Function of Chemokine (CXC Motif) Ligand 12 in Periodontal Ligament Fibroblasts

The periodontal ligament (PDL) is one of the connective tissues located between the tooth and bone. It is characterized by rapid turnover. Periodontal ligament fibroblasts (PDLFs) play major roles in the rapid turnover of the PDL. Microarray analysis of human PDLFs (HPDLFs) and human dermal fibroblasts (HDFs) demonstrated markedly high expression of chemokine (CXC motif) ligand 12 (CXCL12) in the HPDLFs. CXCL12 plays an important role in the migration of mesenchymal stem cells (MSCs). The function of CXCL12 in the periodontal ligament was investigated in HPDLFs. Expression of CXCL12 in HPDLFs and HDFs was examined by RT-PCR, qRT-PCR and ELISA. Chemotactic ability of CXCL12 was evaluated in both PDLFs and HDFs by migration assay of MSCs. CXCL12 was also immunohistochemically examined in the PDL in vivo. Expression of CXCL12 in the HPDLFs was much higher than that in HDFs in vitro. Migration assay demonstrated that the number of migrated MSCs by HPDLFs was significantly higher than that by HDFs. In addition, the migrated MSCs also expressed CXCL12 and several genes that are familiar to fibroblasts. CXCL12 was immunohistochemically localized in the fibroblasts in the PDL of rat molars. The results suggest that PDLFs synthesize and secrete CXCL12 protein and that CXCL12 induces migration of MSCs in the PDL in order to maintain rapid turnover of the PDL.


Introduction
The periodontal ligament (PDL) is one of the connective tissues located between the tooth and bone. It has shock-absorbing properties against mechanical stress [1], and thus, it prevents damage to the tooth and alveolar bone during mastication [2,3]. In addition, the PDL enables teeth to move via periodontal regeneration during orthodontic treatment [4][5][6][7].
The PDL is composed of heterogeneous cell populations; fibroblasts, osteoblasts, cementoblasts, osteoclasts, mesenchymal cells, mast cells and phagocytes [8]. Among these, fibroblasts are predominant. PDL is characterized by rapid renewal and repair, high remodeling capacity [9][10][11], and a remarkable capacity for renewal and repair when compared with other connective tissues, such as subcutaneous tissue [12]. PDL fibroblasts (PDLFs) are responsible for these characteristics of the PDL.
The supply of fibroblasts in PDL with these characteristics is controversial [13][14][15]. PDLFs are probably a source of osteoblasts and cementoblasts for remodeling of alveolar bone and cementum [16][17][18]. In addition, PDLFs are suitable cell sources of induced pluripotent stem cells [19]. These reports indicate that PDLFs are multipotent and may be capable of self-replication. On the other hand, mesenchymal progenitor cells that differentiate into fibroblasts are also present in the PDL [20][21][22].
Our preliminary results indicated that expression of chemokine (CXC motif) ligand 12 (CXCL12) in HPDLFs was much higher than that in human dermal fibroblasts (HDFs). The function of CXCL12 is to induce migration of mesenchymal stem cells (MSCs) [23][24][25]. Therefore, the objective of this study was to investigate the function of CXCL12 in the PDL with rapid turnover.

Ethics Statement
All the experiments were conducted in accordance with the guidelines of the National Institutes of Health, and the Ministry of Education, Culture, Sports, Research or publication ethics, Science and Technology of Japan, and were approved by the Animal Research Committee of Tsurumi University, Kanagawa, Japan. We made every effort to minimize the number of animals used and their suffering.

RNA Extraction and cDNA Synthesis
Total RNAs were extracted from the cell lines described above using Isogen (Nippon Gene Co., Ltd., Toyama, Japan), and 1 mg of total RNA was reverse transcribed to complementary DNA (cDNA) using a PrimeScript II 1st strand cDNA synthesis kit (Takara Bio Inc.) with random hexamers.

Microarray Analysis
Microarray analysis was performed using a Whole Human Genome 4644K (Agilent Technologies, Tokyo, Japan) containing approximately 44,000 transcripts. According to the manufacturer's protocol, total RNAs from HPDLFs and HDFs were labeled with Cy3, respectively, mixed, and hybridized on the microarray. Hybridization data for HPDLFs were compared with data for HDFs.

RT-PCR
RT-PCR analysis was performed with a Sapphire-Amp TM Fast PCR Master Mix (Takara Bio Inc.) using cDNA derived from 100 ng of total RNA in a 25-ml reaction volume. All reactions were carried out in triplicate. PCR conditions were 94uC for 1 min, followed by 25-40 cycles of denaturation at 98uC for 5 s, annealing of specific primers for 5 s, and extension at 72uC for 5 s. Primer sequences were obtained from published DNA sequences of human genes expressed in periodontal connective tissue or osteogenesis markers [26][27][28][29][30][31]. Primer sequences and annealing temperatures for each reaction are shown in Table 1.

Quantification of CXCL12
qRT-PCR analysis was performed with SYBR Premix Ex Taq II Perfect Real Time reagent (Takara Bio Inc.) in a Thermal Cycler Dice Real Time System (Takara Bio Inc.) using a specific primer pair of each cDNA according to the published sequences for COL1A1 [28,32] and COL3A1 [29,32]. PCR conditions were 95uC for 30 s, 40 cycles of denaturation at 95uC for 5 s and   annealing and extension at 60uC for 30 s, followed by dissociation and preparation of a standard denaturation curve. The standard curve for each gene was plotted showing Ct versus the logarithmic value of diluted concentrations of the target cDNA standard. The amount of endogenous mRNA of CXCL12 and GAPDH in these cells was quantified by assumption from the standard curve for each gene, as in our previous study [27]. The mean values and standard deviations were calculated from three independent experiments. Analysis was also performed on other genes.

Quantification of CXCL12 Protein
Culture medium incubated for 24 h was collected using a pipette, and was centrifuged 1000 rpm at 4uC for 10 min. Supernatants were used for protein analysis. Quantification of CXCL12 in culture supernatants of HPDLFs and HDFs was performed using a commercially available Quantkine ELISA kit (R&D Systems, Minneapolis, MN). CXCL12 concentrations were calculated with reference to a standard curve. All samples were analyzed in triplicate.

siRNA Transfection
Small interfering RNA targeting CXCL12 (CXCL12-siRNA) and scramble siRNA as a negative control were transfected into cells with jetPRIME (Polyplus, Illkirch, France) according to the manufacturer's recommendations. Briefly, HPDLFs were seeded in 24-well plates at a density of 1.5610 5 cells per well in 2 ml of Opti-MEM I Reduced-Serum Medium (Life Technologies). HPDLFs were transfected with scrambled siRNA (negative control) or CXCL12-siRNA diluted in transfection buffer at a final concentration of 50 nM. Fifty microliters of transfection medium was added to 24-well plates. After 24 h, transfection medium was replaced with culture medium, and the following day, transfected HPDLFs were used for migration assay and all knockdown studies. CM was also used in the migration assay.

In Vitro Transwell Chamber Migration Assay
A Falcon cell culture insert system and a companion 24-well Falcon tissue culture plate (Becton Dickinson, Franklin Lakes, NJ) were used for migration assay, which was performed according to the manufacturer's instructions. Migration assay was performed using MSCs from three donors. Cell cytoplasm of MSCs was stained with DiI Fluorescent Dye (Becton Dickinson) for 30 min at 37uC in the dark, and then 1610 5 cells in 200 ml of PLUSOID-M were seeded on the upper surface of the filter membrane in the upper wells. The lower wells were seeded with 1.5610 5 HPDLFs from three donors to confirm reactivity to HPDLFs. Chambers were incubated for 24 h at 37uC. Nuclei of migrated cells on the lower surfaces were stained with DAPI and were observed by fluorescence microscopy. The number of migrating cells in each microscopic field (6100) was also counted. There were no significant differences in the number of migrated cells among MSCs from three donors. Therefore, MSCs from one donor were used in subsequent experiments. Lower wells were also seeded with 1.5610 5 CXCL12-siRNA transfected HPDLFs, scramble siRNA transfected HPDLFs or HDFs. Human recombinant CXCL12 (250 ng/ml; PeproTech, Rocky Hill, NJ) in 500 ml of PLUSOID-M was used as a positive control, and was added to the medium in the lower well. Non-migrating cells on the upper surfaces of the filter membrane were removed with a swab. Nuclei of migrated cells on the lower surfaces were stained with DAPI and observed by fluorescence microscopy. The number of migrating cells in each microscopic field (6100) was also counted. Medium without cells served as a negative control. Experiments were performed in triplicate. For blocking of CXCR4, CXCR4 neutralizing antibody (100 mg/ml; Abcam, Cambridge, UK) or IgG antibody (negative control, 100mg/ml; Abcam) were added to upper wells [33]. In addition, to block CXCR4 receptors, MSCs were first incubated in 1.0 mg/ml AMD3100 (Sigma, St. Louis, MO) for 30 min at room temperature, washed in PBS and then resuspended in PLUSOID-M [34].

Exposure of MSCs to HPDLFs CM
HPDLFs, HDFs, CXCL12-downregulated HPDLFs and scramble siRNA transfected HPDLFs were incubated in a 6-well plate (2610 5 cells/ml) with PLUSOID-M culture medium. After 24 h, culture medium was collected using a pipette, and was centrifuged 1000 rpm at 4uC for 10 min. Supernatants from these cells were used for CM. MSCs (2610 5 cells/ml) were exposed to CM under the previously described conditions, and medium only was used for 24 h as a negative control. MSCs were analyzed by quantitative RT-PCR.

Immunohistochemistry
After perfusion with 4.0% paraformaldehyde (0.1 M PBS buffer, pH 7.4, room temperature) for 5 min, upper first molars including its periodontal ligament (PDL) were excised from the maxilla of rats. Specimens were decalcified with 0.3 M EDTA-Na2 (7% sucrose, 4uC) for 2 weeks, and were washed overnight in 0.1 M PBS buffer. After dehydration in graded ethanol, specimens were embedded in paraffin, and were cut serially with a microtome (sliding type, Sakura Seiki, Tokyo) at 6-7 mm thickness. Some deparaffinized sections were stained with HE. Other sections were incubated in a 1:100 dilution of anti-CXCL12 (host: rabbit, polyclonal type, SC-109854; Santa Cruz Biotech- nology, Santa Cruz, CA) as primary antibody in a conventional procedures, followed incubation in peroxidase-labeled anti-rabbit IgG (MP-7500; ImmPRESS Reagent, Vector Lab., Inc., Peterborough, UK) as secondary antibody. After incubation, sections were immunostained with DAB and counterstained with hematoxylin. The sections incubated without primary antibody were used as a negative control.

Statistical Analysis
All data are expressed as means and standard deviation from three independent experiments. Differences among independent groups were analyzed by one-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparison using statistical software (SPSS Statistics ver. 19.0; IBM Japan, Tokyo, Japan). Pvalues of less than 0.05 were considered to be statistically significant.

Microarray Analysis of HPDLFs and HDFs
Microarray analysis demonstrated that 3,807 genes in HPDLFs were expressed at levels more than 2-fold higher than in HDFs. Array data was deposited at the National Center for Biotechnology Information (NCBI) under the accession number GSE52162. The most remarkable result was observed in the expression of CXCL12 (209-fold), which is a migration-associated cytokine ( Table 2). The expression of COL1A1 (26.1-fold) and COL3A1 (41.8-fold) in HPDLFs was higher than in HDFs. Expression of specific markers of connective tissues such as BST1 (15.3-fold), GDF5 (10.3-fold), Gremlin (8.8-fold) and Follistatin (FST) (6.6-fold) were also higher in HPDLFs than in HDFs.

Confirmation of Gene Expression Levels by RT-PCR in HPDLFs, HDFs and HeLa Cell Lines
In order to confirm the results of microarray analysis, RT-PCR was performed for CXCL12 in HPDLFs, HDFs and HeLa cells (Fig. 1A). CXCL12 in HPDLFs showed distinctly higher expression when compared with HDFs and HeLa cells. Expression of COL3A1 and COL1A1 in HPDLFs was higher than in HDFs. On the other hand, Gremlin, BST1, FST and GDF5 were expressed in both HPDLFs and HDFs, but no clear differences in expression were observed between the two types of fibroblast. In HeLa cells as a negative control, only BST1 and FST were detected, which is consistent with our previous report [19].

Examination of Gene Expression of CXCL12 in HPDLFs and HDFs from three Donors for Each Population by qRT-PCR
Expression of CXCL12 in HPDLFs derived from three donors was examined by qRT-PCR in order to confirm common characteristics in HPDLFs. This confirmation was also performed in HDFs from three donors (Fig. 1B). There were no statistically significant differences in the expression of CXCL12 among the fibroblasts from the three donors of HPDLFs (Fig. 1Bi) or among fibroblasts from the three donors of HDFs (Fig. 1Bii).
The effects of CXCL-siRNA on HPDLFs and HDFs were also examined three donors for each population (Fig. 1C). The results showed significant down-regulation in the expression of CXCL12 in HPDLFs from three donors (Fig. 1Ci). But the expression of CXCL12 in HDFs was not influenced by CXCL12-si-RNA in three donors (Fig. 1Cii). Therefore, HPDLFs from one donor of HPDLFs (HPDLFs1) and HDFs from one donor were used in subsequent analysis.

Confirmation of Gene Expression and Protein Expression Levels of CXCL12 in HPDLFs and HDFs by qRT-PCR and ELISA for CXCL12
qRT-PCR demonstrated that expression of CXCL12 in HPDLFs was significantly higher than in HDFs ( Fig. 2A). Expression of CXCL12 in HDFs was as low as in HeLa cells. The expression of CXCL12 protein in the conditioned medium (CM) was examined by ELISA (Fig. 2B). CXCL12 in CM from HPDLFs was significantly higher than that in CM from HDFs and HeLa cells. These results are consistent with the results of RT-PCR. The effects of CXCL12-siRNA caused a marked reduction in the expression of CXCL12 in both mRNA and protein levels. On the other hand, HPDLFs with transfection of scrambled siRNA did not show significant differences in expression of CXCL12 when compared with parental HPDLFs. The effects of CXCL12-siRNA transfection on mRNA levels were also examined from 0 to 72 h. The effects continued until 72 h.

Migration Assay of MSCs by HPDLFs
Prior to comparison among the results, MSCs from three different donors were examined in this assay to confirm the reactivity to HPDLFs (Fig. 3Ai). There were no statistically significant differences in the number of migrated cells among MSCs from three donors (Fig. 3Aii). MSCs from one donor were used in subsequent assays. The chemotactic activity of CXCL12 in HPDLFs, HDFs and CXCL12-siRNA-transfected HPDLFs was examined by transwell chamber migration assay. Fluorescent microscopy clearly demonstrated the results of the assay (Figs. 3Bai). MSCs passing through the filter membrane were observed on the lower surface of the membrane. Migrated MSCs by HPDLFs were much more abundant (Fig. 3Ba) on the lower surface of the filter membrane than those by HDFs (Fig. 3Bb) and medium only (Fig. 3Bc). The results of migration using inhibitor of CXCR4  (Fig. 3Bd) and CXCR4 antibody (Fig. 3Be) showed similar migration as with the results of medium alone. This was also observed in the migration of MSCs, using CXCL12-siRNA (Fig. 3Bg). On the other hand, migrated MSCs by scramble siRNA-transfected HPDLFs (Fig. 3Bh), IgG antibody (Fig. 3Bf) and recombinant CXCL12 (Fig. 3Bi) showed the similar results as HPDLFs.
The number of migrated MSCs was also counted at 24 h after initiation of the assay (Fig. 3C). The number of migrated MSCs by HPDLFs was significantly higher than that by HDFs. Recombinant CXCL12 also induced similar migration as with HPDLFs. The number of migrating cells on by medium alone, CXCR4 inhibitor and CXCR4 antibody was significantly lower than that by HPDLFs.

Examination of Gene Expression in Migrated MSCs by RT-PCR and qRT-PCR
The influence of CM from HPDLFs on MSCs was examined by RT-PCR (Fig. 4A). Results were compared to those with MSCs cultured in normal culture medium. Expression of CXCL12 was up-regulated in the MSCs exposed to CM from HPDLFs at 24 h after incubation. COL1A1 and Alkaline phosphatase were upregulated in MSCs. Up-regulation was also observed in medium containing recombinant CXCL12. These results were supported by qRT-PCR (Figs. 4Ba-d). CXCL12 and COL1A1 expression was also significantly up-regulated in MSCs. Up-regulation was not observed in the expression of COL3A1 and PLAP1. MSCs exposed to CM from CXCL12-siRNA HPDLFs or MSCs alone did not show up-regulation in expression of CXCL12 and COL1A1.

Immunohistochemistry for CXCL12 in PDL in vivo
Periodontal tissue consists of PDL, alveolar bone, cementum and gingiva. Fibroblasts were observed throughout the PDL and scattered along the periodontal fibers (Fig. 5A,B). In gingiva, fibroblasts were scattered and some capillaries were present in the gingival connective tissue (Fig. 5C). CXCL12 was predominantly localized in the fibroblasts in the PDL (Fig. 5 D, E). However, CXCL12 was not observed in most of the fibroblasts in the gingival connective tissue (Fig. 5D, F), and was only detected in cells around the capillaries.

Discussion
HPDLFs and HDFs used in this study were commercially available cell lines. It is important to confirm the characteristics of HPDLFs and HDFs. Both fibroblasts showed a spindle-like shape when grown in culture dishes. These fibroblasts clearly showed expression of several connective tissue-specific markers on RT-PCR [27][28][29][30][31][32], which were different from those in the human epithelial cell line Hela (negative control), except for BST1 and follistatin. The HPDLFs and HDFs used in this study had the characteristics of connective tissue fibroblasts. In this study, three cell lines of HPDL fibroblasts, and HD fibroblasts were used to examine the differences in expression of CXCL12, and consequently, no differences were observed. This indicates that CXCL12 is commonly expressed in HPDLFs, and that its expression is significantly higher than in HDFs. In addition, MSCs from three donors were also examined to confirm the reactivity to the chemotactic activity of HPDLFs, and there were no significant differences in the number of migrated cells among the three donors.
RT-PCR analysis showed that both HPDLFs and HDFs expressed COL1A1 and COL3A1, and that HeLa cells did not express these genes. This indicates that both HPDLFs and HDFs in this study have the characteristics of fibroblasts. However, it appears that the expression of COL1A1 and COL3A1 in HPDLFs was much stronger than in HDFs. In general, the PDL shows rapid turnover and PDLFs play distinct functional roles in the regeneration and repair of the PDL [35,36] and the collagen synthesis is one of the primary functions of PDLFs [37,38], HPDLFs have fundamentally high activity in the formation of collagen fibers [39,40].
Microarray analysis of the whole genome revealed distinctly high expression of CXCL12 in HPDLFs when compared to that in HDFs. qRT-PCR demonstrated significant upregulation of CXCL12 in HPDLFs. These results indicate that CXCL12 is upregulated not only in the cells under pathological conditions such as inflammation and tumors [41,42], but also in normal HPDLFs. The high expression of CXCL12 in HPDLFs was also confirmed at the protein level by ELISA. Therefore, HPDLFs synthesize CXCL12 protein and secrete it extracellularly. In addition, the immunohistochemical findings showed that CXCL12 is located in fibroblasts in the PDL.
The function of CXCL12 in HPDLFs was investigated by migration assay. CXCL12 is involved in cell migration [33,[43][44][45]. Our migration assay demonstrated that HPDLFs induced migration of a significantly higher number of MSCs than did HDFs. HPDLFs in which CXCL12 was downregulated by siRNA induced low migration, which was equivalent to the migration induced by HDFs. Furthermore, recombinant CXCL12 induced migration of MSCs, which is similar to that of HPDLFs. These results precisely define the function of CXCL12 in HPDLFs. CXCL12 plays a critical role in the supply of fibroblasts that are responsible for rapid turnover. This migration may result from the increased mobility of plasma membranes in response to CXCL12, which induces cytoskeletal arrangement, pseudopodia formation [46] and consequently MSC pass through the filter membrane. MSCs usually circulate in the blood vessels [47][48][49] and PDL is rich in blood vessels and shows rapid turnover, when compared with other connective tissues [50,51]. Thus, we postulated that CXCL12 secreted from PDLFs induces migration of MSCs from the blood to the site of rapid renewal and repair. Furthermore, CXCL12 may be a specific marker that distinguishes PDLFs from other fibroblasts.
On the other hand, the treatment with CXCL12 siRNA did not affect the expression of CXCL12 in HDFs population. This might be because the expression level of CXCL12 is intrinsically low in HDFs, which is closely related to low turnover in dermal connective tissue, as compared with that in PDL.
This study also noted an interesting finding regarding MSCs. MSCs incubated in CM from HPDLFs showed some changes in gene expression patterns. The CM from HPDLFs induced upregulation of CXCL12, COL1A1 and ALP in MSCs. However, MSCs incubated in CM from HPDLFs in which CXCL12 was down-regulated by CXCL12-siRNA, did not show the same changes and instead, showed similar results as control MSCs. The up-regulation of CXCL12 in MSCs may induce new migration of MSCs via synthesis and secretion of CXCL12 during migration, which also contributes to the rapid renewal and repair of the PDL. Furthermore, the up-regulation of COL1A1 and ALP in MSCs suggests the possible differentiation of MSCs into fibroblast-like cells during migration. This might be due to the activation of MAP kinase by CXCL12 [52,53], which induces up-regulation of COL1A1 [54][55][56] and ALP [57,58]. In addition, COL1A1 and ALP are requisite factors of PDL fibroblasts [59]. Collagen type I is a primary component of periodontal fibers and ALP is located in collagen fibers and fibrils, which are different characters from other connective tissues. Therefore, COL1A1 and ALP might be given priority over the expression of other genes. In summary, the results in this study suggest that PDLFs synthesize and secrete CXCL12. CXCL12 from PDLFs induces migration of MSCs to supply new fibroblasts in the PDL with rapid turnover.

Author Contributions
Conceived and designed the experiments: Y. Nakamura. Performed the experiments: YY. Analyzed the data: Y. Nomura. Contributed reagents/ materials/analysis tools: MK KN NH. Wrote the paper: YY.