Will Amgen’s PCSK9 loss read-through to patents covering other biologics drugs?

Zachary Silbersher

The long-running patent dispute between Amgen ($AMGN) and Regeneron($REGN) and Sanofi over their competing PCSK9-inhibitors (Repatha® and Praluent®) has reached another milestone.  Despite prior victories, the tides have now shifted against Amgen.  On August 28, the district court vacated the jury’s earlier verdict in Amgen’s favor, and invalidated Amgen’s patents.  With that decision, Amgen’s request for an injunction against Regeneron’s Praluent® as well as financial damages was nullified.  The case will be appealed by Amgen. 

Yet, the case also represents another milestone in the changing landscape for patents covering biologic drugs.  The Amgen decision is at least the second district court decision this year that has invalidated biologic antibody patents under the doctrine of enablement.  The earlier decision related to MorphoSys patents asserted against Janssen related to Darzalex®. 

 These cases are sure to read-through to other patents protecting the exclusivity of other biologic drugs.  The takeaway is clear:  as biologic drugs take up a larger share of the pharmaceutical medications in the U.S., courts are making it harder for drug companies to use overly-broad patents to corner the market on a particular inhibitor.

Background of the PCSK9 dispute

In March 2016, Amgen prevailed at trial against Regeneron before a jury that held Amgen’s patents were not invalid.  Then, in January 2017, the district court ordered that Praluent® should be enjoined from the market.  Before that could happen, however, in October 2017, the Federal Circuit vacated that order and remanded the case for another trial on invalidity.   

Over the next two years, Amgen and Regeneron went back to the district court, collected more evidence, and then had another trial.  In February 2019, Amgen prevailed in front of the jury a second time, and then promptly renewed its bid for a permanent injunction against Praluent.  The court held a multi-day hearing on Amgen’s request in June 2019. 

But then, in yet another twist in the case, on August 28, 2019, the district court reversed the jury’s verdict that Amgen’s patents were not invalid.  In effect, the district court invalidated Amgen’s patents.  That, in turn, mooted Amgen’s request for an injunction against Praluent® as well as any request for past damages or a future royalty.

The district court’s latest decision was surprising to the extent that courts do not typically upset a jury’s verdict.  But, it was less surprising given the context of this case.  The prevalence of biologic drugs is increasing, and courts are increasingly taking a closer look at patents purporting, much as Amgen’s patents do, to corner the market on a particular inhibitor.  We previously wrote about a similar case earlier this year between Morphosys and Janssen relating to Janssen’s drug Darzalex®.  

What is enablement for antibody patents?

 The district court held that Amgen’s patents were not sufficiently enabled, and therefore, invalid.  For the claims of a patent to be sufficiently enabled, a person of skill must be able to recreate the invention without undue experimentation.  More importantly, the enablement doctrine requires that the full scope of the claims are enabled.  If a patent can theoretically cover millions of different permutations of a particular antibody, all of those permutations should be enabled.  It is not enough for only a single example permutation to be enabled.  This has become a critical consideration for biologics patents.

Among patent attorneys, enablement is one of those doctrines that always lurks in the background, but is not necessarily litigated that frequently.  Yet that may be changing, especially for pharmaceuticals.  In the context of biologic drugs, and patents covering antibodies and inhibitors, the enablement doctrine is increasingly defining the landscape of which patents will survive patent challenges and which drugs will see efforts to box out competitors fail.  This is evident from the Amgen decision and the MorphoSys decision, two decisions this year alone that have both knocked out antibody patents under the enablement doctrine.

The hallmark of showing that a patent lacks enablement is that it requires undue experimentation.  In the context of typical antibody patents, this raises a unique challenge.  Amgen’s patent covers antibodies that bind to PCSK9.  Yet, the patent claims the antibody based more on what it does, rather than what it is.  Specifically, Amgen’s patents do not claim a protein with a specific amino-acid sequence or particular three-dimensional structure.  Rather, the claims are directed to any antibody that binds to certain residues on PCSK9 and blocks binding of PCSK9 to LDLR. 

By claiming the PCSK9-inhibitors in this way, namely functionally rather than structurally, there are potentially millions of antibodies that would theoretically fall within the scope of Amgen’s patents.  That would necessarily include scores of antibodies that look nothing like Amgen’s own protein within its drug, Repatha®.  (Indeed, this was one of Regeneron’s arguments at the trial.)

And therein lies the way that drug companies have attempted to corner the market on a particular class of inhibitors.  Drug companies have procured patents covering large swathes of proteins by describing them in functional terms rather than structural terms.  That makes it much easier to peg a competitor with infringing those patents.  That strategy seemed to be working for Amgen.  But courts are now evidently pushing back.  If the net cast by the patent is too broad, then the patent faces the risk of being invalidated for not being adequately enabled.

            How will the Court’s decision read-through to other biologics patents?

There were many reasons the district court found that Amgen’s patents were not sufficiently enabled.  These reasons are likely to read-through to existing biologics patents as well as stand an a cautionary tale for drug companies procuring biologics patents in the future. 

The Court relied extensively on a similar case from earlier this year between MorphoSys and Janssen:  MorphoSys AG v. Janssen Biotech, Inc., No. 16-221 (LPS) (Dkt. 471) (Jan. 25, 2019).  In the MorphoSys case, the patents claimed a CD38-inhibitor antibody manner similar to Amgen’s PCSK9 patents, namely, by their function rather than the structure.  The court invalidated those patents for lacking enablement for reasons similar to those adopted by the Court in Amgen’s dispute with Regeneron. 

First, the Court found that Amgen’s patents were very broad.  They encompassed millions of potential proteins.  That in itself was a reason to question whether the patents were sufficiently enabled.  Even assuming that scientists would only perform “intelligent substitutions” of example proteins, that would still yield an unwieldy number of proteins that would need to be tested to determine if they bound to PCSK9.   

For existing drug patents, red flags are sure to be raised when an antibody patents covers seemingly millions of potential permutations.  This feature—covering a broad swathe of proteins—was deemed a benefit in the past.  Indeed, that was likely the basis for claiming antibody patents through their function (i.e., where they bind on an antigen) rather than their structure.  Yet, now, functional claiming for antibody patents may soon itself be enough to doom biologics patents.

The Court also found that the state of the art when Amgen’s patents were filed—in 2008—remained unpredictable.  Experts for both sides testified that even “conservative” substitutions (presumably covering variants made within the framework region, rather than the CDR region) would also still need to be tested to determine if they block binding of PCSK9 to LDLR. 

This factor may jeopardize more older patents rather than newer patents.  Amgen’s patents were filed in 2008, and the state-of-the-art has undoubtedly matured since then.  What was unpredictable in the 2000’s may be less so today.  The predictability of whether certain substitutions for known antibodies can yield predictable functions may theoretically increase over time.  

Finally, the Court also found no evidence that substituting amino acids within an antibody has predictable affects on its function.  That means that testing all permutations covered by the claims would still be required.  Put another way, the Court essentially found that Amgen’s patents covered millions of potential antibodies.  Yet, for a scientist to determine if any of those antibodies actually infringed the claims—and blocked PCSK9 from binding to LDLR—that would take the same trial-and-error required by Amgen to isolate Repatha®.  In that case, the Court essentially held that Amgen had not invented the full scope of the antibodies covered by its patents, and therefore, the patents were no adequately enabled.

This is really the crux of the issue.  Biologic patents such as those asserted by Amgen and MorphoSys cover millions of antibodies that those companies neither invented, enabled or reduced-to-practice.  Indeed, both Janssen and Regeneron argued, in defense to those patent litigations, that their own proteins did not resemble the example proteins disclosed in Amgen’s or MorphoSys’ respective patents.  Given that, courts in both cases pushed back on the drug company’s attempt to corner the market on a whole class of inhibitor.  This same reasoning could doom any existing drugs whose exclusivity are equally covered by broadly claimed functional patents that cover millions of antibodies those companies did not actually invent.

Going forward, these decisions, assuming they are blessed by the Federal Circuit, are likely to alter the way that drug companies obtain patents covering biologic drugs.  They are likely to focus on claiming less—meaning, defining their patents based more upon the structure of the protein, and less on its function. 

The result will be that the patents are not as broad, thus mitigating the likelihood will capture any competing antibody.  And yet, this may be necessary to avoid years of litigation only to have your patents sunk under an enablement challenge.  Narrower patents are typically easier to design-around.  That is likely to incentivize competition, rather than forestalling it through a cornered market.  These decisions may end up boosting biosimilar competition.

            The Federal Circuit will weigh in shortly 

MorphoSys and Janssen settled their dispute before the case reached appeal at the Federal Circuit.  In the case of Amgen versus Regeneron, that is less likely.  The Federal Circuit is very likely to hear this appeal within the next year.  If the Court affirms the district court’s decision, then its decision is likely lay down important principles that could affect how biologic proteins are protected through patents in the future. 

More importantly, however, it could eviscerate the validity of scores of antibody patents that adopted the same approach as Amgen’s patents covering its PCSK9-inhibitor and MorphoSys’s patents covering CD38-inhibitors.  That, in turn, could impact the exclusivity of several biologics blockbuster drugs, and help usher in biosimilars more quickly.   

A lot hangs in the balance, not just for Amgen and Regeneron, but for all companies wagering years of exclusivity for high-priced biologic drugs with antibody patents.