British Journal of Dermatologij {1992) 126. 450-455.

Paraphenylenediamine, a contact allergen, induces oxidativestress and ICAM-1 expression in human keratinocytes

M.PICARDO. CLAUDIA :/:OMPETTA.* CINZIA MARCHESE.* CHIARA DI- LUCA,A.FAGGIONI,* R.J.SCHMIDTt AND B.SANTUCCISan Galtkano Dermatologicul ItisUuae and 'I'h'iHvtmt'nl of E.vperimental Medicine. Universitu ••'t Sapienza. Home,f Welsh School of Pharinacn. IJniwrsiiy ot Wates Colteitf of Cardiff. Cardiff, U,K.

Accepted for publication 26 November 199 1

Surnniciry In iin investigation ofthe role of keratinocytes in the pre-immunological phase of contact allergy, wehave studied the effect of piimphenylencdiiimine (PPD) on cell proliferation, memhrane lipidperoxidation and the expression of the intercellular adhesion molecule 1 (ICAM-1). Beeause PPDundergoes rapid autoxidation in the culture medium, the effect of PPD-modified medium onkeratinocyto proliferation and ICAM-l expression was also examined.

PPD at low concentrations (up to 10 /(g/ml) and with low exposure times (OS h) enhancedkeratinocyte proliferation, but at high concentrations and with longer exposure times resulted in cellstasis and toxicity. These effects and the enhanced membrane lipid peroxidation that was alsoobserved can be ascribed to the production of superoxide and hydrogen peroxide by the autoxidationof PPD in the medium. At non-cytotoxic concentrations, PPD induced ICAM-1 expression on tfiekeratinocytes. PPD-modified medium was also cytotoxic to the keratinoeytes and induced ICAM-1expression in non-cytotoxic concentrations. It appeared tfiat .superoxide and hydrogen peroxide werenot responsible for the cytotoxicity. These results are consistent with the view that oxidative stressmay be an essential part of the pre-immunological phase in the induction of allergic contactdermatitis.

Althougli mucb progress has been made in tbe under-standing of antigen presentation and the subsequentevents in the immune response of the skin, very fewstudies have been carried out on the early events in thepre-imnuinological phase of allergic contact dermatitis.Recent attention has been focused on the role ofkeratinocytes in the immunological reactivity of theskin.'-' and studies have demonstrated that keratino-cytes exhibit important biochemical changes whenexposed to low molecular weight allergens.-"^ Nickelions can induce the production of interieukin 1 andstimulate lipoxygenase activity.'' Urushiols and relatedmono- and dihydric alkylbenzenes may initiate redoxcycling and thus locally deplete the levels of endogenousreducing equivalents In the skin.'' The application ofpoison ivy/oak mixture in vivo and urushiol in vitro candirectly induce the expression of intercellular adhesionmolecule 1 (ICAM-I) on the keratinocytes in non-sensitized subjects.'

Paraphenylenediamine (PPD) is an aromatic amine

Correspondence: i)r Maum I*icardo, Islituto Dermatologico San Galli-cano. Via Siin tliillicano 25/A. 001 l i Rome, Il.ily.

that is frequently the cause of an occupationai contactdermatitis.*^ This compound does not exhibit protein-reactivity and hence is believed to undergo activationeither in or on the skin and that the hapten is eitherbenzoquinone'' or benzoquinone-imine.^ However, thesemiquinone-imine iree radical intermediate that arisesfrom metabolic activation in either the epidermis ordermis Ls the likely candidate.'' At present, it is notknown where or how activation of PPD occurs, or indeedany other prohapten. or on which cell or cells thehaptenization occurs. The aim of this study was toinvestigate the cflect of PPD on the viability, growth rateand expression of ICAM-1 on human keratinocytes inculture. This was to gain further understanding oftherole of keratinocytes in the induction of the immuneresponse in the skin.

Methods

Keratinocyte cultures

Human keratinocytes derived from neonatal foreskinswere plated on irradiated mouse fibrohlasts'' and cul-

4S0

PIM) INDUCES ICAM-1 EXPRESSION IN HUMAN KERATINOCYTES 451

tured in a keratinocyte growth medium (KGM; Clonetics,San Diego, CA, U.S.A.) with epidermal growth factor andhovinc pituitary gUind extract. The cells at the third orfourth pas.sage were used for these experiments.

Faraphenylenediamine and keratinocytes

The keratinocytes were seeded at 4-5 x 10'' cells/well in24-well dishes (Flow, Irvine, U.K.). Paraphenylenedia-mine hydrochioride IMerck AG. Darmstadt. Germany)wasdissolvedinphosphate-bufferedsaline(PBS)atpH 7 4and diluted In KGM. The spontaneous decomposition ofPPD in KGM was determined using high-performanceliquid chromatography (HPLC) as previously described.'"

1 o evaluate the effects of the autoxidation products.varying concentrations of PPD (1-20 /ig/ml) were pre-incubated for 24 h in the culture medium at 37°C andthe resulting PPD-modified medium (PPD-MM) was thendiluted 1:1 with fresh medium prior to treatment of thecells.

The keratinocytes were incubated for varying periods( iO min, 3 h, 24 h) in concentrations from 1 -20 /ig/mlPPD or with PPD-MM. After 24 h. the cultures were thenpulsed with 0 5 fiCi 'H-thymidinc for 24 h and thentreated with 10% trichloroacetic acid and the precipitateri'siispended in 1 N NaOH. The degree of 'H-thymidineincorporation was evaluated by scintillation counting.

Lipoperoxidative damage

The degree of lipoperoxidative damage resulting fromthe generation of free radical species during the autoxi-dation ofl'PD was evaluated by exposing 5 x lo'" cells for5 h to concentrations of PPD of 5-50 /ig/ml. The viabilityofthe cells after washing was checked with trypan blueexclusion. The cells were extracted in equal parts ofchloroform and methanol that contained butylatedhydroxytoluene as the anti-oxidant. The phospholipidfraction was separated by thin-layer chromatographyand the fatty acids component esteriiied using 10%boron trifiuoride in MeOH (Merck). The fatty acid methylesters were analysed by gas chromatography on a FFA-Pcolumn and the arachidonic acid/palmitic acid ratio wasdetermined.''

whether exposure to PPD or to PPD-MM induced theexpression of ICAM-1.

Keratinocytes (2 x 10"') were incubated for 1 and 3 hwith varying concentrations of PPD from ()• I to 5 /(g/mlor with PPD-MM. The cells were washed and after 24 hculture in KGM subjected to either FACS analysis or theT-cell adherence assay.

FACS analysis

Cells were incubated for 30 min at 4°C with the mousemonoclonal antibody (MAb) 84H10 (Immunotech.Marseille. France) against ICAM-1 or with the MAb toHLA-DR (Becton Dickinson. Grenoble. France). Thespecific binding was revealed using a goat anti-mouseIgG fiuorescein-Iabelled antibody diluted 1:20. Thesamples were analysed by FACS and the results reportedas an increase in the mean channel fiuorescence.

T-cell adherence assay^'

I'eripheral blood mononuclear cells were obtained fromthe blood of healthy donors and isolated by FicollHypaque gradient centrifugation. Interface mono-nuclear cells were cultured in RPMI 1640 medium with10%, foetal calf serum (FCS) and antibiotics and withconcanavalin A at 1 0 ng/ml for 72 h. The T-cell blastswere then washed twice with PBS and labelled with ^'Crusing 100 /(Ci/10" cells at J7°C for 1 h with periodicagitation. After four washes in the complete medium.10^ viable cells were added to the keratinocytes seededinto 96-well plates and incubated for 1 h at 17°C. Rachwell was gently washed three times with PBS containing1% FCS. The adherent cells were then lysed by theaddition of 1 0%, sodium dodecyl sulphate and the platescentrifuged at 800 fl for 15 min. The released "''Cr wasdetermined in 100 /il of supernatant in a gammacounter. The release of''' Cr by non-washed/lysed (maxi-mal) and washed/non-lysed (minimal) wells was deter-mined in each experiment. The data were expressed asthe percentage of adherence relative to controls. Toverify the specificity of T-blast adherence, keratinocyteswere pre-incubated for JO min with the MAb anti-ICAM-1 (Immunotech. Marseille, France) before the assay.

ICAM-1 expression

Because T-cell adherence to keratinocytes in vitro isLFA-l/lCAM-1 dependent'- both the tunctional T-blastadherence assay and indirect immunofluorescencestaining with FACS analysis were used to determine

Statistical analysis

Ail the experiments were performed in triplicate and theresults expressed as mean±SD of at least three experi-ments. Student's t-test was used to determine thesignificance.

452 M.PICARDO et ai

Results

Auto-oxidation of PPD in keratinocyte growth medium

The rate of decomposition ofPPD(l, 5, lOand 20/ig/mlin KtJM) as evaluated by HPLC is shown in Figure 1 andwas related to its concentration. The higher the initialconcentration of PPD. the lower the initial rate ofdecomposition.

Effect of PPD on keratinocyte cuitures

A biphasic response was observed to PPD on the growthrate ofthe cultures (Fig. 2). Exposure to concentrationsof 5-10 /ig/ml of PPD for 30 min gave a signiticantincrease in 'H-thymidine uptake after 24 h of culture.With higher concentrations, a cytotoxic effect wasevident as shown by a lowered 'H-thymidine uptake andthe trypan blue exclusion test. A longer period ofexposure to PPD of 3-24 h produced a significant degreeof cell death even at the low concentrations of 2-5/ig/ml. The presence of either superoxide dismutase orcatalase significantly decreased PPD-induced cytotoxi-city.

PPD-MM produced a significant dose-dependent de-crease in cell viability at concentrations above 10 /ig/ml(Fig. 3). but in this case neither superoxide dismutasenor catalase exerted any protective effect.

Lipoperoxidative damage

Significant lipoperoxidative damage, as evidenced by a

150

0.5 h 3 h 24 hExposure time to PPD (h)

ligure 2. Time- and dose-dependent effects ol' Tl'l) on uptake of '1!-thyniidiru" liy huiiiiiii kcratinocytt's in fullurc. A sigiiiliciinl iiureiisi- of'H-lhyrtiidiiifuplakf wiis detected al O-S h with 5 and lO/iy/nil IM'U.Longer periods of exposure produced a cytotoxic effect. litich resultrepresents Ihe mean ± SD of three experiments in triplicate, a . 2 ;ig/inlPPD: a . 5 /ig/ml PPD: • . lD/^g/mt PPD: S. 2()/(g/ml PPD, " l>< l)-()5:

decrease in the ratio of polyunsattirated/saturated fattyacids of cell membranes (Fig. 4) was found with longerexposure to PPD and this correlated with the cytotoxiceffect.

Expres.sion of ICAM-1

Control keratinocyte cultures exhibited only u lowadherence to T blasts. A significant increase in adher-

120 n

100

80 -QQ_

o> 60

E 40 -

20 -

10 15

Time (hours)

20 25

t'isure t. Rate of decomposition of I. 5. 10 and 20/ig/mt PPDin KGMmedium as evaluated by HPLC. Each result represents the mean ± SDofthree dctcrminitlions. • . I fig/ml PPD: • , 5 ;jg/mi PPD; w, 10I'PD: ^ 20/(g/ml PPD.

120

100 -

3h 24hExposure time to PPD-modified medium (h)

Figure i. Time- anti dose-dependent effects of PPD-moditied medium(PPD-MMl on human keratinocytes in culture. Kach resull representsthe mean ± SD of tiirw experiments in triplicate, a . 5 /^g/ml PPD; • , 10/ig/mIPPD;B. 20/ig/ml PPD. " ( ' < ( ) OS: 'P<()-()01.

PPD INDUCES ICAM-1 EXPRESSION IN HUMAN KERATINOCYTES 453

u

o

— ^

a.u'co

o

<

15 -

10 -

5 -

I/

•Cv

A--

/ •

'V/

vA• //

• ;

'-/ .

^ ^

*

1

A •'

Table 1. Immunofluorescence FACS analyses indicative of ICAM-1expression in huniim keratinocytes following treatments with PPD orPPD-MM

0 2 10 20 50

Initial PPD concentration

Kigurc 4. Mpopcroxidativf damage of huniun koratinocytes by PPD asdclcnniiicd hy llie iinichidonic iicid/ptilmitic add ratio. Each resultrepresents the meiin±SD of three determinations. "" P<0( )5 :

PPD 1h PPD 3h PPD-MM ihPPD-MM 3hExposure time (h)

Figure S. Adherence of T-coll blusl.f to hunuin kcriilinocyk's. Hiinuinkenilinocyies were ireiiled for \ itnd i h wilh PPD and P!'D-modiiiedmedium. After 24 h culture in KOM the percentage adherence of T-cellblasts relative to untreated cells was calculated. Each result representsthe meaniSD of three experiments in quadruplicate, d. 0'1 /ig/mlPPD; m. O'S ;/g/ml PPD: B. I //g/ml PPD. *'/*<()-()5; ' P<()•()()].

ence was observed 24 h after treatment with PPD atconcentrations that did not produce a cytotoxic effect.The anti-ICAM-l antibody specifically inhibited thisadhesion. The greatest increase in adherence (approxi-mately 40% greater than controls) was detected after 5 hof exposure to 1 /ig/ml of PPD. A 20 ± 5% increase inadherence relative to control was observed after 1 h of

.g/ml

0-1

! • ( )

1 h

25 ±5*20 ±6*"

Exposure time

PPD

3h 1 h

U±4 n±5J1 ± 6* 19 ± 5

PPD-MM

3h

19±5*'26 ± 6*

Kfsiilts are expressed as an increase (meaniSD: i i=J) in channeliiuorcscence (arbitrary units) relative to conlroi.' " P < O O 5 : ' P < O D ( ) 1 .No significant fluorescence indicative of HLA-DR expression wasdetected.

treatment with 0-5 \i%lwX of PPD. The percentageincreased after 5 h ot treatment, and by that time morethan 50% of the PPD would have decomposed (Fig. 1).PP[1-MM was also able to induce expression of ICAM-1on keratinocytes (Fig. S). the expression being moredependent on the concentratioti used than the period oftreatment. FACS analysis confirmed the expression ofICAM-1 (Table 1) and that expression of HI,A-DR hadnot occurred.

Discussion

Paraphenylenediamine is an aromatic amine whichundergoes autoxidation with the generation of short-lived free radicals as well as reactive oxygen species suchas superoxide and hydrogen peroxide.'" '"* By analogywith the autoxidation of ascorbic acid in cell culturemedia, which has been studied extensively.'"""'^ wesuppose that the process is catalysed by metal ions suchas Fe'+ present in the culture medium. The co-occur-rence of superoxide, hydrogen peroxide and a transitionmetal catalyst in the culture medium creates conditionsthat favour the superoxide-assisted Fenton reactionwhich generates highly reactive hydroxyl radicals.'""'''This would explain both the observed enhancement ofkeratinocyte growth by low concentrations of PPD andthe cytotoxicity and lipid peroxidation with higherconcentrations. Low levels of 10 ' w of hydrogenperoxide in mammalian cell cultures can stimulate cellrepiication. btit as this is increased cytoxicity occurs.'" Ifhydrogen peroxide, superoxide and a suitable transitionmetal catalyst are all present then the hydroxyl radicalsformed as a result will generate oxidative stress and

454 M.FICARUO el al.

hence enhance lipid peroxidation with inhibition ol'growth and produce cytotoxicity.-" In the presence ofeither superoxide dismutase or catalase. we observed asigniHcantiy decreased PPD-induced cytotoxicity (datanot shown), as expected from an interruption of theFenton cycling process.'** In guinea-pigs the appiicationof 1% PI'i) .solution to the skin resulted in a significantenhancement of lipid peroxidation.-'

Low levels of hydrogen peroxide can activate calcium-dependent protein kinase C'"- ' and therefore producemetabolic activation with an increased growth rate aswe observed with iow concentrations of PPD in thekeratinocyte culture. ICAM-1 expression is also regu-lated by protein kinase C signal transduction.-' Theinduction of ICAM-1 expression by keratinocytesexposed to PPD can again be explained by hydrogenperoxide generated from the autoxidation of PPD. Preli-minary experiments with H7. an inhibitor of proteinkinase C. have shown a decrease in the effect of PPD onICAM-1 expression. Recently it has been reported thatin-vivo application of ()-5% PPD in patch tests in non-sensitized subjects induces the expression of ICAM~1 onkeratinocytt's.-'

The results ofour experiments with PPD-MM are moredifficult to explain. The cytotoxicity of the modifiedmcdliun. as reported for oxidizable phenolic com-pounds,''^ was not inhibited by either superoxide dismu-tase or catalase. This suggests that neither hydrogenperoxide, superoxide nor hydroxyl radicals were respon-sible for the effect. There remains the po.ssibility thateither the oxidation products from PPD or oxidativelydamaged components of the culture medium wereresponsible. Simiiariy. the effect of PPD-MM on theexpression of ICAM-1 by keratinocytes can be ascribed toits oxidation products, suggesting that these productsmay possess significant biological activity.

It Is clear from our results that keratinocytes in culturecan respond to the presence of PPD. Because PPD is aknown allergen, it is of interest to consider whether thebiochemical events we have observed have any signili-cance as to the induction ofthe contact allergic reaction.In the induction of contact sensitivity, both the period ofexposure and the concentration ofthe applied substanceure con.sidered important for the immune reaction.^'' -''Keratinocytt's produce several cytokines and expressintercellular adhesion molecules, upregulate T-cell re-sponses and possess lipoxygenase activity.'*'" Theproduction of epidermal cytokines is also regulated viaprotein kinase C and is stimulated by cell membranedamage.-'^-" Cytokines and lipoxygenase products maybe responsible for initiating an inflammatory reaction

that leads to an antigen independent T-cell activation.'-ICAM-1 expressing keratinocytes bring LFA-1-express-ing T cells, which are continually trafficking in lownumbers in normal epidermis, in close apposition withthe epidennal compartmeiU. thus facilitating the inter-action between haptenized protein and the immuno-competent cells.'"•^''

It has recently been reported that mono- and dihydricalkylbenzenes with contact allergenic activity canundergo redox cycling, and hence induce oxidativestress, in mouse skin.'' Both irritancy and allergenicitycan be lowered by the administration of 2-oxo-4-thiazolidine carboxylatc. a compound known toenhance intracellular glutathione levels.'' ln a moreextensive study comparing the allergenic 2.4-dinitroha-lobenzenes and structurally related non-allergenic ana-logues, it was found that only the allergenic compoiuid.scaused significant alteration ofthe thiol-disulphide stateof the skin, being indicative of oxidative stress.^"Although we have tested only PPD in this study, ourresults are consistent with the view'" that oxidativestress may be an essential prerequisite in the induction ofthe contact allergic reaction. An interesting corollary tothis observation is that intlammation arising fromoxidative stress (through enhanced eicosanoid biosyn-thesis by cyclo-oxygenases and lipoxygenases andthrough superoxidc-mediated attracticin of ntnitrophils/monocytes to the site''*) may well be essential for thedevelopment ofthe immune reaction.

Acknowledgments

This work was partially supported by a grant fromProgetto Finalizzato liiotecnologie e liiostrumentazioni.National Research Council. Italy. We thank PatriziaBirarelli and Orictta Terminali for technical assistance.

References1 Baadsgaiird (). Immunological mechanisms of tonUut di-niiiUitis.

In: ll.vofienoiis DenmUoses: I'nvironmcnliil Dermatitis (Mt'iinc T.Maibiith H. eds). Hoca Ralon, I'loridii: CKC Press. IS9I: 1-20.

2 Barker |NW.N, Mitra RS. (Irilflths CEM ei aL KtTatinocytes asinitiators ot" inflammation. Ijimet 19^1: J37: 211-14.

i Griffiths CKM. Biirkcr |NWM. Ktiiikd S. Nickoloff B]. Modulation ofleucocyte adhesion nuilocuk's. a 'I'-cell cheniotiixin Ill.-S) and aregiiliitory cytokine (TNI's) in alltTitic contact dermatitis iKhusdernuilitis). lir I Dermatol 1991: 124: SI9-2fi.

4 I'iyuet I'l', Cniii CA'., lUiuscr C et al. Tumor necTosis factor is acritical mediator in haplen-induccd irritant and contact hyper-sensitivity reactions. / I'.xf Med 1991; 173: ()7i-9.

5 Pieardo M. /ompetta C. IDe Luca C et al. Nickcl-keratinocyteinteraction: a possible role in sensilization. Hr / Dirnuitol 1990:122: 729-3S.

PPD INDUCES ICAM-1 liXPRESSION IN HUMAN KERATINOCYTES 45S

6 Schmidt K]. Khan L. Cluing LY. Are free radicals and not quinones

the hiiptenic species derived from uriishioh and (ither contactallergenic nuino- and dihydric a Iky I hen/en esr The signiiicance ofNADH. (•liilatliione, and redox cycling in the skin. Arch Dennatol

Hrs 1991): 2H2: Sft-h4.

7 iSrittiths CtiM, Nickoloff 11|. Keratinocyte intercellular adhesionmolecule-1 (ICAM-11 expression precedes dermal T Iymphocyticiniiltration in allergic contact dermatilis (Rhus dermatitis). Ant j

Pathol I9S9: 1 J5: 1045-55.

H Cronin I-'. Contact Vennatilis, Edinburgh: Churchill Livingstone.I9K(): 119-2(1.

9 I )upuis G, lienezra C. Allergic Contact Dennatitis lo Shnple Chemicals,

A n\olecular Approach, New York: Marcel Dekker, 19H2; 66. I I I .

10 Piciirdo M, Cannistraci C, Cristaudo .\ el al. Study on cross-reactivity to the para group. Dermatoiogica 1990: 181: I()4-S.

! I t'assi S, Pieardo M. De Luca Cet ul. Blood levels of polyun.saturatedfiitty acids of phnspholipitls. vitamin li and glutathione peroxidaseactivity in patients with sebnrrheic dermalit is. / Immunol Sci I 991;2: I71-«.

12 Dustin ML. Singer KH, Tuck DJ. Springer 'I'A. Adhesion of Tlymphohlasls to epidermal keratlnncytes is regulated hy interferongamma and is mediated by intercellular adhesion molecule-1llCAM-ll. I ILxpMed I9SS: 167: 1 J2J-4().

M Cauylinuiri SW. Lian-)ie L. DegilK K. Characleri/ation and I'tinc-tioiial analysis olinierleron-v-induced intercellular adhesion [nol-eculc-l expression in human keratinocytes and A-451 cells. /Invest Dermatol 1990: 94 (Suppl.); 22S-6S.

14 Pieardo M. Passi S, Nazziiro-Porro M et al. Mechanism of antitu-moral activity ol'catechols in culture. Bioihen\ Phurmacol 19S7:Id: 417-25.

1 5 Feng ]. Melcher AH. Brunette DM cdi/. lXnermination of i.-ascorbicacid levels in culture medium: concentrations in commercialmedia and maintenance of levels under conditions of organculture. In Viiro 1977: M:91-9 .

16 Peterkofsky 11. Prather W. Cytotoxicity of ascorbate and otiierreducing agents towards cultured librohlasts as a result olhydrogen peroxide formation. / CeU Physiol 1977: 90: 61-70.

17 Koch C], Bitiglow \V., Toxicity. radiation sensitivity modification,and metabolic effects of dehydroascorhate and ascorhate inmammalian cells. / Celt Phusiol 197H: 94; 299-J06.

18 Buettner GK. In the absence yf catalytic metals aseorbate does not

autoxidize at pH7: aseorbate as a test for catalytic metals. / Biochem

Biophys Methods 198S: 16: 27-40.

19 Halliwell B. (lutteridge ]MC. Free Radicals in Biotofiy and Medicine,

2nd edn. Oxford: Clarendon Press. I9K9: Hfi-176. i59. JS9.

20 Burdon RH. Rice-Iivans C. Tree radicals and the regulation ofmammalian cell proliferation. Free Rad Res Comm 1989; 6: 345-58.

21 Mathur AK. Gupta BN. Narang S et aL Biochemical and histopath-ologlciil changes following dermal exposure lo paraphenylenediamine in iiuincii-pigs.//!;>;'//'(u-irii/ 1990: 10: J8J-fi.

22 Shasby DM, Yorek M. Shasby SS. Exogenous o.\idimts initiatehydrolysis of endothelial cell inosito! phospholipids. Blood 19S8:72:491-9.

2 J Griffiths CEM. Esinann J. Fisher G] et al Differential mmlulation ofkeraiinocyte intercellular adhesion molecule-1 expression bygamma interferon and phorbol ester: evidence for involvement ofprotein kinuse C signal transduction. Br / Dermatol 1990: 122:3 3 3-42.

24 Sterry W. Kunne N. Weber-Matthiesen K el al. Cell trafficking inpositive and negalive patch-test reactions: demonstration of astereotypic migration pathway./ Invest Lkrmatol 1991: 96: 4S9-62.

25 Thestrup-Pedersen K. Larsen CG. Rennevig ). 1he immunology ol'contact dermatitis. A review with special reference to the pathn-physiology of eezema. Contact Dtrmaiitis ]':)H9; 20: 81-92.

26 Res P. Kapsenberg ML. Bos JD. Steikama F. The crucial role ol'human dendritic antigen-presenting cell subsets in nickel-speciticT-ccll proliferation. / Invest Dennaiol 1987: 88; 550-4.

27 Ansel]. Perry P. Brown Jet ul. Cytokine modulatiim of keratinocytecytokines.; Invest Dermatol 1990: 94: I01S-7S.

28 Marlin SD, Springer TA, Puritied inlercellular adhesion moiecule-1(lCAM-11 i.s a ligand for lymphocyte function-associated antigen 1(LFA-n. CW/ 1987: 51: 8 H - 1 9 .

29 Nickoloff BJ, Role of interferon gamma in cutaneous trafficking oflymphocytes with emphasis on molecular and cellular adhesionevents. Arch Dermatol 1988: 124: 18J5-43.

30 Schmidt RJ, Chung LY. Oxidative stress occurs in skin in the pre-immunologic phase of the cell mediated immune response toallergenic 2,4-dinitrohalobenzenes. 1-rec Rad Biol Med 1990: 9(SuppL I): 147.