Genomics-Based Patents: Are Prophetic Assertions of Utility Enough?

During the late 1990s and early 2000s large-scale gene sequencing projects led to a wealth of newly identified human gene sequences, and a flood of patent applications attempting to lay claim to them. With so many laboratories cranking out newly discovered gene sequences every day there was huge pressure to file patent applications early, in many cases before the biological role of the gene or gene product had been assessed experimentally. In an attempt to satisfy the utility requirement, which essentially requires a patent applicant to disclose a practical use of a genetic sequence that is specific, substantial and credible, patent applicants resorted to the disclosure of “prophetic utilities.” Often times these predictions were based on sequence similarity (i.e., homology) between the sequence of the newly discovered gene product and a protein or family of proteins of known function, or identification of tissue or disease state-specific expression of the gene. In many cases this limited information was supplemented by providing essentially a laundry list of potential uses of the gene, typically involving use of the gene or genes product as a drug or diagnostic, or in the development of a drug or diagnostic. The hope was that at least one of the proposed uses would actually pan out, and the patent applicant would be able to point to the disclosure of the utility in the patent application as evidence that the utility was disclosed at the time of filing.

Although many of these applications resulted in the issuance of patents directed to the gene, the protein product of the gene, antibodies specific for the protein product, etc., it is not at all clear whether these patents would withstand a determined validity challenge during litigation. I am not aware of any case where one of these patents has actually been successfully asserted in court, either in the US or abroad. In my recent survey of human gene patent litigation in the US (“Trend in Human Gene Patent Litigation,” Science, 322:198-99 (2008)), I identified only one lawsuit in which a genomic-based patent was asserted. The lawsuit was filed by Incyte (a leading genomics company during the peak of the gene patenting frenzy, and one of the leading holders of human gene patents) against Invitrogen. Incyte appears to have filed the lawsuit in retaliation for a patent lawsuit that Invitrogen had previously filed against Incyte. Incyte and Invitrogen settled prior to any substantive briefing by the parties, so there is no indication from this lawsuit as to how Incyte's gene patents would have fared against a validity challenge.

Other major genomic companies that patented many genes based on prophetic assertions of utility, such as Human Genome Sciences, Lexicon Genetics and Millenium Pharmaceuticals, have apparently never asserted any of these patents in court. Many of these turn-of-the- century genomic companies have in recent years sought to reposition themselves as pharmaceutical companies, as the genomic-based business models have lost much of their appeal. They have also chosen not to pay maintenance fees on many of these patents, allowing them to enter the public domain. New patent filings on genes have also reportedly dropped off in recent years (see cite in my Science article).

On a number of occasions, the US Patent Office’s Board of Patent Appeals and Interferences (BPAI) has affirmed rejections of genomic-based patent claims under 35 USC 101 and 112 for lack of utility. For example, in Ex Parte Lee (2004 WL 1967421) claims to a newly discovered gene, and the corresponding protein, were rejected for failure to identify a specific utility in the application as filed. On appeal, the applicant attempted to establish utility based on post-filing date gene knockout experiments that allegedly showed that at least one of the predicted utilities actually panned out. However, the Board rejected this argument and affirmed the rejection of both the protein and DNA claims, finding that the predicted utilities were only speculative at the time the application was filed, and there was no evidence that the utility had been adequately demonstrated as of the filing date of the application.

For another example, see Ex Parte Lal, 2007 WL 1878008, an unsuccessful appeal by Incyte of a rejection based on lack of utility. The application as filed disclosed that the most similar prior art sequence the applicant could find to the protein product of the claimed gene sequence was a taste receptor, and Incyte argued that this constituted a sufficient assertion of utility. The BPAI rejected this argument, and found that post-filing date evidence that the protein actually was a taste receptor was insufficient, since the utility had not been established as of the date of filing.

On July 31, 2008 the British Royal Courts of Justice revoked a Human Genome Sciences (HGS) genomic patent claiming a member of the TNF ligand superfamily, which HGS identified as “Neutrokine-alpha.” (Eli Lilly v. Human Genome Sciences [2008] EWHC 1903 (Pat)). HGS knew at the time it filed its patent application that the newly discovered gene appeared to be a member of the TNF ligand superfamily, based on sequence similarity. The patent application discloses this fact, and includes a host of predicted utilities, many of them based on the known diverse activities of other members of the TNF superfamily. The British court found the claims to be invalid based on lack of industrial applicability (utility), insufficiency (enablement) and obviousness. In particular, the Court found that by disclosing a huge number of speculative (and often contradictory) utilities, the application in effect failed to provide any real teaching of utility and thus fail to solve any technical problem. The court was highly critical of this laundry list approach to satisfying the industrial applicability requirement. The decision cites a number of EPO Technical Board of Appeals decisions relating to patent applications disclosing prophetic utilities for biological molecules and arriving at a similar outcome.

One aspect of the British decision I found interesting, and quite different from US practice, was the Court’s ruling on obviousness. Lilly raised a number of obviousness challenges to the patent, based on prior art disclosures of clones containing partial sequences of the claimed gene, e.g., EST sequences, but the Court held that these partial sequences were insufficient to render HGS’s successful cloning of the full-length gene obvious. However, the Court held that since the specification contains more no more than a range of speculative applications of the gene, which the Court characterized as implausible, it fails to teach a person skilled in the art how to solve any technical problem. In essence, the court finds that because the applicant failed to meaningfully identify a use for the gene, the patent was invalid not only for lack of industrial activity, but also for lack of inventive step, i.e., obviousness.