2 Short Stories: An siRNA Delivery Paper by Merck, Pharmaceutical Interest in Silence’s Gene Target

The most enjoyable part in following RNAi Therapeutics is to look at the rich stream of scientific data and determine the absolute maturity and competitive position of the technologies and companies involved, as well as getting a glimpse at relationship dynamics. I therefore thought to share today two examples of this that I picked up recently. One is a paper by Sirna Therapeutics/Merck shedding some light on their approach towards RNAi pharmacology and RNAi trigger design. The other is some intriguing evidence that Silence Therapeutics’ most important gene target, PKN3, is gaining traction in the pharmaceutical space.

Studying the pharmacology of siRNA delivery

Pei and colleagues from Merck published in RNA a nice paper on better understanding the pharmacology of siRNA delivery [Pei et al. (2010). Quantitative evaluation of siRNA delivery in vivo]. Unlike small molecules or even antibodies, the pharmacology of RNAi Therapeutics is more complex as simply measuring the raw tissue abundance of an RNAi trigger is a poor indicator of successful RNAi delivery. This is because functionally inactive siRNAs may vastly outnumber the active siRNAs loaded into the mammalian Argonaute 2 protein (Ago2), the nuclease responsible for seeking out and destroying complementary target messenger RNAs.

Not surprisingly, the Merck researchers employed the LNP/SNALP delivery technology in rodents and monkeys as their system of choice. After intravenous delivery of these LNPs, siRNA abundance was determined by quantitative PCR both at the tissue (mainly liver) and Ago2 level.

For the LNP aficionados among you, the 1mg/kg ED50 lipid nanoparticle used in this study still involved the CLinDMA lipid that was shown previously by Merck to be associated with immunostimulation (Abrams et al. 2010), something that Tekmira has interpreted as being the result of the strong positive charge of such LNPs.

Although playing too many number games carries the risk of missing biology sometimes, a number of quite interesting findings were made. One is that the vast amount of siRNA in the liver (>99%) is lost in the first 24 hours upon which a slower tissue elimination phase sets in that is apparently dominated by the turnover of guide strand incorporated in Argonaute.

It is generally thought that the longevity of gene silencing often seen in vivo, often on the order of 1-2 months following a single administration, is due to the stability of this complex. Despite that, there was still an approximately 3-5 fold decrease in the abundance of such complexes over a week. Not too fast, but fast enough to make it worthwhile studying in more detail whether the stability of these complexes is limited by Argonaute protein half-life or by a selective removal of the guide strand. If the latter, siRNA structure-chemistry may be able to increase silencing duration still. Such studies should also shed light on what pharmacological advantages siRNA depots might have which could be of particular interest to ocular and oncology applications.

Merck employed Zamore rule in siRNA design

The paper also allowed for some interesting insights into the siRNA trigger design process employed by Sirna/Merck. Supporting the importance of the Zamore end-stability patent recently issued in the US and exclusively licensed to Silence Therapeutics (Intradigm) from UMass [note: this corrects an earlier version that improperly stated the IP had been assigned to Silence], the authors first determined the relative Ago2 incorporation efficiencies of guide and passenger strands and then studied how this was changed following chemical modification of the same sequence. As a reminder, achieving a high ratio of guide to passenger strand in RiSC is widely considered to be beneficial both for reducing passenger strand-mediated off-targeting as well as enhancing siRNA efficacy.

Indeed, the authors find that chemical modification changed (in this case enhanced) guide strand incorporation over passenger strand incorporation. However, the authors argued that this was not due to the application of the Zamore rules, but due to having added inverted caps to the ends of the passenger strand. I agree that since the 5’-modification of the guide strand plays a major role in Argonaute loading, these caps, as also employed e.g. by mdRNA, should have a considerable effect on loading the passenger strand. However, since the modified siRNA with which the comparison to the unmodified siRNA was undertaken contained additional modifications besides the caps, it is not possible to argue that it was only the caps that had the effect on differential loading. In fact, the differential strand loading efficacies of 2 modified siRNAs, distinguished only by the nature and position of backbone modifications, differed by a factor of 2, clearly indicating that siRNA modifications besides the cap have a major influence on differential strand loading.

Wyeth/Pfizer shows interest in PKN3 in cancer

I had always considered it the wrong strategy for Silence, even more so before their merger with Intradigm, to focus so much of the company’s resources on a single drug target: PKN3. One reason is that Silence claims to be an RNAi platform company, and resources would have been better spent on building on their early pioneering position in siRNA design which then seemed to be at the risk of getting stuck in an early 2000’s 'dead-end'.

Even more worrisome is that Silence was essentially the only group really working on the PKN3 gene, and at that early stage it is always a very real possibility that it might turn out to be a useless artifact of no commercial value. However, the scientists stuck to their guns and it now seems that additional data confirms PKN3 to be an interesting oncology target in the angiogenesis field with the rest of the pharmaceutical world slowly paying attention.

Curiously, it is research by Wyeth, now part of Pfizer (!), that confirms that PKN3 plays a role in endothelial biology and that it is upregulated in a number of cancers. Even more intriguing is the fact that one of the co-discoverers of PKN3 as a cancer drug target (Anke K.-G.) is also named as an inventor in a PKN3-related patent application by Pfizer published this year (WO 2010/105128 A2). This patent application is about methods of using PKN3-containing protein complexes for cancer diagnostic purposes, e.g. determining patients with high PKN3 levels which would be candidates for a PKN3-targeting drug just like Silence’s Atu-027 (this candidate has been reviewed here with Tobias Wolfram). A nice validation of Silence’s own results and demonstrating just how close Pfizer’s PKN3 science seems to be to that of Silence is that a number of experiments described in the patent application were based on the same rodent cancer models previously employed by Silence showing that PKN3 silencing leads to an inhibition of tumor growth in mice.

With the PKN3 gene patented by Silence as a cancer drug target, it would make sense for Pfizer to gain access to Silence’s IP and maybe even take on the clinical development of Atu-027 itself for which Pfizer could use their methods as a response biomarker. Maybe Silence’s belief in PKN3 will be financially rewarded after all, and it might also explain why Alnylam seems to be so keen in weakening Silence’s PKN3 patent estate. Until now, I had come to believe that the only purpose of fighting that patent estate was to frustrate Silence by engaging them in yet another patent skirmish. It is interesting to speculate that the strength of PKN3 science and IP could critically inform whether Pfizer will partner with Silence or Alnylam.