Small molecules that block a key protein–protein interaction could change the treatment landscape for many patients with acute leukaemias.

By Ken Garber

Nature Reviews Drug Discovery, 2024 Aug;23(8):567-569.  doi: 10.1038/d41573-024-00106-3.

An important new class of drugs for acute myeloid leukaemia (AML) and acute lymphocytic leukaemia (ALL) could soon secure FDA approval. The agency should decide on Syndax’s first-in-class menin inhibitor revumenib by late September. At least seven small-molecule inhibitors of this scaffold protein are in clinical trials (Table 1). These drugs look to be effective in at least two genetic subsets of leukaemia.

“We’re just getting started with menin inhibitors,” says Lee Greenberger, CSO of the Leukemia & Lymphoma Society. The society funded the earliest menin inhibitor work, at the University of Michigan, and has certain rights to Kura Oncology’s menin inhibitor ziftomenib.

In a phase I trial in relapsed or refractory acute leukaemia, revumenib provided a 30% complete response/complete response with partial haematological recovery (CR/CRh) rate. Registrational phase II trial results, reported at the American Society for Hematology (ASH) annual meeting in December, were not quite as good, but the trial still met its primary efficacy endpoint and was stopped early for efficacy.

The phase II data are similar to what secured approval for IDH inhibitors and FLT3 inhibitors in AML, says Greenberger. “It’s in the same ballpark. That’s the good news. The bad news is, don’t expect these drugs to be curative. We’re going to need to [use them] in combination.”

Guggenheim Partners biotech analyst Paul Jeng sees a potential US$3.5 billion market in the USA for menin inhibitors, assuming eventual front-line and maintenance use in leukaemias. These drugs have proved effective in leukaemias with two genetic alterations: fusions involving KMT2A (also known as MLL1) and mutations in NPM1, the most frequent AML mutation in adults. Together these alterations amount to about 40% of AML cases.

Like IDH and FLT3 inhibitors in AML, menin inhibitors reduce tumour burden by removing a block on differentiation, enabling stalled, self-renewing cells to mature and die.

An unlikely target

The origins of the menin inhibitors date back almost two decades, to Stanford cancer biologist Michael Cleary’s 2005 paper showing that genetic blockade of the interaction between KMT2A and menin prevented leukaemia in mice. At the time, most drug developers were not interested in taking on such protein–protein interactions, which involve large surface areas that are hard to block with small molecules. Instead, some targeted another KMT2A complex component, the enzyme DOT1L. But the Cleary paper caught the attention of Jolanta Grembecka and Tomasz Cierpicki, married chemists from Poland who were then postdocs at the University of Virginia.

“We thought this could be a very good project for us,” says Cierpicki, who later moved with Grembecka to the University of Michigan, where they started separate labs. Menin’s role in leukaemia, Cierpicki says, was “an interesting scientific problem.”

The two scientists proceeded systematically. First, they mapped a 43-amino acid region on the amino terminus of KMT2A that binds, via two separate motifs, to a cavity in menin. Such a large surface would typically be hard to block with a small molecule. But Cierpicki’s subsequent high-resolution crystal structure of human menin bound to the KMT2A motifs revealed that hydrophobic contacts at three residues on KMT2A were critical for binding. “These residues are in close proximity,” says Grembecka. “It seemed to be feasible to develop small molecules that can kind of occupy the same position on menin.”

Their first screen and optimization work resulted in compounds with nanomolar potency, they reported in 2012, although these agents were metabolically unstable. Kura Oncology licensed optimized compounds in December 2014. Further tweaking yielded Kura’s lead menin inhibitor, ziftomenib.

Vitae Pharmaceuticals also launched a menin inhibitor discovery programme in 2014. “We had a little trepidation” blocking a protein–protein interaction, recalls Jerry McGeehan, a Vitae chemist at the time and now menin programme team leader at Syndax. But the menin pocket looked tractable. Vitae preserved a key bond between the Michigan compounds and a menin residue, using structure-based design to arrive at a different scaffold. Allergan acquired Vitae in 2016, and subsequently licensed Vitae’s menin inhibitors to Syndax. By then Grembecka and Cierpicki had shown that menin inhibitors work in mouse models of leukaemia, emboldening other companies to jump in.

Menin inhibitors, the field has learned, act via epigenetic mechanisms. Menin is a scaffold protein, and its partner KMT2A is a histone methyltransferase that is critical for normal haematopoiesis. In AML, KMT2A fusion proteins drive expression of genes that lock in cell self-renewal and block differentiation, such that abnormal immature myeloid cells, known as blasts, accumulate. KMT2A fusions need menin for this oncogenic event. Menin inhibitors release the KMT2A fusion complex from chromatin, removing the blockade on differentiation and letting cells proceed to either apoptosis or to a differentiation programme and eventual turnover. Menin inhibitors do not seem to interfere with most wild-type KMT2A function, so they do not cause massive depletion of normal myeloid or lymphoid cells.

Dana-Farber Cancer Institute haematology researcher Scott Armstrong reported in 2016 that menin inhibitors also work in NPM1-mutated leukaemias, opening up a much larger patient population for these drugs. The disease mechanism in these cancers remains to be fully resolved. NPM1 mutations cause NPM1 to build up in the cytoplasm rather than the nucleolus, which makes some NPM1 partner proteins unable to fulfill their regulatory roles. But Armstrong’s group recently showed that mutant NPM1 also binds directly to chromatin targets and, with wild-type KMT2A, stalls the differentiation programme. Menin inhibitors reverse this activity, restoring differentiation.

A two-drug race

Kura’s ziftomenib began human trials in 2019, with Syndax’s revumenib two months behind. Syndax has since pulled well ahead, at least in patients with KMT2A rearrangements.

The FDA “didn’t let us select [patients],” explains Kura CEO Troy Wilson. “So we dosed in all comers.” Kura’s path through Project Optimus, an FDA programme that ensures that cancer drugs are tested at their optimal doses, was also lengthy. “We lost some time there,” says Wilson.

Syndax, on the other hand, was able to select patients with KMT2A rearrangements or NMP1 mutations almost from the outset, and achieved an effective dose quickly under Project Optimus. The company attributes both outcomes to clean preclinical data. “Physicians got behind our programme early,” adds Syndax CEO Michael Metzger, thanks in part to a compassionate use programme that provided the drug to more than 125 patients.

Syndax is seeking approval first in KMT2A-rearranged leukaemias, because it obtained FDA breakthrough therapy designation for this population, which comprised most of its phase I trial. Patients with KMT2A-rearranged leukaemias fare particularly poorly, with roughly a 25% 5-year survival rate, increasing the urgency for new treatments. At the 2023 ASH meeting, an open label, single-arm phase II trial of revumenib in 57 previously treated patients with KMT2A-rearranged leukaemias yielded a CR/CRh rate of 23%. For the 13 patients whose cancer went into remission, the median response duration was 6.4 months. Median overall survival was 8 months at the time of the interim analysis. The study was stopped early for efficacy. About 40% of responders proceeded to stem cell transplant, which can be curative, with many restarting revumenib as maintenance therapy.

By comparison, current salvage therapies after second relapse produce a 5% complete response rate with 2.4 months median overall survival. “They seem to have cleared the necessary regulatory bar here for approval,” says BTIG biotech analyst Justin Zelin.

“We’ll bring this drug to market ahead of all the competition,” Metzger says. “For KMT2A there will be very little competition for several years.”

The bigger market opportunity is NPM1-mutated leukaemias. “All of our data, at least preclinical data, suggest [this cancer is] similarly responsive,” says Armstrong, who has filed a patent on a menin inhibitor. Syndax finished enrolling a phase II trial in this cancer in March, and is eyeing a possible regulatory filing in the first half of 2025.

Kura, meanwhile, expects to complete enrolment in a registrational trial of ziftomenib in NPM1-mutated leukaemias by mid-2024, which might enable a filing early next year. “We’re catching up quickly,” says Wilson. Last year at the European Hematology Association Congress, Kura reported a 35% composite complete response rate for ziftomenib in 20 patients with NPM1-mutated leukaemias, with a median duration of response of 8.2 months.

Class competition

If both revumenib and ziftomenib are approved, physicians will get to choose which to use. As yet, they appear similarly effective, with manageable side effects.

One early concern was that these drugs might interfere with normal haematopoiesis, a process that depends on the activity of menin partner KMT2A. But this does not happen, possibly because KMT2A has many different domains, and some of its interactions with chromatin appear to be independent of its interaction with menin. “I think that menin is just supporting a very specific set of genes in biology and haematopoiesis, and maybe we’re taking advantage of that,” Armstrong says.

Differentiation syndrome, the rapid accumulation of differentiated cells, has instead emerged as an on-target side effect of these drugs. Symptoms — including shortness of breath, low blood pressure and high fever — can be severe, with some patients requiring oxygen, mechanical ventilation and other interventions. “It can be pretty scary,” says Texas Children’s Hospital paediatric haematologist Michele Redell, who treated several patients with revumenib in phase I and under Syndax’s compassionate use programme. Redell is helping coordinate a basket trial that includes ziftomenib.

Kura has seen severe differentiation syndrome, especially with single agent ziftomenib in patients with KMT2A-rearranged leukaemias. The drug “works too well in KMT2A,” says Wilson. “It works so well that it puts patients into differentiation syndrome, which can become life threatening.” But adding chemotherapy to the drug regimen should minimize differentiation syndrome, he adds, by eliminating blasts before they differentiate, and also by reducing high white blood cell counts, which contribute to the syndrome. Kura has not yet reported any cases of differentiation syndrome in these chemo combination trials, but Wilson expects it will appear in mild form.

KMT2A-rearranged leukaemia blasts permeate most tissues, he adds, “a recipe for differentiation syndrome.” The company has so far seen less differentiation syndrome with NPM1-mutated leukaemia, which Wilson describes instead as a disease of bone marrow and blood.

Differentiation syndrome has been less of a problem for Syndax, which Metzger attributes to the shorter half-life of its drug (about 7 hours for revumenib compared with more than 24 hours for ziftomenib).

The drugs have other distinct safety issues that could prove important. Revumenib’s phase I dose-limiting toxicity was QTc prolongation, a cardiovascular side effect that can cause fatal arrhythmias. Syndax saw medically significant QTc prolongation in 14% of patients in its phase II of the drug. All were able to stay on the drug at a lower dose. QTc prolongation “is something that we’re familiar with monitoring, and comfortable with,” says Redell. “It has not yet been considered a dealbreaker.” But other targeted agents also have this problem, adding risk if they are to be combined with revumenib.

There are also metabolism differences. Revumenib is a substrate of P450 CYP3A4, which complicates the use of the drug in AML. Patients with AML are immunocompromised and so often take azole antifungal drugs that are CYP3A4 inhibitors. These drugs can as a result raise revumenib serum levels, although Syndax has worked out how to mitigate this with a lower revumenib dose. “It’s just something we need to know about, and either avoid it or take advantage of it as appropriate,” Redell says.

Ziftomenib may by contrast be a weak CYP3A4 inhibitor, but Wilson says any ziftomenib CYP3A4 inhibition is “not clinically meaningful.”

Other differences include ziftomenib’s more prolonged tissue penetration and residency, which could boost both efficacy and side effects. And in Kura’s hands, ziftomenib not only blocks the menin–KMT2A interaction, but also drives the degradation of menin — a feature that could prove beneficial. Menin-dependent leukaemias may ultimately respond better to degraders than to inhibitors, believes Wilson. “With rapidly cleared, non-degrading drugs, the menin complex can reform as soon as drug washes out of the cell,” he says. Metzger, for his part, says that continuously blocking the KMT2A interaction is what matters most.

These and other factors will play out over time. “The drugs are pretty comparable at this point,” says Jeng. “The question for us is more about clinical execution.”

Other menin inhibitors are further behind, and some may fall out altogether. Johnson & Johnson’s JNJ-75276617 is in phase I. “The J&J data that we’ve seen have been very underwhelming,” says Zelin. In April Daiichi Sankyo discontinued development of its menin inhibitor after phase I, citing loss of competitiveness and marketability due to timeline delays.

Making the most of menin

Combination options will be a key to the longer-term success of menin inhibitors. “These drugs are intended for and will be used in combination,” says Wilson. Trials of chemotherapy pairings are already underway, and combination trials with various targeted agents are either planned or ongoing.

Combinations could both boost remission rates and help overcome drug resistance. In phase I, revumenib induced the emergence of resistance mutations in the gene that encodes menin, apparently a class effect.

There is clear opportunity in solid tumours too, adds Wilson, although he declines as yet to offer specifics. “We’ll try to roll the first one of those out later this year,” he says. Syndax, too, is looking at solid tumours, with a phase I colorectal cancer trial already underway, based on preclinical evidence that the menin–KMT2A interaction is important in β-catenin-driven cancers.

None of this was anticipated by the field’s originators. “We didn’t know how it would turn out,” says Grembecka. She and Cierpicki still work on menin, but have also moved on to target other protein–protein interactions in leukaemias. “We know it can be done,” says Cierpicki.