Melanoma-intrinsic NR2F6 activity regulates antitumor immunity

Abstract

Nuclear receptors (NRs) are implicated in the regulation of tumors and immune cells. We identify a tumor-intrinsic function of the orphan NR, NR2F6, regulating antitumor immunity. NR2F6 was selected from 48 candidate NRs based on an expression pattern in melanoma patient specimens (i.e., IFN-γ signature) associated with positive responses to immunotherapy and favorable patient outcomes. Correspondingly, genetic ablation of NR2F6 in a mouse melanoma model conferred a more effective response to PD-1 therapy. NR2F6 loss in B16F10 and YUMM1.7 melanoma cells attenuated tumor development in immune-competent but not -incompetent mice via the increased abundance of effector and progenitor-exhausted CD8⁺ T cells. Inhibition of NACC1 and FKBP10, identified as NR2F6 effectors, phenocopied NR2F6 loss. Inoculation of NR2F6 KO mice with NR2F6 KD melanoma cells further decreased tumor growth compared with NR2F6 WT mice. Tumor-intrinsic NR2F6 function complements its tumor-extrinsic role and justifies the development of effective anticancer therapies.

Fig. 1.. Identification of NRs potentially linked to antitumor immunity in patients with melanoma.

(A) Outline of workflow in the selection process. ICT response data were obtained from scRNA-seq of patients with melanoma (59). * denotes the gene expression data of malignant cells prepared based on GSE70256 (57). # denotes P < 0.05, Kaplan Meier analysis. (B) Heatmaps show Spearman’s correlation coefficient (top) and corresponding P values (bottom) in comparisons of expression of 48 NRs (x axis) with 10 IFN-γ signature genes (y axis). Twenty NRs showing a significant correlation with the IFN-γ signature were selected. (C) Those 20 NRs, which were either positively (red) or negatively (blue) correlated (Corr.) with the IFN-γ signature (IFN-γ sign.), were assessed for correlation with patient overall survival (top) and patient responses to ICT (bottom). Relevant to overall survival, red and blue bars represent respective favorable and unfavorable overall survival relative to the expression of corresponding NRs. Red and blue bars are relevant to ICT response, representing the respective “sensitivity” and “resistance” of patients with high NR-expressing tumor cells to ICT. Squares at the bottom show percent expression and expression level of NRs in tumor cells from patients resistant (Res, top) or sensitive (Sen, bottom) to ICT. Expression of eight NRs that correlated with patient response to ICT, were selected (indicated by asterisks). (D) Tumor-intrinsic average and percent expression of those eight NRs was assessed using scRNA-seq data from specimens of patients with melanoma (57). Four NRs (as indicated by asterisks) expressed at higher levels in >10% of tumor cells were either positively or negatively correlated with IFN-γ signature genes (E), response to ICT (F), or overall patient survival (G). TCGA, The Cancer Genome Atlas; OS, overall survival; TPM, transcript per million.

Fig. 2.. NR2F6 control of murine melanoma growth requires an intact immune system.

(A) B16F10 cells were transduced with scramble (Scr) control shRNA or one of two shRNAs (sh194 and sh226) targeting murine Nr2f6. NR2F6 mRNA and protein expression was assessed by qPCR and immunoblotting. n = 3 for each group. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (B) Growth of cultured cells described in (A), as assessed in vitro using CellTiter-Glo. Relative fold differences in luminescence on days 1 and 3 were calculated relative to luminescence on day 0. n = 6 for each group. (C) C57BL/6 mice were then inoculated with cells and treated with anti–PD-1 antibody (RMP1-14) on days 6, 9, 12, and 15 (arrows). Tumor volumes were monitored at indicated time points. n = 5 mice for each group. (D) NR2F6 KO B16F10 cells were established by CRISPR using specific single guide RNAs (sg492 and sg495). NR2F6 mRNA and protein expression was assessed by qPCR and immunoblotting, respectively. n = 3 for each group. (E and F) Cells established in (D) were used to inoculate C57BL6 (E) or NSG (F) mice, and tumor volumes were monitored at indicated time points. n = 8 mice (E) and n = 6 mice (F) for each group. (G) YUMM1.7 cells were transduced with scrambled (B-Scr) control shRNA or one of two shRNAs (B-sh194 and B-sh226) targeting Nr2f6 as in (A). mRNA and protein expression was assessed by qPCR and immunoblotting, respectively. n = 3 for each group. (H and I) Cells established in (G) were then used to inoculate C57BL/6 (H) or NSG (I) mice, and tumor volumes were monitored at indicated time points. n = 9 mice (H) and n = 5 mice (I) for each group. Data are presented as means ± SD. Statistical significance was assessed by one-way analysis of variance (ANOVA) with Dunnett’s test (A, D, and G) and two-way ANOVA with Dunnett’s test (B, C, E, F, H, and I).

Fig. 3.. NR2F6 limits murine melanoma growth by controlling CD8 T cell infiltration.

(A) Control and NR2F6 KO B16F10 cells were engrafted into C57BL/6 mice. Tumors were collected 12 days later and assessed for weight and volume. n = 4 mice for each group. (B) CD45⁺ cell abundance (expressed as a percentage) in cells within the singlet gate, as assessed by FACS. n = 4 mice for each group. (C) Abundance of T cell subtypes within all CD45⁺ cells, as assessed by fluorescence-activated cell sorter (FACS). n = 4 mice for each group. (D and E) Control and NR2F6 KO (sg492 and sg495) B16F10 tumor sections were prepared from tumors collected on day 18 after inoculation and stained with anti-mouse CD8 antibody. The number of CD8⁺ staining was visualized via microscope (×20) and counted in four random regions of each section. n = 3 mice for each group. Scale bar, 50μm. (F) Control and NR2F6 KO B16F10 tumors were prepared as in (A). FACS assessed the abundance of Texh subsets. n = 4 mice for each group. (G to I) Mouse groups were injected with control IgG or anti-CD8A antibodies to deplete CD8⁺ T cells. Eight days later, CD8⁺ T cell abundance was assessed in blood samples (G). Tumor growth (H) and overall animal survival (I) were monitored at indicated time points. n = 8 mice for each group. Data are presented as means ± SD. Statistical significance was assessed by one-way ANOVA with Dunnett’s test (A, B, C, E, and F), Student’s t test (G), two-way ANOVA with Sidak’s test (H), or by long-rank test (I).

Fig. 4.. RNA-seq identifies NACC1 and FKBP10 as NR2F6 effectors.

(A) DEGs between control and CRISPR-KO B16F10 were identified on the basis of RNA-seq analysis of bulk tumors, MACS-sorted tumor cells, or cultured cells. Unique or common genes up-regulated (top) or down-regulated (bottom) are plotted in Venn diagrams. (B and C) Activated or repressed pathways (B) and upstream regulators (C) of identified DEGs were analyzed using Ingenuity Pathway Analysis. (D) Heatmaps show the Spearman’s correlation coefficient (top) and corresponding P values (bottom) in expression analysis of 15 genes commonly up- or down-regulated in RNA-seq data from the three samples described in (A). x and y axes indicate up-/down-regulated genes and 10 IFN-γ signature genes, respectively. CXCL10, FKBP10, and NACC1 expression were significantly correlated with NR2F6 and the IFN-γ signature. (E) Overall survival of melanoma patients with relatively high expression (z score > 2.0) of NACC1, FKBP10, or CXCL10 was compared with all other patients (z score < 2.0) based on Kaplan-Meier analysis. (F) NR2F6 expression was assessed in melanoma specimens from patients with low or high expression of NACC1, FKBP10, or CXCL10 in (E). Data are presented as means ± SD. Statistical significance was assessed by log-rank test (E) or one-way ANOVA with Sidak’s test (F). ERK, extracellular signal–regulated kinase; MAPK, mitogen-activated protein kinase; HIF1, hypoxia-inducible factor 1; IRF, interferon regulatory factor; T_H1, T helper 1; FLT3, fms-like tyrosine kinase 3; BEX2, brain-expressed X-linked 2; PD-L1, programmed death-ligand 1; LPS, lipopolysaccharide; NAD, nicotinamide adenine dinucleotide.

Fig. 5.. Loss of NACC1 or FKBP10 attenuates tumor growth in mice with an intact immune system.

(A and B) Expression of indicated genes was assessed in NR2F6 KO (CRISPR-based) (A) or KD (shRNA-based) (B) B16F10 cells by qPCR. n = 3 for each group. (C) ReMap peaks predicted transcriptional regulatory regions of Nacc1. R1 included promoter and a part of intron1, and the other candidates (R2, R3, and R4) were predicted within intron1. (D) Abundance of NR2F6, RNA polymerase II, and H3K27 acetylation on each candidate region was assessed by ChIP-qPCR using corresponding antibodies and primers. Relative abundance to input (5% of pre-pulldown material) was calculated. n = 3 for each group. (E) Abundance of NR2F6, RNA polymerase II, and H3K27 acetylation on R4 was assessed in control and NR2F6 KO B16F10 cells. n = 3 for each group. (F) B16F10 cells were transduced with control scrambled (Scr) shRNA or two shRNAs (sh334 and sh336) targeting NACC1. mRNA and protein expression was assessed by qPCR and immunoblotting, respectively. n = 3 for each group. (G and H) Cells were then used to inoculate C57BL/6 (G) or NSG (H) mice. Tumor volumes were monitored at indicated time points. n = 9 mice (G) and n = 5 mice (H) for each group. (I to K) As in (F) and (H), B16F10 cells were transduced with control scrambled shRNA or two (sh486 and sh876) shRNAs targeting FKBP10. FKBP10 expression was assessed (I). n = 3 for each group. The growth of tumors emerging from transduced cells was monitored in C57BL/6 (J) or NSG (K) mice. n = 10 mice (J) and n = 5 mice (K) for each group. Data are presented as means ± SD. Statistical significance was assessed by one-way ANOVA with Dunnett’s test (A, B, E, F, and I) or two-way ANOVA with Dunnett’s test (G, H, J, and K).

Fig. 6.. Loss of NACC1 and FKBP10 phenocopies NR2F6 loss.

(A) B16F10 cells were transduced with scrambled shRNAs [Scr (puromycin resistant) and B-Scr (blasticidin resistant)] and one of two shRNAs (sh334 for NACC1 KD and B-sh876 for FKBP10 KD). mRNA and protein expression was assessed by qPCR and immunoblotting, respectively. n = 3 for each group. (B) Growth of transduced cells in vitro, as assessed by CellTiter-Glo. n = 6 for each group. (C and D) Transduced cells were used to inoculate C57BL/6 (C) or NSG (D) mice. Tumor growth was monitored at indicated time points. n = 9 and 10 mice for Scr/B-Scr and sh334/B-876 group, respectively. (E) Overall survival was assessed in patients with melanoma (SKCM in TCGA, n = 428) whose specimens showed both high NACC1 and high FKBP10 expression, high expression of just one, or low expression of both. (F) NR2F6 expression was assessed in groups of specimens in (E). (G and H) Scrambled (Scr/B-Scr) and NACC1/FKBP10 KD (sh334/B-sh876) B16F10 cells were engrafted into C57BL/6 mice. Tumors were collected 12 days later, and their volume was assessed (G). Different shapes of data points indicate the pooled cells for FACS analysis. n = 7 and 10 mice for Scr/B-Scr and sh334/B-876 group, respectively. (H) Shown is the abundance of T cell subtypes within all CD45⁺ cells based on FACS. n = 5 pooled samples for each group. Data are presented as means ± SD. Statistical significance was assessed by Student’s t test (G) with Holm-Sidak’s test (A and H), one-way ANOVA with Dunnett’s test (F), two-way ANOVA with Sidak’s test (B to D), or log-rank test (E).

Fig. 7.. Combined tumor cell-intrinsic and cell-extrinsic NR2F6 loss synergize to inhibit melanoma growth.

(A and B) B16F10 control (NR2F6 WT) or NR2F6 KO (sg492 and sg495) cells were used to inoculate C57BL/6 control (NR2F6 WT) or NR2F6 KO mice. Tumor volumes (A) and overall animal survival (B) were monitored at indicated time points. n = 8 mice for each group. (C) Violin plots compare the expression of NR2F6, NACC1, and FKBP10 in malignant melanoma cells from patients classified as ICT responders (n = 12) and nonresponders (n = 23). Each dot represents a single malignant cell, and gray horizontal lines are the mean expression in each group. Statistical significance was assessed by two-way ANOVA with Tukey’s test (A), log-rank test (B), or Wilcoxon test (C).