Repurposing metformin for cancer treatment: current clinical studies

Abstract

INTRODUCTION

Metformin is one of the most widely prescribed oral anti-diabetic medications. It is the first line therapy for type 2 diabetes mellitus [1]. It has an anti-hyperglycemic effect which is mediated by inhibiting gluconeogenesis, decreasing glucose absorption from the small intestine, increasing glucose uptake in cells, and decreasing plasma free fatty acid concentration [2]. Metformin also increases insulin induced translocation of glucose transporters to the cellular plasma membrane, thus reducing insulin resistance [3]. Use of metformin has been found to be generally safe, with mild gastrointestinal symptoms being the most common adverse effects [4].

There is substantial preclinical evidence suggesting that metformin has anti-cancer properties. In-vitro and in-vivo analysis of metformin has exhibited anti-proliferative activity by inhibiting intracellular pathways. It has also been observed that metformin activates the T cell mediated immune response against cancer cells.

Numerous retrospective studies have reported that metformin is associated with a reduced risk of developing cancer. Meta-analysis of data obtained from cohort and observational studies has revealed that metformin use was associated with a decrease in both cancer related and all-cause mortality.

Here we summarize the available evidence from clinical trials of metformin as part of cancer therapy. We review the landscape of current investigation and suggest directions for future investigation.

PRECLINICAL DATA

Metformin has been extensively studied in preclinical models, which has revealed numerous molecular pathways that it modulates, either directly or through other downstream targets, contributing to reduction in growth and proliferation of tumor cells. The inhibition of mTOR (mammalian target of rapamycin) in tumor cells is one of the potential key mechanisms that facilitates the anti-cancer activity of metformin. Use of metformin in MCF-7 breast cancer cells exhibited reduction in phosphorylation of S6 kinase, ribosomal protein S6 and eIF4E binding protein, along with inhibition of mTOR and reduced translation initiation due to AMPK activation [5]. Animal models of pancreatic cancer fed with metformin showed inhibition of insulin like growth factor-1 (IGF-1) and mTOR, along with an increase in phosphorylated AMPK and tuberous sclerosis complex (TSC1, TSC2) [6]. The AMPK mediated phosphorylation of TSC2 has been observed to increase the activity of TSC2, leading to inactivation of mTOR [7, 8]. AMPK has been described to directly inhibit mTORC1 through phosphorylation of mTOR binding raptor as well [9]. Comparing the effects of metformin with rapamycin, a direct mTOR inhibitor, metformin decreases the activation of AKT in addition to AMPK dependent mTOR inhibition. Thus, metformin renders a better anti-tumor response than rapamycin in breast cancer cells [10]. Metformin has been found to decrease HER2 expression in human breast cancer cells by directly inhibiting p70S6K1, which is a downstream effector of mTOR [11]. In a study using nude mice with acute myeloid leukemia (AML), the use of metformin was correlated with a decrease in proliferation of AML cells. This action was characterized by the activation of LKB1/AMPK/TSC pathway, which led to mTOR inhibition and consequently suppression of mRNA translation [12]. In tobacco carcinogen induced lung cancer mice, the inhibition of insulin like growth factor 1 receptor/insulin receptor (IGF- 1R/IR) by metformin decreased the downstream signaling through Akt pathway. This reduced the activation of mTOR in lung tissue which corresponded to a 72% reduction in tumor burden [13].

Metformin can inhibit the activation of mTOR independent of AMPK pathway as well. Metformin has been shown to escalate the expression of REDD1 by a p53 mediated inhibition of mTOR in prostate cancer cells [14]. Another study described that metformin can prevent mTOR activation through Ras-related GTPase (RagGTPases), independent of AMPK, as well as TSC1/2 [15].

Tumors are known to exhibit the Warburg effect, where tumor cells generate ATP from glycolysis instead of oxidative phosphorylation secondary to low nutrient supply and hypoxia [16]. Metformin blunts the Warburg effect and consequently downregulates the growth of cancer stem cells [17]. In-vivo studies on hepatocellular carcinoma xenografts have shown that metformin improves cellular oxygenation ability and decreases mitochondrial oxygen consumption, thus suppressing hypoxia-induced HIF-1α accumulation. These effects form the basis of anti-cancer activity of metformin, particularly against hepatocellular carcinoma [18].

In addition to the above, various other mechanisms of metformin for cell growth inhibition have been identified in pre-clinical models. Metformin therapy in paired isogenic colon cancer cell lines (HCT116 p53 [+/+] and HCT116 p53 [−/−]) showed an increase in apoptosis of p53 deficient cells [19]. In-vitro and in-vivo analysis of metformin therapy showed that it can inhibit growth of ovarian stem cells [20], glioma initiating cells [21], breast cancer cells [22], endometrial cancer cells [23] and non-small cell lung cancer cells [24]. It has also exhibited synergistic action with VEGF inhibitors to inhibit proliferation of BRAF mutant melanoma cells [25]. Additionally, it has been noted that metformin increases radiosensitivity of cancer cells [22, 24].

Recent experiments on animal models have also suggested that metformin has immune modulatory properties. Metformin inhibits immune exhaustion of CD8+ tumor induced lymphocytes (TIL), thereby enhancing T cell mediated immune response to tumor tissue. It decreases apoptosis of CD8+ tumor infiltrating lymphocytes (TILs), and also shifts the phenotype of CD8+ TILs expressing exhaustion markers (especially PD1 negative Tim3 positive) from central memory T cells (TCM, inactive against tumor cells) to effector memory T cells (TEM, active against tumor cells). The increase in TEM cell population has been found to correlate with regression of tumor cells [26]. In a study evaluating an experimental cancer vaccine, administration of metformin after vaccination in animal models showed an increase in CD8+ memory T cells which conferred protective immunity upon subsequent tumor challenge [27]. Figure ​Figure11 summarizes the effects of metformin on various cellular pathways.

Figure 1

Possible mechanisms of anti-cancer activity of metformin

EPIDEMIOLOGY

There are numerous epidemiological studies that have put forth evidence suggesting utility of metformin as an anti-cancer agent.

Several observational and cohort studies have been conducted to assess the influence of metformin on cancer. Survival analysis in a cohort study on diabetic patients comparing metformin users (n = 4,085) to non-users (n = 4,085) reported a reduced risk of cancer (hazard ratio, HR = 0.63; 7.3% diagnosed with cancer in metformin users versus 11.6% in non-users) [28]. Meta-analysis performed on 37 studies, with a total of 1,535,636 patients comparing metformin users and non-users reported overall cancer incidence summary relative risk (SRR) as 0.73. The results also noted a reduction of cancer incidence for liver (78%), breast (6%), pancreatic (46%) and colorectal cancer (23%). However, no significant correlation could be derived between the use of metformin and incidence of prostate cancer [29]. Meta-analysis of 6 case control studies comparing 39,787 participants that were on metformin to 177,752 participants that were not, exhibited a lower risk of developing lung cancer in the metformin group (odds ratio, OR = 0.55; p < 0.001) [30]. Likewise, another meta-analysis showed a reduction in incidence of prostate cancer (RR = 0.88; p = 0.03) in patients who were on metformin treatment [31]. Meta-analysis of 35 observational studies reported a considerable correlation with using metformin to reduce the risk of developing all-cancer (OR = 0.73), liver cancer (OR = 0.34), colorectal cancer (OR = 0.83), pancreatic cancer (OR = 0.56), gastric cancer (OR = 0.83), and esophageal cancer (OR = 0.90) [32]. A recent meta-analysis of 265 studies showed the use of metformin or thiazolidinediones was associated with a lower incidence of cancer (RR = 0.86 and 0.93 respectively). But interestingly, insulin, sulfonylureas, and alpha glucosidase inhibitor use was associated with an increased incidence of cancer (RR = 1.21, 1.20, 1.10 respectively) [33].

Meta-analysis of retrospective studies has provided substantial evidence associating the use of metformin with a decrease in cancer related mortality. A cohort study, investigating mortality due to cancer in type 2 diabetes patients, showed a decrease in cancer related mortality risk (HR = 0.56) with the use of metformin [34]. Data from meta-analysis of 6 observational studies noted a significant correlation between using metformin and risk of cancer related mortality (OR = 0.65) [32]. Retrospective analysis of data from over 350 primary care practices in the United Kingdom found a decrease in cancer related mortality (HR = 0.85) in diabetic patients who were on metformin monotherapy when compared with those on other drugs for diabetes [35]. Another meta-analysis showed that the use of metformin in diabetic patients diagnosed with cancer was associated with a decrease in risk of all-cause mortality in cancers of breast (pooled relative risk, RR = 0.70; p = 0.003), ovary (RR = 0.44; p < 0.001), endometrium (RR = 0.49; p = 0.001) and colorectal cancer (RR = 0.70; p < 0.001) [36].

In light of the available preclinical and retrospective data suggesting anti-cancer properties of metformin, clinical trials are necessary to further investigate its role in cancer therapy.

METHODS

Information about the study drug was obtained from clinicaltrials.gov (a service of United States National Institutes of Health), using the search query “metformin” and “cancer”. Of the 223 results obtained from the search engine, relevant drug trials were selected. The published clinical trials were obtained from PubMed using the same search query and choosing the “clinical trial” filter. The publications that focused on use of metformin for treatment of cancer in human subjects through a prospective clinical trial were identified and selected. The results of completed clinical trials were obtained from PubMed and online abstract library for professional societies such as the American Association of Cancer Research (AACR) and American Society of Clinical Oncology (ASCO). The AACR and ASCO online libraries were explored using the NCT number associated with the clinical trial. The authors, enrollment number, primary & secondary outcomes and primary location of conducting the research mentioned in abstracts collected from ASCO and AACR database were matched with the information obtained from clinicaltrials.gov to avoid any discrepancies in associating the abstracts to the appropriate trial.

CLINICAL TRIALS

Presently, there are 55 ongoing clinical trials in various stages that are evaluating metformin as a monotherapy (11 trials, 20% of all ongoing trials using metformin as an anti-cancer agent) or in combination with cytotoxic chemotherapy (38 trials, 69%) and/or radiotherapy (6 trials, 11%) for the treatment of cancer (Tables ​(Tables1,1, ​,2,2, ​,3,3, ​,4).4). These trials primarily focus on establishing the effects of metformin on markers of cellular proliferation, pathological response rate, progression free survival, and recurrence free survival. Also, certain trials are directed towards determining the maximum tolerable dose of metformin in specific tumors.

Table 1

Proof of concept for anti-tumor activity of metformin

Table 2

Anti-tumor activity of metformin in locally advanced and hematologic cancers

Table 3

Anti-tumor activity of metformin in metastatic tumors

Table 4

Antitumor activity of metformin in combination with radiotherapy

A considerable amount of focus has been laid on investigating metformin as a potential anti-cancer agent for cases of breast cancer. Eleven trials (20% of all ongoing trials using metformin as an anti-cancer agent; Tables ​Tables1,1, ​,2,2, ​,3)3) are focused on evaluating metformin as a treatment for breast cancer. Of these, two trials are using metformin as monotherapy. There are 9 trials using metformin in combination with other anti-cancer agents. These include capecitabine, cyclophosphamide, docetaxel, doxorubicin, erlotinib, epirubicin, exemestane, ganitumab, letrozole, sirolimus and temsirolimus. One trial is exploring the use of metformin plus atorvastatin combination as a possible treatment for breast cancer. Two trials using metformin combination therapy are also evaluating pathological complete response as a primary endpoint. Apart from the ongoing trials, data obtained from 5 completed trials (all using metformin monotherapy in a pre-surgical window of opportunity trial design) has facilitated a better understanding regarding the effects of metformin in breast cancer. In addition to survival outcomes, several surrogate markers are also being employed to study the effects of metformin on breast cancer cell population. These include Ki67, S6K, 4E-BP-1, AMPK and effects on AMPK/mTOR pathway.

A phase 2 single arm window of opportunity trial of 39 breast cancer cases [37] showed significant reduction in Ki67 (36.5 to 33.5 %, p = 0.016) and an increase in TUNEL staining (0.56 to 1.05, p = 0.004) along with significant fall in HOMA (homeostatic model assessment, used for determining the status of insulin resistance) [38]. A recently published randomized control trial (RCT) also reported a decrease in Ki67 staining (mean = 3.4%, p = 0.027). Additionally, it noted an increase in mean AMPK score, fall in pAKT score and reduced caspase-3 staining in patient samples with the use of metformin when compared to placebo [39]. However, other trials have had conflicting outcomes. A phase II RCT with 200 participants recorded no significant changes in Ki67 on comparing metformin and placebo arms. But, interestingly, the cases with HOMA ≤ 2.8 showed a non-significant increase of Ki67 by 11.1% (95% Confidence interval (CI): −0.6% to 24.2%) and those with HOMA > 2.8 (implying a higher probability of insulin resistance) showed a non-significant mean proportional decrease in Ki67 by 10.5% (95% CI: −26.1% to 8.4%) [40]. Another phase II trial (non-randomized) examined effects of metformin in overweight/obese patients with stage 0-III breast cancer. Though noting a correlation of Ki67 with tumor growth, the calculated ln (Ki67) showed no significant changes when comparing metformin to placebo [41]. A different study with 200 participants randomized to metformin or placebo did not document any major difference in Ki67 and TUNEL levels (used for assessing cellular apoptosis) between the two arms. The study did note that TUNEL levels were higher in women without insulin resistance (metformin: +4%, interquartile range, IQR: 2-14, placebo: +2%, IQR: 0-7) as compared to those who had insulin resistance (metformin +2%, IQR: 0-6, placebo +5%, IQR: 0-15) [42]. The survival benefit with the use of metformin in breast cancer is being evaluated in 3 clinical trials, however there is no data presently available (NCT01627067, NCT0131023, NCT01885013).

Metformin is presently being evaluated as an anti-cancer agent for endometrial cancer as well. There are 6 ongoing trials (10.9% of all ongoing trials using metformin as an anti-cancer agent; Tables ​Tables1,1, ​,2),2), with 3 each for monotherapy and combination chemotherapy. The drugs that are currently being assessed in combination with metformin for treatment of endometrial cancer are carboplatin, everolimus, letrozole, paclitaxel, and megestrol acetate. One trial is assessing the role of metformin as a maintenance therapy. In conjunction with clinical response, the effect of metformin on endometrial cancer is being studied through a wide variety of markers including Ki67, pS6, Akt, pAMPK, ERK1/2, histone H3, telomerase, topoisomerase IIα and p27. The effect of using metformin on expression of estrogen (ER) and progesterone (PR) receptors in cancer tissue of endometrial origin is also being investigated. Three completed trials, all pre-surgical window of opportunity trials using metformin monotherapy, have shown a significant decrease in Ki67 staining [43–45] and in pS6 staining [44, 45]. One trial also reported a reduction in topoisomerase IIα and ERK 1 / 2, along with significant elevation in pAMPK and p27 [45]. A different trial reported a decrease in tumor cell proliferation by 11.75% and a decrease in expression of ER with the use of metformin. PR expression, however, was not affected [43]. Thus far, no trial has provided any data on survival benefit with the use of metformin in endometrial cancer, though one trial is ongoing (NCT02065687).

Metformin is being assessed in combination with various anti-cancer agents for the treatment of pancreatic cancer. Presently, there are 7 ongoing and 2 completed trials (Tables ​(Tables2,2, ​,3,3, ​,4).4). With the exception of one trial where the treatment regimen involves using metformin together with stereotactic radiosurgery (Table ​(Table4),4), all others are using metformin in combination with different anti-cancer agents. These include cisplatin, capecitabine, epirubicin, erlotinib, everolimus, gemcitabine, octreotide, paclitaxel, rapamycin, and FOLFOX (fluorouracil, oxaliplatin, leucovorin). The effects of metformin are being assessed mostly through clinical outcomes including progression free survival (PFS), recurrence free survival (RFS) and toxicity due to chemotherapy combination. The results from a phase II non-randomized trial showed that the combination of metformin plus paclitaxel was not well tolerated, with 42.1% patients experiencing grade 3-4 toxicities. A total of 31.6% cases had to undergo metformin dose reduction secondary to development of diarrhea, while one case experienced febrile neutropenia which was attributed to paclitaxel. This trial reported a median overall survival (OS) of 133 days and median PFS as 43 days, but could not meet the disease control rate endpoint [46]. Another trial, consisting of 120 participants randomized to metformin or placebo arm, noted that although the combination of metformin, gemcitabine and erlotinib was well tolerated, the 6 month survival rate was 55% in metformin arm and 66% in placebo arm. Also, no significant difference was observed in PFS and median OS between metformin users and non-users [47].

Five clinical trials (Tables ​(Tables1,1, ​,2,2, ​,3)3) are presently working to evaluate if metformin may be of value in the treatment of prostate cancer. These include two trials that are using metformin as monotherapy and three in combination with different agents: abiraterone (NCT01677897), docetaxel (NCT01796028), and enzalutamide (NCT02339168). Data made available from one trial, a single arm window of opportunity study, showed a significant reduction in Ki67 index and 4E-BP-1 staining with no changes in pAMPK. Three of 24 patients developed grade 3-4 toxicities, indicating that the treatment was overall well tolerated [48]. The effect of metformin therapy on PFS for prostate cancer is being assessed in two trials (NCT01433913, NCT02339168), with one other trial evaluating PSA response (NCT01796028). However, presently there is no data available on survival benefit.

There is one phase II trial of metformin use in non-small cell lung cancer (NSCLC) combined with stereotactic body radiotherapy that is currently recruiting patients (NCT02285855).

There are two completed trials (Table ​(Table2,2, ​,3)3) on multi-histology solid tumors assessing the dose limiting toxicity (DLT) of various treatment regimens that include metformin. One of the trials used metformin in combination with 26 chemotherapy regimens for 17 tumor types on a total of 100 participants. The study was divided in two stages. In stage one, participants were randomized to receive metformin or placebo with chemotherapy. In stage two, participants of delayed arm would be crossed over to receive metformin with chemotherapy. Results showed that 46% of participants documented stable disease. In the sub-set of patients having quantifiable tumor markers, 28% exhibited favorable changes. The participants receiving metformin together with chemotherapy showed a lower rate of DLT (6.1% in stage one of concurrent arm, including grade 3 anemia, decrease in albumin and elevation in ALT) compared to those who received just the chemotherapy (7.8% in stage one of delayed arm [including grade 3 syncope, dehydration and elevation of bilirubin] and 3.8 % in stage two [including dehydration, vomiting and proteinuria]). The participants reporting DLTs in stage two of delayed arm were known cases of adverse events with chemotherapy [49]. The other trial used metformin and temsirolimus combination in 11 patients. It reported that 100% of participants had grade 1 toxicity and 82% experienced grade 2 toxicity, with DLT being reported in all 3 patients of the first cohort (grade 4 pneumonitis, grade 3 fatigue and grade 3 thrombocytopenia). In the second cohort, the dose of temsirolimus and metformin was reduced and DLT was observed in only two of eight cases (grade 4 dyspnea and grade 3 thrombocytopenia). After 2 months of treatment, 5 patients had stable disease, 1 case had partial response and 2 showed progression [50].

DISCUSSION

There is great excitement surrounding metformin as a potential anti-cancer agent. Epidemiological data has associated the use of metformin with a decrease in the risk of developing cancer and a reduced cancer related mortality. The information that has been gathered from preclinical studies has provided encouraging evidence for anticancer mechanisms of metformin. It has been suggested that metformin may well be used as a radiation sensitizer or an immunotherapy drug, in addition to a direct anti-proliferative agent for the treatment of cancer.

The anticancer mechanism of metformin has been extensively studied and attributed to mTOR inhibition. More recent data has revealed an immunomodulatory effect on cancer cells. Pre-clinical data has demonstrated that metformin can inhibit apoptosis of CD8+ TILs. In addition, it also increases the effector memory T cell population through phenotype switching of CD8+ TILs, thus enhancing the immune response against tumor cells [26]. The use of metformin with an experimental cancer vaccine (LmOVA) showed an increase in the number of CD8+ memory T cells that conferred immunity to cancer [27]. The recent breakthrough developments in immunotherapy for patients with advanced melanoma, triple negative breast cancer, and non-small cell lung cancer with check point blockade monoclonal antibodies (anti-PD-1/L1) have generated excitement in the oncology community [51–53]. Immunomodulatory properties of metformin have yet to be studied in combination with other forms of immunotherapy, in particular with check point blockade monoclonal antibodies. Further investigation into a possible synergistic effect is warranted.

The preliminary results from clinical trials assessing metformin as an anti-cancer agent have shown that metformin can significantly impact markers of tumor proliferation. A total of 19 ongoing and completed trials (Table ​(Table1)1) are using various surrogate markers to assess pro-apoptotic effects of metformin on cancer cells. Although majority of these trials are being performed on breast and endometrial cancer cases, a limited number of trials are also evaluating tumors of head & neck, prostate, bladder, lung, kidney and lymphoma.

Pre-surgical window of opportunity trials in endometrial and breast cancer showed that tumor markers such as Ki67 and TUNEL (indicative of changes in cell proliferation and apoptosis respectively) exhibited favorable anti-tumor effect. Two trials demonstrated favorable changes in Ki67 and TUNEL in a subset of women without insulin resistance as compared to those with insulin resistance. Although the findings were not statistically significant, the direction of change suggests an intriguing hypothesis. The use of metformin in patients without insulin resistance may offer more benefit as compared to those with insulin resistance.

Diabetes is known to be associated with insulin resistance [54] and an impaired immune response against various pathogens [55, 56]. The immune system responds to the proliferating tumor cells by increasing production of tumor specific lymphocytes which check tumor growth by various mechanisms [57]. Evidence from preclinical trials has described that metformin, at least in part, exerts an anti-cancer effect by inhibiting immune exhaustion of CD 8+ TILs [26], thus amplifying the existing immune action against cancer cells. Therefore, it may be hypothesized that the patients with insulin resistance have a compromised immune system, which consequently results in a sub-optimal anti-cancer effect of metformin. It might be rational to stratify outcomes according to the insulin resistance status of participants in future clinical trials in order to better appreciate the anti-cancer activity of metformin.

Fourteen ongoing trials (Tables ​(Tables1,1, ​,2,2, ​,3,3, ​,4)4) are presently amassing evidence to ascertain if a survival benefit is associated with the use of metformin in various malignancies. These include tumors of the breast, pancreas, lung, endometrial, brain, prostate and gynecological cancers. Thus far, there are results available from two clinical trials on metastatic pancreatic cancer, neither of which had favorable outcomes. From this data, it can be speculated that metformin may not be viable option for the treatment of advanced pancreatic cancer. The upcoming clinical trials may need to shift focus towards treating earlier stages of pancreatic cancer or using a different combination of agents with metformin to have better outcomes in advanced disease. Additional data on survival indices from multiple ongoing trials will be pivotal to draw a better conclusion.

One concern with the clinical utility of metformin is its side effect profile, particularly in combination with cytotoxic chemotherapy. Metformin is well known for causing GI upset, sometimes limiting patient compliance due to discomfort. Clinical trials have revealed a low incidence of DLTs with metformin in combination with a wide variety of chemotherapy regimens [48]. With this data, clinicians can be reassured that metformin will most likely be a tolerable addition to a chemotherapy regimen, and should not limit its practical utility.

CONCLUSIONS

A strong base of epidemiological and pre-clinical data has prompted attempts to probe the anti-cancer effects of metformin through clinical trials. Metformin has been shown to have a favorable effect on markers of tumor proliferation but it remains to be seen if that translates to benefit in survival rates. It is prudent to find better histology and the appropriate stage of tumors for utilizing metformin therapy. The potential use of metformin as an immunotherapy agent needs to be substantiated with further evidence to ascertain possible benefits in future.