EHA 2022 – Caribou’s turn to run into allo Car-T relapses

Caribou has hailed its Crispr-edited CD19-targeting lead project, CB-010, as the first ever off-the-shelf Car-T to show a 100% complete response rate. Unfortunately, this does not tell the whole story: as the group’s EHA poster revealed today, few of the remissions are durable.

Caribou thus appears to be the latest allogeneic Car-T player to run into the problem of relapse, after this issue last year derailed Allogene, Crispr Therapeutics and Precision Biosciences. On the plus side, all of Caribou’s responses occurred with the initial CB-010 dose, and with no overwhelming toxicity concerns the company at least has scope to dose higher.

The data come from CB-010’s first-in-human Antler study, which in an EHA abstract was revealed to have put the first five evaluable lymphoma subjects into remission, with one complete response (CR) ongoing at over eight months. At EHA today the data as at a May 13 cutoff were presented.

This showed the situation to have developed: six subjects are now evaluable, five of whom were in CR at their first post-treatment scan. The sixth went into partial response, which a month later deepened to a CR. This is a promising start.

But by six months three of the six patients relapsed with progressive disease. Two of the three remaining CRs were ongoing beyond six months, and Caribou added that the first patient treated remained in CR at their 12-month scan, after the abstract’s cutoff date. All six subjects received the first dose level of 40 million cells, but it will not go unnoticed that the most durable CR is in follicular lymphoma, a relatively slow-growing disease.

CB-010 uses Crispr to knock an anti-CD19 Car transgene into the Trac locus, to avoid graft-versus-host disease (none has been seen in Antler), as well as knocking out PD-1 to limit T-cell exhaustion.

Caribou, which raised $304m in an IPO a year ago, said Antler was now enrolling at the second dose level, 80 million CB-010 cells. CB-010’s safety profile appears good, with grade 3 toxicities including just one neurotoxicity and no cytokine storm, so the hope for more durable remissions with higher dosing is not unrealistic.

Redosing?

This seems a better plan than trying to rescue relapsed patients by redosing, something that an off-the-shelf therapy enables much more easily than an autologous product, and which has been done by Allogene, Crispr and Precision.

However, just this week Precision showed that this strategy brought important concerns of its own, perhaps the biggest of which is the need to lymphodeplete patients every time they get another dose of cells, to ensure that these engraft.

Among 12 selected patients given the company’s Arcus nuclease-edited PBCAR0191 there were four deaths attributed to the fludarabone lymphodepletion regimen. All 12 subjects had relapsed after autologous Car-T therapy, so the data showed the risk of putting patients who do not have durable remissions through another round of lymphodepleting chemo.

PBCAR0191 had disappointed at last year’s Ash meeting, when an initial 71% ORR melted away as most patients relapsed by around six months. This week's update comprised 12 patients: six post-Car-T, CD19-positive responders highlighted in that dataset – two are still in response at around nine and 18 months – and six given new dosing, with reduced lymphodepletion. Four of the latter are in CR, one at nearly five months.

Earlier, allo Car-T contenders from Allogene (ALLO-501) and Crispr (CTX110) were hit with problems over lack of durability. Similar concerns have been hinted at with Car-NK approaches from Fate and Nkarta, while the question of persistence with Adicet’s gamma-delta Car-T cells remains unanswered after the company’s Asco presentation on Monday.

Precision has touted its follow-on project PBCAR19B, which it hopes might evade rejection by the host’s T cells. With Caribou opening off 20% this morning it seems clear that relapse is a problem that is not about to go away.