PRSS2 remodels the tumor microenvironment via repression of Tsp1 to stimulate tumor growth and progression

Abstract

The progression of cancer from localized to metastatic disease is the primary cause of morbidity and mortality. The interplay between the tumor and its microenvironment is the key driver in this process of tumor progression. In order for tumors to progress and metastasize they must reprogram the cells that make up the microenvironment to promote tumor growth and suppress endogenous defense systems, such as the immune and inflammatory response. We have previously demonstrated that stimulation of Tsp-1 in the tumor microenvironment (TME) potently inhibits tumor growth and progression. Here, we identify a novel tumor-mediated mechanism that represses the expression of Tsp-1 in the TME via secretion of the serine protease PRSS2. We demonstrate that PRSS2 represses Tsp-1, not via its enzymatic activity, but by binding to low-density lipoprotein receptor-related protein 1 (LRP1). These findings describe a hitherto undescribed activity for PRSS2 through binding to LRP1 and represent a potential therapeutic strategy to treat cancer by blocking the PRSS2-mediated repression of Tsp-1. Based on the ability of PRSS2 to reprogram the tumor microenvironment, this discovery could lead to the development of therapeutic agents that are indication agnostic.

Fig. 1. Breast and prostate cancer cells repress Tsp-1 in a paracrine manner via PRSS2.

A Immunohistochemistry (IHC) of thrombospondin-1 (Tsp-1) expression in tumors formed in the prostate gland of SCID mice by PC3 and PC3M-LN4 prostate cancer cells (scale bar = 100 μm) (n = 8 per group). B Schematic diagram of the proteomic screening and identification of Tsp-1 repressing protein in PC3M-LN4 conditioned media. C ELISA of Tsp-1 expression (normalized to total protein levels) in PC3M-LN4 fractions eluted from a Heparin-sepharose-Cu²⁺ column (n = 3). D Western blot of PRSS2 and actin protein levels in PC3 and PC3M-LN4 prostate cancer cells (n = 3). E Western blot of PRSS2 and actin protein levels in MDA-MB-231 (231), MDA-MB-231-LM2 (LM2), MCF7 and SUM159 breast cancer cells (n = 3). F Tsp-1 and actin levels in MRC5 fibroblasts and primary human peripheral blood mononuclear cells (PBMCs) that were untreated (–) or treated with recombinant human PRSS2 ( + ) (n = 3). G Plot of Tsp-1 mRNA fold change, measured by real-time quantitative RT-PCR in WI38 cells treated with conditioned media of 293T cells engineered to overexpress PRSS2 (n = 3; P value calculated via two-sided Student’s t test). H Upper panel: Western blot of PRSS2 in SUM159 cells that were untransfected (C), transfected with empty pLKO.1 vector (V), or pLKO.1 vector expressing three independent shRNA sequences directed against PRSS2 (sh1, sh2, sh3). Lower panel: Western blot of Tsp-1 and actin in WI38 cells that were untreated (–) or treated with conditioned media from SUM159 cells with empty vector (V) or PRSS2 shRNA (n = 3). I Western blot of Myc, PRSS2 and actin in SUM159 cells transfected with vector control (V) or three shRNA sequences directed against c-Myc (sh1, sh2 and sh3) (n = 3). J Dot plot of fold change of Myc and PRSS2 in SUM159 cells transfected with vector control (V) or three shRNA sequences directed against c-Myc (sh1, sh2, and sh3 (n = 3). K Dot plot of fold change of Myc and PRSS2 in PC3M-LN4 cells transfected with vector control (V) or three shRNA sequences directed against c-Myc (sh1, sh2, and sh3) (n = 3).

Fig. 2. PRSS2 represses Tsp-1 via enzyme-independent binding to LRP1.

A Western blot of PRSS2 protein levels in conditioned media of 293T cells transfected with WT PRSS2 (blue), PRSS2-G191R (black), PRSS2-S200A (green), PRSS2-S200C (red), and PRSS2-S200T (orange) (n = 3). B Plot of enzymatic activity WT and mutant PRSS2 proteins relative to WT PRSS2 (%) (n = 3) (P value calculated via two-sided Student’s t test; error bars depict SEM). Western blots of: C Tsp-1 and actin expression in WI38 fibroblasts treated with CM from 293T cells transfected with empty pCMV-Sport6 vector or (pCMV-Sport6 -PRSS2 or a mutant version of PRSS2 (GR = G191R, SA = S200A, SC = S200C, ST = S200T) (n = 3; P value calculated via two-sided Student’s t test). D Integrin α2 (ITGA2), Integrin β1 (ITGB1), LRP1, PRSS2, and actin following immunoprecipitation with control IgG, or α-PRSS2 (IP) antibody and unbound protein (UB) in WI38 cells (n = 3). E Dot plot of all replicates of Tsp-1, LRP1, and actin in WI38 cells that were untreated (–) or treated with PRSS2 (+) in the presence (+) or absence (–) of siRNA directed against LRP1 (n = 3; P value calculated via two-sided Student’s t test). F Secreted truncation mutants of LRP1 (sLRP1) containing binding domains 1–4, PRSS2, and actin from an immunoprecipitation experiment with α-myc epitope antibody (n = 4). G Tsp-1 and actin in WI38 cells that were untreated (–) or treated with PRSS2 (+) in the absence (–) or presence of secreted truncated mutants of PRSS2 comprised of binding domains 1, 1a, 1b, 2, 3, and 4 (n = 3). H Dot plot of quantitation of Tsp-1 western blots from all replicates of WI38 cells that were untreated (–) or treated with PRSS2 (+) in the absence (–) or presence of secreted truncated mutants of PRSS2 comprised of binding domains 1, 2, 3, and 4 (n = 3; P value calculated via two-sided Student’s t test).

Fig. 3. PRSS2 represses Tsp-1 by activating Rac downstream of LRP1.

A Dot plot of GTP-bound Rac and Rho in WI38 cells that were untreated (control), treated with CM from 293T cells transfected with empty vector (pCMV) or PRSS2 (PRSS2) (n = 3; P value calculated via two-sided Student’s t test). B Western blot of Tsp-1 and actin expression in WI38 cells that were untreated (–) or treated with PRSS2 (+) in the absence (–) and presence (+) of a small molecule inhibitor of Rac1 NSC23766 (n = 3). C Dot plot of quantitation of Tsp-1 western blots from all replicates of WI38 that were untreated (–) or treated with PRSS2 (+) in the absence (–) and presence (+) of NSC23766 (n = 3; P value calculated via two-sided Student’s t test). D Western blot of phospho c-Jun, total c-Jun and actin in WI38 cells that were treated with saline (–), PRSS2, Rac1 inhibitor NSC23766, or PRSS2 and NSC23766 in combination (n = 3). E Dot plot of quantitation of phospho c-Jun, total c-Jun, and actin in WI38 cells that were treated with saline (–), PRSS2, Rac1 inhibitor NSC23766, or PRSS2 and NSC23766 in combination (P values calculated by two-sided ANOVA; n = 3). F Western blot of phosphor-JNK, total JNK and actin in WI38 cells that were treated with saline (–), PRSS2, Rac1 inhibitor NSC23766, or PRSS2 and NSC23766 in combination (n = 5). G Dot plot of quantitation of phospho-JNK, total JNK and actin in WI38 cells that were treated with saline (–), PRSS2, Rac1 inhibitor NSC23766, or PRSS2 and NSC23766 in combination (P values calculated by two-sided ANOVA; n = 5). H Western blot of Tsp-1 and actin in WI38 cells that were treated with saline (–), PRSS2, JNK1 inhibitor III, or PRSS2 and JNK1 inhibitor III in combination (n = 3). I Schematic diagram of PRSS2 repression of Tsp-1 via LRP1-Rac signaling.

Fig. 4. PRSS2 is required for efficient tumor formation of SUM159 cells.

A Immunohistochemical analysis of breast cancer patient Series 1 (n = 544 patients) for expression of: (i) PRSS2 in tumors with strong (high) PRSS2 expression in tumor cells, (ii) PRSS2 in tumors with weak (low) PRSS2 expression in tumor cells, (iii) PRSS2 in tumors with strong PRSS2 expression in TME, (iv) PRSS2 in tumors with weak PRSS2 expression in the TME; (v) proliferating microvessel density (pMVD) in tumors with strong PRSS2 expression, (vi) pMVD in tumors with weak PRSS2 expression; (vii) proliferation (Ki67) in tumors with strong PRSS2 expression; and (viii) Ki67 in tumors with weak PRSS2 expression (scale bar = 50 μm, original magnification ×400). B H&E staining and immunohistochemical analysis of prostate cancer patient series (n = 458 patients) for expression of: (i) PRSS2 in tumors with strong (high) tumor cell expression of PRSS2 in localized prostatic carcinoma; (ii): PRSS2 in tumors with weak (low) tumor cell expression of PRSS2 in localized prostatic carcinoma; (iii) PRSS2 in tumors with strong expression of PRSS2 in castration-resistant carcinoma (CR); (iv) PRSS2 in tumors with weak expression of PRSS2 in castration-resistant carcinoma; (v) VEGF-A in localized carcinoma with strong PRSS2 expression; (vi) VEGF-A in localized carcinoma with weak PRSS2 expression; (vii) Ki67 in localized carcinoma with strong PRSS2 expression; (viii) Ki67 in localized carcinoma with weak PRSS2 expression (scale bar = 50 μm, original magnification ×400). C Kaplan–Meier curve of clinical progression following radical prostatectomy of prostate cancer patients with strong and weak expression of PRSS2 (P values were determined via two-sided log-rank test). D Kaplan–Meier curve of overall survival following the acquisition of castration resistance of prostate cancer patients with strong and weak expression of PRSS2 (P values were determined via a two-sided log-rank test). E Plot of in vitro proliferation of SUM159 and SUM159shPRSS2 cells over 3 days (no significant difference was determined via two-sided ANOVA). F Plot of in vivo luciferase activity of orthotopic tumors formed by mammary gland injection of SUM159 and SUM159shPRSS2 cells (P value was determined via two-sided Mann–Whitney U test; error bars depict SEM). G In vivo luciferase imaging of mice bearing SUM159 (upper panel) and SUM159shPRSS2 (lower panel) tumors. H Photographs of tumors formed by SUM159 and SUM159shPRSS2 cells. I Plot of the volume of tumors formed by SUM159 and SUM159shPRSS2 cells (P value calculated via two-sided Mann–Whitney U test). J Immunohistochemistry of Tsp-1 in tumors formed by SUM159 and SUM159shPRSS2 cells (n = 8 mice per group) (scale bar = 200 μm).

Fig. 5. Silencing PRSS2 inhibits primary pancreatic tumor growth and metastasis.

A Schematic of tumor implantation strategy. B Western blot of PRSS2 and actin in Pan02 cells transduced with lentiviral vectors specifying 4 shRNA sequences against murine PRSS2. C Photographs of tumors formed by Pan02 and Pan02shPRSS2 (sh1 from 4B) cells in wild-type C57Bl6/J mice and thbs1^−/− C57Bl6/J mice. D Dot plot of mass of tumors formed by Pan02 and Pan02shPRSS2 cells in wild-type C57Bl6/J mice and thbs1^−/− C57Bl6/J mice (n = 6 per group; P value calculated via two-sided Mann–Whitney U test). E Dot plot of ascites volume of tumors formed by Pan02 (WT n = 6; thbs1^−/− n = 6) and Pan02shPRSS2 (WT n = 6; thbs1^−/− n = 6) cells in wild-type C57Bl6/J mice and thbs1^−/− C57Bl6/J mice (P value calculated via two-sided Mann–Whitney U test). F H&E and Tsp-1 immunohistochemical staining of tumors formed by Pan02 and Pan02shPRSS2 cells in wild-type C57Bl6/J mice and thbs1^−/− C57Bl6/J mice (n = 6 mice per group) (scale bar = 200 μm). G Dot plot of the number of macrometastatic lesions identified by gross examination of the peritoneal cavities of mice bearing tumors formed by Pan02 and Pan02shPRSS2 cells in wild-type C57Bl6/J mice (4/6 refers to the number of mice that developed macrometastatic lesions) (n = 6 mice per group; P value calculated via two-sided Mann–Whitney U test). H Dot plot of the area of macrometastatic lesions (in mm²) measured using Image-J (Fiji) analysis of images of the peritoneal cavities of mice bearing tumors formed by Pan02 and Pan02shPRSS2 cells in wild-type C57Bl6/J mice (n = 6 mice per group: P value calculated via two-sided Mann–Whitney U test).

Fig. 6. Myeloid-specific deletion of LRP1 inhibits tumor growth by preventing the repression of Tsp-1.

A Plot of percentage of LRP1 + multiple myeloid and lymphoid cells in wild-type (n = 5) and LysM-Cre/LRP1^fl/fl (n = 5) mice as determined by FACS analysis. B Representative FACS plots of LRP1 expression in myeloid and lymphoid cells in wild-type and LysM-Cre/LRP1^fl/fl mice (n = 5 per group). C Plot of relative abundance of CD45 + myeloid and lymphoid cells in wild-type (n = 5) and LysM-Cre/LRP1^fl/fl (n = 5) mice as determined by FACS analysis. D Western blot of PRSS2 and Actin in Pan02 murine pancreatic cancer cells, SUM159 human breast cancer cells, and E0771 murine breast cancer cells (n = 3). E Plot of average tumor volume (as measured by calipers) of orthotopic mammary tumors formed by E0771 murine breast cancer cells in wild-type (red line), LysM-Cre/LRP1^fl/fl (green line), and LysM-Cre/LRP1^fl/fl/THBS1^−/− (blue) mice (n = 7 per group) (error bars depict SEM). F Dot plot of volume of tumors formed by E0771 cells in LysM-Cre/LRP1^fl/fl mice (LRP1mKO) (n = 7 mice per group). G Photographs of tumors formed by E0771 murine breast cancer cells in wild-type, LysM-Cre/LRP1^fl/fl, and LysM-Cre/LRP1^fl/fl/THBS1^−/− mice. H H&E and Tsp-1 immunohistochemical staining of tumors formed by E0771 murine breast cancer cells in wild-type, LysM-Cre/LRP1^fl/fl, and Tsp-1^−/−/LysM-Cre-LRP1^fl/fl mice (n = 7 mice per group) (scale bars = 100 μm).

Fig. 7. Repression of Tsp-1 by PRSS2 creates and immunosuppressive TME.

A Immunofluorescence staining of CD3 (green) and FoxP3 (red) in tumors formed by Pan02 (n = 4) and Pan02shPRSS2 (n = 3) cells in WT C57Bl6/J mice and Tsp1^−/− C57Bl6/J mice (scale bar = 1 mm). B Plot of the percentage of CD3 + /FoxP3 + T cells out of total CD3 + T cells in tumors formed by Pan02 (n = 4) and Pan02shPRSS2 (n = 3) cells in WT C57Bl6/J mice and thbs1^−/− C57Bl6/J mice (P value calculated via two-sided Mann–Whitney U test). C Immunofluorescence staining of CD3 (green), CD4 (red), and CD8 (white) in tumors formed by Pan02 (n = 4) and Pan02shPRSS2 (n = 3) cells in WT C57Bl6/J mice and Tsp1^−/− C57Bl6/J mice (scale bar = 1 mm). D Plot of percentage of CD3 + /CD8 + T cells out of total CD3 + T cells in tumors formed by Pan02 (n = 4) and Pan02shPRSS2 (n = 3) cells in WT C57Bl6/J mice and thbs1^−/− C57Bl6/J mice (P value calculated via two-sided Mann–Whitney U test). E Plot of ratio of CD3 + /CD8 + T cells to CD3 + /FoxP3 + T cells in tumors formed by (n = 4) and Pan02shPRSS2 (n = 3) cells in WT C57Bl6/J mice and thbs1^−/− C57Bl6/J mice (WT/WT minimum = 0.099; maximum = 0.58; median = 0.37; percentile = 25–75; P value calculated via two-sided Mann–Whitney U test). F Representative pictures of immunohistochemistry of CD8 + and FoxP3 + T cells in prostate tumors (n = 458 patients) with high or weak levels of PRSS2 (scale bar = 50 μm). G Kaplan–Meier curves of: upper left: correlation of high PRSS2 and low CD8 levels with survival after radical prostatectomy. Upper right: correlation of high PRSS2 and high FoxP3 levels with survival after radical prostatectomy; and lower left: correlation of low PRSS2 and high CD8 levels with survival after developing castration resistance; and lower right: correlation of high PRSS2 and high FoxP3 levels with survival after developing castration resistance (P values calculated via two-sided Mann–Whitney U test). H Schematic diagram of PRSS2 regulation of tumor immune microenvironment.