Stat3-mediated Atg7 expression regulates anti-tumor immunity in mouse melanoma

Abstract

Epigenetic modifications to DNA and chromatin control oncogenic and tumor-suppressive mechanisms in melanoma. Ezh2, the catalytic component of the Polycomb Repressive Complex 2 (PRC2), which mediates methylation of lysine 27 on histone 3 (H3K27me3), can regulate both melanoma initiation and progression. We previously found that mutant Ezh2^Y641F interacts with the immune regulator Stat3 and together they affect anti-tumor immunity. However, given the numerous downstream targets and pathways affected by Ezh2, many mechanisms that determine its oncogenic activity remain largely unexplored. Using genetically engineered mouse models, we further investigated the role of pathways downstream of Ezh2 in melanoma carcinogenesis and identified significant enrichment in several autophagy signatures, along with increased expression of autophagy regulators, such as Atg7. In this study, we investigated the effect of Atg7 on melanoma growth and tumor immunity within the context of a wild-type or Ezh2^Y641F epigenetic state. We found that the Atg7 locus is controlled by multiple Ezh2 and Stat3 binding sites, Atg7 expression is dependent on Stat3 expression, and that deletion of Atg7 slows down melanoma cell growth in vivo, but not in vitro. Atg7 deletion also results in increased CD8 + T cells in Ezh2^Y641F melanomas and reduced myelosuppressive cell infiltration in the tumor microenvironment, particularly in Ezh2^WT melanomas, suggesting a strong immune system contribution in the role of Atg7 in melanoma progression. These findings highlight the complex interplay between genetic mutations, epigenetic regulators, and autophagy in shaping tumor immunity in melanoma.

Keywords: Atg7; Autophagy; Melanoma; Tumor-immune response.

Fig. 1

Regulation of Atg7 expression by Ezh2 and Stat3. a Enriched motifs in Ezh2^WT and Ezh2^Y641F melanoma cells. b Gene Set Enrichment Analysis (GSEA) of Stat3 ChIP-seq peaks identifies several signatures associated with autophagy mechanisms (FDR < 0.05). c Transcript expression of Atg7 in Ezh2^Y641F vs. Ezh2^WT melanoma cells measured by RNA-sequencing, in the absence or presence of the Ezh2 inhibitor JQEZ5. d Human ChIP-seq data in various cell lines showing direct binding of both Stat3 (green) and Ezh2 (blue) at the Atg7 promoter and intronic regions that correspond to cis-regulatory elements. Image modified from UCSC Genome Browser. e ChIP-seq tracks for Ezh2 and Stat3 in Ezh2^WT and Ezh2^Y641F melanoma cells at the mouse Atg7 locus indicating binding at the Atg7 promoter and the first intron

Fig. 2

Deletion of Atg7 in melanoma cells has no significant effect on cell-intrinsic cell growth in vitro. a Top: Protein expression of Stat3 measured by western blot after shRNA-mediated stable gene knockdown in melanoma cell line 234Δ (Y641F). Bottom: Quantification of protein expression, N = 3 independent experiments. b Expression of Atg7 and LC3 after Stat3 knockdown in Ezh2^WT and Ezh2^Y641F melanoma cell lines 234 and 234Δ. c Quantification of western blot in b, N = 2. d Immunoblotting for Atg7 and LC3 in control and Atg7 knockout clones in the 234, 234Δ, 27.6-M2, and 28.2-M4 cell lines. NT = non-targeted sgRNA. Quantification of the Atg7/GAPDH, N = 4, and LC3-II/I, N = 5. e In vitro growth curve of Ezh2^WT and Ezh2^Y641F melanoma cell lines 27.6-M2 and 28.2-M4 with and without Atg7 deletion. N.S. = not statistically significant. For all graphs, error bars are standard deviation; *** p value < 0.001, ** p value < 0.01, and * p value < 0.05

Fig. 3

Deletion of Atg7 in melanoma cells results in slower in vivo tumor growth and increased presence of tumor infiltration of lymphocytes. a (Left) In vivo tumor growth in Ezh2^WT (27.6-M2) and Ezh2^Y641F (28.2-M4) melanomas, with and without Atg7 deletion. The group average is displayed, and error bars indicate the standard deviation. Control = non-targeted sgRNA, N = 8 per group, representative of two independent experiments. (Right) Tumor volume at day 5 post-injection. The bars indicate the group mean, and the circles are individual tumor sizes. (Far right) Image of tumors at day 7. The image has been cropped, and the brightness and contrast have been increased to improve viewing. b Flow cytometric analysis of tumor-infiltrating CD45 + hematopoietic cells and CD45- cells. N = 6–8 tumors per group. c Flow cytometric analysis of tumor-infiltrating CD8 + , CD4 +, and NK1.1 + cells. N = 7–8 tumors per group. d Representative flow cytometry plots of the CD4 + and CD8 + data shown in panel c. For the graphs in b and c, each dot on the graph represents an individual tumor, and the bar marks the average for the group. *p < 0.05, **p < 0.01, and ***p < 0.001

Fig. 4

Deletion of Atg7 in melanoma cells results in decreased infiltration of myelosuppressive cells. a Expression of PD-1 on tumor-infiltrating CD8 + and CD4 + T cells in Ezh2^WT and Ezh2^Y641F melanoma cells, with and without Atg7 deletion. N = 7–8 tumors per group. b Expression of the PD-1 ligand (PD-L1) on the melanoma cells from panel a. N = 6–8 tumors per group. c Flow cytometric analysis of tumor-infiltrated CD11c + , Mac1 +, and double Mac1/Gr1 + cells in Ezh2^WT and Ezh2^Y641F melanoma tumors, with and without Atg7 deletion. N = 6–8 tumors per group. d Representative flow cytometry plots for the Mac1 + and double Mac1/Gr1 + data in panel c. For the graphs in a–c, each dot on the graph represents an individual tumor, and the bar marks the average for the group. *p < 0.05, **p < 0.01, and ***p < 0.001