While the approaches of the USPTO and EPO to antibody patents may differ in some respects, there are some similar approaches that may assist with prosecuting in both jurisdictions. Patrick McManamny discusses.
When monoclonal antibodies were first produced in 1974, Kohler and Milstein predicted that they “could be valuable for medical and industrial use”. However, it took 11 years for the first therapeutic monoclonal antibody to be approved. From then until 2012, 30 more monoclonal antibody therapies have been approved, with current estimates of 500 in clinical trials and 50 more predicted to enter clinical trials each year. The global market for therapeutic antibodies has been estimated at $44.6 billion in 2011 and is expected to rise at a compound annual growth rate of about 5.3% to nearly $58 billion in 2016. Clearly, the therapeutic monoclonal antibody area is big business.
As with all drugs, patent protection is an important component of commercial success. However, patents relating to antibodies are treated differently to some other compounds (e.g., small molecule drugs) and are also treated differently in different jurisdictions. This article touches on some of the important similarities and differences in approaches of USA and Europe to antibody patents, and what you may face if you propose to take a patent application directed to an antibody into prosecution or litigation in either region.
Both the US and Europe (together with many other countries) will grant a patent for an antibody to a new antigen, as long as the antigen is structurally described (e.g., by amino acid sequence for a protein-based antigen) and some function has been ascribed to the antigen (and, preferably, the antibody). Such a patent can be grated even if no antibody has actually been produced.
The recent litigation between Eli Lilly and Human Genome Sciences in UK over a patent covering antibodies to BAFF (encompassing HGS’ Benlysta® and Lilly’s tabalumab) has shown that such patents can be upheld even on the basis for a predicted function for the antigen and the antibody.
Both jurisdictions also appear to be willing to grant a patent to any antibody that binds to a previously unidentified epitope in an antigen based on data from only one or two antibodies.
However, following the human genome project, there are few new antigens remaining to be discovered, meaning that most recently filed patents are somewhat narrower in scope (e.g., defining specific antibodies or classes of antibodies), and this is where things become more complicated.
The recent decision in Centocor Ortho Biotech, Inc. v. Abbott Laboratories (Fed. Cir. 2011), suggests that the US approach to granting patents claiming antibodies based on a structurally characterised antigen only applies to “newly-characterised antigens wherein creation of the antibodies is routine”. Thus, this approach may not apply to complex or poorly characterised epitopes (e.g., post-translational modifications), or to claims defining antibodies by functional characteristics. As Centocor discovered in this litigation, the US Court of Appeals for the Federal Circuit also believes that this approach does not apply to antibodies characterised by both an epitope and a series of functional characteristics (e.g., the type of antibody (“fully-human”) and binding affinity), unless the patentee has demonstrated possession of the claimed class of antibodies (e.g., by actually making the antibodies). Thus, the USPTO and US Courts appear willing to grant broad patents with less data required for narrower claims. While this may appear counter to standard practice, it fits with the approach of determining how “routine” it is to make the antibody, i.e., the more limitations are introduced into the claims, the more complex it is to make something satisfying the claim. Of course, actually producing several antibodies within the scope of the claims can show that you are in possession of the class of antibodies claimed. Furthermore, defining your antibody in structural terms (by its sequence) is likely to avoid this issue, since you are generally considered to be in possession of the specific antibody reduced to practice.
The decision in Centocor v Abbott may mean that it is more difficult to obtain valid claims defining an antibody in purely function claims based on a single example.
Consistent with the decision in Centocor v Abbott, standard USPTO practice is to resist granting claims defining an antibody by the sequence of its V regions or complementarily determining regions (CDRs), where the claim defines variation in the sequence (e.g., by defining a percentage identity or a consensus sequence). This approach is based on the USPTO’s opinion that any variation in a CDR can significantly affect the function of an antibody. While claims defining variation in CDRs can be obtained in some circumstances, they generally require substantial data showing which residues can be changed and in which manner without affecting the defined function of the antibody (i.e., a “structure-function relationship”). Pursuing such claims can also lead to a risk that arguing that it is straightforward to make the class of antibodies you are claiming can be used against you later in alleging the class is obvious over similar prior art antibodies.
Consistent with the USPTO’s approach that even single changes in a CDR can dramatically alter the function of an antibody, it is generally straightforward to obtain claims defining a new antibody by all six CDRs or one or both V regions. This is because the USPTO is generally willing to concede structural non-obviousness (i.e., if the CDRs are new, the antibody is novel and non-obvious). Thus, if you know your commercial embodiment and are willing to settle for claims specifically defining that product, it may be relatively straightforward to obtain a granted US patent, which can be important for partnering/licensing or blocking a biosimilar. This focussed approach can also avoid putting arguments on the record which may be used against you later (see related article).
In contrast to the structural non-obviousness approach adopted by the USPTO, the EPO tends to take a claim to an antibody is inventive if the antibody is produced by routine methods and shows unexpected properties (T645/02). The flip side of this statement is that if the antibody is produced by routine methods but does not show an unexpected property, then it is obvious. This analysis ignores any differences in sequence between your antibody and the antibody of the prior art. It also raises difficulties when faced with prior art that describes an antibody prophetically (i.e., without actually producing it), to which you cannot compare to show an “unexpected property”.
To deal with the EPO’s approach to obviousness, it is important to anticipate this form of objection when drafting a specification. Such objections may be based on a prior art reference of which you were unaware or a prior art antibody that has a partially overlapping function or was not characterised in the manner your antibody was, which you considered irrelevant when drafting. When drafting a specification you could include description of multiple properties of your antibodies (e.g., specificity, cross-reactivity, neutralising potency, stability, etc). It is also important to describe as many desirable characteristics as you can when drafting, since it can be difficult to rely on post-filing data to overcome an allegation of lack of inventive step in Europe unless the new data is “plausible” based on the specification as filed.
Another option to overcome an allegation of obviousness by the EPO is to show that while you may have used standard methods to produce the antibody, you had to overcome one or more difficulties or proceed in a counter-intuitive manner. Such an approach may require you to limit the claims to the one (or few) specific antibody actually produced by that method.
Pursuing claims to a group of antibodies defined in functional terms (as opposed to all six CDRs or both V regions) in Europe can be even more difficult that claims defining a single antibody. This is because you may be required to show that you can produce several antibodies falling within the class (two may not be enough) and that all of those antibodies have “unexpected properties”.
While the approaches of the USPTO and EPO to antibody patents may differ, there are some approaches that may assist with prosecuting in both jurisdictions:
In the US, it is also desirable to include some form of structure in your claims (e.g., the sequence of the epitope and/or the sequence of the CDRs/V region(s) of your antibody).