Application of cyclic strain to bladder smooth muscle (SM) cells results in profound alterations of the histomorphometry, phenotype, and function of the cells. The onset of this process is characterized by the activation of a cascade of signaling events coupled to progressive and, perhaps, interdependent changes of gene expression. In particular, externally applied cyclic stretch to cultured bladder SM cells results in the transient expression of the Cyr61 gene that encodes a cysteine-rich heparin-binding protein originally described as a proangiogenic factor capable of altering the gene programs for angiogenesis, adhesion, and extracellular matrix synthesis. In this study, we investigated the effects of mechanical stretch-induced Cyr61 on the expression of potential mechanosensitive Cyr61 target genes and the signaling pathways involved. We showed that suppression of Cyr61 expression with an adenoviral vector encoding an antisense oligonucleotide reduced mechanical strain-induced VEGF, αv-integrin, and SM α-actin gene expression but had no effect on the myosin heavy chain isoforms SM-1 and SM-2. Signaling pathways involving RhoA GTPase, phosphatidyl inositol 3-kinase, and cytoskeletal actin dynamics altered stretch-induced Cyr61 and Cyr61 target genes. Reciprocally, adenovirus-mediated overexpression of Cyr61 in cells cultured under static conditions increased the expression of VEGF, αv-integrin, and SM α-actin, as well as that of SM-1 and SM-2 isoforms, suggesting that the effects of a sustained expression of Cyr61 extend to SM specific contractile function. These effects were dependent on integrity of the actin cytoskeleton. Together, these results indicate that Cyr61 is an important determinant of the genetic reprogramming that occurs in mechanically challenged cells.
- mechanical stretch
- signal transduction
- actin cytoskeleton
mechanical forces are well recognized as major regulatory factors of normal tissue homeostasis and important determinants of tissue fate during development and pathological conditions. Their biological effects are particularly relevant to smooth muscle (SM)-rich tissues such as the bladder wall and the cardiovascular system, which must withstand continuous fluctuations in strain and periodic stretching and adjustment of muscle tonicity. Perturbations of the force equilibrium within these tissues occur often as a result of obstructive diseases or pressure overload triggering pathophysiological changes, i.e., hypertrophy, hyperplasia, and various degrees of fibrosis. The signals orchestrating these changes are complex and may involve selective alterations of the expression of molecular intermediates whose encoded proteins act locally to orchestrate broader molecular and cellular changes.
Recent in vitro studies indicated that, in responding to mechanical stimuli, the cells integrate several different types of information and initiate a cascade of events coupled to progressive and, perhaps, interdependent changes of gene expression and/or protein activity (13, 17). In particular, cellular mechanotransduction pathways induce the rapid and transient expression of a number of transcripts referred to as immediate early (IE) responsive genes, many of which encode regulatory molecules typified by transcription factors, signaling molecules, growth factors, and cytoskeletal proteins (24, 25, 30). However, although the response to mechanical or chemical stimuli of many IE genes of the transcription factor family has been extensively studied, relatively little is known about the mechanical regulation and function of several nontranscription factor types of genes. In this study, we report evidence of the role of one such IE gene, Cyr61, in the mechanotransduction process.
The cysteine-rich protein 61 (Cyr61) is the encoded product of an IE gene, whose expression is rapidly and transiently induced in response to externally applied cyclic mechanical strain to SM cells (33). Mechanical induction of the Cyr61 gene, which occurs with kinetics similar to those of the transcription factor Egr-1, was sensitive to cytoskeletal actin dynamics and required signaling through RhoA GTPase and phosphatidylinositol (PI) 3-kinase. Structurally, Cyr61 protein contains a highly conserved cysteine-rich primary sequence organized into discrete domains with partial identities to insulin-like growth factor binding protein, von Willebrand factor, thrombospondin, and a COOH-terminal cysteine knot domain (5, 28). Functionally, the recombinant Cyr61 protein acts as a ligand for heparin sulfate proteoglycan and various integrins and mediates cell adhesion, migration, and proliferation in fibroblasts and endothelial cells (11, 20, 21). Through these activities, Cyr61 was suggested to regulate inflammation, angiogenesis, tissue regeneration, and tissue development. However, information detailing the actual function of the endogenously produced Cyr61 is still scanty. Targeted disruption of the Cyr61 gene by homologous recombination in mice led to an early embryonic lethality as a result of placental insufficiency and compromised vessel integrity, suggesting that disruption of the Cyr61 gene may affect the expression of other genes involved in angiogenesis, extracellular matrix (ECM) remodeling, and cell-ECM interactions (26).
In this study, we sought to determine the functional significance of Cyr61 expression in bladder SM cells and further define the role of this protein in the phenotypic modulation of the cells dictated by mechanical strain. In particular, two approaches were used to determine whether Cyr61 regulates selectively the expression of secondary responsive genes: 1) a loss-of-function approach by suppressing the expression of Cyr61 in mechanically stimulated cells and analyzing the impact on the expression of potential Cyr61 target genes; and 2) a gain-of-function approach by overexpressing Cyr61 in cells cultured under static conditions and analyzing the expression profile of Cyr61 target genes and the signaling pathways involved in their modulation. Our results are consistent with the notion that Cyr61 does play a key role in reprogramming gene expression in SM cells and that specific signaling pathways concomitantly regulate the expression of both Cyr61 and Cyr61 target genes.
MATERIALS AND METHODS
Chemical inhibitors were purchased from Calbiochem (San Diego, CA). All other chemicals used were of reagent grade. Y-27632 was kindly provided by Dr. T. Kondo (Welfide, Osaka, Japan). The Y-27632 selectively inhibits Rho-associated kinase/ROCK and suppresses, at least in part, Rho-induced reorganization of the actin cytoskeleton (12, 34). Antibodies to Cyr61 were described previously (33). Antibodies to integrin subunits were from Dr. Grzesik (15). Antibodies to VEGF were from Chemicon. Antibodies to hypoxia-inducible factor-1α (HIF-1α), SM α-actin, and GAPDH were from Pharmingen, Upstate Biotech, and Oncogene, respectively.
Cell culture, stretching, and drug treatment.
Primary cultures of SM cells were established from fetal bovine bladders obtained from slaughterhouse as previously described (16). The institutional review board at the University of Pennsylvania and SUNY Downstate had given full approval for use of animal by-products in this study. Primary cultures from several animals were used in our experiments. Cells were maintained in DMEM supplemented with 10% fetal bovine serum and antibiotics in a humidified atmosphere containing 5% CO2 in air at 37°C. Freshly isolated SM cells were phenotypically characterized by using muscle specific antibodies against SM α-actin and myosin heavy chain (MHC). Mechanical stretch of the cells was performed by using a device designed to apply a precise and reproducible biaxial strain to a compliant type I collagen-coated membrane on which the cells were grown (33, 38). Typical experiments were carried out with stretch units in which the cells were seeded at a density of 250,000 cells/well and incubated for 24 h in serum-containing medium. The medium was then removed and replaced by serum-free medium. Cells were subjected to 10% cyclic biaxial strain at a frequency of 0.3 Hz. This strain regimen was shown to produce maximum changes in Cyr61 expression without inducing apparent cell injury (33). For controls, cells were cultured under the same conditions but were not subjected to mechanical strain. After completion of the stretch regimen, cells were harvested and processed for RNA and protein analyses. To test the effects of specific inhibitors on signal transduction pathways, the cells were incubated in the presence of one of the following pharmacological drugs for the duration of the experiment: Y-27632 (10 μM), wortmannin (20 μM), PD-098059 (20 μM), or latrunculin B (0.5 μM).
A recombinant adenovirus (rAd) vector expressing an antisense Cyr61 cDNA fragment (Ad-AS-Cyr61) was used to interfere with Cyr61 protein expression in cultured cells. The Ad-AS-Cyr61 was generated by cloning a 386-nucleotide fragment overlapping the initial Cyr61 ATG translation start site into the pAd.CMV.Link.1 vector. The antisense DNA fragment was initially generated by PCR amplification using the primers 5′-ggca/ggtacc/cg-ccacaatgagctccagcac-3′ and 5′-ggca/agatct/tgcactggtgtttacagttgggc-3′. Recombinant adenoviruses were produced by cotransfecting the adenoviral shuttle vector with a viral backbone cut with ClaI. Creation of recombinant virus and large-scale preparation of rAd was performed by the Vector Core of the Institute for Gene Therapy at the Wistar Institute. Overexpression of Cyr61 protein was performed by infecting the cells with another adenovirus encoding the Cyr61 cDNA driven by the CMV promoter. An adenovirus encoding green fluorescent protein (GFP; Ad-GFP) was used as control. All adenoviruses were replication deficient and used at 20 multiplicity of infection with no apparent cytotoxicity.
Northern blot hybridization.
Total RNA was extracted from cells using TRIZOL (Invitrogen, Carlsbad, CA). Samples containing 12 μg of total RNA were fractionated by electrophoresis in 1% agarose/formaldehyde gels, transferred to Zeta-Probe nylon filters (Bio-Rad, Richmond, CA), and hybridized sequentially to [α-32P]-dCTP-labeled cDNA probes (>108 cpm/μg) as described below in the results section. Membranes were stripped between two successive hybridizations according manufacturer's instructions (Bio-Rad). Total RNA loading and transfer were evaluated by probing with a GAPDH cDNA probe. The filters were analyzed by phosphorimaging, and hybridization signals were quantified to determine the steady-state mRNA levels in each sample (Molecular Dynamics). The mRNA levels were normalized to those of GAPDH to compensate for loading and transfer.
RNase protection assay.
The mRNA levels of MHC isoforms (SM-1 and SM-2) were measured by RNase protection assay relative to those of GAPDH mRNA. Accordingly, two bovine riboprobes were used that were complementary to SM-1 and SM-2 mRNAs. The riboprobes were generated by reverse transcription and PCR amplification using the following primers: for SM-1, 5′-aagatggccgagcagtacaag-3′ and 5′-ccttctagaaggaacgaaaga-3′; for SM-2, 5′-aagatggccgagcagtacaag-3′ and 5′-tttttatgttctggggttgtcc3-′. The PCR products were cloned into the cloning vector pCRII from Invitrogen and further sequenced to verify their orientation and identity. For the RNase protection assay, the plasmids were linearized and used as a template for in vitro transcription by either SP6 or T7 RNA polymerase to generate [32P]uridine triphosphate-radiolabeled RNA probes by using an in vitro transcription kit (Promega). Total RNA (12 μg) was resuspended in hybridization buffer containing 80% formamide, 1 mM EDTA, 40 mM piperazine-N,N′-bis(2-hexanesulfonic acid), pH 6.4, 0.2 M sodium acetate, and 1 × 106 cpm riboprobe and denatured at 85°C for 5 min. After 24 h of incubation at 45°C, nonhybridized RNAs were digested with 40 μg/ml ribonuclease A and 100 units/ml ribonuclease T1. The protected hybrids were then precipitated and separated on a 4% polyacrylamide-urea denaturing sequencing gel followed by autoradiography. The protected bands were quantitated by use of a Phosphorimager.
DNA probes for Cyr61 and GAPDH have been previously described (33). The DNA probes for VEGF, αv-integrin, and SM α-actin were generated by RT-PCR using total RNA from serum-stimulated SM cells as a template together with the primer combinations 5′-caccaaagccagcacatag-3′ and 5′-ccaattccaagaggaaccg-3′ for VEGF, 5′-atgggggtgaagataaaaaag-3′ and 5′-acattggcaggagtaaattgg-3′ for αv-integrin and 5′-tgaagaggaagacagcacag-3′ and 5′-gcagtagtcacgaaggaatag-3′ for SM α-actin. The PCR products were purified, cloned into the expression vector pCRII, sequenced, and used in Northern blot hybridization analyses as probes.
Western blotting and immunodetection.
For detection of Cyr61, integrin subunits, SM α-actin, and GAPDH proteins, cell lysates were prepared by harvesting the cells in 0.1% Triton X-100 lysis buffer. Protein concentration was determined by using the Bradford protein assay (Bio-Rad). Protein samples (20 μg) were separated by 6–12% gradient polyacrylamide gels, transferred to nitrocellulose membranes, and further incubated with specific antibodies to either Cyr61, integrin subunits, SM α-actin, or HIF-1α. For detection of VEGF polypeptide, culture medium was concentrated fivefold by using Ultra Centricon columns (Millipore). Equal amounts of protein concentrates were loaded on and fractioned in SDS-polyacrylamide gel. Immunodetection was performed by using enhanced chemiluminescence (Amersham Bioscience). The protein levels were determined by densitometric scanning of the protein band signals. Obtained values were expressed in arbitrary densitometric units and normalized to those of GAPDH to correct for total protein loading. Note that these values are merely relative indexes.
Data were expressed as mean ± SD from three independent experiments. To compare data from different experiments, steady-state mRNA and protein levels in control cells were set to 100%. A two-factor analysis of variance was used to compare dependent variables. Statistical comparisons were made according to the Student's t-test for unpaired data, and P values of <0.05 were considered significant.
Expression pattern of the Cyr61 and potential Cyr61-regulated genes in mechanically stimulated cells.
To define Cyr61 function under mechanically challenging conditions, we determined the expression profile and kinetic parameters of both Cyr61 and several of its potential gene targets upon application of cyclic biaxial stretch to cultured bladder SM cells. When cells were subjected to 10% cyclic stretch at a frequency of 0.30 Hz in multiple experiments, there was a time-dependent increase of Cyr61 gene expression. The steady-state mRNA levels of Cyr6l increase as early as 10 min after stretch application, peak after 30–60 min, and decline rapidly thereafter (Fig. 1A). The changes in Cyr61 mRNA levels correlate to similar changes in the Cyr61 protein contents that increased in a time-dependent manner in stretched cells compared with cells remaining under static conditions (Fig. 1B). Densitometric scanning of the protein bands showed a 2.2- and 3.1-fold increase of Cyr61 protein content after 1 and 4 h of mechanical stimulation, respectively (data not shown).
We further examined the expression profile of potential Cyr61 target genes including VEGF, αv-integrins, and cytoskeletal proteins such as SM α-actin and MHC isoforms SM-1 and SM-2. Representative experiments are shown in Figs. 1 and 2. Mechanical stretch significantly increased the steady-state transcript levels of VEGF, αv-integrin, and SM α-actin with a delayed expression type of pattern (Fig. 1C). In particular, the increase of VEGF mRNA levels was initially evident after 6 h of stretch, reached a peak after 8 and 12 h when expression was 3.1-fold greater than in control cells, and gradually declined thereafter approaching baseline values after 24 h. The application of cyclic stretch also produced a persistent accumulation of αv-integrin mRNA that began to increase after 6 and 12 h of stretch, resulting in a 2.9-fold increase after 24 h relative to those in cells cultured under static conditions. The expression pattern of SM α-actin mRNA was similar to those of αv-integrin, although the increase was much less pronounced (a 2.4-fold maximum induction after 24 h of cyclic stretch).
To determine whether changes at the mRNA level resulted in corresponding changes in protein levels, cells were exposed to cyclic stretch for either 6, 12, or 24 h. Culture medium was then evaluated for VEGF protein levels, and cell lysates were assessed for αv-integrin and SM α-actin protein expression by Western blot and immunodetection analyses. As shown in Fig. 1D, mechanical stretch significantly increased the protein levels of both VEGF and αv-integrin at all time points compared with control static conditions, whereas the increases in SM α-actin protein levels were significant only after 12 and 24 h after the onset of mechanical stretch. Therefore, the changes seen at the mRNA level correlate well to those at the protein level.
Because mechanical stretch affected SM α-actin expression, we sought to determine whether the transcription levels of MHC isoforms SM-1 and SM-2 were altered. These proteins play an important role in contraction of the actomyosin cytoskeleton and generation of tensional forces. As shown in Fig. 2, the steady-state mRNA levels of SM-1 and SM-2 were not significantly altered upon application of mechanical stretch to the cells for up to 24 h as determined by RNase protection assay. Similarly, mechanical stretch did not significantly alter the mRNA levels of either α2(I) collagen or matrix metalloproteinase-1, suggesting that cyclic stretch caused a selective gene regulation and not merely an overall increase of gene expression (data not shown).
Suppression of Cyr61 expression altered mechanical regulation of VEGF, αv-integrin, and SM α-actin.
The above- described experiments indicated a difference in the expression pattern between the Cyr61 gene that exhibited a rapid and transient expression pattern and that of VEGF, αv-integrin, and SM α-actin genes whose expression was delayed and long lasting in response to mechanical stretch. This is compatible with a model of progressive changes of the genetic program in which the activation of specific sets of genes acting at discrete stages occurs. To determine whether mechanical stretch-induced Cyr61 regulates the expression of VEGF, αv-integrin, and SM α-actin, SM cells were infected with a replication-deficient adenovirus encoding a Cyr61-specific antisense cDNA fragment (Ad-AS-Cyr61), after which the response of the VEGF, αv-integrin, and SM α-actin genes to mechanical stretch was examined. Cells infected with the same adenovirus encoding GFP (Ad-GFP) and either cultured under static conditions or subjected to mechanical stretch were used as negative and positive controls, respectively. As shown in Fig. 3A, mechanical stretch of Ad-AS-Cyr61-infected cells induced little or no detectable levels of Cyr61 protein, whereas Ad-GFP-infected cells expressed the expected large amount of Cyr61 protein after application of mechanical stretch to the cells. The increase of Cyr61 protein levels was effectively suppressed in Ad-AS-Cyr61-expressing cells throughout the time period that mechanical stretch was applied (up to 24 h), indicating that Ad-AS-Cyr61 efficiently interfered with the translation of the Cyr61 protein (data not shown). Under these conditions, cells were subjected to mechanical stretch for either 6 or 12 h and then harvested, and the expression of the potential Cyr61 target genes was assessed by Northern blot analysis. As shown in Fig. 3B, mechanical stretch produced the expected increase in VEGF, αv-integrin, and SM α-actin mRNA levels after 6 and 12 h in the Ad-GFP-infected cells. In contrast, the expression of VEGF and αv-integrin was not significantly altered in Ad-AS-Cyr61-infected cells upon application of stretch for either 6 or 12 h. The SM-α actin protein levels were reduced by 39% upon application of stretch for 12 h. These data clearly indicate a regulatory relationship between Cyr61 and late-responsive genes such as those encoding VEGF, αv-integrin subunit, and SM α-actin in mechanically stimulated cells. Conversely, the steady-state mRNA levels of either SM-1 or SM-2 were not significantly altered in mechanically stimulated cells, and their basal expression was independent of Cyr61 protein suppression (data not shown).
Cyr61-dependent regulation of VEGF is mechanical stretch specific.
Because the Cyr61 protein appeared to be required for the expression of selected late-responsive genes in mechanically stimulated cells, we asked the question whether stimuli other than mechanical strain require Cyr61 as a mediator for the activation of these genes as well. To address this specificity issue, we used as a model the VEGF gene. In addition to being encoded by a mechanosensitive gene, VEGF is also a well-characterized angiogenic factor whose expression occurs as an adaptive response of SM cells to low oxygen tension both in vivo and in vitro (29). In particular, the posttranscriptional upregulation of the HIF-1α plays a major role in VEGF gene induction by either hypoxia or hypoxia mimetic reagents such as cobalt chloride (CoCl2) or desferrioxamine (DFO). The latter exploit the cellular oxygen-sensing mechanisms or the hypoxia signal transduction pathways for gene activation. Therefore, we treated SM cells with CoCl2 and determined the steady-state levels of Cyr61 and VEGF mRNAs by Northern blot hybridization analyses. As shown in Fig. 4A, CoCl2 induced accumulation of both Cyr61 and VEGF mRNA in a time-dependent manner. Maximum increases of 3.8- and 3-fold were observed after 1 and 12 h for Cyr61 and VEGF, respectively, at a metal concentration of 100 μM. Treatment of the cells with DFO had the same effects as CoCl2, and higher concentrations of either CoCl2 or DFO resulted in cellular toxicity and reduced Cyr61 and VEGF mRNA levels (data not shown). Thus, hypoxia mimetic reagents increase Cyr61 and VEGF mRNA levels with kinetic parameters similar to but not identical to that of mechanical stretch. Additionally, as shown in the representative experiment in Fig. 4B, treatment of the cells with CoCl2 strongly increased HIF-1α protein levels as expected from previous studies (4). To determine whether Cyr61 mediates CoCl2-induced VEGF expression, we determined the steady-state mRNA levels of VEGF in CoCl2-treated cells upon suppression of Cyr61 expression with Ad-AS-Cyr61. As shown in Fig. 4C, interference with Cyr61 expression did not significantly alter CoCl2-induced VEGF expression, indicating that the expression of Cyr61 may be sufficient but not necessary for VEGF gene activation by CoCl2 in SM cells. By the same token, interference with Cyr61 expression did not alter in any way the levels of HIF-1α protein in CoCl2-treated cells. Thus Cyr61-induced VEGF gene expression is mechanical stretch specific.
Signaling molecules involved in Cyr61-mediated gene expression.
To characterize the pathways involved in Cyr61-mediated gene expression, we sought to determine whether the signaling events previously shown to mediate mechanical stretch-induced Cyr61 expression were involved. This was performed by preincubating the cells with pharmacological agents that interfere with the activity of known signaling molecules. As shown in Fig. 5A, mechanical stretch-induced Cyr61 expression was substantially reduced in the presence of Y-27632, wortmannin, and latrunculin B, which inhibit RhoA kinase, PI3-kinase, and actin polymerization, respectively. The inhibition of ERK1/2 by PD-098059 did not alter the Cyr61 expression. In parallel experiments, inhibition of Rho kinase, PI3-kinase, and actin polymerization reduced the upregulation of VEGF, αv-integrin, and SM α-actin expression by mechanical stretch, indicating that both Cyr61 and its target genes are commonly regulated through the same mechanotransduction pathways (Fig. 5, B–D). The ERK1/2 inhibitor reduced the expression of SM α-actin but had no effects on the expression of VEGF and αv-integrin, suggesting that both Cyr61-dependent and -independent mechanisms contribute to mechanical stretch-induced SM α-actin. It is noteworthy that although inhibition of either RhoA kinase or PI3-kinase altered Cyr61, VEGF, αv-integrin, and SM α-actin to various extents, latrunculin B, a toxin that disrupts the actin cytoskeleton by sequestering G-actin monomers and therefore inhibiting actin polymerization, nearly abolished the expression of Cyr61 and all Cyr61 target genes, suggesting an important role of actin cytoskeleton integrity in both Cyr61 expression and activity.
Effects of adenovirus-mediated overexpression of Cyr61 on cells cultured under static conditions.
To determine whether the expression of Cyr61 is sufficient to induce the expression of potential target genes, cells were infected with an adenovirus overexpressing Cyr61 cDNA (Ad-Cyr61). Cells were incubated for various periods of time in serum-free medium, which results in minimal expression of the endogenous Cyr61 gene and minimal growth. Cells infected with Ad-GFP were used as control. As shown in Fig. 6A, a marked increase in Cyr61 protein levels was observed in Ad-Cyr61-infected cells and expression was sustained for more than 36 h. Cells infected with Ad-GFP did not express Cyr61 protein as expected. Sustained expression of Cyr61 had no effects on cell proliferation as determined by thymidine incorporation assay (data not shown).
Furthermore, overexpression of Cyr61 increased the steady-state mRNA levels of either VEGF, αv-integrin, or SM α-actin but with slightly different kinetics (Fig. 6B). The mRNA levels of VEGF started increasing after 6 h, peaked at 24 h and decreased progressively thereafter. The expression of αv-integrin subunit and SM α-actin was upregulated after 12 h and showed a sustained increase up to 36 h. The changes at the mRNA levels correlate well with those at the protein level as VEGF, αv-integrin, and SM α-actin protein levels increased proportionally to their mRNAs (Fig. 6C). In contrast, adenovirus-mediated overexpression of Cyr61 did not seem to affect the mRNA levels of α5- and β5-integrin subunits. Surprisingly, the mRNA for β1-integrin subunit was reduced, suggesting a differential regulation of integrin subunit expression by Cyr61. The mRNA levels of α1- and α4-integrin subunits and α2(I) and α1(III) collagen chains were not affected in cells overexpressing Cyr61 (data not shown).
Using RNase protection assay, we determined the mRNA levels of MHC isoforms SM-1 and SM-2 upon overexpression of Cyr61 in the cells. As shown in Fig. 7, the steady-state mRNA levels of both SM-1 and SM-2 significantly increased (2.8- and 2.5-fold, respectively) in Ad-Cyr61-infected cells after 24 and 36 h of incubation time. The ratio of SM-1 to SM-2 decreased as well after 36 h in Cyr61-overexpressing cells. These data clearly indicate that the effects of chronically expressed Cyr61 extend to motor proteins of the contractile apparatus. Thus MHC isoform mRNA levels are dependent on whether Cyr61 is expressed transiently or chronically in the cells because the transient expression of Cyr61 in mechanically stimulated cells did not affect SM-1 and SM-2 whereas its sustained expression did.
Cyr61-dependent regulation of VEGF, αv-integrin, SM α-actin, and MHC gene expression requires an intact cytoskeleton.
Using pharmacological inhibitors, we sought to determine what signaling molecules, if any, were involved in Cyr61-mediated gene expression. As shown in Fig. 8, the expression of Cyr61 target genes in Ad-Cyr61-infected cells was not dramatically affected when the cells were incubated with ERK1/2 inhibitor. In contrast, incubation with latrunculin B induced a substantial reduction of VEGF and αv-integrin transcripts and nearly abolished the increase in SM α-actin and SM-1 and SM-2 mRNA levels, suggesting an important role of the actin cytoskeleton in promoting the effects of chronic expression of Cyr61. Treatment of the cells with latrunculin B induces depolymerization of stress fibers and morphological changes characteristic of an altered cytoskeleton, as previously reported (12). Similarly, incubation with Y-27632 inhibitor induced rounding of the cells as a result of the collapse of the cytoskeletal architecture and reduced the expression of Cyr61 target genes in both GFP- and Cyr61-infected cells (data not shown). Thus an intact cytoskeleton is required for the effects of Cyr61 on the expression of its target genes in SM cells.
Cyr61 belongs to the group of so-called matricellular proteins that appear in the ECM either at specific stages of development or in association with specific pathological conditions. These proteins that include, in addition to Cyr61, the thrombospondins, tenascin C, connective tissue growth factor, osteopentin, and osteonectin play important roles in regulating cell fate during development and in the pathogenesis of several stress-related diseases through yet incompletely understood mechanisms. In particular, Cyr61 was reported to be one of the earliest genes whose expression was deregulated in SM-rich tissues (e.g., aorta and bladder) with the onset of hypertension or bladder outlet obstruction (8, 35). Cyr61 was also described as a potent and early marker of mechanical-load-induced cardiac hypertrophy (18). Highly compliant tissue such as the bladder wall may experience high strains during normal voiding that are biaxial in nature without concomitant alterations of the expression of the pathogenic matricellular proteins (7, 8). However, it is the chronic nature of bladder wall tension in the urodynamically overloaded detrusor muscle during obstructive diseases that is considered the trigger of a genetic reprogramming within the tissue. Cyclic stretch and relaxation applied to cultured bladder SM cells in vitro at a magnitude and frequency ranging from 5 to 25% and 5 to 20 cycles per minutes, respectively, is thought to simulate the strain regimen within an overloaded detrusor muscle layer of obstructed bladders (40). However, although our mechanical stretch system delivers strains within the above-mentioned range, the actual strain experienced by the cells is complex and depends on several parameters, including cell orientation and morphology and loading conditions (3). Additionally, studies have demonstrated that all cells in a pressure-deformed membrane system do not “see” the same strain field as the circumferential strain substantially decreases from the center to the membrane edge (31, 37). Under these conditions, it is unknown whether the alterations of gene expression result from signaling between cells receiving different strain levels. Other data from our laboratory showed that the application of a relatively uniform radial and circumferential strain using the Flexercell Tension Plus System (FlexCell International, McKeesport, PA) upregulated both Cyr61 and its target genes, suggesting that mechanical regulation of gene expression in bladder SM cells is possibly merely sensitive to the strain profile (data not shown). However, further studies are needed to determine the proper effects of strain direction on the expression of Cyr61 and its target genes.
Cyr61 exhibited a transient expression in cultured bladder SM cells, and the rapid reestablishment of its basal expression level after the onset of cyclic stretch might be indicative of an adaptive mechanism in which compensatory signaling pathways are activated to allow gene transcription to return rapidly to normal levels. Concordantly, stretch-induced Cyr61 upregulated, albeit to various extents, the expression of VEGF, αv-integrin, and SM α-actin genes, suggesting the existence of a gene network in mechanically stimulated cells. These data are in line with those of Chen et al. (10), who showed that exposure of cultured human fibroblasts to soluble Cyr61 protein activated a genetic program for wound repair and tissue regeneration that included the expression of angiogenic and inflammatory factors. More importantly, it was shown that Cyr61-deficient mice exhibit a substantial downregulation of VEGF, and striking similarities between Cyr61- and αv-integrin-deficient mice were reported, which suggest a potential regulatory relationship between Cyr61 and either VEGF or αv-integrin genes during development (2, 26). Thus it appears that mechanical stretch revives a genetic program that was active during tissue development. Additionally, mechanical stretch-induced Cyr61 increased, at least in part, the expression of SM α-actin, suggesting that the biologically relevant functions of Cyr61 encompass also structurally related effects through regulation of cytoskeletal protein genes.
Another important finding in our studies is that Cyr61 target genes were commonly regulated through mechanotransduction pathways involved in stretch-induced Cyr61 gene expression. In work that stemmed largely from our laboratory, we showed that the expression of the Cyr61 gene was controlled through multiple signaling pathways involving RhoA GTPase coupled to the organization of the actin cytoskeleton as well as other signaling molecules, including PKC, p38 mitogen-activated protein (MAP) kinase, and PI3-kinase (16, 33). We showed here that these signaling pathways affect, albeit to various extents, mechanical stretch-induced expression of VEGF, αv-integrin, or SM α-actin genes as well in agreement with data reported by other laboratories. In particular, mechanical stretch-induced VEGF gene expression was shown to involve signaling mechanisms dependent on PI3-kinase-mediated activation of PKC-ζ in retinal pericytes and mesangial cells (14, 32). However, hypoxia was by far the only other potent stimulus known to induce VEGF expression and involves signaling mechanisms associated with HIF-1α activation. Although some reports suggested potential effects of mechanical stretch on HIF-1α expression and activation, we have found no evidence of such an effect in our cellular model (19, 29) (data not shown). Our data ruled out also a major role of Cyr61 in hypoxia-induced VEGF expression. Taken together, these observations suggest that separate and stimulus-dependent signaling cascades are likely implicated in VEGF expression. In addition to VEGF, SM α-actin gene was also reported to be sensitive to mechanical strain via mechanotransduction pathways involving Rho GTPase and MAP kinase pathways (1, 22). Other studies have shown that tissue stretch over a 24-h time period initiates dynamic alterations in integrin gene expression, including those encoding αv-, β1-, and β3-integrin subunits through yet-unknown mechanotransduction mechanisms (36). In our hands, the expression of αv-integrin subunit seems to be the most affected by stretch-induced Cyr61, whereas the expression of other integrin subunits remained unaltered.
Furthermore, we showed that the mechanical signals were largely dependent on the actin cytoskeleton integrity because the expression of Cyr61 and its target genes was abolished upon either disruption of the actin cytoskeleton or inhibition of RhoA GTPase/actin signaling. Thus Cyr61-mediated VEGF, αv-integrin, or SM α-actin gene expression clearly involves actin-sensing mechanisms. Actin filaments are highly dynamic structures that are being constantly assembled, disassembled, and reorganized as the cell changes its shape, divides, or adheres to a substratum. They assume the role of a scaffolding structure on which proteins of the signaling machinery and other cytoplasmic structures are docked and become activated. The potential effects of Cyr61 on a common cytoskeletal actin-sensitive factor affecting the expression of distinct mechanoresponsive genes could explain its effects on the expression of the VEGF, αv-integrin, and SM α-actin genes. Interestingly, one interpretation of how mechanical stimulus-specific gene expression is achieved is based on the idea that mechanical loading causes the expression and activation of transcription factors that bind to “mechanoresponsive” promoter elements of mechanosensitive genes. Functional activator protein-1 and NF-κB elements have been found in the promoter region of Cyr61 target genes, and it is not excluded that the Cyr61 protein targets either directly or indirectly the activation and/or expression of these trans-acting factors via actin-dependent mechanisms. There are several observations in support of this possibility, including the demonstration by Lin et al. (23) that Cyr61 promotes resistance of MCF-7 cells to apoptotic agents by a mechanism involving the activation of NF-κB. Other evidence suggests that globular actin monomers shuttle between the nucleus and the cytoplasm and modulate the activity of transcription factors such as activator protein-1, serum response factor, and NF-κB either through direct physical interactions or by sequestering cofactors required for their activation (6, 9, 23). Further studies are needed to delineate precisely the underlying mechanisms of Cyr61-mediated gene expression.
An additional important issue was to determine whether the effects of Cyr61 on the expression of other genes were dependent on the availability of other molecules and/or binding partners that become readily primed in a mechanically challenging environment. Using a recombinant adenoviral gene transfer approach, we showed that overexpression of Cyr61 in cells cultured under static conditions strongly increased VEGF, αv-integrin, and SM α-actin gene expression, suggesting that the mechanical stimulus does not necessarily create a contextual environment for Cyr61 action. However, overexpression of Cyr61 induced additional unanticipated alterations not previously seen in mechanically stimulated cells, including an increase in MHC isoforms, SM-1 and SM-2, and a significant downregulation of the β1-integrin subunit expression. Differential expression of integrins has, indeed, recently been associated with forced expression of Cyr61 in small lung cancer cells, suggesting a good correlation between chronic expression of Cyr61 and that of integrin genes independently of the cell type (39). On the other hand, modulation of the mRNA levels of MHC isoforms suggests that the action of Cyr61 may extend to SM cell-specific functions as well. Indeed, both SM-1 and SM-2 encode motor proteins with unique functional roles in SM cells. Their expression was shown to be particularly upregulated in hypertrophic SM, and alteration of their expression in cultured cells involves both transcriptional and posttranscriptional mechanisms (27). The fact that mechanical stretch-induced Cyr61 cells did not alter the mRNA levels of SM-1 and SM-2 suggests that perhaps the sustained overexpression of Cyr61 induces molecular events and/or generates latent signals more effectively than when it is transiently expressed in mechanically stretched cells. Overall, although it is possible that these genes are not all direct Cyr61 targets, it remains clear that Cyr61 overexpression in the complete absence of exogenous growth factors and mechanical stimuli leads to a substantial and selective reprogramming of gene expression within SM cells. Cyr61 most likely lies downstream of the signaling pathways that normally govern the expression of its target genes.
In summary, Cyr61 is clearly a determinant factor in the genetic reprogramming that occurs in mechanically stimulated cells. Our data showed that this IE gene-encoded polypeptide is an integral part of the mechanotransduction process because it promotes the expression of mechanosensors such as integrins and SM α-actin and propagates the mechanical signal to neighboring cells via the expression of autocrine and/or paracrine factors such as VEGF. When chronically expressed, Cyr61 has the ability to modulate MHC isoform levels and alter SM differentiation. A genome-wide screen of Cyr61 target genes should determine the extent to which Cyr61 affects gene expression and will aid in mapping out additional potential gene networks that are activated by mechanotransduction mechanisms.
This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R01-DK-060572 and R21 DK-068483 (to B. Chaqour).
We are grateful to Dr. W. J. Grzesik for providing us with integrin antibodies. D. Zhou was a Postdoctoral Researcher in B. Chaqour's laboratory previously located at the School of Dental Medicine at the University of Pennsylvania.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society