The role of lipid-lowering drugs in urologic and male reproductive cancers: a Mendelian randomization study

Introduction

Urological and male reproductive system cancers—including malignancies of the kidney, renal pelvis, bladder, prostate, and testis—remain a major source of cancer morbidity and mortality among men, with incidences rising steadily as populations age (1). Parallel to this trend, cardiovascular disease has persisted as a dominant global health burden, particularly in individuals with metabolic disturbances such as dyslipidemia and obesity—conditions that frequently coexist with urological cancers (2). Their shared epidemiologic landscape suggests not coincidence but rather overlapping biological foundations rooted in metabolic dysfunction. Among these shared pathways, lipid metabolism is particularly noteworthy. Tumor cells and immune cells reconfigure lipid synthesis, transport, and oxidation to fuel growth, reprogram signaling, and reshape the tumor microenvironment. Thus, lipid metabolism may be a potent driver—and occasionally a vulnerability—of malignant progression (3). Complementing these mechanistic insights, human genetic studies have begun to link lipid-related processes to cancer susceptibility. Recent lipidomic and metabolomic Mendelian randomization (MR) analyses have linked circulating metabolites and lipid species to the risks of bladder, kidney, prostate, and other urological cancers, suggesting that lipid pathways may influence cancer development in tumor-specific and metabolite-specific ways (4-6).

These findings naturally suggest that perturbation of lipid-lowering pathways may influence cancer risk. However, the current evidence base remains incomplete. Existing studies have mainly focused on selected pathways, particularly 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR) inhibition and prostate cancer, whereas evidence for other lipid-lowering targets across urological and male reproductive cancers remains limited. Prior evidence has also not always clearly distinguished cancer incidence from tumor progression, aggressiveness, or therapeutic response, which may represent biologically distinct processes (5-7). Accordingly, the added value of the present study lies not in reassessing HMGCR inhibition in isolation, but in systematically comparing both established and newer lipid-lowering targets across multiple tumor types within the same causal framework.

Drug-target MR extends this framework by using variants within or near drug-target genes to proxy the long-term perturbation of specific therapeutic pathways (8). This approach is particularly suitable for comparing both low-density lipoprotein cholesterol (LDL-C)-related and triglyceride (TG)-related targets across different tumors. It is also relevant for newer targets such as PCSK9 and ANGPTL3. Although these proteins are best known for their roles in hepatic lipid regulation, they may also influence systemic lipoprotein handling and broader immune-metabolic signaling beyond the liver, making associations with extrahepatic tumors biologically plausible, although still insufficiently defined. Therefore, we applied a drug-target MR framework to evaluate whether genetically proxied modulation of major lipid-lowering targets influences the risks of urological and male reproductive cancers, with a specific focus on cancer risk rather than progression-related outcomes. By integrating cis-expression quantitative trait locus (cis-eQTL)-based target prioritization, lipid genome-wide association study (GWAS) screening, colocalization analyses, and external validation, we sought to identify which lipid-lowering pathways show more consistent genetic support for associations with these malignancies. In this way, we aimed to provide genetic evidence to inform long-term pharmacological safety monitoring and future risk-stratified prevention research (7,8). We present this article in accordance with the STROBE-MR reporting checklist (9) (available at https://tau.amegroups.com/article/view/10.21037/tau-2026-1-0146/rc).

Methods

Study design

This study used cis-eQTL data from the Genotype-Tissue Expression (GTEx) project together with publicly available GWAS summary statistics from European-ancestry populations to evaluate the potential causal associations between lipid-lowering drug targets and urologic and male reproductive system tumors (10). Based on the ESC/EAS and American Heart Association/American College of Cardiology guidelines, lipid-lowering drug target genes were grouped according to their principal lipid-modifying effects, distinguishing genes primarily affecting LDL-C from those mainly regulating TG metabolism (Table S1), in order to facilitate pathway-specific interpretation (11,12). Figure 1 shows the biological mechanisms and metabolic pathways of these genes, and Figure 2 outlines the study workflow.

Figure 1 Schematic diagram of lipid-lowering drug targets and related metabolic pathways. ANGPTL3, angiopoietin-like 3; APOB, apolipoprotein B; APOC3, apolipoprotein C3; HMGCR, 3-hydroxy-3-methylglutaryl-CoA reductase; IDL, intermediate-density lipoprotein; LDL, low-density lipoprotein; LDL-C, low-density lipoprotein cholesterol; LDLR, low-density lipoprotein receptor; LPL, lipoprotein lipase; PPARA, peroxisome proliferator-activated receptor alpha; PCSK9, proprotein convertase subtilisin/kexin type 9; TC, total cholesterol; TG, triglyceride; VLDL, very-low-density lipoprotein.

Figure 2 Overview of the workflow in this drug-target Mendelian randomization study. Candidate lipid-lowering drug classes were mapped to 11 target genes. GTEx v8 cis-eQTLs and UK Biobank lipid GWAS summary statistics were used for single-SNP screening, colocalization analysis, and primary drug-target MR analysis, followed by external validation using GLGC lipid summary statistics and positive-control analysis using CAD. Blue indicates risk-decreasing associations, red indicates risk-increasing associations, and asterisks indicate associations further supported in GLGC validation analyses. In the colocalization panel, H0 indicates no association; H1 indicates association with gene expression only; H2 indicates association with the lipid trait only; H3 indicates association with both traits but with distinct causal variants; and H4 indicates association with both traits with a shared causal variant. ASOs, antisense oligonucleotides; CAD, coronary artery disease; cis-eQTLs, cis-expression quantitative trait loci; EAF, effect allele frequency; FDR, false discovery rate; GLGC, Global Lipids Genetics Consortium; GTEx, Genotype-Tissue Expression; GWAS, genome-wide association study; IVW, inverse-variance weighted; LD, linkage disequilibrium; LDL-C, low-density lipoprotein cholesterol; MR, Mendelian randomization; PPH4, posterior probability of hypothesis 4; SNPs, single nucleotide polymorphisms; TG, triglyceride; UK, United Kingdom.

The study was conducted in two stages. In the first stage, target genes were prioritized by testing the associations of their top GTEx cis-eQTL single-nucleotide polymorphisms (SNPs) with LDL-C and TG in UK Biobank using single-SNP MR analysis (13,14). Significant target-lipid associations were then subjected to colocalization analysis to retain targets with stronger evidence of shared causal signals (15). In the second stage, the prioritized target genes were evaluated in drug-target MR analyses against urologic and male reproductive system tumors using UK Biobank lipid GWAS data and FinnGen cancer outcomes (16). For external validation, summary statistics for LDL-C and TG from the Global Lipids Genetics Consortium (GLGC) were used in place of the UK Biobank lipid GWAS while retaining the same FinnGen cancer outcomes. The main method used was inverse-variance weighted (IVW), supplemented by MR-Egger regression, simple mode, weighted median, and weighted mode approaches (17-20). Robustness was further assessed using Cochran’s Q test, MR-Egger intercept analysis, and leave-one-out analysis (18,21,22). Coronary artery disease (CAD) data from the CARDIoGRAMplusC4D consortium were analyzed as a positive-control outcome (23). False discovery rate (FDR) correction was used to identify significant target genes.

Data sources

This study used instrumental variable, outcome, validation, and positive control data from publicly available GWAS summary statistics, as detailed in Table S2. No individual-level data collection or processing was involved in the secondary analysis, and thus ethical approval was not required. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. GTEx was used as a cis-eQTL resource for target prioritization, whereas cancer outcomes from FinnGen were based on publicly available registry-defined cancer endpoints.

Instrumental variable data

Initially, top SNPs for 11 lipid-lowering drug target genes (PCSK9, HMGCR, NPC1L1, APOB, ABCG5, ABCG8, LDLR, LPL, ANGPTL3, APOC3, and PPARA) were selected on the basis of eQTL and GWAS data to improve target relevance and robustness. GTEx V8 cis-eQTLs were used as an initial multi-tissue target-prioritization resource and were analyzed by single-SNP MR against UK Biobank metabolic marker data for LDL-C and TG. Colocalization analysis [posterior probability of hypothesis 4 (PPH₄) >0.85] was then performed to confirm shared causal signals between target genes and metabolic markers. The use of GTEx cis-eQTLs did not imply equal pharmacological relevance across all tissues. For targets with known tissue specificity, biologically relevant or highly expressed tissues were given greater interpretive weight where appropriate, and colocalization was used to strengthen target specificity and reduce the likelihood of linkage disequilibrium-driven or non-pharmacological signals.

Outcome data

Outcome data were divided according to the two analytical stages. For the single-SNP MR analyses, outcome data consisted of UK Biobank GWAS summary statistics for LDL-C and TG. In contrast, the drug-target MR analyses used FinnGen release 9 registry-defined cancer endpoints for bladder cancer, kidney cancer (except renal pelvis), prostate cancer, renal pelvis carcinoma, and testicular cancer to evaluate the effects of target genes on urologic and male reproductive system tumors. Because these analyses were based on publicly available summary statistics, subtype-specific pathological information was not uniformly available for all cancer endpoints. Accordingly, more granular distinctions, such as invasive versus non-invasive bladder cancer, could not be consistently evaluated in the present study.

Positive control data

CAD data were used as a positive-control outcome to assess whether the direction of effect for lipid-lowering targets was broadly consistent with established cardiovascular biology.

Instrumental variable selection

Instrumental variables were identified from eQTL and GWAS datasets via a two-stage approach.

Stage one entailed the identification of SNPs related to the causal pathway between target genes and metabolic traits. Only cis-acting variants that showed a significant association with gene expression (P<0.005) were included. Only variants within ±1,000 kb of the target gene and with an F statistic >10 were included to minimize weak instrument bias. Single-SNP MR analyses were then performed to identify variants associated with the exposure (P<0.05), which were subsequently taken forward for colocalization analysis.

Stage two involved the construction of drug-target instruments. SNPs that reached genome-wide significance (P<5×10⁻⁸), were located within ±100 kb of the target gene, and had a minor allele frequency >0.01 were included. To control for linkage disequilibrium, LD clumping was applied at r²≤0.1 to retain correlated variants representing the overall genetic signal of the target. SNPs with an F statistic <10 were removed. The Steiger test was used to confirm the causal direction between exposure and outcome, and palindromic SNPs were excluded to prevent strand ambiguity.

MR analysis framework

Drug-targeted MR was applied to assess the causal link between metabolic markers (LDL-C and TG) and urological tumors, based on three assumptions: (I) the instrumental variable is significantly associated with the exposure; (II) it is independent of confounders; and (III) it affects the outcome only through the exposure (24). The analytical framework comprised two stages. First, we assessed the associations of candidate target genes with LDL-C and TG. Genes supported by both MR and colocalization analyses were selected for further evaluation. Second, these genes were examined for their potential causal associations with urologic tumors.

Colocalization analysis was performed to determine whether gene expression and the corresponding metabolic trait shared a common causal variant. For external validation, we repeated the drug-target MR analyses using GLGC lipid summary statistics in place of UK Biobank data while retaining the same tumor outcome datasets. CAD was included as a positive control outcome.

Statistical analysis

In the first stage, the Wald ratio method was used to estimate the effect of the top SNP for each target gene on LDL-C or TG levels. Instrument strength was assessed using the F statistic, and SNPs with an F statistic greater than 10 were considered sufficiently strong to reduce the risk of weak instrument bias. Colocalization analysis was then conducted to assess whether gene expression and the corresponding metabolic trait were driven by a shared causal variant, with PPH₄ >0.85 considered evidence of colocalization.

In the second stage, the causal effects of the validated target genes on urologic tumors were primarily estimated using the IVW method. MR-Egger regression and weighted median analysis were additionally applied as complementary sensitivity methods to examine the robustness of the findings. To account for multiple comparisons, FDR correction was applied (25).

Sensitivity analyses were further conducted to evaluate the robustness and validity of the MR estimates. Leave-one-out analysis was used to assess whether any single SNP disproportionately influenced the overall effect estimate. Heterogeneity among instrumental variables was examined using Cochran’s Q statistic, and potential horizontal pleiotropy was assessed using the MR-Egger intercept test. All analyses were performed in R software version 4.3.1 (The R Foundation for Statistical Computing, Vienna, Austria), mainly using the TwoSampleMR, MRInstruments, and coloc packages. The schematic pathway figure was created using FigDraw (https://www.figdraw.com/).

Results

Validation of the associations between target genes and metabolic markers

A total of 448 SNPs from the GTEx V8 eGenes dataset across 49 tissues were screened as instrumental variables for 11 lipid-lowering drug-target genes and were divided into LDL-C- and TG-regulating groups. These SNPs were tested against UK Biobank LDL-C and TG GWAS datasets via single-SNP MR to evaluate their associations with metabolic markers. The analysis identified 110 SNPs linked to LDL-C and 59 to TG regulation. These SNPs subsequently underwent colocalization analysis to assess shared causal signals between gene expression and metabolic markers.

Bayesian colocalization with UK Biobank GWAS data confirmed the shared genetic signals between SNPs and metabolic markers. Figure 3 shows that 10 target genes (but not APOC3) had strong colocalization signals (PPH₄ >0.85) in at least one tissue, suggesting SNPs may affect metabolic markers through gene expression. Tables S3,S4 provide detailed validation results for the target genes associated with the regulation of LDL-C and TG.

Figure 3 Colocalization significance of target genes for LDL-C and TG across tissues. ABCG5, ATP-binding cassette subfamily G member 5; ABCG8, ATP-binding cassette subfamily G member 8; ANGPTL3, angiopoietin-like 3; APOB, apolipoprotein B; APOC3, apolipoprotein C3; HMGCR, 3-hydroxy-3-methylglutaryl-coenzyme A reductase; LDL-C, low-density lipoprotein cholesterol; LDLR, low-density lipoprotein receptor; LPL, lipoprotein lipase; NPC1L1, Niemann-Pick C1-like 1; PCSK9, proprotein convertase subtilisin/kexin type 9; PPARA, peroxisome proliferator-activated receptor alpha; TG, triglyceride.

Drug-target MR analyses of lipid-lowering targets and urologic and male reproductive system tumors

In the second phase of this study, we evaluated the associations of 10 validated lipid-lowering drug target genes—ABCG5, ABCG8, ANGPTL3, APOB, HMGCR, LDLR, LPL, NPC1L1, PCSK9, and PPARA—on tumors of the urologic and male reproductive systems. This evaluation integrated metabolic marker data from the UK Biobank with tumor outcome data for urological cancers from FinnGen. SNPs associated with each target gene were selected as instrumental variables, as detailed in Table S5, and were applied in the drug-target MR analyses. To validate our methodology, CAD was used as a positive control, which revealed significant inverse causal associations between several of the target genes and CAD. These findings were consistent with the effects observed in clinical trials of drugs targeting these genes (summarized in Table S6), thereby supporting the validity of the instrumental variables and analytical framework.

To investigate the potential causal relationships between genes targeted by lipid-lowering drugs and tumors of the urologic and male reproductive systems, we employed several MR techniques, including IVW, MR-Egger regression, weighted median, and simple mode. Initial MR analyses identified significant associations of ANGPTL3, APOB, and HMGCR with selected tumor outcomes. Specifically, ANGPTL3 showed significant positive associations with both prostate cancer and renal cell carcinoma, while HMGCR was associated with a reduced risk of prostate cancer. Comprehensive estimates of these causal effects, along with corresponding statistical analyses, are detailed in Table S7.

Associations remaining significant after FDR correction

To account for multiple comparisons in the multigene, multi-outcome analysis, FDR correction was applied to all results, including those from the positive controls. In addition to CAD, six target genes were significantly associated with urologic and male reproductive system tumors in the MR analyses. As shown in Figure 4, the estimated causal effects of genetic inhibition of these target genes on tumor risk were as follows: inhibition of LPL was significantly associated with an increased risk of bladder cancer [odds ratio (OR) =1.46; 95% confidence interval (CI): 1.14–1.86; P-IVW=0.003; FDR =0.016]. Inhibition of PCSK9 was significantly associated with an elevated risk of bladder cancer (OR =1.76; 95% CI: 1.15–2.68; P-IVW=0.009; FDR =0.034). Inhibition of NPC1L1 was significantly associated with a reduced risk of prostate cancer (OR =0.27; 95% CI: 0.14–0.55; P-IVW<0.001; FDR =0.002). Inhibition of ANGPTL3 was significantly associated with an increased risk of prostate cancer (OR =1.37; 95% CI: 1.10–1.70; P-IVW=0.007; FDR =0.026). Inhibition of HMGCR was significantly associated with a decreased risk of prostate cancer (OR =0.72; 95% CI: 0.56–0.92; P-IVW=0.010; FDR =0.035). ANGPTL3 inhibition was also significantly associated with an increased risk of renal cell carcinoma (OR =1.97; 95% CI: 1.22–3.20; P-IVW=0.006; FDR =0.027). Inhibition of NPC1L1 was significantly associated with a reduced risk of testicular cancer (OR =0.02; 95% CI: 0.001–0.30; P-IVW=0.005; FDR =0.026). APOB inhibition was significantly associated with a decreased risk of testicular cancer (OR =0.20; 95% CI: 0.09–0.44; P-IVW<0.001; FDR <0.001) (Figure S1). These estimates should be interpreted cautiously given the limited number of testicular cancer cases.

Figure 4 Mendelian randomization forest plot of lipid-lowering gene targets and cardiovascular and urologic cancers. CI, confidence interval; IVW, inverse variance weighting; NSNP, number of single-nucleotide polymorphisms; OR, odds ratio.

Sensitivity analyses were conducted for all significant FDR-adjusted gene-outcome associations. Leave-one-out analysis forest plots (Figure S2) indicated no SNPs that could alter the causal effect estimates. Sensitivity tests showed consistent trends in the estimates, with no statistical evidence of directional horizontal pleiotropy or substantial heterogeneity (Table S8). To further assess the external consistency of the findings, independent validation was performed using LDL-C and TG data from the GLGC database. Using the same instrumental variable selection criteria as applied in the drug-target MR analyses described above, the target genes were further evaluated (Table S9). The validation results showed that target gene associations identified using GLGC data were directionally consistent with those obtained from the UK Biobank (Figure S3). Specifically, NPC1L1 inhibition was significantly associated with a reduced risk of prostate cancer (OR =0.37; 95% CI: 0.24–0.58; P-IVW<0.001; FDR <0.001), APOB inhibition was significantly associated with a reduced risk of testicular cancer (OR =0.36; 95% CI: 0.21–0.63; P-IVW<0.001; FDR =0.002), LPL inhibition was significantly associated with an increased risk of bladder cancer (OR =1.84; 95% CI: 1.42–2.39; P-IVW<0.001; FDR <0.001), and ANGPTL3 inhibition was significantly associated with an increased risk of renal cell carcinoma (OR =2.33; 95% CI: 1.27–4.29; P-IVW=0.007; FDR =0.029). These findings confirmed the consistency between the GLGC and UK Biobank validation results and further supported the robustness and external validity of the study’s conclusions.

Discussion

This MR study integrated data from GTEx, UK Biobank, GLGC, and FinnGen to systematically evaluate the potential causal associations between lipid-lowering drug targets and the risks of urologic and male reproductive system cancers. In our study, genetically proxied inhibition of HMGCR and NPC1L1 was associated with a lower risk of prostate cancer, whereas inhibition of PCSK9 was associated with a higher risk of bladder cancer. Among TG-related pathways, inhibition of ANGPTL3 was associated with a higher risk of prostate and kidney cancer, while reduced LPL activity was associated with an elevated risk of bladder cancer. In addition, inhibition of NPC1L1 and APOB showed inverse associations with testicular cancer; however, given the limited number of cases for this rare outcome, these estimates should be interpreted more cautiously. Overall, these associations were broadly supported by multiple MR methods, colocalization, and sensitivity analyses, with no clear evidence of directional horizontal pleiotropy.

Our findings are broadly consistent with and extend previous research. Observational and mechanistic studies have suggested that statins may reduce prostate cancer risk by suppressing cholesterol synthesis and disrupting lipid raft-dependent oncogenic signaling, which aligns with the protective association observed for HMGCR inhibition (26-28). Similarly, inhibition of NPC1L1, the molecular target of ezetimibe, reduces intestinal cholesterol absorption, suppresses steroid-related biosynthetic pathways, and restrains tumor growth in PTEN-deficient prostate cancer models (29-32). The inverse associations observed for NPC1L1 and APOB with testicular cancer should be interpreted cautiously. Because testicular cancer was a relatively rare outcome in the present analysis, the corresponding effect estimates were extreme and accompanied by wide CIs, indicating limited precision. Accordingly, these findings should be regarded as hypothesis-generating and require confirmation in larger datasets. Our MR results showed that genetically proxied inhibition of PCSK9 was associated with a higher risk of bladder cancer, a direction that is not fully consistent with all prior drug-target MR studies. Such discrepancies may reflect differences in instrument construction, outcome definition, and cancer-specific heterogeneity, as well as the biological distinction between lifelong target perturbation and local expression changes observed in tumor tissue. PCSK9 regulates LDLR turnover and has also been linked to inflammatory and immune-related signaling in experimental studies (33-35). In addition, transcriptomic evidence has associated PCSK9 expression with survival across several cancers, including bladder cancer (36). However, these observations relate more directly to tumor progression and the local tumor microenvironment than to cancer initiation. Therefore, the present finding should not be interpreted as definitive mechanistic evidence that reduced PCSK9 activity promotes bladder carcinogenesis. Rather, it may represent a context-dependent genetic signal, and residual horizontal or biological pleiotropy cannot be excluded. Although earlier evidence linked LDLR variation to renal cell carcinoma susceptibility (37), this association was not replicated in our analysis, possibly because of effect heterogeneity or limited statistical power.

Regarding TG-regulatory pathways, we found that inhibition of ANGPTL3 was associated with higher risks of prostate and kidney cancer, whereas reduced LPL activity was linked to a higher risk of bladder cancer. These findings are directionally compatible with studies suggesting that altered lipolysis, circulating free fatty acid availability, and fatty-acid oxidation may contribute to metabolic adaptation in aggressive tumors (6,38,39). Our analysis further suggests that when LPL-mediated TG breakdown is limited, tumor cells may shift toward other routes of lipid acquisition and fatty-acid oxidation to sustain growth (6,40). However, these associations should not be over-interpreted as fully pathway-specific direct effects. Because ANGPTL3 and LPL are pleiotropic genes involved in broader metabolic regulation, including adiposity, glucose metabolism, and inflammatory signaling, the observed associations may partly reflect correlated metabolic traits rather than isolated TG-pathway effects. In this context, multivariable MR would be valuable for further disentangling these mechanisms. By contrast, genetic proxies for PPARA or APOC3 did not exhibit meaningful associations, a result that aligns with previous evidence suggesting variable, context-dependent effects for PPARA and only sparse data connecting APOC3 to cancer biology (41-43). Overall, these observations suggest that TG-related pathways do not act uniformly across urologic tumors but instead display distinct, target-specific influences.

Taken together, the biological interpretation of our results may be framed within two partially interconnected metabolic axes. In the cholesterol-lipid raft signaling axis, inhibition of HMGCR and NPC1L1 may reduce intracellular cholesterol availability, destabilize membrane lipid rafts, and attenuate AKT/mTOR and MAPK signaling, thereby limiting proliferation in hormone-responsive tumor cells (26-31). By contrast, PCSK9 may influence tumor-related biology through a more context-dependent route that involves systemic lipid handling, LDLR turnover, and immune-metabolic signaling, rather than simply mirroring local tumor expression levels (33-36). In the lipolysis-fatty acid oxidation axis, inhibition of ANGPTL3 may enhance lipolysis, increase circulating free fatty acids, and support membrane synthesis and mitochondrial β-oxidation, whereas reduced LPL activity may trigger compensatory shifts toward alternative lipid acquisition and fatty-acid oxidation pathways (6,38-41). These interpretations remain inferential, but they provide a coherent biological framework for understanding why different lipid-lowering targets may show divergent associations across distinct urologic cancers.

This study also has several methodological strengths. By integrating GTEx-based target prioritization, UK Biobank and GLGC lipid GWAS datasets, FinnGen cancer outcomes, and colocalization analyses, we sought to improve target relevance and reduce the likelihood that the observed associations were driven solely by linkage disequilibrium. The use of a positive-control outcome further supported the validity of the analytical framework. In addition, the comparison of both LDL-C-related and TG-related pathways within the same drug-target MR design allowed a more systematic assessment of pathway-specific heterogeneity than would be possible in single-target analyses (44).

This study has several limitations. First, MR estimates represent lifelong genetic exposure and may not fully reflect the short-term pharmacologic effects of lipid-modifying therapies. Second, all participants in the primary analyses were of European ancestry, and the use of FinnGen may still have introduced population-specific bias because the Finnish population is a genetic isolate within Europe, which may limit comparability with broader European exposure or eQTL datasets. Third, FinnGen cancer outcomes were based on registry-defined endpoints from publicly available summary statistics, and subtype-specific pathological information was not uniformly available. Accordingly, more granular distinctions, such as invasive versus non-invasive bladder cancer, could not be consistently evaluated. Fourth, the limited number of cases for rare cancers, such as testicular and renal pelvic carcinoma, may have reduced statistical power and contributed to imprecise or extreme estimates. Fifth, although we did not observe any clear signal of directional pleiotropy, biological pleiotropy cannot be ruled out, especially for PCSK9, ANGPTL3, and LPL, given their broader metabolic roles. Finally, GTEx cis-eQTLs were used as an initial multi-tissue target-prioritization resource, but not all tissues are equally pharmacologically relevant, and tissue-nonspecific noise cannot be completely excluded despite the use of colocalization. Nevertheless, the consistency of the findings across datasets and analytical approaches provides some reassurance regarding the robustness of the main results.

Conclusions

This study provides genetic evidence that lipid-lowering targets may have heterogeneous associations with urologic and male reproductive system cancers. Among LDL-C-related pathways, statin-related HMGCR inhibition and NPC1L1 inhibition were associated with a lower risk of prostate cancer, while inverse associations with testicular cancer were also observed for NPC1L1 and APOB, although these rare-outcome findings should be interpreted cautiously. In contrast, genetically proxied PCSK9 inhibition was associated with a higher risk of bladder cancer. Among TG-related pathways, ANGPTL3 inhibition was associated with higher risks of prostate and renal cell carcinoma, whereas reduced LPL activity was associated with a higher risk of bladder cancer. Several associations were directionally supported using GLGC lipid data. Overall, these findings support target-specific heterogeneity in the associations between lipid-lowering pathways and tumor risk, and they warrant continued pharmacological safety monitoring as well as further mechanistic and genetic validation.

Acknowledgments

We would like to express our gratitude to the UK Biobank, the FinnGen project, and the GTEx consortium for providing summary statistics data for our MR analyses. We also extend our thanks to the researchers who contributed to these databases and to all the study participants whose involvement made this research possible. During the preparation of this manuscript, the authors used ChatGPT to improve language and readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the final published article.

Footnote

Reporting Checklist: The authors have completed the STROBE-MR reporting checklist. Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-1-0146/rc

Peer Review File: Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-1-0146/prf

Funding: The work was supported by funding from the High Level Talent Project of the Natural Science Foundation of Hainan Province (No. 824RC542).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tau.amegroups.com/article/view/10.21037/tau-2026-1-0146/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. As the analyses were based on publicly available, de-identified summary-level data, ethical approval and written informed consent were not required under local legislation and institutional guidelines.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Siegel RL, Kratzer TB, Giaquinto AN, et al. Cancer statistics, 2025. CA Cancer J Clin 2025;75:10-45. [Crossref] [PubMed]
2. Martin SS, Aday AW, Allen NB, et al. 2025 Heart Disease and Stroke Statistics: A Report of US and Global Data From the American Heart Association. Circulation 2025;151:e41-e660. [Crossref] [PubMed]
3. Liu R, Wang C, Tao Z, et al. Lipid Metabolism Reprogramming in Cancer: Insights into Tumor Cells and Immune Cells Within the Tumor Microenvironment. Biomedicines 2025;13:1895. [Crossref] [PubMed]
4. Chen R, He L, Huang Z, et al. Causal relationship between the plasma lipidome and urological cancers: A two-sample Mendelian randomization study. Medicine (Baltimore) 2025;104:e42577. [Crossref] [PubMed]
5. Liu F, Li K, Zhu Q. Targeting Metabolic Reprogramming in Bladder Cancer Immunotherapy: A Precision Medicine Approach. Biomedicines 2025;13:1145. [Crossref] [PubMed]
6. Lei Y, Cai S, Zhang JK, et al. The role and mechanism of fatty acid oxidation in cancer drug resistance. Cell Death Discov 2025;11:277. [Crossref] [PubMed]
7. Tang M, Chen L. Mendelian randomization in cancer research: opportunities and challenges. Infect Agent Cancer 2025;20:37. [Crossref] [PubMed]
8. Gill D, Georgakis MK, Walker VM, et al. Mendelian randomization for studying the effects of perturbing drug targets. Wellcome Open Res 2021;6:16. [Crossref] [PubMed]
9. Skrivankova VW, Richmond RC, Woolf BAR, et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA 2021;326:1614-21. [Crossref] [PubMed]
10. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 2020;369:1318-30.
11. Mach F, Baigent C, Catapano AL, et al. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J 2020;41:111-88. [Crossref] [PubMed]
12. Grundy SM, Stone NJ, Bailey AL, et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the Management of Blood Cholesterol: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation 2019;139:e1082-143. [Crossref] [PubMed]
13. Bycroft C, Freeman C, Petkova D, et al. The UK Biobank resource with deep phenotyping and genomic data. Nature 2018;562:203-9. [Crossref] [PubMed]
14. Woolf B, Di Cara N, Moreno-Stokoe C, et al. Investigating the transparency of reporting in two-sample summary data Mendelian randomization studies using the MR-Base platform. Int J Epidemiol 2022;51:1943-56. [Crossref] [PubMed]
15. Hormozdiari F, Gazal S, van de Geijn B, et al. Leveraging molecular quantitative trait loci to understand the genetic architecture of diseases and complex traits. Nat Genet 2018;50:1041-7. [Crossref] [PubMed]
16. Kurki MI, Karjalainen J, Palta P, et al. FinnGen provides genetic insights from a well-phenotyped isolated population. Nature 2023;613:508-18. [Crossref] [PubMed]
17. Burgess S, Butterworth A, Thompson SG. Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol 2013;37:658-65. [Crossref] [PubMed]
18. Minelli C, Del Greco M F, van der Plaat DA, et al. The use of two-sample methods for Mendelian randomization analyses on single large datasets. Int J Epidemiol 2021;50:1651-9. [Crossref] [PubMed]
19. Verbanck M, Chen CY, Neale B, et al. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet 2018;50:693-8. [Crossref] [PubMed]
20. Hartwig FP, Davey Smith G, Bowden J. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int J Epidemiol 2017;46:1985-98. [Crossref] [PubMed]
21. Burgess S, Davey Smith G, Davies NM, et al. Guidelines for performing Mendelian randomization investigations: update for summer 2023. Wellcome Open Res 2023;4:186.
22. Gao Y, Zhang J, Zhao H, et al. Instrumental Heterogeneity in Sex-Specific Two-Sample Mendelian Randomization: Empirical Results From the Relationship Between Anthropometric Traits and Breast/Prostate Cancer. Front Genet 2021;12:651332. [Crossref] [PubMed]
23. Nikpay M, Goel A, Won HH, et al. A comprehensive 1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease. Nat Genet 2015;47:1121-30. [Crossref] [PubMed]
24. Sanderson E, Glymour MM, Holmes MV, et al. Mendelian randomization. Nat Rev Methods Primers 2022;2:6. [Crossref] [PubMed]
25. Shuken SR, McNerney MW. Costs and Benefits of Popular P-Value Correction Methods in Three Models of Quantitative Omic Experiments. Anal Chem 2023;95:2732-40. [Crossref] [PubMed]
26. Roy S, Ozay ZI, Guha A, et al. Statin Use in Patients With Advanced Prostate Cancer in the TITAN and SPARTAN Trials. JAMA Netw Open 2025;8:e2527988. [Crossref] [PubMed]
27. Longo J, van Leeuwen JE, Elbaz M, et al. Statins as Anticancer Agents in the Era of Precision Medicine. Clin Cancer Res 2020;26:5791-800. [Crossref] [PubMed]
28. Mollinedo F, Gajate C. Lipid rafts as signaling hubs in cancer cell survival/death and invasion: implications in tumor progression and therapy: Thematic Review Series: Biology of Lipid Rafts. J Lipid Res 2020;61:611-35. [Crossref] [PubMed]
29. Wu Y, Song W, Su M, et al. The Role of Cholesterol Metabolism and Its Regulation in Tumor Development. Cancer Med 2025;14:e70783. [Crossref] [PubMed]
30. You X, Hu X, Sun Z, et al. Dual targeting PPARα and NPC1L1 metabolic vulnerabilities blocks tumorigenesis. Cancer Lett 2025;612:217493. [Crossref] [PubMed]
31. Allott EH, Masko EM, Freedland AR, et al. Serum cholesterol levels and tumor growth in a PTEN-null transgenic mouse model of prostate cancer. Prostate Cancer Prostatic Dis 2018;21:196-203. [Crossref] [PubMed]
32. Okita Y, Sobue T, Zha L, et al. Long-term use of anti-cholesterol drugs and cancer risks in a Japanese population. Sci Rep 2024;14:2896. [Crossref] [PubMed]
33. Hsu CY, Abdulrahim MN, Mustafa MA, et al. The multifaceted role of PCSK9 in cancer pathogenesis, tumor immunity, and immunotherapy. Med Oncol 2024;41:202. [Crossref] [PubMed]
34. Mantovani A, Marchesi F, Malesci A, et al. Tumour-associated macrophages as treatment targets in oncology. Nat Rev Clin Oncol 2017;14:399-416. [Crossref] [PubMed]
35. Hu J, Zhang M, Gui L, et al. PCSK9 Suppresses M2-Like Tumor-Associated Macrophage Polarization by Regulating the Secretion of OX40L from Hepatocellular Carcinoma Cells. Immunol Invest 2022;51:1678-93. [Crossref] [PubMed]
36. Ungvari Z, Menyhart O, Lehoczki A, et al. PCSK9 expression and cancer survival: a prognostic biomarker at the intersection of oncology and geroscience. Geroscience 2026;48:435-49. [Crossref] [PubMed]
37. Zhang GM, Wang MY, Liu YN, et al. Functional variants in the low-density lipoprotein receptor gene are associated with clear cell renal cell carcinoma susceptibility. Carcinogenesis 2017;38:1241-8. [Crossref] [PubMed]
38. Zhao T, Liang X, Chen J, et al. ANGPTL3 inhibits renal cell carcinoma metastasis by inhibiting VASP phosphorylation. Biochem Biophys Res Commun 2019;516:880-7. [Crossref] [PubMed]
39. Zhang YJ, Zhang L, Feng F, et al. ANGPTL3 Overexpression Suppresses the Development of Oncogenic Properties in Renal Cell Carcinoma via the Wnt/β-Catenin Signaling Pathway and Predicts Good Prognosis. Dis Markers 2021;2021:2863856. [Crossref] [PubMed]
40. Kuemmerle NB, Rysman E, Lombardo PS, et al. Lipoprotein lipase links dietary fat to solid tumor cell proliferation. Mol Cancer Ther 2011;10:427-36. [Crossref] [PubMed]
41. Wang Y, Lei F, Lin Y, et al. Peroxisome proliferator-activated receptors as therapeutic target for cancer. J Cell Mol Med 2024;28:e17931. [Crossref] [PubMed]
42. Schmitz J, Gouni-Berthold I. APOC-III Antisense Oligonucleotides: A New Option for the Treatment of Hypertriglyceridemia. Curr Med Chem 2018;25:1567-76. [Crossref] [PubMed]
43. Shah M, Sharma A, Ayyad M, et al. Targeting Apolipoprotein C-III for the Management of Severe Hypertriglyceridemia: Current Research and Future Directions. Cureus 2024;16:e67091. [Crossref] [PubMed]
44. Walker VM, Davey Smith G, Davies NM, et al. Mendelian randomization: a novel approach for the prediction of adverse drug events and drug repurposing opportunities. Int J Epidemiol 2017;46:2078-89. [Crossref] [PubMed]