Patenting antibodies has long been challenging. Although most inventions can be patented based on their functionality, assuming the functionality is new and non-obvious, for antibodies and other biomolecules there is a higher standard. Although the antibody standard was recently challenged, the U.S. Court of Appeals for the Federal Circuit (CAFC) decided to uphold it.
The Historic Antibody Rule
One of the main requirements for a patent is that the invention be described thoroughly enough to enable the reader to make and use the invention, given that the reader is skilled in the relevant field, and given that a certain level of experimentation is expected. Section 112(a) of the Patent Act states (in part) that “The specification shall contain a written description of the invention, and of the manner and process of making and using it… as to enable any person skilled in the art… to make and use the same.” The main function of antibodies is to recognize specific antigens through targeted binding. Thus, when new antibodies are invented or discovered, their valuable function is their ability to specifically bind antigens. In the early days of antibody technology, antibodies were claimed according to this function, by claiming for example “An antibody that binds to antigen X.”
The Modern Antibody Rule
In a landmark case of AbbVie Deutschland GmbH & Co. v. Janssen Bio-tech, Inc., the CAFC decided that claiming an antibody as defined by its antigen was too permissive. It reasoned that it is too difficult to predict whether a given antibody will bind a given antigen, and it would require excessive testing (“undue experimentation”) to screen all possible antibodies for this function. At that time, the field had very little understanding of how the structure of an antibody correlates to its binding target and specificity. Typically monoclonal antibody lines were screened and selected based on binding characteristics, rather than being designed or engineered. Often the structure of the antibody was analyzed after screening. An antibody could be claimed in a patent based on its structure (including its 3D structure or amino acid sequence) and based on its parental monoclonal cell line, but not based on its binding target. Similar logic was used to prevent the patenting of other biomolecules with specific binding characteristics. However, in its In re Wands opinion, the CAFC upheld a patent for “monoclonal high affinity IgM antibodies immunoreactive with HBsAg determinants” due to the specific source and structure included in the claim. Another exception to this rule is when the invention is a nucleic acid (e.g., DNA or RNA); because the mechanism of nucleic acid binding to other nucleic acids is well understood, an inventor is allowed to define a nucleic acid invention based on its binding partner.
The Recent Challenge in Amgen
The antibody rule was recently challenged in Amgen Inc. v. Sanofi, Aventisub LLC. The inventors in that case developed numerous lines of monoclonal antibodies capable of binding to the proprotein convertase subtilisin/kexin type 9 (“PCSK9”) enzyme. PCSK9 plays a role in cholesterol metabolism, and the inventors discovered that certain antibodies that bind at specific locations on the enzyme potentially reduce circulating low density lipoprotein (LDL – sometimes referred to as “bad cholesterol”). Numerous examples of such antibodies were described in the patents. Each of the patents claimed the antibodies based on the ability to bind to specific parts of PCSK9. The patents did not claim the antibodies based on anything about the antibodies’ structures, such as amino acid sequences or 3D structures. When challenged, the inventors argued that the many examples of the inventive antibodies in the patents gave enough information to the reader to make and use the invention, which is all that is required by the law. The inventors also argued that the challenger was unable to identify any particular claimed antibody that a skilled person could not make and use without undue experimentation. On the other hand, the challenger argued that the patent potentially covered millions of individual antibodies, and it was beyond the capabilities of even a skilled person to make and use every single one.
The trial court held that even with many working examples of antibodies capable of binding as claimed, one skilled in the art would not be able to make and use the broad class (or “genus”) of antibodies capable of such binding. On appeal a three-judge panel of the CAFC agreed. The panel agreed with the challenger that “it is important to consider the quantity of experimentation that would be required to make and use, not only the limited number of embodiments that the patent discloses, but also the full scope of the claim.” The court cited past cases in which the claims were limited to biomolecules having certain structures, and so “the full scope of the claim[s]” was also limited; in those cases the courts usually found the claims met the enablement requirement. However, in cases such as the one at bar, where there were no structural limits on the antibodies, the genus of antibodies claimed is too vast to be enabled based only on a few dozen examples.
Amgen requested that the CAFC reconsider the decision by convening every judge on the circuit en banc. When the CAFC decides matters en banc it creates a precedent that must be followed. However, the request was refused. Three of the judges issued an accompanying opinion as to why en banc review was unnecessary was issued, reinforcing the logic of the original decision and mocking those who disagree (“It seems to them that the sky is falling”). The accompanying opinion emphasized that the standard applied in Amgen is an application of prior precedent, and merely maintains the status quo ante. It pointed out that patents for non-biological chemicals typically contain numerous examples of versions of the chemical, including structures and instructions for synthesis. The relevance of this fact is not clear, as Amgen’s patent also contained numerous examples of its antibodies. The opinion also characterized functional claims as “the problem,” although functional claims are the only type of patent claim affirmatively authorized in the Patent Act (in Section 112(f)).
The Lessons for Antibody Inventions
Despite consistent precedent from the courts and the U.S. Patent and Trademark Office on this point, many patent applicants attempt to patent antibodies based on their binding characteristics, either in defiance of the AbbVie rule, or in ignorance of it.
The bottom line is that claiming antibodies and other biomolecules based on binding characteristics is going to remain difficult, especially when there is no understanding of how the molecule’s structure contributes to binding the target. When patenting biomolecules that are useful due to their binding characteristics, make sure to prepare robust claims that do not solely depend in identifying the binding target. The Amgen case is instructive, because the inventor claimed the antibody not only by its target molecule, but more specifically by the portion of the molecule that is targeted. This was not availing, however. The court did not distinguish between a claim to a targeted molecule or a targeted portion of a molecule. Patent applicants should take note, and avoid relying solely on the specificity of the target to support the patentability of the antibody.
Robust claims protecting antibody-based inventions can be drafted for the applications of the antibody. Such applications commonly include methods of diagnosis, Western blotting, non-diagnostic immunoassays, methods of medical treatment, personalized medicine (i.e., methods of diagnosis coupled to methods of treatment), and prophylaxis (e.g., prevention of prenatal hemolytic disease). For example, while a claim for “An antibody capable of binding enzyme X” is very difficult to patent, one can patent “A method of detecting enzyme X comprising binding enzyme X to an antibody.” Specific antibodies can be patented by patenting the hybridoma lines used to produce them, as well. An antibody can be patented by claiming its specific structure, independent of its functionality, although such claims are often too narrow to be very valuable.
Alternatively, antibodies can be protected by other forms of intellectual property. Many commercial antibodies are never patented; instead the hybridoma lines used to produce them are maintained as trade secrets, and the antibodies are either sold to others or used in-house. Selling antibodies usually will not compromise the confidentiality of the hybridoma line, because of the difficulty in reverse engineering the hybridoma line based on the antibody. Some business models involve the distribution of the hybridoma line under a material transfer agreement (MTA) or other form of nondisclosure agreement. This can be preferable to patenting for numerous reasons. No government approval is required, avoiding the need to meet the enablement requirement or the other requirements for novelty, non-obviousness, etc. Preparing and prosecuting a patent application incurs significant legal costs; although protecting trade secrets does also, the costs are typically much lower. Patents have finite lifespans, and require maintenance fees. Trade secrets have no set lifespan, and require no maintenance fees. The one weakness of trade secrecy protection is that it is not easily compatible with regulatory requirements for biologic medical treatments. These require large amounts of disclosure to the regulatory agency that could compromise the confidentiality of the hybridoma lines.
Although antibody technology is not new, it remains one of the most nuanced areas of technology for patenting. Intellectual property strategy for antibody technology should consider whether patenting is practical or desirable, and always look to the option of using trade secrecy instead. When patents are sought, never assume that a patent can be obtained based only on the antibody’s target, but rather formulate alternative strategies for effectively patenting the antibody (such as based on its structure) or its likely uses.