Induction of Tumor Necrosis Factor-a In Vivo by a Skin Irritant, Tributyltin, Through Activation of Transcription Factors: Its Phartnacological Modulation by Anti-inflatntnatory Drugs

Emanuela Corsini, Alessandra Terzoli, Alessandra Bruccoleri, Marina Marinovich, and Corrado Lodovico Galli Ccnter for Cosm cti c Toxico logy, Institute of Pharmacologica l Sciences, Universi ty of Milan, 20133 . Milan. Ita ly

Skin irritant reactions are under the control of a network of cytokines and lipid Inediators. This study characterized the production of tUInor necrosis fac­tor-a (TNF) induced by a skin irritant treatInent, tributyltin (TBT), in wce through transcription fac­tor activation and its pharInacologic Inodulation by anti-inflaInInatory agents. Tbe ears of BALB/c Inice were painted with different aInounts of TBT (67-536 nInol in acetone) or with acetone alone. At different times thereafter, TNF production was analyzed both at the InRNA and protein level, by semiquantitative RT-PCR and L929 cytotoxicity assay, respectively. TBT induced rapid (1 h) TNF gene expression and protein synthesis. MaxiInal TNF production was ob­served 2 h after treatment. The production of TNF was paralleled by accuInulation of skin water; this was partially prevented by intraperitoneal injection

Skin irritation involves resident epidermal ce lls, fibro­blasts of dermis , and endotheljal ce lls as well as invading leukocytes interacting under the control of a network of cytokincs and lipid mediators (Boss and Kapsenberg, 1993). Because of their anatom ical location, a rol e for

cytokines produced by keratinocytes, especially inte rleukin-la (IL-1 a), has been proposed (Kupper, 1990). T he hypothes is is that every time the ep idermis is damaged, active IL-1 a is re leased as a primary event in skin defense (Nickoloff ef ai , 1990). IL-la release stimu lates further produ ction of IL-l a together with synthesis of other cytokines such as lL-6, lL-8, and inte rcellular adh esion molecules, thus amplifying the response (Kupper, 1989) .

W e used tributyl tin (TBT) as an agent to incite the skin irritant reaction . Indeed, occupationa l and accidental exposure to TBT, a biocidal agent used mainly in wood preserva tion and marine antifouling paints, can result in skin and eye irritation, and severe dennatitis has been reported after direct contact with the skin (reviewed by World Health Organization, 1990) . W e have previ­ously demon strated (Corsini ef ai, 1996a) the involvement of IL-1 a

Manuscript received October 2 '1. 1996; revi sed January 30. 1997; ac­cepted fo r publication January 31. '1997.

R eprint requests to: Dr. Em anuel" Corsini. Institute of PhaJ"lnaco logic,d Sc iences . Via Balzarctti 9. 20133 Milano. Italy.

Abbreviations: TOT, tribu tyltin; NF-KB, lIuclea r factor-kJ3; EMSA. elec­trophore ti c mobility shjft assay.

of antibody against murine TNF. These data indicate that skin irritation induced by TBT is attributable, in addition to the actions of other inflaInInatory Inedi­ators, to the action of keratinocyte-derived TNF. TNF production was preceded by a rapid (5 min) activation of nuclear factor-KB (NF-KB), which was also maximal 30 min after treatInent. TBT-induced accumulation of skin water and TNF production were significantly reduced by topical treatment with dexa­methasone and pentamidine, two anti-inflammatory agents. Interestingly, dexamethasone, but not pent­aInidine, decreased TBT -induced NF-KB activation, confirIning in vivo that the glucocorticoid receptor interacts functionally within the nucleus with other transcription factors opposing one another's activity. Key words: orgallotill/NF-KBlskill iJ'Yitatioll. ] ["vest Der­lIIatol 108:892-896, 1997

in TBT -induced skin i.n;tation in the mouse. The prior injection of anti-murine lL-1 a aJltibody, however, red uced skin irritation by only 50'X,. The analysis of the expression of cytokine mRNAs by reverse transcriptase-polymerase chain reaction (RT -PCR) showed that T BT treatment is also able to induce other cytokines such as IL-6 and tumor necrosis factor (TNF) , which supports the hypoth­esis of con com.itant re lease of other inflammatory mediators after TBT treatment; inte res tingly, prior treatment with TL-l a antibody partially inhibited (45%) the TBT-induced express ion of IL-6 but not that of TNF, which may partially account fo r the antibody protection observed.

[n the skin, TNF is stored in dermal mast cells (Gordon and Galli , 1990), but followin g stimulation it may be produced by keratino­cytes (Kock e{ ai, 1990) and Langerhans ceUs (Larrick c{ ai, 1989). An increasing body of evidence suggests that TNF is fun ctionally relevant to a variety of inflammatory skin diseases, both in rodents and humans (Wakefi e ld et ai, 1991; Groves cl ai, 1995) . Antibody to TNF abolishes many inflammatory skin reac tions , including allergic and irritant contact dennatitis (Piguet c { ai, 1991) and graft- lIcrsl/s­host disease (Piguet e{ ai, 1987). An important mechanism by which TN F influen ces the de velopment of an inflammatory reaction i indu ction of the express ion of cutaneous and endothelial adhesion molecules (Fober and Co.mran, 1990; Groves cf ai, 1995). Inducible expression of cytokines is controlled by regulation of the activ ity of transcription fa ctors (Muegge and Durum, 1990). In particular, nuclear factor-KB (NF-I<B) (Shako v e{ ai, 1990) and activator

0022-202X/97/S10.S0 • Copyright 1997 by The Society for In vestigative Dermatology. Inc.

892

VOL. 108. NO.6 JU NE 1997

protein-lIc-jllll (Rhoades et (/1, 1992) have bee n implicated in TNF expression. We have pre viously d e monstrated (Corsini ef nl, 1996b) that TET is able to induce a con centration-dep endent NF-KB activation in a murine keratinocyte cell line.

The purpose oftl-us study was to furtb e r investigate in m o use skin the role of cytokines, il] partic ular TNF, in skin irritation through the ability ofTBT to induce ill Ilifl i! NF-KB activation. We focused our a ttention on this transcription filctor because NP-KE is a primary transcription factor, and its activation occurs without n ew protein synthesis (Eae uerle and Henke l, 1994). Finally , w e modu­lated TET-induced skin irrita tion using d examethasone and pent­amidine. Corticostero ids are the most effective therapy currently available for the treatment of many skin diseases, whereas pentam­idine is an aromatic diamidine currently used to treat Pll el/l/loC)'stis

carinii pneumonia in acquired immunode fi c ie n cy syndrome patients (Sands et (/1, 1985), for which w e have previously demonstrated an ti-inflammatory properties in diff'erent systems, including the skin (Blay loc k cl: (/1, 1991; Corsini cf nl, 1992, Rosenthal ct ai , 1992).

MATEIUALS AND METHODS

CbelDicals Tributyltin chloride was obtained fi'om Aldrich (Steinhcin, Germany) , dexamethasone £i'om Sigma (St Louis, MO), and pentamidine isoduonate £i'om Lyphol11ed (Melrose Park, IL). All reagents were pur­chased at the highest purity available.

Anin2als Nine CO 12-wk-old fema le BALB/c mice (Charles River. Calco, Italy) were used throughout these studies. Mice wcre randomized inm trea tment groups, housed three CO fo ur per cage over woodchip bedding , and all owed food and water ad libi/lIl11 . Before creatment mice were quarantined for 2 wk and were acclimati zed co a 12-h light-dark cycle.

Anin2al Exposure Groups of three 0 0: [our mice received d ill'erent amounts (67-536 I1mol) ofTBT disso lved in 10 /-LI of acetone on both sides of one ear or an equal volume of acetone alone on tbe other car (Dla )'lock (I aI, 1991). T he concentrations were chosen from a previous study (Middleton and Pratt, 1977) . Ears were treated on both sides to maximize indn ction of cytokine production . At various times after treatment. mice were killed . and the cars were removed. Half of each ear was nsed to dctenninc the \va ter content frOlU the corresponding wet and dry weights, the latter measured after heating at 105DC for 24 h . The other half was wash e d once in phosphatc- b\1lfcred saline and sliced (R.osenthal el 01, \ 992), and the slices were incubated overnight in 1 1111 of RJ'M[ '1640 (Sigma) containing 2 mM L-glueamine, 0.1 11Ig streptomycin per 1111 , 100 IU penicilun P CI' ml , and 10% feta l bovine serum (Sigma) (medium). IL-1 C1' and TNF released iJl to the medium wCre dctermiJ1ed by enz)'me-linkcd il11l11u­nosoibent assay (ELISA) and the L929 cytotoxicity assaI', respectively. T he TN F content of wbole ears was determined in skin homogenate prepared as fo llows. Ears were minced with sca lpels and homogenized in 1 ml of medium with a ti.ssuc homogenizer (Camlab, Cambridge, U.K.) for 20 s. The samples were then frcezed-thawed twice , sonica ted for 15 s, and fina\J)' centrifuged for 5 min ,It 15000 rpm. Supernatants were stored at - SODC.

In a n anoeher experiment. ,mimaIs (n = 4) were injected intraperitonea\Jy with 200 ILl of poll' clonal anti-mouse TNF antibody (Gcnzymc, Cambridge, MA) or with rabb it pre-immune serum 'I h and 30 min before TnT (268 runol) Or acetone treatnlent. Animals werc killed 2 h later. and skin water conte n t Was assessed as previously described.

For the pharmacologic modulation ofTBT-induced skin irritation, mice wcre painted w ith 268 nnwl of TBT or acetone and then with dexameth­asone (30 ILg in 10 /-LI of acetone) or pentamidine (80 ILg in 10 I·d of acetone). The doses of pentamidine and dexamethasone we re chosen from previo u s studies (Blaylock et nf, 1991) . Pentamidine was disso lved in distilled water at a concentration of 1 % and then diluted in acetone for topical app lication. Dexamethasone was disso lved directly in aceeone. Mice were !ciUcd 30 min or 2 h l:ltcr fo r the preparation of nuclea r extract or dete rmination of skin water and TNF content, respecti vely.

Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) RT -peR reactions were performed as previously described (Corsini ef of. 1996a). Commercially availab le PC R primers for TNF- c< and glyceralde­h)'de 3-phosplJ<1te dehydrogenase (GPDI-I) were purchased /i'om Clontech laboratories, Inc. (Palo Alto. CA) and contained the fo llowing sequences:

mouse TNF-c<: sense, S' -ATGAGCACAGAAAGCATGACCGC-3' antisellse, 5' -CCAAAGTAGACCTGCCCGGACTC-3' mouse GPDH: sense. 5'-TGAAGGTCGGAGTCAACGGATTT-3' antisense, 5' -C A TGTGGGCCA TGAGGTCCACCAC-3' T h e amplified PC R. products from cytokine RNA arc 692 bp for TNF-c<

and 983 bp for GPDH . ll1 preliminary experiments, RNA concentrations

TNF-u AND TBT-INDUCED SKI N IRR.ITATION 893

and PC R cycles were titrated to establish standard curves to document lincarity ~l11d eo permit semiquantitative an"lysis of sign al strength . Gels were photographed with type S5 positive-negative fi lm (Po laroid , Cam­bridge, MA).

Preparation of Nuclear Extracts and Electrophoretic Mobility Shift Assay (EMSA) Nuclear extracts werc extracted fro111 minced ear skin that had been washed with phosphate-bulfercd saline 'md homogenized in 2 1111 of a hypotonic lySis bu/fer (bu/fer A: '10 lllM HEi'ES, pH 7.8, 10 mM KC1, 2 mM MgCI" 1 mM dithiothre itol, 0.1 mM ethylenediaminecctraacetic acid , 0.1 111M phcl1 ),lmechylsulfon)'1 flu oride) wieh a tissue homogenizer for 20 s. Homogenates were kept Oll icc for 15 min , 12S ILl of a 10% Nonidet P-40 solution was added and mixed for 15 s, and the mixture WaS chen centrifuged for 30 s at 12,000 rpm. T he pe lleted nuclei were washed once with 400 /-L1 of butfer A plus 25 fLl of] 0% NP-40, centrifuged, alld then suspended in 50 /-LI ofbntfer C (50 mM HEPES. pH 7.8, 50 mM KCI, 300 mM NaCl, 0.1 111M ethylenediaminetetraacetic acid . 1 mM dithiothTeitol. O.lmM phenylmctl1ylsulfonyl fluoride, 10'}\, (vollvol) glycerol). mixed for 20 min , and centrifuged for 5 min at 12 ,000 rpm . T he supem atant containing nuclear proteins \vas h.u:vcsted . protein concentratio n \~' as

determined , and the supematant was stored at -80DC nntil used in EMSA . EMSA was performed as follows. Binding reaccion mix cnres (20 iL l ) containing 5 J.l.g protein of nuclea r extract. 0.5 iLl,' poly(d l-dC) . poly(dl-dC) (Sigma). 10,000 cpm 32P_Iabeled probe ill binding buller (10 mM HEPES, pH 7 .9, 50 rnM NaCI, l rI1M cthylenediamine tetraacetic acid, 1. mM dithiothreicol, 10% (vol/vo l) glycerol , 1 % Ficoll and 0.2 J.l.g per mJ albumin) \\fere incubated for 30 111jn ilt rOO I11 tClllpcrature before separation in a 7%1 acrylamide gel in 1 X Tris-boratc-ethylcncdiamiuetetraacetic acid fo llowed by autoradiography. A double- so'andcd o ligonucleotide cont'lining the bUlding site for NF-KB (S ' -GTCTCGCAATTCCCCT CTCT CAG-3 ') was l"bcled with [a-"PJdA TP (Amersham, Buckinghamshirc, U.K.) using T4 polynucleotide kinase (Amersh"m).

Assay for TNF TNF content w"s assayed by determining the cyt,oto"iciry ofTNF against sensitive L929 cells as previonsly described (Rosenthal and Corsini . 1995). R abbit anti-murine TNF (Gcnzyme) polyclonal antibody was llsed to demonstrate that the cytolytic activity Was due to the presence of TNF in the conditioned media (dara not shown) . T he results arc expressed in picogra111s/mg dry tissue o r picogra1l1s/ear for rel ease and total TNF content, respectively. TNF concentration was ca lcnlated against a standard curve with known amo ul1CS of recombinan t murine TN F (Gcn­zyme).

Enzyme Linked Immunosorbent Assay (ELISA) IL-1 C1' content was measured by a specific sandwich ELISA as previously described (Corsini c/

ai, 1996,,) . T he limie of sensitivity for this assay is 15 pg/mL Results are expressed Ul pg/mg dr)' tissue. Preliminary studies showed that the maxi­mnm TET concentratiolls tested did no t interfere with d,e ELI SA.

Statistical Analysis All expm;menrs w ere performed aC least twice; representati ve results arc shown . Statistical significance was determined by Student's t test or Dunnett's multip le comparison test as indica ted. after arwlys is of variance.

RESULTS

TBT Induces TNF Production Figure 1 shows dle RT-PCR analysis of TNF gen e expre ssion after exposure for different times and con centrati.ons ofTBT. In the control animaJs, the low level of constitutive expression ofTNF may b e associate d with tb e presen ce of dermal mast cells, w hich store previously form ed TNF (Gordo n and Galli , 1990) . Afte r treatment w ith TBT, a rapid and dose­re lated upregulation ofTNF mRNA was o bserved , detectable after 1 h of treatment and re maining e levated up to 4 h. Maximal TNF production was detected 2 h after trea tment (Fig 2); b y 24 h th e levels ofTNF w e re b ack to control v alues. As sh own in Fig 2 , the kinetic of TBT-induced IL-l a production p ara lle ls TNF release , w l-uch supports the idea that other inflammatory m ediators are re leased after TBT treatment. After 2 h TBT induced a dose-re lated produc tio n ofTNF, along with accumulation of skin wate r (Table I) . \.ntrape ritoneal injection of a TNF-neutralizing antibo d y par­tiall y prevented TBT-indu ced accumulation of skin water (Fig 3), which indica tes that skin irritation induced b y TBT is in part attribu table to the action of keratinocyte-derived TNF.

l.nducible expression of TNF is controlled b y regulation of the activity of transcription factors N F-KB (Shakov CI (/1, 1990). Indeed, as shown in Figure 4, TBT-induced T NF produ c tion was preced ed

894 CORSINI ET AL THE JOURNAL OF INVESTIGATIVE DEI~ATOLOGY

(a) Table I. Skin Irritant Induces a Dose-Related Effect on 1 h 2h 4h

o 268 o 268 o 268 TBT nmoVside

TNF-a

GPDH

(b)

+ o 134 268 536 TBT nmol/side

TNF-a

GPDH

Figure 1. Skin irritant induces TNF gene expression. T ime course and dose response of TNF gene express ion induced by TBT: RT-PCR ana lysis. (a) Photograph of TNF a.nd GPDH amplification products from murine skin 1-4 h "fter topical application of268 nmol ofTBT or "cetone; (b) the same products 2 h after different doses of TBT or acetone (0-536 omol). C. acetOlle vehicle cOll trol; T . TBT-treated allimals; (+ ). PCR­positive contro l; (- ). no RNA.

;11 II;IJO by a rapid (5 min) N F-KB activation , w hich was max imal 30 min aftcr treatment.

Anti-inflammatory Agents Modulatc TBT -Induced Skin Irritation W c thcn investigated whether TBT -induccd skin irritation is modu lated by dexamethasonc and pentamidine, two anti-inflammatory agen ts. Corticosteroids are the most effective therapy currently ava ilable for the treatment of many skin diseases, although we have previously demonstrated thc an ti-inflammatory properties of pentamidine in different system s, incl uding the skin (Blaylock ct ai , 1991; Corsini cl ai , 1992, Rosenthal el ai, 1992). Table II shows that both compounds are indeed able to reduce

500 ',$

400

-0

.,;

'" 300 E , '" c.

:'. .:, 200

100

\. $ i \ .. [ ....................................... j

12 Time (h)

16 20 24

2 r

.,;

'" E m ~ . z ...

Figure 2. IL-la and TNF show parallel release following skin irritant reac tion. T ime course ofTBT -induced release off L-l a and TNF fro m ear skin slices at different times after topical application of268 IImol of TBT. T ime 0 shows basa l release of IL-l a and T NF from control animals; the release remained constant throughout the times ana lyzed and was not affected by application of vehicle. Eacb va lue represents the mean ::: Sf) of th ree determinations. Statistical analysis by Student's t test; " p < 0.05 for IL-l a release l/CJ'SIiS control; S, p < 0.05 for T NF release "ers lIs control.

TNF-a Production and Accumulation of Skin Water"

TBT T NF-a TNF-a Total Skin Water App lied R.elease Content Content

(ll111ol / side) (pg/mg d.w .) (pg/ear) (mg H20 / mg d.w.)

0 0.4 ± 0.1 8.9 ± 2.5 0.84 ± 0.06 134 2.7 ::: 0.5" n.d. 1.51 ::: 0.04" 268 7.3 ± 0.3" 103 ± 45.6" '1.58 ± 0.24" 536 8.5 ± 1.2" 281 ::: 133" 1.67 ::: 0.08b

,/ Groups of three mi ce were treated with diJrcrcnt concentrations ofTBT or acetone vehicle a ll both side of each ear. Anim:l ls were terminated 2 h late r. and TNF-Q: re lease, total TNF-a . :md skill water content of the cars were measured . Each value represellt.s the I11C<lll ± SO of three determinations.

/. SI":1ristic:lI analysis by D unnett's [cst: • . p < 0.0 1 I't.TSI/S co n trol (treatment 0) ; n.d., not done; d.w .• dry weigh t.

signifi cantly both TBT-induced T NF produ ction and skin water accumulation, which confirm s that the ir anti-inflammatory capacity is associated with a modulation of cytokine production . In Fig 5 w e showed that dexam ethason e is able to prevent TBT-inducec! N F- /(B activation ; 11 11;110, confirming that inhibi tion of transactiva, tion ma y be an important m olecular m echanism for dexam ethasone anti-inA am mato ry action, although this is not th e mechanism fo t pentamid ine.

DlSCUSS ION

T he purpose of this study was to charactel;ze in mice the produc, t ion ofTNF and the activa tion oftransct;ption factor NF-KB during an irritant reaction, using TBT to incitc the reaction , and theiJ: m o dulation by anti-inflammatory drugs. W e chose TBT because it is a biocide that has been extensi ve ly used in the past, and its irritant e ffect is w e ll documented in both human and roden ts (Worle! H ea lth Organ iza tion , 1990). As an ;11 ";110 model we chose the eat swcll.ing test. We have previously es tahlished (Corsini cf ai , 1996a) that TBT applied to mouse ears induccs an irri tant reaction wid, maximum levels w ithin 4 h, a peak time similar to that found by Middle ton and Pratt (1977) for TBT and rat dorsal skin. A s previo usly m entioned , we have also dcmonstrated a direct role Of IL-1 a in the ;11 11 ;110 effect of TBT. Prior treatment with a neutral, izing antibody against IL- 1a reduced skin irritation by only 50%, however, w hich indica tes the involvcm ent of o ther TBT-induced inflammatory mediators such as arachidonic acid metabolites Or other cytokines. In particular , as indicatcd by R T -PC R analysis, T NF expressio n was not affected by prior treatment with IL-1 Q antibody . lndeed, he re we dcmonstra te that the kinetics of TNt:

Figure 3. Protective effect ofanti-TNF antibody in skin irritation. Protection by injected antibody to murinc TNF against TI3T-induced increase in skin water content 2 h afte r treatment. Each va lue represents the mcan ::: SD of three to fO ll r dercn11 inations . Statistical ana lysis by Student' t test; . , p < 0.05 "e/'s IiS vehicle control; S, p < 0.05 l"/'SIIS animals treated with TBT and normal rabbit seru m.

VOL. 1 08 . NO. (, JUNE 1997

o 5 30 60 min 30 min + cold probe

NF-KB

free probe

Figu.re 4. Skin irritant rapidly induces NF-KB activation jll "j"o. Time COurse of TBT -induced N F-KB activation aftcr topical application of 268 1unol ofTBT or acetone vehicle control (0 min) and the effect ofNF-KB competition when a 100-fold excess of cold probe was used to demonstrate NF-KB complex specificity (cold). Equal amounts (5 JLg) of nuclear extracts were analyzed by EMSA. with a UP_labeled DNA probe detecting the binding activity of NF-KB.

and of lL-1a release from ear slices after trea tment with TBT occurred in parallel , which confIrms that TNF production induced by TBT is independent of IL-1 a release and clearly indicates that IL--l a is not necessarily the primary event in the skin defence mechanism but only one of the inflammatory mediators involved.

Recombinant TNF has marked pro-inflammatory effects in nor­mal skin and is capable of recruiting a variety ofinflammatory cells, modulating adhesion expression, and inducing changes suggestive of L3J.lgerhans cell mi gration (Groves c( ai , 1995). It has been previously reported that keratinocyte-derived TNF mRNA is rap­idly upregulated following treatment with irritants and allergens in mice (Enk and Katz, 1992). Also in our experien ce, skin irritants, such as sodium dodecyl sulfate, or skin allergens, such as dinitro­chlorobenzene and oxazolone, all induce TNF production (data not shown). Here, we show that TBT also induces a rapid and dose-dependent TNF production , which was l11<ua.mal 2 h after treatrrlent. Skin water accumulation following TBT treatment could be modulated by systemic treatment of mice with a neutra l-

Table II. Pentamidine and Dexamethasone Modulate Skin Irritant-Induced TNF-a Production and Skin Water

Accumulation"

T rcatnl cnt

Acetone TaT 268 nnlOl Pentatnidine 80 fLg PentaITIidine + TBT Dex3Jnethasone 30 JLg Dexamethasone + TBT

TNF-a: Total Content (pg/ea.r)

1.9 ::':: 0.6

11 8.6 ::':: 34.2" 2.9 ::':: 1.0

65.7 ::':: 9.6" " 2.6 ::':: 0.4

17.9 ::':: 6.3.1· ,.

Skin W ater Content (mg H 20/mg d.w.)

1.27 ::':: 0.05 1.94 ::':: 0.13" 1.13 ::':: 0.05 1.44 ::':: 0.01"' " 1.07 :': 0.03 1.35 :': 0.09'"

Il Groups of fOllf mi ce were painted 011 both cars with acemne or an anti ­inRam nlatory in th e pn.~$cncc or absell ce of TBT. An illlaJs werc killed 2 h late r. Each valu e represents the m ean ::!: SO of follr dctcnnin<ltions. Stati st ical analysis by SCUdCllt' s

t tcst. 'with as follows: b p < 0.0 I " S relati ve cOl1trol. ' p < 0.05. ,I P < 0.05 lIS TBT -treated :l1IimaLs. • p < 0.05.

TNF-" AND T I3T-1 N D U C ED SK IN I IUUTATION 895

+ pentamidine + dexamelhasone

I o 268 o 268 o 268 TBT nmoUside

NF-KB

free probe

Figure 5. Dexamethasone modulates NF-KB activation j" v jl'o .

Modulation of TBT -induced NF-KB activation after 30 min of treatment with acetone (0) or 268 11111 0 1 of TBT in the presence or absence of pentamidine or dexamethasone ( + ). Equal amounts (5 fLg) of nuclear extracts were analyzed by EMSA, with a 32P-labeled DNA probe detecting d,e binding activity of NF-KB.

izing antibody to T NF, so this cytokine must play an important role in irritant reactions and, in particular, in TBT-induced skin irri ta ­tion .

The activation of transcliption factor NF-KB is involved in TNF expression (Shakov c( ai, 1990). In this study, we investigated th e ability of TBT to induce N F-KB activation i/l lIivo. In the resting cell, NF-KB resides in the cytoplasm as a latent form compl exed to the inhibitory proteii, IKB (Baeuerle and H enkel, 1994). Upon cell stimulation, NF-KB is rapidly activated by its release from I KB , and this allows it to migrate into the nucleus and subsequently bind to DNA. NF-KB controls the inducible ex-pression of various gen es that are involved in immune responses and inflammatory and cellular defense m echanisms. We have previously shown (Corsini c(

ai, 1996b) that TBT induces a con centration-dependent NF-I<B activation in a murine keratinocyte cell line (HEL30). Here. we report tl,a t TBT induces a rap id (5 min) activation ofNF- KB ill. uil/o.

An early toxic action of TBT in rat skin is to inhibit oxidative m etabolic pathways (Mjddleton, 1982); trio rganotins, in general. bind to a component of th e ATP synth ase complex , disturb the pro ton gradient, and inhibit mitochondrial ATP synthesis (Snoeij e (

ai, 1987; Marinovich el ai , 1990). We speculated th at TBT is also likely to cause disturbances ill vi l)o in the respiratory chain in mitochondria , with the productio n of reactive oxygen inte rmedi­ates that activate transcription factors and promote th e synthesis of TNF and other pro-inflammatory molecules, as already demon­strated ill I) i(/'O (Corsini c/ ai, 1996b) .

Finally, we investigated the possibility of moderatin g TBT­induced skin initation by th e use of dexamethasone and p entami­dine . D exam etllasone and pentamidine both reduced TBT-induced skin water accumulation and TNF produ ction . Regarding the m echanism of action, recent reports have indicated that stimulation of glucocorticoid receptors antagonizes the action of many inflam­matory m ediators ' second m essenger systems tha t all con verge on th e activation of transcription factors (Schule c/ ai, 1990; Adcock el ai, 1994). T hese i/l vil l'O data indicilte that members of distinct classes of transcription f.'1ctors can oppose o ne anoth er 's ac tivity through a mech anism likely to involve protein-protein interactions. Indeed , we demonstrated i ll villo that inhibition of transactivation may be an important molecular m echanism of dexamethasone action but not that of pentamidine. It is not surprising that pentamidine did not interfere with NF-KB ac tivation because pentamidin e is a specific and efFective inhibito r of the release of cellular IL-l and TNF release fro m m acroph ages i" IIi/ro , acting throughout a post-translational mechanism (Rosenthal el nl, 1991 ;

896 CORSINI ET !IL

Corsini ef ai , 1992) and conseq uently is an effective blocker of the murine response to endotoxin (Rosenthal ci nl, 1992). In addition, pentamidine exhibi ted less broad immunosuppressive action than dexamethasone.

Our data support the direct induction ofTNF production ;11 11;110

by irritant, independently of lL-l ex release, and th e involvement of TNF in skin irritation, which could be modulated by anti- inflam­matory drugs. Fina lly, we confi rmed ;11 11;110 that inhibition of transactivation may be an important molecular mechanism for the action of dexamethasone.

REFERENCES

Adcock 1M. Shirasaki H. Gelder CM , Peters MJ, Brown C R , lJarnes PJ: The e(fects of glucoconicoids on phorbol ester and cytokinc stimulated transcription f.1ctor activacioll in human lung: . Life Sci 55 :1 47- 153, '1994

B:lCucrJc PA, HenkcJ T: Function and activation of NF-K.J3 in the immune system. AIUII/ Rei! 1/1//111/1 /0/12:14 1- 179 , 1994

Blaylock BL, KOllchi Y. Comment CE. Corsini E. Rosenthal GJ . Luster MI : Pcnt:ll11idinc iscthionatc reduces la expression and antigen presentation by Lmgcrhans cell s rind inhibits rile contact hypersensitivity rcacrioll. J 1111111111Wl

147:2 11 6-2121. 199 '1 Boss JD. Kapse l1berg ML: The skjn immune system: progress in cutaneous biology.

I",,,,,,,wl "I"oday 14:75-78. '1993 Corsilli E. Druccolcri A. Marinovich M. Galli CL: Endogenolls intcrleukin-I a is

associated with skin irritation induced by tributyltin. Tf1x;col Appl Pltarl1lncol 138:268-274. 1996.

Corsini E. Craig WA. It.oscnthal GJ: Modulatiol1 of tl1l11or necrosis facror release fro l11 alveo lar macrophages trcated with pentamidine iscthiOlwtc. III f) 1111",,,lIopltnn1larol 14:1 21-130. 1992

Corsini E. Schubert C. Marinovich M. Galli C L. : Kolc (lfmitochondria in tributyltin­induced inrcrlcukiu- 1 a production in IllUrillC kcrorti noc),lcs. J !tlllcst Derma/al 107:720-725. 1996b

Enk HA, K.atz SI: Early molecular events in the induction phase of contact scnsitivity. I'r"c NtI(I !lClld Sci USA 89: ,1398 -1 402, '1992

Gordon JR. Galli SJ: Mast tells as a source o f both preformed and immun ologica.lly inducible TNF-a lpl m/cach ectin. Nafflrc 346:274-276. 1990

Groves R W. All en MI-I. Ross EL. Barker JNWN , Macdonald OM: T lIlllour necrosis factor :llpha is pro-inflmn l'natory in norm:,1 hu m,lI1 skin and modulates cutaneous adhesion molccuJc expression. nrJ DCI'll/lifo/ 132:345-352. 1995

Kock A. Schwarz T, l(jrnbauer R . Urbanski A. Perry P, Allsel J C . Luger TA: Human keratinocytcs are II source for t.UJllor necrosis fa ctor alph a: evidence for synthesis ,md release upon stimulatio n with endotoxin or ultravio let light. J Esp ivied 172: 1609-161 4,1990

Kupper TS: Immune and infialnmamry process in cutaneous tisslies. M echanisms and specu lalions. ) Clill /r",,,s( ~6: I 783-'1789. 1990

Kupper TS: Mechanism of CUl:aIl COUS inflammation. Intcraction helween epidermal cytokincs. adhesion f1lOi eculcs and leukocytes. Arcll De,.,IIII'(I/ 125:'1406- 1412. '1989

THE JOURNAL OF IN VESTIGATIVE DERMATOLOGY

Larrick JW. Morhenn V. C hiang: YL, Shi T: Activated Langcrhans cells release tumor necrosis factor.) Lcllkoc Bioi 45:429-433. 1989

Marinovich M. Viviani B. Galli C L: Reversibility of TllT-illduccd protein synthesis illhibition after· ATP recovery in H EL30 cells. Tox;c,,1 L"" 52:311 - 317.1990

M iddl c roll Me: Ea rly metabolic changes in rat epidermal and dermis following cu taneoliS application ofirri rant doses oftributyltin.J lImcs' Denll ntol79: 163-166. 1982

Mjddleron Me, Pratt I: Skin water contenr as.1 quantitative index of the vascular and histologic ch,H1ges produced ill rat skin by di-n-butyltin and rri-n-bury ltil1 . .lIIl Ilt~SI

0<,.,,11/(01 68 :379-384. 1977

Mucgge K. DlIrul11 SK: Cytokincs and transcription factors. C )'(ok;"c 2: 1-8, 1990 Nickoloff" BJ, Griflith CEM, Barker I N-WN: T he role of adhesion molecules. chemo­

ta ctic fllcto r.s and cytokincs in infl:lJt1m:ltory and neoplastic skin discases-1 990 update.) J,,,,c.'( 0",.,,11/(0194: 1515-1575. 1990

Pig-uet PF. Grau GE. Allet B. Vassa lH P: Tumor necrosis f.1ctor/cachectin is an effector of skin and gut lesions of acute phase of g raft-vs-host disease . J Exp tWed 166: 1280-1289 . 1987

Piguer" PF. Grau GE, Houser C. Vassalli P: Tumor necrosis f.1ctor is a critical mediators in hapten-in duced irritant and contact hypersensitiviry reactions. J Exp l\ ifed 173:673-679, 1991

I'ober J S, Contran RS: Cytokines and endothelial cell biology. I'lty,,;,,1 R"" 70:427-451.1990

n..h oadcs KL, Golub SH, Economou J S: The rcguhtion of the human tulllor necrosis f.1ctor t."t promoter region in macrophage. T cell, ;1IId B cell lines. J W"l Chelfl 3 1 :22 I 02- 221 07, 1992

R osenthal GJ. Corsini E: TUlllor necrosis r.'lcror-cr ill inunun otox.icity assessment. Ill: Burlcnson, Dc,IIl, Munson (cds .) A1ctilOds ill 1"''''"11a/"s), . Vol I. pp 327-343. Wi ley liS!, N ew York. 1995

Rosenthal GJ. Corsini E. Craig WA, COlllment CEo Luster MI: Pentamidine: an inhibitor of intcrlcukin-l that acts via a post-translational event. Tox;col Appl I'I"' rlllor.o/ I07:555-561. 1991

Rosenthal GJ. Craig WA , Corsini E. Taylor M. Luster M.: Pen tamidine blocks the pathophysiologic effects of endotoxclni:1 through inhibition of cytokillc release. T",·;r.ol IIppl 1'1"'l'lIIor.,,1 11 2:222-228. 1992

Sands M. KrOll MA. Brown R..l3: Pentamidine: 01 review. U. CII ltifcct Dis 7:625-634, 1985

Schulc R, Rang;uajah P. JGiewer S, Ransone LJ. UoladoJ . Yang N. Verna 1M, Evans Il..M: Fun ctional ant;'lgonislll between o ncoprolcin c-jun and the glucocorticoid receptOr. Cdl 62:'12 17-1 226, 1990

Slwkov AN. Collart MA , Vass:.tlli P, NcdosJ>asov SA , Nongcllcci CV: Kappa B-type enhancer OIre involved in lipopolysaccharide-mediated transcriptional activation of the tumor necrosis factor alpha gene in primary m:lcrophagcs. J Exp Nfed 171:35-47, 1990

$nocij NJ, Penninks AH . Seincl1 W: Biological :Ictivity oforganotin compounds- an overview. E,win". Res 44:335-53. 1987

Wakeficld PE. J am es WD. Salnlaska C Po Mcttz-cs MS: Tumor necrosis f.1ctor. J Am !Ieod Oemln(ol 24:675-685, 199 I

World I-Iea lrh Organization: Tributyltin compollnds. Environmental I-Iealth C riteria 11 6. WHO. Geneva. 1990