Teva v. Eli Lilly: Why This Genus Antibody Patent Survived When Others Have Not
By Prashant Kotian, PhD Researcher (Chemistry), ICT Mumbai
On April 16, 2026, the Federal Circuit reversed a district court and reinstated a jury verdict finding that Eli Lilly's Emgality willfully infringed three Teva patents covering the use of humanized anti-CGRP antagonist antibodies to treat headache. Read next to the Seagen v. Daiichi Sankyo decision from a few months earlier, in which a broadly worded genus claim collapsed for lack of disclosure, this case looks at first like a contradiction. It is not. It is the same legal test drawing a real line between two different kinds of genus claim.
CGRP is a protein that, when it binds to certain receptors, contributes to headache. Antibodies that bind to and block CGRP were already known in mouse form by the time Teva's headache patents were filed in November 2006, and Teva's own specification said so, citing a product catalog that already sold one such murine antibody. Converting a murine antibody into a "humanized" one safe for use in people was, by the patents' own description, something that could be done "by any method known in the art." Teva's specification actually showed only one real humanized example, an antibody called G1, which is the active ingredient in Teva's own Ajovy product. The patent claims, though, are not limited to G1. They cover using any humanized anti-CGRP antagonist antibody to treat headache.
A district court found that too broad to survive, granting Eli Lilly judgment as a matter of law that the patents were invalid for lacking written description and enablement of the full genus of antibodies covered. That ruling leaned on the same reasoning that had already sunk a structurally similar-looking claim in Amgen v. Sanofi, the 2023 Supreme Court case, and again in Baxalta v. Genentech: a patent that claims an entire genus of antibodies defined by what they bind to, for any and all purposes, has to actually teach how to find or build the members of that genus, not just describe what a member would do once found.
The Federal Circuit drew a real distinction. Amgen and Baxalta claimed antibodies themselves, for any use, and the actual unsolved scientific problem in both cases was identifying which antibodies belonged in the genus at all. Teva's claims are method-of-treatment claims: use an antibody, already known to be makeable by already-known techniques, to treat headache. By November 2006, making and humanizing anti-CGRP antagonist antibodies was not the hard, undisclosed part of the invention; it was already routine, a point Eli Lilly itself had argued elsewhere in the same litigation while separately trying to prove different Teva patents obvious. The only thing Teva's specification needed to teach with real detail was that antibodies in this already-known class would treat headache, a narrower and more supportable claim to have actually demonstrated.
Why it matters: The line between a genus patent that survives and one that does not is not about how the claim is worded or how broad it sounds on its face. It is about which part of the invention actually needed disclosing in detail: what the members of the genus are, which is the hard problem Amgen, Baxalta, and Seagen's linker claims all failed to solve on paper, versus what an already-identifiable class of things does, which Teva's specification supported with real evidence. Two patents can use nearly identical genus language and land on opposite sides of that line entirely.
Lilly's liability for willful infringement of Teva's headache patents stands, for now, on remand.
Nothing here is a comment on any other patent's drafting or validity, and nothing above should be read as legal advice. It is a summary of a public, precedential federal court opinion, based on the opinion's own text.
What we do
See if this applies to one of your matters
We run genus/species (Markush) claim coverage analysis, Section 3(d) compliance screening, and prior-art structural triage on real chemical structures — not keyword search. If a case like this one is live for you, we'll run a free sample on a compound of your choosing.