JOURNAL OF PATHOLOGY, VOL.
165: 129-136 (1991)
MHC ANTIGEN EXPRESSION IN HUMAN ORAL SQUAMOUS CARCINOMA CELL LINES SERDAR MUTLU*, JOHN B. MATTHEWST, MARSH MIDDA*, CRISPIAN SCULLY* AND STEPHEN
s. PRIME*
*Department of Oral Medicine, Surgery and Pathology, University of Bristol, U.K.; ?Department of Oral Pathology, University of Birmingham, U.K. Received 14 January 1991 Accepted 18 April 1991
SUMMARY This study quantified the constitutive and interferon-gamma (IFN-y) stimulated expression of MHC class I (HLAABC and 8,microglobulin) and class I1 antigens (HLA-DR, -DP, -DQ) on normal and malignant oral keratinocytes using radioimmunoassay and immunocytochemical techniques. Normal keratinocytes and three of four malignant cell lines (H 103, HI 57, H314) expressed MHC class I antigens constitutively; IFN-y increased MHC class I expression with significant changes in normals, H 157 and H314. Normal keratinocytes expressed significantly more constitutive MHC class I antigens than H 103 and H 157 and significantly more IFN-y stimulated MHC class I antigens than H103, H 157 and H3 14. MHC class I1 antigens predominantly were not expressed constitutively on normals, H 103 and H 157 but, in H314, HLA-DR, -DP and -DQ antigens were demonstrated on 35, 11 and 5 per cent of cells, respectively, and resulted in a non-coordinated pattern ofexpression (HLA-DR> -DP= -DQ). IFN-y induced HLA-DR on normals, HI03 and H157, whilst HLA-DP and -DQ remained undetectable. In H314, IFN-y enhanced HLA-DR, -DP and -DQ (significant increase of HLA-DQ) but the interrelationship between these antigens was maintained (HLA-DR > -DP = -DQ). Normal keratinocytes expressed significantly more IFN-y stimulated HLA-DR than H 103 and H 157 but significantly less HLA-DR than H314 under similar experimental conditions. One oral malignant cell line (H191) did not express MHC class I and MHC class I1 antigens either constitutively or in response to IFN-y. The results demonstrate aberrant patterns of MHC expression (absence, enhanced, diminished) in the different malignant oral keratinocyte cell lines. KEY WORDS-MHC,
antigens, human, oral, carcinoma, cell lines.
INTRODUCTION A number of different factors are involved in tumour-immune system interactions. The recognition of major histocompatibility complex (MHC) antigens on tumour cells by cell-mediated immune mechanisms is now known to be of critical importance in the early events of neoplasia.’ MHC genes in humans are located on the short arm of chromosome 6 and encode both MHC class I and class IT antigens. MHC class I antigens consist of a heavy chain (HLA-A, -B, -C) linked noncovalently to &-microglobulin (encoded by a gene on chromosome 15) and are expressed predominantly by all cell types.’ MHC class I1 antigens (HLA-DR, -DP, -DQ) consist of two chains, a heavy (a) and a light @) chain and, in healthy tissues, Addressee for correspondence:Dr S. S. Prime, Department of Oral Medicine, Surgery and Pathology, Bristol Dental Hospital, Lower Maudlin Street, Bristol BS1 2LY, U.K.
0022-341 71’9ljl00 1 2 9 4 8 $05.00 0 1991 by John Wiley & Sons, Ltd.
are mainly expressed by the cells of the immune system (antigen-presentin cells, B lymphocytes, activated T lymphocytes). MHC class I antigens serve as restriction elements for T-cell-mediated cytot~xicity,~ whilst MHC class TI molecules are required for the presentation of antigen to helper T cells (HLA-DR),’ serve as up-regulators of cytotoxic T lymphocytes (CTL) (HLA-DP),6 and modulate indirectly the proliferative response of T-helper cells (HLA-DQ).’ There is a vast literature describing the expression of MHC class I and class IT antigens in human malignancy.8 The interpretation of much of this data, however, is limited by the fact that it is not quantitative, sensitivity levels have not been defined and retrospective/follow-up studies have been precluded by the necessity of using frozen material. Recently, we have characterized four new human oral carcinoma cell lines which originated from tumours with a broad range of biological activity.’
5
130
S. MUTLU E T A L .
malignant and 4 x lo4normal (together with 4 x lo4 3T3 fibroblasts) keratinocytes per well were grown in 96-well flat-bottomed culture plates and cultured in the presence or absence of 50 units/ml IFN-y (Amersham, U.K.) which was added to each well for 12-96 h. After removal of the culture media, the plates were washed twice in PBS and incubated in 3 per cent (w/v) bovine serum albumin (BSA) (Sigma, W.K.) in PBS for 1 h at 37°C. The BSA solution was then removed and 50pl of primary antibody, diluted in complete media, was added to each well and the cells incubated overnight. After wasifiing with BSA ( x 2), 10 p1 of 35S-labelled sheep antimouse immunoglobulin (Amersham, U.K.) (diluked in PBS containing 10 per cent (v/v) normal porcihe serum) was added for 5 h to give approximately 50 000 c.p.m./well. Plates were washed with BSA ( x 3), the cells solubilized with two 200-p1 aliquots of 10 per cent (w/v) sodium dodecyl sulphate (BDH) and the aliquots counted in 3.5 ml of scintillant in an LKBP-counter. Values were expressed as counts per minute (c.p.m.). Each assay consisted of the results from four microtitre wells for each antibody and the assays were repeated on at least three separate occasions. MATERIALS AND METHODS In parallel with measurements of MHC expression, the amount of protein in individual wells, with Cell culture or without IFN-y, was estimated using a Bio-Rad The establishment and culture of four human protein assay kit (Bio-Rad). The constitutive expression of MHC antigens oral SCC cell lines (H103, H157, H314, H191) and normal oral keratinocytes has been described in in the malignant keratinocytes was calculated by detail p r e v i o ~ s l y .Briefly, ~ cells were cultured in subtracting the ‘non-specific’ c.p.m. values (using Dulbecco’s modified Eagle’s medium (DMEM) non-specific antibodies to rat MHC class I and rat supplemented with 10 per cent (v/v) fetal bovine MHC class 11) from ‘specific’ c.p.m. values (using serum (FBS), 0.075 per cent (v/v) sodium bicar- specific monoclonal antibodies to the human MHC bonate, 0.6 mg/ml L-glutamine and 0.5 pg/ml hydro- monomorphic determinant of HLA-ABC, pzmicrocortisone. Cholera toxin, a routine constituent of globulin, HLA-DR, -DP and -DQ) and expressing epithelial culture media, was omitted in the present the results per microgram of protein. IFN-y study to prevent competitive inhibition with IFN-y.13 stimulated MHC antigen expression was calculated Cultures were incubated in a humidified atmosphere by subtracting ‘non-stimulated’ c.p.m. values of 5 per cent C02/95per cent air at 37°C. Cell culture (using non-specific monoclonal antibodies) from techniques, such as the preparation of mitomycin ‘stimulated’ c.p.m. values (using specific monoC-treated 3T3 fibroblasts, the removal of degener- clonal antibodies) and expressing the results per ating 3T3 cells and host fibroblasts, subculture microgram of protein. Details of the antibodies are procedures and methods used to check that the shown in Table I. The viability of normal oral keratinocytes was cells were mycoplasma free, have also been documented.14 All components were purchased from only maintained by culturing these cells with mitomycin C-treated 3T3 fibroblasts. MHC expression Flow Laboratories, U.K., unless otherwise stated. in normal controls, therefore, was calculated by subtracting values obtained from cultures of M H C class I and class 11antigens mitomycin C-treated 3T3 cells alone from those in MHC antigens were studied using a modification mixed cultures of normal epithelial cells and 3T3 of the technique described by Crane et al.Is 1 x lo4 cells. Thereafter, a similar method was used to
Interestingly, there was considerable variation in the expression of MHC class I and class I1 antigens in the xenografts of the cell lines transplanted to athymic mice.’ We now wish to extend these initial observations by examining the MHC profile of the original oral keratinocyte cell lines. Interferon-gamma (IFN-y), a cytokine produced by activated T cells, LGL and NK cells, exerts antiviral, antiproliferative and immunomodulatory effects.” The fact that IFN-y modulates MHC expression on normal and malignant keratinocytes’ ’ + I 2 suggests that IFN-y is an important mediator of epithelial lymphocyte interactions. What may be important in tumour development, therefore, is whether changes in the tumour MHC antigenic profile, perhaps occurring as a result of an altered response to IFN-y, lead to a selective growth advantage. The purpose of this study, therefore, was to quantify theconsitutive expression ofMHC antigens on normal and malignant human oral keratinocytes and to investigate the effect of IFN-y on MHC antigenic expression.
MHC EXPRESSION IN ORAL CARCINOMA CELL LINES
131
Table I-Primary antibodies Antibody
Source
Dilution
Specificity
~
W6/32
Sera-Lab
1:1600
(mouse IgG2a) W6/32HLK (mouse IgG2a)
Serotec
1:1000
RI .30
Sera-Lab
1:lOO
Dakopatts
1:50 1:20*
Human HLA-DR
Becton-Dickinson
1:50
Human HLA-DQ
(mouse IgG2b) DK22 (mouse IgG2a) Leu 10 (mouse IgG1) B7/21
(mouse IgGI) MRC OX-I8 (mouse IgGl) MRC OX-6 (mouse IgGl)
Human HLA-ABC (shared determinant) Human HLA-ABC (shared determinant) Human µglobulin
1:40*
Serotec
1:50 1:40* 1:400
Serotec
1:400
Becton-Dickinson
Human HLA-DP Rat MHC class I (monomorphicdeterminant) Rat MHC class I1 (monomorphicdeterminant)
*Immunocytochemical studies.
calculate MHC expression in normal keratinocytes as described previously. Controls included the omission of either the primary or secondary antibodies and the substitution of the primary antibody with an irrelevant antibody (Table I). In each experiment, control wells received IFN-y pre-incubated with mouse anti-human IFN-y monoclonal antibody (75 neutralizing units per ml; Sera-Lab, U.K.) at 37°C for 1 h. Cells were assayed at passages 2&28 (H103), 15-19 (H157), 22-31 (H314), 3 2 4 2 (H191) and 2 4 (normal controls). The expression of MHC adtigens by each cell line showed minimal variation between different culture passages and different time points (12-96 h). All of the results, therefore, were pooled for statistical analysis and analysed using a twotailed Student t-test for unpaired samples.
of normal keratinocytes prior to cell counts by agitation in 0.02 per cent (w/v) EDTA for 2-3 min at room temperature.
Immunocytochemistry Spot preparations of cell lines were prepared from cells previously stored in liquid nitrogen in medium containing 20 per cent (v/v) FBS and 10 per cent (v/v) DMSO. Cells were thawed, washed in PBS ( x 3) and resuspended in PBS to give a cell suspension of approximately 1 x lo6 cells/ml. Ten microtitres of cell suspension was pipetted onto each spot of a multidot slide which had been cleaned by dipping in 70 per cent ethanol and air dried prior to use. Two rows of six spot samples were prepared per slide. The cells were air dried (room temperature, 60 min), fixed in acetone for 10 min a t room temperature, wrapped in foil and stored at - 20°C. Prepared slides, which were stored for a maximum of 5 days, were removed from the freezer 1 h prior to staining Growth studies to avoid condensation. Spot preparations were 5&2000 units/ml IFN-I, were added to 5 x lo4 stained using a triple layer indirect immunoperoxicells in 35 mm petri dishes and after 12-96 h incu- dase techniqueI6 and monoclonal antibodies to bation in 5 per cent C0,/95 per cent air at 37"C, the human class I and class I1 MHC antigens (Table I). Controls included omission of the primary anticells were washed in PBS, dissociated in a solution of 0.025 per cent (w/v) trypsin-0.01 per cent (w/v) body and substitution of the primary antibody with EDTA for 20 min at 37°C and counted using a monoclonal antibody to rat Ia (MRC OX-6). The percentage of cells expressing MHC antigens haemocytometer. Mitomycin C-treated 3T3 cells and host fibroblasts were removed from cultures was determined by cell counts at a magnification
132
S. MUTLU E T A L.
,I
24
1.
12
.I
EXPERIMENTAL PERIOD IhOUIsI
I2
21
4.
7s
,I
EXPERIMENTAL PERIOD l h o w i l
Fig. I-The expression of HLA-ABC (constitutive, D; IFN-y stimulated, 0) and 8, microglobulin (constitutive, W ; IFN-.J stimulated, B) o n normal oral keratinocytes (A), HI03 (B), HI57 (C) and H314 (D). Cells were cultured in the absence (constitutive expression) or presence of IFN-y (IFN-stimulated expression) for 12-96 h. The figure illustrates MHC class I antigen expression at passages 28 (H103), 19 (H157), 31 (H314) and 2 (normal keratinocytes). Bar = standard error of the mean
of x 00 using an eye-piece graticule. A total of 500-800 cells were counted per antigen for each cell line. RESULTS M H C class I antigens Fig. 1 shows the constitutive and IFN-y stimu-
H157 (constitutive HLA-ABC P
165: 129-136 (1991)
MHC ANTIGEN EXPRESSION IN HUMAN ORAL SQUAMOUS CARCINOMA CELL LINES SERDAR MUTLU*, JOHN B. MATTHEWST, MARSH MIDDA*, CRISPIAN SCULLY* AND STEPHEN
s. PRIME*
*Department of Oral Medicine, Surgery and Pathology, University of Bristol, U.K.; ?Department of Oral Pathology, University of Birmingham, U.K. Received 14 January 1991 Accepted 18 April 1991
SUMMARY This study quantified the constitutive and interferon-gamma (IFN-y) stimulated expression of MHC class I (HLAABC and 8,microglobulin) and class I1 antigens (HLA-DR, -DP, -DQ) on normal and malignant oral keratinocytes using radioimmunoassay and immunocytochemical techniques. Normal keratinocytes and three of four malignant cell lines (H 103, HI 57, H314) expressed MHC class I antigens constitutively; IFN-y increased MHC class I expression with significant changes in normals, H 157 and H314. Normal keratinocytes expressed significantly more constitutive MHC class I antigens than H 103 and H 157 and significantly more IFN-y stimulated MHC class I antigens than H103, H 157 and H3 14. MHC class I1 antigens predominantly were not expressed constitutively on normals, H 103 and H 157 but, in H314, HLA-DR, -DP and -DQ antigens were demonstrated on 35, 11 and 5 per cent of cells, respectively, and resulted in a non-coordinated pattern ofexpression (HLA-DR> -DP= -DQ). IFN-y induced HLA-DR on normals, HI03 and H157, whilst HLA-DP and -DQ remained undetectable. In H314, IFN-y enhanced HLA-DR, -DP and -DQ (significant increase of HLA-DQ) but the interrelationship between these antigens was maintained (HLA-DR > -DP = -DQ). Normal keratinocytes expressed significantly more IFN-y stimulated HLA-DR than H 103 and H 157 but significantly less HLA-DR than H314 under similar experimental conditions. One oral malignant cell line (H191) did not express MHC class I and MHC class I1 antigens either constitutively or in response to IFN-y. The results demonstrate aberrant patterns of MHC expression (absence, enhanced, diminished) in the different malignant oral keratinocyte cell lines. KEY WORDS-MHC,
antigens, human, oral, carcinoma, cell lines.
INTRODUCTION A number of different factors are involved in tumour-immune system interactions. The recognition of major histocompatibility complex (MHC) antigens on tumour cells by cell-mediated immune mechanisms is now known to be of critical importance in the early events of neoplasia.’ MHC genes in humans are located on the short arm of chromosome 6 and encode both MHC class I and class IT antigens. MHC class I antigens consist of a heavy chain (HLA-A, -B, -C) linked noncovalently to &-microglobulin (encoded by a gene on chromosome 15) and are expressed predominantly by all cell types.’ MHC class I1 antigens (HLA-DR, -DP, -DQ) consist of two chains, a heavy (a) and a light @) chain and, in healthy tissues, Addressee for correspondence:Dr S. S. Prime, Department of Oral Medicine, Surgery and Pathology, Bristol Dental Hospital, Lower Maudlin Street, Bristol BS1 2LY, U.K.
0022-341 71’9ljl00 1 2 9 4 8 $05.00 0 1991 by John Wiley & Sons, Ltd.
are mainly expressed by the cells of the immune system (antigen-presentin cells, B lymphocytes, activated T lymphocytes). MHC class I antigens serve as restriction elements for T-cell-mediated cytot~xicity,~ whilst MHC class TI molecules are required for the presentation of antigen to helper T cells (HLA-DR),’ serve as up-regulators of cytotoxic T lymphocytes (CTL) (HLA-DP),6 and modulate indirectly the proliferative response of T-helper cells (HLA-DQ).’ There is a vast literature describing the expression of MHC class I and class IT antigens in human malignancy.8 The interpretation of much of this data, however, is limited by the fact that it is not quantitative, sensitivity levels have not been defined and retrospective/follow-up studies have been precluded by the necessity of using frozen material. Recently, we have characterized four new human oral carcinoma cell lines which originated from tumours with a broad range of biological activity.’
5
130
S. MUTLU E T A L .
malignant and 4 x lo4normal (together with 4 x lo4 3T3 fibroblasts) keratinocytes per well were grown in 96-well flat-bottomed culture plates and cultured in the presence or absence of 50 units/ml IFN-y (Amersham, U.K.) which was added to each well for 12-96 h. After removal of the culture media, the plates were washed twice in PBS and incubated in 3 per cent (w/v) bovine serum albumin (BSA) (Sigma, W.K.) in PBS for 1 h at 37°C. The BSA solution was then removed and 50pl of primary antibody, diluted in complete media, was added to each well and the cells incubated overnight. After wasifiing with BSA ( x 2), 10 p1 of 35S-labelled sheep antimouse immunoglobulin (Amersham, U.K.) (diluked in PBS containing 10 per cent (v/v) normal porcihe serum) was added for 5 h to give approximately 50 000 c.p.m./well. Plates were washed with BSA ( x 3), the cells solubilized with two 200-p1 aliquots of 10 per cent (w/v) sodium dodecyl sulphate (BDH) and the aliquots counted in 3.5 ml of scintillant in an LKBP-counter. Values were expressed as counts per minute (c.p.m.). Each assay consisted of the results from four microtitre wells for each antibody and the assays were repeated on at least three separate occasions. MATERIALS AND METHODS In parallel with measurements of MHC expression, the amount of protein in individual wells, with Cell culture or without IFN-y, was estimated using a Bio-Rad The establishment and culture of four human protein assay kit (Bio-Rad). The constitutive expression of MHC antigens oral SCC cell lines (H103, H157, H314, H191) and normal oral keratinocytes has been described in in the malignant keratinocytes was calculated by detail p r e v i o ~ s l y .Briefly, ~ cells were cultured in subtracting the ‘non-specific’ c.p.m. values (using Dulbecco’s modified Eagle’s medium (DMEM) non-specific antibodies to rat MHC class I and rat supplemented with 10 per cent (v/v) fetal bovine MHC class 11) from ‘specific’ c.p.m. values (using serum (FBS), 0.075 per cent (v/v) sodium bicar- specific monoclonal antibodies to the human MHC bonate, 0.6 mg/ml L-glutamine and 0.5 pg/ml hydro- monomorphic determinant of HLA-ABC, pzmicrocortisone. Cholera toxin, a routine constituent of globulin, HLA-DR, -DP and -DQ) and expressing epithelial culture media, was omitted in the present the results per microgram of protein. IFN-y study to prevent competitive inhibition with IFN-y.13 stimulated MHC antigen expression was calculated Cultures were incubated in a humidified atmosphere by subtracting ‘non-stimulated’ c.p.m. values of 5 per cent C02/95per cent air at 37°C. Cell culture (using non-specific monoclonal antibodies) from techniques, such as the preparation of mitomycin ‘stimulated’ c.p.m. values (using specific monoC-treated 3T3 fibroblasts, the removal of degener- clonal antibodies) and expressing the results per ating 3T3 cells and host fibroblasts, subculture microgram of protein. Details of the antibodies are procedures and methods used to check that the shown in Table I. The viability of normal oral keratinocytes was cells were mycoplasma free, have also been documented.14 All components were purchased from only maintained by culturing these cells with mitomycin C-treated 3T3 fibroblasts. MHC expression Flow Laboratories, U.K., unless otherwise stated. in normal controls, therefore, was calculated by subtracting values obtained from cultures of M H C class I and class 11antigens mitomycin C-treated 3T3 cells alone from those in MHC antigens were studied using a modification mixed cultures of normal epithelial cells and 3T3 of the technique described by Crane et al.Is 1 x lo4 cells. Thereafter, a similar method was used to
Interestingly, there was considerable variation in the expression of MHC class I and class I1 antigens in the xenografts of the cell lines transplanted to athymic mice.’ We now wish to extend these initial observations by examining the MHC profile of the original oral keratinocyte cell lines. Interferon-gamma (IFN-y), a cytokine produced by activated T cells, LGL and NK cells, exerts antiviral, antiproliferative and immunomodulatory effects.” The fact that IFN-y modulates MHC expression on normal and malignant keratinocytes’ ’ + I 2 suggests that IFN-y is an important mediator of epithelial lymphocyte interactions. What may be important in tumour development, therefore, is whether changes in the tumour MHC antigenic profile, perhaps occurring as a result of an altered response to IFN-y, lead to a selective growth advantage. The purpose of this study, therefore, was to quantify theconsitutive expression ofMHC antigens on normal and malignant human oral keratinocytes and to investigate the effect of IFN-y on MHC antigenic expression.
MHC EXPRESSION IN ORAL CARCINOMA CELL LINES
131
Table I-Primary antibodies Antibody
Source
Dilution
Specificity
~
W6/32
Sera-Lab
1:1600
(mouse IgG2a) W6/32HLK (mouse IgG2a)
Serotec
1:1000
RI .30
Sera-Lab
1:lOO
Dakopatts
1:50 1:20*
Human HLA-DR
Becton-Dickinson
1:50
Human HLA-DQ
(mouse IgG2b) DK22 (mouse IgG2a) Leu 10 (mouse IgG1) B7/21
(mouse IgGI) MRC OX-I8 (mouse IgGl) MRC OX-6 (mouse IgGl)
Human HLA-ABC (shared determinant) Human HLA-ABC (shared determinant) Human µglobulin
1:40*
Serotec
1:50 1:40* 1:400
Serotec
1:400
Becton-Dickinson
Human HLA-DP Rat MHC class I (monomorphicdeterminant) Rat MHC class I1 (monomorphicdeterminant)
*Immunocytochemical studies.
calculate MHC expression in normal keratinocytes as described previously. Controls included the omission of either the primary or secondary antibodies and the substitution of the primary antibody with an irrelevant antibody (Table I). In each experiment, control wells received IFN-y pre-incubated with mouse anti-human IFN-y monoclonal antibody (75 neutralizing units per ml; Sera-Lab, U.K.) at 37°C for 1 h. Cells were assayed at passages 2&28 (H103), 15-19 (H157), 22-31 (H314), 3 2 4 2 (H191) and 2 4 (normal controls). The expression of MHC adtigens by each cell line showed minimal variation between different culture passages and different time points (12-96 h). All of the results, therefore, were pooled for statistical analysis and analysed using a twotailed Student t-test for unpaired samples.
of normal keratinocytes prior to cell counts by agitation in 0.02 per cent (w/v) EDTA for 2-3 min at room temperature.
Immunocytochemistry Spot preparations of cell lines were prepared from cells previously stored in liquid nitrogen in medium containing 20 per cent (v/v) FBS and 10 per cent (v/v) DMSO. Cells were thawed, washed in PBS ( x 3) and resuspended in PBS to give a cell suspension of approximately 1 x lo6 cells/ml. Ten microtitres of cell suspension was pipetted onto each spot of a multidot slide which had been cleaned by dipping in 70 per cent ethanol and air dried prior to use. Two rows of six spot samples were prepared per slide. The cells were air dried (room temperature, 60 min), fixed in acetone for 10 min a t room temperature, wrapped in foil and stored at - 20°C. Prepared slides, which were stored for a maximum of 5 days, were removed from the freezer 1 h prior to staining Growth studies to avoid condensation. Spot preparations were 5&2000 units/ml IFN-I, were added to 5 x lo4 stained using a triple layer indirect immunoperoxicells in 35 mm petri dishes and after 12-96 h incu- dase techniqueI6 and monoclonal antibodies to bation in 5 per cent C0,/95 per cent air at 37"C, the human class I and class I1 MHC antigens (Table I). Controls included omission of the primary anticells were washed in PBS, dissociated in a solution of 0.025 per cent (w/v) trypsin-0.01 per cent (w/v) body and substitution of the primary antibody with EDTA for 20 min at 37°C and counted using a monoclonal antibody to rat Ia (MRC OX-6). The percentage of cells expressing MHC antigens haemocytometer. Mitomycin C-treated 3T3 cells and host fibroblasts were removed from cultures was determined by cell counts at a magnification
132
S. MUTLU E T A L.
,I
24
1.
12
.I
EXPERIMENTAL PERIOD IhOUIsI
I2
21
4.
7s
,I
EXPERIMENTAL PERIOD l h o w i l
Fig. I-The expression of HLA-ABC (constitutive, D; IFN-y stimulated, 0) and 8, microglobulin (constitutive, W ; IFN-.J stimulated, B) o n normal oral keratinocytes (A), HI03 (B), HI57 (C) and H314 (D). Cells were cultured in the absence (constitutive expression) or presence of IFN-y (IFN-stimulated expression) for 12-96 h. The figure illustrates MHC class I antigen expression at passages 28 (H103), 19 (H157), 31 (H314) and 2 (normal keratinocytes). Bar = standard error of the mean
of x 00 using an eye-piece graticule. A total of 500-800 cells were counted per antigen for each cell line. RESULTS M H C class I antigens Fig. 1 shows the constitutive and IFN-y stimu-
H157 (constitutive HLA-ABC P
