Myeloid cell genome-wide screen identifies variants associated with Mycobacterium tuberculosis–induced cytokine transcriptional responses
Joshua J. Ivie, … , Sarah J. Dunstan, Thomas R. Hawn
Joshua J. Ivie, … , Sarah J. Dunstan, Thomas R. Hawn
Published May 22, 2025
Citation Information: J Clin Invest. 2025;135(14):e179822. https://doi.org/10.1172/JCI179822.
View: Text | PDF
Research Article Genetics Infectious disease

Myeloid cell genome-wide screen identifies variants associated with Mycobacterium tuberculosis–induced cytokine transcriptional responses

Abstract

Immune and clinical outcomes to Mycobacterium tuberculosis (Mtb) infection vary greatly between individuals, yet the underlying genetic and cellular mechanisms driving this heterogeneity remain poorly understood. We performed a cellular genome-wide association study to identify genetic variants associated with Mtb-induced monocyte transcriptional expression of IL1B, IL6, TNF, and IFNB1 via RNA-Seq in a Ugandan cohort. Significantly associated variants were assessed for transferability in an independent Seattle cohort, further validated in vitro, and assessed for clinical phenotype associations. We identified 77 loci suggestively associated with Mtb-induced cytokine expression in monocytes in Uganda. SNPs associated with Mtb-induced TNF were enriched within α-linolenic acid metabolism pathway genes, which was validated in vitro using PLA2 inhibitors. Four loci maintained significant associations in Seattle. We validated a cytokine effect with siRNA knockdown for two of these loci, which mapped to the genes SLIT3 and SLC1A1. Furthermore, exogenous treatment of macrophages with SLIT3 enhanced Mtb intracellular replication. Finally, SLC1A1 and SLIT3 variants were associated with susceptibility to tuberculous meningitis and subsequent survival, respectively, in a Vietnamese cohort. In summary, we identified multiple variants and pathways associated with Mtb-induced cytokine transcriptional responses that were validated in vitro and were associated with clinical tuberculosis susceptibility.

Authors

Joshua J. Ivie, Kimberly A. Dill-McFarland, Jason D. Simmons, Glenna J. Peterson, Penelope H. Benchek, Harriet Mayanja-Kizza, Lily E. Veith, Moeko Agata, Dang T.M. Ha, Ho D.T. Nghia, W. Henry Boom, Catherine M. Stein, Chiea C. Khor, Guy E. Thwaites, Hoang T. Hai, Nguyen T.T. Thuong, Xuling Chang, Sarah J. Dunstan, Thomas R. Hawn

×

Figure 2

MAGMA enrichment identifies Mtb-induced TNF association with α-linolenic acid metabolism.

Options: View larger image (or click on image) Download as PowerPoint
MAGMA enrichment identifies Mtb-induced TNF association with α-linolenic...
(A) Previously calculated SNP P values for each cytokine were used to calculate gene-level P values, mapped to within each gene, accounting for SNP LD using MAGMA multi-model. Selected C2-canonical pathways (KEGG, Reactome, BioCarta, and Pathway Interaction Database) and C5-GO terms were assessed. Significant results after Bonferroni multiple correction are shown. Dashed vertical line indicates threshold for significance (adjusted P value < 0.05). (B) Network of genes within α-linolenic acid metabolism gene set was annotated for interaction using STRINGdb and plotted for MAGMA-calculated gene P value significance showing multiple PLA2 isoforms with increased significance. (C and D) TBWCL (C) and Mtb-induced secretion (D) of TNF, IL1B, and IL6 assessed by ELISA after increasing doses of cPLA2 inhibitor AACOCF3 pretreatment in human macrophages show that PLA2 enzymes do significantly affect TBWCL-induced TNF response and Mtb-induced IL1B and IL6 response. Results shown are from 3 biological replicates in 3 human donors per treatment. Significance was assessed using linear mixed model adjusting for random effect of donor followed by ANOVA and pairwise comparisons. Individual data points were plotted according to model calculated values adjusted for donor intercept. ***P < 0.001, ****P < 0.0001. Individual donor values are presented in Supplemental Figure 6A.