By Kevin E. Noonan --
The Supreme Court's decision to grant certiorari in Amgen v. Sanofi is the first time in almost a hundred years that the Court has deigned to consider sufficiency of disclosure decisions, in this case enablement under 35 U.S.C. § 112(a). While these circumstances themselves might motivate amici to file briefs with the Court to weigh in on the Question Presented, the Federal Circuit's trend in recent years to apply more tightly the strictures of Section 112 to chemical and biotechnology inventions and to pharmaceutical cases has provided its own incentive for such briefing.
A total of thirty-four amicus briefs have been filed, wherein twelve support Petitioner Amgen, seventeen support Respondents Sanofi and Regeneron, and five were filed in support of neither party. One of the latter briefs was filed by the Solicitor General representing the U.S. Government, and it is this amicus brief that is discussed in this post.
The Question Presented in the Supreme Court's certiorari grant frames every brief and states:
Whether enablement is governed by the statutory requirement that the specification teach those skilled in the art to "make and use" the invention, 35 U.S.C. § 112, or whether it must instead enable those skilled in the art "to reach the full scope of claimed embodiments" without undue experimentation ― i.e., to cumulatively identify and make all or nearly all embodiments of the invention without substantial "time and effort," Pet. App. 14a [emphasis in Question].
It will be recalled that the Supreme Court issued a Call for the Views of the Solicitor General (CVSG) last year, and in their responsive brief, the government urged the Court not to grant certiorari (for many of the same reasons set forth in their amicus brief).*
The brief's explication of the proceedings below focused on a few key points. First, the Solicitor General asserted the determination by the District Court that the claimed invention required that the "millions" of antibodies produced according to the methods disclosed in the specification would need to be tested to determine whether they would be functional (i.e., in preventing binding of PCSK9 to LDL receptors in vivo). The SG then asserted that there was great uncertainty in the art as to whether an antibody that bound PCSK9 would prevent LDL receptor binding and what the brief stated was a lack of explicit guidance regarding what sequence alterations would result in having or lacking such neutralizing activity. Making these determinations would take "substantial time and effort" which, the SG contends, could amount to undue experimentation even under the conventional test from In re Wands, particularly at what the Federal Circuit had characterized to be the "far corners of the claimed landscape." Amgen had chosen to disclose its claimed antibodies by function rather than structure, according to the SG, which as a consequence required such extensive characterization. In addition, the SG contended that functional descriptions raise "special problems" for enablement, including that a skilled artisan or member of the public can only know whether an antibody falls within the scope of the genus after it is made and tested. "A specification does not enable a person skilled in the art to make and use a product if the person is compelled to engage in the same trial-and-error process the inventor undertook to produce her innovation in the first place" (emphasis added) according to the brief, and this is the situation here.
In something reminiscent of the "quacks like a duck" test, the SG says that this specification, limited to 26 exemplified antibodies (a calculation Amgen disagrees with in its Reply brief) and perhaps some derivatives cannot support the breadth of the claims, particularly when its "roadmap" essentially instructs to "make the antibody pool and fish for the useful ones." Taking recourse in a baking analogy (to be repeated later in the brief, perhaps thinking it might resonate with the Court), the SG states in this regard that "Petitioners may not evade an undue-experimentation problem merely by baking the need for experimentation into their roadmap." The brief makes a comparison between Amgen's 26 exemplars and "structural diversity within the class" of PCSK9-binding antibodies, asserting that Amgen's 26 exemplars "do not even capture the structural diversity represented among the antibodies that petitioners' competitors have already created."
The SG argues that these factual circumstances support the Federal Circuit's decision that the claims were invalid for lack of enablement. The brief rejects Amgen's characterization of the decision as creating a "full-scope" enablement test, and provides a synopsis of the means an inventor can protect an invention of a genus:
Where an inventor develops an innovative process that may be applied across multiple contexts, she can protect it with a process patent. Where she invents an innovative product and others attempt to market that product with insignificant alterations designed to evade an infringement suit, the inventor may use the doctrine of equivalents to enforce her patent against those copyists. But an inventor whose novel product achieves a widely-shared research goal may not obtain a patent for the entire genus of products that perform the same function, thereby foreclosing others from inventing potentially better products that achieve the same goal, unless she provides the information necessary to enable others to make and use the full range of products within the genus.
The SG's legal arguments parallel others who support the Federal Circuit's application of the enablement standard, emphasizing the quid pro quo relationship between patent exclusivity for a limited time and providing an enabling disclosure, as well as encouraging follow-on innovation and ensuring post-expiry that the public has the benefit of the quid pro quo bargain. The brief cites Supreme Court precedent, including Consolidated Electric Light Co. v. McKeesport Light Co., 159 U.S. 465, 476 (1895) (requiring the disclosure to enable the full scope of the claims prevents an "unwarranted extension of [the] monopoly"), stating that "a patent generally cannot satisfy the enablement requirement when its specification references a broad class of substances without giving a person skilled in the art a means of discerning whether and to what extent each substance will produce a working invention." Regarding undue experimentation, the SG argues that it will "generally depend not on any bright-line rule but on a flexible inquiry that takes into account the nature of the claimed invention and the field in which it arises," making the first of two baking analogies regarding a cake recipe that provides merely a list of ingredients and leaves it to the baker to determine how to apportion them in making a cake (non-enabling). The brief then compares baking a cake with making a stew, where it is permissible to instruct the cook to "season to taste" "because similar instructions are standard in recipes, and a more precise instruction is generally impossible given natural variations in ingredients and sodium tolerances." Finally, the SG argues that seasoning (at least with regard to adding salt) would not be permissible for a bread recipe, because the variable ingredient (salt) has a functional consequence to whether the bread is successfully baked; as a consequence the SG asserts that bread recipes "almost invariably specify the precise quantity necessary to produce the chemical reaction that makes the bread rise."
Returning to situations frankly concerned with enablement, the brief cites Wood v. Underhill, 46 U.S. (5 How.) 1, 4 (1846), for situations where what counts is the "degree of vagueness and uncertainty" of the disclosure to the expected/desired outcome, wherein there can be some accommodation to naturally occurring variances (in Wood, the variable compositions of clay used in making bricks). Minerals Separation, Ltd. v. Hyde, 242 U.S. 261 (1916), is cited for the existence of a "reasonableness" standard in the application of the enablement requirement, again dependent on the subject matter of the claimed invention (backed by an impossibility standard for any more precise definition, where it is "obviously impossible to specify in a patent the precise treatment which would be most successful and economical in each case"). But even with these cases the brief asserts there must be a balance between these realities (of inherent uncertainties arising from the subject matter) and the need that "the degree of experimentation required may not be so great as to defeat the basic command of the enablement requirement by forcing others skilled in the art to retrace the patentee's research steps."
The SG also contends that there should be "some general quality" in the members of a claimed genus, which for example was not found among the claimed light filaments asserted against Edison's bamboo filament in Consolidated Electric, as an illustration of how a genus claim can be adequately supported by a generic disclosure. The SG further uses Consolidated Electric to illustrate undue experimentation arising from the need for Edison to exercise "the most careful and painstaking experimentation" that distinguished his light bulb from what had been disclosed and claimed in his competitor's patent, a good but extreme definition of the undue experimentation principle. The brief also set forth Holland Furniture Co. v. Perkins Glue Co., 277 U.S. 245 (1928), as an example of insufficient disclosure for enablement, where starch-based glues were found to be equivalent to animal-based glues when combined at specific proportions with water. In this case the specification provided no characterization of which starches to use or how to select them except by finding that they work when practiced according to the claims, meaning experimentation was required both to practice and avoid practicing the claimed invention. The SG crafts a rule from these cases, that a "specification cannot satisfy Section 112(a)'s enablement requirement merely by providing an example of the products that fall within the class and a generalized description that captures other products similar to the example." Rather, "the patent must describe, with enough particularity to avoid the need for undue experimentation, the structural features that distinguish the genus." (While perhaps helpful it will be immediately appreciated that the enablement devil will be very much in the details of how the rule is applied.)
With regard to the application of these rubrics to Amgen's patents, the SG perceives two types of undue experimentation here: first to find antibodies that bind to the PCSK9 "sweet spot" comprising 16 specific amino acids in PCSK9 itself and then to identify those antibodies that prevent LDL receptor binding, equivalent to requiring independent invention to determine which antibodies are which.** The SG maintained that Amgen's specification fails the test she has crafted because the specification does not identify common structural characteristics "running through" the claimed antibody.*** The SG's brief discounts Amgen's "roadmap" disclosure on the basis that these methods merely produce antibodies but do not necessarily (or with enough certainty) produce antibodies having the functions of the claimed antibodies. And the SG asserts an unsupported speculation (denied by Amgen in its Reply brief), to the effect that "given petitioners' incentive to create and disclose as broad a range of antibodies as possible to bolster their claims, the specifications' inclusion of only 26 antibodies suggests that petitioners do not know how to produce and describe additional exemplars without undue experimentation." The SG supports this argument with comparisons between Amgen's exemplars and competitor antibodies (and the number of sweet spot amino acids they bind to) which suggests to SG that Amgen wasn't able ("did not know how") to make any others. The SG's conclusion is that even for Amgen production of these other antibody species would have required the exercise of undue experimentation.
The brief concludes by more summarily asking the Court to reject Amgen's other arguments, e.g., on the impropriety of the Federal Circuit's "full scope" test, based on the SG's interpretation of Federal Circuit dicta, and that the policy goals Amgen asserts can be better served through other means, including using method claims (suggesting patenting the roadmap, which has the advantage of permitting antibodies to be made by other methods) and resort to the doctrine of equivalents.
The government's amicus brief explicated the history of the proceedings below, including that after two trials the jury in each rendered a verdict that the patents-in-suit had not been shown to be invalid for failing to satisfy the enablement requirement, and that in both instances the Federal Circuit had held otherwise (in this case upholding the District Court's grant of JMOL that the claims were invalid for lacking enablement). Practitioners before the Supreme Court have noted that there is a belief among some that the Supreme Court respects the jury process and the wisdom that can arise from its exercise. The government's brief emphasizing these procedural aspects of the question may put any such faith to the test.
* Amgen in its Reply brief noted that the government's brief mistakenly used the grant date of at least some of Amgen's patents for their filing dates, creating the incorrect impression that Amgen had filed these patents after at least some of their competitors' application filings.
The actual dates for these Amgen patents are:
USP 8,030,457 (granted Oct. 4, 2011), filed Aug. 22, 2008, priority to Aug. 23, 2007
USP 8,829,165 (granted Sept. 9, 2014) filed 4/10/2013, priority to Aug. 23, 2007
USP 8,859,741 (granted Oct. 14, 2014), filed 4/24/2014, priority to Aug. 23, 2007
To be fair, the SG's brief made the same error in at least one of Sanofi's patents:
USP 8,062,640 (granted Nov. 22, 2011), filed 12/15/2009, priority to Dec. 15, 2008
** There was a certain amount of scientific misunderstanding in this portion of the brief, which asserts: "For the remaining 24 exemplars, the specifications provide only the amino acid sequences. Nonetheless, this kind of structural information is generally sufficient to permit a scientist to 'make and use' any of the exemplars she chooses through standard laboratory techniques that allow scientists to use an antibody's amino acid sequence to reverse-engineer additional antibodies with the same sequence"; perhaps the SG meant antibodies with equivalent amino acid sequences.
*** It should be noted that the specification discloses the crystal structure of two of Amgen's antibodies and it is possible that Amgen might have been able, but did not, to extract sufficient common structural characteristics to satisfy the SG's test.
The earliest published application in the name DABUS in the Esp@cenet database is IL289683A of which claim 1 reads:
A food or beverage container comprising: (a) a generally cylindrical wall defining an internal chamber of the container, said wall having interior and exterior surfaces and being of uniform thickness; and (b) a top and a base disposed on either end of said cylindrical wall; wherein said wall has a fractal profile with corresponding convex and concave fractal elements on corresponding ones of said interior and exterior surfaces; and wherein the wall of the container is flexible, permitting flexing of said fractal profile thereof, said fractal profile of said wall permits coupliing by inter-engagement of a plurality of the containers together, and flexibility of said wall permits disengagement of said or any coupling of a plurality of the containers.
Brief perusal of the drawings and of the description and claims makes it clear that a person who believes that the subject-matter was invented by artificial intellegence is likely also to believe that the Earth is flat, that the Sun revolves around the Earth, and that burning materials evolve phlogiston. It would suffice, in my submission, to dispose of these applications on the ground that the allegation that DABUS was the inventor was not credible on the face of the document. It is surprising that hearing officers and judges can get lost in legalism when the simple answer is that the artificial intellegence inventorship is not credible on the face of the document.
Posted by: Paul Cole | April 07, 2023 at 11:04 AM