Differential expression of interferon responsive genes in rodent models of transmissible spongiform encephalopathy disease

Abstract

Background: The pathological hallmarks of transmissible spongiform encephalopathy (TSE) diseases are the deposition of a misfolded form of a host-encoded protein (PrPres), marked astrocytosis, microglial activation and spongiosis. The development of powerful gene based technologies has permitted increased levels of pro-inflammatory cytokines to be demonstrated. However, due to the use of assays of differing sensitivities and typically the analysis of a single model system it remained unclear whether this was a general feature of these diseases or to what extent different model systems and routes of infection influenced the relative levels of expression. Similarly, it was not clear whether the elevated levels of cytokines observed in the brain were accompanied by similar increases in other tissues that accumulate PrPres, such as the spleen.

Results: The level of expression of the three interferon responsive genes, Eif2ak2, 2'5'-OAS, and Mx2, was measured in the brains of Syrian hamsters infected with scrapie 263K, VM mice infected with bovine spongiform encephalopathy and C57BL/6 mice infected with the scrapie strain ME7. Glial fibrillary acidic expression confirmed the occurrence of astrocytosis in all models. When infected intracranially all three models showed a similar pattern of increased expression of the interferon responsive genes at the onset of clinical symptoms. At the terminal stage of the disease the level and pattern of expression of the three genes was mostly unchanged in the mouse models. In contrast, in hamsters infected by either the intracranial or intraperitoneal routes, both the level of expression and the expression of the three genes relative to one another was altered. Increased interferon responsive gene expression was not observed in a transgenic mouse model of Alzheimer's disease or the spleens of C57BL/6 mice infected with ME7. Concurrent increases in TNFalpha, TNFR1, Fas/ApoI receptor, and caspase 8 expression in ME7 infected C57BL/6 mice were observed.

Conclusion: The identification of increased interferon responsive gene expression in the brains of three rodent models of TSE disease at two different stages of disease progression suggest that this may be a general feature of the disease in rodents. In addition, it was determined that the increased interferon responsive gene expression was confined to the CNS and that the TSE model system and the route of infection influenced the pattern and extent of the increased expression. The concurrent increase in initiators of Eif2ak2 mediated apoptotic pathways in C57BL/6 mice infected with ME7 suggested one mechanism by which increased interferon responsive gene expression may enhance disease progression.

Figure 1

Expression levels of interferon responsive genes in four different models of amyloid disease. Pooled samples of total RNA isolated from the whole brains of diseased or age-matched control animals were used to determine relative gene expression levels by quantitative real-time PCR. The histograms represent the average of the means of three independent quantitative real-time PCR reactions, done in duplicate, for each gene ± standard deviation. All plotted data represent gene expression levels in diseased animals relative to gene expression levels in age and strain-matched control animals. The days post infection (dpi) indicated correspond to the times at which the onset of clinical symptoms and terminal stage disease are observed in the different model systems. At the onset of clinical symptoms (panel A) significantly higher expression levels of the three interferon responsive genes were observed in all the models of TSE disease via the IC route (p ≤ 0.008). No significant increase was observed in the hamster scrapie 263K 63 day IP sample or the similar disease stage transgenic mouse model of Alzheimer's disease. At the terminal stage of disease (panel B) the expression of the three interferon responsive genes was significantly increased in all rodent models of TSE disease (p ≤ 0.002) but not in the Alzheimer's disease model (p ≤ 0.32).

Figure 2

Expression level of GFAP in four different models of amyloid disease. Pooled samples of total RNA isolated from the whole brains of diseased or age-matched control animals were used to determine relative GFAP expression levels by quantitative real-time PCR. The histograms represent the average of the means of three independent quantitative real-time PCR reactions, done in duplicate, ± standard deviation. All plotted data represent expression levels in diseased animals relative to gene expression levels in age and strain-matched control animals. A statistically significant increase in GFAP expression was observed in all terminal disease models (p = 3.44E^-5 for Alzheimer's model, p = 0.000364 for VM BSE mice, p = 3.24E^-5 for Hamster 263K 56 IC, p = 0.0068 for Hamster 263K IP, and p = 0.0014 for C57BL/6 ME7 mice).

Figure 3

IFN responsive gene expression in the brains and spleens of individual ME7 infected and mock infected C57BL/6 mice. cDNA from whole brains and spleens of 9 ME7 infected and 10 mock infected C57BL/6 mice that were analyzed in a blinded fashion. Equal amounts of cDNA from the 10 mock infected samples were pooled to make an eleventh sample. Each data point represents the mean of three independent quantitative real-time PCR reactions for each gene normalized to the average level of expression exhibited by the mock infected samples. Filled diamonds represent those samples obtained from brains of ME7 infected C57BL/6 mice. The shaded circles represent cDNA samples obtained from age-matched mock infected C57BL/6 mice. The lines represent the mean value of the 20 samples for each gene. A clear segregation of the infected and mock-infected samples based upon the relative expression of any of the three interferon responsive genes was observed in brain tissue (A, Eif2ak2 p = 1.39E^-23, 2'5'-OAS p = 7.04E^-21, Mx2 p = 7.08E^-26). A random distribution of infected and mock infected samples derived from the corresponding spleen tissue relative to the mean of the sample set was observed (B, Eif2ak2 p = 0.75, 2'5'-OAS p = 0.80, Mx2 p = 0.48).

Figure 4

Increase pro-apoptotic gene expression in the brains of individual ME7 infected and mock infected C57BL/6 mice. cDNA from whole brains of 9 ME7 infected and 8 mock infected C57BL/6 mice were analyzed in a blinded fashion. Each data point represents the mean of three independent quantitative real-time PCR reactions for each gene normalized to the average level of expression exhibited by the mock infected samples. Filled diamonds represent those samples obtained from brains of ME7 infected C57BL/6 mice. The shaded circles represent cDNA samples obtained from age-matched mock infected C57BL/6 mice. The lines represent the mean value of the 17 samples for each gene. A clear segregation of the infected and mock-infected samples based upon the relative expression of TNFα, TNFR1, Fas/ApoI and caspase 8 was observed in brain tissue. In contrast, FasL was not significantly differentially expressed between infected and mock infected samples (TNFα p = 3.74E^-32, TNFR1 p = 1.03E^-25, FasL p = 0.071, Fas/ApoI p = 6.62E^-27, and caspase 8 p = 1.37E^-15).