Journal of Dermatological Science, 1 (1990) 103-l 10 Elsevier

103

DESC 00017

The effect of cyclosporin A on proliferation and differentiation-associated antigens of normal human skin xenografted onto nude mice * Jean Kanitakis,

Atsumichi

Urabe, Marek Haftek and Jean Thivolet

INSERM U. 209, H6pital Edouard Herriot. Lyon, France (Received

30 September

Key words: Cyclosporin

1989; accepted

A; Epidermal

22 November

proliferation;

Epidermal

1989)

differentiation

Abstract Cyclosporin A (CsA) has been shown to inhibit, in vitro, the proliferation of cultured normal and neoplastic keratinocytes and to exert also in vivo an antiproliferative effect on keratinocytes of normal human skin xenografied onto nude mice. To gain further insight into the effects of CsA on human skin we investigated the immunohistochemical expression of several epidermal proliferation- and differentiation-associated antigens in the same model: six-week-old nude mice received a xenograft of full-thickness normal human skin; six animals subsequently received a daily subcutaneous injection of 50 mg/kg of CsA diluted in olive-oil while the others received an equivalent volume of olive-oil. The rate of epidermal proliferation was evaluated through a BrdU pulse-labelling technique, and was found to be decreased by 56% in the CsA-treated epidermal xenografts as compared to the controls. The xenografts were further examined for the expression of the following antigens: Epidermal Growth Factorand Transferrin-receptors, Ki-67, 56.5 kD keratin polypeptide, Filaggrin, Involucrin, fi2-microglobulin, Ulex Europaeus I- and Peanut-Agglutinin-binding sites. Most of these antigens were unchanged on CsA-treated human xenografts. However, the 56.5 kD keratin polypeptide which was consistently expressed by both basal and suprabasal epidermal keratinocytes in control xenografts showed a normal expression pattern (i.e. suprabasal keratinocytes only) in three out of the six CsA-treated xenografts. These results raise the possibility that, concurrently with a cytostatic effect, CsA may also affect keratinocyte differentiation and that this effect, possibly contributes in the beneficial effect of CsA in diseases of abnormal keratinization.

Introduction Cyclosporin A (CsA), a cyclic undecapeptide of fungal origin is a potent immunosuppressant used with success to prevent allograft rejection.

Correspondence to: J. Kanitakis, INSERM U. 209, Hopital Eduard Herriot, 69437 Lyon Cedex 03, France. * This work was presented at the Tricontinental SID/ ESDR/JSID Joint Meeting, Washington DC, 26-30/4/89. 0923-181 l/90/$03.50

0 1990 Elsevier Science Publishers

Although its main cellular target was considered to be the T-helper subset of blood lymphocytes, more recent data supports that other cell types also constitute targets for interesting biological activities of CsA. It was recently shown, for instance, that CsA exerts in vitro a direct antiproliferative effect on several epithelial cell types in culture, including normal human epidermal keratinocytes (EK) [l-7], and normal and transformed mouse EK [7]. We have recently shown

B.V. (Biomedical

Division)

104

that CsA exerts also in vivo an antiproliferative effect on EK of normal human skin xenografted onto nude mice [ 81. Since in this model there is a deficiency in T-lymphocytes, the cytostatic effect of CsA seems to be directly exerted on EK. Normal EK produce a variety of antigenic markers the expression of which is correlated to the degree of proliferation or differentiation of the epidermis. In the present work we studied the immunohistochemical phenotype of normal human EK xenografted onto CsA-treated nude mice in an attempt to gain further insight into the effects of CsA on human epidermis.

inserted in it, fixed in position with transpore surgical tape and dressed with self-adhesive sticking plaster. Ten days after grafting the dressings were removed. 12 mice bearing welladherent normal-looking NHSX were selected for the study and randomly assigned to the control or the CsA-treated group (each consisting of 6 mice). GA administration An oily solution of CsA (100 mg/ml, Sandoz Inc., Basel, Switzerland) was further diluted in olive oil and a dose of 50 mg/kg of this product was injected subcutaneously every day for 3 weeks on the back area of mice of the treated group. Control mice received a daily injection of the equivalent volume of olive oil.

Material and Methods Animals and grafting technique Six-week-old congenitally athymic female ‘nude’ mice (Swiss, nu/nu, Iffa Credo, Les Oncins, France) were used as recipients of normal human skin xenografts (NHSX). The latter were prepared from breast skin of a healthy woman undergoing plastic surgery, by using an electric keratotome trimming the skin to a depth of approx. 0.25 mm. Mice were anaesthesized with intraperitoneal pentobarbital and a circular graft bed 1 cm in diameter was prepared in the dorsolateral upper thoracic region. A circular NHSX of a slightly larger size than the graft bed was TABLE

Immunohistochemical study During the third week, all mice received intraperitoneally injections of 5-bromo-2’-deoxyuridine, according to a previously-described schedule [ 81. At the end of the study the NHSX were excised, and divided in two parts: one of them was fixed in Baker’s fixative and used for the immunohistochemical detection of BrdU-positive EK, while the other was snap-frozen in liquid nitrogen andstored at - 30 ‘C until used. Subsequently, 4pm-thick frozen sections were cut, air-

I

Antibodies

used in the study

Antibody’

Antigen recognized

Reactivity on NHS’

Ref.

Ab-1 (M) 0KT9 (M) Ki-67 (M)

EGF-receptors Transferrin-receptors Nucleolar ag of cycling cells (late Gl/S/G,M) 56,5 Kd keratin Filaggrin Involucrin Human µglobulin

EK (pl. membr.) Some basal EK (pl. membr.) Basal EK (cyt) (cross-reactivity?)

9, 10 11,12 13, 14

Suprabasal EK (cyt) Granular layer EK (cyt) Upper malpighian/granular EK (pl. membrane)

15 16, 17 18, 19 20

cc-L-fucose

Upper layer EK (pl. membr.), endothelial cells Upper layer EK (pl. membr.)

KLl (M) BT-576 (M) BT-600 (P) IOT 2c (M) Lectins UEA-I PNA i M = monoclonal;

fi-D-galactose P = polyclonal;

2 NHS = normal human skin.

layer EK

21 21

105

dried, fixed in cold acetone and examined, using an amplification biotin-streptavidin-fluorescein technique for the expression of antigens listed in Table 1. Fluorescein-conjugated lectins (UEA-I and PNA) were also used in a direct immunofluorescence technique. For the immunolocalization of differentiation antigens on normal human skin an avidin-biotin-peroxidase technique was also used. Results

CsA levels and BrdU-labelling index Starting from the second week of the experiment CsA-administered mice developed macroscopically visible white hair mostly over the cephalic extremity and the back, thus proving that the drug had been absorbed and effectively reached the skin. As reported previously [8], whole blood levels of CsA measured two hours after the last injection were evaluated by a radioimmunologic assay to 679 + 501 ng/ml. CsA-treated NHSX had a LI decreased by 56% (P < 0.001) compared to control NHSX.

tently labelled, and in addition, a few suprabasal layers were also stained on some specimens. Overall, however, no obvious difference in the staining pattern between control and CsA- treated NHSX could be evaluated (Fig. lb,c). KLl. This antibody consistently labels suprabasal EK of NHS (Fig. 2a). In all NHSX examined, in addition to suprabasal cells, most basal EK were also labelled (Fig. 2b). However, in 3 out of 6 CsA-treated NHSX, the staining pattern was normal, i.e. basal EK remained unlabelled (Fig. 2~). Filaggrin. On NHS, the anti-filaggrin antibody labells the cytoplasm of stratum granulosum EK (Fig. 3a). On CsA-treated-and control-NHSX a similar staining pattern was obtained (Fig. 3b,c,). Involucrin. On NHS, the anti-involucrin antibody labells the cytoplasm of upper malpighian layer EK (Fig. 4a). The same staining pattern was obtained on control and CsA-treated NHSX (Fig. 4b,c).

Immunohistochemical study Epidermal Growth Factor-Receptors (EGF-R). On normal human skin (NHS) the anti-EGF-R antibody labelled the plasma membrane of EK, with an intensity decreasing from the basal-cell layer upwards. Control and CsA-treated NHSX exhibited a similar staining pattern although the labelling intensity was weaker than the one obtained on NHS.

Human fi,-microglobulin. On NHS, the anti-/I,microglobulin antibody recognizes the plasma membrane of EK (excluding the horny layer ones) (Fig. 5a). On both CsA-treated and control NHSX, a similar staining pattern was observed (Fig. 5b,c). The positive labelling furthermore confirmed that the examined epidermis was of human origin, since the adjacent mouse epidermis remained unstained.

Transferrin-receptors (TF-R). On NHS, the antiTF-R antibody occasionally labelled faintly small clusters of basal cells located over dermal papillae. No such specific labelling could be detected on either control or CsA-treated NHSX.

UEA-I lectin. On NHS, UEA-I labels the surface of upper-layer EK and of dermal-vessel endothelial cells (Fig. 6a). A similar staining pattern was obtained on CsA-treated and control NHSX (Fig. 6b,c). Noticeably, compared to NHSX, UEA-I staining of adjacent mouse skin was stronger in intensity in the epidermis but absent from endothelial cells.

Ki-67. On NHS, a cytoplasmic labelling of basal cells was observed (Fig. la). Despite thorough search, no nuclear/nucleolar specific staining could be seen. On control and CsA-treated NHSX, the cytoplasm of basal-cells was consis-

PNA lectin. On NHS, an intercellular staining pattern with a decreasing intensity towards the

06

107

lower epidermal layers was obtained. CsA-treated and control NHSX displayed a similar staining pattern, but the labelling generally extended further down in the epidermis compared to NHS. Discussion Xenografting full-thickness normal human skin (NHS) onto congenitally athymic (nude) mice constitutes an elegant model for studying kinetics, differentiation and pharmacological responses of normal epidermal keratinocytes (EK). Indeed, previous studies have shown that normal human skin xenografts (NHSX) maintain his(concerning immunohistochemical tologic, bullous pemphigoid and the 67 kD keratin antigens) as well as proliferation-response characteristics of human skin [22-241. Therefore, this system allows the long-term maintenance in vivo of human skin that resembles more closely NHS than other in vitro systems (EK or skin explant cultures) do. Furthermore, this model is suitable for pharmacological and toxicological studies that for ethical reasons cannot be carried out in humans. In the present study, we further confirm the overall maintenance of the immunohistochemical profile of NHS by NHSX, as evaluated by a wider panel of antigenic markers related to epiderma1 proliferation and differentiation (EGF-R, involucrin, filaggrin, j?,-microglobulin). The main difference between NHS and NHSX detected concerned the labelling of basal EK in NHSX by the antibody KLl. This antibody recognizes the 56.5 kD acidic keratin polypeptide (number 10 of the Moll catalog) expressed in NHS by suprabasal EK but also the 56 kD basic keratin polypeptide (number 6) characteristic of hyperproliferative epithelia, expressed in vitro by monolayers of cultured EK [25] and in vivo by basal EK in epidermal diseases (our unpublished data).

The reactivity of basal EK in NHSX with KLl is therefore in keeping with the contention that the epidermis of NHSX is in a hyperproliferative condition [ 241, that renders it more susceptible to the antiproliferative effect of CsA. Minor differences were obtained in the staining pattern for transferrin-receptors (being undetectable on NHSX) and for UEA-I and PNA-lectin binding sites. The labelling observed with the Ki-67 monoclonal antibody is intriguing. Indeed, Ki-67 recognizes an as yet biochemically-undetined nucleolar antigen expressed by cycling (late G,/S/G, M) cells [ 131, and is being used in the in situ immunohistochemical evaluation of the proliferative fraction of haemopoietic malignancies [ 141. The labelling of basal EK obtained on NHS and NHSX (mentioned briefly for basal cells of the oral mucosa in ref. number 13) may be due in fact to a cross-reactivity with a cytoplasmic antigen, the nature of which remains to be elucidated. CsA appears to exert a wide range of biological activities that extend far beyond the immunosuppressive one. However, its precise mechanism of action remains elusive l-261. Previous studies carried out in order to investigate the expression of activation antigens showed that CsA, while having little if any effect on unstimulated T-cells, is able to prevent the lectin-triggered expression of transferrin-receptors and of HLA-class II antigens [ 27,281. In a previous work we reported that the in vivo treatment of NHSX with CsA does not modify the number or the distribution of epiderma1 CDla + /HLA-DR + Langerhans’ cells, thus suggesting that the baseline expression of HLA-DR antigens by these cells remains unchanged [29]. Our present work suggests that CsA does not modify either the expression of the light chain of HLA-class I molecules (p,-microglobulin) by EK. Nowadays it has become clear that CsA re-

Fig. 1-6. Expression of various proliferation- and differentiation-associated antigens on normal human skin (a), control normal human skin-xenografted onto nude mice (b) and normal human skin xenografted onto nude mice treated with 50 mg/kg Cyclosporin A for three weeks (c). 1: Ki-67,2: KLI, 3: Filaggrin, 4: Involucrin, 5: j3,-microglobulin, 6: UEA I lectin. (the dotted lines underline the dermal-epidermal junction) (sections immunolabelled by a streptavidin-biotin-fluorescein technique, except from sections shown on Fig. 2, 3, 4, and 6a immunolabelled by an avidin-biotin-peroxidase technique).

108

duces the proliferative capacity of interfollicular EK both in vitro and in vivo through an as yet unravelled mechanism [ 301. In order to investigate the effects of CsA on human skin we chose the system of xenografting of NHS onto nude mice since, as stated above, the skin obtained through this technique closely resembles, in terms of differentiation, NHS. Furthermore in this model, thanks to a deficiency of T-cell-mediated immunity, any possible immunologically-conveyed effects of CsA on EK proliferation and differentiation are circumvented. A three-week CsA-administration period was applied since by that time hypertrichosis developing over the mouse skin provided unequivocal evidence that the drug had not only reached the skin but also exerted its best-known cutaneous biological sideeffect. With respect to CsA administration, the daily dosage of 50 mg/kg may at first glance seem very high compared to the one used therapeutically in humans (5 mg/kg). However, under the conditions of our study, whole-blood CsA levels averaged 679 ng/ml, i.e. they were not very much higher than the ones considered therapeutically effective in humans (200-400 ng/ml), most likely due to the fact that CsA absorption is poorer after subcutaneous injection than after oral administration. In spite of the fact that the immunohistochemical approach to study the effect of CsA on epidermal differentiation is only qualitative and not sensitive enough to detect subtle changes, we were able to observe that CsA tends to normalize the (aberrant) KLl-positivity of basal EK of control NHSX. This finding suggests that CsA, while able to decrease EK proliferation may also influence (normalize) the process of epidermal differentiation (keratinization). This activity merits, in our opinion, further study. Indeed, despite the reported failure of CsA to improve patients with lamellar ichthyosis [ 311, an earlier observation had reported a beneficial effect of CsA in a case of ichthyosis vulgaris [32], a disease due to a genetically-determined disorder of normal epidermal differentiation. With regard to the remaining markers, no obvious influence of CsA treatment could be detected. Noteworthy is the

fact that recently CsA was reported to decrease in vitro the number of 1251-EGF-binding sites on cultured human EK [ 51; this finding, if confirmed in vivo, may account for the antiproliferative activity of CsA. The fact that in our study no clearcut difference in EGF-R expression between control and CsA-treated NHSX could be detected may be due either to the lower sensitivity of our methodology or to the fact that the reduction of ‘251-EGF-binding sites is not actually due to a decreased expression of immunologically active EGF-R. In conclusion, the results of our study suggest that CsA may have an effect on EK differentiation. This activity is not totally unexpectable, since it is known that agents that have an effect on EK proliferation (such as retino’ids or calcium salts) simultaneously modulate EK differentiation. Although our data need confirmation, we feel they suggest the interest in pursuing studies concerning the direct effect of CsA on both healthy and diseased human epidermis. Acknowledgements

We thank manuscript.

Valerie

Rattaire

for typing

the

References Nickoloff B, Fisher G, Mitra R, Voorhees J: Additive and synergistic antiproliferative effects of cyclosporin A and gamma-interferon on cultured human keratinocytes. Am J Path01 131, 12-18, 1988. Fisher G, Duel1 E., Nickoloff B, Annesley T, Kowalke J, Ellis C, Voorhees J: Levels of cyclosporin in epidermis of treated psoriasis patients differentially inhibit growth of keratinocytes cultured in serum free versus containing media. J Invest Dermatol 91, 142-146, 1988. Fairley J, Fing P, Ewing N: Calmodulin content of keratinocytes in vitro: effect of proliferation and cyclosporin A (abstr). J Invest Dermatol 92, 426, 1989. Dykes P, Brunt J, Marks R: Effect of cyclosporin on human epidermal cells in vitro (abstr). J Invest Dermatol 92, 423, 1989. Mitra R, Voorhees J, Nickoloff B: Modulation of 1251EGF ligand binding to cultured keratinocytes (abstr). J Invest Dermatol 92, 482, 1989. Ramirez-Bosca A, Kanitakis J, Haftek M, Fame M,

109 Castells-Rodellas A, Thivolet J: Effects of Cyclosporin A on cultured human epidermal cells: growth and 5-bromo2’-deoxyuridine incorporation. Acta Derm Venereol 70, 6-10, 1990. 7 Furue M, Gaspari A, Katz S: The effect of cyclosporin A on epidermal cells. II. Cyclosporin A inhibits proliferation of normal and transformed keratinocytes. J Invest Dermatol 90, 796-800, 1988. 8 Urabe A, Kanitakis J, Viac J, Thivolet J: Cyclosporin A inhibits directly in vivo keratinocyte proliferation ofliving human skin. J Invest Dermatol 92, 755-757, 1989. 9 Kawamoto T, Sato TD, Le AD, Polikoff J, Sato GH, Mendelsohn J: Growth stimulation of A43 1 cells by epidermal growth factor: identification of high affinity receptor for epidermal-growth-factor by an anti-receptor monoclonal antibody. Proc Nat1 Acad Sci (USA) 80, 1337-1341, 1983. 10 Nanney LB, McKanna JA, Stoscheck CM, Carpenter G, King EL: Visualization of Epidermal Growth Factor receptors in human epidermis. J Invest Dermatol 82, 165-169, 1984. 11 Judd W, Poodry CA, Strominger JL: Novel surface antigen expressed on dividing cells but absent from nondividing cells. J Exp Med 152, 1430-1436, 1980. 12 Soyer Hp, Smolle J, Torne R, Kerl H: Transferrin receptor expression in normal skin and in various cutaneous tumors. J Cutan Path01 14, 1-5, 1987. 13 Gerdes J, Lemke H, Baisch H, Wacker HH, Schwab U, Stein H: Cell-cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J Immunol 133, 1710-1715, 1984. 14 Raltkiaer E, Stein H, Bosq J: Expression of a cell-cycleassociated nuclear antigen (Ki-67) in cutaneous lymphoid infiltrates. Am J Dermatopathol 8, 37-43, 1986. 15 Viac J, Reano A, Brochier J, Staquet MJ, Thivolet J: Reactivity pattern of a monoclonal antikeratin antibody (KLl). J Invest Dermatol 81, 351-354, 1983. 16 Dale BA, Gown AM, Fleckman P, Kimball JR, Resing KA: Characterization of two monoclonal antibodies to human epidermal keratohyalin: reactivity with tilaggrin and related proteins. J Invest Dermatol 88, 306-312, 1987. 17 Kanitakis J, Ramirez-Bosca A, Reano A, Viac J, Roche P, Thivolet J: Filaggrin expression in normal and pathological skin. A marker of keratinocyte differentiation. Virchows Arch A (Path01 Anat) 412, 375-382, 1988. 18 Rice R, Green H: Presence in human epidermal cells of a soluble protein precursors of the cross-linked envelope: activation of the cross linking by calcium ions. Cell 18, 681-694, 1979.

19 Murphy G, Flynn T, Rice R, Pinkus G: Involucrin expression in normal and neoplastic human skin: a marker for keratinocyte differentiation. J Invest Dermatol 82, 453-457, 1984. 20 LiabeufA, Le Borgne de Kaouel C, Kourilsky F, Malissen B, Manuel Y, Sanderson AR: An antigenic determinant of human µglobulin masked by the association with HLA heavy chains at the cell surface: analysis using monoclonal antibodies. J Immunol 127, 1542-1548,198l. 21 Reano, A, Faure M, Jacques Y, Reichert U, Schaefer H, Thivolet J: Lectins as markers of human epidermal cell differentiation. Differentiation 22, 205-210, 1982. 22 Reed ND, Manning DD: Long-term maintenance of normal human skin on congenitally athymic (nude) mice. Proc Sot Exp Biol Med 143, 350-356, 1973. 23 Krueger GG, Selby J: Biology ofhuman skin transplanted to the nude mouse: I. Response to agents which modify epidermal proliferation. J Invest Dermatol 76, 506-510, 1981. 24 Haftek M, Ortonne JP, Staquet MJ, Viac J, Thivolet J: Normal and psoriatic human skin grafts on “nude” mice: morphological and immunohistochemical studies. J Invest Dermatol 76, 48-52, 1981. 25 Regnier M, Schweizer J, Michel S, Bailly C, Prunieras M: Keratin in human keratinocytes cultured on dead de-epidermized dermis. Exp Cell Res 165, 63-72, 1986. 26 Bore1 JF: Basic Science Summary, In: Cyclosporine. Nature of the agent and its immunologic actions. Edited by B Kahan. Grune and Stratton, Philadelphia, 1988, pp 722-730. 27 Leapman S, Strong D, Filo D, Smith E, Brandt GL: Cyclosporin A prevents the appearance of cell surface ‘activation’ antigens. Transplantation 34, 94-95, 1982. 28 Miyawaki T, Yachie A, Ohzeki S, Nagaoki T, Taniguchi N: Cyclosporin A does not prevent expression of Tat antigen, a probable TCGF receptor molecule, on mitogen-stimulated human T-cells. J Immunol 130, 2742-2757, 1983. 29 Urabe A, Haftek M, Kanitakis J, Schmitt D, Thivolet J: Cyclosporin A does not modify Langerhans cell number and distribution in normal human skin. Acta Derm Venereol (Stockh) 69, 249-252, 1989. 30 Kanitakis J, Thivolet J: Cyclosporin A: an immunosuppressant affecting epithelial cell proliferation. Arch Dermatol (in press). 31 Ho V, Gupta A, Ellis C, Cooper K, Nickoloff B, Voorhees J: Cyclosporine in lamellar ichthyosis. Arch. Dermatol 125, 511-514, 1989. 32 Velthuis P, Jesserun R: Improvement of ichthyosis by cyclosporin. Lancet i 335, 1985.