Amicus Briefs in Ariad v. Eli Lilly: Professor Christopher Holman

    By Kevin E. Noonan --

Christopher M. Holman, Associate Professor of Law at the University of Missouri-Kansas City (and proprietor of Holman's Biotech IP Blog), has filed an amicus brief in support of neither party in Ariad Pharmaceuticals, Inc. v. Eli Lilly & Co.  While not arguing on behalf of either party, Ariad would benefit if the Federal Circuit adopts Professor Holman's position:  he not only advocates that there is no separate written description requirement in 35 U.S.C. § 112, first paragraph, but asserts that the requirement is differentially (and improperly) applied to biotechnology inventions.

The brief begins by arguing that the Federal Circuit "established a novel interpretation of the written description requirement" in Regents of the University of California v. Eli Lilly & Co.  His argument is not only that this requirement was new (and unnecessary), but that the Court has applied it in a manner "that is essentially redundant to the enablement requirement" and that acts as a super-enablement requirement for biotechnology claims reciting protein and DNA subject matter.  Citing Enzo Biochem, Inc. v. GenProbe Inc., he argues that the Court has applied the requirement in an arbitrary and inconsistent manner that lacks any basis in science and can, in some instances, preclude adequate patent protection for biotechnology inventions.  The enablement requirement, grounded in the "objective" considerations exemplified by the factors enunciated in In re Wands, is the proper method for assessing whether the specification describes the invention sufficiently, according to Professor Holman.

His argument falters somewhat when he discusses the requirement that at least one species must be disclosed to support a claim, as opposed to the necessity under the written description requirement for "something more" as interpreted under the Lilly precedent.  The problem with this argument is that while priority of invention requires merely that the specification teaches how to make and use "at least one" species falling within the scope of the claim (Hahn v. Wong, 892 F.2 1028 (Fed. Cir. 1989)), enablement requires a claim to be supported by a specification that enables the skilled person in the art to make and use the claimed invention throughout its scope (M.P.E.P. § 2164.08; In re Wright, 999 F.2d 1557 (Fed. Cir. 1993)).  He illustrates what he considers to be a major inconsistency in the Court's written description jurisprudence by contrasting the Court's decisions in Noelle v. Lederman (for proteins) and Falkner v. Inglis (for DNA species) with the strict interpretation of the Lilly decision.  In Noelle, the Court upheld the validity of claims to antibodies where the specification describes only their cognate antigens, while in Falkner, the Court upheld the validity of claims directed to virus species having non-specific deletion mutations resulting in an observable phenotype.  These cases are inconsistent with Lilly, according to Professor Holman, in that Lilly requires "super-enablement" disclosure of human insulin while Noelle and Falkner permit patenting of undisclosed proteins (antibodies in Noelle) or DNAs (viruses in Falkner).

This argument misses the reasoning behind the seeming inconsistency.  In both Noelle and Falkner, the patentees were claiming species that could be reliably produced by one having skill in the art, wherein the impediment to producing the invention was overcome by the patentees.  (For example, in Noelle, the Court required there to be a novel antigen disclosed in the specification in order for generic antibody claims to be patentable.)  In Lilly, on the other hand, the patentees were claiming a species -- human insulin DNA -- that they neither possessed, conceived, nor described.  In Lilly, permitting the patentee to bootstrap its cloning of the rat insulin gene to support a claim for human insulin would violate several long-standing principles, priority of invention being but one of them.

Professor Holman also cites the Enzo debacle, and makes an interesting argument.  He posits that Enzo I was a literal reading of the Lilly precedent, and as a consequence, reversal of the initial panel decision in Enzo II was a repudiation of the "super-enablement" written description requirement enunciated in Lilly.  Another way of viewing the situation is that Enzo I was an over-interpretation of the Lilly precedent, ignoring the portion of that decision that said (emphasis added):

Thus, as we have previously held, a cDNA is not defined or described by the mere name "cDNA," even if accompanied by the name of the protein that it encodes, but requires a kind of specificity usually achieved by means of the recitation of the sequence of nucleotides that make up the DNA . . . .  A description of a genus of cDNAs may be achieved by means of a recitation of a representative number of cDNAs, defined by nucleotide sequence, falling within the scope of the genus or of a recitation of structural features common to the members of the genus, which features constitute a substantial portion of the genus . . . .  This is analogous to enablement of a genus under § 112, P 1, by showing the enablement of a representative number of species within the genus.  We will not speculate in what other ways a broad genus of genetic material may be properly described, but it is clear to us, as it was to the district court, that the claimed genera of vertebrate and mammal cDNA are not described by the general language of the '525 patent's written description supported only by the specific nucleotide sequence of rat insulin.

Considering this language, the decision in Enzo I was in error not for applying the Lilly test properly, but for disregarding the fact that a deposit satisfies both the written description and the enablement requirements because it places the public in possession of the claimed embodiment directly.

Another case Professor Holman cites as evidence of the error of the Lilly decision is In re Wallach, where the applicant had disclosed an enabling method for producing protein embodiments of the claimed invention (which the Court found to be patentable) but did not disclose the nucleic acid encoding the protein (claims to which were unpatentable for failure to satisfy the written description requirement).  According to Professor Holman, the Court's decision in In re Kubin points out the error in the Wallach decision.  In Kubin, the Court held that claims to a DNA encoding a known protein were obvious, in view of the routine nature of cloning methods in the art and the existence of both the known protein and a probe (a monoclonal antibody) specific for the protein.  The Court held that the Supreme Court in KSR International Co. v. Teleflex Inc. had repudiated the Federal Circuit's earlier jurisprudence on the question of the obviousness of a nucleic acid encoding a known protein as set forth in In re Deuel.  That jurisprudence was based on traditional rubrics of chemical obviousness, wherein invalidating prior art disclosed a compound having a structure sufficiently related to the claimed structure to render the latter obvious.  Spurred on by its interpretation of KSR, the Kubin panel instead determined that the claimed nucleic acid was obvious substantially because it satisfied the enablement standard -- that it could be obtained from the prior art without undue experimentation.

The connection between the chemical obviousness cases and written description is made by Professor Holman by way of a quotation from Lilly, citing In re Deuel and In re Bell to the effect that if a chemical structure must be disclosed in the prior art to render a nucleic acid obvious, "a fortiori, a description that does not render a claimed invention obvious does not sufficiently describe that invention for purposes of §112, ¶1."  For Professor Holman, the intervening Kubin decision destroys this rationale, since for the Kubin court, a prior art description no longer requires disclosure of a chemical structure to render a nucleic acid claim obvious.  This argument fails to consider two relevant facts, however.  First, even the Patent Office conceded that the factual underpinnings in Kubin were vastly different from the facts in Deuel, and that the obviousness of the Kubin claims was directly related to those differences (including the existence of a specific antibody probe, advances in the level of skill in the art, and a non-prior art example of cloning the mouse homolog using a combination of the antibody probe and routine methods in the art).  Second, the written description and obviousness prohibitions are not competitive but complementary:  the concept of obviousness precludes patenting for an invention that does not sufficiently "promote the progress" of the relevant art, while the written description requirement, as applied under the Lilly precedent, does not permit a patentee to claim what the patentee has not actually obtained.  The fact that it could have been obtained is deemed insufficient.

The brief castigates the Federal Circuit for inconsistencies in applying the written description requirement, and as a consequence asserts that this has prevented the Patent Office from arriving at proper standards for assessing the patentability of biotechnology claims.  Professor Holman discusses the Written Description Training Materials promulgated in March 2008 in this regard and contrasts the treatment of nucleic acid claims in Wallach with antibody claims (under the Noelle case precedent; Example 13), with regard to the "billions" of variants in the (appropriately-named) variable region and the laxity with which the Office applies the written description requirement (essentially providing no barrier to patentability despite the complete absence of a description of any actually-produced antibody).  Correctly, he characterizes this treatment as being inconsistent and rejects the policy reasons for these differences as discussed above and enunciated by the Office.  He argues that the existence of variable regions in antibodies provides just as much uncertainty as the sequence of an unknown nucleic acid, and permitting antibodies to be patented with little or no disclosure of structure provides the basis for the inconsistency.  This basis, of course, is (ironically) rooted in the enablement requirement, because antibodies inherently select just those (generally unknown) sequences that specifically bind their cognate antigen by the way they are produced.  This specificity is not inherently a part of cloning methods used to produce nucleic acids, thus raising the heightened description requirement for such claims.

The Carnegie-Mellon University v. Hoffman-LaRoche Inc. case is also discussed in the brief with regard to how the written description requirement is inconsistently applied.  In this portion of the brief, Professor Holman makes an interesting argument that turns on their heads traditional rationales against broad variant nucleic acid claims.  He asserts, based on several scientific treatises, that it is "generally possible to substantially alter a protein's structure without necessarily destroying its function.  For example, in some cases 50% or more of the amino acids in a protein can be altered while still maintaining function."  He then acknowledges the more conventional bases for requiring actual written descriptions of such functional variants, because "a single amino acid change can sometimes result in loss of function," citing his own scientific work.  He also correctly identifies the latter situation as the basis for requiring a priori "some level [of] disclosure of the structure-function relationship" that would permit the skilled worker to appreciate "which variants" will retain function.  This he properly characterizes as an "impossible" standard.  Given the very large number of variants for even a small (300 amino acid) protein (10²²⁶), this results in claims that cannot properly protect their inventions, since it is relatively easy for an infringer to produce a (presumably non-infringing) functional amino acid sequence variant.

To illustrate, he contrasts the court's decisions in Capon v. Esshar (regarding recombinant antibodies) and Carnegie-Mellon (regarding DNA polymerase I enzymes).  In Capon, the Court held that a description of "well-known" sequences encoding alternative antigen-binding portions of proteins expressed at the surface of an immune system cell that triggers proliferation thereof was not necessary to satisfy the written description requirement.  In contrast, the Court in Carnegie-Mellon invalidated claims to recombinant constructs for producing bacterial DNA polymerases because the specification disclosed only the polymerase from E. coli, and only two other bacterial DNA polymerases were known in the art at the time the invention was made.  For Professor Holman, the existence of "thousands" of mouse antibody-derived antigen-binding species in the art has but "superficial appeal" as a rationale, in view of the great amino acid sequence variability in antibodies, and the structural similarity of bacterial polymerases.  He maintains that polymerases being enzymes all have substantially the same function, and makes the reasonable presumption that this function is dictated by structure (conferred by the protein's amino acid sequence) and thus "there is a relatively strict correlation between structure and function in the class of bacterial DNA polymerase I proteins."  This he contrasts with the recognized hypervariability of antigen-binding portions of antibodies as further examples of the inconsistent application of the written description requirement.

Professor Holman returns to the PTO Training Materials with regard to Examples 11A and 11B, which he evocatively characterizes as "junk science."  Specifically, he raises the apparent paradox between what the Office considers unpatentable (Example 11A) and patentable (Example 11B) claims to amino acid sequence variants, based on disclosure of structure-function relationships of the variants in the supporting specification.  The claims in both examples recite a nucleic acid encoding an amino acid sequence 85% identical to the disclosed sequence.  The difference in the patentability of the claims is that the specification supporting the claim in Example 11A provides no evidence of any structure-function relationship, while the specification supporting the claim in Example 11B discloses a deletion variant that identifies two domains "critical" for the protein's function.  The "Catch-22" nature of these examples has been discussed on Patent Docs previously (see "An Analysis of the New Written Description Training Materials - DNA Hybridization & Percent Identity" and "Docs at BIO: 'Gotcha' Games Continue at USPTO"); here, Professor Holman asserts that "[t]he PTO's analysis fails to recognize the complex and unpredictable relationship between protein structure and function, and thus fails to comport with scientific reality."  (Of course, the irony is that these examples are an attempt, however flawed, to do just what Professor Holman contends that they do not do.)  The basis for Professor Holman's contention is that the general tendency of non-conservative substitutions to be more likely than not to result in loss-of-function is "anything but hard and fast rules," because while the scientific literature is replete with examples consistent with Patent Office assumptions, it is also the case that there are several examples to the contrary, e.g., where mutations at one site can correct or compensate for mutations at another, spatially-distinct site, or where a conservative substitution results in a greater loss of function than a non-conservative one (citing several scientific references including one of his own).  "The bottom line," according to Professor Holman, is that the deletion studies in Example 11B of the Training Materials are "but the slightest tip of the iceberg, and [do] not identify regions of the protein that could or could not be altered without a loss of function."  Finally, he correctly notes that the "patentable" claim of Example 11B is not limited to conservative substitutions or mutations outside the identified functional domains, and thus are likely to include non-described, inoperative embodiments.

In the penultimate portion of the brief, Professor Holman discusses the history of the development of the written description requirement, from the "simultaneous conception and reduction to practice" concept announced by the court in Amgen Inc. v. Chugai Pharmaceutical Co., to the decision in Fiers v. Revel that a nucleic acid claim must be supported by an actual description of the claimed nucleic acid in order to show possession, and finally to the Lilly decision.  (He also mentions In re Smythe, but dismisses it as being "not even a chemical case.")  He asserts that the Court has "never articulated a rational justification" for the written description requirement as it has been applied to "some" nucleic acid and protein inventions.  In conjunction with the argument (discussed above) that the Court's Kubin decision rebuts the legal basis for the Lilly decision, he also attacks the basic premise of that decision:  according to Professor Holman, a claim to a "gene that controls blood sugar" would be a functional definition, while a claim to "human insulin" is not, since the latter "describes" a "specific gene that was known to exist at the time of the patent application."  While this is true, the argument appears to miss the point that while the gene existed, the patentee did not know or disclose its structure.

Finally, Professor Holman argues that the enablement requirement is sufficient to ensure "adequate disclosure across all technologies."  He notes instances where the enablement requirement has been used to "police" the scope of biotechnology claims, and asserts that the "well-developed" body of enablement case law is more suited to the task than the Court's written description jurisprudence.  (He doesn't mention that, as a question of law, such consistency should be more likely than for questions of fact, such as whether the written description requirement is met.)  He consigns the outcomes of "the Amgen-Fiers-Lilly trilogy" to being appropriate under the "unpredictability of cloning technology at the time," but says this is no longer warranted in view of the technology having become "conventional and relatively predictable," citing Kubin and argues:

Just as in Kubin this Court chose to discard the strict requirement of structural disclosure in the obviousness determination arguably required by Bell and Deuel, it should do away with LWD [Lilly written description] and its irrational focus on structure, and employ the more adaptable and appropriate enablement requirement to assess adequate disclosure of biotechnological and other inventions.

Professor Holman cites almost as many law reviews and opinions of other legal scholars as he does Federal Circuit and CCPA precedent in support of his arguments.  Noticeably missing from the citation list, however, are the cases cited in Lilly to support its formulation of the written description requirement, including In re Wilder, 736 F.2d 1516 (Fed. Cir. 1984); In re Robins, 429 F.2d 452 (CCPA 1970); In re Grimme, 274 F.2d 949 (CCPA 1960); In re Angstadt, 537 F.2d 498 (CCPA 1976) and In re Gosteli, 872 F.2d 1008 (Fed. Cir. 1989).  This deficiency may compromise the persuasiveness of Professor Holman's otherwise thought-provoking brief.

For additional information regarding this topic, please see:
• "Lilly Files Principal Brief for Ariad v. Lilly Rehearing En Banc," November 16, 2009
• "Next Up: Ariad v. Lilly Rehearing En Banc," November 10, 2009
• "Federal Circuit Grants En Banc Review in Ariad v. Lilly," August 21, 2009
• "Ariad Files Petition for Rhearing in Ariad v. Lilly," June 3, 2009
• "Ariad Decision Voids Attempt to Use Broad Claiming to Avoid the Written Description Requirement," April 14, 2009
• "Ariad Pharmaceuticals, Inc. v. Eli Lilly and Co. (Fed. Cir. 2009)," April 6, 2009

Comments

[SKD]

The problem with the patent office examiners is that most of them neither understand the complexities of the written description case law nor the nuances of the scientific differences being claimed by the present day biotech inventions. In part that is because of the Fed. Cir. seemingly higher burden for biotech patents, because it seems, any thing that relates to biotech appears to be unpredictable and any knowledge of the skilled artisan is uncoupled from the analysis.

[Allison Williams Dobson]

In 1911, Judge Hand concluded the opinion in Parke-Davis v. H.K. Mulford Co. with the following paragraph:

I cannot stop without calling attention to the extraordinary condition of the law which makes it possible for a man without any knowledge of even the rudiments of chemistry to pass upon such questions as these. The inordinate expense of time is the least of the resulting evils, for only a trained chemist is really capable of passing upon such facts, e.g., in this case the chemical character of Von Furth's so-called "zinc compound," or the presence of inactive organic substances. In Germany, where the national spirit eagerly seeks for all the assistance it can get from the whole range of human knowledge, they do quite differently. The court summons technical judges to whom technical questions are submitted and who can intelligently pass upon the issues without blindly groping among testimony upon matters wholly out of their ken. How long we shall continue to blunder along without the aid of unpartisan and authoritative scientific assistance in the administration of justice, no one knows; but all fair persons not conventionalized by provincial legal habits of mind ought, I should think, unite to effect some such advance.

Sadly, his call for change went unheeded. As Professor Holman has pointed out, the body of biotechnology case law is rife with inconsistencies. Generalist judges cannot be expected to master the field of molecular biology, so what else can we expect as long as the current model endures?

[Kevin E. Noonan]

Dear SKD:

Yes and no. I think that one of the problems with PTO examination is that the kinds of bright-line rules that are easy to apply with simple technology get harder to apply as the technology gets more complex. It isn't that the knowledge of the skilled worker is disregarded, but that the skilled worker knows less a priori the more unpredictable the technology becomes.

Thanks for the comment.

[Kevin E. Noonan]

Dear AWD:

What I have always found most bracing about Judge Hand's legal analyses is the courage and humility with which he addressed his own lack of knowledge. He did not (as some current members of the Supreme Court have made it a habit of doing) believe that the science needed to bend to the law. His was not a simple "We decide." Your quotation illustrates his commitment to finding the truth, wherever it may lay and using whatever tools he could command, without the arrogance of thinking that finding and using such tools were beneath the court's notice. We live with the great fiction that judges judge and that we must be satisfied with those judgments; Judge Hand points out that there is another way, if we were only smart enough to admit we should take it.

Thanks for the comment.

[Allison Williams Dobson]

Well put, Kevin. Thanks for the reply and for a consistently doing a great job with your blog.