FDA Proactively Questions Eteplirsen Biomarker Methodology

Orphan drugs tailored to address defined genetic lesions account for a significant portion of the current value of RNA Therapeutics.  One of the reasons is that they promise much more economical approval pathways, including shorter timelines, smaller patient populations, and biomarker-based approval.  Following RNAse H antisense Kynamro by ISIS Pharmaceuticals/Genzyme Eteplirsen for homozygous FH, eteplirsen by Sarepta for Duchenne Muscular Dystrophy (DMD) offers us another front-row view of this process.

Currently, the question on most observers' minds is whether the biomarker data gathered so far, most notably from a phase IIb trial in just 12 patients is sufficient for accelerated approval (AA).  Towards this end, Sarepta and the FDA had a meeting a few weeks ago to discuss whether the FDA would accept for review such an AA submission.  According to the Minutes of the meeting which were discussed in a press release by Sarepta on Monday, the FDA does not believe the data it has seen warrants an AA submission and asked the company to come back with more data.

I am one of the very few open critics of eteplirsen and have explained my rationale why I have little confidence in the phase IIb data that have catalyzed much popular support: a manipulated trial conduct where essentially all the supportive evidence was collected after unblinding of the trial and a cherry-picking way of presenting them.

What emerged from the recent FDA interaction appears to be largely consistent with my concerns.  Today, I would like to highlight a critical issue when it comes to pursuing biomarker-based approval strategies and that was the focus of the recent interaction.

How you measure biomarker is as important as the biomarker per se

It does not reflect well on a company seeking accelerated approval when, after two decades or more of morpholino-based drug development, it has yet to establish sufficient manufacturing capacity to satisfy even the small patient population that would be eligible for eteplirsen.  The same poor planning becomes evident when the FDA appears to be questioning the method by which the biomarker (dystrophin) data was collected: what is the value of a biologically strong biomarker if you can’t reliably measure it to draw comparisons?

The dystrophin expression by immunofluorescence and Western blot were from biopsy samples.  Obviously, muscle tissue is everywhere in the body and any drug efficacy will vary depending on the muscle.  Even within a given muscle, (revertant) dystrophin expression is known to be variable in DMD patients, so despite of consistently taking biopsies from the same muscle before and after treatment, you may not be able to tell 2 or 3-fold differences.  This problem was noted by the investigators in the first phase II trial of eteplirsen in discussing the discrepancies of Western blot and immunofluorescence data.

Of course, this sampling issue could be addressed with large patient numbers, but not with the 12 as in this trial.  Given the importance of methodology, it is surprising that Sarepta missed the opportunity to provide the FDA with sufficient related information to be in a position to more conclusively tell whether it’s worth submitting for AA or not.

Issues the FDA might want to know more about could be the selection strategy of biopsy location and whether the antibody was appropriately chosen also for taking into account revertant fibers.  Revertant fibers are a phenomenon due to either secondary mutations or alternative splicing which leads to dystrophin expression in many DMD patients.  This might also shed light on for example why after 12 weeks on drugs, no dystrophin expression was seen, but at 24 weeks it was. 

Another important question that the detailed methods might answer is the quantitation of the absolute amount of dystrophin expression (not percent fibers expressing).  Based on the phase II study publication, it appears to be quite difficult to do even just semi-quantitative calculations on fuzzy Western blots.  Of course, once the real expression is established, the question is how functional such alternatively spliced (Becker-type) dystrophin is.  Maybe what you need is at least 30% of such dystrophin to have a functional impact on DMD.  

Sustained 6 MWT stabilizations common

Supporters of eteplirsen like to point out that the proof that it works is in the apparent disease stabilization over 52 weeks or more when measured in terms of the 6-minute-walk test.  However, it appears from natural history studies of DMD that such walking stabilizations are common.  In fact, depending on the age, you would even expect an increase in walking ability over 52 weeks.  Thus, arguing that the 6 MWT data are proof for drug efficacy is like arguing that Stable Disease in an uncontrolled cancer trial is evidence that a drug works.

Confounding the 6 MWT data is the fact that it is an effort-based test and that the separation between drug and placebo was only observed after unblinding of the data and dropping out the worst performers. In the very words of Sarepta itself and clinical collaborators when discussing clinical data from the competing exon-skipping drug by Prosensa and GSK:

'In terms of the clinical efficacy, the PRO051 study claims that eight of 11 boys who were ambulant at entry to the extension study showed improvement in the 6-min walking test of 35·2 m (SD 28·7) after 12 weeks' treatment; however, this change was not significant. Moreover, several of these children were younger than 7 years and, according to longitudinal observation, boys younger than 7 years with Duchenne muscular dystrophy gain motor function. Additional confounding factors are the variability in the walking test (SD 36 m) and the powerful placebo effect of open-label studies. Despite these limitations, this observation is encouraging.'

I don’t hold out myself to be an expert in DMD, but as a skeptical scientist I believe there remain too many open questions around the eteplirsen data and trial conduct to make me feel comfortable.  Add to this the almost religious support by a strange alliance of investors, patient groups, politicians, and journalists.  Spending half a million per patient per year on a drug that may turn out to be ineffective and have side effects would neither protect patients nor help the cause of RNA Therapeutics.  If patients are so keen on the drug, and I'm all in favor of patient choice, then why not provide them with eteplirsen based on compassionate use?

Last but not least, why newly Cambridge, Mass-based Sarepta had to issue the press release on the Minutes of their FDA meetings just minutes after the Boston marathon bomb blasts, is a mystery to me.

Disclosure: no position in Sarepta.

Register for the GTC RNAi Research and Therapeutics meeting in San Francisco today (June 20-21).  Get a free RNAi Therapeutics blog T-shirt and 20% discount on registration by entering discount code 'RNABLG13'.