Introduction
Many cancers display activation of the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR) pathway, which plays an important role in controlling the proliferation and metabolism of cancer cells [
1]. However, only few malignancies demonstrate clinical response to mTOR inhibitors, such as everolimus. Accordingly, mTOR inhibitors are only FDA- and EMA-approved for use in patients with a few types of cancer, such as mantle cell lymphoma, pancreatic neuroendocrine tumors, advanced renal cell carcinoma, and giant cell astrocytoma [
2‐
6].
In other solid organ malignancies, mTOR inhibitors show clinical responses, though in most cases not enough to warrant monotherapy and with considerable toxicity. A major issue is that most cancers develop resistance against monotherapy with mTOR inhibitors. One of the proposed mechanisms how tumor cells develop resistance to mTOR inhibition is through the mTOR-dependent negative feedback loop that acts to inhibit PI3K/AKT activity. Via this mechanism, drugs that target mTOR inhibit neoplastic processes downstream of mTOR but also the feedback loop. Subsequently, this may activate AKT [
7] and its downstream oncogenic effects [
8], thereby limiting the efficacy of monotherapy with mTOR inhibitors [
9,
10].
A possible way to prevent the development of therapy resistance against mTOR inhibitors is by combining these drugs with the mitochondrial respiration inhibitor metformin, a drug commonly prescribed for the treatment of type 2 diabetes mellitus (T2DM). Several in vitro studies that combined metformin with cytotoxic agents or with everolimus have shown a synergistic inhibition of breast cancer cell growth, compared to single-agent therapy [
11,
12]. Metformin’s synergistic antiproliferative effect in combination with everolimus is a result of activation of 5′ adenosine monophosphate-activated protein kinase (AMPK), which inhibits AKT. Via this mechanism, metformin counteracts the oncogenic AKT activation that is induced by mTOR inhibition [
13,
14]. In short, because metformin may delay or overcome the mechanism of treatment resistance against monotherapy with an mTOR inhibitor, there is a rationale for combining these two drugs [
7]. In a retrospective combined analysis of 94 patients from three clinical trials with various mTOR inhibitors for the treatment of endometrial cancer, self-reported metformin use was associated with a higher response rate to an mTOR inhibitor (18 vs. 7%) and a lower rate of disease progression (12 vs. 33%) [
15]. Also, a clinical trial combining everolimus and letrozole for the treatment of endometrial cancer found that patients that received metformin to treat hyperglycemia (either in the context of pre-existing diabetes or study treatment-related) had a significantly higher response rate (56 vs. 23%) [
16]. Moreover, metformin has attracted interest as an anti-cancer drug [
17] since an association between metformin use in T2DM patients and a reduced risk of breast, colon, pancreas and prostate cancer was acknowledged [
18‐
23], as well as a reduced risk of mortality, as compared with patients treated with insulin or sulfonylureas [
24].
The present clinical trial investigates the safety and maximum tolerated dose (MTD) of a combination therapy with everolimus and metformin in patients with advanced cancer. As secondary objectives, the pharmacokinetics of everolimus and metformin combination treatment and tumor responses to study treatment were assessed.
Discussion
This study explored the safety and pharmacokinetics of the combination of everolimus and metformin in patients with advanced solid malignancies. We found that a combination regimen of everolimus and metformin is poorly tolerated in these patients. Therefore, we were unable to determine the MTD for this combination treatment.
Recently, a retrospective study of 31 patients with pancreatic neuroendocrine tumors (PNETs) was published [
30]. This study showed improved clinical outcomes to treatment with everolimus and octreotide in patients that were also treated with metformin for diabetes as compared with patients that were treated with insulin or as compared with nondiabetic patients. In contrast to our study, Pusceddu et al. [
30] made no mention of intolerable toxicity. By speculation, possible explanations for this apparent difference come to mind. First, diabetic PNET patients that were already being treated with metformin may better tolerate everolimus than cancer patients that are treated with metformin and everolimus for the first time because the former group may have had more time to habituate to metformin. Second, metformin was administered 500 mg bid upfront in our study, whereas the patients described by Pusceddu et al. received metformin after a titration period, starting at 500 mg qd in the first week [
30]. The latter approach mimics metformin treatment schedules in T2DM and decreases the initial side effects of metformin, especially those of gastro-intestinal nature [
31,
32]. What argues against this possible explanation for metformin toxicity is that only two patients of the present study experienced gastro-intestinal side-effects and all gastro-intestinal side-effects were grade 1–2, except one case of grade 3 melaena. Third, octreotide might ameliorate the toxic effects of the combination regimen of everolimus and metformin through mechanisms that are still unknown. Fourth, a retrospective study design such as that of Pusceddu et al. might lead to unintentional under-ascertainment of toxicity when side-effects lead to discontinuation of treatment, which prevents the inclusion of these patients in a retrospective analysis.
In a phase I clinical trial, Khawaja et al. investigated the combination of metformin and another mTOR inhibitor, the intravenously administered temsirolimus [
14]. The authors concluded that this combination was well tolerated with only two DLTs observed among 21 patients and an MTD/recommended dose of 25 mg temsirolimus weekly and 1000 mg metformin bid. In contrast, MacKenzie et al. observed severe toxicity in a similar phase I clinical trial in patients with advanced cancers that were also treated with metformin and temsirolimus, reporting an MTD of 20 mg temsirolimus weekly and 500 mg metformin qd [
33]. The main difference between the two study designs was that the former utilized a titration period for metformin to limit its toxicity, whereas the latter did not.
On one hand, altered pharmacokinetics of everolimus were observed when everolimus was administered in combination with metformin as compared with everolimus single-agent administration. While the maximal everolimus concentration was lower in combination than as single-agent, this was statistically not significant but still might suggest that metformin affects everolimus absorption. Furthermore, everolimus clearance was significantly higher in combination with metformin than as single-agent, while the elimination half-life was unchanged, possibly due to a moderate, but unsignificant, increase in the volume of distribution.
On the other hand, altered metformin pharmacokinetics were observed when metformin was administered in combination with everolimus as compared with metformin single-agent administration, i.e. a lower elimination half-life and elimination rate constant. However, the significantly slower metformin elimination was not accompanied by a significantly slower clearance. This might be explained by a moderate, but unsignificant, increase in the volume of distribution. The slower metformin elimination did not translate to higher AUCs in combination with everolimus versus single-agent administration, which suggest that the observed toxicity of the study treatment cannot be explained by higher metformin AUCs in combination with everolimus versus single-agent administration. A limitation of our pharmacokinetic analysis is that two patients who had an evidently increased exposure to metformin due to slower drug elimination had incomplete time-concentration curves and could thus not be included in our pharmacokinetic analysis. This may explain why the quantitative pharmacokinetic analysis did not show a higher metformin AUC when administered in combination with everolimus relative to single-agent administration. Alternatively, our results may suggest that pharmacodynamic rather than pharmacokinetic interactions between everolimus and metformin explain the high toxicity of this drug combination. To the best of our knowledge, there is no data on pharmaco-interactions between everolimus and metformin at the clinical or molecular level yet.
The combination of everolimus and metformin was tolerated for at least nine weeks in three patients and they all had stable disease. Patients treated with everolimus and metformin for at least nine weeks had a prolonged survival compared with those who had to discontinue study treatment prematurely due to toxicity, but these analyses were conducted in very small groups and should be carefully considered as merely preliminary evidence of anti-tumor efficacy of the combination of an mTOR inhibitor and a biguanide. Another study, investigating a combination of metformin and the mTOR inhibitor temsirolimus, also observed modestly promising anti-tumor efficacy in a cohort with heavily pretreated patients with advanced cancer [
14].
In conclusion, results from a prospective, open-label, single-center phase I study show that the combination of everolimus and metformin is poorly tolerated in patients with advanced cancer. This may be due to pharmaco-interactions between the two drugs, because everolimus delayed and inhibited the elimination of metformin. Our findings have implications for daily practice, especially for diabetic patients using metformin that also have a cancer being treated with everolimus. In addition, our data may be important for the interpretation and design of ongoing and future clinical trials that study the combination of everolimus and metformin (see
ClinicalTrials.gov entries NCT01627067, NCT01797523 and NCT02294006). Although our analyses of tumor responses and overall survival are based on very small groups and should be interpreted with the greatest caution, they could suggest that combining an mTOR inhibitor and a biguanide has anti-cancer activity in patients with advanced cancer. Therefore, alternative combination regimens [
14] could be investigated in future studies with drugs other than metformin and/or everolimus.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.