Mendelian randomization analysis identified serum palmitoylcarnitine as a biomarker for hepatocellular carcinoma

Abstract

Background and aims With the extensive application of metabolomics in hepatocellular carcinoma(HCC),more studies have found that serum metabolites are closely related to the occurrence and development of HCC. However, the causal relationship between them remains unclear. We will use the two-sample Mendelian randomization analysis to explore the causal relationship between 1400 different sources of serum metabolites and HCC at the genetic level in this study,aiming at providing valuable reference for the pathogenesis, diagnosis and treatment of HCC from the metabolic pathway. Methods Two-sample Mendelian randomization analysis was performed to estimate the causal relationship between genetically predicted serum metabolites and HCC.A genome-wide association study (GWAS) of 1400 serum metabolites were used as exposure and HCC as outcome. Both exposure and outcome datasets are available from the publicly published GWAS catalog. The inverse variance weighted method(IVW) was used as the main causality analysis method, and Cochran's Q, MR-Egger intercept, MR-PRESSO and other methods were used to carry out sensitivity analysis of heterogeneity and pluripotency, so as to ensure the accuracy and reliability of the results. In addition, the Bonferroni correction method was used for multiple correction of P-values. Finally, MetaboAnalyst 5.0 software was used for metabolic pathway analysis of significant metabolites. Results IVW results showed that 20 metabolites and 5 metabolite ratios were positively correlated with HCC, which may be risk factors for HCC, and the OR ranged from 1.450-4.036. Among them,palmitoylcarnitine (OR 4.036,95%CI 2.160-7.543,p=1.22×10^-5) was the metabolite with the highest OR and the lowest P_IVW.The other seven metabolites with P_IVW less than 0.010 were Ornithine levels (OR 3.368,95%CI 1.620-7.003,p=0.001),N-acetylaspartate (naa) levels (OR 4.030,95%CI 1.707-9.514,p=0.001),4-ethylcatechol sulfate levels (OR 3.340,95%CI 1.561-7.147,p=0.002),Eicosapentaenoate (EPA;20:5n3) levels (OR 2.713,95%CI 1.431-5.140,p=0.002),Tyramine O-sulfate levels (OR 2.893,95%CI) 1.418-5.901,p=0.003),3-(3-amino-3-carboxypropyl)uridine levels (OR 3.517,95%CI 1.462-8.461,p=0.005),2-naphthol sulfate levels (OR 2.617,95%CI 1.267-5.406,p=0.010). 21 metabolites and 6 metabolite ratios were associated with reduced risk of HCC. OR ranged from 0.321-0.609, The most significant metabolites were lipid metabolites 1-(1-enyl-Palmitoyl) -2-Linoleoyl-GPE (P-16:0/18:2) levels (OR 0.348,95%CI 0.180-0.676,p=0.002).The other six metabolites with P_IVW less than 0.010 were 4-allylphenol sulfate levels (OR 0.346,95%CI 0.175-0.685,p=0.002),Tyrosine to pyruvate ratio (OR 0.336,95%CI 0.165-0.684,p=0.003),1-stearoyl-2-linoleoyl-gpc (18:0/18:2) levels (OR 0.381,95%CI 0.199-0.729,p=0.004),Ascorbic acid 3-sulfate levels (OR 0.345,95%CI 0.158-0.753,p=0.008),Linoleoyl ethanolamide levels (OR 0.411,95%CI 0.214-0.790,p=0.008). However, the Bonferroni correction method found that only palmitoylcarnitine levels (p=1.22×10-5<3.57×10-5) passed the multiple tests. Cochran's Q test showed no heterogeneity (all p>0.05). Although a few metabolites have pleiotropy, no outliers were found in further MR-PRESSO detection, indicating that they were unlikely to be affected by horizontal pleiotropy. The leave-one out test did not find a single SNP to have a significant effect on the overall results. The results of sensitivity analysis proved that results of Mendelian randomization analysis are robust. Conclusions Our findings revealed that elevated levels of 20 serum metabolites and 5 metabolite ratios such as palmitoylcarnitine could increase the risk of HCC. However, the increased levels of 21 metabolites including 1-(1-enyl-Palmitoyl) -2-Linoleoyl-GPE (p-16:0/18:2) and 6 metabolite ratios could reduce the risk of HCC. After multiple correction, only palmitoylcarnitine levels had a causal relationship with HCC, and palmitoylcarnitine levels could be considered as a strong and reliable risk factor for HCC. These findings contributed to a deeper understanding of the pathogenesis of HCC at the metabolic level and provided evidence to support multiple serum metabolites such as palmitoylcarnitine as potential biomarkers for subsequent HCC diagnostic studies. In addition, further exploration of related metabolic pathways of risk factor metabolites may provide a valuable reference for drug target therapy of HCC. However, more research is needed to confirm this in the future.