Exome sequencing-driven discovery of coding polymorphisms associated with common metabolic phenotypes

AIMS/HYPOTHESIS

Human complex metabolic traits are in part regulated by genetic determinants. Here we applied exome sequencing to identify novel associations of coding polymorphisms at minor allele frequencies (MAFs) >1% with common metabolic phenotypes.

METHODS

The study comprised three stages. We performed medium-depth (8×) whole exome sequencing in 1,000 cases with type 2 diabetes, BMI >27.5 kg/m(2) and hypertension and in 1,000 controls (stage 1). We selected 16,192 polymorphisms nominally associated (p < 0.05) with case-control status, from four selected annotation categories or from loci reported to associate with metabolic traits. These variants were genotyped in 15,989 Danes to search for association with 12 metabolic phenotypes (stage 2). In stage 3, polymorphisms showing potential associations were genotyped in a further 63,896 Europeans.

RESULTS

Exome sequencing identified 70,182 polymorphisms with MAF >1%. In stage 2 we identified 51 potential associations with one or more of eight metabolic phenotypes covered by 45 unique polymorphisms. In meta-analyses of stage 2 and stage 3 results, we demonstrated robust associations for coding polymorphisms in CD300LG (fasting HDL-cholesterol: MAF 3.5%, p = 8.5 × 10(-14)), COBLL1 (type 2 diabetes: MAF 12.5%, OR 0.88, p = 1.2 × 10(-11)) and MACF1 (type 2 diabetes: MAF 23.4%, OR 1.10, p = 8.2 × 10(-10)).

CONCLUSIONS/INTERPRETATION

We applied exome sequencing as a basis for finding genetic determinants of metabolic traits and show the existence of low-frequency and common coding polymorphisms with impact on common metabolic traits. Based on our study, coding polymorphisms with MAF above 1% do not seem to have particularly high effect sizes on the measured metabolic traits.

A role for coding functional variants in HNF4A in type 2 diabetes susceptibility

AIMS/HYPOTHESIS

Rare mutations in the gene HNF4A, encoding the transcription factor hepatocyte nuclear factor 4α (HNF-4A), account for ~5% of cases of MODY and more frequent variants in this gene may be involved in multifactorial forms of diabetes. Two low-frequency, non-synonymous variants in HNF4A (V255M, minor allele frequency [MAF] ~0.1%; T130I, MAF ~3.0%)-known to influence downstream HNF-4A target gene expression-are of interest, but previous type 2 diabetes association reports were inconclusive. We aimed to evaluate the contribution of these variants to type 2 diabetes susceptibility through large-scale association analysis.

METHODS

We genotyped both variants in at least 5,745 cases and 14,756 population controls from the UK and Denmark. We also undertook an expanded association analysis that included previously reported and novel genotype data obtained in Danish, Finnish, Canadian and Swedish samples. A meta-analysis incorporating all published association studies of the T130I variant was subsequently carried out in a maximum sample size of 14,279 cases and 26,835 controls.

RESULTS

We found no association between V255M and type 2 diabetes in either the initial (p = 0.28) or the expanded analysis (p = 0.44). However, T130I demonstrated a modest association with type 2 diabetes in the UK and Danish samples (additive per allele OR 1.17 [95% CI 1.08-1.28]; p = 1.5 × 10⁻⁴), which was strengthened in the meta-analysis (OR 1.20 [95% CI 1.10-1.30]; p = 2.1 × 10⁻⁵).

CONCLUSIONS/INTERPRETATION

Our data are consistent with T130I as a low-frequency variant influencing type 2 diabetes risk, but are not conclusive when judged against stringent standards for genome-wide significance. This study exemplifies the difficulties encountered in association testing of low-frequency variants.

Testing of diabetes-associated WFS1 polymorphisms in the Diabetes Prevention Program

AIMS/HYPOTHESIS

Wolfram syndrome (diabetes insipidus, diabetes mellitus, optic atrophy and deafness) is caused by mutations in the WFS1 gene. Recently, single nucleotide polymorphisms (SNPs) in WFS1 have been reproducibly associated with type 2 diabetes. We therefore examined the effects of these variants on diabetes incidence and response to interventions in the Diabetes Prevention Program (DPP), in which a lifestyle intervention or metformin treatment was compared with placebo.

METHODS

We genotyped the WFS1 SNPs rs10010131, rs752854 and rs734312 (H611R) in 3,548 DPP participants and performed Cox regression analysis using genotype, intervention and their interactions as predictors of diabetes incidence. We also evaluated the effect of these SNPs on insulin resistance and beta cell function at 1 year.

RESULTS

Although none of the three SNPs was associated with diabetes incidence in the overall cohort, white homozygotes for the previously reported protective alleles appeared less likely to develop diabetes in the lifestyle arm. Examination of the publicly available Diabetes Genetics Initiative genome-wide association dataset revealed that rs10012946, which is in strong linkage disequilibrium with the three WFS1 SNPs (r(2)=0.88-1.0), was associated with type 2 diabetes (allelic odds ratio 0.85, 95% CI 0.75-0.97, p=0.026). In the DPP, we noted a trend towards increased insulin secretion in carriers of the protective variants, although for most SNPs this was seen as compensatory for the diminished insulin sensitivity.

CONCLUSIONS/INTERPRETATION

The previously reported protective effect of select WFS1 alleles may be magnified by a lifestyle intervention. These variants appear to confer an improvement in beta cell function.

The Pro12Ala variant at the peroxisome proliferator-activated receptor gamma gene and change in obesity-related traits in the Diabetes Prevention Program

AIMS/HYPOTHESIS

Peroxisome proliferator-activated receptor gamma (PPARgamma), encoded by the PPARG gene, regulates insulin sensitivity and adipogenesis, and may bind polyunsaturated fatty acids (PUFA) and thiazolidinediones in a ligand-dependent manner. The PPARG proline for alanine substitution at position 12 (Pro12Ala polymorphism) has been related with obesity directly and via interaction with PUFA.

METHODS

We tested the effect-modifying role of Pro12Ala on the 1 year change in obesity-related traits in a randomised clinical trial of treatment with metformin (n = 989), troglitazone (n = 363) or lifestyle modification (n = 1,004) vs placebo (n = 1,000) for diabetes prevention in high-risk individuals.

RESULTS

At baseline, Ala12 carriers had larger waists (p < 0.001) and, in a subset, more subcutaneous adipose tissue (SAT; lumbar 2/3; p = 0.04) than Pro12 homozygotes. There was a genotype-by-intervention interaction on 1-year weight change (p = 0.01); in the placebo arm, Pro12 homozygotes gained weight and Ala12 carriers lost weight (p = 0.001). In the metformin and lifestyle arms, weight loss occurred across genotypes, but was greatest in Ala12 carriers (p < 0.05). Troglitazone treatment induced weight gain, which tended to be greater in Ala12 carriers (p = 0.08). In the placebo group, SAT (lumbar 2/3, lumbar 4/5) decreased in Ala12 allele carriers, but was unchanged in Pro12 homozygotes (p < or = 0.005). With metformin treatment, SAT decreased independently of genotype. In the lifestyle arm, SAT (lumbar 2/3) reductions occurred across genotypes, but were greater in Ala12 carriers (p = 0.03). A genotype-by-PUFA intake interaction on reduction in visceral fat (lumbar 4/5; p = 0.04) was also observed, which was most evident with metformin treatment (p < 0.001).

CONCLUSIONS/INTERPRETATION

Within the Diabetes Prevention Program, the Ala12 allele influences central obesity, an effect which may differ by treatment group and dietary PUFA intake (ClinicalTrials.gov ID no: NCT00004992).

Association testing of common variants in the insulin receptor substrate-1 gene (IRS1) with type 2 diabetes

AIMS/HYPOTHESIS

Activation of the insulin receptor substrate-1 (IRS1) is a key initial step in the insulin signalling pathway. Despite several reports of association of the G972R polymorphism in its gene IRS1 with type 2 diabetes, we and others have not observed this association in well-powered samples. However, other nearby variants might account for the putative association signal.

SUBJECTS AND METHODS

We characterised the haplotype map of IRS1 and selected 20 markers designed to capture common variations in the region. We genotyped this comprehensive set of markers in several family-based and case-control samples of European descent totalling 12,129 subjects.

RESULTS

In an initial sample of 2,235 North American and Polish case-control pairs, the minor allele of the rs934167 polymorphism showed nominal evidence of association with type 2 diabetes (odds ratio [OR] 1.25, 95% CI 1.03-1.51, p = 0.03). This association showed a trend in the same direction in 7,659 Scandinavian samples (OR 1.16, 95% CI 0.96-1.39, p = 0.059). The combined OR was 1.20 (p = 0.008), but statistical correction for the number of variants examined yielded a p value of 0.086. We detected no differences across rs934167 genotypes in insulin-related quantitative traits.

CONCLUSIONS/INTERPRETATION

Our data do not support an association of common variants in IRS1 with type 2 diabetes in populations of European descent.

Common variants in HNF-1 alpha and risk of type 2 diabetes

AIMS/HYPOTHESIS

Mutations in the hepatocyte nuclear factor 1-alpha gene (HNF-1alpha, now known as the transcription factor 1 gene [TCF1]) cause the most common monogenic form of diabetes, MODY3, but it is not known if common variants in HNF-1a are associated with decreased transcriptional activity or phenotypes related to type 2 diabetes, or whether they predict future type 2 diabetes.

SUBJECTS AND METHODS

We studied the effect of four common polymorphisms (rs1920792, I27L, A98V and S487N) in and upstream of the HNF-1alpha gene on transcriptional activity in vitro, and their possible association with type 2 diabetes and insulin secretion in vivo.

RESULTS

Certain combinations of the I27L and A98V polymorphisms in the HNF-1alpha gene showed decreased transcriptional activity on the target promoters glucose transporter 2 (now known as solute carrier family 2 [facilitated glucose transporter], member 2) and albumin in both HeLa and INS-1 cells. In vivo, these polymorphisms were associated with a modest but significant impairment in insulin secretion in response to oral glucose. Insulin secretion deteriorated over time in individuals carrying the V allele of the A98V polymorphism (n = 2,293; p = 0.003). In a new case-control (n = 1,511 and n = 2,225 respectively) data set, the I27L polymorphism was associated with increased risk of type 2 diabetes, odds ratio (OR) = 1.5 (p = 0.002; multiple logistic regression), particularly in elderly (age > 60 years) and overweight (BMI > 25 kg/m(2)) patients (OR = 2.3, p = 0.002).

CONCLUSIONS/INTERPRETATION

This study provides in vitro and in vivo evidence that common variants in the MODY3 gene, HNF-1alpha, influence transcriptional activity and insulin secretion in vivo. These variants are associated with a modestly increased risk of late-onset type 2 diabetes in subsets of elderly overweight individuals.

Positive natural selection in the human lineage

Positive natural selection is the force that drives the increase in prevalence of advantageous traits, and it has played a central role in our development as a species. Until recently, the study of natural selection in humans has largely been restricted to comparing individual candidate genes to theoretical expectations. The advent of genome-wide sequence and polymorphism data brings fundamental new tools to the study of natural selection. It is now possible to identify new candidates for selection and to reevaluate previous claims by comparison with empirical distributions of DNA sequence variation across the human genome and among populations. The flood of data and analytical methods, however, raises many new challenges. Here, we review approaches to detect positive natural selection, describe results from recent analyses of genome-wide data, and discuss the prospects and challenges ahead as we expand our understanding of the role of natural selection in shaping the human genome.

A candidate gene approach to searching for low-penetrance breast and prostate cancer genes

Most cases of breast and prostate cancer are not associated with mutations in known high-penetrance genes, indicating the involvement of multiple low-penetrance risk alleles. Studies that have attempted to identify these genes have met with limited success. The National Cancer Institute Breast and Prostate Cancer Cohort Consortium--a pooled analysis of multiple large cohort studies with a total of more than 5,000 cases of breast cancer and 8,000 cases of prostate cancer--was therefore initiated. The goal of this consortium is to characterize variations in approximately 50 genes that mediate two pathways that are associated with these cancers--the steroid-hormone metabolism pathway and the insulin-like growth factor signalling pathway--and to associate these variations with cancer risk.

A Y chromosomal influence on prostate cancer risk: the multi-ethnic cohort study

BACKGROUND

A Y chromosomal role in prostate cancer has previously been suggested by both cytogenetic findings and patterns of Y chromosomal gene expression. We took advantage of the well established and stable phylogeny of the non-recombining segment of the Y chromosome to investigate the association between Y chromosomal DNA variation and prostate cancer risk.

METHODS

We examined the distribution of 116 Y lineages in 930 prostate cancer cases and 1208 controls from four ethnic groups from a cohort study in Hawaii and California.

RESULTS

One lineage, found only among the Japanese group in our study, was associated with a statistically significant predisposition to prostate cancer (odds ratio (OR) = 1.63; 95% confidence interval (CI) 1.07 to 2.47), and, in particular, to high severity disease in younger individuals (OR = 3.89; 95% CI 1.34 to 11.31).

CONCLUSIONS

This finding suggests that a Y chromosomal factor contributes significantly to the development of prostate cancer in Japanese men.

Genomewide linkage analysis of stature in multiple populations reveals several regions with evidence of linkage to adult height

Genomewide linkage analysis has been extremely successful at identification of the genetic variation underlying single-gene disorders. However, linkage analysis has been less successful for common human diseases and other complex traits in which multiple genetic and environmental factors interact to influence disease risk. We hypothesized that a highly heritable complex trait, in which the contribution of environmental factors was relatively limited, might be more amenable to linkage analysis. We therefore chose to study stature (adult height), for which heritability is approximately 75%-90% (Phillips and Matheny 1990; Carmichael and McGue 1995; Preece 1996; Silventoinen et al. 2000). We reanalyzed genomewide scans from four populations for which genotype and height data were available, using a variance-components method implemented in GENEHUNTER 2.0 (Pratt et al. 2000). The populations consisted of 408 individuals in 58 families from the Botnia region of Finland, 753 individuals in 183 families from other parts of Finland, 746 individuals in 179 families from Southern Sweden, and 420 individuals in 63 families from the Saguenay-Lac-St.-Jean region of Quebec. Four regions showed evidence of linkage to stature: 6q24-25, multipoint LOD score 3.85 at marker D6S1007 in Botnia (genomewide P<.06), 7q31.3-36 (LOD 3.40 at marker D7S2195 in Sweden, P<.02), 12p11.2-q14 (LOD 3.35 at markers D12S10990-D12S398 in Finland, P<.05) and 13q32-33 (LOD 3.56 at markers D13S779-D13S797 in Finland, P<.05). In a companion article (Perola et al. 2001 [in this issue]), strong supporting evidence is obtained for linkage to the region on chromosome 7. These studies suggest that highly heritable complex traits such as stature may be genetically tractable and provide insight into the genetic architecture of complex traits.

A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms

We describe a map of 1.42 million single nucleotide polymorphisms (SNPs) distributed throughout the human genome, providing an average density on available sequence of one SNP every 1.9 kilobases. These SNPs were primarily discovered by two projects: The SNP Consortium and the analysis of clone overlaps by the International Human Genome Sequencing Consortium. The map integrates all publicly available SNPs with described genes and other genomic features. We estimate that 60,000 SNPs fall within exon (coding and untranslated regions), and 85% of exons are within 5 kb of the nearest SNP. Nucleotide diversity varies greatly across the genome, in a manner broadly consistent with a standard population genetic model of human history. This high-density SNP map provides a public resource for defining haplotype variation across the genome, and should help to identify biomedically important genes for diagnosis and therapy.

The role of PPAR-gamma in macrophage differentiation and cholesterol uptake

Peroxisome proliferator-activated receptor-gamma (PPAR-gamma), the transcription factor target of the anti-diabetic thiazolidinedione (TZD) drugs, is reported to mediate macrophage differentiation and inflammatory responses. Using PPAR-gamma-deficient stem cells, we demonstrate that PPAR-gamma is neither essential for myeloid development, nor for such mature macrophage functions as phagocytosis and inflammatory cytokine production. PPAR-gamma is required for basal expression of CD36, but not for expression of the other major scavenger receptor responsible for uptake of modified lipoproteins, SR-A. In wild-type macrophages, TZD treatment divergently regulated CD36 and class A macrophage-scavenger receptor expression and failed to induce significant cellular cholesterol accumulation, indicating that TZDs may not exacerbate macrophage foam-cell formation.

An SNP map of the human genome generated by reduced representation shotgun sequencing

Most genomic variation is attributable to single nucleotide polymorphisms (SNPs), which therefore offer the highest resolution for tracking disease genes and population history. It has been proposed that a dense map of 30,000-500,000 SNPs can be used to scan the human genome for haplotypes associated with common diseases. Here we describe a simple but powerful method, called reduced representation shotgun (RRS) sequencing, for creating SNP maps. RRS re-samples specific subsets of the genome from several individuals, and compares the resulting sequences using a highly accurate SNP detection algorithm. The method can be extended by alignment to available genome sequence, increasing the yield of SNPs and providing map positions. These methods are being used by The SNP Consortium, an international collaboration of academic centres, pharmaceutical companies and a private foundation, to discover and release at least 300,000 human SNPs. We have discovered 47,172 human SNPs by RRS, and in total the Consortium has identified 148,459 SNPs. More broadly, RRS facilitates the rapid, inexpensive construction of SNP maps in biomedically and agriculturally important species. SNPs discovered by RRS also offer unique advantages for large-scale genotyping.

The common PPARgamma Pro12Ala polymorphism is associated with decreased risk of type 2 diabetes

Genetic association studies are viewed as problematic and plagued by irreproducibility. Many associations have been reported for type 2 diabetes, but none have been confirmed in multiple samples and with comprehensive controls. We evaluated 16 published genetic associations to type 2 diabetes and related sub-phenotypes using a family-based design to control for population stratification, and replication samples to increase power. We were able to confirm only one association, that of the common Pro12Ala polymorphism in peroxisome proliferator-activated receptor-gamma(PPARgamma) with type 2 diabetes. By analysing over 3,000 individuals, we found a modest (1.25-fold) but significant (P=0.002) increase in diabetes risk associated with the more common proline allele (85% frequency). Moreover, our results resolve a controversy about common variation in PPARgamma. An initial study found a threefold effect, but four of five subsequent publications failed to confirm the association. All six studies are consistent with the odds ratio we describe. The data implicate inherited variation in PPARgamma in the pathogenesis of type 2 diabetes. Because the risk allele occurs at such high frequency, its modest effect translates into a large population attributable risk-influencing as much as 25% of type 2 diabetes in the general population.

Androgen metabolism and prostate cancer: establishing a model of genetic susceptibility

The prostate is an androgen-regulated organ, which has led to longstanding interest in the role of androgens in prostate carcinogenesis. Although evidence of a hormonal etiology for prostate cancer is strong, it is almost entirely circumstantial. Much of the problem in proving a causal relationship relates to the continued difficulties in reliably measuring human tissue-specific exposure to endogenous steroid hormones. The international and racial-ethnic variations in prostate cancer incidence, combined with the effects of migration on risk patterns, have suggested that genetic factors play a central role in determining prostate cancer risk. We are developing a polygenic model of prostate carcinogenesis, focused around a series of genes involved in androgen biosynthesis, transport and metabolism. We have begun to develop this model by utilizing sequence variants to study how polymorphic markers in two genes (SRD5A2 and AR) are related to prostate cancer risk within and between racial-ethnic groups. We are now collaborating with the Whitehead Institute/MIT, Center for Genome Research, to screen for single nucleotide polymorphisms in additional genes relevant to the androgen pathway and prostate cell growth. The model when fully developed can potentially provide a basis for targeting populations for screening interventions and for implementing primary preventive strategies.

A novel member of the F-box/WD40 gene family, encoding dactylin, is disrupted in the mouse dactylaplasia mutant

Early outgrowth of the vertebrate embryonic limb requires signalling by the apical ectodermal ridge (AER) to the progress zone (PZ), which in response proliferates and lays down the pattern of the presumptive limb in a proximal to distal progression. Signals from the PZ maintain the AER until the anlagen for the distal phalanges have been formed. The semidominant mouse mutant dactylaplasia (Dac) disrupts the maintenance of the AER, leading to truncation of distal structures of the developing footplate, or autopod. Adult Dac homozygotes thus lack hands and feet except for malformed single digits, whereas heterozygotes lack phalanges of the three middle digits. Dac resembles the human autosomal dominant split hand/foot malformation (SHFM) diseases. One of these, SHFM3, maps to chromosome 10q24 (Refs 6,7), which is syntenic to the Dac region on chromosome 19, and may disrupt the orthologue of Dac. We report here the positional cloning of Dac and show that it belongs to the F-box/WD40 gene family, which encodes adapters that target specific proteins for destruction by presenting them to the ubiquitination machinery. In conjuction with recent biochemical studies, this report demonstrates the importance of this gene family in vertebrate embryonic development.

Characterization of single-nucleotide polymorphisms in coding regions of human genes

A major goal in human genetics is to understand the role of common genetic variants in susceptibility to common diseases. This will require characterizing the nature of gene variation in human populations, assembling an extensive catalogue of single-nucleotide polymorphisms (SNPs) in candidate genes and performing association studies for particular diseases. At present, our knowledge of human gene variation remains rudimentary. Here we describe a systematic survey of SNPs in the coding regions of human genes. We identified SNPs in 106 genes relevant to cardiovascular disease, endocrinology and neuropsychiatry by screening an average of 114 independent alleles using 2 independent screening methods. To ensure high accuracy, all reported SNPs were confirmed by DNA sequencing. We identified 560 SNPs, including 392 coding-region SNPs (cSNPs) divided roughly equally between those causing synonymous and non-synonymous changes. We observed different rates of polymorphism among classes of sites within genes (non-coding, degenerate and non-degenerate) as well as between genes. The cSNPs most likely to influence disease, those that alter the amino acid sequence of the encoded protein, are found at a lower rate and with lower allele frequencies than silent substitutions. This likely reflects selection acting against deleterious alleles during human evolution. The lower allele frequency of missense cSNPs has implications for the compilation of a comprehensive catalogue, as well as for the subsequent application to disease association.

Taurine promotes the differentiation of a vertebrate retinal cell type in vitro

The retina offers a model system for investigating the mechanisms that control cell type determination and differentiation in the vertebrate central nervous system. Previously, rod photoreceptor development in vitro was found to require a diffusible activity released by retinal cells (D. Altshuler and C. Cepko, Development 114, 947–957, 1992). In this report, we show that retinal-cell-conditioned medium and extracts contain two separable activities that influence rod development: a &gt; 10 kDa inhibitory activity, and a stimulatory activity that is &lt; 1 kDa and heat stable. Taurine was found to be a component of the &lt; 1 kDa fraction and to stimulate rod development when added to retinal cultures. Taurine was not the only rod-promoting factor in these retinal preparations, however, as conditioned medium and extracts stimulated a higher level of rod development than did taurine alone. Taurine uptake into cells could be blocked without inhibiting taurine's ability to stimulate rod development, arguing against an osmoregulatory or nutritive mechanism of action. Finally, a competitive antagonist of taurine's bioactivity was identified and shown partially to inhibit rod development in retinal explants, suggesting that taurine may normally act to stimulate rod development in the retina. These results provide evidence for three activities, one of which is taurine, that are candidate regulators of rod photoreceptor development in vivo.

A temporally regulated, diffusible activity is required for rod photoreceptor development in vitro

The retina is a relatively simple and well-characterized CNS structure in which cell-cell interactions have been hypothesized to influence cell type determination. By manipulating cell density in serum-free cultures we show that rat rod photoreceptor development requires a diffusible activity produced by neonatal retinal cells. This effect is not mediated by changes in cell survival or mitosis. Production of the rod promoting activity varies with developmental stage and is temporally correlated with the timing of rod generation in vivo. In low density cultures, which do not support rod development, an increased fraction of cells stain with an antibody specific for another retinal neuron, the bipolar cell. Thus, the diffusible rod promoting activity may influence cell fate determination, and not only terminal differentiation. These results provide an approach for the molecular characterization of developmentally important signals in the vertebrate retina.

Control of photoreceptor development

Recent studies of cell type determination in the vertebrate retina suggest that rod photoreceptor development involves interactions among cells that are mediated, at least in part, by temporally regulated diffusible signals. In this review the strategies used to generate rods in the vertebrate retina are compared with those described for photoreceptor development in the Drosophila retina.