Dichotomous outcomes of TNFR1 and TNFR2 signaling in NK cell-mediated immune responses during inflammation

Abstract

Natural killer (NK) cell function is regulated by a balance of activating and inhibitory signals. Tumor necrosis factor (TNF) is an inflammatory cytokine ubiquitous across homeostasis and disease, yet its role in regulation of NK cells remains unclear. Here, we find upregulation of the immune checkpoint protein, T cell immunoglobulin and mucin domain 3 (Tim3), is a biomarker of TNF signaling in NK cells during Salmonella Typhimurium infection. In mice with conditional deficiency of either TNF receptor 1 (TNFR1) or TNF receptor 2 (TNFR2) in NK cells, we find TNFR1 limits bacterial clearance whereas TNFR2 promotes it. Mechanistically, via single cell RNA sequencing we find that both TNFR1 and TNFR2 induce the upregulation of Tim3, while TNFR1 accelerates NK cell death but TNFR2 promotes NK cell accumulation and effector function. Our study thus highlights the complex interplay of TNF-based regulation of NK cells by the two TNF receptors during inflammation.

Fig. 1. scRNA-sequencing identifies Tim3 upregulation in effector NK cells during S. Typhimurium infection.

Wild-type C57BL/6 mice were infected with S. Typhimurium, and NK cells were analyzed by scRNA-seq or flow cytometry on day four post-infection. a UMAP plot of NK cells colored by Seurat clusters, where each dot represents a single cell. b Heatmap showing the top 20 signature genes identifying Seurat clusters. c Violin plots showing expression of indicated genes across Seurat clusters. d UMAP plots showing normalized expression of Ifng, Zbtb32, Havcr2, and Thy1. e Representative flow cytometry plots showing expression of Tim3 in splenic NK cells from naïve mice or day four post S. Typhimurium infection. f Percentage of NK cells expressing Tim3 in the spleen of naïve or S. Typhimurium infected mice. g Percentage of NK cells expressing Tim3 in the spleen of mice at different time points post S. Typhimurium infection. Data in (a–d) from a single experiment (n = 4 biological replicates) and (e, f) pooled from two independent experiments (n = 15 naïve mice and 14 S. Typhimurium infected). Data in (g) is representative of two independent experiments (n = 4-5 biological replicates). Error bars indicate mean ± SEM. Groups were compared using a two-tailed Mann-Whitney U test, where P < 0.05 was considered statistically significant. Source data are provided as a Source Data file.

Fig. 2. Overexpression of Tim3 restricts NK cell function and accumulation.

Splenocytes from Ncr1^cre (wild-type), Ncr1^creTim3^fl/fl (Tim3-null), or Ncr1^creFSF-Tim3 (Tim3-overexpression) mice were cultured for four hours with 50 ng/ml IL-15 and the indicated stimulations; plate bound anti-NK1.1, recombinant IL-12 (25 ng/ml), and/or recombinant IL-18 (5 ng/ml). a, b Percentage of NK cells expressing intracellular IFN-γ (a) or cell surface CD107a (b). NK cells isolated from Tim3-null or Tim3-overexpressing mice were adoptively transferred into Rag2^-/-γc^-/- mice and recipient mice infected with S. Typhimurium or left naïve. c Schematic showing the experimental procedure. Created in BioRender. McCulloch, T. (2023) BioRender.com/g40v791. d Relative proportions of NK cells pre-transfer and in spleens of adoptively transferred mice with or without infection. Ncr1^cre, Ncr1^creTim3^fl/fl, and Ncr1^creFSF-Tim3 mice were infected with S. Typhimurium, and spleens and livers analyzed at day 4 post-infection. e, f Total numbers of NK cells isolated from the spleen (e) or liver (f). g, h Expression of CD69 on spleen (g) or liver (h) NK cells. i, j Bacterial burden per spleen (i) or liver (j), determined by CFU counts. k Titers of IFN-γ from plasma of infected mice. Data pooled from two independent experiments (a, b: n = 6 Ncr1^cre, 7 Ncr1^creTim3^fl/fl, and 9 Ncr1^creFSF-Tim3; c, d: n = 12 naïve mice and 9 S. Typhimurium infected; e–k: n = 15 Ncr1^cre, 12 Ncr1^creTim3^fl/fl, and 9 Ncr1^creFSF-Tim3). Error bars indicate mean ± SEM. Groups were compared using the Kruskal-Wallis test with Dunn’s multiple comparisons test for (a, b) and (e–k), or the two-tailed Mann-Whitney U test for (d), where P < 0.05 was considered statistically significant. Abbreviations: IFN, interferon; CFU, colony forming units. Source data are provided as a Source Data file.

Fig. 3. TNF upregulates Tim3 and impacts NK cell function and survival in vitro.

NK cells from wild-type C57BL/6 mice were isolated and cultured for three days in the presence of the 10 ng/ml IL-15 plus the indicated cytokines. a Expression of Tim3 on NK cells. b Overall survival of NK cells, normalized to IL-15 only. Total splenocytes from C57BL/6 mice were cultured for four hours in the presence of the indicated stimulations with or without 10 ng/ml recombinant TNF. c Representative flow cytometry plots showing expression of CD107a and IFN-γ in NK cells. d, e Percentage of NK cells expressing IFN-γ (d) or CD107a (e). Data are from two independent experiments (a, b: n = 6 biological replicates, c–e: n = 9 biological replicates). Each dot represents one animal, and error bars indicate mean ± SEM. Groups were compared using one-way ANOVA with Šídák’s multiple comparisons test for (a, b) (comparisons to IL-15 only, *P < 0.05, **P < 0.01, ***P < 0.001) or paired t test for c, d where P < 0.05 was considered statistically significant. Abbreviations: TNF, tumor necrosis factor; IFN, interferon; TGF, tumor growth factor. Source data are provided as a Source Data file.

Fig. 4. Suppression of TNF signaling in NK cells impacts their phenotype during infection.

Transgenic Ncr1^cre and Ncr1^creTnfr1^fl/flTnfr2^fl/fl mice were infected with S. Typhimurium, and spleens and livers analyzed on day 4 post-infection. a Representative flow cytometry plots showing expression of Tim3 on splenic NK cells. b, c Percentage of spleen (b) or liver (c) NK cells expressing Tim3. d Total count of NK cells in the spleen of infected mice. e Percentage of NK cells expressing Ki67. f Percentage of splenic NK cells in each maturation stage based on CD11b and KLRG1 expression (Less mature = CD11b^-KLRG1^-, M1 = CD11b⁺KLRG1^-, M2 = CD11b⁺KLRG1⁺). g Percentage of NK cells expressing CD69. h, i Bacterial burden per spleen (h) or liver (i), determined by CFU counts. Data are from two independent experiments (n = 17 biological replicates). Each dot represents one animal, and error bars indicate mean ± SEM. Groups were compared using a two-tailed Mann-Whitney U test, where P < 0.05 was considered statistically significant. Abbreviations: CFU, colony forming units. Source data are provided as a Source Data file.

Fig. 5. Deletion of either TNFR1 or TNFR2 reduces NK cell terminal maturation.

Transgenic Ncr1^cre, Ncr1^creTnfr1^fl/fl, and Ncr1^creTnfr2^fl/fl mice were infected with S. Typhimurium, and NK cells analyzed by scRNA-seq on day four post-infection. a UMAP plot of NK cells colored by Seurat clusters, where each dot represents a single cell. b Dot plot showing the relative abundance of signature genes from each cluster. c UMAP plot of scRNA-seq differential abundance between WT and Ncr1^creTnfr1^fl/fl by Milo, where sampled neighborhoods are colored according to logFC. d Beeswarm plot showing the distribution of logFC between WT and Ncr1^creTnfr1^fl/fl in neighborhoods separated by Seurat clusters. e UMAP plot of scRNA-seq differential abundance between WT and Ncr1^creTnfr2^fl/fl by Milo, where sampled neighborhoods are colored according to logFC. f Beeswarm plot showing the distribution of logFC between WT and Ncr1^creTnfr2^fl/fl in neighborhoods separated by Seurat clusters. For c–f, colored neighborhoods represent statistically significant differences (FDR < 0.05 by quasi-likelihood F-test). Data from a single experiment (n = 4 Ncr1^cre, 4 Ncr1^creTnfr1^fl/fl, and 3 Ncr1^creTnfr2^fl/fl). Abbreviations: WT, wild-type; Nhood, neighborhood; logFC, log fold change.

Fig. 6. TNFR1 and TNFR2 have overlapping and opposing effects on NK cells.

a, b Volcano plots showing differentially expressed genes between Ncr1^creTnfr1^fl/fl and WT NK cells (a), or Ncr1^creTnfr2^fl/fl and WT NK cells (b) from scRNA-seq. Gray dots represent genes upregulated in WT cells, red dots represent genes upregulated in Ncr1^creTnfr1^fl/fl cells, and blue dots represent genes upregulated in Ncr1^creTnfr2^fl/fl cells (log₂FC > 0.5 and adjusted P < 0.05). c–e Violin plots showing relative enrichment of indicated Hallmark gene sets, calculated by GSEA. Transgenic Ncr1^cre, Ncr1^creTnfr1^fl/fl, and Ncr1^creTnfr2^fl/fl mice were infected with S. Typhimurium, and spleens and livers analyzed at day four post-infection. f Expression of Tim3 on splenic NK cells. g Expression of TIGIT on spleen NK cells. h, i Total NK cell counts from spleen (h) and liver (i). j Representative flow cytometry plots showing AnnexinV and Viability staining. k Percentages of apoptotic NK cells (defined as AnnexinV⁺Viability⁺ and displayed as relative to Ncr1^cre to normalize between experiments). l IFN-γ titers from the serum of infected mice at day four post-infection. m, n Bacterial burden per spleen (m) or liver (n), determined by CFU counts. Data from (a–e) are from a single experiment (n = 4 Ncr1^cre, 4 Ncr1^creTnfr1^fl/fl, and 3 Ncr1^creTnfr2^fl/fl), data from (f–n) are from two independent experiments (n = 17 Ncr1^cre, 10 Ncr1^creTnfr1^fl/fl, and 15 Ncr1^creTnfr2^fl/fl). Each dot represents one animal, and error bars indicate mean ± SEM. Groups were compared using the Wald test with Benjamini and Hochberg adjustment for (a, b), the two-sided Wilcoxon rank-sum test for (c–e), and Kruskal-Wallis test with Dunn’s multiple comparisons test for (f–n), where P < 0.05 was considered statistically significant. Abbreviations: IFN, interferon; CFU, colony forming units. Source data are provided as a Source Data file.

Fig. 7. scRNA-seq reveals Tim3 and TNF signaling in NK cells from human PBMC.

Reanalysis of scRNA-seq from PBMC pulsed with S. Typhimurium. a UMAP showing total cells from PBMC cultures. b UMAP showing HAVCR2 expression in PBMC. c Relative proportions of HAVCR2- or HAVCR2 + NK cells from naïve or S. Typhimurium pulsed PBMC. Reanalysis of scRNA-seq of PBMC from sepsis patients. d UMAP showing total cells from PBMC of sepsis patients and healthy controls, divided cell type. e NK cells divided up the Seurat cluster. f UMAP showing HAVCR2 expression in NK cells. g–j Box plots showing the percentage of NK cells expressing either HAVCR2, TIGIT, TNFRSF1A, or TNFRSF1B. Box represents the upper and lower quartile, and the median and whiskers show range between minima and maxima. Vertical dotted line separates cohorts from different hospitals. k Heatmap showing average expression in NK cells of selected genes from the indicated HALLMARK genesets. All data from published datasets (g–k: n = 19 control, 10 Leuk-UTI, 7 Int-URO, 9 URO, 7 ICU-NoSEP, 8 ICU-SEP, and 4 Bac-SEP). Groups were compared using the Kruskal-Wallis test with Dunn’s multiple comparisons test, where P < 0.05 was considered statistically significant. Abbreviations: UTI, urinary tract infection; URO, urosepsis; ICU, intensive care unit; SEP, sepsis; Bac, bacterial. Source data are provided as a Source Data file.