EXELIXIS, INC. v. MSN LABORATORIES PRIVATE LTD.
Authored by: Jeremy J. Gustrowsky
The Federal Circuit has affirmed that a patent specification describing a salt’s chemical name, formula, and crystalline nature adequately supports claims covering all crystalline forms of that salt, even though only two specific crystal forms were actually described. The decision offers welcome guidance for pharmaceutical companies claiming polymorphs, an area where written description challenges have become a common defense tactic.
The case revolves around Exelixis’s cancer drug Cabometyx®, which contains cabozantinib (L)-malate and treats kidney, liver, and thyroid cancer. Exelixis owns three patents sharing a common specification, United States Patent Nos. 11,091,439, 11,091,440, and 11,098,015 (the “Malate Salt Patents”), which claim crystalline cabozantinib malate salts, formulations containing them, and methods of treating cancer with them. The specification describes two crystalline polymorphs, designated N-1 and N-2, along with methods for preparing both crystalline and amorphous forms of the salt. MSN Laboratories filed an abbreviated new drug application seeking approval for a generic version using a different crystal form it called form S, on which MSN obtained its own patent.
MSN conceded infringement but argued the claims were invalid for lack of written description under 35 U.S.C. § 112(a). Its theory was straightforward. If the inventors only described two polymorphs, they could not have possessed the entire genus of crystalline forms, particularly since different polymorphs have different densities, melting points, solubilities, and stability profiles. The district court disagreed after a bench trial, and the Federal Circuit affirmed, reviewing the written description finding as a question of fact for clear error.
The court applied the familiar framework from Ariad Pharmaceuticals v. Eli Lilly, under which a patent can support a genus claim by disclosing either a representative number of species or structural features common to members of the genus. Exelixis satisfied the second path. The specification disclosed the chemical name and formula of cabozantinib (L)-malate and stated that the structure is crystalline, and a skilled artisan could readily distinguish crystalline material from amorphous material. That combination identified the structural features shared by every member of the claimed genus. The court analogized the situation to GlaxoSmithKline v. Banner Pharmacaps, where describing a complex of dutasteride and solvent molecules sufficiently identified the genus. Notably, the claims contained no functional or performance requirement, so there was no need for the specification to explain how to achieve a particular result.
MSN’s argument about differing physical properties among polymorphs did not carry the day. The court pointed out that those properties are unclaimed and were described only in general terms, and that the district court never relied on the properties of N-1 and N-2 to predict the properties of other forms. It relied instead on structural identity. The court also distinguished AbbVie v. Janssen, where specific evidence showed the accused antibodies shared only 50 percent sequence similarity with the disclosed ones. Here, an unchallenged finding placed the maximum size of the polymorph genus at fourteen forms, meaning this case did not raise the concerns that arise when a genus claim potentially sweeps in vast numbers of species. Cases like ICU Medical, Tronzo, and Eli Lilly v. Teva were also distinguished because each involved claims reaching beyond anything the specification actually disclosed. Here, the claims were no broader than the written description.
A separate issue involved United States Patent No. 11,298,349, which covers cabozantinib compositions “essentially free” of a genotoxic impurity. The district court found MSN’s product did not infringe claim 3 and also rejected MSN’s inherent obviousness challenge. Exelixis initially cross-appealed the noninfringement ruling but then dropped the cross-appeal, which made the noninfringement judgment final and left MSN with no live controversy. Exelixis argued MSN still had standing because the inherency ruling might have collateral consequences in separate litigation over a continuation patent, United States Patent No. 12,128,039. The court found those consequences too speculative. Because Exelixis’s unilateral decision to drop its cross-appeal prevented MSN from obtaining merits review, the court applied the Munsingwear doctrine, dismissed that portion of the appeal, and vacated the district court’s judgment on claim 3.
The practical takeaway for pharmaceutical patent practice is that structural genus claims to crystalline salt forms can survive written description scrutiny without describing every polymorph, provided the specification identifies the compound with precision and the claims stay tethered to structure rather than function. The procedural lesson is equally useful. A party that abandons a cross-appeal may inadvertently strip its opponent of appellate review and open the door to vacatur of a favorable ruling.