Mitogenic signaling mechanisms of human cementum-derived growth factors.

Cementum-derived growth factor (CGF) is a M(r) 23,000 protein, which is sequestered in the mineralized matrix of tooth cementum. We have investigated the mitogenic signaling reactions induced by CGF using quiescent human gingival fibroblasts as target cells. Cells activated with CGF were compared with those treated with CGF plus epidermal growth factor (EGF) and other growth factors. CGF caused a transient increase in cytoplasmic Ca2+ concentration, and this was accompanied by enhancement of membrane protein kinase C activity, myelin basic protein and S6 kinase activities, inositol phosphate levels, and activation of c-fos and jun-B gene expression. Membranes obtained from cells activated with CGF contained several protein bands, which cross-reacted with antiphosphotyrosine antibody; however, proteins corresponding to a putative phosphorylated CGF receptor were not detected. DNA synthesis induced by CGF was inhibited by 65% in cells treated with pertussis toxin but only 25-29% in cultures exposed to H7 or 12-O-tetradecanoylphorbol-13-acetate; these values were different from those obtained when EGF, PDGF, or fetal bovine serum were used as mitogens. CGF and TGF-beta, but not EGF, caused an increase of PDGF-A chain mRNA expression 4 h after mitogen addition. However, while CGF was mitogenic for gingival fibroblasts, TGF-beta was not. Kinetics of DNA stimulation and experiments with anti-PDGF antibodies indicated that PDGF-A expression does not contribute significantly to CGF-induced DNA synthesis. When the stimulation of various signaling pathways induced by CGF and other growth factors was compared, the pattern of stimulation by CGF was different from other growth factors. The characteristic signaling reactions of CGF are likely to be important components of the mechanisms that regulate the formation and regeneration of cementum and adjacent connective tissues.

include PDGF,' TGF-a, TGF-P, IL-1, and interferons, are secreted by platelets, macrophages, and other inflammatory cells during inflammation, and they are believed to influence the events involved in wound healing and tissue repair (1). Other growth factors, especially acidic and basic fibroblast growth factors and TGF-P are sequestered in the extracellular matrix, where the matrix appears to store and protect these molecules from denaturation and degradation (2)(3)(4)(5).
Activation of cells to initiate DNA synthesis and complete cell division is associated with a series of signaling events that include PIP2 hydrolysis and generation of IPS and 1,2-diacylglycerol, mobilization of Caz+, PKC activation, and transcription of c-myc, c-fos, jun, and other protooncogenes (6,7). These reactions are initiated by the binding of growth factors to their specific receptors. So far, two major signal transduction mechanisms have been shown to be associated with initiation of DNA synthesis and cell division. The first is autophosphorylation of some growth factor receptors at specific tyrosyl residues; the receptors then undergo dimerization and ligand-dependent association with SH-2 domains of PI3kinase, GTPase-activating protein, and PI-specific phospholipase C-71 and activate these enzymes (8)(9)(10)(11). This appears to be the major pathway of action of a group of mitogens, which includes PDGF, EGF, and IGF-1. The receptors for these proteins contain an inherent growth factor-dependent protein tyrosine kinase activity (8). In contrast, substances such as bombesin, thrombin, and bradykinin catalyze the breakdown of PIPz by a mechanism independent of receptor phosphorylation. This occurs through a family of GTP-binding proteins, some of which are inactivated by pertussis and cholera toxins (12)(13)(14). Irrespective of apparent triggering of a major signaling pathway, most growth factors activate a multitude of reactions, and no single event appears to be sufficient or necessary for the completion of DNA synthesis and cell division (15)(16)(17)(18)(19)(20)(21)(22)(23). Many of the signaling reactions are also redundant, and they are affected by the type and concentration of growth factors, duration of exposure to them, and the target cell type.
Cementum is the junctional interface through which collagen fibers of adjacent soft connective tissues are inserted into tooth root surfaces. It is a unique mineralized tissue which contains very few cells and is devoid of blood supply. The structural integrity of cementum is essential for the mobility and function of teeth and pathological alterations of cemen-The abbreviations used are: BSA, bovine serum albumin; CGF, cementum-derived growth factor, H7, 1-(5-isoquinolinylsuIfonyl)-2methylpiperazine; IL-1, interleukin-1; IP,, inositol (1,4,5)triphosphate; PDGF, platelet-derived growth factor; PI, phosphatidylinositol; PI3 kinase, phosphatidylinositol 3-phosphate kinase; PIP,, phosphatidylinositol 4,5-biphosphate; PKC, protein kinase C; TGF, transforming growth factor; TPA, 12-0-tetradecano~lphorbol-13-acetate; EGF, epidermal growth factor; HPLC, high pressure liquid chromatography. turn, and adjacent tissues cause the loss of soft tissue attachment to teeth and lead to tooth loss. Attempts to restore normal cementum and reestablish connective tissue attachment are mostly unsuccessful, and results are u n p~d i c~b l e . Recent observations indicate that cementum contains novel components, which regulate the formation and regeneration of connective tissues, and that biochemical changes in cementum components could explain why diseased periodontal tissues do not effectively regenerate (24, 25). We have recently isolated one such component from cementum, which is a M , 23,000 polypeptide growth factor (26,27). This growth factor, referred to as cementum-derived growth factor (CGF), resembles PDGF in its chromatographic properties, susceptibility to reduction, and heat resistance; however, it consists of a single subunit and it is not recognized by anti-PDGF antibodies (27). An examination of its properties indicated that it is also different from other growth factors? and that it may be a novel mitogen present in cementum but not in adjacent structures (28). Because CGF is the major growth factor present in cementum (261, we have performed experiments to understand its mode of action. Our results show that this growth factor resembles other mitogens in inducing a battery of signaling events, yet the type of reactions catalyzed and their degree of stimulation are distinct from other known growth factors.

EXPERIMENTAL PROCEDURES
Materials-c-fos and jun-B cDNAs were generous gifts from Dr.
[32P]dCTP, [3H]inositol, and PKC assay kit were obtained from Amersham Cow. TPA, bombesin, and H7 were from Sigma, and pertussis toxin was from List Biological Laboratories, Campbell, CA. All other chemicals were of analytical grade.
Purification of CGF-CGF was obtained from human cementum as described previously (27). Briefly tissue was harvested, extracted in 1.0 M CHsCOOH containing a battery of proteinase inhibitors, fractionated by hep~in-affinity HPLC, and subjected t o p~i p i t a t i o n by trichloroacetic acid and cation exchange and Cls reverse-phase HPLC. For experiments described here we used pooled CIS HPLC fractions that contained CGF and an additional M, 14,000 component, which was not mitogenic (27).
Assay of DNA Synthesis-Human gingival fibroblasts between 4th and 12th transfers were used as target cells. Assays were performed in q u a~p l i c a t e in 36-well dishes (Falcon) using lo4 cells/well. The modified Eagle's medium for 24-48 h and then mitogens added? After cells were made quiescent by incubation in serum-free Dulbecco's 22 h fresh serum-free medium with 10 pCi/ml [3H]thymidine was added. Six h later the cells were washed with ice-cold 0.1 M NaH2P04, pH 7.2, buffer containing 0.15 M NaCl and 5% trichloroacetic acid, taken in 1.0 N NaOH, neutralized, and counted in a Packard 1500 Tri-Carb liquid scintillation counter (26,27).
Measurement of Intracellular Calcium-Intracellular Ca2+ levels of The CGF is a single polypeptide, the electrophoretic migration of which is not significantly affected by reduction. It is heat-resistant but differs from PDGF in properties (26, 27). Recently we have obtained an internal amino acid sequence for CGF, which had no homology with other growth factors (K. Yonemura, C. Hart, and A. S. Narayanan, unpublished data).
The concentrations chosen induced optimal mitogenic stimulation as determined by dose curves. When CGF and EGF were added together, CGF was present at 20% optimal concentration. The actual CGF concentration is likely to be lower than reported because CIS fractions used for experiments contained an additional component (27). human gingival fibroblasts were measured with the fluorescent Ca" indicator Indo-1. Subconfluent fibroblasts on glass coverslips (Nunc) were incubated in serum-free medium for 24 h, and the medium was replaced with Hanks' balanced salt solution containing 1 mg/ml BSA.
The cells were exposed to 2 p~ Indo-I-AM for 1 h at 37 "C, then rinsed with Hanks' balanced salt solution containing 1 mg/ml BSA, and maintained at 37 "C. Single cells were monitored for their change in Indo-1 fluorescence in response to vehicle (0.1 mg/ml BSA) in the presence or absence of growth factor using a digital imaging fluorescence microscope (ACAS 570, Meridian Instruments, Okemos, MI). Intracellular free calcium in its bound and free states was measured by exciting Indo-1 dye with 351-363-nm lines of the argon laser and simultaneously detecting two emissions a t 405 and 485 nm (29). The ratio of these emissions was computed on a single cell basis using a ratiometric kinetics software package. Intracellularly free Ca2' was calculated by comparing Indo-1 fluorescence ratios in a series of standards with the same detector settings used to acquire cellular data.
Serine/Threoni~ Kinase Assays-The methods and conditions for measuring kinase activation have been described previously (31,32). Briefly, confluent cultures in 100-mm dishes were extracted by sonication for 10 s in 50 mM P-glycerophosphate, 1.5 mM EGTA, 0.1 mM Na3V04, 1 mM dithiothreitol, 10 pg/ml leupeptin, 10 @&/mi aprotinin, 2 pg/ml pepstatin A, and 1 mM benzamidine, After centrifugation at 100,000 X g for 20 min at 4 "C, cytosolic supernatants (12.5 pl) were mixed with substrate (4.2 pl), and phospho~lation reactions were After incubation for 20 min a t 30 "C with substrates (0.25 mM SS peptide, 0.33 mg/ml myelin basic protein, 1 mM syntide-2, or 1 mM casein kinase I1 peptide) 20 pl of the assay mixture were spotted onto P-81 phosph~ellulose filter paper squares ( W~t m~~, washed five times with 175 mM phosphoric acid, and counted in a liquid scintillation counter. Determination of Protein Tyrosine Kinase Activation-Confluent cultures were exposed to mitogens. At indicated times, cells were lysed 20 mM Tris-HC1, pH 8.0, 137 mM NaCl, 10% glycerol, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 0.15 units/mI aprotinin, 10 mM EDTA, 10 pg/ml Ieupeptin, 100 mM sodium fluoride, and 2 mM sodium orthovanadate at 4 "C for 20 min. The lysates were centrifuged at 4 "C for 15 min at 15,000 X g, 100 pg of protein in the lysate were separated by S D S -p o l y a c~l~i d e gel electrophoresis in 7.5% gels, and proteins were electroblotted onto a polyvinylidene difluoride membrane. The membrane was blocked with 1% gelatin for 2 h at 25 "C and incubated for 3 h with 0.25 pg/ml anti-phosphotyrosine antiserum. It was washed in 0.05% Tween 20 and developed with 1.0 pCi/ml high specific activity 12*I-Protein A (33).
RNA Blot Analysis-Total RNA was isolated by the guanidinium thiocyanate method of Chomczynski and Sacchi (34). 15 pg of total RNA/lane was separated in 1% agarose gels containing formaldehyde and transferred to a 0.45-pm Nytran filter (Schleicher & Schuell) by capillary transfer. Filters were hybridized with cDNAs labeled with [32P]dCTP to specific activities of 2 X lo9 cpm/pg using Amersham random prime Iabeling kit. The probes were a 103-base pair AuaI-StuI c-fos fragment, a 180-base pairjun-B cDNA obtained by BamHI digestion, a 1.3-kilobase PDGF-A chain clone Dl, and PDGF-B chain 704-base pair BamHI fragment. After hybridization, the membranes were washed with 0.2 X SSC-0.1% SDS (pH 7.2) at 65 "C and exposed to Kodak XAR-2 film at -70 "C (35).
Assays for PDGF-Media and cells from -10' gingival fibroblasts were collected and concentrated. They were incubated for 1 h in acetic acid containing 1 M NaCl and neutralized, and PDGF was assayed by radioreceptor assay on human dermal fibroblasts using 9 -P D G F -AB (36).

800
Time (sec) FIG. 1. Effect of CGF, EGF, and CGF and EGF ( A ) and 10% fetal bovine serum ( B ) on cytosolic free Cas+ levels in human gingival fibroblasts. Cells were grown on individual glass coverslips and loaded with Indo-1 as described under "Experimental Procedures." Individual cells were monitored for changes in Indo-1 fluorescence by an interactive laser cytometer ACAS570. The data represent a typical profile obtained for three independent experiments. CGF and EGF were added at 10 and 5.0 ng/ml, respectively, at times indicated by the arrows. FBS, fetal bovine serum.

RESULTS
We first examined whether CGF activates early signaling events that are associated with mitogenic stimulation. To determine whether changes in DNA synthesis can be attributed to particular events, we compared cells exposed to CGF alone and CGF plus EGF. The latter treatment was included because, although the CGF is mitogenic by itself, it manifests synergism with EGF, and its activity is potentiated to levels greater than that obtained with 10% serum controls (26-28): The response of other growth factors were also used as controls when necessary. CGF caused a rapid increase in cytosolic Caz+ concentration, which reached 180% of initial value 80 s after mitogen addition and returned to basal level in the next 200 s (Fig. 1A). In contrast, 10% fetal bovine serum caused a 5-fold increase in Caz+ level (Fig. 1B). However, CGF plus EGF treatment, which stimulated DNA synthesis and had mitogenic activity as much as serum, caused only a slight increase over CGF. The latter was roughly equal to that obtained with EGF alone (Fig. 1A).
CGF also increased IPI, IP2, and IP3 levels; the increase began within 1 min and reached a maximum at 10 min (data not shown). This indicated PIPz hydrolysis and generation of 1,2-diacylglycerol. The latter process, along with higher Caz+ levels, activates PKC and causes redistribution of PKC from the cytosol to membranes (37); therefore, cytosolic and membrane PKC activities were measured. CGF induced mobilization of PKC activity from the cytosolic fraction to the mem- branes (Fig. 2 A ) . Membrane activity levels were greater in the presence of CGF than with EGF (Fig. 2B); this increased only slightly when both CGF and EGF were used.
A common growth factor response is activation of serine/ threonine protein kinases; therefore, we measured cytosolic serine/threonine kinase activities using myelin basic protein (MAP-2 kinase), Ss kinase peptide, syntide-2, and casein kinase I1 peptide as substrates. Activities of the kinases for these substrates were increased 10 min after adding CGF or EGF. The levels in the presence of EGF, CGF, and CGF and EGF were not statistically different for the three different combinations (Fig. 3). CGF did not activate casein kinase I1 (Fig. 30).
Expression of c-fos and jun family of cellular protooncogenes is associated with early mitogenic signaling, and the protein products of these genes form the AP-1 transcriptionalactivating factor complex (7, 38, 39). To examine whether CGF activates transcription of these genes, Northern blots were performed using c-fos and jun-B cDNAs as representative probes in cells exposed to mitogens. No c-fos mRNA was detectable in quiescent gingival fibroblasts (Fig. 4, a-c). It was detectable after 30-min exposure to CGF, reached peak   . 4. Northern analysis of c-fos and jun-B mRNA levels in human gingival fibroblasts at the indicated times following treatment with CGF, EGF, and CGF+EGF. In each lane, 15 pg of total RNA was fractionated on a 1% agarose gel and transferred onto a Nytran filter. The filter was hybridized with 32P-labeled c-jos and jun-B probes and exposed at -70 "C for 36 and 48 h, respectively. a-c, c-fos; d-j, jun-B. Rehybridization of the blots with glucose-6phosphate dehydrogenase demonstrated no differences for any of the conditions, confirming an equal load of all lanes. levels after 1 h, and became undetectable after 2 h (Fig. 4a). Fetal bovine serum, PDGF, TGF-@, and IL-1 also increased c-fos mRNA, although in the presence of these substances peak levels were reached earlier at 30 min (data not shown). By densitometry the increase in c-fos mRNA levels was calculated to be IO-, 9-, 19-, and 20-fold, respectively, for CGF, EGF, CGF and EGF, and serum treatments. The expression of jun-B mRNA, another early response gene, was also transiently induced although in this case maximum levels were present 2 h after mitogen addition (Fig. 4, d-f). The mRNA level of glucose-6-phosphate dehydrogenase was comparable at these time points (data not shown).
These data showed that CGF stimulated early events associated with mitogenic stimulation. Growth factors such as PDGF, EGF, colony-stimulating factor, and insulin-like growth factor-I mediate these events through receptor tyrosine autophosphorylation, resulting in the stimulation of respective protein tyrosine kinase activity, which in turn tyrosine phosphorylates several cytoplasmic proteins (7)(8)(9)(10)(11). To examine whether CGF action is also mediated through receptor tyrosine phosphorylation, fibroblasts were exposed to CGF and tyrosine phosphorylation determined by Western blotting as described under "Experimental Procedures." PDGF' and EGF treatments were compared as controls, because products of these mitogens have been identified. Two protein bands at M, 170,000-180,000 and -300,000 (Fig. 5, bands b and a, respectively) were prominent in cells exposed to PDGF, and in addition several other protein bands were also present migrating with M, 140,000, 85,000, 77,000, 60,000, 52,000, 49,000, m d 42,000 (Fig. 5, bands c-i, respectively). Similar bands were also present in the presence of EGF although d, e, and f migrated slightly differently (Fig. 5). In the presence of CGF the prominent bands of M, 180,000 and 300,000 (bands a and 6) were not detectable. All others were present although the degree of stimulation of bands f-i was considerably less Cells were treated with 5, 5, and 10 ng/ml mitogens, respectively, for 5, 10, and 20 min. Proteins were separated by SDS-polyacrylamide gel electrophoresis and subjected to Western blotting using antiphosphotyrosine antibody. Bands were visualized with lZ5I-Protein A as described under "Experimental Procedures." The migration of molecular mass markers is indicated on the left. The bands a-i represent M, 300,000,180,000,140,000,85,000,77,000,60,000,52,000,49,000, and 42,000, respectively.  and only barely detectable. Cells treated with both CGF and EGF had a pattern identical to that obtained for EGF and no additional bands were detected (data not shown).
The absence of any novel tyrosine-phosphorylated protein bands in cells exposed to CGF indicated that receptor tyrosine phosphorylation probably does not participate CGF-induced signaling reactions to a significant extent. Polypeptides such as bombesin and thrombin catalyze PI hydrolysis and subsequent signaling events through a receptor-tyrosine kinase independent mechanism, which is sensitive to pertussis toxin (12)(13)(14). Therefore we examined whether CGF-mediated DNA synthesis is affected by pertussis toxin. Quiescent gingival fibroblasts were treated with 5 ng/ml pertussis toxin as described in Table I, and DNA-synthesis was measured. The results showed that this toxin reduced DNA synthesis by 65% (Table I). The inhibition was 61% for serum. Bombesin had very little mitogenic activity at concentrations from 1 to 100 nM tested, alone, or with EGF: and the small thymidine PDGF alone does not stimulate mitogenesis in human gingival fibroblasts to the maximum possible by 10% fetal bovine serum unless EGF or 1% plasma-derived serum is present. The potencies of PDGF and CGF are roughly similar (26, 27).
To determine if an increase in any of the various activities measured correlated directly with the stimulation in DNA synthesis, the magnitudes of their increase was compared (Table 111). The increase was 1-5-fold for Ca2+, IP1, serinethreonine kinases, and PKC levels in cells treated with CGF or EGF, and it was greater (up to 10-fold) for c-fos and jun-B mRNA. When both mitogens were present the increase was the sum obtained with each growth factor alone. No clear cut correlation was present between any of the activity measured and DNA synthesis. This was most notable in CGF and EGF treatments when the stimulation of DNA synthesis was considerably greater (Table 111), indicating that other mechanisms may contribute to DNA synthesis. Induction of PDGF-A chain expression is one such mechanism that is believed to contribute to mitogenic activity of IL-1, PDGF-AB, and TGF-/ 3 (44-46). We therefore performed experiments to determine the levels of PDGF-A mRNA expression in cells treated with CGF. TGF-B treatment was included as a control as this growth factor activates PDGF-A mRNA expression, but it is not mitogenic for gingival fibroblasts (45-47). As expected, TGF-B caused a significant increase in PDGF-A mRNA level, which reached maximal levels at 2-4 h after its addition (Fig.  6). CGF also caused an increase at 4 h, although it was less. In contrast, EGF had no significant effect either alone or in the presence of CGF (Fig. 6 ) . The increase in PDGF-A chain mRNA levels caused by CGF,and &fold,respectively. PDGF-AB, serum, and IL-1s also enhanced PDGF-A mRNA levels (data not shown). None of these growth factors had any effect on PDGF-B mRNA levels (data not shown). The levels of PDGF in the cells and secreted into the media were also determined in cells treated with different growth factors (Table IV). Although increased levels of PDGF in the media were observed in all cases where PDGF-A chain mRNA expression was induced, the relative increase in PDGF mRNA levels did not correlate with the amount of PDGF detected in the media or cells.
To determine whether the appearance of PDGF-A chain and its release into media leads to a second wave oE mitogenic activity, DNA synthesis kinetics were assessed. However, no evidence for secondary induction was observed as DNA synthesis occurred at 24 h for all mitogenic stimulants (Fig. 7). TGF-8, as expected, had no significant mitogenic activity.

DISCUSSION
Our data show that CGF induces many of the signaling pathways associated with mitogenesis. These pathways include an increase in cytosolic Ca2+ concentration, PI hydrolysis, activation of PKC cascade, and expression of cellular protooncogenes. Several tyrosine-phosphorylated proteins were also formed. While these data show that CGF resembles other growth factors in activating early signaling pathways, there are distinct differences. For example, we were unable to detect any new tyrosine-phosphorylated protein that may represent putative CGF receptor. Although low level receptor expression may prevent detection of the phosphorylated receptor, the data indicate that CGF is different from PDGF-EGF group of growth factors5 in mediating signaling events through a mechanism independent of receptor tyrosine phosphorylation. The absence of receptor tyrosine phosphorylation may be a reason why tyrosine phosphorylation of bands e-i (Fig. 5) is considerably less than that observed with EGF and PDGF.
CGF also appears to differ from thrombin-bombesin-like mitogens. For example, bombesin has very little mitogenic activity toward human gingival fibroblasts! While bombesininduced mitogenic activity is completely susceptible to pertussis toxin in Swiss 3T3 cells (14), 35% of DNA synthesis induced by CGF is not sensitive to pertussis toxin. Thus, CGF signaling appears to involve both pertussis toxin-sensitive guanidine nucleotide-binding proteins as well as pertussis toxin-independent pathways. Although CGF, EGF (Figs. 2 and 3), PDGF, and serum treatments (data not shown) activate protein kinases to a similar extent, the inhibition by PKC inhibitors for CGF is less than for other growth factors (Table 11) (14). This appears contradictory. One possible reason for this may be the ability of CGF to trigger multiple signaling events that complement and substitute the action of protein kinases. While susceptibility to inhibitors may depend on many factors such as target cell type, effector concentration, duration of exposure, etc., the overall pattern of susceptibility of CGF response to these inhibitors is characteristic and different from other growth factors and provides     additional support for the possibility that CGF may be a novel growth factor.2 The stimulation of various activities measured in the presence of both CGF and EGF is roughly equal to that obtained with each growth factor alone, while DNA synthesis is enhanced to a considerably greater extent (Table 111). Tyrosine phosphorylation of several proteins in the presence of CGF and EGF was comparable with EGF, yet the former treatment was considerably more mitogenic. The increase in c-fos and jun-B levels also do not appear to match DNA synthesis, especially with TGF-j3 and IL-1 (data not shown), which induce these genes but not DNA synthesis. These results are consistent with other studies that show that many of the individual signaling events are not rate limiting for DNA synthesis and that DNA synthesis does not depend upon a single pathway (15-22, 43, 48-50). For example, other mechanisms, phosphatidylcholine hydrolysis and arachidonic acid formation for example (51), may also contribute to DNA synthesis.
Although CGF induces the expression of PDGF-A chains, we do not believe that this process contributes significantly to DNA synthesis. This is because (i) TGF-j3 and IL-1, which are more potent in inducing the PDGF-A mRNA levels than CGF, have very little mitogenic activity for gingival fibroblasts ( Fig. 7) (47); (ii) PDGF-A at 5 x concentration detected in conditioned media had negligible mitogenic activity (data not shown); (iii) no significant peak of DNA synthesis was observed at times corresponding to the induction of the PDGF-A mRNA (Fig.  7); (iv) antibodies that recognize PDGF-A do not significantly inhibit thymidine uptake induced by the CGF (27). Interestingly, while EGF increases the expression of PDGF-A mRNA in other cells (45), it has no effect on human gingival fibroblasts.
CGF is the major mitogen present in cementum matrix, and it is not detected in adjacent dentin or soft tissues (28). Its unique localization raises several possibilities about its function. For example, it may be stored in the matrix in a manner similar to FGFs, released at a time of need (during inflammation, for example), and it may participate in the formation and regeneration of soft periodontal connective tissues. Its ability to induce the expression of PDGF-A, a mitogen that appears early in inflammation (52) and its stimulation of collagen synthesis better than BB or AB dimers in human lung fibroblasts: may be significant in this regard. Additionally, the CGF may promote the migration and growth of progenitor cells present in adjacent structures toward the dentin matrix and participate in their differentiation into cementoblasts (53). The variety of signaling mechanisms induced by CGF may contribute to a pleiotropic influence, and G. Raghu and A. S. Narayanan, unpublished data.
in these actions it may be aided by other growth factors that are also present in the cementum matrix (26).