patents, ethics, human life human germ line ...guenin.med.harvard.edu/documents/ethics of patents...

PATENTS, ETHICS, HUMAN LIFE FORMS 1

PATENTS, ETHICS, HUMAN LIFEFORMS

LOUIS M. GUENIN

Harvard Medical SchoolBoston, Massachusetts

OUTLINE

IntroductionExploitation and Monopoly of ClonedDNA Sequences

Anticipated Benefits of TransgenesisReservations Concerning Transgene-sisDefense of Human Germ Line Inter-ventionAutonomy and Patent Claims AgainstParents and ChildrenDistributive Justice and Patent Claimson Extracorporeal Compounds

Product Patents on Human DNASequences

Isolation-Purification Rationale forDNA Product PatentsUnpatentability of Nature and DNAPatentsWhat Should Suffice for a Biotech-nology Product Patent?Adverse Welfare Effects of DNA Patents

Alternative Incentives for Biotechnolog-ical Innovation

Measures for Holding onto the Avail-ability of Product PatentsCategorical ProhibitionsCompulsory Publication of DNASequence Data to Thwart PatentsSubsidiesExclusive FDA Approval for a Termof YearsHuman Methods Patent

Ancillary MechanismsCompulsory LicensingExpert Guidance

Implications of Patents in the ClinicIntroduction of Human SubstancesOutside the Germ Line

Human Germ Line InterventionConsistency of Policy for Plants and Ani-malsAcknowledgmentBibliography

INTRODUCTION

It might be supposed that morality oper-ates as a side constraint on patentability.On this view, even though a process or devicemight meet conceptual and scientific crite-ria for recognition as an invention, moralconsiderations might override so as to deny apatent. Or again, it might be held that moral-ity, genetics, and biotechnology so intertwinethat whenever we construct criteria of paten-tability with respect to ‘‘genetic inventions,’’we perforce impose some moral view.

EXPLOITATION AND MONOPOLY OFCLONED DNA SEQUENCES

Whereas U.S. patent authorities formerlydeclined to issue patents on gambling devicesand phony medicines, the U.S. Patent Code of1952 dissociates law and morality. It leavesto other laws the matter of restraining theuse of inventions. Efficiency alone commendsthis division of labor, since many patentsare never exploited. For example, Pasteurobtained a U.S. patent in 1873 on a yeastfor making beer. But, so far as we know, henever developed a commercial product (1). Ifimmorality of use does not count as an objec-tion to a patent application in general—evenif the proffered device be mischievous—should immorality of use count as an objec-tion to patents on life forms?

We might answer this with another ques-tion. If a patent on a gene or other humanlife form confers ownership over somethinghuman, and if, on moral grounds, we rejectclaims to own humans, are not patents ongenes and human substances illegitimate?This question probes not an invention’s usebut the appropriateness of the patent privi-lege itself. A patent lawyer will reply, reprov-ingly, that a patent does not confer own-ership, that a patent merely grants for aterm of twenty years the privilege to exclude

2 PATENTS, ETHICS, HUMAN LIFE FORMS

others from making, using, or selling aninvention. This reply does not end the dis-cussion. For various circumstantial reasonsany policy on biological patents brings moralcontroversy in its train. In the first instance,allowing commercial entities to wield evenlimited monopolies on things human willseem morally problematic to many observers.Some will regard such privileges as threatsto the autonomy of persons (as discussedbelow for clinical settings). Others will pointto various economic consequences of wield-ing patents, among them high prices andrestricted output of end products. When aDNA sequence patent issues but the paten-tee fails or declines to introduce a productpredicated on the sequence, the only bene-fit of the patent, if one may call it that, isto prevent the patentee’s competitors fromexploiting the sequence. It may be grantedthat for some the welfare loss of squander-ing an opportunity to improve beer produc-tion, especially for a mere scientific career, iscause for lament. But if a patentee shelvesa human gene patent and denies society anopportunity to develop beneficial drugs orto perform gene therapy, the cost may behuman suffering. As we shall see, good rea-sons obtain to resist the generalization thatbiological patents enhance aggregate wel-fare. In respect of the foregoing concerns, onehears not merely the voices of patent examin-ers and courts—unlikely arbiters of moralityin any event—but a variety of moral viewsheld among citizens to whom accountabilityfor governmental decisions is owed.

Because the decision to award a patentmay be publicly perceived as at least implic-itly a decision to condone any and all usesof the invention, it may behoove us first toresolve objections concerning morally prob-lematic uses of certain biotechnological inno-vations before we attempt a consensus onmonopoly of the innovations. If prudencecommends this two-part agenda in the UnitedStates, the European patent system demandsit. The European Patent Convention of 1963,whose criteria of patentability are otherwiseroughly coincident with the American, pro-scribes patents on inventions whose commer-cial exploitation would be ‘‘contrary to l’ordrepublique or morality.’’ This phrase was long

considered so vague as to lack teeth. But asadopted in 1998, the Directive on the LegalProtection of Biotechnological Inventions ofthe European Parliament (the ‘‘EuropeanDirective,’’ or ‘‘ED’’) declares unpatentable,on the ground that their commercial exploita-tion would be contrary to l’ordre publiqueor morality, the following: human germ lineintervention, ‘‘cloning’’ humans, commercialuse of embryos, and both somatic and germline genetic intervention in animals thatis ‘‘likely to cause suffering without anysubstantial medical benefit to man or ani-mal’’ (2). To this the ED curiously adds,‘‘exploitation shall not be deemed to be socontrary merely because it is prohibited bylaw or regulation.’’

Anticipated Benefits of Transgenesis

Mankind has bred plants and animals for mil-lennia. Since Mendel, breeders have exploitedknowledge of dominant and recessive alleles.Moral controversy about genetic engineeringstems not from manipulation by breeding,but from recombinant DNA. It is not thatrecombinant techniques clearly violate anymoral view in particular. Rather it is thecase that recombinant DNA technology posesquestions not previously raised within anytraditional moral theory.

Transgenesis consists in isolating a geneof one living being and inserting the geneat an early embryonic stage, before somaticand germ cells separate, in such a way thatthe gene enters the germ line of another liv-ing being of a different species. The insertionmay be accomplished by introducing the for-eign gene into (i) a retrovirus that infectsan embryo, (ii) a plasmid microinjected intothe pronucleus of a zygote, or (iii) culturedembryonic stem cells injected into the cav-ity of a blastocyst. The inserted genes areusually few and manifest themselves in onlya small subset of an animal’s phenotype.Transgenesis enables improvements in thegrowth, heartiness, and yields of animalsand plants as sources of food, vaccines, andother compounds, affords models for studyof diseases, the immune system, and generegulation and expression, holds promise fordirect therapeutic use in humans, allowsthe ‘‘pharming’’ of animal organs so that,


upon transplant to humans, they will notbe rejected, and allows production of cot-ton, plastics, and other industrially valuablecompounds. A vaccine-enriched transgenicbanana holds promise as a vehicle for sur-mounting economic and practical obstacles tovaccine delivery in many regions of the world.

Reservations Concerning Transgenesis

As encouraging as these prospects may be,they are not without their detractors. Objec-tions to transgenesis include the following.Even if genes insert at a targeted locus,in animals the effect of transgenesis maybe suffering, a theme frequently rehearsedin European discussions. The usual defenseof animal experimentation (as in the ED)adverts to collective human benefit. A netincrease in aggregate human preference sat-isfaction is all that need occur to satisfy autilitarian; the second form of Kant’s categor-ical imperative permits using even humansas means so long as they are not used solely asmeans. But risk-averse humans worry abouttheir own welfare in eating transgenic plantsand animals—even assuming full disclosurein the grocery store. To introduce a vaccineinto a banana crop raises questions aboutimposed risk-taking and paternalism wheninformed consent may not be feasible. Riskof human suffering is sometimes cited as aconsideration against human gene therapy.Risks about where and in how many copiesgenes insert and whether a procedure willotherwise work are chanced by any singlerecipient of somatic cell therapy; to this germline intervention adds the risk that an unto-ward result may burden future generationswho, it may be said, have no voice in what isdone to their ancestors’ genomes. Even whengene therapy achieves an intended result,the long-term effect may be a population lessdiverse, a gene pool that is diminished. Asuite of controllable genetic characteristicsmay eventually generate a canon. By refer-ence to that canon, persons lacking certaintraits may be treated by others as inferior.Perhaps indeed we shall cavort down a slip-pery slope from disease-related therapies tofrivolous enhancements. To engage in germline intervention, it is decried, is to ‘‘playGod.’’

Defense of Human Germ Line Intervention

A defense of human germ line interventionmight run as follows. Genetic engineeringmay be a way to improve man’s contributionas co-creator in God’s work (3). One mightargue that God would wish caring physi-cians to use it. Gene therapy will not inventdiscrimination, a practice already thrivingwith respect to many traits. We should notcount even against enhancement that some-one will be born with a trait more desir-able than another trait. Instead a tempta-tion to be invidious should remind us, as doKant and many religions, to recognize thedignity of each person. As for an effect ondescendants, that is not a new category ofmoral responsibility. Future generations arealready affected by innumerable influenceson one’s germ cells of how one lives. Onemay imagine a complaint of wrongful lifeby someone born with a genome adverselyaffected by something that went wrong in agene therapy procedure, but in another case,the same tort might be committed by fail-ing to attempt gene therapy. It would appeardubious to bar a physician from using anavailable method of averting disease in aconsenting patient’s offspring if, for example,the odds of the method’s success exceed thoseof any treatment after birth. It is also dif-ficult to expect a family or society to forgoeradicating a lethal gene if that be possible.A few decades from now, germ line interven-tion may be considered routine, its provisionthe duty of a competent physician, its inclu-sion a requirement of the health care a justsociety ought to provide.

The slippery slope to enhancement is notfairly ascribed to gene therapy (though per-haps to recombinant DNA). Recombinanthuman growth hormone, for example, isalready dispensed. At least at present, theimagined efficacy of gene therapy is lim-ited to diseases involving single genes, andamong those, to diseases mediated by reces-sive genes, because an inserted gene’s locusmay not be controllable and any dominantdefective gene remains in the patient. Onthe other hand, someday prospective parentsmay be speaking of a ‘‘designer child’’ poly-math 9′ basketball player. In any event a


distinction between ‘‘therapy’’ and ‘‘enhance-ment’’ may not be sharp or necessary. If aslippery slope connects therapy and enhance-ment, the transition from first violation of anyban on genetic enhancement to widespreadviolation thereof may be an avalanche. Atreaty presented for adoption by members ofthe European Union bans enhancement (4).As of this writing, human germ line interven-tion for enhancement purposes is not feasible,but upon its availability, one may expect thefollowing. Though a government may bangenetic enhancement, as soon as one personmanages to procure an enhancement, othersacting rationally will likely rush to procureenhancement in order to remain competi-tive (5). These aspirants will either violatethe ban or migrate to a sovereignty that lacksone. Sovereignties will likely behave the sameway, rushing to follow the first innovator forfear of being dominated by superiors. Unlessone contemplates intrusive ‘‘audits’’ compar-ing parental and progeny DNA, a ban onenhancement seems unenforceable.

The foregoing brief account reveals moralconcerns about uses that insinuate them-selves into discussions whether to approvemonopolies of uses. We may now turn tomoral concerns about a patent privilege itself.

Autonomy and Patent Claims Against Parentsand Children

The effect that a patent might exert on indi-vidual autonomy may be studied through adramatic example. This first requires that weexplain the rationale for patents on transgen-ics.

Patented Transgenic Organisms AlthoughPasteur’s yeast long before gained a patent,modern recognition of a nonplant life form aspatentable occurred in a decision with respectto a bacterium into which were introducedplasmids rendering the bacterium capable ofdecomposing oil (6). Patentability was laterconfirmed for multicellular organisms (poly-ploid oysters) (7). A furor ensued over ethicalconcerns, and for a time, the U.S. Patent andTrademark Office (‘‘PTO’’) imposed a morato-rium on animal patents. Thereafter the PTOissued a patent on the Harvard mouse (8).

Designed for the study of cancer, the Har-vard mouse contains DNA sequences, com-prising an oncogene such as myc and a pro-moter, that effect a high proclivity to formtumors when the mouse encounters carcino-gens. The introduction and expression of themyc gene in the mouse was innovationary.One could not have presumed that a zygoteacquiring the oncogene would survive itsinsertion and expression, or if the zygotedid survive and a mouse were born, thatthe offspring would be fertile. The Harvardmouse invention was exhibited for patentpurposes by deposit of DNA in a plasmid.But the patent extends not merely to theinserted DNA, not merely to the oncogene’sintroduction and expression, but to the whole‘‘oncomouse.’’

Why should a patent embrace an animal?Two arguments might be mustered. First,there has endured throughout the historyof patents the notion that patents shouldnot be available on nature’s extant trea-sures—in a phrase attributed to ThomasJefferson, on any ‘‘product of nature’’—butshould be available only on what humansmanufacture. The discovery of uranium gar-nered no patent, but the PTO issued a patentto Glenn Seaborg on curium and isotopes ofamericum, transuranic elements believed toexist on earth only in a cyclotron or reac-tor as a result of human efforts. Of coursesuch elements may be abundant in stars. Aprecise statement of patent eligibility wouldnot exclude from patentability every natu-rally occurring thing. Rather we may statepatent eligibility by the following proposi-tion, which we may call the ‘‘unpatentabilityof nature’’: to be eligible for a patent, a thingmust be such that there obtains only a verylow probability that, without human inter-vention, the thing exists near the surface ofthe earth or on other astronomical bodiesto which humans travel. What constitutes a‘‘very low’’ probability requires specification.Since patent lore speaks confidently of thingsthat ‘‘exist’’ or ‘‘do not exist,’’ no guidance onprobability may be found within it.

It is logically possible for there to evolvean organism whose genome is identical tosome modern transgenic organism. Exchange


of genes across species occurs and any muta-tion is possible. Yet the probability may beextremely low that a creature will containgenes of two given organisms that do notmate. In such case a patent examiner maytreat a transgenic genome as if it does notnaturally occur. An animal possessing suchgenome is then seen not as an unpatentableproduct of nature but as a patentable ‘‘manu-facture’’ or ‘‘composition of matter’’ (9). Withrespect to the European Patent Convention,it would be said that such an animal is notan unpatentable ‘‘variety’’ (10). In general,the fruits of breeding programs are consid-ered varieties, but transgenic animals arenot considered varieties because transgene-sis was unknown in 1973 when the Europeanconcept of a variety was introduced.

Second, when introduced into a recipient’sgerm cells, transgenes pass to descendants.A transgene will not be expressed in all off-spring of the first generation, but those inwhich it is expressed will be selected forfurther breeding. Were a patent to coveronly cells expressing a trait, it would notcapture the invention, which, by virtue ofbeing genomic, appears in every cell. Werea patent to cover only an inventor’s processof introduction and expression, anyone whopurchased one transgenic animal could breedothers without infringement; natural repro-duction is not the same process as laboratorytransgenesis. A purchaser of transgenic agri-cultural livestock could breed the livestockthrough unlimited generations. Hence inven-tors are accorded the protection of a patent onthe transgenic genome, which is effectivelya patent on the animal. Breeding descen-dants of a patented transgenic animal with-out license would just as clearly constituteinfringement as would duplicating a patentedlaboratory process for inserting transgenesinto an embryo. The patent system assimi-lates reproduction, whether natural or artifi-cially aided, to ‘‘making’’ a duplicate.

The Human Qua Infringement A moment’sreflection reveals that if the foregoing twogrounds (a claim to originate a manufacture;self-reproducibility of a recipient) entitle anarchitect of transgenesis to a patent on recip-ient and progeny, then in the case of human

germ line intervention, infringement claimswill lie against the birth and existence ofhumans. To call human birth or life a ‘‘patentinfringement’’ seems perverse. But on whatprincipled grounds should we reject suchclaims? Even in the somatic cell case, as sci-entists perfect the manufacture of yet morehuman enzymes and other proteins, as theyprogress to substantial tissues, should societycontinue to grant patents on human ‘‘parts’’?Manufacture of a liver or other major organ,or someday even of a brain, may confoundprevious thinking.

Abjuring the Human Qua Infringement Toresolve the solecism of the human quainfringement, we may reason as follows. Wedo not imagine infringement claims againstany plant or animal. Instead we recognizeclaims against people who control breeding.We do so because we recognize a farmer’sownership of plants and animals. When the‘‘designer’’ of a transgenic organism appliesfor a patent, the contest concerns only whichhumans (or corporations they represent) ownproperty in the nonhuman species. When theissue is which of two humans owns a human,we say that humans own themselves. They donot own each other. Human births, we hold,are not analogous to breeding, to manipula-tion by owners of mating subjects. Hence wemay decline to recognize property rights inhumans.

The premise that humans do not owneach other, that we each enjoy a ‘‘bodyright’’(11), is not categorically held in all societies,and given that the common law describes anadolescent’s maturity as ‘‘emancipation,’’ per-haps it is not unequivocally held anywhere.Defense of the premise often comes round tosome distinction between humans and ani-mals. According to a Cartesian distinction,man is a singular creature possessed of rea-son. According to Kant, only man and angelsare capable of reason. To say that a humanbeing’s existence infringes property rightswould seem inconsistent with the second andthird forms of the Kantian categorical imper-ative, which together enjoin that we treateach person not solely as a means but asan end-in-himself in a kingdom of ends. Ifwe allow an ownership claim on a person,


we condone treating the person solely as ameans. We condone interfering with the per-son’s autonomy. Were someone to assert anownership claim that purportedly extendedto only part of a person, the claim wouldappear indistinguishable from a claim onthe whole. Bodily parts are integrated. Forthe same reasons, conception and birth, theinstantiation of human nature, may be heldimmune from claims of others. We may alsosay that conception and birth are private.

Distributive Justice and Patent Claims onExtracorporeal Compounds

Although we may thus deny the permissibil-ity of exerting dominion over, impairing theautonomy of, or disrespecting an individual,a different case, actual in biotechnology, isthe following. There is adduced a substancethat is human in the sense of being found inthe human species, but which has been madeoutside the human body and is not ascribableto any individual. Were a patent to issue onthat substance, the patent would not appearto interfere with any individual.

To this it must be added that it is noteasy to steer clear of the DNA sequences thatdistinguish an individual or that make anyindividual akin to another. Only about threemillion of the three billion base pairs in thehuman genome account for individual dif-ferences, but the genetic code is redundant,the most interesting traits are polygenic andbeyond present understanding, and mutationnever ceases. For now, individual identity, tothe extent it is genetic, is genomic. We havenot demarcated a nonindividuating subset ofthe genome that we may cede. We do knowthat individuality is greatly affected by a rel-atively small number of regulatory sequencesthat control which genes are expressed. Suchsequences are indeed used in biotechnologymanufacturing unless a bacterium’s or otherhost’s regulatory sequences effect expression.

Let us assume for the moment that it ispossible to grant monopolies on proteins andDNA sequences without there resulting anyinterference with the autonomy of any indi-vidual. The PTO effectively allowed as muchwhen it began to grant patents on humanDNA sequences despite its earlier declara-tion that a patent on a human would violate

the prohibition of slavery in the ThirteenthAmendment to the Constitution of the UnitedStates. It contravenes common usage to saythat a nonpossessory interest in a protein orgene constitutes slavery.

Were one confident that a system of lim-ited monopolies would lead to advances thatprevent or alleviate human suffering, onemight decide that the conceptual coherenceof the patent system should give way to thepromotion of aggregate welfare. If patentson genes contravene the unpatentability ofnature, so much the worse for that premise.It seems that one perforce resorts to thatstance for the defense of patents on plantantibiotics. One might go so far as to say thatthere should issue any patent, even if thepatenting process is purely piecemeal, thatresults in net aggregate welfare gains.

Whether compromise with intellectualpurity be systematic or piecemeal, and evenassuming any contingent results that anadvocate of such conceptual indulgence pre-dicts, this talk of welfare effects presupposesa criterion for discerning improvements inwelfare. That in turn implicates some versionof a social welfare function. A social welfarefunction is a function that yields or inducesa positioning of possible resource allocationson which one may predicate a claim such as‘‘α is a welfare improvement over β.’’ Thespecification of a social welfare function isthe main problem of distributive justice. Forthis reason, what begins as a moral problemconcerning respect for personal autonomy,which arguably is tractable by virtue of theability to eschew interfering with any indi-vidual, endures as a challenging problem ofcollective morality.

To the extent that patents are distribu-tive mechanisms, this problem arises withinan economy in respect of any patent. Butconcerns abound with respect to the wel-fare effects of biotechnology patents. Somemay espy unacceptable burdens and risks forthe human species as a whole from variouspatents on molecular or structural humanlife forms. For instance, when a patent ownersets what seems an exorbitant price for avital drug, one observes an arguably undesir-able effect of market power conferred by anunqualified government-created monopoly.


PRODUCT PATENTS ON HUMAN DNASEQUENCES

Supposing that the prospects of collectivebenefit or some other morally persuasive con-sideration have justified the alteration anduse of life forms, why confer exclusive controlon one party? The orthodox quid pro quo ofa patent is that, instead of keeping an inven-tion a trade secret, the patent teaches thedetails. When a patent expires, the inventionwill be in the public domain, and even duringits term, what others learn from its teach-ing may foster other innovations. Whetherthe patent’s revelations are in fact valuablewill depend on whether one may easily inferthe invention by reverse engineering (12).An alternative and more familiar rationaleasserts that a patent provides an incentivethat fosters ingenuity and effort. Or as Ben-tham put it, ‘‘He who has no hope that he shallreap will not take the trouble to sow’’ (13).

Isolation-Purification Rationale for DNAProduct Patents

Organic compounds found in humans are,ipso facto, naturally occurring. Suppose thatan organic chemist discovers a way to synthe-size a protein in a purified form not found inhumans. If the protein appears extractablefrom another organism, then perhaps weshould not regard the protein as distinc-tively human. But in fact the human versionof a given protein is unlikely to be identi-cal with that of another organism. Throughmutations in duplicate genes, species haveevolved a variety of genes coding for dif-ferent versions of proteins that we call bysingle generic names. The notion of isolationand purification (a creature of case law, notstatute) was popularized by product patentson inorganic chemicals. (In patent parlance,a ‘‘product patent’’ is a patent on a thing asopposed to a patent on a process.) The notionwas then borrowed in support of patents onthe products of biological processes, includingpurified human adrenalin, prostaglandins,vitamin B12, and, most recently, human DNAsequences. For the last, investigators’ counselhave persuaded patent examiners that inves-tigators have ‘‘isolated and purified,’’ whichis to say cloned, human genes.

There is reason for scepticism whether apatent must be available in order to induce agiven result. ‘‘The large amount of researchthat has already occurred when no researcherhad sure knowledge that patent protectionwould be available,’’ noted the Supreme Courtof the United States in affirming the patenton the oil-eating bacterium, ‘‘suggests thatlegislative or judicial fiat as to patentabilitywill not deter the scientific mind from prob-ing into the unknown any more than Canutecould command the tides’’ (6). There arrivedfor filing a spate of plant patent applicationsin Europe, many presumably from Europeancompanies, quickly after the first plant patentwas allowed there in 1989, which suggeststhat the research had long before been done.And rivals face effort and expense to fol-low a first entrant into the market. In theUnited States the lead time that an imita-tor of a drug or medical device would needto obtain approval from the Food and DrugAdministration (‘‘FDA’’) for selling the prod-uct provides a period of de facto exclusivityto the product’s originator once the originatorobtains approval. There may also be observeda tendency after a favorable experience forphysicians to continue prescribing, and con-sumers to purchase, the first drug of a genre.Assertions about the necessity of incentivescan be facile, but evidence is lacking.

Inventions concerning nonliving phenom-ena make use of materials that mankindhas long exploited with an aplomb per-haps attributable to the mistaken belief thatwe cannot alter earth’s vastness. Biologi-cal inventions obviously effect alterations ofnature. If we approve the engineering of someprotein, we might view the circumstance thatit is found in humans as a reason againstmonopoly. Hence the isolation–purificationrationale originated for chemistry in gen-eral cannot be assumed to carry the dayfor human compounds. One might add thatto adopt such construct for humans wouldnot follow the model of chemistry faithfullyenough. Patents have been granted to thosefirst to synthesize chemicals, but courts tendto find evidence that chemical patents havebeen infringed only insofar as a patentee’sprocess has been copied.


The probable benefits of recombinant inno-vations may in a given case outweigh theacknowledged detriments. But each wave ofinnovation evokes a new comparison of risks,costs, and welfare gains.

Unpatentability of Nature and DNA Patents

An enduring challenge for the molecular biol-ogist is to understand a disease or bodilyfunction, to identify a protein related to it,to ascertain the nucleotide sequence of agene coding for the protein and the protein’samino acid sequence, to locate the gene on achromosome, and to explain the gene’s reg-ulation and expression. Once sequenced, acloned gene may be preserved as comple-mentary DNA (‘‘cDNA’’) in a vector. Whenvectors transform and infect, not only dothey multiply an inserted gene, but the genecan integrate into the transformant genome,causing such host to produce the protein forwhich the gene codes. It is by growing suchtransformants under suitable conditions thata biotechnology manufacturer may produce aprotein in high volume.

A typical patent claims at least threeinventions: (a) an isolated and purified DNAsequence encoding some protein, (b) any vec-tor that contains that sequence and anytransformed host possessing that sequence,and (c) one or more processes. The patentsystem indulges the notion that the cloningof genes produces ‘‘inventions’’ that do notnaturally occur. By contrast, a detailed exam-ination of what occurs in the laboratory,though providing ample evidence to confirmour admiration for scientific achievements,reveals no entity that mankind creates (14).

A gene encoding any human protein existsin nature. It is embodied in a chromosome. Itstranscript also exists in mature messengerRNA (‘‘mRNA’’), a single strand of DNA-complementary nucleotides. Transcription ofDNA into mRNA, followed by splicing thateliminates introns, is nature’s own ‘‘isolation’’of the coding sequence (with uracil in place ofthe thymine of DNA). This alone seems to tellagainst the argument that only an inventorhas achieved isolation. Is ‘‘purification’’ thenthe inventor’s trump over nature? The pro-cess of making cDNA is not thought to occurnaturally in humans (though many viruses

that infect humans make DNA from RNA).But once a gene is known, the laboratoryprocess of making cDNA can be routine. Per-haps then vector and host deserve credit asthe ingenious embodiments of purification?Where a host and a donor of foreign DNAare members of species that do not mate,asserted Stanley N. Cohen and Herbert W.Boyer in teaching the first recombinant pro-cess, a recombinant host ‘‘could not exist innature’’ (15). This imagined impossibility ofcourse is an exaggeration. Any mutation ispossible. As Bernard D. Davis observed indebating the hazards of recombinant DNAresearch, bacteria absorb foreign DNA fromlysed cells of their host, including the humangut. The rate of bacterial absorption of foreignDNA is low, but the number of bacteria in thegut is enormous, and bacteria have thrivedfor millions of years. There is some proba-bility for any given human gene that it hasalready occurred in a bacterium. We may notobserve that gene today, since the gene mayhave conferred no selective advantage on thebacterium, and the strain vanished (16). Onthe other hand, what is the likelihood thata plasmid or bacterium naturally containsa given human gene? That probability mayapproximate that of unicorns existing. Cohenand Boyer evoked a sense of mythologicalimprobability when they called an alteredplasmid a ‘‘chimera.’’

When a patent claims a chimera andhost, the two are usually notable only inone respect: they contain a DNA sequencethat the applicant purports to have invented.The plasmid and host are effectively thesequence’s housing and factory. Once aninvestigator has selected a sequence asdescribed above, the process of cloning it, andhence the ‘‘invention’’ of the sequence in avector and transformed host, is mechanistic.Advanced techniques for sequencing proteinsmay make straightforward the selection ofprobes and primers, and hence the discoveryof a known protein’s gene. The ‘‘invention’’ ofa protein variant may also follow straight-forwardly from a variant, experimentallyachieved, of a gene sequence. An investigatorwho finds naturally occurring genes and pro-teins merits accolades. But it remains to be


shown why a patent should issue, and if so,on what.

What Should Suffice for a BiotechnologyProduct Patent?

As necessary conditions for award of aproduct patent, consider the following: (1) aclaimed invention is such that it is highlyimprobable that we shall find it as such onearth or on any other astronomical body towhich human travel is possible, and (2) theinvention is ingenious. The ‘‘as such’’ phrasein (1) would allow some particularly conve-nient forms (e.g., a vector with a foreignDNA insert) to gain recognition apart froma natural form. In (2), ‘‘ingenious,’’ whichshares an etymological root with ‘‘genet-ics,’’ is a placeholder for what constitutesan invention, of which more later. Sequencesfail (1) if they are found in chromosomes andmRNA. Chimeras, transformed hosts, andcDNA meet (1) but fail (2) when they aremere mechanistic steps from discovery of asequence. The vectors and hosts of microbiol-ogy are not as fastidious as the species unitedin mythological chimeras: they accept DNAinserts regardless what creature originatesthem.

An ‘‘artificial gene’’ may satisfy both (1)and (2). In theory such a gene may be con-structed of any codon-containing sequenceone likes; in practice the amino acid sequenceof a protein may inspire the sequence. Somecaution may be needed in characterizing asequence as ‘‘artificial’’ because if a sequencecodes for a human protein, then either thatsequence, another differing only by substi-tution of alternative codons for the sameamino acids, or yet another that is insignif-icantly different exists somewhere in thegenome. Because of the phenomenon of over-lapping genes, the sequence may also bepart of another gene that its discover hasnot even envisioned. On the other hand, itmay be that an ‘‘artificial gene’’ meets (1) andthat its gene product is in some sense supe-rior. (The artificial gene product might, forexample, lack contaminants usually foundin the natural gene product.) Not every suchsequence will be ingenious. Courts have oftendeclared DNA sequences inferred from pro-tein sequences to be obvious (17, p. 50).

In view of patents on algorithms—a depar-ture from previous conventional wisdom thatideas are not patentable—one might appealto the notion of patents on information asa defense of DNA patents. But this defensewould seem to fail insofar as any informa-tion encoded in cDNA is encoded in naturallyoccurring DNA and mRNA.

Adverse Welfare Effects of DNA Patents

If the autonomy of no one in particular isthreatened by a product patent, the auton-omy of everyone together might be.

The PTO in 1987 granted a product patenton isolated and purified natural erythropoi-etin. Merely four months later it granteda second patent to another party relatingto a recombinant DNA technique for mak-ing the protein. The second patentee hadcloned the gene after screening a genomicDNA bank with two sets of probes. It thenproduced the protein in transformed ham-ster ovary cells. The first patent blocked theinvention of the second. This portended thatpatients would be deprived of a recombinantmethod of producing erythropoietin in highvolume at low cost. As a group, patientswere saddled with the first patentee’s pro-duction method (extracting extremely lowyields of the protein from thousands of gal-lons of urine). When, four years later, thefirst patent was invalidated on unrelatedprocedural grounds, the second became a bar-rier against any better recombinant processemploying the claimed sequence (18). Simi-larly did a biotechnology firm discover thehuman gene for factor VIII:C by probinga human cDNA bank, inserting the genein plasmids, transforming hamster kidneycells with the plasmids, and producing fac-tor VIII:C. This recombinant advance wasblocked by an earlier patent on factor VIII:Citself (19). The patentee’s process not onlyrequired enormous amounts of donated bloodplasma for a small yield, but in contrast withthe recombinant method, it risked contam-ination. Contamination was a critical riskbecause many hemophiliacs who receivedcontaminated factor VIII:C died of AIDS.The recombinant’s manufacturer protestedunsuccessfully that the patentee had notinvented factor VIII:C (though, given the


chance, the manufacturer might have arguedfor its own invention of the recombinant).The erythropoietin and factor VIII:C episodesillustrate how product patents may frustratesociety’s interest in encouraging, at the samerapid pace at which biomedical research isotherwise moving, helpful innovations in pro-cesses for making therapeutically valuablehuman compounds.

Farmers raise the specter that, burdenedby the high cost of patented animals andcrops, they may turn to cheap unpatentedstrains, that crops will become less diverse,and that more crops will succumb to pests.In transgenesis, often an investigator cannotcontrol the place within a genome at which aforeign gene inserts or the number of copiesthat insert, or in the case of plants, the weedsor other unwanted plants to which a trans-gene may migrate via airborne pollen. Wealso have reason to rue ‘‘blind promotion oftechnological innovation’’ (20). Agriculture,after all, is an industry afflicted with overpro-duction. A patent granted by the EuropeanPatent Office on all manner of geneticallyengineered soybeans has been criticized onthe ground that soybeans are among theworld’s most important crops and monopolyof soybeans will threaten ‘‘world food secu-rity.’’ Suppose that a patent on a critical crop,organism, or substance has been conferred onan enterprise that becomes bankrupt. Or sup-pose that the patent is acquired by some for-eign entity that is involved in internationalintrigue, that uses the patent as leveragefor some disreputable purpose, or that oth-erwise seems to control output contrary tothe common good. As exemplified by experi-ence with the anti-AIDS drug zidovudine (or‘‘AZT’’), the price of a patented product is amonopolist’s price.

Scientists have become acutely aware thatavailability of patents on DNA sequencesmay be generating a patent race that mis-allocates resources and delays publicationof results. This would run contrary to thehope underlying the Human Genome Projectthat disseminating chromosomal mappingand sequence data will foster growth in col-lective knowledge. It took four years aftera gene implicated in breast cancer, BRCA1,was mapped to chromosome 17 before one of

twelve rival collaborations found the gene, afeat they all recognized as a ‘‘discovery’’ (21).Yet the winner immediately sought a patenton BRCA1 and related diagnostic processes.About a year later, one of the competinggroups contributed to a public database thesequence of a large portion of chromosome13 where BRCA2 was thought to repose. ‘‘Itwill not be helpful to medicine,’’ the scientistJohn Sulston was quoted as saying, ‘‘if, bythe year 2003, control of every single geneis tied up by one company or another fortwenty years. That would be an enormousball and chain. . . .[F]or the good of humanity,we should try to keep these things in the pub-licly exploitable domain’’ (22). The group con-tributing the chromosome 13 sequence dataurged that DNA sequences be public infor-mation. Within a month thereafter, BRCA2was found (23). This seemed to exemplify therapidity of progress when results are shared.Thereupon the discoverers of BRCA1 filed fora patent on BRCA2, launching a dispute overwho found BRCA2 first. Seemingly ignoredwas the untenability of claiming to inventparts of nature’s storehouse.

One cannot dismiss objections to prod-uct patents as the outpouring of any sin-gle, disputed moral view. Even without anappeal to morality, it may be argued on exclu-sively scientific and economic grounds thatpatents on human DNA sequences violate theunpatentability of nature. Many moral viewsassign significance to the aggregate welfareconsequences of that violation. In such casethe moral case against human DNA sequencepatents reprises the scientific.

ALTERNATIVE INCENTIVES FORBIOTECHNOLOGICAL INNOVATION

Measures for Holding onto the Availability ofProduct Patents

The legal criteria for patentability are thata ‘‘process,’’ ‘‘manufacture,’’ or ‘‘compositionof matter’’ be ‘‘new,’’ ‘‘useful,’’ and ‘‘nonobvi-ous’’ (24). According to a conservative articleof faith espoused by patent practitioners,these criteria possess such protean qualitiesas to suffice for the resolution of all questionsthat arise from time to time. The criteria


need only be interpreted by the courts. Inreply to this, it must be said that, underprevailing interpretations, ‘‘new’’ and ‘‘use-ful’’ erect only minimal thresholds. ‘‘New’’eliminates from patentability only what hasalready been published. ‘‘Useful’’ eliminatesfrom patentability only the utterly useless,a rare creature among proffered inventionsanyway. (Scientists at the National Institutesof Health, NIH, dramatized the weaknessof the ‘‘utility’’ requirement in 1991–1992when they ostensibly satisfied the crite-rion by citing a seemingly trivial use forparts, ‘‘expressed sequence tags,’’ of cDNAsequences. The applicants conceded igno-rance about the feature of usual biologicalinterest, viz., what the sequences encode orregulate, and ventured only that the tagscould be useful as genetic markers, primers,or probes in diagnostic kits for unnameddiseases. But any DNA sequence may be amarker in genomic mapping.)

To resolve a question of patentability, twosobriquets, ‘‘manufacture’’ and ‘‘nonobvious-ness,’’ must carry most of the load. In factnonobviousness must do all the work. For itis considered settled in U.S. patent law thatcloned DNA sequences, as fruits of ‘‘isolationand purification,’’ constitute a patent eligiblegenre. Being a ‘‘product of nature’’ is now seenas no impediment to patent eligibility; thequestion is whether a sequence constitutes anew, useful, and nonobvious manufacture orcomposition of matter (25). Thus stood on itshead is Jefferson’s use of ‘‘product of nature’’for the unpatentable. But the point is onlysemantic: since every extant thing’s ingredi-ents are naturally occurring raw materials,every extant thing may be called a ‘‘prod-uct of nature’’ in some sense. The semanticpoint entails no practical consequence if someother provision insures the unpatentability ofnature. (As defined earlier, the unpatentabil-ity of nature is the premise that to be eligiblefor a patent, a thing must be such that thereobtains only a very low probability that with-out human intervention, the thing exists nearthe surface of the earth or on other astro-nomical bodies to which humans travel.) Wemight think that the statutory term ‘‘invent’’secures the unpatentability of nature. Butinstead for the domain of biotechnology

though not for others, we observe patentexaminers and courts effectively either reject-ing the unpatentability of nature or exhibit-ing remarkable restraint as they construe thepremise. The DNA sequences that they pro-nounce patentable are sequences that chro-mosomes of living beings contain. The onlyapparent way to reconcile this with someversion of the unpatentability of nature is toemphasize that a given cDNA sequence corre-sponding to a chromosomal sequence differsfrom the chromosomal sequence insofar asthe chromosomal sequence is littered withintrons. Still it must be said that the chromo-somal sequence includes the cDNA sequence.That is to say nothing of the chromosomalsequence’s uninterrupted transcript in theform of mRNA. Courts and patent examinerskeep faith with only a weak version of theunpatentability of nature.

Given that DNA sequences are recog-nized as a patentable genre, whether a givensequence garners a patent turns on whetherthe sequence is deemed obvious. When courtsfirst struggled with arguments about recom-binant DNA technology, it seemed obviousthat what was obvious was not obvious.As courts came to recognize recombinanttechniques as commonplace, they bent overbackwards to conclude that newly discov-ered cDNA sequences were nonobvious. Ifthe prior art did not enable a method offinding a proffered sequence ‘‘with a rea-sonable prospect of success,’’ a court wouldpronounce the sequence nonobvious. (Thismove vindicated a patent on the sequenceencoding erythropoietin, a sequence found byscreening a genomic DNA bank with two fullydegenerate sets of oligonucleotide probes.)As further reasons to sustain a verdict ofnonobviousness, courts have even recognizedcircumstances extraneous to the intellectualprocess of discovery and invention, includingcommercial success, long-felt need, failure ofothers, unexpected results, and the scepti-cism of rivals (17, p. 19). As critics would haveit, one influential judicial decision saves theday for cDNA patents only by tortuously con-struing ‘‘obvious’’ so as effectively to declarepatentable per se any DNA sequence foundto encode a protein (26). According to oneobserver, the obviousness of many purported


cDNA inventions is betrayed ‘‘in the very atti-tude of the persons skilled in the field. Today,if a researcher discovers a new protein and itsprobable properties, he usually does not pub-licize the information until he has found thecorresponding gene. How to explain this ina community whose motto is ‘publish or per-ish’ save that it would be obvious to anotherresearch team to pick up the information,and clone the gene?’’ (17, p. 90). In hopesof securing future DNA patents against atide of progress that may render ever fewercDNA sequences nonobvious, it has been sug-gested that the nonobviousness requirement,to the extent not already emasculated by theaforementioned judicial decision, might beweakened. If the steeplechase jump provestoo high for the average contestant, lower thebar. Where a nucleotide sequence is itself adrug, as with anti-sense RNA or the use of aDNA sequence to achieve expression withoutintegration into the genome, it has been sug-gested that obviousness might be replacedwith superiority over prior art in therapeuticefficacy (17, pp. 141–143, 148). Without somesuch move, it is urged, future application ofthe obviousness standard may thwart theavailability of product patents on the expec-tation of which the biotechnological industryarose. Here an appeal is made to the biotech-nologist’s familiar prediction that a worldwithout DNA patents will be a world withouttherapeutic innovation. But rehearsing thatprediction does not provide evidence for it.

The issue remains whether, all things con-sidered, more DNA product patents shouldissue. By virtue of considerations mentionedearlier, the answer may be in the negative. Insuch case, what might replace such patents?

Categorical Prohibitions

In the Biotechnology Patent Protection Actof 1995 (27), procured at the behest of thebiotechnology industry to allow a biotechnol-ogy process claim to piggyback on a productclaim, precedent was set for legislation thatspeaks in biological parlance about patentsconcerning molecular biology. Despite thesupposed protean generality of the patentconceptual scheme, the door has been thrownopen to discipline-specific rules. What mostcommonly seems to flow through that door

is a stream of ad hoc prohibitions, usuallycategorical and often embracing uses as wellas monopolies on uses. One favorite in theU.S. Congress and the European Parliamentis a moratorium on a given sort of patentor research. Sometimes the rationale for aprohibition will appear to assimilate a prop-erty interest in an extracorporeal molecule toinvasion of personal autonomy, which earlierwe found reason to distinguish. As the por-tion of the human genome claimed by patentsexpands, motivation arises to prohibit anymore such patents. Were a ban imposed, thecontrol experiment of life without patentswould run in real time. Biotechnology firmswould compete with no intellectual propertysave for trade secret protection of whateverthey managed to keep secret. Such competi-tion might produce salutary results. It mightalso diminish aggregate welfare unless somemechanism replaces at least some of theincentives fostered by product patents.

Compulsory Publication of DNA SequenceData to Thwart Patents

When without first filing for a patent, a sci-entist publishes a DNA sequence, no one mayobtain a patent on the sequence. Mere cita-tion of that publication as prior art will spikeanyone’s claim that the sequence is ‘‘new.’’Mindful of this, some scientists acting oftheir own volition and other scientists act-ing in compliance with funding mandateshave promptly and systematically releasedDNA sequence data as discovered. A con-certed effort so to publish could thwartmost new DNA patent applications. There-upon it becomes open season for any andall to explore therapies predicated on allunpatented portions of the genome. The ben-efits of expanding the universe of potentialinvestigators would seem apparent. To theextent that research motivated by profit maycontribute applications that might not flowfrom academic laboratories, incentives mustnow be sought elsewhere.

Subsidies

When a public good is underprovided, asis familiarly the case in perfect compe-tition (e.g., as to education and national


defense), government may step in to pro-vide it. Valuable public ideas are intellectualpublic goods. Suppose then that no fur-ther patents issue on DNA sequences otherthan artificial sequences. Instead, the gov-ernment systematically subsidizes biotech-nological research. Subsidies are awardednot only to academic institutions but to non-profit biotechnology research centers. Specif-ically organized for the pursuit of applied aswell as basic research, these centers tackleapplications that might not be pursued, orpursued with less zeal, in academic labora-tories. This scheme could implement coor-dinated decisions, reached with benefit ofexpert extramural advice, concerning whichfields of fundamental biomedical and biotech-nological research should be pursued and towhat extent. The scheme entails substan-tial expenditures and may importune taxesearmarked for research (20). But the subsi-dies assure that society gains the benefit ofvaluable innovations.

Were it widgets that society sought toencourage, subsidies for institutional labo-ratory research might not succeed in coax-ing the same innovations as would marketincentives for entrepreneurs experimentingin their shops. When the desired innovationsare biotechnological, it happens that aca-demic laboratories constitute society’s mostfertile source of ideas. What academic labo-ratories do not pursue by way of applicationsmay be pursued in the research centers.Were a share of sales revenues promised toany laboratory originating an end product,market incentives could also be brought tobear within both academic laboratories andresearch centers.

In the marketplace, with valuable discov-eries being contributed to the public domainand available for exploitation, firms wouldnow compete less on the basis of their discov-eries and more on their efficiency in produc-tion. As with any subsidy, it may be difficultto ascertain whether the extent of biotech-nological innovation induced is optimal. Butone could at least compare the extent of tech-nological innovation during the present era ofproduct patent availability with the extent oftechnological innovation under a new regime.

Exclusive FDA Approval for a Term of Years

A government may also engraft an incen-tive mechanism upon the process by which,with a view to public safety, the govern-ment grants approval for the sale of medi-cal products and devices. The Orphan DrugAct (28) affords a model for such an incen-tive scheme. According to that statute, ifthe FDA grants a manufacturer approvalto sell a drug targeted at a disease thataffects fewer than 200,000 persons in theUnited States, or whose likely sales cannotreasonably be expected to recoup the costsof development, the agency must refrain, forseven years after such approval, from approv-ing sale of the drug by anyone else for useagainst that disease. Routine delay in obtain-ing FDA approval for any drug affords to thefirst party who gains FDA approval someperiod of de facto postapproval protectionagainst imitators; in respect of an orphandrug, the first party to gain approval enjoysseven years of de jure postapproval exclusiv-ity. The orphan drug scheme is not withoutits complications. For purposes of identifyingwhich compounds are blocked for seven yearsby an approved orphan drug, it has been nec-essary to define what constitutes ‘‘the samedrug.’’ The FDA defines a new drug to bethe same as a previously approved orphandrug if the new drug has the same ‘‘prin-cipal molecular features’’—unless the newdrug is ‘‘clinically superior’’ to the approvedorphan drug (29). This seems to rehearse,though with variations, a judicial patent doc-trine that a claimed invention is obvious ifthe prior art includes a structurally sim-ilar compound—unless the claimed inven-tion possesses an unexpected property (17,pp. 145–148).

This incentive scheme could be extended.From orphan drugs it could be extended toany genre of products that seem likely toserve the public interest—indeed to any andall biotechnology products. To specify thegenre of products for exclusive approval, thegovernment could rely on advice from extra-mural scientific panels. Such a scheme wouldspare the costs, burdens, and uncertaintiesof patents. It would reward the develop-ment of valuable products without tying upthe human genome with property claims. It


would respect the unpatentability of nature.The number of years and other terms of theexclusive sale privilege are of course vari-able. One might also replicate the provisonof the Orphan Drug Act that allows the FDAto approve sale of an orphan drug by a sec-ond applicant if the original manufacturer‘‘cannot assure the availability of sufficientquantities of the drug to meet the needs ofpersons with the disease or condition’’ (30).

Human Methods Patent

Ambivalence between patents on productsvis-a-vis patents on processes has been evi-dent since recombinant DNA technologybegan. The Cohen-Boyer patent protects aprocess. It was followed, as the technol-ogy developed, by many process as wellas product patents (31). The Cohen-Boyerapplication also sought claims on recombi-nants, but no product patents issued until1984 (on plasmids) and 1988 (on plasmid-transformed hosts). Stanford University’slicensing of the process patent thrived begin-ning in 1981 before Stanford acquired anyproduct patent (32).

A possible resolution of the ambivalencewould be to require hereafter that to securea patent pertaining to a DNA sequence, onemust invent some new process that can beperformed in respect of the sequence ratherthan claim to have invented the sequence orits gene product. A new form of patent predi-cated on this principle has been proposed (14)in the following statutory phrasing:

There shall be allowed a patent pertaining toa human life form (a ‘‘human methods patent’’or ‘‘HMP’’), the scope of which patent shall notexceed the least inclusive description of an inge-nious process. Such a process may consist inthe production, use, alteration, amplification,or attenuation of human life forms outside thehuman body. An HMP may include an addi-tional claim on nonhuman reproduction of anytransgenic and its progeny if and only if (a) theingenious process produces such transgenic,(b) such transgenic produces a human life form,and (c) without reference to such human lifeform, the process would not be patentable.

No product patent shall be allowed ona human life form or anything in which itis included. The foregoing shall not preclude

a patent on a synthesized, fully explicatednucleotide sequence or protein that is notpresent, consecutively or otherwise, in thehuman body.

Research in a nonprofit institution for non-profit purposes shall be exempt from any claimof infringement.

The significance of an HMP may be mademore clear by the following observations.

(i) Interspecies homology is only similar-ity to a degree, not identity of nucleotidesequences. Absent evidence of identity with anonhuman form in a given case, ‘‘human lifeform’’ may be assumed distinct.

(ii) The confinement of an HMP to the leastinclusive description of an invention protectsagainst the detriment of overbreadth as illus-trated by experience with erythropoietin andfactor VIII:C. Such limitation would departfrom the law’s tendency to allow contribu-tors a claim on a whole—as when a farmerobtains a claim on grain in an elevator withwhich the farmer’s is commingled, or a secu-rity interest in a part attaches to a mass inwhich the part is commingled or assembled.Reasons for parsimony obtain concerning thehuman genome.

(iii) Suppose that an HMP claims ‘‘amethod for obtaining DNA sequenceb1, b2, . . . bn from genomic DNA as fol-lows: . . . ,’’ and describes the ingeniousmethod by which the sequence was discov-ered. Without more, such a patent wouldafford little protection. Everyone may nowread the sequence disclosed by the patent.Free of infringement, anyone may thenobtain the sequence by employing any pro-cess, including the polymerase chain reac-tion, other than the patented process. Toavoid this vulnerability, the discoverer mightseek to claim ‘‘cloning of the sequence invector v and transformation by v of host hthat results in production of protein pi asfollows: . . . .’’ Perhaps this investigator hasingeniously devised a way to use a new vto produce pi in some mammalian h neverbefore used to produce human substances. Ingeneral, it will not be ingenious to clone anidentified sequence, nor to produce a proteinby means of a known gene. The principle ofleast inclusiveness allows a claim on only so


much of the process as is ingenious. The dis-coverer’s successors may find it unnecessaryto use the process first used to discover thething, and may proceed to ‘‘event around’’the process. This is true about the discoveryof any natural thing. Successors may alsoinvent methods by which to use, alter, orpromote or attenuate the effect of the thing.

Process patent opportunities stillawait—in protein chemistry, insertion offoreign DNA, transformation and infection,gene expression, and protein-manufacturingtechniques. One process might describe atechnique for making a protein, another howto use it. Or a firm might use knowledge of agene not to produce but to curtail the effect ofa given protein, including a newly discoveredprotein.

(iv) It may be possible to state certain min-imal conditions for work to be ingenious. Ifa claimed method is predicated on a humanlife form, it may be unlikely to exhibit anadvance over present knowledge unless thelife form is fully explicated. ‘‘Fully expli-cated’’ entails, in the case of DNA, that aspecific nucleotide sequence (the ‘‘explicatedsequence’’) is identified, including all regula-tory sequences necessary for any exons in theexplicated sequence to be transcribed intoRNA and for a gene to be expressed, thatall such sequences have been inserted intoa vector or maintained in some stable form,that it is known what the explicated sequenceencodes or regulates (or perhaps only that itis implicated in the etiology of a disease),and that the process succeeds in expressingor preventing expression of such gene. For aprotein, full explication would embrace bio-logical function, amino acid sequence, andencoding gene sequence.

(v) One could circumvent a patent thusfar described if, for example, one were topay a royalty in order to perform a patentedtransgenic process of producing a humanhormone in a pig, and then, without pay-ing any more royalties, one were to breed aline of pigs. Thereby one could obtain copiousamounts of the hormone. Natural breeding ofcourse produces naturally occurring progeny,and, except for plants in the United States,such progeny would seem unpatentable. Toprevent the foregoing circumvention, which

would defeat an inventor’s reasonable prop-erty expectations, the HMP allows a claim ongrowing or nonhuman breeding of a trans-genic if the transgenic is a result of theinvented process and the transgenic producessome human life form without reference towhich the process would not be patentable.The additional claim may be defended asa claim on reproduction of an ‘‘unnatural’’organism, one not likely to be found in nature.Such scheme resolves the predicament towhich the self-reproducibility argument forthe Harvard mouse patent is directed. Itallows no claim on a human life form itself.Nor may the additional claim encompasshuman reproduction. Should the inventedprocess happen to be one of artificial humanreproduction, remedies may be provided (asdiscussed in the next section) against infring-ing providers, but never against a parent orchild as such.

Whether an ‘‘artificial gene’’ or the proteinit encodes will qualify for an HMP is contin-gent on how close a variant or equivalent thegene may be to what is found in the humangenome. Will this contingency discouragefruitful research on the genome? Significantdisincentives seem unlikely unless firms sogreatly prefer product to process patentsthat they choose to pursue the more difficulttask of sequencing proteins rather than find-ing naturally occurring genes encoding forproteins. Where proteins may be sequencedautomatically, a disincentive may occur. Butif the therapeutic value of an artificial geneproduct is not sufficient, it will not be anappealing product no matter what the patentavailability. At least the products of naturallyoccurring genes have known worth.

Objections to the HMP and replies theretoinclude the following. Industrialists prefer-ring product patents often contend thatrecombinants are more potent and free ofcontaminants, that recombinants thus differfrom natural isolates and from each other,and hence that product patents will not pre-vent new advances from reaching the market.This conjecture seems belied by the historyof erythropoietin and factor VIII:C in par-ticular, and in general by the hegemonyof any product patent over improvements.Whichever industrialist happens to be first in


time will often hoist another on the petard ofcontradiction. In scientific publications andin advertising, sellers of recombinants arewont to describe their products as virtuallyidentical to the corresponding natural iso-lates. But when forced to defend againsta claim of patent infringement, the samesellers may be heard invoking the ‘‘reversedoctrine of equivalents,’’ which, under U.S.patent law, excuses some literal infringe-ments if the accused product displays dif-ferences in specific activity and purity fromthe patented product. (The doctrine of whichthis is called the ‘‘reverse’’ sustains a claimof infringement against an accused proteinsomehow differing from a patented proteinif no functional differences obtain betweenthem.)

A more orthodox industrialist objection tothe HMP would be to say that without prod-uct patents, businesses will not invest themillions of dollars needed to find a geneand to produce a protein by recombinantmethods. This bluff is handy because it iscounterfactual. As earlier indicated, whenone looks at the relatively scant evidence ofinventive behavior without patents, and thenconjectures about what happens if only pro-cess and not product patents are available,one may be sceptical about the claim thatbiotechnology cannot thrive without productpatents. The effective protection afforded byprocess patents depends on how easy it isto design around a process. Large, complexproteins found in humans may be more dif-ficult to design than, say, pharmaceuticals.Biotechnology patents are replete with pro-cess claims. It appears that firms have foundways to protect their intellectual propertyeven though patent examiners vary in theirview of product vis-a-vis process claims, andeven though, given how often courts invali-date them, the status of any product patentis contingent. It must be granted that processpatents are often less convenient to enforcebecause a patentee must show what tran-spired in a rival’s plant. Even so, if a patenthas been issued on a recombinant process,ordinarily the recombinant result has only avery low probability of naturally occurring.The patent holder may invoke this probabil-ity to refute a defendant’s claim to have bred

transgenics without using a patented processand without using offspring of the patentedprocess.

One previous motivation for a U.S. prod-uct patent is now obviated. When the Har-vard mouse emerged, anyone could avoidinfringement of a U.S. process patent byperforming the patented process in a for-eign country outside the reach of U.S. lawand then importing the end product into theUnited States; no such move would defeata product patent. A statutory amendmentchanged this by declaring that any suchimportation is an infringement of the processpatent (33). By virtue of the BiotechnologyPatent Protection Act, one may obtain a pro-cess patent on a recombinant process thatuses or makes a patented product, althoughthis piggyback rule will be moot if productpatents become unavailable. Instead of thispiggyback rule, it might better be declaredthat a patent is available on an inventedprocess if what the process uses or produceswould be patentable but for the fact thatthe product is a human life form. Such isthe effect of the HMP. It allows a processclaim to be predicated upon a human lifeform while allowing no claim on the life formitself.

It remains necessary to show an inge-nious process. An industrialist may objectthat there seem to be few new processes toinvent, that current biotechnology employsstandard processes that differ only by genesexpressed. Mere substitution of a differentgene in a known process may indeed beperfunctory. It would not seem to state anargument for product patents to say thatinnovation is difficult. Opportunities for pro-cess innovations abound. The Cohen-Boyerpatents expired in 1997. It may simply bethat the challenge of finding genes commandsmore attention at present.

The HMP, subsidies, and a period ofexclusive FDA drug approval could be imple-mented separately or together.


ANCILLARY MECHANISMS

Compulsory Licensing

A patent subjects society to the vagaries ofa monopolist’s choices and fortunes. A possi-ble protection against such risk with respectto biological patents is compulsory licens-ing according to which anyone may use apatented process upon payment of no morethan some reasonable royalty. Another pro-tection is ceilings on the prices of goods madeby patented processes. As early as the fed-erally supported Cohen-Boyer research, theNIH considered seeking patents on fundedinnovations. NIH asserted patent rights toAZT based on the research contributionsof NIH intramural scientists, all with thedeclared purpose of restraining prices of prod-ucts. This prompts the suggestion that a gov-ernment agency other than the patent officebe empowered to determine what events trig-ger, and the royalty rate of, a compulsorylicense established as a condition of anybiotechnological patent. A further conditionmight empower the agency to set maximumprices on goods produced and processes per-formed in the practice of the patent. An idealscheme would foster commercial incentivesand allow a reasonable return on investmentwhile preventing exorbitant prices.

Such a scheme, it may immediately beobjected, would interfere with markets. Theindustrialist might contend that govern-ments should not restrain returns on geneticinventions since they do not restrain pricesof patented artificial hearts or organ trans-plants. One might reply that when a gov-ernment grants the privilege of selling adrug or medical product, or of enjoying amonopoly on anything importantly related tohuman health, the public interest may jus-tify conditioning the privilege on end productprice restraint. Compulsory licensing wouldalso protect against disasters with respectto things other than price. As earlier noted,the patentee of the sole therapy for a seriousdisease could become bankrupt or for otherreasons decline to practice or license theinvention. The common weal may demandthat the invention be available. The indus-trialist’s appeal to the case of an organtransplant does not provide a persuasive

counterexample against a compulsory licensebecause organs are donated and recipientspay only for services. An artificial organ isnot perfectly analogous to a gene since theorgan lacks person-defining genetic informa-tion. In any case there may be good reasons tointerfere concerning any commerce in humanparts.

Expert Guidance

A U.S. patent is only presumptively valid.Since courts often invalidate patents, no oneknows for sure that a patent is valid untiland unless it is upheld in court. Considerhow numerous are the courts within thesovereignties that comprise the internationalbiotechnology market. Trial courts decideonly questions placed before them by a flowof cases that is nearly stochastic. The same istrue for appellate courts on which depend theprospects of resolving conflicts among trialcourts. In contrast to scientists for whom dia-logue is a way of life, judges of different courtsdo not, as a matter of decorum, communicatewith each other on pending cases. The scienceon which they rule is also limited to that prac-ticed a few years, if not a decade, before trial.This obtains because time of invention is thereference point for what is obvious. Hencejudicial decisions provide uncertain guidanceabout patentability of today’s scientific pro-cesses. Moral issues, as we earlier saw, arenot even tackled.

It seems improbable that any one wordsuch as ‘‘nonobviousness’’ or ‘‘ingenuity’’ canbear the load of defining what is a sufficientfeat to merit a monopoly. For instance, aclaimed invention might be a tour de force ofgenetic engineering, even though the investi-gator knows neither a sequence’s chromo-somal locus nor the sequence’s coding orregulatory function, if the investigator cor-rectly infers that the sequence is involved,by homology or otherwise, in the etiology of adisease. To transform ‘‘ingenious’’ from place-holder to admission ticket, we may have tosettle for a notion of family resemblance. Forif ingenuity were to admit of precise defini-tion, would anything be ingenious?

To meet the difficulty of recognizing inge-nuity as science progresses, to overcome thelag between research and adjudication, and to


improve upon the limited expertise broughtto bear in patent adjudication, a mecha-nism could be confirmed for introducing sci-entific expertise. A government agency, oth-erwise involved in scientific research, couldexercise authority continually to revise pub-lished standards for patenting life forms inreliance on recommendations of expert scien-tific panels. For purposes of judicial review,the law could preserve the practice of judg-ing a patent by the standards in effect at thetime of alleged invention. From such expertlyframed standards, the biotechnology indus-try could obtain guidance more current andsystematic than case law or statute is likelyever to be.

IMPLICATIONS OF PATENTS IN THE CLINIC

Introduction of Human Substances Outside theGerm Line

Ex vivo somatic intervention involves remov-ing patient cells (e.g., tissue-infiltrating lym-phocytes or bone marrow stem cells), grow-ing them in culture, transferring genes intothem using nonvirulent retroviruses or oth-erwise, and reintroducing the cells into thepatient’s body, not necessarily at the site oftheir effect. In vivo intervention is exempli-fied by the introduction of retrovirus vectorscontaining human genes at the site of thecondition to be overcome. The ED, whichwould allow patents on substances isolatedfrom the human body, would permit, whilethe HMP would deny, a patent on such cul-tured cells or vectors. They are ineligiblefor an HMP because they are or containhuman life forms. Indeed the cultured cellsare grown from the patient’s. Except forattempted enhancement, the cultured cellswould be unlikely candidates for ‘‘inventions’’anyway. They are not intended to be innova-tions. The goal of therapy is to insert a normalgene. The ultimate achievement is homolo-gous recombination. Thereby a normal genereplaces a defective one rather than enteringthe genome at an indeterminate locus.

Somatic interventions involve medicalprocedures on patients. Medical treatment,surgery, and diagnosis are not patentable inEurope (34). Their eligibility for U.S. patents

has been dubious since 1862 when a patentwas sought on the use of ether. It has seemedto many that it would be wrong to discour-age physicians, on pain of infringement, fromdeploying in the relief of human suffering themost efficacious procedures they can musterunder the exigencies they face. Hence onemight conclude that the only processes ofsomatic intervention that may qualify for anHMP are ancillary laboratory processes. Sim-ilarly might patents be confined to laboratoryprocesses with respect to tissues or organsgrown in cell culture—especially if, as maybe typical to avoid rejection, the cultured cellsare grown from the patient’s. Opportunitiesfor process innovations would appear abun-dant when one apprises present difficultiesin somatic cell therapy and the challenge ofgrowing tissues and organs.

A contrary moral view might be that theforegoing is too generous. Suppose that oneopposed patents on reproduction of any sort.One might assimilate the culturing of cellsto reproduction, thereby reversing the patentlaw’s assimilation of reproduction to man-ufacture. One might add that a laboratoryprocess ancillary to a medical treatment isindistinguishable for these purposes from thetreatment. The contention that human repro-duction cannot be an infringement does notentail any claim about what is human repro-duction. One might conclude that a patent ongrowing cells outside the human body doesnot threaten any patient’s autonomy so longas there is no claim on the cells themselves.The difficulty of developing successful meth-ods of somatic cell therapy, and of cultivatingtissues and organs, suggests the benefit ofpatent incentives. One need not claim thatlaboratory processes ancillary to medical pro-cedures are in general nonmedical. One needonly allow some of them to be patentable.

Human Germ Line Intervention

Germ line intervention affects reproductionin two ways. (a) It alters the genome, an off-spring’s complement of genes that appear inall cells including the gametes. (b) In orderto achieve (a), it is performed before germand somatic cells of an individual differenti-ate, i.e., on zygotes and early stage embryos.A moral objection might be lodged against


a patent on any such method because ofthese links to reproduction. As noted, theED would allow no patent on any methodof human germ line intervention. Again areply may be that collective benefit couldresult from creating patent incentives oncertain laboratory processes. It is also note-worthy that a patent on gene therapy wouldnot be a patent on in vitro fertilization.Therapy is subsequent to fertilization. Thechoice to conceive may be seen as a dif-ferent choice than the choice whether tointervene genetically for the health of a childwhose conception has been chosen, even ifthe former is contingent on the latter. Onthe other hand, the opposite may be thecase if eggs fertilized in vitro are screenedfor genetic defects or traits, thereby exercis-ing a choice of which shall live. Two in vivomethods also merit mention. One consistsin altering an embryo in utero by retrovi-ral infection. Another consists in causingadult testes or ovaries to produce geneti-cally engineered gametes (35). A requestedEuropean patent on the latter technique wascriticized as contrary to l’ordre publique ormorality (36). For these also one may askwhether the prospect of collective benefitsuffices to warrant property claims on ancil-lary laboratory processes of medical proce-dures.

If government grants patents on any germline interventionary process, does that com-port with the stance that human reproductioncannot be infringement? The answer lies instipulating that no remedy will lie against aparent or child as such. Damages and precon-ception injunctive relief could be made avail-able against unlicensed providers of patentedprocesses. If Mr. and Mrs. Thurston, learn-ing of Mendipulate Inc.’s patented techniquefor germ line manipulation, arrange withtheir physician for the technique but no onepays the royalty, a damage remedy maylie against the providers. We can scarcelyimagine a suit by Mendipulate against Mrs.Thurston, her daughter or granddaughter,or their physicians or hospitals, complainingof the conception of a child, not to mentioninjunctive relief, i.e., an order for an abortion.Mere pragmatism makes clear that Mendip-ulate’s interests require no remedy against

a patient. Drug manufacturers do not suepatients who infringe by ‘‘using’’ an infring-ing drug. They sue rival manufacturers anddistributors who ‘‘make’’ and ‘‘sell’’ the drugin quantity.

Mendipulate may protest that if it can-not obtain a product patent, every Thurstondescendant will benefit from Mendipulate’sinvention without paying for it. Mendipu-late is correct that the HMP allows claimson reproducing the progeny of transgene-sis only for nonhuman reproduction. Butconsider that Mendipulate will advertise apatented process of germ line therapy as amethod to remedy a genetic defect. It cannottenably assert that if it had a product patent,many Thurston descendants would becomegood-paying customers when they inherit thedefect! Moreover, whether the process is ther-apy or enhancement, Mendipulate’s twentyyears of monopoly will run before any trans-genic Thurston reaches adulthood. Mendip-ulate may still complain that if Mr. or Mrs.Thurston undergoes a patented Mendipulateprocess that causes them to produce geneti-cally engineered gametes, no more compen-sation will be gotten by Mendipulate if theThurstons have a dozen children than if theyhave one. This of course overlooks the differ-ence between having children and copying apatented contraption for profit. People are notmotivated to have children because they cancopy a gene for free. Mendipulate may antic-ipate fecundity when it prices the royalty forits laboratory process.

Since interventions will be performed byphysicians, enforcement of a process patentwill require showing what happened in thedoctor’s office. To Mendipulate this will seeminconvenient. It would prefer a product patentwhose infringement it could establish by com-parison of parental and progeny DNA. Sucha comparison would be peculiar, to say theleast, as it would be mustered in support of acomplaint that a child is healthy or possessedof some enhancement. It should suffice toprotect Mendipulate that licensed specialistsmay generally be expected to pay royaltieson patented processes. What would be trou-blesome would be the enterprising move ofa patient who sells gametes that contain


altered genes. This concern may be mini-mized for the moment by realizing that onlyenhancement genes, not corrected disease-causing genes, would be likely to be mar-ketable.

Society might deem the collective benefitof enhancement to be insufficient for allowinga patent. If concerns about playing God anddiscrimination prevail, refusing patents onenhancement would be a means to discouragethe practice. A contrary view might be thatif we demarcate certain interventions to beoutside the physician’s armamentarium formaintaining health, no public policy will bedisserved by a patent.

There remains possible a product patenton a synthetic gene nowhere found inhumans. To use such a gene might departfrom the present vision of installing normalin lieu of defective genes. The prospect of suchdepartures no doubt explains the habitualmention of Frankenstein when observers dis-cuss germ line intervention. Regardless, theimmunity of parents and children as suchfrom claims of infringement would control.The inventor of a human genetic interven-tion surrenders the product of the processfor integration into an unownable being. Ifa process alters an early stage embryo, inte-gration occurs into a human in gestation.If alterations are made in gametes or themeans of their production, integration occursinto the body of the patient.

CONSISTENCY OF POLICY FOR PLANTS ANDANIMALS

Unless policies about forms of life evince aconsistent understanding of innovation andreflect generalizable moral principles, a sta-ble consensus seems unattainable. Condi-tions (1) and (2) above stated for a biotechnol-ogy product patent—low likelihood of findingthe claimed invention in nature, and inge-nuity—appear applicable to any life formpatent. Some transgenic plants and animalsmay be improbable of natural occurrence andrecognizable as the products of ingenuity.Others may possess transgenes from mem-bers of their own species for which the oddsof acquisition by mutation are better than

trivial, or as to which the process of transgen-esis is not ingenious. Where a product patentwould be unwarranted, a process patent couldbe available. As may an HMP, a processpatent could claim a process by referenceto an identified plant or animal life form. Itcould add a claim on the breeding of any plantor animal that the patented process producesand without which the process would not bepatentable. Such an additional claim wouldobviate the self-reproducibility rationale fora transgenic product patent.

One may argue for bounding a patent’senforceability by operation of a ‘‘farmers’privilege,’’ a derogation imposed for plantsin the United States and often proposed foranimals there and in the ED. This permitsa farmer to breed patented animals to theextent needed to replenish stocks on the farm,or to plant seeds generated by transgenicplants grown on the farm. A farmers’ privi-lege would avail a typical farmer who doesnot seek to compete with breeders in sellingvarieties as such but who wishes to sell whatis raised on the farm. The derogation wouldentail that, as Mendipulate must do concern-ing the Thurstons, commercial breeders mustcollect their royalties on the first generation.

ACKNOWLEDGMENT

This article is adapted from the author’sNorms for Patents Concerning Human andother Life Forms (14).

BIBLIOGRAPHY

1. P.J. Federico, Science 86, 327 (1937).2. Directive 98/44/EC of the European Parlia-

ment and of the Council of 6 July 1998 on theLegal Protection of Biotechnological Inven-tions, Off. J. Eur. Commun. L213, 13–21,Article 6.1 (1998).

3. M.A. Ryan, J. Med. Philos. 20, 419–438(1995).

4. Council of Europe, Convention for the Pro-tection of Human Rights and Dignity of theHuman Being with Regard to the Applicationof Biology and Medicine, Eur. Treaty Ser. 164,Article 13.

5. W. Gardner, J. Med. Philos. 20, 65–84 (1995).6. Diamond v. Chakrabarty, 447 U.S. 303 (1980).


7. Ex parte Allen, 2 U.S.P.Q. 2d 1425 (1987).8. U.S. Pat. 4,736,866 (April 12, 1988), P. Leder

(to the President and Fellows of Harvard Col-lege).

9. 35 U.S.C. §101.

10. European Patent Convention Article 53.11. M. Tallacchini, Bibl. Della Liberta 33(147),

21–50 (1998).12. M.S. Greenfield, 44 Stanford Law Rev.

1051–1094, 1058 (1992).13. J. Bentham, A Manual of Political Economy

in J. Bowring, ed., The Works of Jeremy Ben-tham, vol. 3, Russell and Russell, New York,1962, p. 71.

14. L.M. Guenin, Theor. Med. 17, 279–314 (1996).15. U.S. Pat. 4,237,224 (December 2, 1980),

S. Cohen and H. Boyer (to Trustees of LelandStanford Jr. University).

16. B.D. Davis, Science 193, 442 (1976); Am. Sci.65, 547–555 (1977); The Genetic Revolution,Johns Hopkins University Press, Baltimore,MD, 1991, pp. 21, 243.

17. P.G. Ducor, Patenting the Recombinant Prod-ucts of Biotechnology, Kluwer, London, 1998.

18. Amgen v. Chugai Pharmaceutical Co., Ltd.,706 F. Supp. 94 (D. Mass. 1989); 13 U.S.P.Q.2d 1737 (D. Mass. 1989); aff’d in part andreversed in part, 927 F. 2d 1200 (Fed. Ct.1991), cert. denied, 112 U.S. 169 (1991).

19. Scripps Clinic and Research Foundation v.Genentech, Inc., 666 F. Supp. 1379 (N. D. Cal.1987); 678 F. Supp. 1429 (N. D. Cal. 1988);707 F. Supp. 1547 (N. D. Cal. 1989); aff’d inpart and reversed in part, 927 F. 2d 1565 (Fed.Cir. 1991).

20. M. Svatos, in S. Sterckx, ed., Biotechnology,Patents and Morality, Ashgate, Aldershot,1997, pp. 291–305.

21. Y. Miki et al, Science 266, 66–71 (1994).22. D. Dickson, Nature 378, 424–425 (1995).23. R. Wooster et al, Nature 378, 789–792 (1995).

24. 35 U.S.C. §101–103; European Patent Con-vention Article 52 (expressing the last twocriteria as ‘‘industrial application’’ and ‘‘inven-tive step,’’ and, corresponding to the notionof ‘‘manufacture,’’ excluding ‘‘discoveries’’ ofnature).

25. Merck & Co., Inc. v. Olin Mathieson ChemicalCorp., 116 U.S.P.Q. 484 (4th Cir. 1958).

26. In re Deuel, 34 U.S.P.Q. 2d 1210 (Fed. Cir.1995), discussed at length in (17), pp. 70–90,141–142.

27. P.L. 104-41, adding 35 U.S.C. §103(b).

28. 21 U.S.C. §§360aa–360cc.29. 21 C.F.R. §316.3 (13).30. 21 U.S.C. §360cc.31. D.W. Plant, N.J. Reimers, and N.D. Zinder,

eds., Patenting Life Forms, Cold Spring Har-bor Lab., Cold Spring Harbor, NY, 1982,pp. 75–78, 95–106.

32. U.S. Pat. 4,468,464 (August 28, 1984) and4,740,470 (April 26, 1988), S. Cohen andH. Boyer (to Trustees of Leland Stanford Jr.University). See N. Reimers, J. Assoc. Univ.Technol. Managers 7, 25–47 (1995).

33. 35 U.S.C. §271 (g) added by P.L. 100-418.34. European Patent Convention Article 52 (4).35. R. Brinster and M.R. Avarbock, Proc. Natl.

Acad. Sci. 91, 11303–11307 (1994).36. Nature 368, 572 (1994).

See other entries Human Genome Diver-sity Project; Medical Biotechnology, UnitedStates Policies Influencing its Development;Ownership of Human Biological Material; seealso Patents and Licensing entries.

patents, ethics, human life human germ line ...guenin.med.harvard.edu/documents/ethics of patents...

Documents