IDH enzymes are crucial in several cellular metabolic processes, including the Krebs cycle, glutamine metabolism, lipogenesis, and regulation of cellular redox status (8, 9). Three IDH isoforms, IDH1, IDH2 and IDH3, encoded by different genes with distinct cellular localization, contribute to the regulation of central metabolic circuits (10–12). IDH1 is located in the cytoplasm and peroxisomes, whereas IDH2 and IDH3 are localized to the mitochondria (10, 22).

IDH enzymes are fundamental to the tricarboxylic acid cycle (TCA, Krebs cycle) (9): a series of metabolic oxidative reactions necessary for the mitochondrial electron transport chain to produce ATP ( Figure 1 ) (6). Nicotinamide adenine dinucleotide phosphate (NADP)-dependent IDH1 and IDH2 catalyze the oxidative decarboxylation of isocitrate into α-ketoglutarate (α-KG), NAPDH, and carbon dioxide (CO2) (3, 9). α-Ketoglutarate is a major modulator of electron transport chain activity and TCA flux (27) and it is a co-factor of numerous important cellular reactions, including fatty acid metabolism (3, 10).

IDH is critical to maintaining a sufficient pool of reduced glutathione (GSH) and supporting the peroxiredoxin system by generating NADPH (12, 28). The NADPH produced from the reactions catalyzed by IDH1 and IDH2 is critical to maintain redox balance and protect cells from reactive oxygen species (ROS) that cause DNA damage (3, 10, 12, 29, 30). IDH is particularly important in the brain, producing 65% of the brain’s NADPH, which is essential for lipid metabolism (31). IDH1 and IDH2 are also involved in glutamine metabolism (32), which is important in tumorigenesis as glutamine deprivation suppresses cancer growth (28).

IDH3 is a holoenzyme located in the mitochondria and it catalyzes the irreversible nicotinamide adenine dinucleotide (NAD+)-dependent α-KG (6, 10, 11).

IDH mutations contribute to tumorigenesis in multiple cancer types by interfering with normal cellular metabolism and through the production of the oncometabolite D2-hydroxyglutarate (D2-HG) ( Figure 2 ). Somatic mutations in IDH1 and IDH2 are heterozygous, and primarily consist of missense variants that result in single amino acid substitutions at key arginine residues within the enzyme’s core active sites (10, 11). IDH1 mutations typically occur at Arginine 132, the R132H and R132C variants are the most prevalent (11). IDH1 G97 is mutated in some colon cancers and pediatric astrocytomas (33). IDH2 mutations occur at Arginine 172 and Arginine 140 (34, 35). There are no reports of tumor-associated mutations in the IDH3 gene (10, 36).

Uncontrolled cell proliferation associated with cancer often leads to metabolic alterations that support rapid cell growth (37). Mutant IDH enzyme activity interferes with normal cellular metabolism by depleting α-ketoglutarate (α-KG) from the Krebs cycle (2). Mutant IDH enzymes also consume NADPH, reducing its availability for maintaining redox balance and de novo lipogenesis (38). Accumulating oxidative damage is a hallmark of cancer biology for IDH-mutated malignancies, as a direct consequence of disruption to redox balance (39). Under hypoxic conditions, IDH1-mutant cells exhibit increased oxidative TCA metabolism and decreased reductive glutamine metabolism (32, 40). Metabolically reduced glutamine serves as the major carbon source for fatty acid synthesis during hypoxia and impaired cellular respiration, a switch that is crucial for sustaining rapid cell proliferation (37).

In addition to losing its normal catalytic activity, the mutant IDH enzyme catalyzes the reduction of α-KG to its (R)-enantiomer, the oncometabolite D2-HG (9–13, 41–43). Accumulation of D2-HG causes profound metabolic dysregulation, including inhibition of normal cellular differentiation, epigenetic modulation, DNA repair, redox balance as well as alterations to the tumor immune microenvironment (11–14).

D2-HG is a homolog of α-KG, and functions as a competitive inhibitor of α-KG-dependent dioxygenases, vital for DNA and histone demethylation (3, 42, 44). Accumulation of D2-HG has been shown to inhibit several key histone demethylases, including KDM7A (demethylates H3K9me2 and H3K27me2), KDM4A/B (demethylates H3K9 and H3K36), and Jumonji C domain-containing (JmjC) histone demethylases (2, 11, 41, 44). D2-HG also inhibits the DNA modifying enzymes in the ten-eleven translocation (TET) family of 5-methlycytosine (5mC) hydroxylases (44). Interference with the normal activity of dioxygenases disrupts histone and DNA methylation patterns, leading to the signature global DNA hypermethylation phenotype seen in IDH-mutant gliomas, known as the Glioma CpG Island Methylator Phenotype (G-CIMP) (2). This epigenetic reprograming results in a block of cellular differentiation, which in turn causes inappropriate activation of growth-promoting signaling (2).

A high concentration of D2-HG also promotes angiogenesis via VEGFR2 signaling and increased matrix metalloproteinase (MMP2) activity (45). Additionally, D2-HG contributes to gliomagenesis by directly stimulating prolyl 4-hydroxylase activity, thereby decreasing hypoxia-inducible factor- α (HIF1α) activity, which enhances the proliferation of human astrocytes in vitro (2, 46). Furthermore, tumor cell-derived R-2-HG is absorbed by T cells, leading to a disruption of nuclear factor of activated T cells (NFAT) transcriptional activity and polyamine biosynthesis, resulting in the suppression of T cell activity (47). D2-HG drives an immunosuppressive tumor microenvironment, acting directly on CD8+ cells by altering their metabolic and cytotoxic signatures (26, 47, 48).

D2-HG has been shown to accumulate in high levels in glioma cells but is absent in normal brain cells (44). D2-HG also structurally and functionally mimics glutamate, thereby contributing to the genesis of seizures in patients with gliomas (36, 49).

Patients with Ollier disease and Maffucci syndrome, non-hereditary skeletal disorders characterized by multiple enchondromas and spindle cell hemangiomas, are associated with somatic mosaic IDH1 and IDH2 mutations (50). These patients are at increased risk of developing IDH-mutant gliomas (51) and require thorough clinical examination along with a full body MRI every second year, from age 25 years (52).

IDH mutations have been implicated in other cancer types, including AML, myelodysplastic syndrome (15–18), cholangiocarcinoma of intrahepatic origin (53), central and periosteal cartilaginous chondrosarcomas (54), and melanoma (55, 56).

IDH1 and IDH2 mutations are important, class-defining mutations in gliomas that carry significant therapeutic and prognostic implications (3, 63). IDH mutations are very early events in gliomagenesis, affecting a common glial precursor cell population in most cases except in patients with replication repair deficiency (RRD) (10, 64, 65), and show almost ubiquitous expression throughout the life-cycle of the disease. Among the three isoforms of IDH, mutations in IDH1 are the most common (66).

The latest WHO Classification of CNS Tumors divides mIDH adult-type diffuse gliomas into astrocytoma grades 2-4 and oligodendroglioma 1p/19q-codeleted. Mutations in IDH1 or IDH2 define WHO grade 2 and 3 diffuse gliomas in adults (5, 34, 66, 67).

Patients with IDH-mutant gliomas tend to be younger than their wild-type counterparts with a median age of 37 years (68). IDH1 gliomas arise from a neural precursor population that is spatially and temporally restricted in the brain (69). Therefore, mIDH tumors tend to be in the frontal lobes compared with other lobes and compared with IDH wild-type tumors (70).

The presence of co-existing mutations also helps subdivide mIDH gliomas and provides prognostic information. Astrocytomas typically harbor ATRX and TP53 mutations (3, 24), whilst oligodendrogliomas are defined by co-deletions in 1p and 19q and may carry CIC and TERT mutations (3). Overall survival (OS) is significantly shorter for astrocytomas harboring CDKN2A/B deletions in both pediatric and adult patients (21, 71), and therefore the CNS5 classification now upgrades astrocytomas with CDKN2A/B homozygous deletion to WHO grade 4 irrespective of high grade histology features with microvascular proliferation or necrosis (58). The role of CDKN2A deletion in patients with oligodendroglioma remains under investigation. The presence of mutations in the PIK3CA, PIK3R1 and amplification of PDGFRA and MYCN also confer a trend toward poorer survival (72). Genetic alterations involving members of the RB1 pathway, including CDK4 amplification and RB1 mutation or homozygous deletion, have also been reported (71).

IDH mutation status is an independent predictor of favorable outcomes among adults with glioma (34, 73). In part, this may be due to increased sensitivity to chemotherapy and radiotherapy, as IDH mutation does not have the same impact on survival in untreated tumors (74).

Given the rarity of IDH mutations in pediatric and AYA tumors, there is less detail on their prevalence and clinical impact (75), with most information found in small cohort studies. The prevalence of IDH mutations in the pediatric and AYA cohort ranges from 1% to over 50%, with IDH1 mutations being more common than IDH2 mutations ( Table 1 ). In a cohort of 43 pediatric patients with newly diagnosed WHO grade 3 or 4 gliomas treated on the Children’s Oncology Group ACNS0423 study, IDH1 mutations were detected in 7 of 43 (16.3%) of children with primary malignant gliomas, and no IDH2 mutations were identified. Older children and AYA are more likely to have mIDH gliomas. In a large AYA case series including patients aged 15-39.9 by Bennet et al. IDH mutations were noted in 57% patients (7) whereas Pollack et al. report IDH mutations in 7 of 20 gliomas (35%) from children ≥14 years and 0 of 23 (0%) in younger children (19). The PRecISion Medicine for Children with Cancer (PRISM), trial has reported two patients with mIDH astrocytoma of the 146 patients enrolled with high-risk pediatric brain tumors who had at least 18 months of follow-up (77).

Recently, a distinct cohort of patients with IDH-mutant astrocytomas has been recognized with hereditary mismatch repair deficiency (MMR) (78, 79). These patients are unique to those with acquired MMR due to alkylating agents. Instead, primary mismatch repair-deficient IDH-mutant astrocytomas (PMMRDIA) are histologically high grade, often displaying a hypermutant genotype and microsatellite instability (78, 80). These tumors primarily present in younger patients and have worse clinical outcomes compared to other IDH-mutant gliomas (78).

Dodgshun et al. first described six patients with RRD HGG with secondary IDH1 mutations that clustered with other IDH1 mutant gliomas on methylation profiling (79). However, while other mIDH gliomas demonstrated a CpG Island Methylator Phenotype (CIMP), these six tumors with secondary IDH1 mutations displayed the inverse pattern of methylation disturbance; a CpG Island Demethylator Phenotype (CIDP) in keeping with other RRD HGG. These findings suggest that the CIDP phenotype generated by RRD glioma is unable to be overcome by secondary IDH mutations.

In a study by Suwala et al., samples from 32 patients with IDH-mutant gliomas and proven or suspected primary MMR deficiency were sequenced. Patients were clinically diagnosed with Lynch syndrome or Constitutional Mismatch Repair Deficiency Syndrome and/or had germline mutations in DNA mismatch repair genes (MLH1, MSH6, MSH2) in all but one case. Results demonstrated a distinct DNA methylation profile which clustered separately from other IDH-mutant glioma subtypes including those with acquired MMR deficiency on t-distributed stochastic neighbor embedded (t-SNE) plots. Tumors displayed a higher proportion (60%) of unmethylated MGMT promoter compared to other IDH-mutant gliomas. Frequent inactivation of TP53, RB1, ATRX and activation of the RTK/PI3K/AKT pathway was also present. The OS of this cohort of patients was significantly worse, with a mean OS of only 15 months despite treatment with surgery, radiation and chemotherapy. Pediatric and AYA patients diagnosed with a mIDH glioma warrant screening for MMRD.

The benefits of safe maximal resection in adults with glioma were well-established prior to the era of molecular characterization. These results have been recapitulated in the modern diagnostics era by Wijenga et al. (81) with a retrospective re-analysis of adult tumor subtypes applying new molecular classification. This study demonstrated that for mIDH astrocytomas, each 1cm³ increase in postoperative volume of disease results in poorer OS, hazard ratio (HR) of 1.01 (p<0.0001). However, the impact of extent of upfront resection is less clear in pediatric patients. In a multi-institutional retrospective analysis of 78 pediatric patients with IDH1/2 mutated gliomas including 45 patients with low grade astrocytoma, the 5Y PFS of patients with gross total resection (GTR) was 45.7% with median PFS of 4.4 years compared to 5Y PFS of patients with subtotal resection (STR) of biopsy of 41% with median PFS of 4.7 years (21). There was no statistically significant difference in OS in this cohort. With only 13 patients in the cohort of low-grade oligodendrogliomas in this study, there was no statistically significant differences in PFS or OS for upfront GTR versus STR or biopsy. The discrepancies in outcomes compared to adults may be reflective of small sample size.

In adults, repeat craniotomy and safe resection is feasible for patients with recurrent LGG (82). Re-operation in eloquent or near eloquent brain areas are not associated with higher risk of neurological sequelae compared to initial surgery (83). Up to 50% of patients can achieve a gross total resection at time of recurrence (84). Mounting evidence demonstrates that there is survival benefit to near total or gross resection at time of recurrence (83–85). Thus, standard of care is moving towards aggressive safe early resection at recurrence.

In adult IDH-mutant grade 2 gliomas a “watch and wait” approach may be taken for “low-risk” patients, historically defined as younger patients under 40 years with GTR (2, 86). For “high-risk” patients, postoperative combination radiotherapy and sequential chemotherapy is standard of care (2, 24, 25). The most common chemotherapy regimens are procarbazine, lomustine and vincristine (PCV) or temozolomide (TMZ). All cooperative groups recommend radiotherapy and sequential chemotherapy for patients with Grade 2 IDH-mutant gliomas aged over 40 years and with STR (2, 24, 25). Patients with Grade 3 oligodendrogliomas with 1p/19q co-deletion should receive radiotherapy with PCV (2, 24, 25). Those with Grade 3 astrocytomas and no 1p/19q co-deletion receive radiotherapy with adjuvant TMZ (2, 24, 25).

TMZ is an oral alkylating agent that results in the methylation of DNA at the O⁶-guanine, as well as N⁷-guanine and N³-adenine residues (87). Methylation of O⁶-guanine results in the O⁶-methylgaunine (O⁶-meg) lesion. O⁶-meg lesions are repaired by O⁶-methylguanine DNA methyltransferase (MGMT), a process that is dependent on the number of MGMT molecules per cell and rate of MGMT regeneration (88). Unrepaired O⁶-meg lesions incorrectly pairs with thymine during DNA replication instead of cytosine, triggering the DNA mismatch repair (MMR) system and repeated cycles of double-stranded breakages in a process called futile cycling (87). The accumulation of these DNA strand breaks eventually results in cell cycle arrest and activation of apoptosis. Thus, cytotoxicity of TMZ is dependent on an intact MMR pathway and low levels of MGMT.

The CATNON trial (89), a phase 3 randomized controlled trial, established the use of adjuvant rather than concurrent TMZ with radiotherapy on 1p/19q non-co-deleted anaplastic gliomas, particularly those with IDH mutations. The CATNON trial used a 2 x 2 factorial design, enrolling 751 patients to receive either radiotherapy alone, or radiotherapy with concurrent and/or adjuvant TMZ. Adjuvant TMZ showed significant improvement in OS (median OS 82.3 months, 5Y OS 56% vs median OS 46.9 months, 5Y OS 44% in control arm). The second interim analysis did not demonstrate a statistically significant benefit of concurrent TMZ compared to radiotherapy alone. TMZ was generally well tolerated with the most common Grade 3 and 4 TRAE being related to hematological side effects, particularly thrombocytopenia and neutropenia.

Importantly, this study holds significant relevance for IDH-mutant gliomas. Mutation status was analyzed in 671 of the 751 patients, with 436 harboring IDH1 or IDH2 mutations, evenly distributed across the four arms of the study. Overall, IDH1/IDH2 mutation status had a significant impact on survival with median OS 19.9 months in the wild-type group vs 98.4 months in the IDH-mutant group (HR 0.14, p<0.0001). Significantly, in patients with IDH1 and IDH2 wild-type tumors, neither concurrent nor adjuvant TMZ improved OS compared with radiotherapy alone, whereas IDH-mutated tumors had significantly improved OS with the addition of adjuvant TMZ. Median PFS in IDH-mutant patients treated with any TMZ was 77.0 months compared to 34.2 months in patients treated with radiotherapy alone (p<0.0001).

The results of the CATNON trial suggest that there is a crucial role for adjuvant TMZ in adult IDH-mutant LGG and this is currently recommended as standard of care in the settings outlined above. However, in the pediatric and AYA population TMZ-induced hypermutation and acquired resistance need to be considered given the potential for prolonged survival and ongoing exposure to alkylating agents. TMZ-resistance may occur through the mutagenic action of TMZ on DNA repair genes resulting in acquired deficiencies in MMR (90). Loss of MMR function leads to ongoing mispairing of guanine with thymine; however, unrepaired DNA damage is no longer detected and cells are unable to activate apoptosis (91). In the absence of MGMT-mediated repair and intact MMR, cells incur a large number of G:C>A:T transitions throughout the genome upon DNA replication, including thousands in coding regions (87). This mutational signature is known as single base substitution (SBS) signature 11 and is commonly seen in hypermutated gliomas after exposure to alkylating agents (92). The acquisition of mutations that inactivate the MMR pathway and continued TMZ exposure results in hypermutation. Despite these considerations, TMZ remains widely used in pediatric HGG.

Another recent trial, the CODEL trial, was initially designed to compare the efficacy of radiotherapy alone, radiotherapy + TMZ, or TMZ monotherapy in patients with 1p/19q codeleted grade 3 oligodendroglioma (93). Initial analysis demonstrated a significantly shorter PFS in patients receiving TMZ monotherapy compared to RT (HR 3.12, p=0.014). Of the patients included that were evaluable for IDH mutation status, 30 (86%) were IDH mutated. Given the results, the trial has been redesigned to compare adjuvant radiotherapy + TMZ and radiotherapy + PCV among patients with grade 2 and 3 oligodendroglioma. The trial is ongoing with results awaited. However, recent results from the French POLA network demonstrate that in a cohort of 306 patients with grade 3 oligodendroglioma, radiotherapy + PCV was associated with significantly improved 5-year and 10-year OS compared to radiotherapy + TMZ (5Y OS 89% PCV vs 75% TMZ, p=0.0014; 10Y OS 73% PCV vs 60% TMZ, p=0.0003) (94).

Understanding of the biological impact of IDH inhibitors in glioma is evolving. Preclinical modelling in mIDH glioma is complicated by the indolent nature of the tumors which makes generation of patient-derived cell lines and xenograft models difficult (49, 133). However, using a combination of RNA single-cell analysis, bulk RNA-sequencing analysis and in vitro modelling, Spitzer et al. were able to demonstrate differentiation of oligodendrogliomas to an “astrocytoma-like” phenotype with decreased cell cycle expression, thus resulting in less proliferation, in patients treated with vorasidenib (134).

The specific role of IDH mutations in driving tumor growth and aggressiveness at the time of recurrence is also unclear. Recurrent tumors are more likely to display a hypermutation phenotype and genome-wide loss of DNA methylation (135, 136). Several recent studies also confirm the present of copy number alterations, particularly CDKN2A/B loss and CDK4 amplification, clinically conferring poorer prognosis in IDH-mutant astrocytomas (137–141). However, the effect of CDKN2A/B deletion in oligodendroglioma is less clear (142). These findings may explain the clinical data demonstrating poorer survival outcomes of high-grade mIDH glioma patients treated at recurrence with mIDH inhibitors (143).

The role of IDH inhibitors in standard of care for IDH-mutant gliomas is yet to be determined. Overall, studies suggest superior activity in patients with non-enhancing tumors compared with enhancing tumors (97, 98, 100, 102), with enhancement typically indicating transformation of LGG to higher grades (98). The suggestion based on the above trials would be that these inhibitors may be useful in the early, indolent course of disease (133). However, many questions remain including the efficacy of IDH inhibitors compared to current standard of care radiotherapy and chemotherapy regimens.

There is a lack of current understanding of how prolonged use of IDH inhibitors alters the biology of mIDH gliomas, leading to resistance (133). IDH inhibitor resistance has been reported in adults with AML through trans or cis dimer-interface mutations (144), receptor tyrosine kinase (RTK) pathway mutations (145) or second site mutations (146, 147). In adults with cholangiocarcinoma resistance mechanisms include a second IDH mutation (129, 148). Understanding the mechanism of resistance is important to guide salvage treatment options and to identify other potential targets.

Several reports and a recently published clinical trial by the International Replication Repair Deficiency Consortium (IRRDC) demonstrating durable responses and prolonged survival in patients with glioma and high mutational burden (149, 150). In contrast, in Suwala’s study of patients with PMMRDIA, limited effect was seen in three patients treated with immune checkpoint inhibitors (ICI) (78). A subsequent IRRDC study of mIDH RRD HGG confirmed lower PFS in mIDH RRD HGG compared to sporadic mIDH HGG treated with ICI monotherapy (151).

The lack of efficacy in this cohort is thought to be attributed to the immune suppressive effects of the IDH mutation on the tumor microenvironment (80). IDH mutations result in down-regulation of genes involved in immune activation of cancer cells likely occurring either by a direct metabolic inhibitory effect of D2-HG or through epigenetic reprogramming by hypermethylation of promotors of immune-related genes (47). Hence, D2-HG strongly represses T-cell activity, acting as a protective mechanism for tumor cells by evading the immune system. Furthermore, hypermethylation has been implicated in the downregulation of immune checkpoints such as PD-L1 and CTLA4 (78, 152).

The combination of IDH inhibitors with ICI is therefore promising, specifically for PMMRDIA where novel therapeutic strategies are required. In this combination, IDH mutation driven immunosuppression could be reversed with IDH inhibitor treatment, enhancing immune checkpoint blockade. In the clinic, Das et al. have reported favorable responses for mIDH-RRD-HGG receiving combination ICI and IDH inhibitor therapy with prolonged survival at 12 months compared to those without IDH mutations (151). To test this further, a number of clinical trials are ongoing including a phase 2 trial combining nivolumab with ivosidenib in patients with advanced IDH-mutant solid tumors (NCT04056910) as well as a phase 1 trial of pembrolizumab and vorasidenib in patients with relapsed/refractory Grade 2 and 3 mIDH glioma (NCT0584622).