Thymomas are thymic epithelial neoplasms, associated with a variety of autoimmune disorders (especially myasthenia gravis), that apparently result from aberrant intra-tumourous thymopoiesis and export of inefficiently tolerized T-cells to the periphery. The autoimmune regulator (AIRE) drives the expression of self-antigens in the thymic medulla and plays an essential role in 'central' tolerance in both humans and mice. However, while inactivating AIRE mutations result in the 'autoimmune polyendocrinopathy syndrome type 1' (APS-1), its major features are not well reproduced in AIRE-knock-out mice. Therefore, alternative human disease scenarios with concomitant AIRE deficiency may be valuable tools to test conclusions drawn from mouse models. Here we show, in a large series, that approximately 95% of thymoma patients are 'chimeric'; expression of AIRE and major AIRE-related autoantigens (eg insulin) were undetectable in their tumours but maintained in their remnant thymic tissue and lymph nodes. Notably, despite the AIRE-deficient thymopoiesis in thymomas, disorders and autoantibodies typical of APS-1 were distinctly uncommon in these patients. The one striking similarity was in the recently observed neutralizing anti-type I interferon (IFN) antibodies, which are found at diagnosis in 100% of patients with APS-1 and in approximately 60% of patients with thymomas, as we show here. We conclude that APS-1 type autoantigens must be protected from autoimmunity by mechanisms that do not extend to the muscle autoantigens so frequently targeted in thymoma patients but so rarely recognized in APS-1. Thus our findings argue strongly for a tolerogenic function of AIRE beyond its role in negative T-cell selection in human thymopoiesis, and/or for specific autoimmunization against muscle in thymomas.

To study the permeability of the blood-brain barrier (BBB) and the blood-nerve barrier (BNB) for immunoglobulin G (IgG) we adapted the avidin-biotin system for postembedding demonstration of the tracer IgG in the central and peripheral nervous system (CNS, PNS). Normal mouse and human IgG were biotinylated and injected daily into the intraperitoneal (i.p.) space of adult BDF1 mice. After 24h, IgG was detected in blood vessels and in the interstitium of various organs, but staining was restricted to the dura mater in the CNS, to the spinal ganglia, and to the perineurium of peripheral nerves. After 4 days, IgG was also present in the endoneurial connective tissue of peripheral nerves, while the brain, spinal cord, and spinal roots remained free of IgG. Our results show a partial permeability of the normal mouse BNB for homologous and heterologous IgG.

Neural-specific T lymphocytes are held to play a pathological role in inflammatory peripheral nerve disorders such as the Guillain-Barré syndrome (GBS) and chronic inflammatory demyelinating polyneuropathy (CIDP). Here, non-neural-specific T-cell-mediated inflammation was studied in peripheral nerves in Lewis rats by systemic transfer of ovalbumin-specific activated T cells followed by intraneural injection of ovalbumin. Rapid endoneurial perivenular infiltration of alpha beta T cells and ED1+ macrophages occurred with ovalbumin injection following transfer of 2 x 10(6) T cells. This cellular infiltration and accumulation produced marked increases in blood-nerve barrier (BNB) permeability. In contrast, control casein injections produced neither significant cell accumulation nor BNB permeability changes. Transfer of a higher number of T cells (5 x 10(6)) induced severe Wallerian degeneration and nerve conduction failure in ovalbumin injected nerves. Fewer T cells (5 x 10(5)) induced conduction block and mild demyelination which were markedly augmented by systemic cotransfer of anti-myelin immunoglobulin. This study demonstrates that activated T cells of non-neural specificity can accumulate in peripheral nerve, produce dramatic changes in BNB permeability and with intravenous anti-myelin antibody orchestrate primary demyelination or axonal degeneration in a dose-dependent fashion.

Mice were injected daily, for up to 10 weeks, with purified monoclonal immunoglobulin G from patients with myelomatous polyneuropathy or benign gammopathy. The animals developed a demyelinating polyneuropathy with slowed nerve conduction velocities. The putative antinerve factor may be an antibody since injection of Fab fragments from the monoclonal immunoglobulin G produced a similar demyelination. This provides evidence of a circulating factor in the serum of myeloma patients with polyneuropathy that reproduces typical features of the human disease on passive transfer. This disorder is thus distinguished from other neuropathies that occur as remote effects of malignant disease but have no identified pathogenic factors associated with them.