Alnylam Launches Pair of Pre-emptive PR Strikes to Cover Patent Losses

When the therapeutic relevance of a patent (Tuschl-I) that once had the potential to be broadly applicable to RNAi Therapeutics has been lost forever in Europe, and when in another US patent proceeding an important claim covering a cancer drug candidate that has just completed a phase I study has been declared invalid due to lack of written description support and may actually belong to your worst nightmare, it seems hardly a time to declare victories. Nevertheless, the PR department at Alnylam has risen to the challenge and done just that: two press releases (here and here) by the company shortly following these patent outcomes announced with great fanfare that Alnylam’s IP position had been bolstered. The reward for the skilled use of small qualifiers such as ‘requested’ and ‘key’ and obedience? Not just keeping your job, but actually being promoted to Vice President of Investor Relations.

Alnylam vs Silence Therapeutics et al.: Tuschl I in Europe

As Silence Therapeutics can attest, the Tuschl I patent estate, licensed by Alnylam and Merck, was once an effective deterrent against competing RNAi triggers of 21-23 nucleotides and/or base-pairs. While certainly an important contribution to the eventual discovery of RNAi triggers for use in human cells, its main vulnerability in IP terms derived from the fact that its research was mainly conducted in Drosophila fly lysates and tissue culture cells.

Consequently, in both the US and Europe its importance for RNAi Therapeutics has been waning to the point of being essentially relegated to the use of RNAi triggers isolated from Dicer processing reactions in the test tube (my bet is that we will never see such a product candidate). Nevertheless, some passages in the written description and claim language in the European T-I patent EP 1309726 was sufficiently ambiguous that it must have worried companies like Silence Therapeutics, BASF, and Sanofi-Aventis that it might be exploited by a legally aggressive company such as Alnylam as a convenient casus belli, causing them to oppose the patent. Specifically (former) independent claim 10 may have raised such concerns:

‘10. A method of producing knockdown cells, comprising introducing into cells in which a gene is to be knocked down isolated double-stranded RNA of from 21 to 23 nucleotides in length and corresponding to a sequence of the gene, that targets the mRNA corresponding to the gene and maintaining the resulting cells under conditions under which RNAi occurs, resulting in degradation of the mRNA of the gene, thereby producing knockdown cells.’

On February 29 and March 1 there was an unusually lengthy, and therefore probably highly contested Oral Hearing on the case, the outcome of which was quickly claimed by Alnylam (and patent co-owner UMass) as a victory. Particularly, it claimed that the ‘requested claims of the ‘726 patent were upheld without modification’ [emphasis mine]. Turns out that a comment poster on this blog last week got it right in that he/she pointed to the qualifier ‘requested’. To back up, essentially what happens in such an Oral Proceeding is that the Defending Party (here: Alnylam/Umass) submits alternative claim sets in case that the original claim language is rejected by the EPO based on the principle that a weaker patent is better than no patent at all (e.g. for the purposes of appearance).

After reading the now published documents from the Oral Hearing (if you have problems accessing them, you can request them by email from me), what happened is that Alnylam has indeed had the boldness to represent that after having been forced to modify the claims, the new claims were upheld in unmodified form. If you are an Alnylam investor/supporter, you may want to ask the new VP whether you are justified in feeling cheated by such an insincere conclusion and may want to re-think the general credibility of that company.

The upshot in terms of the patent: by clarifying the written description with expressions such as ‘isolating from combination’ and ‘in vitro’ and completely re-writing the claims, T-I is confirmed to be without relevance for RNAi Therapeutics. It remains important, however, for the commercialization of in vitro RNAi-based target discovery/validation services such as offered by companies like Cenix, although only for dsRNAs of 21-23nucleotides in length.

Alnylam vs Tekmira: Alnylam about to lose ‘key’ ALN-VSP patent

Similar to the T-I patent confrontation, Alnylam was quick to claim victory in yet another of its patent fights: ‘Alnylam retains key claims for Kinesin Spindle Protein (KSP) RNAi Patent in Interference Proceedings’. Briefly, last year the USPTO had called an Interference between an issued patent by Alnylam (US 7718629) covering an siRNA sequence in ALN-VSP02 (phase I studies completed) and a patent application by Tekmira which may in fact cover the same sequence.

As the UTPTO cannot grant identical claims in two different patents, the Interference proceeding will decide on the validity and priority of these claims. As Tekmira pointed out, this Interference has no relevance to that company’s pipeline. It could, however, mean that a patent covering a ‘key’ feature of ALN-VSP02 (an siRNA sequence) may not only turn out to be invalid, but actually belong to Alnylam’s worst nightmare: Tekmira.

By now, we all know the drill: look for the little qualifiers. ‘Key claims’ would be the obvious candidate here to home in on. But in this case, not even twisted logic can explain Alnylam’s conclusion, as in fact no claim (at least those contested and related to synthetic RNAi triggers) was ‘upheld’. Quite the opposite, the Appeals Board (BPAI) granted Protiva’s/Tekmira’s Motion 2 which requested that Alnylam’s KSP sequence claim be ruled unpatentable. The explanation was that while the claim covered dsRNAs between 15 and 30 nucleotides, only 19bp siRNAs with 2 nucleotide overhangs were presented in the application (--> lack of written description support). Since all the other claims relating to synthetic RNAi triggers are dependent on claim 1, I’m puzzled as to how Alnylam could ever claim that ‘key claims’ had been upheld, when in fact it looks like all the relevant claims will be revoked.

Claim 1 of US patent 7718629:

‘1. A double-stranded ribonucleic acid (dsRNA) for inhibiting the expression of a human kinesin family member 11 (Eg5) gene in a cell, wherein said dsRNA comprises a sense strand comprising a first sequence and an antisense strand comprising a second sequence complementary to SEQ ID NO:1311, wherein said first sequence is complementary to said second sequence and wherein said dsRNA is between 15 and 30 base pairs in length.’

Protiva’s Motion 1 that Alnylam’s provisional applications be declared invalid as priority documents was denied. The importance of this ruling is that it may affect the outcome of the second stage of the Intererence proceeding which will be about determining priority. Because the provisional applications by Tekmira and Alnylam were filed only two months apart, with Alnylam having the earlier date, chances are that Alnylam would enjoy priority IF Alnylam’s ‘key’ claims were held valid. However, with the above ruling and a recent rejection of claim 32 in a pending patent application by Alnylam (serial number 13/165568), this seems highly unlikely (note: I am unable to see the pending patent application, but claim 32 should be directed at the KSP sequence). By contrast, the Appeals Board decided that, unlike Alnylam’s patent, the KSP sequence claims are supported in the written description in Tekmira's competing patent.

The upshot: Tekmira prevailed on all Motions (filed by Alnylam and Protiva), except for on the validity of Alnylam’s provisional applications for priority purposes and the Appeals Board deferring decision on including claim 32 from the pending patent application which, based on the March 6 rejection does not seem to help Alnylam much anyway. This means that not only are the ‘key’ sequence claims about to be rejected, but Alnylam’s worst nightmare may actually end up owning it.

Don’t take my word for it, you can read the relevant documents yourself here (T-I) and here (Interference). If you are somewhat familiar with reading such documents, it will only take you 30-45min to read (and understand) them each. It seems that Alnylam is relying on the fact that most news organizations and analysts cannot be bothered to read the primary documents and conveniently adopt Alnylam’s version of events. If you do a Google search for news related to Alnylam, you will see that this happens more often than not. After all, Alnylam has more money to feed into the biotech machine: why antagonize a lucrative source of income?