Blackwell Science, LtdOxford, UKCEIClinical and Experimental Immunology1365-2249Blackwell Publishing Ltd, 2003 1343 497502 Original Article L. E. Rees et al. MHC class II locus expression on human laryngeal epithelium
Clin Exp Immunol 2003; 134:497–502
doi:10.1046/j.1365-2249.2003.02301.x
Differential major histocompatibility complex class II locus expression on human laryngeal epithelium L. E. REES*†, O. AYOUB*, K. HAVERSON†, M. A. BIRCHALL* & M. BAILEY† *Laryngeal Research Group, Department of Otolaryngology, Division of Surgery and †Division of Veterinary Pathology, Infection and Immunity, Clinical Veterinary Science, University of Bristol, Bristol, UK
(Accepted for publication 4 September 2003)
SUMMARY The survival of a laryngeal allograft will be dependent on the immunological composition of the donor larynx and, in particular, on the expression of major histocompatibility complex (MHC) class II antigens on professional and non-professional antigen-presenting cells. Laryngeal and tonsillar biopsies from normal individuals aged 18–78 years were processed and prepared for quantitative, multiplecolour immunofluorescence using mouse antihuman monoclonal antibodies to human leucocyte antigen (HLA)-DR, HLA-DQ and CD45. The laryngeal epithelium expressed HLA-DR locus products at variable levels, but expression of HLA-DQ was virtually absent. Tonsillar epithelial cells expressed HLADR at the basal layer only, while HLA-DQ was similarly not expressed. In contrast, both HLA-DR and -DQ locus products were present on lamina propria and intraepithelial leucocytes in both laryngeal and tonsillar mucosae, although at varying levels. The finding that laryngeal epithelial cells express MHC class II antigens has implications for the survival of laryngeal allografts and suggests that they may require significant immunomodulation. In addition, antigen presentation by epithelial cells has been hypothesized to contribute to the immunoregulatory function of mucosal tissues, and the finding that HLA-DQ locus products are only expressed at low levels by laryngeal epithelium raises questions about the repertoire of peptides to which the mucosal immune system can respond. Keywords
epithelium
intraepithelial leucocytes
INTRODUCTION Currently, the main treatment for patients with irreversible laryngeal disease, including trauma and advanced cancer, is total removal of the larynx (laryngectomy). The first revascularized human laryngeal transplant was performed by a group at Cleveland, Ohio in 1998 [1] and its success has raised hope of the technique becoming routine within the next few years. However, further progress will require an understanding of the immunological function of the normal larynx in both animal models and man, in order to predict the impact of transplantation on mucosal immune responses and the requirement for immunosuppression. The pig is the preferred large animal model for laryngeal transplantation, due to its similarities in size, anatomy and phonation characteristics to the human larynx [2–4]. Large numbers of major histocompatibility complex (MHC) class II positive leucocytes
Correspondence: Dr Louisa Rees, Division of Veterinary Pathology, Infection and Immunity, Clinical Veterinary Science, Langford House, Langford, Bristol BS40 5DU, UK. E-mail:
[email protected] © 2003 Blackwell Publishing Ltd
lamina propria
MHC class II
transplantation
with dendritic cell morphology occur throughout the pig larynx in close association with CD4+ and CD8+ T cells [3]. These cells can, presumably, present antigen locally and also emigrate from laryngeal tissues, and are likely to trigger strong alloresponses after transplantation. The situation is comparable to the intestine which, in humans and pigs, contains MHC class II+ dendritic cells. However, human intestinal epithelial cells can express MHC class II molecules, while this has been demonstrated not to be the case in pigs [5]. The ability of stromal cells (epithelial cells) to express MHC class II antigens is likely to have significant implications for the maintenance of alloresponses in a transplanted larynx. It is therefore important to determine whether human laryngeal epithelial cells contribute to local MHC class II expression and the degree of variability between ‘normal’ individuals. The aim of the present study was to examine expression of MHC class II locus products on the static (epithelial) and migratory (leucocyte) components of the larynx and to determine to what extent they vary between individuals. In addition, we demonstrate quantitative immunofluorescence as a valid method for measuring expression of immunological markers in tissue, by comparison with standard flow cytometric techniques.
497
498
L. E. Rees et al. MATERIALS AND METHODS
Immunofluorescence Pinch biopsies of human supraglottis were obtained from individuals undergoing routine surgery and who were certified to be free from head and neck cancer (n = 39). Whole larynges were obtained from laryngectomy patients ( n = 2) with cancer, and human tonsils were used as control tissue ( n = 4). All procedures were approved by local ethical review and written consent was obtained from every patient. Specimens were embedded in OCT (Sakura, CA, USA) and snap-frozen in isopentane cooled over liquid nitrogen. Five mm sections were cut on a cryostat (Bright, Huntingdon, UK) and processed for dual colour immunofluorescence using mouse antihuman monoclonal primary antibodies to MHC class II locus products [human leucocyte antigen (HLA)-DR, clone G46-6, HLA-DQ, clone TU169 (Pharmingen, CA, USA)] and leucocyte common antigen (CD45, clone HI30, Pharmingen), together with isotype specific goat anti-mouse FITC and Texas Red secondary fluorochrome conjugates (Southern Biotech, AL, USA). The mouse anti-human collagen IV basement membrane (clone AD 4) was a kind gift from Gordon Paul. All antibodies were titrated to optimal working concentration. Sections were blocked for 1 h with 5% human and goat serum in phosphate buffered saline (PBS) in a humidity chamber (R.A. Lamb, NC, USA). Excess serum was removed and 30 ml primary monoclonal antibody mix diluted in PBS to optimal concentration was applied onto each section and incubated overnight at 4∞C. Slides were washed three times in PBS, and incubated for 1 h at room temperature with 30 ml secondary fluorochrome-conjugated goat anti-mouse monoclonals diluted in PBS. Slides were again washed three times in PBS, mounted on coverslips with Vectashield (Vector Laboratories, CA, USA) and sealed with nail varnish. Multiple fields at 40¥ magnification from each specimen were viewed and digitized with a Leica DMR fluorescence microscope (Leica, Wetzlar, Germany) and analysed using ImagePro Plus software (Media Cybernetics, MD, USA). Separate red and green images were digitized using single Texas Red and FITC filters (TX2 and K3, Leica, respectively). Single colour images were then combined to create a red/green colour image. Background intensities for each colour were obtained by examining an area of tissue without staining, and positive staining was defined as above this intensity. The area of locus-specific MHC class II+ (DR or DQ) CD45– epithelial cell staining was calculated as a percentage of the total area of CD45– epithelium for each field. Similarly, for intraepithelial leucocytes and lamina propria leucocytes, the area of MHC class II+ CD45+ staining was calculated as a percentage of the total area of CD45+ leucocyte staining. In every case, each field was counted three times, and averaged. These values for each field were then totalled and an average percentage area value for the entire biopsy was then obtained. Induction of varying levels of MHC class II on leucocytes Thirty ml fresh blood was taken in EDTA tubes and peripheral blood mononuclear cells (PBMC) were separated using density gradient centrifugation with Histopaque 1077 (Sigma, St Louis, MO, USA). Cells were cultured for 24 h at 2 ¥ 106/ml in 25 cm2 flasks (Corning, NY, USA) in RPMI-1640 medium containing Lglutamine (Sigma) supplemented with 10% fetal calf serum (PAA Laboratories, GmbH), and 100 units/ml penicillin/streptomycin
(Sigma), with 20 U/ml IFN-g (Sigma) to induce MHC class II expression on monocytes. Plastic non-adherent cells were recovered by washing the culture flasks gently and adherent cells were recovered using a cell scraper (Falcon, Becton Dickinson). Adherent and non-adherent cells were recombined at different ratios to give final populations containing variable numbers of cells expressing MHC class II.
Assessment of HLA-DR expression on isolated cells by flow cytometry The mixtures of cultured cells were resuspended at 2 ¥ 107/ml, blocked with 10% human and goat serum for 30 min, washed in PBS-azide and incubated for 1 h with mouse monoclonal antihuman antibodies to CD45 and HLA-DR (Pharmingen), then for a further 1 h with isotype-specific goat anti-mouse fluorescein isothyocyanate (FITC) and PE-conjugated secondary monoclonals (Southern Biotech). A total of 40 000 cells were analysed for HLA-DR and CD45 expression using a Coulter Epics XL flow cytometer and the percentage of HLA-DR + CD45+ cells was calculated as a proportion of the total number of CD45 + leucocytes. Assessment of HLA-DR expression on isolated cells by immunofluorescence The mixtures of cultured cells were resuspended at 1 ¥ 107 cells/ml and spun down onto microscope slides using a cytospin (Shandon Southern). Slides were air-dried for 48 h and then subjected to immunofluorescence staining and analysis exactly as described for patient biopsies. Statistical analysis Statistical analysis was carried out using SPSS (SPSS Inc.). Paired two-sample data were compared using paired and/or unpaired t-tests or Wilcoxon signed rank where data were obviously nonparametric. Three-sample data were compared using ANOVA.
RESULTS Comparison of quantitative immunofluorescence and flow cytometric analysis methods In order to validate comparisons of MHC class II expression using quantitative immunofluorescence, populations containing different numbers of MHC class II+ cells were analysed by flow cytometry and by immunofluorescence. Figure 1 shows a high degree of correlation between the two measures ( R2 = 0·9699). We conclude that quantitative immunohistology is likely to provide as accurate a measure of numbers of cells expressing of MHC class II as does flow cytometry. Expression of MHC class II loci in the human larynx Laryngeal epithelial cells from each patient expressed HLA-DR at variable levels. Figure 2a is a section of supraglottic mucosa stained for HLA-DR (green) and with an antibody to collagen 4 basement membrane (AD 4, in red). Overall, expression of HLADR was significantly greater than HLA-DQ (P < 0·005, paired or unpaired tests), with HLA-DQ virtually absent on laryngeal epithelium (Fig. 2b). When the supraglottis was stained for the presence of leucocytes (CD45, red) and HLA-DQ (green), again there was little or no expression of HLA-DQ locus products on epithelium, but leucocytes within and below the epithelium itself expressed this locus product (Fig. 3).
© 2003 Blackwell Publishing Ltd, Clinical and Experimental Immunology, 134:497–502
MHC class II locus expression on human laryngeal epithelium
% area HLA-DR + CD45 + cells by cytospins and immunofluorescence
Correlation between flow cytometry and percentage area calculations
60
499
(a)
R 2 = 0·9699
50 40 30 20 10 0
10
20
30
40
50
% HLA-DR + CD45 + cells by flow cytometry
In the tonsil, the pattern of MHC class II distribution was somewhat different. Expression of HLA-DR locus products was restricted to the basal layer of tonsillar epithelium (Fig. 4a), and many leucocytes strongly expressing HLA-DR were situated above and below the basement membrane, as in the larynx. Levels of epithelial HLA-DQ were not significantly different to HLA-DR (Fig. 4b), although HLA-DR locus products were present at lower levels on tonsillar epithelium compared to laryngeal epithelium (note the differences in scale between Figs 2b and 4b). Figure 5 demonstrates the expression patterns of HLA-DR and HLA-DQ on larynx and tonsillar CD45 + leucocytes within the epithelium. Levels of HLA-DR were significantly greater than HLA-DQ on laryngeal intraepithelial leucocytes (P < 0·005). HLA-DR locus products were also present at higher levels in laryngeal lamina propria leucocytes (Fig. 6). However, both locus products were not significantly different on leucocytes in tonsil epithelium and lamina propria (Figs 5 and 6). The results from initial analyses demonstrated that, in the larynx, epithelial cells, intraepithelial leucocytes and lamina propria leucocytes all expressed greater amounts of HLA-DR compared to HLA-DQ. Importantly, we wanted to know whether this differential expression of HLA-DR and HLA-DQ was consistent between the cell types. The ratio of HLA-DR to HLA-DQ was calculated for each cell type and is shown in Fig. 7. The ratio of HLA-DR/-DQ was significantly higher on laryngeal epithelium than on intraepithelial leucocytes and both were highly significantly greater than the ratio on lamina propria leucocytes. Thus, the ratio of expression is highest on epithelial cells, followed by intraepithelial leucocytes, followed by lamina propria leucocytes. However, in all sites the ratios were highly variable, particularly so in epithelial cells.
% Area MHC Class ll positive epithelial cells
on peripheral blood mononuclear cells. Shown is the correlation between percentage of HLA-DR+ CD45+ cells by flow cytometry and percentage area of HLA-DR+ CD45+ staining by immunofluorescence analysis of cytospins. Each point indicates a separate sample containing increasing proportions of MHC class II+ monocytes.
P < 0·005
(b)
Fig. 1. Comparison of two methods to measure expression of HLA-DR
100 80 70 60 50 40 30 20 10 0 DR
DQ
Fig. 2. (a) Two-colour immunofluorescent image of supraglottis stained with anti-HLA-DR (green) and anticollagen 4 basement membrane (red). RE: respiratory epithelium; BM: basement membrane; LP: lamina propria. (b) Expression of HLA-DR (left) and HLA-DQ (right) in laryngeal epithelium. Each triangle represents one individual.
DISCUSSION This study demonstrates clearly that MHC class II products are expressed on laryngeal epithelium. If the epithelial cells
Fig. 3. Two-colour immunofluorescent image of supraglottis stained with anti-HLA-DQ (red) and anti-CD45 (green). Blue is autofluorescence. SSE: stratified squamous epithelium; L: leucocyte; LP: lamina propria.
© 2003 Blackwell Publishing Ltd, Clinical and Experimental Immunology, 134:497–502
500
L. E. Rees et al.
(a)
(b) n·s·
% Area MHC Class ll positive epithelial cells
20 18 16 14 12 10 8 6 4 2 0 DR
DQ
Fig. 4. (a) Two-colour immunofluorescent image of tonsil stained with anti-HLA-DR (green) and anti CD45 (red). Blue is autofluorescence. BM: basement membrane; SSE: stratified squamous epithelium; L: leucocyte; LP: lamina propria; HEV: high endothelial venule. (b) MHC class II expression in tonsillar epithelium. Each triangle represents one individual.
themselves are capable of acting as antigen-presenting cells, then both direct and indirect allorejection will be a potential problem in laryngeal transplantation. Antigen presentation by epithelial cells has been demonstrated in vitro and hypothesized to contribute to the immunoregulatory function of mucosal tissues [6–9]. Hershberg [10] described how the polarized human colonic epithelial cell line T84, when transfected with an HLA-DR allele, expressed MHC II molecules without invariant chain (Ii), and that intestinal epithelial cells transfected with an HLA-DR allele could use two distinct MHC class II processing pathways, depending on the presence of IFN-g. Therefore, it is possible that under inflammatory conditions, such as are likely to occur after transplantation, mucosal epithelial cells can be induced to function as typical antigen presenting cells, while the function of constitutive MHC class II expression by epithelial cells may be simply to ‘shuttle’ peptides to deeper immunologically active cells. Thus, epithelial cells may contribute to direct allorecognition by infiltrating recipient T cells. In addition, studies of dendritic cells in intestinal afferent lymphatics have suggested that epithelial debris is present in migrating dendritic cells [11]: if these cells express MHC class II then indirect recognition of allogeneic MHC class II peptides is also likely. The study has also compared the relative expression of different MHC class II locus products on laryngeal epithelial cells, on leucocytes within the epithelium and on leucocytes in the lamina propria, below the basement membrane. The differences in the relative expression of HLA-DR and HLA-DQ products observed here clearly imply differences in regulation of expression between laryngeal epithelium, intraepithelial leucocytes and lamina propria leucocytes. The differences between cell subsets, the high level of interindividual variation and the equivalence of paired and unpaired statistical test results (data not shown) suggest that epithelial and leucocyte HLA-DR and -DQ are controlled by disparate microenvironmental factors. Furthermore, intraepithelial leucocytes also appeared to exhibit different ratios of HLA-DR/ -DQ locus product expression compared to leucocytes within the lamina propria, suggesting that these also comprise two distinctly regulated CD45+ leucocyte populations. The control tonsil biopsies do not indicate any statistical differences in MHC II locus expression, but this is may be due to the analysis of only four biopsies, and is preliminary data only. This study is ongoing so it is anticipated that greater numbers of tonsils will be obtained for comparison.
P < 0·005 % Area MHC Class ll positive epithelial leucocytes
100 90 80 70 60 50 40 30 20 10 0
% Area MHC Class ll positive epithelial leucocytes
100 90 80 70 60 50 40 30 20 10 0 DR
DQ
n.s.
DR
DQ
Fig. 5. MHC class II expression in laryngeal epithelial leucocytes (left); MHC class II expression in tonsillar epithelial leucocytes (right). n.s.: not significant. © 2003 Blackwell Publishing Ltd, Clinical and Experimental Immunology, 134:497–502
MHC class II locus expression on human laryngeal epithelium P < 0·005 100 90 80 70 60 50 40 30 20 10 0
n.s.
100 90 80 70 60 50 40 30 20 10 0
% Area MHC Class ll positive lamina propria leucocytes
% Area MHC Class ll positive lamina propria leucocytes
501
DR
DQ
DR
DQ
Fig. 6. MHC class II expression in laryngeal lamina propria leucocytes (left); MHC class II expression in tonsillar lamina propria leucocytes (right). n.s.: not significant.
P < 0·005 P < 0·05 P < 0·005 3·5 3 Log DR/DQ ratio
2·5 2 1·5 1 0·5 0 -0·5 -1 -1·5 Larynx epithelium
Intra-epithelial leucocytes
Lamina propria leucocytes
Fig. 7. The logged HLA-DR/DQ ratios on laryngeal epithelium, intraepithelial leucocytes and lamina propria leucocytes for 41 individuals.
It has been demonstrated that rodent and human MHC class II genes are generally expressed co-ordinately, but with a few important exceptions. Bland et al. [12] have demonstrated in vivo that expression of MHC class II I–E ak protein on murine gut epithelial cells and lamina propria macrophages seems to be controlled differently to that on lamina propria dendritic cells, consistent with the idea that MHC class II molecule expression is regulated differently in mucosal cell populations. Similarly, MHC class II is present on normal human intestinal epithelium in a hierarchical way: HLA-DR > DP > DQ, with HLA-DQ present either at extremely low levels or undetectable [13,14]. However, all three locus products have been reported to be up-regulated in inflammatory bowel disease [15–17]. In vitro studies in both mouse and man have indicated that the transcription of HLA-DQ locus products is controlled differently to that of HLA-DR and -DP [18], particularly on epithelial cells. HLA-DQ transcription was found to be largely independent of the class II transactivator (CIITA) [19]. Indeed, the nucleotide sequences of transcriptional control regions of HLA-DQB genes are highly polymorphic, suggesting that control even of different alleles may vary [20]. Beaty et al. [20] showed that different
HLA-DQ alleles had different transcriptional efficiencies, attributable to a TG dinucleotide between the W and X elements, which correlated with the allele-specific binding of the trans-acting transcriptional regulatory protein YY1. Together, these observations confirm that HLA-DQ genes may be subject to additional, allelespecific regulatory mechanisms, including differential transcription of the genes encoding the two subunits of the mature heterodimer and post-transcriptional defects such as altered translation and/or altered assembly of the a and b subunits. The data described from in vivo immunohistological studies were all qualitative. The present study demonstrates comparable results for laryngeal cells in a quantitative manner and confirms differences in expression between HLA-DR and -DQ, presumably reflecting the different transcriptional and posttranscriptional control mechanisms. The study also showed a high degree of variability in expression of both molecules in a range of individuals. This raises interesting questions regarding the repertoire of peptides presented by systemic and mucosal antigenpresenting cells, suggesting that mucosal environments may be less efficient at presenting HLA-DQ- than -DR-restricted peptides. Further studies are now required to characterize the leucocyte populations within and below the epithelium, as it is likely that local dendritic cells and T cells are regulating MHC class II locus expression by different mechanisms. Finally, it is necessary to correlate findings with lifestyle risk factors such as smoking and alcohol consumption to allow hypothesis testing on the relationship between sources of inflammation and MHC class II expression. Such a study would also suggest what lifestyle characteristics an ideal laryngeal transplant donor might possess.
ACKNOWLEDGEMENTS We would like to thank the Charles Courtney-Cowlin endowment fund, BBSRC and Wellcome for funding this work.
REFERENCES 1 Strome M, Stein J, Esclamado R et al. Laryngeal transplantation and 40-month follow-up. N Engl J Med 2001; 344:1676–9. 2 Gorti GK, Birchall MA, Haverson K et al. A preclinical model for laryngeal transplantation: anatomy and mucosal immunology of the porcine larynx. Transplantation 1999; 15:1638–42. 3 Barker E, Rees L, Birchall M et al. Abstract given at British Society of Immunology, Harrogate, December 2000 [Abstract no. OP183].
© 2003 Blackwell Publishing Ltd, Clinical and Experimental Immunology, 134:497–502
502
L. E. Rees et al.
4 Jiang JJ, Raviv JR, Hanson DG. Comparison of the phonation-related structures among pig, dog, white-tailed deer, and human larynges. Ann Otol Rhinol Laryngol 2001; 110:1120–5. 5 Wilson AD, Haverson K, Southgate K, Bland PW, Stokes DR, Bailey M. Expression of major histocompatibility complex class II antigen on normal porcine intestinal endothelium. Immunology 1996; 88:98–103. 6 Mayer L. The role of epithelial cells as accessory cells. Adv Exp Med Biol 1987; 216A:209–18. 7 Bland PW, Warren LG. Antigen presentation by epithelial cells of the rat small intestine. I. Kinetics, antigen specificity and blocking by antiIa antisera. Immunology 1986; 58:1–7. 8 Bland PW, Warren LG. Antigen presentation by epithelial cells of the rat small intestine. II. Selective induction of suppressor T cells. Immunology 1986; 58:9–14. 9 Kaiserlian D, Vidal K, Revillard JP. Murine enterocytes can present soluble antigen to specific class II-restricted CD4+ T cells. Eur J Immunol 1989; 19:1513–6. 10 Hershberg RM, Framson PE, Cho DH et al. Intestinal epithelial cells use two distinct pathways for HLA class II antigen processing. J Clin Invest 1997; 100:204–15. 11 Hwang I, Huang JF, Kishimoto H et al. T cells can use either T cell receptor of CD28 receptors to absorb and internalise cell surface molecules derived from antigen-presenting cells. J Exp Med 2000; 191:1137–48. 12 Bland PW, Whiting CV. Differential control of major histocompatibility complex class II, I–Eak protein expression in the epithelium and in subsets of lamina propria antigen-presenting cells of the gut. Immunology 1993: 79:107–11.
13 Madrigal L, Lynch S, Feighery C et al. Flow cytometric analysis of surface major histocompatibility complex class II expression on human epithelial cells prepared from small intestinal biopsies. J Immunol Meth 1993; 158:207–14. 14 Marley NJ, Macartney JC, Ciclitira PJ. HLA-DR, DP and DQ expression in the small intestine of patients with coeliac disease. Clin Exp Immunol 1987; 70:386–93. 15 Natali PG, De Martino C, Pellegrino MA et al. Analysis of the expression of I-Ak-like antigens in murine fetal and adult tissues with the monoclonal antibody 10–2·16. Scand J Immunol 1981; 13:541–6. 16 Horie Y, Chiba M, Iizuka M et al. Class II (HLA-DR-DP, and -DQ) antigens on intestinal epithelia in ulcerative colitis, Crohn’s disease, colorectal cancer and normal small intestine. Gastroenterol Jpn 1990; 25:575–84. 17 Hirata I, Austin LL, Blackwell WH et al. Immunoelectron microscopic localization of HLA-DR antigen in control small intestine and colon and in inflammatory bowel disease. Dig Dis Sci 1986; 31:1317–30. 18 Accolla RS, Tosi G, Sartoris S et al. MHC class II gene regulation: some historical considerations on a still ontogenetic and phylogenetic puzzle. Microbes Infect 1999; 1:871–7. 19 Douhan J, Lieberson R, Knoll J et al. An isotype-specific activator of major histocompatibility complex (MHC) class II genes that is independent of class II transactivator. J Exp Med 1997; 185:1885– 95. 20 Beaty JS, Sukiennicki TL, Nepom GT. Allelic variation in transcription modulates MHC class II expression and function. Microbes Infect 1999; 1:919–27.
© 2003 Blackwell Publishing Ltd, Clinical and Experimental Immunology, 134:497–502
