COVID Patents at the Federal Circuit (Moderna Loses)

by Dennis Crouch

Moderna Therapeutics v. Arbutus BioPharma (Fed. Cir. 2021) (20-1184 Decision) (20-2329 Decision)

Arbutus owns several patents related to a lipid-based delivery system for nucleic-acid-based treatments.  U.S. Patent 8,058,069 and 9,364,435 are the two at issue here. The claims are directed to the creation of a “acid-lipid particle” that includes some RNA along with lipids to stabilize the sequence, including particular percentages of cationic lipids, non-cationic lipids, and conjugated lipids.

COVID: If you have been following the news, you know that the COVID-19 vaccines distributed by Moderna and Pfizer deliver mRNA to the cells using lipid nanoparticles.  Although Arbutus has apparently not threatened Moderna with an infringement action, it has apparently requested a broader license agreement and refused to issue a covenant-not-to-sue.  Moderna recognizes “a substantial risk” that Arbutus will assert the patents in an infringement lawsuit “targeting Moderna’s COVID-19 vaccine.”

Pre-COVID: Moderna had been working on mRNA delivery systems and had previously licensed the patents for research work using a milestone payment scheme.  Moderna challenged the patents via IPR in 2018 and 2019 so that it could use the inventions without royalty payments or other reporting.  IPR2018-00739;  IPR2019-00554. This was all before COVID, but the risk and value have ramped-up exponentially.  The PTAB granted the IPR petitions but substantially sided with the patent holder:

  • ‘069 patent — all claims confirmed (not proven invalid); and
  • ‘435 patent — claims  7–8, 10–11, 13, and 16–20 are confirmed; Claims 1–6, 9, 12, and 14–15 invalid as anticipated.

The parties appealed their respective losses, and the Federal Circuit has now affirmed on all grounds — although a portion of its decision is a procedural shortcut.  Because the patents were left largely intact, we have a good likelihood of further litigation between these parties over the patents (or else a new license in the background). 

The basic idea here is that this area of innovation is extremely tricky and unpredictable.  Thus, although the “general conditions for a nucleic acid-lipid particle were disclosed in the prior art,” the particular claimed lipid profile would not “have been achievable through routine optimization.”

Standing to Appeal: It is pretty clear that at this point the post-COVID relationship between the parties is such that Moderna has standing in court to challenge these patents.  However, as I mentioned, the IPRs were filed pre-COVID and the 2018 IPR was appealed in November 2019 — well before Moderna had an inkling that it would be making a COVID-19 vaccine.  The problem for Moderna here is that the appellate court can only hear cases involving an “actual case or controversy” between the parties.  And, that actual case-or-controversy must persist throughout the entire appeal (from beginning to end).  What Moderna needs to show is that the patent served as a concrete and immediate threat as of that November 2019 date.   I mentioned that there was a license agreement, but the court found that any payments due under that agreement too speculative to generate standing since – at the time – Moderna was not actually doing the planned research.  Further, the patentee (at the time named Protiva) “emphasized that it had never initiated a patent infringement action or directly accused Moderna of infringing its patents.”

Things had changed dramatically by the time Moderna filed its opening brief in May 2020.  But, as I mentioned, courts require continuous standing throughout a lawsuit, and it just wasn’t there in November of 2019.  The result then is that Moderna was not permitted to appeal its loss in the ‘435 case.


30 thoughts on “COVID Patents at the Federal Circuit (Moderna Loses)

  1. 4

    Off topic, but today in In re Alfresco Software the CAFC denied a mandamus to the WD Tex because in the time since the mandamus petition was filed, Judge Albright transferred the case to CD Calif, thus mooting the petition. Is Judge Albright, perhaps, beginning to weary of the barage of writs?

    1. 4.1

      Do simians bring fire-hosed in a cage learn not only to notclimb the ladder in the middle of the cage, but also to pounce on any new simians that do?

  2. 3

    How is

    would not “have been achievable through routine optimization.”

    related to:

    without undue experimentation?

    It would seem that the first is far different (and much lower) than the second.

    1. 3.1

      The In re wands factors define undue experimentaion, and link it to predictability and routine experimentation (plus other stuff).

      See Idenix v Gilead as example of when “this field is highly unpredictable” came back to screw them because of undue experimentation / enablement issues.

    2. 3.2

      Also, see Wyeth v Abbott which basically uses “routine” and “undue” interchangeably:

      “Undue experimentation is a matter of degree. Even “a considerable amount of experimentation is permissible,” as long as it is “merely routine” . . . . Yet, routine experimentation is “not without bounds.”.”

      Note Wyeth has been cited for the proposition that EVEN routine experimentation can be undue– suggesting that “undue” is a lower bar than routine, not the other way around: “Where, as here, practicing the full scope of the claims would have required excessive experimentation,
      even if routine, the patent is invalid for lack of enablement” … (Idenix citing wyeth).

      1. 3.2.1

        Thanks 4M — I have to think that any court equating “routine” with “undue” (or even worse, making “routine” more onerous than “undue” has butchered (yet another) patent decision.

      2. 3.2.2

        Wyeth v Abbott… basically uses “routine” and “undue” interchangeably:

        “Undue experimentation is a matter of degree. Even a considerable amount of experimentation is permissible, as long as it is merely routine”

        Either I am misunderstanding you, or you are misunderstanding Wyeth. The quoted section from that precedent treats “undue” and “routine” as antonyms (as they are), not as “interchangeable.”


          yes, of course, it uses “permissible experimentation” (ie, that which is not undue) and “routine experimentation” interchangeably. the point is that “routine optimization” from prior art/obviousness analysis and the “undue experimentation” from enablement analysis are likely looking at very similar elements


            Thanks for the clarification. I definitely agree that “‘routine optimization’ from prior art… analysis and… ‘undue experimentation’ from enablement analysis are… looking at very similar elements.” Indeed, they are flip sides of the same coin.

            The point of §103 is that Congress does not want to grant patents on things that the prior art has already taught us how to do, even if those things are not precisely exemplified in a single piece of prior art. Therefore, that which is patentable must—by definition—not be enabled by the prior art. If a claimed invention is entirely enabled by the prior art, then the invention is—by definition—unpatentably obvious.

            That which requires only “routine” experimentation is obvious. In re Aller, 220 F.2d 454, 456 (C.C.P.A. 1955). By contrast, that which requires “undue” experimentation is—by definition—not enabled, and it cannot be obvious for the skilled artisan to do that which the skilled artisan does not know how to do.

  3. 2

    For a good laugh, check out Moderna’s patent specifications where they describe a million “innovative” RNA molecules, none of which are demonstrably useful except perhaps as probes … oh but wait someone else found a use so let’s use their data to support a specific claim to that molecule! LOL

  4. 1

    >”The basic idea here is that this area of innovation is extremely tricky and unpredictable. Thus, although the “general conditions for a nucleic acid-lipid particle were disclosed in the prior art,” the particular claimed lipid profile would not “have been achievable through routine optimization.””

    Great, as long as they apply this same standard to enablement.

    1. 1.1

      Did you read those claims? They are really quite narrow. Good luck challenging the enablement on such well crafted claims.

      1. 1.1.1

        I did read the claims. Claim 1 of ‘435:

        A nucleic acid-lipid particle comprising:
        (a) a nucleic acid;
        (b) a cationic lipid comprising from 50 mol % to 85 mol % of the total lipid present in the particle;
        (c) a non-cationic lipid comprising from 13 mol % to 49.5 mol % of the total lipid present in the particle; and
        (d) a conjugated lipid that inhibits aggregation of particles comprising from 0.5 mol % to 2 mol % of the total lipid present in the particle.

        Does the invention function with EVERY combination of any nucleic acid and any cationic lipid and any non-cationic lipid? If it doesn’t (and it probably doesnt), then the claim is only valid if one of ordinary skill could separate working from non-working embodiments without using “undue experimentation.” If this is a highly “tricky and unpredictable field,” then you run into problems.


          I expect that there are precise percentages in those ranges where it works especially well (e.g., the particle is particularly temperature stable, or particularly long lived at -20°C), but I doubt that there is much in those claimed ranges where the particle will not form at all. Enablement is not best mode.


            Not (just) about if particles form at all, but if the particles that form across that entire range have utility. It could be the case, but I’m just pointing out there are big potential downsides to arguing around prior art by saying you are in a highly unpredictable field.


              Utility is a low bar to clear. I am sure that narrow ranges of lipids within the broader claim genus function much better than others. I expect that all the particles across the claimed range, however, have at least some functionality. It might require “undue” effort to find the optimum values in that range, but I doubt that it requires any experimentation at all to work the full claim scope.


                Let me correct Greg’s
                Utility is a low bar to clear.

                Traditionally, Utility is a low bar to clear.

                This is where Greg’s Big Pharma versus his computing arts hypocrisy is noticeable.


                Greg — I’m not sure I understand your statement:

                “It might require “undue” effort to find the optimum values in that range, but I doubt that it requires any experimentation at all to work the full claim scope.”

                Is there such a thing as “undue effort” in biochem patent law?

                1. My apologies. I should have said “undue experimentation,” rather than “undue effort.”


                Utility and enablement requirements are two different things.

                You have to enable the full scope of the claims. Enablement includes how to make and use.

                If the literal scope of the claims includes non-functional embodiments (and it almost always does, because there are generally some combinations, e.g. certain lipids that for whatever reason, wont work), then OOSA has to be able to distinguish functional from non-functional embodiments without undue experimentation. This is about enablement, and particularly the “how to use” aspect, not about utility. If you argue it is a “highly unpredictable field” to get around prior art, then you are taking a risk on enablement.

                1. If you argue it is a “highly unpredictable field” to get around prior art, then you are taking a risk on enablement.

                  I agree that one is taking a risk in making the “highly unpredictable” argument. I feel good about the patentee’s chances in this case, however. They did not get greedy, claiming the moon and stars. This is a defensible claim scope.

                2. Greg’s “this case” comment leaves me thinking that somehow his “touchy-feely” application of law is a wee bit suspect.


          To be fair, I suppose that “nucleic acid”—in its BRI—reads on whole chromosomes. The particle would not form with cargo so large. The written description gives the reader plenty of guidance, however, about cargo sizes, so I am not actually worried about practitioners spending “undue” time trying to form the particle with a 3000 kbp cargo. One must bring a certain measure of common sense to both the claim construction and enablement analyses.


            Sure. I can believe that there is a novelty problem. I am more skeptical about an enablement challenge here.


          Depends on what you mean by “tricky.” The machines necessary to do the work are complicated and expensive. You could not do this compounding in your garage. If you have the machines, however, I am sanguine that you could make (at least minimally functional) particles across all of those claimed ranges.

          The only part of the claim that seems arguably scope-of-enablement problematic to me is “nucleic acid” (see above). The lipid ranges, however, seem wholly defensible to my eye.


            I was thinking that you needed bespoke microfluidic mixing chambers that aren’t off the shelf. If the parts are expensive but available, that’s clearly not the case.

Comments are closed.