Copy number variation in metabolic phenotypes

Despite successes in identifying genetic contributors to common metabolic phenotypes, only part of the heritable component of these traits has thus far been explained. Copy number variation (CNV) is likely to be responsible for some of the unexplained variation. As observed with single nucleotide changes, it is probable that both rare and common CNVs will contribute to susceptibility to metabolic disease. For instance, CNVs in the LDLR gene underlie a substantial portion of disease in patients with heterozygous familial hypercholesterolemia. As well, a common CNV in LPA encoding apolipoprotein(a) is the primary determinant of plasma lipoprotein(a) concentrations, a risk factor for atherosclerosis. Recent efforts to map CNVs in control populations have defined their size, frequency and distribution. Many of the identified CNVs overlap genes with important functions in metabolic pathways. The overlap of CNVs that were found in control datasets with functional candidate genes or genes with previous evidence of association with metabolic syndrome presents an important subset for future CNV association studies. Finally, we describe an approach to search for CNVs in a rare high-penetrance metabolic disorder, namely lipodystrophy. As methods to identify CNVs increase in precision and accuracy, the prospect of identifying their role in both rare Mendelian and common complex metabolic phenotypes will become a reality.

COPY NUMBER VARIATION IN METABOLIC SYNDROME

Metabolic syndrome (MetS) is defined by the presence of abdominal obesity, hypertension, dysglycemia, and dyslipidemia. Many mutations have been discovered that cause rare monogenic components of metabolic syndrome, and association studies have linked common variants with increased risk of MetS and its components. Despite successes in identifying genetic contributors to metabolic syndrome, unexplained heritability exists and copy number variation (CNV) could be responsible for a portion of this variation. As observed with single nucleotide changes, it is likely that both rare and common CNVs will contribute to MetS disease susceptibility. Recent efforts to map CNVs in control populations have given insight into their size, frequency and distribution. However, despite being observed in controls, the reported CNVs could still modulate susceptibility for late-onset complex traitsor produce subtle metabolic phenotypes. Here we examine the overlap between CNVs found in control datasets and genes with functional hypotheses or evidence of previous association to MetS. Secondly, we present the results and methodology of a search for a rare CNV in a high-penetrance Mendelian disorder, namely familial partial lipodystrophy. As methods to identify CNVs increase in precision and accuracy, the prospect of identifying their role in both rare Mendelian and common complex diseases is exciting.

A genome-wide linkage scan for familial partial lipodystrophy susceptibility genes in a German kindreds

Background: In Dunnigan-type familial partial lipodystrophy (FPLD) patients are born with normal fat distribution, but subcutaneous fat from extremities and gluteal regions are lost during puberty. The abnormal fat distribution leads to the development of metabolic syndrome (MetS), a cluster of phenotypes including hyperglycemia, dyslipidemia, hypertension, and visceral obesity. The study of FPLD as a monogenic model of MetS may uncover genetic risk factors of the common MetS which affects ~30% of adult North Americans. Two molecular forms of FPLD have been identified including FPLD2, resulting from heterozygous mutations in the LMNA gene, and FPLD3, resulting from both heterozygous dominant negative and haploinsufficiency mutations in the PPARG gene. However, many patients with clinically diagnosed FPLD have no mutation in either LMNA or PPARG, suggesting the involvement of additional genes in FPLD etiology. Methods: Here, we report the results of an Affymetrix 10K GeneChip microarray genome-wide linkage analysis study of a German kindred displaying the FPLD phenotype and no known lipodystrophy-causing mutations. Results: The investigation identified three chromosomal loci, namely 1q, 3p, and 9q, with non-parametric logarithm of odds (NPL) scores >2.7. While not meeting the criteria for genome-wide significance, it is interesting to note that the 1q and 3p peaks contain the LMNA and PPARG genes respectively. Conclusions: Three possible conclusions can be drawn from these results: 1) the peaks identified are spurious findings, 2) additional genes physically close to LMNA, PPARG, or within 9q, are involved in FPLD etiology, or 3) alternative disease causing mechanisms not identified by standard exon sequencing approaches, such as promoter mutations, alternative splicing, or epigenetics, are also responsible for FPLD.

Brain-derived neurotrophic factor variants are associated with childhood-onset mood disorder: confirmation in a Hungarian sample

Brain-derived neurotrophic factor (BDNF) is a nerve growth factor that has been implicated in the neurobiology of depression. Our group has previously reported an association between a BDNF variant and childhood-onset mood disorder (COMD) in an adult sample from Pittsburgh. We hypothesize that variants at the BDNF locus are associated with COMD. Six BDNF polymorphisms were genotyped in 258 trios having juvenile probands with childhood-onset DSM-IV major depressive or dysthymic disorder. BDNF markers included the (GT)n microsatellite, Val66Met and four other single-nucleotide polymorphisms (SNPs) distributed across the BDNF gene. Family-based association and evolutionary haplotype analysis methods were used. Analysis of linkage disequilibrium (LD) revealed substantial LD among all six polymorphisms. Analyses of the Val66Met polymorphism demonstrated significant overtransmission of the val allele (chi2=7.12, d.f.=1, P=0.0076). Consistent with the pattern of LD, all other SNPs showed significant biased transmission. The (GT)n microsatellite alleles also indicated a trend towards biased transmission (170 bp: Z=2.095, P=0.036). Significant haplotypes involved Val66Met and BDNF2 (P=0.0029). In this Hungarian sample, we found all five BDNF SNPs tested and a haplotype containing the BDNF Val66Met Val allele to be associated with COMD. These results provide evidence that BDNF variants affect liability to juvenile-onset mood disorders, supported by data from two independent samples.

Genome scan of Arab Israeli families maps a schizophrenia susceptibility gene to chromosome 6q23 and supports a locus at chromosome 10q24

Schizophrenia is a complex neuropsychiatric disorder to which an as-yet-unknown number of genes contribute, interacting with each other and the environment. Linkage analyses have implicated several chromosomal regions as harboring schizophrenia susceptibility loci although rarely at levels commensurate with proposed thresholds for genome-wide significance. We systematically recruited Arab Israeli families multiply affected with schizophrenia from the catchment area of a Regional Mental Health Center. Clinical diagnoses were established by semistructured interviews and all other available sources of information under narrow, core and broad categories. Using 350 microsatellite markers, spaced at an average of 10.3 cM, we performed an autosomal scan in 155 subjects from 21 families. Linkage analysis employed affects only, multipoint, nonparametric (model-free) and also parametric (dominant and recessive) approaches. We detected significant evidence for a schizophrenia susceptibility gene at chromosome 6q23 with a nonparametric LOD score (NPL) of 4.60 (P=0.000004) under the broad diagnostic category and a parametric LOD score of 3.33 (dominant model). Under the core diagnostic category the NPL was 4.29 (P=0.00001) and the LOD score 4.16 (dominant model). We also detected suggestive evidence for linkage at chromosome 10q24 under the broad diagnostic category (NPL 3.24, P=0.0008; heterogeneity LOD score, dominant model 2.65, alpha=0.82). Additionally, NPL scores >2.0 were observed at chromosome 2q37, 4p15-16, 7p22, 9q21-22 and 14q11.1-11.2. The linkage we detected at chromosome 6q23 fulfills the criteria for genome-wide significance and is located approximately midway between loci suggested by a previous significant report at chromosome 6q25 and findings located more centromerically at 6q21-22.