This study was designed to examine how smooth muscle (SM) cell (SMC) isolation affects the distribution of some adherens junction (AJ) complex-associated proteins. Immunofluorescence procedures for identifying protein distribution were used on gastrointestinal and tracheal SM tissues and freshly isolated SMCs from dogs and rabbits. As confirmed by force measurements, relaxation, Ca2+ depletion, and cholinergic activation of SM tissues do not cause significant redistribution of the AJ-associated proteins vinculin, talin, or fibronectin away from the plasma membrane. Unlike SMCs in tissue, freshly isolated SMCs show a variable peripheral/cytoplasmic vinculin and talin distribution that is not altered by activation. Enzymatic treatment of SM tissues (as done for the first step of SMC isolation) results in loss of fibronectin immunoreactivity in SMCs still in the tissue but fails to cause redistribution of vinculin, talin, or caveolin away from the periphery. The loss of fibronectin immunofluorescence with enzymatic digestion correlates significantly with loss of tissue force production. These results confirm that the AJ-associated proteins vinculin and talin do not redistribute throughout SMCs in tissues when relaxed, when generating force, or after enzymatic digestion. In addition, in freshly isolated SMCs, the distribution of these proteins is significantly altered in ∼50% of the SMCs. The cause of this redistribution is currently unknown, as is the impact on intracellular signaling and mechanics of these cells. Use of these two systems (SMCs in tissues vs. freshly isolated SMCs) provides an ideal situation for studying the role of the AJ in SMC signaling and mechanics.
smooth muscle (SM) adherens junctions (AJs) are located at the plasma membrane and include a complex association of cytoskeletal, extracellular, and integral membrane proteins, kinases, phosphatases, GTPase modulators, and other enzymes that allow for cell-matrix, cell-cell, and cytoskeletal-contractile protein associations that bind and transmit force between these constituents (see Refs. 5, 12, 19, 23, 32, 39, and 55). The analogous structure in cultured cells is the focal adhesion (13). Integrins are single transmembrane-spanning receptors that connect the extracellular matrix to the cell’s cytoskeleton at the AJ (3, 23, 44) and are involved in inside-out and outside-in signaling (11, 24, 35, 43, 52). On the cytoplasmic face of the AJ a large group of cytoskeletal proteins including vinculin, talin, actin, filamin, calponin, tensin, α-actinin, and plectin are known to be associated with integrins and critical for the function of the AJ (46, 51). On the extracellular face, fibronectin, vitronectin, collagen, and laminin are known to associate with integrins (21, 40, 44). All these cytoskeletal proteins are believed to be involved, along with other proteins, in the linking of actin filaments/stress fibers to the integral membrane proteins and ultimately to the extracellular matrix and neighboring cells (20). SM force generation resulting from the interaction of contractile proteins must be coupled to the AJ at the cell membrane, and ultimately to neighboring cells for hollow organs to function. This may be especially critical in the digestive tract, where there can be major changes in lumen diameter that put enormous strain on intercellular connections.
Interdispersed between the AJs on the plasma membrane are caveolin-rich domains referred to as caveoli. Caveoli are invaginated membrane regions that are rich in glycosphingolipids, cholesterol, and numerous cell signaling molecules that are involved in endocytic and exocytic processes as well as cell regulation (1). Caveoli can be readily identified at the light microscope level by immunoreacting for the protein caveolin, which is localized to these areas of the membrane. Although the unique alternating distribution of caveoli and AJs in SM appears invariant (16–18, 36, 41, 45, 46, 49), the exact interactions between these two structures remain unclear.
The functions of AJ-associated proteins are complex and not well defined. Vinculin has been reported to be in equilibrium between cytosolic and cytoskeletal pools in chick embryo fibroblasts (30). Using tissue homogenization and fractionation, Kim et al. (27) reported that cholinergic stimulation of bovine trachea SM resulted in simultaneous increases in force and recruitment of α-actinin, talin, and metavinculin (but not vinculin) from the cytosolic fraction to the cytoskeletal fraction (indicative of a cytosolic to plasma membrane translocation). Opazo Saez et al. (37), using indirect immunofluorescence, also reported that cholinergic activation of isolated canine tracheal smooth muscle cells (SMCs) results in translocation of vinculin, talin, paxillin, and focal adhesion kinase (FAK) from the cytoplasm to the plasma membrane. These results suggest that vinculin, talin, and other cytoskeletal proteins are highly dynamic in their SM cellular distribution in intact SM tissues and isolated cells. Both talin and vinculin are reported to readily dissociate from focal adhesions in permeabilized cells and show high solubility in the presence of nonionic detergents (5). In addition, numerous proteins exhibit altered expression and changes in localization that occur with cell isolation and culturing (see discussion). This suggests that the association of AJ proteins is labile.
In contrast, we (16) have reported that the AJ-associated proteins vinculin, talin, and fibronectin are stably localized near the membrane of SMCs in gastrointestinal organs under relaxed and activated conditions, suggesting that these proteins are very stable at the cell periphery in intact organs. However, many laboratories (including our own) use enzymatic treatment of SM tissue to isolate SMCs for a wide range of in vitro and cell culture experiments. Thus it is not clear whether the AJ-associated proteins are stably localized at the SMC periphery in isolated SMCs or whether the focal adhesion in cultured SMCs is the equivalent of the AJ in SM tissues.
This study set out to determine 1) whether there are differences between the cellular distributions of some AJ-associated proteins (vinculin, talin, fibronectin) in SM tissues vs. isolated SMCs, 2) whether there are mechanical implications for differences in cellular localization of these proteins in SM tissues, and 3) whether relaxation and/or stimulation alter the cellular distribution of these proteins. The results show that 1) although the distribution of the AJ-associated proteins vinculin and talin remains stably localized near the plasma membrane in SM tissues, these proteins are readily observed to be distributed throughout the cytoplasm of many SMCs after their isolation; 2) the loss of fibronectin immunoreactivity correlates with the loss of force production in digested SM tissues; and 3) in tissues and single SMCs these proteins do not translocate to/from the membrane with stimulation/relaxation.
Both dog and rabbit stomach, ileum, colon, and tracheal SM tissues were used in this study in an attempt to determine the universality of our results. Because of the large differences in protein expression, cell coupling, organization, etc., between different SM organs, this study used four organs from two species that are commonly used in SM research. Despite the major differences between these organs, the results reported here for any given species/organ were consistent between all of them. Figures show representative results from one or more of the species/organ sources and are applicable to all the other organs reported here unless stated otherwise.
Organ Handling and Cell Isolation
Experimental procedures were approved by the Institutional Animal Care and Use Committees of the Medical College of Wisconsin, the Zablocki Department of Veterans Affairs (VA) Medical Center (dogs), and Marquette University (rabbits). Organs were harvested from euthanized dogs (n = 25) used at the VA Medical Center after acute vascular studies (performed on the rear leg). Immediately after the experiment, the animal was euthanized by an overdose of anesthetic and KCl and the trachea, stomach, ileum, and colon were removed from the animal and put in cold physiological salt solution [PSS; in mM: 140.1 NaCl, 4.7 KCl, 1.2 Na2HPO4, 2.0 MOPS (pH 7.4), 0.02 Na2EDTA, 1.2 MgSO4, 1.6 CaCl2, and 5.6 glucose]. Rabbit organs (trachea, ileum, colon, stomach) were removed after CO2 asphyxiation (n = 25). Each experimental procedure discussed was done on at least 5 rabbits and as many as 15. Tissues from the respiratory and digestive system were used to confirm the results in these two often-used model tissues. Organs were cleaned of blood, loose connective tissue, and in some cases mucosa and frozen in isopentane cooled in liquid nitrogen or stored in PSS in the refrigerator for 0–2 days. Some organs were fresh frozen as soon as possible after animal death (30–90 min). Some organs were incubated in PSS or Ca2+-free PSS (PSS with no added Ca2+, twice the MgSO4, and 0.5 mM EGTA) before rapid freezing. Stored tissues were enzymatically digested with papain similarly to the first step used for isolating single SMCs (14) as modified (15). The digesting solution had 20 IU papain/ml and 2 mM DTT in a nominally no-Ca2+ PSS (no added Ca2+ and twice the MgSO4). Digestions were done on tissue with hydrostatic pressure as originally reported by Driska and Porter (14) or done on tissue strips in a tissue bath attached to a force transducer (see Mechanical Measurements below).
After enzymatic digestion, tissue was either frozen immediately (in liquid nitrogen-cooled isopentane), 1) activated with 10 μM carbachol (CCh; Acros Organics), or 2) relaxed in normal PSS solution or Ca2+-free PSS solution at 37°C for 0–3 h before freezing. Antagonists for tissue-activating solutions included phentolamine and propranolol (1 μM each; Sigma, St. Louis, MO), to minimize the effects of neuronal or endogenous activators. Variable incubation times and agonist concentrations were also tested. Digestion times (5–30 min), postdigestion duration before freezing (0–3 h), postdigestion solutions (PSS, Ca2+-free PSS), and CCh agonist concentrations (10 or 100 μM) were all varied to test for a range of effects on the AJ-associated protein distribution. All tissue was stored frozen until sectioned and immunoreacted. Five- to ten-micrometer sections of the frozen tissues were cut on a Leica CM1900 cryostat, picked up on glass slides, and stored frozen (1–3 days) until immunoreacted.
SMCs were isolated from the digested tissue by teasing the tissue apart. SMCs from a given isolation were aliquoted and either stimulated with 10 μM CCh or with 10 μM phorbol 12,13-dibutyrate (PDBu; Sigma) while suspended in solution and then attached to a cover glass (stimulated-attached; see controls below) or attached to a cover glass and then stimulated (attached-stimulated). In the former situation there is no external load to prevent cell shortening on agonist activation, whereas in the latter the cell is attached to the cover glass, providing an external load to limit shortening. In the stimulated-attached SMCs, the cells shortened significantly on stimulation with 10 μM CCh or with 10 μM PDBu (n = 3; unpublished data). The isolated SMCs were then processed for immunohistochemistry in an manner identical to that for the tissue sections as stated below.
The antibodies used were obtained from the following sources: talin (8D4), vinculin (VIN-11–5), and fibronectin (IST-3) from Sigma; caveolin-1-BD from Biosciences/Pharmagen (San Diego, CA); PKC-α (H-7 and C-20) from Santa Cruz Biotechnology (Santa Cruz, CA); Cy2 and Cy3 donkey anti-mouse or -rabbit secondaries from Jackson ImmunoResearch (West Grove, PA); and Alexa Fluor 594-phalloidin and DAPI from Molecular Probes (Eugene, OR).
Frozen tissue sections picked up on glass slides and isolated cells attached to cover glasses were fixed with 2% paraformaldehyde for 10 min, permeabilized in 0.5% Triton X-100 for 10 min, and blocked with 5 mg/ml BSA for 1 h before reacting with the primary antibody for 1 h and then the appropriate secondary antibody for 1 h at room temperature. After the secondary antibody, the tissues were incubated with DAPI (0.5 μM), phalloidin (10–50 nM), or DAPI-phalloidin as appropriate for staining nuclei and/or filamentous actin. Multiple washes were used after the primary and secondary incubations. Cover glasses were mounted with buffered 75% glycerol with 0.2% n-propyl-gallate to minimize fading. All immunoreacting solutions were made in PBS-Tween (in g/l: 8.0 NaCl, 0.2 KH2PO4, 1.15 Na2HPO4, and 0.2 KCl with 1% Tween 20, pH 7.4) with 0.1% BSA. Negative controls included leaving out the primary antibody or using serum in place of the primary antibody.
Sections were observed with an Olympus IX70 microscope with epifluorescence illumination. Digital images were taken with a 16-bit Princeton Instruments (Princeton, NJ) charge-coupled device camera controlled through a PCI board via IPLab for Windows on a PC (version 3.6, Scanalytics; Fairfax, VA). Images were taken with a ×100 [1.3 numerical aperture (NA)] or ×60 (1.25 NA) oil lens or a ×40 (0.9 NA) air lens and stored on the PC. Emission filters used were 405, 490, and 570 nm. Montages were assembled in Adobe Photo Shop (version 6.0, Adobe Systems, San Jose, CA). Sections were also viewed with a Zeiss confocal microscope (Axiovert 200 with LSM 5 Pascal software). The objectives used were ×63 (1.4 NA), ×40 (1.3 NA), and ×25 (0.8 NA) oil lenses and emission filters at 420, 505, and 560 nm. Similar results were observed in immunofluorescence distributions with these two different systems.
Image Analysis of Protein Distribution of Single Cells in Tissues
For cells in tissues, profiles of fluorescence intensity were taken for individual cells in transverse tissue sections. Sections were observed at low magnification (×10–20) to avoid areas of apparent artifacts (tissue folding, freeze damage, etc). In an artifact-free area the magnification was increased to level (×40–100). A picture was taken, and a single cell chosen in a given field with all the cells adjacent to it in a cluster, or all cells that fell on a line drawn across the field, were measured (up to the desired n) to minimize potential subjective bias. The region of interest (ROI) for analysis was a single-pixel-width line drawn across the diameter of the cell and through the AJs on each side of the cell. The ratio of the peak value for vinculin at the AJ divided by the minimum value in the cytoplasm was used. Cells with nuclei present in the section were not included in data analysis.
Image Analysis of Protein Distribution in Isolated Single Cells
An isolated cell was chosen from a given field with a Zeiss confocal microscope. Z-stack series were taken in three color channels (phalloidin, red; AJ-associated proteins or PKC-α, green; and DAPI, blue). Approximately 1-μm-thick Z sections were taken, resulting in 10–15 Z sections being taken for each cell(s) in the field (to cover the entire depth of the cell). Each Z series was examined for each cell to identify the center Z image, and this was converted to a TIFF file, which was then imported into IPLab for Windows on a PC (version 3.6, Scanalytics). Additional cells were chosen by rastering across the slide and taking a complete Z-stack series of pictures of all cells encountered.
The intensity profiles from two ROIs perpendicular to the long axis of the SMC were analyzed to determine protein distribution for each respective protein as identified by its respective fluorophore. Control experiments for identifying a valid measurement protocol included testing 1) the number of pixels for the width of the ROI used for the intensity profile, 2) the position of the ROI in the cell from the nucleus to the tip of the cell, 3) variability from one half of the cell to the other, 4) the number of ROIs per cell to analyze, and 5) how specific locations for “membrane-associated” and “cytosolic” intensities would be chosen. The final protocol used for quantifying the peripheral/cytosolic intensity of a protein in isolated SMCs for this study was as follows. 1) Confocal images were taken of all cells observed when rastering across the slide. 2) Data would only be analyzed from isolated SMC pictures including a complete Z stack (∼background fluorescence in first and last section of the Z stack) with all three fluorophore channels. 3) SMCs with normal morphology were analyzed (i.e., completely rounded up cells or cells that appeared sheared or otherwise mechanically damaged were not analyzed). 4) The center Z stack of the series corresponding with the center of the cell would be used for intracellular protein distribution analysis. 5) Intensity profiles from two ROIs (10 pixels by 150 pixels centered on the cell’s width) were used. The distance from the end of the nucleus to the end of the cell was divided into thirds, and placement of ROIs was made in the middle third region, thereby avoiding the perinuclear region and the very narrow tapered tip of the cell. 6) All pictures of cells or tissues were taken below camera saturation to allow quantitative analysis. The fluorescence intensity for each antibody/phalloidin was plotted as a intensity distribution. The first five pixels (∼0.7 μm) on either side of the cell where the phalloidin intensity increased sharply above background was defined at the edge of the cell, and the immunofluorescence intensities of the two brightest pixels for the protein of interest in these regions were averaged. Eight pixels from the center of the cell were averaged to estimate the cytosolic level. The ratio of the average edge intensity over the average cell center intensity was used to get a number for the peripheral/cytosolic intensity.
For determination of tissue cross-sectional area (CSA) from tissue strips used for force measurements (n = 6 animals, 4 organs), multiple pictures were taken to encompass the entire tissue CSA. Total tissue CSA was determined by thresholding (at a level that gave a profile that matched the fluorescent signal distribution; IPLab software) on the caveolin immunofluorescence, as the peripheral location of this protein is not altered by the papain digestion. Fibronectin immunoreactivity CSA was also measured by thresholding on its respective immunofluorescence channel to determine the CSA of the tissue that still had fibronectin fluorescence. As fibronectin immunoreactivity is diminished/eliminated by the papain digestion, the remaining area without fibronectin immunoreactivity after digestion, relative to the entire tissue CSA (area of caveolin immunoreactivity), allowed calculation of the percent tissue CSA digested. Thresholding for fibronectin was done independent of knowledge of the extent of tissue digestion. When specifically trying to determine the effects of papain digestion on tissue, regions without fibronectin immunoreactivity were chosen as being indicative of regions having been digested. Once identified, these same areas were located in serial sections that were immunoreacted for other proteins of interest. Examination of caveolin counterstaining of these tissue areas allowed elimination of regions with tissue-processing artifacts from analysis.
Immediately before use, tissue strips were cut and clamped on each end, with the clamps secured between hooks on a stationary metal rod and a metal rod hanging on an isometric force transducer (Harvard Apparatus, Holliston, MA), in PSS bubbled with 95% O2-5% CO2 in water-jacketed muscle chambers (Radnoti Glass Technology, Monrovia, CA) at 37°C. The length of each strip was varied by repositioning of the stationary metal rod.
SM tissue strips (stomach, ileum, colon, and trachea) were equilibrated for 1 h and stretched to a passive tension approximating optimal length, using an abbreviated length-tension curve. To contract tissues, PSS was replaced with K+-PSS (109 mM KCl and 70 mM NaCl substituted for 140 mM NaCl). The muscle strip was activated repeatedly with stretching of the tissue between each activation until peak force no longer showed significant increase over the previous contraction. Chambers were flushed three times with PSS after all tissue activations. At least two successive K+-PSS contractions were used to get a standard force trace, with a 20-min rest between each contraction. The tissues were then activated with 10 μM CCh and relaxed again as for the K+-PSS contractions. Subsequent steps varied depending on the question being asked. To determine whether activation or relaxation (as determined by the presence or absence of force generation) affects AJ-associated protein distribution, the tissues were activated a second time with 10 μM CCh, relaxed for 20 min, and then either rapidly frozen while relaxed or activated a third time with 10 μM CCh and rapidly frozen when peak force was obtained (see Fig. 1). To examine the effect of intracellular Ca2+ concentration depletion on the AJ-associated protein distribution, the tissues were activated with 10 μM CCh three additional times in a Ca2+-free PSS solution and then washed in Ca2+-free PSS for 2 h before rapid freezing (see Fig. 2). To examine the effect of papain digestion on AJ-associated protein distribution, the tissues were activated twice in 10 μM CCh, washed in PSS for 20 min, digested with papain solution for 10 min, washed in PSS for 20 min, activated again in 10 μM CCh to measure force production following the digestion, and rapidly frozen while still activated. Additional experiments included variations in postdigestion incubation times, with or without Ca2+ and with or without activation after these digestions. In total, these experiments were done with strips from 15 stomachs, 12 tracheas, 7 colons, and 6 ileums.
Analysis of Force Data
Statistical comparisons were carried out with the program Prism (GraphPad Software, San Diego, CA). Unpaired t-tests were performed on the means of peripheral/cytosolic intensity from tissues and cells that were relaxed or activated before rapid freezing or fixation. A paired t-test was used to compare the decrease in force with the loss of fibronectin immunofluorescence. Sample size is given in the text and figures. Values are means ± SD.
In a previous study (16) we reported that AJ-associated proteins remain localized near the SMC periphery in tissues incubated in Ca2+-depleted PSS or PSS and with CCh activation. However, this work was done on untethered tissues, and thus it was not possible to confirm the contractile status of the tissue. Therefore, we repeated these studies on tissue strips that were hung from force transducers and maintained at 37°C in tissue chambers to measure force. Figure 1 shows the force traces from representative dog tracheal strips used to determine the immunofluorescence distribution of specific AJ-associated proteins when the tissue was relaxed or activated. The tissues were rapidly frozen when the tissue was relaxed (no force generation; Fig. 1A, arrow) or when the tissue was activated (isometric force generation; Fig. 1B, arrow). Figure 1C shows the immunofluorescence distribution of vinculin, talin, fibronectin, and caveolin from the rapidly frozen relaxed (Fig. 1C, left) and rapidly frozen activated (Fig. 1C, right) tracheal strips shown in Fig. 1, A and B, respectively. In both the relaxed (PSS incubated) and contracted (10 μM CCh incubated) strips these four proteins maintained their peripheral distribution in the SMCs.
While SM tissues incubated in PSS do not generate significant force, they can still have basal tone and/or show spontaneous contractions, and thus it is possible that AJ-associated proteins only migrate away from their peripheral distribution when intracellular Ca2+ is below some threshold concentration. To test for this possibility, we did a Ca2+-depletion protocol. Figure 2 shows the force trace of a dog ileal strip that was activated twice with K+-PSS and then once with 10 μM CCh in normal PSS. The tissue was then transferred to a Ca2+-free PSS solution and activated three additional times with 10 μM CCh in a Ca2+-free PSS solution. The force generated by these contractions decreases significantly such that by the third Ca2+-free contraction there is only nominal force being generated. The tissue was then incubated in the Ca2+-free PSS for two additional hours before being rapidly frozen (Fig. 2, arrows). Figure 3 shows the immunofluorescence patterns of four proteins from four different dog SM tissues that were treated in this manner. In all SM tissues tested (trachea, ileum, colon, and stomach) the immunofluorescence pattern of vinculin, talin, fibronectin, and caveolin are localized primarily to the SMC periphery. There is no indication of a shift from the periphery toward the core of the SMC with Ca2+ depletion of these tissues.
In contrast to this consistent peripheral protein distribution in cells in tissues, the peripheral distribution of these same proteins can be altered significantly when SMCs are physically isolated. Figure 4 shows the vinculin distribution (green) in freshly isolated rabbit stomach SMCs in relaxing solution (PSS). About half of the SMCs in a cell isolation on a given day show a peripheral distribution of vinculin similar to that observed in SMCs in intact tissues, while the rest show a more uniform distribution of vinculin throughout the cell. In Fig. 4 the horizontal and vertical lines drawn through the main panel (Fig. 4A) indicate the location of the transverse cell image shown on the top and right (Fig. 4, B and C, respectively) of the main middle panel. The transverse image (made by compiling all the Z-stack sections) shows that these SMCs can have two very different possible vinculin distributions. Thus the uniquely peripheral distribution of these proteins in intact tissue can be altered in some isolated SMCs, becoming more uniformly distributed in many of the isolated cells.
To examine how isolated SMC stimulation might affect the protein distribution, we routinely divided a freshly isolated population of suspended SMCs into two fractions. One was fixed and immunoreacted, while the other was agonist stimulated before fixation and immunoreacting. Figure 5 shows that the vinculin (talin results are similar; data not shown) distributions are variable between SMCs. Figure 5A shows a central Z stack confocal image with parts of three SMCs in the same field that were incubated in PSS before fixation and immunostaining. One cell has a uniform vinculin distribution throughout the cell, while the two other SMCs have a distinctly peripheral distribution of vinculin. Figure 5C is a central Z-stack confocal image of two cells that were activated in PDBu for 15 min before fixation and immunostaining. As observed for isolated SMCs in PSS incubation, some SMCs have a uniform vinculin distribution while others have a distinctly peripheral distribution. Figure 5, B and D, show transverse sections of the cells shown in Fig. 5, A and C, respectively, at the point where the horizontal line is drawn across the cells. The transverse sections (a compilation of all the Z-stack sections) also show that the vinculin is distributed throughout some of the cells (green throughout the transverse section) and peripherally localized in others (green fluorescence ring around the red center) when relaxed (Fig. 5, left) and after PDBu activation (Fig. 5, right). These same distributions are observed for dog stomach smooth muscle cells (Fig. 6). Figure 6A shows freshly isolated dog stomach body SMCs immunoreacted for vinculin (green) and counterstained with phalloidin and DAPI (blue) in relaxing (PSS, Fig. 6A) or activating (PBDu, Fig. 6C) conditions. Figure 6, B and D, show transverse sections of the cells shown in Fig. 6, A and C, respectively, at the point where the horizontal line is drawn across the cells. Some SMCs show a peripheral distribution of vinculin, while others show a peripheral distribution, independent of relaxing or activating conditions.
Because there were apparent quantifiable differences between isolated SMCs relative to their vinculin distribution, we wanted a means of validation of these techniques and some determination of sensitivity of this method. This was done by the widely published translocation of PKC-α from the cytosol to the membrane in freshly isolated SMCs with PDBu stimulation (34, 35, 37, 38, 60, 61, 66). PKC-α immunofluorescence shows a relatively homogeneous distribution throughout the cytoplasm of freshly isolated SMCs when they are fixed without stimulation (Fig. 7A). When isolated SMCs from this same tissue preparation are activated with 10 μM PDBu for 15 min before fixation and immunoreacting, the distribution becomes preferentially localized at the cell’s periphery (Fig. 7C). Similar to Figs. 5 and 6, Fig. 7, B and D, show a transverse section of the cells shown in Fig. 7, A and C, at the point where the horizontal line is drawn across the cells. The transverse sections (a compilation of all the Z-stack sections) confirms that the PKC-α is distributed throughout the cells (green throughout the transverse section, Fig. 7B) and peripherally localized after activation (green fluorescence ring around the red center, Fig. 7D).
Quantitation of vinculin and talin distribution in SMCs in tissues was done to determine whether there are statistically significant changes in the distribution of these proteins with digestion. Figure 8 shows a set of double-labeled immunofluorescence pictures from dog colon longitudinal SM without and with enzymatic digestion before rapid freezing. Caveolin was used as a control for all tissues to verify that changes in immunofluorescence were not an artifact of the tissue section (data not shown). Images of transverse tissue sections for fibronectin (Fig. 8, left), vinculin (Fig. 8, center), and talin (Fig. 8, right) are shown. Figure 8A, top, is from a fresh frozen colon, and Fig. 8A, bottom, is from an enzymatically digested colon. Exposure times were held constant for each antibody, to allow relative changes in signal intensity to be determined. The fibronectin immunofluorescence was significantly altered by the enzymatic digestion, showing a dramatic decrease in signal intensity at the cell periphery where it normally resides in intact tissue. The vinculin and talin panels (Fig. 8, center and right, respectively) were taken from the same region of the tissue where the fibronectin pictures were taken, ensuring that the tissue was digested in this region as shown by the diminished fibronectin immunofluorescence signal. The patterns of the fluorescence signal for vinculin or talin show no apparent difference with enzymatic digestion. Figure 8B shows the fluorophore intensity profile of the distribution of the respective AJ-associated protein across the transverse section of the cell. A white line in each frame of Fig. 8A shows the respective representative cell used for the histogram shown in Fig. 8B. In the papain-digested tissue, the intensity profile for fibronectin shows a loss of signal at the periphery of the cell, with no compensatory redistribution elsewhere (Fig. 8B, bottom left). The intensity profiles for vinculin and talin (Fig. 8B, center and right, respectively) show no changes in the relative distribution from the periphery to the core of the SMC as a result of enzymatic digestion of the tissue.
For quantitation of the fluorophore intensity, signal intensity histograms were determined from control (fresh frozen) and enzymatically digested tissues. The ratio of the peak intensity at the cell periphery over the lowest value in the center of the cell (excluding areas of nuclei) was compared (see methods). Because fibronectin immunoreactivity is readily lost after tissue digestion (Fig. 8A), areas of digested tissue were selected for by the absence of fibronectin signal. Figure 9 shows there was no significant difference in the SMC peripheral/cytosolic vinculin distribution between undigested and digested dog stomach body and colon longitudinal SM tissues when this analytical approach was used. The variability in the vinculin distribution in SMCs in tissues with or without enzymatic digestion was fairly small (SD 0.4–1.0).
Quantitation of fluorescence signal was also done on isolated SMCs. When working with isolated cells it is critical that unbiased sampling techniques be used and that large numbers of cells be analyzed because of the extreme variability of distribution of these proteins between cells. In contrast to the routinely peripheral distribution of vinculin in SMCs in tissues, the vinculin distribution was highly variable in freshly isolated stomach SMCs from dog (Fig. 10A) and rabbit (Fig. 10B), as can be observed by the large standard deviation (SD 0.9–2.6). For each SMC isolation procedure, approximately half of the cells show a peripheral vinculin distribution and the rest show a roughly uniform distribution. Activation of these cells with CCh or PDBu has no significant effect on this distribution. By contrast, the PKC-α distribution in isolated SMCs in PSS is relatively uniformly distributed throughout the cytoplasm of all the SMCs after cell isolation (peripheral/cytosolic intensity ∼1, SD 0.3–0.4). This distribution is not altered with CCh activation but changes significantly with PDBu stimulation to become primarily peripheral (peripheral/cytosolic intensity ∼4, P < 0.05). The phalloidin ratio (indicative of filamentous actin distribution) for this same population of isolated SMCs is 0.80 ± 0.12 and always appears uniformly distributed throughout all the cells.
To determine the effect of the enzymatic digestion on SM tissue function, SM strips were attached to force transducers while maintained in a tissue bath at 37°C. The tissue was activated twice with K+-PSS, followed by two 10 μM CCh contractions. The tissue was then incubated in the papain enzyme solution for tissue digestion, activated again with 10 μM CCh to determine how much isometric force could be generated after the digestion, and then rapidly frozen for immunohistochemistry. The loss of the fibronectin immunofluorescence signal as a result of enzymatic digestion occurs from the periphery to the center of the tissue. Quantitation of the extent of tissue digestion by the enzyme treatment of tissue strips was determined by using the immunofluorescence signal for caveolin (total tissue CSA) and fibronectin (undigested tissue CSA). The caveolin immunofluorescence signal remains essentially unaltered by enzymatic tissue digestion, while the fibronectin immunofluorescence signal is diminished in direct correlation with the length of the digestion. Figure 11 shows the results of comparisons for the loss of isometric force due to enzymatic tissue digestion in dog trachea, stomach (circular), and intestinal (circular and longitudinal) SM tissues and the extent of the tissue digestion as measured by the loss of fibronectin signal. The correlation coefficient for the decrease of force and decrease in fibronectin immunofluorescence with enzymatic digestion in these tissues is r2 = 0.96.
SM force generation resulting from the interaction of contractile proteins must be coupled to the AJ at the cell membrane, and ultimately to neighboring cells for hollow organs to function. The organization of these proteins is affected by cell isolation and culturing. When SMCs from diverse organ systems (circulatory, urinary, respiratory, and gastrointestinal) are dissociated and put into cell culture they show a dramatic transformation from their “contractile” phenotype to a “synthetic” phenotype. The most profound physiological alteration in cultured SMCs is the loss of contractility, resulting in part from changes in contractile and cytoskeletal protein expression (see Ref. 38 for review). The cultured SMC phenotype is similar in many ways to that of SMCs in developing arteries or in pathological conditions such as atherosclerosis (6, 31) and may parallel changes with SMCs in gastrointestinal diseases. These changes have been extensively documented and include a decrease in SM α-actin, SM myosin, calponin, and desmin (to name a few), with a concomitant increase in their nonmuscle isoforms (6, 38). However, changes in contractile protein expression in cultured cells do not always coincide with changes in myofilaments (7), and cytoskeletal reorganization (as a result of altered environment) (56) has been reported to be related to changes in cell function (25). A change in the intracellular distribution/organization of proteins and loss of contractility occurs in cultured SMCs from all tissue sources (8–10, 18, 22, 42, 50). Worth et al. (54) published that cultured rabbit aortic SMCs show not only a phenotypic modulation of contractile and cytoskeletal proteins, but a reorganization of these proteins as well. They proposed that “Since the cytoskeleton acts as a spatial regulator of intracellular signaling, reorganization of the cytoskeleton may lead to realignment of signaling molecules, which in turn, may mediate the changes in function associated with SMC phenotypic modulation.” These phenotypic changes begin during cell dissociation or early in cell culture (2, 50).
Because of the significant role isolated SMCs play in SM research, understanding how cell isolation/culturing affects the cytoskeleton is of critical importance. To this end, we examined the stability of the AJ protein complex because this is important for understanding the temporal function and organization of both the contractile and cytoskeletal domains in SMCs. We reported previously (16) that the AJ-associated proteins vinculin, talin, and fibronectin remain associated with the AJ at the SMC periphery in tissues under relaxed and activated conditions. However, these experiments were performed on untethered tissues, and thus their contractile status was not directly defined. In this study, we repeated the experiments on tissues while force was being recorded (Figs. 1 and 2). Both tracheal and digestive SM tissues were used to confirm the findings in these two commonly used organ models. Immunofluorescence done on organs frozen while activated (producing force), relaxed, or Ca2+ depleted (no active force) showed that there was no significant movement from/to the cell periphery of vinculin, talin, fibronectin, or caveolin in any of these conditions (Figs. 1 and 3). These results confirm our previous study reporting no gross translocation of these proteins throughout the SMC under normal physiological conditions (activation and relaxation). However, when SMCs are freshly isolated from tissues, the distribution of these proteins is altered in many of the cells, with vinculin and talin being as likely to be distributed throughout the cell as they are to be localized to the cell periphery (Figs. 4–6). Approximately half of the SMCs from any single cell isolation show the altered pattern of vinculin being distributed throughout the cytoplasm.
To determine when and how the distribution of the AJ-associated proteins becomes altered in isolated SMCs, we examined the procedures we use to isolate single SMCs (15) and the effect that these procedures have on tissue function. In tissues that were enzymatically digested and then frozen, caveolin immunoreactivity remained at the cell periphery, suggesting a stable association at the membrane. However, its distribution became less punctate and the cells appeared larger in cross section (data not shown). This change may in part be due to enzymatic disruption of the AJ, which could then allow the caveoli greater freedom to move among the phospholipids in the plasma membrane. Increased cellular CSA (Fig. 8A) could result from shortening of the cells during the digestion as a result of disruption of extracellular proteins. Because the peripheral localization of caveolin is not altered by the enzymatic digestion, it was used as a control for tissue artifacts and for calculating tissue CSA.
Fibronectin immunoreactivity is lost with enzymatic digestion (Fig. 8). This is most likely a direct result of proteolysis of this extracellular protein by papain. Consistent with this is the observation that the loss of fibronectin immunoreactivity is directly related to the length of digestion time. When examining tissue sections, fibronectin immunoreactivity is progressively lost from the periphery to the center of the tissue as would be predicted from proteolysis by an enzyme diffusing into the tissue and proteolyzing this protein as it progressed. The loss of fibronectin was therefore used to identify regions of tissues that were digested. In contrast to these changes in fibronectin immunoreactivity, the vinculin and talin immunofluorescence did not change from their peripheral distribution as a result of enzymatic digestion (Fig. 8). Activation or relaxation of SM tissues also failed to alter their peripheral distribution after enzymatic digestion. Thus the peripheral localization of vinculin, talin, and caveolin in cells still in tissues remains, even after enzymatic digestion of the tissue (Fig. 8).
The vinculin distribution measured in SMCs in tissues had little variation (average SD = 0.68), showing the uniformity of the peripheral cellular localization of this protein in SMCs. In contrast, in isolated SMCs the vinculin distribution was much more variable (average SD = 1.42), indicating the change in population distribution, as a large percentage of the isolated SMCs no longer showed the in vivo peripheral vinculin distribution (Figs. 4–6, 10). Exactly why many of the isolated SMCs lose their peripheral vinculin distribution and others do not is not apparent at this time. In addition, there was no significant difference in mean vinculin distribution in isolated SMCs relaxed in PSS relative to those after activation in CCh or PDBu (Fig. 10), suggesting that the nonperipheral distribution is not reversible by activation of these cells.
Quantitation of the distribution of AJ-associated proteins in isolated SMCs was done after empirical measurements were made to provide an objective measure of the edge of the isolated cells. As a control for the immunofluorescence technique used in this study, we tested how well PKC-α translocation could be measured. PKC-α has been reported in numerous articles to translocate to the membrane in isolated cells with PDBu stimulation (25, 26, 28, 29, 47, 48, 53), and so we used this system as a positive control for our methods. Freshly isolated SMCs were relaxed in PSS or activated with 10 μM PDBu for 15 min before fixation and immunoreacting. Quantitation of the immunofluorescence signal from confocal images shows that there was a significant shift of the PKC-α from the cytosol in the relaxed SMCs to the plasma membrane with PDBu activation (Figs. 6, 7, and 10). Unlike the dichotomy of vinculin distribution observed in freshly isolated SMCs (approximately half of the SMCs have peripheral and half have uniform vinculin distribution), the vast majority of isolated SMCs showed a uniform distribution of PKC-α when relaxed and a peripheral distribution after PDBu stimulation. PKC-α has a similar uniform distribution in isolated SMCs (same cell population, 0.96 ± 0.35) when relaxed or when stimulated with CCh (1.00 ± 0.86). When isolated SMCs are PDBu stimulated, however, the PKC-α distribution shifts to being very peripheral (3.90 ± 2.41; Fig. 10). In the same isolated SMCs, filamentous actin (as measured by phalloidin immunofluorescence; see methods) appears uniformly distributed throughout the cell with very little variance (peripheral/cytoplasmic intensity 0.80 ± 0.12; n = 30). These results are consistent with previous reports of translocation of PKC-α to the membrane with PDBu activation in isolated SMCs and validate the use of immunofluorescence for quantifying protein translocation. The translocation of PKC-α occurred with PDBu activation independent of cell shortening. This was observed by using cells activated in suspension (and thus able to shorten) and then stuck to a cover glass (stimulated-attached), or cells stuck to a cover glass (and thus restricted in their ability to shorten) and then stimulated (attached-stimulated). This suggests that this second messenger system may still function in cells where the cytoskeletal organization has been altered.
Force production in enzymatically treated tissue strips was significantly reduced compared with predigested force (Fig. 10). Fibronectin immunofluorescence showed a significant decrease in digested tissue that correlated with the decrease in force production (r2 = 0.96). There was no shift in caveolin, vinculin, or talin peripheral/cytosolic cellular distribution in SMCs in these digested tissues. SMCs in SM tissues that had been enzymatically digested and frozen showed no movement of intracellular AJ-associated proteins away from the cell periphery. However, further mechanical dissociation of enzymatically digested tissue to liberate single SMCs resulted in a dramatic redistribution of vinculin and talin from the AJ at the membrane into the cytosol (Figs. 4–6). We propose that proteolysis of the extracellular AJ-associated protein fibronectin (and likely others) with enzymatic digestion, by itself, does not alter the clustering of integrins and their binding to the intracellular AJ-associated proteins vinculin and talin in tissue. However, when the SMCs are physically isolated from their tissue beds, the proteolyzed fibronectin is allowed to diffuse away from the integrin, resulting in unclustering of integrins and dissociation of the intracellular AJ-associated proteins vinculin and talin. This proposed loss of outside-in signaling and resultant release of vinculin and talin from integrin may significantly alter the mechanical and biochemical function of these freshly isolated cells.
This idea is consistent with that of Miyamoto et al. (33), who reported that integrin receptor occupancy and receptor clustering are involved in determining the response of integrin to extracellular ligands. Integrin receptor occupancy by itself can result in receptor redistribution, while receptor clustering by itself can result in accumulation of tensin and FAK and tyrosine phosphorylation. When both integrin receptor binding and integrin clustering occur simultaneously, there is accumulation of talin, α-actinin, paxillin, vinculin, F-actin, and filamin at the cytosolic membrane. In further work, Miyamoto et al. (34) reported a hierarchy of transmembrane actions including aggregation of integrin receptors (without ligand occupancy) causing translocation of >20 signal transduction molecules to the membrane; tyrosine kinase activity causing ERK and SAPK/JNK pathway activation; integrin receptor occupation and aggregation causing talin, α-actinin, and vinculin accumulation at the membrane; and integrin occupancy, aggregation, and tyrosine kinase activity causing F-actin, paxillin, and filamin aggregation at the membrane. Brown et al. (4) showed that talin is an essential core member of integrin-mediated adhesion whose function is to connect extracellular matrix-bound integrins to the cytoskeleton but is not required for signaling for gene expression. Thus the movement of talin from the AJ in the isolated SMC periphery to the cytoplasm may have a significant effect on the mechanics and cell signaling of freshly isolated and cultured SMCs.
The results of this study are significant in suggesting/eliminating several working hypotheses for SM function. To begin, these results suggest that mechanical plasticity of SM tissue is not a result of translocation of talin and vinculin to/from the AJ. These results do not, however, rule out possible association/dissociation of filamentous actin with a given AJ. This could still occur via filamentous actin severing proteins and/or polymerization/depolymerization of G-/F-actin. The results also suggest a possible explanation for the observation that after enzymatic digestion and mechanical disruption of a SM tissue for isolation of single SMCs a significant number of these isolated cells are mechanically unresponsive to stimulation (unpublished data). This could be because some of the cells “die” as a result of the isolation and thus electrical or agonist stimulation is not possible (the latter could also result from proteolysis of the receptors by the enzyme). However, this should not preclude their direct activation with Ca2+ solutions after permeabilization. In fact, there is still a significant percentage of cells that are mechanically nonresponsive to direct activation (pCa 5.0 solutions) after permeabilization (unpublished results). Another possible explanation could be that the association between vinculin and talin and the AJ has been disrupted and any force generated by the contractile proteins cannot be transmitted to the membrane to cause cell shortening. This is supported (but not proven) by a subset of the isolated SMCs showing translocation of PKC-α to the cell periphery with PDBu stimulation, without cell shortening. Proteolysis of fibronectin, as indicated by the decrease of fibronectin immunoreactivity, could cause this by altering integrin signaling, clustering, and/or intracellular AJ-protein associations. It is also possible that the integrin itself is being proteolyzed and is no longer mechanically functional.
This study shows changes in the distribution of the AJ-associated proteins vinculin, talin, and fibronectin due to SMC isolation. The implications of these changes for isolated SMC mechanics and signaling pathways have not been fully evaluated. However, as discussed above, the formation, stability, mechanics, and signaling of cells are critically dependent on the proper association and positioning of these proteins at and near the cell membrane. Thus this work suggests that the mechanics and signaling of cells isolated for single-cell studies or cells in culture (which normally form tissues with more stable associations) may be altered. It is unclear whether these altered cells can recover from these changes and, if so, what the time frame would be. One immediate result of alterations to the AJ could be a reduction or loss of force generation/cell shortening by these cells. This could be a signal for the dedifferentiation process that occurs when SMCs are put in culture and cell migration in pathological diseases. Further work looking for possible changes in mechanics and cell signaling pathways in freshly isolated/cultured SMCs relative to intact tissues/in vivo studies may prove efficacious in advancing our knowledge of SM function in normal and pathological conditions.
This study was supported by National Heart, Lung, and Blood Institute Grants RO1-HL-62237 (to T. J. Eddinger) and RO1-HL-073828 (to D. R. Swartz).
We thank P. S. Clifford and J. J. Hamann at the Medical College of Wisconsin for allowing us to harvest dog tissues at the end of their experiments and Joel Meehl for some of the data collection and analysis.
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 © 2007 the American Physiological Society