Autoreactive T-cells expand upon CD1c stimulation.

(A) Representative histogram overlays showing CD1c, MHCI, MHCII, and β2m expression on wild type THP1 cells, and engineered THP1-KO and THP1-CD1c APCs. (B) Representative flow cytometry analysis of T-cells from a healthy donor expanded with THP1-CD1c cells in the absence of exogenous lipid antigen and then cultured overnight with THP1 cells. Cells were gated on live CD3+CTV- T-cells and T-cell activation was measured with anti-CD69, anti-CD25 and anti-CD137. Plots show expression of CD69 and CD137 (top), CD25 and CD137 (bottom). Significant numbers of CD1c-autoreactive T-cells were present in expanded cultures. (C) Cumulative data from 14 healthy donors showing frequency of CD3+CTV-CD69+CD137+ T-cells (top), and CD3+CTV-CD25+CD137+ T-cells (bottom) from lines first expanded with THP1-CD1c and then overnight culture with THP1-KO or THP1-CD1c APCs. CD1c-autoreactive T-cells were present in the majority of donors. ** P < 0.01; *** P < 0.001; (C, Wilcoxon matched pairs signed rank test).

CD1c is expressed in human TB and downregulated in APCs by Mtb infection.

(A) Lung biopsies from patients with active TB were stained with anti-CD1c antibodies. i-iii) Anti-CD1c immunoreactivity in a representative TB granuloma. iv) Control staining of the same granuloma with secondary antibody only and avidin biotin–peroxidase complex (ABC) detection shows no immunoreactivity. v) Extensive staining in inflammatory tissue remote from the TB granuloma. vi) CD1c staining in B-cell follicles adjacent to the granuloma. (B) RNA-Seq heat map showing significant reduction in CD1A, CD1B, and CD1C expression in MoDCs from five donors at 48 hours after Mtb infection (MOI=1). Differential gene expression was performed on filtered normalised counts using the voom – limma pipeline in R Studio with p-value adjustment being performed using the Benjamini-Hochberg method. Differentially expressed genes were identified as having Log2FC > +1 (Upregulated) or < −1 (Downregulated) with adjusted p-values <0.05. (C) MoDC histograms of normalised counts demonstrating a reduction in CD1a, CD1b, and CD1c expression on differentiated MoDCs following live Mtb infection (MOI=1). Data representative of an experiment conducted in two donors performed in triplicate. (D) Flow cytometry histograms showing CD1c expression on THP1-CD1c cells at 72 hours following infection with live Mtb (MOI=1). Mtb infection does not change CD1c expression on THP1-CD1c cells. Data representative of two experiments performed in triplicate.

CD1c-autoreactive T-cells are cytotoxic to Mtb stimulated cells.

(A-D) CD1c-endo tetramer staining of T-cell lines generated from two healthy donors. (A) T-cells generated from PBMCs after CD1c-endo streptamer enrichment and one round of CD1c-endo dextramer flow cytometry sorting and subsequent in vitro expansion (Fig. S3). (C) T-cells generated after expansion with THP1-CD1c cells, followed by CD1c-endo-tetramer guided cell sorting and subsequent in vitro expansion. (B) and (D) FACS plots demonstrating that T-cells in (A) and (C), respectively, are αβ TCR and CD4 positive. (E) T-cell lines are activated in response to THP1-CD1c APCs. Rested T-cells and T-cells cultured with THP1-KO APCs served as control. T-cell activity was determined by measuring the upregulation of CD69 and CD25 by flow cytometry. Data is representative of three experiments from the two donor-derived lines performed in triplicate. (F) CD1c-autoreactive T-cells display significant cytotoxicity in a Mtb dose dependent manner against UV-killed Mtb-treated THP1-CD1c APCs. No cytotoxicity was observed by T-cells cultured with untreated THP1-CD1c APCs or irrelevant control T-cells, lacking CD1c restriction. (G) CD1c-autoreactive T-cells displayed significant cytotoxicity in a T-cell dose-dependent manner against UV-killed Mtb-treated THP1-CD1c APCs. No cytotoxicity was observed by T-cells cultured with untreated THP1-CD1c APCs or with irrelevant control T-cells. Cytotoxicity was measured using a ToxiLight assay. Data is representative of two independent experiments, each preformed in triplicate. * P < 0.05; ** P < 0.01; **** P < 0.0001 (E, one-way ANOVA with Tukey’s multiple comparison test; F-G, two-way ANOVA).

CD1c-autoreactive T-cells lyse target cells infected with live Mtb.

(A) Flow cytometry dot plots and (B) bar graphs showing CD1c-autoreactive T-cells are activated by THP1-CD1c APCs but not THP1-KO APCs, with significantly greater activation when THP1-CD1c APCs are infected with live Mtb (MOI=1). (C) Gating strategy for measuring the T-cell lysis of THP1 cells. Prior to T-cell culture, THP1 cells were stained with Tag-it violet to allow identification of THP1 target cells in co-culture. LIVE/DEAD Fixable Near-IR Dead Cell Marker was used to measure the proportion of lysed THP1 cells. (D) CD1c-autoreactive T-cells lyse THP1-CD1c APCs but not THP1-KO APCs. Killing is significantly enhanced when THP1-CD1c cells are infected with live Mtb (MOI=1). (E) Lysis of THP1-CD1c is increased with Mtb infection, but not enhanced with LPS or Pam3CSK4 stimulation. Data are representative of three independent experiments, each preformed in triplicate. ** P < 0.01; ***P < 0.001 (B, D and E, one-way ANOVA with Tukey’s multiple comparison test).

TCR-CD1c interactions mediate the response to Mtb infection.

(A) The alpha (top) and the beta (bottom) chain sequences of EM1 and EM2 TCRs. Variable region (grey), N additions (white) and joining segment (blue) are shown. (B) CD1c-endo tetramer staining of parental Jurkats and Jurkat T-cells transduced to express EM1 and EM2 TCRs. (C) Percentage activation of Jurkat T-cells transduced with EM1 and EM2 TCRs in response to uncoated or CD1c-endo coated wells. T-cell activity was measured by flow cytometry staining with anti-CD69. (D) Jurkat T-cells transduced with EM1 and EM2 TCRs are activated by THP1-CD1c APCs but not THP1-KO APCs, with significantly greater activation when THP1-CD1c APCs are infected with live Mtb (MOI=1). Data are representative of two independent experiments, each preformed in triplicate. ** P < 0.01, *** P < 0.001, **** P < 0.0001 (C, unpaired t-test, D, one-way ANOVA).

CD1c-autoreactive T-cells secrete diverse cytokines and reduce Mtb luminescence.

(A) Cytokines secreted by CD1c-autoreactive T-cells cultured with Mtb-infected THP1-CD1c APCs. CD1c-autoreactive T-cells produced significant amounts of the Th1 cytokines TNF-α, IFN-γ, IL-1α and GM-CSF, and the Th2 cytokines IL-4, IL-5, IL-10 and IL-13 in a T-cell dose dependent manner. Irrelevant control T-cells did not release cytokines. (B) Heat map summarising cytokines released by CD1c-autoreactive T-cells, or irrelevant control T-cells (CD1c unrestricted), in response to Mtb-infected THP1-CD1c APCs. Red indicates high concentrations, and blue indicates low concentrations. Cytokine secretion was measured using a Luminex assay. Data are representative of two independent experiments, each preformed in triplicate. (C) CD1c-autoreactive T-cells reduce Mtb luminescence significantly when cultured with THP1-CD1c APCs relevant to when they are cultured with THP1-KO APCs. T-cells cause some reduction in THP1-KO cells, and this is greater in THP1-CD1c cells. * P < 0.05; ** P < 0.01, **** P < 0.0001 (A, Two-way ANOVA, C, unpaired t-test).

Single-cell profiling reveals that CD1c-endo dextramer positive T-cells are enriched for cytotoxic and effector phenotypes.

(A) Schematic overview of the sample processing workflow. PBMCs were isolated from two donors and CD3+ T-cells were subsequently enriched by negative selection. Cells were stained with CD1c-endo dCODE dextramer, and CD3+CD1c-endo+ (positive) and CD3+CD1c-endo− (negative) T-cells were sorted for single-cell RNA sequencing. (B) UMAP visualisation of 11,804 single T-cells clustered by transcriptional profile. Clusters were annotated as functional T-cell subsets, including CD4⁺ and CD8⁺ naïve, central memory (TCM), effector memory (EM and TEM), cytotoxic, and stress-response populations. Additional subsets included metabolically active T cells and tissue-resident memory-like (Trm-like) CD4⁺ cells. TRM: tissue-resident memory; EM: effector memory; TEM: T effector memory; TCM: T central memory. (C) UMAP projection showing CD1c-endo dextramer binding intensity, reflecting the number of bound dextramer molecules per cell. Cells with higher binding intensities are enriched within cytotoxic and effector CD4+ and CD8+ subsets. (D) Bar plots showing the proportional distribution of T-cell subtypes among CD1c-endo-positive (top) and - negative (bottom) populations. CD1c-endo-positive T-cells were enriched for cytotoxic and effector subsets. (E) Violin plots showing expression levels of the top 10 differentially expressed genes by Wilcoxon rank-sum test. CD1c-endo positive T-cells upregulate genes associated with cytotoxicity (e.g., NKG7, GZMA, GZMK, GNLY, CTSW) and inflammation (e.g., CCL5, NFKBIA, DUSP2), consistent with a distinct effector phenotype. (F) Chord plot illustrating functional enrichment of differentially expressed genes in CD1c-endo positive T-cells. Upregulated genes are associated with biological processes such as cell killing, antigen processing and presentation, and response to other organisms. In contrast, downregulated genes predominantly map to ribonucleoprotein complex biogenesis. The colour gradient indicates the log fold-change (logFC) in gene expression.

(A) Flow cytometry gating strategy analysing the expression of the T-cell activation markers CD69, CD25, and CD137 on proliferated CD3+CTV- T-cells. (B) Cytokine release by T-cells first expanded with THP1-CD1c APCs and then stimulated overnight with THP1-KO or THP1-CD1c APCs. Cytokine secretion was measured by Luminex array. * P < 0.05; ** P < 0.01, *** P < 0.001, **** P < 0.0001 (one-way ANOVA with Tukey’s multiple comparison test). Mean and SD of triplicate measurements are shown and are representative of three individual donors.

(A) Flow cytometry gating strategy of live HLA-DR+/CD11c+ MoDCs (top), and line graphs showing the effect of Mtb infection on the expression of CD1a, CD1b, and CD1c on MoDCs (bottom). (B) Flow cytometry gating strategy depicting live CD14+ THP1-CD1c cells stained with anti-CD1c antibody.

Flow cytometry gating strategy depicting live CD3+ T-cells comprising the negative (cells that did not bind the streptamers) or the CD1c-endo streptamer positive T-cell fraction (containing 1.14% CD1c-endo positive T-cells) stained with CD1c-endo dextramers.

CD1c-endo dextramer positive T-cells were sorted and expanded. After expansion, T-cells were either unstained, or stained with an irrelevant tetramer or with CD1c-endo tetramer. Enriched T-cells brightly stain with CD1c-endo tetramers.

(A) Flow cytometry dot plots showing high expression of CD1c on wild-type JRT3.5 Jurkat T-cells. (B) Flow cytometry dot plots showing the absence of CD1c and TCR expression on β2-microglobulin knock-out JRT3.5 Jurkat T-cells.

(A) CD1c-autoreactive T-cells secrete cytokines in response to untreated THP1-CD1c APCs in an autoreactive manner. Importantly, CD1c-autoreactive T-cells release significantly higher concentrations of cytokine when cultured with Mtb-infected THP1-CD1c APCs. (B) CD1c-autoreactive T-cells secrete diverse cytokines in a CD1c dependent manner. CD1c-autoreactive T-cells release cytokines when cultured with Mtb-treated THP1-CD1c APCs but not when were cultured with Mtb-treated THP1-KO APCs. Data are representative of two independent experiments, each preformed in triplicate. * P < 0.05; ** P < 0.01, **** P < 0.0001 (Two-way ANOVA).

UV Mtb-treated THP1-CD1c APCs secrete chemokines IL-8 and RANTES (pink bars), in comparison to media only (black bars), and untreated THP1-CD1c APCs (blue bars).

Data are representative of two independent experiments, each preformed in triplicate. **** P < 0.0001 (Two-way ANOVA).

(A) Representative flow cytometry gating strategy used to isolate CD3⁺ T-cells stained with CD1c-endo dCODE dextramers. (B) Representative plots from two donors showing distinct CD3⁺CD1c-endo⁺ and CD3⁺CD1c-endo⁻ T-cell populations. Percentages indicate the proportion of positive cells within the CD3⁺ gate.

Dot plot showing the top 5 ranked marker genes for each annotated T-cell population, derived from all high-quality T-cells (n = 11,804).

Dot size represents the proportion of cells expressing each gene within the corresponding cluster, while colour indicates the scaled average expression (z-score). This visualization highlights the transcriptional signatures that define distinct T-cell subsets, including naïve, memory, effector, and metabolically active populations.

Filtering strategy for identifying CD1c-endo dextramer-bound T-cells in scRNA-seq data from sorted populations.

A total of 14,524 cells were initially profiled by scRNA-seq from dextramer-sorted fractions. Following standard quality control, including thresholds for gene and UMI counts, mitochondrial and ribosomal content, doublet detection, and outlier removal, 11,804 high-quality cells were retained. CD1c-endo-negative cells (n = 9,518) were defined by the absence of detectable dextramer signal and further stratified by their sort gate: 7,652 originated from the CD1c-endo-positive gate and 1,866 from the CD1c-endo-negative gate. To minimise false negatives, only the 1,866 cells from the negative sort gate were retained as confidently CD1c-endo-negative. To define CD1c-endo-positive cells, dextramer signal intensities were modelled using kernel density estimation (KDE) and quantile-based filtering (1st and 99th percentiles, IQR). Cells with signal below the lower bound (Q1 – 1.5×IQR) or extreme values were excluded, yielding 11,800 cells. Of these, 2,288 cells with no binding signal were removed. Among the remaining 2,282 CD1c-endo-positive candidates, 2,117 were from the CD1c-endo-positive sort gate and 165 from the negative gate. To ensure specificity, only the 2,117 cells from the positive sort gate were retained as confidently CD1c-endo-positive. This combined quality control and gating strategy enabled robust identification of CD1c-endo-binding T-cells in the single-cell dataset.