Background
While the inflammatory response to injury in the peripheral nervous system (PNS) is complex and involves a number of immune cells, including dendritic, T, and B cells [
1,
2], it is largely dominated by macrophages [
3]. Macrophage responses to a peripheral axonal injury occur in two distinct compartments: the nerve distal to the injury site and the axotomized ganglion. The largest contribution to macrophage accumulation within injured PNS tissues come from monocyte-derived macrophages. In the sciatic nerve 7 days after injury, infiltrating macrophages outnumber the resident population 3:1 [
4]. Infiltration of the sciatic nerve begins 48 h after injury with peak accumulation occurring between 7 and 14 days [
3,
5]. Additionally, macrophage numbers remain elevated in the nerve for at least 30 days after injury [
6].
In peripheral ganglia, monocyte infiltration and macrophage accumulation follow a similar timeline to that seen in the sciatic nerve [
7‐
9]. In the superior cervical ganglion (SCG), monocyte-derived macrophages are present at 48 h after injury and remain for at least 2 weeks following transection of the internal and external carotid nerves [
9,
10]. Likewise, the lumbar dorsal root ganglion (DRG) shows accumulation beginning 48 to 72 h after sciatic nerve injury with an increased number of monocyte-derived macrophages remaining for at least 28 days [
11,
12].
The signal-mediating monocyte/macrophage entry into PNS tissue following injury is thought to be the C-C motif chemokine ligand 2 (CCL2), as a global knockout of its primary receptor C-C class chemokine receptor 2 (CCR2) yields a significant decrease in macrophage accumulation in the sciatic nerve following injury [
13,
14]. CCL2, also known as monocyte chemoattractant protein-1 (MCP-1), recruits monocytes to sites of inflammation, infection, and injury. CCL2 also has the ability to act as a chemokine for neutrophils, T cells, and dendritic cells [
15,
16].
While the function of macrophages in peripheral ganglia following nerve injury has been unknown until recently, evidence in the DRG and SCG indicates a role in axonal regeneration [
11,
14,
17,
18]. Since there is very little cell death and degeneration occurring in peripheral ganglia after injury [
19,
20], the stereotypical phagocytic function of macrophages is likely not their main function in the cell body compartment after axonal injury, suggesting an alternate purpose [
21,
22]. Recent work by our laboratory has demonstrated that macrophage accumulation near injured neuronal cell bodies is both necessary and sufficient for peripheral axon regeneration [
14,
17]. By inhibiting macrophage accumulation within the SCG and DRG following axotomy using a global knockout of the chemokine receptor
Ccr2 (
Ccr2
−/−
), we showed that in vitro axonal regeneration is significantly reduced when the injury-induced immune response in ganglia is impaired [
14].
Numerous studies on peripheral nerve regeneration and the neuronal cell body response to axonal injury exclusively utilize injury to the sciatic nerve and DRG as a model system. However, the DRG is a very unique ganglion in that the sensory neurons lack dendrites and synaptic input within the cell body region, the neurons within the DRG are a heterogeneous population composed of multiple subtypes, and the ganglia lack a blood-ganglion barrier [
23]. The differences between the DRG and other peripheral ganglia could allow for unique mechanisms to arise in response to peripheral axonal injury. Thus, a comparison of the injury responses, and more specifically the axotomy-induced immune responses, between the DRG and other peripheral ganglia could identify cellular and molecular mechanisms that are common elements of peripheral injury or unique elements that are specific to the DRG. The SCG, the largest sympathetic ganglia, has many characteristics that directly oppose those of the DRG such as the presence of presynaptic innervation within the ganglia, a more homogeneous neuronal population, and the existence of a blood-ganglion barrier [
24‐
26].
Nevertheless, very little is known about the complete molecular and cellular components of the injury-induced immune response within peripheral ganglia. Here, we utilize unbiased antibody protein arrays, flow cytometry, and immunohistochemistry (IHC) to identify and characterize the molecular and cellular profile of the immune response after nerve injury within the DRG and SCG of wild-type (WT) and Ccr2
−/−
mice.
Methods
Mice
Male (8–12-week-old) WT (C57BL/6J; The Jackson Laboratory, Bar Harbor, ME, USA) and Ccr2
−/−
(B6.129S4-Ccr2
tm1Ifc
/J backcrossed to C57BL/6J; The Jackson Laboratory) mice were used for this study. All mice had ad libitum access to food and water and were housed under a 12-h light/dark cycle.
Injury model
Mice were anesthetized under isoflurane, and the external carotid nerve (ECN) and internal carotid nerve (ICN) of the right SCG were exposed and transected. For axotomy of neurons in the fourth, fifth, and sixth DRG, the right sciatic nerve was exposed and transected at the hip level, and 1 mm of the nerve was removed. The nerves on the left side of the animal were exposed but not transected, and the ganglia served as sham-operated controls. One, two, three, or seven days after injury, mice were killed by CO2 inhalation, and the SCG and a combination of the L4, L5, and L6 DRG were harvested for analysis. All surgical procedures were approved by the Case Western Reserve University Institutional Animal Care and Use Committee.
Antibody arrays
L4 and L5 DRG and SCG from WT and Ccr2
−/−
mice were collected at 48 h (for chemokine assay) and 7 days (for cytokine assay) after right sciatic nerve transection and right SCG axotomy and flash frozen for chemokine and cytokine determination by Proteome Profiler Array ARY020 and ARY006 (R&D Systems, Minneapolis, MN, USA), respectively. Ganglia from five mice/group were pooled for each sample, and protein was isolated through homogenization in RIPA buffer (Santa Cruz Biotechnology, Dallas, TX, USA). Protein (180 μg per sample) was diluted and mixed with a cocktail of biotinylated detection antibodies and then incubated with mouse chemokine or cytokine array membrane. After washing, streptavidin-horseradish peroxidase and chemiluminescent detection reagents were added. Array images were obtained using a LI-COR Odyssey Fc Imaging System (LI-COR Biosciences, Lincoln, NE, USA) with a 10-min exposure time and analyzed by densitometry for integral optical density using ImageJ software. The optical density of each pair of chemokine or cytokine spots was normalized to the corresponding positive control spots. The data were then represented as a fold increase of the injured chemokine or cytokine expression compared to the WT sham results. Three arrays were run per condition, and the results were averaged. t tests were used to evaluate statistical differences between injured and uninjured tissues within a genotype and in the injured tissue between genotypes. The following chemokines were assessed at 48 h post injury in the DRG and SCG: 6Ckine, BLC, C10, C5/C5a, CCL28, chemerin, CTACK, CXCL16, eotaxin, CX3CL1, interleukin (IL)-16, IP-10, I-TAC, CCL2, CXCL1, LIX, CCL8/MCP-2, MCP-5, MDC, MIG, macrophage inflammatory protein-1 (MIP-1)α/β, macrophage inflammatory protein-1 gamma (MIP-1γ), MIP-2, RANTES, and SDF-1. The following cytokines were assessed at 7 days post injury in the DRG and SCG: CXCL13, C5a, G-CSF, GM-CSF, CCL1, CCL11, soluble intercellular adhesion molecule-1 (sICAM-1), IFN-γ, IL-1α, IL-1β, interleukin-1 receptor antagonist (IL-1ra), IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-10, IL-13, IL-12p70, IL-16, IL-17, IL-23, IL-27, IP-10, CXCL11, KC, M-CSF, CCL2, CCL12, CXCL9, CCL3, CCL4, CXCL2, CCL5, CXCL12, CCL17, tissue inhibitor of metalloproteinase-1 (TIMP-1), tumor necrosis factor alpha (TNF-α), and TREM-1.
Flow cytometry
A pooled sample of L4, L5, and L6 DRG and single SCG was enzymatically digested in 0.125% collagenase for 1 h at 37 °C. Mechanical digestion using a 23-gauge needle attached to a 1-ml syringe produced single cell suspensions which were filtered through a 35-μm cell strainer. Dead cells were labeled using a Live/Dead Fixable Blue Dead Cell Stain kit (Thermo Fisher Scientific, Waltham, MA, USA) for 30 min at 4 °C. Cells were then washed in FACS buffer (PBS, 1% BSA) and blocked with a monoclonal antibody to CD16/CD32 (1:500; Thermo Fisher Scientific) for 10 min at 4 °C. Cells were incubated with fluorophore-conjugated antibodies against F4/80, CD11b, Ly6G (1A8), and p75 ([F4/80] Cat no. 123130; [CD11b] Cat no. 101206; [Ly6G] Cat no. 127610; [p75] Cat no. 113405) for 1 h at 4 °C. Cells were washed and resuspended in FACS buffer and then run on a BD FACSAria (BD Biosciences, San Jose, CA, USA) and analyzed using FlowJo (BD Biosciences). CD45 was not used to gate leukocyte populations, and therefore, non-leukocyte populations may be included in the analysis [
27]. All events were gated based on viable single cells. Compensation and gating were performed using negative, single-stained, and isotype controls. Cell populations were gated as follows: F4/80
+CD11b
+ (macrophages), CD11b
+Ly6G
− (monocytes/macrophages), CD11b
+Ly6G
+ (neutrophils), and p75
+ (Schwann cells).
IHC
L5 DRG and SCG from WT and Ccr2
−/−
mice were removed 1, 3, or 7 days post injury, and the ganglia were desheathed and fixed by immersion in 4% paraformaldehyde. The tissues were cryoprotected in 30% sucrose and embedded in Tissue-Tek O.C.T. compound (Electron Microscopy Sciences, Hatfield, PA, USA). IHC was performed on 10-μm cryostat sections to double label tissue with a macrophage marker, CD68, and an anti-inflammatory macrophage marker, CD206. A goat polyclonal antibody to CD206 (1:200; R&D Systems) was incubated with tissue sections overnight at 4 °C. After washing, the sections were incubated in Alexa Fluor 488 secondary antibody (1:400; Thermo Fisher Scientific) for 1 h and washed in PBS. Next, a rat monoclonal antibody against CD68 (1:200; AbD Serotec, Oxford, UK) was incubated with tissue sections overnight at 4 °C. After additional washes, the sections were incubated with a Cy3 secondary antibody (1:200; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) for 1 h. The sections were then stained with DAPI (1:1000; Thermo Fisher Scientific). IHC was also performed to label neutrophils in the ganglia using a rabbit polyclonal antibody to myeloperoxidase (MPO; 1:100; Abcam, Cambridge, UK) overnight at 4 °C and a Cy3 donkey anti-rabbit secondary antibody (1:400; Jackson ImmunoResearch Laboratories, Inc.). Additionally, IHC was performed on L3, L4, L5, and L6 DRG and SCG to identify the percentage of injured neurons after axotomy using a mouse monoclonal anti-HuC/D (1:100; Thermo Fisher Scientific) and a rabbit polyclonal anti-activating transcription factor 3 (ATF3) antibody (1:100; Sigma-Aldrich, St. Louis, MO, USA) by incubating sections overnight at 4 °C. Prior to the addition of primary ATF3 and HuC/D antibodies, the tissue underwent citrate buffer antigen retrieval and an endogenous mouse antibody blocking step by incubating the tissue with a donkey anti-mouse-unconjugated IgG (1:400; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA). ATF3 and HuC/D were visualized by incubation with goat anti-rabbit Alexa Fluor 594 (1:400; Jackson ImmunoResearch Laboratories, Inc.) and goat anti-mouse IgG2a Alexa Fluor 488 (1:400; Thermo Fisher Scientific). In all experiments, sections not exposed to the primary antibody were included for each experimental group. Images were captured at × 10 or × 25 magnification using HCImage software (Hamamatsu Corporation, Bridgewater, NJ, USA), and then cell counts were performed using ImageJ software. Cell counts were performed to quantify the number of CD68+, CD206+, CD68+CD206+, HuC/D+, ATF3+, and HuC/D+ATF3+ cells. Data are represented as the sum of the cells counted from three distinct images per ganglion per animal. Five samples were included for each group.
Statistical analysis
Data in graphs are presented as mean ± SEM. For experiments involving multiple genotypes, treatments, and time points, a two-way ANOVA was performed, followed by a Tukey’s post hoc test. For independent two-group experiments, an unpaired Student’s t test (two-tailed) was used to determine statistical significance. Statistical analyses were performed using SigmaPlot, version 12.3 (Systat Software, Inc., San Jose, CA, USA). Values were considered statistically significant at p < 0.05. The number of experimental replicates (n), the designation of n belonging to the number of animals, and experiment repetition are indicated in the figure legends. Data from all experiments were included in the analysis. IHC experiments were performed with the experimenter blinded to genotype.
Discussion
CCL2 overexpression in sensory ganglia has been shown to be sufficient to cause macrophage accumulation in the DRG without a nerve injury [
17]. Here, we have shown that CCL2 protein expression is increased in both the WT and
Ccr2
−/−
axotomized DRG and SCG early after injury. CCL2 upregulation is associated with a heightened macrophage response in both ganglia, albeit this response is markedly reduced in
Ccr2
−/−
mice. Indeed, deletion of the receptor alone showed almost complete blockade of macrophage accumulation in the DRG and a significant but only partial reduction in the SCG after nerve injury [
14]. These results suggest that a second chemokine and/or action through a second chemokine receptor is facilitating monocyte-mediated chemotaxis to the injured sympathetic ganglia.
An obvious candidate for this role is MIP-1γ, which we found, using the unbiased chemokine array, to be increased only in the injured SCG of both WT and
Ccr2
−/−
mice. MIP-1γ acts through CCR1 expressed on monocytes, neutrophils, dendritic cells, and T cells [
50‐
53]. Some evidence indicates that MIP-1γ may act in an autocrine fashion on macrophages to promote their activation and survival [
53], while other studies point to dendritic cells as important producers of this chemokine for extravasation of inflammatory cells from the bloodstream [
51]. Regardless of its source, heightened expression of MIP-1γ in the SCG is associated with increased macrophage accumulation. Furthermore, robust expression of both MIP-1γ and CCL2 in the SCG could explain the early and significant accumulation of macrophages that we observed in the axotomized sympathetic ganglia at 1 day post injury, compared to the typical increase in the macrophage response observed in the injured DRG after 3 days [
11,
12]. However, we would need to confirm that MIP-1γ is upregulated in the SCG earlier than the 48-h time point that was used in this study.
What are the consequences of inhibiting macrophage accumulation in these ganglia after nerve injury? The most well-described function of macrophages is their role in the distal nerve segment where they promote regeneration of injured neurons through growth factor release and the removal of inhibitory axonal and myelin debris at the distal nerve that would otherwise obstruct the path for regenerating axons [
54‐
57]. Though peripheral nerve degeneration involves the actions of multiple cells, hematogenous macrophages that infiltrate as monocytes from the blood to the injury site have been considered to be necessary for the phagocytosis of myelin and axonal fragments [
13,
58‐
60]. It is believed that while Schwann cells initiate myelin uptake and removal [
61,
62], infiltrating macrophages complete the clearance process [
60,
63,
64]. However, recent studies have shown that myelin clearance is normal in
Ccr2
−/−
sciatic nerves after an injury [
14], and that neutrophils play a pivotal role in this clearance [
65].
Of late, a new site of macrophage action has been proposed in promoting nerve regeneration. Following injury to the sciatic nerve or ICN and ECN, macrophages accumulate within the DRG and SCG, respectively [
14]. Mice lacking CCR2 exhibited diminished macrophage accumulation in the ganglia and a concomitant decrease in axonal regeneration [
14], while a similar study that administered the antibiotic minocycline to reduce macrophage numbers in the DRG also observed a reduction in regeneration [
11]. Moreover, our lab provided evidence that simply overexpressing CCL2 in sensory ganglia without a nerve injury is sufficient to mount an injury-like macrophage response in the DRG and to prompt robust neurite growth [
17].
What remains to be determined is how these macrophages promote regeneration at the level of the neuronal cell bodies. The most likely mechanism of action is through a macrophage-derived releasable factor, such as a cytokine, that stimulates the neuronal expression of regeneration-associated genes [
45,
66‐
68]. Analysis of regeneration-associated genes in the uninjured DRG in which CCL2 was overexpressed highlighted increased expression of leukemia inhibitory factor (
Lif) mRNA [
17], a gene that had been previously described as promoting sensory and sympathetic neuron regeneration [
69,
70].
In the cytokine array, we observe differential expression of cytokines between genotypes in the injured DRG 7 days after injury, many of which are downregulated in the
Ccr2
−/−
model compared to WT mice. This is not surprising if we consider evidence that the lack of CCR2 and a resulting decrease in macrophage accumulation in the injured ganglia reveal an inhibition of in vitro regeneration of both sensory and sympathetic neurons [
14]. TIMP-1, a protein that we find is reduced in the axotomized
Ccr2
−/−
DRG and SCG compared to their axotomized WT counterparts, has been shown to be important in supporting regeneration in the central nervous system [
71]. TIMP-1 functions to inhibit the activity of metalloproteinases which cleave myelin-associated glycoproteins that accumulate in the central nervous system after nerve injury and obstruct the path of regenerating DRG neurons [
72]. Peripheral nerve injury has been shown to induce TIMP-1 expression in the DRG [
73]. Furthermore, a preconditioning injury to the peripheral nerve process of the DRG followed by an injury to the central nerve process results in TIMP-1 upregulation in the DRG and the spinal cord in addition to increased neurite outgrowth into the spinal cord [
71].
The specific contribution of macrophage releasable factors to the cytokine microenvironment has been shown to be indicative of the macrophage phenotype, i.e., pro-inflammatory or anti-inflammatory [
46]. Pro-inflammatory macrophages are characterized by the increased production of nitric oxide and specifically the enzyme inducible nitric oxide synthase (iNOS) [
74,
75], while anti-inflammatory macrophages upregulate expression of the mannose receptor CD206 [
76]. These distinctions of macrophage activation states are important because of recent work that has shown that conditioned media taken from anti-inflammatory macrophage cultures was able to stimulate DRG neurite outgrowth on both growth-permissive and growth-inhibiting substrates [
45]. We previously reported increased CD206 transcript levels in the uninjured DRG of WT mice where CCL2 was overexpressed, resulting in increased macrophage accumulation and neurite outgrowth [
17]. Here, we find that a majority of macrophages in the axotomized WT DRG and SCG at 1, 3, and 7 days post injury express the anti-inflammatory marker CD206, which suggests that these macrophages are of an anti-inflammatory nature. It should be noted that most macrophages in the axotomized
Ccr2
−/−
DRG and SCG also express CD206, but the overall accumulation of these myeloid cells is drastically reduced relative to WT mice. Furthermore, the hypothesis that monocyte-derived macrophages that infiltrate the DRG and SCG after injury are of an anti-inflammatory type is evidence of the small number of macrophages that express the pro-inflammatory marker iNOS (data not shown).
In addition to the evaluation of macrophage activation states and the evident differences in macrophage accumulation between genotypes in the axotomized ganglia, we also note a genotype-specific alteration in neutrophil chemotaxis to the SCG.
Ccr2
−/−
sympathetic ganglia boast a larger neutrophil population than its WT counterpart 3 days following nerve injury. This increase is associated with heightened expression of CXCL5, a neutrophil chemokine, in the
Ccr2
−/−
SCG [
77]. Interestingly, upregulation of another neutrophil chemokine MIP-2 (CXCL2) was not observed, though this chemokine most likely plays a role in increased neutrophil accumulation in the injured
Ccr2
−/−
sciatic nerve compared to WT nerves [
65].
Interestingly, our data suggest that different subpopulations of neutrophils may exist in the SCG after nerve injury. Flow cytometry analysis demonstrated that CD11b
+Ly6G
+ neutrophils were significantly more abundant in the
Ccr2
−/−
SCG 3 days following injury compared to the WT SCG; however, IHC data showed comparable MPO
+ neutrophil accumulation in the sympathetic ganglia between genotypes at this same time point. The idea of neutrophil heterogeneity has gained considerable traction with evidence that has shown that these polymorphonuclear cells develop distinct phenotypes in response to a range of stimuli [
78].
Unlike in the SCG, we do not identify a significant population of neutrophils in the DRG after sciatic nerve transection in either genotype. However, the role of neutrophils in the DRG has been extensively characterized in experimental paradigms of neuropathic pain. Neutrophils reportedly infiltrate both the DRG and the sciatic nerve in response to chronic nerve constriction [
79] and partial nerve ligation [
2]. These types of injuries manifest neuropathic pain due to the sensitization of neurons by exposure to a prolonged inflammatory response [
1]. Neutrophils and other immune cells including macrophages and T cells have been shown to contribute to neuropathic pain via the release of TNF-α, IL-1β, and IL-6 through directly engaging cognate receptors on sensory neurons (for a summary, see [
80]). Chemotactic disruption of these immune cells using pharmacological or genetic methods has been shown to ameliorate the severity of neuropathic pain [
81‐
83].
Clear differences in chemokine and cytokine expression profiles were observed between DRG and SCG following axotomy. Forty-eight hours after injury, C5/C5a is significantly increased in the WT DRG but is not upregulated in the WT SCG, while MIP-1γ is highly upregulated in the injured WT SCG but is not increased in the DRG after injury. Seven days after injury, IL-1ra is significantly upregulated in the WT SCG but not the WT DRG. The distinct expression of specific immune signaling molecules between ganglia could be a result of the presence of a significant neutrophil population found in the SCG but not the DRG 3 days after injury. However, other ganglion-specific aspects such as the presence of a blood-ganglion barrier in the SCG or the significant heterogeneity of neurons in the DRG could play a contributing factor [
24,
84].