Research ArticleCell Biology

Apoptotic Cells Activate the “Phoenix Rising” Pathway to Promote Wound Healing and Tissue Regeneration

See allHide authors and affiliations

Science Signaling  23 Feb 2010:
Vol. 3, Issue 110, pp. ra13
DOI: 10.1126/scisignal.2000634


The ability to regenerate damaged tissues is a common characteristic of multicellular organisms. We report a role for apoptotic cell death in promoting wound healing and tissue regeneration in mice. Apoptotic cells released growth signals that stimulated the proliferation of progenitor or stem cells. Key players in this process were caspases 3 and 7, proteases activated during the execution phase of apoptosis that contribute to cell death. Mice lacking either of these caspases were deficient in skin wound healing and in liver regeneration. Prostaglandin E2, a promoter of stem or progenitor cell proliferation and tissue regeneration, acted downstream of the caspases. We propose to call the pathway by which executioner caspases in apoptotic cells promote wound healing and tissue regeneration in multicellular organisms the “phoenix rising” pathway.


The ability to repair damaged tissues is essential for metazoan organisms (1, 2). Some organisms (for example, salamanders) have the ability to completely regenerate entire amputated limbs (35). In contrast, other organisms, such as humans, can only partially replace damaged organs (for example, liver regeneration) (68). Wound healing and tissue regeneration are complicated processes involving the coordinated efforts of many different cell types. The initial response in mammals suffering tissue injuries is generally considered an inflammatory one. Subsequently, new tissue formation and tissue remodeling complete the wound healing process. Because inflammatory cells were thought to be the “first responders” at the site of tissue injury, a long-held notion was that cells of the immune system, especially macrophages and neutrophils, were crucial in initiating and coordinating the wound healing–tissue regeneration process (9, 10). However, observation of intact wound healing in PU.1 knockout mice, which lack macrophages and neutrophils, showed that neither cell type is necessary for skin excision wound healing (11). Therefore, the initiating cellular and molecular events in wound healing and tissue regeneration remain unclear.

The stem cells in and around damaged tissues play a critical role in wound healing and tissue regeneration (1). It was generally assumed that factors released from damaged tissues mobilize and recruit these stem and progenitor cells to the damaged site, where they proliferate, differentiate, and eventually replace the damaged tissue. The released factors were thought to arise during the inflammatory process ensuing from the initial tissue damage. Previous studies have focused on the roles of immunoeffector cells, such as macrophages, that are activated by tissue injury and that secrete cytokines and growth factors that promote wound healing and tissue regeneration (1, 12). Growth factors such as hepatocyte growth factor (HGF) (13), fibroblast growth factors FGF7, FGF10, and FGF22 (14, 15), and transforming growth factor–β (TGF-β) (16, 17) play important roles in wound healing and tissue regeneration, as do small-molecule hormones such as acetylcholine (18), catecholamine (19, 20), and polyunsaturated fatty acids (21). However, which cell types and molecular mechanisms initiate the signaling cascades responsible for wound healing is unclear.

In this study, we hypothesized that dying cells in the wounded tissues send signals to stimulate the proliferation of stem or progenitor cells, which in turn start the process of tissue regeneration and wound healing.


Stimulation of stem and progenitor cells by dying cells in vitro and in vivo

We used irradiated mouse embryonic fibroblasts (MEFs) to simulate dying cells in wounded tissues and determined whether these cells stimulated the proliferation of cocultured firefly luciferase (Fluc)–labeled stem or progenitor cells (see fig. S1A for validation that the intensity of luciferase corresponds to cell numbers and fig. S1B for verification that the change in cell numbers is due to proliferation, not inhibition of cell death, of the stem or progenitor cells). In the coculture system, dying MEFs (4 × 104) significantly stimulated the proliferation of a small number (200) of Fluc-labeled murine epidermal keratinocyte progenitor (EKP) cells, neural stem cells (NSCs), or mesenchymal stem cells (MSCs) when compared with Fluc-labeled cells cultured alone or with nonirradiated, live MEF cells (P < 0.05) (Fig. 1A). We observed similar growth-promoting properties from various lethally irradiated human and mouse cells (fig. S1C), which suggests that growth-promoting activity toward stem or progenitor cells is a general property of dying mammalian cells.

Fig. 1

Stimulation of stem or progenitor cell proliferation by dying cells. (A) Lethally irradiated MEF cells were mixed with Fluc-labeled mouse EKP cells, NSCs, or MSCs, and Fluc-labeled cell proliferation was quantified with bioluminescence imaging (fig. S1). (B) EKP-Fluc cells were injected subcutaneously into the hind legs of nude mice either alone (left leg) or with lethally irradiated MEF cells (right leg). The EKP-Fluc cells were then imaged. The relative luciferase activity normalized against day 0, immediately after injection, is graphed (P < 0.01, t test, n = 5 total, days 1, 3, and 5). (C) About 2 × 105 EKP-Fluc cells were injected subcutaneously into the left (unirradiated) or right (irradiated with 18 Gy) hind legs of nude mice and monitored for growth. Signals at later time points were normalized to the signals on day 0 (P < 0.05, the entire course of experiment, one-way ANOVA test; n = 5 total).

We also confirmed the ability of dying cells to support the proliferation of progenitor cells in vivo. When 2 × 105 lethally irradiated MEFs were co-injected with 1000 nonirradiated Fluc-labeled mouse EKP cells into nude mice (subcutaneously into the right hind legs), the intensity of the luciferase signal, representing cell proliferation, was significantly increased relative to that of EKP-Fluc cells injected alone (subcutaneously into the left hind legs) (Fig. 1B). In fact, proliferation was only observed for EKP-Fluc cells that were co-injected with irradiated MEF cells; for EKP-Fluc cells injected alone, no proliferation was observed. In both groups, the luciferase signal weakened over time, presumably because of a generally nonpermissive host tissue environment.

The proliferation-promoting effects of dying cells were also evident from experiments in which EKP-Fluc cells were injected subcutaneously into lethally irradiated (18 Gy) or nonirradiated hind legs of nude mice (Fig. 1C). EKP-Fluc cells injected into irradiated hind legs showed a significant initial proliferation (increase in luciferase intensity) in the first week and, although the luciferase signals subsequently weakened, the cells were still detectable beyond day 103 after injection (Fig. 1C), indicating long-term engraftment. In comparison, EKP-Fluc cells injected into nonirradiated left hind legs showed minimal proliferation (Fig. 1C) and rapid signal attenuation to undetectable. The difference in the luciferase intensity between EKP-Fluc–injected irradiated (right hind legs) and nonirradiated (left hind legs) was statistically significant throughout the entire period of observation [P < 0.05, one-way analysis of variance (ANOVA) test; n = 5 mice].

These results underscore the importance of tissue damage and cell death in stimulating the proliferation of stem or progenitor cells. They also agree with the results of a study showing proliferation of individual myogenic stem cells in radiation- or drug-treated mouse tissues (22) and are consistent with long-established bone marrow transplant procedures in which donor hematopoietic stem cells engraft only into hosts who have been ablated of their original bone marrow cells through irradiation or chemical treatment (23).

A key role for caspases 3 and 7 in cell death–induced proliferation of stem and progenitor cells

Because of the requirement for regenerative capability among different tissues, the factors regulating cell death–stimulated stem and progenitor cell proliferation may be derived from a common process operating in dying cells. Therefore, we hypothesized that the molecular machinery involved in apoptosis, one of the major mechanisms of cell death in many different types of mammalian cells, might play a key role in initiating the paracrine cascades that lead to the stimulation of stem and progenitor cell proliferation.

We examined the possibility that apoptotic caspases contribute to cell death–stimulated progenitor and stem cell proliferation because these enzymes are essential proteases that function as the nexus of the apoptotic process (2430). We used caspase 3–deficient (Casp3−/−), caspase 7–deficient (Casp7−/−), or the doubly deficient Casp3−/−Casp7−/− MEFs (31, 32). Irradiation of any of these caspase-deficient cells reduced clonogenicity of these cells to a similar extent as wild-type cells, despite less induction of apoptosis markers relative to that of wild-type cells; this result indicated that the cells die through other mechanisms. We evaluated the proliferation of various Fluc-labeled stem or progenitor cells when cocultured with lethally irradiated caspase 3–deficient (Casp3−/−), caspase 7–deficient (Casp7−/−), or the doubly deficient Casp3−/−Casp7−/− MEFs (31, 32). Relative to wild-type fibroblasts, both lethally irradiated Casp3−/− (Fig. 2A) and Casp7−/− (fig. S2A) fibroblasts were less effective at stimulating stem or progenitor cell proliferation. The doubly deficient MEFs were even less effective at promoting stem or progenitor cell proliferation than were either of the singly caspase-deficient MEFs, consistent with the overlapping functions of caspases 3 and 7 (Fig. 2A).

Fig. 2

Critical role of caspase 3 in cell death stimulation of stem and progenitor cell proliferation. (A) Proliferation of 200 Fluc-labeled stem or progenitor cells mixed with lethally irradiated MEF cells were monitored in 24-well plates serially with bioluminescence imaging. The error bars represents SEM. In all three groups, the differences between the wild-type and Casp3−/− groups are statistically significant (P < 0.05, t test; n = 4 independent samples). The differences between the Casp3−/− groups and the Casp3−/−Casp7−/− groups are also significant (P < 0.05, t test; n = 4). (B) EKP-Fluc cells injected alone or with lethally irradiated MEFs were monitored with bioluminescence imaging. Relative luciferase activities from co-injected EKP-Fluc (in the right hind legs) normalized against those from cells injected alone (in the left hind legs) are graphed. The differences between the two groups are significant (P < 0.01, one-way ANOVA test; n = 5). (C) Growth of EKP-Flu cells in wild-type and Casp3−/− mice. Relative luciferase activity in the irradiated legs normalized against those in the nonirradiated legs (P < 0.05, n = 5 each in wild-type and knockout mice, one-way ANOVA test). (D) MEF cells induced tissue growth in a modified DIVAA assay. Host tissue growth into silicone cylinders was quantified by FITC-lectin staining of endothelial cells in the vasculature. Signals were normalized against PBS control. Cylinders were prepared with PBS, wild-type MEFs, or Casp3−/− MEFs mixed with BME before implantation. The differences in tissue growth into the cylinders containing the wild-type MEFs versus the PBS control and versus the knockout MEFs were significant (*P < 0.05, t test; n = 6 cylinders in six different mice for each group).

We also examined the importance of caspase 3 and caspase 7 in vivo for promoting stem or progenitor cell proliferation with three paradigms. The first experiment was similar to that described for Fig. 1B. Lethally irradiated wild-type or Casp3−/− MEFs (2 × 105) were co-injected with 1000 EKP-Fluc cells into the right hindlimbs of mice and compared to the left hindlimbs, which were injected with the same pair of cells, only nonirradiated. The animals co-injected with the wild-type MEFs exhibited a significantly greater increase in luciferase intensity, indicating proliferation of the EKP-Fluc cells, compared to the animals co-injected with the Casp3−/− MEFs (P < 0.05, n = 5 mice), which showed minimal proliferation (Fig. 2B). Similar results were also obtained when lethally irradiated wild-type and Casp7−/− MEF cells were co-injected with EKP-Fluc cells, although the difference in the ability to stimulate proliferation was smaller than that observed between Casp3−/− and wild-type MEF cells (fig. S2B).

A second in vivo model that we used to examine the importance of caspases in the promotion of stem or progenitor cell proliferation was a radiation-induced wound healing model with transgenic Casp3−/− and Casp7−/− knockout mice. These mice have decreased tissue apoptosis in response to various stimuli (31, 32). Furthermore, Casp3−/− mice have profound developmental defects in the brain as a result of reduced brain cell apoptosis (31). About 1000 EKP-Fluc cells were injected subcutaneously into irradiated (right) and nonirradiated (left) hind legs of mice. The irradiated tissues in wild-type mice (in C57BL/6 background) showed significantly more capacity to promote EKP-Fluc proliferation (Fig. 2C) relative to irradiated tissues in Casp3−/− mice (also in C57BL/6 background), which showed minimal growth-promoting ability toward the injected EKP-Luc cells during the observation period. Similar results were also obtained for Casp7−/− mice (fig. S2C), indicating that caspase 7 also participates in the same process.

We evaluated the importance of caspase 3 in an experimental model of tissue regeneration: the directed in vivo angiogenesis assay (DIVAA). We placed lethally irradiated wild-type or Casp3−/− MEFs (mixed with basement membrane extract) into silicone cylinders and implanted the cylinders subcutaneously into nude mice. After 2 weeks, we killed the mice, removed the implanted cylinders, and quantified host tissue growth by measuring vascular growth into the cylinders. Whereas irradiated wild-type MEF cells induced significant vascular growth (P < 0.05), irradiated Casp3−/− MEF cells induced minimal vascular growth (Fig. 2D). The vascular growth in cylinders containing irradiated Casp3−/− MEF cells was not significantly different from the growth in the negative control cylinders containing only phosphate-buffered saline (PBS).

Evidence for the involvement of caspases 3 and 7 in skin wound healing

To determine the roles of caspases under more physiologically relevant circumstances, we performed two additional experiments in vivo. In the first experiment (Fig. 3, A and B), we examined the effect of caspase 3 or caspase 7 deficiency on the healing excision wounds in mouse skin. Excision wounds were created in wild-type, Casp3−/−, or Casp7−/− mice in the C57BL/6 background through punch biopsy (Fig. 3 and fig. S4). The rates of wound healing in these mice were quantified by measuring wound diameters. Casp3−/− mice exhibited a reduced rate of repair of dorsal skin wounds relative to that of wild-type mice (Fig. 3A). All of the wounds in wild-type mice were healed by day 9. In contrast, the time for complete wound closure in Casp3−/− mice was 14 days. We conducted immunofluorescence analyses to visualize the rate of wound closure by staining the skin at the wound site with an antibody against cytokeratin 14, which is a marker for skin epithelial cells, a key cell responsible for basal layer formation during epidermal closure. It is clear from the immunofluorescence staining (Fig. 3B) that skin epithelium recovery around the wound was markedly attenuated in caspase 3–deficient mice. By day 6, when there was already complete epidermal closure in wild-type mice, as shown by continuous cytokeratin 14 staining across the wound gap, there was little sign of closure in the wound gap of caspase 3–deficient mice. Staining for cytokeratin 6, which is a marker for keratinocytes and indicates epidermal cell migration, showed a similar reduction in the progression of wound closure in the Casp3−/− animals (fig. S3). The exact pattern of staining was different from that of cytokeratin 14.

Fig. 3

Key roles for caspase 3 in skin wound healing. (A) Healing of 4-mm circular excision wounds in C57BL/6-derived wild-type and Casp3−/− mice. Wound healing was monitored through periodic measurement with a caliper. Data are plotted as the changes in the average areas ± SEM of wounds from the two groups of eight mice with two wounds each (P < 0.01, one-way ANOVA test; n = 16 wounds in each group). (B) Full-thickness skin samples containing entire wound sites were biopsied from wild-type and Casp3−/− mice at different time points after punch biopsy. Paraffin-embedded samples were sectioned and analyzed with antibodies against cytokeratin 14 (K14), a marker for skin epithelial cells. Hematoxylin and eosin (H&E) staining shows basic tissue and cellular structures. Scale bars, 500 μm. (C) BrdU staining in skin biopsies containing full-thickness wounds. Scale bars, 200 μm. BrdU staining was quantified by counting and averaging four randomly chosen 100× fields on six mice. The error bars represent SEM; P < 0.01, t test, n = 6.

Staining with bromodeoxyuridine (BrdU), which was administered to mice shortly before they were killed, confirmed that cellular proliferation was significantly attenuated in wounded skin tissue of Casp3−/− mice (Fig. 3C). Because BrdU incorporation is a direct indicator of DNA synthesis during cellular proliferation, these data are consistent with a defect in Casp3−/− mice of the cellular proliferation that should occur during normal skin wound healing. Quantification of representative fields indicated that there was at least a factor of 10 reduction in BrdU incorporation in Casp3−/− mice (Fig. 3C) on the first day after skin biopsy. Experiments with Casp7−/− mice showed a similar defect in skin wound healing and cellular proliferation (BrdU incorporation) (fig. S4).

Evidence for the involvement of caspases 3 and 7 in liver regeneration

We examined the importance of caspases in liver regeneration in a well-established partial hepatectomy model in mice (33). Two of the four lobes of liver (the median and the left lateral lobes, which account for about two-thirds of total liver weight) were surgically removed from wild-type, Casp3−/−, or Casp7−/− mice and the mice were allowed to recover (Fig. 4 and fig. S5). When measured in terms of liver weight, wild-type mice regenerated lost liver tissue effectively, reaching about 78% of presurgery weight by day 10 (Fig. 4A). In comparison, liver weight in Casp3−/− mice remained at only 60% of presurgery level (Fig. 4A) and in Casp7−/− remained at 57% (fig. S5A), indicating a substantial defect in liver regeneration. Quantification of BrdU staining indicated that, when compared with wild-type mice, there was at least a 50% reduction in cellular proliferation in Casp3−/− mice (Fig. 4B) as well as in Casp7−/− mice (fig. S5B).

Fig. 4

Key roles of caspase 3 in liver regeneration after partial hepatectomy. (A) Assessment of liver regeneration after partial hepatectomy by measuring liver weight. The ratios of the average weight (± SEM) of regenerated livers (from mice that have undergone partial hepatectomy) against unresected liver weight are graphed. (P < 0.01 on and after day 2, t test; n = 5). (B) Cell proliferation (BrdU-positive cells) at different time points after partial hepatectomy. Representative images are shown in the upper panels (scale bars, 100 μm). Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). The lower panel shows ions of BrdU-positive cells by counting the average number of BrdU-positive cells in four randomly chosen 200× fields. Average and SEM are plotted. The difference between the two groups is statistically significant on days 2 and 5 (P < 0.01, t test; n = 4 mice in each group).

The role of calcium-independent phospholipase A2 in cell death–induced tissue regeneration

Thus far, our in vivo results provide evidence for the involvement of caspases 3 and 7 in wound healing and tissue regeneration. We wanted to determine the downstream factors by which caspases 3 and 7 stimulated cellular proliferation and the tissue regeneration process. It seemed likely that the dying cells released factors that promoted stem or progenitor cell proliferation; thus, we focused on caspase-activated pathways that would release growth-promoting factors. One candidate is calcium-independent phospholipase A2 (iPLA2), whose activity is enhanced by caspase 3 or 7 cleavage and the activation of which increases the synthesis and release of arachidonic acid and lysophosphocholine from apoptotic cells (3436). Arachidonic acid is of particular interest because it is a precursor to prostaglandin E2 (PGE2) (37, 38), a stimulator of stem cell proliferation (3944), tissue regeneration, and wound healing (4547). Therefore, we determined the rate of arachidonic acid release from wild-type and Casp3−/− MEFs. Ionizing radiation, which induces increasing amounts of apoptosis in wild-type MEF cells from 24 hours after irradiation, induced a significant amount of arachidonic acid release by 48 hours from wild-type MEFs (P < 0.01 n = 3) (Fig. 5A, left panel). However, this release was significantly less for irradiated Casp3−/− MEFs, which confirms the previous studies (3436). Western blot analysis confirmed a caspase 3–dependent cleavage of iPLA2 (fig. S6A), which is also consistent with previous observations (3436).

Fig. 5

A role for caspase 3–mediated activation of iPLA2 in cell death stimulation of stem or progenitor cell proliferation. (A) Left panel: Arachidonic acid (AA) release from wild-type and Casp3−/− MEF cells with or without irradiation. The differences in AA release between irradiated wild-type and casp3−/− MEF cells are statistically significant at 48 hours (P < 0.01, t test; n = 3). The graph was plotted as relative concentrations of [3H]arachidonic acid in the supernatant, which were derived by normalizing [3H]arachidonic acid concentrations at different time points against those in the supernatant taken immediately before irradiation. Right: PGE2 concentration in the supernatants of nontreated and irradiated wild-type and Casp3−/− MEF cultures. Data are plotted as the average and SEM; statistically significant differences are noted (t test, n = 3). (B) Role of iPLA2 in cell death–mediated stimulation of EKP cell proliferation in vitro. Data are plotted as the average and SEM (n = 5, t test). (C) The importance of iPLA2 in tissue growth in a modified DIVAA assay. Top: Quantification of host tissue vasculature from cylinders containing various MEF cells. All the data were normalized against the PBS control group, which serves as the negative control. Statistically significant differences are noted (t test, n = 6). Bottom: Photographs of representative cylinders after removal from the mice. Tissue growth (yellowish color) into the cylinder is apparent. The red smears are red blood cells that are part of the host tissue that has grown into the cylinder.

Because arachidonic acid is a precursor for PGE2, we examined PGE2 concentration in the supernatants of wild-type and Casp3−/− MEF cultures before and after irradiation. Our results showed a strong correlation of supernatant PGE2 concentration with cellular caspase 3 status (Fig. 5A, right panel). The supernatant from wild-type MEF cultures had a higher basal concentration of PGE2 than did the cultures of Casp3−/− MEFs and showed a strong increase after radiation. In addition, radiation-induced concentrations of PGE2 in the supernatant from Casp3−/− MEF cultures were significantly lower than those from irradiated wild-type cell cultures. A role for iPLA2 in mediating radiation-induced increase in PGE2 concentration was demonstrated in two different ways. Short-hairpin RNA (shRNA)–mediated knockdown of iPLA2 significantly reduced induced concentrations of PGE2 in wild-type MEF cultures (Fig. 5A, right panel). (The effectiveness of the shRNA in knocking down iPLA2 is shown in fig. S6B.) Overexpression of a truncated iPLA2 gene, which encodes a truncated protein (fig. S6B) that is identical to a putative, constitutively active caspase 3 or 7 cleavage product of iPLA2, significantly restored both the basal concentration and the irradiation-induced increase in the concentration of PGE2 in Casp3−/− MEF cultures (Fig. 5A, right panel). These results provide evidence that caspase 3 regulates PGE2 concentrations through the activation of iPLA2 in dying cells.

Exposure of EKP-FLuc cells to either PGE2 or lysophosphatidylcholine (LPC), which are both products of activated iPLA2, showed that PGE2 stimulated cell proliferation, but that LPC did not (fig. S6C). Because PGE2 has been implicated in stem cell mobilization and tissue regeneration, we examined whether caspase 3–mediated iPLA2 activation was important for stimulating stem and progenitor cell proliferation. We evaluated wild-type MEF cells transduced with a shRNA-encoding gene against iPLA2 and Casp3−/− MEF cells transduced with a truncated, constitutively active version of iPLA2 (that was identical to a previously identified caspase-generated fragment; see fig. S6B) for the ability to support proliferation of EKP-Fluc cells. Reducing iPLA2 abundance in lethally irradiated wild-type MEFs significantly diminished their ability to support EKP-Fluc proliferation in vitro, whereas transduction of the constitutively active iPLA2 into Casp3−/− MEFs significantly increased their capacity to support EKP-Fluc proliferation when lethally irradiated (Fig. 5B).

A role for iPLA2 in tissue regeneration was also obtained with a modified DIVAA tissue regeneration assay in nude mice (Fig. 5C). In this assay, silicone cylinders containing cells or growth factors are placed into the host and the growth of the host tissue in the implanted silicone cylinders is quantified. Knocking down iPLA2 significantly attenuated the ability of irradiated wild-type MEF to induce host vascular growth into MEF cell-embedded silicone cylinders, whereas exogenous expression of a constitutively active, truncated iPLA2 significantly enhanced the ability of Casp3−/− MEF cells to stimulate host tissue growth into the cylinders. Therefore, these experiments provide evidence for a pathway involving caspase 3–activated iPLA2 in stem or progenitor cell proliferation and tissue regeneration.


Our study reveals a key role for programmed cell death in tissue regeneration in mice. Furthermore, we found that caspases 3 and 7, the “executioner” proteases that are instruments of cellular death and that function at the terminal stages of apoptosis, hold the key for mobilizing tissue stem and progenitor cells and promoting tissue regeneration. Caspases 3 and 7 cleaved and activated iPLA2 to ultimately trigger the production of the growth signal PGE2, thus showing that these executioner caspases not only contribute to the death process, but also participate in the production of paracrine signals.

Examples of the importance of apoptosis in tissue regeneration have been reported in earlier studies. For example, tail regeneration in Xenopus laevis requires apoptosis (48), as does tissue regeneration in planaria (flat worms) (49). However, the molecular mechanisms by which apoptosis contributed to regeneration were not determined. A study demonstrating the importance of apoptosis in Hydra head regeneration (50) showed that positive growth signals (in this case, Wnt3) were generated from apoptotic cells. However, the molecular mechanism for Wnt3 generation is not clear. It is therefore interesting to note a relationship between PGE2 and Wnt signaling pathways in regulating developmental specification and stem cell growth in mammalian cells (51). From these studies and our own results, it appears that the involvement of apoptosis-caspase-PGE2 (+ Wnt?) pathways in wound healing and tissue regeneration is evolutionarily conserved.

Because our study suggests a key role for cell death in promoting tissue regeneration in mammalian cells, we propose to name this mechanism the “phoenix rising” pathway of tissue regeneration (Fig. 6). We believe that it is a fundamental biological process that metazoan organisms use to sense tissue damage and initiate repair, regeneration, or both. The existence of such a pathway is a manifestation of the inextricable link between the yin and yang of cellular life and death in metazoan organisms.

Fig. 6

A schematic representation of the “phoenix rising” pathway of cell death–induced tissue regeneration. In damaged tissues, apoptotic cells activate caspases 3 and 7 through either the intrinsic pathways, involving Apaf and caspase 9, or the extrinsic pathways, involving caspase 8. Activated caspases 3 and 7 subsequently cleave and activate iPLA2, which generates arachidonic acid. Arachidonic acid is then converted into PGH2 by cyclooxygenases 1 and 2 (Cox1&2). PGE2 synthase converts PGH2 into PGE2, which stimulates stem cell proliferation and tissue regeneration.

Materials and Methods

Cell lines and culture conditions

EKP cells derived from C57BL/6 mice were purchased from Millipore. NSCs, derived from Balb/C mice, were purchased from American Type Culture Collection. MSCs were obtained from freshly isolated bone marrow following established protocols (MSCs, Technical Manual, Stem Cell Technologies). Wild-type, Casp3−/−, Casp7−/−, and Casp3−/−Casp7−/− double knockout MEFs were obtained from R. Flavell of Yale University (New Haven, CT). EPK cells were maintained in epidermal keratinocyte medium from Millipore (catalog #CNT-02). NSC cells were grown in neural stem cell medium (Millipore, catalog #SCM003). Wild-type and caspase-knockout MEF cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with 5% FBS.

Mouse strains

Nude mice were obtained from National Cancer Institute (Bethesda, MD). Wild-type C57BL/6, Casp3−/−, and Casp7−/− mice and enhanced green fluorescent protein (EGFP)–expressing transgenic mice (all in C57/BL6 background) were obtained from the Jackson Laboratory. The transgenic mice were deposited by R. Flavell. All animal procedures were approved by the University of Colorado Denver Institutional Animal Use and Care Committee. Mice were kept in the vivarium located on the Anschutz Medical Campus of University of Colorado Denver (Aurora, CO).

Genetic manipulation of the cell strains

For many experiments, we used the pLEX lentiviral vector system (52) from Open Biosystems for transduction of various genes into MEF cells and various stem or progenitor cells. The genes transduced with pLEX include (i) firefly luciferase gene, which was transferred from the plasmid pGL4.31-luc2 from Promega; (ii) a truncated version of the mouse iPLA2, which was amplified through RT-PCR from murine messenger RNA with the following primers:


Spe I site is underlined. Kozak sequence is bold.


Not I site is underlined.

We used the Pfx polymerase (Invitrogen) for the PCR amplification. The amplified fragment encodes amino acids 453 to 679, which is equivalent to amino acids 514 to 733 of human iPLA2 (35, 36), of murine iPLA2 (accession no. NM-016915). This fragment is a constitutively active, caspase-cleavage product (3436). The fragment was cloned into the Spe I and Not I sites of a modified pLEX plasmid. The modified pLEX plasmid has an influenza hemagglutinin (HA) tag inserted between Not I and Mlu I (two of the unique restriction sites in pLEX) so that genes inserted into the Not I site can be fused with the HA tag.

We obtained lentiviral vectors encoding shRNA-encoding minigenes targeted against the murine iPLA2 gene from Open Biosystems. These minigenes were carried in the pLKO.1 lentiviral vector system. The most effective one had the following targeting sequence: 5′GCGTATGAAGGACGAGGTGTTTCTCGAGAAACACCTCGTCCTTCATACGTTTTTTG-3′ (catalog no. Rmmu534-NM-016915 from Open Biosystems; in bold are the sense and antisense targeting sequences and in underlined text is the loop sequence). We followed the manufacturer’s instructions to produce live recombinant lentiviral vectors in 293T cells.

Bioluminescence imaging of cultured cells

Proliferation of Fluc-labeled stem or progenitor cells on lethally irradiated feeder cells was tracked though bioluminescence imaging. Imaging of the cells was performed with an IVIS200 instrument from Caliper Life Sciences. Luciferin (Caliper Life Sciences) was added to the media at a concentration of 150 μg/ml. After 5 to 10 min of incubation, cells cultured in petri dishes or multiwell plates were imaged. Quantification of signal strength was performed with manufacturer-supplied software.

Bioluminescence imaging of cells implanted in mice

Growth of Fluc-labeled stem or progenitor cells in vivo was followed through noninvasive bioluminescence imaging with the IVIS200 instrument. Mice were injected with d-luciferin (150 mg/kg) in 200 μl of PBS and imaged 10 min later. The time between injection and imaging was kept constant among different batches of mice.

Irradiation of cells and mice

X-ray irradiation of cells and mice was carried out in an RS-2000 Biological Irradiator (Rad Source Corporation) located in the vivarium of University of Colorado Denver Anschutz Medical Campus. The dose rate for the machine is about 1 Gy/min.

Measurement of arachidonic acid release

To measure arachidonic release, we plated cells in six-well plates at a density of 2.0 × 105 cells per well. About 1.0 μCi of [3H]arachidonic acid (GE Healthcare Life Sciences) was added to the cells cultured in about 1 ml of serum-free DMEM with lipid-free bovine serum albumin (0.5 mg/ml; Sigma). After 16 hours, the cells were washed with fresh medium three times and incubated with 3 ml of DMEM supplemented with 5% serum. After 5 hours, when arachidonic acid in the supernatant reached a steady-state level, the cells were exposed to 8 Gy of x-rays. Supernatants were then removed at 4, 24, and 48 hours from the cells and counted with a scintillation counter for quantification of [3H]arachidonic acid.

ELISA measurement of PGE2

To evaluate the PGE2 secretion from cells, we plated about 4 × 104 cells per well in 24-well plates. Cells were cultured in 1 ml of DMEM supplemented with 2% fetal bovine serum. Cells were then irradiated with x-rays (10 Gy). Supernatant from the cells was removed before and 48 hours after cellular irradiation. PGE2 in the supernatants was measured with an ELISA kit from R&D Systems.

Skin excision wound healing assay

To characterize the rate of wound healing in different strains of mice, punch biopsies were created in the dorsal skin of mice. The biopsy procedure was carried out with VisiPunch devices from Huot Instruments. Usually, two 4-mm excision wounds were created in the dorsal skin of each mouse. Monitoring of the rate of healing was performed by measuring the diameter of the wounds with a caliper every other day. For histological analysis, skin biopsies containing the full-thickness wounds were taken from mice killed at different time points after generation of the excision wounds. About 60 min before the killing, mice were injected with BrdU (50 mg/kg). The tissue samples were paraffin-embedded at the Histology Core in the Department of Dermatology at University of Colorado Denver.

For BrdU immunofluorescence staining, an Alexa Fluor 594–conjugated mouse monoclonal antibody against BrdU from Invitrogen was used at a dilution of 1:20. To quantify the number of cells positive for BrdU staining, we counted four randomly chosen 100× fields.

To analyze skin epithelial cells and keratinocytes through immunofluorescence staining, we used antibodies against cytokeratin 14 and cytokeratin 6, respectively. These were rabbit polyclonal antibodies produced in-house and described previously (53).

Partial hepatectomy–liver regeneration assay

To quantify liver regeneration, we adopted a published protocol (33). Briefly, mice (8 to 10 weeks old) were anesthetized with isofluorane. After restraining the mouse, we created a 3-cm incision in the midline of the abdominal skin and muscle to expose the xiphoid process. The median and left lateral lobes were tied around the top of the lobes with 4-0 silk surgical thread. The lobes were then removed surgically. This removes about two-thirds of the total liver weight. The abdomen was closed by suturing, and the mice were placed on a warm pad to recover. At different times after hepatectomy, mice were killed for evaluation. About 60 min before the mice were killed, BrdU (50 mg/kg) was injected intraperitoneally into the mice to label proliferating cells. The regeneration rate of the liver was quantified with two methods. The weights of the two lobes of liver left intact in mice were determined and normalized to the total liver weight before the initial surgery to derive a relative liver weight. Alternatively, immunofluorescence staining of BrdU-positive cells was performed in sections of paraffin-embedded liver tissues. Liver regeneration was then quantified through enumeration of BrdU-positive cells in four randomly chosen 200× fields.

Modified DIVAA assay for tissue regeneration

To quantify the ability of different cells to induce tissue regeneration, we adapted the DIVAA from Trevigen. The original purpose of this kit was to measure the ability of various growth factors to induce angiogenesis in vivo. The basic procedure involves mixing growth factors at different concentrations with basement membrane extract (BME) and loading the mixture (in a total volume of 20 μl) into silicone cylinders called the “angioreactor.” In our experiments, instead of mixing growth factors with the BME, we mixed various MEFs with BME. For each angioreactor, we used about 2 × 105 cells in a final volume of 20 μl. We surgically implanted up to four of the angioreactors subcutaneously into each nude mouse. Because the angioreactor has only one sealed end, the host tissue and vasculature can grow into it. Two weeks after implantation, host mice were killed and the angioreactors were removed. The contents of the angioreactors were then transferred to microtubes and their endothelial cell content was quantified by fluorescein isothiocyanate (FITC)–lectin staining following the manufacturer’s (Trevigen) instructions.


Acknowledgments: We thank R. Flavell of Yale University for providing the MEFs with caspase 3 and 7 deficiencies. We also thank the anonymous reviewers for the excellent suggestions and criticisms.

Funding: This study was supported in part by grants EB001882 from the U.S. National Institute of Bioimaging and Bioengineering, CA131408 and CA CA136748 from the U.S. National Cancer Institute, and grant NNX09AH19G from NASA Space Radiation Biology Research Program (all to C.-Y.L.). It was also supported by US NIH grants AR052263 and AR057212 to D.R.R. Q.H. was supported by grant 2010CB529900 from the National Basic Research Project of China and Outstanding Young Scientist grants 30325043 and 30428015 from China National Natural Science Foundation.

Author contributions: F.L., Q.H., J.C., and Y.P. performed the experiments; F.L., C.-Y.L., D.R.R., and J.S.B. designed the experiments and analyzed the data; C.-Y.L. analyzed the data and wrote the paper.

Competing interests: The University of Colorado has filed a provisional patent related to this work. The authors declare no personal financial interest in the published study.

Supplementary Materials

Fig. S1. Validation of the bioluminescence assay as a measure of cell proliferation.

Fig. S2. Caspase 7 in the dying cells contributes to their stimulation of progenitor cell proliferation.

Fig. S3. Immunofluorescence staining of cytokeratin 6 and cytokeratin 14 during skin wound healing in wild-type and Casp3−/− mice.

Fig. S4. Skin wound healing is compromised in Casp7−/− mice.

Fig. S5. Caspase 7 deficiency compromises liver regeneration.

Fig. S6. Calcium-independent phospholipase A2 in the proliferation-stimulating effects of dying cells.

References and Notes

View Abstract

Stay Connected to Science Signaling

Navigate This Article