Concentrations of liver enzymes in plasma are widely used as indicators of liver disease. We carried out a genome-wide association study in 61,089 individuals, identifying 42 loci associated with concentrations of liver enzymes in plasma, of which 32 are new associations (P = 10−8 to P = 10−190). We used functional genomic approaches including metabonomic profiling and gene expression analyses to identify probable candidate genes at these regions. We identified 69 candidate genes, including genes involved in biliary transport (ATP8B1 and ABCB11), glucose, carbohydrate and lipid metabolism (FADS1, FADS2, GCKR, JMJD1C, HNF1A, MLXIPL, PNPLA3, PPP1R3B, SLC2A2 and TRIB1), glycoprotein biosynthesis and cell surface glycobiology (ABO, ASGR1, FUT2, GPLD1 and ST3GAL4), inflammation and immunity (CD276, CDH6, GCKR, HNF1A, HPR, ITGA1, RORA and STAT4) and glutathione metabolism (GSTT1, GSTT2 and GGT), as well as several genes of uncertain or unknown function (including ABHD12, EFHD1, EFNA1, EPHA2, MICAL3 and ZNF827). Our results provide new insight into genetic mechanisms and pathways influencing markers of liver function.

Background

Genome-wide association (GWA) studies have identified several susceptibility loci for metabolic syndrome (MetS) component traits, but have had variable success in identifying susceptibility loci to the syndrome as an entity. We conducted a GWA study on MetS and its component traits in four Finnish cohorts consisting of 2637 MetS cases and 7927 controls, both free of diabetes, and followed the top loci in an independent sample with transcriptome and NMR-based metabonomics data. Furthermore, we tested for loci associated with multiple MetS component traits using factor analysis and built a genetic risk score for MetS.

Methods and Results

A previously known lipid locus, APOA1/C3/A4/A5 gene cluster region (SNP rs964184), was associated with MetS in all four study samples (P=7.23×10−9 in meta-analysis). The association was further supported by serum metabolite analysis, where rs964184 associated with various VLDL, TG, and HDL metabolites (P=0.024-1.88×10−5). Twenty-two previously identified susceptibility loci for individual MetS component traits were replicated in our GWA and factor analysis. Most of these associated with lipid phenotypes and none with two or more uncorrelated MetS components. A genetic risk score, calculated as the number of alleles in loci associated with individual MetS traits, was strongly associated with MetS status.

Conclusions

Our findings suggest that genes from lipid metabolism pathways have the key role in the genetic background of MetS. We found little evidence for pleiotropy linking dyslipidemia and obesity to the other MetS component traits such as hypertension and glucose intolerance.

Association testing of multiple correlated phenotypes offers better power than univariate analysis of single traits. We analyzed 6,600 individuals from two population-based cohorts with both genome-wide SNP data and serum metabolomic profiles. From the observed correlation structure of 130 metabolites measured by nuclear magnetic resonance, we identified 11 metabolic networks and performed a multivariate genome-wide association analysis. We identified 34 genomic loci at genome-wide significance, of which 7 are novel. In comparison to univariate tests, multivariate association analysis identified nearly twice as many significant associations in total. Multi-tissue gene expression studies identified variants in our top loci, SERPINA1 and AQP9, as eQTLs and showed that SERPINA1 and AQP9 expression in human blood was associated with metabolites from their corresponding metabolic networks. Finally, liver expression of AQP9 was associated with atherosclerotic lesion area in mice, and in human arterial tissue both SERPINA1 and AQP9 were shown to be upregulated (6.3-fold and 4.6-fold, respectively) in atherosclerotic plaques. Our study illustrates the power of multi-phenotype GWAS and highlights candidate genes for atherosclerosis.

Author Summary

In this study, we aim to identify novel genetic variants for metabolism, characterize their effects on nearby genes, and show that the nearby genes are associated with metabolism and atherosclerosis. To discover new genetic variants, we use an alternative approach to traditional genome-wide association studies: we leverage the information in phenotype covariance to increase our statistical power. We identify variants at seven novel loci and then show that our top signals drive expression of nearby genes AQP9 and SERPINA1 in multiple tissues. We demonstrate that AQP9 and SERPINA1 gene expression, in turn, is associated with metabolite levels. Finally, we show that the genes are associated with atherosclerosis using mouse atherosclerotic lesion size (AQP9) as well as tissue from healthy human arteries and atherosclerotic plaques (AQP9 and SERPINA1). This study illustrates that multivariate analysis of correlated metabolites can boost power for gene discovery substantially. Further functional work will need to be performed to elucidate the biological role of SERPINA1 and AQP9 in atherosclerosis.

Objective

Recent genome-wide association studies (GWAS) identified a variant rs7575840 in the apolipoprotein B (APOB) gene region to be associated with LDL-C. However, the underlying functional mechanism of this variant that resides 6.5 kb upstream of APOB has remained unknown. Our objective was to investigate rs7575840 for association with refined apoB containing lipid particles; for replication in a non-Caucasian Mexican population; and for underlying functional mechanism.

Methods and Results

Our data show that rs7575840 is associated with serum apoB levels (P=4.85×10−10) and apoB containing lipid particles, very small VLDL, IDL and LDL particles (P=2×10−5 - 9×10−7) in the Finnish METSIM study sample (n=7,710). Fine mapping of the APOB region using 43 SNPs replicated the association of rs7575840 with apoB in a Mexican study sample (n=2,666, P=3.33×10−05). Furthermore, our transcript analyses of adipose RNA samples from 175 Finnish METSIM subjects indicate that rs7575840 alters expression of APOB (P=1.13×10−10) and a regional non-coding RNA (BU630349) (P=7.86×10−6) in adipose tissue.

Conclusions

It has been difficult to convert GWAS associations into mechanistic insights. Our data show that rs7575840 is associated with serum apoB levels and apoB containing lipid particles as well as influences expression of APOB and a regional transcript BU630349 in adipose tissue. We thus provide evidence how a common genome-wide significant SNP rs7575840 may affect serum apoB, LDL-C, and TC levels.

OBJECTIVE

We investigated the effects of 34 genetic risk variants for hyperglycemia/type 2 diabetes on lipoprotein subclasses and particle composition in a large population-based cohort.

RESEARCH DESIGN AND METHODS

The study included 6,580 nondiabetic Finnish men from the population-based Metabolic Syndrome in Men (METSIM) study (aged 57 ± 7 years; BMI 26.8 ± 3.7 kg/m2). Genotyping of 34 single nucleotide polymorphism (SNPs) for hyperglycemia/type 2 diabetes was performed. Proton nuclear magnetic resonance spectroscopy was used to measure particle concentrations of 14 lipoprotein subclasses and their composition in native serum samples.

RESULTS

The glucose-increasing allele of rs780094 in GCKR was significantly associated with low concentrations of VLDL particles (independently of their size) and small LDL and was nominally associated with low concentrations of intermediate-density lipoprotein, all LDL subclasses, and high concentrations of very large and large HDL particles. The glucose-increasing allele of rs174550 in FADS1 was significantly associated with high concentrations of very large and large HDL particles and nominally associated with low concentrations of all VLDL particles. SNPs rs10923931 in NOTCH2 and rs757210 in HNF1B genes showed nominal or significant associations with several lipoprotein traits. The genetic risk score of 34 SNPs was not associated with any of the lipoprotein subclasses.

CONCLUSIONS

Four of the 34 risk loci for type 2 diabetes or hyperglycemia (GCKR, FADS1, NOTCH2, and HNF1B) were significantly associated with lipoprotein traits. A GCKR variant predominantly affected the concentration of VLDL, and the FADS1 variant affected very large and large HDL particles. Only a limited number of risk loci for hyperglycemia/type 2 diabetes significantly affect lipoprotein metabolism.

Recent genome-wide association (GWA) studies described 95 loci controlling serum lipid levels. These common variants explain ∼25% of the heritability of the phenotypes. To date, no unbiased screen for gene–environment interactions for circulating lipids has been reported. We screened for variants that modify the relationship between known epidemiological risk factors and circulating lipid levels in a meta-analysis of genome-wide association (GWA) data from 18 population-based cohorts with European ancestry (maximum N = 32,225). We collected 8 further cohorts (N = 17,102) for replication, and rs6448771 on 4p15 demonstrated genome-wide significant interaction with waist-to-hip-ratio (WHR) on total cholesterol (TC) with a combined P-value of 4.79×10−9. There were two potential candidate genes in the region, PCDH7 and CCKAR, with differential expression levels for rs6448771 genotypes in adipose tissue. The effect of WHR on TC was strongest for individuals carrying two copies of G allele, for whom a one standard deviation (sd) difference in WHR corresponds to 0.19 sd difference in TC concentration, while for A allele homozygous the difference was 0.12 sd. Our findings may open up possibilities for targeted intervention strategies for people characterized by specific genomic profiles. However, more refined measures of both body-fat distribution and metabolic measures are needed to understand how their joint dynamics are modified by the newly found locus.

Author Summary

Circulating serum lipids contribute greatly to the global health by affecting the risk for cardiovascular diseases. Serum lipid levels are partly inherited, and already 95 loci affecting high- and low-density lipoprotein cholesterol, total cholesterol, and triglycerides have been found. Serum lipids are also known to be affected by multiple epidemiological risk factors like body composition, lifestyle, and sex. It has been hypothesized that there are loci modifying the effects between risk factors and serum lipids, but to date only candidate gene studies for interactions have been reported. We conducted a genome-wide screen with meta-analysis approach to identify loci having interactions with epidemiological risk factors on serum lipids with over 30,000 population-based samples. When combining results from our initial datasets and 8 additional replication cohorts (maximum N = 17,102), we found a genome-wide significant locus in chromosome 4p15 with a joint P-value of 4.79×10−9 modifying the effect of waist-to-hip ratio on total cholesterol. In the area surrounding this genetic variant, there were two genes having association between the genotypes and the gene expression in adipose tissue, and we also found enrichment of association in genes belonging to lipid metabolism related functions.

Background

Variations in the pattern of molecular associations are observed during disease development. The comprehensive analysis of molecular association patterns and their changes in relation to different physiological conditions can yield insight into the biological basis of disease-specific phenotype variation.

Methodology

Here, we introduce a formal statistical method for the differential analysis of molecular associations via network representation. We illustrate our approach with extensive data on lipoprotein subclasses measured by NMR spectroscopy in 4,406 individuals with normal fasting glucose, and 531 subjects with impaired fasting glucose (prediabetes). We estimate the pair-wise association between measures using shrinkage estimates of partial correlations and build the differential network based on this measure of association. We explore the topological properties of the inferred network to gain insight into important metabolic differences between individuals with normal fasting glucose and prediabetes.

Conclusions/Significance

Differential networks provide new insights characterizing differences in biological states. Based on conventional statistical methods, few differences in concentration levels of lipoprotein subclasses were found between individuals with normal fasting glucose and individuals with prediabetes. By performing the differential analysis of networks, several characteristic changes in lipoprotein metabolism known to be related to diabetic dyslipidemias were identified. The results demonstrate the applicability of the new approach to identify key molecular changes inaccessible to standard approaches.

Diabetic kidney disease, diagnosed by urinary albumin excretion rate (AER), is a critical symptom of chronic vascular injury in diabetes, and is associated with dyslipidemia and increased mortality. We investigated serum lipids in 326 subjects with type 1 diabetes: 56% of patients had normal AER, 17% had microalbuminuria (20 ≤ AER < 200 μg/min or 30 ≤ AER < 300 mg/24 h) and 26% had overt kidney disease (macroalbuminuria AER ≥ 200 μg/min or AER ≥ 300 mg/24 h). Lipoprotein subclass lipids and low-molecular-weight metabolites were quantified from native serum, and individual lipid species from the lipid extract of the native sample, using a proton NMR metabonomics platform. Sphingomyelin (odds ratio 2.53, P < 10−7), large VLDL cholesterol (odds ratio 2.36, P < 10−10), total triglycerides (odds ratio 1.88, P < 10−6), omega-9 and saturated fatty acids (odds ratio 1.82, P < 10−5), glucose disposal rate (odds ratio 0.44, P < 10−9), large HDL cholesterol (odds ratio 0.39, P < 10−9) and glomerular filtration rate (odds ratio 0.19, P < 10−10) were associated with kidney disease. No associations were found for polyunsaturated fatty acids or phospholipids. Sphingomyelin was a significant regressor of urinary albumin (P < 0.0001) in multivariate analysis with kidney function, glycemic control, body mass, blood pressure, triglycerides and HDL cholesterol. Kidney injury, sphingolipids and excess fatty acids have been linked in animal models—our exploratory approach provides independent support for this relationship in human patients with diabetes.

Electronic supplementary material

The online version of this article (doi:10.1007/s11306-011-0343-y) contains supplementary material, which is available to authorized users.

A protocol for determination of oxidation susceptibility of serum lipids based on proton nuclear magnetic resonance (1H NMR) spectroscopy is presented and compared to the commonly used spectrophotometric method. Even though there are methodological differences between these two methods, the NMR-based oxidation susceptibility correlates well (r2 = 0.73) with the lag time determined spectrophotometrically. In addition to the oxidizability of serum lipids, the NMR method provides also information about the lipid profile. The NMR oxidation assay was applied to the chocolate study including fasting serum samples (n = 45) from subjects who had consumed white (WC), dark (DC) or high-polyphenol chocolate (HPC) daily for 3 weeks. The oxidation susceptibility of serum lipids decreased in the HPC group, and there was a significant difference between the WC and HPC groups (P = 0.031). According to the random forest analysis, the consumption of the HPC chocolate induced changes to the amounts of HDL, phosphatidylcholine, sphingomyelin, and nervonic, docosahexaenoic and myristic acids. Furthermore, arachidonic, docosahexaenoic, docosapentaenoic and palmitic acids, gamma-glutamyl transferase, hemoglobin, HDL, phosphatidylcholine and choline containing phospholipids explained about 60% of the oxidation susceptibility values.

Electronic supplementary material

The online version of this article (doi:10.1007/s11306-011-0323-2) contains supplementary material, which is available to authorized users.

The lipid–leukocyte (LL) module is associated with, and reactive to, a wide variety of serum metabolites.The LL module appears to be a link between metabolism, adiposity, and inflammation.Serum metabolite concentrations themselves determine the connectedness of LL module.

Comprehensive characterization of human tissues promises novel insights into the biological architecture of human diseases and traits. We assessed metabonomic, transcriptomic, and genomic variation for a large population-based cohort from the capital region of Finland. Network analyses identified a set of highly correlated genes, the lipid–leukocyte (LL) module, as having a prominent role in over 80 serum metabolites (of 134 measures quantified), including lipoprotein subclasses, lipids, and amino acids. Concurrent association with immune response markers suggested the LL module as a possible link between inflammation, metabolism, and adiposity. Further, genomic variation was used to generate a directed network and infer LL module's largely reactive nature to metabolites. Finally, gene co-expression in circulating leukocytes was shown to be dependent on serum metabolite concentrations, providing evidence for the hypothesis that the coherence of molecular networks themselves is conditional on environmental factors. These findings show the importance and opportunity of systematic molecular investigation of human population samples. To facilitate and encourage this investigation, the metabonomic, transcriptomic, and genomic data used in this study have been made available as a resource for the research community.

Subtle metabolic changes precede and accompany chronic vascular complications, which are the primary causes of premature death in diabetes. To obtain a multimetabolite characterization of these high-risk individuals, we measured proton nuclear magnetic resonance (1H NMR) data from the serum of 613 patients with type I diabetes and a diverse spread of complications. We developed a new metabonomics framework to visualize and interpret the data and to link the metabolic profiles to the underlying diagnostic and biochemical variables. Our results indicate complex interactions between diabetic kidney disease, insulin resistance and the metabolic syndrome. We illustrate how a single 1H NMR protocol is able to identify the polydiagnostic metabolite manifold of type I diabetes and how its alterations translate to clinical phenotypes, clustering of micro- and macrovascular complications, and mortality during several years of follow-up. This work demonstrates the diffuse nature of complex vascular diseases and the limitations of single diagnostic biomarkers. However, it also promises cost-effective solutions through high-throughput analytics and advanced computational methods, as applied here in a case that is representative of the real clinical situation.

Background

A key challenge in metabonomics is to uncover quantitative associations between multidimensional spectroscopic data and biochemical measures used for disease risk assessment and diagnostics. Here we focus on clinically relevant estimation of lipoprotein lipids by 1H NMR spectroscopy of serum.

Results

A Bayesian methodology, with a biochemical motivation, is presented for a real 1H NMR metabonomics data set of 75 serum samples. Lipoprotein lipid concentrations were independently obtained for these samples via ultracentrifugation and specific biochemical assays. The Bayesian models were constructed by Markov chain Monte Carlo (MCMC) and they showed remarkably good quantitative performance, the predictive R-values being 0.985 for the very low density lipoprotein triglycerides (VLDL-TG), 0.787 for the intermediate, 0.943 for the low, and 0.933 for the high density lipoprotein cholesterol (IDL-C, LDL-C and HDL-C, respectively). The modelling produced a kernel-based reformulation of the data, the parameters of which coincided with the well-known biochemical characteristics of the 1H NMR spectra; particularly for VLDL-TG and HDL-C the Bayesian methodology was able to clearly identify the most characteristic resonances within the heavily overlapping information in the spectra. For IDL-C and LDL-C the resulting model kernels were more complex than those for VLDL-TG and HDL-C, probably reflecting the severe overlap of the IDL and LDL resonances in the 1H NMR spectra.

Conclusion

The systematic use of Bayesian MCMC analysis is computationally demanding. Nevertheless, the combination of high-quality quantification and the biochemical rationale of the resulting models is expected to be useful in the field of metabonomics.