Leap failure sends Clavis and Clovis crashing back to earth

The phase II study of CP-4126 could have been a leap forward in the notoriously difficult setting of pancreatic cancer, but instead it has turned into a humbling step back for the project’s developers, Clavis Pharma and Clovis Oncology.

The companies insist that the Leap trial had been well designed and executed; the only thing wrong with it seems to have been the result, which showed not even a glimmer of potential for CP-4126. All work on the project, which analysts had touted as a future blockbuster, will now cease. Not surprisingly the Norwegian company’s stock crashed 85% this morning, while Clovis was off 35% in early trade.

Leap had been designed around the relatively simple concept that efficacy of gemcitabine, currently the main drug for pancreatic cancer, is limited by tumour cells’ ability to internalise the molecule. Gemcitabine tends to be taken up via the hENT1 transmembrane protein, and tumours with low hENT1 – present in perhaps 50-65% of patients – are therefore resistant to the drug, it had been postulated.

Unusual design

Leap’s design was unusual in that it recruited 367 pancreatic cancer patients regardless of their hENT1 status, but then looked at only those with low levels of hENT1 to determine the primary endpoint of overall survival (Event – Clovis hopes for a great leap forward with trial results, September 3, 2012).

CP-4126 consists of gemcitabine with a lipid tail attached that effectively enables it to enter the cell irrespective of the hENT1 carrier system, and in Leap all participants were given either CP-4126 or generic gemcitabine.

But in the event all patients, irrespective of hENT1 status or whether they received CP-4126 or active control, survived for around six months – in line with prior studies of gemcitabine in this setting. The two drugs “behaved exactly the same” across several analyses, said Clovis, which had licensed rights to CP-4126 from Clavis three years ago.

It was particularly odd that the hENT1-low population did not do worse on gemcitabine than hENT1-highs – a finding that flew in the face of prior literature.

So what went wrong? The companies insist that the study was well designed and, having carried out various subgroup analyses and failed to spot any obvious flaws, said the finding came as a complete surprise. Still, Leap was the first ever study to test the hENT1 hypothesis prospectively in a large patient group, prior positive data having come either from tiny trials or from retrospective analyses.

If there was nothing wrong with Leap’s design, and assuming that CP-4126 did indeed circumvent the hENT1 carrier, then the result calls into question the entire hENT1 hypothesis. There had been suggestions, for instance, that more advanced tumours tend not to express the carrier, and their intractability is due to their advanced nature rather than hENT1 status.

John Savin, an analyst with Edison Investment Research, said he was surprised that there was no effect at all, but admitted to having had concerns about other resistance mechanisms that might play a role in addition to hENT1.

Another intriguing theory is that owing to molecular weight differences between the molecules the 1,250mg/m2 dose of CP-4126 in Leap actually amounted to only half of the 1,000mg/m2 at which generic gemcitabine was given – an obvious potential confounding factor. Clavis insisted that dosing was checked again, along with many other possible variables, and was “found to be appropriate”.

End of the road

The company admitted that CP-4126 development would be scrapped for all indications, including lung cancer, where a phase I trial is under way. It expects Clovis to return rights to the project.

On a call today Clovis said that based on a forecast $140m of year-end cash it had around a two-year runway, and although a fund raising would be necessary at some point it had enough money to take two other oncology projects – CO-1686 and rucaparib – into pivotal development.

Clavis, meanwhile, will now turn to elacytarabine, which is due to yield data from the phase III Clavela study in acute myeloid leukemia next year. But things for the company look bleak: with a market cap of NOK300m ($52m) – barely above its year’s worth of cash on hand – it is now entirely dependent on a positive result in Clavela, which it hopes will lead to a US licensing deal.

After Leap it is hard to see the Norwegian company surviving another late-stage clinical setback.

To contact the writer of this story email Jacob Plieth in London at [email protected]