Original published on IPKat.
The Central Division of the UPC has issued its first revocation order (Sanofi v Amgen, UPC_CFI_1/2023). The decision is not just remarkable for being the first decision of its kind. The UPC Central Division also takes a strong stance on the patentability of therapeutic antibody inventions in Europe. The Central Division follows (and some may argue, takes even further) the EPO approach to antibodies, according to which the development of new antibodies for a known target is considered routine (EPO Guidelines for Examination, G-II-6.2). The decision of the UPC Central Division in Sanofi v Amgen is in stark contrast with the US Supreme Court decision on Amgen’s corresponding US patents, in which broad functional antibody claims were found to lack enablement because the Court found the task of working the invention over whole scope of the claim to require more than routine efforts (Evolve Insights). The decision of the UPC confirms that the US and European approaches to antibody inventions are diametrically opposed. In Europe, broad antibody inventions are found to lack inventive step because it is considered too routine to make antibodies.
The UPC interpretation of broad functional antibody claims
The decision in Sanofi v Amgen (UPC_CFI_1/2023) concerned Sanofi’s revocation action against one of Amgen’s PCSK9 antibody patents. Claim 1 of the patent as granted (EP 3666797) claimed an antibody for use in treating or preventing hypercholesterolemia or an atherosclerotic disease, wherein the antibody “binds to the catalytic domain of a PCSK9 protein of the amino acid sequence of SEQ ID NO: 1, and prevents or reduces the binding of PCSK9 to LDLR”. Similarly to the corresponding US Supreme Court case, the antibody was not defined in the patent by any structural features (Evolve Insights).
A significant proportion of the UPC decision is devoted to the question of how the broad functional antibody claim should be construed. The US Supreme Court in the corresponding US case construed Amgen’s patents as exceedingly broad, and as covering “an entire universe of antibodies” containing at least “millions of candidates” (Evolve Insights). On the basis of this broad interpretation, the US Supreme Court found the claimed invention to lack enablement because of the undue experimental effort a skilled person would have to put in to find all the antibodies falling under the scope of the claim.
In the present case, the Central Division of the UPC took as its legal framework for claim interpretation the previous UPC Court of Appeal decision in UPC_CoA_335/2023, NanoString/10x Genomics (Evolve Insights). The Central Division found that “[t]he interpretation of a patent claim does not depend solely on the strict, literal meaning of the wording used. Rather, the description and the drawings must always be used as explanatory aids for the interpretation of the patent claim and not only to resolve any ambiguities in the patent claim” (para. 6.4). The UPC Central Division has thus clearly taken a side with respect to the ongoing debate on claim interpretation in Europe, now the subject of a referral to the EBA (G 1/24). The UPC seems keen to give the description more weight than it is usually given by the EPO Boards of Appeal.
The Central Division thus considered the functional claim feature “binds the catalytic domain of a PCSK9”, in view of the description and the claim as a whole. The Court specifically focused on how a skilled person would understand the purpose of the functional limitation: “The skilled person recognises that the binding of the antibody […] is intended to have a consequence. Namely […] the antibody is intended to prevent or reduce the binding of PCSK9 to LDLR […] The binding […] must therefore allow for this result to occur.” (para. 6.23, original emphasis).
Given that there was no indication in the claim or description that it would be necessary for the antibody binding to take place exclusively, or even primarily, within the catalytic domain in order for the technical function to be fulfilled, the Central Division found that this would not be understood as a limitation of the claim. Additionally, the court did not think it necessary to interpret the claim as requiring any particular reduction of the binding of PCSK9 to LDLR, so long as they were suitable for the claimed use, noting that the description indicated the reduction in binding could be as low a 1-20% (para. 6.26). The Central Division also found that the other functional limitation in the claim, namely the therapeutic effect of the antibodies in reducing cholesterol, did limit the scope of antibodies covered by the claim to those with at least some (even if only “a (very) small” therapeutic effect (para. 6.30)).
The Central Division thus concluded that the functional language of the claim should not be construed as covering “all antibodies that may conceivably bind to the catalytic domain”. However, the claim would certainly be understood as being far broader than just the example antibodies described in the description (the jury is metaphorically still out on the interpretation of means-plus function language of antibody claims in the US, and whether these might be interpreted solely as covering example antibodies provided in the description).
The UPC approach to the inventive step of broad functional antibody claims
The Central Division next considered if the claimed invention was inventive pursuant to Article 56 EPC. Interestingly, the UPC chose not to follow the EPO’s problem-solution approach whereby it is necessary to first identify the “closest prior art”. For the Central Division, it was enough that the prior art cited by the claimant was a “realistic starting point”, finding that it was not necessary to determine which prior art document would have been “the most promising” (para. 8.28).
The Central Division found that a skilled person would have understood that, whilst the prior art did not disclose any antibodies blocking the interaction between PCSK9 and LDLR for use in the treatment of hypercholesterolemia, a skilled person would have nonetheless understood from the prior art “that development of anti PCSK9 antibodies that block the LDLR:PCSK9 interaction can be explored for the treatment of hypercholesterolemia” (para. 8.30, original emphasis). The Central Division further found that the skilled person would not have had any “serious doubts” about the potential role for PCSK9 in disease and its consequent attractiveness as a therapeutic target (para. 8.50).
The Central Division therefore concluded that there was explicit direction from the prior art to develop antibodies that blocked the interaction between PCSK9 and LDLR for the treatment of cardiovascular disease. The question thus became whether, following this direction, the skilled person would have arrived at the antibodies defined by the claims without inventive skill.
The patentee argued that, if a skilled person were to pursue new therapeutics for the PCSK9 target, they would not have pursued an antibody approach with a reasonable expectation of success and would further not have arrived at the claimed antibodies. However, the Central Division found that generating antibodies to a known target can generally be considered “a matter of routine” (para. 8.54).
“It is not in dispute that at the priority date, the skilled person generally knew how to generate (monoclonal) antibodies to a given protein target and that such antibodies could be screened functionally for their ability to inhibit the interaction between the target and another protein such as a receptor in an appropriate assay. The skilled person would have realised that making antibodies and setting up the screening methods may require considerable time and resources, but to do so does not constitute an “undue burden” in patent law terms.” (para. 8.68).
The Central Division rejected all the patentee’s arguments on why PCSK9 would have been considered a particularly challenging therapeutic target. The Central Division’s test for whether there was a reasonable expectation of success appears to be whether a skilled person would have had “serious doubts” that a therapeutic antibody could be developed. The onus was thus placed on the patentee to prove that a skilled person would not have expected an antibody screening campaign to succeed. This might be contrasted with a test whereby it must be shown that it was not “obvious-to-try” PCSK9 as a target, compared to all the potential targets available on which a company might expend resources.
The Central Division concluded that a skilled person would have had clear motivation to develop therapeutic antibodies for PCSK9 and would have done so with a reasonable expectation of success in order to arrive at antibodies falling under the scope of the claim without undue burden. The Central Division granted the request for revocation of the patent in its entirety across all UPC states in which the patent was validated (para. 10.1). Amgen is expected to appeal the decision.
Final Thoughts
The decision of the Central Division of the UPC in this case fits in with the pre-existing jurisdictional divide between the US and Europe on the patentability of antibody inventions. Whilst the Central Division deviated somewhat from the EPO’s formal approaches to claim interpretation and inventive step, on the substantive issue of patentability of antibodies, the UPC has taken the EPO approach and run with it. The Central Division followed the EPO view, whereby developing antibodies for a known target, absent some special difficulties associated with that target, is considered routine and non-inventive. The UPC decision thus confirms and potentially increases the difficulty of patenting therapeutic antibody inventions in Europe, regardless of whether antibodies are defined broadly by their function or specifically by their structure. This position imposes a far higher bar for the patentability of antibody inventions in Europe than that imposed by the US enablement requirement.
