No place for Wee1 in Astra’s synthetic lethality work

Nine years after licensing the Wee1 inhibitor adavosertib from Merck & Co, Astrazeneca today pulled the plug on the project – though not, apparently, on the concept behind it.

A big push is instead being made with ceralasertib, a molecule that hits ATR; just like Wee1 this target induces so-called synthetic lethality but without as much toxicity, Astra’s head of oncology R&D, Susan Galbraith, tells Evaluate Vantage. Roche recently paid $125m for Repare’s ATR inhibitor camonsertib, and the small group of competitors still active in Wee1 inhibition might view such moves anxiously.

Chief among these is Zentalis, whose phase 2 Wee1 inhibitor ZN-c3 is its lead asset. At this year’s AACR Zentalis touted promising initial clinical data with ZN-c3 in heavily pretreated ovarian cancer; on the back of this it secured a $25m equity investment from Pfizer and raised $200m from investors – no mean feat in the current biotech market.

Toxicity, not patents

But what Astra has found with adavosertib, which clinicaltrials.gov reveals has been through 19 studies that the group has lead-sponsored, will be of concern, though it must be stressed that each molecule has different properties.

“The challenge with adavosertib has been the tolerability profile,” says Ms Galbraith, citing GI and bone marrow toxicities that “can be dangerous, particularly in vulnerable patients”. This meant that Astra had to lower dosing to minimise this risk, squeezing out the project’s therapeutic window.

Such broad concerns over the Wee1 target could cause alarm because until now the main problem with adavosertib has been its short patent life. Ironically the project’s discontinuation comes shortly after an academic group appeared to find a niche in which it worked.

“Undoubtedly inhibition of Wee1 has activity, and I remain convinced that it is an important target in cancer,” stresses Ms Galbraith. “But we have to prioritise resources.” This has seen Astra advance the ATR inhibitor ceralasertib instead. This can work in some of the same settings as Wee1, and Ms Galbraith describes it as less problematic in terms of toxicity.

Wee1 and ATR inhibitors both work by interfering with a tumour’s DNA repair mechanisms, a concept known as synthetic lethality. Incidentally this also encompasses Parp inhibitors like Lynparza.

Ms Galbraith has been particularly enthused by the Hudson trial, in which Imfinzi plus ceralasertib has shown “intriguing” survival results in NSCLC patients who had already failed PD-(L)1 blockade. This, she says, suggests a close interaction between DNA damage response and the immune system.

While the Wee1 inhibitor field has thinned, there are several other groups active in ATR inhibition, most notably Merck KGaA with berzosertib, licensed from Vertex for $230m in 2017, and Roche with camonsertib, the subject of a $125m tie-up with Repare Therapeutics last month. Roche does not appear to have worked clinically on Wee1 inhibition.

Though ATR inhibition is the domain of small molecules, Astra reckons it shows a way in which immuno-oncology can be segmented based on the immune mechanism involved. As such it makes up part of the “next wave of the I-O space”, according to Ms Galbraith.

But the group continues to work on strict I-O targets too, including bispecifics combining PD-1 inhibition with CTLA-4 or Tigit blockade. The former features a MAb called MEDI5752, which only binds CTLA-4 in the presence of PD-1, something that preclinically showed a better safety profile than hitting these targets using separate MAbs.

MEDI5752 is still causing liver enzyme elevations, and Astra has yet to find an optimal dose. Meanwhile, the PD-1xTigit bispecific AZD2936, derived from Compugen, entered the phase 1 Artemide-01 NSCLC trial last year, before Roche’s setbacks with the pure anti-Tigit MAb tiragolumab in SCLC as well as NSCLC.

Notably, Astra is hedging its bets here too, separately studying a combo of Imfinzi with Arcus’s anti-Tigit domvanalimab in the phase 3 Pacific-8 trial.