Stromal-derived MAOB promotes prostate cancer growth and progression

Abstract

Prostate cancer (PC) develops in a microenvironment where the stromal cells modulate adjacent tumor growth and progression. Here, we demonstrated elevated levels of monoamine oxidase B (MAOB), a mitochondrial enzyme that degrades biogenic and dietary monoamines, in human PC stroma, which was associated with poor clinical outcomes of PC patients. Knockdown or overexpression of MAOB in human prostate stromal fibroblasts indicated that MAOB promotes cocultured PC cell proliferation, migration, and invasion and co-inoculated prostate tumor growth in mice. Mechanistically, MAOB induces a reactive stroma with activated marker expression, increased extracellular matrix remodeling, and acquisition of a protumorigenic phenotype through enhanced production of reactive oxygen species. Moreover, MAOB transcriptionally activates CXCL12 through Twist1 synergizing with TGFβ1-dependent Smads in prostate stroma, which stimulates tumor-expressed CXCR4-Src/JNK signaling in a paracrine manner. Pharmacological inhibition of stromal MAOB restricted PC xenograft growth in mice. Collectively, these findings characterize the contribution of MAOB to PC and suggest MAOB as a potential stroma-based therapeutic target.

Fig. 1.. MAOB levels are elevated in the stroma of human PC.

(A) Representative images and quantitation of MAOB and αSMA IHC staining in the stroma of normal prostates (n = 12) and PC (n = 97) from a US Biomax TMA. Scale bars, 50 μm. (B) Representative images and quantitation of MAOB and αSMA IHC staining in the stroma along with PC progression categorized by low GS (4 to 6, n = 18) and high GS (7 to 10, n = 79) from the TMA in (A). Scale bars, 50 μm. (C) Representative images and quantitation of MAOB and αSMA IHC staining in the stroma along with PC progression categorized by low GS (n = 70) and high GS (n = 125) from the NYU TMA. Scale bars, 50 μm. (D) Western blot of MAOB and its quantitation from three independent blots of three matched pairs of PC patient–derived fibroblasts, normal prostate fibroblasts (PNFs), and PC-associated fibroblasts (PCFs). (E) Western blot of MAOB in NPFs and CAFs derived from WT mouse prostates and Pten-KO mouse prostate tumors, respectively. Statistical analysis was performed using unpaired Student’s t test. Data represent means ± SEM. *P < 0.05, **P < 0.01.

Fig. 2.. MAOB up-regulation in the tumor stroma is associated with worse clinical outcomes in PC patients.

(A) Representative images and quantitation of MAOB IHC staining in the stroma of hormone-naïve (HNPC, n = 38) and castration-resistant (CRPC, n = 16) PCs. Scale bars, 50 μm. (B) Representative images and corresponding Pearson’s correlation analysis of stromal MAOB and adjacent epithelial CHGA expression from the CRPC cohort (n = 16) in (A). Scale bars, 100 μm. (C) Western blot of MAOB in PrSC and PCF2 cells upon ENZ treatment (20 μM) at indicated times. (D and E) Kaplan-Meier recurrence-free (D, n = 195) and cancer-specific (E, n = 161) survival curves of PC patients from the NYU cohort with either low or high stromal MAOB protein levels. (F) Kaplan-Meier recurrence-free survival curves of BC patients from GSE9014 with either low or high MAOB mRNA levels in the tumor stroma. Statistical analysis was performed using unpaired Student’s t test in (A) and log-rank test in (D) to (F). Data represent means ± SEM. **P < 0.01.

Fig. 3.. Stromal MAOB promotes PC cell proliferation, invasion, and tumor growth.

(A) Western blot of MAOB in control and MAOB-OE/MAOB-KD PrSC fibroblasts. (B) Quantitation of PC cells in 2D coculture with control and MAOB-manipulated PrSC cells by Luc assays (n = 3). (C) Representative fluorescence images and quantitation of green fluorescent protein (GFP)–tagged PC-3 or red fluorescent protein (RFP)–tagged C4-2 cells in 2D coculture with control and MAOB-manipulated PrSC cells by fluorescence microscopy (n = 3). Scale bars, 100 μm. (D) Quantitation of PC cells incubated with CM from control and MAOB-manipulated PrSC cells for 3 to 5 days (n = 3). (E) Quantitation of EpCAM⁺ cancer epithelium in 3D cocultures of PC-3 cells with control and MAOB-manipulated PrSC cells by flow cytometry (n = 3). (F) Representative images and quantitation of invasive PC cells in transwell-based coculture with control and MAOB-manipulated PrSC cells (n = 3). Scale bars, 200 μm. (G) Representative fluorescence images and quantitation of LuCaP 147CR and LuCaP 93 PC PDX–derived organoids after 14-day incubation with CM from control and MAOB-manipulated primary fibroblasts (n = 3). Scale bars, 50 or 800 μm. (H) BLI-based growth curves of subrenal capsule xenografts combining Luc-tagged C4-2 cells with control or MAOB-KD PrSC cells in male SCID mice (n = 5). (I) Representative anatomical images of tumor-grown mouse kidneys from each group. Green dashed circles denote size of tumor outgrowth from renal capsules. (J) Tumor weights from each group (n = 5). (K) Representative images of hematoxylin and eosin (H&E) and IHC staining of stromal (Str) MAOB and tumor (T) Ki-67, AR, and SYP and their quantitation in tumor samples from each group (n = 5). Scale bars, 10 μm. Statistical analysis was performed using unpaired Student’s t test for comparisons between two groups and one-way analysis of variance (ANOVA) with Dunnett’s test for comparisons between three groups in (B) to (F), (G), (J), and (K), and two-way ANOVA with Sidak’s test in (H). Data represent means ± SEM. *P < 0.05, **P < 0.01; ns, not significant.

Fig. 4.. MAOB induces development of reactive stroma in a ROS-dependent manner.

(A) GSEA plot of “response to wounding” gene signature enriched in MAOB-OE PrSC cells versus controls. (B) Western blot of αSMA in control and MAOB-OE/MAOB-KD PrSC cells. (C) Representative αSMA IHC staining of C4-2 tumors co-inoculated with control or MAOB-KD fibroblasts in mice. Scale bars, 100 μm. (D) Enzyme-linked immunosorbent assay (ELISA) of TGFβ1 secretion in the culture media of control and MAOB-manipulated PrSC cells (n = 3). (E) Western blot of p-Smad2 and p-Smad3 in control and MAOB-manipulated PrSC cells. (F) GSEA plots of “TGFβ1 targets up” gene signature enriched in MAOB-manipulated PrSC cells versus controls. (G) Quantitation of intracellular ROS levels in control and MAOB-KD PrSC cells (n = 3). (H) GSEA plots of two ROS-related gene sets enriched in MAOB-KD PrSC cells versus controls. (I) Western blot of αSMA in control and MAOB-manipulated PrSC cells upon NAC (5 mM, 48 hours) or H₂O₂ treatment (40 μM, 24 hours). (J) qPCR of indicated reactive stromal markers in control and MAOB-manipulated PrSC cells upon NAC (5 mM, 48 hours) or H₂O₂ treatment (40 μM, 24 hours) (n = 3). (K and L) Determination of collagen levels deposited (K) and collagen-based cell contraction (L) in control and MAOB-KD PrSC cells under H₂O₂ treatment (40 μM, 24 hours) (n = 3). (M) Quantitation of PC-3 cells in coculture with control and MAOB-OE PrSC cells pretreated with NAC (5 mM, 24 hours) (n = 3). Statistical analysis was performed using unpaired Student’s t test for comparisons between two groups and one-way ANOVA with Dunnett’s test for comparisons between three groups in (D) and (G); one-way ANOVA with Tukey’s test in (J), (K), and (M); and two-way ANOVA with Tukey’s test in (L). Data represent means ± SEM. *P < 0.05, **P < 0.01.

Fig. 5.. MAOB activates CXCL12 in prostate tumor stroma.

(A) GSEA plots of “chemokine signaling pathway” and “chemotaxis” gene sets enriched in MAOB-OE PrSC cells compared with controls. (B) ELISA of CXCL12 secretion in the culture media of control and MAOB-KD PrSC cells (n = 3). (C) Representative IHC images and corresponding Pearson’s correlation analysis of stroma-expressed MAOB and CXCL12 protein levels in a PC TMA (n = 37). Scale bars, 100 μm. (D) Pearson’s correlation analysis of MAOB and CXCL12 mRNA levels in patient-derived cultured prostatic stromal cells (left, n = 20) and laser-capture microdissected breast tumor stroma (right, n = 53) from GSE34312 and GSE9014 datasets, respectively. Statistical analysis was performed using one-way ANOVA with Dunnett’s test in (B). Data represent means ± SEM. **P < 0.01.

Fig. 6.. MAOB induces CXCL12 transcription and expression through Twist1 synergizing with TGFβ1/Smads in prostate stromal cells.

(A) Western blot of Twist1 in control and MAOB-manipulated PrSC cells upon NAC (5 mM, 48 hours) or H₂O₂ (40 μM, 24 hours) treatment. (B) ELISA of CXCL12 secretion in culture media of control and MAOB-OE PrSC cells treated with TWIST1 siRNA or NAC (5 mM, 48 hours) (n = 3). (C) qPCR of CXCL12 in indicated PrSC cells upon NAC treatment (5 mM, 48 hours) or Twist1/TWIST1 siRNA expression (n = 3). (D and E) Determination of CXCL12 mRNA (D) and 0.7-kb promoter activity (E) in PrSC cells upon Twist1 expression and/or TGFβ1 treatment (10 ng/ml, 12 hours) (n = 3). (F) Schematic diagrams of WT and mutated CXCL12 E-box/SBE-Luc constructs and determination of their activities in PrSC cells upon Twist1 expression and/or TGFβ1 treatment (10 ng/ml, 12 hours) (n = 3). (G) Representative proximity ligation assay staining and quantitation of indicated Twist1-Smad interactions by per-nucleus fluorescence intensity in control and MAOB-OE PrSC cells. Smad antibody incubation alone served as negative control. Numbers of nuclei included for comparisons between groups are denoted. Scale bars, 50 μm. (H) Co-IP assays of indicated Twist1-Smad interactions in PrSC cells with coexpression of Twist1 and individual Smads. Immunoglobulin G (IgG) was used in IP as negative control. Ten percent input was blotted as positive control. (I) ChIP analysis of chromatin from control and MAOB-OE PrSC cells precipitated with anti-Twist1, anti-Smad4, or a control IgG, followed by qPCR probing the E-box/SBE-centric CXCL12 promoter region (n = 3). (J) ChIP analysis of chromatin from PrSC cells precipitated with anti-Smad4 antibody and then reprecipitated with anti-Twist1 or a control IgG (re-ChIP), followed by qPCR probing the E-box/SBE-encompassing CXCL12 promoter sequence (n = 3). Statistical analysis was performed using one-way ANOVA with Tukey’s test. Data represent means ± SEM. *P < 0.05, **P < 0.01; ns, not significant.

Fig. 7.. Stromal MAOB promotes PC cell proliferation, migration, and invasion via CXCL12-CXCR4/Src/JNK paracrine signaling.

(A) Quantitation of C4-2 and PC-3 PC cells in monoculture and coculture with indicated PrSC fibroblasts upon anti-CXCL12 (0.1 μg/ml) antibody treatment (n = 3). (B) Quantitation of PC cells in coculture with indicated PrSC cells treated with rCXCL12 protein (50 ng/ml; n = 3). (C) Western blot of CXCR4 and CXCR7 in control and CXCR4-KD/CXCR7-KD PC cells. (D and E) Quantitation of control and CXCR4-KD/CXCR7-KD PC cells in coculture with indicated PrSC cells treated without (D) or with (E) rCXCL12 (50 ng/ml; n = 3). (F) Quantitation of PC cells in coculture with indicated PrSC cells treated with 10 nM AMD3100 (n = 3). (G) Representative images and quantitation of PC-3 cell migration and invasion in coculture with indicated PrSC cells upon treatment with anti-CXCL12 antibody (0.1 μg/ml) or 10 nM AMD3100 (n = 3). Scale bars, 200 μm. (H) Representative images and quantitation of PC cell migration in coculture with indicated PrSC cells treated with rCXCL12 (50 ng/ml; n = 3). Scale bars, 100 μm. (I) Phospho-antibody array analysis of PC-3 cells treated with indicated PrSC cell CM. All phosphoprotein levels were normalized to their total forms from a single array with six replicate spots, with significantly activated phosphoproteins (fold change > 1.5, P < 0.05) denoted. (J) Western blot of p-Src and p-JNK in control and CXCR4-KD PC-3 cells treated with indicated PrSC cell CM. (K) Western blot of p-Src and p-JNK in PC-3 cells treated with indicated PrSC cell CM plus rCXCL12 (50 ng/ml). (L) Quantitation of PC cells in coculture with indicated PrSC cells following pretreatment with 40 nM Src inhibitor 1 or 10 μM SP600125 for 24 hours (n = 3). Statistical analysis was performed using one-way ANOVA with Tukey’s test. Data represent means ± SEM. *P < 0.05, **P < 0.01; ns, not significant.

Fig. 8.. Pharmacological inhibition of MAOB in stromal cells restricts PC growth in mice.

(A) Quantitation of C4-2 and PC-3 cell proliferation in monoculture and coculture with control and MAOB-OE PrSC cells upon selegiline treatment (10 nM, 72 hours) (n = 3). (B) BLI-based growth curves of Luc-tagged PC-3 tumors grown in the prostates of male NSG mice treated with selegiline at various doses (0.5, 2, and 10 mg/kg) or saline as a vehicle (n = 5 per group). (C) BLI images of mice from each group at the end point. (D) Determination of tumor weights (n = 5). (E) Determination of MAOA and MAOB enzymatic activities in mouse liver tissue from each group at the end point (n = 3). (F) Representative images of H&E and IHC staining of tumor-expressed Ki-67, p-Src, and p-JNK and stroma-expressed αSMA and CXCL12 and their quantitation in tumor samples from each group (n = 5). Scale bars, 100 μm. (G) Mouse body weights determined weekly (n = 5). (H) Representative H&E images of mouse liver and kidney tissue from each group. Scale bars, 100 μm. (I to L) ELISA of ALT (I), AST (J), BUN (K), and creatinine (L) in mouse sera at the end point (n = 5). (M) Schematic depicting stromal-derived MAOB activation of paracrine CXCL12-CXCR4/Src/JNK signaling through interplay between ROS-dependent Twist1 (via a HIF1α/VEGF-A/AKT/FOXO1 pathway) and TGFβ1/Smads to promote stromal-epithelial interactions for PC growth and progression. Statistical analysis was performed using one-way ANOVA with Tukey’s test in (A), (D) to (F), and (I) to (L) and two-way ANOVA with Tukey’s test in (B) and (G). Data represent means ± SEM. *P < 0.05, **P < 0.01; ns, not significant.