TGF-β neutralization attenuates tumor residency of activated T cells to enhance systemic immunity in mice

Abstract

A tissue resident-like phenotype in tumor infiltrating T cells can limit systemic anti-tumor immunity. Enhanced systemic anti-tumor immunity is observed in head and neck cancer patients after neoadjuvant PD-L1 immune checkpoint blockade (ICB) and transforming growth factor β (TGF-β) neutralization. Using T cell receptor (TCR) sequencing and functional immunity assays in a syngeneic model of oral cancer, we dissect the relative contribution of these treatments to enhanced systemic immunity. The addition of TGF-β neutralization to ICB resulted in the egress of expanded and exhausted CD8⁺ tumor infiltrating lymphocytes (TILs) into circulation and greater systemic anti-tumor immunity. This enhanced egress associated with reduced expression of Itgae (CD103) and its upstream regulator Znf683. Circulating CD8⁺ T cells expressed higher Cxcr3 after treatment, an observation also made in samples from patients treated with dual TGF-β neutralization and ICB. These findings provide the scientific rationale for the use of PD-L1 ICB and TGF-β neutralization in newly diagnosed patients with carcinomas prior to definitive treatment of locoregional disease.

Keywords: Biological sciences; Cancer; Cancer systems biology; Immune response; Immunology; Natural sciences; Systems biology.

Graphical abstract

Figure 1

Detection of peripheral TCRs associated with exhausted CD8⁺ TILs was greater with the addition of TGF-β neutralization to PDL1 blockade (A) Illustration shows the experimental sequencing approach used to measure changes in the frequency of TCRs associated with TILs in spleens and tumor-draining lymph nodes. FACS, fluorescent-activated cell sorting; CDR, complimentary determining region. (B) Scatterplot shows uniform manifold approximation and projection (UMAP) embedding of all CD8⁺ TILs, colored by assigned cluster identify. (C) Scatterplot shows UMAP embedding of CD8⁺ TILs, colored by expression of select genes related to activation and exhaustion (Tox, Entpd1, and Havcr2) or stemness (Tcf7). (D) Dot plot shows expression of select TIL-related genes across CD8⁺ TIL clusters. Circle color corresponds to scaled mean expression; circle size denotes fraction of cells with non-zero gene expression of corresponding gene. (E) Box and whisker plots show the frequency of splenic CDR3 TCRβ sequences matched to CD8⁺ TIL TCRs for individual TIL clusters and colored by treatment condition. Significance between treatment conditions for each cluster, determined by a Wilcoxon test, is listed in the supplemental data file. n = 5 mice per treatment group. (F and G) Heatmaps show the relative expression of (F) Itgae or (G) Znf683 in exhausted CD8⁺ TILs. Significance between the exhausted CD8⁺ TILs from the αTGF-β+αPDL1 and αPDL1 treatment conditions, determined with a Wilcoxon test, is listed in the supplemental data file. n = 5 mice per treatment group. (H) Bar plot shows the median fluorescence intensity (MFI) of pSMAD2^S465/S467/3^S423/S425 expression in CD8⁺ TILs measured by flow cytometry. n = 4–5 mice per treatment group. Significance between treatment groups was determined with a Mann-Whitney test. (I) Bar plot show percentage of CD8⁺ TILs positive for CD103 as measured by flow cytometry. n = 7–13 mice per treatment group. Significance between treatment groups was determined with a Mann-Whitney test.

Figure 2

Kinetics of challenge tumor rejection observed with PDL1 blockade were enhanced with the addition of TGF-β neutralization (A) Illustration shows the experimental approach to functional assessment of systemic anti-tumor immunity using engraftment of a secondary challenge tumor. (B) Line graph shows primary tumor growth curves, colored by treatment. n = 15–20 mice per treatment group. Significance was determined with one-way ANOVA with multiple comparisons, considering tumor volume from the start of treatment (day 9). Black arrows correspond to treatments. The red arrow corresponds to challenge tumor engraftment (2 days after completion of treatment). (C) Line graph shows challenge tumor growth curves, colored by treatment. n = 15–20 mice per treatment group. Significance was determined with one-way ANOVA with multiple comparisons, considering tumor volume from engraftment (day 0). (D) Kaplan-Meier plot shows survival, colored by treatment. n = 15–20 mice per treatment group. Significance was determined with a log rank test. (E) Line graph shows the percentage of mice with engrafted tumors, colored by treatment. Significance between αPDL1 and αTGF-β+αPDL1 engraftment rates was determined at each time point with a Fisher’s exact test. (F) Representative photomicrographs of tumor sections stained with hematoxylin and eosin (H&E, left) or a pan-cytokeratin antibody (right), by treatment condition. (G) Line graph shows tumor growth curves following treatment of mice prior to tumor engraftment. Black arrows indicate pre-treatments. n = 5 mice per treatment group. (H) Line graph shows challenge tumor growth curves, colored by treatment with either full dose αPD-L1 (350 μg/injection) or αPDL1 and αTGF-β+αPDL1 (492 μg/injection) or reduced dose αPDL1 (35 μg/injection) or αPDL1 and αTGF-β+αPDL1 (49.2 μg/injection). Significance was determined with one-way ANOVA with multiple comparisons, considering tumor volume from engraftment (day 0). n = 5–10 mice per treatment group. (I) Line graph shows challenge tumor growth curves, colored by treatment with αTGF-β+αPDL1 alone or in combination with CD8 depletion. Significance was determined with one-way ANOVA with multiple comparisons, considering tumor volume from engraftment (day 0). n = 10 mice per treatment group. (J) Line graph shows challenge tumor growth curves, colored by treatment with αTGF-β+αPDL1 alone or in combination with FTY-720 administered beginning two days before treatment (day 7), 30 μg IP/injection, and continued every other day for three weeks. n = 10 mice per treatment group.

Figure 3

Greater immunity observed with combination PDL1 blockade and TGF-β neutralization using direct measurements of anti-tumor immunity (A) Bar graph shows the percentage of CD8⁺ cells within the total CD45⁺ compartment in challenge tumors, by treatment. Significance between treatment groups was determined with a Mann-Whitney test. n = 10–12 mice per treatment group. (B) Bar graph shows the percentage of FoxP3⁺CD25⁺ cells within the total CD4⁺ compartment in challenge tumors, by treatment. Significance between the treatment groups was determined with a Mann-Whitney test. n = 5–6 mice per treatment group. (C) Bar graph shows the percentage of p15E tetramer⁺ cells within the splenic CD8⁺ compartment, by treatment. Spleens were harvested two days after completion of treatment. Significance between the treatment groups was determined with a Mann-Whitney test. n = 8–10 mice per treatment group. (D) Bar graph shows the percentage of MOC1 tumor-specific splenic CD8⁺ T cells, by treatment. Spleens were harvested two days after completion of treatment. Significance between the treatment groups was determined with a Mann-Whitney test. n = 9 mice per treatment group. (E) Illustration shows the experimental approach for the in vivo cytotoxicity assay. (F) Dot plot shows the percentage of splenic CD45.1⁺ cells positive for cell trave violet (CTV, loaded with p15E antigen) or cell trace yellow (CTY). Fractions of the cell product prior to adoptive transfer are shown as “Injected.” (G) Dot plot shows the percentage of selective killing of CTV-labelled (p15E antigen loaded) adoptively transferred cells. Significance between the treatment groups was determined with a Mann-Whitney test. n = 5 recipient mice per treatment group.

Figure 4

Greater CD8a⁺ T cell clonal expansion and activation in primary tumors observed with combination PDL1 blockade and TGF-β neutralization (A) Line graph shows primary tumor growth curves, colored by treatment with either full or reduced dose αPD-L1 or αTGF-β+αPDL1. Significance was determined with one-way ANOVA with multiple comparisons, considering tumor volume from the start of treatment (day 9). n = 5–10 mice per treatment group. (B) Line graph shows primary tumor growth curves, colored by treatment with αTGF-β+αPDL1 alone or in combination with CD8 depletion. Significance was determined with one-way ANOVA with multiple comparisons, considering tumor volume from the start of treatment (day 9). n = 10 mice per treatment group. (C) Line graph shows challenge tumor growth curves, colored by treatment with αTGF-β+αPDL1 alone or in combination with FTY-720. Significance was determined with one-way ANOVA with multiple comparisons, considering tumor volume from the start of treatment (day 9). n = 5 mice per treatment group. (D) Stacked bar graphs show the cell counts (y axis) and cluster distribution by color of the 10 CD8⁺ TIL clonotypes (x axis) with the greatest frequency within each treatment group. (E) Bar graph shows the number of distinct CD8⁺ TIL clonotypes by treatment. (F) Bar graph shows the Gini index as a measure of CD8⁺ TIL TCR clonality by treatment. (G) Bar graph shows the percentage of CD8⁺ cells within the total CD45 compartment in primary tumors, by treatment. Significance between treatment groups was determined with a Mann-Whitney test. n = 10–12 mice per treatment group. (H) Scatterplot shows UMAP embedding of all CD8⁺ TILs, colored by treatment (left) and a heatmap shows the relative frequency of cluster-associated CD8⁺ TILs within the entire CD8⁺ TILs compartment by treatment (right). n = 5 mice per treatment group. (I) Volcano plot shows the log2 fold change and significance of differentially expressed genes comparing exhausted CD8⁺ TIL from the αTGF-β+αPDL1 and αPDL1 treatment groups. Significance for each gene is included in the supplemental data file. n = 5 mice per treatment group. (J) Violin plots show expression of select genes from I. Significance between treatment groups was determined with a Wilcoxon test. n = 5 mice per treatment group. (K) Bar graph shows the percentage of CXCR3 positive CD8⁺ TILs in primary tumors, by treatment. n = 10 mice per treatment group. Significance between treatment groups was determined with a Mann-Whitney test. (L) Bar graph shows the percentage of CXCR3 positive CD8⁺ splenic T cells, by treatment. n = 5 mice per treatment group. Significance between treatment groups was determined with a Mann-Whitney test.

Figure 5

Increased CXCR3 observed on circulating T cells from patients treated with bintrafusp alfa (A) Illustration shows the collection of PBMC before and after neoadjuvant treatment of patients with newly diagnosed HNSCC with αTGF-β+αPDL1. (B) Representative flow cytometry dot plots show T cell gating and CXCR3 positive CD8⁺ and CD4⁺ peripheral T cells from patients. (C and D) Connected line dot plots show flow cytometry quantification of pre- and post-treatment CXCR3 median fluorescent intensity (MFI, C) or percent CXCR3 positivity (D) of CD8⁺ and CD4⁺ peripheral T cells (n = 14). Lines connect pre- and post-treatment sample values for individual patients. Significance determined with Wilcoxon matched-pairs signed rank tests.

Figure 6

Greater CD4⁺ T_H1 polarization and Treg alterations in primary tumors observed with combination PDL1 blockade and TGF-β neutralization (A) Line graph shows primary tumor growth curves, colored by treatment with αTGF-β+αPDL1 alone or in combination with CD4 depletion. Significance was determined with one-way ANOVA with multiple comparisons, considering tumor volume from the start of treatment (day 9). n = 10 mice per treatment group. (B) Bar graph shows the percentage of CD4⁺ cells within the total CD45 compartment in primary tumors, by treatment. Significance between treatment groups was determined with a Mann-Whitney test. n = 10–11 mice per treatment group. (C) Bar graph shows the mean fluorescence intensity (MFI) of pSMAD expression in CD4⁺ TILs measured by flow cytometry. n = 4–5 mice per treatment group. Significance between treatment groups was determined with a Mann-Whitney test. (D) Bar graph shows CD103 expression on CD4⁺ TILs measured by flow cytometry. n = 4 mice per treatment group. Significance between treatment groups was determined with a Mann-Whitney test. (E) Scatterplot shows UMAP embedding of all CD4⁺ TILs, colored by treatment (left) and a heatmap showing the relative frequency of cluster-associated CD4⁺ TILs within the entire CD4⁺ TIL compartment by treatment (right). n = 5 mice per treatment group. (F) Box and whisker plot shows the Reactome T_H1 differentiation score within the CD4⁺ T_EX cluster by treatment. Significance between treatment groups was determined with a Wilcoxon test. n = 5 mice per treatment group. (G) Violin plot shows Cxcr3 expression on CD4⁺ T cells from the Tex cluster by treatment group. Significance between treatment groups was determined with a Wilcoxon test. n = 5 mice per treatment group. (H) Bar graph shows the percentage of FoxP3⁺CD25⁺ cells within the total CD4 compartment in primary tumors, by treatment. Significance between the treatment groups was determined with a Mann-Whitney test. n = 6–8 mice per treatment group. (I) Bar graph shows the percentage of FoxP3⁺CD25⁺ cells within the total CD4 compartment in the tdLN, by treatment. Significance between the treatment groups was determined with a Mann-Whitney test. n = 6–8 mice per treatment group. (J) Volcano plot shows the log2 fold change and significance of differentially expressed genes comparing FoxP3⁺CD25⁺CD4⁺ TIL (Treg clusters 1–3 considered together) from the αTGF-β+αPDL1 and αPDL1 treatment groups. Significance for each gene is included in the supplemental data file. n = 5 mice per treatment group.