Why Metformin Should Not Be Used as an Oxidative Phosphorylation Inhibitor in Cancer Patients

Oxidative phosphorylation (OXPHOS) is a metabolic process responsible for generating the majority of cellular energy required by cells, in the form of adenosine triphosphate (ATP). OXPHOS takes place within the mitochondria, the energy-producing organelles of the cell. During tumorigenesis, cells frequently shift their metabolism toward glycolysis as the primary pathway for energy production, even in the presence of sufficient oxygen [1]. However, most cancer cells still require mitochondrial activity for various biosynthetic processes and the maintenance of redox balance [2]. OXPHOS inhibition has been explored as a potential antitumor therapeutic strategy. Several preclinical studies suggested that the inhibition of OXPHOS can result in reduced tumor cell proliferation, invasion, and migration, as well as increased vulnerability to anticancer therapy [3, 4]. Although the exact mechanisms involved in these processes are not fully understood, several hypotheses have been proposed, including energy stress due to ATP depletion, increased reactive oxygen species production, and alterations in cellular metabolism, which can interfere with the ability of cancer cells to survive and proliferate [5, 6].

Metformin (N,N-dimethylimidodicarbonimidic diamide hydrochloride) is a water-soluble compound with a molecular weight of 165.62 g/mol [7]. Although primarily used for the treatment of type 2 diabetes mellitus (T2DM), metformin is widely considered to be a prime example of clinically relevant compounds capable of inhibiting OXPHOS [3, 8]. Lactic acidosis is a well-known adverse effect of metformin in T2DM patients. Several review articles have highlighted that metformin can directly inhibit the activity of mitochondrial respiratory chain complex I [3, 9]. Metformin also activates adenosine monophosphate-activated protein kinase (AMPK), which regulates metabolism and cell growth. AMPK activation decreases the rate of ATP-consuming anabolic reactions and increases that of ATP-producing catabolic reactions, resulting in reduced cancer cell proliferation (Fig. 1) [10]. Moreover, metformin-mediated AMPK activation induces autophagy [11]. Studies have also suggested that metformin has other pleiotropic effects, such as inhibition of glucagon-induced elevation of cyclic adenosine monophosphate, mitochondrial glycerol-3-phosphate dehydrogenase variant, mTOR signaling, and immune system activation [12, 13]. It also affects the gut microbiota, exerts an anorexiant effect, decreases gluconeogenesis in the liver, and inhibits the basal secretion of growth, adrenocorticotropic, and follicle-stimulating hormones [14, 15]. Given its favorable safety profile compared to other known OXPHOS inhibitors such as rotenone, oligomycin, or phenformin, metformin has attracted considerable interest as a potential therapeutic option against several types of tumors [6].

Preclinical studies have shown that metformin exerts antitumor effects against various cancers, including pancreatic, colon, breast, and prostate cancers [16‒20]. Metformin also enhances the in vitro activity of various chemotherapeutic agents, such as doxorubicin in MCF-7 and MDA-MB-231 breast cancer cells, gemcitabine in KLM-1R pancreatic cells, and cisplatin in irradiated CNE-1 human nasopharyngeal carcinoma cells [21‒23]. However, human clinical trials have yielded mixed results. In a large phase 3 trial, patients with high-risk operable breast cancer who received adjuvant treatment with metformin did not experience any relapse-free survival prolongation compared to placebo (NCT01101438) [24]. In patients with advanced breast cancer, two phase 2 studies failed to show any progression-free survival or overall survival benefit of metformin use in combination with chemotherapeutic agents, including non-pegylated liposomal doxorubicin plus cyclophosphamide, capecitabine, platinum, or taxanes (NCT01885013) (NCT01310231) [16]. Several studies on lung cancer have investigated treatment strategies combining metformin with chemotherapy, radiotherapy, and epidermal growth factor receptor tyrosine kinase inhibitors, however, early results have been mostly discouraging (NCT01864681) (NCT02115464) [17]. Cumulative evidence from nine cohort studies also did not support any positive effect on pancreatic cancer survival [18]. On the other hand, results from phase 2 clinical trials suggest that metformin can enhance the antitumor effects of chemotherapeutic agents, such as irinotecan or 5-fluorouracil in patients with refractory colorectal cancer (NCT01941953) (NCT01930864) [19]. In patients with advanced prostate cancer, sub-analyses of three randomized, double-blind phase 3 clinical trials suggested that the addition of metformin to enzalutamide did not increase overall survival significantly compared to the administration of enzalutamide alone (NCT00974311) (NCT01212991) (NCT02003924) [20]. In conclusion, metformin may only be beneficial in a small subset of patients with specific tumor types, although more evidence is needed to validate these findings.

These data largely suggest that either metformin-induced OXPHOS inhibition was not an effective antitumor strategy in these patients, or that metformin at the examined doses did not actually pose any clinically meaningful OXPHOS inhibition. In order to address the issue, we need to review the available evidence regarding the effects of metformin at clinically relevant concentrations. For the treatment of T2DM, metformin is usually started at 850 or 1,000 mg per day and titrated to a maximum daily maintenance dose of 2,000 mg [7]. In the STAMPEDE trial (NCT00268476) the daily dose of metformin was 1,000 mg per day, and although other trials used higher doses, none of them exceeded 2,000–2,500 mg per day, which is the maximum recommended dose. The therapeutic concentration of metformin in the blood typically ranges from 0.5 to 2.5 μg/mL for the treatment of T2DM. In the controlled clinical trials, the metformin concentration did not exceed 5 μg/mL (38 μm), even when maximum doses of metformin were used [25].

When metformin is used at therapeutic doses, its plasma concentration in both animals and humans remains within the micromolar range. However, studies have shown that millimolar concentrations are necessary in order to inhibit complex I in isolated mitochondria, with the IC50 ranging from 19 to 79 mm in different laboratories [26].

Therefore, it has been speculated that metformin accumulates inside the mitochondria, and micromolar cytosolic concentrations may result in mitochondrial concentrations within the millimolar range. Metformin is a positively charged (at the internal pH of the body) hydrophilic agent with low permeability through lipid membranes. It enters and leaves cells via several transporters, including multidrug and toxin extrusion transporters and organic cation transporters [26]. This results in a steady-state metformin concentration inside the cells. The inner mitochondrial membrane is also impermeable to a vast majority of hydrophilic molecules, which use specific transporters to move in and out of the mitochondria. Among the several mitochondrial carriers identified, a metformin-specific carrier is yet to be identified. This strongly contradicts the hypothesis that metformin accumulates inside the mitochondria [26]. As highlighted by Fontaine, the positive charge accumulation in the mitochondrial matrix compensated by the respiratory chain-induced proton extrusion should result in the collapse of the mitochondrial membrane potential and an increase in delta pH. However, metformin has never been shown to depolarize isolated mitochondria [26]. In addition, the low volume of metformin distribution and its short half-life are not compatible with large mitochondrial accumulation. Assuming that the total mitochondrial volume represents approximately 20% of the total hepatocyte volume, we can anticipate a 1,000-fold metformin accumulation in the liver mitochondria, to result in at least a 200-fold metformin accumulation in the liver. Such an accumulation is not compatible with that measured by different studies by several orders of magnitude. Moreover, in radioactive [14C] metformin-treated rats, radioactivity did not accumulate in the liver mitochondria [26].

Metformin’s modest effect on the adenosine diphosphate (ADP) to ATP (ADP:ATP) ratio results in AMPK activation, which promotes ATP-generating catabolic pathways and restores cellular energy balance (Fig. 1). Metformin also activates AMPK via a mechanism involving the lysosomes rather than the mitochondria. AMPK and mTOR are master metabolic regulators that control the flux through several pathways. Therefore, they maintain homeostasis during minor metabolic disruptions [10, 27]. These effects are not associated with severe energy depletion, subsequent cellular death, or apoptosis. Metformin also reduces the ADP:ATP ratio through a mechanism that is not mediated by complex I inhibition (Fig. 1) [10, 27].

Metformin has also been shown to inhibit the activity of signal transducer and activator of transcription 3 (STAT3), which is frequently activated in several malignancies. It has been suggested that STAT3-mediated OXPHOS upregulation can contribute to resistance to therapy; hence, it can theoretically serve as a potential marker for response to OXPHOS inhibition [28]. Controversy exists over whether metformin-mediated STAT3 inhibition is actually another result of AMPK activation [29]. STAT3 signaling is complex and directly affects several tumorigenic pathways, including proliferation, evasion of apoptosis, inflammation, metastasis, and angiogenesis [30]. Therefore, STAT3 inhibition is not exclusively associated with metabolic disruption. Moreover, the metabolic effects of STAT3 in cells are complex and likely conflicting. For example, studies have shown that canonical transcriptional activity enhances aerobic glycolysis and lactate production, while suppressing OXPHOS [30]. Hence, currently, there is insufficient evidence to support the hypothesis that metformin-induced STAT3 inhibition (at therapeutic doses) exerts antitumor effects by causing the critical depletion of cellular energy production via the inhibition of OXPHOS.

Despite preclinical evidence, the value of metformin as an agent targeting OXPHOS via complex I inhibition in cancer patients is highly debatable. Hence, we must exercise caution and avoid using metformin to evaluate the antitumor effects of OXPHOS inhibition at this time.

The author reports no financial or nonfinancial conflicts of interest.

No funding was received for this manuscript.

M.S. confirms sole responsibility for the conception, manuscript preparation, and final approval of this paper.