AACR 2023 – Regeneron tries again in multiple myeloma

Regeneron, whose efforts in T-cell engagers have been hit by toxicity, and whose multiple myeloma work has barely distinguished itself in terms of efficacy, is not giving up. Today the group revealed REGN5459, an anti-BCMA construct it has quietly been developing that it hopes could be better than its underwhelming lead BCMA asset, linvoseltamab.

A first-in-human study just unveiled at AACR suggests that REGN5459’s efficacy could be higher than linvo’s, though claims to a vastly improved toxicity profile seem illusory. Still, it is hard to ignore efficacy that at the project’s recommended phase 2 dose seems to beat Johnson & Johnson’s approved Tecvayli, and might even approach that of Car-T therapies.

Tecvayli was approved last October for fifth-line multiple myeloma, on the strength of overall and CR rates of 62% and 31% respectively. It carries a boxed warning of cytokine release syndrome and neurotoxicity, and its label cites a 54% rate of serious adverse reactions.

Against this background numerous groups, including Regeneron, paraded their rival anti-BCMA T-cell engagers at last year’s Ash meeting. None stood out as being meaningfully better, and respective 64% and 24% ORR and CR rates at linvo’s recommended phase 2 dose were accompanied by a 6% rate of deaths due to infection.

Step forward REGN5459

Now the company says a phase 1/2 trial of REGN5459, in at least fourth-line patients, has managed to yield similar efficacy: ORR and CR rates were 65% and 33% respectively across all 43 patients, AACR heard today.

The presenting author, Attaya Suvannasankha from the Indiana University Simon Cancer Center, said REGN5459 had been designed to yield low rates of cytokine release. But was this reflected clinically?

Cytokine release did occur in 54%, but just 5% occurred at grade 3, and none at grades 4 or 5. Only one patient developed neurotoxicity, and that was at grade 2. In comparison, at Ash linvoseltamab resulted in a 44% rate of cytokine release (1% at grade 3, none at grade 4), and a 1% rate of severe neurotoxicity.

As such it is hard to see a meaningful advantage on a cross-trial basis, though Suvannasankha said step-up dosing had additionally been employed for the highest REGN5459 doses, resulting in all of the cytokine release seen (81% rate for these doses) being grade 1 or 2.

And cytokine release is not the whole story. A major problem with linvoseltamab was infection, which was seen in 54% of patients, and resulted in 14 deaths (6% of the treated population). In the REGN5459 trial infection was seen in 63% of patients, and there were two deaths (5%), one from Covid and one from pneumonia.

So perhaps the main advantage of REGN5459 is not safety, but increased efficacy. Suvannasankha highlighted ORR at the highest IV doses, saying this came in at 91%, including a 38% CR rate among 21 patients.

Clearly this kind of efficacy needs to hold up in a larger number of patients, but on the face of it REGN5459 looks better than Tecvayli, and not that far off the 98% ORR and 78% complete response rate yielded by J&J/Legend’s marketed Car-T therapy Carvykti.

Suvannasankha said REGN5459 was designed to bind relatively loosely to CD3 on T cells to reduce cytokine release, and the surprise was that this did not hamper efficacy. “How to mitigate the intensity of the binding of such a molecule is something that needs to be further explored,” she told an AACR press briefing.

This might be an intriguing finding but, as Regeneron’s other T-cell engagers have shown, toxicity is a major problem for the company, despite its tinkering with dosing regimens. The REGN5459 trial will not allay concerns.