Antibodies are proteins made by the immune system to neutralize pathogens through binding to unique features of pathogens called “antigens” with exquisite selectivity. This ability to bind specific targets has been harnessed for innumerable biomedical research applications, diagnostic tests including Covid-19 antigen tests, and antibody-based therapies, making antibodies a valuable, powerful biotechnological tool. Oral arguments for Amgen v Sanofi were recently heard in US Supreme Court, revealing tensions regarding specification requirements that arise from characteristics of antibody preparation, and ultimately, balance between extending adequate protections for inventors and public interest in rejecting overbroad monopolies. Here, I consider patentability of antibody-based inventions in Canada and the significance of Amgen to Canadian jurisprudence.
To be patentable, antibody-based inventions must satisfy requirements of novelty, non-obviousness, usefulness[1] and be accompanied by specification sufficiently detailed to “enable” production of the invention by a skilled technician.[2]
Classic procedures for generating antibodies involved injecting antigens into animals, and then purifying antibodies with desired binding specificity. Antibodies have thus often been defined functionally, by looking to what they bind, as opposed to their amino acid sequence or their structure. This has impacted how inventors of antibody technologies assert novelty and utility. Sometimes novelty is asserted by claiming antibodies binding to a novel antigen, while more recently, claims have shifted to include amino acid sequences of antibodies, particularly their antigen-binding regions. Claims based on antigen-binding might lead to patenting of entire classes (“genuses”) of antibodies with different sequences that bind the same antigen.[3] For sequence-based claims, unpredictability regarding how sequence changes affect binding can mean that showing binding of claimed antibodies will be required to demonstrate utility.[4] For sequence-based claims, it is often unclear whether a previous patent for an antibody with similar sequence would preclude non-obviousness and novelty. Antibodies can also be considered novel if they bind previously-described antigens but have additional biological functions, like changing another protein’s activity. Claims then need to be accompanied by evidence that the antibody has the stated utility, either through demonstration or sound prediction.[5]
Specification requires that antibody inventions be correctly and fully described, and that the method of antibody production be set out in full, clear concise and exact terms to enable a skilled person to make the antibody. Specification in Canada requires that antigens be fully characterized and antibodies must either have been prepared,[6] or when not yet prepared, there must be no indication that skilled technicians would question their likelihood of success in producing an antibody to the antigen.[7]
At issue in Amgen[8]is the extent of enablement required in antibody patents where an entire genus of antibodies is claimed. Like Canada, enablement under US law requires that someone skilled in the art[9] could use specification in a patent to produce the invention. Amgen sought to patent antibodies that bind to a protein called PCSK9 to lower blood cholesterol. Its patent had identified around 400 antibodies theoretically exerting this biological function, but Amgen only provided evidence of enablement of sixteen antibodies. The Federal Circuit Court of Appeals found that Amgen had not met enabling disclosure requirements by failing to enable all embodiments, i.e. antibody species, in the claims. That Amgen had provided instructions to make antibodies with similar properties to those functionally-tested was insufficient to enable preparation of all claimed antibodies. This case also has significance to other types of genus claims, like classes of functionally-similar pharmaceuticals.[10]
Canadian law regarding the extent of enablement across the full claim appears unsettled. In a 2021 decision, the FCA said that it was not enough for the disclosure to enable preferred embodiments, the skilled person must be taught “to put into practice all embodiments of the invention” without undue experimentation.[11] However, earlier decisions indicated that inventors did not need to enable all “modes” of the invention.[12] The Manual of Patent Office Practice acknowledges antibodies defined in functional terms of specific antigen-binding are often a group of structurally different antibodies with the same binding specificity, and further allows enablement that does not require evidence of production of all claimed antibodies, provided there is nothing “peculiar” about that antigen that would call into question the likelihood of a person skilled in the art being able to produce the antibody.[13] Thus, while practice guidelines suggest antibody claims are patentable without enablement across the whole genus in Canada, the recent decision of the FCA shows that the law on enablement across a genus may be shifting in Canada to disallow such patents.
In the US, requiring full enablement is linked to the idea that inventors should be granted a monopoly only for inventions that they have created and described.[14] There is a public interest in not granting overbroad monopolies for genus claims that include unenabled embodiments because this can lead to useful embodiments that require further experimentation never being produced. Yet, for antibodies, focus on antigen binding as defining novelty could suggest that limiting claims to fully-described antibodies would be unfair, if production of a whole genus is the likely outcome of methods described in specification. Inconsistency in FCA holdings on extent of enablement across claims could suggest this issue is ripe for consideration. Overall, Amgen and Canadian law’s approach to antibodies reveal tensions regarding how patent law can accommodate unique features of technological development in a field while also serving public interest.
[1] Patent Act s. 27(3).
[2] Patent Act s. 2.
[3] Manual of Patent Office Practice (“MOPOP”), 23.07
[4] MOPOP 23.07.05; Jackson et al. (1995) J. Imm., 154: 3310-3319.
[5] Ibid.
[6] Ibid.
[7] Re:Immunex Corporation Patent Application No. 583,988 (2011), 89 CPR (4th) 34 (PAB)
[8] Amgen et al. v Sanofi et al. Supreme Court of the United States, No. 21-757.
[9] Patent Act s.27(3)(c)
[10] Brief of Amicus Curiae GlaxoSmithKline PLC in Support of Petitioners in Amgen.
[11] Seedlings Life Science Ventures, LLC v. Pfizer Canada ULC, 2021 FCA 154, para. 68.
[12] Cobalt Pharmaceuticals Co. v. Bayer Inc., 2015 FCA 116, at paras. 68-69
[13] MOPOP, 23.07
[14] Brief for United States of America as Amicus Curiae in Amgen v Sanofi.
Communications Law