In the landmark Amgen v. Sanofi case (previously covered here), the Supreme Court affirmed that patent claims drawn to a genus of monoclonal antibodies, which were claimed in terms of their function and the epitope to which they bind, were invalid due to a lack of enablement. This case and other Federal Circuit decisions, including Baxalta v. Genentech and AbbVie v. Janssen, have led to concerns that there may no longer be a way to broadly claim antibodies without limiting claim scope to specific antibody sequences such as the complementarity-determining region (CDR) sequences.
As far as written description is concerned, until 2018, the United States Patent and Trademark Office (USPTO) had an “antibody exception” which allowed fulfillment of the “newly characterized antigen test” as acceptable written description for antibodies because a novel antigen would allow one of ordinary skill in the art to use conventional techniques to make and screen for antibodies that bound to the antigen. Following the Federal Circuit decision in Amgen v. Sanofi, the use of the “newly characterized antigen test” was discontinued.
On the enablement front, the USPTO issued guidelines in 2024 to clarify that the In re Wands factors continue to be the framework for assessing enablement of genus claims, including broad functional claims for antibodies. However, from the USPTO guidelines and the case law, it remains unclear what specific technologies and data would meet the enablement and written description requirements in order to support antibody claims with scope broader than just specific antibody sequences.
A recent Nature Biotechnology article proposed that CDR scanning, a paratope mapping technology, could be used to provide broad protection by scanning antibody CDR regions to delineate the paratope (ie, specific amino acids in the CDRs that contact specific amino acids on the target antigen) as the common structural feature defining an antibody genus that encompasses all permissible substitutions within the CDRs. During CDR scanning, each CDR residue in a particular CDR sequence is individually mutated to all 19 other amino acids and tested to identify those amino acids that permit antibody binding to the target antigen. An exemplary claim based on CDR-scanning data could read as follows:
An antibody comprising a heavy chain variable region (HVR) and light chain variable region (LVR), wherein the HVR comprises an … HCDR3 comprising an amino acid sequence of X1PX2X3X4WX5, wherein:
X1 is an amino acid selected from the group consisting of F, I, L, and Y
X2 is an amino acid selected from the group consisting of R, C, G, and H
Enablement of the genus claim
The proposed CDR-scanning method starts with known CDR sequences from an already characterized antibody and generates and tests actual antibodies, each having a mutation at one of the CDR residues, while the other CDR residues are unmutated. According to the article, the CDR-scanning data may enable the exemplary claim above, since the actual production and testing of produced antibodies are reasonable using modern high-throughput techniques and would not require the skilled artisan to engage in undue experimentation.
While it is true that with modern techniques one can make and test many antibodies more efficiently than before, it remains to be seen whether the USPTO or a court would agree that no undue experimentation is required to use CDR-scanning data as an enabling disclosure. For instance, CDR scanning proposed in the article only provides antibodies where one amino acid position is mutated to the other 19 amino acid residues while the remaining positions in the CDRs are not mutated. However, the exemplary claim above covers numerous combinations of amino acid residues at positions X1 through X5 where any two or more positions could be mutated simultaneously. Thus, the disclosure may not be commensurate with the scope of protection sought by the claim and undue experimentation may still be required to make and test antibodies including thousands of combinations of amino acid mutations.
In addition, it appears each permissive mutation at a given amino acid position can only be a single amino acid residue, even though the claim recites “HCDR3 comprising” the sequence, which is open-ended. In other words, it appears the CDR sequence allowing mutations may need to be fixed in length. The CDR-scanning data likely do not provide enablement of antibodies having additional amino acid residues on one or both ends of a CDR sequence and/or amino acid insertions/deletions in the original CDR sequence (keeping in mind that CDRs are generally about 6 to 20 amino acids in length). This likely outcome is consistent with the well-settled notion that the more one claims, the more one must enable.
Written description for the genus claim
Written description for genus claims must provide sufficient description of either a representative number of species or common structural features that establish a strong correlation between structure and function to show possession of the full scope of the claimed genus. It can be argued that paratope mapping using CDR scanning provides not only a common structural feature, but also data supporting a correlation between the structure and function of the antibodies in the genus. Again, while CDR scanning is useful to show possession of the antibodies actually generated and tested using the technique, it will likely still be difficult to extrapolate written description support for antibodies not generated using CDR scanning, such as for antibodies having multiple mutations and/or amino acid insertions/deletions in the starting CDR sequences.
Cost-benefit analysis: Is the broader scope worth the cost?
While CDR scanning is undoubtedly a useful tool, it comes with a cost which can be significant when there are a large number of parental monoclonal antibodies (mAbs). Typically, many clones of mAbs having different CDR sequences will be generated from immunizing an animal with the same target antigen, even after extensive screening of the antibodies. If CDR scanning is done on each parental mAb individually, the amount of labor, time, and associated cost can add up quickly.
Similar to claims reciting a certain level of amino acid sequence identity to core sequences in the CDRs of a parental antibody, a genus claim drafted based on CDR scanning is also at risk of being “designed around.” This is because the claim scope will depend on the parental antibody selected as the starting point. As discussed above, many clones of mAbs can be generated against the same target antigen or even the same epitope. Thus, a competitor may use a different parental antibody to design around genus claims based on different parental antibody sequences. Even with the CDR scanning tool, patentees may still need to build a “patent fence” by filing on various parental antibody sequences targeting the same antigen (filed simultaneously or over a period of time) and using the CDR-scanning data to support and expand claim scope. Such efforts may be worthwhile if targeting a particular antigen in a highly competitive field.
In view of the above, whether to perform CDR scanning, how much work needs to be done, and when to do it should be evaluated in conjunction with other business considerations, including, for example, available resources, patent application filing timelines, and the competitive landscape.
Takeaways
CDR scanning is a useful tool and can serve as the basis to provide enablement and written description support for claims scope beyond specific antibody or CDR sequences. However, it is unlikely to provide the same scope of protection as a pre-Amgen claiming strategy, which was based on the epitope to which an antibody binds rather than the antibody sequence itself. Given its dependence on parental CDR sequences, companies may perform CDR scanning on various parental antibodies to obtain consensus CDR sequences and to build a “patent fence” to minimize designing around by competitors targeting the same coveted antigen.
[View source.]
