RNAi Delivery to Vascular Endothelium Increasingly Validated

Less than a month after Napoleone Ferrara from Genentech was recognized with a 2010 Lasker Prize for identifying VEGF as the central actor in blood vessel formation, a prize widely regarded as the stepping stone towards the Nobel Prize in Physiology or Medicine, a press release by Alnylam seems to suggest that RNAi delivery to the vascular endothelium is appropriately reaching critical mass. Before that, the most important body of work in this area probably came from Silence Therapeutics, which, despite its apparent quality, got me a bit worried given what I perceived as a certain lack of enthusiasm in the commercial RNAi Therapeutics space and, more importantly, third party scientific validation.

It is particularly exciting that the silencing of endothelial gene markers following a single dose persisted for two months. Since the duration of silencing is critically dependent on the cell type, for example its proliferation rate, this bodes very well for all delivery technologies targeting the vascular endothelium.

Alnylam mentions that the new LNPs (liposomal nanoparticles) that work for endothelial siRNA delivery stem from their collaboration with the MIT which as we know has involved positively charged ‘lipidoids’. Seen in light of the data by Silence (cationic lipid-siRNA)/Intradigm (RGD-targeted PEI polymers), the picture that is emerging is that LNPs that comprise positively charged lipids have a natural propensity of being taken up by vascular endothelial cells. Although I haven’t seen the details yet, it is to be expected that a number of parameters beyond positive charge, such as the method of formulating the particles, e.g. SNALP-like siRNA encapsulation versus Atuplex-like lipoplex formation, determine the efficiency of the functional uptake of these particles.

Beyond Alnylam and Silence Therapeutics, which with Atu-027 already has an endothelial cell-targeting RNAi Therapeutics in the clinic, Tekmira should also have considerable expertise in this area. Similar to optimizing LNP delivery to the liver, leadership in endothelial RNAi will depend on first empirically determining the structure-function relationships of these particles and then the biological pathway by which the uptake occurs. To my knowledge, it is still unclear whether receptor-mediated uptake, non-specific macropinocytosis ('cell drinking'), or the limited capacity of endothelial cells for phagocytosis is involved. This will also be important to understand as more targeted technologies will be developed.

One trade-off that current technologies might suffer from is that their positive charge at physiological pH may cause them to be slightly more toxic compared to negatively or neutrally charged formulations. This may also tie in with Phil Haworth’s comments in my last interview with him that Silence Therapeutics will initially focus on acute indications with high unmet needs. But again, everything is toxic at sufficiently high concentrations and it is encouraging that Atu-027 is still in its dose-escalating phase according to the last clinical update by the company.

Progress also reported for systemic RNAi delivery to immune cells

Following up on their declared pursuit of vaccine opportunities, Alnylam also highlighted progress in the systemic delivery of siRNAs to immune cells. Previously, LNPs comprising lipids derived from the KC2 series of next-generation ionizable lipids had been described to silence the immune cell marker CD45 with an ED50 of approximately 1mg/kg following intravenous administration (see RNAi delivery roundtable). The new report suggests some improvement over those formulations, stating that a 95% knockdown of CD45 was achieved, with an ED50 of as little as 0.2mg/kg. It will be very interesting to see the scientific details and discussions on these experiments (note: this last paragraph was corrected from an earlier version that mistakenly stated that the 95% knockdown was achieved with 0.2mg/kg).

Uptake of RNAi triggers by immune cells per se is not entirely new. It is a case of turning lemons into lemonades, as the non-specific uptake of nanoparticles by phagocytic cells has been long lamented. However, these cells play a central role in at least as many important diseases as endothelial cells do- so why not harness the efficient uptake of particulates in those cells for therapeutic purposes? Sure, being taken up and being functionally released into the cytoplasm are 2 separate issues, but Tekmira’s Ebola work has shown already that LNPs can trigger gene silencing in macrophages. Again, it will be important to delineate the functional uptake pathways and to employ chemistry to increase the efficiency with which LNPs can harness them.

Once thought of as liver-only formulations, LNPs are showing more and more their considerable versatility. I remember well a talk given by Tekmira’s CSO Ian MacLachlan at Stanford two years ago when he demonstrated that by even only slightly changing the formulation parameters, strikingly different patterns of biodistribution can be achieved. Now it ‘only’ takes one of the LNP/SNALP programs in the clinic to show some efficacy, and the pharmaceutical industry to change its approach towards drug target selection, for the liposome to finally shed its image as the unloved, but necessary stepchild of the industry.

Aptamer-siRNAs: Another Shot at RNAi Therapeutics Delivery

There has been a trickle of papers lately describing the use of aptamers for the functional delivery of siRNAs such as for cancer and HIV. Aptamers are highly folded, 35-100 nucleotide long RNAs that can bind protein targets with relatively high affinities and specificities. One way of thinking about them is as the RNA equivalent of antibodies. Aptamers already are being tested as a therapeutic class of its own where they are typically designed to neutralize extracellular targets, with already one aptamer (Macugen) approved for wet AMD.

As such, aptamers should lend themselves for targeting associated therapeutic siRNAs to cells of interest, in a sense functioning like antibodies and small molecules that have likewise been recruited for targeted RNAi delivery. What distinguishes an aptamer-siRNA combination, however, is the promise of having to simply use only RNA synthesis to generate a pharmacologically viable siRNA therapeutic, obviating the need for complicated formulation technologies. Furthermore, when it comes to repeat-administration such a system may cause inherently little adaptive immunogenicity.

The reason why I have been somewhat skeptical on this technology is that like with so many siRNA targeting approaches, getting to the cell of interest is just a first step, and it is not obvious to me how after e.g. receptor-mediated endocytosis the rather large aptamer-siRNA conjugate would be able to cross the negatively charged lipid bilayer to get into the cytoplasm for incorporation into the RNAi-induced silencing complex (RiSC).

Nevertheless, a recent study in Nature Biotechnology (Dassie and colleagues: “Systemic administration of optimized aptamer-siRNA chimeras promotes regression of PSMA-expressing tumors”) suggests that competitively low mg/kg dosages of intraperitoneally injected prostate-specific membrane antigen- (PSMA) targeted aptamer-siRNAs can efficiently knock down the popular cancer target PLK1 in a mouse xenograft model of prostate cancer. The study is a follow-up of a 2006 paper published in Nature Biotech by the same group from the University of Iowa (McNamara and colleagues: “Cell type-specific delivery of siRNAs with aptamer-siRNA chimeras”) where intratumorally injected- i.e. not systemically administered- PSMA-targeted aptamer-siRNAs showed very efficient inhibition of tumor growth in the same model system.

Since systemic administration is deemed to be necessary for an siRNA therapeutic against prostate cancer, the investigators reasoned that they could achieve such delivery by increasing the potency of the aptamer-siRNA by primarily improving siRNA potency through siRNA design (changing an initially blunt siRNA into a Tuschl-type 3’ overhang type) and then attaching the ubiquitous PEG to increase circulation times so that the aptamer-siRNA would have an increased chance of finding its target. As hoped for, both strategies substantially improved in vivo performance. Impressively, PEG addition increased the half-life of the molecule from less than 35 minutes to over 30 hours (!) and this was accompanied by improved silencing and tumor inhibition. The 3’ overhang siRNA (actually it was a Dicer substrate- more on this later) was also much better than the original blunt-ended version. While most aptamer-siRNAs are bi-molecular which reduces the maximum length of RNA to be synthesized, a unimolecular precursor microRNA mimic performed best. This could due to increased stability of an intramolecular duplex and/or a more “natural” appearance to the RNAi machinery. Practically, however, bimolecular conjugates may be preferable as RNA synthesis becomes exponentially less efficient with size and is also for this reason that the authors further reduced the length of the aptamer from the earlier study.

Overall, all of the many controls that they were probably asked for by the reviewers confirmed the specificity of the results: the therapeutic effect correlated very well with the degree of knockdown, both in vitro and in vivo; binding and silencing was only observed in PSMA-bearing cells; no innate immunostimulation that might explain the anti-cancer effect was detected; 5’ RACE showed that there was in vivo RNAi activity. Finally, only ~21nt siRNAs were detected following administration of the Dicer-substrate RNAi triggers which suggests highly efficient Dicer processing. Generally, it has to be said that while the shorter, traditional siRNAs have many advantages in terms of specificity and immunity, Dicer-substrates may be ideally suited for conjugate approaches such as this, as Dicer-processing would liberate and thereby activate the functional siRNA whereas Argonaute loading, in theory, should be diminished by a direct conjugate to the siRNA (however, strategies such as reversible disulfide bonds might work for such a configuration).

As an aside, the studies are further validation of PLK1 as a very good target for RNAi Therapeutics in oncology. PLK1 is one of the most highly over-expressed genes in cancer, and knockdown studies have shown that cancer cells are very sensitive to the reduction in PLK1 levels while normal/healthy cells, even if transfected with PLK1 siRNA are unaffected. PLK1 is also the target for a SNALP cancer therapeutic candidate developed by Tekmira for solid cancers that is slated for IND next year (Alnylam with a 50:50 opt-in right until start of phase II).

In a sign that there is also commercial interest in aptamer-siRNAs, the leading aptamer company Archemix and Dicer-substrate company Dicerna recently agreed to collaborate on aptamer-siRNA delivery. Archemix, which shares a building with Alnylam, similarly chose to collaborate with heart- and muscle-focussed miRagen on the delivery of microRNA therapeutics. Archemix’ sudden move into small RNA therapeutics is also quite interesting given their failed IPO attempt and speculations of a reverse takeover of Silence Therapeutics.

So where do I think aptamer-siRNA delivery technology stands? I’m still somewhat skeptical and would like to see more of these studies from various laboratories. An important question that was posed by an accompanying News and Views article from Alnylam scientists (which btw makes it very likely that the paper was reviewed by them) is whether the surprising cytosolic uptake of the RNA is a peculiarity of the PSMA antigen or could be a more widely mechanism for presumably endosomal escape that could be exploited. Studies into the precise molecular mechanism of the uptake, as with all RNAi delivery systems, are needed. One could also imagine that to enhance uptake, membrane-active agents may be added to the PEG-aptamer-siRNA, although this would be contrary to the initial concept of a simple design. In summary, the more varied approaches being explored, the better for RNAi Therapeutics. For now, aptamer-siRNAs are just one of those to be watched.

RNAi Therapeutics delivery: not a zero-sum game

Shares of Tekmira briefly fell last week below 80c, a ~20% decline, following Roche’s purchase of fellow RNAi Therapeutics delivery company Mirus Bio. Initially, I did not draw a connection. If a reaction at all, then I would have expected the $125M price-tag to point out the value attached by the industry to promising RNAi delivery technologies in general, which “should have” resulted in a share price increase for Tekmira.

To my surprise, however, it then transpired that the Roche acquisition had been interpreted by some as a vote for DPC and against SNALP RNAi delivery, leading some to sell in a panic. This blog will briefly outline why, unlike is the case in the battle for core RNAi trigger IP, RNAi Therapeutics delivery is not a zero-sum game with room for more than just a handful of technology platforms.

As is becoming more evident by the day, RNAi Therapeutics offers the opportunity to address a very wide range of diseases. As all disease has a genetic element, regulating gene expression should always be able to modify or even correct a disease. To do this, however, one has to be able to knock down the desired genes in various different organs, tissues, and cell types. If there was a universal RNAi delivery technology, it would have to be able to physically contact every cell in the body, yet only knock down genes in the disease-related subset of cells.

Based on the limited publicly avaibable data, Mirus’ DPC technology actually promises to come close to that dream. It is very small in size (~20nm) and therefore satisfies at least one of the pre-requisites of achieving a broad biodistribution. Moreover, it has been shown to be capable of differentiating between the different cell types within an organ, in this case the liver, depending on which sugar had been added- all in the absence of apparent toxicity! It will clearly be exciting to learn more about the delivery profile of DPCs. Nevertheless, I doubt that DPC will necessarily be the delivery system of choice for every indication.

This is because the choice of a delivery system is not only determined by the ability to knock down a gene in a given cell, but also by factors such as overall maturity of the technology, safety as determined by the dose required to achieve such gene knockdown, the biodistribution and cell type-specific gene knockdown on a systemic level, as well as by cost, route of administration, and stability of the formulation.

Consider for example an RNAi Therapeutics against the kinesin spindle protein (KSP) that is being developed by Alnylam as a treatment for liver cancer. Although directly interfering with cell division by down-regulating KSP is seen as a very promising anti-cancer strategy, anti-KSP small molecule programs have indicated considerable dose-limiting toxicities that appear to be target specific. This, however, should not come as that much of a surprise since interfering with spindle function would be predicted to affect normally proliferating cells as well, particularly those of the hematopoietic system. Well, this dilemma is not new at all to the cancer field, and often the benefits outweigh the side-effects, but this example illustrates the value of a delivery technology such as SNALP-RNAi that can be tuned to deliver around 95% of the injected dose to the liver, including hepatic tumor tissue.

Certainly, Roche may elect DPC delivery over SNALP delivery (note: Roche has access to both) for some liver indications, but that type of competition has actually only decreased with Mirus Bio out of play. This is because all the other two dozen or so pharmaceutical companies equally interested in RNAi Therapeutics, will now have access to one less viable delivery technology. The incentive to gain access to the still fairly accessible RNAi delivery platforms with clinical potential has therefore never been greater. Roche, of course, enjoys the luxury of being able to choose between two technologies that, maybe in an effort to avoid the appearance of favoritism, it has valued essentially identically based on their $5M equity investment in Tekmira at $2.4 per share(Tekmira now has 51.6M shares outstanding).

Bottom line, due to the ever growing attraction of RNAi Therapeutics, any company, and that goes beyond Tekmira-SNALP, with clinically relevant RNAi delivery technology and a good IP package to protect it can rest assured that they own a very hot commodity.

RNAi Therapeutics portfolio update: With Roche’s validation of the value attributed to clinically relevant RNAi delivery technologies and with increased clinical visibility, I have slightly increased my position in Tekmira while paring back on the overweight ALNY position. In another attempt to capitalize on the idiosyncracies of the stock market, I have converted the RXi Pharmaceutical holdings back into CytRx shares (CytRx owns about 50% of RXi, but none of that is reflected in CytRx’s market cap) and added some more at the cost of largely cashing out of RNAi trigger competitor Silence Therapeutics (a token $1 investment is maintained to monitor stock performance).

Disclosure: Long Tekmira, Alnylam, Targeted Genetics, Oxford Biomedica.