One Pitfall to Look Out for in RNAi Therapeutics

RNAi Therapeutics have turned the corner to being a generally accepted drug modality.  Following years of ridicule and benign neglect, the industry just celebrated the FDA approval of a 2nd drug, Givlaari for the treatment of acute hepatic porphyria. This was followed in short order by the $9.7B acquisition of The Medicines Company by Novartis for its cholesterol-lowering RNAi drug candidate, Inclisiran, a harbinger for how RNAi medicines will change the game in cardiovascular risk management.

Amid all the understandable investor excitement, it is probably a good time to reflect on the potential pitfalls as we move into a world where RNAi Therapeutics dominate liver-related diseases and advance into new big areas such as the CNS and ocular disease.  Remember, investor, it takes one safety event for the most highly valued programs like HBV RNAi candidate AB-729 by Arbutus Biopharma, Arrowhead Pharmaceutical’s ARO-AAT, or the Lpa program by Silence Therapeutics to decimate your investment by half or more.

Off-targeting safety, off-targeting safety, off-targeting safety

When the discovery of RNAi was rewarded with the Nobel Prize in 2006, the Andy Fire and Craig Mello hammered home a message that would resonate for the years to come: to make RNAi Therapeutics, it’s all about the delivery!

Delivery, delivery, delivery

But as the liver has been conquered and the industry is chewing on a number of liver-related therapeutic opportunities and opening up opportunities in new areas, we still have to fully understand the odds of seeing program failures due to the difficult to predict consequences of off-targeting.

Off-targeting is the moderate knockdown of sometimes dozens of genes sharing sequence homology to the target gene.  This was shown over a decade ago and soon thereafter shown to be addressed by using simple modification strategies.

This, however, was at a time when the delivery mantra was practiced and nobody wanted to risk reducing knockdown efficacy by simultaneously considering off-targeting in their RNAi trigger designs.  I was therefore upset, as it turned out rightfully so, that Alnylam let Nastech essentially go bankrupt a decade ago when it should have acquired rights to Nastech's then industry-leading off-target-mitigation technology.

Alpha-1 Antitrypsin as testbed

Alnylam would have to pay dearly for its pride in never treating a smaller competitor well, even if it served its own interest, with off-targeting manifesting itself in markers of liver toxicity in first-generation clinical programs like chronic HBV and alpha-1 antitrypsin-related liver disease.

These two indications, unlike the now approved AHP indication for Givlaari which addresses symptoms outside of the liver, represent the most exacting off-targeting challenge as they affect liver health directly and any further exacerbation of liver stress should be avoided.  

As a result, Alnylam dug out the old off-targeting literature and in their incarnation added a GNA modification at position 7 of the guide strand while, for experimental purposes, leaving the rest of the first-generation molecule unchanged. Translational science at its best (!) and the industry can be thankful for it.

Indeed, for both the ESC+ chemistry ALN-AAT02 and ALN-HBV02 no liver enzyme elevations (above 3x ULN) were observed in 49 subjects tested, whereas 3 of 33 had such elevations with the ESC-only versions.

In Arrowhead’s case, there also were no ALT elevations >3x ULN.  However, 3 treatment-emergent grade 1 ALT elevations not exceeding 2x ULN were reported in the 28 normal healthy volunteers receiving ARO-AAT.  I am curious to see the individual time curves and the doses at which they occurred.  It is certainly an area to watch for that program as it only now moves into patients. 

Add to this considerable uncertainty about patient selection, dose and duration required to have an impact on AAT liver disease, increasingly analysts topping their price targets by the day may have to re-think how they value this program (see disclosure below).

Further behind, we have to see how Silence Therapeutics and Arbutus have addressed off-targeting with first data expected in 2020.  While Silence has avoided indications directly affecting liver health (iron overload, cardiovascular disease), Arbutus with its HBV indication will have a steeper hill to climb, although the pay-off for investors in this $80 market cap company should be considerable should it climb it come Q1 2020.

Personally, my guesstimate for non-optimized GalNAc designs would be a 50% chance of off-targeting success, also depending on preclinical vetting (after all, the very clean $9.7B Inclisiran is such a design); a 85% off-targeting success for optimized designs. 

Disclosures: long Silence Therapeutics, Arbutus Biopharma; short Arrowhead Pharmaceuticals; no position in Alnylam.

mdRNA and RXi Coming Out with their RNAi Triggers (Part I)

In conference presentations this week, both mdRNA and RXi Pharmaceuticals finally lifted the veil of mystery shrouding the nature and performance of their claimed proprietary RNAi trigger designs. Independent of what they may mean in terms of IP, they are certainly worthy contributions to the science of RNAi triggers: the use of unlocked nucleic acids to increase the specificity of RNAi gene silencing (mdRNA), and the convergence of delivery and RNAi triggers through nucleic acid chemistry (RXi).

At the ‘RNAi and miRNA World Congress’, mdRNA presented data illustrating how selectively spiking siRNAs of the traditional Tuschl structure with a number of unlocked nucleic acid residues (usiRNAs) can reduce the level of off-targeting commonly provoked by ‘naked’ siRNAs. While similar strategies have been reported before and are broadly attempted in the industry, what differentiates mdRNA’s approach is that this particular nucleic acid analogue does not involve the addition of modifying chemical groups. Instead, in an unlocked nucleic acid the ribose ring is simply broken without making the residue any bulkier, but this is enough to significantly affect the interactions of the siRNA with its target mRNA and the Argonaute protein, the effector protein of the RNAi silencing pathway.

The simplest, and therefore widely practiced first step in reducing off-targeting is to modify the 5’ end of the passenger strand with a variety of chemistries. Because the 5’ end is important for the Argonaute protein to recognize and bind an RNA as the guide strand, half the off-targeting can be eliminated based on the sensitivity of this interaction to chemical alterations. It so happens that a UNA at the 5’ end of the passenger strand is similarly not tolerated…as has in effect recently also been shown in a publication by Sirna Therapeutics/Merck (Kenski et al., 2010). In the example provided by mdRNA at the conference, this meant a reduction in the number of genes that were changed in their expression by a factor of more than 2-fold (generally 2-3 fold) from ~390 to ~ 180, while on-target gene silencing activity was maintained at around 95% (data from RNA microarray experiments).

I should add, however, that additional usiRNA residues in the 3’ overhangs of both passenger and guide strands somewhat complicate determining the exact contribution of the passenger strand 5’ modification in reducing off-targeting. This is because it is known from structural studies that the position of the 3’ end of the guide strand relative to the Argonaute protein changes upon the transition from the microRNA-off-target (in the so called ‘PAZ pocket’) to the on-target cleavage (dislodged from the ‘PAZ pocket’) confirmations. 3’ ends that do not fit into the PAZ pocket may therefore change the ease with which the enzyme adopts either conformation and therefore affect the propensity of microRNA-like off-targeting. Structural studies with UNA residues at the 3’ end may be instructive here.

More interesting is mdRNA’s observation that when, on top of the above modifications, one normal base is replaced with its corresponding UNA analogue in the seed (= the first 2-8 nucleotides), the off-target activity is further reduced to about 35 genes that are changed by at least 2-fold, while again maintaining on-target efficacy. This was theorized to be due to the changes in guide RNA-target mRNA hybridization energies within the seed: the usiRNA lowered this energy sufficiently such that a microRNA-like interaction that is reliant on the seed becomes prohibitively weakened. However, the extended length of the guide RNA-target mRNA base-pairing compensates for the lost energy thus still allowing for Argonaute conformational change and on-target cleavage to ensue.

Previously described modifications have similarly been thought to function in reducing microRNA-like off-targeting by modulating seed energies. usiRNAs, however, could be a particularly attractive approach, because its small size should make it less likely to be detrimental to on-target activity. This, of course, would have to be shown with more examples, preferably in a peer-reviewed publication, and the Sirna Therapeutics paper has shown at least some limitations in where the UNA can be placed in the seed. Similarly, it remains to be demonstrated how much of the reduction in the off-target activity was due to eliminating microRNA-like interactions instead of a reduction in the innate immunostimulatory potential of these siRNAs as UNAs were also shown by mdRNA to mitigate such responses.

I believe this story shows how RNAi trigger design has matured to a point that RNAi Therapeutics truly lives up to its original promise of specificity which had temporarily (2003-6 period) been tarnished by the discovery of wide-spread microRNA-like off-targeting and the innate immunostimulatory potential of exogenously delivered synthetic siRNAs. It also demonstrates that at least on the level of RNAi trigger design mdRNA could certainly be a valuable partner to most in Big Pharma.

In my next entry (link here), I will provide my thoughts about RXi Pharmaceutical’s ‘self-delivering rxRNAs’ for which, until now to the best of my knowledge, the company has created so much expectations without actually disclosing the chemical nature thereof.

Journal Club: mRNA Levels are a Good Indicator of Off-targeting

Next to delivery and immune stimulation, sequence-specific off-targeting is considered one of the fundamental challenges in the development of RNAi Therapeutics. Such off-targeting largely occurs by microRNA-like gene silencing following recognition of off-target RNAs that have limited sequence complementarity to the guide strand of the siRNA. Since microRNA gene silencing had long been believed to inhibit translation without affecting mRNA levels, there has been concern that the true extent of RNAi off-targeting may be larger than originally reported in studies assessing changes in global RNA levels. By achieving the feat of simultaneously measuring microRNA-induced changes in thousands of proteins, two high-quality papers now find that, surprisingly, mRNA levels are indeed a very good indicator of off-targeting, validating the use of established high-throughput RNA assays during the siRNA selection stage of RNAi Therapeutics development.

Key to the success of the Bartel lab at MIT (Baek et al.: The impact of microRNAs on protein output) and the Rajewski lab in Berlin (Selbach et al.: Wide-spread changes in protein synthesis induced by microRNAs), was their use of the SILAC technique (Stable Isotope Labeling with Amino acids in Cell culture). In this procedure, newly synthesized proteins are labeled with amino acids containing different isotopes. To determine the impact of microRNAs on global protein synthesis (~5000 proteins were identified in the studies), the samples with and without a given small duplex microRNA were labeled with different isotopes. This results in peptides from the same protein exhibiting slightly shifted peaks during protein mass-spectrometry, and the ratio of the peak intensities belonging to a pair of peptides is taken as a measure of microRNA-mediated gene silencing (1= no change). Changes at the protein level thus determined were then compared to changes on the mRNA level by microarray analysis.

As expected, the authors found that among the down-regulated proteins were highly enriched those that contained microRNA seed target sites in their 3’ untranslated regions (UTRs). Consistent with our notion of microRNAs as regulators of a group of genes rather than individual genes, typically 200-500 proteins were changed by more than 30% following the introduction or deletion of a microRNA, with about half of them likely the consequence of a direct interaction with the microRNA. Remarkably, despite the introduction of rather large amounts of microRNAs into the tissue culture cells (25-100nM using lipofection), most of these changes were rather subtle, and very rarely were changes of more than 3-fold detected. While it is always possible that wide-spread subtle changes may lead to a directed cellular response or cause toxicity in the case of RNAi Therapeutics, this observation is in agreement with microRNAs often acting as rheostats to maintain cellular homeostasis and suggests that the introduction of random short duplex RNAs should not necessarily lead to gene regulatory chaos and distinct phenotypes.

The surprise came when it was found that changes in protein levels could for the most part be accounted by changes in mRNA levels. Only very few proteins were reduced without concomitant mRNA changes. Without going into the details, the studies nevertheless support a model in which translational repression is the primary mechanism of microRNA-mediated gene silencing and mRNA destabilization a consequence thereof. Consistent with this hypothesis,
in the rare instances where 2 microRNAs were located close to each other, silencing was synergistic, yet the mRNA level was not further reduced. This is quite interesting as it may explain how the early use of microRNA reporter systems in which multiple microRNA target sites were closely spaced and which consequently caused strong gene suppression could have led to the notion of translational silencing only. The 20-30% reduction in mRNA could have easily been dismissed as experimental noise.

For practical purposes, however, measuring mRNA levels are highly predictive of proteomic changes and this should allow for the selection of RNAi Therapeutics candidates that have an increased likelihood of not causing toxic changes in gene expression. Merck has long been a proponent of this approach and partly for this reason have closely aligned the systems biology expertise of subsidiary Rosetta Inpharmatics with their RNAi incubator Sirna Therapeutics, in addition to employing systems biology for identifying the right gene targets.

It is very comforting to know that the extent and severity of off-targeting on the protein level is not significantly greater than what we have observed at the RNA level. Moreover, the ability to use high-throughput RNA expression technologies for assessing off-targeting and chemical modification techniques which have proven to significantly further reduce off-targeting potential of siRNAs should minimize the risk of RNAi Therapeutics candidates to fail due to sequence-specific off-targeting. A good siRNA at high concentrations in tissue culture reduces about 5-10 off-target RNAs by more than 2-fold (and as we know now proteins to a similar degree, not more), while the target is the most suppressed gene. By acting at the RNA level, RNAi Therapeutics has the distinct safety advantage in that it allows for off-targeting to be assessed in such a comprehensive manner. Before investing $1 billion dollars in the development of a drug, it appears to be well worth taking advantage of this by trying to understand these changes.

Alternate Use of Two SiRNAs for Chronic RNAi Therapeutics

First, I wanted to write about the pros and cons of developing RNAi Therapeutics as antivirals, or the use of dual-acting siRNA agents, but then decided to sacrifice what remains of my credibility and propose a somewhat unorthodox method to manage off-targeting in RNAi Therapeutics for chronic medication.

As discussed before, both bioinformatics and gene profiling technologies can be applied in the selection of candidate siRNAs with the least likely adverse off-targeting effects. Bioinformatics may avoid sequence elements that occur at high frequencies in mRNAs or in a particular unwanted genes, while microarray profiling may experimentally pick out those genes which are downregulated at least on the RNA level which, despite some caveats, is still a pretty good predictor of off-targeting (note that potentially even more predictive protein mass-spec techniques may also be applied soon for off-targeting). In that manner, it is possible for Merck, for example, to screen out those siRNAs which off-target numerous genes or those genes that you would rather leave untouched. Finally, chemical siRNA modifications may allow for the general reduction in the number off-target interactions.

Still, with all of this in hand, off-target risk remains. Due to the redundancy of the genetic code, it should be generally easier to translate into humans off-target phenotypes from preclinical models for protein-targeting drugs than for RNA knockdown therapeutics. That means, a given drug is more likely to still interact with the same homologous protein in different species because the overall function and therefore usually shape of a protein is more conserved than the underlying mRNA sequence, while even a slight nucleic acid change at even a single position may decide whether an mRNA is an off-target for an siRNA or not at all. So what may be true for off-targeting even in a non-human primate may look quite different in man.

Off-targeting should be a concern particularly for chronic applications of RNAi Therapeutics. A 50% reduction in an important gene due to an siRNA for hypercholesterolemia for example may eventually manifest itself as an off-target toxicity following prolonged drug use. While this is true for essentially any class of drugs, RNAi Therapeutics may have an advantage. Because siRNAs as a class are quite homogenous in terms of pharmacological properties, how about entering two siRNAs targeting the same gene into the clinic, as a single drug development program, but whereby the siRNAs are given alternately?

This would be somewhat similar to how siRNA cocktails are considered for cancer and antiviral RNAi, but with a different rationale, or the idea behind Dharmacon’s smartpools. Just change the siRNA and leave the rest the same, including the delivery formulation. Similar to food where it’s better to consume a varied diet than constantly eat a single healthy dish, that way the siRNA would be changed before an off-target had the chance to manifest itself without having to take a drug holiday. That way, the fact that, unlike protein-targeting therapeutics, two siRNAs even for the same target have completely different off-target spectra, could be turned into an advantage.

I’m not sure how well this would go down with the increasingly defensive FDA, but maybe they could make a start and be more open-minded when it comes to new technologies.

PS: Another advantage of RNA knockdown therapeutics versus protein therapeutics is that it’s much easier to measure changes in RNA and protein abundance (RNA therapeutics) than determine drug-host interactions and their potential consequences, which in the case of RNA therapeutics, as a class, should be quite predictable while they should differ much more widely in the case of say small molecules.

Peter Linsley Leaves Merck to Become Chief Scientific Officer of Regulus Therapeutics

When the Regulus CEO, during Alnylam’s quarterly conference call last week, expressed his satisfaction about their ability to attract top talent to the microRNA therapeutics start-up, Peter Linsley, who it was announced today that he would become the CSO of Regulus, must have been foremost on his mind.

Peter Linsley is the perfect CSO of Regulus in many respects. His group at Merck’s bioinformatics subsidiary Rosetta Inpharmatics was the one who seemed to abruptly end dreams of a perfectly on-target RNAi Therapeutics platform when they revealed in a 2003 paper widespread off-targeting by standard siRNAs using expression profiling (Jackson et al. 2003 “Expression profiling reveals off-target gene regulation by RNAi”). After extensive follow-up work by his group and others, we now know that much of this off-targeting is related to unmodified siRNAs being able to recognize targets based on limited base-pair complementarity in a manner essentially identical to how microRNAs recognize their targets. As such, he will bring with him to Regulus the knowledge and tools needed to predict and characterize the consequences of mimicking and inhibiting microRNAs that Regulus aims to exploit therapeutically. In addition to his bioinformatics background and experience with high-throughput technologies, he has acquired extensive knowledge on the biology of microRNAs through numerous co-authored publications.

Linsley’s know-how may also benefit the parent companies, particularly Alnylam, since his work also involved applying strategies to minimize the microRNA-like behavior of siRNAs, such as by modifying them, so that their activity would be largely limited to targets with perfect complementarity (Jackson et al. 2006 “Position-specific chemical modification of siRNAs reduces “off-target” transcript silencing.”), therefore rendering siRNA therapeutics potentially safer. He should also bring with him expertise in identifying transcript profiles following microRNA and siRNA delivery that either indicate therapeutic promise or, even more importantly, potential harm.

Linsley’s decision to leave Merck further highlights the difficulty of Big Pharma in retaining top talent after acquiring biotech companies, particularly due to a bureaucracy that appears to be so pervasive in today’s Big Pharma and is at risk of choking innovation. Rosetta Inpharmatics and Sirna Therapeutics were part of Merck’s strategy of becoming a powerhouse in RNA therapeutics by combining leading expertise in gene regulatory networks/systems biology with oligonucleotide technologies. It now seems that much of the human capital that made these companies so attractive in the first place have left, and this, together with Alnylam’s decision last fall to terminate their RNAi Therapeutics alliance with Merck, is likely to add to the criticism that Merck overpaid for and mismanaged their acquisitions and is now at risk of losing their leadership position by insisting on going it alone.

This contrasts with Roche’s decision to let Alnylam’s former European subsidiary to operate as a semi-independent “Center of Excellence”, and is echoed by the intention of Pfizer and other Big Pharma and mature biotechs to establish similar center of innovations separate from headquarters. Of course, it is Roche’s success story with Genentech that serves as an example for these attempts.

The most remarkable aspect of today’s announcement to me, however, is the fact that Linsley personifies that by embracing complexity, what has started as a worrisome discovery 5 years ago has now turned into a therapeutic opportunity- the microRNA therapeutics opportunity.

3rd Day of the OTS Annual Meeting: Novartis Demonstrates Anticancer Activity of Orally-administered siRNAs

The third and last day of the conference was another action-packed day with topics ranging from the basic science of RNAi to its clinical applications.

To better understand endogenous microRNA function and siRNA off-targeting, transcript analysis such as high-content microarrays has been widely used. This, however, captures only the changes in RNA levels, which is a problem since microRNAs and off-targeting have been thought to frequently act at the translational level with little changes in transcript abundance. Transcriptional analysis should therefore miss important microRNA targets and siRNA off-targets. Nikolaus Rajewski introduced an intriguing solution to this problem by harnessing a labelling technique coupled to mass-spectrometric analysis that allows changes in protein levels as a consequence of small RNA over-expression or inhibition to be measured in high-throughput. Preliminary results confirm that this technique indeed retrieves those targets that have been down-regulated both on the RNA and protein levels, but also discovers those genes where only the final protein output has been altered. If sufficiently robust, this technology should ultimately facilitate the development of, most importantly, safer, but also more potent RNAi medicines.

Arndt Borkhardt from Duesseldorf, one of the first to publish applications of RNAi in humans in 2002, presented an intriguing correlation between trisomy 21 (Down Syndrome)-related childhood leukaemia and over-expression of microRNAs encoded on chromosome 21. Importantly, only leukaemic cells, but not normal tissues over-express these microRNAs. Using zebrafish as a model system, he was able to show that the inhibition of these microRNAs indeed caused haematological abnormalities, corroborating a causal contribution of these microRNAs to leukaemia. It not strongly supports the rational of using microRNAs as cancer diagnostics, but also suggest these microRNAs to be valid targets for the treatment of childhood leukaemia. Continuing along the lines of microRNAs as therapeutic targets, Hermona Soreq from Israel made a compelling case that miR-132 and miR-182* (star) are important feedback regulators of the cholinergic anti-inflammatory response and may be targeted for modulating inflammatory responses in disease. Her geographical proximity to Rosetta Genomics makes me wonder whether this company will or has already become involved in thinking about clinical applications centered around miR-132 and 182*.

Jost Seibler represented Artemis, an Exelixis subsidiary specialising on generating mouse RNAi transgenic disease models. They have developed technologies that allows them to use constitutive and conditional (tet technology) RNAi transgenics in a matter of around 4 months, which far exceeds timelines for producing knockout animals. This makes these animal attractive target validation tools. Importantly, a number of examples demonstrated that essentially knockout phenotypes can thus be generated with U6 and H1 promoters driving shRNAs. It was also comforting to see that the shRNA hairpins worked in essentially all cell types and tissues with weaker knockdown only in B-cells and the spleen in general. It will be interesting to see whether other RNAi inducers work more efficiently in these cells and whether this is the consequence of some type of selective disadvantage on cells expressing high levels of small RNAs. 34 constitutive, and 48 conditional knockout models have already been successfully generated.

Volker Patzel definitely raised some eye-brows when he reported that specifically siRNA, and not other nucleic acid types, triggered gene silencing in a variety of prokaryotes. Complicating things was that the mechanism of action ranged from transient to persistent knockdown and even gene knockout effects, and that a variety of parameters such as cell transfection competency and host factors made the technique quite tricky. John Rossi certainly made a good point suggesting that bacterial genetics is most likely to uncover the mechanism of the observed silencing effects.

John Rossi himself was up next providing an update on the lentiviral RNA trials for HIV/AIDS that got under way earlier this year at the City of Hope (a collaboration with Benitec). This trial involves a lentivirus expressing 3 types of RNA therapeutics, an RNAi shRNA hairpin against an HIV gene, an HIV-TAR decoy, and a ribozyme directed against the host factor CCR5. It is a particularly important trial since this is the first time in gene therapy that not only a lentivirus is used as a vector, but also DNA-directed shRNAs are employed. Preclinical parameters such as safety and efficacy as measured by vector persistence, expression, HIV knockdown, and the ability of the pluripotent precursor blood cells to normally differentiate after lentiviral transduction were very encouraging. Unfortunately, his hopes to present data at this meeting from treating the first of three enrolled phase I HIV leukaemia patients were dashed due to 65% cell viability after a thawing step in the protocol, when at least 70% was the target. The FDA would have still allowed treatment to continue, but the clinical investigators felt they should not move the goal posts, and it is hoped that the 2nd and 3rd patient cells will fare better, so that they can be re-infused with the HIV-immunised blood progenitor cells. In the second half of his talk, Rossi detailed experiments inspired by a recent report from Bruce Sullenger’s group on successfully using RNA aptamers for targeting siRNAs in vivo. This is an interesting targeting approach since both the targeting agent as well as the therapeutic moiety derive from the same chemical class. As an alternative to the above lentiviral strategy, he would now like to evaluate such aptamer-siRNA RNA hybrids to target anti-HIV siRNAs systemically to infected cells with an gp120 aptamer developed by William James from Oxford University.

Arguably the highlight of the day was a presentation by Francois Watt from Novartis demonstrating for the first time sequence-specific anticancer activity of ORALLY administered siRNAs. These experiments were motivated by preceding mouse cancer model results using local and introperitoneal administration of siRNAs targeting the angiogenesis factors VEGF-R2 and Tie-2. These siRNAs were stabilised with an undisclosed modification at the 3’end for exonucleolytic stabilisation of the siRNA, which was confirmed by incubating them in gastric acids and intestinal lavage as a prelude to the oral delivery experiments. I wonder whether the 3’ end modification also facilitated the presumable uptake of the siRNAs. Clearly, it will be important to quantify the bioavailability of the otherwise unformulated siRNAs. Relatively large amounts of siRNAs had to be administered (100mg/kg), but I expect this to come down with more experience on oral delivery. Since Novartis had been involved in antisense development before their interest in RNAi, it can be speculated that their 3’modification had been informed by previous antisense results and may be covered by relevant modification IP. For all you Alnylam watchers, such positive data augurs well for Novartis exercising the Alnylam adoption license. Overall, my impression is that while companies can be very secretive with their corporate strategies, scientific data are often more believable since the pressure for them to publish glowing results in leading academic journals is less than for academic researchers. Also, they should have little motivation to further pour money into corporate partnerships if the science did not support it.

Francois Watt was a difficult act to follow, but Jeremy Heidel from Calando did a good job by detailing their development of cyclodextrin-based nanoparticles for delivery of siRNAs in cancer applications. Much of the data had already been published, but it was still a useful summary of their preclinical efficacy and tox studies for their first clinical program targeting RRM2 with transferrin-decorated nanoparticles. While transferrin did not affect the biodistribution of these particles, it enhanced siRNA uptake presumably simply by bringing the siRNA into proximity of the target cell which may then take up the siRNAs by transferring-independent mechanisms, a theme similar to the findings by Stoffel and Alnylam (recent Nature Biotech paper) and others. Importantly, these nanoparticles are bioactive at low mg/kg concentrations and can be safely repeat administered up to 27mg/kg. The only immune response observed was directed against the human transferrin component of the nanoparticles in mice, and importantly not the PEG component. This response against the increasingly popular transferrin should therefore not be observed in humans. Everything appears to be now in place for Calando to start their first systemic RNAi therapeutics trial.

Of course, one has to be aware that in the past many promising preclinical results in mouse models of cancer were not held up in human studies. This, however, is more a problem of choosing appropriate model systems and marrying them to the right targets, less a problem of the therapeutic platform per se. I am hopeful that, once efficient targeting is achieved, RNAi therapeutics will have a higher success rate in cancer than other treatment modalities since the choice of the target is not limited by the molecular structure of the protein.

Posters- an interesting poster on flu (influenza A) RNAi was presented by the infectious disease group at the MPI in Berlin showing that LNA-stabilised, but not unmodified siRNAs, when formulated with PEI caused potent knockdown of flu viral titers in a mouse model following intravenous administration. Santaris, the LNA antisense company had a poster dedicated to their ApoB100 antisense program, showing that 12mer LNA gapmers are better at suppressing ApoB100 than longer gapmers. However, I was puzzled by their findings that not only was LDL potently reduced, but that HDL was also suppressed.

Lastly, I wanted to find out more about the rationale for BMS to partner with ISIS for targeting PCSK9 by antisense, when their poster presentation on targeting PCSK9 by RNAi was so positive. I was taught that Big Pharma may use different techniques for their target validation and that the ultimate choice of the targeting technology was determined by other factors. Apparently, the greater experience with systemically administered antisense compounds in the clinic compared to RNAi and the need for formulating siRNAs for efficient delivery to the liver, trumped the need for considerably higher doses of antisense oligos to achieve equivalent knockdown potencies. Ultimately, I came away with the impression that while BMS saw the promise of PCSK9 in the treatment of hypercholesterolemia, Alnylam was just not willing to partner their PCSK9 program at this early stage. In any case, I wish BMS and ISIS well on developing PCSK9-targeted therapies and it would not be unprecedented if two different technologies targeting the same gene will ultimately come to market. I expect, however, that the therapy with the better safety profile will ultimately dominate in the market. While much is being made about the off-targeting potential of siRNAs, it is remarkable that 3 days of observing RNAi next to other nucleic acid-based therapies, demonstrated that while the RNAi field tackles all sorts of potential safety considerations head-on, off-target profiling and the study on the induction of cytokine responses were essentially missing in the presentations on other RNA-targeting technologies.

The 3rd Annual Meeting of the Oligonucleotide Society was a full success. Many thanks to the organisers and presenters for this!!! This is a time when the right ingredients to making oligonculeotide therapies are clearly coming together, and this should become even more evident at the 4th Meeting to be held next year in Boston.

Impressions from the 3rd Annual Meeting of the Oligonucleotide Therapeutics Society: Day 2

The second day of the Meeting was filled with cutting-edge science related to RNAi drug development with emphasis on RNAi mechanism, delivery, and cytokine responses triggered by various nucleic acid classes.

Tuschl kicked off the day by presenting data on the systemic identification of Argonaute interaction partners (mostly published data) and a new approach towards identifying microRNA targets. In contrast to microRNA target identification algorithms which heavily rely on sequence conservation, Tuschl and colleagues identified RNAs pulled down in Argonaute immuno-precipitations and then sequencing. Validating the approach, these RNAs were enriched for sequences with seed targets of the most abundantly expressed microRNAs in the tested cell line (1.8x enrichment over random). He finally reported on small molecule screens aimed at identifying inhibitors of microRNA maturation, however, with less success. It appears to me that he should stick to his guns and work at inhibiting microRNAs and their precursors instead by nucleic acids, such as antagomirs, which have proven to be much more potent and specific inhibitors of microRNA activity.

Ingo Roehl (Roche Gmbh, former Alnylam Europe) gave a nice presentation on progress in the development of analytical chemistries to support DMPK/PD (Drug Metabolism and Pharmacokinetics/dynamics) studies of RNAi Therapeutics. It was encouraging to see significant improvements that now allow siRNA and endogenous microRNA quantitations in the picomole-femtomole range in biological samples at high-throughput. In contrast to methods used by other groups, their coupled HPLC mass-spec setup allows for the discrimination between intact siRNAs and their degradation products. However, further improvements in sensitivity are desired as the potency of siRNAs means that biological activity can be observed long after siRNA levels drop below the level of quantitation.

Peter Linsley from Rosetta Inpharmatics (a subsidiary of Merck), which gained early “notoriety” for being the first group to point out the problem of off-targeting by siRNAs some years ago, highlighted the dilemma posed by the fact that the off-target signature of a given siRNA is extremely different in mouse and man. This means that mice will not function as a safety model for off-target toxicity studies. Furthermore, by limiting oneself to cross-species specific siRNAs for the sake confirming the treatment in the animal model, many siRNAs with better potencies and off-targeting signatures in humans may ultimately be missed. Also of interest is the fact that by leveraging its genomic expression profiling capabilities, Rosetta Inpharmatics has now established microRNA overexpression signatures for over 150 microRNAs, raising the question when Merck will officially announce its entry into the microRNA therapeutics arena.

Following on from his identification of SID-1 in being required for systemic RNAi in worms by acting as an siRNA channel, Craig Hunter (Harvard University) presented data to support a model in which siRNAs enter and exit cells via SID-1 by diffusion, whereas other proteins such as SID-2 are involved in binding RNAs so as to increase their concentration close to the siRNA channel. That this is relevant for the development of RNAi therapeutics is highlighted by the recent Nature Biotech paper by Dr. Stoffel from the ETH Zurich and Alnylam where data suggested a role for SID-1 in taking up siRNAs following docking of siRNA-studded lipoprotein particles to their cellular receptors. In addition to walking his audience the published data-heavy paper, Stoffel stated that he saw no reason why it should not be possible to create artificial lipoprotein particles containing siRNAs and targeting agents.

Instead of Dinah Sah or David Bumcrot presenting Alnylam’s progress on cancer RNAi therapeutics, Rachel Myers stood in to give a general overview of the approach Alnylam takes to RNAi-based drug development. Since little new primary data was presented, I thought it is noteworthy that both Ingo Roehl and Myers gave scientific credit for the development of the hotly contested SNALP development to Protiva Biotherapeutics and that the delivery technology used in last week's Nature paper on microRNA competition by siRNAs was called "Alnylam proprietary”. Also, it was a great relief for me to hear that Alnylam had reproduced all of Sailen Barik’s data on the RSV-RNAi treatment paradigm, including efficacy of siRNAs after RSV infection. Indeed, siRNAs reduce viral titers up to 3 days after infection in a mouse model where maximal viral burden is seen by day 4 of the infection. Moreover, illustrating that the upcoming proof-of-concept studies for RSV in the experimental infection model is not merely an academic RNAi therapeutics de-risking exercise, the patients that Alnylam expects to treat with ALN-RSV01 will have infection in the upper respiratory tract with little or no exposure in the lung. Consequently, ALN-RSV01 will be a combination of treating upper respiratory tract infection and prophylaxis for the lung. Everybody, of course, is quite restless now to learn about the outcome of the phase II experimental infection results which appear to be on track and from which will be presented in detail early next year without precluding a PR on the results in December.

Before the session on RNAi/oligonucleotide delivery, a number of speakers spoke about recent findings on the immunogenic properties of various forms of RNAs, including siRNAs. The importance of considering the expression pattern of the main oligonucleotide receptors (the endosomal TLR-3,7,8,9 and the cytosolic RIG-I and MDA-5), their localization, and their exact ligands became evident and will inform modification strategies aimed at avoiding the induction of unwanted cytokine responses following siRNA administration which is something that Rachel Myers noted Alnylam is still trying to get a better handle on (see the delay in the pandemic flu program). Other groups, however, such as Gunther Hartmann’s group from Bonn, embrace the immunogenic properties of some, particularly unmodified siRNAs and would like to combine it with their silencing activity for treating cancer and viral infections. Some encouraging data in that regard were presented.

Song Li (Pittsburgh) presented the use of neutral lipids for delivering a variety of oligonucleotides to the pulmonary circulation. Like others at the meeting, this approach illustrates a trend away from using cationic liposomes which have been associated with interacting with components of the blood and may trigger certain toxicities. Delivery to endothelial cells of larger vessels and capillaries was particularly efficient, and his group is currently collaborating with ISIS on targeting endothelin-1 for treating hypertension. Encouragingly, he was able to report on good gene knockdowns and quite impressive in vivo efficacy data, such as the reduction of hypoxia-induced right ventricular hypertrophy.

Similarly impressive in vivo efficacy data were then reported by Klaus Giese from Silence Therapeutics, mostly located in the host city of Berlin. Without showing all the controls, tumor burden, metastasis, cell proliferation were all strongly reduced following systemic delivery of siRNAs in a number of mouse tumor models. Although I cannot agree with his repeated claims on the uniqueness of their Atu-siRNAi design and related IP claims, the Atuplex delivery technology certainly deserves more credit. Unlike other lipid-based delivery methods, siRNAs here associate with the liposomes externally thus allowing liposomal charge to be modulated from cationic, via neutral, to anionic. Of practical importance, these particles can be lyophylised for storage and shipping and then resuspended without loss of silencing activity. Progress has also been made on lung delivery and a number of pre-clinical programs are being pushed forward into the clinic.

The founder and CEO of the nucleic acid delivery company Novosom, Steffen Panzner, continued the string of impressive in vivo efficacy data in a mouse model of RA (inflammation of the paws) following delivery of their proprietary siRNA-loaded Smarticles, a charge-reversible liposomal delivery technology. One problem of using the more desirable, in terms of safety, neutral and anionic lipids, is their reduced siRNA binding affinity. Novosome’s amphoteric liposomes circumvent this problem by binding siRNAs at low pHs, at which point they are positively charged, and rapidly shifting them to higher pHs for closing the liposomes. During this process, most of externally bound siRNAs are shaved off, with the additional benefit of consequently minimising endosomal exposure of the siRNAs which may trigger TLR7-mediated cytokine responses.

The scientific day was closed by a talk from Alan Sachs of Merck who set out to emphasise that despite a dearth of recent information on Sirna’s/Merck’s preclinical and clinical RNAi therapeutics pipeline, that they were more than ever committed at developing RNAi therapeutics. Unfortunately, this was not followed by a presentation of primary scientific data and specific examples, but rather how Merck thinks about RNAi drug development in general and repeated appeals of please partnering with Merck, particularly in the area of targeted siRNA delivery. It appears to me that Merck would like to rival Alnylam in accessing the best minds in the RNAi drug delivery field and is offering to send out their siRNAs for free so that others can formulate them and report back on their findings without any strings attached as to the use of these results. Sachs feels that Merck, through their expertise on the genomics and genetics of gene expression through Rosetta Inpharmatics, that it knows best which targets are most suitable to go after for drug development. Since siRNAs allow essentially any gene to be targeted, he like other feels that this is the natural drug development platform to be harnessed and thinks about 21 month development timelines from gene identification to preclinical proof-of-concept given the availability of suitable biomarkers. While searching for more targeted delivery solutions and collaborators, Merck is meanwhile establishing a platform for siRNA delivery to the liver and, encouraged by ISIS' 301012 results, is clearly motivated in targeting ApoB100 with siRNAs for the treatment of hypercholesterolemia. I would not be surprised at all, to see them join the PCSK9 frenzy as well. Being second in a $50-60B should not be that bad.
,

Journal Club: Target Mimicry as a New Way to Regulate microRNA Activity

Just how intricately RNAi-related pathways can be wired, is demonstrated by a study on the fate a nutrient-regulated microRNA in plants [Franco-Zorrilla et al. (2007). Target mimicry provides a new mechanism for regulation of microRNA activity. Nature Genetics, Advance Online Publication, doi: 10.1038/ng2079].

Unlike in animals where microRNAs recognise targets sites of incomplete base-pair complementarity to promote translational repression, microRNAs in plants largely target fully complementary sites in an mRNA. Like with perfectly paired siRNAs in animals, this leads to the destruction of the targeted transcript by endonucleolytic cleavage at the target site. This process is thought to be rapid and the small RNA is then free to recycle and target a new message. By contrast, the turnover rate for animal microRNAs is much less well understood. However, as was often the case in the relatively short history of RNAi-related research, this study in plants may offer us a clue about the kinetics of microRNA activity in animals.

Franco-Zorrilla and colleagues observe that the induction of a microRNA that is regulated based on phosphate availability and that has known target transcripts based on perfect microRNA-target site complementarities around the expected cleavage sites, is accompanied by another non-coding RNA (RNA that is not translated into proteins) that also has high base-pair complementarity to the microRNA, but with telling mismatches around the otherwise predicted cleavage site. Sure enough, this non-coding RNA is recognised by the microRNA, but not degraded. Through a series of elegant genetics, the authors demonstrate that this sufficiently diverts the microRNA, which is of low abundance to start with, so that it cannot act on their “normal” mRNA substrates any more. Hence, the non-coding RNA functions as a sink and regulates microRNA activity by tricking it to bind to itself. This strongly suggests that the turnover of the microRNA complex on incompletely based-paired targets is very slow and suppression requires a one-to-one microRNA-target site relationship.

If you have cared to read all this, I will now tell you what I think the implications are for the development of RNAi Therapeutics. When thinking about off-target effects, we have been largely concerned about the detrimental effects of suppressing unintended mRNAs, largely through microRNA-like translational suppression. However, if the plant system reflects RNA silencing kinetics in humans, then another consequence of off-targeting may be decreased on-target activity due to decreasing the pool of available siRNAs. Consequently, if it were possible to prevent such off-targeting, for example through chemical modifications and bioinformatics, then one could think about lower siRNA doses in the clinic. It is not clear whether the 2’O-methylation strategy pioneered by Dharmacon scientists to limit off-target silencing also prevents microRNA-like binding of an siRNA to incompletely based-paired RNAs, but the strategy certainly points in the right direction.