Loss of Y chromosome: proteomic signatures in human cardiovascular disease

Acquired somatic mutations are a hallmark of aging and are increasingly recognized as contributors to disease susceptibility. Mosaic loss of the Y chromosome (mLOY) in leukocytes represents the most common form of acquired clonal mosaicism in men1 and associates with incident cardiovascular diseases including heart failure,2 atrial fibrillation,3 fatal myocardial infarction in men with CAD, or heart failure-associated death in chronic kidney disease.4,5 While the genetic predisposition to mLOY has been recently investigated,6,7 the effects of mLOY on the systemic proteome signature have not been comprehensively characterized. Therefore, we examined the association between mLOY and the plasma proteome signature in men enrolled in the UK Biobank. Mosaic loss of Y chromosome was measured utilizing single nucleotide polymorphism array intensity data by the mLLRY metric corrected by linear regression, as previously described.1,2 Proteomic profiling as part of the UK Biobank Pharma Proteomics Project (UKB-PPP) was performed using the Olink Exlore 1536 platform. The UKB-PPP cohort includes 54 306 individuals of those N = 20 645 men (mean age 57.1 ± 8.3 years) without prevalent cardiovascular disease, in whom LOY was quantified, and who were included in the present analyses. To test the association between mLOY as continuous variable and plasma protein levels, OLINK-derived normalized protein expression levels were z-transformed. Linear regression models were employed including age at recruitment, ethnicity, smoking, body mass index (BMI), prevalent hypertension, prevalent diabetes, LDL-C, and first 10 principal components of genetic ancestry as covariates. Bonferroni-corrected P < .05/2923 (i.e. 1.7 × 10−5) was considered statistically significant. 332 proteins showed significant associations with mLOY after multiple testing correction, including 127 upregulated and 205 downregulated proteins (Figure 1A). The Akt co-activator TCL1A was among the top upregulated proteins consistent with prior evidence of its role in genetic susceptibility to mLOY.6 In contrast, CSF2RA and CD99, both located in the PAR1 region of chromosome Y and previously reported to be downregulated in mLOY leukocytes7 were among the top downregulated proteins, thereby confirming the validity of the mLOY assessment. Pathway enrichment analysis of the 332 mLOY-associated proteins by using the STRING-db database revealed significant overrepresentation of pathways related to inflammation, immune regulation, and extracellular matrix organization (Figure 1C). Proteomic signature of mosaic loss of Y chromosome (mLOY). (A) Circos plot showing the association between mosaic loss of Y chromosome burden and 2923 plasma proteins in 20 645 men from the UK Biobank proteomics subcohort. P-values are truncated at P < 1 × 10−80. Labels and dots are colour-coded: red for positive β coefficients, blue for negative. Analyses were adjusted for age, smoking, BMI, ethnicity, hypertension, diabetes, LDL-C, and genetic ancestry. (B) Venn diagram illustrating overlap among mosaic loss of Y chromosome-associated proteins significantly linked to hypertensive heart disease, heart failure, atrial fibrillation, and aortic aneurysm. (C) Reactome pathway enrichment analysis of 332 proteins significantly associated with mosaic loss of Y chromosome, highlighting biological processes related to inflammation, fibrosis, and immune regulation. (D) Association between each of the 11 overlapping proteins (per 1-SD increase) and both all-cause and cardiovascular mortality, with models adjusted for age, sex, ethnicity, smoking, BMI, hypertension, diabetes, LDL-C, serum creatinine, and the first 10 principal components of genetic ancestry. (E) Reactome pathway enrichment analysis of 47 DEGs in loss of Y chromosome vs Y-positive mononuclear cells Using multivariable-adjusted models, mLOY-associated plasma proteins were linked to the incident cardiovascular diseases heart failure, atrial fibrillation, hypertensive heart disease, and aortic aneurysm. A total of 223 unique proteins were associated with at least one of the four cardiovascular outcomes at a Bonferroni-corrected significance level (P < 1.5 × 10−4). Heart failure demonstrated the highest number of proteomic associations (N = 198) followed by atrial fibrillation (N = 119), aortic aneurysm (N = 56), and hypertensive heart disease (N = 20). Eleven mLOY-upregulated proteins were significantly associated across all four cardiovascular conditions (Figure 1B). These proteins can be clustered into four functional categories: profibrotic mediators (e.g. WFDC2, GDF15, MSLN, TFF2), general inflammatory markers (e.g. PLAUR, CXCL17), monocyte-specific activation markers (e.g. CD300E, LILRA5), and immune modulators or immune suppressive proteins (e.g. LAMP3, TNFRSF10B, CEACAM5). Importantly, increased plasma levels of these proteins were each individually associated with elevated risk for both all-cause and cardiovascular mortality as determined by using Cox proportional hazard models adjusted for age at recruitment, ethnicity, smoking, BMI, prevalent hypertension, prevalent diabetes, LDL-C, and first 10 principal components of genetic ancestry (Figure 1D). Mortality outcomes were ascertained through linkage with national death registry records. The primary cause of death was extracted from official death certificates and classified using the International Classification of Diseases, Tenth Revision (ICD-10). Cardiovascular deaths were defined as ICD-10 codes I00–I99. Median follow-up was 13.7 (interquartile range: 1.6) years. To determine transcriptional changes associated with mLOY, we performed scRNA-seq of circulating mononuclear blood cells from 10 patients with heart failure comparing cells, which lack expression of Y chromosome-encoded genes (mLOY cells), with their Y-expressing counterparts as previously described.8 Y-associated genes were taken from Ensembl's BioMart (https://www.ensembl.org/biomart/martview/ad92e31dd6744901cd7ec4624cc107a1). Cells with combinatorial loss of all Y-associated genes were deemed mLOY cells. Relative changes in transcription of Y harbouring and mLOY cells were then assessed using the FindMarkers function (Seurat v4). Differentially expressed genes served as input to obtain gene ontology terms from Metascape. mLOY cells revealed concordant upregulation of the coding genes for 47 of the 173 (=27%) increased proteins significantly associated with incident heart failure. The upregulated genes in peripheral blood mononuclear cells demonstrated signature pathway enrichment related to inflammation and the innate and adaptive immune system concordant with the proteomic plasma signature (Figure 1E). Specifically, the upregulated genes in mLOY cells included those genes coding for the proteins PLAUR, CD300E, LILRA5, TNFRSF10B, and LAMP3, which all were significantly associated with the incidence for and mortality of all four cardiovascular disease conditions, heart failure, atrial fibrillation, hypertensive heart disease, and aortic aneurysm. This is the first study to comprehensively assess the plasma proteome signature in men with mLOY. Our results disclose a distinctive upregulation of circulating proteins implicated in fibrosis, immune dysregulation, and inflammation, which is mirrored by transcriptional upregulation of numerous genes encoding for the upregulated proteins in mLOY circulating mononuclear cells. The individual association with increased cardiovascular mortality of the mLOY-upregulated proteins WFDC2, GDF15, PLAUR, LAMP3, TNFRSF10B, CXCL17, and CD300E, which were recently shown as major predictors of incident heart failure,9 suggests that mLOY amplifies pathogenic pathways well established to not only contribute to cardiovascular disease incidence but also to cardiovascular mortality and may, thus, offer novel therapeutic targets in men with mLOY. All authors declare no disclosure of interest for this contribution. UK Biobank data are available upon project application. Analyses were performed under the UK Biobank application number 170256. The Else Kroener-Fresenius Foundation and the Deutsche Forschungsgemeinschaft (SFB TRR 219, project ID 322900939) to T.S.; The Hessian Ministry for Science and the Arts (ENABLE) and the Leistungszentrum TheraNova to S.D.; The Deutsche Forschungsgemeinschaft (FOR 5643/1, DI 600/12-1) to S.D. and A.M.Z. The European Union (ERC-2021-ADG, GAP - 101054899, CHIP AVS) to A.M.Z.; The Deutsche Forschungsgemeinschaft (SFB 1531 Z02, FOR 5643 A02) to W.A. The UK Biobank study received approval from the North West Multi-Center Research Ethics Committee. The analyses in the present manuscript were conducted under UK Biobank application number 170256. Not applicable.