Patents for antibodies defined by discontinuous epitopes have garnered a lot of attention recently. Following the decision of the US Supreme Court in Amgen v Sanofi, it has now been the turn of the European Patent Office (EPO) to consider whether such claims are allowed. It appears that the EPO is to take a much more lenient and patentee-friendly approach to the issue than that adopted in the US, highlighting the significant differences in the approach to biotechnology patents between the two jurisdictions.
The decision has recently been issued in T 326/22, which regular readers of our blogs will recall was highlighted as one to watch in our last annual review of antibody-related EPO case law. The decision upheld the allowability of claims to antibodies defined by a discontinuous epitope, and arguably substantially lowered the bar for them to meet the requirement for sufficiency of disclosure (enablement).
Background to the Case
The case related to antibodies against CD47. CD47 is an innate immune checkpoint protein, sometimes informally referred to as the “don’t eat me” signal. CD47 expressed on the surface of cells interacts with the protein SIRPα on the surface of macrophages and dendritic cells, deactivating phagocytosis. CD47 is expressed in most, if not all, cancers, and is therefore seen as a promising therapeutic target.
The patent underlying the dispute in T 326/22 related to an antibody which binds and inhibits the activity of CD47. The claims defined the antibody by the epitope to which it bound, a conformational epitope comprising 14, non-contiguous amino acids. The claims were interpreted as requiring the antibody to bind all 14 of the specified amino acids. The claims also contained two further functional limitations, requiring the antibody to prevent CD47 from interacting with SIRPα, and not to cause a significant level of agglutination of cells.
The patent disclosed a single antibody which binds the claimed epitope (the “2A1” antibody), which was obtained by immunisation of mice with a CD47 protein, followed by production and screening of hybridomas. The opponent argued that the patent lacked sufficiency of disclosure, because generation of an antibody which bound the specified epitope relied purely on chance. It was also argued that screening for antibodies falling within the claims represented an undue burden. Even if antibodies causing the required inhibition of the CD47/SIRPα interaction without causing cellular agglutination could be identified, the final stage of the screening process required confirmation of the epitope bound by X-ray crystallography, which was argued to constitute an undue burden due to its difficulty and unpredictable nature.
The Decision
The question of enablement of claims to an antibody defined as binding to a discontinuous epitope had previously been considered in T 435/20. In that decision, the Board of Appeal appeared to set out three tests which a patent had to meet in order for a claim to an antibody defined by its epitope to be enabled: (i) a suitable antigen for raising the antibody should be taught; (ii) screening methods to identify the antibody should be described; and (iii) the patent should provide evidence that antibodies binding the claimed epitope can be generated frequently enough and identified reliably enough to guarantee that the skilled person would be able to obtain such an antibody, if they tried (see our discussion of that case here).
The Board in the current case considered the tests set out in T 435/20. The first two tests were considered to be met, because the patent described the antigen used to raise the 2A1 antibody, and the screening methods used to identify it. Notably, the Board agreed that X-ray crystallography would be required as the final screening step, to confirm binding to the specified epitope or otherwise. However, they took the view that identifying the epitope of an antibody by X ray crystallography formed part of the skilled person’s general knowledge, and that it was widely known to be the “gold standard” for this, even if the technique has a high failure rate.
Importantly, they also took the view that the pool of candidates could be narrowed in other ways before applying X-ray crystallography purely as a confirmatory step (rather than needing X-ray crystallography to be applied as a high-throughput screen). For example, the patent taught a cross-competitive assay that could be used to identify antibodies with a high likelihood of binding the same epitope. The patent also showed that of all antibodies screened, only the 2A1 antibody, which bound the claimed epitope, also achieved both the other functional requirements of the claims. Therefore, these functional features could be used to narrow a pool of antibodies to a low number of candidates, before applying X-ray crystallography as a confirmatory step.
However, the Board in the present case did not appear to apply the final test set out in T 435/20. In T 435/20 the Board placed the burden on the patentee to demonstrate that antibodies binding the claimed epitope can be generated frequently enough to guarantee the skilled person success in producing them. This is unusual for the EPO, as normally the burden falls on an opponent to demonstrate that a patent is not enabled. However, the Boards of Appeal have established the principle that, in general, a patent lacks sufficiency of disclosure if the skilled person relies on chance to put it into effect. Given that production of an antibody binding a specific discontinuous epitope is generally reliant on chance, it is arguably consistent with the EPO’s principles for the proprietor to be required to show that the chances of producing such an antibody are high enough not to constitute an undue burden.
In the patent underlying the decision in T 326/22, 73 anti-CD47 antibodies were generated, of which only one (the 2A1 antibody) was found to bind the relevant epitope. On this basis, the opponent argued that the patent did not demonstrate that antibodies binding that epitope could be generated frequently enough to guarantee that the skilled person would be able to produce one. However, the Board rejected this argument out of hand, on the grounds that the opponent had produced no evidence to show that this was the case. They stated that “No reasons are apparent why the skilled person by merely repeating the immunisation protocol of Example 1 as described in the patent application would not reliably arrive at further antibodies that bind to the claimed epitope”. This seemingly shifts the burden of proof on this point from the patentee to the opponent, which is more in line with normal EPO practice on sufficiency.
It is not clear why the Board in T 326/22 took a different approach from the Board in T 435/20 in this respect, though it appears that the argument that the patent did not meet the third test set out in that earlier decision was not a major aspect of the opponent’s case, receiving only a brief mention in the written arguments. One might argue that although the patent only assessed a pool of 73 hybridoma-derived antibodies, techniques are routinely available to provide much larger pools of target-binding antibodies for functional screening.
Summary
It appears that in this case, the Board adopted a considerably lower bar for a patent to provide an enabling disclosure of an antibody against a discontinuous epitope than had previously been set out. It will be interesting to see what approach to the assessment of sufficiency of disclosure is taken by Boards in future cases relating to antibodies defined by their binding to discontinuous epitopes – if the approach taken by the Board in this case is widely adopted it could make it significantly easier to obtain patents with this manner of claims. It will also be interesting to see if a similar approach is adopted by national courts in Europe (including the UPC).
Nonetheless, the decision should not be taken as a green light for claiming antibodies against any discontinuous epitope. The decision highlights the need for the patent to teach straightforward screening methods by which antibodies binding the epitope can be reliably distinguished from those that do not, so that extensive screening by X-ray crystallography is not required. In many cases, functional limitations corresponding to such screening steps may also provide suitable claim language at the EPO.
This decision is also notable for the further evidence it provides of the difference between US and European practice in respect of patents in the Biotech sphere. In the Amgen v Sanofi case the US Supreme Court unanimously found that claims in essentially the same format as those under consideration in the present case lacked enablement. This highlights the potential opportunity for innovators in the antibody sphere to obtain broad protection against competitors in Europe.
Our team has extensive experience in handling patent applications for new antibodies, and defensive and offensive oppositions in this field. If you require any assistance in these areas, we would be delighted to hear from you.
