Proteome-wide Mendelian randomization reveals causal associations between plasma proteins and autoimmune thyroid disease

Abstract

Autoimmune thyroid diseases (AITD) are the most common autoimmune disorders. Identifying new biomarkers and therapeutic targets in plasma proteins is crucial. We conducted a proteome-wide Mendelian randomization (MR) and colocalization analysis to determine plasma proteins causally associated with AITD. Proteome-wide summary-level genome-wide association studies (GWAS) were collected from the UK Biobank Pharma Proteomics Project (UKB-PPP) and deCODE genetics, encompassing 2922 and 4719 plasma proteins, respectively. Genetic associations with AITD were derived from an AITD GWAS meta-analysis study (30,234 cases and 725,172 controls) and the FinnGen database (40,926 cases and 274,069 controls). MR analysis, including summary-data-based Mendelian randomization (SMR), Wald Ratio, and IVW methods, was employed to estimate the causal effects between plasma proteins and AITD. Colocalization analysis was used to assess whether identified proteins and AITD shared the common causal variants. Genetically predicted levels of 11 plasma proteins were found to have a causal association with AITD. Colocalization analysis revealed that five of these proteins had evidence of colocalization, including leukemia inhibitory factor (LIF), interleukin-7 receptor subunit alpha (IL7RA), CD226, tumor necrosis factor ligand superfamily member 11 (TNF11), and transcription factor junD (JUND). Genetically predicted levels of LIF and IL7RA were associated with an increased risk of AITD, whereas CD226, TNF11, and JUND were inversely related to AITD risk. This study has identified multiple candidate plasma proteins causally associated with AITD. Among these proteins, LIF, IL7RA, CD226, TNF11, and JUND are considered to have potential as disease biomarkers and therapeutic targets, but further clinical and experimental validation is still necessary in the future.

Keywords: Autoimmune thyroid diseases; Causal association; Colocalization analysis; Mendelian randomization; Plasma protein.

Fig. 1

Study design overview for identification of plasma proteins causally associated with AITD.

Fig. 2

Causal associations between plasma proteins and AITD identified by SMR. (A,B) Volcano plots displayed the causal plasma proteins identified in the discovery and replication cohort of AITD, using data from UKB-PPP. (C,D) Volcano plots displayed the causal plasma proteins identified in the discovery and replication cohort of AITD, using data from deCODE genetics. (E,F) Forest plots showed the odds ratio, 95% confidence interval and p-value of the causal plasma proteins for UKB-PPP and deCODE genetics, respectively.

Fig. 3

Causal associations between plasma proteins and AITD identified by IVW or Wald ratio. (A,B) Volcano plots displayed the causal plasma proteins identified in the discovery and replication cohort of AITD, using data from UKB-PPP. (C,D) Volcano plots displayed the causal plasma proteins identified in the discovery and replication cohort of AITD, using data from deCODE genetics. (E,F) Forest plots showed the odds ratio, 95% confidence interval and p-value of the causal plasma proteins for UKB-PPP and deCODE genetics, respectively.

Fig. 4

Colocalization analysis between the causal plasma proteins and AITD in the discovery and replication cohorts. (A) Utilizing data from UKB-PPP, high support evidence for colocalization were determined in both AITD cohorts. (B) Utilizing data from deCODE genetics, high support evidence for colocalization were determined in both AITD cohorts.

Fig. 5

Functional enrichment analysis of the AITD-associated plasma proteins. (A) The GO analysis showed the top enriched functions of biological process, cellular components and molecular functions. (B) The top 15 KEGG pathways related to the AITD-associated plasma proteins.