Genetic estimation of correlations between circulating glutamine and cancer

Abstract

The controversy regarding the causal relationship between circulating glutamine and cancer risk remains unresolved. Here, we aim to assess the causal impact of glutamine on the risk of six prevalent cancer types and their respective subtypes including breast, lung, ovarian, thyroid, prostate, and endometrial cancers. A Mendelian randomization (MR) analysis was conducted to evaluate the causal effect of circulating glutamine on cancers risk. Data on circulating glutamine were extracted from the UK Biobank (UKB), comprising 114,750 European patients. To ensure the validity of our findings, we employed several analytical approaches, such as inverse variance weighting, weighted median, weighted mode test, MR-Egger regression, and MR-PRESSO method. Both univariable and multivariable MR analyses were conducted. Additionally, we employed a large-scale summary-level study on circulating glutamine involving 24,925 European participants for validation purposes. Our MR analysis reveals a causal association between circulating glutamine and thyroid cancer in both the UKB cohort (IVW: OR = 0.667, 95% CI [0.541-0.822], P = 1.52×10^-4) and the validated cohort (IVW: OR = 0.577, 95% CI [0.421-0.790], P = 6.14×10^-4). Sensitivity analysis, including multivariable MR analyses, consistently supports this finding (P < 0.05), affirming the reliability and robustness of our study. Our findings indicate an inverse correlation between circulating glutamine and the incidence of thyroid cancer in European populations. However, further research encompassing diverse ancestries is necessary to validate this causal relationship.

Keywords: Circulating glutamine; GWAS; Mendelian randomization; UK Biobank; cancer; genetics.

Figure 1

Schematic overview of the study design. A. Basic assumptions of Mendelian randomization: Assumption 1, the genetic instruments have strong relationships with the exposure; assumption 2, the genetic instruments should not be associated with potential confounders; and assumption 3, the genetic instruments should affect the risk of outcome only through exposure and not through other alternative pathways. B. Main design of this study: Independent SNPs for glutamine were identified as instrumental variables, whereas summary statistics of gene-glutamine associations were retrieved separately from the GWAS performed by Kettunen et al. and UK Biobank. LD, linkage disequilibrium; SNP, single nucleotide polymorphism; IVW, inverse-variance weighted; PRESSO, pleiotropy residual sum and outlier; BMI, body mass index; T2DM, type 2 diabetes; LAC, lung adenocarcinoma; SCLC, squamous cell lung cancer; HGSOC, high-grade serous ovarian cancer; LGSOC, low-grade serous ovarian cancer; UVMR, Univariable Mendelian randomization; MVMR, Multivariable Mendelian randomization.

Figure 2

MR analyses from glutamine to thyroid cancer in UKB dataset. A. Forest plots: the red points showed the combined causal estimate using all SNPs together in a single instrument, using two different methods (MR-Egger and IVW). B. Scatter plots: X axes represent the genetic instruments - glutamine associations in UKB and Y axes represent genetic instruments - thyroid cancer associations. Black dots denote to the genetic instruments included in the primary MR analyses. Light blue: Inverse variance weighted; light green: Weighted Median; blue: MR Egger; green: Weighted mode.

Figure 3

MR analyses from glutamine to thyroid cancer in Kettunen et al. dataset. A. Forest plots: the red points showed the combined causal estimate using all SNPs together in a single instrument, using two different methods (MR-Egger and IVW). B. Scatter plots: X axes represent the genetic instruments - glutamine associations in Kettunen et al. and Y axes represent genetic instruments - thyroid cancer associations. Black dots denote to the genetic instruments included in the primary MR analyses. Light blue: Inverse variance weighted; light green: Weighted Median; blue: MR Egger; green: Weighted mode.

Figure 4

UVMR and MVMR MR analyses of glutamine and thyroid cancer in different datasets. A. UKB dataset (Discovery set). B. Kettunen et al. dataset (Replication set). Red plots (bars) represent OR (95% CI) of IVW for the risk of thyroid cancer associated with each 1 - SD increase of glutamine after unadjusted (UVMR) or adjusted (MVMR) for confounding factors. Blue plots (bars) represent OR (95% CI) of IVW for the risk of thyroid cancer associated with each 1 - SD increase of risk factors after adjusted (MVMR) for glutamine. MVMR, multivariable Mendelian randomization; UVMR, univariable Mendelian randomization; BMI, body mass index; T2DM, type 2 diabetes.