Cell-autonomous IL6ST activation suppresses prostate cancer development via STAT3/ARF/p53-driven senescence and confers an immune-active tumor microenvironment

Abstract

Background: Prostate cancer ranks as the second most frequently diagnosed cancer in men worldwide. Recent research highlights the crucial roles IL6ST-mediated signaling pathways play in the development and progression of various cancers, particularly through hyperactivated STAT3 signaling. However, the molecular programs mediated by IL6ST/STAT3 in prostate cancer are poorly understood.

Methods: To investigate the role of IL6ST signaling, we constitutively activated IL6ST signaling in the prostate epithelium of a Pten-deficient prostate cancer mouse model in vivo and examined IL6ST expression in large cohorts of prostate cancer patients. We complemented these data with in-depth transcriptomic and multiplex histopathological analyses.

Results: Genetic cell-autonomous activation of the IL6ST receptor in prostate epithelial cells triggers active STAT3 signaling and significantly reduces tumor growth in vivo. Mechanistically, genetic activation of IL6ST signaling mediates senescence via the STAT3/ARF/p53 axis and recruitment of cytotoxic T-cells, ultimately impeding tumor progression. In prostate cancer patients, high IL6ST mRNA expression levels correlate with better recurrence-free survival, increased senescence signals and a transition from an immune-cold to an immune-hot tumor.

Conclusions: Our findings demonstrate a context-dependent role of IL6ST/STAT3 in carcinogenesis and a tumor-suppressive function in prostate cancer development by inducing senescence and immune cell attraction. We challenge the prevailing concept of blocking IL6ST/STAT3 signaling as a functional prostate cancer treatment and instead propose cell-autonomous IL6ST activation as a novel therapeutic strategy.

Keywords: Cytotoxic T-cells; IL6ST/STAT3 signaling; Immune cell infiltration; L-gp130; Prostate cancer; Senescence; Senescence-associated secretory phenotype; Tumor microenvironment.

Fig. 1

Prostate epithelium-specific, cell-autonomous insertion of L-gp130 reduces progressive prostate tumorigenesis. a Illustration of wild type IL6ST receptor, which can be activated by binding of the IL-6 ligand and IL-6R receptor (left panel), and Leucine-gp130 (L-gp130) construct (right panel). Wild type IL6ST consists of an extracellular domain comprising an Ig-like domain, a cytokine binding domain, three fibronectin type III-like domains, a transmembrane domain, and a cytoplasmic domain. For generating L-gp130, wild type IL6ST was truncated 15 amino acids above the transmembrane domain and replaced by the leucine zipper region of the human c-JUN gene and a FLAG-Tag. L-gp130 expression can activate downstream signaling cascades identical to stimulated wild type IL6ST. P: phosphorylation. b Illustration of the genetic approach for conditional deletion of Pten (exon 4 + 5) or/and insertion of L-gp130-ZSGreen in prostate epithelial cells under the control of Probasin (PB) promoter after Cre-mediated recombination resulting in PB-Cre4;Pten^fl/fl;L-gp130^+/+ (hereafter Pten^peΔ/Δ), PB-Cre4;Pten^+/+;L-gp130^fl/fl (hereafter L-gp130^peKI/KI) and PB-Cre4;Pten^fl/fl;L-gp130^fl/fl mice (hereafter Pten^peΔ/Δ;L-gp130^peKI/KI). pe: prostate epithelium; fl: floxed site; ex: exon; 2A: 2A peptide; CAG: CAG promoter; KI: knock in; Δ: knock out. c Representative immunohistochemistry (IHC) pictures of phospho-AKT (p-AKT) and immunofluorescence (IF) pictures of co-stainings of ZSGreen (red) and DAPI (blue) in mouse prostates. DAPI is used as a nuclear stain. Scale bar: 40 µm. Scale bar of inset: 10 µm. d Gross anatomy of representative mouse prostates. Scale bar: 0.5 cm. e Prostate weight of wild type (n = 10), L-gp130^peKI/KI (n = 14), Pten^peΔ/Δ (n = 14), and Pten^peΔ/Δ;L-gp130^peKI/KI (n = 14) mice. Individual biological replicates are shown. Data are plotted as the means ± SD and p-values were determined by ordinary one-way ANOVA with Tukey’s multiple comparisons test. f Representative pictures of hematoxylin & eosin (H&E) stains of mouse prostates at low (top) and high (bottom) magnification. Scale bar upper panel: 60 µm, scale bar lower panel: 10 µm. g Quantification of histopathological analysis of prostate tissue from wild type (n = 9), L-gp130^peKI/KI (n = 9), Pten^peΔ/Δ (n = 11), and Pten^peΔ/Δ;L-gp130^peKI/KI (n = 9) mice in regards of histomorphological criteria for aggressive growth patterns: without pathological findings (white); PIN: prostate intraepithelial neoplasia (grey); PCa: prostate cancer (red)

Fig. 2

L-gp130 leads to activation of the STAT3 transcription factor and STAT3 target gene expression. a Representative immunohistochemistry (IHC) pictures of mouse prostates stained for the epithelial marker EpCAM. Scale bar: 40 µm b RNA-Seq workflow showing processing and magnetic bead-based enrichment of EpCAM-positive (EpCAM⁺) mouse prostate tissue. Prostates were dissected and enzymatically and mechanically dissociated to generate single cell suspensions. Cells were labeled with biotinylated anti-EpCAM antibody and enriched from the bulk population using streptavidin-coated magnetic beads. EpCAM⁺ cells were subjected to RNA-Seq analysis. c Heatmap and number of differentially expressed genes (log2norm) based on adj. p-value ≤ 0.05 and fold change ≥ 2 cut-off values comparing Pten^peΔ/Δ and Pten^peΔ/Δ;L-gp130^peKI/KI prostate epithelial cells (n ≥ 5). blue: downregulated, red: upregulated. d Fast pre-ranked gene set enrichment analysis (fGSEA) of the KEGG gene set “Prostate cancer” with genes regulated in Pten^peΔ/Δ;L-gp130^peKI/KI compared to Pten^peΔ/Δ prostate epithelial cells. Genes sorted based on their Wald statistics are represented as vertical lines on the x-axis. NES: normalized enrichment score. e fGSEA of three previously published STAT3 target signatures (“STAT3 targets (Swoboda)”, “STAT3 targets (Azare)”, “STAT3 targets (Carpenter)”) with genes regulated in Pten^peΔ/Δ;L-gp130^peKI/KI compared to Pten^peΔ/Δ prostate epithelial cells. Genes sorted based on their Wald statistics are represented as vertical lines on the x-axis. NES: normalized enrichment score. f Western Blot analysis of prostate protein lysates for phosphoTyrosine705-STAT3 (pY-STAT3) and total-STAT3 (t-STAT3) expression in Pten^peΔ/Δ and Pten^peΔ/Δ;L-gp130^peKI/KI mice (n = 5). β-ACTIN (β-ACT) served as loading control. g Quantification of pY-STAT3 protein levels relative to t-STAT3 protein levels shown in f). h Representative pictures of IHC staining of pY-STAT3 and t-STAT3 expression in prostate sections of Pten^peΔ/Δ and Pten^peΔ/Δ;L-gp130^peKI/KI mice. Scale bar: 40 µm. i-j Quantitative analysis of pY-STAT3 (i) and semi-quantitative analysis of t-STAT3 (j) IHC stainings shown in h) (n = 7). g,i-j Individual biological replicates are shown (g,i). Data are plotted as the means ± SD and p-values were determined by unpaired two-tailed Student’s t-tests (g,i) or Mann-Whitney test (j)

Fig. 3

High IL6ST expression is significantly associated with low-risk groups and better recurrence-free survival in human PCa. a IL6ST gene expression in adjacent (n = 52) and PCa (n = 497) tissue in TCGA-PRAD data set. Statistical analysis of the two risk groups was determined by using the Mann–Whitney test. b Kaplan–Meier plot showing time of disease-free survival in months for IL6ST^low and IL6ST^high risk groups of the TCGA-PRAD data set. Groups were assessed based on the maximally selected rank statistics. blue: high IL6ST expressing group, red: low IL6ST expressing group. The blue and red numbers above horizontal axis represent the number of patients. c Proportion of IL6ST alterations in the TCGA-PRAD data set. Mutation types: deep deletion (n = 21; red), truncating mutations (n = 2; black) and multiple alterations (n = 1; grey). One patient has simultaneous mutations. The data originate from cBioPortal. d IL6ST mRNA expression levels of four different data sets of PCa patient samples compared to healthy prostate sample control. Normalized data and statistical analyses were extracted from the Oncomine Platform. The respective prostate data set and n-numbers are indicated. Representation: boxes as interquartile range, horizontal line as the mean, whiskers as lower and upper limits. e Spearman-correlation analysis of IL6ST and STAT3 expression in TCGA-PRAD data set using cBioPortal analysis tool

Fig. 4

Expression of L-gp130 induces p19^ARF-p53-driven senescence in Pten-deficient PCa. a Fast pre-ranked gene set enrichment analysis (fGSEA) of significantly enriched HALLMARK gene sets with genes regulated in Pten^peΔ/Δ;L-gp130^peKI/KI compared to Pten^peΔ/Δ prostate epithelial cells. Dotted line: adj. p-value (-log10(0.05)), blue: downregulated, red: upregulated; b Representative pictures of immunohistochemistry (IHC) staining of mouse prostates from the indicated genotypes stained for the proliferation marker Ki67. Scale bar: 40 µm. c Semi-quantitative analysis of Ki67⁺ prostate epithelial cells in the indicated genotypes (n = 7) shown in b). d Representative pictures of IHC staining of PML of Pten^peΔ/Δ and Pten^peΔ/Δ;L-gp130^peKI/KI prostates. Scale bar: 40 µm. e Quantification of PML nuclear bodies per high power field (HPF) shown in d) (n ≥ 5). f Cdkn2a mRNA expression levels based on normalized counts from RNA-Seq analysis of Pten^peΔ/Δ and Pten^peΔ/Δ;L-gp130^peKI/KI prostates (n ≥ 5). g qRT-PCR mRNA expression analysis of p19^ARF in mouse prostate tissue of Pten^peΔ/Δ and Pten^peΔ/Δ;L-gp130^peKI/KI mice (n ≥ 5). Signals are relative to the geometric mean of housekeeping genes. h Western Blot analysis of prostate protein lysates of Pten^peΔ/Δ and Pten^peΔ/Δ;L-gp130^peKI/K mice (n = 5) for p53 expression. β-ACTIN (β-ACT) served as loading control. i Quantification of p53 protein levels shown in h) normalized to loading control. j Proposed model of IL6ST signaling induced senescence. L-gp130 activated STAT3 binds to its binding sites in Cdkn2a promoter, followed by upregulation of p19^ARF and p53 expression promoting senescence in PCa. c,e–g,i Individual biological replicates are shown (e–g,i). Data are plotted as the means ± SD and p-values were determined by Mann–Whitney test (c,f), unpaired two-tailed Student’s t-tests (e,g,i)

Fig. 5

Expression of L-gp130 in Pten^peΔ/Δ mice increases infiltration of immune cells mediating anti-tumor defense. a Representative pictures of hematoxylin & eosin (H&E) stains (upper panel) showing immune infiltrate and quantification of histopathological analysis (lower panel) of prostate tissue from Pten^peΔ/Δ (n = 11) and Pten^peΔ/Δ;L-gp130^peKI/KI (n = 9) mice in regards of infiltration (low-grade (grey) and high-grade (red)). Scale bar: 60 µm. b Representative pictures of immunofluorescence (IF) staining of CD45 (red) and DAPI (blue) of mouse prostates with indicated genotypes. DAPI is used as a nuclear stain. Scale bar: 20 µm. c Quantification of CD45⁺ cells of IF stainings shown in b) (n = 5). The percentage of positive cells relative to Pten^peΔ/Δ was calculated. d Representative pictures of immunofluorescence (IF) staining of CD3 (yellow), CD8 (green) and DAPI (blue) of mouse prostates with indicated genotypes. DAPI is used as a nuclear stain. Scale bar: 20 µm. e Quantification of CD3⁺;CD8⁺ cells in the epithelium of IF stainings shown in d) (n = 5). The percentage of positive cells in the prostate epithelium relative to Pten^peΔ/Δ was calculated. f Representative pictures of immunohistochemistry (IHC) staining of NimpR14 (higher panel) and F4/80 (lower panel) of mouse prostates with indicated genotypes. Scale bar: 40 µm. g-h) Quantification of NimpR14⁺ (g) and F4/80⁺ (h) cells in the prostate epithelium of IHC stainings shown in f) (n ≥ 6). The percentage of positive cells in the prostate epithelium relative to Pten^peΔ/Δ was calculated. i) Flow cytometry data showing mean fluorescence intensity (MFI) of F4/80⁺;Cd11b⁺ macrophages in Pten^peΔ/Δ and Pten^peΔ/Δ;L-gp130^peKI/KI prostate tissue (n ≥ 3) for CD86, MHC class II, and CD206. j Representative flow cytomentry blots of data shown in i. k Representative pictures of IHC staining of CD3, NimpR14 and F4/80 (in presented order) of Pten^peΔ/Δ and Pten^peΔ/Δ;Stat3^peΔ/Δ prostates. Scale bar: 40 µm. l Semi-quantitative analysis of CD3, NimpR14 and F4/80 IHC stainings shown in k) (n ≥ 3). c, e, g-i, l Individual biological replicates are shown (c,e,g-i). Data are plotted as the means ± SD and p-values were determined by unpaired two-tailed Student’s t-tests (c,e,g-i) or Mann–Whitney test (l)

Fig. 6

IL6ST signaling in PCa patients actives STAT3 signaling and upregulates senescence, immune score and cytotoxicity. a Fast pre-ranked gene set enrichment analysis (fGSEA) of significantly enriched HALLMARK gene sets with genes regulated in IL6ST^high compared to IL6ST^low expressing patients from the TCGA-PRAD data set. Dotted line: adj. p-value (-log10(0.05)), blue: downregulated, red: upregulated; b fGSEA of the previously described core SASP gene signature upon PICS “Core SASP of PICS (Guccini)” (upper panel) and the curated gene set, class chemical and genetic perturbations (CGP) “CGP: Fridman senescence up” (lower panel) with genes regulated in IL6ST^high compared to IL6ST^low expressing patients from the TCGA-PRAD data set. Genes sorted based on their Wald statistics are represented as vertical lines on the x-axis. NES: normalized enrichment score. c fGSEA of WikiPathways (WP) gene sets “REACTOME: Cell cycle”, “REACTOME: G1/S transition” and “WP: p53 transcriptional gene network” with genes regulated in IL6ST^high compared to IL6ST^low expressing patients from the TCGA-PRAD data set. Genes sorted based on their Wald statistics are represented as vertical lines on the x-axis. NES: normalized enrichment score. d Immune score from the ESTIMATE method for IL6ST^low (red, n = 208) and IL6ST^high (blue, n = 283) patients from the TCGA-PRAD data set, compared with Mann–Whitney test. e fGSEA of the top 20 T-cell-, neutrophil-, and macrophage-associated Biological Processes from Gene Ontology pathways (GO-BP) gene sets with genes significantly regulated in IL6ST^high compared to IL6ST^low expressing patients from the TCGA-PRAD data set. Dotted line: adj. p-value (-log10(0.05)), red: upregulated;

Fig. 7

Proposed roles of active IL6ST signaling in prostate tumorigenesis. Using the genetic mouse model, we showed that cell-autonomous, prostate epithelium-specific and constitutively active IL6ST signaling reduces Pten-deficient tumor growth, enhances the STAT3/p19^ARF/p53-driven senescence and recruits tumor-infiltrating immune cells (T-cells, neutrophils, and macrophages). In human PCa, high IL6ST expression causes active STAT3 signaling, correlates with better survival and is associated with higher level of immune infiltrates