Short Communication. Alan R. Brown, 1 Jeanette Webb, 1 Selma Rebus, 2 Robert Walker, 1 Alun Williams 2 3 and John K. Fazakerley 1

Journal of General Virology (2003), 84, 2605 2611 DOI 10.1099/vir.0.19137-0 Short Communication Inducible cytokine gene expression in the brain in the ME7/CV mouse model of scrapie is highly restricted, is at a strikingly low level relative to the degree of gliosis and occurs only late in disease Alan R. Brown, 1 Jeanette Webb, 1 Selma Rebus, 2 Robert Walker, 1 Alun Williams 2 3 and John K. Fazakerley 1 Correspondence John Fazakerley John.Fazakerley@ed.ac.uk 1 Centre for Infectious Diseases, University of Edinburgh, Summerhall, Edinburgh EH9 1QH, UK 2 Institute of Comparative Medicine, Department of Veterinary Pathology, University of Glasgow, Glasgow, UK Received 3 February 2003 Accepted 26 May 2003 The temporal course of cerebral cytokine gene expression was investigated in the ME7/CV murine scrapie model to determine any association with neuropathological events. Analysis by RNase protection assay (RPA) demonstrated no transcripts for ILs 2, 3, 4, 5, 6, 7, 10, 12p40 and 13, granulocyte macrophage colony-stimulating factor, IFN-c or lymphotoxin-a at any time during the course of this disease. Transcripts for transforming growth factor-b1 were constitutively expressed in both control and scrapie-infected brain and were elevated at terminal disease. RPA and quantitative real-time RT-PCR detected low levels of transcripts for IL-1a, IL-1b and TNFa in scrapie-infected brain but only IL-1b was elevated consistently in all mice studied. Although glial cell activation within the hippocampus was evident from 100 days post-infection (p.i.), elevated IL-1b transcripts (and immunoreactivity) were evident from 180 days p.i., around the time of hippocampal pyramidal neuron loss, and increased steadily thereafter to reach a 3?5-fold increase at terminal disease. Even at their maximum, levels of these transcripts were disproportionately low relative to the degree of glial cell activation. It is concluded that cytokine gene expression in the ME7 scrapie-infected mouse brain, relative to the degree of reactive gliosis, is highly restricted, temporally late and disproportionately low. While the major neuropathological features of transmissible spongiform encephalopathies (TSEs) are well documented, many details of the neuropathogenesis remain obscure. A prominent feature of TSE neuropathology is glial cell activation, which is generally reported to occur after initial PrP Sc deposition but before the onset of neuronal loss. Proinflammatory cytokines are among the many glial cell products shown previously to be elevated in the TSEinfected brain. Three previous studies all showed elevation of IL-1a, IL-1b and TNFa in different scrapie/mouse strain combinations (Campbell et al., 1994; Kordek et al., 1996; Williams et al., 1994b, 1997b), with elevation of transcript typically coinciding with onset of clinical disease. In the 301V/VM scrapie model, elevation of cytokine immunoreactivity for IL-1b, IL-6 and TNFa has also been demonstrated prior to onset of observable clinical signs (Williams et al., 1997b). Previous studies of other cytokines have been more limited, although no increases in ILs 2, 3, 4 3Present address: Department of Pathology and Infectious Diseases, Royal Veterinary College, Hawkshead Lane, North Mymms, Hertfordshire AL9 7TA, UK. 0001-9137 G 2003 SGM Printed in Great Britain IP: 148.251.232.83 2605 and 5 or IFN-c were reported in the Chandler/SWRj scrapie model (Campbell et al., 1994). To determine any relationship between cytokine expression and neuropathological events, we have determined the temporal course of the expression of a large number of cytokines in the brains of C57BLxVM/Dk mice infected with the ME7 strain of scrapie. This (ME7/CV) model was chosen as it is well characterized; the timing and nature of neuronal pathology in the hippocampus has been described in detail (Jeffrey et al., 2000, 2001; Johnston et al., 1998). The temporal and spatial aspects of glial cell activation have received less attention in this model and are also described here. CV mice were inoculated intracerebrally with either control (uninfected) brain homogenate or with brain homogenate from a ME7-infected, clinically affected C57BL mouse, as described previously (Scott & Fraser, 1984). Control and ME7-infected mice were perfused with RNase-free PBS prior to being killed at 40, 70, 100, 130, 160, 170, 180 and 210 days post-infection (p.i.) and at terminal disease (225 235 days p.i.). Brains were removed and divided sagitally, with one-half of the brain being retained for RNA

A. R. Brown and others 2606 IP: 148.251.232.83 Journal of General Virology 84

Cytokine transcripts in ME7 scrapie analysis and the other half being fixed [buffered formol saline or paraformaldehyde/lysine periodate (PLP)], processed and embedded in paraffin wax. Further mice were perfused with 4 % paraformaldehyde for TUNEL (terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labelling) staining, PLP for PrP staining or PLP-containing glutaraldehyde for immunocytochemical studies of microglia. Immunocytochemical studies were performed as described previously (Williams et al., 1994a, 1997a), enabling the temporal progression of neuropathological features to be studied in detail (Fig. 1). To ensure staining specificity, sections in which the specific primary antibody was replaced by the appropriate normal serum (for polyclonal antibodies) or PBS (for monoclonals) acted as controls. Some PrP staining, detected using the 4F2 (a gift from J. Grassi, CEA, France) and 6H4 (Prionics) antibodies, was seen as punctate intracellular immunoreactivity of occasional (rare), scattered neurons in control (uninfected) mice. In infected mice, additional PrP immunoreactivity in the hippocampus was detected first at 100 days p.i. and gradually increased thereafter. At 100 and 130 days p.i., this increased staining was detected as a diffuse extracellular staining, particularly in the molecular layer of the CA1 hippocampal region (Fig. 1A); at later stages, more intense punctate extracellular deposits were evident (Fig. 1B) and increased PrP immunoreactivity was distributed more widely throughout the hippocampus. Increased punctate staining was observed in the thalamus from 70 days p.i.; thereafter, the pattern of PrP staining in both the hippocampus and the thalamus was consistent with previous reports (Jeffrey et al., 2000, 2001). To assess glial cell activation, astrocytes were identified using a rabbit anticow GFAP (glial fibrillary acidic protein) antibody (Dako) and microglia using the monoclonal antibodies FA11 (anti- CD68), F4/80 (both obtained from Serotec) and TIB122 (anti-leukocyte common antigen), the hybridoma for which was obtained from ECC. Activation of both microglia and astrocytes, evaluated as intensity of staining and number of immunopositive cells, was detected first in the hippocampus at 100 days p.i. (Fig. 1C, E), a time at which initial morphological changes in hippocampal neurons have been described previously (Jeffrey et al., 2000). Initial microglial staining was apparent particularly when using the FA11 antibody. Resting microglia do not stain with this antibody. In the hippocampus, focal staining of microglia in the CA1 hippocampal region was observed first at 100 days p.i. (Fig. 1C), although immunoreactivity was evident in the thalamus from 70 days p.i.. Thereafter, the extent of microglial activation, as identified by FA11 immunopositivity and by increased staining for the F4/80 and leukocyte common antigen, increased steadily in both intensity and distribution within the hippocampus until the terminal stages of disease (Fig. 1D). Similarly, the intensity of GFAP staining within the hippocampus was slightly greater than control mice at 100 days p.i. (Fig. 1E) and increased thereafter to show greatest intensity and numbers of GFAP-positive cells by the time of terminal disease (Fig. 1F). These increases in glial cell immunostaining, indicating microglial and astrocyte activation, paralleled increases in PrP immunoreactivity. Initially (100 and 130 days p.i.), glial cell activation was seen as increased intensity of staining and increased number of immunopositive cells with a morphology similar to that of normal (resting) microglia and astrocytes, whereas at later timepoints, increasing numbers of immunoreactive cells showed morphological changes with cell processes becoming reduced in number, broader and shorter and the cell body becoming more prominent and rounded. These morphological changes were consistent with a progressive increase in the extent of glial cell activation. None of the sections stained without primary antibody showed any specific immunoreactivity. Vacuolation within the hippocampus was seen first as scattered individual small vacuoles at 130 days p.i. and more frequent, larger vacuoles from 160 days p.i. onwards; again, this was consistent with previous reports (Jeffrey et al., 2000). TUNEL staining, performed as described previously (Lucassen et al., 1995; Williams et al., 1997a), demonstrated that hippocampal CA1 pyramidal neuron loss was most evident at 170 days p.i. (Fig. 1G), reflecting the known and approximately 50 % loss of these cells between 160 and 180 days p.i. in this model (Jeffrey et al., 2000). To relate the observed course of neuropathology and glial cell activation to any changes in cytokine gene expression, RNA was prepared from half brain samples and analysed by RNase protection assay (RPA). The RPA was performed as described previously (Hobbs et al., 1993), using the multiprobe cdna templates ML-11 [lymphotoxin-a (LTa), TNFa, IL-4, IL-5, IL-1a, IFN-c, IL-2, IL-6, IL-1b, IL-3 and Fig. 1. Hippocampal neuropathology in the ME7/CV mouse model of scrapie. (A, B) PrP immunoreactivity (brown staining) in the middle of the hippocampus at 100 days p.i. (A) and at terminal disease (B). At 100 days p.i., PrP staining is seen as a diffuse immunoreactivity of the CA1 region and shows as a darker staining in the molecular layer of the hippocampal cortex (*). No such staining is seen in hippocampal sections of control mice. At terminal disease, further discrete, more intensely stained foci of PrP reactivity are scattered throughout this region of the hippocampus (arrow). (C, D) FA11 immunoreactivity (brown staining) showing a focus (*) of microglial reactivity (mild activation) at 100 days p.i. (C) and widespread microglial activation at terminal disease (D). (E, F) GFAP immunoreactivity (brown staining) showing astrocytes at 100 days p.i. (E) and at terminal disease (F). The density of astrocytes is greater at terminal disease. (G) TUNEL staining (brown) of apoptotic hippocampal CA1 pyramidal neurons (arrow) at 170 days p.i. (H) IL-1b immunoreactivity at terminal disease. Many of the immunopositive cells have a morphology consistent with that of astrocytes (arrow). Bars, 10 mm (A, B), 5 mm (C, D) or 2?5 mm (E H). http://vir.sgmjournals.org IP: 148.251.232.83 2607

A. R. Brown and others L32) and ML-26 [IL-3, IL-10, granulocyte macrophage colony-stimulating factor (GM-CSF), transforming growth factor-b1 (TGFb1), IL-13, IL-12p40, IL-12p35, IL-7 and L32], kindly provided by M. Hobbs (Scripps Research Institute, La Jolla, USA). Initial analysis, performed with total RNA prepared using TRIzol reagent (Invitrogen), failed to detect the expression of any cytokine transcripts in ME7-infected brain with the exception of TGFb1, which was expressed constitutively. To increase sensitivity, poly(a) + RNA was prepared (Badley et al., 1988) and 9 mg, representing maximum gel loading, was assayed. After extended exposure to the phosphorimager screen, transcripts for IL-1b were detected consistently in all ME7- infected animals from 180 days p.i., which is prior to clinical disease. Thereafter, levels of IL-1b transcript increased steadily to terminal disease (Fig. 2A). At terminal disease, transcripts for IL-1a and TNFa were also detected in some mice (Fig. 2A). Relative quantification between samples was determined by reference of each band to its L32 ribosomal RNA band. Even at their highest (IL-1b at terminal), the levels of these cytokine transcripts did not exceed 1 % of that of the L32 transcripts (Table 1). Transcripts for TGFb1 were readily detectable at all time-points (Fig. 2B) and were increased relative to controls at terminal disease (Table 1). With the exception of TGFb1 and IL-12p35, which are known to be expressed constitutively in the brain (Morgan et al., 1993; Park & Shin, 1996), no cytokine transcripts were detectable by RPA in the brains of normal brain homogenate-inoculated, age- and sex-matched control mice. The gels depicted in Fig. 2(A, B) are representative of the time-course observed by multiple RPA analyses, including at least two animals at each time-point. Brains from SCID mice infected with Semliki Forest virus (SFV) strain A7(74) were also analysed for their levels of cytokine transcripts, to provide a comparison with the scrapie-infected brain. Following intraperitoneal infection, SFV A7(74) enters the brain to produce small perivascular foci of infection, which, in SCID mice, persist for weeks without any influx of inflammatory cells from the blood (Amor et al., 1996; Fazakerley et al., 1993). The small foci of infection have local microglial and astrocyte activation, which is confined to these foci and does not involve large areas of the brain. The levels of cytokine transcripts in the brain of each of two mice at 5, 7 and 9 days p.i. were assessed by RPA using the ML-11 probe set (Fig. 2C). Even at 5 days p.i., the levels of IL-1b and TNFa transcripts in the SFVinfected SCID mice were consistently at least 10-fold higher than the levels in the ME7-infected brains at terminal disease (Table 1). This is despite the RPA analysis of SFV-infected SCID brains being performed on total RNA, in contrast to the poly(a) + RNA from scrapie samples. Levels of IL-1a in Fig. 2. RPA analysis of mrna cytokine levels within brain homogenates derived from scrapie-infected (A, B) and SFVinfected (C) mice. Gels shown are representative of the time-course that was observed by analysis of multiple mice at each time-point, using ML-11 (A, C) and ML-26 (B) riboprobes. Time-matched control samples from mice inoculated with normal brain homogenate (NBH) were analysed also. Each assay included a sense probe control (Ref) to serve as a reference standard. As an engineered part of the assay, the sizes of the protected mrna bands are smaller than the corresponding labelled probe reference bands. In (A), the L32 bands were highly overexposed and are shown separated from the rest of the image and at a lower exposure. 2608 IP: 148.251.232.83 Journal of General Virology 84

Cytokine transcripts in ME7 scrapie Table 1. Relative levels of transcripts as determined by RPA Densitometric analysis of the RPA gels shown in Fig. 2. NBH, Normal brain homogenate; Term, terminal disease; ND, not detected. Inoculum Time (days p.i.) Percentage relative to L32* IL-1a IL-1b TNFa LTa ME7 180 ND 0?20 ND ND 200 ND 0?32 0?07 ND 230 ND 0?40 0?06 ND Term 0?16 0?66 0?23 ND SFV 5 2?71 6?36 2?82 2?31 5 ND 2?43 1?35 0?77 7 3?17 5?72 2?50 1?31 7 ND 2?59 1?46 0?77 9 6?04 9?00 6?10 2?29 9 4?70 8?56 3?50 1?07 Inoculum Time Percentage relative to L32* TGFb1 GFAP Mac-1 EB22/5.3 NBH TermD 4?91 69?36 ND 6?15 Mock TermD 4?67 61?96 2?06 6?27 ME7 Term 17?66 144?32 8?57 67?00 *The intensity of each band is expressed as a percentage of the intensity of the L32 rrna band. DBoth control mice (NBH- or mock-inoculated) were time-matched to the terminal ME7 mouse. the SFV-infected brains were more variable, consistent with the variable number of foci of brain infection observed but, when detected, were at all times considerably higher than the levels observed in ME7-infected brain. To ensure that the low level of expression of specific cytokines observed in the scrapie-infected brain samples by RPA was not due to any degradation or inhibitor activity within these samples, the samples were used to measure the expression of transcripts expected to be present and elevated in the ME7-infected brain using a CNS inflammation probe set (ICAM-1, inos, A20, Mac-1, EB22/5.3, GFAP and L32), kindly provided by I. Campbell (Scripps Research Institute). At terminal disease, significant upregulation of GFAP (a marker of astrocytosis), Mac-1 (a marker for microglial activation) and EB22/5.3 (an acute-phase response gene shown previously to be elevated in the Chandler/SWRj scrapie model; Campbell et al., 1994) was detected readily (Table 1), thus validating the RPA methodology. Further analysis of IL-1a, IL-1b and TNFa cytokine transcripts in ME7-infected brain was performed by quantitative real-time RT-PCR, the greater sensitivity of the assay enabling direct comparisons between transcript levels in control and infected samples. cdna, prepared from http://vir.sgmjournals.org IP: 148.251.232.83 2609 RNA derived from control and terminal stage ME7-infected brain, was used as template in real-time PCR using the LightCycler Real-Time PCR system with the LightCycler FastStart DNA Master kit SYBR Green I (Roche Molecular Biochemicals), according to the manufacturer s instructions. Two pairs of nested PCR primers were designed for each of b-actin, IL-1a, IL-1b and TNFa. Serial dilutions of the first-round PCR product (generated using the external primer pair) acted as standards for second-round PCR, performed on the LightCycler using the internal primer pair. cdna templates were normalized using b-actin and amplified alongside the appropriate standards and controls, enabling a relative quantification of transcript levels in control and infected samples. In accordance with RPA results, only in their IL-1b response did the terminal stage scrapie-infected mice show some degree of uniformity, with IL-1b being upregulated in all mice studied, with a 3?5-fold increase in mean transcript levels relative to controls (P<0?02, Student s t-test). At the group level, there was no significant difference between control and scrapieinfected mice in the levels of IL-1a and TNFa. However, at the level of individual mice, some infected mice did demonstrate reproducibly increased levels of IL-1a and TNFa, suggesting variability between individuals at terminal disease. To determine whether the observed IL-1b gene upregulation was also apparent at the protein level, cryostat brain sections were stained with a sheep anti-mouse IL-1b antibody (NIBSC) and a peroxidase/anti-peroxidase method. IL-1b protein was not detectable at the earliest stages of microglial activation (100 and 130 days p.i.) but became detectable at low levels at 180 days p.i. as a few, weakly positive cells associated with the CA1 hippocampal region. Staining increased progressively thereafter until terminal disease when numerous, scattered, moderately positive immunoreactive glia were detectable in the hippocampus (Fig. 1H), thalamus and, to a lesser extent, cerebral cortex. The morphology of many of the IL-1b-positive cells was consistent with that of astrocytes, again reflecting previous reports of IL-1 immunostaining in other murine scrapie models (Kordek et al., 1996; Williams et al., 1994b, 1997b). These data are in good agreement with the first detection of transcripts at 180 days p.i. (Fig. 2A) with a steady increase thereafter. This study has demonstrated that induction of cytokines in the brains of ME7-infected mice is a late event, occurs at a low level and is restricted to IL-1a,IL-1b, TNFa and TGFb1. Only IL-1b and TGFb1 were elevated consistently in scrapieinfected mice; however, their relationship, causative or consequent, to neuropathology remains unclear. The relatively early (180 days p.i.) upregulation of IL-1b makes it a more likely candidate for cause or contribution to disease than the late upregulation of TGFb1. Elevated IL-1b transcript is a consistent finding in other scrapie models (Campbell et al., 1994; Kim et al., 1999; Williams et al., 1994b, 1997b) and IL-1 receptor (type I)-deficient mice have

A. R. Brown and others an extended disease course following inoculation with the 139A scrapie agent (Baier et al., 2002). Taken together, this evidence indicates that a role for IL-1b in TSE neuropathology cannot be ruled out. In contrast, current studies suggest that other proinflammatory cytokines, including IL-1a, IL-6 and TNFa, are unlikely to have any fundamental role in the initiation or development of neuropathology, since elevation of these transcripts was not detected or was only detectable at terminal disease and only in some mice. This is consistent with the finding that mice devoid of either IL-6 or TNFa do not show prolonged periods of incubation following intracerebral inoculation (Mabbott et al., 2000). A striking feature of the results obtained in this study was the limited extent of upregulation of the proinflammatory cytokines. Even IL-1b, the predominant proinflammatory cytokine in this study, was only increased 3?5-fold in terminally affected animals relative to control animals. In contrast, the extent of gliosis in the ME7/CV scrapie model, described in detail here for the first time, was marked. The current study compared events in ME7-infected brains to those in SFV-infected SCID mouse brains. Despite glial cell activation being significantly less in virus-infected SCID mice than the marked activation observed in the scrapieinfected brain, upregulation of IL-1a, IL-1b, TNFa and LTa (as determined by RPA) was significantly greater. We conclude that the extent of cytokine induction in scrapieinfected brains is disproportionate to the level of gliosis, at least relative to events in a CNS virus infection. A previous study of cytokine induction in ME7 infection of the related C57BL/6 mouse did not detect the upregulation of any proinflammatory cytokines (Walsh et al., 2001), whereas our study observed a slight increase in a limited number of cytokines. This difference may be due to the differing sensitivities of the techniques used in the two studies and the differing time-course of disease in these two mouse models. Furthermore, the study on C57BL/6 mice used stereotaxic intracerebral inoculation of small volumes of brain homogenate for infection, which may result in a different pathogenesis of disease to that generated by our conventional intracerebral inoculation. However, the current finding of TGFb1 upregulation at terminal disease is in agreement with that observed in the ME7/C57BL/6 model (Cunningham et al., 2002). In conclusion, our results suggest strongly that ILs 1a,2,3,4, 5, 6, 7, 10, 12 and 13, GM-CSF, IFN-c, LTa and TNFa have no role in neuropathological events in scrapie. IL-1b and TGFb1 were elevated consistently in scrapie-infected mice, with only IL-1b elevated prior to clinical disease, occurring around the time of hippocampal neuron loss. A direct role for IL-1b in TSE neuropathology cannot be excluded. However, it should be stressed that although IL-1b transcripts are elevated in scrapie-infected brain, their levels are disproportionately low relative to the extent of the glial cell response. 2610 IP: 148.251.232.83 Journal of General Virology 84 ACKNOWLEDGEMENTS We thank Monte Hobbs and Iain Campbell for their kind gifts of RPA probe template sets and Lisa Jarvis and the Scottish National Blood Transfusion Service Research Laboratories for their help and advice with real-time RT-PCR. We are grateful to Paul Lucassen of the Amsterdam Academic Medical Centre for assistance with TUNEL staining. This research was supported by a grant (15/BS410534) to J. K. F. and A. W. from the British Biotechnology and Biological Sciences Research Council. REFERENCES Amor, S., Scallan, M. F., Morris, M. M., Dyson, H. & Fazakerley, J. K. (1996). Role of immune responses in protection and pathogenesis during Semliki Forest virus encephalitis. J Gen Virol 77, 281 291. Badley, J. E., Bishop, G. A., St John, T. & Frelinger, J. A. (1988). A simple, rapid method for the purification of poly A + RNA. Biotechniques 6, 114 116. Baier, M., Schultz, J., Neidhold, S., Simon, D. & Schwartz, A. (2002). Influence of interleukin-1 on astrocytosis and microgliosis during experimental scrapie infection of mice. International Conference on Transmissible Spongiform Encephalopathies (Edinburgh, UK, 15 18 September, 2002). Abstract 1.57. Campbell, I. L., Eddleston, M., Kemper, P., Oldstone, M. B. & Hobbs, M. V. (1994). Activation of cerebral cytokine gene expression and its correlation with onset of reactive astrocyte and acute-phase response gene expression in scrapie. J Virol 68, 2383 2387. Cunningham, C., Boche, D. & Perry, V. H. (2002). Transforming growth factor b1, the dominant cytokine in murine prion disease: influence on inflammatory cytokine synthesis and alteration of vascular extracellular matrix. Neuropathol Appl Neurobiol 28, 107 119. Fazakerley, J. K., Pathak, S., Scallan, M., Amor, S. & Dyson, H. (1993). Replication of the A7(74) strain of Semliki Forest virus is restricted in neurons. Virology 195, 627 637. Hobbs, M. V., Weigle, W. O., Noonan, D. J., Torbett, B. E., McEvilly, R. J., Koch, R. J., Cardenas, G. J. & Ernst, D. N. (1993). Patterns of cytokine gene expression by CD4 + T cells from young and old mice. J Immunol 150, 3602 3614. Jeffrey, M., Halliday, W. G., Bell, J., Johnston, A. R., MacLeod, N. K., Ingham, C., Sayers, A. R., Brown, D. A. & Fraser, J. R. (2000). Synapse loss associated with abnormal PrP precedes neuronal degeneration in the scrapie-infected murine hippocampus. Neuropathol Appl Neurobiol 26, 41 54. Jeffrey, M., Martin, S., Barr, J., Chong, A. & Fraser, J. R. (2001). Onset of accumulation of PrP res in murine ME7 scrapie in relation to pathological and PrP immunohistochemical changes. J Comp Pathol 124, 20 28. Johnston, A. R., Fraser, J. R., Jeffrey, M. & MacLeod, N. (1998). Synaptic plasticity in the CA1 area of the hippocampus of scrapieinfected mice. Neurobiol Dis 5, 188 195. Kim, J. I., Ju, W. K., Choi, J. H., Choi, E. K., Carp, R. I., Wisniewski, H. M. & Kim, Y. S. (1999). Expression of cytokine genes and increased nuclear factor-kappa B activity in the brains of scrapie-infected mice. Brain Res Mol Brain Res 73, 17 27. Kordek, R., Nerurkar, V. R., Liberski, P. P., Isaacson, S., Yanagihara, R. & Gajdusek, D. C. (1996). Heightened expression of tumor necrosis factor a, interleukin 1 a, and glial fibrillary acidic protein in experimental Creutzfeldt Jakob disease in mice. Proc Natl Acad Sci USA93, 9754 9758. Lucassen, P. J., Williams, A., Chung, W. C. & Fraser, H. (1995). Detection of apoptosis in murine scrapie. Neurosci Lett 198, 185 188.

Cytokine transcripts in ME7 scrapie Mabbott, N. A., Williams, A., Farquhar, C. F., Pasparakis, M., Kollias, G. & Bruce, M. E. (2000). Tumor necrosis factor a-deficient, but not interleukin-6-deficient, mice resist peripheral infection with scrapie. J Virol 74, 3338 3344. Morgan, T. E., Nichols, N. R., Pasinetti, G. M. & Finch, C. E. (1993). TGF-b 1 mrna increases in macrophage/microglial cells in the hippocampus in response to deafferentation and kainic acid-induced neurodegeneration. Exp Neurol 120, 291 301. Park, J. H. & Shin, S. H. (1996). Induction of IL-12 gene expression in the brain in septic shock. Biochem Biophys Res Commun 224, 391 396. Scott, J. R. & Fraser, H. (1984). Degenerative hippocampal pathology in mice infected with scrapie. Acta Neuropathol (Berl) 65, 62 68. Walsh, D. T., Betmouni, S. & Perry, V. H. (2001). Absence of detectable IL-1b production in murine prion disease: a model of chronic neurodegeneration. J Neuropathol Exp Neurol 60, 173 182. Williams, A. E., Lawson, L. J., Perry, V. H. & Fraser, H. (1994a). Characterization of the microglial response in murine scrapie. Neuropathol Appl Neurobiol 20, 47 55. Williams, A. E., van Dam, A. M., Man-A-Hing, W. K., Berkenbosch, F., Eikelenboom, P. & Fraser, H. (1994b). Cytokines, prostaglandins and lipocortin-1 are present in the brains of scrapie-infected mice. Brain Res 654, 200 206. Williams, A., Lucassen, P. J., Ritchie, D. & Bruce, M. (1997a). PrP deposition, microglial activation, and neuronal apoptosis in murine scrapie. Exp Neurol 144, 433 438. Williams, A., van Dam, A. M., Ritchie, D., Eikelenboom, P. & Fraser, H. (1997b). Immunocytochemical appearance of cytokines, prostaglandin E2 and lipocortin-1 in the CNS during the incubation period of murine scrapie correlates with progressive PrP accumulations. Brain Res 754, 171 180. http://vir.sgmjournals.org IP: 148.251.232.83 2611