The markets have penalised Nektar heavily for the waning efficacy of its lead project, bempegaldesleukin, in the Pivot-02 study’s melanoma cohort, so the company worked hard to find an explanation. Yesterday it provided one, but instead of being reassured shareholders punished it again.
In the premarket Nektar stock was off as much as 30%. It appears that investors either do not buy the company’s suggestion for bempegaldesleukin’s declining remission rates – that some subjects got “suboptimal” product batches – or that they believe this explanation calls into further question the group’s clinical work.
In the absence of further information either scenario looks possible. On an analyst call yesterday Nektar said new assays had revealed that two of its earliest production batches differed from others, and these two batches correlated with study subjects who yielded relatively low response rates to bempegaldesleukin.
The group also stated that it had identified the cause of the product inconsistencies, putting this down to a “single suboptimal batch of in-process intermediate that was used to produce only these two lots”. It has manufactured 22 lots to date.
The new analysis had been carried out to try and clarify why, as more Pivot-02 subjects yielded data, remissions in the study fell from 64% in 2017 to 53%; in PD-L1-negative subjects, a key focus for bempegaldesleukin, the decline was from 60% to 43% (Asco 2019 – the case for Nektar sweetens, slightly, June 3, 2019).
Releasing suboptimal clinical material raises doubts about the integrity of Nektar’s quality control procedures, and an obvious query must be why the company had allowed the two batches in question to be used in Pivot-02.
In its defence Nektar told analysts that the two dodgy lots were made early in its production campaign, and at the time had fallen within specifications. Only once it had identified that these lots were outliers relative to the others was it able to look for data differences in patients given substandard versus correct product.
Subjects dosed initially with bempegaldesleukin from correct batches achieved an ORR of 75%, including a 44% complete response rate; in those given the two faulty batches ORR was 36% and CR 27%, the company claimed.
Unfortunately for Nektar the sellside did not really go for the story.
Evercore ISI struck the killer blow, pointing out that over the course of a study patients are given product from several batches; while Nektar’s analysis concerned the source of only the first dose, initially non-responding subjects might be expected to respond if later given correctly manufactured product.
As for whether the newly identified issue could affect future readouts, Nektar said some subjects in Propel, a small phase I trial, had also got bempegaldesleukin derived from the dodgy batches.
Data from a first-line NSCLC cohort in Pivot-02 had been expected at Esmo, but this will apparently now not happen. Nektar said there was “some early evidence” that those NSCLC subjects started off on suboptimal product were experiencing relatively small target lesion reductions.
Bempegaldesleukin, a CD122-biased IL-2 receptor agonist, was licensed to Bristol-Myers Squibb for $1.85bn up front, and still carries sellside forecasts amounting to a product NPV of $4.8bn, according to EvaluatePharma consensus.
Though Nektar says it has implemented a comprehensive control strategy, Evercore said the snafu was yet another reason to be sceptical of its results.
“It’s hard to tell if there’s a legitimate issue with the batches or [if this is] another creative way to explain away our concerns that bempegaldesleukin has not been dosed to have meaningful clinical activity,” wrote Evercore’s Josh Schimmer.
This story has been corrected to reflect Nectar's identification of the outlier batches.