Tissue Samples

Tissue sections from experimental wounds created in palatal attached gingival mucosa in three healthy males (mean age 38 years) that have been previously extensively characterized were used [35,42–46]. Briefly, identical, standardized, full-thickness excisional wounds (about 12 mm long, 2 mm wide and at least 10 mm away from each other) were prepared under local anesthesia in the palatal mucosa in an area between the canine and the third molar using a double-bladed scalpel. The tissue obtained from the initial wounds served as the control samples (day 0 sample). After the surgery, subjects were instructed to use standard dosages of acetaminophen or ibuprofen for postoperative pain control. Wound biopsies were collected at days 3, 7, 14, 28, and 60 after wounding. Immediately after collection, the samples were embedded in Optimal Cutting Temperature Compound (Sakura Finetek Inc., Torrance, CA, USA) and frozen in liquid nitrogen. Tissue sections (6 μm) were cut using a 2800 Frigocut Cryostat Microtome (Leica, Nussloch, Germany), placed on 3-aminopropyltriethoxysilane-coated slides and stored at −86°C until use. For the study, minimum of three tissue/mid-wound sections from two to three subjects at each time point were used.

Cell Culture

Four gingival fibroblast lines (GFBLs) were isolated from clinically healthy attached gingiva from healthy human donors, as previously described [47] (S1 Table). Cells were routinely maintained in Dulbecco’s Modified Eagle’s medium (DMEM), supplemented with 1% antibiotic/antimycotic and 10% fetal bovine serum (FBS) (Gibco Life Technologies, Inc., Grand Island, NY, USA) at 37°C and 5% CO₂. Cells were routinely seeded for experiments when they reached about 95% confluence. Experiments were performed at passages 5 to 10.

Ethics Statement

Gingival tissue donors provided a written informed consent, and procedures were reviewed and approved by the Office of Research Ethics of the University of British Columbia, and complies with the ethical rules for human experimentation that are stated in the 1975 Declaration of Helsinki.

Blocking of C×43 Function by Mimetic Peptides or MFA

To block the C×43 function, GFBLs were seeded on 6-well plates (42,000 cells/cm²) in their normal growth medium. At day 2 when cultures became confluent, culture medium was replaced with serum-free medium. At day 3, cells were treated with 150 μM of Gap27 (SRPTEKTIFII; Biomatik, Cambridge, ON, Canada) [3,48,49], equal molar amount of Gap26 (VCYDKSFPISHVR) [3,48–50], or corresponding scrambled Gap27 (TFEPDRISITK) [33] or Gap26 (YSIVCKPHVFDRS) [50] control peptides, respectively, for up to 24 h before total RNA isolation or collection of cell lysates/conditioned medium for Western blotting. In a set of experiments, cells were treated with increasing concentrations (25, 50, 75, 100 μM) of meclofenamic acid (MFA; M4531, Sigma-Aldrich, St. Louis, MO, USA), a pharmacological connexin inhibitor [51], or corresponding amount of MFA diluent (dH₂O) for 24 h before RNA isolation.

Blocking C×43 Expression by siRNA Technique

To block the expression of C×43, GFBLs were seeded in 6-well plates as described above. After 24 h, siRNA transfection was carried out using Lipofectamine RNAiMax reagent (Invitrogen, Carlsbad, CA, USA). To this end, 22 μl of Lipofectamine RNAiMax and 150 picomoles of C×43 siRNAs (siRNA-1; UUUUGCAAGUGUAAACAGC or siRNA-2; AAUGAAAAGUACUGACAGC; Invitrogen) or control siRNAs (control siRNA-1 ACUUCGACACAUCGACUGC or control siRNA-2 ATCGCAAATCCGGACCTAT; Invitrogen) were mixed with 2.4 ml of Opti-MEM medium (Invitrogen), and incubated at room temperature for 5 min. The mixtures of transfection reagent and individual siRNAs were combined and incubated at room temperature for 20 min for complex formation. Then the reagent was diluted with Opti-MEM to yield final siRNA concentration of 30 nM, and added to the cells. After 5 h in cell culture incubator, the transfection reagent was removed, and cells fed with their normal growth medium overnight. Medium was then replaced with serum-free growth medium, and RNA isolation and sample collection for Western blotting was performed after 48 h.

Real-Time PCR

Real-time PCR analysis was performed according to MIQE guidelines [52] as we have described in detail previously [53]. Briefly, total RNA from cultured GFBLs was isolated using NucleoSpin RNA II kit according to the manufacturer’s protocol (Macherey-Nagel). Total RNA (1 μg) was reverse transcribed using iScript Select cDNA Synthesis Kit (Bio-Rad, Mississauga, ON, Canada) and random oligodeoxynucleotide primers according to the manufacturer’s instructions, as described previously. The primers used for real-time PCR are listed in S2 Table. Real-time PCR amplification was performed on the CF×96 System (Bio-Rad) using the following program: 1 cycle at 94°C for 3 min 35 cycles at 94°C for 10 s, 60°C for 20 s, and reaction completion with reading plate and a melt curve analysis from 65°C to 95°C, 5 s for each 0.5°C. Amplification reactions were conducted for target genes with ubiquitin C (UBC), glyceraldehydes-3-phosphate dehydrogenase (GAPDH), hypoxanthine phosphoribosyltransferase I (Hprt1), TATAA-box binding protein (TBP), and Beta-2-microglobulin (B2M) as reference genes. For a given experiment, at least two reference genes were chosen using the integrative RefFinder tool (http://www.leonxie.com/referencegene.php). Non-transcribed RNA samples were used as a negative control. The PCR reactions were performed in triplicate for each sample. The data was analyzed and is presented based on the comparative Ct method (CFX Manager Software Version 2.1, Bio-Rad).

Preparation of Cell Lysates/Conditioned Medium Samples for Western Blotting

Confluent GFBL cultures were treated as described above. The conditioned medium was then collected and immediately treated with Complete Protease Inhibitor Cocktail (Roche Diagnostics, Manheim, Germany). The samples were concentrated (30–40 times) by centrifugation (5,000g) using Centrifugal Filter Units (Amicon Ultra-4 3K, 3000 MWCO; Millipore, Bedford, MA, USA) for 3 h, and stored at −80°C until use. To collect cell lysates, cells were washed with ice-cold phosphate-buffered saline (PBS), and lysed with a buffer containing 25 mM Tris-HCL (pH 7.6), 100 mM Octyl β̃D-glucopyranoside, 5 mM NaF, 1 mM Na3VO4, (Sigma-Aldrich, St. Louis, MO, USA), and the Complete Protease Inhibitor Cocktail (Roche Diagnostics), dissolved in H₂O. Lysates were collected using a rubber policeman, and filtered through a NucleoSpin Filter (Macherey-Nagel) by centrifugation at 5,000g for 10 min.

Western Blotting

Immunoblotting analysis was conducted as described in detail previously [53]. Briefly, total protein concentration in call lysates/conditioned medium samples was determined using the Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad). Equal amount of protein of each sample was solubilized in SDS sample buffer containing 2-mercaptoethanol (5%) and separated by 10–12% SDS-polyacrylamide gel electrophoresis. The proteins were transferred onto a nitrocellulose membrane (Hybond-ECL membrane, GE Healthcare Bioscience, Buckinghamshire, UK). The nonspecific binding sites were blocked by incubating the membranes in Odyssey Blocking Buffer (LI-COR Biosciences; Lincoln, NE, USA) at room temperature for 1 h, followed by incubation with the primary antibody (S3 Table) at 4°C overnight. After washing with TBS containing 0.1% Tween-20 (TBS-T), the membranes were incubated with an appropriate IRdye-conjugated secondary antibody (1:10,000; LI-COR Biosciences). Dried membranes were then detected using the LI-COR Odyssey infrared reader (LI-COR Bioscience, Nebraska, USA). Intensity of the protein bands was quantitated using ImageJ software (NIH).

The activation of signaling pathways by Gap27 treatment was studied in cell lysates obtained as described above. For the experiments, GFBLs were seeded on 6-well plates, treated with Gap27 (150 μM) or equal amount of control peptide for 1, 2, 6, and 24 h, and cell lysates collected, as above. Western blotting was performed with antibodies against total or phosphorylated forms of β-Catenin and GSK3α/β (β-Catenin pathway), SMAD3 (TGF-β pathway), ERK1/2 and p38 (MAPK pathway) (S3 Table). β-Tubulin was used as a loading control.

To identify latent and active MMPs, a set of cell/conditioned medium samples was treated with or without p-aminophenylmercury acetate (APMA; 1.0 mM, pH = 7.4; Sigma-Aldrich) at 37°C for 4 h to activate latent enzymes [54] prior to gel electrophoresis and Western blotting (data not shown).

Use of Chemical Inhibitors to Block Signaling Pathways

To determine the role of key signaling pathways in Gap27-induced gene expression, we blocked TGF-β pathway with SB431542 (20 μM; Selleckchem, Houston, TX, USA), MEK1/2 with PD184352 (10 μM; Sigma-Aldrich), p38 with SB203580 (10 μM; Cell Signaling, Danvers, MA, USA), GSK3α/β with SB415286 (30 μM; Biomol, Hamburg, Germany), AP1 with curcumin (30 μM; Sigma-Aldrich), and SP1 with WP631 (5 nM; Sigma-Aldrich) in Gap27-treated cells, respectively. To this end, confluent GFBL cultures were pre-incubated with inhibitors at 37°C for 1 h, and then treated with Gap27 (150 μM) with or without the inhibitors in serum-free growth medium for 24 h. All inhibitors were dissolved in DMSO, and control samples were treated with respective amounts of DMSO only. Total RNA was collected for real-time PCR as described above.

Dye Transfer Experiments

To assess the GJ function of C×43, dye transfer assays were performed. To this end, GFBL cultures were generated on gelatin-coated glass coverslips in 24-well plates as described previously [55]. Briefly, the coverslips were incubated in 0.2% gelatin in PBS at 37°C for 1 h. After rinsing with PBS, coverslips were incubated in 1% glutaraldehyde at room temperature for 30 min, then washed with PBS, followed by incubation with DMEM at 37°C for 30 min. Coverslips were then washed with PBS and stored at 4°C or used immediately. To assess dye transfer through GJs by scrape loading, cells (GFBL-DC) were seeded on the coverslips in their normal growth medium as described above for 24 h, and then serum-starved in DMEM for another 24 h. Confluent cultures were then pre-incubated with Gap27 or the control peptide (150 μM; 24 h), or with MFA (50 μM; 1 h) or vehicle control (dH₂O; 1 h) in DMEM at 37°C, media was removed, and a scrape wound was created through the cell layer with a 10 μL pipette tip, and cells incubated as above with 0.5% Lucifer Yellow (Molecular Probes Inc., Eugene, OR, USA) in PBS+ (containing 1 mM Ca²⁺ and Mg²⁺) for 5 min at 37°C. Cells were rinsed once with PBS+ and then fixed with 4% formaldehyde at room temperature for 20 min. Nuclei were then stained with 300 nM DAPI (Molecular Probes Inc.) in PBS for 5 min. In a set of experiments, cells were transfected with C×43 siRNA-1, siRNA-2 or control siRNA (30 nM) as above before seeding on the coverslips for 24 h, serum-starved for 24 h, and then subjected to the dye transfer experiment as above. Samples were mounted with Immu-Mount solution (Thermo Scientific, Pittsburgh, PA, USA), examined using the ECLIPSE 80i Microscope (Nikon, Tokyo, Japan), and images captured using NIS-Elements BR software (Version 4.20, Nikon).

Scratch Wound Cell Migration Assay

To assess role of C×43 in cell migration, GFBLs were grown on gelatin-coated glass coverslips, serum-starved, and pretreated with Gap27 or control peptides (150 μM) for 24 h as above. A wound was then created through the cell layer using a 100 μL pipette tip. Cells were then cultured in DMEM with the peptides in a cell culture incubator, and wound closure recorded over time by standardized digital images taken from the samples using a phase contrast microscope (Nikon Eclipse, TS100) and a digital camera (Nikon Coolpix 995). In a set of experiments, cells were transfected with C×43 siRNA-1 or -2 and control siRNA-1 or -2 (30 nM) separately as above, seeded on gelatin-coated glass coverslips for 24 h, and serum-starved for 6 h before wounding. Experiments were performed in triplicate per treatment group, and images were captured from three to four different areas of the wound on each coverslip. Wound closure rate was determined measuring the area of the open wound at each time point relative to the area of the wound at the time of wounding using the Adobe Photoshop for Mac software (https://www.adobe.com).

Immunostaining

For immunostaining, confluent GFBLs were seeded on gelatin-coated glass coverslips in their normal growth medium (24 h) and then serum starved (24 h) as described above, fixed with 4% formaldehyde at room temperature for 20 min, and then permeabilized using 0.5% Triton X-100 in PBS for 4 min. All samples were then blocked with PBS+ containing BSA (10 mg/ml) and glycine (1 mg/ml) at room temperature for 30 min, followed by an incubation with the primary antibody (S3 Table) diluted in PBS containing BSA (1 mg/ml) in a humidified chamber at 4°C overnight. The samples were then washed with PBS containing BSA (1 mg/ml) and 0.01% Triton X-100, and incubated with an appropriate Alexa-conjugated secondary antibody (1:200 dilution; Alexa 488/594; Molecular Probes Inc.) at room temperature for 1 h. Nuclei were then stained with DAPI and samples mounted as above.

For immunostaining of human gingival tissue samples, sections were fixed with cold acetone (−20°C) at room temperature for 5 min. Samples were then washed, blocked, and incubated separately with each primary antibody (S3 Table) overnight, followed by washing, incubation with appropriate Alexa 488 and 594-conjugated secondary antibodies, nuclear staining with DAPI and mounting as above. Images were acquired using optical sectioning at 1 μm (ECLIPSE 80i Microscope; Nikon), and are presented as z-stacks created by the NIS-Elements BR software (Nikon). Control stainings were performed by omitting the primary antibody incubation step.

Statistical Analysis

All data is presented as mean +/− standard error of the mean (SEM) from repeated experiments. Statistical analysis was performed by using Student’s t-test, p<0.05 was considered statistically significant. Values obtained from the real-time PCR by the comparative Ct-method were Log2 transformed for statistical testing [56].

C×43 Is the Major Connexin Expressed by Gingival Fibroblasts

Previous studies have shown that C×43 is present in human skin, periodontal ligament and gingival fibroblasts, but that skin and periodontal ligament fibroblasts also express C×32, C×40 and C×45 [7,17,57]. Therefore, we assessed the expression of C×32, C×40, C×43 and C×45 in four parallel human gingival fibroblast (GFBL) lines from different donors in confluent monolayer cultures that allowed abundant cell-cell contacts to form. Real-time PCR (Fig. 1A) and Western blotting (Fig. 1B) analysis showed that GFBLs expressed C×43 as their major connexin protein, with moderate levels of C×45, very low levels of C×32, and no expression of C×40. Immunolocalization of C×43 and C×45 showed a punctate staining, likely representing connexin plaques. Some of these connexin plaques localized to long cellular processes contacting nearby cells likely representing GJs (Fig. 1C). In addition, they localized to areas with no apparent cell-cell contacts, possibly representing intracellular and/or cell surface non-junctional hemichannel plaques (Fig. 1C). In general, the number of C×43- compared to C×45-postive plaques was markedly higher (Fig. 1C), reflecting the real-time PCR and Western blotting analysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Gingival fibroblasts express C×43 as their major connexin protein.

(A) Results show real-time PCR analysis of major connexins previously described in fibroblasts (C×32, C×40, C×43 and C×45) in cultured human gingival fibroblasts from four different individuals (GFBL-HN, GFBL-CM, GFBL-OL and GFBL-DC). All cell lines expressed C×43 as their major connexin, with moderate levels of C×45, low level of C×32, and no expression of C×40. Range of Ct-values obtained from real-time PCR is indicated below each gene name. (B) Similar findings were found when the same connexins were analyzed in cell lysates using Western blotting. (C) Immunostaining of a representative confluent cell culture (GFBL-DC) for C×43 and C×45. Gingival fibroblasts contained numerous C×43-positive plaques, while much fewer similar structures positive for C×45 were noted. In general, connexin-positive plaques were localized at cell-cell contact areas, possibly representing GJs (arrows), and other areas not associated with cell-cell contacts (arrowheads).

https://doi.org/10.1371/journal.pone.0115524.g001

C×43 Is Strongly Down Regulated in Gingival Epithelium and Fibroblasts During Wound Healing

In order to assess expression of C×43 during gingival wound healing, we compared localization of C×43 in unwounded human gingiva, and in experimental excisional gingival wounds 3, 7, 14, 28 and 60 days after wounding by immunostaining (Fig. 2 and S1 Fig.). Fibroblasts were identified based on their elongated, spindle-shaped morphology, and positive immunostaining for vimentin, a mesenchymal cell marker highly expressed in fibroblasts [58]. C×43 was present in unwounded gingival epithelium where it showed strong, punctate staining in the cell-cell contacts at the spinous layer. In addition, a weaker and sparse staining was noted between some basal and granular layer cells (S1 Fig.). A punctate, positive staining for C×43, likely representing C×43 plaques [23], was also noted throughout connective tissue (Fig. 2A and B), where it associated with vimentin-positive fibroblast-like cells. Many of the C×43-positive plaques were fairly large in size (>1 μm in diameter), and some localized to long processes extending from these cells (Fig. 2B). At day 3 after wounding, abundance of C×43 was strongly reduced in the migrating wound epithelium (S1 Fig.), and in the fibroblasts residing at the wound edge and migrating into the wound (Fig. 2C and D and S1 Fig.). At day 7 when wound epithelium had completely covered the wound, and granulation tissue formation was underway, C×43 was abundantly present in 2–3 most basal cell layers of the wound epithelium, while the spinous layer showed some weak positive staining (S1 Fig.). In general, in the connective tissue cells at the wound edge (Fig. 2E and S1 Fig.), and in the wound granulation tissue (Fig. 2F and S1 Fig.), only very few C×43-positive plaques were noted. At day 14, C×43 was still most abundant in 2–3 most basal wound epithelial cell layers, but its staining was now increased also in more suprabasal cells (S1 Fig.). Our previous analysis of these same wounds has shown that at this stage, the granulation tissue had started contraction, contained abundantly α-SMA-positive fibroblasts, and was being remodeled to wound connective tissue [35,44,45]. At this stage, vimentin-stained fibroblast-like cells at the wound edge showed some C×43-positive plaques (S1 Fig.). In general, the highly cellular wound connective tissue showed only very few such structures in vimentin-positive cells (Fig. 2G and H), or in M2 macrophages (S2 Fig.), also abundantly present in these same wounds at this stage [37], or in α-SMA-positive myofibroblasts (S3 Fig.). At day 28, while the typical gingival epithelial long rete pegs were not yet formed, abundance and localization of C×43 in the epithelium at the wound site was almost similar to normal unwounded tissue, being weakly present in the basal cells, and most strongly stained in the spinous layer (S1 Fig.). At this stage, wound contraction was still underway, but cellularity and the number of myofibroblasts and M2 macrophages had dramatically decreased [35,37,44,45]. Vimentin-stained fibroblast-like cells at the wound edge and within the newly formed wound connective tissue displayed increased number of C×43-positive plaques as compared with earlier time points, but the number and size of the plaques was clearly reduced as compared to unwounded tissue (Fig. 2I and J and S1 Fig.). Few M2 macrophages that were still present in the wounds showed very little C×43 positive staining (S2 Fig.). At day 60 when the epithelium had reformed long rete pegs and connective tissue structure was normalized, the localization and abundance of C×43 was similar to unwounded tissue both at the epithelium and connective tissue at the wound site (Fig. 2K and L and S1 Fig.).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. C×43 is down regulated in gingival fibroblasts during wound healing.

Representative immunostainings of C×43 (red) and vimentin (green; a mesenchymal cell marker) in unwounded human oral mucosal connective tissue (attached gingiva) (A and B), and in gingival granulation and wound connective tissue 3 (C and D), 7 (E and F), 14 (G and H), 28 (I and J) and 60 days (K and L) post-wounding. (A and B) In the unwounded gingival connective tissue, abundant C×43 immunoreactivity was present as punctate staining, likely representing C×43 plaques, in vimentin-positive fibroblast-like cells throughout the tissue. (C and D) At day 3 post-wounding, C×43 was down regulated in fibroblasts at wound edge as compared to unwounded tissue (A and B). First fibroblasts that had migrated into the wound area showed no immunoreactivity for C×43 (D). (E–H) At day 7 (E and F) and 14 (G and H) post-wounding, very few C×43 positive structures were noted in fibroblasts in the highly cellular granulation and connective tissue. (I and J) At day 28 after wounding, abundance of C×43-positive plaques in connective tissue cells in the newly formed connective tissue at the wound area was increased as compared to earlier time points. However, size of these plaques was clearly smaller than in the unwounded tissue (A and B). (K and L) At day 60 after wounding, structure of the connective tissue formed at the wound area was closely similar to unwounded tissue. Size and number of C×43-positive plaques in fibroblast-like cells in the regenerated wound area was similar to the unwounded tissue (A and B). Data shown represents minimum of three sections stained in parallel samples from two to three individual donors at each time point. CT: connective tissue; W Edge: wound edge; W Area: wound area; GT: granulation tissue; WCT: wound connective tissue. Nuclear staining (blue) was performed using DAPI.

https://doi.org/10.1371/journal.pone.0115524.g002

Blocking of C×43 Function Regulates Significantly Expression of a Distinct Set of Wound Healing-Associated Genes in Gingival Fibroblasts

Having established that C×43 is the major connexin expressed by GFBLs, and that abundance of C×43 plaques is strongly reduced during gingival wound healing, we wanted to find out the functional significance of this down regulation. Down regulation of the number of C×43-positive plaques in wound fibroblasts can result in reduced hemichannel and GJ function. Treatment of cells with C×43 mimetic peptides can be used to specifically block these channel functions [59], which promotes connective tissue wound healing in vivo [60]. Therefore, we treated GFBLs with Gap27, a mimetic peptide corresponding to the 11-amino acid sequence in the second C×43 extracellular loop, that blocks C×43 channel functions [59,61]. Treatment with Gap27 (150 μM) did not markedly affect cell morphology (S4 Fig.). However, it reduced GJ- mediated dye transfer, similar to meclofenamic acid (MFA), a pharmacological connexin inhibitor [51], also in this model as expected (Fig. 3). Gap27 treatment also significantly promoted GFBL migration in the scratch wound healing model (S5 Fig.), as previously described for skin fibroblasts [2]. Thus, reduced C×43 abundance/function may promote fibroblast recruitment at the early stages of human gingival wound healing. In addition to being absent from migrating fibroblasts in early wounds, C×43 abundance was also strongly suppressed in fibroblasts abundantly present inside gingival granulation and wound connective tissue at day 7–28 post-wounding. This coincides with resolution of inflammation, angiogenesis, ECM deposition, myofibroblast differentiation, contraction and remodeling stages of wound healing [37]. Therefore, we next assessed the effect of suppressing C×43 function by Gap27 in confluent cell cultures on expression of genes important specifically for these stages of wound healing. Real-time PCR results showed that Gap27 treatment significantly upregulated expression of 15, and down regulated 7, of the 54 genes analyzed (Tables 1–4). The selected genes encoded proteases (MMPs and their inhibitors), important in regulation of inflammation, angiogenesis and ECM remodeling [62,63], molecules involved in intracellular ECM degradation (Endo180 and Cathepsin K) (Table 1), ECM proteins (fibrillar and matricellular proteins, and small leucine-rich proteoglycans), cell contractility and myofibroblast-associated proteins (Table 2), TGF-β signaling associated molecules and cytokines involved in inflammation, angiogenesis and re-epithelialization (CXCL12, FGF-2, VEGF-A, IL1β, IL10 and TNF-α) [37] (Table 3), and cell-cell junction proteins (connexins and cadherins) [64] (Table 4). To rule out significant but potentially biologically irrelevant minor changes, we used minimum of +/− 1.5-fold change threshold for the significantly regulated genes. This yielded 17 significantly regulated genes of which 11 were upregulated (MMP-1, -3, -10 and -14, TIMP-1 and -3, Tenascin-C, TGF-β1, VEGF-A, C×43 and Cadherin-2), and 6 down regulated (Collagen type I, Decorin, Fibromodulin, α-SMA, NMMIIB and CXCL12) (Tables 1–4). The responses were concentration-dependent at least up to 300 μM peptide concentration (data not shown).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Gap27 and MFA suppress GJ-mediated dye transfer in gingival fibroblasts.

(A–F) Confluent GFBL-DC cultures maintained in DMEM were scrape-loaded with Lucifer Yellow (0.5%; green) in the presence of control peptide (A and B; 150 μM), Gap27 (C; 150 μM), vehicle (dH₂O; D and E), or MFA (50 μM; F), and dye transfer was followed for 5 min. Treatment of cells with Gap27 (C) or MFA (F) markedly reduced dye transfer as compared to control samples treated with the control peptide (A and B) or vehicle (D and E). Results show representative images from minimum of three repeated experiments. For the experiments, cells were pretreated with Gap27 and control peptide or MFA and vehicle for 24 h or 1 h before the experiments, respectively. Magnification bars: 50 μm.

https://doi.org/10.1371/journal.pone.0115524.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Blocking of C×43 function with Gap27 treatment modulates significantly expression of genes involved in protein degradation during wound healing in gingival fibroblasts.

https://doi.org/10.1371/journal.pone.0115524.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Blocking of C×43 function with Gap27 treatment modulates significantly expression of extracellular matrix proteins and cell contractility and myofibroblast-associated genes in gingival fibroblasts.

https://doi.org/10.1371/journal.pone.0115524.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Blocking of C×43 function with Gap27 treatment modulates significantly expression of genes involved in TGF-β signaling and encoding VEGF-A and CXCL12/SDF-1α in gingival fibroblasts.

https://doi.org/10.1371/journal.pone.0115524.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Blocking of C×43 function with Gap27 treatment upregulates significantly expression of C×43 and Cadherin-2 expression involved in formation of cell-cell junctions.

https://doi.org/10.1371/journal.pone.0115524.t004

In order to find out, whether the Gap27-mediated regulation of gene expression is a common property of GFBLs, we assessed its effect on expression on a set of the above genes in three parallel GFBL lines in the same experiment (Fig. 4). The findings confirmed a gene expression response to Gap27 treatment in all GFBL lines that was consistent with initial findings using GFBL-DC (Tables 1–4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Effect of Gap27-mediated blocking of C×43 function on gene expression in parallel gingival fibroblast lines.

Confluent cultures of gingival fibroblasts from three different individuals (GFBL-HN, GFBL-CM and GFBL-DC) were treated with Gap27 or control peptide (150 μM) for 24 h, and expression of a set of genes involved in wound healing was analyzed by real-time PCR. Results represent mean +/− SEM of mRNA expression relative to control peptide-treated cells from triplicate samples in one experiment (*p<0.05, **p<0.01, ***p<0.001; Student’s t-test). Horizontal line indicates relative mRNA expression for the control-peptide treated samples. EDA-FN: Extra Domain A-Fibronectin; EDB-FN: Extra Domain B-Fibronectin; TN-C: Tenascin-C; CTGF: Connective Tissue Growth factor (CCN2); α-SMA: α-Smooth Muscle Actin; VEGF-A: Vascular Endothelial Growth Factor-A.

https://doi.org/10.1371/journal.pone.0115524.g004

To confirm specificity of the Gap27 treatment, we also blocked C×43 function by Gap26, another connexin mimetic peptide corresponding to a 13 amino acid sequence in the first C×43 extracellular loop [59,61], and by MFA, and assessed expression of a set of genes as above. Gap27 and Gap26 treatments caused in general similar gene expression changes in GFBLs, although the magnitude of change slightly varied (S6 Fig.). MFA treatment induced a concentration-dependent increase in MMP-1 and -10, Tenascin-C and VEGF-A expression, and down regulated CXCL12 (S7 Fig.), similar to Gap27 (Tables 1–3) and Gap26 treatment (S6 Fig.). Interestingly, Gap27 (Table 4), Gap26 (S6 Fig.) and MFA (S7 Fig.) induced up to about 2-fold increase in expression of C×43 mRNA. Likewise, Gap27 and Gap26 caused about 2-fold increase in C×43 protein level, while MFA had no effect (data not shown).

In a set of experiments, we also suppressed C×43 expression in three parallel GFBL lines using two different C×43 siRNAs (C×43 siRNA-1 and -2), and studied the expression of the above wound healing-associated genes. Treatment of GFBLs with both C×43 siRNAs consistently resulted in about 80% down regulation of C×43 at both mRNA and protein levels, while expression of the two other connexins expressed by these cells, C×32 and C×45, were not affected (data not shown). C×43 siRNA treatment with both C×43 siRNA-1 and -2 effectively suppressed GJ-mediated dye transfer (S8 Fig.), but did not have a significant effect on fibroblast migration in the scrape-wound assay (S5 Fig.). Similar to Gap27 treated cells, C×43 siRNA treatments significantly increased expression of Tenascin-C (1.46 +/− 0.15-fold change; p<0.05) and reduced expression of Collagen type I (0.89 +/− 0.03-fold change; p<0.001) as compared to control treatment. However, possibly due to incomplete C×43 down regulation, the gene expression responses to C×43 siRNA treatment were small with none of the studied genes reaching the +/− 1.5-fold change threshold (data not shown).

Characterization of Proteins Regulated by Gap27 Treatment in Gingival Fibroblasts

As gene expression analysis showed that blocking of C×43 function by Gap27 strongly regulated expression of several genes (Tables 1–4), we further assessed the protein levels of a set of genes that showed significant, minimum of +/− 1.5-fold change relative to control treatment in the real-time PCR analysis. Western blotting analysis showed that treatment of GFBLs with Gap27 (Fig. 5) resulted in markedly increased secretion of active and/or total MMP-1 (Fig. 5A and B) and −10 (Fig. 5H and J) compared to control treatments. In addition, Gap27-treated cells produced significantly elevated levels of pro-MMP-3 in the cell layer (Fig. 5F and G). Expression of mRNA for Decorin (Table 2), a small leucine-rich proteoglycan that regulates cell functions involved in wound healing and fibrosis, and VEGF-A (Table 3), a potent pro-angiogenic growth factor [37], were also strongly suppressed and upregulated, respectively, by Gap27 treatment. Accordingly, Western blotting showed robust down regulation of Decorin and upregulation of VEGF-A levels in the cell culture medium of Gap27-treated GFBL cultures (Fig. 6). Gap27 treatment also significantly increased C×43 expression at mRNA (Table 4) and total protein levels (Fig. 7). However, no changes in the relative intensities of the three bands corresponding to the differently phosphorylated forms of C×43 recognized by the C×43 antibody (P0: pS368; P1:pS279/282 and pS255; P2: pS262) [65,66] were noted in the Western blots (Fig. 7A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Blocking of C×43 by Gap27 resulted in significantly increased secretion of active MMP-1 and MMP-10, and pro-MMP-3 by gingival fibroblasts.

Confluent cultures of gingival fibroblasts (GFBL-DC) were treated with Gap27 or control peptide (150 μM) for 24 h, and abundance of MMP-1 (A–D), MMP-3 (E–G), and MMP-10 (H–J) in the conditioned medium and cell layer was analyzed by Western blotting. (B, D, G and J) Quantitation of MMP levels in Western blots shows mean +/− SEM from three independent experiments (*p<0.05, ***p<0.001; Student’s t-test). Sample loading for cell layer fraction was normalized for β-Tubulin levels. Identity of active and pro-forms of the enzymes was confirmed by pretreatment of a set of samples with or without APMA to activate latent enzymes prior to Western blotting (data not shown).

https://doi.org/10.1371/journal.pone.0115524.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. Blocking of C×43 function by Gap27 promotes secretion of VEGF-A, and suppresses DCN levels in gingival fibroblast cultures.

Confluent cultures of gingival fibroblasts (GFBL-DC) were treated with Gap27 or control peptide (150 μM) for 24 h, and Vascular Endothelial Growth Factor-A (VEGF-A) and Decorin (DCN) levels were analyzed in the conditioned medium by Western Blotting. Representative results from three independent experiments are shown.

https://doi.org/10.1371/journal.pone.0115524.g006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 7. Gap27 treatment increases C×43 protein abundance significantly.

(A) Confluent cultures of gingival fibroblasts (GFBL-DC) were treated with Gap27 or control peptide (150 μM) for 24 h, and abundance of C×43 was analyzed by Western blotting. Gap27 treatment did not affect the relative intensities of three bands corresponding to differently phosphorylated forms of C×43 (P0: pS368; P1:pS279/282 and pS255; P2: pS262). (B) Quantitation of C×43 levels in Western blots shows mean +/− SEM from three independent experiments (**p<0.01; Student’s t-test). Sample loading was normalized for β-Tubulin levels.

https://doi.org/10.1371/journal.pone.0115524.g007

Blocking of C×43 Function by Gap27 Modulates Key Signaling Pathways in Gingival Fibroblasts

Having established that during gingival wound healing, C×43 abundance and/or GJ and hemichannel functions maybe strongly reduced, and that suppressing C×43 function in cultured fibroblasts distinctly regulates expression of a set of wound healing-associated genes, we wanted to find out which intracellular signaling pathways are involved. To this end, we treated GFBLs with Gap27 and assessed phosphorylation changes in TGF-β (SMAD3), MAPK (ERK1/2 and p38), GSK3α/β and β-Catenin pathways that have been previously associated with C×43-mediated signaling [6,67–73], and wound healing and fibrosis [74–78] (Fig. 8). Gap27 induced robust phosphorylation of p38 and ERK1/2 already after 1 h after treatment. These responses lasted at least for 6 h, before returning to the levels of untreated cells by 24 h (Fig. 8C–F). Gap27 treatment also markedly increased GSK3α/β phosphorylation at 1–6 h, returning to the normal level by 24 h (Fig. 8G and H), while phosphorylated and non-phosphorylated β-Catenin levels, downstream targets of GSK3α/β, did not show marked changes (Fig. 8I and J). Gap27 treatment did not noticeably affect SMAD3 phosphorylation until 24 h, when the steady-state p-SMAD3 levels were markedly increased (Fig. 8A and B). Thus, in gingival fibroblasts, blocking of C×43 by Gap27 treatment induced fast activation of MAPK and GSK3α/β signaling pathways, while TGF-β pathway was activated via a slower, possibly indirect, mechanism.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 8. Western blotting analysis of key signaling pathways modulated by Gap27 in gingival fibroblasts.

Confluent cultures of gingival fibroblasts (GFBL-DC) were treated with Gap27 or control peptide (150 μM) for 1, 2, 6, and 24 h. Cell lysates were analyzed for protein levels of total SMAD3 and phosphorylated SMAD3 (p-SMAD3) (A), total p38 and phosphorylated p38 (p-p38) (C), total ERK1/2 and phosphorylated ERK1/2 (p-ERK1/2) (E), total GSK3α/β and phosphorylated GSK3α/β (p-GSK3α/β) (G), and total β-Catenin, phosphorylated β-Catenin (p-β-Catenin) and non-p-β-Catenin (I). (B, D, F, H and J) Quantitation of the phosphorylated or non-phosphorylated signaling molecules relative to their total levels at time 0 (control samples), and at 1, 2, 6 and 24 h after Gap27 treatment. Sample loading was normalized for β-Tubulin levels. Results from one experiment are shown.

https://doi.org/10.1371/journal.pone.0115524.g008

Distinct Involvement of Gap27-Regulated Signaling Pathways in Modulation of Wound Healing-Associated Genes in Gingival Fibroblasts

In order to further assess the role of the above pathways in Gap27-induced gene expression changes, we blocked TGF-β signaling by SB431542, MEK1/2 by PD184352, p38 by SB203580, and GSK3α/β by SB415286 in Gap27-treated cells, and assessed gene expression changes relative to untreated cells by real-time PCR. In addition, we assessed involvement of AP1 and SP1, transcription factors previously linked to C×43 signaling [73,79,80], by treating cells with curcumin and WP631, respectively. We specifically focused on assessing expression of the genes that were significantly modulated (with a minimum by +/− 1.5-fold change threshold) by Gap27 in the above experiments. Results showed that MEK1/2 signaling was most widely involved in Gap27-induced change in gene expression, as its inhibition totally blocked Gap27-induced expression change of 10 genes. These genes included proteases and their inhibitors (MMP-1, -3 and -10, TIMP-1 and -3), ECM molecules (Collagen type I and Tenascin-C), cell contractility-associated genes (α-SMA and NMMIIB), and Cadherin-2 (Table 5 and S9 Fig.). In addition, this pathway partially regulated (inhibited Gap27-induced expression change by at least 50%) expression of growth factors (TGF-β1 and VEGF-A) and C×43. The only genes that were not affected by blocking of MEK1/2 were MMP-14, two leucine-rich proteoglycans (Decorin and Fibromodulin), and CXCL12. The other examined pathways also totally or partially regulated Gap27-modulated expression of several genes. For instance, inhibition of GSK3α/β resulted to total inhibition of Gap27-induced upregulation of TIMP-1 and -3 expression, while it partially blocked MMP-1 and TGF-β1 upregulation. Findings also showed variable and complex interplay between the studied pathways in regulation of Gap27-induced gene expression. For instance, Gap27-induced MMP-10 upregulation and Collagen type I down regulation was only blocked by the MEK1/2 inhibitor. In contrast, Gap27-induced TIMP-1 expression was totally blocked by SP1, TGF-β, MEK1/2 and GSK3α/β inhibitors, and partially by p38 inhibitor. On the other hand, Gap27-induced Tenascin-C expression was totally inhibited by AP1 and MEK1/2 inhibitors, and partially by SP1, TGF-β and p38 inhibitors. Transcription factors AP1 or SP1 were distinctly involved in regulating 7 of the Gap27-modulated genes. Interestingly, inhibition of SP1 totally blocked Gap27 induced upregulation of C×43 expression, while blocking of AP1, p38 and MEK1/2 had a partial effect. Gap27-induced upregulation of VEGF-A and down regulation of CXCL12 were only partially blocked by MEK1/2 and SP1 inhibitors, respectively, suggesting that also other unidentified pathways are involved. In addition to the above effects, interestingly, blocking of GSK3α/β strongly potentiated Gap27-induced upregulation of MMP-3 (S9 Fig.) and Tenascin-C expression (S9 Fig.). Moreover, Gap27 treatment with simultaneous inhibition of MEK1/2 resulted to strongly induced α-SMA expression (by more than 7-fold), while treatment with Gap27 alone reduced its expression by about 50% (S9 Fig.).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Blocking of C×43 function with Gap27 treatment activates distinct signaling pathways that regulate wound healing-associates genes in gingival fibroblasts.

https://doi.org/10.1371/journal.pone.0115524.t005