May 2023 T0435/20 demonstrates the limit of obviousness of the “standard” procedure of generating antibodies for peptide epitopes. T0435/20 also serves to teach us that non-routine methods for generating antibodies for rare or “challenging” epitopes must be fully disclosed and apply to all of the antibodies that are claimed.
In laboratory practice the production of new antibodies is now often considered routine with a desired epitope synthesised and used to raise antibodies to that specific sequence. Accordingly under European practice the EPO views generating antibodies against any known target as being routine. Thus antibodies for such targets are claimed in functional terms. This contrasts markedly with practice in the US where such functional claims are not considered sufficient.
The idea that an antibody can be raised to any epitope without challenging the skilled person is in turn challenged in decision T0435/20.
In this case the target epitope is a group of collocated amino acids within the target protein. However in this case the collated amino acids are not arranged linearly along the peptide chain but are only brought into proximity to one another by the overall 3D structure of the protein. Thus in this case, the epitope is dependent on at least the secondary structure of the protein rather than its primary sequence. Such epitopes are called “conformational” epitopes as indicating the presence of particular protein conformations in the target protein.
The production of antibodies to such conformational epitopes is challenging. This is because a conformational epitope comprised in a folded protein is merely one target epitope amongst a great many. Accordingly, the probability of the immune system recognising a particular epitope and generating an antibody thereto is proportionality lower.
The generation of antibodies to linear epitopes is significantly simpler as the linear epitope can be synthesised and purified to yield, effectively, a single target to which an immune system can react. Such simplification of the problem is much more challenging in the case of conformational epitopes because changing the protein by truncation or other structural changes has a very high probability of affecting the native conformational epitope that is the target of the exercise. Thus the complexity of the screening step to find an antibody that binds a conformational epitope of the target is also greater than for a “standard” epitope.
In the patent considered in T0435/20 (EP1931710) a genus of anti-IL23 antibodies that bind to a defined conformational/discontinuous epitope in IL23 is claimed, i.e. “an epitope comprising residues 82-95 and residues 133-140 of SEQ ID No: 29”.
Thus an epitope is defined and an antibody falling within the scope of the claims is also disclosed. However, for the claim to be sufficient the skilled person must be able to work the invention for all embodiments of the invention having this feature. However, the EPO Technical Boards of Appeal (TBA) in this case considered that the disclosure in the application as filed did not support this level of sufficiency.
In particular, while the application included data demonstrating that the example antibody binds the discontinuous epitope with the region defined in the claim, the broader language meant that the claim was not limited to only this antibody. As noted above, for sufficiency to be demonstrated the skilled person must be able to demonstrate further antibodies for epitopes fulling within the scope of the claim.
However, the TBA considered that the description was insufficient to allow the skilled person to do this. In particular, the TBA noted that there was no guarantee that a single antibody in a pool of antibodies raised against IL23 would bind the same epitope as that of the claimed antibody. Thus there was no guarantee that the skilled person could successfully produce an antibody falling within the scope of the claims. A further difficulty presented to the skilled person would be measuring or testing whether the antibodies produced had the functional characteristic to bind the IL23 target at all, still less whether they could bind the specifically claimed epitope regions.
Accordingly, the TBA considered that the skilled person was presented with an undue technical burden when using the patent specification to attempt reproduce its results. Therefore, without a disclosure of the screening process and how to successfully achieve generation of a conformational epitope the patent was insufficient.
In particular, the TBA noted at paragraphs 58 and 59 of their reasons:
In summary, given the lack of relevant guidance in the patent or in the common general knowledge, the skilled person attempting to carry out the claimed invention is confronted with having to develop an elaborate screening strategy, without a reasonable expectation of success. Indeed such a screening strategy relies on chance, without the skilled person having any knowledge of the likelihood of success.
Finally, if after such a screening process, the antibody taken forward for epitope determination does not have the required specificity, i.e. in case of failure, neither the patent nor the common general knowledge provide adequate information regarding what should be changed or how to guarantee success.
T0435/20, r. 58-59
In other words, production of an antibody is necessarily dependent on the size and complexity of the pool of epitopes that is presented to an immune system. The challenged immune system necessarily produces a large number of different antibodies, and these antibodies are normally screened to extract the most advantageous one. Thus isolating the desired antibody from the pool of antibodies generated is a challenge that is overcome by the use of a screening system. However, in the present application no suitable screening system was disclosed. This inadequacy meant that the TBA had serious doubts as to whether a skilled person would be able to perform the invention over the whole of the scope of the claim.
Accordingly, the patent was considered insufficient and the appeal dismissed with the consequence that the patent remains revoked.
Thus this case contrasts markedly with “standard” European practice regarding claims for antibodies. Generation of “challenging” antibodies must be supported by an enabled screening system in order to yield a reproducible method of producing such challenging antibodies. Such a technique is likely to be non-routine. Thus, in supporting the sufficiency of such antibodies a case is also built for the inventive step of such antibodies because the production of antibodies would not be obvious. As noted above, European practice effectively says that the production of an antibody for a given linear epitope is obvious. This case appears to show that there are non-obvious antibodies because of the challenge in providing and generating antibodies to particular epitopes.
Thus, T0435/20 serves to demonstrate the limit of obviousness with respect to generating antibodies from defined epitopes and also serves to teach us that non-routine methods required to overcome challenging circumstances in epitope generation must be disclosed and apply to all of the embodiments that are claimed.
