a lay health-outreach, worker-delivered diabetes education program for families. The culturally appropriate program addressed family food choices, physical activity, behavior change, communication, and support behaviors. Seventy-two families participated, and pre- and post-evaluations showed an increase in knowledge of diabetes risk factors and an increase in family efficacy to change food and activity behaviors.
Staten et al report their findings after implementing the School Health Index (SHI) in 13 schools in two counties along the U.S.-Mexico border as part of Border Health ¡SI! between 2000 and 2003 (13). The alarming increase in childhood obesity is a contributing factor to the escalating rate of type 2 diabetes among adolescents. Although the school environment is shown to neglect promotion of physical activity (e.g., by eliminating or not offering physical education classes) and good nutrition (e.g., by selling candy in vending machines), it offers opportunities to combat obesity and diabetes. The SHI is a team-based program launched by the Centers for Disease Control and Prevention in 2000 as a self-assessment and planning tool for health promotion. The SHI enables schools to identify strengths and weaknesses of physical activity and nutrition policies and programs and to develop action plans for improving student health. Border Health ¡SI! supported the hiring and training of an external (i.e., not part of the school system) SHI coordinator in each county who worked with the schools to implement the SHI, develop action plans, and monitor progress. Process and participation varied from school to school, but most schools made at least one immediate change in the school environment to promote student health. Analysis of short-term and intermediate outcomes of the SHI at these schools will be of great additional value.
Staten et al also report on Pasos Adelante, a curriculum designed in cultural context aimed at preventing diabetes, cardiovascular disease, and other chronic diseases in Hispanic populations (14). The 12-week program was facilitated by promotoras de salud in two counties along the Arizona-Sonora, Mexico border. Sessions included physical activity. Walking clubs were established that could continue after the program concluded. Approximately 250 people participated in Pasos Adelante. Analysis of pre- and post-program questionnaires demonstrated a significant increase in moderate to vigorous walking among participants as well as positive changes in nutritional patterns. The success of the Pasos Adelante curriculum shows that a culturally appropriate educational program can motivate people in border communities to adopt healthier lifestyle behaviors.
In a related article on original research, Abarca et al (15) illustrate how community indicators were used to assess nutrition in communities targeted by Border Health ¡SI!. Local grocery store purchases were selected as an indicator, and a structured 26-question interview was developed and administered to grocery store managers. In addition, the investigators gathered data from milk distributors serving these communities. Results showed that food items with a higher fat and higher caloric content were favored. The authors suggest that barriers to acceptance of healthier food items include lack of knowledge concerning healthy foods and their prices. They conclude that more interventions are needed to change dietary patterns, improve overall health, and prevent and control diabetes in these communities.
Schachter et al report their findings on implementing national diabetes guidelines in five border-community health centers (two in Arizona and three in Texas) (16). Each center selected their top four or five indicators of diabetes care and performed baseline audits of medical records in a minimal sample of 12 to 15 charts. Percentage level of compliance for each indicator was compared with the average percentage level of overall diabetes care compliance for each community health center. Priorities varied from clinic to clinic, but the majority of indicators showed improvement. All participating centers expressed interest in improving performance. Only three centers, however, were audited again 24 months later: two maintained or increased improvements, and one lost ground. As reported in other studies(17), translating guidelines into practice is easier said than done: "Between the health care we have and the care we could have lies not just a gap, but a chasm"(18).
Although there is increasing evidence of improvements in diabetes care, not all people with diabetes are experiencing these benefits(19). Addressing the complexities of diabetes management, improving the established systems of care, and recognizing the decisive role of personal, social, and economic factors on diabetes care for each individual with diabetes are the greatest health challenges of our time. The U.S.-Mexico border is a unique example of the interaction of global interdependence: the challenges of providing formal diabetes education in border communities are overwhelming (11). It would be desirable for this interdependence to produce better standards of living and health for all, but evidence confirms that this is not the case(1). The Border Health Strategic Initiative is an illustrative example of a long and successful record of collaborative work, with defined goals, including process and outcome analysis. Resolution of the many challenges that the emerging epidemic rates of diabetes presents at the U.S.-Mexico border will certainly apply to other scenarios of health disparity.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Rodríguez-Saldaña J. Challenges and opportunities in border health. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0099.htm.
==== Refs
1 Homedes N Ugalde A 93 12 2003 2016 2022 Am J Public Health Globalization and health at the United States-Mexico border 14652325
2 Zunker CL Cummins JJ 19 1 2004 13 25 J Cross Cult Gerontol Elderly health disparities on the U.S.-Mexico border 14767175
3 Hunter JB de Zapien JG Denman CA Moncada E Papenfuss M Wallace D 28 5 2003 317 333 J Community Health Healthcare access and utilization among women 40 and older at the U.S.-Mexico border: predictors of a routine check-u 14535598
4 Landeck M Garza C 20 1 2002 3 16 Health Mark Q Utilization of physician health care services in Mexico by U.S. Hispanic border residents 12749595
5 O'Rourke K Goodman KJ Grazioplene M Redlinger T Day RS 158 8 10 2003 15 816 824 Am J Epidemiol Determinants of geographic variation in Helicobacter pylori infection among children on the US-Mexico border 14561672
6 Goodman KJ O'Rourke K Wang C Redlinger T Campos A de la Rosa JM 5 3 2003 99 107 J Immigr Health Helicobacter pylori infection in pregnant women from a U.S.-Mexico border population 14512764
7 Weinberg M Waterman S Lucas CA Falcon VC Morales PK Lopez LA 9 1 2003 97 102 Emerg Infect Dis The U.S.-Mexico Border Infectious Disease Surveillance project: establishing bi-national border surveillance 12533288
8 Cohen SJ Ingram M Prev Chronic Dis [serial online] Border Health Strategic Initiative: overview and introduction to a community-based model for diabetes prevention and control 2005 1
9 Meister JS Guernsey de Zapien J 2005 1 Prev Chronic Dis [serial online] Bringing health policy issues front and center in the community: expanding the role of community health coalitions
10 Steinfelt VE. 2005 1 Prev Chronic Dis [serial online] The Border Health Strategic Initiative from a community perspective
11 Ingram M Gallegos G Elenes J 2005 1 Prev Chronic Dis [serial online] Diabetes is a community issue: the critical elements of a successful outreach and education model in the U.S.-Mexico border
12 Teufel-Shone NI Drummond R Rawiel U 2005 1 Prev Chronic Dis [serial online] Developing and adapting a family-based diabetes program at the U.S.-Mexico border
13 Staten L Teufel-Shone NI Steinfelt VE Ortega N Halverson K Flores C 2005 1 Prev Chronic Dis [serial online] The school health index as an impetus for change
14 Staten LK Scheu LL Bronson D Peña V Elenes J 2005 1 Prev Chronic Dis [serial online] Pasos Adelante: the effectiveness of a community-based chronic disease prevention program
15 Abarca J Ramachandran S 2005 1 Prev Chronic Dis [serial online] Using community indicators to assess nutrition in Arizona-Mexico border communities
16 Schachter KA Cohen SJ 2005 1 Prev Chronic Dis [serial online] From theory to practice: challenges to implementing national diabetes guidelines with five community health centers on the border
17 Larme AC Pugh JA 24 10 2001 1728 1733 Diabetes Care Evidence-based guidelines meet the real world: the case of diabetes care 11574433
18 Institute of Medicine Crossing the quality chasm: a new health system for the 21st century National Academies Press Washington (DC) 2001
19 Vinicor F 2004 22 94 96 Clinical Diabetes The future of diabetes: what is there besides new medicines?
|
15670456
|
PMC1323306
|
CC BY
|
2022-01-27 23:20:52
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A03
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0100
Essay
Diabetes and Mexicans: Why the Two Are Linked1
Martorell Reynaldo PhD Department of Global Health, The Rollins School of Public Health of Emory University
1518 Clifton Rd, Room 754, Atlanta, GA 30322 404-727-9854 [email protected]
1 2005
15 12 2004
2 1 A042005
==== Body
The Past
Obesity and diabetes were probably rare before the advent of agriculture. Our ancestors, hunters and gatherers for millennia, had varied but unpredictable diets. Studies of hunter-gatherers of the 20th century suggest that animal sources dominated our ancient food basket, with plants (fruits, vegetables, and nuts) providing only 20% to 40% of total energy (1). Modern and presumably ancient hunter-gatherer populations, despite a high-fat, high-protein diet, were free of the signs and symptoms of noncommunicable diseases — a paradox. Perhaps energy needs were not always met, thus keeping body sizes in check; also, the relative lack of salt and simple carbohydrates, a mix of saturated and good fats, plenty of fiber, abundant micronutrients, a vigorous and active life, and less stress than we now endure may explain this finding. With the food supply uncertain, one would expect individuals with "thrifty" genotypes — genotypes that increase the ability to turn food to fat — to have a survival edge.
Agriculture brought a more predictable food supply but less variety. Crops failed from time to time, bringing on famines when stores of grain were depleted, but over time, agriculture allowed for increasingly larger populations, with thrifty genotypes thriving as before. Super foods — such as corn in Mesoamerica, the substance from which the Mayan gods in their fourth attempt were finally able to make man, according to the Popul Vuh, the sacred book of the Maya — came to provide as much as 80% or more of energy needs. Crowding brought new types of infections, which along with limited diets gave rise to the nutritional deficiencies that have plagued humankind in recent millennia. Agriculture fostered the development of highly stratified societies, and it became possible for a few to lead a life of luxury. Until the 20th century, fatness was a marker of wealth.
The Present
Most of my professional career has been devoted to the study of hunger and malnutrition in developing countries. As rates of child malnutrition decline in Latin America and in other developing countries, the prevalence of obesity is increasing rapidly, and I, like many of my colleagues, have begun to study both ends of the spectrum — namely, deficiency and excess (2).
Economic development and urbanization are the engines of the "nutrition transition" (3). Pathways include increased food security, the availability of cheap sources of fat in the form of vegetable oils, more eating away from home, the less arduous nature of modern jobs, and increases in sedentary recreation (notably television). These pathways have transformed dietary and physical activity patterns and, as a result, tipped the balance in favor of obesity (Figure 1).
Figure 1 Possible causes of the nutrition transition and the emergence of obesity in developing countries. Adapted from Martorell and Stein, 2001 (2), and Popkin, 1994 (3).
Flow chart diagramming the paragraph above.
Some populations may be more susceptible to obesity (e.g., Pacific Islanders, Native Americans) because of thrifty genotypes, as proposed by the geneticist Neel some years ago (4). Thrifty phenotypes may also increase susceptibility to obesity; some evidence suggests that poor intrauterine and infant nutrition may also "program" individuals to be metabolically thrifty, and if later times bring a life of abundance, these individuals will be at risk for developing chronic diseases such as diabetes (5).
The "Supersizing" of the Mexican People
Mexico is a country far along the nutrition transition. The Mexican National Nutrition Survey 1999 showed that obesity (Body Mass Index [BMI] ≥30) among women aged 18 to 49 increased from 9% in 1988 to 24% in 1999 (6). If we add overweight (BMI = 25.0–29.9) to the mix, the percentage of overweight or obese women increased from 33% to 59% in just one decade. The 1999 survey also showed that the prevalence of stunting (low height-for-age, indicative of child undernutrition) among preschool children in the indigenous rural south of Mexico was 42%, as high as in many sub-Saharan African countries. Yet the problem of obesity grew alarmingly among all sectors of society. All socioeconomic groups, rural as well as urban areas, and all regions of Mexico, including the impoverished South, showed equally dramatic increases (Figure 2). Obesity and chronic diseases in Mexico can no longer be dismissed as problems of the rich. However, poor Mexicans have a double burden: child undernutrition in addition to obesity. As the nutrition transition unfolds even further, as it has in Chile, obesity becomes more common among the poor, as it is in the United States.
Figure 2 Levels of overweight (BMI = 25.0–29.9) and obesity (BMI ≥30.0) in 1988 and 1999 in women aged 18 to 49 in Mexico, by region. BMI indicates Body Mass Index. Data from Rivera et al, 2001 (6).
Bar graph North Center México City South
1988 1999 1988 1999 1988 1999 1988 1999
Overweight 26.3 34.0 21.0 36.4 25.6 37.7 22.3 34.5
Obesity 11.8 31.3 8.1 22.2 9.2 21.4 8.2 20.8
Mexican Americans are one of the fattest groups in what is one of the fattest nations on earth. Three out of four Mexican American adults (aged >20 years) were either overweight or obese at the end of the 20th century (7). Plentiful and unhealthy diets, many hours of television watching, and a reluctance to exercise are some of the factors blamed. For example, a study of Mexican children along the Mexico–U.S. border showed low intake of fruits and vegetables and excessive consumption of soft drinks and high-fat snacks (8).
Obesity is an easy, visible marker of the worldwide pandemic of noncommunicable diseases for which considerable data from around the world are available (2). Obesity is also a major risk factor for type 2 diabetes, and where obesity is rising we can expect diabetes to follow (9).
The Type 2 Diabetes Pandemic
Diabetes is a growing problem worldwide. The prevalence of diabetes in adults (aged >20 years) is projected to increase in developed countries from 6.0% in 1995 to 7.6% by 2025 (10). Diabetes in developing countries will also increase from 3.3% to 4.9%, and because of initial population sizes and growth, the increase in the number of people with diabetes will come disproportionately from the developing world. The number of individuals with diabetes will rise from 51 million to 72 million in developed countries, but the number will rise from 84 million to 228 million in developing countries. The three nations with the greatest numbers of individuals with diabetes in 1995 were India (19.4 million), China (16.0 million), and the United States (13.9 million). In 2025, the rankings will be unchanged, but the absolute number will increase dramatically in India (to 57.2 million) and China (to 37.6 million) and less so in the United States (to 21.9 million). Mexico, which was ninth in the world in 1995 (3.8 million), will rise to seventh place by 2025 (11.7 million).
Diabetes is a serious public health problem among Mexicans and Mexican Americans. Diabetes was found in 8.1% of Mexican adults in 2000 (11) compared with 13.1% and 14.5% of Mexican American men and women in 1988–94 (12). In the United States, adults of Mexican origin, particularly men, had higher rates of prevalence of diabetes than non-Hispanic whites or blacks, as well as a greater degree of impaired fasting glucose (Figure 3). The prevalence of diabetes in the United States is rising rapidly. The prevalence of diabetes increased from 8.9% in 1976–1980 to 12.3% in 1988–94 among adults aged 40 to 74 (12). Mexican Americans, the largest Hispanic/Latino subgroup in the United States, are more than twice as likely to have diabetes as non-Hispanic whites of similar age (13).
Figure 3 Age-standardized prevalence of diagnosed and undiagnosed diabetes and impaired fasting glucose in the U.S. population aged ≥20 years by sex and ethnic group, based on the Third National Health and Nutrition Examination Survey (NHANES III). Data from Harris et al, 1998 (12).
Bar graph Men Women
Non-Hispanic white Non-Hispanic black Mexican Americans Non-Hispanic white Non-Hispanic black Mexican American
Diagnosed diabetes 5.2 7.3 7.7 4.5 9.1 10.9
Undiagnosed diabetes 2.9 2.7 5.4 2.0 4.5 3.6
Impaired fasting glucose 8.9 8.9 11.6 4.6 6.4 6.3
Born in Central America, I share a similar ancestry with Mexicans (Spanish and Amerindian). Not surprisingly, diabetes runs in my family. Some statistics should scare me. The lifetime risk of developing diabetes for U.S. individuals born in 2002 is about one in three for the general population, but about one in two for the Hispanic population (14).
Ancestry and Prenatal Exposure
Lifestyle characteristics are primarily responsible for the high levels of obesity and diabetes among Mexicans, but other considerations are also important. The San Antonio Heart Study began in 1979 and is a population-based study of diabetes and cardiovascular disease in Mexican Americans and non-Hispanic whites in San Antonio, Texas (9). One of the interesting findings of the study is that the degree of Native American ancestry is a major risk factor for diabetes, presumably because of inherited thrifty genes (15).
The role of intergenerational mechanisms, specifically the risk of developing diabetes in adulthood as a result of prenatal exposure to diabetes, has become clear from studies of Pima Indians in Arizona (Figure 4). The prevalence of diabetes among adults aged 20 to 24 was found to be 1.4% if the mother was free of diabetes, 8.6% if she was prediabetic (developed diabetes after delivery), and 45.5% if she had gestational diabetes (16). Follow-up studies over three decades reveal a steady rise in diabetes in Pima children and adolescents. From 1967–76 to 1987–96, the prevalence of diabetes in girls aged 10 to 14 years increased from 0.72% to 2.88%. In girls aged 15 to 19 years, the prevalence increased from 2.73% to 5.31% during the same period (17). The percentage of youths (aged 10 to 19 years) who were exposed to gestational diabetes increased during this period (Figure 5). In 1967–76, 2.1% of youths were exposed to gestational diabetes; by 1987–96, exposure had almost quadrupled to 7.5% of pregnancies. The fraction of diabetes attributable to gestational diabetes also rose markedly in youths aged 10 to 19 so that by 1987–96, more than one third of cases of diabetes could be attributed to gestational diabetes. Also, more than 70% of persons with prenatal exposure developed type 2 diabetes at 25 to 34 years of age (18). Clearly, the hyperglycemic intrauterine environment brought on by gestational diabetes is an important determinant of early-onset type 2 diabetes that is above any genetically transmitted susceptibility and is another example of fetal programming (19). An additional consequence is that 50% of women with gestational diabetes will themselves develop diabetes within five years (20). The concern about gestational diabetes is not limited to the Pima population. The incidence of gestational diabetes increased from 4.9% in 1990 to 7.1% in 2000 in California, where Asian and Hispanic women had higher incidences than whites and African Americans (20).
Figure 4 Prevalence of type 2 diabetes among Pima Indian adults, Arizona, aged 20 to 24, by diabetes status of the mother during pregnancy. A prediabetic mother is one who develops diabetes after delivery. Data from Pettitt et al, 1988 (16).
Bar chart Nondiabetic mother Prediabetic mother Diabetic mother
Percentage rate of diabetes among adults aged 20 to 24 1.4 8.6 45.5
Figure 5 Exposure to gestational diabetes (GD) and fraction of diabetes attributed to GD among cohorts of Pima Indian adults, Arizona, aged 10 to 19 years (n = 6902). Data from Dabelea et al, 1998 (17).
Bar chart 1967–76 1977–86 1987–96
Exposure to GD 2.1 4.0 7.5
Attributable fraction 18.1 23.7 35.4
Gestational diabetes is adding fuel to an already raging epidemic of diabetes. The intergenerational component operates through women and begins with the interaction of genetic susceptibility and unhealthy lifestyle practices that precipitate obesity in girls and women of reproductive age, which in turn increases the risk of diabetes prior to or during pregnancy. The percentage of women exposed to diabetes in their intrauterine life then increases in each subsequent generation, driving rates of diabetes in the general population higher and higher with each generation. This scenario is already unfolding in the Mexican populations of North America and deserves serious study.
Where Do We Go From Here?
The costs of diabetes in the United States were estimated at $132 billion for 2002 (21). Meeting the demand for public health care services caused by diabetes will alone cost Mexico $318 million in 2005, 26% more than in 2003 (22). While the monetary costs are staggering, the suffering and disability among those afflicted with diabetes and their families are incalculable.
We need to confront the diabetes pandemic with urgency. Efficacy studies show that lifestyle changes can effectively reduce the incidence of diabetes in persons at high risk (23). We need effective programs that promote healthy lifestyles and make screening and sound case management widely available. We also need to devote significant resources to developing new drugs and therapies. Combating obesity and inactivity must become a national priority. Preventive actions must be undertaken along a broad front, impacting behavior as well as the physical environment — from how we design our cities to promote physical activity to what agriculture and food policies we support to foster a healthier food basket. We need to promote aggressively a love of physical activity and healthy diets, particularly among our children. We need flexible programs that can fit local settings and our diversity of cultures, including the mosaic of Hispanic groups in the United States. Mexico, with far fewer resources, must do all of the above while combating yesterday's unresolved problems of undernutrition. The future will be grim only if we let it become so.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Martorell R. Diabetes and Mexicans: why the two are linked. Prev Chronic Dis [serial online]. Available from URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0100.htm.
==== Refs
1 Cordain L Eaton SB Miller JB Mann N Hill K 56 Suppl 1 2002 S42 S52 Eur J Clin Nutr The paradoxical nature of hunter-gatherer diets: meat-based, yet non-atherogenic 11965522
2 Martorell R Stein AD Bowman BA Russel RM 2001 8th edition 665 685 Present knowledge in nutrition The emergence of diet-related chronic diseases in developing countries ILSI Press Washington (DC)
3 Popkin BM 1994 52 285 298 Nutr Rev The nutrition transition in low-income countries: an emerging crisis 7984344
4 Diamond J 423 5 6 2003 599 602 Nature The double puzzle of diabetes 12789325
5 Hales CN Barber DJP 2001 60 5 20 Br Med Bull The thrifty phenotype hypothesis 11809615
6 Rivera DJ Shamah LV Villalpando HS González de Cossío T Hernández PB Sepúlveda J 2001 Encuesta Nacional de Nutrición 1999. Estado nutricio de niños y mujeres en México Instituto Nacional de Salud Pública Cuernavaca, Morelos, México
7 Flegal KM Carroll MD Ogden CL Johnson CL 2002 288 1723 1727 JAMA Prevalence and trends in obesity among US adults, 1999-2000 12365955
8 Jiménez-Cruz A Bacardí-Gascón M Jones EG 33 2002 74 80 Arch Med Res Consumption of fruits, vegetables, soft drinks, and high-fat-containing snacks among Mexican children on the Mexico-U.S. border 11825635
9 Haffner SM 83 Suppl 1 2000 S67 S70 Br J Nutr Obesity and the metabolic syndrome; the San Antonio Heart Study 10889794
10 King H Aubert RE Herman WH 21 9 1998 1414 1431 Diabetes Care Global burden of diabetes, 1995-2025: prevalence, numerical estimates, and projections 9727886
11 Aguilar-Salinas CA Velazquez Monroy O Gómez-Pérez FJ Gonzalez Chávez AG Esqueda AL Molina Cuevas V 26 7 2003 2021 2026 Diabetes Care Characteristics of patients with type 2 diabetes in México: results from a large population-based nationwide survey 12832306
12 Harris MI Flegal KM Cowie CC Eberhardt MS Goldstein DE Little RR 21 4 1998 518 524 Diabetes Care Prevalence of diabetes, impaired fasting glucose, and impaired glucose tolerance in U.S. adults. The third National Health and Nutrition Examination Survey, 1988-1994 9571335
13 Centers for Disease Control and Prevention 2004 Centers for Disease Control and Prevention National diabetes fact sheet, United States, 2003 Atlanta (GA)
14 Narayan KM Boyle JP Thompson TJ Sorensen SW Williamson DF 290 14 2003 1884 1890 JAMA Lifetime risk for diabetes mellitus in the United States 14532317
15 Gardner LI Stern MP Haffner SM Gaskill SP Hazuda HP Relethford JH 1984 33 86 92 Diabetes Prevalence of diabetes in Mexican Americans. Relationship to percent of gene pool derived from Native American sources 6690348
16 Pettitt DJ Aleck KA Baird HR Carraher MJ Bennett PH Knowler WC 1988 37 622 628 Diabetes Congenital susceptibility to NIDDM. Role of intrauterine environment 3360218
17 Dabelea D Hanson RL Bennett PH Roumain J Knowler WC Pettitt DJ 1998 41 904 910 Diabetologia Increasing prevalence of Type II diabetes in American Indian children 9726592
18 Dabelea D Knowler WC Pettitt DJ 9 1 2000 83 88 J Matern Fetal Med Effect of diabetes in pregnancy on offspring: follow-up research in the Pima Indians 10757442
19 Dabelea D Pettitt DJ 14 8 2001 1085 1091 J Pediatr Endocrinol Metab Intrauterine diabetic environment confers risks for type 2 diabetes mellitus and obesity in the offspring, in addition to genetic susceptibility 11592564
20 Ferrara A Kahn HS Quesenberry CP Riley C Hedderson MM 103 3 2004 526 533 Obstetrics & Gynecology An increase in the incidence of gestational diabetes mellitus: Northern California, 1991-2000 14990417
21 Hogan P Dall T Nikolov P 26 3 2003 917 932 Diabetes Care American Diabetes Association. Economic costs of diabetes in the US in 2002 12610059
22 Arredondo A Zuniga A 27 1 2004 104 109 Diabetes Care Economic consequences of epidemiological changes in diabetes in middle-income countries. The Mexican case 14693974
23 Knowler WC Barrett-Connor E Fowler SE Hamman RF Lachin JM Walker EA 346 6 2002 393 403 N Engl J Med Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin 11832527
|
15670457
|
PMC1323307
|
CC BY
|
2022-01-31 23:21:01
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A04
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0073
Original Research
PEER REVIEWEDNew Mexico’s Capacity for Increasing the Prevalence of Colorectal Cancer Screening With Screening Colonoscopies
Hoffman Richard M MD, MPH New Mexico VA Health Care System
111GIM, 1501 San Pedro SE, Albuquerque, NM 87108 [email protected]
505-256-2727
Stone S. Noell MPH University of New Mexico Cancer Center, Albuquerque, NM
Herman Carla MD, MPH University of New Mexico Cancer Center, Albuquerque, NM
Moore Jung Ann MEd New Mexico Medical Society, Albuquerque, NM
Cotner Jane MS University of New Mexico Cancer Center, Albuquerque, NM
Espey David MD Indian Health Services, Centers for Disease Control and Prevention, Albuquerque, NM
Kozoll Richard MD New Mexico Medical Society, Cuba, NM
Gavin Michael W MD Lovelace Sandia Health System, Albuquerque, NM
1 2005
15 12 2004
2 1 A072005
Introduction
Colorectal cancer screening rates are low throughout the United States. Colonoscopy has been recommended as a cost-effective strategy for colorectal cancer screening and prevention. We evaluated New Mexico's capacity to increase the prevalence of colorectal cancer screening using colonoscopy.
Methods
We identified New Mexican gastroenterologists from state licensing data and from endoscopic manufacturers. We surveyed gastroenterologists on their weekly number of colonoscopies, capacity for additional screening colonoscopies, and barriers to increasing capacity. We used census data, published data on the yield of screening colonoscopy, and professional society guidelines for cancer/polyp surveillance to estimate the additional colonoscopies required to increase the state's prevalence of endoscopic screening.
Results
Forty gastroenterologists, representing all 11 group practices in the state, and nine of 12 solo practitioners responded. They estimated that their weekly procedure capacity could be increased by 41%, from 832 to 1174 colonoscopies. We estimated an annual capacity increase of 14,880 procedures, which could increase the prevalence of endoscopic colorectal cancer screening from the current 35% to about 50% over five years. Lack of support staff, space, and physicians were barriers to increasing screening.
Conclusion
Implementing a screening colonoscopy strategy could achieve the goal of a higher level of colorectal screening. However, achieving more universal screening would require additional testing modalities.
==== Body
Introduction
Colorectal cancer is the third most frequently diagnosed cancer in New Mexico and the second leading cause of cancer death (1). Randomized controlled trials of fecal occult blood testing (FOBT) have shown that screening reduces the incidence and mortality of colorectal cancer (2-4). Flexible sigmoidoscopy has also been shown to reduce colorectal cancer mortality in well-designed case-control studies (5,6). Professional organizations have identified colorectal cancer screening as an effective, high-priority intervention (7-10). Acceptable modalities include FOBT, flexible sigmoidoscopy, colonoscopy, and double-contrast barium enema.
Despite those supportive practice guidelines, colorectal cancer screening rates remain low. National data show that just over 50% of adults aged 50 years and older are considered to be appropriately screened for colorectal cancer with either a FOBT within one year or an endoscopic procedure within 10 years (11,12). In New Mexico, 2001 survey data from the Centers for Disease Control and Prevention's (CDC's) Behavioral Risk Factor Surveillance System (BRFSS) reported that only 23.2% of adults aged 50 years and older had undergone FOBT testing in the previous two years and that 34.5% had undergone a flexible sigmoidoscopy or colonoscopy in the previous five years (13). The BRFSS survey did not obtain information about radiographic screening tests. Overall, only 48% of the adult population was considered currently screened (analysis by the New Mexico BRFSS Unit, July 2003); this likely is an overestimate given the limited concordance of the BRFSS colorectal cancer screening questions with medical records (14) and the potential selection bias introduced by the telephone survey design.
Efforts to improve screening rates have included celebrity endorsements by Katie Couric (15), the CDC's Screen for Life campaign (16), the American Cancer Society's "Polyp Man" public service announcements (17), and President Clinton's 2000 declaration that March would henceforth be National Colorectal Cancer Awareness Month (18). Medicare began reimbursing for colorectal cancer screening with FOBT and flexible sigmoidoscopy in 1998 and has reimbursed screening colonoscopy at 10-year intervals for average-risk adults since July 2001 (19). The National Center for Quality Assurance has established a new Health Plan Employer Data and Information Set measure of colorectal cancer screening performance standards for health care plans beginning in 2004 (20,21).
Although there is no direct evidence for its screening efficacy, colonoscopy is the most accurate diagnostic test and offers the potential to remove premalignant growths. Winawer and colleagues estimated that colonoscopy could reduce the incidence of colorectal cancer by a range of 76% to 90% (22). Economic analyses have also found that colonoscopy is a cost-effective screening strategy for colorectal cancer (23-25). The American College of Gastroenterology practice guidelines recommend colonoscopy to be the first screening option (26). However, experts have questioned the feasibility of increasing screening through colonoscopy because the number of colonoscopists and infrastructure needed to screen the population may be inadequate (9). We conducted a survey of New Mexican gastroenterologists to determine the feasibility of implementing a colonoscopic screening strategy to improve statewide screening rates.
Methods
The Colorectal Cancer Screening Working Group of the Clinical Prevention Initiative (CPI) evaluated screening capacity by conducting a mailed survey of endoscopists in New Mexico. The CPI membership, composed of public health and health care professionals, is supported by the New Mexico Department of Health and the New Mexico Medical Society to promote more effective delivery of practice-based preventive services throughout the state of New Mexico.
Subjects
We identified endoscopists in New Mexico by using data from the Board of Medical Examiners, contacting manufacturers of endoscopic equipment, and obtaining the membership lists of a statewide gastroenterology journal club, the New Mexico Medical Society, and the American Medical Association. Eligible subjects for this analysis were gastroenterologists actively practicing in New Mexico, which included 40 gastroenterologists practicing in one of the 11 group practices and 12 solo practitioners.
Survey
The CPI colorectal cancer group developed a brief survey to obtain information about endoscopic capacity, including colonoscopies and flexible sigmoidoscopies (Table 1). Questions were based on literature review, the BRFSS, and the clinical experience of the CPI colorectal cancer group, which included two gastroenterologists and two internists who performed sigmoidoscopy. Revisions were based on pilot testing the survey with clinical colleagues and other members of the CPI. The survey was conducted between October and December 2001. Subjects were mailed a letter introducing the survey and asking for their participation. The survey was printed on a postcard with a return address and postage. For nonrespondents, we followed up with telephone calls and repeat mailings two weeks after the initial contact.
Statistical analysis
We used simple, descriptive nonparametric statistics to estimate the weekly median number of procedures performed by endoscopists in group practice and solo practice and the estimated weekly potential increase in capacity.
Endoscopic capacity. We determined the number of additional screening colonoscopies that could be performed using survey responses. We averaged responses when multiple members of a group practice completed the survey and provided different estimates for the weekly number of baseline and additional procedures performed by the practice. We imputed the weekly number of baseline and additional colonoscopies for the solo-practitioner nonrespondents using data from the responding solo practitioners. For the annual number of colonoscopies, we assumed that endoscopists performed procedures for 40 weeks. We performed similar estimates for the number of flexible sigmoidoscopies.
Volume of colonoscopies. We modeled the number of procedures required for a statewide screening colonoscopy strategy. To identify the number of subjects potentially eligible for colonoscopic screening, we used data from the 2000 United States Census for New Mexico that reported 468,000 resident adults aged 50 to 85 (27). Based on the census data, we evaluated the additional number of screening colonoscopies required to increase the prevalence of current screening by 5% (23,400 additional people being screened), 10% (46,800), 15% (70,200), 20% (93,600), and 25% (117,000) during a five-year period. We assumed that the additional screening procedures would be performed in equal numbers during the five-year period. We then modeled the number of surveillance procedures that would be required following the initial screening colonoscopy. We used clinical data on the yield of colorectal cancers and adenomatous polyps from a recent large Department of Veterans Affairs (VA) colonoscopic screening trial (28) and consensus guidelines for the timing of surveillance procedures (10).
Colorectal cancer detection level: 1%
Adenomatous polyp detection level: 37%
Advanced (villous, dysplastic, >1 cm, >2 polyps) polyp level: 15%
Surveillance following colorectal cancer detection: 6 months and 3 years
Surveillance following 1–2 adenomatous polyps <10 mm: 5 years
Surveillance following advanced polyp: 3 years
We assumed that half of the cancers diagnosed in the fifth year would have a six-month surveillance colonoscopy that same year. The colonoscopic screening trial had a higher proportion of subjects with positive family history of colorectal cancer than the general population and may have overestimated the yield of screening. Results from an employee-health colonoscopic screening program did show a lower yield than the VA study (29). Therefore, we performed a sensitivity analysis by reducing the expected rates of detected colorectal cancers and adenomatous polyps by approximately 50%.
We entered survey data into a Microsoft Access (Microsoft Corporation, Seattle, Wash) database. We performed statistical analyses with SAS (SAS Institute, Inc, Cary, NC) (30).
Results
We received procedure information from nine of 12 solo practitioners and all 11 group practices, representing 40 endoscopists (two to eight practitioners per group). Physicians and practices were based in 12 different counties. Ten of 11 group practices and six of 12 solo practitioners were located in urban areas, defined by the Census Bureau as having population densities >1000 per square mile (31). Table 2 shows the numbers of procedures currently being performed weekly and the weekly capacity for additional procedures, which were stratified by type of practice. Overall, gastroenterologists reported performing 832 colonoscopies a week; they estimated being able to increase their capacity by an additional 342 (41%) procedures each week.
Assuming a 40-week work year, each endoscopist in group practice could perform an estimated 252 additional colonoscopies every year and solo practitioners could perform an estimated 400 additional colonoscopies. Statewide, endoscopists could perform an estimated 13,680 additional colonoscopic procedures each year. If the nonresponding solo practitioners performed similarly to those completing the survey, the estimated annual additional capacity for colonoscopy would be 14,880 procedures.
We show the estimated number of additional colonoscopies required to increase screening prevalence by 5%, 10%, 15%, 20%, and 25% during a five-year period in Table 3. The total number of procedures includes screening procedures based on the 2000 New Mexico census and surveillance procedures based on the yield of cancer and adenomatous polyps detected with screening. The second column of numbers reflects the yield of advanced neoplasia based on the VA study data from Lieberman and colleagues. The third column is a sensitivity analysis showing the estimated number of colonoscopies if the cancer yield was 0.5% and the overall yield of adenomatous polyps was 20%. If all patients with adenomatous polyps underwent colonoscopic surveillance at three years (rather than just patients with advanced neoplasia), the annual number of procedures would be increased by about 5%. Overall, a screening colonoscopy strategy could increase the prevalence of current colorectal cancer screening by about 15%.
Although our analyses focused on colonoscopies, we also obtained information on flexible sigmoidoscopy. All but one of the group practices performed flexible sigmoidoscopies, but only five of the solo practitioners performed them. Overall, however, only 165 procedures were performed weekly; respondents estimated that they could perform an additional 188 procedures.
The barriers to performing additional endoscopic tests are shown in Table 4. Only one group practice reported no barriers to performing additional procedures, and four solo practitioners reported no barriers. Lack of support staff, space (for procedures and/or recovery room), and physicians were the most frequently cited problems for the group and solo practices.
Discussion
New Mexico gastroenterologists responding to our survey estimated having the capacity to increase their weekly number of colonoscopies by about 41%, from 832 to 1174. This substantial increase could raise the prevalence of current endoscopic screening by approximately 15% within five years. The most recent BRFSS data report that 35% of New Mexican adults are currently screened by endoscopy; thus, the increased endoscopic capacity would be just sufficient to achieve 50% colorectal cancer screening. However, this level of screening would still be far short of the 70% to 90% screening reported for mammography, Papanicolaou (Pap) smears, and prostate-specific antigen (PSA) tests (32,33). Additional recommended screening modalities, including FOBT, flexible sigmoidoscopy, and radiological studies would be needed to achieve a higher level of screening (7,8).
Rex and Lieberman modeled a strategy for implementing colonoscopy as the preferred screening procedure in the United States (9). Based on a 10-year screening interval and assuming that 10% of the adult population aged 50 to 70 would be screened every year, they estimated an annual need for 7.7 million colonoscopies. After reducing this number for patients with significant comorbidities, noncompliance, and current screening, they estimated that approximately 2.56 million additional colonoscopies would need to be performed. Based on a government report that 4.4 million colonoscopies were performed in 1999, Rex and Lieberman concluded that implementing screening colonoscopy would require a 58% increase in capacity. This figure may be an underestimation because they modeled screening only until age 70. Given that the incidence of colorectal cancer increases steadily with age (34) and that screening could appropriately be offered until age 80 (35), the actual number of additional colonoscopies could be quite higher.
Even if Rex and Lieberman correctly estimated the number of additional procedures to fully implement screening colonoscopy, the demand in New Mexico would likely exceed the capacity of the state's endoscopists — despite their already high level of productivity. Endoscopists in New Mexico reported performing about 16 to 20 colonoscopies weekly, which compares quite favorably with data obtained from the National Cancer Institute's (NCI's) nationwide Survey of Colorectal Cancer Screening Practices. The 346 gastroenterologists responding to the survey, conducted between November 1999 and April 2000, performed an average of only 31.7 colonoscopies monthly, including 12.4 for screening (36).
Rex and Lieberman acknowledged that increasing the level of colonoscopies would be challenging (9). One of their solutions was for gastroenterologists to perform 50% fewer flexible sigmoidoscopies to make time to perform colonoscopies (9). They cited Medicare data showing that 543,502 flexible sigmoidoscopies were performed in 2000. The nationwide NCI survey estimated that gastroenterologists performed 25% of sigmoidoscopies, which suggests that nearly 70,000 fewer procedures could be performed in just the Medicare population alone (36). However, this survey indicated that sigmoidoscopies comprised about 30% of the colorectal endoscopic procedures performed by gastroenterologists. Our data showed that sigmoidoscopy comprised only 16% of the lower endoscopic procedures performed by gastroenterologists in New Mexico, implying practice patterns had already changed substantially. Further reductions in performing sigmoidoscopy may not be feasible, especially because many of the sigmoidoscopic procedures are diagnostic.
Another strategy for implementing screening colonoscopy would be to increase the number of procedures performed by other medical providers. The NCI survey reported that general surgeons performed 30% of colonoscopies (36). However, on average, the 251 general surgeons performed fewer than eight colonoscopies monthly, including about three screening colonoscopies. While nearly half of the gastroenterologists performed at least 10 screening colonoscopies monthly, only 6% of general surgeons reached this level. The NCI survey also obtained data on colorectal cancer screening practices by primary care physicians (37). Among the 1235 respondents, fewer than 5% of primary care providers reported performing colonoscopy, and most of them performed fewer than five procedures monthly. Although 29% of primary care respondents performed sigmoidoscopy, fewer than 20% performed more than 10 procedures monthly.
Increasing screening colonoscopy by having general surgeons and primary care physicians perform these procedures does not seem to be a feasible strategy for New Mexico. When we conducted our survey, endoscopic equipment manufacturers provided us information on all practices that had purchased equipment for performing colorectal procedures. In addition to gastroenterologists, we also identified surgeons and primary care physicians as owners of endoscopic equipment. Three of the eight colorectal cancer surgeons in the state identified as performing colonoscopies responded to the survey; they were performing 18 colonoscopies weekly and estimated that they could increase their capacity by 12 weekly. None of the 28 primary care providers who performed endoscopy reported performing colonoscopy. Only six primary care endoscopists reported performing five or more (maximum eight) flexible sigmoidoscopies weekly; the majority performed less than two. Another problem with relying on nongastroenterologists to perform endoscopy is that their low procedure volume may be associated with diminished proficiency (38).
Rex and Lieberman further noted that increasing capacity for screening colonoscopy would require more efficiency in endoscopy suites (9). Our respondents consistently reported that limited space and support staff were barriers to performing more procedures. Our respondents also reported that having more physicians would help improve capacity. Strategies to increase the number of gastroenterologists would likely target a training program; New Mexico does have a university gastroenterology fellowship program. However, Rex and Lieberman questioned the wisdom of training more endoscopists because accurate, cost-effective, noninvasive tests — such as virtual colonoscopy or fecal DNA assays — are likely to be used increasingly, thus reducing the need for screening colonoscopies (9). Another barrier facing New Mexico is a relatively impoverished population with a high proportion of uninsured adults (39); financial incentives may also be necessary to attract and retain specialists. Although there is little data supporting the practice, nonphysicians could also be trained to perform colonoscopy (40).
Our study had some important limitations. We were generally unable to validate the self-reported weekly number of procedures performed by each practice or solo practitioner. However, one group of three gastroenterologists, who estimated that they annually performed 3000 procedures, also reviewed their billing records for the previous year. These data showed that they had overestimated their current capacity by 10% — they actually performed only 2760 procedures. The estimated increased capacity also depends upon the respondents being able to accurately assess the practice's ability to perform additional tests, which could not be validated. However, if other practices similarly overestimated their current capacity, then the estimates for the absolute number of additional procedures could also be inflated.
Another potential limitation was that we used a simplified model. On the demand side, we assumed that patients would be compliant with surveillance-testing recommendations and that the population would be stable. On the supply side, we assumed that the number of gastroenterologists in the state would be stable. We also assumed that the supply of endoscopists would be matched with patients needing procedures. However, New Mexico has problems retaining specialists (41), and even having a sufficient number of gastroenterologists may not ensure comprehensive screening coverage. In New Mexico, access to care may be severely limited by geographic distance. New Mexico is a large, mostly rural state; almost all of the gastroenterologists practice in urban areas. Nonetheless, our intention was not to precisely estimate screening capacity but rather to provide a general assessment of the feasibility of implementing screening colonoscopy, including identifying provider barriers.
We conclude that New Mexico has the colonoscopic capacity to substantially increase the prevalence of adults with current colorectal cancer screening. The state could probably achieve a level of 50% current endoscopic screening by colonoscopy alone. However, New Mexico lacks the capacity to implement a fully comprehensive screening colonoscopy strategy. Efforts to achieve more universal screening would also require additional modalities such as FOBT, flexible sigmoidoscopy, and barium enema in addition to health care policies requiring screening coverage. More efficient use of colonoscopy would also be necessary, including withholding colonoscopic screening from patients with limited life expectancy (7), performing surveillance colonoscopy at appropriate intervals (42), and considering a single lifetime-screening colonoscopy strategy (43).
The project was supported by the New Mexico Department of Health, contract 02/665.4200.0189. We appreciate the comments of Amnon Sonnenberg, MD, MSc and Meg Adams-Cameron, MPH, who reviewed an early draft of the manuscript. We also appreciate the support of the New Mexican endoscopists who responded to our survey. The material in this article was presented in part at the American Society of Gastrointestinal Endoscopy Meeting, Orlando, Fla, May 19, 2003.
Figures and Tables
Table 1 Questions for Survey of Gastroenterologists on Colonoscopy Screening Capacity, New Mexico, 2001
1. How many endoscopists work in your practice?
2. How many perform colonoscopy?
3. How many perform sigmoidoscopy?
4. During an average week, how many colonoscopies do you perform?
5. During an average week, how many colonoscopies are performed in your practice?
6. During an average week, how many sigmoidoscopies do you perform?
7. During an average week, how many sigmoidoscopies are performed in your practice?
8. How many additional screening colonoscopies could your practice perform in a week?
9. How many additional screening sigmoidoscopies could your practice perform in a week?
10. What resources would be required to perform additional endoscopic procedures?
None More equipment More space More support staff More physicians Other
Table 2 Current Volume of Colonoscopies Performed Weekly and Weekly Capacity for Additional Colonoscopies, New Mexico, 2001
Practice Type Total number of endoscopists Current colonoscopies per endoscopista Total current colonoscopies Additional colonoscopies per endoscopista Total additional colonoscopies
Group 40 16.3 (12.9, 26.5) 652 6.3 (1.8, 10) 252
Solo 9 20 (15, 21) 180 10 (5, 15) 90
Combined 49 NAb 832 NA 342
a Values are median (interquartile range).
b NA = not applicable.
Table 3 Number of Colonoscopies Required to Increase the Prevalence of Current Screening During a Five-year Period for New Mexico Adults Aged 50 to 85 Years
Screening increase over five years (%) Annual number of colonoscopies (based on detection rates from VA study)a Annual number of colonoscopies (based on detection rates from sensitivity analysis)b
5 5568 (5983)c 5137 (5360)
10 11,136 (11,966) 10,274 (10,721)
15 16,704 (17,949) 15,411 (16,082)
20 22,272 (23,932) 20,548 (21,442)
25 27,840 (29,915) 25,568 (26,800)
a Includes numbers of screening tests based on 2000 New Mexico census data and numbers of surveillance tests based on applying cancer (1.0%) and adenomatous polyp (37%) detection rates from a Department of Veterans Affairs (VA) study (28).
b Includes numbers of screening tests based on 2000 New Mexico census data and numbers of surveillance tests based on applying cancer (0.5%) and adenomatous polyp (20%) detection rates from sensitivity analysis.
c Numbers in parentheses reflect the strategy of performing a three-year surveillance colonoscopy on all patients with adenomatous polyps compared to five-year surveillance interval.
Table 4 Barriers to Performing Additional Endoscopic Tests, Results of a Survey of Gastroenterologists, New Mexico, 2001a
Barrier Group practices
N = 11 Solo practitioners
N = 9
None 1 4
More equipment 4 2
More space 8 1
More support staff 7 3
More physicians 8 4
Other (lack of referrals) 1 0
a More than one barrier could be reported.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Hoffman RM, Stone SN, Herman C, Jung AM, Cotner J, Espey D, et al. New Mexico’s capacity for increasing the prevalence of colorectal cancer screening with screening colonoscopies. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0073.htm
==== Refs
1 Jemal A Tiwari RC Murray T Ghafoor A Samuels A Ward E 2004 54 8 29 CA Cancer J Clin Cancer statistics, 2004 14974761
2 Hardcastle JD Chamberlain JO Robinson MH Moss SM Amar SS Balfour TW 348 9040 1996 1472 1477 Lancet Randomised controlled trial of faecal-occult-blood screening for colorectal cancer 8942775
3 Mandel JS Church TR Bond JH Ederer F Geisser MS Mongin SJ 343 22 2000 1603 1607 N Engl J Med The effect of fecal occult-blood screening on the incidence of colorectal cancer 11096167
4 Kronborg O Fenger C Olsen J Jorgensen OD Sondergaard O 348 9040 1996 1467 1471 Lancet Randomised study of screening for colorectal cancer with faecal-occult-blood test 8942774
5 Muller AD Sonnenberg A 123 12 1995 904 910 Ann Intern Med Prevention of colorectal cancer by flexible endoscopy and polypectomy. A case-control study of 32,702 veterans 7486484
6 Selby JV Friedman GD Quesenberry CP Jr Weiss NS 326 10 1992 653 657 N Engl J Med A case-control study of screening sigmoidoscopy and mortality from colorectal cancer 1736103
7 Pignone M Rich M Teutsch SM Berg AO Lohr KN 137 2 2002 132 141 Ann Intern Med Screening for colorectal cancer in adults at average risk: a summary of the evidence for the U.S. Preventive Services Task Force 12118972
8 Smith RA Cokkinides V Eyre HJ American Cancer Society 53 1 2003 27 43 CA Cancer J Clin American Cancer Society guidelines for the early detection of cancer, 2003 12568442
9 Rex DK Lieberman DA 54 5 2001 662 667 Gastrointest Endosc Feasibility of colonoscopy screening: discussion of issues and recommendations regarding implementation 11677497
10 Winawer S Fletcher R Rex D Bond J Burt R Ferrucci J 124 2 2003 544 560 Gastroenterology Colorectal cancer screening and surveillance: clinical guidelines and rationale-Update based on new evidence 12557158
11 Seeff LC Nadel MR Klabunde CN Thompson T Shapiro JA Vernon SW 2004 100 2093 2103 Cancer Patterns and predictors of colorectal cancer test use in the adult U.S. population 15139050
12 Centers for Disease Control and Prevention 2003 52 193 196 MMWR Colorectal cancer test use among persons aged ≥ 50 years--United States, 2001
13 Centers for Disease Control and Prevention Atlanta (GA) Centers for Disease Control and Prevention 2002 Behavioral Risk Factor Surveillance System data — 2002
14 Nelson DE Holtzman D Bolen J Stanwyck CA Mack KA 2001 46 Suppl 1 S3 S42 Soz Praventivmed Reliability and validity of measures from the Behavioral Risk Factor Surveillance System (BRFSS) 11851091
15 Cram P Fendrick AM Inadomi J Cowen ME Carpenter D Vijan S 163 13 2003 1601 1605 Arch Intern Med The impact of a celebrity promotional campaign on the use of colon cancer screening: the Katie Couric effect 12860585
16 Centers for Disease Control and Prevention 2004 Screen for life Atlanta (GA) Centers for Disease Control and Prevention
17 Reynolds T 95 8 2003 575 J Natl Cancer Inst Colossal colon, polyp man aimed at increasing screening 12697845
18 Clinton WJ 2000 National colorectal cancer awareness month, 2000 Washington (DC) Office of the Press Secretary
19 Walsh JM Terdiman JP 289 10 2003 1288 1296 JAMA Colorectal cancer screening: scientific review 12633191
20 National Committee for Quality Assurance 2004 HEDIS 2004 Summary Tables of Measures and Product Lines Washington (DC) NCQA
21 Schneider EC Riehl V Courte-Wienecke S Eddy DM Sennett C 282 12 1999 1184 1190 JAMA Enhancing performance measurement: NCQA's road map for a health information framework. National Committee for Quality Assurance 10501126
22 Winawer SJ Zauber AG Ho MN O'Brien MJ Gottlieb LS Sternberg SS 329 27 1993 1977 1981 N Engl J Med Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup 8247072
23 Pignone M Saha S Hoerger T Mandelblatt J 137 2 2002 96 104 Ann Intern Med Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the U.S. Preventive Services Task Force 12118964
24 Lieberman DA 1995 109 1781 1790 Cost-effectiveness model for colon cancer screening Gastroenterology
25 Sonnenberg A Delco F Inadomi JM 133 8 2000 573 584 Ann Intern Med Cost-effectiveness of colonoscopy in screening for colorectal cancer 11033584
26 Rex DK Johnson DA Lieberman DA Burt RW Sonnenberg A 95 4 2000 868 877 Am J Gastroenteroly Colorectal cancer prevention 2000: screening recommendations of the American College of Gastroenterology. American College of Gastroenterology
27 U.S. Census Bureau 2000 American FactFinder; Profile of general demographic characteristics Washington (DC) U.S. Census Bureau
28 Lieberman DA Weiss DG Bond JH Ahnen DJ Garewal H Chejfec G 343 3 2000 162 168 N Engl J Med Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380 10900274
29 Imperiale TF Wagner DR Lin CY Larkin GN Rogge JD Ransohoff DF 343 3 2000 169 174 Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings N Engl J Med
30 SAS Institute, Inc 1999-2001 SAS Institute, Inc Cary (NC) SAS/STAT® User's Guide, Version 8
31 U.S. Census Bureau 2003 Census 2000 urban and rural classification Washington (DC) U.S. Census Bureau
32 Nelson DE Bland S Powell-Griner E Klein R Wells HE Hogelin G 287 20 2002 2659 2667 JAMA State trends in health risk factors and receipt of clinical preventive services among US adults during the 1990s 12020301
33 Sirovich BE Schwartz LM Woloshin S 289 11 2003 1414 1420 JAMA Screening men for prostate and colorectal cancer in the United States: does practice reflect the evidence? 12636464
34 Weir HK Thun MJ Hankey BF Ries LA Howe HL Wingo PA 95 17 200312761299J Natl Cancer InstAnnual report to the nation on the status of cancer, 1975-2000, featuring the uses of surveillance data for cancer prevention and control2003;95(17):1276-99. 12953083
35 U.S. Preventive Services Task Force 137 2 2002 129 131 Ann Intern Med Screening for colorectal cancer: recommendation and rationale 12118971
36 Brown ML Klabunde CN Mysliwiec P 115 2 2003 129 133 Am J Med Current capacity for endoscopic colorectal cancer screening in the United States: data from the National Cancer Institute Survey of Colorectal Cancer Screening Practices 12893399
37 Klabunde CN Frame PS Meadow A Jones E Nadel M Vernon SW 36 3 2003 352 362 Prev Med A national survey of primary care physicians' colorectal cancer screening recommendations and practices 12634026
38 Wexner SD Litwin D Cohen J Earle D Ferzli G Flaherty J 55 2 2002 145 148 Gastrointest Endosc Principles of privileging and credentialing for endoscopy and colonoscopy 11818913
39 New Mexico Health Policy Commission 1998 Health Care in New Mexico. Access, Financing, Delivery and Outcomes Santa Fe (NM) The Commission
40 Lieberman DA Ghormley JM 1992 87 940 943 Am J Gastroenterol Physician assistants in gastroenterology: should they perform endoscopy? 1353660
41 Salem N 1 2000 27 Poll: 21% N.M. doctors want out of state system Albuquerque (NM) Albuquerque Tribune
42 Mysliwiec PA Brown ML Klabunde CN Ransohoff DF 141 4 2004 264 271 Ann Intern Med Are physicians doing too much colonoscopy? A national survey of colorectal surveillance after polypectomy 15313742
43 Sonnenberg A Delco F 162 2 2002 163 168 Arch Intern Med Cost-effectiveness of a single colonoscopy in screening for colorectal cancer 11802749
|
15670460
|
PMC1323310
|
CC BY
|
2022-01-25 23:15:43
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A07
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0087
Original Research
PEER REVIEWEDPrevalence of Overweight, Obesity, and Comorbid Conditions Among U.S. and Kentucky Adults, 2000–2002
Jenkins Todd M MPH University of Kentucky Prevention Research Center, Lexington, Ky
Department of Biostatistics, University of Alabama at Birmingham, Room 327, Ryals Bldg, 1665 University Blvd, Birmingham, AL 35294 [email protected]
205-934-5989
1 2005
15 12 2004
2 1 A082005
Introduction
Obesity rates for adults in Kentucky are regularly among the highest in the nation. Since 1991, adult obesity in Kentucky and the United States has nearly doubled. This trend is of great concern because excess weight has been associated with several chronic diseases and conditions. This paper reports on the prevalence of overweight and obesity among adults in Kentucky between 2000 and 2002. The estimates produced by this study will provide baseline figures for developing Kentucky's statewide obesity action plan.
Methods
A secondary data analysis was performed using the Centers for Disease Control and Prevention's Behavioral Risk Factor Surveillance System. Prevalence estimates and odds ratios were calculated for the United States and Kentucky.
Results
In Kentucky, 24.2% of adults were obese, compared with 21.9% nationally (P < .001). There were also significantly more overweight adults in Kentucky than there were nationwide (P < .001). Logistic regression showed that overweight and obese adults were more likely to report various comorbid conditions.
Conclusion
Overweight and obesity estimates in Kentucky were significantly higher than nationwide figures. However, overweight/obese adults in Kentucky were no more likely than their U.S. counterparts to report selected comorbid conditions.
==== Body
Introduction
The obesity epidemic in the United States worsens with each passing year. From 1991 to 2002, the prevalence of obesity has increased more than 80%, representing an estimated 43 million adults in 2002 (1). In 1991, no state in the nation had an obesity prevalence at or above 20%, but by 2002 there were 39 states with this characteristic (2). The severity of this epidemic has been highlighted in Healthy People 2010, where overweight and obesity have been grouped as leading health indicators for the United States (3). In terms of mortality, an estimated 280,000 to 325,000 adults in the United States die each year from causes related to obesity (4). More importantly, excess weight has been positively correlated with years of life lost (5).
In addition to mortality, substantial morbidity is associated with obesity. For example, in 2000, the total cost of obesity in the United States was estimated to be $117 billion ($61 billion in direct medical costs, $56 billion in indirect costs) (3,6). An estimated 9.1% of annual medical spending in the United States is attributed to overweight and obesity — a figure that rivals medical costs attributable to cigarette smoking (7). Overweight and obesity have been associated with several chronic diseases and conditions, including cardiovascular disease, type 2 diabetes, hypertension, stroke, arthritis, high serum cholesterol, and some cancers (8-10). This is of great concern in Kentucky because the prevalence rates for overweight and obesity continue to increase and are regularly among the highest in the nation. All told, obesity substantially increases morbidity and impairs quality of life (11). Kentucky is developing a statewide action plan to address this public health issue. Estimates produced from this analysis will serve as baseline figures for the action plan.
Methods
A secondary data analysis was performed using data from the Behavioral Risk Factor Surveillance System (BRFSS), 2000–2002 (12,13). Conducted by the Centers for Disease Control and Prevention (CDC), the BRFSS is an annual population-based, random-digit-dialed telephone survey of the noninstitutionalized U.S. civilian population aged 18 or older. This ongoing surveillance system measures health behaviors and preventive practices related to several leading causes of death (12,13). Kentucky data were obtained from the Kentucky BRFSS Program (KY BRFSS) (14). Data from 21,016 adults in Kentucky were collected during this period. U.S. data were retrieved from the CDC's public-use BRFSS datasets (15). Data from 642,924 adults across the nation were collected during 2000–2002. The national dataset included data from Guam, Puerto Rico, and the Virgin Islands, but these areas were excluded from this analysis.
Overweight and obesity classifications used in the analysis were derived from Body Mass Index (BMI) and were consistent with the definitions set forth by the World Health Organization (WHO) and the National Heart, Lung, and Blood Institute (underweight: BMI<18.5; normal weight: BMI = 18.5–24.9; overweight: BMI = 25.0–29.9; obesity-class 1: BMI = 30.0–34.9; obesity-class 2: BMI = 35.0–39.9; obesity-class 3: BMI ⩾40.0) (9,16). BMI (calculated as weight in kilograms divided by the square of height in meters) was calculated using the following questions: 1) "About how much do you weigh without shoes?" and 2) "About how tall are you without shoes?" (17). Respondents with missing or unknown height or weight data were excluded from the analysis. Women who reported they were pregnant at the time of the interview were also excluded from the analysis. After all exclusions, a total of 590,120 respondents for the United States and 19,722 respondents from Kentucky were included in the analysis.
Comorbid conditions were measured using the following questions (17):
Diabetes. Have you ever been told by a doctor that you have diabetes? (2000–2002)
Asthma. Did a doctor ever tell you that you have asthma? (2000) Have you ever been told by a doctor, nurse, or other health professional that you have asthma? (2001–2002)
Arthritis. Have you ever been told by a doctor that you have arthritis? (2000–2001) Have you ever been told by a doctor or other health professional that you have some form of arthritis, rheumatoid arthritis, gout, lupus, or fibromyalgia? (2002)
High blood pressure. Have you ever been told by a doctor, nurse, or other health professional that you have high blood pressure? (2001)
High cholesterol. Have you ever been told by a doctor, nurse, or other health professional that your blood cholesterol is high? (2001)
Health status. Would you say that in general your health is: Excellent, Very Good, Good, Fair, or Poor? (2000–2002).
Women reporting gestational diabetes were coded as having diabetes. Questions assessing blood pressure and cholesterol are asked in alternating years (rotating core questions) and were not selected as modules in most states (including Kentucky) in 2000 or 2002; thus, data from these questions were analyzed for 2001 only.
SAS version 8.2 (SAS Institute Inc, Cary, NC) and SAS-Callable SUDAAN version 8.0.1 (Research Triangle Institute, Research Triangle Park, NC) were used to perform the data analysis and to account for the complex sampling design (18,19). PROC DESCRIPT was used to calculate age-adjusted prevalence estimates and their corresponding standard errors. Prevalence estimates of obesity, overweight, and comorbid conditions were age-adjusted to the 2002 BRFSS. However, figures representing total number of adults were derived from crude estimates. Multivariate logistic regression was performed using PROC RLOGIST to assess associations between BMI and comorbid conditions while controlling for age, race, sex, education, and smoking status. All reported data are weighted, correcting for variation in selection probability and demographic imbalances (20).
Results
During the years 2000–2002, 24.2% of adults (683,000) in Kentucky were obese (BMI ⩾30.0), compared with 21.9% (42 million) in the United States (P < .001) (Table 1). Both men (24.6%) and women (23.8%) in Kentucky had significantly higher levels of obesity compared with men and women nationally (21.9% [P < .001] for men and 21.7% [P < .001] for women). Among race/ethnicity groups, only non-Hispanic whites in Kentucky had a significantly higher obesity estimate compared with the United States (P < .001).
There were also more overweight (BMI = 25.0–29.9) adults in Kentucky than nationwide. During 2000–2002, 62.8% of adults in Kentucky (1.76 million adults) were overweight, compared with 59.7% (115 million adults) nationally (P < .001) (Table 2). As seen with obesity, the prevalence of overweight among men (70.7%) and women (54.8%) in Kentucky was also significantly higher than among their counterparts nationally (67.9% [P < .001] of men and 51.3% [P < .001] of women). Estimates for overweight were also higher among non-Hispanic whites in Kentucky (62.3%) compared with the United States (57.8% [P < .001]). Rates for non-Hispanic blacks in Kentucky and in the United States were not significantly different, but they were significantly higher than for non-Hispanic whites within both regions. By age group, estimates peaked at ages 50–59 for both Kentucky (70.0%) and the United States (67.6%).
Table 3 lists the prevalence of comorbid conditions by BMI category among Kentucky and U.S. adults. As expected, the prevalence of each condition increased with BMI. The largest increases in prevalence were observed with diabetes and fair or poor health status. For comparisons between Kentucky and the United States, differences were most pronounced for arthritis and fair or poor health status. For every BMI category, the prevalence of adults in Kentucky with arthritis was greater than adults nationally. The prevalence of adults in Kentucky reporting fair or poor health status was higher in the United States for all but the highest BMI category, obesity-class 3.
Multivariate logistic regression analysis indicated significant associations for overweight and obesity with each comorbid condition (Table 4). Overweight and obese adults in Kentucky were more likely than those of normal weight to have diabetes, asthma, arthritis, high blood pressure, high cholesterol, and fair or poor health status. As expected, results were strongest for those with obesity-class 3. Using a normal BMI as the reference, the odds of Kentucky respondents with obesity-class 3 were nine times higher (Prevalence Odds Ratio [POR] 9.10) to report diabetes, four times higher (POR 4.26) to report arthritis, more than six times higher (POR 6.83) to report high blood pressure, and more than four times higher (POR 4.59) to report a fair or poor health status. However, none of the results listed in Table 4 for Kentucky was significantly different from U.S. estimates.
Discussion
Since 1991, the prevalence of obesity among adults in the United States and Kentucky has doubled (1,21). When combined with overweight, more than 60% of adults throughout the United States and Kentucky are classified as overweight/obese (BMI ⩾25). Among the fifty states and the District of Columbia in 2002, the obesity rate in Kentucky was the 13th highest (24.4%; 95% confidence interval [CI], 22.8–26.1), 8.5% higher than the U.S. rate (22.5%; 95% CI, 22.2–22.8). For overweight, Kentucky had the sixth highest estimate (63.6%; 95% CI, 61.8–65.4), 5.3% greater than the U.S. rate (60.4%; 95% CI, 60.0–60.7).
Results by age indicate that younger adults in Kentucky (aged 18–29, 30–39, 40–49) had significantly higher obesity estimates than younger adults in the United States. Data for overweight were similar, with estimates for adults up to the age of 60 significantly higher in Kentucky. Comparisons between the United States and Kentucky for youth (<18 years) were also similar. According to the 2001 Youth Risk Behavioral Surveillance System (YRBSS), 12.3% of high school students in Kentucky were overweight, and another 15.2% were at risk for becoming overweight, compared with 10.5% overweight and 13.6% at risk nationally (22). These data suggest that the prevalence of overweight and obesity is unlikely to change in Kentucky in the foreseeable future. Results from this analysis revealed that overweight and obesity are more prevalent in Kentucky, but those with excess weight were no more likely to have other comorbid conditions (e.g., diabetes, arthritis) in Kentucky than observed nationally. However, with its disproportionate share of overweight and obesity, Kentucky will face the costly task of treating and caring for a disproportionately greater number of its population beset with comorbid conditions related to excess weight for many years to come.
The results reported here are subject to several limitations. First, the survey design includes only those noninstitutionalized civilian adults who have a telephone. Therefore, results are generalizable only to this population. According to Census 2000, 2.4% of occupied housing units across the nation and 4.7% in Kentucky do not have telephone service (23). Individuals without telephones are more likely to have a low socioeconomic status, which is associated with obesity (24,25). Therefore, results in this analysis are likely to be underestimated. The use of self-reported height and weight represents another limitation. Respondents in self-reported surveys tend to overestimate their height, while overweight respondents tend to underestimate their weight (1). Compared with studies based on directly measured height and weight, such as the National Health and Nutrition Examination Survey (NHANES), obesity estimates from self-report tend to be lower (26). The prevalence of obesity from NHANES 1999–2000 was 30.5%, compared with 19.8% from the 2000 BRFSS (10,26). There are also drawbacks to using BMI as an indicator for overweight and obesity. BMI can overestimate body fat in persons who are very muscular and underestimate body fat in persons who have lost muscle mass, such as many elderly (10). However, estimates from these potentially misclassified groups likely had little overall impact on the analysis.
The impact of excess weight extends beyond the monetary costs and physical ailments associated with it. Other issues such as social stigma, discrimination, and poor body image all contribute to a lower quality of life for the overweight and obese compared with individuals of normal weight (3,27). If current trends continue, obesity will soon overtake smoking as the primary preventable cause of death (28). These results, in part, serve as baseline figures for Kentucky's initial obesity action plan. Future initiatives addressing diet and physical activity are anticipated to be derived from this plan.
Thanks go to Mark Dignan, Carol White, and Crystal Jenkins of the University of Kentucky Prevention Research Center for their assistance with editing the manuscript.
Figures and Tables
Table 1 Prevalence of Obesity (Body Mass Index ≥30) by Demographic Characteristic, Adults Aged ≥18, United States and Kentucky, 2000–2002 Behavioral Risk Factor Surveillance Systema
Characteristic Kentucky United States
Total 24.2 (23.4-25.1) 21.9 (21.7-22.1)
Sex
Male 24.6 (23.2-25.9) 21.9 (21.7-22.2)
Female 23.8 (22.6-24.9) 21.7 (21.5-22.0)
Race/ethnicity
Non-Hispanic white 23.7 (22.8-24.6) 20.3 (20.1-20.5)
Non-Hispanic black 33.7 (29.2-38.2) 32.9 (32.2-33.6)
Non-Hispanic other 24.4 (17.2-31.5) 16.5 (15.7-17.4)
Hispanic 21.0 (14.2-27.7) 26.2 (25.1-27.2)
Age (years)
18-29 18.4 (16.0-20.9) 14.3 (13.9-14.7)
30-39 26.9 (24.7-29.2) 21.4 (21.0-21.8)
40-49 28.2 (26.1-30.2) 24.6 (24.1-25.1)
50-59 27.0 (24.9-29.1) 26.9 (26.4-27.4)
60-69 25.2 (22.9-27.4) 25.2 (24.6-25.7)
70+ 17.0 (15.2-18.9) 17.4 (16.9-17.8)
Education
<High school 28.2 (26.1-30.3) 29.4 (28.6-30.1)
High school grad 25.1 (23.6-26.5) 24.3 (23.9-24.6)
Some college 25.4 (23.4-27.3) 22.3 (21.9-22.6)
College+ 18.0 (16.1-19.8) 16.2 (15.9-16.5)
Smoking status
Current 19.6 (18.2-21.1) 17.5 (17.1-17.8)
Former 28.6 (26.0-31.3) 24.0 (23.6-24.4)
Never 25.2 (23.9-26.5) 22.5 (22.2-22.7)
a All values represent percentages (95% confidence intervals). Age-adjusted to the 2002 Behavioral Risk Factor Surveillance System.
Table 2 Prevalence of Overweight (Body Mass Index = 25.0–29.9) by Demographic Characteristic, Adults Aged ⩾18, United States and Kentucky, 2000–2002 Behavioral Risk Factor Surveillance Systema
Characteristic Kentucky United States
Total 62.8 (61.8-63.8) 59.7 (59.5-59.9)
Sex
Male 70.7 (69.3-72.1) 67.9 (67.6-68.2)
Female 54.8 (53.5-56.1) 51.3 (51.0-51.6)
Race/ethnicity
Non-Hispanic white 62.3 (61.3-63.3) 57.8 (57.5-58.0)
Non-Hispanic black 71.0 (66.6-75.5) 70.7 (70.0-71.4)
Non-Hispanic other 62.3 (53.4-71.2) 51.3 (50.0-52.6)
Hispanic 58.2 (49.7-66.8) 67.3 (66.3-68.3)
Age (years)
18-29 50.7 (47.9-53.5) 43.4 (42.8-44.0)
30-39 62.0 (59.6-64.4) 58.7 (58.2-59.2)
40-49 66.6 (64.5-68.8) 62.9 (62.4-63.4)
50-59 70.0 (67.9-72.1) 67.6 (67.1-68.1)
60-69 68.2 (65.7-70.7) 67.0 (66.4-67.6)
70+ 57.1 (54.6-59.5) 57.1 (56.5-57.7)
Education
<High school 63.8 (61.5-66.1) 66.3 (65.6-67.0)
High school grad 64.1 (62.5-65.6) 62.4 (62.0-62.7)
Some college 64.0 (62.0-66.1) 59.8 (59.3-60.2)
College+ 57.1 (54.7-59.6) 54.4 (54.0-54.8)
Smoking status
Current 55.4 (53.4-57.3) 52.7 (52.2-53.2)
Former 68.7 (66.4-71.0) 64.2 (63.8-64.7)
Never 64.0 (62.6-65.5) 59.7 (59.4-60.0)
a All values represent percentages (95% confidence intervals). Age-adjusted to the 2002 Behavioral Risk Factor Surveillance System.
Table 3 Prevalence of Comorbid Conditions by Body Mass Index Category, Adults Aged ≥18, United States and Kentucky, 2000–2002 Behavioral Risk Factor Surveillance Systema
Body Mass Index
Normal
(18.5-24.9) Overweight
(25.0-29.9) Obese-class 1
(30.0-34.9) Obese-class 2
(35.0-39.9) Obese-class 3
(≥40.0)
Diabetes
Kentucky 3.7 (3.2-4.3) 7.1 (6.4-7.9) 12.1 (10.6-13.6) 19.3 (16.2-22.3) 23.7 (17.7-29.7)
United States 4.8 (4.7-5.0) 7.5 (7.3-7.7) 13.4 (13.0- 13.8) 19.8 (18.9- 20.7) 26.4 (25.0- 27.8)
Asthma
Kentucky 9.9 (9.0-10.9) 10.2 (9.1-11.2) 12.8 (11.1-14.5) 15.3 (12.5-18.0) 21.8 (16.9-26.7)
United States 9.6 (9.4-9.8) 9.9 (9.7-10.1) 12.6 (12.2-13.0) 16.0 (15.2-16.8) 21.9 (20.5-23.2)
Arthritis
Kentucky 28.8 (27.4-30.2) 32.9 (31.5-34.3) 40.1 (37.8-42.4) 48.6 (44.4-52.7) 57.6 (52.0-63.2)
United States 22.3 (21.9-22.6) 25.6 (25.2-26.0) 32.3 (31.8-32.9) 38.5 (37.4-39.5) 47.1 (45.4-48.7)
High blood pressureb
Kentucky 22.9 (20.7-25.0) 32.5 (30.2-34.8) 45.2 (41.5-49.0) 49.6 (42.6-56.6) 50.1 (41.9-58.3)
United States 19.1 (18.6-19.6) 28.9 (28.3-29.4) 39.5 (38.6-40.5) 46.7 (44.9-48.4) 53.5 (51.0-55.9)
High cholesterolb
Kentucky 24.1 (21.4-26.7) 33.6 (30.6-36.5) 34.8 (30.2-39.4) 36.5 (29.5-43.5) 27.4 (18.9-35.8)
United States 24.7 (24.1-25.3) 32.5 (31.9-33.2) 37.6 (36.5-38.7) 37.0 (35.1-39.0) 36.0 (33.4-38.7)
Fair/poor health
Kentucky 20.2 (18.9-21.4) 22.4 (21.1-23.6) 29.2 (26.9-31.4) 37.7 (33.8-41.6) 47.1 (39.9-54.4)
United States 13.1 (12.8-13.3) 14.7 (14.5-15.0) 20.9 (20.4-21.4) 29.6 (28.6-30.6) 39.4 (37.9-40.9)
a All values represent percentages (95% confidence intervals). Age-adjusted to the 2002 Behavioral Risk Factor Surveillance System.
b 2001 only.
Table 4 Adjusted Prevalence Odds Ratios for Selected Comorbid Conditions by Body Mass Index Category, Adults Aged ⩾18, United States and Kentucky, 2000–2002 Behavioral Risk Factor Surveillance Systema
Body Mass Index
Normal (referent) Overweight Obese-class 1 Obese-class 2 Obese-class 3
Diabetes
Kentucky 1.00 1.83 (1.49-2.26) 3.31 (2.63-4.16) 5.96 (4.51-7.89) 9.10 (6.55-12.65)
United States 1.00 1.56 (1.49-1.64) 2.93 (2.77-3.09) 4.71 (4.38-5.07) 7.13 (6.54-7.76)
Asthma
Kentucky 1.00 1.10 (0.93-1.31) 1.34 (1.10-1.65) 1.61 (1.26-2.05) 2.72 (2.04-3.63)
United States 1.00 1.12 (1.09-1.17) 1.44 (1.38-1.50) 1.83 (1.72-1.94) 2.65 (2.45-2.86)
Arthritis
Kentucky 1.00 1.34 (1.20-1.51) 1.88 (1.63-2.18) 2.84 (2.26-3.58) 4.26 (3.28-5.53)
United States 1.00 1.37 (1.32-1.41) 1.95 (1.87-2.03) 2.62 (2.47-2.79) 4.05 (3.73-4.39)
High blood pressureb
Kentucky 1.00 1.74 (1.43-2.12) 3.26 (2.57-4.13) 4.18 (2.75-6.34) 6.83 (4.08-11.44)
United States 1.00 1.84 (1.76-1.93) 3.19 (3.00-3.39) 4.51 (4.12-4.94) 6.46 (5.75-7.27)
High cholesterolb
Kentucky 1.00 1.65 (1.35-2.03) 1.73 (1.33-2.25) 2.03 (1.39-2.96) 1.60 (0.95-2.71)
United States 1.00 1.48 (1.41-1.55) 1.90 (1.78-2.01) 1.88 (1.71-2.07) 1.88 (1.67-2.13)
Fair/poor health
Kentucky 1.00 1.18 (1.04-1.34) 1.74 (1.48-2.04) 2.67 (2.15-3.31) 4.59 (3.46-6.09)
United States 1.00 1.11 (1.07-1.15) 1.65 (1.58-1.72) 2.64 (2.49-2.80) 4.36 (4.04-4.72)
a All values represent odds ratios (95% confidence intervals). Adjusted for age, race, sex, education, smoking status to the 2002 Behavioral Risk Factor Surveillance System.
b 2001 only.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Jenkins TM. Prevalence of overweight, obesity, and comorbid conditions among U.S. and Kentucky adults, 2000–2002. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0087.htm
==== Refs
1 Mokdad AH Serdula MK Dietz WH Bowman BA Marks JS Koplan JP 282 1999 1519 1522 JAMA The spread of the obesity epidemic in the United States, 1991-1998 10546690
2 Indu B Ahluwalia IB Mack KA Murphy W Mokdad AH Bales VS 2003 52 No. SS-8 MMWR State-specific prevalence of selected chronic disease-related characteristics – Behavioral Risk Factor Surveillance System, 2001
3 U.S. Department of Health and Human Services 2001 U.S. Department of Health and Human Services, Public Health Service, Office of the Surgeon General Rockville (MD) The Surgeon General's call to action to prevent and decrease overweight and obesity
4 Allison DB Fontaine KR Manson JE Stevens J VanItallie TB 282 1999 1530 1538 JAMA Annual deaths attributable to obesity in the United States 10546692
5 Fontaine KR Redden DT Wang C Westfall AO Allison DB 289 2003 187 193 JAMA Years of life lost due to obesity 12517229
6 Centers for Disease Control and Prevention U.S. Department of Health and Human Services Atlanta (GA) Preventing chronic diseases: investing wisely in health. Preventing obesity and chronic diseases through good nutrition and physical activity
7 Finkelstein EA Fiebelkorn IC Wang G W3 2003 219 226 Health Aff National medical spending attributable to overweight and obesity: how much, and who's paying?
8 Must A Spadano J Coakley EH Field AE Colditz G Dietz WH 282 1999 1523 1529 JAMA The disease burden associated with overweight and obesity 10546691
9 National Heart, Lung, and Blood Institute 1998 9 Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report NIH Publication No. 98-4083 Rockville (MD) National Heart, Lung, and Blood Institute
10 Paeratakul S Lovejoy JC Ryan DH Bray GA 26 9 9 2002 1205 1210 Int J Obes Relat Metab Disord The relation of gender, race, and socioeconomic status to obesity and obesity comorbidities in a sample of US adults 12187397
11 Mokdad AH Bowman BA Ford ES Vinicor F Marks JS Koplan JP 286 2001 1195 1200 JAMA The continuing epidemics of obesity and diabetes in the United States 11559264
12 Centers for Disease Control and Prevention Behavioral Risk Factor Surveillance System (BRFSS) Atlanta (GA) Centers for Disease Control and Prevention Accessed 2003 Jul 2
13 Centers for Disease Control and Prevention Health risks in America: gaining insight from the Behavioral Risk Factor Surveillance System Atlanta (GA) U.S. Department of Health and Human Services Revised edition 1997
14 Kentucky Department for Public Health Frankfort (KY) Kentucky Department for Public Health, Surveillance and Health Data Branch Kentucky BRFSS Program
15 Centers for Disease Control and Prevention Atlanta (GA) U.S. Department of Health and Human Services Behavioral Risk Factor Surveillance System Survey data 2000-2002
16 World Health Organization Geneva World Health Organization Obesity: preventing and managing the global epidemic - report of a WHO consultation on obesity 1997 Jun 3-5
17 Centers for Disease Control and Prevention 2000-2002 U.S. Department of Health and Human Services Atlanta (GA) Behavioral Risk Factor Surveillance System survey questionnaire
18 1999-2000 SAS Institute Inc Cary (NC) SAS System. Version 8.1
19 Research Triangle Institute 2001 Research Triangle Institute Research Triangle Park (NC) SUDAAN User's Manual, Release 8.0
20 Centers for Disease Control and Prevention BRFSS data systems prevalence training module Atlanta (GA) Centers for Disease Control and Prevention Accessed 2003 Jul 2 2002
21 Centers for Disease Control and Prevention Behavioral Risk Factor Surveillance System survey data Atlanta (GA) U.S. Department of Health and Human Services 1990-2001
22 Centers for Disease Control and Prevention 51 SS04 2002 1 64 MMWR Youth risk behavior surveillance-United States, 2001
23 U.S. Census Bureau 2000 Washington (DC) U.S. Census Bureau Census 2000 summary file 3 Accessed 2003 Jul 14
24 Mokdad AH Ford ES Bowman BA Dietz WH Vinicor F Bales VS 289 2003 76 79 JAMA Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001 12503980
25 Ford ES 51 1998 55 60 J Clin Epidemiol Characteristics of survey participants with and without a telephone: findings from the third National Health and Nutrition Examination Survey 9467634
26 Flegal KM Carroll MD Ogden CL Johnson CL 288 2002 1723 1727 JAMA Prevalence and trends in obesity among US adults, 1999-2000 12365955
27 Ford ES Moriarty DG Zack MM Mokdad AH Chapman DP 9 2001 21 31 Obes Res Self-reported body mass index and health-related quality of life: findings from the Behavioral Risk Factor Surveillance System 11346664
28 Manson JE Bassuk SS 289 2003 229 230 JAMA Obesity in the United States: a fresh look at its high toll 12517236
|
15670461
|
PMC1323311
|
CC BY
|
2022-01-24 23:45:19
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A08
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0032
Original Research
PEER REVIEWEDThe Volume and Capacity of Colonoscopy Procedures Performed at New York City Hospitals in 2002
Leng Jennifer C.F. MD, MPH New York City Department of Health and Mental Hygieney
125 Worth St, Room 315, CN6, New York, NY 10013 [email protected]
212-788-4637
Thorpe Lorna E PhD New York City Department of Health and Mental Hygiene, New York, NY
Feldman Gabe E MD, MPH, MBA, MHA New York City Department of Health and Mental Hygiene, New York, NY
Thomas Pauline A MD New York City Department of Health and Mental Hygiene, New York, NY
Frieden Thomas R MD, MPH New York City Department of Health and Mental Hygiene, New York, NY
1 2005
15 12 2004
2 1 A092005
Introduction
Colorectal cancer is the second leading cause of cancer death in New York City. In March 2003, the New York City Department of Health and Mental Hygiene recommended colonoscopy every 10 years as the preferred screening test for adults aged 50 years and older in New York City. To screen all eligible adults in New York City would require that approximately 200,000 colonoscopy exams be performed annually. As part of this recommendation, we evaluated current colonoscopy capacity in New York City hospitals.
Methods
We surveyed endoscopy suite nursing or administrative staff at all 66 adult acute care hospitals performing colonoscopy in New York City. Data on colonoscopy procedures performed in 2002 were collected between February and June 2003.
Results
All hospitals and two affiliated clinics responded. The number of hospital-based colonoscopy exams performed in 2002 was estimated to be 126,000. Of these, 53,600 (43%) were estimated to be for screening. Hospitals reported their maximum annual capacity to be 195,200, approximately 69,100 more than current practice. Reported barriers to performing more colonoscopy exams included inadequate suite time and space (31%), inadequate staffing (28%), and insufficient patient referrals (24%).
Conclusion
In 2003, endoscopy suites at New York City hospitals performed approximately one quarter of the estimated citywide need of 200,000 screening colonoscopies. Procedures conducted in outpatient office settings were not assessed. Most endoscopy suites, particularly private hospitals, reported having the capacity to conduct additional procedures. Hospitals and endoscopy suites should prioritize the development of institutional measures to increase the number of persons receiving screening colonoscopy.
==== Body
Introduction
Colorectal cancer is the second leading cause of cancer deaths (after lung) in New York City (NYC) (1) and the leading cause of cancer death among nonsmokers (2). In 2000, 1600 New Yorkers died of colorectal cancer (1). In March 2003, based on findings from an advisory committee on colorectal cancer screening (Citywide Colon Cancer Control Coalition), the New York City Department of Health and Mental Hygiene (NYC DOHMH) recommended colonoscopy every 10 years as the preferred colorectal cancer screening test for average-risk asymptomatic men and women aged 50 years and older in NYC (2,3). Colonoscopy is highly sensitive, examines the entire colon, and allows for screening, diagnosis, and polypectomy in a single visit. While colonoscopy is one of a series of recommended options in all major national colorectal cancer screening guidelines (4-7), few institutions have endorsed it as the preferred screening option. Findings from the National Polyp Study, however, suggest that periodic colonoscopy could prevent 76% to 90% of colon cancers (8). The NYC DOHMH recommendation was based on this estimated effectiveness of colonoscopy in addition to the desire to reduce patient and provider confusion about multiple screening options.
Nationally, concern over increased colonoscopy demand and insufficient capacity has raised the question of whether performing colonoscopies on all eligible adults aged 50 and older is feasible (9). In some U.S. cities, physicians cite waiting lists of up to eight months, and in extreme situations, waiting lists have been closed (9). New York City has a high concentration of specialists and teaching hospitals and therefore may have a greater capacity to perform colonoscopies compared with other communities (10). The NYC DOHMH sought to evaluate current colonoscopy volume and reported maximum capacity in all NYC acute care hospitals.
Methods
Study population
We conducted a telephone survey of nursing or administrative staff at all NYC adult acute care hospitals from February to June 2003 (11). This included 55 voluntary hospitals, three Veterans Administration (VA) hospitals, and 11 public hospitals (which provide care to New Yorkers regardless of their ability to pay). NYC DOHMH staff telephoned hospital endoscopy suites and interviewed the nurse manager or his or her designee. NYC DOHMH staff requested to interview the clinic or nurse manager or the person most able to accurately provide data on the number of colonoscopy exams performed at that endoscopy suite. If key staff were unavailable at the time of the call, a follow-up telephone call was arranged.
At the close of the interview, NYC DOHMH staff also asked if there were other clinical sites within the same facility (or affiliated with the facility) that performed colonoscopy exams. If so, NYC DOHMH staff interviewed staff at these additional sites. Every distinct clinical site that performed colonoscopy procedures and was officially affiliated with the originally targeted facility was included in the survey. Specific departments, such as Gastroenterology and General Surgery, were not separately contacted.
Survey
Using structured survey questions, interviewers asked respondents to identify the type of clinic in which colonoscopy exams are performed, the number and specialty of physicians who perform the procedure, and the waiting time to schedule a patient. Respondents were asked the total number of exams performed during the past year (2002), and the approximate percentage (<25%, 25%–50%, 50%–75%, or >75%) performed for screening purposes. They were then asked to provide the maximum number of exams that could be performed per month. Final questions probed respondents about barriers to performing at full capacity and willingness to receive more colonoscopy referrals. No patient-level information was collected.
Statistical analysis
The citywide need for screening colonoscopy exams was estimated assuming a steady-state population structure. We divided the 2000 census count of New Yorkers aged 50 and older (2,102,578) by the recommended screening colonoscopy interval (10 years) (1). We estimated that to screen all 2 million eligible New Yorkers every 10 years, NYC would need to perform approximately 200,000 colonoscopy exams per year.
Survey variables were analyzed using descriptive statistics. Each hospital reported the total number of exams performed in 2002; the sum of these numbers is the estimated total number of exams performed in hospitals citywide. The estimated number of exams performed for screening purposes in 2002 was calculated by multiplying the midpoint of a given clinic's reported screening percentage category by the total number of reported colonoscopy exams performed in 2002 for that clinic. The same calculation was performed using the low and high endpoints of a given screening percentage category to give an estimated range of the number of exams performed for screening for each clinic. The estimated maximum number of colonoscopy exams an endoscopy suite could perform per year was determined by multiplying the reported maximum number of exams feasible per month by 12.
Potential annual residual capacity was estimated by subtracting the number of exams performed in 2002 from the estimated maximum annual number of exams that could be performed. All estimated values were rounded to the nearest hundred because of the measurement tool's lack of precision. The estimated annual maximum number of exams performed and the estimated annual potential residual capacity were calculated using the exact numbers reported and then rounded to the nearest hundred. Descriptive results were presented according to type of hospital (voluntary/VA vs public). Data were analyzed using SAS, version 8 (SAS Institute, Inc, Cary, NC).
Results
We contacted staff at all 69 acute care hospitals in NYC. Three hospitals did not have endoscopy suites and were excluded from the survey. Two hospitals had additional affiliated endoscopy suites; staff from these sites were included in the survey for a total of 68 sites (Table 1). Most hospitals were voluntary/VA hospitals (84%); the rest (16%) were public hospitals.
The majority of staff interviewed (51%) were nurses in supervisory positions, including nurse managers, charge nurses, nurse directors, and nurse supervisors. Eighteen percent were managers or executive administrators, and 10% were physicians or medical directors. The remaining respondents (21%) were medical and administrative ancillary staff responsible for maintaining endoscopy suite colonoscopy schedules. Respondents reported that a total of 963 physicians were performing colonoscopy exams in these endoscopy suites at the time of the survey. Ninety-three percent of physicians were gastroenterologists, and the remaining 7% were colorectal or general surgeons. The median number of physicians performing colonoscopy exams at public endoscopy suites was five, compared with 12 at voluntary/VA endoscopy suites. The median waiting period to schedule a routine screening colonoscopy was 49 days (range, 21–150 days) at public endoscopy suites and seven days (range, 0–90 days) at voluntary/VA endoscopy suites.
The total number of reported colonoscopy exams performed in acute care hospitals in NYC in 2002 was estimated to be 126,000 (Table 2). Endoscopy suites at voluntary/VA hospitals performed an estimated 117,200 (93%) of the total estimated colonoscopy exams; public hospitals performed 8800 (7%). The total number of exams performed for screening was estimated to be 53,600 (range, 37,800–69,000). Sixty-one endoscopy suites (90%) were able to provide data on both the number of exams performed in 2002 and the estimated maximum annual number of exams that could be performed; of these, 89% reported a maximum capacity that was higher than the volume reported for 2002. If maximum capacity as reported by suites was achieved, approximately 69,100 additional colonoscopy exams could be conducted annually (potential residual capacity). Most of the potential residual capacity (88%) was reported by voluntary/VA endoscopy suites; 12% was reported by public endoscopy suites.
All public hospitals and 50 of 57 voluntary/VA hospitals reported at least one barrier to performing more colonoscopy exams. The most commonly reported barriers were inadequate suite time and space (31%), low physician-staffing levels (28%), low nurse- and/or technician-staffing levels (28%), and insufficient patient referrals (24%). Respondents at public facilities more frequently cited staffing shortages compared with private facilities (45% [public] vs 25% [private] needed more physicians; 64% [public] vs 21% [private] needed more nurses/technicians). At public facilities, respondents also more often described patient no-shows and cancellations (27% [public] vs 4% [private]) as barriers to performing more exams and less frequently described a lack of referrals as a barrier (9% [public] vs 26% [private]). Eighty-eight percent of suites responded yes to the question "Would you like more referrals?" and 71% responded yes to "Would you like to be listed for referrals?" Sixty-three percent of suites reported that they were willing to submit a monthly colonoscopy report to the health department documenting the number of exams performed.
Discussion
We found that approximately 126,000 colonoscopies are being conducted annually in NYC hospitals, almost half (43%) of which are performed for screening purposes. Hospitals reported a potential to conduct an additional 69,100 procedures if barriers could be sufficiently addressed. In 2002, hospital-based endoscopy suites performed approximately one quarter of the estimated annual need of 200,000 screening exams. Effective hospital-based improvements could potentially double this number by enabling endoscopy suites to perform closer to maximum capacity. The actual total volume of colonoscopy procedures being conducted in NYC is unknown, because we currently lack data on the volume of procedures performed in outpatient office settings.
Public hospital endoscopy suites represented 16% of all suites surveyed, yet they conducted only 7% of reported procedures, suggesting a lower overall volume than in the private sector. Nonetheless, public hospitals also reported the capacity to increase the number of procedures they performed (12% of total additional capacity). Although this suggests that public facilities, not just private facilities, may be functioning below full capacity, the long waiting period to schedule a screening colonoscopy, the low number of exams performed (relative to private facilities), the more severe physician staffing shortages, and the more frequent patient cancellations all imply that public hospitals face more obstacles when striving to operate at maximum capacity.
Findings from this survey suggest that, if barriers were adequately addressed, NYC would have sufficient screening capacity in hospital endoscopy clinics to meet much of the demand generated by a focused colonoscopy campaign. Additional capacity is currently concentrated in private facilities, which primarily serve patient populations with health insurance. According to a population-based survey conducted in 2002, only about half of New Yorkers over the age of 50 reported ever having had a colonoscopy or sigmoidoscopy, leaving nearly 1 million adults at greater risk for undetected colon cancer (12). Hospitals and clinics should develop institutional measures to increase the number of persons receiving screening colonoscopy. Regular reminders to primary care physicians to refer patients for colonoscopy, rapid referral systems to expedite the referral process, protocols to bypass the initial visit with the endoscopist, and greater efficiency in colonoscopy procedures could increase the number of colonoscopy exams performed. Community-based organizations, advocacy groups, local government, and the medical community could advocate for legislative changes to increase reimbursement, reduce copays, and mandate insurance coverage for screening colonoscopy exams. These same groups should also work to increase public awareness and further educate providers about colorectal cancer screening. As a result of this study, the NYC DOHMH developed a colonoscopy surveillance system to track the volume of colonoscopy procedures performed in NYC hospitals; this system will allow the NYC DOHMH to monitor the impact of citywide efforts to increase screening colonoscopy rates.
Particular consideration should be given to increasing the number of colonoscopy exams performed on uninsured and low-income patients, who often face significant barriers to health care. Facilities may be able to improve the efficiency of endoscopy suites and decrease patient cancellations by the use of patient navigators (staff members designated to help patients negotiate complex public hospital systems). Shifting some of the need of the uninsured, low-income population to private hospitals may also provide a partial solution and could be accomplished through partnerships among local hospitals.
This study has limitations. We did not attempt to provide data on the complete universe of colonoscopy procedures in NYC, because information on the number and location of outpatient office settings where colonoscopy procedures are performed was unavailable. However, in NYC, colonoscopy procedures performed in hospital endoscopy suites do likely represent a significant proportion of all colonoscopy procedures. One study estimated that 25% of the estimated 35 million outpatient procedures performed nationwide in 2001 were performed in physicians' offices (13). In NYC, this proportion may be similar or even lower, due to the relatively high proportion of uninsured and Medicaid patients in NYC (approximately 19% of persons aged 50 and older in 2002) (14) and the high concentration of hospitals in NYC. Because of low reimbursement, colonoscopy procedures for Medicaid patients are generally only performed at hospital endoscopy suites by hospital-employed salaried physicians (15).
Another limitation was that our estimated citywide need for 200,000 annual screening colonoscopy exams was based on the 2000 census count and assumed a steady-state population structure. This was likely an overestimate, as we did not account for those who have medical contraindications to colonoscopy, those who have already been screened, and those who will absolutely refuse the procedure. This overestimate may be slightly offset by high-risk persons who require more frequent colonoscopy for surveillance; additionally, as more eligible persons undergo screening, more colonoscopy procedures will be needed to perform surveillance on those in whom polyps were identified.
Finally, respondents may have overestimated the annual maximum number of exams that could be performed. Respondents were not asked specifically to report whether they were considering any additional resource investment, such as staff or equipment, in estimating the maximum number of exams that could be performed and therefore may have based their estimate on endoscopy suite availability only.
Our study demonstrates a feasible method for obtaining data on colonoscopy volume and capacity at hospitals in an urban area, either as a one-time study or as a study repeated at regular intervals. In June 2003, the NYC DOHMH began collecting similar data on colonoscopy procedures on a quarterly basis. These data are used to assess the impact of ongoing agency efforts to increase the proportion of eligible persons who undergo screening colonoscopy exams.
In 2003, the NYC DOHMH created a coalition of key individuals and organizations that share an interest in decreasing the incidence and mortality of colon cancer, the Citywide Colon Cancer Control Coalition (C5). The mission of the C5 is to improve citywide colon cancer prevention and control by increasing awareness and screening. As part of these efforts, the C5/NYC DOHMH developed guidelines that state: "Most people 50 years of age and older should undergo colonoscopy every 10 years. Annual fecal occult blood testing (FOBT) is an acceptable, although not optimal, alternative for those unwilling or unable to undergo colonoscopy. Persons at high risk for colorectal cancer should begin screening with colonoscopy at age 40 or earlier" (2,3). Removing institutional barriers to increasing screening capacity, as well as improving access to care, are critical tasks for C5 coalition members to address; a concerted citywide effort is essential to achieving these goals.
The authors are indebted to the members of the Citywide Colon Cancer Control Coalition (C5) for their assistance in establishing guidelines for colorectal cancer screening in New York City, and for their efforts in advancing colon cancer prevention and control. C5 cochairpersons are Harold Freeman, MD (Ralph Lauren Center for Cancer Care and Prevention), and Sidney J. Winawer, MD (Memorial Sloan-Kettering Cancer Center). C5 committee chairs include: Barbara Barrie, Maurice Cerulli, MD (New York Methodist Hospital); Alvaro Genao, MD (North General Hospital); Steven H. Itzkowitz, MD (Mount Sinai School of Medicine); Alfred I. Neugut, MD, PhD (Columbia University Medical Center, New York Presbyterian Hospital); Mark B. Pochapin, MD (Jay Monahan Center for Gastrointestinal Health, New York Presbyterian Hospital, Weill Medical College of Cornell University); Moshe Shike, MD (Memorial Sloan-Kettering Cancer Prevention and Wellness Program, Memorial Sloan-Kettering Cancer Center); Robert Schiller, MD (Beth Israel Medical Center, Institute for Urban Family Health); Thomas Weber, MD (Montefiore Medical Center).
We also thank Drs Kelly Henning and Sidney J. Winawer for their insightful comments on this manuscript.
Figures and Tables
Table 1 Characteristics of Surveyed Endoscopy Suites in New York City Hospitals, 2003
No. (%)
Hospitals contacteda 69
Hospitals performing colonoscopy 66
Additional affiliated clinics 2
Endoscopy suites surveyed 68 (100)
Voluntary/VAb 57 (84)
Type of endoscopy suite
Inpatient and outpatientc 55
Outpatient, free-standingd 2
Public 11 (16)
Type of endoscopy suite
Inpatient and outpatient 11
a One hospital was a hospital center consisting of two separate hospital units; this was analyzed as one hospital.
b VA = Veterans Administration.
c One hospital used both an inpatient/outpatient suite and an operating room to perform colonoscopy exams.
d One outpatient, free-standing endoscopy suite was actually an outpatient office practice affiliated with a hospital.
Table 2 Reported and Estimated Volume of Colonoscopy Exams Performed at New York City Hospitals, 2002, and Estimated Potential Capacity
All endoscopy suites N = 68 Voluntary/VAa N = 57 Public N = 11
Reported number of exams performed, 2002b
Median (range) 1290 (260-7000) 1507 (260-7000) 821 (400-1450)
Estimated total 126,000 117,200 8800
Estimated number of exams performed for screening, 2002c
Midpoint (range) 53,600 (37,800-69,000) 49,800 (35,200-64,200) 3800 (2700-4800)
Reported maximum number of exams per monthd
Median (range) 200 (30-800) 200 (30-800) 150 (45-320)
Estimated total 16,300 14,800 1400
Estimated annual maximum number of examse 195,200 177,800 17,400
Estimated annual potential residual capacityf 69,100 60,600 8500
a VA = Veterans Administration.
b Two voluntary/VA and one public endoscopy suite responded “don't know” to this question.
c Data missing from two voluntary/VA and one public endoscopy suite.
d Four voluntary/VA and two public endoscopy suites responded “don't know” to this question.
e Values were rounded to the nearest hundred; numbers do not add up to the “Reported maximum number of exams per month" x 12.
f Values were rounded to the nearest hundred; numbers do not add up to the “Reported number of exams performed in 2002” subtracted from “Estimated annual maximum number of exams.”
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Leng JCF, Thorpe LE, Feldman GE, Thomas PA, Frieden TR. The volume and capacity of colonoscopy procedures performed at New York City hospitals in 2002. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0032.htm
==== Refs
1 The City of New York 2002 2 Summary of vital statistics 2000 Department of Health, Office of Vital Statistics New York
2 Feldman GE McCord CW Frieden TR 22 2 2003 1 4 Preventing colorectal cancer New York City Department of Health and Mental Hygiene – City Health Information New York
3 Feldman GE McCord CW Bassett MT Frieden TR 290 2 2003 191 JAMA Screening for colorectal cancer – letter to the editor 12851267
4 Smith RA Cokkinides V Eyre HJ American Cancer Society 53 1 2003 27 43 CA Cancer J Clin American Cancer Society guidelines for the early detection of cancer, 2003 12568442
5 U.S. Preventive Services Task Force 7 2002 Screening for colorectal cancer: recommendations and rationale Agency for Healthcare Research and Quality Rockville (MD)
6 Walsh JM Terdiman JP 289 10 2003 1288 1296 JAMA Colorectal cancer screening: scientific review 12633191
7 Winawer S Fletcher R Rex D Bond J Burt R Ferrucci J 124 2 2003 544 560 Gastroenterology Colorectal cancer screening and surveillance: clinical guidelines and rationale-update based on new evidence 12557158
8 Winawer SJ Zauber AG Ho MN O'Brien MJ Gottlieb LS Sternberg SS 329 27 1993 1977 1981 N Engl J Med Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup 8247072
9 Kolata G The New York Times 2003 12 8 50 and ready for colonoscopy? Doctors say wait is often long
10 CNN Money Best places to live, New York, NY Atlanta (GA) Cable News Network LP, LLLP 2004 Available from: URL: http://money.cnn.com/best/bplive/details/3651000.htm
11 The Greater New York Hospital Association (GNYHA) 2002 Membership Directory 2002, and Veterans Administration hospitals GNYHA New York
12 Thorpe LE Mostashari F Feldman G Karpati AM Cobb LK Helgerson SD 2 2 2003 1 4 NYC Vital Signs Cancer screening in New York City: we can do much better
13 Andrews M 2002 8 18 The New York Times In-office surgery: fewer rules apply
14 New York City Department of Health and Mental Hygiene 2002 NYC Community Health Survey 2002 Accessed 2004 Jul 7 New York NYC Department of Health
15 Leng Jennifer 2004 7 Conversation with: M Cerulli, MD
|
15670462
|
PMC1323312
|
CC BY
|
2022-01-24 23:41:42
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A09
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0069
Original Research
PEER REVIEWEDDental Care Use Among Pregnant Women in the United States Reported in 1999 and 2002
Eke Paul I PhD, MPH Surveillance, Investigations, and Research Team, Division of Oral Health (DOH), Centers for Disease Control and Prevention (CDC)
4770 Buford Highway, Mail Stop F-10, Chamblee, GA 30341 [email protected]
770-488-6092
Timothé Peggy DDS, MPH Program Services Team formerly research fellow of the Association of Schools of Public Health, CDC, Chamblee, Ga
Presson Scott M DDS, MPH Program Services Team, DOH, CDC, Chamblee, Ga
Malvitz Dolores M DrPH Surveillance, Investigations, and Research Team, DOH, CDC, Chamblee, Ga
1 2005
15 12 2004
2 1 A102005
Introduction
The purpose of this study was to determine national and state-specific estimates of dental care use among adult pregnant women in the United States using data from two 12-month periods. The study also determined person-level characteristics that may predict a lack of dental care use within this subgroup.
Methods
Responses were analyzed from 4619 pregnant women aged 18 to 44 years who participated in the 1999 and 2002 state-based Behavioral Risk Factor Surveillance System. Dental care use was defined as having a dental visit or a dental cleaning in the 12 months preceding the interview. State-specific estimates were adjusted to the 2000 U.S. population distribution. Multivariable regression analysis was used to evaluate person-level characteristics that may predict not obtaining dental care during this period.
Results
Overall, 70% of pregnant women in 1999 and 2002 had received dental care in the previous 12 months. Age-adjusted estimates ranged from 36% (Nevada) to 89% (Vermont) to 91% (Puerto Rico). In 19 states, 75% or more of pregnant women had obtained dental care in the previous 12 months (age-adjusted figure). Most pregnant women with dental care were non-Hispanic white and married, and they had a greater than high school education. Income and smoking status were significant predictors for not using dental care.
Conclusion
In several states, more than 70% of pregnant women reported a dental visit or dental cleaning during the previous 12 months. Relative to the general population, pregnant women are as likely to receive dental care, but certain subgroups need to do much better. However, these estimates may be biased toward a population with a higher socioeconomic status and may not represent dental care use among pregnant women in the general U.S. population.
==== Body
Introduction
An estimated 6 million women in the United States become pregnant each year (1). Although preventive dental care (e.g., dental cleaning) will improve the overall health of pregnant women and may reduce their risk of adverse pregnancy outcomes, women who are pregnant are known to use dental services less frequently and at lower levels than the general population (2-4). An interrelated set of financial, personal, and social barriers have been identified as possible reasons why subgroups in most need of dental care may be less likely to receive dental care services (5).
Evidence is increasing that poor oral health may be associated with adverse pregnancy outcomes. Several observational studies have reported associations between periodontal infections and increased risk for poor birth outcomes, such as preterm labor or premature rupture of membranes (6-8). These findings are further supported by experimental animal studies that found maternal exposure to periodontal pathogens resulted in abnormal fetal outcomes (9,10). Preliminary findings from intervention studies also suggest that treatment of advanced periodontal infections may reduce the risk of adverse birth outcomes (11,12).
Currently, information is limited at the national and state levels on patterns of dental care use, particularly dental cleaning, among pregnant women. The current literature is limited to estimates from five states (Louisiana, Illinois, New Mexico, Arkansas, Washington) participating in the Pregnancy Risk Assessment Monitoring System (PRAMS); the proportion of new mothers who received dental care during their most recent pregnancy ranged from 23% to 58% in these five states (13,14).
The purpose of the present study was to determine national and state-specific estimates of dental care use (i.e., having a dental visit or a dental cleaning) during two 12-month periods among pregnant women aged 18 to 44 years in the United States. These estimates were generated after combining data obtained in 1999 and 2002 by the Behavioral Risk Factor Surveillance System (BRFSS). In addition, this study examined person-level characteristics that predicted not obtaining dental care during this period.
Methods
The BRFSS is a random, state-based telephone survey of major health risk behaviors, clinical preventive health practices, and health care access that relies on a representative sample of noninstitutionalized adults (aged >18) in the 50 states, District of Columbia, Guam, Puerto Rico, and the U.S. Virgin Islands. Details of the survey are available elsewhere (15).
All female participants (aged ≤44 years) in the BRFSS survey are asked about their pregnancy status with the question, "To your knowledge, are you now pregnant?" In 1999 and 2002, three oral health questions were included in the core module, asked of all participants: 1) "How long has it been since you last visited a dentist or a dental clinic for any reason?" 2) "How many of your permanent teeth have been removed because of tooth decay or gum disease?" and 3) "How long has it been since you had your teeth cleaned by a dentist or dental hygienist?" In the present study, dental care use was defined as having either a dental visit or dental cleaning within the preceding 12 months. BRFSS data for 1999 and 2002 were pooled to increase the samples of pregnant women at the state levels. Analysis was restricted to the dentate, and missing data or persons not responding to the questions were removed from the denominator (<1%). The average nonresponse rate combined across the various characteristics examined in this analysis was 0.80%.
Analysis by SUDAAN (Research Triangle Institute, Triangle Park, NC) (16) was used to account for the complex sampling design of the survey and sampling weights. In separate analyses, estimates were age-adjusted based on the U.S. census population distribution of persons aged ≥18 years in 2000 (17) to provide a sounder basis for comparing estimates among states (18). Estimates of dental care use by pregnant women were stratified by age, level of education, diabetes status, health insurance status, income, marital status, smoking status, and race/ethnicity (Hispanic is a category of ethnicity that may include women of all races). Logistic regression modeling was used to examine characteristics that were significant predictors of pregnant women not receiving dental services within the preceding 12 months, adjusting for other potential explanatory variables. Covariates in the model were selected a priori based on previous evidence that the variable was associated with dental care use and that measures of the variable were available in BRFSS.
Results
National and state estimates for dental care use in the past 12 months among pregnant and nonpregnant dentate women are shown in Table 1. The national estimate for pregnant women having a dental visit or cleaning in the previous 12 months, age-adjusted to the 2000 population, was 70.03% (SE = 1.46%), with state percentages ranging from 48.32% (Nevada) to 87.02% (Vermont). Estimates for nonpregnant women ranged from 62.77% (Texas) to 84.14% (Connecticut). When age-adjusted to the 2000 population, estimates ranged from 36.16% (Nevada) to 91.34% (Puerto Rico). In 19 states, the age-adjusted estimates were 75% or greater.
The distribution of age, race/ethnicity, marital status, education, household income, health insurance status, and smoking status was similar when all pregnant women were compared with those receiving dental care in the past 12 months (Table 2). Most pregnant women reporting a dental visit or cleaning in the preceding 12 months were non-Hispanic white, married, between the ages of 20 and 34 years, and educated beyond high school. In addition, most had health insurance (90.92%, SE = 0.91).
Compared with pregnant women who received dental services, those not receiving dental care were more likely to be aged 20 to 34 years, be active smokers (smoking every day or some days), have less than a high school education, and have diabetes (excluding gestational diabetes). These respondents were also less likely to be married and to have health insurance. Pregnant women who reported not having had dental care in the preceding 12 months were twice as likely to lack health insurance and to use public health clinics or hospital outpatient services. In multivariable logistic modeling, only household income and smoking were significant predictors for not reporting dental services in the previous 12 months (Table 3).
Discussion
This study reports the first national and state estimates for dental care use during two 12-month periods among pregnant women in the United States. Estimates were obtained from a representative sample of pooled data from the BRFSS in 1999 and 2002. For most states, the BRFSS is the only source of information on dental care use and risk factors for chronic diseases. Analyzing this combined dataset, we find about 70% of pregnant women in the U.S. had either a dental visit or a dental cleaning within the preceding 12 months. This estimate was similar to estimates for the general U.S. population (BRFSS 2002 estimate 70.8%).
In the general population, behaviors related to use of dental care are known to be related to demographic characteristics such as level of education and ethnicity (2,19). Among personal characteristics of pregnant women examined in this study, only income and smoking status were significant predictors for not obtaining dental care in the previous 12 months, suggesting that low-income pregnant women may be at higher risk for not receiving dental care. This finding is consistent with reports from PRAMS that pregnant women receiving no dental care were more likely to use tobacco (14). Because low-income women are more likely to smoke, smoking in this subpopulation may be a proxy for low income. Pregnant women who did not receive dental care were skewed toward a younger age, probably because younger women are more likely to have lower incomes. The BRFSS does not collect information on dental insurance, parity, or perceived fears of harm to the fetus, which are important determinants in whether pregnant women obtain dental care. Previous reports also suggest that low-income pregnant women are less likely than their higher-income counterparts to visit the dentist (20). We found that 95% of pregnant women who reported a dental visit in the previous 12 months also had a dental cleaning during that period.
State-specific estimates of dental care use in the previous 12 months among pregnant women varied greatly and generally followed the dental use pattern of the overall population of women aged 18 to 44 in each state. We found relatively higher percentages in states or territories with aggressive preventive dental care programs for pregnant women, such as Puerto Rico. Lower estimates for pregnant women were seen in states such as Virginia, Nevada, and Arkansas and were consistent with lower estimates of dental care use in the general population of these states. It is unclear what factors most influence variation by state. However, the number of community centers with a dentist or dental health program is an important explanatory factor for dental care use among low-socioeconomic status (SES) populations.
Importantly, these estimates do not represent the percentage of women reporting dental care use while pregnant. Depending on the term of pregnancy when surveyed by the BRFSS, there would be a period in the 12 months preceding the interview when women were not pregnant. Health care providers and dentists treat women differently according to pregnancy status, and pregnant women seek dental care differently. A relatively higher or lower percentage of dental visits or cleaning when not pregnant would skew these estimates up or down, respectively.
Notably, state-specific estimates from this study were higher than those published previously from the five states that participated in PRAMS, which ranged from 23% to 58% (11,14). Several factors may account for these differences. First, in PRAMS, questions on use of dental care were restricted to the period when pregnant. Second, while the BRFSS included only adults (i.e., those aged ≥18 years), PRAMS includes all pregnant women (i.e., including those <18 years) and over-samples two or three characteristics, typically low SES. Finally, the BRFSS is a telephone survey and probably includes a higher SES population than PRAMS.
Some limitations should be noted in the use of the BRFSS to obtain estimates for dental care use among pregnant women. First, the survey is based on self reports, which can be influenced by recall bias. Self-reported dental care, however, has been found to be a valid measure for dental care use given adequate sample size and study design (21). Second, the BRFSS is a telephone survey that generally excludes women without residential phones; therefore, the survey might exclude persons of lower SES or households with only cellular phones. Finally, because a relatively small percentage of women are pregnant at any time, samples for pregnant women in most states often were small, sometimes less than 50. We pooled data for 1999 and 2002 to increase the samples and improve estimate reliability, but even then, samples for Maine, Mississippi, and the District of Columbia were small (less than 50), and so estimates from these states may be considered less reliable.
Because preventive dental care may reduce risk for adverse pregnancy outcomes, we must assess how current patterns of dental care use among pregnant women compare with those of the general population. Estimates from this study suggest that dental care use in the previous 12 months among pregnant women is about the same in the general population; in both populations, indicators of SES appear to be important predictors of not using services for those persons (approximately 30%) who have not recently had any dental care. However, we note that lack of health insurance, use of public health clinics, and the use of hospital outpatient services were twice as likely among pregnant women not reporting dental care. One approach to reduce lack of dental care among pregnant women may include providing health insurance. Additionally, health care providers in these health care settings are more likely to come in contact with pregnant women who do not receive dental care. This may present an opportunity to provide important oral health education to these pregnant women.
Barriers to obtaining dental cleaning need to be explored further and be better understood. One approach to addressing dental care use could involve prenatal and professional education on the importance of dental care and the adverse effects of smoking during pregnancy. Overall, these estimates provide baseline information on dental visits and cleaning among pregnant women in the United States and may be useful in formulating oral health policies and programs to improve the health and well-being of pregnant women.
Figures and Tables
Table 1 Distribution of Dental Care Use Among U.S. Pregnant Women Aged 18–44, Behavioral Risk Factor Surveillance System, 1999 and 2002
State Pregnant, Used Dental Care
% (95% Confidence Interval) Pregnant, Used Dental Care (Age- Adjusted)
% (95% Confidence Interval)
Alabama 73.76 (59.5-88.1) 72.63 (51.4-93.8)
Alaska 67.64 (53.3-81.9) 65.40 (48.9-81.9)
Arizona 53.60 (34.0-73.2) 48.41 (30.0-66.8)
Arkansas 60.47 (47.1-73.8) 57.07 (27.6-88.0)
California 75.64 (67.0-84.3) 73.15 (63.4-83.0)
Colorado 67.41 (55.7-79.2) 64.66 (51.8-77.6)
Connecticut 76.86 (66.5-87.2) 75.06 (62.0-88.2)
Delaware 83.38 (70.2-96.5) 85.48 (75.7-95.3)
District of Columbia 85.91 (74.9-96.9)a 83.71 (70.8-96.6)a
Florida 72.11 (63.1-81.1) 46.49 (37.3-55.7)
Georgia 79.33 (68.7-89.9) 82.54 (72.5-92.5)
Hawaii 72.83 (59.5-86.2) 72.77 (57.9-87.7)
Idaho 70.52 (61.1-79.9) 69.91 (53.8-86.0)
Illinoisb 76.65 (67.0-86.3) 75.82 (65.0-86.6)
Indiana 64.94 (50.2-79.6) 56.86 (39.1-74.7)
Iowa 73.85 (62.3-85.4) 77.44 (63.5-91.3)
Kansas 70.84 (61.4-80.2) 73.44 (62.2-84.6)
Kentucky 70.10 (53.2-87.0) 58.46 (40.9-76.1)
Louisiana 82.93 (75.7-90.2) 83.56 (75.4-91.8)
Maine 79.43 (63.4-95.5)a 86.45 (76.3-96.7)a
Maryland 75.49 (65.1-85.9) 73.82 (59.5-88.1)
Massachusetts 75.06 (66.0-84.1) 76.62 (67.6-85.6)
Michigan 79.73 (70.0-88.7) 74.93 (63.5-86.3)
Minnesota 80.67 (73.6-87.7) 81.22 (73.6-88.8)
Mississippi 67.07 (51.6-82.6)a 60.35 (36.7-84.1)a
Missouri 64.02 (50.7-77.3) 64.12 (44.5-83.7)
Montana 74.67 (60.0-89.4) 75.95 (60.3-91.7)
Nebraska 74.38 (64.6-84.2) 80.10 (72.3-87.9)
Nevada 48.32 (30.3-66.4) 36.16 (19.0-53.4)
New Hampshire 71.06 (55.6-86.5) 74.30 (60.8-87.8)
New Jersey 80.18 (66.9-93.5) 81.45 (68.4-94.6)
New Mexico 63.48 (51.1-75.8) 66.35 (49.7-83.1)
New York 70.97 (59.4-82.5) 71.59 (58.7-84.5)
North Carolina 70.54 (58.0-83.1) 77.82 (67.6-88.0)
North Dakota 73.24 (60.1-86.4) 56.43 (36.0-76.8)
Ohio 75.91 (63.6-88.3) 70.35 (53.2-87.6)
Oklahoma 68.70 (58.9-78.5) 64.88 (52.2-77.6)
Oregon 68.78 (55.6-81.9) 64.44 (44.8-84.0)
Pennsylvania 81.71 (73.5-89.9) 83.77 (73.4-94.2)
Rhode Island 80.25 (67.5-93.0) 80.95 (69.4-92.6)
South Carolina 75.65 (63.7-87.6) 83.70 (75.9-91.5)
South Dakota 73.78 (63.8-83.8) 74.91 (63.7-86.1)
Tennessee 71.76 (59.0-84.5) 71.94 (57.4-86.4)
Texas 66.02 (57.0-75.0) 65.27 (52.4-78.2)
Utah 75.85 (66.1-85.7) 66.18 (45.6-86.8)
Vermont 87.02 (78.6-95.4) 89.06 (80.5-97.7)
Virginia 56.02 (41.3-70.7) 51.51 (37.2-65.8)
Washington 71.73 (62.7-80.7) 69.53 (59.3-79.7)
West Virginia 73.47 (59.8-87.2) 77.04 (61.3-92.7)
Wisconsin 84.42 (74.6-94.2) 75.41 (59.5-91.3)
Wyoming 60.36 (43.3-77.4) 58.86 (39.9-77.9)
Guamc 59.83 (24.2-95.5)a 79.58 (61.0-98.2)a
Puerto Rico 86.31 (77.5-95.1) 91.34 (85.8-96.8)
Virgin Islandsc 72.52 (53.9-91.1) a 76.53 (59.1-93.9)a
United States 71.16 (69.0-73.3) 70.03 (67.1-72.9)
a Sample size <50.
b Estimate based on half of sampled population because the state used the dual questionnaire method.
c Estimates from 2002 survey only.
Table 2 Distribution of Person-level Characteristics for U.S. Pregnant Women Aged 18-44, Behavioral Risk Factor Surveillance System, 1999 and 2002
Characteristic All
% (SE) Used Dental Care
% (SE) Did Not Use Dental Care
% (SE)
Age (years)a n=4619 n=3393 n=1226
18-19 8.49 (0.87) 8.74 (1.06) 7.88 (1.53)
20-24 26.29 (1.19) 23.99 (1.36) 31.96 (2.37)
25-29 27.37 (1.05) 28.37 (1.27) 24.90 (1.85)
30-34 23.40 (0.99) 25.16 (1.22) 19.07 (1.64)
35-39 11.24 (0.74) 11.24 (0.84) 11.27 (1.53)
40-44 3.20 (0.45) 2.50 (0.35) 4.92 (1.32)
Race/ethnicity n=2684 n=1979 n=705
Non-Hispanic white 60.71 (1.83) 62.92 (2.17) 55.33 (3.23)
Non-Hispanic black 11.19 (1.23) 10.05 (1.29) 13.96 (2.76)
Other non-Hispanic 6.09 (0.87) 5.15 (0.97) 8.39 (1.83)
Multi non-Hispanic 1.49 (0.37) 1.60 (0.48) 1.23 (0.51)
Hispanic (includes all races) 20.51 (1.81) 20.27 (2.18) 21.09 (3.08)
Marital statusa n=4612 n=3389 n=1223
Married 71.30 (1.24) 73.90 (1.39) 64.88 (2.34)
Divorced 2.97 (0.34) 2.17 (0.29) 4.94 (0.91)
Widowed 0.15 (0.06) 0.14 (0.08) 0.18 (0.10)
Separated 1.42 (0.23) 1.37 (0.24) 1.53 (0.55)
Never married 17.18 (0.99) 16.16 (1.14) 19.71 (1.96)
Unmarried couple 6.97 (0.76) 6.25 (0.79) 8.77 (1.54)
Education levela n=4619 n=3393 n=1226
Less than high school 14.47 (1.14) 12.64 (1.27) 18.99 (2.10)
High school 28.31 (1.19) 27.60 (1.39) 30.05 (2.30)
Greater than high school 57.22 (1.34) 59.76 (1.58) 50.96 (2.42)
Annual household income ($)a n=4055 n=2989 n=1066
<10,000 5.25 (0.60) 4.95 (0.72) 6.01 (1.06)
10,000-14,999 6.46 (0.81) 5.74 (1.02) 8.29 (1.30)
15,000-19,999 10.43 (0.94) 9.36 (1.14) 13.13 (1.76)
20,000-24,999 9.75 (0.78) 8.57 (0.85) 12.73 (1.79)
25,000-34,999 15.98 (0.97) 14.41 (1.05) 19.95 (2.11)
35,000-49,999 17.76 (0.98) 18.34 (1.10) 16.27 (1.98)
50,000-74,999 16.84 (0.90) 18.84 (1.11) 11.77 (1.55)
75,000 or more 17.54 (0.93) 19.79 (1.12) 11.84 (1.58)
Diabetes (not gestational) n=4417 n=3249 n=1168
Yes 1.08 (0.25) 0.75 (0.18) 1.91 (0.77)
No 98.92 (0.25) 99.25 (0.18) 98.09 (0.77)
Health insurance statusa n=4613 n=3390 n=1223
Yes 88.37 (0.84) 90.92 (0.91) 82.07 (1.84)
No 11.63 (0.84) 9.08 (0.91) 17.93 (1.84)
Smoking statusa n=4612 n=3388 n=1224
Yes (every day) 8.74 (0.72) 7.82 (0.78) 11.02 (1.54)
Yes (some days) 3.02 (0.42) 2.61 (0.49) 4.05 (0.83)
Former 21.66 (1.03) 20.93 (1.13) 23.46 (2.06)
Never 66.57 (1.18) 68.64 (1.32) 61.47 (2.37)
Health care accessa n=4617 n=3390 n=1223
Doctor’s office 74.5(1.6) 79.2(1.8) 61.9(3.5)
Public health clinic 8.0(1.1) 6.2(1.1) 12.8(2.6)
Hospital outpatient 3.4(0.9) 2.3(0.5) 6.3(2.9)
Hospital emergency room 4.3(0.8) 3.7(0.9) 5.9(1.5)
Urgent care center 2.7(0.7) 2.6(0.8) 2.9(0.9)
Some other kind of place 1.3(0.5) 1.8(0.4) 2.4(0.8)
No usual place 3.9(3.0) 3.0(0.6) 6.3(1.3)
a Significant at P ≤ .05, based on chi-square test for independence of association between characteristic and dental care use among pregnant women.
Table 3 Possible Predictors for U.S. Pregnant Women Not Having a Dental Visit or Cleaning (N=2226)a
Characteristics Odds Ratio (95% CI) P
Age (years)
⩽19 0.32 (0.10-0.99) .37
20-24 0.86 (0.39-1.91)
25-29 0.84 (0.38-1.83)
30-34 0.75 (0.34-1.69)
35-39 0.78 (0.31-1.97)
40-44 1.00 (ref)
Race/ethnicity
Non-Hispanic white 1.06 (0.57-1.96) .09
Non-Hispanic black 1.71 (0.81-3.62)
Other Non-Hispanic races 1.77 (0.70-4.52)
Non-Hispanic multiracial 0.26 (0.06-1.26)
Hispanic 1.00 (ref)
Marital status
Divorced 1.74 (0.83-3.67) .29
Widowed 0.08 (0.01-1.03)
Separated 0.91 (0.30-2.73)
Never married 1.05 (0.63-1.75)
Unmarried couple 1.04 (0.53-2.02)
Married 1.00 (ref)
Education
<High school 0.63 (0.35-1.34) .27
High school 1.13 (0.76-1.69)
⩾College 1.00 (ref)
Annual income ($)
<10,000 0.65 (0.26-1.61) .047
10,000-14,999 1.34 (0.57-3.19)
15,000-19,999 1.75 (0.70-4.37)
20,000-24,999 1.29 (0.62-2.68)
25,000-34,999 1.43 (0.78-2.60)
35,000-49,999 1.43 (0.80-2.56)
50,000-74,999 0.76 (0.43-1.36)
>75,000 1.00 (ref)
Diabetic status
Yes 2.49 (0.72-8.54) .15
No 1.00 (ref)
Health insurance
Yes 0.69 (0.42-1.14) .15
No 1.00 (ref)
Where you get health care
Doctor’s office 0.43 (0.19-1.00) .08
Public health clinic or community health center 0.92(0.35-2.43)
Hospital outpatient 1.37 (0.38-4.91)
Hospital emergency room 0.88 (0.29-2.67)
Urgent care center 0.65 (0.22-1.87)
Other kind of place 0.50 (0.16-1.63)
Don’t know 0.96 (0.19-4.82)
No usual place 1.00 (ref)
Smoking
Current smoker (every day) 1.53 (0.92-1.14) .03
Current smoker (some days) 2.89 (1.35-6.17)
Former smoker 1.16 (0.80-1.67)
Never smoked 1.00 (ref)
a In this table, all characteristics presented in Table 2 were further evaluated in a multivariable logistic regression model to determine which characteristics retained significant associations with whether or not a pregnant woman received dental care in the last 12 months. Only annual income and smoking were significant at P ≤ .05. CI = confidence interval; ref = reference group.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Timothé P, Eke PI, Presson SM, Malvitz DM. Dental care use among pregnant women in the United States reported in 1999 and 2002. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0069.htm
==== Refs
1 Ventura SJ Mosher WD Curtin SC Abma JC Henshaw S 49 4 2001 1 9 Natl Vital Stat Rep Trends in pregnancy rates for the United States, 1976-97: an update
2 U.S. Department of Health and Human Services 2000 Rockville (MD) Oral health in America: report of the Surgeon General U.S. Department of Health and Human Services, National Institute of Dental and Craniofacial Research, National Institutes of Health
3 Bolden AJ Henry JL Allukian M 57 12 1993 888 900 J Dent Educ Implications of access, utilization and need for oral health care by low income groups and minorities on the dental delivery system 8263236
4 Atchison KA Davidson PL Nakazono TT 11 2 1997 223 234 Adv Denl Res Predisposing, enabling, and need for dental treatment characteristics of ICS-II USA ethnically diverse groups
5 Gilbert GH Shelton BJ Chavers LS Bradford EH Jr 41 1 2003 119 134 Med Care The paradox of dental need in a population-based study of dentate adults 12544549
6 Offenbacher S Katz V Fertik G Collins J Boyd D Maynor G 67 10 Suppl 1996 1103 1113 J Periodontol Periodontal infection as a possible risk factor for preterm low birth weight 8910829
7 Dasanayake AP 3 1998 206 212 Ann Periodontol Poor periodontal health of the pregnant women as a risk factor for low birth weight 9722704
8 Offenbacher S Beck JD Lieff S Slade G 62 10 1998 852 858 J Dent Educ Role of periodontitis in systemic health: spontaneous preterm birth 9847888
9 Collins JG Smith MA Arnold RR Offenbacher S 62 10 1994 4652 4655 Infect Immun Effects of Escherichia coli and Porphyromonas gingivalis lipopolysaccharide on pregnancy outcome in the golden hamster 7927735
10 Collins JG Windley HW 3rd Arnold RR Offenbacher S 62 10 1994 4356 4361 Infect Immun Effects of a Porphyromonas gingivalis infection on inflammatory mediator response and pregnancy outcome in hamsters 7927695
11 Lopez NJ Smith PC Gutierrez J 73 8 2002 911 924 J Periodontol Periodontal therapy may reduce the risk of pre-term low birth weight in women with periodontal disease: a randomized controlled trail 12211502
12 Jeffcoat MK Geurs NC Reddy MS Cliver SP Goldenberg RL Hauth JC 2001 132 875 880 J Am Dent Assocurce Periodontal infection and preterm birth: results of a prospective study
13 Gaffield ML Gilbert BJ Malvitz DM Romaguera R 132 7 2001 1009 1016 J Am Dent Assoc Oral health during pregnancy: an analysis of information collected by the pregnancy risk assessment monitoring system 11480627
14 Lydon-Rochelle M Krakowiak P Hujoel P Peters RM 2004 94 765 771 Am J Public Health Dental care use and self-reported dental pregnancy problems in relation to pregnancy 15117698
15 Mokdad AH Stroup DF Giles WH RR09 1 12 Morb Mortal Wkly Rep Public health surveillance for behavioral risk factors in a changing environment: Recommendation from the Behavioral Risk Factor Surveillance Team
16 Shah BV Barnwell BG Bieler GS Research Triangle Park (NC) Research Triangle Institute 1996 Research Triangle Institute; 1996 SUDAAN: software for the analysis of correlated data. User's manual release 7.00
17 Perry MJ Mackun PJ 2001 U.S. Department of Commerce, U.S. Census Bureau Washington (DC) Population change and distribution: 1990 to 2000
18 Klein RJ Schoenborn CA Healthy People 2010 Stat Notes Age adjustment using the 2000 projected U.S. population 2001 20 1 10 11676466
19 Gilbert GH Shah GR Shelton BJ Heft MW Bradford EH Chavers LS 2002 37 1487 1507 Health Serv Res Racial differences in predictors of dental care use 12546283
20 Mangskau KA Arrindell B 75 6 1996 23 28 Northwest Dent Pregnancy and oral health: utilization of the oral health care system by pregnant women in North Dakota 9487880
21 Gilbert GH Rose JS Shelton BJ 2002 30 352 362 Community Dent Oral Epidemiol A prospective study of the validity of data on self-reported dental visits 12236826
|
15670463
|
PMC1323313
|
CC BY
|
2022-01-31 23:19:00
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A10
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0053
Original Research
PEER REVIEWEDDirect and Indirect Costs of Asthma in School-age Children
Wang Li Yan MBA, MA Surveillance and Evaluation Research Branch, Division of Adolescent and School Health (DASH), National Center for Chronic Disease Prevention and Health Promotion (NCCDPHP), Centers for Disease Control and Prevention (CDC)
4770 Buford Hwy, Mail Stop K-33, Chamblee, GA 30341 [email protected]
770-488-6195
Zhong Yuna MD, MSPH DASH, NCCDPHP, CDC, Atlanta, Ga
Wheeler Lani MD DASH, NCCDPHP, CDC, Annapolis, Md
1 2005
15 12 2004
2 1 A112005
Introduction
Asthma is one of the most common chronic diseases of childhood and is the most common cause of school absenteeism due to chronic conditions. The objective of this study is to estimate direct and indirect costs of asthma in school-age children.
Methods
Using data from the 1996 Medical Expenditure Panel Survey, we estimated direct medical costs and school absence days among school-age children who had treatment for asthma during 1996. We estimated indirect costs as costs of lost productivity arising from parents’ loss of time from work and lifetime earnings lost due to premature death of children from asthma. All costs were calculated in 2003 dollars.
Results
In 1996, an estimated 2.52 million children aged five to 17 years received treatment for asthma. Direct medical expenditure was $1009.8 million ($401 per child with asthma), including payments for prescribed medicine, hospital inpatient stay, hospital outpatient care, emergency room visits, and office-based visits. Children with treated asthma had a total of 14.5 million school absence days; asthma accounts for 6.3 million school absence days (2.48 days per child with asthma). Parents’ loss of productivity from asthma-related school absence days was $719.1 million ($285 per child with asthma). A total of 211 school-age children died of asthma during 1996, accounting for $264.7 million lifetime earnings lost ($105 per child with asthma). Total economic impact of asthma in school-age children was $1993.6 million ($791 per child with asthma).
Conclusion
The economic impact of asthma on school-age children, families, and society is immense, and more public health efforts to better control asthma in children are needed.
==== Body
Introduction
One of the most common chronic disorders in children and adolescents (1), asthma represents a major public health problem of increasing concern in the United States. Between 1980 and 1996, the prevalence of asthma increased by an average of 4.3% per year, from 3.6% to 6.2% among children aged from birth to seven years (2). Since then, asthma prevalence appears to be stable, and in 2001, more than 5 million children aged five to 17 in the United States were reported to have a current diagnosis of asthma (3). Previously published reports (4-8) strongly suggest that asthma not only increases health care use and costs but also places a large burden on affected children and their families. Children with asthma miss out on school, sports, and other childhood activities. Parents or caregivers of children with asthma are affected by missed workdays and decreased job productivity.
In 1992, Weiss et al (4) conducted a study of the asthma-related costs in the U.S. population. They derived estimates of direct medical costs and indirect costs of productivity loss (in 1985 dollars) using data from national surveys, including the National Health Interview Survey (NHIS), the National Hospital Discharge Survey, the National Ambulatory Medical Care Survey, and others. Direct medical expenditures were estimated to be $465 million for children with asthma aged from birth to 17 years, and indirect costs (value of parents' or caregivers' productivity loss associated with school absence days [SADs]) were estimated to be $726 million among school-age children. Because none of the national surveys collects diagnostic and expenditure data, only national expenditure estimates were produced. Although such estimates enable policymakers to understand the economic impact of childhood asthma in the United States, it is necessary to assess the per capita costs for children with asthma to determine the savings or benefits of a successful asthma intervention.
Using the 1987 Medical Expenditure Panel Survey (MEPS), a later study conducted by Lozano et al (6) estimated the per capita health care costs for children with asthma aged one to 17 years, including both asthma-related expenditures only and all care expenditures (both asthma and nonasthma care). For children with asthma aged one to 17 years, the mean asthma-related per capita expenditures totaled $171 per year; the mean all-care per capita expenditures totaled $1129 ($468 for children without asthma). However, those annual cost estimates were based on parent-reported prevalence. Although 8.8% of the study sample was classified as children with asthma (had asthma or wheezing during the past 12 months), 56% reported taking no asthma medication, and 12.3% reported no health care use. To determine the reduction in health costs that would result from an asthma intervention, cost estimates based on treated prevalence or attack prevalence are more appropriate than those using proxy-reported prevalence.
Beginning in 1997, the asthma questions on the NHIS changed the measure of asthma prevalence (2). Now, three measures are used, all restricted to persons with a medical diagnosis of asthma. The first measure is referred to as lifetime asthma prevalence, which includes respondents with a medical diagnosis of asthma at any time in their lives. The second measure identifies persons with a current diagnosis of asthma. The third is a measure of 12-month attack prevalence, which includes the number of persons who had one or more attacks or episodes during the past 12 months.
The objective of this study is to estimate direct and indirect costs of asthma among school-age children using data from the 1996 MEPS. This study is different from previous reports in four aspects: 1) it produces both national estimates and per capita estimates; 2) the 1996 MEPS data are used to derive not only medical cost estimates but also SAD estimates; 3) both medical costs and SADs are estimated based on treated prevalence; and 4) the costs of productivity loss due to premature death among school-age children are estimated. We hope that the results of this study can provide more insights into the economic burden of childhood asthma on society, the individual child, and the family.
Methods
General approach
A societal perspective was used to estimate costs of asthma among school-age children. The direct costs of asthma were estimated as asthma-related medical costs. Indirect costs were estimated as costs of lost productivity, including parents' loss of productivity due to asthma-related SADs and loss of productivity due to premature death of children from asthma. Data from the 1996 MEPS were used to derive medical cost and SAD estimates. Published estimates were used for value of lost productivity (9), and data from the National Vital Statistics System were used for asthma mortality among school-age children (2). Both national and per capita estimates were calculated. All costs were in 2003 dollars.
Data source and data processing
The MEPS is the third in a series of national probability surveys conducted by the Agency for Healthcare Research and Quality (AHRQ)The MEPS is the third in a series of national probability surveys conducted by the Agency for Healthcare Research and Quality (AHRQ) on the financing and use of medical care in the United States. MEPS actually comprises a family of four surveys: 1) a household survey; 2) a survey of medical providers; 3) a survey of health insurance providers; and 4) a survey of nursing home residents. Using the NHIS as its sampling frame, the MEPS household component (MEPS–HC) is designed to provide estimates of health care use, spending, sources of payments, and insurance coverage for the U.S. civilian noninstitutional population. Using an overlapping panel design, self-reports of health care use and spending are collected at the person and household levels through five rounds of in-person interviews that occur during a 30-month period. This yields two full years of data. The MEPS medical provider component (MEPS–MPC) is a survey of medical providers that are directly linked to the respondents in the household survey. The MEPS–MPC is used to replace or to supplement household data to reduce potential bias from relying solely on self-reported data. The MEPS–MPC focuses on medical events and collects information on dates of visit, diagnosis and procedure codes, and charges and payments.
Although MEPS data are generally available from 1996 to 2000, we chose to use 1996 data for this study because only 1996 data have information on SADs. The data used in this article were derived from the 1996 Person-Level Full-Year File (HC-012) and the Person-Level Medical Event Files (HC-001). The 1996 MEPS–HC comprises a sample of 10,597 households and 23,565 individuals. Hispanic households were oversampled at ratios of approximately 2:1; African American households were oversampled at ratios of approximately 1.5:1. The subsample of individuals used in this study consisted of all children aged five to 17 years in 1996 and was further divided into two condition groups: children with asthma and children without asthma. In this study, the definition of children with asthma was further refined to include children who had any type of health care provider visit or prescription medication related to asthma during the year. All other children in the study sample were classified as children without asthma. We identified the children with asthma through diagnosis code 493, according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) (10). An event cost associated with an ICD-9 code of 493 was considered an asthma-related medical care cost and was derived from the event files. Using the full-year file, we derived estimates of all-cause medical spending and SADs for each condition group.
Expenditures in the 1996 MEPS are defined as the sum of direct payments for care provided during the year, including out-of-pocket payments and payments by private insurance, Medicare, Medicaid, and other sources. For this analysis, the medical events were classified and enumerated into the following mutually exclusive categories: purchase of prescribed medicines, hospitalizations, emergency room (ER) visits, outpatient hospital visits, office-based visits, other medical equipments, and home health care.
To derive SAD estimates, we had to address two data issues in the 1996 MEPS. First, Round 3 was conducted in 1996 and 1997; some data from Round 3 pertain to 1997. The number of days lost from school in Round 3 that occurred in each calendar year was not ascertained. We developed an algorithm for deciding what portion of the reported SADs occurred in 1996 (the number of days in Round 3 in 1996 divided by the total number of days in Round 3 and then multiplied by the total number of SADs in Round 3). Second, two variables have missing data — the ending date of Round 3 and SADs in Round 3 (34 children with asthma and 583 children without asthma). We used the SAS Multiple Imputation Procedure to impute the missing values for the two variables (11).
Data analysis
The data were initially processed with SAS software (SAS Institute, Inc, Cary, NC). All estimates produced in this study were weighted to represent the U.S. population. The sampling weights were used to adjust for potential survey response bias. SUDAAN software (Research Triangle Institute, Research Triangle Park, NC) was used to account for the complex sample design in the computation of the final estimates and standard errors for the estimates produced (12). Linear regression analyses were conducted to estimate excess all-cause medical costs and excess SADs for children with asthma compared with children without asthma, controlling for the effects of sociodemographic and access-to-care variables such as age, sex, race, mother's education level, poverty status, and health insurance coverage.
Calculation of direct and indirect costs
Asthma-related expenditures were considered as the direct costs of asthma; estimates were directly derived from MEPS data. The costs of parents' loss of productivity due to SADs were calculated as the product of SADs associated with asthma and the cost of lost productivity (value of a day lost). The costs of loss of productivity due to premature death were calculated as the product of asthma mortality and the cost of lost productivity (discounted value of future total earnings).
According to a previous published study (9), the value of a lost day was estimated to be $108 (in 2000 dollars) or $115 (in 2003 dollars), which was calculated as the sum of annual earnings and annual household services divided by 365. The same study also reported the present value of future total earnings (including fringe benefits) and of the combination of future earnings and household production. Estimates are reported at each exact age in five-year intervals, beginning with birth. The reported estimates (in 2000 dollars and discounted at 3%) are $1,064,530 for a child aged five years, $1,176,371 for a child aged 10, and $1,290,814 for a child aged 15. In this study, we used the average of the three estimates — $1,254,347 (in 2003 dollars) — as the present value of future total earnings of children who are currently aged five to 17 years.
A recent study by Akinbami and Schoendorf used data from the Mortality Component of the National Vital Statistics System and reported that annual asthma mortality (per million) was 2.7 among children aged five to 10 years and 5.6 among children aged 11 to 17 years during 1995–1996 (2). We used these mortality estimates and the 1996 Census data to calculate the total asthma mortality of children aged five to 17 years as 211.
Results
As shown in Table 1, of a total of 4786 children aged five to 17 years in our study sample, 248 had asthma-related care during 1996, representing 2,521,537 children in the United States. The annual treated prevalence rate of asthma was 4.9% (2,521,537/51,605,560). Among all of the children with asthma in the nation during 1996, 68% had at least one health care provider visit of some type, and 94% purchased prescriptions for asthma medication and devices. The percentage of children with asthma who used asthma-related care by category of service was 61% for office-based visits, 10% for ER visits, 3% for outpatient visits, and 2% for hospitalization. Sixty-five percent of children with asthma purchased quick-relief medicine for asthma and 36% purchased controller medicine for asthma. Asthma-related medical costs were $1009.8 million ($401 per child) among all school-age children with asthma in the nation. The distribution of asthma-related expenditures shows that 43% of the total expenditure was in prescribed medicine, followed by 36% in outpatient and office-based visits and 21% in hospitalization and ER care.
Table 2 shows the all-cause medical care costs of children with asthma and children without asthma. The all-cause medical costs of children with asthma totaled $2590 million during 1996, with asthma-related care accounting for 39% of the all-care costs. The weighted per capita costs were $1042 for children with asthma and $618 for children without asthma. Compared with a child without asthma, a child with asthma had an excess of $424 in all-cause medical spending. Comparing the cost distribution of children with asthma to children without asthma, children with asthma had a higher proportion of all-cause medical expenditures in prescriptions, office-based visits, and outpatient visits and a lower proportion of expenditures in hospitalization and ER visits.
Table 3 shows the estimates of SADs in children with and without asthma as well as excess SADs of children with asthma compared with children without asthma. Among 2.5 million school-age children with asthma, a total of 14.5 million SADs occurred. On average, a child with asthma missed 2.48 more days of school than a child without asthma (5.81 SADs per child with asthma and 3.33 SADs per child without asthma). The total number of SADs associated with asthma among all of the children with asthma was 6.3 million, accounting for 43% of the total SADs of all of the children with asthma for all illness and injury.
As shown in Table 4, the direct costs of asthma were estimated to be $1009.8 million ($401 per child). The indirect costs of asthma were estimated to be $983.8 million ($390 per child), including $719.1 million ($285 per child) associated with SADs and $264.7 million ($105 per child) associated with premature deaths due to asthma. The total economic impact of asthma in school-age children was $1993.6 million ($791 per child).
Discussion
This study is the first to produce both national and per capita estimates of direct and indirect costs of children with asthma. Although only 4.9% of U.S. children, or 2.5 million children nationally, had any type of health care provider visit or prescription medication related to asthma during 1996 (parent-reported asthma prevalence was 5.9%), the economic impact of asthma among those children is substantial. The direct costs of asthma were estimated to be $1009.8 million ($401 per child). The indirect costs of asthma were estimated to be $983.8 million ($390 per child), including $719.1 million ($285 per child) associated with SADs and $264.7 million ($105 per child) associated with premature deaths due to asthma. The total economic impact of asthma in school-age children was $1993.6 million ($791 per child).
Unlike most of the previous studies of childhood asthma, this study focuses on children who had any treatment for asthma during 1996. Before 1997, no other national survey collected information to measure treated prevalence or attack prevalence. After the 1997 redesign of the NHIS questionnaire, information to estimate asthma attack prevalence was obtained. The 1996 treated prevalence of asthma derived in this MEPS study (4.9%) is close to the attack prevalence derived from the 1997 NHIS (5.9% among children aged five to 10 years and 6.0% among children aged 11 to 17 years). The annual prevalence of parent-reported asthma in this MEPS study (5.9%) is in the general range of other recent studies, where the prevalence of asthma in 1996 has been reported previously to be 5.5% among individuals aged five to 20 years (2), 7.4% among children aged five to 10 years, and 7.7% among children aged 11 to 17 years (2).
Although two studies by Weiss et al (4,5) have previously estimated direct medical expenditures of asthma in children, their estimates are not comparable to ours because of two major differences. First, they used charges data, which are very different from the payment data we used in this study, to derive cost estimates. Second, they studied costs for all children aged from birth to 17 years, rather than for the school-age children we focused on in this study. Although there is no comparable per capita estimate in the literature to compare with our estimate of asthma-related medical costs, we found that our per capita estimate of asthma-related costs ($401) is very close to our estimate of the excess all-cause medical spending ($424) per child with asthma compared with a child without asthma.
To our knowledge, this is the first study that uses MEPS data to derive SAD estimates for children with asthma. Most early SAD estimates were produced using NHIS data. To compare the results of this MEPS study with those of the most recent NHIS study, one should keep in mind two differences between the studies. First, as noted earlier, the definition of children with asthma is different. In this study, only children who had treatment for asthma were identified as children with asthma. In the NHIS study, the definition of children with asthma was broader: children with asthma did not have to have treatment if their parents believed that they had asthma during the year. Second, the information collected on SADs is different between the two data sets. Although the 1996 MEPS obtained information on the total number of SADs for each child and on the conditions that caused the SADs, it collected no information on the number of SADs associated with a particular condition. Thus, our SAD estimates reflect annual SADs resulting from all conditions. In contrast, the NHIS provided data on the number of SADs resulting from specific conditions (i.e., SADs associated with asthma). Based on the MEPS data, we found that the total number of SADs in 1996 resulting from all conditions was 14.5 million among children with treated asthma (5.75 days per child). On average, a child with asthma had an excess of 2.5 SADs compared with a child without asthma. The study using 1994–1996 NHIS data found that the annual number of SADs associated with asthma was 14 million (3.7 days per child) among children with parent-reported asthma. Although our SAD estimates are different from those of the NHIS study, we found that the proportion of all children with asthma who had asthma-related SADs (49%) is interestingly close to the proportion of all SADs associated with asthma at the individual level (42% = 2.48/5.81). This indicates internal consistency in the MEPS data.
Like many other cost analyses, this study has some clear limitations. The results of this study should be interpreted with some degree of caution. First, as noted earlier, the medical condition data in the MEPS were derived from parents' reports; they may not conform perfectly to diagnoses made by physicians. However, a study by AHRQ staff indicated that, at the three-digit ICD-9 code level, there was agreement between household- and provider-reported conditions in the overwhelming majority of cases (13). In addition, this concern is minimized by the fact that most of the children (94%) who were classified as children with asthma in this study had purchased prescription medicine for asthma. Second, in the MEPS event data file, a health service may occur for multiple reasons; spending associated with specific conditions is not mutually exclusive. However, in this study, we also estimated all medical spending for children with asthma and compared them with children without asthma. Because the estimated per capita asthma-related cost ($401) is very close to the estimated excess cost per child with asthma ($424), it is reasonable to believe that our direct cost estimates are fairly accurate. Third, we did not use asthma medication data as a means of identifying children with asthma; we might have underestimated the treated prevalence as well as the national medical expenditure of children with asthma. There were more probable asthma cases in the 1996 MEPS data. Among children who had neither proxy-reported asthma nor treated asthma in the study sample, 15 probably really had asthma because they purchased albuterol more than once during a year. Fourth, we have significantly underestimated the direct costs of asthma in school-age children, because children with asthma receive a substantial proportion of care in school settings. Because MEPS is limited to expenditures in medical settings, the cost of services provided by the school system, such as nursing care and first-aid care provided by school nurses, school health aids, and school secretaries,were not included in this study. Fifth, the sample size for inpatient, outpatient, and emergency care was small. However, the focus of this study was total medical expenditures across all medical care services, for which we had a sufficient sample size to be confident in both national estimates and per capita estimates.
Even with these limitations, it is reasonable to conclude that the economic impact of asthma to school-age children, families, and society is immense, and more public health efforts are needed to better control asthma in children. Evidence from evaluation studies of asthma intervention programs suggest that asthma among school-age children can be controlled, and significant cost savings in medical care and parents' loss of productivity from asthma-related SADs can be realized, especially when the programs are aimed at children with moderate to severe asthma (14-21). Two economic studies of clinical-based asthma education programs have documented cost savings from $180 per enrolled child with at least persistent asthma to $542 per enrolled child with a history of frequent use of ER services for asthma (14,15). One particularly well-designed randomized controlled trial of an inner-city social-worker–based education program was able to demonstrate cost savings of $2509 per child with one or more hospital visits at baseline, $1050 per child with two or more unscheduled visits at baseline, and $220 per child with more than 50% of days with asthma symptoms (16). Two studies of asthma management programs found a cost savings of $1144 per inpatient child as a result of an inpatient asthma clinical pathway and a cost savings of $1667 per child with one or more hospitalizations or two or more ER visits during a six-month period as a result of a home-based self-management program (17,18). Studies examining the impact of asthma interventions on SADs have also found significant reduction in SADs, including a 1.8-day reduction by a school-based asthma education program, a two-day reduction by a summer asthma camp-based education program for children with moderate to severe asthma, and a 2.5-day reduction by a large-scale population-based asthma management program for asthma patients and their caregivers (19-21).
Based on the medical cost estimates derived in this study, we found that $211.4 million in medical expenditures (21% of the total asthma-related medical expenditures in 1996) are preventable with effective asthma interventions, including $120.8 million in inpatient care and $90.6 million in ER care. In addition, published asthma intervention studies reveal that SADs related to asthma can be reduced by 1.8 to 2.5 days (19-21). Since the cost estimate of parents' loss of productivity derived in this study was based on an additional 2.5 SADs per child with asthma relative to a child without asthma, it is reasonable to believe that more than half of the total indirect costs of parent's loss of productivity ($983.8 million) are preventable with effective interventions. However, to achieve such cost savings, more public health efforts are needed to educate parents and children onthe child's condition and medications, the need for follow-up care, and the importance of avoiding known disease triggers. To ensure the success of such effort, education of primary care providers and school staff and efficient collaboration among primary care providers, school health professionals, and health education professionals are essential.
Figures and Tables
Table 1 Asthma-related Medical Care Among U.S. Children Aged Five to 17 Yearsa
Number of persons Expendituresb
Medical care Unweighted Weighted % of children who used each service (weighted) Per capita weighted expenditure, $
(SE) Weighted expenditure, $
(SE) Weighted expenditure distribution,
%
Prescription 230 2,375,040 94 183
(20) 433,761,955
(58,520,972) 43
Inpatient 5 55,072 2 2193
(337) 120,768,739
(81,768,373) 12
Outpatient 8 76,413 3 635
(377) 48,481,437
(35,537,654) 5
Emergency 24 244,170 10 371
(89) 90,592,184
(31,553,044) 9
Office visit 153 1,532,728 61 206
(47) 316,158,833
(78,720,894) 31
Total 248 2,521,537 NA 401
(54) 1,009,763,148
(147,630,859) 100
a Based on sample size of 4786. Source: 1996 Medical Expenditure Panel Survey. SE = standard error. NA = not applicable.
b Expenditures are presented in 2003 dollars.
Table 2 Comparison of All-cause Medical Expenditures of U.S. Children Aged Five to 17 Years With and Without Asthmaa
Children with asthma
(N = 2,521,537) Children without asthma
(N = 49,084,023)
Medical care Per capita expenditure,
$ (SE) National expenditure,
$ (SE) Expenditure distribution,
% Per capita expenditure,
$ (SE) National expenditure,
$ (SE) Expenditure distribution,
%
Prescription 263
(25) 651,578,798
(78,265,297) 25 146
(21) 3,525,325,923
(510,679,902) 11
Inpatient 3328
(819) 284,511,726
(134,082,066) 11 9254
(1999) 9,657,060,475
(2,439,668,298) 29
Outpatient 898
(271) 260,576,661
(93,150,742) 10 1068
(189) 3,110,152,939
(598,533,120) 9
Emergency 364
(51) 199,370,182
(42,482,913) 8 677
(170) 3,765,138,136
(974,320,040) 11
Office visit 435
(66) 954,241,953
(152,081,219) 37 327
(17) 10,774,764,219
(687,815,760) 33
Other equipment 558
(372) 239,714,979
(157,341,673) 9 231
(18) 1,538,558,287
(158,934,373) 5
Home visit 0 0 0 2,762
(1365) 714,950,437
(365,042,774) 2
Total expenditure 1027
(110) 2,589,994,298
(317,235,615) 100 675
(72) 33,085,950,417
(3,598,962,584) 100
Adjusted per capita expenditureb 1042
(112) NA NA 618
(62) NA NA
a Source: 1996 Medical Expenditure Panel Survey. All expenditures are weighted and in 2003 dollars. SE = standard error. NA = not applicable.
b Adjusted by age, sex, race, mother’s eduction level, poverty status, and health insurance coverage.
Table 3 Estimates of School Absence Days (SADs) Among U.S. Children Aged Five to 17 Years With and Without Asthmaa
Children with asthma Children without asthma Excess SADs per child with asthma compared with a child without asthma
Unweighted
(n = 248) Weighted
(n = 2,521,537) Unweighted
(n = 4538) Weighted
(n = 49,084,023)
Total SAD 1440 14,489,164
(1,448,744) 15,454 166,321,913
(6,935,832) NA
SAD per child 5.8 5.75
(0.41) 3.41 3.39
(0.12) 2.36
Adjusted SAD per childb NA 5.81
(0.40) NA 3.33
(0.12) 2.48
a Data are from 1996 Medical Expenditure Panel Survey. All values are numbers (standard errors) unless otherwise indicated. NA = not applicable.
b Adjusted by age, sex, race, mother's education level, poverty status, and health insurance coverage.
Table 4 Direct and Indirect Costs of Asthma Among U.S. Children Aged Five to 17 Years (in 2003 Dollars)
Direct costs ($) Indirect costs ($) Total costs ($)
Asthma-related medical costs Costs of lost productivity due to asthma-related school absence days Costs of lost productivity due to premature death
National estimate 1,009,763,148 719,142,352 264,667,217 1,993,572,717
Per capita estimate 401 285 105 791
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: JWang LY, Zhong Y, Wheeler L. Direct and indirect costs of asthma in school-age children. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0053.htm
==== Refs
1 2000 49 908 911 MMWR Morb Mortal Wkly Rep Measuring childhood asthma prevalence before and after the 1997 redesign of the National Health Interview Survey — United States 11043644
2 Akinbami LJ Schoendorf KC 110 2 Pt 1 2002 315 322 Pediatrics Trends in childhood asthma: prevalence, health care utilization, and mortality 12165584
3 American Lung Association 3 2003 Trends in asthma morbidity and mortality (Internet) New York The Association Available from: URL: http://lungusa.org
4 Weiss KB Gergen PJ Hodgson TA 1992 326 862 866 N Engl J Med An economic evaluation of asthma in the United States 1542323
5 Weiss KB Sullivan SD Lyttle CS 2000 493 499 J Allergy Clin Immunol Trends in the cost illness for asthma in the United States, 1985–1994 10984369
6 Lozano P Sullivan SD Smith DH Weiss KB 1999 104 957 963 J Allergy Clin Immunol The economic burden of asthma in US children: estimates from the National Medical Expenditure Survey 10550739
7 Fowler MG Davenport MG Garg R 1992 90 939 944 Pediatrics School functioning of US children with asthma 1437438
8 Silverstein MD Mair JE Katusic SK Wollan PC O'Connell EJ Yunginger JW 2001 139 278 283 J Pediatr School attendance and school performance: a population-based study of children with asthma 11487757
9 Haddix AC Teutsch SM Corso PS 2003 Prevention effectiveness: a guide to decision analysis and economic evaluation New York Oxford University Press
10 U.S. Department of Health and Human Services 1988 International classification of diseases, ninth revision, clinical modification Public Health Service Washington (DC)
11 Little RJA Rubin DB 1987 Statistical analysis with missing data John Wiley & Sons, Inc New York
12 Shah BV Barnwell BG Bieler GS 1997 Research Triangle Institute Research Triangle Park (NC) SUDAAN user's manual: software for analysis of correlated data, release 7.5
13 Cohen JW Krauss NA 2003 22 129 138 Health Aff Spending and service use among people with the fifteen most costly medical conditions, 1997
14 Lewis CE Rachelefsky G Lewis MA de la Sota A Kaplan M 1984 Pediatrics A randomized trial of A.C.T. (asthma care training) for kids 74 4 478 486 6384911
15 Kelly CS Morrow AL Shults J Nakas N Strope GL Adelman RD 105 5 2000 1029 1035 Pediatrics Outcomes evaluation of a comprehensive intervention program for asthmatic children enrolled in Medicaid 10790458
16 Sullivan SD Weiss KB Lynn H Mitchell H Kattan M Gergen PJ 2002 110 576 581 J Allergy Clin Immunol The cost-effectiveness of an inner-city asthma intervention for children 12373264
17 Kelly CS Andersen CL Pestian JP Wenger AD Finch AB Strope GL Ann Allergy Asthma Immunol Improved outcomes for hospitalized asthmatic children using a clinical pathway 2000 84 509 516 10831004
18 Axlrod RC Zimbro KS Chetney RR 2001 8 38 42 J Clin Outcomes Manag A disease management program utilizing life coaches for children with asthma
19 Spencer GA Atav SA Johnston Y Harrigan JF 2000 23 20 30 Fam Community Health Managing childhood asthma: the effectiveness of the open airways for schools program
20 Kelly CS Shield SW Gowen MA Jaganjac N Andersen CL Strope GL 1998 35 165 171 J Asthma Outcomes analysis of a summer asthma camp 9576142
21 Georgiou A Buchner DA Ershoff D Blasko KM Goodman LV Feigin J 2003 90 308 315 Ann Allergy Asthma Immunol The impact of a large-scale population-based asthma management program on pediatric asthma patients and their caregivers 12669894
|
15670464
|
PMC1323314
|
CC BY
|
2022-01-27 23:22:41
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A11
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0038
Special Topics in Public Health
PEER REVIEWEDChildhood Obesity — What We Can Learn From Existing Data on Societal Trends, Part 1
Sturm Roland PhD RAND
1700 Main St, Santa Monica, CA 90401 [email protected]
310-393-0411 ext 6164
1 2005
15 12 2004
2 1 A122005
The number of overweight and obese youth has increased in recent decades, and numerous theories on causes have been proposed. Yet almost no data are available to assess how the lives of children have changed during the “obesity epidemic.” What are children and adolescents now doing with their time that they did not do before? Are they participating less in sports? Watching more television? Doing more homework? Without tracking these broader societal changes, it is difficult to identify the most (and least) promising areas for interventions. This two-part report compiles trend data for several areas. Part 1 discusses trends in time use, homework, and media use; part 2 discusses trends in transportation, physical education, and diet.
The main findings of this article are the following: One, the free time of children has substantially declined because of increased time away from home, primarily in school, day care, and after-school programs. Two, participation in organized activities (including sports) has also increased. Three, unstructured playtime has decreased to make room for organized activities. Four, time spent in some sedentary activities like watching television, participating in conversations, or taking part in other passive leisure activities also declined just when obesity became a major concern. Five, increases in homework have not caused decreases in free time, contradicting a common belief in education circles.
Part 1 of 2
==== Body
Introduction
The number of overweight and obese youth has been increasing dramatically in recent decades, and there is no sign that this trend is ending (1). Prevention may be one of the hallmarks of pediatric practice, but office-based counseling offers limited leverage to counter broader changes that affect the daily lives of children. Even though prevention and treatment in clinical settings have been the focus for interventions in the past, researchers now agree that trends in overweight arise from changes in social and environmental factors that need to be understood and modified for effective prevention (2,3). Many factors have been suggested as causes of the "obesity epidemic" among children — reduced physical education at school, increased homework loads, campus vending machines, television, larger portion sizes, fast-food restaurants, video games, and countless others. Yet virtually no data track how the lives of children have changed during the obesity epidemic; in fact, except for the National Health and Nutrition Examination Survey (NHANES), no reliable data exist to track weight increase among children (1).
This two-part report reviews data sources available for time trends, summarizes trend data that have been published (typically outside the health literature), and provides several new calculations. Finding comparable data across several years is difficult; finding such data across several decades is even more difficult. With purely cross-sectional data, we can only compare instantaneous snapshots against an ideal world, but what we really want to know is: What has changed since the time when childhood obesity was not a major public health problem? What are children and adolescents now doing with their time that they did not do before? How are they less physically active? Are they participating less in sports? Without tracking these broader societal changes, we will not easily identify the most promising areas for intervention.
When considering an array of information, it is often useful to start with a systematic review of the primary research literature. Systematic reviews are most useful when multiple studies pose similar research questions. Guidelines on clinical practice are often based on systematic reviews because many studies have identical research questions. An Institute of Medicine committee focused on preventing childhood obesity has taken the systematic review approach (4) and found more than 40,000 citations after searching the topics of obesity, overweight, body weight, dietary patterns, and physical activity. Unfortunately, this approach has not been very successful in identifying information about societal trends. Much of the primary research on societal trends does not focus on health, weight, or physical activity and contains no related key words in the abstracts. A seminal paper on changes in children's time use, for example, contains none of the words or word fragments "obesity," "overweight," "body weight," "diet," or "physical activity" anywhere in the text (5).
The approach taken here is, therefore, more eclectic by necessity. The research literature on childhood obesity has several themes, including transportation, media use, physical education, school hours, and diet. For this report, all data sets on children at the Inter-university Consortium for Political and Social Research — an organization of member institutions that archives social science data — were scanned for articles providing time-series data. Most sources were found, however, by the less systematic approach of identifying the main data sets and surveys used in the fields mentioned above and then determining whether those data could provide information on secular trends.
Analysis of time-use data is a first step toward understanding how economic incentives have altered behavior patterns that lead to weight gain. Indeed, time-use data for adults are so important in tracking societal trends that in 2003 the Bureau of Labor Statistics and the Census Bureau started to collect time-use data. Ideally, we would like to translate time use into energy expenditure and examine how this relationship contributes to weight gain, but the data are not detailed enough for this task. The next section of this study discusses existing time-use data.
When we try to understand the economic and technological forces behind changes in time use, few technological innovations have been more important in the lives of children and adolescents than the emergence and evolution of new communication technologies. Technologies like cable television, videos, computer games, and the World Wide Web have altered how children obtain information and entertainment, but how do they affect children's physical activity patterns or use of time? A later section of this article (Media) discusses available data, but unfortunately we can only calculate trend data for television watching. Only a single cross-sectional national survey (for 1999) provides a full assessment of all media use.
In education circles, the hypothesis that homework, or, more generally, the academic excellence movement, plays an important role in weight gain (and/or declining physical fitness) among children appears to be widely accepted, even outside the United States. In surveys conducted by the Chinese University of Hong Kong, the National University of Malaysia, Thailand's Mahidol University, and the Philippine Food and Nutrition Research Centre, by far the most common reason students gave for not participating more in physical activity was homework, cited 1.5 times more often than the runner-up, heat/weather (6). The final section of this article reviews data sources to investigate the hypothesis that an increase in homework has contributed to weight gain and/or lack of physical activity.
The second part of this report in Volume 2, Issue 2 of Preventing Chronic Disease will look at data for three other areas: transportation, physical education, and diet.
Time Use
We first review existing data on how American children spend their time. There are several ways of assessing time use, but numerous methodological studies show that the best way to collect large-scale time-use data is to use time-diary data, where individuals describe what they have done during the past day (7). Relatively comparable time-use data for adults have been collected approximately every decade since the 1960s, enabling researchers to paint a broad picture of how adult lives and physical activity have changed in the past 40 years (8,9). Time-use data for children are sparse, but researchers from the University of Michigan fielded two surveys in 1981 and 1997 (5,10). Data presented in this article are based on calculations made from data provided in detailed tables published in 2001 (5).
Both University of Michigan surveys used 24-hour time diaries with similar methodology. The 1981 data are based on time diaries of 222 children aged two to 12 years. Each child provided data for one school day and one nonschool day. The data are nationally representative (weighted for sampling probability and post-stratification factor) and have been used in a variety of other reports, primarily in education. The 1997 survey was an addition to the Panel Study of Income Dynamics, a representative sample of U.S. men, women, children, and their families. The study consisted of interviews of 2380 households containing 3563 youth and had a response rate of 88%. Post-stratification weights based on the 1997 Current Population Survey are used to make the data nationally representative, and sampling weights adjust for survey design. Subsetting to children aged three to 12 years with complete data resulted in 2119 observations.
Free/discretionary time has declined
The free time of children as a proportion of the total weekly time of 168 hours declined by approximately 12% from 1981 to 1997, if time spent eating, sleeping, in personal care (e.g., preparing to go places, packing, getting dressed), in school, and in child care is subtracted from the total. The decline in free time — between seven and eight hours per week, depending on how we treat study time or household work — is largely due to increased time spent in school and child care and to a lesser extent to increased personal-care time. Figure 1 shows that this decline occurred across all age groups; data are significant at P < .01.
Figure 1 Decline in discretionary time (in minutes per week) between 1981 and 1997 among U.S. children aged three to 12 years. Calculations based on data from Hofferth and Sandberg (5).
Bar chartA text description of this chart is also available
Minutes per week
Age 3–5 337
Age 6–8 442
Age 9–12 408
All ages 419
Time spent in school increased about two hours per week, or 25 minutes per school day, from 24 hours and 45 minutes to 26 hours and 48 minutes in an average week. Day-care time increased from about 14 minutes to almost three hours per week because both a larger share of children used day care and children in day care spent more time in it. In Figure 2, we add other academic activities that time-use researchers often consider discretionary but that have similar purposes (and physical activity levels): primarily reading and studying at home. Personal-care time also increased over the same period, from more than six hours per week to eight hours per week. Time-use researchers have hypothesized that as children spend more time away from home, they need more time to get ready to do so (5). Smaller changes occurred in time spent eating, which decreased, and sleeping/napping, which increased by the same amount (not shown). The decline in eating time parallels a decline in the frequency with which families sit down together to share a family meal (11). More important changes occurred in the composition of children's free time. While time spent in some categories decreased by more than the average 12%, time spent in other categories increased. Time spent viewing television as a primary activity declined by the greatest percentage — by 23%, or about four hours per week (Figure 2, P < .001). This decline did not take place because of a decrease in the proportion who watched television (because almost all children watched television during both 1981 and 1997); instead, the decline took place because of a reduction of television time among watchers.
Several other sedentary activities declined significantly and proportionately more than discretionary time overall: church attendance, youth-group participation, passive leisure, and other household conversations. In Figure 2, this group of activities represents the second largest group of declines. But two sedentary activities at home, reading and studying (grouped with school/day care in Figure 2), also increased.
Time spent in hobbies, organized sports, and arts activities increased, reflecting an overall trend toward structured activities and a decline in unstructured activities. Figure 2 groups only the more active categories together (sports/outdoors), and they show a substantial increase. Sports increased significantly for children younger than nine years; other categories did not change significantly. Hobbies/arts, which include dance and music lessons (not shown here), are more active than watching television or other passive leisure and are perhaps more comparable to playtime.
Figure 2 Changes in time (in minutes per week) spent on activities between 1981 to 1997 by U.S. children aged three to 12 years. Calculations based on data from Hofferth and Sandberg (5).
Bar chartA text description of this chart is also available
Minutes per week
School, day care, studying, reading, art activity 366
Personal care 107
Sports/outdoors 73
Shopping 61
Playing -138
Other passive leisure (e.g., conversations, church, visiting) -162
Television -246
Hofferth and Sandberg analyzed the role of shifting demographics and maternal employment and found that few of the changes in time use could be linked to either factor (5). While some time use differed according to maternal education, family size, and family composition in 1981, different socioeconomic groups tended to become more similar rather than more different in time use. Thus, the increase in time spent in school (which includes after-school care) reflects changing social preferences for greater use of schools and school activities.
Age differences
Some notable differences in time use across different age groups suggest different levers for interventions by age group. For children aged three to five years (Figure 3) and children aged six to eight (Figure 4), the largest decline in time use is in playtime; for children aged nine to 12 years, the largest decline is in television watching (Figure 5). While playtime declined overall, it actually increased among children aged nine to 12. This increase may reflect more video- and computer-game use among this age group. The largest decline for children aged nine to 12 was in television watching, but household conversations and other passive leisure also declined more in this group than in other age groups. However, these declines may be countered by increases in sedentary playtime; time spent in this category grew by 1.5 hours per week in this age group (Figure 5). The declines and increases may represent trade-offs between different forms of media use, but time diaries do not reveal this level of detail.
Figure 3 Changes in time (in minutes per week) spent on activities between 1981 and 1997 by U.S. children aged three to five years. Calculations based on data from Hofferth and Sandberg (5).
Bar chartA text description of this chart is also available
Minutes per week
School, day care, studying, reading, art activity 405
Personal care 181
Sports/outdoors 134
Shopping 69
Television -82
Other passive leisure (e.g., conversations, church, visiting) -94
Playing -509
While the dominant change is in the increase of time spent in school or day care away from home, time spent studying at home increased significantly among children aged six to eight, and time spent reading increased significantly among children aged three to five. The proportion of children aged three to five who spent time reading or being read to doubled between 1981 and 1997. Increased reading among this age group probably reflects increasing parental concern about preparing children for school. Increased enrollment in day-care centers and preschools may also be associated with children reading at early ages. We will examine homework loads more closely in a later section to investigate claims in the popular press and the education field that the homework burden has increased by so much that it now constitutes an enormous time burden on students and families, preventing them from engaging in other activities. The time-use data from the Michigan group is often used as a key piece in this argument, so it is worth keeping the magnitudes in mind: 76 minutes of the 485 additional minutes per week spent in school or other learning activities among children aged six to eight (Figure 4) were designated toward homework. Among children aged nine to 12, however, the increase in homework was not statistically significant, and the point estimate was an increase of 19 minutes for studying at home out of the 369 additional minutes per week in learning activities (Figure 5).
Figure 4 shows that sports/outdoor time did not increase significantly among children aged six to eight, representing a different pattern than other age groups. The major reason for the significant increase among children aged three to five was the increased proportion engaging in sports, which almost doubled over the period.
Figure 4 Changes in time (in minutes per week) spent between 1981 and 1997 on activities by U.S. children aged six to eight years. Calculations based on data from Hofferth and Sandberg (5).
Bar chartA text description of this chart is also available
Minutes per week
School, day care, studying, reading, art activity 485
Personal care 100
Shopping 99
Sports/outdoors -46
Other passive, conversations, church, visiting, leisure -135
Television -181
Playing -228
Figure 5 Changes in time (in minutes per week) spent between 1981 to 1997 on activities by U.S. children aged nine to 12 years. Calculations based on data from Hofferth and Sandberg (5).
Bar chartA text description of this chart is also available
Minutes per week
School, day care, studying, reading, art activity 369
Sports/outdoors 92
Personal care 92
Playing 90
Shopping 27
Other passive, conversations, church, visiting, leisure -163
Television -385
In summary, the largest change in children's time use in the past two decades was a decline in discretionary/free time, paralleled by an increase in school or day care and personal care. Particularly noteworthy are significant declines in many categories of passive leisure (television, conversations, other passive leisure) and increases (statistically insignificant) in sports/outdoor time. Increases in sports participation were largest among children of non-working mothers; thus, these increases did not result from increases in maternal employment (5,10). Children with less time to play mostly reflect decreased time spent at home. Children may be playing in their preschool programs, and they may have some free time at school, so the level of aggregation presented here provides only a partial picture of children's time and activity levels. Sports/outdoor time outside school is increasing, mainly among preschool children, but the increase in sports is significant for both the three-to-five and nine-to-12 age groups. Given the large increase in time children spend in school or day care, it is also important that children have enough physical activity in those settings. Moreover, the amount of time children participate in more physically challenging activities in school or day-care settings should have increased over time, corresponding to the total increase that children are now away from home. To what extent this has or has not occurred requires different data.
Media
The first complete national data on media use among American youth were collected in 1999 by the Kaiser Family Foundation project, Kids & Media @ The New Millennium (12). Other surveys have examined children's use of selected media (most commonly television watching), but no data allow us to track media use comprehensively over time.
Despite interest in new media (e.g., computers, video games), television remains by far the dominant medium (Table). The impact of computers and video games on sedentary behavior is probably not very large, especially when compared with television, as they together comprise only about 10% of the average daily media budget of children aged two to 18. There are, however, large differences by age and sex. Children younger than eight years spent a negligible amount of time on video games or computers in 1998, but boys aged eight to 13 averaged 47 minutes per day playing video games (13). On the other hand, children aged eight to 12 also experienced the largest decline (approximately 50 minutes per day) in television watching between 1981 and 1997.
No comparable surveys track the latest changes for all children, although it is likely that computer use has increased since the Kaiser Family Foundation survey. Video gaming may or may not have peaked already by 1999. A new survey fielded in 2003 for children younger than seven years found that one in five children aged four to six plays computer games in a typical day (14). A new trend certainly is that some computers and video games target preschoolers.
Television has been around for a much longer time and because of its continuing dominance, it has received more attention. Television may contribute directly to obesity by reducing energy expenditure through displacing physical activity or indirectly by increasing dietary intake — through snacking during viewing or changing eating patterns caused by food advertising. Numerous cross-sectional studies found significant positive associations between television viewing and youth obesity; prospective studies include some null findings, but a randomized trial confirmed that a reduction in television watching can reduce weight gain (15-17).
Assessing time trends, even for television, is difficult because small differences in methods across different surveys create methods effects that far exceed real underlying changes. Nevertheless, we can compile some consistent time trends. The time-diary data suggest that children under 12 are now watching less television than they did in the past, a decline of about 23%, or about six hours per week between 1981 and 1997 (Figure 2). The Monitoring the Future survey confirms that this trend also holds for adolescents (Figure 6): there has been a substantial decline in heavy television watching and an increase in the proportion of adolescents watching one hour or less daily (18). Monitoring the Future is a large and nationally representative study; the documented decline is therefore very likely to represent a true effect. There are no conclusive data on whether the large decline in television viewing is more than offset or only partially offset by new-media use (e.g., video games, computer).
Figure 6 Percentage of teenagers who spend one hour or less, two to three hours, or four hours or more watching television on average weekday, 1991–2001. Analysis based on annual data from Monitoring the Future (18). Reprinted with permission from Child Trends.
Bar chartA text description of this chart is also available
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
1 hour or less 28.9 29.6 29.7 31.4 31.5 31.9 32.7 32.9 32.4 34.8 35.6 34.1
2-3 hours 41.9 41.3 41.8 41.0 41.6 41.6 40.9 40.4 41.3 39.5 38.9 39.9
4 hours or more 29.1 29.0 28.5 27.6 26.9 26.6 26.4 26.7 26.3 25.7 25.5 26.0
Studying at Home
Many news reports and academic and popular books claim that homework overburdens children and limits learning with lack of physical activity and weight gain as major secondary consequences (19). Although most of the evidence for this idea is anecdotal, we noted above that time spent on home study by children aged six to eight did increase between 1981 and 1997. This fact has been cited in a number of news reports and in the book The End of Homework, which is subtitled, How Homework Disrupts Families, Overburdens Children, and Limits Learning (19).
Gill and Schlossman used several national surveys to provide a 50-year perspective on time spent on homework; a summary of their main findings follows (20).
The most systematic evidence on homework time at multiple grade levels across the country is found in answers to background questions asked of students taking the National Assessment of Educational Progress (NAEP). The data are nationally representative and based on large samples. Most variables show little change. Variables that do change, however, tend to show increases in the 1970s and early 1980s and declines in the 1990s (20). Figure 7 shows the trends: increases in the early 1980s were followed by a decline in the 1990s, resulting in figures for 1999 that are similar to those for 1980. Overall, increased time spent studying at home does not appear to be related to the obesity epidemic among youth.
Figure 7 Proportion of U.S. students doing one hour or more of homework. Data available for every other year starting 1978–1996 (except 1986), plus 1999; data not available for children aged 13 in 1978. Data from Gill and Schlossman (20). Copyright 2003 by the American Educational Research Association. Reprinted with permission from publisher.
Bar chartA text description of this chart is also available
1/1/1978 1/1/1980 1/1/1982 1/1/1984 1/1/1988 1/1/1990 1/1/1992 1/1/1994 1/1/1996 1/1/1999
age 13 31 39 38 41 36 39 37 35 34
age 17 33 33 37 40 38 37 36 39 35 35
Homework loads for elementary-level students deserve further analysis, however. The NAEP data collection for children aged nine (third or fourth grade) started only in 1984. Figure 8 shows two indicators: rates of homework assignment and the proportion of students doing more than one hour of homework the night before the survey. Clearly, the number of students with homework assigned the prior day has increased, consistent with an increase in average study time, but the one-hour-or-more trend line — flat or even declining — indicates that the daily time increase cannot have been very large. Regular daily assignments may also distribute work more evenly throughout the week and therefore decrease the probability of working one hour or more on any given day.
Figure 8 Homework trends for children aged nine years. Data available every other year 1984–1996 (except 1986), plus 1999. Data from Gill and Schlossman (20). Copyright 2003 by the American Educational Research Association. Reprinted with permission from publisher.
Bar chartA text description of this chart is also available
1/1/1984 1/1/1988 1/1/1990 1/1/1992 1/1/1994 1/1/1996 1/1/1999
Homework assigned yesterday 64 71 69 68 68 74 74
One hour or more 19 20 18 17 15 17 17
Combining several different data sets, we can track the proportion of high school students doing substantial homework over the half-century from 1948 though 1999. High school students during the late 1940s and early 1950s studied no more or less than their counterparts did in the 1970s, 1980s, and 1990s; only during the 1960s did homework time temporarily increase (20).
Summary
In contrast to adults, who now have more free time than in the past, children have less free time than previously because of increased time away from home, primarily in school, day care, and after-school programs (5,8,9). Participation in organized activities (including sports) also increased. Time spent in many sedentary activities — television viewing, conversations, or other passive leisure — declined just when obesity became a major concern. Unstructured playtime also declined except for in older children, but it is not clear whether this playtime was sedentary or active. The role of new media is not fully clear, although it is unlikely to have played a substantial role prior to 1999, except for children eight to 12 who spent a significant amount of time playing video games (considered unstructured playtime in the time-use data). But this age group also watched much less television at the turn of the century than in the 1980s.
Increased homework burdens and time studying at home have not caused a decrease in free time, contradicting a common belief in education circles. The great majority of American children at all grade levels now spend less than one hour studying on a typical day — an amount that has not changed substantially for at least 20 years. Compared to the large changes in other uses of time, it appears unlikely that changes in homework have altered the activity levels of children.
As time in structured settings away from home increases, so does the importance of physical activity in those settings. A substantial percentage of youth (about one third of high school students) is insufficiently active. An increase in structured time offers opportunities for interventions that may be more successful at expanding the number of youth who meet minimum-guideline criteria for strenuous physical activity than interventions targeted at diverse and unstructured home environments. In Part 2 of this report, we will look at trends in transportation, physical education, and diet.
This report was prepared for the Robert Wood Johnson Foundation. Tania Andreyeva and Hilary Rhodes provided research assistance.
Figures and Tables
Table 1 Media Use Among U.S. Youth (Hours per Day), 1999*
2-18 year-olds 2-7 year-olds 8-18 year-olds
Medium White Black Hispanic White Black Hispanic White Black Hispanic
Total media exposure 6:00a 7:56b 7:05c 4:04a 4:59b 4:25a 7:16a 9:52b 9:02b
Television 2:22a 3:56b 3:31c 1:43a 2:46b 2:20b 2:47a 4:41b 3:50c
Taped television shows 0:09a 0:17b 0:11a 0:04 0:02 0:02 0:12a 0:27b 0:18c
Videotapes 0:28 0:30 0:29 0:28 0:27 0:22 0:28 0:32 0:34
Movies 0:08a 0:19b 0:21b 0:01a 0:04b 0:01ab 0:13a 0:29b 0:35b
Video games 0:17a 0:25b 0:24b 0:08 0:08 0:09 0:23a 0:35b 0:35b
Print media 0:45a 0:45a 0:37b 0:47aa 0:42ab 0:38b 0:43a 0:47a 0:35b
Radio 0:38 0:40 0:43 0:22a 0:32b 0:25ab 0:49 0:45 0:56
CDs and tapes 0:50 0:43 0:49 0:22a 0:13b 0:23ab 1:09 1:03 1:08
Computer 0:22 0:20 0:19 0:07a 0:04b 0:04ab 0:31 0:31 0:29
* Within each row and age subgroup, only those mean times that do not share a common superscript differ from one another with statistical reliability. Those mean times without a superscript, or those that share a common superscript, do not differ by a large enough margin to ensure statistical reliability. Total media exposure is the sum of the amount of time children spend with each type of media. Data from Table 8c in Kids & Media @ The New Millennium (12) reprinted with permission from the Henry J. Kaiser Family Foundation.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Sturm R. Childhood obesity — what we can learn from existing data on societal trends, part 1. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0038.htm
==== Refs
1 Ogden CL Flegal KM Carroll MD Johnson CL 288 14 2002 1728 1732 JAMA Prevalence and trends in overweight among US children and adolescents,1999-2000 12365956
2 Hill JO Peters JC 280 5368 1998 1371 1374 Science Environmental contributions to the obesity epidemic 9603719
3 Hill JO Wyatt HR Reed GW Peters JC 299 5608 2003 853 855 Science Obesity and the environment: where do we go from here? 12574618
4 Institute of Medicine, Committee on Prevention of Obesity in Children and Youth 9 2004 National Academies Press Washington (DC) Preventing childhood obesity: health in the balance forthcoming
5 Hofferth SL Sandberg JF Owens T Hofferth S 2001 Elsevier Science New York Children at the millennium: where have we come from, where are we going? Advances in life course research Changes in American children's time, 1981-1997
6 Straits Times [Internet] Asian kids blame homework, heat for not exercising Singapore Singapore Press Holdings 2004 updated 2003 Nov 5; cited 2004 Mar 12 Available from: URL: http://straitstimes.asia1.com.sg/health/ story/0,4395,209633,00.html
7 Juster F Stafford FP 1985 Ann Arbor (MI) Institute for Social Research Time, goods, and well-being
8 Robinson JP Godbey GG Pennsylvania State University Press University Park (PA) 1999 2nd ed Time for life: the surprising ways Americans use their time
9 Sturm R 27 3S 2004 126 135 Am J Prev Med The economics of physical activity: societal trends and rationales for intervention 15450623
10 Hofferth SL Sandberg JF 2001 63 3 J Marriage Fam How American children use their time
11 Kinney DA Dunn JS Hofferth SL Paper presented at Conference on Work and Family: Expanding the Horizons 2000 Mar 3-4 San Francisco, CA Family strategies for managing the time crunch
12 Kaiser Family Foundation Kids & media @ the new millenium [Internet] Menlo Park (CA) The Henry J. Kaiser Family Foundation 2003 updated 2003 Dec 10; cited 1999 Nov 15 Available from: URL: http://www.kff.org/entmedia/1535-index.cfm
13 Kaiser Family Foundation Key facts: children and videogames [Internet] Menlo Park (CA) The Henry J. Kaiser Family Foundation 2003 updated 2004 March 23; cited 2002 Oct 24 Available from: URL: http://www.kff.org/entmedia/3271-index.cfm
14 Kaiser Family Foundation Zero to six: electronic media in the lives of infants, toddlers and preschoolers [Internet] Menlo Park (CA) The Henry J. Kaiser Family Foundation 2003 updated 2004 March 23; cited 2003 Oct 28 Available from: URL: http://www.kff.org/entmedia/3378.cfm
15 Robinson TN 279 12 1998 959 960 JAMA Does television cause childhood obesity? 9544774
16 Robinson TN 282 16 1999 1561 1567 JAMA Reducing children's television viewing to prevent obesity: a randomized controlled trial 10546696
17 Robinson TN 48 4 2001 1017 1025 Pediatr Clin North Am Television viewing and childhood obesity 11494635
18 Child Trends Data Bank Monitoring the Future [Internet] Washington (DC) Child Trends Available from: URL: http://www.childtrendsdatabank.org/ indicators/55WatchingTV.cfm
19 Kravolec E Buell J Beacon Press Boston 2000 The end of homework: how homework disrupts families, overburdens children, and limits learning
20 Gill BP Schlossman SL 25 3 2003 319 337 Educational Evaluation and Policy Analysis A nation at rest: the American way of homework
|
15670465
|
PMC1323315
|
CC BY
|
2022-01-31 23:29:44
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A12
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0066
Review
PEER REVIEWEDThe Vital Link Between Chronic Disease and Depressive Disorders
Chapman Daniel P PhD Emerging Investigations and Analytic Methods Branch, Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention
4770 Buford Highway NE, Mailstop K-45, Atlanta, GA 30341 [email protected]
770-488-5463
Perry Geraldine S DrPH Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, Ga
Strine Tara W MPH Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, Ga
1 2005
15 12 2004
2 1 A142005
Introduction
Chronic diseases have assumed an increasingly important role in public health research and intervention. Without treatment, depressive disorders characteristically assume a chronic course and are expected, by 2020, to be second only to heart disease in the global burden of disease. Thus, understanding the relationship between depressive disorders and chronic disease appears vital to public health assessment and health care delivery.
Methods
Articles for review were primarily identified by a Medline search emphasizing the subject headings mental disorders or depression crossed with selected chronic diseases and conditions including asthma, arthritis, cardiovascular disease, cancer, diabetes, and obesity.
Results
Mental illnesses — most specifically, depressive disorders — were associated with increased prevalence of chronic diseases. This association between depression and chronic disease appears attributable to depressive disorders precipitating chronic disease and to chronic disease exacerbating symptoms of depression. The complex interrelationship between depressive disorders and chronic disease has important implications for both chronic disease management and the treatment of depression.
Conclusion
Depressive disorders assume an important role in the etiology, course, and outcomes associated with chronic disease. Multivariate community-based research and intervention fostering the detection and treatment of depressive disorders is needed, as is further examination of the role exerted by mental illnesses other than depression in the pathogenesis of chronic disease.
==== Body
Introduction
Recent research indicates that seven out of 10 office visits to a primary care practitioner concern chronic diseases (1). As the management of chronic diseases has assumed an increasingly vital role in health care delivery, recognition of the importance of depressive disorders has also grown. By 2020, depression is expected to be second only to heart disease as a source of the global burden of disease (2). As chronic disease and depressive disorders are increasingly recognized as major impediments to health, understanding the connection between them becomes of utmost importance to providing quality health care.
Despite the growing recognition of the importance of both chronic disease and depressive disorders to the health of individuals and communities, research examining their interrelationship has been the subject of surprisingly little empirical review. A Medline search for literature reviews emphasizing both chronic disease and depression yielded only two publications. These articles addressed factors germane to health-service costs (3) and individual characteristics precipitating the onset of depressive disorders subsequent to the development of chronic disease (4). While raising important concerns, previous reviews were deliberately limited in scope and did not generally address disease-specific variables potentially underlying the associations between depressive disorders and a number of chronic diseases. To better address this issue, we reviewed the research literature examining the relationships between depressive disorders and prevalent chronic diseases that are also of programmatic relevance to the work of the Centers for Disease Control and Prevention.
Methods
Articles included in this review were primarily identified through a Medline search of the terms mental disorders or depression crossed with the chronic diseases and conditions asthma, arthritis, cardiovascular disease, cancer, diabetes, and obesity. These chronic diseases were selected because they have been identified as highly prevalent and as major sources of morbidity and mortality among U.S. adults. Studies included for review were generally limited to empirical investigations that provided definitional or diagnostic criteria for both depression and relevant chronic diseases and that featured a specified time course.
Results
Asthma
Nearly 50% of asthma patients may suffer from clinically significant depressive symptoms (5), which have been, in part, attributed to the stress of having a chronic illness (6). In particular, persons with asthma who experience disruptive symptoms, such as dyspnea and nighttime awakening, are at increased risk for major depression (7). The presence of depression among persons with asthma assumes particular gravity because increased depressive symptoms have been associated with poorer asthma outcomes (8), such as impaired voluntary activation of the diaphragm (9). In clinical samples of children and adolescents, asthma has been associated with the presence of an anxiety disorder (10) and with anxious depressive symptoms among youth with moderate and severe persistent asthma (11).
It appears that the symptoms — rather than the diagnosis — of asthma are associated with depression or anxiety (12): 87.5% of persons with frequent asthma attacks manifest psychopathology, compared with 25% of persons with less frequent attacks (13). Similarly, among persons whose asthma is difficult to control, psychopathology — primarily anxiety and depressive symptoms — was associated with more frequent visits to primary care providers and emergency departments and with more hospitalizations (14). Psychological morbidity is associated with poor adherence to medication regimens (15), and mothers of children with asthma are at risk for increased depressive symptoms (16). Early assessment and intervention addressing depressive disorders improves treatment adherence and outcomes and may also decrease mortality (17).
Cognitive behavioral therapy (CBT) — in which the individual is instructed to monitor and challenge self-negating thoughts — has yielded a significant decrease in asthma symptoms and depression (18). Likewise, physical inactivity has been speculated to augment the strength of the association between perceived stressors and depression in persons with asthma, suggesting that exercise may ameliorate this association and decrease the likelihood of depression in this population (19).
Arthritis
Depression and/or anxiety are among the most commonly reported concerns by persons with arthritis (20). Screening of patients with arthritis revealed that depression was associated with activity restriction, further suggesting that nonpsychiatric physicians should be aware of the mental health status of patients with chronic illnesses (21). Persons with arthritis experiencing arthritis-based disability (22) and the recurrence of arthritic symptoms (23) reported greater depression. Research on adolescents and young adults with arthritis has found that functional status is significantly correlated with depression, self-esteem, and loneliness (24); significantly greater depression was reported among those experiencing more severe symptoms (24).
Research on the Arthritis Self-Management Program found that participation in this intervention had a positive effect on perceptions of self-efficacy, communication with physicians, fatigue, anxiety, pain, and depression (25). A randomized trial of an intervention designed to improve mood (antidepressant medications and/or six to eight sessions of psychotherapy) improved depression and fostered improvements in functional status and quality of life (26). CBT has proven particularly effective in ameliorating depressive symptoms when initiated early in the course of rheumatoid arthritis (27) and when tailored to the concerns reported by persons with rheumatoid arthritis, such as fatigue or mood (28). Similarly, antidepressant medication has been associated with significant improvements in both psychological status and health status in persons with rheumatoid arthritis (29).
Rest and inactivity were previously considered to be reasonable therapeutic approaches in the management of osteoarthritis until it was recognized that physical inactivity contributed to disability and impaired functioning. Subsequent research suggests that a tailored program of aerobic or resistance-based exercise may be an important component of self-managing osteoarthritis (30). Aerobic exercise has been found to both ameliorate depressive symptoms and to reduce disability and pain among persons with arthritis (31).
Cardiovascular disease
Depressive disorders have been associated with risk factors for cardiovascular disease (CVD), such as smoking and physical inactivity (32), and mental illness, in general, has been associated with increased mortality due to CVD (33). In general, persons who are depressed are much more likely to develop coronary artery disease (34), and meta-analyses reveal that the relative risk for developing heart disease in individuals with depression or depressive symptoms is approximately 1.6 times greater than among nondepressed persons (35,36), which is more than the risk conferred by passive smoking (36). A stronger effect size was reported for clinical depression than for depressive symptoms, suggesting the presence of a dose-response relationship (35). Depression has been positively associated with the metabolic syndrome among women (but not men) younger than 40 years (37), suggesting that early detection and treatment of depression may potentially forestall the risk of cardiovascular disease among women.
Depression or depressive symptoms are also predictive of stroke (38): persons with significant depressive symptoms are approximately twice as likely as those with few depressive symptoms to have a stroke within 10 years (39). Moreover, depression is associated with an increased risk for stroke morbidity and mortality (40).
In addition to being a predictor of stroke, depression commonly develops after a stroke, especially after a stroke affecting the left hemisphere of the brain (41). More than half of patients experiencing a stroke report depressive symptoms within 18 months of having a stroke (42). Post-stroke depression has been associated with impairments in response to rehabilitation (43) and with increased mortality up to two years following the stroke (44). Antidepressant treatment of post-stroke depression is warranted and, in addition to alleviating depression, may foster recovery of cognitive function (45) and significantly increase survival (46).
Depressive disorders also appear related to the occurrence of heart attack, or myocardial infarction (MI). Persons with a history of major depression are more than four times as likely to have an MI than those with no history of depression (47), and high levels of depressive symptoms are associated with an increased risk of MI (48).
Approximately one in six persons who have experienced an MI suffer from major depression, and at least twice that many experience significant depressive symptoms (49). Patients who have had an MI and are also depressed have more medical comorbidities (50) and cardiac complications (51) and are at greater risk for mortality (52) than their nondepressed peers. Increased mortality is also evident in persons who had an MI and who manifest very low levels of depressive symptoms (53), underscoring the importance of mental health to physical health outcomes.
Persons with depression following an MI are less likely to adhere to recommended lifestyle and behavioral changes, potentially increasing their risk for subsequent cardiac events (54). This is particularly unfortunate because cardiac rehabilitation has been found to improve depressive symptoms (55). However, the use of a specific class of antidepressant medications — the selective serotonin reuptake inhibitors (SSRIs) — may, in addition to their beneficial effect on depression, exert antiplatelet effects protecting against MI (56). In addition to being safer in overdose (57), SSRIs are also less likely to induce arrhythmia than other classes of antidepressant medications (58). It has further been concluded that the combination of CBT with an SSRI is frequently the most effective treatment of depression in persons with CVD (59).
Cancer
Estimates of the prevalence of psychiatric disorders among persons with cancer vary widely, depending on the type of cancer and its clinical stage. Previous research indicates that nearly 50% of patients newly admitted to a cancer center met diagnostic criteria for a psychiatric disorder. Adjustment disorders — distress related to a specific precipitant — comprised 68% of these diagnoses, although many of those diagnosed reported anxiety or depression as a central symptom (60). Among cancer patients judged terminally ill, 53% met psychiatric diagnostic criteria, with delirium — a fluctuating change in cognition and disturbance in consciousness — being the most frequently diagnosed disorder (61).
In addition to delirium, cancer patients also suffer from depression and anxiety (62); 21% of cancer patients are reported to be depressed (63). Depression assumes particular significance in the care of individuals with cancer, because it has been associated with a desire for hastened death among terminally ill cancer patients (64), and increased depressive symptoms are inversely related to survival (65). Of cancer patients in an intensive care unit who were assessed as being at high risk for hospital death, 40% reported depression (66), suggesting that diagnosis and treatment of depression are inadequate. Strikingly, among cancer patients undergoing chemotherapy and experiencing anemia-related fatigue, improved hemoglobin levels have been reported to reduce depressive symptoms (67), further suggesting the importance of physical health to mental health status.
A previous survey of psychotropic prescription practices at five major oncology centers revealed hypnotics to be the most widely prescribed drugs, with antidepressants comprising only 1% of psychotropic prescriptions (68). Subsequent research, however, has indicated an increase of antidepressant use in community cancer care, with 19.2% of breast, 11% of colon, and 13.7% of lung cancer patients receiving antidepressants during a two-year interval (69).
Despite the observation that both antidepressants and psychotherapy are effective in treating depression in patients with cancer, research on antidepressant pharmacotherapy and psychotherapy among persons with cancer has been characterized as largely lacking randomized placebo-controlled trials (70). Moreover, antidepressant prescription has been found to be associated with factors not specifically related to psychopathology, such as patient age or the presence of pain (69), and some speculate that most depressed patients with cancer do not need medication (71). This belief, however, may reflect the misconception that depression is a "natural" response to cancer and does not merit systematic diagnosis and treatment (72).
Research suggests that depression in persons with cancer is amenable to treatment. Among cancer patients with a life expectancy of at least 12 months, CBT has been associated with significantly decreased depressive symptoms across a four-month interval (73). CBT has also been associated with decreased pain, reduced symptomatic distress (74), and subsequent improvement in cellular immune function (75).
Adoption of a depression screening program and antidepressant algorithm by oncologists resulted in significant improvements in mood and quality of life among cancer patients (76). Similarly, a placebo-controlled trial of antidepressant medication in advanced cancer patients demonstrated that antidepressant therapy decreased depressive symptoms and improved patient assessments of quality of life (77). In addition to reducing the risk of depression, data suggest that physical activity may also decrease the risk of colon, breast, and lung cancer (78).
Diabetes
Elevated rates of depression have consistently been associated with diabetes (79), with results of a meta-analysis indicating depression is twice as prevalent among persons with diabetes than it is among persons without diabetes (80). While it has been proposed that depressive symptoms may be a risk factor for the development of diabetes, this association is most pronounced at high levels of depressive symptoms and, interestingly, only observed among persons with less than a high school education (81). These findings suggest that factors associated with low socioeconomic status may contribute to the development of diabetes among persons with substantial depressive symptoms.
Comorbid depressive symptoms or depression among persons with diabetes have been associated with adaptation to the illness (82), diabetic-related complications (79), unemployment (83), and illness intrusiveness, a construct defined as the degree to which diabetes disrupts valued activities and interests (84). As is true in the general population, depression was more prevalent among women than among men with diabetes (80,85) and among younger adults (85). Depressive symptoms are more likely to persist among persons with multiple diabetic-related complications and those with less than a high school education (79). In a prospective community-based study, baseline depressive symptoms were positively associated with fasting insulin levels and physical inactivity (86). A diagnosis of diabetes and self-reported depression were positively associated with sedentariness in both bivariate and multivariate analyses (87). Compared with their nondepressed peers, patients with diabetes who were diagnosed with depression were more likely to report frequent overeating of sweets and high-fat foods and were less satisfied with their ability to adhere to a diabetic diet away from home (88).
Despite the availability of measures to screen for depression, it is estimated that less than 25% of those with depression are diagnosed and treated (89). This is particularly disconcerting because the treatment of depression appears to be associated with improved glycemic control (90). Furthermore, because depression is associated with diabetic complications (91), treatment of depression may also reduce diabetes-related disability. Compared with their nondepressed peers, persons with diabetes and depression have higher ambulatory-care use and fill more prescriptions. Total health expenditures for persons with diabetes and depression were 4.5 times higher than for those without depression: $247 million compared with $55 million (85).
Research has revealed that both CBT (90) and antidepressant pharmacotherapy (92) are associated with decreased severity of depression among persons with diabetes and with improved glycemic control. Thus, in addition to preventing needless suffering, the treatment of depression among persons with diabetes offers the added promises of substantial financial savings and improved medical care of these individuals.
Obesity
Several studies have indicated an association between psychopathology, including depressive symptoms, and high body mass index (BMI), or obesity. The relationship between obesity and psychopathology differs among men and women, with a BMI ≥30 among women associated with nearly a 50% increase in the lifetime prevalence of depressive disorders compared with nonobese women (93). In contrast, while BMI has not been found to be related to measures of mental well-being among men, abdominal obesity or a high waist/hip ratio has been associated with an increased prevalence of both depressive symptoms (94) and antidepressant medication use (95).
Although it is important to note that most overweight or obese persons do not suffer from mood disorders (96), significant positive associations have been reported between BMI and depressive symptoms (97). It has been posited that a common pathophysiology may underlie both obesity and depression. The neurotransmitters serotonin and norepinephrine are involved in regulating both mood and body weight and, logically, in the treatment of both depression and obesity (98). Antidepressant medications available before the development of SSRIs frequently induced weight gain; newer agents generally do not stimulate appetite, thus making them potentially useful in depressed patients who do not wish to gain weight (99).
Previous longitudinal research has examined the relationship between depressive symptoms or psychological well-being and weight gain. Women who were either normal weight or overweight at baseline and who had experienced a recent weight gain scored lower on a measure of psychological well-being than women who had not gained weight (100). Similarly, persons who were overweight and depressed at baseline demonstrated a significantly increased likelihood of subsequent weight gain relative to those who were not depressed. Among the highest quintile of baseline BMI, this relationship was stronger among women (with an odds ratio of 2.2) than men (with an odds ratio of 1.3) (101).
Cognitive behavioral interventions have been useful in managing obesity, largely by modifying eating behaviors and dietary choices in addition to decreasing psychological distress and sedentariness (102). In addition to fostering weight loss, CBT has been found to improve self-reported mental health among obese persons (103). However, psychosocial difficulties have been associated with weight gain following initial weight loss among obese individuals who had received CBT (104), with long-term CBT compliance being particularly low among persons with binge-eating behaviors (105).
Strikingly, children and adolescents with major depressive disorder appear to manifest an increased risk for subsequently becoming overweight (96), suggesting that both depressive disorders and their treatment are relevant to the prevalence of obesity. The relationship between obesity and depressive disorders thus appears to be reciprocal, with advances in the recognition and treatment of each of these diseases potentially fostering improved mental and physical health.
Discussion
Research examining the association between depressive disorders and chronic disease suggests that timely diagnosis and treatment of psychiatric disorders could greatly affect the impact of chronic disease. The presence of mental illness may be an important contributor to the etiology of chronic disease. Thus, the promotion of mental health would likely result in reducing a considerable proportion of the burden of chronic disease. Similarly, the presence of depressive disorders often adversely affects the course and complicates the treatment of chronic disease. It is important to remember that untreated depressive disorders characteristically assume a chronic course (106), thereby adding to the burden of chronic disease in their own right. Multivariate investigation of the associations among depressive disorders, chronic disease, and a variety of medical and sociodemographic characteristics would provide valuable insights into contemporary notions of health and quality of life.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Chapman DP, Perry GS, Strine TW. The vital link between chronic disease and depressive disorders. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0066.htm
==== Refs
1 Veale BM 2003 179 247 249 Med J Aust Meeting the challenge of chronic illness in general practice 12924971
2 Murray JL Lopez AD 1996 Summary: the global burden of disease Boston (MA) Harvard School of Public Health
3 Katon WJ 2003 54 216 226 Biol Psychiatry Clinical and health services relationships between major depression, depressive symptoms, and general medical illness 12893098
4 Burke P Elliott M 1999 40 5 17 Psychosomatics Depression in pediatric chronic illness: a diasthesis-stress model 9989116
5 Mancuso CA Peterson MG Charlson ME 2000 15 301 310 J Gen Intern Med Effects of depressive symptoms on health-related quality of life in asthma patients 10840265
6 Mrazek DA 2003 12 459 471 Child Adolesc Clin N Am Psychiatric symptoms in patients with asthma causality, comorbidity, or shared genetic etiology
7 Goldney RD Ruffin R Fisher LJ Wilson DH 2003 178 437 441 Med J Australia Asthma symptoms associated with depression and lower quality of life: a population survey 12720509
8 Mancuso CA Rincon M McCulloch CE Charlson ME 2001 39 1326 1338 Med Care Self-efficacy, depressive symptoms, and patients' expectations predict outcomes in asthma 11717574
9 Allen GM Hickie I Gandevia SC McKenzie DK 1994 49 881 884 Thorax Impaired voluntary drive to breathe: a possible link between depression and unexplained ventilatory failure in asthmatic patients 7940427
10 Ortega AN Huertas SE Canino G Ramirez R Rubio-Stipec M 2002 190 275 281 J Nerv Ment Dis Childhood asthma, chronic illness, and psychiatric disorders 12011605
11 Vila G Nollet-Clemencon C de Blic J Mouren-Simeoni MC Scheinmann P 1998 7 137 144 Eur Child Adolesc Psychiatry Asthma severity and psychopathology in a tertiary care department for children and adolescent 9826300
12 Janson C Bjornsson E Hetta J Boman G 1994 149 930 934 Am J Respir Crit Care Med Anxiety and depression in relation to respiratory symptoms and asthma 8143058
13 Erhabor GE Kuteyi F Obembe F 2002 94 987 993 J Natl Med Assoc Asthma: the psychosocial impact among a sample of south western Nigerians 12443002
14 ten Brinke A Ouwerkerk ME Zwinderman AH Spinhoven P Bel EH 2001 163 1093 1096 Am J Respir Crit Care Med Psychopathology in patients with severe asthma is associated with increased health care utilization 11316641
15 Cluley S Cochrane GM 2001 95 37 39 Respir Med Psychological disorder in asthma with poor control and poor adherence to inhaled steroids 11207015
16 Bartlett SJ Kolodner K Butz AM Eggleston P Malveaux FJ Rand CS 2001 155 347 353 Arch Pediatr Adolesc Med Maternal depressive symptoms and emergency department use among inner-city children with asthma 11231800
17 Galil N 2000 12 331 335 Curr Opin Pediatr Depression and asthma in children 10943812
18 Grover N Kumaraiah V Prasadrao PS D'souza G 2002 50 896 900 J Assoc Physicians India Cognitive behavioural intervention in bronchial asthma 12126343
19 Hurwitz EL 2003 13 358 368 Ann Epidemiol Do asthma and physical inactivity influence the associations of personal and job stressors with perceived stress and depression? Findings from the 1998-1999 California Work and Health Survey 12821275
20 Sotosky JR McGrory CH Metzger DS DeHoratius RJ 1992 157 162 5 Arthritis Care Res Arthritis problem indicator: preliminary report on a new tool for use in the primary care setting 1457491
21 Vali FM Walkup J 1998 36 1073 1084 Med Care Combined medical and psychological symptoms: impact on disability and health care utilization of patients with arthritis 9674624
22 Husaini BA Moore ST 1990 15 253 260 Health Soc Work Arthritis disability, depression, and life satisfaction among black elderly people 2276687
23 Musil CM Morris DL Haug MR Warner CB Whelan AT 2001 52 1729 1740 Soc Sci Med Recurrent symptoms: well-being and management 11327144
24 Taal E Rasker JJ Timmers CJ 1997 24 994 997 J Rheumatol Measures of physical function and emotional well-being for young adults with arthritis 9150098
25 Barlow JH Turner AP Wright CC 2000 15 665 680 Health Educ Res A randomized controlled study of the Arthritis Self-Management Programme in the UK 11142075
26 Lin EH Katon W Von Korff M Tang L Williams JW Jr Kroenke K IMPACT Investigators 2003 290 2428 2429 JAMA Effect of improving depression care on pain and functional outcomes among older adults with arthritis: a randomized controlled trial 14612479
27 Sharpe L Sensky T Timberlake N Ryan B Allard S 2003 42 435 441 Rheumatology (Oxford) Long-term efficacy of a cognitive behavioural treatment from a randomized controlled trial for patients recently diagnosed with rheumatoid arthritis 12626793
28 Evers AW Kraaimaat FW van Riel PL de Jong AJ 2002 100 141 153 Pain Tailored cognitive-behavioral therapy in early rheumatoid arthiritis for patients at risk: a randomized controlled trial 12435467
29 Parker JC Smarr KL Slaughter JR Johnston SK Priesmeyer ML Hanson KD 2003 766 777 49 Arthritis Rheum Management of depression in rheumatoid arthritis: a combined pharmacologic and cognitive-behavioral approach 14673962
30 Clyman B 2001 3 520 523 Curr Rheumatol Rep Exercise in the treatment of osteoarthritis 11709115
31 Penninx BW Rejeski WJ Pandya J Miller ME Di Bari M Applegate WB 2002 57 P124 P132 J Gerontol B Psychol Sci Soc Sci Exercise and depressive symptoms: a comparison of aerobic and resistance exercise effects on emotional and physical function in older persons with high and low depressive symptomatology 11867660
32 Hayward C 1995 17 129 138 Epidem Rev Psychiatric illness and cardiovascular disease risk
33 Davidson S Judd F Jolley D Hocking B Thompson S Hyland B 2001 35 196 202 Aust N Z J Psychiatry Cardiovascular risk factors for people with mental illness 11284901
34 Nemeroff CB Musselman DL Evans DL 8 Suppl 1 1998 71 79 Depress Anx Depression and cardiac disease
35 Rugulies R 2002 23 51 61 Am J Prev Med Depression as a predictor for coronary heart disease. A review and meta-analysis 12093424
36 Wulsin LR Singal BM 2003 65 201 210 Psychosom Med Do depressive symptoms increase the risk for the onset of coronary disease? A systematic quantitative review 12651987
37 Kinder LS Carnethon MR Palaniappan LP King AC Fortmann SP 2004 66 316 322 Psychosom Med Depression and the metabolic syndrome in young adults: Findings from the Third National Health and Nutrition Examination Survey 15184689
38 Jonas BS Mussolino ME 2000 62 463 471 Psychosom Med Symptoms of depression as a prospective risk factor for stroke 10949089
39 Ohira T Iso H Satoh S Sankai T Tanigawa T Ogawa Y 2001 32 903 908 Stroke Prospective study of depressive symptoms and risk of stroke among Japanese 11283390
40 Ramasubbu R Patten SB 2003 48 250 257 Can J Psychiatry Effect of depression on stroke morbidity and mortality 12776392
41 Narushima K Kosier JT Robinson RG 2003 15 422 430 J Neuropsychiatry Clin Neurosci A reappraisal of poststroke depression, intra- and inter-hemispheric lesion location using meta-analysis 14627768
42 Berg A Palomaki H Lehtihalmes M Lonnqvist J Kaste M 2003 34 138 143 Stroke Poststroke depression: an 18-month follow-up 12511765
43 Gillen R Tennen H McKee TE Gernert-Dott P Affleck G 2001 82 1645 1649 Arch Phys Med Rehab Depressive symptoms and history of depression predict rehabilitation efficiency in stroke patients
44 House A Knapp P Bamford J Vail A 2001 32 696 701 Stroke Mortality at 12 and 24 months after stroke may be associated with depressive symptoms at 1 month 11239189
45 Kimura M Robinson RG Kosier JT 2000 31 1482 1486 Stroke Treatment of cognitive impairment after poststroke depression: a double-blind treatment trial 10884441
46 Jorge RE Robinson RG Arndt S Starkstein S 2003 160 1823 1829 Am J Psychiatry Mortality and poststroke depression: a placebo controlled trial of antidepressants 14514497
47 Pratt LA Ford DE Crum RM Armenian HK Gallo JJ Eaton WW 1996 94 3123 3129 Circulation Depression, psychotropic medication, and risk of myocardial infarction. Prospective data from the Baltimore ECA follow-up 8989119
48 Barefoot JC Schroll M 1996 93 1976 1980 Circulation Symptoms of depression, acute myocardial infarction, and total mortality in a community sample 8640971
49 Ziegelstein RC 2001 286 1621 1627 Depression in patients recovering from a myocardial infarction JAMA
50 Watkins LL Schneiderman N Blumenthal JA Sheps DS Catellier D Taylor CB ENRICHD Investigators 2003 146 48 54 Am Heart J Cognitive and somatic symptoms of depression are associated with acute myocardial infarction 12851607
51 Lauzon C Beck CA Huynh T Dion D Racine N Carignan S 2003 168 547 552 CMAJ Depression and prognosis following hospital admission because of acute myocardial infarction 12615746
52 Carney RM Blumenthal JA Catellier D Freedland KE Berkman LF Watkins LL 2003 92 1277 1281 Am J Cardiol Depression as a risk factor for mortality after acute myocardial infarction 14636903
53 Bush DE Ziegelstein RC Tayback M Richter D Stevens S Zahalsky H 2001 88 337 341 Am J Cardiol Even minimal symptoms of depression increase mortality risk after acute myocardial infarction 11545750
54 Ziegelstein RC Fauerbach JA Stevens SS Romanelli J Richter DP Bush DE 2000 160 1818 1823 Arch Int Med Patients with depression are less likely to follow recommendations to reduce cardiac risk during recovery from a myocardial infarction 10871976
55 Milani RV Lavie CJ 1998 81 1233 1236 Am J Cardiol Prevalence and effects of cardiac rehabilitation on depression in the elderly with coronary heart disease 9604957
56 Serebruany VL Glassman AH Malinin AI Nemeroff CB Musselman DL van Zyl LT Sertraline AntiDepressant Heart Attack Randomized Trial Study Group 2003 108 939 944 Circulation Platelet/endothelial biomarkers in depressed patients treated with selective serotonin reuptake inhibitor sertraline after acute coronary events: the Sertraline AntiDepressant Heart Attack Randomized Trial (SADHART) Platelet Substudy 12912814
57 Glassman AH 59 suppl 15 1998 13 18 J Clin Psychiatry Cardiovascular effects of antidepressant drugs: updated 9786306
58 Cleophas TJ 1997 11 111 118 Drugs Aging Depression and myocardial infarction. Implications for medical prognosis and options for treatment 9259174
59 Guck TP Kavan MG Elsasser GN Barone EJ 2001 64 641 648 Am Fam Physician Assessment and treatment of depression following myocardial infarction 11529263
60 Derogatis LR Morrow GR Fetting J Penman D Piasetsky S Schmale AM 1983 249 751 757 JAMA The prevalence of psychiatric disorders among cancer patients 6823028
61 Minagawa H Uchitomi Y Yamawaki S Ishitani K 1996 78 1131 1137 Cancer Psychiatric morbidity in terminally ill cancer patients. A prospective study 8780554
62 Massie MJ Holland JC 1987 71 243 258 Med Clin North A The cancer patient with pain: psychiatric complications and their management
63 Bodurka-Bevers D Basen-Engquist K Carmack CL Fitzgerald MA Wolf JK de Moor C 2000 78 302 308 Gynecol Oncol Depression, anxiety, and quality of life in patients with epithelial ovarian cancer 10985884
64 Breitbart W Rosenfeld B Pessin H Kaim M Funesti-Esch J Galietta M 2000 284 2907 2911 JAMA Depression, hopelessness, and desire for hastened death in terminally ill patients with cancer 11147988
65 Meyer HA Sinnott C Seed PT 17 2003 604 607 Palliat Med Depressive symptoms in advanced cancer. Part 2. Depression over time; the role of the palliative care professional 14594151
66 Nelson JE Meier DE Oei EJ Nierman DM Senzel RS Manfredi PL 2001 29 277 282 Crit Care Med Self-reported symptom experience of critically ill cancer patients receiving intensive care 11246306
67 Kallich JD Tchekmedyian NS Damiano AM Shi J Black JT Erder MH 16 9 Suppl 10 2002 117 124 Oncology Psychological outcomes associated with anemia-related fatigue in cancer patients 12380961
68 Derogatis LR Feldstein M Morrow G Schmale A Schmitt M Gates C 1979 44 1919 1929 Cancer A survey of psychotropic drug prescriptions in an oncology population 40688
69 Ashbury FD Madlensky L Raich P Thompson M Whitney G Hotz K 2003 11 278 285 Support Care Cancer Antidepressant prescribing in community cancer care 12720073
70 Fisch M 2004 32 105 111 J Natl Cancer Inst Monogr Treatment of depression in cancer 15263050
71 Schuler US 2002 325 1115 BMJ Most patients depressed by cancer do not need drugs 12424180
72 Endicott J 1984 53 2243 2249 Cancer Measurement of depression in patients with cancer 6704912
73 Greer S Moorey S Baruch JD Watson M Robertson BM Mason A 1992 304 675 680 BMJ Adjuvant psychological therapy for patients with cancer: a prospective randomised trial 1472184
74 Dalton JA Keefe FJ Carlson J Youngblood R 2004 5 3 18 Pain Manag Nurs Tailoring cognitive-behavioral treatment for cancer pain 14999649
75 McGregor BA Antoni MH Boyers A Alferi SM Blomberg BB Carver CS 2004 56 1 8 J Psychosom Res Cognitive-behavioral stress management increases benefit finding and immune function among women with early-stage breast cancer 14987957
76 Passik SD Kirsh KL Theobald D Donaghy K Holtsclaw E Edgerton S 2002 24 318 327 J Pain Symptom Manage Use of a depression screening tool and a fluoxetine-based algorithm to improve the recognition and treatment of depression in cancer patients. A demonstration project 12458113
77 Fisch MJ Loehrer PJ Kristeller J Passik S Jung SH Shen J Hoosier Oncology Group 2003 21 1937 1943 J Clin Oncol Fluoxetine versus placebo in advanced cancer outpatients: a double-blinded trial for the Hoosier Oncology Group 12743146
78 Varo Cenarruzabeitia JJ Martinez Hernandez JA Martinez-Gonzalez MA 2003 121 665 672 Med Clin (Barc) Benefits of physical activity and harms of inactivity 14642230
79 Peyrot M Rubin RR 1997 20 585 590 Diabetes Care Levels and risks for depression and anxiety symptomatology among diabetic adults 9096984
80 Anderson RJ Freedland KE Clouse RE Lustman PJ 2001 24 1069 1078 Diabetes Care The prevalence of comorbid depression in adults with diabetes: a meta-analysis 11375373
81 Carnethon MR Kinder LS Fair JM Stafford RS Fortmann SP 2003 158 416 423 Am J Epidemiol Symptoms of depression as a risk factor for incident diabetes: findings from the National Health and Nutrition Examination Epidemiologic Follow-up Study, 1971-1992 12936896
82 Lernmark B Persson B Fisher L Rydelius PA 1999 16 14 22 Diab Med Symptoms of depression are important to psychological adaptation and metabolic control in children with diabetes mellitus
83 Friis R Nanjundappa G 1986 23 471 475 Soc Sci Med Diabetes, depression, and employment status 3764497
84 Talbot F Nouwen A Gingras J Belanger A Audet J 1999 18 537 542 Health Psychol Relations of diabetes intrusiveness and personal control to symptoms of depression among adults with diabetes 10519470
85 Egede L Zheng D Simpson K 2002 25 464 470 Diabetes Care Comorbid depression is associated with increased health care use and expenditures in individuals with diabetes 11874931
86 Golden SH Williams JE Ford DE Yeh HC Paton Sanford C Nieto FJ Atherosclerosis Risk in Communities study 2004 27 429 435 Diabetes Care Depressive symptoms and the risk of type 2 diabetes: the Atherosclerosis Risk in Communities study 14747224
87 Yancey AK Wold CM McCarthy WJ Weber MD Lee B Simon PA 2004 27 146 152 Am J Prev Med Physical inactivity and overweight among Los Angeles County adults 15261902
88 Ahlgren SS Shultz JA Massey LK Hicks BC Wysham C 2004 13 819 832 Qual Life Res Development of a preliminary diabetes dietary satisfaction and outcomes measure for patients with type 2 diabetes 15129892
89 Rubin RR Ciechanowski P Egede LE Lin EH Lustman PJ 2004 4 119 125 Curr Diab Rep Recognizing and treating depression in patients with diabetes 15035972
90 Lustman PJ Griffith LS Freedland KE Kissel SS Clouse RE 1998 129 613 621 Ann Intern Med Cognitive behavior therapy for depression in type 2 diabetes mellitus. A randomized, controlled trial 9786808
91 de Groot M Anderson R Freedland KE Clouse RE Lustman PJ 2001 63 619 630 Psychosom Med Association of depression and diabetes complications: a meta-analysis 11485116
92 Lustman PJ Freedland KE Griffith LS Clouse RE 2000 23 618 623 Diabetes Care Fluoxetine for depression in diabetes: a randomized double-blind placebo controlled trial 10834419
93 Becker ES Margraf J Turke V Soeder U Neumer S 25 Suppl 1 2001 S5 S9 Int J Obes Relat Metab Disord Obesity and mental illness in a representative sample of young women 11466578
94 Rosmond R Bjorntorp P 2000 26 90 94 Behav Med Quality of life, overweight, and body fat distribution in middle-aged men 11147295
95 Rosmond R Lapidus L Marin P Bjorntorp P 1996 4 245 252 Obes Res Mental distress, obesity and body fat distribution in middle-aged men 8732958
96 McElroy SL Kotwal R Malhotra S Nelson EB Keck PE Nemeroff CB 2004 65 634 651 J Clin Psychiatry Are mood disorders and obesity related? A review for the mental health professional 15163249
97 Johnston E Johnson S McLeod P Johnston M 2004 95 179 183 Can J Public Health The relation of body mass index to depressive symptoms 15191118
98 Licinio J Wong ML 2003 25 196 197 Rev Bras Psiquiatr The interface of obesity and depression: risk factors for the metabolic syndrome 15328542
99 Schatzberg AF Cole JO DeBattista C 200337157Manual of clinical psychopharmacology4th edAntidepressantsAmerican Psychiatric Publishing, Inc Washington (DC) Schatzberg AF, Cole JO, DeBattista C. Antidepressants. In: Manual of clinical psychopharmacology. 4th ed. Washington (DC): American Psychiatric Publishing, Inc; 2003. p. 37-157.
100 Rumpel C Ingram DD Harris TB Madans J 1994 18 179 183 Int J Obes Relat Metab Disord The association between weight change and psychological well-being in women 8186816
101 Barefoot JC Heitman BL Helms MJ Williams RB Surwit RS Siegler IC 1998 22 688 694 Int J Obes Relat Metab Disord Symptoms of depression and changes in body weight from adolescence to mid-life 9705031
102 Wisotsky W Swencionis C 2003 14 37 48 Adolesc Med Cognitive-behavioral approaches in the management of obesity 12529189
103 Marchesini G Natale S Chierici S Manini R Besteghi L Di Domizio S 2002 26 1261 1267 Int J Obes Relat Metab Disord Effects of cognitive-behavioural therapy on health-related quality of life in obese subjects with and without binge eating disorder 12187405
104 Golay A Buclin S Ybarra J Toti F Pichard C Picco N 2004 9 29 34 Eat Weight Disord New interdisciplinary cognitive-behavioural-nutritional approach to obesity treatment: a 5-year follow-up study 15185831
105 Melchionda N Besteghi L Di Domizio S Pasqui F Nuccitelli C Migliorini S 2003 8 188 193 Eat Weight Disord Cognitive behavioural therapy for obesity: one-year follow-up in a clinical setting 14649781
106 Keller MB Boland RJ 1998 44 348 360 Biol Psychiatry Implications of failing to achieve successful long-term maintenance treatment of recurrent unipolar major depression 9755357
|
15670467
|
PMC1323317
|
CC BY
|
2022-01-24 23:45:14
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A14
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0078
Community Case Study
PEER REVIEWEDDiabetes Is a Community Issue: The Critical Elements of a Successful Outreach and Education Model on the U.S.-Mexico Border1
Ingram Maia MPH Mel and Enid Zuckerman Arizona College of Public Health
2501 E Elm, PO Box 24177, Tucson AZ 85721 [email protected]
520-626-7946, ext 242
Gallegos Gwen MS Carondelet Health Network, Holy Cross Hospital, Nogales, Ariz
Elenes JoJean Mariposa Community Health Center, Nogales, Ariz
1 2005
15 12 2004
2 1 A152005
Background
Diabetes is reaching epidemic proportions on the U.S.-Mexico Border, and culturally competent diabetes education is not available in many communities.
Context
People with diabetes often do not have access to regular medical care, cannot afford medication, and lack the community infrastructure that supports self-management practices. Self-management education and support have great potential to impact diabetes control in this environment.
Methods
To address this need, partners of the Border Health Strategic Initiative (Border Health ¡SI!) collaboratively developed a culturally relevant diabetes outreach and education program. The model included a five-week series of free diabetes education classes that assisted participants in gaining the knowledge and skills necessary to be physically active, control diet, monitor blood sugar, take medications, and be aware of complications. Central to the model was the use of community health workers — or promotores de salud — to conduct outreach, participate in patient education, and provide individual support.
Consequences
Program participants achieved significant improvements in self-management behaviors and HbA1c, random blood glucose, and blood pressure levels.
Interpretation
Quantitative and qualitative evaluation helped to identify the essential elements of a successful program, including partnership of providers, community diabetes classes, promotores outreach and support, linkage between diabetes education and clinical care, and program evaluation.
==== Body
Background
The impact of diabetes is devastating along the U.S.-Mexico Border. The rate of diabetes mortality in the border region is nearly 50% higher than in the rest of the country (1), and Hispanics are two to three times more likely to suffer from serious secondary complications (2,3). Self-management behaviors, such as diet, physical activity, and glucose self-monitoring are fundamental to avoiding the long-term complications of diabetes (4). For many individuals, however, self-management behaviors constitute drastic lifestyle changes for which there is little external support. In a managed-care setting, Hispanics were shown to exhibit poor diabetes control when compared with non-Hispanic whites (5).
Diabetes education can have a positive impact on self-management behaviors and glycemic control, particularly when accompanied by intensive follow-up support (6). Diabetes classes delivered in a community setting have been shown to be effective in achieving glycemic control among adults with type 2 diabetes, and this mode of delivery is likely to increase the cultural relevancy and appropriateness of educational techniques in addition to providing greater access to hard-to-reach populations (7). Community partnerships also have the potential to enhance cultural relevance and positively impact self-management and clinical outcomes (8).
There are overwhelming challenges to providing formal diabetes education in border communities. Individuals without insurance do not have access to diabetes education services. For individuals with insurance, few certified diabetes educators (CDEs) live and work in border communities and they may not speak Spanish. Programs that provide interpretation or translation are often not culturally relevant to Hispanics.
This paper describes the patient component of the Border Health Strategic Initiative (Border Health ¡SI!) funded by the Centers for Disease Control and Prevention (CDC), which used the community health worker model to provide culturally competent diabetes education in two Arizona border communities in Yuma and Santa Cruz counties. A detailed description of Border Health ¡SI! is included in this issue of Preventing Chronic Disease (9) along with several companion papers addressing other components of the model (10-18). More information on the rationale and effectiveness of the community health worker model in addressing diabetes can be found in the CDC Division of Diabetes Translation's position statement (available from http://www.cdc.gov/diabetes/projects/comm.htm).
Context
Individual ability to manage diabetes cannot be separated from community context and support for diabetes care (19). Both Yuma and Santa Cruz Counties are rural and more than 90% Hispanic; Yuma County has a large migrant/farmworker community. The region is medically underserved. Lack of insurance, seasonal employment of farmworkers, and fear and discrimination related to immigration present challenges to establishing a regular source of care (20). Patients with diabetes often cross the border to Mexico for medical care, making it difficult to maintain continuity of care.
Residents not eligible for Medicaid programs can rarely afford diabetes medication. Individuals with insurance often do not have pharmaceutical coverage and must decide whether to buy food or medicine. Patients share medication or resort to taking it only when they are feeling badly. While diabetes programs may make glucose monitors available, few resources cover the cost of glucose-monitoring strips.
The border environment does not support good nutrition and physical activity. Few recreational areas, parks, or sidewalks exist in these rural areas to facilitate walking. Summer heat, inadequate lighting, dangerous walking surfaces, and wild dogs pose additional challenges. Although southern Yuma County is a farming community, and the city of Nogales (in Santa Cruz County) is a throughway for produce from Mexico, healthy foods such as fresh fruits and vegetables are high-priced and often unavailable. Furthermore, the health messages taken for granted in urban areas rarely reach farmworkers who work 12-hour days in isolated areas.
The social network that can potentially support self-management is often not in place. The elderly may have family members who migrate to follow the harvesting season or move to urban areas. Many extended family members live in Mexico. Diabetes patients may become isolated and depressed as they experience increasing health problems.
Because of these barriers, education programs must be culturally competent. Vital to the diabetes education program was the use of promotores de salud. Promotores are indigenous to the communities in which they work and provide a bridge between the health care delivery system and the community. In addition to health information, they provide social support and advocate for patients to gain access to health and social services (21). In one diabetes education program, the use of promotores in a Hispanic community was shown to increase the rate of completion (22).
The program
The diabetes outreach and education program was created in Santa Cruz County under a Health Resources and Services Administration Rural Health Outreach Grant (RHOG) in 1997 and adapted by the Yuma community in 2000 under its own RHOG. The programs were supported logistically under the comprehensive framework of Border Health ¡SI! over a three-year period, although Yuma County had additional resources. An investigation of both programs allowed us to define the essential elements of the outreach and education model, which are described below and illustrated conceptually (Figure).
Partnership of providers. Both the Yuma and Santa Cruz programs relied upon a consortium of community providers to implement the patient education component. The community health centers (CHCs) administered the programs and provided a program coordinator. Both programs involved first-time collaboration between the health center and local hospital. The hospital in each county provided a CDE to facilitate classes, train promotoras in diabetes care, and work individually with participants. In Yuma, a grassroots farmworker advocacy organization provided the promotoras, while in Santa Cruz, the promotoras were provided by the CHC. Each program had an academic partner who provided evaluation and technical assistance. The collaborative aspect of the program was crucial in building broader community support for diabetes care.
Figure 1 The roles and responsibilities of partners in the diabetes outreach and education program, Border Health ¡SI!, Yuma and Santa Cruz counties, Arizona. Promotores de salud are community health workers.
The four partners in the Border Health Strategic Initiative are 1) the community health centers, 2) the local hospital, 3) grassroots organization (Yuma County, only), and 4) academic institution. Each is named in a box that also identifies the roles and responsibilities of each partner. The community health centers are responsible for program coordination, patient identification, health care access, and medical examinations. The local hospital is responsible for providing certified diabetes educators, diabetes training for promotores de salud, and individualized services. The Yuma County grassroots organization is responsible for providing the promotores de salud, community outreach, and case management. Arrows connect all three boxes, showing that each interacts with the other. The fourth box represents the academic institution, which is responsible for evaluation and technical assistance. Dotted lines connect the academic institution to each of the three other partners.
Community diabetes classes. Very few participants had prior diabetes education, although many had had diabetes for years. The programs used a culturally competent curriculum that employed a variety of teaching methods to educate participants on how diabetes affects the body and how self-management controls the disease. The curriculum was developed prior to the initiation of the programs by the CDE working in Santa Cruz County using the American Diabetes Association (ADA) Standards of Care. The curriculum followed the content areas set by the ADA and adapted them to the border communities. The curriculum included five two-hour classes held once a week over a five-week period. The sessions included the following topics: 1) understanding diabetes; 2) meal planning; 3) monitoring, medications, and movement; 4) avoiding complications and maintaining health; and 5) foot/eye clinics. In both sites, participants were encouraged to bring family members. The class formats included presentation and discussion and used handouts, videos, and other teaching aids, such as food models. Participants engaged in activities such as creating a balanced plate of food to achieve dietary goals and dancing to achieve physical activity goals. Each class began with a review of the previous session. In addition, program staff measured blood glucose, weight, and blood pressure at each class to demonstrate to participants the progress they were making over the course of the program.
In Santa Cruz, a bicultural CDE based in the local hospital taught the classes. In Yuma, classes were taught by a health educator and eventually by the promotoras under the supervision of a non-Spanish–speaking CDE located in the hospital. Class structure varied between communities. The Santa Cruz community embraced the importance of an open-door program so that classes were available on a rotating basis and class size was maintained at about 20 participants. Participants attended them in any sequence and as often as they wished. In Yuma, the partners recruited a group for each round of classes and encouraged them to complete the program during this time period. Growing interest in the Yuma program resulted in class sizes of up to 40 people.
Promotores outreach and support. The role of the promotoras was to provide outreach, assist participants in incorporating self-management behaviors into their lifestyles, and offer ongoing support and follow-up. There was some disparity in resources between the two programs, and the Yuma community had the advantage of being able to fully implement the promotora model. Four promotoras took responsibility for recruitment, support, and follow-up for a caseload of participants. Potential participants were identified through the health center database. The promotoras personally invited potential participants to the program, provided support to the learning process both during and outside of the classes, and followed up with participants for a six-month period following completion of the classes. The promotoras assisted patients in accessing health insurance, medications, and other social services.
In Santa Cruz, one promotora was available on the day of class to provide telephone follow-up. However, because this program had been initiated several years prior to the initiation of the Border Health ¡SI! patient education component, local providers were aware of the benefits of the program and regularly referred their patients to the classes.
Linkage between diabetes education and clinical care. In both communities, the program was based in a CHC, increasing opportunities for patient-provider communication on patient care. During the program, providers in both programs increased referrals as they recognized the benefits of patient participation. In Santa Cruz, the program added a patient diabetes "empowerment card" to track clinical care and increase patient-provider communication. The trifold card included a form to track the last five physician visits, current medications, participation in diabetes classes, and personal goals.
Many participants did not have access to regular care, and many could not recall a past eye examination. The programs assisted participants in identifying insurance options. A foot exam was included in both programs, and in Yuma, ophthalmologists volunteered their time for eye clinics on Saturdays.
Regardless of insurance status, many participants could not afford medication. While program resources to provide medication were not available, participants were linked to insurance or special programs when possible. Both programs accessed samples from pharmaceutical companies.
Program Evaluation. Program partners engaged in a participatory model of evaluation under the guidance of the academic institution. Under the participatory model, all stakeholders are involved in each phase of evaluation, ensuring a continuous exchange of knowledge, skills, and resources (23). Partners collaboratively developed quantitative and qualitative instruments and shared responsibility for data collection. The academic partner was responsible for analyzing and compiling program data on a cyclical basis to allow for integration of program findings over time. Evaluation efforts were hindered, however, by a lack of resources, which resulted in gaps in data and at times forced promotoras to choose between serving clients (always the first priority) and collecting evaluation information.
Self-management practices were assessed through pre- and follow-up questionnaires administered by the promotoras prior to initiation in the program and six months after graduation. The academic partners trained promotoras in administering the questionnaire, which asked participants if they engaged in self-management practices, including diet, physical activity, foot care, and regular glucose monitoring. The questionnaire also asked participants about their most recent visit with their doctor and whether they had received diabetes health exams in the past year. The initial questionnaire included information on demographics and health history.
Health outcomes included random blood glucose, blood pressure, weight, and HbA1c. Program staff took measurements at three points: initiation of classes, upon graduation from the program, and six months afterwards. HbA1c was measured only twice: before classes and at six-month follow up. In Yuma, the data set is much more complete than in Santa Cruz, and all post-measures were made six to 12 weeks after participants entered the program. In Santa Cruz, the timing of post-measures varied because participants graduated at different points, and attempts to collect HbA1c data at follow-up were unsuccessful because of a lack of staff and financial resources.
Qualitative evaluation took place in Yuma and consisted of in-depth interviews with a random sample of participants in the second and third years of the program. Program partners developed the questionnaire, and academic partners who were not engaged in service delivery conducted the interviews. The interviews explored perceptions of diabetes before and after the program, the role of the family in self-management, changes in self-management practices, and ongoing barriers to diabetes control.
Consequences
The process of implementing the Border Health ¡SI! patient education component over three years in two communities provided a rich opportunity to learn from successes and challenges. In spite of diminishing resources, both programs maintained a strong commitment to providing diabetes education to the underserved. Both communities expressed increased demand for the classes, which was difficult to manage in Yuma because the program moved one group of participants through one series of classes before starting another. At times, classes in Yuma had more than 40 people. Santa Cruz began offering classes in the evening to respond to those who worked during the day.
Santa Cruz had the advantage of a CDE who had worked in the community for years. The Yuma health educator left halfway through the program. The promotores then took responsibility for teaching the classes under the supervision of the hospital CDE. Participant outcomes were maintained when the promotores began teaching.
Evaluation results
Evaluation results generated by the Border Health ¡SI! patient education component are extensive; this paper attempts only to highlight key findings. Table 1 describes the characteristics of individuals who enrolled in the diabetes education classes. In Yuma, 376 individuals enrolled in classes and 306 (81%) graduated. Of graduates, 243 (79%) were reached for the follow-up interview. In Santa Cruz, 406 people enrolled in classes, and 135 (33%) graduated. Of graduates, 40 (30%) were reached for follow-up. Demographic information revealed that the programs did reach the targeted populations. In both counties, participants were more likely to be female and older than 50 years. The majority did not graduate from high school, and approximately two thirds had family members with diabetes. In Yuma, participants were slightly older and experienced more diabetes-related illness; however, they had better access to insurance through Medicare. Few participants had received prior diabetes education, and many had never had an eye exam. Approximately one half reported having high blood pressure and, in Yuma, 59% experienced numbness and burning in their feet.
Health outcomes
Health measures were taken pre- and post-class and at six-month follow-up. Paired t-tests performed on pre- and post-data revealed a significant decrease in the average random blood glucose measurement among participants in both programs (Table 2). In Yuma, levels dropped from 224 mg/dL to 201 mg/dL, and, in Santa Cruz, levels dropped from 197 mg/dL to 151 mg/dL. Both programs also achieved modest but significant decreases in diastolic blood pressure among all participants. Among high-risk participants in Yuma, systolic blood pressure fell from 151 mg/dL to 137 mg/dL, and diastolic blood pressure fell from 100 mg/dL to 84 mg/dL. Among-high risk participants in Santa Cruz, systolic blood pressure fell from 153 mg/dL to 139 mg/dL, and diastolic blood pressure fell from 102 mg/dL to 91 mg/dL. There were no significant changes in health outcomes at the six-month follow-up measure. In Yuma, follow-up results demonstrated a significant 0.7 decrease in HbA1c from 9.4 to 8.7 among those who initiated the program with HbA1c >6.9.
Self-management outcomes
Self-management practices were evaluated in the six-month follow-up interview. Paired t-tests were used to determine significant changes in self-management behaviors. As seen in Table 3, a significant proportion of participants in both counties reported increasing self-management behaviors, including diet, foot care, and glucose monitoring. In Santa Cruz, the percentage of individuals following a diabetes diet increased significantly. In Yuma, where HbA1c and eye exams were provided as part of the Border Health ¡SI! patient education component, the percentage of individuals who had ever received these examinations increased significantly from 53% to 96% (HbA1c) and 57% to 91% (eye exam).
In-depth interviews
Quality of life is as important as clinical outcomes, and in-depth interviews in Yuma demonstrated the impact of the program on program participants. Participant attitude toward diabetes changed from ignorance and fear to acceptance and control, which seemed pivotal in improving their emotional well-being, regardless of self-management practices. Comments included:
"I take care of myself better. I know what is bad for me. I don't feel angry now."
"They tell you how to care for yourself. You can adapt and live a normal life."
The promotoras were also vital to the process because participants felt that the promotoras cared for them and were willing to do whatever they could to help them.
"They are concerned about me. I am motivated because they are worried about me and helped me. "
"My promotora is marvelous. I have a thousand good things to say about her."
Both programs used findings to pursue and secure additional funding to sustain services.
Interpretation
This program responded to a need for accessible, culturally competent diabetes education and demonstrated how communities can galvanize local capacity to respond to an overwhelming lack of resources. Local providers contributed free eye and foot exams and promotoras took over the diabetes education classes when the health educator left the community.
Partnership of providers. Crucial to success was the partnership of diverse organizations that enabled the programs to confront challenges of the border environment on multiple levels. The CHCs had access to the target population, but they would not have been able to recruit and retain participants without the promotores. In both communities, the hospital was critical in providing expertise and in accessing resources.
Community diabetes classes. Holding classes at a community site in a series with a specific group of participants appears to contribute to program completion. This may be because participants have a greater sense of commitment and enjoy belonging to a group. Santa Cruz was extremely fortunate to have a committed, culturally competent and expert CDE. In rural communities where CDEs are not available, promotoras can be trained to provide diabetes education. It is vital, however, that they have backup and support from a qualified person.
Promotores outreach and support. Program outcomes would not have been achieved without promotores. Promotores are fundamental in ensuring that participants initiate and complete classes, gain access to resources, and adopt self-management practices.
Linkage between diabetes education and clinical care. Providing access to health care, examinations, and medications is a challenge that should be addressed early on. For this reason alone, community collaboration is essential. Creating formal relationships with clinical providers may enhance health outcomes. The patient empowerment card was one attempt to establish a formal relationship, and the card was popular with program participants. Strategies to ensure that providers use the card need to be implemented and the impact on care needs to be evaluated.
Program evaluation. Conducting meaningful program evaluation — especially with limited resources — was a challenging but key element of the patient education component. Consistent with the participatory model of evaluation, the academic partner was not an outsider to but rather an integral member of the team and a stakeholder in its success. Within this framework, evaluation became a tool of program development, encouraging partners to define concretely the desired outcomes of the program, to make the effort to collect the necessary information, and to integrate feedback into program strategies. The influence of evaluation on Border Health ¡SI! included 1) designing a series of diabetes education classes (rather than an open-door policy) to create group cohesion and support, 2) establishing a greater focus on including family members in the education and care process, and 3) developing strategies to increase patient-provider communication. Both Border Health ¡SI! communities used evaluation results to sustain program activities beyond the funding period, one through institutional support and the other through other grant funding.
In these two marginalized border communities, the Border Health ¡SI! diabetes education and outreach program had a positive influence on the ability of individuals to adopt self-management practices and improve health outcomes. It is important to note that as a component of the comprehensive Border Health ¡SI!, the education and outreach program was linked to a policy action group that addressed challenging environmental issues related to diabetes (15,16). Participation in a policy-focused group enabled program partners and community leaders to discuss systemic problems, leverage additional resources, and address prevention on a community level.
The authors thank Sunset Community Health Center, Mariposa Community Health Center, and Carondelet Health Network.
Figures and Tables
Table 1 Characteristics of Participants in Diabetes Patient Education Program, Border Health Strategic Initiative, Arizona, 1999–2002
Santa Cruz County
N = 406
(%) Yuma County
N = 376
(%)
Female 284 (70) 250 (66)
Aged >50 years 203 (50) 262 (70)
Graduated from high school 170 (42) 72 (19)
Insured 268 (66) 281 (75)
Diabetes in family 276 (68) 275 (73)
HbA1c >6.9 Data not available 212a (58)
Prior diabetes education 28 (7) 64 (17)
High blood pressure 191 (47) 196 (52)
Numbness/burning in feet 138 (34) 218 (59)
Hospitalized in the last year for diabetes 46 (11) 67 (18)
Graduated from program 135 (33) 306 (81)
a N = 290 because of missing data.
Table 2 Changes in Health Measurements Among Participants Who Completed Diabetes Patient Education Program, Border Health Strategic Initiative, Arizona, 1999–2002a
Santa Cruz County
N = 135 Yuma County
N = 306
Among all participants
Pre-program Post-program Pre-program Post-program
HbA1c level (N = 198) No data available 8.7 8.2**
Random blood glucose (mg/dL) 196.9 151.1*** 224.5 200.6***
Systolic blood pressure (mm Hg) 130.3 128.2 131.5 127.9***
Diastolic blood pressure (mm Hg) 80.9 78.0* 77.9 76.5*
Weight (lbs) 184.6 182.7* 174.3 173.0*
Among high-risk participants
HbA1c level (N = 132) No data available 9.4 8.7***
Random blood glucose (mg/dL) 225.9 159.6*** 246.6 212.0***
Systolic blood pressure (mm Hg) 152.6 138.8*** 150.8 137.3***
Diastolic blood pressure (mm Hg) 101.6 91.3*** 99.6 84.5***
a Post-program measurements were taken upon completion of the program, with the exception of HbA1c, which was taken at six-month follow-up. High-risk is defined as HbA1c level >6.9. P = *<.05; **<.01; ***<.001.
Table 3 Changes in Self-Reported Diabetes Self-Management Outcomes, Border Health Strategic Initiative, Arizona, 1999–2002a
Santa Cruz County
N = 40 Yuma County
N = 243
Pre-program Post-program Pre-program Post-program
Exercises regularly 50 70* 67 83***
Follows diet 45*** 80*** Data not complete
Checks feet regularly 60 88** 86 98***
Monitors blood sugar 38 63 51 96***
Ever had HbA1c 33 45 53 96***
Knows what HbA1c is 40* 40* 22 64***
Ever had eye exam 33 47 57 91*
a All values are percentages. Post-program measurements were taken six months after program graduation. P = *<.05; **<.01; ***<.001.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Ingram M, Gallegos G, Elenes J. Diabetes is a community issue: the critical elements of a successful outreach and education model on the U.S.-Mexico border. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0078.htm
==== Refs
1 American Diabetes Association 1996 Diabetes vital statistics American Diabetes Associatio Alexandra (VA)
2 Haffner SM Fong D Stern MP Pugh JA Hazuda HP Patterson JK 1988 37 878 884 Diabetes Diabetic retinopathy in Mexican Americans and non-Hispanic Whites 3384186
3 Hanis CL Ferrell RE Barton SA Aguilar L Garza-Ibarra A Tulloch BR 1983 118 659 672 Am J Epidemiol Diabetes among Mexican Americans in Starr County, Texas 6637993
4 Diabetes Control and Complications Trial Research Group 1993 977 986 N Engl J Med 329 The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitu 8366922
5 Brown AF Gerzoff RB Karter AJ Gregg E Safford M Waitzfelder B 93 10 2003 1694 1698 Am J Public Health Health behaviors and quality of care among Latinos with diabetes in managed care 14534224
6 Norris SL Engelgau MM Narayan KM 24 3 3 2001 561 587 Diabetes Care Effectiveness of self-management training in type 2 diabetes: a systematic review of randomized controlled trials 11289485
7 Norris SL Nichols PJ Caspersen CJ Glasgow RE Engelgau MM Jack L 22 4 Suppl 5 2002 39 66 Am J Prev Med Increasing diabetes self-management education in community settings: a systematic review 11985934
8 Gerber JC Stewart DL 3 1998 48 52 J Assoc Acad Minor Phys Prevention and control of hypertension and diabetes in an underserved population through community outreach and disease management: a plan of action 9747058
9 Cohen SJ Ingram M Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Border heath strategic initiative: overview and introduction to a community-based model for diabetes prevention and control
10 Abarca J Ramachandran S Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Using community indicators to assess nutrition in Arizona-Mexico border communities
11 Schacter KA Cohen SJ Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] From research to practice: challenges in implementing national diabetes guidelines with five community health centers on the border
12 Teufel-Shone NI Drummond R Rawiel U Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Developing and adapting a family-based diabetes program at the U.S.-Mexico border
13 Staten LK Scheu LL Bronson D Peña V Elenes J Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Pasos adelante: the effectiveness of a community-based chronic disease prevention program
14 Staten LK Teufel-Shone NI Steinfelt VE Sanchez N Halverson K Flores C Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] The School Health Index as an impetus for policy change
15 Meister JS de Zapien JG Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Bringing health policy issues front and center in the community: expanding the role of community health coalitions
16 Steinfelt VE Prev Chronic Dis [serial online] 2005 Jan [2005 Dec 15] The Border Health Strategic Initiative from a community perspective
17 Rodríguez-Saldaña J Prev Chronic Dis [serial online] 2005 Jan [2005 Dec 15] Challenges and opportunities in border health
18 Martorell R Prev Chronic Dis [serial online] 2005 Jan [2005 Dec 15] Diabetes and Mexicans: why the two are linked
19 U.S. Department of Health and Human Services Healthy People 2010: understanding and improving health 2nd ed Washington (DC) U.S. Government Printing Office 2000 11
20 Ruiz-Beltran M Kamau JK 26 2 2001 123 132 J Cmty Hlth The socio-economic and cultural impediments to well-being along the US-Mexico border
21 Love MB Gardner K Legion V 1997 24 510 522 Health Educ Behav Community health workers: who they are and what they do 9247828
22 Corkery E Palmer C Foley ME Schechter CB Frisher L Roman SH 20 3 1997 254 257 Diabetes Care Effects of a bicultural community health worker on completion of diabetes education in a Hispanic population 9051367
23 Springett J Minkler M Wallerstein N 11 2002 268 288 Community-based participatory research for health Issues in participatory evaluation John Wiley & Sons, Inc Hoboken (NJ)
|
15670468
|
PMC1323318
|
CC BY
|
2022-01-27 23:16:24
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A15
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0080
Community Case Study
PEER REVIEWEDBringing Health Policy Issues Front and Center in the Community: Expanding the Role of Community Health Coalitions1
Meister Joel S PhD Associate Professor of Public Health and Director of the Public Health Policy and Management Concentration Mel and Enid Zuckerman Arizona College of Public Health
1501 N Campbell Ave, PO Box 210228, Tucson, AZ 85724-5163 [email protected]
520-318-7270, ext 19
Guernsey de Zapien Jill Associate Dean for Community Programs Mel and Enid Zuckerman Arizona College of Public Health, Tucson, Ariz
1 2005
15 12 2004
2 1 A162005
Background
Systemic, environmental, and socioeconomic conditions create the context in which community members deal with their health concerns. Comprehensive, community-based chronic disease prevention interventions should address community-wide or regional policy issues that influence lifestyle behaviors associated with chronic diseases.
Context
In two communities along the Arizona-Mexico border, community coalitions that administered a comprehensive diabetes prevention and control intervention expanded their membership to become policy and advocacy coalitions with broad community representation. These coalitions, or Special Action Groups (SAGs), identified and prioritized policy issues that directly or indirectly affect physical activity or nutrition.
Methods
Local schools were one focus of advocacy. The Centers for Disease Control and Prevention's School Health Index was implemented as part of the overall intervention; the SAGs supported schools in advocating for more physical education programs, removal of vending machines, substitution of more healthful options in vending machines, and changes in health education curricula. In the broader community, the SAGs promoted opportunities for walking and bicycling, long-term planning by their cities and counties, and healthy food choices in local grocery stores.
Advocacy tactics included attending and making presentations at city council, school board, parks and recreation, and planning and zoning commission meetings; participating on long-range planning committees; organizing an annual community forum for elected and appointed officials; and presenting healthy food and cooking demonstrations in local markets.
Consequences
After three years, SAGs were able to document changes in local policies and practices attributable to their activities.
Interpretation
The SAGs contributed to systems changes in their communities and were able to obtain new resources that support protective behaviors. Also, the advocacy process itself provided strong positive reinforcement to all participants in this comprehensive diabetes intervention.
==== Body
Background
Approaches to preventing and controlling chronic diseases, such as diabetes, must focus on broad lifestyle issues. Such an approach to preventing and controlling diabetes may include patients, their families, providers, and the entire community (1-3).
More recently, and with increasing recognition of the extent to which individual health-related behavior is shaped by social and cultural norms and by the physical and policy environment of a community (4), attention is being given to the systems and environmental- and community-level factors that contribute to the behaviors that affect health status and outcomes (4-7). The Centers for Disease Control and Prevention's (CDC's) Racial and Ethnic Approaches to Community Health (REACH) 2010 program illustrates the increasing emphasis on changing systems factors using a logic model that includes changes in change agents and environmental and policy shifts as precursors of more distal changes in health-related behaviors and health status (Figure) (8).
Figure 1 The Racial and Ethnic Approaches to Community Health (REACH) 2010 model of change, adapted by the Southwest Center for Community Health Promotion (8)
This flowchart shows the Racial and Ethnic Approaches to Community Health (REACH) 2001 model adapted by the Southwest Center for Community Health Promotion. Arrows show a hierarchy of six factors, starting at the bottom with "Community awareness of issue." Second from the bottom is "Develop needed community capacity." Third is "Changes in change agents." Fourth is "Environmental Shift (i.e., policy)." Fifth is "Changes in risk factors and protective factors." Sixth, and at the top of the list of factors is "Changes in health." Targeted actions take place at the lower four levels, beginning with "Community awareness of issues" and ending with "Environmental shift (i.e., policy). Arrows point from "Targeted Actions" to these four factors.
The Border Health Strategic Initiative (Border Health ¡SI!) was a comprehensive diabetes prevention and control program that focused on border communities along the Arizona-Mexico border (9). The authors adapted the REACH 2010 model so that Border Health ¡SI! included a significant policy component. The community coalitions, originally formed to bring together community partners and the University of Arizona, were challenged to become Special Action Groups (SAGs) with their own unique role — to effect policy changes that would promote health in the community.
Context
The U.S.-Mexico border has several singular features relevant to diabetes prevention and control. It is a poor region with fragmented services, and residents often cross the border — in both directions — for health care (10). The border region has a large Hispanic population, with diabetes prevalence approximately twice the average for non-Hispanic whites (11,12). Many residents are undocumented and therefore have no access to health care except for private fee-for-service, which they can rarely afford, or for emergency services (13).
Along the Arizona-Sonora border, the University of Arizona and numerous community partners have been working together for the last twenty years to create health promotion programs and the joint capacity and infrastructure to address a wide spectrum of health issues. Based on this ongoing, evolving, and positive history of collaboration among community-based agencies and the University of Arizona and our common recognition of the need for systemic change, the partners in both communities responded positively to the recommendation that the programmatic partners of Border Health ¡SI! (those responsible for specific intervention components) continue to meet as a technical team while the coalition expand to include other community members and agencies with a stake in policy. These coalitions — the SAGs — would be dedicated to planning and advocating for policy change.
One of the university partners, the Cooperative Extension Service, was asked and agreed to be the facilitator for the SAGs. The Mel and Enid Zuckerman Arizona College of Public Health's collaboration with Cooperative Extension pre-dates this project, and the relationship was expanded and strengthened by the decision to have Cooperative Extension serve as SAG facilitator. As SAG facilitator, Cooperative Extension used its everyday, longstanding connection between community and university to strengthen the SAGs.
SAG membership included organizational leaders, program directors, community health workers (promotores de salud), and other concerned citizens. Promotores de salud were critical to forming SAGs (14-16). They provided the outreach and leadership in every component of the intervention except the provider component. They brought to the SAGs their knowledge of what was actually happening in the community day to day. They also provided the potential leadership for any community mobilization that might become part of the SAGs' action plans.
Methods
Fitting policy into the picture
SAG members met first to become familiar with the REACH 2010 model of change. The model's most novel features were emphasis on the changes in "change" agents and changes in local policies that were posited to contribute to changes in behaviors such as physical activity and nutrition. The "targeted activities" that drive the model would thus have to consist not only of the health education programs with which all partners were familiar and comfortable but with new capacity-building activities and advocacy interventions that at first seemed somewhat threatening or exotic. SAG minutes and participant observation data show that most SAG members, including organizational leaders, had never appeared before a city council or other elected body (17).
Distinguishing between program and policy
Our community partners were highly skilled at delivering health promotion and education, but they had much less experience dealing with broader policy issues that were not part of traditional health promotion culture. These issues included, for example, the physical environment of the community and whether it supported walking or bicycle-riding or other forms of exercise, the availability of low-fat, low-sugar foods in grocery stores, the food products available in school vending machines, and the use of candy for school fundraising.
Identifying and prioritizing policy issues
As each SAG began to identify and prioritize policy issues in its community, sustaining the distinction between programs and policies was the most challenging aspect of developing a policy agenda. For instance, in initial discussions about changing food choices, some SAG members suggested a health fair. Others, more cognizant of the policy issues, wanted to go straight to market owners or managers and attempt to influence their decisions on which food products to stock and promote and how food products were displayed.
As the policy focus became clearer, the SAGs prioritized and selected issues to be addressed over the following one to three years. Community A divided its policy goals into short- and long-term goals. The short-term goal was defined as increasing opportunities and places for physical activity, and the long-term goals were defined as making an impact on the county's long-range parks and recreation planning and resource allocation. Community B selected the following policy goals: 1) develop more parks and recreation areas, 2) work with grocery stores to offer and promote more healthful foods, and 3) work with schools to emphasize health curricula and to change the use of candy and other junk food in the fundraising and reward structure.
Redefining health as a community-wide issue
Health came to be seen among SAG members as an array of policy issues that extend well beyond the purview of the experts in the county health department, the community health center, or school nurses. SAG members realized that they needed to reach a number of change agents that included elected officials, business people, members of the faith community, and educational leaders. They also needed to bring this broader vision to other health professionals.
Bringing new members to the coalition
Identifying and then recruiting new SAG members was a critical step in promoting a policy agenda. Convincing some of them that health should be one of their issues was a major achievement in recruiting and retaining them as SAG members (18). These new recruits included the following (some in Community A, others in Community B): a chamber of commerce executive director, county interfaith council director, city manager, parks and recreation department director, public works department director, planning and zoning director, hospital administrator, school superintendent, town librarian, newspaper editor, and police officer.
Developing an action plan
Once issues were identified and prioritized, the SAGs formed subgroups to develop action plans for each major issue. Community A decided to make the SAG indispensable to the county's long-range development planning effort by volunteering to serve on the planning committee, offering the SAG's own recommendations for open space, parks and recreation, and walking/bicycle paths development, and offering data gathered by its university partner.
Community A also adopted a short-term action plan that designated a three-month period for mounting a series of health promotion activities that would culminate in a presentation to the city council, stressing the need for reallocating (not increasing, at this time) parks and recreation resources to promote physical activity among the entire community, and attending to neighborhood safety, including lighting, sidewalks, and animal control. The SAG in Community A contracted with a consultant to design a compelling fact sheet that would be used in its presentation to the city council and other policy-making bodies.
In Community B, the SAG initiated an annual community forum designed to educate policy makers, advocate for policy change, and hold elected officials accountable for their support, or lack thereof, of policies to promote health. The forum was designed so that representatives of the SAG and other community groups could first present their activities and policy agendas to public officials who were invited to attend. After the community presentations, elected and appointed officials were invited to respond, and then the forum was opened to discussion.
The promotores in this community's SAG mobilized their constituents to advocate for new parks in one of the small towns near the border and in an unincorporated area of the county that provided few public services to its residents. These promotores had been leading the community walking groups and nutrition classes that were one component of Border Health ¡SI!. Now they and members of these groups went before the county board of supervisors to advocate for parks in their neighborhoods.
The SAG in Community B also worked with the schools component of Border Health ¡SI! to promote changes in the curriculum and the use of junk foods. While the schools component of Border Health ¡SI! worked with the School Health Index and the school health teams, the SAG also kept in close contact with the school superintendent and individual principals to promote change and monitor progress.
Consequences
Results that can be traced directly to the actions of the SAGs are described below.
Community A
New walking paths were incorporated into the county's development plan.
A new Wal-Mart Supercenter added a perimeter walking path to its construction plan.
Plans to terminate physical education at a local school were halted.
Health-related articles now appear regularly in the local newspaper.
Community B
Two Community Development Block Grants were obtained for parks and walking paths. The SAG also succeeded in convincing the local school district to donate land for one of the parks. This donation made it possible to use the grant to fund landscaping and to purchase exercise equipment and other amenities.
Grocery stores in the target communities initiated healthy food demonstrations one to two times per month. These demonstrations were organized and conducted by promotores.
Stores began stocking more healthy products.
Sales of food featured in the healthy food demonstrations increased.
The SAG received the 2002 Mayor's Physical Activity Leadership Award.
Of the many lessons learned from the SAGs, the following are among the most salient:
A comprehensive approach to community health promotion requires a policy component.
Commitment and organizational involvement of the key community-based health organizations are necessary.
Promotores must be involved as change agents.
Social action focused on policy change can energize a coalition, giving it a raison d’être beyond merely coordinating activities, and can contribute to its sustainability.
The SAG created an engine for change on community health issues.
Short-term successes contribute to long-term effectiveness of SAGs.
Consciousness-raising about public health issues among those who are not public health practitioners is important to effecting policy change. Convincing people that health is their business regardless of what they do professionally is critical to recruiting opinion leaders to join a SAG and to activating local or regional policy makers.
Sustainability is made possible by a SAG in several ways. SAG action motivates members to continue their advocacy efforts as new issues arise and successes are achieved. SAG advocacy creates links between programs and policies that may result in local or regional agencies incorporating successful programs and new policies into their standard mode of operation. SAGs create strategic alliances with non-health specific groups that may lead to new funding opportunities that help sustain multiple components of a community health intervention. SAGs provide an opportunity for promotores to serve as community change agents.
The experience of the SAGs and the results of their advocacy have been reported to the community in a variety of ways. Foremost has been the publication of numerous articles in local newspapers — made possible, no doubt, by SAG membership of newspaper editors or reporters in each community. Presentations at conferences, including the U.S.-Mexico Border Health Association, Arizona Public Health Association, CDC Diabetes Translation Conference, and others provided a mechanism for dissemination of lessons learned to other border communities throughout the region. SAG activities are also reported regularly to the Community Action Board (CAB) of the Southwest Center for Community Health Promotion. The CAB is, in effect, a super-SAG for all communities involved in Border Health ¡SI! and other border community health interventions of the Mel and Enid Zuckerman Arizona College of Public Health.
Interpretation
After the fact, it is difficult to imagine the Border Health ¡SI! program without its SAG policy-change component. This is so not only because the SAGs contributed to systems changes in their communities and were able to obtain new resources that support protective behaviors but also because the advocacy process itself provided such strong positive reinforcement to all participants in this comprehensive diabetes intervention. The results of evaluation interviews with SAG members and the administration of the Wilder Collaboration Factors Inventory (19) strongly suggest that participation in the SAG resulted in:
Improved health behaviors within members' own organizations.
Better understanding of community needs.
Closer relationships with other agencies represented on the SAGs.
SAG members also took credit for:
Building awareness among policy makers.
Influencing community-wide resource allocation.
Gaining support for SAG initiatives by city, county, and school-governing bodies.
Working collaboratively with decision makers in the planning process.
Context always plays an important role in defining the issues to be addressed and the boundaries of possible action and change in a given community. In this case, context included the border geography and demography, especially the preponderance of Hispanics in these communities, the persistent poverty and lack of formal education among much of the population, and the pervasiveness of diabetes. One might suppose that such a context would militate against effective organization for policy change. We did not find this to be true. On the contrary, the brief history of the SAGs confirms our prior experience — that in these have-not communities along the U.S.-Mexico border, there is a largely untapped reservoir of intelligence and thirst for knowledge, concern about community conditions, desire for change and willingness to take risks, and, most important, a willingness to act collectively for the common good.
From the perspective of university-based participatory-action researchers, creating a collaborative policy-change initiative, whether stand-alone or as part of a broad health intervention, requires a strong, positive university-community partnership (20,21). Those partnerships take time to build and require mutual trust (22-25). To that we would add that the researcher's goal is to be a partner in the fullest sense, not merely to provide technical assistance, advise, and evaluate but to be an integral part of planning, decision making, and action — without inadvertently assuming the leadership of what is, after all, a community coalition. It is the action taken by all of the partners that results in the kind of impact that lives on in the community.
Funding for this project comes from the Division of Diabetes Translation, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Meister JS, Guernsey de Zapien J. Bringing health policy issues front and center in the community: expanding the role of community health coalitions. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0080.htm
==== Refs
1 Teufel-Shone NI Drummond R 2002 The University of Arizona Tucson (AZ) La diabetes y la unión: familiar curriculum for the Border Health Strategic Initiative, a comprehensive community based diabetes prevention and control program
2 Veazie MA Teufel-Shone NI Silverman GS Connolly AM Warne S King BF 7 2 2001 21 32 J Public Health Manag Pract Building community capacity in public health: the role of action-oriented partnerships 12174397
3 Cohen SJ Meister JS de Zapien JG 2004 119 40 47 Public Health Rep Special action groups for policy change and infrastructure support to foster healthier communities on the Arizona-Mexico border 15147648
4 Wilcox A Knapp A 2000 115 139 143 Public Health Rep Building communities that create health 10968745
5 Butterfoss FD Goodman RM Wandersman A 8 3 1993 315 330 Health Educ Res Community coalitions for prevention and health promotion 10146473
6 Merzel C D'Afflitti J 93 4 2003 557 574 Am J Public Health Reconsidering community-based health promotion: promise, performance, and potential 12660197
7 Roussos ST Fawcett SB 2000 21 369 402 Ann Review Public Health A review of collaborative partnerships as a strategy for improving community health
8 Gerberding JL 2004 Racial and ethnic approaches to community health (REACH 2010): addressing disparities in health – at a glance Atlanta (GA) Centers for Disease Control and Prevention
9 Cohen S Ingram M Prev Chronic Dis [serial online] 2005 1 Border Health Strategic Initiative: overview and introduction to a community-based model for diabetes prevention and control
10 U.S. Department of Health and Human Services 1999 Assuring a healthy future along the U.S.-Mexico border Washington (DC) U.S. Department of Health and Human Services
11 Flood T Lebowitz MD De Zapien J Staten L Rosales C 1999 Douglas community health survey: diabetes and health care in Arizona on the Mexican border ADHS Phoenix (AZ)
12 Centers for Disease Control and Prevention 2004 Diabetes: disabling, deadly, and on the rise U.S. Department of Health and Human Services Atlanta (GA)
13 De Zapien JG 2004 Feb 18-20 18th National Conference on Chronic Disease Prevention and Control Washington (DC) Making a difference for prevention: policy change in communities at the US-Mexico Border
14 Meister JS 2000 The community health worker evaluation tool kit Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona Tucson (AZ) Available from: URL: http://www.publichealth.arizona.edu/chwtoolkit
15 Meister JS 1997 Community outreach and community mobilization: options for health care at the U.S.-Mexico border U.S. Department of Health and Human Services Washington (DC)
16 Rosenthal EL 1998 The national community health advisor study Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona and The Annie E. Casey Foundation Tucson (AZ) Available from: URL: http://www.aecf.org
17 Reinschmidt K Dunne A 2003 Personal communication and meeting minutes
18 Castro FG Elder J Coe K Tafoya-Barraza HM Moratto S Campbell N J Natl Cancer Inst Monogr 1995 18 127 135 Mobilizing churches for health promotion in Latino communities 8562213
19 Mattessich P Murray-Close M Monsey B 2001 The Wilder Collaboration Factors Inventory: assessing your collaboration's strengths and weaknesses Saint Paul (MN) Amherst H. Wilder Foundation
20 Israel BA Schulz AJ Parker EA Becker AB Allen A Guzman JR Minkler M Wallerstein N Critical issues in developing and following community-based participatory research principles 2003 56 73 Community-based participatory research for health San Francisco (CA) Jossey-Bass
21 National Association of County and City Health Officials Mobilization for action through partnerships and planning (MAPP) National Association of County and City Health Officials Washington (DC) Available from: URL: http://www.naccho.org/project77.cfm
22 Israel BA Schulz AJ Parker EA Becker AB 1998 19 173 202 Annu Rev Public Health Review of community-based research: assessing partnership approaches to improve public health 9611617
23 Weiss ES Anderson RM Lasker RD 29 6 2002 683 698 Health Educ Behav Making the most of collaboration: exploring the relationship between partnership synergy and partnership functioning 12456129
24 Alexander JA 12 2 2001 159 175 Nonprofit management and leadership Leadership in collaborative community health partnerships
25 Harris E Wills J 21 4 Spec No 1997 403 412 Aust N Z J Public Health Developing healthy local communities at local government level: lessons from the past decade 9308206
|
15670469
|
PMC1323319
|
CC BY
|
2022-01-24 23:45:16
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A16
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0079
Community Case Study
PEER REVIEWEDFrom Research to Practice: Challenges to Implementing National Diabetes Guidelines With Five Community Health Centers on the U.S.-Mexico Border1
Schachter Kenneth A MD, MBA Specialist, Clinical Research and Clinical Quality Improvement Mel and Enid Zuckerman Arizona College of Public Health
2231 E Speedway Blvd, Tucson, AZ 85719 [email protected]
520-906-4388
Cohen Stuart J EdD Mel and Enid Zuckerman Arizona College of Public Health, Tucson, Ariz
1 2005
15 12 2004
2 1 A172005
Background
Given the dramatic increase in type 2 diabetes in the United States, the development of effective strategies to prevent and control this potentially devastating illness is more important than ever. In the Southwest, diabetes is a far too common and rapidly growing problem among Mexican Americans living near the U.S.-Mexico border. A project designed to address this problem enabled faculty from the University of Arizona to work with community health centers to evaluate and improve diabetes care in border communities.
Context
This project was a component of the Border Health Strategic Initiative (Border Health ¡SI!) and Racial and Ethnic Approaches to Community Health 2010 (REACH 2010), both funded by the Centers for Disease Control and Prevention. University of Arizona faculty worked in partnership with five community health centers funded by the Health Resources and Services Administration. The goal of the faculty was to assist the community health centers with 1) development of measures of diabetes care based on national clinical practice guidelines, 2) identification of gaps in care based on those measures, and 3) implementation of strategies for closing those gaps.
Methods
All five centers prioritized their top four or five indicators of diabetes care (e.g., annual dilated eye examination). Different community health centers selected different indicators. Baseline medical record audits were performed using the chosen indicators. Individual results were shared confidentially with providers; overall center results were shared and discussed with providers and staff.
Consequences
Each clinic chose its own strategies for closing gaps in care. At one-year follow-up, there was evidence of improvement for the majority of indicators in all community health centers. However, some gaps remained. Of the three community health centers having a second-year evaluation, two maintained or increased the improvements made, but one lost ground.
Interpretation
Our experience with these five border clinics was that translating guidelines into practice is easier said than done. Factors that favored success included an onsite champion, staff buy-in, a willingness to see systems change, and the availability of additional resources, particularly for chart reviews.
==== Body
Background
Between 1990 and 2000, the Mexican population in the United States increased by 52.9%, from 13.5 million to 20.6 million. By 2050, it is estimated that there will be 97 million Hispanic Americans in the United States, comprising about one quarter of the total population (1). In 2000, more than 43% of Hispanics lived in the West. Half of all Hispanics lived in just two states, California and Texas. The largest Mexican populations were in California, Texas, Illinois, and Arizona. In the three border states (California, Texas, and Arizona), Hispanics were in the majority in 50 counties along the U.S.-Mexico border (2).
Diabetes was the sixth leading cause of death in the United States in the year 2000. More than 17 million Americans (about 2 million Hispanic Americans) have been diagnosed with diabetes, and approximately 1 million more individuals, aged 20 years and older, are diagnosed with diabetes each year (3). At any given age, Mexican Americans are twice as likely to have type 2 diabetes as non-Hispanic whites (4).
In 2002, researchers at the Centers for Disease Control and Prevention (CDC) developed a Diabetes Report Card to examine quality of diabetes care in the United States during the 1990s based on nationally accepted guidelines for care. Their research revealed that 18% of persons with diabetes aged 18 to 75 years had very poor glycemic control (HbA1c values >9.5%) and that 34% had elevated blood pressures (⩾140/90 mm Hg). Left untreated or inadequately treated, both conditions will lead to increased morbidity and mortality (3). Additionally, 45% of patients with diabetes had not had a foot examination during the previous year and 37% had not had a dilated eye exam (3).
In a September 2004 report on the quality of care among its member health plans, The National Committee for Quality Assurance (NCQA) asserts that many Americans do not receive adequate preventive care and/or care for chronic conditions like diabetes and hypertension. It further asserts that the gap between the less-than-optimal health care that most Americans receive and the care that some receive from the best health plans results in anywhere between 42,000 and 79,000 premature deaths per year (5).
There is increasing evidence of programs and treatment strategies that are effective in controlling diabetes and preventing its complications; however, the translation of that evidence into medical practice continues to lag. The Institute of Medicine estimates that the time between the discovery of an effective treatment and its incorporation into routine care is as long as 17 years and that more than 50% of patients with such common conditions as diabetes, hypertension, tobacco addiction, hyperlipidemia, congestive heart failure, asthma, depression, and chronic atrial fibrillation are inadequately managed (6).
Context
Community Health Centers (CHCs) first received federal funding as part of the War on Poverty in the mid-1960s. Approximately 100 CHCs (known at the time as neighborhood health centers) had been funded under the Economic Opportunity Act (EOA) by the early 1970s. These centers made culturally appropriate health care accessible to many low-income families. In 1969, the Public Health Service (PHS) also began funding neighborhood health centers. When the EOA was phased out in the early 1970s, centers previously supported under it were transferred to the Public Health Service (PHS). Today, CHC funding is authorized under section 330 of the PHS Act through the Health Resources & Services Administration (HRSA). CHCs exist in areas where economic, geographic, or cultural barriers limit access to primary health care. Their mission is to provide family-oriented, primary, and preventive health care services for people living in rural and urban medically underserved communities (7).
This paper describes our experiences as university faculty with the primary care providers (medical doctors, doctors of osteopathy, nurse practitioners, and physician assistants) and support staff in five community health centers participating in two federally funded projects, The Border Health Strategic Initiative (Border Health ¡SI!) and Racial and Ethnic Approaches to Community Health 2010 (REACH 2010). All five centers are located in the United States near the U.S.-Mexico border and care for large Hispanic populations with a high prevalence of diabetes. The border region includes four states in the United States (Arizona, California, New Mexico, and Texas) and six states in Mexico (Baja California, Chihuahua, Coahuila, Nuevo Leon, Sonora, and Tamaulipas).
In 2000, approximately 11.3 million people lived on both sides of the border — 6,268,107 individuals on the U.S. side, and 5,054,516 on the Mexico side (8). The U.S. side is approximately 70% Hispanic, with a higher population growth rate (1.8%) than the national rate (0.9%). Five of the seven poorest counties in the United States are on the border, and more than 30% of Hispanics living on the border are uninsured (9).
Methods
In all five centers, we met with the medical directors and other clinical management staff. First, we explained the project and our approach. Next, we worked with medical directors to develop target guidelines, or indicators. We used an "indicators of care" form to help them identify five aspects of diabetes care they most valued. We limited the medical directors to five aspects out of concern that they might try to accomplish too much, too soon. One center tracked only four indicators after a mid-course change. We encouraged the medical directors to involve their medical and other health care staff in this selection process, and we had conversations with the medical directors and/or their staff about their choices, especially when there was not good evidence to support a selected intervention or intervention frequency. However, the centers' choices always prevailed, even if they selected a measure for which there was not good evidence. Some medical directors were more interested than others, some had more participatory management styles than others, and some delegated more than others. We adapted our procedures to the local characteristics of each center. Each CHC was given the option of having record reviews performed and reported only at the center level or at both the center and provider levels, with the provider-level results being shared confidentially with individual providers. All CHCs opted for both provider and center-level results.
Primary care providers for each clinic were eligible to have their patients' charts included in the record review if the provider had practiced at the CHC for the 12 months prior to the review. Not all providers at each CHC participated in this project. In some CHCs with multiple offices, participation was limited to one site. In addition, while many of the same providers participated from the beginning to the end of the project, staff turnover led to changes in those individuals being reviewed from one year to the next. As planned, we completed three rounds of data collection (baseline, one-year follow-up, and two-year follow-up) in three clinics and two rounds (baseline, one-year follow-up) in the other two. In 2001, we audited 22 providers' records from five participating CHCs; in 2002, 19 providers' records from five CHCs; and in 2003, nine providers' records from three CHCs.
After indicators were selected, we developed indicator-specific training manuals for medical records reviewers. The manual was designed to allow us to train staff with little or no medical records review experience and to serve as a reference for questions reviewers might have during the review process. Reviewers were instructed to begin auditing charts soon after their training was completed. Charts were randomly selected from up-to-date listings of patients with diabetes assigned to each primary provider. For a chart to be eligible, the patient had to be at least 18 years of age, had to have a diagnosis of diabetes based on Current Procedural Terminology 250.XX codes, and had to have visited his/her primary provider at least once during the 12 months under review. If there were multiple visits during the 12 months, the primary provider had to have seen the patient for a majority of those visits.
Obtaining provider-level data required a larger sample than would have been necessary for clinic-level data. Evidence from prior studies indicated that 12 to 15 charts per provider are needed to obtain a stable estimate of provider performance while imposing the lowest possible burden on center staff. To ensure and improve the reliability of our reviewers, we used two reviewers in every center for each review cycle, asking that they assign a primary and secondary reviewer for each provider's records. The primary reviewer reviewed all of that provider's records. The secondary reviewer randomly selected and reviewed two of that provider's records while avoiding discussing them with the primary reviewer and/or viewing his/her audit results. The secondary reviewer was then instructed to compare both reviews and mark all inconsistencies on the secondary review form. Both reviewers were then asked to review the disagreements and, where indicated, correct any mistakes on the primary review form. The primary review forms were used to calculate the level of compliance with selected indicators. The marked secondary review forms, which showed primary and secondary reviewer errors, were used to calculate interrater agreement. Our goal was to achieve interrater agreement of greater than or equal to 90%. We missed that mark only twice in a total of thirteen reviews. The chart selection process was repeated for each round of reviews. These cross-sectional samples included only those randomly selected patients who met eligibility criteria for that year.
Following the initial reviews, we met with the medical and key office staff to present a table showing baseline center-level results. All participants received their CHC's results. Each individual provider also received a table in a sealed envelope comparing his/her results to center results. We promised a repeat audit in about 12 months.
We returned to reaudit charts, as promised, in approximately 12 months for two centers and in both 12 and 24 months for three centers. After each audit, we presented our results to the medical staff and discussed strategies for further improvement. In the satellite center that showed the least improvement at year one, we changed some indicators at the request of the primary provider, who had different priorities for diabetes care and had not been able to participate in the initial selection process.
Consequences
For most indicators, overall center performance was higher at the one-year assessment than it was at baseline. In addition, two of the three CHCs having year-two assessments showed generally improved results from year one. The third center's year-two results showed worsening performance in most areas compared with year one. Given the cross-sectional nature of the samples, these results should be interpreted with caution.
Only two of the five centers prioritized the same five diabetes indicators, and only two indicators were selected by all five CHCs — namely, annual assessment of urine for microalbuminuria and HbA1c testing. For HbA1c testing, there were differences in the desired frequency, with some CHCs wanting at least two HbA1c tests per year and others wanting three tests per year. Factors that seemed to influence indicator selection and adherence included whether the selection was based on consensus or made by the medical director; provider training, experience, and beliefs; and CHC staff and organizational issues. In one CHC where the medical director chose the indicators, we later revised them midstream to reflect the priorities of a physician who had not been involved in the initial process and was the sole CHC physician participant in our initiative. From this experience and others, we learned that it was important to recognize and address local issues that could adversely affect indicator selection and/or staff buy-in and participation.
All five of our CHCs used paper records. While there is evidence that provider reminder systems such as diabetes flow sheets helped improve diabetes care, not all of our centers used them (10,11). Some were understandably resistant to adding yet another flow sheet to their already complicated charts. One CHC already had incorporated its diabetes measures into its adult health maintenance flow sheet. The majority of its patients did not have diabetes, and providers were only infrequently using that portion of the flow sheet. After some discussion, we arrived at the solution of placing colored stickers inside the charts on the adult health maintenance/diabetes flow sheets of their patients with diabetes. This change resulted in improved recognition of patients with diabetes and improved performance on the indicators. The use of flow sheets, in general, was associated with improved recognition and performance.
Any new initiative dependent on the participation of providers must compete with many other demands on their time during usual patient encounters (e.g., patient expectations and requests, professional concerns, diverse and sometimes conflicting practice guidelines and prevention recommendations, local and national initiatives, interruptions, emergencies). For example, even though four of our five partner CHCs were participating in the HRSA/CDC Diabetes Collaborative — whose members agreed to adopt local shared quality-improvement measures consistent with national guidelines — the level of participation still varied considerably from site to site. This taught us that participation in other diabetes programs was no guarantee of success.
In most centers, providers reacted to our initial presentation of results with disbelief, as both their individual and CHC levels of compliance were typically lower than they expected. During our meeting, they appeared to be comparing their results with center results and sometimes with another provider's results. We addressed the skepticism in several ways. First, we described our methods during our presentation (i.e., the comprehensiveness of the chart reviews, the use of two reviewers for quality control, the levels of interrater reliability). Second, we also asked the reviewers, who could be project and/or local office staff, to be present to respond to any questions. Third, we put the results in context by comparing each center's results with available national statistics that were typically about the same or worse. Generally, these strategies overcame barriers to acceptance, and we were able to move on to a more substantive discussion on what steps could be taken for improvement. We then facilitated discussions on what behavioral and structural changes providers could make as a staff to improve their results, and we offered technical assistance, such as help with the development of flow sheets or telephone consultation. By the end of the meeting, centers had usually developed a tentative plan for improvement. From this, we concluded that while obtaining provider-level data was more work, it generated a healthy interest and sense of competition among participants.
Given how busy providers often are, we looked for other ways to improve care. When feasible, we recommended implementing measures via "systems change" as an alternative to assigning a new responsibility to already overburdened providers. In one center, the medical director agreed with our recommendation that medical assistants take more responsibility for charting and ordering certain diabetes screening tests under standing orders, such as annual urine testing for microalbumin, annual lipid panel, and periodic HbA1c testing. We conducted a special training session for those staff. However, it took several visits before we noticed a change, and we were not confident that it would persist. We learned from this and other experiences that systems change at the practice level can be quite difficult to achieve and sustain.
Interpretation
Despite the many competing demands on CHCs, our project did achieve some success, and we believe that it was worthwhile. We helped the CHCs focus on interventions that matter but are sometimes neglected. Our CHCs often chose indicators based on national guidelines. They were motivated to review and, in many cases, improve their performances, thereby closing the gap that exists between research and practice. Three CHCs elected to continue beyond their original three-year commitment. For the CHCs that chose not to continue after their initial commitment, the availability of resources, particularly for medical audits, was an important issue.
We want to be careful about generalizing, since we worked with only five CHCs and no two were alike. Further, as consultants, we were not always privy to the activities and interventions that took place between our visits. Nonetheless, we observed that six factors were most important to overall success in our initiative: 1) the presence of an onsite champion, 2) broad staff and managerial support and participation, 3) the willingness of providers to delegate authority to ancillary staff via standing orders for routine tasks and testing, 4) the use of flow sheets, 5) the presence of a full-service diabetes clinic, and 6) access to provider-level data. In our experience, these are the factors that most favor success. We hope that these observations will prove useful to those contemplating similar initiatives.
We appreciate the support and cooperation of Sunset Community Health Center, Somerton, Ariz; Mariposa Community Health Center, Nogales, Ariz; Brownsville Community Health Center, Brownsville, Tex; Su Clinica Familiar, Harlingen, Tex; and Nuestra Clinica del Valle, Pharr, Tex.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Schachter KA, Cohen SJ. From research to practice: challenges to implementing national diabetes guidelines with five community health centers on the U.S.-Mexico border. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0079.htm
==== Refs
1 U.S. Census Bureau Population projections of the United States by age, sex, race, and Hispanic origin: 1995 to 2050 (current population reports no. P25-1130) Washington (DC) U.S. Government Printing Office 1996
2 U.S. Census Bureau The Hispanic population: census 2000 brief Washington (DC) U.S. Dept. of Commerce, Economics and Statistics Administration 2001 5
3 Centers for Disease Control and Prevention Fact sheet: a diabetes report card for the United States: quality of care in the 1990s Atlanta (GA) National Center for Chronic Disease Prevention and Health Promotion
4 National Institute of Diabetes and Digestive and Kidney Diseases Diabetes in Hispanic Americans Bethesda (MD) National Institutes of Health, National Diabetes Information Clearinghouse 2002 5
5 National Committee for Quality Assurance State of health care quality 2004 Washington (DC) National Committee for Quality Assurance 2004 9 Available from: URL: http://www.ncqa.org/
6 Institute of Medicine The chasm in quality: select indicators from recent reports Washington (DC) Institute of Medicine 2004 Available from: URL: http://www.iom.edu/subpage.asp?id=14980
7 Health Resources and Services Administration Community health centers Rockville (MD) U.S. Department of Health and Human Services 2003
8 Border collaboration [PowerPoint slide] Washington (DC) Pan American Health Organization, U.S.-Mexico Border Field Office [Accessed 2004 Jan] slide 18 Available from: URL: http://www.paho.org/%20english/ad/dpc/nc/dia-camdi-2003-mex-usa-border.ppt
9 Diabetes Project, diabetes along the border [Internet] Washington (DC) Pan American Health Organization Adapted 2004 May 5 Available from: URL: http://www.fep.paho.org/newdiabetes/english/
10 Cohen SJ Halvorson HW Gosselink CA 23 3 1994 284 291 Prev Med Changing physician behavior to improve disease prevention 8078848
11 Norris SL Nichols PJ Caspersen CJ Glasgow RE Engelgau MM Jack L 22 4 Suppl 2002 15 38 Am J Prev Med The effectiveness of disease and case management for people with diabetes 11985933
|
15670470
|
PMC1323320
|
CC BY
|
2022-01-26 23:15:54
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A17
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0075
Community Case Study
PEER REVIEWEDPasos Adelante: The Effectiveness of a Community-based Chronic Disease Prevention Program1
Staten Lisa K PhD Southwest Center for Community Health Promotion, Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona, Division of Health Promotion Sciences
2231 E Speedway Blvd, Tucson, AZ 85719 [email protected]
520-321-7777
Scheu Linda L MS, MPH Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona, Tucson, Ariz
Bronson Dan MS Division of Epidemiology and Biostatistics, Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona, Tucson, Ariz
Peña Veronica Regional Center for Border Health/WAHEC, Somerton, Ariz
Elenes Jo Jean Platicamos Salud, Mariposa Community Health Center, Nogales, Ariz
1 2005
15 12 2004
2 1 A182005
Background
Implementing programs that target primary prevention of chronic diseases is critical for at-risk populations. Pasos Adelante, or "Steps Forward," is a curriculum aimed at preventing diabetes, cardiovascular disease, and other chronic diseases in Hispanic populations. Pasos Adelante is adapted from the National Heart, Lung, and Blood Institute's cardiovascular disease prevention curriculum, Su Corazón, Su Vida, and includes sessions on diabetes and community advocacy and incorporates walking clubs.
Context
The Pasos Adelante curriculum was implemented in two Arizona, United States-Sonora, Mexico border counties. Key issues in these communities are safety, access to recreational facilities, climate, and cultural beliefs.
Methods
Pasos Adelante is a 12-week program facilitated by community health workers. The program includes interactive sessions on chronic disease prevention, nutrition, and physical activity. Evaluation of the program included precurriculum and postcurriculum questionnaires with self-reported measures of physical activity and dietary patterns. Approximately 250 people participated in the program in Yuma and Santa Cruz counties.
Consequences
Postprogram evaluation results demonstrate a significant increase in moderate to vigorous walking among participants and shifts in nutritional patterns.
Interpretation
The Pasos Adelante program demonstrates that an educational curriculum in conjunction with the support of community health workers can motivate people in Arizona/Sonora border communities to adopt healthy lifestyle behaviors.
==== Body
Background
The leading causes of death in U.S. Hispanic populations are heart disease/stroke, cancer, accidents, and diabetes mellitus (1). Health-compromising behaviors such as physical inactivity and poor nutrition are clearly linked to increased risk for many of these chronic diseases. Studies have demonstrated that increases in physical activity and a healthy diet can prevent or delay the onset of diabetes (1,2). This is of particular importance to individuals living along the U.S.-Mexico border, where diabetes prevalence is twice that of the rest of the nation (3). In general, U.S. Hispanics are less physically active than non-Hispanic whites (4). Although the typical eating habits of U.S. Hispanics are comparable to those of the general U.S. population, the U.S. Hispanic population consumes significantly fewer than the recommended five to nine daily servings of fruits and vegetables (5).
Effective primary prevention programs are crucial to reversing the high diabetes rates along the U.S.-Mexico border. Use of community health workers, or promotores de salud, is ideal for primary diabetes prevention programs that focus on lifestyle changes. Promotores provide enhanced community knowledge, dedication to health promotion, assistance with culturally appropriate program adaptation and supplementation, and personal knowledge of the disease experiences of their communities. Furthermore, promotores are shown to be effective in increasing access to care and preventive screening (6) and are frequently used in Hispanic community programs.
In 2000, the Mel and Enid Zuckerman Arizona College of Public Health received a legislative appropriation to fund a comprehensive diabetes prevention and control program, the Border Health Strategic Initiative (Border Health ¡SI!), in Cochise, Santa Cruz, and Yuma, Arizona counties. The Border Health ¡SI! program included behavioral intervention components targeting providers (7), people with diabetes (8), their families (9), the general community, and school environments (10). The program also included a policy component (11). For a detailed discussion of Border Health ¡SI!, read the overview by Cohen et al (3).
This paper describes the community component, Pasos Adelante, or "Steps Forward." The purpose of the project was to collaboratively develop and implement a community-based chronic disease prevention program and to demonstrate the effectiveness of the program in reducing risk factors for cardiovascular disease, diabetes, and other chronic diseases related to diet and physical activity. This article provides details about the unique communities involved, the participatory development and evaluation of the program between community organizations and university personnel, and the community and university perspectives about the value and impact of the program.
Context
The Pasos Adelante program was implemented in southwestern Yuma County and southern Santa Cruz County, Arizona, from 2000 to 2003. The communities share a border with Mexico. While each is unique in its economy, employment rates, demographic structure, and ethnic composition, as described by Cohen et al (3), the population of both communities is predominantly Hispanic, with high rates of poverty and unemployment. Various factors within these communities may contribute to health-compromising behaviors, including the availability of recreational facilities, climate, nutrition, and cultural beliefs.
There are virtually no recreational facilities, walking paths, or safe areas in which to exercise in these communities. Many neighborhoods do not have sidewalks or crosswalks. Additional concerns identified over the project period included lack of lighting and roaming dogs. The extreme heat acts as another barrier to physical activity. Temperatures range from an average low of 29°–41° F in December to 96°–106° F in July.
A survey conducted in 1998 in a neighboring Arizona border community reported a diabetes prevalence of 18%, obesity or overweight in 74%, and no regular physical activity among 67.4% of respondents (12). Many additional factors contribute to this high rate of overweight. Healthy food choices such as low-fat milk are not often readily available in rural communities (13). Because 63% to 78% of the population reports incomes less than 200% of the federal poverty level (14), many people do not have the resources to purchase the frequently more expensive healthier items.
An additional barrier to promoting healthy lifestyle choices may be cultural beliefs. Some participants expressed beliefs that people can develop diabetes by having it wished upon them or by putting the body through hot and cold extremes. These participants may not be receptive to the idea that physical activity and a healthy diet will in any way affect their chances of developing diabetes. Other cultural factors must be addressed also, such as the belief that walking is indicative of low socioeconomic status and therefore not desirable.
Methods
Formative phase
Two key concepts of the Border Health ¡SI! were partnership and collaboration between community partners and university personnel, as well as use of the promotor(a) model whenever appropriate. At the onset of Border Health ¡SI!, a need was identified to develop and implement a program facilitated by promotores to educate community members about diabetes and what can be done to prevent or delay its onset. The key partners in the development and implementation of the community component were Mariposa Community Health Center (MCHC) in Nogales, Ariz, and Regional Center for Border Health/Western Arizona Health Education Center (WAHEC) in Somerton, Ariz, with technical assistance from the University of Arizona (UA). Both community agencies have had award-winning promotor(a) programs for more than a decade.
During the formative phase of Border Health ¡SI!, diabetes prevention programs implemented by the two community agencies were documented and reviewed. Simultaneously, a comprehensive search for culturally appropriate prevention programs was performed to take advantage of existing curricula. Promotores and staff members from the two agencies suggested using Su Corazón, Su Vida, a cardiovascular disease prevention manual developed for Latinos by the National Heart, Lung, and Blood Institute (NHLBI) (15). After reviewing other curricula, the partners agreed to adapt the NHLBI curriculum to include diabetes. They chose the NHLBI curriculum for several reasons. First, the curriculum was developed by a trusted source. Second, cardiovascular disease is a major complication of diabetes and is therefore an appropriate part of a diabetes prevention program. Third, the partners supported the idea of expanding the target audience beyond persons at risk for diabetes. Fourth, most other curricula identified focused on patients with diabetes and not on disease prevention. Fifth, the curriculum was viewed as culturally competent; it targets the appropriate level of literacy and is available in Spanish. Finally, the curriculum was designed to be facilitated by promotores, and a number of them had previous training in the curriculum and liked the implementation style.
The main drawback to Su Corazón, Su Vida was that it did not focus on diabetes. A second drawback was the lack of a published program evaluation. Because Border Health ¡SI! would be adapting and evaluating the program, however, these concerns were not viewed as strong reasons not to use it.
Curriculum design
To take advantage of the Su Corazón, Su Vida curriculum, the manual was modified to broaden the focus to include an emphasis on diabetes prevention. Two sessions addressing diabetes and a single session on community health advocacy were created and inserted. Diabetes prevention materials used by the two community agencies were also incorporated. The revised curriculum is called Pasos Adelante.
The Pasos Adelante curriculum (available from: URL: http://www.borderhealthsi.org/steps_pasos.htm*) consists of a manual and free-standing flip charts created for Pasos Adelante in addition to a flip chart, telenovela magazine, and video available from Su Corazón, Su Vida. The Pasos Adelante curriculum was written in both English and Spanish, although it was only implemented in Spanish in our two communities. It follows the Su Corazón, Su Vida scripted teaching style; the promotores were given a script to follow if they desired, thereby enhancing the consistency of each session among promotores and from group to group. Additional background information was included in each session so that promotores were prepared for more in-depth questions about each topic. The manual consists of 12 two-hour sessions (Figure). The sessions "Are you at risk for diabetes?," "Glucose and sugar," and "Is your community healthy?" were designed for the Pasos Adelante curriculum. Each session consists of five main components: an introduction, the session in action, a weekly promise, a review of the day's most important points, and the close of the session. Furthermore, as part of the session in action, participants engage in a physical activity, such as dancing or aerobics, to reinforce the importance of physical activity.
Figure Pasos Adelante program session topics in Spanish and English (from introductory handout).
Sesiones de Pasos Adelante
image of a heart 1 ¿Está usted en riesgo de desarrollar enfermedades del corazón?
image of a family 2 Manténgase físicamente más activo.
image of a kidney 3 ¿Está usted en riesgo de desarrollar la diabetes?
image of a blood pressure cuff 4 Todo lo que necesita saber acerca de la presión arterial alta, la sal y el sodio.
image of a letter 5 Coma menos grasa, grasa saturada y colesterol.
image of a scale 6 Mantenga un peso saludable.
image of a community 7 Nuestra comunidad, ¿es saludable?
image of a sugar bowl 8 La glucosa y el azúcar.
image of a bowl of fruit 9 Goce con su familia de comidas saludables para el corazón.
image of money 10 Coma más saludable por su corazón — aun cuando tenga poco tiempo o dinero.
image of a cigarette 11 Goce de la vida sin el cigarrillo.
image of a diploma 12 Repaso y graduación.
Steps Forward Sessions
image of a heart 1 Are you at risk for heart disease?
image of a family 2 Be more physically active.
image of a kidney 3 Are you at risk for diabetes?
image of a blood pressure cuff 4 What you need to know about high blood pressure, salt, and sodium.
image of a letter 5 Eat less fat, saturated fat, and cholesterol.
image of a scale 6 Maintain a healthy weight.
image of a community 7 Is our community healthy?
image of a sugar bowl 8 Glucose and sugar.
image of a bowl of fruit 9 Make healthy eating a family affair.
image of money 10 Eat healthier — even when time or money is tight.
image of a cigarette 11 Enjoy living smoke-free.
image of a diploma 12 Review and graduation.
In addition to weekly classroom sessions, walking clubs were incorporated into the Pasos Adelante program. The walking clubs were designed to engage participants in recreational walking in a coordinated, socially supportive effort to increase physical activity. The walking club was designed so that participants would initially walk together outside of class for at least 20 minutes once a week with the promotor(a) at a mutually agreed-upon location, such as a park or local school track. Gradually, the group would build up to walking at least 20 minutes three times a week. At week seven of the program, the promotores start to withdraw from the groups but continue to encourage them during class sessions so that the groups can be self-sustained after the program ends. Walking clubs were incorporated into all Pasos Adelante sessions to move from the didactic focus of physical activity into actual behavior change. Staff and promotores from MCHC and WAHEC and UA personnel met monthly to discuss curriculum development.
Facilitator training
MCHC and WAHEC were contracted to implement the Pasos Adelante program. Each agency hired or reorganized existing promotores to participate in the Border Health ¡SI! program. The selection of promotores was left entirely to the agencies. Nine promotores employed by the two centers participated in curriculum design and received approximately six hours of manual training. During this training, evaluation instruments and protocols were also discussed and reviewed. All discussions and trainings were conducted in Spanish. Two additional promotores were hired during the project period and were trained individually by other promotores with technical assistance from UA personnel.
During the training, promotores were encouraged, but not required, to use the script. Emphasis was placed on the content and flow of each session. If promotores were unsure of themselves, they tended to rely on the script. Those who were comfortable making public presentations preferred a less formal style. Additional training was conducted when necessary throughout the program.
In addition to attending the Pasos Adelante training, many promotores had attended week-long trainings for Su Corazón, Su Vida at an annual community health worker conference cosponsored by WAHEC and therefore understood the fundamental design of the Pasos Adelante program. Promotores who had not attended Su Corazón, Su Vida training prior to starting the Pasos Adelante program did so during Pasos Adelante implementation. All promotores worked in pairs, with senior promotores paired with junior promotores. In addition, the promotores attended a variety of trainings on diabetes, including Diabetes: La Comunidad en Accion, sponsored by the Diabetes Today National Training Center and Diabetes Training for Lay Health Workers, sponsored by MCHC.
Program implementation
Eleven promotores (10 women, one man) led the sessions working in pairs. (The one male promotor facilitated one 12-week session of the program.) Sessions were scheduled for two-hour periods but ranged from 90 minutes to 150 minutes. At times the physical activity portion at the end of the class was eliminated to complete the educational portion.
To address some of the previously identified barriers, classes were conducted at centrally accessible public locations, such as schools, churches, the MCHC, and other public multipurpose rooms. One agency provided onsite childcare services. If necessary, participants were encouraged to bring their children or grandchildren. Class members decided the times of the class and walking groups. In the Yuma area, where the weather is the most extreme, the walking groups frequently met around 5:00 AM or in the late evening to avoid the heat. The promotores indicated that long-term residents did not have problems with walking so early or late, but newer residents did. According to the promotores, long-term residents regularly used those hours to avoid the heat.
Process evaluation
The community partners were essential to adapting and developing the Pasos Adelante manual. The promotores provided feedback on sessions and walking clubs using program-specific feedback forms. The feedback forms asked if information was missing and whether the information made sense, was adequate, and was presented in a style and manner easily understood by the group. After both agencies had completed one 12-week session, a meeting was held with all promotores to discuss what worked and what did not work. The promotores praised the curriculum and offered some minor grammatical corrections but little constructive criticism. Although the UA personnel were gratified to hear that the promotores liked the program, they were skeptical of the response and afraid that the promotores might have a cultural bias against expressing anything that sounded like criticism. So a second strategy for feedback was developed. All the manuals of the promotores were collected and examined. We found extensive notes in the margins of the manuals and additional handouts indicating where more information or clarification was needed. These notes allowed us to initiate a more direct conversation on the promotores' interpretation of the materials: "You wrote in the margin that. . . . Would you like to share with us what you mean?" This enabled us to avoid putting responsibility on the promotores for pointing out problems.
UA personnel were occasionally able to observe the sessions and meet with promotores afterwards. UA personnel contributed feedback on the actual presentation of the material and on effective communication styles. They also offered additional information, if needed.
Recruitment
Program participants represented a convenience sample recruited by promotores through presentations at schools, church groups, internal agency programs, and health fairs and by going door-to-door. Classes were offered year round from 2001–2003, except during holidays. Group size averaged 10 to 15 participants.
After a consent form was signed, each participant was asked to answer a standardized physical activity risk-assessment questionnaire to ensure that the individual was physically able to participate without any serious physical or medical risk. If an individual indicated any risk, he or she was then required to obtain a provider's permission to participate. One site offered screenings for all participants.
Participants then completed a questionnaire consisting of nutrition questions based on the Centers for Disease Control and Prevention's Behavioral Risk Factor Surveillance Survey (BRFSS) (16) and physical activity questions targeting moderate to vigorous activities using a one-month format adapted from the Minnesota Leisure Time Physical Activity Questionnaire (17). The BRFSS and adapted Minnesota questionnaire have been used among low-income Hispanic populations and have been shown to be reliable indicators of behavior. The intake form was designed to identify changes in the frequency and exertion levels of moderate to vigorous physical activity and in dietary consumption patterns that would reflect curriculum content. The questionnaire focused on changes in frequency in weekly consumption of fruit, vegetables, and sweet beverages. It also asked about the changes in the type of milk consumed and the type of fat used for cooking. Questionnaires were repeated at 12 weeks (the end of class).
Statistical methods
Intercooled STATA 7.0 (StataCorp LP, College Station, Tex) was used to analyze the quantitative data. McNemar chi-square tests were used for categorical data, Wilcoxon rank sum tests were used for independent continuous data, and Wilcoxon sign ranked tests were used for paired continuous data. Nonparametric tests were used because variables were not normally distributed. Matched pairs t-tests were calculated to compare age and body mass index (BMI). Significance was assigned at P ≤ .05. Because of small sample size, trends are identified if P ≤ .10.
Consequences
A total of 248 participants began the program and 216 completed it, for a completion rate of 87%. The participants who completed the program were mainly Hispanic women born in Mexico who had not graduated from high school (Table 1). The average participant age was 49.5 years. Compared to those who completed the program, the 32 individuals who did not complete the program were significantly more likely to be employed full- or part-time and to have asthma; more smokers were among those who did not complete the program.
As Table 2 shows, self-reported changes were found in levels of physical activity and in nutrition from preclass to postclass. The number of participants walking and the number of minutes per week of moderate to vigorous walking significantly increased. There were significant reductions in the weekly consumption of sweetened soda and sweetened hot drinks and an increase in the consumption of fruit juice. The number of servings of salads, vegetables, and fruits eaten per week also increased significantly.
Results differ between the two sites. Fewer significant changes were seen in Santa Cruz participants than in Yuma area participants. For example, in Santa Cruz, only the average number of minutes walking at a moderate pace and the average number of salads eaten per week increased significantly. These differences may be due to slight demographic differences, as shown in Table 1. Santa Cruz participants were significantly younger (45.1 ± 15.5 years vs 52.7 ± 13.4 years) and more educated; fewer were born in Mexico, and more had health insurance.
In addition to statistical evidence of positive changes, promotores frequently commented on seeing people walking and observing that some were losing weight. Anecdotal comments were overheard or recorded in the end-of-session evaluation sheets. For example:
"A person commented in my class that for the first time in her life she is walking for 15 minutes."
"One woman said that she felt really embarrassed to go out and walk, so she didn't. Now she's happy because she walks and feels comfortable doing it."
"One woman explained that she would eat a whole can of corn not imagining the number of portions and amount of sodium in it. She's going to pay more attention."
"One woman said she has lost six pounds since the beginning of the class. She is very happy. Also, her mother's blood sugar levels have dropped."
Interpretation
One of the key findings from this project is that while it is difficult to get people walking when the temperature is extremely high or when no sidewalks exist, it is not impossible. Residents in the Yuma area were not reluctant to walk during the summer months. Many residents have been farm workers and are used to an early-morning lifestyle. Participants were also able to make changes in their diet. Many of the participants initially indicated that they had no idea how to eat healthier.
The formal evaluation of community-based programs is difficult. Without funds to support more systematic evaluation, organizations frequently rely on self-reported data. Our evaluation instrument was feasible and effective for our agency. It provided statistical evidence for positive changes that at the very least indicate an increased awareness of healthy lifestyle behaviors among participants. The self-reported increases in walking matched the observations of the promotores. They report seeing a number of their participants continuing to walk without them. The promotores also report individuals losing weight and participants telling them that their providers are happy with their health improvements.
Another important outcome was the integration of the Pasos Adelante program and the Special Action Groups, community-based coalitions formed as part of the Border Health ¡SI! model (11). Promotores reported to the coalitions monthly to quarterly about issues raised during their class sessions. Based on reports from the promotores, the coalitions worked to have parks, playgrounds, and walking paths incorporated into city development plans. The creation of a new park in the Yuma area resulted directly from the coalition network and a motivated promotora. The Yuma coalition also targeted grocery stores as arenas for promoting and increasing the availability of healthier food options.
One of the drawbacks of this program was its lack of male participants. The Pasos Adelante program is based on the theoretical foundation of social support, including organized group activities to promote physical activity. Men may not be as responsive as women to programs emphasizing social support. In the future, it is important to determine how the Pasos Adelante program could be tailored to appeal to a male audience. Additionally, the program reached a generally older population. In communities where more than half the population is under the age of 35, programs need to be developed that target primary prevention and appeal to younger people.
The Pasos Adelante program has demonstrated that an educational curriculum in conjunction with the support of promotores can motivate people to adopt healthy lifestyle behaviors. The integration of classroom sessions and walking clubs allowed for increased interactions among participants and helped create social support for nutritional and physical behavioral change. In areas with abundant educational programs and sources of information, these kinds of programs may not result in behavior change; however, in areas with relatively few resources where residents have not repeatedly been exposed to prevention messages, these programs may have much greater impact.
Other communities can use the Pasos Adelante curriculum. Educational sessions have occurred with promotores in Mexico, and work has begun on the adaptation of the curriculum for a Native American health department. Prior to implementing the program, we would suggest that an advisory committee of local community members review the curriculum and decide what changes should be made to ensure that it is culturally appropriate. It is critical that the review committee include individuals who are truly part of the target community and share its cultural beliefs.
This project was funded by contract 200-2000-10070 from the Centers for Disease Control and Prevention. The authors thank all the promotores and participants for making the program so successful.
Figures and Tables
Table 1 Demographic and Health Status Comparison by Community of Pasos Adelante Participants (n=216) Who Completed Preprogram and Postprogram Assessments, Arizona, 2000–2003a
Characteristic Yuma County
Community
N=128 Santa Cruz County
Community
N=88 P
Demographics
Age, mean (SD) 52.7 (13.4) 45.1 (15.5) <.001
Hispanic 127 (99.2) 88 (100) .40
Female 113 (88.3) 84 (96.5) .03
Marital status
Married 96 (80.7) 55 (67.9) .002
Single/divorced 10 (8.4) 22 (27.2)
Widowed 13 (10.9) 4 (4.9)
Education
Some elementary 88 (69.8) 26 (29.9) <.001
Elementary 27 (21.4) 25 (28.7)
Some high school 7 (5.6) 26 (29.9)
High school 4 (3.2) 10 (11.5)
Currently employed 24 (19.0) 12 (13.8) .23
Health insurance 40 (32.3) 61 (69.3) <.001
Preferred speaking language is Spanish 121 (96.8) 76 (97.4) .80
Preferred reading language is Spanish 122 (97.6) 76 (96.2) .56
Length in community
<5 years 27 (21.1) 21 (23.1) .03
>10 years 84 (65.6) 54 (62.1)
Born in Mexico 124 (97.6) 75 (87.2) .003
Health Status
Diagnosed with diabetes 21 (16.4) 24 (27.6) .06
How long been a diabetic
<1 year 4 (19.1) 8 (36.4) .43
1-5 years 8 (38.1) 4 (18.2)
6-10 years 3 (14.3) 4 (18.2)
10+ years 6 (28.6) 6 (27.3)
Family member diagnosed with diabetes 53 (41.4) 54 (62.1) .002
Participant with a heart condition 6 (4.7) 10 (11.5) .17
High cholesterol 37 (29.1) 24 (27.3) .93
Diagnosed with cancer 6 (4.7) 3 (3.4) .66
Diagnosed with osteoporosis 12 (9.5) 5 (5.8) .21
Diagnosed with asthma 1 (0.8) 4 (4.6) .07
Current smoker 4 (3.1) 7 (8.0) .11
a Values are numbers (percentages) unless otherwise indicated. Percentages are based on the number of participants who responded to the question; not all participants (n=216) responded to every question.
Table 2 Preprogram and Postprogram Comparison of Self-reported Physical Activity and Dietary Intake, Pasos Adelante Participants (n=216), Arizona, 2000–2003
na Median (range) Mean (SD) P
Physical Activity
Fast walking, minutes/week 198 .002
Preprogram 0 (0-1800) 77.5 (±204.5)
Postprogram 0 (0-780) 108.9 (±160.0)
Moderate walking, minutes/week 191 <.001
Preprogram 0 (0-840) 73.7 (±117.7)
Postprogram 120 (0-840) 138.10 (±145.4)
Slow walking, minutes/week 202 .81
Preprogram 0(0-840) 45.7 (±107.1)
Postprogram 0(0-420) 40.5 (±82.1)
Dietary Intake
Soda, servings/week 204 <.001
Preprogram 0.5(0-49) 2.6 (±5.7)
Postprogram 0(0-21) 1.4 (±2.9)
Diet soda, servings/week 204 .20
Preprogram 0(0-35) 1.7 (±4.7)
Postprogram 0(0-35) 1.7 (±4.3)
Sweetened drink, servings/week 208 .24
Preprogram 2(0-56) 4.7 (±7.5)
Postprogram 2(0-28) 4.3 (±5.8)
Sports drink, servings/week 203 .07
Preprogram 0(0-28) 1.6 (±3.9)
Postprogram 0(0-21) 1.0 (±2.4)
Sweetened hot drink, servings/week 205 .03
Preprogram 7(0-49) 7.5 (±7.8)
Postprogram 7(0-28) 6.5 (±6.0)
Fruit juice, servings/week 205 .01
Preprogram 3.7(0-42) 5.8 (±6.0)
Postprogram 7(0-28) 6.6 (±5.7)
Salad, servings/week 208 <.001
Preprogram 3(0-35) 4.5 (±5.2)
Postprogram 5.5(0-28) 6.2 (±5.1)
Vegetables, servings/week 205 <.001
Preprogram 3(0-21) 5.3 (±5.0)
Postprogram 7(0.25-28) 7.8 (±5.9)
Fruits, servings/week 204 <.001
Preprogram 7(0-49) 8.8 (±7.3)
Postprogram 7(0-49) 11.7 (±7.9)
Fruits and vegetables, servings/week 201 <.001
Preprogram 12(0-63) 14.2 (±9.7)
Postprogram 16(3-63) 19.6 (±11.8)
a n = number of participants who answered the question in both the preprogram and postprogram questionnaire. Not all participants (n=216) responded to every question. SD = standard deviation.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Staten LK, Scheu LL, Bronson D, Peña V, Elenes J. Pasos Adelante: the effectiveness of a community-based chronic disease prevention program. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0075.htm
==== Refs
1 American Heart Association 2002 Heart disease and stroke statistics–2003 update Dallas (TX) American Heart Association
2 Tuomilehto J Lindstrom J Eriksson JG Valle TT Hamalainen H Ilanne-Parikka P N Eng J Med Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance 344 2001 18 1343 1350
3 Cohen SJ Ingram M 2005 1 Prev Chronic Dis [serial online] Border health strategic initiative: overview and introduction to a community-based model for diabetes prevention and control
4 1996 U.S. Department of Health and Human Services Atlanta (GA) Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion Physical activity and health: a report of the Surgeon General
5 Potter JD Finnegan JR Jr Guinard J-X Huerta EE Kelder SH Kristal AR 2000 Washington (DC) National Institutes of Health, National Cancer Institute 5 A Day for Better Health program evaluation report
6 Swider SM 19 1 2002 11 20 Public Health Nurs Outcome effectiveness of community health workers: an integrative literature review 11841678
7 Schachter KA Cohen SJ 2005 1 Prev Chronic Dis [serial online] From research to practice: challenges to implementing national diabetes guidelines with five community health centers on the border
8 Ingram M Gallegos G Elenes J 2005 1 Prev Chronic Dis [serial online] Diabetes is a community issue: the critical elements of a successful outreach and education model on the U.S.-Mexico Border
9 Teufel-Shone NI Drummond R Rawiel U 2005 1 Prev Chronic Dis [serial online] Developing and adapting a family-based diabetes program at the U.S.-Mexico border
10 Staten LK Teufel-Shone NI Steinfelt VE Ortega N Halverson K Flores C 2005 1 Prev Chronic Dis [serial online] The School Health Index as an impetus for change
11 Meister JS Guernsey de Zapien J 2005 1 Prev Chronic Dis [serial online] Bringing health policy issues front and center in the community: expanding the role of community health coalitions
12 Flood T Lebowitz MD De Zapien J Staten L Rosales C 1999 Douglas Community Health Survey: diabetes and health care in Arizona on the Mexican border Phoenix (AZ) ADHS
13 Abarca J Ramachandran S 2005 1 Prev Chronic Dis [serial online] Using community indicators to assess nutrition in Arizona-Mexico border communities
14 Arizona Department of Health Services Community Health Profiles for Nogales, San Luis and Somerton, AZ, 2001 Division of Public Health June 5, 2004 Accessed online at http://www.hs.state.az.us/hsd/chpweb/2001/
15 U.S. Department of Health and Human Services 5 2000 accessed 2004 Jun 14 Su corazón, su vida NIH Publication No. 00-4087 Bethesda (MD) National Institutes of Health, National Heart, Lung, and Blood Institute
16 Centers for Disease Control and Prevention Atlanta (GA) U.S. Department of Health and Human Services 2000 Behavioral Risk Factor Surveillance System Survey Questionnaire
17 Richardson MT Leon AS Jacobs DR Jr Ainsworth BE Serfass R 47 3 1991 271 281 J Clin Epidemiol Comprehensive evaluation of the Minnesota leisure time physical activity questionnaire
|
15670471
|
PMC1323321
|
CC BY
|
2022-01-25 23:15:13
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A18
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0076
Community Case Study
PEER REVIEWEDThe School Health Index as an Impetus for Change1
Staten Lisa K PhD Division of Health Promotion Sciences and Southwest Center for Community Health Promotion, Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona
2231 E Speedway Blvd, Tucson, AZ 85719 [email protected]
520-321-7777
Teufel-Shone Nicolette I PhD Division of Health Promotion Sciences and Southwest Center for Community Health Promotion, Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona, Tucson, Ariz
Steinfelt Victoria E MS Cooperative Extension, College of Agriculture and Life Sciences, University of Arizona, Yuma, Ariz
Ortega Nohemi Cooperative Extension, College of Agriculture and Life Sciences, University of Arizona, Yuma, Ariz
Halverson Karen Southeast Arizona Area Health Education Center, Nogales, Ariz
Flores Carmen Southeast Arizona Area Health Education Center, Nogales, Ariz
Lebowitz Michael D PhD Arizona Prevention Center and Southwest Center for Community Health Promotion, Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona, Tucson, Ariz
1 2005
15 12 2004
2 1 A192005
Background
The increase in childhood obesity and prevalence of chronic disease risk factors demonstrate the importance of creating healthy school environments. As part of the Border Health Strategic Initiative, the School Health Index was implemented in public schools in two counties along the Arizona, United States-Sonora, Mexico border. Developed in 2000 by the Centers for Disease Control and Prevention, the School Health Index offers a guide to assist schools in evaluating and improving opportunities for physical activity and good nutrition for their students.
Context
Between 2000 and 2003, a total of 13 schools from five school districts in two counties participated in the School Health Index project despite academic pressures and limited resources.
Methods
The Border Health Strategic Initiative supported the hiring and training of an external coordinator in each county who was not part of the school system but who was an employee in an established community-based organization. The coordinators worked with the schools to implement the School Health Index, to develop action plans, and to monitor progress toward these goals.
Consequences
The School Health Index process and school team participation varied from school to school. Individual plans were different but all focused on reducing in-school access to unhealthy foods, identified as high-fat and/or of low nutritional value. Ideas for acting on this focus ranged from changing the content of school lunches to discontinuing the use of nonnutritious foods as classroom rewards. All plans included recommendations that could be implemented immediately as well as those that would require planning and perhaps the formation and assistance of a subcommittee (e.g., for developing or adopting a district-wide health curriculum).
Interpretation
After working with the School Health Index, most schools made at least one immediate change in their school environments. The external coordinator was essential to keeping the School Health Index results and action plans on the agendas of school administrators, especially during periods of staff turnover. Staff turnover, lack of time, and limited resources resulted in few schools achieving longer-term policy changes.
==== Body
Background
Adult U.S. Hispanic populations living along the Arizona, United States-Sonora, Mexico border experience type 2 diabetes prevalence rates that are double the rate of the general U.S. population (1,2). The rate of type 2 diabetes is also rising among youth, especially in Mexican American children (3,4). School nurses in the border region report that the number of children with diabetes in their schools is increasing rapidly. Risk factors contributing to these rates are ethnicity, family history, obesity, physical inactivity, and poor nutrition (5).
Increases in rates of diabetes are closely associated with obesity rates. Obesity rates among U.S. children have been escalating rapidly over the past three decades (4,6,7). Data from the 2003 Centers for Disease Control and Prevention (CDC) Youth Risk Behavior Surveillance System show that 14% of Arizona high school students were at risk for becoming overweight, and 11% were overweight. While this was slightly lower than the U.S. estimates of 15% at risk and 14% overweight (8), no data are available for children living along the border. If extrapolations are made from adult data from the region (1,2), more children living along the border are at risk and are overweight than the general U.S. population. To reverse this trend of increasing obesity and diabetes in youth along the U.S.-Mexico border, interventions must target two modifiable risk factors: physical inactivity and poor nutrition.
Schools are ideal environments for promoting physical activity and good nutrition (9,10). Unfortunately, U.S. schools face many barriers to having healthy environments. The reduction or elimination of physical education (PE), the transfer of school food service to outside vendors, and reliance on vending machine revenues for extracurricular activities all contribute to a less-than-optimal health environment for children. In addition, these factors may be contributing to the dramatically increasing rate of childhood obesity in the United States (4). Policies and resources shaping the school environment impact students' patterns and levels of physical activity (9,10) and patterns of food and nutrient intake (11).
To address physical activity and nutrition in the school environment, the CDC developed the School Health Index for Physical Activity and Healthy Eating: A Self-Assessment and Planning Guide (SHI) in 2000 (12)To address physical activity and nutrition in the school environment, the CDC developed the School Health Index for Physical Activity and Healthy Eating: A Self-Assessment and Planning Guide (SHI) in 2000 (12). The SHI enables schools to 1) identify strengths and weaknesses of physical activity and nutrition policies and programs; 2) develop action plans for improving student health; and 3) involve teachers, parents, students, and the community in improving school services. The SHI manual consists of eight modules drawn from the CDC Coordinated School Health Program model. The SHI is a team-based assessment process. Recommended team members include administrators, teachers, school health workers, food service personnel, parents, and community health agencies. Team members respond to a series of questions in each module, and the questions are scored to yield an index reflecting their school's strengths and weaknesses. The SHI also includes a planning section that helps schools use the index scores to develop action plans (12).
Between 2000 and 2003, the SHI was implemented in 13 schools in two Arizona-Sonora border counties as part of the Border Health Strategic Initiative (Border Health ¡SI!) (13). Border Health ¡SI! was a legislative appropriation for a comprehensive diabetes prevention and control program in Cochise, Santa Cruz, and Yuma, Arizona counties. Border Health ¡SI! consisted of policy coalitions and interventions targeting providers, people with diabetes, their families, the general community, and schools in two of the border counties. This paper describes the schools component of Border Health ¡SI!. It provides a case study of the SHI implementation process for seven elementary schools and the barriers to change encountered in the school environment.
Context
Schools were recruited from the Nogales area of Santa Cruz County and the communities of Somerton and San Luis in Yuma County, Arizona. Nogales had a population of approximately 21,000 in 2001 and is predominantly Hispanic (97%) (14). The majority of individuals (64%) had incomes less than 200% of the federal poverty level. Most adults (52%) did not have a high school diploma, and 17% were unemployed (14).
Border Health ¡SI! recruited eight schools from three public school districts in the Nogales area. Combined, these districts serve approximately 9256 students and have 10 elementary schools, three middle schools, and three high schools. Of these 16 schools, six did not meet the federal Leave No Child Behind criteria in 2003, and two were underperforming (15-17). No schools were classified as excelling. During the first two years of Border Health ¡SI!, one district was on a year-round calendar. In 2002, this district resumed a traditional calendar. A small district (made up of one school) kept the year-round schedule.
In the Yuma area, five schools were recruited from the communities of Somerton and San Luis, which are 100% Hispanic. Combined, the two communities had a population of approximately 24,610 in 2001 (18,19). Between 63% and 78% of the population had incomes less than 200% of the federal poverty level. The majority of adults (62%–65%) did not have a high school diploma, and a large percentage (44%–66%) was unemployed (18,19).
Border Health ¡SI! worked with two public school districts in Yuma County. At that time, the districts served approximately 6524 students from K–12, with six elementary schools, two middle schools, and one high school. Of the nine schools, four were ranked by the state of Arizona as underperforming, and four did not meet the federal Leave No Child Behind criteria in 2003 (20,21). No schools were classified as excelling. Because of exploding population growth, two schools in one district were running double sessions (7:00 AM–12:25 PM and 12:30 PM–6:00 PM) with two sets of principals and teachers. This schedule did not allow time for extra activities or even a vacant meeting room.
Methods
Selection of SHI
The community-based agencies involved with Border Health ¡SI!, along with technical assistance from the Mel and Enid Zuckerman Arizona College of Public Health (MEZACOPH), selected the recently released SHI as a tool that would enable schools to start thinking about creating healthier environments. Despite the lack of published evaluation results, the Border Health ¡SI! group felt that it was a reasonable tool to focus schools on physical activity and nutrition policy. The SHI and follow-up were the only school-based interventions as part of Border Health ¡SI!.
Project design
The University of Arizona Cooperative Extension in Yuma County and Southeast Arizona Area Health Education Center (SEAHEC) in Santa Cruz County facilitated implementation of the SHI. These two community-based agencies had strong existing relationships with local schools. SEAHEC was involved in a variety of school nutrition education programs, and Cooperative Extension was responsible for the 4-H clubs for children and thus also worked closely with area schools.
The community agencies and others involved in Border Health ¡SI! expressed concern that the schools were overburdened and that health promotion and chronic disease prevention might not be high priorities. We believed, however, that an outside advocate could discuss the serious issues related to chronic disease and how they impact children. We also believed that resource-stressed schools would accept external coordinators from established and trusted agencies to provide assistance and support. Cooperative Extension and SEAHEC identified staff members who could serve as external coordinators to assist schools in completing the SHI assessment and planning process and in coordinating and compiling the SHI materials. MEZACOPH staff provided education to external coordinators on the SHI and the relationship between adolescent health and chronic disease. External coordinators were responsible for documenting recruitment efforts, the SHI process within the schools, team member activities, and action plans.
School recruitment
External coordinators initiated the recruitment process by presenting the SHI to school district superintendents or assistant superintendents, the school boards, or directly to principals. In addition, external coordinators contacted schools where they had personal connections. By the end of the third year, all schools in the area were approached, and any that expressed interest were contacted. The external coordinator provided a verbal overview and copy of the SHI, and if received positively, made a presentation to school personnel. When schools were hesitant to participate, additional assistance was sought from two directors of health services, a school board president, and a registered nurse at a school-based clinic to encourage schools to participate. Schools were offered a financial incentive of $1500 upon completion of the SHI. The incentive was provided by Cooperative Extension and SEAHEC. Schools were encouraged to apply the funds toward their action plans but were not required to do so. At the end of the three-year period, the SHI was implemented in 10 elementary schools, two middle schools, and one high school, about half of the 25 schools approached.
Implementation of the SHI
Once a school agreed to complete the SHI, the principals identified an internal SHI coordinator. The internal coordinator recruited team members, and the external coordinator met with the team to provide materials and an overview and outline the assessment process. The external coordinator then followed up with the internal coordinator to ensure that the teams were meeting, collected scorecards, produced a summary document, and scheduled and helped conduct the action planning session.
Evaluation plan
Because Border Health ¡SI! was a legislative appropriation and was not funded as a research project, resources were not available to do an in-depth study of the SHI. However, the community-based agencies and MEZACOPH reached an agreement on an evaluation plan. Documentation of the plan would include information on the interest schools showed in completing the SHI, the ability of SHI teams to understand and complete the SHI, the process for completing the evaluation instrument, whether or not the teams created action plans, and finally, whether or not schools were able to make any changes suggested in the action plans.
Detailed quarterly reports by Cooperative Extension and SEAHEC were used to document school interest and the SHI process. MEZACOPH staff conducted in-depth interviews with the external coordinators and with a convenience sample of 14 SHI team members from the first five schools to complete the SHI. The purpose of the interview was to identify barriers to implementing the SHI, to find out whether or not SHI team members believed an external coordinator was necessary, and to identify changes in the school environment that could be attributed to the process. The external coordinators were in frequent contact with the schools after completion of the SHI. A year after completion of the SHI, external coordinators contacted the schools to determine whether or not SHI action plans were being implemented.
Consequences
The SHI process was different in each of the schools. The time frame for complete implementation ranged from four weeks to almost two academic years. Most schools completed the SHI within one semester. Schools were encouraged to complete the SHI during the fall semester to allow for time to work on action plans. Unfortunately, all schools completed the SHI during the spring semester.
The SHI team composition varied by school. Team size ranged from six to 34. All schools included at least one parent. As described earlier, the border communities are predominantly Hispanic; therefore, at several schools, parents emphasized the need for a Spanish version of the SHI. A health educator and social worker were not included for Yuma area schools because these positions do not exist in these schools.
An in-depth interview with 14 SHI team members revealed that team members felt that the SHI helped to build awareness of school commitment, identify changes that do not require resources, encourage policy and action, bring health issues to the schools' attention, and raise awareness of federal policies. The team members also identified the four key barriers to implementing the SHI: 1) time, 2) getting people to meetings, 3) initial buy-in, and 4) perceived lack of expertise. The SHI team members believed that the key roles played by the external coordinators were facilitation and guidance. The external coordinators also assisted in overcoming barriers.
No school included representatives from community health agencies on the SHI team. Although outside community members were suggested by SHI guidelines, both internal and external coordinators felt that these individuals did not have the in-depth knowledge of the school environment necessary to answer the detailed SHI questions. The external coordinators frequently filled this role as representatives of the community. As part of the Border Health ¡SI!, community coalitions or Special Action Groups (SAGs) were established to focus on policy change to create healthier communities (22). The external coordinators regularly updated these coalitions on SHI progress in schools, and the coaltions served as resources to the schools. Coalition members included the external coordinators, school administrators, nurses, teachers, and a wide variety of community agencies (22).
Santa Cruz County
By the end of Border Health ¡SI!, four elementary schools, one K-8 school, two middle schools, and one high school completed the SHI in Santa Cruz County. During the final quarter of the project period, the director of health services for one district (two elementary schools, one middle school, and one high school) used SHI results to develop the district's comprehensive health plan. Results were not available from the individual schools, and the district health plan is not complete at this writing. The information reported here is based on reports from three of the elementary schools representing two school districts. Overall, half of the schools in the area completed the SHI. The main reasons for schools refusing to participate were lack of time and lack of a school champion. When a school champion was identified, time became less of a barrier.
The intent of the SHI is for SHI teams to collaboratively develop action plans based on the results from the eight SHI modules. In two of the Santa Cruz County schools, the SHI was completed at the end of the semester, and the teams were unable to develop action plans. The principals, who had participated on the teams, created the action plans in isolation. This process was especially problematic because one of the principals resigned shortly after creating a plan. In fact, by the beginning of the next academic year, only two of the seven SHI team members were still at their school (hereafter referred to as Santa Cruz School 1 [SC1]). The second principal to develop action plans promoted many school-level changes (Santa Cruz School 2 [SC2]). Unfortunately, she also left her position.
Case Study 1
SC1 made several changes as it implemented the SHI, including the decision to move all candy sales out of the cafeteria. Most notably, the school hired a full-time PE teacher and developed a policy that prohibited the selling of junk food for fundraisers. The fundraising policy came into question the following academic year when the loss of almost all members of the SHI resulted in very little institutional memory. This case offers an example of how important the external coordinator's role is in the process. The remaining SHI team member was uncomfortable promoting the action plans to the new principal. The external coordinator contacted the principal early in the school year to discuss the SHI and action plans. The long-term goal of the SHI team was to increase the length of the lunch period. The principal was unfamiliar with the SHI but expressed verbal support for the action plans. When resistance developed over the fundraising policy, the principal enlisted the aid of a remaining SHI team member to explain the policy to the staff and parent-teacher organization. The policy was upheld. SC1 had a new principal the following year. The external coordinator continued to follow up and provided the impetus for the principal to create a new SHI team that completed the SHI a second time. SC1 has had a full-time PE teacher for three years. They have not lengthened the school lunch period. Fundraisers now include gifts, wrapping paper, and magazines as well as chocolate. Items for sale at lunch include pencils, notebooks, pickles, oranges, peanut butter crackers, and salditos (salted dried prunes) instead of baked goods and junk food.
Case Study 2
The principal/superintendent at SC2 was deeply committed to the SHI process. The school made several immediate changes. The school bake sales were converted to healthy snack sales. Graduation cookies and punch were cancelled and replaced with lemonade and baked chips and salsa. As a result of the SHI, the school hired a full-time PE teacher and developed a PE course. The goal was to have a letter grade for the course, not a pass/fail grade. An existing staff member was certified as a PE instructor. The school also organized track and basketball teams for the first time. The following summer, the school board attempted to drop the program. The principal called the Border Health ¡SI! coalition for assistance. Coalition members, teachers, and parents attended the school board meeting. The principal presented the SHI results, and the teachers and parents strongly supported the program. The PE program was not cut. Unfortunately, before the next academic year, the principal accepted a job as a district superintendent in another community. In that role, she was able to add PE into all elementary schools. The PE course is pass/fail. Individual teachers are offering low-fat snacks and less sweet snacks at parties, but this is not a coordinated school effort. Carrots, orange juice, milk, graham crackers, and cheese sandwiches are served at the school open house.
Case Study 3
The third Santa Cruz school (SC3) to complete the SHI took one and a half years to complete the process. The school was undergoing major renovations when it was first contacted and despite expressing interest, it was unable to commit to the process. Once the school committed, however, the external coordinator reported that the SC3 SHI team was the most enthusiastic. The school removed the vending machine from the cafeteria and decided to remove all other soda and candy vending machines and replace them with healthier choices. The beverage machines include fruit juices, water, tea, and lemonade. The SHI team also recommended removing the school store from the cafeteria. The SHI team reported their results to the Border Health ¡SI! coalition. The team was encouraged by the coalition to write a newspaper article for the local paper describing the changes they were making. Students were interviewed by a local television station when the soda was replaced by fruit drinks. The students said they missed the soda "but juice was okay."
The school store has not moved, but it is no longer selling candy. It replaced candy with healthier options, and profits have stayed the same. Bake sales now have oranges, cucumbers, and carrots. Some sales include candy but only small candy bars.
Yuma County
A total of five schools, four elementary schools and one middle school, implemented the SHI as part of the Border Health ¡SI! in the southern Yuma area. All three elementary schools in one district participated. In the second district, only one elementary school participated. It was the only elementary school not running double shifts. Overall, five of the nine schools in the area completed the SHI. The main reason for refusing to participate was lack of time. In Yuma, the external coordinators actively facilitated the action planning process. Documentation from the Yuma area focuses on the action plans.
Case Study 1
All teachers participated on the SHI teams at the first school (YC1) in the Yuma area. Not all teachers were enthusiastic. A school policy was developed to prohibit the use of food as a reward in the classroom, and students were limited to one snack per day. The SHI team also set goals of 1) teaching more nutrition information, 2) gathering and disseminating information on all food and drink items sold at the school, and 3) shifting from selling junk food items and sports drinks to healthier options. The school offered granola and fruit bars as alternatives, but they did not sell, and the school reverted to selling junk food. YC1 established walking groups two to three days per week for students. Teachers organized structured fitness breaks. The school also implemented an annual field day for staff and students.
Case Study 2
YC2 also established a school policy prohibiting nonnutritious food as a reward and limiting snacks to one per day per student. The SHI team was especially interested in eliminating outdated health education material and incorporating a sequential health education curriculum. Teachers began standing at the salad bar to encourage children to eat more fruit and vegetables. The school also began to increase use of community facilities and to offer a swimming class at the local community pool. The school's part-time PE teacher left the school for a full-time position out of the district. The SHI team presented the SHI results at a staff workshop.
Case Study 3
YC3 established a lengthy list of goals, including requiring hand washing, adopting a sequential health education curriculum, and incorporating activity breaks into the classroom. Two years after implementation of the SHI, no progress has been made.
The three Yuma area elementary schools described here are from one school district. The superintendent and SHI team members requested that the external coordinators from Cooperative Extension present the SHI results to the school board. The perception was that the board would view the external coordinators as community members and not as school personnel, and they would thus have a greater impact. The SHI team members then presented the results to the Border Health ¡SI! coalition.
Case Study 4
The fourth Yuma area elementary school (YC4) was from the second school district and focused on similar issues as the other three schools. This school chose to focus on increasing instruction time for health and PE, adopting new health education textbooks, and presenting health-related information to parents and staff. At this point, no progress has been made toward these goals.
The action plans of these schools emphasized modifying health-related curricula and adopting a sequential health education curriculum. Upon exploring the issue at the district level, it was determined that the district is not considering modifications to health education curriculum until 2005. To provide the teachers with resources, Cooperative Extension presented a resource event where teachers selected educational materials focusing on health education, nutrition, and physical activity and provided suggestions on ways to incorporate nutrition messages and physical activity into the existing curricula. In addition, Cooperative Extension also provided one school library with more than 50 books on physical activity and nutrition (fiction and nonfiction).
Interpretation
Seven Arizona elementary schools along the Arizona-Sonora border assessed their school environments and developed action plans using the CDC's SHI. Despite some difficulties resulting from time constraints and human resources, all schools were able to complete the assessment. The schools focused their action plan priorities on nutrition and physical activity. School action plans included items that could be addressed immediately (e.g., a school policy prohibiting use of candy as rewards) as well as policies that would necessitate implementation at the district level (e.g., adopting a sequential health curriculum or seeking funds to hire a full-time PE instructor). The individual school action plans varied, but all seven shared one component: to reduce internal access to unhealthy foods. Plans included changing the content of school lunches, discontinuing the use of nonnutritious foods as classroom rewards, moving candy sales and a snack bar away from the cafeteria, and choosing healthy alternative fundraisers. Two schools were able to use SHI results to hire PE teachers.
The goal of Border Health ¡SI! was for schools to disseminate the results of the SHI at least to the teachers at each participating school. In addition, we hoped that the results would be presented at parent-teacher and school board meetings. High rates of staff turnover highlight the importance of disseminating SHI results and action plans to a larger audience. It is crucial for schools to have an advocate for physical activity and nutrition. The external coordinator and Border Health ¡SI! coalitions served as those advocates and were able to keep the issue of school health prominent in school administrators' minds.
The number of underperforming schools in these districts and the lack of highly performing or excelling schools results in stress on the educational system. Principals and school board members are under pressure to improve the academic performance of their schools. When approached for funds or time to implement new programs, school administrators frequently cite the need for scholastic improvement above all other issues. External coordinators returned repeatedly to meet with school administrators to discuss the importance of physical activity and good nutrition. An additional tactic for educating administrators was to recruit them to participate in the Border Health ¡SI! SAGs.
The current educational system is responsible for a wide variety of activities: academic performance, social services, childhood immunizations, and the nutrition of students through the federal school breakfast and lunch programs. Involving schools in health promotion and disease prevention activities is critical, but we must recognize that most school systems are under extreme pressure to demonstrate academic progress. The fact that half of schools approached in two low socioeconomic areas participated in the SHI process indicates the commitment of school personnel to the overall health and well-being of their students and communities. The support of an external coordinator from a local agency can assist in removing some logistical barriers.
While most schools were extremely committed and enthusiastic about SHI action plans, staff turnover, time, and limited resources were barriers to progress even with the support of an external facilitator. Implementation of new programs is limited further by the low number of certified PE and health education specialists employed by the districts. Change in staff in one district occurred during the first year, and staff had to be educated about the SHI action plans and school goals. Some schools were undergoing major renovations and building projects, and in some cases, new schools were built. These schools found it challenging to take on new projects like the SHI.
One cautionary note is that publicity may present a barrier to acceptance of the SHI process. In the opinion of the external coordinators, publicity surrounding the removal of soda machines at one school may have discouraged other schools from participating in the SHI process. When funding for schools is so tight that vending machine sales are used to support photocopying expenses, field trips, graduation ceremonies, and extracurricular events, the threat of losing those funds can deter schools from eliminating this source of revenue, even though they are interested in improving the health of their students. The SHI and our project did not push schools to remove vending machines; instead, we encouraged schools to identify the best priorities for their own schools. These messages were lost by the news media. The external coordinator sought advice from the Border Health ¡SI! SAGs on how to convey our messages. The negative impressions receded after a brief time, and five more schools completed the SHI by the end of Border Health ¡SI!.
This project showed the value of having an external coordinator to help with continuity and with keeping the project top-of-mind with school officials, especially during periods of high staff turnover. In addition, the external coordinators acted as a resource beyond coordination. External coordinators created resource manuals for alternative ideas for school fundraisers, linked schools with other community resources, found funds to provide teachers with educational materials, and volunteered at school events. Although the SHI process does not include incentives, we found that monetary incentives for carrying out policy priorities seemed to encourage participation and gave schools resources to implement policies. Overall, most schools were able to implement immediate changes. Policies requiring a longer-term process and additional resources were more difficult to carry out. Future projects should focus on documenting whether students increased their physical activity or improved their eating habits as a result of SHI policies.
This project was funded by Contract 200-2000-10070 from the Centers for Disease Control and Prevention. We thank school personnel and district administrators in the Gadsden, Nogales, Santa Cruz, Santa Cruz Valley, and Somerton school districts.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Staten LK, Teufel-Shone NI, Steinfelt VE, Ortega N, Halverson K, Flores C, et al. The School Health Index as an impetus for change. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0076.htm
==== Refs
1 West SK Klein R Rodgriguez J Muñoz B Broman AT Sanchez R Snyder R 2001 24 1204 1209 Diabetes Care Diabetes and diabetic retinopathy in a Mexican-American population: Proyecto Ver 11423503
2 Flood T Lebowitz MD De Zapien J Staten L Rosales C 1999 Douglas community health survey: diabetes and health care in Arizona on the Mexican border ADHS Phoenix (AZ)
3 Bobo N Evert A Gallivan J Imperatore G Kelly J Linder B 2004 114 259 263 Pediatrics An update on type 2 diabetes in youth from the National Diabetes Education Program 15231940
4 Miller J Rosenbloom A Silverstein J 2004 89 4211 4218 J Clin Endocrinol Metab Childhood obesity 15356008
5 American Diabetes Association Accessed 9/15/04 http://www.diabetes.org/
6 Ogden CL Flegal KM Carroll MD Johnson CL 2002 288 1728 1732 JAMA Prevalence and trends in overweight among US children and adolescents, 1999-2000 12365956
7 Strauss RS Pollack HA 2001 286 2845 2848 JAMA Epidemic increase in childhood overweight, 1986-98 11735760
8 Centers for Disease Control and Prevention Arizona Youth Risk Behavior Survey 2003 state fact sheet [Internet] Accessed 2004 Oct 15 Atlanta (GA) Centers for Disease Control and Prevention 2004
9 Sallis JF Conway TL Prochaska JJ McKenzie TL Marshall SJ Brown M 91 4 2001 618 620 Am J Public Health The association of school environments with youth physical activity 11291375
10 Datar A Sturm R 2004 94 1501 1506 Am J Pub Health Physical education in elementary school and body mass index: evidence from the early childhood longitudinal study 15333302
11 Cole SM Teufel-Shone NI Ritenbaugh CK Yzenbaard RA Cockerham DL 101 7 2001 802 806 J Am Diet Assoc Dietary intake and food patterns of Zuni adolescents 11478480
12 Centers for Disease Control and Prevention 2000 Atlanta (GA) Centers for Disease Control and Prevention School Health Index for physical activity and healthy eating: a self-assessment and planning guide. Elementary version
13 Cohen SJ Ingram M 2005 1 Prev Chronic Dis [serial online] Border Health Strategic Initiative: Overview and introduction to a community-based model for diabetes prevention and control
14 Division of Public Health, Arizona Department of Health Services Nogales community health profile 2001 [Internet] Accessed 2004 Sep 1 Phoenix (AZ) Arizona Department of Health Services 2004
15 Arizona Department of Education Arizona district report card 2003-04, Nogales unified district [Internet] Phoenix (AZ) Arizona Department of Education Accessed 2005 Sep 15
16 Arizona Department of Education Arizona district report card 2003-04, Santa Cruz Valley unified district [Internet] Phoenix (AZ) Arizona Department of Education Accessed 2005 Sep 15
17 Arizona Department of Education Arizona district report card 2003-04, Santa Cruz elementary district [Internet] Phoenix (AZ) Arizona Department of Education Accessed 2005 Sep 15
18 Division of Public Health, Arizona Department of Health Services Somerton community health profile 2001 [Internet] Phoenix (AZ) Arizona Department of Health Services 2004 Accessed 2004 Sep 1
19 Division of Public Health, Arizona Department of Health Services San Luis community health profile 2001 [Internet] Phoenix (AZ) Arizona Department of Health Services 2004 Accessed 2004 Sep 1
20 Arizona Department of Education Arizona district report card 2003-04, Gadsden elementary district [Internet] Phoenix (AZ) Arizona Department of Education Accessed 2005 Sep 15
21 Arizona Department of Education Arizona district report card 2003-04, Somerton elementary district [Internet] Phoenix (AZ) Arizona Department of Education Accessed 2005 Sep 15
22 Meister J Guernsey de Zapien J Prev Chronic Dis [serial online] 2005 1 Bringing health policy issues front and center in the community: expanding the role of community health coalitions
|
15670472
|
PMC1323322
|
CC BY
|
2022-01-24 23:35:54
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A19
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0083
Community Case Study
PEER REVIEWEDDeveloping and Adapting a Family-based Diabetes Program at the U.S.-Mexico Border1
Teufel-Shone Nicolette I PhD Mel and Enid Zuckerman Arizona College of Public Health
PO Box 245158, University of Arizona, Tucson, AZ 85724 [email protected]
520-321-7777, ext 16
Drummond Rebecca MA Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona, Tucson, Ariz
Rawiel Ulrike MS Mel and Enid Zuckerman Arizona College of Public Health, University of Arizona, Tucson, Ariz
1 2005
15 12 2004
2 1 A202005
Context
The prevalence of diabetes among Hispanics is more than twice that of non-Hispanic whites in communities along the U.S.-Mexico border. The University of Arizona and two community health agencies on the Arizona border, Campesinos Sin Fronteras and Mariposa Community Health Center, collaborated to design, pilot and assess the feasibility of a lay health-outreach worker- (promotora-) delivered diabetes education program for families. La Diabetes y La Unión Familiar was developed to build family support for patients with diabetes and to teach primary prevention behaviors to family members.
Method
Community and university partners designed a culturally appropriate program addressing family food choices and physical activity, behavior change, communication, and support behaviors. The program offers educational content and activities that can be presented in home visits or multifamily group sessions. Community partners led the implementation, and university partners guided the evaluation.
Consequences
Seventy-two families (249 total participants) including children and grandchildren participated. Preintervention and postintervention questionnaires completed by adults (n = 116) indicate a significant increase in knowledge of eight diabetes risk factors (P values for eight factors range from <.001 to .006) and a significant increase in family efficacy to change food (P < .001) and activity behaviors (P < .001). Interviews with participants highlight the program's positive psychosocial impact.
Interpretation
Community and university collaboration involved building upon thepromotoras' expertise in engaging the community and the university's expertise in program design and evaluation. A promotora-delivered family-based diabetes prevention program that emphasizes family support, communication, and health behaviors is feasible and can yield change in family knowledge, attitude, and behavior relative to diabetes risk factors.
==== Body
Background
Among Hispanic populations in the United States, rates of type 2 diabetes and secondary complications (e.g., retinopathy, neuropathy, renal failure) are more than twice those reported for non-Hispanic whites (1-3). The high prevalence of associated conditions corresponds to the poor rates of adherence to American Diabetes Association recommendations for diabetes self-management documented for this population (4,5). Hispanic adults in the United States with diabetes and cardiovascular disease cite attitudes, perceptions, and preferences of family members as significant barriers to making recommended changes in their diet and exercise patterns (4,6,7). Yet family members of Hispanic patients with diabetes are at a particularly high risk for developing the disease themselves because of their family history of diabetes and because of high rates of risk factors (e.g., gestational diabetes, obesity, physical inactivity) in this population (1,8). Providing diabetes education to the entire family would address both prevention and treatment.
Because of a strong cultural emphasis on family connectedness among Hispanics, family members may be particularly influential in the development of health behaviors in this population (7,9-11). Family behaviors and attitudes can support or challenge a patient's psychosocial adaptation to illness and subsequently a patient's confidence, intent, and willingness to implement disease-management strategies (12-15). In a study of 76 Hispanic patients with diabetes, Chesla et al (16) reported that 23% of the patients noted troubling changes in family relations since their diagnosis, and 32% feared that more changes created by the social and financial stress of the patient's increasing dependency on other family members were imminent.
In several studies from Mexico, family support has been shown to have significant positive association with glucose control in patients with diabetes (17,18). Family diabetes education studies in Cuba (19) and Costa Rica (20) have demonstrated improved glycemic control and treatment compliance in patients whose families had information about diabetes and were supportive of recommended health behaviors. In the United States, family education and health-behavior change among Hispanic populations has been related to weight loss. Foreyt et al (21) and Cousins et al (22) demonstrated that interactive family-oriented weight-loss interventions yield greater success than approaches that target the individual. Both studies highlight the importance of family support within Hispanic culture and call for additional efforts to evaluate the effectiveness of family-oriented health interventions.
Despite the documentation of family influence on health behaviors, particularly within Hispanic populations, the literature offers no U.S. examples of diabetes interventions that target the family and address the family's collective behaviors. Brown et al (23,24) encourage the involvement of a family member in support groups in a diabetes self-management intervention but do not specifically target the family or family behaviors.
La Diabetes y La Unión Familiar (Diabetes and the Family) is the family component of the Border Health Strategic Initiative (Border Health ¡SI!). Border Health ¡SI! is a comprehensive community-based diabetes prevention and control project developed by the Mel and Enid Zuckerman College of Public Health at the University of Arizona (MEZACOPH) and various community partners serving Arizona residents at the U.S.-Mexico Border. The various components and multiple community-university partnerships of Border Health ¡SI! are described in this issue (25).
The objectives of this family-based diabetes education intervention are the following: 1) to enhance family members' social support of patients with diabetes and 2) to increase the range of primary prevention behaviors associated with diabetes in family members of patients with diabetes. This community case study describes the collaborative development, delivery, and outcome of the initial implementation of La Diabetes y La Unión Familiar.
Context
Geographically and demographically, the context of Border Health ¡SI! and the family component are the underserved Hispanic communities of Yuma and Santa Cruz counties, located along Arizona's southern border with Mexico. Both counties report an annual median household income of less than $33,000 with greater than 19% of residents living below the poverty level (26). In addition, 43.5% of households in Yuma County and 79.2% in Santa Cruz County report that Spanish is the predominant language spoken at home (26). Many residents in these counties do not seek regular health care because they lack access to Spanish-speaking health care providers and do not have access to health insurance (20% of Arizona residents do not have health insurance); some are fearful about their immigration status (26,27).
To adapt to the needs of Hispanic families in these counties and to the skills of local promotoras, MEZACOPH, Campesinos Sin Fronteras (CSF) in Yuma County, and Mariposa Community Health Center (MCHC) in Santa Cruz County collaborated to develop, implement, and evaluate La Diabetes y La Unión. The promotoras from MCHC had extensive experience in providing education and support to Hispanic patients with diabetes. The promotoras from CSF had received national attention for their health education efforts extended to Hispanic families and migrant farm workers, but they had limited experience with diabetes education. In both communities, the promotoras reported the need for a diabetes intervention that addressed family support and education. Program development used promotoras' knowledge of local context and experience in providing outreach services to this population.
Methods
Development
MEZACOPH investigators, CSF and MCHC directors and promotoras, and the MCHC-certified diabetes educator met in a day-long work session to discuss local strengths, challenges, and educational needs. The group reached consensus on module subject areas, order, format, appropriate instructional styles, and community differences that might influence recruitment, retention, and program delivery. After this initial meeting, MEZACOPH investigators met separately with promotoras from each site in monthly meetings over an eight-month period to review draft instructional materials and to gain feedback on the approach, format, and translation of the developing curriculum. In these meetings, promotoras from both sites also provided examples of health education materials that had been well received in their communities.
Training
Once the curriculum was developed, a bilingual MEZACOPH investigator conducted a day-long training session in Spanish with CSF and MCHC promotoras. Since the promotoras had contributed to the development of many of the materials incorporated into the curriculum, the training offered an introduction to the overall flow and format of the curriculum and instruction in the use of educational materials. Promotoras gained familiarity and comfort with the curriculum during and after the training through playing roles, practicing delivery, and having coworkers critique their style.
Intervention
La Diabetes y La Unión Familiar is a 12-week program with 10 points of contact: three home visits, five educational sessions, and two celebratory events. Drawing on key concepts of Social Learning Theory, a theoretical model was developed to guide the program design process (Figure). Key concepts include the influence of the social environment in behavior change, the need for knowledge and skills to change behavior (behavioral capability), and the importance of building confidence in the ability to take action (self-efficacy). Intervention activities include teaching team-building and communication skills to build and reinforce intrafamily communication, collective esteem (so all family members accept and value the family as a group), and collective efficacy (which promotes confidence in the family's ability to make changes). Intervention activities also include providing information on food choices and physical activity so families can make informed choices. Key concepts related to diet, exercise, and family support are introduced and discussed through the use of pictorial flipcharts, educational games, food sampling and preparation, and low-level physical activities.
Family social behaviors such as cohesion, adaptation (exhibited by resiliency and problem solving), and support within the family influence family food choice and physical activity behaviors. The proposed outcome of targeting family social behaviors to change health behaviors is improved nutrient intake, activity level, and diabetes management or prevention for all family members. The program encourages family members to collectively set health-behavior goals, to overcome obstacles hindering healthy behaviors, and to develop a plan to sustain behavior changes.
Figure A family-based diabetes control and prevention program at the U.S.-Mexico border.
A logic model illustrates the family component of the diabetes control and prevention program at the U.S.-Mexico border. An arrow points from 'Intervention Activities' to the box 'Family Skills and Chacteristics' with the words 'Build' and 'Reinforce' to show how the intervention seeks to build and reinforce skills and characteristics as communication, collective esteem and collective efficacy (as described in text). Another arrow points from 'Intervention Activities' to the box 'Family Behaviors' with the word 'information' to show how providing information can impact family behaviors on food choices and physical activity. The arrow leading from 'Family Skills and Characteristics' to the arrow between 'Family Social Behavior' and 'Family Health Behaviors' shows how family heath behaviors are shaped by family social behaviors, skills, and characteristics. An arrow between 'Family Social Behaviors' (described in text) and 'Family Health Behaviors'show how social behaviors impact behaviors such as food choices and physical activity. An arrow leads from 'Family Health Behaviors'to 'Family Health Outcomes,' which include nutrient intake, activity level, and diabetes prevention or management (described in text).
Figure. A family-based diabetes control and prevention program at the U.S.-Mexico border.
The promotora instructional manual provides an overview of the goals and format of the program, a description of the objectives, and an outline of the activities and supplies needed for each point of contact. Each family is given a notebook that includes copies of the flipchart materials and pockets for handouts, recipes, and other memorabilia such as photographs and a graduation certificate.
To accommodate differences in promotora skills and community characteristics, curriculum delivery is flexible. Educational flipcharts and games are prescribed for specific modules, but foods/snacks, exercises, and supplemental activities (e.g., additional games and stories) are selected by the promotoras from appendices in the instructional manual.
The program sequence and content are described below:
Conocer a la Familia (Meet the Family) is the first home visit to explain the length and format of the program to the family, register interested family members (name, age and relationship to the family member with diabetes), gain informed consent for participation in evaluation activities, and administer a preintervention questionnaire to all participating family members aged 18 years and older. All consent and evaluation procedures were approved by the University of Arizona Institutional Review Board.
Bienvenidos! (Welcome!) is a kick-off event for all families.
Five weekly interactive educational modules are described below: Familias y Diabetes (Families and Diabetes) is a general introduction addressing diabetes risk factors, symptoms, and complications.
Ser Saludable (Being Healthy) addresses the relationship between physical activity, food choices, and diabetes control and prevention.
Crear Metas (Creating Goals) asks families to evaluate their own current health behaviors and provides steps for creating, reaching, and maintaining health-behavior goals.
La Unión Familiar (Working Together) encourages families to discuss their progress and goals and suggests ways they can enhance their success through family support and unity. This module expands the discussion of support by teaching all family members to recognize, avoid, and remedy low and high blood sugar.
Seguir Saludable (Staying Healthy) encourages families to continue to support one another in reaching goals and to create new goals as previous goals are obtained. In this module, family communication and support skills useful in recognizing and coping with the stress and depression that often accompany diabetes are discussed.
Felicidades! (Congratulations!) is a celebratory event for all families to acknowledge completion of educational modules.
Cómo Están? (How Are You?) is the second home visit and provides an opportunity for promotoras to meet with each family to discuss progress and challenges to family health goals.
Evaluación (Evaluation) is the third and final home visit and provides an opportunity for an outside evaluator to administer the postintervention questionnaire to all participating family members aged 18 years and older.
All training, intervention, and evaluation materials and activities were produced and delivered in Spanish.
Recruitment
Two CSF and two MCHC promotoras implemented the program at their respective sites. Promotoras contacted patients who had completed the Border Health ¡SI! patient education classes and extended an invitation to participate in a family diabetes education program. If a patient expressed interest, one or two promotoras made a home visit, scheduling a time when interested family members might be present. See previous description of this first meeting underConocer a la Familia (Meet the Family).
The term
family was defined by the patient and included spouses, children, parents, siblings, and friends. Even family members and friends not living in the same household with the patient could be identified as family if they had weekly contact with the patient. To avoid barriers created by child-care needs, no age limit was imposed, and even infants were permitted to attend. In the group-delivery format, child care was provided. Children aged less than 18 years participated in the intervention but were not included in assessment activities. Once registered, all family members were invited to the kick-off event (Bienvenidos!) to meet other families participating in the intervention to play games and to enjoy a healthy meal.
Delivery
Each agency chose a different delivery style for the five educational modules based on their experience with family education efforts in their communities. In Yuma County, two promotoras delivered the five sessions through a series of weekly home visits. In Santa Cruz County, two promotoras delivered the modules in weekly evening classes to five to 10 families in a group format at a central location; if requested during the initial home visit, promotoras provided transportation and child care. Home visits and celebratory events were implemented comparably at the two sites.
Retention
In the group format, families who missed an education session were called or visited by a promotora to determine if transportation, illness, lack of interest, or other barriers had prevented their attendance. The promotora would offer to help solve any problem, such as providing transportation or reassuring participants who felt uncomfortable with the format or information. In the home delivery, retention was not an issue. Families were consistently home at the prearranged time.
Assessment
Impact of the intervention was assessed using the following: 1) a written preintervention and postintervention Knowledge, Attitudes, Beliefs, and Behaviors (KABB) questionnaire with 15 close-ended items that documented adult participants' self-reported knowledge of risk factors, dietary and exercise habits, perception of need to eat healthy foods and be active, collective (family) exercise habits, and collective efficacy to make behavior changes; and 2) postgraduation interviews. Questions were designed to track changes in the stated learning objectives of each of the educational sessions. Promotoras reviewed the questionnaires for readability and comprehension. A team of university Border Health ¡SI! investigators reviewed the individual questions and entire questionnaire for content and face validity. The design of the questionnaire was driven by community and university collaborators' intent to develop a user- and administrator-friendly instrument that the promotoras would continue to use to document program impact.
The McNemar test for paired categorical data and the Wilcoxon signed rank test for paired continuous data were used to compare preintervention and postintervention responses of 116 adult participants (48 Santa Cruz County and 68 Yuma County).
In the initial visit, promotoras administered the preintervention questionnaire. In the final home visit, an evaluator not identified with the intervention administered the postintervention questionnaire. To accommodate variations in literacy levels and to avoid bias, all administrators were instructed to use a straightforward objective style when reading questionnaires aloud in Spanish. Participants were instructed to make their selection independently.
In the second home visit (Cómo Están?) ) two weeks after graduation, promotoras were impressed by participants' comments emphasizing the psychosocial importance of the program. To further explore an outcome not assessed by the preintervention and postintervention questionnaire, MEZACOPH investigators decided to conduct follow-up interviews with a sample of participants. One year after graduation, a sample of 18 participants consented to an open-ended interview with a MEZACOPH investigator not identified with the intervention. Given the small sample and an interest in documenting all experiences within this case study, participants' statements, as recorded in writing by the interviewer, were grouped by common themes.
Consequences
The following outcomes of the initial implementation of La Diabetes y La Unión Familiar are offered in support of the feasibility and potential impact of a promotora-delivered family-based diabetes education intervention.
Participation
Four rounds of La Diabetes y La Unión Familiar were implemented in each intervention county, yielding 72 patients with diabetes and 177 support people, including children and grandchildren. Table 1 illustrates the distribution of participants by sex and age group; figures include all patients with diabetes, family members, and supporters. Twenty-five percent of participants were children (younger than 18 years). Depending on their age and interest, children participated in the games, listened to stories, and participated in discussions generated by the flipcharts. Younger children and infants played or were cared for in a separate child-care area in the multifamily group sessions.
Adult daughters and wives were the predominant participating supporters. Percentage of supporters by relationship to the family member with diabetes was: 22% daughters, 20% spouses (with 54% wives and 46% husbands), 15% sons, and 9% friends. Of the total adult participants, 87% attended three or more of the five educational modules, and 43% of the youth and adults attended eight or more of the 10 points of contact.
Preintervention and Postintervention Knowledge, Attitudes, Beliefs, and Behaviors (KABB) Outcomes
Sixty-one (53% of total participants) of the preintervention and postintervention response pairs were from the family member with diabetes.
Table 2 provides the percent of yes responses to a list of possible diabetes risk factors listed in the questionnaire. A family history of diabetes (heredity) and being Hispanic, overweight, inactive, and older than 45 years were introduced in the first educational session (Familias y Diabetes) as known risk factors for diabetes. Based on their outreach experience, promotoras reported that some clients believed that stress, fear, and contact with an individual with diabetes were also risk factors. This first session provided an opportunity to discuss these perceptions. Preintervention and postintervention test comparisons indicate a significant increase in participants who identified the five known risk factors and a significant decrease in those who indicated yes to stress, fear, and contact after the intervention.
Table 3 provides a comparison of participants' preintervention and postintervention responses to family efficacy questions. Participants responded to the following questions on a five-point scale with 1 = not very confident and 5 = very confident: How confident are you that your family can become more physically active? How confident are you that your family can eat healthier? After the intervention, participants report a significant increase in their perception of their family's efficacy to make specific behavior changes.
Questions of food intake and activity were changed during the course of the program in response to questionnaire administrators' report that participants may be misinterpreting questions. The initial questions asked about general intake of specific foods and participation in specific activities. Questionnaire administrators indicated that respondents did not believe that they had regular food and activity behaviors and answered based on their behavior over the previous week. To create a questionnaire responsive to the administrators' observations and to provide an evaluation instrument that allowed participants to report behavior change in their own terms, these questions were changed after two rounds of administration from "in general" to "in the last week." Given the change, responses to these two different sets of questions were analyzed separately (separate data table not provided). Despite the difference in question wording, these separate data sets do reflect similar patterns of change, which are described below:
The frequency of sweetened drink consumption decreased significantly (P < .001 for both response to question on general intake as well as question on previous week). These drinks included fruit drinks distinctive within Hispanic culture (e.g. horchata, tamarindo, jamaica, Tampico™ as well as Gatorade™ and Sunny Delight™, but not carbonated soft drinks.
No consistent change was noted in reported fruits, vegetables, soft drinks, or low- and nonfat milk consumption.
A nonsignificant trend in respondents reportedly exercising five or more times per week for 30 minutes or more.
A significant increase in family members participating together in a physical activity (P = .002).
A significant increase in participants reporting that family members help and support each other (P = .01).
A nonsignificant trend toward greater communication and cohesive behaviors, such as talking about food choices, going to the doctor with the family member with diabetes, and agreeing to eat out or buy food from places with healthy choices.
Changes in knowledge, attitudes, behaviors, and beliefs were not different in a comparison of family members with diabetes and family members without diabetes.
Interviews
Eighteen individuals (both patients and family members) from both sites were interviewed individually for approximately one hour. All statements could be grouped into one of three themes:
Program participation had a positive psychosocial impact on participants. Those with diabetes and family members reported feeling better and being less depressed and isolated.
Family communication, particularly about food choices and understanding of depression, improved; communication was more frequent and/or less emotional.
The social interaction provided by the promotoras was the best part of the program.
No other follow-up data were collected to assess long-term behavior change or retention of knowledge.
Interpretation
La Diabetes y La Unión Familiar, a Spanish language family diabetes education intervention that targets family support, communication, and family health behaviors, implemented by promotoras in two Arizona border communities, yielded changes in family members' knowledge, attitudes, behaviors, and beliefs relative to diabetes prevention and control.
The development of the program content, delivery format, and even evaluation methods was a collaborative process among a university, MEZACOPH, and two community health agencies, CSF and MCHC. The description of the collaborative process illustrates how standard research practices and community experience, observation, and interests contributed to the final intervention. Program outcomes demonstrate that teaching the family as a group can influence health behaviors, yielding an increase in family-based physical activity and ameliorating family member feelings of depression and isolation. This community case study supports the use of a family-based approach to diabetes prevention and control. This study indicates that family involvement should go beyond diabetes support groups that tend to focus only on the behaviors of the person with diabetes. Addressing the family's collective behaviors as well as patterns of cohesion and communication can yield change in the family environment, an important influence in chronic disease management and prevention.
Certain preexisting factors and limiting conditions of this case study should be acknowledged. The promotoras had previous experience and training in community outreach services. They were uniquely familiar with the curriculum as they collaborated in its development and adapted its delivery for their communities' needs. The small sample size in this case study limits the authors' ability to project the applicability and impact of La Diabetes y La Unión Familiar in other communities. Yet these results are promising and warrant continued implementation of the program in these counties and piloting in similar communities. The program is available to other agencies by accessing its Web site (available from: http://www.borderhealthsi.org/). Furthermore, the preintervention and postintervention evaluation instrument did not capture the psychosocial impact of the program as revealed by promotoras' observations and a small number of interviews. Future implementation should consider revising the evaluation instrument or supplementing evaluation activities with a formal guided interview conducted by an evaluator not identified with the intervention.
Figures and Tables
Table 1 Participation in Family-based Diabetes Program From Two Intervention Sites in Yuma and Santa Cruz Counties, Arizona
Male Female Total
Adults (≥18 years) 49 (27%) 135 (73%) 184 (75%)
Youth (<18 years) 36 (55%) 29 (45%) 65 (25%)
Total 85 (34%) 164 (66%) 249 (100%)
Table 2 Adult Participants’ (N = 116) Preintervention and Postintervention Responses to the Knowledge, Attitude, Beliefs, and Behaviors (KABB) Questionnaire: Knowledge of Risk Factors, Yuma and Santa Cruz Counties, Arizona
Yes n (%)
Diabetes risk factorsa Preintervention Postintervention Pb
Heredity (family history) 86 (74.1) 104 (89.7) .002
Hispanic 39 (33.6) 94 (81.0) <.001
Overweight 85 (73.3) 106 (91.4) .001
Inactive 69 (59.5) 109 (94.0) <.001
>45 years of age 51 (44.0) 102 (87.9) <.001
Stress 68 (58.6) 46 (39.7) .002
Fear 81 (61.8) 50 (38.2) <.001
Contact 3 (75.0) 1 (0.9) .006
a Participants were asked to answer yes if they believed that the characteristic put them at greater risk for diabetes and to answer no if they did not believe that the characteristic put them at greater risk.
b Determined by two-tailed test.
Table 3 Adult Participants’ (N = 116) Preintervention and Postintervention Responses to the Knowledge, Attitude, Beliefs, and Behaviors (KABB) Questionnaire: Family Efficacy, Yuma and Santa Cruz Counties, Arizona
Confidence in ability of family to change behavior Mean ± SD Pa
Preintervention Postintervention
To eat healthier foods 3.42 ± 1.22 4.13 ± 1.19 <.001
To be more physically active 3.46 ± 1.20 4.00 ± 1.20 <.001
a Determined by two-tailed test.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Teufel-Shone NI, Drummond R, Rawiel U. Developing and adapting a family-based diabetes program at the U.S.-Mexico border. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0083.htm
==== Refs
1 National Institute of Diabetes and Digestive and Kidney Disease Diabetes in Hispanic Americans [Internet] Accessed 2004 Oct 18 Bethesda (MD) The Institute, National Diabetes Information Clearinghouse
2 Harris MI Klein R Cowie CC Rowland M Byrd-Holt DD 21 8 1998 1230 1235 Diabetes Care Is the risk of diabetic retinopathy greater in non-Hispanic blacks and Mexican Americans than in non-Hispanic whites with type 2 diabetes. A U.S. population study 9702425
3 Harris MI Cowie CC Eastman RC 16 11 1993 1446 1452 Diabetes Care Symptoms sensory neuropathy in adults with NIDDM in the US population 8299433
4 von Goeler DS Rosal MC Ockene JK Scavron J De Torrijos F 29 4 2003 663 672 Diabetes Educator Self-management of type 2 diabetes; a survey of low-income urban Puerto Ricans 13677177
5 Stern MP Mitchell BD National Diabetes Data Group 1995 631 659 Diabetes in America 2nd ed Bethesda (MD) National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health Diabetes in Hispanic Americans NIH Publication No. 95-1468
6 Juarbe T 20 6 1998 765 782 West J Nurs Res Cardiovascular disease-related diet and exercise experiences of immigrant Mexican women 9842291
7 Lipton RB Losey LM Giachello A Mendez J Girotti MH 24 1 1998 67 71 Diabetes Educ Attitudes and issues in treating Latino patients with type 2 diabetes: views of healthcare providers 9526327
8 Khan A Lasker SS Chowdhury TA 26 3 2003 710 712 Diabetes Care Are spouses of patients with type 2 diabetes at increased risk of developing diabetes? 12610026
9 Brown SA Garcia AA Winchell M 2 2 2002 166 176 Curr Diab Rep Reaching underserved populations and cultural competence in diabetes education 12643136
10 Anderson RM Goddard CE Garcia R Guzman JR Vazquez F 24 5 1998 618 625 Diabetes Educ Using focus groups to identify diabetes care and education issues for Latinos with diabetes 9830959
11 Hunt LM Valenzuela MA Pugh JA 46 7 1998 959 969 Soc Sci Med Porque me toco a mi? Mexican American diabetes patients' casual stories and their relationship to treatment behaviors 9579748
12 Schafer JC McCaul KD Glasgow RE 9 2 1986 179 185 Diabetes Care Supportive and non-supportive family behavior: relationships to adherence and metabolic control in persons with type 1 diabetes 3698784
13 Trief PM Himes CL Orendorff R Weinstock RS 24 8 2001 1384 1389 Diabetes Care The marital relationship and psychosocial adaptation and glycemic control of individual with diabetes 11473074
14 Trief PM Grant W Elbert K Weinstock RS 21 2 1998 241 245 Diabetes Care Family environment, glycemic control, and the psychosocial adaptation of adults with diabetes 9539989
15 Fisher L Chesla CA Skaff MM Gilliss C Mullan JT Bartz RJ 23 2 2000 267 272 Diabetes Care The family and disease management in Hispanic and European-American patients with type 2 diabetes 10868849
16 Chesla CA Skaff MM Bartz RJ Mullan JT Fisher L 23 12 2000 1780 1785 Diabetes Care Differences in personal models among Latinos and European Americans 11128352
17 Rodríguez-Morán M Guerrero-Romero JF 39 1 1997 44 47 Salud Publica Mex Importancia del apoyo familiar en el control de la glucemia. [Importance of family support in the control of glycemia] 9092097
18 Valadez-Figueroa I Aldrete-Rodríguez MG Alfaro-Alfaro N 35 5 1993 464 470 Salud Publica Mex Influencia de la familia en el control metabolico del paciente diabético tipo II. [Family influence on metabolic control of patients with type II Diabetes] 8235892
19 Valenciaga Rodríguez J.L.G.d.l.V. F Pons Bravet P Sanchez Valdes O 11 2 1995 144 149 Rev Cuba Med Gen Integr Repercusion en el control de la diabetes mellitus de una técnica educativa dirigida a la familia. [Repercussion of an educational technique aimed to the family for the control of diabetes mellitus.]
20 Aráuz AS Padilla G Fernandez G Rosello M Guzman S 9 3 3 2001 145 153 Rev Panam Salud Publica Intervención educativa comunitaria sobre la diabetes en el ambito de la atencion primaria. [Community diabetes educational intervention at the primary care level] 11349349
21 Foreyt JP Ramirez AG Cousins JH 53 6 Suppl 6 1991 1639S 1641S Am J Clin Nutr Cuidando El Corazon: a weight-reduction intervention for Mexican Americans 2031499
22 Cousins JH Rubovits DS Dunn JK Reeves RS Ramirez AG Foreyt JP 107 5 Sep–Oct 1992 549 555 Public Health Rep Family versus individually oriented intervention for weight loss in Mexican American women 1410236
23 Brown SA Hanis CL 25 2 Mar–Apr 1999 226 236 Diabetes Educ Culturally competent diabetes education for Mexican Americans: the Star County Study 10531848
24 Brown SA Hanis CL 21 3 1995 203 210 Diabetes Educ A community-based, culturally sensitive education and group-support intervention for Mexican Americans with NIDDM: a pilot study of efficacy 7758387
25 Cohen SJ Ingram M Prev Chronic Dis [serial online] Border health strategic initiative: overview and introduction to a community-based model for diabetes prevention and control 2005 1
26 U.S. Census Bureau State and county quick facts. 2000 census of population and housing, small area income and poverty estimates U.S. Census Bureau Washington (DC) Accessed 2004 Oct 18
27 Centers for Disease Control and Prevention Behavioral Risk Factor Surveillance System, 2000 Atlanta (GA) National Center for Chronic Disease Prevention and Health Promotion
|
15670473
|
PMC1323323
|
CC BY
|
2022-01-25 23:20:14
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A20
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0059
Community Case Study
PEER REVIEWEDThe Cancer Prevention and Control Research Network
Harris Jeffrey R MD, MPH, MBA Health Promotion Research Center, University of Washington Health Promotion Research Center
1107 NE 45th St, Suite 200, Seattle, WA 98105 [email protected]
206-616-8113
Brown Pamela K MPA Mary Babb Randolph Cancer Center, West Virginia University, Morgantown, WV
Steven Coughlin PhD Division of Cancer Prevention and Control, National Center for Chronic Disease Prevention and Health Promotion (NCCDPHP), Centers for Disease Control and Prevention (CDC), Atlanta, Ga
Wilson Katherine PhD Division of Cancer Prevention and Control, National Center for Chronic Disease Prevention and Health Promotion (NCCDPHP), Centers for Disease Control and Prevention (CDC), Atlanta, Ga
Fernandez Maria E PhD Center for Health Promotion and Prevention Research, University of Texas Health Science Center at Houston School of Public Health, Houston, Tex
Hebert James R ScD University of South Carolina, Columbia, the Hollings Cancer Center at the Medical University of South Carolina, and the Palmetto Health South Carolina Cancer Center, Columbia, SC
Kerner Jon PhD Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, Mda
Prout Marianne MD, MPH Department of Epidemiology, Boston University School of Public Health, Boston, Mass
Schwartz Randy MSPH New England Division, American Cancer Society, Boston, Mass
Simoes Eduardo J MD, MSc, MPH Division of Adult and Community Health, NCCDPHP, CDC, Atlanta, Ga
White Carol MPH University of Kentucky Center for Prevention Research, Lexington, Ky
1 2005
15 12 2004
2 1 A212005
The Cancer Prevention and Control Research Network is a national network recently established to focus on developing new interventions and disseminating and translating proven interventions into practice to reduce cancer burden and disparities, especially among minority and medically underserved populations. Jointly funded by the Centers for Disease Control and Prevention and the National Cancer Institute, the Cancer Prevention and Control Research Network consists of sites administered through Prevention Research Centers funded by the Centers for Disease Control and Prevention. The five sites are located in Kentucky, Massachusetts, South Carolina, Texas, Washington State, and West Virginia. The Cancer Prevention and Control Research Network's intervention areas include primary prevention of cancer through healthy eating, physical activity, sun avoidance, tobacco control, and early detection of cancer through screening. The Cancer Prevention and Control Research Network uses the methods of community-based participatory research and seeks to build on the cancer-relevant systematic reviews of the Guide to Community Preventive Services. Initial foci for the Cancer Prevention and Control Research Network's research work groups include projects to increase screening for breast, cervical, and colorectal cancers; to promote informed decision making for prostate cancer screening; and to validate educational materials developed for low-literacy populations.
==== Body
Background
The Cancer Prevention and Control Research Network (CPCRN) is a federally funded, national network of academic, public health, and community partnerships that work together to reduce the burden of cancer, especially among those disproportionately affected. The CPCRN was initiated in October 2002, with funding from the Centers for Disease Control and Prevention (CDC) and the National Cancer Institute (NCI) as part of their effort to more effectively translate research into practice. The five CPCRN sites were selected through a competition among the CDC-funded Prevention Research Centers (PRCs). Three sites are operated by individual universities: the Universities of South Carolina, Texas-Houston, and Washington. Two sites are operated jointly by pairs of universities: Boston and Harvard Universities; and the University of Kentucky and West Virginia University. This paper introduces the CPCRN; outlines the context for its creation, along with its goals, structure, and operations; and summarizes progress to date.
Context
Although the CPCRN sites carry out most of their work locally, the CPCRN is a national network and was developed in a national context. The CPCRN is a further step in efforts by two federal agencies, the CDC and the NCI, to translate research into practice with potential for reducing the cancer burden in the United States, especially among populations that are disproportionately affected. The context for the CPCRN, discussed in detail below, consists of four factors. First, the cancer burden in the United States remains high, and disparities in incidence and mortality persist. Second, one of the best opportunities to reduce these disparities is through community-based participatory research. Third, recently published syntheses of research, such as the CDC's Guide to Community Preventive Services (Community Guide) (1), suggest specific areas where carefully evaluated dissemination research is needed. Finally, the CDC's PRC Program (2), with its focus on community-based participatory research (CBPR) and translation, provides a unique combination of trained, experienced investigators and infrastructure to support a network like the CPCRN.
Cancer burden in the United States
The creation of the CPCRN is, in part, a response to the growing magnitude of and persistent disparities in cancer burden. Cancer is the second leading cause of death in the United States as well as a leading cause of morbidity. Cancer accounts for one of every four deaths (3), and, in 2004, 563,700 people are expected to die of cancer. In 2004, 1,368,030 new cancer cases are expected to be diagnosed in the United States (3), not including carcinoma in situ or basal and squamous cell skin cancers. The top four cancer sites (with expected numbers of cases) are prostate (230,110), breast (217,440), lung (173,770), and colorectal (150,950). In addition, more than 1 million cases of basal and squamous cell skin cancer are expected to be diagnosed in 2004
Disparities in cancer incidence and mortality persist. For example, the incidence of cervical cancer is four times as high among Vietnamese women as it is among other Asian American and Pacific Islander women (3). Overall, cancer mortality among African American men is 1.4 times higher than among whites, and cancer mortality among African American women is 1.2 times higher than among whites (4). Additional disparities in cancer incidence and mortality rates across major racial and ethnic groups in the United States are highlighted in recent reports (3,5) and below in descriptions of the CPCRN sites.
Those disparities are the result of a complex array of economic, social, and cultural factors, and these factors are also reflected in disparities in preventive behaviors. For example, smoking prevalence is now highest among American Indian (38.6%) and Alaska Native (27.4%) men and women (3,5). Screening prevalence for colorectal cancer is the lowest among Hispanics and Latinos (3,5). To optimize the effect of cancer prevention efforts relative to expenditure, we need to be clear about 1) the efficiency of intervening on known risk factors early in the natural history of the carcinogenic process (e.g., reducing use of tobacco products, improving access to fresh fruits and vegetables), 2) the utility of various preventive services (e.g., screening), and 3) the willingness of communities to be engaged in cancer prevention and control.
The importance of community-based participatory research
The CPCRN embraces the principles of CBPR (6) as core values. Research and evaluation developed with communities in a participatory way are more likely to reflect the needs, interests, and values of the community. Also, after research funding has ended, the results from such research and evaluation are more likely to be widely disseminated and the interventions to be sustained (7). The CPCRN has made a commitment in each of its sites to implement activities in a manner that is community-based and participatory to strengthen local resources and to build the capacity of community organizations to conduct and translate research. The CPCRN's commitment to CBPR is consistent with the strong and growing commitment of funding agencies to support this type of research partnership (8).
The need for research dissemination and translation
The need for the CPCRN is highlighted by recent reports of both progress against cancer and remaining challenges in disseminating and translating knowledge gained from efficacy and effectiveness research. During the late 1990s, death rates from the four leading cancers — lung, colorectal, breast, and prostate — declined nationally and in most states (5). Two important prevention strategies have contributed to this decline but remain underused: 1) primary prevention by reducing risk behaviors and 2) early detection by increasing the use of screening services (3).
Both the NCI and the CDC have given high priority to bridging the dissemination gap. The NCI's Translating Research into Improved Outcomes program has identified the dearth of dissemination research as a key impediment to the adoption of evidence-based cancer control intervention. The program has also identified the need to expand research/practice partnerships as critical to both the adoption and evaluation of evidence-based interventions in public health and clinical practice settings (9).
The Community Guide, developed under the aegis of an independent, nonfederal Task Force on Community Preventive Services and maintained at the CDC, has provided a partial summary of the state of the art of community-based cancer prevention and control. Based on a systematic review of the literature, the Community Guide currently recommends 14 interventions aimed at increasing physical activity, reducing exposure to and use of tobacco, and reducing exposure to ultraviolet light (Table 1) (10-13). In addition to these recommendations, the Task Force has recently completed recommendations for increasing informed decision making regarding cancer screening (14) and plans to publish its reviews of interventions to increase cancer screening in 2005.
The CDC has a ready dissemination outlet for proven intervention strategies through its state-based cancer prevention and control programs: the National Breast and Cervical Cancer Early Detection Program and the Comprehensive Cancer Control Program.
The CDC Prevention Research Centers program
The prior investment of the CDC in its PRC program facilitated the creation of the CPCRN by providing a well-established administrative home for each of the five sites. In 1984, Congress created the PRC program at the CDC by authorizing the funding of academic health centers for innovative research and demonstration projects to prevent chronic disease. In 1986, the CDC established three PRCs for two years. Since then, the funding period has increased to five years, and the number of centers has grown to 28, which are located in 25 states.
Three aspects of the PRC program strengthen the ability of the participating CPCRN sites to design, implement, and evaluate cancer prevention research with immediate application to public health practice. First, the PRC peer-review mechanism involves researchers experienced with CBPR. The peer review process provides assurance that applications are scientifically sound and that the research conducted is of practical use to communities. Second, the program's focus on CBPR enhances the likelihood that the PRCs will evaluate acceptable and sustainable community-based interventions. Partnerships and collaborations of PRCs with various entities (e.g., businesses, community coalitions, grassroots organizations, private health care providers, state and local public health agencies, voluntary health organizations) increase the likelihood that PRCs will produce evaluations of interventions that are likely to be translated and sustained. Third, research conducted among the most disadvantaged and underserved populations in the nation provides the PRCs with the opportunity to evaluate the external validity of interventions among diverse populations, including the rural poor of Appalachia, African Americans in South Carolina, public housing residents in Boston, residents of the U.S.-Mexico border, and loggers and pulp-mill workers in Washington State.
Methods
The overall goal for the CPCRN is to conduct community-based cancer prevention and control intervention and dissemination research that extends the knowledge base, addresses critical gaps, and leads to adoption, replication, implementation, and dissemination of successful programs in communities.
The CPCRN addresses gaps and builds on recommendations in the Community Guide by conducting site-specific and multisite intervention and dissemination research. The four specific research areas include 1) research on the effectiveness of community-based interventions for which evidence is insufficient to justify a Community Guide recommendation; 2) research replicating Community Guide-recommended interventions in populations and settings where they have not been adequately evaluated; 3) research on how to disseminate and implement Community Guide-recommended interventions into communities by public health and community-based organizations; and 4) evaluation of community programs to determine their effectiveness.
The initial funding of the CPCRN for a two-year period supported the development of both local network centers and the larger national network of the CPCRN sites. At both levels, expected outcomes are evidence of the networks' existence and viability, including mission and vision statements, short- and long-term objectives, and active working groups. Of special emphasis in the research arena are efforts to develop strong partnerships with communities bearing the greatest burden from cancer, where community-based participatory research projects are likely to contribute to the reduction and/or the elimination of disparities in cancer burden.
Cancer Prevention and Control Research Network structure and operations
The CPCRN is a national network of five sites, each of which is a local network of academic, community, and other organizations with an interest in cancer prevention and research. The work of the CPCRN is led by a Coordinating Center that organizes, among other things, cross-site work groups on research topics of mutual interest to the sites, the CDC, and the NCI. The six sections that follow describe the work of the Coordinating Center and of each of the five local sites.
Cancer Prevention and Control Research Network Coordinating Center
The University of Kentucky and West Virginia University share the coordinating center role for the CPCRN. The Coordinating Center guides discussions on developing research tools for community interventions in cancer prevention and control, organizes collaborative activities with CPCRN members and their partners, fosters relationships among CPCRN members and national/state/local partners to ensure that CPCRN objectives are being achieved, and provides leadership in developing and managing the CPCRN operational structure.
Specific activities include 1) developing and implementing a plan and system for effective communication among CPCRN centers; 2) implementing a collaborative planning process resulting in a seven-year plan for CPCRN research, dissemination, and evaluation; 3) implementing processes and procedures for encouraging PRCs to develop collaborative cancer prevention and control research projects; and 4) ensuring that external evaluation is conducted and is focused on the Coordinating Center's performance.
Alliance for Reducing Cancer, Northwest
The site based at the University of Washington is the Alliance for Reducing Cancer, Northwest (ARC NW; available from: URL: http://www.arcnw.org). The mission of the ARC NW is to increase primary-preventive and early-detection behaviors to prevent and control cancer in the Puget Sound region, Washington State, and the Pacific Northwest. ARC NW is a collaborative effort among the University of Washington PRC, the American Cancer Society Great West Division; Fred Hutchinson Cancer Research Center; Group Health Cooperative of Puget Sound, a health maintenance organization; Public Health Seattle and King County, a local health department; the Puget Sound Neighborhood Health Centers, an organization of several community health centers in the region; Qualis Health, the Medicare quality improvement organization for Alaska, Idaho, and Washington State; the Washington State Department of Health; and the Weyerhaeuser Company, a large timber products company.
Data from the Washington State Behavioral Risk Factor Surveillance System reveal underuse of primary-preventive and early-detection behaviors. In 2002, 21% of Washingtonians aged 18 or older smoked, 15% were physically inactive during leisure time, 76% ate inadequate quantities of fruits and vegetables, and 60% were overweight or obese. Also in 2002, among appropriate age groups, 45% had never received a flexible sigmoidoscopy or colonoscopy, 40% had never received a fecal occult blood test, 26% had not received a mammogram within two years, and 13% had not received a Papanicolaou (Pap) test within three years (15).
One important factor in the underuse of these behaviors is the lack of support for prevention at the worksite and in employer-based health insurance. At the worksite, employers of all sizes reported in a 2001 national survey the following offerings: 11% offered fitness services and 5% offered tobacco-cessation services (personal communication, Maris Bondi, Partnership for Prevention, November 2003). Employers of all sizes nationwide reported in the same survey the following health insurance offerings: 80% covered mammograms, 79% covered Pap smears, 68% covered colorectal cancer screening, and 10% covered smoking cessation treatment that included both prescription medications and counseling.
The ARC NW focuses on employed populations and on underserved communities. Five current activities include 1) a pilot test, involving the Weyerhaeuser Company, of a policy intervention to promote primary prevention and early detection via the worksite and employment-based health insurance; 2) development of a work site-based, team-oriented intervention to promote primary prevention and early detection of cancer; 3) a pilot test of a tool to increase informed decision making regarding prostate cancer screening; 4) assistance to the Washington State Department of Health in designing and evaluating its colorectal and prostate cancer screening programs; and 5) a review of the literature regarding the quality of life after treatment of prostate cancer.
Appalachian Cancer Research Consortium
The site based at the University of Kentucky and West Virginia University is the Appalachian Cancer Research Consortium (ACRC). The target population of the ACRC includes the poor, medically underserved, and primarily rural residents of West Virginia and the 51 counties in Appalachian Kentucky. The two universities have a long history of collaboration, with extensive experience in working with communities throughout Appalachia on critical health issues.
The U.S. Department of Health and Human Services considers the rural residents of Appalachia a "special population" (16). These residents tend to be older, poorer, less educated, and more likely to be uninsured than urban Americans. Rural communities have higher rates of chronic illness and disability and report poorer overall health status than urban communities (16). Residents of rural areas generally have fewer visits with physicians and lower levels of preventive care. In addition to factors related to rural health status and practices, there are systemic factors related to rural life that may contribute to less than optimal preventive care (17). These factors include lack of public transportation, lack of health care providers, and lower levels of community services.
As a result, West Virginia and the Appalachian regions of Kentucky have higher total cancer mortality rates than the national average (18,19). Both states rank among the top 10 U.S. states for total, male, and female cancer mortality. Lung cancer is a significant problem for residents, accounting for approximately 30% of all cancer deaths in West Virginia and Kentucky and resulting in a higher lung cancer mortality rate than the U.S. rate. Kentucky and West Virginia have invasive cervical cancer incidence and mortality rates that are significantly higher than the U.S. rates. West Virginia and Appalachian Kentucky also have higher colorectal cancer mortality rates than the United States and Appalachia as a whole. Breast cancer mortality rates are similar to national rates, but breast cancer mortality in several rural counties exceeds the national rate by more than 50%.
The ACRC focuses its efforts primarily on four cancer sites — lung, cervix, colorectal, and breast — with high disease burden, high behavioral risks, and high importance to community members in the region. Current activities of the ACRC include 1) developing a standardized assessment tool to evaluate readability, format, illustrations, and content of cancer prevention and control materials; 2) developing a protocol for colorectal cancer intervention for men and women aged 50 and older; and 3) conducting work site focus groups to identify barriers to colorectal screening for public employees aged 50 and older.
Latinos in a Network for Cancer Control
The site based at the University of Texas (UT) is Latinos in a Network for Cancer Control (LINCC; available from: URL: http://www.sph.uth.tmc.edu/research/lincc). The mission of the LINCC is to reduce cancer-related health disparities among Hispanics/Latinos through community-based intervention, replication, and dissemination research. LINCC is a collaboration among 1) academic researchers at the UT School of Public Health, the UT M.D. Anderson Cancer Center, and the Baylor College of Medicine; 2) cancer control organizations, including the American Cancer Society, Cancer Information Service, Sanchez Cancer Center, Texas Cancer Council, Texas Comprehensive Cancer Coalition, and the Texas Department of Health; and 3) community-based organizations, including the Center for Border Health Research, Hispanic Health Coalition, Migrant Health Promotion, the National Center for Farmworker Health, and the Racial and Ethnic Approaches to Community Health coalition.
Hispanics/Latinos in Texas account for approximately 25% of the total U.S. Hispanic population and 32% of the total Texas population (20). Along the U.S.-Mexico border where LINCC has focused its initial research efforts, Hispanics comprise roughly 80% of the population (21). Many border residents experience high rates of poverty and live in colonias, unincorporated areas where environmental pollution, inadequate wastewater systems, and inadequate access to public drinking water compound socioeconomic influences on health behavior.
Hispanics in the United States experience higher incidence rates of cervical cancer per 100,000 (16.3) compared with non-Hispanics (7.8) and higher rates of mortality per 100,000 (3.7 compared with 2.6) (22). Along the U.S.-Mexico border, the disparity is even greater: the incidence rate of cervical cancer per 100,000 among Hispanics (18.7) is higher than the rate among non-Hispanics (8.2), and the mortality rate among Hispanics (6.2) is higher than the rate among non-Hispanics (3.4) (23). In addition, Hispanics have lower rates of cancer screening. Only 27% of the older Hispanic adults in Texas reported having a recent fecal occult blood test for colorectal cancer (compared with 34% among non-Hispanic whites), and only 50% reported regular mammography use (compared with 60% for non-Hispanic whites) (24). Use of Pap tests for cervical cancer screening among Hispanics (83%) was also lower compared with non-Hispanic whites (87%) (22).
Current LINCC activities include 1) new research on factors influencing colorectal cancer screening among Hispanics and the development of a community-based intervention to increase this screening; 2) research on informed decision making for prostate and colorectal cancer screening; 3) an evidence review and new research on lay health-worker- (promotora-) based interventions for increasing cancer screening; and 4) research on the effectiveness of small media interventions to increase cancer screening. Another major focus of LINNC is to identify important factors and effective strategies for replicating and disseminating effective cancer control interventions in Hispanic communities. To this end, LINCC is conducting research on the replication and dissemination of an evidence-based, effective breast and cervical cancer screening intervention for Hispanic women: Cultivando la Salud (Cultivating Health) (25).
Massachusetts Cancer Prevention Community Research Network
The site based at Boston and Harvard Universities is the Massachusetts Cancer Prevention Community Research Network (MCPCRN). The MCPCRN's mission is to foster a network of partnerships among cancer prevention researchers and community collaborators to support CBPR and to reduce social disparities in cancer risk. The MCPCRN is a collaboration of the Dana-Farber/Harvard Cancer Center Risk Reduction Program, the Harvard Prevention Research Center (HPRC), and the Boston University Prevention Research Center, with participation from the American Cancer Society's New England Division and the Massachusetts Cancer Control Coalition.
Massachusetts has 6.5 million residents, 82% of whom are non-Hispanic white (26). In Boston, however, because of recent immigration, non-Hispanic whites are no longer in the majority (27). The total cancer incidence rate per 100,000 in Massachusetts (501.2) is higher than the national rate (468.9) in the Surveillance, Epidemiology, and End Results Program, and so are the incidence rates for prostate, breast, lung, and colorectal cancers (4,28). The total cancer mortality rate per 100,000 in Massachusetts (211.3) is just slightly higher than the U.S. rate (206.0) (4,28); and colorectal and breast cancers are the major contributors with higher mortality rates. Smoking rates have fallen to less than 19%; 20.8% of the population is sedentary, and 54.4% is overweight or obese (15).
To reduce these risks, the MCPCRN is approaching four priority community sectors: 1) schools and youth, 2) work sites and labor unions, 3) health care providers, 4) and low-income housing. Among schools and youth, the HPRC faculty direct a range of school and community-based research to improve youth nutrition and physical activity. Work sites and union partners include the Massachusetts AFL-CIO (American Federation of Labor-Congress of Industrial Organizations), the Massachusetts Coalition on Occupational Safety and Health, and individual local unions. Approximately 50 community health centers, many with strong ties to MCPCRN partners, facilitate access to health care providers in Massachusetts. The cost of housing in Massachusetts ranks third nationally; MCPCRN collaborators have identified more than 100 housing developments in Boston, Cambridge, and Somerville as potential partners.
The MCPCRN's current objectives are to strengthen ties with communities and to conduct pilot and developmental studies as a foundation for future research. An upcoming conference on CBPR will emphasize engaging community organizations in cancer prevention research opportunities. Collaborative community efforts support Health Ambassadors for African American and African immigrant women in Boston and train tobacco advocates in housing developments. Developmental research includes a work site protocol to increase informed decision making for prostate cancer screening; materials to promote timely follow-up for abnormal mammograms among low-income, ethnic minority women; and methods to improve decision making on colorectal cancer screening. Pilot studies include an intervention aimed at weight reduction and increased physical activity through the Young Men's Christian Association and data collection in low-income housing developments.
South Carolina Cancer Prevention and Control Research Network
The site based at the University of South Carolina is the South Carolina Cancer Prevention and Control Research Network (SCCRN). The SCCRN was created to address the large and growing cancer burden among African Americans living in South Carolina. Its aim is to serve the entire state, with a population of just more than 4 million people, comprising an area of 31,000 square miles, and ranging from a long, broad coastal plain to the Piedmont region of southern Appalachia. The SCCRN builds on a strong network of existing programs that have coalesced recently in the South Carolina Cancer Alliance (SCCA), which consists of more than 750 institutional and individual members. The constituent bodies of the SCCA include the South Carolina Department of Health and Environmental Control and numerous grassroots organizations in addition to all academic, clinical care, and nongovernmental organizations with cancer-related missions.
South Carolina is a relatively rural state, with very high (>40%) African American representation in rural areas. It is also a poor state, where the average personal income is about 81% of the national average (29). Cancer rates of African Americans, who represent 31% of South Carolina's total population, diverge from the U.S. average, in many instances markedly (4,30). For example, prostate cancer incidence among African American men in South Carolina is more than 70% higher than in white men, whereas the difference is 55% nationally (30). Nearly all cancers have higher mortality in South Carolina than in the United States as a whole (4). Illustrative of the pattern of increased mortality, breast cancer incidence in South Carolinian African American women is 18% lower than the incidence in white women (as opposed to being 15% lower nationally), but mortality is 47% higher (vs the national differential of 32%) (4,30).
Research at the SCCRN focuses on investigating ways to implement programs that complement existing cancer prevention and control infrastructure and through which we can anticipate risk reduction based on changes in individual and organizational behavior. The SCCRN focuses on breast, cervix, colorectal, oropharyngeal, prostate, and thoracic cancers. Ongoing projects include 1) investigation of small media approaches to increase breast and cervical cancer screening in low-income, rural women at highest risk of aggressive forms of these cancers, 2) research on informed decision making for prostate cancer screening and treatment, and 3) identification of geographical determinants of prostate cancer. Formative work includes 1) exploration of a community-based, statewide program of research in oral cancer precancerous lesions, 2) development of a mammography registry to understand patterns of use in low-income, predominantly African American populations, and 3) a church-based participatory intervention of lay health advocate-delivered cancer education and referral.
Consequences
In its first year of operation, the CPCRN has focused on a strategic planning process. From the beginning, community partners from each of the five sites have played strong and active roles in these processes. The strategic planning process produced vision and mission statements; a set of operating structures, principles, and plans; and four research work groups (see below).
Vision statement
Communities and researchers working together to reduce the burden of cancer, especially among those disproportionately affected.
Mission statement
The mission of the CPCRN is to conduct cancer prevention and control research that extends the knowledge base, addresses critical gaps, and leads to adoption, replication, implementation, diffusion, and evaluation of successful programs in communities.
Operating structure, principles, and plans
The CPCRN has developed a governing structure with a steering committee, guiding principles, and a seven-year strategic plan. Further information on each of these documents is available from: URL: http://ukprc.uky.edu/CPCRN/home.htm.
Research work groups
The strategic planning process also suggested the development of work groups to initiate and carry out CPCRN research. The CPCRN currently has work groups focusing on 1) screening for breast and cervical cancers; 2) screening for colorectal cancer; 3) informed decision making and quality-of-life issues for prostate cancer screening and treatment; and 4) validating low-literacy educational and media materials. The work groups involve cross-site collaboration among scientists from the network centers and have established research goals (Table 2).
Interpretation
The CPCRN represents a new and innovative approach for addressing the challenge of identifying effective interventions and promoting dissemination and adoption of these interventions into communities. The CPCRN sites are geographically distributed across the nation, enhancing opportunities to develop community partnerships and to conduct community-based assessments, evaluation, and research with populations that represent nearly all types of medically underserved racial and ethnic groups in the continental United States. A strong commitment to CBPR increases the likelihood that CPCRN research will benefit the underserved communities in greatest need. The CPCRN also provides an opportunity for the sites to collaborate in addressing research gaps, including dissemination research and research translation, and to build on recommendations provided in the Community Guide. Finally, the location of the CPCRN within the CDC's PRC program enables its research findings to be easily translated, both nationally and locally, through long-existing partnerships with other prevention organizations.
Fostering the optimal results from the CPCRN will require that its members maintain a delicate balance between coordinated, centralized efforts and retaining and enhancing the critical, locally responsive nature of its individual members. Within the tension between centralization and decentralization lies the exciting opportunity to create new strategies for successfully reducing the burden of cancer, especially among those disproportionately affected.
Figures and Tables
Table 1 Cancer Prevention Interventions Recommended by the Guide to Community Preventive Services (10-12)
Increasing Physical Activity Reducing Tobacco Use and Exposure to Environmental Smoke Reducing Exposure to Ultraviolet Light
Behavioral and social approaches Individually adapted programs
School-based physical education
Group programs that foster social support
Environmental and policy approaches Enhanced access to facilities, with informational outreach
Stair-use reminders
Informational approaches Community-wide campaigns
Stair-use reminders
Increasing cessationIncreasing the price of tobacco products
Mass media campaigns
Provider reminders
Reducing treatment out-of-pocket costs
Telephone quit lines
Reducing exposure to environmental tobacco smoke Smoking bans and restrictions
Reducing initiation Increasing the price of tobacco products
Mass media campaigns
Setting-specific approachesPrimary schools: education and policies
Recreation/tourism: education and policies
Table 2 The Cancer Prevention and Control Research Network (CPCRN) Research Work Groups and Their Goalsa
Work Group Research Goals
Colorectal cancer Develop a protocol for a community-based intervention trial to increase colorectal cancer screening and promote informed decision making for colorectal cancer screening among Hispanics.
Develop a protocol for implementing an intervention to increase colorectal cancer screening among men and women aged 50 and older.
Breast and cervical cancers Design a small-media community-intervention trial to increase the use and awareness of the CDC’s National Breast and Cervical Cancer Early Detection Program among program-eligible African American women.
Replicate small-media interventions to increase breast and cervical cancer screening among Hispanics (both farm-working and non-farm-working populations) living in the Texas-Mexico border area.
Prostate cancer Develop and pretest a work site intervention protocol to increase informed decision making for prostate cancer screening among men aged 50 and older.
Develop an interactive and innovative decision-making tool to promote informed decision making for prostate cancer screening among men aged 40 to 70.
Conduct and write a review of available literature regarding the effect of treatment on health-related quality of life among prostate cancer survivors, with an emphasis on African American men.
Identify a care provider network that serves African American men and supports informed and shared decision making for prostate cancer screening; assess the network’s acceptance and perception of the usefulness and relevance of NCI materials for informed decision making, and field test these materials with African American men.
Low-literacy materials validation Provide CPCRN sites and others with an extensive collection of tested materials for individuals who are among several minority and ethnic populations and have limited reading skills, and develop tools the CPCRN sites can use to validate materials.
Conduct a review of existing materials that address the need for timely follow-up of mammographic abnormalities among low-income ethnic minority women and, where needed, adapt materials to better meet the needs of these women.
a CDC indicates the Centers for Disease Control and Prevention; NCI indicates National Cancer Institute.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Harris JR, Brown PK, Coughlin S, Fernandez ME, Hebert JR, Kerner J, et al. The Cancer Prevention and Control Research Network. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0059.htm
==== Refs
1 Truman BI Smith-Akin CK Hinman AR Gebbie KM Brownson R Novick LF 18 1 Suppl 2000 18 26 Am J Prev Med Developing the Guide to Community Preventive Services--overview and rationale. The Task Force on Community Preventive Services 10806976
2 Doll L Kreuter M 7 2 2001 x xii J Public Health Manag Pract The prevention research centers program: linking science and practice through community collaborations 12174406
3 American Cancer Society Cancer facts & figures 2004 [Internet] Atlanta (GA) American Cancer Society 2004 8 9 cited 2004 Aug 9 Available from: URL: http://www.cancer.org/docroot/STT/stt_0.asp
4 Ries LAG Eisner MP Kosary CL Hankey BF Miller BA Clegg L SEER Cancer Statistics Review, 1975-2000 [Internet] Bethesda (MD) National Cancer Institute 2003 cited 2004 Aug 9
5 Weir HK Thun MJ Hankey BF Ries LA Howe HL Wingo PA 95 17 2003 1276 1299 J Natl Cancer Inst Annual report to the nation on the status of cancer, 1975-2000, featuring the uses of surveillance data for cancer prevention and control 12953083
6 Israel BA Schulz AJ Parker EA Becker AB 1998 19 173 202 Annu Rev Public Health Review of community-based research: assessing partnership approaches to improve public health 9611617
7 Israel BA Schulz AJ Parker EA Becker AB 14 2 2001 182 197 Educ Health (Abingdon) Community-based participatory research: policy recommendations for promoting a partnership approach in health research 14742017
8 Minkler M Wallerstein N Minkler M Wallerstein N 2003 Community based participatory research for health Introduction to community based participatory research Jossey-Bass San Francisco (CA)
9 National Cancer Institute Designing for dissemination: conference summary report [Internet] Washington (DC) Center for the Advancement of Health, Robert Wood Johnson Foundation 2002 cited 2004 Aug 9
10 Task Force on Community Preventive Services 20 2 Suppl 2001 10 15 Am J Prev Med Recommendations regarding interventions to reduce tobacco use and exposure to environmental tobacco smoke
11 Task Force on Community Preventive Services 22 4 Suppl 2002 67 72 Am J Prev Med Recommendations to increase physical activity in communities 11985935
12 Saraiya M Glanz K Briss P Nichols P White C Das D Task Force on Community Preventive Services On reducing Exposure to Ultraviolet Light 52 RR-15 2003 1 12 MMWR Recomm Rep Preventing skin cancer: findings of the Task Force on Community Preventive Services On reducing Exposure to Ultraviolet Light 14561953
13 18 1 Suppl 2000 92 96 Am J Prev Med Recommendations regarding interventions to improve vaccination coverage in children, adolescents, and adults. Task Force on Community Preventive Services 10806981
14 Briss P Rimer B Reilley B Coates RC Lee NC Mullen P 26 1 2004 67 80 Am J Prev Med Promoting informed decisions about cancer screening in communities and healthcare systems 14700715
15 Centers for Disease Control and Prevention Behavioral Risk Factor Surveillance System, 2002 data [Internet] Atlanta (GA) Centers for Disease Control and Prevention 2002 cited 2004 Dec 9
16 Appalachian Regional Commission 2000 The Appalachian Region; income rates in Appalachia, 1997; poverty rates in Appalachia, 1990; employment rates in Appalachia, 1998; ARC-designated distressed counties, FY 2000 cited 2004 Apr 23
17 Institute of Medicine 1999 The unequal burden of cancer: an assessment of NIH research and programs for ethnic minorities and the medically underserved Washington (DC) National Academies Press
18 Kentucky Cancer Registry 2002 1996-2000 Cancer incidence report Kentucky Cancer Registry Lexington (KY)
19 West Virginia Cancer Registry 2002 Cancer in West Virginia: Incidence and Mortality, 1993-2000 West Virginia Cancer Registry Charleston (WV)
20 Ramirez RR de la Cruz GP The Hispanic population in the United States: March 2002 [Internet] Current Population Reports Washington (DC) U.S. Census Bureau 2003 P20 545
21 Texas Department of Health 2001 Selected Facts for Border Area, 1999 Austin (TX) Texas Department of Health, Bureau of State Health Policy and Data Analysis
22 American Cancer Society 2003 Report No. 8623 Cancer facts & figures for Hispanics/Latinos, 2003-2005 Atlanta (GA) American Cancer Society
23 2003 Texas Cancer Registry thesource Cervical cancer incidence: Lower Rio Grande family females, 1995-1999 and other Texas females, 1995-1999 Austin (TX) Texas Department of State Health Services
24 American Cancer Society Texas Division, Inc Texas Cancer Facts & Figures 2002-2003: A sourcebook for planning and implementing programs for cancer prevention and control Austin (TX) The Society 2002
25 Fernandez ME Gonzales A Tortolero-Luna G Saavedra-Embesi M Evaluation of Cultivando La Salud: A breast and cervical cancer screening promotion program for Hispanic farmworker women Final Report submitted to the CDC, Feb. 2003
26 U.S. Census Bureau Massachusetts: 2000 - Census 2000 profile [Internet] Washington (DC) U.S. Census Bureau 2002 cited 2004 Apr 4
27 Brookings Institution Center on Urban and Metropolitan Policy Racial change in the nation's largest cities: evidence from the 2000 Census [Internet] Washington (DC) The Center 2001 cited 2004 Apr 19 Available from: URL: http://www.brook.edu/es/urban/census/citygrowth.htm
28 Bureau of Health Statistics Research and Evaluation 2003 Cancer incidence and mortality in Massachusetts, 1995-1999 Massachusetts Department of Public Health Boston (MA)
29 Bureau of Economic Analysis 2003 Personal income by state U.S. Department of Commerce Washington (DC)
30 U.S. Cancer Statistics Working Group 2002 United States cancer statistics: 1999 incidence Atlanta (GA) U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Cancer Institute
|
15670474
|
PMC1323324
|
CC BY
|
2022-01-25 23:20:05
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A21
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0077
Step-by-Step: Making Your Communities Healthier
The Border Health Strategic Initiative From a Community Perspective1
Steinfelt Victoria E University of Arizona, Cooperative Extension
2200 W 28th St, Yuma, AZ 85364-6928 [email protected]
928-726-3904
1 2005
15 12 2004
2 1 A232005
==== Body
A Community Coordinator's Perspective
As local community coordinator of the Border Health Strategic Initiative (Border Health ¡SI!), a program in Yuma County to develop a comprehensive, community-based approach to diabetes prevention and control, I was responsible for facilitating communication among five community partners, between community partners and campus-based faculty, and developing and coordinating a Special Action Group (SAG). My 30 years of experience as a cooperative extension agent, 27 of those years in Yuma County, helped to prepare me for this role.
As the Yuma County Cooperative Extension Agent, I had collaborated previously with each of the five community-based partners of Border Health ¡SI! on past programs, and I also had experience working with some of the campus-based faculty. Border Health ¡SI! provided funding to continue these partnerships and to build new partnerships with new campus faculty and community members. Twenty-five percent of my time was assigned to Border Health ¡SI!, and funding provided a full-time program assistant and half-time secretary for the cooperative extension component of the program. In addition to community coordination and the SAG, the Cooperative Extension Office was responsible for implementing the Centers for Disease Control and Prevention's School Health Index (SHI) in Yuma County (1).
The Cooperative Extension Office is located in the community, not on the university campus, and it involves local residents in identifying, planning, and implementing local programs. The Cooperative Extension Agent is a University of Arizona faculty member, thus providing a link between local community needs and university faculty expertise. This article addresses the details of coordinating a comprehensive diabetes prevention and management program from the viewpoint of a local community coordinator.
Purpose and Setting of Border Health ¡SI!
Border Health ¡SI! was implemented in Arizona along the border between the United States and Mexico. In Yuma County, Arizona, the initiative served the municipalities of Somerton, Gadsden, and San Luis, representing a total population of 25,563 — 92.1% of which is Hispanic. Approximately 35% of the residents in these agricultural communities have incomes below the 200% poverty level, and 30% are without medical insurance (2). Additional information about the purpose and setting of Border Health ¡SI! can be found in the introductory article of this issue of Preventing Chronic Disease (3).
A team of 10 faculty members from the Mel and Enid Zuckerman Arizona College of Public Health worked with five community-based partners to plan, implement, and evaluate Border Health ¡SI!. The community partners were the Yuma County University of Arizona Cooperative Extension, Western Arizona Health Education Center, Campesinos Sin Fronteras, Sunset Community Health Center, and Puentes de Amistad. The intervention components, outcomes, and conclusions are described in the other articles in this issue of Preventing Chronic Disease (1,3-9).
Recruiting Community Partners
Timetable for Yuma County, Arizona, Special Action Group (SAG), The Border Health Strategic Initiative
November 2000 through January 2001: Recruitment of SAG members
January 2001: First SAG meeting
January to April 2001: Monthly meetings to identify issues
April to June 2001: Subcommittees developed action plans
June 2001: SAG approved action plans
June 2002 to September 2003: Action plans implemented through Border Health ¡SI!
My first task as the community coordinator for the Border Health initiative was to organize a Special Action Group (1,9). Its mission was to determine and prioritize community policy issues surrounding physical activity and healthy eating. In November 2000, our small team (the Cooperative Extension program assistant and I) started the process of developing the SAG by making face-to-face contacts with key community leaders to explain the purpose and goal of Border Health ¡SI! and recruit them as members.
We did not set a goal for number of members to recruit because we were more interested in achieving broad community representation than meeting a numeric goal. For example, we recruited parks and recreation directors, city managers, community planners, school personnel, business owners, police officers, health care providers, county health department employees, elected officials, community residents, and others. The community partners were members of the SAG as well, and two campus-based faculty members participated in meetings and provided technical assistance to the SAG.
The face-to-face contacts were time consuming, but this strategy provided the best opportunity for a two-way dialogue about Border Health ¡SI! and the function of the SAG. During these dialogues, we asked for names of other key community people to contact. We also realized that community members lacked knowledge about diabetes prevention and discovered, ironically, that about half of those contacted had a family member with diabetes. We used this valuable information to prepare agenda items for the first SAG meeting, which took place in January 2001.
Once our team was established, our next major step was to schedule our first meeting, which took place two months after recruitment began. We met monthly for four months to select policy issues (9). Meeting notices were mailed each month to the 36 SAG members, and meeting attendance ranged from 20 to 28 members. During our fourth meeting, we formed two subcommittees to develop action plans. Each subcommittee had 10 to 12 members. One subcommittee developed the action plan to increase physical activity through advocating for more parks and walking paths, and the second committee addressed promoting healthier food choices in grocery stores and in schools (1). The entire group continued to meet monthly, and three months later we had action plans in place.
The Challenge of Systematic Problem-solving
In my experience, when you bring together community members to address a problem or concern, the group wants to do something immediately. It is often a challenge to get community members to take the time to follow a systematic problem-solving process (10). This process includes 1) examining in depth the issues at hand; 2) identifying alternative ways to address the issues; 3) writing an action plan, including ways to evaluate and implement the plan; 4) evaluating progress; 5) modifying the plan if needed; and 6) assessing results.
Fortunately, we were able to follow the systematic problem-solving process with the SAG for several reasons. By the time the SAG first met, some Border Health ¡SI! interventions were already being implemented. Recruitment for the walking clubs and family component had started, and patient education classes were being taught, so our members had a feeling that something was happening in their community. Our university partners presented information about the incidence and burden of diabetes at our meetings and led a discussion about the difference between community interventions and policy issues. Once our members grasped the need to address policy, the group understood the benefit of taking time to plan.
One of the action steps in our plan was to address food selection by South Yuma County grocery stores. Low-fat and nonfat milk, diet soda, and yogurt, for example, were available in limited quantities or not available at all in some stores. We planned to provide healthy food-promotion booths in stores, featuring food choices highlighted in Border Health ¡SI! community nutrition classes. Our hope was that if more customers requested healthier food choices, the stores would stock these items. The idea of providing nutrition information at grocery stores broke new ground in South Yuma County. It took several months to schedule an appointment with the storeowner because of his busy schedule. The storeowner was somewhat reluctant to allow our program to provide samples and nutrition information in his stores. He did not want us to tell his customers not to buy certain foods. He asked for a written plan and list of foods the program would promote. Again, it took several months to schedule another appointment to present the plan and list of foods. After reviewing the plan, the owner allowed us to provide samples and nutrition information in his stores with the conditions that we purchased the supplies in his stores and that we would not tell customers not to buy certain foods. This process took approximately eight months.
Surprises
Political changes in one community provided surprises. One of our action steps was to attend city council meetings in this community to support the open space and parks segment of the city plan, which was under discussion (9). A group of SAG members, walking club participants, and promotoras gathered to attend a city council meeting in which the city plan was listed as the first agenda item. When the group of 15 people arrived for the meeting, last-minute agenda issues arose because of a local political controversy, and the city plan was moved to the end of the meeting. Our group waited patiently for several hours but eventually went home before the city council introduced the item. The city plan item was scheduled for a later city council meeting, but we were able to gather only five people to attend.
When we implemented the SHI, we were surprised to discover the differences in beliefs among school principals regarding the role of schools in health promotion for students and staff. Some principals believed that the school has a very important role, and some believed that the school has no role. Obviously, it was easier to implement the SHI in the schools where the principal was supportive than in the schools where the principal was allowing us to implement the SHI but was not supportive. Through the SHI process, an action plan is developed, and the principal's support and encouragement for changes in the school is a critical element for changes to occur.
We were also surprised to learn about differences among community agencies regarding operating procedures, workplace culture, and funding sources. For example, operating procedures at the University of Arizona are different from those of a community-based nonprofit agency. The university's hiring procedures and expenditure process are more complicated, so it took longer for the university to hire program personnel and to obtain approval for program expenditures.
The differences in workplace culture revolved around the methods by which nonprofit agencies assigned personnel to Border Health ¡SI!. Some agencies cross-trained personnel who were funded by several different grants, while other agencies assigned total program responsibility to one person. Both strategies were effective, but at first it was confusing to know who exactly was working on Border Health ¡SI!.
We also discovered that the university could not serve as the lead organization for some grant applications. Some funds were available only to community-based nonprofit agencies. Small communities like ours tend to lack experienced proposal writers among nonprofit agencies, city offices, schools, and Cooperative Extension offices. The partnership between the local communities and the University of Arizona — newly strengthened by Border Health ¡SI! — provided community agencies with access to the university's proposal-writing expertise. Nonprofit agencies were able to apply for funding with technical assistance from the campus-based partner. Additionally, some nonprofit agencies and schools contracted with professional proposal writers, and the local community foundation offered proposal-writing training to community agencies.
Results-yielding Synergy
Community groups begin simply as a collection of people. It takes time to evolve into a working team that generates synergy (11). Over the three years of the Border Health ¡SI!, our SAG members developed into a working team that had an impact on the community in a way that no one agency or organization could have accomplished working alone.
For example, in one small community, a community leader had been working for many years to renovate an existing park. He formed a community group and experienced some success. This individual was identified as a potential member of the parks and open space subcommittee of our SAG, and he agreed to join. At his first meeting, he described the frustrations of trying to obtain funding for the park and mentioned that he was just about ready to give up. The group listened to his concerns and empathized with his frustration. Members pointed out the accomplishments of his group and made a commitment to work together with this community leader and other community groups to increase the number of parks and renovate existing parks in South Yuma County. About a year later, after our first Border Health ¡SI! Park Development Community Forum, which was a component of the SAG's action plan, this person shared with me his appreciation for being a member of the SAG. He said that he and his community group would not have been able to reach their goals without the assistance of other SAG members.
During one meeting, our university partners led the SAG in a discussion to identify the outcomes the members felt they had accomplished. One comment made during the meeting summarizes the synergy that developed: "The SAG has been instrumental in bringing key government people to meetings and networking. Education of the government entities from this awareness and the funding for two parks has been acquired. Yuma County officials would not be open to listen or cooperate as much if the SAG hadn't been involved on a big scale."
Conclusion
The Border Health Strategic Initiative was a three-year program that ended September 2003. The basic model that was developed will be continued and expanded through Steps to a HealthierUS, which is an initiative of the U.S. Department of Health and Human Services.
Although coordinating a comprehensive community-based health promotion program is time consuming, the synergistic relationship that evolves will yield exciting and rewarding results for you and the communities involved.
Our special focus this issue is on the Border Health Strategic Initiative (Border Health ¡SI!)) along the U.S.-Mexico border in Arizona. Related articles are indicated with the icon. Selected articles and abstracts are available in both English and Spanish.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Steinfelt VE. The Border Health Strategic Initiative from a community perspective. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0077.htm
==== Refs
1 Staten LK Teufel-Shone NI Steinfelt VE Sanchez N Halverson K Flores C Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] The School Health Index as an impetus for policy change
2 U.S. Census Bureau 2000 The Bureau; 2000 United States Census 2000 Washington (DC) The Bureau
3 Cohen SJ Ingram M Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Border heath strategic initiative: overview and introduction to a community-based model for diabetes prevention and control
4 Abarca J Ramachandran S Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Using community indicators to assess nutrition in Arizona-Mexico border communities
5 Ingram M Gallegos G Elenes J Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Diabetes is a community issue: the critical elements of a successful outreach and education model on the U.S.-Mexico Border
6 Schachter KA Cohen SJ Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] From research to practice: challenges in implementing national diabetes guidelines with five community health centers on the border
7 Teufel-Shone NI Drummond R Rawiel U Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Developing and adapting a family-based diabetes program at the U.S.-Mexico border
8 Staten LK Scheu LL Bronson D Peña V Elenes J Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Pasos adelante: the effectiveness of a community-based chronic disease prevention program
9 Meister JS de Zapien JG Prev Chronic Dis [serial online] 2005 Jan [2004 Dec 15] Bringing health policy issues front and center in the community: expanding the role of community health coalitions
10 Klein G 1982 The S-T-P model of problem solving family community leadership regional resource notebook Corvallis (OR) FCL at the Western Rural Development Center
11 Senge PM Kleiner A Roberts C Ross R Smith B Day 1994 48 77 The fifth discipline fieldbook New York Doubleday
|
15670476
|
PMC1323326
|
CC BY
|
2022-01-27 23:20:38
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A23
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0108
Letter to the Editor
Reengineering Vital Registration and Statistics Systems
Nasseri Kiumarss DVM, PhD, MPH Public Health Institute, Tri-Counties Cancer Surveillance Program, Cancer Center of Santa Barbara, Santa Barbara, Calif
1 2005
15 12 2004
2 1 A252005
==== Body
To the Editor:
In his timely essay, "Reengineering Vital Registration and Statistics Systems for the United States," Charles J. Rothwell raises important issues about the registration of vital events in the United States (1). There is no question about the value of taking advantage of electronic technology to improve the quality and increase the utility of the registration system. The partnership mentioned by Rothwell is a step in the right direction. Currently, only a small portion of the data collected on paper death certificates is captured in machine-readable format. The rest — including valuable address and occupation information — is lost forever for population-based research and public health purposes.
My primary concern, however, is about the issues that were not discussed by Rothwell. First, the paper death certificate is a document whose legality precedes its public health importance. Unless electronic documents are legally accepted, we must have the paper documents.
Second, some important data are "mutilated" when captured in machine-readable format. According to National Center for Health Statistics (NCHS) guidelines, the place of birth of the deceased is coded only to the 50 states, the U.S. territories, and a few foreign countries like Canada and Mexico. All the other nations of the world are grouped together as one code! Granted, individuals born in foreign lands (i.e., first-generation immigrants) do not make up a large segment of U.S. death certificates, but from a public health point of view, they are an important group. Some states defy NCHS guidelines and collect the exact place of death for each deceased person in machine-readable format, but the utility of these data is limited to the state collecting the information; national organizations are forced to accept NCHS format. Also, information on the place of birth of a decedent's parents is of major epidemiological value because it identifies second-generation immigrants. Some states collect this information, but following NCHS guidelines, they do not capture it in machine-readable format.
The issue of accuracy is significant. Technically, we can evaluate the accuracy of information only against an independent document. Yet we accept many data on official, technical, and research documents at face value without verification and use the data to draw important conclusions. Information on occupation, for example, can be complex. Most people change their occupation several times, and upon death, a simple choice of occupation may not seem so clear. Or, perhaps the next of kin wishes to "upgrade" the occupation of the deceased. "Retired" and "housewife" are the most common occupations given for men and women, despite what careers they may have pursued. Nevertheless, with proper instructions and smart algorithms, we can improve the process of collecting information on occupation so that it can be used effectively for research purposes.
My main point in writing this letter is to draw attention to some important details that might be lost during the process of upgrading the current vital registration system to an electronic one. I would also like to suggest that the partnership of the National Association of Public Health Statistics and Information System, NCHS, and Social Security Administration be expanded to include a fourth member that represents the consumer side of these data — epidemiologists or other public health officials who have closely worked with these data for many years and are intimately aware of their utility and shortcomings.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Nasseri K. Reengineering vital registration and statistics systems [letter to the editor]. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0108.htm.
==== Refs
1 Rothwell CJ Prev Chronic Dis [serial online] 2004 10 Accessed 2004 Sep 15 Reengineering vital registration and statistics system for the United States
|
15670477
|
PMC1323327
|
CC BY
|
2022-01-27 23:18:42
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A25
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0108
Letter to the Editor
Reengineering Vital Registration and Statistics Systems
Nasseri Kiumarss DVM, PhD, MPH Public Health Institute, Tri-Counties Cancer Surveillance Program, Cancer Center of Santa Barbara, Santa Barbara, Calif
1 2005
15 12 2004
2 1 A252005
==== Body
To the Editor:
In his timely essay, "Reengineering Vital Registration and Statistics Systems for the United States," Charles J. Rothwell raises important issues about the registration of vital events in the United States (1). There is no question about the value of taking advantage of electronic technology to improve the quality and increase the utility of the registration system. The partnership mentioned by Rothwell is a step in the right direction. Currently, only a small portion of the data collected on paper death certificates is captured in machine-readable format. The rest — including valuable address and occupation information — is lost forever for population-based research and public health purposes.
My primary concern, however, is about the issues that were not discussed by Rothwell. First, the paper death certificate is a document whose legality precedes its public health importance. Unless electronic documents are legally accepted, we must have the paper documents.
Second, some important data are "mutilated" when captured in machine-readable format. According to National Center for Health Statistics (NCHS) guidelines, the place of birth of the deceased is coded only to the 50 states, the U.S. territories, and a few foreign countries like Canada and Mexico. All the other nations of the world are grouped together as one code! Granted, individuals born in foreign lands (i.e., first-generation immigrants) do not make up a large segment of U.S. death certificates, but from a public health point of view, they are an important group. Some states defy NCHS guidelines and collect the exact place of death for each deceased person in machine-readable format, but the utility of these data is limited to the state collecting the information; national organizations are forced to accept NCHS format. Also, information on the place of birth of a decedent's parents is of major epidemiological value because it identifies second-generation immigrants. Some states collect this information, but following NCHS guidelines, they do not capture it in machine-readable format.
The issue of accuracy is significant. Technically, we can evaluate the accuracy of information only against an independent document. Yet we accept many data on official, technical, and research documents at face value without verification and use the data to draw important conclusions. Information on occupation, for example, can be complex. Most people change their occupation several times, and upon death, a simple choice of occupation may not seem so clear. Or, perhaps the next of kin wishes to "upgrade" the occupation of the deceased. "Retired" and "housewife" are the most common occupations given for men and women, despite what careers they may have pursued. Nevertheless, with proper instructions and smart algorithms, we can improve the process of collecting information on occupation so that it can be used effectively for research purposes.
My main point in writing this letter is to draw attention to some important details that might be lost during the process of upgrading the current vital registration system to an electronic one. I would also like to suggest that the partnership of the National Association of Public Health Statistics and Information System, NCHS, and Social Security Administration be expanded to include a fourth member that represents the consumer side of these data — epidemiologists or other public health officials who have closely worked with these data for many years and are intimately aware of their utility and shortcomings.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Nasseri K. Reengineering vital registration and statistics systems [letter to the editor]. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0108.htm.
==== Refs
1 Rothwell CJ Prev Chronic Dis [serial online] 2004 10 Accessed 2004 Sep 15 Reengineering vital registration and statistics system for the United States
|
0
|
PMC1323328
|
CC BY
|
2022-01-27 23:15:39
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A26
|
latin-1
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0120
Errata
Erratum, Vol. 1, No. 4
1 2005
15 12 2004
2 1 A272005
==== Body
In the PDF version of the article “Considerations for Using a Geographic Information System to Assess Environmental Supports for Physical Activity,” Figure 2 was incorrectly labeled. The correct figure legend is “Figure 2. Using a half-mile buffer to represent a neighborhood around a survey respondent’s home address, GIS can be used to identify a sidewalk or recreation facility in a survey respondent’s neighborhood.” The correct legend for Figure 2 appeared in the original HTML version, and the PDF version was corrected on our Web site on October 12, 2004. It appears online at http://www.cdc.gov/pcd/issues/2004/oct/04_0047.htm. We regret any confusion this error may have caused.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
|
0
|
PMC1323329
|
CC BY
|
2022-01-26 23:18:03
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A27
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
Prev Chronic Dis
Preventing Chronic Disease
1545-1151
Centers for Disease Control and Prevention
PCDv21_04_0105
Book Review
Latina Health in the United States: A Public Health Reader
Beckles Gloria L.A. MD, MSc Medical Epidemiologist Division of Diabetes Translation, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention Atlanta, GA
1 2005
15 12 2004
2 1 A24Aguirre-Molina Marilyn Molina Carlos W
Latina Health in the United States: A Public Health Reader.
Jossey-Bass, John Wiley & Sons, Inc. 2003. 9 ISBN: 0-7879-6579-0 Price: $45.00 688 pages 2005
==== Body
Latina Health in the United States: A Public Health Reader is a collection of articles drawn from journals and books published in the last decade about the health status and health needs of Latin American adolescent and adult females in the United States. In the preface, the editors state that the goals of the anthology are to identify a number of critical issues of importance and relevance to Latina health; to present an overview of the existing literature on Latina health; to highlight the leading indicators of morbidity and mortality affecting various subgroups of Latinas; and to identify gaps in research, policy issues, program planning, and practice. To achieve these goals, the editors organized the book's 29 chapters into nine parts that explore the population's demographics; risk factors, socioeconomic disparities, and race/ethnic disparities among Latinas; sexual and reproductive health issues; chronic conditions such as heart disease, cancer, and diabetes; alcohol, tobacco, other drug use, and mental health; patterns of risk behaviors among Latina adolescents; and the health needs of rural and migrant workers.
In Part 1, the editors describe the process used to identify the essential topics for inclusion in the book. Content experts were interviewed, morbidity and mortality data were reviewed, and a comprehensive review of papers published during 1984–2002 was undertaken to identify the extent to which published research reflects the health needs of Latinas. The editors conclude that research on Latinas has several limitations, including lack of generalizability, small sample size, misclassification bias (when Spanish surnames are used to identify sampling frames), and lack of comparison groups. Furthermore, they report that few journals publish studies related to Latina health issues. In Parts 2 through 6, contributors document the principal health conditions that beset Latinas and the persistent racial and ethnic differences in biological factors, behavioral risk factors, and use of health services. Parts 3, 7, and 9 focus on the health and health care experience of women at different life stages (adolescence, the reproductive years, and midlife).
Latina Health in the United States is primarily intended for four groups of people: students, practitioners, decision makers, and researchers who seek to address the challenge of the growing health disparities among communities of color. The book should be required reading in all schools of public health; it synthesizes and presents in a single source the major issues on the health of women of Latin American origin. For public health practitioners and decision makers, it should be a necessary general reference because it calls attention to issues that have important implications for health policy and planning of programs and services (e.g., the heterogeneity in geographic origins of Latinas, the lack of data to provide timely assessment and monitoring of the health status of Latinas). Additionally, the population dynamics (e.g., high fertility and immigration rates, regional concentration) that lead to rapid population growth raise the question about the ability of the health care system, at least regionally, to meet the current and future health needs of Latina women.
The gaps in knowledge of Latina health issues documented by the articles in the book pose a challenge to researchers. In particular, the articles demonstrate the necessity to understand the mechanisms and pathways by which structural and contextual factors (e.g., poverty, low-wage employment, social capital, cultural norms) may protect against or determine the health deficits that often result from international migration and acculturation. The findings from such research could be used to develop intervention strategies to preserve protective behaviors and to contribute to risk reduction in Latinas and other women.
Overall, the editors have realized several of their stated goals. Latina Health in the United States is a much needed and timely collection that chronicles a wealth of information — information that can sharpen the focus and guide the direction of efforts to improve the health of Latinas.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
Suggested citation for this article: Beckles GLA. Latina health in the United States: a public health reader [book review]. Prev Chronic Dis [serial online] 2005 Jan [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/jan/04_0105.htm.
|
0
|
PMC1323330
|
CC BY
|
2022-01-31 23:31:35
|
no
|
Prev Chronic Dis. 2004 Dec 15; 2(1):A24
|
utf-8
|
Prev Chronic Dis
| 2,004 |
nan
|
oa_comm
|
==== Front
BMC Med EthicsBMC Medical Ethics1472-6939BioMed Central London 1472-6939-6-91622974310.1186/1472-6939-6-9Research ArticleThe activity of French Research Ethics Committees and characteristics of biomedical research protocols involving humans: a retrospective cohort study Decullier Evelyne [email protected]éritier Véronique [email protected] François [email protected] UF de méthodologie en recherche clinique, Département d'Information Médicale des Hospices Civils de Lyon, Lyon, France2 Hôpital Edouard Herriot, Lyon, France3 French National Conference of Research Ethics Committees, Lyon, France2005 17 10 2005 6 9 9 21 3 2005 17 10 2005 Copyright © 2005 Decullier et al; licensee BioMed Central Ltd.2005Decullier 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
Clinical trials throughout the world must be evaluated by research ethics committees. No one has yet attempted to clearly quantify at the national level the activity of ethics committees and describe the characteristics of the protocols submitted. The objectives of this study were to describe 1) the workload and the activity of Research Ethics Committees in France, and 2) the characteristics of protocols approved on a nation-wide basis.
Methods
Retrospective cohort of 976 protocols approved by a representative sample of 25/48 of French Research Ethics Committees in 1994. Protocols characteristics (design, study size, investigator), number of revisions requested by the ethics committee before approval, time to approval and number of amendments after approval were collected for each protocol by trained research assistant using the committee's files and archives.
Results
Thirty-one percent of protocols were approved with no modifications requested in 16 days (95% CI: 14–17). The number of revisions requested by the committee, and amendments submitted by the investigator was on average respectively 39 (95% CI: 25–53) and 37 (95% CI: 27–46), per committee and per year. When revisions were requested, the main reasons were related to information to the patient (28%) and consent modalities (18%). Drugs were the object of research in 68% of the protocols examined. The majority of the research was national (80%) with a predominance of single-centre studies. Workload per protocol has been estimated at twelve and half hours on average for administrative support and at eleven and half hours for expertise.
Conclusion
The estimated workload justifies specific and independent administrative and financial support for Research Ethics Committees.
==== Body
Background
Clinical trials in biomedical research throughout the world must be evaluated by research ethics committees. No one has yet attempted to clearly quantify the activity of ethics committees and describe the characteristics of the protocols submitted, and only five studies[1-4] have dealt with this subject, two in England and the others in the USA, Australia and Spain. However, they each focused on only one or two research ethics committees, and no knowledge is available on a national or even a regional level. In addition, four of these studies[2-4] concentrated on the fate of protocols; some characteristics were gathered but they were presented only for protocols whose investigator was not lost to follow-up. None of these studies evaluated research ethics committees' workload, but one study[1] described the activity in terms of number of meetings, approvals and reasons of queries.
France was one of the first country to affirm these principles through the Huriet-Sérusclat Act of 1988 [5], which launched administrative and financially independent research ethics committees (REC) in 1991, called CCPPRBs (committees for the protection of human beings involved in biomedical research). Each protocol involving human beings in France has to be assessed by one of these committees.
A European harmonization was needed since a long time and was decided in 2001 [6]. All European governments were asked to change national laws according to this directive. Research ethics committees will be created where they do not exist currently and it is of upmost importance to anticipate activity, workload, financial and administrative support needed by these committees.
We describe one year of activity in a sample comprising half of the French RECs and the full characteristics of the protocols approved in these RECs during this year.
Methods
All biomedical research protocols approved by a representative random sample of every other French RECs the participating committees between 01 January 1994 and 31 December 1994 were included. This year was chosen in order to allow studies to be completed and so to collect data on all amendments, information and serious adverse effects.
Definitions
The definitions used in this study are drawn from French law on the subject (table 1).
Table 1 Definitions
Area covered by the law Every trial or experimentation involving human beings based on a procedure formalized in a protocol in order to develop biological or medical knowledge.
CCPPRB "Comités Consultatifs de Protection des Personnes se Prêtant à la Recherche Biomédicale", i.e., committees for the protection of human beings involved in biomedical research.
Composition of REC The committees gather 12 members and 12 replacements. The REC is made up of four persons competent in biomedical research (at least three physicians), a general practitioner, two pharmacists (at least one working in a hospital, a nurse, one person qualified in ethics, one social worker, one person qualified in psychology and one person qualified in the legal aspects of research.
Studies without direct individual benefit None of the participants could expect any individual and immediate benefit, e.g. research in physiology or phase I studies are typically considered as studies with no therapeutic benefit for subjects.
Studies with direct individual benefit Patients can potentially expect a therapeutic benefit from the research.
Decision RECs have four possibilities in their decisions: rejecting the protocol, accepting the protocol without any modifications, accepting the protocol with minor modifications or asking for revisions (major modifications) before approval.
Time to decision RECs must give their first answer within 5 weeks after the protocol application.
Revision Mandatory modification requested by the REC before protocol approval.
Amendment Modifications proposed by the investigator to the committee after approval
Information Investigators can send information concerning their study (state of inclusions, article, etc.). This information does not lead to a decision by the REC.
Data collection
Ethics committee activity and protocol characteristics were gathered into four thematic groups. The first two comprised committee activities: (i) the process of approval (modifications requested before approval, approval date, etc.), (ii) the modifications requested after approval and the information transmitted. The other two comprised protocol characteristics: (iii) the legal and administrative characteristics of protocols (investigator, sponsor, etc.) and (iv) the proposed scientific characteristics (duration, sample size, etc.). Research assistants in charge of data collection were trained in order to ensure homogeneous results. The questionnaires were completed using the committee's archives, and were then sent to the coordinating centre. An identification number was given to each protocol in order to ensure anonymity of the investigator.
Assessment of activity and expertise workload
Activity
Each formal step was studied, i.e., the number of protocols approved, revisions requested by the ethics committee and amendments submitted by the investigator were studied. The number of revisions, amendments and total exchanges (defined as the sum of revisions and amendments) were calculated per approved protocol.
Administrative and expertise load
At the REC level, each protocol is registered and assessed for approval. A list of the different tasks was established by the Lyon B REC members and administrative staff and submitted to a national panel of REC members (table 2). The time required to accomplish each task was estimated based on individual experience and checked on site. For this study, the estimated times devoted to administrative work and expert assessment were then accepted by consensus among administrative staff and REC members respectively. The workload is estimated from a protocol perspective, i.e. each protocol is valued. We did not assess neither fix activities, such as or training of the members or structure management (accounting, legal, insurances, structure organisation,...), nor the committee meetings.
Table 2 Development of REC activities
Main tasks administrative time (h) expert time (h)
New file Receipt 0.25 0.25
Computer registration 0.25 0,25
Content and conformity checking 1 1
Acknowledgement of receipt 0.5 -
Sending to the experts 0.5 -
Inscription of committee's meeting agenda 0.5 -
Sending the reply ready for posting 1 -
Filing 0.5 0.5
Valuation+writing and sending valuer's report - 4
TOTAL NEW FILE 4.5 6
Revision Request for information to the investigator 1 -
Disfiling 0.5 0.5
Reply receipt 0.5 0.25
Sending to the experts 0.5 -
Inscription of committee's meeting agenda 0.5 -
Sending the reply ready for posting 1 -
Filing 0.5 0.5
Valuation+writing and sending valuer's report - 2
TOTAL REVISION 4.5 3.25
Amendment Receipt 0.5 0.5
Disfiling 0.5 0.5
Sending to the experts 0.5 -
Inscription of committee's meeting agenda 0.5 -
Writing and sending the reply 1 -
Filing 0.5 0.5
Valuation+writing and sending valuer's report - 1
TOTAL AMENDMENT 3.5 2.5
-: not applicable
Statistics
As this was mainly a descriptive study, we presented frequency distributions calculated with SAS software®. A Kaplan-Meier survival analysis [7] was performed with SPSS software® to establish probability of approval curves based on cumulative hazard function in order to study the time between the submission of the protocol and final approval (in days). A log-rank test [7] was then computed to compare protocols with direct approval, protocols with minor modifications and protocols with revisions.
A multiple correspondence analysis [8] was also performed on protocol characteristics. To determine the number of axes we used the scree test [9] and the total inertia was computed with Benzecri formulae [10]. To explain the meaning of each axis, modalities of variables were cumulated until 80% of inertia was explained. The individual coordinates were then used to obtain a hierarchical clustering with the Ward minimum variance method [11]. The number of clusters was chosen using the cubic clustering criterion [12].
To compare number of revisions, amendments and time per protocol between the clusters an ANOVA was performed [13].
Results
A total of 25/48 RECs throughout France participated in this national study. There were 1143 declared protocols approved by these committees during the year 1994. One hundred sixty-seven had to be excluded because the inclusion criteria were not fulfilled (Table 3). Thus 976 protocols were eligible (mean per committee, 39; median, 37; range, 17–81).
Table 3 Reasons for non-inclusion (167 protocols out of 1143)
n %
Approved in 1995 82 49
Withdrawn before approval 48 29
Rejected 16 9
Not within the scope of the law 12 7
Missing file 6 4
Approved in 1993 3 2
REC activity
Approval
Only 31% of protocols were approved with no request for modifications. Minor or major modifications were requested for the remaining protocols.
Time from submission of application to approval
Because of missing registration and/or approval dates, seven protocols were dropped from the study, and the analysis was performed on 969 studies. The median time to approval for protocols with direct approval, with minor changes requested, and request for major changes (revisions) was, respectively, 16 days (95% confidence interval: 14–17), 27 days (24–30) and 48 days (43–52). The comparison between the three groups resulted in a significant log-rank test (figure 1).
Figure 1 Probability of approval from time of submission, according to the initial decision of Research Ethics Committees.
Number of modifications studied
On average, the number of revisions required was 39 per committee and per year (median, 30; range, 4–132, 95% CI, 25–53) and the number of amendments was 37 (median, 28; range, 3–91, 95% CI, 27–46). Thus the average number of exchanges (revisions and amendments) was 76 (median, 60; range, 19–212, 95% CI, 55–96) per year and per committee, ranging from 13 exchanges for phase I studies on drugs, through 19 for studies not investigating drugs and to 45 for studies on drugs other than phase 1.
Revisions (major changes)
Revisions were requested for 555 (57%) protocols, but only 181 (19%) required a second revision or more. A revision could contain one or more points to be modified by the investigator. In total 1438 points for modification were cited (Table 4). Per protocol, the average number of reasons per revision was 2.01 (the same reasons could have required more than one revision). On average, one reason (or less) was mentioned in 45% of revisions, between one and two reasons in 25%, between two and three 15%, and more than three reasons in 15% (data not shown).
Table 4 1211 main reasons for revision before approval (out of a total of 1438)
n %
Patient information 399 27.7
Consent modalities 257 17.8
Inclusion criteria 116 8.1
Scientific prerequisite 105 7.3
Legal requirements 91 6.3
Sample size 71 4.9
Insurance 49 3.4
Information on treatments and exams 48 3.3
Study objectives 44 3.1
Information on methodology and statistics 31 2.2
The main reasons for revisions were information to the patient (28% of responses), consent modalities (18%), inclusion criteria (8%), scientific factors (7%) and legal and administrative requirements (6%). This global ranking of reasons remained the same when considering the sub-group of protocols with only one revision and the sub-group of protocols with two revisions.
The ranking of reasons for protocols with three or more revisions showed that legal and administrative points, which were quite easy to solve, became less important, falling from the 4th position (for protocols with one or two revisions) to the 9th position. However, requests for more specific methodological and statistical information, which were more complex to solve, were more frequently cited, rising from the 11th rank (for protocols with one or two revisions) to 7th rank.
Amendments (after approval)
Over half of the investigators submitted no amendments (57%), although few protocols were not implemented and therefore could not have led to amendments. For the 416 protocols requiring at least one amendment, 875 reasons were mentioned. An average of 1.2 reasons per amendment was cited. A single reason was cited in 52% of cases.
The main reasons for which protocols had to be modified after approval were a change in the number of subjects to be included (18%), a modification of inclusion criteria, a modification of the timetable and changes in examinations and treatments (Table 5). Regarding the legal and administrative points cited, 43% concerned the update of the list of investigators (data not shown). The number of amendments submitted did not modify the ranking of these reasons.
Table 5 657 main reasons for amendment after approval (out of a total of 875)
n %
Sample size 156 17.8
Inclusion criteria 143 16.3
Timetable 132 15.1
Treatment or exam modifications 89 10.2
Patient information 76 8.7
Legal and administrative modifications 61 7.0
Additional information
As information on the progress of research is not mandatory, 81% of investigators did not send any further information regarding their study. When they did it was mainly to warn of the end (premature or normal) of the study (30%), to declare adverse side effects (25%) and to inform the committees of the intermediate or final study results (18%).
RECs' administrative and expertise workload
The administrative staff panel estimated that administrative tasks (registration, postal service, copying, classification, letters, etc.) required 4.5 hours per initial protocol recording, 4.5 hours per revision and 3.5 hours per amendment (table 2). The experts panel estimated that the expertise process required 6 hours per protocol, 3.25 hours per revision and 2.5 hours per amendment. This was confirmed during the test on-site.
These figures show, per year and per month respectively, an estimated administrative workload of 480.50/40 hours and an estimated expertise workload of 453.25/38 hours per REC (based on observed mean annual activity per committee: 39 protocols, 39 revisions and 37 amendments on average per REC and per year).
Protocol characteristics
Table 6 summarizes the main legal and administrative characteristics of protocols and Table 7 the main scientific and technical characteristics.
Table 6 Administrative characteristics of approved protocols
n %
Main investigator's status
Professor 567 58.1
Assistant 255 26.1
Other 154 15.8
Sponsor
Pharmaceutical firm 628 64.3
Tertiary teaching hospital 161 16.5
Industry 45 4.6
Other public organisation 102 10.4
Other 40 4.1
Place of research
Phase I specialized unit 162 16.6
Tertiary teaching hospital 543 55.6
Tertiary teaching hospital + private hospital 121 12.4
Tertiary teaching hospital + phase I specialized unit 18 1.8
Other 132 13.5
Table 7 Technical characteristics of approved protocols
n %
Topic
Drug 667 68.3
Phase I 163 24.4
Phase II 173 25.9
Phase III 226 33.9
Phase IV 105 15.7
Cosmetics and nutrition 50 5.1
Physiology 65 6.7
Medical equipment & prothesis 64 6.6
Surgical and/or diagnostic act 52 5.3
Other 78 8.0
Design
Descriptive, analytic 150 15.4
Experimental, non-randomized 319 32.7
Experimental, randomized 507 51.9
Scope
National-monocentric 463 47.4
National-multicentric 311 31.9
International 161 16.5
Not available 41 4.2
Expected sample size
Fewer than 21 patients 293 30.0
21 to 50 patients 251 25.7
51 to 150 patients 192 19.7
More than 150 patients 228 23.4
Not available 12 1.2
Expected duration
Less than 2 months 115 11.8
2 to 6 months 159 16.3
6 to 18 months 286 29.3
More than 18 months 163 16.7
Not available 253 25.9
Legal and administrative characteristics
Sponsors were mainly pharmaceutical firms (64%). The research setting was a tertiary teaching hospital in 55% of cases; another 15% of protocols were conducted simultaneously in a tertiary teaching hospital and another type of institution. The item "other" included other combinations of these settings and also private offices, etc.
Technical characteristics
Drug evaluation was the object of research in 68% of cases. For the research topic, the item "other" was cited in 8% of cases, usually for studies on genetics or vaccines, i.e., other types of clinical trials. It was also pointed out that participating RECs encountered problems in reporting expected duration since 26% of the answers on questionnaires neglected to mention this information.
Reasons for revisions varied according to investigators' status (p < 0.0001; revisions on scientific prerequisite were more frequent when investigators were neither professor, nor assistant) and to place of research (p = 0.005; revisions on inclusion criteria were more frequent when research was conducted in phase I units) and reasons for amendments according to sponsor (p = 0.01; amendements on sample size were more frequent when the sponsor was a public hospital), place of research (p = 0.03; amendments on treatment was more frequent for research conducted in phase I units), study phase (p = 0.04; amendements concerned more often treatment for phase I studies), study scope (p = 0.002; sample size was more often modified when the study was national and multicentre) and study duration (p = 0.001; amendments were more often about treatment when study duration was less than 2 months).
Multiple correspondence analysis
the scree test resulted in the selection of two axes explaining 40% of the inertia. The modalities most contributing to the construction of each axis were represented on figure 2. The first axis (dashed line) showed that study size differentiated protocols the most. On one side were phase I studies, protocols studying nutrition, cosmetics, protocols on physiology, without direct benefit to the patient, of short duration, single-centre studies, and those with fewer than 20 patients expected (small studies). On the other side were phase III studies, those with direct benefit to the patient, more than 200 patients to be included, national multicentre or international studies, and those with a planned intermediate analysis (large studies).
Figure 2 Graphical representation of multiple correspondence analysis on protocols characteristics*. *graphical representation of the modalities which contributes the most to the construction of axes 1 and 2.
The second axis separated descriptive studies sponsored by the public sector from experimental studies on drugs, mostly sponsored by private pharmaceutical firms.
Hierarchical clustering using Ward's minimum variance
This analysis yielded four clusters explaining 75% of the variance. The first cluster included protocols on drug testing (98% in this cluster vs. 68% in the global population), with direct individual benefit, and phase II and phase IV studies (respectively, 35% and 22% vs. 18% and 11%). Cluster 2 grouped phase I drugs trials (89% vs. 17%), without direct individual benefit (97% vs. 34%), in a single centre (99% vs. 47%), lasting less than 2 months and sponsored by pharmaceutical firms. Cluster 3 also concerned drugs, but protocols were phase III studies (64% vs. 23%), with direct individual benefit, multicentre and international (42% vs. 16%) and with a randomized design (87% vs. 52%). Finally, cluster 4 grouped protocols not evaluating drugs (89% vs. 32%), the studies were descriptive (32% vs. 15%) and the sponsor was in the public sector (58% vs. 27%).
Activity and workload in relation with protocol characteristics
We have linked the different clusters to the number of revisions and amendments (table 8). The mean number of revisions is homogeneous through clusters with a minimum of 0.94 for cluster I and a maximum of 1.08 for cluster III. Number of amendments varied more: from 0.46 for cluster II to 2.14 for cluster III.
Table 8 Tendendy in characteristics of 976 protocols according to clusters defined by a hierarchical analysis
CLUSTER I CLUSTER II CLUSTER III CLUSTER IV
N 280 144 209 343
Main characteristics -drug testing
-phase II
-pharmaceutical sponsor
-with direct benefit
-phase IV -phase I
-without direct benefit
-investigator: other
-single centre
-less than 2 months -international scope
-phase III
-more than 200 patients
-drug testing
-with direct benefit -not drugs
-public setting
-descriptive design
-public sponsor
-industrial sponsor
PER PROTOCOL ANOVA p-value
Mean number of revisions 0.94 1.05 1.08 0.98 0.70
Mean number of amendments 0.78 0.46 2.14 0.53 <0.0001
Estimated average time (initial evalution, revision, amendment) 22.5 21.4 31.7 21.3 <0.0001
The mean time spent per protocol (workload) was estimated in each cluster, cluster III protocols are the most time-consuming.
Time spent by each cluster was obtained by multiplying number of protocols by mean time per protocol. At a national level 31% of committees' time is spent for cluster IV protocols, 28% for cluster III, 27% for cluster I and 13% for cluster II.
Discussion
On average, each REC studied 39 protocols, 39 revisions and 37 amendments per year, representing an annual workload of 934 hours (480.50 hours for administrative tasks and 453.25 for expertise tasks, on a monthly basis respectively 40 and 38 hours) including neither committee meetings nor training of the members since there is too much variability across committees. Most protocols evaluated drugs (68%), were experimental (85%) and were monocentric (47%).
The main advantage of our study is to have collected exhaustive information for half of French RECs over a full year (whether or not the principal investigator was lost to follow-up). Moreover, homogeneous data collection was guaranteed by the initial training followed by research assistants. The choice of the year 1994 was based on initial studies conducted by the French Ministry of Health [14-16] showing that on average six years were needed to reach publication, and that some studies were still ongoing eight years later. This choice enabled us to gather all the information during the whole protocol's duration. To our knowledge our study is the first to have evaluated the workload and the activity of RECs for each protocol through the collection of the number of revisions, amendments and additional information. The combination between 1994's activity data and 2004's workload assessment is justified since committees' missions have remained the same and there is no reason why the registration or the expert evaluation should take more or less time in 1994 compared to 2004. A 2001 French Senate report [17] showed similarities across years in the number of protocols evaluated by French RECs (total number of protocol, type of research,...). Moreover, this would make more sense to use the current procedure and workload (2004) rather than the old ones (1994). To our mind, this evaluation was closest to the reality, since a retrospective cohort was mandatory to evaluate characteristics and fate of biomedical protocols, whereas retrospective data collection was not possible to assess workload.
Moreover the French law is broader than EC 2001/20 European Directive [6], since it aims to protect human beings in all biomedical research (drugs, biomedical devices, physiology, vaccines,...), and therefore the protocols included in our study are not only on drugs evaluation.
The major difficulties encountered concerned definitions. Although the law clearly defines what are direct approvals and revisions, some RECs evocated direct approval with revision and postponed approval without revision. However this issue is not related to our study methods but to a difficulty in understanding and putting into effect the French law. Despite their training, research assistants had difficulties to retrieve information on study design and to categorise it. Some data were often missing in the protocols, such as the expected duration.
The four articles previously published collected information on only one or two ethics committees[1-4]. These studies did not describe the general situation at a country level and they provided information only when the investigator was not lost for follow-up. Only one study gave information on RECs' activity [1], namely some information on the number of new applications, correspondence, decisions, and the number of meetings needed to obtain approval were given.
One of our major results is that 46% of revisions concerned patient information and consent.
The above article showed that when studies were conditionally approved, deferred or rejected, most queries also concerned the patient information leaflet (85%). But only protocols with five or more centres were assessed; consequently these protocols were not representative of all biomedical research on human beings.
A French report [17] also gave such information, but the aim was more to point out problems of putting the law into effect in France.
Almost all protocols anticipate recruitment in terms of sample size rather than in terms of study duration. We think that the first point shows compliance with good clinical practices [18], and the latter expresses the need for a better training of prime investigators for clinical research study management and anticipation.
The workload theme in terms of numbers of protocols, number of revisions and number of amendments is very relevant since French RECs' members work for free during or after their workday, and at least administrative and financial support of the REC structure is obviously needed in order to guarantee independence in each country. The results of our study are fully supported by the recommendations of the Ad Hoc advisory group on the operation of NHS Research Ethics Committees [19], asking for independent opinion, managed operating system, time recognition and protection, as well as finances.
Across countries, RECs may not have the same structure (frequency of meetings, number of people involved in the committee), the same remit (only studies with intervention on human or also studies on physiology, for instance) and the same functioning since the number of protocols reported in the above article was very heterogeneous in the different countries. When looked at in terms of protocols approved per year and per committee, clear differences appeared: 110 protocols approved in the multicentre REC of London [1], 158 protocols in Barcelona [4], 180 protocols in Oxford [2], 94 protocols in Sydney [3] and on average 39 protocols in France.
It would be of great importance to launch a similar study in more recent years and in other settings at country level to see if things have changed, and if so in which way. Studies in different European countries will allow to collect information which would be very useful to ease harmonisation. It would also be interesting to launch a prospective recording of all tasks and time needed on a random sample of committees.
Conclusion
Up to now, one study described the activity of one REC specialised in multicentre trials and three other studies were carried out on the fate of protocols approved by one committee but nothing was known at a nation level.
Our study showed that 976 protocols were approved in one year by half of the French RECs and that median approval time ranges from 16 days (if no modification) to 48 days. Moreover, most protocols were carried out in France only (79%) for drugs evaluation (68%)
Revisions before approval relates first and foremost to patient information and consent modalities (46%). For a protocol evaluation (first evaluation, revisions and amendments), scientific and administrative workload varied on average from 21.3 hours up to 31.7 hours according to protocols' characteristics.
Competing interests
The author(s) declare that they have no competing interest.
Authors' contributions
ED coordinated the study, managed the data, performed the statistical analysis and drafted the manuscript.
VL participated in the design of the study, coordinated the study and managed the data.
FC designed, submitted and coordinated the study, interpreted data, and helped to draft the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Pr Yves Matillon and Pr Christian Hervé for their advice on drafting the protocol, Pr François Lemaire from the French National Clinical research Program (PHRC), Pr Terry Stacey, Central Office for Research Ethics Committees UK, for his advices on this paper, Marie-Pierre Rochette and Françoise Leclet, administrative staff of REC Lyon, Patricia Darnand and William Banga, members of the logistic staff; and the chairpersons and members of the participating French RECs: Alsace, Aulnay, Auvergne, Bordeaux, Boulogne, Brest, Dijon, Loire, Lyon A, Lyon B, Lyon C, Marseille1, Marseille2, Montpellier, Nice, Normandie, Paris-Bicetre, Paris-Creteil, Paris-Hotel dieu, Paris-Necker, Paris-Versailles, Poitou, Toulouse1, Toulouse2, Tours. Our views are not necessarily theirs.
Grant: French Ministry of Health (Programme Hospitalier de Recherche Clinique 1998 -065) and Hospices Civils de Lyon. French Ministry of Research and Higher Education and Claude Bernard University.
==== Refs
Boyce M Observational study of 353 applications to London multicentre research ethics committee 1997-2000 Bmj 2002 325 1081 12424169 10.1136/bmj.325.7372.1081
Easterbrook PJ Matthews DR Fate of research studies J R Soc Med 1992 85 71 76 1538384
Stern JM Simes RJ Publication bias: evidence of delayed publication in a cohort study of clinical research projects Bmj 1997 315 640 645 9310565
Pich J Carne X Arnaiz JA Gomez B Trilla A Rodes J Role of a research ethics committee in follow-up and publication of results Lancet 2003 361 1015 1016 12660062 10.1016/S0140-6736(03)12799-7
Loi 88-1138 du 20 Décembre 1988 dite Loi Huriet-Sérusclat. Loi relative à la protection des personnes qui se prêtent à des recherches biomédicales
Directive 2001/20/EC of the European Parliament and of the Council of 4 april 2001 on the approximation of the laws, regulations and administrative provisions of the Member States relating to the implementation of good clinical practice in the conduct of clinical trials on medicinal products for human use 2001 , Official journal L121, 01/05/2001 34 44
Kaplan-Meier Non parametric estimation from incomplete observations J Am Stat Assoc 1958 53 457 481
Benzécri JP L'analyse des données. 2, L'analyse des correspondances : introduction, théorie, applications diverses notamment à l'analyse des questionnaires, programmes de calcul 1981 Paris, Dunod
Cattel RB The scree test for the number of factors Mult Behavioral Research 1966 1 p245 276
Benzécri JP Sur le calcul des taux d'inertie dans l'analyse d'un questionnaire Cahiers de l'Analyse des Données 1979 3 p377 378
Ward JH Hierarchical grouping to optimize an objective function J Am Stat Assoc 1963 58 236 244
Cubic Clustering Criterion 1983 SAS technical report A-108 , SAS institute Inc., Cary, North Carolina 56
Zar JH Biostatistical analysis 1998 Fourth , Pearson Higher Education
Pico F Analyse à 6 ans de 168 projets financés dans le cadre du premier Programme Hospitalier de Recherche Clinique (PHRC 93) Thèse de médecine 1999 Université Paris XI,
Duffet JP Rapport d'évaluation du Programme Hospitalier de Recherche Clinique 1994 2001 , Ministère de l'emploi et de la solidarité- Ministère délégué à la santé
Duffet JP Rapport d'évaluation du Programme Hospitalier de Recherche Clinique 1995 2003 , Ministère de la santé, de la famille et des personnes handicapées
Huriet C Rapport d'information fait au nom de la commission des affaires sociales sur le fonctionnement des comités consultatifs de protection des personnes dans la recherche biomédicale 2001 , Sénat
Guideline for Good Clinical Practice 1996 , International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use
Report of the Ad Hoc advisory group on the operation of NHS Research Ethics Committees 2005 , Department of Health
|
16229743
|
PMC1323331
|
CC BY
|
2021-01-04 16:31:58
|
no
|
BMC Med Ethics. 2005 Oct 17; 6:9
|
utf-8
|
BMC Med Ethics
| 2,005 |
10.1186/1472-6939-6-9
|
oa_comm
|
==== Front
PLoS Comput BiolPLoS Comput. BiolpcbiplcbploscompPLoS Computational Biology1553-734X1553-7358Public Library of Science San Francisco, USA 1673870410.1371/journal.pcbi.001007605-PLCB-PV-0321plcb-01-07-11PerspectivesBioinformatics - Computational BiologyResourcesBioinformaticsOntologySemantic WebTime to Organize the Bioinformatics Resourceome PerspectivesCannata Nicola Merelli Emanuela Altman Russ B ** To whom correspondence should be addressed. E-mail: [email protected] Cannata is at the Centro Ricerca Interdipartimentale Biotecnologie Innovative, Università di Padova, Padova, Italy and in the Dipartimento di Matematica e Informatica, Università di Camerino, Camerino, Italy. Emanuela Merelli is in the Dipartimento di Matematica e Informatica, Università di Camerino, Camerino, Italy. Russ B. Altman is in the Department of Genetics, Stanford University Medical Center, Stanford, California, United States of America.
12 2005 30 12 2005 1 7 e76Copyright: © 2005 Cannata 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. Citation:Cannata N, Merelli E, Altman RB (2005) Time to organize the bioinformatics resourceome. PLoS Comput Biol 1(7): e76.
==== Body
We will be witnessing the birth of the artificial, or in-silico, scientist. —J. D. Wren [1]
The field of bioinformatics has blossomed in the last ten years, and as a result, there is a large and increasing number of researchers generating computational tools for solving problems relevant to biology. Because the number of artifacts has increased greatly, it is impossible for many bioinformatics researchers to track tools, databases, and methods in the field—or even perhaps within their own specialty area. More critically, however, biologist users and scientists approaching the field do not have a comprehensive index of bioinformatics algorithms, databases, and literature annotated with information about their context and appropriate use. We suggest that the full set of bioinformatics resources—the “resourceome”—should be explicitly characterized and organized. A hierarchical and machine-understandable organization of the field, along with rich cross-links (an ontology!) would be a useful start. It is likely that a distributed development approach would be required so that those with focused expertise can classify resources in their area, while providing the metadata that would allow easier access to useful existing resources.
The growth of bioinformatics can be quantified in many ways. The Intelligent Systems for Molecular Biology Meeting began in 1993, and numerous other meetings have been established. The International Society for Computational Biology (ISCB) was formed in 1995, and recent membership numbers have reached 2,000. The field has gone from having one or two journals to having more than a dozen—if one considers “-omics” (i.e., subjects relating to high-throughput functional genomics, where computation plays a central role) and the emerging field of systems biology. Because bioinformatics has a strong element of engineering, the creation and maintenance of tools provide value only insofar as they are used. These tools may be databases that hold biological data, or they may be algorithms that act on this data to draw inferences. Access to these artifacts is currently uneven. Of course, the published literature is the archival resting place for the initial description of these innovations, but it only contains a snapshot of most tools early in their lifetime. The literature does not use any standard classification system to describe tools, so the sensitivity of searches for specific functions is not generally high. Indeed, the bibliome itself is idiosyncratically organized, and finding the right article is often like searching for a needle in a haystack [2]. Finally, the published literature does not contain reliable references to the location and to the availability of most bioinformatics resources [3,4]. One could also argue that Google (http://www.google.com) provides adequate access to tools based on keyword searching [5]. However, the lack of standard terms makes sensitive and specific searches difficult. In addition, most search hits confound papers, Web sites, tools, departments, and people in a manner that makes extracting useful information very difficult.
Recognizing this limitation, there have been some grassroots attempts to organize the bioinformatics resourceome. Among the most famous are the “archaeological” Pedro's List—a list of computer tools for molecular biologists (http://www.public.iastate.edu/~pedro/research_tools.html)—and the Expasy Life Sciences Directory, formerly known as the Amos's WWW links page (http://www.expasy.org/links.html). The Bioinformatics Links Directory (http://www.bioinformatics.ubc.ca/resources/links_directory/) today contains more than 700 curated links to bioinformatics resources, organized into eleven main categories, including all the databases and Web servers yearly listed in the dedicated Nucleic Acids Research special issues [6]. The National Center for Biotechnology Institute has tried to make access to its suite of tools transparent, with moderate success. Many Web sites can be found listing “useful sites,” especially concerning special interest or limited topics (e.g., microarrays, text mining, and gene regulation). But all of these efforts are limited by the difficulty in maintaining currency and by the lack of a uniformly recognized classification scheme. Yet our colleagues in bioinformatics and biology are constantly asking about the availability of tools or databases with certain characteristics. The lack of a useful index, thus, routinely costs time and opportunities. In addition, there is no “peer-review” system for bioinformatics tools so that the most useful ones can be highlighted by happy users. A secure and reliable system for rating (similar to that used by Amazon.com, for example) would also be an important prerequisite.
An “ontology” is a specification of a conceptual space, often used by computer programs. The field of ontology engineering has matured in the last 20 years, making fundamental contributions in computer science and establishing applications in biology. The success of the Gene Ontology Project (it is used by multiple model organism databases, and is used to annotate high-throughput data routinely [8]) is one example of an ontology that was developed for the narrow purpose of supporting comparative genomics, but which has found a multitude of other uses. A primitive bioinformatics-specific ontology is available in Google Directory (http://directory.google.com/Top/Science/Biology/Bioinformatics), assembled in the collaborative Open Directory effort (http://www.dmoz.org), but it, too, mixes all different classes of objects (personal Web sites, organization Web sites, databases, and tools) in a way that is not transparent. It seems clear that a well-organized and intuitive ontology of bioinformatics resources would provide a very valuable framework on which a fully distributed system of registration and annotation of biology-related computational resources could be constructed. The Transparent Access to Multiple Bioinformatics Information Sources (TAMBIS) [9] work was a bold attempt to describe bioinformatics concepts, including resources, using formal description languages. Unfortunately, it has not been widely used, perhaps because it was ahead of its time or because the underlying knowledge representation techniques are somewhat sophisticated and complex.
In the foreseeable future the web of links between documents, databases, and programs can provide a new level of interaction among scientific communities. —J. Hendler [10]
Ontologies are important, but their use is often hindered by the lack of “killer apps” for using them. It is often unclear how to exchange information about ontologies, and how to link them to other resources on the Web. Emerging technologies that contribute important infrastructures to the resourceome are represented by the semantic Web and Web services. It is now possible to have standardized descriptors of Web resources, using an ontology, in order to “publish” the availability of tools or simply to announce their existence. Thus, the vision for using an ontology to support the resourceome becomes clear: each individual who has created or who is maintaining a resource uses a standard ontology to describe the basic features of that particular resource using the semantic Web, and these are automatically included in a distributed index of resources. Thus, the index is created by querying the semantic net for descriptions of all available tools, which can then be registered and updated on a regular basis. The development of a browser for this index could be the final step (or “killer app”) in building a self-sustaining, distributed index of bioinformatics resources. Adoption of agent technology may be helpful in overcoming the inherent complexity of this challenge [11].
We believe that the need for a bioinformatics resourceome project and the technical requirements for it are both present. We therefore urge the community to come together to start the process of creating a simple distributed system for describing resources, announcing their availability, and presenting this information to biologists and bioinformaticians in an easy-to-navigate manner. The World Wide Web Consortium already launched its first workshop on Semantic Web for Life Sciences, bringing together more than 100 participants from academia, industry, and international organizations. Another important event is the recent creation of the National Center for Biomedical Ontology (http://www.bioontology.org).
The initial steps toward a bioinformatics resourceome are clear. First, an overall ontology with the high-level concepts (algorithms, databases, organizations, papers, people, etc.) must be created, with a set of standard attributes and a standard set of relations between these concepts (e.g., people publish papers, papers describe algorithms or databases, organizations house people, etc.). The initial ontology should be compact and built for distributed collaborative extension. Second, a mechanism for people to extend this ontology with subconcepts in order to describe their own resources should be designed. The precise location of a tool within a taxonomy is not critical—the author will place it somewhere based on the location of similar/competing resources or based on a best-informed guess. Others may create links to the resource from other appropriate locations in the taxonomy in order to ensure that competing interpretations of the appropriate conceptual location for the resource are accommodated. Third, the formats for the ontologies and the resource descriptions should be published so enterprising software engineers can create interfaces for surfing, searching, and viewing the resources. The resulting distributed system of resource descriptions would be extensible, robust, and useful to the entire biomedical research community.
Who can take leadership in this effort? We believe that a coalition of publishers with an open-access ethic, funding agencies, and scientists who want to contribute to an improved computational infrastructure for biomedicine would be most effective. Companies with an interest in cost-effective research and development may also want to be involved. Most likely, a small group of devoted scientists with both biological domain knowledge and understanding of semantic Web technologies must take the lead. A critical mass of resources must be indexed so that the value of the effort can be assessed. Most likely, the initial indexing will not include all possible resources, but rather algorithms and databases. The community can decide later if Web sites, publications, people, and institutions should also be indexed. The system should also include from the start a capability for routinely evaluating sites for availability (no 404s!). There is increasing discussion of the requirements and technologies for the resourceome at bioinformatics conferences, including Intelligent Systems for Molecular Biology (http://ismb2006.cbi.cnptia.embrapa.br), Pacific Symposium on Biocomputing (http://psb.stanford.edu), and others (see http://www.iscb.org).
==== Refs
References
Wren JD 2004 The emerging in-silico scientist: How text-based bioinformatics is bridging biology and artificial intelligence IEEE Eng Med Biol Mag 23 87 93 15264477
Grivell L 2002 Mining the bibliome: Searching for a needle in a haystack? EMBO Rep 3 200 203 11882534
Wren JD 2004 404 not found: The stability and persistence of URLs published in MEDLINE Bioinformatics 20 668 672 15033874
Schilling LM Wren JD Dellavalle RP 2004 Letter to the editor: Bioinformatics leads charge by publishing more Internet addresses in abstracts than any other journal Bioinformatics 20 2903 15231533
Brin S Page L 1998 The anatomy of a large-scale hypertextual web search engine Comput Netw 30 107 117
Fox JA Butland SL McMillan S Campbell G Ouellette BF 2005 The bioinformatics links directory: A compilation of molecular biology web servers Nucleic Acids Res 33 W3 W24 15980476
Berners-Lee T Hendler J 2001 Publishing on the semantic web Nature 410 1023 1024 11323639
The Gene Ontology Consortium 2000 Gene ontology: Tool for the unification of biology Nat Genet 25 25 29 10802651
Stevens R Baker P Bechhofer S Ng G Jacoby A 2000 TAMBIS: Transparent access to multiple bioinformatics information sources Bioinformatics 16 184 185 10842744
Hendler J 2003 Science and the semantic web Science 299 520 521 12543958
Berners-Lee T Hendler J Lassila O 2001 The Semantic Web Sci Am 284 34 43 11396337
Neumann E 2005 A life science Semantic Web: Are we there yet? Sci STKE 283 pe22
|
16738704
|
PMC1323464
|
CC BY
|
2021-01-05 09:18:23
|
no
|
PLoS Comput Biol. 2005 Dec 30; 1(7):e76
|
utf-8
|
PLoS Comput Biol
| 2,005 |
10.1371/journal.pcbi.0010076
|
oa_comm
|
==== Front
PLoS Comput BiolPLoS Comput. BiolpcbiplcbploscompPLoS Computational Biology1553-734X1553-7358Public Library of Science San Francisco, USA 1675800410.1371/journal.pcbi.001007705-PLCB-MI-0335plcb-01-07-12Message from ISCBBioinformatics - Computational BiologyNoneThe Eighth Annual Bio-Ontologies Meeting Message from ISCBLord Phillip *Stevens Robert Butler James A McEntire Robin * To whom correspondence should be addressed. E-mail: [email protected] Stevens is at the University of Manchester, Manchester, United Kingdom. Phillip Lord is at the University of Newcastle upon Tyne, Newcastle, United Kingdom. Robin McEntire and James A. Butler are at GlaxoSmithKline, King of Prussia, Pennsylvania, United States of America.
12 2005 30 12 2005 1 7 e77Copyright: © 2005 Stevens 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. Citation:Stevens R, McEntire R, Lord P, Butler JA (2005) The eighth annual Bio-Ontologies Meeting. PLoS Comput Biol 1(7): e77.
==== Body
Scientific Mission
The annual Bio-Ontologies Meeting has been operating for eight years; during this time, it has stimulated discussion about the role of ontologies and their associated technologies for structuring, sharing, analysing, and searching knowledge about biological systems. The format varies; usually, it has a mixture of talks and structured discussions such as panel sessions. It regularly draws a hundred participants, showing a consistently high interest across the Intelligent Systems for Molecular Biology (ISMB) community.
The meetings were initially characterised by evangelism for the use of ontoloiges and by arguments about the nature of ontology and which representation was best. Now the debate centres on how to use and how to improve new ontologies.
Hot Topics
Ontologies have an increasing presence in bioinformatics, particularly since the Gene Ontology Project demonstrated the value of supplementing genomic annotation with ontology terms. This interest from the biomedical sciences is intersecting with the interest from scientists and technologies in general. The Semantic Web vision now being actively promoted as the “next generation” Web—which aims to make the Web open to automatic computational use—makes heavy use of ontological technology (http://www.w3.org/2001/sw); this, in turn, is leading to increasing provision of mature tools and experience, lessening the activation energy for those wishing to develop or use ontologies.
One of the themes of this year's meeting, highlighted by Mark Musen's keynote address, are the attempts to move bio-ontology development to a much larger scale (further information about this talk and others mentioned in this article are available from the Bio-Ontologies Web site). Ontology development is often carried out by small groups of individuals, mirroring much of biology before high-throughput technologies. With the large numbers of ontologies now available (for examples, see http://obo.sourceforge.net), orthogonality, maintenance, and consistency are becoming key issues. Currently, there is a relatively poor understanding of both appropriate best practises and the technology that will be required to support these best practices: what, for instance, are the best mechanisms for peer review of ontologies; is centralised management necessary or are more decentralised approaches possible?
Many of the research talks also touched on these issues. Several people discussed applications of automated reasoning techniques: enabling complex querying over data from the yeast community [1] or moving towards the automation of protein classification as part of the process of genome annotation [2]. Ontologies are increasingly being used statistically, often to augment or refine experimental results—in the case of Vailaya et al. [3], this was microarray data. Several talks described new techniques for ontological engineering, new logics better able to describe change in biological systems [4], or more formal treatment of pathological anatomical features [5].
This year, the Bio-Ontologies Meeting was also able to hold a well-attended poster session. Several of the posters described new warehousing environments [6,7] or tools [8,9]. Finally, a number of posters described existing projects from several viewpoints [10,11,12].
Historically, the main use of ontologies within biology was to enable a de facto integration between different data sources by providing a common vocabulary. There are, however, now recognised to be many uses beyond the provision of vocabulary. The use of ontologies within bioinformatics started as complex schema for knowledge databases such as EcoCyc (http://www.ecocyc.org)—a use which continues to this day. The use of ontologies in data analysis is also becoming commonplace.
As with much of bioinformatics, we see a great interest in bio-ontologies from the computer science community. Bio-ontologies offer a variety of large, rapidly changing examples of ontologies with which knowledge representation techniques, methodologies, and tools can be developed and tested. At the Bio-Ontologies Meeting, this relationship has grown from mutual incomprehension to, by and large, a useful symbiosis.
Breaking News
Pacific Symposium on Biocomputing 2006.
This year, the Pacific Symposium on Biocomputing will feature a new Semantic Web track, called Semantic Webs for Life Sciences. This includes a tutorial, a paper session, and panel/discussion session. The conference will be held on 3–7 January 2006 in Maui (http://psb.stanford.edu/cfp.html; http://www.cs.man.ac.uk/~stevensr/events.html).
ISMB 2006.
One of the most pleasant surprises about the 2005 Bio-Ontologies Meeting was the increase in the number of paper submissions, about 30 papers. We were lucky to be able to accommodate so much excellent work with the late introduction of a poster session. Given this increase in submissions, for next year, we hope to modify the publication process—we would like authors to have more space to explain their science than we have currently been able to provide.
Finally, we are currently investigating ways to interact better with other special interest groups. The programme at ISMB is now very full, with many different special interest groups providing excellent science, although with conflicting schedules. For 2006, we are investigating coordinating with BioLink; the synergy between ontological representation and natural language techniques is a natural one. This should ensure that attendees get maximal benefit from both programmes. We welcome any input (E-mail: [email protected]).
The organisers gratefully acknowledge GlaxoSmithKline for funding the keynote speaker for this workshop.
Abbreviation
ISMBIntelligent Systems for Molecular Biology
==== Refs
References
Baker C Witte R Shaban-Nejad A Butler G Haarslev V The FungalWeb ontology: Application scenarios [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Wolstencroft K Lord P Tabernero L Brass A Stevens R Using ontology reasoning to classify protein phosphatases [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Vailaya A Kuchinsky A Kincaid R Adler A Tabibiazar R Ontology-based statistical analysis of microarray data [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Ramakrishnan N Antoniotti M Mishra B Reconstructing formal temporal logic models of cellular events using the GO process ontology [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Smith B Kumar A On the proper treatment of pathologies in biomedical ontologies [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Dewey C Downes A Chou H Zhang S ExperiBase—An object model implementation for biology [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Karp P BioWarehouse: A bioinformatics database warehouse toolkit [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Stephens S Musen M A novel ontology development environment for the life sciences [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Zheng W Shegogue D Integration of the Gene Ontology into an object-oriented architecture [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Cary M Luciano J New and views: The BioPAX pathway data exchange format [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Harris M Why go there? Ensuring that the Gene Ontology meets biologists' needs [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
Whetzel P Parkinson H Sansone A Taylor C Stoeckert C FuGO: Development of a functional genomics ontology [abstract] 2005 8th Annual Bio-Ontologies Meeting 2005 24 June Detroit, United States of America Available: http://bio-ontologies.man.ac.uk . Accessed 13 December 2005.
|
16758004
|
PMC1323465
|
CC BY
|
2021-01-05 09:18:23
|
no
|
PLoS Comput Biol. 2005 Dec 30; 1(7):e77
|
utf-8
|
PLoS Comput Biol
| 2,005 |
10.1371/journal.pcbi.0010077
|
oa_comm
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.