Forward Looking Personalized Medicine, Patent Law and Science
|Written by Lori Pressman
Technology Transfer Consultant
Posted: April 15, 2013 @ 11:25 am
Today the Supreme Court will hear oral arguments in Association for Molecular Pathology v. Myriad Genetics, Inc. Myriad Genetics was founded in 1991, and licensed, exclusively in the field of diagnostics, and also invented a group of patents associated with the five to ten percent of breast cancers known to have a strong hereditary component. The biomarkers disclosed, discussed, and claimed in these patents are also associated with an increased lifetime risk of ovarian and other cancers. See this U.S. Government website Genetics of Breast and Ovarian Cancer, updated 3/04/2013, and accessed April 14, 2013, for timely information on the significance of these biomarkers.
Social policy concerns have influenced the AMP v Myriad debate. This article is intended to address certain misstatements of fact and draw attention to other facts not generally considered.
Petitioners have cast the case as “patients versus patents”. Petitioners comprise primarily molecular pathologists and women who assert they had to provide, and receive, respectively, inferior medical care because of Respondent’s patents. In part, because the SACGHS found that so called “genetic” diagnostic tests which were covered by patents cost no more than such tests which were not covered by patents, alleged negative impacts on future medical research have been more forcefully added into the mix of policy concerns creating a biased context for the debate.
Petitioners elected not to try to invalidate the offending patent claims by asserting anticipation, obviousness, or lack of written description, enablement, or specific utility. In a post Rochester v. Searle and post in re: Fisher world, they would have likely have gotten traction with such an approach, and thus clarified and presumably limited the acknowledged scope of the claims.
|Vote for IPWatchdog
We are in the ABA top 100 blogs for the 5th year in a row! Voting now begins to crown the top IP blog. We need your help. Please take a moment to vote for IPWatchdog. CLICK HERE to Vote.
Wanting to engage and sway public opinion, Petitioners and their advocates instead sought to invalidate the offending patent claims based on a legal finding that they claim a “product of nature,” and thus not are not patent eligible subject matter under 101.
While the 2009 case filed in the Southern District of New York used the traditional “prayer for relief” language, requesting the court declare listed claims invalid or unenforceable under 101, the more visible 2013 case before the Supreme Court kicked the rhetoric up a notch, and asks: “Are human genes patentable”?
False Equalities: A “gene” = DNA = the patent claims at issue = information.
A “gene” of course, is an evolving concept and not a molecule. See e.g., the 2007 Gerstein article “What is a gene, post-ENCODE? History and updated definition.” Just as important for an updated understanding of “gene”, read Nessa Carey’s accessible new book © 2012: “The Epigenetics Revolution: How modern biology is rewriting our understanding of genetics, disease and inheritance”, which explains how molecules other than nucleic acids, and even diet, alter gene expression, sometimes even inheritably.
One hundred years ago a “gene” was conceptualized as a “unit of heredity.” Fifty years ago, following the discovery and characterization of the double helix and deoxyribonucleic acid “DNA”, it became a molecule. In 2001, the human genome was sequenced, and scientists reported that only about 1% of it coded for proteins, but kept trying to understand what the other 99% did. A first glimpse of the role of non protein coding DNA appeared in 2007 by the ENCODE pilot project, and motivated the Gerstein article cited above to propose a revised definition of a “gene”.
September 2012 marked a set of comprehensive publications on the role of non protein coding nucleic acids, described by Gina Kolata in a September 5, 2012 NY Times article (accessed April 14, 2013) thus: “The human genome is packed with at least four million gene switches that reside in bits of DNA that once were dismissed as ‘junk’ but that turn out to play critical roles in controlling how cells, organs and other tissues behave.” On the NY Times website with the Kolata article, is this quote from Mark Gerstein of Yale University: “It is like opening a wiring closet and seeing a hairball of wires.” There are many other quotes from similarly surprised and intrigued scientists, and it is doubtful that this will be the last surprising result on human biology.
So post ENCODE and epigenetics, “gene” is a concept once again. Unfortunately DNA had meanwhile assumed a near mythic status, – and it serves Petitioners’ interests, but not ours, and not patients’, and not science itself, to keep the molecule exalted and improperly conflated with “gene.”
Exalting the molecule does this because the legal heart of Petitioner’s case is to argue that DNA is no ordinary molecule, it is “information”, and the “information” is the same whether it has been isolated, -ripped out of the dynamic living system in which switches, assembly and shape matter, and the proteins and enzymes which decorate it (by altering methylation or acetylation) matter-, or not. Yet never has DNA alone been less sufficient information than now. At the moment, LEGO® block is probably a better analogy than blueprint.
Thus was “gene” improperly conflated with DNA in general, and from there to the molecules in the patent claims, which are allegedly “information”, and thus not “patent eligible” enabling Petitioners to present the rather more dramatic question on “human genes” to the Supreme Court and the public.
Bad data, frequently cited.
Meanwhile, the Center for Public Genomics at Duke University received a total of $4.8M of federal funding over a 5 year period, in part, to investigate “gene patents”, and maintain a database, which runs a bioinformatic algorithm and counts patents which reference DNA sequences in their claims., so called “DNA Patents.” While the algorithm is laudably objective, it has been shown, objectively, that it does not find, with a reasonable degree of specificity and sensitivity, patents associated with clinical diagnostic tests informed by an understanding of the underlying genetics.
See particularly figure 2 of the BNA paper, one of the results of the federal funding . As a convenience, the figures from that paper are posted in color here. Figure 2, the Radio Operator Characteristic “ROC” plot shows the same rigorous analysis used by physicians to determine if medical diagnostic tests are good enough, i.e. i) tell you there is a problem when there really is one, ii) tell you there is no problem when there really isn’t one. The bioinformatic algorithm, the one on the Georgetown website, finds and counts patents unrelated to diagnostic tests, and misses others which are and clearly isn’t good enough.
Disappointed in the failure of the algorithm to yield a useful tool for “screening” patents, the team asked expert curators to identify patents said to have the potential to interfere with the practice of medicine by virtue of covering clinical diagnostic tests. The experts read the claims of hundreds of patents and similarly failed to sensitively and specifically a priori identify the “bad” patents.
As readers of IPWatchdog know, this is not surprising, as often diagnostic tests inspired by an understanding of the underlying genetics use proteins, not nucleic acids, or other biomarkers entirely. Patents based on a shared specification can also have claims, issuing in an unpredictable order, on nucleic acid sequences, and proteins, and antibodies to mutated proteins. No wonder you can’t find “diagnostic” patents, or even “genetic diagnostic patents”, by looking for patents which reference nucleic acid sequences in the claims.
The algorithm has a reasonable use as a selector of patents for comparative studies, -such as comparing licensing and commercialization outcomes for DNA Patents managed by Universities with DNA Patents managed by the NIH, but by itself is not a meaningful category of patent.
Thickets: An issue manufactured by Petitioner’s misuse of language and persistent misquotes.
“Gene patent” has regrettably entered every day usage, and an important paper on this topic is so routinely misquoted that even Chris “The-name-of-the-game-is-the-claim” Holman’s heroic attempts to set the record straight may not succeed. The particularly bad misquotes and out-of-context excerpts start from a paper published in 2005 by Kyle Jensen and Fiona Murray “The Intellectual Property Landscape of the Human Genome.” Jensen and Murray carefully, albeit quaintly in view of ENCODE, define a “gene” as “a set of cotranscribed protein-encoding exons,” but “gene patent” as “a patent disclosing and claiming a human gene sequence or some fraction thereof. [emphasis added] Thus, most of the Jensen-Murray 4,270 ‘‘gene patents,’’ as they define them are not to “genes” as they define them! Pity the poor bewildered nonscientist citizen, science writer, non biological scientist social policy researcher, or jurist trying to make sense of the policy implications of the Jensen Murray paper.
The misunderstanding of the Jensen Murray work, inspired by the theoretical publications of eloquent law professor Rebecca “tragedy-of-the-anti-commons” Eisenberg, has fed the “thicket” theory, -that there are just too many of these patents, and they will get in the way of many things, including research. However, the empirical, systematic and thorough work on this topic, the 2005 Walsh Cho Cohen Final Report to the National Academy of Sciences’ Committee Intellectual Property Rights in Genomic and Protein-Related Inventions, and more recently Chris Holman’s work show this is not the case.
The Walsh work emphasized academic’s disregard for patents generally, and their understandable self interest in appropriately attributed credit, -which is often easier to assure with a patented material than an unpatented one. One of the remarkable not fully appreciated findings of the Walsh work was they observed a greater delay associated with sharing of unpatented material than patented materials! Chris Holman, for his part, determined that “Analysis of the claims of 533 of the of the patents identified in the Science [Jensen Murray article referenced above] article as ‘covering’ human genes reveals that most do not include a single claim that would be infringed by whole genome sequencing and other forms of genetic testing.”
Examples of companies with diagnostic products and patents which help patients.
Important empirical evidence, based on AUTM data and Better World Reports “BWRs” is found in then AUTM President Todd Scherer’s letter on behalf of AUTM to the USPTO. This letter discusses eleven companies and their products which rely to some degree on i) patents which either have nucleic acid sequences in the claims, in the “DNA Patent” sense, or ii) which make and sell diagnostics, and sometimes, but not always, both.
First, the AUTM BWRs clearly show that patents with nucleic acid sequences in the claims, and patents on diagnostics can and do benefit patients.
Second, consistent with the ROC analysis discussed earlier, the patents (found by typing the inventor and institution names in the AUTM BWRs into public patent databases) likely associated with the tests have a mix of nucleic acid claims and analytes and other biomarker claims and analytes.
Third, these examples were selected from AUTM Better World Reports because they described diagnostic or nucleic related products and not because they were developed or sold by start-ups, yet more than half were developed or sold by start-ups, and still more worked with start-ups or small companies to bring products to market. None of the AUTM examples started with a license to a large company.
As known to readers of IPWatchdog, patents play a key role in attracting talent and capital to start-ups and small companies and in gaining them a bargaining position with investors and business partners, – and start-ups plant the seeds which renew our innovation ecosystem.
My research showed that i) diagnostics take longer to develop than reagents, though not as long as the generally accepted time for therapeutics, ii) a consensus on clinical utility, and thus insurance reimbursability is likely the more significant gate to diagnostic product availability than is FDA approval, and iii) patent incentives play a positive and role for development of certain diagnostic tests, including, possibly making such tests available sooner. See in particular figure 6 and table 2 of the BNA paper. As the line between diagnostics and therapeutics continues to blur, overly doctrinaire approaches to patent eligibility could perversely harm, and not help patients. This empirical analysis was done using the Bradford Hill criteria for investigating causation, from his classic 1965 paper and included studying product commercialization timelines of hundreds NIH managed “DNA Patents” and thousands of university managed “DNA Patents.”
The only alternative to patents is not “open source”, -it is also trade secrets.
Virtually unmentioned by anyone is the fact that absent patents on isolated nucleic acids, and post Prometheus v. Mayo, fewer phenotype-biomarker associations will be published at all. Already there are companies based on proprietary tissue samples, from 23 and Me, to µBiome, to Dognition to Champions Oncology. Proprietary medical data and tissue collections provide a never expiring first mover advantage analogous to the customer and subscriber lists of Facebook, Google, Amazon and Target.
The past is past.
The 2001 publication of the Human Genome, combined with the Rochester v. Searle and in re Fisher decisions, mean that claims of the scope in this case are in the well in the past. Even now, one wonders at the results of a simple Markman ruling on the disputed claims. 89 of 99 of the patents numbers referenced in the case studies the Genetics in Medicine articles resulting from the SACGHS study have priority dates before the 2001 publication of the human genome and 96 of the 99 have priority dates before the September 7, 2005 the in re Fisher decision on specific utility, which of course post dates the 2004 Rochester v. Searle decision on written description and enablement.
The future awaits.
The authors of the 2007 Gerstein article who felt a need to revise the definition of “gene”, start by asserting that it must be “backward compatible.” Perhaps they felt it was not even necessary to state that this new definition be forward compatible as well. The Supreme Court, to the extent it must make a ruling for our times, informed by societal context, should dispassionately consider all the available empirical evidence, from the academic work cited here, to the claim scope limits resulting from massive sequence publication projects and recent court cases, and the thriving innovation ecosystem in personalized medicine at and among for profit and not for profits, and render a clear forward compatible decision for us all.