Résumé : In previous studies we have induced TSH binding-inhibiting Igs and thyroiditis in BALB/c mice and thyroiditis alone in NOD mice immunized with the extracellular domain of the human TSH receptor produced as a maltose-binding protein fusion in bacteria (MBP-ECD). In this study, our aim was to determine whether thyroiditis can be transferred to syngeneic naive recipients with in vivo and in vitro primed spleen cells. Groups of 6-week-old female BALB/c and NOD mice were immunized ip with MBP-ECD in an adjuvant of alum plus attenuated Bordetella pertussis toxin, on days 0 (100 micrograms), 14, 28, and 35 (50 micrograms). These mice (in vivo primed) and nontreated age- and sex-matched controls were killed on day 43, and their spleens and thyroids were removed, the latter to verify the induction of thyroiditis in the antigen-treated mice. Splenocytes were disrupted mechanically and cultured at 3 x 10(6)/ml in RPMI supplemented with 20 micrograms/ml MBP-ECD for 48-64 h. After this in vitro priming, some of the splenocytes received no further treatment, but a portion was fractionated into a CD4+-enriched population. Groups of 6-week-old female BALB/c and NOD mice were immunized into the tail vein with 100-200 microliters PBS containing approximately 10(5)-10(7) unfractionated T cells (both in vivo primed and not) and CD4+-enriched (in vivo primed) splenocyte populations. The animals were killed 16 days later, and their thyroids were examined histologically and by immunohistochemistry. In addition, levels of antibody to the MBP-ECD priming antigen were assessed by enzyme-linked immunosorbent assay in the antigen- and spleen-treated mice. In the donor animals, in vivo priming resulted in an extensive lymphocytic infiltration of the thyroids in both BALB/c and NOD mice and follicular destruction in the latter. There was no evidence of thyroiditis in all 9 BALB/c mice and all 4 NOD mice who received unfractionated T cells from mice that had not been primed in vivo. In contrast, transfer of MBP-ECD in vivo primed unfractionated T cells resulted in thyroiditis in 9 of 13 BALB/c mice and 5 of 6 NOD mice; similarly, the equivalent CD4+-enriched population produced extensive thyroiditis in 2 of 3 BALB/c mice and all three NOD mice. The most striking difference between the antigen- and spleen-treated mice was in the quantity of the infiltrate, which was much greater in the latter and extended throughout the thyroid glands of these animals. In common with mice treated directly with the MBP-ECD antigen, the infiltrates of both BALB/c and NOD recipient mice contained large numbers of activated T cells expressing the receptor for interleukin-2, and macrophages and dendritic cells were plentiful, particularly in the BALB/c mice, in which B cells and interleukin-10-positive T cells were also present. The most abundant infiltrates, containing numerous CD8+ T cells and follicular destruction, were observed in NOD mice receiving primed unfractionated T cells or CD4+-enriched T cells. In contrast to the donors, none of the recipient animals had circulating antibodies to the MBP-ECD antigen. In conclusion, we have shown that it is possible to transfer thyroiditis with spleen cells from mice primed in vivo with a human TSH receptor preparation. Furthermore, the thyroiditogenic activity appears to reside in the CD4+ population.