Increased toll-like receptor 4 expression in thymus of myasthenic patients with thymitis and thymic involution

Abstract

Thymic abnormalities are present in approximately 80% of myasthenia gravis (MG) patients, and the thymus seems to be the main site of autosensitization to the acetylcholine receptor. In view of findings that the innate immune system can generate an autoimmune response, we studied the expression of Toll-like receptors (TLRs) 2 to 5, key components of innate immunity signaling pathways, in 37 thymuses from patients with autoimmune MG. TLR4 mRNA levels were significantly greater in thymitis (hyperplasia with diffuse B-cell infiltration) and involuted thymus than in germinal center hyperplasia and thymoma. By immunohistochemistry and confocal microscopy, cells positive for TLR4 protein were rarely detected in thymoma. However, in thymitis TLR4 protein was mostly found on epitheliomorphic (cytokeratin-positive) cells located in close association with clusters of acetylcholine receptor-positive myoid cells in thymic medulla and also at the borders between cortical and medullary areas. B cells were never TLR4-positive. TLR4 protein was also present in remnant tissue of involuted thymus. This is the first finding of a possible link between innate immunity and MG. We speculate that in a subgroup of MG patients, an exogenous or endogenous danger signal may activate the innate immune system and give rise to TLR4-mediated mechanisms contributing to autoimmunity.

Figure 1

Steady-state TLR mRNA levels in MG thymitis, thymoma, involuted thymuses, and hyperplasia, and in healthy thymuses. The results are expressed as band intensities relative to β-actin and are plotted as means ± SEM. In general mRNA levels of TLR2, TLR3, and TLR5 are higher in healthy than pathological thymuses, with significant differences for hyperplasia for TLR2 and TLR3 (*P < 0.005, Bonferroni/Dunn test). By contrast, TLR4 mRNA expression was much higher in thymitis and thymoma than other specimens, with a statistically significance difference between thymitis and hyperplasia (**P < 0.005, Bonferroni/Dunn test).

Figure 2

Overexpression of TLR4 mRNA in thymitis and involuted thymus. The transcript was analyzed by real-time PCR starting from total RNA extracted from MG thymuses. TLR4 levels were higher in thymitis and involuted thymus compared with hyperplastic thymus (P = 0.0047 and P = 0.0026, respectively). Horizontal bars are means with SDs. Values obtained from healthy thymuses were used as calibrator as required by the 2^−ΔΔCT method.

Figure 3

Photomicrographs showing histopathological characteristics of thymitis. A: H&E staining shows thymus gland with darker cortex (left) and paler medullary region (center and right). Note preserved lobular structure of cortex. B and C: Immunostainings with anti-CD20 antibody reveal B cells (CD20⁺), which are infiltrating perivascular spaces and surround Hassall’s corpuscles. D: Note Hassall’s corpuscles (center) surrounded by rounded and spindle-shaped myoid cells (desmin⁺). E: Clusters of desmin-positive myoid cells associated with indentated border within perithymic adipose tissue. Original magnifications: ×10 (A); × 20 (B–D); ×40 (E).

Figure 4

Immunostaining of thymitis specimens. A and B are serial sections: A, TLR4 immunostaining; B, cytokeratin immunostaining. Note that the TLR4-positive cytokeratin-positive TECs form a band adjacent to the cortex. C: TLR4 stains pericortical epithelial cells interspersed between cortical lymphoid cells. D: TLR4-positive cells clustered around a Hassall’s corpuscle. E: Pericortical and medullary staining of epithelial cells by TLR4. F: Immmunostaining to reveal myoid cells shows that these occasional cells are often associated with Hassall’s corpuscles (arrow). Tissue sections counterstained with hematoxylin. Original magnifications: ×20 (A, B, F); ×40 (C–E).

Figure 5

Serial sections of an involuted thymus. A: CD1a (thymocyte) staining; B: CD22 staining; C: cytokeratin staining; D: TLR4 staining. Note that TLR4 positivity is closely associated with epithelial (cytokeratin-positive) cells and that B cells CD22⁺ are negative for TLR4. All counterstained with hematoxylin. Original magnifications, ×10.

Figure 6

A, B: Serial sections of thymoma; C–E: healthy young thymus. In A, cytokeratin staining reveals abundant epithelial cells, whereas B show few cells positive for TLR4. In normal thymus cytokeratin immunostaining (C) and macrophage immunostaining (E) reveal that these cells are fairly abundant. D: By contrast TLR4-positive cells are uncommon. Tissue sections counterstained with hematoxylin. Original magnifications: ×20 (A, B); ×10 (C–E).

Figure 7

Expression of TLR4 (B, E, I, L) on cytokeratin-positive cells (C, F, J, M) in thymitis (A–G) and involuted thymus (H–N), viewed under confocal microscope. Merged images (D, G, K, N) of TLR4 (green) and cytokeratin (red) show co-localization (yellow) of TLR4 and cytokeratin positivity. In thymitis samples, the double labeling is restricted to around Hassall’s bodies as shown by single immunostaining (Figure 4D). In involuted thymus, double immunofluorescence is positive on residual TECs. Original magnifications: ×10 (A, H); ×20 (B–D, I–K); ×40 (E–G); ×60 (L–N).