Celyad — Gaining value on stable disease observations

The THerapeutic Immunotherapy with NKR-2 (THINK) study (NCT03018405) is an open-label, multiple-dose US and European study currently in a dose escalation phase. It is assessing at higher dose levels the safety and clinical activity of autologous NKR-2 CAR T-cells (CYAD-01) in seven refractory cancers: the two hematological cancers with Phase I data (AML and MM) and five solid tumors (colorectal, ovarian, bladder, triple-negative breast, and pancreatic).

The trial is testing three dose levels: 3x108 cells (completed), 1x109 cells (currently being treated) and 3x109 cells. At each dose, a patient receives three successive administrations, two weeks apart, of CYAD-01. It is likely that the highest dose will be needed as Celyad does not use a preconditioning regimen – that is patients are not given a cytotoxic therapy that ablates existing immune cells. The avoidance of preconditioning is based on preclinical work published by Professor Sentman at Dartmouth College who originated the NKG2D CAR T-cell therapy.

Use of preconditioning in B-cell cancers allows the infused CAR T-cells to grow rapidly and become, for a while, a high proportion of the T-cells. As the tumor regresses and normal T-cells recover, the proportion of CAR T-cells drops rapidly. Without preconditioning, the NK CAR T-cells will remain as a small proportion of the overall T-cell population.

However, as noted in the July 2017 Novartis FDA advisory committee presentation on CD19 therapy in paediatric ALL, there were no specific markers or dose response correlations with patient responses.

A key aspect of NK CAR T-cell therapy is its ability to recruit an endogenous T-cell response against the cancer. This could enable successful long-term immune control of the cancer. It is also notable that, so far, NK CAR T-cell therapy has been free of common side effects seen with B-cell targeted CAR T-cells like cytokine release syndrome (CRS) and neurotoxicity (NT). This might be due to the lack of preconditioning, which limits the expansion of the NK CAR T-cells. However, as noted, CRS and NT side effects did not correlate with response in CD19 CAR therapy.

So far, Celyad has seen stable disease over three months in two colorectal patients treated at the lowest dose. There was a previous stable disease response in an AML patient in the previous study at a lower dose still. Note that in solid cancer, the use of any CAR T-cell therapy is experimental and response rates are potentially different to those seen in more tractable B-cell cancers.

The dose escalation part of the study will enroll up to 24 patients. Specifically, the trial is split into hematological and solid tumor arms with 12 patients each. The dose escalation cohorts are of three patients dosed sequentially. If a dose-limiting toxicity is noted, a fourth patient is tested at that dose, otherwise the next dose is tested. At the highest tolerated dose, a further three patients are tested to confirm safety. Overall six-month dose data are possible in H218. The two-year primary endpoint data could be due in mid-2020. However, there should be no delay in starting the expansion of different cancer types in H218. In the expansion stage, individual cancer types will be recruited with up to five solid cancer arms plus two possible hematological arms. This phase requires 14 patients per cancer: 98 in total made of 86 further patients plus the 12 at the highest dose cohort. As safety will have been shown by this stage, patients can be dosed as they are enrolled rather than sequentially as at present. This will potentially speed up the trial, perhaps allow further expansion of some indications and therefore lead potentially to BLA filings.

The Celyad natural killer (NKG2D) receptor CAR T-cell approach, CYAD-01, targets the ubiquitous stress ligands expressed by many cancers. These stress ligands are upregulated in response to chemotherapy. However, they are also expressed by normal cells exposed to toxic agents, if only for a short period. Chemotherapy might make the tumor more stressed and so more susceptible to CYAD-01 targeting. It might also expose some normal tissues to CYAD-01 so timing of dosing is important to allow enough normal tissue recovery post chemotherapy. In the real world, most solid cancer patients will receive chemotherapy and knowing how to combine standard chemotherapy with CAR T-cell therapy is crucial. If it proves to be synergistic, it would be a major cancer therapy breakthrough as most chemotherapy regimens show limited survival gains in a minority of patients.

In SHRINK, CYAD-01 doses will be adjusted to body weight and escalate from 3x108 to 1x109 to 3x109 cells. Patients will receive three CYAD-01 doses two weeks apart. The dose escalation phase will be 18 patients (six/cohort) with 21 in any expansion phase. The trial is being run in Belgium. It has not yet been posted on the clinical trials databases but is approved by the Belgium regulator.

The patients enrolled will suffer from colorectal cancer with potentially resectable liver metastases. These patients routinely receive FOLFOX chemotherapy. FOLFOX is a combination of folinic acid (leucovorin), fluorouracil (5FU) and Oxaliplatin. About 20-30% of colorectal cancer cases are found to have metastatic disease on diagnosis. FOLFOX is not a regimen that targets the immune system, unlike the preconditioning regimen used in B-cell CAR T-cell therapies. We would therefore not expect to see dramatic immediate effects of adding CAR T-cell therapy. FOLFOX is effective at shrinking tumors and multiple treatments are given. Shorter endpoints are likely to be used in initial studies. Overall survival will be crucial for widespread use, but will take some years to determine as FOLFOX gives median overall survival of about two years vs six months on supportive care.

On 4 August 2017, Celyad announced that it was paying $25m in cash and shares to receive an “increased share of future revenues” generated by intellectual property licensed from Dartmouth College. The NK CAR T-cell IP was originally licensed in 2010 by OnCyte, a subsidiary of Celdara LLC, a US private venture company. Celyad acquired OnCyte in 2015 from Celdara. The value is being paid as $12.5m cash and $12.5m in shares at $38.17/share, a 14% premium.

To reflect the change in expected revenues, we have made two major adjustments:

There are two technical adjustments:

This gives a revised indicative value estimated by Edison of US$616m, implying US$61 per ADR, Exhibit 3. This is based on only one solid indication reaching the market. This is very cautious as these estimates are already risk-adjusted.

As a scenario, if all five current solid indications are included, the value would be US$1.5bn and US$153 per ADR; note that developing, producing and marketing five solid CAR T-cell cancer indications would require extra capital, implying substantial dilution. The C-Cure cardiac therapy value of $191m requires a partner to fund the CHART-2 study needed.