What is the prospect of indoleamine 2,3-dioxygenase 1 inhibition in cancer? Extrapolation from the past

With increasing recognition of evading immunological destruction as a hallmark of cancer acquired during its development[1], immunotherapy has emerged as a novel and important pillar of cancer treatment in addition to surgery, radiation, chemotherapy, and targeted therapy. Instead of directly killing cancer cells or suppressing the abnormal signal transduction within cancers, cancer immunotherapy aims to leverage the patient’s immune system to eliminate tumor cells by enhancing tumor immunity mediated by blocking immune inhibitory pathways and/or inhibitory cells in the tumor microenvironment (TME) or enhancing the specificity of anti-tumor immunity by inducing the expansion of T cells and antibodies directed to well-defined tumor antigens[2]. The launch of immune checkpoint inhibitors (ICPIs) including the first cytotoxic T-lymphocyte antigen 4 (CTLA-4) blocking antibody ipilimumab in 2011[3] and programmed death receptor 1 or ligand 1 (PD-1/PD-L1) blocking antibodies pembrolizumab and nivolumab in 2014[4, 5] represents the substantive progress in the field and boosts immunotherapy as a new standard therapy for many cancer types. To date, the US Food and Drug Administration has approved the use of PD-1/PD-L1 and CTLA-4 targeted therapies for the treatment of more than 15 cancers including melanoma, non-small cell lung cancer (NSCLC), small cell lung cancer, squamous cell carcinoma of the head and neck, renal cell carcinoma (RCC), hepatocellular carcinoma (HCC), classical Hodgkin lymphoma, urothelial carcinoma, colorectal cancer (CRC) and other cancer types[3‐5]. However, limited response due to innate and acquired resistance as well as serious adverse effects especially life-threatening immune-related toxicities of anti-CTLA-4 and anti-PD-1/PD-L1 antibodies call for a need to identify other immunotherapy options targeting different mechanisms of immunosuppression in tumors that could be used alone or in combination with existing treatments. Currently over three thousand drugs directing against more than four hundred active targets are being investigated under different stages from preclinical to post-approval[6]. In this context indoleamine 2,3-dioxygenase 1 (IDO1) is one of the most studied target of cancer immunotherapies by virtue of its effects on immune suppression in the TME[6].

IDO1, a monomeric heme-containing enzyme, is one of the three enzymes that catalyze the first and rate-limiting step in the kynurenine pathway (KP), leading to the degradation of the essential amino acid tryptophan (TRP) and generation of kynurenine (KYN) and other downstream metabolites, in addition to tryptophan 2,3-dioxygenase (TDO) and indoleamine 2,3-dioxygenase 2 (IDO2)[7, 8]. TRP is required for protein synthesis and TRP metabolites can act as neurotransmitters and signalling molecules. Thus, TRP and its metabolites play important roles in diverse physiological processes ranging from cell growth and survival to the coordination of responses to internal and external environmental changes[9]. IDO1 is expressed in various tissues, organs and cell types, catalyzing the conversion of a wide range of substrates including TRP and 5-hydroxytryptamine[10, 11]. TDO mainly exists in the liver and brain[12, 13], specifically catalyzing the conversion of TRP and some of its derivatives substituted in the 5- and 6-positions of the indole ring[14, 15]. TDO is thought to be the main modulator of TRP catabolism and is responsible for regulating systemic TRP levels[16], while IDO1 plays an important role in TRP catabolism under pathological conditions[17, 18]. The host-protective effect of IDO1 in parasitic infection has been first discovered in 1990s, which is attributable to a reduced availability of TRP in the inflammatory environment[19, 20]. In addition, as many KP metabolites are neuroactive, dysfunction of KP enzymes including IDO1 and TDO, often caused by inflammatory insults can trigger or facilitate diseases of the central nervous system, such as depression, Alzheimer’s disease, and Huntington’s disease[21‐23]. IDO1 was first discovered to have a potent role in immune tolerance between a mother and fetus, since the inhibition of IDO1 in pregnant mice caused spontaneous immune rejection of allogeneic fetuses[24]. Later, the IDO1-mediated immune tolerance in tumors has been noted, experimental models suggest that IDO1 expression prevents rejection of tumor cells in immunogenic mice[25‐30]. The imbalances in the level of TRP and KYN as well as the overexpression of IDO1 have been reported in many cancers[25‐29, 31‐47]. High IDO1 expression was associated with a decreased survival[48] and the extent of IDO1 overexpression depends on the tumor type and risk factors[49]. Three downstream effector pathways, including the general control nonderepressible 2 (GCN2), mammalian target of rapamycin (mTOR), and aryl hydrocarbon receptor (AhR) pathways, have been implicated in the biological responses to IDO1-mediated TRP depletion or KYN accumulation. In addition, some studies have shown that these effector pathways mediate immunosuppression in the TME[50‐53]. TRP depletion promotes the accumulation of uncharged tRNA, resulting in a GCN2-dependent inhibition of protein synthesis that is accompanied by cell cycle arrest and irresponsiveness to immunological challenges[54]. TRP depletion leads to inhibition of the immunoregulatory kinases mTOR and protein kinase C, along with the induction of autophagy[55]. KYN accumulation promotes the translocation of AhR from the cytosol to nucleus, where it binds target genes and activates their transcription, inducing tolerogenic immune responses with the consequent benefit to tumor progression and migration[53, 56]. The exact contribution of each effector function to tumor immunosuppression remains to be determined. And it is likely that each pathway counts for much or less in different tumors although all of them cooperate synergistically in the development of immunosuppression in the TME. Based on these data, the inhibition of IDO1 has become an exciting approach as cancer immunotherapy and IDO1 inhibitors have been intensively investigated during the recent years. Multiple IDO1 inhibitors with different structural skeleton have been developed[57‐67] and some have entered clinical development (Table 1). However, the further development of IDO1 inhibitors has experienced a significant setback due to the recent failure of the phase III trial (ECHO-301/KEYNOTE-252) of epacadostat (INCB024360), the would-be first-in-class IDO1 inhibitor, which had been anticipated to receive regulatory approval on the basis of promising results from initial studies (Fig. 1) [24, 50, 68‐72].

Several reviews discussing the biochemical properties and physiological functions of IDO1 and the therapeutic applications of IDO1 inhibition in cancer and other diseases have been published [8, 16, 73‐75]. The aims of this article are, (i) to introduce the current status of clinical testing of IDO1 inhibitors (ii) to review representative data of epacadostat and (iii) to discuss some open questions raised from the failure of epacadostat in advanced malignant melanoma.

Currently, as a focus of research and drug discovery efforts, there are at least eight small molecule IDO1 inhibitors under investigation in clinical trials indicated by the clinical trial registry website ClinicalTrials.gov. These include epacadostat (INCB024360; developer: Incyte), indoximod (D-1MT; developer: NewLink Genetics), navoximod (NLG919; developer: NewLink Genetics), BMS-986205 (Ono-7701/linrodostat; developer: Bristol-Myers Squibb/Ono), PF-06840003 (EOS-200,271; developer: Pfizer), KHK-2455 (developer: Kyowa Hakko Kirin), LY-3381916 (developer: Eli Lilly) and DN1406131 (developer: Jiangxi Qingfeng).

The first clinical trial of IDO1 inhibitor was started in 2007, which is the first-in-human (FIH) trial of indoxiomd with the purpose of assessing its safety and pharmacokinetics profile in patients with advanced solid tumors (NCT00567931). Since then to the cutoff date of 16 June 2020, a total of 98 clinical trials aiming to characterize the safety and efficacy profile of different IDO1 inhibitors have been initiated and the majority of them (88/98, 90 %) focuses on the exploration of treatment combinations (Table 1). Because one trial may have been tested in multiple cancer types, total 104 studies of 98 clinical trials have been carried out (Table 2). Seventy studies have been performed in 6 cancer types including lung cancer (16/70), head and neck cancer (12/70), gynecological tumors including ovarian and endometrial carcinoma (12/70), malignant melanoma (11/70), urothelial carcinoma (11/70) and brain tumors especially glioblastoma (8/70). Twenty studies did not specifically indicate the target cancer type or enrolled healthy volunteers only, and the rest studies have been performed in 15 cancer types including RCC, HCC and CRC (Table 2). In all studies, ICPIs, in particular anti-PD-1/PD-L1 antibodies were the most popular class selected as a combination agent, while others include chemotherapy, cancer vaccines, and other immunotherapies targeting CTLA-4, lymphocyte-activation gene 3, C-C motif chemokine receptor type 4 are also under investigation.

An analysis of the trial initiation (i.e. First posted date disclosed in ClinicalTrials.gov) over time (Fig. 2) shows that the clinical development of IDO1 inhibitors started from 2007, and the expansion significantly accelerated after 2014 with its peak in 2017. In 2017, the number of all initiated trials has grown to 27, representing a 125 % increase from 2016 to 2017 and the proportion of phase III trials has grown from 8 % (1/12) to 30 % (8/27) from 2016 to 2017. However, a rapid fade of interest in IDO1 inhibitors came up in 2018, the year that the failure of ECHO-301/KEYNOTE-252 was announced[72].

Epacadostat is an IDO1 selective inhibitor, which is the most advanced in the clinical development and would be supposed to be the first IDO1 inhibitor to get registration approval. Epacadostat showed preliminary promising anti-cancer activity in its early phase I/II trials (ECHO-202/KEYNOTE-037 and ECHO-204) when combined with anti-PD-1 drugs including pembrolizumab and nivolumab in patients with advanced malignant melanoma[76, 77]. However, its clinical value was not further validated in subsequently pivotal phase III trial (ECHO-301/KEYNOTE-252)[78] when the number of enrolled patients increased – the addition of epacadostat to pembrolizumab failed to improve progression-free survival (PFS) in patients with advanced malignant melanoma compared to pembrolizumab monotherapy.

Consequently, the unexpected setback of epacadostat has had a major impact on the entire field and tremendously slowed the pace of clinical development of all other IDO1-targeting compounds. In turn, total 27 trials involving 5 compounds, including epacadostat, BMS-986205, indoximod, PF-06840003 and LY-3381916, have been terminated, suspended or withdrawn and none of the other phase III programs of epacadostat is enrolling patients as originally planned. Just as every coin has its two sides, this failure also contributes to the understanding of the role of IDO1 inhibition and changes the clinical development program of inhibitors. Since the second half of 2018, seventeen new trials have been initiated and a large proportion of the clinical development is taking place in phase I/II, with only one agent-BMS-986205 proceeded to Phase III trial in bladder cancer (Fig. 2). Of note, the target patient populations in these new trials were narrowed down to certain tumor types, such as head and neck cancer, glioblastoma and bladder cancer. This may reflect that IDO1 is still regarded as a potential target to break the immune suppressive TME and restore the immune surveillance system. The critical challenges would be what we can learn from the failure and how to address the outstanding questions.

In preclinical studies, epacadostat selectively and competitively inhibited the catalytic activity of IDO1 in multiple cell-based assays with little activity against IDO2 and TDO2[65, 79](Table 3). In co-cultures of human allogeneic lymphocytes with dendritic cells (DCs) or tumor cells, epacadostat promoted the growth of effector T cells and NK cells, reduced the conversion of naïve T cells to regulatory T cells (Tregs) and increased the number of CD86^high DC[79]. Along with these, administration of epacadostat to CT26 tumor-bearing mice reduced tumor growth with up to 57 % tumor growth control (TGC) in a dose-dependent and lymphocyte-dependent fashion and similarly decreased kynurenine levels in tumor tissues, plasma and draining lymph nodes ranging from 78–87 %[65]. As the most advanced IDO1 inhibitor, epacadostat was selected as a positive control in several preclinical studies of other IDO1 inhibitors in some animal models and showed similar IDO1 inhibitory activities and variable anti-tumor activities (Tables 3 and 4). Meanwhile, a synergy of INCB23843, an analogue of epacadostat, with anti-PD-L1 or anti-CTLA-4 antibodies, respectively, was investigated in B16 melanoma models. The combination with either anti-PD-L1 or anti-CTLA-4 antibody could suppress tumor growth more effectively than either agent alone, primarily through reactivation of anti-tumor immunity[80]. The combination of epacadostat with a MerTK inhibitor or an agonistic CD40 antibody also exhibited positive anti-tumor activities but epacadostat alone not[81, 82]. Furthermore, anti-PD-1 antibody, anti-CTLA-4 antibody and agonistic CD40 antibody therapies have been found to induce IDO1 expression[82, 83], which might be one reason for the limited response rate of these therapies. Therefore, the combination of IDO1 inhibitor with these therapies to defeat cancer has been paid high attention.

With the support of the results generated from the above summarized preclinical data, the activity of epacadostat as single agent was evaluated in a FIH phase I trial (NCT01195311) and subsequently in a phase II trial (NCT01685255). In the FIH phase I trial, the safety, pharmacokinetics (PK), pharmacodynamics (PD) and preliminary anti-tumor efficacy profile of epacadostat were characterized in 52 patients with advanced solid malignancies. The results showed that the safety profile of epacadostat was manageable and it can be generally well tolerated at doses of up to 700 mg twice daily (BID). The PK of epacadostat was characterized by a time of maximum concentration at around 2 hours, a dose-independently biphasic disposition with an apparent terminal-phase disposition half-life of 2.9 hours and a relatively longer effective half-life of 4 to 6 hours in the context of systemic accumulation following BID dosing[91]. Regarding PD assessments, in addition to the direct evaluation of the plasma concentration of TRP and KYN in patients at different time points, the levels of TRP and KYN in patients blood pretreated with interferon gamma (IFN-γ) or lipopolysaccharide (LPS) were also determined. IFN-γ or LPS was used to induce IDO1 expression and minimize the confounding effect of TDO. Tracking KYN levels in plasma has been considered as a sufficient method for assessing IDO1 activity and selected as a valuable pharmacodynamic marker in the clinical studies of epacadostat[65]. The percentage of IDO1 inhibition was finally defined as the reduction in plasma KYN levels from the predose to the postdose value. It was found that epacadostat could reduce KYN levels at doses ≥ 100 mg BID and a 90 % inhibition was achieved when the plasma concentration of epacadostat was greater than 500 nmol/L[91]. Epacadostat’s on-target potency, i.e., its concentration exerting 50 % of the maximal inhition effect (IC₅₀) against IDO1, is important for dose selection but complicated by the bioconversion of TRP to KYN catalyzed by both IDO1 and TDO. In vitro and ex vivo, the IC₅₀ was estimated following the selective induction of IDO1, rendering the TDO activity relatively insignificant; however, it was desirable to determine the in vivo IC₅₀ without inducing IDO1. Based on the data from the FIH trial, a mechanism-based population PD model was developed, and epacadostat in vivo IC₅₀ was estimated to be ~ 70 nM, consistent with the ex vivo value independently determined. And this model also suggests that approximately 60 % of TRP to KYN bioconversion was attributed to IDO1 in the cancer patients at baseline, the mean concentration of plasma KYN was 2–3 % of that of TRP, and no significant changes in TRP concentrations were observed before or during epacadostat treatment[92]. Based on PK and PD data, 300 mg BID was selected as the recommended phase II monotherapy dose to provide sufficient drug exposure that would inhibit > 90 % of IDO1 activity[91]. The therapeutic efficacy profile from this FIH trial suggested a limited anti-tumor activity of epacadostat in human, only stable disease (SD), which is one of response category[93], was observed as the best overall response in 18 of 52 patients, and SD lasting ≥ 16 weeks was seen in 7 patients [91].

The following phase II trial aiming to compare the efficacy of epacadostat versus tamoxifen as therapy for biochemically recurrent (CA-125 relapse)–only epithelial ovarian cancer reported similar results. Epacadostat was generally well tolerated, but its efficacy was not satisfactory enough, i.e. no significant difference in efficacy was found between epacadostat and tamoxifen, and thus the trial was terminated accordingly[94]. These results are not that unexpected, since multiple mechanisms are involved in the evasion of tumors from immune surveillance and IDO1 inhibition strategy alone may be insufficient. Given that the synergy of IDO1 inhibitor and ICPIs was confirmed in B16 melanoma models[80], several phase I/II clinical trials were then initiated to investigate epacadostat in combination with ipilimumab[95] (NCT01604889), pembrolizumab[76] (NCT02178722) and nivolumab[77] (NCT02327078) for the treatment of unresectable or metastatic malignant melanoma.

In the phase I/II trial investigating epacadostat in combination with ipilimumab (NCT01604889), epacadostat ≤ 50 mg BID was demonstrated to bear clinical and pharmacologic activity and was generally well tolerated in patients with advanced melanoma[95]. Objective responses were observed in 9 of 39 immunotherapy-naive patients (23 %), while no objective response was seen in the rest 11 patients who previously received immunotherapy although 3 of them (27 %) had SD. Using the same method and definition of IDO1 inhibition as in the previous FIH trial, PD data from this trial showed that a dose-dependent inhibition of IDO1 by epacadostat and the average IDO1 inhibiton exceeded 50 % at doses of 25 mg BID and 50 mg BID[95]. These results were generally consistent with previously reported data in patients with advanced solid malignancies[91]. The emerging success of anti-PD-1 antibodies for the treatment melanoma led to an early termination of this trial (NCT01604889, only phase I part was conducted) and shift more attention to exploring epacadostat in combination with anti-PD-1 antibodies.

ECHO-202/ KEYNOTE-037 (NCT 02178722) was a phase I/II trial aiming to evaluate epacadostat plus pembrolizumab in patients with advanced solid tumors. The results from its phase I part showed that the combination regimen of epacadostat and pembrolizumab was well tolerated and a maximum tolerated dose (MTD) of epacadostat did not reach within pre-defined dose levels up to 300 mg BID; ≥50 % time-averaged IDO1 inhibition (i.e. the level of pharmacodynamic activity associated with inhibition of tumor growth seen in nonclinical models) determined with the above mentioned PD model was yielded in all patients treated with epacadostat 100 mg BID or 300 mg BID; objective responses were observed in 25 of total 62 patients (42 %) and 12 of 22 patients with advanced malignant melanoma (55 %)[76]. The phase II trial of epacadostat in combination with nivolumab, i.e. ECHO-204 (NCT02327078) having similar study design and patient population demonstrated consistently promising anti-tumor activity with objective response at 62 % and tolerable safety profile in 50 patients with advanced malignant melanoma[77]. On the basis of these results, epacadostat 100 mg BID was selected to combine with anti-PD-1 antibodies for additional investigation in phase III trials including ECHO-301/KEYNOTE-252 (NCT02752074)[78].

A non-randomized trial using a small number of patients is prone to biased selection that may confound the result and be misleading. Thus, a further confirmation with a randomized phase II trial as gatekeeper before entering into pivotal phase III trial might be necessary. However, regardless of the lack of evidence from at least one randomized, active controlled (anti-PD-1 antibody monotherapy) phase II trial as a validation step, the phase III trial ECHO-301/KEYNOTE-252 was initiated to further clarify the clinical benefit of epacadostat plus pembrolizumab compared to pembrolizumab alone in patients with advanced malignant melanoma previously untreated with a ICPI using PFS as primary endpoint. In this trial, the testing of IDO1 expression status by use of in situ hybridization RNA scope technology was retrospectively completed after patients were enrolled, but was not taken into account in the treatment assignment or primary analysis of the endpoints. Among all enrolled 706 patients, IDO1 expression was positive in 451 (90 %) of 502 patients with evaluable tumor specimens, and the IDO1 expression positivity was determined as higher than 1 % of tumor or immune cells expressed IDO1. The positive expression of IDO1 in tumor cells, intratumoral immune cells and both kinds of cells was found in 393 of 502 samples (78 %), 422 of 472 samples (89 %) and 364 (77 %) of 472 samples, respectively. As of data cutoff, surprisingly, the trial failed to prove the clinical value of epacadostat in terms of PFS, overall survival (OS), and objective response with a median PFS of 4.7 months (95 % CI 2.9–6.8) in the epacadostat plus pembrolizumab group and 4.9 months (2.9–6.8) in the placebo plus pembrolizumab group (HR 1.00, one-sided P = 0.52). Furthermore, the absence of a significant PFS benefit for epacadostat was evident in all prespecified and post-hoc subgroups examined including IDO1 status (Table 5).

The failure of ECHO-301 has been discussed in terms of the sufficiency of the dose of epacadostat and drug combination selection[96, 97]. Furthermore, the signaling activity of IDO1 besides its catalytic activity may be attributed to another reason for the failure[97]. It has been described that the immunoregulatory activity of IDO1 is related to its signaling activity in addition to its catalytic activity[98, 99]. It could be the case and would offer a hypothesis that some IDO1 inhibitors, which have been designed for blocking the catalytic activity of IDO1 only, may not in reality block its signaling activity and thus could not be sufficient to “neutralize” the effects of IDO1. Although the failure of ECHO-301 trial upset the whole field as the benchmarker of IDO1 inhibitors, all preclinical and clinical data from epacadostat discussed above is like a treasure for the successors to excavate and get potential inspiration for future direction.

Despite the failure of epacadostat to demonstrate clinical benefits in the pivotal phase III trial, there is much to be learned from the available data, which is likely to ultimately benefit the field. It is too arbitrary and premature to deny the future of the entire field as the result of a single trial failure in the context of some reasonable doubt raised accordingly.