Last week, the full US Appeals Court for the Federal Circuit upheld an earlier ruling by a partial panel, invalidating a biotech patent that originated in research at MIT and Harvard. The patent covered any of three ways to disable a signaling pathway involved in the immune response, and would have enabled its licensee, Ariad Pharmaceuticals, to go after companies that already have drugs on the market. The court held, however, that simply specifying different ways of interfering with a protein without any written description of how to do so constituted insufficient grounds for granting a patent.This case, and a similar one (University of Rochester v. Pharmacia) that served as precedent, both followed a similar pattern. In each case, basic research in a university context identified a key protein involved in inflammatory responses. For Rochester, it was they enzyme Cox-2; drugs that inhibit it included Celebrex and Vioxx, both painkillers with lower risk of stomach irritation than aspirin. In the new case, it was the NF-kappaB signaling pathway, which is involved in the immune response to pathogens. Excessive activation of the NF-κB creates chronic inflammation. In this case, Eli Lilly had two drugs already on the market.
In each case, the patent that was granted contained a generic description of how to inhibit the protein involved. For the Ariad patent, three methods were mentioned: blocking the signaling pathways that activate NF-κB, reducing the protein's activity, and preventing the protein from binding DNA. Neither patent specified the actual biochemical mechanism for performing any of these inhibitory functions, nor did they describe any substances that could actually do so. In short, they covered any possible method of targeting a specific biochemical pathway.
As soon as the patents were granted, the grantees turned around and sued the pharmaceutical companies that had actually done the hard work of finding an implementation of these generic concepts.
The first suit, which was decided in 2003, didn't go well for the patent holders, as a judge ruled that the patent, as granted, didn't include a sufficient description of an actual invention. The new case resulted in a jury trial, which Lilly lost, leading its lawyers to request that the verdict be vacated as a matter of law. The presiding judge declined, leading to the appeals process. An initial ruling by a three-judge panel overturned the Ariad patent on the grounds of an insufficient written description, but the full court decided that the issue was significant enough to merit consideration.
The relevant passage of the US patent code was quoted in full in the decision:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
To a certain extent, the argument is over where to put the commas in the statute, which determines what, precisely, needs to appear in the written description. Based on its reading, the Court has determined that "nothing in the statute’s language or grammar that unambiguously dictates that the adequacy of the 'written description of the invention' must be determined solely by whether that description identifies the invention so as to enable one of skill in the art to make and use it." So, even though a skilled biochemist could view the rough descriptions in the patent and know how to develop an inhibitor, that alone is not enough to mean that the patent is valid.
With that interpretation of the law stated, the court focused on whether the patent defined anything that could reasonably be described as an invention. "Every patent must describe an invention," the decision states. "It is part of the quid pro quo of a patent; one describes an invention, and, if the law’s other requirements are met, one obtains a patent."
Its conclusion is that this sort of description is nowhere to be found. Using biological terminology, the Court concluded that the patent describes what it terms a "genus" of inventions, but neglects to specify any "species"—meaning actual chemicals—that are sufficient to show that the patent holder can actually claim to have implemented the genus. Instead, using language from an earlier decision, it dismisses this approach as "no more than a 'wish' or a 'plan'." Just as the court decided in the Rochester case, without a description of an actual chemical that performs these functions, a patent cannot be considered to have provided an adequate written description.
Although it simply reaffirms the Rochester decision, the new opinion attempts to explicitly set a standard for written descriptions that apply generally. And the standard it sets is a rather significant one. Based on the last several decades of biochemistry and molecular biology, it's really easy to identify proteins that help regulate a variety of essential processes, and suggest a variety of routes to inhibiting them. Actually developing something that does so successfully is where the hard work and creativity—the inventiveness, as it were—comes in. In short, this interpretation of the statute appears to bring patent law more in line with the general intent of the patent system.
The decision is also a good sign that the Appeals Court (or at least its clerks) have come to grips with modern biology. The decision includes a decent description of the NF-κB signaling process, and casually discusses testimony regarding whether the protein has a transcription activation domain that's distinct from its DNA binding domain. It's not a stretch to expect that a court that finds the biology mysterious will be more likely to grant deference to technically complex but flawed patents, to the detriment of both research and society. The fact that the court is comfortable discussing biochemistry suggests that this risk is receding.
Source: Arstechnica.com
