There is No Doubt: Splice Modulator Drug for Spinal Muscular Atrophy Works

The fairy-tale story of the splice modulation for spinal muscular atrophy (SMA) continues.  This morning, Isis Pharmaceuticals provided an update on the phase II study of ISIS-SMNRx in type I SMA infants.  The data built on already highly promising data as of last September, showing that a doubling (~9 to ~18 months) of the median ‘event-free survival’ compared to the Natural History has now been reached with numbers still increasing as more than half the infants remain event-free.

Only one out of 12 infants still on study suffered an event (permanent ventilation) over the last 9 months.  This one in 108 month event rate compares to 6 events in ~200 months in the prior phase of the study, suggesting that if babies can be diagnosed and treated early enough so that they are covered during a critical period of development (e.g. maturation of the neuromuscular synapse) chances are that they will enjoy a very significant treatment benefit from ISIS-SMNRx.

This is thus consistent with the biomarker data showing that ISIS-SMNRx increases the missing functional full-length SMN protein by 2-3 times essentially turning a type I SMA infant (usually 2 copies of SMN2) with an 80% chance of dying or going on permanent ventilation by 18 months into a much milder form of the disease where patients have 4 or more copies of SMN2 and have an almost normal life expectancy (note: those with 3 copies, usually type II SMA, live into teens/early adulthood). 

While as a parent, I would almost do anything for my child to get access to the drug and I do understand there to be calls for immediate (à once diagnosed, the window of treatment opportunity may be quite narrow) regulatory action, the first consequence of today’s data should be getting SMA on mandatory genetic panels for newborn screening.  Only then will there be maximal benefit once the ongoing blinded phase III study reads out in late 2016/early 2017.

Antisense Technology Is Feasable for Neurodegenerative Drug Development

Last week at World Muscle, ISIS Pharmaceuticals provided an update on their phase II study results for ISIS-SMNRx for the infant- and child-onset forms of spinal muscular atrophy (SMA).  Importantly, biomarker and biodistribution results were reported that clearly showed that ISIS-SMNRx is doing exactly at what it was designed to do, namely meaningfully increase target gene expression in the central nervous system (CNS).

The data not only greatly de-risk ISIS-SMNRx, but open up phosphorothioate-based antisense technology for a whole range of other, largely severe CNS-based diseases of high unmet need, including Huntington's disease, the spinal cerebellar ataxias, Alzheimer's, Parkinston's- you name it!

Biodistribution

Specifically, the data showed that despite only a focal, intrathecal infusion of the antisense drug into the lower spine, it readily distributed throughout the CNS up to the brain and at concentrations (10-30ug per gram tissue) that are strongly predicted to support both steric blocking and RNaseH antisense mechanism of actions in the CNS for 2' MOE chemistry.  Further chemistry improvements such as cET are opening the therapeutic window even more so.  What is more, these concentrations were maintained for months, thus further supporting the apparent therapeutic benefits seen in these open-label studies.

Note that the effective concentration for antisense mechanisms will differ according to target tissues; e.g. in the liver, largely due to competition from phagocytic Kupffer cells, the effective concentrations are 100ug/g and above with 2' MOE chemistry.

With the generous support of SMA families, the company was also able to look for the drug and the SMN protein in tissue sections from 3 deceased infants.  These investigations showed that the phosphorothioate oligo had been taken up pretty much in every neuronal and non-neuronal cell types. 

Such broad-based uptake may be quite important according to the opening keynote address of ISIS collaborator Don Cleveland last night at the annual OTS meeting in San Diego, given that expressions of disease-causing genes in various cell types, not just the neurons, seem to contribute to most neurodegenerative diseases.


Biomarker

In terms of drug action, the SMN protein was found to be re-expressed in the corresponding cells as intended for the splice-modulating approach of ISIS-SMNRx.  This was shown by immunofluorescent analysis.  Moreover, quantitative PCR showed that the expression of the intended full-length SMN2 mRNA was increased by 2 to 3-fold, consistent with the 2 to 3-fold increases in SMN2 proteins found from cerebrospinal fluid (CSF) samples in the child-onset studies.

No dose-limiting safety issues were seen and the intrathecal infusions which are predicted to be needed on a ~6 month-basis for many of the anticipated CNS-related antisense applications could be performed without having to resort to general anesthesia.

For SMA, genetically speaking all this essentially turns a type I infant-onset SMA baby into a less severe type II/III child, and a type II/III SMA child into a normal one, with the caveat that this benefit obviously only accrues from the time the drug is given which, unfortunately, may be too late for many type I SMA babies.  I was e.g. somewhat disappointed that no apparent correlation was seen between onset of antisense administration and therapeutic outcomes in the infant study, although clearly the numbers may well have been too small (n=20). I am very hopeful, however, that those infants making it out to say 18 months and beyond with ISIS-SMNRx may see very good outcomes indeed.

Despite the caution, all this was accompanied by apparent therapeutic benefits in terms of survival and muscle strength.  While highly intriguing, due to the open-label nature of the studies and the small patient numbers, it is not my intention to delve more into that aspect of the data and instead focus today on the truly mind-blowing pharmacodynamic data.  These should provide hope for many patients and families with neurodegenerative diseases. If not, we might as well give up on rational drug development.

 

Black Box Therapeutics Finds Small Molecule Cure for Spinal Muscular Atrophy

After curing Cystic Fibrosis and Duchenne Muscular Dystrophy with nonsense drug ataluren, PTC Therapeutics is moving on to relieving the sufferings of many more children afflicted with severe genetic diseases.  This time it is spinal muscular atrophy (SMA).  In a publication that recently appeared in SCIENCE, PTC Therapeutics has stumbled upon another orally bioavailable small molecule that is claimed to modulate an RNA processing event for therapeutic purposes.

The reason why I’m interested in the PTC story is that their small molecule approach to RNA modulation is counter-intuitive as it instinctively calls for a nucleic acid-based solution.  On the other hand, an oral alternative to what are usually more invasive routes of administration could have certain advantages such as patient convenience and access.

Naryshkin paper

The paper by Naryshkin and colleagues is on the discovery of RG7800, a compound partnered with the new LNA antisense owner Roche.  It is an illustration of the general strategy taken by PTC Therapeutics of finding RNA sequence-specific small molecule modulators of genes of interest. 

In this example, the company began the panning process by hooking up RNA elements from the SMN2 gene that are involved in the splicing of an exon of interest with a luciferase reporter gene so as to enable large-scale small molecule library screening.  If left untouched, this reporter construct will produce very little light emission due to luciferase expression.  On the other hand, if the small molecule is successful in biasing splicing towards the inclusion of exon 7, the intended outcome of this particular SMA treatment approach, then light is produced.

As one might expect there are numerous ‘hits’ that come out of such primary screens.  In this case, 2000 small molecules from the library increased luciferase expression.  Since such expression changes can be due to a myriad of causes, selected candidates were put through a number of tests such as whether the increase was dependent on the SMN2 sequence elements (reporter constructs without the SMN2 sequences would take care of this), whether the selected small molecules change the expression or splicing of other RNAs (à off-targets; e.g. by RNA seq) etc.

RG7800 apparently survived all these tests and was found to increase the desired SMN2 splice form by about 70% in a number of cell models (including patient-derived cells) and in a mouse model.  It also has entered clinical development.

Comparison with ISIS drug

Since RG7800 is in direct competition with antisense drug candidate ISIS-SMNRx by ISIS and Biogen as both aim to increase SMN2 exon inclusion, a brief comparison is warranted.

While RG7800 has the obvious advantage of being orally bioavailable versus the need for intrathecal administration of ISIS-SMNRx, in terms of molecular outcome, exon inclusion in spinal motor neurons, it appears to be lacking: a ~70% increase compared to ~100-150% increases in the good SMN2 isoform achieved by ISIS-SMNRx in two clinical trials in SMA infants and children, and even more than that in rodent studies before (Passini et al. 2011).    However, RG7800 achieves SMN2 splice modulation not just in the motorneurons of the CNS, but in many other places, in and outside the CNS. 

There is ongoing debate as to whether such body-wide modulation is required, a question also prompted by a study by ISIS and collaborators on the systemic/subQ application of the antisense drug (Hua et al. 2011).  Parenthetically, this also means that if ISIS/Biogen wanted global SMN2 regulation, they have an option with subcutaneous administration in addition to intrathecal administration.  Interestingly, with the subcutaneous administration of ISIS-SMNRx, the mice lived much longer than when the oligo was given intrathecally (days/weeks versus months).  Indeed, consistent with those studies, the PTC small molecule also prolonged the lives of SMA mice considerably.  

I guess it will have to be the clinical trials which will be most informative as to whether this is an artefact of the mouse model or not.  If so, systemic drug exposure would add no benefit and only increase the risk of adverse events from off-targeting.

At the end of the day, what I find remarkable is that it is apparently possible to find small molecules that can modulate gene expression in a fairly sequence-selective manner.  While I don’t doubt that you can change splicing with small molecules, it is the apparent specificity of a simple molecule such as RG7800 that perplexes me.  

PTC Therapeutics did not disclose how many compounds they had to sift through to find RG7800.  If they really just picked one or a few after the initial screen that yielded ~2000 hits and then got lucky, I’d be quite skeptical.  Adding to my skepticism is that drug concentrations in the CNS were reported that greatly (>10x) exceeded those in the plasma (supplementary figure S7).  I’m no small molecule guy, but for an oligo guy who has been following drug development in general, that seems very unique. The blood-brain-barrier apparently does not exist for PTC.  

As a result, and also given the controversy around the discovery of ataluren (artefact or not) and the fact that PTC Therapeutics itself can only remotely speculate on the mechanism of action of RG7800, I acknowledge the publication as interesting, but am not ready to jump on board just yet.

ISIS Wasting No Time to Find Out Clinical Utility of Spinal Muscular Atrophy Drug

Yesterday at the 2014 Annual Academy of Neurology (AAN) meeting, ISIS Pharmaceuticals provided an update (presentation here, press release here) on their splice modulation antisense drug candidate ISIS-SMNRx for the treatment of spinal muscular atrophy (SMA).  The data were largely consistent with those presented about two months ago.  

Disease progression, including deaths in the infant trial, however, dashed hopes that this drug would be on its way to Accelerated Approval.  In light of the small number of subjects (15) in this trial, these events make it difficult to determine the impact of ISIS-SMNRx.  Similarly, the open-label nature of both the infant (type I SMA) and children (type II/III SMA) studies leaves open the possibility that apparent treatment benefits seen in motor-function outcomes could have been a placebo effect.

As a result of these concerns, the company has taken the bold step to run placebo-controlled, blinded pivotal phase III trials in infants and children to ascertain the tantalizing signs of efficacy seen thus far.  Most importantly, the drug more than doubled target therapeutic SMN protein levels (based on measuring surrogate levels in the cerebrospinal fluid) thereby genetically converting type I into type II/III SMA and similarly type II/III into near-healthy status.   In addition, the motor function increases seen were not only clinically relevant, but were also dose-related and hardly, if ever seen in the natural history of the disease.

Clearly, the apparent disease progression in the infant study would seem contrary to this conclusion.  Type I SMA babies have a mere life expectancy of 10 months, whereas type II/III children are faced with only somewhat shortened life expectancies.  The reason for this discrepancy is probably that in a treatment paradigm, SMN protein levels only catch up with type II/III SMA a few months after birth, whereas in normal development the SMN gene is already expressed in the womb.  The fact, however, that in the mouse model of the disease, ISIS-SMNRx restored mice to near-healthy status when given after birth, supports the notion that the drug should be efficacious when given post-delivery.

In the case of the infant study, drug administration commenced around months 4-5.   Considering that rapid functional decline and death occur soon after this in the natural history of the disease, the more optimal use of the drug would be when given starting right after birth.  However, until whole genome sequencing at birth becomes a routine screening tool for genetic disease (it should be routine already in my opinion), such optimal use has to wait for probably another 7-10 years.  Nevertheless, since ISIS-SMNRx seems safe and was shown to successfully address the disease at its root in babies and children- which also means that it could synergize with other future treatments- regulators should consider the full potential of this drug, present and future, when making a decision as to its approvability in 2016/2017.

One thing is clear, ISIS and partner BiogenIdec want to conduct the most rigorous trials possible to do ISIS-SMNRx justice when it could have taken the present data and pressed for accelerated approval in type I SMA largely based on the increases in SMN protein.  Knowing the real benefit of a drug can only benefit SMA patients and their families, and as we have found out in the case of Sarepta and DMD, the timelines ultimately do not have to be that different.     

Oligonucleotide Therapeutics for Spinal Muscular Atrophy Impresses in Clinical Study

In addition to the continued validation of RNAi in Man, the other big winner of 2013 in the field of oligonucleotide therapeutics was single-strand phosphorothioate chemistry in the CNS.  Today, ISIS Pharmaceuticals announced clinical data from the most prominent candidate in that effort, namely ISIS-SMNRx for spinal muscular atrophy (SMA), a genetic muscle-wasting disease.  Following intrathecal administration of the splice modulating oligonucleotide, time- and dose-dependent improvements were observed not only in muscle function, but also in SMN protein production (biomarker), thus laying the foundation for an accelerated approval pathway.

In the open-label phase Ib/IIa study in ~30 children with the ‘less severe’ form of type II and III SMA, functional improvements of 1.5, 2.3, and 3.7 points on the HFMSE scale were seen at the 3mg (3 doses), 6mg (3 doses), and 9mg (2 doses) cohorts, respectively.  The changes were thus largely consistent with results from a previous similar, but single-dose phase I study where a 3.1 point increase could be observed at the 9mg dose.

Despite the generally positive news, the data raise a number of questions.  For example, optimal dosing frequency remains uncertain as there were similar functional improvements regardless of whether a single dose had been given or 2-3 doses.  This could have been due to the  long half-life of the drug and the time it takes from SMN protein production (as a  result of the splice modulation) to impacting motor neuron function.  Similarly, in the prior phase I study no positive changes in HFMSE scores were observed at the 3mg and 6mg doses whereas in the present study, improvements were reported.  Clearly, larger patient numbers are required  to settle on the optimal dose, and in fact this dose may not have been reached yet (note: a 12mg cohort has been initiated and children from the phase Ib/IIa trial are allowed to roll over to an additional dose of 12mg).

Case for accelerated approval?

Possibly foreseeing such issues due to small patient numbers, ISIS Pharmaceuticals and BiogenIdec recently developed an assay that allows them to measure SMN protein abundance in the cerebral spinal fluid (CSF).  It is the results from these measurements that provide a strong case for why ISIS-SMNRx should be made available (pending the 12mg results) before a larger phase III study will have been completed.  This is because the functional improvements were accompanied by increases in the SMN protein which also were dose-dependent with a more than doubling of SMN protein at 9mg.

In SMA, the SMN1 protein is missing due to mutations.  The therapeutic approach of ISIS-SMNRx takes advantage of the fact that humans have a pretty much identical gene to SMN1, SMN2.  The problem with SMN2, however, is that only ~10% of its precursor messenger RNAs is spliced into a functional SMN protein due to a difference in essentially just one nucleotide in exon 7.  The severity of the disease, i.e. whether somebody belongs to type I (most severe), type II, type III, or type IV of the disease depends on the copy number of SMN2 genes: 2 copies in type I, ~3 copies in type II and III, and at least 4 copies in type IV.

Therefore, doubling the protein output for type II and III patients (the patient population in the present phase Ib/IIa study) would appear to put the children into the type IV category in terms of protein output (correponding to ~6 SMN2 copies).  In contrast to type I-III, type IV results in no differences in life-expectancy and only in rare cases causes patients to be wheelchair-bound late in life.  

Severe disease of high unmet need, strong biomarker data with highly suggestive functional results, all dose-proportional…the ingredients for an accelerated approval. 

Just in: ISISis reporting preliminary data from a parallel multi-dose phase II study in the most severe, infant form of SMA (type I).  Although small in numbers, the fact that the 4 babies at the starting 6mg dose are still alive and without permanent respiratory support at an average age of 12.5 months appears to be much better than expected.  According to natural history data, you would have expected 2 babies either dead or on permanent ventilation by month 10.  

And finally...the ISIS-SMNRx results increase the value of Marina Biotech's CRN chemistry.  This chemistry competes with the ISIS 2'MOE chemistry employed in ISIS-SMNRx and appears to be of higher potency/affinity, but much less defined safety.