Thursday, December 23, 2010

Lawsuit Alleging Infringement of Canine Genotyping Patent Raise Interesting Patent Eligibility Issues

On December 17 Fred Hutchinson Cancer Research Center and its licensees (Argus Genetics and Mars, Inc.) sued MMI Genomics for infringement of a patent directed towards methods of using genetic testing to identify the breed of the dog (US patent 7,729,863). According to the complaint, MMI Genomics is a company based in Davis, California that provides dog breed identification services associated with its Canine Heritage XL Breed Test product.

Some of the claims in the asserted patent raise interesting issues of patent eligibility, particularly in view of the recent Bilski and Prometheus decisions, both discussed extensively on this blog, and the pending appeal of AMP v. PTO (the ACLU challenge to Myriad’s BRCA gene patents). For example, consider claim 1.

1. A method for determining the contributions of canid populations to a canid genome, comprising: (a) genotyping a sample obtained from a test canid to determine the identity of one or both alleles present in the test canid genome for each of a set of markers, wherein the set of markers is indicative of the contributions of canid populations to the genome of the test canid; (b) using a specifically programmed computer comprising an algorithm to compare the identity of one or both alleles for each of the set of markers determined to be present in the test canid genome to a database comprising a plurality of canid population profiles, wherein each canid population profile comprises genotype information for the set of markers in the canid population; and (c) determining the contributions of canid populations to the test canid genome.

The heart of the claimed method seems to reside in applying an algorithm to genotype information using information derived there from to determine the contribution various dog breeds to dog being tested. If that were all the claim covered, I think under Bilski and Prometheus a court would likely find the claim patent ineligible for embodying a fundamental principle. The fundamental principle might be characterized as an abstract idea, or as a natural phenomenon, either would work.

But the claim includes two elements that might be sufficient to render the claim patent eligible: (1) the step of genotyping a sample taken from the dog, and (2) the use of a "specifically programmed computer." In Prometheus, the Federal Circuit relied heavily on the machine or transformation test in arriving at the conclusion that the claims at issue in that case were patent eligible. I believe that even though the Supreme Court clarified in Bilski that the machine or transformation test is not the only test for patent eligibility, it did acknowledge that the test can be a useful and informative tool for assessing patent eligibility, and I predict that the Federal Circuit will continue to rely heavily on the machine or transformation test when faced with claims such as this, as it did in Prometheus.

Under Prometheus and Bilski, I think this one could come out either way. On the one hand, a court could find that genotyping is inherently transformative, because it necessarily involves physical manipulation of biological samples, in the same way that the step of "determining drug metabolite level” was found transformative in Prometheus, and sufficient to render the claim patent eligible. A court could also find that the use of a "specifically programmed computer" satisfies the machine prong of the machine or transformation test.

On the other hand, a court could characterize the genotyping step as mere data-gathering, i.e., "insubstantial extra-solution activity,” and thus ignore it in the analysis. This is what the Federal Circuit did in In re Grams, and as noted in my previous blog post the Federal Circuit explicitly held in Prometheus that Grams is still good law. In Prometheus, the Federal Circuit distinguish the two cases by pointing out that in Grams the focus of the invention is on the algorithm itself, and that the data-gathering step generically covered many different types of clinical diagnostic tests, whereas the Prometheus claims were targeted to specific drugs and specific diseases. The dog breed testing claim seems to fall somewhere in between. I think a court could go either way, by either stressing that the claim focuses on the algorithm and is not limited to any specific genotype or genetic test, in which case the claim is patent ineligible under Grams, or by stressing that the genotyping step is "central to the purpose" of the claim and thus sufficient to render the claim patent eligible under Prometheus.

Regarding the use of a computer, it is important to note that under In re Bilski the machine or transformation test requires the involvement of a "particular" machine, and many have speculated that a general-purpose computer would not satisfy this requirement of "particularity." Whoever drafted the dog breed testing claim wisely specified that the computer used in the method is "specifically programmed," but I'm not sure if that will be sufficient. A court could dismiss the "specifically programmed" language and find that the mere involvement of the computer does not satisfy the "particularity” aspect of the machine or transformation test.

Claim 35 is similar to claim 1, but instead of reciting the use of a computer, includes a step of "applying a computer-implemented statistical model.” This claim could be more vulnerable than claim 1, since it does not appear to explicitly recite the use of a particular machine.

There are also a number of dependent claims limited to specific SNP markers, specific dog breeds, etc., and these claims should face less of a threat of invalidation for lack of patent eligibility. Patent eligibility analysis often hinges on an assessment of whether the claim preempts all substantial practical uses of an abstract idea or natural phenomenon, so logically the narrower a claim is the more likely it is not preemptive, and thus patent eligible. Although in Bilski the Supreme Court apparently rejected this logic, because it held even relatively narrow dependent claims to be patent ineligible even though they clearly did not preempt all uses of risk hedging (the fundamental principle at issue in that case).

Claim 36 appears to be a Beauregard claim, a claim directed toward computer readable medium on which a computer program is inscribed. It would be interesting to see how this claim holds up, it is my understanding that the status of this sort of claim has not been entirely resolved post-Bilski, I saw some posts on this topic on Patently-O a while back.

36. A computer readable medium comprising stored thereon: (a) a data structure stored thereon for use in distinguishing canid populations, the data structure comprising: (i) a marker field, which is capable of storing the name of a marker or of an allele of the marker; and (ii) a genotype information field, which is capable of storing genotype information for the marker in a canid population, wherein a record comprises an instantiation of the marker field and an instantiation of the genotype information field and a set of records represents a canid population profile; and (b) computer-executable instructions for implementing a method for determining the contributions of canid populations to a canid genome, comprising: (i) obtaining the identity of one or both alleles in a test canid genome for each of a set of markers; and (ii) determining the contributions of canid populations to the test canid genome by comparing the alleles in the test canid genome to a database comprising canid population profiles, wherein each canid population profile comprises genotype information for the set of markers in the canid population.

Claim 37 seems even more problematic to me. I don't know that it can be called a Beauregard claim, it appears to basically be directed toward a computer readable medium on which a data structure is stored, which I am guessing is not the same thing as a computer program. I would be particularly interested in seeing how court deals with this claim.

37. A computer-readable medium comprising a data structure stored thereon for use in distinguishing canid populations, the data structure comprising: (a) a marker field, which is capable of storing the name of a marker or of an allele of the marker; and (b) a genotype information field, which is capable of storing genotype information for the marker in a canid population, wherein a record comprises an instantiation of the marker field and an instantiation of the genotype information field and a set of records represents a canid population profile, wherein the marker field comprises a set of markers indicative of the contributions of canid populations to the genome of a test canid.

Friday, December 17, 2010

On Remand, Federal Circuit (Once Again) Decides Prometheus v. Mayo in Favor of Patent Eligibility for Methods of Treatment and Diagnostic Tests

Today a panel of the Federal Circuit issued a decision in Prometheus v. Mayo, an important case involving the application of the patent eligibility doctrine to biotechnology, and more particularly molecular diagnostic testing and personalized medicine. I have discussed this case in numerous previous blog posts, see for example here.

In a nutshell, the case involves claims directed towards methods for optimizing the amount of drug administered to a specific patient, an example of personalized medicine. The claims are limited to certain categories of thiopurine drugs used to treat immune-mediated gastrointestinal disorders such as Crohn's disease, the use of which is often accompanied by serious adverse side effects, including hepatotoxicity. Because different patients metabolize the drug differently, it had been difficult for doctors to ascertain the proper dosage, and some doctors had reportedly been hesitant to prescribe the drugs at all for fear of toxic side effects.

The challenged claims are directed towards methods that all comprise a step of determining the level of thiopurine drug metabolite in a patient sometime after administering the drug to the patient, and based on the result of that determination recognizing that the dosage of drug administered to the patient should be increased or reduced, i.e.,a determination that the level of drug metabolite exceeds a certain threshold would indicate that the dosage should be decreased, and vice versa. Many of the claims also explicitly recite a step of administering the drug to the patient.

A district court judge, in an opinion clearly influenced by Judge Breyer's provocative dissent in LabCorp, held the claims patent ineligible for allegedly wholly preempting a "natural phenomenon." In its analysis, the judge dismissed the steps of administering drug and determining drug metabolite level as mere "data-gathering steps," irrelevant to the patent eligibility of the claim.

The case was appealed to the Federal Circuit, but before the appeal was decided the Federal Circuit issued its en banc In re Bilski decision (Bilski I) which held the machine or transformation test to be the definitive test for patent eligibility of method claims.

On appeal the Federal Circuit panel reversed and held that the Prometheus claims satisfied applied the newly minted Bilski machine or transformation test. In particular, the panel held that the steps of administering drug to the patient and determining drug metabolite level were both inherently and independently sufficient to satisfy the transformation prong of the machine or transformation test. In contrast with the district court, which dismissed these steps as mere data-gathering, the Federal Circuit characterized the administering and determining steps as central to the purpose of the claim and thus relevant to the question of patent eligibility.

But shortly thereafter, in Bilski v. Kappos (Bilski II), the Supreme Court overturned Bilski I to the extent that it had characterized the machine or transformation test as the definitive test for patent eligibility, and reiterated that the sole and fundamental test for patent eligibility is whether the claim "patents" a fundamental principle, i.e., laws of nature, physical phenomena and abstract ideas. The Supreme Court then vacated the Federal Circuit's decision in Prometheus v. Mayo and remanded the case to the Federal Circuit to decide in a manner consistent with Bilski II.

Today's Post-Bilski II Prometheus Decision
Today's Federal Circuit decision on remand comes to the same conclusion as the first, reversing the district court and holding all of the challenged claims to be patent eligible. The court's analysis is also virtually identical to the original decision, focusing primarily on the transformative nature of the administering and determining steps.

It is important to bear in mind that while Bilski II rejected the notion that the machine or transformation test is the only test for patent eligibility of method claims, it did acknowledge that the machine or transformation test is in many instances highly probative of the fundamental inquiry. Significantly, the Supreme Court did not offer up any other alternative test for determining whether a claim "patents a fundamental principle." As noted by the en banc Federal Circuit in Bilski I, as a practical matter it can be extremely difficult to apply this fundamental test to actual claims, which is why the Federal Circuit tried to institute the use of a definitive machine or transformation test as a more objective and easier to administer proxy for the ultimate inquiry. As a practical matter of expedience, I have predicted that the courts and PTO will continue to rely heavily on the machine or transformation test post-Bilski II.

In today's decision, the Federal Circuit panel pointed out that under Bilski II the scope of patent eligible subject matter remains quite broad, and that while fundamental principles are not patent eligible "particular applications" of fundamental principles are. Applying this standard, the panel held that Prometheus' claims do not preempt the correlations between drug metabolite level and optimal dosage, but rather utilize them in "the treatment of a specific disease by administering specific drugs and measuring specific metabolites."

The panel also emphasized the continuing relevance of the machine or transformation test, pointing out that the Supreme Court had characterize it as a "useful and important clue, an investigative tool" for determining patent eligibility. It then came to the same conclusion as it did in its pre-Bilski II decision, finding that both the administering and determining steps were each inherently and independently sufficient to confer patent eligibility on the claims. In view of this determination, the panel found it unnecessary to address the question of whether the claims satisfy the machine prong of the test.

The panel emphasized that if the claims had not included a step of administering a drug or determining the level of drug metabolite, but had merely recited recognizing the correlation between drug metabolite level and the desirability of modifying the dosage (i.e., the "wherein" clauses of the claims), then the claims would have been patent ineligible for claiming nothing more than "mental steps." But by including a step of administering a drug and/or determining the level of drug metabolite, transformative steps which are "central to the purpose" of the method, the claim is rendered patent eligible.

In effect, Bilski II had essentially on effect on the outcome of the Prometheus appeal. Fortunately, the Supreme Court decided Bilski II in a manner allows a great deal of discretion for the Federal Circuit and lower courts. This panel of the Federal Circuit clearly sees the importance of permitting the patenting of diagnostic methods and methods of treating patients with drugs, and sought to maintain their patent eligiblity by emphatically establishing that method claims reciting a step of administering a specific class of drugs to a patient, or of performing a molecular diagnostic test of the patient, are patent eligible per se.

However, the panel is also emphatic that claims that could be potentially infringed by merely thinking about or recognizing a physiological correlation cannot be patented, a hypothetical scenario that some critics of gene patents and patents on diagnostic tests often point to.

The one aspect of the decision that I found less than entirely convincing was the panel's attempt to distinguish over the holding in In re Grams, a 1989 Federal Circuit decision. Grams was a case involving the patent eligibility of claims reciting methods that basically involved (1) performing a clinical diagnostic test on an individual patient, and (2) applying an algorithm to the data generated by the test in order to determine whether an abnormality exists, and possible causes of the abnormality. In Grams, the data-gathering steps were disregarded in the court's patent eligibility analysis, and the claims held patent ineligible for "in essence" claiming nothing more than a patent ineligible algorithm.

In a sense, the Grams claims are highly analogous to the Prometheus claims. The determining step in the Prometheus claims involves nothing more than performing a clinical diagnostic tests on individual patient, and the step of using the results to determine whether drug dosage should be altered is nothing more than a simple algorithm. According to the Prometheus panel, the distinction lies in the fact that in Grams the "essence" of the claimed process is the algorithm, and that the Gram patent "focused only on the algorithm rather than a clinical test." I'm not sure how convincing this distinction is, perhaps the panel could have been more straightforward and acknowledged the tension between Grams and Prometheus.

However, I have read the Grams decision and think a plausible distinction is that the Prometheus claims are limited to diagnostic tests for a limited number of specified of drug metabolites, while in Gram the claim appears to broadly encompass any of a host of clinical tests to which the algorithm could be applied. This suggests that if the Prometheus claims had been drafted more broadly to encompass any drug and drug metabolite, rather than being restricted to specific classes of thiopurine drugs, the claims could have been found patent ineligible. In effect, the distinction between the Grams and Prometheus claims is a matter of claim scope, something I think would be better addressed using the enablement requirement rather than patent eligibility, as I argued in the amicus brief I filed with Robert Cook-Deegan in the Myriad gene patent case.

Finally, what are the implications of today's Prometheus decision for the patent eligibility of the genetic diagnostic method claims at issue in the Myriad gene patent case (Association for Pathology v. US Patent and Trademark Office)? Today's decision does not alter my previous prediction, which is that the outcome will depend upon how the Federal Circuit interprets the scope of Myriad's patent claims. If the Federal Circuit adopts the ACLU's broad interpretation the claims (which is the interpretation the district court judge accepted for his analysis), pursuant to which the claims broadly cover the purely mental process of analyzing genetic information, then under Prometheus I believe the claims must be found patent ineligible. Alternatively, if the Federal Circuit adopts Myriad’s interpretation of its claims, under which the claims are limited to methods that involve physically isolating and analyzing DNA molecules, I think under Prometheus the claims should be found patent eligible by virtue of comprising a transformative step (highly analogous to the administering step in the Prometheus claims) central to the purpose of the claimed methods.

Tuesday, December 7, 2010

The American Medical Association et al. File Amici Brief Arguing for Affirmance of AMP v. PTO

An amici brief was filed yesterday by Professors Lori Andrews and Joshua Sarnoff on behalf of the American Medical Association, American Society of Human Genetics, American College of Obstetricians and Gynecologists, American College of Embryology, and the Medical Society of the State of New York, arguing for affirmance of the district court's decision in AMP v. PTO. Professor Andrews has for years been one of the most outspoken critics of gene patents, and was involved with a bill introduced in Congress in 2007 that if enacted would have broadly prohibited the patenting of DNA-based inventions.

Their brief is more alarmist than the brief submitted by the ACLU/PubPat, arguing that gene patents have cost patient's their lives, and the healthcare system billions of dollars. For example, allege that gene patents are being asserted against physicians across the country, and that physicians have to worry about infringing patents based on nothing more than mere "conscious thoughts." They also contend that gene patents are killing patients. For example, they state that drug companies have used these patents to prevent use of genetic testing to determine whether a drug will help or harm patients, to block the availability of genetic testing (which they allege resulted in the of a 10-year-old patient), and to force patients to use poor quality diagnostic tests rather than improved alternatives (which they allege resulted in a patient undergoing unnecessary removal of ovaries based on erroneous BRC genetic test result).

The AMA et al. adopt a very broad definition of the diagnostic method claims, under which they would be infringed by anyone merely "reading and thinking about the sequence data disclosed in the patent." As discussed in a previous post, this is a much broader interpretation of the claims than Myriad has proffered.

They also argue that patents are not necessary to incentivize the discovery of genetic mutations correlated with disease or the commercialization of diagnostic testing based on these discoveries, citing to the Sec.'s Advisory Committee on Genetics, Health and Society (SACGHS) Report.

In the amici brief submitted by me and Robert Cook-Deegan, we argued that regardless of whether patents are necessary to incentivize the development of genetic diagnostic testing for Mendelian traits, in which mutations in a single gene are highly correlated with a specific disease, they quite likely will be necessary for the optimal development of next-generation genetic diagnostic testing involving much more complex relationships and multiple genes, or if FDA imposes heightened regulation on diagnostic testing. We also argued that a decision broadly rendering gene patents patent ineligible could have serious unintended negative consequences for biotechnology, particularly outside the realm of Mendelian genetic diagnostic testing, and that there are other more focused approaches for dealing with problematic gene patents or gene patent enforcement activities. The AMA brief does not address these points, but rather seems to assume that the only available mechanisms for dealing with problems associated with gene patents is a broad prohibition extending far beyond Myriad and BRCA testing.

Monday, December 6, 2010

Myriad and the ACLU Disagree over Claim Scope, But Apparently Not over Patent Eligibility of Genetic Diagnostic Methods

After reading the briefs submitted by Myriad and the ACLU/PubPat in AMP v. PTO (the challenge to Myriad’s gene patents, described in previous posts to this blog), one thing that struck me was that the parties do not appear to disagree over the patent eligibility of genetic diagnostic methods. In fact, with respect to these method claims, the only dispute is with respect to the proper interpretation of the claims.

The challenged diagnostic method claims recite processes of either "analyzing" or "comparing" nucleotide sequences. For example, claim 1 of US patent number 6,033,857 claims:

A method for identifying a mutant BRCA2 nucleotide sequence in a suspected mutant BRCA2 allele which comprises comparing the nucleotide sequence of the suspected mutant BRCA2 allele with the wild-type BRCA2 nucleotide sequence, wherein a difference between the suspected mutant and the wild-type sequence identifies a mutant BRCA2 nucleotide sequence.

Myriad argues for a relatively narrow interpretation of the claims, under which the claims are limited to methods which require actual physical analysis of DNA molecules. Under their interpretation, the term "nucleotide sequence," as used in the claims, refers to actual polynucleotides, i.e., DNA or RNA molecules, and the step of comparing these molecules necessarily requires physically isolating polynucleotides from a patient's tissue sample, and then processing and analyzing the molecules. Myriad argues that these processes are all clearly transformative of physical molecules, and because the transformations are “central to the purpose of the claims,” the claims satisfy the machine or transformation test. They cite to Prometheus for the proposition that transformative steps that are central to the purpose of the claims cannot be disregarded in the analysis as mere "data-gathering steps,” and argue these transformative steps render the claim patent eligible.

The ACLU, on the other hand, argues for a much broader interpretation of the method claims. They argue that the term "nucleotide sequence" in the claims refers to sequence information, not to molecules, and that the claims cover the purely mental process of "comparing" or "analyzing" DNA sequence information. Under this broad interpretation, the claims clearly do not require any sort of physical transformation, and on this basis the ACLU argues they fail the machine or transformation test, which they correctly note remains "a useful and important clue" to patent eligibility (using the language of the Supreme Court).

Interestingly, the ACLU never suggests in their brief that the diagnostic method claims would be patent ineligible if limited to methods involving actual isolation and processing of DNA molecules, i.e., the only processes covered by the claims under Myriad's interpretation of the claims. Conversely, Myriad never argues that a method claim that encompasses purely mental processes for comparing DNA sequence information would be patent eligible. Thus, both parties seem implicitly to agree on a consensus approach under which a claim directed toward a genetic diagnostic method is patent eligible if limited to diagnostic methods that involve actual isolation and manipulation of DNA molecules, but patent ineligible if the claim would also cover merely comparing DNA sequence information.

Notably, the district court adopted the ACLU's broad interpretation of the claims, and held that the claims cover a process of comparing or analyzing DNA sequence information. However, in dicta the district court went even further, stating that "[e]ven if the challenged claims were read to include the transformations associated with isolating and sequencing human DNA, these transformations would constitute no more than ‘data-gathering step[s]’ that are not central to the purpose of the claimed process.. . . . Consequently, even if the method claims-in-suit were construed to include the physical transformations associated with isolating and sequencing DNA, they would still fail the ‘machine or transformation’ test under Section 101 for subject matter patentability."

In my analysis of the district court decision, I pointed out that this dicta seems clearly wrong, and is inconsistent with the Federal Circuit's decision in Prometheus. Clearly, these data-gathering steps are central to the purpose of the claim and should be included in the machine or transformation analysis. I suspect that the ACLU agrees with my assessment, and is not even trying to argue for more extreme position taken by the district court.

It seems to me that the implicit consensus between Myriad and the ACLU is correct, i.e., method of genetic diagnostic claims are patent eligible if they include steps involving the actual physical manipulation of DNA molecules, but patent ineligible if they would cover the wholly mental process of analyzing DNA sequence information.

As a side note, another point of contention regarding claim interpretation exists with respect to whether some of the claims are limited to cDNA molecules. In its amicus brief, the United States government has argued that claims to isolated genomic DNA are patent ineligible, but claims limited to cDNA molecules are patent eligible (as reported in an earlier post to this blog). For some reason, in its brief the ACLU argues that the United States is mistaken, and that none of the challenged claims is limited to cDNA. ACLU seems to be clearly mistaken on this point--for example, Claim 2 of US patent 5,747,282 (one of the challenged claims) recites "DNA [having] the nucleotide sequence set forth in SEQ ID NO:1.” The Sequence Listing section of the patent specification explicitly identifies SEQ ID NO:1 as a cDNA.

Tuesday, November 23, 2010

Boston Globe Editorial Acknowledges Importance of Gene Patents, and Suggests Issue Should Be Decided by Congress, Not the Courts

In an apparent shift of position, the Boston Globe has parted company with much of the mainstream media by acknowledging the importance of gene patents to biotechnology, the potential deleterious impact if Judge Sweet's decision finding gene patents ineligible were to be upheld on appeal. The editorial points out that:

Whatever the principle at stake, one practical effect of Judge Sweet’s ruling, if it isn’t reversed, will be to eliminate part of the legal structure upon which the biotechnology industry was founded. The US might leave the company of Japan, Canada, Australia, and the European Union to become the biggest industrialized nation to consider a gene that’s been discovered, characterized, and isolated through scientific inquiry as categorically non-patentable.
. . .
There is still, of course, the very legitimate, and clearly urgent, question of whether gene patents serve the larger goal of advancing medical research and promoting the creation of therapies. But the best answer to this question is unlikely to arrive through a patent challenge. Determining the benefit to society of gene patents involves much more than just the narrow legal consideration of whether DNA isolated through the process of scientific discovery fits the definition of an invention under the US Patent Act.
. . .
Perhaps the best policy would be to simply do away with these patents. But the issue is of such magnitude that it would be best addressed outside the court system, by an act of Congress.

It is good to see that at least some in the mainstream media are waking up to the reality of what is at stake in the ACLU's challenge to gene patents.

Regeneron Sues for Declaratory Judgment That VEGF Trap Does Not Infringe Genentech Patents

On November 19, Regeneron sued Genentech in the US District Court for the Southern District of New York seeking a declaration that no activities relating to the Regeneron VEGF Trap infringe any claim of US patent numbers 5,952,199; 6,100,071; 6,383,486; 6,897,294; and 7,771,721 (all assigned to Genentech according to the complaint). The complaint is available here, compliments of

Vascular Endothelial Growth Factor, or VEGF, is a protein growth factor which is required for formation of blood vessels, a process known as angiogenesis. According to Wikipedia:

In the adult human, angiogenesis is only consistently active in the gut - however in pathological cases such as cancer, where a solid tumor grows rapidly and therefore requires a blood supply, there is also active angiogenesis. Therefore, the blockade of VEGF is a potentially viable strategy for treating solid tumors.
Regeneron has developed a chimeric (i.e., fusion) protein, which it refers to as VEGF Trap (aflibercept), that comprises the extra-cellular domain of VEGF Receptor (the growth factor's natural receptor). The extra-cellular domain of the receptor is used to bind, or "trap" VEGF, and thereby keep it from binding receptors on the surface of blood vessels. Regeneron is working towards FDA approval to market VEGF Trap for a number of ophthalmologic and oncology indications. According to the complaint, the company intends to submit a Biologic License Application before the end of the second quarter of 2011 to market the product for a wet age-related macular degeneration. Yesterday, Regeneron stock rose 19.7% after the company announced positive Phase 3 clinical trial results for the drug (see here).

Genentech has developed an antibody called Avastin that binds VEGF. Avastin is already approved for the treatment of colon cancer. The five Genentech patents which are the subject of the declaratory judgment action are directed toward chimeric VEGF receptor proteins capable of binding VEGF.

Wednesday, November 10, 2010

Amicus Brief Filed By Alnylam Pharmaceuticals Argues That Funk Bros. Is About Obviousness, Not Patent Eligibility

Alnylam Pharmaceuticals is a biopharmaceutical company based in Cambridge, Massachusetts focused on developing synthetic "small interfering RNAs" (SiRNAs) as therapeutic agents. It has filed an amicus curiae brief in the Myriad gene patent case (AMP v. US PTO) that makes a number of insightful points, particularly in attempting to "debunk myths of Funk Brothers Seed Co. v. Kalo Inoculant Co."

Funk Brothers is generally assumed to be a case about patent eligibility. However, Alnylam points out that Funk Brothers was decided by the Supreme Court in 1948, prior to the enactment of the 1952 patent statute and thus prior to 35 USC 101, which is the statutory basis for the patent eligibility doctrine. Prior to 1952, the Court used the term “invention” to refer to "inventiveness," a concept that subsequent to 1952 and the enactment of 35 USC 103 we refer to as "nonobviousness." The Court in Funk Brothers found the claimed subject matter was not the “product of invention,” but Alnylam argues that post-1952 this statement should be interpreted a finding of obviousness, not patent ineligibility.

I have always found it extremely difficult, if not impossible, to find a principled distinction between the inventions claimed in Funk Brothers and Diamond V. Chakrabarty that would justify finding only Dr. Chakrabarty's invention patent eligible. In Funk Brothers, the alleged invention was a novel and useful combination, in a single inoculant, of naturally occurring microorganisms in a combination that did not exist naturally. Chakrabarty’s invention was a novel and useful combination, in a single microorganism, of naturally occurring plasmids in a combination that did not exist naturally. It is important to bear in mind that Chakrabarty's invention did not involve any genetic engineering at the molecular level.

In Funk Brothers, it has been pointed out that while the combination of bacteria in a single inoculant was novel, the bacteria function in the same manner and have the same characteristics as they would have in nature. But the same can be said regarding the plasmids Dr. Chakrabarty introduced into a single microorganism. These were naturally occurring plasmids, and in the claimed microorganism they encoded the exact same proteins, and performed the exact same function as they did in their natural state.

In a nutshell, Funk Brothers was a novel combination of naturally occurring bacteria capable of achieving novel and useful function, with the bacteria functioning in the same manner as they do in nature. Chakrabarty was a novel combination of naturally occurring plasmid DNA capable of achieving novel and useful function, with the plasmid DNA functioning in the same manner as they do in nature. I don't see how one invention can be patent eligible and the other not. However, if Funk Brothers was really decided based on obviousness rather than patent eligibility, as argued by Alnylam, the two decisions are easily reconciled.

Friday, November 5, 2010

Superficial Appeal of DOJ Brief Proves Illusory

I can understand why many people find the position advanced by the US government in AMP v. PTO (the DOJ amicus brief, discussed in an earlier post) quite appealing. On its face, it seems to offer a reasonable compromise: patent eligibility for "engineered" cDNA, but not for genomic DNA merely "excised" from a human chromosome. However, the distinction is superfical, and does not in my view withstand closer scrutiny.

In particular, the DOJ conveniently ignores the fact that in almost all cases isolation of genomic DNA involves amplification, either through laboratory techniques such as PCR, or by replication in recombinant cells. This distinguishes DNA and RNA from other biomolecules. Prior to the development of recombinant molecular biology in the 1970s, isolation of a naturally occurring biomolecule, such as a protein, involved removing the biomolecule produced in a naturally occurring from other biological constituents. Importantly, all of the purified biomolecule originated in the natural source, and was purified by separating it from other cellular constituents.

In contrast, because DNA can serve as a template for its own replication, DNA is normally isolated by amplifying it many times ("million-fold" is the term used in the amici brief filed by BIO and AUTM). As a result, a typical preparation of "isolated genomic DNA" actually consists primarily of synthetic copies of the genomic DNA molecule.

The DOJ makes much of the fact that a cDNA molecule does not contain the introns, which exist in most human genes, and implies that molecular biologists have "engineered" the introns out. In fact, the introns are naturally removed in the body when genomic DNA is transcribed into mRNA. There is no human engineering involved; production of a cDNA molecule simply entails making a DNA copy of a naturally occurring mRNA and then amplifying it.

In most instances a preparation of isolated genomic DNA comprises synthetic DNA molecules, produced in the laboratory by amplification, that correspond in sequence to a naturally occurring genomic DNA. A preparation of cDNA comprises synthetic DNA molecules, produced in the laboratory by amplification, that correspond in sequence to a naturally occurring mRNA. Considered in this light, I think that the distinction DOJ makes between "excised" genomic DNA versus "engineered" cDNA proves largely illusory.

True, a cDNA molecule has a slightly different chemical structure than the mRNA on which it is based, but the differences are small. The sugar group (ribose) in DNA is lacking an oxygen atom (hence the “deoxy”), and one of the four base groups in RNA (uracil) is slightly different than the corresponding base in DNA (thymine), due to replacement of a methyl group with a hydrogen atom. But mRNA and its corresponding cDNA both encode the same information, and both hybridize to the same complementary strand.

In many instances, a synthetic copy of a genomic DNA molecule is also chemically different from the original. Most naturally occurring genomic DNA is methylated, a form of epigenetic modification, while synthetic copies are often not methylated.

In short, the minor structural difference between cDNA and mRNA is not so qualitatively different than the difference between amplified genomic DNA and genomic DNA of natural origin to warrant using this distinction as the basis for drawing a line between patent eligible and patent ineligible subject matter.

Thursday, November 4, 2010

Summary of Myriad Patent Claims Under Attack in AMP v. PTO

Today I published a short article in Managing IP "discerning" Myriad's patent claims, currently under attack in AMP v. PTO . The article is available here for subscribers to Managing IP, along with other materials relating to the Myriad case.

The substance of my article is as follows:

As a preliminary matter, it bears noting that the precise contours of Myriad's patent claims are not entirely clear. Claim interpretation is a notoriously unpredictable exercise, and no court has ever formally construed the claims, or considered their scope in the context of infringement litigation. In fact, no US court has ever interpreted a human gene patent claim in a case involving genetic diagnostic testing. In an amicus brief filed with the Federal Circuit in AMP v. PTO, the author (Holman) and Robert Cook-Deegan explain that if the claims were ever enforced in an infringement suit there are reasons to believe that the claims would not be interpreted as broadly as many critics of gene patents have assumed. The following analysis is based on a facial reading of the claims in view of the patent specifications, as well as statements made by the district court and by Myriad.

The challenged patent claims recite subject matter falling under general three categories: isolated DNA molecules claimed as compositions of matter, methods of diagnostic testing, and cell-based assay methods. Importantly, none of the claims encompass DNA as it exists naturally in the human body, and none of the claims recite mere genetic information. Thus, popular allegations that the patent claims confer ownership on people's bodies, or can be used to prevent a doctor from communicating the existence of a genetic predisposition to cancer to a patient, would appear to be unfounded.

Isolated DNA claims
All of the composition of matter claims are limited to "isolated" DNA molecules, and hence clearly do not encompass genomic DNA as it exists naturally in the human body. This is consistent with long-standing patent office policy which requires claims directed to naturally occurring DNA molecules to be limited to isolated, purified and/or recombinant embodiments of the molecule, in order to prevent the patent from covering native genes in the human body

There is, however, some uncertainty with respect to exactly how broadly the term "isolated" should be interpreted. Myriad's patent specifications define the term to encompass any DNA sequence that has been “removed from its naturally occurring environment, and includes recombinant or cloned DNA isolates and chemically synthesized analogs or analog biologic synthesized by heterologous systems." Many in the biotechnology community have assumed that claims to "isolated" DNA molecules broadly encompass the claimed sequence in any context outside its native environment, such as residing in a recombinant DNA vector, cell or organism. However, there is little US case law directly addressing the proper interpretation of the term "isolated" in this context, and there is reason to believe that if litigated the term might be given a narrower construction, at least in some cases (depending, for example, on how the term "isolated" is used in the patent specification and in communications to the patent office during patent prosecution).

Some of the composition of matter claims encompass any isolated DNA that codes for a full length BRCA1 or BRCA2 protein. These claims are relatively broad in the sense that they would appear to cover any of the astronomical number of redundant DNA sequences that would encode the BRCA1 protein. They also arguably cover genomic DNA coding for the protein, i.e., the isolated gene as it exists in the human body. Assuming the courts gives the term "isolated" an expansive reading, the claims arguably cover very large recombinant molecules that comprise the BRCA coding sequence (i.e., large DNA vectors), as well as recombinant cells or organisms comprising the coding sequence in a non-native context.

Although many have assumed that claims directed towards DNA encoding full-length BRCA proteins can be used to block genetic diagnostic testing, in fact these claims would not appear to be infringed by standard genetic diagnostic testing. In conventional BRCA diagnostic testing, segments of the BRCA gene are amplified and sequenced, not the entire full-length genomic sequence. As a consequence, the full-length coding sequence is not made or used, and thus there would seem to be no infringement (see Holman/Cook-Deegan amici brief). The claim would however appear to cover a recombinant DNA molecule encoding a full-length BRCA protein, which could be used in drug discovery research or, potentially, gene-based therapy.

There are also a variety of narrower isolated DNA claims at issue. For example, some of the isolated DNA claims are limited to a specific cDNA molecule encoding a BRCA protein. These claims are relatively narrow, and could be readily circumvented in many instances by simply using a different DNA sequence that encodes the same protein (which is trivial given the redundancy of the genetic code and the general availability of whole gene synthesis). Although the district court found the cDNA claims patent ineligible, the government argues in its amicus brief that while isolated genomic DNA is patent ineligible, cDNA is patent eligible because it does not exist in nature (cDNA is essentially a DNA version of a naturally existing RNA molecule).

Some of the claims recite any isolated DNA having least 15 nucleotides of a DNA sequence encoding full-length BRCA protein. These claims appear on their face to be extremely broad, encompassing conventional BRCA genetic diagnostic testing (because the testing involves amplification of fragments of the BRCA gene exceeding 15 nucleotides in length). However, a recent bioinformatics study suggests that the breadth of these claims could render them invalid based on anticipation by genetic sequences published in GenBank more than one year before the patents’ filing dates (see Holman/Cook-Deegan amici brief).

Method of diagnosis claims
The method of diagnosis claims are generally directed towards processes of "analyzing" a person's BRCA gene sequence for the existence of a naturally occurring mutation or genetic variation, or “comparing” two BRCA sequences in order to identify the existence of a variation. Critics of gene patents have argued that these method claims broadly cover the mere mental process of analyzing or comparing genetic sequence information, and this was the interpretation of the district court in its summary judgment opinion. However, in their appellant brief Myriad argues for a narrower interpretation, explaining that when the claims are read in light of the specification it is clear that the claims require "extracting, processing and analyzing” DNA molecules, not merely an analysis or comparison of genetic information.

The scope of these method claims could be significant. For example, in the not-too-distant future it is predicted that many people will have their whole genome sequenced. With this information in hand, one could then look for medically significant genetic variations, such as variations in the BRCA gene. Under the district court's broad interpretation, the method claims might be infringed by a doctor or patient merely analyzing the genetic information, which is one of the fears voiced by critics of gene patents. On the other hand, under Myriad's proposed interpretation merely analyzing genetic information would not constitute infringement.

Cell-based assay methods
One of the challenged patent claims recite a method of using recombinant cells engineered to express a BRCA protein to screen for potential cancer drugs. This claim clearly has no implications for genetic testing, and it is unclear why the ACLU chose to challenge its patent eligibility.

Monday, November 1, 2010

Links to Holman/Cook-Deegan and DOJ Amicus Briefs Are Now Operational (I think)

On Friday I thought I posted links to amicus briefs filed by Holman/Cook-Deegan and DOJ in AMP v. PTO, but turns out they were not operational, I believe the problem has been corrected. Just in case, the link to Holman/Cook-Deegan brief is here. Please let me know if still not working (they work on my computer).

Friday, October 29, 2010

Two Amicus Briefs Filed in ACLU Gene Patent Case: Holman/Cook-Deegan and DOJ

Robert Cook-Deegan and I just submitted an amicus brief in AMP v. PTO, the ACLU challenge to Myriad's gene patents, available here. The brief was filed on behalf of neither party, but we ask the court to reverse the district court on its determination that the claims at issue are patent ineligible.

Robert Cook-Deegan directs the Center for Public Genomics, Duke University, which conducted eight case studies of the impact of patenting and licensing on genetic testing for ten clinical conditions commissioned by the US Secretary’s Advisory Committee for Genetics, Health and Society, published in April 2010.

A brief submitted by the DOJ on behalf of neither party is available here.

DOJ asks the court to reverse the district court’s invalidation of the composition claims that are limited to cDNAs and similar man-made constructs, but affirm the district court’s conclusion that the claims encompassing isolated human genomic DNA are invalid.

Here are Summaries of the Arguments:

In their zeal to address perceived public policy concerns associated with
Myriad’s gene patents, and more particularly Myriad’s controversial business and
patent enforcement practices, plaintiffs have invoked the recently re-invigorated
patent eligibility doctrine in a manner that threatens to wreak substantial collateral
damage on future innovation in genetic diagnostic testing, personalized medicine,
and biotechnology in general. DNA patents have created incentives critical in
attracting the substantial investment necessary to fuel the discovery and
development of life-saving products produced by the biotechnology industry.
Although plaintiffs have identified numerous potential concerns with gene patents
in the context of some types of genetic diagnostic testing, to date there is
insufficient evidence that harms attributable to patents on genes justify broad,
subject matter-based invalidation of all patents made of or based on DNA. More
appropriate and targeted legal and policy solutions to problems associated with
some gene patents and patent enforcement practices are preferable to the blunt
doctrinal instrument of patent eligibility. The decision below should be reversed in
order to prevent substantial unintended negative consequences for innovation in
this increasingly important technology sector, and to enable adjudication of
patentability using other tools that are more appropriate to the task

Section 101 marks the “threshold” of the patent system. Bilski v.
Kappos, 130 S.Ct. 3218, 3225 (2010). It not only “defines the subject
matter that may be patented,” ibid., but simultaneously defines what
must remain in “‘the storehouse of knowledge of all men * * * free to all
men and reserved exclusively to none,’” ibid. (quoting Funk Brothers
Seed Co. v. Kalo Inoculant Co., 333 U.S. 127, 130 (1948)); see Bonito
Boats, Inc. v. Thundercraft Boats, Inc., 489 U.S. 141, 151 (1989) (the
patent laws “determine not only what is protected, but also what is free
for all to use”). The boundary between eligible and non-eligible subject
matter is defined, in significant part, by the settled principle that the
patent laws do not embrace laws of nature, physical phenomena, or
abstract ideas. See Bilski, 130 S.Ct. at 3225.
In attempting to apply that principle here, the district court
erroneously cast doubt on the patent-eligibility of a broad range of manmade
compositions of matter whose value derives from the informationencoding
capacity of DNA. Such compositions — e.g., cDNAs, vectors,
recombinant plasmids, and chimeric proteins, as well as countless
industrial products, such as vaccines and genetically modified crops,
created with the aid of such molecules — are in every meaningful sense
the fruits of human ingenuity and thus qualify as “‘human-made
inventions’” eligible for patent protection under section 101. J.E.M. Ag
Supply, Inc. v. Pioneer Hi-Bred Int’l, Inc., 534 U.S. 124, 130 (2001)
(quoting Diamond v. Chakrabarty, 447 U.S. 303, 313 (1980)). The
district court therefore erred in invalidating the challenged composition
claims, such as claim 2 of the ‘282 patent, that are directed solely to
The district court correctly held, however, that genomic DNA that
has merely been isolated from the human body, without further
alteration or manipulation, is not patent-eligible. Unlike the
genetically engineered microorganism in Chakrabarty, the unique
chain of chemical base pairs that induces a human cell to express a
BRCA protein is not a “human-made invention.” Nor is the fact that
particular natural mutations in that unique chain increase a woman’s
chance of contracting breast or ovarian cancer. Indeed, the relationship
between a naturally occurring nucleotide sequence and the molecule it
expresses in a human cell — that is, the relationship between genotype
and phenotype — is simply a law of nature. The chemical structure of
native human genes is a product of nature, and it is no less a product of
nature when that structure is “isolated” from its natural environment
than are cotton fibers that have been separated from cotton seeds or
coal that has been extracted from the earth.
The scope of Section 101 is purposefully wide and its threshold is
not difficult to cross. See Bilski, 130 S.Ct. at 3225. New and useful
methods of identifying, isolating, extracting, or using genes and genetic
information may be patented (subject to the prohibition against
patenting abstract ideas), as may nearly any man-made transformation
or manipulation of the raw materials of the genome, such as cDNAs.
Thus, the patent laws embrace gene replacement therapies, engineered
biologic drugs, methods of modifying the properties of plants or
generating biofuels, and similar advanced applications of biotechnology.
Crossing the threshold of section 101, however, requires something
more than identifying and isolating what has always existed in nature,
no matter how difficult or useful that discovery may be.

Tuesday, October 26, 2010

The ACLU Gene Patent Challenge: Analysis of Myriad's Brief

Yesterday I posted a link to the brief filed last Friday by Myriad Genetics with the Federal Circuit in the case of AMP v. PTO (ACLU and Public Patent Foundation challenge to gene patents). In their brief, Myriad argues that (1) the district court should have dismissed the case for lack of jurisdiction, because there was no evidence of any real or immediate dispute between Myriad in any plaintiff; (2) composition of matter claims directed towards isolated DNA molecules are patent eligible, in part because they have "markedly different characteristics" than their naturally occurring counterparts; and (3) method of diagnosis claims are patent eligible, in part because they all require physical transformation of a DNA sample (i.e., isolation, processing and analysis) that satisfies the Bilski machine-or-transformation test.

Lack of Jurisdiction

Myriad argues that the 20 plaintiffs were all recruited by the ACLU and Public Patent Foundation (referred to herein collectively as ACLU) in order to further those group’s agenda of abolishing gene patents, and that there is no evidence of any current dispute between Myriad and any of the plaintiffs. According to Myriad, only three of the plaintiffs have ever been contacted by Myriad, and those three were contacted more than a decade before the filing of the complaint. Myriad argues that none of the plaintiffs has a controversy "of sufficient immediacy and reality to warrant the issuance of a declaratory judgment" under the MedImmune standard.

Composition of Matter Claims

Myriad begins by arguing that the district court was incorrect in its conclusion that "products of nature" are patent ineligible, pointing out that such a "sweeping exception" would bar the patenting of important inventions like pharmaceuticals derived from natural sources (e.g., Taxol), and would be inconsistent with cases finding isolated natural products patentable that date back nearly 100 years.

Myriad goes on to argue that even if the district court is correct, and isolated products of nature are only patent eligible if they display "markedly different characteristics" compared to the product as exists in nature, Myriad's isolated BRCA gene sequences clearly satisfy the test. Myriad points out that it is inconsistent for plaintiffs to argue on the one hand that the claimed isolated BRCA molecules are necessary for performing BRCA diagnostic testing, while at the same time arguing that the claimed isolated molecules have no "markedly different characteristics" from their naturally occurring counterparts. They note that native DNA is useless for the diagnostic and detection applications for which the isolated molecules must be utilized.

Myriad addresses the much discussed dissent by Judge Dyk in his separate opinion in Intervet, which many have interpreted as signaling his skepticism with respect to the patent eligibility of isolated gene sequences. In his opinion, Judge Dyk suggested that it "would be difficult to argue, for instance, that one could patent a leave of a plant merely because the leaves do not occur in nature in an isolated form." Myriad points out that not only would Judge Dyk’s hypothetical leave likely fail under sections 102 and 103 for lack of novelty and obviousness, it would fail the "markedly different characteristics" test "because the plucked leaf would have exactly the same properties as the unplucked leaf-unlike here, where isolated DNA molecules possess significantly different structural and functional characteristics from native DNA."

Method Claims

Critics of Myriad's gene patents have complained that the claims are much too broad, claiming the information content of the BRCA genes, and totally precluding doctors and patients from looking at or discussing the presence of genetic variations in the BRCA genes. For example, some of the challenged method claims could be interpreted broadly to cover comparing a patient's BRCA sequence with wild-type BRCA in order to identify genetic variations (which could be clinically significant).
However, in their brief Myriad argues for a narrower interpretation, under which all the method claims could only be infringed by a party that isolates and analyzes DNA molecules (as opposed to genetic sequence information). This could be a significant limitation that avoids some of the fears that have been raised with respect to the inability to design around these patent claims. For example, I am reading a book called The $1000 Genome, by Kevin Davies, which describes new technology that is rapidly coming online which will enable people to inexpensively determine the sequence of their entire genome. Salzberg and Pertea recently published an article in Genome Biology called "Do-It-Yourself Genetic Testing,” (described recently on the Patent Docs blog), which purports to allow a person who knows her BRCA sequences to analyze for variations predictive of cancer. It seems to me that in the not-too-distant future method claims such as Myriad’s could be readily circumvented by having one's whole genome sequenced (without asking the sequencer to analyze for BRCA mutations), and then using a computational screen such as the one provided by Salzberg and Pertea to analyze for clinically relevant variations.

Myriad argues that because its method claims all require isolation and processing of DNA molecules, they satisfy the transformation prong of Bilski's machine or transformation test. They correctly note that the Federal Circuit's Prometheus decision explicitly held that diagnostic methods that involve processing a biological sample "necessarily involve a transformation,” and thus pass the machine or transformation test. The Supreme Court in Bilski v. Kappos clarified that the machine or transformation test is not the exclusive test for patent eligibility, but acknowledged that in most instances the existence of a machine or transformation is highly relevant to the question of patent eligibility. The Supreme Court vacated the Federal Circuit's original Prometheus decision, and remanded for the Federal Circuit to reconsider in light of Bilski v. Kappos. I predict that when the Federal Circuit issues a new decision in Prometheus, it will maintain the holding that diagnostic methods involving physical manipulation of biological samples are transformative and generally patent eligible.

Myriad goes on to point out that Bilski v. Kappos "remove any suggestion that the rigid 'machine-or-transformation' test provides the exclusive test for patent eligibility, particularly as applied to "Information Age" technologies like the advanced diagnostic techniques claimed injury and patents." In contrast with the method of hedging claims found to be unpatentable abstract ideas in Bilski, Myriad argues that its methods represent "very real ways of diagnosing and treating cancers."
Interestingly, Myriad does not address the question of whether its method claims impermissibly encompass a natural phenomenon. Supreme Court precedent makes clear that natural phenomena are patent ineligible, and prior to the Federal Circuit's In re Bilski decision most patent eligibility challenges to biological method claims (particularly diagnostic claims) focused on arguments that the method claim preempted the practical use of a natural phenomenon. I have discussed this often on this blog, particularly with respect to the Prometheus and Classen cases, and Justice Breyer's often cited dissent from LabCorp.


I don't have a good sense as to how the Federal Circuit will decide the issue of jurisdiction, it seems to me that the threat of being sued for patent infringement was nonexistent for many of the plaintiffs, and highly speculative for others.

However, I do think that if the Federal Circuit finds jurisdiction it will uphold the validity of isolated polynucleotide claims and methods of diagnosis claims, so long as those method claims require the performance of physically transformative processes (isolation, processing and analysis of molecules).

On the other hand, a broad method claim that could be infringed by the mere analysis of genetic information, without requiring any physically transformative step, would I think face a significant threat of invalidation for patent eligibility under Bilski. I suspect this is why Myriad is arguing for a narrower interpretation.

Monday, October 25, 2010

Myriad Files Brief in AMP v. PTO

Myriad filed its brief with the Federal Circuit last Friday in its appeal of the district court decision in AMP v. PTO, which found Myriad's gene patents to be patent ineligible. The brief is available here, I will provide some commentary tomorrow.

Sunday, October 3, 2010

Yeda Appeals Board Decision Favoring Abbott in TBP-II Interference

Abbott and the Yeda Research and Development Co. have been involved in a patent interference since 1996 over the right to patent TNF-alpha Binding Protein II (TBP-II, also known as TNF receptor 2). This is a therapeutically and commercially relevant protein. ENBREL (etanercept), a biologic drug developed and marketed by Immunex (now Amgen) for the treatment of autoimmune disease, is a fusion protein comprising the Fc component of human immunoglobulin G1 and TBP-II. I don't know what the specific commercial implications of the patent interference are, but clearly the parties view the patent as sufficiently valuable to justify the cost of a prolonged interference contest. A corresponding patent application claiming the genetic sequence encoding TBP-II was the subject of an important 2004 Federal Circuit decision relating to the written description requirement (In re Wallach).

The interference involves a patent issued to Abbott (5,344,915) and Yeda’s US patent application. The inventor on the Yeda application is Wallach (claiming common priority to the application at issue in In re Wallach). The interference was declared by the patent office in 1996, and resulted in a "Final Decision" by the Board of Patent Appeals and Interferences (the Board). In this decision, the Board invalidated Abbott's patent as anticipated by a prior art reference. Abbott claimed priority to two German applications that would have predated the invalidating prior art reference, but the Board denied Abbott the benefit of these two earlier filing dates. The Board essentially held that the German applications did not satisfy the written description requirement of section 112 with respect to the claimed invention.

The German applications disclosed an isolated protein, defined in terms of its approximate molecular weight and an incomplete N-terminal sequence. The applications also disclosed the source of the protein (the urine of patients with fever) and the method used to purify the protein. Abbott's US patent claims the protein as a "purified and isolated TNF-alpha binding protein which has a molecular weight of about 42,000 daltons and has at the end terminus the amino acid sequence [the specific amino acid sequence of the N-terminus of TBP-II]” The critical difference between the patent claims and the disclosure in the German applications is the N-terminal amino acid sequence. In the German applications, there was some uncertainty about the exact N-terminal sequence, and some amino acids in the sequence were not disclosed. These ambiguities were addressed and resolved in the subsequent filed US patent.

The Board’s decision was based on its determination that the earlier incomplete disclosure failed to adequately describe the more complete amino acid sequence provided in the ultimately issued claims. Abbott argued that the proteins described in the US patent claims and the German applications are the same, as demonstrated by the fact that they have the same molecular weight, bind specifically to TNF-alpha, are isolated from the same source (urine from patients with fever), and isolated and purified using the same methods. In essence, Abbott argued that the German applications inherently disclosed the later claimed protein. However, the Board rejected this argument, finding that the record failed to establish to the necessary degree of certainty that TBP-II was inherently disclosed in the German applications, and noting that "inherency may not be established by probabilities or possibilities."

Abbott appealed the Board decision under 35 USC 146 to the District Court for the District of Columbia. In 2008 the District Court granted Abbott's motion for summary judgment, vacated the Board's final decision and remanded to the Board for further proceedings. The District Court was obliged to review the Board's decision under the "clearly erroneous" standard, but even under this deferential standard of review the court reversed the Board, holding that the German applications inherently disclosed the later claimed protein to a degree of certainty sufficient to satisfy the written description requirement. The court rejected alleged ambiguities raised by Yeda (and accepted by the Board) with respect to the identity of the protein disclosed in the German applications.

Yeda appealed the district court decision to the Federal Circuit, but in 2009 the court dismissed the appeal, holding that they did not have jurisdiction to hear the case prior to a final decision in the interference.

On remand, the Board granted Abbott priority benefit of the German applications, and in May 2010 awarded judgment in the interference to Abbott.

On September 9, 2010, Yeda filed its appeal under 35 USC 146, again in the District of Columbia, challenging the Board's decision. It seems likely the case will continue on for at least several more years, since the District Court's decision will begin be subject to appeal to the Federal Circuit (which should have jurisdiction to hear the case since a final decision will have been rendered in interference).

It interesting to compare this case with Goeddel v. Sugano, a September 7 Federal Circuit decision regarding a patent interference over mature human fibroblast interferon (beta interferon). At issue were claims directed to the mature form of the protein, and the DNA encoding the mature form of the protein.

The mature form of fibroblast interferon is produced by cleaving a 21 amino acid presequence from the N-terminal of precursor fibroblast interferon. Sugano attempted to claim priority to a Japanese application that disclosed the full length precursor protein, and also disclosed the N-terminal sequence of the mature protein. Based on this disclosure, clearly anyone that knows anything about molecular biology would know the sequence of the mature protein - it is simply the portion of the disclosed precursor protein sequence that begins with the N-terminal sequence of mature protein.

However, whoever drafted Sugano’s Japanese patent application (filed more than 30 years ago) apparently did not see the need to point out the obvious, and thus there is no explicit disclosure in the application specifically pointing out the sequence of the mature protein. Because the Japanese application failed to connect the dots, the Federal Circuit held that the Japanese patent application did not satisfy the requirements of section 112 first paragraph (enablement and/or written description), and as a result awarded priority to Goedell.

It seems to me a good argument could be made that the Sugano Japanese application inherently but unambiguously disclosed the sequence of the mature protein. This strict application of the written description requirement is arguably inconsistent with the relatively lenient standard applied by the District Court in reversing the first Board decision in the interference involving Abbott and Yeda.

I say arguably, because even though both interferences involve claims directed to proteins, the facts are quite distinct. In Yeda v. Abbott, the court found that the claimed protein was the same protein disclosed in the priority applications, even though the earlier disclosure provided less detail with respect to the amino acid sequence of the protein. In Goeddel v. Sugano, on the other hand, the priority document provided the full amino acid sequence of the later claimed protein, but failed to explicitly point out the distinction between the mature protein and the precursor protein from which it is derived.

Wednesday, September 29, 2010

Glaxo Fires Latest Shot in Patent Battle with Genentech over Therapeutic Molecule Antibodies

On September 20, Glaxo Smith Kline (Glaxo) sued Genentech for infringement of US patent numbers RE40,070 and RE41,555 in the District of Delaware. The patents claim methods of purifying monomeric IgG antibody using hydrophobic interaction chromatography. GSK alleges in its complaint that Genentech's production of therapeutic monoclonal antibody products, specifically Herceptin, infringes the patents. Herceptin is an important monoclonal antibody therapeutic used to treat HER2 positive breast cancer. The complaint is available here.

The lawsuit is the latest volley in an ongoing patent war between GSK and Genentech over recombinant monoclonal antibody technology. Most of the patents that have been asserted are directed towards general methodologies used in the production and formulation of monoclonal antibody therapeutics.

For example, in 1999 Glaxo sued Genentech for allegedly infringing US patent numbers 5,545,403 and 5,545,405 (both generally claiming methods of treatment using recombinant antibody glycosylated by Chinese hamster ovary (CHO) cells), and US patent numbers 5,654,403 and 5,792,838 (both directed towards methods of stabilizing antibody formulations using copper chelators). Glaxo alleged that the patents were infringed by Genentech's production of Herceptin and Rituxan, two cancer drugs that target breast cancer whose tumors over express the HER2 protein and CD20 positive, B-cell non-Hodgkin's lymphoma, respectively. A district court held for Genentech, finding all of the asserted Glaxo claims invalid and/or not infringed. The case settled while on appeal to the Federal Circuit.

In 2000, Glaxo sued Genentech for allegedly infringing US patent number 5,633,162, which claims methods of growing CHO cells in a serum-free medium. While the complaint does not specifically identify the allegedly infringing Genentech products, Genentech reported in a filing to SEC that it assumed that the lawsuit was directed towards Herceptin and Rituxan. This case settled early in the litigation.

In 2009, Glaxo received FDA approval to market Arzerra, an anti-CD20 approved for the same indication as Genentech's Rituxan (chronic lymphocytic leukemia). Thus, the two products can be viewed as direct market competitors, perhaps biosimilars (we are still awaiting guidance from FDA as to criteria for biosimilarity under the new biosimilarity legislation passed by Congress as part of healthcare reform).

In 2009, Glaxo filed a declaratory judgment action in the Southern District of Florida seeking a judgment of invalidity, unenforceability and noninfringement with respect to Genentech's patent number 6,331,415. This patent is widely referred to as the “Cabilly II” patent, and came out of a highly publicized interference between Genentech and Celltech. The patent broadly covers the co-expression of immunoglobulin heavy and light genes in a single host cell. Genentech has been quoted as stating that the patent is "the backbone of recombinant antibody production in the biotech industry." Genentech has already asserted the patent against MedImmune and Centocor, alleging that those companies’ production of monoclonal antibodies products Synagis, ReoPro and Remicade infringe the patent. The litigation between Genentech and MedImmune resulted in a Supreme Court decision on standing of a patent licensee to bring a declaratory judgment action action challenging the patent. Genentech successfully moved to have the litigation transfer to the Central District of California, where the Centocor and MedImmune litigations are located.

In 2010, Genentech (along with Biogen Idec, its partner in the marketing of Rituxan) sued Glaxo in the Southern District of California for allegedly infringing US patent number 7,682,612. The patent claims methods of treating chronic lymphocytic leukemia using an anti-CD20 antibody, and thus invalid would appear on its face to be infringed by Arzerra. This case is somewhat noteworthy, since it is the only one in which the patent is specifically directed towards an antibody recognizing a specific antigen (CD20), as opposed to the other patents all relating to general methods of antibody production and formulation.

Tuesday, September 28, 2010

Jury Finds Scruggs Willfully Infringed Monsanto Patents, Award at Least $9 Million in Damages

10 years ago Monsanto sued farmers Mitchell and Eddie Scruggs (collectively referred to as "Scruggs") for infringing patents relating to genetically modified Roundup Ready soybean seeds and Bollgard-containing cotton seeds. The case has resulted in numerous judicial decisions, including three Federal Circuit decisions, most notably Monsanto v. Scruggs, 459 F.3d 1328 (Fed. Cir. 2006). Scruggs raised numerous arguments against Monsanto, including first sale/patent exhaustion, implied license, lack of sufficient written description and enablement, as well as allegations of patent misuse and antitrust violations.

Nonetheless, Monsanto has prevailed, and on September 21 a jury in the Northern District of Mississippi found that Monsanto is entitled to $2.6 million for Scruggs unauthorized planting of saved crop seeds contain the patented genes, and $6.3 million for their sale of saved soybean seeds containing the Roundup Ready gene (a practice referred to as "brown bagging"). Not only that, the jury found that the infringement was willful, which means that the court can enhance the damages up to triple the amount found by the jury, or close to $30 million. Based on history, I would not be surprised if Scruggs continues to appeal, perhaps stretching this out a few more years.

District Court Denies Novozymes Motion for Preliminary Injunction in Battle for Control of the Alpha-Amylase Market

On September 24, 2010, a federal district court in the Western District of Wisconsin denied a motion for preliminary injunction in a patent lawsuit between two of the leading companies in the field of genetically engineered industrial enzymes, Novozymes and Genencor (a division of Danisco). The two companies are essentially the only competitors in the market for alpha-amylases, used to convert corn into fuel ethanol. According to the order (available here) denying preliminary injunction, Genencor dominated the alpha-amylase market until 1999, when Novozymes entered the market with Liquozyme, a genetically engineered thermostable alpha-amylase. At one point, Liquozyme accounted for more than 80% of the alpha-amylase sold in the fuel ethanol market. In 2008, Genencor responded by introducing its own genetically engineered thermostable alpha-amylases (called GC358). Since then, Novozymes market share has dropped approximately 60%.

The patent being asserted by Novozymes, US patent number 7,713,723 (the ‘723 patent) claims priority to a patent application originally filed in 2000. The original patent application was broadly directed toward alpha-amylase mutants having increased stability at “high temperature and/or low pH conditions, in particular at low calcium concentrations.” However, the claims that ultimately issued in the ‘723 patent, which recite thermostable alpha-amylase variants comprising "a substitution of serine at position 239 relative to the parent alpha-amylase," were not added until 2009, apparently in response to Genencor's marketing of GC358, which presumably includes this mutation. Prior to the introduction of these claims in 2009, the patent application contained no disclosure specifically directed towards substitution of serine at position 239.

Importantly, the '723 patent does not cover Liquozyme, over for that matter any Novozymes product. Liquozyme presumably possesses improved thermostability over the wild type enzyme as a result of a different mutation or set of mutations.

In deciding a motion for preliminary injunction, courts essentially consider four factors: (1) the likelihood the patent owner will ultimately prevail on the merits in the case (i.e., the patent will be found infringed and not invalid); (2) the extent to which the patent owner will experience "irreparable harm" if the infringement is not stopped immediately; (3) the balance of the interests of patent owner an accused infringer; (4) and the public interest. In this case, the district court found that all these factors weighed in favor of Genencor, and denied Novozymes’ motion for preliminary injunction.

Irreparable Harm
The district court decision began by addressing the issue of irreparable harm, the factor which the court found likely to be dispositive by itself. Novozymes argued that it would suffer irreparable harm in the form of diminished reputation, loss of market share and price erosion if Genencor's product were allowed to remain on the market during the course of the litigation. However, the court pointed out that Genencor had entered the market with GC358 in 2008, but that the patent did not issue until May 2010, so that most of Novozymes’ loss in market share occurred when it was perfectly legal for Genencor's product to be on the market. The court surmised that most of the customers who would be inclined to switch from Liquozyme to GC358 had probably already done so, and thus most of the alleged harm identified by Novozymes had already occurred prior to issuance of the patent.

In assessing irreparable harm, the district court found the fact that the patent does not cover any product marketed by Novozymes to be quite significant, pointing out that while Novozymes’ argument that sale of GC358 would harm its reputation "could be persuasive if plaintiffs actually sold a product that practiced the ‘723 patent[,] it is difficult for plaintiffs to argue that their good reputation is contingent on their ability ‘to uniquely make and sell [their] patented alpha-amylase’ when they are not even using the claim technology themselves and identify no plans to do so."

The district court went on to speculate that if "plaintiffs are not trying to protect their own right to provide consumers a product that embodies the patent, then it may be that the patent is nothing more than weapon to prevent defendants from competing with them." In other words, the court seems put off by the fact that Novozymes is not practicing the technology claimed in their patent, treating them as something akin to a non-practicing entity (sometimes referred to as a patent troll).

Likelihood of Success on the Merits
The court also expressed doubt with respect to the likelihood that Novozymes would succeed on the merits of the case, based on the court's determination that there was a substantial likelihood that the asserted patent claims are invalid for lack of enablement and lack of sufficient written description. Both the enablement and written description issues arise out of the failure of the originally filed patent application to sufficiently point out and identify the later claimed substitution of serine at position 239. The patent application identified 33 amino acid positions (including position 239) in alpha-amylase which allegedly can be modified by amino acid substitution, deletion or insertion to result in some increased stability under high temperature and/or low pH conditions. It does not, however, particularly point out position 239, nor does it particularly point out substitution with serine (one of 20 amino acids commonly found in proteins), and it does not correlate the specific mutation particularly with thermostability. Genencor pointed out that the patent specification identified 8.589 x 1042 possible amino acid substitutions, deletions and insertions, and now Novozymes was attempting to claim one of those possibilities that had subsequently been shown by Genencor to provide thermostability.

This case illustrates the difficulty the inventors of genetically engineered proteins face when attempting to obtain adequate patent protection for their inventions. Protein sequence space is vast, and in most cases in which a sequence variant of a protein has been found to have some desirable functional characteristic, there are an astronomical number of alternate sequence modifications that will result in the same functional outcome. I have described this problem in a law review article (available here), and more recently in my amicus brief filed in Ariad v. Lilly (available here). Essentially, the written description and enablement requirements limit the ability of protein engineers to claim protein variants in functional terms, resulting in claims limited to enzyme sharing some degree of structural similarity to variants explicitly disclosed in the patent application. This allows competitors to design around the patent by screening for alternate structural modifications to the amino acid sequence that result in the desired function, as exemplified Genencor’s product that apparently shares the desirable functional characteristics of Novozyme’s product but employs a different structural modification to the amino acid sequence.

Balance of Harms
The court found that the balance of harms in this case favored Genencor, finding that if Genencor were forced off the market the company might "very will be crippled and unable to recover even if the injunction is lifted after the trial." On the other hand, the court found that Novozymes was essentially "asking for assistance in maintaining a monopoly at least until the end of trial."

Public Interest
The fact that Novozymes is not marketing a product covered by the asserted patent also counted against them in the court’s analysis of the effect of an injunction on the public interest. The court found that if Genencor's GC358 were taken off the market customers would be deprived of the patented invention entirely for the course of the litigation proceedings, since Novozymes is not marketing any product covered by the patent. The court found it "somewhat inconsistent for plaintiffs to be arguing on one hand that the ‘723 patent represents an important new invention and then argue on the other hand it should make no difference if no one is allowed to actually use it.”

Thursday, September 9, 2010

BIO Files Amicus Brief Supporting Eli Lilly Petition for En Banc Rehearing of Sun . Lilly

Yesterday the Biotechnology Industry Organization filed an amicus brief supporting Eli Lilly in its petition for en banc rehearing of Sun v. Lilly, a recent Federal Circuit decision pertaining to the doctrine of obviousness-type double patenting, and discussed in a previous post. I participated in the drafting of the BIO amicus brief, a copy of which can be found here.

Thursday, September 2, 2010

Sun Pharmaceuticals v. Eli Lilly: The Creeping Expansion of the Doctrine of Obviousness-Type Double Patenting

In Sun v. Eli Lilly, a panel of the Federal Circuit held a Lilly patent claiming use of the drug gemcitabine (GEMZAR) for the treatment of cancer invalid for obviousness-type double patenting, in view of an earlier Lilly patent claiming the drug per se and its originally discovered use as an antiviral agent. The invalidated patent was set to expire on November 7, 2012, 2.5 years after the expiration of the patent claiming the drug active ingredient (May 15, 2010), so this could result in a substantial reduction in Lilly’s period of marketing exclusivity. Lilly is currently seeking en banc reconsideration by the Federal Circuit.

Sun represents the most recent in a series of Federal Circuit decisions expanding the scope of the judge made doctrine of obviousness type double patenting. In this article, I briefly summarize the history of the expansion, and outline why I think the court should take this opportunity to reconsider the implications of these cases en banc.

Eli Lilly's patent prosecution decisions that resulted in a finding of obviousness type double patenting

It is informative to review the facts of the case. In the early 1980s, a Lilly scientist invented a method for synthesizing a genus of chemical compounds including gemcitabine, something others had previously attempted but failed to accomplish. He also showed that the compound had antiviral activity. Lilly filed a patent application disclosing the genus of chemical compounds and their use as antiviral agents.

The original inventor of the compound, along with a second Lilly inventor, then discovered that gemcitabine has anticancer activities, so these inventors filed a second application claiming methods of using the drug to treat cancer. This application did not claim priority to the first application.

On the same day Lilly filed the second application, it also filed a CIP of the first application. In retrospect, this was a critical mistake. The CIP was apparently filed to expand upon the definition of the disclosed genus of compounds, basically by adding to it species wherein certain R groups could be hydrogens. Significantly, this added disclosure was not necessary in order to patent gemcitabine, since gemcitabine was disclosed in the originally filed patent application.

At that time, there was some uncertainty with respect to the extent to which it was necessary to update the best mode when filing a CIP application (this uncertainty was ultimately addressed in 1994 in Transco). In any event, presumably in an attempt to ensure compliance with the best mode requirement, Lilly also added disclosure of the anticancer properties of the chemical compounds in the CIP.

Ultimately, the CIP application resulted in a patent claiming the drug active ingredient as a composition of matter, and also methods of using it as an antiviral agent. Later, the second patent application issued with claims to methods of using the drug to treat cancer. In Sun v. Lilly, it was the disclosure of anticancer activity added to the first patent by means of the CIP application which was used as the basis for invalidating the second patent claiming that use. In the next section, I explain why this represents a significant expansion of the doctrine of obviousness type double patenting.

The creeping expansion of obviousness type double patenting

The Federal Circuit has established a two-step process for analyzing obviousness type double patenting. First, the court construe the claims in the earlier patent and the claims in the later patent, and determines the differences. Second, it determines whether those differences render the claims patentably distinct.

One fundamental distinction between analysis for obviousness under section 103 and obviousness type double patenting is that when analyzing for double patenting only the claims of the two patents are to be considered. As the Federal Circuit stated in 1992 in General Foods v. Studiengesellschaft Kohle, its "precedent makes clear that the disclosure of a patent cited in support of a double patenting rejection cannot be used as though it were prior art.”

However, in the 2002 decision of Geneva v. GSK, the Federal Circuit began chipping away at this prohibition against using the written description of the earlier patent to invalidate the second patent, and embarked upon what has become a creeping expansion of the doctrine. In that case, the earlier patent claimed a drug active ingredient. The written description portion of the earlier patent also describes a method of using the drug to treat a disease, but significantly, this use of the drug is never mentioned in the patent claims. Nonetheless, the panel used this disclosure in the written description of the first patent to invalidate a second patent claiming that method of use.

This seems to fly in the face of General Foods and the controlling case law. However, the Geneva panel justified its decision to go outside the claims of the first patent in its analysis by reasoning that the claim in the earlier patent "is drawn to a compound having a certain physical property," and thus, "[s[tanding alone, that claim does not adequately disclose the patentable bounds of the invention. Therefore, this court examines the specifications of those patents to ascertain any overlap in the claim scope for the double patenting comparison."

Although the Geneva panel states that it needed to refer to the written description of the first patent to determine the extent of overlap between the claims in the two patents, this explanation does not appear to survive close scrutiny. The claims in the first patent broadly recite chemical compound, with absolutely no limitation respecting the use of the compound. It is black letter law that a patent claiming a chemical compound dominates all uses of the compound as defined by the claims, including methods of using the compound that are not disclosed in the patent specification, and methods that were nonobvious and not even contemplated at the time the patent was filed. While it is undoubtedly appropriate to consult the entire specification in construing the scope of a patent claim, it is unnecessary to consult the specification for methods of using a compound when ascertaining the scope of a patent claim directed to the compound as a composition of matter.

Although the Geneva panel’s use of disclosure in the written description as a basis for invalidating the second patent is questionable, at least the panel provided a policy justification for his decision. The panel correctly noted that in order for chemical compound to be patentable, the patent applicant must disclose a utility for the compound. Since the earlier patent only identified a single utility for the claimed drug, the disclosure of that use was necessary for the patentability of the chemical compound. Thus, while it was incorrect for the panel to suggest that the disclosure of the method of use in the written description was relevant to the question of claim scope, it was clearly relevant to the question of patentability.

The Geneva panel stressed that since the earlier patent disclosed only a single utility of the compound, the claims of the second patent reciting nothing more than that disclosed utility as a method was not patentably distinct. In essence, the panel's decision can be rationalized as a policy of not allowing a second patent on a method of use if that method of use was critical to the patentability of the first patent claiming the molecule per se.

Five years later, the Federal Circuit again was faced with an allegation of double patenting involving a first patent claiming a drug active ingredient and a second patent claiming methods of using the drug active ingredient. In Pfizer v. Teva, the panel once again went beyond the language of the patent claims, and invalidated the second patent because the method of use claimed was disclosed in the written description section of the earlier patent.

Significantly, in Pfizer the first patent disclosed multiple uses of the composition, and the invalidated claims encompassed only some subset of those uses. Thus, the facts of Pfizer are distinguishable from those in Geneva, where the method of use claim was directed to the sole utility disclosed in the earlier patent, and thus presumably necessary takedown to the validity of the first patent. Thus the rationale to justify the Geneva decision, i.e., that the claimed method of use was critical to the patentability of the first patent, does not appear to exist in Pfizer. In effect, Pfizer expanded the rule set forth in Geneva, by holding that the disclosure of multiple uses of a molecule in the first patent renders unpatentable later claims directed towards any of those methods.
The Pfizer decision provides no policy justification for this expansion of the Geneva doctrine, and fails to even acknowledge the expansion. Nonetheless, one might justify it on policy grounds by pointing out that by disclosing multiple methods of use in the original patent application, the inventor was able to stake a claim of priority to these multiple uses by securing an early effective filing date. By choosing to avail itself to this earlier effective filing date, the patent applicant (arguably) should be limited to a single term for all of the patents issuing out of this single priority document.

We then come to the Sun v. Lilly decision, in which the panel further expands the holdings in Geneva and Pfizer by interpreting those decisions as creating a bright line rule that any use disclosed in a patent claiming a drug active ingredient precludes the owner of the patent from obtaining any second patent on one of the disclosed method of use, regardless of whether the claimed method of use was disclosed in the original patent application which provides the effective filing date for the first patent. Unlike the earlier cases, the method of use claimed in the second patent was not disclosed in the originally filed application to which the first patent claims priority.

Note that putative policy justifications for the Geneva and Pfizer decisions do not appear to be present in Sun. Unlike Geneva, the second patent does not claim a method of use that was critical to the patentability of the first patent. The first patent disclosed use of the compound as an antiviral agent, and the patent office issued patent claims directed towards use of the compound as an antiviral, raising a presumption that this is a valid practical utility which would confer patentability on the compound, regardless of the later disclosure of the anticancer activity. Furthermore, unlike the case in Pfizer, the later claimed method of use (anticancer) was not disclosed in the originally filed patent application that led to the first patent, so this first patent application was not available to Lilly to provide an effective priority date for the method of treating anticancer.

In fact, Lilly only ran afoul of obviousness type double patenting because it voluntarily chose to update and supplement the disclosure of the first filed patent application with the disclosure of anticancer activity. They could have avoided the problem by simply prosecuting the originally filed patent application to obtain patent claims covering the drug active ingredient, which it seems clear did not require disclosure of the additional material added by amendment to the CIP. If they wanted to expand the genus of chemical compounds, they could have filed the CIP as a divisional, thereby avoiding the need to add disclosure of the anticancer activity to the gemcitabine composition of matter patent.

In short, if Lilly had a crystal ball and could have foreseen the recent expansion of obviousness type patenting, it could have easily avoided obviousness type double patenting by simply altering its patent filing procedure. But this illustrates how the holding in Sun elevates form over substance.

The decision seems to conflict with 35 USC 103(c)

Congress amended the patent statute by introducing 35 USC 103(c), which basically shields companies from having the inventions of a company’s inventors rendered obvious by prior art created by other inventors at the same company. Congress’s intent was to encourage follow-on innovation and intra-company collaboration.

Originally, 103(c) only included 102(f) and 102(g) prior art, but it was subsequently amended to also include 102(e). In effect, Congress explicitly decided that commonly owned inventions should not be subject to invalidation for obviousness based on subject matter disclosed in a commonly assigned earlier filed patent application. Sun stands congressional intent on its head. By ignoring General Foods, it effectively allows the court to invalidate a patent based on a finding that it claims subject matter that is obvious in view of subject matter disclosed in the specification of an earlier filed commonly assigned patent application. But it goes even farther than that, for in Sun invalidity was based on subject matter added after the effective filing date, and thus not available as 102(e) prior art.

The biotech industry would benefit from en banc review of this issue

A number of issues raised by this and decision that would seem to warrant en banc review. There is clearly a tension between General Foods and Sun and its predecessors. While General Foods states that only the patent claims are to be compared, Geneva opened up the door to using the written description of the earlier patent to invalidate a later patent. In Pfizer and Sun, the door has opened wider and wider. The Federal Circuit should acknowledge this expansion of judge made law and consider whether it is justified by public policy considerations. If the court chooses to go down this route, it should delineate the boundaries of the expanded doctrine.

For one thing, it is unfair to patent owners to change the laws of patentability in a manner that invalidates a valuable patent for patent prosecution decisions made based on the case law at the time. In Festo, the Supreme Court stressed the importance of not altering the law in a way that undercuts the investment backed expectations of the inventive community.

Moving forward, some clarity on the issue would be very helpful. Today's inventors should understand the consequences of adding disclosure to a patent application, in order to rationally decide when to file a patent application, what to include in the disclosure, and when to invest in follow-on research.

For example, based on the Sun decision the wisest course for drug companies would be to file their original composition of matter patent application with a minimal disclosure of methods of using the compound sufficient to satisfy the utility requirement. There is an incentive to prioritize follow-on research on uses not disclosed in the original application, because presumably those uses will not be precluded from being patented separately. Companies might be less inclined to pursue research and development of methods of use disclosed in the earlier patent application, but perhaps that is a reason for the court to rethink the course it has embarked upon.

But who is to say for sure that the doctrine will not be expanded in subsequent decisions to cover methods of use that are not even disclosed in the earlier specification? It would certainly be consistent with the progressive expansion we have seen in going from Geneva to Pfizer to Sun. Uncertainty with respect to the availability of patent protection is a disincentive to future investment in innovation, and it would be nice if the en banc Federal Circuit would clarify the boundaries of the obviousness type double patenting doctrine sooner rather than later.