Single-cell RNA sequencing highlights the immunosuppression of IDO1+ macrophages in the malignant transformation of oral leukoplakia

Abstract

Rationale: Immunosuppressive tumor microenvironment (iTME) plays an important role in carcinogenesis, and some macrophage subsets are associated with iTME generation. However, the sub-population characterization of macrophages in oral carcinogenesis remains largely unclear. Here, we investigated the immunosuppressive status with focus on function of a macrophage subset that expressed indoleamine 2,3 dioxygenase 1 (Macro-IDO1) in oral carcinogenesis. Methods: We built a single cell transcriptome atlas from 3 patients simultaneously containing oral squamous cell carcinoma (OSCC), precancerous oral leukoplakia (preca-OLK) and paracancerous tissue (PCA). Through single-cell RNA sequencing and further validation using multicolor immunofluorescence staining and the in vitro/in vivo experiments, the immunosuppressive cell profiles were built and the role of a macrophage subset that expressed indoleamine 2,3 dioxygenase 1 (Macro-IDO1) in the malignant transformation of oral leukoplakia was evaluated. Results: The iTME formed at preca-OLK stage, as evidenced by increased exhausted T cells, Tregs and some special subsets of macrophages and fibroblasts. Macro-IDO1 was predominantly enriched in preca-OLK and OSCC, distributed near exhausted T cells and possessed tumor associated macrophage transformation potentials. Functional analysis revealed the established immunosuppressive role of Macro-IDO1 in preca-OLK and OSCC: enriching the immunosuppression related genes; having an established level of immune checkpoint score; exerting strong immunosuppressive interaction with T cells; positively correlating with the CD8-exhausted. The immunosuppression related gene expression of macrophages also increased in preca-OLK/OSCC compared to PCA. The use of the IDO1 inhibitor reduced 4NQO induced oral carcinogenesis in mice. Mechanistically, IFN-γ-JAK-STAT pathway was associated with IDO1 upregulation in OLK and OSCC. Conclusions: These results highlight that Macro-IDO1-enriched in preca-OLK possesses a strong immunosuppressive role and contributes to oral carcinogenesis, providing a potential target for preventing precancerous legions from transformation into OSCC.

Keywords: 3-dioxygenase 1.; Oral leukoplakia; Oral squamous cell carcinoma; immunosuppressive microenvironment; indoleamine 2; macrophage.

Figure 1

Single-cell transcriptomic landscape of paired human PCA, preca-OLK, and OSCC tissues (A) Overview of the workflow and the experimental design for scRNA-Seq. (B) Representative H&E staining of tissue samples biopsied for scRNA-seq. Scale bar: 100μm. (C) Dot plot showing the highly expressed marker genes in each major cell type. (D) UMAP plot showing the clustering results of 12 major cell types for 42291 high-quality single cells from all the samples. (E) Stacked histogram showing the percentages of individual cell types/total cells from PCA, preca-OLK, and OSCC tissues.

Figure 2

The function enrichment and differentiation trajectory of fibroblast subpopulations (A) UMAP plot of the total fibroblast subsets. (B) UMAP plot showing the distribution of all fibroblasts in PCA, preca-OLK, OSCC. (C) UMAP plots showing the subpopulations of fibroblasts. (D) Heatmap showing representative GO pathways enrichment of the gene set expressed in fibroblasts subsets predicted by GSVA. (E) UMAP plot showing the differentiation trajectory of fibroblasts and the distribution of each cell subset on the trajectory.

Figure 3

T cell dysfunction and cell state transitions in carcinogenesis (A) UMAP plot showing the distribution of T cell subsets. (B) Dot plot showing expression of relative markers in indicated T cell clusters. (C) Proportion of T cell subtypes was shown in bar plots in different tissues. (D-E). Proportion of CD4⁺T cells (D) or CD8⁺T cells (E) subtypes were shown in bar plots in different tissues respectively. (F) Statistical analysis comparing the proportions of PD-1⁺CD3⁺CD8⁺cells (exhausted CD8⁺T cells) in tissues of OLK (n = 7), preca-OLK (n = 8) and OSCC (n = 7). (G) Representative IF staining of human tissues (40x). DAPI (blue), Pan-CK (orange), CD3(purple), CD8(red), PD-1(green). Bar, 20 μm. (H) UMAP plot showing the differentiation trajectory of CD4⁺T cells or CD8⁺T cells and the distribution of each cell subset on the trajectory.

Figure 4

Annotation and distribution of macrophage subpopulations in PCA, preca-OLK and OSCC (A) UMAP plot showing the composition of macrophage colored by cluster. (B) Bar plot showing proportion of each macrophage subtype in each tissue. (C) Statistical analysis comparing the proportions of IDO1⁺CD68⁺cells (IDO1⁺macrophages) in tissues of OLK (n = 25), preca-OLK (n = 15) and OSCC (n = 11). (D) Representative IF staining of human tissues (40x). DAPI (blue), IDO1 (red), CD68 (green). Bar, 50 μm. (E) Heatmap showing representative GO pathways enrichment of the gene set expressed in macrophage subsets predicted by GSVA. (F) Violin plot showing the immunosuppression scoring of macrophage subsets. (G) Heatmap showing the expression of immunosuppressive ligand and receptor molecules in macrophage subsets in PCA, preca-OLK and OSCC.

Figure 5

IDO1⁺ macrophages exert immunosuppressive effects. (A) Heatmap showing the interacting relationship pairs between macrophage subsets and T cell subsets according to CellPhoneDB analysis in preca-OLK. (B) Dot plot showing the interaction intensity of immune checkpoint ligand/receptors between macrophage subsets and CD4/CD8_exhuausted according to CellPhoneDB analysis. (C) mIF results showing the spatial distribution of PD-1⁺CD3⁺CD8⁺ cells (exhausted CD8⁺T cells) around IDO1⁺CD68⁺ cells (IDO1⁺macrophages), red area indicates the area with a radiu≤50 μm, as green area indicates the area with a radiu >50μm from the IDO1⁺CD68⁺ cells (IDO1⁺macrophages). (D-E). Representative image showing the spatial distribution of PD-1⁺CD3⁺CD8⁺ cells (exhausted CD8⁺T cells) in the area ≤50 μm from the IDO1⁺CD68⁺ cells (IDO1⁺macrophages) in preca-OLK (D) and OSCC (E). White arrows indicate examples of IDO1⁺CD68⁺cells; yellow arrows indicate examples of PD-1⁺CD3⁺CD8⁺ cells.

Figure 6

IDO1 is upregulated in HNSC, and activation of the IFN-γ-JAK-STAT axis is the potential cause for differentiation to IDO1⁺macrophages. (A) Umap plot showing the differentiation trajectory of macrophages and the distribution of each cell subset on the trajectory. (B) Comparison of IDO1 expression between tumor (n = 519) and normal (n = 44) in TCGA database. (C) The boxes showing the expression distribution of IDO1 in PCA (G1), OLK (G2) and OSCC (G3) in GEO data. (D) Gene set enrichment analysis (GSEA) results showing the enrichment of JAK-STAT gene sets in Macro_IDO1. (E) Information flow of feature signaling pathway in different tissues. (F) The CellPhoneDB communication analysis demonstrating that IFN-II signaling pathway mediates communication between T cells and macrophages in preca-OLK and OSCC. (G) Scatter plot showing the correlation between IFNG and IDO1 expression across 4 independent datasets with OLK in GEO data. (H) Heatmap showing RAS activity of regulons in macrophage subtype predicted by SCENIC.

Figure 7

The IDO1 inhibitor significantly reduces 4NQO induced oral carcinogenesis in mice (A) Schematic plot showing the induction of OSCC by 4NQO in C57BL/6 mice. (B) Macroscopic lesions on the tongues of each group of mice. The dotted circles indicate macroscopic cauliflower-like lesions. (C) Statistical results of macroscopic lesions on the tongues of each group of mice. (D) Statistical results of the tongue lesions (carcinoma) in mice of each group. (E) Representative microscopic images of typical pathological images following H&E staining. Scale bar: 100 μm. (F) Graphical illustration of the working model. Abbreviations: OLK, oral leukoplakia; preca-OLK, precancerous OLK; OSCC, oral squamous cell carcinoma; Trp, tryptophan; Kyn, kynurenine.