|Home | About | Journals | Submit | Contact Us | Français|
Heart failure is a leading cause of death of people in South Asia, and cardiomyopathy is a major cause of heart failure. Myosin binding protein C (MYBPC3) is expressed in the heart muscle, where it regulates the cardiac response to adrenergic stimulation and is important for the structural integrity of the sarcomere. Mutations in the MYBPC3 gene are associated with hypertrophic or dilated cardiomyopathies. A 25-base-pair deletion in intron 32 causes skipping of the downstream exon and is associated with familial cardiomyopathy. To date, this deletion is found primarily in India and South Asia, although it is also found at low frequency in Southeast Asia. In order to better characterize the distribution of this variant, we determined its frequency in 447 individuals from 19 populations, including 10 populations from India and neighboring populations from Pakistan and Nepal. The deletion frequency is over 8% in some of our Indian samples, and it is not present in any of the populations we sampled outside of India. The differences in the deletion frequencies among populations in India are consistent with patterns of variation previously reported and with patterns we observed among Indian populations based on high-density SNP chip data. Our results indicate the MYBPC3 deletion is primarily found among Indian populations, and that its distribution is consistent with genome-wide patterns of variation in India.
Heart failure is a global disease. One important cause of heart failure, cardiomyopathy, leads to premature death among individuals over 40 years of age and affects an estimated 1–2% of the Western population (Berry et al., 2001). Studies based on human subjects, animal models, and gene expression analyses have elucidated hundreds of cardiomyopathy-associated mutations in more than 20 genes (Liew & Dzau, 2004). Several of these genes are involved in pathways essential for the development and structural integrity of the heart muscle. Despite ongoing genetic investigation, little is understood about the causes and prevalence of cardiomyopathy among individuals in developing countries.
Studies of previously underrepresented populations have helped elucidate the genetic basis of cardiomyopathy (Kubo, 2005), and indicate that it is relatively common among individuals of South Asian ancestry (Joshi, 2007, Reddy et al., 2005). Recently, Dhandapany et al. (2009) reported that a specific cardiomyopathy-associated genetic variant, a deletion in myosin binding protein C 3 (MYBPC3), is primarily limited to South Asia. As with most cardiomyopathy-associated mutations, many carriers of this deletion exhibit late onset of disease (Dhandapany et al., 2009). MYBPC3 is expressed in the heart muscle, where it regulates the cardiac response to adrenergic stimulation and is important for organization of the myocardium. The South-Asian-specific MYBPC3 mutation is a 25-basepair (bp) deletion in the intron upstream of exon 33 (Figure 1). Transcriptional studies indicate that this deletion leads to skipping of exon 33 (Dhandapany et al., 2009), and it is possible that accumulation of the altered protein may contribute to late-onset cardiomyopathy (Kubo, 2005, Niimura, 1998).
In order to further characterize patterns of variation of this deletion, we determined its frequency in 447 Asians, including populations from India and surrounding regions, East Asia, and Southeast Asia (Figure 2). A previous report indicated a high prevalence of this mutation in India (approximately 4%) (Dhandapany et al., 2009). Here, we provide a description of the South-Asian-specific distribution of the MYBPC3 deletion in populations from and near India, highlighting differences in its prevalence among samples from 10 different tribal and caste populations from the southern state of Andhra Pradesh in India. To better place this variation in perspective, we compare it to genetic variation assayed using high-density single nucleotide polymorphism (SNP) microarrays.
DNA samples were obtained from unrelated individuals with informed consent as previously described (Watkins et al., 2005, Watkins et al., 2003, Xing et al., 2009). Subjects included in this study are permanent residents of their respective regions with ancestors having lived in that region. Each population is designated according to language or geographic location, and in the case of all Indian samples, by caste or tribal group. The DNA used for this study was collected from the following Indian populations from Andhra Pradesh: Brahmin (52), Kshatriya (11), Vysya (8), Kapu (51), Yadava (44), Madiga (27), Mala (24), Relli (19), Khonda Dora (25), and the Irula (26). Asian individuals from the following populations were included for additional analysis: Pakistan (25), Nepal (25), Iraq (25), Kyrgyzstan (25), Thailand (25), Vietnam (5), Cambodia (7), China (10), and Japan (13). We examined mitochondrial DNA and Y chromosome polymorphisms for each of the populations to verify ancestry. Genome-wide single nucleotide polymorphism (SNP) analysis was performed on Brahmin, Mala, Madiga, and Irula populations within India and all individuals from regions outside of India.
We amplified the 25-bp deletion region using polymerase chain reaction (PCR) under standard conditions with 5’-GTTTCCAGCCTTGGGCATAGTC-3’ and 5’-GAGGACAACGGAGCAAAGCCC-3’ primers. The amplicons were run on 3% MetaPhor Agarose gel at 120 volts for two hours to determine presence or absence of the deletion. Allele frequencies were determined by direct counting and upper and lower binomial confidence limits were determined at the α = 0.05 level. Fisher’s exact tests were used to compare allele frequencies between each population and all the remaining populations.
Approximately 240,000 SNP genotypes for Andhra Brahmin, Madiga, Mala, and Irula and all populations outside of India were obtained from Xing et al. 2009. Per-locus pairwise FST was calculated for the set of SNPs genotyped in all populations as previously described (Xing et al., 2009).
The deletion is present in most of our Indian population samples but absent in the populations we sampled outside of India. Interestingly, the allele frequency varies considerably within India. The deletion is absent in two of ten Indian populations and is present with an allele frequency greater than 8% in three of the populations. The allele frequency in the Yadava (middle caste) population are significantly higher than the remaining Indian populations (Fisher’s exact test, P = 0.01). However, the allele frequencies in the other Indian populations do not significantly deviate from the rest of the populations (Table 1).
The Khonda Dora tribal population, which exhibit genetic similarity to Southeast Asians, has a relatively high allele frequency of the deletion (10%), whereas the mutation is completely absent in the Irula sample, another tribal group from Andhra Pradesh. More than 8% of the Mala and Yadava (lower and middle castes, respectively) also have the deletion; all other Indian populations sampled, including upper-caste Brahmins, exhibit heterozygote frequencies less than five percent. The distribution of the deletion allele frequencies in all of our populations sampled from India is shown in Table 1.
In order to determine whether the variation in deletion allele frequencies within India is unusual, we calculated pairwise FST between four Indian populations using ~240,000 genome-wide SNPs and compared these values to the deletion pairwise FST in the same four Indian populations. All deletion FST values were less than those calculated for the genome-wide average, with the exception of comparisons made with the Mala. While pairwise FST values were higher than the genome average in all Mala comparisons (0.013–0.069), values fell within the 95% confidence interval of the genome-wide range of FST (the upper 95% confidence interval is between 0.10 and 0.165). Therefore, the FST for the deletion is not significantly outside the normal range of genetic variation for populations within India, even in the most extreme case of the Mala, and the population allele frequency distribution of the 25-bp deletion is typical of the variation observed throughout the genome in India.
The MYBPC3 25-bp deletion is similar in effect to other MYBPC3 mutations, which manifest later in life, lead to mild hypertrophy, and exhibit incomplete penetrance (Kubo, 2005, Niimura, 1998). Based on a family study of this deletion, ~90% of the oldest family members who carried the deletion were symptomatic (Dhandapany et al., 2009). Although individuals with this deletion typically present with cardiomyopathy at more than 30 years of age, homozygous individuals are affected with severe phenotypes approximately a decade earlier (Dhandapany et al., 2009). The risk of developing cardiomyopathy is probably also increased by other genetic and environmental factors (Dhandapany et al., 2009).
It is interesting that the MYBPC3 25-bp deletion is relatively common in India and its frequency varies across this limited region. The time to the most recent common ancestor (TMRCA) estimate, based on previous haplotype analysis, is ~33,000 (−/+23,000) years before present (Dhandapany et al., 2009), suggesting that the deletion might have arisen in India and was possibly not present in the initial inhabitants that arrived in this region from Africa some 50,000 to 20,000 years ago (Misra, 1992). Previous reports indicate it reaches highest frequency in southern and western Indian states (2% to 8%) and is absent from India populations sampled in the Northeast, recent migrant populations originally from Africa, and a group from the Andaman Islands (Dhandapany et al., 2009).
The MYBPC3 deletion allele frequency we report here (~5%) and the variation we observe among our Indian samples is consistent with the previous study (Dhandapany et al., 2009). One of the middle caste groups, the Yadava, has significantly different allele frequencies compared to other populations, potentially due to the effects of genetic drift. Interestingly, the Irula tribal group we sampled does not have the deletion, whereas the Khonda Dora, another tribal group from Andhra Pradesh, exhibits the second highest allele frequency among our ten populations (10%). The dramatic difference among tribal populations is possibly due to genetic drift and, in the case of the Irula samples, extreme isolation from other Indian populations.
The MYBPC3 deletion patterns we describe here are also concordant with reported patterns of genetic variation and divergence based on neutral genetic markers. The genetic affinity of Indian populations has been previously analyzed using Alus, short-tandem repeats (STRs), and genome-wide SNP data (Bamshad et al., 2001, Basu et al., 2003, Roychoudhury et al., 2001, Thanseem et al., 2006, Watkins et al., 2005, Watkins et al., 2003, Xing et al., 2009, Reich et al., 2009). SNP microarray and STR studies show relative homogeneity among caste populations, whereas tribal Indian populations are more distinct from caste populations and each other (Watkins et al., 2005, Xing et al., 2009). The MYBPC3 deletion FST values lie in the normal genome-wide range, suggesting that the same factors that shape genome-wide patterns are also shaping the distribution of this deletion.
It is somewhat surprising that a deleterious mutation would be found at such a high frequency. Considering the region surrounding the MYBPC3 locus is gene-rich (over 15 genes in the 500 Kb region), it is possible that the MYBPC3 deleterious variant hitch-hiked along with a linked region that was subjected to positive selection. Dhandapany et al. (2009) reported that there is no evidence of deviation from neutral evolution for the 25-bp deletion region. Their conclusion is based on two analyses performed in individuals both with and without the deletion, which indicate: 1) similar amounts of variation observed for five STR loci spanning 3.4 megabases (Mb) in more than 250 individuals examined, and 2) no significant difference in sequence variation 5 kilobases on either side of the deleted region in nine heterozygous individuals.
Interestingly, there are a relatively large number of mutations in the MYBPC3 locus (over 150 identified thus far) that have deleterious effects in Western and Asian populations (http://cardiogenomics.med.harvard.edu/home Harvard CardioGenomics). The recombination rate in this region is exceedingly low (0.08 cm/Mb) compared to the genome-wide average (1.3cM/Mb) (McVean et al., 2004). It is possible that the disadvantageous MYBPC3 mutations accumulated in these populations faster than they can be removed by purifying selection due to the low recombination rate (Gillespie, 2004, Muller, 1964). While it is possible that variants in this region have been influenced by positive selection or have not been removed by purifying selection, such hypotheses require further investigation.
We thank J.M. Naidu, B.B Rao, B.V. Prasad, and M.J. Bamshad for assistance with sample collection and A.R. Rogers for helpful discussion. This work was supported by NSF grants SBR-9514733, SBR-9512178, and NIH grant GM-59290.