Empirical Research Fails to Support Gene Patenting Ban

    By Kevin E. Noonan --

Bans on gene patenting and the possibility of having a court, Congress, or czar declare them banned are all the rage recently, spurred in part by the lawsuit against Myriad Genetics and the University of Utah (among other defendants) in the Federal District Court for the Southern District of New York (see "Association for Molecular Pathology v. U.S. Patent and Trademark Office").  Whether that lawsuit will go forward is expected to be decided by the end of the month, when a decision on standing for many of the plaintiffs is expected from the District Court.  No matter how the judge rules, however, interests groups like the American Civil Liberties Union are unlikely to refrain from efforts to have gene patenting banned.

Under the circumstances, it is reasonable to wonder whether gene patents are in fact inhibiting innovation or otherwise harming patients, physicians, or researchers (since the existential harm asserted by those against gene patents -- that it is fundamentally "wrong" -- is harder to assess quantitatively).  The latest report from academia on the question is consistent with several other reports refuting negative effects predicted on theoretical grounds (such as the "tragedy of the anticommons" idea).  Indeed, the report's conclusions are that the greatest difficulty surrounding gene patents arises from the legal uncertainties created in part by efforts to ban gene patenting.

The report, from Isabelle Huys, Nele Berthels, Gert Matthijs & Geertrui Van Overwalle, researchers at the Catholic University of Leuven, Belgium, is found in the October 2009 edition of Nature Biotechnology (see "Legal uncertainty in the area of genetic diagnostic testing").  The authors' express intent was to assess whether gene patents, either as "blocking patents" or as a "patent thicket," were inhibiting genetic technology.  The concern stemmed from the belief that such patents would be "difficult or impossible to circumvent."  Their research is current as of February 6, 2009 (and thus, it does not reference the ACLU-backed lawsuit filed earlier this year).

Defined as a "large-scale empirical study," the authors describe the following "heart of the problem:  which types of claims occur in disease-specific patents and to what extent are these claims essential for carrying out genetic diagnostic tests?"

The sample set was U.S. and European (EP) patents and patent applications directed at twenty-two inherited, monogenic diseases (see Table 1 below; click on table to enlarge).  The legal status of the patent documents reviewed included patent applications, granted patents, patent applications "dropped" (i.e., abandoned during prosecution) and abandoned, withdrawn, revoked, or expired patents.  267 claims from 145 patents were assessed, with regard to whether they were "easy," "difficult," or almost "impossible" to circumvent, based on comparing the claimed methods to the "best practices" guidelines for genetic susceptibility screening; claims deemed "impossible" to circumvent were characterized as "blocking" claims.  The claims studied were independent claims for genes, oligonucleotides, methods, and kits.  However, they did not review the file history nor does the report compare the commercial embodiments with the claims or their licensing status.  Finally, U.S. claims were assessed with regard to the "written description," "enablement," and "best mode" requirements, and the European patents with regard to EPC Art. 83 and 69.

The patents and applications (250 patent documents) were grouped into 72 different patent families (where a "family" is "a group of patents or applications ('patent documents') taken in various countries to protect a single invention").  U.S. patents were found in 66 families whereas EP patents were found in only 26 families.  Twenty-one families had both U.S. and EP patents, while 46 families had only U.S. patents and only 5 families had EP but not U.S. patents.  There were 42 patent families containing U.S. applications versus 22 families with EP applications pending.  "Remarkably," according to the authors, "about twice as many patent applications are dropped during the patent prosecution procedure in Europe (34) in EP than in US (16)."  Also, in Europe there were about the same number of granted patents (26) as pending applications (22), whereas in the U.S. there are 66 granted patents and 42 pending applications.

There were 56 different applicants for these patents/applications, with 62.5% of the applicants coming from non-profit sector, predominantly universities and research organizations, 58.9% of applicants from the U.S., 25% from EP, and 16.1% from JP or CA, with U.S. applicants owning 55.8% of the EP patents and 67% of the U.S. patents.  There were also 11 different co-ownerships with 17 different partners.  The top applicants from the U.S. were Johns Hopkins University (6 patent families) and Baylor College of Medicine (6 families); the top European applicant was Leiden University (3 patent families).

The vast majority of genetic diseases (19/22) have patented tests (Huntington's disease a surprising exception); the status of the diseases and the patents or applications relating to each are shown in Table 2 below (click on table to enlarge).  Four diseases were covered by granted patents in Europe, while six diseases have U.S. patent coverage.

(Huys et al., 2009, Nature Biotechnology 27: 903-909, Table 2, p. 907).

The "[m]ost heavily patented [diseases] in Europe are hereditary hemochromatosis and familial breast cancer testing," while in the U.S., "most patents are for spinocerebellar ataxia, Charcot-Marie-Tooth neuropathy, hereditary nonpolyposis colon cancer and familial breast and ovary cancer."

145 granted patents were reviewed (118 U.S. patents and 27 EP patents) and 267 claims from these patents characterized:  of these 38% were method claims, 25% gene claims, 23% oligonucleotide/probe claims, and 14% kit claims.  For the patents having claims to genes (66/145), the majority were claims for cDNA copies of genes (although there were 22 patents having claims to genomic DNA embodiments of the claimed genes).  When assessed under the "easy," "difficult," or "impossible" (i.e., "blocking") standard, only 3% of these gene claims were "blocking."  Accordingly, the authors concluded that these claims did not represent a "patent thicket" or otherwise provide an impediment to innovation.

For method claims, on the other hand, 30% impossible, 47% difficult, and 23% easy to circumvent, and 35/145 patents contained at least one blocking method claim.

Claim 1 of U.S. Patent No. 5,693,470 was recited as exemplary of a "blocking" method claim (for hereditary nonpolyposis colon cancer):

1.  A method of determining a predisposition to cancer comprising:
testing a body sample of a human to ascertain the presence of a mutation in a gene identified as hMSH2 (human analog of bacterial MutS and Saccharomyces cerevisine MSH2) which affects hMSH2 expression or hMSH2 protein function, the presence of such a mutation indicating a predisposition to cancer.

60% of patent families (21/35) had no kit claims, and 15/22 diseases have claims impossible to circumvent.

The authors reported a number of observations that they characterized as "surprising."  These included the supremacy of U.S. patents on genetic inventions persists, almost 30 years after Diamond v. Chakrabarty.  They also noted that, despite the requirement in the U.S. since 2001 that gene patents can be obtained in the U.S. only if "biochemical, biological, or genetic data describing the function are included" (they reported 63 such patents), 11 "key" patents were directed to genes without this information.  The situation is different in Europe, with many more applications abandoned during prosecution or otherwise not pursued.  The authors speculate that this many be due "perhaps" to a more restrictive and expensive patent system.  The authors also noted that 60% of applicants were U.S. universities and research institutes, due to the "progressive attitude" of the U.S. for universities to be able to patent their inventions.

With regard to the claim analysis, the data clearly showed no evidence of a patent thicket for gene claims.  The situation is different for diagnostic method claims, with most of these claims being difficult or impossible to circumvent.  Because the authors found there to exist claims to the same disease owned by different applicants, they noted the possibility that a patent thicket could exist but provide no evidence for any.  They also noted that for method claims, there was a failure to link the biological observation to specific method steps, which may make it difficult to find infringement for these claims.  (For example, a claim reciting an isolated cDNA encoding a human gene is unlikely to be infringed by a test for identifying a mutation in DNA from a human sample.)

The authors present these conclusions:

In conclusion, the present analysis and accompanying observations do not point to the existence of a wide patent thicket in genetic diagnostic testing.  Rather, they highlight a problem of lack of transparency and clarity, leading to legal uncertainty.  Neither case law nor patent legislation resolves the legal uncertainty related to patents on genetic inventions.  In 2006, the US Supreme Court dismissed a judicial review on the case Laboratory Corporation of America Holdings v. Metabolite Laboratories Inc.  . . .  As a substantial number of patent claims would have been affected by a decision in this case, this refusal further increased legal uncertainty, including in the genetic field.  The consequence of this high level of legal uncertainty is that either enormous risks are taken if genetic tests are performed without knowingly infringing a specific patent, or much time and energy goes into establishing patent landscapes and freedom-to-operate analyses or to efforts to use different techniques and methods that may eventually be below the state of the art that is clinically requested.

As this study shows that not that many blocking gene patents exist, proposals aiming at banning patents on human genes do not provide a plausible solution, unless the ban would be on patents for broad genetic diagnostic methods as such.  For instance, the European Society of Human Genetics (Vienna) has recently recommended avoiding patenting of the pure link between a mutation and disease.  More attention should be paid to the licensing practices in a 'responsible' way.  Otherwise, the risk exists that the control by owners of patents containing those broad claims with respect to genetic diagnostic testing may in the future dissociate actual genetic diagnostic testing from genetic counseling and clinical investigation, which is to the detriment of progress of the genetic diagnostic service and public healthcare system.

(Emphasis added)

These conclusions reflect two realities.  The first is that researchers have identified practices that may, in the future, lead to negative outcomes, but have not detected any evidence of such outcomes using current data.  The second is that the legal uncertainty occasioned by "decisions" such as Justice Breyer's dissent in the LabCorp case have consequences for research, investment, and healthcare.  Judge Rader may note, in the Prometheus decision, that this dissent is "not controlling law," but these sentiments, combined with the risk of a gene patenting ban, are (according to these authors) on balance having an effect more negative on innovation than the gene patents that are at risk.

(click on table to enlarge)

Comments

[Gena777]

This is an interesting article. However, I'm not sure that taking an unwavering position regarding gene patenting is advisable at this point, because so many significant considerations are still unknown. The issue is so new, and future technologies are so uncertain, that specific policy details will likely have to be hashed out in future patent litigation. Please keep us posted as things progress.
http://www.GeneralPatent.com

[Kevin E. Noonan]

Dear Gena:

One thing to keep in mind. The vast majority of human genes will no longer be eligible for patenting after about 2020. Most of these genes were disclosed, either in patent applications or in the Human Genome Project database, prior to 2000. Since the U.S. changed its patent term to 20 years from the earliest filing date, all human gene patent applications filed at that time cannot extend past 2020, and if these genes have not been filed by that time, their disclosure in public databases would also preclude patenting.

It will be interesting to see whether alternative claiming strategies, espoused by some participants in the gene patenting debate, fill the breach. I suspect the same voices will remind us that, having identified a gene sequence, expressing that sequence or using the encoded protein would be obvious and therefore unpatentable.

But you are right: the most prudent course is simply "stayed tuned.' Thanks for the comment.

[Bob Cook-Deegan]

Kevin,

Just to point out, your response to Gena is at serious variance with your own blog of just a few days ago. You wrote in laudatory terms about Myriad exclusively licensing a patent (turns out it's a Johns Hopkins PCT patent application). It's 2009. That patent will expire in 2029. If granted, it will be a "gene patent" of some sort, since it will be based on the mutations newly associated with familial pancreatic cancer. Either there will continue to be such patents or there won't. The number of total DNA patents peaked in 2000 and dropped until 2006, but it is back up again. The argument that "it's not so bad, these patents will just go away," is not compelling for many reasons. Many gene-based inventions will surely continue to be both patent-eligible and patentable.

[Kevin E. Noonan]

Dear Bob:

What I said was "the vast majority" of gene patents. So "serious variance" is a serious overstatement.

But you raise an important point. While the vast majority of gene patents will expire by 2020, what is being worked out now are mechanisms of disease and how genetic polymorphisms are associated with disease. There will continue to be claims such as "a method for identifying a risk of disease Y, by detecting a polymorphism in gene X." And that was the subject of the piece on the pancreatic cancer gene.

I think these are different claims, impacting different aspects of the gene patenting debate. Those who believe gene patenting is "wrong" don't care about the distinction, but there is a very practical difference. If I cannot protect the gene for protein A, which treats disease Y, then I think my chances of taking the risk of bringing a therapeutic drug to market is essentially zero. These drugs won't exist.

On the other hand, if I cannot protect the correlation between polymorphism Z in gene X related to disease Y, then as I have said these polymorphisms will be identified but not disclosed, in order to justify the expense of bringing them to market. I think that is bad public policy, but unlike in the therapeutic arena the tests will be available.

I frankly don't know what will happen once the "vast majority" of genes go off-patent in 2020 with regard to therapeutic drugs. But it may not be pretty.

Thanks for the comment.

[Gary Johnston]

Kevin:

Where is the evidence for your statement that without patent protection, the biomarker correlations will be identified but not disclosed? You do understand that almost all biomarker correlations that are identified are done so by researchers at universities, non-profits, and government labs, right? You're telling me that these entities are going to start behaving like Coca-Cola and keep all their correlation discoveries secret? I don't think so. Moreover, no physician will order a test unless he/she knows the biological basis for the test (I know that most physicians are not scientists, but they're not dummies either). And I haven't even mentioned the fact that there are federal and state regulations (i.e., CLIA) which would prohibit secret tests. Finally, the notion that companies will not undertake the expense of getting a test based on a correlation up and running without patent protection is not only inaccurate, it's bizarre. Go look at the roster of tests offered by the largest reference laboratories such as Quest and LabCorp. The overwhelming majority of these tests are not patented. These companies make their money by volume and efficiency. I know that those terms are heresy to some patent attorneys out there, but that's the way most American companies make their money.

[Kevin E. Noonan]

Dear Gary:

No, I don't think that universities will act in that way. Indeed, looking over my hypothetical, I said it would be companies that would do so. If you doubt that companies can compete in this space with academia, revisit the Human Genome Project versus Craig Venter.

I think physicians will accept a test that has a good prognostic track record. If a salesman could come into a physician's office and show that a chip could reliably predict disease, I think there would be less concern about the specifics.

I note that both Quest and Labcorp have licensed patents to protect their tests, and while it is true that these companies make their money by volume and efficiency, there is a place for patenting in their business models, too (and these are established companies, not start-ups, just the companies that could be expected to adopt my hypothetical approach).

Indeed, your argument about "efficiency and volume" sounds a lot like what used to be said about Japanese and German companies in the 1970's and 1980's - before it we widely recognized that the FTC had forced many US companies to give licenses on its technology to foreign companies due to antitrust concerns. The fact is that nascent industries need patent protection, and if we preclude that protection it will favor just those established companies least likely to do cutting edge innovation. I don't think that is the kind of behavior we should be promoting.

Thanks for the comment.

[David Koepsell]

Dear Kevin,

Craig Venter made enormous progress in the HGP, as we know, by piggybacking off the publicly-funded science (his map was only 50% his data, the rest was taken from the public domain HGP). Moreover, the conclusion of the article you cite indicates that the uncertainties caused by gene patents are, in fact, holding up research to a degree, even if not to a large degree). In the case of BRCA 1 and 2 as you know, and as you more or less admitted last week at the forum in NYC, the exclusive licensing by Myriad is a poor policy, and bad PR (it also is hindering science). You might also point out that in Europe, there is mandatory licensing for basic research, so the same sort of thickets do not occur in Europe. I also agree with the commentators above, as you know, that the science would proceed just fine, and disclosure would occur as part of the institutions of science, without patents. You should have attended the screening and panel the following night of the movie "In The Family" because there we heard from a genetic counselor and a clinical physician, both of whom had received cease and desist letters for their clinical work, or had been unable to conduct their work appropriately due to these sorts of patents. The crime is that it's not even just the basic science that is being hindered where firms choose to hold exclusive patents and won't license, but health care is also impacted.

It was good to meet you finally, and it was great to hear you at a public forum being civil, rather than throwing around words like "lies" as you did some time ago, both in reference to the Myriad suit and to my book.

best,
David

[Kevin E. Noonan]

Dear Dave:

It was also good to meet you. I think you are very well aware of the difference between misusing a right, and not being entitled to the right in the first place.

And the funny thing (that I mentioned) is that it the legal uncertainty, not the patents, that is causing the present difficulties.

Thanks for the comment.

[Ron Katznelson]

David Koepsell said: “You might also point out that in Europe, there is mandatory licensing for basic research, so the same sort of thickets do not occur in Europe. I also agree with the commentators above, as you know, that the science would proceed just fine, and disclosure would occur as part of the institutions of science, without patents.”

Of course, this enlightened European mandatory licensing policy was adopted with the same European certitude that drove the Netherlands to abolish the patent system in 1869 (only to restore it 43 years later). They too, thought that science and technology would progress “just fine without patents.”

[David Koepsell]

and, in fact, it did:
http://www.atypon-link.com/doi/abs/10.1257/0002828054825501

[David Koepsell]

here's the full Moser article, which contradicts the notion that without patents, somehow Dutch innovation languished: http://www.google.com/url?sa=t&source=web&ct=res&cd=7&ved=0CCYQFjAG&url=http%3A%2F%2Fweb.mit.edu%2Fmoser%2Fwww%2Fpat501.pdf&ei=OXDxSrr0GuDOjAfV6_yVAQ&usg=AFQjCNFK7aRz7FuMw2PmmfCMDgFmtTt4aw