Injury-induced MAPK activation triggers body axis formation in Hydra by default Wnt signaling

Abstract

Hydra's almost unlimited regenerative potential is based on Wnt signaling, but so far it is unknown how the injury stimulus is transmitted to discrete patterning fates in head and foot regenerates. We previously identified mitogen-activated protein kinases (MAPKs) among the earliest injury response molecules in Hydra head regeneration. Here, we show that three MAPKs-p38, c-Jun N-terminal kinases (JNKs), and extracellular signal-regulated kinases (ERKs)-are essential to initiate regeneration in Hydra, independent of the wound position. Their activation occurs in response to any injury and requires calcium and reactive oxygen species (ROS) signaling. Phosphorylated MAPKs hereby exhibit cross talk with mutual antagonism between the ERK pathway and stress-induced MAPKs, orchestrating a balance between cell survival and apoptosis. Importantly, Wnt3 and Wnt9/10c, which are induced by MAPK signaling, can partially rescue regeneration in tissues treated with MAPK inhibitors. Also, foot regenerates can be reverted to form head tissue by a pharmacological increase of β-catenin signaling or the application of recombinant Wnts. We propose a model in which a β-catenin-based stable gradient of head-forming capacity along the primary body axis, by differentially integrating an indiscriminate injury response, determines the fate of the regenerating tissue. Hereby, Wnt signaling acquires sustained activation in the head regenerate, while it is transient in the presumptive foot tissue. Given the high level of evolutionary conservation of MAPKs and Wnts, we assume that this mechanism is deeply embedded in our genome.

Keywords: MAPK signaling; Wnt/beta-catenin; axis formation; injury response; regeneration.

Fig. 1.

MAPK activation is injury dependent and required for the initiation of regeneration. (A) Experimental scheme showing decapitation by cutting (wounding) or careful ligation (without wounding). Ligated animals cannot regenerate properly due to preserved epithelial integrity, as indicated by propidium staining 30 min after injury/ligation. (B) Phosphorylation of MAPKs is reduced in ligated polyps. Regenerates from cut or tied animals were analyzed by western blot after the indicated time points using phospho-specific MAPK antibodies and ɑ-tubulin specific antibody (ɑ-tub as control (Materials and Methods). (C) Equal activation of MAPKs in head and foot regenerates as detected by phospho-specific antibodies, as in B. (D and E) Confocal images of regenerating and injured polyps with phospho-specific antibodies for ERK, p38, and JNK 30 min after amputation/injury/ligation. (D) Lateral view on the wound site (asterisk) of regenerating polyps postamputation shows a cytoplasmic distribution for activated ERK and JNK, while p38 localizes to nuclei. Scale bar: 50 µm. (E) IF analysis of MAPK activation at the injury site in lateral views 30 min after injury; the column with the oral top view additionally shows the direct view of the wound site. Phospho-specific antibodies revealed gradient-like activation MAPKs peaking at the site of amputation or incision as indicated. Note that the tissue completely lacked MAPK activation in tied polyps. Nuclei were stained using DAPI (blue). Scale bars: 250 µm each. (F and G) Inhibition of MAPK activity prevents head and foot regeneration. Polyps were bisected and the proximal and distal parts were allowed to regenerate in the presence of inhibitors specific for ERK, p38, and JNK. Foot regenerates were evaluated by peroxidase staining, as shown in G. Error bars indicate the mean of three independent experiments with SD. Head regeneration (head reg.): n (control [ctrl]) = 108, n (ΔERK) = 61, n (Δp38) = 58, n (ΔJNK) = 72. Foot regeneration (foot reg.): n = 45 per condition. (G) Representative pictures from the experiment described in F. The white arrow indicates the basal disk stained with peroxidase. Scale bars: 500 µm.

Fig. 2.

MAPK activation acts downstream of redox signaling. (A and B) Analysis of ROS in regenerating tissue within the first 30 min postamputation. Polyps were bisected at 50% body length and allowed to regenerate in the presence of HM (ctrl) or HM supplied with either H₂O₂ or GSH. Asterisks indicate decreased staining intensity upon GSH treatment. Specimens were analyzed by western blot (A) or IF (B) using phospho-specific antibodies as indicated. (A) Exposure to H₂O₂ increased activated MAPKs levels as compared with untreated ctrl, while regenerates in reduced GSH showed less MAPK activation. (B) IF analysis of polyps fixed at 30 min postamputation showed an increase in MAPK activation after H₂O₂ treatment and a decrease after reduced GSH treatment. Scale bars: 250 µm. (C) Quantitative analysis of polyps bisected at 50% body length. 72 h after the onset of regeneration, ctrl polyps and H₂O₂-treated animals regenerated normally, whereas GSH exposure inhibited both head and foot regenerates. Error bars indicate the mean of three independent experiments with SD. n = 45 per condition. (D) Representative images of the experiment described in C. Regeneration of a head or foot was evaluated as before (Fig. 1E). Scale bars: 500 µm.

Fig. 3.

MAPK activation is calcium dependent. (A) Injury causes increased and prolonged calcium release. Using transgenic ectodermal and endodermal GCaMP reporter strains, Ca²⁺ release was measured in vivo by fluorescence increase (arbitrary units [AU], i.e. fold change increase of basic reporter activity) after cutting (injury) or touch-induced contraction (noninjury). The time point of the poke or cut is indicated by an arrowhead. Mechanically stimulated ctrl polyps show a brief increase in fluorescence in both germ layers, whereas cut polyps show a prolonged increase in signal activity. Note that this increase occurs at both the oral and the aboral sites of section. (B) Representative pictures of transgenic Ca²⁺ reporter strains at 5 s postamputation. Dashed boxes indicate regions used for measurements shown in A. Scale bars: 500 µm. (C and D) Changes in intracellular Ca²⁺ levels affect MAPK activation, as analyzed by western blots (C) or by IF (D). (C) Animals were sectioned in the presence of the intracellular calcium chelator BAPTA, the ionophore A23187, or DMSO (ctrl). Whole-body lysates were prepared at 30 min pi and subjected to western blot analysis using phospho-specific MAPK antibodies. A decrease of Ca²⁺ (BAPTA) resulted in a decrease in activated MAPK levels, while an increase of Ca²⁺ (A23187) resulted in elevated phosphorylation events of all MAPKs. (D) IF analysis of whole-mount polyps after exposure to DMSO (ctrl), A23187, or BAPTA. Animals were fixed 30 min after amputation and stained with antibodies specific for pERK, pp38, and pJNK. Increasing cytoplasmic calcium levels transformed the gradient-like distribution of activated MAPKs (ctrl) into a homogenous staining throughout the tissue (A23187). Calcium capture by BAPTA resulted in a reduction of signal intensity upon injury. Scale bars: 250 µm. (E and F) Calcium release is essential for regeneration. Bisected polyps were exposed to DMSO, BAPTA, or BAPTA/EGTA for 3 h after cutting, and the regeneration capacity was evaluated at 72 hpa. The transient repression of Ca²⁺ signaling by BAPTA inhibited head and foot regeneration significantly (n = 40 each; error bars indicate the mean of three independent experiments with SD). Note that regeneration was further reduced when BAPTA was used in combination with EGTA. (F) Representative images of the experiment described in E. Scale bars: 500 µm.

Fig. 4.

Wound- and patterning-specific functions of Wnts. (A) Activation of Wnt expression in head and foot regenerates of bisected polyps. RNA was isolated from tissue of bisected polyps at the indicated time points and used for qPCR analysis with specific primers for HyWnt9/10c, HyWnt3, and HyWnt7. Values obtained were normalized to the GAPDH ctrl, and fold changes were calculated. (B) Expression levels of HyWnt3 and HyWnt9/10c increased about 20-fold within the first 6 h in both head and foot regenerates upon injury, while the HyWnt7 increase was only twofold. Note that the foot-specific decrease of Wnt expression is already initiated at 6 hpa. Significance was tested using Student’s t test. **P < 0.005. Error bars without asterisks did not pass the significance threshold. (C–E) siRNA-mediated knockdown of Wnt3, Wnt9/10c, and Wnt7 results in decreased regeneration capacity. (C) Silenced polyps were bisected, and head regeneration was evaluated at 72 hpa. The regeneration capacity of polyps dropped to about 50%, with siWnt3 and siWnt9/10c treatment affecting the onset of regeneration, while siWnt7 caused patterning defects. (D) Regeneration capacity upon Wnt3, Wnt9/10c, or Wnt7 knockdown equally decreased to less than 50% in foot regenerates determined by peroxidase assay (perox-). Error bars indicate the mean of five independent experiments with SD. Head regeneration: n (siGFP) = 163, n (siWnt3) = 153, n (siWnt7) = 140, n (siWnt9/10c) = 154. Foot regeneration: n (siGFP) = 95, n (siWnt3) = 92, n (siWnt7) = 92, n (siWnt9/10c) = 90. Significance of regeneration capacity was tested using logistic regression analysis, ***P < 0.001. (E) Representative pictures of C and D are shown. siWnt9/10c blocked regeneration completely, while siWnt3-treated head regenerates formed a characteristic monotentacle and siWnt7 treatment resulted in misaligned tentacles. Foot regeneration was abrogated in all Wnt-silenced polyps, as determined by peroxidase assay. Scale bars: 500 µm.

Fig. 5.

MAPKs influence Wnt expression levels in head and foot regenerates. (A) Activation of Wnt expression in head and foot regenerates of bisected polyps. RNA was isolated from tissue of bisected polyps at the indicated time points and used for qPCR analysis with specific primers for HyWnt3 and HyWnt9/10c. Head (A) and foot (B) regenerating polyps were incubated with inhibitors specific to each MAPK, and log2 fold change was calculated. While inhibition of ERK activity resulted in reduced Wnt expression, inhibition of stress-induced MAPKs yielded higher expression levels for Wnt3 but down-regulation for Wnt9/10c. Error bars indicate the mean of three independent experiments with SD. Significance was tested using Student’s t test. *P < 0.05; **P < 0.005. Error bars without asterisks did not pass the significance threshold. (C) HyWnt9/10c partially rescues regeneration defects after MAPK inhibition. Polyps were bisected and regenerated in the presence of inhibitors specific to either ERK, p38, or JNK. During the first 24 h, regenerates were incubated with recombinant HyWnt3, HyWnt9/10c, or BSA (ctrl). At 72 hpa, an increased regeneration capacity was found after HyWnt9/10c treatment for ERK- and p38-inhibited polyps, whereas the effect of HyWnt3 treatment was smaller. Error bars indicate the mean of three independent experiments with SD. Significance of regeneration restoration was tested using logistic regression analysis. *P < 0.05; **P < 0.005; ***P < 0.001. Absolute numbers and representative pictures of animals used are shown in SI Appendix, Fig. S7. (D) Schematic representation of the polarity reversal experiment performed in E and F. A central piece of gastral tissue was excised from budless animals at 30% and 70% body length, exposed to recombinant Wnt protein for 24 h, and afterward allowed to regenerate in HM for 6 days. (E) Polyps were scored at 7 days postamputation. Pieces incubated with BSA regenerated animals with the original oral-aboral polarity. Pieces incubated with HyWnt3 and/or HyWnt9/10c regenerated animals with morphological defects (e.g., ectopic tentacles) and animals in which a head formed at the original oral and aboral pole, indicating a transformation of positional identity (polarity reversal) at the aboral end. Pieces with polarity reversal form foot structures between the two heads, as visualized by peroxidase assay. Note that polarity reversal is less pronounced in polyps exposed to HyWnt9/10c only. Error bars indicate the mean of three independent experiments with SD. Significance was determined by binomial regression model. *P < 0.05; **P < 0.005; ***P < 0.001. n (BSA) = 61, n (Wnt3) = 72, n (Wnt9/10c) = 78, n (Wnt3/Wnt9/10c) = 79. (F) Representative pictures of the experiment described in E. Regenerates incubated with BSA (ctrl) show a correct oral-aboral axis, while regenerates treated with Wnt proteins show polarity reversal. Scale bars: 500 µm.

Fig. 6.

MAPK cross talk balances apoptosis. (A) Scheme showing the relationships between MAPK pathways based on data shown below and in SI Appendix, Fig. S8. While stress-activated MAPKs (p38 and JNK) activate the respective pathway, the ERK pathway inhibits stress-activated MAPKs. Antagonism between ERK and JNK pathways is reflected on the level of apoptosis. (B) Comparison of protocols provided by the manufacturer and as published in Vogg et al. (25) and Chera et al. (29). Specificity of the TUNEL reaction was tested by comparing polyps that had been fixed overnight, followed by amputation to animals that had been first bisected and subsequently fixed. Upper: Incubation step at 70 °C yielded the same rate of TUNEL-positive cells in animals fixed before and after amputation. Therefore, the signal was considered to be unspecific, and incubation of polyps in citrate buffer was executed exclusively at room temperature to prevent an unspecific stain despite the decreased signal intensity (Lower). Scale bars: 250 µm. (C) Time lapse of apoptosis visualized by TUNEL assay from bisected polyps fixed at the indicated time points. The maximum of apoptotic cells is detectable between 1 and 2 hpa in both head and foot regenerates. Scale bars: 250 µm. (D) Visualization of apoptotic cells by TUNEL assay analysis in bisected polyps upon inhibition of MAPKs at 1.5 hpa. While p38 inhibition has no effect on apoptotic events, the frequency clearly increased upon JNK inhibition and was diminished in the presence of the ERK inhibitor. Note that there is no difference between head and foot regenerates. Scale bars: 250 µm. Nuclei were stained with DAPI (blue) in all shown images.

Fig. 7.

Injury triggers default Wnt expression to initiate patterning. (A) Polyps were labeled with green fluorescent latex beads and subsequently incised in the budding region at the side of the label at 2 h postlabeling (cut), while the uncut population remained unaffected (ctrl). After 3 days, the number (ctrl/cut) and position of new buds were determined, as depicted by schematic representation. While noncut polyps exhibited random bud development, incised animals showed positional bias of budding at the labeled side. Significance of biased budding from three independent experiments was tested using logistic regression analysis with a P value of 0.0001. n (ctrl) = 48, n (cut) = 55. (B) Gastric pieces were excised and exposed to DMSO (ctrl) or to azakenpaullone (AZP). Stabilization of β-catenin yields 85% of animals showing head structures at both extremities upon AZP treatment, while ctrl polyps preserve the original polarity. Right Lower Corner: Numbers indicate the frequency of normal morphology/morphology defects/polarity reversal. Scale bars: 500 µm. n (ctrl) = 96, n (AZP) = 95. (C) Schematic representation of how source density is related to Wnt/β-catenin signaling and how it finally yields head and foot regeneration upon injury. Injury causes initial up-regulation of Wnt/β-catenin at either end that induces the tissue to regenerate. The positional specification of the regenerating tissue is dependent on the relative value of the source density along the body column. While high source density commits the tissue to head regeneration, lower values of the source density yield foot regeneration.