Antibodies to thyroid peroxidase arise spontaneously with age in NOD.H-2h4 mice and appear after thyroglobulin antibodies

Abstract

Hashimoto's thyroiditis, a common autoimmune disease, is associated with autoantibodies to thyroglobulin (Tg) and thyroid peroxidase (TPO). TPO, unlike abundant and easily purified Tg, is rarely investigated as an autoantigen in animals. We asked whether antibodies (Abs) develop to both TPO and Tg in thyroiditis that is induced (C57BL/6 and DBA/1 mice) or arises spontaneously (NOD.H-2h4 mice). Screening for TPOAbs was performed by flow cytometry using mouse TPO-expressing eukaryotic cells. Sera were also tested for binding to purified mouse Tg and human TPO. The antibody data were compared with the extent of thyroiditis. Immunization with mouse TPO adenovirus broke self-tolerance to this protein in C57BL/6 mice, but thyroiditis was minimal and TgAbs were absent. In DBA/1 mice with extensive granulomatous thyroiditis induced by Tg immunization, TPOAbs were virtually absent despite high levels of TgAbs. In contrast, antibodies to mouse TPO, with minimal cross-reactivity with human TPO, arose spontaneously in older (7-12 months) NOD.H-2h4 mice. Unexpectedly, TgAbs preceded TPOAbs, a time course paralleled in relatives of probands with juvenile Hashimoto's thyroiditis. These findings demonstrate a novel aspect of murine and human thyroid autoimmunity, namely breaking B cell self-tolerance occurs first for Tg and subsequently for TPO.

Figure 1

Expression of mouse TPO measured by flow cytometry. A, COS-7 cells transiently infected with mTPO-Ad (upper panels) and hTPO-Ad (lower panels) and tested with human TPO recombinant antibody Fab that recognize the A-domain (Fab WR1.7) or the B-domain (Fab TR1.8). The oval indicates the percentage of positive cells in the M2 fraction. B, CHO cells stably expressing mouse TPO: normal mouse serum (NMS, left panel) and mouse monoclonal antibody no. 24 (right panel) generated against human TPO (40).

Figure 2

TPOAbs, but not TgAbs, develop in C57BL/6 mice immunized with mTPO-Ad. BALB/c mice immunized with hTPO-Ad (29) develop Abs that cross-react with mouse TPO but not TgAbs. Sera were tested from mice euthanized 4 wk after the third immunization with mTPO-Ad, hTPO-Ad, or control (Con)-Ad. Data are shown for individual mice. Shaded areas represent the mean ± 2 sd for Con-Ad-immunized mice. A and B, TPOAbs measured by flow cytometry with mouse TPO-CHO cells (sera 1:50) in C57BL/6 (A) and BALB/c (B) mice; data are expressed as the geometric mean (Geo Mean). **, P < 0.001, Statistically significant differences; #, P < 0.002, (rank sum tests). C and D, TgAbs measured by ELISA (sera 1:100) in C57BL/6 (C) and BALB/c (D); values are reported as the OD 490 nm. Positive control values (OD 1.33 ± 0.12) were obtained for sera from A-subunit transgenic mice that developed TgAbs after Treg depletion and immunization with TSHR adenovirus (33). E and F, Normal thyroid histology in mice immunized with Con-Ad (E) and lymphocytic infiltrate in a few animals immunized with mTPO-Ad (F). Magnification, ×100.

Figure 3

Granulomatous thyroiditis induced in DBA/1 mice is not associated with TPOAbs despite high levels of TgAbs and extensive thyroiditis. Thyroiditis was induced by transferring primed splenocytes activated in vitro with mouse Tg + IL-12 (35). Observations were pooled for WT and Flip-2 transgenic DBA/1 mice because they were similar (Supplemental Fig. 1B). Sera from mice without thyroiditis (grade 0) provided negative controls. Mice were tested at the time of maximum disease (d 18 and 20) and in the resolution phase (d 40 and 55). Thyroiditis scores were high (4–5+) at all time intervals in affected animals. Data are shown for individual mice; shaded areas represent the mean ± 2 sd for mice lacking thyroiditis. A, TPOAbs measured by flow cytometry (sera 1:50). Positive control: Geo mean 54.1 (serum from C57BL/6 mouse immunized with mTPO-Ad). B, TgAbs measured by ELISA (sera diluted 1:3000). C–F, Thyroid histology depicting lymphocytic thyroiditis grade 5+ (C and D) and 4–5+ (E and F). Magnifications, ×100 (C and E) and ×400 (E and F).

Figure 4

Development of TPOAbs in NOD.H-2h4 mice is associated with increasing age. Sera were tested from WT mice (solid circles) and IL-4^−/− and CD28^−/− mice (stippled circles) all on the NOD.H-2h4 background. Mice received regular water (−I) or water supplemented with 0.05% NaI (+I) for times ranging from 3 to 12 months. Shaded areas represent mean ± 2 sd for mice without thyroid lesions. It should be noted that the IL-4^−/− mice tested were a subgroup with severe thyroiditis; most IL-4^−/− animals develop mild thyroiditis as in WT mice (37). A, TPOAbs measured by flow cytometry with mouse TPO-CHO cells (sera 1:50). *, P < 0.02, **, P < 0.001, values significantly different from mice without thyroid lesions (rank sum test). B, TgAbs measured by ELISA (sera 1:100). C and D, Thyroid pathology in NOD.H-2h4 mice not exposed to NaI (C) or after exposure to NaI (0.05% NaI in drinking water). [Reproduced with permission from S. M. McLachlan, et al.: Endocrinology 46:294–300, 2005. ©The Endocrine Society.]

Figure 5

TPOAbs that develop in mice immunized with mTPO-Ad (C57BL/6) or that arise spontaneously (NOD.H-2h4) cross-react poorly with human TPO. In contrast, TPOAbs in mice immunized with hTPO-Ad (BALB/c), as well as some monoclonals from mice immunized with human TPO (40), bind to mouse TPO. Binding data are compared for mouse vs. human TPO (black bars) for the following strains: C57BL/6 immunized with mTPO-Ad; BALB/c immunized with hTPO-Ad; NOD.H2h4 not exposed (−I) or exposed (+I) to NaI water for 7 or 12 months; monoclonals from BALB/c immunized with human TPO and adjuvant (40). Control values (white bars): Control (C) adenovirus immunized C57BL/6 or BALB/c; mice without thyroid lesions (NOD.H2h4); group × BALB/c monoclonals. [Note: our subdivision of the monoclonals (40) into groups X and Y based on their recognition of mouse TPO: nonbinders, group X (monoclonals no. 40, 47, 53, and 64) and binders, group Y (no. 18, 24, 30, and 59)]. A, Antibody binding to mouse TPO CHO cells; data are shown as the mean (+sem) of values for individual mice in Figs. 2, 3, and 4. B, Antibody binding in ELISA to human TPO (mean + sem). Inset, Correlation between TPOAbs measured using mouse or human TPO for NOD.H2h4 mice; note the high proportion of values outside the 95% confidence limits.

Figure 6

Relationship between age and development of TgAbs or TPOAbs. A, Prevalence of TgAbs and TPOAbs in NOD.H2h4 mice at 2, 7, and 12 months. B, Proportion of Caucasians aged 12–80 yr positive for TgAbs and TPOAbs (plotted from the on-line data from Ref. 43). C–E, TPOAbs and/or TgAbs in siblings (sibs) or parents of juvenile Hashimoto’s thyroiditis probands. Horizontal dashed line (C and D, negative cutoff, <0.5 U/ml). Inset (C), Proportions of siblings positive or negative for TgAbs and TPOAbs. E, mean + sem. #, P = 0.003, TPOAb or TgAb levels significantly different (rank sum test); *, P < 0.001, TPOAb or TgAb levels significantly different (t test).