Activated Microglia Mediate the Motor Neuron-, Synaptic Denervation- and Muscle Wasting-Changes in Burn Injured Mice

Abstract

Background: Muscle wasting (MW) of burn injury (BI) remains unresolved. Microglia-mediated inflammatory cytokine/chemokine release, motor neuron loss (MNL) and MW is observed after BI but connection of the central changes to synaptic-denervation and MW is unelucidated. Stimulation of microglia α7acetylcholine receptors (α7AChRs), a Chrna7gene-product, exhibits anti-inflammatory properties (decreased cytokine/chemokines). Hypothesis tested was that exploitation of the microglia α7AChR anti-inflammatory properties mitigates cytokine inflammatory responses, MNL, synaptic-denervation and MW of BI.

Methods: Wild-type or α7AChR knockout (A7KO) mice received 30% body BI or Sham BI (SB) under anaesthesia with and without selective α7AChR agonist, GTS-21. Lumbar spinal cord tissue and hindlimb muscles were harvested. Immunohistochemistry, TUNEL assay for apoptosis and/or Nissl staining were used to examine microglia (Iba1 staining), MNL (NeuN staining) and synapse morphology (synaptophysin for nerve and α-bungarotoxin for muscle α7AChR). Spinal cytokine/chemokine transcripts, inflammatory transducer-protein expression and tibialis, soleus and gastrocnemius muscle weights were measured.

Results: BI to Wild-type mice caused significant microgliosis (5.8-fold increase, p < 0.001) and upregulated TNF-α, IL-1β, CXCL2, MCP-1 transcripts, and inflammatory transducer-protein (STAT3 and NF-κB, p < 0.01) expression together with increased transcripts of iNOS (p < 0.01) and CD86 (p < 0.01) at day 14 reflective of inflammatory M1 microglia phenotype. Significant apoptosis, MNL (32.2% reduction, p < 0.05), increased spinal caspase-3 expression (> 1100-fold, p < 0.05) and synaptic denervation were observed with BI. The tibialis muscle endplates (synapse) of SB had a smooth pretzel shaped appearance with good apposition of presynaptic nerve to postsynaptic muscle. In BI mice, the normal pretzel-like appearance was lost, and the endplates were fragmented with less nerve to muscle apposition. Tibialis, soleus, and gastrocnemius mass were decreased 31.7% (p < 0.01), 23.4% (p < 0.01) and 27.5% (p < 0.01) relative to SB. The A7KO mice with SB showed significant MNL loss (61.5% reduction, p < 0.05), which was aggravated with BI, accompanied by significantly higher expression of STAT3 and Nf-kB (p < 0.05). GTS-21 ameliorated the spinal expression of above enumerated cytokines/chemokines, inflammatory transducer-proteins (p < 0.05) together with mitigated MNL (p < 0.05), synaptic denervation (p < 0.05) and decreased MW of tibialis (25%), gastrocnemius (15%) and soleus (20%) relative to untreated wild type BI mice (p < 0.01). GTS-21 beneficial effects were absent in the A7KO mice.

Conclusions: Microglia-mediated inflammatory responses play pivotal role in MNL as decrease of inflammatory responses improved MNL; α7AChR stimulation also mitigated synaptic denervation and MW changes of BI. α7AChRs have a role in spinal homeostasis even in uninjured state.

Keywords: burn injury; inflammatory cytokines/chemokines; microglia activation; muscle wasting; neuronal apoptosis.

FIGURE 1

BI leads to microgliosis together with neuronal apoptosis and decreased neuronal numbers in the ventral horn of spinal cord. a: Confocal images (20X magnification) of triple immunofluorescence staining of L3–4 segments of spinal cord ventral horn with microglia‐marker, Iba1 (red), neuron‐marker, NeuN (green), nuclear‐marker, DAPI (blue) at 14 days after sham‐burn (SB) or burn injury (BI). Motor neurons decreased in numbers and were smaller in size together with prominent proliferation of microglia (microgliosis) at day 14 after BI (inset of dotted white box and magnified at 63x bottom of Figure 1a). Compared to SB, the microglia (red) in BI appeared juxtaposed to the motor neurons (white arrow in magnified image below Figure 1a). b: The numbers of neurons were significantly (p < 0.001) decreased in the ventral horn area (200x magnification) at day 14 after BI as compared to SB (Figure 1b). (Analysis on the motor neuron numbers used a threshold of 25 μm or higher as a motor neuron). The microgliosis was analysed by intensity of the staining of Iba1 from 5 sections/mouse, 4–6 animals/group. Compared to SB, microgliosis was significant (p < 0.001) c: In situ TUNEL apoptosis assay (magenta) together with counter‐staining with neuronal maker (NeuN, green) and nuclei (DAPI, blue) of the ventral horn area was performed. As compared to SB, spinal cord of BI showed prominent increased TUNEL positive nuclei in the neurons (magenta). White arrows point to positive magenta staining contained within neurons in the merged Figure 1c. d: Immunostaining against activated caspase‐3 (magenta) in the ventral horn area is shown with counterstaining against neurons (NeuN, green) and nuclei (DAPI, blue). The merged figure shows overlap of neurons and capase‐3 (Figure 1d). BI group showed significantly increased level of activated caspase‐3 staining. e: RT‐PCR against caspase‐3/GAPDH showed increased level of capase‐3 transcripts in the ventral horn of the spinal cord in the BI mice as compared to SB at day 7 and 14. **p < 0.05 vs. SB group. ns: not statistically significant.

FIGURE 2

BI leads to inflammatory chemokines cytokines, and M1 macrophage phenotype in the ventral horn of spinal cord. a: Quantitative PCR mRNA expression of chemokines (CXCL2 and MCP1) and cytokines (IL‐1β, TNFα) were measured and normalised to GAPDH. The mRNA expression of pro‐inflammatory and chemokine molecules, IL‐1β (p = 0.0003, p < 0.0001), TNF‐α (p = 0.0057, p < 0.0001), and CXCL2 (p < 0.0001, p < 0.0001), MCP‐1 (p < 0.0001, p = 0.0023), was significantly increased at day 7 and/or 14 post‐injury in BI relative to SB (p‐values in the parenthesis in the order of day 7 and 14, respectively). b: Immunoblots of lumbar spinal cord (L3‐L4 segments) isolated from BI and SB mice at day 7 and 14 post‐injury probed for STAT3 and NFκB and normalised to GAPDH as loading control and fold change of the immunoblot was determined using Image J to measure band intensities of target protein to GAPDH (left panel). Immunoblot was quantified using Image J to measure band intensities of target proteins (phosphorylated vs. total molecule, right panels). The phosphorylated STAT3 (p < 0.0001, p < 0.0001) at day 7 and 14 and phosphorylated NFκB at day 14 (ns: p = , p < 0.0001), markers of upstream inflammatory signal transduction proteins, were significantly upregulated in BI. c: Lumber spinal cord samples (L3‐L4 segments) were cryo‐sectioned for the analysis of microglia phenotypes by immunofluorescent staining against the pro‐inflammatory marker, CD86 (classical M1 subtype, yellow staining in the left panel with white arrows pointing to microglia co‐stained by Iba1 and CD86). Compared to SB, BI showed increased numbers of CD86 positive cells; some of them colocalized with Iba1 positive microglia (indicated by the white arrows of merged Figure 2 c). d: RT‐PCR for transcripts of pro‐inflammatory microglia subtypes (iNOS and CD86 for classical M1 phenotype) were performed on the ventral horn of lumber spinal cord and shown by normalising to the internal control, GAPDH. BI increased iNOS at day 7 (p = 0.0003) and day 14 (p < 0.0001) and CD86 (p < 0.0001) at day 14. The transcripts increased in a time‐dependent manner at 7 towards day 14 (p = 0.024, 0.0012, and 0.0003 for iNOS, and CD86, respectively). Data are means ± SD; n = 5–8, *p < 0.05, **p < 0.01.

FIGURE 3

Burn injury induced neuromuscular synapse disintegration and muscle wasting. a: Tibialis anterior muscles were stained with α‐BTX (red) to detect acetylcholine receptors at the post‐synaptic membrane of the muscle. Presynaptic nerve terminal neurons were stained by anti‐synaptophysin antibody (green). The endplates in the SB group had a smooth and pretzel shaped appearance. The merged image shows perfect apposition of the pre‐and post‐synaptic nerve and muscle, respectively, suggesting normal undisturbed innervation of the synapse in SB. In the BI mice, the normal pretzel‐like shaped appearance was lost, and the endplate appeared fragmented with irregular edges. The presynaptic nerve component (green) was partially lost in BI, leading to an imperfect apposition to muscle (pointed by white arrows), suggesting partial denervation in BI. (Scale bar = 30 μm). b: The changes in muscle weight in SB, and BI group mice at day 14. Dry muscle weights of TA, SOL, and GC muscle) are expressed as a ratio to body weight pre‐BI or SB. The body weight at day 0 was used in view of body weight loss induced by BI at day 14. Those histograms show that tibialis anterior, gastrocnemius and soleus muscles masses were decreased by 31.7%, 23.4%, and 27.5%, respectively, at day 14 after BI. Data are means ± SD; n = 5–8 mice/group. **p < 0.01. c: The figure shows the absolute weights of tibialis, gastrocnemius and soleus relative to body weight. All three muscles had a significant decrease in muscle mass measured mg/g.

FIGURE 4

Co‐staining of neurons, microglia and nuclei after Burn injury with or without GTS‐21 in wild type and A7KO (α7 AChR) knockout mice. a: At day 14 after BI or sham‐burn (SB), lumbar 3–4 spinal cord ventral horn segments were cryo‐sectioned and stained for motor neurons (NeuN, green), microglia (Iba1, red), and nuclei (DAPI, blue). Ventral horn motor neurons (diameter ≥ 25 μm) were detected as green staining. Similar to observations in Figure 1, BI induced microgliosis in the ventral horn. There were increased numbers of microglia interacting with neurons in the ventral horn in BI, as compared to SB (pointed by white arrows in the merged figure for Iba1 and NeuN). Microgliosis and the increased interaction between microglia and neurons were mitigated by GTS‐21 treatment in the wild type mice (WT). Compared to wild type SB, the α7 knockout mice with SB showed a higher (84%) microgliosis. The beneficial effects of GTS‐21 against BI were not observed in the α7 knockout (A7KO) mice. This suggests that even in the basal uninjured state the α7AChRs have a role in spinal homeostasis. b: At day 14 after BI, the ventral horn of the lumber spinal cord was co‐stained with microglia marker, Iba1 (red), and pro‐inflammatory CD86 (yellow) expressed in microglia and monocytes/macrophages. Areas for double‐positive cells, suggestive of the pro‐inflammatory (classically M1) microglia phenotype, are pointed by white arrows. c: Quantification of microglia stained by Iba1 in the ventral horn, is shown as a bar graph based on image segmentation and the cell area size measurement (200X magnification) to confirm the findings in Figure 4a. BI caused as significant increase in microgliosis, which was normalised by GTS‐21 in wild type but not in the A7KO mice. d: Transcripts of markers of microglia pro‐inflammatory phenotype iNOS and CD86, and anti‐inflammatory, Ym‐1 were quantified in the lumbar spinal ventral horn by RT‐PCR and normalised to the internal control, GAPDH. In the WT, BI caused increase of pro‐inflammatory markers (iNOS and CD86), but not of anti‐inflammatory marker (Ym‐1). GTS‐21 treatment mitigated the increment of pro‐inflammatory microglia in the wild type mice with BI, but not in the A7KO mice. GTS‐21 treatment increased the expression of the alternatively activated Ym‐1 of wild type BI mice, reflecting M2 phenotype. Basal level of pro‐inflammatory marker (CD86) was high in the A7KO mice than WT (A7KO‐SB vs. WT‐SB), which was exacerbated by BI (A7KO‐BI vs. WT‐BI). Both pro‐ and anti‐inflammatory markers (CD86 and Ym‐1, respectively) were higher in the A7KO‐BI than in WT‐BI. The marker for anti‐inflammatory subtype of microglia, Ym‐1 was increased by GTS‐21 treatment in the WT‐BI animals, but not in A7KO mice with BI. *p < 0.05, **p < 0.01.

FIGURE 5

Anti‐inflammatory effect of α7 AChR stimulation against BI induced inflammatory response in the spinal cord. a: Quantitative PCR mRNA expression of pro‐inflammatory cytokines (IL‐1β, TNFα) and chemokines (CXCL2 and MCP1) were measured and normalised to GAPDH in the ventral horn of the spinal cord samples from BI mice with or without GTS‐21 treatment. We showed that the mRNA expression of pro‐inflammatory and chemokine molecules, IL‐1β, TNF‐α, and CXCL2, MCP‐1 were significantly (p < increased compared to SB at day 7 and/or 14 post‐BI). In these experiments, the effects of GTS‐21 were studied only in BI WT and A7KO mice. The administration of GTS‐21 significantly ameliorated inflammatory chemokine and cytokine expression in the wild type mice (WT). The beneficial effect of GTS‐21 was, however, nullified in the α7 AChR knockout mice (A7KO). b: the immunoblots of lumbar spinal cord (L3‐L4 segments) isolated from BI mice at day 7 and 14 post‐injury showed increased STAT3 and NFκB expression when normalised to GAPDH as loading control. In the following experiments, the effects of GTS‐21 inflammatory proteins were evaluated in BI WT and KO mice only were tested: The probed immunoblots were quantified to measure band intensities of target proteins (phosphorylated vs. total molecule). The increased levels of the phosphorylated STAT3 and of phosphorylated NFκB of BI were significantly reduced by GTS‐21 treatment in the wild type (‘WT’). GTS‐21 treatment of BI A7KO mice showed no mitigation of the STAT3 and NFkB phosphorylation. Data are means ± SD; n = 5–6, for BI and SB, * p < 0.05, **p < 0.01.

FIGURE 6

Motor neuron numbers quantified by Nissl staining. a: Nissl staining of spinal cord dorsal and ventral horn of the Sham Burn (SB) and burn injured (BI) mice with or without GTS‐21 treatment were examined in wild type (WT) or α7 AChR knockout (A7KO) mice. In the WT mice with SB (WT‐SB), the typical dark blue staining of the cytoplasmic portion due to Nissl bodies was observed (see magnifications of the ventral horn area from the insets enclosed by solid black line, and for the representative neurons from the insets enclosed by dotted line). In the WT‐BI mice, Nissl body was less obvious with the lighter staining of the cytoplasm and the overall cell shape swollen in BI with no GTS‐21. BI decreased the motor neuron numbers compared to WT SB (31.6% reduction). GTS‐21 ameliorated the motor neuron loss of in the ventral horn of the WT mice (improved to 12.3% reduction from SB), but not in the A7KO mice. Note that in the SB A7KO mice, there was decreased neuron numbers compared to SB WT (61.5% of WT‐SB) and the BI exacerbated the neuronal loss in the A7KO mice. b: The motor neuron numbers (diameter ≥ 25 μm, 200X magnification) in the ventral horn of each group were quantified. BI caused a significant decrease in neuron numbers, which was significantly mitigated by GTS‐21. There was significant neuron loss in the SB A7KO compared SB WT; the presence of BI aggravated the neuron loss in the A7KO. Data are means ± SD; n = 6. *p < 0.05.

FIGURE 7

GTS‐21 prevents motor neuron loss by blocking neuronal apoptosis. a: In situ TUNEL assay for apoptosis (green) in the ventral horn area of the spinal cord with counter‐staining for neurons using NeuN (NeuN, red) and nuclei (DAPI, blue) in the burned injured (BI) and sham‐burn (SB) mice with or without α7AChR agonist, GTS‐21. As compared to BI without treatment, GTS‐21 treatment reduced the TUNEL positive nuclei in the neurons in the wild type mice. In the α7AChR knockout mice, however, the therapeutic effect by GTS‐21 was not observed. b: Immunostaining against activated caspase‐3 (green) of the ventral horn area is shown with counterstaining against neurons (NeuN, red) and nuclei (DAPI, blue). Compared to BI group without treatment, GTS‐21 administration lowered the expression of caspase‐3. The therapeutic effect by GTS‐21 was nullified in α7AChR knockout mice.

FIGURE 8

GTS‐21 stimulation of α7 AChR maintains synaptic integrity and mitigates muscle wasting. a: The morphology of synaptic membrane and the apposition of pre‐ and post‐synaptic apparatus was compared in BI mice with or without α7AChR agonist, GTS‐21 (‘+G’) in WT and α7AChR knockout (‘A7KO’) mice. Tibialis anterior muscles of BI mice were stained with α‐BTX (red) for post‐synaptic muscle AChRs on muscle membrane and anti‐synaptophysin antibody for presynaptic neuron to monitor the disintegration of the synapse denervation status, with or without GTS treatment. The disintegrated AChR morphology (red) and the partial denervation phenotype (poor apposition of a‐BTX and synapse, pointed by white arrows) in the BI group was improved by GTS‐21 treatment in the wild type mice (‘WT‐BI+G’). The therapeutic efficacy of GTS‐21 was lacking in α7AChR KO mice. (Scale bar = 30 μm). b: The changes in muscle weight of SB, and BI group mice at day 14 in the WT or α7AChR KO mice. Dry muscle weights TA, SOL, and GC muscle] are expressed as a ratio to body weight pre‐burn. The body weight at day 0 was used in view of body weight loss induced by BI at day 14. Muscle wasting by BI was significantly improved by GTS‐21 treatment in the WT, but not in α7AChR KO mice. Data are means ± SD; n = 5–8 mice/group. *p < 0.05, **p < 0.01. c. The figure shows the absolute weights of tibialis, gastrocnemius and soleus relative to body weight. All three muscles had a significant decrease in muscle mass measured mg/g.