Viral pre-challenge increases central nervous system inflammation after intracranial interleukin-1β injection

Adult 8-week-old (200 g), male Wistar rats (Harlan, Bicester, UK) were housed in specific pathogen-free (SPF) facilities under a standard light/dark cycle with food and water ad libitum. In each experiment, at least three animals were used per treatment group and per time point. All procedures were carried out in accordance with the UK Animals (Scientific Procedures) Act, 1986 and with approval of local and European ethical committees, and all reasonable efforts were made to minimise discomfort. Animals were randomly assigned to experimental, groups as per the ARRIVE guidelines [26].

Animals were anaesthetised with 2 to 3% isoflurane in oxygen. All surgical procedures were performed under an operating microscope (Wild M650, Leica, Milton Keynes, UK). Stereotaxic surgery was performed as described previously [27]. Briefly, the head of the anaesthetised animal was shaved and held in a stereotaxic frame (Kopf Instruments, Tajunga, California). The skull was then exposed with a midline sagittal incision and dental burr drill was used to make a small hole for the injection (co-ordinates from bregma: A/Pr +1.2 mm, M/L +3.0 mm, D/V -4.5 mm). One microlitre of rrIL-1Β (recombinant rodent-IL-1Β; National Institute for Biological Standards and Control (NISBC), Potters Bar, UK; 1  μg/μl) or vehicle (0.9% saline) was microinjected into the left striatum through a pulled glass microcapillary needle (tip <50  μm), as our standard laboratory model of brain inflammation [28]. The cytokine was delivered over a period of 2 minutes and the micropipette was then left in place for a further 5 minutes to allow the injected cytokine to diffuse away from the injection site and prevent reflux up the injection tract. Local anaesthetics were applied to wound sites prior to recovery. Body temperature was maintained on a heated blanket throughout the period of anaesthesia and the animals were allowed to recover in a heated chamber thereafter. For pre-challenge studies, animals were briefly anaesthetised using 2% isoflurane via a nose-cone and injected intravenously with either a luciferase-expressing adenovirus (AdLuc kindly provided by Prof. Len Seymour, Oxford) or vehicle (0.9% saline) 24 hours prior to stereotaxic surgery. Virus was injected intravenously (iv) at a concentration of 5 - 10⁹ pfu (plaque-forming unit)/animal.

Animals were surgically anaesthetised and blood was collected via cardiac puncture into EDTA-coated tubes (Teklab, Durham, UK), or into serum separators (Southern Syringe Services Ltd, London, UK). Animals were then transcardially perfused with 0.9% saline containing 5,000 U/l heparin followed by a periodate lysine paraformaldehyde solution (PLP: 2% paraformaldehyde, lysine, periodate and 0.05% glutaraldehyde). Tissues were removed, post-fixed in PLP for 4 hours and further fixed in 30% sucrose for >12 hours. Ten-micron-thick sections were cut on a cryostat (Leica, Bucks, UK) and mounted on gelatine-coated slides. One-microlitre volumes of whole blood were taken from the EDTA coated tubes and films were produced on clean glass slides. Blood films were allowed to dry at room temperature for 5 minutes before being stored at -20°C.

Immunohistochemistry was performed as described previously [29]. Briefly, tissue sections were quenched for endogenous peroxidase activity in alcoholic hydrogen peroxide (H₂O₂) solution (0.03% H₂O₂ in methanol). Using the Shandon-Sequenza system, non-specific Fc-dependent binding was blocked with 10% normal serum (from the species in which the secondary antibody was raised) and primary antibody incubation took place. These included anti-neutrophil serum (HB-199, made in house; 1:10,000, 2 hours at room temperature); anti-ED-1 (mouse monoclonal, AbD Serotec, Oxford, UK; 1:200 overnight at room temperature); anti-ICAM-1 (mouse monoclonal, BD Pharmingen, Oxford, UK; 1:50, 2 hours at room temperature); anti-luciferase (goat polyclonal, Promega, Southampton, UK; 1:1,000 overnight at room temperature) and anti-CD11b (OX-42 clone, Serotec, Oxford, UK; 1:200 overnight at room temperature). Sections were then incubated with the appropriate biotinylated secondary antibody (Vector Laboratories, Peterborough, UK) before an avidin-biotin-peroxidase solution (Vectastain Elite ABC, Vector Laboratories, Peterborough, UK) was added. Peroxidase labelling was visualised by exposure to the chromogen diaminobenzidine hydrochloride (0.5  μg/ml) in 0.1 M phosphate buffer (0.05% H₂O₂) for varying times until optimal contrast was achieved. To control for endogenous biotin present in the liver, a separate avidin-biotin blocking procedure using an avidin-biotin blocking kit (Vector Laboratories, Peterborough, UK) was employed. For CNS tissue, samples were taken serially through the injection site. For blood counts, smears of 1  μl of blood were taken in triplicate and stained as above. For each analysis, four representative fields were chosen for quantitation and the average number of positive cells was calculated and expressed as number of cells per mm². Tissue sections were counterstained with cresyl violet and immuno-positive labelling was quantitated only when associated with a cell nucleus. Spatial maps of positive immunoreactivity were created using plate 15 from the rat brain atlas [29]. Specifically, the sections were viewed at 1.6 magnification using a microscope attached to a camera lucida; positive staining was dotted onto white paper using a black marker pen. Images were scanned and superimposed onto plate 15 of the rat brain atlas to form spatial maps of overall immunoreactivity. All immunohistochemistry was performed blind, as per ARRIVE guidelines [26].

One cubic millimetre punches of striatal tissue at the level of the injection were removed for RNA extraction. RT-PCR assays were performed as previously described [30]. Samples were run against absolute copy number standard curves generated from serially diluted cDNA from LPS-challenged rat liver. Primer and probe sets for rat IL-1Β, CXCL-1, CCL-2, CXCL-10, CCL-3 (MIP-1α), and CCL-4 (MIP-1β) were designed using the Roche universal probe library assay design centre (Roche.com). Samples were analysed using a Roche Light Cycler 480® (Roche Diagnostics, Welwyn Garden City, UK) and all reagents were used according to manufacturer's instructions. Briefly, gene-specific primers were combined with a FAM/TAMRA labelled hybridisation probe (for sequences see Additional file 1: Table S1). PCR was run according to standard conditions [31]. Analysis was performed using the Pfaffle method, taking the standard curve to determine reaction efficiency followed by a comparative-cycle-threshold method. Results are expressed as absolute copy number.

Data are presented as mean ± standard error of the mean (SEM) for a minimum of three animals at each time point. Data were analysed using Student's t-test and two-way ANOVA with Bonferroni-Dunn post hoc testing for multiple groups where appropriate. Results were considered statistically significant when P <0.05 as compared to vehicle-treated or naïve controls where appropriate.

In keeping with previous observations, the microinjection of IL-1Β into the brain caused an up-regulation of cytokines and chemokines known to be involved in leukocyte migration and recruitment (CXCL-1, IL-1Β, and CCL-3) at 12 hours post-injection (Additional file 2: Figure S1). The microinjection of IL-1Β did not induce the expression of TNF in the brain (data not shown), which, while surprising, supports previously reported findings [32]. While cytokine and chemokine expression profiles were examined at all time points, significant differences in expression were found between vehicle/IL-1Β and AdLuc/IL-1Β at 3 days post-injection (Figure 1). At this time point the microinjection of IL-1Β into the brain had caused an elevation in mRNA for CXCL-1, CCL-2, IL-1Β and CCL-3 relative to naïve levels (Figure 1). However, the presence of the viral load caused an up-regulation in CXCL-1, CCL-4 and CXCL-10 production following the microinjection of IL-1Β when compared to vehicle/IL-1Β injection alone (Figure 1). It was of note that the microinjection of vehicle, in the presence of AdLuc also induced an up-regulation in CXCL-10 production in the brain. By 7 days, only CCL-2 and CXCL-1 remained elevated in all IL-1Β injected animals (with virus or without) compared to naïve or AdLuc/saline-injected animals (Additional file 3: Figure S2); at this time none of the other chemokine mRNA species measured in the CNS were elevated above naïve levels. It is important to note that immunohistochemistry for luciferase revealed the presence of luciferase protein in the liver, but not in the brain of the challenged animals (Additional file 4: Figure S3).

The microinjection of vehicle, in the presence of intravenous AdLuc, did not cause any observable changes in the expression of ICAM-1 relative to baseline levels observed in naïve animals (Figure 3A and B). As expected, the microinjection of IL-1Β into the brain induced widespread up-regulation of ICAM-1 expression (Figure 3C) that was most prominent around vessels actively involved in recruitment of leukocytes to the injected striatum. Some staining was also evident, to a lesser extent, in the cortex in the ipsilateral hemisphere. When animals were microinjected with IL-1Β into the brain, in the presence of intravenous AdLuc, the up-regulation in ICAM-1 expression appeared more widespread (Figure 3D). Again, the strongest ICAM-1 positivity was observed in the ipsilateral hemisphere, with further up-regulation above naïve in other areas actively involved in recruitment. The most notable observation was that in the striatum pre-conditioned IL-1Β animals there appeared to be two populations of ICAM-1+ vessels. There were small vessels positive for ICAM-1, but not actively involved in recruitment (arrowhead, Figure 3D), and those heavily positive for ICAM-1, and cuffed by large numbers of neutrophils (arrow, and inset, Figure 3D).

The microinjection of IL-1Β into the brain in the presence of intravenous vehicle resulted in elevated numbers of neutrophils present in the ipsilateral hemisphere from 12 hours post-injection (Figure 1A). By 3 days, the number of neutrophils had dropped to approximately 35% of those seen at 12 hours, and by 7 days had reduced to baseline. The presence of intravenous AdLuc administered 24 hours prior to the microinjection of IL-1Β caused an elevation in the numbers of neutrophils observed in the injected hemisphere at 12 hours as compared to vehicle (P <0.01), which persisted to 3 days post-injection. At this time, neutrophils were presents in large cuffs around vessels in the injected striatum in animals receiving AdLuc that was not evident in other groups (Figure 4A, arrowheads). By 7 days, neutrophil numbers were for the most part back to baseline, although slightly more were observed in animals that received AdLuc, than the IL-1Β alone group. In animals microinjected with vehicle into the brain, in the presence of intravenous AdLuc, no measurable numbers of neutrophils were observed in the brain at any time point studied. The microinjection of IL-1Β into the brain in the presence of intravenous vehicle also resulted in some recruitment of neutrophils to the contralateral hemisphere 12 hours after the striatal injection, although at a much lower level than in the injected hemisphere (Figure 4A). These were mainly observable in the cortex, and very few were present in the striatum. By 3 days, the numbers of neutrophils had returned to baseline levels. The presence of AdLuc administered 24 hours prior to the microinjection of IL-1Β caused an elevation in the numbers of neutrophils observed in the contralateral hemisphere at 12 hours, but by 3 days the numbers of neutrophils were again back to baseline.

The microinjection of IL-1Β into the brain, in the presence of intravenous vehicle, also elevated the number of ED-1-positive cells present in the ipsilateral hemisphere from 12 hours to 7 days (Figure 4B). The number peaked at 3 days, but remained well above baseline until day 7. The presence of intravenous AdLuc administered 24 hours prior to the microinjection of IL-1Β caused a further significant elevation in the numbers of ED-1-positive cells all time points. In the contralateral hemisphere, there was a small increase in the number of ED1-positive cells at 12 hours (P <0.05). By 3 days, the number of ED1 cells was slighted elevated over baseline, but similar in all the treatment groups.

In order to better represent the pattern of recruited leukocytes into the CNS, spatial maps were created to show the patterns and density of cellular recruitment over time. When animals were microinjected with vehicle, in the presence of intravenous AdLuc, no neutrophils were observed in the brain at any time point (Figure 5, left column). The microinjection of IL-1Β in the presence of intravenous vehicle caused large numbers of neutrophils to be observed in the ipsilateral striatum and cortex at 12 hours (Figure 5, middle column). Limited numbers of neutrophils were observed in the contralateral cortex, and bilateral meningitis was present. By 3 days, significant numbers of neutrophils remained present in the injected striatum, but there was much less recruitment occurring in the cortex, and only isolated cells were observed in the contralateral hemisphere. At this time point, the meningitis was dramatically reduced in both spread and magnitude, being largely ipsilateral in nature. By 7 days, only isolated cells were observable in the injected striatum, the rest of the brain resembled naïve tissue. When the microinjection of IL-1Β was combined with intravenous administration of AdLuc 24 hours beforehand, the numbers of neutrophils observed in the brain at 12 hours post-IL-1Β was significantly increased (Figure 5, right column). Within the injected hemisphere, there were no dramatic changes in the manner of recruitment observed, merely that the vessels actively engaged in recruitment was spread over a larger area, contributing to the increase in numbers. The most striking changes were observed in the contralateral hemisphere, where the area in which active recruitment was occurring was much more widespread within the cortex, and to a lesser extent, within the striatum. By 3 days, a similar pattern was observed to that seen in animals injected with IL-β alone, but with some spatial differences. The area over which recruitment was occurring in the injected hemisphere was much larger, predominantly in the cortex. Furthermore, in the injected striatum, large numbers of neutrophils were cuffed around vessels, five to seven cells thick, indicating that the process of diapedesis was occurring to a much greater extent than in the IL-1Β/vehicle group. By 7 days, the differences between the groups were less defined. In the group administered AdLuc there were marginally elevated numbers of neutrophils in the injected striatum, and there was still evidence of meningitis, although the majority of the section was devoid of neutrophils.