Some Thoughts on Amgen v. Sanofi

By Chris Holman

Much has been already been written about the Supreme Court’s recent decision in Amgen v. Sanofi, including Dennis’s prompt response. Speaking for myself, I read Amgen as an endorsement of the Federal Circuit’s current interpretation and application of the enablement requirement, and maintenance of the status quo. The enablement requirement is an important (and I would argue the primary) non-prior art-based doctrinal tool for policing claim scope, and as such it plays an important role in preventing inventors of pioneering inventions from laying too much claim to subsequent innovation.

As the Supreme Court observed in Amgen, although Sawyer and Man might have been the first to invent an electric lamp filament of “carbonized fibrous or textile material,” this achievement did not necessarily entitle them to a patent encompassing all electric lamps filaments constructed of this broadly defined class of material, including Edison’s subsequent, but apparently vastly superior, bamboo-based filament. Similarly, Samuel Morse’s invention of the telegraph did not mean he was allowed a patent encompassing any use of electricity “for marking or printing intelligible characters, signs, or letters, at any distance” (sounds like the internet). And in the realm of biopharmaceuticals, the discovery of an antibody capable of specifically binding a particular epitope does not necessarily justify a patent claim encompassing all antibodies specific for that epitope, including antibodies having different (or even superior) pharmaceutical function, a point raised by Sanofi in its briefing of Amgen.

For years, this has been Federal Circuit’s interpretation of the enablement requirement, which requires the scope of the patent’s disclosure to be “commensurate” with the scope of the claims. In Amgen the Supreme Court essentially paraphrased this long-standing standard, stating that “we understand [the Federal Circuit] to have recognized only that the more a party claims for itself the more it must enable. [T]hat much is entirely consistent with Congress’s directive and this Court’s precedents.”

At the same time, it has also long been understood that a patentee is not required to literally enable a person of skill in the art to make and use every embodiment falling within the scope of a patent claim in order to satisfy the enablement requirement. For example, it is uncontroversial that a patent claim broadly directed towards pharmaceutical compositions comprising a specified active ingredient is not invalid for lack of enablement simply because the claim encompasses improved formulations of that active ingredient (e.g., timed release or combination products) that are not literally enabled by the original patent.  In many cases, such formulations are sufficiently innovative to justify their own patent, sometimes referred to as an “improvement patent.”  I do not think that anything in Amgen should be interpreted as altering this fundamental understanding of the enablement requirement.

Over the last few years I have written quite a bit about the enablement requirement, particularly in the context of therapeutic antibodies and other pharmaceuticals. For those who might be interested, here are some links to those articles, along with a brief description of some of the points raised in those articles.

I will begin with two articles I co-authored with Professor S. Sean Tu, entitled Antibody Patents: Use of the Written Description and Enablement Requirements at the Patent & Trademark Office and Antibody Claims and the Evolution of the Written Description / Enablement Requirement.  These articles report the results of an empirical study wherein we found that over the last couple of decades patent examiners faced with antibody claims have steadily increased the use of enablement and written description rejections, and have shifted from the issuance of broad, functionally-defined antibody claims (such as the claims struck down in Amgen) to narrower, structurally-defined claims (such as the sequence-based patent claims Amgen and Sanofi have obtained on their respective products, which the Federal Circuit and Supreme Court both pointed to with seeming approval).

In another article, Amgen v. Sanofi: The Supreme Court Takes up the Enablement Requirement in the Context of Therapeutic Monoclonal Antibodies, I offered some thoughts on Amgen prior to the Supreme Court deciding the case.   The article summarizes some of the main arguments that were raised in Amgen’s successful petition for certiorari, as well as Sanofi’s brief in opposition and a number of amici curiae briefs filed by various interested parties, including the U. S. government’s brief recommending that the Court not grant certiorari.

Amgen’s petition for certiorari relied heavily upon a 2020 law review article entitled The Death of the Genus Claim (“Death”), which purports to document a dramatic shift in the Federal Circuit’s interpretation of the enablement and written description requirements that has rendered it nearly impossible to obtain a chemical genus claim that will withstand judicial scrutiny.  Perhaps Death played some role in the Court’s decision to grant certiorari.  In any event, I responded to Death with my own two-part article, entitled Is the Chemical Genus Claim Really “Dead” at the Federal Circuit?, Parts I and II.

In Part I, I analyzed the judicial decisions upon which Death bases its claim, and explain why, in my view, the article does not substantiate its claim of a marked shift in the Federal Circuit’s interpretation and application of 112(a).  In Part II, I push back against one particular assertion made in Death, i.e., the suggestion that patentees can circumvent the Federal Circuit’s purported heightened application of 112(a) to chemical genus claims by drafting broader claims that define chemical genuses solely in structural terms, without the inclusion of any functional limitations. Part II then reviews a substantial number of judicial decisions involving chemical genus claims, and finds little evidence of a pronounced change in the application of 112(a) to chemical genus claims over the time span which Death identifies as corresponding to a purported dramatic shift in the law.

In another article, The Enablement and Written Description Requirements Through the Lens of the Federal Circuit’s Actions (As Opposed to Its Words), I observe that, when applying patent law’s enablement and written description requirements to police the scope of chemical genus claims, there can at times be a marked divergence between what the Federal Circuit says the law is and what the court actually does. For example, although the Federal Circuit has repeatedly stated that a patentee must enable the “full scope” of a patent claim in order to comply with 112(a), that is clearly not literally true – rarely if ever does a patent enable every conceivable embodiment that would fall within the scope of a patent claim, a task that would be impossible in most (if not all) cases, given the open-ended nature of most patent claims.  In Amgen, the Supreme Court expressly observed that “we review judgments of the lower courts, not statements in their opinions,” which I interpret as tacit recognition of the importance of focusing more on the actions of the Federal Circuit, as opposed to statements in the opinions that can sometimes be misleading when divorced from context.

10 thoughts on “Some Thoughts on Amgen v. Sanofi

  1. 5

    The first sentence of that paragraph is as follows — “At the same time, it has also long been understood that a patentee is not required to literally enable a person of skill in the art to make and use every embodiment falling within the scope of a patent claim in order to satisfy the enablement requirement.”

    But, that is essentially what the Federal Circuit and the Supreme Court have held for genus antibody claims.

    Your excipient example is not the same. It is not a modification of the claimed antibody; it is an additive.

    1. 5.1

      But, that is essentially what the Federal Circuit and the Supreme Court have held for genus antibody claims.

      I am not sure that it makes sense to extend Amgen to all antibody genus claims. It seems to me to be more properly limited to antibody genus claims defined functionally. It is perfectly possible to define a genus of antibodies structurally (e.g., “… having at least 85% sequence identity to SEQ ID NO:1…”), and that might still hold up under Amgen.

      Indeed, the Amgen court fairly well shouts that structurally defined genus claims are fine when it explains that the difference between Morse’s valid claims and his invalid claims was the difference between structural and functional definitions of the claimed invention. Slip op. at 9 (“Morse’s patent included eight claims, and this Court had no trouble upholding seven of them—those limited to the telegraphic structures and systems he had designed”).

      1. 5.1.1

        ‘al”l”…’

        A rather odd inclusion – so the dude likes art….!?

        As for the patent side, “having at least” is NOT ‘objective physical structure,’ as it is an OPEN ENDED clause, and is even worse than the (supposedly) ‘bad’ functional description as there are NO constraints to the supposedly “physical” description.

  2. 4

    I’m noticing a tendency among critics of the Amgen decision to reduce the enablement question to “could you make these antibodies?”. Obviously, the answer to that question is generally yes for any antibody, and therefore a patent for new antibodies is not earned because you showed the public how to synthesize them. SCOTUS seemed to be more concerned with the how-to-use prong, and that Amgen claimed exclusivity over embodiments without giving the public any clue whether they work. If the public is expected to perform those tests itself, then why is this even an invention?

    The biggest weakness of the SCOTUS decision imo is that it simply took as a given that lab bench assays enable a drug invention. The shakiness of that assumption is the underlying reason for the skepticism of genus claims.

    Second, the DOE has been applied recently to great effect, in the Viiv v. Gilead case, where the structural similarity was possibly even more attenuated than it would be in the antibody context. I agree that CAFC precedent on DOE can be wobbly, so patentees aren’t likely to be enthusiastic about resting their case on it, but it is arguably better structured to answer the essential question posed by first-in-class vs. follow-on drug patent suits: did the follower free-ride on the pioneer’s invention, and would its commercialization rob the pioneer of exclusivity over its contribution to the art? Where the mechanism of action is exactly the same (e.g., another TNF-a blocker), that might be an easy question.

    In other words, the inquiry can be limited to the follow-on molecule itself and its origin, rather than having to fully map out the abstract boundaries of the invention for Section 112 purposes.

  3. 3

    ” For example, it is uncontroversial that a patent claim broadly directed towards pharmaceutical compositions comprising a specified active ingredient is not invalid for lack of enablement simply because the claim encompasses improved formulations of that active ingredient (e.g., timed release or combination products) that are not literally enabled by the original patent. In many cases, such formulations are sufficiently innovative to justify their own patent, sometimes referred to as an “improvement patent.” I do not think that anything in Amgen should be interpreted as altering this fundamental understanding of the enablement requirement.”

    Really? How does that position even make sense after Amgen? The Supreme Court was concerned that one could not make an antibody that bound to 14 out of 15 residues of the sweet spot without undue experimentation (something no one in this field would set out to do), yet you think they would allow the claims to encompass some hypothetical improvement antibody? What would that improvement be? Conservative amino acid substitutions in the hypervariable regions? The Supremes did not understand the concept. Nor did some of the attorneys.

    One of the fallacious arguments raised, including by the SG, was “one could always rely upon the doctrine of equivalents.” Except, the DOE has never been applied to antibody claims, and the framework usually used is “insubtantial change” not “substantially the same function, way, and result.”

    1. 3.1

      I wonder if you and Prof. Holman are talking past each other. Your comment seems focused on antibodies, but the portion of Prof. Holman’s post that you cite does not appear so limited to my eye.

      Imagine that I file an application in 2019 that includes claim 1: “an antibody having a heavy chain according to SEQ ID NO:1 and a light chain according to SEQ ID NO:2.” This claim is entirely structural, with no functional recitations at all. It should be unproblematic under Amgen.

      Imagine that the application also has a claim 2: “a composition comprising the antibody of claim 1 and a pharmaceutically acceptable excipient.” This is a genus claim, because it reads on many different formulations. It reads on formulations of the antibody with a stabilizer, the antibody with a tonicity modifer, the antibody with a lyoprotectant, etc.

      Imagine in 2021, someone invents a totally new excipient that makes antibodies more shelf-stable at room temperature. Obviously the 2019 application did not enable compositions that include an excipient that was only invented in 2021. Does Amgen mean, then, that claim 2 fails for want of enablement.

      I take Prof. Holman’s answer to be “no” (although I hope that he will correct me if I am wrong in that). Certainly I do not read Amgen to answer this question. I think that if one is trying to extrapolate from the facts of Amgen to the very different facts of my hypo above, then one is laying oneself out for embarrassment, because the factual circumstances are too different to support the reasoning by analogy.

      There are, in other words, plenty of generic claims that are not obviously foreclosed by Amgen. Anyone reading Amgen as a death blow to all genus claims is over-reading it.

      I will even go so far as to say that anyone who tries to apply Amgen to computer implemented inventions is over-straining the analogy. The Amgen holding is much narrower than some had feared, and even than some presently believe.

  4. 1

    These words:

    clearly not literally true – rarely if ever does a patent enable every conceivable embodiment that would fall within the scope of a patent claim, a task that would be impossible in most (if not all) cases, given the open-ended nature of most patent claims.

    misplay the ‘open-ended nature.’

    Further, taking the plain meaning, one would hazard a guess that you are against MOST ALL patents.

    Would that be accurate?

Comments are closed.