Cyclocreatine Suppresses Creatine Metabolism and Impairs Prostate Cancer Progression

Abstract

Prostate cancer is the second most common cause of cancer mortality in men worldwide. Applying a novel genetically engineered mouse model (GEMM) of aggressive prostate cancer driven by deficiency of the tumor suppressors PTEN and Sprouty2 (SPRY2), we identified enhanced creatine metabolism as a central component of progressive disease. Creatine treatment was associated with enhanced cellular basal respiration in vitro and increased tumor cell proliferation in vivo. Stable isotope tracing revealed that intracellular levels of creatine in prostate cancer cells are predominantly dictated by exogenous availability rather than by de novo synthesis from arginine. Genetic silencing of creatine transporter SLC6A8 depleted intracellular creatine levels and reduced the colony-forming capacity of human prostate cancer cells. Accordingly, in vitro treatment of prostate cancer cells with cyclocreatine, a creatine analog, dramatically reduced intracellular levels of creatine and its derivatives phosphocreatine and creatinine and suppressed proliferation. Supplementation with cyclocreatine impaired cancer progression in the PTEN- and SPRY2-deficient prostate cancer GEMMs and in a xenograft liver metastasis model. Collectively, these results identify a metabolic vulnerability in prostate cancer and demonstrate a rational therapeutic strategy to exploit this vulnerability to impede tumor progression.

Significance: Enhanced creatine uptake drives prostate cancer progression and confers a metabolic vulnerability to treatment with the creatine analog cyclocreatine.

Figure 1.

Concomitant PTEN and SPRY2 deficiencies drive prostate cancer progression. A,PTEN and SPRY2 genomic alterations in patients with metastatic prostate cancer (SU2C/PCF Dream Team, PNAS 2019; ref. 41). B, Kaplan–Meier plot for overall survival of indicated mice Spry2^pc−/−, n = 5; Pten^pc−/−, n = 12; Pten^pc−/− Spry2^pc−/+, n = 8; Pten^pc−/− Spry2^pc−/−, n = 14 (*, P < 0.0001, Pten^pc−/− Spry2^pc−/− compared with Pten^pc−/−; P = 0.0021, Pten^pc−/− Spry2^pc−/− compared with Spry2^pc−/−; P = 0.0274, Pten^pc−/− Spry2^pc−/− compared with Pten^pc−/− Spry2^pc−/+; log-rank Mantel-Cox test). C, Noncystic prostate tumor weights from indicated mice at clinical endpoints (n = 7 mice for each group. Mean values ± SD are shown. *, P < 0.05 compared with Pten^pc−/−; one-way ANOVA with Tukey multiple comparison test. D, Representative hematoxylin and eosin (H&E) and Ki67 images of prostate tumor sections from Pten^pc−/−, Pten^pc−/− Spry2^pc−/+, and Pten^pc−/− Spry2^pc−/−mice (n = 5 mice for each group). Scale bar, 100 μm. E, IHC quantification of Ki67 in prostate tumors as indicated (n = 5 mice for each group). Mean values ± SD are shown. *, P < 0.05 compared with Pten^pc−/−; one-way ANOVA with Tukey multiple comparison test. F, Relative mRNA levels of Spry2 in primary murine prostate cancer cells as indicated. P1 derived from Pten^pc−/−; SP1 and SP2 derived from two independent Pten^pc−/− Spry2^pc−/+prostate tumors (n = 4 independent experiments). Mean values ± SD are shown. *, P < 0.05 compared with P1; one-way ANOVA with Dunnett multiple comparison test. G, Growth of indicated cells (n = 3 independent experiments). Mean values ± SEM are shown. *, P < 0.05 compared with P1; two-way ANOVA with Tukey multiple comparison test. H, Relative mRNA levels of SPRY2 in PC3 human prostate cancer cell lines as indicated. Nsi is PC3 cells with stable expression of nonsilencing vector control; CL1 and CL10 are PC3 clones with stable knockdown (KD) of SPRY2 (n = 3 independent experiments). Mean values ± SD are shown. *, P < 0.05 compared with Nsi; one-way ANOVA with Dunnett multiple comparison test. I, Soft agar colony quantifications of the indicated cells. Data shown are from a single representative experiment, performed three times with three technical replicates each. Mean values ± SD are shown. *, P < 0.05 compared with Nsi; one-way ANOVA with Dunnett multiple comparison test.

Figure 2.

Creatine availability affects proliferation of prostate cancer cells and is increased upon PTEN and SPRY2 deficiency. A and B, Relative cellular creatine levels in human (A) and murine prostate cancer cells (n = 3 independent experiments; B). Mean values ± SD are shown. *, P < 0.05 compared with Nsi and P1, respectively; unpaired t test. C, Intracellular abundance of creatine isotopes (M+0, M+1) in PC3 cells (Nsi, CL1, and CL10) after 24-hour incubation in the presence of unlabeled arginine or ¹³C₆-arginine. Data from single experiment with three technical replicates per condition. Mean values ± SD. D, Relative levels of creatine transporter Slc6a8 mRNA in murine prostate cancer cells. Data shown are from a single representative experiment, performed three times with three technical replicates each. Mean values ± SD are shown. *, P < 0.05 compared with P1; one-way ANOVA with Dunnett multiple comparison test. E, Relative levels of creatine transporter SLC6A8 in human prostate cancer cells (n = 3 independent experiments). Mean values ± SD are shown. *, P < 0.05 compared with Nsi; one-way ANOVA with Dunnett multiple comparison test. F, Relative levels of SLC6A8 mRNA in PC3 Nsi and SPRY2 KD CL10 with SLC6A8 KD (n = 3 independent experiments). Mean values ± SD are shown. *, P < 0.05 compared with Nsi CTRL and #, P < 0.05 compared with respective CTL cells; one-way ANOVA with Tukey multiple comparison test. G, Relative cellular creatine levels in PC3 Nsi and SPRY2 KD CL10 with SLC6A8 KD (n = 3 independent experiments). Mean values ± SD are shown *, P < 0.05 compared with Nsi CTRL and #, P < 0.05 compared with respective CTL cells; one-way ANOVA with Tukey multiple comparison test. H, Number of soft agar colonies of PC3 Nsi and SPRY2 KD CL10 with SLC6A8 KD (n = 3 independent experiments). Mean values ± SD are shown. *, P < 0.05 compared with Nsi CTRL and #, P < 0.05 compared with respective CTRL cells; one-way ANOVA with Dunnett multiple comparison test. I, IHC quantification of Ki67 in prostate tumors from Pten^pc−/− Spry2^pc−/−mice treated for 2 months with vehicle (VC) or 1% creatine (n = 5 mice for each group). Mean values ± SD are shown. *, P < 0.05; Mann–Whitney test. J, Representative hematoxylin and eosin (H&E) and Ki67 images of prostate tumor sections from Pten^pc−/− Spry2^pc−/−mice treated for 2 months with vehicle (VC) or 1% creatine. n = 5 mice for each group. Scale bar, 100 μm. K, Kaplan–Meier plot for progression‐free survival in MSKCC, Cancer Cell 2010 (24) prostate cancer dataset showing cases with high (more than z-score = 1.8; n = 6) or low (less than z-score = 1.8; n = 74) expression of SLC6A8; log‐rank Mantel-Cox test.

Figure 3.

Cyclocreatine suppresses creatine metabolism and proliferation in prostate cancer cells. A and B,In vitro growth of human PC3 (A) and murine (B) cells treated with control (CTRL) or 1% cyclocreatine (CC; n = 3 independent experiments). Mean values ± SD are shown. **, P < 0.005; ***, P < 0.0005 comparing cyclocreatine treatment with respective CTRL cells; unpaired t test. C, Soft agar colony quantifications of PC3 Nsi and SPRY2 KD clones (CL1 and CL10) treated with control or 1% cyclocreatine containing medium. Data shown from a single representative experiment performed twice, with three technical replicates each. Mean values ± SD are shown. *, P < 0.05 compared with Nsi CTRL and #, P < 0.05 compared with respective CTRL cells; one-way ANOVA with Tukey multiple comparison test. D–F, Relative levels of cellular creatine (D), phosphocreatine (E), and ATP in control (CTRL; F) or 1% cyclocreatine-treated PC3 Nsi and SPRY2 KD clones CL1 and CL10 (n = 3 independent experiments). Mean values ± SD are shown. *, P < 0.05 compared with Nsi CTRL and #, P < 0.05 compared with respective CTL cells; one-way ANOVA with Tukey multiple comparison test.

Figure 4.

Cyclocreatine suppresses creatine uptake and metabolism. A, Intracellular abundance of cyclocreatine in PC3 CL1 cells after 24 hours treatment with various concentrations of cyclocreatine and ¹³C-creatine. Data are from single experiment with three technical replicates per condition. Mean values ± SD are shown. B–D, Intracellular abundance of different isotopes of creatine (B), creatinine (C), and phosphocreatine (M+0, M+1; D) in PC3 CL1 cells after 24 hours treatment with various concentrations of cyclocreatine and ¹³C-creatine. Data are from single experiment with three technical replicates per condition. Mean values ± SD are shown. E–G, Oxygen consumption rate in PC3 Nsi (E), SPRY2 KD CL1 (F), and SPRY2 KD CL10 (G) treated as indicated (n = 4 independent experiments). Mean values ± SEM are shown. *, P < 0.05 compared with CTRL; two-way ANOVA with Tukey multiple comparison test.

Figure 5.

Cyclocreatine treatment affects prostate cancer progression in vivo.A, Representative Ki67 images of prostate tumor sections from Pten^pc−/− Spry2^pc−/−mice treated for 1 month with vehicle (VC) or 1% cyclocreatine (CC; n = 5 mice for each group). Scale bar, 100 μm. B, IHC quantification of Ki67 in prostate tumors from Pten^pc−/− Spry2^pc−/−mice treated for 2 months with vehicle or 1% cyclocreatine (n = 5 mice for each group). Mean values ± SD are shown. *, P < 0.05; Mann–Whitney test. C, Noncystic prostate tumor weights from Pten^pc−/− Spry2^pc−/−mice treated for 1 month with vehicle (n = 5) or 1% cyclocreatine (n = 6); mean values ± SD are shown. D, Cyclocreatine abundance detected in whole blood of Pten^pc−/− Spry2^pc−/−mice treated with 1% cyclocreatine (n = 6) or vehicle (n = 5). Data show mean ± SD. **, P < 0.01; Mann–Whitney test. E–J, Abundance of cyclocreatine (E), creatine (F), phosphocreatine (G), creatinine (H), arginine (I), and guanidinoacetate (J)detected in tumor tissues of Pten^pc−/− Spry2^pc−/−mice treated with 1% cyclocreatine (n = 6) or vehicle (n = 5). Data show mean ± SD. **, P < 0.01; Mann–Whitney test. K, Percentage PC3M liver metastases burden in CD-1 nude mice treated with control or 1% cyclocreatine water (ad libitum) for 1 month (n = 5 mice for each group). Mean values ± SEM are shown. *, P < 0.05; Mann–Whitney test. L, Representative images of PC3M liver metastases in CD-1 nude mice treated with control or 1% cyclocreatine water (ad libitum) for 1 month (n = 5 mice for each group). Arrows, liver metastases.