Any CAR-T company with aspirations to dominate must look beyond CD19 as a target, and Juno’s chief scientific officer, Hy Levitsky, makes no secret of the fact that this will call for the standard constructs to come with “bells and whistles” attached.
One of these is a suicide gene that can be incorporated into the T cells to destroy them in the event of serious toxicity. In an interview with EP Vantage at this week’s AACR meeting Mr Levitsky spelled out his desire to own something better than Juno’s current suicide switch, raising an interesting business development question: should Juno buy Bellicum?
After all, Bellicum owns an especially clean and simple suicide technology based on the infusion of the small molecule rimiducid to trigger apoptosis. In contrast the EGFRt-based technology used by some of Juno’s constructs requires the infusion of Erbitux – something doctors dare not do for fear of an inflammatory response.
Valuing the asset
Mr Levitsky accepts the severe limitations of Juno’s EGFRt switch, and declares an interest in licensing or perhaps even buying Bellicum’s: “It all comes down to how the two parties view the value of the asset, but if that can be broached then I think [a deal] would make a lot of sense.”
Bellicum has in the past floated the idea of licensing its CAR-T suicide switch to other industry players (A CAR-T suicide switch for hire, March 12, 2015). It might appear to make little sense to license it to an arch rival like Juno, but then Bellicum has lost over half its value in the past year, so its bargaining power is much reduced.
Juno could get a first-hand look at Bellicum’s suicide technology from a trial its partner, Memorial Sloan Kettering Cancer Center, is running that paradoxically incorporates it, though Mr Levitsky stresses that this is under a purely academic arrangement.
But he cautions that there are some issues with using rimiducid, adding: “There is a lot of interest in using other kinds of drugs for this. We’re working on our own drug-inducible technology.”
One of the projects that will need to include such a technology is Juno’s anti-Ror1 CAR-T project, JCAR024, whose phase I trial is about to start enrolling. This will also need to incorporate features to improve T-cell function in an immunosuppressive microenvironment to overcome resistance to TGF-beta signalling, or adenosine pathway or PD-1-mediated suppression.
Juno is already making progress on what it calls “armoured CARs” that secrete IL-12, for instance (CAR-T meeting – To hit solid tumours use a supercar, March 14, 2016). However, Mr Levitsky, who at the AACR presented preclinical data with JCAR024, said some of these “bells and whistles... would be built into [JCAR024] only once we have an assessment of safety”.
This fits its phase I plan, which initially targets CLL and mantle cell lymphoma – indications that the science chief calls mere proof of concept. “The real motivation of course is the diseases that CD19 is not on,” he states, stressing that a second cohort will look at Ror1-positive triple-negative breast and lung cancers.
An earlier focus is Juno's anti-CD19 CARs, with a regulatory filing planned in late 2017 for JCAR015 in adult ALL. JCAR017, a slightly different construct that also targets CD19, is targeting lymphomas and CLL, and it will probably be late 2017 or early 2018 before there is sufficient data to file it.
Thus Mr Levitsky accepts that Juno could be behind Kite, whose filing in lymphoma that company claims is possible late this year, and Novartis, which is targeting paediatric ALL. Interestingly, he puts Juno’s choice of construct to indication largely down to legacy issues at the three academic institutions on which Juno’s work is based.
Another issue is that the advanced CAR constructs still use murine antigen-binding domains, though all three leading companies are pursuing humanised or fully human binders. The current version of Juno’s JCAR024 uses a rabbit-derived anti-Ror1 domain, and Mr Levitsky says the group is in lead selection for a fully human binder.
Another Juno scientific founder, Memorial Sloan Kettering’s Dr Michel Sadelain, earlier told the AACR meeting that future CAR-T work “should now only make use of human binding domains”.
The search for improved binding domains, activation boosters and suicide switches thus looks set to occupy much of Juno’s time between now and market launch. It will not go unnoticed that JCAR017 uses the EGFRt suicide switch.