Overexpression of miR-146a Might Regulate Polarization Transitions of BV-2 Cells Induced by High Glucose and Glucose Fluctuations

Microglia are critical in neuroinflammation. M1/M2 polarization transitions of microglial phenotypes determine the states of neuroinflammation and are regulated by multiple pathways, including miRNAs and other epigenetic regulations. This study investigated the polarization transitions of microglia induced by high glucose and glucose fluctuations, and the role of miR-146a in regulating M1/M2 polarization transitions of microglia. BV-2 cells were cultured with 25 mmol/L glucose, 75 mmol/L glucose, and 25 mmol/L−75 mmol/L glucose fluctuation for 48 h. BV-2 cells overexpressing miR-146a were generated using a lentiviral vector. Quantitative real-time polymerase chain reaction (qRT-PCR) was used to measure mRNA expression of miR-146a, CD11b, iNOS, Arg-1, IRAK1, TRAF6, and NF-κB. Immunofluorescence was used to measure CD11b expression. Western blot was used to measure protein expression of CD11b, iNOS, and Arg-1. Compared with those in the 25 mmol/L glucose control group, expression of CD11b, iNOS, TNF-α, and IL-6 in the 75 mmol/L glucose or glucose fluctuation groups of cultured BV-2 cells were significantly increased, while Arg-1 and IL-10 was significantly decreased. These effects were reversed by overexpression of miR-146a. Furthermore, expression of IRAK1, TRAF6, and NF-κB was significantly increased in the high glucose and glucose fluctuation groups; this was reduced after miR-146a overexpression. In sum, high glucose and glucose fluctuations induced polarization transitions from M1 to M2 phenotype in BV-2 cells. Overexpression of miR-146a might protect BV-2 cells from high glucose and glucose fluctuation associated with M1/M2 polarization transitions by downregulating the expression of IRAK1, TRAF6, and NF-κB.


INTRODUCTION
Diabetic encephalopathy (DE) is a chronic complication of diabetes mellitus (DM), which can cause cognitive decline, dementia, and mental disorders, which significantly affects the quality of life of diabetic patients (1). Since the concept of DE was proposed by Nieleson in 1965, our understanding of DE has deepened. Its pathogenesis involves various aspects, including neuroinflammation (2, 3), blood-brain barrier (BBB) (2), vascular factors (4), insulin resistance, insulin deficiency (5,6), and oxidative stress (3,6). However, controversy in the literature remains. Neuroinflammation plays an important role in the pathogenesis of DE (7). Neuroinflammation is a complex innate immune response of the central nervous system that inhibits infection; removes pathogens, cell debris, and misfolded proteins; and plays an important role in nerve repair and evolution (8,9). However, the persistence or excessive activation of neuroinflammation can lead to various neurological diseases.
Microglia are critical nervous system-specific immune cells involved in the response to neuroinflammation. Microglial polarization states comprise M1 and M2 phenotypes. M1 polarized microglia have pro-inflammatory effects and phagocytic functions (10), mainly secreting pro-inflammatory factors such as inducible nitric oxide synthase (iNOS), tumor necrosis factor (TNF)-α, interleukin (IL)-6, and nitric oxide (NO) to further aggravate neuronal damage. M2 polarized microglia inhibit inflammatory responses, regulate neuroinflammation (7), nourish nerves, and promote nerve repair (8,11), with overexpression of arginase-1 (Arg-1), IL-10, IL-4, and transforming growth factor (TGF)-β. Polarization transitions to M1 phenotype induced by high glucose and glucose fluctuations have been observed in microglia both in vivo and in vitro. The expression of iNOS in the hippocampus of diabetic mice was significantly increased by inflammatory factors induced by high glucose (11). Compared with that in the control group, the expression of iNOS in microglia was upregulated after glucose fluctuations. The expression of iNOS also significantly increased after high glucose induction (12). Activation of nuclear factor (NF)-κB and signal transducer and activator of transcription 1 (STAT1) may regulate M1 polarization of microglia, whereas STAT6 and STAT3 activation may regulate M2 polarization (13). Nevertheless, it remains unclear how hyperglycemia and glucose fluctuations regulate polarization transitions of microglia.
In recent years, studies have reported that miRNAs are involved in the regulation of microglia polarization. miR-29b and miR-125a induce polarization transitions to M1 phenotype by inhibiting the expression of TNF-α-inducible protein 3 (TNFAIP3) downstream of the NF-κB signaling pathway (14). miR-146a plays an important role in the regulation of inflammatory responses (15). Previous studies confirmed that miR-146a directly regulates IRAK1 and TRAF6, thereby regulating the pro-inflammatory factor NF-κB and expression of inflammatory factors (16).
The present study therefore aimed to evaluate the role of miR-146a on high glucose and glucose fluctuation-associated polarization transitions of microglia. We hypothesized that miR-146a was involved in the regulation of polarization transitions of microglia induced by hyperglycemia and glucose fluctuations.
For glucose fluctuation BV-2 group culture, after cell attachment, the cell culture medium was replaced with 25 or 75 mmol/L glucose in DMEM every 6 h and cultured for 48 h. High glucose was the last fluctuation.
BV-2 at 3 × 10 4 /well (six well plates) were seeded and cultured for 12 h. A volume of 2 mL enhanced infection solution containing Lv-mmu-miR-146a (Genechem Company, Shanghai, China) with the corresponding viral load and control virus expressing spontaneous green fluorescent protein (GFP) with puromycin acetyltransferase was incubated with cells. The load of viral particles was quantified and normalized before the addiction to the cells. At 12 h after infection, the medium was replaced with conventional culture medium. After 72 h, GFP expression was observed under a fluorescence microscope (Leica, Germany). Complete medium containing 2 µg/mL puromycin (Sigma, Aldrich, USA) was used to screen virus-infected cells, and 1 g/mL was used for further screening. Quantitative realtime polymerase chain reaction (qRT-PCR) was used to assess relative expression levels of miR-146a in the Lv-mmu-miR-146a infection group.

Immunofluorescence Staining of CD11b
Cells grown on glass coverslips were fixed with 4% paraformaldehyde, 0.1% Triton X-100, permeabilized, blocked with 1% bovine serum albumin (BSA), and incubated with anti-CD11b antibody (rat monoclonal antibody; Abcam, USA; 1: 500) at 4 • C overnight. After washes with 1× phosphate-buffered saline (PBS), cells were incubated with the corresponding secondary antibody conjugated with Alexa Fluor R 647 (donkey polyclonal secondary antibody to rat IgG, H&L; Abcam, USA; 1:100) in the dark for 1 h and analyzed using an inverted fluorescence microscope (Leica, Germany). 4 ′ ,6-diamidino-2phenylindole (DAPI) was used to label cell nuclei. To assess BV-2 cell activation. ImageJ was used to calculate the expression of CD11b fluorescence intensity per scaffold. At least five slides were examined per treatment group for each experiment. A comparison of the fluorescence intensity of CD11b between the indicated groups was performed.

Quantitative RT-PCR (qRT-PCR)
Total RNA was extracted with TRIzol (RNAiso Plus) (Takara, Japan). RNA was reversed transcribed into cDNA using the

Gene Primers sequence
CD11b twostep method with PrimeScript TM RT reagent Kit with gDNA Eraser (Takara, Japan). Genomic DNA of miRNA was removed using Recombinant DNase I (RNase-free) (Takara, Japan). miRNA was reversed transcribed into cDNA using the Mir-X R miRNA FirstStrand Synthesis and SYBR R qRT-PCR (Takara, Japan). mRNA qRT-PCR was performed with the TB Green TM Premix Ex Taq TM (Tli RNaseH Plus) (Takara, Japan). miRNA qRT-PCR was performed with Mir-X TM miRNA FirstStrand Synthesis and SYBR R qRT-PCR (Takara, Japan). The primers used are shown in Table 1.

Enzyme-Linked Immunosorbent Assay (ELISA)
The levels of TNF-α, IL-6, and IL-10 in cell culture supernatants were evaluated with an ELISA kit (ABclonal, China), according to the manufacturer's instructions.
Statistical Analysis SPSS 22.0 software (SPSS, USA) was used for data analysis. Data are expressed as mean ± standard deviation (SD). Group pairs were compared by t-tests. Normally distributed data were analyzed 292 using one-way analysis of variance (ANOVA), and non-normal 293 distributions were analyzed with the Kruskal-Wallis H-test. Statistical differences were 297 considered as significant if the P < 0.05.

DISCUSSION
This study demonstrated that high glucose and glucose fluctuations induced polarization transitions to M1 phenotype in BV-2 cells. M1 phenotype parameters including CD11b and iNOS were significantly increased while the expression of M2 phenotype polarizing parameter Arg-1 was significantly decreased. These effects were reversed by overexpression of miR-146a. Furthermore, IRAK1, TRAF6, and NF-κB expression was significantly increased in the high glucose group and glucose fluctuation group. Similarly, these effects were reduced after miR-146a overexpression.
Neuroinflammation is a double-edged sword. Neuroinflammation plays an important role in suppressing infection, removing pathogens, clearing cell debris and misfolded proteins in the nervous system, and nerve repair. However, persistent neuroinflammation induces damage to the nervous system. The role of neuroinflammation depends primarily on the polarization transitions of M1/M2 phenotypes of microglia. Microglia are critical cells that mediate neuroinflammation and have different effects according to polarization phenotype. The M1-polarized phenotype secretes inflammatory cytokines and damages nerve cells, while the M2-polarized phenotype exerts protective effects on nerve cells. Most microglia in the central nervous system are in a relative "quiescent state." Upon changes to the surrounding environment, microglia are activated, which induces M1/M2 polarization transitions. An in vitro study suggested that hyperglycemia induced polarization transitions to M1 polarization in microglia (17). In the cortex of diabetic mice, M1 phenotype polarization of microglia was increased, whereas M2 phenotype polarization of microglia was decreased. Similarly, microglia in the hypothalamus of streptozotocininduced diabetic rats demonstrated M1 phenotype polarization transition (18). The deleterious effects of glucose fluctuations on diabetic macroangiopathy has been reported (19,20). However, few studies on the effects of diabetic encephalopathy have been performed. An in vitro study indicated that glucose fluctuations significantly activated microglia (12). Nevertheless, more studies are required to confirm these findings.
In this study, we use 25 mmol/L glucose DMEM for BV-2 cell culture. We found that 25 mmol/L glucose didn't impact the level of miR-146a, while in the high glucose group, miR-146a expression was reduced. However, Chen et al. demonstrated that 25 mmol/L of glucose decreased the expression of miR-146a in human retinal microvascular endothelial cells (21).
We measured the microglial activation marker CD11b as well as M1 and M2 phenotype markers. We observed that both hyperglycemia and glucose fluctuations induced polarization transitions to M1 phenotype in microglia, leading to increased expression of pro-inflammatory factors including TNF-α and IL-6. Previous studies suggested that excessive M1 phenotype polarization induces inflammatory factor secretion in diabetic encephalopathy, leading to central nervous system damage in diabetes. The expression of iNOS was up-regulated in the hippocampus of diabetic mice and aggravated neuronal damage (22). Our study verified that high glucose and glucose fluctuations induced M1 phenotype polarization transitions in microglia in vitro, which is a possible mechanism underpinning diabetic encephalopathy.
Hyperglycemia and blood glucose fluctuations induce cell activation through multiple pathways (23). Of these, the TLR4/NF-κB signaling pathway is one of the main pathways involved. Dasu et al. reported that high glucose induced upregulation of toll-like receptors (TLR; including TLR2 and TLR4) in human monocytic THP-1 cell lines and initiated intracellular myeloid differentiation factor 88 (MyD88)/IRAK-1/NF-κB signaling pathway, which induced inflammatory responses (24). In recent years, a growing number of studies has reported that epigenetic regulation is an important regulatory mechanism underscoring neuroinflammation, especially microRNAs (25)(26)(27)(28). Different miRNAs are expressed depending on phenotype polarization of microglia; in turn, miRNA affect polarization transitions of microglia (29). For example, high expression of miR-155 in microglia induced M1 phenotype polarization and inhibited M2 phenotype polarization (30). miR-124 induced M2 phenotype polarization by targeting inhibition of M1 phenotype polarization to maintain the resting state of microglia and macrophages (31). miR-689, miR-124, and miR-155 are involved in M1 phenotype polarization. In contrast, miR-124, miR-711, and miR-145 are implicated in M2 phenotype polarization in lipopolysaccharide (LPS) or IL-4 stimulated primary mouse microglia based on miRNA expression profiling and bioinformatics analysis (29). To date, there have been no reports on miR-146a in the context of microglial phenotype polarization. Previous studies reported that miRNA-146a played an important role in the regulation of inflammation mainly through the NF-κB pathway. Taganov et al. reported that miR-146a regulated two key adapter molecules, TRAF-6 and IRAK1, which were downstream of the TLR signaling pathway, thereby regulating the activation of the pro-inflammatory factor NF-κB and negatively regulating the release of pro-inflammatory factors (16). Yang et al. reported that miRNA-146a negatively regulated mouse T-cell activation, and the regulatory mechanism also affected the expression of NF-κB by regulating the expression of TRAF-6 and IRAK1 (32). Studies have indicated that miR-146a inhibited M1 polarization and induced M2 polarization in mouse alveolar macrophages induced by LPS (33). Further, the M1 phenotype polarizing factors IL-6 and TNF-α were increased, while the M2 phenotype polarizing factor MGL-1 was decreased in alveolar macrophages by inhibiting the expression of miR-155 or miR-146a. miR-146a and miR-155 may be involved in the anti-inflammatory activity of macrophages and participate in the mechanisms underpinning M1 and M2 polarization transitions (34).
miR-146a induced M2 phenotype polarization and upregulated tumor-associated macrophages expression to regulate breast cancer tumor growth (35). Our results showed that high glucose increased TNF-α and IL-6 expression while reduced IL-10 expression, which was reversed by overexpression of miR-146a. Similar results was observed in another study. Overexpression of miR-146a attenuated the inhibitory effects of NIFK-AS1 M2 phenotype polarization in macrophages, while the expression of IL-10 and Arg-1 were increased (36). Altered expression of miR-146a in the brains of patients with Alzheimer's disease (AD) underpinned inflammatory senile plaque lesions in AD brains (37,38). miR-146a expression was decreased in DRG cells in hyperglycemia studies of diabetic peripheral neuropathy; cell survival rate was decreased, but this was attenuated by miR-146a mimics (39,40). Rong et al. reported that the expression of miR-146a in plasma of patients with type 2 diabetes was significantly decreased (29). Our study revealed that the expression of miR-146a was significantly decreased in cell cultures by high glucose and glucose fluctuations, which partially verified that low levels of miR-146a expression were induced by hyperglycemia (41). Both in vitro and in vivo studies suggest that miR146a is involved in the regulation of neuroinflammation and plays an important role in various diseases such as diabetes and neurodegenerative diseases. Therefore, we speculated that miR-146a was involved in regulating the polarization transitions of microglia induced by high glucose and glucose fluctuations. Overexpression of miR-146a regulated polarization transitions induced by high glucose and glucose fluctuations promoting M1 to M2 phenotype polarization.
A negative feedback loop between miR-146a and NF-κB may be at play in this process. The expression of miR-146a was decreased in the hippocampus of diabetic rats, while increased expression of IRAK1, TRAF6, and NF-kB aggravated hippocampal inflammation and apoptosis (42). However, compared with that in the control group, the expression of miR-146a and NF-κB were increased, expression of TRAF6 was decreased, and expression of TNF-α, IL-6, and IL-1β were increased in the sciatic nerve of DM rats, indicating that loss of NF-κB/miR-146a in the negative feedback loop may underpin the pathogenesis of diabetic peripheral neuropathy (43). Further, miR-146a regulates the expression of target genes IRAK1, TRAF6, and NF-κB. Thus, the regulation may be involved in the NF-κB pathway. In this study, overexpression of miR-146a regulated M1/M2 polarization transitions by downregulating the expression of IRAK1, TRAF6, and NF-κB in microglia. This could be a regulatory mechanism in the response of microglial inflammation induced by high glucose and glucose fluctuation.
Central nervous system inflammation is a complex process involving various glial cells, macrophages, neuronal cells, and vascular endothelial cells. The results of this study were only observed in vitro, which precludes inferences on the neuroinflammatory effects of polarization transitions of microglia in diabetic encephalopathy. However, this study provides insight into the pathogenesis of diabetic encephalopathy. We observed that overexpression of miR-146a contributed to polarization transitions from M1 to M2 phenotype in microglia. Our results provide a potential strategy for the treatment of diabetic encephalopathy. Future studies should further investigate how miR-146a regulates polarization transitions in microglia induced by high glucose and glucose fluctuations.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the manuscript.

AUTHOR CONTRIBUTIONS
YH, ZL, and XL conceptualized and designed these studies, performed them, and wrote the manuscript. XW, XC, XB, YZ, YY, and JZ contributed through data analyses, data interpretation, and manuscript preparation. All authors contributed to manuscript revision and read and approved the submitted version.