We study both single cell wound healing and the healing of epithelial cell layers. Each of these is described in turn below.
Single cell wound healing basics
Following damage, cells can repair themselves. This repair is thought to entail at least two complementary processes, both of which depend on inrush of calcium through the site of damage (Wound figure 1): First, the plasma membrane is believed to be "patched" by fusion of organelles with each other at with the plasma membrane at the site of the wound. This patching membrane can serve as a temporary barrier that prevents further inrush of calcium and loss of cell contents through the hole created by wounding. Second, the cortical cytoskeleton (which contains among other things actin filaments and myosin-2) undergoes repair as a result of accumulation of actin filaments and myosin-2 around the wound site in a "contractile ring" (Wound movie 1). The contractile ring closes inward over time, replenishing actin filaments and myosin-2 at the site of the former hole and also expelling the patching membrane from the cell.
Wound figure 1--A schematic diagram representing the process of single cell wound repair as if viewed from the side.
Wound movie 1: A fluorescence, time-lapse movie showing actin filaments (in green) and myosin-2 (in red) during single cell wound repair. Note that the view is from the outside of the cell, looking in, which differs from that depicted in the diagram above. The dark area in the middle of the field of view is the wound. The field of view is about 110 um across and the movie shows about 3 minutes of the healing process. Consistently, the relative concentration of myosin-2 is higher at the wound edge while the relative concentration of actin filaments is lower, yielding a characteristic concentric ring pattern. This process was originally described in papers published by Craig Mandato and Bill Bement in Current Biology (http://www.sciencedirect.com/science/article/pii/S0960982299802619) and the Journal of Cell Biology (http://jcb.rupress.org/content/154/4/785.long).
The discovery of Rho GTPase activity "zones"
How is the local activation and accumulation of actin filaments and myosin-2 around wounds controlled? Helene Benink found that the small GTPases, Rho and Cdc42, are rapidy (within 15-20s) activated around wounds and then sort into concentric activity "zones", with the active Cdc42 circumscribing the active Rho (Wound movie 2). In this study, which was published in the Journal of Cell Biology (http://jcb.rupress.org/content/168/3/429.long), Helene also found that formation of the Rho and Cdc42 zones is necessary for formation of the contractile ring of actin filaments and myosin-2. Specifically, the active Rho is responsible for stimulating myosin-2 activation and recruitment at the wound site, while the active Cdc42 stimulates the assembly of actin filaments around the wound. Thus, the concentric pattern of the Rho and Cdc42 zones explains the concentric pattern of the myosin-2 and the actin. As described under the cytokinesis section of the "research interests" portion of this site, we subsequently found that zones of Rho and, in some cases, Cdc42 activity, also accompany cytokinesis.
Wound movie 2: A fluorescence, time-lapse movie showing active Rho (in green) and active Cdc42 (in red) during single cell wound repair. The field of view is about 110 um across and the movie shows about 3 minutes of the healing process. For more details, see the paper in the Journal of Cell Biology (http://jcb.rupress.org/content/168/3/429.long).
How is wounding coupled to activation of Rho and Cdc42? One obvious possibility is that the inrush in calcium is somehow stimulates Rho and Cdc42 GEFs (GEFs are proteins that stimulate GTPases). However, this remains an untested hypothesis. However, Emily Vaughan has identified one player that regulates Rho and Cdc42 during single cell wound repair. This is Abr, a dual GEF-GAP (Wound figure 2). Abr has been demonstrated to have GEF activity for Rho, Rac and Cdc42 (see, for example, http://www.pnas.org/content/92/22/10282.long) in vitro, implying that it can activate these proteins in vivo. Abr has also been demonstrated to have GAP activity for Rac and Cdc42 in vitro, implying that it can inactivate these proteins in vivo.
Wound figure 2--A schematic representing the domain organization of Abr.
As described in a 2011 paper in Current Biology (http://www.sciencedirect.com/science/article/pii/S0960982211000376), Emily found that Abr is recruited to wounds and localizes inside the zone of active Cdc42, where active Rho is concentrated (Wound figure 3). The recruitment of Abr is dependent on active Rho, suggesting that Abr and Rho engage in a positive feedback loop.
Wound figure 3--Fluorescence image showing the distribution of Abr (green) relative to active Cdc42 (red) at a single cell wound.
Remarkably, overexpression of wildtype (normal) Abr expands the Rho zone at the expense of the Cdc42 zone, expression of a GEF-dead Abr mutant reduces both the Cdc42 zone and the Rho zone, and expression of a GAP-dead Abr mutant results in mixing of the two zones (Wound figure 4). These results suggest that Abr has the job of both locally amplifying Rho activity and enforcing segregation of the Rho zone from the Cdc42 zone.
Wound figure 4--Fluorescence image showing the distribution of active Rho (green) and active Cdc42 (red) in control cells (Control) or cells subject to expression of wild type (WT Abr OE), GEF-dead (GEF-Dead Abr) or GAP-dead (GAP-Dead Abr) Abr.
What makes the contractile ring (and the associated Rho and Cdc42 zones) close inward? The obvious answer is that it is myosin-2 powered contraction. However, in a recent study published in Developmental Cell (http://www.sciencedirect.com/science/article/pii/S1534580712002791) Brian Burkel found that in fact the Rho and Cdc42 zones themselves provide information essential for the inward closure of the rings. That is, each of these GTPases is organized as a gradient, with preferential activation occurring near the leading (inward) edge of the zone, and preferential inactivation occurring at the trailing (outward) edge of the zone. One of the experiments that lead to this finding is shown in Wound movie 3. A photoactivatable probe for active Rho was employed to allow the visualization of the turnover of active Rho. At increasing times after wounding, the Rho activity persists longer and longer at the leading edge of the zone.
Wound movie 3: A fluorescence, time-lapse movie showing actin (in red) and active Rho (in green) during single cell wound repair made by Brian Burkel and George von Dassow. The active Rho is visualized by flashing the cell with a laser. This photoactivates the probe (ie makes it visible in the region irradiated with the laser). The disappearance of the probe reflects Rho inactivation. The field of view is about 90 um across and the movie shows about 3 minutes of the healing process. For more details, see the paper in Developmental Cell (http://www.sciencedirect.com/science/article/pii/S1534580712002791) or the accompanying video abstract on Youtube (http://www.youtube.com/watch?v=btCY4InNmsc).
The results above raise a number of fascinating questions: What other players and processes control the activation of Rho and Cdc42? What, besides Abr, is involved in the characteristic pattern concentric ring pattern exhibited by Rho and Cdc42? What occurs in the late stages of the healing processes? Do the membrane fusion events hypothesized to patch the membrane contribute to the cytoskeletal response?
Healing of epithelia
While the healing of epithelial cell layers has been studied for decades, much of the earlier work focused on analysis of epithelial cells cultured in vitro. More recently, however, it has become possible to study epithelial wound healing at high resolution in living organisms. We have begun such studies using the embryonic frog epithelium. Remarkably, Andy Clark found that the epithelium, when wounded, recapitulates many of the same processes seen in a single cell, but at a multicellular scale. Specifically, when a small wound is made in an epithelial cell (a wound too small to kill that cell) not only does the wounded cell response by activating Rho, and Cdc42, as seen in single cells, the adjacent cells respond as well by elevating Rho and Cdc42 activity on the sides of the cell nearest the wound in their neighbor. An example of this behavior is shown below (Wound movie 4); to get the whole story see Andy's paper published in Current Biology in 2009 ((http://www.sciencedirect.com/science/article/pii/S0960982209013219).
Wound movie 4: A fluorescence, time-lapse movie showing actin filaments (in red) and active Rho (in green) during wound repair of an epithelial cell layer. Note that the wounded cell responds by activating Rho and accumulating actin filaments around the wound while some of the neighbor cells respond as well. Specifically, they also elevate Rho activity and accumulate actin filaments at their cell-cell junctions that face the wounded cell. These junctions then move inward to help with the healing process. The field of view is about 200 um across and the movie shows about 3 minutes of the healing process.
When the wound completely destroys one or more cells in the epitheilium, all of the neighbors elevate Rho (and Cdc42) activity at their junctions nearest to the lost cell, recruit actin filaments and myosin-2, and then close inward over the space previously occupied by the ablated cell. An example of this is shown below in Wound movie 5.
Wound movie 5: A fluorescence, time-lapse movie showing actin filaments (in red) and active Rho (in green) during wound repair of an epithelial cell layer. The wound completely destroys the cell in the center of the field of view. The neighboring cells elevate Rho activity and accumulate actin filaments at the cell-cell junctions facing the ablated cells. This forms a continuous ring of actin filaments and myosin-2 (not shown) that closes over the space previously occupied by the ablated cell. The field of view is about 200 um across and the movie shows about 3 minutes of the healing process. For more details, see the paper in Current Biology (http://www.sciencedirect.com/science/article/pii/S0960982209013219).
Thus, some kind of signal is conveyed from the intial wound site to the neighboring cells. Other experiments show that this signal can be transmitted many cells away. What is the signal? We do not know but we are fairly sure that it works at least in part through calcium, just as for single cell wound healing except in this case the calcium is elevated not only at the wound site, but at cell-cell contacts distant from the wound. A simple example of this is shown in Wound movie 6, wherein a wound made in the center of one cell elicits calcium elevation at the cell-cell junctions of all of the neighboring cells and a slightly more complex example is shown in Wound movie 7, where it is seen that junctions not immediately border.
Wound movie 6. A fluorescence, time lapse movie showing calcium elevation in epithelial cells following wounding. Note that not only does calcium rise at the immediate edge of the wound, but also at the cell-cell junctions of cells bordering the wounded cell. For more details, see the paper in Current Biology (http://www.sciencedirect.com/science/article/pii/S0960982209013219).
Wound movie 7. A fluorescence, time lapse movie showing calcium elevation in epithelial cells following wounding. Note that not only does calcium rise at the immediate edge of the wound, but also at the cell-cell junctions of cells bordering the wounded cell. The mosaic labeling of the embryo (ie the introduction of different amounts of fluorescent probe into different cells) makes this particularly easy to see. For more details, see the paper in Current Biology (http://www.sciencedirect.com/science/article/pii/S0960982209013219).
Here again, a number of questions arise: how, exactly, do the epithelia transmit information from a wound in one cell its immediate neighbors? To the neighbors farther away? How is the increase in junctional calcium is triggered? Is the junctional calcium increase responsible for the subsequent rise in Rho and Cdc42 activity at the junctions? If so, how it is converted into Rho and Cdc42 activation.