Abstract
ABSTRACT. Disruption of normal actin organization in renal tubular epithelial cells is an important element of renal injury induced by ischemia. Studies of fixed cells indicate that the cytoskeleton is disrupted by both ischemia and ATP depletion in a site-specific manner. However, few studies have examined these effects in living cells, and the relationship between the time course of ATP reduction and alteration of the cytoskeleton remains unclear. Here, time-lapse video images of cultured renal epithelial cells expressing an enhanced green fluorescent protein (EGFP)-actin fusion protein were obtained, and the kinetics of fluorescence actin distribution before and during ATP depletion is quantified and compared with measured ATP levels. This study found that assembly of lamellar actin is inhibited rapidly as cellular ATP levels are reduced, whereas disruption of actin in stress fibers is more gradual and persistent. Actin associated with focal adhesions is largely resistant to ATP depletion in these experiments, and, consistent with previous studies, particulate aggregates of actin were formed within the cytoplasm of ATP-depleted cells. Most surprisingly, time-lapse imaging of EGFP-actin distribution, quantitative fluorescence imaging of phalloidin-stained cells, and ultrastructural studies indicate that assembly of actin filaments occurs at sites of epithelial cell-cell attachment in ATP-depleted cells. This assembly is initiated early during ATP depletion and continues after ATP levels are maximally reduced. Assembly of actin at sites of cell-cell attachment may be an element of the pathology of injury induced by ischemia, or alternatively, could reflect the function of a protective mechanism. These studies directly demonstrate site-specific alteration of actin assembly in living epithelial cells during ATP depletion. The results also reveal that actin reorganization continues after ATP levels are maximally decreased and that epithelial cell-cell attachments are sites of actin assembly in ATP-depleted cells. Email: shelden@umich.edu
Normal epithelial function is dependent on the integrity of actin cytoskeletal arrays and complexes mediating cell-cell and cell substrate attachment. In the kidney, disruption of these arrays in renal tubular epithelial cells (RTE) is thought to be an important mediator of ischemic acute renal failure (for review, see references 1–3). For example, disruption of microvillar actin arrays in RTE can be detected within 5 min of renal artery occlusion in vivo (4). Dissolution of basal actin filament bundles, or stress fibers, has also been observed in a variety of ATP-depleted cells (5–9), and this may weaken cell-substrate attachment. In the kidney, it is thought that the loss of cell-substrate attachment during ischemia results in the shedding of cells into tubule lumens and contributes to impaired renal function (10–12). Similarly, disruption of actin filament at sites of cell-cell attachment during ATP depletion has been described (9,13,14) and may play a role in compromising epithelial barrier function during ischemia. Evidence from recent in vitro studies indicates that disruption of actin filament arrays during ischemia may be mediated by upregulation in ADF/cofilin actin severing activity (15) as well as loss of Rho kinase-mediated assembly of actin at cell-substrate attachments (13) and other elements of cell-cell junctions (16).
Because many actin filament arrays found in cells under control conditions are disrupted by ATP depletion, it is surprising that the total cellular content of filamentous actin (F-actin) in epithelial cells increases during ATP depletion (8,14,17,18). Several groups have shown that actin aggregates appear within the cytoplasm of ATP-depleted cells, but the mechanism of their formation and their significance remain under investigation. Recently, we have also documented that hsp27, a putative stress-activated actin-associated protein, is recruited to sites of cell-cell adhesions during ATP depletion of renal epithelial cells (19). Overexpression of hsp27 has previously been shown to increase stability of actin arrays in cells subject to a variety of injuries (for review, see references 20–22), and previous investigators have noted that actin arrays associated with cell junctions are more resistant to disruption during ATP depletion than those at other sites (9,23). These studies generally indicate that site-specific alteration of actin stability or assembly occurs in ATP-depleted cells and support the hypothesis that preferential stabilization of actin filaments at epithelial cell junctions may be an important aspect of the cellular response to ATP depletion. However, previous studies examining fixed, fluorescently stained actin arrays have been limited in their ability to address this issue because they could not correlate the distribution of actin both before and during ATP depletion in individual cells.
In the present study, we conducted time-lapse fluorescence imaging of a renal epithelial cell line (LLC-PK1) stably expressing an enhanced green fluorescent protein (EGFP)-actin fusion protein before and during ATP depletion and during ATP repletion. We find that actin in lamellar protrusions is rapidly disrupted by ATP depletion, whereas disruption of stress fibers occurs more gradually. Actin associated with terminal focal adhesions was resistant to disruption by ATP depletion. As expected, aggregates of actin were formed within the cytoplasm of ATP-depleted cells. Most surprisingly, our videos of EGFP-actin in living cells reveal that fluorescence actin accumulates at sites of epithelial cell-cell attachment in ATP-depleted cells. The time course of this accumulation correlates well with a quantitative reduction in background cytosolic fluorescence, and presumably actin monomer, quantified here and noted in previous studies (8), suggesting that the two events may be causally linked. Accumulation of fluorescence actin at sites of cell-cell attachment continued after ATP levels were maximally reduced. Ultrastructural analysis of cell junctions and association of phalloidin with sites of cell-cell attachment in fixed cells before and during ATP depletion further indicate that the observed accumulation of fluorescence actin in ATP-depleted cells represents actin filament assembly.
Together, these results reveal that individual types of actin filament arrays are distinctly altered by ATP depletion. Our results demonstrate that the inhibition of actin assembly in lamellae is an early consequence of ATP depletion and that assembly of actin filament at sites of cell-cell attachment can play a role in the formation of actin polymer in ATP-depleted cells. We propose that assembly of actin at epithelial cell junctions could be one element of a protective response of epithelial cells to ischemic injury or, alternatively, may be an aspect of the pathology of renal injury induced by ischemia.
Materials and Methods
Cell Culture
LLC-PK1 epithelial cells were purchased from American Type Culture Collection (Manassas, VA) and cultured in Dulbecco’s Modified Eagle Medium (Invitrogen Life Technologies, Carlsbad, CA) containing 25 mM glucose, 10% fetal bovine serum, and antibiotics at 37°C in an environment containing 5% CO2. Cells were transfected with 2 μg of an expression vector coding for a fusion protein of actin and EGFP (Clontech, Palo Alto, CA) using lipofectamine (Invitrogen) according to the manufacturer’s instructions. Transfected cells were selected and subcloned as described in our previous study (19), generating a stable cell line expressing the EGFP-actin fusion protein.
Measurement of ATP Levels
ATP levels were measured by plating LLC-PK1 into 24 well plates and allowing them to grow to 90% confluence over a 2-d period. Triplicate wells were rinsed with HEPES-buffered saline (HBS; 20 mM HEPES, 135 mM NaCl, 4 mM KCl, 1 mM Na2HPO4, 2 mM CaCl2, 1 mM MgCl2, pH 7.2) and treated with ATP depletion medium (HBS containing 1 μM antimycin A and 10 mM 2-deoxyglucose, pH 7.2) for indicated times. All reagents were purchased from Sigma Chemical Co., St. Louis, MO, except as indicated. Triplicate control wells were left untreated, and a second triplicate set of wells was rinsed twice with HBS and incubated for 1 h with HBS containing 25 mM glucose. ATP was extracted in a 6% solution of TCA, acidity was neutralized by vortexing with tri-N-octylamine/Freon, and ATP content of samples was analyzed using HPLC as described previously (24).
Imaging of Living Cells
Observation chambers for live cell imaging were made by drilling a 15-mm-diameter hole in the bottom of 35-mm petri dishes and gluing number 1 microscope coverslips over the hole using Sylgard elastomer (Dow Corning Corp.). Chambers were sterilized by treatment with 70% ethanol before use. Before the conduction of experiments, cells were trypsinized and plated in observation chambers at a density sufficient to reach approximately 50% confluence after 1 to 2 d of culture. Cells were imaged with a Zeiss Axiovert 135 microscope (Carl Zeiss Inc., Thornwood, NY) equipped with a 40 × 1.4 NA oil immersion objective lens. Temperature was maintained at 37°C by using an Airtherm air stream incubator (World Precision Instruments, Sarasota, FL). Images of cells were obtained at 2-min intervals by using a cooled integrating CCD camera (DAGE RT3000, DAGE-MTI Inc. Michigan City, IN), using a 0.5-s integration time. Illumination was provided with an Attoarc 100W mercury arc lamp (Carl Zeiss Inc.) attenuated using neutral density filters and shuttered using a Uniblitz shutter and controller (Vincent Associates, Rochester, NY). Camera integration times, shutters, and image capture were coordinated by macro command sets using NIH-Image running on an Apple Macintosh G4 computer equipped with an image capture board (LG3; Scion Corp, Frederick, MD). For initial imaging of cells, culture medium was replaced with HBS, pH 7.2, containing 25 mM glucose. After imaging cells in this medium, ATP levels were depleted by rinsing chambers twice with HBS, pH 7.2, without glucose followed by addition of HBS containing 1 μM antimycin A and 10 mM 2-deoxyglucose at pH 5.5 or pH 7.2. ATP repletion was conducted by removing solution with inhibitors, rinsing cells twice with HBS, pH 7.2, and adding HBS, pH 7.2, containing 25 mM glucose. Imaging of cells was continued during medium changes. Representative movies in Quicktime format may be found on the Internet at http://www.umich.edu/~shelden/JASN2002b.html.
Confocal Imaging of EGFP-Actin and Total Actin
Cells were cultured for 24 to 48 h on glass coverslips until confluent and subjected to ATP depletion with or without recovery as described above, then fixed in 4% paraformaldehyde at room temperature and stained with rhodamine phalloidin, as described in our previous study (19), and 1 μg/ml Hoechst to reveal nuclear morphology. Coverslips were mounted for observation on microscope slides using Prolong mounting medium (Molecular Probes, Eugene, OR). Imaging of EGFP-actin and rhodamine phalloidin was conducted using a Zeiss LSM510 confocal microscope (Carl Zeiss Inc.) equipped with a 63 × 1.2 NA water immersion objective. Laser output and detector gain and black level settings were optimized using a preparation of ATP-depleted cells and then held constant for all imaging. Images of all three fluorescence probes were obtained simultaneously using a multichannel scanning procedure in which each line of the final image was scanned three times, using excitation and imaging filters specific for each individual fluorophore.
Quantitative Analysis and Statistics
Actin assembly kinetics in lamellar protrusions were quantified using an “image difference analysis” developed by our laboratory for analyzing lamellar ruffling dynamics from phase contrast images (25). Briefly, sequential video images were digitally subtracted from each other and regions varying by more than 5% selected. Camera noise was reduced using a median filter, and the final area of difference was measured for each image pair. We believe that this study represents the first use of this method to quantify actin array turnover in living cells.
Fluorescence intensity of stress fibers and background fluorescence of the cytoplasm was calculated essentially as described previously by others (8,13). For each image, a duplicate image was created, and background fluorescence was removed using a 2D rolling ball background subtraction algorithm. The resultant image of stress fibers (and other sharp detail) was used as a digital mask and multiplied by the original image, creating an image in which only stress fiber fluorescence was retained. The region of each cell containing stress fibers was selected with a cursor, and the average brightness of the region was measured. The brightness of the background cytoplasm in the same region was obtained using the inverse of the stress fiber mask to select regions lacking stress fibers. To measure the brightness of cell-cell attachments in live cells, a duplicate image of each video frame was created and sites of cell-cell attachment were traced with a digital brush set to a unique value. The trace was used as a mask to select grayscale values in the original image, and the area and brightness of the final selected region in each image was measured. This approach was also used to measure the fluorescence intensity of cell junctions in fixed, triple-labeled cells imaged using confocal microscopy. The fluorescence intensity of actin in non-junctional regions was measured from confocal images by selecting each cell with a cursor, excluding junctional areas, and computing the average fluorescence intensity of selected regions. For analysis of confocal images, only the rhodamine-phalloidin staining intensity (red channel in Figure 5) was analyzed, thus, this analysis specifically examines only polymerized actin. Measurements of junctions and of cytoplasmic actin in non-junctional regions were analyzed separately for cells expressing and lacking detectable expression of EGFP-actin. In all of our studies, maximum brightness would correspond to a measured value of 256, while a black background would have a grayscale value of zero. Confocal images of control and ATP-depleted cells were also scored for the presence of actin aggregates within the cytoplasm by a blinded observer, but no attempt was made to analyze changes in the size or number of aggregates in cells in the present study.
Figure 5. Distribution of EGFP-actin and total filamentous actin in confluent monolayers of LLC-PK1 cells before, during, and after ATP depletion. EGFP-actin (left column, green) is incorporated into rhodamine phalloidin-stained actin filament arrays (middle column, red) in control cells (A), after 1 h of ATP depletion at pH 5.5 (B) and after 90 min of recovery (C). Incorporation of EGFP-actin into stress fiber bundles is particularly evident in cells marked with asterisks. However, not all cells express detectable EGFP-actin (arrowheads). Comparison of the brightness of junctional staining in control cells (A) expressing (arrows) and lacking (arrowheads) EGFP-actin with the intensity of junctions in corresponding ATP-depleted cells (B) suggests that junctional regions increase in thickness and intensity of actin fluorescence during ATP depletion. Actin aggregates are found in the cytoplasm of cells under all conditions (arrowheads, right panels) but were most evident in ATP-depleted cells (B). Enhanced actin fluorescence at site of cell-cell attachment persists after 90 min of ATP repletion in some cells (arrowhead, C). Scale bar = 20 μm.
Finally, the fluorescence intensity of focal adhesions was measured by first outlining each focal adhesion using a cursor. For the purposes of this study, focal adhesions were defined as oblong, often somewhat triangular, areas of actin fluorescence at the bottom of cells, which terminated and were slightly larger and brighter than an attached stress fiber. Isolated, stationary fluorescence actin structures similar in size, orientation, and brightness to neighboring focal adhesions with attached stress fibers were also considered to be focal adhesions (Figure 8). The area and average intensity of each selected region was quantified using software commands available within the NIH-Image program. For each focal adhesion, the average brightness was calculated from four images obtained before the onset of ATP depletion, and four images obtained after 1 h of ATP depletion.
Figure 8. Fluorescence intensity of cytoplasm and stress fibers but not focal adhesions is reduced by 1 h of ATP depletion. Actin distribution in a cell before (A and C) and after 1 h (B and D) of ATP depletion, showing loss of stress fibers (arrow, A) but not focal adhesions (C and D). Scale bar = 20 μm in panels A and B. Quantitative analysis shows the time course of fluorescent intensity decreases within stress fibers (E; four cells from two videos) and the background fluorescence of the cytoplasm (F), but a plot of focal adhesion fluorescence before (horizontal axis) versus after 1 h (vertical axis) of ATP depletion (G) reveals no average change in focal adhesion fluorescence intensity.
Statistical analyses of data were conducted using Microsoft Excel. Comparison of population means was conducted using a t test assuming equal variance.
Electron Microscopy
Cells were cultured in 35-mm dishes containing 200-mesh nickel grids with a Formvar/carbon coating (Electron Microscope Sciences, Fort Washington, PA) and either ATP depleted at pH 5.5 or processed without ATP depletion (controls) as described above. Cells were detergent-lysed for 5 min to remove non-cytoskeletal components essentially as described by Svitkina and Borisy (26) in an actin-stabilizing lysis buffer (50 mM imidazole, 50 mM KCl, 0.5 mM MgCl2; 0.1 mM EDTA; 1 mM EGTA, 4% polyethylene glycol [8000 MW], and 200 μg/ml rhodamine phalloidin, pH 6.8), then fixed in 0.1 M phosphate buffer containing 2.5% glutaraldehyde, pH 7.2. Grids were rinsed three times with distilled water and stained for 3 min with 2% aqueous phosphotungstic acid. Excess stain was removed and grids dried slowly for about 10 min in a humid chamber. Cells were imaged using a Phillips CM100 transmission electron microscope operated at 60 kV equipped with a Kodak Megaplus camera, model 1.6.
Results
Time Course of ATP Depletion
To characterize the effects of our procedures on ATP levels in LLC-PK1 cells expressing EGFP-actin and to permit the direct comparison of changes in actin cytoskeletal organization with ATP levels, ATP levels were measured in cells treated for up to 1 h with 1 μM antimycin A and 10 mM 2-deoxyglucose. As expected, ATP levels were rapidly reduced after the application of these inhibitors. Relative to the control time zero, values (± SD) were 71.4 ± 6.2% at 2.5 min, 34.9 ± 1.6% at 5 min, 12.7 ± 1.2% at 10 min, 3.9 ± 0.5% at 20 min, 1.2 ± 0.2% at 40 min, and 1.0 ± 0.1% at 60 min. (Figure 1). In contrast, replacement of culture medium with HBS containing 25 mM glucose, pH 7.2, produced no change in ATP levels after 1 h of treatment.
Figure 1. ATP levels in LLC-PK1 cells expressing enhanced green fluorescent protein (EGFP)-actin during ATP depletion and control experiments. ATP levels fall rapidly in cells treated with HEPES-buffered saline (HBS) containing 1 μM antimycin A and 10 mM 2-deoxyglucose, but not when incubated with HBS containing 25 mM glucose. Values shown are means and SD calculated from three replicate experiments.
Lamellar Protrusion Is Inhibited Rapidly during ATP Depletion
EGFP-actin was imaged at 2-min intervals in LLC-PK1 cells under control conditions, during ATP depletion at an extracellular pH of either 7.2 or 5.5, and during ATP repletion at pH 7.2 in the presence of 25 mM glucose. In total, actin fluorescence and the kinetics of actin array turnover were examined in 19 videos showing 143 ATP-depleted cells and 3 videos showing 18 control cells. To quantify effects of ATP depletion on actin assembly dynamics, an analysis of image difference was applied to videos (see Materials and Methods). Figure 2 shows two sequential images of fluorescent actin in a cell before ATP depletion (Figure 2A) with the resultant difference image (Figure 2C). Regions undergoing more than 5% change in grayscale value are observed at sites of lamellar protrusion at the leading edge of the cells (arrows, Figure 2B) and are black in the final difference image (Figure 2C). After 15 min of ATP depletion, no changes in actin distribution are detected in this same cell over a 2-min interval (Figure 2, D and E), and no area of change is detected in the difference image (Figure 2F). In Figure 2G, difference values obtained over the time course of ATP depletion for four cells and for one control cell are shown. Measured dynamic turnover of the actin cytoskeleton is greatly reduced within 10 min of the addition of metabolic inhibitors (transient increases at the start of ATP depletion [arrow, Figure 3G] and other points are due to focus and stage position changes [data not shown]), whereas replacing the initial medium with additional glucose containing medium (cntrl, Figure 2G) has no effect on the dynamics of actin reorganization.
Figure 2. Analysis of actin array turnover in lamellar protrusion using image difference calculations. Images of a cell under control conditions taken 2 min apart (A and B) are subtracted and areas changing by more than 5% in grayscale value selected to produce a resultant image (C) in which sites of array turnover (arrows, B) are black. The same method applied to images of this cell after 15 min of ATP depletion (D and E) produce a difference image (F) showing no regions of change. Scale bar = 20 μm. (G) The time courses of normalized difference values (area of black regions in panels C and F) calculated before and during ATP depletion for four cells and a control cell (open circles) in which medium was exchanged without ATP depletion are shown. The peak at the time of medium change reflects shifts in culture chamber position (arrow). Axes are min of ATP depletion (horizontal) and proportional change in image difference value (log scale, vertical).
Figure 3. Rapid inhibition of actin assembly in lamellar protrusions during early ATP depletion. (A) Video images taken 2 min apart of lamellae formed between two attached cells were combined to show areas of actin array extension or assembly in green and areas of retraction or disassembly in red. A diagram of the first image (panel I) shows junctional actin arrays that are stable over the 2-min interval in yellow. Dynamic reorganization of actin arrays in lamellar protrusions is seen before the onset of ATP depletion but is inhibited within 4 min of the start of ATP depletion. (B) ATP depletion also induces loss of actin fluorescence within lamellar protrusions (arrow) without retraction. Numbers are min before (negative) or after (positive) application of metabolic inhibitors. Scale bar = 10 μm.
Figure 3A shows images of lamellae formed at a site of cell-cell attachment between two LLC-PK1 cells. Images obtained 2 min apart were combined such that the later image is green and the earlier image is red (Figure 3A). Sites of actin assembly are green in the resultant image, whereas structures that disappear during this interval are red. The site of cell-cell attachment and other stable actin-containing structures are unchanged in both images and are therefore yellow (quiescent cell boundaries are shown in gray in the accompanying diagram of the first image). Both green (arrows, Figure 3A) and red areas are seen in these images before the addition of inhibitors, indicating dynamic turnover. In contrast, images obtained within 4 min of inhibitor addition show complete overlap of red and green, indicating that time-dependent change in the distribution of fluorescent actin is no longer occurring. Figure 3B also shows grayscale images of fluorescent actin in a lamellar protrusion (arrows) in which loss of actin fluorescence is observed without lamellar retraction. Together, data shown in Figures 1, 2, and 3 reveal that actin assembly ceases in lamellar protrusions within minutes of the start of ATP depletion and at time points before maximal reduction of ATP levels.
Accumulation of EGFP-Actin at Sites of Epithelial Cell-Cell Attachment in ATP-Depleted Cells
Figure 4A shows images of fluorescent actin at a site of epithelial cell-cell attachment in a cell treated with inhibitors of ATP production. Unlike lamellar protrusions, EGFP-actin fluorescence associated with this structure increased gradually and persistently in fluorescence intensity and apparent thickness over about 2 h of ATP depletion. Figure 4B shows higher magnification images of an attachment site between two cells (shown at low power in the inset) after the application of inhibitors of ATP production at time zero. The image series illustrates that the increase in apparent thickness of fluorescent cell-cell junctions seen in Figure 4A is due, at least in part, to the formation or elongation of fluorescent structures resembling microspikes or filopodia. In Figure 4C, normalized fluorescent actin intensities quantified for five sites of cell-cell attachment in ATP-depleted cells and a control cell are shown. Increased actin fluorescence is detected at cell-cell attachments during ATP depletion, and comparison of these graphs with data shown in Figure 1 reveals that much of this increase occurs after ATP levels have fallen to less than 2% of control. No change in actin fluorescence is detected at site cell-cell attachment in cells before the start of ATP depletion (Figure 4B) and in a 1 h control experiment (cntrl 1, Figure 4B).
Figure 4. EGFP-actin fluorescence intensity at epithelial cell junctions during ATP depletion. (A) Images of a site of cell-cell attachment (from a series at 2-min intervals) after the start of ATP depletion. Increased brightness of the junctional region is observed. (B) Higher magnification of the region between two cells expressing EGFP-actin (inset, panel B) showing assembly of structures resembling filopodia or microspikes. Numbers are min after the start of ATP depletion. Scale bars: 10 μm in panel A; 4 μm in panel B. (C) Graph of normalized EGFP-actin fluorescence intensity at sites of cell-cell attachment (n = 5 cells from two videos) during ATP depletion and from a control experiment in which medium was changed at time 0 without ATP depletion (open circles). Axes are time after medium exchange (horizontal) and relative change in average brightness (vertical).
ATP Depletion Induces Accumulation of Actin Filaments at Sites of Epithelial Cell-Cell Attachment
Videos of cells expressing EGFP-actin described above were made of well-spread cells cultured at moderate (50%) confluence to clearly visualize cell junctions, lamellar protrusions, and stress fibers in the same image. Additionally, structures containing fluorescence EGFP-actin might contain either assembled actin filaments or monomeric actin. Therefore, to determine if accumulated EGFP-actin represented the presence of actin filaments, and to address whether junctional actin accumulated during ATP depletion in cells cultured at higher densities, cells were plated at confluent cell densities, experimentally treated, and then fixed and stained with rhodamine phalloidin, a specific marker for assembled actin filaments. Multichannel confocal fluorescence imaging was used to examine and compare the distribution and intensity of the EGFP-actin and rhodamine phalloidin probes at sites of cell-cell attachment (Figure 5). Inspection of these images shows that EGFP-actin (left column and green, Figure 5) is incorporated into all structures stained with rhodamine phalloidin (middle column and red, Figure 5). Incorporation of EGFP-actin into basal stress fibers is particularly evident in cells marked with asterisks. Junctional regions in control cells expressing EGFP-actin (arrows, center, Figure 5A) and neighboring cells, which lack detectable EGFP-actin (arrowhead, center, Figure 5A), are thin and stain relatively dimly with rhodamine phalloidin when compared with cells fixed after 1 h of ATP depletion (Figure 5B). Junctional regions in ATP-depleted cells expressing EGFP-actin (arrow, center, Figure 5B) and lacking detectable EGFP-actin (arrowhead, center, Figure 5B) are comparatively bright, thick, and fibrous. Images of cells obtained after 2 h of recovery from ATP depletion (Figure 5C) show some junctional areas with normal fluorescence intensity (arrow, center, Figure 5C), while other junctional areas remain fibrous and brightly fluorescent (arrowhead, center, Figure 5C). Aggregates of phalloidin stained actin were observed within the cytoplasm of cells under all conditions (representative examples are indicated by arrowheads in the color panels, right, Figure 5), but they were more common and numerous in cells after 1 h of ATP depletion than in control cells. For example, 25.8% of control cells (73 of 283) expressing EGFP-actin and 22.6% of control cells (51 of 226) lacking detectable EGFP-actin displayed some actin aggregates, whereas 57.6% cells (177 of 307) expressing EGFP-actin and 51.4% of cells (126 of 245) lacking detectable EGFP-actin displayed aggregates after 1 h of ATP depletion. Because our imaging parameters were optimized for detection of our probes in ATP-depleted cells (Figure 5B), phalloidin staining in controls cells (Figure 5A) is comparatively low (Figure 5A, center column; red, right column, Figure 5A). Additionally, background fluorescence of monomeric EGFP-actin would not be expected to stain with rhodamine phalloidin, and the presence of monomeric EGFP-actin probably accounts for the overall green color in the color images of fixed control cells and cells fixed during recovery from ATP depletion. All images shown in Figure 5 were obtained using identical imaging parameters, and brightness and contrast levels were adjusted for all EGFP-actin images and rhodamine phalloidin images together.
Quantitative analysis of the brightness of phalloidin staining at cell junctions was conducted from confocal images of cells fixed without ATP depletion and cells fixed after 1 h of ATP depletion (Figure 6), and results obtained from cells expressing EGFP-actin (green in Figure 5) compared with those obtained from the analysis of cells lacking detectable EGFP-actin (red in Figure 5). Triplicate experiments were conducted, and at least ten random fields of cells were imaged for each trial. Fluorescence phalloidin intensities of all visible junctional regions associated with well-spread, non-mitotic cells were measured. The average brightness (± SD) of junctional regions for control cells expressing EGFP-actin was 73.3 ± 20.7 (n = 820) and 99.5 ± 28.4 (n = 641) for ATP-depleted cells expressing EGFP-actin (Figure 6A). For cells lacking detectable EGFP-actin, the average fluorescence intensity of junctional regions for control and ATP-depleted cells was 75.3 ± 20.8 (n = 275) and 111.3 ± 31.3 (n = 331), respectively (Figure 6B). The increases in fluorescence intensity observed in ATP-depleted cells expressing EGFP-actin and cells lacking EGFP-actin were both significantly higher than values measured for corresponding control cells (P ≤ 0.01). We conclude that the increased phalloidin intensity, and thus actin filament content, of junctional regions in ATP-depleted cells is independent of EGFP-actin expression. Indeed, cells lacking detectable EGFP-actin showed a small but significantly greater increase in fluorescence intensity of junctional regions after ATP depletion (P ≤ 0.01). For comparison, the average fluorescence intensity of rhodamine phalloidin staining was analyzed for the total cell area, excluding cell junctions (Figure 6, C and D). In contrast to actin staining at sites of cell-cell attachment, a significant decrease (P ≤ .01) in the average rhodamine phalloidin brightness of non-junctional areas was measured for cells after ATP depletion. The average brightness of the non-junctional regions (± SD) of control cells was 32.1 ± 8.2 (n = 194) for EGFP-expressing cells and 37.5 ± 8.4 (n = 180) for cells lacking EGFP-actin. The average brightness of the non-junctional regions (± SD) of ATP-depleted cells was 27.0 ± 6.4 (n = 191) for EGFP-expressing cells and 31.7 ± 8.4 (n = 165) for cells lacking EGFP-actin. Because actin aggregates are formed in non-junctional regions of ATP-depleted cells, the decrease in average phalloidin staining measured in ATP-depleted as compared with control cells was unexpected, but it may reflect a comparatively large decrease in the polymer content of stress fibers in ATP-depleted cells (Figure 8E).
Figure 6. Quantitative analysis of phalloidin staining intensity in control and ATP-depleted cells expressing and lacking detectable EGFP-actin. Normalized histogram distributions of phalloidin staining intensity are shown. (A) Intensity of rhodamine phalloidin staining was measured at sites of cell-cell attachment in EGFP-actin expressing cells fixed under control conditions (black bars, n = 820) or after 1 h of ATP depletion (gray bars, n = 641). A population shift toward higher (brighter) values is observed after ATP depletion. (B) Analysis of junctional phalloidin staining in cells lacking detectable EGFP-actin also shows an increase in brightness of junctions in ATP-depleted cells (gray, n = 331) as compared with control cells (black, n = 275). In contrast, normalized histogram distributions of the average intensity of actin staining in non-junctional regions in control cells (black bars) and ATP-depleted cells (gray bars) show no overall increase in brightness of ATP-depleted cells versus control cells in cells expressing (C) or lacking (D) EGFP-actin.
Actin Assembly at Sites of Cell-Cell Attachment during ATP Depletion Is Independent of the Degree of Cellular Injury
Previous studies have shown that cell survival during recovery from ATP depletion is strongly inhibited when ATP depletion is conducted at neutral or greater extracellular pH, and enhanced when ATP depletion is conducted at acidic extracellular pH (27,28). Therefore, to determine whether actin assembly at cell junctions could be correlated with the amount of cellular injury induced during ATP depletion, we analyzed and compared the increase in actin brightness at cell-cell attachments from videos of LLC-PK1 cells undergoing ATP depletion at extracellular pH values of 7.2 and 5.5. Most cells undergoing ATP depletion at pH 5.5 were subsequently observed to recover normal lamellar protrusion behavior after addition of HBS containing glucose, whereas those undergoing ATP depletion at pH 7.2 failed to recover lamellar ruffling behavior. These cells instead underwent dramatic loss of actin cytoskeletal integrity after the addition of HBS and glucose and may have undergone necrotic cell death during the experiment (see videos available on our web site, as described in Materials and Methods). The analysis of the brightness of actin at cell-cell attachments in these experiments was also complicated by the substantial reduction of fluorescence that GFP exhibits at acidic pH (29). We therefore determined the average fluorescence intensity of cell-cell attachments in the first four images obtained during ATP depletion with the average intensity of these same junctions in four images taken after 1 h of ATP depletion (Table 1). The increase in brightness of cell-cell attachments in cells observed to recover lamellar protrusion (pH 5.5), and those that did not recover during a similar observation period (pH 7.2) did not differ significantly (P > 0.1).
Table 1. Average brightness increases during ATP depletion of actin arrays at sites of cell- cell attachment but notfocal adhesionsa
Electron Microscopy of Actin Filaments at Sites of Cell-Cell Attachment in Control and ATP-Depleted Epithelial Cells
The increases of EGFP-actin fluorescence intensity in living cells during ATP depletion and phalloidin staining in fixed cells after ATP depletion (above) support the hypothesis that actin polymer mass increases at sites of epithelial cell-cell attachment during ATP depletion. To further assess the morphology and distribution of actin polymer at these sites, we conducted ultrastructural studies of cells after detergent extraction using an actin stabilizing lysis buffer (Figure 7). Initial studies conducted by thin sectioning epon-embedded cells followed by uranyl acetate and lead citrate staining were less informative than we hoped, perhaps because of the difficulty of visualizing three dimensional actin filament arrays in 70-nm-thick sections (not shown). However, examination of whole cells cultured on EM grids after detergent lysis, fixation, and negative staining using phosphotungstic acid reveals the extensive presence of negatively stained filaments at sites of cell-cell attachment in control cells (Figure 7, A through C). Similar filaments are observed along the margin of sites of cell-cell attachment (white arrows, Figure 7F) and in fibrous protrusions associated with sites of cell-cell attachment (black arrows, Figure 7F) in cells fixed after 1 h of ATP depletion at pH 5.5. In both cases, filaments observed here are morphologically similar to those observed by previous investigators at the leading edge of fibroblasts, stress fibers, and microspikes (30) and at these sites in the present study (not shown). Images shown are representative of 11 junctional regions of ATP-depleted cells and 9 junctional regions of control cells, and no attempt was made to distinguish between cells that expressed or lacked detectable EGFP-actin in these studies. The width of filaments shown in Figures 7C and 7F measured between 7 and 8 nm, the predicted thickness of actin filaments in negatively stained preparations (data not shown), and the orientation of fibers is consistent with that expected for actin filaments associated with epithelial adherens junctions (31). Comparison of the electron density of staining in cell junctions of control cells (Figure 7C) and ATP-depleted cells (asterisk, Figure 7F) also suggests that an increase in electron-dense material characterizes ATP-depleted cell junctions.
Figure 7. Electron microscopy of negatively stained actin filaments at regions of cell-cell attachment in control and ATP-depleted cells. Low magnification view (×2350, panel A) and medium magnification (×13,500, panel B) of the junctional region between two control cells, circled in panel A. (C) High magnification view (×130,000) of the region indicated with an arrow in panel B. The junctional region contains negatively stained filaments oriented along the length of the junctional region. Particularly clear examples are indicated with arrows. The width and orientation of these filaments is consistent with individual actin filaments associated with epithelial adherens junctions. Scale bar = 100 nm. Low magnification view (×2600, panel D) and medium magnification view (×25,000, panel E) of the junctional region between two ATP-depleted cells (circled in panel D). The junctional region is more electron-dense than that of control cells and displays numerous lateral protrusions that appear to be attached to the junctional region. (F) High magnification view of the region indicated by the arrow in panel E. Aligned, negatively stained filaments are detected in protrusions associated with junctional regions (black arrows, F) and at the margin of electron-dense junctional cell borders (white arrows, F). The inset shows the central region of this process after contrast enhancement. The electron density of the junctional region (asterisk) can be directly compared with that of control cells shown in panel C. Scale bar = 100 nm.
Focal Adhesions Are More Resistant to Early ATP Depletion than Stress Fibers
To determine if actin arrays at sites other than cell-cell attachments display altered actin assembly during ATP depletion, we analyzed the intensity of EGFP-actin fluorescence in stress fibers and focal adhesions. Images of a cell before ATP depletion (Figure 8A) and after 1 h of ATP depletion (Figure 8B) show that stress fibers are partially disrupted by ATP depletion in our experiments. Reduced fluorescence intensity of stress fibers can be observed as well as the disappearance of some thin stress fibers (arrow, Figure 8A and surrounding region). These results are in agreement with a previous analysis of stress fiber disruption in living renal epithelial cells during ATP depletion (8). In contrast, the fluorescence intensity of some individual focal adhesions increased during the same time interval (compare insets in Figures 8A and 8B shown at higher magnification in Figures 8C and 8D). As expected, the measured, average fluorescence intensity of stress fibers declined in ATP-depleted cells (Figure 8E). However, when the fluorescence intensities of focal adhesions before and after 1 h of ATP depletion are compared, both decreases and increases in fluorescence intensity of focal adhesions are seen (Figure 8G), but no average change in the fluorescence of focal adhesions is detected (Figure 6G and Table 1).
Finally, previous studies have correlated a decrease in fluorescence actin intensity of the cytoplasm during ATP depletion with a reduction in total actin monomer content (8). To confirm that the behavior of cells in our studies replicated that reported previously and to compare the time courses of the increase in fluorescence of cell-cell attachments in our studies with the reduction in fluorescence of nonpolymerized actin, the fluorescence intensity of the cytoplasm was analyzed here. We find that time-dependent decreases in the fluorescence intensity of the cytoplasm occur in ATP-depleted but not control cells (Figure 8F), and that the time course of this reduction is similar to that observed for the increase in intensity at cell-cell attachments (compare with Figure 4B).
Discussion
Results of the present study provide new data on actin distribution in living renal epithelial cells (RTE) during ATP depletion and permit the comparison of changes in actin organization with measured reductions in ATP levels. The most immediate effect of ATP depletion observed in our studies is the inhibition of lamellar turnover and the accompanying loss of actin from these structures (Figures 2 and 3). Presently, relatively little is known about the contribution of processes mediating lamellar protrusion to maintenance of epithelial tissues in vivo. Direct visualization of lamellar protrusion by RTE in vivo has not been accomplished, and the possibility that the generation of these structures by cells in our studies is an artifact of our in vitro cell culture conditions should not be ruled out. However, lamellar protrusion by RTE is likely to play a significant role in wound healing and re-epithelialization in vivo during recovery from ischemic and other injuries. Additionally, recent evidence suggests that mechanisms involved in mediating lamellar protrusion also play roles in the maintenance of normal epithelial function. For example, lamellar ruffling is induced as a consequence of activation of the Rho family small GTPase Rac1 (32–34), and results of recent studies indicate that Rac1 plays a role in the maintenance of epithelial cell junctions (35,36). Results of the present study indicate that such functions may be highly sensitive to disruption as a consequence of ATP depletion and that even very brief periods or modest degrees of ischemia could inhibit epithelial functions that are dependent on lamellar protrusion.
Perhaps the most novel aspect of this study is the demonstration of an increase in brightness of fluorescent actin probes at sites of cell-cell attachment during ATP depletion. The increase in EGFP-actin intensity observed in living ATP-depleted cells (Figure 4A), the increased intensity of phalloidin staining after ATP depletion, both in cells expressing and lacking detectable EGFP-actin (Figure 6), the formation of microspike-like structures at sites of cell-cell attachment in ATP-depleted cells (Figure 4B), and the presence of 7- to 8-nm-diameter filaments at these sites (Figure 7) all lead us to conclude that this increase in fluorescence intensity is due, at least in part, to actin filament assembly during ATP depletion. Our findings agree well with results from previous analysis of fixed cells showing greater resistance of junctional actin to ATP depletion than actin in stress fibers (9,23). Indeed, although to our knowledge our studies provide the first direct demonstration of actin assembly at sites of epithelial cell-cell attachment during ATP depletion, such assembly can be inferred from careful comparison of actin distribution patterns in fixed cells published in previous studies (23,28). Additionally, like results from previous studies of fixed cells (28), our results show that actin cytoskeletal alteration is independent of extracellular pH during ATP depletion. These similarities indicate that the actin assembly observed in our experiments is likely representative of changes in actin assembly occurring in other experimental models of ischemia and in vivo.
The assembly of actin in ATP-depleted cells has been previously demonstrated using biochemical assays of actin polymer and monomer (8,14,17,18) and has been inferred from an increase in cytoplasmic actin aggregates within the cytoplasm and perinuclear area in several previous studies of ATP-depleted renal epithelial cells and tissues (9,14,17,28). Unfortunately, although actin aggregates were observed in fixed, phalloidin-stained cells in our studies (Figure 5), these structures were not observed in living cells in the present study, probably because our observations of living cells focused only on actin arrays close to the cell substrate. Because our analysis of actin polymer in fixed cells was limited to individual confocal images focused at the level of cell junctions (Figures 5 and 6) these results also do not address the relative contribution of assembly in aggregates and at sites of cell-cell attachment to the accumulation of actin polymer in ATP-depleted cells. Additionally, although our ultrastructural studies demonstrate the presence of morphologically normal actin filaments in structures assembled at sites of cell-cell attachment in ATP-depleted cells, it also remains unclear whether actin in ATP-depleted cells at either cell attachments or cytoplasmic aggregates polymerizes through a normal assembly mechanism; for example, through incorporation of ATP-associated actin monomer at filament barbed ends or through some other mechanism. However, because we determined that ATP levels fell to less than 2% of control levels within 20 to 40 min of the application of inhibitors in our study (Figure 1), we conclude that much of this assembly is either ATP-independent or requires extremely low ATP concentrations. Thus, actin assembly observed in ATP-depleted cells may occur as a consequence of mechanisms that do not normally play a role in actin assembly in control cells.
Assembly of actin filaments in junctional regions may be an important aspect of the epithelial cell response to prolonged ATP depletion. However, examination of the graphs shown in Figure 4C also suggests that there is no significant lag period between the start of ATP depletion and the time-point at which actin fluorescence begins to increase at sites of cell-cell attachment. Therefore, actin assembly at cell junctions also appears to be an early consequence of ATP depletion. Interestingly, actin assembly occurs at cell-cell attachments in cells, but stress fibers and focal adhesions in the same cells did not exhibit assembly during ATP depletion in our study (Figure 8). Thus, not all actin filament arrays in cells are capable of promoting actin assembly during ATP depletion. These differences may reflect the diversity of actin-associated proteins found at cell adhesion complexes or other cortical actin arrays. Our discovery that actin filament content increases can occur at epithelial cell-cell attachments during ATP depletion raises the question of the functional significance of this behavior. Because cell adhesion complexes are a highly ordered assembly of cytoskeletal, regulatory, and transmembrane proteins, it is possible that abnormal actin assembly at these sites could play a role in the disruption of epithelial barrier function accompanying ischemic injury. Alternatively, assembly of actin and recruitment of actin-associated proteins to cell junctions may reflect the function of mechanisms involved in preserving epithelial integrity during injury.
Finally, it is of importance to consider whether the EGFP-actin probe is an appropriate marker for actin cytoskeletal reorganization. Previous investigators have shown normal actin-dependent cell behavior in cells expressing the EGFP-actin construct used in our studies (37). Additionally, Herget-Rosenthal et al. (8) have recently published an extensive analysis of the behavior of EYFP-actin in LLC-PK1 renal epithelial cells and have concluded that EYFP-actin expression is a faithful marker for the total cellular pool of actin and that its expression does not alter actin-dependent cellular responses. The actin coding region of the EGFP-actin construct used in our studies is identical to that of the EYFP-actin construct used in this previous report; in total, the two expressed proteins differ by only 5 of 621 amino acids. It seems likely that the assembly characteristics of the EGFP-actin probe used in the current study are very similar to that of the EYFP-actin probe used in the previous study. Additionally, our examination of the phalloidin-staining intensity of cells expressing detectable EGFP-actin and those lacking EGFP-actin reveal similar increases in ATP-depleted cells as compared with control cells (Figure 6). We conclude that EGFP-actin expression is not a causal factor in generating the observed changes in actin assembly in our studies.
In summary, we have quantified the kinetics of actin distribution in cultured renal epithelial cells before and during ATP depletion and correlated these data with measured ATP levels. Loss of actin assembly in lamellar protrusion is an immediate consequence of reducing ATP levels, and actin turnover in lamellae is completely inhibited when ATP levels are reduced to less than 2% of control values. Actin associated with stress fibers was also disrupted during ATP depletion, albeit more slowly. In contrast, actin assembly is detected in cytoplasmic aggregates and observed at sites of epithelial cell-cell attachment. Assembly at sites of cell-cell attachment is initiated early during ATP depletion but persists after ATP levels are maximally reduced. These results illustrate that actin assembly is altered in a site-specific manner during ATP depletion and suggest that actin assembly at sites of epithelial cell-cell attachment is an important aspect of the cellular consequence of ATP depletion.
Acknowledgments
We thank N. Roeser and R. Senter for assistance with ATP measurements, and Drs. S.A. Ernst and B. Margolis at the University of Michigan for critical reading of this manuscript. Grant support from the National Institute of Environmental Health Science (ES11196–01) and the National Institute on Aging (AG19847–01) to Eric A. Shelden and from the National Institute of Diabetes and Digestive and Kidney Diseases (DK-34275 and DK-39255) to Joel M. Weinberg is gratefully acknowledged.
- © 2002 American Society of Nephrology
References
Jump to section
More in this TOC Section
Cited By...
- Human Cytomegalovirus Directly Induces the Antiviral Protein Viperin to Enhance Infectivity
- Cytopathic effects of the cytomegalovirus-encoded apoptosis inhibitory protein vMIA
- Endocytosis of Cadherin from Intracellular Junctions Is the Driving Force for Cadherin Adhesive Dimer Disassembly
- Nucleoplasmic {beta}-actin exists in a dynamic equilibrium between low-mobility polymeric species and rapidly diffusing populations.
- Hsp72 Interacts with Paxillin and Facilitates the Reassembly of Focal Adhesions during Recovery from ATP Depletion
- Mechanism of Actin Polymerization in Cellular ATP Depletion
- Transitioning to Therapy in Ischemic Acute Renal Failure