Santaris Terminates PCSK9 Hypercholesterolemia Trial

Antisense company Santaris seems hardly able to catch its breath these days. After being sued by ISIS Pharmaceuticals for patent infringement, reporting first clinical proof-of-concept for a MicroRNA Therapeutic, it has now been revealed that the company prematurely terminated a phase I trial of its PCSK9 phosphorothioate LNA RNase H antisense inhibitor SPC5001 for the treatment of hypercholesterolemia. This study had been initiated in May of this year with an estimated enrollment of 40 healthy and familial hypercholesterolemia subjects.

Reasons for the trial termination were not revealed. However, since this was a phase I safety trial, it is likely that the trial termination was due to some safety issue. Supporting this is that, coinciding with the PCSK9 trial termination, enrollment has been halted in another phase I hypercholesterolemia trial sponsored by Santaris, this time targeting ApoB (SPC4955).

Consequently, after the premature trial termination last year of BMS/ISIS’ phosphorothioate 2’ MOE RNase H PCSK9 antisense inhibitor (BMS-844421), Alnylam’s SNALP-delivered ALN-PCS has suddenly taken the lead in the race to develop a PCSK9-targeting RNA Therapeutic, pitting it against the monoclonal antibodies by the likes of Regeneron and Amgen.

Without explanations for the trial terminations, it is difficult to pin down the exact cause(s) of the presumed toxicities. What the compounds by Santaris and BMS/ISIS Pharmaceuticals obviously share is that they both use phosphorothioate chemistries and work via an RNaseH mechanism. Phosphorothioates are widely used in antisense development because they bind to proteins in the plasma and various tissues, thereby allowing one to achieve tissue concentrations high enough such that mass action will allow for cell penetration and target gene knockdown. In my opinion, phosphorothioate oligonucleotides are promising for a number of local applications. The problem with systemic applications, however, is that the high concentrations reached in tissues such as liver, spleen, and kidney, ultimately limit their therapeutic window. Some industry experts have referred to this issue as the ‘hazardous waste’ problem of systemic phosphorothioate oligonucleotides.

ISIS supporters will point out that Santaris’ specific LNA chemistry is the most likely culprit, whereas ISIS’ chemistries are relatively safe as supported by ISIS’ comparative studies. Of course, these studies were published to make Santaris look bad, just as the recent lawsuit was aimed at blunting Santaris’ competitive threat. Conspiracy theorists may even suspect that the sudden turmoil around Santaris is no coincidence and was orchestrated to inflict maximum damage at a time that Santaris seemed to be on a roll with a number of trial initiations, positive HCV data, and depriving RNAi Therapeutics of a deal with Pfizer (Pfizer apparently being attracted to ‘naked’ oligonucleotide approaches over RNAi nanoparticle ones). However, if Santaris intended to go public to satisfy its increasing cash needs (maybe less so with the trial terminations), it can probably scratch that now.

RNA Therapeutics is intensely competitive as they compete for pretty much the same investment dollars. Investors, including Big Pharma, are easily influenced by public opinions and fashion trends. 4 years ago it was all about the fear of being left behind in RNAi Therapeutics, with little attention being paid to true enablement and expertise in a noisy field. This made antisense look like the ugly cousin. In the last 2 years, however, antisense has regained favor as the RNAi delivery ‘problem’ became widely publicized which made antisense look deceptively simple. Santaris especially made a living of that by advertising, e.g. at conferences and press releases, the delivery problem of RNAi Therapeutics and getting a few laughs by stating their motto of 'staying naked’ and ‘staying short’.

Needless to say, my favored technology is RNAi Therapeutics: ‘natural’ and ‘potent’ is my motto. The hypercholesterolemia arena will be a particularly good battleground for antisense and RNAi Therapeutics to test their metals: same targets, easy biomarkers, chronic applications. The next chapter will be top-line data from Alnylam’s ALN-PCS phase I trial which are expected by year-end.

Merck Seeks to Optimize Value of ApoB as Target for Hypercholesterolemia with Combinatorial RNAi Therapeutics

A series of recent papers (e.g. Ason et al. 1; Tadin-Strapps et al.; Ason et al. 2) shows that Merck wishes to use RNAi Therapeutics for the treatment of hypercholesterolemia, a precursor of cardiovascular disease. Despite the success that widely prescribed drugs such as statins have had in lowering bad cholesterol, there are still many patients in need of additional treatment options, patients that either have very high cholesterol levels to start off with or those that do not respond to or tolerate these drugs. As a result, interest in the pharmaceutical industry remains high in developing new approaches to treat these underserved patient populations.

The Pros and Cons of ApoB as a Target

ApoB has emerged as a very attractive, hitherto undruggable target in this endeavor, and indeed ISIS in partnership with Sanofi-Aventis are currently knocking on regulators’ doors to get their ApoB-targeting RNaseH-antisense drug mipomersen (commercial name: Kynamro) approved for familial hypercholesterolemia. Meanwhile, Tekmira and apparently Merck are in the early stages of developing ApoB-based RNAi Therapeutics for hypercholesterolemia.

What makes ApoB such an attractive target is that it is the critical protein component of bad LDL cholesterol (LDLc) and knocking it down therefore very potently reduces LDLc levels in the blood. Not only that, it seems that through the wonders of sophisticated feedback control mechanisms of lipid metabolism, essentially all other atherogenic lipids are reduced, too, following ApoB knockdown (e.g. Tadin-Strapps). There remain, however, two concerns with ApoB as a target.

The first one relates to the observation that in most rodent models, not only all the atherogenic lipids are reduced, but also the ‘good’ HDL-cholesterol which is responsible for reverse cholesterol transport from the plaques (where they are dangerous) back to the liver for excretion in bile (which is where they belong to). Research by Merck, of course using LNP technology, shows that when ApoB is knocked down by ~95%, both HDLc and LDLc where reduced by more than 2/3 (Tadin-Strapps et al. 2011; 79-90% non-HDLc lowering and 67-78% HDLc lowering in Ason et al). This was highly unlikely due to an off-target effect as various ApoB-targeting siRNAs exhibited this phenotype while non-targeting LNP formulations did not.

Nevertheless, it is unclear whether these rodent and similar non-human primate findings translate into humans, and what ApoB knockdown levels would need to be achieved to start seeing an effect on HDL. Mipomersen e.g. reduces LDLc by about a third and does not seem to affect HDLc in humans. The Merck scientists also speculate that the HDLc reduction simply reflects that in the absence of LDLc, ApoE redistributes to HDLc leading to their more rapid uptake in the liver. Therefore, despite the mantra that it is all about the HDL:LDL ratio, HDLc reductions via this route would actually be positive.

The second, probably more pressing concern is that ApoB inhibition leads to a failure to export lipids from and their accumulation in the liver, a condition known as hepatosteatosis or ‘fatty liver’. This has not only been observed in pre-clinical studies of ApoB knockdown, but was also observed in the mipomersen clinical studies (Visser et al. 2010). ISIS Pharmaceuticals, the discoverers behind mipomersen, argue that this accumulation is likely to be temporary only as compensatory genetic circuits get switched on to reverse the phenotype, a mechanism that is supported by Merck's own gene expression analysis. Moreover, it has yet to be shown that ApoB-related fatty liver predisposes to the development of liver fibrosis and ultimately liver failure or cancer which is really why we care about fatty liver in the first place.

Enhancing the Therapeutic Profile of ApoB-targeting Drugs

Giving up on ApoB in hypercholesterolemia because of the fatty liver concerns would mean forfeiting the potential of one of the most if not the most potent target in the hypercholesterolemia space. I therefore fully agree with the strategy by Merck to exploit the combinatorial potential of RNAi Therapeutics to optimize the profile of an ApoB-targeting RNAi Therapeutic, a strategy that I would fully expect of Tekmira to be evaluating as well.

The combinatorial potential of RNAi Therapeutics is one of the major attractions of this technology. Because of the almost identical pharmacological behaviors of siRNAs, it is relatively simple to employ multiple instead of just a single siRNA payload in an RNAi Therapeutic. This is particularly useful for complex diseases such as metabolic syndromes and diseases that involve resistance/escape such as cancer and viral infections. ALN-VSP02 is a dual-targeting example in oncology that is already in the clinic, and Tekmira’s Ebola RNAi Therapeutic candidate slated to enter the clinic in early 2012 will also involve at least two different RNAi triggers.

You can thus imagine that knocking down a gene along ApoB that leads to increased lipid excretion via the bile, increased fat oxidation in the liver, or redu­ced hepatic fat synthesis or reduced uptake of dietary fats in the liver, would greatly enhance the therapeutic profile of an ApoB-based drug by countering the development of fatty liver. It is the latter approach that Ason and colleagues from Merck took in their recent paper by targeting fatty acid transport protein 5 (Fatp5) alongside ApoB as Fatp5 had been described, also through the elegant application of ddRNAi, to reverse diet-induced hepatosteatosis.­

To study the effect of Fatp5 co-knockdown on ApoB-induced fatty liver, the researchers formulated both siRNAs into LNPs and infused them into mice. Both genes were knocked down efficiently (89-95%) and as you can imagine, at these ApoB knockdown levels, the fatty liver phenotype was quite robust. Predicted Fatp5-dependent phenotypes, such as an almost 1000-fold increase in the ratio of unconjugated to conjugated bile acids in the bile, were also observed (Fatp5 plays a role in bile acid conjugation) confirming the functional knockdown of both ApoB and Fatp5.

Unfortunately, despite the potent knockdown of Fatp5, the ApoB-dependent fatty liver phenotype was not reversed in the mice which were fed a ‘Western low-fat diet’. It therefore appears that Fatp5 intervention is only useful for diet-induced fatty liver, and that approaches specific to fat excretion or fatty acid oxidation in the liver will be more promising. Nevertheless, the scientists seem to be on the right track, and with SNALP siRNA delivery, it should be relatively easy to characterize other candidate genes. Indeed, due to competition, the Mercks, Tekmiras, and Alnylams may not necessarily want to disclose their magic siRNA cocktail.

So as we are on the eve of seeing mipomersen being approved as the first ApoB-targeting compound for the treatment of hypercholesterolemia, a second generation of ApoB-targeting RNAi Therapeutics are being readied that not only aim at incremental improvements in potency and dosing frequency, but completely rehaul the therapeutic profile of ApoB-based therapeutics.

Alnylam Says Adverse Event in Tekmira ApoB Trial Was ‘Life-Threatening’

One apparent tactic employed by Alnylam in the ongoing litigation with Tekmira is not to win on arguments, but by coming out with 'revelations' aimed at unsettling Tekmira investors and potential collaborators. A weakened share price and financial position would supposedly make Tekmira more amenable to a light settlement or lowball takeover offer. Much of this has been discussed on this blog before, but today I would like to bring to the attention of the readers here one particularly grave and damaging remark that Alnylam made in their Response to the Amended Complaint, namely that the one case of immunostimulation in the phase I ApoB study that caused Tekmira to stop the trial was ‘life-threatening’:

From the Response:

‘Further, Alnylam invited Tekmira to join a joint development committee that Alnylam had formed with one of its pharmaceutical partners with the goal of assisting Tekmira and the whole field in advancing LNP technology. Moreover, it is through clinical trials conducted, mainly by Alnylam, that critical elements of Tekmira’s siRNA delivery technology have been validated in the clinic, and Alnylam has provided critical advice and counsel to Tekmira related to their pre-clinical and clinical development activities for their own products. For example, Alnylam’s chief medical officer provided critical and urgent counsel to clinicians attending to a patient in a Tekmira clinical trial that experienced a serious, life-threatening adverse reaction to Tekmira’s drug [emphasis mine].'

Not only does Alnylam make it sound like they are holding little Tekmira’s hands, have the new Chief Medical Officer rush to inept Tekmira's help, and more or less run the trials for them, this statement essentially accuses Tekmira of lying to investors and in front of regulators about the adverse event. These are very serious allegations, some of which by the way were part of the rumor mill even before Alnylam made these statements now in public (!).

These characterizations are in sharp contrast to the way Tekmira used to describe them: serious enough to be taken into consideration for future SNALP development, but not a show-stopper and certainly not life-threatening. Here is how:

Disclosure of adverse event in the phase I ApoB press release:

‘The primary endpoints of the ApoB SNALP Phase 1 clinical trial were measures of safety and tolerability. ApoB SNALP was well tolerated overall in this study with no evidence of liver toxicity, which was the anticipated dose-limiting toxicity observed in preclinical studies. Of the two subjects treated at the highest dose level, one subject experienced flu-like symptoms consistent with stimulation of the immune system caused by the ApoB siRNA payload. The other subject treated at the highest dose level experienced no side effects. Based on the potential for the immune stimulation to interfere with further dose escalation, Tekmira decided to conclude the trial [emphasis mine].’

If the adverse event had been ‘life-threatening’, then not disclosing this in the press release and subsequent investor discussions would obviously have been misleading.

The clinical details of the adverse event were then disclosed at the Drug Information Association '3rd Oligonucleotide-based Therapeutics Conference' held in Bethesda on March 23, 2010. An important function of these DIA meetings is to further the dialogue between industry and the US regulatory authorities and I assume that the FDA was in attendance and that Tekmira assumed the same when it reported the following:

‘Adverse Event in Suject 190:

- 53 Year old 103 kg male w/ hyperlipidemia

- Second in cohort to be treated with 0.6 mg/kg

- Tolerated infusion well

- ~2.5 hours felt ‘wobbly in legs’

- ~4 hours rigors, vomiting, fever, hypotension, hypoxia, HR 110 BPM

- Treated with Ibuprofin, O2, saline infusion

- Fever resolved in 3 hours

- ~5 hours BP reached nadir

- Patient continued to receive fluids

- IV methylprednisolone (120mg)

- BP, HR, O2 Sat improved

- Patient normal that night

- Returned to study unit the next morning

- Discharged the next day as per protocol

- Patient maintained perfect cogntion throughout episode

Event described as moderate in severity and related to study drug [emphasis mine].'

The account above is consistent with characterizing the event as ‘moderate’ and in no way suggests a ‘life-threatening’ event. While not exactly encouraging, such safety findings are common in clinical trials.

I’m confident that if this comes to trial, this damaging characterization of the ApoB trial will add another few millions to the damages awarded to Tekmira. I simply cannot believe that Tekmira’s management would risk getting in trouble with the SEC and FDA in such a way. There simply is no middle-ground. I know Tekmira shareholders are getting tired of hearing this, an attitude attesting to the success of Alnylam's tactics.

In related News: Alnylam guides down cash as ALN-VSP02 wraps up

Alnylam reported Q2 financial results after the close yesterday. Months after the departure of the former CFO and following the $150M shelf filing, the replacement had the pleasure of announcing that Alnylam is guiding down cash from ‘greater than $275M’ guided 3 and 6 months ago to now ‘greater than $250M’. No explanations were provided for the implied 50% increase in net burn. I hope that the new CFO has retained his ‘sense of humor’- apparently a job requirement at Alnylam [correction 5 August, 2011: in the conference call that followed the press release of the financial results the increase in expenditures were attributed largely to the 5x15TM efforts; not very illuminating really, but there was an explanation].

On the day of the financial results, Alnylam also disclosed that it has completed the phase I study in liver cancer with ALN-VSP02. There was no real new development in this study from the comprehensive data presented at this year’s ASCO (related blog here). What I noticed though is that this drug candidate has now been fully renamed ALN-VSP, dropping the '02' which may have been a painful reminder that Alnylam got all its clinically-relevant LNP delivery from Tekmira [hint: my strong impression based on the VSP conflicts that emerged 4 years ago with the remarks by David Bumcrot is that ‘01’ was a lipidoid formulation, and the ‘02’ formulation resulted from Alnylam giving up and seeking help from Protiva (now part of Tekmira, of course)].

Confidence in SNALP Safety Building

Alnylam Pharmaceuticals disclosed in their Q1 2011 results that it will expand the dose escalation of the phase I study of ALN-TTR01 in TTR amyloidosis, increasing the target dose in that study from 0.4mg/kg to now 1mg/kg. This puts the company in a good position to demonstrate, unambiguously, that systemically administered SNALPs can knock down genes in the human liver. This dose expansion is an important indication that, as we learn more about the clinical performance of SNALP delivery, confidence in its safety is growing.

Given yesterday's news, the phase I study being conducted in Europe should have almost fully enrolled its first 24 patients up to 0.4mg/kg. Regulatory applications for this study were filed in late 2009/early 2010, yet it took over half a year for the first patient to be dosed in July 2010. At that time, Alnylam revealed the 0.4mg/kg target dose. This was somewhat confusing and made me even speculate that TTR01 may involve a ‘generation 1.5’ formulation as it had been Alnylam’s stated goal to demonstrate proof-of-concept of gene knockdown with this study. 0.4mg/kg, depending on the gene, is about where you’d expect to start seeing efficacy with the initial SNALP formulations, so stopping at 0.4mg/kg would have been slicing it quite thinly. Indeed, the ED50 (dose of 50% drug efficacy) for ALN-TTR01 in non-human primates is around 0.4mg/kg.

In retrospect, the original conservative dose escalation schedule might have been due to safety concerns surrounding SNALP delivery. Around the time of the regulatory submissions, Alnylam knew of one death in their liver cancer trial with ALN-VSP02 (also using SNALP delivery) which occurred some time after a patient with extensive liver mets from pancreatic neuroendocrine cancer received a second infusion of 0.7mg/kg SNALP. Dose-limiting toxicities were not observed in that trial until then up to 0.4mg/kg. Although patients in that trial are quite fragile and autopsy revealed that the patient may have in fact died as a result of too much drug efficacy (extensive necrosis of the liver mets), it was a potentially drug-related severe adverse event nevertheless. ALN-VSP02 was subsequently escalated to the robust dose of 1.5mg/kg, with 1.25mg/kg being my predicted maximally-tolerated dose.

Shortly thereafter, in January 2010, Tekmira stopped early their phase I hypercholesterolemia trial for TKM-ApoB after one trial participant at the 0.6mg/kg dose level became hypotensive and suffered from general flu-like symptoms after receiving a SNALP formulation. Because the necessary safety/tolerability profile for a hypercholesterolemia drug differs from that of a cancer drug, Tekmira did the right thing and stopped the trial and take advantage of the technology progress in the SNALP delivery field with the aim of re-entering clinical development with a formulation with presumably improved potency and robust long-term safety. Just as for ALN-VSP02, 0.4mg/kg was the dose level up to which no significant adverse event was observed with TKM-ApoB.

Since no good deed goes unpunished in the competitive area of siRNA delivery, these events were quickly made out to prove that ‘SNALP delivery is toxic’ adding to the SNALP confusion caused by Alnylam's lipidoid line of work. Regulators may not be as emotional, but it is understandable that they and Alnylam took a cautious approach in selecting their initial doses. They were, however, clever enough to design an adaptive trial that would allow for further dose escalation depending on the safety findings up to the 0.4mg/kg level.

The enrollment of the anticipated additional 8 patients should therefore be a very good sign that the safety profile so far has been highly satisfactory, setting the scene for convincingly demonstrating TTR knockdown in the liver as measured by plasma protein levels. Alnylam plans to present top-line data from the completed study in the third quarter, hopefully building on good ASCO data on SNALP-based ALN-VSP02.

There has been some debate around whether the clinical development of SNALP technology was started pre-maturely. Yes, you can try and aim for perfection (that may be what Merck and Roche are waiting for), but by taking the necessary clinical precautions, Alnylam and Tekmira were able, in a timely manner, to gather much clinical data on the SNALP delivery platform that is really starting to pay off now as their synergistic value becomes obvious. Important months for RNAi Therapeutics are ahead of us.

RNAi is not Antisense

From the abstract of an important paper in RNAi:

"Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans.Fire A, Xu S, Montgomery MK, Kostas SA, Driver SE, Mello CC.
Carnegie Institution of Washington, Department of Embryology, Baltimore, Maryland 21210, USA. fire@mail1.ciwemb.edu

Experimental introduction of RNA into cells can be used in certain biological systems to interfere with the function of an endogenous gene. Such effects have been proposed to result from a simple antisense mechanism that depends on hybridization between the injected RNA and endogenous messenger RNA transcripts...Here we investigate the requirements for structure and delivery of the interfering RNA. To our surprise, we found that double-stranded RNA was substantially more effective at producing interference than was either strand individually. After injection into adult animals, purified single strands had at most a modest effect, whereas double-stranded mixtures caused potent and specific interference..."


RNAi is not antisense- This very finding formed the basis for awarding the 2006 Nobel Prize committee to Andy Fire and Craig Mello. Despite this authoritative decision, as the promise of the RNAi Therapeutics platform is getting closer to reality by the day, the promise also of financial fortunes has led to efforts to re-brand RNAi as yet another antisense technology. Although this in a way could be considered as some kind of validation of the competitive nature of RNAi, it has important implications for RNAi intellectual property.

In the past, ISIS Pharmaceuticals, undoubtably a leader in RNA therapeutics and with nucleic acid modification IP very relevant for RNAi Therapeutics, has made a number of statements I cannot agree with. Essentially, the company claims having discovered RNAi and has long regarded Alnylam as one of their many “satellite companies”. This, however, does not reflect reality in as much as repetitively claiming that RNAi is an antisense technology does not make it a true statement.

The latest example can be found in ISIS Pharmaceutical’s press release on the issuance of an ApoB-related antisense patent summarizing the claims as covering “the use of both single-stranded and double-stranded (siRNA) antisense drugs complementary to any site of the mRNA of human apoB regardless of their chemistry or antisense mechanism of action.” “Double-stranded antisense drug”- quite an oxymoron. But note that nowhere in the claims of the patent are duplex RNAs mentioned, instead “12 to 30
nucleobases in length targeted to a nucleic acid molecule encoding apolipoprotein B wherein said compound is 100% complementary to the nucleic acid sequence set forth in SEQ ID NO: 3” are claimed.

RNAi being antisense, having discovered RNAi first, and Alnylam being a satellite company, just does not agree with the fact that the very reason why Fire and Mello’s discovery is considered a major breakthrough, is because they found that RNAi is induced by double-stranded RNAs and not by antisense molecules as had been widely believed, and this indeed was the very working hypothesis that Fire and Mello famously disproved and for which they were awarded with the Nobel Prize. Come to think of it, RNAi as an “anti-antisense” technology may be a more apt description.

The claim why RNAi is supposed to be an antisense technology is largely based on the observation that during RNAi, the double-stranded siRNA is unwound, and the single-stranded guide RNA strand incorporated into RiSC which it then guides to target mRNA partly, but not exclusively, based on base-pair complementarity. Just because base-pairing interaction is involved, however, does not make it antisense. What is important for a biotechnology that harnesses an endogenous biological pathway are the features that make it being recognized as a bona fide substrate. In the case of RNAi, this clearly is double-stranded RNA with a couple of additional characteristics, and not single-stranded antisense RNA. This is the result of RNAi having evolved as a way to recognize and fight RNA viruses which are characterized by double-stranded RNA intermediates. Re-branding RNAi after one of the components of its inducers would be like saying the wheels of a car are the car.

While single-stranded antisense RNAs may also induce RNAi, research by ISIS itself and others has shown that they do so much efficiently, about 20- to 50-fold less efficient than double-stranded RNAs, that is about as efficient as a conventional RNaseH antisense molecule (probably another reason for ISIS’ growing interest in RNAi).

With so much “new fundamental” trigger IP being claimed, RNAi now becoming an antisense technology etc., are we about to see a different approach by the RNAi Therapeutics companies in possession of the real fundamental IP? I would not necessarily expect legal proceedings as long as it does not start to affect their business dealings. However, such rhetoric has the potential to lead to market inefficiencies, hurt smaller companies with a less active PR department (see Tekmira and their ApoB program), and lead to a situation in which as the first RNAi Therapeutics are approved and patent disputes to be resolved, the debate will have been re-framed to the disadvantage of RNAi Therapeutics companies.

PS: The RNAi-is-Antisense claims also get to the important issue of target IP. Target-specific IP should be encouraged as it incentivizes the identification of suitable RNAi Therapeutic gene targets, as long as certain minimum validation standards are satisfied. I would therefore agree that if the ISIS antisense patent had been a milestone in demonstrating the suitability of apo-B inhibition for the treatment of hypercholesterolemia (note: ribozymes targeting Apo-B were specifically exempted from the claims which already suggests that the present patent is not the seminal ApoB-hypercholesterolemia patent), this should also have consequences for drugs based on other mechanisms of actions such as RNAi, monoclonal antibodies etc. In cases where there is a proven benefit in targeting at the RNA level, the royalty rate should climb commensurately, although I do not believe that in practice this means that an antisense therapeutic will be able to stop an RNAi Therapeutics from reaching the market and vice versa.

Note: The abstract of the Fire and Mello paper has been added after the first publication of the blog entry.

ENHANCE Ripple Effects Felt at ISIS Pharmaceuticals

The waves started by the ENHANCE tsunami have finally come to engulf ISIS Pharmaceutical’s stock this Friday as the FDA indicated that it wants to see costly and more time-consuming outcome studies for all but the most at-risk cardiovascular patient population in the approval of ISIS’ lead product mipomersen, an antisense compound directed against ApoB-100 in the liver. As I had discussed earlier in the wake of two high-profile studies that failed to correlate the responses of some of the most trusted biomarkers for diabetes and cardiovascular disease with disease outcome, most notably Merck/Schering-Plough’s ENHANCE study, this was almost to be expected.

The 29% drop in the stock that day was caused by a discussion between ISIS and their partners at Genzyme with the FDA on the approval path for mipomersen. Instead of being able to gain accelerated approval by merely showing LDL-cholesterol lowering as a surrogate biomarker for a broader patient population, this will only be accepted for the small homozygous familial hypercholesterolemia (ho FH) population, and it was not clear from the conference call whether this would be a full or an accelerated approval. For all other indications, mipomersen now has to show tangible benefits with regards to cardiovascular events, which, although it would eventually increase the market value of the drug considerably, means longer and more expensive outcome trials before the real commercial value of mipomersen can be realized (in addition, ho FH will be delayed by at least one year, because the FDA requires 2 rodent carcinogenic studies, instead of the one they had planned for).

What can the RNAi Therapeutics space learn from this experience? One lesson is that it is very important to be clear early on with what the almighty FDA wants to see in approving a drug, something particularly relevant for novel technologies and novel drug targets such as will be the case with RNAi. It is probably not a coincidence that clarity has come now that Genzyme has joined the mipomersen effort, as it is probably THE company experienced with gaining regulatory approval for unique, high-margin drugs with small, but very well-defined patient populations, a category that a number of RNAi Therapeutics will fall into as part of the personalized medicine revolution.

Another lesson is that the fortunes of an entire technology platform may be subject to the woes of a single dominant development program. While the cardiovascular disease market is enormous and the potential rewards substantial, this is not without a reason and substantial investments have to be made before gaining approval. A number of RNAi Therapeutics companies such as Alnylam, Tekmira/Protiva, RXi, Mirus Bio, and Merck have all indicated an interest in targeting very promising liver targets for hypercholesterolemia by RNAi, and although it may limit the ultimate financial reward, a company like Alnylam that carries much of the hope for realizing the therapeutic promise of RNAi, may have to think twice about to what degree it wants to make itself dependent on a single development program such as ALN-PCS01.

Companies targeting ApoB by RNAi (probably including Tekmira/Protiva, RXi, and Merck) should benefit considerably by learning from ISIS’ pioneering ApoB experience, and actually should have a chance at gaining accelerated approval should ISIS be able to show that lowering LDL-cholesterol by targeting ApoB is associated with cardiovascular benefits. For companies like Alnylam though interested in other, previously untested targets, the decision may only confirm that, yes, outcome studies will be required.

I welcome the trend towards evidence-based medicine and addressing some of the problems associated with direct-to-consumerism, but again, the FDA and politicians should not blindly destroy the drug development industry by forgetting that at the same time there need to be clear rewards for innovative (patent protection!) and efficacious (reimbursement!) drugs. This way the pie could stay around the same size, and yet everybody including patients, payers, and the industry would be winners.

Also this week: In a catastrophic financing, Nastech raised just shy of $8M in a registered direct offering at a probably historically low $1.73 a share, and may raise another $3M at $2.17 a share. One has to wonder why this relatively small fund-raising round was done at all at these so unfavorable conditions- unfavorable at least for the present shareholders. In any case, this may be the best time to cut their losses and fully commit to developing RNAi Therapeutics while monetizing what is left over from TJT nasal delivery. Their experience in drug delivery and peptide technology in particular may position them to be a respectable player in peptide-facilitated RNAi delivery (direct conjugation of RNAi trigger to membrane-penetrating peptides and/or using peptides as ligands for targeted delivery). But there is one advice that I would give them: please do not waste any more of your shareholder money and your own credibility on senseless and blatant patent workaround efforts such as three-stranded siRNAs. As the 36% drop on Friday illustrates, investors are sophisticated enough to see through this and rather than claiming to own everything under the RNAi sun, or universe, it is more credible to specialize in an area of your expertise and be good at it. Develop clinically relevant delivery for example, and the market will more than generously reward you for it.

More Breaking News: ISIS and Genzyme in Heart Attack Biotech Deal

Dwarfing Tacere’s Pfizer collaboration announced earlier today, ISIS disclosed after the close of the markets Genzyme to be their long-expected partner for mipomersen, an antisense compound targeting ApoB100 for lowering bad cholesterol. Mipomersen is currently in phase III clinical development for the small, but high-risk group of people with familial hypercholesterolemia. Similar studies are planned to apply mipomersen for the statin-refractory general population. Pre-clinical data so far have shown the drug to be safe and surprisingly well tolerated.

The blockbuster deal, including a $175M upfront license payment plus $150M at a 103% premium to today’s closing price and more significant milestone and profit-sharing opportunities downstream, is another validation of how highly innovative RNA-targeting therapies are being valued. Genzyme appears to be the ideal partner for ISIS in that it has demonstrated an amazing ability to target relatively small patient populations for genetic disorders, such as familial hypercholesterolemia, yet be extraordinarily profitable. ISIS on the other hand is opting for an IP licensing model whereby it develops a series of antisense compounds to certain value inflection points to then partner them, in addition to receiving significant royalties on licensed IP from a wide range of RNA technology companies, including Alnylam. More antisense drugs for liver indications are in development, including a phase I HCV drug licensed by Merck [note corrigendum below].

Today’s announcement will focus the attention to the entire field of RNA-based therapeutics, including RNAi. The markets are likely to react very positively to the news, and Alnylam’s plans of two more significant platform licensing deals will sound a lot more real with this new development.

PS: It seems odd that today's announcement was preceded by a dramatic, over 15% drop in ISIS shares at the opening of the session. It is likely that "shorts" familiar with the matter pushed the stock down to create panic selling to create volume for covering their positions. Unfortunately, this may have achieved its purpose and some honest investors will have been stopped out by this action.

Corrigendum (15 January, 2008): A reader correctly noted that the Merck HCV drug mentioned in this blog is a nucleoside polymerase inhibitor, not an antisense compound which I assumed based on its origin in a Merck-ISIS collaboration. This, however, further illustrates ISIS' considerable leverage in nucleotide-based drug development.

The Race to Knocking Down Cardiovascular Disease

Given the burden of cardiovascular disease in the Western world representing a multibillion dollar drug market, finding a drug to complement statins in reducing complications due to high levels of bad cholesterol is naturally high on the priority list of many drug developers. The recent OTS Meeting and a Press Release by Alnylam emphasising their leadership by having obtained first-ever data on safely and effectively knocking down PCSK9 with RNAi in non-human primates, illustrate home the promise of RNA-based therapies for CVD. The interest is largely rooted in the fact that targets such as PCSK9, ApoB100, and potentially microRNA-122, well known determinants of blood cholesterol levels, but which have proven impossible to target by traditional small molecule approaches. Moreover, these targets are expressed in the liver, and it is clear by now that current systemic oligo delivery technologies allow them to be knocked down in vivo. Hence, the race is on to who will be first to develop a safe and efficacious oligonucleotide-based therapy for hypercholesterolemia and stand to reap the benefits of a potential blockbuster in the first phase of RNA-based drugs.

Assuming that it is a safe bet that cholesterol levels can be reduced with oligo-based strategies, what will determine regulatory success? Given that low cholesterol is a life-long effort, any drug taken over a long period of time, even before disease onset, will have to be safe first of all. Risk can be largely grouped into four categories: target risk, risks inherent to the therapeutic platform, sequence risk, and risks associated with route of delivery and drug formulation. Arguably the target best validated on the grounds of human genetics is PCSK9, a protease that degrades LDL-receptors and therefore inhibits clearance of bad cholesterol from circulation. Research mostly from the University of Texas Southwestern has shown that mutations that increase the activity of PCSK9 increase cholesterol levels, whereas individuals with nonsense mutations in PCSK9 that reduce PCSK9 activity have lower cholesterol levels and, importantly, a much reduced risk for cardiovascular events. Moreover, the absence of any functional PCSK9 throughout life has no obvious adverse side-effect while retaining the health benefits of low cholesterol.

Before PCSK9 came to the fore, ApoB100, a protein required for the assembly of LDL-cholesterol, used to be the target of choice. Indeed, the development of PCSK9-based treatment strategies have extensively made use of ApoB100 as a marker protein for evaluating RNAi delivery and knockdown in the liver. Pioneering research mostly by ISIS Pharmaceuticals has shown that indeed ApoB100 knockdown has the ability to lower LDL-cholesterol. Although ISIS has not seen fatty liver in clinical trials and preclinical research of their lead antisense compound ISIS 301012 (currently in late phase II) to be a problem, various other groups have observed this side-effect following ApoB100 knockdown, which would not be that surprising given the role of ApoB100 in fat metabolism. However, even if fatty liver will be observed in larger phase III trials and post-approval, ISIS has made the right decision to test 301012 first for patient populations most at risk for CVD.

Similar to ApoB100 and PCSK9, inhibition of microRNA-122 by antisense technologies has been now shown numerous times to also have LDL-cholesterol lowering effects. Strangely, despite the fact that this is by far the most abundant microRNA in the liver, no obvious toxicities have been associated with miR-122 inhibition. Consequently, a number of groups such as Regulus and Santaris hope to develop this into a treatment for hypercholesterolemia.

Taken together, my bet is on PCSK9 knockdown to lead the way in oligo-based therapies for the long-term treatment of hypercholesterolemia. New targets, however, should emerge, partly as a result of now being able to apply RNAi itself for target identification, for example by transiently targeting essentially any gene of interest in the liver in vivo and the use of transgenic RNAi mice (Artemis), a combination of the two latest Nobel prize-winning technologies.

Next to target choice, the nature of the knockdown technology, antisense versus RNAi, itself will also have important safety implications. As I am quite fascinated about the prospect of RNAi for various reasons, please keep in mind that my natural inclination is to favour RNAi any time. In terms of potency, once equal amounts of oligos get delivered into the cell, RNAi has been shown frequently to be generally superior to antisense oligos (ASO), although antisense technologies can be quite diverse. Lower dosages will not only reduce cost of a treatment that has to be taken long-term, but, more importantly, allow for dosages that fall well within therapeutic windows. Moreover, in the case of RNAi, I feel quite comfortable with a technology where the risks such as immuno-stimulation, off-targeting, and potential interference with the endogenous microRNA pathway are reasonably well understood, intensely studied, bioinformatics- and chemistry-based solutions devised, and well taken into account in current RNAi-based drug development efforts. This in fact reflects a new awareness in RNA-targeted therapies, largely driven by the renewed interest generated by the discovery of RNAi. Accordingly, the therapeutic utility of any two RNAi compounds, or antisense compounds for that matter, may differ dramatically due to sequence-dependent toxities.
These toxicities may also be linked to route of delivery and related oligo formulation. A technically quite uncomplicated approach, as taken by 301012, is to simply administer relatively large amounts of unformulated oligos (200mg/week in the case of 301012) to make sure that enough of it ends up in the liver. By contrast, liver uptake of siRNAs is thought to require additional formulation. Indeed, liposomal formulations that are set to enter the clinic within the next year increase liver uptake of siRNAs from less than 1% of injected material to over 30%, allowing for lower dosages to be used. Some toxicities, however, were observed at relatively high dose levels with some of the cationic liposomes, and it remains to be seen whether lipidoids and other “not-so-cationic” liposomes will come to dominate the liver delivery field. Also, while most of the disclosed liposomal delivery vehicles efficiently enhance liver uptake, they are often not specific for uptake into the hepatocyte population in the liver, the cell type of interest. Particularly uptake into Kupffer cells, a type of immune cell in the liver, can lead to dosing and safety complications, and ultimately the path taken recently by scientists from Mirus, which by the way has an RNAi delivery collaboration with Pfizer, to specifically target formulated siRNAs to hepatocytes, but not other liver cell types, may substitute non-specific liposomes in the second wave of RNAi-based therapies for hypercholesterolemia. While delivery is often described as the Achilles Heel for RNAi therapeutics, the charge (ironically) and chemical similarity of siRNAs as a class makes them ideally suited to devise drug targeting strategies that can be broadly applied and should lead to safer therapies, something that is nearly impossible for say small molecules.

ISIS’ ApoB100-targeting antisense 301012 has good chances of becoming the first oligo-based therapy for CVD, at least for people with familial hypercholesterolemia and for whom statins don’t work. Although only a fraction of the overall market, the sheer size of the cholesterol market makes this a lucrative goal nonetheless. I am somewhat surprised that, to my knowledge and despite potential target risk, there is little talk of other ApoB100-targeting therapies. It will be interesting to see what companies like Merck, which has clearly stated their admiration for 301012 at the last OTS Meeting, are willing to pay for rights to 301012. PCSK9-targeting therapies are in late preclinical development and therefore about 3 years behind 301012, but I believe these to be the safest bet for a widely applied oligo-based drug for hypercholesterolemia with a number of organisations ramping up their PCSK9 programs.

Alnylam appears to be leading this race with the recent announcement of first-ever non-human primate data of an RNAi compound that safely and effectively knocked down PCSK9 with concomitant reductions in total and LDL-cholesterol. An IND is planned for the end of this year, or early next year, and probably will depend on finding the delivery solution that most importantly is safe for long-term administration. Importantly, Alnylam enjoys a particularly strong IP position and know-how in targeting PCSK9 by RNAi, due to their own position in fundamental RNAi technology, and important collaborations on the biology of PCSK9 with UT Southwestern, which has been leading in the genetics of PCSK9, as well as in delivery with the Anderson/Langer lab at the MIT and exclusive access to Tekmira’s cationic liposomal delivery IP for RNAi. Sirna-Merck may want to dispute this with an patent on targeting the same PCSK9 by RNAi that issued recently and was filed in July 2006 as part of their brute-force approach to patenting genes for RNAi. Alnylam, however, presented their first PCSK9 RNAi data in mice at last year’s 2nd Annual OTS Meeting, and it is anybody’s guess when their or rather UT Southwestern’s first lab-book entry on PCSK9 RNAi occurred. Probably at a similar stage to Alnylam is the PCSK9-antisense collaboration of ISIS with Bristol-Myers Squibbs for which mouse data have been published earlier this year. Santaris’ antisense compounds for PCSK9/ApoB100 and miR-122 should also be heading soon towards the clinic.

New delivery technologies, including oral formulations, and targets should ensure that the oligo-CVD field will remain lively in the years to come. Also, since there have been a number of recent data demonstrating efficient targeting of RNAi to the endothelia of blood vessels, new RNAi strategies aimed directly at the atherosclerotic plaques may emerge.

It would not be the first time that several similar compounds, small molecule, antibody or recombinant protein, with essentially the same molecular targets, would co-exist in a market, a concept also very familiar to the hypercholesterolemia field. IP, careful clinical development involving the best scientists in both oligonucleotide technology, delivery and the biology of the drug targets, together with a bit of luck, will decide who will reap the largest benefits from the potentially first knockdown blockbuster.

Impressions from the 3rd Annual Meeting of the Oligonucleotide Therapeutics Society: Day 1

The 3rd Annual Meeting of the Oligonucleotide Society (OTS) got under way in the stylish Langenbeck-Virchow Auditorium situated in the centre of a culturally vibrant Berlin. The OTS brings together the many disciplines in the rapidly expanding field of oligonucleotide-based drug development, ranging from aptamers, ribozymes, antisense, immunostimulatory nucleic acids to RNAi. Reports from clinical studies of these treatment modalities and oligonucleotide chemistries were at the centre of today’s lecture sessions.

Brett Monia from ISIS Pharmaceuticals chaired the first session and set the stage by reporting that while there are only four approved oligo-based therapies, more than 50 are currently in clinical development. Many of these are designed for cancer applications, but infectious, metabolic, and inflammatory disease applications are emerging as major new focus areas. This may not only be due to considerable unmet medical needs, but also because these may offer better success rates for clinical development. A few talks on cancer therapies with ASOs (antisense oligos) documented the challenges of designing smooth clinical trials for cancer therapies. One problem is the heterogeneity of cancers and matching the gene target with the right cancer. Early trials on a given compound typically involve small patient numbers with a wide spectrum of cancer manifestations. Although safety trials, almost anecdotal evidence of biological activity in a few patients then often guide the design of later phases of clinical development. Clearly, for cancer, there is a need for the development of more reliable biomarkers as early indicators of therapeutic efficacy and for patient stratification, which may be something that Rosetta Genomics’ microRNA diagnostics may be able to facilitate.

Focusing on diseases like viral infection and metabolic disease where efficacy can be routinely evaluated in early stages of clinical development (viral titers, LDL-cholesterol, glucose) should therefore be lower hanging fruits for the next wave of oligonucleotide-based therapeutics. This transformation in the industry is illustrated by ISIS’ new focus on metabolic disease and its out-licensing of the more challenging ASO applications. Rosanne Crooke recounted the clinical experience with 301012, ISIS’ lead compound for the treatment of hypercholesterolemia. As I wrote earlier, one major concern with targeting ApoB100 is the risk of steatosis of the liver due to inhibition of a protein required for the export of triglycerides and which has indeed been observed by a number of groups targeting ApoB100 by RNAi. Possibly motivated by the existence of individuals with null mutations for PCSK9, the new target of choice for hypercholesterolemia, with no evidence for adverse effects due to lack of the LDL-receptor processing protein, a conference participants asked about the existence of ApoB100 knockout mice. Interestingly, no such mice exist, because ApoB100 would have additional roles during development.

Despite reports of steatosis due to knockdown of ApoB100 with RNAi, Sirna Therapeutics is yet another biotech company interested in targeting it for lowering cholesterol. In a talk about siRNA delivery to the liver, Barry Polisky, CSO of Sirna Therapeutics/Merck also noted that in Sirna’s experience, there are genes that can be knocked down extremely well (e.g. ApoB100) and some that are relatively refractory to knockdown by RNAi. Although anecdotal, given the extensive experience of Sirna Therapeutics with screening for effective siRNAs by tiling them across entire genes, and similar findings by Qiagen and others, it will be interesting to determine what makes a gene a good target for RNAi, particularly whether it involves its normal regulation by microRNAs.

Polisky further focused on the need for siRNA modifications to make RNAi therapeutics a reality. Currently, Sirna’s standard siRNA design seems to involve si(R)NAs that lack essentially all 2’OH groups, except for the three 5’ nucleotides of the guide strand, while the passenger strand has 5’ and 3’ modifications that avoid it to be loaded into RiSC to reduce off-targeting. Other speakers and posters also noted that since most nucleotide modifications had been developed with antisense therapies in mind, new modifications may exist for optimal RNAi efficacy, particularly in vivo. Indeed, new synthesis strategies are being developed that will allow this to be tested, such as the 6-membered carbohydrate substitutions of ribose reported by Piet Herdewijn from the University of Leuven, Belgium.

Polisky further characterized siRNA delivery to the liver as efficient, albeit not optimal. Progress is clearly illustrated by the fact that >30% of siRNA formulated into liposomal nanoparticles now end up in the liver with first-pass kinetics with no evidence for toxicity in the 1-9mg/kg range (mostly rodent studies), whereas >99% of unformulated siRNAs are rapidly filtered out by the kidney.

The liver is also the organ where the first microRNA-based therapies are likely to be targeted. Like Regulus and Rosetta, I was amazed, or maybe not, at how many groups are currently working at abrogating miR-122 function for treating either hypercholesterolemia or HCV infection. LNAs, owned by the privately held Danish company Santaris, appear to be a particularly promising oligonucleotide class for achieving potent microRNA down-regulation in vivo. Similarly, antisense strategies based on LNA appear to rival siRNA potency in a number of settings, and it will be interesting to learn more about the toxicologies of these compounds, which would certainly be facilitated if Santaris decided to enable to field by allowing better access to their technology. It also highlights the need for specialized chemistries for therapeutic targeting of microRNAs which in my mind is something that Regulus ought to consider.

With regard to delivery outside the liver following systemic administration, PEI-formulated siRNAs (see recent PolyPlus-Alnylam deal) are now being used successfully by a number of groups for lung delivery. Other interesting delivery data was presented by a Portuguese group in a poster on RNAi in the brain. By non-covalently adding transferrin protein to cationic liposome-siRNA particles, they were able to efficiently target neuronal cells both in vitro and in vivo which, similar to cancer cells (see Calando’s cancer cell targeting with transferrin), are studded with transferrin receptors.

PS: A nuisance to most scientists in the field, immunostimulation by oligonucleotides is exploited by Dynavax as an adjuvant for vaccines. Poor adjuvants activity lead to insufficient immune responses and the need for multiple booster injections affecting compliance. HBV vaccination is a case in point, and as a consequence around 50% of patients in the US fail to successfully complete vaccination. Very high HBsAg antibody titers after only two injections of recombinant HBsAg with an immunostimulatory oligo from Dynavax as adjuvant, instead of alum, suggests that they may be able to greatly improve on those numbers.

Mirus Scientists Publish Elegant Paper on Targeted siRNA Delivery to Hepatocytes

The ability to systemically administer siRNAs and functionally modulate gene expression in tissues of interest is considered by many an important step to opening up the therapeutic potential of RNAi to a wide range of diseases. The liver is arguably the best example where intravenously administered siRNAs have already been shown to potently knockdown target genes from mice to non-human primates.

The most promising delivery technologies so far (next to gene therapy vectors such as AAV which I will not discuss here) involve liposomes such as the SNALP particles which were first pioneered by Protiva scientists and subsequently became the subject of intense legal battles involving Protiva, Tekmira, Sirna Therapeutics, and Merck. While facilitating highly efficacious gene knockdown, toxicities such as liver enzyme elevations and non-linear relationships between knockdown and siRNA dosage have been repeatedly reported with these chemistries. While the toxicities were observed at relatively high dose levels, non-linear dose responses complicate the choice of the right dose for entering clinical trials.

One step in the right direction was taken when Daniel Anderson from the MIT reported earlier this year at the Keystone meeting the identification of a slightly different class of compounds which they termed “lipidoids”. Importantly, this class of chemistries appear to efficiently promote RNAi gene knockdown with little if any apparent toxicities and linear dose responses.

The just released paper by Rozema and colleagues from Mirus Bio Corporation (Rozema et al. PNAS Early Edition 24 July 2007: Dynamic PolyConjugates for targeted in vivo delivery of siRNA to hepatocytes) sheds light on some of the issues above and holds out a new paradigm for achieving safe and efficacious therapeutic RNAi knockdown in select cell populations of the liver. Rather than regarding drug delivery to the liver as a passive process given that the bulk of intravenously injected drugs will pass through it with a relatively high chance of entering resident cells there, the present approach involves attaching a simple galactose-derived ligand thereby actively targeting asialoglycoprotein receptors (ASGPr) displayed on hepatocytes.

Knockdown of ApoB100 and microscopic analysis of fluorescently tagged small double-stranded nucleic acids confirmed that hepatocytes were indeed efficiently transfected. Equally important, however, was their observation that Kupffer cells, a major population of macrophages in the liver that are implicated in many cases of drug-related liver toxicities, did not take up the siRNA mimicks, whereas non-ASGPr targeted particles were able to transfect surrounding cells, including Kupffer cells.

Taken together with the SNALP and lipidoid data, this study supports the hypothesis that Kupffer cells may act as a sink for certain siRNA formulations such as SNALPs, thereby not only causing non-linear dose responses, but also an immunogenic response and related toxicities. It also shows that it should be possible to design simple and small siRNA nanoparticles for RNAi delivery from relatively cheap materials. Although lipidoids appear to be a clinically viable technology already, it is comforting to know that alternative routes exist and gratifying to see almost daily improvements being made in the RNAi delivery field.

PS: An interesting aspect of this publication was data the authors obtained on ApoB knockdown. In addition to serving as a target for early proof-of-concept studies for systemic RNAi delivery, ApoB has been a favourite for treating hypercholesterolemia using both antisense and RNAi. ISIS Pharmaceuticals in particular has an anti-ApoB antisense compound in late phase II clinical trials. However, only a month ago, Alnylam reported that in their hands ApoB knockdown led to unacceptably high levels of fat accumulation in the liver (fatty liver phenotype), a finding that was confirmed in the present study. Moreover, both reports indicate that this was a siRNA sequence specific effect and was achieved using two different delivery methods. This physiologic response makes a lot of sense, since ApoB’s main role is in the export of cholesterol and triglycerides from the liver. Failure to export them should accumulate them in the liver.

This is a good example where RNAi can serve both as a target validation tool AND a platform for developing innovative drugs. Consequently, I fully support Alnylam’s decision to target PCSK9 in their hypercholesterolemia program instead, which is a genetically well validated target for reducing LDL-cholesterol and heart disease (see also Blogs from 6 May, 2007: “Preventing Heart Disease with RNAi Therapeutics”, and 9 May, 2007: “ISIS Copies Alnylam’s Heart Disease Strategy“). It is curious then that ISIS maintains and has published the absence of such a phenotype. For the sake of patients and stakeholders in ISIS, I can only hope they are right. The explanation for the apparent discrepancy? I cannot really offer a good one except to speculate on a fortuitous ISIS 301012 antisense off-target effect.

Alnylam Chooses Lipidoids over Cationic Liposomes for their First Systemic RNAi Clinical Studies

[Important update at end of this entry]

In an interesting twist, Alnylam announced at the Beyond Genome conference held last week in San Francisco the use of MIT’s lipidoid technology for their first systemic RNAi programs. These formulations will be used for knocking down PCSK9 for the treatment of hypercholesterolemia and the dual siRNA cocktail ALN-VSP01 for liver cancer. This was somewhat surprising, following a proof-of-concept Nature study last year that demonstrated efficient systemic RNAi delivery in primates using cationic liposomes. These were developed by Protiva/Tekmira, and Alnylam consequently established a broad alliance with Inex Pharmaceuticals, now Tekmira, that comprised an exclusive license to Alnylam to Tekmira’s liposomal delivery IP estate. It was therefore largely expected that Alnylam would use Tekmira’s cationic liposome SNALP technology in their systemic RNAi programs which are scheduled to enter the clinic by the end of this year.

As I pointed out in some of my earlier blogs, I was concerned that Alnylam’s plans to move into systemic clinical trials so quickly were too aggressive. This concern was largely based on the considerable, albeit transient elevation in liver enzymes, a measure of liver toxicity, at the 2.5mg/kg dose range reported in last year’s Nature study. Moreover, combined with the non-linear dose response for cationic liposomal siRNA delivery that Alnylam reported at the Keystone RNAi meeting earlier this year, this made choosing the right dose range for human studies less than certain.

Interestingly, at the same Keystone meeting, Dan Anderson together with Rob Langer, a world authority on drug delivery and scientific advisor for Alnylam, presented impressive systemic delivery data using so called “lipidoids”. “Lipidoids” were discovered as part of a library approach as an apparently new class of lipid-like molecules that were very effective in delivering siRNAs systemically to the liver and were structurally sufficiently distinct to conventional lipids and cationic polymers to give it a new name. Only a few months after that, Alnylam and the same groups at MIT announced a broad systemic RNAi delivery initiative in which Alnylam would fund 10 post-doctoral researchers for 5 years in return for exclusive rights to the lipidoid technology and an exclusive option for any new RNAi delivery technologies resulting from the sponsored fellowships.

Importantly, the tiny lipidoid formulations were reported to have a favourable safety profile and showed dose-dependent gene suppression without compromising RNAi knockdown efficacy.

In hindsight, it is very reassuring to see that Alnylam did not go out on a limb by promising an aggressive systemic RNAi timeline and gamble an early program on a potentially unsafe delivery technology. It speaks to the quality of the management and Alnylam’s reputation as the leader in RNAi Therapeutics that it always had valid options outside cationic liposomes.

Where does this leave the Tekmira-Alnylam alliance? Although “lipidoids” appear to be somewhat distinct to cationic liposomes, it is certainly a good insurance to be covered by Tekmira’s important IP estate in the field of liposomal drug delivery. Moreover, as part of the alliance Alnylam invested in Tekmira’s manufacturing capabilities, and it appears this investment will pay off as Tekmira is manufacturing the lipidoid-siRNA formulations for the PCSK9-hypercholsterolemia and liver cancer trials. According to David Bumcrot of Alnylam, IND-enabling studies for these programs are well underway.

PS: In another interesting twist, David Bumcrot noted that part of Alnylam’s decision to target PCSK9 in their hypercholesterolemia program is due to a fatty liver problem seen in targeting the former front-runner ApoB100. This appears to be a target-specific phenotype as this phenomenon has been seen with a number of siRNAs targeting ApoB100. Interestingly, ISIS Pharmaceuticals which has an ApoB100 antisense compound in late phase II trials, but has not reported on that problem has followed Alnylam’s lead in targeting PCSK9 in a new hypercholesterolemia program (see May 9, 2007 post). Once again, Alnylam has demonstrated a characteristically circumspect development approach that includes bringing together the best scientists in a given disease area to carefully characterise RNAi knockdown phenotypes.

[Update May 1, 2008: According to a report by RNAiNews , an Alnylam spokesperson indicated that the company had not given guidance on the specific liposome formulations to be used for their hypercholesterolemia and liver cancer clinical programs. This is in contrast to an earlier report by RNAiNews from last year’s Beyond Genome conference in San Francisco which indicated that Alnylam had chosen “choose lipidoids over SNALPs” for these indications (also discussed in a blog entry here). It therefore seems that Protiva/Tekmira's 0.1mg/kg IC50 liposomal nanoparticle formulations may be the current frontrunners in entering the clinic (see also a recent PR).

The sometimes imprecise use of the terms SNALPs and lipidoids may be partly to blame for the confusion. SNALP refers to a liposomal formulation technique, while lipidoids are a new class of lipids generated by combinatorial chemistry which can now be evaluated for liposomal drug delivery. As such, lipidoids could be used within the context of SNALPs, and both of these approaches are therefore complementary to each other.]

Preventing Heart Disease with RNAi Therapeutics

High blood cholesterol, a major risk factor for heart disease, already is an enormous public health problem that is likely to worsen given current lifestyle and demographic trends. Fittingly, statins, a class of small molecules that inhibit cholesterol synthesis and therefore have a wide range of effects that together lower the risk of developing heart disease, are probably the best-selling drugs on the market today. However, many patients still do not achieve their cholesterol goals and there is significant demand for new innovative treatment options.

An RNAi Therapeutic alternative, due to the ability to target almost any gene, is particularly interesting because of well validated targets, but which have proven refractory to targeting by the existing drug classes. These genes are not directly involved in cholesterol synthesis and targeting them should be synergistic with statins. Currently the most interesting gene targets are ApoB100 and PCSK9. ApoB100 is the sole protein component of “bad” LDL cholesterol and is produced in the liver. Notably, targeting ApoB100 by RNAi in the using cholesterol-conjugated siRNAs in 2004 by Alnylam scientists was also the first demonstration of gene silencing following systemic administration of siRNAs. This study not only showed considerable reductions of ApoB mRNA and protein, but also the hoped for decrease in blood LDL cholesterol. I should add that clinical data from phase I and II trials conducted by ISIS Pharmaceuticals using antisense oligo technology further document the promise of ApoB100 as a target for hypercholesterolemia. It will be interesting to follow their further clinical progress, but I expect siRNAs to do even better, because of increased specificity and potency thus allowing for lower amounts of nucleic acids to be administered.

Two years later after the demonstration of systemic RNAi in mice, Alnylam scientists then reported even enhanced ApoB-100 silencing and improved lipid profiles in monkeys, this time using liposomal formulations originally developed by Protiva Biotherapeutics. These and similar liposomal formulations have proven to be very efficient for liver gene knockdown in general and are now being pursued by a number of companies in preparations for the first systemic RNAi clinical trials. Unfortunately, however, their promise has also led to legal haggling as to who owns the IP behind this delivery technology. Companies involved in this dispute involve Protiva, Inex Pharmaceuticals, and Sirna Therapeutics/Merck and I hope that legal issues will not do further damage to the development of this promising delivery technology. No matter who owns the commercial rights to the technology, Protiva scientists have to be credited with this major achievement.

Interestingly, despite their publication record on ApoB100, Alnylam decided to target PCSK9 for the treatment of hypercholesterolemia. Although I cannot exclude that this move is partly due to a deal with one of their collaborators in siRNA delivery, PCSK9 has a lot riding for it. In fact, they are pursuing this program in collaboration with scientists from UT Southwestern Medical Center that arguably are world-leading in the genetics of hypercholesterolemia. PCSK9 itself is a protease that degrades LDL-receptors (LDL-R). LDL uptake by the liver is important for clearing LDL in circulation and it is expected that increasing LDL-R levels by suppressing their inhibitors should lower LDL cholesterol. Indeed, data presented at this year’s Keystone Meeting support this thesis. However, it should be kept in mind that many drug development projects fail, not because the drug fails to reach its target, but because of side-effects. Side-effects are a particularly important consideration for drugs that have to be taken chronically as is often the case for hypercholesterolemia. So one of the major questions here is whether long-term downregulation of PCSK9 can have adverse consequences. Here, the genetics of PCSK9 are compelling: Naturally occurring mutations in the human population that increase PCSK9 activity have been shown to increase LDL and lead to hypercholesterolemia, while those that inactivate it lower LDL dramatically- without any obvious detrimental consequences! Of course, compensation mechanisms cannot be excluded, but this is probably as good as you can get with choosing a target based on human genetics.

In summary, due to the availability of excellent “non-druggable” targets and the ability to knockdown genes in the liver with current delivery technologies, RNAi Therapeutics are a promising strategy for treating hypercholesterolemia. Alnylam is expected to initiate phase I studies in the second half of this year, and I would not be surprised to see further studies being initiated in the near future by Protiva or Inex (mere speculation here though). The major obstacle for these trials that I see are side-effects due to the liposomal formulations, and my advice would be to carefully characterise them in animal models before committing to phase I instead of simply bowing to investor expectations.