Another gene therapy safety scare hits

Anyone hoping that last week’s benign FDA advisory committee meeting about gene therapies might restore confidence in this troubled space will have been disappointed this morning. Biomarin is the latest group to be hit by a safety scare, with its phenylketonuria project BMN 307 being put on clinical hold by the agency today.

There are reasons to believe that this could be a temporary blip: the hold came after tumours were seen in mouse studies, and a previous link between other AAV gene therapies and liver cancer in mice has not been replicated in humans. But, after a recent string of bad news, the latest development reinforces the concern that patients might choose to steer clear of gene therapies, especially if other options are available.

Unmet need

In PKU Biomarin already has two approved therapies, Kuvan and the newer Palynziq, which are forecast to sell $86m and $619m respectively by 2026, according to Evaluate Pharma sellside consensus. Both drugs are designed to reduce blood phenylalanine, which in PKU is not broken down owing to deficiencies in the phenylalanine hydroxylase (PAH) enzyme, caused by mutations in the PAH gene.

Still, these drugs leave a lot to be desired: Kuvan is only suitable for a fraction of patients and must be used alongside a restrictive diet, while Palynziq comes with a black box warning and is given by daily subcutaneous injection.

This unmet need explains why various companies are trying to get into the PKU market, although Rubius’s cell therapy RTX-134 failed to make a mark in the disease, while Sangamo has discontinued its gene therapy ST-101.

Whether Biomarin’s BMN 307 will join these projects on the scrapheap depends on the latest findings in mice being shown to be relevant to humans.

The preclinical trial found that six of seven animals receiving the highest dose of the project, 2x10¹⁴vg/kg, had tumours a year after dosing, with evidence of integration of the AAV vector into the genome.

This raises the spectre of insertional mutagenesis, wherein a vector-delivered gene tampers with a target cell’s chromosomal arrangement, triggering cancer. This has long been one of the big theoretical worries about gene therapy in humans, and was discussed at last week’s adcom.

Of mice and men

Other AAV gene therapies have led vector-induced liver cancers in neonatal mice, the panel heard.

Biomarin’s study used mice bearing two genetic mutations: one eliminating the PAH gene and another rendering the animals immunodeficient. The company noted that these mutations might predispose the mice to developing cancer.

Biomarin also stressed that cancers due to AAV integration have not been seen in larger animals or humans. But Leerink analysts covering the panel wrote that there is evidence that AAV vectors can integrate into the dog and non-human primate genomes, although there has been no link so far between integration and tumourigenesis.

Late last year, concerns about insertional mutagenesis came to a head when a case of liver cancer was seen in the pivotal study of Uniqure’s AAV5-based haemophilia B gene therapy, etranacogene dezaparvovec. This was later deemed unlikely to be due to the project, but the group’s stock has still not recovered.

BMN 307, meanwhile, is already in a phase 1/2 study, in which it is being dosed at 2x10¹³vg/kg or 6x10¹³vg/kg – notably lower than the dose in the mouse study. There is scope to add another dose group, but the company said during its second-quarter call that, based on the evidence so far, it could take forward 6x10¹³vg/kg.

Unless the hold is lifted, though, the question of the ideal dose is academic. In terms of the broader gene therapy space the FDA panel recommended against capping the dose of gene therapies, something that had been mooted ahead of the meeting.