Mechanism of inert inflammation in an immune checkpoint blockade-resistant tumor subtype bearing transcription elongation defects

Abstract

The clinical response to immune checkpoint blockade (ICB) correlates with tumor-infiltrating cytolytic T lymphocytes (CTLs) prior to treatment. However, many of these inflamed tumors resist ICB through unknown mechanisms. We show that tumors with transcription elongation deficiencies (TE^def+), which we previously reported as being resistant to ICB in mouse models and the clinic, have high baseline CTLs. We show that high baseline CTLs in TE^def+ tumors result from aberrant activation of the nucleic acid sensing-TBK1-CCL5/CXCL9 signaling cascade, which results in an immunosuppressive microenvironment with elevated regulatory T cells and exhausted CTLs. ICB therapy of TE^def+ tumors fail to increase CTL infiltration and suppress tumor growth in both experimental and clinical settings, suggesting that TE^def+, along with surrogate markers of tumor immunogenicity such as tumor mutational burden and CTLs, should be considered in the decision process for patient immunotherapy indication.

Keywords: CCL5; CP: Cancer; CTL; ICB; TBK1; TIL; Treg; chemokine; dsRNA; immune exhaustion; transcription elongation defects (TE(def)).

Figure 1.. TE^def+ tumors have high, not low, tumor lymphocyte infiltration

(A) T cell population as measured by GZMA expression in human TE^def+ and TE^def− tumors (H-TE^def+ and H-TE^def−tumors) across TCGA datasets of kidney renal clear cell carcinoma (KIRC) (left) and skin cutaneous melanoma (SKCM) (right). (B) T cell population as measured by GZMB expression in H-TE^def+ and H-TE^def− tumors across TCGA datasets of SKCM. (C) Left, representative IHC sections stained for tumor-infiltrating CD8⁺ T cells in CT26 tumors with (TE^def+) or without (TE^def−) flavopiridol treatment. Right: average numbers of tumor-infiltrating CD8⁺ T cells/μm² on an IHC area (n = 3 areas). **p < 0.01. (D) Left: representative CD8⁺ T cell percentages measured by flow cytometry in CT26 flavopiridol pre-treated (murine TE^def+ model) subcutaneous (s.c.) tumors grown in BALB/c mice challenged with combination therapy (100 μg/mouse of anti-CTLA4 and anti-PD1 for 10 days after the onset of a palpable tumor) (n = 3 mice). Right: quantification of tumor-infiltrating CD8⁺ T cells on flow cytometry analysis. **p < 0.01.

Figure 2.. TE^def+ cells have increased endogenous viral RNA expression and faulty mRNA and dsRNA

(A) Left: Boxplot of ERV mRNA derepression in H-TE^def+ and H-TE^def− tumors across TCGA datasets of KIRC. Right: Bar graphs of ERV mRNA derepression in H-TE^def+ and H-TE^def− cell lines. (B) Ratio of defective mRNA/normal mRNA in terms of 5′-capping and 3′-poly A-tail in H-TE^def+ versus H-TE^def− and TE^def+ (CT26 flavopiridol treated) versus control (CT26 without flavopiridol treatment). (C) PicoGreen-based estimation of dsDNA in ng/μL. (D) Left: confocal images of dsRNA (in green) in the cytoplasm of human (H-TE^def+ and H-TE^def−) and murine (with and without flavopiridol treatment) TE^def+ and TE^def− cell lines. Right: quantification of dsRNA signal intensity in 2d (n = 6). **p < 0.01.

Figure 3.. TE^def+ tumors have increased signaling through TBK1-IRF3/7 pathway and pro-inflammatory chemokine expression

(A) Immunoblots of cytokine DNA/RNA sensing pathway in human (H-TE^def+) and murine (with and without flavopiridol treatment) TE^def+ and TE^def− cell models. (B) ELISA readout of the 24-h conditioned media from H-TE^def+ and H-TE^def− cells (n = 3). (C) Intracellular cytokine staining of pro-inflammatory chemokines in murine (CT26 with and without flavopiridol treatment) TE^def+ and TE^def− cell lines (n = 3). (D) qPCR mRNA fold change readout of pro-inflammatory chemokines in CT26 murine TE^def+ (with flavopiridol treatment) and TE^def− (without flavopiridol treatment) cells (n = 3). (E) Intracellular cytokine staining of pro-inflammatory chemokines in TE^def+, TE^def−, and shTBK1 CT26 murine cells (n = 3). (F) Chemotaxis assay of splenic CD8⁺ T cells from tumor-bearing BALB/C mice tested against the conditioned media from TE^def+, TE^def−, and shTBK1 versions of CT26 murine cells; anti-CCL5-, anti-CXCL9-, and anti-CXCL10-(1 μg/mL) treated conditioned media (n = 3). **p < 0.01.

Figure 4.. scRNA-seq reveals early CD8⁺ T cell infiltration in TE^def+ TME offset by immune dampening effects

(A) Tumor growth curves of TE^def+ (with flavopiridol treatment) and TE^def− (without flavopiridol treatment) CT26 murine cells inoculated in BALB/c mice; 100 μg/mouse given every 2 days for 10 days after the onset of palpable tumors (n = 3). (B) Immunoblots confirming the TE^def phenotype in TE^def+ (with flavopiridol treatment) and TE^def− (without flavopiridol treatment) CT26 murine cells before tumor inoculation and post-excision tumor lysates. (C) Representative uniform manifold approximation and projection (UMAP) plot of scRNA-seq 3′v3 assay at resolution 0.2 from 16,000 CT26 murine tumor cells per condition with a coverage of 100,000 reads per cell. (D) Analysis of cytotoxic, helper, and regulatory T cell subsets infiltrating CT26 TE^def+ (with flavopiridol treatment) and TE^def− (without flavopiridol treatment) tumors challenged with anti-CTLA4; resolution at 0.3. **p < 0.01.

Figure 5.. Flow cytometry confirms that TE^def+ tumors attract FOXP3⁺ Treg cells and exhausted CD8⁺ cytotoxic T cells

(A) Tumor growth curves of CT26 TE^def− (without flavopiridol treatment) versus CT26 TE^def+ (with flavopiridol treatment) challenged with anti-CTLA4 or IgG; 100 μg/mouse given every 2 days for 10 days after the onset of palpable tumors (n = 3). (B) Representative flow cytometry gates of CD8⁺ and CD4⁺ T cells infiltration of CT26 murine tumors from (A). (C) Quantification using flow cytometry data of relevant markers in CT26 TE^def− and CT26 TE^def− tumors challenged with or without anti-CTLA4 (n = 3). (D) Quantification using flow cytometry data of CCR5 marker in CT26 TE^def− and CT26 TE^def+ tumors challenged with or without anti-CTLA4 (n = 3). **p < 0.01.

Figure 6.. TBK1 and cGAS knockdown partially reverse TE^def+ resistance to immune checkpoint blockade through distinct tumor extrinsic effects

(A) Western blot characterization of NAS pathway activation in stable short hairpin knockdown of genes TBK1 and cGAS in CT26 cells compared with their scramble controls after a pre-treatment with flavopiridol for 7 days at 25 nM in culture. (B and C) Tumor growth curves of CT26 TE^def+ tumors (with flavopiridol treatment) and TE^def− tumors (without flavopiridol treatment) with either cGAS shRNA, TBK1 shRNA, or scramble inoculated in BALB/c mice; TBK1 shRNA and cGAS shRNA show partial rescue of immune checkpoint blockade (ICB) sensitivity in TE^def+; 100 μg/mouse anti-CTLA4 or IgG given every 2 days for 10 days after the onset of palpable tumors (n = 8); statistical significance based on t test on days 7, 11, 13, 15, 17, 19, 21, 23, and 25. (D and E) Quantification using flow cytometry data of relevant markers in CT26 TE^def+ and CT26 TE^def− tumors with and without cGAS shRNA or TBK1 shRNA tumors challenged with or without anti-CTLA4 (n = 8). **p < 0.01.

Figure 7.. The correlates of inert inflammation of TE^def+ tumors are observed in clinical datasets

(A) Heatmap correlating mRNA level derepression of ERVs in human TE^def+ tumors across multiple cancers in TCGA datasets. (B) Heatmap correlating indicated processes in human TE^def+ tumors across multiple cancers in TCGA, with box highlighting the CD8⁺ T cells, T follicular helper cells, NK cells, B cells, and Treg cells; the indicated processes in the heatmap were scored for each tumor sample in TCGA by the iAtlas study (Thorsson et al.), (C) Boxplot of TE^def+ enrichment in treatment naive and post-ICB in scRNA-seq dataset of patients with melanoma. (D) Left: Correlation of pre-therapy TE^def+ tumor score versus anti-PD1 therapy response in a clinical cohort of Riaz et al. Right: boxplots of pre-therapy TE^def+ tumor score versus anti-PD1 therapy naive and progressing disease. (E) Left: correlation of pre-therapy TE^def+ tumor score versus change in GZMA levels (surrogate of TIL load) across tumor samples in the clinical cohort of Riaz et al. Right: correlation of pre-therapy TE^def+ tumor score versus change in CCNB1 (surrogate of tumor progression) across tumor samples in the clinical cohort of Riaz et al. (F) Inert inflammation presents as phenotypically hot TME at baseline that remains functionally cold upon ICB therapy. Figure made using BioRender.