When triggered appropriately, dental follicle cells are considered to be able to differentiate toward a cementoblast/osteoblast phenotype. However, factors and mechanisms regulating follicle cell differentiation remain undefined. This study focused on determining the ability of bone morphogenetic protein (BMP) 2 to promote the differentiation of follicle cells and periodontal ligament (PDL) cells along a cementoblast/osteoblast pathway. Follicle cells and PDL cells were isolated from the first molar region of CD-1 mice and immortalized with SV40. Both cell types expressed BMP-4 and BMP receptors (BMPR) IA and II, but only follicle cells expressed BMP-2 mRNA. Cells were exposed to recombinant human BMP (rhBMP)-2 (0–100 ng/ml) and Northern blots were used to determine the expression of mineral-associated markers. BMP-2, in a dose- and time-dependent manner, induced cementoblast/osteoblast differentiation of follicle cells, as reflected by enhanced core binding factor α1 (Cbfa1), bone sialoprotein (BSP), and osteocalcin (OCN) mRNA expression and enhanced mineral formation. U0126, a specific inhibitor of MEK-1/2 members of the MAPK family, abolished BMP-2-mediated expression of BSP and OCN. In contrast, exposure of PDL cells to BMP-2 resulted in modest expression of OCN and minimal promotion of mineralization. These results suggest that BMP-2 triggers follicle cells to differentiate toward a cementoblast/osteoblast phenotype and that the MAPK pathway is involved.
PERIODONTAL DISEASES are characterized by destruction of periodontal support, that is, periodontal ligament (PDL), cementum, and bone, with subsequent tooth loss if left untreated. The recognition that regeneration of these tissues can be achieved has resulted in increased attempts to understand the cellular and molecular mechanisms and factors regulating formation of these tissues during development and regeneration.(1–3) Knowledge gained from such studies has led to clinical investigations targeted at designing regenerative therapies.(4–6) During the course of these studies, both at the clinical level and at the basic research level, bone morphogenetic proteins (BMPs) have emerged as attractive factors to use for promoting regeneration of periodontal tissues.(7) Many of the BMPs, which belong to the transforming growth factor (TGF) β superfamily, are found in high concentration in mineralized tissue and exhibit chondrogenetic/osteogenetic activity.(8,9) Mesenchymal cells exposed to BMP-2 exhibit increased expression of genes and proteins associated with the osteoblastic phenotype, including alkaline phosphatase, type I collagen (Col I), and osteocalcin (OCN).(10–12) Also, exposure of several cell lines, including PDL cells to BMP-2, in combination with factors that increase cyclic adenosine monophosphate (cAMP) levels, such as parathyroid hormone (PTH), results in increased levels of cAMP beyond those observed in cells exposed to such factors alone.(10–13) Data from animal studies indicate that recombinant human BMP (rhBMP)-2 can stimulate regeneration of periodontal tissues,(14–16) thus suggesting that rhBMP-2 may have clinical applications. To better understand the mechanism(s) by which BMPs promote regeneration of periodontal tissues, we have focused on determining the responsiveness of cells associated with the periodontium both developmentally and in the mature tissue, for example, dental follicle cells and PDL cells, respectively, to BMPs.
Cells within the dental follicle region, a loose connective tissue surrounding the developing tooth, play a critical role in the process of tooth eruption.(17–22) In addition, substantial evidence exists indicating that follicle cells are progenitors of periodontal cells including cementoblasts, PDL fibroblasts, and alveolar osteoblasts.(23–25) However, the specific events, mechanisms, and factors controlling follicle cell activities during tooth eruption and formation of the periodontium remain undefined. The ability of rhBMP-2 to promote osteogenic activity suggests that this protein is a promising candidate for stimulating follicle cell maturation along a “cementoblast/osteoblast” pathway. This study was devoted to determining the effect of rhBMP-2 on follicle cell differentiation. In addition, one of the putative descendants of follicle cells, PDL cells, considered to play a key role in maintenance of periodontal structure and function and also in regeneration of periodontal tissues, were included to compare putative progenitor cells with mature cells of the periodontium. As reported here, the results indicate that BMP-2 triggers follicle cells to differentiate toward a cementoblast/osteoblast phenotype including expressing markers associated with the mature phenotype, for example, bone sialoprotein (BSP) and OCN, and promoting mineral nodule formation. In PDL cells, BMP-2 induced modest expression of OCN and minimal promotion of mineralization. Analysis of the possible signaling pathways involved in this process suggests that the MAPK pathway is necessary for BMP-mediated effects on follicle cells.
MATERIALS AND METHODS
Two murine cell lines SV F4 (follicle cells) and SV11.03 (SV-PDL cells), were used. Methods for obtaining these cell populations have been published previously.(26–28) The method for isolation of follicle cells was a modification of that used by Wise's group.(17) Briefly, follicle cells were obtained from the first molars of CD-1 mice on days 22-24 of development, where vaginal plug date is designated as day 0 and mice usually are born on day 19. Selection of this time point for isolation of follicle cells was based on in situ data, indicating that at this time, root formation (cementogenesis) has not been initiated.(29,30) Using a dissecting microscope, the tooth germs of first mandibular molars were removed and then subjected to enzymatic digestion (0.6 mg/ml of collagenase type VII + 0.25% trypsin; Sigma, St. Louis, MO, USA) for 1 h. To immortalize cells after enzymatic digestion, cells were resuspended in 25 ml of DMEM (Gibco BRL, Gaithersburg, MD, USA) containing 2% FBS and SV40 at a multiplicity of infection of 100. After infecting for 2 h in suspension, cells were plated with an equal volume of media containing 20% FBS. Media were removed the next day and replaced with DMEM containing 10% FBS. To confirm infection with virus, expression of SV40 large TAg was determined by Northern analysis using a large TAg probe (obtained from Dr. Jolene Windle).(31) Immortalized follicle cells were designated SV F4 cells, and were maintained in DMEM plus 10% FBS containing 100 U/ml of penicillin and 100 μg/ml of streptomycin (Gibco BRL) in a humidified atmosphere of 5% CO2 at 37°C. Cells of passages 5-10 were used. Under these culture conditions, epithelial cells did not survive, as confirmed by lack of expression for keratins for all cells within the culture (DAKO-AE1/AE3 antibody; data not shown; Dako Corp., Carpintena, CA, USA). In certain experiments, primary murine dental follicle cells were analyzed to ensure that immortalization did not alter follicle cell function.
PDL cells were obtained from first mandibular molars of CD-1 mice on day 41 of mouse development(26) and then immortalized with SV40 as described previously. The SV11.03 subclone was selected from the total population with a limiting dilution technique and maintained under the same condition as the SV F4 cells. Cells of passages 5-10 were used. All SV-PDL subclones did not express transcripts for BSP or OCN and were not responsive to PTH-related protein (PTHrP).(27)
Reverse-transcription polymerase chain reaction: evaluation of BMPs and BMP receptors
Expression of transcripts for BMP-2, BMP-4, and BMP receptors (BMPRs) IA and II was determined by reverse-transcription polymerase chain reaction (RT-PCR), using the GeneAmp RNA PCR Kit (Perkin Elmer, Boston, MA, USA). Total RNA was isolated from cells, digested with DNAse I to remove genomic DNA contamination, and reverse-transcribed by murine leukemia virus (MuLV) reverse transcriptase, with random hexamers, followed by PCR with appropriate primers. RNA from a preosteoblastic cell line MC3T3-E1 cultured in the presence of ascorbic acid (AA) was used as a positive control. For negative controls, distilled water alone or RNA without RT reaction was used. Products of PCR were subjected to electrophoresis on a 1.2% agarose gel and visualized by staining with ethidium bromide.
The 5′ and 3′ primers used were designed for mouse tissue according to Otsuka et al.(32) as follows: BMPR IA (448 bp), 5′-TAGCACCAGAGGATACCTTGC-3′ and 5′-AATGCTTCATC CTGTTCCAAA-3′; BMPR II (474 bp), 5′-AATCAAGAACGGCTGTGTGCA-3′ and 5′-CATGCT GTGAAGACCCTGTTT-3′; BMP-2 (403 bp), 5′-CGGGAACAGATACAGGAAGC-3′ and 5′-GCTGTTTGTGTTTGGCTTGA-3′; BMP-4 (493 bp), 5′-ACTCACCTCCACCAGACACG-3′ and 5′-CCTCTACCACCATCTCCTGA-3′; GAPDH (311 bp), 5′-GGCAAATTCAACGGCACAGTC-3′ and 5′-AAGCAGTTGGTGGTGCAGGA-3′.
Northern blot analysis
To determine the effects of BMP-2 on gene expression, cells were plated in 60-mm cell culture dishes, at a density of 20,000 cells/cm2 in DMEM with 10% FBS. At confluence, media were changed to DMEM with 2% FBS, ±AA (50 μg/ml) and ±BMP-2 and incubated for 5 days. Media were changed on day 2 and day 4. For initial experiments, BMP-2 was used at a concentration of 100 ng/ml. Total RNA was isolated using Trizol reagent (Gibco BRL) and RNA concentration was quantified by spectrophotometer. RNA (6 μg) was denatured, fractionated on a 6% formaldehyde and 1.2% agarose gel, transferred to nylon membrane (Duralon-UV; Stratagene, Inc., La Jolla, CA, USA), and cross-linked by a Stratalinker (Stratagene, Inc.). rhBMP-2 protein was a generous gift from the Genetics Institute (Cambridge, MA, USA). To confirm that immortalized follicle cells responded to BMP-2 in a similar fashion to primary cells, primary dental follicle cells were examined in parallel experiments for responsiveness to BMP-2.
Blots were hybridized with random primed32P-labeled probes (Rediprime; Amersham-Pharmacia Biotech, Arlington Heights, IL, USA) and were exposed to Eastman Kodak Co. (Rochester, NY, USA) X-omat or Biomax film with intensifying screens at −70°C for 24-72 h. Probes used for Northern blots were BSP, M-BSP consisting of 1 kb of mouse cDNA in PCR II(33); OCN, 400 bp of mouse OCN cDNA in pSP65(34) (obtained from Dr. J. Wozney, Genetic Institute); Col I, bovine α2(I) procollagen cDNA; osteopontin (OPN), MOP-3 consisting of 1 kb of mouse OPN cDNA in PCR II(35) (gifts from Dr. M. Young and Dr. L. Fisher, National Institute of Dental and Craniofacial Research/National Institutes of Health [NIDCR/NIH]); core binding factor α1/osteoblast-specific factor 2 (Cbfa1/OSF2), mouse Cbfa1 cDNA (from Dr. G. Karsenty).(36) Hybridized blots were also scanned using a Packard A2024 Instantimager and all values were normalized for RNA loading by probing blots with cDNA to 18S rRNA. Experiments were carried out two times with comparable results.
Mineralization assay in vitro
In parallel with examining markers for osteoblast/cementoblast phenotype, the effect of BMP on mineral nodule formation was determined. Follicle cells and PDL cells were plated at the density of 20,000 cells/cm2 in 24-well culture plates in DMEM containing 10% FBS. On confluence, designated day 0, media were removed and cells were incubated in mineralizing media (DMEM with 2% FBS, 50 μg/ml of AA, and 5 mM of β-glycerophosphate),(26) supplemented with rhBMP-2 (0-100 ng/ml). Media were changed every other day. On designated days up to 14 days for follicle cells and 21 days for PDL cells, the ability of cells to promote mineral nodule formation was determined by von Kossa staining.(37) Experiments were performed twice with comparable results.
To quantitate the relative intensities of cell mineralization with or without BMP-2 treatment, scanned images were analyzed with Scion Image software (Scion Corp., Frederick, MD, USA). The intensity of mineral formation by follicle cells and PDL cells, without BMP-2 treatment, was designated as 1.00.
Dissection of signal transduction pathways
Next, assays were designed to establish the signaling pathway(s) by which BMP regulates genes associated with the osteoblast/cementoblast phenotype, BSP and OCN by using inhibitors of cAMP-dependent protein kinase A (PKA), protein kinase C (PKC), and MAPK pathways. Follicle cells, at confluence, were treated with or without rhBMP-2 (100 ng/ml) for 5 days and total RNA was extracted for Northern blot analysis. Inhibitors were added for the last 24 h. A cell-permeable adenylyl cyclase inhibitor 9-(2-tetrahydrofuryl) adenine (THFA; 100 μM; Sigma) was used to inhibit PKA pathway activity and GF109203X (3 μM; Calbiochem, San Diego, CA, USA) was used to repress the PKC activity.(38) The compound U0126 (Calbiochem), a direct and potent inhibitor of MAPK/MEK-1/2,(39) was used to block MAPK activity. For negative controls, DMSO (used as a solvent for all inhibitors) and U0124 (a structural analog of U0126 without MEK inhibition activity) were used. Experiments were repeated twice with comparable results.
Western blot analysis of Erk-1/2 phosphorylation
Western blot analysis was performed as described previously.(40) Briefly, whole cell extracts were prepared using a lysis buffer containing 1% NP-40, 5% sodium deoxycholate, 1 mM of phenylmethylsulfonyl fluoride (PMSF), 100 mM of sodium orthovanadate, and 1:100 protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO, USA). The proteins were resolved in a 12% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membrane using a semidry transfer apparatus (Bio-Rad Laboratories, Hercules, CA, USA). The membrane was probed with primary rabbit antibodies against p44/42 MAPK (Erk-1/2) or phospho-p44/42 MAPK (phosphorylated Erk-1/2; Cell Signaling Technology, Inc., Beverly, MA, USA) at a dilution of 1:1000, followed by horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG; 1:5000). The response was visualized using an enhanced chemiluminescence (ECL) reagent, according to the manufacturer's recommendation (Amersham Pharmacia Biotech).
Expression of BMPs and BMPRs
To determine whether or not unstimulated follicle cells or PDL cells had the capacity to respond to BMPs, cells were evaluated for expression of BMP-2, BMP-4, and BMPRs mRNAs using RT-PCR. As shown in Fig. 1, both cell lines exhibited transcripts for BMP-4 and BMPR IA and II, while only follicle cells expressed BMP-2. Murine primary follicle cells (data not shown) and the murine osteoprogenitor cell line MC3T3-E1 cells, used as a positive control, expressed BMP-2, BMP-4, and BMPRs.
Effects of BMP-2 on gene expression
Having shown that both PDL and follicle cells express receptors for BMPs, studies were designed to determine the responsiveness of cells to BMP-2. Cells at confluence were treated with BMP-2 (100 ng/ml) for 5 days and RNA was extracted for analysis of mineral-associated genes. As seen in Fig. 2, exposure of follicle cells to BMP-2 resulted in induction of BSP mRNA and an increase in OCN mRNA, regardless of the presence of AA, while OPN transcripts were decreased. When cells were exposed to BMP-2 and AA, in combination, BMP-2 exhibited a more dramatic effect on increasing transcripts for OCN and BSP. In the presence of AA, BMP-2 up-regulated Col I expression. In contrast to follicle cells, BMP-2 induced only modest expression of OCN mRNA in PDL cells, while having no effect on mRNA levels for Col I and OPN. BSP mRNA was not detected in PDL cells under any of the conditions examined here. Based on these results, only follicle cells were used in all subsequent experiments except for in the mineralization assay.
Effects of BMP-2 on gene expression: dose response
To determine the dose of BMP-2 required to increase transcripts for specific markers, follicle cells were treated with BMP-2 at concentrations from 0 to 100 ng/ml, with or without AA for 5 days. Results from this experiment revealed that follicle cells responded to BMP-2 in a dose-dependent fashion (Fig. 3). In the absence of AA, strong expression of BSP transcripts was noted at 100 ng/ml of BMP-2. This response was much more dramatic in the presence of AA, with BSP transcripts noted at 10 ng/ml of BMP-2 and peaking at 50 ng/ml of BMP-2. A similar pattern was noted for OCN mRNA levels in follicle cells exposed to BMP-2, but increased levels of OCN mRNA were noted up to 100 ng/ml of BMP-2. In addition, Col I mRNA was enhanced when cells were exposed to BMP-2 and AA in combination, but not to BMP alone. In the absence of AA, BMP-2 had a minimal effect on OPN mRNA levels; however, in the presence of AA, BMP-2 down-regulated OPN expression, with a marked effect noted at 100 ng/ml. Based on these results, 100 ng/ml of BMP-2 was selected for use in further studies described here.
Effects of BMP-2 on gene expression: time course
Next, the time required for BMP-2 to alter the expression of BSP and OCN in follicle cells was determined. Follicle cells were exposed to BMP-2 (100 ng/ml) for 1, 3, and 5 days, followed by RNA extraction and Northern blot analysis for expression of Cbfa1 as well as BSP and OCN. Cbfa1, an osteoblast-specific transcription factor, has been shown to regulate bone markers and to promote osteoblastic differentiation.(36,41,42) Thus, Cbfa1 transcripts were examined in parallel with BSP and OCN to search for possible correlations between the effects of BMP-2 on BSP and OCN expression versus Cbfa1.
As shown in Fig. 4, Cbfa1 mRNA levels were enhanced in cells exposed to BMP-2 at all the time points examined. AA did not appear to enhance the BMP-2-mediated increase in Cbfa1 transcripts. Consistent with Cbfa1 being a master switch required for osteoblast maturation,(36,41,42) increased expression for BSP and OCN paralleled the expression of Cbfa1. This finding suggests that Cbfa1 may, in part, control the downstream genes associated with the differentiation of follicle cells. Cells exposed to BMP-2 exhibited a minimal increase in BSP and OCN mRNAs at 24 h but with time a marked increase was noted.
To determine whether protein synthesis was required for the BMP-2-induced response, confluent follicle cells were treated with BMP-2 (100 ng/ml) with or without AA, and cycloheximide (5 μM) was added on day 3 for 15 h. In comparison with untreated control samples, cycloheximide-treated cells exhibited a decrease in transcript for both BSP and OCN (data not shown), suggesting that the BMP-2-mediated effect requires protein synthesis.
Biomineralization in vitro
In previous studies, we reported that follicle cells cultured long-term (18 days) were capable of inducing mineral nodule formation to a limited extent.(28) The data showing that BMP-2 enhanced expression of genes associated with the mature osteoblast/cementoblast prompted us to determine the effect of BMP-2 on mineral nodule formation by follicle cells. As shown in Fig. 5, consistent with the effects of BMP-2 on gene expression, BMP-2 enhanced mineral nodule formation by SV F4 cells in a time- and dose-dependent manner. Quantitative analysis, as measured by intensity of staining, revealed that at all time points assayed, BMP-2 at 50 ng/ml and 100 ng/ml greatly increased mineral nodule formation in comparison with controls (without BMP-2; Fig. 5B). In contrast, PDL cells exposed to BMP-2 exhibited minimal mineral formation, even at the higher doses of BMP-2 (Fig. 5). In the absence of BMP-2, no mineral nodule formation was noted in PDL cells, even when studies were extended for 21 days.
Effect of noggin on BMP-2-mediated gene expression
Noggin, an extracellular BMP binding protein, forms an inactive complex with BMP-2, BMP-4, and BMP-7 and thus prevents such BMPs from binding to their receptors.(43) To confirm the specificity of BMP-2 action on follicle cells, cells were cultured in DMEM with 2% FBS, +BMP-2 (100 ng/ml), ±AA (50 μg/ml), and ±noggin (1 μg/ml) for 5 days and total RNA was extracted for Northern blot analysis. As seen in Fig. 6, noggin abolished BMP-2-induced expression of BSP and dramatically suppressed OCN expression. Furthermore, noggin also inhibited the additive action of AA and BMP-2 on OCN gene expression.
Dissection of signal transduction pathways
As an initial determination of the signaling pathways involved in BMP-2-mediated effects on follicle cells, specific inhibitors of PKA, PKC, and MAPK were used. For these experiments, the ability of inhibitors to suppress rapidly gene expression induced by prior exposure to BMP-2 was determined. As shown in Fig. 7, neither the PKA inhibitor THFA (100 μM) nor the PKC inhibitor GF109203X (3 μM) inhibited BSP or OCN gene expression, suggesting that these two pathways are not involved in BMP-2-induced differentiation of follicle cells. In marked contrast, the specific inhibitor of MEK-1/2 (U0126 [20 μM]) reduced the BMP-mediated BSP and OCN expression to near baseline levels when compared with vehicle-treated control cells or with cells treated with U0124, a structural analogue of U0126 without MEK inhibitory activity.
Dose-dependent effect of U0126
Next, the dose of U0126 required to inhibit BMP-2 activity was determined. As shown in Figs. 8A and 8B, U0126 inhibited BMP-2-induced expression of BSP and OCN at a dose as low as 2.5 μM. U0126 slightly reduced the mRNA levels of Cbfa1 to 75-90% of the untreated control (result not shown). To examine the effect of U0126 on phosphorylation of Erk-1/2, downstream molecules of MEK-1/2, follicle cells exposed to 100 ng/ml of BMP-2 for 4 days were treated with U0126 for 15 minutes and cell extracts were harvested for Western blot analysis. U0126, at a concentration as low as 2.5 μM, inhibited phosphorylation of Erk-1/2 in follicle cells (Fig. 8C). Findings from these experiments confirm that U0126 is a rapid and potent inhibitor of the MAPK pathway and that activation of the MAPK pathway is essential for BMP-2-mediated gene expression in follicle cells.
The inhibitors did not appear to be toxic to cells at the doses used, in that none of the compounds affected cell number or morphology (data not shown), except for a slight increase in cell death at 40 μM of U0126.
BMPs, both native protein and the recombinant form, are known to have osteoconductive ability.(9,44) As a result, several BMPs have been evaluated for their ability to promote fracture healing and periodontal regeneration using in vivo models.(4,8,16,45) Results from such studies indicate some success, although results are not predictable and often the extent of regeneration is not 100%.(7) To improve the outcome of regenerative therapies, it is necessary to understand the response of cells within the local environment to specific factors. In the present studies we have focused on determining the responsiveness of follicle cells and cells of the mature periodontium, PDL cells, to BMP-2. Notably, BMP-2 promoted follicle cells to differentiate along a cementoblast/osteoblast pathway as measured by increased expression of Cbfa1, BSP, and OCN mRNA, as well as by enhanced mineral nodule formation in vitro, in a time and dose-dependent manner. Also, BMP-2 increased alkaline phosphatase activity, an early marker of osteoblastic differentiation, in follicle cells in a dose-dependent manner (data not shown). These results support existing data suggesting that follicle cells have the ability to differentiate into cementoblasts and osteoblasts; however, specific factors promoting their differentiation were not identified in past studies.(46,47) In fact, our group reported that unstimulated follicle cells begin to express transcripts for BSP and OCN with time in culture and also promote mineral nodule formation.(28) Because dental follicle tissues contain a heterogeneous population of mesenchymal cells, it is possible that with time in culture cells having the capacity to differentiate along the cementoblast/osteoblast pathway were selected out, and/or that with time in culture BMPs secreted by the cells promoted expression of BSP and OCN. In this study, BMP-2 stimulated expression of BSP and OCN transcripts in follicle cells as early as 1 day in vitro; however, we do not know whether all of the cells responded to BMP-2 treatment or only a selected population responded. Currently, these questions are being addressed using clonal follicle cell populations. Importantly, to confirm that immortalization did not modify the properties of follicle cells, primary follicle cell cultures were examined for expression of BMP-2, BMP-4, and their associated receptors and for responsiveness to BMP-2. Results showed that primary cultures expressed BMP-2, BMP-4, and their receptors and responded in a similar fashion to BMP-2 as SV follicle cells, as measured by increased expression of BSP and OCN (data not shown).
In contrast to the effect of BMP-2 on follicle cells, the mature PDL cell population exhibited a minimal response to BMP-2. Because a clonal population of PDL cells was used, it is possible that other PDL clones may be more responsive to BMP-2, although this is highly unlikely. In previous studies using several PDL subclones and also primary cultures of human PDL cells, we showed that SV-PDL subclones had identical properties with human PDL cells, for example, lack of BSP and OCN expression, no significant PTHrP-mediated cAMP response, and inability to promote mineralization,(27) suggesting that PDL cells have a limited capacity to act as osteoblasts/cementoblasts, even when exposed to “differentiation” factors.
Based on these results, we focused on defining the specific genes and signaling pathways involved in BMP-2-mediated follicle cell activities. Existing data indicate that BMPs increase expression of Cbfa1 and subsequently BSP and OCN expression.(36,48,49) Therefore, Cbfa1 was examined in parallel with expression of BSP and OCN mRNA (Fig. 4). In comparison with control cells, in the presence of BMP-2, Cbfa1 mRNA levels were elevated in a time-dependent manner. More importantly, elevation of Cbfa1 mRNA paralleled BMP-2-induced/enhanced expression of BSP/OCN mRNA, thus indicating that BMP-2-mediated increase in Cbfa1 expression might be responsible, in part, for the differentiation of follicle cells. This finding is consistent with results from other groups that have reported increased levels of Cbfa1 mRNA expression in the immortalized human bone marrow stromal cell line,(48) in C2C12 cells(49) and in 2T3 cells(50) in response to BMP-2. However, Cbfa1 is constitutively expressed in follicle cells and, in addition, although an increase in expression of Cbfa1 over basal levels was observed in cells treated with BMP-2, in three separate experiments, the increase was minimal. This finding suggests that other transcription factors are involved in the ability of BMP-2 to promote BSP and OCN gene expression and/or that Cbfa1 phosphorylation is involved. The latter possibility is supported by studies of Xiao et al.(51) in which they showed that phosphorylation of Cbfa1 enhanced OCN promoter activity in MC3T3 E1 cells and that this activity was mediated through the MAPK pathway.
Analysis of the signaling mechanisms responsible for the ability of BMP-2 to promote differentiation of follicle cells identified the MAPK pathway as an essential pathway, while neither the PKA nor PKC pathway appeared to be involved (Fig. 7). U0126 inhibited both BSP and OCN expression and Erk-1/2 phosphorylation at a similar range of doses (Fig. 8). Interestingly, U0126 treatment only slightly reduced Cbfa1 mRNA (data not shown), providing additional evidence that BSP and OCN gene expression is regulated by factors beyond changes in mRNA levels of Cbfa1, for example, phosphorylation of Cbfa1 as discussed previously. It has been shown that collagen-integrin interactions are important for differentiation of osteoblast progenitor cells MC3T3-E1 cells(52) and also for BMP-2-induced differentiation of SV 40 TAg immortalized 2T3 osteoblast clonal cells.(53) In this study, cells exposed to BMP-2 and AA in combination exhibited a greater increase in BSP and OCN mRNA levels when compared with cells exposed to either AA or BMP-2 alone. Thus, our data support the findings that collagen-integrin interactions are important for BMP-2-mediated cell differentiation(53) and future experiments are planned to confirm this issue.
Although evidence that activation of the MAPK pathway is required for BMP-2-induced osteoblast differentiation was provided by experiments of Gallea et al. on C2C12 cells,(54) the specific mechanisms by which BMP-2 activates the MAPK pathway remains to be defined. It is well established that BMPs trigger their responses by binding and bringing together a class of transmembrane serine/threonine kinase receptors, types IA, IB, and type II, and we now show that receptors IA and II are present on follicle cells (Fig. 1). The BMPR interaction results in phosphorylation of Smads, which translocate from cytoplasm into the nucleus and act as transcriptional activators.(55,56) Smads have been shown to participate in osteoblastic differentiation of C2C12 cells induced by BMP-2 or by constitutively active forms of BMPR I.(57–59) In contrast to Smads, the role of the MAPK pathway in BMP signaling is less defined and appears to vary with experimental system. Direct induction of Erk-1/2 phosphorylation by BMP-2 has been reported to be related to enhanced Col I synthesis in ROS 17/2.8 cells(60) and with osteoblastic differentiation in C2C12 cells.(54) However, BMP-2-induced MAPK activity in C3H10T1/2 cells appears to be a latent and sustained response, suggesting both direct and indirect pathways are involved in BMP-2-MAPK activities.(61) In our study with follicle cells, BMP-2 had no effect on Erk-1/2 when compared with untreated cells (results not shown). This result indicates that other signaling factors, possibly Smads, may play a direct and critical role in BMP-2 induction of follicle cell differentiation, and the MAPK signaling pathway may play a permissive role. There is evidence that TGF-β and BMP-2 can phosphorylate endogenous Smad1 in intestinal epithelial cells (IECs) and this process can be inhibited partially by inactivation of the MAPK pathway through expression of a dominant-negative mutant of Ras or addition of a MEK inhibitor PD98059.(62) Also, existing data indicate that Smad5 is required for Cbfa1 induction and the cooperation between Cbfa1 and BMP-2-activated Smad5 induces osteoblast-specific gene expression in C2C12 cells.(63) In contrast to the positive effects of Smad5 on Cbfa1, Smad3, activated by TGF-β, represses Cbfa1 transcription and, subsequently, activation of osteoblast differentiation.(64) These data suggest that BMP/TGF-β-activated Smads are involved in complex signaling events, including regulation of Cbfa1 activity, that control expression of specific genes associated with osteoblast/cementoblast differentiation. A plausible explanation for the BMP-2-mediated response in follicle cells may be that BMP-2, by binding to BMP receptors on these cells, results in phosphorylation of Smad1, Smad5, and/or Smad8 and, subsequently, activation of specific promoters required for differentiation of follicle cells and that this process requires participation of the MAPK pathway molecules and specific transcription factors such as Cbfa1. The relationship of Smads, MAPK, and Cbfa1 in BMP-2-mediated response of follicle cells is now under investigation in our laboratory.
In contrast to BMP-2 effects on follicle cells, PDL cells exposed to BMP-2 did not express a cementoblast/osteoblast phenotype, although a modest induction of OCN was consistently observed. The SV-PDL cells used here were a cloned cell line established based on lack of BSP and OCN expression.(65) Thus, it is possible that we selected for a population of PDL cells having more fibroblastic versus osteoblastic properties. There has been considerable controversy as to the properties of PDL cells. Data from some groups suggest that all cells within the PDL region have the capacity to differentiate toward an osteoblast/cementoblast phenotype,(66,67) and other groups suggested that most cells within the PDL region have factors that protect against mineralization.(68) Using a heterogeneous population of human PDL cells, Kogayashi et al. reported that BMP-2 promoted osteoblastic differentiation,(13,69) and others have shown that heterogeneous PDL cells, in vitro, exhibit some but not all the properties of an osteoblast/cementoblast phenotype.(66,67,70–72) A logical explanation for these apparent discrepancies is that the PDL region contains a specific population of undifferentiated mesenchymal progenitor cells that are responsive to BMPs and that in addition to cells from the PDL region, cells come from the marrow at the healing site. To address these questions, specific biomarkers for PDL cells need to be identified.
In summary, this study identified BMP-2 as a trigger factor for promoting dental follicle cells to differentiate along the cementoblast/osteoblast pathway. This BMP-2-mediated response was dose- and time-dependent and required participation of the MAPK. In contrast, PDL fibroblastic cells showed only a modest response to BMP-2 treatment in both gene expression and in vitro mineralization. These findings provide valuable insight into the factors, cells, and mechanisms involved in triggering periodontal precursor cells to function as osteoblasts/cementoblasts. This information is important for determining the specific factors and cell types required to promote regeneration and designing predictable therapies for use in regenerating periodontal tissues lost to disease.
We thank the Center for the Biorestoration of Oral Health (CBOH) at the University of Michigan, School of Dentistry, for providing an atmosphere conducive to promoting excellence in oral health research. Special thanks go to Genetics Institute for providing rhBMP-2. This study was supported by the NIH/NIDCR grants DE09532 (to M.J.S.) and DE11723 (to R.T.F.).