By Robert Cook-Deegan

Robert Cook-Deegan co-authored an amicus brief in the Myriad case with the author of the preceeding article, Christopher M. Holman.


Professor Holman and I found common ground because we are both concerned about the implications of eliminating all patents on DNA-based inventions. To my eyes, the business model pursued by Myriad Genetics is problematic, and I would like to see many of Myriad's broad claims invalidated and the scope of the patents narrowed considerably-probably more so that Prof. Holman-but I share his concern that the logic of Judge Robert Sweet's March 29, 2010, ruling pushes in the direction of eliminating the possibility of patents on DNA molecules.


Is DNA Patentable Subject Matter or Not?

The intuition that Judge Sweet captured is that DNA is an embodiment of biological information, and one of its key properties is self-replication. This feature of DNA is the reason that Watson and Crick became household names. Judge Sweet recognized DNA's unique role as the storage and transmission medium for genetic information. It does not follow, however, that nothing made of DNA can be patented. DNA can also be engineered, and forms of it can be discovered, even in unaltered form, that have value in altering biology and possibly treating disease. If the higher courts agree with Judge Sweet, then the problems associated with gene patents in diagnostics will indeed be gone, but some good things will be gone with them. Moreover, there is another way to get the same result without the secondary consequences of eliminating all DNA patents.

The Department of Justice amicus curie  brief argues that Judge Sweet has been too broad, and that some DNA molecules should be patentable-those that have been engineered or altered-but that DNA sequences found in nature should not be patentable subject matter. I am sympathetic to this view, but I do worry that it will be difficult to hold that line. Some things that could create social benefit if patentable would be swept away, and I suspect courts will be sympathetic to conferring the patent incentive for such products and services. I am thinking in particular of DNA vaccines, small interfering RNAs and DNAs, and other objects that may be found in nature but might also be molecules of therapeutic value in and of themselves. If so, they will be expensive to develop into biologic therapeutics, in which case the patent incentive would serve the same role it plays in small molecule drugs and protein therapeutics (including some "gene patents" underlying those products).

The problem with Judge Sweeet's logic, and probably even the Department of Justice position, is collateral damage to socially valuable patents underlying therapeutics, and possibly also to future diagnostics that will be expensive to develop. That risk would be much less if traditional doctrines of patent law-such as novelty, obviousness, enablement, and adequate written description-were used to root out destructively broad patent claims on BRCA and other genes relevant to diagnosis, instead of a sweeping rule about whether DNA is patentable subject matter.


"Discoveries" Are Sometimes Patentable in U.S. Law

One argument marshaled against "gene patents" is that they are discoveries and not inventions. Genes associated with disease (or biological function) are indeed discoveries that are not invented by their discoverers, but so are many other patentable things such as antibiotics, vaccines, and hormones which are items found in nature. Social benefit arises from patenting such inventions when they require substantial investment to develop into marketable products or services. Genetic diagnostics have generally not required that kind of investment, but future diagnostics might, and some DNA-based inventions clearly will if they prove therapeutically useful. Examples include DNA vaccines (including perhaps even naked purified DNA derived from natural pathogens), inhibitory RNAs and DNAs, or naturally occurring genes that may eventually become treatments through gene transfer. If so, they will have to be clinically tested in expensive trials, and the patent system is one way (admittedly not the only way) to induce investment in such clinical studies.

Much of the argument about the BRCA patents claiming DNA molecules centers on what it means to "isolate" DNA; but DNA sequencing by definition entails "isolating" a DNA molecule in one way or another by clever laboratory manipulation. So I agree with Judge Sweet's conviction that the words "isolated" or "purified" in patent claims are basically "a lawyer's trick," but I do not agree with his logic that this implies DNA cannot be patented. Manipulation entailed in sequencing by definition entails the hand of man, and it is therefore hard to distinguish this use of DNA from the other valuable molecules found in nature that are now and have long been patentable. I don't think his line will hold in law; I do think, however, that the traditional criteria for patentability (as opposed to patent-eligibility or patentable subject matter) if properly applied would reduce DNA inventions to the scope of what is disclosed in the patent.


Applying Criteria of Patentability Can Reach the Same End

If banning DNA patents were the only way to solve the problems in DNA diagnostics, I might agree with Judge Sweet's ruling and the Department of Justice position. But that is not the only remedy to problematic claims in gene patents that have been granted. Many of the broadest claims in the patents being contested could be found invalid if subject to challenge in court. Before this case, there was no case law in the diagnostic context, and courts could narrow patent claims relevant to genetic testing-particularly from the U.S. Patent and Trademark Office (which has been much more prone to granting broad patent claims than the European Patent Office). The problem is much more about overly broad patent claims than whether patents on DNA exist at all.

My argument is that rather than eliminate DNA as patentable subject matter, it would be better to allow DNA molecules to be patentable, but then apply the doctrines of patent law rigorously to eliminate the broad claims that have been granted. Current claiming practice in gene patents allows claims covering discoveries that take years of hard work, long after the patent is granted, to reduce to practice. Those practices need to change. Case law would address the problems of diagnostic patents by limiting their scope to what "inventors" actually contribute, and not more.

Two examples are directly pertinent to the BRCA case. One is about claims on DNA molecules and the other about methods for detecting DNA sequence changes. Claims 5 and 6 of the original BRCA DNA molecule patent (US patent 5,747,282) claim any 15-base pair segments of DNA that have a sequence that would encode a BRCA peptide. Only a court can decide patent claim validity, but that claim seems likely to fall afoul of two patent criteria: novelty and enablement. What is claimed is not novel because it in effect covers any DNA sequence that encodes any five amino-acid sequence in any protein that is the same as any five-amino acid sequence in BRCA1. It turns out there are a lot of DNA sequences that encode at least five amino acids that can be found in the BRCA1 peptide, and some of them were in GenBank more than a year before the patent application was filed. Moreover, while Myriad et al., did make it possible to determine those sequences, they have not found a use for the vast majority of the millions of 15-mers they claim. If they had restricted the scope of the claim to mutations they demonstrated to confer risk of breast and ovarian cancer, they would have some grounds for having enabled their invention to be useful. Claims 5 and 6 as granted are extremely broad and thus probably invalid, on grounds of novelty and enablement. If they fall, would other claims block BRCA1 diagnosis? The main independent claims (1 and 2) would not be infringed, because no one makes and sequences a full-length cDNA for diagnosis. To the degree that other claims are infringed, they might be vulnerable to the same problems of novelty and/or enablement as claims 5 and 6, unless they were restricted to the very specific mutations disclosed in the patent. And if they were so restricted, they should not confer a monopoly for genetic testing.

Myriad also has method claims that are quite broad. Claim 1 of their main BRCA1 "method" patent is basically a very wordy and convoluted way to claim comparing a sample sequence to a reference sequence and noticing if there is a difference. The utility, however, comes from knowing which alterations are associated with disease. The claim does not restrict itself to useful comparisons described in the patent. Again, if the claim were only on a method for detecting differences that were known to confer risk and disclosed in the patent, it might be specific enough to pass muster, and be enabled by the patent disclosure. But if it were restricted to what was enabled at the time of the patent application, it would not allow Myriad to block others from testing for new mutations, and while Myriad would have earned a right to licensing royalties for the mutations they worked hard to correlate with disease risk, they would not be in position to assert a monopoly for mutations they never discovered or had not yet discovered. Others might well find new mutations, and if they patented them, Myriad would have to license them to do genetic testing. This is how PGxHealth's monopoly on testing for long-QT syndrome was broken, when GeneDx secured countervailing exclusive rights to long-QT mutations. Myriad's U.S. business model would have to change (and the United States is the only jurisdiction in which Myriad has been able to establish dominance as the sole provider of BRCA testing). Their business model is the problem, not their patents, if properly interpreted. The extraordinary breadth of the claims and the cost of pushing back against Myriad's initially aggressive patent enforcement were indeed enabled by their patents, but it is not necessarily to kill all gene patents to rein in the business model. Indeed, ACLU and the Public Patent Foundation may have already achieved much of what they desired by making clear that patent-holders trying to enforce gene patents in diagnostics may well lose, especially if claims are unduly broad.

I truly believe there are serious problems in DNA diagnostics that are associated with patent rights, especially when those rights have been exclusively licensed to a single provider. But if legal doctrines of patentability were properly applied to gene patents in light of how diagnostics work in the real world, monopoly business models would not be possible.


DNA Patents Are Not the Problem, but Unduly Broad Claims

One virtue of gene patents in diagnostics is that they reward inventors and their institutions for some of the value they confer through their discoveries. I believe the Universities of Michigan and Toronto deserve the royalties they have earned from uncovering the cystic fibrosis gene, for example-and other universities and inventors who found mutations later have also earned royalties from CF testing. Moreover, if the CFTR gene had proven useful for gene therapy, it is hard to see how it would have been developed into a therapeutic without exclusive patent rights in the gene. So it was right to patent the CFTR gene, and Michigan has licensed it for diagnostics in a way that has not hindered innovation or clinical access.

My view of the problem with the BRCA story is not that patents are the root of the problem, but that unduly broad claims were sought and granted, and then Myriad pursued a monopoly business model enabled by those mistakes.

Rather than proscribe DNA patents, it seems wiser to allow them, but restrict their scope by applying criteria for patentability. This would reward universities, companies, and inventors with royalties they justly earn through their valuable contributions, but not more than what they disclose and deserve. Allowing DNA claims that meet criteria for patentability would also preserve the exclusive patent rights that may prove important in developing therapeutic molecules based on DNA that will be expensive to prove safe and effective in clinical studies.

Judge Sweet and the Department of Justice have done a great service by forcing a reexamination of patent policy as it pertains to DNA diagnostics. I am sympathetic to their goals, but their means are unduly broad.

Professor Holman and I thus agree that it is a mistake to stipulate that DNA is not patentable subject matter as a matter of law. It would be far better to admit DNA molecules to patent eligibility, but let the other doctrines of patent law do their work. By narrowing patent scope, we should get the same result in diagnostics, but without the collateral damage to DNA patents that may be therapeutically useful.                    


Robert Cook-Deegan, MD, is Director of the Center for Genome Ethics, Law and Policy at the Institute for Genome Sciences & Policy, Duke University.

Search: GeneWatch
The Council for Responsible Genetics’ Genetic Privacy Manual: Understanding the Threats- Understanding Your Rights will be a comprehensive, electronic source of information for the consumer on these issues.
View Project
Cloning and Human Genetic Manipulation
View Project