The latest changes to the EPO Guidelines for Examination provide several updates to (i) the patent protection of antibodies, and (ii) the interpretation of “identity” and “similarity” in relation to amino acid or nucleic acid sequences. These updates entered into force on 1 March 2021.
(i) Patent protection of antibodies
A new section – G-II, 5.6 – has been added to the Guidelines in the part defining what is able to be claimed as an “invention” under the EPC. G-II, 5.6 outlines some of the ways antibody inventions may be defined in the claims and also clarifies some specific requirements for the inventiveness of new antibodies.
Notably, this section is additional to the existing EPC requirements that must be satisfied for a claim to be allowed, including novelty, inventive step, and sufficiency of disclosure.
Antibody Definitions
At the EPO, antibodies – including conventional antibodies, recombinant derivatives (fragments, fusions, etc.), and alternative antibody formats (camelid VHHs, VNARs, other sdAbs, etc.) – can in principle be defined:
– Structurally
– Functionally
– Both structurally and functionally
– With reference to the method of manufacture
In practice, however, it can be particularly challenging to obtain patents at the EPO to antibodies defined without including structural features such as the sequences of the complete set of CDRs, or the entire VH and VL domains.
Structural features, or a combination of structural and functional features
When claiming an antibody solely by structure, all the CDRs that interact with the target must be defined in the claims for compliance with Article 84 EPC. For a claim to a conventional antibody, the EPO will usually require six CDRs to be defined in the claim. However, fewer CDRs may be acceptable where it can be experimentally shown that one or more of the CDRs is not essential, or if the claims are directed to antibodies having formats which characteristically employ fewer CDRs (e.g. VHHs, VNAR antibodies) (Guidelines, G-II, 5.6.1.1).
Where a claim does not define the CDRs by their sequence but instead refers to e.g. the CDRs of a particular VH or VL sequence, the numbering scheme used to define the CDRs will typically be required (e.g. Kabat, Chothia, etc.).
It is theoretically permitted to define antibodies by variable domain(s) or CDR sequences having less than 100% sequence identity by combining those structural features with a clear functional limitation (Guidelines, G-II, 5.6.1.4); however the applicant must be able to demonstrate that such claims are properly enabled by the application.
Functional features
The EPO has clarified that an antibody can be functionally defined by the antigen it binds to or a specific epitope on the antigen. When defining the antigen or epitope, closed language (e.g. “consisting of”) should be used without any sequence variability. Otherwise, the claim will lack novelty over any known antibody on the basis that existing antibodies are able to bind to the undefined region of the target (Guidelines, G-II, 5.6.1.2; 5.6.1.6).
If the epitope is non-linear, the specific amino acid residues of the epitope need to be clearly identified, and the method for determining this discontinuous epitope must also be indicated in the claim. Moreover, the application must enable the skilled person to determine whether further antibodies bind this epitope, and enable the production of additional antibodies binding to the same epitope without undue burden (Guidelines, G-II, 5.6.1.6).
If an antibody is claimed exclusively by functional features, it will lack novelty over any prior art disclosing an antibody directed to the same target antigen and obtained using the same immunisation protocol, because the prior art antibody is considered to inherently possess the same functional properties as the claimed antibody.
Furthermore, when defining an antibody exclusively by functional properties, care should be taken to ensure an enabling disclosure is provided across the whole scope claimed and the limits of the claim can clearly be determined by the skilled person (Guidelines G-II, 5.6.1.3). For example, a claimed antibody can be defined by epitope in principle, but the application must also enable (substantially) all other antibodies that bind to the same epitope and thus fall within the claim. If the application does not satisfy this burden of enablement, the claim will not be allowed.
Definition by production process
Antibodies can be defined as a product-by-process. However, for antibodies produced by an immunisation protocol of a non-human animal using a well characterised antigen, the antigen must be specified with 100% identity to a defined sequence for compliance with Article 84 EPC (Guidelines, 5.6.1.5).
Inventive step
For inventive step to be acknowledged for a new antibody that binds to a known antigen for which there exists alternative antibodies, a surprising effect over the existing antibodies must be demonstrated in the application. Examples of a surprising effect may include improved affinity, reduced toxicity or unexpected therapeutic activity.
For antibodies having an improved property over prior art antibodies, the method for determining the property may need to be recited in the claim or indicated by reference to the description. Where the improved property relates to binding affinity, the claims must include the regions that influence binding, including the CDRs and the framework regions.
When arguing for non-obviousness, it is not sufficient that the specific amino acid sequences of the novel antibody are not predictable. In the absence of a surprising or unexpected effect over the prior art antibodies, the EPO will take the view that it would have been obvious to the skilled person to obtain further antibodies to the target antigen using standard techniques known in the art.
Nevertheless, inventive step could be established if the application overcomes technical difficulties in producing or manufacturing the claimed antibodies (Guidelines G-II, 5.6.2).
(ii) Interpretation of terms such as “identity” and “similarity” in relation to amino acid or nucleic acid sequences
Amino acid or nucleic acid sequences can be defined by percent identity to another sequence (i.e. the number of identical residues over a defined length in a given alignment). Amino acid sequences can also be defined by percent similarity to another sequence (e.g. including conserved substitutions of amino acid residues having the same physiochemical properties). Where an algorithm or similarity-scoring matrix for determining percent identity or percent similarity, respectively, is not defined in the application as filed, the broadest interpretation will be applied using any reasonable algorithm or scoring matrix known at the relevant filing date.
For amino acid sequences, if the degree of homology is the only feature distinguishing the subject-matter of a claim from the prior art, then the method of determining or calculating percent homology must be clearly defined in the application as filed to satisfy Article 84 EPC (Guidelines F-IV, 4.24).
If you wish to know more about these updates or would like further advice, please contact Ed Ronan or your usual Boult Wade Tennant LLP advisor.
Find out more about our latest updates to the EPO Guidelines in relation to plants and animals, human-derived cells and methods of treatment by therapy.
