Cementum protein 1 (CEMP1) activates p38 and JNK during the mineralisation process by cementoblast-like cells in vitro

Summary

Abstract:
We recently presented evidence showing that cementum protein 1 (CEMP1) promotes periodontal ligament (PDL) cell migration, proliferation, expression of bone, and cementum-matrix proteins and mineralisation. In other words, it induces PDL precursor cells commitment toward a cementoblast-like cells phenotype. The intracellular signalling pathways involved in cementoblast differentiation and mineralisation have not been well characterised. JNK and p38 protein kinases (MAPKs) are intracellular signalling pathways and key mediators of cellular processes such as proliferation and differentiation. Since signalling pathways involving MAPKs have been associated with osteoblastic phenotype, in this study we investigated the effect of hrCEMP1 and mineralising media containing β-glycerophosphate and ascorbic acid on the activation of p38-MAPK and JNK–MAPK in cementoblast-like cells. Our results show that mineralising media and hrCEMP1 induced phosphorylation of p38 and JNK kinases. Mineralising media containing hrCEMP1 increased the activation of p38-MAPK and its translocation to the cell nucleus; increased phosphorylation of JNK–MAPK and induced the phosphorylation of the protein C-JUN. We also demonstrate that hrCEMP1 regulates the expression of BSP, OCN, and ALP specific activity. We found that hrCEMP1 and mineralising media promote nodule formation. These findings give an insight into the signalling pathways activated by hrCEMP1 and suggest likely components of the mechanisms that regulate the formation and regeneration of cementum and surrounding connective tissues.

Keywords:
cell signalling pathways; cementoblasts; CEMP; MAPK; mineralisation

Introduction

Cementum is a unique avascular mineralised tissue that covers the roots of the teeth and provides the interface through which the root is anchored to alveolar bone via collagen Sharpey's fibers of the periodontal ligament. The cementum organic matrix consists of several proteins and glycoproteins like collagens types I and III, fibronectin, osteopontin (OPN), bone sialoprotein (BSP), osteocalcin (OCN), alkaline phosphatase (ALP), these proteins play an important role during the cementogenesis process particularly during the mineralisation events, chondroitin sulphate, dermatan sulphate, hyaluronic acid, and several growth factors like BMPs, FGFs, etc. (Narayanan et al., 1995; Cho and Garant, 2000; Shimono et al., 2003; Nanci and Bosshardt, 2006). More recently, a new protein present in cementum, named Cementum Protein 1 (CEMP1; GenBank NM_001048212, gene ID 752014) has been characterised and it has been shown that this protein is specific for cementum and it is expressed by cementoblasts and cementoblast-progenitor cells present in the periodontal ligament (Alvarez et al., 2006). Functional studies have shown that human recombinant cementum protein 1 (hrCEMP1) induces differentiation of non-mineralising cells to a mineralised-like phenotype (Carmona et al., 2007; Kémoun et al., 2007) and controls periodontal ligament cells commitment to become cementoblast. Additionally, CEMP1 is preferentially expressed in ALP-positive periodontal ligament cells, possible precursors of cementoblasts and knockdown of CEMP1 indicates that BSP expression in cementoblasts is regulated by CEMP1 and OCN is expressed selectively by cells lining tooth root surface cementoblasts (Komaki et al., 2012). Furthermore, hrCEMP1 plays a role during the biomineralisation process by promoting octacalcium phosphate (OCP) crystal nucleation (Villarreal et al., 2009). Although it is well accepted that cell differentiation is dependent upon regulation of multiple interacting signalling pathways, the mechanisms regulating cementoblast cell proliferation and differentiation, and in particular the mechanism by which hrCEMP1 plays a role during the mineralisation process by cementoblasts, are not well known.

The mitogen-activated protein kinases (MAPKs) are the family of secondary messengers that transmit signals from the cell surface to the nucleus (Yang et al., 2013). MAPKs are extracellular signal-regulated kinase (ERK), p38 kinase, and c-JUN terminal kinase (JNK). MAPKs are activated via phosphorylation of tyrosine and threonine residues. Activated MAPKs subsequently phosphorylate their specific substrates at serine and/or threonine residues in order to regulate substrate and thus the entire signalling cascade activity (Yang et al., 2013). The mitogen-activated protein kinases (MAPKs), p38 and JNK play a pivotal role in many essential cellular processes such as proliferation, differentiation and apoptosis (Johnson and Lapadat, 2002; Krens et al., 2006; Aouadi et al., 2006; Raman et al., 2007; Paula-Silva et al., 2010). In vitro studies have shown that p38-MAPK and JNK–MAPK act as signalling pathways in osteoblastic cell differentiation (Guicheux et al., 2003; Caverzasio and Manen, 2007; Greenblatt et al., 2010). Molecules such as estrogen, platelet-derived growth factor (PDGF) and fibroblast growth factor (FGF), which act as bone-active agents, have been shown to induce ERK signalling in osteoblasts (Fulzele et al., 2007; Chau et al., 2009). Therefore, in the present study, we wanted to determine if hrCEMP1 triggered the activation of p38-MAPK and/or JNK–MAPK pathways and regulates the expression of BSP, OCN and ALP since these proteins may play a key role during the mineralisation process in human cementoblastoma-derived cells (HCDC) in vitro.

Materials and methods

Production of human recombinant cementum protein 1

Production of hrCEMP1 has been previously described in detail (Alvarez et al., 2006; Hoz et al., 2012). Human recombinant CEMP1 protein was purified by Ni²⁺ affinity chromatography (Invitrogen, Carlsbad, USA). Purity was determined using a SDS/12% PAGE and Western blotting as described elsewhere (Alvarez et al., 2006).

Cell culture

Human cementoblastoma-derived cells (HCDC) used in these studies have been previously isolated, grown and characterised as putative cementoblastic cell line as described in detail elsewhere (Arzate et al., 1992, 2000, 2002, Alvarez et al., 2003). Cells between the 2nd and 5th passage were used for the experiment (Arzate et al., 2002). The cells were grown in medium supplemented with 10% fetal bovine serum (FBS) or in ‘mineralising media’ (1% FBS, 10 mM β-glycerophosphate and 50 µg/mL of freshly prepared ascorbic acid).

Western blotting and kinase assays

HCDC were plated at high density (2 × 10⁵) in 24-well plates and cultured for 3, 7 and 14 days. Cells were divided in four groups and treated as follows: (1) controls: DMEM + 1% FBS, (2) DMEM + 1% FBS + 10 mM β-glycerophosphate + 50 µg/mL ascorbic acid, (3) DMEM + 1% FBS + 10 µg/mL hrCEMP1 and (4) DMEM + 1% FBS + 10 mM β-glycerophosphate + 50 µg/mL ascorbic acid + 10 µg/mL hrCEMP1. After treatment, cell lysates were prepared and the total protein content was determined according to the Lowry method (Lowry et al., 1954). Equal amounts of total protein (10 µg) were electrophoretically separated using 12% SDS–PAGE and electroblotted into an Immobilon-P (PVDF) membrane (Millipore Corp., Bedford, MA, USA) and probed with properly diluted primary antibodies in PBS plus 2 mg/mL of BSA; rabbit anti-rat p38 (1:300), rabbit anti-human phospho-p38 (1:300), (Sigma Chemical Co., St. Louis, MO, USA), rabbit anti-human JNK (1:250), rabbit anti-mouse phospho-JNK (1:300), (Millipore Corp.), rabbit anti-human BSP (1:500), (LF-100, courtesy of Dr. Larry W. Fisher, NIH, Bethesda, MD, USA), rabbit anti-human hrCEMP1 (produced in house, 1:300). Peroxidase-conjugated-goat anti-rabbit and goat anti-mouse IgG (Santa Cruz Biotech., Inc., CA, USA) were used and secondary antibody detection was performed as described elsewhere (Alvarez et al., 2006). The reference proteins (non-phosphorylated p-38, JNK and GAPDH) were used as internal controls. Blots were scanned and analysed with a Kodak Electrophoresis Documentation and Analysis System (EDAS) 290. The relative level of each protein was assessed by measuring the integrated intensity of all pixels in each band, excluding the local background. Results are expressed as percentages of protein intensity obtained in control cultures.

Alkaline phosphatase activity

HCDC were plated at 2 × 10⁴ in 24-well culture plates and cultured for 3, 7 and 14 days in the conditions described above. Alkaline phosphatase activity (ALP) was determined as described elsewhere (Lowry et al., 1954). The activity was expressed as nanomoles of p-nitrophenol/min/mg of protein. Protein content was determined using BSA as standard as described elsewhere (Bradford,1976).

Enzyme-linked immunoassay

Osteocalcin release by HCDC was directly measured in supernatants (control and experimental conditions as referred above) by ELISA according to the manufacturer's instructions (ALPCO Immunoassays, Salem, NH, USA). The plates were read at 490 nm using a Microplate Reader (Awareness Technology, Inc., USA). The limits of detection for these immunoassays were 8 pg/mL (Funaoka et al., 2010).

Immunostaining

HCDC were plated at 0.5 × 10³ onto glass coverslips and cultured in the conditions described above for 3, 7 and 14 days and then fixed with 2% paraformaldehyde (PFA) in PBS at 4°C for 30 min. Cells were immunostained with primary antibody rabbit anti-human phospho-p38 and subsequently incubated with a second antibody goat anti-rabbit FITC-conjugated antibody (Invitrogen). The slides were analysed using a confocal microscope (Eclipse 80i, Nikon Instruments Inc., USA).

Mineralisation assays

Mineralisation assays were performed in HCDC according to the different culture conditions described in detail in Western Blotting and Kinase Assays Section (Alvarez et al., 2003). The cultures were fixed as described above, and calcium nodules were detected using alizarin red as described elsewhere (Carmona et al., 2007). Residual stain was removed with PBS and the presence of nodules was analysed using light microscopy.

Statistical analysis

All experiments were performed in triplicate and repeated three times. Data were analysed using one-way analysis of variance (ANOVA) followed by Dunnett's or Tukey's test. Data are expressed as means ± SE from three independent experiments. P < 0.05 was considered statistically significant. Statistical analysis was performed with Sigma Stat V 3.1 software (Jandel Scientific Ashburn, VA).

Results

Effect of hrCEMP1 on the phosphorylation of MAPK p38 and JNK

In order to evaluate if hrCEMP1 had any effect on the activation of MAPK-p38 and MAPK–JNK, HCDC were incubated in DMEM plus hrCEMP1 (10 µg/mL) for 15 and 30 min; 1, 2, 3, 6 and 12 h. Our results show that hrCEMP1 increased phosphorylation of p38 by 17% and JNK by 9% after 15 min and it continued to increase slowly up to 12 h after treatment (Figure 1A). To evaluate the effect of hrCEMP1 on the activation of MAPK-p38 and MAPK–JNK under mineralisation conditions, HCDC were incubated with different concentrations of hrCEMP1 for 3 days, The results indicate a dose-dependent increase in phosphorylation of p38, reaching its maximum at 10 µg/mL of hrCEMP1 (49%) and decreasing with higher concentrations (Figure 1B). Similar patterns were observed for JNK, although its level of phosphorylation was lower than p38. These results indicate that indeed hrCEMP1 activates phosphorylation of p38-MAPK and JNK-MAPK.

Effect of hrCEMP1 on phosphorylation of p38, C-JUN, JNK, translocation of p38 to the nucleus and formation of mineralised nodules in HCDC

Once we determined the optimal concentration of hrCEMP1 to induce phosphorylation of p38 and JNK, we wanted to determine if there was a difference in the effect of hrCEMP1 in the presence of regular media or mineralising media in order to induce cell differentiation and mineralisation. Cells were grown for 3, 7 and 14 days, and proteins were extracted and analysed as described above. After 3 days in culture, phosphorylation of p38-MAPK increased in both normal and mineralising media plus with hrCEMP1 as compared to controls (Figure 2A). However, phosphorylation in mineralising media containing hrCEMP1 was considerably higher (40%) than normal media with hrCEMP1 (26%). The pattern of p38-MAPK phosphorylation was similar after 7 days in culture and there was no significant difference with any treatment after 14 days in culture (Figure 2A).

The effect of hrCEMP1 in mineralising and normal media was determined for JNK–MAPK. Figure 2B shows that after 3 days in culture, the phosphorylation pattern was similar to that found for p38-MAPK. Nevertheless, the level of phosphorylation is less for JNK (Figure 2B). The translocation of p38 to the nucleus was observed under the same culture conditions. A strong staining in HCDC treated with hrCEMP1 plus ‘mineralising media’ was observed (Figure 3A). The results showed that hrCEMP1 and ‘mineralising media’ acting alone or together triggered the phosphorylation of C-JUN (Figure 3B). Our results show that hrCEMP1 as well as the mineralising media promoted the formation of mineralised nodules as expected (Figure 3C).

Effect of hrCEMP1 on the expression of CEMP1, BSP, OCN and alkaline phosphatase activity

To evaluate the effect of hrCEMP1 on the expression of proteins associated with the mineralisation process, HCDC were incubated as previously described in the presence or absence of hrCEMP1 for 3, 7 and 14 days. The results shown indicate that BSP expression appears to be regulated in early stages of mineralisation and the addition of both hrCEMP1 and mineralising media decreases its expression by 54%. In longer periods of culture addition of hrCEMP1 appears to accelerate this decrease (Figure 4A). CEMP1 expression increased in all samples after 7 days and it disappeared after 14 days in culture, indicating that hrCEMP1 plays a role at the initial stages of mineralisation. The expression of OCN was determined using an ELISA assay and the results show that addition of mineralising media increased expression of OCN at initial stages of mineralisation and addition of hrCEMP1 decreased this effect (Figure 4B). ALP specific activity increased at early stages of mineralisation in the presence of mineralising media or hrCEMP1 and increased fivefold when they were combined as compared to controls. Nevertheless, in late stages of mineralisation, ALP activity slightly decreased as compared to controls when treated with hrCEMP1 (Figure 4C).

Discussion

The molecular mechanisms that regulate cementoblast cell differentiation and cementum deposition remain poorly understood. In the present study we describe in vitro conditions that lead to the activation of proteins involved in proliferation and differentiation events, which are mediated by CEMP1, and provide experimental evidence of the involvement of p38- and JNK–MAPKs in these cellular responses. Studies in vitro established that β-glycerophosphate and ascorbic acid are the most widely used supplements to induce osteogenic differentiation of various cell types (Temu et al., 2010; Khanna-Jain et al., 2010). It has also been shown that MAP kinases activation is involved in the differentiation of osteoblastic cells, and that p38-MAPK mediates the stimulation of ALP (Rey et al., 2007). This molecule is thought to play a role in phosphate metabolism and cementum formation (Beertsen et al., 1990). In this study, we determined that hrCEMP1 induces phosphorylation of p38-MAPK in cementoblast cells maintained in vitro. We also showed that ALP activity was increased in the presence of hrCEMP1 at early stages of the mineralisation process. Therefore, we suggest that p38-MAPK phosphorylation cell signalling is involved in mediating ALP stimulation by hrCEMP1 during the mineralisation process. This signalling pathway has been shown in the process of osteoblastic differentiation and mineralisation as well as cementoblastic differentiation (Guicheux et al., 2003; Aouadi et al., 2006; Caverzasio and Manen, 2007; Paula-Silva et al., 2010). It has also been reported that ascorbic acid is associated with ERK activation. However, attenuated ascorbic acid induces differentiation in chondrogenic and periodontal ligament cells (Temu et al., 2010; Khanna-Jain et al., 2010), although the mechanisms are not completely understood. The use of osteoinductors is related to the activation of the family of MAPK ERK1/2, p38 and JNK, signalling pathways (Kakita et al., 2004). Results presented in this study also indicate that HCDC grown in the presence of mineralising media showed activation of both p38- and JNK–MAPKs in a time-dependent manner, suggesting a role of these pathways in mediating mineralising media-induced cell proliferation and/or differentiation and/or cellular survival of HCDC. Here we demonstrated that both hrCEMP1 as well as mineralising media activated p38-MAPK and JNK–MAPK signalling pathways. Our studies showed that stimulation with mineralising media and/or hrCEMP1 promotes translocation of p38 to the nucleus and the activation of C-JUN, component of the transcription factor AP-1. These results suggest that hrCEMP1 and/or mineralising media could activate transcription factors such as AP-1 and ATF2 which are involved in the differentiation or proliferation process of cementoblastic cells. The translocation of these kinases to the cell nucleus suggests that the transcription factor-like AP-1, which is involved in cell proliferation and differentiation processes (Suzuki et al., 2002; Raman et al., 2007; Turjanski et al., 2007), might be activated during cementoblast differentiation induced by hrCEMP1. However, further experimental work is necessary to confirm this hypothesis.

In this study control media does not promote the activation of p38 and JNK. However, these proteins show phosphorylation with media supplemented with β-glycerophosphate and ascorbic acid. Also it is observed that phosphorylation is increased by the addition of hrCEMP1. These results show that b-glycerophosphate, ascorbic acid and hrCEMP1 are responsible for the activation of MAPK p38 and JNK.

We have previously demonstrated that CEMP1 induced phenotypic changes in non-mineralising cells. Human gingival fibroblasts overexpressing CEMP1 display a mineralising phenotype and express OCN, BSP, OPN and ALP (Carmona et al., 2007). Our results also showed that hrCEMP1 regulates expression of OCN in early and intermediate stages of mineralisation and decreases the levels of BSP at late stages of mineralisation.

It has been demonstrated that ALP is expressed by HCDC (Arzate et al., 1992; Alvarez et al., 2003) and the enzymatic activity of this protein is associated with the mineralisation process. In our studies, ALP activity was increased in the presence of hrCEMP1 in early and intermediate stages of cell-induced mineralisation. These findings suggest that hrCEMP1 plays an important role during the mineralisation process perhaps by the initial activation of ALP and induction of expression of mineralised-tissue associated proteins like BSP.

Our findings show that p38 is activated by mineralising media and hrCEMP1 when they act together; however hrCEMP1 might regulate the mineralisation of HCDC using more than one mechanism. The findings of these studies demonstrate that hrCEMP1 activates this signal pathway. However, the activation of intracellular kinases has a strict control of scaffolding proteins, and these would seem to have the answer for the regulation on the activation of the signalling pathways (Suzuki et al., 2002; Raman et al., 2007).

In conclusion, data obtained from this study indicate that p38 and JNK MAPKs are essential signalling pathways for CEMP1-induced cementoblastic differentiation. The stimulation of these pathways are important for the expression of ALP, BSP and OCN and therefore matrix mineralisation. Our results point to the key role for CEMP1-mediated signalling during cementoblast differentiation and matrix mineralisation and provide impetus to define the mechanisms that regulate cementoblast differentiation from uncommitted periodontal ligament stem cells toward the cementoblastic lineage and to determine whether p38 and JNK are essential signalling pathways in this respect. According to our results CEMP1 plays an important role during initial mineralisation since biologically activehrCEMP1 responsible for OCP crystal nucleation activity. OCP is a transient phase during the growth of biological crystals toward hydroxyapatite (Villarreal et al., 2009). Therefore the evidence obtained from this study, strongly indicates that CEMP1 is an important factor during the initial stages of cementoblast differentiation and initial stages of mineralisation. Understanding the role of hrCEMP1 in the signal pathway activated and mineralisation processes provides powerful tools that might allow to use this novel protein as a new therapeutic alternative in order to achieve regeneration of the periodontal structures and other mineralised tissues.

Acknowledgments and funding

This work was supported by DGAPA-UNAM, IN216711, IT200414, CONACyT: 130950.

Abbreviations

CEMP1 cementum protein 1

hrCEMP1 human recombinant cementum protein 1

HCDC human cementoblastoma-derived cells

BSP bone sialoprotein OCN osteocalcin

ALP alkaline phosphatase

MAPK mitogen activated protein kinases

JNK C-JUN N-terminal kinases

Received 26 September 2013;

Accepted 25 November 2013.

Final version published online 16 December 2013.

^* Corresponding author: e-mail: harzate@unam.mx