First-in-human study of IM156, a novel potent biguanide oxidative phosphorylation (OXPHOS) inhibitor, in patients with advanced solid tumors

Oxidative phosphorylation (OXPHOS) is an important source of energy and metabolic precursors for tumor cells and offers an attractive target for development of anticancer therapies [1]. Oxidative phosphorylation can be targeted by inhibition of mitochondrial protein complex 1 (PC1), a critical component and the first step in the electron transport chain that binds and oxidizes nicotinamide adenine dinucleotide hydride (NADH) before the subsequent transfer of electrons resulting in formation of adenosine triphosphate (ATP) [2]. Preclinical experiments in cancer demonstrated that biguanides such as metformin, commonly used to treat Type 2 diabetes mellitus (DM), and phenformin, also used to treat Type 2 DM but withdrawn from the market due to the risk for lactic acidosis, can inhibit OXPHOS via inhibition of PC1 [3]. Despite most cancers having increased rates of glycolysis, they also require OXPHOS [1]. Therefore, energetic stress in cancer cells due to inhibition of OXPHOS by biguanides may result in cancer cell death [4‐6]. Metformin demonstrated anticancer activity in preclinical in vitro and in vivo models; however, an important caveat was that many in vitro models with anticancer activity used metformin concentrations higher than those which can be achieved in the plasma of patients treated with conventional doses [7, 8]. Indeed, most clinical studies and randomized trials with metformin failed to show meaningful anticancer activity in patients with advanced cancers [9, 10]. Nevertheless, since many in vivo cancer models demonstrated remarkable antineoplastic activity of biguanides, the development of novel biguanides with improved pharmacokinetic (PK) and toxicity profiles would potentially be of clinical benefit [4, 11‐13].

IM156 is a novel biguanide that inhibits PC1 [3]. Metformin is relatively hydrophilic and thus requires active transport (e.g., through OCT1) to enter cells. IM156, on the other hand, is more hydrophobic than metformin and thus potentially more bioavailable to cancer cells. In addition, at equal concentrations IM156 was more potent at decreasing the oxygen consumption rate (OCR) of tumor cells compared to phenformin and metformin, and in reducing cellular ATP production versus phenformin without an increase in the extracellular acidification rate (ECAR) (Immunomet – Data on File). In addition, IM156 inhibits mTOR through AMPK activation resulting in inhibition of cancer cell growth and proliferation (Immunomet – Data on File). Furthermore, preclinical in vivo models demonstrated anticancer activity of IM156 in glioblastoma, gastric cancer and EGFR-mutated lung cancer [14]. Therefore, we designed a first-in-human dose-escalation study to determine the maximum tolerated dose (MTD) or recommended phase 2 dose (RP2D) and to evaluate the safety, tolerability, pharmacokinetics (PK), exploratory pharmacodynamic (PD) markers, and early signals of efficacy of IM156 in patients with advanced solid tumors refractory to standard therapies.

To our knowledge, this first-in-human study of IM156 is the first successfully completed phase 1 study of a PC1 inhibitor in patients with cancer with a RP2D determined for further clinical development. Other PC1 inhibitors, such as BAY 87–2243 [17], ASP4132 [18], and IACS-010759 [19], have all discontinued clinical development for oncology indications due to poor tolerability, primarily fatigue and AEs of the GI and nervous systems. By contrast, IM156 appears to have a favorable safety and tolerability profile, with treatment-related Grade ≥ 3 AEs occurring infrequently (14% of patients). The toxicity profile of IM156 is primarily gastrointestinal AEs, including nausea and vomiting in 68% and 41% of patients, respectively. Despite these AEs being primarily low-grade, their frequency, persistence, and limited resolution with anti-emetic drugs such as 5-hydroxytryptamine (5HT3) receptor antagonists ultimately precluded further dose escalation. Though no DLTs were reported, the RP2D of IM156 for patients with cancer is 800 mg orally daily, a dose that resulted in exposure above the therapeutic target predicted from preclinical cancer models.

Gastrointestinal AEs have also been observed with other biguanides and drugs targeting PC1 [9, 19, 20]. For instance, in a Phase 1 study of 21 patients of a selective non-biguanide inhibitor of PC1, IACS-010759, nausea and vomiting of any grade were reported in 67% and 29% of patients, respectively. Although diarrhea of any grade was more common in patients treated with IM156 versus IACS-010759 (50% vs. 10%), Grade ≥ 3 AEs were more frequent with IACS-010759 compared to IM156 (43% vs 14%). Additionally, IACS-010759 led to debilitating AEs not observed with IM156, such as Grade ≥ 3 visual impairment and peripheral neuropathy, and elevations of blood lactate (71% vs. 9%). The latter may be a result of a differences in the drugs’ binding to PC1 and degree of target inhibition, with a steeper dose–effect relationship of IACS-10759 than for IM156.

Two patients, 1 each at 800 and 1,200 mg QD, had Grade 1 asymptomatic elevation in blood lactate. Increased blood lactate is potentially an on-target effect of IM156; one could hypothesize that inhibition of mitochondrial OXPHOS would increase the rate of glycolysis with a subsequent rise in lactate production; however, lactate elevation is nonspecific.

Pharmacokinetic data demonstrated dose proportionality for both C_max and exposure at doses ≥ 200 mg. IM156 exposures following QD dosing were predicted from modeling of the data obtained with QOD dosing, supporting the switch to QD administration. Plasma exposures of IM156 at 800 mg QD demonstrated modest accumulation at steady-state (Day 1 C_max 829 ± 260 ng/mL, AUC_0-24 7,520 ± 3,820 h*ng/mL; Day 27 C_max 1,870 ± 485 ng/mL; AUC_0-24 33,600 ± 9,780 h*ng/mL) (Table 4). These values are in the range of those expected to demonstrate response based on robust efficacy in preclinical in vivo mouse models dosed at 15 – 30 mg/kg (Immunomet, Data on file); the AUC_0-24 in single-dose oral PK studies in mice administered a 30 mg/kg dose was approximately 3.0 – 3.5 h*ng/mL. Our recent publication reported that IM156 is highly distributed to major metabolic organs such as lung, liver, kidney, etc. and that tissue concentrations of IM156 are 30–80 fold higher compared to plasma [21]. As such, measurements of the concentrations and activities of IM156 in plasma are not informative and significantly underestimate target tissue exposures and associated activity. In short, measurement of circulating plasma IM156 levels is a poor surrogate for target tissue exposures. A recently completed Phase 1 study in healthy volunteers using appropriate tissue-based biomarkers was designed to further explore and confirm the PK and target engagement of IM156 (Immunomet, Data on file).

IM156 metabolic profiling was completed for human samples following oral QOD administration of 400 mg of IM156 to cancer patients. One major metabolite detected in the pooled human plasma samples. Similarly, this metabolite was identified as a major circulating metabolite in rat and dog following oral administration of a single- or multiple-dose of IM156 in preclinical toxicity studies. This metabolite was negative in an in silico computational prediction of genetic toxicity, and no inhibition were greater compared to IM156 against off-target inhibition safety pharmacology profiling panel. Finally, this metabolite did not inhibit OXPHOS with no effect on cell viability up to concentration of 50 μM. These data demonstrate that this metabolite showed limited biological activity in mechanism of action and safety profile.

No clinical responses were observed but 32% of patients had SD as the best response, including 1 with gastric neuroendocrine carcinoma and 1 patient with gastric adenocarcinoma who remained on IM156 for 444 and 169 days, respectively. In an unselected population of patients whose cancer has not demonstrated dependence on OXPHOS, this level of anticancer activity is not supportive for further development as a monotherapy. However, IM156’s favorable safety profile suggests the potential for further development in rationally designed combinations with other anticancer agents, and as monotherapy for selected patients whose tumors have molecular profiles demonstrate a therapeutic vulnerability to OXPHOS inhibition [22].

Similar to IM156, the efficacy results of IACS-010759 showed one (5%) PR in a patient with prostate cancer and 8 (38%) patients with SD [19]. A switch to glycolysis has been speculated as a plausible mechanism for insufficient single-agent activity of other small molecule inhibitors of PC1. An adaptive switch to glycolysis in a chronic lymphocytic leukemia mouse model treated with IACS-010759 suggested that simultaneous use of a glycolysis inhibitor may be required to demonstrate efficacy [23], a combination that, due to the essential nature of the OXPHOS, may plausibly lead to significant toxicity.

In summary, our study demonstrated that IM156 is tolerable at RP2D of 800 mg QD. These data suggest that clinical development of PC1 inhibitors may be feasible. However, IM156 demonstrated rather limited single agent clinical activity in the unselected population, and we did not identify any biomarker associated with therapeutic response. Therefore, the study did not advance into monotherapy expansion cohorts at the RP2D. Further development in oncology will focus on rational combinations with other anticancer agents such as targeted therapies or chemotherapy and on continuing efforts to identify molecular markers predicting anticancer activity of IM156.