Dysregulated cellular energetics as a signature of cancer cells were noted by Nobel laureate Otto Heinrich Warburg, MD, PhD, in 1924.1 It has taken nearly a century, as well as advances in modern genomics, to fully recognize the importance of cancer metabolism as a biological concept, with the potential to bring clinical benefits through novel, targeted therapeutics. Indeed, dysregulated metabolism is now established as one of the key hallmarks of cancer, initiating and driving tumor growth.2 Recent promising clinical data around inhibitors of mutant isocitrate dehydrogenase (IDH) have generated a great deal of excitement and optimism for cancer metabolism research. Although this has increased awareness of the cancer metabolism field, researchers have known for quite some time that mutations in proteins that are key regulators of cellular energetics are often associated with cancer. The earliest examples of bona fide metabolic tumor suppressors were the mitochondrial enzymes fumarate hydratase and succinate dehydrogenase, where loss-of-function mutations result in hereditary leiomyomatosis, renal cell carcinoma, and paragangliomas.3 When missense mutations in IDH were first described structurally as a codon-substituted residue within the active site, they were originally hypothesized to be loss-of-function mutations.4 But, as is often the case in science, the underlying cancer biology is much more complex, and we have since learned that mutant IDH is a potent oncogene.
Mutant IDH as an Oncogene
The isocitrate dehydrogenase family consists of three critical metabolic enzymes: IDH1, IDH2, and IDH3. All three wild-type enzymes catalyze the oxidative decarboxylation of isocitrate to produce carbon dioxide and alpha-ketoglutarate (aKG), an intermediate in the tricarboxylic acid (TCA) cycle (Fig. 1). The nicotinamide adenine dinucleotide phosphate (NADP+)–dependent family members, IDH1 and IDH2, are homodimeric enzymes that share significant structural similarity. However, one fundamental difference is their subcellular localization, with IDH1 localized to both peroxisomes and cytosol, and IDH2 to mitochondria.5,6 Unlike IDH1 and IDH2, the IDH3 holoenzyme is heterotetrameric and is the canonical mitochondrial member of the TCA cycle, producing aKG and nicotinamide adenine dinucleotide.
The seeds were sown for the therapeutic targeting of IDH1/2 mutations when landmark genomic studies revealed recurrent hot-spot mutations in IDH1 and IDH2 in glioma,7,8 followed by a report of recurrent IDH1 mutations in normal-karyotype acute myeloid leukemia (AML).9 Subsequently, other genomic studies identified IDH mutations in several other cancer types,10,11 as well as premalignant proliferative diseases,12 strengthening the hypothesis that mutant IDH1/2 plays a key role in oncogenesis. But the breakthrough discovery distinguishing mutant IDH as a potential therapeutic target was reported by scientists at Agios in 2009. Using a combination of cellular metabolic profiling and structural biochemistry, the team uncovered the gain-of-function, neomorphic activity of mutant IDH: production of the oncometabolite (R)-2-hydroxyglutarate (2-HG) (Fig. 1).13 Studies show that when IDH1 and IDH2 enzymes are mutated at one of the highly conserved arginine residues in the substrate-binding site, as observed in various cancers, the normal function of catalytic conversion of isocitrate to aKG is lost and they instead produce 2-HG.14 Interestingly, the production of 2-HG was demonstrated for both IDH1 and IDH2, irrespective of codon substitution at this conserved site. To date, IDH3 has not been implicated in cancer biology.
|IDH mutations were originally hypothesized to be loss-of-function mutations, but it is now understood that mutant IDH is a potent oncogene.|
|Somatic IDH mutations occur in a subset of patients across a broad range of solid and hematologic tumors, including glioma, chondrosarcoma, AML, IHCC, angioimmunoblastic T-cell lymphoma, and giant-cell tumor of the bone, melanomas, and prostate, colon, and lung cancers.|
|Multiple preclinical studies have provided proof-of-concept that mutant IDH is a valid target for cancer drug development.|
|Early clinical data from the mutant IDH inhibitor phase I trials indicate that this new drug class has activity in patients with advanced hematologic malignancies, including relapsed/refractory AML and myelodysplastic syndromes.|
2-HG Is a Driver of Cancer
In wild-type cells, 2-HG is produced at only very low levels as a result of metabolic errors by hydroxyacid–oxoacid transhydrogenase or malate dehydrogenase. It is a terminal metabolite with no known physiological role, and it is cleared by 2-HG dehydrogenase. Surprisingly, 2-HG was demonstrated to be present in IDH-mutated cells at millimolar concentrations far exceeding those found in normal cells (up to 100-fold higher).15 This makes 2-HG among the most abundant measureable cellular metabolites in these cells. In cells with IDH1/2 mutations, it appears that the rate of 2-HG production by the mutant enzymes far exceeds the rate of homeostatic clearance, leading to pathological accumulation of 2-HG.
The identification of 2-HG was important, as it provided a mechanism for how IDH mutations could initiate and drive multiple cancer types. There are a number of hypotheses tying IDH mutations and the production of 2-HG to tumorigenesis. The leading one implicates 2-HG as a competitive inhibitor of a class of aKG-dependent DNA and histone-modifying enzymes because of its structural similarity to aKG, including JmjC domain-containing histone demethylases and ten-eleven translocation (TET) family 5-methylcytosine hydroxylases.16,17 These enzymes play a key role in regulating the epigenetic state of cells, and their inhibition by 2-HG leads to epigenetic alterations, and thus, a block in cellular differentiation and unchecked proliferation of undifferentiated cells.18 Other hypotheses involve the effects of 2-HG on EglN prolyl-4-hydroxylases, hypoxia-inducible factor 1 signaling, and extracellular matrix homeostasis.19
In vitro studies have confirmed that overexpression of mutant IDH1 is associated with a high level of 2-HG accumulation, leading to differentiation block and cellular transformation.20 As knowledge of the mechanisms and prevalence of mutant IDH-driven carcinogenesis increases, work is underway to translate these findings into clinical benefit.
IDH Mutations Occur in Multiple Tumor Types
Genotyping efforts have identified somatic IDH mutations in a subset of patients across a broad range of solid and hematologic tumors,21,22 including glioma, chondrosarcoma, AML, intrahepatic cholangiocarcinoma (IHCC), angioimmunoblastic T-cell lymphoma, and giant-cell tumor of the bone, as well as a small percentage of melanomas and prostate, colon, and lung cancers. Somatic IDH mutations have also been identified in premalignant conditions, such as Ollier disease/Maffucci syndrome, in which the associated benign cartilaginous neoplasms (enchondroma) may transform into malignant chondrosarcomas,23 and myelodysplastic syndromes (MDS)/myeloproliferative neoplasms (MPN), which can progress to AML.12
Heterozygous germline IDH2 mutations give rise to the rare neurometabolic disorder D-2-hydroxyglutaric aciduria, in which 2-HG accumulates in the central nervous system, serum, and urine, and patients suffer diverse symptoms, including neurological deficits and reduced lifespan.24
In patients with AML, IDH1 mutations are found in 6%-10% of samples, and IDH2 mutations are found in 9%-13% of samples and are generally associated with normal cytogenetics. The prognostic effect of IDH mutations has been studied, with the majority of reports concluding either a negative or no prognostic effect on overall survival (OS) for either IDH mutation.
IDH1/2 mutations have been reported in 3%-6% of MDS/MPN cases. Two series have indicated that IDH1 mutations are associated with an inferior OS and a higher rate of transformation to AML.25,26
Central nervous system tumors
IDH1 mutations have been identified in approximately 80% of cases of diffuse glioma (predominantly grades II-III) and secondary glioblastoma multiforme (GBM), with IDH1 mutations at the R132 codon occurring most frequently. The rarer IDH2 mutations (approximately 5% gliomas) are associated with the R172 codon. Several series have reported that the presence of an IDH mutation in diffuse glioma is associated with younger age and improved survival independent of age, grade, and MGMT methylation status.21 Although gliomas harboring IDH mutations are histologically similar to gliomas with wild-type IDH (IDHwt), several studies suggest that gliomas harboring IDH mutations represent a distinct disease entity that arises from a different cell type. IDH1-mutated GBM tends to be restricted to frontal locations and displays less contrast enhancement and necrosis than IDHwt GBM. Indeed, clonal studies suggest that gliomagenesis occurs following acquisition of an IDH1 mutation in a common tumor progenitor cell in the presence of other molecular lesions, such as TP53 mutation or 1p/19q co-deletion.27 As gliomas progress, additional genetic events (such as citrate carrier mutation) arise, which cooperate to transform a previously low-grade glioma into a high-grade glioma, with IDH1 mutation persisting throughout progression. This supports the central role of mutant IDH1 in initiating and driving progression of cancer, and the potential therapeutic importance of mutant IDH inhibition.28
Biliary tract carcinomas include epithelial malignancies of the gall bladder and bile ducts (cholangiocarcinoma). IHCCs have been demonstrated to harbor both IDH1 (11%-24%) and IDH2 mutations (2%-6%). Recent work has shown that IDH mutation contributes to IHCC tumorigenesis through 2-HG production and suppression of the hepatocyte nuclear factor 4a regulator, inhibiting liver progenitor cell differentiation.29
Cartilaginous and periosteal malignancies have been found to harbor IDH mutations with a frequency of 40%-52% and 6%-11% for IDH1 and IDH2, respectively. In one study, IDH mutations were discovered in enchondromas and central and dedifferentiated chondrosarcomas but not in peripheral chondrosarcomas or osteochondromas, and they were thought to occur early in tumorigenesis.11
Development of Mutant IDH Inhibitors
The observations that IDH mutations are initiators and drivers of tumorigenesis in multiple tumor types prompted efforts to discover inhibitors of the mutated IDH protein for therapeutic purposes. Targeted inhibition of mutant IDH enzymes and the subsequent reduction of 2-HG has the potential to provide clinical benefit in patients with these mutations. Many of the tumor types discussed above currently have poor outcomes and limited effective treatment options.
Multiple preclinical studies have provided proof-of-concept that mutant IDH is a valid target for cancer drug development. In vitro and in vivo studies have shown that expression of mutant IDH1 or IDH2, 2-HG accumulation, and cellular differentiation block could be reversed with specific mutant IDH1 or IDH2 inhibitors.20,30-33
Mutant IDH inhibitors have entered clinical trials for patients with IDH mutations (NCT02074839, NCT01915498, and NCT02381886). AG-120 and AG-221 are first-in-class, oral, potent, reversible, selective inhibitors of the IDH1 and IDH2 mutant enzymes, respectively. These compounds represent a novel drug class for the targeted treatment of multiple cancers that harbor IDH mutations. Preclinical studies have shown that they inhibit mutant IDH activity and 2-HG accumulation.34-37 Importantly, treatment with AG-221 significantly improved survival in an IDH2-mutant AML primary xenograft mouse model.34
Separate first-in-human, phase I, dose-escalation studies of AG-120 and AG-221 are underway in patients with IDH-mutated hematologic malignancies (NCT02074839 and NCT01915498). Similar phase I dose escalation studies of AG-120 in patients with IDH-mutated gliomas and other solid tumors (NCT02073994), and of AG-221 in patients with IDH-mutated gliomas, other solid tumors, and angioimmunoblastic T-cell lymphoma (NCT02273739), are now open. The primary objective of these studies is to establish the safety and tolerability profile of AG-120 and AG-221, while secondary objectives are to characterize the pharmacokinetics, pharmacodynamics, and clinical efficacy.
Preliminary unpublished clinical data from the ongoing phase I trials for AG-120 and AG-221 have demonstrated clinical proof-of-concept and an acceptable safety profile in patients with relapsed or refractory AML. The maximum tolerated dose has not yet been reached in either trial at doses of up to 300 mg per day for AG-221 and 800 mg once daily for AG-120. Preliminary results from both trials have demonstrated antitumor activity, with durable complete and partial remissions observed with a daily dosing schedule. Consistent with preclinical studies, differentiation of myeloblasts into mature myeloid cells has been observed. Both compounds demonstrate high, dose-dependent exposure levels after multiple oral doses and a long plasma half-life. Additionally, the significantly elevated 2-HG levels seen at study entry were largely reduced after 15 days of dosing to levels seen in healthy volunteers. These trials are ongoing, and dose escalation continues. The phase I trial for AG-221 in hematologic malignancies has expanded enrolment into additional relapsed/refractory AML cohorts defined by age, prior stem cell transplantation, and fitness for chemotherapy.
Mutant IDH Inhibitors: Implications for Clinical Practice
The promising preliminary data from the mutant IDH inhibitor trials indicate that this new drug class has the potential to significantly affect treatment paradigms across multiple tumor types. The novel mechanism of action demonstrated preclinically, whereby the inhibition of 2-HG leads to normal maturation of malignant cells, has also been detected in patients with AML. The detection of IDH mutations allows selection of a genetically-defined patient population for whom mutant IDH inhibitor treatment is appropriate, regardless of tumor pathology, and thus is a predictive molecular marker for response to these agents. There is also scope for combining mutant IDH-targeting agents with existing standard of care, and integration into sequential regimens.
IDH mutations can be detected using standard sequencing and genotyping methods if tumor DNA can be extracted and mutant proteins can be detected in biopsy samples using immunohistochemistry. Assessment of 2-HG levels using medical imaging or mass spectrometry of tumor, blood, or body fluid samples may also be used to determine presence of IDH mutations, depending on the tumor type. Further work is needed to determine the most appropriate IDH mutation detection techniques. The incorporation of IDH mutation detection into standard work-up for tumors that may harbor these mutations should be considered to facilitate early identification of patients who may benefit from inhibitors. Furthermore, a number of studies investigating the use of 2-HG as a quantitative biomarker of disease burden in IDH-mutated tumors indicate that 2-HG can be useful in monitoring treatment response.38-40 Levels of 2-HG circulating in the serum of patients with AML and IHCC can be measured and correlate with disease status,38-40 and advances in imaging allow 2-HG to be quantified noninvasively in patients with glioma using magnetic resonance spectroscopy.41
Advances in cellular metabolism and cancer research led to the discovery that IDH mutations result in a gain of function, namely production of the oncometabolite, 2-HG. A new class of drugs targeting mutant IDH has been shown to induce differentiation in preclinical models, and clinical proof of concept has been achieved in early phase I trials in adults with relapsed or refractory AML. These promising early results are driving expansion of trials in AML and other IDH mutation–positive advanced hematologic malignancies, with data from solid tumors anticipated in the near future. These agents have the potential to transform clinical management across all tumors harboring IDH mutations, regardless of pathology, and the development of combination strategies using standard-of-care and scientifically-driven approaches are on the horizon.
Agios Pharmaceuticals would like to acknowledge all of the patients, their caregivers, nurses, and physicians participating in the clinical trials mentioned.
Medical writing assistance was provided by Christine Tomlins, PhD, Excel Scientific Solutions, Horsham, U.K., and supported by Agios.
About the Author: Dr. Schenkein is chief executive officer of Agios Pharmaceuticals. He has been an ASCO member for 21 years.