Hedgehog Signaling Regulates Treg to Th17 Conversion Through Metabolic Rewiring in Breast Cancer

Abstract

The tumor immune microenvironment dynamically evolves to support tumor growth and progression. Immunosuppressive regulatory T cells (Treg) promote tumor growth and metastatic seeding in patients with breast cancer. Deregulation of plasticity between Treg and Th17 cells creates an immune regulatory framework that enables tumor progression. Here, we discovered a functional role for Hedgehog (Hh) signaling in promoting Treg differentiation and immunosuppressive activity, and when Hh activity was inhibited, Tregs adopted a Th17-like phenotype complemented by an enhanced inflammatory profile. Mechanistically, Hh signaling promoted O-GlcNAc modifications of critical Treg and Th17 transcription factors, Foxp3 and STAT3, respectively, that orchestrated this transition. Blocking Hh reprogramed Tregs metabolically, dampened their immunosuppressive activity, and supported their transdifferentiation into inflammatory Th17 cells that enhanced the recruitment of cytotoxic CD8+ T cells into tumors. Our results demonstrate a previously unknown role for Hh signaling in the regulation of Treg differentiation and activity and the switch between Tregs and Th17 cells in the tumor microenvironment.

Figure 1.. Hedgehog signaling blockade diminishes Treg abundance and activity in tumor and spleen of mammary tumor bearing mice.

A. Schematic of in vivo experimental timeline. Briefly, tumor-bearing mice (E0771, 4T1, or EMT6) were treated with Smo-i or Veh 3 times per week for three weeks via oral gavage. B. UMAPs of immune and tumor cell clusters (left) and T cell subsets (right) identified from scRNA-seq analysis of 4T1 mammary tumors from mice treated +/− Smo-i. C. Heatmap depicting relative expression of Hh pathway gene expression in T cell subsets from scRNA-seq analysis. D. Foxp3 expression in T cell cluster from 2 representative Smo-i or vehicle treated mice. E. Flow cytometric contour plots and quantitation of 4T1 tumor infiltrating (top) and splenic (bottom) Treg populations (% positive Foxp3 out of CD4⁺ T cells). Independent experiments were performed at least twice. N=15 mice per group for tumors and 8 mice per group for spleens. F. Flow cytometric contour plots and quantitation of EMT6 tumor infiltrating (top) and splenic (bottom) Treg populations (% positive Foxp3 out of CD4⁺ T cells). Independent experiments were performed at least twice. N=12 mice per group for tumors and 5 mice per group for spleens. G. Line plots representing Teff proliferation and CFSE intensity of CD4⁺ Teffs from Treg suppression assay from tumors (top) or spleens (bottom) of 4T1 tumor bearing mice or H. EMT6 tumor bearing mice treated +/− Smo-i. Treg suppression assays were performed twice from 4T1 model and once from EMT6 model. For Treg suppression assay, Tregs were pooled from 10 tumor bearing mice per treatment group. All data is from Tregs differentiated within the tumor or spleens of tumor bearing mice.

Figure 2.. Hedgehog signaling is upregulated in iTregs, and blocking Hh signaling diminishes Treg differentiation and activity.

A. Transcript levels of Gli1 and Gli2 in Th0s vs Tregs, B. Flow cytometry quantitation of Th0s cultured with 4T1 tumor CM +/− 5E1 or isotype control (% positive Foxp3 out of CD4+ T cells) C. Flow cytometry quantitation of Th0s and iTregs treated with Hhi or Veh for 24 hours post differentiation (% positive Foxp3 out of CD4⁺ T cells). D. Transcript levels of Il2ra and Tgfb1 in iTregs, where Hhi is incorporated into the differentiation cocktail, or E. after differentiation. F. Transcript levels of Ebi3, and Il12a in iTregs differentiated for 72 hours, followed by a 24 hour treatment with Hhi post differentiation. G, H. Transcript levels of Il2ra, Tgfb1, Ebi3, and Il12a in iTregs treated with exogenous Shh for 24 hours following differentiation. I. Transcript levels of Il2ra and Tgfb1 in iTregs treated with 4T1 CM +/− 5E1 or isotype control for 24 hours after differentiation. All transcript levels in this figure were determined by qRT-PCR. qRT-PCRs and flow cytometry in A, C-H were repeated at least twice, qRT-PCR in I and flow cytometry in B was performed once. All data is from iTregs differentiated ex vivo. Hhi (GANT61) was used at 20μM for all experiments. N=3 per treatment group for all panels and data plotted as +/− SD.

Figure 3.. Hh signaling blockade alters iTreg bioenergetics.

A. Flow cytometry of 4T1 splenic Tregs differentiated in vivo expressing Ki-67 from mice treated with Veh or Smo-i bearing 4T1 tumors. N=5 per treatment group +/− SEM. B. Seahorse extracellular flux analyzer used to calculate OCR of Th0 and iTregs treated with Hhi for 24 hours following differentiation, from mitochondrial stress test C. Seahorse extracellular flux analyzer was used to calculate basal OCR, ATP-linked respiration, proton leak, and reserve capacity of Th0s and iTregs +/− Hhi. D. Seahorse extracellular flux analyzer used to calculate ECAR of Th0 and iTregs +/− Hhi E. Seahorse extracellular flux analyzer was used to calculate basal ECAR, oligomycin sensing respiration, maximal respiration, and non-glycolytic respiration of iTregs +/− Hhi. Seahorse data (B-E) was collected once and is from iTregs differentiated ex vivo and Hhi (GANT61) was used at 20μM. Technical replicates for Seahorse experiment include the following: n=3 for Th0 group, n=5 for iTreg + DMSO group, and n=3 for iTreg + Hhi group. T cells from 30 mice were pooled to contribute to the various groups for Seahorse assay and all data is plotted +/− SD.

Figure 4.. Hh signaling blockade diminishes Treg O-GlcNAcylation of Foxp3 and STAT3, potentiating the Treg to Th17 switch.

A. iTregs were treated with Veh or Hhi for 24 hours following differentiation followed by metabolomics assessment. Schematic of HBP depicting fold change of metabolites involved in UDP-GlcNAc biosynthesis from iTregs + Hhi (purple bar) and iTregs + DMSO (blue bar). B. Western Blot for O-GlcNAcylated proteins (RL2Ab), OGT, and OGA expression in Th0s, or iTregs +/− Hhi for 24 hours following differentiation. C. Transcript levels of Ogt and Mgea5 in Th0s, or D. iTregs treated with Shh for 24 hours following differentiation. E. Transcript levels of Ogt and Mgea5 in iTregs +/− Hhi. F. iTregs +/− Hhi for 24 hours following differentiation were immunoprecipitated for O-GlcNAcylated (RL2Ab), then immunoblotted for Foxp3 or STAT3. Densitometry values for Ips were determined by normalizing by the average of empty beads and isotype control lanes, and additionally normalized to input. G. Western Blot for p-STAT3 (Y705), and STAT3 expression in Th0s, or iTregs treated with Hhi or Veh for 24 hours post differentiation. H. Luciferase assay to assess Foxp3 activity in EL4 cells differentiated into Tregs and treated +/− Hhi for 24 hours following differentiation. N=3 per treatment group. I. Transcript levels of Il17a, Rorc, and Ifng in iTregs, where LLL12 is added after differentiation for 24 hours. J. Transcript levels of Il17a, Rorc, and Ifng in iTregs treated with 4T1 either SHH, or CM +/− 5E1 or isotype control for 24 hours after differentiation. All transcript levels in this figure were determined by qRT-PCR, and data is plotted as fold change +/− SEM. Immunoblots in B-E, G, and luciferase assay in H were performed at least twice. Immunoprecipitation in F and qRT-PCR in J was performed once. All data is from iTregs differentiated ex vivo and Hhi (GANT61) was used at 20μM. N=3 per treatment group for all qRT-PCR assessments and data plotted as +/− SD.

Figure 5.. Hh signaling blockade promotes the switch from Tregs to Th17 cells.

A. Flow cytometric quantitation of 4T1 tumor infiltrating Tr17 populations from 4T1 tumor-bearing mice treated with Smo-i or Veh for 3 weeks (% positive IL-17a⁺ Rorγt⁺ out of Tregs, n=9 per treatment group), and B. Inflammatory Tr17 populations from the 4T1 tumor (% positive IFN-γ⁺ or GM-CSF⁺ out of Tr17s, n=7 per treatment group), and Ki-67 expression in these populations (n=5 per treatment group) and C,D spleen (n=8 per treatment group for Tr17s and inflammatory Tr17s, n=5 per treatment group for Ki-67+ Tr17s and inflammatory Tr17s. E. Flow cytometric quantitation of EMT6 tumor infiltrating Tr17 populations (n=7 per treatment group), and F. inflammatory Tr17 populations from the EMT6 tumor (n=7 per treatment group), and G,H. Spleen (n=7 per treatment group). Data in A-H plotted as +/− SEM. I. Flow cytometric quantitation of Th17 markers in iTregs treated with Hhi during differentiation (% positive IL-17a⁺ Rorγt⁺ out of Tregs, n=3 per treatment group). J. qRT-PCR analysis of transcript levels of Il17a and Rorc in iTregs, where Hhi is incorporated into the differentiation cocktail (n=3 per treatment group). K. Flow cytometric quantitation of Th17 inflammatory markers in iTregs treated with Hhi during differentiation (% positive IFN-γ⁺ or GM-CSF⁺ out of Tr17s, n=3 per treatment group). L. Flow cytometric quantitation of Th17 markers in iTregs treated with Hhi for 24 hours after differentiation (n=3 per treatment group). M. qRT-PCR analysis of transcript levels of Il17a and Rorc in iTregs, where Hhi is incorporated for 24 hours after differentiation (n=3 per treatment group). N. Flow quantitation of inflammatory markers in iTregs treated with Hhi after differentiation (n=3 per treatment group). All qRT-PCR data is plotted as fold change +/− SD. In vivo experiments A-D were repeated at least twice. In vivo experiments E-H were performed once. Ex vivo experiments I-N were performed once and T cells were pooled from 20 mice. Hhi (GANT61) was used at 20μM for all ex vivo experiments.

Figure 6.. Tumor derived Hh ligand promotes iTreg differentiation and blunts Th17 differentiation.

A. Il17a expression from scRNA-seq analysis in the lymphocyte cluster from representative Smo-i treated or Veh treated 4T1 tumor-bearing mice. B. Heatmap of Th17 related genes in Th17s from scRNA-seq analysis from vehicle or Smo-i treated 4T1 tumor-bearing mice. C. Flow cytometric quantitation of 4T1 tumor infiltrating populations from mice treated with Veh or Smo-i from left to right: i) Th17 populations (% positive IL-17a+ Rorγt+ out of CD4+ T cells, n=12 per treatment group), ii) inflammatory Th17 populations from the tumor (% positive IFN-γ⁺ or GM-CSF⁺ out of Th17s, n=7 per treatment group), and Ki-67 expression in iii) Th17s (n=5 per treatment group) and iv) inflammatory Th17s (n=5 per treatment group). D. Flow cytometric quantitation of 4T1 tumor bearing mice splenic populations from left to right: i) splenic Th17 populations (% positive IL-17a⁺ Rorγt⁺ out of CD4⁺ T cells, n=7 per treatment group), ii) splenic inflammatory Th17 populations from the tumor (% positive IFN-γ+ or GM-CSF+ out of Th17s, n=7 per treatment group), and Ki-67 expression in iii) splenic Th17s (n=5 per treatment group) and iv) splenic inflammatory Th17s (n=5 per treatment group). E. Flow cytometric quantitation of EMT6 tumor infiltrating Th17 populations (left, n=7 per treatment group) and (right) inflammatory Th17 populations (n=7 per treatment group) and F. Splenic Th17 populations (left n=7 per treatment group) and (right) inflammatory Th17 populations (n=7 per treatment group). In vivo experiments C-D were performed at least twice. In vivo experiments E-F were performed once. Data plotted as +/− SEM.

Figure 7.. Hh signaling blockade increases the abundance and activity of tumor CD8⁺ T cells.

A. Fold change of CD8⁺ T cell recruiting chemokine transcript expression in 4T1 tumor infiltrating Th17s from scRNA-seq analysis of tumor from 4T1 tumor-bearing mice treated with Veh or Smo-i. B. Fold change of CD8⁺ T cell recruiting chemokine receptor transcript expression in 4T1 tumor infiltrating CD8⁺ T cells and C. cytotoxic T cells from scRNA-seq analysis. D. Flow cytometric quantitation of 4T1 tumor infiltrating CD8⁺ T cells from mice treated with Veh or Smo-i (n=15 per treatment group) and splenic total CD8⁺ T cells (n=12 per treatment group). E. Flow cytometric quantitation of EMT6 tumor infiltrating CD8⁺ T cells (n=5 per treatment group) and splenic total CD8⁺ T cells (n=7 per treatment group) F. 4T1 tumor infiltrating (n=10 per treatment group) and splenic cytotoxic CD8⁺ T cells (n=7 per treatment group) (% positive GZMB+ out of CD8⁺ T cells), and G. EMT6 tumor infiltrating (n=12 per treatment group) and splenic (n=7 per treatment group) cytotoxic CD8⁺ T cells. H. GO pathway analysis of migration pathways significantly enriched in CD8⁺ T cells and I. cytotoxic CD8⁺ T cells from Smo-i treated mice. Rich factor is calculated using the following formula: # of differentially-expressed genes divided by total genes within GO term or pathway. J. Flow cytometric quantitation of 4T1 tumor and spleen infiltrating inflammatory CD8⁺ T cells (% positive IFN-γ+ out of CD8⁺ T cells). K. Schematic for in vivo adoptive transfer experiment. L. Flow cytometric quantitation of 4T1 tumor (n=5 Vehicle and 7 treated) and spleen infiltrating (n=5 vehicle and 7 treated) total CD8⁺ T cells, CD8⁺ CD45.1⁺ cells, CD8⁺ CD45.2⁺ cells, and CD8⁺ CD45.2⁺ GZMB⁺ cells, represented as number of cells per million total cells. In vivo experiments D-G were performed at least twice. In vivo experiments in J and L were performed once. Data plotted as +/− SEM.