Migrating Cells Retain Gap Junction Plaque Structure and Function

Abstract

Cell migration is an essential process in organ development, differentiation, and wound healing, and it has been hypothesized that gap junctions play a pivotal role in these cell processes. However, the changes in gap junctions and the capacity for cell communication as cells migrate are unclear. To monitor gap junction plaques during cell migration, adrenocortical cells were transfected with cDNA encoding for the connexin 43–green fluorescent protein. Time-lapse imaging was used to analyze cell movements and concurrent gap junction plaque dynamics. Immunocytochemistry was used to analyze gap junction morphology and distribution. Migration was initiated by wounding the cell monolayer and diffusional coupling was demonstrated by monitoring Lucifer yellow dye transfer and fluorescence recovery after photobleaching (FRAP) in cells at the wound edge and in cells located some distance from the wound edge. Gap junction plaques were retained at sites of contact while cells migrated in a “sheet-like” formation, even when cells dramatically changed their spatial relationship to one another. Consistent with this finding, cells at the leading edge retained their capacity to communicate with contacting cells. When cells detached from one another, gap junction plaques were internalized just prior to cell process detachment. Although gap junction plaque internalization clearly was a method of gap junction removal during cell separation, cells retained gap junction plaques and continued to communicate dye while migrating.

• connexins
• gap junction dynamics
• time-lapse microscopy

INTRODUCTION

Gap junction channels are cylindrical units composed of proteins called connexins (Goodenough et al. 1996). The protein sequences of several connexin gap junction proteins have been determined (Kumar and Gilula, 1986, 1996), and it is now evident that many cells express more than one of the 20 known members of the connexin family (Goodenough et al. 1996). Connexin 43 gap junction protein (Cx43), the most abundant gap junction protein in mammalian vertebrates (Goodenough et al. 1996; White and Paul 1999), is synthesized in the endoplasmic reticulum, oligomerized into hexameric hemichannels (connexons) in the Golgi, and then transported to the cell surface (Evans 1994; Goodenough and Musil 1993; Musil and Goodenough 1993). They unite with similar connexons from apposing cells to form gap junction channels, which, in turn, aggregate to form gap junction plaques. Gap junction channels provide pathways for the direct exchange of small molecules, including cyclic adenosine monophosphate (cAMP), Ca^2 +, and inositol triphosphate (Bedner et al. 2003; Bruzzone et al. 1996; Loewenstein 1981).

By regulating the passage of small molecules, gap junction channels are thought to play a pivotal role in orderly and effective cell migration, which includes the regulation of proper direction and rate of cell movement. For example, the trajectory of glioma cell migration in a two-dimensional culture appears to be dependent on cell-cell communication through gap junction channels (Aubert et al. 2006). Enhanced gap junction communication in neural crest cell populations is associated with an increased rate of cell migration (Huang et al. 1998) and inhibition of gap junctions has the opposite effect, in reducing the rate of cell migration (Ashton et al. 1999; Huang et al. 1998; Lo et al. 1999). It has been shown that inhibition of gap junction communication reduced migration of striatal subventricular zone cells to the olfactory bulb (Menezes et al. 2002), which would suggest the need for gap junction communication during coordinated cell movement. In contrast, it has been suggested that the process of uncoupling may actually trigger migration (Batten and Haar 1979).

Although the pattern of gap junction plaque distribution and coupling in adult and embryonic tissues has been extensively studied (Andries et al. 1985; Becker et al. 2002), and the life cycle of gap junction plaques in nonmigrating cell populations has been described (Laird 1996; Lauf et al. 2002; Lopez et al. 2001), the presence and functionality of gap junction plaques in migrating populations has not been demonstrated. It is clear, however, that communication during cell movement requires that the gap junctions must be retained or continually reassembled at new sites of cell-cell contact as cells migrate.

In this study, we provide the first characterization of the relationship between cell movement and gap junction plaque dynamics revealed by time-lapse microscopy in migrating cells. Adrenal-derived epithelial cells were used as a model system because they form large gap junctions and numerous annular gap junctions and are capable of migration. The internalization of gap junction plaques to form cytoplasmic annular gap junction packets was demonstrated during cell migration when cells detached from one another. The morphology, origin, and fate of these annular cytoplasmic structures have been previously studied by us and others (Jordan et al. 1999, 2001; Lopez et al. 2001) but not during cell detachment and migration. Our findings establish for the first time the presence of gap junction communication during migration, the retention of gap junction plaques at the initial sites of formation, and the internalization of plaques during cell process detachment.

MATERIALS AND METHODS

Cell Line

Human adrenal cortical tumor cells (SW-13), an immortalized adrenal epithelial cell line known to form numerous and large gap junction plaques, were used in this study. Cells were grown under standard conditions at 37°C with 5% CO₂ in Leibovitz's L15 medium (Gibco, Grand Island, NY) supplemented with 200 U/ml penicillin, 200 µg/ml streptomycin, 5 µg/ml fungizone (amphotericin B), and 10% fetal calf serum (Gemini Bio-Products, West Sacramento, CA).

Transfection with cDNA

To visualize gap junction trafficking in living cells, adrenal cells were transfected with cDNAs encoding for a fluorescent Cx43-GFP (green florescent protein) or GFP control vector (provided by Dr. M. Falk, Lehigh University). The Cx43-GFP was constructed by linking the GFP fluorescent reporter protein to the C-terminus of the rat Cx43 cDNA. This Cx43-GFP fusion protein has been characterized and found to assemble into gap junction plaques similar to wild-type Cx43 (Falk 2000; Lopez et al. 2001). Lipofectamine 2000 (Invitrogen, Carlsbad, CA) transfection reagent was used to establish cell populations that transiently expressed fluorescently tagged Cx43, empty vector, or GFP.

Cell populations, at 70% to 80% confluence grown in 35-mm culture dishes, were transfected in 2 ml of Opti-MEM medium (Gibco, Carlsbad, CA) containing 10 µl Lipofectamine 2000 transfection reagent and 4 µg of plasmid DNA (Cx43-GFP DNA or GFP DNA) for 6 h at 37°C in an atmosphere of 5% CO₂. The transfection medium was removed by gentle aspiration and the cells were washed with phosphate-buffered saline (PBS). Fresh L15 complete cell growth medium was added to the dishes, and the cells were incubated at 37°C in an atmosphere of 5% CO₂ for 48 h before imaging.

Migration Assay

Migration was measured with a wound healing migration assay technique (Rodriguez et al. 2005). Briefly, cells were plated in 6- or 12-well plates and grown to confluency, with care taken to insure that the cell density did not become great enough to allow the cells to form a secondary layer. A mechanical scratch was made down the center of the confluent cell monolayer with a 0.5 µl sterile pipette tip, which resulted in a denuded area within the well, and the monolayer was extensively rinsed with PBS to remove nonadherent cells and debris, after which fresh medium was added. A portion of the denuded area was imaged with phase microscopy at a location defined by a mark drawn on the underside of the plate. Subsequent images were taken at the exact same location, at 1 min and at 2, 4, 6, and 24 h post wounding. Throughout the imaging period the cells were maintained at 37°C. Multiple fields were photographed and the width of the denuded space was measured with computer-assisted analysis, using the MetaMorph program, at each selected time point. These experiments were performed in triplicate to produce a minimum of nine areas for analysis. The distance the cells migrated was calculated by measuring the width of the denuded space at each experimental interval. A minimum of two fields from three different wells at each time point were analyzed. The data are expressed as mean distance±SEM. A statistical comparison of means was calculated by a Student's t test and a p value of ≤0.05 was considered to be statistically significant.

Immunocytochemistry

Immunocytochemistry of gap junction proteins was performed as previously described (Oyoyo et al. 1997). Cells were grown to confluency on sterile coverslips. For the assessment of migration, cell populations were mechanically wounded and incubated at 37°C to allow the cells to migrate. The coverslips were then fixed, at various times (t = 1 min, 2, 6, or 24 h) after wounding, in 4% paraformaldehyde for 20 min at room temperature, permeabilized with cold acetone for 7 min, and washed several times with PBS. Primary antibody, rabbit polyclonal anti-connexin 43 (Zymed, Carlsbad, CA), was diluted (1:100) in a blocking solution (3% bovine serum albumin [BSA]), and then applied to the cells for 1 h at 37°C in a humidity chamber. After washing 5 times in PBS, the coverslips were incubated with a secondary anti-rabbit antibody (Alexa-488 or Alexa-594 [Molecular Probes, Carlsbad, CA]) diluted in 3% BSA blocking buffer (1:1000) for 1 h at 37°C. After five washes in PBS, Hoescht nuclear stain was applied for 2 min at room temperature, the coverslips were mounted in fluoromont G (Southern Biotechnology Association, Birmingham, AL) and imaged.

Quantitation of Gap Junction Plaques and Annular Cytoplasmic Vesicles

Gap junction plaque and cytoplasmic annular gap junction vesicle numbers were assessed in migratory cells at 1 min and 2, 4, and 24 h after wounding. Cells on coverslips were fixed, processed for immunocytochemistry, and captured with fluorescence microscopy. Cells at the wound margin and four to six cells distant from the margin were analyzed for the average number of gap junction plaques or cytoplasmic annular gap junction vesicles per cell. In some experimental populations, to aid in distinguishing annular and gap junction plaques, clathrin adaptor protein, AP-2, was detected with immunocytochemical techniques or the presence of calcein AM within the cells was used to aid in delineating the cell borders. The number of cells per field was determined by counting the number of Hoescht-stained nuclei present. These experiments were performed in duplicate and 10 fields were captured at the migrating edge and at four to six cells distant from the edge for analysis.

Gap junction plaques and annular gap junction vesicles were characterized with either an Olympus IMT2 or Nikon Microphot FXA fluorescence phase contrast microscope and the numbers were quantified directly from fluorescence microscopy images with the MetaMorph software program. A statistical comparison of means was calculated by analysis of variance (ANOVA) followed by a Student's t test or by the Student's t test alone and a p value of ≤ 0.05 was considered to be statistically significant. The data are expressed as mean±standard error of the mean (SEM).

Gap Junction Functional Assay

Scrape load dye communication

The functional state of gap junctions was assessed with an established gap junction–permeable dye, Lucifer yellow, cell-cell communication assay (Ed-Fouly et al. 1987). Cells were plated on coverslips and allowed to adhere and reach confluency. The medium was removed and cells were washed with Earl's buffered saline solution (EBSS). A gap junction–permeable Lucifer yellow dye solution (5 mg/ml in medium) or –impermeable rhodamine dextran in solution (5 mg/ml) was added to the cell monolayer and a scratch was made in the confluent monolayer with a sterile pipette tip. Following incubation in Lucifer yellow for 5 min at 37°C, the dye was removed, cells were washed, and then imaged immediately and at 20 and 40 min post wounding. Images were obtained with phase-contrast and fluorescence microscopy, and these images were overlaid to show the extent of Lucifer yellow uptake in the “wound” margin. Subsequent images were obtained at the same location to determine any additional Lucifer yellow dye transfer among adjacent cells.

Fluorescence recovery after photobleaching (FRAP) assay

The adrenal cells were grown to confluency on a glass coverslip mounted to cover an opening in the bottom of a culture dish (Lab-Tek, Naperville, IL). Cells were loaded with 2.0 µM green-fluorescent calcein AM (Molecular Probes, Eugene, OR) for 15 min at 37°C. Calcein AM is a membrane-permeable dye, which once in a viable cell is cleaved into a fluorescent membrane-impermeable form capable of passing through gap junctions but not through other plasma membrane channels (Pollmann et al. 2005). The calcein AM–loaded cell populations were scratched with a pipette tip to initiate migration and the debris and nonattached cell monolayers were removed by rinsing the monolayer three times with L15 medium (Invitrogen) supplemented with 200 U/ml penicillin, 200 µg/ml streptomycin, 5 µg/ml fungizone (amphotericin B), 10% fetal calf serum (Gembio), and 10 mM HEPES (Sigma) at pH 7.2. The cells were maintained at 37°C in an open-chamber microincubator (Harvard Apparatus) in medium. The FRAP assay was performed on cells visualized with a 40×(0.75 numerical aperture) objective on an inverted Olympus Fluoview1000 confocal microscope and images were captured with the FluoView software. Real-time bleaching was done with a 405-nm line on a laser diode and excited with a 488-nm line on the argon laser. Regions of interests were drawn around cells at the wound edge and cells located away from the edge, and these areas were bleached for 10 s with 100% laser output. The bleached and nonbleached (control) areas were sequentially imaged (3-s scan, every 10 s for 3 min). The FRAP was performed in duplicate at multiple times (36, 49, 55, 60, 72 112, 331, 339, 349 369, and 375 min) over a 6-h period after wounding. The extent of bleaching and postbleach recovery of fluorescence was analyzed with the MetaMorph software by measuring change in fluorescent intensity over time within the bleached region and a control region far removed from the bleached region. For statistical analysis, 20 cells located 4 or more cells away from the wound margin, 20 bleached cells at the margin, and 40 unbleached cells (controls) were monitored. In addition, three cells that were not in contact with other cells and therefore did not have gap junction plaques (negative control) were bleached and monitored. The data are expressed as the percent fluorescent recovery±SEM after bleaching.

Imaging of Cx43-GFP in Living Cells

Coverslips with cells expressing Cx43-GFP or GFP were placed into a temperature controlled FCS2 Bioptechs chamber (Butler, PA) and maintained at 37°C on a Zeiss Axovert 135 microscope. The culture medium was supplemented with 10 mM HEPES pH 7.2. Both differential interference contrast (DIC) and fluorescent microscopic images, obtained with standard fluorescein isothiocyanate (FITC) filters (chroma) with a 40×(numerical aperture 0.75) or 63×oil objective (numerical aperture 1.4), were collected at intervals of 3 or 5 min with a Hamamatsu ORCA ER camera. Focus was maintained by algorithms, which established maximum contrast on the DIC image prior to each exposure. The focal plane thickness (z-ordinate) was limited to 456 nm. The data sets were analyzed as pseudocolor movies and as individual TIF image files. Sequential images have been converted to movies.

RESULTS

Characterization of Cell Migration

Cell migration was induced by wounding the cell monolayer with a sterile pipette. Cells progressively migrated into the denuded space and after 24 h cells nearly refilled the wound area (Figure 1A and B). The cells appeared, with DIC and phase-contrast microscopy, to remain in contact with one another and to move in a “sheet-like” manner, rather than as single cells or loosely associated cell clusters. The suggestion that cells at the wound edge were in contact as they migrated into the wound area was supported by visualizing cortical actin in cells near the wound edge (Figure 1C). In these populations, the intimate contact between migrating cells was demonstrated and even in the cases where the cell bodies of cells were somewhat separated from one another, the cells still remained in contact at cell processes (Figure 1C).

Figure 1.  Characterization of cell migration. Cell migration was assessed with the wound-healing assay and then captured with phase-contrast microscopy. The wound area was decreased as cells migrated into the denuded space (time = 2, 6 and 24 h are shown) (A). The width of the denuded space was quantitated (time = 2, 4, 6, and 24 h) and graphed (B). Note that the migrating cells remain in contact with one another (arrows) and remain in a sheet-like formation and that the actin cytoskeleton (green) is present in protrusions of cells at the leading edge (C). Bar = 240 µm (A), 20 µm (C).

Gap junction plaques were seen between cells at the wound edge and their adjacent neighbors as they migrated into the denuded area (Figure 2A to D). The functionality of the gap junctions was demonstrated by Lucifer yellow dye communication studies in which cells were scrape loaded, and the dye was observed to spread from the cells at the cut edge to adjacent cells (Figure 2E, F). In these studies, rhodamine dextran, which was too large to move through gap junctions, did not spread to adjacent cells. In addition to the Lucifer yellow studies, gap junction function in migrating cells was monitored by analysis of fluorescence recovery after photobleaching (FRAP), in which photobleaching of cells loaded with green-fluorescent calcein AM, a gap junction channel–permeable dye, resulted in rapid recovery of fluorescence in the photobleached cells at the migrating cell front (Figure 3). Cell-cell communication was robust in cells at the leading edge in migrating cell populations evaluated at 3 different times post wounding ranging from 5 to 40 min for dye scrape loading analysis (Figure 2 E, F) and 11 different times ranging from 36 to 375 min for FRAP analysis (Figure 3I).

Figure 2.  Characterization of gap junctions and cell-cell communication during migration. Immunocytochemical localization of Cx43 at 1 min (A), 2 h (B), 6 h (C), and 24 h (D) after the onset of migration reveal that plaques are retained between cells during migration. The wound edge is indicated by the white dotted line (A to C) and nuclei (n) are stained with Hoescht dye. The gap junction-mediated communication of Lucifer yellow dye (green) was evident at 5, 20, and 40 min following creating the wound (E), whereas rhodamine dextran, too large to transfer via gap junctions, remains in cells at the wound site (F). Bar = 20 µm (A to D), 200 µm (E, F).

Figure 3.  Analysis of gap junction protein and fluorescence recovery after photobleaching (FRAP). Cell populations were wounded, which stimulated migration. The cells were allowed to migrate before incubation of the populations for 15 min in 2.0 µM green-fluorescent calcein AM in the medium. After removal of excess dye, cells located at the migrating edge and some distances from the edge were selected for photobleaching. The cells were allowed to migrate for 2 (A to D) or 6 (E to H) h. Fluorescence images were taken before (Pre-bleached), immediately afterward (bleached), and at 3 min after photobleaching (Recovery). Note that all photobleached cells rapidly recovered dye after photobleaching, whether they were located at the migrating edge (cells 1 and 3) or some distance from the wound edge (cells 2 and 4). Following the FRAP analysis, cells were prepared for immunocytochemical analysis of gap junction protein (D and H). Note gap junctions are present at sites of cell contact between cells at the leading edge as well as cells located distant from the leading edge. Cx43 gap junction protein (red, arrows) and calcein AM (green). Graph of average percent recovery in cells at the leading edge after photobleaching over a 6-h period after wound initiation (I). Bar = 30 µm.

We found no evidence that cell-cell communication was significantly reduced in the cell at the leading edge of the migrating population. For example, the average percent recovery of fluorescence at 180 s post bleaching in cells at the migrating front was not significantly different from that in cells located some distance from the leading edge (78.76%±0.04% versus 81.32%±0.05% at 2 h and 79.19%±0.04% versus 84.47%±0.18% at 6 h). As expected, cells that lacked gap junction plaques (isolated single cells in the culture) did not recover fluorescence after photobleaching (data not shown). Thus, the recovery of fluorescence following photobleaching is a clear indication of cell-cell communication via gap junction plaques in migrating cell populations and is in agreement with the immunocytochemical observation of gap junction plaques in migrating cell populations (Figures 2 and 3). Gap junctions were maintained at sites of cell contact, both at the leading front and in cell located a distance from the front (Figures 2 and 3; Table 1). At the times evaluated post wounding, no significant reduction was observed in gap junction numbers at the margin compared to cells in the interior (Table 1).

Table 1.  Gap junction plaque number

Plaques per cell	1 minute	2 hours	4 hours	24 hours
Margin	3.97±0.43	3.39±0.33	4.14±0.48	4.86±0.88
Recess	4.45±0.44	3.43±0.39	5.23±0.99	5.47±0.54

There were no significant differences in the average number of gap junction plaques between cells at the leading edge (margin) compared to those located four to six cells away from the edge (recess). These experiments were performed in duplicate and 10 fields were captured for analysis at 1 min and 2, 4, and 24 h after wounding. A student's comparison of means was calculated by ANOVA followed by a student t test and a p value of ≤ 0.05 was considered to be statistically significant. The data are expressed as mean±standard error of the mean (SEM).

Monitoring Cx43-GFP in Transfected Cells

To determine whether gap junction plaques are retained as cells migrate, multimode time-lapse imaging was used to analyze Cx43 gap junction protein dynamics in migrating cells. As anticipated from immunocytochemical data, the fluorescent Cx43 protein, Cx43-GFP, was localized to areas of cell-cell contact or within cytoplasmic Cx43-GFP puncta (Figures 4 and 5) in cells viewed 48 h post Cx43-GFP transfection. Cytoplasmic Cx43-GFP packets were similar in size as that seen for endogenous Cx43 gap junction packets and Cx43-GFP gap junction plaques, either small puncta or in longer linear arrays, were seen at sites of cell contact throughout the cell population.

Figure 4.  Visualization of gap junctions during cell migration. Two different cell populations shown with fluorescence and DIC microscopic overlay techniques (A to D and E to H). Cells remain connected though spatially altered during migration and gap junction plaques (green at an arrow tip between cells 1 and 2) are retained. Time (T) is in minutes following the initiation of imaging. Cells were imaged 48 h post transfection. Bar = 30 µm.

Figure 5.  Gap junction plaque internalization during cell detachment. Gap junction plaque internalization (arrow) was observed as a cell process detached from the cell body of another cell. Pseudocolor (A, C, E, G, I) and corresponding DIC-fluorescent overlay with (B′, D′, F′, H′, J′) and without (B, D, F, H, J) the plasma membrane of one of cell outlined are demonstrated. Cells were imaged 48 h post transfection. Bar = 5 µm. In the movie, gap junction plaque internalization can be followed as a cell process detaches from the cell body of another cell.

We evaluated gap junction plaque dynamics in the cell monolayer and found that gap junction plaques appeared initially as small puncta at the cell surface. Small Cx43-GFP particles (<150 nm) were observed to stream toward these existing plaques throughout the viewing period. Gap junction plaques were invariably remodeled, either by fusion with one another or by fragmentation (data not shown). Similar remodeling of gap junctions has been previously observed with GFP-tagged Cx43 in other systems (Jordan et al. 1999; Lauf et al. 2002).

In cell cultures wounded 48 h after Cx43-GFP transfection, gap junction plaques were observed to be retained at sites of cell contact in the migrating cell population, even though the cells shifted their relationship to one another as they migrated. As can be seen in Figure 4, for example, the Cx43-GFP gap junction plaque between cells 1 and 2 could be followed as the cells moved to fill the denuded area. In these time-lapse images, cell 1 immediately after the wounding was situated superior to cell 2. As the cells continued to migrate, they changed their relative position to each other, such that cell 1 became situated inferior to cell 2, but remained attached. The gap junction plaque connecting the two cells was retained as the cells continued to move toward and finally filled the denuded area (Figure 4). Once the denuded area was filled with cells, migration was halted and the cells no longer changed their relative position to one another (Figure 4C and D). In some cases, the cells invaded the denuded area without changing their relative positions to one another, and as in cells that shifted positions, the gap junction plaques were also retained (Figure 4E–H).

Cells of the monolayer were connected to one another by relatively long processes. This was true in both migrating and nonmigrating cell populations. Cells were observed to lose contact with one another by retracting these processes (Figure 5). In the instances where Cx43-GFP gap junction plaques between cell processes and the cell body of contacting cells were monitored with time-lapse imaging techniques, gap junction plaques were observed to internalize as cells detached from one another at such processes (Figure 5 and Figure 5 movie). As the processes were detached, the linear gap junction plaques could be seen to become cytoplasmic annular gap junction packets (Figure 5). Gap junction plaque endocytosis, rather than plaque component (hemichannel) dispersion within the membrane, was the method used for gap junction removal as cells lost cell-cell contact. It appeared that cells throughout the population were capable of removing gap junction plaques from the cell surface in this manner, because the number of Cx43 cytoplasmic gap junction packets, measured at the leading edge of migrating cells, were not significantly different from that in cells located a distance from the leading edge at 2, 4, 6, or 24 h after wounding. For example at 4 h, the number of cytoplasmic packets was (7.95±0.72) in cells leading edge and (10.1±1.1) in cells located a distance from the edge.

With time-lapse microscopy, the gap junction plaques were observed to internalize and form annular cytoplasmic vesicles, presumed to be annular gap junctions. These cytoplasmic Cx43-GFP–positive packets were observed over time to (a) fragment to form smaller packets, (b) join with other Cx43 positive puncta, or (c) disappear from view.

DISCUSSION

Gap junction communication has been extensively studied over the last decade and its role in multicellular processes, which depend on cell migration, is just beginning to be understood. In reviewing the literature concerning gap junctions and cell migration, one finds some controversy in that the relationship between migration and connexin expression has been demonstrated to be both direct (increasing gap junction expression correlated with increased migration) (Huang et al. 1998; Xu et al. 2006; Lo et al. 1999; Menezes et al. 2008) as well as indirect (increased gap junction expression correlated with decreased migration) (McDonough et al. 1999; Batten and Harr 1979; Momiyama et al. 2003). The association is most likely dependent on cell type and may reflect different aspects of migration (directionality versus rate). McDonough and colleagues demonstrated that glioma cell lines, which expressed more surface connexin 43, had a slower migration rate than cell lines with fewer surface gap junction plaques . Similarly, in MCF-7 breast cancer cells, an increase in connexin 26 surface junctions produced by transfection techniques was associated with a decrease in chemotaxis (Momiyama et al. 2003).

Developmental defects and severe cardiac problems are associated with alterations in the migration of sheets of cardiac neural crest cells in mice lacking Cx43 (Huang et al. 1998; Xu et al. 2006). Conversely, transgenic mice that overexpress Cx43 show enhanced neural crest cell migration, which would suggest that Cx43 gap junction plaques, and presumably cell-cell communication, enhance the motility of cells in vivo (Huang et al. 1998). Although gap junctions and cell-cell communication may play an important role in the regulation of cell movement, little is known about the capacity for continuing gap junction-mediated communication during migration.

One would presume that if cell-cell communication is important in regulating direction, speed, or the sequence of behaviors needed for cell migration, the mechanisms for direct intercellular communication must persist as cells migrate. In this study, we have used immunocytochemistry and time-lapse photography to monitor the presence and position of gap junction plaques at various time points during in vitro two-dimensional migration. Gap junction plaques were evident between cells at the leading edge and their contacting neighbors. In fact, the average number of gap junction plaques was not significantly different in cells at the leading edge and in those some distance from the leading edge. This was a somewhat unexpected result because at the wound edge there were fewer neighboring cells for potential cell-cell contact. This finding may suggest that there was an increase in gap junction plaque formation or a decrease in gap junction turnover in cells located at the leading front. In monolayer cultures, for example, there is an increase in the number of gap junction plaques as the number of neighboring cells increases. However, gap junctions do not occur at every site of possible contact. It is possible that the number as well as size of gap junctions are regulated by factors that interact with the connexin molecule. For example, it has been demonstrated that gap junction plaques formed with tagged Cx43, which fail to bind ZO-1, are larger than those formed from native Cx43 in which ZO-1 binding is intact (Girao and Pereira 2007). This observation has led to the suggestion that ZO-1 plays a role in gap junction size and turnover (Giepmans et al. 2001c). The cell may be therefore capable of autonomously regulating gap junction plaque number and size, independent of the number of neighboring cells. Regardless of these considerations, our data suggest that gap junction plaques with in retained in migrating populations and are thus is position to facilitate migration.

Although morphological techniques can be used to demonstrate the presence of gap junction plaques, they fail to clarify their functional status. It might be suggested, for example, that cells simply tolerate gap junctions during migration, whereas the junctions themselves do not participate in any meaningful way in the coordination of multicellular migration. We report, however, that molecules continue to diffuse between cells during migration, as assessed by FRAP analysis. Transfer of regulatory molecules is therefore possible, even at the leading edge where cells were most active in changing their relative position. It should be noted that calcein AM, the dye used in FRAP analysis, is a membrane-permeable substrate that becomes hydrolyzed by nonspecific esterases into non–membrane-permeable fluorescent polyanionic calcein, which is retained in the cytoplasm (Czyz et al. 2000). This fluorescent form of calcein is capable of passing through gap junctions, but not through other plasma membrane channels, and therefore has been widely used to study gap junction–mediated communication (Czyz et al. 2000; Pollmann et al. 2005; Tenopoulou et al. 2007; Wade et al. 1986). In our study, all cells in the population were initially labeled by calcein and all cells including those at the leading edge were able to recover fluorescence. Cells not in contact with other cells (and thus lacking the capacity for gap junction–mediated cell communication) did not recover fluorescence following photobleaching. The failure of such isolated cells to recover fluorescence confirms the need for cell-cell contact and the role of gap junction–mediated communication in our FRAP assays.

It is of interest to consider the capacity of the migrating cells to maintain the dynamics of gap junction plaque removal as a mechanism for regulating plaque number. Many studies have documented the presence of annular gap junctional profiles within the cytoplasm of normal and malignant cell types (Larsen 1977, 1983; Sasano et al. 2007). Similar to our live-cell imaging observations that gap junction plaques are dynamic, it has been previously demonstrated in a number of other cell types that Cx43 gap junction plaques separate into smaller plaques or coalesce with one another and that small secretory vesicles similar to the ones we observed could be seen streaming toward preexisting gap junction plaques (Gaietta et al. 2002; Jordan et al. 1999; Lopez et al. 2001). Furthermore, investigators have demonstrated Cx43 gap junction packets subsequently exiting from the central region of the plaques to enter the cytoplasm (Lauf et al. 2002; Piehl et al. 2007; Sosinsky et al. 2003). Some of these cytoplasmic packets were demonstrated to be the equivalent of the double-membrane “annular” gap junctions described with electron microscopy (Jordan et al. 2001; Laird 1996), thereby confirming the earlier suggestions that entire gap junction plaques or plaque fragments were internalized to form cytoplasmic annular gap junction vesicles. Most recently, quantum-dot microscopic techniques, which use the resolution of the electron microscope, were employed to demonstrate that the annular gap junctions seen with transmission electron microscopy (TEM) were identical to the Cx43 gap junction packets identified with immunocytochemical techniques (Giepmans et al. 2005). We now report that gap junction plaques are removed from the cell surface as cell processes disconnect. It should be noted that cell detachment was not a prerequisite for gap junction plaque internalization, because internalization was observed in cells that remained attached to one another.

We conclude that gap junction plaques are capable of being internalized as cells migrate and that internalized gap junction plaques may be subjected to additional regulatory processes. For example, the reported intimate relationship of lysosomes with cytoplasmic gap junction packets would be consistent with degradation of the internalized gap junction components by lysosomes (Laird 1996; Larsen 1977, 1983; Murray et al. 1981). Gap junction plaque internalization may be a post-translational means of regulating gap junction function in migrating cells as well as in nonmigrating cell populations.

There is evidence that precise regulation of gap junction number and function may be especially critical for control of cell behaviors that coordinate cell migration. For example, investigators studying Cx43 knockout mice and transgenic mice either overexpressing Cx43 or expressing a dominant-negative form of Cx43 noted the need for a precise level of Cx43 function for the regulation of coordinated neural crest cell motility (Huang et al. 1998; Xu et al. 2006). Inhibition of Cx43 in these studies inhibited neural crest cell migration, reduced directional migration, and decreased the number of cells targeted to the outflow tract. Conversely, in transgenic mice overexpressing Cx43, neural crest cell migration was enhanced, and this was associated with an excess of cardiac neural crest cells in the outflow tract (Huang et al. 1998).

The physical structure of the gap junction could serve as a platform for assembly of cytoplasmic elements needed during cell movement. In support of such a possibility, the cytoplasmic domains of the gap junction proteins have been demonstrated to interact with cytoskeletal proteins such as the actin-binding protein drebrin (Butkevich et al. 2004; Giepmans 2006; Majoul et al. 2007) and microtubules (Giepmans et al. 2001b; Levy and Holzbaur 2007), as well as with such molecules as the tight junction protein ZO-1 (Giepmans and Moolenaar 1998; Li et al. 2004) and src (Chung et al. 2007; De Vuyst et al. 2007; Duffy et al. 2007; Giepmans et al. 2001a; Lin et al. 2006). In addition to gap junction interactions with cytoskeletal and membrane structures to produce a platform for binding of components needed for migration, gap junction hemichannels in the membrane can be opened or closed by a number of factors (Ramachandran et al. 2007; Retamal et al. 2007; Spray et al. 2006; Srinivas et al. 2006), and have been shown to release molecules such as adenosine triphosphate (ATP) from the cell (Stout et al. 2002; Zhao et al. 2005). Thus, the release of molecules from the hemichannels, the physical structure of the gap junction, and the well-known role of gap junction channel cell-cell communication could all play a pivotal role in providing information or structural platforms needed by migrating cells. Our data do not rule out the possibilities of either hemichannel or physical structure as mechanisms. However the data are consistent with the gap junction being capable of allowing cell-cell communication as cells migrate.

In summary, the results of our correlated immunocytochemical, time-lapse, and migration studies support the hypothesis that gap junction plaques persist between cells migrating as a sheet. The fact that migrating cells retain their gap junctions would suggest the capacity for cell-cell movement of regulatory molecules critical to the process of wound healing. The observations of gap junction internalization and the presence of annular cytoplasmic gap junction packets suggest that migrating cells maintain their capacity for gap junction turnover. We conclude that all of the functions associated with gap junctions appear to be conserved in the migrating cell population.

Acknowledgements

The authors gratefully acknowledge Dr. Nalin Kumar's gift of antibodies and Dr. Matthias Falk's contribution of plasmids used in this study. This research was supported by NSF grant MCB-0444398 and 0647748.