Integrin Dependence of Brain Natriuretic Peptide Gene Promoter Activation by Mechanical Strain*

Expression of the brain natriuretic peptide (BNP) gene in cultured neonatal rat ventricular myocytes is activated by mechanical strain in vitro. We explored the role of cell-matrix contacts in initiating the strain-dependent increment in human BNP (hBNP) promoter activity. Coating the culture surface with fibronectin effected a dose-dependent increase in basal hBNP luciferase activity and amplification of the response to strain. Preincubation of myocytes with an RGD peptide (GRGDSP) or with soluble fibronectin, each of which would be predicted to compete for cell-matrix interactions, resulted in a dose-dependent reduction in strain-dependent hBNP promoter activity. A functionally inert RGE peptide (GRGESP) was without effect. Using fluorescence-activated cell sorting, we demonstrated the presence of β1, β3, and αvβ5 integrins in myocytes as well as non-myocytes and α1 only in non-myocytes in our cultures. Inclusion of antibodies directed against β1, β3, or αvβ5, but not α1, α2, or cadherin, was effective in blocking the BNP promoter response to mechanical strain. These same antibodies (anti-β3, -β1, and -αvβ5) had a similar inhibitory effect on strain-stimulated ERK, p38 MAPK, and, to a lesser extent, JNK activities in these cells. Cotransfection with chimeric integrin receptors capable of acting as dominant-negative inhibitors of integrin function demonstrated suppression of strain-dependent BNP promoter activity when vectors encoding β1 or β3, but not β5, α5, or a carboxyl-terminal deletion mutant of β3(β3B), were employed. These studies underscore the importance of cell-matrix interactions in controlling cardiac gene expression and suggest a potentially important role for these interactions in signaling responses to mechanical stimuli within the myocardium.

Expression of the brain natriuretic peptide (BNP) gene in cultured neonatal rat ventricular myocytes is activated by mechanical strain in vitro. We explored the role of cell-matrix contacts in initiating the strain-dependent increment in human BNP (hBNP) promoter activity. Coating the culture surface with fibronectin effected a dose-dependent increase in basal hBNP luciferase activity and amplification of the response to strain. Preincubation of myocytes with an RGD peptide (GRGDSP) or with soluble fibronectin, each of which would be predicted to compete for cell-matrix interactions, resulted in a dose-dependent reduction in straindependent hBNP promoter activity. A functionally inert RGE peptide (GRGESP) was without effect. Using fluorescence-activated cell sorting, we demonstrated the presence of ␤ 1 , ␤ 3 , and ␣ v ␤ 5 integrins in myocytes as well as non-myocytes and ␣1 only in non-myocytes in our cultures. Inclusion of antibodies directed against ␤ 1 , ␤ 3 , or ␣ v ␤ 5 , but not ␣ 1 , ␣ 2 , or cadherin, was effective in blocking the BNP promoter response to mechanical strain. These same antibodies (anti-␤ 3 , -␤ 1 , and -␣ v ␤ 5 ) had a similar inhibitory effect on strain-stimulated ERK, p38 MAPK, and, to a lesser extent, JNK activities in these cells. Cotransfection with chimeric integrin receptors capable of acting as dominant-negative inhibitors of integrin function demonstrated suppression of straindependent BNP promoter activity when vectors encoding ␤ 1 or ␤ 3 , but not ␤ 5 , ␣ 5 , or a carboxyl-terminal deletion mutant of ␤ 3 (␤ 3 B), were employed. These studies underscore the importance of cell-matrix interactions in controlling cardiac gene expression and suggest a potentially important role for these interactions in signaling responses to mechanical stimuli within the myocardium.
Brain natriuretic peptide (BNP) 1 is a vasoactive hormone that, despite its name, is produced primarily in the heart. Like atrial natriuretic peptide (ANP), it possesses potent natriuretic, diuretic, and vasorelaxant activities, properties that position it physiologically as a potential antagonist of vasoactive systems (e.g. the renin-angiotensin and sympathetic nervous systems) associated with intravascular volume expansion and increased blood pressure (1).
Like ANP, BNP is expressed in both atrial and ventricular myocardia, although differential expression is not so marked as that for ANP (the atrial/ventricular ratio for BNP expression is ϳ3:1, whereas that for ANP is in the range of 40:1) (2). As with ANP, ventricular expression of BNP is activated in pathophysiological states associated with hemodynamic overload in vivo (3,4) and following exposure to hypertrophy-promoting maneuvers in vitro (5)(6)(7)(8)(9)(10). In the latter group are a number of biochemical (e.g. ␣-adrenergic agonists, endothelin, angiotensin II, and cardiotrophin-1) as well as physical (e.g. mechanical strain and hypoxia) stimuli that trigger increases in cell size and protein synthesis, reactivate a program of fetal gene expression, and reorganize sarcomeric structure of myocytes in culture. These phenotypic changes closely resemble those associated with myocyte hypertrophy in vivo (11).
A number of recent studies have documented the importance of the extracellular matrix in establishing cellular responses, both qualitatively and quantitatively, to a variety of biochemical and physical stimuli (12,13). In the case of mechanical strain, it has been hypothesized that key matrix-integrin attachments may participate in the signal transduction process linking the strain signal to changes in gene expression, cytoskeletal reorganization, and DNA synthesis (14).
We have recently shown that application of cyclical, passive mechanical strain (i.e. stretch) to cultured neonatal rat ventricular myocytes in vitro results in stimulation of immunoreactive BNP secretion, increased steady-state levels of the BNP gene transcript, and activation of a transfected human BNP (hBNP) gene promoter (8). In the present study, we show that both basal and strain-activated BNP promoter activities are dependent upon specific contacts that the ventricular myocyte makes with proteins in the extracellular matrix, implying a potential signaling function for these contacts in the subsequent activation of gene expression. (Lake Placid, NY). Glutathione S-transferase-c-Jun was prepared as described (15). Purified hamster anti-rat/mouse ␣ 1 , ␣ 2 , ␣ 5 , and ␤ 1 monoclonal antibodies; purified mouse anti-rat ␤ 3 monoclonal antibody; FITC-labeled anti-hamster IgG; and FITC-conjugated hamster anti-rat ␤ 1 monoclonal antibody were obtained from Pharmingen (San Diego, CA). Anti-pan cadherin monoclonal antibody was purchased from Sigma. Anti-␣ v ␤ 5 monoclonal antibody and phycoerythrin-labeled antimouse IgG were provided by Dean Sheppard (University of California, San Francisco). Other reagents were obtained though standard commercial suppliers.

Materials-Fibronectin
Plasmid Constructions-The construction of Ϫ1595 hBNPLUC has been described previously (16). Cytomegalovirus-driven chimeric ␤ 1 , ␤ 3 , ␤ 3 B, ␤ 5 , and ␣ 5 integrin expression vectors were provided by Susan E. LaFlamme (Albany Medical College, Albany, NY) (17,18). These vectors link the DNA sequence encoding the intracellular domains of the individual integrins to the extracellular and transmembrane domains of the interleukin-2 receptor (gp55 subunit). Plasmid containing the Rous sarcoma virus-␤-galactosidase reporter was provided by Weijun Feng (University of California, San Francisco).
Cell Culture and Application of Mechanical Strain-Ventricular myocytes were prepared from 1-2-day-old neonatal rat hearts by alternate cycles of 0.05% trypsin digestion and mechanical disruption as described previously (19). Cells (1 ϫ 10 6 ) were cultured on collagencoated FLEX plates in Dulbecco's modified Eagle's/H-21 medium containing 10% bovine calf serum (Hyclone Laboratories, Logan, UT), 2 mM glutamine, 10 units/ml penicillin, and 100 mg/ml streptomycin. A glass cloning cylinder (1-cm diameter) was placed in the middle of each well to preclude cell attachment, thereby placing the majority of adherent cells on the outer 75% of the culture surface where distension is maximal (20). We used the Flexcell Strain apparatus to apply mechanical strain to our cultured myocytes. This is a commercially marketed device that applies vacuum (at variable levels) to the undersurface of a collagen-coated silicone elastomer disc positioned in the bottom of a six-well culture dish, which, in turn, is secured in an airtight manifold. Computer-regulated application of negative pressure to the underside of the culture dish results in downward displacement of the disc and stretch of adherent cells cultured on its upper surface (21)(22)(23)(24). The medium was changed 24 h prior to initiation of the experiment. Cells were subjected to cyclical strain (60 cycles/min) on the Flexcell Strain apparatus at a level of distension sufficient to promote ϳ20% increment in surface area at the point of maximal distension on the culture surface (20). Preliminary experiments indicated that this level of distension provided nearmaximal induction of BNP gene promoter activity (data not shown).
Transfection and Luciferase and ␤-Galactosidase Assays-Freshly prepared ventricular myocytes were transiently transfected with the indicated reporters and expression vectors by electroporation (Gene Pulser, Bio-Rad) at 280 mV and 250 microfarads. DNA content in individual cultures was normalized with pUC18. After transfection, cells were plated and cultured as described above. Cells were harvested and lysed in 100 l of cell culture lysis reagent (Promega, Madison, WI). Protein concentration of each cell extract was measured using Coomassie protein reagent (Pierce). Cell lysates were processed (20 g of protein/sample) and assayed for luciferase as described using a commercially available kit (Promega). Measurements of ␤-galactosidase activity were made using the Galacto-Light Plus kit from Tropix Inc. (Bedford, MA). To ensure reproducibility, experiments were repeated three to seven times.
Coating of Plates with Fibronectin-Varying concentrations (0.0001-1 g/cm 2 ) of fibronectin in aqueous solution were overlaid onto FLEX plates and allowed to air-dry in a laminar flow hood. All plates were washed with phosphate-buffered saline and rinsed with 10% enriched calf (EC) serum-containing medium to remove unadsorbed protein prior to addition of cells.
Flow Cytometry-Cells were harvested in 0.005% trypsin (1:10 dilution) and placed in Dulbecco's modified Eagle's/H-21 medium containing 10% enriched calf serum. Cells were pelleted, washed with phosphate-buffered saline, and incubated with the anti-integrin antibodies for 30 min at 4°C. Cells were washed with phosphate-buffered saline three times and incubated with FITC-labeled anti-hamster IgG or phycoerythrin-labeled anti-mouse IgG secondary antibody for an additional 30 min at 4°C. Cells were washed, resuspended in phosphate-buffered saline, and subjected to flow cytometric analysis.
Immunoprecipitation and Kinase Assay-The cells were harvested in lysis buffer (0.1 ml/well; 20 mM Tris-HCl, pH 7.9, 137 mM NaCl, 1% Triton X-100, 5 mM EDTA, 1 mM EGTA, 10% glycerol, 10 mM NaF, 1 mM ␤-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, and 1 g/ml pepstatin) and centrifuged at 12,000 ϫ g for 30 min. 200 g of supernatant protein was incubated with 1 g of anti-ERK2, anti-JNK1, or anti-p38 antibody and 10 l of protein G-Sepharose for 2 h at 4°C. The immunoprecipitates were recovered by centrifugation and washed three times with cell lysis buffer and once with kinase reaction buffer (25 mM HEPES, pH 7.6, 10 mM MgCl 2, 10 mM ␤-glycerophosphate, 1 mM sodium vanadate, and 2 mM dithiothreitol). The activities of ERK and p38 were measured by adding 20 g of myelin basic protein to the immunoprecipitates in 30 l of kinase reaction buffer containing 2 Ci of [␥-32 P]ATP. The activity of JNK was measured by adding 5 g of glutathione S-transferase-c-Jun to the immunoprecipitates in 30 l of kinase reaction buffer containing 2 Ci of [␥-32 P]ATP. Reactions were incubated for 15 min at 30°C. The samples were electrophoresed on 15% SDS-polyacrylamide gels, which were then dried and subjected to autoradiography. Signals were identified and quantified by NIH Image.
Data Analysis-Data were subjected to analysis of variance, and the Newman-Keuls test was used to assess statistical significance. Results are presented as means Ϯ S.D.

RESULTS
As noted above, there is a growing body of evidence that the cell may sense physical distortion and signal downstream events in response to this distortion through perturbation of matrix-integrin interactions at the cell surface (14). Fibronectin, one of the major matrix proteins in the myocardium (25,26), has been shown to exert profound effects on gene expression in other systems (27,28). With this in mind, we explored the effects of exogenous fibronectin on the activity of the BNP gene promoter in our myocyte cultures. When culture surfaces were precoated with increasing concentrations of fibronectin, a dose-dependent increase in basal hBNP luciferase activity and amplification of the strain-dependent response (Fig. 1A) were observed, implying that interaction of the myocyte with this matrix component is an important determinant of the magnitude of the response to mechanical strain.
Disruption of cell-matrix interactions in the myocyte cultures negatively impacted the magnitude of the BNP promoter response to strain. Addition of a soluble peptide containing the matrix protein sequence RGD (i.e. GRGDSP) to the myocyte cultures, a maneuver that is known to interfere with selected matrix-integrin interactions, effected a dose-dependent reduction in strain-stimulated BNP gene promoter activity (Fig. 1B). When the same experiment was carried out with the functionally inert RGE peptide (i.e. GRGESP), no reduction in promoter activity was observed. Similar inhibition was observed when soluble fibronectin was included in the culture medium (Fig.  1C). As with the RGD peptide, a dose-dependent reduction in BNP promoter activity was seen, presumably reflecting impaired generation of matrix-integrin contacts in the cultures. Collectively, these findings imply that intact matrix-integrin attachments are required for strain-dependent induction of BNP gene transcription.
Next, we examined the nature of the integrins expressed in our cultured cells using antibody-driven fluorescence-activated cell sorting analysis. As shown in Fig. 2, we detected the presence of ␤ 1 , ␤ 3 , and ␣ v ␤ 5 integrins in our primary cultures of ventricular myocytes. Since these cultures are routinely contaminated to some degree with non-myocytes (primarily fibroblasts), we reexamined the integrin profile in myocytes cultured in the presence of bromodeoxyuridine. In addition, we examined the profile in non-myocytes (predominantly fibroblasts) isolated from the same neonatal hearts and cultured in parallel. The myocytes cultured in the presence of bromodeoxyuridine displayed a nearly identical integrin profile when compared with cells cultured in the absence of the drug, indicating that the integrins identified in these cultures reside almost exclusively on the cardiac myocytes. Non-myocytes also expressed ␤ 1 and ␤ 3 integrins in abundance. ␤ 3 integrin expression was actually more pronounced in non-myocytes than in the myocyte population. They also expressed ␣ v ␤ 5 and ␣ 1 integrins; ␣ 1 integrins were not demonstrated in the myocyte cultures.
Addition of these same antibodies to viable transfected myocyte cultures revealed differential effects on hBNP promoter activity. As shown in Fig. 3, antibodies against ␤ 1 , ␤ 3 , and ␣ v ␤ 5 clearly inhibited the promoter response to mechanical strain. Antibody against the ␣ 5 integrin effected a modest reduction in the response to strain, whereas antibody against ␣ 1 , ␣ 2 , or cadherin was without effect.
These results were supported by additional studies employing transient overexpression of chimeric integrin molecules that link the extracellular and transmembrane domains of interleukin-2 to the intracellular domains of different integrin monomeric proteins. The latter, in the absence of a bona fide extracellular domain, do not respond to stimuli requiring engagement of the extracellular matrix, but continue to associate with and compete for intracellular proteins responsible for signaling integrin-dependent activity. In effect, when the chitreated with increasing concentrations (5-50 g/ml) of soluble fibronectin for 48 h prior to cell collection. Luciferase measurements were made as described under "Experimental Procedures." *, p Ͻ 0.01 versus strain control; #, p Ͻ 0.01 or 0.05 versus static control. Neonatal rat ventricular myocytes were cultured for 48 h with or without bromodeoxyuridine (BUdR; 0.1 mM). Non-myocytes were collected from neonatal rat hearts at the preplating step and cultured for a 2-week period. Cells were detached by limited trypsin digestion and incubated with antibody against ␣ 1 , ␣ 2 , ␣ 5 , ␤ 3 , or ␣ v ␤ 5 integrin or cadherin or with FITC-labeled antibody directed against ␤ 1 integrin. After extensive washing, cells were incubated with FITC-labeled anti-hamster or phycoerythrin (PE)-labeled anti-mouse secondary antibody (Ab) and subjected to flow cytometric analysis as described under "Experimental Procedures." The x axis demonstrates fluorescence (FL) intensity, and the y axis indicates cell number. meric proteins are expressed at high levels, such competition promotes a dominant-negative inhibition of integrin-dependent activity in the cell (17,18). As shown in Fig. 4, cotransfection of increasing amounts of vector encoding chimeric ␤ 1 or ␤ 3 integrin was effective in reducing basal promoter activity and partially blocking the response to mechanical strain. Forced expression of ␤ 5 , ␣ 5 , or a carboxyl-terminal deletion mutant of ␤ 3 (␤ 3 B) that is defective in the signaling functions normally attributed to ␤ 3 failed to effect a systematic reduction in BNP promoter activity. These findings lend additional credence to the hypothesis that ␤ 1 and ␤ 3 play key roles in signaling straindependent activity in these cultures.
We have previously demonstrated that the response to mechanical strain is heavily dependent upon signaling activity transduced through the ERK (8), p38 (29), and, to a lesser extent, JNK (8) pathways. To probe the connection between the integrins and activation of these different pathways, we examined the relative abilities of the antibodies described in the legend to Fig. 3 to prevent activation of these signaling cascades. As shown in Fig. 5A, antibodies against ␤ 1 , ␤ 3 , and ␣ v ␤ 5 , all of which display the ability to suppress the strain-dependent activation of the BNP gene promoter (see Fig. 3), each effected a significant reduction of strain-dependent ERK activ-ity. Antibody directed against cadherin, which failed to perturb strain-dependent BNP gene promoter activity, similarly was without effect on ERK. A similar inhibition was observed when p38 MAPK (Fig. 5C) and, to a lesser extent, JNK (Fig. 5B) were examined in parallel. Thus, interference with integrin function cannot be linked to suppression of BNP gene promoter activity exclusively through any one downstream MAPK signaling pathway. Rather, it appears that multiple pathways are impacted, supporting the hypothesis (29,30) that there is a functional cooperativity among these pathways that links hypertrophic stimuli to enhanced transcriptional activity of the BNP gene. DISCUSSION A number of studies (27,28,31) have documented an important role for the extracellular matrix in the regulation of gene expression. Such regulation has important implications for development where cell movement and, by inference, sequential formation and disruption of matrix attachments play a major role in morphogenesis. It is also important in differentiated cells where physiological function is dependent upon specific cell-matrix attachments (e.g. leukocyte adhesion to blood vessel walls or osteoclast attachment to remodeling bone surfaces).
In the cardiovascular system, the extracellular matrix is known to play an important role in modulating cellular function (12,32,33). In addition, because of the dynamic (i.e. pulsatile) environment in which these cells reside, considerable interest has focused on the possibility that formation/disruption of cell-matrix attachments could play an important role in signaling information regarding cellular deformation (e.g. that resulting from increased stretch or shear stress) from the cell surface/cytoskeletal apparatus to the nuclear compartment.
The major findings of this study include demonstration that 1) matrix proteins, including fibronectin, play an important role in supporting both basal and strain-induced BNP gene transcription in ventricular myocytes; 2) selected integrins present on the surface of cardiac myocytes participate in the activation of a number of signaling pathways that have been linked to the strain response as well as myocyte hypertrophy; and 3) interference with the function of these integrins leads to suppression of strain-dependent BNP gene transcription.
FIG. 4. Effect of chimeric integrins on hBNP promoter activity in neonatal rat ventricular myocytes. Cells were cotransfected with 1 g of Ϫ1595 hBNPLUC, 1 g of Rous sarcoma virus-␤-galactosidase, and different concentrations (0.1-5 g) of chimeric ␤ 1 , ␤ 3 , ␤ 3 B, ␤ 5 , or ␣ 5 integrin. After 24 h of culture, cells were subjected to strain for 48 h. Cells were then collected and assayed for luciferase and ␤-galactosidase activities. Luciferase measurements were normalized for ␤-galactosidase activity within a given sample. Data represent the means Ϯ S.D. from five separate experiments. *, p Ͻ 0.01 versus strain control; #, p Ͻ 0.05 versus strain control. ␣ 5 , and ␤ 1 integrins are up-regulated after induction of cardiac hypertrophy in the adult rat (34). ␤ 1 integrin has been shown to be important for the hypertrophic response of neonatal cardiac myocytes to ␣-adrenergic agonists in vitro (33). ␤ 3 integrin has been linked to assembly of a Src signaling complex (along with focal adhesion kinase) during ventricular hypertrophy in vivo (35). ␤ 1 has also been implicated as playing an important role in cardiac development (32).
We have identified ␤ 1 , ␤ 3 , and ␣ v ␤ 5 on both myocytes and fibroblasts in our cultures. Of note, unlike MacKenna et al. (38), we did not identify ␣ 5 in our fibroblast cultures. This could reflect differences in the integrin expression profile in the two experimental systems or, alternatively, limited efficacy of our anti-␣ 5 antibody to interact with the surface integrin. Of note, high concentrations of our anti-␣ 5 antibody appeared to effect a modest reduction in strain-dependent BNP promoter activity, implying that at least small amounts of this integrin are present in our cultures.
The presence of ␤ 1 , ␤ 3 , and ␣ v ␤ 5 integrins on both myocytes and non-myocytes in our cultures raises the obvious possibility that the antisera or dominant-negative mutants might be targeting integrins present on the non-myocytes, secondarily suppressing a paracrine activator of BNP gene expression in the myocytes. Such paracrine activators exist (39,40); however, they account, at best, for ϳ50% of the hBNP promoter response to mechanical strain. The fact that the inhibition with antiintegrin antisera exceeded 50% implies that a significant portion of the integrin-dependent activity operates directly at the level of the cardiac myocyte. Furthermore, we have employed the same anti-integrin antibodies in cultures depleted of nonmyocytes through treatment with bromodeoxyuridine, and we continue to see near-total inhibition of the response to strain (data not shown), again supporting an effect directly at the level of the cardiac myocyte.
It appears that activation of the hBNP gene promoter in our cultures is heavily dependent upon interaction of cellular integrins with fibronectin or fibronectin-like proteins in the extracellular matrix. Both the RGD peptide, which harbors a soluble integrin-binding site, and soluble fibronectin itself inhibited the response to strain, presumably by competing for matrix attachments. Morphologically, the cells appeared normal in each case and continued to adhere to the culture surface, indicating that attachment to the surface alone does not confer sensitivity to strain and implying that formation and/or disruption of selected matrix attachments may trigger specific signals that lead to changes in cellular phenotype. Although fibronectin clearly appears to be important, other matrix proteins (e.g. vitronectin and collagen) also utilize the RGD recognition sequence for integrin association. In addition, cardiac myocytes under hypertrophy-promoting conditions have been shown to produce fibronectin, vitronectin, and collagens I and III (26,41); and rat cardiac fibroblasts have been shown to increase synthesis of fibronectin and collagen III in response to biaxial stretch (42). Thus, one or more of these matrix proteins may establish the key integrin attachments that are required to signal the response to strain.
Ingber (43) has suggested that generation of tension in the matrix-integrin-cytoskeletal network activates the signaling cascade(s) that lead to changes in cell shape, mitogenesis, and gene expression. The fact that functional neutralization of several different integrins (through exogenous antibodies or transfected dominant-negative mutants) effected a substantial reduction of the hBNP transcriptional response to strain supports the notion that tension within the matrix-integrincytoskeletal assembly per se drives the response, presumably through integrin-dependent signal transduction pathways. The data would suggest that there is a quantitative threshold of integral cellular tension that must be maintained to support the transcriptional response. If this threshold is not maintained, the relevant signaling mechanisms fire imperfectly or not at all, thereby abrogating the response.
Numerous signal transduction pathways have been shown to activate in response to integrin ligation, including Ras (44), non-receptor tyrosine kinases (e.g. Src and Fyn) (45,46,53), focal adhesion kinase (45,47,53), ERK (46, 48 -50), JNK (47,50), and p38 MAPK (51). Importantly, we have show previously that ERK, JNK, and p38 MAPK are each activated following application of strain to our myocyte cultures (8,29), and each of them appears to play an important role in mediating the transcriptional response to strain. Clerk et al. (30) have noted similar activation of these three pathways following ␣-adrenergic induction of neonatal rat cardiac myocytes, providing support for a shared role in controlling downstream transcriptional activity. This suggests that there may be a functional interaction among these pathways that requires that all be operative to achieve an optimal response. It is noteworthy that MacKenna et al. (38) found that stretch activated ERK and JNK, but not p38 MAPK, in adult rat cardiac fibroblasts. In this case, the ERK response was mediated by RGD-directed, integrin-dependent activity, whereas the JNK response was not. The latter finding stands in contrast to our findings, implying important differences in the myocyte versus non-myocyte integrin-dependent signal transduction pathways. It is clear, however, that strain-dependent induction of JNK was only partially inhibited by integrin neutralization. Li et al. (52) reported that although shear stress induction of ERK activity in bovine aortic endothelial cells was completely blocked by inhibition of the integrin-associated focal adhesion kinase, the induction of JNK was only partially blocked. This may suggest that activation of ERK and possibly p38 involves an integrin/ focal adhesion kinase-signaled event, whereas JNK is activated by this as well as a second focal adhesion kinase-independent pathway.
In summary, we have shown that strain-dependent increments in hBNP gene promoter activity are critically dependent on specific matrix-integrin attachments established at the cell periphery. It is conceivable that perturbation of these attachments may serve as the primary stimulus that the cell senses in response to cellular deformation. Processing of this signal through the integrins, cytoskeletal proteins, or the various proteins associated with these structures (43) presumably triggers the multiple cellular events that are evoked in response to strain. Careful elucidation of the individual molecular participants in this process may provide us with a better understanding of the cellular and molecular events linking hemodynamic overload to cardiac hypertrophy.