Teva v. Eli Lilly — Federal Circuit Revives Teva’s Migraine-Drug Patents Against Lilly

Background

Teva sued Eli Lilly in 2018 in the District of Massachusetts, claiming that Lilly’s migraine drug Emgality (galcanezumab) infringed three Teva patents covering methods of treating headache by administering a “humanized anti-CGRP antagonist antibody.” CGRP (calcitonin gene-related peptide) is a protein that triggers the blood-vessel dilation associated with migraine; antibodies that block CGRP are now a leading class of migraine medications. Teva’s own product in the same class, Ajovy (fremanezumab), also targets CGRP.

At trial in 2022, a jury found that Lilly willfully infringed and rejected Lilly’s invalidity defenses. But the district court (Judge Allison Burroughs) then granted Lilly judgment as a matter of law (JMOL), holding the asserted claims invalid for failing both the written-description and enablement requirements of 35 U.S.C. § 112. The court reasoned that because the genus of possible humanized anti-CGRP antibodies was enormous and Teva’s specification disclosed only one such antibody (“G1,” the active ingredient in Ajovy), the patents improperly claimed more than they taught. Teva appealed.

The appeal took place in the shadow of Amgen v. Sanofi, 598 U.S. 594 (2023), in which the Supreme Court invalidated antibody claims directed to an entire functional genus of PCSK9-binding antibodies because the specification left skilled artisans to do the “research assignment” of finding them. Lilly argued that Teva’s claims suffered the same defect.

The Court’s Holding

In a unanimous precedential opinion by Chief Judge Prost, the Federal Circuit reversed the JMOL and remanded, reinstating the jury’s verdict. The panel drew a careful line between antibody-genus claims and method-of-use claims. Teva’s asserted claims, the court emphasized, did not claim the antibodies themselves — they claimed only the use of humanized anti-CGRP antagonist antibodies to treat headache.

On written description, the court held that a reasonable jury could have found Teva’s disclosure sufficient. The specification was allowed to rely on known building blocks: anti-CGRP antagonist antibodies (in murine form) were well known in the prior art — Lilly itself had taken that position in earlier IPR proceedings — and humanization was a routine technique by the 2006 priority date. Against that background, disclosing one humanized example (G1) plus the structural and functional traits needed for the method of treatment was “sufficiently representative” of the claimed genus. The court rejected the view that AbbVie Deutschland v. Janssen imposes a bright-line requirement that the specification disclose a composition “structurally similar to” the accused product.

On enablement, the court similarly concluded that JMOL was improper. The panel distinguished Amgen and Baxalta v. Genentech: those cases involved claims to the antibody genus itself, which forced the patentee to teach how to make the full scope of those antibodies. Teva’s claims, by contrast, covered only the method of treating headache. Because a reasonable jury could have found that all humanized anti-CGRP antagonist antibodies treat headache, there was no open-ended “research assignment” left for skilled artisans — any such antibody would work in the claimed method. The court distinguished Idenix v. Gilead on the same basis: there, not every candidate compound was effective, so screening was necessary; here, all candidates were.

Key Takeaways

• Method-of-use claims get more breathing room than composition claims. After Amgen, antibody-genus composition claims face very tough §112 scrutiny. Teva shows that the analysis shifts when the claim is to a method of using a known or easily made class of antibodies for a specific therapeutic purpose.
• Known-genus building blocks count. When the underlying class (here, anti-CGRP antagonist antibodies) and the processing technique (humanization) are already routine in the art, the specification need not re-teach them. Prior-art knowledge fills gaps in the disclosure.
• “Research assignment” analysis is claim-specific. The enablement question turns on what skilled artisans must do to practice the claim — not to map the entire underlying chemical space. If every member of the relevant subset works, screening effort doesn’t create an undue-experimentation problem.
• AbbVie is not a bright-line structural-similarity rule. The Federal Circuit clarified that AbbVie‘s language about antibodies “structurally similar to” the accused product is fact-bound, not a categorical requirement for antibody cases.
• Jury verdicts on §112 get real deference on JMOL. The panel repeatedly emphasized that substantial evidence supported findings underlying the verdict, and that the clear-and-convincing burden belongs to the challenger on JMOL review.

Why It Matters

This decision is a significant win for biologics patent holders and a roadmap for how to draft antibody-adjacent patents after Amgen. By focusing on method-of-use claims backed by a known antibody class, Teva preserved enforceable patent coverage where pure composition claims would have failed. Expect more life-science patentees to structure their portfolios around therapeutic-use claims tied to well-characterized prior-art antibody classes, rather than broad functional genus claims that courts have grown skeptical of.

For the migraine market specifically, the ruling revives Teva’s infringement case against Lilly’s Emgality — a product with hundreds of millions of dollars in annual sales — and sends the dispute back to the district court for further proceedings consistent with the restored jury verdict. More broadly, the opinion gives the biotechnology industry a much-needed signal that Amgen does not sweep away all functional-antibody patenting, and that careful claim architecture can still deliver robust protection for novel therapeutic uses.