Centocor v. Abbott Labs. – “Antibody Exception” To Written Description Requirement Under Fire
Abbott Laboratories markets a recombinant human antibody, HUMIRA, as a treatment for rheumatoid arthritis. This antibody binds to a receptor on TNF. NYU and Centocor own US Pat No 7070775 which claims an isolated recombinant anti-TNF-a antibody (Ab) comprising a human constant region, where said Ab competitively inhibit binding of the mouse A2 antibody to human TNF-a and binds to a neutralizing epitope of human TNF-a in vivo with the recited affinity. The A2 mouse antibody is the only species disclosed in the specification, and is not the subject of claim 1.
The string of 13 patent applications that yielded this patent began with an application filed in 1991. The ‘775 patent issued in 2006 and Centocor sued Abbott for infringement. The district court denied Abbott’s motion for JMOL that the specification of the ‘775 patent failed to meet the written description requirement of s.112, and Abbott has appealed to the Federal Circuit. Centocor v. Abbott Labs., No. 2:07CV139-TJW, 2009 U.S. Dist. Lexis 102427 (E.D. Tex. Nov. 4, 2009).
A little bit of history. As pointed out by Lilly in its Amicus Brief supporting Abbott (a PDF is provided at the end of this posting), after years of R&D, antibody-based therapies have come of age: “Rapid growth in the field of antibody research has led to the introduction of nineteen monoclonal antibody [MCA] drugs into the U.S. market since 1994, while only one such MCA drug was approved by the FDA prior to 1994.” When I started prosecuting biotech applications in the early 80’s, MCAs were seen primarily as “research tools,” that might have value in diagnostic assays. If a researcher discovered a new polypeptide that might be useful as a disease marker, broad claims were routinely allowed by the PTO, e.g.: “A purified antibody which specifically reacts with protein x [a possible marker for, e.g., prostate cancer], wherein said antibody does not significantly react with prostate specific antigen.” Put simply, if you knew the structure of the antigen, you could get a claim to any antibody that bound to that antigen. In fact, if you could get a claim to the antigen, the antibody claims pretty much got tossed in gratis. Maybe we have the legal death of In re Durden in 1995 to “blame” for this, but that is a story for another day.
The inventors often did not have a lot of working examples, but it didn’t matter, even after UC v. Lilly and Enzo v. Gen-Probe, because the PTO explicitly recognized an “antibody exception” to the heightened requirements for the written description of biomolecules, and the Fed. Cir. endorsed it, at least in dicta. In Enzo v. Gen-Probe, 323 F.3d 956, 964 (Fed. Cir. 2002), Judge Lourie wrote:
“It is not correct, however, that all functional descriptions of genetic material fail to meet the [WDR]. The PTO has issued Guidelines governing its internal practice for addressing that issue. The Guidelines … are not binding on this court, but may be given judicial notice to the extent they do not conflict with the statute [citation omitted]. In its Guidelines, the PTO has determined that the [WDR] can be met by ‘show[ing] that an invention is complete by disclosure of … functional characteristics when coupled with a known or disclosed correlation between function and structure’… Guidelines 66 Fed. Reg. at 1106… For example, the PTO would find compliance with 112, 1, for a claim to an isolated antibody capable of binding to antigen X, notwithstanding the functional definition of the antibody, in light of the well-defined characteristics for the five classes of antibody, the functional characteristics of antibody binding, and the fact that antibody technology is well-developed and mature.”
The “antibody exception,” as many practitioners have come to reference it, arose, in my opinion because the court did not want the newly minted WDR to lead to challenges of a generation of patents with functional claims to antibodies. Well, the technology is even more well-developed and mature eight years post-Enzo, but Lilly and Abbott don’t feel like cutting it that much slack when it comes to its duties under the WDR post-Ariad. Lilly has filed a convincing – in view of Ariad v. Lilly – amicus brief in support of Abbott’s appeal, arguing that the approval by the Fed. Cir. of the antibody exception is dicta, and that the PTO Guidelines are not in accord with the law. Lilly argues:
“the sequence of one antibody capable of binding to an antigen [The ‘775 patent discloses two Abs with the same variable region that could bind to TNF-a] does not lead to the identification of other antibodies encompassed within these potentially vast generic claims, even using the best current scientific tools. These basic scientific facts should be dispositive of this appeal.… No relevant identifying characteristics, no structure-function relationship, and no representative number of species is disclosed.”
In other words, “no correlation between function – e.g., the binding affinity to TFN-a – and structure –e.g., of a given Ab that binds to TFN-a – equals no valid generic patent claim.” In other words, even if the structure of the antigen is disclosed, this should only be relevant “if there were some universal antibody structure correlation that allows one to relate an antibody’s structure with [to?] its antigen binding function. There is no such correlation….the variable region of Humira [which has little sequence identity with A2] could not be predicted based on the structure of [the ‘775 patent’s] A2 [Ab].” Lilly also notes that that the generic claim to human Abs that bind to human TNF-a is not adequately supported “because the structure of a representative number of species is necessary to support the claims.”
The key word in that last sentence is “structure” and it modifies “representative number.” In other words, if Centocor/NYU had disclosed 100 human polyclonal preparations and monoclonal Abs that bound to human TNF-a and inhibited A2 binding, the specification would still fail to meet the WDR, without a structural characterization of each of them. I don’t think that any of the case law supports a requirement for structural characterization of the representative species. Enzo certainly does not. No institutional inventor is going to be able to leap that hurdle any time soon. Still, Lilly urges that “[a]t the very least [Least?] a completed conception of antibody compounds involves knowledge of the structural parts of the antibody that are critical for its biological function.”
But what if the claim in question is directed to a method involving the therapeutic use of one of a genus of monoclonal antibodies that necessarily vary widely in structure? Your homework assignment is to go back and read In re Alonso, 535 F.3d 1015 (Fed. Cir. 2008), a decision cited repeatedly in the Lilly brief as illustrative of an overreaching antibody claim. Then let’s talk.