`zures (intrathalamic administration) [Z. Liu, M.
`Vergnes, A. Depaulis, C. Marescaux, Neuro(cid:173)
`science 48, 87 (1992)] and (ii) two pharmacologic
`models of absence seizures (i.p. administration)
`[O. C. Snead, Eur. J. Pharmacol. 213, 343 (1992)].
`26. G. Karlson, M. Schmutz, C. Kolb, H. Bittiger, H.-R.
`Olpe, in GABA8 Receptors in Mammalian Func(cid:173)
`tion, N. G. Bowery et al., Eds. (Wiley, Chichester,
`United Kingdom, 1990), pp. 349-365.
`27. P. C. Waldmeier, P. Wicki, J.-J. Feldtrauer, P. A.
`Baumann, Naunyn-Schmiedeberg 's Arch. Phar(cid:173)
`makol. 337, 289 (1988) ; R. A. Deisz and D. A.
`Prince,.). Physiol. (London) 412, 513 (1989) ; D. D.
`Mott and D. V. Lewi s, Science 252, 1718 (1991) .
`28. B63HF1
`from C57Bl/6JEi
`females
`is the F1
`crossed with C3H/HESnJ males. We refer to the
`
`B63HF1 strain, with which lh/lh is congenic, as
`wild or +/+ . We maintain lh!lh and +/+ mouse
`colonies at the Duke University Vivarium , using
`stocks from Jackson Laboratory (Bar Harbor,
`ME) .
`29. Supported by grants from the Veterans Adminis(cid:173)
`tration (D.A.H. and WA.W.) and NIH (D.A.H .,
`W.A.W., and RAM.). We thank H. Bittiger for the
`gift of CGP 35348 [European patent 463560-A 1 (2
`January 1992); priority CH-002092 (22 June
`1990)] ; V. H. MacMillan for HPLC measurements;
`N. Lambert, D. V. Lewis, J. 0. McNamara, D. D.
`Mott, C. Shin, and H. S. Swartzwelder for critical
`reading and discussion; and B. Worrell and S.
`Sneed for administrative assistance.
`
`21January 1992; accepted 22 May 1992
`
`Nitric Oxide: A Physiologic Mediator of
`Penile Erection
`Arthur L. Burnett, Charles J. Lowenstein, David S. Bredt,
`Thomas S. K. Chang, Solomon H. Snyder*
`Nitric oxide (NO) is a cytotoxic agent of macrophages, a messenger molecule of neurons,
`and a vasodilator produced by endothelial cells. NO synthase, the synthetic enzyme for
`NO, was localized to rat penile neurons innervating the corpora cavernosa and to neuronal
`plexuses in the adventitial layer of penile arteries. Small doses of NO synthase inhibitors
`abolished electrophysiologically induced penile erections. These results establish NO as
`a physiologic mediator of erectile function.
`
`N itric oxide mediates bactericidal and tu(cid:173)
`moricidal actions of macrophages (1) and
`blood vessel relaxation of endothelial cells
`(2) . NO may also be a major neuronal
`messenger (3) . lmmunohistochemical stud(cid:173)
`ies localize NO synthase (NOS) to neurons
`in the brain as well as to discrete popula(cid:173)
`tions of autonomic nerves in the periphery
`(4) , where NO fulfills most characteristics
`of a neurotransmitter. For instance, NOS is
`highly localized to cell bodies and fibers of
`the myenteric plexus of the gastrointestinal
`pathway
`(4). The nonadrenergic, non(cid:173)
`cholinergic relaxation evoked by physiolog(cid:173)
`ic stimulation of myenteric plexus neurons
`is potently and selectively blocked by NOS
`inhibitors (5) .
`Penile erection is thought to involve
`parasympathetic, neuronally mediated re(cid:173)
`laxation of the blood vessels as well as of
`the trabecular meshwork of smooth muscle
`that comprises the corpora cavernosa (6).
`The neuronal chemical mediator of erec(cid:173)
`tion has not been established. Vasoactive
`intestinal polypeptide (VIP) occurs in lim-
`
`A. L. Burnett and T. S. K. Chang, Department of
`Urology, Johns Hopkins University School of Medi(cid:173)
`cine, 725 North Wolfe Street, Baltimore, MD 21205.
`C. J. Lowenstein, Department of Cardiology, Johns
`Hopkins University School of Medicine, 725 North
`Wolfe Street, Baltimore, MD 21205.
`D. S. Bredt and S. H. Snyder, Departments of Neuro(cid:173)
`science, Pharmacology, and Molecular Sciences, and
`Psychiatry, Johns Hopkins University School of Medi(cid:173)
`cine, 725 North Wolfe Street, Baltimore, MD 21205.
`
`'To whom correspondence should be addressed.
`
`-
`
`ited populations of nerve fibers in the penis
`(7), but direct administration of VIP does
`not fully mimic physiologic erection (8). In
`isolated smooth muscle from the corpus
`cavernosa of several species, relaxation
`evoked by electrical field stimulation could
`be blocked by NOS inhibitors, as reported
`in some studies (9) but not in others (10).
`Blockage of relaxation by NOS inhibitors
`can establish NO as a mediator of caver(cid:173)
`nosal muscle relaxation but does not permit
`conclusions as to whether it is a neuronal,
`transmitter-like messenger and a physiolog(cid:173)
`ic mediator of erection.
`Several portions of the genitourinary
`tract of rat displayed substantial NOS ac(cid:173)
`tivity, monitored by the conversion of
`[3H]arginine to [3H]citrulline (Table 1) _
`High concentrations in the pelvic plexus,
`referred to in the rat as the major pelvic
`ganglion, suggest a neuronal role for NOS _
`Amounts of NOS in the membranous ure(cid:173)
`thra exceeded amounts in the pelvic plexus
`and were three to four times larger than
`those in the penis and the bladder neck and
`considerably larger than those in the pros(cid:173)
`tate. This regional distribution of NOS
`activity was confirmed by protein immuno(cid:173)
`blot (1 J) _
`immunohistochemical
`We conducted
`staining of rat penile tissue (Fig. I) with an
`antiserum that is highly selective for NOS
`and stains NOS specifically in a variety of
`rat peripheral tissues and in brain tissue (4).
`All immunohistochemical staining of NOS
`
`SCIENCE
`
`• VOL. 257
`
`• 17 JULY 1992
`
`in penile tissues was blocked by preabsorp(cid:173)
`tion with recombinant NOS protein ( 11) .
`The antibody to NOS stained the pelvic
`plexus and its axonal processes that form
`the cavernous nerve (Fig. IA), located
`immediately adjacent to the deep caver(cid:173)
`nosal artery, the major arterial source of the
`corpus cavernosum. In the proximal penis,
`the nerve plexus in the adventitia of the
`deep cavernosal arteries stained prominent(cid:173)
`ly as did neuronal processes in the sinusoids
`and the periphery of the corpora cavernosa
`(Fig. lB) . This staining circumscribed the
`· corpora cavernosa directly below their fi(cid:173)
`brous capsules, the tunica albuginea_ Neu(cid:173)
`ronal staining of the deep cavernosal arter(cid:173)
`ies continued as the arteries subdivided into
`the intracorporal network of helicine arter(cid:173)
`ies (Fig. ID). In the most distal portion of
`the corpora, staining diminished as the
`helicine arteries were replaced with cavern(cid:173)
`ous spaces (Fig. IE). The distal part of the
`penis superficial to the corpora cavernosa
`contained dorsal penile nerve fibers that
`stained for NOS (Fig. I, E and F)- Dorsal
`penile and cavernosal arteries stained for
`NOS both in the adventitial and endothe(cid:173)
`lial layers, although endothelial staining
`was faint in the cavernosal vessels. NOS
`staining in the urethra was associated with
`the
`neuronal
`fibers
`coursing
`through
`smooth muscle or the submucosal vascula(cid:173)
`ture or both (Fig. IE) and fits with the
`substantial urethral NOS catalytic activity
`(Table 1).
`The neural specificity of NOS staining
`was established by bilateral cavernous nerve
`transection, after which we no longer ob(cid:173)
`served penile neurons stained for NOS (Fig.
`IC), although endothelial staining persist(cid:173)
`ed (11). To ensure that this distribution
`was not species-specific, we conducted im(cid:173)
`munohistochemical localizations of NOS
`from the penes of dogs and demonstrated
`essentially identical localizations to nerve
`plexuses in the adventitial layers of penile
`arteries and the dorsal nerves of the dog
`penes.
`The localization of NOS to neuronal
`fibers innervating blood vessels and the
`corpora cavernosa of the penis suggested a
`possible role for NO as a neuronal mediator
`of erection. We examined this possibility in
`a rat model of penile erection in which we
`electrically stimulated the cavernous nerves
`of intact rats by using optimal parameters
`that evoked physiologic erection (12) _
`L-Nitroarginine, a potent and selective in(cid:173)
`hibitor of NOS, markedly diminished pe(cid:173)
`nile erections (Table 2) . As little as I mg
`per kilogram of body weight (mg/kg) ad(cid:173)
`ministered intravenously (i. v.) significantly
`reduced erection, and 2.5 mg/kg produced
`more than a 50% reduction. At 5 mg/kg,
`nitroarginine almost completely inhibited
`erection. Intravenous bolus injections of
`
`401
`
`1 of 3
`
`LA BioMed EXHIBIT
`
`2031
`
`Lilly v. LA BioMed
`IPR2014-00693
`
`
`
`L-arginine (25 mg/kg) partially reversed the
`L-nitroarginine (2.5 mg/kg) inhibition of
`penile erection. The physiologic L-isomer
`of N-methylarginine, another selective in(cid:173)
`hibitor of NOS that is less potent than
`L-nitroarginine, significantly inhibited pe(cid:173)
`nile erection at 10 mg/kg with a larger effect
`of 40 mg/kg. By contrast the D-isomer,
`which does not inhibit NOS, also failed to
`block erection even at 40 mg/kg. Large
`doses of atropine (1 mg/kg i.v.) did not
`inhibit electrically stimulated erections.
`These results fit with recent experiments
`that show inhibitions of erection by caver(cid:173)
`nosal nerve stimulation in rabbits after in(cid:173)
`jections of L-nitroarginine directly into the
`corpus cavemosum (13). The dose required
`for maximal effect, about 2 mg, is effective(cid:173)
`ly several hundred times greater than the
`parenteral doses we administered to rats.
`Also, L-arginine failed to reverse effects of
`nitroarginine in rabbits ( 13) .
`The stereospecificity for inhibition of
`penile erection displayed by N-methylargi(cid:173)
`nine as well as the very substantial potency
`of nitroarginine indicates that the blockage
`of penile erection comes from the inhibi(cid:173)
`tion of NOS activity. This conclusion is
`supported by observations that relaxation of
`isolated corpus cavemosum muscle strips in
`vitro after electrical field stimulation is
`blocked by NOS inhibitors (9), although
`
`Table 1. NOS activity in the urogenital system.
`We measured the activity of NOS by monitoring
`the conversion of [3 H)arginine to 13 H)citrulline
`as described (21) on specimens obtained by
`anatomical dissection of adult male Sprague(cid:173)
`Dawley rats. Tissue was homogenized in 1 O
`volumes (w/v) 50 mM tris (pH 7.4), 1 mM EDTA,
`1 mM EGTA, and centrifuged at 10,000g for 1
`min at 4°C. Enzyme assays contained 25 µ.I of
`tissue supernatant and 50 µ.I of 100 nM
`[3H)arginine (53 Ci/mmol; 1 Ci = 37 gigabec(cid:173)
`querels), 10 mM nicotinamide adenine dinucle(cid:173)
`otide phosphate (reduced form), and 10 mM
`CaCl2 . After a 15-min incubation at room tem(cid:173)
`perature , the assays were terminated with 3 ml
`of 20 mM Hepes (pH 5.5) with 2 mM EDTA and
`applied to 0.5-ml Columns of Dowex AG50WX8
`(Na+ form) . [3 H]Citrulline was quantified by
`liquid scintillation spectroscopy of the 3-ml
`flow-through . The data are expressed as mean
`values ± SEM for five experiments and normal(cid:173)
`ized to cerebellar NOS activity assayed in par(cid:173)
`allel. According to the Duncan multiple range
`test, the amounts of NOS catalytic activity in the
`pelvic plexus and urethra are different from the
`amounts in the prostate, bladder neck, and
`penis .,
`
`Structure
`
`[3 H)Citrulline
`formation [cpm min - 1
`(mg protein) - 1 ± SEM]
`
`Pelvic plexus
`Membranous urethra
`Penis
`Bladder neck
`Prostate
`
`408 ± 36
`857 ± 105
`212 ± 44
`214 ± 31
`36 ± 15
`
`others report a failure of the inhibitor
`N-methylarginine to prevent such relax(cid:173)
`ation (JO). The selective localization of
`NOS in penile neurons that subserve erec(cid:173)
`tion, as well as the ability of NOS inhibi(cid:173)
`tors to block physiologic erection selective(cid:173)
`ly, potently, and completely, imply that
`NO is the major if not sole neuronal medi(cid:173)
`ator of erection. As in the myenteric plexus
`of the gastrointestinal system, NO in the
`
`penis appears to fulfill the principal criteria
`of a neurotransmitter. For instance, it is
`localized to the neurons that innervate the
`smooth muscle of the penis. Furthermore,
`direct application of NO or its precursors
`relaxes the muscle in a manner similar to
`the relaxation produced by nerve stimula(cid:173)
`tion (9), and the effects of neuronal stimu(cid:173)
`lation are blocked by inhibitors of the
`formation of NO.
`
`·.
`:cs
`
`.. , ...
`cs
`
`\
`
`Fig. 1. lmmunohistochemical localization of NOS in the rat penis . lmmunohistochemistry was
`performed as described (4) on slide-mounted pelvic tissue sections from adult male Sprague(cid:173)
`Dawley rats. The primary antibody was an affinity-purified NOS antiserum (1 :50 dilution), which was
`subsequently bound with the use of an avidin-biotin-peroxidase system (Vector Laboratories) with
`diaminobenzidine as a chromogen. (A) The pelvic plexus (PP) containing neural cell bodies
`adjacent to glands ot the lateral prostate (LP) . (B) Obl ique section through the crus of the corpus
`cavernosum that depicts staining primarily localized to nerves in the · adventitia of the deep
`cavernosal artery and its major tributaries (arrows) and to nerves extending into erecti le tissue (C}
`Duplicate section as (B) that was obtained in an animal 1 week after bilateral transection of the
`cavernous nerves. Faint staining can be observed in the adventitia of the major arterial divisions. (D}
`Coronal section through the pelvis at the level of the proximal penis that shows the corpora
`cavernosa merging in midline. The deep cavernosal arteries have tapered (arrows}, whereas arterial
`subdivisions, the helicine arteries, have extensively arborized . The capsule containing the erectile
`tissue, the tunica albuginea, does not stain (thick arrow) . (E) Cross section through the visible penis
`distally showing prominent nerve fiber staining of the dorsal penis and staining of the urethra
`(arrow). Bilateral cavernous spaces (CS) within the corporal bodies are shown. (F) A magnified view
`of the dorsal penis from (E) showing discrete nerve fibers (NF) . Staining is also localized to the
`adventitia (arrows) and endothelium (arrowhead) of dorsal arteries. [Scale bars in (A) to (C) , and (F}
`= 300 µ.m; bars in (D) and (E) = 1 mm].
`
`402
`
`SCIENCE
`
`• VOL. 257
`
`• 17 JULY 1992
`
`..
`
`2 of 3
`
`
`
`Table 2. Effects of NOS inhibition on penile erection in intact rats. Penile erection was induced
`electrically with a Grass S48 square wave stimulator in anesthetized (pentobarbital , 50 mg/kg,
`administered intraperitoneally) male Sprague-Dawley rats with optimal stimulation parameters ( 12) .
`Bipolar silver wire electrodes were attached unilaterally to the cavernous nerve that arises from the
`ipsilateral pelvic plexus situated dorsolateral to the prostate. lntracavernous pressures were
`measured (Gould Polygraph, Cleveland, Ohio) with a 25-gauge needle inserted unilaterally at the
`base of the penis and connected to an lsotec pressure transducer. Neurostimulation was performed
`until a 10-s maximal pressure recording was achieved, but no stimulation lasted longer than 90 s.
`At least 1 O min elapsed between repeated stimulations. Arginine derivations were administered into
`the jugular vein. Data represent mean values ± SEM as a percentage of the baseline pressure
`(range 35 to 50 mmHg) recorded 15 min after agents were administered.
`
`Agent
`
`Dose
`(mg/kg)
`
`% lntracavernous
`pressure (± SEM)
`
`L-Nitroarginine methyl ester
`
`N-Methyl-L-arginine
`
`N-Methyl-o-arginine
`
`1.0
`2.5
`5.0
`10.0
`40.0
`10.0
`20.0
`40.0
`40.0
`
`75 ± 7
`47 ± 5
`16 ± 1
`10 ± 10.3
`0
`63 ± 3
`17 ± 2
`15 ± 5
`128 ± 5
`
`n
`
`3
`10
`4
`3
`5
`11
`2
`4
`2
`
`Acetycholine is the classical neurotrans(cid:173)
`mitter for the parasympathetic inne"CVation
`of penile nerves responsible for erection.
`However, penile erection does not appear
`to require either cholinergic or adrenergic
`mechanisms (14). VIP was advanced as a
`candidate transmitter for the mediation of
`erection on the basis of its immunohisto(cid:173)
`chemical localization in penile neurons (7).
`However, the density of VIP-containing
`penile neurons in several species (8) is
`substantially less than that of the NOS(cid:173)
`containing neurons that we have observed
`in rats and dogs. Injections of VIP into the
`dog and human penis produce some erec(cid:173)
`tion, but responses are relatively modest
`and may be elicited primarily by increased
`venous outflow resistance rather than by
`the dilation of penile arteries or by the
`relaxation of corpora cavernosa muscle (8).
`Lesion studies demonstrate that NOS-con(cid:173)
`taining fibers in the adventitia of cerebral
`arteries come from parasympathetic cell
`bodies in the sphenopalatine ganglia, many
`of which also contain VIP ( 15). In the
`myenteric plexus, NOS-containing neurons
`also contain VIP. Whereas NOS inhibitors
`can almost completely block neuronally
`mediated gastric relaxation, antibodies to
`VIP can produce up to a 30% blockage of
`this relaxation with NOS inhibitors block(cid:173)
`ing the remaining relaxation (16). Thus, in
`various portions of the parasympathetic
`nervous system, VIP and NO might func(cid:173)
`tion as cotransmitters.
`The immunohistochemical visualization
`of NOS in penile neurons clarifies function(cid:173)
`al penile innervation as described in hu(cid:173)
`mans by Walsh and Donker (17) and as
`described by others in rats (18) . NOS-
`
`containing cavernous nerve processes pen(cid:173)
`etrate the corpora cavernosa, appear to
`envelop the centrally situated cavernosal
`arteries, and also extend into the corporal
`bodies radially and circumferentially. This
`implies a direct neural modulation of the
`vasoactivity of penile arteries, intracorporal
`sinusoids, and the entire tubular-shaped
`corpora cavernosa. Besides well-character(cid:173)
`ized neuroregulation of arterial dilation in
`the penis, our findings suggest that penile
`erection also involves an active process of
`neurally regulated sinusoidal and corporal
`expansion rather than passive engorgement
`of cavernous spaces with blood supplied by
`the penile arteries. The immunohistochem(cid:173)
`ical localization of NOS in the urethra
`parallels its high NOS catalytic activity,
`which suggests that NO may be involved in
`the urethral functions that regulate urinary
`continence or micturition.
`The involvement of NO in erection has
`clinical implications. Priapism, a condition
`of painful, prolonged erections unassociated
`with sexual arousal or desire, occurs in sev(cid:173)
`eral clinical situations, including as many as
`40% of patients with sickle cell anemia ( 19).
`NOS inhibitors such as nitroarginine might
`have therapeutic utility in priapism. Further(cid:173)
`more, some varieties of impotence may re(cid:173)
`sult from a dysfunction in the NOS-contain(cid:173)
`ing neuronal system. Our delineation of
`NOS neuronal localizations within the cor(cid:173)
`pora cavernosa may facilitate surgical ap(cid:173)
`proaches to correct penile abnormalities and
`injuries, such as hypospadias and Peyronie's
`disease, respectively congenital and acquired
`conditions conspicuous for disabling penile
`curvature. Impotence associated with their
`operative management (20) might be avoid-
`
`ed with surgical techniques designed to pre(cid:173)
`serve NOS innervation.
`
`REFERENCES AND NOTES
`
`1. J. B. Hibbs, Z. Vavrin, R. R. Taintor, J. lmmunol.
`138, 550 (1987) : J. B. Hibbs, Jr., R. R. Taintor, Z.
`Vavrin, E. M. Rachlin, Biochem. Biophys. Res.
`Commun. 157, 87 (1988); C. F. Nathan and J.B.
`Hibbs, Jr. , Curr. Opin. Jmmunol. 3, 65 (1991).
`2. R. F. Furchgott and J. V. Zawadzki, Nature 288,
`373 (1980); S. Moncada, R. M. J. Palmer, E. A
`Higgs, Biochem. Pharmacol. 38, 1709 (1989) ; R.
`M. J. Palmer, A G. Ferrige, S. Moncada, Nature
`327, 524 (1987); R. F. Furchgott and P. M. Van(cid:173)
`houtte, FASEB J. 3, 2007 (1988); L. J. lgnarro,
`Annu. Rev. Pharmacol. Toxico/. 30, 535 (1990).
`3. D. S. Bred! and S. H. Snyder, Neuron 8, 3 (1992);
`J. Garthwaite, Trends Neurosci. 14, 60 (1991).
`4. D. S. Bredt, P. M. Hwang, S. H. Snyder, Nature
`347, 768 (1990) .
`5. K. M. Desai, W. C. Sessa, J. R. Vane, ibid. 351, 477
`(1991); T. M. Cocks and J. A Angus, Naunyn(cid:173)
`Schmiedeberg's Arch. Pharmacol. 341, 364
`(1990); A T01trup, D. Svane, A Forman, Am. J.
`Physiol. 260, G385 (1991); H. Bult et al., Nature
`345, 346 (1990); J. S. Gillespie, X. Liu, W. Martin,
`Br. J. Pharmacol. 98, 1080 (1989); M. V. Ramago(cid:173)
`pal and H. J. Leighton, Eur. J. Pharmaco/. 174, 297
`(1989); A Gibson, S. Mirzazadeh, A J. Hobbs, P.
`K. Moore, Br. J. Pharmacol. 99, 602 (1990).
`6. R. Blanco et al., Am. J. Physiol. 254, 468 (1988).
`7. J.M. Polak, S. Mina, J. Gu, S. R. Bloom, Lancet ii,
`217 (1981); E. A Willis, B. Ottesen, G. Wagner, F.
`Sundler, J. Fahrenkrug, Ute Sci. 33, 383 (1983).
`8. E. A Kiely, S. R. Bloom, G. Williams, Br. J. Ural.
`64, 191 (1989) ; J. B. Roy, R. L. Petrone, S. I. Said,
`J. Ural. 143, 302 (1990); K. P. Juenemann et al.,
`ibid. 138, 871 (1987); W. D. Steers, J. McConnell,
`G. S. Benson, ibid. 132, 1048 (1984).
`9. R. S. Pickard, P. H. Powell, M. A Zar, Br. J.
`Pharmacol. 104, 755 (1991); L. J. lgnarro et al.,
`Biochem. Biophys. Res. Commun. 170, 843
`(1990) ; N. Kim, K. M. Azadzoi, I. Goldstein, S.
`deTejada, J. Clin. Invest. 88, 112 (1991); F.
`Holmquist, H. Hedlund, K. E. Andersson, Acta
`Physiol. Scand. 141, 441 (1991); J. Rajfer, W. J.
`Aronson , P. A Bush, F. J. Dorey, L. J. lgnarro, N.
`Engl. J. Med. 326, 90 (1992).
`10. J. S. Gillespie and L. Xiaorong , Br. J. Pharmacol.
`97, 453P (1989); N. 0. Sjostrand, J. Eldh, V. E.
`Samuelson, S. Alavanta, E. Klinge, Acta Physiol.
`Scand. 140, 297 (1990).
`11 . A L. Burnett, C. J. Lowenstein, D. S. Bred!, T. S.
`K. Chang, S. H. Snyder, data not shown.
`12. D. M. Quinlan, R. J. Nelson, A W. Partin, J. L.
`Mostwin, P. C. Walsh, J. Urol. 141, 656 (1989).
`13. F. Holmquist, C. G. Stief, V. Jones, K. E. Ander(cid:173)
`son, Ac ta Physiol. Scand. 143, 299 (1991) .
`14. G. S. Benson, World J. Ural. 1, 209 (1983).
`15. K. Nozaki et al., J. Cereb. Blood Flow Metab., in
`press.
`16. C. G. Li and M. J. Rand, Eur. J. Pharmaco/. 191,
`303 (1990) .
`17. P. C. Walsh and P. J. Donker, J. Ural. 128, 492
`(1982) .
`18. W. G. Dail, D. Trujillo, D. De la Rosa, G. Walton,
`Anat. Rec. 224, 94 (1989) .
`19. A M. Emond, R. Holman, R. J. Hayes, G. R.
`Serjeant, Arch. Intern. Med. 140, 1434 (1980) .
`20. J. M. Palomer, H. Halikiopoulos, R. Thomas, J.
`Ural. 123, 680 (1980).
`21 . D.S. Bred! eta/., Nature351, 714 (1991).
`22. Supported by U.S. Public Health Service grants
`MH-18501, DA-00266, and DK-19300, contract
`DA-271-90-7408, and Research Scientist Award
`DA-00074
`(S.H.S.),
`training grant DK-07552
`(A.LB.). training grant GM-07309 (D.S.B.), NIH
`Physician Scientist Award HL-02451 (C.J.L.), and
`a gift from Bristol-Myers Squibb.
`
`3 March 1992; accepted 28 May 1992
`
`SCIENCE
`
`• VOL 257
`
`• 17 JULY 1992
`
`403
`
`3 of 3
`
`