30 July 2024

Sanofi v. Amgen: the UPC's first revocation action and therapeutic antibody patentability, from revocation to reversal

The PCSK9 dispute, a worldwide litigation

The case belongs to a confrontation under way since 2014 around Amgen's patents covering evolocumab (Repatha®), asserted against alirocumab (Praluent®) marketed by Sanofi and Regeneron. The litigation unfolded across several continents, up to the United States Supreme Court which, in May 2023, invalidated Amgen's US claims for lack of enablement. Few patent families have been tested before so many courts, under so many doctrines, and with such divergent outcomes.

As soon as the UPC opened, on 1 June 2023, Sanofi and Regeneron brought a revocation action against patent EP 3 666 797 before the Munich Central Division, while Amgen brought an infringement action before the Munich Local Division. The first patent revocation in the Court's history would follow.

16 July 2024: the first revocation ordered by the UPC

The Central Division first rejected the strictly functional reading of the claim language: absent a clear indication of an exclusive binding site, the skilled person would not have understood the claim as covering all antibodies binding to the catalytic domain of PCSK9. In doing so, the division gave the description significant weight in claim interpretation, departing from the then dominant practice at the EPO.

On inventive step, the division set aside the classic problem-solution approach in favour of a "realistic starting point" in the prior art. Noting the prior art's explicit orientation towards antibodies blocking the PCSK9-LDLR interaction, it concluded that the skilled person had a reasonable chance of success, without inventive effort despite the scale of the work required: developing antibodies against a known target was routine. The message to the market was severe: functional antibody claims directed at a known target entered the UPC era under a presumption of fragility. The patent was revoked in its entirety, underArticle 56 EPC.

Claim interpretation: the weight of the description

The first lasting contribution of the 2024 decision: before the UPC, the description is not a subsidiary resort reserved for ambiguous claims, it always plays a part in interpretation. The Alexion appeals (UPC_CoA_402/2024 and UPC_CoA_405/2024) confirmed that line for antibodies: strict textual construction, with no judicial rescue of defective wording.

The gap with the EPO has since closed: in decision G 1/24 of 18 June 2025, the Enlarged Board of Appeal held that the description and drawings must always be consulted when interpreting the claims. On this ground, it is the UPC's practice that set the standard.

25 November 2025: the reversal on appeal

Ruling on the appeals (UPC_CoA_529/2024 and UPC_CoA_528/2024), the Court of Appeal set aside the revocation and upheld patent EP 3 666 797. The judgment, available in the UPC register of decisions, adopts a holistic approach to obviousness: the skilled person's motivation and the reasonable expectation of success are assessed together, starting from an objective technical problem, providing an effective treatment of hypercholesterolemia, and from a realistic starting point in the prior art.

The tipping point is evidentiary: for the Court of Appeal, the skilled person could not reasonably predict that an anti-PCSK9 antibody would be therapeutically effective, given the scientific uncertainties then surrounding PCSK9's pathways of action. The Court also gave full effect to the claim language: the claimed antibodies must be genuinely effective for the intended treatment, any reduction in cholesterol not being enough.

The outcome converges with that of the opposition proceedings before the EPO, which had likewise left the patent standing. The divergence between the two UPC instances thus concerned neither the method of interpretation nor the rejection of the problem-solution approach: it concerned the threshold of the reasonable expectation of success, that is, the plausibility of therapeutic success at the priority date. That threshold, and the evidence available to meet it, will now be the true battleground of European antibody litigation.

Therapeutic antibodies: the European benchmark after the saga

The transatlantic contrast remains instructive. In the United States, Amgen's functional claims perished through enablement: too broad, the monopoly exceeded what the disclosure allowed to be reproduced. In Europe, the debate has shifted to inventive step and its reasonable expectation of success: the path to an antibody patent remains narrow, but it exists, provided the uncertainties overcome and the therapeutic efficacy achieved are documented from filing.

On the infringement side, the Düsseldorf Local Division completed the picture on 13 May 2025 (UPC_CFI_505/2024) by subjecting second medical use claims to a two-limb test, causality and the defendant's knowledge. For patentees, the overall lesson is clear: the value of an antibody family turns on the quality of the scientific and evidentiary record, from drafting through to the preservation of evidence before the UPC, and can flip at every level of the proceedings.

Key takeaways

The first revocation action before the UPC succeeded on 16 July 2024: the Munich Central Division revoked Amgen's anti-PCSK9 patent (UPC_CFI_1/2023) for lack of inventive step.
On 25 November 2025, the Court of Appeal set aside that decision (UPC_CoA_529/2024) and upheld patent EP 3 666 797, for want of a reasonable expectation of therapeutic success for the skilled person.
The UPC interprets claims in the light of the description; the EPO aligned itself with decision G 1/24 of 18 June 2025.
The Court of Appeal assesses obviousness holistically, without a mechanical problem-solution approach: motivation and reasonable expectation of success are examined together.
Developing antibodies against a known target is not necessarily routine: documented scientific uncertainties can support inventive step.
The comparison with the 2023 US judgment reveals two distinct gatekeepers: enablement in the United States, inventive step and plausibility in Europe.

Frequently asked questions

Why was the revocation of Amgen's patent reversed on appeal?

Because the Court of Appeal, assessing obviousness holistically, held that the skilled person could not reasonably predict the therapeutic efficacy of an anti-PCSK9 antibody in view of the scientific uncertainties of the time: inventive step was therefore established.

Does the UPC apply the EPO's problem-solution approach?

No, not mechanically: the UPC starts from a realistic starting point in the prior art and assesses motivation and the reasonable expectation of success together, a method confirmed by the Court of Appeal in its judgment of 25 November 2025.

What are the consequences for drafting antibody patents?

Document from filing the technical uncertainties overcome and the genuine therapeutic efficacy of the claimed antibodies, and align claims strictly with the description, since both the UPC and the EPO now interpret claims in its light.

Does a UPC revocation apply to the whole of Europe?

It applies to all Contracting Member States where the patent is in force, which makes the centralised revocation action a powerful lever, but one whose outcome can be reversed on appeal, as this case illustrates.

Dhenne Avocats conducts revocation actions and validity defences before the Unified Patent Court and the French courts, notably in biotechnology and therapeutic antibody matters. Talk to us.

This article is an original and substantially updated adaptation of an analysis by Matthieu Dhenne first published on Kluwer Patent Blog on July 24, 2024: Sanofi vs. Amgen: Successful first UPC revocation action and strict assessment of therapeutic antibodies patentability.

Facts of the case

In this case, an action for revocation was brought against the patent EP 3,666,797, one of Amgen's patents concerning PCSK9 antibodies. The first claim of this title concerned an antibody used to treat or prevent hypercholesterolaemia or arteriosclerotic disease, in which the antibody «binds to the catalytic domain of a PCSK9 protein of the amino acid sequence SEQ ID NO: 1, and prevents or reduces the binding of PCSK9 to LDLR». Three Sanofi companies filed a principal revocation action on 1 June 2023, the day the UPC entered into force, before the Munich Local Division of the Central Division of the UPC, while on the same day, a few minutes later, Amgen filed an infringement action against these three companies and Regeneron before the Munich Local Division. Regeneron subsequently filed a counterclaim for revocation, which the Sanofi companies did not do. These are therefore parallel and similar revocation actions before the Central Division and the Local Division, so all parties agreed that the counterclaim for revocation should be referred to the Central Division. These actions take place in a context that is at least rich. Amgen, Sanofi and Regeneron have been fighting since 2014 over Amgen's PCSK9 antibody patents, which reportedly cover both its Repatha and Praluent marketed by Sanofi and Regeneron. Initially, the disputes were brought before national European courts, before being brought before the UPC. A case relating to the US patent corresponding to EP’797 was also brought before the Supreme Court of the United States.

Key takeaway of this decision

The JUB's decision's main contribution lies in its teaching on the patentability of therapeutic antibodies. The Court first focused on interpreting claim 1, and therefore its interpretation method. First, the Court considers that the person skilled in the art interpreting a claim determines the technical meaning of the terms used with the help of the description and drawings. Then, the Court examines the functional characteristic of the claim «binds to the catalytic domain of a PCSK9». This is about understanding the functional limitation of the claim. The Court considers that for the person skilled in the art, the binding of an antibody has a function, namely to prevent or reduce the binding of PCSK9 to the LDLR. However, in the absence of any indication in the claim or the description that the binding must take place exclusively or mainly within the catalytic domain for the technical function to be achieved, it follows that the person skilled in the art would not have understood it. Furthermore, the Court also considered that the other functional limitation, namely the therapeutic effect of the antibodies (i.e., cholesterol reduction), implied that the claim only concerned antibodies with a therapeutic effect (even if it is «very weak»). Ultimately, the functional language of the claim should not be interpreted as covering «all antibodies capable of binding to the catalytic domain». Such an interpretation shows that the Court appears to want to give more weight to the description than the EPO's Boards of Appeal generally do (e.g., T 169/20, «the support of the description for interpreting the claims should only be used in exceptional cases where the subject matter of the invention and/or its technical context needs to be clarified, and can only be applicable when the invention in the description corresponds to the invention as claimed»).

Second input of the decision

Secondly, the Court develops its analysis of the inventive step requirementCBE, art. 56It is interesting to note here that the Court chose not to follow the problem-solution approach conventionally used at the EPO, by not starting from the closest, i.e. most promising, prior artEPO guideline G-VII, point 5.1), but only from a realistic starting point. The central division considered that there was an explicit direction in the prior art towards the development of antibodies capable of blocking the link between PCSK9 and LLDR, as a person skilled in the art would have understood from the prior art that blocking this link could be explored for the treatment of hypercholesterolemia. Thus, given this direction, the person skilled in the art would have had a reasonable chance of success in obtaining the antibodies defined by my claims without exercising inventive activity, even if they would have realised that the manufacture of the antibodies and the implementation of screening methods might require considerable time and resources. The court therefore revoked the patent in its entirety (i.e. here for all EPC member states), considering that the person skilled in the art was motivated to develop therapeutic antibodies against PCSK9 and that they would have arrived at the claimed antibodies with a reasonable chance of success without «undue burden» (referring to the time and resources required).

Conclusion

As we have already noted above, the Munich Central Division deviates slightly from the EPO's positions on the interpretation of claims and the assessment of inventive step. Nevertheless, the outcome regarding the patentability of antibodies is close to that of the EPO and therefore also differs from the US approach. The Court's position is close to that of the EPO, which considers that «the subject matter of a claim defining a new antibody that binds to a known antigen does not involve an inventive step, unless a surprising technical effect is demonstrated in the claim or there was no reasonable expectation of obtaining antibodies with the required properties» (EPO guideline G-II-5.6.2Therefore, according to the German local division, in the absence of particular difficulties, the development of antibodies for a known target is routine and thus not inventive. This contrasts with the revocation in the United States, based on the enablement requirement, and where enablement was deemed insufficient because, even from the description, the person skilled in the art had to make an inventive effort to arrive at the millions of antibodies covered, according to the Supreme Court, by Amgen's patents (21-757 Amgen Inc. v. Sanofi (05/18/23)). Ultimately, the UPC establishes a much higher patentability threshold for antibodies in Europe than in the United States.

Author : Dhenne Avocats.

31 August 2026