Robert Mullan Cook-Deegan, M.D.
Author, The Gene Wars: Science, Politics and the Human Genome (New York: W.W. Norton, 1994)

Doll's analysis rests on a few historical premises, by analogies to chemical patents and other experience with past patent issues. Doll asserts, in particular, that arguments for DNA patents "resemble those voiced 30 to 40 years ago when polymer chemistry was an emerging technology." Yes and no. The concerns are similar in that DNA patents pertain applied to compositions of matter (big chemical molecules) and some worry that patents may impede subsequent innovation. Those were indeed concerns with polymers. The disanalogies are almost as strong, however. There is a bounded set of human genes, selected from DNA transmission through evolutionary history, but no similar limit to chemical embellishments. In polymer chemistry, however, the variety is not as vast and this part of the analogy might still hold.

The technology is at hand to "tag" the vast majority of genes quickly. That is precisely what Incyte (Palo Alto, CA) and Human Genome Sciences (Rockville, MD) have done. The bounded set of human genes (estimated as 70,000 to 100,000, although this could be off by a factor of two or more in either direction) has combined with new technologies for rapid DNA sequencing to produce proprietary databases that purportedly contain fragments of the vast majority of human genes. Only the USPTO knows the exact figures, because patent applications in the United States are confidential until issued, but patents are reportedly pending on over a half million sequence tags, suggesting that most human genes are the subject of more than one patent application. Out of this mass of information will surely grow immensely useful drugs, diagnostics, and other inventions.

If those working on genes must license rights to use, make and sell them in recombinant or other useful form, then those two companies and others who pursued a sequence-tagging strategy will be in the cat bird seat, exerting considerable control over molecular genetics for the next decade or more, until their patents expire. The key concern is that, as Doll notes, "once a product is patented, that patent extends to any use, even those that have not been disclosed."

How many researchers working with human genes must obtain licenses from those holding sequence tag patents, gene patents, and method patents will be determined by the claims in patents that are issued and how courts interpret them in accumulating case law. The boundaries of infringement will be particularly important for gene fragments, but will also pertain to many other kinds of DNA patents (methods, full-length genes, cell lines, vectors, etc.). Until some sequence tag patents issue, the projected benefits and harms to research from such patents are speculative, but the speculation can be grounded in experience with other DNA patents. That experience is not as uniformly rosy as the Doll article implies.

DNA is not just a long polymer of A's, C's, G's and T's. It is an informational molecule in quaternary code that specifies protein and RNA structure and consequently cellular and organismal function. While DNA alone is not a complete specification of cell function, it is clearly important. Molecules with the same nucleotides but in different order specify completely different biological functions; the sequence dependence of information contrasts with chemical polymers in plastics. The requirements for disclosure of an invention in a patent must be correspondingly more precise for DNA than for polymer chemistry. The exact sequence of an entire molecule matters, not just a partial sequence plus a method for, say, using a sequence tag to fish a gene out of the genome. Federal courts have recognized this, as Doll notes in reference to the recent insulin patent litigation. Indiana's federal district court and the Court of Appeals for the Federal Circuit ruled that cloning rat insulin and describing how to clone it in humans did not permit a claim for human insulin because its cloning was not actually demonstrated. The new examination guidelines to be issued in the wake of this ruling should clarify the ground rules, but considerable uncertainty about the boundaries defining infringement of gene patents will persist.

The Doll article appears to presage action of the USPTO to allow patents on sequence tags; if so, the precise language of allowed claims will be critical. Doll notes that in some cases, a "second patent holder may have to obtain licenses from or pay fees to the primary patent holder, but is not prevented from obtaining the second patent." The degree to which subsequent inventions are subject to constraints by first patents will turn on the detailed claims language. If those studying genes must obtain licenses to pursue research, patents may well slow rather than promote progress in medical research. It will either cost more because of the "discovery tax" or future discoveries will be encumbered by past patent rights.

Focusing only on whether a second patent can be issued misses the point that scientists are most worried about. Molecular biologists may or may not care about whether they can secure patents on their work, but they are concerned about paying fees, and can become apoplectic about having to obtain licenses if those licenses seriously hinder their work. The vexing policy question is whether patent and other rights on current inventions will help or hinder progress; this does not reduce to whether a subsequent discovery can be patented. The balance here is between induced investment for research that would not take place without patents against the costs and constraints of licensing; the severity of the problem depends entirely on the discretion of patent holders. This is a collective problem but it will be settled one decision at a time.

If investigators or their institutions cannot abide the licensing terms offered by patent holders, as Heller and Eisenberg point out, the cost could be slower progress in research and fewer inventions to exploit commercially. Patents induce investment in promising technologies, but they can also be used to impede science. It is a question of balance, not a simple argument that patents arising from molecular genetics are good or bad, which is where so much of the public rhetoric about DNA patenting has long been mired.

The role of patents is vastly more important in pharmaceuticals and biotechnology than chemistry, microelectronics, or other sectors that are being used to guide patent policy. Few who work in or study the pharmaceutical sector deny that patents are essential to sustaining the long-term investment to move down the protracted and perilous path from drug discovery (research) through clinical trial and FDA approval (development) to commercial success. It is no accident that strong patent protection parallels strong performance among U.S. pharmaceutical firms, although foreign firms can also gain advantage from American policies friendly to patents. Other sectors depend far less on patents, and so patent policies are correspondingly a less decisive factor in commercial competition and innovative advance.

Pharmaceuticals and biotechnology are likewise much more dependent on federally funded academic research than any other major industrial sector. The medical research ecosystem thus displays more interdependence between industry, academe, and government. The chemical industry and microelectronics have long relied on a combination of trade secrecy, first-mover advantages, and patents with the balance tilting more heavily toward trade secrets and first-mover advantage in other sectors compared to pharmaceuticals and biotechnology. The importance of patents in pharmaceuticals and biotechnology and the unique mutualism between industry and academe imply that net social benefit is more sensitive to patent policy than in other economic sectors.

Doll takes solace that generic polymer patents issued three decades ago did not devastate the chemical industry. Molecular biologists may take less comfort in the historical comparison. Absence of calamity is but a weak argument for a policy, particularly if harms can be foreseen. Heller and Eisenberg's picture of an emerging anticommons is a clear expression of one reason biologists might be worried. The number of patents in molecular biology began to rise even before 1980, but has sharply increased since. At the same time, the prevalence and complexity of material transfer agreements, licensing terms, and constraints on disclosure of research results have also proliferated. A lot more molecular biology is going on, and it is increasingly encumbered by licensing and nondisclosure agreements. Patenting is part, but by no means all, of the problem.

The Heller and Eisenberg article is not just about DNA patents; it is about other patents as well as licensing of unpatented methods and reagents used in research. Their arguments do not reduce to patent questions. In order to get access to a valuable database or unpatented cell line, for example, scientists may sign agreements that concede rights to future discoveries. Their concern is much broader than patenting of sequence tags, although that controversy is a salient example.

No brief commentary can do justice to Heller and Eisenberg's intricate arguments, except to emphasize that they have the ring of truth. The terms of licensing property and intellectual property are at least as important as what gets patented. The USPTO does not determine licensing of patents, only what gets patented, so many important decisions are well beyond the reach of both the patent office and courts rendering decisions about patent litigation. The Doll and Heller/Eisenberg articles thus overlap, but Heller and Eisenberg's points will hold even if the debate about DNA patenting dissipates.

Norms governing licensing vary tremendously among companies and academic research institutions, and appear to be changing toward more aggressive enforcement of property and intellectual property rights, imposing more constraints on future discoveries based on licensing agreements. Many scientists labor under the myth that their research activities are in a category exempt from patent infringement allegations. Some are beginning to learn otherwise, as they get letters asking them to sign up for licenses on transgenic animals or research methods lest they risk being sued for patent infringement.

Terms in licensing agreements can seem deceptively innocent, constraining only future "commercial" uses of research results. Those terms, however, matter precisely when the research is likely to have direct benefit in the form of products and services, that is, have some prospect of improving health or other practical use, which usually entails some commercial activity. The only discoveries that encounter the "reach-through" provisions may be the ones that are actually useful. If those making individual decisions about licensing agreements are oblivious to their systemic consequences, medical researchers are apt to find themselves fighting an escalating battle, locked into a complex multi-player prisoner's dilemma.

Heller and Eisenberg are not arguing against patents, in fact they clearly value patents for end products such as therapeutic pharmaceuticals. They are concerned, however, about consequences of licenses that constrain medical research, and how patents are used by some who hold patent rights. Their arguments are not against patenting per se, and are thus distinct from most writing critical of patenting in biotechnology and drug discovery, but a caution that exclusive licensing, royalty-stacking, and reach-through rights may prove an impediment to what patent law was intended to promote--progress in science and the useful arts. Their concern, like Garret Hardin's, is about stewardship of resources needed to reach collective goals.

The upshot of both the Doll and Heller/Eisenberg articles is far from clear. Doll presents a case that DNA patents need not be disastrous, although the benefits of patenting are presumed rather than exhibited. The policies that encourage patenting for drug discovery and enabled polymer plastics to develop may not have the same salutary effects if the market is medical research rather than medical products and services. While most who study systems of innovation have long since abandoned the "pipeline" model in which research dollars are pumped in and commercial products flow out, Heller and Eisenberg's repeated allusions to "upstream" and "downstream" suggest that the general metaphor still pervades policy analysis. A river has many tributaries and is a pipeline open to the sky, but it still flows in only one direction. But producing new knowledge and developing new technology depends on bidirectional flow.

Sixty years ago, Warren Weaver of the Rockefeller Foundation coined the term molecular biology to describe a program for introducing techniques from the physical sciences into biology. Molecular biology has been highly dependent on technical advances from its inception, and most molecular biology has long had strong, if indirect and unforeseeable, connections between new knowledge and practical application. The patents claiming erythropoietin were research tools for years, covering vectors, protein production, and the gene for a protein of uncertain therapeutic import, but those patents subsequently became the source for manufacturing a drug generating over a billion dollars in annual revenue. The bivalent status of some molecular biological inventions threatens to complicate the upstream-downstream distinction central to Heller and Eisenberg's thesis. The process of innovation may be closer to a complex tidal estuary than a river, replete with reversing currents and eddies. Heller and Eisenberg do well to point us to some adverse and unintended consequences of conferring rights on materials and methods used in research; but further analysis will be needed to guide us in distinguishing upstream from downstream.

To Advertise   |   Find Products