In vivo relevance of CD30 in atopic dermatitis

Allergy 1997: 52:1063-1070Printed in UK - oil rights reserved

Copvrighl © Munksgaard 1997

ALLERGYISSN 0105-4538

In vivo relevance of CD30 in atopicdermatitis

Caproni M, Bianchi B, D'Elios M, De Carli M, Atnedei A, Fabbri P. In vivorelevance of CD30 in atopic dermatitis.Allergy 1997: 52: 1063-1070. © Munksgaard 1997.

CD30 expression was evaluated by immutTohistochemistry in lesional skinbiopsies of eight patients with active atopic dertnatitis (AD) and threepatients with allergic contact (niekel-induced) dertnatitis (ACD). CD30expressioti was also assessed in a large panel of CD4 -i- and CD8 -i- T-cellelones generated frotn the skin biopsies of four patients with AD. Finally,the levels of soluble CD30 (sCD30) were measured in the serurn of 41patients with AD, 19 patients with ACD, and 60 healthy eontrols. In allspecimens of lesional AD skin, where the great rnajority of infiltrating cellswere CD4-H T cells, remarkable numbers of cells were CD30-H, whereasvirtually no CD30 -t- cells were found in the skin of patients with ACD. InCD4H- T-cell clones generated from the lesional AD skin, tnost of whichproduced both interleukin (IL)-4 and interferon-gatiima (IFN-y) (ThO-likecells) or IL-4 and IL-5, but not IFN-y (Tli2-like cells), CD30 expressiondirectly correlated with the ability to produce IL-4 and IL-5, but wasinversely related to lFN-y production. High levels of sCD30 (correlated withdisease activity: /• = 0.618) were detected in the serum of most AD patients,whereas there was no increase of sCD30 levels in the serum of patients withACD. Tliese data support the view that ThO/Th2-type responses predominatein the skin of patients with AD and suggest that the presence of CD30+ Tcells in tissues and/or increased levels of sCD30 in biologic fluids areindicative of Tli2-dotTiinated responses.

M. Caproni\ B. Bianchi\M. M. D'Elios^ M. De Carli^A, AmedeP, P. Fabbri'Department of Dermatology and 'Institute ofInternal Medicine and Immunoallergology,University of Florence, Florence, Italy

Keywords: atopic dermatitis; human Thl/Th2cells; interleukin-4; membrane CD30; solubleCD30.

Prot. Paolo FabbriIstituto di Clinica DermatologicaVia della Pergola 5850121 FirenzeItaly

Accepted for publication 20 May 1997

Atopic dermatitis (AD), a chronic inflamtiiatoryskin disease characterized by severe itching, age-dependent skin manifestations, and a fluctuatingclinical course, represents the cutaneous tiianifes-tation of atopic allergy. Most patients with ADhave elevated serutn IgE levels, with specificity tovarious cornmon environmental allergens to whichthey also show imtnediate-type skiti reactions (1,2). Lesiotial AD skiti is characterized histologicallyby derrnal perivascular infiltrates of mononuclearcells, mainly consisting of CD4-I- T helper (Th)cells (3, 4) and eosinophils (5). The predominanceof CD4 + Th cells in the AD skin infiltrates hasraised the question of a possible pathogenic role ofthese cells.

CD4-)- Th cells can be classified into differenttypes by their cytokine profile. Type 1 (Thl) cellsproduce interferon-gatiirna (IFN-y), interleukin(IL)-2, and tumor necrosis factor-beta (TNF-P),and promote the production of opsonizing andcomplement-fixing antibodies, macrophage activa-tion, and delayed-type hypersensitivity (DTH),

Type 2 (Tli2) cells produce IL-4, IL-5, IL-9, IL-10,and IL-13 and provide optimal help for humoralresponses, including IgE isotype switching, andmast-cell and eosinophil growth and differentia-tion. In the absence of clear polarizing signals,CD4-I- T-cell subsets with a tnore heterogeneousprofile of cytokine production than Till or Th2 aredesignated TliO and mediate intertnediate effects,depending upon the ratio of lymphokines producedas well as the nature of the responding cells (6, 7).

In vitro cloning studies have provided valuableinformation about the functional nature of skin-infiltrating CD4 + T cells in patients with AD. Mostof these cells express a Tli2-like phenotype, andsome of them are specific to common environmen-tal allergens (8-11). However, other reports sug-gest a potential role for IFN-y in AD, IFN-y mRNAexpression, and not IL-4 mRNA, beitig increasedin chrotiic AD lesional skin (12, 13), Moreover,chronic lesions seem to have significantly fewerIL-4 mRNA-expressing cells than acute AD skinlesions, but significantly more IL-5 tiiRNA (14).

1063

Caproni et al.

It has recently been shown that membraneexpression of CD30, a member of the TNF-recep-tor superfamily (15), is preferentially associatedwith the production of Th2-type cytokines in bothCD4-h and CD8+ human T-cell clones (16, 17). Inaddition, costimulation of CD4 -i- T cells, via CD30,has been found to favor the in vitro developmentof antigen-specific T cells into the Th2-hke profile,whereas blockade of CD30-CD30 ligand inter-action inhibited the development of Th2-like cells(18).

In apparent contrast to this selective nature ofCD30 expression, the ability of activated CD30-t-T cells to produce high levels of both IFN-y andIL-5 has also been reported, suggesting that CD30expression is related not only to Tli2-like cells (19).Finally, the extracellular portion of CD30, which isproteolytically cleaved to produce an 88-kDa sol-uble form of the molecule (sCD30), was found tobe released by CD30-expressing T cells both invitro and in vivo (16, 17, 20, 21).

This study investigated the expression of CD30by infiltrating T cells present in the lesional skin ofpatients with active, acute AD. In addition, thesCD30 levels in the serum of AD patients were alsoassessed. Patients with nickel-induced allergic con-tact dermatitis (ACD), in which Thl-like cells havebeen thought to play a pathogenic role (22, 23),provided "disease control" skin specimens andsera. Results showed remarkable CD30 expressionin the lesional skin, and high levels of sCD30 in theserutn of patients with AD. In contrast, neitherCD30 -I- T cells nor elevated levels of sCD30 werefound in patients suffering from nickel-inducedACD.

and two women; mean age 36 years, range 24-56)in three different phases of disease, i.e., acute,subacute, and chronic. At the time of the study,none of the patients were receiving systemic corti-costeroids or other immunosuppressive drugs, butmost patients with AD were using topical steroidson an irregular basis.

For measurement of sCD30 levels, sera wereobtained from a total number of 41 patients withacute, active AD (disease activity score rangingfrom 15 to 84) and 19 nonatopic patients with activenickel-induced CD. Sixty healthy donors (32 menand 28 women; mean age 31 years) with normalserum IgE levels and no history of atopy providedcontrol sera.

Reagents

Monoclonal antibodies (mAb) used for immuno-histochemistry on cryostat sections of skin biopsiesincluded anti-CD3, anti-CD4, anti-CD8, and anti-CD30 (BerH2) (Dako, Glostrup, Denmark). Anti-CD30 agonist M44 and M61 niAbs were preparedby Imrnunex (Seattle, WA, USA), and their activityis detailed elsewhere (18). Recombinant humanIL-2 was a kind gift of Eurocetus (Milan, Italy),and recombinant human IL-5 was purchased fromAmgen Biologicals (Tliousand Oaks, CA, USA).Phytohemagglutinin (PHA) was purchased fromGibco Laboratories (Grand Island, NY, USA).Phycoerythrin-conjugated anti-CD4 and anti-CD8 mAb were purchased from Becton Dickinson(Mountain View, CA, USA); FITC-conjugatedanti-CD30 mAb (BerH2) was purchased fromDako.

Material and methodsPatients

For the immunohistochemical study, upon in-formed consent, punch biopsies (6 mm), weretaken from acute erythetnatous lesions of eightpatients (seven men and one woman; mean age 28years, range 19-44), in whom active AD had beendiagnosed according to the diagnostic criteria ofHanifin & Rajka (24). All patients showed elevatedserum IgE levels and positive skin tests to three ormore allergens, as confirmed by the detection ofallergen-specific IgE in the serum. Clinical diseaseseverity was assessed according to a previouslydescribed system (25), with a total body diseaseactivity maximum score of 110. At the time of thestudy, disease activity scores ranged frorn 10 to 77(mean 42). Upon informed consent, punch biopsieswere also taken from lesional skin of three non-atopic patients with nickel-induced ACD (one man

Immunohistochemistry

Skin biopsy specimens were immediately frozen inliquid nitrogen, and stored at -80°C until used.Sequential cryostat sections (5 (xm) were cut fromeach specimen, air-dried, and fixed in 100% ace-tone. Sections were washed in TRIS-bufferedsaline, coated with nonimmune goat serum for 20min, washed again, and then incubated with atiti-CD3, anti-CD4, anti-CD8, or the different anti-CD30 mouse mAbs (primary antibodies, allbelonging to the IgGl isotype) for 60 min. Afterwashing for 10 min, sections were incubated for40 min with rabbit anti-mouse Ig antiserum (RAM,1 : 30) (Dako) and then processed with alkalinephosphatase anti-alkaline phosphatase cornplex(APAAP 1 : 50) (Labometrics, Milan, Italy). Foretihancement of the labeling intensity, incubationswith RAM and APAAP were repeated for anadditional cycle. Binding of complex was revealedby fuchsin as chromogenic substrate (Merck,

1064

CD30 in atopic dermatitis

Darmstadt, Germany). Sections were then counter-stained with Mayer's hematoxylin, cleared, andmounted in top resin (Pabish, Milati, Italy).Negative control sections were iticubated with ananti-human IgG2 mouse IgGl mAb as primaryantibody. Well-stained cells were coutited by twoindependent and "blinded" observers, who countedskin-infiltratitig cells imtnediately below and abovethe dertiioepidermal junction in at least threeadjacent high-power fields (x250).

Generation of T-eell lines and clones

To favor the expansion of activated T cells presentin lesional skin, fragments of bioptic specimens offour (Table 1: S.A., R.M., B.V., and D.I.) AD patientswere cultured for 9 days in RPMI 1640 mediurnsupplemented with 2 mM L-glutamine, 2x10"' M2-ME, 3% heat-itiactivated hutiian serum, and 10%fetal calf serutn (cotnplete tnediutn). Recombinant(r)IL-2 (30 U/ml) was added on days 0 and 6. Tissuespecimens were then disrupted, and the viable T-cellblasts obtained were seeded under litiiititig dilutionconditions (0.3 cells/well) in round-bottomed micro-well plates containing lOVwell irradiated peripheralblood rnotionuclear cells (as feeder cells), PHA(0.5% v/v), and IL-2 (20 U/ml), as detailed else-where (26). High-mol.-wt. DNA was prepared fromT-cell clones derived from lesional skin of patient 3and assayed for diversity of T-cell receptor generearrangements by Southern blot analysis, asdescribed elsewhere (27).

Characterization of surface markers and cytokineprofile of T-cell clones

Cell surface marker analysis of T-cell cloneswas performed on a Cytoron Absolute cytofluoro-nieter (Ortho Pharmaceuticals) with PE-labeledanti-CD3, anti-CD4, or anti-CD8 tiiAb (BectonDickinson, San Jose, CA, USA) and FITC-labeledanti-CD30 (BerH2 Dako, Glostrup, Denmark),as detailed elsewhere (17, 18). For itiduction ofcytokine production by skin-derived T-cell clonesin response to PHA, T-cell blasts were resuspendedat 10''/7 ml complete medium and cultured in thepresence of PHA (1% v/v). After 48 h, culturesupernatants were collected, filtered, and stored inahquots at -70°C until used. Supernatants wereassayed for their IFN-y, TNF-p, IL-4, and IL-5content. Tlie quantitative detertninations of IFN-yor TNF-P and IL-4 were perforrned by cotntnercialassays (BioSource International, Inc., Catnarillo,CA, USA, or Quantikine R & D Systems,Minneapolis, MN, USA, respectively). For meas-urement of IL-5, the rnurine LyH7.B13 cell line wasused as a source of indicator cells, as detailed

Table 1. CD4, CD8, and CD30 expression by skin-infiltrating T cells in patients with

atopic dermatitis |AD) and allergic contact dermatitis lACD)

Diagnosis

Atopic dermatitis

Patients

S.A.

R.M.

B.V.

D.I.S.F.A.A.B.P.C.S.Contact dermatitis

Patients

LR. (acute)

M.S. (siibacute)

B.E. (chrcnicl

CD8

27232117213

5029

2815610

Mean number of positive cells*

CD4

937378748922

11097

2617946

CD30

5454756246126965

1133

" Mean cell counts in three fields | x 250).

elsewhere (25), Supernatants showing IFN-y, TNF-p,IL-4, or IL-5 levels 5 SD over the mean levels incontrol supernatants derived from irradiatedfeeder cells alone were regarded as positive. T-cellclones able to produce IFN-y and TNF-p, but notIL-4 or IL-5, were categorized as Till; clones ableto produce IL-4 and/or IL-5, but not IFN-y and/orTNF-p, were categorized as Tli2; and clones produc-ing both IFN-y (and/or TNF-p) and IL-4 or lL-5were categorized as TliO. Tlie clotiing tnethod usedin our study has already been adopted to character-ize the functional profile of T cells infiltratitig thethyroid of patients with Hashitnoto's thyroiditis (27),the retroorbital tissue of patients with Graves'ophthaltnopathy (28), and the bronchial mucosa ofpatients with allergic bronchial asthma (29).

Determination ofsCD30 serum levels

sCD30 serum levels were determined by the com-tnercial sandwich ELISA Dako CD30 (Ki-1 anti-gen, Dako), based on the use of two mAbs reactivewith two different epitopes of the 88-kDa sCD30molecule.

ResultsHigh CD30 expression by infiltrating T cells inlesional skin of AD patients

In all specimetis of lesional skin from AD patients,most of the infiltrating cells were CD3-t- T cellslocated iti large perivascular infiltrates, as wellas scattered throughout the dermis. In all

1065

Caproni et al.

' ' r

Fig. 1. CD4 (A) and CD30 (B) staining in skin biopsy fromconsecutive sections of representative patient with atopic der-matitis (x200). In vivo relevance of CD30 expression can beobserved in papillary areas.

fields examined, CD4 -i- T cells were predominant(Table 1, Fig. lA), with a mean CD4 to CD8 ratioof 4.0. High CD30 expression, detected by anti-CD30 BerH2 mAb, was observed in a remarkableproportion of infiltrating cells, with a prevalentperivascular distribution in the superficial dermis(Fig. IB). Double staining with anti-CD4 and anti-CD30 was not feasible, but the distribution pat-tern of CD30 + cells roughly corresponded to thatof CD4 + cells in consecutive sections. CD30expression by a remarkable proportion of skin-infiltrating T cells was confirmed by staining withtwo other anti-CD30 mAbs (M44 and M67), butwith lower intensity. In the three samples oflesional ACD skin, the degree of T-cell infiltrationwas variable (Table 1, Fig. 2A), the CD4 to CD8ratio being higher (4,6) in the patient with chronicdisease and lower (0.9) in the patient with acutedisease. CD30 + cells were very rare in the skinof both these patients, whereas in the patientwith subacute disease and larger skin infiltrates(Fig. 2A), only a few scattered CD30-I- cells werecounted within the epidermis and dermis(Fig. 2B).

/'

it

'r^^ir^^M^M' M.,h •

B

Fig. 2. CD4 (A) and CD30 (B) staining in skin biopsy fromconsecutive sections of representative patient with contactdermatitis (x200). Only few, scattered CD30+ T cells werefound within epidermis and dermis.

High proportions of T-cell clones derived from ADlesional skin secrete Th2-type cytokines andexpress membrane CD30

T cells recovered from skin fragments of four ADpatients were cloned by limiting dilution accordingto a high efficiency protocol using PHA stimulationin the presence of irradiated feeder cells and IL-2.A total of 189 CD4-F and 20 CD8+ T-cell cloneswere obtained and analyzed for their cytokinesecretion profile in response to PHA. In all cases,most CD4 + clones expressed the ThO profile, and26-38% were Th2, whereas only 10-24% of cloneswere Thl (Table 2). Among the few CD8 clones,45% were Till-like, 35% were TliO-like, and 20%were Th2-like. Southern blot atialysis of TCR-Pgene configuration, performed on the DNA prep-arations of the 19 ThO and 10 Th2 CD4+ T-cellclones derived from patient 3, showed individualpatterns of rearrangement (data not shown), thusexcluding the possibility that the series of clonesobtained represent merely the progetiy of only afew infiltrating T cells.

All the 32 CD4-f clones showing the Thlprofile were compared with 32 TliO and 32 Tli2

1066

CD30 in atopic dermatitis

Table 2. Phenotype and cytokine secretion profile of T-cell clones derived from

lesionai skin of four patients with ADA) 100

Source

of cells(patient)

1. (S.A.)

2. (R.M.)

3. (B.V.)

4. (D.I.)

Ail patients

No. of cell

clonesobtained

CD4+ 41 (93)*CD8+ 3(7)

CD4+ 52 (90)

CD8+ 6 (10)

CD4+ 38(91)

CD8+ 4(9)

CD4-H 58 189)

CD8-I- 7(11)

CD4-I- 189(90)

CD8-f- 20(10)

Th1§

7(17)

210(19)

39(24)

• 2

6(10)

232(17)

9(45)

Cytokine secretion

ThO

23 (56)

127 (52)

219 (50)

130 (52)

399 (52)

7(35)

profile

Th2

11 (27)

015 (29)

110(26)

122 (38)

258 (31)

4(20)

* Numbers in parentheses indicate percentages. § T-cell clones able to produce

IL-4 and/or IL-5, but not IFN-y and/or TNF-p, were categorized as Th2, and clones

producing both IFN-y (and/or TNF-p) and IL-4 or lL-5 were categorized as ThO.

randomly selected CD4-I- clones for their ex-pression of rnernbrane CD30 as well as for IL-4,IL-5, IFN-y, and TNF-P production after 48-hstimulation with PHA. As shown in Fig, 3, thepercentages of clonal T-cell blasts expressingCD30 were positively correlated with the amountsof either IL-4 (r: -t-0.812, 7 = 0.0001) or IL-5(r: -F 0,734, P = 0.0001; data not shown) produced,but inversely correlated with production of IFN-y (r: -0.604, F = 0.0001) or TNF-P (n -0.562,P = 0.0001, data not shown). In agreetiient withdata previously reported (15), the ability tosecrete Tli2-type cytokities expressed by the fewTliO- and Th2-like CD8 T-cell clones was alsopositively correlated with the degree of mem-brane CD30 expression.

Serum levels of sCD30 are increased in most ADpatients

Since sCD30 is the proteolytically cleaved, 88-kDasoluble form of CD30, which is released by CD30-expressing cells in vitro and in vivo (30), serutnlevels of sCD30 were measured in 41 patients withacute, active AD. As control, sera from 19 non-atopic patients with active, nickel-induced ACDand from 60 healthy donors were also assessed.The mean value (±SE) of sCD30 in healthy sub-jects was 6.8 + 0.8 U/tnl, all individual values beinglower than, and in two donors equal to, 15 U/ml.As Fig. 4 shows, only 10% of ACD patients hadsCD30 serum levels higher than 15 U/ml (18 and19 U/ml, respectively), whereas in 71% of ADpatients, sCD30 serum levels were increased(ranging from 17 to 300 U/ml, mean 50.5 -)-11 U/ml) and correlated with disease activity scores(/• = 0.618),

50-CO

8

0.812

B) 100-1

2 4IL-4 (ng/ml)

50-CO

O

=-0.604

100 200

IFN-7 (U/ml)

300

Fig. 3. Correlation between membrane CD30 expression andproduction of different cytokines by CD4-I- T-cell clonesderived from skin infiltrates of patients with AD. A) Thirty-two Tl-iO randomly selected clones were compared with 32randomly selected Tli2 clones for membrane expression ofCD3() and lL-4 production upon VHA stimulation. B) Tliirty-two Till clones were compared for CD30 expression and IFN-7production with same 32 ThO clones. Membrane CD30 expres-sion and cytokine secretion were assessed as described inMaterial and methods.

DiscussionAlthough the pathogenetic mechanisms of AD arestill under debate, recent research on cytokines hasgradually elucidated some aspects of its pathophys-iology. AD is frequently associated with high IgElevels in the serum and with the presence in theskin of mast-cell-bound IgE specific to commonenvironmental allergens, whose production isstrictly regulated by IL-4 and/or IL-13 (32, 33). ADis also frequently associated with peripheral bloodeosinophiha, and blood eosinophil counts roughlycorrelate with disease severity (33). Moreover,although eosinophils are rarely detectable in ADskin lesions, several products of activated eosi-nophils, including tnajor basic protein, eosinophil-derived neurotoxin, eosinophil peroxidase, and

1067

Caproni et al.

300-

200-

1 100-

o

50-

15-oooooooooooooooooooooooooooooooooooooooooooooooooooooooooo

Atopiedermatitis

Contactdermatitis Healthy subjects

Fig. 4. Serum levels of sCD30 in AD patients, compared withACD patients and healthy subjects. sCD30 levels were assessedby CD30 ELISA in sera from 41 patients with active AD, 19patients with nickel-induced ACD, and 60 healthy controls.Dotted line indicates highest sCD3() values detected in controlsera. In AD patients, sCD30 levels correlated (r = 0.618) withdisease activity seores.

eosinophil cationic protein, have been shown to bedeposited in AD lesions (34). TTiree cytokines,GM-CSF, IL-3, and especially IL-5, have beenshown to influence greatly eosinophil differentia-tion in bone marrow and prolong eosinophil sur-vival in tissues. In addition, these cytokines canactivate eosinophils, as indicated by increased sur-face receptor expression and production of inflam-matory mediators (35-37). The mechanismsaccounting for the joint involvement of IgE-pro-ducing B cells, mast cells, and eosinophils in ADhave become clearer with the demonstration ofpolarized forrns of the specific immune response.Indeed, CD4-I- Th2 cells, which produce IL-4 andIL-5, in addition to GM-CSF, TNF-a, IL-3, IL-6,IL-9, IL-10, and IL-13, are likely to hold the keyto the question of why the trtast cell/eosinophil/IgE-producing cell triad is involved in the pathogenesisof allergic reactions, including those peculiar to AD(38).

CD30 is a rnember of the TNF-receptor super-family (15), and exists as a membrane glycoproteinchain of 105 and 120 kDa, derived frorn a 90-kDaprecursor and a 57-kDa intracellular form (20). Theextracellular portion of CD30 is proteolyticallycleaved to produce an 88-kDa soluble forrn of themolecule (sCD30), which is released by CD30-expressing cells in vitro and in vivo (21,30). Recently,

CD30 expression and sCD30 release have beenfound to be preferetitially associated with the pro-duction of Tli2-type cytokitie in vitro by both CD4 4-and CD8-I- human T-cell clones (16, 17). Moreover,high proportions of CD30-F T cells have beetiobserved in vivo in the lymph tiodes and skiti ofpatients with Omenn's syndrorne, a rare congetiitalimmunodeficiency syndrome in which a pathogenicrole of Th2 cells has been suggested (39). Finally,elevated levels of sCD30 have been found in patho-logic conditions characterized by marked activationof Tli2/TliO cells, such as systemic lupus erythemato-sus (40), rneasles virus (21), and HIV infection (41).

In contrast to this selective expression of CD30on Th2-type cells, it has been suggested that CD30can be expressed by Till like-cells, contributing toIFN-y secretion (42). Furtherrnore, Alzona et al.(19), by analyzing regulation of the CD30 expres-sion on blood T cells, found that CD30 defities asubset of activated T cells that can produce bothIFN-y and IL-5, and exhibit potent helper activityfor B-cell Ig production.

The results of our study provide further supportfor the hypothesis that ThO/Tli2-type responsespredorninate in the skin of patients with AD,inasmuch as most CD4-FT-cell clones generatedfrom the lesional skin of our patients with activeAD showed a TliO/Tli2 profile. The possibility thatthis outcome does not reflect the real functionalattitude of effector Tli cells in vivo, but is the resultof in vitro artifacts of the cloning procedure, wasconsidered. However, Thl-like clones were getier-ated by the same cloning protocol frotn the thyroidof patients with Hashimoto's thyroiditis (27) andfrom the retroorbital tissue of patients with Graves'ophthalmopathy (28).

Nevertheless, the most interesting finding of thisstudy was the demonstration of remarkable num-bers of CD30-I- cells in the lesional skin of ADpatients, as well as high values of its soluble fonrt(sCD30) in the serutii of most of these patients.Although double staining with anti-CD4 and anti-CD30 or with anti-IL-4 and anti-CD30 was notperforrned iti the present study, it is very likely thatthe CD30-)- cells foutid in AD skin are indeedCD4-I- T cells able to produce Th2-type cytokines.In fact, the great majority of skin-infiltrating cells inAD were CD4+ T cells (CD8-^ T cells being veryfew) that showed the same prevalent petivasculardistribution of CD30 -i- cells in the papillary areas.

Moreover, CD30 expression in T-cell clones gen-erated from the skin of our AD patients stronglycorrelated with the ability of these cells to produceIL-4 and IL-5, whereas it was inversely related totheir ability to produce IFN-y. Other in vitro clon-ing studies (9) have supported the view that skin-infiltrating CD4-I- T cells in patients with AD

1068

CD30 in atopic dermatitis

mainly belong to the Th2-like phenotype; panels ofrandomly cloned CD4 + T cells frotn lesional ADskin produced high amounts of IL-4, with no or lowlevels of IFN-y, thus showing a Th2/ThO-like pro-file. However, many notiallergen-specific clonesderived from atopy patch test (APT) sites or clonesgenerated from natural skin lesions produced IFN-Y in additioti to IL-4, or even IFN-7 alotie (12,43-44). These discrepancies have recently beenexplained by the dernonstration, in inhalatit-aller-gen APT, of a biphasic response to aeroallergen inAD, characterized by an initial Th2-dominatedresponse shifting into a subsequent predotninantTill response in situ. Immunocytochemical doublestaining with anti-IL-4 and anti-IFN-y in combina-tion applied to APT skin specimens revealed apredotninant IL-4 expression in the initial phase ofthe skin inflammatory response, followed by a shiftto predominantly IFN-y in the late and chronicphase, possibly due to the influx of IL-12-producingmacrophages (45).

On the whole, the retnarkable expression ofCD30 in the skin and the high levels of sCD30 inthe serum of our AD patients further support theconcept that AD is a Th2/TliO-dominated disorder.In fact, the high sCD30 levels docutnented in ourstudy may reflect the local release by CD4-I-CD30 -I- infiltratitig T cells, which, diffusing into thecirculation, may contribute, together with that prob-ably released in lymph nodes, to the elevation ofsCD30 levels in the serutn. Tliese data agree withthe recent observation that serum levels of sCD30were significantly elevated in children with AD andcorrelated closely with the degree of disease sever-ity (46). Our findings are of even greater value inview of the demonstration that neither CD30 -1- cellsin the lesional skin nor elevated serum levels ofsCD30 could be observed in any of the patients withcontact (nickel-induced) dermatitis. As is known,both contact dennatitis induced by exposure tochemicals in experimental animals and dermatitisinduced by contact with nickel iti about 10% of thehuman population are based on a chronic delayed-type (or type IV) hypersensitivity, which requiresthe participation of CD4-I- and CD8-h T cellsand the prevalent activation of Tlil-dominatedresponses (22). These latter findings provide furtherevidence that demonstration of CD30-I- cells intissues and/or increased levels of sCD30 in biologicfluids may be a simple and useful tool for thedetection of immune responses dominated by Tli2-type cytokines in the active phase of AD.

AcknowledgmentsThis work was supported in part by grants from the ConsiglioNazionale delle Riccrche (Targeted project "Prevention and

control of disease factors"), the Associazione Italiana per laRicerca sul cancro, and the European Union of Biotechnology"Transplantation" Network. We are indebted to Prof. MarioRicci for helpful discussion.

References1. Juhlin L, Johansson S, Bennich H, Hogman C, Tliyresson

N. Imniunoglobulin E in dermatoses. Arch Dermatol 1969;f00:12-16.

2. Mackie RM, Cobb S, Cochrane R, Tliompson J. Total andspecific IgE levels in patients with atopic dermatitis. ClinExp Dermatol 1979:4:187-95.

3. Zachary CB, Macdonald DM. Quantitative analysis of Tlymphocyte subsets in atopic eczema, using monoclonalantibodies and flow cytometry Br J Dermatol 1983;108:401-22.

4. Sillevis Smitt JH. Bos JD, Hulsebosch H.1, Krieg SR. /;; situimniunophenotyping of antigen presenting cells and T cellsubsets in atopic dermatitis. Chn Exp Dermatol 1985;112:149-56.

5. Bruijnzeel-Koomen CAMF, van Wichen DF, Spry CJF.Venge P, Bruijnzeel PLB. Active participation of eosi-nophils in patch test reaction to inhalant allergens inpatients with atopic dermatitis. Br J Dermatol 1988;118;229-38.

6. Mosmann TR, Coffman RL. Till and Th2 cells: differentpatterns of lymphokine secretion lead to different func-tional properties. Ann Rev Immunol 1989:7:145-73.

7. Romagnaiii S. Lymphokine production by human T cellsin disease states. Annu Rev Immunol 1994:12:227-57.

8. Wierenga EA, Snoek M. Bos .ID, Jansen I-IM,Kapsenbcrg ML. Comparison of diversity and function ofhouse dust mite-specific T lymphocyte clones trom atopicand non-atopic donors. Eur ,1 Immunol 1991;20:15t9-26.

9. van der tJeijden FL. Wierenga EA, Bos JD, Kapsenberg ML.High frequency of lL-4 producing CD4+ allergen-specificT lymphocytes in atopic dermatitis lesional skin. J InvestDermatol 1991:97:389-94.

10. Reinhold U, Kukel S, Goeden B, Ncuman U, Kreysel HW.Functional characterization of skin-infiltrating lymphocytesin atopic dermatitis. Clin Exp Immunol 1991: 86:444-8.

f 1. van Reijscn FC, Bruijnzeel-Koomen CAMF, Kalthoff FS.et al. Skin-derived aeroallergen-specific T-cell clones olTln2 phenotype in patients with atopic dermatitis. J AllergyClin Immunol 1992:90:184-93.

t2. Grewe M. Gyufko K, Schopf E, Krutmann J. Lesionalexpression of interferon-y in atopic eczema. Lancet1994:343:25-6.

13. Ohmen JD, Hanifin JM, Nickoloff BJ, et al. Overexpressionof lL-10 in atopic dermatitis. Contrasting cytokine patternswith delayed-type hypersensitivity reactions. J Immunol1995:154:1956-63.

14. Hamid Q, Boguniewicz M, Leung DYM. Differential in situcytokine expression in acute versus chronic atopic derma-titis. J Clin Invest 1994:94:870-6.

15. Smith C, Davis T, Anderson D, et al. A receptor for tumornecrosis factor defines an unusual family of cellular andviral proteins. Science 1990:248:It) 19-23. '

16. Del Prete G, De Carli M, Almcrigogna F, et al. Preferentialexpression of CD30 by human CD4+ T cells producingTli2-type cytokines. FASEB J 1995:9:81-6.

17. Manetti R, Annunziato F, Biagiotti R, et al. CD30 expres-sion by CD8 + T cells producing type 2 helper cytokines.Evidence for large numbers of CD8 + CD30 + T-cell clonesin human immunodeficiency virus infection. J Exp Med1994:180:2407-12.

1069

Caproni et al.

18. Del Prete G, De Carli M, D'Elios MM, et al. CD30-mediated signalling promotes the development of humanT helper type 2-like T cells. J Exp Med 1995;182:t655-61.

19. Alzona M, Jack HM, Fisher RI, Ellis TM. CD30 defines asubset of activated human T cells that produce IFN-y andtL-5 and exhibit enhanced B cell helper activity. J tmmunol1994;153:286f-7.

20. Smith CA, Gruss HJ, Davis T, et al. CD30 antigen, a markerfor Hodgkin's lymphoma, is a receptor whose ligand de-fines an emerging family of cytokines with homology toTNF Cell 1993:73:1349-60.

2L Del Prete G, Maggi E, Pizzolo G, Romagnani S. CD30, Th2cytokines and HIV infection: a complex and fascinatinglink. Immunol Today 1995;16:76-80.

22. Kapsenberg ML, Wierenga EA, Bos JD, Jansen HM.Functional subsets of allergen-reactive human CD4 + Tcells. Immunol Today 1991 ;12:392-5.

23. Kapsenberg ML, Wierenga EA, Stiekema FE,Tiggelman AM, Bos JD. Thl lymphokine production pro-files of nickel-specific CD4+ T lymphocyte clones fromcontact allergic and non-allergic individuals. J Invest Der-matol 1992:98:59-63.

24. Hanifin JM, Rajka G. Diagnostic features of atopic(iermatitis. Acta Derm Venereol SuppI (Stockh)1980:92:44-7.

25. Sowden J, Bert-Jones J, Ross JS. Double-blind, controlledcrossover study of cyclosporin in adults with severe refrac-tory atopic dermatitis. Lancet 1991:338:346-50.

26. Del Prete G, De Carli M, Mastromauro C, et al. Purifiedprotein derivative of Mycobacterium tuberculosis and ex-cretory/secretory antigen(s) of Toxocara canis expand invitro human T cells with stable and opposite (type 1 Thelper or type 2 T helper) profile of cytokine production.J Clin Invest 1991:88:346-50.

27. Del Prete G, Tiri A, De Carli M, et al. High potentialtumor necrosis factor (TNF)-a production of thyroid infil-trating T lymphocytes in Hashimoto's thyroiditis: a peculiarfeature of destructive thyroid autoimmunity. Autoimmu-nity 1989:4:267-76.

28. De Carli M, D'Elios MM, MarioUi S, et al. Cytolytic T cellswith Till-like cytokine profile predominate in retroorbitallymphocytic infiltrates of Graves' ophthalmopathy J ClinEndocrinol Metab 1993;77:1120-4.

29. Del Prete G, De Carli M, D'Elios MM, et al. Allergenexposure induces the activation of allergen-specific Tli2cells in the airway mucosa of patients with allergic respi-ratory disorders. Eur J Immunol 1993:23:1445-9.

30. Josimovic-Alasevic O, Durkop H, Schwarting R, Backe E,Stein H, Diamantstein T. Ki-1 (CD30) antigen is releasedby Ki-1-positive tumor cells in vitro and in vivo. Partialcharacterization of soluble Ki-1 antigen and detection ofthe antigen in cell culture supernatants and in serum by anenzyme-linked immunosorbent assay. Eur J Immunol1989;19:157-62.

31. Pene J, Rousset F, Briere F, et al. IgE production by normalhuman lymphocytes is induced by interleukin and sup-pressed by interferon a and y and prostaglandin E2. ProcNatl Acad Sci U S A 1988;85:6880-4.

32. Punnonen J, Aversa G, Cocks BG, et al. Interleukin 13induces interleukin 4-independent IgG4 and IgE synthesisand CD23 expression by human B cells. Proc Natl AcadSciU S A 1993;90:3730-4.

33. Uehara M, Izukura R, Sawai T. Blood eosinophilia inatopic dermatitis. Clin Exp Dermatol 1990:15:264-6.

34. Leiferman KM. A current perspective on the role ofeosinophils in dermatologic diseases. J Am Acad Dermatol1991:24:1101-12.

35. Sanderson CJ, Warren DJ, Strath M. Identification of alymphokine that stimulates eosinophil differentiation invitro, its relationship to interleukin-3, and functional prop-erties of eosinophils produced in cultures. J Exp Med1985:162:60-74.

36. Kita H, Ohnishi T, Okubo Y, Weiler D, Abranis JS, GleichGJ. Granulocyte/macrophagc colony-stimulating factor andinterleukin-3 release from human peripheral blood eosi-nophils and neutrophils. J Exp Med 1991; 174:745-8.

37. Caproni M, D'Agata A, Cappelli G, Fabbri P. Modulationof serum eosinophil cationic protein levels by cyclosporinin severe atopic dermatitis. Br J Dermatol 1996;135: 336-7.

38. Romagnani S. Atopic allergy and other hyperscnsitivities.Technological advances and new insights in pathogenesisprelude to novel therapeutic strategies. Curr Opin Immu-nol 1995:7:745-50.

39. Schandene L, Ferstcr A, Mascart-Lcmone F, et al. T helpertype 2-likc cells and therapeutic effects of interferon-ain combined immunodeficiency and hypereosinophilia(Omcnn's syndrome). Eur J Immunol 1993:23:56-60.

40. Caligaris-Cappio F. Bertero MT, Converso M, et al. Circu-lating levels of soluble CD30, a marker of cells producingTh2-type cytokines, are increased in patients with systemiclupus erythematosus and correlate with disease activity.Clin Exp Rheumatol 1995:13:339-43.

41. Pizzolo G, Vinante F, Morosato L, et al. High serum levelof soluble form of CD30 molecule in the early phase ofHIV-1 infection as an independent predictor of progressionto AIDS. AIDS 1994;8;741-5.

42. Hamann D, Hilkens CMU, Grogan JL, et al. CD30 expres-sion does not discriminate between human Thl- and Tli2-type T cells. J Immunol 1996;156:1387-91.

43. Sager N, Feldmann A, Schilling G, Kreitsch R NeumannC. House dust niitc-specilic T cells in the skin of subjectswith atopic dermatitis: frequency and lymphokine profilein the allergen patch test. J Allergy Clin Immunol1992:89:801-10.

44. Virtanen T, Maggi E, Menetti R, et al. No relationshipbetween skin-infiltrating Tli2-likc cells and allergen-spe-cific IgE response in atopic dermatitis. J Allergy ChnImmunol 1995:96:411-20.

45. Thepen T, Langevckl-Wildschut EG, Bihari IC, et al.Biphasic response against acroallergcn in atopic dermatitisshowing a switch from an initial Th2 response to a Thlresponse in situ: an immunocytochemical study. J AllergyClin Immunol 1996:97:828-37.

46. Frezzolini A, Paradisi M, Ruffelh M, Cadoni S, De Pita O.Soluble CD30 in pediatric patients with atopic dermatitis.Allergy 1997;52:106-9.

1070