This work was presented at the Annual Meeting of the American Society for Biochemistry and Molecular Biology, April 24, 1998.
Cementum attachment protein (CAP) is a collagenous protein present in the matrix of tooth cementum that mediates preferential attachment of some mesenchymal cell types, and CAP binding capacity is related to mineralizing tissue-forming capacity in culture. We have examined if adhesion to surfaces containing CAP as the only attachment protein permits human fibroblasts to escape G1 arrest and synthesize DNA, and if adhesion to CAP modulates the levels of cyclins D1 and E. Human gingival fibroblasts (HGFs) were serum-starved, trypsinized, and added to plates coated with CAP or bovine serum albumin (BSA). Cells were then exposed to either 10% fetal bovine serum (FBS) or to cementum-derived growth factor (CGF), an insulin-like growth factor I (IGF-I)-like molecule sequestered in tooth cementum, plus epidermal growth factor (EGF). DNA synthesis was measured as [3H]thymidine uptake, and cyclin D1 and E levels were determined by Western analysis. Cyclin E-dependent kinase (Cdk) activity was assessed in terms of H1 kinase activity in immunoprecipitates of cyclin E. Cells adhering to CAP synthesized DNA, whereas on BSA they remained unattached and did not synthesize DNA. Protein levels of cyclin D1 were higher in cells adhering to CAP in the absence and presence of growth factors. Cyclin E levels were not affected by adhesion alone, but they increased in the presence of growth factors. Cyclin E-associated kinase activity was higher in cells adherent on CAP, and it increased further in the presence of growth factors. Our results indicate that adhesion to CAP increases cyclin D1 levels and cyclin E-associated Cdk activity, and that these increases contribute to cell cycle progression. We previously observed that the signaling reactions induced during adhesion are characteristic of the CAP; together these observations indicate that specific matrix components present in the local environment can contribute to recruitment and differentiation of specific cell types for normal homeostasis and wound healing.
ADHESION TO a substratum is a requirement for anchorage-dependent cells to escape G1 arrest and traverse the cell cycle. In healthy tissues extracellular matrix (ECM) components such as collagen and fibronectin (FN) serve as substrates for the attachment of these cells.(1–4) Cells attach to the ECM components through specific integrin receptors, and binding to the integrins initiates a series of biochemical signaling events that include tyrosine phosphorylation of signaling molecules, cytoskeletal assembly, and activation of transcription factors such as c-fos and c-jun.(3–6) The attachment also leads to sustained activation of the extracellular signal-regulated kinases (ERKs) of the mitogen-activated protein (MAP) kinase cascade family of proteins; these enzymes participate in the transcription of AP-1 protein genes and cyclin D1 expression, which are key events necessary for cell cycle progression.(7–12) Adhesion to substrate also is necessary for activation of cyclin-dependent kinase 2 (Cdk-2) and cyclin A expression.(13–15)
Inflammation and wound healing are associated with destruction of damaged tissues, granulation tissue formation, its remodeling and formation of repair, or regenerated tissue.(16) This process involves participation of several cell types and molecules and requires the recruitment, proliferation, and differentiation of specific cell types at specific stages. Signals for the recruitment of specific cell types and regulation of their activities are provided by substances such as growth factors and ECM components.(16, 17) These substances may be nonspecific and capable of interacting with many cell types, or specific and restricted to certain cell types. For example, FN can mediate the attachment of several cell types whereas laminin promotes the attachment of only some cell types. Thus, the composition of ECM in the local environment could determine the types of cells participating in wound healing and determine whether a damaged tissue heals by regeneration or repair.
There is evidence indicating that local environment in cementum matrix may play a role in regulating the type of attachment that forms on teeth from which connective tissue attachment has been lost because of chronic periodontitis.(18, 19) The most common outcome is epithelial attachment rather than connective tissue attachment; this is because the biochemical composition of cementum is affected by disease and diseased cementum does not support adhesion and growth of fibroblasts.(20–22) We have shown that tooth cementum contains a collagenous attachment protein that promotes the attachment and spreading of mesenchymal cells.(23, 24) This protein, referred to as cementum attachment protein (CAP), appears to be restricted to cementum and it promotes adhesion and spreading of mesenchymal cells, not epithelial cells.(23, 25) It promotes the adhesion and spreading of some cell types better than others, whereas such differential activity is not manifested by FN.(26) Subpopulations of fibroblasts bind CAP with different capacities and cells binding with highest capacity form mineralized tissue in culture.(27)
To understand the molecular mechanisms involved in the action of CAP, we have studied the biochemical signaling mechanisms induced during attachment of human fibroblasts to CAP-containing surfaces. We previously showed that the CAP promotes tyrosine phosphorylation of signaling proteins such as focal adhesion kinase pp125FAK, assembly of cortical actin fibers, and activation of ERK-2 and c-fos messenger RNA (mRNA).(10) The pattern and degree of activation of these reactions were characteristic of CAP and different from FN. We have extended these observations to study whether adhesion to surfaces containing CAP as the only adhesion substrate can permit cells to escape G1 arrest and enter “S” phase and to determine the biochemical signaling events involved in this process.
MATERIALS AND METHODS
Mouse monoclonal antibodies against human cyclin D1 (HD11) and cyclin E (HE12 and HE111), rabbit polyclonal antibodies against Cdk inhibitors p21cip1/waf1 (p21; C-19) and p27kip1 (p27; N-20), and protein G agarose beads were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Horseradish peroxidase (HRP)-conjugated anti-mouse and anti-rabbit immunoglobulin G (IgG) antibodies were purchased from Amersham Corp. (Arlington Heights, IL, USA). Human FN, bovine serum albumin (BSA) fraction, poly-L-lysine (PL), human recombinant epidermal growth factor (EGF), and trypsin inhibitor were purchased from Sigma Chemical Co. (St. Louis, MO, USA). [3H]thymidine and [γ-32P]-adenosine triphosphate (ATP) were obtained from Dupont New England Nuclear (Boston, MA, USA).
Purification of CAP and cementum-derived growth factor
These proteins were purified from bovine cementum as described previously.(10, 24 28 29) Briefly, cementum tissue was scraped and harvested from bovine teeth, extracted in 1.0 M acetic acid containing proteinase inhibitors, and cementum-derived growth factor (CGF) was fractionated by heparin-affinity chromatography. Proteins eluted by 0.6 M NaCl were precipitated by 20% trichloroacetic acid, separated by high-performance liquid chromatography (HPLC) using TSK CM-5PW cation exchange column, and finally by reversed-phase HPLC using a C18 column and a gradient of 0-80% CH3CN in 0.1% trifluoroacetic acid (TFA). Each fraction was monitored for mitogenic activity by [3H]thymidine uptake using human gingival fibroblasts (HGFs) as target cells. CGF concentration in HPLC fractions was determined by absorbance at 280 nm, based on the absorbance of BSA standard separated under the same conditions.
CAP was obtained from guanidine-HCl extracts of cementum after the acid extraction.(10, 24) The guanidine extract was fractionated by diethylaminoethyl (DEAE)-cellulose chromatography, and proteins eluted by 0.5 M NaCl were then separated by reversed-phase HPLC using a C18 reversed-phase column with a gradient of 0-80% CH3CN in 0.1% TFA. Fractions were monitored for attachment activity by cell adhesion assay.(10, 24)
HGFs were grown and maintained in Dulbecco-modified Eagle's medium (DMEM) supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 100 mM sodium pyruvate, and 10% fetal bovine serum (FBS) as described previously.(28) Cells between transfer numbers 8 and 12 were used for the experiments.
Preparation of cell lysates
The preparation of plates coated with adhesive legends has been described previously.(10, 24) In brief, Costar nontissue culture plastic dishes (COSTAR Corp., Cambridge, MA, USA) were incubated with 5 μg/ml CAP or 2 mg/ml BSA for 2 h at 37°C and then blocked with 2 mg/ml of BSA. Before mitogenic assay, cells were serum-starved for 48 h in serum-free DMEM, trypsinized, washed twice, and added to plates. Incubations were done with or without growth factors. The growth factors used were 10 ng/ml CGF, 5 ng/ml EGF, 10 ng/ml CGF plus 5 ng/ml EGF, or10% FBS. We examined CGF because this growth factor is a major mitogen present in the local environment of cementum and in the presence of EGF its mitogenic activity is comparable with 10% FBS.(28, 29) Unattached cells were discarded. Attached cells were lysed with lysis buffer containing 10 mM Tris-HCl (pH 7.4), 1% Triton X-100, 150 mM NaCl, 5 mM EDTA, 50 mM NaF, 0.1 mM sodium orthovanadate, 0.5 mM phenylmethylsulfonyl fluoride, and 10 mM β-glycerophosphate, disrupted by sonication and then centrifuged 12,000 rpm for 15 minutes.(10) Protein concentration was determined by using a bicinchoninic acid (BCA) protein assay kit, (Pierce, Rockford, IL, USA).
Assay for DNA synthesis
DNA synthesis was evaluated in terms of [3H]thymidine incorporation. Cells, which had been serum-starved for 48 h, were trypsinized and plated at a cell density of 1.0 × 104/well in 48-well nontissue culture plates coated with the substrates specified and incubated with specified growth factors in serum-free medium containing 5 μCi/ml [3H]thymidine. Unattached cells were not removed for this experiment and incubation was continued. Because DNA synthesis in these cells begins after ∼18 h and lasts for 36 h, cells were harvested after 36 h and [3H]thymidine incorporation was measured by liquid scintillation counter as described previously.(28, 29)
The levels of cyclins and Cdk inhibitors were determined by immunoblotting 12 h after adding cells to plates; this time point was chosen for optimal levels of these proteins based on kinetics experiments done previously.(30) Equal amounts (5 μg) of total cell lysate from each treatment was loaded and separated by 12% (for cyclin D1 and cyclin E) or 15% (for p21 and p27) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions. Separated proteins were electrotransfered to nitrocellulose membranes (Immobilon-NC; Millipore Corp., Bedford, MA). The membranes were blocked with TBST (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.05% Tween 20) containing 5% nonfat dry milk (NFDM) for 2 h at room temperature. They were then probed with 0.1 μg/ml primary antibody in TBST containing 0.1% NFDM for 2 h, washed for 30 minutes with changing TBST, and then incubated with HRP-conjugated anti-mouse or anti-rabbit IgG antibody for 1 h at room temperature.(30) Bands were visualized using enhanced chemiluminescent system (ECL; Amersham). Exposed films were subjected to densitometry to compare the levels of cross-reacted protein bands in each lane. The Western analyses were repeated at least three times.
Measurement of histone H1 kinase activity
Cell lysates containing 100 μg of protein were precleared with protein G agarose beads for 2 h at 4°C and then incubated with equal amounts from antibodies to cyclin E and protein G agarose beads for 4 h at 4°C. Immunoprecipitated complex on the beads was washed three times with lysis buffer and twice with kinase buffer containing 50 mM HEPES (pH 7.5), 10 mM MgC12, 2.5 mM ethylene glycol-bis(β-amino ethyl ether)-N,N,N′,N′ tetracetic acid (EGTA), 10 mM β-glycerophosphate, 1 mM NaF, and 0.1 mM Na3VO4. Preparations containing equal amounts of cyclin E were then incubated at 30°C for 30 minutes in kinase buffer containing 1 mg of histone H1, 20 mM of ATP, and 10 μCi of [γ-32P]-ATP (specific activity, 6000 Ci/mmol). Reactions were stopped by boiling in SDS sample buffer and separated on 12.5% SDS-polyacrylamide gel. The gel was dried and exposed to X-ray film for autoradiography.(30) Each exposed film was subjected to densitometry and kinase activities of individual lanes were compared based on densitometry units.(30)
We first determined if adhesion to surfaces containing CAP as the only substrate is sufficient for DNA synthesis by human fibroblasts in the presence of growth factors. The cells were serum-starved for 48 h and added to nontissue cultures plates coated with CAP or BSA or to conventional tissue culture plates (positive controls), and DNA synthesis was determined as described in the Materials and Methods section. Plates coated with BSA, in which cells remained in suspension, served as negative controls. The results showed that cells adhering to CAP synthesized DNA in the presence of CGF and EGF or FBS, and the synthesis on CAP was comparable with positive controls (Fig. 1). [3H]thymidine uptake on CAP-coated plates also was comparable with plates coated with FN or type I collagen (data not shown). In contrast, thymidine uptake was minimal on BSA-coated plates even in the presence of mitogenic factors and it was similar to uncoated plates (Fig. 1). Adhesion to CAP alone was not sufficient for optimal DNA synthesis because [3H]thymidine uptake was lower in the absence of growth factors (Fig. 1). The CAP did not promote DNA synthesis when added in solution to monolayers (data not shown).
Accumulation of D1 and E cyclins is a crucial event necessary for crossing the restriction point and entering the S phase of the cell cycle.(1, 2, 31) We previously showed that the levels of D1 and E cyclins increase before DNA synthesis in HGFs(30); therefore, we examined if the increase in these cyclins requires adhesion. First, we added the cells to CAP-coated dishes, incubated in medium with and without growth factors, and examined D1 cyclin levels after 12 h by Western blotting. This incubation period was chosen because maximum levels of D1 levels were attained ∼12 h after adding growth factors.(30) The results showed that in cells adherent on CAP cyclin D1, levels were two times as much as 0 time controls in the absence of growth factors and in the presence of CGF + EGF and FBS (Fig. 2A, lanes 5-7). D1 levels did not increase in cells remaining in suspension on BSA-coated plates (the levels were 1.0, 0.9, and 1.0 times as much as 0 time controls, respectively, in the presence of no growth factors, CGF + EGF and FBS; Fig. 2A, lanes 2-4). In the next experiment, we compared cyclin E levels. The level of this cyclin in cells attaching to CAP was 1.2 times as much as 0 time control (Fig. 2B, lane 5 vs. lane 1), and the levels were higher in cells exposed to growth factors (2.2 and 2.2 times in presence of CGF + FBS, respectively; Fig. 2B, lanes 6 and 7). The cyclin E levels on BSA-coated plates were 1.1, 1.1, and 1.2 times as much as 0 time controls in the presence of no growth factors, CGF + EGF and FBS, respectively (Fig. 2B, lanes 2-4). We also examined D1 and E cyclin levels of monolayer cultures, which were exposed to 5 μg/ml CAP in solution; however, CAP in solution did not alter the levels of these cyclins (data not shown).
The cyclin E-dependent kinase Cdk-2 plays a crucial role in cell cycle progression and the activity of Cdk-2 has been shown to be adhesion dependent.(1, 2, 13, 14), (30, 31) Changes in cyclin E-dependent kinase activity also correlate with mitogenic activity.(30) Therefore, we examined whether adhesion to CAP-coated plates affects E-Cdk-2 activity. Lysates of cells incubated on CAP or BSA-coated plates with or without growth factors were immunoprecipitated by anticyclin E antibody and then histone H1 kinase activity of immunoprecipitates was determined. The results showed that H1 kinase activity increased 2.5 times in cells adhering to CAP alone relative to 0 time and BSA controls, and it increased 4.5-fold in the presence of CGF + EGF (Fig. 3). On BSA, the enzyme activity level in the absence and presence of growth factors was 1.0 and 1.1 times as much as 0 time controls, respectively.
The activities of Cdk-2 and Cdk-4 are regulated by two Cdk inhibitors p21 and p27.(31) The levels of p21 are affected by adhesion to ECM and adhesion-dependent regulation of p21 in turn regulates E-Cdk activity.(14, 32, 33) We therefore examined if the levels of these two inhibitors are affected in cells adhering to CAP. The p21 levels decreased relative to basal levels in adhered cells (Fig. 4A, lanes 1, 4, and 5), whereas they were similar to 0 time controls on BSA. In contrast, the p27 levels increased slightly in cells adhered to CAP (Fig. 4B).
Our results show that human fibroblasts attached onto surfaces containing CAP as the only adhesion substrate that can synthesize DNA, whereas cells remaining in suspension in plates coated with BSA do not. These data indicate that fibroblasts adhered on surfaces containing CAP as the only substrate can synthesize DNA efficiently. However, the synthesis requires engagement of integrins, presumably the α5β1 to which CAP binds,(34) because DNA synthesis was only 35% as much as positive controls on plates coated with PL to which cells adhere without integrins (data not shown). Adhesion alone is not sufficient, and growth factors are necessary because cells do not synthesize DNA in the absence of growth factors, and CAP added in solution did not promote [3H]thymidine uptake in monolayer cultures. These data indicate that in vivo tooth-root surfaces containing CAP as the matrix component is conducive for cell proliferation. The CAP does not permit attachment of epithelial cells; thus, the destruction and loss of CAP in tooth cementum may be one reason why it is difficult to form connective tissue on diseased tooth surfaces and epithelial attachment forms instead of connective tissue.(20, 22)
We previously showed that attachment to CAP induces immediate early G1 phase events.(10) The present studies show that further downstream events also are induced by adhesion to CAP. Thus, cyclin D1 levels increase in the cells adhered on CAP alone, even without growth factors. In contrast, cyclin D1 levels did not increase in cells suspended on plates coated with BSA even in the presence of growth factors. These results are consistent with others' observations showing that cyclin D1 expression is regulated by adhesion in the presence of growth factors(14, 35) and indicate that signal reactions generated by binding of CAP to CAP receptors induce cyclin D1 expression. Our results also showed that cyclin E levels were not affected by adhesion, but it increased in response to growth factors. However, cyclin E-Cdk-2 activity was activated by adhesion. These results are consistent with those of Fang et al.(13) and Zhu et al.(14) who observed that E-Cdk-2 activation by growth factors is adhesion dependent. Our data show that the cyclin E-Cdk-2 activity increases by adhesion alone in the fibroblasts even in the absence of growth factors. However, thymidine uptake was less in cells adherent on CAP in the absence of growth factors, indicating that adhesion to CAP alone and resultant increase in cyclin D1 and E-Cdk activity are not sufficient for optimal DNA synthesis.(30) Other investigators have reported that adhesion increases the levels of cyclin A in other cells(14, 15); however, we did not detect any significant increase in cyclin A levels because of adhesion to CAP (data not shown).
G1 to S phase transition and cell cycle progression depend on not only accumulation of cyclins and Cdk activity, but also on down-regulation of Cdk inhibitors, such as p21 and p27.(31, 32) The protein levels of p21 decreased in cells adhering to CAP, regardless of whether growth factors are present or not; this indicates that down-regulation of p21 is one, and not the only, factor contributing to cell cycle progression. However, p27 levels are not affected, and, previously, we showed that changes in p27 levels do not correlate with mitogenic activity.(30)
We have previously shown that attachment to CAP induces sustained activation of ERK-2, which lasts for up to 12 h.(10, 36) ERK-2 activation has been shown to be necessary for expression of AP-1 proteins and cyclin D1 expression.(11, 12, 37–39) Thus, it appears possible that attachment to CAP-containing surfaces induces cyclin D1 expression through sustained activation of ERK-2. The enhanced D1 expression requires sustained ERK-2 activation because transient activation by growth factors alone in suspended cells does not result in enhancement of D1 levels.(10, 28) These results are consistent with the recent report by Bottazzi et al. showing that in anchorage-dependent cells immediate ERK-2 activation leads to induction of p21 while sustained ERK-2 induction is important for cyclin D1 expression and control of E-Cdk-2 activity.(33)
Our results show that adhesion to CAP affects cell cycle progression through mechanisms that are common to other molecules. Nevertheless, differences occur in the type and degree of induction of these events.(10) Cells differing in capacity to bind CAP also differ in ability to form mineralized tissue in culture and produce CAP.(27, 40) These observations indicate that substances such as CAP, which are present in the local environment, could determine which cells are recruited and how they differentiate during normal homeostasis and wound healing and whether healing response is regeneration or repair. This is especially important in periodontal regeneration when regeneration requires new cementum formation and restoration of connective tissue attachment by the insertion of newly formed collagen fibers into cementum.(18, 20)
This work was supported by the National Institutes of Health grants DE-10491 and DE-08229.