INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE EXTRACELLULAR MATRIX (ECM) in the heart consists of interstitial collagens, proteoglycans, glycoproteins, and proteases that are arranged in a precise, three-dimensional network associated with myocytes and capillaries (4). Of these components, the arrangement of interstitial collagen has received the most attention because of the important roles of the collagen network in maintaining cardiac function (4). Cellular attachments to collagen, fibronectin, laminin, and other ECM components are mediated primarily by integrins (5, 17, 18). Integrins are transmembrane, heterodimeric proteins consisting of an - and -subunit. The -chains show considerable variability in their association with -subunits, which in the heart is primarily ₁. This variability of /-chain pairing is important in determining the specificity of the integrins for different ECM components (15, 18). The expression of integrins has been shown to vary with different stages of development and physiological condition (5, 7) and appears to be coordinated with the expression of ECM components (7). Whereas much is known concerning the expression and distribution of integrin receptors, little is known concerning the turnover of these receptors under different physiological conditions. However, recent data suggest that part of integrin, the ectodomain, may be proteolytically shed into the extracellular environment during times of rapid cell growth (10).

The outside-in signaling of integrins is critical to numerous cellular functions such as adhesion, proliferation, survival, differentiation, and migration (15). Clearly, the number and type of integrin receptors together with the availability of specific ECM substrates are important in determining which cellular functions are affected. The synthesis and insertion of new integrins into the membrane, removal from the cell surface, or both are possible mechanisms for controlling the number of available integrin receptors. New synthesis would require upregulation of expression and sorting of specific -chains to pair with excess ₁ in the cytoplasm and presentation of the new /-heterodimer in a precise location on the cell surface, which is not a very targeted mechanism. Cleavage at the cell surface, or shedding, would be an immediate method for removal of specific integrin-ECM contacts and would provide a focused mechanism for regulating specific functions. In addition, the shed fragment could bind to cells or ECM components or be involved in signaling biological events involved in cellular growth and remodeling.

Increasing biochemical and morphological evidence of ectodomain shedding from membrane-anchored proteins indicates that this may be a relatively common process (21, 26, 27). However, the physiological role of many of these shed ectodomains remains to be defined. The ectodomains can be detected in fluids associated with tissue injury and repair, which has led to the hypothesis that these proteins can function in wound healing, host defense, arthritis, and development (11). Extracellular proteinases may potentially be responsible for proteolytic cleavage of membrane proteins; however, only a few specific types have been identified. Recent studies indicate that matrix metalloproteinases (MMPs) and a disintegrin and metalloproteinase (ADAM) are potential candidates (2, 3, 19, 27). These studies, together with recent immunohistochemical data (10), have lead to the hypothesis that integrin shedding resulting from proteolysis is critical to cell growth. The goal of this study is to characterize the ₁-integrin shed fragment observed to be associated with the ECM (10) and to begin experiments aimed at defining a physiological role for the shed fragment.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Antibodies. Multiple monoclonal and polyclonal (AB314, AB1292, AB1134, and AB1501) antibodies were used to examine integrin shedding (Table 1). To confirm that the 55-kDa fragment was indeed a portion of ₁-integrin, two mouse anti-human ₁-integrin monoclonal antibodies (MAB2251 and MAB2252, Chemicon; Temecula, CA) that recognize amino acids 648-670 and 15-54, respectively, in the extracellular region of ₁-integrin were used in Western blotting.

View this table: [in this window] [in a new window]   Table 1.   Anti-1-integrin antibodies

Confocal microscopy and image analysis. The degree of ₁-integrin staining in the ECM of hearts undergoing hypertrophy and dilation was assessed using confocal fluorescence microscopy as previously described (10). Briefly, 100-µm vibratome sections from mouse hearts subjected to aortic stenosis (AS) (10) or hearts from tropomodulin overexpressing transgenic (TOT) mice (24) were stained with AB1134 (diluted 1:200) (16). Sections were rinsed with PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 2 mM KH₂PO₄; pH 7.4) followed by staining with a secondary antibody conjugated to Alexa 488 (Molecular Probes; Eugene, OR) to detect ₁-integrin and rhodamine phalloidin (1:20, Molecular Probes) to visualize F-actin. A final PBS rinse was followed by sections being mounted in a 1:3 mixture of glycerin-PBS and 10 mg/ml 1,4-diazabicyclo(2.2.2)octane (Sigma). Z-series images were collected on a Bio-Rad MRC 1024 confocal scanning laser microscope (Hercules, CA).

Cell isolation and culture. Cardiac myocytes were enzymatically dissociated from 3- to 4-day-old neonatal rat hearts as previously described (6). Briefly, myocytes were isolated by collagenase digestion (100 U/ml, Worthington; Lakewood, NJ) and separated from fibroblasts using a percoll (Sigma; St. Louis, MO) density gradient. Cells were resuspended in DMEM (Sigma) supplemented with 8% horse serum (GIBCO-BRL; Rockville, MD), 5% newborn bovine serum (GIBCO-BRL), 100 U/ml penicillin G (Sigma), 100 µg/ml streptomycin (Sigma), 1 µg/ml amphotericin B (Sigma), and 2 µg/ml cytosine arabinoside (Sigma). Myocytes plated on aligned collagen gels were cultured for 1 wk. Before experimentation, myocytes were serum starved in PC-1 media (Biowhittaker; Walkersville, MD) for 24 h.

Collagen substrate preparation. Thin, aligned collagen gels were prepared on 60-mm² culture dishes as previously described (22). A neutral collagen solution was prepared by combining 200 mM HEPES (pH 9), 10× minimum essential medium (Sigma), and bovine dermal collagen type I (Cohesion Technology; Palo Alto, CA) [1:1:8 (vol/vol/vol)] and placed on ice. The collagen stock solution was coated on dish surfaces and allowed to flow in one uniform direction. Gels were tilted at an angle in the direction of the collagen flow and allowed to polymerize for 1.5 h at 37°C, after which time they were rinsed with Moscona's buffer and dialyzed overnight.

Immunoprecipitation and purification of the ₁-integrin fragment. Serum-free medium (10 ml) was collected and concentrated to a final volume of 2 ml, followed by immunoprecipitation with 5 µl/ml AB314 and 100 µl/ml Protein G Sepharose (Amersham Pharmacia Biotech; Piscataway, NJ) before analysis by SDS-PAGE and Western blotting.

Phorbol ester perturbation. To determine the extent to which integrin shedding can be stimulated, myocytes and fibroblasts were plated on 60-mm² dishes and incubated for 24-48 h. Cells were serum starved for 24 h at 37°C before experiments. For both myocytes and fibroblasts, serum-free media were treated with phorbol 12-myristate 13-acetate (PMA; 100, 250, and 500 ng/ml, Sigma) for 24 and 48 h at 37°C. Controls consisted of treatment with only the vehicle. Media were collected as described above and analyzed for the shed fragment by Western blotting with AB314. The resulting bands were quantified using the GelDoc system (Bio-Rad) and changes in shedding are expressed as a percentage of the control.

MMP inhibition. To determine the extent to which integrin shedding can be inhibited, myocytes and fibroblasts were plated on 60-mm² dishes and incubated for 24-48 h. Cells were serum starved for 24 h at 37°C before experiments. For both myocytes and fibroblasts, the MMP inhibitor GM6001 (Chemicon) was added to serum-free media at concentrations of 5 nM, 50 nM, 500 nM, and 5 µM. After 24 h of culture, media were collected and analyzed for the shed fragment as described above.

Isolation of membrane proteins from AS mouse hearts. Atria were removed from freshly isolated hearts, and membrane proteins were isolated as previously described (16). Membrane proteins were solubilized in PBS containing 1% Triton X-100, 1 mM MnCl₂, and CompleteMini protease inhibitor (Roche; Mannheim, Germany). Protein concentrations were determined using the BCA assay described above. Equal amounts of protein (10 µg/sample) were loaded onto 4-15% gradient gels (Bio-Rad) and subjected to SDS-PAGE. For Western blotting, proteins were transferred to 0.45-µm nitrocellulose membranes (Bio-Rad) for analysis of ₁-integrin using AB1292.

Antibody production and purification. Polyclonal antibodies directed against the isolated 55-kDa protein were generated as previously described (25). New Zealand White rabbits were first bled to obtain preimmune antibodies and subsequently injected with 100 µg of the isolated protein. Antibodies were purified using a Protein A-Sepharose (Amersham Pharmacia Biotech) affinity column and the antibody concentration was determined spectrophotometrically.

Adhesion assay. Adhesion assays were used to measure the ability of myocytes and fibroblasts to attach to various ECM components (6). Wells of a 24-well plate were coated with ECM proteins (50 µg/ml collagen, 10 µg/ml laminin, and 10 µg/ml fibronectin) by incubation at 4°C overnight and subsequently rinsed with buffer 3 (137 mM NaCl, 5 mM KCl, 0.6 mM MgSO₄, 2 mM CaCl₂, and 10 mM HEPES; pH 7.4). Before cells were seeded, each well was incubated in 2 mg/ml BSA in buffer 3 for 1 h at 37°C to prevent nonspecific attachment. Myocytes and fibroblasts were preincubated with 100 µg/ml of the 55-kDa ₁-integrin fragment for 30 min at 4°C before being plated. Plates were rinsed again with buffer 3, and cells (500,000 myocytes and 200,000 fibroblasts) were added to each well and allowed to attach for 1 h at 37°C. Nonattached cells were removed by rinsing with buffer 3, after which 500 µl were added to each well. A standard curve was generated by plating 10,000-500,000 myocytes and 10,000-200,000 fibroblasts into wells. Substrate buffer [50 mM sodium citrate (pH 5.0), 0.25% Triton X-100, 4 mM p-nitrophenyl-N-acetyl--D-glucosaminide; Sigma] was added to each well and the plate was incubated for 3.5 h at 37°C (1.5 h for fibroblasts). After incubation, aliquots were transferred to wells in a 96-well plate and development/stop buffer (50 mM glycine and 5 mM EDTA; pH 10.4) was added. The absorbance at 405 nm was read immediately using a Bio-Rad Benchmark microtiter plate reader. The number of cells attached was determined by comparison to the standard curve and expressed as the total number of cells attached.

ELISA to examine 55-kDa fragment binding to the ECM. The wells of a 96-well plate were coated with various ECM proteins (50 µg/ml collagen, 10 µg/ml laminin, and 10 µg/ml fibronectin) overnight at 4°C, followed by PBS-T washes and blocking in 2 mg/ml casein for 1 h at 37°C. The 55-kDa shed fragment (100 µg/ml) was added and incubated with ECM substrates for 1 h at 37°C. The wells were rinsed with PBS-T and incubated with AB1501 diluted 1:250 in PBS-T for 1 h at 37°C. Wells were washed twice with PBS-T and incubated with donkey anti-rabbit IgG antibody conjugated to horseradish peroxidase (1:5,000, Amersham Pharmacia Biotech) for 1 h at 37°C. Wells were washed twice in PBS-T, and detection was accomplished using 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) tablets (Sigma) dissolved in citrate-phosphate buffer (pH 5.0). Absorbance was read at 405 nm using the Benchmark microtiter plate reader (Bio-Rad).

Statistical methods. Statistical significance was determined using Student's t-test with significance at P < 0.05. All values are reported as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Confocal microscopy demonstrating ₁-integrin immunoreactive material in the ECM. By confocal microscopy, ₁-integrin staining was observed at the Z lines on cardiac myocytes and in the extracellular space of mouse hearts from two different models (Fig. 1). Stacked Z-series images of hearts stained with AB1134 suggest that the immunoreactive material in the extracellular space is not bound to any cell surface. An increase in staining was observed in mouse hearts that developed cardiac hypertrophy in response to AS (10) (Fig. 1, A and B). In the early stages of hypertrophy (4 wk post-AS), ₁-integrin localization was increased in the extracellular space compared with control animals (data not shown). During this time, myocyte size has been previously shown to increase compared with control (10). After 8 wk post-AS with continued hypertrophy, increased staining was observed prominently around areas of myocyte branching (Fig. 1A) and the intensity of staining in the extracellular space was visibly greater than age-matched controls (Fig. 1B). The amount of immunoreactive material detected in the ECM of 8 wk post-AS animals was greater than that observed at 4 wk post-AS (Table 2).

View larger version (177K): [in this window] [in a new window]   Fig. 1.   Confocal microscopy analysis of 1-integrin shedding. Representative compressed Z-series confocal micrographs demonstrating the presence of 1-integrin immunoreactive material (green) in the extracellular matrix (ECM) of aortic stenosis (AS; A and B) and tropomodulin-overexpressing transgenic (TOT; C and D) mouse hearts stained with AB1134. Post-AS hearts (8 wk; A) showed both increased myocyte size and elevated amounts of 1-integrin immunoreactive material in the ECM compared with age-matched sham controls (B). TOT mice (C) demonstrated an elongation and thinning of myocytes compared with transgene-negative littermate controls (D) with a lower prevalence of 1-integrin immunoreactive material in the ECM.

View this table: [in this window] [in a new window]   Table 2.   Quantitation of 1-integrin shedding during cardiac disease

Biochemical characterization of the shed ₁-integrin fragment. By Western blotting with AB1292, a low-molecular-mass ₁-integrin positive material was observed in the initial pellet of membrane preparations from 8-wk AS hearts, whereas only intact integrins were associated with the membrane fraction (Fig. 2A). To examine integrin shedding in vitro, the media from cultures of neonatal cardiac fibroblasts and myocytes were immunoprecipitated with AB314 to isolate the shed ₁-integrin fragment. Media from myocytes grown on aligned collagen or from fibroblasts showed the presence of a ₁-integrin immunoreactive component with a relative molecular mass (M_r) of 55 kDa under nonreducing conditions (Fig. 2B). This 55-kDa material was detected with AB1292 prepared against intact ₁-integrin and AB1501 against the 55-kDa fragment (Fig. 2B). The higher-molecular-mass proteins observed in these blots are attributed to intact ₁-integrin. To confirm that the 55-kDa fragment was shed from ₁-integrin, commercial monoclonal antibodies with defined epitopes at the NH₂-terminal (amino acids 15-54, MAB2252) and COOH-terminal (amino acids 648-670, MAB2251) regions of the ₁-integrin extracellular domain were used (Fig. 2C). Both antibodies recognized the purified shed fragment in Western blotting, indicating that this fragment corresponded to the extracellular region of rat ₁-integrin. In addition, two-dimensional SDS-PAGE, followed by Western blotting with AB314 was also used to confirm the 55-kDa fragment was derived from ₁-integrin (Fig. 2D). ₁-Integrin fragments from two-dimensional gels were subsequently subjected to trypsin digestion and mass analysis, producing molecular masses that would be expected for ₁-integrin (data not shown).

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Western blot analysis for shed 1-integrin. A: Western blot of membrane preparation fractions from AS and control mice using AB1292 demonstrates the presence of a low-molecular-mass 1-integrin fragment in the pellet fraction not observed in the membrane fraction. B: myocyte and fibroblast culture media were subjected to immunoprecipitation with AB314 and subjected to Western blot analysis for the shed 1-integrin fragment. C: Western blotting of the purified 55-kDa shed fragment using commercial monoclonal antibodies that recognize amino acids 15-54 (MAB2252) or 648-670 (MAB2251) in the extracellular domain of human 1-integrin. D: Western blot of shed fragment subjected to two-dimensional SDS-PAGE with AB314 detected one spot with an approximate molecular mass of 55 kDa.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Effect of phorbol 12-myristate 13-acetate (PMA) on integrin shedding by cardiac myocytes and fibroblasts. Myocytes (A) and fibroblasts (B) were treated with increasing concentration of PMA or control with vehicle only (c w/v), and the degree of shedding was determined after 24 and 48 h of culture. Although PMA did not stimulate integrin shedding from fibroblasts at either time point, myocytes demonstrated statistically significant (P < 0.05) increases in 1-integrin shedding at both time points.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Role of matrix metalloproteinases (MMPs) in mediating 1-integrin shedding. Myocytes (A) and fibroblasts (B) were treated with varying concentrations of the MMP inhibitor GM6001 and the amount of the shed 1-integrin fragment was determined. Whereas GM6001 had no significant effect on myocyte-mediated integrin shedding, treatment of fibroblasts resulted in a statistically significant (*P < 0.05) dose-dependent inhibition of 1-integrin shedding.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Role of the 55-kDa shed 1-integrin fragment in mediating cellular adhesion. Myocytes (A) and fibroblasts (B) were incubated in the presence of the 55-kDa 1-integrin fragment at 100 µg/ml to determine whether the fragment would alter cellular attachment to ECM components. In the case of myocytes (A), the fragment promoted adhesion to collagen (P < 0.05) but did not significantly affect adhesion to laminin or fibronectin. Although the fragment appears to decrease fibroblast adhesion to all substrates tested (collagen, laminin, and fibronectin), these decreases were not significant compared with control (P > 0.05) using Student's t-test.

View larger version (11K): [in this window] [in a new window]   Fig. 6.   1-Integrin fragment specificity for ECM components. Collagen, laminin, and fibronectin were assayed for their ability to bind the shed 1-integrin fragment using AB1501. Collagen clearly demonstrated the highest degree of interaction with the shed fragment, followed by laminin and then fibronectin. Binding to all substrates was significantly increased compared with control (no substrate), with *P < 0.05 as determined by Student's t-test.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

While receptor-mediated cellular interactions with the ECM are dynamic events, the orderly movement and organization of cells during development and cellular adaptation to pathophysiological signals requires that cell-ECM contacts change in a very concerted manner; however, the mechanisms that regulate such behavior are unknown. Recent studies (10) suggest that the loss of integrins from the cell surface may provide one mechanism to modulate cell-ECM interactions. The data presented here demonstrate that the extracellular domain of ₁-integrin can be shed into the extracellular environment both in vivo and in vitro. Both cardiac fibroblasts and myocytes demonstrate shedding in vitro; however, the regulation and function of the shed ectodomain may be different in each cell type.

During cardiac disease, myocytes undergo shape and size changes. Interstitial collagen attaches to the cell at or near the Z line (8), and for the cell to change in size or shape these sites of collagen attachment must be modified. Analysis of the confocal data presented indicates a significant difference in the amount of shed integrin in the extracellular space surrounding hypertrophying myocytes (AS model) compared with elongated myocytes from the TOT model (Table 2). During cardiac hypertrophy, integrin expression increases (25) and, although these cells are changing in size, the increased amount of integrin present on the cell surface could accommodate the increased shedding of ₁-integrin observed in the AS model. As myocytes elongate during cardiac dilation, the surface area of these cells increases (unpublished observations), which could spatially segregate integrins from cell surface enzymes that may be responsible for integrin shedding. This protein segregation would lead to the reduced levels of integrin shedding observed in the TOT model. The data presented demonstrate that integrins are shed into the ECM, allowing for modulation of cell shape (Figs. 1 and 2). These data indicate that myocyte cell growth does not involve total cellular release from the ECM, but in fact these cells maintain some degree of constant contact with the ECM. This contact with the ECM is necessary for cell viability as release from the ECM can lead to anoikis, an anchorage-dependent form of apoptosis (13). It has recently been proposed that anoikis may be responsible for the slight increases in myocyte apoptosis observed in the late stages of cardiac hypertrophy when abnormal myocyte connections with the ECM were observed (10).

The enzyme(s) responsible for ₁-integrin shedding remains unknown. Two protein families believed to be essential in shedding are MMPs and ADAMs. Several members of both families have been shown to be present in the heart, but which enzymes are involved in integrin shedding remain unclear. Fibroblast shedding of the ectodomain was significantly reduced by GM6001, which inhibits MMP-1, -2, -3, -8, and -9, whereas myocyte shedding was not affected (Fig. 4, A and B). This would allow each cell type to respond differently to the same signal. Further experimentation is clearly necessary to determine the role of specific proteases on each cell type. Recently, ADAM-12-mediated cleavage of heparin-binding epidermal growth factor has been demonstrated to have a role in cardiac hypertrophy (1). Although integrins have been shown to associate with specific members of the ADAM protein family through the disintegrin domain (9), no evidence for cleavage of an integrin by ADAM proteins has yet been reported. A related metalloproteinase containing a disintegrin domain, jararhagin, has been shown to directly bind ₂₁-integrin and specifically cleave the ₁-integrin subunit (20).

The function of the 55-kDa shed ₁-integrin fragment remains unclear. Potentially, the fragment may provide a feedback signal to regulate receptor distribution on the cell surface. Another potential function supported by our data is that the shed fragment binds to specific ECM components in turn enhancing or inhibiting some cellular functions. The data presented demonstrate that, whereas the 55-kDa ₁-integrin fragment slightly increased myocyte adhesion to collagen, this fragment has no significant affect on fibroblast adhesion to ECM proteins (Fig. 5, A and B). These observed differential effects of the shed fragment may be due to the presence of different ECM receptors on the cell surfaces of myocytes and fibroblasts through which the fragment may be altering cellular adhesion. The increase in myocyte attachment may be due to the 55-kDa ₁-integrin fragment binding collagen and then interacting with other receptors on the surface of myocytes, therefore increasing the number of cell-ECM interactions. In addition, the shed ₁-integrin fragment is capable of interacting with collagen, laminin, and fibronectin (Fig. 6) independent of association with other proteins. Previous studies examining ADAM-17-mediated shedding of TNF- demonstrated that as a result of shedding and binding to fibronectin, ECM-bound TNF- was capable of inhibiting chemotatic stimulated migration of T cells (12). Whether similar mechanisms occur that regulate fibroblast movement or some aspect of myocyte-ECM interactions remains unclear.

Integrin shedding represents an exciting, novel mechanism by which cell-ECM interactions can be modulated. The data presented demonstrate potential functional roles that the shed ₁-integrin fragment may exert on both cardiac myocytes and fibroblasts. Further experiments are necessary to clarify the functional role of shed integrins in the modulation of myocyte and fibroblast interactions with the ECM, to determine whether specific -chains are involved in the shedding process and to identify the enzyme(s) responsible for ₁-integrin shedding.