Chloe Sullivan

January 2026

To briefly recap the key events leading up to this decision, the US Supreme Court found a lack of enablement when assessing the PCSK9 antibody claims in Amgen’s US patent and the Munich Central Division of the UPC initially revoked the European equivalent based on lack of inventive step. The first instance decision of the Munich Central Division was appealed, and we now have the decision of this appeal.

In short, the Court of Appeal appear have taken an approach closely resembling that which we would expect to see from the EPO in their assessment of both sufficiency and inventive step. The decision therefore indicates a willingness of the UPC to keep in step with the EPO when it comes to broad, functionally defined antibody inventions.

Amgen’s European patent in question relates to anti-PCSK9 monoclonal antibodies for use in treatment and, in particular, claim 1 of the patent defines the antibody solely in terms of its function, i.e. not by any specific sequence or structure.

As mentioned above, the Munich Central Division revoked the patent at first instance, their reasoning being that the target, PCSK9, was known and, thus, generating antibodies against this target was a matter of routine experimentation.

Amgen appealed the first instance decision, arguing that the court had applied the wrong legal framework regarding the skilled person’s reasonable expectation of success, and Sanofi (along with their co-respondent, Regeneron) cross-appealed, maintaining that the patent was also invalid for lack of sufficiency.

Sufficiency

On appeal, Sanofi argued that the claimed functional definition, which required binding to the catalytic domain and blocking of the LDLR interaction, imposed an undue burden on the skilled person. Sanofi also argued that the patent failed to teach how to produce the full breadth of antibodies covered by the claims.

Sanofi relied on expert declarations to point out that there were specific embodiments falling within the scope of the claims which would be very difficult to make. These embodiments are the so-called EGFa mimics. The Court of Appeal was not convinced by Sanofi’s argument and noted that the patent disclosed several methods for identifying antibodies that bind to the catalytic domain, including X-ray crystallography and scanning mutagenesis.

Sanofi then argued that X-ray crystallography was onerous and unpredictable, but the Court of Appeal dismissed this, noting that occasional failure is part of scientific work and that a reasonable amount of trial and error does not necessarily equate to undue burden. The Court of Appeal also clarified that the non-availability of certain embodiments was immaterial to sufficiency, provided that the skilled person could obtain a suitable number of embodiments falling within the scope of the claims.

In particular, the Court of Appeal noted that “tricky to make” does not equate to “impossible to make”, nor does it mean that there is undue burden placed on the skilled person. The Court of Appeal thus found that a patent need not enable each and every conceivable embodiment in order to be sufficient. Provided that the skilled person is able to deduce at least one way to perform the invention and can obtain suitable embodiments within the functional scope of the claims, the non-availability of specific variants, like EGFa mimics in this case, is immaterial to sufficiency in the eyes of the Court of Appeal.

The Court of Appeal thus concluded that the burden of proof lay with Sanofi to demonstrate that the skilled person would be unable to obtain antibodies with the claimed functional properties. The court found that Sanofi had failed to demonstrate this and, thus, the insufficiency attack was dismissed.

This contradicts the finding of the US Supreme Court, which took the view that the sheer amount of effort required to produce all of the antibodies falling within the scope of a broad functional antibody claims led to a lack of enablement.

By contrast, the EPO is generally content to accept that making new antibodies falling under a functional definition is entirely within the capabilities of the skilled person. In the eyes of the EPO, it generally doesn’t matter that it might take the skilled person a very long time, as long as it is possible. Successful sufficiency attacks against broad functional antibody claims at the EPO are therefore uncommon.

Both the first instance decision from the Munich Central Division of the UPC and the presently discussed Court of Appeal decision suggest that the UPC intend to follow the EPO’s lead when assessing sufficiency of broad functional antibody inventions.

Inventive Step

As mentioned, at first instance, the patent fell down for lack of inventive step.

The Court of Appeal agreed with the Munich Central Division that a realistic starting point for inventive step was a paper disclosing that PCSK9 reduces LDLR levels and teaching that inhibitors, including antibodies, should be explored for use in treatment. However, where the Munich Central Division saw a clear path to the invention starting from this paper, the Court of Appeal was not as easily convinced. The Court of Appeal acknowledged it was known at the priority date that PCSK9 could function through two pathways, namely an intracellular pathway and an extracellular pathway. The court also confirmed it is well known that antibodies being large molecules cannot enter the cell and, thus, can only target the extracellular pathway.

The Court of Appeal thus considered that, for a skilled person to have a reasonable expectation of success, they would have needed to know that blocking the extracellular pathway alone would be sufficient to achieve a therapeutic effect in vivo. Accordingly, the Court of Appeal found that the prior art left the relative contribution of these pathways unresolved, not least because the paper which they had selected as the closest prior art explicitly stated, “we do not know which pathway predominates”, and it was also noted that other literature in the field shared this uncertainty.

The court further concluded that the mere fact pharmaceutical companies were working on PCSK9 antibodies was not sufficient to warrant a reasonable expectation of success. Instead, the court clarified that this simply indicated PCSK9 was an interesting target to explore but did not render the invention obvious.

Thus, the court found that the skilled person could not have reasonably predicted such an antibody approach would result in a therapeutically effective treatment.  The Court of Appeal therefore set aside the first instance decision to revoke the patent and, instead, the patent is maintained as granted.

Although the UPC Court of Appeal did not follow the EPO’s strict problem-solution approach and instead favoured of a more holistic approach, taking into account the context of the specific case in hand, the subtle difference in approach does not appear to have significantly influenced the decision.

Therefore, although the UPC’s approach to inventive step may be slightly different to the well-known problem-solution approach employed by the EPO, this case supports that the UPC’s decisions on inventive step will generally be aligned with the EPO’s.

Conclusion

This decision therefore offers us some reassurance, in that we can now see a certain amount of harmonisation maintained between the UPC and the EPO, particularly in relation to broad, functionally defined antibody claims.

Nevertheless, however reassuring the apparent harmonisation between the UPC and the EPO may be, we cannot overlook the significant divide between the approaches taken by the US and European courts, when it comes to assessing broad antibody inventions. Particularly in relation to sufficiency, this decision highlights that the UPC and the EPO stand together on one side of the fence, while the US Supreme Court is very clearly sitting on the other side.