Centocor v. Abbott: The Federal Circuit Struggles to Articulate a Coherent Standard for Compliance with the Lilly Written Description Requirement

I have repeatedly pointed out (on my blog, in a law review article, and in an amicus brief filed in Ariad v. Eli Lilly), that the Lilly Written Description Requirement (LWD) lacks any coherent doctrinal foundation, which has led the courts and PTO to flounder in their attempts to apply the doctrine in a principled and consistent manner. The Federal Circuit's recent decision in Centocor v. Abbott underscores and exemplifies some of the points I've been trying to make.

For example, prior to Centocor the Federal Circuit and PTO had adopted a standard for compliance with LWD for antibodies entirely inconsistent with the standard applied to other biomolecules, including DNA and even other proteins. Strangely, the courts and PTO have refused to acknowledge that antibodies are simply a category of protein, and have never provided any sort of convincing rationale for treating antibodies so differently than other proteins.

Basically, for DNA and other proteins the PTO and courts have held that compliance with LWD requires some adequate level of disclosure of chemical structure, i.e., DNA or amino acid sequence. Although the requirement of structure has not been enforced as stringently as many commentators feared after UC v. Eli Lilly came down in 1997, inadequate disclosure of structure has resulted in claim invalidation in multiple Federal Circuit decisions, such as In re Wallach and Carnegie Mellon University v. Hoffman-La Roche.

However, the PTO has taken the position that a broad genus claim reciting an "isolated antibody capable of binding to [a protein identified as] antigen X” satisfies LWD, even in a case where the specification indicates that not one single antibodies falling within the scope of the claim has never been made, and provides no description of the structural, physical or chemical properties of any antibody falling within the scope of the claim. (See example 13 of the 2008 revised written description training materials; essentially the same example appears in the original 1999 interim guidelines). In other words, an antibody can be purportedly claimed in solely functional terms, without any disclosure of structure for the antibody variable region (the part that binds the antigen), nor even any structural description of the antigen. In Noelle v. Lederman, the Federal Circuit endorsed this approach, characterizing the antibody example in the PTO training materials as "precedent."

In my Ariad v. Eli Lilly amicus brief, I explained why the PTO's attempt to rationalize this dissonant treatment of antibodies under LWD was absolute nonsense as a matter of science.

In Centocor, the Federal Circuit held that a claim reciting a genus of antibodies that bind to a specific epitope of human TNF-alpha was invalid under LWD for failure to disclose sufficient structure. Significantly, the antigen was disclosed, so LWD seemed to be satisfied under the incongruosly permissive approach set forth in Example 13 of the PTO training materials, and endorsed by the Federal Circuit in Noelle. However, the Federal Circuit distinguished the Centocor claim, stating that "while our precedent suggests that written description for certain antibody claims can be satisfied by disclosing a well-characterized antigen, that reasoning applies to disclosure of newly characterized antigens were creation of the claimed antibody is routine." (Emphasis added). The court went on to find that since TNF-alpha was previously characterized, and the production of the claimed antibody "comprising a human constant region" was not routine, the claim failed to satisfy LWD.

I don't necessarily disagree with the court's decision to invalidate the claim, but I do disagree with the court's decision to invalidate the claim under the murky LWD doctrine. This is not an original claim, nothing like it appeared in the priority patent application, and I think traditional written description requirement could have been used to invalidate the claim. The claim also could have been invalidated for lack of enablement, based on the Court's finding that the methodology for producing the claimed antibody was not routine.

Centocor continues the trend of Federal Circuit decisions applying LWD in inconsistent and incoherent manner, lacking any grounding in science, as set forth in greater detail in my Ariad amicus brief. For example, the Centocor panel places great emphasis on the fact that TNF-alpha was not a newly characterized antigen, clearly implying that the requirement for compliance with LWD depends substantially upon whether an invention relates to a newly characterized or previously known biomolecule. But in previous decisions, the Federal Circuit has come to the opposite conclusion, rejecting the argument that the standard for compliance with LWD varies depending upon whether or not the claim involves a newly discovered biomolecule.

For example, in the 2008 Carnegie Mellon University v. Hoffman LaRoche decision, the appellant had argued that LWD’s requirement of structural disclosure only applied to novel DNA sequences, but the Federal Circuit rejected this argument, stating that “nothing in Eli Lilly indicates that that holding was limited to inventions involving novel DNA sequences. Indeed, in University of Rochester, we rejected a similar argument.” (emphasis added)

Another point I have often raised is that the Federal Circuit has failed to articulate any principled distinction between the standard for determining compliance with LWD and the enablement requirement, and this trend continues in Centocor. Recall that the Centocor panel's decision to invalidate the claim under LWD was based on the fact that the antigen had previously been characterized, and the finding that the creation of the claimed antibodies was not "routine." There is nothing in the decision to indicate any distinction between "routine" and "enabled."

For example, the Centocor panel points out that "Centocor simply failed to support its contention that generating fully-human antibodies with the claim properties would be straightforward for a person of ordinary skill in the art given the state of human antibody technology in 1994." (emphasis added). It also found that producing the claimed antibodies "was not possible in 1994
using 'conventional,’ 'routine,' 'well developed and mature' technology."

This is essentially the enablement standard, simply substituting words such as "straightforward," for "without undue experimentation," but failing to articulate any distinction between the level of disclosure necessary to satisfy LWD and enablement.

Neither does the Centocor panel articulate any policy rationale for LWD distinct from that behind the enablement requirement. The panel states that, under LWD, "the scope of Centocor's right to exclude cannot 'overreach the scope of [its] contribution to the field of art as described in the patent specification.’" But this simply paraphrases In re Fisher, a 1970 CCPA decision that held (with respect to the enablement requirement) that the scope of the claim must bear a “reasonable correlation to the scope of enablement provided by the specification to persons of ordinary skill in the art.”

I think it would be better for patent law if the court were to use the enablement requirement to address the patentability of claims such as this, instead of continuing to try to muddle through with the highly flawed LWD.