Date Published: February 2, 2017
Publisher: Public Library of Science
Author(s): Jihoon Kim, Chisato Shimizu, Stephen F. Kingsmore, Narayanan Veeraraghavan, Eric Levy, Andre M. Ribeiro dos Santos, Hai Yang, Jay Flatley, Long Truong Hoang, Martin L. Hibberd, Adriana H. Tremoulet, Olivier Harismendy, Lucila Ohno-Machado, Jane C. Burns, Ludmila Prokunina-Olsson.
Kawasaki disease (KD) is the most common acquired pediatric heart disease. We analyzed Whole Genome Sequences (WGS) from a 6-member African American family in which KD affected two of four children. We sought rare, potentially causative genotypes by sequentially applying the following WGS filters: sequence quality scores, inheritance model (recessive homozygous and compound heterozygous), predicted deleteriousness, allele frequency, genes in KD-associated pathways or with significant associations in published KD genome-wide association studies (GWAS), and with differential expression in KD blood transcriptomes. Biologically plausible genotypes were identified in twelve variants in six genes in the two affected children. The affected siblings were compound heterozygous for the rare variants p.Leu194Pro and p.Arg247Lys in Toll-like receptor 6 (TLR6), which affect TLR6 signaling. The affected children were also homozygous for three common, linked (r2 = 1) intronic single nucleotide variants (SNVs) in TLR6 (rs56245262, rs56083757 and rs7669329), that have previously shown association with KD in cohorts of European descent. Using transcriptome data from pre-treatment whole blood of KD subjects (n = 146), expression quantitative trait loci (eQTL) analyses were performed. Subjects homozygous for the intronic risk allele (A allele of TLR6 rs56245262) had differential expression of Interleukin-6 (IL-6) as a function of genotype (p = 0.0007) and a higher erythrocyte sedimentation rate at diagnosis. TLR6 plays an important role in pathogen-associated molecular pattern recognition, and sequence variations may affect binding affinities that in turn influence KD susceptibility. This integrative genomic approach illustrates how the analysis of WGS in multiplex families with a complex genetic disease allows examination of both the common disease–common variant and common disease–rare variant hypotheses.
Although the ability to generate whole genome sequences (WGS) from individual subjects has existed for over a decade, the use of such methods to discover novel disease-causing variants in multiplexed families affected by complex genetic disease has been limited . However, robust methods have been developed to identify rare, disease causative variants in WGS of families with monogenic diseases [2, 3]. Furthermore, such methods have proven useful in identifying individual patients with common diseases that are caused by rare, highly penetrant variants . Individuals with rare Mendelian forms of common complex diseases are often distinguished by extreme phenotypes–such as very early onset, or disease refractory to usual treatments–and by multiple affected members in single families . Here we sought to test this hypothesis in a multiplexed family with Kawasaki disease (KD).
The genomes of all six family members were sequenced with paired, short reads to an aligned mean read depth of 33.2-fold. Unique nucleotide variants (8,018,553) were identified in the family, of which 7,592,729 were of high quality (Fig 1). We created three filter pipelines for further analysis. First, we applied the following filters: recessive homozygous only in the affected children, potentially deleterious (located in an exon, promoter region, splice site, or 3’UTR), and rare (allele frequency <1% or not available in 1,000 Genomes database). To these 34 variants in 31 genes, we applied two additional filters: gene found in KD pathway (defined by Ingenuity Pathway Analysis) and differentially expressed (p<0.05) in our KD transcriptome database . This identified a CAG repeat variant (CAG10 homozygosity in the two affected siblings) in myocyte enhancer factor 2A (MEF2A). The reference allele for this variant is CAG11 and the variant was predicted to be deleterious. The deletion was confirmed by resequencing all six family members (S1 Table). Analysis of WGS of an African Americans family with two affected and two unaffected siblings and their unaffected, biologic parents highlighted genetic variation in TLR6 in KD susceptibility. These TLR6 variants included both compound heterozygosity for two rare, likely deleterious SNVs and homozygosity for common KD risk SNVs. Subsequently, using an acute KD whole blood transcriptome data set, eQTL analysis of the common SNVs suggested decreased transcript levels of IL6 and higher ESR at diagnosis in individuals homozygous for the risk allele. This integrative genomic approach illustrates how WGS in families with multiple members affected with a complex genetic disease can yield insights into both the common disease–common variant and common disease–rare variant hypotheses. Source: http://doi.org/10.1371/journal.pone.0170977