Identification of Functional Platelet-Activating Factor Receptors on Human Keratinocytes

Jeffrey B. Travers, J. Clark Huff, Marek Rola-Pleszczynski,t Erwin W. Gelfand,* Joseph G . Morelli, and Robert C. Murphy* Department of Dermatology, University of Colorado Health Sciences Center, and ' Division of Basic Sciences, Dcpartmcnt of

Pediatrics, National J ewish Center for Immunology and Respiratory Medicine, Denver, Colorado , U .S.A.; and ,! Immuno logy Division, Department of Pediatrics, Universite de Sherbrooke, Shcrbrooke, Qucbcc, Canada

Platelet-activating factor (P AF) is a potent inflamma­tory Inediator that has been shown to be produced by human keratinocytes and is thought to playa role in cutaneous inflammation. Immunofluorescence and radioligand binding studies were used to characterize P AF receptors (P AF-R) on human keratinocytes and the human epidermoid cell lines A-431 and HaCaT. Indirect immunofluorescence studies demonstrated anti-P AF-R staining of primary cultures of human keratinocytes, A-431 cells, and HaCaT cells. Primary cultures of human fibroblasts and the melanoma cell line SK-30 failed to show immunostaining above that seen with control antiserum. With indirect immuno­fluorescence studies of sections of normal human skin, a granular anti-P AF-R staining pattern was noted on the keratinocyte cell membranes. A-431 cells readily metabolized P AF by deacetylation­reacylation at 37°C, but not at 4°C. Binding studies on crude membrane preparations of A-431 cells con­ducted at 4°C demonstrated specific binding that

P latelct-activating facto r (1-alkyl-2-acetyl-glycero-3-phosphocholine; PA F) is a potent activator of many cell types including plate lets, vascular endothelium , n eutro­phi ls, mast cells, and monocytes [1, 2]. By acting as both a chemoattractant to recruit and as a stimulus to activate

granulocytic cells, PAF h as profound proinflammatory effccts . In addition to the proinflammatory effects found with the stimu lation of the above cell types, PAF h as mitogenic efFccts on cultured smooth muscle and lymphoblastic cell lin es [3-5] , The majority of PAl" actions are thought to be exerted through inte raction with a specifi c receptor, because radioligand binding studies usin g [3H]PAF have demonstrated high-affinity binding sites o n PAF­responsive cells, and various structurally dissimilar PAF receptor antagonists can inhibit PAl" effects [6]. The PAF receptor (PAF-R)

Manuscript received December 28, 1994; final revision received June 12, 1995; accepted for publication July 16, 1995 .

Presented in part in abstract form at The Society fo r In vestigative Dermatology m eeti ng, Ap ril 1994,

R eprint requcsts to: Dr. Jeffrey B. Travers, Departments of Dermatology and Pharmacology, Indiana Univers ity School of Medicine, 550 N o rth University Boulevard. Sui te 3240. Indianapo li s. IN 46202 .

Abbreviations: GPC, glycerophosphocholine; KLH. keyhole limpet he­mocyanin; PAF, platelet-activating factor; PAF-R, pAF recepto r.

reached saturation by 120 min. Scatchard analysis of PAF binding data revealed a single class of high­affinity (Ko = 6.3 ± 0.3 nM) PAF binding sites. The immunofluorescence and radioligand binding sites were shown to be functional PAF-Rs, as 10 pM to 1 #-tM P AF increased intracellular calciuln in primary cultures of human keratinocytes, A-431 cells, and HaCaT cells, whereas PAF treatment of primary cul­tures of human fibroblasts or the m,elanoma cell line SK-30 did not result in changes in the intracellular calcium concentration. The structurally dissimilar PAF-R antagonists CV-6209, Ro19-3704, and alpra­zolam all inhibited the PAF-induced calcium changes in A-431 cells. The CV -6209 inhibition was seen at doses that competed with the P AF binding to these cells. These studies provide the first evidence for the pres­ence of a functional P AF-R expressed on human kera­tinocytes, suggesting d"lat tIns lipid mediator may play an important role in normal keratinocytes or in inflam­matory dennatology.] [,,,,est DemratoI105:816-823, 1995

has been cloned fi'om guinea pig lung [7] and hum an leukocytes [8]. The elucidated structure of the PAF-R places it in the G-protein­coupled rhodopsin class receptor r.,mi ly. with seven putative trans­membrane domains .

The PAF-R is functionally lin ked to phosphoinosito l hydrolysis, and activatio n resu lts in the production of inosito l-l ,4,5-trisphos­phate and diacylglycerol [9]. lnositol-l ,4,5-trisphosp hate mobilizes intracellular calcium, and diacylg lycero l activates prote in kinase C . PAF also affects cyclic nucl eotides by inhibi ting cyclic adenosine m onophosphate accumulation [1 0]. PA F activates microtubule­associated protein-2-kinase [11] and induces express ion of the earl y response gene c:los [12).

Several lines of evidence suggest that PAF m ay playa role in epidermal pathophysio logy. First, PAF has been found in psoriatic scale [13] and in blister fluid from both traumatic [14) and buUous pemphigoid [15] blisters. Primary cultures of human keratinocytes have been shown to synthesize PAl" when sti mulated with the calcium iOllophore A23187 [16] . Second, intradermal injections of PAF resul t in wheal and flare reactions [17] . Third, PA F-R antagonists administered either systemically o r topicall y have anti­inflammatory effects in murine model system s of contact dermatitis [1 8,19].

Although there is much ev idence suggesting a role for PAF in

0022-202X/95/S09 .50 • SSD I0022-202X(95)00373-8 • Copyright 1995 by T hc Society for Investigative Dermatology. Inc.

816

VOL. 105. NO. 6 DECEM13ER 1995

epidermal fun ction, a keratinocyte PAF-R h as not yet bee n dem­onstrated. In tills study, we uscd immuno fluorescen ce and radioli­gand binding studies to ide ntify PAF-R immunoreacti v ity and sp ecific binding sites in human ke ratinocytes. C h arac te rization of P AF-induced changes in k e ratinocyte intra cellular free calcium u sing the calcium-sensitive fluorescent dye indo-l indicated that these binding sites w e re fun ctional receptors. T h e demonstration of a fun c tional k e ra tinocyte PAF-R su ggests that PAF may have direc t e ffects on keratinocyte function.

MATEIUALS AND METHODS

Reagents Labeled pAF (1-[' H]octadecyl-2-acetyl-glycerophosphocho­line [GpC]. 161 C i/ mmol) was obtained fi'om Amcrsham Corp. (Arlington Heig hts, IL) . Unlabeled PAF (1-hexadecyl-2-(R)acetyl-G PC) and other phosph olipids used for chromatograph y were purchased fro m Biomol R esear ch Labs (Plymouth, pA). Lyso PAF (1-hexadecyl-2-(R)lyso-GPC). S-PAF (3-hexadccyl-2-(S)ace tyl-1-GpC), f.,tty acid-frec bovinc serum al­bunlln (BSA). l-lEpES, phenanthroline. phcnylmcthylsulfonyl Ruoride , leu­peptin. and aprotinin A were plll'chascd rrom Sigma C hcmica l Co. (St. Louis, MO) . C V-620 9 was a kind gift of Dr. Hajimc Toguchi (Takcda Chel11.ica l Industrics, Japan). Ro 19-3704 was kindly provided by Dr. Petcr Sorter (Hoffinan-La Roche, Inc., Nutlcy, NJ). Alprazolam was a genero us gift from the UpJohn Co. (Kalamazoo, MI).

Cell Culture Primary cultures of human ncollatal kcratinocytes and fibroblasts werc isolatcd from fo reskins as described prev iously [20]. All tissue culturc media and supplcments were obtaincd from Gibco (Grand Island , NY) . Human neona ta l keratinocytes were grown in keratinocyte­serun" free medium supplcmcntcd with antimicro bials (pcnicillin 10 ,000 U/rrtL, streptomyc in 10 mg/mL, fun gizonc 25 J..Lg/ mL) and werc used aftcr the second passage (days 18-22). Human nconatal foreskin fibrobla sts were gro""n in M-1 99 mcdia witll antimicrobia ls and wcre used within the first month of cultlll'e. T hc epidermoid cell line A-431 was obtaincd from Am e rican Type C ulture Collcction (Rockville , MD) , and HaCaT cells [21] wer e a kind gift from Professor Norbert Fusenig (German Cancer Research Center, Heidelberg, Germany) . The keratinocytc cell lines were grown in Dulbecco's modified Eagle's medium supplemcnted with 5 mM HEPES , L-gi u tamine, and antimicrob ials. For use in the mctabo li sm. binding, and calci um studies . adherent cell s were harvestcd while 80% to 90')1" conRuent by treatmcnt with 3 mM cthylenediaminete traacctic acid in phosphatc­bufFered sa linc (PBS) at 37°C for 30 min. with gcntl c sc raping. T he ccll s were w as hed at least twice with Hanks ' balanced sa lt solution (HBSS) wi thout phcno l rcd, which contained (in mM) : 138 NaC I, 5 KC I. 0.3 KH 2 PO." 0.3 Na2 HPO., 0.5 MgCl,-6H 2 0, 0.4 MgSO.,-7H 20, 4 NaHC03, 1.3 aC12, 5.6 glucose, and 25 HEPES at pH 7.4.

Measuremcnt of PAF-R Irnmunofluorcsccncc ImmunoRuorescencc studies to characterize the keratinocyte PAF-R used a rabbit polyclollal antibody directcd aga inst a decapeptidc cxpressed on the extracellular portion ofthc human PAF-R. T his antibody (anti-hPAF-P'.' ''·-I7') has been sho""n to be specific for the PAF-R by both Ruoresccnce-activated cell sorter ana lysis and Western blotting [22]. ImmunoRuorescence studies used eithe r anti-hPAF-R' ''''- ' 73 or a control antibody 6'om rabbits immunizcd agains t thc carricr keyhole limpct hemocyanin (anti-KLH) alonc, purified as described [22], diluted 1 :100 with PBS. Keratinocytes were cultured in Vitrogen 100 (C elt";'x , Santa C lara, CA) prctreated Lab-tek Chamber slides (N unc, Naperville, IL) . Subconfiuent culturcs wcre wash cd twice with cold PBS a nd incubated with either anti-hPAF- R",·,- ' 73 or anti-KLH for 2 h at 4°C. T he primary antibodies were then rcmovcd. and the slidcs were wash e d with co ld PBS. The cell s were tllCn incubated with a 1:1 00 dilution of affi nity-iso lated Ruorescein isothiocyanate (FITC)-Iabclcd goat F(ab ' )2 an ti-rabbit IgG (Biosource Int .. Carrillo. CA) for 2 h at 4°C . thcn washed twice with co ld PBS and dried brieRy befo re acetone fixation. Skin speci.n1ens were taken from extra tissue fi'om yo ung adults undcrgoing cosn~eti c procedurcs. The tissucs wcrc embcddcd in OCT (Mi lcs Inc .. El.kha rt, IN) , frozen in liquid n.itrogen . and storcd at - 70°C until used. Thc skin was sectioned at 5 J..Lm on a cryostat. SectiollS wcrc placcd on gela tin-coated slidcs and incubated with anti_hPAF_R" ,"- ' 7J o r anti-KU-I antibody for 2 h at room temperature, rinsed with PBS, thcn incubated with me FITC-Iabeled secondary antibody for '1 h at room tcmpcrature. Covcr­slips ""ere mounted and the slidcs werc examined on an O lym pus Ruores­cen ce microscope.

Measurement of ['HlPAF Mctabolism A-431 cc lls (2.5 X 10") wcre wash e d and resuspended to a final volume of 0.95 1111 of HBSS. To the above ce ll suspcnsion. wc added 50 J..L I of aqueo us solution containing ['H]PAF complexcd to BSA (2.5 mg/ ml). bringing thc final concentration to 1 X 10- 0 M . The incubation mixturc was shaken gcntl y at either 37°C

PAF R.ECEPTO R S O N KEKATINOCYTES 817

or 4°C for '1 20 min. The rcaction was tcrminated by addition of3.75 ml of chloro form / mcthanol (1:2 v /v ) to thc ce ll suspcnsion, and the lipids wcre cxtracted using the technique of Bligh and Dycr [23]. T he recovercd pro ducts \vere those present in bo th the cell s and 111CdiuI11 , and the recovery of tritium label in the lipid fraction was greater than 85%. Solvents were removcd by a stream of nitrogen, and the lipids were resuspend cd in chlorofonn/ mcthanol (1: 2 v/v). The phospholipids and phosphol ipid stan­dards werc scparated On sili ca gel G th in-Iaycr chromatograph y (TLC) platcs (A nal tech, Newark, DE) using an acid.ic so lvent sys tem containing chloroform/ methano l/glacia l acctic acid / watcr (65 :35:8: 1, v/v). Thc dis­tribution of radiolabel was determined using an automatic TLC Iincar ana lyzer (Bcrtho ld. Germany).

Measurement of ['HlPAF Binding SubconRuent cultures of A-431 ce lls (150-200 X 10" ce ll s) were washed and resuspended to a final volumc of 20 ml with PBS containing the protcasc inhibitors phenanthro linc (I mM) , phcnylmethylsulfonyl Ruo ridc (100 J..LM). aprotin.in A (10 J..Lg / mL). and leupeptin (10 J..Lg/ mL). The ce llular suspension was sonicated on icc. thcn centrifuged at 100 X g to rcmovc whole cells. Thc supcrnatant was rcmovcd and centrifuged at 100 .000 X S for 60 min at 4°C. The rcsulting pellct was resuspended in binding bufFc r. which contained (in mM): 10 NaC I, 5 MgC I2-6H20, 5 MgSO.,-7H2 0 . 10 CaCI2• 5. 6 glucose, and 10 Tris-HCI at pH 7.0. Aliquots of crudc mcmbrane (50 J..Lg protein ; measurcd by protein assay [24]) wcre placed in 15 X 85-mm glass tubes prctrca tcd with Sigmacote (Sigma), with a fi nal volumc of 0.4 ml. Suspensions were thcn simultaneously exposed to 50 J..LI of aqucous solution con ta ining [' H]PA F complexed to BSA (10 mg/ml ), plus eithcr 50 J..Ll unlabe lcd PA F. lyso PAF, o r C V -6209 complcxed to BSA, or BSA alone. T he mixture of crudc membranes, [' H]PAF, and inhibitors was shaken gently at 4°C for thc appropdatc ti111CS . Each sanlpic \vas rUIl in tripli ca te. Suspensio ns \verc harves ted by suction through GF/C filters (Whatman) premoistcncd with 10 mg/ ml BSA using a Hocffcr Filtration apparatus (Hocffer Scientifi c. San Francisco, CA). Thc tubcs were thcn was hcd threc times witll 5 ml of cold binding buffer. The filters were air dri ed and placcd in 10-ml scintillation vials, to which 7 ml of ScintiveL'sc (Fisher, Fairlawn, NJ) was added. The vials were coun tcd in a Beckman LS6800 scintillation CO untcr that was programmcd to measure each sam pIc 's quench and to determine dpm fi'om cpm using tritium standards.

Intracellular Calcium Measuremcnts Adhercnt cclls werc harvcstcd while 80% to 90% conRuent by treatment with 3 mM ethylenediaminc­tetraacetic acid in PBS at 37°C for 30 min. with gcntl c scraping. T he ce lls were washcd at Icast twice with HBSS witho ut phenol red . Cells (0.5-'1.0 X 10"/ml) werc thcn loaded with thc Ca + + -sensitivc indicator, indo- I AM ('I J..LM; Calbiochem, San Dicgo, CA), in HBSS at 37"C for 45 min. After loading. thc cells were spun down , resuspcnded in HBSS . and maintaincd at room temperature. Beforc each mcasurement, an aliquot of the cell suspension (1.0-1.5 X 106

) was pclleted rapidly by centrifugation and rcsuspended in 3 ml ofHBSS. PAF and PAF-R antagonists were addcd in 2.5- 5 J..LI cthanol. Endothelin-l was added in PBS buffer. Indo-I Ruorcscence was mOl~itored in a Hitachi F-4010 spectrophotometer. Mea­surcments were performcd at 37°C witll constan t stirring. T he cxcitation and cmission wavclengths of the Ruorescent dye w cre 331 and 410 nm. respectivcly, and [Ca++]; was calculatcd as described [25].

RESULTS

PAF-R Expression at the Cell Surface of Cultured Kerati­nocytes and Skin Slices To charac tc l;ze the expression of PAF-R at the prote in level, we used a rabbit polyclon a l antibody against a decap c ptide (hPAF_R' 64- ' 7, ), corresponding to a segm ent in the second extracellular loop of the putative struc ture of the human PAF-R., in immunofluorescen ce studies as d escribe d in Materials alld I'dethods. The anti-hPAF-R 1(,.1- 173 antibody bound to the surface of c ultured hum an keratinocytes and to thc keratinoC)'te ce ll lines A-431 and HaC aT, as sh own in Fig lA-D. The granu lal' p a ttern of fluores cen ce oudin ing the cell s w as consiste nt with that of a membranc receptor. Indirect immunofluo rcscen ce with control antibody (anti-K.LH) did not produce' similar staining (Fig iF). Inc ubation of anti- hPAF-R' 64 - , 73 antibody with primary c ultures

of human fibroblasts (Fig lE) or the m e lanoma cell tinc SK-30 (data not shown) did not demonstra te diffcren ces in fluorescen ce above the background found with anti-KLH. In c ubation of slcin slices with anti-PAF-RI(>·' - ' 73 antibody resulted in selective bind­ing to the epidermis (Fig 2) in a granular patte rn on keratinocyte

m e mbranes . most intense on the supra basal cell s .

818 TRAVERS ET AL T HE JOU RNAL OF INVESTIGATIVE DERMATOLOGY

Figure 1. Indirect immunofluorescence labe ling of human keratinocytes. C ultured keratinocytes were incubated with either anti_hPAF_R' 64- ' 73 o r anti-KLH, washed, and incubated w ith FITC-conjugated goat anti-rabbit IgG, as described in Mnlerin /s nm/ Melfl"ds. The anti_ hPAF_R' 64- ' 73 antibody gave granular staining o f the cell m embranes on (A) human neo natal foreslcin keratinocytes (600 X), (8) A-43 1 cell s (400 X). (C) HaCaT ce lls (600 X), and (D) H aCaT ce ll s (200 X), but did not stain (E) human neonatal foreslci n fibrob lasts (400 X). T he an ti-KLH control antibody fililed to stain primary cultures of human neonatal keratinocytes, A-431 cell s, HaCaT cell s, or primary cultures of human neonatal fo reskin fibroblasts. F. an example of the anti-KLH anti body sta ining in A-431 cell s (400 X).

Metabolism ofPAF by A-431 Cells Because it was not known whether keratinocytes metabolize PAF, [3H]PAF was incubated w ith A-431 cell s and the products were examill ed by TLC, as described in Materials Il lld Methods. Figure 3 shows the distribution of label in PAF, lyso PAF, l - alkyl-2-acyl-GPC, and neutral lipids after incubation with A-43 1 ceUs for 2 h at 37°C and 4°C. A-431 ceUs readily metabolized PAF at 37°C , with l-alkyl-2-acyl-GPC the major produ ct, indicating that these cells metabolize PAF by the deacetyla tion-reacylation pathway found in other tissues [26,27].

Tlus metabolism was temperature dependent, with milumal PAF degradation at 4°C. The products shown in Fig 3 were extracted from the complete incubation mixture and thus represent products derived from both cells and m edium. Incubation of CH]PAF under standard conditions at 37°C in the absence of cell s resul ted in negligible me tabolism of [3H]PAF (resul ts not shown).

Binding ofPAF to A-431 Membranes Because ligand metab­oli sm can result in spurious binding resul ts, PAF binding studies

VOL. lOS, NO. 6 DEC EMBER 1995

were conducted at 4°C . The kinetics of PAF binding to A- 431 membranes were examined. C rude membrane preparations of A-431 cells (approximately 50 JoLg protein) were incubated in triplicate with 1 nM of [3H]PAF, with or without 1 JoLM of unlabeled PAF, and the binding was q uantified as described in Materials alld Methods. As indicated in Fig 4, specific binding reached saturation by 120 min. T he affi ni ty and number of PAF binding sites in A-431 cells were determined by constructing saturation binding isotherms. As demonstra ted in Fig 5, PAF specific binding increased with the concentration of labeled ligand and reached saturatio n. The data were then subjected to Scatchard analysis [28] to quantitate the number of binding sites present and their apparent affinity. Scatchard analysis of four separate experi­ments gave a linear plot (representative experiment shown in Fig 5), indicatin g a single population of PAF binding sites. The equilibrium dissociation constant (KD ), representing the affin ity of the PAF binding site, was calculated to be 6.3 ~ 0.3 nM . The Bmax ,

representing the total number of specific binding sites present, was calcula ted to be 1.0 ~ 0.1 pmollmg protein . Assuming an equinlo­lar ligand-receptor complex and 100 X 10(' ce ll s yielding 2.1 m g of

PAF REC EPTORS ON KERATINOC YTES 819

Figure 2. Indirect itnmullofluorescence l a ­beling of sections of normal human skin. Five-micron-thick cryostat sections were incubated with either anti_hPAF_R" ,4- 173 o r anti- KLH , washed , and incubated with FITC-conjugated goat anti-rabbit IgG as desc ribed in Materials alld IIIlct"­ods. Granular staining of the keratinocyte ce ll m em­branes was seen w;th the anti- hPA F-R 164 -1 73 an­tibody (A) , but not with the anti- KLH antibody (B) (400 X).

crude m embrane protein under o ur experimental conditions, the above B",," x corresponds to 120 ~ 12 binding sites per A- 431 cell . The specificity of PAF binding was examined by conducting competition binding studies with unlabeled PAF, the PAF-R antagonist CV-6209, and lysoPAF (Fig 6) . Both unlabeled PAF and CV- 6209 (50% inhibitory concentration = 4.9 and 10.5 nM , respectively) competed with eH]PAF for binding to A-431 mem­branes, whereas lysoPAF was inactive.

PAF-Induced Intracellular Calcium Mobilization To de­termine whether the binding sites on human keratinocytes were functionally ac tive, we measured changes in rCa + +1 induced by PAF using the fluorescent probe indo-·1. PAF caused a transient increase in intracellular free calcium in a dose-dependen t m anner. As illustrated in Fig 7 , [Ca++ l began to increase 5 to 10 seconds after the addition of P AF and returned to baseline within several minu tes. Repeat stimulation with PAF resulted in little change in rCa ++ ]; , suggesting homologous desensitization (results not shown). Dose-response curves in Fig 8 charactet;ze the relation between maximal changes in rCa + +1 and PAF concentration for

820 TRAVERS ET AL

o

" " " " " " II: i III , , , ,

5 10 15 Migration Distance (cm)

20

F igure 3. Distribution of labe l in products after incubation of 1- ['H]octadecyl- 2- acetyl-GPC with A - 431 cells. T he cell s were incu­bated for 120 min at 4°C (das" ed lill e) and 37°C (solid lille) , and the lipids were extracted as described in Malerials alld Metll"ds . An aliquot of the labeled products was separated on TLC and radioassayed. T he peak numbers identifY the following lipids: 1-[3H]octndecyl- 2-lyso-GPC (peak J, RF = 0.26) , 1-[' H]octadecyl-2-acetyl-GPC (peak II, RF = 0.3 1), l -['H]octa­decyl-2-acetyl-GPC (peak 111 , IUO = 0.51), unkJlOwn component (peak lV, RF = 0.77), and neutral lipid comignlting with 1- [3H]octadecyl-2-acctyl glycerol (peak V).

primary cultures of human keratinocytes and A-431 cells. PAF treatment of indo- I-loaded HaCaT cells resulted in calcium re­sponses similar to those in A-431 cells (data not shown) . T he change in [Ca ++]; was evident at a PAF concentration as low as 10 pM. T he m edian effective concentration for the Cllrves generated were 780 pM and 1.5 nM for primary cultures of human keratino­cytcs and A-431 cells, respectively. S-PAF, the biol ogically less active enantiomer of PAF, was a much less potent stimulus for A-43I calcium m obilization than was nati ve PAF. T he structurally dissimilar PAF-R antagonists CV -6209, Ro19-3704, and alprazo­lam all inhibited PAF-induced calcium flu x in a dose-dependent

,..... c: . iii ..., 0

OlL c: Q

'6 Ol c: ::1 iii .........

E Q

:::,

200

150

100

50 .. '

0 0

"

60

.. ' f-----------· ....... iIi~------- !

120 Minutes

180

F igure 4 , Kinetics of ['H]PAF binding to A-431 cell membranes at 4°C, C rude membrane prote in (50 p.g) was incubated w ith 1 nM [3H )PAF for the indicated time and harvested as desc ribed in Malerials ami Meillods. Specific binding (closetl circles) was the difl:c rence between PAF binding in the absence (oPell s1"ares) and presence (closed s1""res) of excess unlabeled PAF (1 p.M). Each point is the mean :!: SEM PAF bind ing of two separate experiments using triplicate samples.

THE JOURNA L OF IN VESTIGATIVE DERMATOLOGY

Figure 5 . Saturation kinetics and Scatchard analysis of [JH]PAF binding to A-431 membranes. Saturation binding iso therms were con­structed (illset) by equilibrium (120 min) binding of 0.05 to 10 nM [3H]PAF to A-431 membranes. The data were subjected to Scatchard anal ysis. T he da ta pictured are from a single experiment using triplicate samples and is represen tative of four separate experiments. T he binding affin ity (Ko) and maxim um binding (B",,,xl were ca lcu lated from the Scatchard plot and were found to be 6.3 :!: 0.3 nM and 1.0 :!: 0.1 pmol/mg protein, respectively (m ean :!: SEM £i'om fo ur separate experiments).

maImer, as illustra ted in Fig 9. This inhibition of the PAF-induced increase in [Ca + +]; by these PAF-R antagon.ists was specific for PAF, because pretreatment of A-431 cells or primary cultures of human keratinocytes with 10 nM CV- 6209, 1 IJ.M Ro19-3704, or 100 IJ.M alprazolam did not affect 10 nM endothelin-l -induced calcium mobilization (Fig 7) . The concentrations of PAF, S-PAF, or PAF-R antagonists did not affect cellular integrity as measured by trypan blue dye exclusion . PAF (100 nM) treatment of indo-l­loaded primary cultures of fibroblasts, primary cultures of human m elanocytes, and the m elanoma cell line SK-30 did not affect the baseline fluorescence (data not shown) .

DISCUSSION

These studies demonstrate that human keratinocytes express func­tional PAF-Rs. Indirect immllnoflu orescence studies using a poly-

Ol 4- c: 0 '5 c: c: 0 iii ..., ii . ~ E ~ c: u ~ Q)

Q

~ (f)

125

100

75

50

25

1,

I 1 I I

.L

.. _---- -. o~_..~--~---..~--~--~~--~--~~

10- 11 10-10 10-9 10-8 10- 7 10-6 10- 5

[Inhibitor] eM)

Figure 6. Compctition of[JH]PAF binding to A-431 mcmbranes b y PAF and CV- 6209 . A-43 1 membranes were incubated w ith 1 nM ['H]PAF and difFerent conccntrations of unlabeled l - hcxadecyl-2-acetyl­GPC (circles), CV-6209 (s1"ares), lyso PAF (tri""gles) , or J3SA alone fo r 120 min and harvested as described in Mnterials alld Melhads. The data are expressed as the m ean :!: SEM percentage of specific binding observed at each concentration of analogue from three or four separate experiments.

VOL. 105, NO.6 DECEMD ER 1995

4 00 A

r--. 350 ::E c ...........,.

,---0

+ 3 00 P C\J ~

~ U L---J

25 0

200~~~~~~~~~~'-~~~

o 1 2 3 4 Minutes

4 00 B

r--. ::E

350 E c C p ...........,.

1 ,---0 1 1 + 300 C\J ~

U L---J

250

200+-~~~~-.~~~~-,----

o 5 10 Minutes

Figure 7. PAF-induced intracellular Ca ++ mobilization in human neonatal foreskin keratinocytes. Indo- I-loaded keratinocytes were incuba ted in HBSS containing 1.4 mM CaCl2 and challenged with (A) 1 nM PAF or (B) 10 nM C V-6209 before PAP treatment. As shown (B), pretreatment of ce ll s with CV-6209 did not affect the calcium mobilization indu ced by 10 nM cndothclin-l. These arc typical resu lts fi·om at least four o r five separate cxpcri lll cnts.

clo n a l anti-peptid e antibody (anti_hPAF_RI 64 - 173), w hich h as been shown to be specific for the PAF-R in m o nocytic, neutro­philic, and B-Iymph o blastic cell lin es [22], showed binding of anti- hPAF-R 1 (,·1- 173 an tibody to primary cultures of human kera­tinocytes and the keratinocyte cell lines A-431 and Ha CaT. Mic­rotome sections of normal human skin also demonstrated a granu­lar, cell ular membrane pattem of anti_hPAF I6'1- 173 antibody binding localized to the epidermis (Fig 2). Anti_hPAF1(,.1- 173 immunoreactiv­ity was slightly more prominent in the squamous and granular layers, with l ess in the basal layer. Anti_hPAF_R 1(,.1- 173 staining spared the horny layer (stratum com eum), suggesting that cellular viability is necessary for PAF-R expression .

PAF RECEPTORS ON KERATINOCYTES 821

250

,..... ::E:

200 c

,-'-'; A + (\J 150 ro u '--'

c 100

ClJ OJ c ro

L 50 u

• 0

10-13 10-12 10-11 10-10 10-9 10- 8 10- 7 10- 6

[PAF] (M)

,..... ::E:

1000 c

'--'

,--, B + (\J

ro U '--'

c 500 ClJ OJ c ro

" L U

j

0 10 - 1310-1210-1110-10 10- 9 10- 8 10- 7 10- 6 10- 5

[PAF Agonist] (M)

Figure 8. Concentration dependence of intracellular Ca++ mobi­lization. Indo-I-loaded (A) human neonata l foreskin keratinocytes or (B) A-431 ce lls were treated with 100 I'M to 1 fLM PAF (circlrs) or 1 nM to 5 fLM S-PAF (squares). and the peak change in [Ca ++); was calculated. T he values arc mean :t SEM frol11 three to eight separate experiments.

Because of the large number of cells n eeded and the advantages inherent in using a homogeneous source of cells instead of primary cultures (which can contain small amounts of oth er nonkeratino­cyte skin cells), w e u sed the cell li n e A- 431 in both the m etabolism ;lI1d binding experiments. A-431 cells m etabo lized PAF at 3 rc by transacylation , similar to that found in other cells such as n eutro­phils [26], lymphocytes [27], and plate lets [29] . To avoid binding artifacts due to ligand m etabolism , we conducted binding assays on crude membran e preparations of A-431 cells at 4°C, a temperature at which th ese cells do n ot appreciably m etabolize PAF. [3H]PAF specific binding in creased over time and reach ed equilibrium by 120 min. Scatchard analysis of saturation binding isotherms re­vea led a sing le class of high-affinity binding sites, which had a KD of 6.3 ± 0.3 11M (Fig 5) . This Ko found for A- 43 1 cells is similar to reported va lues for hum an platele ts [30] and Raji lymphoblasts [31], and is somewhat highe r than repo rted va lues for human neutrophils [32]. The number of bindin g sites was estim ated to b e 120 per cell. PAF and CV- 6209 , but n ot the biolog ically inactive PAF precursor lyso PAF, competed w ith CH]PAF for binding to A-431 m embranes . .

The anti_PAF_R 1(,4 - 173 immun oreactivity and r3H]PAF binding sites fo und on human keratinocytes were shown to be functional PAF-Rs because PAF treatment resulted in intracellular calcium mobilization . The m edian effec tive concentration of the PAF­indu ced calcium mobi lization was similar to the Ko derived from the binding studies . T h e pattern of PAF-induced increase in ke ratinocyte [Ca ++ ]; was similar to th at described in o the r cell

822

c Q ...,

0 '" L N ..., ii c

0 0 U :::E

"- E 0 .2 ;,.e .!d

'" u

TRAVERS ET AL

125

100 .. ... -------i_

75 f I\

'f 1

50 I \1

25 I •

O~_r--_,~--~--_r--~~--~--~L_--._-10- 11 10-10 10- 9 10- 8 10- 7 10-6 10-5 10- 4

[Inhibitor] (M)

Figure 9. Concentration dependence of the inhibition of PAF­induced intracellular calcium mobilization in A-431 cells by PAF-R antagonists CV-6209, Ro19-3704, and alprazolam. Indo-l­loaded A-431 cells were pretreated with various doses ofCV-6209 (circles), Ro19-3704 (sqllal'es). or alprazolam (triallgles) for 1 min. followed by 500 pM PAF. The peak change in [Ca++ J; was calculated. Each point is the mean :t SEM of the percentage of peak change of[Ca++]; induced by PAF alone from three to five separate experiments.

types [31]. The structurally dissimilar PAF-R antagonists CV -6209. R019-3704, and alprazolam all inhibited this PAF-induced re­sponse (Fig 9), yet had no effect on endothelin-1-induced intra­cellular calcium mobilization (Fig 7) . Primary cultures of human fibroblasts or the melanoma cell line SK-30 did not display anti_hPAF_RI 64-173 immunoreactivity, and PAF treatment of these cells did not result in intracellular calcium mobilization (results not shown).

In platelets, the PAF-R is functionally linked to inositol-specific phospholipase C, with resultant production of inositol-1,4,5 trisphosphate and diacylglycerol [9]. The intracellular calcium mobilization that is seen with PAF-R activation is due at least in part to inositol-1,4,5 trisphosphate-stimulated release ofintracellu­lar (nonmitochondrial) stores of calcium [33]. Our finding that PAF-R activation stimulates an increase in keratinocyte rCa + + 1. is consistent with a previous report that PAF treatment of cultured human keratinocytes (but not fibroblasts) resulted in phosphoino­sitol hydrolysis [34]. The neutral lipid diacylglycerol formed by phosphoinositol hydrolysis is known to activate protein kinase C by facilitating translocation of this enzyme from the cytosol to the extracellular membrane [35] .

We detected 1-alkyl neutral lipids when A-431 cells were incubated with PAF (Fig 3). The 1-alkyl-2-acetylglycerols, which are neutral lipid products ofPAF degradation by phospholipase C, also activate protein kinase C [3]. Alternatively, 1-alkyl-2-acyl­glycerols, which can also be produced by PAF degradation (transacylation followed by phospholipase C degradation), can inhibit protein kinase C activation [36]. These l-alkyl neutral lipids are thought to exert their effects independent of the PAF-R and thus would not be blocked by PAF-R antagonists. Inasmuch as PAF acring through its receptor and PAF degradation products could theoretically have either similar or opposite biochemical effects, demonstration of a functional keratinocyte PAF-R (inhibited by PAF-R antagonists) is an important first step in the future evalua­tion of PAF effects on this cell type .

The role of the PAF-R in kerarinocyte function is not clear. However, the finding by Pignol et al [37] that PAF potentiates interleukin-1 production by lipopolysaccharide-stimulated guinea pig kerarinocytes suggests that activation of the keratinocyte PAF-R may augment inflammation. PAF-R antagonists have been shown to have antiinflammatory effects in animal models of cuta­neous inflammation, including allergic and irritant contact derma­titis in the mouse [17,18] and burn edema formation in the rabbit

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

[38] . The availability of PAF-R antagonists may allow further characterization of tills potent autacoid in kerarinocyte pathophys­iology, information that could lead to better treatment strategies for inflammatory dermatoses.

We ackllolVledge the tecl",ical assistallce oj Ms. Tall/ara Zekll/all alld The New Lab [IIC. , Dellvel', CO. This stlldy ",as sllpp0l1ed ill pal't by Natiollal [llstitlltes oj Health grallfs AR074011 alld HL34303 .

REFERENCES

1. Braquet P, Touqui L. Shen TY. Vargaftig BB: Perspectives in platelet-activating f.,ctor research . Phnm,ncol Rev 39:97-145. 1987

2. Prescott SM. Zimmerman GA, Mcintyre TM: Plate let-activating facto r . J Bioi Cl,ell ' 265:17381-17384. 1990

3. Stoll LL, Figard PH. Yerram NR, Yorek MA. Spector AA: 1-0-alkyl-2-acetyl glycero l: a platelet-activating factor metabolite with biological activity in vascular smooth muscle cells. Cell Rec~IIll :13-25, 1989

4. LevesquejP. Hatzfeld A. Hatzfeldj: Fibrinogenmirogenic elfect on hemopoietic cell lines: control via receptor modulation. Proc Natl Acnd Sci USA 83:6494-6498 , 1986

5. Leprince C. Vivier E, Treton D, Galanaud p. Benveniste j, Richard Y, Thomas Y: Immul1orcgulatory functions of paf-acether. VI. Dual effect on human B cell proliferation . Lipids 26: 1204 -1208, 1991

6. Saunders R.N. Handley DA: Platelet-activating factor receptor antagonists. AIIII" ReI' Plrnm,ncol ToxicoI27:237-250. 1987

7. Honda Z I. Nakamura M, Miki I. Minami M. Watanabe T, Seyama Y. Okado H. Toh H. Ito K, Miyamoto T, Shimuzu T: C loning by functional expression of pla te let- activating factor receptor from guinea-pig lung. Nflttl re 349 :342-346. 1991

8. Nakamura M, Honda ZI, Izumi T, Sakanaka C. Mutoh H. Minami M. Bito H. Seyarna Y. M.ltsumoto T. Noma M . Shimuzu T: Molecular cloning and expression of platelet-activating factor receptors from human leukocytes.) Bioi C l, ell' 266:20400-20405, 1991

9. Shukla SO: Platelet-activating factor-stimulated fonnation of inositol-tris phos­phate in platelets and its regulation by various agents including calcium. indomethacin, CV3988 and forskolin. Arclr Bioclrell' Biopl,ys 240:674-679, 1985

10. Haslam Rj, Vanderwel MJ: Inhibition of platelet adenylate cyclase by 1-0-alkyl-2-0-acetyl-sn-glyceryl-3-phosphorylcholine (platelet-activating facror) . ) BioI Cl,elll 257:6879 - 6884,1982

11. Mazer B, Domenico J, Sawami H, Gelfand EW: Platelet-activating factor induces 3n increase in intracellular calcium and expression of regulatory genes in human B-Iymphoblastoid cell lines.) 1111111''''01 14 6:1 914 -1 919.1991

12. Tripathi YB, KandalajC. Gu ntaka R V. Lim R W, Shukla SD: Platelet-activating factor induces expression of early response genes c-fos and TIS in human epidermoid carcinoma A-43 1 cells. Life Sci 49:1761-1765,1991

13. Mallet AI, Cunningham FM: Structural identification of platelet-activating factor in psoriatic scale. Bioe/,elll Biopl,ys Res CO""""" 126:192-198. 1985

14. Barr RM, Lawlor F,judge MR, Courcney P, Barlow R. Kobza Black A, Mallet AI, Greaves MW: Platelet-activating facto r, lyso-platelet-activating factor and arachidonic acid release in normal human skin and the influence of topical steroid treatment. Br} Cli" PI,nm,"col 35:637-641, 1993

15. Okubo K. Karashima T, Hachisuka H. Sasai Y: The existence of a platelet­activating factor (paf-acether)-like substance in blister fluids derived from patients with bullous pemphigoid as demonstrated by human platelet aggre­gation teclmiques. K,m""e Med) 35 :1 73-179, 1989

16. Michel L, Dcnizot Y, Thomas Y, jean-Louis F, Heslan M, Benveniste j . Dubertret L: Production of paf-acether by human epidermal cells.) ["vesl Dom,nlol 95:576-581, 1990

17. Archer CB, Page CP, Morley j, MacDonald DM: Inflammatory characteristics of platelet-ac tivating factor (Paf-accther) in human skin. Brj Dcnllatol11 0 :45-50, 1984

18. Csato M, Csametzki BM: Elfect of BN52021, a platelet-activating factor antagonist. on experimental murine contact dermatitis. Brj DenuatoI118 :475-479, 1988

19. Kemeny L, Csato M, Braquet P, Dobozy A: Elfect of BN52021. a platelet­activating f.,ctor antagonist. on dithranol- induced inflammation. Br J Derlllatol 122:539-544. 1990

20. Yohn jj, Morelli jG, Walchak Sj, Rundell KB , Norris DA, Zamora MR: Cultured human keratinocytes synthesize and secrete endothelin-1 . ) II/vest Dem,atol 100:23-26, 1993

21. Boukamp P, Petrussevska R T. Breitkreutz D, Hornung j, Markham A, Fusenig NE: Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line .} Cell Bioi 106:761-778, 1988

22. Muller E, Dagenais P, Alami N, Rola-Pleszczynski M: Identification and functional characterization of platelet-activating f.,ctor receptors in human leukocyte populations using polyclonal anti-peptide antibody. Proc Nnll Acad Sci USA 90:5818-5822, 1993

23. Bligh EG. Dyer W]: A rapid method of total lipid extraction and purification. Cnll) Bioe/lell' PI,),siol 37:911-915. 1959

24. Bradford M: A rapid and sensitive method for the quantitation of microgram

VOL. 105. N O . (, DECEMUER 1995

q uall tities or prote in urili z ing the principle o f prote in- d ye binding. Alltli lJj"dll'lII 72:248-254. 1976

25. M ;u.; Do ug:dl SL. Gril1srci ll S. Gelfand EW: Dctct.:rion o f ligand-acrivalcd COIl-

duc ti ve calciul1l chan lie Is in human B IYl11phocyl"cs. CI'" 54:229-235. 1988 26. C hil ton FH. O'Flah erty J T. Elli s JM . Swendsen C L. Wykle R.L: Me tabo li c f:lte

o f plarclcr-acti v'ltin g f:lcro r in IlclItrophi ls. J fiilll C /u'lII 258:6357-6362. 1983 27. Tr:lvcrsJB. SprcL"hcr I-I. Fc rtcl ltJ-l: T h e IlIcw bo lisl1l of p!:lIclct-;J(.: tivaring f:lcto r

ill human T-IYlll phocytcs. Bhlr/Iilll lJjeJpilys ...-ierfl 1042: 193- 197 . 19CJO 28 . Scatchill'cI G: Th e artractio lls of protei ns fo r small m olecules ;!l ui ions. A IIII N )I

I i<"nd Sri S I:f,60-fJ(,·I. 1959 29. M"l o lic B. Lee T-C. Snydcr F: Ina ctiva t io n of platclc [-a cri vatin~ t:lclOr hy rahhit

platele t·s. ) /Jiol CII/'II, 260: I 53 1- 153(,. 19H5 30. rnarre;1 P. GOlllez-Ca lllhro llero J. N ie to M . Sallc hez-Crespo M : C haracte ristics

o f the bind ing of pl:ll c ler-a ctivating f;l cwr to pJ:ltcle ts o f different animal species. /ill)" ) 1'IIIII"/II nl"<ll IO;; :3()9-3 15. 1984

31. Travers JU . Li Q. Kniss DA. Ferrcl IU-{: Id c nti ti coltio ll o f fUll ctio nal platclct-;Jcri varing f:lClOr rcccptors in Il:tii lympho hlasts.) 1111111/11 /0' 143:3708 - 37 1.1. 1989

32. O' Flaherty JT. Surles JR. Iledlllall J. J acobsell D. I'iantadosi C. W ykle RL llinciin 1! and m c taho lism of pknclc t-.II': I"jvalin g: f.1 c fo r h y hU\I1 ;111 lIe uITophi ls . .J C lill /1I 1"'SI 7H:3H 1-38(,. 19H6

PAF R.ECEPTO llS ON KER.ATINOCYTES 823

33. llerr idgc MJ. Irv in g: RF: CYlosolic cakiullI o scillat ors.. FASEB J 2:3074 -30S3 . I 9HH

3-1. I=isltc r GJ . Ta lwar H S. ltydcr NS. Voorhccs J.I : Ditrcrentia l activatio n ofhulI1:1Il skin celh .. hy phHclc t-activatin g: (;lctor: stimulation of p hos.pho in ositidc Iurn­

over and arachidonic acid m ohilization in kcrat inoc), lcs bu r not fihrohln sts. IJiodlCl lI IJioplt )'s /{t'S C Ol/III11'1I 163: 13 .... - 1350. 1989

35 . N ishiz tlka Y: T he ro le of p I' ore ill kin asc ill cell surfa ce sign al tran sd uctio n a nd

tumo r promotion. N t/TIm: 308:693-697. 198-1 3r.. lIa ss DA. M cPhail LC . SchmittJ D. N orris-Natschke S. M cCall CEo W ykle R.L:

Selcerive primi ng ofrarc and durat io n of rile resp ira lory hurst of ne ll tfophi ls h y

1.2-cliacyl and I-O-alk yl- 2-acyl d iglyceridcs. Poss ihle ro le of prote in kin ase C . J ifi,,1 C holl 2(,3: 1% 10-1% 17. 19RH

37. Pignol B. LOII Ch a1l1pr MO. M Cllc ia-Hlierra JM . Uraqlle r P: Pl:uc le t-ani ,'atin g

f;lclor porc lltiates intcrie ukill - l / cp idcrt ll :1J cell de rived I'I lymocytc-acnvat in g fac tor release by g uinea-pig ker:t tinoc),fes stillH1latcd w ith Iipo po lys:1Ccllaride .

) Lipid JIII'''i", 2(S uppl):SH3-S9 1. 1990 3g. O no 1. C lIllj i H . J'hrad:1 H . K;1Jlcko F. Matsu zaki M : EtIcc[s of a platelet­

:u.: ti vating f.1C (Or an tagoni st o n oedcm a form ation foll owin g hUn1 s. BII"".~

19:202-207. 1993