Immunoglobulin G (IgG) attenuates neuroinflammation and improves neurobehavioral recovery after cervical spinal cord injury

This study used 135 female Wistar rats (250 to 300 g) from Charles River and followed the animal-use protocol approved by the Animal Use Committee (AUC) of the University Health Network (#979) for all procedures. All rats underwent a C7-T2 laminectomy and were then randomly and blindly assigned to three groups (sham, saline, and IgG). A 35-g modified aneurysm clip was applied extradurally to animals in the saline and IgG groups at the level C7-T1 for 60 s to cause a moderate to severe spinal cord injury [29, 30]. The spinal cords of sham animals were not injured. Animals were given 1 mL of buprenorphine (0.05 mg/kg) and 5 mL of saline immediately and subsequently every 12 h for 3 days after surgery.

IgG (I4506) from pooled human serum was purchased from Sigma Aldrich. After dissolving 500 mg of IgG in 4 mL of sterile saline solution (0.9% NaCl, Baxter), a single bolus of IgG at a dosage of 0.4 g/kg was injected intravenously through the tail-vein at 15 min after SCI. Control animals were injected intravenously through the tail-vein with 1 mL of sterile saline (0.9% NaCl). The dosage of 0.4 g/kg was chosen based on the reported dosage by Gok et al. and based on the clinical dosages that are currently used to treat Guillain-Barré syndrome and other autoimmune disorders [25, 28]. There was no noticeable difference in the general behavior and appearance of animals that received either IgG or saline. There was also no difference in mortality rate between saline and IgG-treated animals. All assessors were blinded to the treatment groups throughout the study.

Following SCI, animals were sacrificed at 4 h for multiplex enzyme-linked immunosorbant assays (ELISA) and at 24 h for western blot and myeloperoxidase (MPO) activity analyses. Each animal was overdosed with pentobarbital and perfused with 180 mL of iced-cold saline (0.9% NaCl). The spinal cord was isolated in ice-cold Ringer’s solution and the meninges were removed. A 0.5 cm length of the spinal cord centered at the injury epicenter was dissected, immediately frozen with liquid nitrogen and crushed with a frozen mortar and pestle. The tissue was then stored at −80°C until use.

Animals were sacrificed at 24 h after SCI for immunohistochemical analysis and at 6 weeks after SCI for histological analysis. Each animal was overdosed with pentobarbital and then perfused with 60 mL of ice-cold phosphate-buffered saline (PBS) and 180 mL of para-formaldehyde (4% w/v in PBS, pH 7.4). The spinal cord of each animal was dissected and post-fixed overnight with 10% sucrose 4% para-formaldehyde PBS solution. The spinal cord was then cryo-protected in 20% sucrose PBS solution. The spinal cord tissue (1.0 cm centering on the injury epicenter) was embedded in tissue-tek and stored at −80°C. Spinal cords embedded into tissue-tek were cryosectioned at 20 μm thickness. Tissue sections were mounted on glass slides and stored at −80°C.

All statistical analyses were completed using SigmaStat Software. One-way analysis of variance (ANOVA) followed by a Bonferroni post-hoc test was used to determine the difference between treatments with regard to the biochemical, immunohistochemical, molecular and electrophysiology analyses. A two-way ANOVA followed by Bonferroni post-hoc test was used to determine significant differences between treatments in BBB and inclined plane tests over several time points. Since MMP-9 expression was not detectable in the sham group, the Student t-test was used to test the difference between saline and IgG-treated animals. All data are reported as mean ± SEM. Differences were considered significant at P <0.05.

Fluorescence microscopy was used to show that IgG was present in the spinal cord of the injured rat 24 h following injury (Figure 1A). IgG (shown in red) could be seen caudal and rostral to the injury site and in the white and grey matter. The fluorescence signal associated with IgG’s presence was very weak in the non-injured spinal cord with IgG injection (Figure 1A, right panel). This is supported with confocal microscopy (Figure 1B). Images were acquired using a 60X objective at approximately 1000 μm caudal to the injury site. At 24 h after injection, IgG was found outside the vasculature (marked by RECA-1 positive staining) in the spinal cords of injured animals. In contrast, non-injured animals with IgG injection contained no parenchymal IgG (Figure 1B, right panel). This suggests that SCI enabled IgG to cross the blood-spinal cord barrier (BSCB).

Microglia and astrocytes are two important CNS cell types involved in the orchestration of the inflammatory response [37, 38]. Having demonstrated that IgG appears to cross the BSCB at 24 h after SCI, we assessed whether IgG would co-localize with microglia and astrocytes using confocal microscopy. Iba-1 is a marker for microglia/macrophages. Since monocytes do not infiltrate the injured spinal cord until 2 to 3 days after SCI, most of the Iba-1 positive cells are likely microglia. IgG (red), although in the same vicinity, did not co-localize with Iba-1 (green) positive cells in IgG-injected animals (Figure 1C). Importantly, IgG was not observed in saline injected animals even though the secondary antibody specific to human IgG was applied to these sections (Figure 1C, lower panel).

Astrocytes were stained with anti-glial fibrillary acid protein (GFAP). We observed co-localization between GFAP (green) and IgG (red) in animals with IgG treatment (Figure 1D). Again, there was no IgG signal in saline injected animals, as expected (Figure 1D, bottom panel). This observation suggests potential interaction between IgG and astrocytes in IgG-treated animals.

Neutrophils, which represent an important component of the innate immune response, infiltrate the spinal cord within hours of injury, and maximal cell numbers are reached at 24 h after SCI [7]. Neutrophils can increase the extent of tissue injury after SCI by producing reactive oxygen species (ROS) and reactive nitrogen species (RNS) [41‐43] that can damage proteins, DNA, and lipids. Neutrophils can further increase the extent of the inflammatory response by producing pro-inflammatory mediators such as TNF-α, IL-1β, and IL-8 [44]. MMP-9 secreted by neutrophils can degrade the collagen matrix of the blood-spinal cord barrier and increase leukocyte infiltration [45, 46]. Blockade of neutrophil infiltration using various strategies has been demonstrated to reduce oxidative damage and to improve neurobehavioral recovery in several rat models of SCI [17, 18, 39, 40]. In the current study, we found that intravenous IgG treatment was associated with a significant reduction in MPO activity at 24 h after SCI (Figure 2A). MPO is an enzyme produced by neutrophils and stored in the azurophilic granules, and as such, MPO activity was used as an indirect measure of neutrophil presence in the injured spinal cord [47]. Our finding is consistent with a previous report by Gok et al., in which IgG was associated with a significant reduction in MPO activity when it was injected intraperitoneally immediately following SCI [28]. We also observed a significant reduction in neutrophil infiltration to support the MPO data (Figure 2B and C).

Immediately after SCI, a rapid increase in the levels of TNF-α, IL-1β, and IL-6 is observed in the spinal cords of humans and animal models of SCI [7, 48‐50]. It has been suggested that these cytokines exacerbate the initial damage and contribute to neurodegeneration following CNS injury [51]. Following CNS injury, pro-inflammatory cytokines play an important role in activating and recruiting leukocytes including neutrophils, monocytes, and lymphocytes. Pro-inflammatory cytokines promote leukocyte infiltration to the site of injury by increasing the expression of vascular cell adhesion molecules on the surface of endothelial cells [52, 53] and by regulating the expression and secretion of chemokines [38, 54, 55]. IL-6 has previously been shown to regulate the production of IL-8 and MCP-1 expression in a mouse model of skin pouch inflammation [54]. Furthermore, IL-1β has previously been shown to regulate the production of CINC-1 and MCP-1 in a mouse model of SCI via the IL-1R/MyD88 axis, and astrocytes were shown to be the primary cellular source of the aforementioned cytokines [38].

IgG treatment was associated with a significant reduction in IL-1β and IL-6 levels in the injured spinal cord. This finding may explain why neutrophil infiltration is attenuated upon intravenous administration of IgG after SCI. In fact, previous studies in SCI have shown that reductions in the expression level or disruption in the signaling pathway of either TNF-α, IL-1β, or IL-6 is associated with a significant reduction in neutrophil infiltration in SCI models as well as other models of acute inflammation [38, 54, 56, 57]. Along the same lines, if TNF-α or IL-1β is injected into a non-injured spinal cord or IL-6 is over-expressed in an injured spinal cord, then leukocyte infiltration has been shown to be enhanced [58‐60]. Our findings suggest the observed reduction in neutrophil infiltration may be due to IgG’s mechanism of action in reducing the level of pro-inflammatory cytokines in SCI. In our future experiments, using our established experimental paradigms, we could employ antibodies to block Fc receptors as an approach to elucidating the immuno-modulatory mechanism of IgG.

We observed that IgG treatment was associated with a significant reduction in MCP-1 levels but not CINC-1. In addition to astrocytes, endothelial cells have been shown to co-localize with CINC-1 after SCI in the perivascular region [18]. Currently, there is insufficient scientific evidence describing the primary cellular source of these chemokines and the signal transduction pathways that regulate the production of these chemokines. The lack of strong supporting evidence in the literature presents major challenges in outlining the relationship between IL-1β, IL-6, MCP-1, and CINC-1 in animal models of SCI. The heterogeneity in species and injury mechanisms of animal models of SCI make it difficult to draw correlations between published findings and the observations stemming from the present body of work. Future experiments examining the cellular targets of IgG and the cellular sources of cytokines and chemokines in rats will illuminate the complex interaction and regulation of cytokine and chemokine expression.

Although qualitative in nature, our immunohistochemistry experiments using astrocyte and microglia as specific markers provide valuable insights into the cellular target(s) of IgG in SCI. We observed that IgG only co-localized with astrocytes and is associated with significant reduction in the level of IL-1β and IL-6. Previous studies on cellular sources of pro-inflammatory cytokines might explain why IgG had only a modest attenuation in the level of TNF-α after SCI. Microglia/macrophages and astrocytes are the major cellular sources of TNF-α, IL-1β, and IL-6 immediately after SCI [37]. The level of IL-1β but not TNF-α is maintained by neutrophils and astrocytes at later time points after SCI [37, 48]. Although determining the exact cellular target and the underlying mechanism of IgG is beyond the scope of this study, the selective reduction of cytokine expression and the specific astrocytic co-localization of IgG suggest some degree of cellular specificity in IgG’s mechanism of action. Although we did not observe co-localizations between IgG and Iba-1 positive cells in this study, we cannot rule out the potential interaction between microglia and IgG in the acute phase as well as the sub-acute phase of SCI. The underlying mechanism and the molecular target of IgG on astrocytes and/or microglia can be examined in future studies.

Previous studies have reported that attenuation of IL-6 signaling [61, 62], blocking neutrophil infiltration [15, 16, 39, 40, 63], or reducing MMP-9 expression [46] separately after SCI results in tissue preservation and improvement in hind-limb locomotor function. In this study, we show that IgG treatment is associated with a decrease in the area of scar and cavity (Figure 5A), an increase in tissue sparing (Figure 5B) and a significant improvement in neurobehavioral recovery (Figure 6A and B). Importantly, spinal cord axonal conduction changes (Figure 6C) were in agreement with our neurobehavioral outcomes. The results suggest that axonal conduction is severely impaired following injury and IgG treatment provides some degree of axonal preservation or improvement in axonal signaling. This is in agreement with the histological, molecular and behavioral data. Although the magnitude of tissue preservation in the IgG-treated animals may appear modest compared to the saline control group, it has previously been demonstrated that only 10% of axons are needed to achieve significant functional recovery after SCI [30]. Our tissue preservation and neurobehavioral findings are consistent with findings by Gris and colleagues [15], in which they attenuated leukocyte infiltration with anti-CD11d antibodies and observed modest levels of tissue sparing and a two-point improvement in the BBB score after moderate to severe SCI. In conjunction with other groups’ findings, our data reinforce the detrimental role of the acute inflammatory response in SCI. Our results also support the underlying rationale of immuno-modulatory therapy in attenuating the effects of the acute inflammatory response after SCI.