Retinoic Acid Upregulates Human Langerhans Cell Antigen Presentation and Surface Expression of HLA-DR and CD11c, a P2 Integrin Critically Involved in T-Cell Activation

Laurent Meunier, Kim Bohjanen, John]. Voorhees, and Kevin D. Cooper Department of Dermatology, University of Michigan, Ann Arbor, Michigan, U.S.A.

Ittununomodulatory effects of retinoids may be part of their anti-carcinogenic and anti-inflammatory prop­erties. We studied the in vivo effects of retinoic acid (&A) on antigen-presenting activity of human epider­lUal Langerhans cells and on accessory cell activity of keratinocytes. Two skin sites from each volunteer were treated in vivo with 0.1% RA or vehicle, respectively, once a day for 4 d. RA-treated epidermal cell (RA-EC) alloantigen presentation to CD4+ T cells in each vol­unteer tested was consistently greater than that in­duced by vehicle EC. However, this increased antigen­presenting activity did not lead to autoreactive CD4+ T -lymphocyte proliferation. Elevated unfractionated epidermal antigen-presenting activity of RA-EC was not due to increased keratinocyte major histocompati­bility complex (MHC) or intercellular adhesion mole­cule expression or to other keratinocyte accessory signaling, because incubation o~ CD1a- flou~os~e~ce­activated cell sorter (FACS)-purtfied RA-EC lDhlblted alloantigen presentation, presumably through in­creased keratinocyte transforming growth factor-po By contrast, Langerhans cell function was upregu-

Although the immunomodulatory effects of retinoids are not well defined [1 -4], the beneficial anti-carci­nogenic properties of retinoids [5 - 7] may be attrib­uted in part to modulation of the host tumor Immune response [8] . Similarly, modulation of the .human

anti-papilloma virus immune response [9] ~a~ be a mec~al1lsm un­derlying the beneficial clinical effect of ret1l10~ds 111 treat1l1g bem~n and precancerous viral papillomas of the shn, larynx, or cervix [10-12].

Cellular immune responses in the skin are initiated by presenta­tion of antigen to antigen-specific T cells by cutaneous Langerhans cells (LC) [13] . Like murine LC, human epidermal LC function as potent antigen-presenting cells for a variety of T-cell-dependent

Manuscript received March 11, 1994; revised June 17, 1994; accepted for publication June 24, 1994. . .

Reprint requests to: Dr. Kevll1 D. Cooper, Immunodermatology Unit, University of Michigan, R5548 Kresge I/Box 0530, Ann Arbor, MI 48109.

Abbreviations: APC, allophycocyanin; EC, epidermal cells; FI, fluores­cence index; L90LS, log 90· light scatter; LC, Langerhans ce ll ; LFlTC, log fluorescein isothiocyanate; LFLS, log forward light scatter; PE, phycoer­ythrin; RA, retinoic acid.

lated; FACS-purified CDla+ Langerhans cells derived from RA-EC displayed a markedly increased ability, relative to Langerhans cells from vehicle EC, to present alloantigen to T cells. Triple color flow-cyto­metric analysis ofRA-EC and vehicle EC suspensions revealed that RA treatment did not modify the num­ber of DR + and CD la+DR + EC, but did result in statis­tically significant increases in Langerhans cells expres­sion of HLA-DR, CD11c, and CDlc. Another novel finding was that HLA-DR-dependent Langerhans cells antigen-presenting activity in both normal and RA-treated skin was completely blocked by anti­CD11c antibody. Thus, retinoid upregulation of anti­gen-presenting activity may be due to upregulation of Langerhans cell CD11c, as well as class II MHC. Up­regulation of cutaneous immune responsiveness in human skin without autoreactivity has not (to our knowledge) been reported previously, and the Langer­hans cell phenotypic and functional state achieved is distinct from previously reported states of Langer hans cell activation.] Invest Dermato1103:775 -779, 1994

immune responses to haptens and antigens [14,15]. We postulated that topically administered retinoids may modulate the antigen­presenting activity of epidermis, either as a direct effect on LC or as an indirect effect on keratinocytes.

Therefore, we assessed the i/1 lI illo effects of retinoic acid (RA) on the antigen-presenting activity of human epidermal LC and on the antigen-presenting accessory cell activity of keratinocytes. hI lIillO, RA treatment of human skin upregulated critical antigen-present­ing cell molecules on LC, such as CDl1c and HLA-DR, resulting in an increased immunologic LC functional activity. This occurred without induction of auto reactivity, possibly because of a concomi­tant increase of keratinocyte inhibition of antigen-presenting activ­ity. Selective upregulation of human immunologic molecules and cells of dendritic antigen-presenting cell lineage may be of rele­vance to clinical states in which immune stimulation by LC may be beneficial.

MATERIALS AND METHODS

Subjects Two sites from each vol unteer were treated ill "ivowith 0.1 % RA (R. W . Johnson Pharmaceutical Research Institute, Raritan, NJ) dissolved into a vehicle composed of 70% ethanol and 30% propylene glycol or with the ethanoljpropylene glycol vehicle alone, respectively, once a day for 4 d,

0022-202X/94/S07.00 Copyright © 1994 by The Society for Investigative Dermatology, Inc.

775

776 MEUNIER ET AL

at which time keratome biopsies were performed. After each application, the treated sites were covered with a semiocclusive plastic wrap (Telfa-Pad). None of the volunteers (male/female ratio = 5: 4) displayed redness or irri­tation with this treatment, in contrast to RA application under full occlusion [16].

Antibodies The panel of monoclonal antibodies employed is as follows: anti - HLA-DR (Becton Dickinson, Mountain View, CAl, anti-CDla (Bec­ton Dickinson), anti-CDlb and CDlc (Amac, Westbrook, ME), OKMl (Ortho, Raritan, NJ), Leu M5 (Becton Dickinson), anti-CDlla (Amac), anti-HLA-l (Atlantic Antibodies), OKM5 (Ortho), Leu M3 (Becton Dick­inson), anti-CD54 (Amac), anti-HPCA-l (Becton Dickinson), anti-Fclg­GRII (Medarex, West Lebanon, NH), and anti-FcIgGRl (Medarex). Poly­clonal antibodies included fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse immunoglobulin (Ig)Gl (Boehringer, Indianapolis, IN) and phycoerythrin (PE) goat anti-mouse IgG2b (Cal tag, South San Fran­cisco, CAl. Streptavidin conjugated with allophycocyanin (APe) was pur­chased from Biomeda (Foster City, CAl ·

Flow-Cytometrie Analysis Epidermal suspensions, prepared as de­scribed previously [17], were FcIgGR-blocked with purified heat-aggre­gated (65 'C, 20 min) human IgG (Sigma Chemical) at a concentration of 0.5 mg/ml for 30 min at 4'C and triple stained as described previously [17]. Flow cytometry was performed using an Epics Elite (Coulter Cytometry, Hialeah, FL). Signals for log forward light scatter (LFLS) and log 90' light scatter (L90LS), as well as fluorescence from FITC, PE, and APC were collected on a four-decade logarithmic scale (LFITC, LPE, LAPC), as de­scribed previously.

Based upon isorype control values, HLA-DR+ or CD1a+ epidermal cells (EC) were electronically selected from single-parameter histograms of cell counts at each level of APC or FITC fluorescence intensity, respectively. The selected cells were then analyzed on two additional histogram parame­ters, such as LFITC versus LPE, to determine isorype background and spe­cific staining intensities on LC themselves, in the absence of contaminating autofluorescent keratinocytes. The mean fluorescence intensity of particular markers was determined by subtracting the nonspecific background (isotype controls) from the fluorescence intensity of cells stained with specific mono­clonal antibodies. Fluorescence index (FI) of a surface marker on a subset population was defined as the ratio of mean fluorescence intensity of the specific antibody to mean fluorescence intensity on that specific subset stained with the isotype control. The percent of cells expressing a given marker was determined as the number of cells within a given area of a histogram displaying positive staining with the isotype control antibody subtracted from the number of cells within the same area of the histogram displaying positive staining with the specific marker.

Functional Assessment of Alloantigen-Presenting Capacity of EC Eighty thousand EC were mixed with 50,000 allogeneic CD4+ T lymphocytes (prepared by negative selection in round-bottom microtiter wells for 7d) to assess [3H]Tdr (ICN Radiochemicals, Irvine, CAl incorpora­tion over the last 18 h [18]. In some experiments, freshly separated EC were stained with anti-CDla (Ortho), and fluorescence-activated cell sorting was performed to obtain separately CD1a+ and CD1a- EC. In other experi­ments, before co-culture with CD4+ T lymphocytes, EC were preincubated for 30 min at 37 ' C with one of the following antibodies: LeuM5 (anti­CDllc) (Becton Dickinson) (25 fLg/ml), anti-CDlc (Amac) (50 fLg/ml), mouse IgG2b (Pharmingen, San Diego, CAl (25 fLg/ml), and mouse IgG1 (Sigma) (50 fLg/ml).

For assessment of keratinocyte accessory antigen-presenting activity, CDla- fluorescence-activated cell sorter (FACS)-isolated EC were incu­bated at various concentrations with gamma-irradiated JY B cells, an anti­gen-presenting cell line, and allogeneic CD4+ T cells. 3H-Tdr uptake was then assessed at 7 d as above.

RESULTS

RA Upregulates Major Histocompatibility Complex (MHC) Class II Mean Fluorescence Intensity on Epidermal LC But Does Not Modify the Number ofHLA-DR+ or CDla+DR+ EC After 4 d of ifl vivo treatment with RA, the mean percentage of LC in RA-treated sites was not statistically different from the per­centage of LC in vehicle-treated areas (Table I) . No evidence of interceUular adhesion molecule (ICAM+) cells, or DR-CD11c+, DR-CD1c+, or DR-CD36+ infiltrating cells was seen in any of the subjects (0% of total EC for each) (data not shown). To compare accurately the mean surface molecule density on a cell-per-cell basis for group comparisons from multiple individuals, we determined the FI of specific monoclonal antibody binding to cell subsets rela-

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Table I. In Vivo RA Treatment of Human Skin Does Not Modulate LC Numbers or Induce

an Influx of Non-LC Leukocytes"

Vehicle RA (% of Total EC) (% of Total EC)

Total DR+ 2.4 ± 0.3% 2.1 ± 0.2% DR+CD1a+ 1.7 ± 0.2% 1.6 ± 0.2% DR+FcIgGRiI+ 2.0 ± 0.2% 1.7 ± 0.3%

n

5 5 4

• Data expressed as mean percentage of total vehicle and RA-EC ± SEM of indicated number of subjects. Antigen-presenting cell subset infiltrations not induced by topical RA(O.l%).

tive to isotype control binding [19,20). The group mean FI of anti­HLA-DR on LC obtained from RA-treated sites (81.7 ± 9.3) (n = 5) was almost twice the mean FI of anti-HLA-DR on vehicle LC (44.9 ± 6.7) (n = 5) (p < 0.02) (Table II).

The class II MHC fluorescence increase was not associated with an increase in overall cell size because the LFLS of RA-EC was similar to that of vehicle EC; the mean ratio of RA-LC LFLS to vehicle LC LFLS was 1 .06 ± 0.05 light-scatter units. These data indicate that increased RA-LC class II MHC expression is due to an increase in class II MHC molecule density on LC.

RA Specifically Upregulates CDlc, but Not CDla or CDlb Expression on LC Constitutive human LC expression of the dendritic antigen-presenting cell lineage markers CD1a, CD1b, and CD1c [17,18,21- 23] may be functionally important for T-cell activation [24,25). However, RA treatment exerted no effect on CD1a or CD1b expression by LC (vehicie-CD1a Fl, 220 ± 41.5 versus RA-CDla Fl, 183 ± 38) (n = 5) (p> 0.3) (Table II) . By contrast, the mean Fl of anti-CD1c on RA-treated LC (4.9 ± 0.8) was slightly but consistently (four of five subjects tested) higher than the CD1c FI on vehicle-treated LC (3.2 ± 0.5) (p < 0.04) (n = 5) (Table II).

Upregulation ofCDllc, but Not CDlla Or CDllb Expres­sion On LC by RA The monocyte/macrophage CD11c (aXfJ2 integrin) is also expressed 011 normal LC [17,26). In vivo retinoid treatment resulted in a consistent upregulation of CD11c expres­sion on epidermal LC: CD11c-FI on RA-treated CD1a+DR+ LC (16.1 ± 4.4) was almost double that of CDllc-Fl on vehicle­treated LC (9 ± 2.4) (p < 0 .03) (n = 5) (Table II). This upregula­tion was selective for CD11c because RA did not induce de novo expression of CD11b (aM [32 integrin) and CDlla (LFA-1) (aLfJ2 integrin) on freshly prepared epidermal LC (not shown).

A representative experiment demonstrates raw data on LC CD11c and CD1a intensity (Fig 1). DR+ cells from vehicle-EC (panels A, B) and from RA-EC (panels C, D) were selected for second- and third-marker analysis. EC stained with anti- HLA-DR,

Table II. Fluorescence Intensity Index of Human LC Molecules and Keratinocyte Class I MHC and ICAM

Molecules After In Vivo RA Exposure"

HLA-DR CLla CDle CDltc FcIgGRII On unfraetionated EC

(96-98% keratinocytes) Class I MHC I CAM

Vehicle (Mean FI ± SEM)

44.9 ± 6.7 220.0 ± 41.5

3.2 ± 0.5 9.0 ± 2.4

20.5 ± 6.5

106.0 ± 25.7 1.0 ± 0.08

RA (Mean FI ± SEM)

81.7 ± 9.3b

183.0 ± 38.1 4.9 ±0.8'

16.1±4.4J

14.0 ± 5.0

133.0 ± 30.7 1,,1 ± 0.09

n

5 5 5 5 4

5 4

• Data expressed as mean FI of specific antibodies on HLA-DR+ vehiclc-EC and RA-EC relative to isotype ± SEM of indicated l1umber of subjects.

• p < 0.02. ' p < 0.04. J P < 0.03.

VOL. 103. NO.6 DECEMBER 1994

HLA-DR+ Veh-EC

HLA-DR+ RA-EC

.1 (FITC) IgG1

.-t

2

c

4

leee

-;'-""""'1""'" 1I",-rTTTII~III1...JI"'Tl mJ",TTll' ,J . 1 (FITC) IgG1 leoe

u .­.-Q U

... ell ell .-t

-w a. -

u .­.-Q U -w a. -

... ... CIl .-t

in combination with FITC IgGl and PE IgG2b isotype controls, aJlowed cursors for defining positivity and negativity to be placed specifically according to LC nonspecific fluorescence, without in­terference by keratinocyte fluorescence (Fig 1, left pallels). Upon combining FITC OKT6 and PE anti-CD11c with the APC anti­DR, all DR+ cells moved from quadrant 3 to quadrant 2 (Fig 1, right panels), indicating positive expression of both CD1a and CD11c. In contrast to CDla intensity, which was unchanged, the mean chan­nel PE (CD11c) fluorescence intensity of DR+ vehicle-EC was 14.8, whereas the mean channel CD11c fluorescence of DR+ RA­Ee was 42.3.

By contrast to the modulation of CD 11 c and DR, the LC version of the FcIgGRII receptor (CD32) was not affected by retinoid ther­apy (Table II). RA treatment did not induce ICAM and CD36 expression on LC, or CD34, CD14, CD64, or CD1b (data not shown). RA also did not produce epidermotropism of cells bearing these surface markers (data not shown) . Keratinocytes did not mod­ulate their expression of class I MHC (Table II), and ICAM was not induced on keratinocyte surfaces (data not shown).

RA Upregulates EC Alloantigen Presentation But Does Not Induce Autoreactivity Incubation of CD4+ T cells with vehicle versus RA-EC revealed an increase in the allogeneic T-cell prolifer­ation in response to RA-EC relative to vehicle-EC in each of the three subjects tested (Fig 2). The fold increase in cpm of RA-EC relative to vehicle-EC in each subject was 3.8, 2.6, and 3.2, respec­tively. We next asked whether RA-induced upregulation ofLC was linked to induction of an autoreactive response, as occurs in diseased skin [18,27] or upon ill vitro culture of epidermal LC [27 - 29]. However, coculture of CD4+ T lymphocytes with either autolo­gous vehicle- or RA-treated EC did not lead to T-cell proliferation (Fig 2, n = 2) .

Purified RA-LC, and Not Keratinocytes, Are Responsible for Increased Epidermal Antigen-Presenting Activity Whereas both FACS-purified CDla- vehicle-EC and CDla- RA-EC were ineffective in inducing alloreactivity, increasing concentrations of purified CDla+ EC (2.5 X 103 to 20 X 103/ml) resulted in a dose­dependent increase in antigen-presenting activity (Fig 3). In comparison to purified vehicle-LC, RA-CDla+ EC reproducibly displayed an increased activity at each antigen-presenting cell

RETINOID MODULATION OF HUMAN ANTIGEN PRESENTATION 777

1 2

c .!""

3 4

-rrmm~nnnm"~TI~-rrmrn

• 1 \FITC) OKT6 Ieee

1 ',2 ....

.... ; ' ;;~ ... =f.--------l--~

3 4

.. 1·111T1nr-Tl1T1mr rTTll 'lI - ,T

. 1 (FITC) OKT6 leee

Figure 1. Flow-cytometric analysis of electronically gated HLA-DR + EC expressed on two-parameter histo­grams. Isotype controls: Mouse IgG I (FITC goat ami-mouse IgGl) and IgG2b (PE goat anti-mouse IgG2b) binding on HLA-DR+ vehicle-treated (upper/eft) and RA-treated (lolVer left) EC (representative experiment). DR+ vehicle EC (upper right, qlladrallts 2, 4) and DR+ RA-EC (lolVer right, quadrants 2, 4) subsets co­express CDla (OKT6) and the ax fJ2 integrin CDllc. Relative to CDllc+ DR+CDla+ vehicle EC (upper right. quadrant 2), the CD1Ic+DR+CDla+ RA-EC population is shifted to a higher CDllc intensity on the Y axis (lolVer right, quadrant 2). indicating a higher ex­pression of the CDllc marker on DR+ RA-treated LC.

concentration tested above 100 LC/well (mean fold increase = 5.9 ± 2.1) (Fig 3, representative experiment). These data indicate that RA-LC are more potent in their antigen-presenting activity than vehicle-Ee.

To determine whether upregulation of RA-EC antigen-present­ing activity was due not only to LC upregulation, but also to modu­lated keratinocytes. CD1a- RA-EC and CD1a- vehicle-EC were purified by FACS as above, and cultured with autologous CD4+ T cells and allogeneic B-antigen - presenting cells (JY B cell line). At this concentration, CDla- vehicle-EC exhibited little effect on an­tigen-presenting activity (Fig 4), but CD la- RA-EC demonstrated a potent ability to inhibit alloantigen presentation (Fig 4). Thus, CD1a- FACS-purified RA-EC inhibited alloantigen presentation by the JY B cell line, indicating that elevated antigen-presenting activity of unfractionated RA-EC was not due to enhanced kerati­nocyte accessory signals, but was due to upregulated LC antigen­presenting activity alone.

100

80 ~

M

6 60 ~ :E 40 Q, (,)

20

0 CD4+TC CD4+TC + VEH-EC + RA-EC

(AUTO) (AUTO)

Figure 2. RA upregulates allogeneic CD4+T-cell (TC) proliferation induced by normal EC. Coculture of vehicle or RA-treated EC with autologolls CD4+ TC does not lead to auroreactiviry. Results represent the mean ± SEM cpm of[3H]Tdr uptake after 7 d culture of three (a11o) and two subjects (auto).

778 MEUNIER ET AL

35

30

o VEH-CD1a+ LC

• RA-CD1a+ LC

~ 25 M

6 20

~ :E 15 Q.

0 10

5

0 CD4+T

+ 100 LC CD4+T

+ 500 LC CD4+T

+ 1000 LC CD4+T

+2000 LC CD4+T

+4000 LC

Figure 3. RA upregulates antigen-presenting cdl activity of puri­fied epidermal LC. CDla+ and CDla- RA and vehIcle EC were sorted by FACS and cultured with allogeneic CD4+ T lymphocytes (50 X 103 per well). Purified LC derived from RA-EC displayed a markedly increased ability, relative to vehicle LC, to present antigen to T cells . Results are expressed as the mean ± SEM cpm of triplicate samples (n = 2).

The fJ2 Integrin CDllc (p 150/95) is Critical for the Anti­gen-Presenting Activity of Both Vehicle- and RA-Treated Epidermal LC To determine whether the upregulated CD1c and CD11c may, in addition to class II MHC, participate in the elevated RA-LC antigen-presenting activity, we performed block­ing antibody studies. Preincubation of ex vi~o vehicle- or RA-treated EC with anti-CD11c (50 ,ug/ml) resulted In a blockade of allogen­eic CD4+ T-cell responses (Fig 5), whereas incubation with mouse IgG2b isotype did not modify the upregulation of foreign alloanti­gen presentation induced by retinoid treatment (Fig 5). By contrast, incubation with anti-CD1c, relative to its mouse IgG1 isotype con­trol, had no effect on the allogeneic T-cell response (data not shown).

DISCUSSION

We have demonstrated a pharmacologic method that upregulates human antigen-presenting activity in vivo without concomitant in­duction of autoreactivity. The upregulated antigen-presenting ac­tivity of LC was associated with upregulated expression of class II MHC and CD11c. Class II MHC is critical for LC antigen presen­tation [13], and its upregulation is known to be associated with upregulated antigen-~resenting activity [29]. We demonstrate here that CD11c (gp150j95) is also critical for LC antigen-presenting activity (Fig 5); its upregulation by RA (Fig 1, Table II) may thus

25

20 ....... b 15 .... ~ ~ 10 Q.

0 5

0 JYB PLUS: +TC

APC: JY B CELL LINE (2.5x1()3/wELL) TC: APe DEPLETED CD4+ KC: VEH va RA CD1a- EC

VEH RA

1 X 1()4 KC

VEH RA

2 X 1()4 KC

Figure 4. Elevated whole RA-EC antigen-presenting activity is not due to enhanced keratinocyte accessory signals. CDta- RA- and vehi­cle EC were sorted by FACS to obtain LC-depleted keratinocytes (KC) and cultured with allogeneic CD4+ T cells and 2.5 X 103 JY B cells. Data are expressed as mean ± SEM [3HJTdr uptake of triplicate samples at 7 d.

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

100

....... 80 '" 6

~ 60

:::E 40 Q.

()

20

0 T+ T+ T+ T+ Veh-EC Veh-EC Veh-EC RA-EC RA-EC RA-EC

+ IgG2b + aCD11c +lgG2b +aCD11c

Figure 5. COllc expression on LC is critical for both normal and retinoid-modulated antigen-presenting activity. Preincubation of EC with anti-CDllc (50 Ilgjml) markedly inhibits, relative to IgG2b isotype, the alloantigen presentation of both vehicle and RA-EC to CD4+ T cells. Results are expressed as mean ± SEM [3HJTdr uptake at 7 d.

also account for increased epidermal LC activity. CD11c (aX P2 integrin) is expressed on normal LC [26] , on normal dermal LC [17], and on dendritic interdigitating cells in the lymph nodes [30] . Our results show that RA treatment markedly increases CD11c expression on LC and that this adhesion molecule plays a crucial role in antigen-presenting T-cell interactions (Fig 5) .

The elevated capacity of purified RA-LC to activate T lympho­cytes was not reflected in the keratinocyte population. LC-depleted EC were not able on their own to activate T cells (RA or vehicle) and, in fact, CD1a- RA-EC (keratinocytes) exhibited increased in­hibitory effects on T-cell activation induced by the JY APC line. Retinoids may cause keratinocytes to release inhibitory soluble cy­tokines, although topical RA treatment does not modify human epidermal interleukin-1 (IL-1) levels [31-33] or production of ara­chidonic acid products such as the T-cell inhibitor prostaglandin E2 [32]. Topical RA does, however, increase human epidermal trans­forming growth factor-p levels [16,34], a potent inhibitor ofT-cell activation. The opposing effects of RA-treated keratinocytes versus RA-treated LC may account for the lack of induction of autostimu­latory capacity of the RA-EC, despite upregulated LC ability to present alloantigen.

Our in vivo results differ from the inhibition of epidermal anti­gen-presenting activity by synthetic retinoids upon addition to an itl vitro co-culture of BC and allogeneic lymphocytes [35] . Although human cells were used in both cases, RA was not tested in the study of Dupuy et al. Retinoids can alter T-cell receptors for mitogens [4,36,37] or for IL-2 and thus also reduce proliferative responsive­ness at the T -cell level [38 - 41].

Our study'S phenotypic and functional changes after in vivo RA treatment are to some extent similar to those observed after itl lIivo epicutaneous hapten sensitization in mice (i.e., increased I-A and I-E class II MHC expression) [42] and during the itl I/itro maturation of LC into potent dendritic cells (i.e., increased class II MHC and antigen-presenting activity) [28,43,44]. However, the effect of RA differs from the effect of in vitro culture on LC in that RA does not result in the reduced CD1a and FcIgGRII expression observed upon culture of LC and does not induce ICAM-1 or autoreactivity as occurs in cultured LC [43]. These differences indicate that RA does not induce a state of differentiation that represents maturation into dendritic antigen-presenting cells oflymphatics and regional nodes. RA-LC also differ from hyperstimulatory LC present in atopic der­matitis and mycosis fungoides [18,45] in that RA-LC lack CD36 and CD1b. These comparisons indicate that RA-LC represent a previously undescribed state of activation/differentiation of the human LC.

In conclusion, ill vivo RA treatment of human skin results in an increased ability to present alloantigen to T cells, but not in a man­ner that breaks the threshold for autoreactive responsiveness, per-

VOL. 103, NO.6 DECEMBER 1994

haps because ?f a conc.omitant incr~ase in keratinocyte immunosup­pressive activi ty. The IIlcreased actiVity appears related to changes 111

surface-marker expression on LC such as HLA-DR and CDllc. Thus, we have demonstrated that in addition to class II MHC, CDllc is a molecule critically involved in T-cell activation by LC. The ability to upregulate cutaneous antigen-presenting function without modulating autoimmunity may be important in improving cutaneous immune responsiveness to microbial infections, vaccines, or tumor antigens.

This work was sllpporled ill parI by Ihe Cilag Marc Chaptal Fellowship, the Phar­macetltical Research Irlslilute, JohllsO Il & Joh 1150 II, alld the N IH UM-MAC Core Facility Grallt (AR20557).

We gratefully thallk Pallia Hochmall, PhD, Biogetl, Cambridge, Massachtlsetts, Jar the JY B cell line, alld Patricia Hopkins alld Marsha Gamble Jar their assistatlCe ill the preparatioll of this lIIanliscripl.

1.

2.

3.

4 .

5.

6.

7.

8.

9 .

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

REFERENCES

H aftek M, Faure M, Schmitt D , Thivolet J: Langerhans cells in skin from patients w ith psoriasis: quantitative and qualitative study ofT6 and HLA-DR antigen­expressing cell s and changes with aromatic retinoid administration. J blllesl Dermato/81:10-14,1983

Fernandez-Bussy RA, Schmitt D, Maudu it G, Thivolet J: Effects of aromatic retinoid (Ro 109359) on Langerhans' cells in lichen planus. Arch Dermatol Res 275:105-108, 1983

Shapiro PE, Edelson RL: Effects of retinoids on the immune system. In: SauratJH (ed.). New Trellds ill R esea rch alld T herapy. Reti'lOid Symposi"m. Gelleva, 1984. Karger, Basel, 1985, pp 225-235

Sidell N , Rieber P, Golub SH: Immunological aspects of retinoids in humans. l. Analysis of retinoic acid enhancement of thymocyte responses to PHA. Cell IrIll,"mo/87:118-125,1984

Eccles SA: Effects of retinoids on growth and dissemination of ma.lignant tumors: immunological considerations. Bioe/,em Pha rmacoI34:1599- 1610, 1985

Smith MA, Parkinson DR, C heson BD, Friedman MA: Retinoids in cancer ther­apy. J Clill 011<0110:839-864,1992

DiGiovanna JJ: Retinoids for the future: onco logy.J Am Acad DermatoI27:S34-S37,1992

Halliday GM, Ho KK-L, St C Barnetson R: Regulation of the skin immune system by retinoids during carcinogenesis.] [lI vestDermatoI 99:8?S -86S, 1992

Cooper KD, Androphy EJ, Lowry D, Katz SI: Antigen presentation and T-cell activation in epidermodysplasia verruciformis. J b lVest Dermatol 94:769 - 776, 1990

Lutwer MA, Blanchet-Bardon C, Orth G: Clinical observations, virologic stud­ies, and treatment trials in patients with cpidcrmodysplasia vcrruciformjs, a disease induced by specific human papiliomaviruses.J blllesl Der",ato/83:18S-25S, 1984

Euvrard S, Verschoore M, Touraine JL, Dureau G, Cochat P, Czernielewski J, Thivolet J: T opical retinoids for warts and keratoses in transplant recipients (letter). Lall ,et 340:48-49,1992

Tallman MS, Wiernik PH: Retinoids in cancer treatment. J Clill phar",acol 32:868-888,1992

Sting I G, Tamaki K, Katz SI: Origin and function of epidermal Langerhalls cells. [rm,wllol Rev 53:150-174,1980

Katz SI, Cooper KD, Iijima M, Tsuchida T : The role of Langerhans cells in antigen presentation. J [liVest Der",ato/85:96S-98S, 1985 . .

Czernielewski J, Faure M, Schmitt D, Thivolet J : In vItro mIxed skill ceillym­phocyte cu lture reaction (MSLR) in man: analySIS of the epidermal cel l and T cell subpopulations. Clill Exp [""mmo/50:426-433, 1982 . .

Fisher GJ, T avakkol A, Griffiths CEM, Elder JT, Zhang Q-Y, Fmkel L, Damel­pour D, Glick AB, Higley H , Ellingsworth L, Voodlees JJ: DIfferential ~odu­!ation of transforming growth factor-beta 1 expreSSIon and mucm deposmon by retinoic acid and sodium lauryl sulfate in human skin.] [liVest Der",atoI98: 102-108, 1992

Meunier L, Gonzalez-Ramos A, Cooper KD: Heterogeneous popu lations of class II MHC+ cells in human dermal cell suspensions. Identification of a small subset responsible for potent dermal antigen-presenting cell activiry with fea­tures analogous to Langerhans cell s. J [",,,"1t101151 :4067-4080, 1993

Taylor RS, Baadsgaard 0, Hanunerberg C, Cooper KD: Hyperstimulatory CD l a+CDlb+CD36+ Langerhans cells arc responsible for increased autolo­gous T lymphocyte reactivity to lesional epidermal cells of patients with atopic dermatitis. J [,,"'"1t10/147:3794-3802, 199 1

Muirhead KA, Schmitt T C, Muirhead AR: Determination of linear fluorescence intensities from flow cytometric data accumulated with logarithmic amplifiers. Cytometry 3:251 - 256, 1983

RETIN OID MODULATION OF HUMAN ANTIGEN PRESENTATIO N 779

20. Hulette BC, Banavali SD, Finke DP, Gopal V, Preisler HD: Protooncogene expression in subpopulations of cells from leukemia patients. Cytometr)' 13:653-658, 1992

21. Knowles RW, Bodmer WF: A monoclonal antibody recognizing a human thymus leukemia-like antigen associated with beta 2-microglobulin. E .. r J [ m­""ltIo/1 2:676-681,1982

22. van de Rijn M, Lerch PG, Bronstein BR, Knowles RW, Bhan AK, T erhorst C: Human cutaneous dendritic cells express two glycoprotei ns T6 and M241 which are biochemicall y identical to those fou nd on cortical thymocytes. H .. m [",,,, ulloI9:20 1-210, 1984

23. Elder JT, Reynolds NJ, Cooper KD, Griffiths CEM, Bleicher PA: COl gene expression in human skin. J D er",alol Sci 6:206 - 213, 1993

24. Theodorou ID, Boumsell L, Calvo C-F, Gouy H, Beral HM, Debre P: CDI stimulation of human T cell lines induces a rapid increase in the intracel lular free Ca2+ concentration and the production of IL-2.J ["''''''"0/144:2518-2523, 1990

25. Faure F, Jitsukawa S, Miossec C, Hercend T: CDle as a target recognition struc­ture for human T lymphocytes: "lalysis with peripheral blood gamma/delta cells. E,,, J [",,,,,mol 20:703 - 706, 1990

26. De Panfilis G, Soligo D, ManaraGC, Ferrari C, Torresani C: Adhesion molecules on the plasma membrane of epidermal cell s. I. human resting Lallgerhans cells express two members of the adherence-promoting CDll /CD I 8 family, H­MAC-l (CDllb-CD18) and gp 150,95 (CDllc/CDI8). J [liVest Dermatol 93:60 -69, 1989

27. Delllidem A, Taylor JR, Grammer SF, StreileinJW: T-Iymphocyte-activating properties of epidermal antigen-presenting cells from normal and psoriatic skin: evidence that psoriatic epidermal antigen-presenting cells resemble cul­tured nonnal Langerhans cell s.J blllest DermatoI97:454-460, 1991

28. Schuler G, Steinman RM: Murine epidermal Langerhans cells mature into potent inununostimulatory dendritic cells in vitro. J Exp Med 161 :526 - 546, 1985

29. Shimada S, Caughman SW, Sharrow SO, Stephany D, Katz SI: Enhanced anti­gen-presenting capacity of cultured Langerhans cells is associated with mark­edly increased expression of Ia antigen. J ["''''''"0/139:2551 -2555, 1987

30. Hart DN, McKenzieJL: Isolation and characterization of human tonsil dendritic cel ls. J Exp Med 168: 157 -170, 1988

31. Chatellard GD, Didierjean L, Gumowski-Sunek D, Saurat J-H: Effect of topical retinoic acid on the intedeukin I and beta immunoreactive pool in normal human epidermis. Br J Der",atoI1 23:283-289, 1990

32. Fisher GJ, Esmann J, Griffiths CEM, Talwar HS, Duell EA, Hammerberg C, Elder JT, Karabin GO, N ickoloff BJ, Cooper KD, Voorhees JJ: Cellular, immunological and biochemical characterization of topical retinoic acid­treated human skin.J blllest D er",aloI96:699-707. 1991

33. Tokura Y. Edelson RL, Gasparro FP: Retinoid augmentation ofbioactivc inter­leukin-l production by murine keratinocytes. Br J Der",atoJ 126:485 - 495, 1992

34. Epstein SP, Baer RL, Thorbecke J , Belsito DV: Immunosuppressive effects of transforming growth factor beta: inhibition of the induction of Ia antigen on L.lngerhans cells by cytokines and of the contact hypersensitiviry response. J [liVest Der",atoI96:832 -837, 1991

35. Dupuy P, Bagot M, Heslan M, D ubertret L: Synthetic retinoids inhibit the anti­gen presenting properties of epidemtal cells in vitro.] IrI1'est Der",atoI 93:455-459, 1989

36. Valone FH, Payan DG: Potentiation of mitogen-induced human T-Iymphocyte activation by retinoic acid. Callcer R es 45:4128 - 4131 , 1985

37. Dillehay DL, Cornay WJ, Walia AS, Lamon EW: Effects of retinoids on human thymus-dependent and thymus-independent mitogenesis. Clill [mtn,mol [m mll­IIopatiI0150:100 - 108, 1989

38. Sidell N , Ramsdell F: Retinoic acid upregulates interleukin-2 receptors on acti­vated human thymocytes. Cell [",,,"mo/I15:299-309 , 1988

39. Bollag W: Modulation of human immune function in vitro by temarotene and its metabolite. Skill Pltar",aco/ 4:437-443, 199 1

40. Felli MP, Vacca A, Meco D, Screpanti I, Farina AR, Maroder M, Martinotti S, Petrangeli E, Frati L, Gulino A: Rctinoic acid-induced down-regulation of the interlcukin~2 promoter via cis-regulatory sequences containing an octamer rnotif. MoICellBio/ll :4771 - 4778, 199 1

4 1. Prabhala RH, Garewal HS. Hicks MJ, Sampliner RE, Watson RR: The effects of 13-cis-rcrinoic acid and beta-carotene on cellular ilnmuniry in humans. Catrcer 67:1556- 1560, 1991

42. Aiba S, K.l tz 51: Phenorypic and functional characteristics of in vivo-activated Langerhans cel ls. ] ["'''',11101 145:2791-2796, 1990

43. Romani N , Lenz A, Glassel H, Schuler G: Cultured human Langerhans cells resemble lymphoid dendritic cells in phenorype and function.] b,,'est D ermatol 93:600-609, 1989

44. Teunissell MBM, Wormmeester J , Kreig SR, Peters PJ, Vogels IMC, Kapscnberg ML, Dos JD: Human epidermal Langerhans cells undergo profound morpho­logic and phenorypical chan ges during in vitro culture. J IrlVest Dermatol 94:166- 173, 1990

45. Lisby S, Baadsgaard 0, Cooper KD, HansenER, Mchreg.n D, Thomsen K, Allen E, Vejlsgaard GL: Phenotype, ul trastructure. and function ofCDl+DR+epi­dermal cells that express CD36 (OKMS) in cutaneous T -cell lymphoma. Scand J [m""IIIoI32:111-119 , 1990