Author Correction: Infection of microglia with Porphyromonas gingivalis promotes cell migration and an inflammatory response through the gingipain-mediated activation of protease-activated receptor-2 in mice

A correction to this article has been published and is linked from the HTML and PDF version of this paper. The error has been fixed in the paper.


Infection of microglia with P. gingivalis induces cell migration through gingipains. To further
address that microglial accumulation around the injection site of P. gingivalis in the somatosensory cortex is due to cell migration but not to cell proliferation, in vitro migration assay using a Boyden chamber was performed. P. gingivalis infection induced the cell migration of MG6 cells (Fig. 2a) and primary cultured microglia (Fig. 2b) through the polycarbonate membrane to a greater degree than seen in untreated controls. The potential involvement of extracellular nucleotides, including ADP or ATP, in the cell migration of microglia was next examined, as cortical microglia extended their processes towards focally injected P. gingivalis through the UDP-P2Y 6 receptor system 15 . However, neither MRS2578, a selective inhibitor of P2Y 6 receptors, nor PSB0739, a selective inhibitor of P2Y 12 receptors, had any effect on the P. gingivalis infection-induced MG6 cell migration ( Supplementary  Fig. S1). KYT1 and KYT36 significantly inhibited the P. gingivalis infection-induced cell migration of MG6 cells when administered separately, and their combined administration almost completely inhibited the migration (Fig. 2a,c). A combined administration of KYT1 and KYT36 also significantly inhibited the P. gingivalis infection-induced cell migration of primary cultured microglia (Fig. 2b,d). On the other hand, P. gingivalis LPS failed to induce cell migration of MG6 cells, whereas E. coli LPS induced a significant cell migration of MG6 cells ( Supplementary Fig. S1).

Gingipains induce the proteolytic activation of protease-activated receptor (PAR) 2 in microglia
after infection with P. gingivalis. The four PAR family members (PAR1, 2, 3 and 4) are G-protein-coupled receptors with a unique mechanism of activation. These receptors carry their own tethered ligands and are activated by serine proteases such as trypsin and tryptase. PARs in epithelial cells is also activated by trypsin-like cysteine proteases and gingipains [20][21][22] . The effects of infection with P. gingivalis on the activation of PAR2 in MG6 cells was examined, as human microglia only express PAR2 23 . S-19, which specifically recognizes the activated form of PAR2, was used to examine the effect of P. gingivalis infection on the activation of PAR2 expressed on MG6 cells and primary cultured microglia. Infection with P. gingivalis significantly increased the levels of the activated form of PAR2 in MG6 cells (Fig. 3a) and primary cultured microglia (Fig. 3b). KYT1 and KYT36 significantly inhibited the PAR2 activation when administered separately, and their combined administration almost completely inhibited the P. gingivalis infection-induced activation of PAR2 in MG6 cells (Fig. 3a,c). A combined administration of KYT1 and KYT36 also significantly inhibited the P. gingivalis infection-induced activation of PAR2 in primary cultured microglia (Fig. 3b,d).
PAR2 activation and subsequent phosphorylation of Akt and ERK1/2 are involved in the P. gingivalis infection-induced microglial migration. Next, the potential involvement of PAR2 in the P.
gingivalis-induced cell migration of MG6 cells was examined. The P. gingivalis infection-induced cell migration ScientiFic RepoRTs | 7: 11759 | DOI:10.1038/s41598-017-12173-1 of MG6 cells was significantly inhibited by the administration of the PAR2 neutralization antibody SAM11 and by PAR2 siRNA. In contrast, the control antibody (Ab) and siRNA had no effect (Fig. 4a,b).
The signaling pathways that are triggered by PAR2 activation may activate several downstream intracellular signaling pathways, including phosphoinositide 3-kinase (PI3K)/Akt 24 and mitogen-activated protein kinase/ extracellular signal-regulated kinase (ERK) kinase (MEK)/ERK pathways 25 , which are responsible for a variety of microglial cellular activities including cell migration 26 . We then examined the potential activation of PI3K/ Akt and MEK/ERK pathways after the infection of MG6 cells and primary cultured microglia with P. gingivalis. Akt is a serine/threonine protein kinase activated by stimuli that induce the production of phosphatidylinositol (3,4,5)-trisphosphate through activation of PI3K 27 . Akt and ERK1/2 were phosphorylated after the infection of MG6 cells with P. gingivalis ( Supplementary Fig. S2). Blockade of PAR2-mediated intracellular signaling using the neutralizing antibody SAM11 or siRNA significantly inhibited the P. gingivalis infection-induced phosphorylation of Akt and ERK1/2 in MG6 cells (Fig. 4c,d) and primary cultured microglia (Fig. 4e,f). In contrast, control Ab and siRNA had no effect.
in a dose-dependent manner (Fig. 5a,b). Furthermore, ERK inhibition with U0126 also significantly inhibited the P. gingivalis infection-induced cell migration of MG6 cells in a dose-dependent manner (Fig. 5c). MG6 cells were stained with phalloidin to visualize actin polymerization after P. gingivalis infection, as actin polymerization and membrane ruffling are essential steps for cell migration. After the infection of MG6 cells with P. gingivalis, the F-actin of MG6 cells was polymerized to induce membrane ruffling (Fig. 5d arrows). Regarding the mechanism of membrane ruffling after the infection of MG6 cells with P. gingivalis, the possible involvement of the PI3K/Akt and MEK/ERK pathways was examined. The treatment with LY294002, Akti and U0126 significantly inhibited the P. gingivalis infection-induced membrane ruffling of MG6 cells (Fig. 5d,e).
Infection of microglia with P. gingivalis promotes expression of proinflammatory mediators through activation of PAR2 by gingipains. Infection of MG6 cells with P. gingivalis significantly increased the mRNA expression levels of proinflammatory mediators, including interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α) and inducible nitric oxide synthase (iNOS) (Fig. 6a), without affecting the mRNA expression of anti-inflammatory mediators, including IL-10, arginase-1 and IL-4 ( Supplementary Fig. S3). Both KYT1 and KYT36 significantly inhibited the P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS in MG6 cells (Fig. 5a). The combined use of KYT1 and KYT36 further inhibited the P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS in MG6 cells. P. gingivalis infection also induced expression of IL-6, TNF-α and iNOS in primary cultured microglia. A combined administration of KYT1 and KYT36 significantly inhibited the P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS in primary cultured microglia (Fig. 6b). Furthermore, the infection of primary cultured microglia with P. gingivalis also significantly increased the secretion of IL-6 and TNF-α, and the accumulation of NO metabolites (Fig. 6c). A combined administration of KYT1 and KYT36 significantly increased the mean amounts of IL-6 and TNF-α in the culture medium after infection of P. gingivalis. On the other hand, their combined administration significantly decreased the mean amount of NO metabolites in the culture medium (Fig. 6c).
Next, a possible involvement of PAR2 in the P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS was examined. The PAR2 neutralization antibody SAM11 or PAR2 siRNA significantly inhibited P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS in MG6 cells (Fig. 7a). In contrast, control Ab and siRNA had no effect. Furthermore, either treatment with LY294002 as well as Akti significantly inhibited P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS in MG6 cells (Fig. 7b).

Infection of microglia with P. gingivalis induce microglia-mediated neuroinflammation
through the activation of PAR2 and Toll-like receptor 2 (TLR2). Next, we examined a possible crosstalk of PAR2-and TLR2-regulated inflammatory activation of microglia, because P. gingivalis LPS activates microglia to produce proinflammatory mediators through TLR2 14 . TLR2 Ab significantly inhibited the P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS in MG6 cells (Fig. 8a) and primary cultured microglia (Fig. 8b). A combined administration of TLR2 Ab and PAR2 siRNA further inhibited P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS in MG6 cells (Fig. 8a) and primary cultured microglia (Fig. 8b).
Finally, a possible link between cell migration and inflammatory responses of microglia in response to P. gingivalis infection was examined by treatment with cytochalasin D, which can suppress cell migration by inhibiting actin polymerization. Treatment with cytochalasin D significantly inhibited the mRNA expression of proinflammatory mediators in microglia ( Supplementary Fig. S4).

Discussion
We herein show that the infection of microglial cells with P. gingivalis induced cell migration through secretion of gingipains and the subsequent proteolytic activation of PAR2. Furthermore, the combined use of KYT1 and KYT36 almost completely suppressed the P. gingivalis infection-induced cell migration and the PAR2 activation, suggesting that gingipains, Rgp and Kgp, are wholly responsible for these events. Although the potential involvement of extracellular nucleotides, including ADP or ATP, in the cell migration of microglia was widely known, neither P2Y 6 receptors nor P2Y 12 receptors was not involved in the P. gingivalis infection-induced microglial cell migration. Furthermore, the cell migration of MG6 cells was induced by LPS derived from E. coli, but not from P. gingivalis. These observations were consistent with a previous report to demonstrate that none of the TLR2-specific agonists stimulated migration in cultured microglia 28 . Therefore, it is considered that TLR4, but not TLR2, is involved in cell migration of microglia.
In our in vivo experiments, we utilized CX3CR1 +/GFP mice to evaluate microglial migration towards the injection site of P. gingivalis. It has been reported that brain-resident microglia expressed significantly higher level of CX3CR1 than macrophages 29 . Furthermore, Yamasaki et al. have utilized CX3CR1 +/GFP CCR2 +/RFP knock-in mice to show that CCR2 single-positive cells (infiltrating macrophages) exhibited elongated or spindle shape, whereas CX3CR1 single-positive cells (brain-resident microglia) showed a process-bearing morphology 30 . In the present study, most CX3CR1-positive cells that accumulated around the injected sites had a process-bearing morphology. Therefore, it is considered that brain-resident microglia could be delineated from infiltrating macrophages by counting process-bearing bright CX3CR1-positive cells. Furthermore, process-bearing bright CX3CR1-positive cells accumulated around the injection site were negative for Ki67, but only few CX3CR1-positive cells with spindle shape, presumably infiltrating macrophages, were positive for Ki67. These observations suggest that the accumulated cells around the injection site were mainly associated with cell migration rather than cell proliferation.
To address the intracellular signalings associated with PAR2 activation, we have focused on PI3K/Akt and MEK/ERK pathways because these two signaling pathways can induce microglial migration 31,32 . In the present study, infection of cultured microglia with P. gingivalis induced phosphorylation of Akt and ERK1/2. Our pharmacological studies showed that activation of PI3K/Akt and MEK/ERK pathways plays essential roles in the P. gingivalis infection-induced migration and membrane ruffling of microglia. The P. gingivalis infection-induced phosphorylation of Akt in microglia was decreased by the inhibition of PAR2. Furthermore, we provided evidence supporting the notion that Rgp and Kgp may regulate the cell migration of microglia through the proteolytic cleavage of PAR2 and the subsequent activation of the PI3K/Akt pathway. However, it was previously reported that Rgp and Kgp modulate neutrophil chemotaxis through the proteolytic cleavage of IL-8 33 . Furthermore, Rgp and Kgp promote monocyte migration through the proteolytic activation of matrix metalloprotease-9, which induces monocyte migration 34 . Therefore, the involvement of proteolytic cleavage other than the activation of PAR2 in the gingipain-mediated microglial migration cannot be totally ruled out, as the inhibition of PAR2-mediated signaling using either neutralizing Ab or siRNA significantly but not completely inhibited P. gingivalis infection-induced microglial migration.
In addition to the cell migration, the infection of microglia with P. gingivalis significantly increased the mRNA expression of proinflammatory mediators, including IL-6, TNF-α, and iNOS, without affecting the mRNA expression of anti-inflammatory mediators, including IL-10, arginase-1 and IL-4. The combined use of KYT1 and KYT36 significantly but incompletely inhibited the P. gingivalis infection-induced increased mRNA expression of proinflammatory mediators. These observations suggest that virulence factors other than gingipains, such as LPS and fimbriae, are also involved in the P. gingivalis infection-induced inflammatory responses. The infection of microglia with P. gingivalis also significantly increased the secretion of IL-6 and TNF-α, and the accumulation of NO metabolites. On the other hand, the combined administration of KYT1 and KYT36 significantly increased the extracellular levels of IL-6 and TNF-α, but significantly decreased NO metabolites, after infection of microglia with P. gingivalis. It has been reported that the cooperative action of Rgp and Kgp is also involved in the proteolytic degradation of various cytokines including TFN-α 35 and IL-6 36,37 . Therefore, it is considered that Rgp and Kgp may limit the microglia-mediated neuroinflammation through proteolytic degradation of proinflammatory cytokines.
In the present study, the siRNA of PAR2 strongly inhibited the P. gingivalis infection-induced expression of proinflammatory mediators. A similar observation was reported by Uehara et al. 32 , who observed that Rgp and Kgp stimulate the secretion of proinflammatory cytokines by monocytes through the proteolytic cleavage of PAR-1, PAR-2 and PAR-3 with synergistic effects when used in combination with pathogen-associated molecular patterns. Furthermore, the inhibition of the PI3K/Akt pathway significantly inhibited the P. gingivalis infection-induced expression of IL-6, TNF-α and iNOS in microglia. Besides P. gingivalis LPS/TLR2 system, gingipain/PAR2 system was also involved in P. gingivalis infection-induced expression of inflammatory mediators by microglia. In the present study, TLR2 Ab and PAR siRNA had additive inhibitory effects, suggesting that infection of microglia with P. gingivalis activates gingipain/PAR2 system to induce neuroinflammation in a manner parallel to P. gingivalis LPS/TLR2 system. The precise mechanism by which microglia mediate chronic inflammation, which is responsible for the dementia associated with AD, remains unclear. One possible mechanism is that gingipain-and/or LPS-mediated acute neuroinflammation may induce the neuronal production of Aβ 5 , which may advance the acute neuroinflammation to the chronic state.
In this study, we have provided the first evidence that Rgp and Kgp cooperatively contribute to the cell migration of microglia towards the infected site and induction of neuroinflammation after infiltration into the brain. The cell migration associated with actin polymerization may be necessary for the subsequent inflammatory responses of microglia after activation of PAR2. Considering that Rgp and Kgp degrade components of the basal membrane, including type I collagen and fibronectin, and enhance the vascular permeability 38 , they may also contribute to the invasion of P. gingivalis into the brain (Supplementary Fig. S5). Therefore, our observations in this study may support "the infection hypothesis of Alzheimer's disease" 1,9 .  Bacterial strains and culture conditions. P. gingivalis ATCC33277 and Lys gingipain mutant P. gingivalis KDP129 were used. P. gingivalis was maintained on blood agar plate containing 40 mg/ml trypto-soya agar (Nissui Pharmaceutical, Tokyo, Japan), 5 mg/ml brain heart infusion (Becton, Dickinson and Company, Franklin Lakes, NJ, USA), 1 g/ml cysteine (Wako Pure Chemical Industries, Osaka, Japan), 5 μg/ml hemin (Sigma-Aldrich, St. Louis, MO, USA), 1 μg/ml menadione (Sigma-Aldrich), 5% defibrinated sheep blood (Nippon Bio-test laboratories, Tokyo, Japan) in Bactron anaerobic chamber (Shel Lab, Cornelius, OR, USA) with 10% CO 2 , 10% H 2 , 80% N 2 15 . KDP129 was maintained on blood agar plate with 20 μg/ml chloramphenicol (Wako Pure Chemical Industries). P. gingivalis was grown in enriched BHI broth containing 37 mg/ml brain heart infusion, 2.5 mg/ ml yeast extract (Becton, Dickinson and Company), 1 g/ml cysteine, 5 μg/ml hemin and 1 μg/ml menadione. KDP129 was grown on BHI broth with 20 μg/ml chloramphenicol. Before P. gingivalis was cocultured with MG6 cells and primary cultured microglia, P. gingivalis culture medium was centrifuged (6000 × g, 10 min) and the supernatant was replaced with DMEM without FBS or penicillin-streptomycin.

Animals. CX3CR1
Microinjection and data acquisition. CX3CR1 +/GFP mice were anesthetized via an intraperitoneal injection of urethane and atropine (1.7 g/kg and 0.4 mg/kg, respectively) and positioned in a stereotaxic apparatus (Narishige, Tokyo, Japan), A cranial window was made in the somatosensory cortex, and the dura was carefully removed. A stab wound was made by inserting a 30-G needle into the somatosensory cortex (anterior, −1.7 mm; lateral, 2.0 mm, dorsoventral, 0.5 mm). Culture medium BHI, P.gingivalis in BHI or KDP129 in BHI (1 μl, 1 × 10 6 CFU/ml) was injected. The flow rate of the injection was 10 μl/h maintained by a microsyringe pump (As One, Osaka, Japan), followed by waiting 10 min for total infiltration. Then skin was then sutured with 6.0-mm silk thread. The mice were killed 24 h after the injection. Migration assay in vivo. The microinjected mice were deeply anesthetized with an overdose of sodium pentobarbital (120 mg/kg, i.p.) and perfused transcardially with phosphate-buffered saline (PBS, pH 7.4) followed by 4% paraformaldehyde. The brains were obtained, cryoprotected for 2 days in 30% sucrose in PBS and then were embedded in an optimal cutting temperature compound (Sakura Finetechnical, Tokyo, Japan). Serial coronal frozen sections (14 μm) of the samples were prepared and mounted in Vectashield anti-fading medium (Vector Laboratories, Burlingame, CA, USA). Fluorescence images were taken using the confocal laser-scanning microscope (CLSM, 2si Confocal Laser-scanning Microscope; Nikon, Tokyo, Japan). All of the images were processed using the ImageJ 1.47 h software program (NIH). To account for any differences in the fluorescence intensity among the experiments, every image was set at the maximum threshold value (255), and all of the background was set to 0. To count the CX3CR1-GFP-positive cells, five squares (300 × 300 μm) were placed around the injection site of 3 independent sections per mouse, and the cells in those 5 squares were counted and statistically analyzed.
Primary cultured microglia. CD11b+ cells were isolated from the mouse brain by the MACS method.
Twelve mice were anesthetized and transcardially perfused with PBS. The brains were obtained and cut into small pieces. After enzymatic digestion using a Neural Tissue Dissociation Kit (Miltenyi Biotec), the cell suspensions were further mechanically dissociated using a gentle MACS Dissociator (Milteny Biotec). The single-cell suspensions were obtained after being move to a 30mm cell strainer. After magnetic labeling with CD11b MicroBeads, the cell suspension was loaded onto a magnetic (MACS) column placed in the magnetic separator (Milteny Biotec). After rinsing the MACS column with PBS, the CD11b-positive fraction was collected according to previously described methods 39,40 . Cell migration assays in vitro. Cell migration was determined by using a Boyden chamber assay. A 12-well Boyden chamber (NeuroProbe, Gaithersburg, MD, USA) was used for the measurement of cell migration, in accordance with the manufacturer's instruction. E. coli LPS, P. gingivalis LPS (Invivogen) and 6 × 10 5 CFU/ml P. gingivalis in DMEM were placed into the lower wells, which were separated from the upper wells by polyvinylpyrrolidone-free polycarbonate filters (8-μm pore size, 25 × 80 mm; NeuroProbe). MG6 cells and primary cultured microglia were harvested by trypsinization, resuspended in DMEM, and added to the upper chamber at a density of 1.8 × 10 5 cells/ml. Cells were incubated at 37 °C under cell culture conditions (95% air/5% CO 2 ) for 12 h. At the end of the incubation, any non-migrating cells on the upper side of the membrane were removed with a cotton swab. Migrated cells on the lower part of the membrane were fixed in methanol for 10 min and stained with diff-quik (Sysmex, Kobe, Japan). Photomicrographs of randomly chosen fields were taken (BX-41; Olympus, Tokyo, Japan), and all cells were enumerated to calculate the average number of cells that had migrated.
PAR2 knockdown with small interfering RNAs. MG6 cells and primary cultured microglia were seeded on 6-well plates at a density of 2 × 10 5 cells/well in 2 ml of antibiotic-free DMEM. After 12 h, the cells were transiently transfected with control siRNA (sc-37007; Santa Cruz Biotechnology, Dallas, TX, USA), or PAR2 siRNA (sc-36187; Santa Cruz Biotechnology), using siRNA Transfection Reagent (sc-29528; Santa Cruz Biotechnology) in accordance with the manufacturer's protocol. At 12 h after transfection, the cells were harvested for cell migration and RT-PCR. Immunofluorescent staining. Microinjection brain sections of CX3CR1 +/GFP mice were obtain for the proliferation assay, the samples were incubated with rabbit anti-Ki67 (SP6, 1:500; abcam) overnight at 4 °C, after being washed with PBS, they were incubated with donkey anti-rabbit Cy3 (1:500; Jackson ImmunoResearch, West Grove, PA, USA) for 2 h and then with Hoechst stain (1:200) and mounted in Vectashield anti-fading medium. Fluorescence images were taken using CLSM.
For the PAR2 activation assay, the infected MG6 cells and primary cultured microglia were fixed with 4% paraformaldehyde for 30 min. After washing with PBS, cells were then incubated with the goat anti-PAR2 (S-19, 1:1000; Santa Cruz Biotechnology) overnight at 4 °C. After being washed with PBS, they were incubated with donkey anti-goat Cy3 (1:500; Jackson ImmunoResearch, West Grove, PA, USA) for 2 h and then with Hoechst stain (1:200) and mounted in Vectashield anti-fading medium. Fluorescence images were taken using CLSM. An immunofluorescent intensity analysis of cleaved PAR2 staining was performed using the ImageJ 1.47 h software program (NIH).
For the actin polymerization assay, the infected MG6 cells were fixed with 4% paraformaldehyde for 30 min. After washing with PBS, cells were incubated with Texas Red-X Phalloidin (1 unit/well; Thermo Fisher Scientific, Waltham, MA, USA) for 1 h and then with Hoechst stain (1:200) and mounted in Vectashield anti-fading medium. Fluorescence images were taken using a CLSM. All of the images were processed using the ImageJ software program, and Z-stack images were projected along the Z-axis to recreate a two-dimensional (2D) representation of three-dimensional (3D) structures 39,40 . Immunoblotting analyses. The immunoblotting analyses were conducted as described previously 40 . In brief, each specimen was electrophoresed using 12% SDS-polyacrylamide gels. The proteins on the SDS gels were then electrophoretically transferred to nitrocellulose membranes. Following the blocking, the membranes were incubated at 4 °C overnight under gentle agitation with rabbit anti-Akt antibody and rabbit anti-pAkt (Ser 473) antibody (1:1000; Cell Signaling, Danvers, MA, USA), rabbit anti-ERK1/2 (1:1000; Cell Signaling Technology) and rabbit anti-pERK1/2 (1:1000; Cell Signaling Technology). After being washed, the membranes Enzyme-linked immunosorbent assay (ELISA) and NO 2 − /NO 3 − assay. The mean amounts of IL-6, TNF-α and NO 2 − + NO 3 − in the culture medium were measured after the infection of primary cultured microglia (2 × 10 6 cell/well, 6-well plates) with P. gingivalis (MOI 1:5). IL-6 and TNF-α were measured by ELISA kits (R&D Systems) following the protocol provided by the manufacturer. The absorbance at 450 nm was determined using a microplate reader. NO 2 and NO 3 were measured by NO 2 -/NO 3 assay kits (Griess reagent kit; Dojindo, Tokyo, Japan) following the protocol provided by the manufacturer. The absorbance at 540 nm was determined using a microplate reader.
Reagents. P2Y 12 receptor inhibitor PSB0739 was obtained from Tocris Bioscience (Bristol, UK). P2Y 6 receptor inhibitor MRS2578, MEK inhibitor U0126 and cytocharasin D were obtained from Sigma-Aldrich. PAR2 neutralization antibody (SAM11) was obtained from Santa Cruz Biotechnology, TLR2 neutralizing antibody from eBioscience (San Diego, CA, USA), and control Ab (mouse IgG2a) from BioLegend (San Diego, CA, USA). Rgp inhibitor KYT1 and Kgp inhibitor KYT36 were obtained from Peptide Institute (Osaka, Japan). PI3K inhibitor LY294002 was obtaind from EMD Milipore (Billerica, MA, USA), Akt inbibitor Akti from EMD Milipore. For in vivo studies, KYT1 and KYT36 were preincubated with P. gingivalis for 1 h. For in vitro studies, inhibitors and neutralizing enzymes were preincubated with MG6 cells and primary cultured microglia for 1 h. Statistical analysis. The data are represented as the mean ± standard error of the mean. The statistical analyses were performed using a two-tailed unpaired Student's t-test and a one-way analysis of variance (ANOVA) with a post hoc Tukey's test using the GraphPad Prism 7 Software package (GraphPad Software Inc., San Diego, CA, USA). A value of p < 0.05 was considered to indicate statistical significance.