Transcriptional signature of human pro-inflammatory TH17 cells identifies reduced IL10 gene expression in multiple sclerosis

Abstract

We have previously reported the molecular signature of murine pathogenic T_H17 cells that induce experimental autoimmune encephalomyelitis (EAE) in animals. Here we show that human peripheral blood IFN-γ⁺IL-17⁺ (T_H1/17) and IFN-γ^-IL-17⁺ (T_H17) CD4⁺ T cells display distinct transcriptional profiles in high-throughput transcription analyses. Compared to T_H17 cells, T_H1/17 cells have gene signatures with marked similarity to mouse pathogenic T_H17 cells. Assessing 15 representative signature genes in patients with multiple sclerosis, we find that T_H1/17 cells have elevated expression of CXCR3 and reduced expression of IFNG, CCL3, CLL4, GZMB, and IL10 compared to healthy controls. Moreover, higher expression of IL10 in T_H17 cells is found in clinically stable vs. active patients. Our results define the molecular signature of human pro-inflammatory T_H17 cells, which can be used to both identify pathogenic T_H17 cells and to measure the effect of treatment on T_H17 cells in human autoimmune diseases.

Fig. 1

Transcriptionally distinct human T_H17 subsets in peripheral blood. a IFN-γ and IL-10 expression in human T_H17 cells. Isolated PBMCs were stimulated with PMA and ionomycin for 4 h. Production of indicated cytokines in CD4⁺ T cells were assessed by flow cytometry with intracellular cytokine staining assay. Dotplots shown were gated on CD4⁺ lymphocytes. Data are representative of two independent experiments with similar results. b Isolation of live T_H1/17, T_H17, T_H1, and DN cells from human PBMC for nCounter analysis. CD4⁺ T cells isolated from the peripheral blood of healthy donors were stimulated with PMA and ionomycin for 3 h. CD3⁺CD4⁺-T_H1/17 (IFN-γ⁺IL-17⁺), T_H17 (IFN-γ⁻IL-17⁺), T_H1 (IFN-γ⁺IL-17⁻), and DN (IFN-γ⁻IL-17⁻) cells were sorted after being stained with fluorescence-conjugated anti-CD3 and CD4 in combination with cytokine secretion detection kits (Miltenyi) (n = 5). b Isolated CD4⁺ T subsets were stimulated and stained as in a. c–f CD4⁺ T-cell subsets treated as in a were measured using the nCounter (nanoString Technologies) CodeSet HuT_H17 and subsequently analyzed (hereafter abbreviated as nCounter analysis). c Differential expression analysis of mRNA levels of IL17A and IFNG. *p < 0.05, **p < 0.005, ***p < 0.0005, One-way ANOVA with Tukey’s multiple comparison test (mean ± s.d.). d Differential expression analysis of mRNA levels of IL10. Two tailed, paired Student’s t test p-value was shown (mean ± s.d.). The 326 out of the 418 measured genes in the HuT_H17 CodeSet that showed unsupervised variation across the sample population were used for e hierarchical clustering of the individual samples (individual donors: A, B, C, D, and E), f hierarchical clustering of the Pearson’s linear correlations between the samples (individuals n = 5; individual donors: A, B, C, D, and E), and g principal component analysis of the samples (individuals n = 5)

Fig. 2

Gene expression comparison between human T_H1/17 vs. T_H17 cells and mouse pathogenic vs. non-pathogenic T_H17 cells. a The expression of previously reported murine and human T_H17 signature genes in purified ex vivo T_H1 cells. The mRNA gene expression levels in T_H1/17, T_H17, T_H1, and DN cells were measured as described in Fig. 1. *p < 0.05, repeated measures one-way ANOVA; **p < 0.05, pairwised group comparison with Tukey’s multiple comparison test (mean ± s.d., n = 5). b, c Gene set enrichment analysis comparing human T_H1/17 vs. T_H17 cells with mouse pathogenic vs. non-pathogenic T_H17 cells. b Heatmap of upregulated (upper panels) and downregulated (lower panels) “leading edge” genes of comparison Scenario 1: human T_H1/17 vs. T_H17 cells vs. mouse TGF-β3 plus IL-6-induced T_H17 cells vs. TGF-β1 plus IL-6-induced T_H17 cells (Kolmogorov–Smirnov test p < 0.0001; FDR q < 0.0001 for upregulated genes; Kolmogorov–Smirnov test p = 0.0007; FDR q = 0.001 for downregulated genes). c Heatmap of upregulated (upper panels) and downregulated (lower panels) “leading edge” genes of comparison Scenario 2: human T_H1/17 vs. T_H17 cells vs. mouse IL-1, IL-23 plus IL-6-induced T_H17 cells vs. TGF-β1 plus IL-6-induced T_H17 cells (Kolmogorov–Smirnov test p < 0.0001; FDR q < 0.0001 for upregulated genes; Kolmogorov–Smirnov test p = 0.004; FDR q = 0.004 for downregulated genes). Each column represents one donor in human (n = 5) or one sample in murine (n = 4). d The robust predicted pathogenic signature (PreP-Signature) of human T_H1/17 cells. Signature genes are those identified as differentially expressed between human T_H1/17 and T_H17 cells that are identified as enriched “leading edge” genes when assessing these human genes in the mouse profiles in b, c and that are additionally curated for robustness based on supervised absolute fold change >1.5. Two tailed, paired Student’s t test p-value < 0.05. *p < 0.05, **p < 0.05, ***p < 0.0005, n = 5

Fig. 3

Gene expression comparison between human T_H1/17 vs. T_H17 cells and IL-10^– vs. IL-10⁺ T_H17 clones. a Quantitative RT-PCR analysis of gene expression in human IL-10^– and IL-10⁺ T_H17 clones isolated from healthy donors (two tailed, paired Student’s t test, mean ± s.d., n = 3). For IFNG, IL17A, and IL23R, resting T_H17 clones were stimulated with anti-CD3 and anti-CD28 for 4 h before RNA extraction. For IL10, resting T_H17 clones were stimulated with anti-CD3 and anti-CD28 for 5 days, then cells were re-stimulated with anti-CD3 and anti-CD28 for 4 h before RNA extraction. b Hypergeometric enrichment test between human T_H1/17 vs. T_H17 cells and IL-10^– vs. IL-10⁺ T_H17 clones. Genes differentially expressed between human T_H1/17 and T_H17 cells (Supplementary Data 3) were analyzed for enrichment in those of human IL-10^– vs. IL-10⁺ T_H17 clones (Supplementary Data 5). Heatmap shows the overlapping genes (one-sided Fisher’s exact test p < 0.0001, FDR q = 0.0001). Each column represents one donor (n = 5). c Predicted upstream regulators for T_H1/17 and IL-10^– T_H17 clone differentiation. The differentially expressed genes with corresponding fold changes and p-values from the T_H1/17 vs. T_H17 comparison (Supplementary Data 3) and IL-10^– vs. IL-10⁺ T_H17 clone comparison (Supplementary Data 5) were analyzed using the IPA upstream regulator analysis. Ex vivo, T_H1/17 vs. T_H17 comparison; Clone, IL-10^– vs. IL-10⁺ T_H17 clone comparison. d, e Gene set enrichment analysis comparing human IL-10^– vs. IL-10⁺ T_H17 clones with mouse pathogenic vs. non-pathogenic T_H17 cells. d Heatmap of upregulated (upper panels) and downregulated (lower panels) “leading edge” genes of comparison Scenario 1: human IL-10^– vs. IL-10⁺ T_H17 clones vs. mouse TGF-β3 plus IL-6-induced T_H17 cells vs. TGF-β1 plus IL-6-induced T_H17 cells (Kolmogorov–Smirnov test p = 0.004; FDR q = 0.005 for upregulated genes; Kolmogorov–Smirnov test p = 0.004; FDR q = 0.003 for downregulated genes). e Heatmap of upregulated (upper panels) and downregulated (lower panels) ‘‘leading edge” genes of comparison Scenario 2: human IL-10^– vs. IL-10⁺ T_H17 clones vs. mouse IL-1, IL-23 plus IL-6-induced T_H17 cells vs. TGF-β1 plus IL-6-induced T_H17 cells (Kolmogorov–Smirnov test p = 0.005; FDR q = 0.010 for upregulated genes; Kolmogorov–Smirnov test p = 0.016; FDR q = 0.009 for downregulated genes). Each column represents one donor in human (n = 5) or one sample in murine (n = 4)

Fig. 4

Predicting STAT3 as upstream transcription factor by the PreP-Signatures of T_H17 cells. a Venn diagram representations of signature genes upregulated (left) and downregulated (right) in human T_H1/17 cells and IL-10^– T_H17 clones. Complete PreP-Signatures from GSEA comparison Scenarios I and II were merged for ex vivo cells (Fig. 2b, c) and T_H17 clones (Fig. 3d, e), respectively. Genes with supervised absolute fold change >1.5 in either ex vivo cells or T_H17 clones were shown in italic bold letters. Ex vivo, differentially expressed “leading edge” genes from T_H1/17 vs. T_H17 GSEA comparisons presented in green circles; Clone, differentially expressed “leading edge” genes from IL-10^– vs. IL-10⁺ T_H17 clone GSEA comparisons presented in yellow circles. b Predicted upstream transcription factors for T_H1/17 and IL-10^– T_H17 clone differentiation. The molecular signatures of T_H1/17 cells and IL-10^– T_H17 clones in a were analyzed using the Enrichr ChEA2016 analysis and the predicted transcription factors with Benjamini–Hochberg adjusted p-value <0.05 were shown. TF transcription factor

Fig. 5

Frequency of T_H17 subsets and differential expression of PreP-Signature genes in T_H1/17 vs. T_H17 cells in MS. Peripheral CD4⁺ T cells isolated from the PBMC of untreated RRMS patients (n = 19) and age- and sex-matched healthy controls (HC) (n = 16) were stimulated, stained, and sorted for T_H1/17, T_H17, T_H1, and DN cells as described in Fig. 2b. RNA isolated from sorted cell subsets was subjected to low-input qPCR analysis. a Frequencies of T_H1/17, T_H17, and total T_H17 cells in total peripheral CD4⁺ T cells (Welch’s t test, p-values, mean ± s.d.). b Frequency of T_H1/17 in total T_H17 cells (Welch’s t test, mean ± s.d.). c, d qPCR analysis of gene expression in isolated CD4⁺ T-cell subsets. c Comparison of IL17A expression between HC and patients in T_H1/17 or T_H17 cells (Welch’s t test, mean ± s.d.). d Differential expression of PreP-Signature genes between T_H1/17 vs. T_H17 cells in HC and MS patients. Two tailed, paired Student’s t test, *p < 0.05, **p < 0.001, ***p < 0.0001

Fig. 6

Expression of PreP-signature genes of human pro-inflammatory T_H17 cells in RRMS. RNA isolated from T_H1/17, T_H17, and DN cells (Fig. 5) was subjected to qPCR analysis. Comparison of PreP-Signature gene expression between HC and patients with RRMS in a T_H1/17 cells and b T_H17 cells. (Welch’s t test, mean ± s.d.). Comparison of c TBX21 and d IL10 expression between HC and patients with RRMS in T_H1/17 and T_H17 cells. e Comparison of STAT3 (upper panel) and STAT5A (lower panel) expression between HC and patients with RRMS in T_H1/17, T_H17, and DN cells. Welch’s t test p-values were shown (mean ± s.d.)

Fig. 7

Correlation of IL10 and STAT3 expression with disease activity in MS. The mRNA levels of signature genes with altered expression in MS (Fig. 6) were compared between active and stable patients. Signature gene expression in a T_H1/17 (active, n = 8; stable, n = 7), b T_H17 (active, n = 12; stable, n = 6), and c DN cells (active, n = 12; stable, n = 7) (Welch’s t test, mean ± s.d.)