Role of Serum Metabolites in Mediating the Effect of Circulating Inflammatory Proteins on NK/T Cell Lymphoma

Introduction:

NK/T Cell Lymphomas (NKTCL) are highly aggressive typesof Non-Hodgkin’s Lymphoma (NHL) that have a predilection for Asian and South Americanpopulations. Identifying new drug targets or prognostic markers is essential for developingfuture strategies for NKTCL treatment.

materials and methods:

Data sourcesData on 91 circulating inflammatory proteins were obtained from a proteomic study that included 14,824 participants(13). Comprehensive per-protein GWAS summary statistics are available for download at Cambridge University’s website and the EBI GWAS Catalog (accession numbers GCST90274758 to GCST90274848). Genetic associations with NKTCL were sourced from the FinnGen study (finngen_R10_CD2_TNK_LYMPHOMA), which included 363 cases and 324,650 controls (https://www.finngen.fi/en). Information on 1,091 blood metabolites and 309 metabolite ratios was obtained from the GWAS Catalog (accession numbers Ebi-a-GCST90199621 to Ebi-a-GCST90201020)(14). Importantly, there is no overlap in population samples between the exposure and outcome groups. All GWAS summary statistics used in this study are based on European populations.Study designThis study employed a two-step, bidirectional Mendelian randomization (MR) approach using an inverse variance weighted (IVW) method to explore the relationship between CIF and the genetically predicted risk of NKTCL, as well as the potential mediating role of serum metabolites in this association. Initially, the study assessed the causal effect of CIF on NKTCL using two-sample, bidirectional MR to identify CIF variants strongly associated with NKTCL risk. Next, the causal effect of serum metabolites on NKTCL was evaluated using a two-sample MR. Finally, the study examined the causal effect of the identified CIF on selected serum metabolite signatures and calculated the mediation proportion for each metabolite in CIF’s effect on NKTCL. The study design is illustrated in Fig. 1A-B.Genetic instrumental variables (IVs) selectionThe selection criteria for SNPs as instrumental variables (IVs) for CIF and serum metabolites were set to p < 1e>10 indicating a robust association between IVs and exposure factors(15). Additionally, PhenScanner V2 was utilized to filter out potential confounding SNPs(16).Statistics analysisStatistical and sensitivity analyses were performed using R software (version 4.4.0) along with several R packages, including “Two Sample MR,” “ieugwasr,” “MR-PRESSO,” and “MR Instruments.” The primary MR analysis used to estimate causal effects was the random-effects inverse variance weighted (IVW) method. To assess potential pleiotropic effects, the weighted-median (WM) and MR-Egger regression methods were applied. A p-value of less than 0.05 was considered indicative of a significant association between exposure and outcome. Leave-one-out analyses were also conducted by sequentially removing each instrument to identify influential instruments and evaluate their effect on the overall estimates. The indirect effect of CIF on NKTCL risk through serum metabolites was assessed using the “product of coefficients” method, and the standard errors for these indirect effects were calculated using the delta method.

Methods:

This study employed a two-step and bidirectional Mendelian Randomization(MR) study with an Inverse Variance Weighted (IVW) approach to investigate the potentialassociation of Circulating Inflammatory Proteins (CIPs) with NKTCL and the mediationeffect of serum metabolites in the relationship between CIPs and NKTCL. Data on91 CIPs from a proteomic study of 14,824 individuals conducted with the Olink Targetplatform were collected. Genetic information for NKTCL was obtained from the finngen_R10_CD2_TNK_LYMPHOMA, including 363 cases and 324,650 controls. Datafor 1,091 plasma metabolites and 309 metabolite ratios (Ebi-a- GCST90199621 to Ebia-GCST90201020) were derived from the GWAS Catalog.

Results:

MR analyses identified 4 plasma proteins significantly associated with NKTCL.Increased levels of TGF-α and TNFRSF9 were associated with increased NKTCLrisk, with Odds Ratios (OR) of 1.6388 (95% CI 1.0158−2.6440) and 1.6863 (95% CI1.1757−2.4187), respectively. Conversely, LIF-R (OR 0.6265, 95% CI 0.4427−0.8866)and SULT1A1 (OR 0.5425, 95% CI 0.3758−0.7833) exhibited protective effects. Notably,none of these proteins showed evidence of reverse causality. Furthermore, the mediationMR suggested that the effect of TNFRSF9 on NKTCL (total effect IVW: OR =1.6863, 95% CI 1.1757−2.4187, P = 0.010) was partially mediated by Isobutyrylcarnitine(C4) levels (0.739, 95% CI 0.609−0.895, P = 0.002), accounting for 5.77% of the totaleffect.

Discussion:

This MR study suggests that CIPs may be associated with NKTCL pathogenesisand proposes that metabolite-mediated pathways represent a potential novel mechanisticlink. It provides preliminary evidence consistent with serum metabolites playinga role as mediators in this relationship. These findings highlight the therapeutic potentialof targeting specific metabolic pathways in NKTCL.

Conclusion:

Our proteome-wide MR analysis suggests an association between CIPs andNKTCL, which may be mediated by serum metabolites.