Recent Federal Circuit decisions call into question the value of patents broadly claiming inventions on antibodies and their function in treating debilitating diseases. The decisions in these cases originated in district courts and arguably swept aside the merits of scientific breakthroughs because the inventions claimed were not enabled or were otherwise insufficiently described to justify their broad breadth.* In Teva Pharmaceuticals Int’l GmbH v. Eli Lilly and Company, Appeals 2020-1747, -1748, and -1750 (Fed. Cir. Aug. 16, 2021), the Federal Circuit again dealt with patents broadly claiming antibodies.But here the court confronted Patent Trial and Appeal Board decisions that canceled patent claims as obvious. Specifically, in Teva, the court affirmed Board decisions canceling claims to antibodies that target a central nervous system neurotransmitter, known as “CGRP” and known to cause migraine headaches.
Teva’s challenged patents claim humanized antibodies that antagonize CGRP and, thus, interfere with its ability to cause headaches. An exemplary claim in one of these patents simply recites a “humanized monoclonal anti-CGRP antagonist antibody that preferentially binds to human α-CGRP as compared to amylin.” The claim’s broad breadth likely could be challenged, consistent with the developing precedents,* based on potential frailties in its written description and enablement. But those types of challenges are not permitted before the Board in inter partes reviews. Instead, before the Board, Eli Lilly successfully argued that the broad breadth rendered the claims obvious over the combination of three earlier publications. Two are journal publications from the mid-1990s (and predating Teva’s patent filings, circa 2005). Collectively, they explain the biology and therapeutic potentials when antagonizing CGRP, and report lab studies demonstrating an exemplary antibody could inhibit the activity of CGRP in rats. The third publication is a patent filed in 1990, teaching scientists how to make partially-non-human (humanized) antibodies.
These publications teach, at least in isolation, every aspect of what Teva is broadly claiming. Further, Teva’s patents concede that anti-CGRP antagonist antibodies were known and commercially available before it filed the applications for the challenged patents. It’s not difficult therefore to imagine how these three publications might render obvious the broad claim. But to render a claim obvious, a person ordinarily skilled in this science as of 2005 needed a motivation to combine these teachings and have a reasonable expectation of success. Teva argued neither existed. But based on the deferential standard of review, the court concluded substantial record evidence contradicted Teva’s arguments.
A few points are worth noting. First, Teva complained the Board satisfied itself that the prior art’s suggestion to study or use humanized antibodies was sufficient motivation, whereas Eli Lilly argued that the prior art’s suggestion of therapeutic use of such antibodies in humans was the relevant motivation. The court saw no meaningful distinction, noting that “[c]ommon sense and scientific reality dictate that scientists do not ‘study or use’ humanized antibodies with an end goal of treating diseases in test tubes or in rats.” Slip Op. at 12. Second, Teva also complained that the skilled person would have expected the treatment to be unsafe or ineffective. But the court was unpersuaded because the Board found from the record evidence that the potential safety and efficacy concerns would not discourage the skilled person. Id. at 12–14. Third, the court also rejected Teva’s alternative argument that was premised on its complaint that the Board misread the prior art publications to conclude a motivation to make the antibodies was present when, under Teva’s interpretation it wasn’t. According to the court, the record evidence supported the Board’s interpretation and, in any event, the Board’s “decision to favor one conclusion over the other is the epitome of a decision that must be sustained upon review for substantial evidence.” Id. at 15 (internal quotations and citations omitted).
Teva also argued that the commercial success of two antibody products—its own AJOVY® product and Eli Lilly’s Emgality® product—renders its claims unobvious. Under the law, the Board and court are to accord substantial weight to this commercial success where there is a legally and factually sufficient connection (“nexus”) between the evidence of that success and the claimed invention. The nexus is presumed where the commercially-successful product is coextensive with the claimed invention. The court clarified that the nexus may be presumed even where an unclaimed feature might materially affect the functionality of the allegedly-coextensive product. The Board found and the court acknowledged that such features (i.e., sequence-specific affinity, picomolar binding affinity, full-length antibodies versus fragments, or IgG antibody classes) exist in each product.
A claim to ‘anything that works’ hardly has a nexus to any particular product.
The court explained that “it is not hard to imagine a presumption of nexus between a structurally claimed genus of chemical compounds and a commercial product that meets each claimed limitation.” Id. at 20 (citing Immunex Corp. v. Sandoz Inc., 964 F.3d 1049, 1067–68 (Fed. Cir. 2020), which presumed a nexus because “the patent claims recited the molecular weight and amino acid sequence of the ‘protein’ to which they were directed”). Here, however, the antibodies were not, according to the court, claimed with reference to their structure, but rather with reference to their functional ability to bind CGRP: “A claim to ‘anything that works’ hardly has a nexus to any particular product.” Id. at 21. Because of the claims’ broad breadth and the relevance of unclaimed features contributing to the products’ success, the court concluded that no nexus could be presumed and no nexus exists. And Teva apparently presented no further evidence that the commercial success warranted a conclusion of nonobviousness.
The Board and court also rejected Teva’s argument that its license of the challenged patents, among a license of 188 patents, to a third party is objective evidence of nonobviousness. Ordinarily, the existence of a license is significant because a licensee apparently has chosen to avoid litigation over the patent. But “given that 188 patents were licensed,” said the court, “the nexus between the license and the validity of any particular claim is rather tenuous to say the least.” Id. at 23. In rejecting Teva’s licensing argument, the court also noted “the true purpose of the nexus requirement  is to consider whether the fact-finder can infer that the licensing arose out of recognition and acceptance of the subject matter claimed in the patent.” Id. at 24 (internal quotations and citation omitted). Here, that inference wasn’t possible from the record evidence.
The court’s Teva decision adds to the headaches in broadly claiming antibodies based on their functional abilities to preferentially bind.* In hindsight, maybe the claims could have survived these prior art attacks had they specified “picomolar binding affinity, full-length antibodies versus fragments, or IgG antibody classes”—all features that are apparently critical to the ability of the commercial Teva and Eli Lilly products to function. Id. at 21–22. Maybe recitation of one or more of these features might have supported a different conclusion on whether a nexus to the commercial success is present. But any potential certainty accompanying such recitations may well be fleeting where, in the courts, the enablement and written description requirements are creating other headaches for patentees broadly claiming antibody-based inventions.
* See, e.g., Juno Therapeutics, Inc. v. Kite Pharma, Inc., Appeal 2020-1758, Slip Op. at 9 (Fed. Cir. Aug. 26, 2021) (regarding claims to a chimeric antigen receptor (for a T cell) that includes any single-chain antibody fragments capable of binding any targets); Amgen Inc. v. Sanofi, Aventisub LLC, 987 F.3d 1080 (Fed. Cir. 2021) (regarding claims to antibodies that bind to the PCSK9 protein); see also, Idenix Pharms. LLC v. Gilead Sciences Inc., 941 F.3d 1149 (Fed. Cir. 2019) (regarding a hepatitis C virus treatment method that includes administration of any one of a genus of certain nucleoside compounds).