[1] 17w Inducing inflammation following subacute spinal cord injury in female rats: A double-edged sword to promote motor recovery

[1] 208w The inflammatory response following spinal cord injury is associated with increased tissue damage and impaired functional recovery. However, inflammation can also promote plasticity and the secretion of growth-promoting substances. Previously we have shown that inducing inflammation with a systemic injection of lipopolysaccharide in the chronic (8 weeks) stage of spinal cord injury enhances neuronal sprouting and the efficacy of rehabilitative training in rats. Here, we tested whether administration of lipopolysaccharide in female rats in the subacute (10 days) stage of spinal cord injury would have a similar effect. Since the lesioned environment is already in a pro-inflammatory state at this earlier time after injury, we hypothesized that triggering a second immune response may not be beneficial for recovery. Contrary to our hypothesis, we found that eliciting an inflammatory response 10 days after spinal cord injury enhanced the recovery of the ipsilesional forelimb in rehabilitative training. Compared to rats that received rehabilitative training without treatment, rats that received systemic lipopolysaccharide showed restored motor function without the use of compensatory strategies that translated beyond the trained task. Furthermore, lipopolysaccharide treatment paradoxically promoted the resolution of chronic neuroinflammation around the lesion site. Unfortunately, re-triggering a systemic immune response after spinal cord injury also resulted in a long-term increase in anxiety-like behaviour.

[1] 445w Rehabilitation is currently one of the most effective treatment options to promote motor recovery following incomplete spinal cord injury (SCI), with early intervention generally providing more favourable outcomes (Nam et al., 2017;Norrie et al., 2005;Scivoletto et al., 2005;Sumida et al., 2001). This observation corroborates research indicating that the acute injury environment increases the capacity for plasticity (Biernaskie, 2004;Ding et al., 2005;Scivoletto et al., 2005;Sumida et al., 2001). Neuroinflammation likely plays a role in this process, as early components of the inflammatory process have been shown to be beneficial for the natural but limited repair process following SCI (Anderson et al., 2016;Arnett et al., 2001). Indeed, Chen et al. showed that over-expression of neurotrophin-3 promoted sprouting of corticospinal tract (CST) axons in the acutely lesioned, but not chronically lesioned or unlesioned spinal cord (Chen et al., 2006). This enhanced growth of the CST is important for motor recovery as the CST is critical for fine motor control (Martin, 2005;Piecharka et al., 2005). Neuronal sprouting of the CST could be re-established in the chronically lesioned spinal cord by triggering an immune response with lipopolysaccharide (LPS), implicating immune activation as a key component of neurotrophin-3 induced axonal growth (Chen et al., 2008). Further evidence for the link between injury-induced inflammation and plasticity comes from research in the optic nerve. Using a model of optic nerve crush, Benowitz et al. showed that oncomodulin, a protein released from macrophages, is a formidable growth promoting signal from the immune system (Yin et al., 2009). Further support of the association between inflammation and neuronal plasticity is the finding that the window of opportunity for effective rehabilitative training can be reopened by injecting LPS in rats with chronic SCI (Torres-Espín et al., 2018a). LPS treatment resulted in increased CST sprouting and a robust increase in rehabilitative training-induced motor recovery (Torres-Espín et al., 2018a). This points towards a dichotomous role of inflammation as it can both exacerbate tissue damage and yet is an essential promotor of plasticity following SCI (Gensel and Zhang, 2015;Jones et al., 2005;Rust and Kaiser, 2017). Although the above evidence suggests that the acute neuroinflammatory response may promote plasticity, there is also substantial research implicating inflammation as a key factor in secondary damage following SCI (Gris, 2004;Zhu et al., 2014). We therefore hypothesized that enhancing inflammation in the subacute stage of SCI using LPS, at a time point when the lesion environment is still in a proinflammatory state (Popovich et al., 1997), would not have the same beneficial effect as chronic administration. To test this, rats received a systemic injection of LPS 10 days following a mild cervical SCI to trigger an immune response followed by 6 weeks of rehabilitative training.

[1] 177w Rats received a cervical SCI that selectively impaired the reaching and grasping ability of their preferred forepaw. Ten days after injury rats received either saline or LPS injection and were monitored for sickness behaviour, weight and temperature right before the treatment administration and 4, 8, 24, 48, and 72 h later. LPS treatment resulted in visible sickness behaviour beginning at 4 h and lasting up to 36 h postinjection (time × treatment effect (F (6, 156) = 125.0, p < 0.0001), time effect (F (6, 156) = 125.0, p < 0.0001) and treatment effect (F (1, 26) = 517.4, p < 0.0001)) (Fig. 1A). This sickness behaviour was accompanied by a significant reduction in weight relative to saline controls between 24 and 72 h post injections (time × treatment effect (F (7, 182) = 43.30, p < 0.0001), time effect (F (4.003, 104.1) = 86.51, p < 0.0001) and treatment effect (F (1, 26) = 8.947, p = 0.001) (Fig. 1B). There were no significant differences measured in body temperature between LPS or saline groups (Fig. 1C).

[2] 334w One week following SCI, all rats experienced a drastic (approximately 40%) reduction in success rate in the SPG task when compared to baseline/preinjury (Fig. 2A). After receiving LPS or saline treatment and undergoing 6 weeks of rehabilitative training, both groups followed a similar trend and exhibited an improved success rate over the first 3 weeks of training. However, at this same time point the saline group plateaued in their recovery, whereas in weeks 4 through 6 the LPS treated group showed a modest increase in success rate compared to the saline group (Fig. 2A). There was a significant effect of treatment over time (F (6, 156) = 2.716, p = 0.016), as well as a significant time effect (F (4.161, 108.2) = 28.76, p < 0.0001) (Fig. 2A). There were no differences between LPS or saline groups in the number of attempts made to reach for the pellet throughout the rehabilitative period (Fig. 2B). To discriminate between compensatory strategies and true functional recovery, a gap was introduced in the pellet dispenser at the offset of rehabilitative training to prevent the rats from scooping the pellets into their mouths (Fig. 2C). Once this compensatory behaviour was eliminated, it revealed that the LPS treated animals performed significantly better than the saline group (p = 0.046) (Fig. 2D). To further analyze the skilled reaching and grasping pattern, these movements were decomposed and analyzed in 11 components (Fig. 2E). Prior to SCI, both LPS treated and untreated groups displayed a similar reaching and grasping pattern (Fig. 2F). At the offset of rehabilitative training, rats that received LPS had generally better performance in the grasping, supination and release movements compared to rats that received saline only (significant treatment effect F (1, 275) = 5.397, p = 0.021) (Fig. 2G). In summary, these results reveal that untreated rats develop compensatory strategies to improve their success rate in the SPG task, whereas rats that received a single dose of LPS 10 days after SCI recovered grasping ability of their uninjured paw.

[3] 242w Non-trained behavioural tasks were used to determine whether subacute LPS treatment had an effect outside of the trained SPG task. There was no effect of LPS treatment on performance in the horizontal ladder (Fig. 3A), however rats that received LPS injections showed significantly improved recovery in the cylinder test relative to saline controls (p = 0.015) (Fig. 3B and 3C). To assess general locomotor activity, the distance travelled in an open field arena was analyzed at baseline, post SCI (before LPS or saline treatment), 1 day, 3 days and 6 weeks after treatment. As a result of the LPS-induced sickness response, rats in the LPS group travelled significantly less distance in the open field compared to the saline group on day 1 post treatment; this difference became insignificant by 3 days and there was no difference in activity in the open field by 6 weeks (p < 0.0001 at 1 day, time × treatment effect (F (4, 104) = 6.779, p < 0.0001) and time effect (F (3.251, 84.54) = 38.06, p < 0.0001) (Fig. 3D). There was an effect of injury on mechanical sensitivity such that the contralesional forepaw had increased sensitivity and the ipsilesional forepaw had reduced sensitivity (Post op: saline p = 0.041, LPS p = 0.043; Post LPS: saline p = 0.0002, LPS p = 0.0009; Offset (6 weeks post LPS): saline p < 0.0001, LPS p = 0.0003), however there was no effect of LPS (Fig. 3E).

[4] 482w To determine whether subacute LPS treatment induced plasticity of the injured CST, BDA was injected into the contralesional forelimb motor cortex and the sprouting of CST fibres into the grey matter was normalized to the number of BDA labelled descending CST axons (Fig. 4A) (Bareyre et al., 2004;Lindau et al., 2014;Mitchell et al., 2016;Torres-Espín et al., 2018a). The density of BDA + CST axon collaterals projecting into the cervical grey matter was quantified rostral to the lesion site (C2-C3, Fig. 4B) and immediately above the maximum lesioned area (C4, Fig. 4C). Both saline and LPS groups displayed a reduction of CST collaterals extending into the grey matter at the injury site compared to above the lesion at C2 -C3 (Fig. 4D, F, E, G). The overall density of CST collaterals projecting into the cervical grey matter was not significantly different between groups above (Fig. 4L) or at the injury site (Fig. 4M). There was no correlation between the grey matter CST density and success rate in training (Supplementary Fig. 1). To determine whether there were statistical differences between groups in the distribution of CST projections, the grey matter was sectioned into 75 μm 2 bins. Differences were observed between groups both above (Fig. 4H) and at the injury site (Fig. 4I). However, there were a similar number of bins where either the saline or LPS group was significantly increased, suggesting no overall increase of one group over another. To further analyze the distribution of CST projections in a different manner, the grey matter was sectioned into 7 evenly spaced rings originating at the center of the spinal cord and propagating outwards. In both groups, the density of CST collaterals was highest in the first few ring sections closest to the central canal. Compared to the saline group, rats that received LPS displayed decreased density of CST collaterals near the central canal above the injury site, but a slightly increased density of axons at the lesion site in the first ring section (Fig. 4J, K). Overall, these data indicate that there was no significant effect of LPS treatment on CST projections into the cervical grey matter. Furthermore, there was no difference in lesion severity between saline or LPS treated animals (I), respectively. Colours denote which group is significantly increased. To further examine CST axonal projections, the grey matter was split into 7 sections of rings propagating from the center of the spinal cord. Quantification of BDA + pixel area above the lesion site (J) and at the lesion site (K) show that the majority of CST collaterals project to the intermediate grey matter. Quantification of the total normalized CST density above (L) and at the lesion site (M) revealed no significant differences between groups. (N) The lesioned area was calculated as a percentage of the total cross-sectional spinal cord area. LPS treatment had no effect on the lesion size. Error bars represent SEM.

[5] 122w Because inflammation has frequently been linked to anxiety and depression (Dantzer et al., 2008;De La Garza, 2005;Raison et al., 2006;Vogelzangs et al., 2013;Yirmiya, 1996), we wanted to determine whether enhancing inflammation in the subacute stage after SCI had any long-term consequences on anxiety-like behaviour. Therefore, rats were tested in the EPM 4 weeks following LPS or saline injections (Fig. 5A). LPS treated rats travelled significantly less distance in the maze compared to saline controls (p = 0.015). Rats that received LPS spent significantly less time in (p = 0.027) and made less entries into (p = 0.028) the open arms of the maze. Thus, a single dose of LPS in the subacute period following SCI induces a long-lasting increase in anxietylike behaviour.

[6] 153w 10 weeks following SCI, IBA1 and GFAP expression was assessed immediately rostral, at the lesion epicenter and immediately caudal to the injury (Fig. 6A-I). Rats that received LPS displayed significantly reduced density of IBA positive cells rostral to the lesion (p = 0.016) and at the maximum injury location (p = 0.009) (Fig. 6J, L). Caudal to the injury, LPS treated rats also displayed reduced IBA1 expression, however this did not reach significance (Fig. 6N). There was no effect of LPS on the morphology of the microglia, assessed by the length of microglial processes and number of process endpoints per cell (Supplementary Fig. 2). LPS treatment resulted in a long-term decrease of GFAP expression rostral, at (p = 0.003) and caudal to the lesion (p = 0.002) (Fig. 6K-O). These data suggest that a single dose of LPS 10 days after SCI can attenuate immune cell expression long-term (i.e., 8 weeks) after injection.

[1] 43w Previously we have shown that inducing inflammation with LPS 8 weeks after SCI amplifies the efficacy of rehabilitative training, which is often less effective in these chronic stages of injury (Norrie et al., 2005;Scivoletto et al., 2005;Sumida et al., 2001;Torres-Espín et al., 2018a).

[2] 141w In the present study we sought to determine whether LPS treatment earlier after SCI would also be able to increase training efficacy. We initially hypothesized that enhancing inflammation at this subacute time point would not have the same effect as chronic application since levels of inflammation and the capacity for motor recovery are already relatively higher (Popovich et al., 1997;Scivoletto et al., 2005;Sumida et al., 2001). Contrary to our hypothesis, we found that a single dose of LPS given 10 days following SCI had a beneficial effect on improving functional recovery of the injured forepaw that translated beyond the trained grasping task. Paradoxically, inducing inflammation with LPS resulted in a chronic decreased expression of microglia and astrocytes around the injury site. The beneficial effects of LPS came at a cost however, since LPS treatment induced a long-term increase in anxietylike behaviour.

[3] 224w We show that triggering an immune response in the subacute period following SCI in combination with rehabilitative training can enhance functional recovery. This recovery may in part be due to the paradoxical chronic resolution of neuroinflammation at the lesion site following subacute LPS treatment. However, inflammation can be a double-edged sword and therefore its manipulation should be considered cautiously. Although eliciting inflammation with LPS promoted functional recovery following SCI, it also caused a long-term increase in anxiety-like behaviour. Inducing neuroinflammation with LPS may generally enhance the plasticity of a variety of neural substrates, as evidenced by LPS-induced changes in the limbic system, pain sensitivity, and motor recovery (Calil et al., 2014;Guo and Schluesener, 2006;Torres-Espín et al., 2018a;Yirmiya, 1996). Given the widespread immune response triggered by LPS, future research should explore the temporal systemic and local immune response to LPS treatment for SCI. Furthermore, the timing of treatment intervention may still be an important factor. Although we have shown that inducing inflammation 10 days or 8 weeks following SCI both have a beneficial effect on functional recovery without exacerbating lesion size, it is very likely that inducing a systemic immune response in the acute (i.e., within days) lesion environment would have a detrimental effect on SCI pathology. Future work should be considered to optimize inflammation-induced plasticity by separating its beneficial aspects from detrimental behavioural consequences.

[1] 91w Adult female Lewis rats (n = 60 in 2 cohorts, Charles River Laboratories) were group housed (n = 5 per cage, treatment groups housed separately) with ad libitum access to water and 12 h on-off light cycle. Rats were food restricted during training periods (10 g per rat per day) and otherwise had ad libitum access to standard rat chow. The study was approved by a local animal care and use committee (Health Sciences) at the University of Alberta and complies with the guidelines of the Canadian Council for Animal Care.

[2] 116w Rats were anaesthetized using isoflurane (3% in 50:50 air:oxygen mix) and their dorsal neck shaved and cleaned with 10% chlorhexidine digluconate (Sigma-Aldrich). An incision was made in the skin above vertebrae C2-C5, the muscles above C3-C4 were split and a laminectomy was performed at C4. A dorsolateral quadrant SCI was performed at C4 on the side of the preferred paw using custom made blades. The muscle layers were sutured with 5-0 Vicryl and the skin was stapled with 9 mm surgical clips. Animals received 4 ml saline for hydration and 0.2 ml of buprenorphine (0.03 mg/ml) as analgesic immediately postoperatively and a second dose of 0.1 ml buprenorphine (0.03 mg/ml) was given 8 h after injury.

[3] 59w LPS was derived from Escherichia coli endotoxin (serotype 055:B5, Sigma-Aldrich) and dissolved in sterile saline for injection. Rats received a single intraperitoneal dose of 0.5 mg/kg LPS or saline 10 days following SCI. Skin temperature, weight and general sickness behaviour (piloerection, social isolation and reduced activity) was monitored preinjection, 4, 8, 24, 36, 48 and 72 h after injection.

[4] 64w Animals were placed in the centre of a black acrylic open field arena (100 × 80 × 30 cm) and video recorded from above for 5 min. The total distance moved was analyzed using custom motion-tracking software. This test was performed at baseline (before SCI), after SCI (before LPS/ saline injections), 1-and 3-days post LPS/saline injections, and again at the offset of rehabilitative training.

[5] 139w Rats were filmed as they traversed a horizontal ladder (100 cm long, 12 cm wide, 12 cm high, 3 mm diameter cross bars spaced between 2 and 3 cm with a 45 • mirror underneath). If a rat paused or turned around, the trial was considered invalid. 6 continuous ladder crosses per rat (3 per side) were used for analysis and the average between the three videos was used for each animal. For each paw, the total number of correct paw placements, paw slips (the paw contacted the ladder rung but slipped) and paw misses (the paw did not make contact with the ladder rung) were calculated. The final outcome measure was the percentage of correct paw placements for each paw. This test was performed at baseline, after SCI (before LPS/saline injections) and at the offset of rehabilitative training.

[6] 51w Rats were placed in an acrylic cylinder (21 cm wide × 23 cm tall) and video recorded for 3 min or until a minimum of 10 rears were made. The number of left and right forepaw placements were counted, and forepaw asymmetry was expressed as a percentage of ipsilesional paw placements.

[7] 101w Rats were acclimatized to the testing chamber prior to testing (IITC Life Science, CA, USA). Tactile sensitivity was assessed on both forepaws; if the rat was placing weight on the forepaw the score was not recorded. The Von Frey rigid tip probe was applied gradually in increasing pressure until the rat displayed a defined nociceptive response (paw retraction, licking) and the maximum pressure that elicited a withdrawal was noted. This test was repeated 3 times per paw, with a minimum of 3 min between measures. For each rat the average of the 3 measures for each paw was used for analysis.

[8] 119w Anxiety-like behaviour was assessed in the elevated plus maze (EPM) four weeks after LPS or saline injections in the second cohort of rats (n = 14 per group included regardless of participation in rehabilitative training). Rats were placed in the junction of two open arms and two closed arms, facing towards an open arm and allowed to explore the arena (100 × 100 cm and elevated 65 cm above ground) for 10 min. Time spent and entries into the open and closed arms as well as the total distance travelled were recorded from above as measures of anxiety-like behaviour. This test was used only once to avoid habituation to the maze. Offline video analysis was performed using customized software.

[9] 114w At the offset of rehabilitative training and after all behavioural testing, the anterograde tracer biotinylated-dextran amine (10% BDA; 10 000 MW, Life Technologies, New York, USA) was injected into the contralesional forelimb motor cortex to trace the ipsilesional corticospinal tract (CST). Using a dental drill, a 1.5 mm square window over the motor forelimb cortex was made (1-2.5 mm rostral and 1-2.5 mm lateral to bregma). Three injections of 1 µl BDA were made at a depth of1.5 mm into the cortex using a Hamilton syringe. Following tracing, the skin was sutured with 5-0 Prolene and the animals received 0.1 ml buprenorphine (0.03 mg/ml). Rats were euthanized and perfused 12 days following tracing surgeries.

[10] 268w To visualize BDA traced CST axons, frozen sections were acclimatized at 37 • C for 1 h and rehydrated in TBS (2 × 10 min) and TBS-TX (TBS with 0.5% Triton X-100) (2 × 10 min). The sections were incubated for 2 h at room temperature with 1:200 Streptavidin, Alexa Fluor™ 488 conjugate (Invitrogen) diluted in TBS-TX. The sections were then washed in TBS (4 × 10 min) and coverslipped with Fluoromount (Southern Biotech). For lesion analysis, slides were thawed at 37 • C for 1 h and rehydrated in TBS (2 × 10 min). Slides were then placed in 0.5% Cresyl Violet solution for 3 min followed by a serial dehydration in 50%, 75%, and 99% EtOH for 2 min each. Slides were then cleared in Xylene for 2 × 2 min and coverslipped using Permount mounting media. For IBA1 and GFAP analysis, slides were thawed for 1 h at 37 • C and rehydrated in TBS for 10 min followed by TBS with 0.3% Triton™ X-100 (TBS-T) for 10 min. A blocking buffer of 5% normal goat serum in TBS-T was applied for 1 h at room temperature. Sections were then incubated overnight at room temperature in rabbit-anti-IBA1 (1:500, Wako) and mouse-anti-GFAP (1:500, Sigma) antibodies with blocking buffer. 20 h later, sections were washed with TBS (3 × 10 min) and incubated with goat-anti-rabbit AF488-conjugated (1:500, Life Technologies) and goatanti-mouse AF555-conjugated (1:500, Life Technologies) antibodies in blocking buffer for 2 h at room temperature. Sections were then rinsed in TBS-T (2 × 10 min) followed by TBS (2 × 10 min) and cover slipped with Fluoromount™.

[11] 314w For lesion analysis, Cresyl Violet stained sections were imaged with an epifluorescence (Leica DM6000B, camera Leica DFC350 FX) microscope. The maximum lesioned area was calculated as the percentage of damaged tissue using ImageJ software (National Institute of Health, USA). For BDA analysis, a confocal (Leica DMi8 and TCS SP8) microscope was used and 5 spinal cord sections from the C2-C3 block and 5 sections from the C4 block were imaged to quantify the total number of traced CST axons and CST collaterals using imageJ software. The number of descending BDA + CST axons were manually counted and the BDA + CST collaterals in the grey matter were manually outlined in each image. All of the images were aligned with one another and the xy coordinates of the BDA + pixels were extracted using customized ImageJ macros. Then, using a customized R script, the xy coordinates were summed for each group and heat maps were generated using the kde2d function in the R MASS package (R Core Team, 2013). The number of BDA + pixels divided by the number of labelled descending CST axons (for each respective spinal segment) was calculated as a measure of normalized CST sprouting into the grey matter. IBA1 and GFAP stained sections were imaged with an epifluorescence microscope immediately rostral to the injury, at the maximum injury site, and immediately caudal to the injury. 10x magnification was used to image the entire spinal cord cross section and imageJ was used to quantify the IBA1 and GFAP optical density, calculated as the percentage area of positive staining in the selected ROI area. 40x magnification was used to image microglia cells in the ventral grey matter for morphological analysis. 6 representative microglia cells per cross section (3 ipsilesional and 3 contralesional were chosen and the number of endpoints and process length were measured using the imageJ plugin NeurphologyJ (Ho et al., 2011).

[12] 147w Graphpad prism (version 8.0.0 for Mac, GraphPad Software, California USA) was used for statistical analysis. Normality was assessed using the D'Agostino-Pearson omnibus test. For time-course data, a repeated measures two-way ANOVA was used followed by Sidak's multiple comparison test, with a single pooled variance. For the highspeed analysis of reaching and grasping movements, an ordinary twoway ANOVA was used. A parametric unpaired t-test was used for data analyzed at a single time point. 20 animals were excluded based on lack of participation in rehabilitative training. 9 animals were excluded based on deviation of pre-defined lesion size (spared corticospinal tracts and rubrospinal tract). 3 animals died following SCI. Final analysis included 13 animals in the LPS group and 15 animals in the saline group. One rat in the LPS group did not participate in the gap test (made no attempts) and therefore was excluded from that particular analysis.

[1] 290w The single pellet grasping (SPG) enclosure, motorized pellet dispenser and training protocols were used as previously described (Torres-Espín et al., 2018b). First, the rat's preferred paw was established by manually presenting a pellet and recording the number of left and right paw attempts. Once the preferred paw had been established, the pellet dispenser was positioned in a way to enable the rat to only use this paw. A high-intensity dual-window enclosure system was used to train the rats; once the rat had completed an attempt on one side of the enclosure, a pellet was presented on the other side, and so forth. Training consisted of 10 min sessions per rat per day, 5 days a week for 6 weeks before SCI. Ten days following SCI, rats received intraperitoneal injections of LPS/saline (see below for details) and rehabilitative training started 4 days after (14 days post-SCI). After 6 weeks of training the final assessment was conducted. Performance on the SPG task was analyzed once a week from video recordings. At the offset of rehabilitative training, rats were tested in a modified SPG task with a 7 mm wide gap between the pellet and the opening of the enclosure. This set up prevented 'scooping' of the pellet into the mouth, which is a common compensatory strategy. The parameters used to analyze the SPG task (both regular and gap) were the number of attempts the rat made to reach for a pellet and the success rate. Success rate was defined as the number of successful attempts divided by the total number of attempts, expressed as a percentage. An attempt was defined as each time the rat reached for a pellet, and a success as an attempt that resulted in the pellet being eaten.

[2] 99w At the offset of pre-training (baseline) and at the offset of rehabilitative training after SCI, the pattern of movements to successfully grasp and retrieve a pellet was analyzed as previously described (Metz and Whishaw, 2009). Rats were placed in the training enclosure and 3 successful reaching attempts were recorded at high speed (120 fps, Panasonic DMC-FZ200; resolution of 1280 × 720 pixels). These 3 successful attempts were scored and averaged for each animal. This skilled reaching analysis consisted of 11 components each rated from 0 (movement is absent), 0.5 (movement is present but abnormal) to 1 (movement is normal).

[3] 105w Twelve days following CST tracing, rats were euthanized with Sodium Pentobarbital (240 mg/kg) and transcardially perfused with saline containing 0.02 g heparin/l followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate-buffered with 5% sucrose as fixative. Spinal cord and brain tissue were extracted, post-fixed in 4% PFA overnight at 4 • C and cryoprotected in 30% sucrose for 5 days. Spinal cord tissue was cut into a 0.5 cm block above the injury (cervical levels 2-3, C2-C3) and a 0.5 cm block around the lesion site (cervical level 4, C4). Spinal cord blocks were embedded in O.C.T. and frozen in 2-methylbutane at -50 • C.

[4] 25w Spinal cord cross sections were cut at 25 μm on a CryoStar™ NX70 cryostat (Thermo Scientific) and stored at -20 • C until further processing.

[5] 284w Although we found similar effects of inducing inflammation in both the current study and our earlier work (chronic SCI, (Torres-Espín et al., 2018a), there were some important differences in the methods and results. First, a single injection of LPS was used for the subacute time point (10 days post injury), however two injections were given in the chronic setting (8 and 11 weeks post-injury) (Torres-Espín et al., 2018a). Our results suggest that a single injection is sufficient to produce a treatment effect. However, in comparison to chronic administration, LPS given in the subacute time point had a more modest effect on training efficacy. This may be due to the lack of a second injection, the subacute time point being less effective as inflammation is still present, and/or the saline group displaying a more robust recovery due to the earlier training onset. Regardless, inducing inflammation in either the subacute or chronic stages of SCI promoted the restoration of grasping function of the ipsilesional forepaw. In the present study, the beneficial effects of LPS were not associated with significant changes in CST sprouting rostral to injury. When applied chronically after SCI, LPS treated rats displayed increased CST collateral density and further projection of CST collaterals into the cervical grey matter both at the lesion site and one to two spinal segments rostral to the lesion site (Torres-Espín et al., 2018a). Since we did not observe such structural plasticity of the CST fibres, the beneficial effects of subacute LPS treatment on rehabilitative training could otherwise be due to functional plasticity (ex. synaptic plasticity) or modification of other descending motor tracts involved in the control of reaching such as the rubrospinal and reticulospinal tracts (Morris and Whishaw, 2016).

[6] 175w One of the most robust effects of LPS treatment was its ability to produce meaningful, functional recovery of the rat's forepaw. This finding was demonstrated at the end of rehabilitative training when a gap was introduced into the pellet dispenser to discourage compensatory scooping of the pellet. Rats that did not receive LPS performed poorly in this task, displaying a similar success rate as they did immediately following SCI. This suggests that the saline group's improved success rate over the 6 weeks of rehabilitative training was largely due to learning a compensatory strategy. In comparison, rats that received LPS treatment in either the subacute or chronic setting after SCI displayed restorative grasping and supination movements to retrieve the pellet, without relying on compensatory movements (Torres-Espín et al., 2018a). Notably, when applied in the subacute setting, LPS treatment promoted recovery in the cylinder test. Therefore, LPS-induced motor recovery was not task-specific and effectively translated to an untrained task. This finding is significant as it indicates that LPS-induced motor recovery increased injury-induced recovery likely via enhanced neuroplasticity.

[7] 436w It is well recognized that upon binding to the CD14/TLR4/MD2 receptor complex on immune cells in the periphery, LPS triggers an immune response therefore promoting the secretion of nitric oxide, reactive oxygen species and pro-inflammatory cytokines (Lu et al., 2008;Qin et al., 2007). Once LPS-induced inflammation reaches the brain, it initiates a self-propagating process that can last for months after peripheral injection (Qin et al., 2007). This LPS-induced neuroinflammation is characterized by the activation of macrophages and microglia, and the upregulation of a variety of proinflammatory factors such as tumor-necrosis factor alpha, interleukin-1B, nuclear factor kappa B, nitric oxide and cyclooxygenase-2 (Zhao et al., 2019). There is extensive literature on the destructive nature of neuroinflammation, and many immunosuppressive therapies have been shown to be effective for central nervous system disorders and damage including multiple sclerosis, Parkinson's disease, depression, stroke, and SCI (Kohler et al., 2016;Liebigt et al., 2012;Rocha et al., 2015;Thompson et al., 2018;Wells, 2003). However, neuroinflammation and adverse CNS outcomes do not go hand-in-hand, and multiple studies report significant benefits of neuroinflammation (Schwartz et al., 1999b, 1999a, Yong et al., 2019). For example, serial prophylactic injections of LPS can promote a reactive and neuroprotective microglia phenotype that promotes recovery and protects against neuronal loss following SCI (Freria et al., 2020). Furthermore, transplantation of peripheral nerve-activated macrophages into the injured spinal cord has been shown to promote tissue repair and functional recovery (Rapalino et al., 1998;Schwartz et al., 1999b). The reparative effects of macrophages may be dependent upon oncomodulin, as this macrophage-derived protein has been shown to promote regeneration of retinal ganglion cells (Yin et al., 2009). Immune cells other than macrophages have also been implicated in the promotion of CNS repair. Leukocytes and microglia are well known to promote the secretion of a variety of neurotrophic factors that are important for neurogenesis and remyelination (Sousa-Victor et al., 2018;Yong and Rivest, 2009). The role of the LPS receptor, TLR4, which is found on a variety of cell types including dendritic cells, neutrophils, mast cells, macrophages, microglia and neurons, has also been implicated in CNS repair. Antagonizing TLR4 inhibited neurological recovery following intracerebral hemorrhage in a rat model, while another group found that agonizing TLR4 improved Alzheimer's pathology in mice (Lei et al., 2016;Michaud et al., 2013). Histamine may also play a role in the neuromodulator effects of LPS. LPS stimulates an increase in mast cells which can pass the blood brain barrier and release histamine (Silverman et al., 2000;Wang et al., 2020), which has been shown to modulate neuronal and central inflammatory circuits (Coslovich et al., 2018;Dong et al., 2014;Wei et al., 2016;Zhu et al., 2014).

[8] 218w We have previously shown in chronic SCI that LPS enhances the expression of microglia at the lesion site within hours after administration (Torres-Espín et al., 2018a). Although LPS-induced acute microglia and astrocyte activation is well characterized (Ryu et al., 2019), we describe a novel finding in which LPS resulted in a chronic (i. e., measured 8 weeks after injection) attenuation of microglia and astrocyte expression. This result may or may not be particular to SCI, in which immune cells can persist at the lesion site chronically (Beck et al., 2010;Fleming et al., 2006;Sroga et al., 2003). It is possible that exposure to LPS following SCI resulted in a compensatory anti-inflammatory response syndrome, whereby excessive inflammatory stimuli produces an adaptive immune suppression (Adib-Conquy and Cavaillon, 2009;Vergadi et al., 2018). The beneficial effects of LPS for treatment following subacute SCI may therefore be due to the promotion of immune resolution at the lesion site, which has previously been associated with improved recovery following SCI (Francos-Quijorna et al., 2017). The improvements in motor recovery with LPS treatment took time to develop (approximately 4 weeks following injection) in both subacute and chronic applications (Torres-Espín et al., 2018a). This time point may coincide with the resolution of inflammation observed in LPS treated rats, however future work would be required to support this hypothesis.

[9] 475w There is a strong link between inflammation and mental health disorders (Miller and Raison, 2016;Raison et al., 2006). Increasing evidence shows that this holds true after SCI, which causes a drastic posttraumatic immune response, prolonged neuroinflammation, and an increased prevalence of depression and anxiety (Hausmann, 2003;Williams and Murray, 2015). This has been shown to be independent of lesion severity or location, and does not necessarily improve over time after injury (Craig et al., 1994;Dryden et al., 2005). Reducing levels of blood proinflammatory cytokines with a 12 week anti-inflammatory diet is effective in decreasing symptoms of depression after SCI, implicating inflammation as a key factor in the development of mental health disorders in the context of SCI (Allison and Ditor, 2015). Further evidence from preclinical SCI research in rodents has shown an association between anxiety-and depressive-like behaviours and increased levels of inflammation in the brain, spinal cord and blood (do Espírito Santo et al., 2019;Maldonado-Bouchard et al., 2016;Wu et al., 2014). Furthermore, alterations in the intestinal microbiota composition and a leaky gut (which can allow bacterial matter such as LPS to enter the circulation (Fukui, 2016;Liu et al., 2004;Valentini et al., 2014)) have been linked to motor outcome and the development of anxiety-like behaviour following SCI in rodents (Kigerl et al., 2016;Schmidt et al., 2020). Outside of SCI research, inducing inflammation with LPS is commonly used as a model for anxiety-and depressive-like behaviour in rodents (De La Garza, 2005;Yirmiya, 1996). Following intraperitoneal LPS administration, rats display sickness behaviour characterized by decreased motor activity, decreased appetite, and social isolation. This LPS-induced behavioural response is considered acute and transient and is therefore studied within 24 h after injection (Nava and Carta, 2001;Salazar et al., 2012). Indeed, in the present study, rats experienced decreased locomotion in the open field, weight loss and sickness behaviour lasting up to 3 days after LPS injection. More importantly, in our study a single injection of LPS elicited long-term (i.e. 4 weeks after LPS) effects on anxiety-like behaviour in the EPM. It is unclear whether uninjured rats would have a similar long-term increase in anxiety-like behaviour following LPS administration. It is possible that the combination of SCI and LPS acted similar to the two-hit hypothesis suggested for other mental health disorders, where previous immune activation can prime the immune system to be more susceptible to a second adverse event (Feigenson et al., 2014). Nonetheless, LPS-induced anxiety-like behaviour did not interfere with the rat's ability or motivation to participate in rehabilitation training as evidenced in the similar attempt rates between groups. Furthermore, in line with research indicating that LPS causes mechanical allodynia only in male but not female rats (Sorge et al., 2011), our female rats did not experience LPS-induced pain behaviours. Given the sex differences in response to the LPS-induced inflammatory response (Kuo and Fürnsinn, 2016), the present research should be replicated in males.