Small Molecules and Gene Therapy Muscle Out Oligonucleotide Therapeutics

Over the last two days, breathtaking data were reported for the treatment of two severe, inherited muscle-wasting diseases affecting children.  The investigative agents were a small molecule splicing modulator and a gene therapy both of which appear to achieve superior results compared to approved Oligonucleotide Therapeutics agents. 

The developments highlight the risk that while gain-of-function changes (here by splice modulation) may have proved to be low-hanging fruits for Oligonucleotide Therapeutics, they, unlike gene knockdown approaches, face increased challenges from other technology platforms.

Small molecule splicing modulation for Spinal Muscular Atrophy (SMA)

In 2011, Roche started collaborating with PTC Therapeutics on small molecule splice modulators for the treatment of SMA.  The idea is to screen small molecules for their ability to bring about changes in RNA processing that would hopefully be gene-specific enough so as not to cause widespread off-targeting.

I had always considered this to be a monumental, if not insurmountable task.  This is because a given splicing event brings together a set of proteins that each in turn also function at other genes.  So surely a small molecule that may bias splicing from SMN1 to SMN2, as does antisense oligonucleotide SPINRAZA from Ionis and Biogen through highly specific base pairing, would also affect a range of other genes.

If that were not enough of a challenge, a small molecule carries the extra baggage of being more widely available across tissue types such that off-targeting is a risk to not just the CNS as with SPINRAZA, but many other cell types where there may be no benefit from SMN upregulation.

Accordingly, the first compound in the PTC-Roche collaboration to enter clinical development, RG7800, had to be discarded last year due to retinal tox concerns.  Another small molecule competitor, branaplam from Novartis, had similarly been put on hold due to tox concerns although this compound has resumed development late last year.

It was therefore amazing to see updated results from the FIREFISH study of the follow-on compound RG7916 in type I SMA infants.  They show that 90% of children had an improvement in the CHOP-INTEND measure of physical functioning after 6 months on the drug.  The results are particularly impressive considering that treatment had been initiated relatively late compared to the new standard of care with SPINRAZA and the soon-to-be-approved gene therapy by Novartis (àAvexis).   

Not only that, there had been no treatment discontinuations due to safety issues with RG7916.

Given that SPINRAZA has to be given intrathecally while RG7916 can be given orally, and given that both the gene therapy and the orally available RG7916 appear to be somewhat more efficacious than the oligonucleotide, the focus of Biogen and Ionis should now be on testing combinations of SPINRAZA with both modalities.  Ideally, there is added efficacy from using the agents together either because due to higher achievable SMN protein levels and/or due to complementary biodistribution (note: the value of SMN increases outside motor neurons is debated).  If not, the SPINRAZA franchise may have a limited shelf-life.

Fake-it-‘til-you-make-it Sarepta with gene therapy breakthrough

The other piece of great news for families dealing with neuromuscular disease came yesterday at the Sarepta Therapeutics R&D Day. 

To wit, Sarepta had used dubious data and a lot of political lobbying to get the controversial exon skipper eteplirsen approved under accelerated approval.  While delaying the confirmatory study that is supposed to be part-and-parcel of an accelerated approval, Sarepta has been raking in billions in sales and added market capitalization.  This has allowed the company to build a veritable DMD powerhouse with a number of candidates that look much more promising than ordinary PMO-based eteplirsen.  They include peptide-conjugated PMOs and especially gene therapies.

If you are involved in drug development, better get used to the dubious morals of the industry.  If things go well, you behave like the paragon of virtue, if things don’t go so well you fake it until you get another chance at succeeding.  I digress…

Before the initial gene therapy data were to be presented by Jerry Mendell from Children’s Nationwide of SMA fame, I had dreaded the thought of having a hyped-up R&D Day being about divining the meaning of a biopsy slide or two on the barely-above-background expression of the microdystrophin transgene.

However, what was presented was anything but borderline.  Unlike with eteplirsen where we were dealing with debatable 1%-type absolute expression levels, there was robust microdystrophin expression: ~75% of cells expressed the transgene (by IF) with roughly 30% absolute expression of microdystrophin relative dystrophin from a normal person (by Western blot).

Not only this, the microdystrophin was functional at the molecular level as judged by restoring dystrophin-related protein complexes serving to protect the muscle from damage by acting as shock absorbers.  Accordingly, CK levels in the blood, a marker of muscle damage and elevated in children with DMD, were robustly (9x) and uniformly lowered in all 4 boys between the ages of 4 and 7.  Add to this the obligatory before-and-after videos and there is little doubt already at this relatively early stage already (~1-3 months after gene transfer) that AAVrh74.MHCK7.microdystrophin is a powerful agent applicable to essentially all types of DMD.

On the safety side, there were considerable, but transient and manageable increases in liver enzymes.  This was to be expected, however, considering the very high doses of AAVRh74 needed to achieve widespread transgene expression in muscles throughout the body and treating physicians know to look for it.

If the safety holds up and expression continues to be long-lived, AAVrh74.MHCK7.microdystrophin could render many exon-specific oligonucleotide splice modulators obsolete. The duration of action is the most concerning issue to me at this point given the attendant cell turnover and attendant risk of losing episomal gene therapies in patients with muscle damage.    

Arrowhead Cuts Trial Short And The Market Loves It

Yesterday, Arrowhead Pharmaceuticals announced that it is cutting short its phase I study of ARO-AAT as it had observed what appeared to be maximal drug activities at lower than predicted doses.  ARO-AAT is the GalNAc-targeted investigational RNAi agent for the treatment of alpha-1-antitrypsin-related liver disease.  This just 3 months after it had initiated dosing in the (healthy volunteer) study.  

The stock market greeted the announcement, sending shares up another 24% turning it into a 10-bagger in the span of one year!

Interpretation of AAT program status

This is what was said to describe the general results seen so far:

”We have escalated above a dose that we believe achieves maximal activity, and all doses to date appear to be generally well-tolerated.”

And this was the trial design:

AROAAT1001 (NCT03362242) is a Phase 1 single- and multiple-ascending dose study to evaluate the safety, tolerability, pharmacokinetics, and effect of ARO-AAT on serum alpha-1 antitrypsin levels in healthy adult volunteers. The study includes 7 cohorts in which 16 subjects receive placebo and 28 subjects receive single or multiple doses of ARO-AAT at doses of 35, 100, 200, or 300 mg. Additional cohorts were planned at a dose of 400 mg, but were deemed unnecessary based on observed activity at lower doses.

The company said that data from the study would be presented at AASLD late this fall.  However, based on the original study design involving single and multiple dose cohorts of 35, 100, 200, 300, and 400mg, we can roughly infer the following from the press statements: single and multiple dose cohorts of the first 3 doses up to 200mg had been completed (= 6 cohorts) and the 300mg single-dose cohort had started when they determined that 300mg is already above where maximal activity is reached.  Although the statement does not exclude the possibility that maximal activity appeared to have been seen already at 100mg, we can conservatively 200mg (~3mg/kg) as the dose with which the company plans to move into patients.

In monkeys, the company reported that AAT expression in the liver was almost completely abolished following 2 doses of 3mg/kg spaced 28 days apart based on last year's AASLD poster.

Moving pipeline and technology speedily along

This fast-tracking of the program is consistent with Arrowhead’s overall rapid-fire execution in bringing new programs into the clinic and then use modern, parallel-cohort trial designs and disease experts to move them rapidly along in the clinical development bowel. 

Its goal- and there is every indication they will make good on it- is to bring around half a dozen clinical candidates into the clinic in just over year.  If that weren’t challenging enough for a small biotech company like Arrowhead, these programs are not limited to being targeted to a single tissue (the liver), but could also involve the lung and cancer.

In a way, the liver pipeline is being moved along in a similar fashion to when Alnylam selected its current late-stage clinical crop of candidates 5 or so years ago in a cook-off between different target genes.  Once the technology is ready for a certain tissue, you focus on populating the pipeline with related candidates.  This allows you to grow your market cap and become an established biotech player.  With a market cap now well above $1B, Arrowhead certainly is close to reaching that dream.  It only needs to add to balance sheet cushion.

Overpromising

On the other hand, claimed breakthroughs not in just one (lung), but two (à cancer) additional tissue types, make me wonder whether Arrowhead has started to overpromise again in a fashion similar to the DPC experience when it and investors chose to whistle over safety signals ('DPC has no side effects').  It is understandable if RNAi companies dangle the carrot of additional tissue types in front of the investor community- investors need dreams to hang their hats on- but the speed in light of the size of the company make me wonder if they are taking a short-cut or two too many and investors choose to ignore red flags again.

How times have changed: a year ago, still overly punished by the markets for the DPC fiasco, an cutting-a-trial-short announcement like yesterday’s would have sent shares down further.  Now, Arrowhead is getting the benefit of the doubt.

Knowing that a statement like ‘all doses to date appear to be generally well-tolerated’ can mean pretty much anything bar a drug-related patient death, I choose to remain neutral with regard to the news and will await the AASLD data until further judgement on the quality of Arrowhead’s approach.

Alnylam's Primary Hyperoxaluria RNAi Drug Clears Patient Hurdle

Alnylam today provided an update on its clinical program of Lumasiran for primary hyperoxaluria (PH), an ultra-rare disease of organ damage and failure due to elevated levels oxalate (here for OxalEurope presentation slides).

The new data include safety and efficacy from all three studied dose cohorts (1 and 3mg/kg monthly, 3mg/kg quarterly) and involved early-stage PH type 1 (PH1) patients without end-stage renal disease and systemic oxalosis.  Accordingly, RNAi knockdown of the target glycolate oxidase (GO) enzyme reduced (urinary) oxalate by ~2/3.

Importantly, this was uniformly achieved across all dosing cohorts, perhaps with somewhat less inter-patient variability for the 3mg/kg cohorts.  Accordingly, oxalate levels were reduced in all patients below a threshold where natural history studies point to an almost complete protection from progression to ESRD. 

Of note, despite the caution that RNAi competitor Dicerna Pharmaceuticals had voiced that GO as a target might be prone to a rebound effect- based on the preclinical observations that oxalate reductions only kick in after profound, ~85% GO gene knockdown is achieved- the new data provide no evidence for this.  Instead, quarterly dosing with 3mg/kg was as good as with monthly dosing of the same dose, consistent with the sustained target-gene knockdown profiles seen with most current GalNAc-RNAi candidates.

The safety profile was unremarkable for a disease like this.  The only obvious drug-related side effect were transient injection site reactions in 2 out of the 9 patients receiving the investigational agent.

With these data in hand, Alnylam is about to embark on a registrational study of Lumasiran in ~25 PH1 patients.  This study will further include younger patients and those with more advanced disease than in the present study.  Given the similarity of the oxalate knockdowns across the cohorts, it is difficult to predict what the dose and dosing schedule will be, but testing both 1mg/kg and 3mg/kg quarterly could be a good idea.

Dicerna competition

Lumasiran is not the only gene knockdown agent in the clinic.  Its closest competitor, ~1.5 years behind based on the May 30 announcement that the 1^st PH patient has now been dosed, is Dicerna Pharmaceuticals’ DCR-PHXC.  DCR-PHXC is differentiated from Lumasiran in that it targets LDHA instead of GO1 and is expected to have a more linear target gene knockdown-oxalate knockdown relationship.  

It is also applicable beyond the PH1 patient population, i.e. also for the less severe (and much less commonly diagnosed) PH2 and PH3 forms of primary hyperoxaluria.

Lumasiran has thus now set a high bar for the commercially currently most valuable patient population.  It will be thus be important for Dicerna to improve diagnosis rates of PH2 and 3 as it intends to embark on a pivotal study in 2019.