When Functional Claims Outrun Your Disclosure: Seagen ADC Patent Invalidated under § 112(a)

Three Takeaways

For Applicants: Don't sit on broad early patent applications hoping to later claim whatever competitors develop. If your initial application discloses a large umbrella without specifically identifying the important commercial embodiment you later want to claim, your competitors’ intervening publications can destroy your priority chain, especially in unpredictable technical fields. File new applications promptly after conceiving the most valuable commercial embodiments, and ensure (as best you can) that your earliest applications explicitly discuss preferred sub-genuses alongside longer lists of possibilities.

For Patent Prosecutors: Functional claiming in unpredictable arts carries severe enablement risk post-Amgen. When claims define inventions by what they do (here, "intracellularly cleaves") rather than what they are, and the specification provides no common structural or chemical quality predicting which embodiments will satisfy that function, you're inviting risk. Draft claims around specific structures supported by working examples. If you must include functional limitations, the specification needs to teach a predictive principle: some quality running through all functional embodiments that lets skilled artisans select candidates without assay-by-assay trial and error. "Test it and see" is not adequate guidance when millions of candidates exist.

For Litigators: Inventor testimony about conception timing can be outcome-determinative on written description. The inventor’s admission that he never saw a Gly/Phe-only tetrapeptide until eleven years after the priority date gave the Federal Circuit a clean basis for reversal. In discovery and at deposition, probe rigorously: When did the inventors actually conceive of the specific claimed embodiments? Did they recognize the claimed subgenus as distinct before seeing the accused product? Admissions that conception followed the accused infringer's public disclosure can convert a jury loss into a JMOL win on appeal.

The Decision

In Seagen Inc. v. Daiichi Sankyo Company, Ltd. (Fed. Cir. Dec. 2, 2025), the Federal Circuit reversed a jury verdict and district court judgment, invalidating claims 1-5, 9, and 10 of Seagen's U.S. Patent No. 10,808,039 under 35 U.S.C. § 112(a). The panel, Lourie (writing), Reyna, and Chen, held that the patent's 2004 priority application could not support claims to a specific subgenus of 81 Gly/Phe-only tetrapeptide linker units when that application disclosed over 47 million possible tetrapeptide variants without any guidance pointing to the claimed species. The court also extended Amgen v. Sanofi to hold that functional claim limitations covering an enormous universe of possible drug moieties, combined with unpredictable ADC science, render claims unenabled.

The decision wipes out a $41.8 million damages award plus an 8% running royalty. More significantly, it represents a continued tightening of both written description and enablement doctrine in biotechnology patent law.

The Technology and Timeline

The '039 patent claims antibody-drug conjugates (ADCs), cancer therapeutics that combine a monoclonal antibody, a cytotoxic drug payload, and a linker connecting them. The claimed innovation specifies that the linker contain a "tetrapeptide" composed exclusively of glycine and phenylalanine in any combination. Because each of four amino acid positions can be either amino acid in multiple stereoisomeric forms, this "Gly/Phe-only" specification encompasses exactly 81 different molecular species.

The '039 patent claims priority to a November 2004 provisional application that describes ADCs with linker peptides potentially including dipeptides through dodecapeptides with 39 different possible amino acids, generating over 47 million possible tetrapeptide combinations. Daiichi Sankyo independently developed Enhertu® (DS-8201), containing the Gly-Gly-Phe-Gly tetrapeptide sequence, in 2011. Enhertu's structure became publicly known in December 2015.

Seagen did not file the '039 patent application until July 2019, four years after Enhertu's public disclosure.

Written Description: Finding 81 Trees in a Forest of 47 Million

The court's written description analysis applies the familiar "blaze marks" framework from In re Ruschig (CCPA 1967). When a specification discloses an enormous genus, claims to a specific species require that the specification provide "reasonably specific supporting disclosure,” i.e., blaze marks directing skilled artisans to the particular claimed species.

Seagen argued that its 2004 application disclosed tetrapeptides generally, listed glycine and phenylalanine among possible amino acids, and exemplified GFLG (glycine-phenylalanine-leucine-glycine). This, Seagen contended, was sufficient under In re Driscoll, where the CCPA upheld written description when an earlier application recited 14 specific moieties and later claims selected one.

The Federal Circuit rejected this analogy. In Driscoll, the 14 moieties were explicitly enumerated, and the claimed species was "clearly discernible in the generalized formula." Here, the 2004 application never mentions Gly/Phe-only tetrapeptides. The 81 claimed species represent "merely an infinitesimal fraction" (approximately 0.0002%) of 47 million possibilities, with nothing in the specification directing attention to that particular combination.

The court found Seagen's expert testimony "self-defeating." Dr. Bertozzi testified that a skilled artisan could make a "straightforward leap" from the disclosed GFLG tetrapeptide to an all-glycine-and-phenylalanine combination. But as the court observed, if a leap is required, the destination cannot already be disclosed. Written description demands that the inventor demonstrate actual possession, not that skilled artisans could predict modifications the inventor might have envisioned.

Perhaps most damaging: lead inventor Dr. Senter admitted that "the first time... [he] ever saw a G/F-only tetrapeptide in an ADC... was in Daiichi Sankyo's Enhertu," publicly disclosed in 2015, eleven years after the 2004 priority application. The court quoted the foundational principle: "one cannot describe what one has not conceived."

Enablement: Extending Amgen to ADC Linker Patents

The enablement analysis builds directly on the Supreme Court's 2023 decision in Amgen Inc. v. Sanofi. The '039 patent claims require that the drug moiety "intracellularly cleave, "a functional limitation. Under the district court's claim construction, "D is a drug moiety" encompasses any drug (not limited to dolastatin/auristatin types), creating claims covering essentially unlimited drugs meeting a functional criterion.

This parallels Amgen, where antibody claims defined by functional binding and blocking properties covered millions of structurally diverse embodiments. The Supreme Court held that such claims require either (1) disclosure of a common quality running through all embodiments enabling selection without trial-and-error, or (2) sufficient working examples that reasonable experimentation will enable the full scope.

The '039 patent specification satisfies neither requirement. It does not disclose any common structural or chemical quality predicting which drug moieties will intracellularly cleave. Inventor Dr. Kline testified that assay testing is necessary to determine whether a particular drug moiety meets the functional limitation. Without a predictive principle, skilled artisans face precisely the "trial-and-error discovery" that Amgen forbids.

The unpredictability of ADC science compounds the problem. The court emphasized that "a skilled artisan would be required to use an assay to test whether any given ADC with a given drug moiety meets that functional limitation." When millions of candidates must be tested with no reliable method for selecting likely successes, that constitutes undue experimentation—even if individual assays are routine.

The Bigger Picture

Seagen v. Daiichi Sankyo joins Amgen, Novozymes v. DuPont, and Gilead in establishing doctrinal barriers to broad genus claims. Especially in chemical and biotechnology fields, the Court continues to signal claims must be supported by specific disclosures demonstrating what the inventors actually possessed, not by vast genera from which claimed species might be derived.