A novel direct activator of AMPK inhibits prostate cancer growth by blocking lipogenesis

Abstract

5'AMP-activated kinase (AMPK) constitutes a hub for cellular metabolic and growth control, thus representing an ideal therapeutic target for prostate cancers (PCas) characterized by increased lipogenesis and activation of mTORC1 pathway. However, whether AMPK activation itself is sufficient to block cancer cell growth remains to be determined. A small molecule screening was performed and identified MT 63-78, a specific and potent direct AMPK activator. Here, we show that direct activation of AMPK inhibits PCa cell growth in androgen sensitive and castration resistant PCa (CRPC) models, induces mitotic arrest, and apoptosis. In vivo, AMPK activation is sufficient to reduce PCa growth, whereas the allelic loss of its catalytic subunits fosters PCa development. Importantly, despite mTORC1 blockade, the suppression of de novo lipogenesis is the underpinning mechanism responsible for AMPK-mediated PCa growth inhibition, suggesting AMPK as a therapeutic target especially for lipogenesis-driven PCas. Finally, we demonstrate that MT 63-78 enhances the growth inhibitory effect of AR signaling inhibitors MDV3100 and abiraterone. This study thus provides a rationale for their combined use in CRPC treatment.

Figure 1

The novel small molecule MT 63–78 induces a direct activation of AMPK and prevents its dephosphorylation.

1. Molecular structure of MT 63–78 (MW = 326 Da).

2. Dose-dependent phosphorylation of GST-ACC peptide (1–150) by human recombinant AMPK α1β1γ1, after 30-min incubation with MT 63–78. AMP (25 μM) was used as positive control.

3. AMPK dephosphorylation assay, as described in Supplementary materials and methods. Recombinant AMPK α1β1γ1 (100 ng) was incubated with 100 ng of upstream kinase calcium/calmodulin-dependent protein kinase kinase-β (CaMKKβ). Phosphorylation of AMPK was then detected in the presence or absence of protein phosphatase 2C alpha (PP2Cα, 26 ng) and in the presence or absence of MT 63–78 (5 μM) using an antibody against the residue Thr172 on the α subunit.

Source data are available for this figure.

Figure 2

MT63–78 activates AMPK in PCa cells without altering the energy levels.

1. Dose-response activation of AMPK and phosphorylation of its direct targets Acetyl CoA Carboxylase (ACC) and Raptor in PCa cells.

2. AMPK activity assay as described in Supplementary materials and methods. Results are expressed as normalized average ± s.d. of three independent samples. One-way ANOVA test, followed by Dunnett's post hoc test for multiple comparisons was performed and adjusted P values were calculated (LNCaP: **P = 0.0074 MT 5 μM versus DMSO; ***P < 0.0001 MT 25 μM versus DMSO; ***P < 0.0001 MT 50 μM versus DMSO; ***P = 0.0002 AICAR versus DMSO. PC3: ***P = 0.00022 MT 5 μM versus DMSO; ***P < 0.0001 MT 25 μM versus DMSO; ***P < 0.0001 MT 50 μM versus DMSO; ***P < 0.0001 MT 100 μM versus DMSO; ***P = 0.000173 AICAR versus DMSO). RLU = relative luminescence units..

3. Measurment of ATP. 24-h incubation with metformin (Metf, 2.5 mM), 2-deoxyglucose (2-DG, 10 mM), and olygomycin (OM, 1 μM) was used as control. Results are expressed as means ± s.d. of three independent samples. One-way ANOVA test, followed by Dunnett's post hoc test for multiple comparisons was performed and adjusted P values were calculated (LNCaP: **P = 0.0093 Metf versus DMSO; ***P = 0.0002 2-DG versus DMSO; ***P = 0.0002 OM versus DMSO. PC3: ***P = 0.0002 2-DG versus DMSO).

4. Measurment of ADP. Twelve hour incubation 2-deoxyglucose (2-DG, 10 mM), and olygomycin (OM, 1 μM) was used as control. Results are expressed as means ± s.d. of three independent samples. One-way ANOVA test, followed by Dunnett's post hoc test for multiple comparisons was performed and adjusted P values were calculated (LNCaP: *P = 0.0172 2-DG versus DMSO; *P = 0.0208 OM versus DMSO).

5. Expression levels of phosphorylated ACC, Raptor, AMPK and their correspondent total forms in wild-type (wt) and AMPK α1^−/− and α2^−/− MEFs.

6. AMPK activity in LNCaP and PC3 cells transfected with AMPK β1 subunit siRNA, following treatment with MT 63–78 (MT) or DMSO. Results are expressed as normalized average ± s.d. of three independent samples. One-way ANOVA test, followed by Bonferroni post hoc test for multiple comparisons was performed and adjusted P values were calculated. (LNCaP scramble: ***P = 1.92E-13 MT versus DMSO; LNCaP β1 siRNA P = 0.1179 MT versus DMSO. PC3 scramble: ***P < 0.0001 MT versus DMSO; PC3 β1 siRNA: P = 0.2455 MT versus DMSO. n.s. = non significant. Western blot analysis shows the rate of AMPK β1 silencing.

Data information: Analyses in (A–F) were performed after 30-min incubation with MT 63–78, at indicated concentrations. Source data are available for this figure.

Figure 3

LKB1 is not essential for MT 63–78-mediated activation of AMPK.

1. LKB1 and CaMKKβ expression in PCa cells. HeLa cells were used as negative control.

2. Dose-response activation of AMPK, phosphorylation of ACC and Raptor in LKB1 null HeLa (left) and DU145 (right) cells, after 30-min incubation with MT 63–78, at indicated concentrations.

3. AMPK activity in HeLa cells treated with CaMKKβ inhibitor STO-609. HeLa cells were incubated with or w/o STO-609 (10 μg/ml) for 2.5 h prior to incubation with MT 63–78 (25 μM) for 30 min. Results are expressed as normalized average ± s.d. of three independent samples. One-way ANOVA test, followed by Bonferroni post hoc test for multiple comparisons was performed and adjusted P values were calculated (HeLa: ***P < 0.0001 MT versus DMSO; HeLa + STO-609: P = 0.0605 MT versus DMSO), n.s. = non significant. One sample for each condition was used for western blot analysis.

4. AMPK activity and western blot analysis in wt and LKB1^−/− MEFs treated as in C. Results are expressed as normalized average ± s.d. of three independent samples. One-way ANOVA test, followed by Bonferroni post hoc test for multiple comparisons was performed and adjusted P values were calculated (MEF LKB1wt + STO-609: ***P < 0.0001 MT versus DMSO; MEF LKB1^−/− + STO-609: P > 0.09999 MT versus DMSO). n.s. = non significant.

Source data are available for this figure.

Figure 4

MT63–78 inhibits PCa cell growth and soft agar colony formation.

1. Growth curves of LNCaP and PC3 cells treated with MT 63–78 and AICAR for 4 days, at indicated concentrations.

2. Relative growth of CL1, 22Rv1, C4-2, C4-2B cells, following 48-h treatment with MT 63–78. Results are expressed as percentage of cells compared to control (DMSO) ± s.d. of three independent samples. For CL1 experiments, six independent samples were used. One-way ANOVA test, followed by Dunnett's post hoc test for multiple comparisons was performed and adjusted P values were calculated (CL1: ***P = 0.00047 MT 10 μM versus DMSO; ***P < 1E-04 MT 25 μM versus DMSO; ***P < 1E-04 MT 50 μM versus DMSO. 22Rv1: *P = 0.0471 MT 10 μM versus DMSO; **P = 0.0021 MT 25 μM versus DMSO; ***P = 0.0004 MT 50 μM versus DMSO. C4-2: **P = 0.0011 MT 10 μM versus DMSO; ***P < 0.0001 MT 25 μM versus DMSO; ***P < 0.0001 MT 50 μM versus DMSO. C4-2B: ***P < 1E-08 MT 10 μM versus DMSO; ***P < 1E-08 MT 25 μM versus DMSO; ***P < 1E-08 MT 50 μM versus DMSO).

3. Relative growth of wt and AMPK α1^−/− and α2^−/− MEFs, following 3-day treatment with MT 63–78 (25 μM), AICAR (1 mM), and metformin (Metf, 2.5 mM). Results are expressed as percentage of cells compared to control (DMSO) ± s.d. of three independent samples. One-way ANOVA test, followed by Tukey's post hoc test for multiple comparisons was performed and adjusted P values were calculated. (MEFs wt: ***P = 5.91E-07 MT versus DMSO; ***P = 3.75E-06 AICAR versus DMSO; ***P = 6.14E-06 Metf versus DMSO. AMPK α1^−/− and α2^−/− MEFs: P = 0.1373 MT versus DMSO non-significant (n.s.); ***P = 8.79E-09 AICAR versus DMSO; ***P = 4.27E-08 Metf versus DMSO; ^###P = 1.41E-07 AICAR versus MT; ^###P = 1.37E-06 Metf versus MT.

4. Number of PC3 colonies in soft agar after 3-week treatment with 1 and 5 μM (positive control) MT 63–78. Results are expressed as number of colonies ± s.d. of three independent experiments. One-way ANOVA test, followed by Dunnett's post hoc test for multiple comparisons was performed and adjusted P values were calculated (*P = 0.027 MT 1 μM versus DMSO; **P = 0.0029 MT 5 μM versus DMSO). Photographs of colonies (10 ×  magnification) are shown.

5. Growth curve of PC3 cells treated with 1 μM MT-63–78 or vehicle for 9 days. Results are expressed as number of viable cells ± s.d. of three independent samples. Two way ANOVA test, followed by SIDAK's multiple comparison test was performed. No significant difference was observed.

Figure 5

MT63–78 induces mitotic arrest in PCa cells.

1. Cell cycle analysis in LNCaP and CRPC cells treated for 24 h with 25 μM MT 63–78 (MT). Percentage of cells in G1, S, and G2/M phases is indicated.

2. Phase contrast and immunofluorescence (Hoechst staining) images of LNCaP and CRPC cells, following 24-h treatment with MT 63–78 (25 μM). Magnification is indicated.

3. Western blot analysis of mitotic proteins in PCa cell lysates and conditioned media, following 24-h treatment with 25 μM MT 63–78.

4. Western blot analysis of DNA damage signaling in PCa cell lysates, following 24-h treatment with 25 μM MT 63–78. Treatment with 100 ng/ml Nocodazole (Noc) for 14 h was used as positive control.

Source data are available for this figure.

Figure 6

MT63–78 induces apoptosis in PCa cells.

1. Western blot analysis of apoptosis markers, pro-apoptotic, and anti-apoptotic proteins in PCa cells lysates, after 24-h treatment with MT 63–78.

2. Flow cytometry plots of Annexin V/PI assay in PCa cells, following 48-h treatment with MT 63–78. LNCaP treatment with AICAR was used as positive control. Percentage of apoptotic cells (early and late apoptosis) is indicated in red color. Percentage of necrotic cells is depicted in blue.

3. Western blot analysis of autophagic flux in LNCaP and PC3 cells, following treatement with MT 63–78 (25 μM). Lysosomal inhibitor Cloroquine (CQ, 10 μM) was added together with MT 63–78, whereas Bafilomycin A1 (Bafilo, 400 nM) was added 6 h before harvesting. Treatment with rapamycin (10 nM) was used as positive control. Contr+= 10  l of positive control for anti-LC3 antibody (PM036-PN).

Source data are available for this figure.

Figure 7

Inhibition of mTORC1 is only partially responsible for the anti-growth effects of MT 63–78.

1. Western blot analysis of lipogenic and mTORC1 pathways, after incubation with MT 63–78 for the indicated time points. ACC = Acetyl-CoA carboxylase, FASN = fatty acid synthase, S6RP = S6 ribosomal protein, 4EBP-1 = 4 element binding protein 1. AICAR (1 mM) was used as positive control.

2. Expression levels of precursor and mature forms of SREBP-1, following MT 63–78 (25 μM) treatment for the indicated time points.

3. Relative growth of LNCaP and PC3 cells after 3-day treatment with MT 63–78 (MT) or rapamycin (Rapa). Results are expressed as mean ± s.d. of three independent samples. One-way ANOVA test, followed by Tukey's post hoc test for multiple comparisons was performed and adjusted P values were calculated (LNCaP: ***P = 5.11E-08 MT versus DMSO; ***P = 1.45E-06 Rapa 1 nM versus DMSO; ***P = 2.47E-06 Rapa 100 nM versus DMSO; ^###P = 0.000152 MT versus Rapa 1 nM; ^###P = 0.000393 MT versus Rapa 100 nM. PC3: ***P = 1.63E-07 MT versus DMSO; ***P = 6.09E-06 Rapa 1 nM versus DMSO; ***P = 1.53E-06 Rapa 100 nM versus DMSO; ^###P = 0.000438 MT versus Rapa 1 nM; ^##P = 0.00763 MT versus Rapa 100 nM).

4. Expression levels of phosphorylated ACC, Raptor, AMPK, S6RP, and their total forms after 3-day treatment with MT 63–78 (25 μM) and rapamycin (Rapa).

5. Relative growth of LNCaP and PC3 cells transfected with empy vector (pKH3-EV) or with a construct containing constitutively active, rapamycin insensitive, S6 kinase 1 (pKH3-HA-S6K1 CA). Count was performed 3 days after incubation with MT 63–78 (25 μM) or rapamycin (1 nM). Drugs were added 12 h post-transfection. Results are expressed as a percentage of control (DMSO-treated cells) ± s.d. of three independent samples. One-way ANOVA test, followed by Tukey's post hoc test for multiple comparisons was performed and adjusted P values were calculated (LNCaP-pKH3-EV: ***P < 0.0001 MT versus DMSO; **P = 0.0089 Rapa versus DMSO. LNCaP-pKH3-HA-S6K1 CA: ***P = 0.0008 MT versus DMSO. PC3-pKH3-EV: ***P = 0.0003 MT versus DMSO; **P = 0.0088 Rapa versus DMSO. PC3-pKH3-HA-S6K1 CA: ***P = 0.0005 MT versus DMSO). n.s. = non significant.

6. Feedback activation of Akt and p42/p44 MAPK after 72-h incubation with MT 63–78 (25 μM) or rapamycin (Rapa).

Source data are available for this figure.

Figure 8

The anti-growth effects of MT 63–78 are mediated by inhibition of lipogenesis.

1. Incorporation of ¹⁴C-acetate into lipids, following treatment with MT 63–78 (MT). Results are expressed as a percentage of ¹⁴C-acetate incorporation into lipids compared to control, normalized to protein content. Mean values ± s.d. of three independent experiments are shown. One-way ANOVA test, followed by Dunnett's post hoc test for multiple comparisons was performed and adjusted P values were calculated (LNCaP-12 h: **P = 0.0018 MT 25 μM versus DMSO; ***P = 0.0004 MT 50 μM versus DMSO. LNCaP-24 h: *P = 0.0230 MT 25 μM versus DMSO; **P = 0.0014 MT 50 μM versus DMSO; LNCaP-48 h: ***P = 0.0003 MT 25 μM versus DMSO; ***P = 5.04E-05 MT 50 μM versus DMSO).

2. ¹⁴C-acetate incorporation in different lipid species. Data are expressed in cpm normalized to protein content. PL = Phospholipids, TG = Triacylglycerols, CE = Cholesterol esters, Chol = Cholesterol, DAG = Diacylglycerols.

3. Relative cell growth after 3-day treatment with MT 63–78 (MT) alone or in combination with 75 μM palmitate (Pal) and/or 100 μM mevalonate (Mev). Results are expressed as mean ± s.d. of five independent samples. One-way ANOVA test, followed by Tukey's post hoc test for multiple comparisons was performed and adjusted P values were calculated (***P = 2.4E-09 MT versus DMSO; ^###P = 0.000573 MT+Pal versus MT; ^###P = 6.96E-05 MT+Mev versus MT; ^###P = 4.97E-05 MT+Pal+Mev versus MT).

4. Relative cell growth after 3-day treatment with MT 63–78 (25 μM), tofa (10 μg/ml), simvastatin (Simv, 10 μM), and their combination. Results are expressed as mean ± s.d. of three independent samples. One-way ANOVA test, followed by Tukey's post hoc test for multiple comparisons was performed and adjusted P values were calculated (***P = 1.13E-08 MT versus DMSO; ***P = 2.26E-08 Tofa versus DMSO; ***P = 3.17E-08 Simv versus DMSO; ***P = 4.82E-09 Tofa+Simv versus DMSO).

5. Working model for MT 63–78. LKB1 = liver kinase B1, CaMKKβ = calcium/calmodulin-dependent protein kinase kinase-β, ACC = Acetyl-CoA carboxylase, HMG-CoA = 3-Hydroxy-3-methyl-glutaryl-CoA, PI3K = phosphatidylinositol 3-kinase, TSC1/TSC2 = tuberous sclerosis complex 1/2, Rheb = Ras homolog enriched in brain, mTORC1 =  mammalian target of rapamycin complex 1, 4EBP-1 = 4E-binding protein 1, S6K = S6 kinase, eiF4E = Eukaryotic translation initiation factor 4E, PL = Phospholipids, TG = Triacylglycerols. Color code: blue = activation; red = inhibition.

Figure 9

MT63–78 enhances the growth inhibitory effect of ADT.

1. Relative growth of LNCaP and CRPC cells C4-2 and 22Rv1, following 3-day treatment with MT 63–78 (MT), AR antagonist bicalutamide (Bic), and combined treatment (MT+Bic). Results are expressed as mean ± s.d. of three independent samples. One-way ANOVA test, followed by Tukey's post hoc test for multiple comparisons was performed and adjusted P values were calculated (**P < 0.01, ***P < 0.001 MT versus MT+Bic; ^##P < 0.01, ^###P < 0.001 Bic versus MT+Bic). Exact P values for each comparison are reported in Supplementary materials and methods. Expression levels of AR and PSA following single or combined treatments are shown under the bar graphs.

2. Relative growth of LNCaP and CRPC cells C4-2 and 22Rv1, following 3-day treatment with MT 63–78 (MT), AR antagonist MDV3100 (MDV), and combined treatment (MT+MDV). Results are expressed as mean ± s.d. of three independent samples. One-way ANOVA test, followed by Tukey's post hoc test for multiple comparisons was performed and adjusted P values were calculated (***P < 0.001 MT versus MT+MDV; ^###P < 0.001 MDV versus MT+MDV). Exact P values for each comparison are reported in Supplementary materials and methods. Expression levels of AR and PSA following single or combined treatments are shown under the bar graphs.

3. Relative growth of LNCaP and CRPC cells C4-2 and 22Rv1, following 3-day treatment with MT 63–78 (MT), CYP17A1 inhibitor abiraterone (Abi) and combined treatment (MT+Abi). Results are expressed as mean ± s.d. of three independent samples. One-way ANOVA test, followed by Tukey's post hoc test for multiple comparisons was performed and adjusted P values were calculated (**P < 0.01, ***P < 0.001 MT versus MT+Abi; ^###P < 0.001 Abi versus MT+Abi). Exact P values for each comparison are reported in Supplementary materials and methods. Expression levels of AR and PSA following single or combined treatments are shown under the bar graphs.

Figure 10

In vivo effects of AMPK activation.

1. Average tumor volume of LNCaP xenografts during 14-day treatment with MT 63–78 (30 mg/kg, i.p.). Experiments were performed twice for a total number of 17 mice under treatment and 15 controls. Results are expressed as n fold the mean initial volume (equal to 1) ± s.e.m. (*P = 0.049, end of treatment, Unpaired t-test).

2. Representative image of treated and control mice (with similar pre-treatment tumor volume) and their excised tumors at sacrifice. Correspondent western blot analysis of key proteins belonging to lipogenic and mTORC1 pathways is shown.

3. Densitometric analysis of phosphorylated ACC and Raptor in tumor homogenates of all treated mice (n = 16) and controls (n = 15). One treated mouse was not included in the analysis because frozen material was not available. Highlighted (black circle) are the two tumors with the highest ACC phosphorylation and volume reduction. Results are expressed in arbitrary units, after normalization to Vinculin. Unpaired t-test and Wilcoxon Mann–Whitney t-test were performed. Significant P values are reported on the graphs.

4. Average tumor volume of LNCaP xenografts during 21-day treatment with MT 63–78 (60 mg/kg, i.p., five mice) and vehicle (six mice). Results are expressed as n fold the mean initial volume (equal to 1) ± s.e.m. (**P = 0.004, end of treatment, Wilcoxon Mann–Whitney test). One mouse from treated group was removed from the study since biochemical analysis showed that the drug for unknown reasons did not penentrate in the tumor.

5. Densitometric analysis of phosphorylated ACC and Raptor performed as in C. Unpaired t-tests were performed. Significant P values are reported on the graphs.

6. Serum PSA levels measured at the end of the 21-day treatment with MT 63–78 (60 mg/kg, i.p.) in five treated and five control mice. Results are expressed as mean ± s.e.m. ELISA was performed twice (each sample in duplicate) with similar results. A representative experiment is shown. Serum from one control mouse was not available. Unpaired t-test was performed (P = 0.27).

7. PIN incidence in FASN-Tg α2^+/+ (n = 34), FASN-Tg/AMPK α2^+/− (n = 28), FASN-Tg/AMPK α2^−/− (n = 7) age-matched mice. Logistic regression P = 0.077, with post hoc calculated power of 0.44.

Source data are available for this figure.

Figure 11

¹¹C-acetate PET in LNCaP xenograft treated with AMPK activators.

1. ¹¹C-acetate micro-PET imaging before and after i.p. treatment with vehicle (5% hydroxypropyl beta-cyclodextrine) or AMPK activators MT 63–78 and AICAR. FASN inhibitor C75 was used as positive control. Fire LUT images of tumor-bearing mice, after 15 min of i.v. injection of ¹¹C-acetate, are shown. Arrows mark tumor location. Maximal and mean standardized uptake volumes (SUVs) are indicated.

2. ¹⁴C-acetate oxidation in LNCaP cells, following 6-h treatment with MT 63–78 (MT) or AICAR. ¹⁴C-CO₂ was measured in cpm and normalized to protein content, after background subtraction. Results are expressed as mean ± s.d. of three independent samples. One-way ANOVA test, followed by Dunnett's post hoc test for multiple comparisons was performed and adjusted P values were calculated (*P = 0.0109 MT 25 μM versus DMSO; **P = 0.0051 MT 50 μM versus DMSO; **P = 0.0016 AICAR versus DMSO).