Inositol kinase and its product accelerate wound healing by modulating calcium levels, Rho GTPases, and F-actin assembly

Abstract

Wound healing is essential for survival. We took advantage of the Xenopus embryo, which exhibits remarkable capacities to repair wounds quickly and efficiently, to investigate the mechanisms responsible for wound healing. Previous work has shown that injury triggers a rapid calcium response, followed by the activation of Ras homolog (Rho) family guanosine triphosphatases (GTPases), which regulate the formation and contraction of an F-actin purse string around the wound margin. How these processes are coordinated following wounding remained unclear. Here we show that inositol-trisphosphate 3-kinase B (Itpkb) via its enzymatic product inositol 1,3,4,5-tetrakisphosphate (InsP₄) plays an essential role during wound healing by modulating the activity of Rho family GTPases and F-actin ring assembly. Furthermore, we show that Itpkb and InsP₄ modulate the speed of the calcium wave, which propagates from the site of injury into neighboring uninjured cells. Strikingly, both overexpression of itpkb and exogenous application of InsP₄ accelerate the speed of wound closure, a finding that has potential implications in our quest to find treatments that improve wound healing in patients with acute or chronic wounds.

Vertebrate embryos have remarkable capacities to heal wounds quickly, efficiently and without scarring, through mechanisms that are not fully understood (1). Xenopus embryos, for example, can heal perfectly and completely within a few hours following injury through mechanisms that include constriction of an actin purse-string and protrusive activity of the leading edge cells at the wound margin (2, 3). We have been exploiting the efficient wound healing ability of Xenopus embryos and oocytes to identify therapeutic targets aimed at improving the speed and quality of wound healing in human patients.

Calcium (Ca²⁺) is a key player during wound healing (4). In particular, it acts upstream of the activation of the Rho family GTPases (RhoA, Rac1, and Cdc42), which are essential for controlling the assembly and contraction of the F-actin ring at the wound edge (5, 6). Although it has long been appreciated that Ca²⁺ plays important roles during both single-cell and multicellular wound healing, less is known about the mechanisms that modulate intracellular Ca²⁺ levels following injury.

Inositol (1,4,5)-trisphosphate (InsP₃) is a second messenger that mediates the release of Ca²⁺ from intracellular stores. The role of its downstream and chemically related metabolites is less known but inositol 1,3,4,5-tetrakisphosphate (InsP₄) has been shown to function synergistically with InsP₃ and modulate the strength and persistence of intracellular Ca²⁺ levels (7–11). InsP₄ is made via the phosphorylation of InsP₃ in a reaction mediated by the enzyme inositol 1,4,5-trisphosphate 3-kinase (Itpk) (12).

Here, we use Xenopus embryos and oocytes to investigate whether inositol-trisphosphate 3-kinase B (Itpkb) and its product InsP₄ provide a mechanistic link between the release of Ca²⁺ and the reorganization of the cytoskeleton during wound healing. Our results revealed that itpkb is expressed in the embryonic epidermis and its protein product rapidly accumulates around the wound edges, where it colocalizes with activated Cdc42. Overexpression of itpkb and exogenous treatment with InsP₄ promotes the activation of Cdc42, Rac1, and RhoA; increases the assembly of F-actin around the wound margin; and accelerates the speed of wound closure. In contrast, itpkb knockdown impairs the formation of an F-actin purse string and delays the speed of wound closure. This delay can be rescued by addition of InsP₄ to the medium, suggesting that InsP₄ is the critical molecule downstream of Itpkb in this process. Finally, we show that itpkb is essential for the propagation of a rapid Ca²⁺ wave that begins from the site of injury and spreads into the neighboring uninjured cells. These data identify Itpkb and its metabolite InsP₄ as important cellular responders to injury and as potential targets in our quest to enhance the quality and speed of wound healing.

Results

Itpkb-Positive Cells Show a Scattered Epidermal Distribution.

We first characterized the temporal and spatial expression of itpkb in Xenopus tropicalis embryos to determine whether itpkb was expressed at the right time and place to participate in embryonic wound healing. RT-quantitative (q)PCR data showed that itpkb is expressed in oocytes and that its expression level increases throughout early development (Fig. S1A). RT-qPCR analysis on animal cap explants, which are fated to give rise to the embryonic epidermis (13), showed an enrichment in the animal pole versus whole embryos at the blastula stages (Fig. S1B). Itpkb transcripts were detectable by whole-mount in situ hybridization starting from the blastula stage in the animal cap and around the anterior neural plate at the gastrula stages (Fig. 1A and Fig. S1C). By the tailbud stage, itpkb expression was clearly evident in the epidermis [Fig. 1A, stage 23–32; Fig. 1B).

Fig. 1.

Itpkb gene is maternal and presents an epidermal pattern with strong expression on scattered cells. (A) Whole-mount in situ hybridization (WMISH) for itpkb at different stages of development. Dashed line on st. 10 embryo indicates the dorsal (d) lip. Dashed lines on st. 12.5 and st. 13 embryos mark the neural tube, which divides the embryo into left and right halves. (B) WMISH section for itpkb confirms the epidermal expression in scattered cells. Section is 25 μm thick. (Bar, 100 μm; inset bar, 25 μm.) a, anterior; an, animal pole; L, lateral view; p, posterior; v, ventral; veg, vegetal pole.

Itpkb Modulates the Speed of Healing in Ectodermal Explants.

To address whether itpkb and InsP₄ facilitate multicellular wound healing, we developed a quantifiable wound healing assay system using animal cap explants isolated from blastula stage Xenopus laevis embryos (Materials and Methods). Freshly isolated animal cap explants are initially flat but they round up and heal within a few hours (Fig. 2A). To compare relative speed rates of wound healing, we determined the time taken for control animal cap explants to reach half closure (i.e., 0.5) and we set this time as relative time 1 (Fig. 2B, control). Animal caps overexpressing myc-itpkb reached half closure significantly faster than control animal caps (Fig. 2B, uninjected and prolactin injected). We also found that animal caps incubated with 0.5 D-myo-inositol-1,3,4,5-tetraphosphate, sodium salt (aIP₄), the primary product of the Itpkb reached half closure significantly faster than control animal caps (Fig. 2B).

Fig. 2.

Itpkb modulates the speed of wound closure. (A) Animal cap assay. Area inside and area outside of the animal cap explants were measured and relative closure was calculated. The control and experimental time at 0.5 closure was extracted from each relative closure graph and represented as a bar graph. Each experimental time was relative to its respective control. (B) Itpkb and IP₄ accelerate multicellular wound closure. Graph of time at 0.5 closure when 0.5 ng myc-itpkb mRNA was injected or 0.5 μM aIP₄ treatment was done, relative to controls (uninjected and prolactin injected embryos). Uninjected control n = 12; prolactin control n = 5; itpkb n = 6; aIP₄ n = 6. (C) Knock down of itpkb delays multicellular wound closure. Graph of time at 0.5 closure when 40 ng itpkb-mo–injected with or without 0.4 ng myc-itpkb or 0.5 μM aIP₄, relative to controls (moc-injected embryos). Control n = 7, itpkb-mo n = 7, itpkb-mo + itpkb n = 4, itpkb-mo + aIP₄ n = 3. Each experimental condition was analyzed by Student t test, nonparametric Mann-Whitney test. Results are shown as means ± SEM. Asterisks indicate *P < 0.05. Each N is one experiment done with five animal caps. (D–F) Epidermal itpkb knock down delays wound closure. (D) Embryos were coinjected with microruby and 20 ng moc or itpkb-mo into the two ventral animal blastomeres in stage 4 (eight-cell stage) embryos. Embryos were wounded and treated with or without 0.5 μM aIP₄. (E) Quantification of the epidermal wound closure over time, with the different experimental condition, moc (black spot), itpkb-mo (red spot), itpkb-mo + aIP₄ (orange spot), moc + aIP₄ (green spot). N moc = 10, N itpkb-mo = 7, N itpkb-mo + aIP₄= 4, N moc + aIP₄ = 8. Each time point was analyzed by Student t test, nonparametric Mann-Whitney test. Results are shown as means ± SEM. Asterisks indicate *P < 0.05, **P < 0.01, ***P < 0.001. (F) Example of fluorescent images showing a time lapse of embryos wounded in the epidermis, taken with the different experimental condition. From top to bottom: moc, itpkb-mo, itpkb-mo + aIP₄, moc + aIP₄.

In contrast, animal cap explants isolated from embryos injected with an ATG itpkb-mo, which blocks de novo translation of itpkb-C′eGFP mRNA (Fig. S2A) and decreased the level of endogenous Itpkb protein in blastula and neurula stage embryos (Fig. S2D), healed significantly slower than the mo control (moc) animal caps (Fig. 2C). Importantly, the delayed speed in wound healing in itpkb-mo injected animal caps was fully rescued by coinjection of the myc-itpkb construct, which is insensitive to the itpkb-mo (Fig. S2C), or by incubation with aIP₄ (Fig. 2C).

Itpkb Function Is Required for Epidermal Wound Healing in the Embryo.

To address whether Itpkb and InsP₄ play a role during epidermal wound healing in vivo, we coinjected the fluorescent lineage label (micro ruby) together with itpkb-mo or moc into the two animal ventral blastomeres at the eight-cell stage, which are fated to give rise to the embryonic epidermis (Fig. 2D). These embryos were raised until the tailbud stages (stages 20–25) and wounded in the epidermis (Fig. 2D). Consistent with our previous animal cap wound-healing experiments, injection of itpkb-mo into embryos resulted in delayed epidermal wound healing, an effect that could be rescued by incubating the itpkb-mo wounded embryos with 0.5 μM of aIP₄ (Fig. 2 E and F). Interestingly, we saw a significant acceleration in the speed of wound healing when moc-injected embryos were treated with 0.5 μM aIP₄ (Fig. 2 E and F). In summary, these data suggest that itpkb is required for epidermal wound healing and that InsP₄, the product of Itpkb, is not only an important modulator of epidermal wound healing, but it is capable of accelerating the speed of wound closure of epidermal wounds in Xenopus embryos.

Itpkb Colocalizes with F-Actin During Wound Closure.

To test whether the subcellular localization of Itpkb changes during wound healing, we generated an antibody against Xenopus Itpkb (Fig. S2 A and B) and used it to assess its localization after injury (Fig. 3A). To identify the wound margin, we injected embryos with mRNA encoding GAP43-GFP, which labels the plasma membrane (14). Immunofluorescence of wounded embryos showed that endogenous Itpkb is present at the wound margin (Fig. 3A). To visualize the change in localization in living embryos, we injected an itpkb-3xC′eGFP fusion construct (itpkb-eGFP) or eGFP as control. We wounded the injected blastula stage embryos in the animal pole and imaged them using time-lapse, multiple focal plane (4D) microscopy. Although eGFP did not change localization after wounding, Itpkb-eGFP accumulated at the edges of the multicellular wounds within 2 min of injury (Fig. 3B and Movie S1). To compare the distributions of Itpkb localization with the assembly of the F-actin ring around the wound edge, we injected embryos with mRNA encoding both Itpkb-3xC′eGFP and Cherry-dBDmoesin (an F-actin probe). Following wounding, we found that Itpkb-eGFP colocalized with F-actin at the wound edge (Fig. 3C). Disruption of F-actin assembly using Latrunculin B (LatB) resulted in the loss of Itpkb-eGFP accumulation at the wound edge, whereas disruption of microtubules using nocodazole (Noc) had no effect on its localization (Fig. 3D), suggesting an intimate association of Itpkb and the actin cytoskeleton during wound healing. These findings are consistent with previous reports showing that Itpkb physically associates with cortical actin filaments (15, 16).

Fig. 3.

Itpkb accumulates around the wound edge and modulates F-actin assembly in multicellular wounds. (A) Immunofluorescence of wounded embryos shows the accumulation of endogenous Itpkb at the wound edge. Embryos were injected with membrane GFP probe (gap-43-gfp mRNA), wounded, fixed, and stained with antibodies against GFP and Itpkb. Green, GFP; red, Itpkb. n = 6. (B) Itpkb accumulates at the leading edge of wounded cells. (Left) An animal cap wound in an embryo injected with 0.2 ng egfp mRNA as a control. (Right) An animal cap wound in an embryo injected with 5 ng itpkb-3xC′egfp mRNA (itpkb-eGFP). Control n = 7, itpkb-eGFP n = 8. (C) Itpkb colocalizes with the F-actin array around the wound edge. Embryos were coinjected with 1.5 ng cherry-dbmoesin mRNA and 5 ng itpkb-egfp mRNA. Green, itpkb-eGFP accumulation around wound edge; red, F-actin array around the multicellular wound; yellow, merged image of F-actin and itpkb-eGFP postwounding (p.w.). The image represents an example at 2 min p.w. n = 7. (D) The accumulation of Itpkb at the leading edge of wounded cells is F-actin–dependent. Embryos were injected with itpkb-egfp mRNA, treated with DMSO (Left), LatB (2.5 μM) (Center), or Noc (50 μM) (Right) for 30 min before wounding. Live images were taken 3 min after wounding to observe the localization of itpkb-eGFP. N DMSO = 11; N LatB = 9; N Noc = 8. (E) Itpkb knock down impairs F-actin assembly after wounding. Embryos were coinjected with 0.5 ng F-actin probe and 40 ng, itpkb-mo, or moc. Itpkb-mo embryos were also coinjected with myc-itpkb mRNA. (Left) F-actin array in moc-wounded animal cap. Middle panel shows disrupted F-actin array on itpkb-mo wounded animal. (Center) Rescued F-actin array in itpkb-mo + myc-itpkb injected embryos. N moc = 14; N itpkb-mo = 20; N rescue = 9. (Bars, 50 μm.)

Knock Down of itpkb Impairs F-Actin Assembly after Wounding.

As Itpkb colocalized with F-actin at the wound edges, we asked whether disrupting itpkb affects F-actin assembly during wound healing. To do this, we injected embryos with mRNA encoding the F-actin probe eGFP-dBDmoesin with itpkb-mo or moc and cultured the embryos until stage 9, at which point they were wounded and imaged. Although moc-injected embryos showed normal F-actin assembly around the wound edge, itpkb-mo–injected embryos showed disarrangement in the normal assembly of F-actin purse string around the wound edge (Fig. 3E and Movie S2). Furthermore, overexpression of myc-itpkb mRNA was able to restore the assembly of the F-actin purse string in itpkb-mo–injected embryos (Fig. 3E). These data were confirmed with experiments in oocytes: myc-itpkb overexpression or incubation with aIP₄, but not cIP₄ [Ins(1,2,3,5)P₄, control], led to brighter F-actin array compared with controls (Fig. S3 A–D and Movies S3 and S4), acceleration in the speed of F-actin ring contraction at the wound margin (Fig. S3 E and F), and overall acceleration of wound closure (Fig. S3 G and H). These data further suggest a role for the Itpkb in regulating the cytoskeleton during wound healing.

Itpkb Modulates the Activation of the Cdc42, Rac1, and RhoA in Wounded Embryos and Oocytes.

To address the mechanism by which Itpkb and InsP4 facilitate wound closure, we asked whether Itpkb activity affects wound-induced activation of the Rho family GTPases (17, 18). To measure Cdc42 and Rac1 activity on embryos, we performed GST pull-down assays on wounded embryos using a fusion protein of GST and the p21-binding domain from PAK1, which binds both activated Cdc42 and Rac1 (19) (Fig. S4). For RhoA activity, we generated a GST fusion protein with the Rho-binding domain of Rhotekin (20) and used this construct for GST pull-down assays (Fig. S4). We found that knocking down itpkb function, with the itpkb-mo, impairs injury-induced activation of Cdc42 and Rac1 in multicellular wounds (Fig. 4 A and B), whereas RhoA activation was increased in itpkb-mo embryos following wounding (Fig. 4C). In contrast, myc-itpkb overexpression and aIP₄ treatment resulted in hyperactivation of Cdc42, Rac1, and RhoA in both wounded and unwounded embryos (Fig. 4 D–F). This finding was confirmed by fluorescent intensity quantification studies in oocytes, where either overexpression of myc-itpkb or incubation with 0.5 μM aIP₄ resulted in a significant increase in the intensity of fluorescence and thus the levels of active RhoA and Cdc42 relative to controls (Fig. S5 A, C–F and Movies S5 and S6).

Fig. 4.

Itpkb modulates the activation of Rho GTPases during wound repair. (A) Knock down of itpkb impairs the activation of Cdc42 after wounding. Western blot shows the activation of Cdc42 before and after wounding in control (lanes 1–2), moc (lanes 3–4), and itpkb-mo–injected embryos (lanes 5–6). (B) Knock down of itpkb impairs the activation of Rac1 after wounding. Western blot shows the activation of Rac1 before and after wounding in control (lanes 1–2), moc (lanes 3–4), and itpkb-mo–injected embryos (lanes 5–6). (C) Knock down of itpkb enhances the activation of RhoA after wounding. Western blot shows the activation of RhoA before and after wounding in control (lanes 1–2), moc (lanes 3–4), and itpkb-mo (lanes 5–6). (D–F) Itpkb and IP₄ promote the activation of Cdc42 (D), Rac1 (E), and RhoA (F) in embryos. Western blots show the activation of Cdc42 (D), Rac1 (E), and RhoA (F) before and after wounding in control (lanes 1–2), myc-itpkb mRNA-injected (lanes 3–4), and aIP₄-treated (lanes 5–6) embryos. For each experiment, 5% input of respective GTPase was taken as the starting level of total protein before pull down, and the GST levels were assessed as loading controls. (−) Indicates unwounded and (+) indicates wounded. Results are representative of three independent experiments.

To determine whether Itpkb colocalizes with active Cdc42, we coinjected Xenopus oocytes with mRNA encoding eGFP (as control) or Itpkb-3xC′eGFP together with a fusion construct containing monomeric red fluorescent protein (mRFP) in frame with the Cdc42-GTPase binding domain (GBD) of N-WASP (mRFP-wGBD) to visualize activated Cdc42 (aCdc42) (20). We then laser wounded the oocytes and imaged them on the confocal. Although eGFP did not localize around the wound margin (Fig. S5B, Bottom), we observed Itpkb-eGFP localized as a broad ring around the wounds, overlapping with aCdc42 (Fig. S5B, Top and Movie S7). The distribution of Itpkb-eGFP and aCdc42 narrowed as wound closure progressed, supporting a role of Itpkb on cytoskeleton rearrangement upon wounding. Thus, Itpkb, through the generation of InsP₄, increases the basal level of active Rho GTPases, Rac1, Cdc42, and RhoA and further activates these proteins following injury in multicellular and single-cell wounds.

Knock Down of itpkb Impairs the Propagation of a Ca²⁺ Wave Following Wounding.

Calcium plays an essential role during embryonic wound healing (4, 5) and is required to activate the Rho GTPases during wound repair (5, 20). In addition, epidermal wound healing in Caenorhabditis elegans triggers a Ca²⁺-dependent signaling cascade that promotes wound closure (21). Previous reports have shown that InsP₄ can modulate Ca²⁺ signaling in Xenopus oocytes and in cultured cells (8–11); however, whether this also occurs during the complex process of wound healing was not known. Thus, we next asked whether Itpkb modulates Ca²⁺ levels during embryonic wound healing in Xenopus embryos in vivo. For these experiments, we used two different Ca²⁺ sensors; GEM-GECO1, an improved genetically encoded Ca²⁺ sensor based on GCaMP3 (22), and C2-mRFP, a genetically encoded Ca²⁺ sensor, which translocates to the plasma membrane in the presence of elevated Ca²⁺ (5, 23). We coinjected the mRNA encoding both of these Ca²⁺ sensors together with 40 ng of moc or itpkb-mo, and at stage 9, we generated laser wounds, followed by high-speed resonant scanning confocal imaging. To avoid influx of extracellular Ca²⁺ after wounding, we performed these experiments in Ca²⁺ free 1× Marc's Modified Ringer's (MMR) culture medium. The embryos were washed with Ca²⁺ free 1× MMR medium immediately before wounding and they remained in Ca²⁺-free medium for the duration of the imaging, which lasted less than a minute.

The GEM-GECO1 Ca²⁺ sensor in moc-injected embryos displayed a robust and rapid intracellular Ca²⁺ response within seconds following laser wounding, which propagated into uninjured neighboring cells as a wave (Fig. 5 A and B and Movies S8 and S9). In contrast, itpkb-mo–injected embryos exhibited a much reduced Ca²⁺ wave propagation following injury (Fig. 5A and Movie S9). However, the Ca²⁺ wave was restored in itpkb-mo–injected embryos after treatment with aIP₄, suggesting that the product of Itpkb, InsP₄, was essential for the propagation of the Ca²⁺ wave from the site of injury through the tissue (Fig. 5A and Movie S9). We quantified the Ca²⁺ signal in the injured cells, extending radially into the second and third rows of neighboring uninjured cells, to get a sense of the speed and extent of wave propagation. In moc-injected embryos, the Ca²⁺ level reached a maximum in the first cell after 6 s, the second cell after 7.5 s, and the third cell after 9 s (Fig. 5B). In itpkb-mo–injected embryos, the Ca²⁺ signal takes much longer to reach a maximum in the second and third cells, and the maximum is considerably lower that in the moc control cells (Fig. 5 C–E). In contrast, the aIP₄ rescued embryos not only restored the Ca²⁺ wave to normal levels, but the speed of the wave was slightly faster, such that the wave reached a maximum in the first cell after 3 s (as opposed to 6 in the moc-injected embryos), 6 s in the second cell (as opposed to 7.5 s in the moc-injected embryos), and 7.5 in the third cell (as opposed to 9 seconds in the moc-injected embryos) (Fig. 5 C–E). Interestingly, we also found that aIP₄ increased the speed of wave propagation in moc-injected embryos (Fig. S6 A and B and Movie S8).

Fig. 5.

Itpkb is required for Ca²⁺ wave propagation following wounding. (A) Knock down of itpkb impairs Ca²⁺ wave propagation after wounding. Embryos were coinjected with 2.5 ng mRNA encoding GEM GECO and 40 ng of itpkb-mo or moc. Examples of Ca²⁺ ratio high/low over time in a wounded moc embryo (Top), wounded itpkb-mo embryo (Middle), and wounded itpkb-mo embryo treated with 0.5 μM aIP₄ (Bottom). (Bars, 20 μm.) Asterisks indicate wound site. (B) 3D spatiotemporal quantification of Ca²⁺ ratio high/low on moc embryos after wounding, in the first cell or wounded cell (solid black line), second cell (dark gray line), and third cell (light gray line). For first cell n = 11, second cell n = 17, and third cell n = 10. (C–E) Temporal quantification of Ca²⁺ ratio high/low after wounding in embryos treated with moc (black line), itpkb-mo (red line), and itpkb-mo + 0.5 μM aIP₄ (green line), in the first cell line or wounded cell (C); second cell line (D); and third cell line (E). For moc first cell n = 11, second cell = 17, third cell = 10; for itpkb-mo first cell n = 5, second cell = 24, third cell = 23; for itpkb-mo + aIP₄ first cell n = 5, second cell = 13, third cell = 19. (F–M) Temporal quantification of Ca²⁺ ratio high/low in second cell line after wounding embryos. (F and G) Knock down of itpkb impairs intracellular Ca²⁺ wave propagation after wounding. Embryos wounded in medium without (F) or with calcium (G). Gray, moc embryos; red, itpkb-mo embryos. (H and I) XeC delays Ca²⁺ wave propagation after wounding in moc embryos and has an additive inhibitory effect on itpkb-mo–injected embryos. (H) Moc-injected embryos. (I) itpkb-mo–injected embryos. Green, DMSO-treated embryos; pink, 4 μM XeC-treated embryos. (J and K) InsP₃ rescues the delayed Ca²⁺ wave propagation on itpkb-mo–injected wounded embryos. (J) Moc-injected embryos. (K) Itpkb-mo–injected embryos. Fuchsia, water-treated embryos; purple, 0.5 μM inositol-1,4,5-triphosphate (IP₃)-treated embryos. (L and M) Ry delays Ca²⁺ wave propagation after wounding in moc embryos, but does not have an additive inhibitory effect on itpkb-mo embryos. (L) Moc-injected embryos. (M) Itpkb-mo–injected embryos. Blue, ethanol-treated embryos; orange, 30 μM Ry-treated embryos. N is as in Fig. S6. Each graph was analyzed by two-way ANOVA. Results are shown as means ± SEM. **P < 0.01, ***P < 0.001. NS, not significant.

As an independent measure of intracellular Ca²⁺ levels, we also used an additional Ca²⁺ sensor, C2-mRFP, which translocates to the plasma membrane upon Ca²⁺ binding (5). The C2-mRFP sensor also spread several cell diameters (∼61.5 μm radial perimeter) from the site of injury within seconds following wounding (Fig. S6C and Movie S10). As with the GEM-GECO1 sensor, the itpkb-mo embryos failed to show a membrane-localized C2-mRFP Ca²⁺ signal beyond the injured cell (Fig. S6C and Movie S10). We quantified the spread of the Ca²⁺ wave by measuring the change in fluorescence intensity of the C2-mRFP signal from the wounded cell to the third row of neighboring uninjured cells (Fig. S6 D and E). Although both moc- and itpkb-mo–injected embryos showed a response in the wounded cell by 3 s (compare solid black line in Fig. S6D with black dotted line in Fig. S6E), the wound-induced Ca²⁺ signal failed to carry into the second or third row of neighboring cells in itpkb-mo–injected embryos (compare dotted yellow and red lines in Fig. S6E versus the solid yellow and red lines in Fig. S6D). Thus, Itpkb activity is required for injury-induced Ca²⁺ wave propagation from the site of injury into neighboring uninjured cells in multicellular wounds, and addition of InsP4 to the medium is sufficient to restore the injury response in embryos lacking Itpkb activity.

We next wished to determine whether the Itpkb-dependent Ca²⁺ wave was mediated by the release of intracellular Ca²⁺ stores from the endoplasmic reticulum (ER) or whether it might also involve Ca²⁺ influx from membrane Ca²⁺ channels, such as transient receptor potential ion channels (TRPM channels) (21). Therefore, we performed the Ca²⁺ imaging studies on moc–injected and itpkb-mo–injected embryos in the presence and absence of extracellular Ca²⁺. From these experiments, we found that the Itpkb-dependent Ca²⁺ wave was only apparent when the medium lacked Ca²⁺, suggesting that Itpkb was primarily involved in the release of intracellular Ca²⁺ stores (Fig. 5 F and G and Fig. S6 F and G). We then assessed whether Itpkb acts through the activation of ER Ca²⁺ channels. Treatment of embryos with Xestospongin C (XeC), which inhibits the InsP₃ receptors (24), delayed the Ca²⁺ wave in moc-injected embryos, suggesting that the Ca²⁺ wave was dependent on the release of ER Ca²⁺ stores. In addition, treatment of itpkb-mo–injected embryos with XeC resulted in a further decrease in both the speed and intensity of the Ca²⁺ wave (Fig. 5 H and I; Fig. S6 H and I), suggesting that Itpkb works together with InsP₃ receptors. To confirm this, we added InsP₃ to itpkb-mo–injected embryos and found that InsP₃ could restore the Ca²⁺ wave in itpkb-mo–injected embryos (Fig. 5K and Fig. S6K). However, addition of InsP₃ failed to accelerate the speed of the Ca²⁺ wave in moc-injected control embryos, unlike InsP₄, which is able to accelerate the speed of the Ca²⁺ wave in moc control embryos. (Fig. 5J and Fig. S6 A, B, and J). This suggests that InsP₄ acts in conjunction with InsP₃ in activating the InsP₃ receptors. Finally, we asked whether the Itpkb-dependent Ca²⁺ wave involved ryanodine (Ry) receptors as well. We added Ry to moc and itpkb-mo–treated embryos and found that, although Ry could delay the Ca²⁺ wave in moc-treated embryos, it had little additional effect on the Ca²⁺ wave in itpkb-mo–injected embryos (Fig. 5 L and M and Fig. S6 L and M). This suggests that Itpkb and its product InsP₄ act in combination with InsP₃ in inducing Ca²⁺ release from intracellular ER stores following wounding. To assess whether these effects were relevant to wound healing, we returned to the animal cap wound healing assay and found that XeC and Ry delayed wound healing in animal caps, and InsP₃ was able to rescue the speed of wound healing in itpkb-mo–injected animal caps (Fig. S6 N–P).

Discussion

There is great interest in understanding the cellular and molecular bases of epithelial repair as a means of identifying unique therapeutic targets to improve the speed and quality of wound healing (25). Following injury, a cell needs to repair its membrane and close the wound. It does so by mechanisms that use Ca²⁺ signaling, remodeling of the cytoskeleton, and the activity of the Rho family GTPases (5, 26–29). To preserve the life of the cell and perhaps the organism, these primary repair mechanisms have to be activated within seconds after injury. However, how the speed of wound healing is regulated was not well understood. Here, we show a role for Itpkb activity and its product InsP₄ during the rapid and efficient repair of both single-cell and multicellular wounds. Furthermore, we show that increasing the level of itpkb expression or incubation with InsP₄ results in a marked acceleration in the speed of wound healing in both single-cell and multicellular wounds, including epidermal wounds in embryos.

Previous studies have linked Itpk activity and InsP₄ with the modulation of intracellular Ca²⁺ levels (7–11). In addition, both single-cell and multicellular wound healing depend on a rapid rise in intracellular Ca²⁺ levels (5, 6, 20, 21). An immediate outcome of this increase is the fusion of internal membrane vesicles and their coordinate exocytosis, thus restoring the integrity of the plasma membrane (30, 31). Another outcome of this increase in intracellular Ca²⁺ is the activation of Rho family GTPases, which then regulate the generation and contraction of an F-actin purse string around the wound edge (5, 6, 20). Our findings expand on this by showing that InsP₄ plays a critical role in the activation of the Rho family GTPases and the formation of F-actin around the edge of both single-cell and multicellular wounds. It is notable that the precise mechanism by which Ca²⁺ regulates the Rho family GTPases during wound healing is not yet known. Interestingly, many Dock-related Rho guanine nucleotide exchange factors (GEFs) and a few Dbl-related Rho GEFs contain Ca²⁺ binding (C2) domains, which provide a potential link with Ca²⁺ (32). Another link between Ca²⁺ and Rho family GTPases are through Ras-GRF proteins. In particular, Ras-GRF1 has been shown to activate Rac in calcium-dependent manner (33).

It is remarkable that repair mechanisms used by single-cell wounds and multicellular wounds are so similar. Perhaps this is not surprising if one considers that, despite any differences in relative wound size, in both cases the wound has to close rapidly and the membrane has to be repaired. However, multicellular wounds have additional requirements because multiple cells need to coordinate their response and an injury signal has to be transmitted to cells further away from the wound site. We have found that for these events, Itpkb and its product InsP₄ are also important. Through mechanisms that are not yet clear, InsP₄ works in conjunction with InsP₃ in facilitating the release of intracellular calcium stores as a wave, propagating from the site of injury into neighboring uninjured cells. In this way, InsP₄ with InsP₃ play essential roles in transmitting the injury signal from the wound site into neighboring uninjured cells, thereby initiating the wound repair response.

In conclusion, our findings provide a previously unappreciated regulatory intracellular element, InsP₄, which facilitates and indeed has the capacity to accelerate the speed of single-cell and multicellular wound healing. Thus, InsP₄ is an important factor, which in combination with other molecules, such as InsP₃, helps coordinate the diversity of cellular processes that are needed for wound repair. Thus, our findings provide a unique avenue for therapeutic targets aimed at improving wound healing in patients suffering from acute or chronic wounds.

Materials and Methods

Plasmid DNA Construction and mRNA Synthesis.

RFP-wGBD, eGFP-Rhotekin GBD (rGBD), and C2-mRFP constructs were gifts from William (Bill) Bement, Madison, WI (5, 20). dBDmoesin plasmid (F-actin probe) (34) was a gift from Tom Millard, Manchester, UK and was PCR-amplified and subcloned into pDEST-N eGFP. Full-length Xenopus itpkb (XM_002938844) cDNA was PCR-amplified and subcloned into pCS2-C′3xGFP or pCS2-N′ myc-tag, resulting in itpkb-3xGFP (three copies of eGFP at the C terminus) or Myc-itpkb, respectively. For GST pull-down assay, pGEXTK-PAK1 70–117 aa (N′GST-PAK) plasmid was purchased from Addgene (Addgene plasmid 12217). PAK, rGBD, and GST were PCR-amplified and subcloned into pCS107. For calcium probe, GEM-GECO plasmid was purchased from Addgene (Addgene plasmid 32442) and subcloned into pCS2. All cDNA constructs were linearized and mRNA was generated by in vitro transcription using Ambion’s mMESSAGE kit (5′capped). The primers used for the different subcloning steps are as described in SI Materials and Methods.

Embryos Manipulations.

For the animal cap wound healing assays, X. laevis embryos were injected into the animal pole at stage 2, and animal cap explants were excised at stages 8–9. Caps were incubated and cultured inside up in 75% (vol/vol) Normal Amphibian Medium (NAM) containing 0.2% BSA. Embryos were preincubated 0.5 μM aIP₄ or 0.5 μM cIP₄ 30 min before animal cap explants were excised. Image acquisition was taken every 30 min. For embryonic epidermal wound healing assay, embryos were injected at stage 4 in both ventral animal blastomeres, and wound was done at the embryonic epidermis at tailbud stage with Gastromaster equipment (Xenotek Engineering) to perform consistent sizes wound. Thirty minutes before wounding, embryos were preincubated in 75% (vol/vol) NAM containing 0.2% BSA with or without aIP₄. After wounding, images were taken every 1 min. For both wound assays, we used Leica MZFLUOIII stereoscope and Northern Eclipse software (Empix Imaging Inc).

For live fluorescence imaging, embryos were injected and animal cap explants were excised as described previously. Following animal cap explants extraction, superficial wound was done on them with forceps and confocal imaging was taken with Olympus IX81 FluoView FV1000 microscope (Olympus Inc) through a 20× objective.

For Ca²⁺ imaging (C2-mRFP and GEM-GECO), embryos were coinjected with Ca²⁺ probe and itpkb-mo or moc. They were raised until stage 9 and wounded with an ablation laser, followed by high-speed resonant scanning confocal imaging.

GST Pull Down.

For active Rac1 and Cdc42 pull down, embryos were injected with 500 pg of PAK1-GST (70-117) at the one- or two-cell stage to all cells. Embryos were raised until st. 10 in 0.1 × MMR and transferred to 75% (vol/vol) NAM before wounding. Forceps were used for wounding the animal pole of embryos, and wounded embryos were collected 4 min after wounding and put on ice immediately. Fifty embryos were collected per condition. The GST pull-down procedure was based on ref. 35 with modifications (SI Materials and Methods). Immunoblot was done to determine the amount of active Rac1, Cdc42, RhoA, and GST was pulled down; 5% input was taken as starting control.