Will Merck’s subcutaneous Keytruda infringe Halozyme's patents?
Zachary Silbersher
Merck is planning to sell a subcutaneous version of its blockbuster Keytruda that can be injected rather than administered intravenously. Yet, standing in the way of that is a growing portfolio of patents recently acquired by Halozyme Therapeutics. In 2024, Halozyme was awarded numerous patents covering modified PH20 hyaluronidase polypeptides. Merck immediately went to the Patent Office and filed petitions for the patents to be cancelled. What is the likelihood Merck’s patent challenges will succeed? Or Merck will be compelled to sign a license for Halozyme’s patents.
Background
Beginning in November 2024, Merck started filing petitions for Post-Grant review (PGR) of patents recently issued belonging to Halozyme. The patents cover modified PH20 hyaluronidase polypeptides. A listing of the patents is below.
PGR 1: U.S. Patent No. 11,952,600, which issued in April 2024.
PGR 2: U.S. Patent No. 12,018,298, which issued in June 2024.
PGR 3: U.S. Patent No. 12,152,262, which issued in November 2024.
PGR 4: U.S. Patent No. 12,123,035, which issued in October 2024.
PGR 5: U.S. Patent No. 12,110,520, which issued in October 2024.
PGR 6: U.S. Patent No. 12,054,758, which issued in August 2024.
PGR 7: U.S. Patent No. 12,060,590, which issued in August 2024.
The seven patents challenged by Merck each belong to the same patent family. Each patent is also directed to the same general field, namely, structurally-altered forms of the human PH20 hyaluronidase enzyme that retain enzymatic activity.
What are Post-Grant Review proceedings, or PGR’s? And why did Merck file them?
Companies seeking freedom to operate around an existing do not have to wait to be sued by the patent owner. They can go to the Patent Office and ask for the patent to be cancelled. That is what Merck has done in this instance. The most popular type of proceeding for doing so is called inter partes review. Merck has filed a more specialized proceeding known as Post-Grant review, or PGR.
A PGR can only be filed within nine months after a patent issues. That is why Merck’s PGR petitions were filed so soon after the patents issued in mid- and late-2024. Therefore, if Merck wanted to pursue a PGR against Halozyme’s patents, there was a short deadline to do so.
On the other hand, PGRs are generally a stronger vehicle for invalidating a patent. In an inter partes review, the challenger is essentially limited to arguing the patent should be cancelled because the invention was already taught within the prior scientific literature. Yet, in a PGR, there are more arguments at the challenger’s disposal. In particular, a PGR allows a challenger to argue the patent should be cancelled for lacking written description or not being enabled.
As discussed below, Merck’s primary argument for why Halozyme’s patents should be cancelled (or invalidated) is because they lack written description and are not enabled. Merck essentially argues that Halozyme has patented literally billions of different versions of mutated PH20 polypeptides without doing any testing or providing any guidance to determine which versions will be useful. Had Merck waited past the nine-month deadline, it would have been precluded from making these written description and enablement arguments at this time—at least, not until it was affirmatively sued in court for patent infringement by Halozyme.
Accordingly, by filing the PGR’s, Merck is admittedly tipping its hand that Halozyme’s patents may cover Merck’s enzyme. From Merck’s perspective, however, that’s not likely a grand admission given that Halozyme’s patents potentially cover billions of different PH20 modified enzymes. To fend off the risk of a future lawsuit from Halozyme, Merck opted for a pre-emptive attack. It went directly to the Patent Office and petitioned for the patents to be cancelled.
Halozyme cannot likely sue Merck yet. If Merck has not yet launched its subcutaneous Keytruda product, then Halozyme currently lacks standing to bring suit. Yet, even if Halozyme does sue, the case is likely to be stayed pending the outcome of the PGR’s.
Merck argues that Halozyme’s patents lack written description and are not enabled.
Merck’s primary argument for why Halozyme’s patents should be cancelled is that they cover a staggering number of modified PH20 polypeptides. According to Merck, the patented PH20 mutants can includes sequences of more than 400 amino acids. Yet, the patents do not claim specific amino acid sequences, but rather something much broader.
For instance, the first independent claim of the ‘600 patent (the patent challenged in the first PGR) is directed to a modified PH20 polypeptide that has one amino acid replacement at position 320 and may contain additional modifications elsewhere as long as the resulting PH20 polypeptide retains at least a certain percentage (e.g., 95%) of sequence identity to one of 36 reference PH20 proteins. The reference amino acid sequences for the reference PH20 proteins are identified in the patents themselves.
According to Merck, the ‘600 patent allows for so many possible substitutions or changes to the reference PH20 proteins that Halozyme’s patents technically cover billions of different enzymes, each with a different amino acid sequence. Merck estimates that the first claim alone of the ‘600 patent covers approximately 1.35 x 1066 different distinct PH20 polypeptide mutants.
Merck argues this necessarily contravenes the written description and enablement requirements of patent law. In order to satisfy the written description requirement, Halozyme would have had to identify a representative sample of PH20 mutants that not only satisfied the structural requirements of the patent, but also test those sample mutants to see if they were efficacious. In this case, according to Merck, that would mean testing for improved hyaluronidase activity compared to the human PH20 polypeptide (otherwise known as “active” mutants). This would be opposed to mutants that satisfy the claimed structural criteria, but do not exhibit hyaluronidase activity (otherwise known as “inactive” mutants.)
The patents disclose testing for about 5,900 PH20 polypeptide mutants. Among those tested, only about 13% were found to have improved enzymatic activity compared to the unmodified PH20 polypeptide. More importantly, among the 5,900 PH20 mutants tested, they only contained a single amino acid change. In other words, none of disclosed examples in the patents contain multiple amino acid changes. PH20 mutants with multiple amino acid changes, ranging generally from 2 to upwards of at least 19 substitutions, for example, are clearly encompassed by the patented claims. These versions with multiple amino acid changes drives the number of versions covered by the patents into the billions. For this reason, Merck is essentially arguing that the actual testing disclosed in the patents (i.e., the examples) cover only a tiny fraction of what Halozyme claims to be covered by its patents.
By the same token, Merck argues that the patents also fail enablement because it would require “undue experimentation” to identify and test all the billions of mutants covered by the patent. (Merck suggests that if anyone were to actually do that, the resulting proteins would be heavier than the weight of the earth.)
Halozyme, by contrast, argues that Merck has this all wrong. Merck assumes that the patents require that the resulting mutants exhibit hyaluronidase activity. Yet, according to Halozyme, the patents do not require that at all. Indeed, the patents explain that even inactive mutants, i.e., those that do not possess hyaluronidase activity, have therapeutic uses, such as for contraceptives. Accordingly, because the patented PH20 mutants do not purport to claim any utility or purpose, then there is no need to test them for that function (i.e., hyaluronidase activity.) And then, according to Halozyme, because there’s no need to test them for hyaluronidase activity, Merck’s arguments all collapse.
Halozyme emphasizes that the patents are very clear about the structural criteria required to fall within the patented claims—i.e., a specific change at certain positions and then other changes that nevertheless retain upwards of 95% identity with reference polypeptides. Scientists in this field, according to Halozyme, can clearly envision the scope of mutants that would fall within that patented structural criteria. And the science required to identify those proteins is fairly routine and straightforward. Thus, according to Halozyme, no undue experimentation is involved.
A key issue in this dispute is whether or not the patents require improved enzymatic or hyaluronidase activity. Merck argues they do; Halozyme argues they don’t. Yet, resolution of that issue—which is really a claim construction dispute—could impact how the PTAB rules on the written description and enablement issues. On the one hand, the patents clearly contemplate modifying the PH20 polypeptides for improved enzymatic activity. Indeed, the Abstract for the ‘600 patent describes the disclosed invention as: “Modified PH20 hyaluronidase polypeptides, including modified polypeptides that exhibit increased stability and/or increased activity, are provided.”
On the other hand, the patent claims themselves do not recite that the modified PH20 polypeptides are for any specific purpose. The disclosure in the abstract, or elsewhere within the patent’s specification, is not dispositive. Halozyme also emphasizes that inactive PH20 mutants may also have therapeutic efficacy in themselves for contraceptive purposes. Merck contests that.
That said, Halozyme faces risk here. Pharmaceutical patents do not always recite efficacy as an expressly limitation or requirement of the invention, but courts have nevertheless presumed efficacy under certain circumstances. Otherwise, what exactly is the invention? In Halozyme’s telling, it has essentially roped off and cornered the market on an arbitrary selection of billions of different PH20 mutants that serve no essential claimed purpose.
Important to the resolution of this issue is the evolving caselaw regarding patenting antibodies. Merck identifies recent caselaw holding that antibodies that claim a patent based on its function, i.e., any amino acid sequence that binds to a particular antigen at specific epitopes, fails the enablement requirement. The most important case is the Supreme Court’s recent decision in the PCSK9 dispute, which I’ve previously written about previously.
By contrast, Halozyme has identified caselaw suggesting that its formulae for patenting mutant polypeptides, by claiming any substitutions that that preserve a degree of identity with reference polypeptides, is focused on structure, not function. Halozyme further points to prior cases upholding the written description or enablement of patents based on the approach taken by Halozyme. Two of those cases, however, are PTAB decisions from nearly 20 years ago, which is long before the Supreme Court’s Amgen decision. A more recent decision cited by Halozyme, Boehringer v. Kansas State, PGR2022-00021 (P.T.A.B.) is not as authoritative as a Federal Circuit decision, and therefore, there’s room for Merck to attempt to distinguish that case.
Merck also argues the patents are invalid as obvious
Merck also argues the patents should be cancelled as obvious over the prior art. The patents contemplate multiple substitutions in the human PH20 protein, but only technically require a single substitution at position certain positions. For instance, the ‘600 patent requires the D320K as a single substitution in the PH201-447. (The other patents require a single substitution at different positions within the same region, e.g., 307, 317, etc.) Merck identifies prior art teaching that single substitutions in certain non-essential regions may not substantially compromise biological activity, such as hyaluronidase activity. Merck identifies additional prior art teaching that a scientist would have been motivated to make a single amino acid substitution at position 320 by studying which amino acid substitutions have been tolerated among a finite sample of known homologous, stable and active naturally occurring hyaluronidase enzymes.
Halozyme, for its part, disputes Merck’s prior art analysis. The dispute is highly factual and depends to a large degree on differing expert opinions regarding what was generally known about PH20 single amino acid substitutions in non-essential regions in 2011, and more importantly, how routine it would have been to study such substitutions.
Yet, one of Halozyme’s arguments is interesting compared with its arguments against Merck’s written description and enablement challenges. Halozyme argues that Merck fails to prove that a scientist in 2011 would have been motivated to make a D320K amino acid substitution in PH20. For instance, Halozyme argues, “Merck fails to establish that the ’429 Patent combined with
Chao provides the requisite reasonable expectation of success that a D320K substitution in PH20 would not only be tolerated, but would result in a protein that exhibits at least comparable hyaluronidase activity to unmodified PH201-447.” (PGR2025-00003, Patent Owner Preliminary Response at 78)(Mar. 7, 2025).
Halozyme’s prior art arguments are slightly in tension with its written description and enablement arguments. For written description and enablement, Halozyme argues the patents do not require improved enzymatic activity. Yet, to rebut Merck’s obviousness argument, Halozyme argues that Merck failed to show that the prior art would have motivated a scientist to modify a PH20 polypeptide to achieve improved enzymatic activity. If the patented polypeptides require no therapeutic purpose, then how could a scientist ever be motivated from the prior art to arrive at the claimed invention? But in that case, what exactly is the claimed invention other than an arbitrary collection of billions of mutant polypeptides with arbitrary substitutions.
***
Barring an order from the PTAB, Merck will not have the opportunity to respond to Halozyme again before the PTAB determines whether to institute the PGR petitions. Those decisions should start rolling in beginning in early June 2025.