Genome-wide Significance and Replication of the Chromosome 12p11.22

Summary

Peripartum cardiomyopathy (PPCM) is a rare condition of unknown etiology that occurs in late pregnancy (last month) or early postpartum (within 5 months after delivery) (ref 1). Incidence of PPCM is estimated to be one in 3,000-4,000 live births in the US. Presumed risk factors include maternal age >30 years, multiparity, multiple gestation, history of preeclampsia, and use of tocolytics. PPCM pathophysiology may be distinct from idiopathic dilated cardiomyopathy, but this is uncertain and recent studies suggest that genetic factors may influence PPCM.

This study evaluated and replicated genome-wide association of single nucleotide polymorphisms (SNPs) with PPCM. The study was conducted first in a population of 41 cases and 49 local controls (post-menopausal age-discordant females) as well as 654 additional controls (aged 30-84). Genotyping was performed using the Illumina 550K Bead chip microarray (550.000 SNPs). A replication study was performed in a second population of 30 cases of PPCM and a population of 124 age discordant controls or 89 younger controls or 90 obstetrical controls.

 In the genome-wide association study (GWAS) one SNP (rs258415) met genome-wide significance for PPCM versus controls (p=2.06 x 10-8, OR=5.96, 95% CI 3.13 to 11.38). The result was replicated in the second independent populations (p=0.009 to 0.029 according to the control populations, OR=2.15 to 2.44). This locus is observed at chromosome 12p11.22, the region of linkage disequilibrium being a 30 kilobase region. The SNP of interest (rs258415) is located in an intergenic region near two genes, KLHDC5 (possibly involved in immune response), and PTHLH (a member of parathyroid hormone family involved in calcium transfer in placenta and uterus where it regulates blood flow, and also is involved in the modulation of ventricular contraction and pacing of sinus node).
 The authors concluded that this genome-wide association study discovered and replicated a locus associated with PPCM, at chromosome 12p11.22.

Comments

The pathophysiology of peripartum cardiomyopathy remains unclear but recent studies suggested that it may partially overlap with dilated cardiomyopathy (DCM).   Several studies in 2010 reported on a substantial number of DCM families including PPCM patients. Alternatively, the screening of relatives of PPCM patients may identify a number of DCM patients. Finally a mutation screening in PPCM patients may identify some causal mutations in genes responsible for conventional DCM such as MYH7 and TNNC1 genes (ref 2, 3).  These cases are however quite rare.

The present study of Horne et al. provides new data strongly suggesting a role for genetic factors, as a component of a multifactorial determinant, in peripartum cardiomyopathy. The frequency of the minor allele is quite high in the control population (0.16 to 0.21 in Caucasian) and the related odds ratio for PPCM is very high (>5) which may appear as unusual since high OR are typically related with rare SNP whereas frequent alleles are associated with low OR. An important limitation of the study is in fact the low sample size that may have allowed uncontrollable GWAS lab testing, allele assignment, or genotype clustering/calling errors to drive the results. We should notice however that several GWAS were previously reported in relatively small populations including one of the first GWAS results in macular degeneration (ref 4) that indicates well characterized patients and also an important effect (OR) of the genetic polymorphisms. Obviously such a study performed in a small population cannot identify a number of SNPs associated with a mild effect on the disease.

Conclusion:

Several conclusions can be drawn from the present study. (1) Peripartum cardiomyopathy may be a good example of interactions between genetic background and acquired / environmental factors (pregnancy and related hormonal factors). (2) GWAS study performed in unorthodox population of very limited size may allow statistically significant results. (3) As frequently observed the SNP of interest is located in an intergenic region without clear evidence at that stage for a precise underlying gene (as the original publication about chromosome 9p21 in coronary artery disease, ref 5) in contrast with rare GWAS reports of coding SNP directly in a precise gene (as the BAG3 gene recently reported as associated with DCM, ref 6). The present study provides a new chromosomal locus for further pathophysiological and clinical investigation.