22-22X/83/813-236$2./ Tli JOURNAL OF INVSTIGATIV DRMATOLOGY, 81.:236-24, 1983 Copyright 1983 by The Williams & Wilkins Co. Vol. 8 1, No.3 Printed in U.S. A. Forskolin Ativates Adenylate Cylase Ativity and Inhibits Mitosis in In Vitro in Pig pidermis JUNJI TAKDA, M.D., KNJI ADACHI, M.D., PH.D., KNNTH M. HALPRIN, M.D., SATOSHI ITAMI, M.D., VICTOR LVIN, B.S., AND CLYD WOODYARD, B.SC. Veterans Administration Medial Center, Miami, and Department of Dermatology, University of Miami Shool of Mediine, Miami, Florida, U.S.A. The novel adenylate ylase ativator forskolin aused rapid and high intraellular aumulation of yli AMP in a floating skin (epidermal) slie system. Inreased AMP levels were also deteted in the media. Addition of a phosphodiesterase inhibitor to forskolinontaining medium aused only a slight inrease in the intraellular AMP level and forskolin itself did not inhibit phosphodiesterase ativity. Ka of forskolin for epidermal adenylate ylase was about 2-3 x 1-5 M. This forskolin ativation was rapidly reversed after washing. The forskolin stimulation (Ka 5 x 1-5 M) was also found when tested with an epidermal membrane preparation whih ontained the atalyti unit of adenylate ylase but laked either the GTP or reeptor stimulation. With the epidermal slie system, the ombination of forskolin and epinephrine (or histamine) stimulated adenylate ylase synergistially. The data suggest that forskolin ativates not only the atalyti unit but also the nuleotide regulatory protein or the reeptor-regulatory protein omplex of the adenylate ylase system. The AMP aumulation aused by forskolin produed a dose-dependent mitoti inhibition of epidermal ells in an in vitro outgrowth system. This inhibitory effet was reversible 48 h after washing out the forskolin. Forskolin is a novel diterpene isolated from the roots of Coleus forskohlii [1). Although t he therapeuti use of extrats from plants of a related family was desribed in anient Hindu medial texts, the isolation and identifiation of fo rskolin as a unique adenylate ylase ativator are only reent events [1-3]. As far as t he mehanism of ation is onerned, forskolin has been reported to ativate diretly the atalyti unit of the adenylate ylase system [3,4]. The latest report [5] suggests a n additional mehanism, suh as partiipation of the nuleotide regulatory protein of the adenylate ylase unit for full ativation. The ation of forskolin on the adenylate ylase system of different ells and tissues has been reported [2-1] but its ation on the skin adenylate ylase system has not yet been desribed. Our laboratory has reported on various surfae reeptor-adenylate ylase omplexes and their possible relationship to psoriasis [11-15]. Sine the {3-adrenergi-yli AMP system is defetive in psoriasis lesions, any drug or hemial that inreases the AMP level in skin may have potential therapeuti value. The fat that forskolin is effetive in intat (unhomogenized) t issues [3,5,9,1] further enouraged us to use a simple epidermal slie system in addition to the membrane preparation to study its ation. MATRIALS AND MTHODS Materials [2,8-3 H]AMP (sp at 36.4 Ci/ mmol) was purhased from New ngland Nulear (Boston, Massahusetts), [a - 32 P]ATP (sp at 4 Ci/ mmol) from Amersham (Arlington Heights, Illinois), AMP radioimmunoassay kit from Collaborative Researh, In. (Waltham, Massahusetts), and forskolin from Calbiohem (LaJolla, California). All other hemials used were of the highest grade available. Methods Skin slies (approximately 8% epidermis) were taken from the baks of domesti pigs using a keratome with the utting blade adjusted to.2 mm or.3 mm. The skin slies were ut into 5 X 5 mm squares at 4 and floated keratin layer up on Hanks' balaned salt solution. After preinubation for 3 min at 37, the squares were inubated in forskolin and other hemials. After inubation, they were frozen between piees of dry ie to stop the reation and AMP ontents were measured by radioimmunoassay [16] with iniromodifiation [17]. When AMP levels in the epidermis were high, we omitted the aetylation step of the radioimmunoassay. Crude membrane preparations were obtained by homogenization with an all glass homogenizer for 3 min in 5 mm HPS buffer (ph 8.) with.1 M DTA and 1 mm MgCb. The homogenate was entrifu ged at 2, g for 2 min and the membrane preparation was was hed 3 times by reentrifugation. The adenylate ylase ativity was assayed by slight modifiation [15] of Salomon's method [18]. whih is based on the 2-step hromatographi separation of [ 32 P]AMP. Protein was measured by the method of Lowry eta! [19] with human serum albumin as a standard. Phosphodiesterase assay was as previously desribed [2]. The basi proedures for the explant ulture of pig skin were essentially those for mouse skin [21 ] with adaptation to pig skin [22-24]. In short, keratome-slie epidermis was ut into 2-mm squares. The 2 rumsquare explants were allowed to attah to glass overslips in air for 15 min. RPM! medium ontaining antibiotis and 1% fetal alf serum was then added to eah ulture and the ultures were allowed to grow for 4 days. Cultures with good growth (>3 JLID from the explant) on at least 3 sides were then transferred to RPMI medium ontaining 5% dialyzed fetal alf serum, antibiotis, and nuleosides for testing the effets of various forskolin onentrations. They were grown for 4 h in the presene of olhiine (.1!Lg/ml), fixed in aloho l, stained with hematoxylin, and the number of mitoses ounted. Manusript reeived Otober 29, 1982; aepted for publiation April 25, 1983. Supported in part by grants AM 17179 from the National Institutes of Health and by the Dermatology Foundation of Miami. Reprint requests to: Dr. Kenji Adahi, Veterans Administration Medial Center, 121 N.W. 16th Street, Miami, Florida 33125. Abbreviations: C: atalyti unit IBMX: 3-isobutyl-1-methylxanthine H PS: N-2-hydroxyethylpiperazine-N' -2-ethanesulfon i aid N: nuleotide regulatory protein R: reeptor site RSULTS xperiments with pidermal Slies The inubation of pig skin squares with forskolin aused marked stimulation of epidermal adenylate ylase (Fig 1). The AMP aumulation ourred in both t he t issue and medium. The intraellular AMP levels reahed 3 pmol/mg protein after 45-min inubation and the levels in the medium were over 1 pmoljmg protein after 1-h inubation. There was, however, a delay in transferene of AMP from the tissues to the medium. After 1 min of inubation, AMP in the medium that had 236
Sept. 1983 ADNYLAT CYCLAS ACTIVATION BY FORSKOLIN IN PIDRMIS 237 " +J 3... > 2... Q.) " +J... OJ.5 1 3 6 min FIG 1. Time ourse of the effet of forskolin (15 JLM) on the intraellular (- ) and medium (- ) AMP levels. Details of experimental onditions are desribed in Materials and Methods. Intraellular AMP levels are averages ± S of 4 determinations. Medium AMP levels are averages of 2 determinations. leaked from the tissues was minimal while the intraellular AMP level had reahed over 1 pmoljmg protein. To rule out the possibility that the AMP inrease was due to an inhibition of its breakdown rather than ativation of adenylate ylase, the effet of a phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine (IBMX), was tested. Fig 2 shows time ourses of the forskolin ativation in the presene and absene of IBMX. The addition of IBMX aused only a minimal additional stimulatory effet. In order to examine a diret inhibitory effet of forskolin on phosphodiesterase, we measured phosphodiesterase ativity in epidermal homogenates with and without forskolin. Fig 3 shows that forskolin (1 p.m) did not have an inhibitory effet on the low Km phosphodiesterase (AMP onentration at 1p.M). This was also true at the high K.n (AMP onentration at 1 p.m, data not shown). Forskolin was also not inhibitory to a ommerial phosphodiesterase preparation (Boehringer). Fig 4 shows the onentration urve for forskolin ativation of adenylate ylase. This ativation by forskolin was dosedependent with a maximal response ourring at 1 p.m and an apparent Ka value for AMP at 2-3 p.m. This ativation was reversible (fig 5). When the epidermal slies were inubated for 45 min with forskolin (15 p.m) and then transferred to plain Hanks' balaned salt solqtion medium at 37 C, the AMP level in the epidermis dereased very quikly. However, even at 2 h after the transfer, the AMP level in the epidermis was still higher than the basal level (6 pmol/mg protein vs the basal level of 1 pmol/mg protein). To examine the effets of a reeptor stimulator with forskolin, we inubated epinephrine or histamine with forskolin. Sine the AMP inrease aused by epinephrine or histamine is 3 6 min FIG 2. ffets of IBMX (1 mm) on forskolin stimulation: forskolin (15 JIM) (-), forskolin + IBMX (e- e). Data are averages ± S of 4 determinations. "5.._. lo. 6 ' 3oo.... 5 1 15 min FIG 3. Phosphodiesterase assay with forskolin (1 JIM) (- ) or with IBMX (1 11M) (L'. - L'.) or ontrol (-). Skin homogenate (14 1-'g) was used for this assay. AMP onentration was 1 J-<M for low Km enzyme. transient and reahes a peak by 5 min, we hose a 3-min inubation period. The onentrations of stimulators were at maximum stimulatory doses to examine the possible additive effets. As shown in Table I, the ombination of forskolin and epinephrine yielded a synergisti effet whih was muh more
238 TAKDA T AL Vol. 81, No. 3........ '> -.. 15 1 5 1:Jo-7 1o-s 1-3 M FIG 4. Stimulation of epidermal adenylate ylase by various forskolin onentrations. Details of experimental onditions are desribed in Materials and Methods. Inubation t ime for forskolin ativation was 1 min. Data are averages of 2 different skin slies in dupliate for eah assay.... 3 e 2oo... ) -... 1..! : \ \. ble II). Propranolol itself had a slight additive effet to forskolin ativation. A speifi histamine (H 2 ) bloker, metiamide, had no effet on forskolin ativation nor on forskolin synergism when ombined with epinephrine. xperiments with pidermal Homogenates To examine a diret ation of forskolin on the atalyti unit of the adenylate ylase system, a rude membrane fration was prepared by a glass-glass homogenization of epidermal slies. With this preparation, in whih t he guanine nuleotide stimulation of t he adenylate ylase system was no longer demonstrable, forskolin showed onsistent stimulation of the atalyti unit, i.e., the AMP inreases were both time- and dose-dependent. The time ourse showed an inrease for t he 9 min thus far tested (with a linear rate up to 6 min). Forskolin ativation of adenylate ylase at different onent rations with the membrane preparation was nearly idential to the onentration urve obtained with epidermal slies (data not shown). An apparent Ka value for AMP was 5 X 1-5 M. In one membrane preparation we ompared the degree of stimulation due to forskolin (1 JIM), guanyl-{3, -y-imidodiphosphate (1 JIM), sodium fluoride (1 mm), and epinephrine (5 JIM) at their maximum stimulatory onentrations. Only fo rskolin and sodium fluoride stimulated the membrane preparation to 11% and 45% respetively and the simultaneous addition of both forskolin and sodium flu oride did not exeed the 11% stim ulation due to forskolin alone. xperiments with pidermal xplant Outgrowth Culture System Fig 6 shows t he effets of a 4-h exposure to forskolin on epidermal ell mitoses. Dose-dependent inhibit ion of mitoses down to a onentration of 1 x 1-6 M was seen. The reve rsal of this inhibition was tested, and Fig 7 shows omplete reovery of mitoti figu res at 48 h after the washing. It is of interest that during t he first 4 h after removal of t he fo rskolin, mitoti inhibition ontinued to inrease. TABL I. The ombinations of forsholin and epinephrine or histamine on the epidermal adenylate ylase AMP level Addition None Forskolin (15 I'M) pinephrine (5 I'M) Histamine (55,uM) Forskolin + epinephrine Forskolin + histamine Forskolin + epinephrine + histamine AMP level (pmol/ mg protein for 3-min inubation) <1. 4.7 ± 2.5 14.5 ± 2.4 8. ± 1. 9.3 ± 6.9 67.3 ± 8.6 12.5 ± 14.9 a Details of experimental onditions a re desribed in Materials and Methods. T he ski n slies were inubated for 3 min. T he values shown represent averages + S of 4 determinations. I 6 12 I 1 18 min FIG 5. Reversal of the effet of forskolin. After forskolin (15 I'M ) stimulation for 45 min, skin slies were extensively washed and transfe rred to plain Hanks' balaned salt solution at 37. Data are averages of 2 different skin slies in dupliate for eah assay. than the mathematial sum of the two separately, as did the ombination of forskolin and histamine. The synergisti effet of epinephrine and forskolin was partially, but not ompletely, inhibited by the speifi {3-adrenergi bloker, propranolol (Ta- TABL II. The effets of speifi blokers on forshol(n-stimulated adenylate ylase Addition None Forskolin (15 I'M) Forskolin + propranolol (1 mm) Forskolin + metiamide (1 mm ) Forskolin + epinephrine (5,uM ) Forskolin + epinephrine + propranolol Forskolin +epinephrine+ metiamide AMP level (pmol/mg protein fo r 3-min in ubation) <1. 31.6 ± 12.9 64. ± 5.6 41.3 ± 7.7 152 ± 15.4 17 ± 12.3 f79 ± 18.7 a The values shown represent averages + S of 5 determinations.
Sept. 1983 ADNYLAT CYCLAS ACTIVAT ION BY FORS KOLIN IN PIDRMIS 239 X '" () "+=i -~ ~ 3 2 1 'I I tf1o-s 1-4 M FIG 6. Inhibition of epidermal ell mitosis by forskolin. Mitoti index shown in the ordinate is the number of olhiine-arrested mitoses seen per 1 ells in ulture. The absissa indiates onentrations of forskolin added in a log sale. ah point represents an average from 1 outgrowth ultures ± S. Mitoti figures (numbers of mitoses per ulture) also yielded essentially the same urve (data not shown ). The ontrol mitoti index (in the absene of fo rskolin) was 23.7 ± 2.8. ativate t he atalyti unit diretly. The result is ompatible wit h studies on other mammalian t issues [3,4] whih have shown t he diret stimulation of solubilized C as well as of a variant, S49 lymphoma ells whih lak N. In addition to the diret ation, however, another mehanism has reently been proposed, i. e., Darfler et al [5] reported that full stimulation of t he adenylate ylase system in wild type S49 ells was due to the st imulation of C and potentiation of hormone- mediated response required an intat N. Their data suggested that fo r skolin altered t he interation of N with both Rand C. Our data with epidermal slies have shown that the ombinations of epinephrine and forskolin, or of histamine and fo rskolin stimulated adenylate ylase synergistially (T able I). The synergisti effets an arise when two agents at at different (perhaps sequential) sites along a pathway and hene effets are multipliative and not merely additive. Sine N has two funtional omponents [28-3] it is reasonable to assume that t he additional ativating site is a omplex form of R and N, or t hat of N and C. Thus fo rskolin ativation of the adenylate ylase system has been shown in all mammalian t issues thus far tested [2-1]. Both the poteny and mehanisms of stimulation appear to be ommon to all. Membrane preparations of rat erebellum striatum, ortex, heart, and liver exhibited 2- to 4-fold ativation by sodium fluoride and 8- to 15-fo ld ativation by forskolin, but those from skeletal musle, adrenal, panreas, lung, and stomah showed lesser ativation by forskolin than the ativation by sodium fluoride. Our data indiate that the epidermis belongs to the former group (sodium fluoride < fo rskolin). One of the unique ations of forskolin is its ability to stimulate in intat ells and organs suh as S49 ells [ 4,5], platelets 3 DISCUSSION Forskolin is a extremely potent and unique ativator of epidermal adenylate ylase. T he intraellular AMP aumulation aused by fo rskolin was muh different from that aused by stimulators of ell surfae reeptors suh as epinephrine [11], histamine [12], prostaglandin [13], and adenosine [14]. The AMP aumulation aused by forskolin was extremely high (up to 3 pmol/ mg protein) and persisted fo r 6 min, whereas the AMP aumulation aused by t he stimulators noted above was usually less t han 2 pmol/ mg protein and transient with a peak at 5 min. When the phosphodiesterase inhibitor (IBMX) was added to the medium wit h t he above stimulators, the AMP level reahed a muh higher level (up to 15 pmol/ mg protein), but persisted for only lo to 2 min and dereased thereafter [11,12]. However, IBMX, when added to the forskolin-ontaining medium (Fig 2), aused only a minimal additional stimulatory effet. The fat that forskolin does not inhibit phosphodiesterase (Fig 3) indiates its diret ation on adenylate ylase as a stimulator. The forskolin stimulation is reversible after washing and this is in ont rast to holera toxin ation on skin slies [1 5] whih is irreversible and requires a lag time (1 h) for ativat ion of adenylate ylase in the skin slies (15]. What is the me hanism of this stimulation? The adenylate ylase system is omposed of at least 3 omponents, a reeptor site (R), a nuleotide regulatory protein (N), and a atalyti unit (C). Hormones or neurotransmitters bind to t heir speifi Rand subsequently ativate N and C [25-27]. Sine it stimulates AMP prodution in epidermal homogenates (C) whih annot be ativated by guanine nuleotides, obviously it must 2 u 1-- I- 1 ~ - ~~~~'----'~--~'--~'~--~'--~', 1 2 3 4 5 HOURS POST WASHOUT FIG 7. Reversal of forskolin effet in ulture. Forskolin (1 X 1-4 ) was washed out after a 4- h exposure and new medium was added. Using olhiine, mitoti fi gures were olleted at - 4 h, 2-24 h, and 44-48 h after the forskolin was removed ( time). ah point is the average of 1 ultures fo r eah ontrol and for eah forskolin time period. Figu res at 48 h for the washout vs ontrol ex plants are statistially not signifia nt. Mitoti index also shows an idential urve (data not shown). - = Forskolin washout, - = ontrol.
24 TAKDA T AL [6), adipoytes [7], brain slies [3,1), and epidermis (this paper). This is in ontrast to the ations of sodium fluoride and guanine nuleotides whih do not penetrate intat membranes. Perhaps the diterpene struture is suited for membrane penetration, and topial appliation of forskolin ould be an effetive stimulator of epidermal adenylate ylase. In the explant ulture system all agents that inrease AMP and ativate AMP-dependent protein kinase also ause inhibition of mitoti ativity in t he G 2 phase of the ell yle [24,31-33]. This pharmaologi effet was also shown by forskolin in the 4-h assay used. Also, onsistent with its biohemial reversal was t he reovery of the ultures after 48 h. Thus, further studies with forskolin to ounterat hyperplasti skin diseases suh as psoriasis may be justifiable. RFRNCS 1. Bhat SV, Bajwa BS, Dornauer H, desouza NJ, Fehlhaber HW: Struture and stereo hemistry of new labdane-diterpenoids from Coleu.s [orskohlii. Tetrahedron Letters 19:1669-1972, 1977 2. Metzger H, Lindner : Forskolin- a novel adenylate ylase ativator. IRCS Med Si Libr Compend 9:99, 1981 3. Seamon KB, Padgett W, Daly JW: Forskolin: unique diterpene ativator of adenylate ylase in membranes and in intat ells. Pro Natl Aad Si USA 78:3363-3367, 1981 4. Seamon KB, Daly JW: Ativation of adenylate ylase by the diterpene fo rskolin does not require the guanine nuleotide regulatory protein. J Biol Chern 256:9799-981, 1981 5. Darner FJ, Mahan LC, Koahman AM, lnsel PA: Stimulation by forskoli n of intat S45 lymphoma ells involves t he nuleotide regulatory protein of adenylate ylase. J Biol Chern 257:1191-1197, 1982 6. Siegl AM, Daly JW, Smith JB: Inhibition of aggregation and stimulation of yli AMP ge neration in intat human platelets by the diterpene forskolin. Mol P harmaal 21:68-687, 1982 7. Litosh I, Hudson TH, Mills 1, Shih-Ying L, Fain JN: Forskolin as an ativator of yli AMP aumulation and lipolysis in rat adipoytes. Mol P harmaal 22:19-115, 1982 8. Insel PA, Stengel D, Ferry N, Hanoune J : Regulation of adenylate ylase of human platelet membranes by fo rskolin. J Bioi Chern 257:7485-749, 1982 9. Seamon KB, Daly JW: Forskolin: a unique diterpene ativator of yli AMP-generating system. J Cyli Nuleotide Res 1:21-224, 1981 1. Daly JW, Padgett W, Seamon KB: At ivat ion of yli AMPge nerating systems in brain membranes and slies by the dieterpene forsko lin: augmentation of reeptor-mediated responses. J Neurohem 38:532-544, 1982 11. Yoshikawa K, Adahi K, Halprin KM, Levine V: The effet of ateholamine and related ompounds on the adenylate ylase system in the epidermis. Br J Dermatol 93:29-36, 1975 12. Iizuka H, Adahi K, Halprin KM, Levine V: Histamine (H 2) reeptor-adenylate ylase system in pig skin (epidermis). Biohim Biophys Ata 437:15-157, 1976 13. Adahi K, Yoshikawa K, Halprin KM, Levine V: Prostaglandins and AMP in epidermis. Br J Dermatol 92:381-388, 1975 14. Iizuka H, Adahi K, Halprin KM, Levine V: Adenosine and adenine Vol. 81, No.3 nuleotide stimulation of skin (epidermal) adenylate ylase. Biohim Biophys Ata 444:685-693, 1976 15. Takeda J, Adahi K, Halprin KM, Levine V, Woodyard C: Adenylate ylase ativation by holera toxin in pig epidermis: an obligatory role of the GTP-regulatory protein. J Invest Dermatol 81:131-136, 1983 16. Steiner AL, Parker CW, Kipnis DM: Radioimmunoassay for yli nuleotide. I. Preparation of antibodies and iodinated yli nuleotides. J Biol Chern 247:116-1113, 1972 17. Adahi K, lizuka H, Halprin KM, Levine V: pidermal AMP is not dereased in psoriasis lesions. J Invest Dermatol 74:74-76, 198 18. Salomon Y: Adenylate ylase assay. Adv Cyli Nuleotide Res 1:35-55, 1979 19. Lowry OH, Rosebrogh NJ, Farr AL, Randall RJ: P rotein measurement with the Folin phenol reagent. J Biol Chern 193:265-275, 1951 2. Adahi K, Levine V, Halprin KM, Iizuka H, Yoshikawa K: Multiple forms of yli nuleotide phosphodiesterase in pig epidermis. Biohim Biophys Ata 429:498-57, 1976 21. Halprin KM, Lueder M, Fusenig N: Growth and differentiation of postembryoni mouse epidermal ells in explant ultures. J Invest Dermatol 72:88-98, 1979 22. Hammar H, Halprin KM: On the in vitro growth of epidermis, The pidermis in Disease. dited by R Marks, Christophers, La naster, ngland, M.T.P. P ress, 1981, pp 243-271 23. T aylor JR, Halprin KM, Levine V, Woodyard C: ffets of methotrexate in vitro on epidermal ell proliferation. Br J Dermatol 18:45-61, 1983 24. Halprin KM, T aylor JR, Levine V, Woodyard C, Adahi K, Comerford M: Age nts whih ativate yli AMP-dependent protein kinase inhibit explant ulture growth and mitoti ativity. J Invest Dermatol, in press 25. Rodbell M: The role of hormone reeptors and GTP regulatory proteins in membrane t ransdution. Nature 284:17-21, 198 26. Pfeuffer T : GTP-binding proteins in membranes and t he ontrol of adenylate ylase ativity. J Biol Chern 252:7224-7234, 1977 27. Limbird L, Gill DM, Lefkowitz RJ: Agonist promoted oupling of the {J-adrenergi reeptor wi th the guanine nuleotide regulatory protein of t he adenylate ylase system. Pro Nat l Aad Si USA 77:775-779, 198 28. Lad PM, Welton AF, Rodbell M: videne for distint guanine nuleotide sites in the regulation of the gluagon reeptor and of adenylate ylase ativity. J Bioi Chern ~52:5942-5946, 1977 29. Iyengar R, Abramowitz J, Bordelon-Riser M, Blume AJ, Birnbaumer L: Regulation of hormone-reeptor oupling to adenylyl ylase. J Biol Chern 255:1312-1321, 198 3. Iyengar R, Bhat MK, Riser M, Birnbaumer L: Reeptor-speifi desensitization of the S49 lymphoma ell adenylyl ylase. J Biol Chern 256:481-4815, 1981 31. Ha rper RA, Flaxman BA, Chopra DP: ffet of pharmaologial agents on human keratinoyte mitosis in vitro. I. Inhibition by adenine nuleotides. Pro So xp Bioi Med 146:132-136, 1974 32. Flaxman BA, Harper RA: In vitro analysis of t he ont rol of keratinoyte prolife ration in human epidermis by physiologi and pharmaologi agents. J Invest Dermatol 65:52-59, 1975 33. Harper RA, Flaxma n BA: ffet of pharmaologi agents on human keratinoyte mitosis in vitro. III. Inhibition by histamine and methylated analogs. J Invest Dermatol 65:4-43, 1975