Fig 2: In vitro and in vivo expression of ARG1 and ARG2 in PCa.PCa cell lines (LNCaP, 22Rv1, DU145 and PC3) were maintained in RPMI supplemented with 10% FBS. A) Gene expression of ARG1 (left panel) and ARG2 (right panel). Mean relative expression (n = 3) with standard error of the mean (error bars). B) Top panel: Arginase activity of PCa cell lines quantified in mU/mg of proteins. Bottom panel: Western blot of ARG1 and ARG2. Ran served as loading control. C) Representative image of immunohistochemistry staining of ARG2 expression in prostatic tissue. Note that the expression of ARG2 was confined to the epithelial cells with no stromal staining. D) Quantification of ARG2 expression by immunohistochemistry in prostate specimens. *Statistically significant difference in ARG2 expression between PIN and HR tissues (p = 0.033, Mann-U). **Statistically significant difference in ARG2 expression between tumor tissues and normal tissues (p<0.001, Mann-U), non-malignant tissues adjacent to tumor (p<0.01, Mann-U) and PIN tissues (p<0.001, Mann-U).

Fig 3: ARG1 and ARG2 are metabolically active.LNCaP cells were transfected with either a siCTRL or a cocktail of three siRNA against the ARG1 or ARG2. Post-transfection (24 hours), cells were plated in charcoal-stripped serum supplemented media for 72 hours and then stimulated for 72 hours with 10 nM R1881. A) siRNA inhibition of ARG1 and ARG2 expression was evaluated by Western Blot. Representative experiment is shown, (n = 4). B) Decreased arginase activity following transfection with siARG1 or siARG2 in LNCaP cells. The corresponding Western blot is shown in the bottom panel. Representative experiment is shown, (n = 3). C) Decreased metabolism of l-arginine in the absence of arginase expression. Conditioned media of LNCaP cells transfected with siCTRL, siARG1 or siARG2 were analyzed by HPLC for l-arginine concentration. The conditioned media analyzed by HPLC were from the LNCaP cells presented in Figure 4A. *Statistically significant difference (p<0.05, Mann-U). D) Decreased proliferation of LNCaP cells in the absence of arginase expression. LNCaP cells were transfected as previously described. Proliferation was measured by cell count 96 hours post-transfection. *Statistically significant difference (p<0.05, Mann-U), (n = 3). E) Inhibition of ARG2 expression causes increased PBMC proliferation and activation. PBMCs from normal donors were activated with anti-CD3 (OKT3, 1 µg/ml) with or without IL-2 in the presence of fresh media or conditioned media of LNCaP cells transfected with either control, siCTRL, siARG1 or siARG2 as previously described. Control PBMCs were incubated with IgG isotype control (1 µg/ml) and with PBS instead of IL-2. Left panel: PBMC proliferation was quantified by BrdU incorporation following 120 hours of OKT3 and IL-2 stimulation. Mean absorbance (n = 4) is shown with standard error (error bars). Right panel: PBMC secretion of IFN-? quantified by ELISA. Same experiment as previously described, but the PBMCs were activated for 24 hours without IL-2. Representative expression is shown (n = 4) with standard error of the mean (error bars).

Fig 4: BCKAs in cancer cell CM induce murine BMDM polarization. (A), Relative proliferation rate of BMDM grown in DMEM complete medium or indicated cancer cell CM. (B), Relative proliferation rate of BMDM grown in DMEM complete media or equivalent media containing 200 µM KIV or 200 µM KIC or 200 µM KMV (n = 3). (C), mRNA expression of Arg1, Mrc1, Mgl1 and Mgl2 in BMDM was measured by qPCR after incubation with AE17 CM for 24 hours. (D), ARG1 protein level was measured by western blot after incubation with AE17 CM for 24 hours. (E), mRNA expression of Arg1, Mrc1, Mgl1, Mgl2, Vegf, Il1ß, Il6, Nos2 and Slc16a1 was measured in BMDM after stimulation with DMEM (Control), 200 µM KIV, 200 µM KIC and 200 µM KMV, respectively for 24 hours. (F), Cytokines TGF-ß, TNF-a, IL6 and growth factor VEGF in BMDMs after the indicated treatments were measured by ELISA kits. (G), mRNA expression of Arg1 and Vegf was measured after pretreatment with the MCT1 inhibitor 250 nM AZD3965 for 1 hour, and stimulation with 200 µM KIV, 200 µM KIC and 200 µM KMV, respectively for 24 hours. Data show the mean ± SEM of n = 3 biological experiments. *p < 0.05, **p < 0.01, ***p < 0.001 (unpaired two-tailed t test).

Fig 5: Androgen-regulated expression of ARG1 and ARG2.A-B) LNCaP cells (left panels) and 22RV1 (right panels) were stimulated over a period of 72 hours with 10 nM R1881 following a 72 hour incubation period in charcoal-stripped media and the gene expression of A) ARG1 and B) ARG2 analyzed by qPCR. Control (light gray bars) and R1881-stimulated (black bars). *Statistically significant difference (p<0.05, Mann-U). Mean relative expression (n = 4) with standard error (error bars). C) Increased protein expression of both ARG1 and ARG2 following R1881 stimulation by Western blot. LNCaP cells were stimulated with 10 nM R1881 as previously described. PSA served as positive control. Representative experiment, (n = 6). D) Inhibition of AR activity with bicalutamide (Casodex). LNCaP cells were stimulated with R1881 for 72 hrs as previously described in the presence of increasing doses of bicalutamide (0, 10, 20 and 40 µM). ARG1 and ARG2 expression levels were evaluated by Western blot. Representative experiment is shown, (n = 3). Note the agonist effect of bicalutamide in the absence of R1881 illustrated by an increased PSA and ARG2 expression. E) Inhibition of AR expression by siRNA. LNCaP cells were transfected as previously described. AR, ARG1 and ARG2 expression levels were evaluated by Western blot. Representative experiment is shown, (n = 4). F) Arginase activity was quantified in mU/mg of proteins. LNCaP cells were transfected with siCTRL (light gray bars) or siAR (black bars) and then stimulated with R1881 for 72 hrs as previously described. Representative experiment, (n = 3).