One of the most difficult aspects of patenting antibodies and other biologics is obtaining sufficiently broad patent protection. In order to claim a genus of antibodies, rather than specific sequences, the applicant must satisfy the written description requirement (35 U.S.C. § 112), generally by providing a sufficient number of species that are representative of the full variety or scope of the genus. This standard is particularly difficult to meet given that neither the United States Patent and Trademark Office (“USPTO”) nor the Court of Appeals for the Federal Circuit (“Federal Circuit”) has stated what a sufficient number of species is and because, by its very nature, the determination is fact-specific.
Recent court decisions from the District of Delaware indicate that claims that cover only conservative variants of antibodies known to be effective and/or claims that define the binding location of the antibodies by specific and limited number of residues may be sufficient to satisfy the written description requirement. At this time, however, neither the USPTO nor the Federal Circuit has provided definitive guidance on this issue.
Clarification of guidance for claiming of antibodies
The USPTO used to permit claiming of antibodies by disclosing the targeted antigen (the “newly characterized antigen test”). In 2017, the Federal Circuit rejected the newly-characterized antigen test in Amgen Inc. v. Sanofi. 872 F.3d 1367 (Fed. Cir. 2017). The Supreme Court denied certiorari review in January 2019.
The Patent Office had issued a clarification of the written description guidance for claims drawn to antibodies in February 2018, noting that, “in view of the Amgen decision, adequate written description of a newly characterized antigen alone should not be considered adequate written description of a claimed antibody to that newly characterized antigen, even when preparation of such an antibody is routine and conventional.” Examiners were instructed to continue following the guidance in the MPEP regarding the written description requirement (i.e., MPEP 2161.01 and 2163), “except insofar as MPEP 2163 indicates that disclosure of a fully characterized antigen may provide written descriptive support of an antibody to that antigen.” Examiners were instructed to continue to follow 2015 and 2016 training materials related to 35 U.S.C. § 112(a) , including the guidance “Antibody Decisions and Their Compliance with the Written Description Requirement” (“Antibody Training Materials”). The USPTO is still in process of updating the Antibody Training Materials based on the Federal Circuit’s decision in Amgen Inc. v. Sanofi.
The MPEP provides that the written description requirement for a claimed genus may be satisfied through sufficient description of a representative number of species by actual reduction to practice, reduction to drawings, or by disclosure of relevant, identifying characteristics, i.e., structure or other physical and/or chemical properties, by functional characteristics coupled with a known or disclosed correlation between function and structure, or by a combination of those characteristics. A “representative number of species” requires the adequately-described species provided to be representative of the entire genus.
With respect to the invention of antibodies, which is generally recognized as an unpredictable art, disclosing one or even a few species will not provide adequate written description of a genus which embraces widely variant species. “The disclosure must adequately reflect the structural diversity of the claimed genus, either through the disclosure of sufficient species that are ‘representative of the full variety or scope of the genus,’ or by the establishment of ‘a reasonable structure-function correlation.’” Such correlations may be established in the specification or may be “known in the art at the time of the filing date.”
Federal Circuit precedent for claiming of antibodies
At this time, the USPTO has not provided additional guidance on the claiming of antibodies or on the written description requirement for the biological arts. The Federal Circuit has stated that, to show possession of a genus, a patentee must disclose (1) “a representative number of species falling within the scope of the genus” or (2) “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.” Ariad Pharm., Inc. v. Eli Lilly & Co., 598 F.3d 1336, 1350 (Fed. Cir. 2010). An adequate written description “requires a precise definition, such as by structure, formula, chemical name, physical properties, or other properties of species falling within the genus sufficient to distinguish the genus from other materials.” Id. Functional claim language “can meet the written description requirement when the art has established a correlation between structure and function.” Enzo Biochem, Inc. v. Gen–Probe Inc., 323 F.3d 956, 964 (Fed. Cir. 2002). While the Federal Circuit has provided plenty of examples where a disclosure does not meet these tests, there are not many cases where has it has blessed “functional” language.
The lack of guidance leaves numerous open questions: can applicants still:
(1) claim antibodies with respect to the structure of the antigen (even when the entire epitope is mapped); or (2) effectively claim the genus of antibodies that bind to a particular antigen?
And, even if the USPTO will accept functional language (binds to epitope X or binds to antibody at residues a, b or c), will the patent hold up in litigation? What is a sufficient number of disclosed examples to claim a genus of antibodies? How does the specification establish a sufficient structure-function correlation?
Morphosys AG v. Janssen Biotech
Recent district court cases, while not definitive, may provide guidance on these issues. The first decision, Morphosys AG v. Janssen Biotech, addressed the written description requirement in the context of three patents on antibodies that bind to the CD38 protein and can be used to treat blood cancer. 358 F.Supp. 3d 354 (D. Del. 2019). The district court granted summary judgment, finding that the patents were invalid for failure to comply with the enablement requirement of 35 U.S.C. § 112. The court found an issue of fact, however, as to whether the applicant disclosed a representative number of species sufficient to support the claimed genus. Id. at 365. While the claims covered a large number of actual and potential antibodies, the size of the claimed genera was in dispute. One particular antibody disclosed (MOR03080) and the accused antibody (daratubumab) had a 90 percent sequence similarity, and the accused antibody could have been made employing the teachings of the patents. Further, the defendant failed to identify any antibodies that fell within the scope of the claims but were not variants of the disclosed antibodies. Id. at 366.
The district court did, however, find that the patentee failed to disclose structural features common to the members of the genus. The claims covered CD38 antibodies that satisfied functional tests (binding within specific locations or certain cell-killing activities) but did not correlate those functions to structural characteristics (amino acid sequence) of antibodies having those properties. Id. at 366. Plaintiff admitted that the antibody sequence could not be used to predict binding properties, but that additional screening was necessary. While the claims cited certain structural limitations (such as “human” or “humanized” antibodies, “IgG” or “IgG1” antibodies, or “VH3 heavy chains” or “kappa light chains”), those structural features were present in many antibodies that do not bind to CD38. Id. at 367.
The district court also found lack of enablement, stating that the full scope of a claim is not enabled when there is an embodiment within the claim’s scope that a person of ordinary skill, reading the specification, would be unable to practice without undue experimentation. Id. at 368-69. (A literal reading of this language would invalidate almost any claim to a genus; because almost every genus claim may have some inoperative embodiment). The evidence demonstrated that there were anywhere from ten quintillion to millions of anti-CD38 antibodies. Id. at 369. Looking at the various other limitations of the claims (such as (1) being IgG or IgG1 antibodies; (2) having a specified antibody dependent cellular cytotoxicity (ADCC) or cell dependent cytotoxicity (CDC) effectiveness; (3) binding to specific epitopes; (4) being used in treatment; and (5) having specific structural characteristics), even the most limiting claim covered at least 0.5% of all human anti-CD38 antibodies. Id.
The Court noted that “conservative” variants of antibodies that were “known to be effective” are themselves “reasonably expected” to be effective. Id. at 370. Conservative variants result from “conservative point substitutions,” whereby substitutions are made only within the framework regions of an antibody and would have a “’small impact” on the antibody’s functional properties, including whether it binds to CD38. Id. On the other hand, an antibody made via nonconservative changes, particularly by changes to the complimentary determining regions (CDRs), would have to be screened in order to determine its effectiveness. Id. The accused antibody was not a conservative variation of any antibody disclosed in the patent, and its CDR was only 35 percent similar to that of the most similar disclosed antibody, MOR03079. A person of ordinary skill in the art would have had to devote substantial time and effort to discover antibodies that fall within the scope of the claims that were not conservative variants of the disclosed antibodies. Id. at 370-71. Undue experimentation would be required to practice the full scope of the claims. Id. at 371. The Court did note, however, that “certain conservative amino acid substitutions would create new antibodies while preserving a known antibody’s properties.” Id. While the Court addressed conservative variants in the context of enablement, its language appears to apply equally to the written description requirement and how to disclose structural features common to the genus.
The case subsequently settled, so the Federal Circuit will not review this decision.
Amgen v. Sanofi
On remand in Amgen v. Sanofi, discussed above, the district court denied a motion for summary judgment requesting a determination that claims to a genus of monoclonal antibodies that inhibit PCSK9 were invalid for lack of written description and enablement. 2019 W.L. 259099 (D. Del. Jan. 18, 2019). The court found that genuine disputes of material fact existed under both “the common structural features test and representative species test.” Id. at *2.
The parties’ experts disputed whether “the function of binding correlates to the structure of the antibody.” Id. The patents included claims to: (1) a pharmaceutical composition comprising an isolated monoclonal antibody, wherein, when bound to PCSK9, the isolated monoclonal antibody binds to at least two of the following residues S153, I154, P155, R194, D238, A239, I369, S372, D374, C375, T377, C378, F379, V380, or S381 of PCSK9 listed in SEQ ID NO:3; and (2) a pharmaceutical composition comprising an isolated monoclonal antibody, wherein the isolated monoclonal antibody binds to at least two of the following residues S153, I154, P155, R194, D238, A239, I369, S372, D374, C375, T377, C378, F379, V380, or S381 of PCSK9 listed in SEQ ID NO:3 and blocks the binding of PCSK9 to LDLR by at least 80 percent.
The plaintiff claimed that the 15 amino acids (out of PCSK9’s 700) form a small region of PCSK9’s surface having a unique structure and chemical properties. Antibodies that bind to one or more residues in that sweet spot block PCSK9’s ability to bind to LDL receptors. The plaintiff’s experts provided testimony that the claims were “very narrow” and that a small number of antibodies bind to the particular amino acids specified in the claims. The plaintiff’s expert testified all antibodies that bind to the sweet spot have common structures—both their three-dimensional shape and chemical structural features. The evidence also allegedly demonstrated that the claimed antibodies had a common structural feature of greasy and polar/charged residues complimentary to the unique shape and chemical features of the sweet spot. (In that respect, the evidence regarding the structural commonality of the claims that bound to certain epitopes appears to be different in this case than in Morphosys AG v. Janseen Biotech – it does not appear that Morphosys pointed to any structure other than the sequence.)
As to whether there were a sufficient number of representative species, the patents described “at least 32 antibodies by sequence that fall within the claimed genus by binding to the fifteen amino acid sweet spot on PCSK9.” Id. at *3. The defendants argued that the disclosed species were not representative because the disclosed antibodies would bind to no more than eight PCSK9 residues, whereas the competitor antibodies bind to different combinations of residues, and the patents disclosed “only four antibodies that share sixty percent or more of [their] heavy or light chain CDR sequences with any of the competitor antibodies.” Id. The plaintiff, however, presented expert testimony that a person of ordinary skill in the art would understand that the exemplary antibodies were representative of the claim genus due to sequence characteristics common to the accused antibodies. Id.
On February 25, 2019, the jury found that certain claims were not invalid for lack of adequate written description. The case is expected to be appealed.
