Staphylococcal Enterotoxin B Upregulates Expression of ICAM-1 Molecules on IFN-y-Treated Keratinocytes and Keratinocyte Cell Lines

Hisashi Wakita, Yoshilci Tokura, Fukumi Furukawa, and Masahim Takigawa

Department of Dermatology, Hamamatsu University School of Medicine, Hamamatsu, Japan

The effects of staphylococcal enterotoxin B (SEB), a Staplrylococws aureus- derived bacterial superantigen, on expression of intercellular adhesion molecule-1 (ICAM-1) were examined in cultured normal and

transformed (DJM-1 cells) human keratinocytes by flow cytometry, confocal microscopy, digital image

processing, and reverse transcriptase-polymerase chain reaction.

SEB significantly upregulated ICAM-1 expression in the interferon-y (IFN-y)-pretreated, HLA-DR­

positive normal keratinocytes and DJM-1 cells in a dose-dependent manner, but not in the untreated, HLA-DR-negative cells. Other toxins such as diph­

theria and pertussis toxins did not have the effect. The distribution of SEB and HLA-DR molecules

was identical on the IFN-y-treated, HLA-DR-posi­tive DJM-1 cells by confocal microscopy. Digital image processing analysis demonstrated that SEB induced a transient increase of intracellular calcium

c ertain inAammatory skin diseases such as atopic

dermatitis (1 ,2] and psoriasis [3] are frequently colonized by Stnphylococws nurcus. l11 vitro studies

have shown that bacte rial superantigens such as

staphylococcal enterotoxin B (SEB) <md exfoliative

toxin induce T-cell proliferation in the presence ofLangerhans cells

and in terferon-)' (IFN-y)-treated, major histocompatibility com­

plex (MHC) class 11-positi ve kcratinocytes [ 4,5]. Cytokin es re­

leased from these cells and regulated expression of adhesion

molecules may in part be responsible for the prolife ration, beca use

binding of bacterial supcrantigens to MJ-I C class II molecul es [6]

upregu lates interleukin-1 13 and tumor necrosis factor-a (TN F- a )

m.RNA levels in antigen-presenting cells [7]. These observations

sugges t that bacterial superantigens penetrate the cutaneous lesions

through defective skin barriers, thereby stimulating infiltra ting

T cells and modifYing the course of the disease. However, direct

Manuscript received September 16, 1995: final revision received June 17. '1995: accepted for publication June 21, 1995.

Preliminary portions of this work were presented at the 43rd annual

symposium on the Biology of the Skin, Snowmass, Colorado, August 6-1 1, 1994.

Reprint requests to: Dr. Hisashi W:~kit:~, Department of Dermatology,

Hamamatsu University School of Medicine. 3600 1-land:~-ch o, Hamam:~tsu

43 1-3 1, Japan. Abbreviations: PE, phycoe rythrin: SEI3, staph ylococca l enterotoxin H.

concentration only in the IFN-y-treated DJM-1 cells. Pretreatment of the IFN-y-treated DJM-1 cells with anti-major histocompatibility complex class II monoclonal antibody completely blocked the effect

ofSEB. Furthermore, ICAM-1 mRNA was detected in the IFN-y-pretreated, SEll-exposed normal keratino­

cytes by reverse transcriptase-polymerase chain reac­tion.

Our results demonstrate that SEB binds to kerati­nocytes, presUJnably via. major histocompatibility complex class II molecules such as HLA-DR, triggers calcium mobilization, and induces the synthesis of ICAM-1 molecules. We speculate that, in various cutaneous disorders, SEB penetrates the epidermis and interacts with HLA-DR-positive keratinocytes to

upregulate ICAM-1 expression, thus modulating the course of the infia1nn1.atory process. Key rvords: cell

adlresion moleculelsuperautigen!signal tmnsductionleytoki11e. J Invest Der·matol 105:536-542, 1995

inAuenccs ofbactcrial superantigens o n keratinocytes have not been

fully examined , except for one report showing that SEB augments

the production ofT-cell-activating cytokincs by keratinocytes (5] .

In addition, Nickoloff cl n/ [4] suggested that n on-MHC-class

molecul es on c ultured human keratinocytes could bind superanti­gcns.

In several cutaneous diseases, intercellul ar adhesion molecule-1

(ICAM-1) is one of the important adhesion m o lec ules for cell-cell

interactions [8] such as leukocyte trafficking (9-12] . Recently,

Cornelius et nl have shown the existence of both tissue- and

cytokinc-specific responsive clements in the 5' flanking region of

the lCAM-1 gene and the dependence on cellular context for

regulatory effects by such elements [13]. In the present study, we

examined the effect of SEB on ICAM- 1 expression in cultured

human keratinocytes . By Aow cytometry, confo cal microscopy,

digital image processing, and reverse transcrip tasc-polymerase

chain reaction, it was shown that SEB actually bound to MHC clas

II molecules on ke ratinocytcs, mobi lized intracellular calcium , and

indu ced surface expression of lCAM-1. These resu lts suggest that

bacterial superantigens inAuence the outcome of cutaneous inAam­

matory disorders by modulating the expression of cell adhesion

molecules on keratinocytes .

MATERIALS AND METHODS

Reagents and Antibodies SEB was purchased from Sigma Chemical

Co., St. Louis, MO ; biotinylatcd SEB from Tox.in Technology h1c.,

0022-202X/95/S09.50 • SSDI0022-202X(95)00325-F • Copyri ght © '1995 by The Society for Investigative Dermatology, Inc.

536

VOL. 105. NO. 4 OCTOBER 1995

Sarasota, FL; fluorescein isothiocyanate (FJTC)- conjugated m onoclonal antibody (MoAb) 841-1 .10 (anti-human IC AM-1. mouse lgG I) from lmmu­notech, Marseille, France; divalent anti-MHC class II MoAb (CR3 /4 3) fro m DAKO A/S, Glostrup , Denmark; FITC-strcptaviclin , diphtheria toxin , pertussis toxin , epidermal g rowth f.,cto r , and whole bovine pituitary extract fro m Gibco l3RL, Gaithersburg. MD ; recombinant human IFN-y and TN F - a from Genzymc Corp .. lloston , MA; FITC-conjugated mouse lgG I isocyp e control antibody. phycoerythrin (PE)- conjugatcd anti-HLA-DR MoAb (L243) and PE-conjugated mouse lgG2 isotypc control antibody fro m Becton Dickinson, Mountain View. CA; non-labeled anti-1-lLA-.D.R MoAb (L243) fi·om Leinco Technologies, Inc .. Ballwin, MO ; and anti­MHC class I MoAb (B9.12.1) and PE-conjugated anti-mo use IgG goat poly clonaJ antibodies from Cosmo 13io Co. Ltd. , Tokyo, J apan .

•' Cell Culture DJM-1 cell s, a human squamous cell ca rcinoma line originating from skin [1 4] (kindly provided by Dr. H. Yaoita , Jichi Medical School, Tochigi , 'md Dr. Y. IGtajima, Gifu University, Gifu, Japan) and normal human keratinocytes . derived from foreskim (Epipack , C lonetics Corporation, C A). we re g rown in scrun1-frcc keratinocyte gro\vth n1 cdium (KC-SFM, Gibco Blli). supplemented with 10 ng/ ml of epidermal g rowth fac tor and 60 ng/ml of who le bovine pituitary extract in 24-well Rat­bottomed culture pl;1tcs (Nunclon. Nunc, Denmark) in a humidi fied incubator with 5% C02 in air at 37°C. Cells for confoca l microscopi c ana lysi s and measurcn1cnt of intrace llul ar calciun1 concen tration were grown on cover g lasses .

Because cytokine-mcdiated modulation of ICAM-1 expression is not different in transformed and normal human kcratinocytcs [15 ,16]. DJM-1 ce1ls were pri1narily used in the present study , and norn1al hurnan ke rati­nocytcs were used o nl y when specificall y m entioned.

., Induction of the ICAM-1 Molecule Cells at 90% conRuency were refed with 1 ml of KC-SFM without epidermal growth factor o r whole bovine pituitary extract to inhibit proliferation, treated with vaoious con­centra tions of I FN-y in the same medium for 48 h. washed in phosphate­buffe red saline (PBS, 10 mM sodium phosphate , 133 mM sodium chloride, pH 7 .4) to remove IFN-y, and incubated with various concentrations of SEB for 48 h .

Conditioned Medium DJM-1 ce lls were pre-treated with 25 U / ml of ( JFN--y for 48 h, washed with PI3S, and further incubated with 600 ng/ ml of

SEB for 48 h. Equal volumes of the supernatant obtained fi·om the cul ture and fresh KC -SFM were mixed and used as "conditioned medium." Although the condi tio ned medium contains 300 ng/ ml ofSEB, the biologic acti vity of SEB required for induction ofT-cell proliferation (one of the

, most sensitive methods to eva luate th e activity ofSEB [4,5]) was lost during the 48-h incubation (data not sho wn).

Flow Cytometry Single-cell suspensions were prepared b y treatm ent of the c ultured cell s with 0.25% trypsin and ·1 mM ethylcncdiaminetctrnacetic acid in PBS . After being washed with PI3S. 105 cell s we re incubated with 5 J.Ll of FITC-conjugatcd anti-I CAM-1 MoAb and 5 p.l of PE-conjugated an ti-HLA-D R MoAb (L243) o r m ouse isorype control antibodies for 30 ntin on icc. To ana lyze SEll binding to D.JM-1 ce ll s, we used a previously described method [1 7] with some modifi cation . In brief, cell suspensions were prepared for two experiments. Cell suspensions were incubated with FJTC-conjugated ami-ICAM-1 MoAb and PE-conjugatcd anti-HLA-DR MoAb (L243) as mentioned above . T he other suspensions were incubated

" with 600 ng/ml of biotinylated SEB fo r 1 h and then with 2.5 /Lg/ ml of FJTC- streptavidin. Binding of SEB to HLA-Dilmolecules was blocked by the addition of 10 p.g/ml of anti-MHC class ll MoAb (CR3 /·13) o r anti-HLA-DR MoAb (L243) before incubation of the cells w ith biotin y­lated SEB. The sta inccl ce ll s were analyzed in a FACSc;on (Becton Dickin­so n). M ean Au o rcscencc intensity on the abs t: issa was expressed as n logarithm.

Confocal Microscopic Assessment of the Localization of SEll­Binding Sites DJM-J ce ll s were incubated with 25 U / ml of IFN-y for 48 h , wash eel with PBS, and then incuba ted with 600 ng/ml of biotinylated SEB for 30 min in a C02 incubator at 37°C. After being washed with cold PBS, the cell s were fix ed with methano l at - 20°C, ;o ir-clried. rehyd rated with PBS, ancl incubated with 2.5 p.g/ml of FITC-streptavidin and 5 p.l of PE-conjugated anti-1-lLA-Dil MoAb (L243) for 30 min. T he specimens mounted in glyce rin-PBS we re observed in a Biorad Ml'lC600 ( ;ombridgc.

" MA) confoca l laser scanning microscope equipped with an argo n ion laser exciting maximally at 488 11111 and 5 14 nm and operated under CoMaS software. 131ue (BHS) and red (GHS) hi gh-sensitivity filter blocks containecl excitation fdters 488 nm DFIO and 51411111 DFIO, dichroic filters DR 510 nm LP and D R 540 nm LP, and emiss ion fi lters SIS 11111 LP and 550 nm LP, respectively. All images were genera ted fro m a frame of 51:! X 768 pixels.

SEB UPR.EGULATES KEflATINOCYTE !CA M-I EX PRESS ION 537

Measurement of Intracellular Calcium Concentration D.JM- ·1 cell s were e ithe r un treated or pre-treated \:vith 25 U / n1l of I FN-y for 48 h and then incubated with 20 p.M acctoxymethylester of Fura-2 (Doj in Labora­tories, Kumamoto. Japan) in phenyl red- free keratinocyte growth medium (Cionetics Corporation) in till! clark for I h at 3 7°C. Fo ll owing extens ive washing, 600 ng/ml of SEB, I 0 p.g/ ml of anti-MHC class II MoAb (CR3 /43 ). o r 10 p.g/ ml ofanti-MH C class I MoAb was adcled to tl1e cells in the snn1e n1 ediun1. Five rninutcs later the M oAb-treated cell s were exposed to 600 ng/ml ofSEB. Fluorescence intensity was rn easured using a dig it;o l image processing system, Argus-1 00/CA, designed by 1-lmnamatsn Photonics Co. Ltd . (1-lamamatsu , Japan) .

Reverse Transcriptase-Polymerase Chain Reaction DJM-1 ce lls and normal human keratinocytes w ere incubated with 100 U / ml of IFN-y fo r 48 h . Ce ll s we re washed with PBS and left in medium alone for 48 h to minimize the influence of IFN-')' and then incubated with 800 ng/ m l of SEB. Tota l RN A was iso lated 4 h later by the single-step guanidinium thi OC)'anate method ofChom czyu ski and Sacchi 11 8]. After lys ing of cells in gnanidinium thiocyanate , RNA was iso lated by pheno l chloroform extrac­tio n and e thanol precipitation. Tota l RNA ('I p.g) was re verse transcribed to eDNA in the presence of 2.5 p.M of random hcxamers, l mM of deoxy­nucleotide triphosphatcs, and 2.5 U / p.l of reverse transcriptase (aU fi·om Perkin Elmer Cetus, N orwalk. CT) and then amplified for 35 cycles w ith a GeneAmp l'l.NA PC R kit (Perkin Elmer Ce tus). accord ing tO the manu folc­turer's descrip tion. Each cycle consisted of denaturation at 94°C for I min . an nealing at 55°C for 1 rnin and extension at 72°C for 1.5 min . The sequences of primer pairs , 5 ' and 3 '. were as fo llows : ICAM-1, GTGA­CATGCAGCACC T CCTG and TCCATGGTGATCT C TCCTCA: and /3-acri n. GATTCCTATGTGGGCGACGA and GTGTCCATCACGAT­GCCAGT. The product size was 408 bp for ICA M-1 and 3 '14 bp for /3-actin. The PCR products and DNA m olecular weight m arker VI (Boehringer Mannheim GmbH Biochemica. Mannheim) were separated in 2% agarosc gels.

Statistical Analysis Data from th ree independent experiments were ana lyzed with the Student t test.

RESULTS

IFN-y Induces ICAM- 1 and HLA-DR on DJM-1 Cells First , the time course and con centration dependency for the IFN-y­indu ced express ion of l CAM-1 and HLA-DR w e re determined . ICAM-l was con stitutive ly expressed o n D JM-1 ce ll s. IFN-y at 25 U / ml promptly augmented lCAM-1 expression, w hic h reached a ma;..:imum intensity afte r 24 h inc ubation and then gradually d ecl ined to approximately 60'}1, of the peak intensity at 96 h (data not shown). Although not detec ted in untreated DJM-1 ce ll s, the expressio n of I-ILA-.DR was gradually indu ced with 25 U / ml of IFN-y, peaked at 48 h inc ubation , and plateau ed shortly the rea fte r with a minimal loss of the inte nsity b y 9 6 h (Fig la) . In contrast ro ICAM-1, HLA-DR was onl y weakly expressed at 24 h in c uba tion. T h e in tens ity of both ICAM-1 and HLA-DR (Fig 1/1) expression increased sh arp ly up to con centrations of 50 U / ml IFN-y and plateau e d at concentration s up to 200 U / ml. Th ese results were consistent with those previously describe d 1'1 9].

SEB Upregulates ICAM-1 Expression on IFN-y-Treated Keratinocytes ICAM-1 was constitutive ly expressed on DJM-1 cells even w h e n g rowth was inhibited b y the d epletion of epide rmal g rowth f.1c tor and w h o le bovine pituitary extract fi·om the medium (Fig 2, e xperiment n). SED at 600 n g/ml did not augment ICAM- l expression in ce ll s t hat were n ot previously treated w ith IFN-y (Fig 2, expcrirncnt b). W h en the cells were in ubate d with 25 U / m l of IFN-y for 48 h and left un t reated for an addition al 48 h. ICAM-1 express io n was still e levated (Fig 2, experiment c). W h en the lFN-y-treated cell s w e re inc ubated w ith 600 ng/ ml ofSEB for 48 h (Fig 2 , exp e rimen t d). the Au o rescen ce inte nsity was further inc re ased w h en compared with th e SEI3-untreated cells (Fig 2. e xperiment c) .

ICAM-1 express ion on ce ll s p re-treated w ith 25 U / ml of IFN-y for 48 h was augmented even at 200 ng/ m l of SED. increased in a con centration-dependent manner, and plateaued at 600 ng/ml of SEB. Pretreatment of the ce lls w ith IFN-y at 25 U / ml for 24 h instead of 48 h did not augment ICA M-1 expression in the presence of SEB, suggesting t hat re latively long-te rm inc ubation with I FN--y

538 WAI< ITA ET AL

90

..: c RO -1: ..J :r ... 0

..: Q

-!: ..J X 'o .~ · ~

-~ ., "' ~ "' E 0 :I t:

; Ei ,;., 0

..J

sn

1411

120

100

RO

(>0

40

a

b

0

24 4!1 72

Incuhation timc(h)

100 zoo

IFN- y (U/ml)

Figure 1. IFN-y induces HLA-DR molecules on DJM-1 cells. T ime course (11) and concentration dependence (I>) of the induction of HLA -DR. on OJM-·1 ce ll s by LFN- y were dete rmined by Aow-cytomctric analys is. Mean flu o rescence inwnsity was expressed as the tn can va lu e ± SD fron1

three independent experiments . 25 U/ ml of lFN-y was used in (a).

was necessa ry fo r SEll-re lated augm entatio n of IC AM-1 expres­

sion . N either diphthe ria tox in no r pe rtussis toxin at concentratio ns

betwee n 200 and 600 ng/ml afFec ted IC AM-1 expressio n on the

LFN-y-un treated and IFN - y-treated DJM-1 cells (data not sho wn).

In no rmal hum an ke ratinocytes , neithe r IC AM-1 no r HLA-DR

was constitu tively expressed (Fig 3). SEll alon e did no t induce

I CAM- t expression . ICA M-1 expressio n induced in ce ll s incubated

w ith 100 U/ml ofl FN- y fo r 48 h w as significantl y upregulated wi th

exposure to 800 ng/ml o f SEll fo r an addi tio nal 48 h (p < 0.01 ) .

Hig he r concentratio ns o f IFN-y and SEB w ere needed to induce

and upregula te ICAM-1 ex pressio n in no rm al human ke ratinocytes

than in OJ M - 1 cells.

T H E JOU R NAL OF INVESTIGATI VE DERMATOLOGY

lncub:a tlon {48 hl

'" KC-St"M KC-Snt lfN·'f( B U/ml) IFN-yl!5 U/ ml)

lnd

KC·SfM SE IJ t600 ngl mll KC-S rM S£ 1)( 600 ng/ mll

Figure 2. SEB upregulates ICAM-1 expression on IFN-y-treated DJM-1 cells. The ce ll number is shown on the ordinate and mean fluo­rescence in tensity (log scal e) on the abscissa. T he ce ll s we re incubated with lFN-y (25 U/ ml ) or KC-SFM, washed. and then incubated with SEI3 (600 ng/ml) or KC-SFM. T he shaded nrcn represents an iso rype-matchcd control.

SEB Binds to IFN-y-Treatcd DJM-1 Cells To elu cidate the

binding of SEll to ke ratinocytes in associati o n w ith HLA-DR

m o lecules, we d o uble stai11 ed DJM-·1 cdls wi th biotinylated SEll

and PE-anti-H.LA-D R M oA b (L243) . W e st;tincd the cells with

biotin ylated SE13 be fore fi xa tio n to preserve binding sites for SEll

and th en with PE- anti-HLA-DR M oAb (L243 ) . T his procedure

w as used beca use a pre vio us repo rt [20] and o ur p reliminary study

showed th at L243 bl ocks SEB bin d in g to HLA-D R m o lecules an d

because pri o r inc ubatio n with bi o tinylated SEB did not di sturb

an ti- HLA-DR M oAb binding to DJM-1 cells. N o immuno Auo res-

"' v

g L. 0

95

H5

:I 75 c

~ E 0, 0

..J

65

~ E~ l' u

~ ~ ;: 0 -u ~

~ ~ ::>: .~ ~ ... ~ ;..-, "' "" o u ~ ~

z. ::::. ::.:

li xp b Exp c Ex p d

"E "E "E

~ 3 3 1-0 0

~ 0 ~ 0

~ >- 00 ;;; "' i . z. "' "' ~ ... "' - +

Figure 3. SEB upregulates ICAM-1 expression on IFN-y-treated normal human keratinocytes. Experiments n to d were done as in Fig 2. Mean Au o rcscc ncc intensity was expressed as the rn can vnluc :!: SO !Tom

three independent cxperimcms.

VOL 105. NO. 4 OCTOBER 1995 SEll UPREG ULATES KERAT INO YTE ICAM-1 EX PRESS ION 539

. I

Figure 4 . SEB and HLA-DR are identically distributed on DJM- 1 cells. DJM-1 cell s treated wi th (c, rl) or without (a,b) IFN-y (600 ng/ ml) fo r 48 h ' were sequenti all y double stained w ith biotin ylatcd SEB plu s FITC-strcptav idin and then PE-anti-HLA-DR MoAb. Neither SEB (a) no r HLA-DR (b) was

detec te d on untreated cells. SEB (c) and HLA-DR (d) were detected on IFN-y-trcated DJM-1 ce lls. MHC class I molecules on IFN- y-treated DJM-1 cell s were v isuali zed by anti-MHC class I MoAb p lus PE-conjugatcd anti-mo use lgG (e). Scale bnr, 20 IJ.m (a,b,e) or 10 IJ.m (c,rl).

cence for SEB and HLA-DR molecules was observed in the untrea ted DJM-1 cells (Fig 4a,b). On the other hand, both SEB and HLA-DR were observed in lFN--y-trea ted DJM-1 cells (Fig 4c,d). Staining patterns w ith the tv.ro molecules were indistin guishable (Figs 4c,d) . ln contrast, an irre levant mouse MoAb, anti-MH C class I, showed an in tense and continuous staining pattern on

, fFN - -y-created DJM- 1 cells (Fig 4e), appa rently distinct fi·om that ofSEB and anti-HLA-DR MoAb. These results indicated the close, if not identical, binding sites o f SEB and anti-HLA-DR MoAb.

We then compared the expression of ICAM-1 and HLA-DR as well as SEB binding to DJM-1 cells by Aow cytom etry (Fig 5). T he

• mean fluorescence intensity fo r SEB, H LA-DR, and ICAM-1 was significantly increased in the IFN--y-treated DJM-1 cells w hen compared with the untrea ted DJM-1 cells. Because SEB is ab le to

1 bind not only HLA-DR, but also other MHC-class II molecules such a s HLA-DQ [20), and because lFN--y induces both molecules on keratinocytes [2 '1 ), we used not only anti-HLA-DR MoAb (L243) but also anti-MHC class II MoAb (CR3/43) as blocking

antibodies for SEB binding to DJM-1 cells. Although both MoAbs signifi cantly inhibited the SEB binding (p < 0.01) and C R 3/43 inhibited SEB binding m ore e lfe ctively than L243, they did not inhibit SEB binding completely.

Because SEB is reported to stimul ate keratinocytes to produce cytokines such as T NF-a [5). which is capable of indu cin g ICAM-1 express ion on th ese cell s, th e efrect of keratinocyte­deri ved solubl e factors on ICAM-1 e xpression was examined . First, DJM-1 cells t reated with 10 ng/ ml o f TN F- a n either e xpressed H LA-DR no r had an affi nity for SEB desp ite the high level of expression of ICA M-1 (Fig 5) . In addition , SEB could not augment ICAM-1 express io n on TNF-a -pretreated kerati ­nocytes (data not shown). Second. the in cubation of eith er IFN-y-treated o r untreated DJ M-1 cell s w ith conditione d me­dium for up to 48 h did not upregulate ICA M-1 exp ress io n . These fi ndings confirmed that H LA-DR molecu les were critica l fo r SEB bi.ndin g to keratinocytcs and SEB-mcdiated upregula­tion of ICA M-1 express ion o n kcra tin ocytes and also that

540 WAKITA ET AL

c.: '? ~

"' < < ~ ..J

~ "' :I:

IIIII • 0 90

150

1 10

-~ 140 c "' -~ ... 130 u 90 c

5 70 ... 120 0 = r:: c 70

110

" ... E 0. 100 0

..J

50 50 90

blocking mAb

Figure 5. SEB binds to DJM-1 cells pre-incubated with IFN-y but not TNF-a. After 48 h incubation with 50 U/ ml of lFN--y or 10 ng/ml of TN F-a, cell suspensions from the same we ll were divided in to two samples. O ne sample was sta ined with biotinylatcd SEB (800 ng/ml) plus FITC streptavidin and the other was sta ined with FITC-anti-I CAM-1 MoAb and PE--anti-1-lLA-DR. MoAb. Some IFN- -y-treated cells were incubated with anti-H.LA-DR MoAb, L243 or anti-M J-I C-- class II MoAb, C R3/43 before sta i11ing. Mean flu orescence intensity was expressed as the 111 ean value :t S.D from three independent experiments.

so lubl e factors present in th e conditioned m edium were not responsibl e for the upregulati on.

SEB and Anti-MHC Class II MoAb Mobilize Intracellular Free Calcium We examined the effect of SEB on mobilizatio n of intracellular calcium in DJM-1 cells, whjch was based on the idea that A23187, well known to increase intracellular free c al cium, stron gly augm ents lCA M-1 expression on keratinocytes [1 9]. In the cells that expressed HLA-DR mol ecules after 48 h exposure to IFN-y, the additi on of SEB at 600 ng/ ml or anti-MHC-class JJ M oAb at 10 fLg/ml led to a transient in crease in the l evel of cytoplasmic fi·ee ca lcium (Fig 6a.,IJ, respectively). The levels of cytoplasmic free calcium in untreated DJM-l cells were not alte red by SEB (Fig 6a) or anti-MHC class IT M o Ab (data not show!l). Five minutes after calcium m obilization by the additio n of anti-MHC­class II MoA b, th e IFN-y-pretreated cells did not respo nd furth er to the addition ofSEB (Fig 6/1). Pt;o r calcium mobilization by SEB also inhibited the effect of anti-MHC-class II MoAb (Fig 6c). In addition , anti-MH C class I MoAb could neith er mobiljze i~1tracel­

lular fi:ee calcium nor di sturb calcium mobiliza tion induced by SEB (Fig 6d) . These results suggested th at SEB triggered the calcium mobilization, possibl y tlin binding to MHC-class II m olecules.

ICAM-1 mRNA Is Induced by SEB Because ICAM-1 mol e­cules were constitutively expressed on DJM-1 cells, LCAM-1 mRNA was detected by reverse transcriptase-polymcrase chain reaction (data not shown) in all samples including non-treated, IFN-y-treatcd (both short in cubation for 4 h and long incubation for 48 h fo llowed by medium alone for 52 h) , and IFN-y- and SEB-trea ted DJM-1 cell s. T hus, it is difficult to compare mRNA levels between SEB-treated and non- treated DJM-1 cells by thi s method . Next, we used normal, ICAM-1-negative keratinocytes to investigate w hether SEB induced JCAM-1 mRNA. As shown in Fig 7, mRNA for ICAM-1 was detected in IFN-y-pretreated/ SEE-stimulated (incubated with lFN-y for 48 h and then left untreated for 48 h fo llowed by the incubation with SED for 4 h) ce iJ s (Fig 7 II), but not in IFN-y-pretrea ted /SEB-non-stimulated

THE JOURNAL OF INVESTIGATIVE DERMATOLOG\'

(incuba ted with TFN-y for 48 h and then left untreated for 52 h) cells (Fig 7c) o r contro l cells (Fig 7a). In contrast, mRNA for ICAM-1 was detected in keratinocytes cultured with IFN-y for 4 h (Fig 7d). mRN A for /3-actin was detected at the comparative !eve) in all groups (Fig 7 L~ /r) .

DISCUSSIO N

Our results demonstrate that SEB m odula tes the expression of JC AM-1 by bindin g to MHC class II molecules on kerati.nocytes because 1) fluorescence images of the binding sites were virtuaiJ;. identica l for SEB and anti-HLA-DR MoAb, 2) SEB mobilized intracellular free calcium in HLA-DR-positivc but not in HLA­DR-negative keratinocytes, and 3) an ti-MHC class II M oAb partiaiJy prevented bindjng of SEB to l-ILA-OR-positive keratino. cytes, and completely blocked calcium m obilization by SED. Be­cause anti-HLA-DR MoAb (L243) and anti-MHC class II M oAb (C R3 /43 ) bind to /3-chai.n ofHLA-DR, and because the amino acid sequence of the a - chajn is critica l fo r SED binding to HLA-DR.. [22], partial inhibition of SEB binding by these MoAbs m ay h ave been due to steric hindrance but probably not to topographic overlappin g of the binding sites. In addition, it is plausible that SEB binds to IFN-y-treated transformed keratinocytes that had been pretreated with L243, whjch bl ocks HLA-DR molecules only, beca use of the binding ability ofSEB fo r HLA-DQ mol ecules other than DR [20] and the indu ction of both mol ecules on keratinocytes by IFN-)' [21. ]. In contrast, MoA b obtained from clone C R 3/43 reacts with the /3-chai.n of all products of the gene subregions DP, DQ, and DR. T hus, C R 3/43 blocks SED binding to DJM-1 cells more elti ciently than does L243.

Cross-linking o fSEB by other reagents such as anti-SEB antibody was not necessary to mobilize calcium . T he present data are inconsistent with the previous finding that calcium mobilizatiot1

occurs on ly after ct·oss-lin.king of SEB on the ce ll surface [23- 25]. However, even without cross-Jjnking, SEB activates protein ty­rosine kinase and phosphoinositide- specific phospholipase C [26 -28], which cleaves phosphoinositide into diacylg lycerol [29) and inositol triphosphate, which in turn re leases free calcium from the endoplasmjc reticulum [30] . These observa tions suggest that mono­valent binding of SEB to HLA-DR molecules induces calciun1

m obilization and signal transduction. Alteration of adhesion molecul es expression by SEB binding to

tbe MH C class ll molecu le has also been previously shown in B cells. For exa mple, SEB induces B cell-B cell adhesion [31) tha t is caused by quaLitative changes of lymphocyte function-associated antigcn-l molecule expression without upregulation of JCAM-1 molecules. Thus, express ion of adhesion m olecules is differentially regulated in kerat.inocytes and B cell s 11it1 the binding of SED to MHC class 11 molecules .

At least three mechanisms could expl ain SED-mediated upregu­lation ofiCAM- 1 expression in the IFN-y-treatcd DJM-l cells: 1) SEB binds to I-ILA-D.R. molecules and transduces signals that directl y upregulate synth esis of lCA M- 1 molecules, 2) SEB protect ICA M-1 m olecul es on the cell surface from degradation or shed­ding, and 3) SEB stimulates keratinocytes to secrete soluble fac tors, such as T N F-a, that upregul ate ICA M- 1 expression in an autocrine fashion. T he first mechanism is most plausible beca use ICAM-1 mRNA was detected 4 h after the addition of SEB, during which time T N F- a was not de tec ted by enzyme-linked immunosorbent assay in culture medium (unpubli shed observation). In addition, conditioned medium that might contain biologicall y active mole­cules did not augment ICAM-1 expression on IFN-y-treated keratinocytcs. This indica ted that under our experim ental condi­tio ns , soluble molecules were not responsible fo r upregulation of IC AM-1 expression on IFN-y-treated keratinocytes. The second mechanism is also possible in the present study and not mutu ally I

exclusive with the first one. Levels of soluble l CA M-1 in culture medium could be measured to test this.

One of the critical immunologic and inflammatory events within th e epidermis is binding of infiltrating leukocytes, such as T cells. monocytes, and possibl y eosinophils , to epidermal keratinocytes

VOL 105, NO. 4 OCTOBER 1995

~ 1.6

~ a ~ ~ ~

~ I.S g ;:::

.... 1.4 0

~

"-, I.J

~- "' -~ "'

I.Z <J.

~

~ > 1.1 0 -.;

"

1.6 c

~ ~

~ I.S

5 1.4 <= .... 0

~ 1.3

"' 'o 1.2

::-·~

1.1 .~ ~

1.0 > 0 o;

"

100 Time(scc)

Tint(.' (sec)

SEI3 U I'R.EGULATES KEI'-ATINOCYTE ICA M-1 EX PRESSION 541

2 00

Figure 6. SEB and anti-MHC class II MoAb mobilize cytoplasmic free calcimn. Fura-2-loaded DJM-1 cell s were either un treated (closed squares) or pre-treated with 50 U / ml of IFN--y for 48 h (op<:u squares) and then incubated with 600 ng/ml of SEB (n) . JFN--y-pretrcatcd cell s were stimulated with 10 11-g/ml of anti-MHC class II MoAb (b), 600 ng/ml of SEB (c) . o r 10 JJ.g/ m l of anti-MH C class I MoAb (d) and exposed to SEB (/J,d) or anti-MH C class Il MoAb (c) 5 min later.

Figure 7. SEB induces ICAM- 1 mRNA in IFN--y-trcatcd kcratino­cytcs . Normal human kcratinocytes were either untreated (n,e) or treated (b,cJ,g) for 48 h w ith 100 U / m l of IFN--y. Afte r a 48-h incubation witho ut stimul ant, medium was replaced with fres h KC-SFM contain.ing either no stimulants (n ,c,e,,~) or 1 JJ.glml of S£13 (/!J) fo r an additional 4 h. Normal human keratinocytcs without pretre<l tment were incubated with 100 U / ml of IFN--y for 4 h as a control (rl,h). Tota l RNA was isola ted and 1 JJ.g was subjected to reverse transcripta sc-po lyn1crasc chain rc:.tction uslng prirn crs for ICAM-1 (n-d) or /3- actin ( <~ h) .

through the interaction of ICAM-1 and lymphocyte fun ction­associated antigen-1 [32,33]. In 11itro studies have shown that such interactioH is a prerequisite for MHC class IT-bearing keratinocytes to serve as <J ccessory cells in the T - cell response to ba cterial superan tigens [4]. T he present findings indicate that SEB f.1cilitates the accessory cell function of keratinocytes by indu cing and up­regu lating ICAM-1 express ion . Superantigens fi·om superficial ly colonized S tnpll ylococcus nureus may penetrate viable epidermal layers through the disrupted skin barrier, augmenting and prolong­ing inflammation in va rious cutaneous diseases by upregul ating ICAM-1 expression of keratinocytcs. Furthermore, chronic expo­sure to exogenous superan tigens induces profound deletion of reactive mature T cells [34], thus downregulating cutaneous in­flammation.

VVc tltnuk !VIs. Kcikt> Sugn yn _fin It er teclmiw l nssisttwcc. 'Fhis work 111ns Sll)JJ1011l.'d

by grnurs Jimu the iV/iuisll)' of Etlucnliou, Scirurc nud C.llturr ".f Jnpnu nud tlu­Miuisll)' of l¥eljnrc nud Hcn /1h of]npnu.

R EFERE N CES .

1. Lcydc nJJ, Marples R.R .. Kligmollt AM : Swplt)'lowrm s nun·11s in da! lesions of;~topic dermatitis. Ur / Dcrwnrol 90:525-530. 1974

2. Aly R. Maihacl; 1-11 . Shincficld Hll: Mi crobial Ao ra of :Hopic dermatitis. A rd• Dm 11 ntol 11 3:780-782. 1977

542 WAKITA ET AL

3. M arples Jl..R ... H eaton C L. K ligman AM : Stnplt )l/oa,cws n u rt' IIS in psoriasis. Arr/1 Ocmwtol I 07:568-570 , 1973

<1. Nicko lo ff Bj , Mitra R S, Green J , Z hcng XG, Shimi zu Y, T ho mpson C . Turka

LA: Accessory cell fun ction of kcratinocyLcs fo r supcran tigcns. J fllllllltt/0/

150:2 148 - 2 159, 1993 5. Tokura Y . Yngij, O'Malley M. Lewis JM , Tnkignwn M , Ede lson R.L. Tigclnnr

lli: Supc rantig:cnic staphylococcal cxotoxins induce T -ce ll pro li fe ration in th e

presence of Langcrhans cells or class 11-bcaring keratinocytcs and stimulate kcrati nocytcs co produce T -cell-activa ting cytokines. J flwc.H Dermnwl 102:3 I-

38, 1994 (j_ C hat.il:l T. Gcha R.S: Signal transduction by microbial supcnmd gcns via MH C

cl"ss I! molecules . 1111111111 /0 I Ret! 13 1 :43-59, 1993 7. Trcde N , Gch ~1 R S, C hatil<a T: T ranscrip tional activation of m onokinc genes by

MH C class II ligands. ) illll llllllo/14 6:23 10-23 15, 1991 H. Springer TA: Adhesion receptors of the immune system. N nture 346:425- 434 ,

1990 9. Ku pper TS: Mech anism s o f cutaneous inflammatio n. Intcr<~ ction between epi­

dermal cyrokincs, adhesio n molecules, and leukocytes. Jl rclt Dcr/1/atol 125:

l 4UG-14 12, 1989 10. Wnkita H . Sakamoto T. T okura Y, Takigawa M: E-sclcctin and vascular cell

adhesion molcculc-1 as cri tical adhesion molecules for infil tra tion ofT lym­

phocytes and eosinophils in atopic dermatitis. J C utn11 l'ntlllll 21:33- 39. ·1994 II . Wakita 1-1. Takigawa M : E-sclccrin nnd vascular cell adhesion m olcculc-1 arc

critical fo r initial traffi cking of helper-indu cer/ memory T cells in psoriatic

plaques. Arclt Dcmtntol 130:457-463, 1994 12. GriffHhs CEM: C utaneo us le ukocyte trafli ckin g and psoriasis. A rch Dt:nn awl

130:494- 499 , 1994 13. Corne lius LA, Taylor JT. Degitz K, Li L-J. Lawley TJ , Caughman SW: A 5'

portion of the ICA M- 1 gene co nfers ti ssu c-sp cci.fic diff"crcntia l expression

levels :1 nd cytokin c responsive ness. J lu vcst Dt.·rlll nto/1 00:753-758. 1993 14 . Kitaj ima Y. Ino ue S. Yaoita 1-1: Elfccts of pemphigus anti body o n the regene ra­

tion of cell-ce ll contac t in keratinocytc cul tures grown in low to normal ca ++

concentration . J ltwc.st Dermatol 89: 167- .1 71. 1987

15. Ca ughman SW, Lian-Jic L. Degitz K: C haracteri za tion and fun ctiona.l analys is of intc rfcron-y-in d uccd inte rcellula r adhesio n molecule- I expressio n in human

keratinocytes and A-43 1 cell s. J 11111cs r Dcn11ntol 94 :22S- 2GS, 1990

I (J. W inisky AI'. Foster CA: IC AM-·1 expressio n in spontaneously transform ed human kcrati nocytc cell line: chara cterization by a simple ce ll-ELI SA assay. J lll ttl:st Dcnll n/C/I 99 :48-52 , 1992

17. Brausten NS. Weber DA , Wo ng X-C, Long EO, Karp D: Sequen ces in bo th class II rn ;ajor histocompatibility compl ex :1 and b chain s contribute ro the bindin g o f

the supcrantigcn tox ic shock syndrom e toxin 1. J Exp J\ltcd 175: ·t 30 L-1 305,

1992 18. C homczynskj P. Sacchi N: Single-step method of R.NA isola tion by acid

guanidinium th iocyanatc-pllcno l-chlo ro form extractio n. Al/fll Bim:hcm 162:

156-159, I %7 19. Griffith CEM. Esrnnnn J . Fishe r GJ. Voorhees JJ , Nickoloff l3J: Differential

THE J OURNAL OF INVESTIGATIVE DERMATOLOGY

m od ulation of kcra tinocyrc intercellular adhesion nwlcculc-1 expression by gamma intcd "'c ro u and p h o rbo l c!itcr: evidence fo r in vo lvem en t of protein

kinase C signa l transduction. Brj Oermntct /1 22 :333- 342, 1990

20. Scholl PR, Diez A. Gcha R S: Staphylococca l cnrcro toxin l3 and tox ic shock syndrome tox in- I bind ro distinct sites on HLA-DR :md I-ILA-DQ m olecules.

J illll/11111(1/ 143:2583-2588. 1989 2 1. G riffath s CEM, Voorhees JJ . N ick o loff BJ : Gamma interfe ro n ind uces dilferem

kcratinocyrc expression of I-lLA-D R . DQ. and interce llul ar adhesion mole­cul e- I (ICA M- 1) antigens. Brj Oermttw/1 20: 1- 8, 1989

22. Thibodeau J, Labrecque N , Denis F, Huber BT, Sckoly R - P: Binding sites for bacteria l nnd endogenous rc rrovi.ra l supcranti g:cns can be dissociated on mitior

histocompati bili ty complex class llmolccules .J Exp Mcd 179: I 029 -1 034, 1994

23. Lone PJ , Me on nell FM. Schievcn Gl., C lark EA. Ledbetter JA: T he role of class

II molecules in hum:m B cell nc tivari on. J Immlll tol 144:3684 - 3692. 1990

24. Mooney NA, Grillo t-Courvolin C, Hi vroz C, Ju LY, C harron 0 : Early biochem­

ical events after MJ-lC cbss 11-mcdiatcd signaling on hum an B lymphocytes. j l11111111110l H 5:2070- 2076, 1990

25. Damaj D, Mournd W , Naccachc PH: Supcrantigcn-mcdiatcd human mono­

cyte- T lymphocyte int.crncti ons are associated with MH C class II -, TCIV C D 3- and C D4-dcpendcn t mobilization of ca lcium in monocyrcs. j lmmrmol

149:1497-1503, 1992 26. Scholl PR, T rcde NS . C hnLiln T, Geha RS: R o le of protein tyros in e phosphor­

ylntion in monokine induction by staphylococca l superantig:cn toxic shock

syndrome toxin-1. J 11111111/tw l 148:223 7- 224 1, ·1992 27. Gchn R S, Trcde NS, Scholl PR, C hatilo T: Signal tr"nsdu ction by microbial

superanrigens in human peripheral blood monocyrcs (abstr) . FASEB J G:A 1076, 1992

28. Matsuyam a S, Koidc Y , Yoshid a TO: l-ILA class II mol ecule-mediated signal transduction mechanism responsibl e for the expressio n of interlcuk.in-1 {3 and

tumor necrosis f.1c to r- cr genes induced by a staphylococca l superantigen. Eur J 1111111/lllil l 23 :3194 -3 202. 1993

29. N ishizuka Y: Studies and perspectives of protein kinase C. Science 233:305- 312,

•J986 30. Berridge MJ , Irvine Rl:: : Inositol phosplmtcs and cell signaling. Ntr ttii'C 34 '1 :197-

205, '1989 31 . Mournd W , Gehn R S, C hatila T: Engagement of m:~ or histocompatibility

compl ex class I I molecul es induces sustain ed. lympl1 ocyte fun ction-associated molecule 1-depcndcnt cell ad hes ion . ) Exp Ml'd 172 :1 5 13-1 516. 1.990

32. Kupper TS: Mechanisms of cutaneous inflammation. Interaction between epi­dermnl cytokines, adhesio n molecules, and leukocytes. Arcb Dermmol '1 25: 1406-141 2, 1989

33. Wa lsh LJ . Murph y GF: .R o le of adlu.:sion molecules in cutnncous inflammation and neoplasia. ) C 1111111 Pntlwl 19:161- 17 1, 1992

34. McCormack J E, Collohon J E, Kappler J. Marrock PC : Profound deletion of mature T ce lls in vivo b y chronic exposure to exogenous supcrantigens. J lllllll/111 01 150:3 785- 3 792. 1993