Last week the District Court issued an important memorandum and order in Amgen v. Roche, a patent infringement litigation of great importance to biotechnology. The 150 page memorandum seeks to address and resolve all matters pending before the court, putting the case in condition for full appellate review. When reading the decision, I came away with the impression that Judge Young has been overwhelmed by the many complex issues brought up in the case. Early in the opinion, he acknowledges that due to the massive volume of post-trial briefing he was unable to address every motion. Nonetheless, he ends by noting that the order should be sufficiently "final" to allow for an immediate appeal, welcome appellate guidance, and expresses his hope that in this case he will be able “to avoid the process of appeal-remand-appeal-remand that unfortunately has characterized the related Amgen v. TKT/HMR litigation.”
I found a number of Judge Young's conclusions surprising, and will summarize a few of them here.
The Patents and Allegedly Infringing Product
Amgen's asserted patents claim methods and reagents for expressing and purifying recombinant erythropoietin (EPO), one of the earliest and still most successful biotechnology products. The patents come out of Amgen's groundbreaking work at the dawn of the biotechnology era which resulted in the successful cloning and recombinant expression of human EPO. Amgen sells two recombinant EPO biologic drugs (referred to as “erythropoiesis stimulating agents,” or ESAs), EPOGEN and Aranest. It also licenses Johnson & Johnson to sell its ESA, Procrit.
Roche’s allegedly infringing product is a pegylated-version of recombinant EPO, which goes under the trade name MIRCERA. Pegylation refers to the covalent attachment of PEG (a non-biological polymer) to the protein, and is a common technique used to modify the functional characteristics of a protein. In the case of MIRCERA, pegylation reportedly increases the half-life of the EPO protein in the body, allowing the drug to be administered at less frequent intervals. MIRCERA is produced by Roche in Europe, but they have sought FDA approval to market the drug in the US, sparking the infringement suit by Amgen.
Obviousness Double Patenting
Roche apparently pinned much of his hope of escaping liability on its argument that Amgen's patents are invalid for violating the judge-made prohibition against obviousness-type double patenting (ODP). Unfortunately for Roche, the court did not find this argument very persuasive. There appears to have been some confusion as to whether the determination of ODP is for the judge or jury. Judge Young noted that he was "surprised (and concerned) that Roche focused heavily in its opening to the jury on double patenting issues, as though it were some sort of equitable defense.” Ultimately, he decided it was a decision for the judge, not jury, and he rejected all of Amgen's ODP arguments.
At times, Judge Young appeared to confuse obviousness double patenting with statutory double patenting. For example, he states that “at its essence the inquiry ODP asks: “Is the same invention being claimed twice?”
Based on the facts as I understand them, Roche seemed to have some pretty strong arguments for ODP. For example, Amgen’s earliest EPO patent (4,703,008; expired now but successfully asserted in Amgen v. Chugai and Amgen v. TKT/HMR) includes a claim 23 which recites a “[eucaryotic] host cell transformed or transfected with a DNA sequence [encoding erythropoietin] in a manner allowing the host cell to express [the EPO protein].” The judge acknowledges this claim in his memorandum , but for some reason chooses to ignore the more relevant Claim 24, which depends from Claim 23 and specifies that “the host cell is capable of glycosylating [the EPO protein].”
The only practical utility for the cell of Claim 24 is to express the glycosylated EPO and to purify it. Methods for purifying glycosylated EPO were available at the time, as evidenced by the earlier success of others in purifying glycosylated EPO from the trace amounts present in human urine. Nonetheless, the judge found that the claim to the cell did not render obvious Amgen’s later issued 5,441,868 patent claims reciting a process for the production of glycosylated [EPO protein] comprising the steps of growing up the cell claimed in the ‘008 patent and isolating the glycosylated [EPO].
The judge held that “simply having a starting material [i.e., the cell capable of expressing glycosylated EPO] and knowing that, in theory, it can be used to create proteins is not the equivalent of having an actual process of successfully does so.” But this implies that the earlier patent was not enabled, since he is essentially arguing that there was no reasonable expectation of successfully expressing and isolating glycosylated EPO from the claimed cells, which means the earlier patent did not adequately teach one of skill in the art to use the invention. He does not provide any positive explanation as to why the use of the cells for their intended purpose would not have been obvious given the disclosure of the cells.
He was not persuaded by some seemingly relevant statements made by Amgen in an interference proceeding before the PTO involving some of the asserted patents. For example, in the past Amgen has argued that “clearly, the whole purpose and intent of the purified and isolated DNA sequence encoding human EPO (and cells transfected therewith) at issue in the litigation was to express in vivo biologically active human EPO.” In its decision, the Interference Board ruled that Amgen's work “relating to expression of the EPO gene in mammalian host cells and isolation of the resulting product involved [nothing] more than the exercise of ordinary skill by practitioners in the field.”
The judge concludes his ODP analysis by finding “that the ‘868 claims are not anticipated by the ‘008 patent and [] that the ‘868 claims are different from the claims of the ‘008 patent, again seemingly confusing anticipation with obviousness.
Anticipation
Roche also argued that Amgen's claims directed to pharmaceutical compositions comprising recombinant EPO were anticipated by the earlier purification of the EPO from human urine. The court rejected this argument, finding that a claim limitation requiring that the EPO be purified from mammalian cells grown in culture distinguished over the prior art. This determination was based on evidence that there are structural differences between recombinant EPO and EPO purified from human urine. In particular, it had been shown that the glycosylation of recombinant EPO differs from urinary EPO, so that the limitation to EPO purified from cell culture reflected a real structural difference relative to the prior art, although the structural difference was not determined with any specificity.
This ruling seems correct, based on the apparent structural differences between the claimed and prior art proteins, assuming that the claims would not cover a recombinant EPO that is structurally identical to the prior art EPO derived from urine. There was one interesting statement from Judge Young that did stand out. In addressing the anticipation argument, he noted that "it is significant that the source [i.e., from recombinant cell culture] is what enables mass production and commercial viability.” In effect, he appears to be suggesting that recombinant EPO protein would be patentable over the prior art even if the structure of recombinant and prior art EPO is identical. This clearly cannot be correct. The invention of a new process for producing a prior art product can be patentable, but it does not render product produced by the new process patentable unless there is some structural difference from the prior art, regardless of the commercial significance of the new process.
Indefiniteness
Roche also argued that certain claims were indefinite because "the breadth of the claim term ‘human erythropoietin’ makes it impossible to determine what is and what is not within the claim." Roche present expert testimony which opined that “the patent specification contemplates dozens of ‘polypeptides of the invention’ that fall within the scope of ‘human erythropoietin,’ including mutants, analog and allelic variants." However, the jury sided with Amgen, apparently crediting Amgen's expert, who testified that one of skill in the art would have understood the meaning of the term, and the court found this was enough to support the jury’s verdict.
Infringement
Roche argued that the EPO claimed in the Amgen patents does not encompass pegylated EPO, and thus MIRCERA does not contain the claimed EPO. The court found that these assertions were inconsistent with Roche's own internal communications, and communications with FDA in connection with seeking regulatory approval for MIRCERA. Prior to litigation, Roche referred to the active ingredient in MIRCERA as “peg-EPO” in internal communications. In its biologic license application (BLA), submitted to FDA to gain approval to market a product in the US, Roche stated that pegylation did not alter the amino acid sequence or glycosylation of EPO.
While acknowledging structural differences between EPO and pegylated EPO, the court found these differences to be irrelevant because they did not affect the underlying amino acid sequence of the protein. According to the court, the only structural limitations set forth in Amgen's patent relate to amino acid sequence, and therefore attachment of other chemical groups to the amino acid sequence (i.e., post-translational modifications) are irrelevant to the question of infringement. The problem with this is that structural modifications to the amino acid chain were critical to the patentability of Amgen's recombinant EPO. The amino acid sequence of recombinant EPO was not novel, but inherently existed in the prior art EPO purified from human urine. In order to overcome this prior art, Amgen limited its EPO claims to EPO derived from cultured cells (i.e, recombinant EPO), which it argued had a different chemical structure than the prior art EPO. But these patentable differences did not reside in the amino acid sequence, but rather in the carbohydrate groups attached to the string of amino acids, i.e., the glycosylation pattern. I think the judge is confused when he states that the attachment of carbohydrate groups to the amino acid chain are irrelevant to infringement, when it is these very modifications that distinguish the patented recombinant EPO over the prior art.
Judge Young engages in some circular logic in his infringement analysis. He concedes that the source limitation in the claims (derived from cell culture) confers patentability because they result in a structural difference relative to the prior art. If recombinant EPO derived from cell culture had identical structure to the prior art EPO, then recombinant EPO would not be a novel chemical and would not be patentable as a product. However, he concludes that “the structural differences are merely evidence that the source limits the claim; they are not themselves limitations.” He concludes that once a determination has been made that a source limitation confers patentability, the issue of structure becomes immaterial. But he seems to forget that since this is a claim directed to the protein itself, not the process used to produce it, the source limitation is relevant only to the extent it serves as a proxy for a definition by structure.
Injunctive Relief
Judge Young thought seriously about denying Amgen permanent injunctive relief, based on its initial view that the introduction of competition in the EPO market would benefit the public. Under the 2006 Supreme Court decision in eBay v. Mercexchange, the public interest is one of the four factors to be considered in weighing whether to grant a permanent injunction after infringement of a valid patent has been found. Nonetheless, he ultimately decided that entering a permanent injunction is the proper remedy under eBay.
The judge surveyed post-eBay case law and found it very rare for a court not to grant injunctive relief in cases where not granting the injunction would result in competition between the plaintiff and infringer.
He then found that the first three eBay factors strongly favored permanent injunctive relief for Amgen. In particular, he found that the entry of Roche in the EPO market would cause substantial market erosion and loss of revenue for Amgen, and could encourage other would-be infringers to attempt to gain access to the market. This would weaken Amgen's patents, which the court characterized as "the foundation of Amgen's business." He opined that the Amgen stock price would fall, along with its ability to attract investment and to engage in research and development.
Regarding the fourth eBay factor, Judge Young found that the public interest would not be disturbed by permanent injunction. He determined that Amgen would be able to adequately satisfy the current demand for EPO products, and that Roche had failed to adequately prove that MIRCERA offers significant clinical advantages over Amgen's products.
He also found that declining to enter injunctive relief would harm the public interest by weakening the patent system, which he characterized as critical for incentivizing the highly expensive and risky business of drug development.
Finally, he found that the competition Roche’s product would introduce into the EPO market would probably not result in any savings to the US consumer (or more particularly, to Medicare, which ultimately fits the bill for most EPO sold in the US). To the contrary, he concluded that it was not unlikely that the introduction of competition would actually result in higher prices, a counterintuitive conclusion that I discuss in more detail in an earlier post.
Judge Young concludes by noting that a modified injunction that he had originally considered, which would have in effect amounted to a compulsory license to Roche under terms dictated by the court, would be difficult to enforce and manage. Not only would it be difficult to set and monitor entry prices for MIRCERA, but the court would almost certainly be called on to monitor pricing and resolve disputes that would occur during the remaining life the patents. In short, the compulsory licensing scheme he considered would result in “needless, protracted involvement in the affairs of these two companies.”
Judge Young's View of PTO Operations
The judge expressed some cynicism regarding the functioning of the PTO, stating that it is an "indisputable fact" that the Office is "perennially underfunded and slow."
Thursday, October 9, 2008
Wednesday, October 8, 2008
Judge in Amgen v. Roche Finds That Competition by Follow-On Biologic Will Likely Not Lead to a Reduction in Price, and Might Result in Price Increase
In a February post, I summarized a new report by economist Robert Shapiro which purported to identify $378 billion in savings if the US creates accelerated pathway for the approval of follow-on biologics. A critical assumption underlying this projection is that competition will inevitably drive down prices. Apparently, however, at least one federal district court judge does not buy into this assumption. Last week, in a much anticipated decision by the district court in Amgen v. Roche, Judge Young entered a permanent injunction against Roche, based in part on his conclusion that the introduction of competition into the erythropoietin (EPO) market would likely result in no reduction in Medicare costs, and might actually cause prices to rise. Most EPO purchased in the US is ultimately paid for by Medicare.
This counterintuitive economic analysis was based primarily on testimony from Amgen's expert, Prof. Douglas Bernheim, who testified that under the current Medicare reimbursement scheme there are perverse incentives at play such that competition can actually lead to higher prices. This conclusion would presumably not apply to most drugs, but depends upon the fact that EPO is a physician-dispensed drug, and the actual purchaser is the physician or healthcare provider. According to Prof.. Bernheim, a healthcare provider is not incentivized to purchase the least expensive drug, but rather the drug that offers the largest difference between the amount a provider is reimbursed under Medicare and what the provider actually paid for the drug. Because the amount of Medicare reimbursement is determined by the drugs average sale price (ASP), a higher drug price can actually increase provider reimbursement and thus lead to higher sales, contrary to the classic free market where increase in price is assumed to result in a decrease in demand.
Roche offered the testimony of its own expert, Prof. Einer Alhauge, who argued that Prof. Bernheim's analysis “really assume[s] Medicare is irrational." Prof. Elhauge opined that Medicare would rationally approve reimbursement at a rate higher than that charged by Amgen only if Roche’s product (MIRCERA, a pegylated version of EPO) is of higher quality than Amgen's product, or if Medicare thought that MIRCERA’S entry was going to lead to lower prices over time and wanted to help fund the entry to make sure that it happened. However, ultimately judge Young appears to have been persuaded that Medicare is indeed irrational, and that at least for physician-administered drugs like EPO the introduction of follow-on competition will not likely lead to lower prices. This conclusion is not only inconsistent with the Shapiro report, but with the strong movement in Congress towards providing a statutory abbreviated approval process for follow-on biologics.
It should also be noted that Judge Young also cited more convincing reasons for denying Roche's request for a compulsory license. In particular, he pointed out that allowing Roche to enter the market would cause "immense, immeasurable, irreparable harm" to Amgen, there was no compelling evidence that Roche's product provides substantial clinical benefits not provided by Amgen's products, and permitting Roche to enter the market would severely undermine the incentives for innovation provided by patents, and so crucial to the success of the pharmaceutical and biotechnology industries.
More on the decision to follow.
This counterintuitive economic analysis was based primarily on testimony from Amgen's expert, Prof. Douglas Bernheim, who testified that under the current Medicare reimbursement scheme there are perverse incentives at play such that competition can actually lead to higher prices. This conclusion would presumably not apply to most drugs, but depends upon the fact that EPO is a physician-dispensed drug, and the actual purchaser is the physician or healthcare provider. According to Prof.. Bernheim, a healthcare provider is not incentivized to purchase the least expensive drug, but rather the drug that offers the largest difference between the amount a provider is reimbursed under Medicare and what the provider actually paid for the drug. Because the amount of Medicare reimbursement is determined by the drugs average sale price (ASP), a higher drug price can actually increase provider reimbursement and thus lead to higher sales, contrary to the classic free market where increase in price is assumed to result in a decrease in demand.
Roche offered the testimony of its own expert, Prof. Einer Alhauge, who argued that Prof. Bernheim's analysis “really assume[s] Medicare is irrational." Prof. Elhauge opined that Medicare would rationally approve reimbursement at a rate higher than that charged by Amgen only if Roche’s product (MIRCERA, a pegylated version of EPO) is of higher quality than Amgen's product, or if Medicare thought that MIRCERA’S entry was going to lead to lower prices over time and wanted to help fund the entry to make sure that it happened. However, ultimately judge Young appears to have been persuaded that Medicare is indeed irrational, and that at least for physician-administered drugs like EPO the introduction of follow-on competition will not likely lead to lower prices. This conclusion is not only inconsistent with the Shapiro report, but with the strong movement in Congress towards providing a statutory abbreviated approval process for follow-on biologics.
It should also be noted that Judge Young also cited more convincing reasons for denying Roche's request for a compulsory license. In particular, he pointed out that allowing Roche to enter the market would cause "immense, immeasurable, irreparable harm" to Amgen, there was no compelling evidence that Roche's product provides substantial clinical benefits not provided by Amgen's products, and permitting Roche to enter the market would severely undermine the incentives for innovation provided by patents, and so crucial to the success of the pharmaceutical and biotechnology industries.
More on the decision to follow.
Wednesday, September 24, 2008
In re Kubin: Update on the Briefing of This Important Federal Circuit Case Relating to the Application of the Nonobviousness and Written Description
In re Kubin is an appeal of a Board of Patent Appeals and Interferences (BPAI) decision affirming the rejection of claims directed to a genus of isolated nucleic acid molecules. The claim at issue encompasses polynucleotides encoding the human NAIL protein(e.g., cDNA), along with polynucleotides encoding functional variants sharing at least 80% sequence identity with an extracellular domain of the disclosed NAIL protein. The real party in interest is Amgen, a leading biotechnology company.
Kubin involves the classic biotechnology “invention,” i.e., the isolation, cloning and characterization of a human gene (I use quotes because the Patent Office (PTO) has taken the position that the Amgen’s discovery is obvious and thus not a patentable invention.) Essentially, at the time of the alleged invention, the prior art had reported the existence of a protein on the surface of natural killer (NK) cells (which was ultimately identified as the NAIL protein), and the creation of a monoclonal antibody capable of recognizing that protein. The prior art did not describe the purification or isolation of the protein, nor did it teach the protein’s chemical structure, i.e., amino acid sequence. Against this backdrop, Amgen scientists (Kubin and Goodwin) succeeded in cloning and sequencing the cDNA sequence encoding the protein, thereby allowing them to deduce the amino acid sequence of the protein. They also determined that the protein (NAIL) binds to the CD48, a protein expressed on the surface of several types of cells involved in the immune process, including NK cells and various white blood cells such as T cells, B cells, monocytes and granulocytes . According to Amgen, the NAIL-CD48 interaction has important biological consequences, including an increase in natural killer (NK) cells cytotoxicity and stimulation of interferon gamma production. They assert that the isolation and characterization of NAIL and the cDNA encoding it allows for the generation of molecules that can modulate the activation of NK and T cells, with potential therapeutic benefits.
In claiming their invention, Amgen did not limit itself to the specific cDNA sequence encoding NAIL, but instead sought patent coverage for the entire genus of polynucleotides having the ability to encode any polypeptide (protein) sharing at least 80% identical to the extracellular domain of the human NAIL protein and which retains NAILS ability to bind CD48. This sort of broad claiming strategy is standard in biotechnology, and highly desirable in order to provide effective coverage against a competitor designing around a narrow patent limited to the actual cDNA sequence discovered, or to the genus polynucleotides encoding the native NAIL protein. It is generally possible to introduce multiple sequence alterations in at both the polynucleotide and polypeptide level without substantially affecting the protein’s function, e.g., the ability to bind CD 48. (For an extended discussion of genus claiming, the interested reader is directed to my law review article on this point).
The patent examiner rejected the claim for failure to comply with the nonobviousness, enablement and written description requirements, and on appeal the BPAI reversed on the issue of enablement but affirmed on the issues of obviousness and written description. The case is currently on appeal to the Federal Circuit, and many in the biotechnology sector view it as an important opportunity for the Federal Circuit to clarify the proper application of the nonobviousness and written description requirements to this ubiquitous type of biotechnology invention. At this point Amgen and the PTO have filed briefs with the Federal Circuit, along with a number of amicus curiae, including Eli Lilly and the Biotechnology Industry Organization (BIO), Glaxo Smith Kline, Johnson & Johnson and Novartis.
With respect to nonobviousness, a primary question to be addressed is whether, and to what extent, In re Deuel sets the standard for assessing the obviousness of newly isolated genetic sequences. In Deuel, decided in 1995, the Federal Circuit reversed a BPAI decision which had found claims similar to those at issue in Kubin to be obvious. The facts of the two cases are quite analogous. In Deuel, the BPAI found that prior art teaching a partial amino acid sequence for a protein, when combined with standard methodology for cloning a cDNA based on the knowledge of the partial amino acid sequence for a protein encoded by the cDNA, rendered obvious not only the methodology for cloning the cDNA, but also the resulting cDNA. In reversing the BPAI, the Federal Circuit essentially stated that in assessing the obviousness of a biomolecule, it is insufficient to find that the methodology for generating the biomolecule would have been obvious. Instead, it is necessary that the prior art render the biomolecule itself obvious. Many have interpreted Deuel as establishing an extremely low obviousness threshold for the patentability of biotechnology inventions, or more particularly cloning inventions. In particular, Amgen and others would argue that a newly isolated polynucleotide is generally nonobvious unless the prior art discloses a polynucleotide sequence that is sufficiently similar in chemical structure to render the newly isolated polynucleotide obvious.
In Kubin, the BPAI found that a prior art reference disclosing the NAIL protein in an unpurified state and a monoclonal antibody capable of binding the NAIL protein, along with a prophetic example describing how to clone the corresponding cDNA, rendered the claim obvious in view of what it characterized as the routine nature of isolating and sequencing a cDNA molecule once the corresponding protein has been identified and a monoclonal antibody generated. To a large extent, the BPAI is relying on the same sort of rationale that the Federal Circuit rejected in Deuel. Amgen argues that the BPAI’s decision violates precedent established by Deuel and subsequent Fed Cir decisions relating to chemical obviousness, by focusing on the obviousness of the methodology for obtaining the NAIL cDNA rather than the obviousness of the chemical structure of the cDNA. Amgen also argue that the, in any event, the cloning of this particular gene was not routine, and the prior art provided no reasonable expectation of success.
At this point there is some question as to the continuing precedential authority of Deuel, particularly in light of substantial advances in cloning methodology subsequent to date of invention in Deuel, and the Supreme Court's recent decision in KSR which effectively raised the nonobviousness bar for inventions in general. Morevover, the BPAI has previously seemed to interpret the holding in Deuel as effectively limited to the specific facts of that case. For example, in Ex parte Goldgaber, decided shortly after Deuel, the BPAI affirmed an obviousness rejection of claims directed to newly isolated cDNA sequences based on prior art teaching the encoded protein and general methodologies for cloning and sequencing cDNA, i.e., facts very analogous to both Deuel and Kubin. Thus, Kubin is an important opportunity for the Federal Circuit to address the current significance of Deuel in light of these developments. Does it broadly establish that a biomolecule is only rendered obvious by the disclosure of a structurally similar molecule in the prior art, as argued by Amgen and others in the biotechnology industry, or should Deuel be read narrowly, as superseded by developments in the technology of cloning and/or by KSR?
With respect to written description, it is noteworthy that the examiner rejected the claims for violating both the enablement and written description requirements, a common rejection in view of the substantial overlap between the enablement and Lilly written description requirements (Note that I am referring to the Lilly written description requirement, not the traditional written description requirement, which is applicable only to amended claims or claims added after the initial filing date.
A couple of years ago I published an article describing the results of an exhaustive search for decisions coming out of the BPAI and federal courts applying the Lilly written description requirement (LWD), and at that time I found only one case in either the BPAI or courts in which a claim was explicitly found to satisfy the enablement requirement but violate LWD. That case was Capon v, Esshar, decided by the BPAI. Significantly, the Federal Circuit subsequently reversed the BPAI on this point, and essentially held that the standards for compliance with the enablement and LWD overlap substantially, and that it will be a rare case in which a claim will satisfy the enablement requirement but fail to satisfy LWD. Other Federal Circuit decisions (e.g., Lizardtech) have also treated enablement and Lilly written description as essentially redundant. Nevertheless, the PTO has recently been pushing for an interpretation of LWD that would it with a role distinct from the enablement requirement. This new expanded interpretation of LWD by the PTO is reflected in Kubin, and also In re Alonso a similar case in which the BPAI found claims directed to an antibody to satisfy the enablement requirement but fail LWD and which is currently on appeal to the Federal Circuit. The PTO’s expanded interpretation of LWD is also reflected in its revised written description guidelines, discussed in a previous post. In the Federal Circuit, Judge Lourie (the architect and main proponent of LWD) recently breathed new life into LWD by affirming a district court’s invalidation of claims directed to genetic constructs for failure to comply with LWD (Carnegie Mellon v. Hoffman-LaRoche, decided 8 September 2008).
Amgen argues that the board’s application of LWD is flawed and inconsistent not only with Federal Circuit precedent, but also with the patent offices own written description guidelines (both the original and revised versions). For example, Amgen points out that example 14 of the original written description guidelines would find a claim valid which is essentially analogous to Amgen's rejected claim, albeit with a 95% identity limitation instead of Amgen’s 80%. The PTO responds that 80% encompasses many more variants than the 95% limitation in the example, but provides no guidance with respect to how one is to determine when a percent identity limitation crosses the threshold between compliance with the written description requirement and invalidity. I think this is a fundamental flaw with LWD jurisprudence – so far, neither the courts nor the PTO have provided any meaningful guidance with respect to the permitted breadth of a polynucleotide genus claim such as this under either the enablement or LWD requirements.
Turning to some of the amicus briefs, BIO, SKB and Novartis all argue in support of Kubin on the question of nonobviousness, but none weigh in on the written description issue (which I consider to be at least as important as the nonobviousness issue). Eli Lilly supports Kubin on nonobviousness, but backs the PTO on written description (not surprising perhaps, since Eli Lilly was directly involved with the creation of LWD in UC v. Eli Lilly).
Kubin involves the classic biotechnology “invention,” i.e., the isolation, cloning and characterization of a human gene (I use quotes because the Patent Office (PTO) has taken the position that the Amgen’s discovery is obvious and thus not a patentable invention.) Essentially, at the time of the alleged invention, the prior art had reported the existence of a protein on the surface of natural killer (NK) cells (which was ultimately identified as the NAIL protein), and the creation of a monoclonal antibody capable of recognizing that protein. The prior art did not describe the purification or isolation of the protein, nor did it teach the protein’s chemical structure, i.e., amino acid sequence. Against this backdrop, Amgen scientists (Kubin and Goodwin) succeeded in cloning and sequencing the cDNA sequence encoding the protein, thereby allowing them to deduce the amino acid sequence of the protein. They also determined that the protein (NAIL) binds to the CD48, a protein expressed on the surface of several types of cells involved in the immune process, including NK cells and various white blood cells such as T cells, B cells, monocytes and granulocytes . According to Amgen, the NAIL-CD48 interaction has important biological consequences, including an increase in natural killer (NK) cells cytotoxicity and stimulation of interferon gamma production. They assert that the isolation and characterization of NAIL and the cDNA encoding it allows for the generation of molecules that can modulate the activation of NK and T cells, with potential therapeutic benefits.
In claiming their invention, Amgen did not limit itself to the specific cDNA sequence encoding NAIL, but instead sought patent coverage for the entire genus of polynucleotides having the ability to encode any polypeptide (protein) sharing at least 80% identical to the extracellular domain of the human NAIL protein and which retains NAILS ability to bind CD48. This sort of broad claiming strategy is standard in biotechnology, and highly desirable in order to provide effective coverage against a competitor designing around a narrow patent limited to the actual cDNA sequence discovered, or to the genus polynucleotides encoding the native NAIL protein. It is generally possible to introduce multiple sequence alterations in at both the polynucleotide and polypeptide level without substantially affecting the protein’s function, e.g., the ability to bind CD 48. (For an extended discussion of genus claiming, the interested reader is directed to my law review article on this point).
The patent examiner rejected the claim for failure to comply with the nonobviousness, enablement and written description requirements, and on appeal the BPAI reversed on the issue of enablement but affirmed on the issues of obviousness and written description. The case is currently on appeal to the Federal Circuit, and many in the biotechnology sector view it as an important opportunity for the Federal Circuit to clarify the proper application of the nonobviousness and written description requirements to this ubiquitous type of biotechnology invention. At this point Amgen and the PTO have filed briefs with the Federal Circuit, along with a number of amicus curiae, including Eli Lilly and the Biotechnology Industry Organization (BIO), Glaxo Smith Kline, Johnson & Johnson and Novartis.
With respect to nonobviousness, a primary question to be addressed is whether, and to what extent, In re Deuel sets the standard for assessing the obviousness of newly isolated genetic sequences. In Deuel, decided in 1995, the Federal Circuit reversed a BPAI decision which had found claims similar to those at issue in Kubin to be obvious. The facts of the two cases are quite analogous. In Deuel, the BPAI found that prior art teaching a partial amino acid sequence for a protein, when combined with standard methodology for cloning a cDNA based on the knowledge of the partial amino acid sequence for a protein encoded by the cDNA, rendered obvious not only the methodology for cloning the cDNA, but also the resulting cDNA. In reversing the BPAI, the Federal Circuit essentially stated that in assessing the obviousness of a biomolecule, it is insufficient to find that the methodology for generating the biomolecule would have been obvious. Instead, it is necessary that the prior art render the biomolecule itself obvious. Many have interpreted Deuel as establishing an extremely low obviousness threshold for the patentability of biotechnology inventions, or more particularly cloning inventions. In particular, Amgen and others would argue that a newly isolated polynucleotide is generally nonobvious unless the prior art discloses a polynucleotide sequence that is sufficiently similar in chemical structure to render the newly isolated polynucleotide obvious.
In Kubin, the BPAI found that a prior art reference disclosing the NAIL protein in an unpurified state and a monoclonal antibody capable of binding the NAIL protein, along with a prophetic example describing how to clone the corresponding cDNA, rendered the claim obvious in view of what it characterized as the routine nature of isolating and sequencing a cDNA molecule once the corresponding protein has been identified and a monoclonal antibody generated. To a large extent, the BPAI is relying on the same sort of rationale that the Federal Circuit rejected in Deuel. Amgen argues that the BPAI’s decision violates precedent established by Deuel and subsequent Fed Cir decisions relating to chemical obviousness, by focusing on the obviousness of the methodology for obtaining the NAIL cDNA rather than the obviousness of the chemical structure of the cDNA. Amgen also argue that the, in any event, the cloning of this particular gene was not routine, and the prior art provided no reasonable expectation of success.
At this point there is some question as to the continuing precedential authority of Deuel, particularly in light of substantial advances in cloning methodology subsequent to date of invention in Deuel, and the Supreme Court's recent decision in KSR which effectively raised the nonobviousness bar for inventions in general. Morevover, the BPAI has previously seemed to interpret the holding in Deuel as effectively limited to the specific facts of that case. For example, in Ex parte Goldgaber, decided shortly after Deuel, the BPAI affirmed an obviousness rejection of claims directed to newly isolated cDNA sequences based on prior art teaching the encoded protein and general methodologies for cloning and sequencing cDNA, i.e., facts very analogous to both Deuel and Kubin. Thus, Kubin is an important opportunity for the Federal Circuit to address the current significance of Deuel in light of these developments. Does it broadly establish that a biomolecule is only rendered obvious by the disclosure of a structurally similar molecule in the prior art, as argued by Amgen and others in the biotechnology industry, or should Deuel be read narrowly, as superseded by developments in the technology of cloning and/or by KSR?
With respect to written description, it is noteworthy that the examiner rejected the claims for violating both the enablement and written description requirements, a common rejection in view of the substantial overlap between the enablement and Lilly written description requirements (Note that I am referring to the Lilly written description requirement, not the traditional written description requirement, which is applicable only to amended claims or claims added after the initial filing date.
A couple of years ago I published an article describing the results of an exhaustive search for decisions coming out of the BPAI and federal courts applying the Lilly written description requirement (LWD), and at that time I found only one case in either the BPAI or courts in which a claim was explicitly found to satisfy the enablement requirement but violate LWD. That case was Capon v, Esshar, decided by the BPAI. Significantly, the Federal Circuit subsequently reversed the BPAI on this point, and essentially held that the standards for compliance with the enablement and LWD overlap substantially, and that it will be a rare case in which a claim will satisfy the enablement requirement but fail to satisfy LWD. Other Federal Circuit decisions (e.g., Lizardtech) have also treated enablement and Lilly written description as essentially redundant. Nevertheless, the PTO has recently been pushing for an interpretation of LWD that would it with a role distinct from the enablement requirement. This new expanded interpretation of LWD by the PTO is reflected in Kubin, and also In re Alonso a similar case in which the BPAI found claims directed to an antibody to satisfy the enablement requirement but fail LWD and which is currently on appeal to the Federal Circuit. The PTO’s expanded interpretation of LWD is also reflected in its revised written description guidelines, discussed in a previous post. In the Federal Circuit, Judge Lourie (the architect and main proponent of LWD) recently breathed new life into LWD by affirming a district court’s invalidation of claims directed to genetic constructs for failure to comply with LWD (Carnegie Mellon v. Hoffman-LaRoche, decided 8 September 2008).
Amgen argues that the board’s application of LWD is flawed and inconsistent not only with Federal Circuit precedent, but also with the patent offices own written description guidelines (both the original and revised versions). For example, Amgen points out that example 14 of the original written description guidelines would find a claim valid which is essentially analogous to Amgen's rejected claim, albeit with a 95% identity limitation instead of Amgen’s 80%. The PTO responds that 80% encompasses many more variants than the 95% limitation in the example, but provides no guidance with respect to how one is to determine when a percent identity limitation crosses the threshold between compliance with the written description requirement and invalidity. I think this is a fundamental flaw with LWD jurisprudence – so far, neither the courts nor the PTO have provided any meaningful guidance with respect to the permitted breadth of a polynucleotide genus claim such as this under either the enablement or LWD requirements.
Turning to some of the amicus briefs, BIO, SKB and Novartis all argue in support of Kubin on the question of nonobviousness, but none weigh in on the written description issue (which I consider to be at least as important as the nonobviousness issue). Eli Lilly supports Kubin on nonobviousness, but backs the PTO on written description (not surprising perhaps, since Eli Lilly was directly involved with the creation of LWD in UC v. Eli Lilly).
Wednesday, September 10, 2008
Federal Circuit Decides Important Biotechnology Written Description Case
In Carnegie Mellon University v. Hoffman-La Roche, decided September 8, 2008, the Federal Circuit affirmed a district court’s holding on summary judgment that claims directed to a recombinant plasmid containing a DNA polymerase I gene, isolated from any bacterial source, under the control of the conditionally controllable foreign promoter were invalid for failure to comply with the written description requirement.
The decision is significant in a number of respects. For one thing, it is the first time the Federal Circuit has invalidated a claim directed to a DNA sequence (or any biological molecule for that matter) in a situation not involving the discovery of a novel DNA sequence. For example, in UC Regents v. Eli Lilly (1997), the first case in which the Federal Circuit applied the written description requirement to invalidate an originally filed, non-amended claim, the invalidated claims were directed towards novel genetic sequences encoding insulin. Likewise, in In re Wallach (2004) the court invalidated claims reciting newly discovered TBP-II genes for failure to comply with what I refer to as the Lilly written description requirement (in contrast with the traditional written description requirement, which is applicable only to amended claims or claims added after the initial filing date).
In Carnegie Mellon, on the other hand, the essence of the invention was the discovery of a means of regulating the expression of the gene encoding DNA polymerase I (polA) to avoid lethality problems associated with over expression of the gene. The polA gene had previously been isolated from three different species of bacteria, and conditionally controllable promoters were also known in the prior art. Carnegie Mellon argued that the holding in Lilly was limited to inventions involving novel DNA sequences, but the Federal Circuit explicitly rejected this argument.
The court also voiced its approval of the USPTO’s 2001 “Written Description Guidelines,” (which were recently superseded by updated guidelines, as discussed in a previous post) finding it to be “persuasive authority” providing guidance in interpreting the written description requirement.
Carnegie Mellon is arguably inconsistent with the Federal Circuit's 2005 decision in Capon v. Eshhar, in which the Federal Circuit reversed a Board of Patent Appeals and Interferences holding that a claim directed to a chimeric DNA sequence failed to satisfy the written description requirement. For example, in Capon the court found that the Board had erred in holding that the written description requirement was not met because the specification failed to recite the structure or formula or chemical name for the DNA sequence components of the claimed chimeric genes. In that case, the court held that it was sufficient that sequences that could function as the segments of the chimeric gene were known in the prior art, and that the applicant was not required to "reiterate" those sequences in the application. In contrast, in Carnegie Mellon the court found it significant that the patent specification only discloses a DNA polymerase gene from a single bacterial species, even though the corresponding gene from two other bacterial species were known at the time of invention.
More striking, the claims in Capon define the chimeric gene in terms of two genetic elements that are defined in much broader, generic terms than the DNA polymerase I gene at issue in Carnegie Mellon. In particular, the Capon claims recite a genus of chimeric genes comprising a first segment encoding some portion of an antibody capable of binding an antigen, and the second segment encoding at least some portion of a protein that (1) is expressed on the surface of cells of the immune system and (2) triggers activation and/or proliferation of the cells. The Carnegie Mellon court distinguished the two cases by noting that in Capon the prior art contained “extensive knowledge of the nucleotide structure of the various and immune-related segments of DNA, including over 785 mouse antibody DNA light chains and 1327 mouse antibody DNA heavy chains,” while only three bacterial DNA polymerase I genes “out of thousands” had been cloned at the time of Carnegie Mellon's invention. However, the court fails to acknowledge the extreme structural variability in antibodies (Capon) compared to DNA polymerases (Carnegie Mellon). Furthermore, the court entirely ignores the extremely broad functional language defining the second element of the Capon chimeric gene, i.e., any segment encoding a protein expressed on the cell surface of cells of the immune system (there are a host of such cells) and capable of triggering activation or proliferation of the cells.
Carnegie Mellon is a significant affirmation of a continuing distinct role for the written description requirement in policing the scope of patent claims, typically in the area of biotechnology. However, it provides little guidance in applying the doctrine, particularly when one considers the seeming inconsistency between the outcomes in this case and Capon. My position is that the court is misguided in its application of the written description requirement in this manner, and the more appropriate tool for policing claim scope is the enablement requirement, but particularly over the last few years Lilly written description is increasingly taking on a life of its own.
The decision is significant in a number of respects. For one thing, it is the first time the Federal Circuit has invalidated a claim directed to a DNA sequence (or any biological molecule for that matter) in a situation not involving the discovery of a novel DNA sequence. For example, in UC Regents v. Eli Lilly (1997), the first case in which the Federal Circuit applied the written description requirement to invalidate an originally filed, non-amended claim, the invalidated claims were directed towards novel genetic sequences encoding insulin. Likewise, in In re Wallach (2004) the court invalidated claims reciting newly discovered TBP-II genes for failure to comply with what I refer to as the Lilly written description requirement (in contrast with the traditional written description requirement, which is applicable only to amended claims or claims added after the initial filing date).
In Carnegie Mellon, on the other hand, the essence of the invention was the discovery of a means of regulating the expression of the gene encoding DNA polymerase I (polA) to avoid lethality problems associated with over expression of the gene. The polA gene had previously been isolated from three different species of bacteria, and conditionally controllable promoters were also known in the prior art. Carnegie Mellon argued that the holding in Lilly was limited to inventions involving novel DNA sequences, but the Federal Circuit explicitly rejected this argument.
The court also voiced its approval of the USPTO’s 2001 “Written Description Guidelines,” (which were recently superseded by updated guidelines, as discussed in a previous post) finding it to be “persuasive authority” providing guidance in interpreting the written description requirement.
Carnegie Mellon is arguably inconsistent with the Federal Circuit's 2005 decision in Capon v. Eshhar, in which the Federal Circuit reversed a Board of Patent Appeals and Interferences holding that a claim directed to a chimeric DNA sequence failed to satisfy the written description requirement. For example, in Capon the court found that the Board had erred in holding that the written description requirement was not met because the specification failed to recite the structure or formula or chemical name for the DNA sequence components of the claimed chimeric genes. In that case, the court held that it was sufficient that sequences that could function as the segments of the chimeric gene were known in the prior art, and that the applicant was not required to "reiterate" those sequences in the application. In contrast, in Carnegie Mellon the court found it significant that the patent specification only discloses a DNA polymerase gene from a single bacterial species, even though the corresponding gene from two other bacterial species were known at the time of invention.
More striking, the claims in Capon define the chimeric gene in terms of two genetic elements that are defined in much broader, generic terms than the DNA polymerase I gene at issue in Carnegie Mellon. In particular, the Capon claims recite a genus of chimeric genes comprising a first segment encoding some portion of an antibody capable of binding an antigen, and the second segment encoding at least some portion of a protein that (1) is expressed on the surface of cells of the immune system and (2) triggers activation and/or proliferation of the cells. The Carnegie Mellon court distinguished the two cases by noting that in Capon the prior art contained “extensive knowledge of the nucleotide structure of the various and immune-related segments of DNA, including over 785 mouse antibody DNA light chains and 1327 mouse antibody DNA heavy chains,” while only three bacterial DNA polymerase I genes “out of thousands” had been cloned at the time of Carnegie Mellon's invention. However, the court fails to acknowledge the extreme structural variability in antibodies (Capon) compared to DNA polymerases (Carnegie Mellon). Furthermore, the court entirely ignores the extremely broad functional language defining the second element of the Capon chimeric gene, i.e., any segment encoding a protein expressed on the cell surface of cells of the immune system (there are a host of such cells) and capable of triggering activation or proliferation of the cells.
Carnegie Mellon is a significant affirmation of a continuing distinct role for the written description requirement in policing the scope of patent claims, typically in the area of biotechnology. However, it provides little guidance in applying the doctrine, particularly when one considers the seeming inconsistency between the outcomes in this case and Capon. My position is that the court is misguided in its application of the written description requirement in this manner, and the more appropriate tool for policing claim scope is the enablement requirement, but particularly over the last few years Lilly written description is increasingly taking on a life of its own.
Wednesday, August 27, 2008
Courts grapple with the question of whether biological phenomena constitute unpatentable “natural phenomona”
It is black letter law that “natural phenomena, principles of nature and abstract ideas," as exemplified by fundamental physical relationships such as E=mc2 and the law of gravity, are unpatentable under section 101 of the patent statute. It is less clear what the criteria is for determining when a patent claim sufficiently embodies a natural phenomenon to be unpatentable; a series of recent decisions have interpreted Supreme Court precedent as establishing that a patent claim which wholly preempts a natural phenomenon is unpatentable. In LabCorp v. Metablolite (548 U.S. 124 (2006), Justice Breyer (joined by Souter and Stevens in a dissent from the Court’s decision to dismiss the case as improvidently granted) expressed his view that a patent claim that “embodies” a naturally existing biological phenomenon (in this case a correlation between homocysteine and vitamin B in the body) is unpatentable for improperly claiming a natural phenomenon.
Breyer’s Labcorp dissent has opened the door to challenging the validity of claims directed to biological phenomena under section 101. Of particular note, at least three subsequent district court decisions have weighed in on the question of whether a biological or physiological phenomenon is properly considered an unpatentable “natural phenomenon” – a question of particular importance to the biotechnology sector in the wake of LabCorp.
In one case, Ariad v. Lilly (529 F.Supp.2d 106 (D.Mass 2007)), the district court rejected Lilly’s argument that Ariad’s claim (directed to inhibition of the NF-kB pathway, a naturally occurring signal transduction pathway in humans) embodies a natural phenomonen. The court’s reasoning in Ariad was quite bizarre – first the court identified the relevant natural phenomenon narrowly as the so-called NF-kB “Autoregulatory Loop,” and then held that the Autoregulatory Loop is “an incomplete model ... subject to a significant amount of ambiguity and inconsistency” and “that Lilly has failed to prove by clear and convincing evidence that the Autoregulatory Loop exists in living cells in a way that is encompassed by Ariad's claims.” The existence of the NF-kB pathway is well established, as is the fact that its inhibition can have a physiological effect, and the claim at issue broadly covers inhibition of the pathway. The court appears to have been bamboozled into adopting an unduly narrow definition of the relevant natural phenomenon, and then found that the evidence supporting this narrowly defined phenomenon was insufficient to overcome the statutory presumption in favor of patent validity.
In two other recent decisions the district court came to the opposite conclusion, finding the claims at issue invalid for impermissibly claiming a physiological natural phenomenon. What is interesting about these cases is that the supposed natural phenomena actually relate to biological correlations that only exist as the result of human, medical intervention in the body, by either the administration of a vaccine or a drug. In contrast, the natural phenomena at issue in LabCorp and Ariad involve correlations and pathways that exist absent any human intervention.
In Classen v. Biogen (not reported), the patents are based on Dr. Classen’s purported discovery of a relationship between vaccination schedule (the timing of the administration of a series of vaccinations) and the development of a “chronic immune-mediated disorder,” and broadly claim methods of comparing vaccine schedules to determine the safest schedule. Defendant’s include manufacturers and distributors of vaccines, who allegedly infringe the patents (directly and/or indirectly) by participating in post-approval vaccination safety studies. In that case, the district court held that the relationship between vaccination schedule and autoimmune disorders is without question a natural phenomonen, and that the claims at issue impermissibly embodied that natural phenomonen. The court did not address the issue of whether a biological phenomenon that exists only as a result of human intervention (the introduction of vaccine in the human body) is accurately characterized as a natural phenomenon. Arguably, a phenomena that does not exist absent human intervention should not be considered a natural phenomenon in the same way as fundamental principles of nature that exist outside of human intervention, such as inhibition of the NF-kB pathway, or the correlation between homocysteine and vitamin B at the center of the claims at issue in LabCorp.
To my mind, perhaps the most troubling decision is the most recent, Prometheus v. Mayo (2008 WL 878910). In that case, the court held that the correlation between the level of thiopurine drug metabolites in the human body and therapeutic efficacy and safety is a “natural phenomonen,” and that Prometheus’ asserted claim were invalid under section 101 for “wholly preempting” this supposed natural phenomenon. Prometheus argued that since the drug metabolites only exist in the human body as the result of human intervention (i.e., by administration of the thiopurine drugs), the correlation cannot be considered a natural phenomonen. However, the court ruled that because thiopurine drugs “are converted naturally by enzymes within the patient’s body to form an agent that is therapeutically active, . . . the correlation results from a natural body process,” and is thus is an unpatentable “work of nature.”
Thus, the court in Prometheus finds that the mere involvement of a natural process renders a correlation that exists only as the result of human intervention an unpatentable natural phenomenon. This seems nonsensical – after all, what invention, in any area of technology, does not at some level involve natural processes? Any electronic invention relies on the fundamental nature of electrons and materials such as silicon. Mechanical inventions rely on the law of gravity and friction. And what biological invention does not involve natural biological processes?
All three decisions are on track for review by the Federal Circuit. I think it is particularly important that the court reverse the Prometheus courts unduly broad definition of unpatentable natural phenomenon, which in combination with its “wholly preempted” test could render many important biotechnology or pharmaceutical patents susceptible to challenge under section 101. For example, a patent claiming a drug product could be characterized as wholly preempting the “natural phenomenon” of the drugs pharmacologic effect in the human body. A method of treatment claim might be challenged for wholly preempting the correlation between administration of the drug and therapeutic effect.
The approach adopted by the Prometheus court would seem to bode particularly ill for the patentability of personalized medicine inventions. In a broad sense, personalized medicine refers to assessment of an individual patient’s safety and efficacy profile for a drug, and selecting drugs and dosage regimens based on this assessment. Fundamentally, personalized medicine relies on the discovery of correlations between biomarkers and the safety and efficacy of drugs in an individual. But under the Prometheus court’s approach, such correlations would seem to be natural phenomena, since they are based on natural processes, even though the correlations would not exist absent human intervention by administration of a drug to a patient. This approach would render questionable the validity of any patent broadly claiming the use of such a correlation.
It is unclear to what role the patenting of such man-made correlations will play in the development of personalized medicine. Perhaps some would argue that broad personalized medicine patents will actually deter innovation and commercialization in this important area. But I question whether the blunt tool of the patentable subject matter doctrine is the appropriate mechanism for determining the patentability of such inventions. This interpretation of Section 101 would seem to open up the door to challenging a number of important issued patents in the biological realm, leading at the very least to increased litigation and uncertainty in this important technology sector.
Breyer’s Labcorp dissent has opened the door to challenging the validity of claims directed to biological phenomena under section 101. Of particular note, at least three subsequent district court decisions have weighed in on the question of whether a biological or physiological phenomenon is properly considered an unpatentable “natural phenomenon” – a question of particular importance to the biotechnology sector in the wake of LabCorp.
In one case, Ariad v. Lilly (529 F.Supp.2d 106 (D.Mass 2007)), the district court rejected Lilly’s argument that Ariad’s claim (directed to inhibition of the NF-kB pathway, a naturally occurring signal transduction pathway in humans) embodies a natural phenomonen. The court’s reasoning in Ariad was quite bizarre – first the court identified the relevant natural phenomenon narrowly as the so-called NF-kB “Autoregulatory Loop,” and then held that the Autoregulatory Loop is “an incomplete model ... subject to a significant amount of ambiguity and inconsistency” and “that Lilly has failed to prove by clear and convincing evidence that the Autoregulatory Loop exists in living cells in a way that is encompassed by Ariad's claims.” The existence of the NF-kB pathway is well established, as is the fact that its inhibition can have a physiological effect, and the claim at issue broadly covers inhibition of the pathway. The court appears to have been bamboozled into adopting an unduly narrow definition of the relevant natural phenomenon, and then found that the evidence supporting this narrowly defined phenomenon was insufficient to overcome the statutory presumption in favor of patent validity.
In two other recent decisions the district court came to the opposite conclusion, finding the claims at issue invalid for impermissibly claiming a physiological natural phenomenon. What is interesting about these cases is that the supposed natural phenomena actually relate to biological correlations that only exist as the result of human, medical intervention in the body, by either the administration of a vaccine or a drug. In contrast, the natural phenomena at issue in LabCorp and Ariad involve correlations and pathways that exist absent any human intervention.
In Classen v. Biogen (not reported), the patents are based on Dr. Classen’s purported discovery of a relationship between vaccination schedule (the timing of the administration of a series of vaccinations) and the development of a “chronic immune-mediated disorder,” and broadly claim methods of comparing vaccine schedules to determine the safest schedule. Defendant’s include manufacturers and distributors of vaccines, who allegedly infringe the patents (directly and/or indirectly) by participating in post-approval vaccination safety studies. In that case, the district court held that the relationship between vaccination schedule and autoimmune disorders is without question a natural phenomonen, and that the claims at issue impermissibly embodied that natural phenomonen. The court did not address the issue of whether a biological phenomenon that exists only as a result of human intervention (the introduction of vaccine in the human body) is accurately characterized as a natural phenomenon. Arguably, a phenomena that does not exist absent human intervention should not be considered a natural phenomenon in the same way as fundamental principles of nature that exist outside of human intervention, such as inhibition of the NF-kB pathway, or the correlation between homocysteine and vitamin B at the center of the claims at issue in LabCorp.
To my mind, perhaps the most troubling decision is the most recent, Prometheus v. Mayo (2008 WL 878910). In that case, the court held that the correlation between the level of thiopurine drug metabolites in the human body and therapeutic efficacy and safety is a “natural phenomonen,” and that Prometheus’ asserted claim were invalid under section 101 for “wholly preempting” this supposed natural phenomenon. Prometheus argued that since the drug metabolites only exist in the human body as the result of human intervention (i.e., by administration of the thiopurine drugs), the correlation cannot be considered a natural phenomonen. However, the court ruled that because thiopurine drugs “are converted naturally by enzymes within the patient’s body to form an agent that is therapeutically active, . . . the correlation results from a natural body process,” and is thus is an unpatentable “work of nature.”
Thus, the court in Prometheus finds that the mere involvement of a natural process renders a correlation that exists only as the result of human intervention an unpatentable natural phenomenon. This seems nonsensical – after all, what invention, in any area of technology, does not at some level involve natural processes? Any electronic invention relies on the fundamental nature of electrons and materials such as silicon. Mechanical inventions rely on the law of gravity and friction. And what biological invention does not involve natural biological processes?
All three decisions are on track for review by the Federal Circuit. I think it is particularly important that the court reverse the Prometheus courts unduly broad definition of unpatentable natural phenomenon, which in combination with its “wholly preempted” test could render many important biotechnology or pharmaceutical patents susceptible to challenge under section 101. For example, a patent claiming a drug product could be characterized as wholly preempting the “natural phenomenon” of the drugs pharmacologic effect in the human body. A method of treatment claim might be challenged for wholly preempting the correlation between administration of the drug and therapeutic effect.
The approach adopted by the Prometheus court would seem to bode particularly ill for the patentability of personalized medicine inventions. In a broad sense, personalized medicine refers to assessment of an individual patient’s safety and efficacy profile for a drug, and selecting drugs and dosage regimens based on this assessment. Fundamentally, personalized medicine relies on the discovery of correlations between biomarkers and the safety and efficacy of drugs in an individual. But under the Prometheus court’s approach, such correlations would seem to be natural phenomena, since they are based on natural processes, even though the correlations would not exist absent human intervention by administration of a drug to a patient. This approach would render questionable the validity of any patent broadly claiming the use of such a correlation.
It is unclear to what role the patenting of such man-made correlations will play in the development of personalized medicine. Perhaps some would argue that broad personalized medicine patents will actually deter innovation and commercialization in this important area. But I question whether the blunt tool of the patentable subject matter doctrine is the appropriate mechanism for determining the patentability of such inventions. This interpretation of Section 101 would seem to open up the door to challenging a number of important issued patents in the biological realm, leading at the very least to increased litigation and uncertainty in this important technology sector.
Tuesday, August 5, 2008
Federal Circuit Interprets the 271(e)(1) Regulatory Approval Exemption Narrowly, a Welcome Development for the Developers of Research Tools
Today in Proveris Scientific v. Innovasystems the Federal Circuit construed the Section 271(e)(1) regulatory approval exemption narrowly, holding that the sale of a device known as an Optical Spray Analyzer (“OSA”)" does not fall within the 271(e)(1) safe harbor. As discussed in a previous post, the defendant argued that because the device is intended solely for use in generating data for submission to FDA under the FDCA and is only sold to pharmaceutical companies or FDA, sales of the OSA falls squarely within the plain language of the statute, which states in relevant part that “[i]t shall not be an act of infringement to . . . offer to sell, or sell . . . a patented invention solely for uses reasonably related to the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs or veterinary biological products.” However, in a unanimous decision penned by Judge Schall, the Federal Circuit rejected this argument, holding that the OSA is not a “patented invention” because it is not a product that requires FDA premarket approval.
Citing to Eli Lilly v. Medtronic, a 1990 Supreme Court decision which held that the 271(e)(1) safe harbor applies to class III medical devices, Judge Schall noted that Congress had enacted 271(e)(1) primarily as a means of addressing two unintended “distortions” of the effective patent term of drugs. As the result of the second of these distortions, prior to the enactment of 271(e)(1) (as part of the Hatch-Waxman Act) drug patent owners enjoyed a de facto patent term extension resulting from the lag between patent expiration and the time required to secure FDA approval for a generic version of a drug (or other FDA regulated product.” However, the court held that “[b]ecause the OSA device is not subject to FDA premarket approval, and therefore faces no regulatory barriers to market entry upon patent expiration, [the defendant] is not a party who, prior to enactment of [271(e)(1)], could be said to have been adversely affected by the second distortion. For this reason, we do not think Congress could have intended that the safe harbor of section 271(e)(1) apply to it.”
Of course, in its 2005 Merck v. Integra decision, the Supreme Court held that the 271(e)(1) exemption applied to patented peptides used as research tools in preclinical drug discovery. Thus, the “patented invention” in Merck was not subject to FDA premarket approval but nevertheless fell within 271(e)(1). So although on its face Proveris suggests that 271(e)(1) only applies to products that require premarket approval, this would appear to be inconsistent with Merck - although in principle one might argue that one could seek FDA approval for Integra’s patented peptides, from a practical point of view these were never true drug candidates. Notably, Proveris does not mention the nature of the “patented invention,” in Merck, or attempt to reconcile the two decisions. But if the court had held that 271(e)(1) applied to the patented OSA, the decision would have raised serious doubts as to the value of a host of patents claiming instruments used in drug development research, and research tool patents in general. In his dissent to the Federal Circuit’s decision in Merck on remand, Judge Rader suggested that in that case the majority had effectively eliminated patent protection for research tools, but judging from the court’s decision in Proveris this clearly does not seem to be the case. A great deal of uncertainty remains with respect to where the court’s will draw the line between the “patented inventions” falling within the exemption (such as the peptides in Merck), and research tools falling outside the safe harbor.
Citing to Eli Lilly v. Medtronic, a 1990 Supreme Court decision which held that the 271(e)(1) safe harbor applies to class III medical devices, Judge Schall noted that Congress had enacted 271(e)(1) primarily as a means of addressing two unintended “distortions” of the effective patent term of drugs. As the result of the second of these distortions, prior to the enactment of 271(e)(1) (as part of the Hatch-Waxman Act) drug patent owners enjoyed a de facto patent term extension resulting from the lag between patent expiration and the time required to secure FDA approval for a generic version of a drug (or other FDA regulated product.” However, the court held that “[b]ecause the OSA device is not subject to FDA premarket approval, and therefore faces no regulatory barriers to market entry upon patent expiration, [the defendant] is not a party who, prior to enactment of [271(e)(1)], could be said to have been adversely affected by the second distortion. For this reason, we do not think Congress could have intended that the safe harbor of section 271(e)(1) apply to it.”
Of course, in its 2005 Merck v. Integra decision, the Supreme Court held that the 271(e)(1) exemption applied to patented peptides used as research tools in preclinical drug discovery. Thus, the “patented invention” in Merck was not subject to FDA premarket approval but nevertheless fell within 271(e)(1). So although on its face Proveris suggests that 271(e)(1) only applies to products that require premarket approval, this would appear to be inconsistent with Merck - although in principle one might argue that one could seek FDA approval for Integra’s patented peptides, from a practical point of view these were never true drug candidates. Notably, Proveris does not mention the nature of the “patented invention,” in Merck, or attempt to reconcile the two decisions. But if the court had held that 271(e)(1) applied to the patented OSA, the decision would have raised serious doubts as to the value of a host of patents claiming instruments used in drug development research, and research tool patents in general. In his dissent to the Federal Circuit’s decision in Merck on remand, Judge Rader suggested that in that case the majority had effectively eliminated patent protection for research tools, but judging from the court’s decision in Proveris this clearly does not seem to be the case. A great deal of uncertainty remains with respect to where the court’s will draw the line between the “patented inventions” falling within the exemption (such as the peptides in Merck), and research tools falling outside the safe harbor.
Thursday, July 24, 2008
To what extent do protein patents cover posttranslationally modified proteins?
People often treat proteins as simply a string of amino acids as specified by a gene, and in particular this is often how they are represented in patents, i.e., essentially as a string of letters. For example, many patents claim proteins as defined by an amino acid sequence referenced by SEQ ID NO. However, while this simplistic view of proteins holds true for most bacterial proteins, and some very small proteins such as human growth hormone, most eukaryotic proteins (and in particular naturally occurring human proteins) are posttranslationaly modified. That is, the process of protein translation in a cell results in a string of amino acids as specified by the corresponding gene, but this polypeptide backbone is subsequently chemically modified to result in a final protein product having a markedly different structure than the originally translated polypeptide backbone. These posttranslational modifications involve the chemical addition of sugars and other molecules to certain amino acids in the polypeptide backbone. Common examples of posttranslational modifications include phosphorylation, sialation, glycosylation, and acetylation. Posttranslational modifications can dramatically affect the structure and function of the protein. Some modifications, particularly glycosylations, dramatically increase the size of the posttranslational modified protein relative to the protein as initially translated. Furthermore, these modifications entail the removal of one or more atoms from the modified amino acid, e.g., most of these modifications cause the removal of a hydrogen atom from the original amino acid and replacement by the modifying chemical group.
As a side note, posttranslational modification is one of the primary rationales offered by the biotechnology industry as to why an abbreviated approval process for generic biologics would raise safety and efficacy issues not presented by traditional small molecule generic drugs. Posttranslational modifications can be quite complex, and purified protein preparations are typically heterogeneous mixtures of a variety of different posttranslationally modified forms of the protein. The specific nature of posttranslational modifications for any given protein preparation can vary based on the particular cell line in which the gene is expressed, and the production conditions under which the cells are grown, the protein is expressed, and the protein is harvested and purified. The potential for substantial alteration in safety and/or efficacy based on minor changes to the production process, or use of a different cell line, is cited as necessitating a full scale assessment of safety and efficacy for any follow-on biologic, even if it has the same amino acid backbone sequence as a currently approved originator biologic.
Back to my main point, which is, to what extent do patent claims directed to proteins encompass posttranslationally modified proteins? The issue has been presented to the Federal Circuit in a brief filed by Roche on May 13, 2008 appealing the decision by the lower court in Amgen v. Roche (2008 WL 2231038). The allegedly infringing Roche product is MIRCERA, a pegylated version of erythropoietin (EPO) that has been on the market in Europe since 2007, and has been approved in the US by FDA. Pegylation is a form of non-natural posttranslational modification, which is accomplished outside of the cell in a chemical process that results in a covalent attachment of polyethylene glycol (PEG) to one or more amino acids on a protein. Pegylation can dramatically affect the biological and pharmacologic function of the protein, particularly by altering binding characteristics and increasing the half-life of the protein in the human body. Increased half-life translates into less frequent administration of the drug to the patient, with obvious clinical benefits.
Roche's challenge is directed particularly at claim 3 of U.S. Patent No. 5547933, which claims a “non-naturally occurring glycoprotein product of the expression in a mammalian host cell of an exogenous DNA sequence comprising a DNA sequence encoding human erythropoietin said product possessing the in vivo biological property of causing bone marrow cells to increase production of reticulocytes and red blood cells.” Roche argues that MIRCERA is not and cannot be produced by a mammalian host cell, as required by the claim, but can only produced by means of the cell free pegylation reaction. Roche also stresses the very different chemical and functional characteristics of MIRCERA in comparison to EPO produced in a cell. In particular, Roche argues that MIRCERA has twice the molecular weight, a lower molecular charge, different hydrodynamic size, different chromatographic profiles, different binding, distinct carbohydrate structures, and a much longer half-life. Roche also emphasizes the functional distinctions between MIRCERA and Amgen’s EPO products, arguing that MIRCERA provides substantially superior outcome for at least some subset of patients, and thus fulfills a medical need that is not currently being satisfied in the US.
It will be interesting to see how the Federal Circuit responds to Roche’s argument. In a previous decision involving a related Amgen EPO patent, Amgen v. HMR (314 F.3d 1313) the Federal Circuit strictly limited another claim directed towards the EPO protein to the amino acid sequence referenced in the claim. In that case, the claim defined EPO by reference to an amino acid sequence listing containing 166 amino acids. This amino acid sequence was derived from the DNA sequence of the corresponding EPO gene. However, subsequent to the filing date of the patent it was learned that the initially translated 166 amino acid EPO was posttranslational processed in the cell, resulting in the removal of a terminal amino acid, such that the final EPO product actually only contains 165 amino acids. The Federal Circuit held that this unanticipated posttranslational modification resulted in the claim not covering mature human EPO literally or under the doctrine of equivalents.
Looking beyond Amgen v. Roche, what are the ramifications of this issue for other protein patents? Many patents claim proteins solely in terms of amino acid sequence, with no explicit mention of posttranslational modifications. Commercially relevant products, such as biologic drugs, are often posttranslationally modified, and one might argue that many proteins claims are not literally infringed by posttranslationally modified versions of the claimed sequences. To find literal infringement, at the very least a court would have to find that the amino acids recited in the claim include derivatives of an amino acid lacking a hydrogen atom, and that the claim encompasses any of a host of chemical additions not defined in the specification. And if the court interprets the claim this broadly, potential 112 issues are raised. The Federal Circuit has repeatedly stated that biomolecules (like proteins and DNA) are chemical compounds, and that patent law should be applied to biomolecules in the same manner as it is to other chemicals. But in other contexts, an inventor is not permitted to claim a genus of chemical compounds sharing a common structural backbone and encompassing any chemical additions to that backbone. If the court were presented with this argument, and chose to actually treat a protein claim like any other chemical claim (as it did in Amgen v. HMR), then perhaps it would find that an interpretation of the claim that would encompass any posttranslationally modified version of the recited amino acid sequence violates the disclosure requirements of section 112.
As a side note, posttranslational modification is one of the primary rationales offered by the biotechnology industry as to why an abbreviated approval process for generic biologics would raise safety and efficacy issues not presented by traditional small molecule generic drugs. Posttranslational modifications can be quite complex, and purified protein preparations are typically heterogeneous mixtures of a variety of different posttranslationally modified forms of the protein. The specific nature of posttranslational modifications for any given protein preparation can vary based on the particular cell line in which the gene is expressed, and the production conditions under which the cells are grown, the protein is expressed, and the protein is harvested and purified. The potential for substantial alteration in safety and/or efficacy based on minor changes to the production process, or use of a different cell line, is cited as necessitating a full scale assessment of safety and efficacy for any follow-on biologic, even if it has the same amino acid backbone sequence as a currently approved originator biologic.
Back to my main point, which is, to what extent do patent claims directed to proteins encompass posttranslationally modified proteins? The issue has been presented to the Federal Circuit in a brief filed by Roche on May 13, 2008 appealing the decision by the lower court in Amgen v. Roche (2008 WL 2231038). The allegedly infringing Roche product is MIRCERA, a pegylated version of erythropoietin (EPO) that has been on the market in Europe since 2007, and has been approved in the US by FDA. Pegylation is a form of non-natural posttranslational modification, which is accomplished outside of the cell in a chemical process that results in a covalent attachment of polyethylene glycol (PEG) to one or more amino acids on a protein. Pegylation can dramatically affect the biological and pharmacologic function of the protein, particularly by altering binding characteristics and increasing the half-life of the protein in the human body. Increased half-life translates into less frequent administration of the drug to the patient, with obvious clinical benefits.
Roche's challenge is directed particularly at claim 3 of U.S. Patent No. 5547933, which claims a “non-naturally occurring glycoprotein product of the expression in a mammalian host cell of an exogenous DNA sequence comprising a DNA sequence encoding human erythropoietin said product possessing the in vivo biological property of causing bone marrow cells to increase production of reticulocytes and red blood cells.” Roche argues that MIRCERA is not and cannot be produced by a mammalian host cell, as required by the claim, but can only produced by means of the cell free pegylation reaction. Roche also stresses the very different chemical and functional characteristics of MIRCERA in comparison to EPO produced in a cell. In particular, Roche argues that MIRCERA has twice the molecular weight, a lower molecular charge, different hydrodynamic size, different chromatographic profiles, different binding, distinct carbohydrate structures, and a much longer half-life. Roche also emphasizes the functional distinctions between MIRCERA and Amgen’s EPO products, arguing that MIRCERA provides substantially superior outcome for at least some subset of patients, and thus fulfills a medical need that is not currently being satisfied in the US.
It will be interesting to see how the Federal Circuit responds to Roche’s argument. In a previous decision involving a related Amgen EPO patent, Amgen v. HMR (314 F.3d 1313) the Federal Circuit strictly limited another claim directed towards the EPO protein to the amino acid sequence referenced in the claim. In that case, the claim defined EPO by reference to an amino acid sequence listing containing 166 amino acids. This amino acid sequence was derived from the DNA sequence of the corresponding EPO gene. However, subsequent to the filing date of the patent it was learned that the initially translated 166 amino acid EPO was posttranslational processed in the cell, resulting in the removal of a terminal amino acid, such that the final EPO product actually only contains 165 amino acids. The Federal Circuit held that this unanticipated posttranslational modification resulted in the claim not covering mature human EPO literally or under the doctrine of equivalents.
Looking beyond Amgen v. Roche, what are the ramifications of this issue for other protein patents? Many patents claim proteins solely in terms of amino acid sequence, with no explicit mention of posttranslational modifications. Commercially relevant products, such as biologic drugs, are often posttranslationally modified, and one might argue that many proteins claims are not literally infringed by posttranslationally modified versions of the claimed sequences. To find literal infringement, at the very least a court would have to find that the amino acids recited in the claim include derivatives of an amino acid lacking a hydrogen atom, and that the claim encompasses any of a host of chemical additions not defined in the specification. And if the court interprets the claim this broadly, potential 112 issues are raised. The Federal Circuit has repeatedly stated that biomolecules (like proteins and DNA) are chemical compounds, and that patent law should be applied to biomolecules in the same manner as it is to other chemicals. But in other contexts, an inventor is not permitted to claim a genus of chemical compounds sharing a common structural backbone and encompassing any chemical additions to that backbone. If the court were presented with this argument, and chose to actually treat a protein claim like any other chemical claim (as it did in Amgen v. HMR), then perhaps it would find that an interpretation of the claim that would encompass any posttranslationally modified version of the recited amino acid sequence violates the disclosure requirements of section 112.
Subscribe to:
Posts (Atom)
Posts
