Safety worries around targeting CD123 will only intensify with news of a patient death with Cellectis’s off-the-shelf CAR-T therapy UCART123. Others working on projects against CD123, including Johnson & Johnson and Novartis, will be hoping that this is not a problem common to this target, something that now looks like a distinct possibility (see table below).

Two ongoing phase I studies of UCART123 have now been suspended and, while Cellectis hopes to restart them with a lower dose, the FDA will surely be cautious after a string of problems with similarly acting agents.

It is unclear whether the CD123 antigen is the issue, but three safety scares over the past year could be more than a coincidence. First, a study of Genmab and Johnson & Johnson’s bispecific antibody JNJ-63709178 was suspended last September after a unnamed serious adverse event – although the clinical hold was lifted in April.

Then came two deaths in a pivotal study of Stemline Therapeutics’ SL-401, but this trial has continued (Stemline patient death becomes secondary investors’ problem, February 03, 2017).

The death with UCART123, in the first patient treated in a phase I study in blastic plasmacytoid dendritic cell neoplasm (BPDCN), was blamed on cytokine release syndrome and capillary leak syndrome. While the former is a known side effect of CAR-T therapy, the latter has also been associated with Stemline’s SL-401, suggesting that it could be an issue with hitting CD123.

And this was not the end of the bad news for Cellectis. The first patient in a second phase I trial of UCART123, in acute myeloid leukaemia (AML), also experienced grade 3 cytokine release syndrome and grade 4 capillary leak syndrome that put her into intensive care, though these adverse events were resolved.

Cellectis’s project incorporates a Rituxan-based “suicide switch” to inactivate the T-cell therapy in the event of safety problems, but it seems that the trial investigators had not attempted to use this. One drawback with this approach is that infusing Rituxan can trigger an inflammatory response, so would be risky in frail patients.

This has raised questions about whether this switch is a real safety solution. In any case, Cellectis executives had told EP Vantage at last year’s Ash meeting that the company was working on different suicide switches, which could encompass small molecules or monoclonal antibodies.

Not as easy as 123

Targeting CD123 could be problematic because, while it is overexpressed in haematological cancers, it is also present on many healthy cells, which could lead to on-target, off-tumour effects.

When asked about this potential issue at last year’s Ash conference, Cellectis’s chief medical officer Loan Hoang-Sayag seemed unconcerned and said that the company remained confident in spite of the problems with JNJ-63709178.

Nonetheless, the latest death is yet another worry for others attempting to hit CD123. Johnson & Johnson is well represented here with two candidates: the aforementioned JNJ-63709178, and talacotuzumab, the most advanced CD123-targeting asset, which is in a phase III trial in AML due for completion next year.

However, it is smaller companies like Macrogenics that have more to lose. This group will present phase I data with its bispecific antibody flotetuzumab in AML and myelodysplastic syndrome at this week’s Esmo meeting in Madrid. Based on the abstract released safety appears to be acceptable so far, but this is something that will remain closely watched.

As for Cellectis, it believes it has a way forward for UCART123: reducing the dose from 6.25x10⁵ to 6.25x10⁴ cells per kilogram. However, the French group’s share price was down as much as 31% today, suggesting that investors do not fancy its chances of finding a therapeutic window here.

The company has a more advanced allogeneic CAR-T project, UCART19, in phase I – this has a better-validated target in CD19, but this area is already crowded, and in any case the asset is now in the hands of Pfizer and Servier.

If CD123 turns out to be a dud, it would be bad news for many companies – but it would be a disaster for Cellectis.

This story has been updated to reflect the fact that talacotuzumab is a MAb, and the addition of Seattle Genetics' SGN-CD123A.

To contact the writers of this story email Madeleine Armstrong or Jacob Plieth in London at news@epvantage.com or follow  @ByMadeleineA or @JacobPlieth on Twitter