ABSTRACT Inflammation increases the risk of cardiometabolic disease. Delineating specific inflammatory pathways and biomarkers of their activity could identify the mechanistic underpinnings

of the increased risk. Plasma levels of kynurenine, a metabolite involved in inflammation, associates with cardiometabolic disease risk. We used genetic approaches to identify inflammatory

mechanisms associated with kynurenine variability and their relationship to cardiometabolic disease. We identified single-nucleotide polymorphisms (SNPs) previously associated with plasma

kynurenine, including a missense-variant (rs3184504) in the inflammatory gene _SH2B3/LNK_. We examined the association between rs3184504 and plasma kynurenine in independent human samples,

_SH2B3_ missense variant associated with plasma kynurenine levels and _SH2B3_−/− mice had significant tissue-specific differences in kynurenine levels.LPS, an acute inflammatory stimulus,

increased plasma kynurenine in humans. Mendelian randomization showed increased waist-circumference, a marker of central obesity, associated with increased kynurenine, and increased

kynurenine associated with C-reactive protein (CRP). We found 30 diagnoses associated (FDR q < 0.05) with the _SH2B3_ variant, but not with SNPs mapping to genes known to regulate

tryptophan-kynurenine metabolism. Plasma kynurenine may be a biomarker of acute and chronic inflammation involving the _SH2B3_ pathways. Its regulation lies upstream of CRP, suggesting that

kynurenine may be a biomarker of one inflammatory mechanism contributing to increased cardiometabolic disease risk. SIMILAR CONTENT BEING VIEWED BY OTHERS GENOME-WIDE ASSOCIATION STUDY OF

PLASMA AMINO ACIDS AND MENDELIAN RANDOMIZATION FOR CARDIOMETABOLIC TRAITS Article Open access 25 April 2025 A GENOME-WIDE ASSOCIATION STUDY IDENTIFIES 41 LOCI ASSOCIATED WITH EICOSANOID

LEVELS Article Open access 31 July 2023 METABOLOME-WIDE ASSOCIATION IDENTIFIES FERREDOXIN-1 (FDX1) AS A DETERMINANT OF CHOLESTEROL METABOLISM AND CARDIOVASCULAR RISK IN ASIAN POPULATIONS

Article 13 May 2025 INTRODUCTION Epidemiological studies have identified numerous biomarkers associated with clinical disease1. However, the etiological relationship between the biomarker

and disease, and the molecular mechanisms contributing to their association often cannot be determined by traditional association approaches. This knowledge gap limits the opportunities to

identify clinical applications for a putative biomarker. Genetic approaches, which link molecular mechanisms with disease outcomes, can ascertain for possible causal relationships between a

biomarker and a disease, and can identify the discrete disease-associated mechanisms captured by the biomarker2. Here, we leverage such genetic approaches to more fully characterize the

plasma metabolite kynurenine. Kynurenine is a modulatory biomolecule synthesized from the essential dietary amino acid tryptophan, in part, by the inducible enzyme indoleamine

2,3-dioxygenase (IDO)3. Plasma kynurenine has been associated with a range of phenotypes including cardiovascular diseases (CVD) [heart disease, atherosclerosis, and endothelial

dysfunction], hypertension, diabetes, obesity and neuropsychiatric disorders3,4. Animal and cell models suggest that kynurenine pathway metabolites are linked to disease through modulation

of inflammation5,6. Therapies that decrease inflammation lower levels of non-specific inflammatory biomarkers such as c-reactive protein (CRP)7, reduce cardiovascular disease risk8 and

improve cardiometabolic indices including levels of glucose and insulin which would subsequently lead to disease reduction9. However, findings relating the kynurenine pathway to inflammation

and cardiometabolic disease are inconsistent, and the mechanistic relationship between plasma kynurenine levels and cardiometabolic disease risk is not well-defined. We hypothesized that

characterizing the genetic determinants underlying plasma kynurenine levels would define its role in health and disease risk. We used single nucleotide polymorphisms (SNPs) associated with

plasma kynurenine levels as instrumental variables to determine how genetic regulation of kynurenine contributes to disease risk. We leveraged data from observational and interventional

human studies and mouse models and showed that plasma kynurenine is a biomarker associated with multiple cellular mechanisms, and that only one of these mechanisms appears to associate with

clinical disease. RESULTS SNPS IN FOUR INDEPENDENT GENOMIC REGIONS ASSOCIATE WITH PLASMA KYNURENINE Of all SNPs reaching genome-wide significance (_p_ < 5 × 10−8), we identified six lead

SNPs associated with plasma kynurenine in the Cooperative Health Research in the Region of Augsburg (KORA) and TwinsUK GWAS meta-analysis at (Table 1). Based on Linkage disequilibrium (LD)

and genomic location, we concluded that these 6 lead SNPs represent 4 independent loci. Three SNPs are cis-expression quantitative trait loci (eQTLs) for genes involved in

tryptophan/kynurenine metabolism: the _IDO_ 1/2 genes (rs10085935, Chr8), which are the rate limiting enzymes in the kynurenine extrahepatic pathway3,10; and _SLC7A5_ (rs750950 and

rs8051149, Chr16), which participates in tryptophan transport across the cell membrane11. Two SNPs mapped to a region on Chr 12. The strongest association was with rs3184504, a missense

variant in a known inflammatory gene, _SH2B3_, and has been associated with multiple phenotypes including myocardial infarction and hypertension12. The final SNP (rs16924894, Chr10) is an

intergenic variant located near the _KIAA1217_ and _ARHGAP21_ genes. Tryptophan 2,3-dioxygenase, encoded by the TDO2 gene, is another key regulatory enzyme in the kynurenine pathway13. We

assessed the SNPs located in or near the TDO2 genes and none were significantly associated with plasma kynurenine levels. Of 6 downstream metabolites (3-hydroxy anthranilic acid (3HAA), and

quinolinic acid (QA), NAM (nicotinamide), NA (nicotinic acid), KA (kynurenic acid), HK (hydroxykynurenine)) in the kynurenine pathway, GWAS data of QA and NAM levels14,15 showed no

significant associations with the _SH2B3_ rs3184504 variant. (Supplementary Table 1). _SH2B3_ AND RS3184504 ASSOCIATE WITH PLASMA KYNURENINE To confirm the _SH2B3_ rs3184504 SNP association

with kynurenine levels, we examined the genetic association with plasma kynurenine using data from the ABO Glycoproteomics in Platelets and Endothelial Cells (ABO) Study (N = 66 young

healthy volunteers). The pattern of association was consistent with a recessive effect (Fig. 1), with a significant association between homozygosity for the T allele and increased kynurenine

(β = 0.17, _p_ = 0.04, recessive genetic model). We further characterized regulation of kynurenine by _SH2B3_ by measuring kynurenine levels in an _Sh2b3__−_/_−_ mouse, as compared to

wild-type controls. The most marked differences in kynurenine levels were observed in white adipose tissue where there were more than 2.5-fold (_p_ = 0.009) higher kynurenine levels in

_Sh2b3_−/− compared to wild type mice (Fig. 2). In contrast, significantly lower levels of kynurenine were observed in plasma (≈ 0.25-fold, _p_ = 0.02) and brain (≈ 0.13-fold, _p_ = 0.02) of

_Sh2b3_−/− mice. There was no difference in kynurenine levels in the spleen and modestly lower, but non-significant, levels in the liver in KO mice compared to controls. (Supplementary Fig.

 1). These data suggest that _SH2B3_ is a regulator of kynurenine metabolism, but that the effects on the pathway may differ by tissue. PLASMA KYNURENINE AND INFLAMMATION Inflammation is a

risk factor for cardiometabolic disease. To understand a possible genetic relationship between kynurenine and CRP, an inflammatory biomarker associated with cardiometabolic disease risk, we

conducted bi-directional Mendelian randomization analyses. Genetically determined kynurenine levels were associated with increased CRP (inverse-variance weighted average meta-analysis [IVWA]

β = 0.21 (0.10), _p_ = 0.04), but the reverse association was not significant (_p_ = 0.71), suggesting that kynurenine may increase CRP levels. We also examined the effects of

lipopolysaccharide (LPS) challenge (an acute inflammatory stimulus) on kynurenine levels in healthy human subjects (N = 24). In the Genetics of Evoked responses to Niacin and Endotoxemia

(GENE) study, plasma kynurenine significantly increased by 25% (_p_ = 0.0008) 2 h post-LPS challenge (Fig. 3), confirming that an inflammatory stimulus causes rapid increases in plasma

kynurenine levels. This timepoint is concurrent with peak cytokine responses to LPS (e.g. TNFα, IL-6), but precedes an increase in CRP, which peaked 24 h post-LPS16, highlighting kynurenine

as an early marker of acute inflammation. KYNURENINE AND BODY COMPOSITION Obesity is an inflammatory state, and is associated with elevated CRP levels17. We tested whether genetically

determined central adiposity, measured by waist circumference is associated with kynurenine levels, and found a significant positive association (IVWA β = 0.04 (0.01), _p_ = 0.0002).

Similarly, in the ABO Study, measured waist circumference was positively correlated with plasma kynurenine levels (r = 0.279, _p_ = 0.02). Collectively, these results suggest that central

adiposity is associated with higher kynurenine levels. THE KYNURENINE ASSOCIATION WITH CARDIOMETABOLIC DISEASES RISK IS SPECIFIC TO THE _SH2B3_ PATHWAY The individual SNPs associated with

plasma kynurenine represent distinct mechanisms of kynurenine regulation and may have distinct patterns of disease associations. We conducted PheWAS for each of the six SNPs individually.

The SNPs near the _IDO1/2_ and _SLC7A5_ genes, which directly contribute to either kynurenine synthesis or metabolism, had no significant associations with disease. In contrast, 30

phenotypes significantly associated with rs3184504 in _SH2B3_ at FDR q < 0.05 (Supplementary Table 2). Among these associations were three distinct collections of diagnoses related to:

hypothyroidism, hypertension and heart diseases including myocardial infarction, all of which were associated with higher kynurenine levels. The associations were similar in the Electronic

Medical Records and Genetics (eMERGE)/BioVU (Fig. 4) and the UK Biobank data sets (Supplementary Fig. 2). In sum, the _SH2B3_ rs3184504 variant associates with a range of diseases, and that

previously reported epidemiological disease associations with plasma kynurenine may be driven by genetic mechanisms associated with _SH2B3_. We performed phenome-wide association testing

between the 6-SNP kynurenine polygenic risk score and 894 phenotypes in the BioVU and eMERGE datasets to explore whether genetically-predicted plasma kynurenine associated with clinical

disease risk. We did not identify any significant associations (FDR q < 0.05). (Supplementary Fig. 3). Thus, plasma kynurenine is likely not a causal mediator of cardiometabolic diseases.

THE _SH2B3_ RS3184504 VARIANT MAY MODULATE DISEASE THOUGH MODULATION OF METABOLIC RATE We hypothesized that the mechanism linking _SH2B3_ to disease may be through effects on energy

metabolism. In the ABO study the rs3184504 risk variant was associated with significantly lower heart rate (β = −4.98, _p_ = 0.0004, additive genetic model and β = −4.98, _p_ = 0.017,

recessive genetic model) and body temperature (β = −0.002, _p_ = 0.0007, additive genetic model and β = −0.003, _p_ = 0.004, recessive genetic model), suggestive of physiological

differences. In the UK biobank (UKBB), we observed significant inverse relationship between the rs3184504 variant and whole body fat-free mass (β = −0.91, FDR _p_ = 1.39E−23), and basal

metabolic rate (β = −0.91, FDR _p_ = 3.38E−21). However, this variant was not associated with body fat percentage (β = −0.05, FDR _p_ = 0.89), suggesting a relationship with mitochondrial

energy metabolism rather than a direct effect on adiposity. Figure 5 summarizes the findings of the current study. DISCUSSION We leveraged genetic information to characterize the

relationship between plasma levels of kynurenine and cardiometabolic diseases. Plasma kynurenine levels are determined by genetic variation associated with genes involved in tryptophan

metabolism as well as genes without a clear previous link to kynurenine biosynthesis or catabolism. We confirmed a role for _SH2B3_ regulation of kynurenine in a knock-out mouse model.

Additional human studies (the GENE evoked endotoxemia study) also confirmed that kynurenine is induced by inflammatory stimuli and is likely up-stream of the inflammatory biomarker CRP,

which is also associated, but not causally related, with multiple inflammatory diseases. A polygenic predictor that assimilated the genetics from all of the kynurenine-modulating mechanisms

was not associated with clinical phenotypes, suggesting that plasma kynurenine itself does not cause disease. Indeed, the individual SNPs near genes with well-characterized roles in

kynurenine metabolism (_IDO, SLC7A5_) were not associated with disease. However, we observed significant associations linking a kynurenine-modulating missense variant in _SH2B3_ to a range

of diseases including atherosclerosis, hypertension and hypothyroidism, as well as with white blood cell and platelet counts. Genetic predictors of plasma kynurenine included some which

impact established regulatory pathways; kynurenine is generated by tryptophan degradation by _IDO_, the first and rate-limiting enzyme in tryptophan catabolism, whose activity is stimulated

by inflammation3,10. _SLC5A7_ encodes a protein involved in tryptophan transport across the cell membrane11. _IDO_ and _SLC5A7_ are thought to predominately regulate kynurenine production.

However, neither _IDO_ or _SLC5A7_ individually, or polygenic risk score analyses, showed genetic associations with cardiometabolic disorders or other disease. These data suggest that plasma

kynurenine levels per se do not have a causal relationship to these diseases. The _SH2B3_ rs3184504 variant was associated with many diseases, including cardiometabolic and thyroid disease.

This variant is a nonsynonymous SNP located in exon 3 of the _SH2B3_ gene which leads to a R262W amino acid change and has been shown to be involved in controlling immune responses. _SH2B3_

(_LNK_) is important in hematopoiesis, regulates the expansion of dendritic cells in lymph nodes, acts as a negative regulator of cytokine signaling, and has been associated with increased

susceptibility to aortic dissection18,19,20. It is thought that the rs3184504 variant in _SH2B3_ approximates a loss of function21. Thus, to confirm the relationship between _SH2B3_ and

kynurenine, we measured kynurenine levels in _SH2B3__−_/_−_mice, and observed significant plasma and tissue-specific changes in kynurenine. Up to now, no mechanistic relationship between

_SH2B3_ and kynurenine metabolism had been reported. _SH2B3_ may regulate kynurenine levels indirectly through modulation of inflammatory signaling, and subsequent activation of _IDO_22.

However, our data supported an association between rs3184504 and kynurenine in healthy individuals, without concurrent inflammation, suggesting _SH2B3_ may directly modulate the kynurenine

pathway. The kynurenine pathway controls production of NAD + , with important consequences for energy metabolism, and a known relationship to acute kidney injury23. Our data support a link

between the SH2B3-kynurenine axis and energy homeostasis. We found that the _SH2B3_ rs3184504 variant associates with lower whole body fat-free mass and basal metabolic rate, in addition to

lower body temperature and heart rate. There may be an important distinction between kynurenine that is elevated due to increased synthesis from tryptophan (controlled by _IDO_ and

_SLC7A5_), versus kynurenine that is elevated due to reduced breakdown of kynurenine or altered downstream pathway flux (potentially controlled by _SH2B3_). This hypothesis remains to be

further interrogated, and there may be other mechanisms linking _SH2B3_ to kynurenine, which remain to be explored. Epidemiological studies have shown that higher plasma kynurenine levels

associate with increased inflammation and CVD prevalence in patients with end-stage renal disease. Individuals with higher kynurenic acid levels were more likely to have higher risk of

coronary artery disease mortality and myocardial infarction24. Furthermore, kynurenine has been associated with the risk of all-cause mortality, particularly death from CVD25. We found that

plasma kynurenine was acutely regulated during evoked endotoxemia, suggestive of a role in acute inflammatory responses. Kynurenine has recently been reported to be upregulated during

COVID-19 infection, further highlighting a potential role for kynurenine as a biomarker of inflammatory activation26. Waist circumference, a marker of central obesity, was positively

associated with kynurenine. We also observed an inverse association between the rs3184504 variant and whole body fat-free mass and basal metabolic rate. PheWAS analyses revealed that the

rs3184504 variant was directly associated with some obesity associated co-morbidities including hypertension and heart-related diseases, but not some other obesity-related comorbidities

including obstructive sleep apnea and type 2 diabetes. Consistent with these findings, previous research demonstrated that _SH2B3_-related genetic alterations contribute to the development

of hypertension and hematological disorders27,28. Consistently, the _SH2B3_ rs3184504 variant is associated with increased risk of CAD29, increased platelet counts and leukocytosis30,

diastolic blood pressure31, atherosclerosis and thrombosis21. Thus, these data suggest that this _SH2B3_-related mechanism, which regulates both disease risk and kynurenine levels, may

account for the observed associations between plasma kynurenine and the risk of cardiometabolic diseases. Traditional prospective studies seeking to link a metabolomic biomarker to a disease

are limited based on sample sizes, follow-up times and do not provide insights into the molecular mechanisms account for candidate associations. We circumvented these limitations by using

an approach which associated a putative biomarker with clinical traits based on their shared genetic structures. Thus, we were able probe for kynurenine associations in large, deeply

phenotyped populations. Along with other strengths (increasing sample sizes and the number of diseases), this approach can also identify molecular mechanisms underlying an association.

Another strength of our study is that we replicated findings in multiple populations (BioVU, eMERGE and the UKBB). Our study had considerable strengths, but also some limitations. Using EHR

billing codes rather than a systematic ascertainment of a diagnosis for classifying PheWAS phenotypes in BioVU and eMERGE databases might lead to both false-positive and false-negative

associations. To address this, we replicated the study in other databases including UKBB and published data from large disease GWAS. Negative associations, such as the absence of an

association between the kynurenine polygenic score (PRS) and disease, leave open the possibility of a false negative due to lack of power. However, we obtained very strong associations for

the SNP in _SH2B3_ in the same datasets, indicating that the kynurenine genetic instruments should have been suitably powered. While experimental evidence is provided for showing kynurenine

increases associated with acute inflammatory stimuli, additional studies are needed to confirm sustained elevations associated with inflammatory states. Furthermore, tissues in the heart,

lungs and cardiopulmonary vasculature were not available for analyses and differences in levels in KO mice could not be assessed. In conclusion: in this virtual biomarker study, we explored

the association between kynurenine genetic predictors and clinical diagnoses derived from large datasets. Our findings suggest diverse molecular mechanisms regulate plasma kynurenine. The

_SH2B3-_rs3184504 variant associates with both plasma kynurenine and diseases; our data suggest this is independent of kynurenine production. Plasma kynurenine, upregulated during

inflammation, is upstream of the inflammatory biomarker CRP. The _SH2B3_ rs3184504 variant, which regulates kynurenine levels, associates with increased cardiometabolic disorders,

potentially in a kynurenine-independent manner. In sum, although targeting plasma kynurenine directly is unlikely to be effective in disease treatment, interrogation of the _SH2B3_ pathways

during inflammation may identify novel causal disease mechanisms. MATERIALS AND METHODS We identified SNPs associated with plasma kynurenine from published data32, and tested their

associations with disease phenotypes. Identified associations were validated and probed using animal models and in independent populations. KYNURENINE GWAS SUMMARY STATISTICS SNPs associated

with plasma kynurenine were identified from a GWAS meta-analysis of the KORA-TwinsUK studies32. Analyses were based on 7824 adult individuals of European descent33,34. Summary statistics

were obtained from the Metabolomics GWAS server (http://metabolomics.helmholtz-muenchen.de/gwas/)32. An independent (r2 < 0.05 within 1000 kilobases) set of SNPs significantly associated

(_p_ < 5 × 10−8) with circulating kynurenine levels were selected from the KORA-TwinsUK meta-analysis GWAS summary statistics. ELECTRONIC HEALTH RECORD-LINKED GENETIC DATASETS BIOVU BioVU

is Vanderbilt University Medical Center’s DNA biobank, which is linked to de-identified EHR phenotype data35. A subset of BioVU (n = 62,303) participants of European Ancestry have SNP

genotype data acquired on the Illumina MEGAEX platform. Quality control steps for the BioVU population have been previously described36. Genotypes were imputed with IMPUTE437, version 2.3.0

(University of Oxford), using the 10/2014 release of the 1000 Genomes cosmopolitan reference haplotypes and variants imputation quality scores less than 0.3 were excluded. One participant

from each related pair (pi-hat > 0.2) was randomly excluded. Analyses were restricted to subjects of European ancestry, defined by principal components analyses in conjunction with HapMap

reference populations. Quality control analyses used PLINK v1.938. The use of BioVU and other de-identified data presented in these analyses was approved by the VUMC Institutional Review

Board (IRB), in accordance with the informed consent guidelines. EMERGE The eMERGE Phase I, II and III Network, a consortium of medical centers using EHRs as a tool for genomic research,

included39 participants (n = 32,324) who were born prior to 1990 and were recruited from Geisinger Health System, Marshfield Clinic, Northwestern University, Mayo Clinic, and Kaiser

Permanente/University of Washington. Consent was collected based on each site’s IRB protocols. eMERGE data were genotyped on multiple SNP arrays. QC procedures and imputation protocols for

these data were conducted based on the established protocols developed by the eMERGE Genomics Working Group41. Use of de-identified eMERGE data was approved by the IRB at each site39, in

accordance with the site-specific informed consent guidelines. THE UKBB STUDY UKBB is a British population-based self-reported study which is composed of approximately 0.5 million

participants aged 37–73 at recruitment42. GWAS summary statistics for 2173 UKBB phenotypes43 were downloaded from the study by Bycroft et al.44. PheWAS results for individual

kynurenine-associated SNPs were obtained from http://geneatlas.roslin.ed.ac.uk/43. DISEASE GENOME-WIDE ASSOCIATION STUDY DATASETS Summary statistics for CRP were obtained from a GWAS

meta-analysis of 204,402 European individuals45. Additional summary statistics were downloaded from published GWAS of waist circumference20. PHENOME-WIDE ASSOCIATION ANALYSIS (PHEWAS) Single

SNP and multi SNP PheWAS analyses were conducted by testing associations with either individual SNPs (single SNP) associated with kynurenine or all kynurenine SNPs (a PRS comprising multi

SNPs) PheWAS phenotypes. Analyses used the R PheWAS package46. PheWAS were performed in BioVU and eMERGE using clinical phecode phenotypes (https://phewas.mc.vanderbilt.edu/) based on

ICD-9-CM and ICD-10 diagnosis codes47,48. Individuals with two or more instances of a PheWAS diagnosis existing in their medical documents were considered as cases49. Clinical phenotypes

with ≥ 300 cases were included and those affecting a single sex (like uterine prolapse and prostate cancer) were excluded. After exclusions, there were 894 phenotypes. Controls were

individuals without any closely related PheWAS codes, and were matched to the age (BioVU) or decade of birth (eMERGE) ranges among the cases. EQTL ANALYSIS eQTL data for selected SNPs were

obtained from the Genotype Tissue Expression (GTEx) portal, https://gtexportal.org. ANIMAL MODELS SH2B3−/− MOUSE Plasma and tissues (white adipose tissue, brain, liver, spleen) were obtained

from 14–15 week-old C57BL/6 J mice (wild type [WT, N = 9 plasma, N = 8 tissues)]) and from Lnk−/− (Sh2b3 knock out [KO], N = 5), as previously described50. Samples were frozen immediately

following collection, and stored at − 80° C prior to analysis. Tissue samples were homogenized (Tissue Lyser LT) and stored at − 80° C prior to metabolite measurement. Mice were housed and

taken care of in accordance with the Guide for the Care and Use of Laboratory Animals, US Department of Health and Human Services. All animal procedures were approved by the Vanderbilt

Institutional Animal Care and Use Committee. ENZYME-LINKED IMMUNOSORBENT ASSAY (ELISA) OF KYNURENINE Kynurenine in mouse plasma and tissue extracts was analyzed by ELISA (Cod. LS-F56401,

_LifeSpan BioScience Inc_., Seattle, WA, USA), according to manufacturer’s instructions. All samples were run in duplicate, with an even distribution of samples from KO and WT animals across

the three plates used. The intra-assay coefficient of variance was 9.8%. Due to relatively high inter-assay variability, likely attributable to differences by lot in the three ELISA kits

used, we analyzed data as fold difference between WT and KO mice for each plate, rather than absolute values. CLINICAL STUDIES GENE STUDY Healthy volunteers (294 non-pregnant/lactating women

and men, age 18–45, BMI 18–30 kg/m2) were recruited to an evoked endotoxemia study (LPS, 1 ng/kg) at the University of Pennsylvania, as previously described16. Plasma metabolites, including

kynurenine, were measured in a subset of individuals (n = 24) at baseline, and two hours post LPS-challenge, by mass spectrometry in positive and negative ion modes using well-established

protocols51,52. ABO STUDY The ABO Study recruited healthy volunteers (non-pregnant/lactating women and men, age 18–50) to a single study visit at the University of Pennsylvania from 2012 to

2014, as described53. Plasma metabolomics profiling, including measurement of kynurenine, was carried out at Metabolon (Metabolon Inc, Morrisville, NC; global metabolomics platform).

Genotyping was performed using the Exome chip (Illumina, CoreExome, N > 540,000 variants, including rs3184504) at the Center for Applied Genomics at the Children's Hospital of

Philadelphia. We analyzed data for a subset of individuals with overlapping metabolite and genetic data (N = 66). The GENE and ABO studies were approved by the IRB of the University of

Pennsylvania and Vanderbilt University. All participants provided written informed consent. STATISTICAL ANALYSES GENOTYPED POPULATIONS (UKBB, BIOVU, EMERGE AND OTHER LARGE DISEASE GWAS)

SINGLESNP PheWAS analyses for each kynurenine-associated SNP were performed in the BioVU and eMERGE populations. Associations were tested assuming an additive genetic model and used

multivariable logistic regression models adjusting for 5 PCs, sex and either birth decade (eMERGE) or age (BioVU). BioVU and eMERGE data were combined using meta-analyses encoded by the

METAL package (default settings were used)54. MULTISNP A PRS for plasma kynurenine levels was computed for each individual in the BioVU and eMERGE populations by summing their (Allele dosage

x change in kynurenine levels per allele) for each kynurenine-associated SNPs. Association testing between the PRS and each phenotype was then tested and combined, as described above.

Odds-ratios (ORs) represent the risk of disease per standard deviation (SD) increase in the PRS. For UKBB phenotypes, and additional phenotypes based on GWAS summary statistics, mutliSNP

association tests were conducted. Associations were tested using the IVWA, MR-Egger and Weighted Median methods, as implemented in the Mendelian Randomization R package55. Heterogeneity _p_

values are based on the Cochran’s Q statistic, and a low _p_ value may indicate horizontal pleiotropy. Association estimates represent the change in the log odds-ratio per standard deviation

change in plasma kynurenine. MULTIPLE TESTING CORRECTIONS We applied a strict Benjamini–Hochberg (B–H) false discovery rate (FDR)56 to adjust for multiple testing, and associations with a

q-value < 0.05 were considered significant40. All analyses were performed in accordance with relevant guidelines and regulations. The study was carried out in compliance with the ARRIVE

guidelines. DATA AND CODE AVAILABILITY eMERGE data are available through dbGaP (phs000360.v3.p1). Upon acceptance, the complete findings from the PheWAS analyses will be made available

thought a publicly-available website. ABBREVIATIONS * B–H: Benjamini–Hochberg * CRP: C-reactive protein * CVD: Cardiovascular diseases * ELISA: Enzyme-linked immunosorbent assay * eMERGE:

Electronic medical records and genetics consortium * eQTL: Expression quantitative trait loci * FDR: False discovery rate * GENE: Genetics of evoked responses to niacin and endotoxemia *

GTEx: Genotype tissue expression * 3HAA: 3-Hydroxy anthranilic acid * HK: Hydroxykynurenine * IDO: Indoleamine 2,3-dioxygenase * IRB: Institutional Review Board * IVWA: Inverse-variance

weighted average * KORA: Cooperative Health Research in the Region of Augsburg * KA: Kynurenic acid * LD: Linkage disequilibrium * LPS: Lipopolysaccharide * NA: Nicotinic acid * NAM:

Nicotinamide * OR: Odds-ratio * PheWAS: Phenome-wide association analysis * PRS: Polygenic score * QA: Quinolinic acid * SD: Standard deviation * SNP: Single nucleotide polymorphism

REFERENCES * Vasan, R. S. Biomarkers of cardiovascular disease: molecular basis and practical considerations. _Circulation_ 113, 2335–2362 (2006). Article  PubMed  Google Scholar  * Maher,

B. S. Polygenic scores in epidemiology: risk prediction, etiology, and clinical utility. _Curr. Epidemiol. Rep._ 2, 239–244 (2015). Article  PubMed  PubMed Central  Google Scholar  *

Michelhaugh, S. K., Guastella, A. R. & Mittal, S. Overview of the kynurenine pathway of tryptophan metabolism. In _Targeting the Broadly Pathogenic Kynurenine Pathway_ (ed. Mittal, S.)

3–9 (Springer, Berlin, 2015). https://doi.org/10.1007/978-3-319-11870-3_1. Chapter  Google Scholar  * Song, P., Ramprasath, T., Wang, H. & Zou, M.-H. Abnormal kynurenine pathway of

tryptophan catabolism in cardiovascular diseases. _Cell Mol. Life Sci._ 74, 2899–2916 (2017). Article  CAS  PubMed  PubMed Central  Google Scholar  * Pawlak, K., Domaniewski, T., Mysliwiec,

M. & Pawlak, D. The kynurenines are associated with oxidative stress, inflammation and the prevalence of cardiovascular disease in patients with end-stage renal disease.

_Atherosclerosis_ 204, 309–314 (2009). Article  CAS  PubMed  Google Scholar  * Mangge, H. _et al._ Disturbed tryptophan metabolism in cardiovascular disease. _Curr. Med. Chem._ 21, 1931–1937

(2014). Article  CAS  PubMed  PubMed Central  Google Scholar  * Cozlea, D. L. _et al._ The impact of C reactive protein on global cardiovascular risk on patients with coronary artery

disease. _Curr. Health Sci. J._ 39, 225–231 (2013). CAS  PubMed  PubMed Central  Google Scholar  * Aimo, A. _et al._ Colchicine for the treatment of coronary artery disease. _Trends

Cardiovasc. Med._ https://doi.org/10.1016/j.tcm.2020.10.007 (2020). Article  PubMed  Google Scholar  * Esser, N., Paquot, N. & Scheen, A. J. Inflammatory markers and cardiometabolic

diseases. _Acta Clin. Belg._ 70, 193–199 (2015). Article  CAS  PubMed  Google Scholar  * Badawy, A.A.-B. Kynurenine pathway of tryptophan metabolism: regulatory and functional aspects. _Int.

J. Tryptophan. Res._ 10, 1178646917691938–1178646917691938 (2017). Article  PubMed  PubMed Central  CAS  Google Scholar  * Scalise, M., Galluccio, M., Console, L., Pochini, L. &

Indiveri, C. The human SLC7A5 (LAT1): the intriguing histidine/large neutral amino acid transporter and its relevance to human health. _Front. Chem._ 6, 243 (2018). Article  PubMed  PubMed

Central  ADS  CAS  Google Scholar  * Yao, C. _et al._ Integromic analysis of genetic variation and gene expression identifies networks for cardiovascular disease phenotypes. _Circulation_

131, 536–549 (2015). Article  CAS  PubMed  Google Scholar  * Breda, C. _et al._ Tryptophan-2,3-dioxygenase (TDO) inhibition ameliorates neurodegeneration by modulation of kynurenine pathway

metabolites. _Proc. Natl. Acad. Sci. USA_ 113, 5435 (2016). Article  CAS  PubMed  PubMed Central  ADS  Google Scholar  * Long, T. _et al._ Whole-genome sequencing identifies common-to-rare

variants associated with human blood metabolites. _Nat. Genet._ 49, 568–578 (2017). Article  CAS  PubMed  Google Scholar  * Rhee, E. P. _et al._ A genome-wide association study of the human

metabolome in a community-based cohort. _Cell Metab._ 18, 130–143 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Ferguson, J. F. _et al._ Race and gender variation in

response to evoked inflammation. _J. Transl. Med._ 11, 63 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Choi, J., Joseph, L. & Pilote, L. Obesity and C-reactive protein

in various populations: a systematic review and meta-analysis: obesity and CRP in various populations _Obes_. _Rev._ 14, 232–244 (2013). CAS  Google Scholar  * Mori, T. _et al._

_Lnk_/_Sh2b3_ controls the production and function of dendritic cells and regulates the induction of IFN-γ–producing T cells. _J. Immunol._ 193, 1728 (2014). Article  CAS  PubMed  Google

Scholar  * Laroumanie, F. _et al._ LNK deficiency promotes acute aortic dissection and rupture. _JCI Insight_ 3, e122558 (2018). Article  PubMed Central  Google Scholar  * Shungin, D. _et

al._ New genetic loci link adipose and insulin biology to body fat distribution. _Nature_ 518, 187–196 (2015). Article  CAS  PubMed  PubMed Central  Google Scholar  * Wang, W. _et al._

LNK/SH2B3 loss of function promotes atherosclerosis and thrombosis. _Circ. Res._ 119, e91–e103 (2016). Article  CAS  PubMed  PubMed Central  Google Scholar  * Jiang, X. & Karlsen, T. H.

Genetics of primary sclerosing cholangitis and pathophysiological implications. _Nat. Rev. Gastroenterol. Hepatol._ 14, 279–295 (2017). Article  CAS  PubMed  Google Scholar  * Poyan Mehr, A.

_et al._ De novo NAD+ biosynthetic impairment in acute kidney injury in humans. _Nat. Med._ 24, 1351–1359 (2018). Article  CAS  PubMed  Google Scholar  * Yu, E. _et al._ Increases in plasma

tryptophan are inversely associated with incident cardiovascular disease in the Prevención con Dieta Mediterránea (PREDIMED) study. _J Nutr._ 147, 314–322 (2017). CAS  PubMed  PubMed

Central  Google Scholar  * Zuo, H. _et al._ Plasma biomarkers of inflammation, the kynurenine pathway, and risks of all-cause, cancer, and cardiovascular disease mortality: the hordaland

health study. _Am. J. Epidemiol._ 183, 249–258 (2016). Article  PubMed  PubMed Central  Google Scholar  * Thomas, T. _et al._ COVID-19 infection alters kynurenine and fatty acid metabolism,

correlating with IL-6 levels and renal status. _JCI Insight_ https://doi.org/10.1172/jci.insight.140327 (2020). Article  PubMed  PubMed Central  Google Scholar  * Dale, B. L. & Madhur,

M. S. Linking inflammation and hypertension via LNK/SH2B3. _Curr. Opin. Nephrol. Hypertens_ 25, 87–93 (2016). Article  CAS  PubMed  PubMed Central  Google Scholar  * Maslah, N., Cassinat,

B., Verger, E., Kiladjian, J.-J. & Velazquez, L. The role of LNK/SH2B3 genetic alterations in myeloproliferative neoplasms and other hematological disorders. _Leukemia_ 31, 1661–1670

(2017). Article  CAS  PubMed  Google Scholar  * CARDIoGRAMplusC4D Consortium _et al._ Large-scale association analysis identifies new risk loci for coronary artery disease. _Nat. Genet._ 45,

25–33 (2013). * Gudbjartsson, D. F. _et al._ Sequence variants affecting eosinophil numbers associate with asthma and myocardial infarction. _Nat. Genet._ 41, 342–347 (2009). Article  CAS 

PubMed  Google Scholar  * Newton-Cheh, C. _et al._ Genome-wide association study identifies eight loci associated with blood pressure. _Nat. Genet._ 41, 666–676 (2009). Article  CAS  PubMed

  PubMed Central  Google Scholar  * Shin, S.-Y. _et al._ An atlas of genetic influences on human blood metabolites. _Nat. Genet._ 46, 543–550 (2014). Article  CAS  PubMed  PubMed Central 

Google Scholar  * Moayyeri, A., Hammond, C. J., Hart, D. J. & Spector, T. D. The UK adult twin registry (TwinsUK Resource). _Twin Res. Hum. Genet._ 16, 144–149 (2013). Article  PubMed 

Google Scholar  * Wichmann, H.-E., Gieger, C. & Illig, T. KORA-gen—resource for population genetics, controls and a broad spectrum of disease phenotypes. _Gesundheitswesen_ 67, 26–30

(2005). Article  Google Scholar  * Roden, D. _et al._ Development of a large-scale de-identified DNA biobank to enable personalized medicine. _Clin. Pharmacol. Ther._ 84, 362–369 (2008).

Article  CAS  PubMed  Google Scholar  * Ruderfer, D. M. _et al._ Significant shared heritability underlies suicide attempt and clinically predicted probability of attempting suicide. _Mol.

Psychiatry_ https://doi.org/10.1038/s41380-018-0326-8 (2019). Article  PubMed  PubMed Central  Google Scholar  * Howie, B. N., Donnelly, P. & Marchini, J. A flexible and accurate

genotype imputation method for the next generation of genome-wide association studies. _PLoS Genet._ 5, e1000529 (2009). Article  PubMed  PubMed Central  CAS  Google Scholar  * Purcell, S.

_et al._ PLINK: a tool set for whole-genome association and population-based linkage analyses. _Am. J. Hum. Genet._ 81, 559–575 (2007). Article  CAS  PubMed  PubMed Central  Google Scholar 

* Gottesman, O. _et al._ The electronic medical records and genomics (eMERGE) network: past, present, and future. _Genet. Med._ 15, 761–771 (2013). Article  PubMed  PubMed Central  Google

Scholar  * Mosley, J. D. _et al._ A study paradigm integrating prospective epidemiologic cohorts and electronic health records to identify disease biomarkers. _Nat. Commun._ 9, 3522 (2018).

Article  PubMed  PubMed Central  ADS  CAS  Google Scholar  * Zuvich, R. L. _et al._ Pitfalls of merging GWAS data: lessons learned in the eMERGE network and quality control procedures to

maintain high data quality: pitfalls of Merging GWAS data: lessons learned. _Genet. Epidemiol._ 35, 887–898 (2011). Article  PubMed  PubMed Central  Google Scholar  * Sudlow, C. _et al._ UK

Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. _PLoS Med._ 12, e1001779 (2015). Article  PubMed  PubMed Central 

Google Scholar  * Jiang, L. _et al._ A resource-efficient tool for mixed model association analysis of large-scale data. _Nat. Genet._ 51, 1749–1755 (2019). Article  CAS  PubMed  Google

Scholar  * Bycroft, C. _et al._ The UK biobank resource with deep phenotyping and genomic data. _Nature_ 562, 203–209 (2018). Article  CAS  PubMed  PubMed Central  ADS  Google Scholar  *

Ligthart, S. _et al._ Genome analyses of >200,000 individuals identify 58 loci for chronic inflammation and highlight pathways that link inflammation and complex disorders. _Am. J. Hum.

Genet._ 103, 691–706 (2018). Article  CAS  PubMed  PubMed Central  Google Scholar  * Carroll, R. J., Bastarache, L. & Denny, J. C. R PheWAS: data analysis and plotting tools for

phenome-wide association studies in the R environment. _Bioinformatics_ 30, 2375–2376 (2014). Article  CAS  PubMed  PubMed Central  Google Scholar  * Denny, J. C. _et al._ Systematic

comparison of phenome-wide association study of electronic medical record data and genome-wide association study data. _Nat. Biotechnol._ 31, 1102–1111 (2013). Article  CAS  PubMed  PubMed

Central  Google Scholar  * Denny, J. C. _et al._ PheWAS: demonstrating the feasibility of a phenome-wide scan to discover gene–disease associations. _Bioinformatics_ 26, 1205–1210 (2010).

Article  CAS  PubMed  PubMed Central  Google Scholar  * Wei, W.-Q. _et al._ Evaluating phecodes, clinical classification software, and ICD-9-CM codes for phenome-wide association studies in

the electronic health record. _PLoS ONE_ 12, e0175508 (2017). Article  PubMed  PubMed Central  CAS  Google Scholar  * Saleh, M. A. _et al._ Lymphocyte adaptor protein LNK deficiency

exacerbates hypertension and end-organ inflammation. _J. Clin. Invest._ 125, 1189–1202 (2015). Article  PubMed  PubMed Central  Google Scholar  * Wang, T. J. _et al._ Metabolite profiles and

the risk of developing diabetes. _Nat. Med._ 17, 448–453 (2011). Article  PubMed  PubMed Central  CAS  Google Scholar  * Kimberly, W. T. _et al._ Metabolite profiling identifies anandamide

as a biomarker of nonalcoholic steatohepatitis. _JCI Insight_ 2, e92989 (2017). Article  PubMed Central  Google Scholar  * Tang, Z.-Z. _et al._ Multi-omic analysis of the microbiome and

metabolome in healthy subjects reveals microbiome-dependent relationships between diet and metabolites. _Front. Genet._ 10, 454 (2019). Article  CAS  PubMed  PubMed Central  Google Scholar 

* Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efficient meta-analysis of genomewide association scans. _Bioinformatics_ 26, 2190–2191 (2010). Article  CAS  PubMed  PubMed

Central  Google Scholar  * Mahajan, A. _et al._ Fine-mapping type 2 diabetes loci to single-variant resolution using high-density imputation and islet-specific epigenome maps. _Nat. Genet._

50, 1505–1513 (2018). Article  CAS  PubMed  PubMed Central  Google Scholar  * Benjamini, Y. & Yekutieli, D. The control of the false discovery rate in multiple testing under dependency.

_Ann. Stat._ 29, 1165–1188 (2001). Article  MathSciNet  MATH  Google Scholar  Download references ACKNOWLEDGEMENTS The authors are grateful to the staff and participants of the BioVU,

eMERGE, UK biobank and all GWAS studies used in this paper. The research is solely the responsibility of the authors and do not necessarily represent the views of Vanderbilt University

Medical Center. FUNDING This research was supported by the NIH, R01 HL142856 (Ferguson), and the American Heart Association; 16FTF30130005 (Mosley). Datasets used for the analyses described

were obtained from Vanderbilt University Medical Center’s BioVU which is supported by institutional funding, the 1S10RR025141-01 instrumentation award, and by the CTSA grant UL1TR000445 from

NCATS/NIH. The eMERGE Network was initiated and funded by NHGRI through the following grants: U01HG006828 (Cincinnati Children’s Hospital Medical Center/Boston Children’s Hospital);

U01HG006830 (Children’s Hospital of Philadelphia); U01HG006389 (Essentia Institute of Rural Health, Marshfield Clinic Research Foundation and Pennsylvania State University); U01HG006382

(Geisinger Clinic); U01HG006375 (Group Health Cooperative/University of Washington); U01HG006379 (Mayo Clinic); U01HG006380 (Icahn School of Medicine at Mount Sinai); U01HG006388

(Northwestern University); U01HG006378 (Vanderbilt University Medical Center); U01HG006385 (Vanderbilt University Medical Center serving as the Coordinating Center); U01HG004438 (CIDR) and

U01HG004424 (the Broad Institute) serving as Genotyping Centers. AUTHOR INFORMATION Author notes * These authors contributed equally: Jonathan D. Mosley and Jane F. Ferguson. AUTHORS AND

AFFILIATIONS * Division of Cardiovascular Medicine, Department of Medicine, Vanderbilt University Medical Center, 2220 Pierce Ave, PRB 354B, Nashville, TN, 37232, USA Minoo Bagheri, Chuan

Wang, Ali Manouchehri, Katherine T. Murray & Jane F. Ferguson * Department of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN, USA Mingjian Shi, Christian M.

Shaffer & Jonathan D. Mosley * Division of Clinical Pharmacology, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA Ali Manouchehri, Katherine T. Murray, 

Matthew B. Murphy & Jonathan D. Mosley * Division of Genetic Medicine, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA Kritika Singh & Lea K. Davis *

Departments of Medicine (Medical Genetics) and Genome Sciences, University of Washington, Seattle, WA, USA Gail P. Jarvik * Division of Nephrology, School of Medicine, Harborview Medical

Center Kidney Research Institute, University of Washington, Seattle, WA, USA Ian B. Stanaway * Center for Precision Medicine Research, Marshfield Clinic Research Institute, Marshfield, WI,

USA Scott Hebbring * Irving Institute for Clinical and Translational Research and Division of Cardiology, Columbia University Medical Center, New York, NY, USA Muredach P. Reilly * Division

of Cardiovascular Medicine, Beth Israel Deaconess Medical Center, Boston, MA, USA Robert E. Gerszten * Department of Internal Medicine, University of Texas Southwestern Medical Center,

Dallas, USA Thomas J. Wang Authors * Minoo Bagheri View author publications You can also search for this author inPubMed Google Scholar * Chuan Wang View author publications You can also

search for this author inPubMed Google Scholar * Mingjian Shi View author publications You can also search for this author inPubMed Google Scholar * Ali Manouchehri View author publications

You can also search for this author inPubMed Google Scholar * Katherine T. Murray View author publications You can also search for this author inPubMed Google Scholar * Matthew B. Murphy

View author publications You can also search for this author inPubMed Google Scholar * Christian M. Shaffer View author publications You can also search for this author inPubMed Google

Scholar * Kritika Singh View author publications You can also search for this author inPubMed Google Scholar * Lea K. Davis View author publications You can also search for this author

inPubMed Google Scholar * Gail P. Jarvik View author publications You can also search for this author inPubMed Google Scholar * Ian B. Stanaway View author publications You can also search

for this author inPubMed Google Scholar * Scott Hebbring View author publications You can also search for this author inPubMed Google Scholar * Muredach P. Reilly View author publications

You can also search for this author inPubMed Google Scholar * Robert E. Gerszten View author publications You can also search for this author inPubMed Google Scholar * Thomas J. Wang View

author publications You can also search for this author inPubMed Google Scholar * Jonathan D. Mosley View author publications You can also search for this author inPubMed Google Scholar *

Jane F. Ferguson View author publications You can also search for this author inPubMed Google Scholar CONTRIBUTIONS All authors reviewed, commented on and approved the paper for the

submission. M.B., J.F.F., J.D.M. designed the study. M.B., J.F.F., J.D.M., C.W., M.S., A.M., K.S. conducted research. M.B., J.F.F., J.D.M., C.W., M.S., C.M.S., K.S. performed the statistical

analysis. M.B., J.F.F., J.D.M. interpreted the results. M.B., J.F.F., J.D.M. wrote the paper. CORRESPONDING AUTHOR Correspondence to Jane F. Ferguson. ETHICS DECLARATIONS COMPETING

INTERESTS The authors declare no competing interests. ADDITIONAL INFORMATION PUBLISHER'S NOTE Springer Nature remains neutral with regard to jurisdictional claims in published maps and

institutional affiliations. SUPPLEMENTARY INFORMATION SUPPLEMENTARY INFORMATION 1. SUPPLEMENTARY INFORMATION 2. SUPPLEMENTARY INFORMATION 3. RIGHTS AND PERMISSIONS OPEN ACCESS This article

is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give

appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in

this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's

Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Reprints and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE Bagheri, M., Wang, C., Shi, M. _et al._ The

genetic architecture of plasma kynurenine includes cardiometabolic disease mechanisms associated with the _SH2B3_ gene. _Sci Rep_ 11, 15652 (2021). https://doi.org/10.1038/s41598-021-95154-9

Download citation * Received: 19 January 2021 * Accepted: 21 July 2021 * Published: 02 August 2021 * DOI: https://doi.org/10.1038/s41598-021-95154-9 SHARE THIS ARTICLE Anyone you share the

following link with will be able to read this content: Get shareable link Sorry, a shareable link is not currently available for this article. Copy to clipboard Provided by the Springer

Nature SharedIt content-sharing initiative