- - -- -- - - - -- · -- - - - --
`
`--LETTERSTONATURE- - -- -- - - -~N~A~11J=R=E~ V=O=L.---'3"'-13C-.c-3 ~JA= NccU=..cA.=R=Yc-=19=85
`
`4-
`-5
`
`process and different cells may use different mechanisms to
`resist injury. For example, surface epithelial cells secrete bicar(cid:173)
`bonate and mucus, and mucus may provide an 'unstirred layer'
`at the cell surface to stabilize a pH gradient generated by HCO3
`11
`secretion 10
`• Furthermore, such cells replicate rapidly and are
`•
`sloughed off from regions of mucosa! injury, with epithelial
`2
`integrity quickly being restored by migration of adjacent cells 1
`•
`As a result of these other mechanisms, surface epithelial cells
`may be less dependent on an H+ -impermeable apical surface
`than are chief cells, which replicate slowly, lack a surface mucus
`coat and appear not to secrete bicarbonate. (Neither trans(cid:173)
`mission3 nor scanning electron microscopy reveal a mucus coat
`on the apical surface of chief cell monolayers. Although cells
`
`staining with periodic acid Schiff (PAS+) constitute 40% of the
`plated cells, only I% of monolayer cells are PAS+.)
`The study of the mechanisms underlying gastric mucosa!
`defence has been hampered by the cellular heterogeneity of the
`tissue, which makes intact mucosa a complex model. As chief
`cell monolayers in primary culture display barrier function, this
`system provides a simple model for studying this element of
`mucosa! defence. Our present findings indicate that chief cells
`in culture re-form an apical surface virtually impermeable to
`protons; we suggest that this mechanism underlies chief cell
`resistance to acid-peptic injury.
`This work was supported by NIH grant AM-30444 and a
`grant from the Stiftung Volkswagenwerk.
`
`Recci 11ed 3 ~plcmber; accepted 21 September 1984.
`I. Davenport, H. W., Warner, H. A. & Code. C. F. Gastroenterofoty 41, (42-152 (1964 ).
`2. Davenport, H. W. in Progress in Ga.stroenierology Vol. 2 (ed. Glass, G . 8. J.) 42- 56 (Grune
`& Stratton, New York, 1970).
`3. Ayalon, A., Sanders, M. J., Thomas, L. P .• Amirian, D. A. & Soll, A. H. Proc. natn. Acad.
`S<L U.S.A. 79, 7009-7013 (1982).
`4. Sanders, M. J., Amirian, D. A., Ayalon, A. & Soll, A. R Am. J. Physinl. 245, G641-G6'6
`(1983).
`5. Powell, D. W. An:i, J. PhysioL 141, 0275-0288 (1981 ).
`6. Ma.chcn, T. E., Silen, B. & Forte, J. Am. J. Physiof. 2.14, E228-E2J5 (1978).
`
`7. Wright, E. M. 4'. Diamond, J.M. BiocJiim. biophys. Ac1a 163, 57-74 ( J978 ).
`8. Mandel, L. J. Am. J. ffi vyKJ/. 23S, C35-C48 (1978).
`9. Schiessel , R., Matthew~,',.. Ban:ilai, A., MerhaY, A. & Silcn, W. Nature 213,671-673 ( 1980).
`to. Flcmstrom. G .• Hcylings, J. R. & Gamer, A. Am. l. Physiol. l-42. GIOO- GI to (1982).
`l I. Grossman, M. l . in Basic Mei·hanisms o/ Oo.strointt'slinal Mucosal Cell Tnju.ry and Prott!ctiot1
`(ed. Harmon, J.) 323'.-326 (Williams & Wilkins, New York, 1981).
`12. Rutten, M. J. & Ito, S. Am. J. PhysioL 244, Gl71-Gl82 (1983).
`13. Soll, A.H . J. clin. Invest. 61, 370-380 (197R).
`
`Calcitonin gene-related
`peptide is a potent vasodilator
`S. D. Brain*, T. J. Williams*, J. R. Tippinst,
`H. R. Morrist & I. MacIntyre*
`Department of Pharmacology, Institute of Basic Medical Sciences,
`Royal College of Surgeons of England, Lincoln's Inn Fields,
`London WC2A 3PN, UK
`t Department of Biochemistry, Imperial College,
`London SW? 2AZ, UK
`t Department of Chemical Pathology, Royal Postgraduate Medical
`School, DuCane Road, London Wl2 OHS, UK
`
`concentration in the medium4
`• Intracerebroventricular injection
`of CGRP stimulates sympathetic outflow in vivo as evidenced
`by increased levels of circulating noradrenaline, tachycardia
`and elevated blood pressure3
`• Thus, CGRP seems to be a
`neuropeptide. However, in these experiments, control injections
`of CGRP given intravenously (i.v.) were shown to induce tachy(cid:173)
`cardia accompanied by lowered blood pressure, suggesting
`peripheral vasodilatation3
`• Our experiments offer an explanation
`for these control observations.
`Human CGRP was extracted from medullary thyroid car(cid:173)
`cinoma tissue and purified to homogeneity, as described pre(cid:173)
`viously'. Chemically synthesized rat CGRP was obtained as a
`commercial preparation (Bachcm). The structures of human
`CGRP' and rat CGRP1 are shown below.
`
`Human CGRP rs
`
`s7
`NH 2- Ala-Cy s-Asp-Th r-A l a-T'nr-Cy s-Va 1-T h r-H i s-Arg-Leu-
`Al a-Cly-Leu- Leu-Se r-Ar,g- Se r- Gl y-Gl y-Val- Va l-Lys-Asn(cid:173)
`
`Ag n- Ph e-V a 1-Pro-Th r- As n-V a 1-G 1 y-Se r - Ly ti-Ala-Ph e-AMI DF.
`
`Rat CGRP
`
`7
`1 5
`~ 112 - Ser-C:y s- As n-Th r-Al a-Th r-C y s-Va 1-Th r-H 1 s-Arg-Leu
`Al a- Cl y-Leu- Leu- Ser-Arg-Ser-Gly-G l y-Va 1-Va 1-Lys- As p(cid:173)
`
`As n-Pl1~V a 1-Pro-Th r-As n- Va l-Gl y-Ser-G lu- Ala- Phe- ,\Ml DE
`
`A novel peptide, calcitonin aene-related peptide (CGRP), has been
`predicted to result from alternative processlna of the primary RNA
`1
`transcript of the calcitonin gene in the rat 1
`• Several lines of
`'
`evideace sugest that CGRP is a transmitter in the central and
`peripheral nervous system2-4. Hum(cid:127) n CGRP has been Isolated
`and characterizect5, and shown to have potent effects on the heart6
`•
`The obsen-ations prese(cid:127) ted here indicate that human and rat
`CGRP also have potent effects on blood vessels. Intradermal
`injection of CGRP in femtomole doses induces microvascular
`dilatation resulting In Increased blood flow, which we have detected
`in the rabbit by usin11 a 133Xe clearance technique7
`,8. 1(cid:127) human
`skin, CGRP induces persistent local reddening. Microscopic
`observation of the hamster cheek pouch9 in vivo revealed that
`topical application of CGRP induces dilatation of arterioles.
`Furthermore, CGRP relaxes strips of rat aorta in vitro by an
`endothelial cell-dependent mechanism. Therefore, we suggest that
`local extravascular release of CGRP may be Involved in the
`physiological control of blood flow and that circulatina CGRP
`may co(cid:127) trlbute to hyperaemia in certain pathological conditions.
`The primary RNA transcript of the calcitonin gene can be
`processed to give alternatively spliced messenger RNA (mRNA)
`products: one of these encodes the precursor bf the 32-amino
`acid thyroid hormone, calcitonin, and another the precursor of
`2
`a putative neuropeptide, the 37-amino acid CGRP1
`• The
`•
`primary structure of rat CGRP has been predicted by analysis
`of the calcitonin gene coding sequence1
`• lmmunoreactivity to
`the C-terminus of the predicted CGRP peptide has been detected
`in discrete areas of the central and peripheral nervous system
`of the rat2. Rat trigeminal ganglionic cells release immunoreac(cid:173)
`tive CGRP in vitro, which is increased by a high potassium
`
`• Present address: Section of Vascular Biology, MRC Clinical Research Centre, Watford Road,
`Harrow. Middlese,; HAI 3UJ, UK.
`
`Both peptides were tested for vasodilator activity using a
`8
`multiple-site 133Xe clearance technique in rabbit skin 7
`• The
`•
`peptides were mixed with a solution of 133Xe and injected rapidly
`in 100-µl volumes into marked sites on the clipped dorsal skin
`of New Zealand White (NZW) rabbits. Intradermal injections
`were given in random-block order according to a predetermined
`balanced site pattern, with six replicates for each of seven
`treatments per animal. CGRP doses were tested against standard
`doses of a known potent vasodilator, prostaglandin E2 (PGE2 ),
`and saline controls. Animals were killed after 15 min, the dorsal
`skin excised and injected sites removed with a 17-mm diameter
`punch. Skin samples and samples of injection fluids were
`measured for radioactivity in an automatic y-counter. Changes
`in blood flow induced by the test substances were determined
`as described in Fig. I legend. Both human and rat CGRP
`increased local blood flow with potencies similar to that of the
`highly active PGE2• Blood flow was significantly increased with
`
`
`© Nature Publishing Group1985
`
`

`

`:..:N:..:.ATIIRE==.=---cV--=O=L·c..c30-"13'----3'----J=ANU=AR~Y~19~85'----- - - ---LETTERSTONATURE-----------------------""H
`
`,;;-
`)l!OO
`~ .,
`.!l
`<> 80
`>( t 60
`C c:: "8 40
`
`~
`
`C
`:C 20
`. =!
`ii
`~ 0
`u
`.5
`
`a
`
`b
`
`---------------*
`-13
`- 14
`-13
`-14
`-II
`-12
`log dose (mo! per 0.1 ml)
`
`Fi11. 1 Effect of human and rat CGRP on blood flow in rabbit
`skin. a, A comparison of increased blood flow induced by human
`CGRP (e) with that induced by PGE2 ((cid:127))
`in seven rabbits. Both
`CGRP and PGE2 increased blood flow in a dose-dependent man(cid:173)
`ner, compared with the control (phosphate-buffered saline (PBS),
`broken line). Human CGRP at 25 fmol induced a significant change
`in blood flow (P < 0.02, paired Student's t-test). Human calcitonin
`(IOpmol, •) and katacalcin (IOpmol, A) were also tested in four
`rabbits but neither had any detectable effect on blood flow. b, A
`comparison of increased blood flow induced by rat CGRP (0)
`and PGE2 ((cid:127))
`in five rabbits. Rat CGRP at 250 fmol induced a
`significant change in blood flow (P < 0.01, paired Student's I-test).
`Methods: Experiments were carried out at room temperature (22-
`25 °C) with NZW rabbits (3.5-4 k~) anaesthetized with Saffan
`{0.5 ml per kg). Agents mixed with 1 3Xe (5-10 µCi per injection)
`were injected intradermally in 100-µI volumes into the clipped
`back skin in random-block order according to a balanced site
`pattern. The agents tested were CGRP (0.025-2.5 pmol), PGE2
`(0.1-10 pmol) for comparison, and PBS as a control; six replicates
`of each dose were tested in each rabbit. After 15 min, the animal
`was killed with an i.v. barbiturate overdose (Expiml), the back
`skin was removed and a 17-mm diameter punch was used to excise
`injection areas. Samples of skin and injection fluid, under paraffin
`oil in sealed tubes, were counted immediately in an automatic
`y-counter. The change in local blood flow was calculated as the
`per cent increase above PBS controls from the equation: I 00
`(In 133Xe count of saline-injected skin - In 133Xe count of agent(cid:173)
`injected skin)/(ln 133Xe count of 0.1 ml injection fluid-In '"Xe
`count of saline-injected skin). Results are expressed as mean±
`s.e.m. for n rabbits.
`
`.
`C .
`
`80
`
`60
`
`40
`
`20
`
`.!!
`
`. .
`"
`Fig. 2 Effect of the prostaglandin synthesis inhibitor, indo(cid:173)
`methacin, on increased blood flow induced by human CGRP
`and arachidonic acid (AA) in rabbit skin. lndomethacin (Indo,
`2.8 nmol) or PBS was injected intradcrmally 15 min before the test
`compounds. 133Xe was mixed with the test compounds, and human
`CGRP (2.5 pmol) or arachidonic acid (3.3 nmol) was then injected
`into the same sites with a further dose of indomethacin (2.8 nmol)
`as required. We determined the responses over a 15-min period as
`described for Fig. l . lndomethacin had no effect on the response
`to CGRP, but caused a large inhibition of increased blood flow
`induced by arachidonic acid (P < 0.02, paired Student's t-test).
`Indomethacin alone had no significant effect on blood flow. Results
`are expressed as mean± s.e,m. for four rabbits.
`
`G~m i n l
`0.5g
`A
`
`.d
`
`,tJ.
`
`. .
`~ ~ : .
`.
`
`A
`
`g 4Q
`-~
`
`-8
`NA
`
`-9 wa,h
`-IO
`-12 -II
`log CGRP x 2.5 (M)
`
`- 12
`
`-9
`- 10
`-II
`log CGRP (M)
`
`-8
`
`Fig. 3 Response of rat aortic rings to human and rat CGRP. Rat
`thoracic aorta was cleared of fat and connective tissue and cut
`transversely into rings 3 mm wide. Each ring was cut to form a
`the
`taking care not to damage
`strip of circular muscle,
`endothelium13
`• The strip was then held in a 10-ml organ bath in
`oxygenated Krebs' solution containing EDTA (30 µM) 13
`• The
`tissue was suspended from an isometric transducer (Lectromed
`UFJ) under lg tension and was allowed to equilibrate for 90 min
`before the experiment. a, Responses of aorta with undamaged
`endothelium; b, responses of aorta after the endothelium has been
`gently rubbed with a microspatula before mounting in the organ
`bath. The aortae were induced to contract by adding noradrenaline
`(NA, 10-• M) to the bath, which sometimes induced some spon(cid:173)
`taneous activity in preparations with intact endothelium. At the
`point of maximum contraction, a control response was obtained
`by adding cumulative concentrations of ACh (Io-• - 10-• M) to the
`bath (results not shown). Only when the endothelium was intact
`did ACh produce a dose-related relaxation of the preparation;
`-50% relaxation was achieved with 10-7 M and 100% relaxation
`with IO"" M ACh. No relaxation was induced by A Ch after removal
`of the endothelium. Human CGRP, added in cumulative con(cid:173)
`centrations, produced a dose-related relaxation of the aorta with
`an intact endothelium; the aorta in which the endothelium had
`been destroyed was not affected. Responses to CGRP took longer
`to reach equilibrium than those to ACh ( - 3 min for CGRP com(cid:173)
`pared with l min for ACh). c, Dose-response curves for human
`and rat CGRP on NA-contracted rat aorta to show the effect of
`indomethacin. Indomethacin (2.8 x 10"" M final concentration)
`was added to the bath IS min before the addition of NA and the
`cumulative doses of CGRP. e, Relaxation induced by human
`CGRP ; •, human CGRP in the presence of indomethacin; 0,
`rat CORP in the presence of indomethacin. Each
`rat CGRP; (cid:144),
`point is the mean± s.e.m. of 4-7 observations. Indomethacin did
`not affect relaxation induced by Ach, but partially inhibited that
`induced by CGRP (• P < 0.05; t P < 0.005, unpaired Student's t-
`test) .
`
`Fl&, 4 Effect of intradermal injection of CGRP into human skin.
`Synthetic human CORP (15 pmol; Sandoz) was prepared in 50 µI
`sterile PBS. Histamine (HA, 500 pmol per 50 µI) was injected with
`PBS as a control. The figure shows the appearance of the injected
`arm 60 min after injection, at which ti.me the response to histamine
`had faded. The dilator response to CGRP is clearly visible as local
`reddening, spreading slowly, with pseudopodia. This response
`persisted for 5-6 h.
`
`
`
`© Nature Publishing Group1985
`
`

`

`56~------------------LETTERSTONATURE-----------'-'N'--'ATI.Jc.=R.:::E:_V::..;Oc'.!L"-.. _:c31c::3...::3...::1,::,AN'--'U"'AR""'-'Yc...!.".19~85
`'-'
`comparison of endothelium-dependent relaxation induced by
`a 25-fmol dose of human CGRP (P< 0.02, Student's I-test). No
`human and rat CGRP is shown in Fig. 3c; each produced
`changes in blood flow were detected in risponse to either human
`as 2.5 x 10- •1 M.
`low
`calcitonin or katacalcin, a 21-amino acid peptide derived from
`as
`at concentrations
`relaxation
`Indomethacin pretreatment partially suppressed the relaxation
`• Human or rat CGRP, at doses up to
`the same precursor' 0
`induced by both CGRP preparations, but not that induced by
`2.5 pmol, did not cause microvascular plasma protein leakage,
`ACh (see Fig. 3 legend). Thus, prostaglandin generation may
`as measured by the local accumulation of 1251-albumin injected
`be responsible for a part of the response to CGRP. A contribu(cid:173)
`• This shows that the response to CGRP does not involve
`i.v.8
`tion by prostaglandins should be considered in future studies
`the release of histamine, which increases vascular permeability.
`of CGRP effects, although there was no evidence of such a
`To test whether the dilator responses to CGRP in rabbit skin
`contribution in the rabbit skin experiments.
`were caused by release of endogenous prostaglandins (both
`The potent vasodilator effects of CGRP were also observed
`PGE 2 and prostaglandin 12 (PGii) are very potent dilators in
`in human volunteers (Fig. 4). Synthetic human CGRP (Sandoz,
`), human CGRP was injected into skin sites preinjec(cid:173)
`the rabbit8
`Basie, Switzerland) was used, and eighteen doses injected
`ted with indomethacin, which inhibits prostaglandin synthesis.
`intradermally into the volar surface of the forearm of three
`Arachidonic acid was used as a positive control; intradermally
`non-atopic subjects (S.D.B., J.R.T. and T.J.W.) . CGRP (15
`injected arachidonic acid is converted to a dilator prostaglandin
`f mol-15 pmol per 50 µI) induced vasodilation that was seen as
`. Figure 2 shows that increased blood How
`11
`in this system8
`•
`local reddening within I min after injection. The reddening
`induced by human CGRP in the skin was unaffected by
`became more intense, while spreading slowly. The response
`indomethacin, whereas responses to arachidonic acid were
`lasted for 1 h with a 15-fmol dose and 5-6 h with a 15-pmol
`inhibited. In this respect the effect of CGRP is similar to that
`dose. At the higher dose, local reddening spread in streaks,
`of another l?utative neuropeptide, vasoactive intestinal polypep(cid:173)
`• A transient
`suggesting lateral dispersion via lymphatic vessels 14
`tide (VIP)'·.
`patchy 'flare' surrounding the local reddening was sometimes
`We also investigated the possibility of a secondary release of
`observed, but there were no obvious weals over this dose range.
`other vasodilators in the skin by CGRP. The vasodilator
`A higher dose of CGRP, 250 pmol, induced a weal and flare
`responses to CGRP were found to be unaffected by the
`response comparable with that induced by histamine at
`acetylcholine (ACh) antagonist, hyoscine (0.1 nmol per site)
`500 pmol. However, local reddening induced by histamine was
`and the f!-adrenergic antagonist, propranolol (0.1-1 nmol per
`less intense and faded with the weal and flare within an hour,
`, ACh, isoprenaline,
`site). Moreover, as reported previously 12
`whereas local reddening induced by the high dose of CGRP
`purines and substance P are weak vasodilators in rabbit skin:
`persisted for I 0-12 h.
`CGRP was at least 1,000 times more potent than ACh, ATP,
`These experiments demonstrate that CGRP is a potent vasodi(cid:173)
`ADP, adenosine, 5-hydroxytryptamine and substance P, and
`lator. Relaxation of the aorta in vitro seems to depend on
`I 0-100 times more potent than the synthetic f!-adrenergic
`endothelial cells, which may be relevant to the responses of
`stimulant isoprenaline.
`major vessels and microvessels to circulating CGRP. Circulating
`Direct microscopic observation of the hamster cheek pouch
`CGRP could be involved in hyperaemia such as the Hushing
`preparation in vioo 9 confirmed that increased 133Xe cfearance
`sometimes found in patients with medullary thyroid car(cid:173)
`induced by CGRP was caused by dilatation of the major resist(cid:173)
`cinoma 15, where CGRP has been detected in the plasma5.
`ance vessels, the arterioles. The microvascular bed was super(cid:173)
`Whether arteriolar dilatation induced by local cxtravascular
`solution containing
`saline
`fused with modified Krebs
`CGRP (as in the skin and hamster cheek pouch experiments)
`indomethacin (2.8 x 10-6 M). Topical application of CGRP
`is endothelial cell-dependent remains open to question for this
`(2.5 pmol per 10 µl) to preparations with a high arteriolar tone
`and other dilators such as ACh. It is possible that CGRP
`(spontaneous, or induced with a superfusion of noradrenaline
`receptors are present on the abluminal surface of endothelial
`(1.5 x 10-7 M) in two experiments) caused arteriolar dilatation
`cells or, alternatively, arteriolar smooth muscle cells may have
`to 39.3 ± 6.3 µm
`from an initial diameter of 12.8 ± 3.5 µm
`their own CGRP receptors. Immunoreactive CGRP has been
`(mean±s.e.m., n = 4 animals; P < 0.02, paired Student's I-test).
`observed in nerve fibres associated with blood vessels2
`, which
`The arteriolar dilatation in response to CGRP was maximal
`raises the possibility that CGRP is involved in neurogenic
`2 min after application and persisted for at least 5 min.
`. CGRP may be the endogenous mediator of the
`inflammation 16
`lmmunoreactive CGRP has been detected in extracts of rat
`skin flare response, as an alternative to substance P which has
`aortae (20 pmol per g in a pooled extract of 12 aortae; S.
`been proposed to be released from nerve terminals by antidromic
`Wimalawansa, unpublished observation). We have discovered
`impulses passing down branches of afferent nerves 17
`•
`that CGRP has a relaxant effect on artery preparations in vitro
`We know of no other reports of vasodilators producing sig(cid:173)
`by an endothelial cell-dependent mechanism. It has been repor(cid:173)
`nificant effects at low femtomole doses. The very high dilator
`ted previously that some substances induce vasodilatation by
`potency of CGRP, shown here in four unrelated species, implies
`stimulating the release, from vascular endothelial cells, of an
`a function for this novel peptide. We consider it worthwhile to
`unknown secondary mediator which causes smooth muscle
`investigate the possibility that CGRP has a role in the physiologi(cid:173)
`• CGRP was tested in this manner using 3 mm-wide
`relaxation13
`cal control of vascular tone and blood flow.
`transverse strips of rat thoracic aorta in vitro. Noradrenaline
`We thank Dr S. Girgis for extraction of human CGRP, Dr
`(lo-• M) was added to the bathing medium to induce tone;
`M. Panico and Mr T. Etienne for purification and quantitation
`human CGRP was then added in increasing concentrations. A
`of CORP. The human skin experiments were carried out with
`progressive relaxation of the tissue was seen (Fig. 3a), which
`permission of the Ethical Committee of the Royal College of
`was slower than that induced by ACh (sec Fig. 3 legend) .
`Surgeons of England. We thank the Wellcome Trust, the British
`Relaxation was not seen in aortic strips from which the
`Heart Foundation and the MRC for financial support.
`endothelial cells had been removed, as shown in Fig. 3b. A
`
`Received 30 June ; accc:ptcd 28 September 1984.
`I. Amara, S. 0 ., Jonas, V., Rosenfeld, M. G ., Ong. E. S. & Evans, R. M. Nattm! 2911, 240....:244
`(1982) .
`2. Rosenfeld, M. G. et al. Na ture JOA, 129-135 ( 1983 ).
`3. Fi!lbcr, L.A. et al. Nature 305, 534- '36 ( 1983).
`4. Mason, R. T. et al. Nature• 653 - 655 (1984).
`5. Mortl!i, H. R. et al. Nature 308, 746- 748 (1984).
`6. Etienne, T . et al J. PltysioL, Umd. 351, 48P ( 1984).
`7. Williams, T. / . J. l'l,ysiol., Lmd. 25"1, '-5P ( l976).
`
`8. Willii&ots, T. J. Br. J. f'flarma c. 6!, 5 17- 524 ( 1979).
`9. Duling, 8. R. Microva.sc. Res. 5, 423-42!1'(1973)
`IO. Hillyard, C. J. et aL Lancet I, 846- 848 (1983).
`11. William!i, T. J. & Peck, M. J. Norure 270. 530- 532 (1977 ).
`12. Williams. T. J. Br. J. Pf,annac. '17, 505-509 (1982) .
`13 , Furch1o tt, R. F. & Zawadzki, J. V. Naiu,e 1:18, 373- 376 ( 1980).
`J4. Solomon, L. M .• Juhlin, L. & Kirschenbaum, M. 8. J_ i11vitst. Denn. Sl, 280- 282 (1968).
`15. Cohen, S. L., MacIntyre, I., Grahame-Smith, D. & Walker,J. G. Lan,·et II, 1172-1174 (1973 ).
`16. Jancso, G., K.Jl"aly, E. & Jan~ o-Gabor, A. Nature :?70, 741 -743 (1977 ).
`17. Lcmbcct, F. &. Holzer, P. No.1111y n.Schmi~d~rgJ Archs Pltarmllk. 310.. 175- 18:) ( 1979 ).
`
`
`
`© Nature Publishing Group1985
`
`