The role of inflammatory proteins in mediating the effect of cathepsins on cholelithiasis: A mendelian randomization study

Abstract

Objective

Previous studies have suggested potential associations between cathepsins and cholelithiasis. This study aimed to investigate the potential causal relationships among cathepsins, inflammatory proteins, and cholelithiasis using Mendelian randomization (MR), and to explore whether inflammatory proteins mediate the association between cathepsins and cholelithiasis.

Methods

We conducted a two-sample MR study using genome-wide association study (GWAS) summary statistics for cathepsins, inflammatory proteins, and cholelithiasis. Genetic associations for cathepsins were obtained from the INTERVAL study of European-ancestry individuals. GWAS data for inflammatory proteins and cholelithiasis were obtained from publicly available datasets, including European-ancestry cholelithiasis datasets from FinnGen and GCST90044196. The inverse variance weighted method was used as the primary MR analysis, with complementary and sensitivity analyses performed to assess the robustness of the findings. Multivariable and mediation MR analyses were performed to evaluate whether inflammatory proteins mediated the pathway from cathepsins to cholelithiasis.

Results

After correction for multiple testing, genetically predicted cathepsin B was associated with an increased risk of cholelithiasis. Mediation analyses suggested that part of the effect of cathepsin B on cholelithiasis may be mediated by fibroblast growth factor 19 (FGF19) and interleukin-6 levels. In the reverse MR analysis, genetically predicted cholelithiasis showed potential associations with two cathepsins and twelve inflammatory proteins.

Conclusions

This MR study provides genetic evidence for a potential association between cathepsin B and increased cholelithiasis risk. FGF19 and interleukin-6 may serve as partial mediators in this association.

1. Introduction

Cholelithiasis is a common chronic digestive system disorder with an increasing incidence, contributing to a growing economic burden on society. ¹ The rising prevalence of cholelithiasis has been partly attributed to changes in dietary habits, particularly increased consumption of high-calorie foods and saturated fat. In contrast, adherence to a Mediterranean dietary pattern may reduce the likelihood of cholelithiasis. ¹ Anatomically, cholelithiasis can involve the intrahepatic bile ducts, hilar bile duct region, extrahepatic bile ducts, and gallbladder. ² Although many patients remain asymptomatic, cholelithiasis may lead to severe complications, including pancreatitis, which can be necrotizing, hemorrhagic, and life-threatening. It may also contribute to malignant transformation in the biliary tract. ³ Cholelithiasis is also closely associated with genetic susceptibility and metabolic risk factors, including obesity, dyslipidemia, and type 2 diabetes.^4,5 These considerations highlight the need to better understand the biological mechanisms underlying cholelithiasis.

Cathepsins, including serine, cysteine, or aspartic residue-based proteases, play crucial roles in a series of functions such as digestion, blood clotting, immune responses, lipid synthesis, hormone release, and peptide synthesis. However, their normal functions are fundamentally affected by the inflammatory state. Many diseases are attributed to the dysregulation of cathepsins, including arthritis, periodontitis, pancreatitis, atherosclerosis, obesity, and more. ⁶ A previous study reported a 25%-50% lower activity of cathepsin L, D, and B in gallbladder mucosa of cholelithiasis patients in comparison to those without cholelithiasis. ⁷ Nevertheless, further research is warranted, and the causality between cathepsins and cholelithiasis remains unclear.

Basic research results indicates that inflammation might be involved in the formation of gallstones.^8,9 Population-based studies also indicate an association between circulating inflammatory factors and the occurrence of cholelithiasis. ¹⁰ Su and colleagues validated a unidirectional causal relationship between four cytokines and growth factors and cholelithiasis. ¹¹ Currently, therapeutic options for cholelithiasis are limited. Exploring potential targets related to inflammatory regulation may provide useful clues for future prevention or treatment strategies for cholelithiasis.

Mendelian randomization (MR) can infer causal relationships between exposure and outcomes and is less susceptible to confounding factors. ¹² In this study, we undertook a comprehensive MR analysis to investigate the causality between cathepsins, inflammatory proteins, and cholelithiasis. Subsequently, we investigated the potential role of inflammatory proteins as mediators in the pathway from cathepsins to cholelithiasis. Additionally, employing reverse causality analysis, we examined whether the genetic susceptibility to cholelithiasis risk influences the levels of cathepsins and inflammatory proteins.

2. Materials and methods

2.1. Study design

The study comprises three primary parts, as illustrated in Figure 1: the analysis of causality of 9 cathepsins on cholelithiasis (step 1A); the analysis of causality of 91 inflammatory proteins on cholelithiasis (step 2A); and the mediation analysis of inflammatory proteins in the pathway from cathepsins to cholelithiasis (step 3). Single-nucleotide polymorphisms (SNPs) were defined as instrumental variables (IVs) in this research. This MR analysis utilized data from publicly genome-wide association study (GWAS) consortia, as detailed in Table S1. The analysis relied on three fundamental assumptions: (1) the IVs are closely linked to the exposure factors; (2) IVs are not correlated with confounding factors; (3) IVs do not directly influence the outcome but only through the impact on the exposure. ¹² All the studies included in the analysis had received approval from their respective ethical review boards. Furthermore, as this study solely relied on aggregated data and did not involve individual-level data, the approval of additional ethical review board was not deemed necessary.

Figure 1.

Study design. In analysis process, Step 1A represents the causal effects of cathepsins on cholelithiasis. Step 1B represents the causal effects of cholelithiasis on cathepsins. Step 2A represents the causal effects of inflammatory proteins on cholelithiasis. Step 2B represents the causal effects of cholelithiasis on inflammatory proteins. Step 3 represents the mediating analysis of inflammatory proteins in the pathway from cathepsins to cholelithiasis: path a was the causal effect of cathepsins on inflammatory proteins; path b was the causal effect of inflammatory proteins on cholelithiasis; path c was the total effect of cathepsins on cholelithiasis. MR, Mendelian randomization; MR-PRESSO, Mendelian randomization pleiotropy residual sum and outlier.

2.2. Data source

Genetic instruments for assessing the levels of 9 cathepsins were obtained from the INTERVAL study, encompassing 3,301 European ancestry. ¹³ Detailed data can be obtained at https://gwas.mrcieu.ac.uk. The data for inflammatory proteins were sourced from a prior GWAS involving 14,824 individuals, covering 91 inflammatory proteins. ¹⁴ A portion of the GWAS summary data for cholelithiasis was obtained from the tenth version of the FinnGen consortium, accessible at https://www.finngen.fi/fi. ¹⁵ The remaining data was obtained from the research conducted by Jiang et al., available in the GWAS Catalog database with the accession number GCST90044196 (https://www.ebi.ac.uk/gwas/). ¹⁶ In the GWAS study of cholelithiasis, the FinnGen consortium had a total of 361,641 European ancestry controls and 40,191 European-ancestry cases, while the GCST90044196 study included 7,426 cases and 448,922 controls.

2.3. Instrumental variable selection

To obtain sufficient instrumental variables for each cathepsin and inflammatory protein, we adopted a relaxed threshold of P < 5 × 10^-6 for SNP selection, as only a limited number of genome-wide significant variants were available for several circulating proteins. Subsequently, SNPs demonstrating linkage disequilibrium (LD) were excluded from the analysis. For SNPs strongly associated with cholelithiasis, LD was considered acceptable if the condition r² < 0.001 within a 10,000 kb window was met. ¹⁷ The exposure and outcome datasets were harmonized to align effect alleles. SNPs absent from the outcome datasets or defined as ambiguous, including palindromic SNPs with minor allele frequencies between 0.42 and 0.58, were excluded to avoid potential allele misalignment.

Relevant information, including SNP, standard error (SE), effect sizes (Beta), other allele (OA), effect allele (EA), effect allele frequency (EAF), and P-value, was extracted. Finally, we calculated F-statistic parameters and the explained variance (R²) to assess whether the identified IVs were strongly connected with the exposure. SNPs with F-statistic parameters below 10 were generally regarded as weak instruments and were subsequently removed. ¹⁸ In this study, R² was computed as 2 × EAF × (1-EAF) × Beta ² /(2 × EAF × (1-EAF) × Beta ² + 2 × EAF × (1-EAF) × N × SE ² ), where N is the sample size of the GWAS, and F was calculated as R² × (N-2)/(1-R²). ¹⁹

To further evaluate the potential influence of weak instrument bias, we summarized the mean, median, minimum, and maximum F-statistics of the final instrumental variables retained after LD clumping, harmonization, removal of ambiguous SNPs, and exclusion of weak instruments. We also assessed the number of available SNPs under stricter thresholds of P < 1 × 10^-6 and P < 5 × 10^-8. However, because the number of eligible SNPs was markedly reduced under these stricter thresholds, with several cathepsins retaining only one or no SNP, formal stricter-threshold multi-SNP MR sensitivity analyses were not performed.

2.4. Primary MR analysis

To assess the causal impacts of cathepsins and inflammatory proteins on cholelithiasis, we conducted two-sample MR analyses separately for step 1A and step 2A, as illustrated in Figure 1. The Inverse Variance Weighted (IVW) under a random effects model served as the primary analysis method. ²⁰ Complementary analyses included the MR-Egger and Weighted Median methods. Associations were considered statistically significant when the false discovery rate (FDR)-adjusted P value of the primary IVW analysis was <0.05 and the effect estimates from the weighted median and MR-Egger analyses showed concordant directions. Estimates for each outcome from various sources were pooled using random effects meta-analysis. Results from the MR analyses were presented as odds ratios (ORs) with corresponding 95% confidence intervals (CIs) for categorical variables and as Beta (95% CI) for continuous variables. The Benjamini-Hochberg method, controlling the FDR, was used for multiple testing correction. Associations with a nominal P < 0.05 but a Benjamini–Hochberg adjusted P > 0.05 were considered suggestive, while associations with a Benjamini–Hochberg adjusted P < 0.05 were considered significant.

2.5. Bi-directional causality analysis

To assess the bidirectional causation effect among cathepsins, inflammatory proteins, and cholelithiasis, we adopted a reverse MR analysis approach, treating cholelithiasis as the “exposure” and cathepsins and inflammatory proteins as the “outcome” (step 1B and step 2B in Figure 1). In this analysis, cholelithiasis-associated SNPs were selected as IVs using the genome-wide significance threshold of P < 5 × 10^–8. This stricter threshold was used because a sufficient number of genome-wide significant SNPs were available for cholelithiasis, which helped reduce the risk of weak instrument bias in the reverse MR analyses.

2.6. Mediation analysis

In the two-sample analysis (step 1A and step 2A in Figure 1), cathepsins and inflammatory proteins exhibiting meaningful causal impacts on cholelithiasis were selected for inclusion in the mediation analysis. Specifically, the MR effect estimates for cathepsins on each inflammatory protein were obtained using the IVW method (step 3, path a, in Figure 1). Subsequently, a multivariable MR analysis was carried out to assess the impact of inflammatory proteins on the risk of cholelithiasis while adjusting for cathepsins (step 3, path b, in Figure 1). The indirect effect was calculated by multiplying the MR estimate for the cathepsin–mediator association by the multivariable MR estimate for the mediator–cholelithiasis association. Lastly, the mediating effect was divided by the total effect to calculate the proportion of the mediating effect in the total effect (step 3, path c, in Figure 1). Mediation MR requires valid instruments that are strongly associated with the exposure, independent of confounders, and affect the outcome only through the exposure–mediator pathway. The mediated proportion was calculated as the indirect effect divided by the total effect, and all mediation estimates were calculated on the beta/log-odds scale. The 95% CIs for the mediated proportions were estimated using the delta method.

2.7. Sensitivity analysis

Cochran’s Q test was conducted to assess the heterogeneity of each SNP. ²¹ Additionally, Mendelian Randomization Pleiotropy Residual Sum and Outlier (MR-PRESSO) and MR-Egger intercept were employed to examine possible horizontal pleiotropic effects. MR-PRESSO was utilized to identify notable outliers and address horizontal pleiotropy by excluding such data points. ²² All analyses were conducted using R statistical software (version 4.3.1). The MR analysis was executed using the “TwoSampleMR” package. The “MR-PRESSO” package was employed to detect horizontal pleiotropy and identify outlier SNPs, and the “mediation” package was used to assess the intermediation effect.^23,24

3. Results

3.1. Causal effects of cathepsins and inflammatory proteins on cholelithiasis

Initially, we identified 133 SNPs associated with 9 cathepsins (Table S2). After harmonization and exclusion of weak instruments, 125 cathepsin SNPs were retained in the FinnGen consortium and 124 cathepsin SNPs were retained in the GCST90044196 study. In the FinnGen consortium, the overall mean and median F-statistics were 31.2 and 22.5, respectively, with a minimum F-statistic of 20.8. Similar instrument strength was observed in the GCST90044196 study. These results suggested that weak instrument bias was unlikely to substantially influence the MR estimates (Table S3). As depicted in Figure 2, genetically predicted cathepsin B was significantly associated with an increased risk of cholelithiasis after FDR correction (OR 1.06, 95% CI: 1.03 to 1.10, P = 3.01E-04; Figure 2, Table S4). The outcomes from the weighted median and MR-Egger were consistent (Table S5). MR-PRESSO identified 2 outliers in the analysis conducted in the FinnGen consortium; however, even after removing these outliers, the association persisted (Table S5).

Figure 2.

Associations of genetic liability to 9 cathepsins with cholelithiasis. *Significant association after multiple testing. The estimate of 9 cathepsins with cholelithiasis was meta-analysis by combining estimates from the FinnGen consortium and the GCST90044196 study. IVW, Inverse variance weighted; OR, odds ratio; CI, confidence interval.

For inflammatory proteins, we initially identified 1,817 SNPs associated with 91 inflammatory proteins (Table S6). As illustrated in Figure 3, genetically predicted fibroblast growth factor 19 (FGF19) levels were significantly associated with a decreased risk of cholelithiasis after FDR correction (OR 0.81, 95% CI: 0.72 to 0.92, P = 0.001; Tables S7 and S8). In addition, seven inflammatory proteins showed suggestive associations with cholelithiasis based on nominal significance. Delta and Notch-like epidermal growth factor-related receptor levels, leukemia inhibitory factor receptor levels, and TNF-related activation-induced cytokine levels were suggestively associated with a decreased risk of cholelithiasis, whereas C-X-C motif chemokine 10, interleukin-15 receptor subunit alpha, interleukin-6, and matrix metalloproteinase-1 were suggestively associated with an increased risk of cholelithiasis (Figure 3, Tables S7 and S8). The sensitivity analysis results demonstrated overall consistency (Table S7). Cochran’s Q test indicated SNP-level heterogeneity in 5 of the 8 inflammatory proteins. MR-Egger intercept tests indicated horizontal pleiotropy for TNF-related activation-induced cytokine levels in the FinnGen consortium but not for any other inflammatory proteins in either of the sources. MR-PRESSO identified 1–3 outliers in the analyses for FGF19 levels; nevertheless, the association persisted after removing these SNPs (Table S7). The association between FGF19 levels and cholelithiasis also remained significant after multiple comparison corrections (Table S8).

Figure 3.

Associations of genetic liability to 8 inflammatory proteins with cholelithiasis. *Significant association after multiple testing. The estimate of 8 inflammatory proteins with cholelithiasis was meta-analysis by combining estimates from the FinnGen consortium and the GCST90044196 study. IVW, Inverse variance weighted; OR, odds ratio; CI, confidence interval.

3.2. Reverse MR analysis assessing the causal impact of cholelithiasis on cathepsins and inflammatory proteins

As shown in Table S9, genetic susceptibility to cholelithiasis showed suggestive associations with higher levels of cathepsin L2 (Beta 0.07, 95% CI: 0.01 to 0.14, P = 0.022) and cathepsin O (Beta 0.09, 95% CI: 0.02 to 0.15, P = 0.006). However, these associations did not remain significant after FDR correction (Table S10).

As shown in Table S11, genetic susceptibility to cholelithiasis showed suggestive associations with increased levels of 11 inflammatory proteins, including CD40L receptor (Beta 0.04, 95% CI: 0.01 to 0.18, P = 0.017), C-X-C motif chemokine 1 (Beta 0.03, 95% CI: 0.00 to 0.06, P = 0.048), C-X-C motif chemokine 10 (Beta 0.04, 95% CI: 0.00 to 0.07, P = 0.039), C-X-C motif chemokine 11 (Beta 0.03, 95% CI: 0.00 to 0.06, P = 0.035), eotaxin (Beta 0.05, 95% CI: 0.02 to 0.08, P = 0.002), interleukin-15 receptor subunit alpha (Beta 0.04, 95% CI: 0.00 to 0.07, P = 0.038), matrix metalloproteinase-1 (Beta 0.05, 95% CI: 0.02 to 0.08, P = 0.003), monocyte chemoattractant protein-4 (Beta 0.05, 95% CI: 0.02 to 0.08, P = 0.003), programmed cell death 1 ligand 1 (PD-L1) (Beta 0.06, 95% CI: 0.03 to 0.10, P = 3.30E-04), stem cell factor (Beta 0.04, 95% CI: 0.00 to 0.08, P = 0.037) and tumor necrosis factor ligand superfamily member 12 (Beta 0.04, 95% CI: 0.01 to 0.07, P = 0.019). In addition, genetic susceptibility to cholelithiasis showed a suggestive association with decreased FGF19 levels (Beta -0.08, 95% CI: -0.14 to -0.02, P = 0.005). After FDR correction, only the association between cholelithiasis and PD-L1 levels remained significant (Table S12).

3.3. Mediation analysis

In this study, cathepsin B and eight inflammatory proteins showed potential causal associations with cholelithiasis. Demonstrating that inflammatory proteins mediate the pathway from cathepsin B to cholelithiasis requires the premise that cathepsin B was significantly associated with inflammatory proteins. Further analysis showed that genetically predicted cathepsin B was significantly associated with higher interleukin-6 levels after FDR correction and was suggestively associated with lower FGF19 levels based on nominal significance (Figure 4, Tables S13 and S14). Our multivariable MR analysis results suggested that FGF19 and interleukin-6 may serve as partial mediators in the association between cathepsin B and cholelithiasis. The cathepsin B–cholelithiasis effect was reduced from 1.06 (95% CI: 1.03 to 1.10) to 1.03 (95% CI: 0.98 to 1.08) after adjustment for FGF19 levels and to 1.04 (95% CI: 1.01 to 1.07) after adjustment for interleukin-6 levels. After simultaneous adjustment for both FGF19 and interleukin-6 levels, the effect estimate was further reduced to 0.99 (95% CI: 0.95 to 1.03) (Table 1).

Figure 4.

Associations of genetic liability to cathepsin B with 8 inflammatory proteins. *Significant association after multiple testing. IVW, Inverse variance weighted; CI, confidence interval.

Table 1.

Estimates of the mediating effects of inflammatory proteins on the association between cathepsin B and cholelithiasis.

Model	OR (95% CI)	P	Mediated proportion, % (95% CI)
MR-IVW regression, crude	1.06(1.03 to 1.10)	3.01E-04	​
Multivariate model
(1) Adjusted for FGF19 levels	1.03(0.98 to 1.08)	0.294	16.7 (-0.6 to 33.1)
(2) Adjusted for interleukin-6 levels	1.04(1.01 to 1.07)	0.010	1.7 (-8.9 to 12.2)
(3) Adjusted for FGF19 levels + interleukin-6 levels	0.99(0.95 to 1.03)	0.640	-

Mediated proportions were calculated on the beta/log-odds scale, and 95% CIs were estimated using the delta method. The combined mediated proportion was not calculated because the simultaneous adjustment model does not yield a directly interpretable single mediated proportion. Abbreviations: CI, confidence interval; FGF19, fibroblast growth factor 19; IVW, inverse variance weighted; MR, Mendelian randomization; OR, odds ratio.

4. Discussion

In this MR study of 9 cathepsins and 91 inflammatory proteins, genetically predicted cathepsin B was associated with an increased risk of cholelithiasis. Four inflammatory proteins showed positive associations with cholelithiasis risk, while another four showed inverse associations. Multivariable MR analyses suggested that FGF19 and interleukin-6 may partially mediate the association between cathepsin B and cholelithiasis. Intriguingly, the inverse MR analysis suggested potential effects of genetically predicted cholelithiasis on the levels of two cathepsins and twelve inflammatory proteins.

Cathepsin B is a lysosomal cysteine protease involved in lysosome-mediated cell death, protein degradation, inflammation, lipid metabolism, and tumor progression. ²⁵ In this study, genetically predicted circulating cathepsin B was associated with an increased risk of cholelithiasis, which appears inconsistent with a previous observational study reporting 25%–50% lower activities of cathepsin L, D, and B in gallbladder mucosa from patients with cholelithiasis. ⁷ This discrepancy may be explained by differences in tissue source, disease stage, and study design. The previous study measured local cathepsin activity in gallbladder mucosa from patients with established cholelithiasis, whereas our MR analysis evaluated the lifelong genetic predisposition to circulating cathepsin B levels. Therefore, reduced cathepsin activity in gallbladder mucosa may reflect local tissue remodeling or secondary changes after cholelithiasis development, rather than the initial causal process. Circulating cathepsin B may influence cholelithiasis through extra-gallbladder pathways, such as systemic inflammation, lipid metabolism, or biliary tract microenvironmental changes. In addition, cholelithiasis has been associated with an increased risk of biliary tract malignancies, although the magnitude and pattern of risk may vary according to the anatomical location of calculi. Mehra and colleagues demonstrated significantly elevated cathepsin B and L levels in patients with gallbladder cancer compared with controls. ²⁶ Given the multifaceted roles of cathepsin B, our findings suggest that cathepsin B may represent a potential target for future studies on cholelithiasis prevention or treatment. Further studies are also needed to clarify whether cathepsin B contributes to biliary tract carcinogenesis through its potential role in cholelithiasis development.

Among the factors contributing to cholelithiasis, inflammation plays a crucial role. ¹ In our MR analysis, we found that FGF19 reduces the risk of cholelithiasis, while interleukin-6 increases the risk. FGF19 is an intestinal hormone that regulates glucose, lipid, and energy balance in the body. ²⁷ It also regulates bile acid homeostasis and gallbladder filling. After meals, the release of substances from food and bile increases the levels of intestinal bile acids, activating the Farnesoid X receptor (FXR) in intestinal epithelial cells. This induces the upregulation and secretion of FGF19, which enters the liver to inhibit the level of cytochrome P450 family 7 subfamily A member 1 (CYP7A1), forming a negative feedback loop to regulate bile acid synthesis.^28–30 FGF19 also relaxes gallbladder smooth muscles, promoting gallbladder filling. The interplay between FGF19 and Cholecystokinin (CCK) controls the gallbladder’s cycle of filling and emptying, as well as the flow of bile into the intestine to aid in digestion. ³¹ The factors contributing to cholesterol gallstone formation may include bile composition (with cholesterol supersaturation), gallbladder motility disorders, inflammation, excessive secretion of gallbladder mucin gel, slow colon motility, and increased intestinal cholesterol absorption. ³² Gallbladder emptying after meals is mainly regulated by CCK, and in patients with high triglyceride levels, there is reduced sensitivity of the gallbladder to CCK, leading to impaired gallbladder motility. This sensitivity improves after using bezafibrate or fish oil to lower serum triglyceride levels. ³³ Previous studies have reported that under high-fat diet conditions, the application of recombinant FGF19 protein and FGF19 transgenic mice effectively control mouse weight by reducing fat content. These mice also show decreased concentrations of circulating triglycerides, cholesterol, and insulin in the blood, as well as decreased liver triglyceride content. ³⁴ Kim and colleagues found that postprandial FGF19 and Small Heterodimer Partner (SHP) inhibit sterol regulatory element-binding transcription factor 2 (SREBF2), thereby suppressing the expression of intestinal NPC1-like intracellular cholesterol transporter 1 (NPC1L1) and cholesterol absorption. ³⁵ Therefore, we hypothesize that FGF19 may reduce the risk of cholelithiasis by lowering plasma cholesterol and triglyceride concentrations. However, owing to the complete overlap of the receptor, fibroblast growth factor receptor 4 (FGFR4), in both bile acid regulation and hepatocyte proliferation, modulation of the FGF19/FGFR4 axis for diseases involving bile acid dysregulation and gallstone formation will need to be considered carefully. ²⁷

Since being identified over 40 years ago as a soluble factor secreted by T cells, the interleukin-6 pathway has become a key player in health immune regulation and the dysregulation seen in various diseases. ³⁶ Targeting the interleukin-6 pathway has emerged as an innovative therapeutic approach for various rheumatic diseases and cytokine release syndromes. Previous reports have also explored the relationship between interleukin-6 and cholelithiasis. ¹⁰ A study utilizing a guinea pig model to explore the preventive effects of Gallbladder Emptying Granules (LDG) on gallstones indicates that LDG treatment can improve the functioning of the hypothalamus-pituitary-adrenal (HPA) axis, decrease the production of inflammatory proteins like IL-1, IL-6, and tumor necrosis factor-alpha (TNF-α) in serum, thereby preventing cholelithiasis. ³⁷ Moreover, previous studies have indicated that in patients with cholelithiasis, both circulating interleukin-6 protein levels and liver interleukin-6 mRNA expression levels are elevated.^38,39 However, Ebadi and colleagues studied the gene polymorphisms of various cytokines in the bile of 254 normal individuals and 158 gallstone patients, evaluating their impact on gallstone formation, but the study did not establish a notable association between polymorphisms in TNF-α and IL-6 and the development of gallstones. ⁴⁰ We speculate that the inconsistencies among these studies may be attributed to differences in the composition and function of plasma, bile, and gallbladder mucosal proteins.

The present study evaluated the potential mediating roles of FGF19 and interleukin-6 in the association between genetic predisposition to cathepsin B and cholelithiasis. Our results suggest that these two inflammatory proteins may partly mediate the effect of cathepsin B on cholelithiasis, providing genetic clues for future preventive and therapeutic exploration. However, statistical significance in MR analyses does not necessarily indicate immediate clinical applicability. Given the modest effect sizes observed in this study, the clinical relevance of these associations remains uncertain and should be evaluated in future experimental and prospective clinical studies. With the increasing application of artificial intelligence (AI) in clinical practice, AI-based models may help integrate genetic, inflammatory, metabolic, and clinical information to support individualized risk prediction and treatment strategies for patients at risk of cholelithiasis. ⁴¹

In the reverse MR analysis, PD-L1 was the only inflammatory protein that remained significantly associated with genetically predicted cholelithiasis after FDR correction. This finding suggests that genetic susceptibility to cholelithiasis may be linked to increased circulating PD-L1 levels. One possible explanation is that cholelithiasis-related chronic biliary inflammation may promote immune activation and upregulation of immune checkpoint signals such as PD-L1. However, this result should be interpreted cautiously, because reverse MR reflects genetic liability rather than clinically established disease status. Further experimental and clinical studies are needed to clarify whether cholelithiasis can induce PD-L1 upregulation and to determine its potential clinical implications.

The primary robustness of the present research lies in its study design, which effectively reduces potential biases arising from reverse causality and residual confounding. Additionally, we conducted analyses using two independent data sources, and the consistency observed in results between these sources makes it less likely that the observed associations were coincidental. The exploration of mediating pathways through MVMR analysis not only enhanced mechanistic insights but also provided clues for future investigation of potential prevention strategies.

Indeed, this study has several limitations. First, although we performed multiple sensitivity analyses, the possibility of horizontal pleiotropy could not be completely excluded, as genetic variants may affect the outcome through pathways other than the exposure. Second, although all included instrumental variables had F-statistics greater than 10, the use of a relaxed threshold for instrument selection may still increase the risk of weak instrument bias. Third, the observed effect sizes were modest; therefore, the clinical significance of these associations should be interpreted cautiously. Fourth, although FGF19 and interleukin-6 may partially mediate the association between cathepsin B and cholelithiasis, the magnitude of effect reduction was modest, and the clinical significance of these mediation effects remains uncertain. In addition, the mediation analysis was based on MR assumptions and statistical decomposition of effects. Therefore, further experimental studies are needed to confirm whether FGF19 and interleukin-6 are true biological mediators in the pathway from cathepsin B to cholelithiasis. Fifth, because MR estimates reflect the lifelong effects of genetic predisposition, they may not be directly comparable with findings from observational or interventional studies. Finally, this study was mainly based on individuals of European ancestry, which may limit the generalizability of our findings to other populations. Given potential differences in genetic background, dietary patterns, and metabolic risk profiles across populations, further studies in diverse ancestries are needed to validate these associations.

5. Conclusion

This MR study provides genetic evidence for a potential causal association between cathepsin B and increased cholelithiasis risk. FGF19 and interleukin-6 may partially mediate this association. Future experimental studies and prospective clinical research are needed to validate these findings and clarify their potential value in risk assessment or therapeutic exploration.

Appendix.

Abbreviations

MR

Mendelian randomization

SNPs

Single-nucleotide polymorphisms

IVs

instrumental variables

GWAS

genome-wide association study

LD

linkage disequilibrium

SE

standard error

EA

effect allele

EAF

effect allele frequency

IVW

Inverse Variance Weighted

ORs

odds ratios

CIs

confidence intervals

FDR

false discovery rate

MR-PRESSO

Mendelian Randomization Pleiotropy Residual Sum and Outlier

FGF19

Fibroblast growth factor 19

PD-L1

Programmed cell death 1 ligand 1.

ORCID iD

Tong Yuan https://orcid.org/0009-0005-2585-4154

Ethical considerations

This study used publicly available, de-identified GWAS summary-level data. Ethical approval and informed consent were obtained in the original studies; therefore, no additional ethical approval or informed consent was required for the present analysis.

Data Availability Statement

Data described in the manuscript are provided within the article. The UKBiobank HRC-imputed can be obtained via https://pheweb.org/UKB-SAIGE/. The summary-level data for FinnGen can be obtained via https://www.finngen.fi/en/access_results.