DICAM in the Extracellular Vesicles from Astrocytes Attenuates Microglia Activation and Neuroinflammation

Cross-talk between astrocytes and microglia plays an important role in neuroinflammation and central sensitization, but the manner in which glial cells interact remains less well-understood. Herein, we investigated the role of dual immunoglobulin domain-containing cell adhesion molecules (DICAM) in the glial cell interaction during neuroinflammation. DICAM knockout (KO) mice revealed enhanced nociceptive behaviors and glial cell activation of the tibia fracture with a cast immobilization model of complex regional pain syndrome (CRPS). DICAM was selectively secreted in reactive astrocytes, mainly via extracellular vesicles (EVs), and contributed to the regulation of neuroinflammation through the M2 polarization of microglia, which is dependent on the suppression of p38 MAPK signaling. In conclusion, DICAM secreted from reactive astrocytes through EVs was involved in the suppression of microglia activation and subsequent attenuation of neuroinflammation during central sensitization.


Background
Central sensitization is a condition of the nerve system, which is in a state of persistent high reactivity, caused by a pathologic process that occurs at the level of the dorsal horn of the spinal cord, and is critical for the development and persistence of chronic pain, such as complex regional pain syndrome (CRPS), postherpetic neuralgia, and fibromyalgia [1][2][3]. Accumulating evidence suggest that central sensitization is driven by neuroinflammation that is characterized by the infiltration of leukocytes, activation of glial cells, and increase in inflammatory mediators [4][5][6]. In the process of neuroinflammation, microglia play a critical role as the tissue-resident macrophages of the central nervous system (CNS) [4]. In physiologic conditions, microglia regulate CNS development and homeostasis via neuronal-microglial interactions, apoptotic cellular debris clearing, trophic factor secretion, and synaptic modeling [7]. However, prolonged stimulation of microglia leads to the release inflammatory cytokines, such as TNF-α, IL-6, IL-1β and prostaglandin E2, reactive oxygen species (ROS), and excitotoxins, including glutamate [8]. These neuromodulators ultimately enhance pain signal transmission by facilitating glutamate-mediated nociceptive neurotransmission and contributing to the loss of gamma-aminobutyric acid and glycinsecreting inhibitory interneurons [9][10][11].
Astrocytes, as the most abundant glial cell type in CNS, have diverse functions, such as structurally forming the blood-brain barrier and supporting neuronal cells, and

Hind Paw Volume Measurement
The hind paw volume was measured every week after cast-off as follows: paw volume (mm 3 ) = 1/2(length of long axis × length of short axis × thickness of paw) [6]. The lengths of the long axis, short axis, and thickness of the ipsilateral and contralateral hind paws were measured using an electronic digital caliper [6].

Threshold Punctate Mechanical Stimulation (von Frey Test)
The von Frey test was conducted to evaluate tactile allodynia using calibrated monofilaments (von Frey hairs; Stoelting, Wood Dale, IL), which were applied to the plantar surface of the ipsilateral and contralateral hind paws placed on an elevated maze in the acrylic cage. Paw withdrawal was considered as a positive response, and the 50% withdrawal threshold upon six repeated applications of varying force with the von Frey filament was measured using the up-down method [29].

Spontaneous Weight-Bearing Test (Incapacitance Test)
Spontaneous weight bearing on the hind limbs, also known as the incapacitance test, was conducted to assess the downward force applied by each hind limb using the incapacitance meter (SangChung commercial, Seoul, Korea). Briefly, mice were placed in the restrainer, and the hind limbs were rested on the two weight averaging platform pads. The paw pressure of each hind limb underwent 10-s measurements approximately 10 times, and the results were averaged. Data were expressed as the percentage of weight distributed on the ipsilateral hind limb.

Rotarod Test
The rotarod test was performed to measure the locomotor performance of mice using the LE8205 Accelerating Rota Rod (Havard Apparatus, Holliston, MA, USA). Briefly, mice were placed on a rotating cylinder with an increasing speed of 0-30 rotations per minute (rpm) for 60 s, which was followed by an additional 240 s at 30 rpm [30]. Latency to fall measurements were repeated five times with 30 min breaks, and the results were averaged [30].

Immunofluorescence Staining
Mice were perfused with phosphate buffered saline (PBS) through the aorta to remove blood, and the lumbar spinal cord was subsequently dissected. After fixation with 4% paraformaldehyde, the spinal cord specimens were cryoprotected in 30% sucrose for 2 days, embedded in the optimal cutting temperature compound, sliced into 20 µm-thick sections and mounted on gelatin-coated slides. Tissue sections were washed with tris-buffered saline with Tween-20 (TBST), blocked with 10% donkey normal serum (cat. #ab7475, Abcam, Cambridge, UK) in 0.3% Triton X-100 for 60 min, and then incubated with primary antibodies against Iba-1 (rabbit, 1:200; Wako, Osaka, Japan) and glial fibrillary acidic protein (GFAP) (rabbit, 1:500; cat. Z0334, DAKO, Carpinteria, CA, USA) overnight at 4 • C. Subsequently, the sections were incubated with fluorescein isothiocyanate (FITC-) or Cy5conjugated secondary antibodies (1:500, Jackson ImmunoResearch, West Grove, PA, USA) for 2 h, and slides were washed and mounted with Fuoromount aqueous mounting medium (cat.#F4680, Sigma) [5]. Four square areas of the ipsilateral and contralateral dorsal horn of the spinal cord were randomly selected and photographed at 100 magnifications under a KI-3000F fluorescence microscope (Korealabtech, Gyeonggi-do, Korea). The Iba-1 and GFAPpositive cells were quantified using ImageJ software (NIH; https://imagej.nih.gov/ij/ accessed on 2 February 2020). In order to distinguish the background and positive cells, the binary threshold was set to 50%, and positive cells were defined as more than 5 pixels.

Primary Cell Isolation and BV2 Microglia Cell Line Culture
Primary mixed glial cells were prepared from the brain cortices of 2-to 3-day-old newborn pups among the C57BL/6J mice. After removing the meninges, brain tissues were homogenized and mechanically disrupted with the nylon mash, which were triturated to the cell suspension and centrifuged at 200 g for 15 min. After resuspension, cells were plated onto T-75 flasks in the DMEM with 10% FBS medium, cultured for 14 days and used for the in vitro experiment. Primary mixed glial cells were then stimulated with lipopolysaccharides (LPS) (100 ng/mL) and interferon-gamma (IFN-γ) (10 ng/mL) in 96-well plates for 24 h.
For the isolation of primary astrocytes, mixed glial cells that were cultured for 14 days were agitated by shaking them at 250 rpm overnight to remove weakly attached microglia. Culture media containing cells were discarded, and astrocytes were dissociated using trypsin-EDTA and then collected by centrifuging them at 1500 rpm for 5 min. For the isolation of primary microglia, mixed glial cells that were cultured for 14 days were separated from the glia layer by shaking them at 200 rpm for 3 hours. Afterwards, the floating microglia were collected by centrifuging them at 1500 rpm for 5 min. The purities of primary astrocytes and microglia exceeded 95%, as determined by GFAP and CD11b, using traditional reverse transcription polymerase chain reaction (RT-PCR).
Murine microglial cell line BV2 was obtained from Xiehe Medical University (Beijing, China). The cells were cultured in DMEM supplemented with 10% heat-inactivated fetal bovine serum, 100 IU/mL penicillin, 100 mg/mL streptomycin and 2 mmol/L glutamine.

Nitric Oxide (NO) Quantification
Nitric oxide colorimetric assay was performed as described previously. The quantity of nitrite, a stable metabolite of NO, in the culture medium was measured as an indicator of NO production. Briefly, 100 µL of cultured media was mixed with the same amount of Griess reagent (Cell Signaling Technology, Danvers, MA, USA) in 96-well plates, incubated at room temperature for 10 min, and the absorbance at 540 nm was measured in a microplate reader (BioTek, Winooski, VT, USA). Fresh culture medium was used as a blank in every experiment, and the quantity of nitrite was determined from a sodium nitrite standard curve.

Real-Time Quantitative PCR and Reverse Transcription PCR
Total RNA was isolated from cultured primary mixed glia, astrocytes, microglia, and the BV-2 microglial cell-line using the TRIzol reagent (Invitrogen). Briefly, cDNA was synthesized from 2 µg of total RNA using the Superscript II reverse transcription kit (Invitrogen, Waltham, MA, USA), with a random hexameter and oligo d(T) primer. Real-time qPCR was then performed using the Luna ® Universal One-Step RT-qPCR Kit (New England Biolabs, Ipswich, MA, USA), which was followed by detection using the ViiA™ 7 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Additionally, the 2-∆∆CT method was used to calculate the relative changes in gene expression, with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as the control. All the RT-qPCR reactions were performed in triplicate and repeated two to three times. Among them, the representative results are shown. The primer nucleotide sequences are listed in Table 1.
Reverse transcription PCR amplification using specific primer sets was carried out at an annealing temperature of 55-60 • C for over 20-30 cycles using the Applied Biosystems Veriti™ Thermal Cycler (Applied Biosystems). To analyze the PCR products, 10 µL of each PCR was electrophoresed on 1% agarose gel stained with ethidium bromide and was detected using the Criterion Stain-Free Imager (Bio-Rad). For cytokine measurements, primary mixed glial cells were stimulated, and the supernatant was subjected to ELISA analysis for mouse TNFα (Abcam; #ab208348, Cambridge, MA, USA), IL-6 (Abcam; #ab46100, Cambridge, MA, USA) and IL-10 (Abcam; #ab100697, Cambridge, MA, USA). The absorbance read was performed using a microplate reader at 450 nm, and the results were compared against a standard curve. Furthermore, the DICAM levels of cultured supernatant and plasma were measured using the DICAM ELISA kit (MyBioSource; #MBS907149, San Diego, CA, USA) according to the manufacturer's instructions.

Western Blot Analysis
Total cell lysates or nuclear and cytoplasmic extracts were separated by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene difluoride (PVDF) membranes (Amersham Biosciences, Buckinghamshire, UK). The membranes were blocked with 5% bovine serum albumin at room temperature for 1 h and were incubated with the aforementioned primary antibodies for another hour with shaking. Primary antibodies against p38 MAPK, p-p38 MAPK, JNK, p-JNK, ERK1/2, p-ERK1/2, Akt, p-Akt, IκBα, p-IκBα, p65, p-p65, Stat3, and p-Stat3 were purchased from Cell Signaling Technology (Beverly, MA, USA); antibodies against CD14, CD68, and CD63 were purchased from Cusabio Technology (Houston, TA, USA); the ERK antibody was purchased from BD Biosciences (Franklin lakes, NJ, USA); and the β-actin antibody was purchased from Sigma Aldrich (St. Louis, MO, USA). After the membranes were washed, incubation with the horseradish peroxidase (HRP)-conjugated secondary antibody was carried out for 1 h. The membrane signal was then visualized using Supersignal Chemiluminescent Substrates (Thermo Fisher Scientific, Waltham, MA, USA), and protein-specific signals were detected using the LAS-3000 (Fujifilm, Tokyo, Japan). Western blot band intensities were quantified with ImageJ software.

Blood Sampling from CRPS Patients and Controls
A pilot study to compare the serum DICAM level in CRPS patients with controls was conducted on 5 CRPS patients who visited the Rehabilitation clinic of Daegu Fatima hospital between May 2019 and June 2020 and fulfilled the Budapest diagnostic criteria for CRPS. The 5 controls were matched to cases on age and sex. The average disease duration of CRPS patients was 20.6 ± 9.7 months and numeric rating scale of pain intensity was 7.4 ± 1.1 ( Table 2). The plasma samples drawn from CRPS patients and controls were Cells 2022, 11, 2977 6 of 18 stored in a refrigerator at −80 degree Celsius, and were subjected to ELISA analysis within 1 month of blood collection. The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board of Daegu Fatima Hospital (DFH18ORIO358, 2018-10-10).

Isolation of Extracellular Vesicles
To obtain astrocyte-derived exosomes, the primary astrocytes were cultured at a density of 2 × 10 7 cells in a T175 flask and stimulated with or without LPS (100 ng/mL) and IFN-γ (10 ng/mL) by incubation for 24 h at 37 • C. After reaching 80% confluency, astrocytes were rinsed with PBS and cultured with serum-free medium for 72 h. The obtained supernatants were then filtered through a 0.22-µm filter and centrifuged at 10,000× g for 30 min at 4 • C to remove cellular debris. Subsequently, supernatants were collected and ultracentrifuged at 100,000× g for 1 h. The remaining supernatants were subsequently discarded, and the pellets were washed and resuspended in 200 µL of PBS.

Astrocyte-Derived Conditioned Media Treatment
To obtain astrocyte-derived conditioned media, primary astrocytes were plated at a density of 2 × 10 6 cells/well in 6-well plates and stimulated with or without LPS (100 ng/mL) and IFN-γ (10 ng/mL) for 24 h. The cells were then washed with PBS and cultured in fresh DMEM for an additional 24 h. Afterwards, conditioned media were collected and centrifuged at 320× g for 2 min to remove cellular debris, and the resulting astrocyte-conditioned media were treated with mixed glia cells for 24 h.

Statistical Analysis
All data are presented herein as mean ± standard error of the mean (SEM). Statistical analysis to compare the mean values of two groups was performed using the Mann-Whitney U test, a nonparametric test, as the sample size was small and it could be assumed as normal distribution. p-values ≤ 0.05 were considered statistically significant. Statistical analyses were performed with Prism software, version 8.0 (GraphPad Software, La Jolla, CA, USA).

DICAM KO Mice Showed Enhanced Nociceptive Behavior and Increased Number of Astrocytes and Microglia in the Dorsal Horn of the CRPS Model
To elucidate whether DICAM is involved in the central sensitization associated with neuropathic pain, the tibia fracture-cast immobilization (fracture/cast) model of CRPS was induced in 12-week-old DICAM KO mice and their WT littermates. The ipsilateral hind paw volume reached its maximum at three weeks after the fracture/cast procedure and then reduced and the DICAM KO mice showed a significant increase in ipsilateral hind paw volume at 3 weeks compared to WT ( Figure 1A). The tactile allodynia of the hind paws assessed by von Frey filaments revealed that the von Frey withdrawal threshold was lower in DICAM KO mice than WT mice at 3 weeks, which became similar in 4th and 5th week, and then worsened again in the DICAM KO mice in the 6th and 7th week ( Figure 1B). The weight-bearing of the ipsilateral hind limb was not different in WT and DICAM KO mice at the acute phase (3rd week after fracture), but it significantly decreased in DICAM KO Cells 2022, 11, 2977 7 of 18 mice compared to WT at the chronic phase of CRPS at 7 weeks ( Figure 1C). The rotarod test examining motor coordination behaviour revealed that DICAM KO mice showed a shorter latency to fall compared to naïve mice, suggesting an impaired motor coordination, while WT mice could hang on to the rotarod for a similar time compared to the naïve mice ( Figure 1D).
showed a shorter latency to fall compared to naïve mice, suggesting an impaired motor coordination, while WT mice could hang on to the rotarod for a similar time compared to the naïve mice ( Figure 1D).
As this enhanced pain-associated behavior suggests a sensitization of nociceptive pathways, we checked the activation of glial cells in the spinal cord dorsal horn (SCDH). The immunostaining of a microglia marker, Iba-1, revealed that DICAM KO mice had a significant increase in Iba-1 positive cells in both the ipsilateral and contralateral side of SCDH compared to WT mice ( Figure 1E). The GFAP immunostaining for astrocytes that are the most abundant glial cells in the CNS also showed an increase in GFAP-positive cells in ipsilateral SCDH of DICAM KO mice, but not in the contralateral side ( Figure 1E). These results demonstrate that DICAM is involved in glial cell activation during central sensitization associated with CRPS. As this enhanced pain-associated behavior suggests a sensitization of nociceptive pathways, we checked the activation of glial cells in the spinal cord dorsal horn (SCDH). The immunostaining of a microglia marker, Iba-1, revealed that DICAM KO mice had a significant increase in Iba-1 positive cells in both the ipsilateral and contralateral side of SCDH compared to WT mice ( Figure 1E). The GFAP immunostaining for astrocytes that are the most abundant glial cells in the CNS also showed an increase in GFAP-positive cells in ipsilateral SCDH of DICAM KO mice, but not in the contralateral side ( Figure 1E). These results demonstrate that DICAM is involved in glial cell activation during central sensitization associated with CRPS.

DICAM Deficiency Aggravates Proinflammatory Glial Cell Responses to Inflammatory Stimuli
Given the increase in microglia and astrocytes in DICAM KO mice in SCDH of the CRPS model, the inflammatory glial cell phenotype was evaluated under the proinflammatory stimuli by LPS and IFN-γ in vitro. Nitrite, a stable metabolite of nitric oxide that is a major mediator of innate immune responses, was significantly increased in the mixed glial cells from DICAM KO mice in the presence of LPS and IFN-γ, but not in unstimulated glial cells (Figure 2A). This pattern was reproduced in the RNA expression of proinflammatory cytokines, such as IL-1β and CXCL10. The expression of IL-10 was reduced to almost no expression in mixed glial cells from DICAM KO mice under unstimulated conditions, and in both DICAM KO and WT glial cells under the LPS and IFN-γ stimulated condition ( Figure 2B). The RNA of CD68, a marker for activated microglia, and CD86, a marker for M1 microglia, were upregulated in DICAM KO mixed glial cells compared to WT in the stimulated condition, while that of Arg-1, a marker for M2 macroglia, decreased in DICAM KO glial cells in both stimulated and unstimulated conditions. GFAP, a marker for astrocytes, increased in DICAM KO mixed glial cells compared to WT in unstimulated conditions, but not in LPS and IFN-γ stimulated conditions ( Figure 2C). When the pro-and anti-inflammatory cytokines were assessed using ELISA analysis, IL-6 increased in the supernatants from DICAM KO mixed glial cells, compared to WT in stimulated conditions. TNF-α level showed an increasing trend in DICAM KO glial cells in stimulated conditions, but failed to show statistical significance. The anti-inflammatory cytokine, IL-10, was decreased in DICAM KO compared to WT mixed glial cells in stimulated conditions ( Figure 2D). The protein level of CD14, a marker for primed microglia, was upregulated in DICAM KO glial cells in the stimulated condition but not in the unstimulated condition, while CD68, another marker for activated microglia, failed to show any difference between DICAM KO and WT mixed glial cells ( Figure 2E). These results collectively suggest that DICAM is involved in the anti-inflammatory phenotypes of mixed glial cells and in the M2 polarization of microglia, especially under inflammatory milieu.

DICAM Is Involved in the Activation of p38 MAPK in Mixed Glial Cells but Not in Astrocytes
To elucidate the signaling mechanisms that are involved in the DICAM-mediated antiinflammatory phenotypes of mixed glial cells, we analyzed the major signaling pathways activated by LPS and IFN-γ in DICAM KO and WT mixed glial cells and astrocytes. Mixed glial cells derived from DICAM KO mice revealed an enhanced phosphorylation of p38 MAPK compared to WT by LPS and IFN-γ, but did not show significant difference in other types of signaling, including ERK1/2, JNK, AKT, NF-kB, and STAT3 ( Figure 3). In contrast, astrocytes from DICAM KO and WT mice did not show any significant difference in the signaling cascade by LPS and IFN-γ ( Figure 3). Considering that mixed glial cells are mainly composed of astrocytes and microglia, these results suggest that the altered p38 MAPK activation of microglia was responsible for the proinflammatory phenotypes in DICAM KO mixed glial cells.

Activated Astrocytes Secrete DICAM through Extracellular Vesicles
We then checked the serum level of DICAM with ELISA assay in CRPS patients to assess the role of DICAM as a serum biomarker compared to healthy participants. It showed a tendency to decrease in CRPS patients compared to the healthy controls (p = 0.059) ( Figure 4A). To further examine the production of DICAM in glial cells, we assessed the RNA expression in primary astrocytes, primary microglia, and murine microglia cell-line BV2. When compared to GFAP, an astrocyte marker, and CD11b, a microglia marker, DICAM was selectively expressed in astrocytes but not in microglia ( Figure 4B). These findings indicate that DICAM mainly produced by astrocytes affects the activation of microglia in Figure 3 and also suggest that DICAM acts in a paracrine manner. To prove this hypothesis, we checked the DICAM protein level in the supernatant of the primary astrocyte culture. Stimulation with LPS and IFN-γ significantly increased the level of DICAM in the culture supernatant of primary astrocytes at 48 and 72h ( Figure 4C). Subsequently, we assessed whether this increase in DICAM in the culture supernatant of primary astrocytes was through extracellular vesicles (EVs). Immunoblot for CD63 and CD9, a marker for EVs, revealed that the activated astrocytes by LPS and IFN-γ secreted more exosomes than unstimulated astrocytes. Interestingly, DICAM was significantly enriched in the EVs isolated from the supernatant of activated astrocytes compared to unstimulated exosomes or total cell lysate ( Figure 4D). Taken together, these results suggest that activated astrocytes secret DICAM-enriched EVs, which are involved in the regulation of microglia activation.

DICAM from Activated Astrocytes Acts as Coupling Factor That Regulates Mixed Glial Cell Activation
We further examined the influences of the conditioned media from DICAM KO and WT astrocytes on the inflammatory phenotypes of the mixed glial cell culture. Conditioned media from DICAM KO astrocytes activated by LPS and IFN-γ significantly increased mRNA expression of IL-1β and CXCL10 from mixed glial cells, while it decreased that of IL-10 ( Figure 5A,B). It also decreased the expression of Arg-1, a marker for M2-microglia, but failed to affect CD68, CD86 and GFAP expression ( Figure 5C). In the ELISA analysis of pro-and anti-inflammatory cytokines, conditioned media from stimulated DICAM KO astrocytes significantly reduced secreted IL-10 levels in glial cells, but did not affect TNFα and IL-6 ( Figure 5D). When we assessed microglia activation with the CD14 and CD68 immunoblot, the stimulated conditioned media from DICAM KO astrocytes increased both CD14 and CD68 expressions in mixed glial cells, but the unstimulated media showed opposite effects on CD14 and CD68 ( Figure 5E).  . (B,C) The Western blot band intensities of phosphorylated proteins were quantified by densitometric analysis on the Image J program and normalized to nonphosphorylayed total protein, which was displayed as relative densitometric bar and dot graphs. * p < 0.05 by Mann-Whitney U test. n = 3 per group.

DICAM from Activated Astrocytes Acts as Coupling Factor That Regulates Mixed Glial Cell Activation
We further examined the influences of the conditioned media from DICAM KO and WT astrocytes on the inflammatory phenotypes of the mixed glial cell culture. Conditioned media from DICAM KO astrocytes activated by LPS and IFN-γ significantly increased mRNA expression of IL-1β and CXCL10 from mixed glial cells, while it decreased that of IL-10 ( Figure 5A,B). It also decreased the expression of Arg-1, a marker for M2microglia, but failed to affect CD68, CD86 and GFAP expression ( Figure 5C). In the ELISA analysis of pro-and anti-inflammatory cytokines, conditioned media from stimulated DICAM KO astrocytes significantly reduced secreted IL-10 levels in glial cells, but did not affect TNFα and IL-6 ( Figure 5D). When we assessed microglia activation with the CD14 and CD68 immunoblot, the stimulated conditioned media from DICAM KO astrocytes

Discussion
A growing body of evidence has shown the cross-talk between astrocytes and microglia in neuroinflammation, but the manner in which glial cells interact remains unclear. Our findings demonstrated the critical role of the astrocyte-specific protein, DICAM, as a paracrine modulator of microglia activation. DICAM is selectively secreted in reactive astrocytes via EVs and contributes to the regulation of neuroinflammation through the M2 polarization of microglia, which is dependent on the suppression of p38 MAPK signaling. Indeed, DICAM KO mice show enhanced nociceptive behavior and glial cell activation in the spinal cord of the CRPS model, suggesting its significant role in the central sensitization of pain ( Figure 6).

Discussion
A growing body of evidence has shown the cross-talk between astrocytes and microglia in neuroinflammation, but the manner in which glial cells interact remains unclear. Our findings demonstrated the critical role of the astrocyte-specific protein, DICAM, as a paracrine modulator of microglia activation. DICAM is selectively secreted in reactive astrocytes via EVs and contributes to the regulation of neuroinflammation through the M2 polarization of microglia, which is dependent on the suppression of p38 MAPK signaling. Indeed, DICAM KO mice show enhanced nociceptive behavior and glial cell activation in the spinal cord of the CRPS model, suggesting its significant role in the central sensitization of pain ( Figure 6). Cross-talk between astrocytes and microglia critically contributes to CNS homeostasis in response to damage or degeneration of neurons [31]. Although DICAM is predominantly expressed in astrocytes, immunologic stimuli with LPS and IFN-γ did not show any differences in the signaling cascade of DICAM KO and WT astrocytes. Instead, our data showed an activation of p38 MAPK by LPS and IFN-γ in DICAM KO mixed glial cells, which consisted of 20% of microglia, and 80% astrocytes [32]. Together with the Cross-talk between astrocytes and microglia critically contributes to CNS homeostasis in response to damage or degeneration of neurons [31]. Although DICAM is predominantly expressed in astrocytes, immunologic stimuli with LPS and IFN-γ did not show any differences in the signaling cascade of DICAM KO and WT astrocytes. Instead, our data showed an activation of p38 MAPK by LPS and IFN-γ in DICAM KO mixed glial cells, which consisted of 20% of microglia, and 80% astrocytes [32]. Together with the increase in M1 microglia markers observed in our DICAM KO glial culture, this result suggests the altered activation of microglia in DICAM KO mice and a role of DICAM as a cross-talk molecule that regulates the activation of microglia. An increasing amount of evidence revealed that cross-talk between astrocytes and microglia is mainly mediated by secretory mediators, such as cytokines, growth factors and neurotransmitters [31]. Moreover, EVs from reactive astrocytes play crucial roles in this reciprocal cross-talk, which can display a combination of neurotoxic or neuro-protective properties in a context-dependent manner [33]. However, very little is known about the molecules that regulate the EV-mediated cross-talk, which can act as a double-edged sword. Our results suggest that DICAM-enriched EVs from astrocytes can play a neuro-protective role by regulating the activation of microglia.
From the viewpoint of interaction with microglia, we need to investigate the ways in which DICAM-enriched EVs from astrocytes work. Given that DICAM is mainly located at the membrane fraction of cells [25], it is more likely to be present in the membrane of EVs rather than in the cytoplasm. Our group reported that DICAM is involved in cell adhesion through homophilic interaction with DICAM itself and heterophilic interaction with αVβ3 integrin [22]. Recent studies have shown that αVβ3 integrin increased in IL-1β-stimulated astrocyte-derived EVs, and blocking of αVβ3 integrin suppressed the uptake of astrocytederived EVs into neurons, suggesting the role of αVβ3 integrin in EVs transfer [34,35]. This evidence suggests that DICAM expressed in the exosomal membrane can interact with the αVβ3 integrin of microglia. This interaction between DICAM and αVβ3 integrin can promote the internalization of EVs into microglia, or is directly involved in the activation of αVβ3 integrin signaling. It needs further investigation to confirm the interaction between DICAM and αVβ3 integrin in the EV-mediated microglia regulation.
DICAM-deficient mixed glial cells showed significantly enhanced phosphorylation of p38 MAPK by LPS and IFN-γ stimulation compared to WT, which can be responsible for the proinflammatory phenotypes in the mixed glial cell culture and CRPS model in DICAM KO mice. As this phenotype was not observed in that of the astrocyte culture, the microglia may contribute to this enhanced p38 signaling. Indeed, p38 MAPK signaling was activated in microglia of the spinal cord after peripheral nerve injury, and pharmacologic inhibition of p38 MAPK suppressed tactile allodynia [36,37]. A recent study by Perea et al. revealed that p38 activation in aged mice occurred mainly in microglia, but not in astrocytes and neurons [38]. Moreover, they identified a subpopulation of microglia with low levels of p38 activation, which were rod-shaped and had a less activated phenotype [38]. In an inflammatory milieu, the activation of microglial p38 MAPK is critical for the production of inflammatory cytokines, such as IL-1β and TNFα, through its downstream targets of MAPK-activated protein kinase 2 (MK2) and mitogen-and stress-activated kinase 1 (MSK1) [39]. The inflammatory role of p38 MAPK is also mediated by autophagy suppression in microglia through the inactivation of UNC51-like kinase-1 (ULK1), which is involved in the initiation of the autophagic cascade [40,41]. This raises the following question: how is DICAM-exosome machinery involved in the inhibition of p38 MAPK? Although sufficient data on this topic do not exist, microRNAs such as miR-21 or miR-451 are known to suppress p38 MAPK activation [42,43]. Our findings imply that DICAM-rich EVs can easily transfer these p38-inhibitory miRs into microglia, and consequently promote anti-inflammatory M2-polarization of microglia.
Neuroinflammation characterized by "glial activation" is crucially involved in synaptic plasticity through inflammatory mediators, and consequently in the development of central sensitization [44]. The central glia, especially microglia and astrocytes, have different patterns of activation; microglial activation in the spinal cord is very rapid and dramatic, whereas astrocyte activation is more persistent and occurs in more painful conditions [4,45]. These findings suggest a critical role of microglia in the initiation of central sensitization. Indeed, a study by Ikeda et al. reported that the pharmacologic inhibition of microglia and p38 MAPK with minocycline and SB203580, respectively, attenuated the development of central sensitization in rats subjected to peripheral nerve injury [46]. This result suggests that microglial activity control with DICAM can be an effective strategy for the inhibition of neuroinflammation and subsequent central sensitization.
Our data showed declining trends of the serum DICAM levels in CRPS patients compared to the controls. A previous study investigated the serum level of inflammatory mediators and revealed that inflammatory cytokines, such as IL-8 and sTNFR, and substance P were signifi-cantly increased in CRPS type I patients compared to the controls, but adhesion molecules, such as E-selectin, L-selectin, and P-selectin, decreased in CRPS patients [47]. Although it is unknown why the serum levels of adhesion molecules decrease in CRPS patients, DICAM, as an immunoglobulin superfamily cell adhesion molecule, showed a decreasing trend in the serum of CRPS patients. Congruent with our results of the microglia-inhibition effects of DICAM, the decreased serum level of DICAM may be associated with microglia activation in CRPS patients. Further study is needed to confirm whether DICAM levels as a serum marker can predict the severity or clinical course of CRPS.
In conclusion, our findings highlighted that DICAM secreted via astrocyte-derived EVs plays an important role in the M2 polarization of microglia through the suppression of p38 MAPK signaling, and consequent attenuation of neuroinflammation during central sensitization. Further investigations are warranted to evaluate not only its significance as a surrogate marker, but also its potential as a therapeutic target for central sensitization.