For years, researchers have been perplexed by the differences in regeneration capabilities between animal species [1-6]. On the one hand, some species such as Planaria, Hydra or starfish can readily regenerate lost organs or body parts in a remarkable way. In contrast, other species, such as most vertebrates, Drosophila or C. elegans have limited regenerative capacities. Nonetheless, all species have some capacity to heal wounds produced by external factors encountered during their existence. What, if any, is the relationship between these two processes? There are some indications that these processes might differ at the cellular and molecular levels. For example, in sea urchins it has been suggested that healing of broken spines occurs by a morphallactic mechanism involving recruitment of differentiated cells, while regeneration of removed spines and pedicellaria occurs by an epimorphic process involving undifferentiated precursors [7]. And in zebra fish, studies on fin regeneration have found that a fish mutant (dob- devoid of blastema) fails to regenerate the fin but has normal wound healing responses [8]. More recently it has been shown that both wound healing and regeneration in the axolotl are dependent on epithelial/mesenchymal interactions, the formation of the wound epidermis, the restructuring of the extracellular matrix and other cellular/molecular events [9]. These similarities led these authors to suggest that the regeneration capacities of these amphibians is basically a process of superhealing[9].

A related question that surfaces is whether regenerative processes in animals where regeneration is a common occurrence that follows autotomy or programmed organ loss, differ from those associated with wound healing? This has been suggested to occur in some animals where autotomy usually occurs in localized breakage zones giving rise to the possibility that the regeneration capacities of cells within these zones differ from those of cells in other parts of the animal. In fact, it has been proposed that animals that autotomize organs or appendages, such as echinoderms and arthropods, show a higher potency for regeneration [10,11].

In answering these questions, it is crucial to determine what mechanisms are shared or differ between the processes of organ regeneration and that of wound healing. We have now used the holothurian or sea cucumber, Holothuria glaberrima, to study the relationship between the two processes. Holothurians are echinoderms, a group that comprise animals renowned for their regeneration capabilities [12,13]. When exposed to noxious stimuli, H. glaberrima undergoes a programmed evisceration process where most of the digestive tract and associated organs are expelled from the body cavity [14,15]. Following evisceration, it then regenerates new viscera, being the digestive system the first to be formed. Our group has used this model system extensively, showing that, following evisceration, a new intestine forms from a thickening in the remaining mesentery [14,15]. The initial intestinal primordium consists of a solid tube that connects the esophagus with the cloaca. Eventually, a lumen is formed and the tube is organized into a functional intestine. The process of intestinal regeneration in H. glaberrima includes, cell division, cell migration, extracellular matrix remodeling and muscle de-differentiation [14,16-19].

We now extend the use of this model system to study wound healing and to compare the events that take place during wound healing with those previously shown to occur during organ regeneration. To this end, a technique was developed to be able to cause an incision in the bodywall. In particular, the incision was made from the coelomic cavity side of the bodywall; a region that would normally not be subjected to wounds generated from the external environment and therefore cannot be considered a normal programmed event. The incision damaged nerve, muscle, dermal tissues and coelomic epithelia among others. Healing of the wound was then studied during the following 4 weeks. Changes in cellular and tissue elements during healing were evaluated using classical dyes, immmuohistochemistry and phalloidin labeling. Our results show that wound healing in these organisms makes use of the same mechanisms previously shown to occur during organ regeneration, to restore epithelial, dermal, and muscular tissues.

Figure 1 Bodywall tissue organization. Longitudinal sections of H. glaberrima bodywall depicting its major components, (A) stained with Toluidine blue and (B) double labeled with rhodamine-labeled phalloidin (red) and monoclonal antibody against holothurian collagen (more ...)

Using Toluidine Blue, we were able to analyze the changes that occurred in the bodywall tissues during the wound healing process (Fig. 2). At 2 dpi the wound site presented a V-shaped structure with a mean distance across the dense connective tissue most proximally to the coelom of 386 μm; the inside edges of this wound had been layered with scarring tissue. The longitudinal muscles and the inner connective tissue of the dermis were exposed and tissue debris and tattoo ink were present within the surrounding area. The vicinity of the wound had been infiltrated by a considerable number of basophilic spherule-containing cells. Interestingly, the staining pattern near the edges had changed, staining darker shades of fuchsia than in non-injured animals. At 6 dpi, the staining near and at the wound edges reached its uppermost intensity coinciding with a substantial decline in the amount of debris at the injury site and its periphery. This stage was also characterized by having an accumulation of provisional matrix in the wound that served as scaffolding, bridging the separation created in the dermis as a result of the injury. This scaffold stained light fuchsia and was infiltrated by basophilic cells; also noticeable, was the presence of small nuclear bodies. In addition, threads of connective tissue appeared to anchor the muscle terminals to the scar. Fusion of the dermal gap at the injury site was completed by 12 dpi, when the dermal substratum beneath the longitudinal muscles was complete. The location where the circular muscle band had been ruptured was occupied by a dense basophilic connective tissue with large number of tightly packed cells. The segment of connective tissue that was in direct contact with the coelomic contents was nearly completely covered by coelomic epithelium that was contiguous with that of adjacent terminals from the circular muscle bands. Nevertheless, dermal tissue architecture was not similar to that of a normal bodywall. It is noteworthy to mention that from 6–12 dpi the laminae of dermal connective tissue in the areas peripheral to the wound appeared separated from one another, suggesting a diminished cohesiveness. Tissue organization was noticeably improved by 20 dpi, when a solid bridge of connective tissue between the muscle stumps had formed and some muscle fibers were present within it; however, there were areas that still consisted mostly of connective tissue. Additionally, dermal tissue architecture improved dramatically by this stage, appearing much more homogeneous in texture and in cellular composition, resembling more closely that of a normal bodywall. Restoration of all tissue components to a practically normal state was attained at 28 dpi; even so, it is important to mention that even at this stage some tattoo ink can still be observed within the bodywall dermis.

Figure 2 Stages in wound healing following transection of H. glaberrima bodywall. Longitudinal tissue sections of injured body wall at (A) 2, (B) 6, (C) 12, (D) 20 and (E) 28 days post injury (dpi) were stained with Toluidene Blue. (A) At 2 dpi the area is filled (more ...)

Figure 3 Wound healing of longitudinal muscle following transection. Body wall sections from animals at (A) 2, (B) 12, (C) 20 and (D) 28 days post-injury (dpi) were labeled with rhodamine-labeled phalloidin to determine the temporal changes in the longitudinal (more ...)

Figure 4 Patterns of cell proliferation in response to wound healing. Longitudinal sections were labeled with an antibody against BrdU (green) and Hoescht dye (red) to determine cell proliferation in the longitudinal muscle (LM) and coelomic epithelia (CE). (A) (more ...)

Figure 5 Changes in collagen immunoreactivity during wound healing. Longitudinal bodywall sections of (A) non-injured animals and animals at (B) 6, (C) 12, and (D) 20 days post-injury (dpi) labeled with an antibody against holothurian collagen. The normal distribution (more ...)

Figure 6 Changes in spherule-containing cells during wound healing. Tissue sections (6 dpi) stained with (A) Toluidine blue and (B) Monoclonal antibody Sph2, show two sub-populations of spherule-containing cells: (A) morulas and (B) spherulocytes within the body (more ...)

We observed changes in both cellular populations during the wound healing process. The spherulocyte population appeared to increase near the injury site. Spherulocytes were also observed to go through a striking alteration in their morphology. Cells that were present distant to the injury site, similar to those found in controls and non-injured animals, were round or oval and possessed well defined spherules within their cytoplasm. When these cells were in close proximity to the injury their spherules were no longer discernable and the labeling appeared less intense and seemed to extend into the extracellular matrix. To quantify the changes in the spherulocyte population, the number of cells were counted near the injury site and compared with cell numbers distant to the site and in control animals (Fig. 6E). By 2 dpi, the spherulocyte population already showed a twofold increase in numbers. The changes in cell number and morphology continued during 6–12 dpi. At latter stages (20–28 dpi) cell numbers returned to levels comparable to those of non-injured animals, however, some cells still showed morphological alterations; their spherules were not clearly discernable. An important observation that needs to be emphasized was that while the number of spherulocytes near the injury site increased during the wound healing period, a 32 % decline in their numbers was detected in areas of the dermis distal to the injury site, suggesting that cells were migrating from non-injured areas to the injury site.

The second spherule-containing cell population, the morulas, was also observed to increase in number in response to injury (Fig. 6F). Clustering of these cells near and at the dermal edge created by the incision began as early as in the 2 dpi stage, when a three-fold increase in their numbers could be observed. This peak in cell number was maintained until 12 dpi, after which cell numbers declined to quantities similar to those in normal animals. These cells also showed a parallel decline in their numbers distally to the injury site at stages where an increase from basal levels was registered proximally. A distinctive characteristic of this cell population was that, although they appeared to migrate to the injury site, only a few showed a change in morphology.

Figure 7 Formation of spindle-like structures (SLS) by muscle cells during wound healing. (A) Double labeling of muscle tissue with rhodamine-labeled phalloidin (green) and Hoescht (red) shows the presence of SLS at the injury site. The lack of nuclei within or (more ...)

Quantification of SLS formation was performed to determine the temporal aspects of changes in the muscle structure (Fig. 7H). SLS's were counted per field of view at different wound healing stages. The peak in SLS formation was shown to occur at 6 dpi. Not only were there more SLS's counted per area, but these could be detected as far as 6 mm from the lesion point. By 12 dpi muscle fibers organization was greatly improved and the presence of SLS's confined once more to muscle terminals. From 12 dpi and forth SLS's continued declining, where at 28 dpi little or none could be seen, and muscle organization was virtually comparable to that of an unlesioned animal. If SLS's were present at this stage these were mostly restricted to the injured muscle or to where the overlapping muscle bundles from opposing stumps were in contact. At all stages, formation of SLS's in the longitudinal muscle was predominantly restricted to the coelomic surface of the muscle.

Here we have described the process of wound healing in the echinoderm H. glaberrima and some of the cellular events that take place during this process. We will first discuss the healing process itself in view of what has been described in the literature for other echinoderm species. However, our main interest in undergoing this study was to compare the cellular events that occur during wound healing to those that occur during organ regeneration. By far what has attracted the attention of scientists to the echinoderms has been their capacity to regenerate following autotomy, where the detachment of the limbs or organs serves a defensive function [20]. In most cases regeneration following autotomy involves not only the restitution of particular tissues but the complete formation of an organ or arm. Nonetheless, certain aspects of regeneration, particularly those that occur early in the process and are involved with the healing of the cut or severed surface, can serve as points of comparison to our results.

Therefore, the main part of this discussion will focus on the four events studied: (1) cell proliferation (2) increase in spherule-containing cells (3) ECM remodeling and (4) SLSs formation (and its association with muscle dedifferentiation) and the comparison to intestinal regeneration.

Our results are in accordance with an important role for morula cells in the wound healing process of holothurians. The increase in number at the injury site during the early stages of wound healing surely suggests that they are somehow involved in the process. Moreover, the fact that their numbers in nearby (non-injured) areas diminish also suggest that morula cells are being actively recruited into the injured area. Differences of these studies with others [21] could be due to the latter only doing superficial injuries or to species differences.

When wound healing is compared to intestinal regeneration one observes, that in contrast to their role in wound healing, morula cells do not appear to play an important role in intestinal regeneration. Although they are present within the regenerating intestine, their numbers actually decrease during the initial stages of regeneration [14].

Spherulocytes appear to be also involved in the process of wound healing. Moreover the changes observed in these cell types during wound healing parallel those observed during intestinal regeneration [24]. In both wound healing and regeneration the number of these cells, localized in the connective tissue, increases. In addition, during both processes they also undergo a dramatic change in their morphology. Complementary to this change in morphology is the alteration seen in the labeling pattern, which is no longer restricted to the interior of the cells but seems to be on the extracellular matrix suggesting that they are releasing their vesicular components into the extracellular matrix. Finally, during wound healing spherulocytes appear to migrate from the adjacent dermis into the wound site. Similarly, during intestinal regeneration, spherulocytes appear to migrate into the regenerating intestine via the mesentery connective tissue [17].

Finally, our results provide for some interesting comparisons between the morula and spherulocyte populations. It has been proposed that the different spherule-containing cell populations reflected cells at different stages of maturation, where the deeply stained cells were younger, and as they matured, their granules were only weakly stained [23]. Based on this, we had previously proposed that spherulocytes were morula cells at a later stage in their maturation [24]. However, the present data questions this interpretation since they show that both morula and spherulocyte populations increase concomitantly, and both appear to migrate into the wound area. Moreover, less than 5% of the cells are labeled with both the antibody and the histological stain. Thus, the data is more consonant with both cell types being independent cell types rather than being the same cell type at different maturation stages.

Although the small amount of cellular proliferation that takes place during the early stages of wound healing might seem surprising, this is somewhat similar to what has been found to occur during regenerative events in holothurians. In our previous studies using BrdU labeling during intestinal regeneration in H. glaberrima, a small level of proliferation was documented during the early days of intestinal regeneration, while the increase in cellular proliferation occurs between 7–14 days [14]. Similar results have been documented in other species using other techniques; in Stichopus mollis by counting mitotic figures [25] and in Eupenctata fraudatrix using radioactive thymidine [26-29]. Thus, once again in the sea cucumber, the wound healing events are similar to those found during organ regeneration.

Another resemblance between wound healing and intestinal regeneration, is that in both processes the coelomic epithelium stands out as a center of cell proliferation. This was observed in our previous studies of intestinal regeneration in H. glaberrima where the division of the coelomic epithelial cells was much higher than those within the connective tissue [14]. Similar percentages in cellular division of the coelomic epithelia were found in this study following wound healing. The important role of the coelomic epithelia in cell proliferation during regenerative events seems to be a common theme to many echinoderm species, having been shown to occur in crinoid arm regeneration [30] and in arm and pyloric caeca regeneration in starfish [31,32]. Moreover, the proliferation of the coelomic epithelia appears to play a significant role in the restitution of echinoderm tissues since it has been proposed these cells give rise to other cellular types and not only coelomic cells [18,29,33,34].

Cell proliferation also plays an important role in the wound healing of other animals. In mammals, where it has been best studied, a proliferative phase takes place usually 3–5 days following injury [35,36]. Some of the proliferating cells are fibroblast that migrate and proliferate in response to growth factors and eventually play a role in the synthesis and deposition of ECM.

Previous experiments from our laboratory have shown that during regeneration of the intestine in H. glaberrima there is a striking remodeling of the extracellular matrix within the mesentery [16]. Among the features of this ECM remodeling are changes in ECM components, activation of matrix metalloproteases and the increased presence of cells involved in matrix degradation and matrix deposition [16,19,24]. One of the hallmarks of the ECM remodeling is the disappearance of ECM collagen, particularly from the mesenterial ECM close to the regenerating intestine. Our results with wound healing show noteworthy similarities to the events that occur during intestinal regeneration. In particular, the disappearance of collagen immunoreactivity in the wounded area and in areas adjacent to the wound, and the temporal scheme in which this occurs. Collagen immunoreactivity diminishes greatly within the two weeks that follow the injury and this decrease occurs not only in the tissue that has been wounded but extends to nearby tissue. During regeneration, collagen also disappears from the mesentery adjacent to the regenerating intestine and a gradient in its disappearance is also observed as one moves from the mesentery edge towards the bodywall.

That ECM changes occur during wound healing of the sea cucumber bodywall is not surprising, since such changes have been associated with the healing of wound in all, if not most, animals where it has been studied. For example, during skin wound healing in mammals, the ECM collagen is degraded and replaced with a temporary matrix [35,37]. This temporary matrix is eventually degraded once the tissues are regenerated and collagen is once again restituted [38]. Similar events have been documented in other organisms following injury [39]. However, one of the consistent findings that differentiate the holothurians from other species is the timing of the response. Collagen disappearance occurs during a 1–2 week period following either injury or evisceration and the reposition of collagen occurs 1–2 weeks afterwards. In contrast in mammals and other species, collagen is degraded and restituted much faster.

There are various similarities between the process of muscle dedifferentiation in intestinal regeneration and wound healing. First, cellular dedifferentiation as determined by SLS's formation occurs during the first and second week following the traumatic event. Second, some of the SLS's are eliminated into the coelomic cavity. An additional similarity is that the dedifferentiation process appears to occur in a gradient, showing high levels close to the injury site and lower levels as one moves distant from the injury site. Such a gradient suggests the presence of a soluble diffusing molecule that might be responsible for muscle dedifferentiation. In this respect, it is interesting that a soluble factor from serum that appears to play a similar role in the dedifferentiation of amphibian myocytes has been partially characterized [40].

Muscle dedifferentiation and formation of SLS's is not restricted to holothurian species. It has been documented in various other echinoderm species undergoing regenerative processes including visceral and longitudinal muscles of sea cucumbers [27,28,34,41,42] and crinoid arm muscle [34,42]. Likewise, elimination of the contractile apparatus and formation of de-differentiated cells from muscle cells has been documented in many other invertebrate and vertebrate species during regenerative procedures. Examples of these are the muscle cell dedifferentiation that occurs during annelid regeneration [43,44] and in amphibian tail and limb regeneration [45-47]. Thus, available data strongly suggest that muscle dedifferentiation is an important component of regeneration processes.

To summarize the comparison between the cellular events that occur during wound healing and those that occur during regeneration, we have prepared graphs showing the relative extent of the event during four weeks following the injury or evisceration process (Fig. 8). The similarities in terms of magnitude and timing of these events are striking. Except for the increase in spherulocytes all other events follow almost identical temporal sequences in regeneration and in wound healing. The difference in the spherulocytes can be ascribed to the fact that these cells appear to be migrating from the bodywall during the regenerative process and thus, take longer to reach the intestinal primordia. In contrast, these cells are quite numerous within the bodywall and therefore are readily available near the wound site. Nonetheless, it is important to emphasize that spherulocytes are still numerous during the second week of wound injury, thus, although the increase in number begins earlier than in regeneration, they still are present in high numbers at the same time their number peak during regeneration, suggesting a similar role in both processes.

Figure 8 Comparison of the cellular events that occur during wound healing and organ regeneration. The four cellular events highlighted in this study were compared to those that occur during intestine regeneration [14, 16, 19, 24]. For this the magnitude of the (more ...)

The graphs also serve to highlight the temporal sequence of the evisceration/wound healing related events. It seems, that the cellular event that occurs earlier in both processes is the dedifferentiation of muscle cells. The highest numbers of SLS's are usually observed within the first two weeks and decay thereafter. Other events peak during the second week of regeneration/wound healing. Of particular interest is the fact that little cell division occurs during the days following injury or evisceration and that most of the cell division occurs at the end of the first week or during the second week, thus suggesting that the initial cells involved in wound healing and regeneration processes are mainly cells already found within the tissues or cells that result from the dedifferentiation event. Experiments are in progress to determine the role that the dedifferentiated cell population plays in the regeneration and wound healing processes.

In contrast, there appears to be some notable differences among regenerating and non-regenerating animals in terms of ECM remodeling. In H. glaberrima ECM remodeling extends over weeks rather than days and the peak in both collagen disappearance as well as in its deposition appears to be much later than during wound healing in mammals. ECM remodeling has been highlighted as one of the mechanisms that is found to occur widely in animals capable of organ and limb regeneration. It has been shown to occur during newt limb regeneration [39,49-52] rat liver regeneration [53], Hydra head regeneration [54,55] and zebrafish fin regeneration [56]. In mice, ECM remodeling, and in particular the activity of the matrix metalloproteinases, has been shown to correlate with the regeneration ability of two mice species: those where the activity of the enzymes is higher and last longer show better healing of ear wounds [57]. Therefore, the timing and extent of ECM remodeling might be an important factor associated with the enhanced regeneration capacities.

Comparisons between regenerating and non-regenerating animals, in terms of the population of spherulocytes are not practical at this time. First, because, their role in holothurians is not clear. They have been associated to the production of a transitory ECM [23,24] but have also been implicated in defense mechanisms [58-60]. Second, because it is uncertain if they have analogous cell type in other species. Some investigators have suggested they are similar to mast cells [21,22,24] but this possibility remains undetermined. Nonetheless, the association of this cell type with the process of wound healing and regeneration in holothurians is worth further exploration, particularly in view of their capacity to release their vesicular contents which might include not only ECM components but growth factors, cytokines and other molecules associated with the cellular events that have been observed.

Out of the four cellular events that occur during holothurian wound healing and regeneration, the only one that appears to be prominent in regenerating species but not to occur in poor regenerators is the dedifferentiation of muscle cells. As discussed above, this process has been documented in various animal species ranging from annelids to amphibians. Moreover, the process could be the source of stem cells that eventually provide the different cell phenotypes necessary for the formation of the new organ [61,62]. Such a possibility has been explored by the laboratory of Tanaka [6,47] showing that, in amphibian tadpoles, cells that originate from muscle cell dedifferentiation gave rise to a variety of cell types. The process of dedifferentiation is not limited to muscle cells. In fact, other cells have been shown to dedifferentiate during the process of regeneration. The best documented is the dedifferentiation of the dorsal iris cells to produce the new lens in salamanders [5]. While recently, it has been shown that dedifferentiation of digestive cells into blastema-like cells occurs in Hydra during head regeneration [63]. Therefore, cell dedifferentiation might be one of the key processes that separate good regenerators from poor regenerators. Thus as has been suggested by Han and colleagues [61] "gaining access to developmentally regulated programs (dedifferentiation) lies at the heart of the regeneration problem in mammals".

The cellular events that occur during wound healing in H. glaberrima, namely the pattern of cell division, increase in the spherulocyte population, ECM remodeling, and muscle dedifferentiation are similar to those that take place during intestinal regeneration. Thus, our results show that, at least in H. glaberrima, but probably in most echinoderms that undergo organ or arm regeneration following autotomy, the same processes that are used for regeneration of complex structures are used for wound healing. Of the four events, muscle dedifferentiation appears to be a common mechanism to animals that posses a high regeneration capacity thus suggesting that this mechanism is of great importance to the regeneration process in metazoans.

Also, the extent of muscle regeneration was quantified by measuring the length of the regenerated fiber from the muscle stump to the farthest point in the regenerating front. For each animal, five to six measurements were taken and the mean value was calculated for each animal, pooled with other values for that same stage, then a final mean was calculated.

The labeling index was obtained by counting the number of cells labeled with BrdU divided by the total number of Hoescht-labeled nuclei per microscope field of view. For the longitudinal muscles and their epithelia a mean value for two corresponding areas in each animal was calculated. At least 3 different animals were used for each wound healing stage. Measurements were made at the injury site as well as proximal and distal from the injury site.

JSMR carried out most of the experiments and helped in the analyses of data and preparation of the manuscript. JEGA conceived the study, participated in its design, data analyses and helped in manuscript preparation. All authors read and approved the final manuscript.

Funded by NSF (IBN-0110692), NIH-MBRS (S06GM08102), NIH-RCMI (RRO-3641-01) and the University of Puerto Rico. JSMR is a fellow of the NIH-MARC Program.