EPO pharma case law trends 2025: Antibodies and biologics
Rose Hughes
EPO pharma case law trends 2025: Antibodies and biologics
Antibodies and other biologics are challenging and expensive to develop. They can simultaneously be difficult to protect in view of strict and conflicting patentability requirements in the US and Europe. The science of biologics is rapidly progressing, with the development of ever more complex protein structures, incorporation of molecules into cell therapies and the increasing use of AI-assisted design and in silico modelling. Patent law must respond to these new challenges. What better time to take a look at the trends from the EPO case law on biologics and antibodies this year?
Antibodies and biologics: Therapeutic, commercial and legal importance
Biologics are both extremely important therapeutics and commercial products in the pharmaceutical industry. Biologics can be defined as therapeutic biological molecules, such as proteins, peptides and nucleic acids. They include monoclonal antibodies, bispecific and multispecific antibodies (including T cell engagers), and the targeting molecules in cell therapy (e.g. chimeric antigen receptors and T cell receptors).
The complexity of biologics is mirrored by the complexity of the IP strategy for protecting these important therapeutics. There are a variety of different inventions relating to biologics that may be protected, including the target, epitope, sequence, structure, therapeutic use and manufacture of the molecule. However, the case law on what can be patented is also complex and constantly developing. Innovators have to navigate the sufficiency and inventive step squeeze, whilst balancing the opposing approach to antibody patentability of different jurisdictions. In Europe, the patentability of biologics follows the general EPO approach, namely a laser focus on the problem solved by the invention (IPKat). This approach permits broad functional claiming (IPKat), but can lead to significant patentability challenges with respect to inventive step and demonstrating a surprising technical effect of a new biologic performing a known function.
Patent law must now also grapple with inventions relating to ever more complex biologics, including cell therapies. We have already seen the strict US written description requirement applied to CAR-T cell therapy (IPKat), and we are now seeing the first CAR T cell therapy cases before the Boards of Appeal. Despite their complexity, the patenting of CAR molecules in Europe can be as challenging as for traditional monoclonal antibodies, given that both are often based on optimisation or even repurposing of a known antigen binding domain.
Added matter (Article 123(2) EPC): Intermediate generalisations
The EPO approach to added matter remains ruthlessly strict, in particular with respect to "intermediate generalisations" where applicants are found to have selected a pick-and-mix of features from different lists disclosed in the application.
The patent in T 0250/24 (MND promoter/2SEVENTY BIO) (EP 3689899) related to a CAR-T cell vector. Claim 1 defined the CAR construct by a very specific combination of five structural domains: (a) an anti-BCMA scFv, (b) a CD8alpha hinge, (c) a CD8alpha transmembrane domain, (d) a CD137 co-stimulatory domain, and (e) a CD3zeta primary signalling domain. The Opponent argued that the claimed combination was not disclosed in the original application.
The Board of Appeal agreed with the Opponent. The Board of Appeal particularly found that to arrive at this claim, the skilled person had to make at least three selections from different lists. The Board of Appeal found that the selection of BCMA as a target had to be selected from a long list of possible antigens, the CD8alpha hinge had to be selected from other options, and the CD137 domain had to be selected from a list of co-stimulators. For the Board of Appeal, the application as filed contained no pointer toward this specific combination. Furthermore, the specific examples provided in the application as filed that did disclose this combination also included other limiting features that had been omitted from the claim, creating what the Board of Appeal considered to be an impermissible "intermediate generalisation". The patent was thus revoked for added matter.
In a similar vein, T 1403/24 (Sost antagonist/OSSIFI-MAB) concerned a patent (EP 3345607) for a sclerostin antagonist antibody for increasing bone density. The claim specified that the antibody was "serially administered with an antiresorptive drug". The Board of Appeal found this was also an unallowable combination. In particular, the original application was found to have disclosed "increasing bone density" as one of several distinct therapeutic uses, and separately disclosed "serial administration" as one of several administration modes (alongside co-administration, etc.). For the Board of Appeal, there was no specific disclosure linking the particular selected therapy with the particular administration mode. The appeal was therefore dismissed.
 |
| Biologics |
Both of these added matter cases demonstrate the dangers of filing a patent application too early, before the relevant clinical product or lead has been finalised. Aside from the sufficiency issues covered above, another risk of filing too early is that, despite the Patentee's best efforts, the specific combination of features that make up the final product may not be clearly disclosed as a combination in the application as filed. This risk is particularly acute in a field such as biologics, where the claimed product may be very complex and comprise multiple elements, as in CAR-T cell therapy products (T 0250/24). Avoiding added matter is also more challenging in crowded fields such as that of therapeutic antibodies, where it is difficult to predict what features will need to be introduced into the claims in order to avoid the prior art (T 1403/24).
Sufficiency (and value) of process and manufacturing claims (Article 83 EPC)
The patent in T 1462/22 (Antibody screening method/CHUGAI) related to an antibody screening process. The patent (EP 2708559), owned by Chugai, particularly claimed a method of producing a pharmaceutical composition comprising an antibody isolated by a particular screening method for identifying antibodies with pharmacokinetically advantageous properties. The claims of the patent were not limited to any particular antibody. The screening method involved finding antibodies with a specific range of pH-dependent binding. The patent was challenged by multiple Opponents on the grounds that the skilled person couldn't carry out the invention.
The Opposition Division (OD) agreed with the Opponents and found the claims insufficient. The Board of Appeal, however, overturned the OD decision. The Board of Appeal particularly noted that the claim was for a method, and the steps, namely binding to a column at pH 7.4, eluting at pH 5.8, and collecting the antibody, were perfectly clear and enabling. The Board of Appeal further noted that the Opponents' arguments about other features, like "improved antibody recycling," were not relevant because these features were not actually included in the claim. The Board of Appeal thus found the claim sufficient and remitted the case for discussion of novelty and inventive step.
The question remains, however, as to the value and enforceability of this process patent compared to a product claim, as in the other antibody manufacturing case T 1913/21 (IPKat). Antibody manufacturing patents can be valuable if they protect a generally applicable method that is adopted by the industry. Given the number of opponents in T 1462/22, the Patentee seems to have identified an industry-relevant screening process. However, process patents can also have significant drawbacks, given that competitors may be able to easily design around process claims or use the method without this being readily apparent to the patentee.
Sufficiency and post-published data (Article 83 EPC)
T 0709/23 (Anti-IL-31 in cats/ZOETIS) is an interesting example of when post-published data can undermine the sufficiency of the invention as disclosed in the application as filed (IPKat). The Patentee, Zoetis, had a patent (EP 3219729) for an antibody to treat itching in cats. The claims were broad, defining the antibody by its function: binding feline IL-31 and inhibiting pSTAT signalling. The original application was based on data from dogs, which the Patentee argued could be credibly extrapolated to cats. The patent's central teaching was that inhibiting pSTAT signalling correlated with a therapeutic reduction in itch.
Unfortunately for the Patentee, their own post-published data submitted during the opposition demonstrated that the lead antibody from the original patent (11E12) did, in fact, inhibit pSTAT signalling but failed to reduce itching in cats. A different antibody (15H05) was shown in the post-published data to be effective at treating itch. However, this antibody was also shown to bind to a completely different and previously unknown epitope. For the Board of Appeal, the new evidence undermined the original application. In particular, for the Board of Appeal, the data demonstrated "that there is no clear correlation between the functional features... and a therapeutic effect of reducing itching in cats" (T 0709/23, r.13). The Board of Appeal thus concluded that a skilled person following the teaching of the original application would be unable to perform the invention across the whole scope of the claim, given that finding the new, working epitope would require a whole new research programme. For the Board of Appeal, this was a more important flaw in the supporting data than the dog versus cat issue.
Inventive step (Article 56 EPC): More than routine?
Another way of protecting biologics is to protect their therapeutic use. However, this can also face obviousness challenges. In T 1505/23 (SIRPalpha signalling blocking antibodies/OSE), OSE Immunotherapeutics' patent (EP 3209691) on anti-SIRPalpha antibodies for treating melanoma or HCC was revoked for lack of inventive step. For the melanoma indication, the closest prior art was identified as a PNAS paper that disclosed the principle. The paper showed that in mice with metastatic melanoma, blocking SIRPalpha signalling (in that case, via a mutation) dramatically improved the effect of an anti-melanoma antibody. The same paper also showed that an anti-SIRPalpha antibody could enhance the killing of breast cancer cells. The Board of Appeal concluded that the problem was simply to provide a reagent to achieve the effect shown in the mouse model, and using an anti-SIRPalpha antibody was an obvious solution suggested by the prior art itself.
For the HCC indication, the closest prior art was identified as a conference abstract showing that anti-SIRPalpha antibodies enhanced macrophage phagocytosis of HCC cells in vitro and in vivo. The abstract concluded it was a "promising strategy". The Board of Appeal found that this gave the skilled person a "reasonable expectation of success" for using the claimed antibodies to treat HCC. The Patentee's attempts to argue that the invention represented a new mechanism of action (MDSC differentiation) were dismissed, as the mechanism was not in the claim.
Inventive step (Article 56 EPC): A surprising technical effect over the whole scope of the claim
Demonstrating inventive step for an antibody or biologic invention can be very difficult in Europe, given the prevailing EPO view that developing new antibodies for a known target is generally a routine matter (IPKat). The Patentee in T 2552/22 (Anti-LAG-3 antibodies) faced this familiar inventive step challenge, with an extra twist from negative post-published data. The patent in this case (EP 2905030) claimed an anti-LAG-3 antibody binding a specific epitope defined by its sequence. The closest prior art was identified as an older antibody (17B4) that bound an adjacent epitope on the very same 30-amino-acid "extra loop". The Patentee argued that their new epitope provided an improved effect, namely better T cell stimulation (IL-2 production).
Unlike the challenges faced in the US to broad functional antibody claims (IPKat), it is possible in Europe to obtain antibody patents based on a newly identified epitope (IPKat). However, the patent in this case was yet another casualty this year of post-published data. The Opponent submitted data showing that another antibody (8B7), which was also disclosed in the patent and also fell within the scope of the claim, did not show this improved effect. In fact, this other antibody was shown to be no better than the prior art. The Board of Appeal was convinced by this data. Since the alleged technical effect was not found to be achieved over the whole scope of the claim, the Board of Appeal found that the purported technical effect could not be considered for inventive step. The technical problem collapsed to simply providing an alternative antibody. The Board of Appeal found that finding an alternative antibody on the same small, known, immunogenic loop was obvious, routine work. The patent was thus revoked.
This case is therefore reminiscent of the sufficiency issues in T 0709/23, which also fell down in view of negative post-published data. The difference being in the present case that the technical effect was not a claimed feature, and was therefore considered under inventive step. Both of these cases are a reminder that applying G 2/21 can go both ways for patentees.
Novelty (Article 54 EPC): The benefit of complexity
The case in T 0358/22 (HER2 Antibody Compositions/GENENTECH) was one in which the complexity of the invention was the Patentee's advantage. The decision also provides an interesting summary of the test for an implicit disclosure, relevant to both novelty and added matter. The patent (EP 2238172) claimed a pharmaceutical formulation of the HER2 antibody pertuzumab which included "a disulfide reduced variant". The Opponent argued this was not novel over a prior art document that also disclosed pertuzumab formulations. The Opponent's case was that the prior art definition of "intact antibody" included "half-antibodies" (HL), that an electropherogram in the prior art implicitly showed this variant; and that it was an "inevitable result" of the general production method disclosed in the prior art.
The Board of Appeal was not persuaded and rejected all three arguments and provided some interesting disclosure on the test for an implicit disclosure for the purposes of a novelty assessment. Firstly, the Board of Appeal found that a mere definition in a document, e.g. of an "intact antibody", was not equivalent to the disclosure of a composition (T 0358/22, r.10). Secondly, the Board of Appeal found that the electropherogram in the prior art was not a "direct and unambiguous" disclosure of the claimed invention. In fact, the Board of Appeal noted that the prior art stated that there was "No evidence of significant product fragments" (T 0358/22, r.13). The Board of Appeal was thus convinced that the skilled person would have had alternative technical explanations for the tiny peak, such as an artifact or a non-glycosylated variant. Finally, the Board of Appeal found that biologic production is complex, and the methods described in the prior art were far too general to inevitably lead to this specific variant. The claim was found novel, and the case was remitted to the OD.
Clarity (Article 84 EPC): Relevance of G 1/24
Finally, T 1351/23 (Blinatumomab/AMGEN) related to the clarity of an antibody second medical use claim. The patent (EP 3134095) concerned the use of a glucocorticoid to prevent adverse events from Amgen's CD19 x CD3 T cell engager (TCE), blinatumomab. T cell engagers are bispecific antibodies that bind to both the target cell (e.g. cancer cells) and cytotoxic T cells, and thereby bring the T cell close to its target. An auxiliary request was challenged for lacking clarity. The Board of Appeal, consulting the description, found however that terms like "prophylaxis" were well-understood and that the dosing regimen ("step-wise") was clearly explained in the description. Unlike in other recent cases, the application of G 1/24 was not found to justify a finding of lack of clarity in view of inconsistent definitions in the specification (IPKat) (T 1351/23, r. 3.2.2). The claim was found to be clear, and the case was remitted to the OD.
Final thoughts
For this Kat, the main theme of the EPO decisions relating to biologics this year has been the adaptation of the law to inventions with ever increasing complexity. In line with the scientific development in the field, the biologics patent cases being considered by the Boards of Appeal now no longer relate to just monoclonal antibodies, and perhaps an update of the Guidelines for Examination to reflect this is needed? (G-II-6. Antibodies). It is clear that the case law for antibodies is applicable to other types of biologic, including cell therapies. Whilst chimeric antigen receptors, T cell engagers and other new forms of biologics are more complex and difficult to develop than traditional monoclonal antibodies, it is clear that the EPO may be as unforgiving with respect to both sufficiency and inventive step. The decision of when to file, and ensuring the application has the right amount and type of data, remains of utmost importance.
The relevance of post-published data is also a theme this year. We are now more than two years on from G 2/21, which confirmed that post-published data may be submitted in support of inventive step (IPKat). The ability to submit post-published data is often considered a patentee favourable approach. However, there were a number of cases this year in which post-published data proved to be the downfall of the patent, as it provided evidence that the invention did not work across the full scope of the claim. This reveals the potential pitfalls of broad functional language, whereby a third party (or in the case of T 0709/23, the Patentee themselves), may invalidate the patent with new data showing that the invention does not work in certain embodiments. We may even see challenges of this type increase, with the increased use of in silico experiments that may lower the burden to screening embodiments of an invention across its full scope. On the other hand, the increased availability of these techniques may enable inventors themselves to bolster their applications with additional and more relevant data. A trend to watch for 2026.
Further reading
- Patenting Antibodies: The epitope claim is dead, long live the epitope claim (Nov 2021)
- US Supreme Court decision in Amgen v Sanofi: The European Perspective (May 2023)
- EPO Board of Appeal maintains functional epitope antibody genus claim (T 1964/18) (Apr 2022)
- When generating antibodies for a target is more than routine (T 0435/20) (Apr 2023)
- Broad functional claiming at the EPO (T 0835/21) (29 Sep 2023)
- UPC takes strong stance on therapeutic antibody inventions (Sanofi v Amgen, UPC_CFI_1/2023) (July 2024)
- Epitope claims stand firm: Board of Appeal upholds functional antibody patent despite insufficiency attack (T 0326/22) (19 Jan 2025)
- Use versus process patents: Implications for novelty and scope (T 1913/21) (April 2025)
- Patentee's own post-published data undermines the credibility of their broad cat antibody patent (T 0709/23) (Oct 2025)