Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Oct 15.
Published in final edited form as: Brain Res. 2019 Jul 8;1721:146328. doi: 10.1016/j.brainres.2019.146328

Elimination of microglia in mouse spinal cord alters the retrograde CNS plasticity observed following peripheral axon injury

Jessika M Hutchinson 1,2, Lori G Isaacson 1,2,*
PMCID: PMC6875753  NIHMSID: NIHMS1534900  PMID: 31295468

Abstract

Following the transection of peripherally located sympathetic preganglionic axons of the cervical sympathetic trunk (CST), transient retrograde neuronal and glial responses occur in the intermediolateral cell column (IML) of the spinal cord, the location of the parent neuronal cell bodies. The role of microglia in this central response to peripheral axon injury was examined in mice fed the PLX5622 diet containing colony-stimulating factor-1 receptor (CSF-1R) inhibitor for 28 days, which eliminated approximately 90% of spinal cord microglia. Microglia elimination did not impact baseline neurotransmitter expression in the IML neurons, and the typical neuronal plasticity observed following CST transection was unaffected. Oligodendrocyte precursor cells (OPCs) were significantly increased at one week post injury in the IML of mice fed the control diet, with no change in mature oligodendrocytes (OLs). Following microglia elimination, the baseline population of OPCs in the IML was increased, suggesting increased OPC proliferation. Injury in the microglia depleted mice resulted in no additional increase in OPCs. Though baseline astrocyte activation and GFAP protein expression were unaffected, microglia elimination led to increased activation and GFAP protein post injury when compared with mice fed the control diet. These results reveal that microglia regulate the baseline OPC population in the uninjured spinal cord and that activated microglia influence the activities of OL lineage cells as well as astrocytes. The regulatory roles of microglia observed in this study likely contribute to the long term survival of the IML neurons observed following the distal axon injury.

Keywords: colony-stimulating factor-1 receptor (CSF-1R) inhibitor, oligodendrocyte precursor cells (OPCs), oligodendrocytes (OLs), astrocytes, glial fibrillary acidic protein (GFAP), adenomatous polyposis coli clone CC1 (CC1)

Graphical Abstract

graphic file with name nihms-1534900-f0008.jpg

1. INTRODUCTION

Following the transection of preganglionic axons in the cervical sympathetic trunk (CST), retrograde influences arising from the peripheral injury site lead to changes in the parent neuronal cell bodies in the intermediolateral cell column (IML) of the spinal cord as well as plasticity in nearby glial cell populations. At one week post injury, neuronal expression of choline acetyltransferase (ChAT) is decreased, and the injured parent neurons express activating transcription factor-3 (ATF3), a marker for injured neurons (Coulibaly et al., 2013). In addition, robust glial cell plasticity is observed, which includes the aggregation and activation of microglia in the vicinity of the parent preganglionic cell bodies in the IML (Coulibaly and Isaacson, 2012a), astrogliosis (Coulibaly and Isaacson 2012a), and changes in mature oligodendrocytes (OLs; Coulibaly and Isaacson 2012a), as well as oligodendrocyte progenitor cells (OPCs; Coulibaly and Isaacson, 2012b; Isaacson et al., 2017).

Activated microglia can affect OL lineage cells and astrocytes following injury. Signals released by activated microglia can result in astrocyte activation, and it is possible that astrocytes are dependent on the early signaling cascades of microglia following injury (Holm et al., 2012; Sofroniew et al., 2014; Liddelow et al., 2017). Moreover, activated microglia may affect OL lineage cells and play a role in the differentiation of OPCs into mature OLs (Agresti et al., 1996). Therefore, understanding the delicate balance of the complicated inter-communication between glial cells could lead to therapeutic tools to prevent axonal and neuronal death in the case of injury.

The overall objective of this study was to understand the role of microglia in influencing the neuronal and glial plasticity typically observed in the spinal cord following CST transection. In order to accomplish this, microglia were eliminated using an inhibitor to colony-stimulating factor-1 receptor (CSF-1R; PLX5622; Plexxikon, Inc; Berkeley, CA), which specifically targets microglia (Elmore et al., 2014). The CSF-1R inhibitor blocks the binding of CSF-1, a ligand in the platelet-derived growth factor family, to CSF-1R, a tyrosine kinase receptor specific to monocytes and peripheral macrophages (Hamilton et al., 2013) and has been successfully used to eliminate microglia in brain (Acharya et al., 2016) and spinal cord (Lee et al., 2018).

The specific goals of this study were to: 1) determine whether the PLX5622 diet successfully eliminated microglia in the spinal cord at rest as well as following CST transection, and 2) understand how microglia elimination affected typical neuronal and glial plasticity following CST transection. Because microglia are key players in central nervous system (CNS) immune regulation, and OLs and astrocytes can respond to microglia-derived cytokines (Liu et al., 2011; Peferoen et al., 2014 ; Norden et al., 2016), we hypothesized that glial cell plasticity would be affected by the absence of microglia. The findings of this study further the understanding of cell-cell communication in the microenvironment in hopes that it could lead to therapies promoting neuron repair following injury.

2. RESULTS

2.1. Daily food consumption and body weight were unaffected by PLX5622 diet

Mice were placed on either a control diet (Cont; provided by Plexxikon Inc.) or the PLX5622 diet (PLX; provided by Plexxikon Inc.) for a total of 28 days. Mice in the Cont and PLX groups received either a sham surgery (Sham) or CST transection (Inj) on day 21 of the diet and the animals survived for an additional week. Food consumption was monitored for the 28 days of the experiment, and body weight was monitored for 1 week following surgery. A slight drop in food consumption was observed following administration of the diet (Days 2-3) and following the day of surgery (Day 22), possibly as a side effect of recovery, but no significant differences in daily food consumption were observed across the four treatments (Sham+Cont; Inj+Cont; Sham+PLX; Inj+PLX; Fig. 1.A.). Mice were weighed on the day of the surgery and then weighed daily during the week following surgery. No significant differences in weight were observed across the four treatment groups (Fig. 1.B.).

Figure 1: Food consumption and body weight were unaffected by the PLX5622 diet.

Figure 1:

Open in a new tab

Mice were fed control diet and underwent sham surgery (Sham+Cont) or CST transection (Inj+Cont) at Day 21 and survived one week post-surgery, or mice were fed PLX5622 diet and underwent sham surge ry (Sham+PLX) or CST transection (Inj+PLX) at Day 21 and survived one week postsurgery. A. Daily food consumption over the 28-day period across treatment groups in young adult (10-12 weeks) female C57BL/6 mice used for microscopy. A slight dip in food consumption was observed following the initial administration of the diet (Days 2-3) and following the day of surgery (Day 22). Overall, no change in food consumption was observed. One way ANOVA (F(3,115)=0.08, p=0.969). Significance reported at p < 0.05. B. Mice were weighed on the day of surgery (Day 0) and then daily during the one week post-surgery period. Weight was unaffected across treatments. One way ANOVA (F(3,31)=1.77, p=0.176). Significance reported at p < 0.05.

2.2. PLX5622 eliminated the majority of microglia in uninjured and injured spinal cord

In the Sham+Cont group, spinal cord microglia expressed ionized calcium binding adaptor protein molecule 1 (Iba1), a marker for microglia, and displayed an inactivated ramified appearance, with a small soma and thin processes that extended into the microenvironment. These cells were evenly distributed across the IML and the adjacent lateral funiculus (LF), location of the axons of the IML cell bodies (Fig. 2.A.). In line with previous findings from our lab (Coulibaly and Isaacson, 2012a), we observed a robust microglia response to CST transection in the IML of the spinal cord in animals fed the control diet (Inj+Cont). Microglia exhibited the ameboid-like morphology typically associated with activation and numerous microglia aggregated around the injured neuronal cell bodies of the IML (Fig. 2.B.). When quantified, the number of Iba1 positive cells in the IML was significantly increased by 114% following injury. In contrast to the control group, the sham group fed the PLX5622 diet (Sham+PLX) exhibited a near complete (~90%) elimination of microglia (compared with Sham+Cont) in both the IML and LF (Fig. 2.C.) with the few microglia remaining exhibiting an enlarged size (Fig. 2.C). An almost complete elimination of microglia also was observed in the Inj+PLX group (Fig. 2.D.). ATF3 immunolabeling, which was absent in the Sham+Cont and Sham+PLX groups, was expressed in the nucleus of the injured neuronal cell bodies in both the control and PLX groups following CST transection (Inj+Cont, Inj+PLX) and was used to locate the injured neuronal cell bodies in the IML (Figs. 2.B., 2.D.).

Figure 2: PLX5622 diet eliminated most microglia in the uninjured mouse spinal cord as well as following injury.

Figure 2:

Open in a new tab

A.-D. Merged confocal micrographs of mouse spinal cord show microglia (green, white arrows), activating transcription factor 3 (ATF3; red, red arrows), and DAPI (blue) in the intermediolateral cell column (IML) and lateral funiculus (LF) from mice fed the control diet and then receiving sham surgery (A.; Sham+Cont) or one week CST transection (B.; Inj+Cont) or mice fed the PLX5622 diet and receiving sham surgery (C.; Sham+PLX) or CST transection (D.; Inj+PLX). A. Note the homogenous population of microglia present in the IML and LF of the Sham+Cont group and lack of ATF3 neurons. B. At one week following injury the typical microglia activation and ATF3 expression were present in the Inj+Cont group. Note two ATF3 positive neurons surrounded by microglia processes (yellow asterisks). C.-D. Microglia were absent in both groups receiving the PLX diet, but ATF3 neurons were present in the Inj+PLX group. Insets (A.’-A.” - D’-D.”) show individual green and red channels so that ATF3 neurons can be clearly visualized in the injury treatments. A.-D., Scale = 100μm. Insets, Scale = 100μm. E. Quantification of Ibal immunolabeling revealed a significant increase in Ibal cells in Inj+Cont (compared with respective sham), but few microglia were present in either of the PLX treatments (Sham+PLX, Inj+PLX). One way ANOVA (F(3,47)=97.37, p=0.000). *, significance at p < 0.05.

2.3. Elimination of microglia did not affect neuronal plasticity

The effects of microglia elimination on neuronal plasticity typically observed in the IML following CST transection were examined using antibodies to ATF3 and ChAT. At one week following injury, ATF3 positive cells were significantly increased compared to their respective sham treatments (Figs. 3.B., 3.D.). This increase was similar in both injury treatments (Inj+Cont, Inj+PLX; Fig. 3.E.), suggesting that elimination of microglia did not affect the ATF3 expression typically observed at one week following injury.

Figure 3: Elimination of microglia had no effect on typical neuronal plasticity in the IML at one week following injury.

Figure 3:

Open in a new tab

A.-D. Confocal micrographs of mouse spinal cord from mice fed the control diet and then receiving sham surgery (A.-A.”; Sham+Cont) or one week CST transection (B.-B.”; Inj+Cont) or from mice fed the PLX diet and receiving sham surgery (C.-C.”; Sham+PLX) or CST transection (D.-D.”; Inj+PLX). A.-A.” ChAT positive neurons (ChAT; green, green arrows) were observed in IML, but no neurons expressing ATF3 (red) were present in Sham+Cont mice. B.-B.” Following CST transection, ChAT positive neurons were decreased in the IML and ATF3 positive neurons (red arrows) were present. C.-C.” The number of ChAT positive neurons appeared similar in Sham+PLX mice compared to the Sham+Cont group. D.-D.” ChAT positive neurons were decreased in Inj+PLX mice (similar to Inj+Cont), and numerous ATF3 positive neurons were present. Scale = 100μm. E. The typical decrease in neuronal ChAT expression following CST transection was unaffected by the elimination of microglia. One way ANOVA (F(3,42)=75.39, p=0.000). F. Elimination of microglia had no effect on the expression of ATF3 by injured neurons. One way ANOVA (F(3,208)=236.18, p=0.000). *, significance at p < 0.05.

Previously we reported a significant decrease in ChAT expression by injured rat IML neurons at one week following CST transection (Coulibaly et al., 2013). This decrease also was observed in the present study in both Inj+Cont and Inj+PLX groups (Figs. 3.B., 3.D.). When quantified, no difference in the number of ChAT positive neurons was observed, suggesting that microglia elimination did not affect the typical decrease in ChAT expression in the injured neurons at one week following injury (Fig. 3.E.).

2.4. Elimination of microglia altered typical OPC plasticity in the spinal cord

To determine whether microglia elimination affected OPCs, this population was investigated using neural-glial antigen 2 (NG2) immunolabeling. Colocalizing NG2 with oligodendrocyte transcription factor 2 (Olig2), which labels all cells in the OL lineage (Ligon et al., 2004), aided in identification and was used to distinguish OPCs from pericytes (Schallek et al., 2013) and macrophages (Tripathi and McTigue, 2007) which can also label with NG2. Pericytes that labeled with NG2 were distinguished from OPCs based on their morphology, which appeared in a crescent like shape, lacking processes of any kind, as well as their lack of colocalization with Olig2. OPCs expressed NG2 in the cell soma, as well as throughout the ramified processes extending out into the nearby microenvironment, while Olig2 labeled nuclei of all cells in the OL lineage. The morphology of OPCs appeared consistent across the treatment groups (Fig. 4).

Figure 4: Elimination of microglia increased baseline oligodendrocyte progenitor cell (OPC) population in the IML.

Figure 4:

Open in a new tab

A.-D. Confocal micrographs of mouse spinal cord from mice fed the control diet and then receiving sham surgery (A.-A.’; Sham+Cont) or one week CST transection (B.-B.’; Inj+Cont), or mice fed the PLX5622 diet and receiving sham surgery (C.-C.’; Sham+PLX) or CST transection (D.-D.’; Inj+PLX). A.-A.’ Spinal cord from Sham+Cont mice show a homogenous population of OPCs (NG2, red, white arrows) distributed throughout the IML. All OPCs included in the analysis colocalized with O1ig2 (green), confirming their OL lineage (yellow arrows). B.-B.’ As observed previously OPCs were increased in the Inj+Cont mice. C.-C.’ and D.-D.’ OPCs were present in the IML and LF of spinal cord from Sham+PLX and Inj+PLX treatments. Boxed area in C.’ depicts OPC morphology at high magnification. A.-D., Scales = 100μm. E. OPCs were increased post injury in the IML and LF of the control group compared with shams (Inj+Cont vs Sham+Cont). The baseline population of OPCs was increased in the IML when microglia were eliminated (Sham+Cont vs Sham+PLX). No additional increase was observed post injury in the IML of the microglia depleted mice (OPCs in IML; Inj+PLX vs Sham+PLX). However, similar to the control group, OPCs in the LF were increased post injury in the PLX treatment (Sham+PLX vs Inj+PLX). One way ANOVA (F(3,47)=6.23, p=0.001). *, significance at p <0.05.

Microglial elimination in the treated, uninjured mice (Sham+PLX) resulted in an increase in OPCs in the IML, but not the LF, when compared to untreated, uninjured mice (Sham+Cont; Fig. 4.E), indicating that OPC proliferation was increased at the location of the neuronal cell bodies in the IML when microglia were absent (Fig. 4.E.). In mice that were fed the control diet and received injury, the number of OPCs was increased in both the IML and LF (Sham+Cont vs Inj+Cont; Fig. 4.E.). This finding is consistent with that previously reported in the rat (Coulibaly and Isaacson, 2012b). The OPC increase that was observed in the IML following CST transection in the untreated mice (Inj+Cont) was no longer present following microglial elimination (Inj+PLX), but the OPC increase in the LF remained significant (Fig 4.E.).

2.5. Elimination of microglia increased the number of mature OLs following injury

Mature OLs were labeled with a monoclonal antibody anti-adenomatous polyposis coli clone CC1 and are referred to here as CC1 cells. CC1 positive cells exhibited a ring of CC1 immunoreactivity around the nucleus of the cells and colocalized with cells having O1ig2 nuclear labeling (Fig. 5). No change in the number of CC1 cells was observed following CST transection in the untreated groups (Sham+Cont vs Inj+Cont; Fig. 5.E.). The number of CC1 positive cells was increased in the LF in the treated injured mice compared to the treated sham group (Inj+PLX vs Sham+PLX; Fig. 5.E.). The number of O1ig2 cells was examined to assess any changes in the entire population of cells in the OL lineage (Fig. 5.F.). O1ig2 positive cells also were increased in the LF following injury in the Inj+PLX group (compared to Sham+PLX; Fig. 5.F.).

Figure 5: Elimination of microglia increased mature oligodendrocytes (OLs) in the LF at one week post injury.

Figure 5:

Open in a new tab

A.-D. Confocal micrographs of mouse spinal cord from mice fed the control diet and then receiving sham surgery (A.-A.’; Sham+Cont) or one week CST transection (B.-B.’; Inj+Cont), or mice fed the PLX5622 diet and receiving sham surgery (C.-C.’; Sham+PLX) or CST transection (D.-D.’; Inj+PLX). A.-A.’ Mature OLs (CC1, red, red arrows) expressed O1ig2 (green, yellow arrows), a marker for all cells in the OL lineage, in Sham+Cont mice. B.-B.’ Elimination of microglia had no obvious effect on the expression of CC1 or O1ig2 in Sham+PLX (compared with Sham+Cont). C.-C.’ No change in the number of mature OLs or O1ig2 was observed in the Inj+Cont group. D.-D.’ Elimination of microglia resulted in an increase in mature OLs present following CST transection (compared with Sham+PLX). A.-D., Scale = 100μm. E. Quantification of CC1 cells revealed no change post injury in the control groups (Inj+Cont vs Sham+Cont). However, CC1 cells were increased in the LF post injury when microglia were eliminated (Inj+PLX vs Sham+PLX). One way ANOVA (F(3,40)=2.48, p=0.077). F. Quantification of O1ig2 cells revealed that the number of OL lineage cells was increased in the LF following injury in the microglia depleted mice (Inj+PLX vs Sham+PLX). One way ANOVA (F(3,40)=2.45, p=0.079). *, significance atp < 0.05.

2.6. Elimination of microglia increased astrocyte activation and GFAP protein post injury

The effects of microglia elimination on typical astrocyte plasticity in the IML were examined using confocal microscopy. No changes were observed when astrocyte activation was compared between the sham groups, suggesting that the elimination of microglia did not induce astrocyte activation in the uninjured animal in either the IML or LF (Figs. 6.A., 6.C.). At one week following CST transection astrocyte activation in the IML was observed in both the Inj+Cont and Inj+PLX groups compared to respective shams (Figs. 6.B., 6.D.), and the elimination of microglia did not appear to impact this typical activation (Inj+Cont vs Inj+PLX; Fig. 6.E.). However, though no astrocyte activation was observed in the LF of the control group post injury, activation was present in the LF of the PLX group (Sham+PLX vs Inj+PLX), suggesting an astrocyte response post injury in the LF in the absence of activated microglia.

Figure 6: Eliminating microglia resulted in increased astrocyte activation in the LF following injury.

Figure 6:

Open in a new tab

A.-D. Merged confocal micrographs of mouse spinal cord show GFAP (green, white arrows), ATF3 (red, red arrows), and DAPI (blue) in the IML and LF from mice fed the control diet and then receiving sham surgery (A.; Sham+Cont) or one week CST transection (B.; Inj+Cont), or mice fed the PLX5622 diet and receiving sham surgery (C.; Sham+PLX) or CST transection (D.; Inj+PLX). Note activation of GFAP astrocytes (larger soma, thickened processes and aggregation) in the IML following injury in both control (Inj+Cont) and PLX (Inj+PLX) treatments and in the LF in the PLX treatment (Inj+PLX). Note ATF3 positive neurons surrounded by astrocyte processes (yellow asterisks). A.’-A.” - D’-D.” Insets show single green and red channels so that ATF3 immunolabeling in the injury treatments is visualized. A.-D., Scales = 100μm. E. Quantification of GFAP immunolabeling in the IML revealed significant increases following injury in both Inj+Cont and Inj+PLX treatments (compared with respective shams). GFAP was significantly increased in the LF post injury when microglia were eliminated (compared to Sham+PLX). One way ANOVA (F(3,47)=5.35, p=0.003). *, significance at p < 0.05.

GFAP western analysis of biopsy cores taken from the IML and adjacent LF revealed no change in baseline GFAP protein (Sham+PLX vs Sham+Cont; Fig. 7), and no increase was detected in GFAP protein following CST transection in the control group. However, GFAP protein was increased by ~37% post injury in the PLX group (compared to the respective sham; Sham+PLX vs Inj+PLX; Fig. 7.B.). In addition, GFAP was significantly increased post injury in the PLX group compared to controls (Inj+Cont vs Inj+PLX), indicating a more robust response by astrocytes in the absence of activated microglia.

Figure 7: Eliminating microglia resulted in increased GFAP in the spinal cord following injury.

Figure 7:

Open in a new tab

A. Western blot analysis of GFAP protein (55 kDa) from spinal cord cores of the IML and LF taken from mice fed the control diet and receiving sham surgery (Sham+Cont) or CST transection (Inj+Cont), or mice fed the PLX5622 diet and receiving sham surgery (Sham+PLX) or CST transection (Inj+PLX). Blots were re-probed for GAPDH (36 kDa) as a loading control. 5μg protein loaded in each lane. All blots shown originated from the same membrane. B. Semi-quantitative analysis revealed no change in the control group following injury (Sham+Cont vs Inj+Cont). However, GFAP protein was increased by ~29% in the Inj+PLX group (compared to Sham+PLX) and by ~37% compared to Inj+Cont. One way ANOVA (F(3,31)=5.5, p=0.004); *, significance at p < 0.05.

3. DISCUSSION

3.1. Microglia play important roles in regulating other cells in the spinal cord

Our results confirmed previous studies in the rat (Coulibaly and Isaacson 2012a, 2012b; Coulibaly et al., 2013) that transection of the CST in the periphery resulted in a robust activation of glial cells in the IML of the mouse spinal cord near the vicinity of the injured cell bodies. In mice fed the control diet, microglia in the IML were activated at one week following injury, characterize d by increased Iba1 expression, cellular aggregation, and amoeboid-like morphology. The administration of the CSF-1R inhibitor diet was successful in eliminating ~90% of the microglia from the spinal cord of young adult C57BL/6 mice in shams as well as following CST transection. Because microglia-derived cytokines can modulate the immune response in the CNS (Cammer et al., 2010; Probert et al., 2015), we hypothesized that microglia elimination would impact the delicate balance and cross-talk that exists between neurons and glial cell populations in the spinal cord.

Indeed, although there were no changes in neuronal plasticity observed in the current study, OL lineage cells as well as astrocytes were affected by the elimination of microglia. Importantly, the loss of the microglia population in the spinal cord affected the baseline population of OPCs, and post injury the mature OL population was increased in the microglia depleted mice. While the elimination of microglia did not affect baseline GFAP or astrocyte activation, a more robust astrocyte activation and GFAP protein expression were observed following injury when microglia were eliminated.

3.2. Microglia depend on CSF-1R signaling in the mouse spinal cord

Following the inhibition of CSF-1R, which successfully blocks the binding of CSF-1, most microglia were eliminated from the spinal cord. Our results demonstrate that microglia are dependent on CSF-1R signaling for survival and are consistent with the results of others (Elmore et al., 2014; Valdearcos et al., 2014). The unique dependence of microglia, but not peripheral macrophages (Valdearcos et al., 2014), on CSF-1R signaling highlights the value of this approach as an experimental tool for studying microglia.

CSF-1 is a ligand of the CSF-1 receptor and is key in the maintenance of the microglia population. Though the origin of CSF-1 is still debated (Luo et al., 2013), recent studies suggest that CSF-1 is localized to neurons (Guan et al., 2016), but also can be produced by endothelial cells, monocytes/macrophages, fibroblasts, and bone marrow stromal cells (Luo et al., 2013; Chitu et al., 2016). Furthermore, the expression of CSF-1 is exacerbated following brain injury. The upregulation of CSF-1 in neurons following injury may prevent neurodegeneration and inflammation (Luo et al., 2013; Guan et al., 2016), providing evidence that CSF-1R signaling aids in the support and survival of neurons following injury, and may serve a protective function (Luo et al., 2013).

Following the administration of the CSF-1R inhibitor, few microglia remained in the spinal cord. Interestingly, microglia still present exhibited an enlarged size. This unique morphology may be attributed to a highly phagocytotic state, given so few microglia left to clear cellular debris, a main function of the microglia population (Elmore et al., 2014), or alternatively the hypertrophy might be characteristic of cells undergoing apoptosis (Haschek et al., 2010).

3.3. Microglia do not play a role in neuronal plasticity following CST transection

It was unknown whether the presence of microglia influenced the neuronal plasticity typically observed at one week following CST transection. To investigate this question, both ATF3 and ChAT were used as markers for neuronal plasticity in the IML following injury. ATF3 was upregulated in IML neurons at one week following injury in our injury model (Coulibaly et al., 2013), and we found a similar ATF3 upregulation in the absence of microglia, suggesting that microglia do not influence the upregulation of neuronal transcriptional factors such as ATF3.

The neuronal expression of ChAT is normally downregulated in the IML following CST transection (Coulibaly et al., 2013) as well as in other models of axon injury (Caldero et al., 1992; Borke et al., 1993). Decreased levels of ChAT can be attributed to poor neurotransmission in the injured neurons (Chang et al., 2004), and reduced ChAT expression suggests that the injured neurons have shifted from a functional state to a regenerative one (Frizell and Sjostran, 1974; Navarro et al., 2007). Loss of microglia did not affect the typical plasticity in ChAT expression, suggesting that microglia do not play a role in regulating ChAT expression in the injured neurons. Findings of the current study suggest that any effects of microglia elimination occur downstream of the neuronal plasticity. One scenario would involve the injured neurons first activating microglia and then the activated microglia in turn influencing other glial cells in the microenvironment of the injured neurons.

3.4. Microglia regulate the OL lineage cell population

The results of the present study revealed changes in OL lineage cells in the absence of microglia. Because NG2 cells in the spinal cord are evenly distributed throughout the grey matter and white matter (Horner et al., 2003), we examined both the IML, the location of the injured neuron cell bodies and the LF, the location of the injured axons as they exit the spinal cord. In microglia depleted mice the number of NG2 cells colocalizing Olig2 was increased specifically in the IML. These findings suggest that microglia can influence the baseline OPC population in the IML, possibly by regulating typical OPC proliferation in the spinal cord. At baseline, ~3% of the NG2 pool is undergoing cell division (Horner et al., 2003), and the loss of microglia may have increased this baseline proliferation rate. In support of this theory, others have shown that signals released from nonreactive microglia can attenuate OPC proliferation without triggering OPC death (Taylor et al., 2010). Interestingly, no such changes were observed in the LF, so this regulatory role for microglia seems specific to NG2 cells in the IML near the neuron cell bodies. Because we quantified OPCs only in the IML, whether the NG2 increase occurred in other grey matter areas of the spinal cord is unknown.

Similar to previous studies in our lab (Isaacson et al, 2017), NG2 cells were increased in mice fed the control diet at one week following CST transection. The number of NG2 cells was not increased following injury when microglia were absent. Because the baseline OPC population was already increased without injury when microglia were absent, it is possible that no further increase was possible following the injury due to a ‘ceiling’ effect. Alternatively, loss of microglia may have slowed the rate of OPC proliferation and therefore an increase was not yet detectable at the one week post injury time point. However it is also possible that proliferation occurred at a faster rate and thus took place earlier in the survival period. The latter explanation could result in earlier OPC differentiation into mature OLs.

Both CC1 cells and O1ig2 cells were increased in the LF following injury when microglia were depleted. However, CC1 and O1ig2 cells in the microglia depleted treatment were not increased compared to the untreated injury group (Inj+PLX vs Inj+Cont), and so the increase in the PLX treatment (Inj+PLX vs Sham+PLX) represents a trend that needs further investigation. It is possible that, without microglia present to regulate OPC differentiation, the typical OPC increase in the IML occurred earlier in the one week post injury period, and that newly formed OPCs migrated to the LF and matured into OLs by the one week post injury time point.

Both extrinsic factors and intrinsic factors contribute to OPC differentiation, and it is likely that a combination work together to push OPCs into a more mature state. Microglia release a number of factors and cytokines following injury, some of which correlate with a decrease in OL differentiation. Cammer et al. (2000) showed that OPC differentiation and maturation are slowed in a time-dependent manner with the incubation of tumor necrosis factor alpha (TNF-α) along with the absence of growth factors in rat brain cultures. Consistent with this finding, inflammatory profiling revealed a robust decrease in inflammatory genes such as TNF-α in the brains of mice fed the PLX diet (Elmore et al., 2014). Furthermore, interferon gamma, a cytokine released following microglia activation, can block OL differentiation (Agresti et al., 1996). Removal of this cytokine resulted in a reversal of this effect, and OPC proliferation and differentiation returned to normal (Agresti et al., 1996).

3.5. Role for microglia in astrocyte plasticity following CST transection

Due to the complex and intricate morphology exhibited by astrocytes, it was not possible to confidently obtain counts of these cells. Therefore, two different approaches, confocal microscopy of astrocyte activation and western analysis of GFAP protein expression, were taken to assess changes in the astrocyte population. These analyses revealed that resting microglia do not affect baseline astrocyte activation state or baseline GFAP expression in the IML or the LF. In addition, the typical astrocyte activation state observed following injury in the IML was unaffected when microglia were absent, suggesting that activated microglia do not directly impact the morphological changes that occur during astrocyte activation in the IML. This would suggest that the injured neurons (and not the activated microglia) play an important role in activating astrocytes in the IML following peripheral axon injury. In support of this idea, injured neurons can directly induce the upregulation of cytokines such as TNF-α to promote astrocyte activation, without having to first signal through the microglia (Probert et al., 2015). Though astrocyte activation state in the IML was similar between the two injury groups, the microglia depleted group showed robust activation in the LF post injury, while no activation was observed in the control group. If increased damage and demyelination in the LF indeed occurred in the microglia depleted group (as described above to explain increased mature OLs), one might expect enhanced astrocyte activation in the LF. This could also explain the robust increase in GFAP protein expression post injury when microglia were absent, and suggests that activated microglia may act to minimize damage to axons in the LF following CST transection.

The lack of any change in GFAP expression following CST transection in the control injury group was surprising when astrocytes were clearly activated when viewing with confocal microscopy. It is possible that changes in GFAP protein were below the level of detection in the western analysis, and a more sensitive assay such as ELISA might better detect subtle changes in GFAP protein.

3.6. Conclusions

The goal of the current study was to determine whether eliminating microglia using the PLX5622 diet influenced the CNS neuronal and glial plasticity typically observed in the mouse spinal cord following peripheral axon injury. In our model, eliminating the majority of microglia did not alter neuronal plasticity, leading to the conclusion that activated microglia do not typically play a role in the downregulation of ChAT or the upregulation of ATF3 that occurs in the injured neuron cell bodies in IML.

Our findings suggest that microglia do play important roles in regulating both OL lineage cells and astrocytes in the spinal cord. Though ‘resting’ microglia do not influence astrocyte activation state or baseline levels of GFAP in the uninjured animal, non-reactive microglia appear to play a critical role in regulating baseline OPC proliferation in the IML, the location of neuronal cell bodies. Following injury, activated microglia seem to influence the axonal response to CST transection. Indeed, our data indicate a trend where activated microglia post injury may regulate OPC differentiation and also suggest a scenario where activated microglia affect the level of astrocyte activation within the LF.

These results suggest an intricate balance exists between microglia and nearby neuronal and glial populations. Indeed, intercellular communication in the microenvironment is likely important for the survival of the IML neurons following CST transection. Although more work needs to be done to elucidate the role of microglia, findings from this research provide insight for future therapies to target microglia in hopes to promote neuronal survival and recovery following injury.

4. EXPERIMENTAL PROCEDURES

4.1. Animals and PLX5622 diet

A total of 50 female C57BL/6 mice (18-25 gm), approximately 8-10 weeks of age, were used in this study. Four treatment groups were used for microscopic analyses: Sham+Cont (n=6), Inj+Cont (n=6), Sham+PLX (n=6), Inj+PLX (n=6). For western blot analysis, this paradigm was repeated in a second set of 26 mice: Sham+Cont (n=6), Inj+Cont (n=6), Sham+PLX (n=7), Inj+PLX (n=7). Food was aliquoted (167 gm) and distributed on day 1 and day 14 of the diet plan. Mice in similar treatments were housed in pairs in the Miami University Animal Facilities on a 12:12 light dark cycle. Each pair of mice received similar aliquots of food to ensure that all cages received similar amounts of food over the four-week period. Animals had access to the food ad libitum and food consumption was weighed on a daily basis. All experiments were approved by the Miami University Institutional Animal Care and Use Committee Protocol 941.

4.2. Surgical procedures

After consuming either PLX5622 or control diet for a total of 21 days, mice were anesthetized with 2.5% isoflurane and an incision was made on the ventral side of the neck. In the sham treatments, the right SCG was exposed but the incoming preganglionic axons were not transected. This procedure was repeated on the left side. In the injury treatments, the right SCG was exposed and the CST was transected as it entered the SCG and this was repeated on the left side. Immediately following surgery, animals were given an intraperitoneal injection of the analgesic Buprenorphine (0.07cc; 0.3mg/cc). One quarter of a Rimadyl tablet (analgesic; 1.25 mg; BioServ) was provided to each animal daily for two days following the surgery. Either the PLX5622 or control diet continued to be administered for a total of seven days following surgery until the time of sacrifice.

4.3. Tissue collection for microscopy and western blot analysis

For immunohistochemistry, mice were anesthetized (Beuthansia-D, 80mg/kg) and perfused with 20ml of 0.9% saline followed by 50 ml of fixative containing 4% paraformaldehyde in 0.1M phosphate buffer (PB). The spinal cord was removed and post-fixed for 1 hr in perfusion fixative and then stored in 0.1M PB at 4°C.

Segments C7-T3 of the spinal cord, identified by counting the cord rootlets, were blocked, infiltrated in 30% sucrose in 0.1M PB for 48-72 hrs and sectioned in series along the coronal plane (18μm), such that one section was mounted onto each of 8 slides (A-H). The mounting was repeated starting with Slide A and continued, so that the entire segment was represented on each slide.

To obtain fresh frozen tissues, animals were decapitated using a Harvard guillotine and the spinal cord was removed, snap-frozen with liquid nitrogen, and stored at −80°C. Prior to western analysis, the C7-T3 levels were blocked from the cord and individual segments were positioned vertically on a freezing stage. Biopsy punches (0.5 mm; World Precision Instruments) that contained the IML and a portion of the adjacent LF were collected from each the right and left sides of each segment and stored at −80°C until ready for use.

4.4. Immunohistochemistry procedure

Sections mounted on slides were incubated overnight in 0.1M phosphate buffer saline (PBS) containing 0.06% Triton-X and 0.1% normal donkey serum (NDS). The sections then were incubated in 0.06% Triton-X and 0.07% NDS in 0.1M PBS for 30 min and then for 48 hrs in primary antibody diluted in 0.1M PBS at 4°C. Combinations of antibodies were used: goat anti-Iba1 (1:200, Abcam), mouse anti-adenomatous polyposis coli (CC1; 1:200, Abcam), chick anti-glial fibrillary acidic protein (GFAP; 1:1000, Abcam), rabbit anti-ATF3 (1:150, Santa Cruz Biotech), goat anti-ChAT (1:200, Millipore Sigma), rabbit anti-NG2 (1:200, Millipore Sigma), goat anti-O1ig2 (1:200, R&D Systems). Sections were rinsed four times (5 min each) with 0.1M PBS, incubated for two hrs in AlexaFluor conjugated secondary antibodies (Jackson Immunoresearch) directed toward the host of the primary antibody, rinsed in 0.1M PBS and then in 0.1M PB. Sections were cover slipped with Vectashield mounting medium with DAPI (Vector Labs). Slides were viewed using a Zeiss 710 confocal laser scanning microscope.

4.5. Assessment of activation in the IML of the spinal cord following CST transection

In order to assess whether PLX5622 successfully eliminated microglia, the spinal cord was examined using an antibody directed against Ibal to label microglia, and an antibody directed against ATF3, which is expressed exclusively in injured neurons of the IML (Coulibaly et al., 2013) to aid in locating the injured neuronal cell bodies. Four levels of spinal cord, C7-T2, were screened for the most robust microglia activation characterized by aggregation, shortened microglial processes, and a larger soma giving the cell an amoeboid like morphology. For the treatment groups that received the PLX5622 diet, the level with the most ATF3 positive neurons in the IML (since microglia were eliminated) was deemed to be most activated and used for all further analyses. Two confocal Z-stacks (3μm intervals; ~4 scans per stack; one from left and one from right side) of the IML were collected using the 20x objective. Images were blinded and the number of Ibal positive cells and ATF3 neurons was counted from maximum intensity projection images using the manual tag tool on ImagePro 6.3 software. Data were compared across the four treatments using a one-way ANOVA, followed by a Fisher post-hoc comparison test. Significance was reported at p < 0.05. Once the level with the most activation was determined for each animal, sections from this level were utilized for all further analyses.

4.6. Assessment of neuronal plasticity

For the analysis of neuronal plasticity, ChAT positive and/or ATF3 positive neuronal cell bodies were counted from each section on the slide while conducting on-site viewing with the confocal microscope. This approach captured the extent of response in the spinal cord segment. Data were compared across treatment groups using a one-way ANOVA followed by a Fisher post-hoc comparison test. Significance was reported at p < 0.05.

4.7. Assessment of glial plasticity using microscopy

For the assessment of glial cell plasticity, two confocal Z-stacks (3μm intervals; ~4 scans per stack; one from left and one from right side) of the IML were collected using the 20x objective. Images were blinded prior to quantitative analysis. To analyze changes in OPCs and mature OLs, the number of cells exhibiting colocalized expression of NG2 and O1ig2 or CC1 and O1ig2 respectively was assessed in the IML and LF. To determine changes in astrocyte activation, confocal images of the IML and adjacent LF immunolabeled with GFAP and ATF3 (to determine level of activation) were blinded and ranked by three independent evaluators based on thickness of astrocyte processes and abundance of the GFAP expression in the IML or LF. Images were unblinded and the three rankings for each image were averaged. The average was expressed as a percent of sham controls. Data were compared across treatments using a one-way ANOVA followed by a Fisher post-hoc comparison test. Significance was reported at p < 0.05.

4.8. Western blot procedure

For western blot analysis, protein was isolated from spinal cord cores by sonicating tissue in 0.01M Tris-HCl buffer (pH 7.4) with 1% sodium dodecyl phosphate and 1% protease inhibitor. Protein concentration for each case was determined using the BCA protein analysis kit (Pierce Biotech) and samples were prepared according to Laemmli (Laemmli, 1970). Protein was loaded (5μg). separated using SDS-PAGE (5% stacking/12% resolving), and transferred to PVDF membrane (0.45μm pore size) in a 4°C chamber for 16 hrs at a constant 150mA and 500V. Following transfer, membrane was briefly rinsed in 100% methanol, dried for 15 min and then cut down the center of the blue standard to separate the proteins of interest from the ladders. Membranes were rewet in 100% methanol for 10 min, and blocked for 4 hrs in 4% nonfat dry milk in Tris-buffered saline with Tween (TBST). The proteins of interest were incubated overnight at 4°C in rabbit anti-GFAP (Abcam, 1:50,000), rinsed four times with TBST, and incubated in goat anti-rabbit HRP IgG (2 hrs; Millipore, 1:100,000). Membrane containing the ladder was incubated in TBST, rinsed and incubated in streptactin-HRP (1:300,000; Bio-Rad). Membranes were then rinsed four times (5 min each) in TBST and submerged in Supersignal West Pico Chemiluminescent Substrate for five min and developed using CL-X Posure X-ray film. Membranes were stripped and re-probed for GAPDH (1:160,000; Fitzgerald Industries), incubated in goat anti-mouse HRP IgG (1:100,000, Millipore Sigma), rinsed, and submerged in Supersignal West Pico Chemiluminescent Substrate for five min and developed using CL-X Posure X-ray film. Films were scanned and band intensities were quantified with ImageQuant 5.2. Each GFAP band was expressed as a ratio to its corresponding GAPDH band. Ratios were compared across treatments using one-way ANOVA followed by a Fisher post-hoc comparison. Significance was reported at p < 0.05.

Highlights.

  • CSF-1R inhibitor eliminated ~90% microglia in the mouse spinal cord

  • Neuronal plasticity in cord following peripheral injury was unaffected by microglia loss

  • Baseline OPCs in uninjured spinal cord were increased when microglia were absent

  • Baseline GFAP protein and astrocyte activation state in spinal cord unaffected by microglia loss

  • GFAP protein and astrocyte activation were increased post injury when microglia were absent

ACKNOWLEDGEMENTS

This study was supported by the National Institutes of Health (NS R15NS095314) awarded to LGI. The confocal usage was supported by NSF DBI-0821211 awarded to the Center for Advanced Microscopic Imaging, Miami University. We thank Matt Duley and Richard Edelmann for their assistance with confocal microscopy and image analysis. We are grateful to Plexxikon, Inc. for providing the PLX5622 chow and the control chow used in this study.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  1. Acharya M, Green KN, Allen BD, Najafi AR, Syage A, Minasyan H, Limoli CL, 2016. Elimination of microglia improves cognitive function following cranial irradiation. Scientific Reports 6, 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Agresti C, D’Urso D, Levi G, 1996. Reversible inhibitory effects of interferon-γ and tumor necrosis factor-α on oligodendroglial lineage cell proliferation and differentiation in vitro. European Journal of Neuroscience 8(6), 1106–1116. [DOI] [PubMed] [Google Scholar]
  3. Borke RC, Curtis M, Ginsberg C, 1993. Choline acetyltransferase and calcitonin gene-related peptide immunoreactivity in motoneurons after different types of nerve injury. Journal of Neurocytology 22(3), 141–53. [DOI] [PubMed] [Google Scholar]
  4. Caldero J, Casanovas A, Sorribas A, Esquerda JE, 1992. Calcitonin gene-related peptide in rat spinal cord motoneurons: subcellular distribution and changes induced by axotomy. Neuroscience 48, 449–461. [DOI] [PubMed] [Google Scholar]
  5. Cammer W, 2000. Effects of TNFα on immature and mature oligodendrocytes and their progenitors in vitro. Brain Research 864(2), 213–19. [DOI] [PubMed] [Google Scholar]
  6. Chang EL, Wei E, Tseng C, Lue J, Wen C, Shieh J, 2004. Differential expression of calcitonin gene-related peptide (CGRP) and choline acetyltransferase (ChAT) in the axotomized motoneurons of normoxic and hypoxic rats. Journal of Chemical Neuroanatomy 28(4), 239–51. [DOI] [PubMed] [Google Scholar]
  7. Chitu V, Gokhan S, Nandi S, Mehler M, Stanley R, 2016. Emerging roles for CSF-1 receptor and its ligands in the nervous system. Trends in Neurosciences 39(6), 378–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Coulibaly AP, Gannon SM, Hawk K, Walsh BF, Isaacson LG, 2013. Transection of preganglionic axons leads to CNS neuronal plasticity followed by survival and target reinnervation. Autonomic Neuroscience 179(1-2), 49–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Coulibaly AP, Isaacson LG, 2012a. Transient changes in spinal cord glial cells following transection of preganglionic sympathetic axons. Autonomic Neuroscience 168(1–2), 32–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Coulibaly AP, Isaacson LG, 2012b. Changes in mature and progenitor oligodendrocyte populations in the spinal cord following distal preganglionic axon injury. Program No. 234.15. 2012 Neuroscience Meeting Planner New Orleans, LA: Society for Neuroscience, 2012. Online. [Google Scholar]
  11. Elmore MRP, Najafi AR, Koike MA, Dagher NN, Spangenberg EE, Rice RA, Green KN, 2014. CSF1 receptor signaling is necessary for microglia viability, which unmasks a cell that rapidly repopulates the microglia-depleted adult brain. Neuron 82(2), 380–397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Frizell M, Sjostrand J, 1974. The axonal transport of slowly migrating [3H] leucine labelled proteins and the regeneration rate in regenerating hypoglossal and vagus nerves of the rabbit. Brain Research 81(2), 267–83. [DOI] [PubMed] [Google Scholar]
  13. Guan Z, Kuhn J, Wang X, Colquitt B, Solorzano C, Vaman S, Guan A, Evans-Reinsch Z, Braz J, Devor M, Abboud-Werner S, Lanier L, Lomvardas S, 2016. Injured sensory neuron-derived CSF1 induces microglial proliferation and DAP12-dependent pain. Nature Neuroscience 19(1), 94–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hamilton JA, Achuthan A, 2013. Colony stimulating factors and myeloid cell biology in health and disease. Trends in Immunology 34(2), 81–89. [DOI] [PubMed] [Google Scholar]
  15. Haschek WM, Rousseaux CG, Wallig MA, 2010. Chapter 2 - Manifestations of toxic cell injury: cell injury/death and chemical carcinogenesis Fundamentals of Toxicologic Pathology (Second Edition), San Diego: Academic Press, 9–42. [Google Scholar]
  16. Holm TH, Draeby D, Owens T, 2012. Microglia are required for astroglial toll-like receptor 4 response and for optimal TLR2 and TLR3 response. Glia 60, 630–38. [DOI] [PubMed] [Google Scholar]
  17. Horner PJ, Thallmair M, Gage FH, 2002. Defining the NG2-expressing cell of the adult CNS. Journal of Neurocytology 31(6-7), 469–480. [DOI] [PubMed] [Google Scholar]
  18. Isaacson LG, Henkel ND, Schneider LJ, Sullivan D, Kolias A, Maloney J, Hutchinson JM, 2017. Glial proliferation in rat and mouse spinal cord in response to peripheral axon injury. Presentation #43.14. Society for Neuroscience Meeting Planner, Washington DC. [Google Scholar]
  19. Laemmli UK, 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 (5259), 680–685. [DOI] [PubMed] [Google Scholar]
  20. Lee S, Shi XQ, Fan A, West B, Zhang J, 2018. Targeting macrophage and microglia activation with colony stimulating factor 1 receptor inhibitor is an effective strategy to treat injury-triggered neuropathic pain. Molecular Pain 14, 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Liddelow SA, Guttenplan KA, Clarke LE, Bennett FC, Bohlen CJ, Schirmer L, Barres BA, 2017. Neurotoxic reactive astrocytes are induced by activated microglia. Nature 541(7638), 481–487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Ligon K, Alberta J, Kho A, Weiss J, Kwaan M, Nutt C, Louis D, Stiles C, Rowitch D, 2004. The oligodendroglial lineage marker OLIG2 is universally expressed in diffuse gliomas. Journal of Neuropathology & Experimental Neurology 63(5), 499–509. [DOI] [PubMed] [Google Scholar]
  23. Liu Z, Fan Y, Won SJ, Neumann M, Hu D, Zhou L, Liu J, 2007. Chronic treatment with minocycline preserves adult new neurons and reduces functional impairment after focal cerebral ischemia. Stroke 38(1), 146–152. [DOI] [PubMed] [Google Scholar]
  24. Luo J, Elwood F, Britschgi M, Villeda S, Zhang H, Ding Z, Wyss-Coray T, 2013. Colony stimulating factor 1 receptor (CSF1R) signaling in injured neurons facilitates protection and survival. Journal of Experimental Medicine 210(1), 157–172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Navarro X, Vivo M, Valero-Cabre A, 2007. Neural plasticity after peripheral nerve injury and regeneration. Progress in Neurobiology. 82(4):163–201. [DOI] [PubMed] [Google Scholar]
  26. Norden DM, Trojanowski PJ, Villanueva E, Navarro E, Godbout JP, 2016. Sequential activation of microglia and astrocyte cytokine expression precedes increased Iba-1 or GFAP immunoreactivity following systemic immune challenge. Glia 64(2), 300–316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Peferoen L, Kipp M, Van der Valk P, Van Noort JM, Amor S, 2014. Oligodendrocyte microglia cross-talk in the central nervous system. Immunology. 141(3), 302–313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Probert L, 2015. TNF and its receptors in the CNS: the essential, the desirable and the deleterious effects. Neuroscience 302, 2–22. [DOI] [PubMed] [Google Scholar]
  29. Schallek J, Geng Y, Nguyen H, Williams D, 2013. Morphology and topography of retinal pericytes in the living mouse retina using in vivo adaptive optics imaging and ex vivo characterization. Investigative Ophthalmology & Visual Science 54(13), 8237–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Schreiner B, Romanelli E, Liberski P, Ingold-Heppner B, Sobottka-Brillout B, Hartwig T, Chandrasekar V, Johannssen H, Zeilhofer H, Aguzzi A, Heppner F, Kerschensteiner M, Becher B, 2015. Astrocyte depletion impairs redox homeostasis and triggers neuronal loss in the adult CNS. Cell Reports 12(9), 1377–84. [DOI] [PubMed] [Google Scholar]
  31. Sofroniew MV, 2014. Multiple roles for astrocytes as effectors of cytokines and inflammatory mediators. The Neuroscientist 20(2), 160–172. [DOI] [PubMed] [Google Scholar]
  32. Taylor DL, Pirianov G, Holland S, McGinty CL, Norman AL, Reali C, Diemel LT, Gveric D, Yeung D, Mehmet H, 2010. Attenuation of proliferation in oligodendrocyte precursor cells by activated microglia. Journal of Neuroscience Research. 88(8), 1632–44. [DOI] [PubMed] [Google Scholar]
  33. Tripathi R, McTigue DM, 2007. Prominent oligodendrocyte genesis along the border of spinal contusion lesion. Glia. 55(7), 698–711. [DOI] [PubMed] [Google Scholar]
  34. Valdearcos M, Robblee M, Benjamin D, Nomura D, Xu A, Koliwad SK, 2014. Microglia dictate the impact of saturated fat consumption on hypothalamic inflammation and neuronal function. Cell Reports. 9(6), 2124–38. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES