T-lymphocyte subset changes in recurrent aphthous stomatitis

oral pathology

--

Editor: JAMES

J. SCIUBBA,

D.M.D.,

Ph.D.

American Academy of Oral Pathology Department of Dentistry Long Island Jewish-Hillside Medical Center New Hyde Park, New York 11042

T-lymphocyte subset changes in recurrent aphthous stomatitis N. W. Savage,* G. J. Seymour,** L4IVERSITY

and B. J. Kruger, * St. Lucia, Queensland, Australia

OF QI;EEYSLAND

lmmunocytochemical techniques using monoclonal antibodies directed against T-lymphocyte surface antigens (OKT3, OKT4, and OKT8) were used to investigate changes in the T-cell subpopulations throughout the natural time course of recurrent aphthous stomatitis (RAS). All lesions contained large numbers of OKTS-positive T ceils. Preulcerative lesions were characterized by many OKTd-positive inducer/helper cells and few OKT8-positive suppressor /cytotoxic cells (T4 : T8 = 2 : 1). The ulcerative lesion contained large numbers of OKTB-positive suppressor/cytotoxic cells and only a very small number of OKT4-positive cells (T4:T8 x 1 : 10). The healing lesion once again became dominated by OKT4-positive cells, with only a small number of OKT8 positive cells present (T4:T8 ^I 10: 1). The results support the role of lymphocytotoxicity in the establishment of the RAS lesion and may suggest a local immunoregulatory imbalance. NK cells, Leu-7 positive, were also identified in the preulcerative lesions and very early ulcerative lesions. 10na1 SURG.ORAL MED. ORAL PATHOL. 60:175-181. 1985)

A

number of studies have defined the histology and histologic changes that occur in the life span of the recurrent aphthous stomatitis (RAS) lesion.‘-3A recent article4 has shown that the composition of the gross mononuclear cell infiltrate does not appear to change with the age of the lesion. An area that has received minimal attention, however, is the in situ identification of immunocompetent cell phenotypes in the RAS lesion. The initial, as yet unrecognized, event in the RAS lesion causesintra- and/or subepithelial cellular activity that can, by direct observation, reasonably be indicted as responsible for the development of mucosal ulceration. The localization and topographic relationships of lesional lymphocytes in the developing, established, and healing RAS lesion constitute an important area in dermatopathology that is yet to be investigated. Aberrations

*Department of Oral Biology and Oral Surgery. **Department of Social and Preventive Dentistry.

in immunoregulatory T-cell subsets exist in a number of human diseases,5so it is equally important to define possible imbalances between effector and regulatory cells in RAS lesions. The purpose of the present study is to examine the phenotype of immunocompetent cells in RAS lesions and to document the changes in cell populations occurring with lesion progression and healing. This was done with a panel of monoclonal antibodies directed against T-lymphcyte surface antigens and examined by immunofluorescence and immunoperoxidase techniques. It was found that there are definite changes in the T-cell subset ratios from the stage of the early preulcerative lesion through to established and healing ulcers. MATERIALS

AND METHODS

Sixteen excisional biopsy specimens of untreated and uncomplicated minor aphthous ulcers were taken with the aid of local anesthesia. The lesions were divided according to whether they showed prodromal 175

176

Savage, Seymour, and Kruger

I. Mouse monoclonal antibodies against human immunocompetent cells Table

Antibody

Dilution

OKT3* OKT4* OKT4A* OKT8* Anti-Leu-7t

1:30 1:50 I:50 I : 100 I:50

Specificity Pan peripheral T lymphocytes’ Inducer/helper T lymphocyte? Inducer/helper T lymphocyte9 Suppressor/cytotoxic T lymphocytes”’ NK/K lymphocytes”

*Ortho Diagnostic Systems Inc., Raritan, N.J. fBecton and Dickinson Monoclonal Centre Inc., Mountain View. Calif.

symptoms or clinical signs of ulcer formation. They were further classified as preulcerative, ulcerative, and healing by the criteria of Stanley.2 This gave three preulcerative, seven ulcerative, and six healing lesions, all of which were taken from either the buccal or labial mucosa. Each specimen was immediately divided transversely and again longitudinally, with one piece being fixed for future ultrastructural investigations. The remainder were embedded on the cut face in cylinders of OCT (Lab-Tek, Naperville, Illinois) and quick-frozen in liquid nitrogen. Cryostat sections (4 pm) were cut from each block and labeled serially, air-dried for 2 hours, fixed in chloroform-acetone (50:50) for 5 minutes and stored at -20” C until use.6The initial sections in each series were stained with hematoxylin and eosin and toluidine blue for histologic assessmentand for confirmation or adjustment of the estimation of the clinical age of the lesion. Clinical age was judged on cell density and distribution and the state of the disintegrating or repairing epithelium. A range of monoclonal antisera with well-defined specificities was used. These are detailed in Table I. All of the monoclonal antibodies were tested on both tonsillar tissue and normal oral mucosa before use. OKT4 and OKT4A were used together to enhance cell identification, as they both define the inducer/ helper T-cell subset and differ only in the specific epitope they are directed against.‘* All incubations were completed with the same polyclonal goatantimouse immunoglobulin in a titrated dilution of one-in-five (Dako Immunochemicals, Copenhagen, Denmark). This antibody was extensively absorbed with rat liver powder to reduce nonspecific or electrostatic binding to tissue components. Immunoelectrophoresis was used to check cross reactivity of the conjugate with human serum proteins. No reactivity was found. Lymphocyte surface antigens were detected by the technique of Poulter and co-worker@ and doublelayer/indirect fluorescence. All incubations and washings were done for a period of 30 minutes. The

Oral Surg. August, 1985

sections were examined in a Zeiss fluorescent microscope with incident illumination. Control sections in which the primary layer was omitted and also normal tonsil were included in each batch of slides. The results were assessedblindly with sections and clinical data being subsequently matched. To complement the immunofluorescence studies and to obtain permanent slides, the avidin-biotinperoxidase complex system (ABC) was used. Following application of the primary layer and washing, a second layer of goat-antimouse IgG biotin conjugate (1:20; Sigma Chemical Company, St. Louis, Missouri) was applied and incubated for 30 minutes. The washed sections were incubated with avidin-biotinperoxidase complex (1: 20; Sigma Chemical Company, St. Louis, Missouri) for 30 minutes. The peroxidase was then developed with diaminobenzidinel H,O,.” For assessment,the lesions were grouped as far as possible into preulcerative, ulcerative, and healing ulcers. Sections from each lesion stained with hematoxylin and eosin and/or toluidine blue were examined and a brief record was made of the histopathologic findings. The immunofluorescent sections were examined for total lymphocyte population and for T-cell subsets in specific areas. The two areas of interest were the intra- and subepithelial areas and the more deeply situated perivascular cells. The distribution of reactive cells and their relative ratios was noted. These figures were then compared with similar recordings made in each of the three groupings. RESULTS Light microscopy

Sections of the earliest preulcerative lesions in this series showed the universal presenceof a mononuclear cell infiltrate (Fig. 1, a). At the deeper tissue levels the early lesion was characterized by perivascular mononuclear cell aggregations (Fig. 1, b) which increased in intensity with lesion progression. Smaller numbers of these cells were present in the subepithelial “reticular” layer. Superficially, the cells formed a subepithelial band of varying density with the migration of small numbers of cells intraepithelially at several separate foci (Fig. 1, c). With progression, the intensity of the cell population increased, particularly in the number of intraepithelial cells. At this stage, edema, acantholysis, and rete peg disintegration occurred with the formation of the typical ulcerative defect. As the ulcerative stage was approached, a corresponding increase in this cellular infiltrate occurred, with the eventual blending of the two areas. There did not appear to be a similar ordered regression in the healing lesion, but there

Volume 60 Number 2

T-cell subsets in aphthous stomatitis

177

Fig. 1. Preulcerative lesion. a, Mononuclear cell infiltrate is present perivascularly (*) and superficially (arrow). b, Section showing the intensity of the perivascular cell aggregations. c, Intraepithelial mononlxlear cells. (Original magnificationr. 9. X100; b and c, X400.)

was a gradual overall decrease in the mononuclear cell population. The final stage showed the epitheliurn to be cell-free, but the deep perivascular cell clusters were still present although greatly lessened in intensity. lmmunohistology

The majority of mononuclear cells at both the deep and superficial levels were OKT3 positive at all three stages. In preulcerative lesions, tightly packed clusters of OKT4/4Apositive cells surrounded deeper blood vesselsand in many areas formed multiple small subepithelial groups. In slightly more advanced areas these cells were seen to form broken chains along the basement membrane zone and also appeared as isolated intraepithelial cell clusters (Fig. 2, a). With the appearance of the epithelial break, some sections contained small clusters immediately adjacent to the break but within the epithelium. In real numbers the OKT4/4A-positive cells found perivascularly far outnumbered the superficial group. The ulcerative-stage OKT4/4A cell numbers

dropped considerably at all levels (Fig. 2, b), and in some sections they formed only a very small percentage of the total mononuclear cell population. With the initiation of the healing phase (judged by both clinical and histologic criteria), the deep perivascular cells were once again OKT4/4A positive and in many sections formed tightly compacted perivascular aggregations. Superficially, the OKT4/4A cell numbers increased, but less dramatically {Fig. 2, c). Positive cells were present intraepithelially, but the majority were within the papillary layer. In the early preulcerative lesion, occasional single OKT%positive perivascular cells were present. The numbers soon increased (Fig. 3, a), but not to the level of the OKT4/4A-positive cells. The papillary layer always contained a significant number of OKT8-positive cells, often (T4/T8 ratios of 2: 1 perivascularly and 1 : 1 subepithelially) in numbers nearly equal to the OKT4/4A-positive cells. These cells, however, were almost always single and did not form clusters, as did the OKT4/4A-positive cells. The number of intraepithelial OKT8positive cells

178

Savage, Seymour, and Kruger

Fig. 2. Intraepithelial lymphocytes. The decreasein OKT4/4A-positive cells from the preulcerative (a) to the ulcerative (b) phase and then an increase again in the healing phase (c) are associated with isolated OKTS-positive cells in the preulcerative phase (d), which increase grossly in the ulcerative phase(e) and then once again decreasein the healing phase (I). (Original magnification, x400.)

was initially restricted to isolated cells (Fig. 2, d), but the influx of more positive cells occurred early and rapidly so that the OKT8-positive cells soon outnumbered the OKT4/4A-positive cells, particularly in sites where rete peg degeneration was occurring. The ulcerative lesion was dominated by OKT8positive cells. They appeared in large numbers at all levels in the lesion. In well-established ulcers, the cells ranged from perivascular aggregations (Fig. 3, b) through the entire lamina propria to form dense pockets of intraepithelial activity (Fig. 2, e). In these

lesions OKT4/4A cells were seen only occasionally (T4/T8 ratio of 1: 10). The healing phase presented a dramatic change in the deep perivascular areas, with OKT4/4A cells far outnumbering of OKT8 cells (Fig. 3, c) once healing became established. Superficially, a significant but lesser number of OKT8 cells were seen subepithelially and, in many areas, intraepithelially. In the one almost completely healed lesion examined, intraepithelial OKT8 cell activity was almost absent (Fig. 2,j) and the overall scene was dominated by OKT4/4A-positive cells (T4/T8 ratio of 10: 1).

Volume Number

T-cell subsets in aphthous .rtomatitis

60 !

I 79

Fig. 3. Perivascular OKT%positive cells. The preulcerative phase shows few OKT%positive cells (a), but a rapid rise occurs to the ulcerative phase (b). The healing phase is again characterized by only kolated OKTS-positive perivascular cells (c). (Original magnification, X400.)

These results are summarized in Table II. Cells showing membrane fluorescence with the anti-Leu 7 antisera were a distinct population. They were present only in preulcerative and very early ulcerative lesions. The Leu-7-positive cells were restricted to the lamina propria and appeared as single fluorescing cells scattered evenly across the sections. Once the ulcer became established, only very occasional positive cells were seen. Not only were the Leu-7 cells absent from the epithelium; they were present only in very small numbers perivascularly. DISCUSSION

A series of monoclonal antibodies was employed to study the T-cell populations at varying stages in the RAS lesion. The basic histopathologic appearance of the lesions conformed to previously reported patterns.‘v2 None of the preulcerative specimens was free of lymphocytes, 3 but this may be a result of “late” tissue biopsy. The preulcerative lesion was characterized by both deep perivascular and superficial aggregations of T cells of the OKT4-positive subset. These were rapidly overtaken by OKT8positive T cells, almost to the exclusion of the OKT4-positive lymphocytes in the ulcerative lesion. With the onset of the healing phase, the dominance was once again reversed, with decreasing numbers of OKTS-positive cells and a marked increase in OKTCpositive cells. While not a part of this study, the importance of the monocyte/macrophage family and their interactions with T lymphocytes has not been ignored and will form the subject of a forthcoming article. The importance of this study lies in the recognition

II. Comparative study of immunocompetent cells in RAS lesions*

Table

-------p-pqq,,,,;;

-,,/,,,-

I

Prculcerative Superficial Perivascular

+t++ t+++

++ +++

+i t

7: I

tt++ ++++

+ +

+++ +++

I IO

UIceratiw Superficial Perivascular

Healing ++++ ++ Superficial ++++ +++ Perivascular _-----*t = 0X-25’? of total cell population. ++ = X%-50% of total cell population. +++ = 50%75% of total cell populatNxl. ++++ = 76’8%loo’% of tot31 cell population.

+t +

IO: I

of a consistent pattern of change in the lymphocyte subsetsduring the course from preulcerative through to clinical ulceration and, finally, the commencement of healing in RAS lesions. The literature concerning RAS leaves little doubt that cellular responsesplay an important effector part in recurrences.‘4 Cells expressing the T4 surface antigen are generally classed as inducer/helper lymphoyctes, but these cells also display a marked degree of functional heterogeneity and can function as both suppressor and cytotoxic effector cells.15The local progression to a lesion dominated by OKT8 cells and their almost certain attack on morphologically normal epithelial cells is indicative of at least a local breakdown in immune regulation. The phase change that marks the initiation of recognizable healing is intriguing. The influx of OKT4-positive cells appears to mark the return to normal immunologic reactivity, with lesion resolution occurring soon afterward.

180

Savage, Seymour, and Kruger

Delayed hypersensitivity (DTH) reactions in allergic and irritant contact dermatitisI show more similarity to the very early aphthous lesion. They are characterized by perivascular and subepithelial infiltrates and intraepithelial vesicles. The mixed cellular infiltrate is dominated by OKT4-positive cells, although it appears that only a small percentage of the cells are specifically sensitized. The predisposition in RAS patients, in combination with local trauma, prolonged influence under an as yet unrecognized initiating agent, and the local immunodysregulation, might modify the basic DTH reaction to such an extent as to lead to epithelial ulceration. The T4/T8 ratio of 2: 1 in the initial lesion supports this premise, while the subsequent elevated numbers of OKTS-positive cells may reflect the local regulatory defect. It is possible that these cells are performing a protective and beneficial function, but in the absence of any supportive evidence this seemsunlikely. The normally tight regulation of DTH by suppressor cells, with the prevention of nonspecific arming of host macrophages,” does not seem to have been achieved in RAS. OKT8-positive cells are presumed cytotoxic in RAS lesions by virtue of their position and cellular relations. Cells showing membrane fluorescence following section incubation with anti-Leu-7 were present only in the preulcerative and early ulcerative phases. These cells are commonly termed large granular lymphocytes (LGL) and are active in antibodydependent cell-mediated cytotoxicity (ADCC) or natural killer (NK) activities. Greenspan and coworkers18 showed that patients with early RAS lesions have a significantly higher ADCC than matched controls. In prodromal, established, and healing lesions, however, there were no differences. This seemsto exclude ADCC as an initiating mechanism in RAS, although the heterogeneity and maturational changes in this population of LGL19 does not fully rule out earlier activity in RAS initiation. In RAS, current knowledge places the NK cells as a consequenceof the initiating events or their sequelae. The question arises as to the possible cytotoxic activity of the NK cells and their targets. Epithelial cells as shown by the present study are geographically separate from the NK cells in RAS lesions. Fibroblasts in the underlying connective tissue are possible NK cell targets. This sort of activity does, in fact, occur in atopic dermatitis.*O In summary, the RAS lesion is characterized by consistent changes in the T4/T8 ratios extending from the preulcerative to the ulcerative and healed phases. The possible significance of these changes

Oral Surg. August, I985

has been examined, and support for an altered delayed hypersensitivity reaction is given. The presence of NK cells in the early RAS lesion is also discussed, although as yet no precise role can be assigned to these cells, We thank Mrs. E. Gemmell for her assistance avidin-biotin-complexperoxidasesystem.

with the

REFERENCES I Lehner T: Pathology of recurrent oral ulceration and oral ulceration in Behcet’s syndrome: light, electron and fluoresence microscopy. J Path01 97: 481-494, 1969. 2. Stanley HR: Aphthous lesions. ORAL SURG ORAL MED ORAL PATHOL 33: 407-416, 1972. 3. Stenman G, Heyden G: Premonitory stages of recurrent aphthous stomatitis. 1. Histological and enzyme histological investigations. J Oral Path01 9: 155-162, 1980. 4. Schroeder HE, Miiller-Glauser W, Sallay K: Stereologic analysis of leukocyte infiltration in oral ulcers of developing Mikulicz aphthae. ORAL SURG ORAL MED ORAL PATHOL 56: 629-640, 1983. 5. Bach MA, Bach JF: The use of monoclonal anti-T cell antibodies to study T cell imbalances in human diseases. Clin Exp lmmunol 45: 449-456, 198 1. 6. Poulter LW, Chilosi M, Seymour GJ, Hobbs S, Janossy G: lmmunofluorescence membrane staining and cytochemistry, appplied in combination for analysing cell interactions in situ. In Polak JM, van Noorden S (editors): Immunocytochemistry: practical applications in pathology and biology, Bristol, 1983, John Wright & Sons Ltd., pp. 233-248. 7. Kung PC, Goldstein G, Reinherz EL, Schlossman SF: Monoclonal antibodies defining distinctive human T-cell surface antigens. Science 206: 347-349, 1979. 8. Reinherz EL, Kung PC, Goldstein G, Schlossman SF: Further characterization of the human inducer T-cell subset defined by a monoclonal antibody. J Immunol 123: 28942902, 1979. 9. Kung PC, Talle MA, De Maria M, et al: Creating a useful panel of anti-T-cell monoclonal antibodies. Int J lmmunopharmacol 3: 175-181, 1981. 10. Reinherz EL, Kung PC, Goldstein G, Schlossman SF: A monoclonal antibody reactive with the human cytotoxic/ suppressor T-cell subset previously defined by a heteroantiserum termed TH2. J Immunol 124: 1301-1307, 1980. Il. Abo T, Balch CM: A differentiation antigen of human NK and K cells identified by a monoclonal antibody (HNK-I). J lmmunol 127: 1024-1029, 1981. 12. Rao PE, Talle MA, Kung P, Goldstein G: Five epitopes of a differentiation antigen on human inducer T cells distinguished by monoclonal antibodies. Cell lmmunol80: 3 1O-3 19, 1983. 13 Petrusz P, Di Meo P, Ordronneau P, Weaver C, Keefer DA: Improved immunoglobulin-enzyme bridge method for light microscopic demonstration of hormone-containing cells of the rat adenohypophysis. Histochemistry 46: 9-26, 1975. 14 Lehner T: Progress report: Oral ulceration and Behcet’s syndrome. Gut 18: 491-511, 1977. 15 Goldstein G: Monoclonal antibodies: use in the diagnosis and therapy of human disease. In Otterness I, Capetola R, Wong S (editors): Advances in inflammation research, New York, 1984, Raven Press, vol. 7, pp. 115-I 21. 16 Scheynius A, Fischer T, Forsum U, Klareskog L: Phenotypic characterization in situ of inflammatory cells in allergic and irritant contact dermatitis in man. Clin Immunol 55: 81-90, 1984. 17. Liew FY: Lymphocyte subsets involved in delayed-type hypersensitivity. In Otterness I, Capetola R, Wong S (edi-

T-cell subsets in aphthous rfomatitis

Volume 60 &umber 2 torso: Advances in inflammation research, New York, 1984. Raven Press, vol. 7, pp. 135-147. 18. Greenspan JS, Gadol N, Olson JA, Talol N: Antibodydependent cellular cytotoxicity in recurrent aphthous ulceration Clin Immunol 44: 603-610, 1981. 19. Ferrarini M, Pistoria V. Nocera A, Ghio R, Crossi CE: Large granular lymphocytes: a new lineage with special functions? Monogr Allergy 18: 112-l 19, 1983. 20. Leung DYM. Parkman R, Feller .I, Wood N. Geha RA:

Cell-mediated cytotoxicity against skin lihroblasts dermatitis. J lmmunol 128: 1736- 17Ji 1’)‘ .!

Reprim reque.tts to. Dr. N. W. Savage Department of Oral Biology and Oral Surger:. University of Queensland St. Lucia. Queensland 4067, Australia

ERRATUM

In the article entitled “Three-Dimensional Imaging of the Temporomandibular Joint” by Roberts et al., which appeared in the October, 1984, issue of the JOLRNAL, the following footnote should have appeared on page 46 1: This research was supported by the National Institutes of Health, Heart and Lung Institute, Grant HL-28438, BRSG Grant RR-05337-2 1, and the Division of Research Resources, G.C.R.C. Grant 5 MOl-RR01224. L-

-_-...-A

181

in aropic