INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

IN THE HEALTHY KIDNEY, mesangial cells in the glomerulus are generally in a quiescent state, nonproliferating, and removed from the cell cycle (29), with at most a few percent cycling at any given time (47). They maintain the surrounding mesangium with a balanced and regulated synthesis and turnover of extracellular matrix (30, 50). Mesangial cell proliferation and net accumulation of extracellular matrix in the mesangium are critical features in a number of human renal diseases characterized by glomerulosclerosis and end-stage renal failure (32, 54, 12). Mesangial cell proliferation alone is insufficient to account for the glomerular changes in disease and may not be prominent in many instances. Rather, the mesangial cell undergoes a phenotypic change marked by several features, including proliferative capacity and active secretion of extracellular matrix. In this phenotype, it has dedifferentiated into a myofibroblast, with characteristics of both smooth muscle cells (e.g., -actin production) and fibroblasts (e.g., synthesis of interstitial collagens) (15, 29).

The extracellular matrix is a key factor in maintaining the mesangial cell phenotype (16), and changes in its structure and composition that accompany changes in phenotype may themselves lead to further departure from the normal state. The mesangial cell proliferating in culture represents the myofibroblast phenotype and so has been studied extensively in attempts to understand mechanisms of cell turnover and matrix accumulation in the injured glomerulus. In the quiescent state in vivo, mesangial matrix contains a number of matrix glycoproteins including laminin, fibronectin, collagen types IV and V (34, 30, 37, 52), and probably the proteoglycans perlecan, biglycan, and decorin (6, 20, 19). In glomerulosclerosis, de novo synthesis of collagen type I (14, 13) and deposition of interstitial collagens (53, 2, 1) occurs. Collagen type I is a prominent matrix component synthesized by proliferating mesangial cells in culture (5, 45).

Mesangial cells have been cultured both on collagen type I-coated plates and in collagen gels. Growth on collagen gel diminished the synthesis of other extracellular matrix components compared with growth on laminin (27), suppressed their proliferation compared with growth on thin collagen films (25), and markedly decreased their mitogenic response to serum (41). Growth in three-dimensional collagen type I gels markedly decreased [³H]thymidine incorporation into DNA by mesangial cells, compared with growth on glass (67). Prolonged growth of mesangial cells in culture leads to the formation of nodules of extracellular matrix rich in collagens. Kitamura et al. (31) showed that mesangial cells associated with these nodules have a more differentiated phenotype than their counterparts not associated with the nodules, as indicated by lower mitogenesis, a reduced collagen type I-to-IV ratio, and decreased -smooth muscle actin expression.

We observed earlier that growth on collagen type I inhibited serum-dependent proliferation of quiescent mesangial cells while enhancing their contractility (41). An initial proliferative response of mesangial cells in some renal disease is followed by deposition of new matrix, and we suggested that deposition of collagen type I might represent an attempt at a reparative response to restore a more differentiated, nonproliferating phenotype. In light of our observations that mitogen-activated protein kinase (MAPK) (40, 58) and Ca²⁺/calmodulin-dependent protein kinase II (CaMK II) (39) are involved in the antiproliferative effects of heparin on mesangial cells, the present study was undertaken primarily to determine the role these pathways and additional kinase activities play in suppressing proliferation on collagen.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture. Rat mesangial cells were cultured from outgrowths of glomeruli harvested by sieving from kidneys of male Wistar rats and were characterized as previously described (57, 58). They were maintained in 10-cm petri dishes in a 5% CO₂ environment at 37°C, in RPMI 1640 medium with 10% FBS, penicillin, and streptomycin (all from GIBCO/BRL, Burlington, Ontario, Canada), passaged by trypsinization, and used between the fifth and fifteenth passages for all experiments reported here. To coat plates with collagen type 1, Vitrogen-100 (Celtrix Laboratories, Santa Clara, CA) was mixed 9:1 (vol/vol) with 10× RPMI 1640 on ice, and the pH was adjusted to 7.4 with 1 M NaOH. The solution was then transferred to culture vessels (0.5 ml/well in 12-well plates, 4 ml/plate in 10-cm petri dishes) and allowed to gel at 37°C. Human plasma fibronectin (Collaborative Biomedical Products, Bedford, MA) and poly-L-lysine (Sigma, St. Louis, MO) were dissolved in sterile water at 0.1 mg/ml and 1 mg/ml, respectively, spread on culture dishes (1.5 ml/10-cm petri dish, 75 µl/well in 24-well plates), and allowed to air dry in a sterile environment.

Quiescence was induced by exchanging medium on cells at 60-80% confluence (5 × 10⁴ cells/cm²) with medium containing 0.4% FBS followed by a 48-h incubation. In experiments to study the initiation of signaling, cells were then stimulated with either 5% NuSerum IV (Collaborative Biomedical Products) or 200 nM 12-O-tetradecanoylphorbol-13-acetate (TPA). Cell viability was checked by the thiazolyl blue [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H tetrazolium bromide (MTT); Sigma] assay modified from Hansen et al. (23). After being washed with serum-free medium, cells were incubated for 1 h at 37°C with medium containing 1 mg/ml MTT. Cell layers were then solubilized in DMSO, and the optical density was recorded at 570 nm. DNA content was measured by suspending sonicated cell homogenates at 1:10 dilution in a solution of 1 µg/ml Hoechst 33258 (Sigma) in phosphate-saline buffer and comparing the resulting fluorescence (_ex = 350, _em = 455) with a standard curve constructed with calf thymus DNA.

RNA isolation and Northern blotting. Total RNA was isolated using TRIzol Reagent, as described by Chomczynski and Mackey (9). Equal amounts of RNA (~10 µg) were denatured by the method of Gong (17), separated by electrophoresis on agarose-formaldehyde gels, and transferred to Hybond-N nylon membrane (Amersham, Oakville, Ontario, Canada) for hybridization with a c-fos cDNA that was labeled with [-³²P]dCTP. Levels of mRNA were quantitated by densitometry of the Northern blot autoradiographs and normalized to 18S rRNA after probing with labeled cDNA to rat 18S rRNA. ³²P labeling of the probes was carried out with a random primer DNA labeling kit from Boehringer-Mannheim (Laval, Quebec, Canada). The rat c-fos cDNA was cloned by T. Curran (11), and a cDNA for mouse 18S rRNA was obtained from J. Koropatnick (University of Western Ontario).

MAPK activity. Cells were washed with ice-cold PBS and scraped into 800 µl of lysis buffer containing 50 mM Tris · HCl (pH 7.4), 1% (vol/vol) Nonidet P-40, 0.25% (wt/vol) sodium deoxycholate, 150 mM NaCl, 5 mM EGTA, 1 mM EDTA, protease inhibitors [1 mM phenylmethylsulfonyl fluoride (PMSF), 1 µg/ml aprotinin, and 1 µg/ml leupeptin], and phosphatase inhibitors (1 mM Na₃VO₄ and 1 mM NaF). Cells were sonicated and centrifuged at 100,000 g for 15 min. Cytosol was assayed for protein, and aliquots containing ~500 µg of protein were precleared by adding 1.0 µg normal rabbit IgG with 20 µl of protein A-Sepharose and centrifuging. The supernatant was incubated with 2 µg of polyclonal rabbit anti-rat Erk-2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) for 3 h at 4°C, and immunoprecipitates were recovered by incubating with a 50% slurry of protein A-Sepharose for a further 2 h. Portions of the immunoprecipitates were subjected to 12% SDS-polyacrylamide gel electrophoresis according to Laemmli (35) and transferred to polyvinylidene difluoride membranes for Western blotting with anti-Erk-2 as described below. MAPK activity was determined by phosphorylation of the specific substrate myelin basic protein (MBP) (3). Immunoprecipitates were mixed with 20 mM HEPES buffer, pH 7.4, containing 10 mM MgCl₂, 2 mM MnCl₂, 0.5 mM EGTA, 10 mM NaF, 0.5 mM Na₃VO₄, 1 mM dithiothreitol, 0.5 mg/ml MBP, 2 µM AMP-dependent protein kinase inhibitor, 50 µM ATP, and 5 µCi of [-³²P]ATP, and incubated at 30°C for 20 min. Reaction mixtures were mixed with 2× sample buffer for electrophoresis according to Laemmli (35) and separated on 16% SDS-polyacrylamide gels for silver staining and autoradiography.

CaMK II activity. Cells were washed twice with ice-cold PBS and lysed by several freeze-thaw cycles in buffer containing 50 mM HEPES, 50 mM sodium pyrophosphate, 1 mM EGTA, 25 mM NaF, 1 mM Na₃VO₄, 1 mM dithiothreitol, 0.5% (vol/vol) Nonidet P-40, aprotinin (1 µg/ml), leupeptin (1 µg/ml), pepstatin (1 µg/ml), and 1 mM PMSF. After brief sonication (3 times for 5 s each) and centrifugation (10 min at 17,000 g), the protein content of the supernatant was determined by the method of Peterson (48). Fractions of the supernatant containing 5 µg protein were incubated with a 10-fold volume of CaMK II assay buffers at 30°C for 3 min. The assay buffer for autonomous activity was 50 mM HEPES, pH 7.5, containing 10 mM MgCl₂, 0.1 mM ATP, 10 µM autocamtide-2, 1 mM EGTA, and [-³²P]ATP (5 µCi/ml). For total CaMK II activity, EGTA was replaced with 3 mM CaCl₂ and 1 mM calmodulin.

Protein kinase C activity. Cells in 24-well plates were washed with medium and incubated for 10 min at 30°C in 100 µl of a solution of 137 mM NaCl, 5.4 mM KCl, 10 mM MgCl₂, 0.3 mM Na₂HPO₄, 0.4 mM KH₂PO₄, 25 mM -glycerophosphate, 5.5 mM D-glucose, 5 mM EGTA, 1 mM CaCl₂, 20 mM HEPES, 100 µM ATP, and 50 µg/ml digitonin, pH 7.2, containing 120 µg/ml of the peptide VRKRTLRRL and 10 µCi/ml [-³²P]ATP, as described (62). This peptide represents the protein kinase C (PKC) phosphorylation site of the epidermal growth factor (EGF) receptor and is a substrate for the conventional isoforms of PKC. Reaction was terminated by adding 50 µl of ice-cold 30% (wt/vol) trichloroacetic acid on ice and spotted onto phosphocellulose circles prewashed sequentially with water, buffer, and 75 mM phosphoric acid. After standing for 15 min at room temperature, the circles were washed (4 min each with gentle shaking) three times in 75 mM phosphoric acid and once in 2.75 mM sodium phosphate, pH 7.5, before liquid scintillation counting.

Immunoblotting. Cells were lysed in the same lysis buffer described above for preparation of lysates for MAPK activity measurements. Lysates were subjected to electrophoresis on 12% SDS-polyacrylamide gels according to Laemmli (35), and proteins were transferred to polyvinylidene difluoride membranes in 25 mM Tris and 192 mM glycine (pH 8.3) containing 15% methanol. The membranes were blocked with 5% bovine serum albumin and 5% Carnation milk in 30 mM Tris · HCl (pH 7.4) containing 137 mM NaCl, 2.6 mM KCl, and 0.05% Tween 20. Membranes were then probed with either anti-Erk-2 antibody, a polyclonal rabbit anti-mouse focal adhesion kinase (FAK) antibody (Santa Cruz Biotechnology), or monoclonal mouse anti-phosphotyrosine antibody PY20 (Transduction Laboratories, Lexington, KY), and immunoreactive bands were detected with an Amersham enhanced chemiluminescence detection system. For immunodepletion of MAPK or FAK, aliquots of lysate containing 500 µg of protein were precleared with 1 µg of normal rabbit IgG and 20 µl of protein A-Sepharose and incubated with 2 µg of anti-Erk-2 or anti-FAK antibodies for 3 h at 4°C. Immunoprecipitates were recovered by incubation for a further 2 h with a 50% slurry of protein A-Sepharose, and the supernatants were subjected to electrophoresis and Western blotting with antiphosphotyrosine.

Alternatively, for immunoblotting of PKC isoforms, quiescent cells were harvested in 20 mM Tris · HCl (pH 7.5) containing 2 mM EDTA, 2 mM EGTA, 6 mM -mercaptoethanol, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 5 µM pepstatin, and 1 mM PMSF. Lysates were electrophoresed and transferred to polyvinylidene difluoride membranes as above, and probed with either anti-PKC- or anti-PKC- antibodies (both from Sigma).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

We showed previously that quiescent mesangial cells grown on a collagen type I gel responded to serum stimulation with decreased [³H]thymidine incorporation, compared with cells grown on plastic (41). This was not due to decreased cell viability. There were no significant differences in trypan blue exclusion, color development with MTT, or cell number as indicated by DNA content per well, between cultures on plastic and collagen gels (Table 1). Furthermore, cells on collagen showed a higher cytosolic Ca²⁺ signal and more rapid myosin light chain phosphorylation than cells on plastic (41). Therefore, we undertook to determine which signaling pathways were affected by collagen type I that might account for decreased mitogenic responsiveness.

Induction of the protooncogene c-fos is an early indicator of entry of cells into the cell cycle, and c-fos mRNA levels are maximal at 30-60 min after stimulation of quiescent mesangial cells with a number of mitogenic stimuli, including serum, phorbol ester, and Ca²⁺ ionophores (40). Cells growing on collagen likewise showed a transient increase in c-fos mRNA in response to serum, but the magnitude of the response was greatly diminished compared with cells grown on plastic (Fig. 1). In additional experiments, we further compared c-fos mRNA levels 30 min after serum stimulation in cells plated on an inert substrate (poly-L-lysine) and on fibronectin, a matrix that supports attachment and proliferation of mesangial cells (Fig. 2). Induction of c-fos was comparable on plastic and fibronectin. Again the response was diminished on collagen. As with collagen, cells grown on poly-L-lysine responded with only low levels of c-fos induction.

View larger version (39K): [in this window] [in a new window]   Fig. 1.   Effect of collagen type I on c-fos mRNA in mesangial cells. Rat mesangial cells growing on plastic petri dishes or dishes coated with collagen type I were made quiescent by growth for 48 h in the presence of 0.5% serum. Cells were then treated with 5% NuSerum, and total RNA was extracted at the indicated time points. Control cells at time 0 were not treated with NuSerum. Level of c-fos mRNA was determined by Northern blotting and autoradiography. Intensity of the signal representing hybridization of the c-fos cDNA was normalized to the intensity of hybridized 18S rRNA probe, determined by laser densitometry. Histogram shows the ratio of the c-fos signal relative to the 18S signal in arbitrary units. The experiment was repeated once with similar results.

View larger version (41K): [in this window] [in a new window]   Fig. 2.   Matrix effects on c-fos mRNA. Quiescent mesangial cells were stimulated at time 0 with 5% NuSerum, and total RNA was collected immediately or after 30 min. A typical Northern blot probed with cDNAs to c-fos and 18S rRNA is shown, and the histogram shows the mean ratio of signal intensities ± SD from 3 separate experiments. Cells were grown on either plastic (PL), fibronectin (FN), collagen type I (COL), or poly-L-lysine (PLY).

Several pathways are known to be involved in c-fos induction, with one acting through the Erk family of MAPK and requiring a cytoplasmic Ca²⁺ signal, another acting through a cAMP response element and requiring intranuclear Ca²⁺ (24), and a third involving an upstream sis-inducible element (26). The cAMP response element is also activated by CaMK (39). Because we showed previously that Erk and CaMK are involved in Ca²⁺-dependent signaling in rat mesangial cells (61) and that collagen enhances Ca²⁺ signaling (41), we examined the effects of growth on collagen on activation of these pathways. When quiescent mesangial cells on plastic were stimulated with serum, an increase in MAPK activity occurred beginning at 1-2 min that was maximal at 5 min and was sustained for at least 1 h (Fig. 3A). This increase was sensitive to pretreatment with the MAPK kinases (MEK1/2) inhibitor, PD-98059. This response to serum was greatly diminished in quiescent cells grown on collagen, compared with the response on plastic (Fig. 3B). The slight increase in MAPK activity was again sensitive to PD-98059, indicating that what little activation does occur on collagen is still mediated through the serum factor-to-MEK signaling pathway. The total amount of MAPK protein detected by Western blotting was the same in cells grown on collagen or plastic.

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Effect of collagen type I on serum-dependent mitogen-activated protein kinase (MAPK) activity. Cells were grown on plastic or collagen type I-coated petri dishes. Quiescent cells were treated with 5% NuSerum, and cytosolic extracts were obtained from cells on plastic by scraping them into lysis buffer as described in MATERIALS AND METHODS, at the indicated times. Cells grown on collagen were released first by a brief exposure to collagenase A before transfer to lysis buffer. Cytosolic extracts were immunoprecipitated with anti-Erk-2 antibody, and portions of immunoprecipitates were used either for Western blotting or for in vitro kinase assay. A: mesangial cells grown on plastic. Top autoradiogram shows a Western blot of the immunoprecipitate with a major band corresponding to Erk-2 (42 kDa); bottom autoradiogram shows the phosphorylation of myelin basic protein (MBP) by immunoprecipitates in the presence of [-32P]ATP. Prominent MBP band was visualized by autoradiography and was also demonstrated on silver-stained gels of the starting material (not shown). Histogram shows the relative intensity of the MBP band determined by densitometry, in arbitrary units.  Presence of the MEK inhibitor PD-98059 (50 µM) 30 min prior to and during 2 min of serum stimulation. First lane (RM) contains the kinase assay reaction mixture alone. Second lane (PA) represents protein A-Sepharose treatment of cytosolic extract in absence of antibody, followed by elution and subsequent use in the kinase assay. Last lane (PAB) is as for PA except that cytosolic extract was replaced by lysis buffer and antibody was included. B: mesangial cells grown on collagen type I are compared with those stimulated on plastic for 5 min. Top and bottom autoradiograms show Western blotting with anti-Erk-2 and MBP phosphorylation, respectively, and the histogram shows MBP signal intensity, as in A. , RM, PA, and PAB are same as in A. Autoradiograms are representative of 2 similar experiments.

Stimulation of quiescent mesangial cells with 5% NuSerum as in the above experiments does not activate PKC, nor is the response to serum influenced by downregulation of PKC (40). Therefore, exposure of the cells to phorbol esters provides an additional pathway for activation of MEK/MAPK in these cells. We therefore examined whether growth on collagen could also suppress MAPK activation by TPA. Similar results to those obtained with serum were observed; the PD-98059-sensitive activation of MAPK was greatly diminished on collagen compared with plastic, but the total MAPK protein was unaffected (Fig. 4).

View larger version (36K): [in this window] [in a new window]   Fig. 4.   Effect of collagen type I on 12-O-tetradecanoylphorbol 13-acetate (TPA)-dependent MAPK activity. Experiments of Fig. 2 were repeated in an identical manner except that 200 nM TPA was used as the stimulus instead of NuSerum. Autoradiograms, histograms, and all symbols are exactly as in Fig. 3. Again, the experiment was repeated twice with similar results.

To determine whether the effect of collagen on TPA-dependent MAPK activation was due to an effect at the level of PKC, or possibly at or below convergence of the serum and PKC signaling pathways at Raf-1 (51), we next measured PKC activity directly with the EGF receptor peptide as a specific substrate. On plastic, TPA caused a marked elevation in PKC activity in quiescent cells by 1 min that was abolished when PKC was downregulated by 24 h pretreatment with 200 nM TPA (Fig. 5A). On collagen, only a minimal increase above basal activity was observed. However, the slight increase was again abolished by pretreatment with TPA, suggesting that collagen did not interfere with availability of the phorbol ester. These results indicate that PKC is either decreased or not activable in these cells grown on collagen. Five minutes after TPA treatment, cells grown on fibronectin showed PKC activity comparable to cells grown on plastic, and this response was abolished by prolonged pretreatment (Table 2). In contrast, TPA caused no significant increase in PKC activity at 5 min in cells grown on poly-L-lysine.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Effect of collagen type I on protein kinase C (PKC) activity. PKC activity was determined as described in MATERIALS AND METHODS and is expressed as nmol 32P incorporated into peptide substrate from [-32P]ATP, per min per µg DNA. A: mesangial cells were plated onto plastic or collagen type I-coated 24-well plates and rendered quiescent. Cells were treated with TPA (200 nM) for the indicated times, and then the phosphorylation reaction was carried out for 10 min;  24-h pretreatment with TPA to downregulate PKC activity. Inset: time course of PKC activation on plastic. Values are means ± SD from 3 independent experiments. B: time course of PKC activity after cell attachment onto plastic () and collagen (). Cells were detached from confluent 10-cm petri dishes by trypsinization and passaged in a split ratio of 1:3 onto plastic or collagen type I-coated 24-well plates. PKC activity was determined at the indicated times after cell attachment as described in A. Level of PKC activity was corrected for the amount of DNA. Values at time 0 were determined from suspended cells prior to plating. Values are means of duplicate wells.

From the above results, we considered that collagen could be causing a time-dependent downregulation of PKC, during initial plating and the subsequent 48-h period of serum deprivation required to render cells quiescent. Therefore, we examined the time course of PKC activity in cells passaged onto either substrate, expecting to see a time-dependent loss of PKC activity on collagen. Cells were grown to confluence on plastic in 10% FBS, passaged by trypsinization onto either plastic or collagen type I, and allowed to attach and proliferate in the continued presence of 10% FBS. Surprisingly, PKC activity was nearly completely lost in the trypsinized cells in suspension, but rapidly returned when cells were plated on plastic, plateauing around 6 h later (Fig. 5B). When plated on collagen, only minimal activity returned, even after 48 h. This was not due to a difference in attachment or proliferation on collagen, with attachment (41) and cell number (e.g., as indicated by DNA content, Table 1) being similar on both substrates. Therefore, collagen does not cause a time-dependent loss of PKC activity, but rather suppresses the restoration of activity lost during trypsinization and detachment.

To determine whether the decreased activity of PKC on collagen was due to decreased levels of the protein, the Ca²⁺- and diacylglycerol-dependent isoform PKC- was examined by Western blotting (Fig. 6). The protein was decreased by ~65-75% in cells grown on collagen or poly-L-lysine, compared with cells grown on plastic or fibronectin. Furthermore, the atypical isoform PKC-, whose activity would not be detected in our assay using the EGF receptor-based substrate, was also decreased ~60-70%, suggesting a general decrease in PKC isoforms on collagen.

View larger version (36K): [in this window] [in a new window]   Fig. 6.   Western blotting of PKC isoforms. Western blots were prepared from cells grown on either plastic (PL), fibronectin (FN), collagen type I (COL), or poly-L-lysine (PLY), probed with antibodies to either PKC- or PKC-, and visualized with enhanced chemiluminescence detection. Control lane (C) contains PKC isoforms from a rat brain lysate (Transduction Laboratories). Twenty micrograms total protein were loaded in each lane. Presence of 3 bands in the control lane for PKC- is consistent with the manufacturer's (Sigma) specifications for the antibody using a rat brain preparation, with the major band at 78 kDa running as a doublet, and the minor band at ~50 kDa thought to represent a degradation product occurring in brain.

Activation of CaMK II signaling was likewise suppressed when cells were grown on collagen. We previously showed that increasing intracellular Ca²⁺ concentration by treatment with 100 nM ionomycin, a concentration of ionophore that gives a transient rise in intracellular Ca²⁺ concentration with a subsequent return to basal levels after several minutes, is accompanied by an increase in autonomous CaMK II activity (see MATERIALS AND METHODS) that is maximal between 15 and 45 s after addition of ionomycin (39). Total CaMK II activity assayed in the presence of excess Ca²⁺ and calmodulin is constant. In the present series of experiments, ionomycin caused an increase of ~10-fold in autonomous activity at 30 s in cells on plastic, which was inhibited by the CaMK II-specific inhibitor, KN-93 (Fig. 7). Growth on collagen completely abolished activation by ionomycin (Fig. 7), although it did not affect the uptake of Ca²⁺ by the cells (not shown; see also Ref. 41). Collagen did not affect the expression of CaMK II, however, as demonstrated by a nearly constant level of total CaMK II activity under all experimental conditions.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Effect of collagen type I on ionomycin-dependent Ca2+/calmodulin-dependent protein kinase II (CaMK II) activation. Mesangial cells were seeded onto plastic or collagen type I-coated petri dishes and rendered quiescent, and cytosolic extracts were obtained after treatment with ionomycin (100 nM) for the indicated periods of time, as described in Fig. 2. CaMK II activity was determined as the ability of the cytosolic extracts to phosphorylate autocamtide-2 in presence of [-32P]ATP, as described in MATERIALS AND METHODS. Autonomous CaMK II activity (A) was determined in the absence of free Ca2+, and calmodulin and total activities (B) were determined in the presence of added Ca2+ and calmodulin. Both are expressed as nmol 32P incorporated into autocamtide-2 per min per mg cell protein.  Autonomous and total CaMK II activities were also determined in cells pretreated with KN-93, a specific CaMK inhibitor. Values are the range from 2 independent experiments.

We next considered whether the effect of collagen was specific to Ser/Thr kinases like CaMK II and PKC or involved kinases more generally. Because serum factors activate the Ras pathway through protein tyrosine kinase receptors and subsequently activate MAPK through the dual specificity MEK, extracts of cells rendered quiescent on plastic were stimulated with serum and subjected to Western blotting with an antiphosphotyrosine antibody (Fig. 8A). Four major bands at ~42, 86, 135, and 210 kDa were detected. Immunodepletion of the extract with anti-Erk-2 antibody demonstrated the lower band was actually p42 MAPK (Fig. 8B). Immunodepletion with anti-FAK antibody revealed the band at 135 kDa to consist of two bands, the lower of which was removed by anti-FAK. The other bands were not identified. Phosphorylated MAPK was only detected after 2 min, appeared maximally activated at about 5 min, and remained increased at 1 h. The other bands all showed basal phosphorylation that increased after serum stimulation and remained elevated. Genistein completely inhibited phosphorylation of MAPK and the 210-kDa band and diminished phosphorylation of the other bands. When cells were plated on collagen type I, no basal tyrosine phosphorylation was detected (Fig. 9); nor was phosphorylation of either MAPK or the 210-kDa band visible on gels loaded with amounts of protein that gave prominent signals from cells grown on plastic. Phosphorylation of the 86-kDa band, now resolved into a doublet, was maximal at 5 min and undetectable by 30 min, and another band in this region at ~65 kDa appeared only transiently at 20 min. The signal centered on 135 kDa appeared to be missing the lower FAK component. Furthermore, no immunodepletion with anti-FAK was observed (not shown). Therefore, growth on collagen suppresses serum-dependent phosphorylation of MAPK, FAK, and an unidentified protein at 210 kDa and decreases and delays phosphorylation of other major phosphorylated proteins in mesangial cells.

View larger version (55K): [in this window] [in a new window]   Fig. 8.   Time course of serum-dependent tyrosine phosphorylation. A: quiescent mesangial cells grown on plastic were treated with 5% NuSerum for the indicated times, and total cell extracts were analyzed by Western blotting with an anti-phosphotyrosine antibody as described in MATERIALS AND METHODS.  Some cultures were pretreated with 50 µg/ml genistein for 30 min prior to activation. Arrows show the positions of molecular mass markers. B: a portion of the sample from 5-min serum exposure was electrophoresed on a separate gel alone (lane C) or after immunodepletion with anti-FAK antibody (lane F) or anti-Erk-2 antibody (lane E), then analyzed by Western blotting with an anti-phosphotyrosine antibody. Blots are representative of 3 separate experiments each. FAK, p125 focal adhesion kinase.

View larger version (43K): [in this window] [in a new window]   Fig. 9.   Effect of collagen type I on serum-dependent tyrosine phosphorylation (representative of 3 separate experiments). Quiescent mesangial cells grown on collagen-coated petri dishes were treated with 5% NuSerum for the indicated times, detached as described in the legend to Fig. 2, and subjected to Western blotting with an anti-phosphotyrosine antibody as described in Fig. 6. An extract from cells grown on plastic and treated with NuSerum for 5 min is included immediately to the left of the collagen lanes for comparison. Arrows on right show the positions of molecular mass markers. Tyrosine-phosphorylated epidermal growth factor receptor (EGFR), supplied with the commercial antibody and included as a positive control, is run in the far left lane and migrates at the position indicated by EGFR.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Several previous studies have demonstrated that collagen type I gels suppress the mitogenic responsiveness and proliferation of cultured renal mesangial cells (67, 25, 41). The present results indicate that this may be due to inactivation of several major kinase signaling cascades when the cells are attached to collagen. Thus collagen inhibited restoration of PKC activity after plating trypsinized cells, inhibited MAPK activation and MAPK-dependent c-fos induction through PKC-independent pathways, prevented the autophosphorylation of CaMK II, suppressed tyrosine phosphorylation of MAPK and FAK, and decreased and delayed serum-dependent phosphorylation of several unidentified tyrosine kinase substrates. Induction of c-fos by stimuli such as serum factors and phorbol esters operates through MAPK and is sensitive to inhibition of MEK. Activation of MAPK leads to phosphorylation of the transcription factor Elk, which drives transcription of c-fos through a serum response element in the c-fos promoter (49, 24). A second region of the c-fos promoter contains a cAMP response element, placing activation of transcription additionally under control of cAMP and the cAMP response element-binding protein, CREB, independent of MAPK/Elk (49). The mechanism of c-fos induction by CaMK II remains somewhat controversial. For instance, whereas both constitutively active CaMK II and CaMK IV induce c-fos in some cell lines (4), it has also been shown that CaMK II, but not CaMK IV, can phosphorylate CREB on a serine residue that prevents its activity as a transcription factor (55). The CaMKs may also phosphorylate the serum response factor (42), a factor that binds with Elk to the serum response element. Nevertheless, overexpression of CaMK II induced c-fos in cultured rat mesangial cells lacking CaMK IV (59). That the inhibition of both MAPK-dependent and CaMK II-dependent c-fos induction by heparin correlates with the antiproliferative effect of heparin on mesangial cells (40, 58, 39) suggests that suppression of these two pathways by growth on collagen is sufficient to account for the antimitogenic effects of this substratum.

Although the effects of collagen involve a range of kinase signals, they are not universal. Importantly, Ca²⁺-dependent contraction is not inhibited and is in fact enhanced by this substrate (41). This contraction is accompanied by myosin light chain phosphorylation, which also is unaffected by collagen. Therefore, collagen does not affect the myosin light chain kinase. Since myosin light chain kinase is itself a CaMK (7), this suggests that collagen does not affect the activation of calmodulin by Ca²⁺ but shows selectivity in inhibiting Ca²⁺/calmodulin targets. Preservation of contractility with suppression of the proliferative phenotype may suggest a contribution by collagen to maintaining the differentiated mesangial cell phenotype.

Several mechanisms have been suggested for the growth suppression of mesangial cells by collagen. Marx et al. (38) demonstrated a marked downregulation of platelet-derived growth factor -receptors in mesangial cells in three-dimensional collagen gels. However, no such downregulation was observed on two-dimensional collagen type I surfaces (38). Transforming growth factor- is a growth suppressant for mesangial cells, and its accumulation by binding to collagen in nodular cultures has been invoked as a possible explanation for the more quiescent, differentiated phenotype of nodule-associated cells (31). However, such a time-dependent accumulation would not seem to account for very early effects of collagen, such as failure of cells to restore PKC shortly after plating. Attachment-mediated signaling is perhaps the most appealing mechanism for substratum specificity at present.

Integrin-mediated attachment of cells to fibronectin, and probably to secreted matrix proteins in cultures on plastic, is associated with focal adhesion assembly and mitogenic signaling (8, 66). On the other hand, attachment to poly-L-lysine does not involve integrins and does not allow spreading or activation of signaling responses (56, 46, 60) and, in the present studies, does not support PKC activation or c-fos induction. The situation with collagen is less clear. Mesangial cells express integrins of the ₁ and ₃ classes (16), and the organization of specific subtypes is dependent on adhesion to specific matrices. On collagen type I, mesangial cells organize integrins ₁₁, ₂₁, and ₃₁ into focal adhesions (16). Therefore, focal adhesion formation represents a way in which substrate-specific signals can arise. For instance, FAK can become phosphorylated upon integrin clustering (16), and the absence of FAK phosphorylation on collagen may be a clue to changes in integrin-based structures. However, in mesangial cells, phosphorylation of FAK appears also to involve PKC (22), so the decrease in both conventional and atypical PKC isoforms and absence of conventional PKC activity may account for the observed lack of FAK phosphorylation. Decreased FAK phosphorylation may lead to decreased MAPK activation, although recently ₁-integrin-mediated activation of MAPK has been demonstrated to be independent of FAK in fibroblasts (36). Further suggesting involvement of integrin species, cell cycle arrest of vascular smooth muscle cells on polymerized collagen type I has been attributed to upregulation of the cyclin-dependent kinase-2 inhibitors p27^Kip and p21^Cip/Waf (33). An integrin-mediated signal is implicated because cell cycle arrest and inhibitor upregulation could be mimicked in control cells by anti-₂ integrin subunit antibodies. Growth on collagen type I has also been shown to increase cAMP levels (18), which can decrease MAPK activation (21). Thus, at present the mechanisms of mitogenic suppression of mesangial cells by collagen type I are unclear but probably involve a complex interplay of integrin signaling, PKC, and MAPK that may be further complicated by cross talk with other kinase cascades.

The lack of PKC activity on collagen was itself somewhat surprising: collagen type I increased membrane-bound PKC in osteoblasts (18), activated the PKC- isoform in fibroblasts (65), and triggered translocation of PKC- in HeLa cells (10). However, it clearly suppressed PKC activity in mesangial cells as measured with an EGF receptor peptide as substrate in the present studies. This substrate will not detect novel or atypical isoforms of PKC, and so their participation is not ruled out by the present studies. However, levels of the atypical isoform PKC- were also found to be decreased on collagen. Recently, it has been shown that syndecans are required for the localization and activation of PKC (44, 43). Furthermore, fibroblasts attach to and migrate upon plates coated with the cell binding domain of fibronectin but require addition of the heparin (syndecan) binding domain to form focal adhesions (64), a process accompanied by FAK activation and requiring PKC (63). The cell surface syndecans may be cleaved by trypsinization when cells are passaged and indeed are shed even when cells are detached from culture plates by lowering the Ca²⁺ concentration of the medium (28). Since syndecans are primarily involved in attachment to the heparin-binding domain of fibronectin, their expression may be disfavored on a pure collagen substrate, and this may provide a mechanism for the lack of reactivation of PKC on attachment to collagen.