Roche takes the Tim-3 battle to Novartis

One day after Novartis delayed the results of sabatolimab, the industry’s most advanced anti-Tim-3 asset, Roche has highlighted its move of RG7769/RO7121661, a bispecific combining this mechanism with PD-1, into phase 2.

This is not the only pipeline change it revealed in its second-quarter update today, with earlier industry developments likely spurring the discontinuations of BCMA, CD40 and ophthalmology assets and disappointing investors in Affimed and 4D Molecular. Meanwhile, a hepatitis C antiviral licensed from Atea and repurposed for Covid-19 has moved into phase 3.

That last move is especially interesting as industry’s Covid-19 antiviral efforts have come off the boil, with Merck & Co canning MK-7110, and limiting molnupiravir work to outpatients. Roche had licensed RG6422/AT-527 from Atea last October, and phase 2 trials have been enrolling for some time; the new phase 3 study highlighted today seeks to recruit 1,386 outpatients, and compares the antiviral against placebo.

Novel IO

Meanwhile, Roche moved the Tim-3/PD-1 bispecific RG7769 into phase 2 before Novartis yesterday disappointed followers of novel immuno-oncology mechanisms and put the spotlight on Glaxosmithkline’s Tim-3 bet with cobolimab. Today’s Roche presentation spells out the design of the new trial, called Talios, in refractory oesophageal squamous cell carcinoma.

This has a three-arm design, also including a second active cohort testing the Lag3/PD-1 bispecific RG7769, and comparing this and RG7769 against Opdivo in a blinded fashion. With overall survival as the primary endpoint Telios might even be registrational, and its design appears geared specifically towards showing what Lag3 or Tim-3 blockade can add on top of PD-1, though whether a fixed bispecific is the way to show an incremental benefit will be a big question.

Roche’s oncology pipeline has also seen the discontinuation of selicrelumab, which had been in several phase 1 combo studies. This molecule is a CD40 agonist MAb, and its shelving might be linked to Johnson & Johnson’s decision two years ago to discontinue the similarly acting ADC-1013/mitazalimab, which it had licensed from Alligator Bioscience.

Still, Roche is not ending its pursuit of this pathway; RG6189, a previously undisclosed bispecific that induces CD40 stimulation solely in the presence of fibroblast activation protein α (FAP), has moved into a phase 1 Tecentriq combo trial in solid tumours, Roche said today.

These developments will be closely followed by the private biotech Apexigen, whose lead asset, APX005M/sotigalimab, is in phase 2, making it the industry’s most advanced CD40 agonist. Mitazalimab also remains in development as Alligator’s most advanced project.

The additional discontinuation of RG6296 will disappoint fans of Affimed, which has impressed with its bispecific approach to engaging NK cells. The project is the result of a 2018 tie-up between Affimed and Roche that focused on NK cell engagement.

However, since the antigen-targeting moiety of RG6296 hits BCMA its canning makes perfect sense: BCMA is now a hugely competitive space, involving late-stage Car-T assets showing strong efficacy against multiple myeloma, not to mention other bispecifics and conjugates, and this is probably too much to make an early-stage project worth the investment.

Finally, Roche has also confirmed the scrapping of two gene therapies licensed from 4D Molecular, 4D-110 for choroideraemia and 4D-125 for X-linked retinitis pigmentosa. 4D had announced the termination of the Roche deal last month.

This was driven by a “change in risk-benefit profile”. It will not escape anyone’s notice that gene therapies for eye disease have had a tough time of it, with Biogen, Adverum and Regenxbio all stumbling on toxicity or lack of efficacy.