The expression and regulation of matrix metalloproteinase-3 is critically modulated by Porphyromonas gingivalis lipopolysaccharide with heterogeneous lipid A structures in human gingival fibroblasts

Background Porphyromonas gingivalis lipopolysaccharide (LPS) is a crucial virulence factor strongly associated with chronic periodontitis which is the primary cause of tooth loss in adults. It exhibits remarkable heterogeneity containing tetra-(LPS1435/1449) and penta-(LPS1690) acylated lipid A structures. Human gingival fibroblasts (HGFs) as the main resident cells of human gingiva play a key role in regulating matrix metalloproteinases (MMPs) and contribute to periodontal homeostasis. This study investigated the expression and regulation of MMPs1-3 and tissue inhibitors of MMP-1 (TIMP-1) in HGFs in response to P. gingivalis LPS1435/1449 and LPS1690 and hexa-acylated E. coli LPS as a reference. The expression of MMPs 1–3 and TIMP-1 was evaluated by real-time PCR and ELISA. Results The MMP-3 mRNA and protein were highly upregulated in P. gingivalis LPS1690- and E. coli LPS-treated cells, whereas no induction was observed in P. gingivalis LPS1435/1449-treated cells. On the contrary, the expression of MMP-1 and −2 was not significantly affected by P. gingivalis LPS lipid A heterogeneity. The TIMP-1 mRNA was upregulated in P. gingivalis LPS1435/1449- and E. coli LPS-treated cells. Next, signal transduction pathways involved in P. gingivalis LPS-induced expression of MMP-3 were examined by blocking assays. Blockage of p38 MAPK and ERK significantly inhibited P. gingivalis LPS1690-induced MMP-3 expression in HGFs. Conclusion The present findings suggest that the heterogeneous lipid A structures of P. gingivalis LPS differentially modulate the expression of MMP-3 in HGFs, which may play a role in periodontal pathogenesis.


Background
Periodontal disease is a bacterially induced and highly common chronic inflammatory condition in humans, and severe periodontal disease (periodontitis) remains the major cause of tooth loss in adult population worldwide [1]. Dysregulated host response to pathogenic plaque biofilm critically contributes to destructive inflammation resulting in tissue damage and alveolar bone loss [2]. Porphyromonas gingivalis is a keystone periodontal pathogen in the mixed microbial community and it releases copious amount of lipopolysaccharide (LPS) which perpetually interacts with host cells, thereby significantly contributing to periodontal pathogenesis [1][2][3][4].
LPS is a potent immuno-inflammatory modulator which causes serious complications in host. It is comprised of three major components viz. outermost O-antigen, core oligosaccharide regions and innermost lipid A [3]. Lipid A is the biologically most active component of LPS that imparts the endotoxin activity. Its structure differs widely among Gram-negative bacteria species depending on the differences in composition of attached fatty acids, number of phosphorylation sites and substituted groups attached to the phosphate residues [3]. The canonical lipid A structure in Escherichia coli LPS is a hexa-acylated diphosphorylated glucosamine disaccharide. Previous studies have shown that P. gingivalis possesses highly heterogeneous lipid A structures containing penta-acylated LPS 1690 and tetraacylated LPS 1435/1449 , and this structural discrepancy may critically account for contrasting biological activities induced by P. gingivalis LPS [3,4].
Human gingival fibroblasts (HGFs) are the major cell type in human gingiva [5][6][7]. They play a key role in maintenance and remodeling of extra cellular matrix (ECM) by producing various structural components, such as collagen, elastin, glycoprotein and glycosaminoglycans. In addition, HGFs also synthesize and secrete various members of matrix metalloproteinases (MMPs) in response to P. gingivalis LPS challenge, which ultimately contribute to periodontal tissue destruction [8].
MMPs are a family of structurally and functionally related proteolytic enzymes containing a zinc-binding catalytic domain and they are active against the components of ECM [8][9][10]. The activity of MMPs is largely regulated by several naturally occurring inhibitors like tissue inhibitors of metalloproteinases (TIMPs) [11]. Overall, there is a remarkable balance between MMPs and TIMPs in periodontal connective tissues and disturbance of this balance is therefore critically implicated in the destruction of periodontal tissues [12,13]. In normal conditions, MMPs are involved in the remodeling and turnover of periodontal tissues under the strict control of TIMPs, which bind specifically to the active site of the enzyme thereby maintaining the equilibrium between degradation and regeneration of ECM [8,14]. Increased production of MMPs 1-3 is observed in chronic inflammatory condition such as periodontitis that results in excessive connective tissue breakdown [14,15]. MMPs such as MMP-1, -2, -3, -9 and −13 are synthesized in periodontal tissues in response to periodontopathic bacteria like P. gingivalis. Previous studies have suggested that LPS could regulate the MMP expression in various host cell types including HGFs [10,16].
Currently, there are no studies on the role of P. gingivalis LPS lipid A heterogeneity with respect to expression of MMPs in HGFs. The present study therefore aimed to investigate the expression and regulation of MMPs 1-3 and TIMP-1 in HGFs in response to the different isoforms of P. gingivalis LPS 1435/1449 and P. gingivalis LPS 1690 as well as E. coli LPS as a reference. This study sheds light on the regulation of MMP expression and underlying signal transduction pathways in HGFs in response to heterogeneous P. gingivalis LPS, which could have important implications in the pathogenesis of periodontal disease.

Results
Heterogeneous P. gingivalis LPS lipid A structures differentially modulate MMPs 1-3 and TIMP-1 mRNAs The dose-dependent experiments showed that both P. gingivalis LPS 1435/1449 and LPS 1690 differentially modulated the expression of MMP-3 transcript. The latter (0.1-10 μg/ml) markedly upregulated the expression of MMP-3 mRNA while the former did not affect the expression (Figure 1c). Similarly, E. coli LPS (0.1-10 μg/ml) significantly upregulated MMP-3 expression. Both isoforms of P. gingivalis LPS upregulated to different extent the expression of MMP-1 and MMP-2 mRNAs, while E. coli LPS significantly upregulated the expression of these transcripts (Figures 1a and b). TIMP-1 mRNA expression was significantly induced in P. gingivalis LPS 1435/1449 -and E. coli LPS-treated cells, and no significant induction was observed following P. gingivalis LPS 1690 stimulation ( Figure 1d).
Notably, MMP-3 transcript was differentially expressed in the cells treated by the two isoforms of P. gingivalis LPS. P. gingivalis LPS 1690 significantly upregulated MMP-3 mRNA expression at 24 and 48 h, while E. coli LPS showed prompt expression at 12 h (Figure 2c). MMP-2 mRNA was significantly upregulated by both P. gingivalis LPS 1435/1449 and LPS 1690 at 48 h (Figure 2b), and MMP-1 transcript was significantly upregulated by P. gingivalis LPS 1690 (Figure 2a). E. coli LPS significantly upregulated both MMP-1 and MMP-2 mRNA expression. TIMP-1 transcript was differently modulated by P. gingivalis LPS 1435/1449 and LPS 1690. The former significantly upregulated its expression at 24 and 48 h, so did E. coli LPS at 48 h.

P. gingivalis LPS 1690 significantly upregulates MMP-3 protein expression
Both dose-and time-dependent experiments showed that MMP-3 protein was differentially modulated by P. gingivalis LPS 1435/1449 and LPS 1690 in consistent with its transcript expression profile ( Figure 3). P. gingivalis LPS 1690 at 1 μg/ml and 10 μg/ml significantly upregulated MMP-3 protein expression in a timedependent manner (12-48 h) (Figure 3c). The MMP-3 level detected in the culture supernatant was greatly higher than that in the cellular fraction (Figures 3a and  b). Similar observations occurred in E. coli LPS-treated cells. Moreover, the MMP-3 level induced by P. gingivalis LPS 1690 was significantly greater than that stimulated by P. gingivalis LPS 1435/1449 (Figures 3a-c).
Next, western blot analysis confirmed that MMP-3 protein markedly increased in P. gingivalis LPS 1690 -and E. coli LPS-treated cells at 48 h, while P. gingivalis LPS 1435/1449 did not induce MMP-3 at a notable level (Figures 4a and c).
The MMP-2 protein expression is not significantly affected by P. gingivalis LPS and E. coli LPS Basal expression of MMP-2 was observed at 24 h, and increased at 48 h ( Figures 4 and 5). With reference to the control, P. gingivalis LPS and E. coli LPS did not significantly affect the expression levels of MMP-2 proteins (Figures 4a and b). Gelatin zymograms revealed that the MMP-2 presented in two forms including pro-MMP-2 (72 kDa) and active-MMP-2 (68 kDa). In both culture supernatant (Figure 5a and b) and cellular fraction (Figure 5c  Notably, U1026 inhibited MMP-3 expression to a greater extent with reference to SB202190. The expression of MMP-3 was not significantly reduced by IKK-2 inhibitor IV in P. gingivalis LPS 1690 -treated cells, whereas it significantly suppressed MMP-3 in E. coli LPS-treated cells ( Figure 6).

Discussion
Periodontal disease is a complex inflammatory disease initiated by pathogenic plaque biofilms and results in destruction of tooth-supporting tissues and alveolar bone [17,18]. Proteolytic enzymes like MMPs play a major role in the degradation of collagens in periodontal tissues. The expression and regulation of MMPs and TIMPs in HGFs are therefore crucial for maintenance of tissue homeostasis and periodontal health. Although many studies have been performed to elucidate the mechanisms involved in the synthesis and regulation of MMPs in periodontal research, no studies are available (1 ng/ml, 10 ng/ml, 100 ng/ml, 1 μg/ml and 10 μg/ml) for 24 h. The expression of mRNAs was measured by real-time qPCR. Each bar represents the mean ± SD of three independent experiments with three replicates. *Significant difference (p < 0.05) as compared with the controls without LPS treatment.
on the effect of P. gingivalis LPS structural heterogeneity on the expression of MMPs and the underlying regulatory mechanisms.
MMP-3 is known as stromelysin which has both elastinolytic and collagenolytic activities that degrade basement membrane components such as laminin, elastin fibronectin as well as collagen types II, III, IV, V, IX, X and XI [8,19]. Its level could significantly increase following the stimuli of pro-inflammatory cytokines, growth factors and LPS [14,[20][21][22]. It has been shown that HGFs could upregulate the expression of MMP-3 due to the effects of pro-inflammatory cytokines such as IL-1β and TNF-α [23][24][25]. The current study showed that the expression of MMP-3 mRNA and protein was markedly upregulated by P. gingivalis LPS 1690 , whereas no induction was observed in cells treated with P. gingivalis LPS 1435/1449 , indicating that the heterogeneous lipid A structures of P. gingivalis LPS may differentially modulate the expression of MMP-3 in HGFs. Moreover, TIMP-1 expression was differently modulated by the two isoforms of P. gingivalis LPS as well. It functions as an inhibitor of MMPs by forming non-covalent complexes with MMPs. It has recently been shown that MMP-3 and TIMP-1 variants may significantly contribute to chronic periodontitis and disease progression [26]. The imbalance between MMPs and TIMPs has been implicated in periodontal tissue destruction [27].
P. gingivalis has long been recognized as a major periodontopathogen [28]. Recently, it is regarded as a keystone pathogen due to its ability to significantly influence the oral microbial community by modulating the innate host response [29,30]. Moreover, this bacterium adopts multiple pathogenic mechanisms to evade or subvert the host immune system [31][32][33]. Notably, P. gingivalis LPS exhibits significant structural heterogeneity with both isoforms of LPS 1435/1449 and LPS 1690 , and our recent studies show that they differentially affect the innate host defense and underlying signaling pathways, thereby contributing to the pathogenesis of periodontal disease [4,34,35]. The current observation that the different isoforms of P. gingivalis LPS modulate the expression of MMP-3 and TIMP-1 may represent an additional pathogenic mechanism adopted by this noxious species to disturb the physiological tissue remodeling and tissue homeostasis, leading to the initiation of periodontal disease.
P. gingivalis and its virulence attributes such as LPS can stimulate various cells types to secrete MMPs including MMP-3 [36,37]. On the contrary, some studies have suggested that P. gingivalis LPS may not induce MMPs such as MMP-1, -2 and −9 [38]. A study performed on gingival epithelial cells using P. gingivalis LPS and E. coli LPS showed that neither LPS nor IL-1β induced MMP-2 or MMP-9 [39]. Studies on tissue models such as synovial membranes dissected from rat knee joints showed induction of MMP-1, -3 and −9 mRNA levels but not MMP-2 in response to LPS stimulation [40]. However, foregoing studies have not considered the heterogeneous nature of bacterial LPS lipid A structures. Therefore, the conflicting findings of the previous studies could to some extent be due to different isoforms of P. gingivalis LPS as demonstrated in the present study.
In the present study, E. coli LPS-treated HGFs exhibited rapid and significant induction of MMPs 1 and 2 mRNAs with reference to the cells treated with P. gingivalis LPS 1690 . One possibility for this observation may be the higher responsiveness of HGFs to hexaacylated nature of the E. coli LPS as compared to the penta-acylated structure of P. gingivalis LPS 1690 . This notion is consistent with previous findings that E. coli LPS is a potent inducer of the production of MMPs in fibroblast-like synovial cells and rat chondrocytes, as well as other innate host response molecules in HGFs and gingival/oral epithelia [41,42]. Moreover, it was noted that both P. gingivalis LPS 1435/1449 and E. coli LPS significantly upregulated the expression of MMP-2 mRNA but not its protein as compared to the controls. A number of factors may account for this finding, such as the stability of mRNA, its processing and splicing patterns, half-life of the target protein and posttranslational modifications [43,44]. Therefore, in the present study increase in MMP-2 mRNA expression level may not be necessarily reflected at its protein level.
TIMPs exhibit high affinity for binding with MMPs and lead to inhibition of their activities. In the present study, TIMP-1 mRNA was upregulated by P. gingivalis LPS 1435/1449 -treated HGFs, while no significant upregulation was observed in P. gingivalis LPS 1690 -stimulated cells. The current results may not be comparable with previous studies in which the structural heterogeneity of LPS was not fully considered [45][46][47][48][49]. This omission may account for the conflicting reports in the literature. Hence, some studies have observed lower TIMP-1 levels in the conditioned media of HGFs in The protein expression levels were measured by ELISA. Each bar represents the mean ± SD of two independent experiments with three replicates. Significant difference as compared with the controls without LPS treatment, *p < 0.05. Significant difference between the cells treated with P. gingivalis LPS 1435/1449 and LPS 1690 respectively, † p < 0.05. Significant difference between the cells treated with P. gingivalis LPS and E. coli LPS respectively, # p < 0.05. response to P. gingivalis LPS [49]. In contrast, other studies have noted the increased expression level of TIMP-1 in gingival crevicular fluid of periodontitis patients [45,47]. Moreover, periodontal treatment could alter the balance between MMP-3 and TIMP-1 [46,48]. Based upon the current findings, further study may be warranted to explore the association of different isoforms of P. gingivalis LPS with periodontal conditions in periodontal patients and the possible effect of periodontal treatment on the expression of these LPS isoforms by P. gingivalis. In addition, the discrepancy observed in TIMP-1 mRNA and protein expression following the stimulation of both P. gingivalis LPS 1435/1449 and E. coli LPS in HGFs could be due to the complex regulation of transcription and translation [43,44].
LPS is the major immuno-stimulatory component of P. gingivalis which has shown to be capable of interacting with TLRs. Binding of LPS to TLRs activates the downstream signal transduction pathways such as NF-ĸB and MAPK [50,51]. Previous studies have suggested that the activation of MMPs could be through both NF-ĸB and MAPK signaling [23,[52][53][54]. The present study demonstrated that p38 MAPK and ERK are critically involved in P. gingivalis LPS 1690 -and E. coli LPS-induced expression of MMP-3 in HGFs. This finding is supported by a previous study that p38 MAPK and ERK1/2 pathways are essential for the expression and regulation of MMPs in various cell types in response to LPS [54]. ERK, JNK and p38 MAPK pathways play vital roles in regulating the expression of MMPs induced by various stimulants such as cytokines [53,55,56]. It is noteworthy that the nature of the stimuli could result in specific signal transduction pathway in the same cell type. For instance, MAPK inhibitor significantly reduced the MMP-3 production in HGFs stimulated with IL-1β, but not with epidermal growth factor [23]. In addition, NF-ĸB pathway may be involved in regulation of MMP-3 expression in rabbit dermal fibroblasts, human saphenous vein and rabbit aortic smooth muscle cells [57,58]. The present study showed that NF-ĸB signaling is not critically involved in LPS-induced MMP-3 expression in HGFs. Notably, the MAPK pathway but not NF-κB was significantly involved in the regulation of MMP-3 expression in HGFs in both mRNA and protein levels. Previous studies have also proven that the expression of MMP-3 is mainly mediated through P38 MAPK, ERK and tyrosine kinase pathways, but not through NF-κB pathway [23,59,60]. Moreover, although a study reported that the activation of NF-κB could be important for MMP-3 secretion, no consensus NF-κB binding site was identified in the MMP-3 gene promoter [61,62]. It suggests that NF-κB may regulate the expression of this gene through different binding sites or interacting with other transcription factors [59]. Therefore, within the context and limitations of the present study, it is tempting to speculate that MAPK pathway may be crucial for MMP-3 expression in HGFs in response to P. gingivalis LPS 1690 . Furthermore, it would be interesting to extend the study to other cells types in human gingiva like gingival epithelial cells to ascertain whether MAPK pathway plays a predominant role in the expression and regulation of MMP-3 in other cells of oral tissues.

Conclusions
The present study reveals that HGFs significantly express MMP-3 in response to penta-acylated P. gingivalis LPS 1690 and hexa-acylated E. coli LPS, but not to the tetra-acylated P. gingivalis LPS 1435/1449 in HGFs. Blocking p38 MAPK and ERK pathways significantly downregulates P. gingivalis LPS 1690 -and E. coli LPS-induced expression of MMP-3. These findings indicate that the heterogeneous lipid A structures of P. gingivalis LPS differentially modulate the expression of MMP-3 in HGFs, which may play a role in periodontal pathogenesis.

Methods
Preparation, purification and identification of P. gingivalis LPS P. gingivalis LPS was isolated from P. gingivalis ATCC 33277 (the American Type Culture Collection, Rockville, MD). LPS was prepared by the cold MgCl 2 -Ethanol procedure followed by lipid extraction and conversion to sodium salts as previously described [63,64]. Optical densities were measured at 280 nm and 260 nm to verify the nucleic acid and protein contamination. LPS preparations were further treated to remove the endotoxin protein and the final protein contamination was less than 0.1% [65]. The fatty acid composition of P. gingivalis LPS was further analysed by Gas chromatographic-mass spectroscopy. Then two separate extractions of P. gingivalis LPS with tetra-(LPS 1435/1449 ) and pentaacylated (LPS 1690 ) lipid A structures were generated, and their structures were verified by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. The canonical hexa-acylated LPS of Escherichia coli JM 83-wild type strain was used as the reference [66].

Cell culture
HGFs were obtained from Sciencell research laboratories (Carlsbad, CA, USA) and cultured according to the manufacturer's instructions [67,68]. Continuous subcultures up to 10th passage contained homogeneous, slim and spindle-shaped cells growing in characteristic swirls. Third to fourth passages of HGFs without any signs of senescence were used for all experiments as described in our previous study [4]. Cells were pretreated with IKK-2 inhibitor IV (NF-ĸB inhibitor), SB202190 (p38 MAPK inhibitor) and U1026 (ERK inhibitor) in serum free medium for 1 h, and then treated with P. gingivalis (Pg) LPS 1690 (1 μg/ml) and E. coli LPS(1 μg/ml) for additional 12 h. Total RNA was harvested and MMP-3 mRNA levels were determined by real-time qPCR. Cell culture supernatants were collected and the protein expression level was measured by ELISA. The histogram shows quantitative representations of the MMP-3 mRNA levels of three independent experiments. Each value represents the mean ± SD. *Significant difference (p < 0.05) as compared with the controls. # Significant difference (p < 0.05) as compared with the cells treated with P. gingivalis LPS 1690 or E. coli LPS alone.

Stimulation of HGFs by heterogeneous P. gingivalis LPS
The cells suspended at 10 5 cell/ml were seeded on sixwell-plates and grown until confluent at 37°C with 5% CO 2 in a culture medium for fibroblasts consisting of basal medium with 2% fetal bovine serum, penicillin/ streptomycin (0.01% w/v) and fibroblast growth supplement. Once the cells were over 90% confluent, fibroblast medium (FM) was replaced entirely with serum free and animal component free-medium (FM-acf ) for the doseand time-dependent experiments. In the dose-dependent assay, cells were stimulated with P. gingivalis LPS 1435/1449 , P. gingivalis LPS 1690 or E. coli LPS in the media containing various doses of LPS (0.001 μg/ml −10 μg/ml). Subsequently, 1 μg of LPS was selected as the appropriate dose for the following time-dependent experiments. Cells were incubated with P. gingivalis LPS or E. coli LPS at 1 μg/ml and harvested at 2, 12, 24 and 48 h. Cells without LPS treatment were designated as the controls. Culture supernatants were collected and centrifuged to remove the cellular debris and stored at −70°C for subsequent protein assays. Cellular fraction was then washed with PBS and collected for mRNA and protein extraction.

RNA extraction, cDNA synthesis and real-time qPCR
Total RNA extraction, cDNA transcription and real-time qPCR for MMPs1-3 and TIMP-1 were performed as previously described [17]. In brief, total RNA was extracted from the homogenized HGFs using RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions [35]. cDNA was synthesized by reverse transcriptase-PCR at 43°C for 90 min in a 20 μl of reaction mixture containing 1 μg of total RNA, 1 μl Real-time primer pairs were designed using ABI software to amplify a sequence that contains two or more exons whenever possible. The amplification efficiencies of the primers used were above 90%. The specific sequences for each pair of primers are listed in Table 1. β-actin was amplified as an internal control. The real-time qPCR reaction conditions were set at 95°C for 10 min followed by 40 cycles at 95°C for 15 s and 60°C for 60 s. The results were analyzed using the comparative cycle threshold (Ct) method as previously described [35]. The expression level of each gene was normalized to a β-actin (ΔCt) and the fold changes for each gene were calculated by comparing the test and control samples from the ΔΔCt values.

Total protein extraction and detection of MMP-3 by ELISA
Total proteins were extracted from homogenized HGFs using CellLytic TM MT-mammalian cell lysis extraction reagent (Sigma, USA). Protein concentrations in both of the cell-bound fraction and culture supernatant were measured respectively by BCA protein assay kit (Pierce, Thermo Scientific, USA) according to the manufacturer's instructions. Enzyme-linked immunosorbent assay (ELISA) was performed to confirm the expression of MMP-3 proteins (BioRad Laboratories, Hercules, CA, USA). The protein expression in both cell lysate and culture supernatants were measured following manufacturer's instruction with the minimal detectable concentration of 0.009 ng/ml. No cross reactivity or no interference was observed with recombinant MMP-3. The absorbance values were determined by a micro-plate reader (Victor, Vienna, VA, USA) at optical absorbance of 450 nm and the final concentration was determined with reference to a standard curve. Experiments were repeated two times with three biological replicates.

Western blot analysis for MMP-2, -3 and TIMP-1 proteins
Total cell lysates were prepared and 40 μg of cellular extracts were separated by 10% SDS-PAGE gel and subsequently transferred onto a polyvinylidene difluoride membrane (PVDF). The proteins were then blocked against the protein-free blocking buffer (Pierce, Thermo Scientific) for 1 h. Afterwards, membranes were incubated overnight at 4°C with primary antibodies against polyclonal rabbit anti-human IgG; MMP-2 (1:1000; Cell signaling), MMP-3 (1:1000; BioVendor) and TIMP-1 (1:1000; Cell signaling), and incubated with horseradish peroxidase (HRP) conjugated anti-rabbit secondary antibodies (1:10000). Visualization of the immunoreactive proteins were accomplished by the use of SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Rockford, IL, USA) and exposed to X-ray films. α-Tubulin was used as the internal loading control (1:1000; Cell signaling). The detected bands were scanned on a calibrated densitometer, GS-800 and assessed by the imageJ software-based analysis (http://rsb.info.nih.gov/ij/) to quantify the integrated density.
Gelatin zymography for enzymatic activity of MMP-2 SDS-PAGE gelatin zymography was performed to observe the enzymatic activity of MMP-2. Supernatants and cellular proteins were collected from cells grown in serum-free medium at 24 h and 48 h as described above. Centrifugal filter devices (Amicon Ultra-0.5-Millipore USA) with a cut off value of 30000 NMWL (Nominal Molecular Weight Limit) were used to concentrate the supernatants. Culture supernatants or cellular extracts (40 μg) were mixed with 2 × non-reducing sample buffer without β-mercaptoethanol (0.125 M Tris-HCl at pH 6.8, 4% SDS, 20% glycerol and 0.05% bromophenol blue). Proteins were separated by 10% Tris-glycine polyacrylamide gel copolymerized with 0.1% gelatin as a substrate. After electrophoresis, gels were washed in renaturation buffer (2.5% Triton X-100 in 50 mM Tris-HCl at pH 7.5) for 1 h and incubated for 20 h at 37°C in incubation buffer (0.15 M NaCl, 10 mM CaCl2 and 0.02% NaN3 in 50 mM Tris-HCl at pH 7.5). Gels were stained with 5% Coomassie blue and destained with 7% methanol and 5% acetic acid to reveal zones of lysis within the gelatin matrix. Areas of enzymatic activity appeared as clear bands over the dark background.

Signal transduction pathways involved in LPS-induced MMP-3 expression in HGFs
Specific pharmacological inhibitors for NF-κB activity, IKK-β inhibitor (IKK-2 inhibitor IV), p38 MAPK activity (SB202190) and ERK activity (U1026) were used to investigate two major signaling pathways potentially involved in the expression and regulation of MMP-3 in HGFs in response to heterogeneous P. gingivalis LPS. Each inhibitor was first dissolved in dimethyl sulfoxide (DMSO) and diluted in DPBS. Cells were pretreated with kinase inhibitors, including 10 μmol/L of IKK-2 inhibitor IV (Merck, USA), 10 μmol/L of SB202190 (Calbiochem Biosciences Inc, La Jolla, CA, USA) and 15 μmol/L of U1026 (Cell Signaling, USA) respectively for one hour, prior to stimulation with LPS. Afterwards, 1 μg/ml of LPS was added to the medium and cells were incubated for another 12 h. Culture supernatants were collected for analyzing the MMP-3 expression by ELISA. Extracted RNA was subjected to real-time qPCR to detect the MMP-3 transcript expression. Positive controls were the supernatants from the cells treated with LPS alone, whereas the negative controls were incubated with the culture medium alone. In addition, the cells treated with DMSO alone were considered as the vehicle control (data not shown).

Statistical analysis
All experiments were repeated in three assays for realtime qPCR and two assays for ELISA. Results of the experiments were presented as the mean ± SD. The statistical significance of difference between the data sets from the dose-dependent assay was evaluated by student t-test, one-way analysis of variance (ANOVA) and post hoc testing with Bonferroni and LSD methods. Additionally, the repeated-measures of ANOVA were used to determine the differences between data sets from the time-dependent assay. A p-value < 0.05 was considered statistically significant. All statistical analysis was performed using a software program (SPSS 19.0, SPSS Inc, Chicago, IL, USA).