Last week the Federal Circuit heard oral argument in four cases that attracted amicus briefs. In one of the patent cases, Amgen Inc. v. Sanofi, Aventisub LLC, the court considered the enablement requirement with respect to antibody claims. This is our argument recap.
Jeffrey A. Lamken argued for the appellant, Amgen. Before he could proceed with his first point, Judge Lourie interjected, noting that the claims at issue were composition of matter claims that were claimed by function rather than structure. Further, he noted, the district court found that no structure-function relationship would eliminate the need for undue experimentation and therefore lack of enablement. In this regard, Judge Lourie did not ask a question, but instead commented that Lamken had an uphill battle.
Despite disagreeing that the claims were claimed by function and not structure, Lamken responded that it did not matter. He asserted that two “anchor antibodies” spanned the full area of the PCSK9 antibody “sweet spot,” meaning that “skilled artisans can use them to identify all of the at most 400 distinct antibodies that bind anywhere on that sweet spot.” Further, he argued, whittling down the antibodies to those that bind to the sweet spot could be done for less than $300 and with the basic research tools of the field of antibody research.
Judge Lourie pointed out that the district court was concerned that the claim did not provide guidance on predicting whether an antibody would bind. Lamken responded by arguing that antibody scientists would understand that once you make the sequence you know they will bind, and argued that the specification here told exactly how to make each of the 400 distinct antibodies.
Judge Prost asked how the patent’s roadmap encompassed Sanofi’s alleged infringing antibodies, which she said seemed to function differently from those claimed by binding a different number of antibodies. Lamken explained that expert testimony indicated no antibody scientist would consider the competitor antibodies to be of a different class from those created by the patent’s roadmap.
Matthew M. Wolf argued for Sanofi. He opened his argument by pointing to the number “400,” repeated frequently throughout Lamken’s argument, and emphasized that this number was not in the district court’s opinion because this number was never presented to the district court. When asked how many antibodies would accomplish the function if the patent’s roadmap was used, Wold said Amgen’s lead inventor replied “I don’t know a specific number” and Amgen’s expert answered “I can’t give you a number on what the total is.”
Judge Lourie asked why the enablement requirement has not been met in this case when the written description requirement has been met. Wolf responded by arguing that when you have a functional limitation, an enormous number of candidates, and you would have to test each and every one to see which work, that is a classic example of undue experimentation. According to the district court, he argued, “the fact that you knew there [was] gold in the hills and that you knew how to use a pan to find it, [that] doesn’t mean you are entitled to every ounce of gold in every square mile of the California countryside.”
Judge Hughes asked why requiring a large quantity of experimentation would be considered undue experimentation if qualitatively the experimentation required was minimal and easy. Wolf responded by arguing that Amgen’s own expert testified that testing “millions and millions of antibodies to see whether they would work . . . would be ‘an enormous amount of work’ and more than any scientist would even contemplate doing.”
Judge Hughes then asked whether a genus claim with regard to antibodies should be able to be claimed functionally at all. Wolf replied that he was not sure he would go so far as to say that antibodies may not be covered by genus claims. He argued there may be a time when function dictates structure sufficiently in the antibody field to cross the threshold of predictability, but contended that it was not the case in this set of facts.
In rebuttal, Lamken emphasized that, given the structure and small size of the “sweet spot,” you would expect a small number of antibodies to result from the patent’s roadmap. Moreover, he argued, an expert predicted the roadmap would produce somewhere around 100 antibodies, and Amgen has conservatively argued that number could extend to 400.
Judge Hughes asked why, if millions and millions of tests were required to see if the antibody binds and blocks, that situation would not be undue experimentation. Lamken asserted that, while he believed that number to be greatly exaggerated, experimentation would only be required because of the “risk of an occasional dud,” and the risk of failure defeats the enablement of a process only when such failures are pervasive and frequent. In comparison to prior cases, he maintained, claims failed when thousands of tests were expected to fail, and you were searching for one that might work. Here, he argued by contrast, thousands of tests would be expected to succeed with a possibility of a few duds.
We will keep track of this case and report on its disposition.