Blood Gene Study Nominally Flags SLC51A and TPMT as Aneurysm Markers

Thoracic aortic aneurysm, a silent widening of the body’s largest artery that can end in fatal dissection or rupture, has long lacked a reliable blood test. A new integrative study, published in Artery Research, now nominally prioritizes two genes, SLC51A and TPMT, as candidate diagnostic markers detectable in peripheral blood, and packages them into a simple two-gene risk score. The authors themselves stress that the evidence is exploratory: none of the findings survived false discovery rate correction, and the model has not been tested in any independent cohort.

The research team, led by Junchao Huang and colleagues with senior author Weidong Li, combined three layers of data. First, they analyzed a publicly available gene expression dataset, GSE9106, containing peripheral blood samples from 59 patients with thoracic aortic aneurysm and 34 age- and sex-matched controls from a United States cohort. After standard normalization, batch correction, and probe-to-gene mapping, differential expression analysis with the limma package identified 1,108 differentially expressed genes: 765 significantly upregulated and 343 significantly downregulated in patients, using a false discovery rate adjusted threshold of 0.05 and an absolute log2 fold change above 0.5.

Expression differences alone cannot distinguish genes that drive disease from genes that merely respond to it. To add genetic leverage, the investigators turned to summary data-based Mendelian randomization, a technique that uses genetic variants as instruments to test whether genetically predicted gene expression is associated with disease risk. They drew cis-acting expression quantitative trait loci from the eQTLGen consortium, which covers 31,684 individuals, and combined them with genome-wide association summary statistics for thoracic aortic aneurysm comprising 1,351 cases and 18,295 controls of predominantly European ancestry. Instruments were restricted to strongly significant variants, and the heterogeneity in dependent instruments test was applied to help separate true vertical causality from linkage or pleiotropy.

The Mendelian randomization scan across 15,600 genes flagged 351 genes whose higher genetically predicted expression appeared to increase aneurysm risk and 357 genes whose higher expression appeared protective. Intersecting these with the differentially expressed genes, in a directionally consistent manner, produced a compact list of 14 candidates. Eight risk-direction genes, RBM15, FAHD2B, SLC51A, TPMT, PRPH2, TLCD5, CLCN7 and TNRC6A, overlapped with upregulated transcripts, while six protective-direction genes, AK5, DLEC1, SIDT1, CRYGS, NOC3L and WDR89, overlapped with downregulated ones. All 14 showed nominally significant associations and passed the heterogeneity test, but the authors are explicit that none survived false discovery rate correction, with all adjusted q-values exceeding 0.5.

From this shortlist, the team deployed two machine-learning algorithms to find the most informative markers. Least absolute shrinkage and selection operator logistic regression, tuned by ten-fold cross-validation with the conservative one-standard-error rule, and a random forest classifier with 500 trees, ranked by Gini impurity, converged on the same two genes: SLC51A and TPMT. Individually, each gene showed moderate discrimination between patients and controls. SLC51A achieved an area under the receiver operating characteristic curve of 0.733, with a sensitivity of 72.9 percent and a specificity of 70.6 percent at the optimal cut-off, while TPMT reached 0.742, with 74.6 percent sensitivity and 73.5 percent specificity.

Combining both genes into a multivariable logistic nomogram, with the equation logit(P) = −18.99 + 0.71 × SLC51A + 0.79 × TPMT, yielded an apparent area under the curve of 0.757 in the discovery cohort. Calibration, assessed with 1,000 bootstrap resamples, tracked the ideal diagonal closely, and decision curve analysis suggested theoretical net benefit across a broad range of risk thresholds compared with treating everyone or no one. The authors caution, however, that because feature selection, model fitting, and evaluation all occurred in the same 93-sample dataset, this is an apparent, within-cohort estimate rather than validated performance, and the decision curve results reflect theoretical rather than real-world clinical utility.

To probe what the two genes might be doing biologically, the researchers performed single-gene gene set enrichment analysis, ranking all genes by their correlation with SLC51A or TPMT expression. Low SLC51A expression was associated with reduced activity of RNA-related pathways, including mRNA processing, non-coding RNA transcription, mRNA metabolic processes, and macromolecule catabolism. Low TPMT expression tracked with downregulation of fatty acid beta-oxidation, monocarboxylic acid catabolism, mitotic cell cycle, and mitochondrial RNA catabolic process pathways. Together, the enrichment patterns point toward coordinated alterations in RNA metabolism, macromolecular catabolism, energy utilization, and cell-cycle programs in the blood of aneurysm patients, though the authors label these associations strictly hypothesis-generating.

The two genes are unusual suspects in aortic disease. SLC51A encodes organic solute transporter alpha, a subunit of a transporter complex best studied in bile acid and steroid handling in liver and intestine. Because bile acid signaling through receptors such as FXR and TGR5 is increasingly linked to vascular inflammation, lipid metabolism, and endothelial function, the authors speculate that altered SLC51A expression could reflect changes in bile acid-related metabolic signaling in circulating immune cells or the vessel wall. TPMT encodes thiopurine S-methyltransferase, an enzyme famous for predicting drug metabolism, which also participates in methylation-dependent detoxification and cellular redox balance, processes relevant to the oxidative stress central to aneurysm biology.

The study carries substantial limitations that the authors enumerate candidly. The expression cohort is small, lacks detailed clinical metadata such as hypertension, diabetes, or medication history, and could not be adjusted for confounders. Whole-blood expression may reflect shifts in circulating immune cell composition rather than cell-intrinsic transcriptional changes, and no cell-type deconvolution was performed. The Mendelian randomization used whole-blood expression quantitative trait loci rather than aortic tissue, formal colocalization analysis was not performed, and the heterogeneity test provides only limited evidence against pleiotropy rather than proof of its absence. Ancestry matching was also constrained by the absence of genotype data in the expression dataset.

The overall message is one of disciplined caution. By triangulating blood transcriptomics, large-scale expression genetics, and machine learning, the study offers a reproducible framework for nominating candidate biomarkers in a disease where current risk stratification relies almost entirely on imaging aortic diameter and growth. SLC51A and TPMT now join the shortlist of molecules worth testing in larger, prospectively collected cohorts, alongside functional experiments in vascular smooth muscle, endothelial, and immune cell models to determine whether they actively contribute to aneurysm formation or simply mark disease state. Until such replication succeeds, the two-gene nomogram remains an internally corrected, hypothesis-generating tool, not a test ready for the clinic.

Subject of Research: Candidate blood-based diagnostic genes for thoracic aortic aneurysm identified through integrative transcriptomic and Mendelian randomization analysis

Article Title: Integrative Transcriptomic and Genetic Analysis Nominally Prioritizes SLC51A and TPMT as Candidate Diagnostic Genes for Thoracic Aortic Aneurysm

Article References: Huang, J., Zhou, J., Liu, Y., Yu, J., Chen, S., & Li, W. (2026). Integrative Transcriptomic and Genetic Analysis Nominally Prioritizes SLC51A and TPMT as Candidate Diagnostic Genes for Thoracic Aortic Aneurysm. Artery Research, 32(1), Article 39. https://doi.org/10.1007/s44200-026-00133-y

Image Credits: AI Generated

DOI: 10.1007/s44200-026-00133-y

Keywords: thoracic aortic aneurysm, SLC51A, TPMT, Mendelian randomization, biomarkers, transcriptomics, eQTL, machine learning, nomogram, gene set enrichment analysis, GWAS, diagnostic genes