In Juno Therapeutics v. Kite Pharma, the Court of Appeals for the Federal Circuit (CAFC) reversed a $1.2 billion dollar judgment to Juno, finding that Juno’s asserted claims, which they alleged were infringed by Kite’s CAR-T therapy YESCARTA® (axicabtagene ciloleucel), lacked written description. The decision follows a line of cases involving antibodies and written description, extending recent antibody-related jurisprudence into the realm of cell-based immunotherapies, such as chimeric antigen receptor (CAR) T cells. The court determined that while the binding element of CARs (in this case, an scFv) and methods of making them were generally known in the art, the scope of the claimed CARs was too vast to be sufficiently supported by the limited number of exemplary binding elements in the specification, as well as the absence of common structural features defining the claimed genus. The decision means that those seeking claims that aim to broadly protect CARs with functional definitions will face significant challenges based on written description.
Kite manufactures YESCARTA®, a CD19-directed CAR-T cell-based therapy that was approved by the US Food and Drug Administration (FDA) in 2017 for certain B-cell malignancies and gained rapid adoption, providing revenues of more than $264 million in 2018 and double that in 2019. Treatment with YESCARTA® involves removing a patient’s own T cells and genetically modifying them to express a CAR which causes the T cells, when returned to the patient’s blood, to specifically target and kill the desired cancer cells. The CAR includes three components: an intracellular signaling domain, a co-stimulatory domain, and an extracellular binding element (e.g., an scFv) that targets the T cell to the desired cancer cells by binding to a specific marker on the target cells (e.g., an FMC63 ScFv targeting CD19).
Juno sued Kite for infringement of claims in US 7,446,190 (the ‘190 patent) in district court. Kite filed counterclaims seeking declaratory judgments of noninfringement and invalidity of the ‘190 patent. The jury disagreed with Kite, finding that the claims were not invalid for lacking written description and enablement, and found that Kite willfully infringed the ‘190 patent, awarding Juno $585 million in damages. In post-trial motions, Kite unsuccessfully moved for JMOL, asserting that Juno’s claims lacked written description and enablement. In the end, the district court agreed with the jury and awarded $1.2 billion in damages. Kite appealed, arguing that the district court erred in denying its JMOL on the issue of written description and enablement.
The Federal Circuit reversed the district court, finding that the record lacked substantial evidence of sufficient written description support for the claims, holding the claims invalid on that ground. The court did not reach the other issues on appeal.
The asserted claims included claims 3 and 5, reproduced below with intervening claim 2 and base claim 1.
1. A nucleic acid polymer encoding a chimeric T cell receptor, said chimeric T cell receptor comprising
(a) a zeta chain portion comprising the intracellular domain of human CD3 zeta chain,
(b) a costimulatory signaling region, and
(c) a binding element that specifically interacts with a selected target,
wherein the costimulatory signaling region comprises the amino acid sequence encoded by SEQ ID NO:6.
2. The nucleic acid polymer of claim 1, wherein the binding element is an antibody.
3. The nucleic acid polymer of claim 2, wherein the antibody is a single chain antibody.
5. The nucleic acid polymer of claim 3, wherein the single chain antibody binds to CD19.
9. The nucleic acid polymer of claim 8, wherein the antibody is a single chain antibody.
11. The nucleic acid polymer of claim 9, wherein the single chain antibody binds to CD19.
Significantly, the point of novelty/nonobviousness of Juno’s claimed invention relates to the “costimulatory signaling region” and not the “binding element.”
The Federal Circuit’s Decision
In a precedential opinion, the CAFC reversed on the issue of written description, finding that no reasonable jury could find that the ’190 patent’s written description sufficiently demonstrated that the inventors possessed the full scope of the claimed invention.
The CAFC applied the standard set forth in Ariad v. Eli Lilly 598 F.3d 1336 (Fed. Cir. 2010), which held that written description is met if the specification “reasonably conveys to those skilled in the art that the inventor had possession of the claimed subject matter as of the filing date.” The court noted that for a genus defined by functional language, such as the binding function of scFvs, adequate written description support under Ariad may be demonstrated by disclosing “either a representative number of species falling within the scope of the genus or structural features common to the members of the genus so that one of skill in the art can ‘visualize or recognize’ the members of the genus.” Citing Regents of the Univ. of Cal. V. Eli Lilly 119 F.3d 1559 (Fed. Cir. 1997), the court also stated in relation to inventions involving a “chemical genus,” that written description support requires a “precise definition,” such as a description of the “structure, formula, [or] chemical name” of the claimed subject matter. According to the court, written description analysis is a question of fact, the adequacy of which under Ariad depends on factors, such as existing knowledge, extent of prior art, and maturity and predictability of the science involved in the claimed subject matter.
Kite contended that Juno’s patent disclosed neither representative species nor common structural features of the claimed scFv genus to identify which scFvs would function as claimed. Kite also argued that the scFv field is unpredictable and that the claims at issue cover “an enormous number (millions of billions) of scFv candidates, only a fraction of which satisfy the functional binding limitation for any given target,” as is required by certain claims of the ‘190 patent.
As to the claims that require the scFv to bind CD19 (i.e., claims 5 and 11 of the ‘190 patent), Kite argued that the universe of possible sequences for scFvs in the scope of the claims is in the range of “millions of billions,” yet only four or five CD19-specific scFvs were known in the art at the priority date and only a single anti-CD19 scFv was disclosed (identified as SJ25C1) in the patent. Kite accordingly contended that the patent thus failed to disclose species that are representative of the claimed genus.
To the contrary, Juno contended that scFvs and how to make them were well-known for nearly 15 years. Juno further argued that it was well-known that scFvs could be incorporated in CARs and that they are generally interchangeable with one another, having common structural features consisting of variable light and heavy chain regions joined by a linker. Juno also pointed to the two disclosed scFv species in the disclosure (J591 and SJ25C1), arguing that “these two working embodiments are representative of all scFvs in the context of a CAR.” As support, Juno pointed to testimony of its expert who stated that the two disclosed scFv species were “representative” of the entire claimed genus “because [scFvs] all do the same thing. They bind to the antigen.”
The CAFC held that Juno’s arguments were not supported by the evidence and that the specification failed to disclose either a representative number of species or common structural features of the “vast” genus of “a binding element that specifically interacts with a selected target.” Regarding Juno’s expert, the court remarked that “nothing about that testimony explains which scFvs will bind to which target” and without disclosure of more characteristics, such as identifying relevant binding sequences, the testimony did not demonstrate possession.
The court, however, did not go so far as to require the disclosure of entire nucleotide or amino acid sequences for known scFvs in the genus, consistent with its earlier decision in Capon v. Eshhar, 418 F.3d 1349 (Fed.Cir. 2005). According to the court, that the disclosure lacked the amino acid sequences of the two scFvs was not itself fatal, “as long as the patent provided other means of identifying which scFvs would bind to which targets, such as common structural characteristics or shared traits.” Since other such identifying characteristics of the species representing the claimed genus were not disclosed, the ‘190 patent lacked a sufficient written description under the “representative species” test.
In relation to the “common structural features” test, the court found that the ’190 patent failed to disclose structural features common to scFvs capable of binding to the vast genus of targets. The court further noted that the specification also failed to disclose a way to distinguish those scFvs capable of binding from those scFvs incapable of binding those targets or any structural aspects attributable to those differences. As viewed by the court, in the absence of these structural characteristics that define binding or non-binding, particularly against a vast number of targets, the ‘190 patent merely claims a “problem to be solved while claiming all solutions to it . . . cover[ing] any compound later actually invented and determined to fall within the claim’s functional boundaries.” Consequently, the specification cannot be said to provide sufficient written description under the “common structural features” test.
Accordingly, the court held that no reasonable jury could find that the ‘190 patent satisfies the written description requirement.
When drafting applications for CARs or CAR-Ts, remember that antibody case law will apply to the CAR binding element when it comes to written description. Understand the scope of functionally defined binding elements needing protection (e.g., important clinical embodiments), draft claims limited to that scope, and prepare the specification to satisfy the representative species and/or common structural features tests for written description. Note that disclosing a large number of species representative of a genus alone may still be insufficient to satisfy the written description requirement without an explanation of the structural characteristics common to the representative species (e.g., as was the case for Idenix).
Accordingly, build in as much support as possible for a structure-function relationship and include as many representative examples as possible for binding elements. Although yet to be tested for success, consider including results of AI/ML approaches to predict binding, and methods (e.g., particular assays) a skilled person could use to distinguish binders from non-binders.
Alternatively, focus claims on the point of novelty/nonobviousness and, to the extent possible, eliminate generic terms that might not be adequately supported. For example, a claim directed to a costimulatory domain for use in a CAR may survive a written description challenge if it does not incorporate the other CAR elements that are not part of the inventive contribution.
Include claims of different types and scope. As demonstrated in Juno v. Kite, case law can change between the time claims are written/issued and the time they are asserted. Method claims, e.g., method of making a protein having a particular property or method of treatment claims, especially where there are known antibodies that bind to a target, may receive more flexible treatment with respect to written description as compared to the treatment of composition claims.
When asserting CAR patents, consider the vulnerability of your claims to a lack of written description attack by assessing the extent of structural characterization of functionally-defined terms in the claims.