https://orcid.org/0000-0002-7585-1913
Abstract
Prostate cancer (PCa) is one of the leading causes of cancer-related death among the male population worldwide. Unfortunately, current medical treatments fail to prevent PCa progression in a high percentage of cases; therefore, new therapeutic tools to tackle PCa are urgently needed. Biguanides and statins have emerged as antitumor agents for several endocrine-related cancers.
To evaluate: (1) the putative in vivo association between metformin and/or statins treatment and key tumor and clinical parameters and (2) the direct effects of different biguanides (metformin/buformin/phenformin), statins (atorvastatin/simvastatin/lovastatin), and their combination, on key functional endpoints and associated signalling mechanisms.
An exploratory/observational retrospective cohort of patients with PCa (n = 75) was analyzed. Moreover, normal and tumor prostate cells (normal [RWPE-cells/primary prostate cell cultures]; tumor [LNCaP/22RV1/PC3/DU145 cell lines]) were used to measure proliferation/migration/tumorsphere-formation/signalling pathways.
The combination of metformin+statins in vivo was associated to lower Gleason score and longer biochemical recurrence-free survival. Moreover, biguanides and statins exerted strong antitumor actions (ie, inhibition of proliferation/migration/tumorsphere formation) on PCa cells, and that their combination further decreased; in addition, these functional parameters compared with the individual treatments. These actions were mediated through modulation of key oncogenic and metabolic signalling pathways (ie, AR/mTOR/AMPK/AKT/ERK) and molecular mediators (MKI67/cMYC/androgen receptor/cell-cycle inhibitors).
Biguanides and statins significantly reduced tumor aggressiveness in PCa, with this effect being more potent (in vitro and in vivo) when both compounds are combined. Therefore, given the demonstrated clinical safety of biguanides and statins, our results suggest a potential therapeutic role of these compounds, especially their combination, for the treatment of PCa.
Prostate cancer (PCa) is the most common endocrine-dependent tumor among men in developed countries and represents one of the leading causes of cancer-related death in this population (1). Although PCa treatment has improved during the past decade, advanced stages of the disease are still difficult to manage. Indeed, the most aggressive phenotype of this pathology, named castration-resistant prostate cancer remains lethal nowadays (2). Therefore, the discovery and establishment of new therapeutic approaches to tackle PCa are urgently needed. An emerging approach against several tumor types is based in repositioning drugs already approved for other pathologies because they are faster and easier to translate to clinical practice than are new drugs. In this scenario, treatment of patients with drugs such as biguanides (eg, metformin) and statins (eg, simvastatin), commonly used to treat metabolism-related pathologies (ie, type 2 diabetes, obesity, and hypercholesterolemia, all of which have been linked to PCa progression (3-5)) has been associated to a lower cancer-specific death and even lower risk to develop different cancer types, including PCa (6-8). However, some of these actions are still controversial (9-13), and, more importantly, as far as we know, the potential direct antitumor effects that different biguanides (ie, metformin, buformin, and phenformin), statins (ie, atorvastatin, simvastatin, and lovastatin), and the simultaneous combination of both types of drugs, may exert on PCa cells (androgen-dependent and androgen-independent) and in normal prostate cells have not been yet compared side by side in previous reports.
The systemic mechanisms related to the actions exerted by biguanides (especially metformin) and statins are the modulation of the circulating levels of glucose and cholesterol, respectively (14, 15); however, the precise molecular mechanisms underlying the antitumor effects of metformin and statins are still controversial. For instance, both metformin and statins have been reported to exert antitumor effects on PCa cells in vitro (16, 17) through the modulation of different signalling pathways, which include the inhibition of mTOR pathway (both AMPK-dependent and AMPK-independent) (18), or the blockade of the mevalonate pathway (resulting from a specific blockade of HMG-CoA reductase) (19), respectively. But these studies are somehow limited, fragmentary, and unclear and, to the best of our knowledge, no studies have explored how metformin or statin alone, and especially in combination, can directly modulate the expression and/or activity of androgen receptor (a critical effector of PCa development and progression (20)) and other key signalling pathways and molecular mediators involved in tumor aggressiveness (eg, cyclin-dependent kinase inhibitors (CKIs) (21), etc.).
Based on all the information mentioned here, this study was devised to explore and compare side by side the direct antitumor effects of different biguanides (ie, metformin, buformin, and phenformin) and statins (ie, atorvastatin, simvastatin, and lovastatin) alone, as well as the combination of both types of drugs, on different PCa cells (androgen-dependent [LNCaP] and androgen-independent [22Rv1, PC-3, and DU145]) and in normal prostate cells (RWPE-1 and normal prostate primary cell cultures). In addition, we also analyzed the ability of metformin and statin (ie, simvastatin) to influence expression levels of critical effectors (ie, androgen-receptor, c-MYC, and CKIs) and the activity of different oncogenic signalling and metabolic pathways (ie, AR, AMPK, AKT, ERK, mTOR) linked to PCa aggressiveness. Moreover, we aimed to assess the putative in vivo association between metformin and/or statins treatment and key tumor and clinical parameters (Gleason score and biochemical recurrence, respectively) using a well-characterized cohort of PCa patients.
Materials and Methods
Human Samples
This study was approved by the Reina Sofia University Hospital Ethics Committee and was conducted in accordance with the principles of the Declaration of Helsinki. The biobank of the public health system of Andalusia (Cordoba Node) coordinated the collection, processing, management, and assignment of the biological samples used in this study, according to the standard procedures established for this purpose. Written informed consent was obtained from all patients. We carried out a cross-sectional pilot study designed to assess the putative in vivo association between metformin and/or statins treatment and key tumor and clinical parameters (Gleason score and biochemical recurrence, respectively) at the moment of diagnosis, which included patients (n = 75) who were diagnosed with PCa and immediately submitted to radical prostatectomies. Specifically, clinical data analyzed here have been obtained at the moment of PCa diagnosis and tumor samples were obtained during radical prostatectomies. Particularly, formalin-fixed, paraffin-embedded (FFPE) PCa tissues (n = 75) were obtained. In all cases, patients were diagnosed with clinically localized and hormone-naïve PCa, representing a relatively homogenous cohort of patients (Table 1), wherein other confounding factors such as androgen dependence or independence, modification of tumor behavior, disease duration, or previous antitumor treatments, could not be considered in this cohort of patients. Use of metformin and statins, as well as data of the time to biochemical recurrence (defined by 2 consecutive prostate-specific antigen values > 0.2 ng/mL and rising, after radical prostatectomy) and Gleason score (analyzed by specialist uro-pathologists following the modified 2005, 2010, and 2014 International Society of Urological Pathology criteria (22), based on the sample collection date) were collected from all patients included in this study. Significant PCa (SigPCa) was defined as Gleason score ≥7. Nonsignificant PCa (NonSigPCa) was defined as Gleason score = 6. Ten patients were being treated with other metabolic-related medication; however, there was a substantial variability in the consumption of these additional drugs, with a very low number of patients taking 1 of these additional medications (ie, 2 patients using insulin and only 1 patient taking other drugs such as sitagliptin, gliclazide, linagliptin, glimepiride, etc.).
Demographic, biochemical, and clinical parameters of patients with PCa
| Patients, n | 75 | |
| Age, y, median (IQR) | 61 (56-66) | |
| BMI, median (IQR) | 28.2 (26.8-31.5) | |
| PSA, ng/mL, median (IQR) | 5.1 (4.2-8.0) | |
| Recurrence, n (%) | 12 (16.0%) | |
| Time to recurrence, mo, median (IQR) | 23.5 (14.0-33.0) | |
| Gleason score, n (%) | 6 | 7 (9.33%) |
| 7 (3 + 4) | 42 (56%) | |
| 7 (4 + 3) | 21 (28%) | |
| 8 | 3 (4%) | |
| 9 | 2 (2.67%) | |
| Stage, n (%) | T2 | 24 (32.0%) |
| T3A | 49 (65.3%) | |
| T3B | 2 (2.7%) | |
| Treatment, n (%) | Untreated | 38 (50.7%) |
| Metformin | 7 (9.3%) | |
| Statins | 21 (28.0%) | |
| Both | 9 (12.0%) |
| Patients, n | 75 | |
| Age, y, median (IQR) | 61 (56-66) | |
| BMI, median (IQR) | 28.2 (26.8-31.5) | |
| PSA, ng/mL, median (IQR) | 5.1 (4.2-8.0) | |
| Recurrence, n (%) | 12 (16.0%) | |
| Time to recurrence, mo, median (IQR) | 23.5 (14.0-33.0) | |
| Gleason score, n (%) | 6 | 7 (9.33%) |
| 7 (3 + 4) | 42 (56%) | |
| 7 (4 + 3) | 21 (28%) | |
| 8 | 3 (4%) | |
| 9 | 2 (2.67%) | |
| Stage, n (%) | T2 | 24 (32.0%) |
| T3A | 49 (65.3%) | |
| T3B | 2 (2.7%) | |
| Treatment, n (%) | Untreated | 38 (50.7%) |
| Metformin | 7 (9.3%) | |
| Statins | 21 (28.0%) | |
| Both | 9 (12.0%) |
Abbreviations: BMI, body mass index; IQR, interquartile range; PSA, prostate-specific antigen.
Demographic, biochemical, and clinical parameters of patients with PCa
| Patients, n | 75 | |
| Age, y, median (IQR) | 61 (56-66) | |
| BMI, median (IQR) | 28.2 (26.8-31.5) | |
| PSA, ng/mL, median (IQR) | 5.1 (4.2-8.0) | |
| Recurrence, n (%) | 12 (16.0%) | |
| Time to recurrence, mo, median (IQR) | 23.5 (14.0-33.0) | |
| Gleason score, n (%) | 6 | 7 (9.33%) |
| 7 (3 + 4) | 42 (56%) | |
| 7 (4 + 3) | 21 (28%) | |
| 8 | 3 (4%) | |
| 9 | 2 (2.67%) | |
| Stage, n (%) | T2 | 24 (32.0%) |
| T3A | 49 (65.3%) | |
| T3B | 2 (2.7%) | |
| Treatment, n (%) | Untreated | 38 (50.7%) |
| Metformin | 7 (9.3%) | |
| Statins | 21 (28.0%) | |
| Both | 9 (12.0%) |
| Patients, n | 75 | |
| Age, y, median (IQR) | 61 (56-66) | |
| BMI, median (IQR) | 28.2 (26.8-31.5) | |
| PSA, ng/mL, median (IQR) | 5.1 (4.2-8.0) | |
| Recurrence, n (%) | 12 (16.0%) | |
| Time to recurrence, mo, median (IQR) | 23.5 (14.0-33.0) | |
| Gleason score, n (%) | 6 | 7 (9.33%) |
| 7 (3 + 4) | 42 (56%) | |
| 7 (4 + 3) | 21 (28%) | |
| 8 | 3 (4%) | |
| 9 | 2 (2.67%) | |
| Stage, n (%) | T2 | 24 (32.0%) |
| T3A | 49 (65.3%) | |
| T3B | 2 (2.7%) | |
| Treatment, n (%) | Untreated | 38 (50.7%) |
| Metformin | 7 (9.3%) | |
| Statins | 21 (28.0%) | |
| Both | 9 (12.0%) |
Abbreviations: BMI, body mass index; IQR, interquartile range; PSA, prostate-specific antigen.
Cell Culture and Reagents
Normal prostate (RWPE-1) and PCa (LNCaP and androgen-independent [22Rv1, PC-3, and DU145]) cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured according to manufacturer instructions, as previously described (23-25). These cell lines were validated by analysis of short tandem repeats sequences using GenePrint 10 System (Promega, Barcelona, Spain), and checked for mycoplasma contamination by PCR as previously reported (23-26). Biguanides (metformin [Merck KGaA, Darmstadt, Germany], buformin [Santa Cruz Biotechnology, Heidelberg, Germany], and phenformin [Merck KGaA]) and statins (atorvastatin [Merck KGaA], lovastatin [Merck KGaA], and simvastatin [Merck KGaA]) were used. The concentrations of the biguanides (metformin [5 mM], buformin [1 mM[, and phenformin [1 mM]) and statins (atorvastatin [10 µM], lovastatin [10 µM], and simvastatin [10 µM]) were selected based in studies previously reported by our group and others (11, 27-30).
Primary Cultures
Nontumor prostate tissues from cystoprostatectomies were dispersed into single cells as previously reported (23). Briefly, tissue was mechanically and enzymatically digested using the protocol described by Goldstein et al. (31), with modifications (32). One piece was included in paraffin to be analyzed by expert anatomopathologists to ensure the absence of tumor tissue and the other piece was immediately placed in sterile cold (4°C) DMEM High Glucose with D-valine (Seralab, Oviedo, Spain) and transported to the laboratory for dispersion and primary culture experiments.
Cell Proliferation and Cell Viability
Cell proliferation and cell viability were measured using Alamar Blue reagent (Bio-Source International, Camarillo, CA, USA) as previously reported (25). Briefly, cells were seeded in 96-well plates at a density of 3000 to 5000 cells/well and serum-starved for 24 hours. Then, cell proliferation/viability was evaluated every 24 hours using FlexStation III system (Molecular Devices, Sunnyvale, CA, USA) until 72 hours. Results were expressed as percentage referred to control.
Cell Migration
Cell migration was evaluated by wound-healing assay, as previously reported (23), in all cell lines except 22Rv1 because of its inability to migrate. Briefly, images of the scratch were taken at 0 and 12 hours and wound healing was calculated as the area observed 12 hours after the wound was made vs the area observed just after wounding. Results were expressed as percentage referred to control.
Tumorspheres Formation
Tumorspheres formation assay was carried out in representative cell-lines of LNCaP and hormone-refractory PCa (PC-3), as previously reported (23-33). Briefly, 2000 cells/well were seeded in Corning Costar 24-well ultra-low attachment plates (Merck KGaA) with DMEM F-12 medium supplemented with 20 ng/mL EGF (Merck KGaA). Treatments were added while plating the cells and refreshed every 3 days. The number and the size of tumorspheres was determined after 14 days of incubation and analyzed with ImageJ software. Results were expressed as percentage of tumorspheres number referred to control.
RNA Extraction, Retrotranscription, and Quantitative PCR
Total RNA from FFPE samples was isolated and DNase-treated using the Maxwell 16 LEVRNA FFPE Kit (Promega, Madison, WI, USA) according to manufacturer instructions in the Maxwell MDx 16 Instrument (Promega). Additionally, total RNA was extracted from fresh samples using the Qiagen AllPrep DNA/RNA/Protein Mini Kit (Thermo Fisher Scientific, Waltham, MA, USA), and from PCa cell lines using TRIzol Reagent (Thermo Fisher Scientific), followed, in both cases, by DNase treatment using the Qiagen RNase-Free DNase Kit (Thermo Fisher Scientific). RNA from PCa cells was isolated after 24 hours of incubation with metformin, simvastatin, and combined treatment. Total RNA concentration and purity was assessed using the Nanodrop One Spectrophotometer (Thermo Fisher Scientific). Total RNA was retrotranscribed using random hexamer primers and the cDNA First Strand Synthesis kit (Thermo Fisher Scientific). Details regarding the development, validation, and application of the quantitative RT-PCR to measure expression levels of the transcripts of interest have been previously reported by our laboratory (23, 34, 35). Specific and validated primer sets used to measure the expression levels of genes of interest in this study are described in Jiménez-Vacas et al. (36). Expression levels were adjusted with a normalization factor calculated from the expression levels of ACTB and GAPDH housekeeping genes, using Genorm 3.3 (37), wherein the expression levels of the housekeeping genes did not differ among groups (data not shown).
Western Blot Analysis
Proteins from whole cell lysates were extracted after 12 hours of incubation with metformin, simvastatin, and combined treatment, separated on 10% SDS-polyacrylamide gels and transferred onto nitrocellulose membranes, as previously described (23, 38). Membranes were incubated overnight at 4°C with specific primary antibodies against proteins of interest, and then for 1 hour with the corresponding horseradish peroxidase-linked secondary antibody. Primary and secondary antibodies used in this study are summarized in Jiménez-Vacas et al. (36). Immunoreactive bands were detected using ECL chemiluminescence substrate solution (GE Healthcare, Madrid, Spain) in an enhanced chemiluminescence detection system (GE Healthcare). Observed bands were quantified using ImageJ software and results were expressed as percentage of control.
Statistical analyses
Statistical differences between 2 variables were calculated by unpaired parametric t test and nonparametric Mann-Whitney U test, according to normality, assessed by Kolmogorov-Smirnov test. For differences among more than 2 variables, 1-way ANOVA analysis was performed (followed by a Tukey post hoc analysis). Chi-squared tests and Fisher’s exact tests were performed to determine associations between treatment groups and Gleason score (NonSigPCa and SigPCa). Significant association between treatments and biochemical PCa recurrence was studied using the long-rank P-value method. All the statistical analyses were assessed using GraphPad Prism 7 (GraphPad Software Inc, La Jolla, CA, USA). All the experiments were performed in at least 3 experiments (n ≥ 3). Statistical significance was considered when P < 0.05.
Results
The Combination of Metformin and Simvastatin Improves Clinical Parameters of PCa Aggressiveness
To study the association between the treatments with metformin, any statin, or the combined treatment of both drugs with clinical parameters of PCa aggressiveness (ie, time to biochemical recurrence and Gleason score), an available pilot cohort of 75 patients was analyzed. Gleason score (not categorized) was not different among untreated patients, treated with metformin or statins alone and cotreated with both drugs (data not shown). However, when tumors were categorized in SigPCa and NonSigPCa, statistically significant differences were found among patients that were untreated, treated with metformin, treated with statins, and cotreated with both drugs (P = 0.028; Fig. 1a). Specifically, cotreatment with metformin and any statin was associated to higher probability to develop NonSigPCa compared with: (1) untreated patients + patients treated with metformin or statins alone (P = 0.033; Fig. 1b), or (2) patients treated with metformin or statins alone (P = 0.011; Fig. 1c). Moreover, treatment with metformin, any statin, or the cotreatment of both drugs was not significantly associated with biochemical recurrence when comparing each group individually (36). However, patients cotreated with metformin and any statin tended to be associated (P = 0.13) with longer free biochemical recurrence survival compared with the group of patients without treatment, treated with metformin, and treated with any statin (Fig. 1d). Additionally, patients cotreated with both drugs also tended to be associated to longer free biochemical recurrence survival (P = 0.08) compared with those individually treated with metformin or with any statin (Fig. 1e).
Effects of metformin, statins, and combined treatment in patients with PCa. (A) Percentage of patients untreated (n = 38), treated with metformin (n = 7), statins (n = 21), and both drugs (n = 9) with NonSigPCa and SigPCa. (B) Percentage of NonSigPCa and SigPCa in untreated patients and treated with metformin or statins alone (n = 66) and patients cotreated with metformin and statins (n = 9). (C) Percentage of NonSigPCa and SigPCa in patients treated with metformin or statins alone (n = 28) and patients cotreated with metformin and statins (n = 9). Kaplan-Meier survival curve of biochemical recurrence-free survival of PCa patients untreated (n = 11) vs treated with metformin (n = 5) vs treated with simvastatin (n = 22) vs treated with both treatments (n = 9). (D) Kaplan-Meier survival curve of biochemical recurrence-free survival of PCa untreated patients and treated with metformin or statins alone (n = 62) and patients cotreated with metformin and statins (n = 9). (E) Kaplan-Meier survival curve of biochemical recurrence-free survival of PCa patients treated with metformin or statins alone (n = 24) and patients cotreated with metformin and statins (n = 9). Patients with coadjuvant treatment were excluded from the Kaplan-Meier analysis (n = 4). Asterisks indicate values that significantly differ from untreated patients (*P < 0.05). NonSigPCa, nonsignificant PCa (defined as Gleason score = 6); PCa, prostate cancer; SigPCa, significant PCa (defined as Gleason score ≥ 7).
Biguanides Exert Antiproliferative Effects on Prostate Cells
In in vitro assays, the 3 biguanides tested, metformin, buformin, and phenformin, reduced the proliferation rate of all the PCa cell lines assayed in a time-dependent manner (Fig. 2a). Specifically, metformin effect was evidenced at 48 and 72 hours of incubation, whereas buformin progressively decreased the proliferation rate of LNCaP cells at 24, 48, and 72 hours of incubation, and in 22Rv1, PC-3, and DU145 cells at 48 and 72 hours of incubation (Fig. 2a). Phenformin reduced cell proliferation in all tested PCa cell lines at 24, 48, and 72 hours of incubation (Fig. 2a). The antiproliferative effect exerted by phenformin was significantly more pronounced than that exerted by metformin and buformin at 24, 48, and 72 hours of incubation in LNCaP cells, as well as at 48 and 72 hours of incubation in 22Rv1 and PC-3 cells (Fig. 2a).
Effect of biguanides (metformin, buformin, and phenformin) on cell proliferation/viability of prostate-derived cells. Cell proliferation was assessed by Alamar Blue after 24, 48, and 72 hours of treatment in (A) PCa cell lines (LNCaP, 22Rv1, PC-3, and DU145) and (B) normal prostate-like cell line RWPE-1. (C) Cell viability was assessed by Alamar Blue after 24, 48, and 72 hours of treatment in primary normal prostate cells. The data are expressed as percentage of proliferation rate compared with vehicle-treated control cells (set at 100%). Values represent the mean ± SEM. Asterisks (comparison to control), “a” (comparison to metformin), and “b” (comparison to buformin) (*P < 0.05; **P < 0.01; ***P < 0.001) indicate statistically significant differences. PCa, prostate cancer.
In normal prostate RWPE-1 cells, phenformin increasingly reduced proliferation rate at 24, 48, and 72 hours, whereas metformin and buformin reduced this parameter at 48 and 72 hours (Fig. 2b). The antiproliferative effect of phenformin on RWPE-1 cells was significantly more pronounced than that exerted by metformin and buformin at 48 and 72 hours of incubation (Fig. 2b). Importantly, the use of primary nontumor prostate cell cultures revealed that buformin and phenformin, but not metformin, were able to significantly reduce the viability rate in these primary cells, being the effect of phenformin significantly higher than that exerted by buformin at 24, 48, and 72 hours (Fig. 2c).
Statins Exert Antiproliferative Effects on Prostate Cells
In the case of statins, atorvastatin also reduced time dependently the proliferation rate in LNCaP, 22Rv1, and PC-3 cells at 48 and 72 hours (Fig. 3a). Similarly, lovastatin decreased cell proliferation at 24, 48, and 72 hours in LNCaP, at 72 hours in 22Rv1, and at 48 and 72 hours in PC-3 cells (Fig. 3a). Likewise, simvastatin reduced at 24, 48, and 72 hours, in a time-dependent manner, the proliferation rate of LNCaP, 22Rv1, and PC-3 cells (Fig. 3a). In contrast, no changes were observed in response to any of the statins in DU145 cells. The antiproliferative effect of simvastatin was significantly higher than (1) the effect exerted by atorvastatin and lovastatin at 24, 48, and 72 hours of incubation in LNCaP cells, at 72 hours in 22Rv1 and PC-3 cells, and (2) the effect exerted by lovastatin at 48 hours in 22Rv1 cells (Fig. 3a).
Effect of statins (atorvastatin, lovastatin, and simvastatin) on cell proliferation/viability of prostate-derived cells. Cell proliferation was assessed by Alamar Blue after 24, 48, and 72 hours of treatment in (A) PCa cell lines (LNCaP, 22Rv1, PC-3, and DU145) and (B) normal prostate-like cell line RWPE-1. (C) Cell viability was assessed by Alamar Blue after 24, 48, and 72 hours of treatment in primary normal prostate cells. The data are expressed as percentage of proliferation rate compared with untreated control cells (set at 100%). Values represent the mean ± SEM. Asterisks (comparison to control), “a” (comparison to atorvastatin), and “b” (comparison to simvastatin) (*P < 0.05; **P < 0.01; ***P < 0.001) indicate statistically significant differences. PCa, prostate cancer.
In normal prostate RWPE-1 cells, atorvastatin, lovastatin, and simvastatin decreased the proliferation rate at 48 and 72 hours of incubation, with this effect being less pronounced in the case of simvastatin compared with that of atorvastatin at 72 hours of incubation (Fig. 3b). Subsequently, lovastatin and simvastatin reduced the viability rate of nontumor primary prostate cultures at 48 and 72 hours compared with vehicle-treated cells (Fig. 3c), whereas atorvastatin only decreased this functional parameter significantly at 72 hours (Fig. 3c).
The Combined Treatment With Biguanides and Statins Exerts Additive Antiproliferative Effects in Normal Prostate and PCa Cells
First, we evaluated the direct antiproliferative effects of metformin and simvastatin alone or in combination in different PCa cell lines. We found that the reduction in the proliferation rate of LNCaP, 22Rv1, PC-3, and DU145 cells was more pronounced in response to the combination of metformin and simvastatin compared with each individual treatment after 48 and 72 hours (Fig. 4a). Similarly, the combination of metformin and atorvastatin also exerted an additive antiproliferative effect in all the PCa cell lines tested at 48 and 72 hours, and also at 24 hours in the case of LNCaP and PC-3 cells (36). Likewise, the combination of metformin and lovastatin also evoked an additive effect at 48 and 72 hours in 22Rv1, PC-3, and DU145 cells (36). The effect of the combination of metformin and lovastatin was also seemingly more pronounced than the effect of metformin and lovastatin alone at 48 and 72 hours in LNCaP cells, but this difference only reached statistical significance when comparing with the treatment of lovastatin, but not metformin alone (36).
Effects of metformin, simvastatin, and its combination on functional parameters of PCa cells. (A) Proliferation rate of PCa cell lines (LNCaP, 22Rv1, PC-3, and DU145) and (B) normal prostate-like cell line RWPE-1 after 24, 48, and 72 hours of metformin (5 mM), simvastatin (10 µM), and combined treatment. The data are expressed as percentage of proliferation rate compared with untreated control cells (set at 100%). (C) Migration rate of PCa cell lines (LNCaP, PC-3, and DU145) and (D) normal prostate-like cell line RWPE-1 after 12 hours of metformin (5 mM), simvastatin (10 µM), and combined treatment. The data are expressed as percentage of migration rate compared to untreated control cells (set at 100%). Representative images of migration assay in PC-3 are shown (similar representative images for LNCaP, DU145, and RWPE-1 cells are represented in Jiménez et al. (36)). (E) Number of tumorspheres produced by PCa cell lines (LNCaP and PC-3) in response to metformin (5 mM), simvastatin (1 mM), and combined treatment after 14 days of incubation (treatment refreshed every 3 days). The data are expressed as percentage of number of tumorspheres compared with untreated control cells (set at 100%). Representative images of tumorspheres assay in PC-3 cells are shown (similar representative images for LNCaP cells are represented in Jiménez-Vacas et al. (36)). Values represent the mean ± SEM. Asterisks (comparison to control), “a” (comparison to metformin), and “b” (comparison to simvastatin) (*P < 0.05; **P < 0.01; ***P < 0.001) indicate statistically significant differences. CMB, combined treatment; MF, metformin; PCa, prostate cancer; SMV, simvastatin.
Additionally, we also analyzed the antiproliferative effects of the combination of buformin or phenformin with all the statins (simvastatin, atorvastatin, or lovastatin) and compared with the corresponding individual treatments (36). Similar to that found with the combination of metformin and simvastatin in PCa cells, our results revealed that the combination of buformin or phenformin with different statins resulted in apparently more pronounced/additive antiproliferative effects compared with the individual treatments; however, the strongest additive effects were observed when combining metformin and simvastatin (results in Fig. 4a and in Jiménez-Vacas et al. (36)).
In normal prostate RWPE-1 cells, incubation with metformin and simvastatin also decreased, in an additive manner, cell proliferation at 48 and 72 hours (Fig. 4b). In contrast, although the combination of metformin with atorvastatin or lovastatin also reduced cell proliferation in RWPE-1 (36), these effects did not differ from those exerted by atorvastatin or lovastatin (36) alone. Similarly, the combination of buformin or phenformin with all the statins (simvastatin, atorvastatin, or lovastatin) rendered comparable results (36).
Effects of Metformin, Simvastatin, or Their Combination on Additional Functional Parameters of Aggressiveness in Normal Prostate and PCa Cells
Based on the in vitro results obtained (Figs. 2 and 3), we chose metformin (biguanide with the lowest effect on the viability of normal primary prostate cell cultures, but with strong effects of PCa cells, and the only US Food and Drug Administration-approved biguanide) and simvastatin (the statin with the strongest antitumor effect on different PCa cells) to further evaluate the antitumor effects of the combination of both drug types and the underlying mechanisms of action in PCa cells.
Treatment with metformin and simvastatin significantly decreased the migration capacity of PCa cells (LNCaP, PC-3; Fig. 4c) and normal prostate cells (RWPE-1; Fig. 4d). Similar results were found in response to metformin, but not simvastatin, in DU145 cells (Fig. 4c). The combined treatment of metformin and simvastatin reduced in an additive manner the migration rate in all these cell lines compared with the individual treatments (36). Moreover, metformin and simvastatin, alone and in combination, significantly decreased tumorspheres formation in LNCaP and PC-3 cells, with this effect being significantly more pronounced (ie, additive) in response to the combined treatment in PC-3 cells [Fig. 4e; (36)]. In the case of LNCaP cells, the combined treatment of metformin and simvastatin completely blocked the formation of tumorspheres (Fig. 4e; (36)). On the other hand, metformin, simvastatin, or their combined treatment did not alter the size of the tumorspheres formed by LNCaP or PC-3 cells (36).
Metformin and/or Simvastatin Treatment Modulates Androgen Receptor Expression/Activity and Other Key Oncogenic and Metabolic Pathways Involved in PCa Aggressiveness
Based on the overall results reported here, we chose LNCaP cells to explore the effects of metformin and simvastatin, alone or in combination, on the expression and/or activity of well-known key effectors/elements associated to aggressiveness of PCa that might be linked to the antitumor effects observed in this study. Remarkably, our results revealed that the treatment with metformin, simvastatin, and their combination reduced the expression levels of androgen receptor (AR) in LNCaP, as well as its activity (measured by the ratio p-ARSer213/total AR ratio) (Fig. 5a/b).
Molecular consequences of treatment with metformin, simvastatin, and its combination in LNCaP cells. (A) Expression levels of AR, (B) protein levels of p-ARSer213, (C) expression levels of MKI67, (D) c-MYC, (E) IGF1R, (F) IGF1, (G) INSR, (H) INS, (I) protein levels of p-AMPKThr172, (J) p-AKTSer473, (K) p-ERKThr202, and (L) p-mTORSer448 after treatment with metformin (5 mM), simvastatin (10 µM) and the combined treatment. Cells were treated for 24 hours and 12 hours for protein and RNA isolation, respectively. mRNA expression levels were adjusted by NF (calculated by ACTB and GAPDH expression). Phospho-protein levels were adjusted by protein levels of total target protein (AR, AMPK, AKT, ERK1/2, and mTOR, respectively). Representative images of western blot results are shown. Data are expressed as percentage of vehicle-treated cells (set at 100%). Values represent the mean ± SEM. Asterisks (comparison to control), “a” (comparison to metformin), and “b” (comparison to simvastatin) (*P < 0.05; **P < 0.01; ***P < 0.001) indicate statistically significant differences. CMB, combined treatment; MF, metformin; NF, Normalization factor; SMV, simvastatin.
In addition, metformin, simvastatin, and their combination also decreased the expression of MKI67 (Fig. 5c), whereas metformin, but not simvastatin, significantly reduced the expression levels of c-MYC compared with vehicle-treated cells (Fig. 5d). Moreover, IGF1R expression was reduced in response to metformin, simvastatin, and the combined treatment (Fig. 5e). However, only metformin, but not simvastatin or the combined treatment, reduced the expression levels of IGF1 and INSR (Fig. 5f/g). On the other hand, INS expression levels were not altered in response to metformin, simvastatin, nor the combined treatment (Fig. 5h). As expected, metformin treatment, and also its combination with simvastatin, but not the latter alone, increased the phosphorylation levels of AMPKα (p-AMPKα Thr172/total AMPKα ratio, which is a surrogate marker of AMPK activation; Fig. 5e). In contrast, simvastatin treatment, alone or combined with metformin, but not metformin alone, decreased the activity of AKT and ERK signalling pathways (p-AKTSer473/total AKT ratio and p-ERKThr202/total ERK ratio, respectively) compared with vehicle-treated control cells (Fig. 5f/g). Interestingly, although treatment with metformin or simvastatin alone did not significantly alter the activity of mTOR (p-mTORSer2448/total mTOR ratio), the combined treatment of both drugs significantly reduced its activity (Fig. 5h).
Metformin and/or Simvastatin Treatment Modulates the Expression of Cell-Cycle Inhibitors
Next, we interrogated the effect of metformin, simvastatin, and their combined treatment on the expression of critical cell-cycle inhibitors. Remarkably, although metformin or simvastatin alone did not significantly alter the expression of CDKN1A, CDKN1B, CDKN2A, or CDKN2D, the combination of metformin and simvastatin drastically increased the expression of all these cell-cycle inhibitors in PCa cells in vitro (Fig. 6a). Notably, these in vitro data nicely resembled the results observed in samples derived from human samples subject in vivo to similar treatments, in which the expression of CDKN1A, CDKN1B, CDKN2A, or CDKN2D was elevated in PCa samples from patients treated with both metformin and statins, but not with the drugs alone, being these differences statistically significance for both CDKN1B and CDKN2A (Fig. 6b).
Effects of metformin, simvastatin, and combined treatment on expression levels of cyclin-dependent kinase inhibitors. Expression levels of (A) CDKN1A, (B) CDKN1B, (C) CDKN2A, and (D) CDKN2D after 24 hours of incubation with metformin (5 mM), simvastatin (10 µM), and combined treatment in LNCaP cells. Expression levels of (E) CDKN1A, (F) CDKN1B, (G) CDKN2A, and (H) CDKN2D in PCa samples taken from patients treated with metformin, simvastatin, or both treatments. mRNA expression levels were adjusted by NF (calculated by ACTB and GAPDH expression). Data are expressed as percentage of untreated cells or untreated patients (set at 100%). Values represent the mean ± SEM. Asterisks (comparison to control), “a” (comparison to metformin) and “b” (comparison to simvastatin) (*P < 0.05; **P < 0.01; ***P < 0.001) indicate statistically significant differences calculated by ANOVA. “pt” indicates P value (CMB vs control) calculated by t-test. CMB, combined treatment; MF, metformin; NF, Normalization factor; SMV, simvastatin.
Discussion
PCa remains an unmet clinical problem worldwide because of its high prevalence and mortality (1). Unfortunately, the therapeutic approaches currently available in clinical practice (ie, surgery, hormone therapy, radiotherapy, and chemotherapy) ultimately fail in a high proportion of patients. As a result, quality of life is detrimentally affected in patients who are often diagnosed at advanced stages, and the associated health care costs are significantly elevated. Therefore, identification of alternative therapeutic options deems necessary to properly confront PCa. In this sense, earlier studies have suggested that drugs such as biguanides and statins exert antitumor actions in different endocrine-related cancers, including PCa (6-8, 11, 39-42). However, our current understanding of the direct effects that different types of biguanides and statins exert on normal prostate and PCa cells is quite scarce, partial, and unclear. To the best of our knowledge, the present report provides the first comprehensive analysis of the direct side-by-side comparison of the antitumor effects that different biguanides (ie, metformin, buformin, and phenformin), statins (ie, atorvastatin, simvastatin, and lovastatin), and the combination of both classes of compounds, exert on PCa cells (androgen dependent and androgen independent) and in normal prostate cells. In addition, we explored their possible underlying mechanisms as well as the relation of these treatments to key clinical parameters of the patients in vivo.
First, we performed an observational retrospective study to assess the key clinical outcomes of patients with PCa treated (or not) with metformin or any type of statin, alone or in combination. Our results provide the first evidence indicating that patients treated with the combination of metformin and statins have less aggressive PCa disease (NonSigPCa) as compared with (1) those untreated or (2) treated only with metformin or any statin, and that this cotreatment was also associated with longer free biochemical recurrence survival compared with those treated only with metformin or statins. To the best of our knowledge, only 3 studies have aimed to examine the effects that the combination of metformin and statins caused on PCa outcomes (43-45). Specifically, 1 study showed that this combination leads to synergistic effects to lower risk of biochemical recurrence in patients with PCa after radical prostatectomy (43), whereas another showed negative results (44). Most recently, in a large epidemiological study, Li et al. observed that the use of statins alone or in combination with metformin might be associated with lower PCa mortality among high-risk patients, particularly in postdiagnostic settings (45); however, the differences between the treatment using statin alone and the combination of both drugs were not statistically different, suggesting that these effects were mainly ascribed to the use of statins. Therefore, although a well-designed clinical trial is warranted to investigate the real effects of metformin or statins alone and the combination of metformin and statins to reduce key clinical aggressiveness features linked to PCa, our data strongly suggest that the combination of metformin and statins might be associated to more beneficial effects in key clinical parameters of aggressiveness of patients with PCa (ie, lower Gleason score, longer free biochemical recurrence survival) than the individual treatment of both drugs. The interpretation of our data should be taken with caution because they are derived from a cross-sectional retrospective study design analyzing a limited cohort of patients (n = 75). These results are supported by 2 previous reports using preclinical models indicating that the combination of metformin and atorvastatin caused stronger inhibition on tumor growth (immunodeficient mice bearing PC-3 xenograft tumors (46), and on tumor formation, bone metastasis and biochemical failure [C4-2B4 orthotopic NCr-nu/nu mice (47)]) than either drug used individually.
The antitumor clinical associations previously observed in vivo were further supported by the in vitro data included here because we provide solid evidence indicating that the treatment with different biguanides and statins significantly decrease the aggressiveness features of PCa cells (ie, proliferation rate, migration capacity, and tumorspheres formation), and that the combinations of different biguanides and statins strongly inhibit (ie, in an additive manner) these antitumor effects compared with the individual treatment of both drug types. In line with a previous report (48), DU145 cells were found to be resistant to statin treatment. Specifically, and based on the results reported here, cotreatment with metformin and simvastatin might represent the best combination because metformin did not reduce the viability rate of nontumor primary cultures and simvastatin exerted higher antiproliferative effects compared with atorvastatin and lovastatin in all PCa cells tested. In line with the present results observed in PCa cells, a similar pattern of response to biguanides, statins, and the combination of both in terms of reduction of aggressiveness features has been found in other endocrine-related cancers (ie, endometrial cancer, pituitary and neuroendocrine tumors) (11, 40-42, 49), suggesting that the utility of these drugs, and especially their combination, might be an effective therapeutic tool for the management of different human tumor pathologies, including PCa. Obviously, further work will be required to complete our understanding of this complex process and to fully elucidate the molecular mechanisms underlying these potential antitumor effects in PCa cells. However, consistent with previous results, we found that metformin, simvastatin, and the combined treatment decreased the expression levels of the well-established PCa proliferation marker MKI67 (50). Moreover, metformin, but not simvastatin treatment, reduced the expression levels of c-MYC proto-oncogene, which is consistent with a previous report indicating that metformin can act as a chemopreventive agent to restrict prostatic neoplasia initiation and transformation by downregulating c-myc (51). Furthermore, we explored the capacity of metformin, simvastatin, and their combination to modulate the expression of key regulators of prostate pathophysiology as the IGF1/INS-system because it has been broadly reported to be critically involved in PCa development and progression in human and animal models (52-54), especially in patients with adverse metabolic conditions (55, 56). Consistent with previously reported results, we found that metformin downregulated IGF1R, IGF1, and INSR in PCa cells, probably leading to a disruption of the IGF system (57, 58). However, although simvastatin and the combined treatment also downregulated IGF1R, their inhibitory effects were not higher than those exerted by metformin. Therefore, according to our results, the additive antitumor effects of the combination of metformin and simvastatin do not seem to be mediated by the dysregulation of the IGF1/INS system.
However, one of the most relevant observations in our study is that metformin and simvastatin treatments decreased the expression of AR, which is in line with previous results in PCa cells in vitro (59, 60). Yet, to the best of our knowledge, this is the first study reporting that the combination of both drugs further downregulates the expression of AR. Moreover, our results are also the first to document that metformin, simvastatin, and especially their combined treatment not only reduce AR expression but also AR activity (ie, phosphorylation-ARSer213), which is a more critical endpoint in PCa cells, thus leading to a lower transactivation of AR (61). Although the precise mechanistic underpinnings of this observation are still unknown, these results might be relevant from the (patho)physiological point of view in PCa, in that AR expression and activity play a pivotal role in prostate tumorigenesis and PCa aggressiveness (20). In this scenario, it is tempting to speculate that the striking change in the expression and activity of AR found in PCa cells in response to metformin, simvastatin, and especially to the combined treatment may translate into a relevant clinical implication in this pathology (eg, by decreasing the aggressiveness potential and sensitizing PCa cells to antihormonal agents (20, 62)). Consequently, these findings suggest a putative utility of these drugs, especially their combination, as a potential new therapeutic tool for the management of human PCa, especially castration-resistant prostate cancer. Obviously, further work will be required to complete our understanding of this complex process and to fully elucidate the molecular mechanisms behind these interesting and potentially relevant observations.
To further interrogate the mechanisms underlying the antitumor actions of metformin, simvastatin, and especially their combination, we analyzed additional signaling pathways important in the pathophysiology of PCa cells. Specifically, we initially selected AMPK, AKT, and ERK based on previous reports from our group indicating that these signaling routes might be modulated by metformin and/or simvastatin in other endocrine-related cancer cellular systems (40-42). Our data revealed that metformin and simvastatin might modulate distinct intracellular signaling pathways to exert their antitumor actions in PCa cells. As expected, we found that AMPK signaling, considered the central mediator of metformin actions in different endocrine-related tumors (63), was significantly activated in response to metformin, but not simvastatin, in PCa cells. In contrast, simvastatin, but not metformin, decreased p-AKTSer463/AKT and p-ERKThr202/ERK ratio, presumably leading to an inactivation of PI3K/AKT and MAPK/ERK signalling pathways, which have been associated with the antitumor actions of different drugs in PCa cells (64, 65). Taken together, these observations could support the idea that metformin and simvastatin exert their antitumor actions in PCa cells through different signaling pathways (AMPK vs AKT/ERK); however, they may not be sufficient to explain why the coadministration of both drugs exerts additive effects in PCa cells. To ascertain this question, we next studied mTOR signalling (a critical oncogenic route in PCa cells that represents a key downstream effector of PI3K/AKT pathway (66)) based on previous observations indicating: (1) that the combined treatment of metformin and simvastatin strongly decreased AR expression/activity compared with the individual treatments (results from this study), and (2) that the activation of AR enhances and reprograms mTOR chromatin-binding profiles and that nuclear mTOR activity is essential for androgen-mediated transcriptional reprogramming of metabolism in PCa cells (67). Interestingly, we found that the combination of metformin and simvastatin, but not any of the individual treatments, was able to significantly decrease the mTOR signaling pathway (ie, by inhibiting Ser2448 phosphorylation). Moreover, because the expression levels of several CKIs are regulated by the pathways found to be altered in response to the combined treatment of metformin and simvastatin (ie, AR, mTOR, AMPK, PI3K/AKT, and MAPK/ERK) (68-74), we tested the expression of relevant CKIs in PCa (ie, CDKN1A, CDKN1B, CDKN2A, and CDKN2D (75-78)). Notably, we found that the expression levels of all these CKIs were upregulated in response to combined treatment in vitro. Consistently, in the samples from our PCa patient cohort, tumors from patients treated with metformin and statins in combination had higher CKIs expression levels (namely, CDKN1B, CDKN2A) than those who did not receive any of the treatments. Hence, based on all these results reported here, it seems reasonable to suggest that the coadministration of metformin and simvastatin might exert additive antitumor actions through a hyperinactivation of mTOR and AR, as well to an increase of the levels of CKIs.
Taken together, the present study provides the first detailed description of the direct effects that different biguanides and statins exert in PCa cells. Interestingly, our results clearly demonstrate that the combination of metformin and statins strongly reduces key parameters of aggressiveness in vitro (ie, proliferation rate, migration capacity, and formation of tumorspheres) and in patients (ie, lower Gleason score, longer free biochemical recurrence survival), and that these effects are likely mediated through distinct signalling pathways (ie, modulation of AMPK/mTOR, AR, and CKI levels) that may result in additive effects. Consequently, considering the clinical safety of metformin and statins, results from this study suggest that the combination of these compounds could represent an attractive and effective therapeutic approach to tackle PCa.
Abbreviations
- AR
androgen receptor
- CKI
cyclin-dependent kinase inhibitor
- FFPE
formalin-fixed paraffin-embedded
- NonSigPCa
nonsignificant prostate cancer
- PCa
prostate cancer
- SigPCa
significant prostate cancer
Acknowledgments
Financial Support: This research was funded by Instituto de Salud Carlos III; cofunded by European Union (ERDF/ESF, “Investing in your future”) (PI16/00264, PI17/02287, CD16/00092); MINECO/MECD (PID2019-105564RB-I00, FPU16/06190, FPU17/00263, FPU18/02485, BFU2016-80360-R); Junta de Andalucia (BIO-0139); and CIBERobn. E.M.Y.-S. was the recipient of the Nicolas Monardes Programme from the “Servicio Andaluz de Salud, Junta de Andalucia,” Spain (C1-0005-2019). CIBER is an initiative of Instituto de Salud Carlos III, Ministerio de Sanidad, Servicios Sociales e Igualdad, Spain.
Additional Information
Disclosures: The authors have nothing to disclose.
Data Availability
All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.
References
