SKIP TO CONTENT
Teva v. Lilly: The “Well-Known Genus Used as Part of a Different Invention” Doctrine Gets a Federal Circuit Revival
On April 16, 2026, the Federal Circuit reversed a District of Massachusetts JMOL that had erased a $176.5 million jury verdict in Teva Pharmaceuticals International GmbH v. Eli Lilly and Company, restoring Teva’s infringement win against Lilly’s Emgality (galcanezumab) and, along the way, dusting off a CCPA-era written description doctrine that has not been front-and-center in this court’s § 112 ¶ 1 jurisprudence for years. The principle, stated simply, is this: when the genus at issue was well understood in the art before the priority date and is not itself what the applicant invented, the specification does not have to exhaustively describe that genus.
That principle applies with equal force to therapeutic method of use claims built on known target classes to platform technologies whose inventive contribution lies in the method rather than in the universe of targets to which the method can be applied. For life-sciences counsel reading Ariad’s progeny with increasing discomfort — Amgen v. Sanofi on enablement, Baxalta on antibody genera, and Idenix, AbbVie, and Juno each reinforcing the demand for deeper species disclosure for written description — Teva is not a transformation. It is, however, a meaningful counter-weight in the recurring fact pattern where the operative compound class, target class, or substrate class pre-exists the claimed invention and the claimed invention is something else. This article unpacks what the decision does, what it does not do, and what it may change for prosecution, litigation, and portfolio management.
The Decision in Brief
Teva’s asserted claims — the representative ones being claim 17 and claim 30 of U.S. Patent No. 8,586,045 — are method-of-treatment claims. Claim 17 recites:
A method for reducing incidence of or treating headache in a human, comprising administering to the human an effective amount of an anti-CGRP antagonist antibody, wherein said anti-CGRP antagonist antibody is a human monoclonal antibody or a humanized monoclonal antibody.
Claim 30, which depends from claim 17, narrows the antibody to a humanized monoclonal:
The method of claim 17, wherein said anti-CGRP antagonist antibody is a humanized monoclonal antibody.
A Massachusetts jury found the claims infringed and not invalid. Judge Burroughs then entered JMOL of invalidity on both § 112 ¶ 1 grounds. On written description she held that the asserted claims reached a new, functionally-defined genus of humanized anti-CGRP antagonist antibodies; that the specification disclosed only one species within the scope of the claims (G1, with the 84 disclosed variants excluded as fragments or unproven antagonists and the murine antibodies excluded as unhumanized); that the accused galcanezumab differed from G1 in multiple structurally significant respects (different epitope, ~50–65% sequence identity, different V-gene family, different CDR sequences and lengths); and that the structural features Teva advanced as common to the genus were generic to full-length antibodies. Reading AbbVie, Juno, and Regents of the Univ. of Cal. v. Eli Lilly as a continuous line, she concluded that the disclosure amounted to an impermissible “research plan, leaving it to others to explore the unknown contours of the claimed genus.” Teva Pharms. Int’l GmbH v. Eli Lilly & Co., 2023 WL 6282898, at *19 (D. Mass. Sept. 26, 2023) (quoting AbbVie, 759 F.3d at 1300).
On enablement she applied Wands in light of Amgen v. Sanofi and Baxalta and held — critically — that even assuming a reasonable jury could have found that humanized anti-CGRP antagonist antibodies as a class would treat headache, that finding did not save the claims under Idenix, because each additional candidate antibody would still have to be synthesized and screened to confirm operability, and the in vitro / in vivo / antibody-acquisition / humanization pathway required months and tens of thousands of dollars per antibody. Id. at *23–*24.
The Federal Circuit, applying First Circuit JMOL standards de novo, reversed on both grounds and reinstated the verdict on liability and the $176.5 million damages award.
The “Different Invention” Lineage
The panel traces this doctrine though four cases, treating them as continuous rather than distinct rules.
The origin point is In re Fuetterer, 319 F.2d 259 (CCPA 1963) — a plurality opinion by Judge Rich. The claim was to a composition for tire-tread stock, one component of which was “an inorganic salt that is capable of holding a mixture . . . in colloidal suspension in water.” 319 F.2d at 260–61. The specification disclosed only four such salts. Id. at 265. The Patent Office rejected the claims as unduly functional and unduly broad. Judge Rich reversed with a passage that has done most of the work in every subsequent case of this lineage:
Appellant’s invention is the combination claimed and not the discovery that certain inorganic salts have colloid suspending properties. We see nothing in patent law which requires appellant to discover which of all those salts have such properties and which will function properly in his combination.
Fuetterer, 319 F.2d at 265–66. Fuetterer‘s plurality status matters for strict stare decisis but is mitigated by its later adoption; the Teva panel candidly acknowledges that “only two judges of the five-judge panel joined it (a third concurred in the result),” slip op. at 13 n.9, but treats it as persuasive authority carried forward through binding CCPA precedent.
That adoption came in In re Herschler, 591 F.2d 693 (CCPA 1979), where Judge Baldwin sustained a generic claim to “physiologically active steroidal agents” delivered transdermally with DMSO even though the great-grandparent application had disclosed only a single corticosteroid working example. Herschler distinguished claims drawn to “novel ‘steroidal agents’” (where species-level disclosure would be essential) from claims drawn to “the Use of Known chemical compounds in a manner auxiliary to the invention,” which need be only “so specific as to lead one having ordinary skill in the art to that class of compounds.” 591 F.2d at 702. Crucially, Herschler relied on the Fuettererplurality as “the line first clearly drawn.” Id. at 701.
Ajinomoto Co. v. ITC, 932 F.3d 1342 (Fed. Cir. 2019), brought the same logic forward into the genomic era. The claim covered engineered E. coli producing aromatic amino acids via a yddG gene under the control of a “more potent promoter”; the specification disclosed four named potent promoters (PL, lac, trp, trc) and cited a 1986 Deuschle article disclosing fourteen more. CJ CheilJedang Corp. argued the claimed genus was inadequately described. The Federal Circuit affirmed the Commission’s rejection of that challenge, explaining that “the genus of more potent promoters was already well explored in the relevant art by the time of the ’655 patent’s invention” and that the invention “was identifying the yddG gene . . . not the well-known techniques for performing the amplification or expression enhancement.” 932 F.3d at 1359.
Finally, the panel credits retired Judge Bryson’s opinion in Erfindergemeinschaft UroPep GbR v. Eli Lilly & Co., 276 F. Supp. 3d 629 (E.D. Tex. 2017) (Bryson, J., sitting by designation), as distilling the principle:
When a genus is well understood in the art and not itself the invention but is instead a component of the claim, background knowledge may provide the necessary support.
UroPep, 276 F. Supp. 3d at 648.
Juno and the Unresolved Tension
Lilly’s principal counter-citation was Juno Therapeutics, Inc. v. Kite Pharma, Inc., 10 F.4th 1330 (Fed. Cir. 2021), and the interplay between Juno and Teva is worth a moment of reflection because the two cases touch the same doctrinal nerve. Ariad established two independent disclosure pathways for satisfying the written description requirement as to a genus claim: disclosure of a representative number of species within the genus, or disclosure of structural features common to the members of the genus that permit a skilled artisan to visualize or recognize members of the genus. 598 F.3d at 1350. Either suffices and a patentee need not satisfy both.
Juno involved a three-part CAR-T construct claim in which the CD3ζ intracellular domain combined with a CD28 costimulatory region formed the “backbone” — characterized at trial as “the birth of the CAR-T field,” 10 F.4th at 1335 (J.A. 32976) — and a single-chain antibody fragment (“scFv”) served as the binding element. The backbone was the inventive contribution. The scFv was drawn from a well-known class; Juno argued precisely that, and the panel acknowledged the point in crediting Dr. Sadelain’s testimony that scFvs were “not new in the art” and that “scFvs in general were well-known.” Id. at 1339, 1341.
The Juno panel rejected the argument. Working on Ariad’s structural-features prong, the panel held that the ’190 specification’s disclosure of two scFvs — one derived from the SJ25C1 antibody, one from the J591 antibody, with no amino acid sequences disclosed for either — failed to describe the vast structural diversity of the scFv genus. Id. at 1341–42. The panel’s sharpest statement for present purposes sits at page 1343: “The test is the same whether the claim is to a novel compound or a novel combination of known elements.”
Teva distinguishes Juno in footnote 16 on the ground that the portion of Juno Lilly relied on addressed structural features rather than representativeness, and because Teva itself did not argue structural commonality of the humanized anti-CGRP antagonist antibody genus. See Teva, slip op. at 9, n.8 & n.19 & n.16. That distinction is technically accurate at the level of which Ariad prong each appellee invoked, but it does not resolve the deeper question: Juno confronted the archetype of a well-known component genus embedded in a novel combination and declined to apply what Tevanow characterizes as a settled line of authority. Juno’s “same test” sentence stares directly at Teva’s framing, and the Federal Circuit has not yet spoken to how the two opinions coexist.
For in-house counsel, the practical implication is that Teva’s reasoning, if applied faithfully, creates a doctrinal argument for construct claims in which the structural backbone is the novel contribution and the functional binding element is drawn from a well-known class — exactly the fact pattern that died under Juno. Whether that argument succeeds may depend on how carefully the specification frames the “different invention,” and on whether the patentee travels on representativeness or on structural features. Juno remains a substantial obstacle when the patentee has to rely on the structural-features pathway; it is simply inapposite as a universal structural-similarity rule for representativeness, which is where Lilly over-read it. Counsel reading Teva as a green light for construct claims should be candid about the Juno headwind, and counsel defending Juno-era patents may now have a doctrinal path that previously looked closed. Either way, the conflict is live, and this is a question the Federal Circuit will need to revisit.
The Narrow Rochester/Ariad Circularity Distinction
The more vulnerable part of the opinion — and the one most worth reading closely — is the panel’s distinction of Univ. of Rochester v. G.D. Searle & Co., 358 F.3d 916 (Fed. Cir. 2004), and Ariad on the circularity objection. Lilly argued that the distinction between claiming an antibody class and claiming a method of using the class is “a semantic distinction without a difference,” quoting Rochester, 358 F.3d at 926. Teva, slip op. at 16. The panel disagreed. The specific grounds matter because the rest of the analysis turns on them.
Rochester and Ariad have a particular architecture: in Rochester, the claim was to a method of selectively inhibiting PGHS-2 activity “by administering a non-steroidal compound that selectively inhibits” PGHS-2 activity, 358 F.3d at 918; in Ariad, the claim encompassed the use of “all substances” that reduce NF-κB activity, 598 F.3d at 1341. In each case, the claimed act and the functional definition of the operative compound were the same verb, and no operative compound was disclosed or known. Those claims were, in the panel’s words, “methods of doing X using ‘something that does X.’” Teva, slip op. at 17.
Here, in contrast, the asserted claims are not to “methods of antagonizing CGRP using humanized antibodies that antagonize CGRP”; they are to methods of treating headache using such antibodies — something different from the function that characterizes these antibodies.
Id. The circularity inquiry is therefore a tautology inquiry rather than a separability inquiry. Rochester‘s claim recited the same verb twice — selectively inhibiting PGHS-2 by administering “a non-steroidal compound that selectively inhibits” PGHS-2 — making the claim “do X with something that does X.” Ariad‘s claims to the use of “all substances” that reduce NF-κB activity occupied the same circular position. Teva‘s claim, by contrast, recites one verb (treating headache) and defines the genus by another verb (antagonizing CGRP); the two verbs are causally linked but they are not the same verb, and the claim is therefore directed to a particular therapeutic application of the mechanism rather than to the mechanism itself. Read rigorously, then, the Teva exception requires that the class pre-exist the invention in the art and that the claim be directed to a specific therapeutic or technical application of the class rather than to the function that defines the class. If either condition is missing — if the class is brand-new, or if the claim verb is identical to the genus-defining function — the Rochester critique would still bite. Counsel relying on Teva‘s circularity distinction should confirm both conditions on the record.
The Evidentiary Mechanics
The doctrinal rules matter only because the record allowed them to operate, and practitioners should pay close attention to what filled in the factual gaps. The ’045 specification incorporated prior art: col. 25 lines 59–63 names three specific references establishing that anti-CGRP antagonist antibodies were known (Tan et al., Clin. Sci. 1995; Sigma product number C7113, clone #4901; Plourde et al., Peptides 1993). At col. 4 lines 50–51 the specification confirms that “[i]n some embodiments, the anti-CGRP antagonist antibody is humanized,” and at col. 27 lines 41–42 and col. 28 line 55 through col. 29 line 28 it lays out that antibodies “may be made recombinantly and expressed using any method known in the art,” followed by the canonical four-step humanization protocol. The murine antibody panel in Tables 2 and 3 at col. 51 lists eleven characterized species (including 4901 and six others), and cols. 72–73 provide the G1 heavy- and light-chain variable sequences (SEQ ID NOs: 1 and 2) plus six CDRs.
The trial record, however, may be what made the spec’s references operative. Teva introduced the Sigma product catalog at trial, showing that clone 4901 was commercially available for $250 before the priority date, and elicited testimony at trial that a researcher had received seven such antibodies as a gift from another researcher — direct evidence that the spec’s “any method known in the art” language was not a vacuous incantation but pointed to actually accessible species. Teva, slip op. at 16.
Perhaps most important for litigators thinking about the risk-reward of post-grant proceedings, Lilly’s own PTAB record deserves careful attention. Between August and October 2018, Lilly filed IPR petitions challenging two groups of Teva patents: the “antibody patents,” containing composition claims to humanized anti-CGRP antagonist antibodies themselves, and the “headache patents” at issue here, containing the method-of-treatment claims. See Teva, slip op. at 4 & n.4. To knock out the antibody-patent claims for obviousness, Lilly told the Board, emphatically and repeatedly, that anti-CGRP antagonist antibodies “were well known in the art” by November 2006, that the prior art was “replete with exemplary disclosures” of such antibodies, that techniques for making them were “extensively described in the prior art,” and that humanizing such antibodies “was a well-established and routine procedure” by the same date. J.A. 21407, 21417, 21442, 21513; Teva, slip op. at 4. Lilly won at the PTAB on the antibody patents, and those determinations of unpatentability were affirmed by the Federal Circuit. See Teva Pharms. Int’l GmbH v. Eli Lilly & Co., 8 F.4th 1349 (Fed. Cir. 2021); Teva Pharms. Int’l GmbH v. Eli Lilly & Co., 856 F. App’x 312 (Fed. Cir. 2021). It lost on the headache patents, whose method claims survived. See Eli Lilly & Co. v. Teva Pharms. Int’l GmbH, 8 F.4th 1331 (Fed. Cir. 2021). And in the resulting district-court trial, Teva put Lilly’s own obviousness admissions to work establishing the “well-known class” predicate underpinning its § 112 ¶ 1 defense. The admissions Lilly needed to kill one patent family were precisely the admissions Teva needed to save another. That is a cautionary note worth internalizing: admissions made in a litigation or prosecution to secure reasonable likelihood of prevailing under § 102 or § 103 can cut strongly the other way on written description and enablement in a parallel district-court proceeding, particularly when both sets of claims reference the same compound, target, or substrate class.
The composition case worth understanding in more detail is Teva Pharms. Int’l GmbH v. Eli Lilly & Co., 8 F.4th 1349 (Fed. Cir. 2021) (Lourie, J.), which affirmed PTAB determinations of unpatentability for claims of U.S. Patent Nos. 9,340,614; 9,266,951; and 9,890,210 over a three-reference combination: Olesen et al., the 2004 New England Journal of Medicine clinical trial establishing that BIBN4096BS, a nonpeptide CGRP-receptor antagonist, was safe and effective for acute migraine; Tan et al., disclosing full-length anti-CGRP monoclonal antibodies that blocked CGRP binding to its receptor in rats; and Queen et al., teaching the canonical humanization methodology. The Board found — and the Federal Circuit affirmed — that a person of ordinary skill in the art would have been motivated to combine those teachings to pursue a humanized anti-CGRP antagonist monoclonal antibody for therapeutic use in humans, and would have had a reasonable expectation of success at making such an antibody. The composition claims therefore fell. In the parallel appeal, Eli Lilly & Co. v. Teva Pharms. Int’l GmbH, 8 F.4th 1331 (Fed. Cir. 2021), the same Olesen/Tan/Queen combination was held insufficient to render the headache patents’ method-of-treatment claims obvious, because the skilled artisan lacked a reasonable expectation of success that the resulting antibody would therapeutically treat migraine in humans – open questions about whether receptor-antagonist efficacy translated to ligand-antagonism, and whether an antibody-format inhibitor would achieve therapeutic effect given pharmacokinetic realities, intervened between class membership and therapeutic outcome. The architecture that emerges from the two 2021 appeals is precisely what the 2026 panel relied on in distinguishing Rochester: the compound class was not merely known but obvious, while the claimed therapeutic use was separable and non-obvious. For portfolio counsel, the pairing is also a lesson in claim strategy: a composition-claim loss in post-grant proceedings is not necessarily fatal to a companion method-of-use program if the therapeutic result remains separable from the mechanism that defines the class, and drafters should deliberately build that separability into the specification at the earliest filing date.
The Parallel Enablement Holding
A reader could be forgiven for thinking Teva is a written-description case and no more. It is not. Judge Burroughs entered JMOL on both § 112 ¶ 1 grounds, and the Federal Circuit reversed on both; Part II of the opinion — slip op. pages 21–24 — is a full enablement analysis that tracks the same doctrinal path and reaches the same result. For any practitioner thinking about the reach of the decision, the enablement reversal is as important as the written-description reversal.
The panel opens with Ariad’s observation that enablement and written description “often rise and fall together,” 598 F.3d at 1352, and states “[s]o it is here.” Slip op. at 21. Lilly argued that the humanized anti-CGRP antagonist antibody genus is too large, that the specification fails to teach which candidates would antagonize CGRP, and that the screening needed to identify operative species would constitute undue experimentation. See id. at 21–22. The panel accepted two propositions for argument’s sake — that a reasonable jury could have found a large number of candidate antibodies would need to be screened, and that the screening would have constituted undue experimentation — and nonetheless rejected the argument.
The reasoning mirrors the written-description analysis and deserves careful reading for its interaction with Amgen v. Sanofi, 598 U.S. 594 (2023), and Baxalta Inc. v. Genentech, Inc., 81 F.4th 1362 (Fed. Cir. 2023). Both of those decisions invalidated antibody claims on enablement grounds, and both are distinguished here on the ground that they involved claims to the genus of antibodies themselves. In Amgen, the claims covered “that entire genus (for any and all purposes),” and the specification functioned as a “research assignment” of “painstaking” trial and error. 598 U.S. at 614; Teva, slip op. at 22. Baxalta followed the same path on facts “materially indistinguishable from those in Amgen.” 81 F.4th at 1366. Teva’s claims, by contrast, are to a method of use for a “different, limited purpose” — treating headache — and not to the antibody genus itself. Because anti-CGRP antagonist antibodies were already well-known, because humanization was routine, and because the record supported a jury finding that all humanized anti-CGRP antagonist antibodies treat headache, the research assignment that Amgen condemned was simply not at issue. The panel’s framing is worth pulling in full:
Given that anti-CGRP antagonist antibodies (and methods of making them) were already well known, humanizing them would have been routine, such humanized antibodies themselves are not claimed here, and all of them work in the claimed method, undertaking to find or make all of them would — in the context of these claims — be more akin to extra credit than a necessary research assignment left to others to complete.
Slip op. at 23.
Idenix Pharmaceuticals LLC v. Gilead Sciences Inc., 941 F.3d 1149 (Fed. Cir. 2019), was distinguished on a different axis. There, the key enablement question was whether a skilled artisan would know, without undue experimentation, which 2’-methyl-up nucleosides would treat hepatitis C; the record did not support a finding that all such nucleosides worked, and the claim accordingly failed because the screening burden fell on the public. Id. at 1156, 1159. Teva is distinguishable on the specific factual finding that “all humanized anti-CGRP antagonist antibodies treat headache, thereby obviating any extensive screening to determine which ones do.” Slip op. at 24.
That single finding does outsized doctrinal work throughout the opinion — it supplies the Idenix distinction here and reinforces the representativeness analysis in Part I — so practitioners considering whether Teva’s pathway is available on their facts should understand how it was built on this record. Four layers converge. First, the procedural posture. The Federal Circuit reviewed the JMOL de novo and treated the jury as having made every verdict-supporting finding that substantial evidence could support. See slip op. at 6. The question was not whether Teva had proved the proposition conclusively but whether a reasonable juror could have found it on the record, which is a materially lower threshold. Second, Lilly’s non-contest on appeal. The panel expressly noted at slip op. 13 that “as Lilly does not dispute, a reasonable jury could have found that a skilled artisan would have understood from the specification that all humanized anti-CGRP antagonist antibodies treat headache,” pinning the citation to J.A. 4174 and J.A. 4272 in Teva’s trial record. Litigators should note the posture carefully — Lilly chose to direct its appellate fire at the representative-species and screening questions rather than at the class-wide-efficacy finding, and that choice had real downstream consequences for the way the circularity and enablement sections read. Third, the district court’s own acknowledgment. Judge Burroughs, in the JMOL opinion, accepted that a person of ordinary skill would have understood from the specification that all humanized anti-CGRP antagonist antibodies would treat headache, see Teva Pharms. Int’l GmbH v. Eli Lilly & Co., No. 18-cv-12029, 2023 WL 6282898, at 4, 12, *19 n.23 (D. Mass. Sept. 26, 2023), and granted JMOL on separate grounds (representative-species and screening burden) rather than by rejecting class-wide efficacy as a matter of fact. That acknowledgment gave the appellate panel unusually clean footing because the fact was not in dispute at any stage. Fourth, the underlying trial testimony. Footnote 12 of the opinion collects multiple independent sources for the well-known status of anti-CGRP antagonist antibodies — Teva’s expert Dr. Hill (J.A. 4059, 4063–67, 4119–26) and Teva’s expert Dr. Hale (J.A. 4209–14) — and footnote 13 collects the same for the routineness of humanization, including Dr. Hale (J.A. 4220–24) and, significantly, Lilly’s own expert Dr. McDonnell (J.A. 3295–96).
The ’045 specification’s disclosure of the CGRP-antagonism mechanism linking class membership to the clinical endpoint then supplied the logical bridge the experts walked across. The one Lilly attempt to narrow the class-wide efficacy finding — an argument that galcanezumab treats cluster headache while the disclosed G1 species had not been shown to do so — was dispatched in footnote 17, where the panel credited Dr. Blumenfeld’s testimony on study-design effects at J.A. 4018–21 to hold that a reasonable jury could have found Lilly’s cluster-headache evidence less than clear and convincing. The practical lesson is straightforward: what made the doctrine available here was the trial record. Teva built a record showing that the class and its properties were well known in the art — through its own experts, through the specification’s mechanistic disclosure, and through Lilly’s own expert. That raises a candid question about how much the specification alone could have carried the day without the findings and admissions that accumulated during litigation.
The enablement holding matters for two reasons. First, it confirms that Teva’s “different invention” doctrine does not stop at written description; it reaches enablement on parallel logic. A practitioner drafting therapeutic method-of-use claims that fit the known-class-plus-downstream-use architecture now has both § 112 ¶ 1 hooks reinforced by the same opinion. Second, it confirms the Amgen v. Sanofi limiting principle: Amgen still controls composition claims to functionally-defined antibody genera, but it does not reach method-of-use claims built on pre-existing antigen-directed classes. The enablement case law is accordingly bifurcating — Amgen and Baxalta on one side for composition claims, Teva on the other for use claims premised on a known class.
Beyond Antibodies: Broader Therapeutic Classes and Platform Technologies
Nothing in the panel’s reasoning confines the doctrine to antibodies. The Teva opinion speaks in terms of “genus,” “different invention,” and “well-known in the art”; the case facts are antibody facts, but the test is formal. That matters for three categories of life-sciences claim that have been receiving similar § 112 ¶ 1 pushback during prosecution and in litigation.
The first is small-molecule method-of-use claims built on a known target class. When the target (for example, a defined kinase, GPCR, or nuclear receptor) and a class of known ligands or modulators for that target pre-exist the priority date, a claim to a new therapeutic use of that class occupies the same doctrinal position as Teva’s humanized anti-CGRP antagonist antibodies. UroPep itself was a small-molecule case — PDE V inhibitors used to treat benign prostatic hyperplasia — and it sits squarely on the lineage Teva now reaffirms. Counsel drafting new-indication claims for a defined pharmacological class should cite Teva and UroPep in tandem and build a record that tracks the two-factor circularity inquiry: the class pre-exists the invention, and the claim is directed to a specific therapeutic application of the class rather than to the function that defines the class.
The second is cell-, gene-, and nucleic-acid-based therapeutics in which the effector class or chemistry class is known and the claimed contribution is a new use, a new combination, or a new architectural feature built around that class. Teva does not, of course, answer Juno’s structural-features holding for novel combinations of known elements — a point developed above — but it does reinforce the representativeness pathway of Ariad when the specification can anchor the claim to a pre-existing genus and a separable downstream use.
The third, and the one most often under-protected, is platform technologies that are generally applicable across target classes — novel amplification chemistries, new isolation or capture methods, improved sequencing or detection architectures, engineered enzymes usable on wide substrate genera. These are the inverse of the Teva fact pattern: the invention is the method and the universe of possible targets is the “well-known genus used as part of a different invention.” Practitioners prosecuting such claims regularly receive § 112 ¶ 1 rejections demanding enumeration of every target molecule to which the platform can be applied. The doctrine Teva rehabilitates answers that directly. Fuetterer did not require the applicant to identify every inorganic salt capable of holding a carbohydrate-protein colloid; Ajinomoto did not require disclosure of every “more potent promoter” capable of driving yddG expression. The invention in each was the combination, not the discovery of the background class. A method claim to “amplifying a target nucleic acid” or “isolating a target protein” using a new technique should be defensible on the same logic, provided the target class is well-characterized in the prior art and the specification ties the method to that class with sufficient anchoring citations. Counsel on the receiving end of such rejections now have a modern Federal Circuit peg on which to hang the Fuetterer/Herschler/Ajinomoto response.
Practical Implications
The decision may be harder to translate into practice than its result suggests, because much of what carried the appeal lived outside the four corners of the specification. The “well-known class” predicate was established in significant part by Lilly’s own admissions in the parallel IPRs against Teva’s antibody composition patents, where Lilly told the Board that anti-CGRP antagonist antibodies “were well known in the art” and that humanizing them “was a well-established and routine procedure.” Teva, slip op. at 4–5. The class-wide-efficacy predicate was built in the trial record by experts on both sides — Teva’s Dr. Hill and Dr. Hale, and Lilly’s own Dr. McDonnell on the routineness of humanization — and was then accepted by Judge Burroughs as a finding the jury could have made on the evidence. None of those evidentiary inputs is available at the time an application is being drafted, and a candid reading of Teva should resist over-claiming what the specification by itself accomplished.
What can be done in the specification is foundational, and the framing of the invention rewards careful attention. The case illustrates the value of identifying the inventive contribution with precision and framing it in language that distinguishes the contribution from the class on which it operates: the inventive matter is the new therapeutic application, the new combination, or the new technical method, not the class of operative compounds or the universe of targets to which the method applies. Anchoring the prior-art status of the class through specific citations to published references, catalog entries, sequence databases, or other anchoring sources rather than generic “known in the art” recitations is a choice that matters at trial: the three references at col. 25 of the ‘045 specification did real work before the jury, and the equivalent evidentiary anchors for other modalities deserve the same attention. Describing the canonical how-to-make-it or how-to-perform-it methodology in enough detail that the spec is operable on its own terms reduces reliance on expert testimony that may be unavailable or contested at a critical reduction-to-practice step. And including as broad a panel of representative species as the program can support at the earliest filing date strengthens both the written-description foundation and the evidentiary record on which that foundation will be tested.
The second and more difficult nuance is the inverse problem that aggressive prior-art characterization creates under §§ 102 and 103. The same record that helps a § 112 ¶ 1 defense — operative compounds were a well-defined class, methods of generating them were routine — is precisely the record that buried Teva’s parallel antibody composition patents in the 2021 PTAB campaign and the resulting Federal Circuit affirmances. The “well-known class” predicate is therefore not a costless disclosure choice; it strengthens § 112 support to the extent the inventive contribution sits somewhere other than in the class itself. Where the inventive contribution is a non-obvious therapeutic outcome or a non-obvious technical application of the class — Teva’s headache patents survived obviousness despite class membership because the therapeutic outcome of antibody-format ligand-antagonism was independently uncertain in 2006 — the prior-art characterization helps under § 112 without prejudicing patentability. Where the inventive contribution reduces to use of the class for its already-known purpose, characterizing the class as well-known invites a § 112 ¶ 1 “research plan” objection and a § 102/§ 103 “obvious to try” rejection on the same record. The strongest § 112 ¶ 1 profile is one that simultaneously preserves a § 102/§ 103 differentiator that does not reduce to the class itself, and characterizations made in any one forum — prosecution, IPR, district-court litigation — travel to the others, so the case file is best understood as a single cross-proceeding record from the outset.
In litigation, the same architecture works as both a defensive shield and an offensive tool. It supplies a written-description and enablement defense for method-of-use and platform claims that might otherwise look vulnerable under AbbVie or Juno read broadly, and it identifies the categories of evidence — adversary admissions in parallel proceedings, expert testimony from both sides on the routineness of operative steps, and a jury-level finding of class-wide efficacy reviewed under a favorable standard — that have done the doctrinal work in this kind of fact pattern. The same logic that powers the defense also reinforces the cross-proceeding caution above: positions taken in one forum become available in another, and the Teva / parallel-IPR pairing is the cleanest available cautionary tale.
Closing
Teva is a meaningful affirmation that written description and enablement doctrine have room for inventions that use well-known compound, target, or substrate classes to achieve new therapeutic or technical ends, and it is the clearest Federal Circuit signal in some time that the genus-disclosure inquiry is fact-sensitive rather than mechanical. For counsel drafting or prosecuting therapeutic applications across any modality — antibody, small molecule, cellular, nucleic acid — and for counsel prosecuting platform technologies whose inventive contribution is a method generally applicable across known target classes, the decision is most useful read alongside the three CCPA-era cases it builds on; for counsel defending those patents in litigation, it supplies a doctrinal framework that finally has a modern Federal Circuit reaffirmation to anchor it. The next several years of prosecution will test whether panels continue to treat the “different invention” line as a stable disclosure pathway or as a narrow carve-out; the Juno/Teva tension in particular is one the Federal Circuit will eventually have to address directly.
For questions about how this decision intersects with your portfolio, contact Brent Nix at Johnson, Marcou, Issacs & Nix — bnix@jmin.com.
Citations
Teva Pharm. Int’l GmbH v. Eli Lilly & Co., No. 24-1094 (Fed. Cir. Apr. 16, 2026) Ariad Pharm., Inc. v. Eli Lilly & Co., 598 F.3d 1336 (Fed. Cir. 2010) (en banc) In re Fuetterer, 319 F.2d 259 (CCPA 1963) (Rich, J., plurality) In re Herschler, 591 F.2d 693 (CCPA 1979) Ajinomoto Co. v. Int’l Trade Comm’n, 932 F.3d 1342 (Fed. Cir. 2019) Erfindergemeinschaft UroPep GbR v. Eli Lilly & Co., 276 F. Supp. 3d 629 (E.D. Tex. 2017) (Bryson, J., sitting by designation) Univ. of Rochester v. G.D. Searle & Co., 358 F.3d 916 (Fed. Cir. 2004) Juno Therapeutics, Inc. v. Kite Pharma, Inc., 10 F.4th 1330 (Fed. Cir. 2021) AbbVie Deutschland GmbH & Co., KG v. Janssen Biotech, Inc., 759 F.3d 1285 (Fed. Cir. 2014) Regents of the Univ. of Cal. v. Eli Lilly & Co., 119 F.3d 1559 (Fed. Cir. 1997) Amgen Inc. v. Sanofi, 598 U.S. 594 (2023) Baxalta Inc. v. Genentech, Inc., 81 F.4th 1362 (Fed. Cir. 2023) Idenix Pharm. LLC v. Gilead Scis. Inc., 941 F.3d 1149 (Fed. Cir. 2019) Capon v. Eshhar, 418 F.3d 1349 (Fed. Cir. 2005) Bos. Sci. Corp. v. Johnson & Johnson, 647 F.3d 1353 (Fed. Cir. 2011) MagSil Corp. v. Hitachi Glob. Storage Techs., Inc., 687 F.3d 1377 (Fed. Cir. 2012) Trs. of Bos. Univ. v. Everlight Elecs. Co., 896 F.3d 1357 (Fed. Cir. 2018)
U.S. Patent No. 8,586,045 (key passages at cols. 4, 25, 27, 28–29, 51, 72–73) U.S. Patent No. 8,586,045, claims 17 and 30 D. Mass. Case No. 1:18-cv-12029-ADB, Dkt. 535 (Oct. 17, 2022) (MIL rulings) J.A. 18046 (Sigma catalog exhibit); J.A. 4122 (gift-of-antibodies testimony); J.A. 4212–14, 4250 (routine-humanization testimony)