Abstract
A recent metabolite genome-wide association study (mGWAS) investigated the relationship between genetic factors and the urine metabolome in kidney disease. The findings demonstrate that mGWAS hold promise for identifying novel genetic factors involved in adsorption, distribution, metabolism and excretion of metabolites and pharmaceuticals, as well as biomarkers for disease progression.
Genome- wide association studies (GWAS) have had limited success in identifying genetic factors that predict kidney disease progression, despite major financial investment and the study of more than 10,000 patients. This lack of success is likely due to inter-individual heterogeneity in kidney disease progression, even in disease subtypes characterized by similar aetiology1. The progression of diabetic kidney disease, for example, is highly variable and the rate of decline in kidney function only partly correlates with urine albumin to creatinine ratio (ACR), blood pressure and glycaemic control2,3.
“Urinary metabolites [are] a readout of systemic processes of metabolite ADME”
In the new era of multi-omics, the identification of genetic variants that are statistically linked to the urinary metabolome is a promising approach for identifying predictive biomarkers for kidney disease4–6. In a recent study, Schlosser et al.7 used a metabolite genome-wide association study (mGWAS) approach to identify genes that are associated with the absorption, distribution, metabolism and excretion (ADME) of endogenous metabolites and pharmaceuticals in patients with chronic kidney disease (CKD). They performed a GWAS in tandem with a robust analysis of the urine metabolome from 1,627 patients with reduced kidney function from the German chronic kidney disease (GCKD) cohort8 and identified 240 unique metabolite quantitative trait loci (mQTLs) — genomic intervals containing at least one single nucleotide polymorphism (SNP) associated with urinary metabolite concentrations.
The threshold for significance in the study was set at P < 4.2 × 10−11, well below the threshold for GWAS alone (usually P < 10−7), establishing that these associations are likely to be more reliable than those from GWAS. The most significant mQTLs — those for which urine metabolite levels were most strongly associated with SNPs — corresponded to the genes PYROXD2 (P = 3.6 × 10−574), NAT8 (P = 2.4 × 10−570) and AKRD7A (P = 2.3 × 10−412), which were associated with metabolites that have yet to be characterized. Untargeted metabolomics studies will likely identify many chemical entities that have not yet been annotated and those with highly significant mQTLs will warrant further investigation and characterization. Indeed, of the 211 unique metabolites linked to the 240 mQTLs identified, 75 have not yet been named. Importantly, the variance in the urine metabolite levels that could be attributed to a single index nucleotide polymorphism ranged from 2.0% to as much as 63.1%, indicating that the genome and urine metabolome are closely related. Finding SNPs strongly linked with urinary metabolites could provide the basis for identifying SNPs linked to disease phenotypes.
The researchers performed further validation in an independent cohort of 977 healthy participants from the population- based SHIP- Trend study. Ninety mQTLs were matched across the GCKD and SHIP- Trend data sets. The researchers found that 70 of the mQTLs identified in the GCKD population were also significant in the SHIP- Trend data, although genetic effects were on average 1.35-fold greater in the population with CKD, indicating that many of these mQTLs may be linked to CKD aetiology.
The researchers then went on to score genes associated with the mQTLs based on their proximity and expression profile. Ninety genes linked with significant mQTLs were strongly enriched for ADME processes, with 35 genes linked to biotransformation reactions involved in drug metabolism (TABLE 1). This result demonstrates the utility of urinary metabolites as a readout of systemic processes of metabolite ADME. Using the published literature and the gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases as well as manual curation, the researchers determined that many of these genes were associated with detoxification and drug metabolism pathways (P < 1 × 10−8) or with the metabolism of small molecules such as carboxylic acids, amino acids and fatty acids. Using an RNA-Seq database that included 38 human tissues (the Genotype- Tissue Expression (GTEx) project), the authors found significant expression of RNAs for the 90 genes in kidney, liver, small intestine and pancreatic tissue, likely reflecting the role these organs play in ADME. Interestingly, expression of these genes was also increased in heart left ventricle tissue, suggesting a new role for this mitochondria- rich tissue in ADME. In the kidney, various segments of the proximal tubule also showed highly enriched expression of these genes, which is consistent with the prominent role of proximal tubular cells in the handling and metabolism of amino acids and organic acids. The study fell short of describing how the enzymes encoded by these genes may impact the metabolism of medications and how this may be linked to kidney disease progression; however, future studies will be able to use data from mGWAS to answer these crucial questions.
“Several metabolites were significantly associated with end- stage kidney disease”
Table 1 |.
Genes linked to biotransformation reactions involved in drug metabolism
| Gene | Protein activity |
|---|---|
| Phase I | |
| ACOT2, ACOT4 | Acyl-CoA thioesterase |
| ADH1A | Alcohol dehydrogenase |
| AKR7A2 | Aldo-keto reductase |
| ALPL | Alkaline phosphatase |
| CYP2C8, CYP2D6, CYP3A5, CYP3A7, CYP4A11, CYP4B1, CYP4F2 | Cytochrome P450 monooxygenase |
| FMO4 | Dimethylaniline monoxygenase |
| Phase II | |
| ACSM2A, ACSM6 | Acyl-CoA synthetase |
| COMT | Catechol methyltransferase |
| GLYATL3 | Glycine N-acyltransferase |
| GSTM2 | Glutathione S-transferase |
| NAT2, NAT8 | N-acetyltransferase |
| NQO2 | NADPH dehydrogenase |
| SULT2A1 | Sulfotransferase |
| UGT2B15, UGT2B7, UGT3A1 | UDP glucuronosyltransferase |
| Phase III | |
| ARSA | Arylsulfatase |
| CDA | Cytidine deaminase |
| DPEP1, NAALAD2 | Dipeptidase |
| FOLH1 | Folate hydrolase, dipeptidase |
| GGT1 | γ-GLutamyltransferase |
| SLCO1B1, SLC22A1, SLC22A10, SLC28A2 | Solute transporter |
Genes selected from the GCKD and SHIP- Trend cohorts. Data from REF.8.
The researchers were able to compare the mGWAS data with well-powered studies of genetic association with disease (UK Biobank). mQTL analysis revealed an association between the gene encoding alkaline phosphatase, ALPL, and urinary phosphoethanolamine, and an SNP in ALPL was associated with urolithiasis in the UK Biobank, suggesting a link between alkaline phosphatase and kidney stone formation. Among 30 NAT8-associated metabolites evaluated for potential links with CKD progression, several metabolites were significantly associated with end- stage kidney disease; however, it is unclear whether the metabolites might prove to be a more reliable predictive biomarker than those currently used clinically.
The mGWAS framework presented in this work provides an example of how pathophysiological elements can be identified and used to better understand the genetic factors that influence kidney disease progression and the potential adverse effects of medication. This study indicates that genes associated with ADME are highly expressed in the kidney and are associated with urine metabolites in CKD and healthy populations. Further endophenotyping based on urine metabolites may identify stronger genetic risk scores in subcategories of patients with CKD using trans-omic analysis. These results also indicate that the metabolism of medication is an important variable that will need to be considered in risk stratification strategies for patients with CKD.
Acknowledgements
This work was supported in part by National Institutes of Health (NIH) grant UH3 DK114920 and Department of Defence grant W81XWH-19-1-0659.
Footnotes
Competing interests
The authors declare no competing interests.
References
- 1.Sandholm N & Groop PH Genetic basis of diabetic kidney disease and other diabetic complications. Curr. Opin. Genet. Dev 50, 17–24 (2018). [DOI] [PubMed] [Google Scholar]
- 2.Lash JP et al. Chronic Renal Insufficiency Cohort (CRIC) study: baseline characteristics and associations with kidney function. Clin. J. Am. Soc. Nephrol 4, 1302–1311 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Reichel H et al. Chronic kidney disease progression and mortality risk profiles in Germany: results from the Chronic Kidney Disease Outcomes and Practice Patterns Study. Nephrol. Dial. Transplant 10.1093/ndt/gfz260 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Köttgen A, Raffler J, Sekula P & Kastenmüller G Genome-wide association studies of metabolite concentrations (mGWAS): relevance for nephrology. Semin. Nephrol 38, 151–174 (2018). [DOI] [PubMed] [Google Scholar]
- 5.Suhre K, Raffler J & Kastenmüller G Biochemical insights from population studies with genetics and metabolomics. Arch. Biochem. Biophys 589, 168–176 (2016). [DOI] [PubMed] [Google Scholar]
- 6.Long T et al. Whole-genome sequencing identifies common-to-rare variants associated with human blood metabolites. Nat. Genet 49, 568–578 (2017). [DOI] [PubMed] [Google Scholar]
- 7.Schlosser P et al. Genetic studies of urinary metabolites illuminate mechanisms of detoxification and excretion in humans. Nat. Genet 52, 167–176 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Titze S et al. Disease burden and risk profile in referred patients with moderate chronic kidney disease: composition of the German Chronic Kidney Disease (GCKD) cohort. Nephrol. Dial. Transplant 30, 441–451 (2015). [DOI] [PubMed] [Google Scholar]