Investigating the shared genetic information between serum concentration levels of liver enzymes and cholelithiasis

Abstract

Background: Liver injury is associated with cholelithiasis, with changes in liver enzyme levels potentially influencing cholelithiasis risk. This study investigates the shared genetic basis between serum levels of four liver enzymes and cholelithiasis using summary data from large-scale genome-wide association studies (GWAS).

Methods: We assessed genetic correlation between liver enzymes and cholelithiasis, and performed local genetic correlation analysis to identify shared genomic regions. A cross-trait meta-analysis identified significant SNPs (single nucleotide polymorphisms) shared between the enzymes and cholelithiasis. To explore causal effects, we applied both two-sample Mendelian randomization and multivariable Mendelian randomization. Heritability-based enrichment analysis was employed to identify tissues and cells jointly associated with liver enzymes and cholelithiasis, while summary-data-based Mendelian Randomization (SMR) was utilized to identify shared genes.

Results: Alanine aminotransferase (ALT) and gamma-glutamyl transferase (GGT) showed relatively stronger genetic correlations with cholelithiasis compared to the other liver enzymes. Shared SNPs were identified among ALT, GGT, alkaline phosphatase (ALP), and cholelithiasis. Mendelian randomization indicated that a tenfold increase in ALT could raise cholelithiasis risk by 203.4%. The liver was identified as the primary tissue linking these enzymes to cholelithiasis, but no shared cell types were implicated. Several candidate genes, such as SPTLC3, may jointly influence liver enzyme levels and cholelithiasis risk.

Conclusions: Elevated ALT levels may increase cholelithiasis risk. Genetic associations across tissues, genes, and SNPs suggest that liver enzymes could mediate the relationship between liver injury and cholelithiasis risk, providing insights into shared genetic mechanisms with potential implications for future research.

Keywords: Cholelithiasis; GWAS; Liver enzymes; Liver injury; Shared genetic information.

Fig. 1

Overview of the analytical framework. This study integrated multi-omic datasets and GWAS summary data for liver enzymes and cholelithiasis. Genetic correlation (LDSC), local genetic correlation (LOGODetect), cross-trait meta-analysis (MTAG), two-sample Mendelian randomization (CAUSE), and multivariable Mendelian randomization (MR-BMA) were used to identify shared genetic loci, causal effects, and phenotype-associated tissues, genes, and cell types. (Abbreviations in parentheses denote the analysis tools used.)

Fig. 2

Genetic correlation between traits estimated by LDSC. The bar chart represents the estimated genetic correlations between each trait, with error bars indicating the 95% confidence intervals (CI) for these estimates. CHO Cholelithiasis

Fig. 3

Mirror Manhattan plots showing genomic regions with local genetic correlations between liver enzyme levels and cholelithiasis identified by LOGODetect. Panels (A–D) display the local genetic correlation signals between cholelithiasis and ALT, GGT, ALP, and AST, respectively. The x-axis indicates chromosomal positions across autosomes (chr1–chr22). The y-axis represents–log₁₀(p) values for each locus. Each plot is mirrored around the horizontal axis: signals related to liver enzymes are shown above, and those related to cholelithiasis are mirrored below. SNPs within regions of significant local genetic correlation—highlighted in green—were used to pinpoint the locations of shared genetic loci

Fig. 4

Shared genomic regions between liver enzymes and cholelithiasis. A-D Genomic regions showing significant genetic correlation between cholelithiasis and each liver enzyme: ALT (A), GGT (B), ALP (C), and AST (D). Each dot represents a genomic region, plotted at its midpoint in megabases (Mb) along the x-axis, with statistical significance indicated by -log₁₀(q-value) on the y-axis. Vertical segments denote the full span of each region. Regions with positive or negative effect directions are colored in blue and red, respectively. Chromosomes are shown in separate panels with free x-axis scales

Fig. 5

Causal inference of the effect of liver enzyme levels on cholelithiasis using two-sample Mendelian randomization. A Causal estimates obtained using five classical MR methods: inverse-variance weighted (IVW), MR-Egger, weighted median, weighted mode, and simple mode. Odds ratios (ORs) and 95% confidence intervals (CIs) are shown for each liver enzyme. B Causal estimates derived using the CAUSE (Causal Analysis Using Summary Effect estimates) method. The number of instrumental SNPs used in each analysis is indicated. In CAUSE analysis, “Share q” is a model parameter indicating the assumed proportion of shared variants, whereas “CAUSE q” is the posterior estimate inferred from the data, reflecting the strength of correlated pleiotropy. OR odds ratio, CI confidence interval, SNP, n number of instrumental variants

Fig. 6

Tissue-specific heritability enrichment of liver enzymes and cholelithiasis using LDSC-SEG. (A-E) Tissue-specific enrichment of SNP heritability for ALT (A), GGT (B), ALP (C), AST (D), and cholelithiasis (E) across 53 human tissues. The x-axis represents the statistical significance of enrichment as -log₁₀(p), and the y-axis lists the corresponding tissues. The dashed vertical line indicates the nominal significance threshold of P = 0.05 (-log₁₀(p) ≈ 1.3); values to the right of the line (i.e., P < 0.05) are considered statistically significant

Fig. 7

Cell-type-specific enrichment of AST based on LDSC-SEG using human single-cell transcriptomic data. (A-E) Tissue-specific enrichment of SNP heritability for ALT (A), GGT (B), ALP (C), AST (D), and cholelithiasis (E) across 53 human tissues. The x-axis represents the statistical significance of enrichment as -log₁₀(p), and the y-axis lists the corresponding tissues. The dashed vertical line indicates the nominal significance threshold of P = 0.05 (-log₁₀(p) ≈ 1.3); values to the right of the line (i.e., P < 0.05) are considered statistically significant