Oligonucleotide Strategies beyond the Liver

As the first major wave of pivotal trial successes involving gene knockdown in the liver has reached the shores, oligonucleotide therapeutics are quickly establishing themselves as the dominant modality in this important target organ for new drug development.  While the industry is milking this organ for therapeutic applications, my thoughts are directed towards the next tissue opportunities that should further feed this revolution in drug development we are witnessing now.

After not having been part of the conference circus for 2 years, this week’s Oligonucleotide Therapeutics Society annual meeting in Bordeaux, France, was just what the doctor ordered for me to get a clear perspective on this issue.  Importantly, lessons learned from GalNAc-targeted oligonucleotide delivery to hepatocytes, but also LNP delivery to the liver prior to this, now allow the field to take the next step in extending delivery beyond the liver.

Chemical stability

All these efforts essentially share their use of highly modified oligonucleotides which have particularly changed the philosophy around the RNAi modality.  This allows the oligonucleotide to not only reach the target tissue intact, but also to remain trapped in endosomal compartments which serve as slow-release depots for long duration of action.

Chemical modification also has greatly reduced the immunogenicity of RNAi triggers and obviated the need for protective nanoparticle formulations.  These often came with the added liability of amplifying the immunostimulatory potential of these molecules.  Consequently, decade-old approaches are now being revisited with more fully modified RNAi trigger versions (e.g. self-delivering RNAi trigger structures as pursued by RXi Pharmaceuticals and the Khvorova group at UMass).

Ironically, after all the song and dance by RNAi bellwether Alnylam about the utility of exotic modifications in their conference presentations (one can also call it willful misleading of the field- not really the purpose of scientific conferences), the strong trend is towards maximizing 2’-O-methyl content, in addition to some 2’-F and phosphorothioation at the RNAi trigger termini.

PK enhancers

One of the reasons why the liver became the first major oligonucleotide target organ is that it is readily accessible from the blood.  This allows it to soak up oligonucleotides before they get removed by renal filtration.  In an effort to fight this tendency, the use of PK enhancers, in particular lipophilic groups is frequently seen.  Ionis and Alnylam are testing these for example for getting better distribution to the muscle and potentially also better functional uptake.

Receptors

PK enhancers, however, are largely about shifting around biodistribution, but it is hepatocyte ASGPR-type receptors that are the most valuable assets the industry is striving to identify.  My highlight of the conference therefore was a talk on a collaboration by AstraZeneca and Ionis demonstrating strikingly selective and effective targeting of beta cells in pancreatic islets. Think diabetes! 

It is the GLP1-receptor that does the magic here and which can be targeted by GLP1-peptides for effective oligonucleotide uptake.  While the in vivo validation was limited to rodent models, including an elegant GLP1-receptor knockout mouse model, I am convinced that the findings will translate to larger animals and humans.  It is one of those things you just know when seeing such data.

This example illustrates the value of knowing your target cell type really well, as this may allow you to identify additional ASGPR-type receptors which had been thought of elusive.  But even if they are lacking in some tissues, a nice, yet simple strategy to overcome this was illustrated by MPEG LA and Axolabs: by linking more than one RNAi trigger to a small scaffold, they were able to show that cellular oligonucleotide uptake capacity can be increased beyond the limits of receptor amount on the cell surface.

 

While the liver certainly does not need this strategy, it should definitely be applied to new targets like the beta cells.  Let free market competition do its magic and have oligonucleotide therapeutics solve diabetes now that you can effectively reach hepatocytes, adipocytes, and now also beta cells. Yes, I love the free markets, but I digress… 

After beta cells, it was a collaboration between Alnylam and Johnson & Johnson on overcoming the long-held dream of oligonucleotide therapeutics addressing gene regulation in cancer cells.  Here, small and stable ~2nm peptide scaffolds referred to as centyrins were coupled to the RNAi trigger and directed towards different receptors like PSMA and EGFR.  Perhaps the most striking aspect of centyrin-siRNA conjugates was their effective tumor penetration where prior RNAi delivery attempts like LNPs had fallen short.

Endosomal release

Sometimes getting to the endosomes alone is not enough when the rate of cytosolic release therefrom is insufficient.  So despite of the DPC fiasco last year and despite of aborted arginin-based endosomal release attempts prior to this, active endosomal release is still embraced in some delivery efforts.  Most notably, Sarepta has shown dramatic increases in dystrophin exon skipping in non-human primates with new peptide-PMOs (PPMOs) compared to their unconjugated parent molecules.

Of course, everybody now wants to know what the therapeutic window really is.  While the Sarepta representative at OTS was a bit cagey when asked about it, Sarepta’s CEO noted in a recent investor presentation that the filing of an IND by the end of the year would be a major positive signal in that regard.

Finally, all of the above developments are aided by more general progress in technologies interrogating biology such as single cell technologies (cell type isolation from complex tissues like the kidney), reduced chemistry costs allowing for much larger numbers of oligonucleotides to be screened, and ubiquitous low-cost and high-throughput sequencing.

 

Sometimes I pinch myself asking whether all this is real and not just a figment of my imagination, but at least in my mind the stars just keep aligning allowing for RNAi and oligonucleotide therapeutics to take the next step up the value ladder.

RNAi Therapeutics Become Real

It’s been a long ride.

The lives of quite a few of us has been consumed now for close to two decades dreaming about a world in which RNAi Therapeutics have real-world clinical impact.  Yesterday, that dream has officially materialized.

TTR knockdown improves disease state

With the conclusion of the most comprehensive clinical study conducted to date in the severe orphan disease ATTR amyloidosis, Alnylam and investigators have found that prolonged knockdown of the causative transthyretin (TTR) gene improved both objective (mNIS+7) and subjective (quality of life) measures of the main manifestation of the disease, peripheral neuropathy.

There is no doubt that this was the result of on-target TTR gene knockdown since this was a strictly controlled study in which the only difference to the placebo group was the administration of the RNAi formulation.  Furthermore, results from a similar study for this indication (NEURO-TTR) by Ionis earlier this year in which related RNaseH antisense oligonucleotide technology was utilized for TTR knockdown also demonstrated disease benefit, albeit a ‘mere’ halt of neuropathy progression in QoL as opposed to the improvement seen here (note: mNIS+7 was reported only with regard to placebo, not versus baseline).

Importantly, this extra benefit is likely explained by the fact  that the Patisiran RNAi formulation had been shown to be slightly more potent than antisense drug candidate Inotersen in earlier-stage studies (here and here).  While considering a ~80-85% vs ~70-75% knockdown may not seem much at first glance, protein deposition, clearance and in particular misfolding that is at the root of the disease are higher-order concentration-dependent processes (think about crystallization) so a difference of ~17.5% vs ~27.5% (>50% more remaining than with Patisiran) remaining insulting protein may well be highly meaningful. 

Of note, phase II data from Patisiran have demonstrated that the ~80-85% knockdown was able to somewhat tilt the tables in favor of TTR tissue clearance, but also that it was not able to fully do so.

It is therefore of utmost importance to push forward with the development of the even more potent GalNAc-enabled RNAi candidate ALN-TTRsc02.  Preliminary phase I data from that candidate suggest that it should be able to reduce TTR levels to 5% or less with subcutaneous dosing as infrequently as every 3-6 months (Patisiran: intravenous every 3 weeks; Inotersen: weekly).

Why this is such a big deal for RNAi Therapeutics

As I had indicated in last week’s post, the phase III APOLLO results represented a make-or-break moment for RNAi Therapeutics.  If it turned out that RNAi- and delivery-related side effects were to outweigh the benefits of gene knockdown at the conclusion of this decade-long, high-visibility program, the financial markets would have reacted violently and starved the industry of the cash necessary to more fully develop the technology (Alnylam with its cash reserves at least would have mounted a comeback eventually).

If it’s one thing I’ve learned about biotech money matters over the last 15 years of riding the RNAi rollercoaster, it is that the perception of a technology is as important for its continued development as its intrinsic technological validity.

Personally, I am most relieved in that the small, but often pervasive transcriptomic off-target changes introduced by an RNAi trigger (à microRNA-type off-targeting) did not have an apparent detrimental effect on the target organ (here: liver) after prolonged, 18-months treatment.  This should not be entirely surprising based on what we’ve learned about microRNA biology- not all targets of a microRNA are biologically relevant- but is still a great relief to see play out in practice.

To wit, the RNAi industry has been overly focused on potency when selecting RNAi trigger sequences for clinical development and mostly relied on bioinformatics for specificity. Indeed, Alnylam has paid the price for this negligence in that certain RNAi triggers, such as the one for its original alpha-1-antitrypsin program, have caused liver tox most likely due to microRNA-type off-targeting.  It is therefore finally employing chemical strategies such as the incorporation of modified nucleotides in the seed region of the trigger to bias the RNAi apparatus towards RNAi cleavage instead of microRNA-type message destabilization. 

This strategy was first reported by Rosetta Inpharmatics (now Merck) and had then been developed further in the commercial realm by Marina Biotech which then licensed the IP to Roche (then acquired by Arrowhead) and possibly others. 

It will be some time until we have complete certainty that microRNA-type off-targeting won’t rear its ugly head again, but the odds should be getting better and better with the employment of best practices. More generally, Patisiran is only the beginning of a long string of real-world impactful RNAi Therapeutics.

The Wait is Finally Over

For some time now, I’ve been fascinated by how the RNAi Therapeutics sentiment cycle has been progressing in 3-year intervals.  Before I left off for my baby-break, I referred to 2014-17 as The Wait. The Wait was a reference to the ‘market’ having been convinced by early clinical data that RNAi can bring about deep and sustained gene knockdowns in WoMan, but that the ultimate clinical utility and subsequent commercial value remained to be proven.

We are now on the cusp of finding out with results from the registrational phase III study of the TTR-lowering Patisiran in familial amyloidotic polyneuropathy (FAP) imminent.

The wind is clearly in RNAi’s back with supporting developments in the related oligonucleotide therapeutics and, indeed, drug regulatory spaces.  In particular, the successful commercial launches of splice modulators EXONDYS51 (for Duchenne muscular dystrophy) and SPINRAZA (for spinal muscular atrophy) and promising phase III data from the triglyceride-lowering Volanesorsen (gene target: ApoCIII) should have removed the long-held doubt that oligonucleotides can be clinical practice-changing and profitable drugs.  And boy, are EXONDYS51 and SPINRAZA changing the clinical practice of two devastating, previously orphan diseases!

Equally important, drug regulation is strongly moving infavor of patient choice and access.  As a result, we are going to see many more drug candidates benefiting from accelerated approval pathways similar to EXONDYS51.  A number of RNAi Therapeutics programs such as Givosiran by Alnylam (àmetabolite ALA lowering as reasonably predictive endpoint) and DCR-PHXC by Dicerna (àoxalate lowering) are poised to benefit from this development.

But let’s not get ahead of ourselves and await the Patisiran APOLLO study outcome for which sponsor Alnylam has guided a ‘late September/summer’ read-out.  Drug development is full of surprises, so despite all the data* pointing towards a positive outcome, a number of variables (e.g. steroid use, lipid nanoparticle delivery) remain unknown in how they impact overall outcomes.  I for one will breathe a big sigh of relief should Patisiran be able to overcome this Make-or-Break event for RNAi Therapeutics.

 *

à positive open-label extension data from phase II Patisiran;

à positive phase III data from IONS-TTRRx targeting same gene and indication;

à the high roll-over rate into the open-label extension) indicating a positive outcome 

What the thrombocytopenia findings mean for Ionis Pharmaceuticals

Yesterday, Ionis Pharmaceuticals disclosed that severe reductions in platelets had been observed in phase III clinical trials of both IONS-TTRRx for the treatment of TTR amyloidosis and IONS-ApoCIIIRx for conditions related to highly elevated triglycerides.  Severe platelet reductions are dangerous since it can lead to occult, uncontrolled bleeding and poor blood clotting following injury.

Since the conference call was a PR disaster as the CEO of Ionis has major issues with speaking his scientific mind, and since competitor Alnylam has seemingly become the original source and interpreter of the Ionis thrombocytopenia issues (one wonders how they come into possession of these Ionis trade secrets...), I thought it may be useful to briefly come out of blogging hibernation and lay out my thoughts about what these events mean for the technology and the company.

Thrombocytopenia likely limited to systemically administered, unconjugated PS-oligos >200mg per injection

As a hematological abnormality that has historically been observed with phosphorothioate (a ‘sticky’ chemistry) oligonucleotides when given at high doses (>200mg/injection)  I’ve always considered it likely that such thrombocytopenia will be associated with measures of plasma exposure of the oligonucleotides.  Notable examples for thrombocytopenia with phosphorothioate oligos include the DMD exon skipper drisapersen by Biomarin/Prosensa (6mg per kg per week, i.e. around 300mg/week for 50kg boy) and telomerase inhibitor imetelstat by Geron (~10mg per kg per week, i.e. around 700mg/week for average adult).  Actually, isn’t it ironic, or maybe even curious that imetelstat is being developed for conditions where elevated thrombocytes is the problem (see related blog entry)???

Consistent with this notion, there was a study by Flierl et al. in 2015 that looked at the mechanism of platelet activation which may lead to platelet consumption and explain lowered thrombocyte counts.  Without going into the details of the mechanistic aspects of the study, the authors find a strong correlation with peak plasma exposure (c max) of the oligonucleotides and platelet activation.   

So why hasn’t Ionis seen severe cases of thrombocytopenia in the past (excluding use of PS-oligos in cancer patients which frequently suffer from potentially confounding bone marrow suppressions from other drugs)?  The most probable explanation is a) these events are quite rare events and b) that their experience with PS-ASOs at 300mg/week and above has been limited.  At 300mg and especially 400mg per week, safety has always looked a bit dicey such that the 300mg per week dose e.g. for TTRRx was only adopted after 200mg per week was not competitive with the knockdown results produced by ALN-TTR02 from Alnylam. 

Similarly, the initial studies with ApoCIIIRx did not include the 300mg per week dose and was adopted in favor of the very impressive triglyceride reductions seen at doses higher than 200mg.  Usually the dose escalation of the prototypical Ionis phase I studies involved 50, 100, 200, then 400mg per week with 400mg per week never being chosen for the phase II and/or pivotal studies.

What I find highly interesting is that the pharmacokinetics data from the healthy volunteer study of ApoCIIIRx reported by Graham and colleagues in 2013 (see only Table IV) reported a non-linear, 4.5x increase in cmax when doubling the dose from 200mg to 400mg per week.  This could mean that at doses of 200mg per week and higher, the risk of severe thrombocytopenia is dramatically elevated by going past the threshold where platelets become critically activated (à clotting cascade).  

If the cmax theory holds true, then the following should be the impact of the new findings on the Ionis platform.  The summary takes into account the clinical observations by Ionis that the platelet reductions are reversible upon stopping dosing and can be prevented and also treated by steroid use (just as ALN-TTR02 involves steroid use):

1)      Unconjugated, systemically administered antisense at 300mg per week and above (incl. phase III assets TTRRx and ApoCIIIRx): need for tight platelet monitoring.  May involve temporary halt of studies to amend protocols.  Commercially, need for tight platelet monitoring could be a problem for less severe diseases due to convenience and competitive issues.  

Note that for all the liver-targeted programs, backup GalNAc-conjugated versions are in development which should not suffer from thrombocytopenia (see below).  However, systemic programs targeting other tissues such as DMPKRx for myotonic dystrophy will have to be under continued scrutiny depending on the dose.

2)      Unconjugated, systemically administered antisense at 200mg and below per week and below: little impact.  Start collecting data more systematically to learn more about platelet interactions, otherwise no big impact.

3)      GalNAc-conjugated antisense: no impact. Essentially all the Ionis pipeline, including ApoCIIIRx, has been re-engineered for some time now to be GalNAc-conjugates.  This is because of their 10-100 fold increased potency over the unconjugated versions thus decreasing the doses to well below those expected to cause severe thrombocytopenia.  Even at the same doses, plasma exposures will be much reduced due to the rapid clearance into the hepatic compartment as demonstrated by Shemesh et al in one of the most recent publications by Ionis.  No thrombocytopenia events to my knowledge were seen with RG-101 (for HCV) by Ionis' 'satellite company' Regulus Therapeutics, where a up to 8 mg/kg of GalNAc-conjugated phosphorothioate oligonucleotide has been administered.

4)      CNS programs: no impact. Peak plasma exposures are insignificant for intathecally administered oligonucleotides as used in Ionis’ CNS franchise, a franchise which includes exciting drug candidates such as phase III asset nusinersen for the treatment of spinal muscular atrophy (SMA) and candidates for other severe neurodegenerative diseases.

In summary, the only programs which could be significantly impacted by the thrombocytopenia findings are the programs that target tissues outside the liver and which involve systemic administration.  The liver franchise remains intact especially with the new GalNAc versions although there could be some minor delays and increased competitive impact in those diseases that Alnylam is free to go after according to the Ionis-Alnylam IP agreements.  The important CNS franchise remains fully intact. 

Disclosure: long Ionis and doubled down yesterday.

Cholesterol-lowering RNAi Therapeutic Shocks Monoclonal Antibody Establishment

Not too long ago, RNAi Therapeutic got dumped by Big Pharma in a big way not least because of the monoclonal/recombinant protein pedigree and corresponding bias among the top dogs of these organizations.  This was most obvious with Roche and Merck when changes in their overall R&D organizations led to the loss of their last internal RNAi champions.

How times have changed. Yesterday, The Medicines Company (who have now become a natural acquisition target) and Alnylam presented data (press release here, data here) strongly suggesting that an RNAi Therapeutic will push aside the temporally more advanced monoclonal antibody competition to become the best-in-class agent in the potentially top pharmaceutical category over probably the next two decades: the inhibition of PCSK9 for the prevention of cardiovascular morbidity and mortality.

The data in support of this claim were presented yesterday at the 2015 ESC congress in London that in retrospect was apparently named in honor of the delivery technology underlying ALN-PCSsc, a so-called Enhanced Stabilization Chemistry-based RNAi conjugate.

Treatment adherence

Importantly, the single-dose part of the study showed that starting with a dose of 300mg of ALN-PCS, PCSK9 levels were flat-lined to ~25% of normal levels for at least 4-5 months and haven’t started to perk up yet by the data cut-off date for this presentation.  It is to be expected that the knockdown will be even more pronounced with repeat dosing as supported by the initial repeat-administration data (2/3 doses) showing mean PCSK9 reductions to ~15% of normal.

Since in the PCSK9 category, it is PCSK9 that is driving LDLc lowering, the ultimate aim of this therapeutic approach, similar kinetics were seen in terms of LDLc levels in the blood with reductions (and safety/tolerability profiles) comparable to that seen with the recently approved monoclonal antibodies PRALUENT (by Regeneron/Sanofi) and REPATHA (by Amgen), ~55-60%.

In the case of the monoclonal antibodies, dose administrations every two weeks is really what it takes to consistently suppress PCSK9/LDLc because their inhibitory ability is directly correlated to their amount in the blood which declines rather precipitously after drug administration.  In the case of RNAi, however, you only need minute amounts to clamp down gene expression and at least for the liver, it appears that quarterly/semi-annual dosing schedules are realistic (it also depends on target and how much it needs to be repressed; e.g. with CC5 you may need much more target gene knockdown than 58-90%).

Sticking a needle into you just 2 or 4 times a year instead of 26 times, of course, has great advantages when it comes to treatment adherence. Keeping patients on drugs is a major issue for such life-long therapies especially since the disease is not felt acutely.  This point was made repeatedly by cardiovascular disease thought leader Dr. Kastelein on the companies’conference call.  By being able to co-ordinate drug administration with routine doctor visits, it would be possible to achieve very high compliance rates thereby preventing intermittent LDL cholesterol spikes that are believed to be particularly harmful.

In other words, assuming cardiovascular outcomes to be almost entirely driven by LDLc lowering, ALN-PCSsc would/should be best-in-class in the PCSK9 category.  There are numerous examples such as Eylea in the wet AMD space where injection frequency is the main competitive driving force among competing agents (here VEGF inhibitors) that exemplify how being a best-in-class follower can be very profitable.  Let the monoclonals build the PCSK9 market for ALN-PCSsc to then take it.

Outcomes

Last but not least, the ultimate value from being different will come from the results of the cardiovascular outcomes (and actually overall survival) studies that will really unleash the wide adoption of the PCSK9 class.  Due to their similarities, there is every reason to believe that the results from the monoclonal antibodies will cluster tightly.  By contrast, for better or worse, the outcome studies from ALN-PCSsc should be notably different and given that an RNAi agent mimics the compelling human genetics behind the PCSK9 story (extreme LDLc lowering in PCSK9-mutant individuals without other apparent untowards effects such as elevated liver triglycerides etc) much more closely, I like my chances here.  

We all know about the intricate feedback mechanisms of lipid biology so that binding a player merely in the serum as the monoclonals do as opposed to removing it from both inside and outside the cell could have unanticipated consequences.  Albeit early, the preliminary data from ALN-PCSsc support that in that the percent LDLc knockdown is the same whether in the presence or absence of high-dose statins whereas that of the monoclonal antibodies becomes muted.

Having said that, expect the monoclonal antibody establishment to play the 'RNAi is different from monoclonal antibody card' lest ALN-PCS piggy-backs on the MAb CVOT results expected to come out starting in 2017.

Back to my self-imposed exile, but I couldn't resist on commenting on what could be a perfect Oligonucleotide Therapeutics storm that is building. Next up is (maybe) ARC-520 for HBV.  And yes, I'm long MDCO as if that's not obvious.

The RNAi Therapeutics Blog is Taking a Break

When it feels that everything has been said, it may be time to be silent for a while.

This point has come for me and the RNAi Therapeutics blog and I look forward to take part in the conversation with renewed energy and ideas.  Until then, you can follow the 'light' version of this blog on Twitter @RNAiAnalyst. 

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.