The RNAi Trigger Marketplace in the Post-Tuschl World

This entry is the second of a 2-part series on the upcoming decision of who will control key intellectual property for therapeutic applications of RNAi. In the first part, I tried to provide an outline of the developments causing ownership of certain data in the Tuschl patent applications to become such an important issue. Here, I will try and delve more into the technical details of the scientific milestones that made RNAi a conceivable new class of human therapeutics, and based on that understanding make an educated guess about the outcome of the Tuschl Tussle and how this could shape the RNAi trigger IP marketplace in the future.

Back to Science.

Could RNAi be used as a therapeutic? That was the sort of topic of wild speculation in the lab where I worked as an undergrad in 2001 on a plant gene silencing project. Hey, Fire and Mello reported this cool stuff in worms 3 years ago and as we can see double-stranded RNAs can trigger the same process so beautifully also in plants thanks to some nice work by the Baulcombe group and another one in Australia. But humans? Well, unfortunately vertebrates seem to represent the exception when it comes to the existence of RNAi. That darn interferon response system...All this would change in a watershed moment when Tuschl and colleagues at the Max Planck reported in Nature the very existence of RNAi in human cells and taught a captivatingly simple technology to induce it there: siRNAs. The story therefore seemed quite simple until then. First it was Fire-Mello, then Tuschl's siRNAs.

Fire-Mello

Fire-Mello coins RNAi. The critical contribution by Fire and Mello in 1998 was their realization that it was in fact double-stranded RNA that was the effective inducer behind a variety of strange gene silencing phenomena in worm genetics and quite likely beyond (e.g. variegated Petunia flower color). While it was not necessarily obvious at the time that this would be applicable to humans as it was still very much doubted that RNAi existed in humans, the deeply influential nature of this eureka moment of the field of gene silencing and the non-exclusive licensing approach taken by the Carnegie Institution, the owner of Fire-Mello, established it as a widely respected patent. Add to this the endorsement by the scientific community as evidenced by the Nobel Prize in Physiology and Medicine this work entailed, there should be little doubt in the mind of patent examiners about the therapeutic relevance of that work. Prohhhhbably a fundamental patent.

Biochemical work in fly cell extracts by the inventors behind Tuschl I (Tuschl, Zamore, Bartel, Sharp) and involving the MIT, the Whitehead, UMass, and Max Planck on the same gene silencing phenomenon in flies aimed at the elucidation of the molecular fate of these long dsRNA RNAi triggers. Their main finding was that during RNAi, long dsRNA gave rise to 21-23 nucleotide small RNAs and that target RNA was cleaved at 21-23 nucleotide intervals also. This strongly indicated that it was the 21-23 nucleotide RNAs that were guiding the destruction of the target RNA. This, however, is different from demonstrating that the 21-23 nucleotide small RNAs are able to trigger RNAi themselves, something one would think would be important for claims to this effect to be considered enabled. To test this hypothesis, they therefore isolated and then reintroduced the 21-23 nucleotide mix of RNAs into fresh fly cell extract and asked whether those were able to induce RNAi gene silencing, too.

What may come as a surprise to a few: the silencing with these purified 21-23nt RNAs was actually quite mediocre, about 50% silencing compared to >>95% silencing with the long dsRNA (Figure 12 of US Tuschl I application). Similarly, when the dsRNA length dependency of RNAi was tested, the shorter the dsRNA, the worse the silencing. Together, these types of findings described in Tuschl I seriously calls into question claims that Tuschl I technically enabled human RNAi. Some may even cite such data as proof to the opposite, namely that this work made it even less likely that short RNAs would be useful RNAi triggers.

What is the explanation for this somewhat surprising finding? In retrospect, it is most likely the fact that when the 21-23nt small RNAs were introduced they were single-stranded and not double-stranded and demonstrates that at that time, the authors did not know about the requirement for double-strandedness also of the small RNA intermediates for triggering RNAi . Consequently, the patent contemplates both single-stranded and double-stranded RNAs as candidate RNAi triggers. Thus, while an important piece of the puzzle of RNAi molecular biology history with ~1600 citations to the underlying Year 2000 paper, it by no means was the catalyst leading to the adoption of RNAi in humans. I would not even be surprised if the authors did test the hypothesis of whether such isolated 21-23nt small RNAs were able to silence genes in humans cells (not very difficult to do) and came up short. It is also worth noting that their discovery of small RNAs during RNAi was not entirely new to gene silencing scientists then, a year after Hamilton and Baulcombe reported such an observation in Science, something that also has not escaped the patent examiner.

Tuschl II

Tuschl II coins ‘siRNA’. The critical insight that, first of all, proved the existence of RNAi in Man and even more importantly in terms of enablement, taught a straightforward method for triggering this process in humans, came from very elegant work led by Tom Tuschl at the Max Planck in Goettingen and forms the basis for Tuschl II.

The spark of ingenuity by the people at Max Planck, not involving those at the MIT, Whitehead, or UMass, was that the small RNAs had to be in double-stranded form to serve as useful triggers of RNAi. To prove this, they generated short dsRNAs, which they coined siRNAs, through chemical synthesis, also a first, and found them to be potent triggers of gene silencing not only in fly lysates, but subsequently also in human cells. The fact that the fly lysate work was reported separately by the Max Planck group and temporally between the Tuschl I 21-23 nucleotide RNA paper and the human RNAi findings, further illustrates the temporal, geographic, and intellectual separation of Tuschl's work in Massachussetts and then as a group leader in Germany.

Importantly, these siRNAs allowed for gene silencing that was specific and independent of the interferon response, again something speculated about, but not clearly proven in Tuschl I. The Tuschl siRNA-template is now used by thousands of laboratories around the world, with an amazing 6000 citations to the underlying paper further illustrating its importance.

In the Max Planck vs Whitehead case, the Whitehead argues that 3’ overhang siRNAs that are at the core of the Tuschl II patent application were already part of Tuschl I. On the surface this is true. This is because for some strange reason and that is the biggest mystery to me in all of this and that I hope the next months will shed some light on, the human siRNA data miraculously appear at the end of the Tuschl I application, as does the term ‘siRNA’ emerge without prior definition. If this data were to remain part of Tuschl I, there is the real possibility that Tuschl II could be declared invalid on a technical basis because of Tuschl I’s priority status (in a temporal sense) and double-patenting laws. Not good for Alnylam!

The two related questions of which the answer will rock the RNAi Therapeutics universe are therefore: a) Has the human siRNA data that in light of the weak activity of the ’21-23 nucleotide RNA’ in fly lysates and very uncertain translation of those results into humans now form the inventive basis for the broad human RNAi claims in Tuschl I, been rightfully included? b) In doing so, has the Whitehead, responsible for prosecuting Tuschl I also on behalf of Max Planck, fulfilled its fiduciary duty towards all its partners?

First of all, as I explained in my previous post, the data critical for the siRNA claims of Tuschl I had been generated by the inventors behind Tuschl II (most importantly, in addition to Tuschl, Elbashir and Lendeckel; both of them also at the Max Planck then), but who are not named as inventors on Tuschl I. On this technical ground already, Tuschl I in its present form is invalid. Beyond that, there appears to be early communication in which Max Planck confirmed with the Whitehead that the human siRNA data were the domain of Tuschl II. So even if Max Planck and the inventors of Tuschl I had been wrongly convinced by the Whitehead and their hired patent attorneys that this should not pose a problem for the approval of both patents, any patent attorney worth his salt should have known this to be a fundamental omission. It would therefore seem to be wise to remedy this deficiency either by including the inventors on the Tuschl I patent or by leaving out the data as stipulated by Max Planck, before the specter of 'malpractice' was raised. And obviously, Whitehead now is clearly not acting on Max Planck’s behalf and this should be sufficient cause to give back Max Planck de facto veto power in Tuschl I by confirming that the Whitehead does not have Max Planck's power of attorney any more.

Because some of these issues are civil ones that are not the domain of the USPTO, it is important to sort them out before it goes back to the patent office and can cause lasting damage to the patents. For the stated reasons, I am quite confident that Max Planck and Alnylam will prevail and regain control of the human RNAi data and some sort of declaratory judgement that the way that data had been used in Tuschl I cannot be construed to contest the validity of Tuschl II in the future.

Before I consider the ramnifications of the two main outcomes of the Tuschl Tussle for the RNAi Therapeutics RNAi trigger IP space, it should be noted that Tuschl II already disclosed the observation that blunt-ended siRNAs can silence, too, just not as efficiently as 3' overhung siRNAs on average. 3' overhangs were therefore taught to be a preferred characteristic of siRNAs when used for mammalian RNAi applications. Hence, with many more reports confirming that RNAi in fact is so robust that all sorts of exogenously introduced small dsRNAs can efficiently induce RNAi in humans, it will become more and more difficult to convince the patent offices of a proprietary nature of not only overhung siRNAs, but also those without overhangs. One exception may be Silence Therapeutics’ blunt ‘Atu-siRNAs’ which because it was a relatively early disclosure Silence/Atugen was able to convince the US and European patent offices of their arguably surprising stability, an important feature for most RNAi Therapeutics approaches.

Outcome 1: Tuschl II becomes dominant, Tuschl I essentially irrelevant

The most likely outcome. This will confirm Alnylam to be the most desirable partner based on RNAi trigger IP alone and leave Merck and RXi Pharmaceuticals empty-handed. Clear freedom-to-operate and exclusivity for the most efficient RNAi trigger that is also highly competitive with regards to other challenges such as innate immune activation. However, since Tuschl II does not claim blunt siRNAs, something I believe should have been done at least initially and maybe wasn’t because of an integrated Tuschl I-II strategy that now clearly has fallen apart, there remains scope for plenty of blunt-ended workarounds in the important 19-24 base-pair range. These workarounds, however, are not very attractive for licensing purposes if they cannot be protected by patents. Again, the exception here is Silence Therapeutics which, assuming that Kreutzer-Limmer's staying power is questionable, would be a beneficiary of such an outcome since it would now free Atu-siRNAs from the Tuschl I threat. There are, however, some significant limitations with Atu-siRNAs, since the scope of the patents is quite narrow in terms of allowed siRNA patterns and chemical modifications. Thus, while Tuschl II will offer a platform that should be applicable to RNAi Therapeutics for many years to come, Atu-siRNAs may not be able to adapt to the evolution in cutting-edge siRNA modification technology and hence its value should decline over time relatively quickly.

The IP position of Dicerna is probably least affected by the Tuschl outcome among the synthetic siRNA Therapeutics companies. There may be some uncertainties with whether and how the Tuschl patents may be applicable to Dicer substrates, but unless there will be a messy outcome in which both Tuschl’s go up in fire (highly unlikely), neither outcome 1 or 2 should change this much.

Under outcome 1, the market would have to balance the luxuries that Tuschl II offers, namely patent protection and overhangs, with the lower price, but added liabilities of the Silence Therapeutics and Dicerna platforms or even non-patented siRNA workaround designs. Last but not least, due to its use of overhangs, mdRNA’s overhung ‘usiRNAs’ would be a loser under this scenario.

Outcome 2: Tuschl I becomes (almost) gate-keeping, supersedes Tuschl II

If Tuschl I were allowed in the US in the form now proposed by Whitehead, then Tuschl II may go up in flames with Tuschl I covering blunt and overhang siRNAs comprising RNA strands of 21-23 nucleotides. Under this, albeit very unlikely scenario, Alnylam would have to share gate-keeper privileges for the most direct route to RNAi Therapeutics with Merck and RXi. Further risking to put pressure on price would be RXi selling such rights for a pittance as well as uncertainty about UMass’s ability and willingness to further grant rights to Tuschl I. Not all would be lost for the rest of the field even under this scenario. Silence Therapeutics, for example, would still be able to operate in the 15-20 base-pair range, with maybe 19 and 20 base-pair offering quite good opportunities of discovering efficacious and non-immunostimulatory siRNAs with acceptable efficiency. Similarly, 19 to 20 base-pair siRNAs may also become the preferred space for other non-patented siRNA designs, though all of this is dependent on what happens to Kreutzer-Limmer. Again, Dicerna would be little affected by all of this, and mdRNA may be well advised to try its luck with ‘usiRNAs’ outside the 21-23 nucleotides range, although I still feel chances are slim that one or two supposedly ‘non-nucleotide’ nucleotides will allow them to call what look and behave like siRNAs by another name.

Where does Big Pharma stand in all of this? Those interested in taking broad platform licenses to RNAi trigger IP can probably be classified into into two categories: 1) those like Pfizer and GSK that have diligently done their homework and will already have made up their minds about what type of RNAi triggers are required, including whether they consider overhangs to be an essential feature or not. Such companies can simply await the outcome of the trial and then choose the most economical option everything else (e.g. access to delivery and other know-how) being equal; 2) those companies that have shied away from heavy investments thus far and would prefer to get started with a pure-play RNAi Therapeutics partner providing patent-protected siRNAs and other basic RNAi capabilities. These companies may be most swayed by the outcome of the Tuschl Tussle, since they may be more relaxed in terms of what they consider acceptable siRNA designs.

Tekmira is a pure-play RNAi Therapeutics company that should be uniquely affected by the outcome, because it does not tout having invented unique siRNA triggers, although it certainly could make up such claims to the same degree that others do in the space, and because of its complex relationship with Alnylam. For one, it may determine whether potential partners consider it to be necessary to access SNALP delivery via Alnylam or whether they can go directly to Tekmira and get the same for probably considerably less. Moreover, in the unlikely case of a Max Planck/Alnylam loss, it may become even more difficult to insist on controlling SNALP delivery for RNAi Therapeutics all the while it is obvious that Alnylam is intent on minimizing the importance of Tekmira for their delivery efforts (the whole issue of what is called a SNALP which clearly differs between Alnylam and Tekmira). If SNALP is old and first-generation where is the harm in letting Tekmira fully exploit this technology by partnering it out ex-ALNY instead of letting it wither on the vine? I guess something ought to be worked out here to the satisfaction of both companies.

After 8 years of spending enormous efforts on confusing the investor world with what is valuable RNAi trigger IP, it looks like everything will come down to basic science. The collective scientific community based on the number of citations a paper gets and how it has recognized critical inventions for example in the form of scientific awards, would have been a much more straight-forward, fairer, and infinitely cheaper way of determining ‘good’ RNAi IP. Who else was better suited to spot critical contributions in technically demanding areas such as RNAi than scientists themselves? One would hope that the judge will concur, also in the interest of what RNAi Therapeutics could do for society.

Disclaimer: The above are my own interpretations of the case, based on publicly available documents from the USPTO and court sites, interviews, and press releases. Accuracy cannot be guaranteed as I may have overlooked critical elements of the case, and am neither trained in intellectual property nor contract law. Information provided herein cannot be relied upon for making investment decisions. Investments in RNAi Therapeutics are very risky and not suited for most. Consult with your own professional advisor before doing so.