Although ivosidenib and olutasidenib inhibit the same target, namely the mutant IDH1 protein dimer, they are structurally distinct molecules with different aspects to their respective specificity and binding mechanisms (Fig. 1).

Molecular structure of olutasidenib and ivosidenib. (left) Olutasidenib has a chemical formula of C18H15CIN4O2. The molecular weight of olutasidenib is 355 g/mol. (right) Ivosidenib has a chemical formula of C28H22CIF3N5O3. The molecular weight of ivosidenib is 583 g/mol

In comparing the individual pharmacokinetic/pharmacodynamic profiles of the two molecules, one key difference between olutasidenib and ivosidenib involves their selectivity. While both molecules are effective at inhibiting mutant IDH1, olutasidenib exhibits greater selectivity towards targeting the mutant form without inhibiting wild-type IDH1. In vitro studies showed that olutasidenib inhibited mutated IDH1 R132H, R132L, R132G, and R132C proteins; however, wild-type IDH1 and mutated IDH2 proteins were not inhibited. The inhibition of mutant IDH1 by olutasidenib led to reduced 2-HG levels. In contrast, ivosidenib potently inhibits both mutant and wild-type isoforms of IDH1. The IC50 of olutasidenib for wild-type IDH1 is 22,400 nM in stark contrast to the IC50 of ivosidenib for wild-type IDH1, which is 24-71 nM [15, 16]. Thus, ivosidenib inhibits wild-type IDH1 at an almost 1000-fold lower concentration [15].

Inhibition of wild-type IDH1 has implications for its downstream metabolic functions. The normal function of wild-type IDH1 is illustrated in Fig. 2. Wild-type IDH1 enzyme catalyzes the oxidative decarboxylation of D-isocitrate to α-ketoglutarate (α-KG). Inhibition of wild-type IDH1 in cells can deplete α-KG and by extension NADPH. Since the reductive power of NADPH is protective against reactive oxygen species (ROS), reduced NADPH production leads to increased oxidative stress [17]. It has been shown that IDH1 knockout (KO) mice have increased levels of ROS in the liver compared to wild-type mice, which affected their survival after sublethal lipopolysaccharide injection [18]. Furthermore, changes in NADP + /NADPH ratio and associated ROS have been linked to ion channel dysfunction, which may have implications on QT prolongation. It is interesting to note that QT prolongation was observed in vivo for ivosidenib. The QTc predicted prolongation observed in patients treated with ivosidenib was 16.1 ms [19]. In contrast, the largest mean increase in QTc interval observed for olutasidenib was 6.2 ms [20]. In clinical trials, the incidence of grade 3 or higher prolongation of QT interval as measured by ECG was 7.8% (14/179) in the R/R AML overall population who received ivosidenib, and < 1% (1/153) in the R/R patients who received olutasidenib.

Reactions catalyzed by wild-type and mutant IDH1. Wild-type IDH1 (w.t. IDH1) converts isocitrate to α-ketoglutarate (α-KG) in a reaction that uses NADP + as an electron acceptor and produces NADPH. A point mutation in the IDH1 gene at the Arg132 residue (R132) results in a gain-of-function mutant IDH1 (mut. IDH1) that reduces α-KG to the oncometabolite 2-hydroxyglutarate (2-HG), consuming NADPH in the process. The inhibitors ivosidenib and olutasidenib potently bind mut. IDH1. Ivosidenib also inhibits w.t. IDH1 at an IC50 of 24–71 nM, whereas the IC50 of olutasidenib for w.t. IDH1 is 22,400 nM, indicating no inhibitory activity

Depletion of α-KG also has oncogenic consequences as it is a rate-limiting substrate of 2-oxoglutarate-dependent dioxygenases and is an essential cofactor for histones and DNA demethylases. Additionally, α-KG is required for amino acid driven gluconeogenesis, the loss of which is linked to poor cell viability in low glucose conditions and worsened survival in animals under fasted conditions [21]. Alpha-ketoglutarate also plays a role in erythropoiesis by affecting heme production. It is apparent that wild-type IDH1 maintains several important functions that can affect cell viability and survival under stress conditions. Therefore, treatments with greater selectivity towards the mutant form of IDH1 without inhibiting the normal function of wild-type IDH1 may have clinical benefits without undue effects on hematopoiesis.

As with most mIDH1 inhibitors, olutasidenib and ivosidenib do not bind at the active site of each protein monomer, but instead bind to the interface of an IDH1 dimer. The dimer interface forms an allosteric site demarcated in part by alpha helices (alpha helix 10 from each monomer) which are adjacent to the catalytic site in both mutant and wild-type IDH1 monomers. The normal IDH1 enzymatic activity of converting isocitrate to α-KG with corresponding reduction of NADP + to NADPH requires wild-type IDH1 homodimers [22]. In cell culture and in vitro assays, both mutant IDH1 homodimers and wild-type/R132 IDH1 heterodimers demonstrate the ability to catalyze reduction of α-KG to 2-HG with consumption of NADPH [23, 24]. However, in cancer cells bearing a mutant IDH1 R132 allele, the wild-type/R132 heterodimer configuration is considered the predominant driver of neomorphic activity, possibly due to its more abundant form, availability of α-KG substrate, or in vivo efficiency [25, 26].

Olutasidenib has a lower molecular weight than ivosidenib (FW 355 vs FW 583, respectively) and occupies a smaller space within the allosteric pocket formed by an IDH1 dimer. The smaller size of olutasidenib allows it to bind independently to each monomer unit of IDH1 within the dimer interface, potentially at a 2:1 ratio of olutasidenib to IDH1 dimer. In contrast, due to its larger size, a single ivosidenib molecule straddles the allosteric pocket of the dimer, resulting in a ratio of 1 molecule of ivosidenib per IDH1 dimer, regardless of whether the dimer is composed of wild-type homodimer, wild-type/R132 heterodimer, or R132 IDH1 homodimer [27].

The binding properties of the 2 molecules have implications for how each drug affects interactions and kinetics within the binding pocket of either mutant or wild-type IDH1 dimers. Both molecules form interactions with Asp279, found on the alpha 10 helix that makes up one side of the allosteric binding pocket. This same alpha helix sits in between the allosteric binding site and the catalytic site. Notably, residues from alpha 10 helix normally facing the catalytic site form interactions with a magnesium ion via Asp279, Asp275, and Asp252. The magnesium ion binding plays a key role in enzyme catalytic reaction. Both olutasidenib and ivosidenib interact with Asp279, which pulls away the aspartate residue from interacting with the magnesium ion. The weakened magnesium ion binding is thought to be the mechanism by which both inhibitors target mutant and wild-type IDH1 enzymatic activity.

Reports indicate that ivosidenib binds slowly and tightly to wild-type IDH1. It is possible that ivosidenib stays bound longer and occupies the allosteric pocket in a more constant state, which may result in an increased interaction with wild-type IDH1 and its subsequent inhibition. This is consistent with ivosidenib binding to both wild-type and mutant IDH1 in isothermal calorimetry (ITC) assays [28]. In contrast, the smaller olutasidenib molecule has weaker binding affinity to wild-type IDH1, making it a more selective inhibitor of mutant IDH1.

Choe et al. first reported that the acquisition of a second-site mutation is one potential mechanism of ivosidenib resistance [29]. Reinbold et. al. further screened IDH1 inhibitors, including ivosidenib and olutasidenib, for inhibitory activity against the IDH1 R132 hotspot mutation, with or without a second site mutation in cis [27]. It was reported that although ivosidenib demonstrated potency against IDH1 single-mutation variants R132C and R132H, no inhibition was detected against the R132C/S280F and R132H/S280F double-mutation variants [27]. In contrast, olutasidenib retained binding affinity to both single- and double-mutated IDH1 variants (Table 3) [27]. Furthermore, in a cellular assay, ivosidenib failed to suppress 2-HG production in R132C/S280F or R132H/S280F-positive cells whereas olutasidenib reduced 2-HG levels in the same cell types (Fig. 3; adapted from Reinbold et al.) [27]. Structural analysis indicated that the increased size of the phenylalanine in S280F, adjacent to the key Asp279 residue, blocks binding of ivosidenib via steric hindrance. Interestingly, the crystal structure of the smaller olutasidenib bound to IDH1 suggests that it can accommodate the bulky phenylalanine, consistent with biochemical data (Fig. 4).

Influence of inhibitors on 2-HG levels in cells bearing IDH1 mutations [27]. (Adapted from Reinbold et al. [27].) LN18 cells were treated with the inhibitors ivosidenib, olutasidenib in DMSO (final concentrations: 5 μM). 2-HG levels were determined by anion-exchange chromatography couple to MS (n = 4 independent replicates). Control cells were generated by transduction with lentiviral vectors containing no IDH1. Box-and-whisker plots: The center line is the median and the bounds are 25th and 75th percentile values. The whiskers are the min and max measured 2-HG levels for each experimental group

Binding of olutasidenib in single- or double-mutant IDH1. Left: Crystal structure of the binding pocket of single-mutant IDH1 R132H. Olutasidenib (blue) interacts with Asp279 in the IDH1 dimer interface. There is ample space remaining near the adjacent Ser280 site. Right: Generated image of the binding pocket of double-mutant IDH1 R132H/S280F based on the crystal structure of IDH1 R132H and grafting on the phenylalanine substitution. In the presence of a second site mutation whereby the serine is replaced with phenylalanine (S280F), the bulky benzene ring of phenylalanine increases steric hindrance and prevents larger molecules from interacting with Asp279 in the binding pocket. Due to its small size, olutasidenib is able to bind Asp279 even in the presence of the second site S280F mutation