Kite investors see an uncomfortable parallel with Juno

The death of a severely ill patient in a trial of a CAR-T therapy is hardly earth-shattering, but when, as Kite revealed with KTE-C19 this morning, it involves cerebral oedema it sets off some very loud alarm bells.

This is because of the precedent of Juno’s JCAR015, which was discontinued after two recent clinical trial halts due to deaths from this specific side effect. Kite stressed that in its case this was the first time it had ever seen cerebral oedema in a trial of KTE-C19, and there is no clinical halt; however, investors’ jitters are understandable, and the stock opened down 9%.

For now, Kite says its KTE-C19 clinical programme is going ahead as planned, with a protocol modification. KTE-C19’s US filing for aggressive lymphoma is under review, and Kite says it expects an advisory panel to be held, though it has yet to receive confirmation of this from the agency.

Aggressive disease

It says the death occurred in late April in a safety study of KTE-C19, its lead CAR-T project, in a patient with aggressive non-Hodgkin’s lymphoma who was in a highly inflammatory state before being dosed with the CAR-T cells.

The patient had had rapidly progressing disease, having responded inadequately to two lines of therapy, and fulfilled all the criteria to receive KTE-C19, also known as axicabtagene ciloleucel. He developed cytokine release syndrome and neurotoxicity, leading to multiple organ failure and ultimately death from cerebral oedema two days later.

Cytokine storm and neurotoxicity are well-documented adverse events of CAR-T therapy, but it was the specific occurrence of cerebral oedema – swelling of the brain – that caused Juno such problems with JCAR015.

That project’s pivotal Rocket trial was halted in June after two deaths due to this complication, but was swiftly restarted after Juno agreed with the FDA that these were due to fludarabine chemo conditioning. But despite a protocol amendment two further cerebral oedema deaths occurred in November, and it was curtains for JCAR015 (Why this could be strike three for Juno’s lead, November 23, 2016).

While patient numbers are far too small to draw firm conclusions, an uncomfortable fact is that KTE-C19 is an extremely similar CAR construct to JCAR015, differing only in the binding region it uses. The precise cause of cerebral oedema with CAR-T is not understood, but it had surprised some that none had previously been seen with KTE-C19, given its similarity with JCAR015.

No less an authority than Dr Stephan Grupp, of University of Pennsylvania, Novartis’s partner, had put the cerebral oedema problem squarely on the CD28 co-stimulatory domain used by JCAR015 and KTE-C19, but not Novartis’s CTL019.

Others had suggested that it was the specific setting of adult acute lymphoblastic leukaemia in JCAR015’s Rocket study that had contributed. As of today investors know that this is an unlikely theory, since the KTE-C19 death occurred in lymphoma.

At one point Juno had suggested that the relatively young ages of the adult ALL patients affected in its Rocket study might have been important too.

Safety record

In Kite’s defence, KTE-C19’s safety record is better than JCAR015’s: Kite told EP Vantage it had treated some 200 patients and seen just one case of cerebral oedema, with a roughly 2% incidence of KTE-C19-related deaths; including NCI studies over 300 patients have been treated with KTE-C19, it says.

However, with so many unknowns about CAR-T, investor nervousness is understandable. And keenly awaited lymphoma data from Novartis’s Juliet trial of CTL019, to be presented at the International Conference on Malignant Lymphoma in Switzerland in June, have just become even more important.

Juno followers were relieved that when their company canned JCAR015 it already had an advanced alternative asset, JCAR017, to pivot to. It is still early days, but cynical investors will note that Kite has no such luxury.

To contact the writer of this story email Jacob Plieth in London at [email protected] or follow @JacobPlieth on Twitter