Each of the following cases involves an important research tool in molecular biology, and each was chosen to illustrate a form of protection of intellectual property and a pattern of development involving both the public and the private sector. For each case, we present background material and a summary of the discussion that raised issues peculiar to the case.
The ideal strategies for the handling of intellectual property in molecular biology are not always immediately obvious, as these case studies illustrate. For most, final decisions have not been made about how access to these research tools will be controlled. Such decisions might be modified in response to both scientific and legal developments.
RECOMBINANT DNA: A Patented Research Tool, Nonexclusively Licensed With Low Fees
The Cohen-Boyer technology for recombinant DNA, often cited as the most-successful patent in university licensing, is actually three patents. One is a process patent for making molecular chimeras and two are product patents—one for proteins produced using recombinant prokaryote DNA and another for proteins from recombinant eukaryote DNA. Recombinant DNA, arguably the defining technique of modern molecular biology, is the founding technology of the biotechnology industry (Beardsley 1994). In 1976, Genentech became the first
company to be based on this new technology and the first of the wave of biotechnology companies, which in fifteen years has grown from one to over 2000.
The first patent application was filed by Stanford University in November 1974 in the midst of much soul-searching on the part of the scientific community. Stanley Cohen and Herbert Boyer, who developed the technique together at Stanford and the University of California, San Francisco (UCSF), respectively, were initially hesitant to file the patent (Beardsley 1994). Several years of discussion involving the National Institutes of Health (NIH) and Congress followed. By 1978, NIH decided to support the patenting of recombinant DNA inventions by universities; in December 1980, the process patent for making molecular chimeras was issued. The product patent for prokaryotic DNA was issued in 1984. The patents were jointly awarded to Stanford and UCSF and shared with Herbert Boyer and Stanley Cohen. The first licensee signed agreements with Stanford on December 15, 1981. As of February 13, 1995, licensing agreements had generated $139 million in royalties, which have shown an exponential increase in value since their beginning. In 1990–1995 alone, the licensing fees earned $102 million.
This case has three key elements. First, the technology was inexpensive and easy to use; from a purely technical standpoint, there were only minimal impediments to widespread dissemination. Second, there were no alternative technologies. Third, the technology was critical and of broad importance to research in molecular biology.
The technology was developed in universities through publicly funded research. The strategy used to protect the value of the intellectual property was to make licenses inexpensive and attach minimal riders. The tremendous volume of sales made the patent very lucrative. Every molecular biologist uses this technology. However, not all inventions are as universally critical. Only a few university patents in the life sciences, such as warfarin and Vitamin D, have been even nearly as profitable as the Cohen-Boyer patent. Clearly, had this technology not been so pivotal for molecular biology or had an equally useful technology been available, the licenses would not have been sold so widely and the decision to license the technology might have met with more resistance.
The Cohen-Boyer patent is considered by many to be the classic model of technology transfer envisaged by supporters of the Bayh-Dole Act, which was intended to stimulate transfer of university-developed technology into the commercial sector. Ironically, it presents a different model of technology than that presumed by advocates of the Bayh-Dole act (for discussion, see chapter 3). Lita Nelsen, director of the Technology Licensing Office at the Massachusetts Institute of Technology (MIT), noted that the premise of the Bayh-Dole Act is that exclusivity is used to induce development and that universities should protect their intellectual property because without that protection, if everybody owns it, nobody invests in it. ''The most-successful patent in university licensing, in the entire history of university licensing, is the Cohen-Boyer pattern which is just the
reverse. It is a nonexclusive license. It provides no incentive, just a small tax in the form of royalties on the exploitation of the technology.''
The biotechnology boom that followed the widespread dissemination of recombinant DNA techniques transformed the way universities manage intellectual property. It also fundamentally changed the financial environment and culture of biological research.
Nelsen described two ways in which this patent was so successful in fostering the aims of the Bayh-Dole Act. First, it got the attention of biologists by showing the advantages of protecting intellectual property. Stanford earned respectability for the venture by involving NIH and discussing in a public forum how this technology could be disseminated in a way that would not impede research. Second, it got the attention of university chancellors. They began to see that licensing, patenting, and technology transfer might have some financial benefits for the university. Nelsen commented that "that went a little too far. Everybody was waiting for $100 million per year out of their technology transfer offices. Most of them did not get it, and most of them are never going to get it." In the meantime, technology transfer managers developed more experience and became professionalized. They began to learn how to decide what to patent, how to market technology, and how to close deals at reasonable prices and with reasonable expectations. And industry learned how to negotiate licenses with universities.
Nelsen concluded that the whole biotechnology industry came out of the Cohen-Boyer patent, not only because Cohen-Boyer developed gene splicing, but because universities learned how to do biotechnology and early technology licensing—even if the first example was paradoxical.
The decision to negotiate nonexclusive, rather than exclusive, licenses was critical to the industry. If the technology had been licensed exclusively to one company and the entire recombinant DNA industry had been controlled by one company, the industry might never have developed. Alternatively, major pharmaceutical firms might have been motivated to commit their resources to challenging the validity of the patent.
Nelsen noted that at most major universities, it has become standard in industry-sponsored research agreements that the university will retain ownership of any resulting patents but almost without exception will grant the sponsor a first option to an exclusive license. With the increase in university-industry partnerships this applies to more research than in past years. Moreover, the Bayh-Dole Act encourages universities to grant exclusive licenses to companies even if the research was publically sponsored. But as the next case study shows, even when a company holds exclusive rights to a fundamental technology, it might choose to disseminate the technology broadly.
PCR AND TAQ POLYMERASE: A Patented Research Tool for Which Licensing Arrangements Were Controversial
Polymerase chain reaction (PCR) technology presents an interesting counterpoint to the Cohen-Boyer technology. Both are widely used innovations seen by many as critical for research in molecular biology. However, the licensing strategies for the two technologies have been quite different, and they were developed in different contexts.
PCR allows the specific and rapid amplification of targeted DNA or RNA sequences. Taq polymerase is the heat-stable DNA polymerase enzyme used in the amplification. PCR technology has had a profound impact on basic research not only because it makes many research tasks more efficient, in time and direct cost, but also because it has made feasible some experimental approaches that were not possible before the development of PCR. PCR allows the previously impossible analysis of genes in biological samples, such as assays of gene expression in individual cells, in specimens from ancient organisms, or in minute quantities of blood in forensic analysis.
In less than a decade, PCR has become a standard technique in almost every molecular biology laboratory, and its versatility as a research tool continues to expand. In 1989, Science chose Taq polymerase for its first "Molecule of The Year" award. Kary Mullis was the primary inventor of PCR, which he did when he worked at the Cetus Corporation. He won a Nobel Prize for his contributions merely 8 years after the first paper was published in 1985, which attests to its immediate and widely recognized impact. Tom Caskey, senior vice-president for research at Merck Research Laboratories and past-president of the Human Genome Organization, attributes much of the success of the Human Genome Project to PCR: "The fact is that, if we did not have free access to PCR as a research tool, the genome project really would be undoable. . . Rather than bragging about being ahead, we would be apologizing about being behind."
Whereas recombinant DNA technology resulted from a collaboration between university researchers whose immediate goal was to insert foreign genes into bacteria to study basic processes of gene replication, PCR was invented in a corporate environment with a specific application in mind—to improve diagnostics for human genetics. No one anticipated that it would so quickly become such a critical tool with such broad utility for basic research.
Molecular biology underwent considerable change during the decade between the development of recombinant DNA and PCR technologies (Blumenthal and others, 1986). The biotechnology industry emerged, laws governing intellectual property changed, there was a substantial increase in university-industry-government alliances, and university patenting in the life sciences increased tenfold (Blumenthal and others 1986, Henderson and others). There was virtually no controversy over whether such an important research tool should be patented
and no quarrel with the principle of charging licensing fees to researchers. The controversy has been primarily over the amount of the royalty fees.
Cetus Corporation sold the PCR patent to Hoffman-LaRoche for $300 million in 1991. In setting the licensing terms for research use of PCR, Roche found itself in a very different position from Stanford with respect to the Cohen-Boyer patent. First, it was a business, selling products for use in the technology. That made it possible to provide rights to use the technology with the purchase of the products, rather than under direct license agreements, such as Stanford's. This product-license policy was instituted by Cetus, the original owner of the PCR patents. An initial proposal to the scientific community by the president of Cetus for reach-through royalties—royalties on second-generation products derived through use of PCR—was met with strong criticism. Ellen Daniell, director of licensing at Roche Molecular Systems, noted that the dismay caused by the proposal has continued to influence the scientific community's impression of Roche's policy.
Roche's licensing fees have met with cries of foul play from some scientists who claim that public welfare is jeopardized by Roche's goals. Nevertheless, most scientists recognize that Roche has the right to make business decisions about licensing its patents. The fact that Roche had paid Cetus $300 million for the portfolio of PCR patents led some observers to think that Roche intended to recoup its investment through licensing revenues, a point that Daniell disputed. She pointed out that Roche's business is the sale of products and that licensing revenues are far less than what would be needed to recoup the $300 million over a time period that would be relevant from a business viewpoint. Daniell listed Roche's three primary objectives in licensing technology:
Expand and encourage the use of the technology.
Derive financial return from use of the technology by others.
Preserve the value of the intellectual property and the patents that were issued on it.
Roche has established different categories of licenses related to PCR, depending on the application and the users. They include research applications, such as the Human Genome Project, the discovery of new genes, and studies of gene expression; diagnostic applications, such as human in vitro diagnostics and the detection of disease-linked mutations; the production of large quantities of DNA; and the most extensive PCR licensing program, human diagnostic testing services. Licenses in the last-named category are very broad; there are no upfront fees or annual minimum royalties, and the licensees have options to obtain reagents outside Roche.
Discussion about access to PCR technology centered on the costs of Taq polymerase, rather than on the distribution of intellectual property rights. Tom Caskey's view was that "the company has behaved fantastically" with regard to
allowing access to PCR technology for research purposes. Bernard Poiesz, professor of medicine at the State University of New York in Syracuse and director of the Central New York Regional Oncology Center, agreed that he knew of no other company that had done as well as Roche in making material available for research purposes. But he also argued that the price of Taq polymerase is too high and has slowed the progress of PCR products from the research laboratory to the marketplace. Poiesz stated that the diagnostic service licenses "are some of the highest royalty rates I have personally experienced." He cited the example of highly sensitive diagnostic tests for HIV RNA, which he said are too expensive for widespread use, largely because of the licensing fees charged by Roche.1 Caskey felt that Roche should have expanded the market by licensing more companies to sell PCR-based diagnostic products and profited from the expansion of the market, rather than from the semiexclusivity that it has maintained.
Nor are all university researchers satisfied with their access to Taq polymerase. Ron Sederoff commented that—in contrast to the human genomics field, in which funding levels are much higher than for other fields of molecular biology—many academic researchers do not find easy access to the technology. Several workshop participants noted that the high cost Taq polymerase made many experiments impossible for them.
What is the effect of the Cetus-Roche licensing policy on small companies? Tom Gallegos, intellectual property counsel for Oncorpharm, a small biotechnology company, stated that most small companies cannot afford the fees charged by Roche. He noted that the entry fee for a company that wants to sell PCR-based products for certain fields other than diagnostics ranges from $100,000 to $500,000, with a royalty rate of 15%. By comparison, a company pays about $10,000 per year and a royalty fee of 0.5–10% for the Cohen-Boyer license. The effect is an inhibition of the development of PCR-related research tools, with consequent reductions or delays in the total royalty stream and possibly litigation.
Sidney Winter, professor of economics at the Wharton School of Business, suggested that in asking whether the price of some technology is too expensive, one should consider "compared with what?" Compared with licensing and royalty fees for Cohen-Boyer, PCR might seem excessive. If one imagines that the cost of the PCR patent were financed by a tax on the annual US health-care expenditure which was about $1 trillion in 1995 (Source: Congressional Budget Office), that tax would be roughly equal to 0.03% and might be a price worth paying for the advances made possible by PCR technology.
During the workshop, several people distinguished between research tools
that are commercial products and tools that have little market value but are important tools for discovery. In the case of PCR, the research tool is both a commercial product and a discovery tool. As such, it raises questions. Are the PCR patents an example of valuable property that would have been widely disseminated in the absence of patent rights? Is PCR an example of a technology that has been more fully developed because of the existence of patent rights? Daniell stated that Roche has added considerable value to the technology, in part through the mechanism of patent rights. There was vigorous discussion and disagreement as to whether the licensing fees justify the value added by Roche.
PROTEIN AND DNA SEQUENCING INSTRUMENTS: Research Tools to Which Strong Patent Protection Promoted Broad Access
This case study was selected because it provides a clear example of how patent protection promoted the development and dissemination of research tools. By most standards, this would be considered a successful transfer of technology. The possibility of automated, highly sensitive DNA and protein sequencers was developed in the public sector by Leroy Hood's group at California Institute of Technology (Cal Tech). However, it was only with the help of substantial private investment that these research tools were widely disseminated.
The ability to synthesize and sequence proteins and DNA revolutionized molecular biology; automating these tasks promised to consolidate the revolution. Indeed much of the achievement of the Human Genome Project is attributable to the development of automated sequencing instruments, which greatly reduced the time and cost needed to sequence DNA. Because the effects of genes depend on the proteins that they encode, protein sequencing has been a key step in deciphering gene function. Until automated sequencing instruments were widely available, only a few laboratories had access to this technology.
The prototypes for these instruments were developed in Hood's laboratory during the years 1970–1986. Over a period of six or seven years, the team of scientists assembled by Hood increased the sensitivity of protein sequencing instruments by a factor of about 100. That transformed a difficult and uncertain task into one that could be reliably accomplished with the minute quantities of purified proteins that so often limited the scope of the analysis. Hood's laboratory was the first to sequence lymphokines, platelet-derived growth factor, and interferons. After those successes, he was approached by many scientists who asked why the technology could not be made available to the whole research community. Since the middle 1990s, the technology has become widely available.
The broad availability of sequencing technology is due, in no small part, to Hood's perseverance in the face of widespread skepticism. His 1980 manuscript
that described, for the first time, automated DNA sequencing was delayed by the journal Nature on the grounds that this technology sounded like "idle speculation." Hood wrote three or four proposals to NIH and the National Science Foundation but was unable to obtain funding for his instrumentation work. The bulk of the support for this technology came from the private sector, and even then companies were reluctant to invest in developing the sequencing instrumentation. He approached nineteen companies, all of which declined to support the development of the sequencers. Eventually, he obtained funding from Applied Biosystems (ABI), but even this support required difficult negotiations between Cal Tech and ABI. ABI insisted on, and received, an exclusive license. As Hood told it, the argument that convinced Cal Tech to support the arrangement was that "if the scientific community wants these instruments, it is our moral obligation to make them commercially available."
At the time of this workshop, ABI had sold more than 3,000 DNA sequencers and more than 1,000 protein sequencers worldwide (although some elements of the technology, such as peptide synthesis, were not protected by patents, most of the instrumentation was patented by ABI). Sequencing facilities that serve multiple investigators are now standard features at research universities. That is not to say that licensing of this technology has been without controversy. Cal Tech licensed the technology to ABI with the stipulation that ABI would sublicense it under what Cal Tech considered reasonable terms. A number of companies have argued that ABI's terms are not reasonable. As with PCR, the situation is complicated in that the primary licensee claims that its license fees reflect what it needs to charge to earn a reasonable return on its investment in developing the technology.
ABI is clearly the leader in the world market for DNA sequencers. But other companies, such as Pharmacia and LI-COR, have important market shares. LI-COR has established a niche in the market with its infrared fluorescence DNA sequencer; infrared light has low background fluorescence, which allows for the development of more robust, solid-state instrumentation than is possible with other DNA sequencing technology. LI-COR is typical of many small biotechnology companies in its reliance on its patent portfolio. Harry Osterman, director of molecular biology at LI-COR, noted that "DNA sequencing is more than just an instrument, it is a system. To make a viable product, all the disparate pieces need to be integrated. That makes for a challenging intellectual property and licensing exercise, unless you have the internal funds to do everything. You require instrumentation, software, chemistry, and microbiology." Patent protection allows a small company to negotiate cross-licenses, which are critical in systems technologies, such as sequencing instrumentation. It can provide an opportunity that a small company would not otherwise have to compete in a market.
One might argue that patent protection served both the large company (ABI) and the small company (LI-COR) in bringing their sequencing technology to the market. In the case of ABI, patent protection afforded them the opportunity to
develop a complex system of technology in an orderly and efficient manner, as proposed by the prospect development theory presented by Richard Nelson in chapter 3. In the case of LI-COR, patent protection of sequencing systems enabled it to negotiate the cross-licenses needed to develop its product fully. In both cases, private support has driven the development and dissemination of a research tool. The public and private sectors seem to have gained equally.
RESEARCH TOOLS IN DRUG DISCOVERY: Intellectual Property Protection for Complex Biological Systems
Research tools in drug discovery present an example of the difficulties in protecting intellectual property when technologies involve complex biological systems that lack discrete borders. The information is often broad and refers to general categories of matter, such as a class of neural receptors, rather than finite entities, such as the human genome, or specific techniques, such as PCR or recombinant DNA techniques. Controversies have emerged over broad patents, which some see as stifling research on and development of useful drugs and others see as critical to the translation of research knowledge into useful products. The focus of the discussion in the workshop was the tension between the dependence of small biotechnology companies on patents and the difficulties created when research on complex biological systems is restricted by a thicket of patents on individual components of the systems.
When research on a complex system—for example, receptor biology or immunology—requires obtaining multiple licenses on individual components of the system, the potential for paying substantial royalty fees on any useful application derived from that product can be daunting. "Royalty stacking" can swamp the development costs of some therapies to the point where development is not economically feasible. That is a problem particularly in gene therapy, where the most promising advances now are related to rare genetic diseases that present small markets.
Bennett Shapiro, vice-president for worldwide basic research at Merck Research Laboratories, argued that the central issue is not about patenting, but about access, about encouraging the progress of biomedical research. Problems can arise when access to related components of biological systems is blocked. For example, schizophrenia is often treated with compounds that suppress dopaminergic neurotransmission. Many such compounds, for example haloperidol, act nonspecificially and suppress the entire family of dopamine receptors. People who take those compounds for schizophrenia often develop other disorders some of which resemble Parkinson's disease, another disease involving the dopamine system. A rational approach to discovery of improved schizophrenia drugs would be to target specific dopamine receptors. But if different companies hold patents
on different receptors, the first step on the path to an important and much needed therapeutic advance can be blocked.
Shapiro commented that when only one company starts along the path of discovering a particular type of drug, its chance of discovering it is very low. Merck supports only a tiny fraction of total biomedical research, and it benefits enormously from research going on elsewhere in the world. It is in Merck's interest to share the results of its research with the understanding that they can be even more useful if placed in the pool of worldwide research resources.
It is interesting to compare that perspective on drug discovery with the early history of radio and television, other examples of complex systems of which many components were patented individually. In chapter 3, Richard Nelson noted that it was not until cross-licensing practices became widespread in the early development of radio and television that important advances that enabled broad access to the technology took place. When the intellectual property was sequestered in the hands of a few companies, the entire electronics industry remained sluggish. Of course, the progress of the industry overall must be balanced by the financial needs of individual companies. Shapiro noted that Merck has felt the need to become more energetic about patenting than it was years ago. For example, carrageenan footpad assays were used to develop non-steroidal anti-inflammatory drugs. The assays were in the public domain, and many companies used them to develop new drugs. Today, Merck would patent such an assay and use its patent position to trade with other companies for access to other research tools.
James Wilson, director of the Institute for Human Gene Therapy at the University of Pennsylvania, described his experience with the different ways in which patents on research tools are used. One is to block others from using the tools—to protect one's proprietary use—which he did not see as economical. Genetic therapy patents might not generate enough financial return to offset the investment costs. Wilson also suggested that genetic therapy patent files are only going to waste money in lawsuits brought against those patents. Second is to generate revenues for universities to support their infrastructures, although, as Lita Nelsen noted, most universities are not likely to earn much from patent revenues. The third is to barter so as to continue development without creating an economic disadvantage.
Like previous panelists, Larry Respess of Ligand Pharmaceuticals, argued that the chances of survival of a small biotechnology company would be slight without patents. He noted that the biotechnology industry is composed of small companies that have grown through venture capital and public offerings and that finance research through equity, not product revenues. The goal is to develop products and then evolve into an independent company.
Wilson also pointed to a dramatic increase over the last two to three years in the difficulty in transferring material between universities. Nelsen emphasized that university technology transfer managers are still learning. And many deci-
sions of the US Patent and Trademark Office (PTO) are controversial and under close scrutiny by those charged with managing intellectual property.
In commenting that "it is hard to know what the proprietary landscape is going to be, but it will be complex, whatever it is," Wilson summarized many of the workshop participants' comments.
Changes in Biotechnology Strategies
Respess discussed how R & D strategies for biotechnology have changed over the last twenty years. The biotechnology industry was born in about 1975 by Genentech, and most of the companies that followed Genentech pursued a similar strategy. Their objective was to produce and sell therapeutically-active large protein molecules, which was made possible by the availability of the Cohen-Boyer technology. The strategy was to discover and try to patent a gene for such a protein; it was hoped that the gene could be used to express abundant quantities of the protein. Some of the early examples are insulin, growth hormone, erythropoietin, and the interferons.
The advantages of that approach were that everyone knew that the products would be useful and that recombinant techniques were efficient for production, compared with earlier techniques of extraction from cadavers and tissue. Another advantage—albeit not from a scientific viewpoint—is that it is easy to sell to the investment community; it was a simple, easily understood model. Respess described the raising of capital in the early days of biotechnology as "unbelievable. You could found a company and, within a relatively short time, go public and raise many millions of dollars." However, those days are now past, in part because of the intrinsic limitation of large protein molecules: they are expensive to produce and to deliver to patients (they must be delivered by injection). The drug targets that are easy to identify have already been exploited.
A newer biopharmaceutical strategy emerged—not to discover large proteins or other large-molecule drugs, but to find other therapeutically active small molecules. These are the traditional targets of pharmaceutical research, but a biopharmaceutical company uses modern biotechnology and insights from molecular biology to get to the ultimate target product more quickly and efficiently. This approach has several advantages. The drugs are conventional and can typically be given orally, as well as by injection; they are relatively easy to manufacture; and the Food and Drug Administration is very familiar with such drugs, which makes it easier to get a new drug approved. The problem from a small company's perspective, however, is that it takes a very expensive infrastructure. Ultimately, synthesizing small molecules means making many molecules, and medicinal chemistry is very expensive. You have a tool, but you do not have any products in hand.
EXPRESSED-SEQUENCE TAGS (ESTs): Three Models for Disseminating Unpatented Research Tools
An expressed-sequence tag (EST) is part of a sequence from a cDNA clone that corresponds to an mRNA (Adams and others 1991). It can be used to identify an expressed gene and as a sequence-tagged site marker to locate that gene on a physical map of the genome. In 1991 and 1992, NIH filed patent applications for 6,800 ESTs and for the rapid sequencing method developed by Craig Venter, who was a scientist at NIH. The PTO rejected NIH's application and when Harold Varmus became director of NIH, he decided not to appeal. But controversy caused by the initial patent application continued. In 1992, Venter left NIH to form The Institute for Genome Research (TIGR), a nonprofit company, and William Haseltine joined the newly established private company, Human Genome Sciences (HGS), a for-profit company that initially provided almost all of TIGR's funding. The focus of the controversy then moved from the public to the private sector, and it changed from an issue about patenting research tools to an issue of access to unpatented research tools. Like many other research tools, ESTs fill different roles and some of the controversy has involved disputes of the relative importance of ESTs for uses other than research.
Two factors have contributed to the controversy over intellectual property issues in this particular setting. First is the perception that some of the participants have been staking out intellectual property claims that extend beyond their actual achievements to include discoveries yet to be made by others. There is no question that ESTs constitute a powerful research tool. Questions about the patenting of ESTs have focused on the criteria of utility. ESTs are of limited value without substantial and nonobvious development. Initially a public institution, NIH, proposed to patent discoveries that both scientists and some representatives of industry felt belonged in the public domain. More recently a private institution, Merck, has assumed the quasigovernment task of sponsoring a university-based effort to place information into the public domain. While other private companies have provided funds for public sector research, such as in the Sandoz-Scripps agreement, these efforts have not been with the expressed purpose of putting information into the public domain.
This is a particularly interesting case study, in part because it began as a controversy over patents—over what could be patented, what should be patented and what would be the effect of patenting. It has evolved into a controversy over the dissemination of unpatented information and the terms on which that information will be made available.
Different firms have taken different approaches to the dissemination of these unpatented research tools, thus providing a natural experiment with which to study three models for disseminating the same sort of information. The models all arose in the private sector, and we can assume that although each firm adopted a different strategy, they had the same ultimate goal of maximizing the value that
they could obtain from the information. Merck has put the information in the public domain, Human Genome Sciences (HGS) initially adopted an exclusive-licensing model, and Incyte adopted a broad licensing approach of offering nonexclusive licenses to its database to as many firms as would sign up. Putting information in the public domain limits opportunities to exploit it as a trade secret by controlling access to it. Patents, or the patent applications of private database owners, potentially limit the ability to use the information that is in the public domain if any patent rights are ultimately obtained.
HGS. The strategy of HGS has been to form a major partnership with the pharmaceutical firm SmithKline Beecham (SKB)2, with which it agreed to provide a three year exclusive license to its EST database. SKB has sublicensed its rights to a major Japanese pharmaceutical company, Takeda Chemical, Ltd., and HGS also has 200 restricted-licensing arrangements with university researchers. The TIGR database contains a limited portion of the data created by HGS and all of the data created by TIGR before April 1, 1994 which is when TIGR stopped work on human cDNAs. TIGR provides two levels of access to its EST databases. At the first level, an investigator is allowed access to sequences that are owned by HGS that overlap or are identical with sequences already in the public domain and for which public databases are available. At the second level, investigators are allowed access to about 70,000 sequences that are not listed in the publicly available databases (Genbank or the European databases). To obtain the second level, an investigator must agree to disclose any invention that is made at any time after access is gained. Furthermore, HGS or the Institute for Genome Research (TIGR) must be allowed at least six months to negotiate a licensing agreement.3 The public does not have access to the much larger HGS cDNA database.
Merck. Merck is interested in using the information from ESTs for furthering its research efforts. The Merck Gene Index was established to fill a public-access gap and was developed in partnership with established genome centers. Sequencing is carried out at Washington University, and the data are handled at the Los Alamos Laboratories. The international databases are a direct source of the information. A biotechnology company has taken all the clones into its distribution system and will freely distribute its materials. Other institutions, such as TIGR and Genethon, have entered sequences into this public database.
Incyte. Incyte's strategy has been to offer nonexclusive licenses to its database. As of the time of the workshop, six companies (Pfizer, Upjohn, Novo Nordisk, Hoechst, Abbott Laboratories, and Johnson & Johnson) have contributed in the aggregate, around $100 million, exclusive of contingency payments and royalty payments for access to this database. Even as the Merck data continue to be placed in the public domain, Incyte continues to sign up new subscribers; there seems to be continuing value for the subscribing firms to obtain access to one of the private databases. This strategy is interesting not only for what it says about the nonexclusive-licensing strategy but because this is the most current information as to the relative values of the private databases versus the public-domain database.
The Informational Value of ESTs Is Rudimentary
None of the participants disputed the value of ESTs as research information, but several commented on the rudimentary nature of the information. Having an EST in hand does not guarantee a practical strategy for obtaining the identity of the gene of which the EST is but a fragment. Furthermore, if the gene identified is unknown, there remains substantial investment in understanding its function. It has been successfully accomplished in many cases, and many specific strategies have been developed over the years for approaching this task. Nonetheless, it remains fraught with uncertainty. In 1995 the Human Genome Organization (HUGO) issued a statement on ''Patenting of DNA sequences'' arguing that the nature of sequence information is so rudimentary that to limit access to it is to impede development of medical advances.
Several uses have been suggested for genes and gene fragments to claim utility requirement for patent protections. They include the use of genes or gene fragments for categorizing, mapping, tissue typing, forensic identification, antibody production, or locating gene regions associated with genetic disease. However, each of those suggested uses may not be carried out without considerable further effort and additional biological information that is not inherent in the sequence alone. Many of the workshop participants concurred with the HUGO statement that without databases to provide further information, the informational value of ESTs themselves is very limited.
William Haseltine, CEO of HGS, noted that patent applications filed by HGS for ESTs involve considerably more than simply identification of the gene fragments and involve information about the stage of development and tissue type in which those genes are expressed. He further commented that the importance of the EST database is not simply that the fragments are identified, but that the database itself provides a high level of information.
The Value of ESTs Could Be Reduced by Limiting Access
Many of the workshop participants echoed the HUGO statement of concern that "the patenting of partial and uncharacterized cDNA sequences will reward those who make routine discoveries but penalize those who determine biological function or application. Such an outcome would impede the development of diagnostics and therapeutics." Both Harold Varmus and Gerald Rubin suggested that some researchers are likely to be discouraged from working on patented ESTs for fear that the patent holders would lay claim to their future discoveries, particularly discoveries about gene function, which are clearly of far greater biological utility than the identification of anonymous fragments and are more likely to have useful applications for human health.
Several previous reports have stated that research-tool claims should not be so broad as to block the discoveries outside of the patent (House of Commons Science and Technology Committee 1995, National Academies Policy Advisory Group 1995). No one at the workshop argued otherwise.
Fragile X syndrome, which is the most-common form of mental retardation, provides an example of how ESTs can contribute to human disease. The name refers to the fact that the X chromosome is easily broken. Caskey described how he, Steve Warren, and Ed Benustra used an EST to discover that the genetic defect involves multiple repeats of the nucleotide triple CGG. They went on to characterize the gene, and that provided the information necessary to develop what is now the most widely used diagnostic test for fragile X syndrome. When they made their discovery, the sequence information on the gene involved gave no information on function. It was investigators like Bob Nussbaum, and Dreyfus, at Philadelphia, who went on to identify the gene's function.
Caskey suggested that if speculative claims were permitted among a certain set of ESTs the rights of investigation to discover that gene would be denied.
James Sikella cited the example of the HIV patent, which is jointly held by the US and French governments. The patent has not been tightly restricted for investigational use. At the time of its filing, its sequence and functions were not described. Many discoveries about HIV have evolved from that sequence information, and Sikella noted that it would have been a disservice to the public if the sequence information had not been available as a general research tool.
The Human Genome Is Finite
As of this workshop, some 27,000–5,000 human genes were represented in the database. Humans are estimated to have about 80,000–100,000 genes, so that represents about one-fourth to almost half of the total. Tom Caskey predicted that as the database begins to be flooded with sequence information, there will be a higher stringency on patents and patent claims will be directed more toward functional aspects of the genes, rather than being primarily descriptive.
Caskey also described how the usefulness of the gene index has improved with the addition of more sequences. When the general location of a disease gene is known from genetic mapping, limited sequencing is an important strategy for finding the gene. By sampling the critical region, the small bits of sequences can be used to search for homologies in the gene sequence database. In this way, a previously sequenced gene or gene fragment can be identified as being located in a critical region. Such a gene is then a prime suspect for more detailed studies in those individuals carrying the disease. Initially, the success rate for this technique of finding disease genes in positionally cloned regions was only about 40%. As the size of the gene database increases, so does the success rate. This is, therefore, becoming a fast and facile method for identifying a disease gene in a critical region identified by genetic mapping.
Sikella suggested that the success of the Human Genome Project may be measured, in part, by how the knowledge that it generates benefits society. He emphasized the importance of making these benefits available in a cost-effective way.
The Advent of DNA Sequencing Presents Important Questions about Patentability
Leon Rosenberg commented that "although the debate seems to have cooled a bit, the issues surely have not been resolved." Tom Caskey of Merck and William Haseltine of HGS both commented that they have no quarrel with the current criteria for patents, but they express different views as how those criteria should be interpreted. Since the workshop, HGS has received patents on a number of ESTs with broader claims of utility than the initial EST patent applications filed by NIH in 1974. Whether this will influence the debate over ESTs is an open question. Caskey noted that after one has an EST, identifying the full length sequence cDNA is the obvious next step. And yet this rarely leads to precise knowledge of that gene's function. He predicted that the complete cDNA sequences might become the 1997 version of ESTs—that is, research tools which many people do not believe meets the full potential criteria of novelty, nonobviousness, and utility. Rosenberg suggested that "the biomedical research community has not yet truly grappled with the possibility that a large number of genes could be controlled by the rights of a relatively small number of parties who could not possibly hope to fully exploit their potential value." He suggested that if research tools are not made available to the scientific community and others, we will have to confront this issue directly, whether that requires changes in patent law or other equally drastic directions.
Adams MD, Kelley JM, Gocayne JD, Dubnick M, Polymeropoulos MH, Xiao H, Merril CR, Wu A, Olde B, Moreno RF, Kerlavage AR, McCombie WR, Venter JC. 1991. Complementary DNA sequencing: expressed sequence tags and human genome project. Science 252(5013): 1651–1656.
Beardsley T. 1994. Big time biology. Sci Amer. November: 90–97.
House of Commons Science and Technology Committee. 1995. Human genetics: the science and its consequences, Vol.1. London, UK: House of Commons
National Academies Policy Advisory Group. 1995. Intellectual property and the academic community. London: The Royal Society. 65p.