Therapeutic targeting of isocitrate dehydrogenase mutant AML

KEYWORDS:

• Isocitrate dehydrogenase
• IDH inhibitors
• AG221
• AG120
• AG881
• ketoglutarate
• hydroxyglutarate
• acute myeloid leukemia

1. Introduction

The last decade has witnessed rapid advances in our understanding of molecular leukemogenesis providing valuable insights from which to explore new therapeutic avenues. Characterization of the mutational landscape in acute myeloid leukemia (AML) has enabled the identification of recurrent, prognostically informative mutations against which to develop targeted therapies. The therapeutic potential of highly specific small molecule inhibitors, such as against FLT3 and IDH, has paved the way for individualization of treatment strategies in an effort to improve upon outcomes in higher risk, biologically distinct AML subsets, an increasing reality as we march forward in the era of personalized medicine.

Figure 1. Mutant IDH enzymes convert α-KG to 2-HG with consequent widespread dysregulation in epigenetic machinery. Altered enzyme activity is indicated with curved red arrow. CB839- glutaminase inhibitor, venetoclax-BCL2 inhibitor, AG221, AG120, IDH305- mutant IDH inhibitors. Full color available online.

IDH exists in three isoforms IDH1, IDH2, and IDH3, of which mutations have only been identified in the former two. The IDHs are tricyclic carboxylic acid cycle enzymes which convert isocitrate to α-keto glutarate via oxidative decarboxylation, producing NADPH from NADP⁺ (IDH1 and 2) and NADH from NAD⁺ (IDH3). IDH1 and 2 are single-gene proteins and are homodimers while IDH3 is a heterotetramer and a multigene protein. While IDH3 is primarily involved in generating NADH for energy production, generation of NADPH by IDH1 and 2 may function toward limiting oxidative damage through reduction of glutathione and thioredoxins. Mutations in genes encoding IDH1 and IDH2 gene isoforms occur in 6–16% and 8–19% of AML cases, respectively [1]. These mutations are enriched in AML with normal cytogenetics, but may be present along with other cytogenetic abnormalities in 30% of the cases. The frequency of mutations in these genes increases with age at diagnosis. Published reports on the prognostic impact of IDH1/2 mutations in AML have thus far been conflicting [2,3]. Early data suggested adverse impact of IDH mutations in AML [4] while recent data from a large cohort suggest that site of mutation and cooccurrence with other mutations influences outcome [5]. Cooccurrence of IDH2^R140 with DNMT3a mutation confers worse prognosis while IDH2^R172 mutation, though uncommon, is associated with favorable outcome. Again while survival is better in IDH1/2-mutated AML with concurrent NPM1 mutations, it compares less favorably against cytogenetically normal AML with NPM1 mutation alone [3,4,6]. Biochemical elucidation of disease mechanisms has implicated the generation of an intermediary ‘oncometabolite’, 2-hydroxyl glutarate (2-HG), to play a key contributory role in leukemic tumorigenesis [7]. In this context, single allele mutations occurring at critical amino-acid residues of the IDH enzymatic active binding site (R132 in IDH1, and R140 and R172 in IDH2) confer a gain of function with neomorphic enzymatic activity. Interestingly, these mutations must occur in a heterozygous form to be able to generate mutant heterodimers, the wild-type enzyme producing α-KG, and the mutant subunit monomer swiftly reducing α-KG to its structurally similar functional antagonist, 2-HG. 2-HG actively binds and competitively inhibits a number of αKG‐dependent dioxygenases, including histone demethylases and 5-methylcytosine hydroxylases, which are important regulators of the oxygen sensing machinery and epigenome [8]. In this context, IDH and TET2 mutations are mutually exclusive (cooccurrences reported in exceptional cases) in AML and attest to their overlapping biological effects in leukemogenesis [1,9]. Epigenetic profiling has revealed IDH1/2-mutated AML to display a global histone- and DNA-hypermethylation phenotype characterized by an increased methylation of promoter sites of target genes associated with myeloid differentiation thereby creating a maturation block [9].

In this editorial, emphasis is laid upon mIDH inhibitors currently undergoing active development in clinical trials. We also incorporate a section outlining agents targeting other simultaneously dysregulated pathways uniquely operational in IDH-mutated AML.

1.1. Specific inhibitor of mutant IDH

Preclinical experiments with AGI-5198 (targeting IDH1R132 H) and AGI-6780 (targeting IDH2R140Q) demonstrated these two small molecule inhibitors to reduce 2-HG to physiologic cellular levels and promote differentiation of tumor cell lines by reversing aberrant epigenetic remodeling, thus providing a proof of concept validating mutant IDH as a clinically tractable target [10].

First-generation inhibitors of mutant IDH1 (AG120, IDH-305, FT-2102), IDH2 (AG221), and a second-generation pan-inhibitor (AG881) are currently undergoing clinical development in phase I trials (Tables 1 and 2). Unlike other chemotherapeutics, these drugs are not cytotoxic and thus expected to be associated with lower rates of cytopenic complications.

Table 1. Interim results of Phase 1 studies on AG120, AG221, and IDH305.

Drug	Study phase	Disease condition	Study number of patients	Reported response rates	Toxicity profile (deaths)	Comments
AG120	Phase 1	IDH1-mutant positive advanced hematologic malignancies	N = 78 (63 with R/R AML)	ORR = 33%, CR = 16%	30-Day mortality: 12%, 60-day mortality: 21%	Safe with minimal toxicities. Durable response durations of a median of 10.2 months
IDH305	Phase 1	Advanced malignancies including R/R AML/MDS with IDH1^R132 mutation	N = 81; AML/MDS = 24	ORR = 7 (33%) AML; CR in 2 (9.5%)	Not reported	Very favorable safety profile. Durable response durations in responders
AG221	Phase 1/2	mIDH2- AML	R/R-AML n = 138 (70%); untreated AML/MDS/other 30%	74 (41%); RR-AML = 52 (41%)	Not reported	Safe with few side effects, including indirect hyperbilirubinemia and nausea. Good response rates in IDH2 AML.

ORR: objective response rate; CR: complete remission; AML: acute myeloid leukemia; MDS: myelodysplastic syndromes.

Table 2. Ongoing Phase I/II studies evaluating AG-221 and AG-120 either as monotherapy or in combination regimens in AML.

Drug	Mutation	Disease condition	Study arm intervention	Outcome measures	Phase, current status	NCI identifier
AG221	IDH2	R/R AML + IDH2 > 60 years	AG221 vs. aza+BSC/LDAC+BSC/IDAC+BSC	Efficacy-OS	3, recruiting	NCT02577406
AG221	IDH2	R/R AML + IDH2 > 18 years not eligible for intensives	AG221	Safety, tolerability, maximum tolerated dose	1/2, active	NCT01915498
AG120	IDHI1	R/R AML + IDH1 > 18 yrs	AG120	Safety, tolerability, maximum tolerated dose	1, recruiting	NCT02074839
AG120	IDHI1	R/R AML + IDH1 > 18 yrs not eligible for intensives	AG120+azacitidine vs. azacitidine	Efficacy-ORR	2, recruiting	NCT02677922
AG120 & AG221	IDH1/2	Newly diagnosed AML with IDH mutation; in combination with standard induction and consolidation therapies	AG120+ standard induction (cytarabine/daunorubicin) and consolidation; AG221+ standard induction (cytarabine/daunorubicin) and consolidation	Safety, tolerability	1b/2, recruiting	NCT02632708

AML: acute myeloid leukemia; OS: overall survival; ORR: objective response rate; BSC: best supportive care; LDAC: low dose ara-C; IDAC: intermediate dose ara-C.

Enasidenib (AG-221/Agios/Celgene) is a highly selective, first-in class IDH2 (R140Q and R172) inhibitor that is currently being investigated in clinical trials after showing potent in vivo efficacy in murine IDH2 mutant leukemia models [11]. Treatment with AG-221 reverses aberrant hypermethylation in multiple genes known to be involved in hematopoietic differentiation. A phase 1/2 dose-escalation and expansion study of AG-221 in patients with advanced hematological malignancies is currently ongoing (NCT01915498). Interim study results, last presented at the Annual Society of Hematology Meeting in 2015, reported objective response rates in 41% of patients (n = 74/181 efficacy evaluable patients), including 52/128 (41%) evaluable R/R AML patients [12]. Of the 128 patients, 33 (26%) patients achieved complete remission (CR) and 19 (15%) patients had a partial remission (PR). Additionally, 57 (45%) patients had a clinical stable disease (SD), defined by stable or decreased blast percentage in the bone marrow. Median duration of therapy for R/R AML patients was 6 months with the maximum tolerable dose, not reached. The drug was tolerable with the most common grade 3 treatment-related adverse events being indirect hyperbilirubinemia and nausea. Response rates were consistent across subgroups irrespective of number of prior treatment regimens or mIDH2 type (R140Q vs. R172K). Importantly, it was observed that mIDH2 variant allelic frequency did not decrease with response or ANC improvement suggesting that eradication of the mutant clone was not a prerequisite for response, and that AG-221 effectuated responses, not by eliminating the malignant clone, but through blast differentiation. Evaluation of AG-221 as monotherapy and in combinatorial regimens in phase II/III trials have begun.

AG-120 is a first in-class, oral, selective mIDH1 inhibitor undergoing active study in phase I trials involving advanced hematological malignancies. Safety and efficacy data, reported as of December 2016, found the drug to be well tolerated with the majority of adverse events being mild to moderate, the most common being fatigue, nausea, diarrhea, and pyrexia. Among the 63R/R AML (of 78 hematological malignancy) patients, 21 (33%) had objective responses (CR = 10, CRi/CRp = 8, marrow CR = 2, PR = 1). Importantly, mIDH1 clearance by next-generation sequencing was observed in 36% of CRs and 4% of non-CRs, the first demonstration that this single agent could result in mIDH1 clearance [13]. Dose escalation continues and evaluation in expansion cohorts and in combination trials has been initiated.

IDH305 is an oral, selective IDH1 inhibitor that specifically targets R132* IDH1 mutation. The drug is currently being evaluated a phase I study involving advanced malignancies including R/R AML/ myelodysplastic syndromes (MDS) (NCT02381886). As of the data cutoff in March, 2016, 81 patients have been enrolled, 24 patients with AML/MDS (MDS, n = 3) [14]. Objective responses were reported in seven (33%) AML patients: CR/CRi in three (14.3%), and partial remission in four (19.0%) patients. Across all diseases, the most common grade ≥3 adverse events reported as suspected of being related to IDH305 included bilirubin and transaminase elevations, and nausea. Dose escalation for AML/MDS continues, with safety, and antitumor activity of IDH305 as a single agent and in combinatorial regimens undergoing further evaluation.

FT-2102 is a mIDH1 inhibitor under investigation in a phase I dose-finding study as a single agent and in combination with azacitidine in patients with AML or higher-risk MDS or who are refractory to prior treatment or ineligible for standard intensive therapy (NCT02719574). At the time of writing this manuscript, no clinical data associated with this agent havebeen reported.

AG881 is a potent pan-IDH inhibitor and fully penetrates the blood–brain barrier [15]. It is currently under investigation in a phase 1, open-label, dose-escalation and expansion study in patients with advanced hematological malignancies that had progressed after prior mIDH inhibitor therapy (NCT02492737).

1.2. Other modes of inhibitory activity in IDH AML

1.2.1. Venetoclax (ABT-199)

Suppression of cytochrome c oxidase in the electron transport chain, by elevated 2-HG levels, has been shown to lower mitochondrial threshold for triggering apoptosis through BCL-2 inhibition, such as by ABT-199 (Figure 1). This is further supported by data from clinical trials investigating venetoclax as a single agent or in combination regimens in R/R AML. Patients with IDH mutations are more responsive to venetoclax suggesting that IDH mutation may identify a patient subgroup more responsive to this class of agents [16].

1.2.2. CB-839

In contrast to normal cells, tumor cells are highly dependent on glutamine to fuel the tricarboxylic acid cycle and promote cellular growth and survival. IDH mutations induce a glutamine addiction by inducing a dependency in cancer cells for glutamine as the main source of α-KG to produce 2-HG [17]. This glutamine dependence is being exploited as a therapeutic option with CB-839, a potent glutaminase inhibitor, which is currently being tested in a phase 1 study in patients with AML (NCT02071927). One of the study end-points in this trial includes assessing response in IDH-mutated AML; the trial is currently recruiting patients.

1.2.3. ATRA

2-HG production skews hematopoietic differentiation favoring myelomonocytic lineage while also sensitizing mutant cells to ATRA-induced differentiation. In this context, Lyn is a negative regulator of ATRA-induced granulocytic differentiation [18]. The sensitizing effect is brought about by the reduced Lyn expression in the presence of IDH1R132 H mutation. Since ATRA induces a transient Lyn activation, thereby hindering its own activity, Lyn inhibitors such as dasatinib may prove synergistic when used with ATRA [19]. This therapeutic strategy holds promise, but is yet to be tested in clinical trials.

1.2.4. BRD4 inhibitors

Brd4 is an epigenetic ‘reader’ protein and IDH-mutated AML appears to be addicted to Brd4-driven Myc expression, and in preclinical models Brd4 inhibitors can produce differentiation and cell death of IDH 1- and 2-mutated AML cells [20]. Clinical data regarding activity of these agents in IDH-mutated AML are not available yet.

1.2.5. PARP inhibitors

IDH1/2 cell mutants are defective in homologous recombination (HR)-mediated DNA double-strand break repairs, a product of DNA exposure to 2-HG. This HR defect however renders mutant cells exquisitely sensitive to poly adenosine 5ʹ-diphosphate-ribose (PARP) inhibition [21]. The role of PARP inhibitors in IDH-mutated AML is yet to be explored in clinical trials.

Hypomethylating therapies: Given that IDH1/2-mutated AML cells demonstrate global hypermethylation profile, treatment with hypomethylating agents (HMA) may be effective against this AML subgroup. A retrospective case series evaluating IDH-mutated AML patients did not demonstrate any further therapeutic benefit from HMA over other therapies [22]. However, the clinical role of HMA in AML patients carrying IDH mutations has not been prospectively tested.

2. Expert opinion

Early clinical data emerging from phase I trials suggest mIDH inhibitors to have an impressive clinical activity with minimal toxicity. Response rates with single agent therapies in IDH-mutated AML range from 31% to 41%. It can be argued that targeting single mutation defects with monotherapies will achieve only part of the therapeutic purpose. In this regard, phase 2 trials implementing IDH inhibitors in combination with HMA are currently recruiting patients, and results are eagerly awaited. Apart from inducing complete and partial remissions, a sizeable proportion of AML patients treated with IDH inhibitors have clinically SD. Some of these SD patients achieve transfusion independence and improvements in absolute neutrophil counts despite persistence of bone marrow blast disease. Survival benefit derived from attaining SD with IDH inhibitors needs further substantiation and is currently an area of active investigation. IDH inhibitors are differentiation inducers and do not typically eliminate the malignant clone, as is the case with AG-221. In contrast, most recently updated data suggests that AG-120 is able to effect IDH1 mutational clearance. Nevertheless, clonal persistence and its therapeutic and prognostic implications on the need for continued IDH inhibitor therapy, and survival and future therapy responses beyond IDH-inhibitor failure remain unknown. Differentiation induction with IDH inhibitors may manifest clinically as leukocytosis and other nonspecific symptoms such as fever, edema, hypotension, malaise, and effusions. These reactions should be recognized promptly and may require instituting hydroxyurea, steroids, and potentially treatment discontinuation. Clinical trial experience with AG-221 has shown responses to be consistent across AML subsets, including in R/R AML patients who have failed multiple prior therapies. Also, IDH inhibitors are relatively noncytotoxic and can thus provide a valuable alternative salvage options in IDH-mutated R/R AML. Clinical trials with AG-221 and AG-120 as frontline treatment in AML are in progress and will hopefully shed light on their clinical effectiveness as monotherapy and in combinatorial regimens in the treatment-naïve setting. The widespread derangements in the epigenetic, apoptotic, and metabolic machineries, induced by high 2-HG levels, have provided opportunities for targeting with agents like ABT-199, CB839, ATRA. While these agents are yet to find their niche in IDH-AML management, they may offer viable alternative salvage options after mIDH inhibitor failure. With mounting evidence documenting promising clinical activity, the next step forward is the introduction of these therapies into existing treatment algorithms. Phase Ib combination studies evaluating AG120 and AG221 in combination with standard AML induction and consolidation therapies are underway (NCT02632708).

Declaration of Interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Additional information

Funding

This manuscript was supported in part by the MD Anderson Cancer Centre Leukemia Support Grant (CCSG) CA016672 and generous philanthropic contributions to the MD Anderson Moon Shots Program. And also supported by National Institute for Health Research.