Exosomal αvβ6 integrin is required for monocyte M2 polarization in prostate cancer

Abstract

Therapeutic approaches aimed at curing prostate cancer are only partially successful given the occurrence of highly metastatic resistant phenotypes that frequently develop in response to therapies. Recently, we have described αvβ6, a surface receptor of the integrin family as a novel therapeutic target for prostate cancer; this epithelial-specific molecule is an ideal target since, unlike other integrins, it is found in different types of cancer but not in normal tissues. We describe a novel αvβ6-mediated signaling pathway that has profound effects on the microenvironment. We show that αvβ6 is transferred from cancer cells to monocytes, including β6-null monocytes, by exosomes and that monocytes from prostate cancer patients, but not from healthy volunteers, express αvβ6. Cancer cell exosomes, purified via density gradients, promote M2 polarization, whereas αvβ6 down-regulation in exosomes inhibits M2 polarization in recipient monocytes. Also, as evaluated by our proteomic analysis, αvβ6 down-regulation causes a significant increase in donor cancer cells, and their exosomes, of two molecules that have a tumor suppressive role, STAT1 and MX1/2. Finally, using the Pten^pc-/- prostate cancer mouse model, which carries a prostate epithelial-specific Pten deletion, we demonstrate that αvβ6 inhibition in vivo causes up-regulation of STAT1 in cancer cells. Our results provide evidence of a novel mechanism that regulates M2 polarization and prostate cancer progression through transfer of αvβ6 from cancer cells to monocytes through exosomes.

Keywords: Exosomes; M2 polarization; Monocytes; Prostate Cancer; STAT1; αvβ6 integrin.

Fig. 1. Exosomal α_vβ₆ is transferred from prostate cancer cells to PBMC

(A), Left, nanoparticle size distribution analysis of PC3 exosomes (Exo) by NTA. Right, IB analysis of β₆ integrin, exosomal markers CD63, CD81 and calnexin (CANX) in lysates of PC3 Exo and cells (TCL). (B), PBMC (3.0×10⁵ cells) derived from β₆-null mice (pool of 3 mice) were incubated with PC3 Exo (30 μg/mL at a concentration of 2.4×10⁹ vesicles/μg) for 24 hours. The cells were washed with acid wash buffer twice followed by IB analysis of cell lysates for expression of β₆ integrin and actin (loading control). (C), PBMC (3.0×10⁵ cells) from two different human healthy donors were incubated with indicated PC3 Exo concentrations (0, 12, 30 μg/mL at a concentration of 2.4×10⁹ vesicles/μg) for 24 hours and cell lysates were analyzed by IB for expression of β₆ integrin and actin (loading control). (D), THP1 cells (3.0×10⁵ cells) were incubated with the indicated PC3 Exo concentrations (0, 30, 60 μg/mL at a concentration of 2.4×10⁹ vesicles/μg) for 24 and 48 hours and analyzed by IB for expression of β₆ integrin and actin (loading control).

Fig. 2. Transfer of GFP-tagged α_vβ₆ integrin to PBMC

(A), IB analysis for expression of GFP, CANX and exosomal marker CD9 in parental PC3, EGFP-PC3 (Mock) and β₆-EGFP-PC3 derived exosome lysate (Exo) and total cell lysate (TCL). (B), Left, Iodixanol gradient analysis of β₆-EGFP-PC3 derived Exo was performed as described in the Experimental Procedures. Expression of β₆, CD63, and CD81 analyzed by IB is shown. The expected density range for Exo is 1.11–1.14 g/mL. Right, NTA for the fifth fraction (density 1.14 g/mL) from the iodixanol gradient of β₆-EGFP-PC3 derived Exo is shown. (C), Left, human normal PBMC (2.0×10⁵ cells) incubated with 8 μg/mL of β₆-EGFP-PC3 Exo for 36 hours, were immuno-stained for CD14. Flow cytometric contour plots of cells gated as CD14⁺ monocytes are shown. Middle, flow cytometric analysis of comparative expression of GFP measured as mean fluorescence intensity (MFI) in PBMC incubated with parental PC3 or β₆-EGFP-PC3 Exo (8 μg/mL for 36 hours). Right, flow cytometric analysis of comparative expression of CD163 (M2 macrophage marker) measured as MFI in PBMC incubated with or without parental PC3 (8 μg/mL for 36 hours). Two graphs showing GFP and CD163 MFI include data from 4 biological replicates tested in 2 different experiments. **: P >0.01, student’s T test.

Fig. 3. α_vβ₆ integrin down-regulation in exosomes inhibits recipient monocyte M2 differentiation

Human PBMC (2.0×10⁵ cells) derived from two healthy donors were incubated for 48 hours with different concentrations of exosomes (Exo) (hi: 8 μg/mL; lo: 0.8 μg/mL) derived from PC3 cells incubated with no siRNA, a non-silencing siRNA (NS) or with one of two different siRNAs specific to β₆ mRNA (D1 and D2). Cultures were harvested at 48 hours for M2 monocyte polarization analysis by flow cytometry. (A), Contour plots depict live gating of CD14⁺ monocytes and representative expression of M2 polarization markers CD163 and CD204 in this population. Numbers indicate the percentage of gated cells. (B), There are 4 different PBMC treatment groups: untreated cells (-); NS, PBMC treated with exosomes from PC3 cells transfected with non-silencing siRNA; D1, PBMC treated with exosomes from PC3 cells transfected with a β₆-siRNA duplex designated D1; D2, PBMC treated with exosomes from PC3 cells transfected with a different β₆-siRNA designated D2. Mean fluorescence intensity (MFI). The percentages of CD14 gated CD163+/CD204+ cells in D1-hi, and D2-hi are statistically lower than NS, as assessed by Dunnett’s test. **: P > 0.01. (C), IB analysis of β₆ integrin expression in TCL and Exo derived from PC3 cells transfected with NS siRNA or with β₆ siRNA (D1 and D2). Flotilin 1 (FLOT1) was used as loading control for TCL and Exo, whereas CANX was used as loading control for TCL but not Exo.

Fig 4. Proteomics analysis of Exosomes from PC3 cells upon β₆ integrin knockdown reveals increased expression of STAT1 and MX1/2 proteins

(A) Results of a proteomics experiment comparing exosomes from non-treated cells (NT), cells treated with non-silencing siRNA (NS), or two β₆ integrin siRNAs (D1 and D2). Unsupervised hierarchical clustering based on label-free quantification of protein intensities is shown. Each horizontal bar is a unique protein and a total of 2,037 proteins are shown in this heat map. Red = higher, blue = lower protein abundance. (B) Heatmap of relative protein levels for the proteins most significantly affected by D1 and D2, detected with at least 10 MS/MS counts and 10 unique peptides.

Fig 5. β₆ integrin down-regulation results in increased expression of STAT1 and MX1/2 in PC3 cells and exosomes

(A), IB analysis of STAT1 expression in 30 μg of exosome lysate (Exo) and total cell lysate (TCL) derived from PC3 cells incubated either with a non-silencing siRNA (NS siRNA) or with β₆ mRNA directed siRNA (D1 or D2). ERK was used as loading control for TCL and TSG101 was used as loading control for Exo lysate. (B), IB analysis of the expression of exosomal markers CD81, CD9, CD63 in PC3 cell Exo lysate and TCL derived from PC3 cells incubated either with a NS siRNA or with D1 and D2 siRNA. CANX was used as loading control found in TCL but not in Exo. (C), IB analysis of the expression of β₆ integrin and MX1/2 in TCL and Exo derived from parental PC3 cells or PC3 cells transfected with shβ₆ or shβ5 retroviral constructs. PC3-shβ5 transfectants are used as a negative control. CANX was used as a loading control for TCL and CD9 was used as a loading control for Exo lysates.

Fig 6. Inhibition of αβ increases STAT1 levels in prostate cancer cells and Pten^pc-/- mice

(A), Evaluation of STAT1 and MX1/2 levels by IB in TCL from PC3 cells either untreated or treated with α_vβ₆ monoclonal antibody 6.3G9 or isotype control antibody 1E6 (both at 10 μg/mL). Actin was used as loading control. (B), Immunohistochemical analysis of STAT1 and α_vβ₆ expression in prostate tumors from Pten^pc-/- mice (sacrificed at 10-13 weeks) treated with 6.3G9 or 1E6 antibodies (10 mg/Kg/week × 5 weeks; n = 5). Representative IHC and H&E images are shown (Scale bar, 100 μm). Arrow, STAT1 expression in nuclei of prostate tumor cells. Arrowhead, nuclei of prostate tumor cells lacking STAT1 expression.

Fig 7. α_vβ₆ Integrin is expressed in PBMC from prostate cancer patients

(A), Flow cytometric analysis of α_vβ₆ expression in PBMC from prostate cancer patients and healthy subjects. Left, monocytes and lymphocytes are gated by SSC and FSC. Right, FACS analysis of α_vβ₆ cell surface expression in monocytes and lymphocytes from healthy subjects and prostate cancer patients respectively, utilizing 6.3G9 monoclonal antibody to α_vβ₆ and mouse IgG, as isotype control. Representative data are shown. (B), The schematic diagram shows that transfer of exosomal α_vβ₆ integrin from prostate cancer cells to monocytes results in down-regulation of STAT1 and MX1/2 levels and increased M2 polarization of monocytes, and has a pro-tumorigenic effect, whereas transfer of α_vβ₆ integrin–negative exosomes results in increased M1 polarization of monocytes, and inhibits cancer growth