Prion Infection of Mouse Brain Reveals Multiple New Upregulated Genes Involved in Neuroinflammation or Signal Transduction

ABSTRACT

Gliosis is often a preclinical pathological finding in neurodegenerative diseases, including prion diseases, but the mechanisms facilitating gliosis and neuronal damage in these diseases are not understood. To expand our knowledge of the neuroinflammatory response in prion diseases, we assessed the expression of key genes and proteins involved in the inflammatory response and signal transduction in mouse brain at various times after scrapie infection. In brains of scrapie-infected mice at pre- and postclinical stages, we identified 15 previously unreported differentially expressed genes related to inflammation or activation of the STAT signal transduction pathway. Levels for the majority of differentially expressed genes increased with time postinfection. In quantitative immunoblotting experiments of STAT proteins, STAT1α, phosphorylated-STAT1α (pSTAT1α), and pSTAT3 were increased between 94 and 131 days postinfection (p.i.) in brains of mice infected with strain 22L. Furthermore, a select group of STAT-associated genes was increased preclinically during scrapie infection, suggesting early activation of the STAT signal transduction pathway. Comparison of inflammatory markers between mice infected with scrapie strains 22L and RML indicated that the inflammatory responses and gene expression profiles in the brains were strikingly similar, even though these scrapie strains infect different brain regions. The endogenous interleukin-1 receptor antagonist (IL-1Ra), an inflammatory marker, was newly identified as increasing preclinically in our model and therefore might influence scrapie pathogenesis in vivo. However, in IL-1Ra-deficient or overexpressor transgenic mice inoculated with scrapie, neither loss nor overexpression of IL-1Ra demonstrated any observable effect on gliosis, protease-resistant prion protein (PrPres) formation, disease tempo, pathology, or expression of the inflammatory genes analyzed.

IMPORTANCE Prion infection leads to PrPres deposition, gliosis, and neuroinflammation in the central nervous system before signs of clinical illness. Using a scrapie mouse model of prion disease to assess various time points postinoculation, we identified 15 unreported genes that were increased in the brains of scrapie-infected mice and were associated with inflammation and/or JAK-STAT activation. Comparison of mice infected with two scrapie strains (22L and RML), which have dissimilar neuropathologies, indicated that the inflammatory responses and gene expression profiles in the brains were similar. Genes that increased prior to clinical signs might be involved in controlling scrapie infection or in facilitating damage to host tissues. We tested the possible role of the endogenous IL-1Ra, which was increased at 70 days p.i. In scrapie-infected mice deficient in or overexpressing IL-1Ra, there was no observable effect on gliosis, PrPres formation, disease tempo, pathology, or expression of inflammatory genes analyzed.

INTRODUCTION

Prion diseases are infectious progressive neurodegenerative disorders that can affect both humans and animals. These disorders include sporadic Creutzfeldt-Jakob disease (sCJD), variant Creutzfeldt-Jakob disease (vCJD), and Gerstmann-Sträussler-Scheinker syndrome (GSS) in humans and bovine spongiform encephalopathy (BSE), chronic wasting disease (CWD), and scrapie in animals. As a group, these diseases are often referred to as transmissible spongiform encephalopathies (TSEs). A prominent feature of this class of neurodegenerative diseases is the early onset of gliosis in the brains of infected hosts (1, 2). The direct cause of this gliosis is unclear, but evidence of microglial and astroglial activation coincides with the detection of disease-associated protease-resistant prion protein (PrPres) (3). However, gliosis and PrPres deposition precede morphological evidence of neuronal damage and neuropil vacuolation in the brain (4, 5), suggesting that both PrPres and gliosis might contribute to neuronal damage in prion disease.

Numerous scrapie strains have been cloned through multiple passages in animals by using limiting dilution techniques. Though many of the clinical signs are similar when mice are infected with various scrapie strains, several strains contrast with regard to incubation period as well as the lesion profile in the brain when injected intracerebrally (6–8). Mice clinically infected with strain 22L present with greater vacuolation and PrPres deposition in the dorsal medulla, the superior colliculus, the hypothalamus, and the septum than mice similarly infected with strain RML (6, 9). Moreover, 22L-infected mice also demonstrate severe vacuolation and PrPres deposition in the cerebellum, but infection with strain RML rarely results in cerebellar attack (9, 10). In contrast, mice clinically infected with strain RML exhibit increased vacuolation in the hippocampus relative to 22L-infected mice (6, 9). This suggests that scrapie strains 22L and RML can infect different regions of the brain with varied clinical outcomes.

We and several other investigators have attempted to analyze the progression of prion disease by using mouse-adapted prion diseases as model systems. Various high-throughput techniques, such as microarray expression profiling (11–19) and quantitative bead-based suspension array systems (3, 20, 21), have allowed elucidation of transcriptional and protein changes in brains of prion-infected mice relative to controls. These studies supported the notion that prion diseases have a neuroinflammatory component that may play a critical role in neurodegeneration (22), with increases in numerous cytokines and chemokines, such as interleukin-1α (IL-1α) and IL-1β, IL-12p40, tumor necrosis factor (TNF), CCL2 to CCL6, and CXCL10 in the brains of mice with clinical disease. Interestingly, the pattern and magnitude of cytokine expression in the brain during prion infection is unlike what is typically observed in bacterial infections, viral infections, or traumatic injury of the central nervous system (CNS) (3, 21, 23, 24), as fewer cytokines are elevated and the levels are lower than in conventional CNS infections.

Many of the inflammatory mediators produced during prion-induced gliosis, such as CCL2, CCL3, CCL4, CCL5, and IL-12, are known activators of the JAK-STAT pathway (25, 26). Na et al. detected phosphorylation (p) of STAT1 and STAT3 in scrapie-infected mice, and it has been suggested that JAK2-STAT1 signaling may be important in facilitating astrogliosis during infection with scrapie strain ME7 (27). Furthermore, those same authors showed an association of pSTAT1 and pSTAT3 with activated astrocytes in scrapie-infected brain. However, beyond this information, there is little known about the transcriptional changes that accompany STAT activation in the brain during prion disease.

Several laboratories have identified global changes in transcription at various times in preclinical and postclinical scrapie-infected brains by using hybridization arrays (12–14, 16, 17, 19, 28, 29). Global microarray technology can be used to assess a large number of transcriptional changes simultaneously, but this approach may overlook genes that encode cytokines, chemokines, and transcription factors, due to their low levels of expression (30, 31). To search for differentially expressed genes at early times after prion infection, we used a highly sensitive quantitative reverse transcription-PCR (qRT-PCR) array assay. Key genes in inflammation and signal transduction were analyzed in brains of scrapie-infected mice at various days postinfection. We also compared alterations in transcription between mice infected with either scrapie strain 22L or RML. Fifteen genes that were previously unidentified by global microarray technology were significantly upregulated in our study in scrapie-infected mice. In addition, we assessed the influence of one of these newly identified genes (Il1rn), which encodes IL-1Ra, a receptor antagonist of the IL-1 signaling pathway, on scrapie pathogenesis by using knockout and transgenic mice.

MATERIALS AND METHODS

Mice and scrapie inoculations.

All mice were housed at the Rocky Mountain Laboratories (RML) in an AAALAC-accredited facility, and experimentation followed NIH RML Animal Care and Use Committee-approved protocols. Female C57BL/10 or C57BL/6 mice were anesthetized with isoflurane and inoculated intracerebrally (i.c.) by one of two methods. For time course analysis with C57BL/10 mice, the left and right hemispheres of mice at 4 to 6 weeks of age were obtained along with 20 μl per hemisphere of a 1.0% (wt/vol) 22L scrapie brain homogenate stock (final concentration, 8.0 × 10⁵ of the 50% lethal dose [LD₅₀]) or 1.0% (wt/vol) normal brain homogenate stock (mock control) in phosphate-buffered balanced saline with 2% fetal bovine serum. Following infection, mice were euthanized at preclinical time points or monitored for onset of scrapie signs. Six to eight mice were euthanized at 21, 44, 70, 94 (collectively preclinical) and 131 (postclinical) days postinfection (p.i.), and brains were collected.

For comparison of strains 22L and RML in C57BL/6 mice, only the left hemisphere was inoculated with 30 μl of a 1.0% (wt/vol) 22L scrapie brain homogenate stock (final concentration of 6.0 × 10⁵ LD₅₀), RML brain homogenate stock (final, 2.4 × 10⁴ LD₅₀), or normal brain homogenate stock (mock control) in phosphate-buffered balanced saline with 2% fetal bovine serum. Mice were euthanized when they displayed persistent signs of ataxia, kyphosis, somnolence, and hind leg weakness as described previously (32). Collected hemispheres were placed in formalin fixative for histological analysis, flash-frozen in liquid nitrogen for processing for protein analysis, or directly homogenized in 3.0 ml ZR RNA buffer (Zymo Research), and stored at −80°C.

Mice heterozygous for the disruption of Il1rn (strain B6.129S-Il1rn^tm1Dih/J; stock number 004754) and hemizygous transgenic mice that overexpress and secrete mouse IL-1Ra [strain B6.Cg-Tg(Il1rn)1Dih/J; stock number 004753] were purchased from The Jackson Laboratory. Mice were bred in-house at the Rocky Mountain Laboratories to obtain mice homozygous for the disruption of Il1rn (IL1rn^−/−), hemizygous transgenic Il1rn overexpressors (TgIL1rn), and normal C57BL/6 littermate controls. IL1rn^−/−, TgIL1rn, and C57BL/6 littermate controls were inoculated with 30 μl in the left hemisphere by using a 1.0% 22L, RML, or normal brain homogenate stock. Mice were euthanized when they displayed persistent clinical signs, described above.

Preparation of brain homogenates for protein analysis from mice.

Brains were homogenized (20%, wt/vol) by using a Mini Bead Beater (BioSpec Products) as previously described (21) in ice-cold cell lysis buffer (Bio-Rad) supplemented with 2× Complete EDTA-free protease inhibitor cocktail (Roche) and 1× cell lysis factors 1 and 2 (Bio-Rad), consisting of sodium orthovanadate and sodium fluoride to prevent dephosphorylation of proteins. Homogenates were stored in 200-μl aliquots at −80°C until use. Protein concentrations were estimated by using the modified Lowry procedure (33), using bovine serum albumin (BSA) as a standard.

RNA isolation.

Mouse brain halves were homogenized in 3.0 ml ZR RNA buffer (Zymo Research) and stored for up to 5 days at −80°C before processing. Total RNA was isolated by using the Quick-RNA MidiPrep system (Zymo Research) and treated with 4 U of DNase I (Ambion) for 1 h at 37°C. High-quality RNA was purified using the RNA Clean & Concentrator-100 system (Zymo Research) per the manufacturer's instructions and eluted from the columns in 300 to 400 μl RNase-free water.

Cytokine and chemokine quantification.

Levels of IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12 p40, IL-12 p70, IL-13, IL-17A, CCL11 (Eotaxin), granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage CSF (GM-CSF), gamma interferon (IFN-γ), CXCL1 (KC), CCL2 (monocyte chemoattractant protein 1 [MCP-1]), CCL3 (macrophage inflammatory protein 1α [MIP-1α]), CCL4 (MIP-1β), CCL5 (RANTES), and TNF-α in brain tissue homogenates were analyzed using the Bio-Plex suspension array system as previously reported (21), with the exception that 50 μl of 10% brain homogenate was used in the assay mixtures and brains were homogenized in cell lysis buffer as described above. Suspension array reagents in brain tissue were purchased from Bio-Rad. Suspension array reagents to quantify CXCL9 (MIG) and CXCL10 (IP-10) release in brain tissues were purchased from Invitrogen. For all enzyme-linked immunosorbent assays (ELISAs), 100 μl of 10% brain homogenate was used in the assay mixture. Quantikine ELISAs to quantify IL-1Ra, CXCL13 (BLC), and CCL22 were purchased from R&D Systems. The Oncostatin-M (OSM) ELISA was purchased from American Research Products, Inc. ELISA plates were read using a SpectraMAX 190 microplate spectrophotometer with SoftMax Pro v5 software (Molecular Devices). Graphing and statistical analysis were performed using the unpaired Student's t test (GraphPad Prism 6) to compare mock versus infected mice at each time point, with a P value of ≤0.05 considered significant.

qRT-PCR analysis.

For quantitative analysis of changes in transcription using qRT-PCR arrays, 400 ng of high-quality RNA from each sample was reversed transcribed to synthesize cDNA by using the RT² First Stand kit (Qiagen) per the manufacturer's instructions. Each cDNA reaction mixture was combined with 2× RT2 SYBR green master mix purchased from Qiagen with RNase-free water to a final volume of 1.3 ml. Ten microliters was then added to the appropriate wells of a 384-well format plate from the following 4 pathway-focused qRT-PCR arrays from Qiagen: mouse signal transduction pathway finder PAMM-014ZE (STPF), mouse JAK-STAT signaling pathway PAMM-039ZE, mouse IL-6/STAT3 signaling pathway PAMM-160ZE, and mouse Inflammatory Cytokine and Receptors PAMM-011ZE (ICR). The analysis was carried out on an Applied Biosystems ViiA 7 real-time PCR system with a 384-well block under the following conditions: 1 cycle for 10 min at 95°C, 40 cycles of 15 s at 95°C, then 1 min at 60°C with fluorescence data collection. Melting curves were generated at the end of the completed run to determine the quality of the reaction products. Raw threshold cycle (C_T) data were collected, with a C_T of 35 as the cutoff. C_T data were analyzed using the Web-based RT² Profiler PCR array data analysis program (Sabosciences). All C_T values were normalized to the average of the C_T values for the housekeeping genes Actb, Gapdh, and Hsp90ab1. Changes in transcription were calculated via the software using the ΔΔC_T-based method (34). Statistical analysis was performed using the unpaired Student's t test of the replicate 2^−ΔΔCT values for each gene in the control group and treatment groups, with P values of ≤0.05 considered significant. Each treatment and control group consisted of a minimum of 3 independent RNA samples.

To determine changes in Il12b transcription during disease, qRT-PCR was performed using the iScript one-step RT-PCR system with SYBR green (Bio-Rad). Changes in fluorescence were monitored by using the MyiQ single-color real-time PCR detection system (Bio-Rad). Twenty-five-microliter reaction mixtures were prepared in triplicate in a 96-well format, and reaction mixtures contained 264 nM of each Il12b primer (Il12b.1-625f, AGCAGTAGCAGTTCCCCTGA; Il12b.1-712r, AGTCCCTTTGGTCCAGTGTG) or Gapdh primer (Gapdh1-506f, TGCACCACCACCTGCTTAGC; Gapdh1-683r, TGGATGCAGGGATGATGTTC). RNA was then added to a final amount of 50 μg per reaction mixture. qRT-PCR conditions were as follows: 1 cycle at 50°C for 30 min, 1 cycle at 95°C for 10 min, and 40 cycles of 95°C for 15 s and 55°C for 60 s with data collection. Melting curves were generated by a cycle of 95°C for 1 min, 55°C for 1 min, and 80 cycles of 50°C for 10 s with 0.5°C increments. The results were calculated using the ΔΔC_T method described above, where relative amounts of RNA were normalized to GAPDH amounts and statistical analyses were performed using the unpaired Student's t test of the replicate 2^−ΔΔCT values.

Immunoblotting for PrPres.

For PrPres immunoblotting, tissue samples were analyzed as described previously (3, 35). Briefly, 0.36 mg of whole-brain equivalent was treated with proteinase K, separated by SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and probed with a 1:100 dilution of monoclonal human anti-PrP antibody D13 derived from cell culture supernatants produced in our laboratory from CHO cells expressing the D13 antibody construct (36). The secondary antibody was peroxidase-conjugated anti-human IgG, used at 1:10,000 (Sigma), and immunoreactive bands were visualized using an enhanced chemiluminescence (ECL) detection system (GE Healthcare).

Detection of STAT proteins.

Fifty micrograms of brain homogenate was solubilized in Laemmli sample buffer (Bio-Rad) with 5% β-mercaptoethanol (Sigma) and separated by SDS-PAGE using 10% Novex Tris-glycine prepoured acrylamide gels in 1× Novex Tris-glycine SDS running buffer in a Novex minicell (Invitrogen) (90 V, 23°C, 2 h). Separated proteins were transferred to PVDF membranes (Bio-Rad) (15 V, 4°C, 16 h) in a Bio-Rad Trans-Blot cell with plate electrodes by using Towbin buffer (0.192 M glycine, 0.025 M Tris base, 0.0013 M SDS, 5% methanol; pH 8.3) (37). Transferred proteins were visualized on membranes by staining for 1 min in amido blue black stain (0.1% amido blue black dye in 1.0% acetic acid), and standards were marked. Broad-range molecular weight standards were purchased from Bio-Rad Laboratories.

Immunoblots were blocked (2 to 3 h at 23°C) in 4.0% (wt/vol) BSA in Tris-buffered saline (Bio-Rad) supplemented with 0.1% Tween 20 (TBS-T₂₀). Blocked immunoblots were probed with a 1:1,000 dilution of anti-STAT1, anti-pSTAT1(Tyr701), anti-pSTAT1(Ser727), anti-STAT3, anti-pSTAT3(Tyr705), and anti-pSTAT3(Ser727) in 4.0% BSA in TBS-T₂₀ (overnight, 4°C, with gentle rocking). Blots were rinsed 3 times in TBS-T₂₀ (50 ml; 15 min each) and probed with a 1:3,000 dilution of horseradish peroxidase (HRP)-linked anti-rabbit IgG in 4.0% BSA in TBS-T₂₀ (45 min, 23°C, with gentle rocking). After probing with the HRP-linked secondary antibody, immunoblots were rinsed 3 times in TBS-T₂₀ (50 ml; 15 min each), and reactive bands were visualized with an ECL kit (Amersham) in accordance with the manufacturer's specifications. All antibodies used in immunoblotting for STAT proteins were purchased from Cell Signaling Technology.

After STAT immunoblot analysis, immunoblots were stripped in 2 M glycine, pH 2.3 (1 h, 23°C, with shaking), rinsed 3 times in TBS-T₂₀ (50 ml; 15 min each), and probed with a 1:2,000 dilution of anti-β-actin (Cell Signaling Technology) in 4.0% BSA in TBS-T₂₀ (overnight, 4°C, with gentle rocking). Anti-β-actin-probed immunoblots were rinsed 3 times in TBS-T₂₀ (50 ml; 15 min each), probed with a 1:3,000 dilution of HRP-linked anti-rabbit IgG in 4.0% BSA in TBS-T₂₀ (45 min, 23°C, with gentle rocking), and visualized using ECL as described above.

Densitometry on reactive bands was performed using the Bio-Rad ChemiDoc MP system. Reactive band density in a sample was normalized to the β-actin for that sample and density values were compared between mock and 22L infected mouse groups at a given time. Statistical analysis was performed using the unpaired Student's t test (GraphPad Prism 6) comparing mock to infected mice, with a P value of ≤0.05 considered significant.

Immunohistochemistry.

Mice were euthanized, brains were removed, and half of the brain was placed in 3.7% phosphate-buffered formalin for 3 to 5 days before dehydration and embedding in paraffin. Serial 5-μm sections were cut using a standard Leica microtome, placed on positively charged glass slides, and dried overnight at 56°C. Slides were stained with a standard protocol of hematoxylin and eosin (H&E) for observation of overall pathology. For the detection of microglia, sections were probed with antibodies at 1:2,000 against ionized calcium-binding adapter-1 (IBA-1), and for astrocytes, sections were probed with antibodies at 1:3,500 against glial fibrillary acidic protein (GFAP). Antigen retrieval, primary and secondary antibody probings, and detection were carried out as previously described (36).

RESULTS

Analysis of key genes mediating the inflammatory response in brains of mice infected with scrapie strain 22L at various days postinfection.

We and others demonstrated previously that expression or release of a limited subset of cytokines and chemokines can be detected in the brains of prion-infected mice (3, 11, 20, 21, 38–41). To enhance detection and comparison of inflammatory response genes expressed at lower levels in the brains of mice in the present study, we analyzed RNA from scrapie-infected mice at different time points by utilizing an Inflammatory Cytokines and Receptors (ICR) array. This array is composed of 84 primer sets that target expression of 25 chemokines and their receptors, 17 interleukins and their receptors, and 22 additional cytokines important in inflammation. We also analyzed the expression of Il12b, encoding IL-12p40, that was not included on the ICR array, as we previously found IL-12p40 to be elevated in scrapie-positive brains (3, 21).

At 44 days p.i., only four genes were significantly increased when RNA from brains of scrapie-infected and control mice were compared using the ICR array (Table 1). These were Ccl4, Cxcl1, Cxcl10, and Il1b. All four demonstrated just slightly over a 2-fold increase in transcription. At the 70-day p.i. time point, 18 of the 85 genes analyzed were significantly increased, and at 94 days p.i. 28 genes demonstrated a significant increase in expression (Table 1). Postclinical samples at 131 days p.i. displayed the largest number of genes (38 of 85) that were increased in the brain relative to controls. Interestingly, we identified 10 genes encoding inflammatory mediators that had not been previously described to be increased during scrapie infection. The majority of these genes were upregulated in brain well before clinical signs (i.e., at 70 or 94 days p.i.). These included Il1rn (encoding IL-1-Ra), Ccl8, Cxcl5, Tnfsf11 (encoding the RANKL receptor activator of the NF-κB ligand), and Osm (encoding oncostatin-M, or OSM) that were all increased in expression, some as early as 70 days p.i., with levels ranging from 2.9- to 58-fold above uninfected brains at 131 days p.i. Overall, in our kinetic analysis of gene expression during scrapie infection with strain 22L, transcription of many inflammatory mediators progressively increased during the course of the disease, with peak values found at the clinical endpoint, but there were many inflammatory genes that were unchanged (see Table S1 in the supplemental material).

TABLE 1.

Genes increased in strain 22L scrapie-infected C57BL/10 mouse brains, based on the Inflammatory Cytokines and Receptors array assay results

New or previously observed gene	Results at:								Description
	44 days p.i.		70 days p.i.		94 days p.i.		131 days p.i.
	FCa	P valueb	FC	P value	FC	P value	FC	P value
Newly identified genes
Il1rnd	—c	—	8.0	8 × 10−6	16.7	2 × 10−4	58.2	2 × 10−6	IL-1RA
Ccl8	—	—	6.5	2 × 10−5	13.5	2 × 10−5	41.3	1 × 10−6	Chemokine (C-C motif) ligand 8
Tnfsf11	—	—	3.9	1 × 10−2	3.7	2 × 10−2	32.7	<1 × 10−6	TNF (ligand) superfamily, member 11
Osm	—	—	2.6	NS	5.8	6 × 10−4	6.8	3 × 10−4	Oncostatin M
Il27d	—	—	—	—	3.7	2 × 10−2	6.5	1 × 10−3	Interleukin-27
Cxcl5d	—	—	—	—	3.0	1 × 10−3	2.9	2 × 10−6	Chemokine (C-X-C motif) ligand 5
Ccl22d	—	—	—	—	2.3	NS	3.0	1 × 10−2	Chemokine (C-C motif) ligand 22
Tnfsf10d	—	—	—	—	—	—	2.4	3 × 10−4	TNF (ligand) superfamily, member 10
Il6ra	—	—	—	—	—	—	2.1	2 × 10−3	IL-6Rα
Tnfsf13bd	—	—	—	—	—	—	2.0	3 × 10−4	TNF (ligand) superfamily, member 13b
Previously observed genes
Cxcl10d	2.4	1 × 10−3	19.2	<1 × 10−6	50.9	<1 × 10−6	96.6	<1 × 10−6	Chemokine (C-X-C motif) ligand 10
Ccl4d	2.1	1 × 10−3	8.4	<1 × 10−6	15.2	<1 × 10−6	46.2	<1 × 10−6	Chemokine (C-C motif) ligand 4
Cxcl13	—	—	8.0	<1 × 10−6	20.8	1 × 10−3	207.5	<1 × 10−6	Chemokine (C-X-C motif) ligand 13
Ccl5d	—	—	6.5	<1 × 10−6	10.2	1 × 10−6	36.0	<1 × 10−6	Chemokine (C-C motif) ligand 5
Ccl2d	—	—	6.4	1 × 10−5	12.8	<1 × 10−6	48.4	<1 × 10−6	Chemokine (C-C motif) ligand 2
Ccl12	—	—	5.7	<1 × 10−6	10.8	<1 × 10−6	47.3	<1 × 10−6	Chemokine (C-C motif) ligand 12
Cxcl9d	—	—	5.5	2 × 10−2	24.1	2 × 10−5	69.4	1. × 10−6	Chemokine (C-X-C motif) ligand 9
Tnfd	—	—	5.4	6 × 10−3	6.7	2 × 10−2	33.4	<1 × 10−6	TNF
Ccl3d	—	—	5.2	<1 × 10−6	9.2	<1 × 10−6	28.7	1 × 10−6	Chemokine (C-C motif) ligand 3
Ccl9	—	—	5.1	3 × 10−4	7.7	<1 × 10−6	17.8	<1 × 10−6	Chemokine (C-C motif) ligand 9
Il1bd	2.2	2 × 10−6	2.9	3 × 10−3	3.2	2 × 10−4	8.8	1 × 10−6	IL-1β
Ccl7	—	—	2.9	1 × 10−4	4.0	3 × 10−5	10.2	<1 × 10−6	Chemokine (C-C motif) ligand 7
Ccl6	—	—	2.7	3 × 10−5	4.1	<1 × 10−6	11.0	<1 × 10−6	Chemokine (C-C motif) ligand 6
Il12bd	—	—	2.4	8 × 10−3	5.0	5 × 10−3	12.1	1 × 10−4	IL-12B(p40)
Il2rg	—	—	—	—	3.4	<1 × 10−6	9.2	5 × 10−6	IL-2R, gamma chain
Ccr1	—	—	—	—	3.1	5 × 10−5	5.9	2. × 10−5	Chemokine (C-C motif) receptor 1
Cxcl1d	2.7	4 × 10−2	—	—	3.0	NS	10.3	1. × 10−6	Chemokine (C-X-C motif) ligand 1
Il1a	—	—	—	—	2.3	5 × 10−5	5.7	<1 × 10−6	IL-1α
Csf1d	—	—	—	—	2.2	3 × 10−4	2.3	3 × 10−4	Colony-stimulating factor 1 (macrophage)
Ccr3	—	—	—	—	2.1	3 × 10−4	3.8	1 × 10−6	Chemokine (C-C motif) receptor 3
Spp1d	—	—	—	—	2.1	1 × 10−2	6.3	3 × 10−6	Secreted phosphoprotein 1
Cxcl11d	—	—	—	—	—	—	6.0	2 × 10−3	Chemokine (C-X-C motif) ligand 11
Cxcr3	—	—	—	—	—	—	3.1	2 × 10−2	Chemokine (C-X-C motif) receptor 3
Il10ra	—	—	—	—	—	—	2.6	5 × 10−6	IL-10Rα
Ccr5d	—	—	—	—	—	—	2.9	<1 × 10−6	Chemokine (C-C motif) receptor 5
Il1r1	—	—	—	—	—	—	2.5	9 × 10−6	IL-1R, type I
Ccl1d	—	—	—	—	—	—	2.3	4 × 10−2	Chemokine (C-C motif) ligand 1
Il10rb	—	—	—	—	—	—	2.1	2 × 10−6	IL-10Rβ

^aFC, fold change of 22L-infected mice (n = 3) relative to mock-infected controls (n = 4).

^bP values were calculated by using a two-tailed t test. NS, not significant.

^c—, FC and/or P value did not meet our criteria (≥2-fold change, P ≤ 5 × 10⁻²).

^dGenes that is a known target for NF-κB.

We next determined the protein concentration of several newly identified and previously observed cytokines/chemokines from homogenates from 22L-infected brains at various times postinoculation (Fig. 1). Similar to our earlier published findings, IL-12p40 and CCL5 (RANTES) increased in a time-dependent manner in scrapie-infected brain homogenates relative to age-matched controls (3, 21). Five additional proteins that have been previously identified as increased during the clinical phase of scrapie infection (CCL4 [Mip-1β], CXCL9, CXCL10 [IP-10], CXCL13, and IL-1α) were also detected in greater amounts in brains of scrapie-infected mice as early as 94 days p.i. One chemokine, CXCL10 (IP-10), was significantly increased in brain tissues at 70 days p.i. as well. Production of three of the newly identified genes (CCL22 [MDC], IL-1Ra [encoded by Il1rn], and OSM) also increased in a time-dependent manner. The release of soluble TNF-α was not significantly elevated at any time point, even though Tnfa transcript levels at 70, 94, and 131 days p.i. were significantly increased (Table 1). This was also reported previously by our lab from samples tested at the clinical endpoint (21). Though for technical reasons we could not quantify protein amounts for all differentially expressed genes listed in Table 1, for those that were analyzed (Fig. 1), most had a strong correlation between increased gene expression and increased protein production at both 94 and 131 days p.i.

FIG 1.

Detection of cytokines in brain tissue of C57BL/10 mice at various times after infection with scrapie strain 22L. Protein levels of 11 cytokines in the brains of scrapie-infected mice were measured by multiplex assay (IL-12p40, CCL4, CCL5, IL-1a, CXCL9, CXCL10, and soluble TNF-α) or by quantitative ELISA (CCL22, CXCL13, IL-1Ra, and OSM) at 70, 94, and 131 days p.i., as indicated. Each data point represents the result from an individual mouse, and the bar in each data set represents the mean. Black circles with gray bars indicate 22L scrapie-infected groups, and open circles with white bars represent mock-infected (M) groups. dpi, days postinfection. The error bars indicate 1 standard deviation from the mean. Statistical analysis was performed using the Student t test to compare infected and mock-infected results at a given time point. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001.

JAK-STAT signal transduction pathway analysis of 22L scrapie-infected mouse brain at various days postinfection.

Many cytokines and chemokines released prior to clinical signs during scrapie infection (i.e., CCL2, CCL3, CCL4, CCL5, and IL-12) are known to signal through the JAK-STAT signal transduction pathway (25, 26, 42). A preliminary analysis using the Signal Transduction Pathway Finder (STPF) array, which targets genes associated with 10 commonly studied signal transduction pathways (TGF-β, WNT, NF-κB, JAK-STAT, p53, Notch, Hedgehog, PPAR [peroxisome proliferator-activated receptors], oxidative stress, and hypoxia) suggested that only the STAT and NF-κB pathways were substantially activated in the brain during scrapie infection with strains 22L or RML (Table 2; see also Table S4 in the supplemental material for unchanged genes). The STPF array also suggested that similar signal transduction pathways were triggered in the brain when mice were infected with either 22L or RML scrapie strains. It has been shown that pSTAT1 and pSTAT3 are increased when mice are infected with scrapie strain ME7 (27). Similar to these findings, we identified an increase in total STAT1α, as well as an increase in pSTAT1α and pSTAT3, in our 22L scrapie model (Fig. 2A and B).

TABLE 2.

Comparison of genes altered in C57BL/6 mouse brain infected with scrapie strains 22L and RML assayed using the Signal Transduction Pathway Finder array

Gene group and name	22L (at 131 days p.i.)		RML (at 151 days p.i.)		Description
	FCa	P valueb	FC	P value
NF-κB associated
Ccl5	26.6	4.0 × 10−4	11.0	8.5 × 10−4	Chemokine (C-C motif) ligand 5
Tnf	20.0	1.6 × 10−5	18.7	3.7 × 10−3	TNF
Olr1d	9.8	1.0 × 10−2	18.6	5.9 × 10−4	Oxidized low-density lipoprotein (lectin-like) receptor 1
Bcl2a1a	7.3	5.2 × 10−3	4.5	3.8 × 10−5	B-cell leukemia/lymphoma 2-related protein A1a
Fasd	5.5	9.2 × 10−4	3.9	1.0 × 10−4	Fas (TNF receptor superfamily member 6)
Icam1	4.0	1.2 × 10−4	3.0	3.4 × 10−5	Intercellular adhesion molecule 1
Hmox1d	3.9	1.6 × 10−3	4.3	1.0 × 10−5	Heme oxygenase (decycling) 1
Birc3	3.0	1.2 × 10−3	2.0	7.3 × 10−4	Baculoviral IAP repeat-containing 3
Serpine1d	2.4	3.2 × 10−2	—c	—	Serine (or cysteine) peptidase inhibitor, clade E, member 1
Cdkn1ad	2.1	3.8 × 10−3	—	—	Cyclin-dependent kinase inhibitor 1A (P21)
Csf1	2.0	9.6 × 10−3	2.3	5.5 × 10−4	Colony-stimulating factor 1 (macrophage)
JAK/STAT associated
Cebpde	6.7	2.3 × 10−5	5.0	1.0 × 10−6	CCAAT/enhancer-binding protein (C/EBP), delta
Socs3	3.7	2.0 × 10−6	5.4	6.0 × 10−6	Suppressor of cytokine signaling 3
Stat1e	3.1	3.9 × 10−4	2.0	4.6 × 10−5	Signal transducer and activator of transcription 1
WNT associated
Fosl1	3.8	4.2 × 10−2	3.3	3.4 × 10−2	Fos-like antigen 1
TGF-β associated
Emp1	3.6	3.8 × 10−5	2.6	4.2 × 10−4	Epithelial membrane protein 1
Notch associated
Hes5	−2.6	4.0 × 10−3	—	—	Hairy and enhancer of split 5 (Drosophila melanogaster)
Heyl	−2.6	1.1 × 10−3	—	—	Hairy/enhancer-of-split related with YRPW motif-like
Hey2	—	—	2.3	2.7 × 10−4	Hairy/enhancer-of-split related with YRPW motif 2
p53 associated
Bbc3e	—	—	2.6	3.5 × 10−4	BCL2-binding component 3

^aFC, fold change of 22L-infected mice (n = 3) or RML-infected mice (n = 4) relative to mock-infected controls (n = 4).

^bBased on a two-tailed t test.

^c—, the FC and/or P value did not meet our criteria (≥2-fold change, P value ≤ 5.0 × 10⁻²).

^dOlr1 is also PPAR associated, Hmox1 is also oxidative stress and hypoxia associated, Serpine1 is also hypoxia associated, and Cdkn1a and Fas are also p53 associated.

^eGene that is a known target for NF-κB.

FIG 2.

Immunoblot analysis of mouse brain homogenates demonstrated increased STAT1, pSTAT1, and pSTAT3 in the brains at various times after scrapie infection. Proteins (50 μg/lane) from brain homogenates of mock-infected and 22L-infected mice were separated by SDS-PAGE, transferred to a PVDF membrane, and probed with the indicated anti-STAT1, anti-pSTAT1, or anti-actin antibody (A) or probed with the indicated anti-STAT3, anti-pSTAT3, or anti-actin antibody as a control (B). The immunoblots are for representative mice at each time point, and the molecular masses of (p)STAT1α, (p)STAT1β, and (p)STAT3 are indicated to the left of each immunoblot (in kDa). In panel A, the fold change measured by densitometry of detectable STAT1α (middle) and pSTAT1α(S727) (lower), normalized to actin and relative to mock controls, is shown in a bar graph, where each data point represents the result for an individual mouse and is plotted on a log₂ scale. In panel B, the fold change measured by densitometry of detectable STAT3 (middle) and pSTAT3(Y705) (lower), normalized to actin and relative to mock controls, is shown as a bar graph, where each data point represents the result for an individual mouse and is plotted on a log₂ scale. Statistical analysis was performed using the Student t test to compare infected to mock-infected controls at a given time point. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001.

To better characterize the potential influences of the JAK-STAT pathway during scrapie infection, we analyzed RNA isolated from brains of 22L scrapie-infected mice at various times postinfection by using two qRT-PCR arrays to analyze numerous aspects of the JAK-STAT pathway, as well as the more specific IL-6/STAT3 signaling pathway. Of the individual 150 genes on the combined STAT signaling arrays, a subset of 54 genes was increased at one or more time points assessed (Table 3). Five previously unreported genes associated with the STAT pathway were upregulated in 22L-infected mice as early as 94 days p.i. Fifteen genes were detectable at 70 days p.i. in the brains of scrapie-infected mice, where several genes overlapped with those represented on the ICR array in Table 1 (Table 3, see footnote), and nine of these (Ccl5, Ccl12, Cxcl9, Cxcl10, Oas1a, Isg15, Osm, A2m, and Socs3) were genes influenced or regulated by phosphorylation of specific STAT protein complexes. Several genes encoding cytokines known to predominantly activate the STAT3 pathway (i.e., Ccl4, Ccl3, Tnf, Osm, and Tnfsf11) were significantly increased at 70 or 94 days p.i. (Table 3).

TABLE 3.

Kinetic analysis of genes altered in C57BL/10 mouse brain infected with scrapie strain 22L assayed with the JAK/STAT and IL-6/STAT3 arrays^a

New or previously observed gene	At 70 days p.i.		At 94 days p.i.		At 131 days p.i.		Description
	FCb	P valuec	FC	P value	FC	P value
Newly identified genes
Tlr4	—d	—	7.2	2 × 10−2	11.4	3 × 10−3	Toll-like receptor 4
Gbp1e	—	—	4.2	1 × 10−6	11.5	<1 × 10−6	Guanylate-binding protein 1
Tnfrsf10b	—	—	—	—	2.7	2 × 10−3	TNFR superfamily, member 10b
Il4ra	—	—	—	—	2.1	2 × 10−4	IL-4Rα
Cd80e	—	—	—	—	2.0	5 × 10−2	CD80 antigen
Previously observed genes
Oas1a	3.1	9 × 10−4	8.8	6 × 10−6	24.6	<1 × 10−6	2′-5′ oligoadenylate synthetase 1A
Isg15	3.0	3 × 10−5	6.6	4 × 10−6	22.9	<1 × 10−6	ISG15 ubiquitin-like modifier
Osmr	2.8	9 × 10−4	7.0	1 × 10−4	18.2	1 × 10−3	Oncostatin M receptor
A2 m	2.1	1 × 10−3	5.7	5 × 10−3	19.5	<1 × 10−6	α2-Macroglobulin
Socs3	2.0	1 × 10−3	2.6	6 × 10−6	6.6	3 × 10−4	Suppressor of cytokine signaling 3
Tnfrsf1a	2.0	7 × 10−3	3.1	1 × 10−4	6.1	1 × 10−6	TNFR superfamily, member 1a
Fcgr1e	—	—	3.6	1 × 10−6	7.9	<1 × 10−6	Fc receptor, IgG, high affinity I
Irf9	—	—	3.5	<1 × 10−6	6.7	<1 × 10−6	Interferon regulatory factor 9
Cebpde	—	—	3.2	4 × 10−5	8.1	6 × 10−6	CCAAT/enhancer-binding protein, delta
Ptprc	—	—	2.9	2 × 10−6	6.1	6 × 10−5	Protein tyrosine phosphatase, receptor type C
Stat1e	—	—	2.7	<1 × 10−6	5.0	<1 × 10−6	Signal transducer and activator of transcription 1
Lif	—	—	2.4	3 × 10−3	4.4	1 × 10−2	Leukemia inhibitory factor
Cxcr4	—	—	2.3	2 × 10−2	2.3	3 × 10−3	Chemokine (C-X-C motif) receptor 4
Fas	—	—	2.2	1 × 10−5	5.2	1 × 10−6	Fas (TNFR superfamily member 6)
Sfpi1	—	—	2.2	3 × 10−6	4.5	<1 × 10−6	SFFV proviral integration 1
Csf3r	—	—	2.0	1 × 10−3	5.8	<1 × 10−6	Colony-stimulating factor 3 receptor (granulocyte)
Tnfrsf1be	—	—	2.0	5 × 10−4	3.3	8 × 10−5	TNFR superfamily, member 1b
Nos2e	—	—	—	—	2.9	2 × 10−3	Nitric oxide synthase 2, inducible
Stat3e	—	—	—	—	2.9	1 × 10−6	Signal transducer and activator of transcription 3
Irf1e	—	—	—	—	2.8	4 × 10−5	Interferon regulatory factor 1
Csf1r	—	—	—	—	2.8	<1 × 10−6	Colony-stimulating factor 1 receptor
Myce	—	—	—	—	2.6	7 × 10−6	Myelocytomatosis oncogene
Il1r1	—	—	—	—	2.6	5 × 10−4	IL-1R, type I
Cd40e	—	—	—	—	2.5	6 × 10−3	CD40 antigen
Ccnd1	—	—	—	—	2.3	6 × 10−6	Cyclin D1
Jak3	—	—	—	—	3.0	5 × 10−4	Janus kinase 3
Socs1	—	—	—	—	2.3	3 × 10−3	Suppressor of cytokine signaling 1
Stat2	—	—	—	—	2.2	2 × 10−4	Signal transducer and activator of transcription 2
Hgfe	—	—	—	—	2.0	4 × 10−3	Hepatocyte growth factor
Ifngr1	—	—	—	—	2.0	2 × 10−5	Interferon gamma receptor 1
Nfkbiae	—	—	—	—	2.0	1 × 10−3	NF-κB light polypeptide gene enhancer in B cells inhibitor, alpha

^aValues for Ccl3, Ccl4, Ccl5, Ccl12, Csf1, Cxcl9, Cxcl10, Il1a, Il1b, Il2rg, Il6ra, Il10ra, Il10rb, Osm, Tnf, Tnfsf10, and Tnfsf11, which were all increased in these arrays, have been omitted from this table; the values are presented in Table 1.

^bFC, fold change of 22L-infected mice (n = 3) relative to mock-infected controls (n = 4).

^cCalculated by using a two-tailed t test.

^d—, FC and/or P value did not meet our criteria (≥2-fold change, P value ≤ 5 × 10⁻²).

^eA gene that is a known target for NF-κB.

Additional genes upregulated by 70 or 94 days p.i. were receptors that signal through the JAK-STAT pathway (Il2rg, Fcgr1, Ptprc, and Il10ra). Furthermore, several receptors that specifically activate STAT3 (i.e., Osmr, Csf3r, Tnfrsf1a, Fas, and Cxcr4) and many pSTAT3 target genes (i.e., Cxcl10, Ccl5, Ccl12, Il1b, and Cebpd) were also increased in infected brains between 70 and 94 days p.i. We also detected increased expression of several STAT pathway-associated transcription factors (Irf9, Cebpd, Stat1, and Sfpi1) by 94 days p.i., with transcripts of Stat1, Stat2, and Stat3 significantly increased in the brain by 131 days p.i. Only one of the four Janus kinase transcripts (Jak3) was significantly increased during the course of scrapie disease. Overall, results with RNA from preclinical (70 and 94 days p.i.) and clinical (131 days p.i.) mice suggested that multiple genes related to the occurrence of pSTAT complexes were upregulated during the course of scrapie disease.

In contrast, several genes involved in the JAK-STAT pathway and encoding canonical STAT3 activators and/or pSTAT3 targets (i.e., Cxcl12, Il10, Il11, Il12a, Il17a, Il21, Il23a, or Il6) were not increased during scrapie infection (see Tables S2 and S3 in the supplemental material). Taken together, the results suggested that expression of a select group of STAT-associated genes was altered during the neuroinflammation observed in scrapie infection, and many of these genes are increased in the brain well before clinical signs are observed.

Comparison of key genes mediating the inflammatory response at the clinical endpoints for mice infected with strain 22L or RML.

Scrapie strains can differ in clinical manifestations, regional distribution of vacuolation, incubation time, and host response (6–8). To study whether dissimilarities in neuroinflammation might account for clinical differences, we compared alterations in transcripts from mice infected with strains 22L and RML at the clinical endpoints of 131 days p.i. and 150 days p.i. in mice, respectively (Table 4). Mice infected with scrapie strain RML shared 28 out of the 34 (82%) upregulated genes seen when mice were infected with strain 22L. Seven previously unreported genes (Ccl8, Cxcl5, Il1rn, Il6ra, Il27, Osm, and Tnfsf11) were increased in the brains of mice during scrapie infection and were in agreement with our kinetic analysis of 22L-infected mice. Five of the newly identified genes were significantly increased in mice infected with 22L and RML strains, but Cxcl5 was increased in 22L-infected brains and not RML-infected mice. Il27 was increased in both infected cohorts but was only statistically significant in 22L-infected brains.

TABLE 4.

Comparison of genes altered in 22L- and RML-infected C57BL/6 mouse brain assayed with the Inflammatory Cytokines and Receptors array

New or previously observed gene	22L (131 days p.i)		RML (150 days p.i.)		Description
	FCa	P valueb	FC	P value
Newly identified genes
Tnfsf11	30.6	2 × 10−2	12.1	4 × 10−3	TNF (ligand) superfamily, member 11
Il1rnd	27.9	1 × 10−2	36.8	3 × 10−6	IL-1Ra
Ccl8	15.5	1 × 10−5	6.9	3 × 10−3	Chemokine (C-C motif) ligand 8
Osm	7.0	4 × 10−3	9.2	1 × 10−3	Oncostatin M
Il27	4.5	3 × 10−2	3.6	NS	IL-27
Cxcl5d	3.8	4 × 10−3	—c	—	Chemokine (C-X-C motif) ligand 5
Il6ra	2.5	3 × 10−3	2.9	4 × 10−4	IL-6Ra
Previously observed genes
Cxcl13	54.2	3 × 10−2	18.2	3 × 10−2	Chemokine (C-X-C motif) ligand 13
Tnf	50.4	6 × 10−4	52.3	4 × 10−4	TNF
Ccl4d	46.0	1 × 10−5	36.5	6 × 10−5	Chemokine (C-C motif) ligand 4
Ccl3d	38.2	8 × 10−6	33.7	8 × 10−6	Chemokine (C-C motif) ligand 3
Cxcl10d	38.2	2 × 10−5	15.7	2 × 10−5	Chemokine (C-X-C motif) ligand 10
Cxcl9d	23.5	5 × 10−3	12.0	6 × 10−4	Chemokine (C-X-C motif) ligand 9
Ccl12	18.0	5 × 10−6	4.7	6 × 10−3	Chemokine (C-C motif) ligand 12
Ccl2d	15.8	2 × 10−4	6.6	2 × 10−3	Chemokine (C-C motif) ligand 2
Ccl9	15.2	2 × 10−3	10.8	1 × 10−4	Chemokine (C-C motif) ligand 9
Ccl5d	15.1	2 × 10−4	6.1	9 × 10−5	Chemokine (C-C motif) ligand 5
Ccl6	11.2	4 × 10−6	7.2	7 × 10−4	Chemokine (C-C motif) ligand 6
Il1a	5.8	3 × 10−4	4.6	2 × 10−4	IL-1α
Ccl7	5.6	6 × 10−5	2.6	2 × 10−3	Chemokine (C-C motif) ligand 7
Il2rg	5.2	6 × 10−4	4.1	2 × 10−5	IL-2R, gamma chain
Il1bd	4.7	1 × 10−3	4.6	8 × 10−4	IL-1β
Cxcl11d	4.5	3 × 10−2	3.9	NS	Chemokine (C-X-C motif) ligand 11
Ccr1	4.2	8 × 10−5	—	—	Chemokine (C-C motif) receptor 1
Ccr3	3.4	4 × 10−6	—	—	Chemokine (C-C motif) receptor 3
Il10ra	3.2	1 × 10−5	2.7	8 × 10−6	IL-10Rα
Ccr5d	3.1	9 × 10−5	—	—	Chemokine (C-C motif) receptor 5
Il1r1	3.0	1 × 10−5	2.1	2 × 10−3	IL-1R, type I
Il21	2.7	3 × 10−2	2.4	NS	IL-21
Spp1d	2.6	1 × 10−3	2.5	2 × 10−3	Secreted phosphoprotein 1
Csf1d	2.6	4 × 10−5	3.0	2 × 10−4	Colony-stimulating factor 1 (macrophage)
Cxcr5	2.3	3 × 10−2	2.8	3 × 10−2	Chemokine (C-X-C motif) receptor 5
Il2rb	2.2	2 × 10−2	—	—	IL-2R, beta chain
Il10rb	2.2	2 × 10−5	—	—	IL-10R, beta chain
Cd40lgd	—	—	2.1	1 × 10−2	CD40 ligand

^aFC, fold change of 22L-infected mice (n = 4) or RML-infected mice (n = 4) relative to mock-infected controls (n = 4).

^bCalculated by a two-tailed t test (results are shown in normal font) or by a Mann-Whitney test (results are shown in italics). A P value of ≤5 × 10⁻² was considered significant. NS, not significant.

^c—, FC was not ≥ 2-fold.

^dThe gene is a known target for NF-κB.

Overall, the majority of the genes increased in mice infected with strain 22L were also increased when mice were challenged with strain RML. However, 10 of these upregulated genes involved in inflammation (Ccl2, Ccl5, Ccl7, Ccl8, Ccl12, Cxcl5, Cxcl9, Cxcl10, Cxcl13, and Tnfsf11) were transcribed in greater amounts (≥2.0-fold) in the brains of mice infected with 22L than in those infected with RML. Of the genes that have been previously identified to increase during scrapie infection, there were five (Ccr1, Ccr3, Ccr5, Il2rb, and Il10rb) that were increased during 22L infection but unchanged when mice were infected with strain RML. Only Cd40lg, encoding the CD40 ligand (CD154), was altered in mice infected with strain RML but unchanged in mice infected with strain 22L. These data indicate that the majority of the altered genes encoding key inflammatory mediators are shared between mice challenged with each strain, but there are a few scrapie strain-specific differences in transcription of host inflammatory genes.

We also studied the protein levels of 14 cytokines and chemokines, including two newly identified, from RML and 22L-infected mouse brain homogenates at the clinical endpoint and compared them to mock-infected mice (Fig. 3). Analysis of these samples indicated that the mean levels for most of the cytokines were similar for the two scrapie strains. The exceptions were IL-12p40, CXCL9, CXCL10, and CXCL13, in which the mean levels were more elevated in 22L-infected mice than in RML-infected mice.

FIG 3.

Comparisons of cytokines in the brain tissue of C57BL/6 mice infected with either scrapie strain 22L or strain RML. Protein levels of 14 cytokines in the brains of scrapie-infected mice were measured by multiplex assay (IL-12p40, CCL2, CCL3, CCL5, GM-CSF, IL-1α, IL-1β, IL-13, CXCL1, CXCL9, and CXCL10) or by quantitative ELISA (CCL22, CXCL13, and IL-1Ra) as indicated. Each data point represents the result from an individual mouse, and the bar in each data set represents the mean. The error bars indicate 1 standard deviation from the mean. Statistical analysis was performed using the Student t test to compare infected to mock-infected groups at a given time point. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001.

Numerous inflammatory response genes were unchanged or undetected during the course of infection (see Table S1 in the supplemental material). For example, of the 25 interleukin-related genes assayed, 14 were either unchanged or not detected when mice were infected with strain 22L or RML. We were also unable to detect any change in transcription of the gene encoding IFN-γ in mice infected with either of the scrapie strains, although changes in many interferon-stimulated genes (such as Isg15, Oas1a, Cacl10, and Irf9) were observed in analyses of JAK-STAT signaling pathways using RNA samples derived from C57BL/10 mice infected with strain 22L (Table 3). These data suggest an alternative regulatory pathway independent of IFN-γ signaling in the brain during scrapie infection. Thus, in brain tissue there appeared to be a distinct subset of upregulated inflammatory cytokines and receptors associated with infection by both of the scrapie strains studied here.

Assessment of IL-1-Ra production on scrapie pathogenesis.

One of the newly identified inflammatory genes upregulated during infection with both scrapie agents was Il1rn. Il1rn encodes the protein IL-1Ra and is an endogenous antagonist of the IL-1 cell surface receptor for IL-1 (IL-1R1). IL-1Ra can occupy the IL-1R1-binding site and inhibit IL-1 signaling. Recently, IL-1Ra has been observed to have agonist activity in the mouse hippocampus that is independent of its function as an antagonist of the IL-1R1 receptor (43). Since IL-1 has been described as a master regulator of neuroinflammation (44) and the loss of IL-1R1 receptor gene expression has been shown to increase the survival time of mice during infection with scrapie (45), we assessed the effect of IL-1Ra expression on scrapie pathogenesis.

Mice lacking the gene for IL-1Ra production (IL1rn^−/−) or transgenic mice that overexpress IL-1Ra (TgIL1rn) were infected intracerebrally with either the 22L or RML strain of scrapie. Neither the IL-1Ra-deficient mice nor the transgenic overexpressor mice had significant changes in tempo of disease or in survival time compared to littermate controls (Fig. 4A to D). Furthermore, gliosis or the accumulation of PrPres during infection was not altered in the brains of infected IL1rn^−/− or TgIL1rn mice (Fig. 5). We quantified the IL-1Ra production from a random selection of mice from these experiments (Fig. 4E). IL-1Ra was undetectable in the brains of uninfected mice but was significantly increased in the brains of scrapie-infected mice. As expected the IL1rn^−/− mice infected or uninfected did not produce detectable levels of IL1Ra in the brain. Conversely, uninfected TgIL1rn mice produced a mean of >6,500 fg IL-1Ra/mg of brain tissue, and scrapie-infected TgIL1rn mice produced a mean of >30,000 fg IL-1Ra/mg of brain tissue. We also analyzed the production of 23 inflammatory cytokines by using the Bio-Plex system, and results indicated no significant change in cytokine production during scrapie infection of IL1rn^−/− or TgIL1rn mice relative to C57BL/6 controls (data not shown). Therefore, the loss or the overexpression of a major negative regulator of IL-1 signaling during scrapie infection had no discernible effect on disease or expression of the cytokines tested.

FIG 4.

Challenge of IL-1Ra-deficient and IL-1Ra-overexpressing mice with two different strains of scrapie. Survival curves of IL-1Ra-deficient (IL1rn−/−) relative to C57BL/6 littermate controls inoculated i.c. with scrapie strain 22L (A) or RML (B) indicated no difference in scrapie disease progression when the endogenous IL-1 receptor antagonist was absent. Survival curves of IL-1Ra-overexpressing (TgIL1rn) and C57BL/6 littermate controls inoculated i.c. with scrapie strain 22L (C) or RML (D) suggested that blocking IL-1 signaling through IL-1RI has no observable effect on scrapie pathogenesis. To ensure that IL-1Ra levels in the brains were absent in IL1rn^−/− mice but elevated in TgIL1rn mice, IL-1Ra amounts in brain tissue were quantified in age-matched mock-treated (M) and 22L- and RML-inoculated mice at the clinical endpoint (E).

FIG 5.

Detection of gliosis, vacuolation, and PrPres formation in scrapie-infected mice. (A) Astrogliosis in 22L-infected and age-matched uninfected IL-1Ra-deficient (IL1rn^−/−) and IL-1-Ra-overexpressing (TgIl1rn) mice was assessed in representative mice by using anti-GFAP antibody in sagittal brain and thalamus sections. Immunohistochemical analysis showed the density of activated astrocytes in 22L-infected IL1rn^−/− mice was similar to that of infected TgIL1rn mice but differed from results in uninfected mice. Similar GFAP results were obtained with brain sections from mice inoculated with strain RML (data not shown) and when sections from mice inoculated with strain 22L or RML were stained with anti-IBA-1 to demonstrate microglial activation (data not shown). H&E staining of thalamus sections showed vacuolation associated with scrapie infection that was absent in uninfected mice. Vacuolation was similar for infected IL1rn^−/−, TgIL1rn, and C57BL/6 control mice. All thalamus sections stained with anti-GFAP or H&E are shown at the same magnification. Bar, 100 μm. (B) Immunoblot detection of PrPres in the brains of representative IL1rn^−/−, TgIL1rn, and C57BL/6 littermate control mice inoculated with scrapie strains 22L and RML. All samples were treated with proteinase K as described in Materials and Methods. Each lane was loaded with 0.36 mg whole-brain equivalent and probed with anti-PrP antibody D13. Protein mass markers (in kDa) are indicated to the left of the immunoblots.

We expanded our studies by assessing the expression of the 84 key inflammatory genes in brains of 22L scrapie-infected IL1rn^−/− or TgIL1rn mice in an ICR array. There were no significant changes in the gene expression profile of IL1rn^−/− mice relative to controls during scrapie infection with strain 22L, with the exception of Il1rn, that was undetectable in the knockout (Fig. 4E and data not shown). Furthermore, we evaluated changes in inflammatory gene transcripts in the brains of TgIL1rn mice relative to littermate controls infected with strain 22L. Again, the only significant alteration in transcription between 22L-infected TgIL1rn and controls was the level of Il1rn, which was >100-fold greater in the infected TgIL1rn mouse brains and led to an extreme overproduction of IL-1Ra in response to scrapie infection in the brain (Fig. 4E). These array data suggest that with respect to mouse-adapted scrapie, the lack of or the overproduction of IL-1Ra in the brain has little effect on the key inflammatory genes tested, including genes encoding IL-1α, IL-1β, and the receptor IL-1R1.

DISCUSSION

Our analysis identified 10 new inflammatory genes and 5 new STAT-associated genes upregulated in the brains of mice infected with scrapie strain 22L. Four of these newly identified inflammatory genes (Il1rn [IL-1Ra], Ccl8 [CCL8/MCP-2], Tnfsf11 [Tnfsf11/RANKL], and Osm [OSM]) were increased preclinically as early as 70 days p.i. and joined a list of 14 previously identified inflammatory markers that were similarly upregulated (Table 1). The function of these 4 newly identified genes in the brain during scrapie disease is unknown, but IL-1Ra is principally described as an antagonist of the receptor IL-1R1 and functions to reduce IL-1 signaling. Thus, we suggest that IL-1Ra may be increased to limit IL-1 signaling and control inflammation during scrapie infection. CCL8 is a chemotactic factor (similar to CCL2, CCL7, and CCL13) that can attract monocytes (46) and may contribute to the recruitment of macrophages into the brain during scrapie infection (47). Tnfsf11 (RANKL) activates brain regions involved in thermoregulation to induce fever (48), is produced by microglia in a model of ischemia, and can be neuroprotective in vitro (49). OSM is a growth regulator, is a member of the IL-6 family, and has been shown to inhibit neural precursor cell proliferation in vitro and in vivo (50). These new inflammatory markers, combined with the 20 other upregulated genes at 70 days p.i. (Tables 1 and 2), indicate that insult or cellular damage is likely being perceived by the CNS immune surveillance system, which contributes to gliosis and monocyte recruitment well before clinical signs.

Several of the genes/proteins found to be chronically increased during scrapie infection could be damaging to the host CNS. Expression of Oas1a, Isg15, Tnfsf11, Olr1, and Ccl5 are associated with triggering apoptosis in cells (51–56), and expression of Cxcl10, Ccl2, A2m, and Tnf can contribute to neurotoxicity in other disease models (57–62). Additionally, others have reported that deletion of Ccl2 (20) or Cxcr3 (the receptor for CXCL10, CXCL9, and CXCL11, which are increased in our model) (46) can increase survival time in mice after scrapie infection, suggesting that signaling through these chemokines and their receptors can lead to damage. IL-1Ra has more recently been demonstrated to increase JNK phosphorylation in hippocampal synaptosomes (36), potentially contributing to neurodegeneration (23). Furthermore, a persistent inflammatory environment, similar to that observed in prion disease, with increased expression of cytokines has been shown to direct neuronal precursor cells to differentiate into astroglia, reducing the percentage of new neurons and leading to an excess of astroglia (10, 16, 40). The cumulative effect of chronic exposure of neurons and glia to these cytokines/chemokines in prion disease likely produces an environment leading to functional and structural neuronal damage.

It is unclear which cell types in the brain are responsible for the changes in gene transcription, but some of the altered expression levels of cytokines and their receptors may occur in activated astroglia and/or microglia. Astrocytes are capable of producing chemokines like CCL2 in vitro when exposed to scrapie-infected brain homogenate (21) or when exposed to Toll-like receptor agonists (63), and they can produce CXCL10 in the brain during scrapie infection (64). In contrast, cytokines such as IL-12 (65) and OSM (66) are known to be preferentially produced by activated microglia. Yet, it is clear both microglia and astrocytes are capable of expressing many proinflammatory effectors, like Tnf, Il1a, Il1b, Ccl6, Cxcl1, and Cxcl10, when similarly stimulated with the appropriate TLR agonists (63) or when glia are analyzed ex vivo from Alzheimer's mice (67). Other cell types in the brain, including neurons, oligodendrocytes, and endothelial cells, have all been shown to produce and respond to cytokines and chemokines (68–70), and these cells may also contribute to the alterations in transcription observed in prion disease. Furthermore, macrophages are recruited to the brain during scrapie infection (47) and likely contribute to the chronic inflammatory milieu. We attempted to use dual staining to detect which cell types produced many of these molecules, but we were unsuccessful. This result might be due to the dispersion of these secreted immune effectors throughout the neuropil, as well as a low signal-to-noise ratio.

The chronic and increasing neuroinflammation observed was not restricted to scrapie infection with strain 22L. Though strains 22L and RML share an origin and were derived from Moredun Institute's sheep scrapie brain pool 1 (7, 8), they differ in incubation period and in neuropathology (6–8). Mice infected with either strain 22L or RML gave strikingly similar profiles for increased inflammatory genes and proteins. However, there were also several differences between the inflammatory responses that might be attributed to the individual strains (Table 4 and Fig. 3). The biological relevance of these differences in determining the distinct neuropathologies for each of the strains is unknown.

Using the Signal Transduction Pathway Finder array, we measured the activation of 10 specific pathways (Table 2). Of these 10 commonly studied signal transduction pathways, our data suggested that only the STAT and NF-κB pathways were substantially activated in our scrapie-infected mice. The increase in transcription of many NF-κB- and STAT-responsive genes, including many cytokines and chemokines in prion-infected brains (indicated in Tables 1 and 3), preceded the time frame of reported neuronal damage and vacuolation in the brain (4, 5). Over half of the genes increased in expression in our analyses could be activated by NF-κB (see the footnotes for Tables 1 through 4 and also the website http://www.bu.edu/nf-kb/gene-resources/target-genes/ for a compendium of NF-κB-activated genes), and several other genes are known to be regulated by specific STAT complexes. Both of these transcription factor complexes are activated in the scrapie-infected brain (27, 71) and likely contribute to neuroinflammation during scrapie pathogenesis.

In addition to acting independently, STAT1 and STAT3 can also act synergistically with NF-κB. pSTAT3 and NF-κB have been shown to affect transcription at the promoters controlling many of the genes described (i.e., Cxcl10, Ccl4, and A2m) (72–76), and together they also strongly influence the expression of acute-phase proteins, such as haptoglobin, ceruloplasmin, α₁-antichymotrypsin, and serum amyloid A (75, 77), which are increased in the serum and brain during scrapie infection (unpublished data) (41, 78, 79). Moreover, components of the NF-κB complex, like RelA, can interact directly with STAT3 to alter transcriptional activity (80–82). In addition, evidence for synergism of NF-κB and STAT1 has also been shown for the expression of many inflammatory genes, such as Ccl5, Cxcl9, Nos2, and Icam1 (83–87), which are also increased during scrapie infection. Thus, synergy might be important in our system. However, not all NF-κB- and/or STAT-responsive genes were altered in our study (see Tables S1, S2, S3, and S4 in the supplemental material), suggesting that selectivity exists at specific promoters during scrapie infection.

Inflammation in the CNS often occurs by activation of glia prompting the synthesis and release of numerous inflammatory mediators. One prominent pathway that is activated during neuroinflammation in the CNS is the JAK-STAT signaling pathway (88). In a Parkinsonian monkey model, STAT1 is activated in the brain years after disease onset (89). In the mouse model of Alzheimer's disease, activation of STAT3, but not STAT1, is increased in the brain; studies with cultured neurons have indicated that STAT3 activation contributes to neuronal death after β-amyloid exposure (90). These findings and our studies suggest that STAT activation is common during neuroinflammation and degeneration, but it is likely that the downstream responses vary in the different neurodegenerative diseases. More thorough comparisons of signal transduction and gene transcription in neurodegenerative diseases using similar platforms are warranted.

Many of the newly identified genes from our analysis were increased prior to clinical signs of disease. One of these genes is Il1rn, which encodes the endogenous IL-1Ra. Mice deficient in IL-1RI have a delayed onset of disease with increased survival times when challenged intracerebrally with scrapie strain 139A (45). These data suggest that alteration of IL-1 signaling by either loss or overproduction of the receptor antagonist IL-1Ra might change the course of scrapie pathogenesis as well as potentially modify the inflammatory response in the brain. Our results indicated that loss or overexpression of the receptor antagonist IL-1Ra had no measurable effect on disease tempo, pathology, measured inflammatory markers, or the expression of genes in the brain that are key to inflammation during scrapie infection. IL-1Ra overexpression in TgIL1rn mice has been reported to positively regulate serum IL-1 levels when mice are stimulated with endotoxin (91), but this positive regulation of IL-1 was not seen in brains during scrapie infection.

The lack of an observable change in inflammation in TgIL1rn mice infected with scrapie was surprising, since mice deficient in the receptor IL-1RI not only demonstrate an increase in survival but also display a concomitant decrease in Cxcl9 and Cxcl10 expression during the infection (45). A similar effect would be expected since overexpression of IL-1Ra or loss of the IL-1 receptor would similarly disrupt IL-1 signaling, but we observed no change in Cxcl9 or Cxcl10 gene expression in the brain in infected TgIL1rn mice versus infected littermate controls (data not shown). One possibility is that the levels of IL-1Ra achieved during scrapie infection in TgIL1rn mice were insufficient to elicit the desired reduction in IL-1 signaling. However, this was unlikely, since mock-challenged TgIL1rn mice produced IL-1Ra levels of ≥6,500 fg/mg of brain tissue, whereas scrapie-infected TgIL1rn mice had levels of ≥30,000 fg/mg of brain tissue. We calculated that scrapie-infected TgIL1rn mice had IL-1Ra:IL-1α and IL-1Ra:IL-1β ratios of 370:1 and 15:1, respectively, in the brain (data not shown). These levels are above the IL-1Ra:IL-1 protein ratio of 10:1 reported to affect IL-1 signaling in vitro (92). Therefore, a lack of sufficient IL-1Ra in our studies in the TgIL1rn mice does not appear to explain our results.

The upregulation of genes encoding proinflammatory cytokines and neurotoxic protein products during scrapie infection of the CNS likely contributes to additional glial activation and further increases in neuronal damage, thereby facilitating a self-perpetuating cycle of increasing neuroinflammation. This unchecked neuroinflammatory environment might contribute to pathogenesis by leading to increased neuronal loss and host death. Infection and/or production of de novo PrPres might directly activate glia, which could lead to neuronal damage through the activation of NF-κB and STAT complexes prompting neuroinflammation and production of toxic products. Furthermore, PrPres might directly damage neurons, resulting in indirect glial activation and leading to further neuronal damage through a similar mechanism.

Although gliosis and escalating neuroinflammation are evident during scrapie infection, the primary causes of these events remain unclear. PrPres accumulation in the infected brain is a common feature of prion disease, and there is usually, but not always, a spatial correlation between PrPres deposition and reactive gliosis in the brain during prion disease. However, PrPres may not be solely responsible for triggering the initial gliosis and neuronal damage observed. In rapid models of prion disease in tga20 and tg7 transgenic mice, which overexpress PrP, severe clinical signs occur with the presence of significant gliosis but with less PrPres deposition and vacuolation relative to levels in nontransgenic mice (93–95), suggesting that additional host factors other than PrPres may also contribute to prion disease progression.