The Journal of Dermatology Vol. 18: 314-323, 1991

Immunohistochemistry of Cutaneous Graft-versus-Host Disease after Allogeneic Bone Marrow Transplantation Giampiero Girolomoni, Carlo Pincelli, Giovanna Zambruno, Marco Andreani*, Claudio Giardini*, Guido Lucarelli* and Alberto Giannetti Abstract

Graft-versus-host disease (GVHD) is an immunologically mediated disease occurring most frequently after allogeneic bone marrow transplantation. The aim of this study was to evaluate the contribution of immunohistochemistry in the diagnosis of cutaneous GVHD. Patients transplanted for either leukemia or ,8-thalassemia were included in the study. Skin lesions of acute and chronic GVHD were examined both by direct immunofluorescence to detect immunoglobulin deposits and by an avidin-biotin-peroxidase complex technique to evaluate the inflammatory cell infiltrate. Epidermal and dermal fluorescent bodies (IgG and IgM) were frequently found in both acute and chronic GVHD. Most of the infiltrating cells were CD3+ T lymphocytes, with CD8+ cells representing the major cell population invading the epidermis both in acute GVHD and in chronic lichenoid GVHD. A small proportion of the dermal cells were CDI4+macrophages; no B cells were detected. RIA-DR, but not RIA-DQantigens, were variably expressed by keratinocytes in all cases of acute GVHD and in chronic lichenoid GVHD. KL-I, a monoclonal antikeratin antibody specific for the 56.5 KD acidic polypeptide usually present in suprabasal keratinocytes, stained all epidermal layers, including the basal layer. Langerhans cells were dramatically reduced in number in the epidermis of both acute and chronic lichenoid GVHD. It is concluded that immunohistologic analysis may be supportive in the diagnosis of acute and early chronic lichenoid cutaneous GVHD. Abbreviations: GVHD: graft-versus-host disease, BMT: bone marrow transplantation, ABCPx: avidin-biotin peroxidase complexes, mAb: monoclonal antibody, LC: Langerhans cells

Key words: cutaneous graft-versus-host disease; diagnosis; immunohistochemistry Introduction Graft-versus-host disease (GVHD) is a major complication following allogeneic bone marrow transplantation (BMT). The skin is a preferential target organ both in acute and chronic GVHD (1). However, clinical signs and symptoms, particularly in acute GVHD, may mimic those caused by viral and other infecReceived January 31, 1991; accepted for publication May 16, 1991. Clinica Dermatologica, Universita di Modena and *Divisione Ematologica, Centro Trapianto Midollo Osseo di Muraglia, Ospedale di Pesaro, Italy. Reprint requests to: Giampiero Girolomoni, M.D., Department of Dermatology, University of Modena, Via del Pozzo 71, 41100 Modena, Italy.

tions, which are relatively common in immunosuppressed patients or those caused by drugs (2). Furthermore, certain histopathological features in acute GVHD can also be produced by irradiation or cytotoxic drugs (3-5). Therefore, more reliable criteria are required for the diagnosis of acute GVHD in order to allow the institution of appropriate treatment as early as possible. The importance of immunohistologic findings for the diagnosis of GVHD has been debated in the literature in recent years, and the specificity of the immunological abnormalities has been questioned (6-10). GVHD is an immunologically mediated disease. All the immunohistologic studies published so far have been performed in patients

Immunohistochemistry of Cutaneous GVHD receiving BMT for diseases such as leukemia or severe aplastic anemia which are characterized by profound immunologic changes. These abnormalities might be of relevance to the development and immunohistologic appearance of GVHD lesions. In the present study, we immunohistologically evaluated skin biopsies from patients with a clinical course and skin histologic features characteristic of GVHD. Four patients received BMT for leukemia and eight for .a-thalassemia, a condition not associated with immunologic abnormalities (11). The results were similar in the two groups of patients and indicate that immunopathological findings may represent supportive tools rather than pathognomonic features for the diagnosis of GVHD.

Materials and Methods Patients Seven patients with acute GVHD and six with chronic GVHD were included in the study. Eight out of thirteen patients were affected by homozygous ,B-thalassemia, four by chronic myeloid leukemia and one by acute lymphoblastic leukemia. All bone marrow donors were siblings who were identical for HLA-A, B, C, and DR, and had nonreactive mixedleukocyte cultures with cells of their respective recipients. Conditioning regimen consisted of busulfan and cyclophosphamide in all cases but one (patient n. 13), who received cyclophosphamide plus total body irradiation. GVHD prophylaxis was attempted with cyclosporin A alone in patients with thalassemia and with cyclosporin A in association with short term methotrexate in patients with leukemia (9, 10). Patients with acute GVHD were graded according to the Seattle classification (1). Chronic GVHD were categorized according to the time of onset, type and extension of skin lesions. Biopsies from patients with acute and chronic lichenoid GVHD were performed on recently developed lesions. Additional details are given in Table 1. Skin biopsies were divided into two specimens: one was processed for routine histopathology, and the other was included in OCT compound (Miles Laboratories, Elkhart, IN), immediately snap frozen in liquid nitrogen, cut at 4 u, and fixed in 3% paraformaldehyde at 4°C for 10 min,

315.

Immunohistochemistry techniques For direct immunofluorescence, frozen sections were air-dried, preincubated with diluted goat serum, washed in phosphate buffered saline (PBS), and then incubated with fluoresceinated antihuman IgG,IgM,IgA, C3c, and fibrin (Behring Institute, Scoppito, L'Aquila, Italy) at 1:20 dilution. Sections were then washed and evaluated using a Zeiss epifluorescent microscope. Immunoperoxidase staining was performed using an avidin-biotin-peroxidase complex amplification system (ABCPx), as previously reported (15). The following monoclonal antibodies (mAb) were employed: OKT6 (anti-CDla) was procured from Ortho Diagnostic System (Raritan, ~); LeuM3 (antiCDl4), Leu12 (anti-CDl9), Leu4 (anti-CD3), Leu7 (anti-CD57), anti-HLA-DR, anti-interleukin 2 receptor (CD25) were purchased from Becton Dickinson (Mountain View, CA); T4 (anti-CD4) and T8 (antiCD8) were obtained from Coulter Immunology (Hialeah, FL), anti-HLA-DQ from Biotest (Dreiech, Germany), and KL-l, which recognizes a 56.5 KD keratin polypeptide, from Immunotech (Marseille, France). Serial sections were fixed, incubated with 1% solution H 20 2 in PBS to block endogenous peroxidase, and incubated with normal swine serum (20% in PBS). Sections were then washed, stained with the above mentioned mAb at room temperature, and washed in PBS. Biotin-labeled anti-mouse Ig (Vector Laboratories, Burlingame, CA) were used as a second-layer antiserum. After further washing, sections were incubated with freshly prepared avidin-biotin peroxidase complex (Vector). Peroxidase was developed with 3 amino-9 ethyl carbazole and the slides were counterstained with Mayer's hematoxylin. A section from each specimen was processed as a negative control by omitting the firstlayer mAb. The percentage of positively stained cells was recorded by two independent observers using a semiquantitative method.

Results Direct immunofluorescence showed a moderate number of fluorescent bodies (mostly IgM and IgG) in five out of eight biopsies of acute ?"VHD and in six out of seven biopsies of chrome GVHD; they were located in the basal epiderm.is, at the dermo-epidermal junction and/or III the superficial dermis. Fluorescent

Girolomoni et al

316

Table 1. GVHD patients included in the study

A

Acute GVHD patients

Patient

Age/sex

Disease

Grade

1 2 3 4

8M 16 M 10 M 5M 35 F 38 M 33 F

ThaI" Thai Thal Thai e CML CML CML

III IV II I

5

6

7

I IV I

Day of onset +24 +42 +14 +11 +9 +21 +44

Day of biopsy +35 +42 +49 +16 +28 +25 +50

Therapy pb,ATGc P, Aza d P P P P,ATG

B. Chronic GVHD patients Patient

Age/sex

Disease

Clinical features

Day of biopsy

8

18 F

ThaI

9

11M 10M 10 F 18 F 20 M

ThaI ThaI Thai CML

(a) progressive, ext., poikiloderma (b) progressive, ext., poikiloderma quiescent, ext., lichenoid, mild quiescent, ext., lichenoid, severe progressive, ext., poikiloderma progressive, ext, poikiloderma (a) progressive, ext., poikiloderma (b) progressive, ext., poikiloderma

+360 +630 +95 +79 +156 +960 +369 +640

10 11 12 13

ALLg

Therapy

P,Aza P,ATG P, CSAf P,Aza P,Aza

&,B-thalassemia; bprednisone; "anti-thymocyte globulin; dazathioprine; "chronic myeloid leukemia; 'cyclosporin A; gacute lymphoblastic leukemia.

bodies were seen in both acute and chronic GVHD and their number was in agreement with the degree of epidermal necrosis observed in histologic sections. Granular and/or linear deposits of immunoglobulins, C3c, and fibrin were sometimes present at the dermo-epidermal junction (data not shown). In patients with acute GVHD (Table 2), a mild to discrete dermal cell infiltrate invading the epidermis either focally or diffusely was seen in all cases. The vast majority (>90%) of dermal cells were T lymphocytes (CD3+) which intensely expressed major histocompatibility complex (MHC) class II antigens (HLA-DR+, HLA-DQ+). Cells of the monocyte/macrophage lineage (CD14+) were observed in moderate number in the dermis, while dermal CDla+ dendritic and/or round cells, CD57+ cells, and CD25+ cells were rarely detected. No B-cells were observed in any of the sections. CDW cells predominated in 4 out of 7 biopsies in the dermal infiltrate, while, in the epidermis, CD8+

cells were more numerous than CD4+ cells in all cases (Fig. 1). Langerhans cells (LC) (CD1a+) in the epidermis were either dramatically decreased or had completely disappeared. Keratinocytes displayed HLA-DR antigens on the cell membrane in all cases, although to different extents, either limited to the basal layer or involving the entire epidermis. Furthermore, HLA-DR expression by keratinocytes was more evident in those sites characterized by a higher number of T lymphocytes close to and/or infiltrating the epidermis (Fig. 2). Keratinocytes did not exhibit HLA-DQ antigens. The whole epidermis, including the basal cell layer, was intensely stained with the mAb KI.r 1 directed against the 56.5 KD acidic keratin polypeptide, which is normally present only in the suprabasal cells of human epidermis (16) (Fig. 3). Apoptotic keratinocytes in the papillary dermis were also stained by KI.r 1 mAb. In late chronic GVHD (Table 3), poikilodermatous lesions showed a scant perivascular

Immunohistochemistry of Cutaneous GVHD

317

Fig. 1. Acute GVHD. The majority of the dermal infiltrate around the hair follicle and invading the epidermis consists of CDWT lymphocytes (ABCPx, x128).

infiltrate with a few CD3+/CD8+ or CD3+/CD4+ cells scattered in the epidermis. In the dermis, CD3+ lymphocytes were again the predominant cell type; CD4+ cells were higher in number than CD8+ cells. HLA-DR and HLA-DQ were strongly expressed by T cells in the dermis. CD14+ cells were the second most numerous cell type, while CDla+, CD25+, and CD57+ were only rarely detected. CDI9+ cells were not observed at all. Keratinocytes did not exhibit MHC class II antigen expression, and the number of epidermal LC was unchanged as compared to pretransplant biopsies (data not

shown). In the two cases of early chronic lichenoid GVHD, there was a heavy band-like lichenoid infiltrate invading the epidermis, composed mostly of T lymphocytes. Intraepidermal CD3+ cells were predominantly CD8+ (Fig. 4A). As in acute GVHD, keratinocytes were found to be diffusely HLA-DR+ (Fig. 4B); KL-I mAb stained all cell layers and epidermal LC were markedly reduced in number. Patients receiving transplants for either leukemia or thalassemia exhibited similar histologic as well as immunohistologic findings.

Girolomoni et al

318

Table 2. Immunoperoxidase findings in acute GVHD Patient

Dermis"

T8

Leu12

OKT6

IL-2R

HlA-DR

OKT6

IL-2R

HlA-DR

HlA-DQ ND 1 ± ± ++++ +++ ++ +++ ++ ND 2 ± ++++ +++ +++ +++ + 3 ND + ++++ +++ ++ ++ + 4 ± ± ++++ +++ ++ + ++++ ++++ + 5 +++ +++ ++++ ++ ++ + 6 +++ ++ +++ ++++ +++ + 7 ± ± +++ ++ + + ++++ +++ "Percentage of positive cells: -, negative; ±, :::;5%; +, 5-25%; ++, 26-50%; +++,51-75%; ++++, ~76%. Leu4

T4

Leu7

LeuM3

Table 3. Immunoperoxidase findings in chronic GVHD Patient

Dermis"

Leu4

T4

T8

Leu12

HlA-DQ 8 (a) ± ND +++ ++ + + + ++++ (b) ± +++ ++ + ++ + ++++ ++ 9 ± ± ND +++ +++ ++ + + ++++ 10 ± ± ND ++++ ++ +++ ++++ 11 ± ± +++ ++ + +++ ++++ ++ 12 ± ++ ++ + ++ ++++ ++ 13 (a) ± +++ ++ + ++ ++++ ++ (b) ± +++ ++ + + +++ ++ 'Percentage of positive cells: -, negative; ±, :::;5%; +,5-25%; ++, 26-50%; +++, 51-75%; ++++, ~76%. Leu7

Discussion The direct immunofluorescence findings observed in both acute and chronic GVHD are quite common in many skin diseases and probably represent nonspecific depositions secondary to the inflammatory state (17). Four immunopathological changes have been consistently observed in acute GVHD in the present study: I) a predominant infiltration of CD8+ T cells at the dermo-epidermal junction and within the epidermis, 2) a marked depletion of epidermal LC, 3) an abnormal expression of HlA-DR antigens on keratinocytes, and 4) KL-1 mAb staining of the epidermal basal layer. The first feature is in agreement with previous reports (5, 8, 9). On the other hand, CD8+ cells at the dermo-epidermal junction have also been observed in other skin diseases characterized by lymphocyte aggres-

LeuM3

sion to the epidermis, including erythema multiforme (18) and pityriasis lichenoides (15). In addition, CD8+ cells are the major subpopulation ofT lymphocytes present in normal human skin (19). The functional significance of these cells and their pathogenetic role in the epidermal damage occurring in GVHD is not known at present. CDW T cells are known to be triggered by antigens bound to MHC class I molecules and to exert cytotoxic activity (20). Inasmuch as keratinocytes express class I MHC molecules and are severely damaged during the course of acute GVHD, it is conceivable that donor CDW T cells participate in the induction of the epidermal necrosis. However, the nature of the antigen(s) involved remains elusive (21). The isolation and clonal expansion of these CD8+ cells in vitro might greatly help in elucidating their functional properties and specificities.

319

Immunohistochemistry of Cutaneous GVHD

T8/T4 ratio Dermis Epidermis 0.8 0.9 1.7 0.8

1.4 2 2

2 1.2 1.2 1.7 1.5 1.8 1.3

bll, markedly reduced; -, negative; membranes; ND, not done.

T8/T4 ratio Dermis Epidermis 0.5 0.3 1.2 2 0.8 0.5 0.2 0.3

OKT6 b

KL-Ic

II II

ND

II II

II

Epidermis HlA-DRd

++ ++++ + ++ ++ +++ ++

+ + + + + +

HlA-DQd ND ND ND

c+, staining of the basal layer; "Expression on keratinocyte

OKT6 b

KL_I c

Epidermis HlA-DRd

HlA-DQd ND

1.5

1.4

I II

0.6

++++ +++ +

+ + ND

b=, normal number; I, slightly reduced; II, markedly reduced; dExpression on keratinocyte membranes; ND, not done.

The second point in our study is a dramatic reduction in LC number in the epidermis in acute GVHD in all patients. This finding, however, is also observed after BMT in patients who do not develop GVHD (22). Data from our laboratory confirm a marked depletion of LC in normal appearing skin during the course of the first month after BMT (23). This alteration is thought to be the result of the conditioning regimen, which can damage LC directly in the epidermis, and/or block the continuous replacement from the marrow. In acute GVHD, keratinocytes constantly expressed HlA-DR antigens. Whether this feature is also induced in patients before the onset of clinical lesions is still controversial. In most studies, HlA-DR+ keratinocytes have been found only in GVHD lesions, but not in normal appearing skin after BMT (8, 9). On the other hand, other authors have detected a few HLA-

+ c+, staining of the basal layer;

DR+ keratinocytes in patients undergoing BMT, but without GVHD (24, 25). However, this latter finding did not appear to predict the developmentofGVHD (24, 25). Furthermore, keratinocyte expression of MHC class II molecules has been reported in a variety of skin diseases, all ofwhich are characterized by lymphocyte infiltration within the epidermis (26). Indeed, cytokines, such as y-interferon (y-IFN) and tumor necrosis factor a (TNF-a) secreted by activated T cells, induce MHC class II molecule synthesis by keratinocytes (27). Some hypotheses explaining the functional relevance of this phenomenon have been recently discussed. Both keratinocytes and pancreatic beta cells bearing MHC class II molecules are able to promote antigen specific unresponsiveness in T cells (28, 29). Thus, the expression of MHC class II molecules may serve as a mechanism to downregulate immunologically mediated inflamma-

320

Girolomoni et al

Fig. 2. Acute GVHD. Keratinoeytes express HLA-DR antigens, particularly where a higher number ofinflammatory cells is present in the underlying papillary dennis (t) (ABCPx, x128).

Fig. 3. Acute GVHD. All epidennallayers, includingthiebasaI cell layer, are posi~~~I;';~~~;d with KIA mAb (ABCPx, x200). Apoptotic keratinocytes in the superficial dermis are also stained (t).

Immunohistochemistry of Cutaneous GVHD

Fig. 4. Chronic GVHD, lichenoid type. (A) Keratinocytes show diffuse expression of RIA-DR antigens (ABCPx, x128); (B) Infiltration of the epidermis by CD8+ T lymphocytes (ABCPx, x200).

321

Girolomoni et al

322

tion. Release of cytokines such as TNF-a and y-IFN by activated T cells may also playa direct role in the induction of epidermal necrosis (30).

KL-l rnAb, stained all epidermal cell layers, including the basal layer in acute GVHD. Dreno et al. (9) observed a similar staining of the basal layer associated with discrete alterations in KL-l expression of the suprabasal layers, particularly in acute GVHD. The significance of such an abnormal staining is at present unknown, but it is interesting to note that basal cell expression of this keratin has been observed in hyperproliferative conditions of the epidermis (31) and that, in neonatal rats with GVHD, an increase of epidermal basal cell labeling index was detected by autoradiographic studies (32). In chronic GVHD with poikilodermatous appearing lesions, the immunohistologic patterns were totally nonspecific, whereas in lichenoid lesions they resembled those of acute GVHD: a predominant CD8+ cell infiltrate, KL1 basal epidermis staining, and a significant depletion of epidermal LC. Such a reduction in LC number is possibly secondary to the epidermal damage since, by this time, there is a complete repopulation of the epidermis by donor-derived LC (33). In conclusion, in the present study, no immunohistologic changes were found to be specific for acute or for chronic GVHD; the findings observed in patients undergoing BMT for either leukemia or thalassemia were similar. The individual immunohistologic parameters we studied are, therefore, of limited value in the diagnosis of GVHD. Nonetheless, they may support, together with the clinical and histologic criteria, the diagnosis of cutaneous GVHD. Further studies are needed to discover immunohistologic markers which may be more helpful in the early diagnosis of this lifethreatening disease. References 1) Glucksberg H, Storb R, Fefer A, et al: Clinical manifestations of graft-versus-host disease in human recipients of marrow from HLA-matched sibling donors, Transplantation, 18: 295-304, 1974.

2) Hood AF, Soter NA, Rappeport J. Gigli I: Graftversus-host reaction: Cutaneous manifestations following bone marrow transplantation, Arch Dermatol, 113: 1087-1091, 1977. 3) Sale GE, Lerner KG, Barker EA, Shulman HM, Thomas ED: The skin biopsy in the diagnosis of acute graft-versus-host disease in man, Am J Pathol, 89: 621-636, 1977. 4) Hymes SR, Simonton SC, Farmer ER, Beschorner WB, Tutschka Pj, Santos GW: Cutaneous busulfan effect in patients receiving bone-marrow transplantation,] Cut Palhol, 12: 125-129, 1985. 5) Lever R, Turbitt M, Mackie R, et al A prospective study of the changes in the skin in patients receiving bone marrow transplants, BrJ Dermatol, 114: 161-170, 1986. 6) Lampert lA, janossy G, Suitters AT, et al: Immunological analysis of the skin in graft versus host disease, Clin Exp Immunol, 50: 123-131,1982. 7) Guyotat D, Mauduit G, Chouvet B, et al: A sequential study of histological and immunological changes in the skin after allogeneic bone marrow transplantation, Transplantation, 41: 340-342, 1986. 8) Atkinson K, Munro V, Vasak E, Biggsj: Mononuclear cell subpopulations in the skin defined by monoclonal antibodies after HLA-identical sibling marrow transplantation, BrJ Dermatol, 114: 145-160, 1986. 9) Dreno B, Milpied N, Harrousseau jL, et al: Cutaneous immunological studies in diagnosis of acute graft-versus-host disease, Br J Dermatol, 114: 7-15, 1986. 10) Vole-Platzer B, Rappersberger K, Mosberger I, et al: Sequential immunohistologic analysis of the skin following allogeneic bone marrow transplantation,] Invest Dermaiol, 91: 162-168,1988. 11) Lucarelli G, Polehi P, Izzi T, et al: Allogeneic bone marrow transplantation for thalassemia, Exp Hematol, 12: 676-681, 1984. 12) Lucarelli G, Polchi P, Galimberti M, et al: Marrow transplantation for thalassemia following busulfan and cyclophosphamide, Lancet, 1: 1355-1357,1985. 13) Lucarelli G, Galimberti M, Polehi P, et al: Marrow transplantation in patients with advanced thalassemia, N EnglJ Med, 316: 1050-1055,1987. 14) Thomas ED, Storb R, Clift RA, et al: Bone marrow transplantation, N EnglJ Med, 292: 832-843,895-902, 1975. 15) Giannetti A, Girolomoni G, Pincelli C, Benassi L: Immunopathologic studies in pityriasis Iichenoides, ArchDermatol Res, 280: s61-s65, 1988. 16) Staquet Mj, Albert J. Lawrence lJ, Thivolet J: Flow cytometric sorting of human epidermal pure basal cell suspensions using a specific antikeratin monoclonal antibody,] Invest Dermatol, 87: 792-794, 1986. 17) Claudy AL, Schmitt D, Freycon F: Graft versus host reaction in the skin: histological, immunological and ultrastructural studies, Acta Derm Venereol, 59: 7-25,

Immunohistochemistry of Cutaneous GVHD 1979. 18) Margolis RJ, Tonnesen MG, Harrist TJ, et al: Lymphocytes subsets and Langerhans cells/indeterminate cells in erythema multiforme,j Invest Dermatol, 81: 403-406, 1983. 19) Bos JD, Zonneveld I, Das PK, Krieg SR, van der Loos CM, Kapsenberg ML: The skin immune system (SIS): distribution and immunophenotype of lymphocyte subpopulations in normal human skin,] Invest Dermatol, 88: 569-573, 1987. 20) Norment AM, Salter RD, Parham P, Engelhard VH, Littman DR: Cell-cell adhesion mediated by CD8 and MHC class I molecules, Nature, 336: 79-81, 1988. 21) De Gast GC, GratamaJW, Ringden 0, Gluckman E: The multifactorial etiology of graft-versus-host disease, Immunol Today, 8: 209-212, 1987. 22) Murphy GF, Merot Y, Tong AKF, Smith B, Mihm MC Jr: Depletion and repopulation of epidermal dendritic cells after allogeneic bone marrow transplantation in humans,] Invest Dermatol, 84: 210-214, 1985. 23) Girolomoni G, Zambruno G, Manca V, et al: Effect of chemotherapy conditioning regimen on Langerhans cells in thalassemic and leukemic patients undergoing allogeneic bone marrow transplantation,] Invest Dermatol, 92: 434A, 1989. 24) Beschorner WE, Farmer ER, Saral R, Stirling WL, Santos GW: Epithelial class II antigen expression in cutaneous graft-versus-host disease, Transplantation, 44: 237-243, 1987. 25) Sviland L, Pearson ADJ, Green MA, et a1: Class II antigen expression by keratinocytes and enterocytes-an early feature of graft-versus-host disease, Transplantation, 46: 402-406, 1988.

323

26) Aubock], Romani N, Grubauer G, Fritsch P: HLA-DR expression on keratinocytes is a common feature of diseased skin, Br] Dermatol, 114: 465-472, 1986. 27) Breathnach S, Katz SI: Keratinocytes synthesize Ia antigen in acute cutaneous graft-vs-host disease,] Immunol, 131: 2741-2745, 1983. 28) Gaspari AA, Jenkins MK, Katz SI: Class II MHCbearing keratinocytes induce antigen-specific unresponsiveness in hapten-specific Thl clones,] Immuno~ 141: 2216-2220, 1988. 29) MarkmannJ, Lo D, Naji A, Palmiter RD, Brinster RL, Heber-Katz E: Antigen presenting function ofclass II MHC expressing pancreatic beta cells, Nature, 336: 476-479, 198B. 30) Piguet P-F, Grau GE, Allet B, Vassalli P: Tumor necrosis factor/cachectin is an effector of skin and gut lesions of the acute phase of graft-vs.-host disease,] Exp Med, 166: 1280-1289,1987. 31) KanitakisJ, Urabe A, Haftek M, ThivoletJ: The effect of cyclosporin A on proliferation and differentiationassociated antigens of normal human skin xenografted onto nude mice,] Dermatol Sci, 1: 103-110, 1990. 32) Stuart SP, Klein RM, Clancy J: Kinetics of mast cell, fibroblast and epidermal cell proliferation during acute graft-versus-host disease in the neonatal rat,] Invest Dermatol, 88: 369-374, 1987. 33) Perrault C, Pelletier M, Belanger R, et al: Persistence of host Langerhans cells following allogeneic bone marrow transplantation: possible relationship with graft-versus-host disease, Br] Hematol, 60: 253-260, 1985.