Abstract
Inflammation resolution is regulated by specialized pro-resolving lipid mediators (SPMs) and the levels of SPMs are found decreased in Alzheimer’s disease (AD) brain. We have previously found that one of the SPMs, Maresin1 (MaR1), improved neuronal survival and increase microglial phagocytosis of amyloid-β 1-42 (Aβ42); however, the mechanisms underlying the protective mechanism remain further investigation. We aim to investigate the effects of MaR1 on microglial chemotaxis and activation in this study. Both indirect and direct primary neuron and microglia co-culture system was used in this study. Our results showed MaR1 downregulated the increased microglial chemotaxis induced by Aβ42. The microglial inactivation marker CD200R was downregulated by Aβ42 and upregulated by MaR1. Pro-inflammatory cytokines secretion such as tumor necrosis factor (TNF)-α were increased by Aβ42 and these changes were revised by MaR1 treatment. In addition, the levels of chemokine monocyte chemoattractant protein (MCP)-1 were increased while the levels of anti-inflammatory factor IL-10 secretion were decreased by Aβ42, and these changes were abolished by MaR1 treatment. Moreover, by proteomics analysis, we identified cell signaling pathways affected by MaR1 were not only limited to inflammation-related pathways such as P38, but also in pathways involved in cell survival, autophagy, axon formation, and apoptosis, including PI3K/AKT, mTOR, ERK, caspase3, Cdc42, and p75NTR. In conclusion, MaR1 promoted inflammation resolution by inhibiting chemotaxis and regulating activation of microglia. MaR1 played a neuroprotective role by affecting cell signaling pathways involving inflammation, cell survival, autophagy, axon formation, and apoptosis inhibition.
INTRODUCTION
Alzheimer’s disease (AD) is the most common cause of dementia. Besides aggregated extracellular amyloid-β (Aβ), intracellular paired helical filaments of hyperphosphorylated tau, and loss of neurons [1], microglia, and astrocyte activation, as well as increased production of inflammatory factors are also a typical pathologic changes in postmortem AD brain [2–4]. Failure of inflammation resolution leads to chronic inflammation and may contribute to AD progression [5].
Resolution of inflammation is an active and highly regulated biochemical process mediated by specialized pro-resolving lipid mediators (SPMs), which are biosynthesized from omega (ω)-3 and ω-6 polyunsaturated fatty acids (PUFAs) via lipoxygenases (LOXs) and cyclooxygenases (COXs) [6]. SPMs are synthesized by leukocytes, macrophages, and organs and tissues during the resolution stage of inflammation [7]. Four classes of SPMs have been discovered, including lipoxins (LXs), resolvins (Rvs), protectins (NPDs), and maresins (MaRs) [8]. The biological functions of SPMs were partly elucidated in the peripheral inflammatory disease models, including cessation of neutrophil infiltration, decreasing production of pro-inflammatory mediators, increasing macrophage uptake of apoptotic neutrophils and cellular debris, and promoting macrophage transformation from M1 to M2 phenotype, and so on [9, 10].
Evidence from recently published studies suggest that resolution of inflammation is impaired in AD: The levels of MaR1, RvD5, and NPD1 were lower in AD brain [5, 11–13] compared to mild cognitive impairment (MCI) and subjective cognitive impairment (SCI) patients; the levels of LXA4 hippocampus were significantly lower in the cerebrospinal fluid (CSF) of AD patients compared to controls. Moreover, the levels of LXA4 and RvD1 in CSF were correlated with Mini-Mental State Examination (MMSE) scores, suggesting that promoting resolution of inflammation might prevent AD-related cognitive decline [5].
Indeed, treatment studies applying SPMs on both animal and cellular studies have been carried out in several laboratories. LXA4 has been shown protective in several AD-related mouse models including the intracerebroventricular injection of Aβ42, Tg2576 amyloid precursor protein (APP) transgenic [14], and in the 3xTg-AD mice models [15]. Combination treatment with RvE1 and LXA4 has been shown reversed the inflammatory process, and decreased the neuroinflammation associated with Aβ pathology in 5xFAD mice [16]. Protective effects were also found in several cellular models. LXA4 was reported downregulated the production of pro-inflammatory cytokines IL-1β and TNF-α, inhibited translocation of NF-κB p65 subunit into the nucleus in Aβ-stimulated BV2 cells. RvD1 was found enhanced the phagocytosis of Aβ and inhibited fibrillar Aβ-induced apoptosis in by macrophages derived from AD patients. PD1 has been reported improved neuronal survival in neuro-glia co-culture challenged with Aβ [11]. In addition, all of the four types of SPMs have been showed improved neuronal survival and only MaR1 could increase microglial phagocytosis of Aβ42 [7]. All these results suggest that promoting inflammation resolution by MaR1 may be a new therapeutic method for AD.
However, studies investigating the effects of MaR1 on AD animal and cellular models and its protective mechanisms are still lacking; therefore, we aim to investigate the effect of MaR1 on the chemotaxis, activation, phenotype, and inflammatory cytokines secretion of microglia. The cell signaling pathways influenced by MaR1 were also investigated.
MATERIALS AND METHODS
Primary culture of neurons
Primary cortical neurons were obtained from 16–18 days’ embryos of pregnant C57BL6/J mice. Pregnant mice were euthanized by isoflurane and cerebral cortices of fetuses were removed and dissociated by mild trypsinization (0.125% concentration), followed by termination of digestion with complete microglia culture medium which contained fetal bovine serum. The cells were suspended in neurobasal medium supplemented with 2% B-27 supplement, 1% L-glutamine, and 0.5% penicillin/streptomycin (all from Life Sciences, USA). Cell suspension was seeded in 6 multi-well plates pre-coated with poly-D-lysine (Beyotime, China) at a concentration of 1×105 cells/well, and then cells were cultured in a 95% air and 5% CO2 atmosphere incubator at 37°C. Half culture medium was replaced on the second day and then on every 3-4 days. Neurons were used for experiments after 6 days.
Primary culture of microglia
Mouse brain mixed glial cells were prepared from whole brains of 1–3 day postnatal C57BL6/J mice and dissociated with a mild mechanical trituration. Cells were seeded in the cell culture bottles (75cm2) pre-coated with poly-D-lysine. The culture medium was DMEM-High glucose supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin (all from Life Sciences, USA), and 0.8% insulin. Microglia were further extracted from the mixed glial cell cultures by mild trypsinization digestion method as previously described [17].
Microglia chemotaxis method
Pure microglia cells were harvest and stained with 5-(and-6)-Carboxyfluorescein Diacetate, Succinimidyl Ester (CFSE, Life Sciences, USA) by incubating for 30 min at 37°C at the final concentration of 10 μmol/L. Afterwards, the cells were washed once with 1X PBS, and then the cells were re-suspended with phenol red free DMEM/F12 (Life Sciences, USA) at a concentration of 1×105 cells/ml. The transwell cell culture inserts (EMD Millipore, USA) were placed in each well of the 6 multi-well plates which contained primary cultured neurons on the bottom. Then the microglial cells were stained with CFSE and seeded into the transwell cell culture inserts. The ratio of microglia to neurons was 2:1.
To investigate the effects of MaR1 on microglial chemotaxis, different doses of MaR1 (Cayman Chemical, USA) whose final concentration ranged from 0.005–0.5 μmol/L were added to the upper culture medium. Medium contains the same concentration of ethanol was used as vehicle. Aβ42 (Abcam, UK, ab120301) was dissolved with DMSO at the final concentration of 1 mg/ml. To further analyze the effect of MaR1 on microglial chemotaxis under the inflammation stimulation of Aβ42, Aβ42 (final concentration 5 mg/L), and different concertation of MaR1 (final concentration ranged from 0.005–0.5 μmol/L) were added to the upper culture medium. After 6, 16, and 24 h culture, the microglia that migrated to the lower part of the neuronal culture including the cells in the cultured medium and the cells that were attached to the bottom of 6-multi-well plates were harvest and washed once with sterile 0.1 M phosphate-buffered saline (PBS). Then the cells were fixed in 300 μl 1% formaldehyde and the fluorescence of CFSE was analyzed by flow cytometry (LSRFortessa, BD Biosciences, USA).
Co-cultures of neurons and microglia and stimulation by Aβ42 and MaR1
Microglia cells were suspended with neuron culture medium which consists of neurobasal medium supplemented with 2% B-27 supplement, 1% L-glutamine, and 0.5% penicillin/streptomycin at a concentration of 1×105 cells/ml, and the cell suspension was added to each neuron cultured well. The ratio of microglia to neurons was 2:1. Co-cultures were grown for 24 h, and the medium was replaced with fresh medium before the drug treatment. The cells were treated with Aβ42 (5 mg/L), MaR1 (0.5 μmol/L), combination of Aβ42 (5 mg/L), and MaR1 (0.5 μmol/L) separately, and medium containing the same concentration of DMSO and ethanol were used as vehicle. Six hours later, the supernatant was collected for cytometric beads array (CBA) analysis, and the cells were also harvested for flow cytometry analysis as described below.
Immunocytochemistry
Primary neurons and/or microglia were fixed with 4% formaldehyde contained with 0.2% Triton X-100 for 30 min. After blocking with 1% bovine serum albumin (BSA) at room temperature for 1 h, cells were incubated first with primary antibodies overnight at 4°C. Microglia and neurons were stained with rabbit polyclonal antibody to CD11b (Abcam, UK, ab128797) and chicken monoclonal antibody to MAP-2 (Abcam, UK, ab92434) respectively, then followed by Alexa-conjugated (–488, or –647) anti-rabbit or anti-chicken secondary antibodies for 2 h, lastly with DAPI for 30 min. Cells were then analyzed by confocal microscopy (Olympus Fluoview, FV1000, Japan).
Flow cytometry analysis
For analysis of membrane-associated cellular markers, the co-cultured cells were harvested and then re-suspended with PBS containing 1% BSA. Cell suspensions were incubated for 30 min on ice with fluorophore-conjugated antibodies against CD40, CD183, CD200, and CD200R (all from Abcam, UK) at their respective working concentrations. The cells were then washed once with PBS and then re-suspend with 1% paraformaldehyde at room temperature for 45 min. The cells were analyzed by the LSRFortessa flow cytometer machine.
Cytometric beads array
The levels of tumor necrosis factor (TNF)-α, interleukin (IL)-6, IL-10, monocyte chemoattractant protein (MCP)-1, interferon (IFN)-γ, and IL-12p70 in the co-culture cell medium were measured by a CBA kit (BD Biosciences, USA, 552364) according to the manufacturer’s manual. Cytokine levels were then quantified by flow cytometry. In brief, a total sample volume of 50 μl supernatant of the four groups were mixed with 50 μl of antibody-coated capture beads and 50 μl of PE-conjugated and FITC-conjugated detection antibodies. This mixture was incubated for 2 h in dark at room temperature. The samples were washed with 1 ml buffer at 200 g for 5 min and resuspended in 300 μl wash buffer. The cytometric beads were analyzed by a LSRFortessa flow cytometer machine (BD Biosciences, USA) and the data were analyzed using FCAP array software (BD Biosciences, USA).
Proteomics
The proteomics was done as previously described [17]. Firstly, the samples were prepared. The cell samples were frozen in liquid nitrogen until use and there were three replicates for each group. Samples containing 100 μg ground materials were mixed with an equal volume of 3X protein extraction solution mix (BestBio, China), containing solution A (protein extraction solution) and solution B (protease inhibitor) at a ratio of A to B 500:1 (v/v). The mixture was further ground to fine and transferred into 1.5 ml Eppendorf Protein LoBind Tubes. After centrifugated at 4°C for 20 min at 150 rpm, the samples were further centrifuged at 14000 rpm for 15 min, and total protein content in the supernatant was measured using BCA protein assay kit (Beyotime, China). The crude protein extracts underwent purification of protein by trichloroacetic acid (TCA)-acetone method and reduction acetylation. Then acetylated protein solution was treated with Sequencing Grade Modified Trypsin (enzyme to protein ratio about 1:40) to generate protein tryptic fragments and filtrated with membrane (Sartorius, Germany) to remove peptides larger than 10 kDa. The final protein concentration was about 1 μg/μl and was stored at –80°C until use. Secondly, shotgun proteomics were measured. Peptides of all samples were pooled and measured by AB Sciex 5600+ TripleTOF platform configured with Eksigent 400nano-HPLC system and 5600 + TripleTOF mass spectrometer (Eksigent, AB Sciex, USA). Thirdly, proteomic analysis was done with Sequential Window Acquisition of all Theoretical fragment ion spectra (SWATH)–Mass spectrometry (MS) method. Every sample repeated SWATH data collection for three times. Finally, data were processed and bioinformatics analyzed. The pooled peptide data acquired in data dependent acquisition (DDA) model were analyzed by Paragon (ProteinPilot software, AB Sciex, v. 5.0.0.0, 4769) against database Mus musculus sequences from UniProt (version 2015_04). An ion library was generated from data acquired in DDA model for targeted extraction in SWATH model. Linear regressions were conducted on the retention time (RT) from different data sources between DDA data and the SWATH data using PeakView SWATH Processing Micro App (AB Sciex, v.1.0.0.1409). All of the RTs of the peptides from DDA were globally corrected resulting in a new ion library consisting of the peptides with corrected RTs. Protein peak area data were exported and processed with Markerview (AB Sciex, 1.2.1.1). Principle component analysis (PCA) and T-test were performed among the four groups. MS in which fold change was set as ≥1.5 or ≤0.067, and p-value less than 0.05 was chosen after area normalization and differentially expressed proteins were obtained. Then differentially expressed proteins were annotated by Gene Ontology Consortium (GO) and their systematic information was computed by Kyoto Encyclopedia of Genes and Genomes (KEGG).
Western blot
The outcomes of proteomics were verified by western blot. The co-cultured cells were lysed with radio-immunoprecipitation assay (RIPA) buffer, supplemented with 1% protease inhibitor cocktail and 1% phosphatase inhibitor cocktail (Sigma-Aldrich, USA), and centrifuged at 12,000 rpm at 4°C for 20 min. The supernatant was collected and total protein concentration was determined by a BCA assay kit. Briefly, samples containing 40 μg protein each were mixed with equal volume of 2X Laemmli sample buffer, and were boiled at 95°C for 5 min. The denatured samples were then loaded on a 10% SDS-PAGE gel, and then the proteins were transferred to a nitrocellulose membrane under 85 mA current over night at 4°C. The membranes were blocked with 5% non-fatty dry milk at room temperature for 45 min, and then incubated with the following antibodies at their working concentrations: anti-actin (1:2500; Abcam, UK, ab8224), anti-LaminA (1:1000; Abcam, UK, ab26300), anti-GAPDH (1:10000; Abcam, UK, ab181602), anti-phospho-PI3K (1:1000; Abcam, UK, ab182651), anti-PI3K (1:1000; Abcam, UK, ab191606), anti-phospho-AKT (1:5000; Abcam, UK, ab81283), anti-AKT (1:2000; Abcam, UK, ab28422), anti-phospho-p38 (1:1000; Abcam, UK, ab195049), anti-p38 (1:2000; Abcam, UK, ab170099), anti-phospho-ERK1+ERK2 (1:1000; Abcam, UK, ab201015), anti-ERK1+ERK2 (1:10000; Abcam, UK, ab184699), anti-p75NTR (1:50000; Abcam, UK, ab52987), anti-Cdc42 (1:1000; Abcam, UK, ab64533), anti-caspase3 (1:5000; Abcam, UK, ab184737), anti-phospho-mTOR (Ser2448) (1:1000; Cell signaling, USA, 2971), and Anti-mTOR (1:1000; Cell signaling, USA, 2972) in Tris-buffered saline with 0.1% Tween20 (TBS-T) at 4°C overnight. After incubation with appropriate secondary antibodies, antigens were revealed by ECL chemiluminescence system (Amersham, UK) and signal quantification was achieved using ImageJ software (NIH, Bethesda, MD, USA).
Statistical analysis
All values are expressed as the mean±SEM. Normal distributions and homogeneity of variance were found for all analyzed categories. All statistical analyses were conducted using GraphPad Prism 7 software (GraphPad Software, La Jolla, CA, USA). Unmatched t-test was used in comparison between the two groups, and oneway-ANOVA analysis was used in multi-group pairwise comparison. In all instances, statistical significance was defined as follows: *p < 0.05, **p < 0.01, ***p < 0.001.
RESULTS
Aß42 increased the chemotaxis of microglia and MaR1 decreased the chemotaxis of microglia both with or without Aß42 stimulation
Compared with Vehicle group, the number of chemotactic microglia in Aβ42 group increased significantly at 6 h after Aβ42 administration (p < 0.05). The numbers of chemotactic microglia in Aβ42 + MaR1 groups (MaR1 final concentrations ranged from 0.005–0.5 μmol/L) were lower than that in Aβ42 group (p < 0.05) (Fig. 2A). Compared with Vehicle group, the numbers of chemotactic microglia in different concentrations of MaR1 groups (MaR1 final concentrations ranged from 0.005–0.5 μmol/L) without Aβ42 decreased, but only in low-concentration MaR1 group (0.005 μmol/L) and high-concentration MaR1 group (0.5 μmol/L), the number of chemotactic cells reached statistical significance (p < 0.05) and the decrease of chemotaxis in high concentration MaR1 group (p < 0.01) was more obvious than that in low concentration group (p < 0.05) (Fig. 2B). At 16 and 24 h, there was no significant difference in the degree of microglia chemotaxis among the four groups (data not shown).
Primary neuron and microglia cells culture. A) Micrograph of primary cultured neurons stained with antibodies against MAP-2. B) Micrograph of primary cultured microglia stained with antibodies against CD11b. C) Micrograph of co-cultured neurons and microglia. Scale bar = 10 μm.
The effect of MaR1 on chemotaxis of microglia. A) Aβ42 increased the chemotaxis of microglia and MaR1 decreased the chemotaxis of microglia induced by Aβ42 in a concentration-independent manner (p < 0.05). B) In the absence of Aβ42 stimulation, MaR1 also reduced the chemotaxis of microglia. The decrease of microglial chemotaxis by MaR1 was more obvious in high concentration (0.5 μM) (p < 0.01) than in low concentration (0.005 μM) (p < 0.05). Data were expressed by mean±SEM, and statistical significance was defined as follows: *p < 0.05, **p < 0.01, ***p < 0.001. Independent experiments n = 3 for each group.
MaR1 increased CD200R expression
In the experiment described above, we found the all concentrations of MaR1 tested had effect on the chemotaxis of microglia; therefore, the highest concentration of MaR1 was used in the following experiment. The differences of chemotactic microglia numbers among the groups reached statistical significance only at 6 h, therefore, only 6 h time point was selected in the following experiment. The surface markers of microglia and neurons in Vehicle group, Aβ42 group, MaR1 group, and Aβ42+MaR1 group were detected by flow cytometry. Compared with Vehicle group, the expression of CD200R in Aβ42 group decreased (p < 0.05). The expression of CD200R in Aβ42+MaR1 group was significantly higher than that in Aβ42 group (p < 0.05) (Fig. 3). The expression of CD40, CD183, and CD200 did not show any significant difference among the four groups (data not shown).
The effect of MaR1 on CD200R expression of microglia. Aβ42 decreased CD200R expression, and MaR1 increased CD200R expression in the presence of Aβ42 stimulation (p < 0.05). Data were expressed by mean±SEM, and statistical significance was defined as follows: *p < 0.05, **p < 0.01, ***p < 0.001. Independent experiments n = 3 for each group.
MaR1 decreased the pro-inflammatory cytokine and chemokine secretion and increased anti-inflammatory cytokine secretion in the co-cultured cell model
We used CBA to analyze cytokines in the supernatant of the four groups (Fig. 4). Compared with Vehicle group, the secretion of pro-inflammatory cytokines TNF-α, IL-6, and chemokine MCP-1 increased and anti-inflammatory cytokine IL-10 decreased in Aβ42 group (p < 0.05). Compared with Aβ42 group, the secretion of TNF-α and IL-6 decreased in Aβ42 + MaR1 group (p < 0.05) and there was a trend of decrease of MCP-1 and trend of increase of IL-10 in Aβ42 + MaR1 group, but did not reach statistical significance. Compared with Vehicle group, IL-6 secretion decreased in MaR1 group (p < 0.01). IFN-γ and IL-12p70 did not show any significance difference among the four groups (data not shown).
The effect of MaR1 on the secretion of cytokines. A) Aβ42 increased TNF-α secretion and MaR1 decreased TNF-α secretion induced by Aβ42 (p < 0.05). B) Aβ42 increased IL-6 secretion (p < 0.01) and MaR1 decreased IL-6 secretion with or without Aβ42 stimulation (p < 0.05). C) Aβ42 increased MCP-1 secretion (p < 0.05). Compared with Aβ42 group, the level of MCP-1 showed a downward trend in Aβ42+MaR1 group, but did not reach statistical significance. D) Aβ42 decreased IL-10 secretion (p < 0.05). In the presence of MaR1, the level of IL-10 showed an upward trend, but did not reach statistical significance. Data were expressed by mean±SEM, and statistical significance was defined as follows: *p < 0.05, **p < 0.01, ***p < 0.001. Independent experiments n = 3 for each group.
Quantitative proteomics revealed molecular changes in the four different stimulus groups and western blot verification
By proteomic analysis of the four groups of co-cultured cells, we found many proteins were altered (Fig. 5). Then we verified the results of proteomics by western blot (Fig. 6). Compared with Vehicle group, the ratio of p-PI3K/t-PI3K and p-AKT/t-AKT were downregulated (Fig. 5A, B), while the ratio of p-mTOR/t-mTOR, p-p38/t-p38, caspase-3, p-ERK/t-ERK, p75NTR, as well as the levels of Cdc42 were upregulated in Aβ42 group (p < 0.05) (Fig. 5C-H); however, these Aβ42 induced changes were reversed by Aβ42 + MaR1 treatment except p-ERK/t-ERK (p < 0.05). The ratio of p-ERK/t-ERK was further increased by Aβ42 and MaR1 co-stimulation as compared to Aβ42 group (p < 0.01) (Fig. 5F).
Proteomics revealed the protein pathways affected by MaR1. A) Altered proteins in Aβ42 group compared with Vehicle group. Compared with Vehicle group, the increased proteins in Aβ42 group were expressed in red spots and the decreased proteins were expressed in blue spots. B) Altered proteins in Aβ42+MaR1 group compared with Aβ42 group. Compared with Aβ42 group, the increased proteins in Aβ42+MaR1 group were expressed in red spots and the decreased proteins were expressed in blue spots. Independent experiments n = 3 for each group.
Verification of proteomic results by western blot. Compared with Vehicle group, the ratio of p-PI3K/t-PI3K and p-AKT/t-AKT were downregulated (A, B), while the ratio of p-mTOR/t-mTOR, p-p38/t-p38, caspase3, p-ERK/t-ERK, p-75NTR as well as the levels of Cdc42 were upregulated in Aβ42 group (p < 0.05) (C-H); however, these Aβ42 induced changes were revised by Aβ42+MaR1 treatment except p-ERK/t-ERK. The ratio of p-ERK/t-ERK was further increased by Aβ42 and MaR1 co-stimulation compared to Aβ42 group (p < 0.01) (F). Data were expressed by mean±SEM, and statistical significance was defined as follows: *p < 0.05, **p < 0.01, ***p < 0.001. Independent experiments n = 3 for each group.
DISCUSSION
Evidence has shown that failure of inflammation resolution leads to chronic inflammation and contribute to AD. Inflammation resolution is conducted by SPMs, and SPMs are decreased in different areas of the AD brain [11]. The levels of LXA4 and RvD1 in cerebrospinal fluid has correlation with MMSE score in AD patients, therefore, SPMs may play an important role in maintaining the cognitive function. All these data suggest that stimulation of pro-resolving activities may be a therapeutic strategy for AD. In the present study, we aim to further explore the effects of MaR1 on chemotaxis, activation, secretion of inflammatory cytokines by microglia and the protective mechanisms of MaR1.
Microglia are the resident immune cells in the central nervous system. In AD, microglia initiate the inflammatory response by migrating to surround Aβ plaques [18, 19] and these microglia mainly play a pro-inflammatory role [18, 20]. The inflammatory cytokines produced by microglia and Aβ all can cause microglia to be dysfunctional in phagocytic capacity [21, 22]. In this experiment, we found Aβ42 increased the chemotaxis of microglia and MCP-1 secretion. MaR1 could decrease microglia chemotaxis with or without the presence of Aβ42. Consistent with our results, using Ccr2RFP/+Cxcr1GFP/+ mice for macrophage and microglia, MaR1 has been demonstrated decreased macrophage infiltration in the hippocampus and attenuated inflammation in a perioperative neurocognitive disorder model [23]. Therefore, we speculate that MaR1 alleviate inflammation by decreasing chemokine and further inhibiting chemotaxis of microglia. Although the reduction of MCP-1 level did not reach statistical significance, the reason may due to other chemokines may affect microglial chemotaxis besides MCP-1.
Consistent with our findings, Aβ is known to stimulate microglia to secret pro-inflammatory cytokines, including IL-1, IL-6, and TNF-α [24, 25]. TNF-α can reduce neuronal viability, cause blood-brain barrier impairment and amyloidosis [26]. Abrogation of TNF signaling receptors can improve cognitive impairments [27]. IL-6 can accelerate plaque formation [28] and tau hyperphosphorylation [29]. We found MaR1 could decrease TNF-α and IL-6 induced by Aβ42, and MaR1 could even decrease IL-6 without Aβ42 stimulation. IL-10 is an anti-inflammation factor that is involved in decreasing Aβ42 load and increasing phagocytosis of Aβ42 [29]. In a rat spinal cord ligation model of neuropathic pain, Gao et al. showed intrathecal MaR1 attenuated mechanical allodynia associated decreased pro-inflammatory cytokine levels including TNF-α and IL-6, indicating that MaR1 could downregulate inflammation via inhibiting pro-inflammatory cytokines secretion [30]. In the present study, we also found IL-10 decreased by Aβ42 stimulation and MaR1 could possibly reverse this change; therefore, MaR1 may play an immunomodulatory role and promotes inflammation resolution through regulating IL-10 secretion.
CD200 is expressed on neurons in the central nervous system [31], while CD200R which is the receptor of CD200 is predominantly expressed on microglia [32, 33]. CD200 exerts inhibitory effect on CD200R, so as to downregulate activation of microglia. Blocking CD200R result in neurodegeneration in Parkinson’s disease mouse models [34]. CD200 and CD200R expression levels are decreased in AD hippocampus and inferior temporal gyrus [35]. In our co-cultured model of neurons and microglia, Aβ42 also decreased CD200R expression, but has no obvious effect on neuron CD200 expression. The enhanced CD200R expression due to MaR1 may limit deleterious neuroinflammation caused by microglia. Above all, MaR1 reduce inflammatory response and promote inflammation resolution through reducing microglia chemotaxis, inhibiting microglia activation, decreasing the secretion of pro-inflammatory cytokines and chemokines and increasing the secretion of anti-inflammatory factor. The changes of chemotaxis among the four groups described above occurred only at 6 h, but not at 16 and 24 h; the reason may be that MaR1 is a lipid molecule which may be degraded quickly and can only work for a short time.
Even though MaR1 has been shown protective different disease models including perioperative cognitive disorders, spinal cord injury, and neuropathic pain [23, 36]. The underlying mechanisms of its protective effects still needs to be further investigated. Unlike other types of SPMs, the receptor of MaR1 is still unknown. A recent study using high fat diet induced nonalcoholic steatohepatitis model showed MaR1 is a ligand to Retinoic acid-related orphan receptor α (RORα). Interestingly, RoRα can increase MaR1 synthesis in a 12-LOX dependent manner [37]. However, the protective mechanism in AD related models is largely unexplored. To further elucidate the protective mechanisms of MaR1, we used proteomic analysis to identify protein alterations among the different stimulus groups and then verified the key results using western blot. Multiple proteins pathways were altered, including phosphoinositide 3-kinases (PI3K), protein kinase B (AKT), mechanistic mammalian target of rapamycin (mTOR), p38, extracellular signal-regulated kinase (ERK)1/2, caspase3, cell division cycle (Cdc)42, and p75 neurotrophin receptor (p75NTR). Besides inflammation, these proteins are involved in cellular survival, proliferation, autophagy, apoptosis, and axon formation as well.
We observed Aβ42 inhibited PI3K/AKT signaling, but MaR1 could reverse this change in the present study. PI3K/AKT signaling is involved in cell survival and has been reported involved in neurodegenerative diseases including amyotrophic lateral sclerosis, Parkinson’s disease, and AD [38, 39]. PI3K/AKT signaling has been found downregulated in the brains of AD patients [40]. PI3K/ AKT can activate its downstream factors mTOR, Nrf2 [41], which are involved in collateral axon branching, autophagy, and alleviating oxidative damage, and inhibit downstream factors caspase3 and glycogen synthase kinase (GSK)-3β which are involved in apoptotic and tau protein phosphorylation [42]. Hence, PI3K/AKT can play a therapeutic role in neurodegenerative diseases [43]. Moreover, all the four kinds of SPMs can enhance PI3K/AKT expression to alleviate the pathological injury and promote cell survival in both in vivo and in vitro disease models [44, 45]. Therefore, our results suggest that MaR1 could promote neuron survival by enhancing PI3K/AKT signaling.
mTOR hyperactivity has been observed in AD brains. Hyperactivation of mTOR results in reduced autophagy leading to the accumulation of protein aggregates [46, 47]. Treatment by mTOR inhibitors could alleviate AD pathology and improve cognitive performances [48, 49]. The SPMs’ precursor omega-3 PUFAs could inhibit mTOR and beneficial for ischemia [50]. Consistently, we observed the protein levels of mTOR were upregulated by Aβ42 and MaR1 reversed this change, indicating that MaR1 could enhance autophagy by inhibiting mTOR. The reason why mTOR was enhanced in AD might attributed to its upstream factor Ras/ERK. We found Aβ42 could activate ERK, and MaR1 could further enhance this change. ERK is a member of the MAPKs family, while the other two are p38 and c-JunNH2-terminal kinase (JNK). ERK1/2 activation regulates synaptic protein and promotes new dendritic spines formation. It also has an impact on determining LTP induction in the hippocampus [51, 52]. Many studies have shown that activation of the ERK/MAPK and PI3K/AKT pathways could revert learning and memory impairments in AD [53–55]. But Aβ42 activating ERK pathway in the hippocampus can also lead to caspase activation [56], aberrant hyperphosphorylation of tau [57] and mTOR enhancement as discussed above, therefore, the enhancement of ERK expression induced by Aβ42 and MaR1 may be a double-sword in AD, can bring both beneficial and detrimental effects.
In addition, we found Aβ42 activated p38 MAPK and MaR1 revised the upregulated p38 induced by Aβ42. p38 MAPK activation is significantly involved in microglia and astrocyte activation and subsequently promotes neuroinflammation in AD [58–60]. Inhibition of p38 MAPK can effectively alleviate chronic inflammatory diseases such as rheumatoid arthritis, cardiovascular disease, and inflammatory pain [61–64], and p38 MAPK inhibitors have been considered as novel and potential therapeutics for neurodegenerative diseases including AD [65]. Therefore, MaR1 play a protective role via inhibiting p38 MAPK mediated inflammation. All the pathway mentioned above, including PI3K/AKT, ERK, and p38 can affect caspase3 which plays an extremely important role in neuronal apoptosis and is considered the terminal event preceding cell death. Caspase3 levels are higher in AD brains than in age-matched controls [66]. Our results showed Aβ42 induced the upregulation of caspase3, and MaR1 reversed this change, suggesting that MaR1 could protect neurons by anti-apoptosis. It is highly likely that MaR1 inhibited p38 pathway that was associated with apoptosis, and enhanced PI3K/AKT pathway which could protect cells from apoptosis [67].
p75NTR and Cdc42 are other two significantly altered proteins in this experiment. p75NTR is expressed exclusively on basal forebrain cholinergic neurons and is a low affinity receptor for neurotrophins such as nerve growth factor (NGF), neurotrophin (NT)-3, NT-4 and brain-derived neurotrophic factor (BDNF), and high-affinity receptor for precursor NGF (pro-NGF). p75NTR is upregulated in AD brain and Aβ is a positive signal to upregulate p75NTR expression [68, 69]. In AD brain, neurodegenerative substances such as Aβ, pro-NGF, Nogo, and other factors exert detrimental effects via p75NTR signaling, resulting in Aβ overexpression, neuronal death, neurite degeneration, tau hyperphosphorylation, cell cycle re-entry, and cognitive decline [68, 69]. p75NTR antagonists [70, 71], vaccines [72, 73], and recombinant extracellular domain of p75NTR [69] are able to block the interaction of Aβ, pro-NGF, and other neurodegenerative ligands with p75NTR, thus to protect neurons from neuronal toxicities. In this experiment, we found that MaR1 could decrease p75NTR expression induced by Aβ42. Therefore, MaR1 could protect neurons from degeneration by reducing P75NTR expression.
Cdc42 belongs to the Rho-family GTPases family and is involved in axon formation, growth cone filopodia formation, and cell cycle re-entry [69]. Expression of constitutively active Cdc42 or hyperactivated Cdc42 in rat hippocampal neurons result in inhibition of neurite extension and formation [74]. Cdc42 levels have been reported to be elevated in select neuronal populations in AD brains compared with age-matched controls [75] and in hippocampal neurons treated with Aβ [76]. Recently, a study showed circulating Cdc42 in the plasma was altered in frontotemporal lobar degeneration (FTLD) and a specific decrease of Cdc42 expression had correlation with the behavioral improvement [77]. We found MaR1 could decrease the elevated levels of Cdc42 induced by Aβ42, therefore, MaR1 may play an important role in neuron repair.
In conclusion, MaR1 could alleviate inflammation by decreasing microglia chemotaxis, modifying microglia activation, decreasing chemokine and pro-inflammatory cytokines whereas increasing anti-inflammation cytokines production. MaR1 could also play neural protective roles by promoting cellular survival, autophagy, axon formation and inhibiting apoptosis signaling pathway.
