Abrogated inflammatory response promotes neurogenesis in a murine model of Japanese encephalitis

Abstract

Background: Japanese encephalitis virus (JEV) induces neuroinflammation with typical features of viral encephalitis, including inflammatory cell infiltration, activation of microglia, and neuronal degeneration. The detrimental effects of inflammation on neurogenesis have been reported in various models of acute and chronic inflammation. We investigated whether JEV-induced inflammation has similar adverse effects on neurogenesis and whether those effects can be reversed using an anti-inflammatory compound minocycline.

Methodology/principal findings: Here, using in vitro studies and mouse models, we observed that an acute inflammatory milieu is created in the subventricular neurogenic niche following Japanese encephalitis (JE) and a resultant impairment in neurogenesis occurs, which can be reversed with minocycline treatment. Immunohistological studies showed that proliferating cells were replenished and the population of migrating neuroblasts was restored in the niche following minocycline treatment. In vitro, we checked for the efficacy of minocycline as an anti-inflammatory compound and cytokine bead array showed that production of cyto/chemokines decreased in JEV-activated BV2 cells. Furthermore, mouse neurospheres grown in the conditioned media from JEV-activated microglia exhibit arrest in both proliferation and differentiation of the spheres compared to conditioned media from control microglia. These effects were completely reversed when conditioned media from JEV-activated and minocycline treated microglia was used.

Conclusion/significance: This study provides conclusive evidence that JEV-activated microglia and the resultant inflammatory molecules are anti-proliferative and anti-neurogenic for NSPCs growth and development, and therefore contribute to the viral neuropathogenesis. The role of minocycline in restoring neurogenesis may implicate enhanced neuronal repair and attenuation of the neuropsychiatric sequelae in JE survivors.

Figure 1. Decreased production of proinflammatory cyto/chemokines across various regions of the adult JEV-infected brain by minocycline administration.

Protein isolated from cortex, hippocampus, SVZ and striatum of Control, JEV-infected and JEV+M animals were analyzed using CBA. The graphs depict the fold increase in TNF-α (A), IFN-γ (B), IL-6 (C) and MCP-1 (D) levels in the JEV-infected and JEV+M treated brain areas compared to those from control brain. A profound increase in the cytokine level of TNF-α (A), IFN-γ (B), IL-6 (C), and MCP-1 (D) was observed in all the areas of JEV-infected brain, which was dramatically reduced upon minocycline administration in JEV+M animals. (Values represent means ± SEM from five animals in each group; ^* significant change from JEV, p<0.01).

Figure 2. Attenuation of inflammatory milieu in the infected adult SVZ by minocycline administration.

Cryostat sections of brains from control, JEV and JEV+M animal groups were stained for cells of monocyte/macrophage lineage (CD11b), T-cells (CD3) and neutrophils (Ly6/C), (FITC, green) (A). Images were acquired using oil immersion lens of confocal microscope. These peripheral immune cells were completely absent in the SVZ of control and JEV+M brains. In SVZ of JEV-infected animals, a large number of these cells were observed; scale bar is 50 µ. Protein isolated from SVZ of control, JEV and JEV+M animals were analyzed by immunoblot (B). Increased expression of iNOS and Cox-2 in JEV-infected SVZ was reduced significantly in JEV+M SVZ. The graphs represent densitometric quantification of proteins bands of iNOS and Cox-2 normalised to β-actin (C). Values represent the means ± SEM from five animals in each group (* significant change from control p<0.01; ^# significant change from JEV-infected samples p<0.01).

Figure 3. Replenishment of proliferating cells in the infected adult SVZ by minocycline administration.

Cryostat sections of brains from control, control+M, JEV and JEV+M animal groups were stained for markers of proliferation- Ki-67 (A) and PCNA (B) and developed using DAB substrate. BrdU was also administered to animals for 5 days at 50 mg/kg body weight and then animals were sacrificed 6 h after the last BrdU injection. Cryostat sections from BrdU administered animals from all groups were stained using anti-BrdU antibody (C). Discrete population of cells in control SVZ show localisation of all the antigens. While reduced population of all three cell types were observed in JEV-infected SVZ, however, they were replenished in the SVZ from JEV+M animals. The graphs represent the number of Ki-67 (D), PCNA (E) and BrdU (F) positive cells in the SVZ from the different treatment groups. Cell counting was done using5 serial sections from 3 animals with Leica IM50 software. Values represent means ± SEM from three animals in each group (* significant change from control, p<0.05; ^#significant change from JEV-infected animals, p<0.05); scale bar is 50 µ.

Figure 4. Population of migrating neuroblasts restored in adult infected SVZ by minocycline administration.

Cryostat sections of brains from control, JEV and JEV+M animal groups were labelled for proliferating cells of neuronal lineage (BrdU+DCX double staining) (A) or for migrating neuroblasts (DCX positive cells) (B). Control SVZ show marked co-localization of BrdU (Alexa Fluor, red) and DCX (FITC, green). These double positive cells were distinctly lesser in the JEV-infected SVZ compared to control SVZ, which was however restored in JEV+M SVZ (A). DCX staining (FITC, green) show migrating neuroblasts in control SVZ and in SVZ from JEV+M animals with neurite extending outwards (shown by arrowheads), was completely absent in JEV-infected SVZ (B); scale bar is 50 µ. Protein isolated from SVZ of control, JEV and JEV+M animals were analyzed by immunoblot (C). Expression of DCX, PSD-95 and PCNA declined in JEV-infected SVZ, which in JEV+M SVZ was elevated almost to control levels. The graph represents densitometric quantification of proteins bands of DCX, PSD-95 and PCNA, normalised to β-actin (D). Values represent the means ± SEM from five animals in each group (* significant change from control p<0.05; ^# significant change from JEV samples p<0.05).

Figure 5. Expression of inflammatory molecules in JEV activated BV2 is attenuated by minocycline treatment.

Mouse microglia BV2 were either control (C), JEV-infected (JEV) or JEV-infected and treated with minocycline (JEV+M). After 6 h of minocycline treatment p.i., fresh DMEM-F12 media was added to all conditions. After additional 6 h, CM from BV2 cells from all conditions was collected for CBA and protein extracted for immunoblot. The graphs depict the concentration in pg/ml of cyto/chemokines TNF-α (A), IL-6 (B) and MCP-1 (C) in Control, JEV-infected and JEV+M treated BV2 cells. Values represent means ± SEM from 3 independent experiments performed in duplicate (^* significant change from control, p<0.05; ^# significant change from JEV, p<0.05). Protein isolated from BV2 cells from C, JEV and JEV+M conditions were analyzed by immunoblot (D). The graphs represent the densitometric quantification of protein bands- iNOS and Cox-2 normalised to β-actin (E). Values represent mean ± SEM from 3 independent experiments performed in duplicate (^* significant change from control, p<0.05; ^# significant change from JEV, p<0.05).

Figure 6. Impaired neurosphere formation by soluble mediators from JEV activated microglia is reversed by minocycline treatment.

Single cell suspensions of NSPCs were cultured alone (C) or in presence of BV2-derived CM under 3 conditions- BV2(C), BV2(JEV), BV2(JEV+M) for 3 days. Phase-contrast micrographs of neurospheres were acquired under 10× magnification (A). Scale bar is 100 µ. The number of neurospheres was counted from 10 different fields for each condition and the average diameter of the spheres was measured using Leica IM50 software. The graphs represent the average number of spheres (B) and their average diameter in microns (C). The numbers of colony forming neurospheres do not show any significant difference between groups (B). The reduction in size of neurospheres grown in CM from BV2(JEV) was prominent compared to CM-BV2(C) or CM-BV2(JEV+M) conditions (C). Values represent mean ± SEM from 3 independent experiments performed in duplicate (^** significant change from control, p<0.05; ^* significant change from CM-BV2(C), p<0.01; ^# significant change from CM-BV2(JEV), p<0.01).

Figure 7. Induction of cell cycle arrest by soluble mediators from JEV activated microglia is reversed by minocycline treatment.

Cell cycle analysis of single cell suspensions of neurospheres grown alone or in presence of BV2-CM was performed using RNase/PI buffer and detected by FACS. Gates were set to assess the percentages of G0/G1 (2n DNA), S (>2n DNA) and G2/M (4n DNA) cells using Cell Quest Pro Software, and the representative histograms have been shown (A). Data is representative of 3 independent experiments. Protein isolated from control and BV2-CM treated neurospheres were immunoblotted for checkpoint proteins p21, p27, p53 and p107 (B). The expression of all these proteins increased in neurospheres grown in CM from BV2(JEV) compared to that from BV2(C) or from BV2(JEV+M) (C). Densitometric analysis of immunoblot normalised to β-actin was plotted as a bar graph. Values represent mean ± SEM from three independent experiments. (^** significant change from control, p<0.05; * significant change from CM-BV2(C), p<0.05; ^# significant change from CM-BV2(JEV), p<0.05).

Figure 8. Decreased neuronal differentiation of NSPCs by soluble mediators from JEV activated microglia is reversed by minocycline treatment.

Phase contrast micrographs of differentiating neurospheres from control and BV2 derived CM treated NSPCs 2 days after induction of differentiation on PDL-coated plates (A). Impaired migration and differentiation of cells from the periphery of the spheres is observed in CM-BV2(JEV) conditions. Control and BV2-Cm treated neurospheres were dissociated and single cell suspensions of NSPCs were differentiated for 2 days. The cells were fixed and double immunocytochemistry for β-III tubulin (neuronal phenotype) and GFAP (astrocytic phenotype) was performed and mounted with Vectashield containing DAPI. Distinct β-III tubulin positive cells (FITC, green) and GFAP positive cells (Alexa 594, red) were observed, indicating NSPCs undergoing differentiation (B). The percentage of β-III tubulin positive cells were calculated from 5 different fields and plotted as a representative graph (C). RNA was isolated from these NSPCs after 2 days of differentiation and qRT-PCR was performed for β-III tubulin expression. The graph represents relative mRNA expression of β-III tubulin normalised to the expression of constitutive 18S rRNA (D). Values represent mean ± SEM from three independent experiments performed in duplicate. (^** significant change from control, p<0.05; ^* significant change from CM-BV2(C), p<0.01; ^# significant change from CM-BV2(JEV), p<0.01). Scale bar corresponds to 50 microns.