Specimens from the Graves’ disease patients were obtained during thyroidectomy from 11 patients who were in euthyroidism with the antithyroid drug and those of normal controls, from the adjacent normal portion of 6 patients with nodular goiter. The specimens were immediately frozen below −20°C and thawed within 1 month of storage to be cut into 6 μm sections for double indirect immunofluorescence staining.

Monoclonal antibody to human HLA-DR antigen was produced at the Department of Microbiology, College of Medicine, Seoul National University²³⁾ (1BD9-2). Sera containing antimicrosomal antibody were obtained from 2 patients with Graves’ disease whose titer of antimicrosomal antibody was over 1:1280² without detectable antithyroglobulin antibody. FITC-labeled or unlabeled rabbit antimurine IgG (heavy & light chain specific), and rhodamine-labeled rabbit antihuman IgG (heavy & light chain specific) were purchased from Cappel (Malvern, PA, USA). Monoclonal anti-CD5, CD4, CD8 antibodies (T1, T4, T8 antibody, respectively) and AP-AAP complex were purchased from DAKO Corp. (Santa Barbara, CA, USA). Bovine serum albumin, Fast Red TR Salt (5-chloro-2-toluidinediazonium chloride hemizine chloride), levamisole and Naphthol AS-MX sodium salt (3-hydroxy-2-Naphthoic acid 2,4-dimethylanilide) were purchased from Sigma (St. Louise, MO, USA).

The appropriate cut sections of frozen tissue were incubated for 30 min in a humid chamber at 37°C with the culture supernatant of the 1BD9-2 cell line and a 1:100 dilution of the sera containing antimicrosomal antibody. Then the sections were washed 3 times with phosphate buffered saline for a total of 30 min, and were reincubated with a 1:20 dilution of FITC-labeled rabbit antimurine IgG and rhodamine-labelled rabbit antihuman IgG in a phosphate buffered saline containing 0.1 % bovine serum albumin for 30 min in a humid chamber at 37°C. The sections were washed 3 times with phosphate buffered saline again for a total of 30 min, and examined under a fluorescent microscope (American Optical) after mounting with glycerol in phosphate buffered saline. The negative control slide was made by incubating the adjacent cut sections with the second antibodies only. Autofluorescence was excluded by scrutinizing unstained cut sections under the fluorescent microscope, and the cross-immunoreactivities between human IgG & antimurine IgG, or murine IgG & antihuman IgG were tested by cross-staining of the cut sections to yield negative results. The number of prominently bright-green follicles compared to the negative control slide was counted in over 50 thyroid follicles and the ratio was rounded to one decimal.

The specimen of one patient which showed large clusters of DR antigen positive cells in the vicinity of the thyroid follicles with double immunofluorescence staining was subjected to AP-AAP (Alkaline Phosphatase-Antialkaline Phosphatase) double bridge staining^24,25) using T1, T4 or T8 monoclonal antibody as the first antibody. The cut frozen sections were incubated with a 1:40 dilution of the first antibody in phosphate buffered saline for 30 min at room temperature and washed with phosphate buffered saline 3 times for a total of 6 min subsequently. Then they were incubated with a 1:50 dilution of antimurine IgG for 30 min at room temperature and washed again 3 times with phosphate buffered saline for a total of 6 min followed by incubation with 1:50 dilution of AP-AAP complex and rewashing. The incubation with the second antibody and the AP-AAP complex were recapitulated to make a “double bridge” and the processed sections were incubated with a color substrate solution made by dissolving 10 mg of Fast Red TR Salt, 2.4 mg of levamisole and 2 mg of Naphthol AS-MX sodium salt in 0.1 M Tris buffer (pH 8.2), 10 ml. The final sections were counterstained with hematoxylin.

The titers of antimicrosomal antibody and antithyroglobulin antibody were measured with the hemagglutination method using the kits produced by Fujirebio (Japan). The activity of TBII (TSH Binding Inhibitory Immunoglobulin) was measured using the kits produced by R.S.R. Limited (Cardiff, Wales, UK).

Ten fields of hematoxylin-eosin stained sections were examined in low-power microscopy to grade the degree of lymphoid follicle formation from − to + + + +. The presence of interstitial lymphocyte infiltration was searched apart from lymphoid follicle formation.

Kendall’s W test of concordance was employed to study the correlation between the degree of DR antigen expression on the thyrocytes, and clinical or laboratory indices, the degree of lymphocyte infiltration in the thyroid gland.

All of the thyroid follicles of both the normal controls and the patients with Graves’ disease were stained with rhodamine suggesting the presence of microsomal antigen on the thyrocytes (Fig. 1).

DR antigen was not found at all on the frozen sections of the normal thyroid tissues from 6 subjects. Strong or weak autofluorescence that could be discerned from true staining by the yellow color instead of green, was observed occasionally along the inner side of large or small vasculatures (Fig. 2).

Of 11 patients with Graves’ disease, 8 patients (73%) showed DR antigen expression on their thyroid follicles, and the ratio of DR antigen-positive follicles varied between 10–80% (Fig. 3, Table 1). The staining pattern was patch, i.e. DR-positive thyrocytes were rather aggregated. There was no apparent correlation between the degree of HLA-DR antigen expression on the thyroid follicles and the clinical or laboratory indices including age, duration of the disease, the activity of TBII, and the titers of antimicrosomal antibody & antithyroglobulin antibody (p>0.1 in all tests).

On the several frozen sections from the patients with Graves’ disease, large or small clusters of microsomal antigen-negative, DR antigen-positive cells were observed after double immunofluorescence staining (Fig. 1, 3). The cells in those clusters were positively stained with the T1 antibody except in the central area (Fig. 4). The lymphocytes stained with T1 antibody also stained with T4 antibody or T8 antibody in the staining of the adjacent sections of the tissue (Fig. 5, 6). There was no apparent discrepancy in the distribution patterns of T4-positive helper/inducer T-lymphocytes and T8 positive suppressor/cytotoxic T-lymphocytes but the quantitative ratio was about 3:1. The lymphocytes of both subsets were also scattered in the thyroid follicles (Fig. 7). Meanwhile, the cells in the central area of those DR antigen-positive clusters showed positive rhodamine staining not only in case of the regular double immunofluorescence staining but also in the control staining, i.e. in the absence of the antimicrosomal antibody, suggesting the presence of a surface immunoglobulin on the cells in the central area (Fig. 1, 3). So the cells in the central area of those clusters were thought of as B lymphocytes considering that they had a surface immunoglobulin in addition to the DR antigen and were not stained with T1 antibody (Fig. 4). Germinal centers were also observed occasionally in the central area of those DR-positive clusters with light microscopy.

There was no apparent spatial or quantitative relation between the lymphoid follicle formed and the DR antigen expression on the thyroid follicles (p>0.1). But the degree of lymphoid follicle formation and the presence of interstitial lymphocyte infiltration did not coincide, and all 4 cases with interstitial lymphocyte infiltration showed DR antigen expression on 50% or above of their thyroid follicles examined (Table 1).