Scand. J. Imtnunol. 35, 321-325, 1992
la Expression in Keratinocytes Following Ultraviolet Radiation A. GILHAR, A. ETZIONI & S. E I D E L M A N Skin Research Laboratory, Faculty of Medicine., Technion-Israel Institute of Technology, Haifa, Israel
Gilhar A. Etzioni A, Fidelman S. la Expression in Keratinocytes Following Ultraviolet Radiation-Scand J Immunol 1992:35:321-5 Injections of murine gamma interferon (IFN-7) into BALB.c nude mice induced la expression by keratinocytes. The aim of the present study was to use this murine model to determine the effect of ultraviolet radiation (UV) or cyclosporine A (CyA) on la expression by keratinocytes. Two sets of experiments were performed. In the first, mice were injected intraperitoneally with IFN-y for 6 days. The mice were divided into three groups. One group, wilh one ear protected by electrical tape, was exposed to UVB radiation for 15 days starting 4 days before the injection. The second group received subcutaneous injections of CyA simultaneously with the IFN-7 and during the 10 days following the IFN-;' injection. The third group received only IFN-y injections. Fifteen days after the IFN-7 injection all mice were killed and evaluations of la positive cells were performed. In the second set of experiments the nude mice were treated with CyA or UVB only 10 days after the last IFN--; injections. In both experiments UVB inhibited and down-regulated la expression by keratinocyles. This effect on keratinocytes was not observed in the protected ears. Thus it appears that the effect of UVB on keratinocytes is local and nol systemic. CyA failed to inhibit or down-regulate la expression. This study may shed some light on understanding the mechanism effect of UV radiation in a variety of skin diseases. Anw.s Gilhar MD. Skin Research Lahoratorv. Faculty af Medicine. Teelinion. P.O.B. 9649. Haifa 31096. Israel
la antigens (la antigeti in mouse, HLA-DR in human) are polymorphic membrane glycoproteins coded by genes within the major histocompatibility complex (MHC). The expression of la antigens on a wide variety of immunological cells suggests that these antigens have a clear association with the immune system [I]. Non-lymphoid cells of various organs can be induced to express la antigen [2, 3]. It has been shown that in graft versus host disease (GVHD) la antigens appeared on epidermal cells [4]. A variety of immunological stimuli can also induce the expression of la antigen at these sites [5]. la positive keratinocytes occur frequently in a variety of skin diseases characterized by lymphoid infiltration within the upper dcrmis [6-9]. Most of these diseases respond well to ultraviolet radiation (UV)[ 10-14] and cyclosporine A (CyA) therapy [15, 16]. Recently, Jcphthah-Ocholaf?c;/. [17] have shown that systemic treatment with
CyA inhibits the induction oC la antigen by nude and SCID mice injected with bacterial lipopolysaccharidcs. The ituthors speculated that inhibition of NK-cell-like activity by CyA prevents the induction of la antigen. We have reported similar results using normal mouse serum instead of bacterial lipopolysaccharidcs as the inducer of la expression [18], Jun et al. [19] demonstrated that injections of murine gamma interferon into nude mice induced la expression by keratinocytes and enterocytes. The aim of the present study was to use this murine model to determine whether UVB or CyA alters expression of la in keratinocytes. Two specific questions were asked: (I) do UVB or CyA inhibit the induction of la expression in keratinocytes. and (2) do UVB or CyA down-regulate keratinocyte expression of la antigen in nude mice that were previously injected with gamma interferon (IFN-7)? 321
322
A. Githaretat.
MATERIALS AND METHODS Animals. Thirty-five mice. 2 3 monihs of age. were obtained from the pathogen-free auima! racilily al the Faeulty of Medicine. Tech n ion-Israel Institute of Technology. Haifa, Ail mice were injected intraperitoneally with recombinant murine IFN-y (Gen/yme BiochemicaKsLtd. UK).5xlO^U/IOOmlPBS(O,l ml/day), every day. for six consecutive days, Vltraiiolet radiation. The nude mice were UV radiated under a bank of Westinghouse FS-40 sun lamps emitting between 2S0 and 320 nm, (Previous studies in the laboratory have shown depletion of la positive Langerhans cells (LC) following UV radiation of 740 mJ/cm'/day for four consecutive days). The distance between the light source and the mice was 25 cm, Cyclosporine A (CyA). CyA (Sandimmune oral suspension, 100 mg/ml: Sandoz, Basel. Switzerland) was diluted in olive oil and a dosage of 50 mg. kg was injected daily subcutaneously on the back area. Evaluation of la positive cells. This technique was performed as previously deseribed [20], Far skin biopsies were taken from the mice. Dorsal and ventral ear skin was separated mechanically with forceps and incubated in a phosphate-buffered EDTA solution as described by Juhhn & Shelley |2[i. After 2 h at 37 C, epidermal sheets were peeled off the dermis and cul into 4 x 4 mm pieces. The epidermal pieces were fi,xed in acetone, rehydraled with PBS and ineubated for I h with tissue culture medium (control) or appropriate monoclonal antibody containing hybridoma tissue eulture supernatant. Tissue bound antibody was assessed by an indireet immunopero,\idase method employing the Vectastain Avidin-biolLii immunoperoxidase staining procedures and reagents (Vector Laboratories Inc. Burlingame, Calif,, USA) with a 3-amino-9ethylcarba7ol as the developing substrate. Monoclonal aniibodies MKD6 (Beeton-Dickinson, Calif,, USA) with specilicity for I-A'' (expressed by BALB/c) and MCA-179 lor I-E (Ia-m'' correlates to conventional la'; Serotec Kidlington, Oxford. UK) were used to identify la positive cell populations. Stained sections were mounted on slides in glycerol and examined by light microscopy. Sections were evaluated for keratinoeyle expression of la and the presence of LC, Patterns of keralinoeyte la staining were recorded as either focal, where patches of ia posilive keratinocytes were observed lo cover at least 50",, ofthe epidermal sheet, or dilfuse, where the keratinocyles throughout the entire epidermal sheet expressed la. The percentage of la positive eells was determined by counling 20 random fields per biopsy under a 45 objective with the use of an ocular grid of known area. The number of ATPase positive cells was determined as described by Jhulin & Shelley [211, Briefly, the epidermal sheets were washed in cold cacodylate formaldehyde pH 4 for 20 min at 4 C, They were then incubated for 20 min in a substrale composed of II) ntg ATP (Sigma Chetnical Co,. Si, Louis, Mo,. USA). 5 tni 5% MgSOa, 3 ml 2% PBNO., in 42 ml trismal buffer. pH 7,3, Thereafter the sheets were washed with trismal buffer and treated with a 5"';. ammonium sulphide solution for 10 min. The mean
nutnber of Ia and ATPase positive dendritie eells per unit was determined by counling five random fields per biopsy. All slides were examined for the number of ATPase positive cells, la positive LC and Ia positive keratinocytes by an observer who was unaware of ihe serum injeeted and type of treatmeni given to the tnice. Student's i test was used for comparison of groups. E.xpcriments. The purpose of the lirst set of experiments was to determine whether UV versus CyA is capable of inhibiting the induction of la positive keratinocytes in nude mice. Fifteen mice were divided into ihree groups. The first group of mice, with one ear protected by electrical tape, were exposed to UV from day I until day 15, From day 1 lo day 4 the dosage was 740 mj/cm-.tiay, whereas from day 4 to day 15 ihe dosage was 369 mj/cm'- From day 4 until day 10 the animals were injected with IFN-y for six eonseculive days to induce keratinoeyle expression of la. The second group of miee was injected with IFN-y from day 1 until day 6, From day I until day 15 the miee were injected with CyA, The third group received only injections of lFN-yand served as la positive induction eontrols. On day 15 epidermal sheets ofall three groups were analysed for la expression. The second set of experiments was designed to determine whether U VB or CyA is capable of down-regulating keratinocyte expression of la. To address ihis question, flve nude mice were injoeted again with IFN--; for six consecutive days, but were exposed lo UVB (369 mJ'cnr/day) only from day 16 (i.e, 10 days after the last IFN-y injeetion). until day 21, On day 21 epidermal sheets were analysed for la expression, Fiveother animals treated with IFN-yfrom day I lo day 6 received injection of CyA from day 16 until day 26, On day 26 epidermal sheets were analysed for Ia expression. Five nude mice whieh were neither exposed to UVB nor treated with CyA served as la posilive induction controls. Five other nude mice were injected with sterile waler inslead of IFN-y and were no! exposed lo UVB or treated with CyA,
RESULTS In the first set of experiments expression of Ia on keratinocytes was observed in UVB protected ears and non-UVB-ireated tnice, as well as in the CyA-lrcaled group. However, no Ia expression was observed in the unprotected ears of UVBexposed animals (Table I), Only a few Ia positive LC were noted through the epidermal sheets. In the CyA-treated group. 2 out of 5 mice showed a focal pattern of la expression in keratinocytes (Fig, I), whereas the retnaining three mice showed a diffuse pattern of staining (Fig, 2), The results of this experiment suggest that UVB, in contrast to CyA. inhibits the induction of Ia expression in keratinocytes. In the second set ofexperiments, la expression was observed in the keratinocytes of
Response of la Aniigen to UV Radiation TABLF I- UVB inhibition of Ihe induction of la expression by keratinocytes in nude mice
323
as compared wilh the non-exposed mice (Table II).
Keratinocyte expression of la
DISCUSSION Taped
-t
IFN-;
UVB CyA
D"
Total 5/5 0/5
-l-t * D = difTuse; ** P = patchy. t One ear of ihe UVB exposed mice was protected with electrical tape, whereas the second ear was exposed.
all miee except those that were exposed lo UVB (Table II). Only one of Ihe five animals that were exposed to UVB showed la positive keralinoeytes, with focal staining only in epidermal sheets. In the CyA-lreated group. 4 out of 5 mice showed diffuse staining of la whereas one mouse had focal staining. No la positive keratinocytes were observed in the animals thai were injected with sterile water instead of !FN-y. This experiment suggests that UVB may down-regulate keratinocyte expression of la antigen in nude mice. A significant reduction of LC population was observed in the epidermis of UVB-exposed mice.
In recent years the importance of la expression in various tissues in the pathogenesis of several autoimmune conditions has emerged [22]. Therefore, elass II positive epithelial eells may be involved in certain immunological skin disorders. We have previously described an enhanced expression of la antigen in keratinocyles following injection of normal mouse serum (NMS)[18]. We and others have postulated that this effect was due to production of IFN-y from NK-like cells whieh exist in the nude mouse [17, 18]. When we looked in our model at the effect of CyA, using NMS, we found an inhibition of la expression. We were unable, however, to conclude whether it was due to inhibition of the production of IFN-7 or a direct effect on this cytokine. In the present study we used IFN-y directly to induce la expression and as no change was observed whether or not CyA was given, we can conclude that CyA exerts its effecl before IFN-y is produced, but plays only a minor role once it is produced. As many skin disorders whieh are associated with la-like positive keratinocytes respond well to
FIG. t. Indirect immunopcroxidase staining showing focal pattern of la-like positive keratinocytes (arrow) and numerou:^ Langerhans cells ( x 120).
324
A. Gilhar et al.
FIG, 2. Diffuse pattern of la-likc positive keratinocytes ( x 120).
TABLE II. UVB down-regulalion of kcralinocyte expression ^}i la in nude mice la in keratinocytes
IFN- V UVB CyA D+ + -1- 1
+ -
+
5 0 4 0
p.*
Total
0 I 1 0
5/5 1/5 5/5 0/5
lat LC
ATPase LC
786 ±124 98126"
925±113
+
847+137
178 + 28^ 1023+197 984+t21
* D = diffuse; • * P = patchy. t = Because of la expression by Ihe keratinocyles il was impossible to detect la positive Langerhans cetts (LC). " = P
la Expression in Keratinocytes Following Ultraviolet Radiation A. GILHAR, A. ETZIONI & S. E I D E L M A N Skin Research Laboratory, Faculty of Medicine., Technion-Israel Institute of Technology, Haifa, Israel
Gilhar A. Etzioni A, Fidelman S. la Expression in Keratinocytes Following Ultraviolet Radiation-Scand J Immunol 1992:35:321-5 Injections of murine gamma interferon (IFN-7) into BALB.c nude mice induced la expression by keratinocytes. The aim of the present study was to use this murine model to determine the effect of ultraviolet radiation (UV) or cyclosporine A (CyA) on la expression by keratinocytes. Two sets of experiments were performed. In the first, mice were injected intraperitoneally with IFN-y for 6 days. The mice were divided into three groups. One group, wilh one ear protected by electrical tape, was exposed to UVB radiation for 15 days starting 4 days before the injection. The second group received subcutaneous injections of CyA simultaneously with the IFN-7 and during the 10 days following the IFN-;' injection. The third group received only IFN-y injections. Fifteen days after the IFN-7 injection all mice were killed and evaluations of la positive cells were performed. In the second set of experiments the nude mice were treated with CyA or UVB only 10 days after the last IFN--; injections. In both experiments UVB inhibited and down-regulated la expression by keratinocyles. This effect on keratinocytes was not observed in the protected ears. Thus it appears that the effect of UVB on keratinocytes is local and nol systemic. CyA failed to inhibit or down-regulate la expression. This study may shed some light on understanding the mechanism effect of UV radiation in a variety of skin diseases. Anw.s Gilhar MD. Skin Research Lahoratorv. Faculty af Medicine. Teelinion. P.O.B. 9649. Haifa 31096. Israel
la antigens (la antigeti in mouse, HLA-DR in human) are polymorphic membrane glycoproteins coded by genes within the major histocompatibility complex (MHC). The expression of la antigens on a wide variety of immunological cells suggests that these antigens have a clear association with the immune system [I]. Non-lymphoid cells of various organs can be induced to express la antigen [2, 3]. It has been shown that in graft versus host disease (GVHD) la antigens appeared on epidermal cells [4]. A variety of immunological stimuli can also induce the expression of la antigen at these sites [5]. la positive keratinocytes occur frequently in a variety of skin diseases characterized by lymphoid infiltration within the upper dcrmis [6-9]. Most of these diseases respond well to ultraviolet radiation (UV)[ 10-14] and cyclosporine A (CyA) therapy [15, 16]. Recently, Jcphthah-Ocholaf?c;/. [17] have shown that systemic treatment with
CyA inhibits the induction oC la antigen by nude and SCID mice injected with bacterial lipopolysaccharidcs. The ituthors speculated that inhibition of NK-cell-like activity by CyA prevents the induction of la antigen. We have reported similar results using normal mouse serum instead of bacterial lipopolysaccharidcs as the inducer of la expression [18], Jun et al. [19] demonstrated that injections of murine gamma interferon into nude mice induced la expression by keratinocytes and enterocytes. The aim of the present study was to use this murine model to determine whether UVB or CyA alters expression of la in keratinocytes. Two specific questions were asked: (I) do UVB or CyA inhibit the induction of la expression in keratinocytes. and (2) do UVB or CyA down-regulate keratinocyte expression of la antigen in nude mice that were previously injected with gamma interferon (IFN-7)? 321
322
A. Githaretat.
MATERIALS AND METHODS Animals. Thirty-five mice. 2 3 monihs of age. were obtained from the pathogen-free auima! racilily al the Faeulty of Medicine. Tech n ion-Israel Institute of Technology. Haifa, Ail mice were injected intraperitoneally with recombinant murine IFN-y (Gen/yme BiochemicaKsLtd. UK).5xlO^U/IOOmlPBS(O,l ml/day), every day. for six consecutive days, Vltraiiolet radiation. The nude mice were UV radiated under a bank of Westinghouse FS-40 sun lamps emitting between 2S0 and 320 nm, (Previous studies in the laboratory have shown depletion of la positive Langerhans cells (LC) following UV radiation of 740 mJ/cm'/day for four consecutive days). The distance between the light source and the mice was 25 cm, Cyclosporine A (CyA). CyA (Sandimmune oral suspension, 100 mg/ml: Sandoz, Basel. Switzerland) was diluted in olive oil and a dosage of 50 mg. kg was injected daily subcutaneously on the back area. Evaluation of la positive cells. This technique was performed as previously deseribed [20], Far skin biopsies were taken from the mice. Dorsal and ventral ear skin was separated mechanically with forceps and incubated in a phosphate-buffered EDTA solution as described by Juhhn & Shelley |2[i. After 2 h at 37 C, epidermal sheets were peeled off the dermis and cul into 4 x 4 mm pieces. The epidermal pieces were fi,xed in acetone, rehydraled with PBS and ineubated for I h with tissue culture medium (control) or appropriate monoclonal antibody containing hybridoma tissue eulture supernatant. Tissue bound antibody was assessed by an indireet immunopero,\idase method employing the Vectastain Avidin-biolLii immunoperoxidase staining procedures and reagents (Vector Laboratories Inc. Burlingame, Calif,, USA) with a 3-amino-9ethylcarba7ol as the developing substrate. Monoclonal aniibodies MKD6 (Beeton-Dickinson, Calif,, USA) with specilicity for I-A'' (expressed by BALB/c) and MCA-179 lor I-E (Ia-m'' correlates to conventional la'; Serotec Kidlington, Oxford. UK) were used to identify la positive cell populations. Stained sections were mounted on slides in glycerol and examined by light microscopy. Sections were evaluated for keratinoeyle expression of la and the presence of LC, Patterns of keralinoeyte la staining were recorded as either focal, where patches of ia posilive keratinocytes were observed lo cover at least 50",, ofthe epidermal sheet, or dilfuse, where the keratinocyles throughout the entire epidermal sheet expressed la. The percentage of la positive eells was determined by counling 20 random fields per biopsy under a 45 objective with the use of an ocular grid of known area. The number of ATPase positive cells was determined as described by Jhulin & Shelley [211, Briefly, the epidermal sheets were washed in cold cacodylate formaldehyde pH 4 for 20 min at 4 C, They were then incubated for 20 min in a substrale composed of II) ntg ATP (Sigma Chetnical Co,. Si, Louis, Mo,. USA). 5 tni 5% MgSOa, 3 ml 2% PBNO., in 42 ml trismal buffer. pH 7,3, Thereafter the sheets were washed with trismal buffer and treated with a 5"';. ammonium sulphide solution for 10 min. The mean
nutnber of Ia and ATPase positive dendritie eells per unit was determined by counling five random fields per biopsy. All slides were examined for the number of ATPase positive cells, la positive LC and Ia positive keratinocytes by an observer who was unaware of ihe serum injeeted and type of treatmeni given to the tnice. Student's i test was used for comparison of groups. E.xpcriments. The purpose of the lirst set of experiments was to determine whether UV versus CyA is capable of inhibiting the induction of la positive keratinocytes in nude mice. Fifteen mice were divided into ihree groups. The first group of mice, with one ear protected by electrical tape, were exposed to UV from day I until day 15, From day 1 lo day 4 the dosage was 740 mj/cm-.tiay, whereas from day 4 to day 15 ihe dosage was 369 mj/cm'- From day 4 until day 10 the animals were injected with IFN-y for six eonseculive days to induce keratinoeyle expression of la. The second group of miee was injected with IFN-y from day 1 until day 6, From day I until day 15 the miee were injected with CyA, The third group received only injections of lFN-yand served as la positive induction eontrols. On day 15 epidermal sheets ofall three groups were analysed for la expression. The second set of experiments was designed to determine whether U VB or CyA is capable of down-regulating keratinocyte expression of la. To address ihis question, flve nude mice were injoeted again with IFN--; for six consecutive days, but were exposed lo UVB (369 mJ'cnr/day) only from day 16 (i.e, 10 days after the last IFN-y injeetion). until day 21, On day 21 epidermal sheets were analysed for la expression, Fiveother animals treated with IFN-yfrom day I lo day 6 received injection of CyA from day 16 until day 26, On day 26 epidermal sheets were analysed for Ia expression. Five nude mice whieh were neither exposed to UVB nor treated with CyA served as la posilive induction controls. Five other nude mice were injected with sterile waler inslead of IFN-y and were no! exposed lo UVB or treated with CyA,
RESULTS In the first set of experiments expression of Ia on keratinocytes was observed in UVB protected ears and non-UVB-ireated tnice, as well as in the CyA-lrcaled group. However, no Ia expression was observed in the unprotected ears of UVBexposed animals (Table I), Only a few Ia positive LC were noted through the epidermal sheets. In the CyA-treated group. 2 out of 5 mice showed a focal pattern of la expression in keratinocytes (Fig, I), whereas the retnaining three mice showed a diffuse pattern of staining (Fig, 2), The results of this experiment suggest that UVB, in contrast to CyA. inhibits the induction of Ia expression in keratinocytes. In the second set ofexperiments, la expression was observed in the keratinocytes of
Response of la Aniigen to UV Radiation TABLF I- UVB inhibition of Ihe induction of la expression by keratinocytes in nude mice
323
as compared wilh the non-exposed mice (Table II).
Keratinocyte expression of la
DISCUSSION Taped
-t
IFN-;
UVB CyA
D"
Total 5/5 0/5
-l-t * D = difTuse; ** P = patchy. t One ear of ihe UVB exposed mice was protected with electrical tape, whereas the second ear was exposed.
all miee except those that were exposed lo UVB (Table II). Only one of Ihe five animals that were exposed to UVB showed la positive keralinoeytes, with focal staining only in epidermal sheets. In the CyA-lreated group. 4 out of 5 mice showed diffuse staining of la whereas one mouse had focal staining. No la positive keratinocytes were observed in the animals thai were injected with sterile water instead of !FN-y. This experiment suggests that UVB may down-regulate keratinocyte expression of la antigen in nude mice. A significant reduction of LC population was observed in the epidermis of UVB-exposed mice.
In recent years the importance of la expression in various tissues in the pathogenesis of several autoimmune conditions has emerged [22]. Therefore, elass II positive epithelial eells may be involved in certain immunological skin disorders. We have previously described an enhanced expression of la antigen in keratinocyles following injection of normal mouse serum (NMS)[18]. We and others have postulated that this effect was due to production of IFN-y from NK-like cells whieh exist in the nude mouse [17, 18]. When we looked in our model at the effect of CyA, using NMS, we found an inhibition of la expression. We were unable, however, to conclude whether it was due to inhibition of the production of IFN-7 or a direct effect on this cytokine. In the present study we used IFN-y directly to induce la expression and as no change was observed whether or not CyA was given, we can conclude that CyA exerts its effecl before IFN-y is produced, but plays only a minor role once it is produced. As many skin disorders whieh are associated with la-like positive keratinocytes respond well to
FIG. t. Indirect immunopcroxidase staining showing focal pattern of la-like positive keratinocytes (arrow) and numerou:^ Langerhans cells ( x 120).
324
A. Gilhar et al.
FIG, 2. Diffuse pattern of la-likc positive keratinocytes ( x 120).
TABLE II. UVB down-regulalion of kcralinocyte expression ^}i la in nude mice la in keratinocytes
IFN- V UVB CyA D+ + -1- 1
+ -
+
5 0 4 0
p.*
Total
0 I 1 0
5/5 1/5 5/5 0/5
lat LC
ATPase LC
786 ±124 98126"
925±113
+
847+137
178 + 28^ 1023+197 984+t21
* D = diffuse; • * P = patchy. t = Because of la expression by Ihe keratinocyles il was impossible to detect la positive Langerhans cetts (LC). " = P
