Seeking alpha in PI3K inhibition

It seems like the PI3K mechanism has received the last rites several times now, but nevertheless it refuses to die. And Novartis’s success last week with alpelisib in the pivotal Solar-1 trial renewed hope that the approach could have promise against breast cancer.

Of course, it is all about specificity now, and it is thought that in certain uses it is the alpha subtype of the kinase that needs to be targeted. The Swiss firm says alpelisib is “the only alpha-specific PI3K inhibitor” – a claim that an analysis of the competitor pipeline reveals to be incorrect.

In fact, alpelisib is one of four alpha-specific PI3Ks in clinical trials. These are TAK-117, a phase II agent from Takeda, and two phase I assets: Shanghai Haihe Pharmaceutical’s HH-CYH33, which is in a Chinese trial listed on Clinicaltrials.gov, and RG6114 from Roche.

The industry pipeline also includes a further two agents specific for the alpha as well as the beta isoform: Bayer’s BAY1082439 recently completed phase I for unspecified advanced cancers, while Onconova’s preclinical ON 146040 is up for licensing out as the company focuses on rigosertib.

While the Novartis claim is thus wrong, the company is correct in stating that alpelisib is the first alpha-specific PI3K inhibitor to show a potential benefit with acceptable tolerability.

Roche's taselisib illustrated why lack of specificity is problematic. The asset initially showed promise in the phase II Lorelei trial in ER-positive/Her2-negative breast cancer; though the effect was borderline it was driven by subjects carrying PIK3CA mutant cancer cells, supporting the mechanistic rationale for hitting the alpha isoform (Esmo 2017 – Roche’s Lorelei gives PI3K agents another second chance, September 8, 2017).

However, it all went wrong at Asco, when taselisib’s Sandpiper trial showed very modest efficacy. Roche blamed an unacceptable level of toxicity on the fact that the compound also hit delta and gamma subtypes. Taselisib was subsequently abandoned in favour of the more targeted RG6114.

Setbacks

Novartis too has had its setbacks, studying numerous PI3K inhibitors over the years. For instance, buparlisib, a pan-PI3K inhibitor, flunked breast cancer trials but is still in the clinic in haematology.

Its most advanced asset, however, is alpelisib, and the design of Solar-1 shows that lessons from Roche are being learned. The trial specifically recruited HR-positive/Her2-negative breast cancer patients who had PIK3CA mutations, which Novartis says drive some 30% of HR-positive breast cancers.

The full Solar-1 data have not been revealed, and all Novartis disclosed last week was that the trial met its primary endpoint of progression-free survival, with side effects “generally consistent” with previous studies. The group says the data are strong enough for it to begin regulatory filing discussions.

In the meantime, others focused on this mechanism will have to wait a little longer: TAK-117 is being studied in triple-negative breast cancer in combination with sapanisertib, a Torc1/2 inhibitor, but this trial will not read out for a couple of years, a similar timeframe to phase I studies of HH-CYH33 and RG6114.

It has been quite a journey from the disappointing launch of Gilead’s idelalisib, a delta-specific agent with effects on B-cell signalling, through Infinity’s sale of develisib to Verastem, to the current interest in the alpha isoform.

Novartis says mutations in beta, delta and gamma PI3K subtypes are not normally associated with breast cancer. It will be hoping that with alpelisib a therapeutic window has been found at last.