Cerebral expression of interleukin-12 induces neurological disease via differential pathways and recruits antigen-specific T cells in virus-infected mice

Abstract

Transgenic expression of interleukin-12 (IL-12) in astrocytes causes a spontaneous inflammatory central nervous system disorder in aged mice. Here we show that spontaneous disorder developed only when both mature lymphocytes and interferon (IFN)-gamma were present. Infection with noncytolytic Borna disease virus (BDV) did not affect wild-type mice but accelerated disease of IL-12 transgenic mice. Infection of transgenic mice lacking lymphocytes did not result in neurological symptoms. In contrast, BDV infection of transgenic mice lacking IFN-gamma induced neurological disease with delayed onset of symptoms that resembled those in infected transgenic mice with a functional IFN-gamma gene. In BDV-infected transgenic mice devoid of IFN-gamma no cerebellar calcification was observed, and multiplication of BDV was not inhibited. To determine the antigen specificity of lymphocytes in brains of diseased animals, the IL-12 transgene was introduced into an H-2k genetic background. Infection of IL-12 transgenic H-2k mice resulted in extensive lymphocytic infiltration into the cerebellum but not into other brain regions that also contained viral antigen but expressed the transgene at lower levels. Tetramer analysis revealed that most CD8 T cells in the cerebellum of such mice were BDV-specific. Our results thus demonstrate that IFN-gamma secreting lymphocytes are responsible for disease of IL-12 transgenic mice. They further suggest that expression of IL-12 in the central nervous system may lead to localized recruitment of T cells that recognize antigens expressed in the brain.

Figure 1

Development of neurological disorder in BDV-infected, IL-12 transgenic C57BL/6 mice with and without defective RAG2 or IFN-γ genes. Newborn mice were infected intracerebrally with BDV as described in Materials and Methods. The animals were examined daily for neurological symptoms. Group sizes varied between 10 and 12 animals. The onset of disease in infected GF-IL12xIFNγ^−/− mice was significantly delayed as compared to infected GF-IL12 mice (*, P < 0.0005; Mann-Whitney) while all infected GF-IL12xRAG2^−/− mice stayed healthy.

Figure 2

Histological alterations in the cerebellum of BDV-infected GF-IL12 mice with and without functional IFN-γ or RAG2 genes. Newborn mice were infected intracerebrally with BDV as described in Materials and Methods. The animals were examined daily for neurological symptoms. Infected GF-IL12 and GF-IL12xIFNγ^−/− mice were symptomatic at the time of sacrifice while GF-IL12xRAG2^−/− mice were healthy. A and B: Cerebellum of diseased GF-IL12 mice. Note the meningeal, parenchymal, and perivascular (arrows) mononuclear infiltrates, granule cell loss, calcification (arrowheads) and extensive demyelination. C and D: Cerebellum of diseased GF-IL12xIFNγ^−/− mice. Note the similar but less pronounced alterations and the absence of calcifications. E and F: Cerebellum of healthy GF-IL12xRAG2^−/− mice. No overt histopathological alterations were found as compared to wild-type mice. gcl, granule cell layer; me, meninges; ml, molecular layer; wm, white matter. H&E stain: A, C, E; Klüver-Barrera myelin stain: B, D, F. Original magnifications, ×20.

Figure 3

Expression of NOS-2 (A) and BDV-N (B) in the cerebellum of BDV-infected IL-12 transgenic and nontransgenic mice with or without a functional IFNγ^−/− gene. Newborn mice were infected intracerebrally with BDV as described in Materials and Methods. Transgenic mice showed neurological symptoms. Healthy nontransgenic control mice were sacrificed at similar ages. Cerebellar RNA was prepared from individual mice (n = 4) and subjected to RNase protection assay as described in Materials and Methods. Autoradiographs were scanned and optical densities of the bands were measured. Graphs show mean values, standard deviations, and the statistical significance (Mann-Whitney).

Figure 4

Histological alterations in the cerebellum (A) and hippocampus (B) and flow cytometric analysis of brain lymphocytes (C–G) from BDV-infected IL-12 transgenic mice with pure C57BL/6 and mixed C57BL/6xB10.BR genetic background. Newborn mice were infected intracerebrally with BDV as described in Materials and Methods. The animals were examined daily for neurological symptoms. Mice were symptomatic at the time of sacrifice. A: There were mononuclear infiltrates in the meninges (arrows) as well as perivascular cuffing (arrowheads) and granule cell loss in the cerebellum of diseased GF-IL12 mice. B: There were no leukocyte infiltrates or signs of neurodegeneration in the hippocampus. C–F: CNS mononuclear cells isolated from brains of diseased BDV-infected GF-IL12 mice on a pure C57BL/6 (C, D) or mixed C57BL/6xB10.BR background (E, F) were stained for flow cytometric analysis with anti-CD8-APC alone (C, E) or double stained with anti-CD8-APC and H-2K^k/TELEISSI-PE tetramer (D, F). The percentages of CD8 T cells co-staining with the MHC I tetramer in the CD8 T-cell population are indicated by the numbers in the corresponding quadrant. Note that the majority of CD8 T cells were specific for TELEISSI, which is the immunodominant BDV epitope in mice with an H-2^k haplotype. G: CNS mononuclear cells isolated from the brain of a diseased BDV-infected C57BL/6xB10.BR animal were stained for flow cytometric analysis with anti-CD4-phycoerythrin and anti-CD8-fluorescein isothiocyanate. Numbers in the respective quadrants refer to the percentage of stained cells from all cells in the lymphocyte/monocyte gate. Note that CD4 and CD8 T cells were present at comparable frequencies. gcl, granule cell layer; ml, molecular layer; wm, white matter; ca, Cornu ammonis; dg, dentate gyrus. H&E stain. Original magnifications, ×10.