Characterization of TGFβ signaling during tail regeneration in the leopard Gecko (Eublepharis macularius)

Authors

  • Richard W.D. Gilbert,

    1. Department of Biomedical Sciences, Ontario Veterinary College, University of Guelph, Guelph, Ontario, Canada
    Search for more papers by this author
  • Matthew K. Vickaryous,

    Corresponding author
    1. Department of Biomedical Sciences, Ontario Veterinary College, University of Guelph, Guelph, Ontario, Canada
    • Correspondence to: Dr. Matthew K. Vickaryous, 50 Stone Road East, Guelph Ontario, Canada N1G2W1. E-mail: mvickary@uoguelph.ca or Dr. Alicia M, Viloria-Petit, 50 Stone Road East, Guelph Ontario, Canada N1G2W1. E-mail: aviloria@uoguelph.ca

    Search for more papers by this author
  • Alicia M. Viloria-Petit

    Corresponding author
    1. Department of Biomedical Sciences, Ontario Veterinary College, University of Guelph, Guelph, Ontario, Canada
    • Correspondence to: Dr. Matthew K. Vickaryous, 50 Stone Road East, Guelph Ontario, Canada N1G2W1. E-mail: mvickary@uoguelph.ca or Dr. Alicia M, Viloria-Petit, 50 Stone Road East, Guelph Ontario, Canada N1G2W1. E-mail: aviloria@uoguelph.ca

    Search for more papers by this author

Abstract

Introduction: The transforming growth factor beta (TGFβ)/activin signaling pathway has a number of documented roles during wound healing and is increasingly appreciated as an essential component of multi-tissue regeneration that occurs in amphibians and fish. Among amniotes (reptiles and mammals), less is known due in part to the lack of an appropriate model organism capable of multi-tissue regeneration. The leopard gecko Eublepharis macularius is able to spontaneously, and repeatedly, regenerate its tail following tail loss. We examined the expression and localization of several key components of the TGFβ/activin signaling pathway during tail regeneration of the leopard gecko. Results: We observed a marked increase in phosphorylated Smad2 expression within the regenerate blastema indicating active TGFβ/activin signaling. Interestingly, during early regeneration, TGFβ1 expression is limited whereas activin-βA is strongly upregulated. We also observe the expression of EMT transcription factors Snail1 and Snail2 in the blastema. Conclusions: Combined, these observations provide strong support for the importance of different TGFβ ligands during multi-tissue regeneration and the potential role of TGFβ/activin-induced EMT programs during this process. Developmental Dynamics 242:886–896, 2013. © 2013 Wiley Periodicals, Inc.

INTRODUCTION

Among amniotes, the most common response to traumatic injury is the rapid formation of fibrotic tissue. Although fibrotic (or scar) tissue is often capable of stabilizing the injury and re-establishing tissue homeostasis, it is a dysfunctional and structurally dissimilar substitute (Martin, 1997; Ferguson and O'Kane, 2004; Gurtner et al., 2008). In contrast, some amniotes, along with various urodeles (newts, axolotls) and teleosts (zebrafish), are capable of scar-free wound healing (Tanaka and Reddien, 2011). Significantly, scar-free wound healing is often followed by tissue regeneration and, in many instances, the restoration of function. Some of the best-known examples of scar-free wound healing and regeneration in amniotes include: early embryos (Whitby and Ferguson, 1991; reviewed by Ferguson and O'Kane, 2004); the digit tips of neonate and adult mice (Han et al., 2008; Fernando et al., 2011; Rinkevich et al., 2011); and large segments of the integument of spiny mice (Acomys) (Seifert et al., 2012). Furthermore, several strains of mice possess enhanced regenerative capacity including, MRL (Murphy Roths Large) mice (Heber-Katz, 1999; Leferovich and Heber-Katz, 2002), and mice with a point mutation in TGFβRI (Liu et al., 2011). In addition, among many species of lizards the tail is capable of undergoing scar-free wound healing and regeneration (Bellairs and Bryant, 1985; Alibardi, 2010; Delorme et al., 2012). This naturally evolved ability is typically associated with caudal autotomy, an anti-predation strategy whereby the tail can be self-detached. Tail loss yields a large cross-sectional wound, with the spinal cord, skeletal muscle, dermis, and tail vertebrae all exposed (Bellairs and Bryant, 1985; McLean and Vickaryous, 2011). Spontaneously, this wound is rapidly sealed without the formation of fibrotic tissue, and a replacement tail is soon developed (Alibardi, 2010; McLean and Vickaryous, 2011; Delorme et al., 2012). The replacement tail arises from a blastema, an accumulation of highly proliferative, mesenchymal-like cells that rapidly populate the site of tail loss (McLean and Vickaryous, 2011). Whereas the endogenous source of blastemal cells in lizards remains uncertain, evidence from the study of various other scar-free wound healing and regenerating species, including urodeles (Kragl et al., 2009) and mice (digit tips: Rinkevich et al., 2011), strongly supports dedifferentiation of lineage-restricted progenitor cells and not pluripotent stem cell populations.

An emerging model for the study of scar-free wound healing and regeneration in lizards is the leopard gecko, Eublepharis macularius (Whimster, 1978; McLean and Vickaryous, 2011; Delorme et al., 2012). Previous work using the leopard gecko has established that wound healing and regeneration are highly conserved events and that the entire process can be divided into seven discrete morphological stages (McLean and Vickaryous, 2011; see Experimental Procedures section). Stages I to III primarily involve closure of the wound, including wound site contraction, the formation of an exudate clot, re-epithelialization of the wound site, and the initiation of blastema formation. Stages IV to VII involve tissue outgrowth and differentiation, including axonogenesis, spinal cord outgrowth, and muscle and skeletal development (McLean and Vickaryous, 2011). Although the molecular regulation of these events remains poorly understood, evidence from various other regeneration-competent models supports the involvement of various cytokines (Stoick-Cooper et al., 2007; Stocum and Cameron, 2011).

The common canonical signaling pathway shared by transforming growth factor beta (TGFβ) and activin ligands, referred to as the TGFβ/activin pathway (Ogunjimi et al., 2012), has been shown to be essential for the regenerative response in other models of multi-tissue regeneration (Jazwinska et al., 2007; Levesque et al., 2007; Ho and Whitman, 2008). TGFβ and activin are members of the TGFβ superfamily, and signal through specific sets of serine/threonine kinase receptors (Moustakas and Heldin, 2009). Upon ligand binding, TGFβ/activin receptors become activated and subsequently recruit and phosphorylate specific signaling intermediates known as receptor-regulated SMADs (R-SMADs) (Wu and Hill, 2009). TGFβ and activin cause the phosphorylation of a particular subset of R-SMADs, SMAD2 and SMAD3 (Ross and Hill, 2008). Once phosphorylated, SMAD2/3 then form a complex with a common mediator, SMAD4, and subsequently translocate to the nucleus to drive a variety of gene expression programs (Ross and Hill, 2008). Previous studies in zebrafish, axolotls, and Xenopus have shown that by inhibiting the kinase domains of TGFβ and activin receptors (via the use of small molecule inhibitors), multi-tissue regeneration can be blocked (Jazwinska et al., 2007; Levesque et al., 2007; Ho and Whitman, 2008). Similar results were obtained by using knockdown morpholinos to silence the activin arm of the TGFβ/activin signaling pathway during zebrafish tail fin regeneration, with treated tails demonstrating a 50–80% reduction in regenerative capacity (Jazwinska et al., 2007). Taken together, these findings suggest that different TGFβ/activin ligands (such as TGFβ1–3, and/or activin βA/βB) are playing vital, ligand-specific roles during the regenerative process. To date, however, specific expression patterns for the different TGFβ/activin ligands have yet to be characterized. For example, earlier work investigating tail regeneration in the leopard gecko determined that TGFβ3 is expressed exclusively during the later stages of regeneration (stages VI and VII) in chondrocytes and fibroblasts, and that this protein was absent during the process of wound healing (Delorme et al., 2012). These observations, combined with those of previous studies demonstrating variable TGFβ ligand expression throughout regeneration in zebrafish (Jazwinska et al. 2007), axolotl (Levesque et al., 2007), and Xenopus (Ho and Whitman, 2008), indicate that differential and dynamic expression of individual TGFβ/activin ligands acts to mediate unique cellular events during tissue restoration.

TGFβ/activin signaling has been demonstrated to play important roles in driving pluripotency and self-renewal in numerous stem cell populations, including embryonic stem cells, tissue specific stem cells, as well as cancer stem cells (Oshimori and Fuchs, 2012). Additionally, TGFβ/activin signals can confer stemness unto non-stem-like cells through a process known as the epithelial to mesenchymal transition (EMT). During EMT, particular transcription factors known to be upregulated in response to TGFβ mediate the transition of cells from a stationary epithelial phenotype to a motile, mesenchymal-like phenotype. Interestingly, some cells transitioning from epithelial-to-mesenchymal states acquire self-renewing traits typical of stem cells (Mani et al., 2008; Scheel et al., 2011). During multi-tissue regeneration, it has been demonstrated that the blastema arises from lineage-restricted progenitors that through an unknown mechanism acquire a mesenchymal-like morphology and self-renewal traits reminiscent of tissue-specific stem cells (Stocum and Cameron, 2010). Thus, it is possible that TGFβ/activin signaling-mediated EMT-like processes drive regeneration in the gecko.

With this in mind, we conducted a spatiotemporal investigation of TGFβ/activin signaling during multi-tissue regeneration in the leopard gecko. In particular, we focused on the TGFβ/activin signaling intermediate SMAD2 as well as the prototypic TGFβ cytokine TGFβ1, a ligand known to be associated with scar formation and fibrosis (Bielefeld et al., 2012). Following the observation that TGFβ1 expression was notably low in early regenerate tissue despite robust pSMAD2 induction, we conducted a qRT-PCR screen of ligands capable of activating the TGFβ/activin pathway and observed an upregulation of activin-βA. We also screened for TGFβ/activin target genes associated with EMT and stemness and observed the EMT transcription factors Snail1 and Snail2 to be significantly upregulated in regenerate tissue.

RESULTS

TGFβ/activin Signaling Is Active in the Regenerate Blastema

The blastema is a highly proliferative, mesenchymal-like population of cells that accumulates under the wound epidermis and contributes to the formation of replacement tissue (Stoick-Cooper et al., 2007; Stocum and Cameron, 2011). To determine if blastema cells are actively responding to TGFβ/activin signals we performed dual immunofluorescence with pSMAD2, indicating TGFβ/activin activation, and the proliferation marker proliferating cell nuclear antigen (PCNA), which was used as a marker of proliferation.

During early blastema formation (stage III of tail regeneration), mesenchymal-like blastema cells aggregate at the site of tail loss, beginning at the distal end of the torn spinal cord (Fig. 1A–D). This population of cells is strongly positive for PCNA as well as showing moderate levels of pSMAD2 expression suggesting active TGFβ/activin signaling within the cells of the early blastema.

Figure 1.

Immunofluorescence of PCNA (red) and pSMAD2 (green) expression during four stages of gecko tail regeneration and associated schematic representations of longitudinal sections. A–D: Stage III; dotted line surrounds putative blastema cell population distal to spinal cord. E–H: Stage IV. I–L: Stage V. M–P: Stage VI; dotted line surrounds cartilaginous cone. All images taken at 20× magnification. Scale bars = 100 μm. sc, spinal cord; cc, cartilaginous cone.

As regeneration continues, a cone-like outgrowth is formed (stages IV and V of regeneration; Fig. 1E–H and I–L). During these stages, most blastema cells are PCNA positive and demonstrate robust pSMAD2 expression (Fig. 1H, L). It is also worth noting that unlike the original epidermis, the newly regenerated wound epithelium lacks a basally restricted proliferative population (Fig. 1F, J).

During late regeneration (stage VI), differentiation of skeletal muscle and cartilage is well underway (Fig. 1M). Overall, there are fewer cells co-localizing PCNA and pSMAD2, and fewer cells positive for PCNA. With few exceptions, PCNA expression is now localized to tissue-specific compartments, such as the chondroblasts lining the outer margin of the regenerate skeleton (Fig. 1M). In contrast, pSMAD2 staining remains robust among regenerating and regenerated tissues (Fig. 1O). These findings support the predicted role of TGFβ/activin signaling throughout the regenerative program.

Phosphorylated-SMAD2 Is Upregulated in Regenerating Tails

To confirm that the canonical TGFβ/activin pathway is activated during tail regeneration, we investigated the protein expression of pSMAD2, native SMAD2, and TGFβ1 using Western blot analysis. Tails at each of stages III, IV, V, and VI were subdivided into regenerate, proximal original, and distal original segments and run in separate lanes (Fig. 2). Across stages IV–VI, pSMAD2 expression is most intense in the blastema/regenerating tissue segment, as well as in the stage-III distal original segment (Fig. 2A–D). In contrast, pSMAD2 expression in original tissue segments is weak. Native SMAD2 expression is relatively uniform across all tissues examined, as is TGFβ1, with one notable exception: TGFβ1 was conspicuously absent from regenerate tissue at stage IV (Fig. 2B). Combined, these observations confirm that pSMAD2 is strongly induced in regenerating tissues and suggest that TGFβ1 may not be the primary pSMAD2 inductive ligand in stage IV of regeneration.

Figure 2.

Representative Western blots analysis of segmented gecko tails at four stages of tail regeneration as well as tissue from original tails (n = 3). A: Stage III. B: Stage IV. C: Stage V. D: Stage VI. Note the induction of pSMAD2 in most distal regenerate tissue at all stages of regeneration and the conspicuous absence of TGF μI protein expression in stage IV regenerate tail tissue. Native SMAD2 and α-tubulin remain relatively constant and serve as loading controls.

TGFβ1 Expression During Tail Regeneration

To verify the absence of TGFβ1 from stage-IV tails, we used immunohistochemistry (Fig. 3). Whereas TGFβ1 is strongly expressed by cells of the original epidermis and wound epithelium at stages III, V, and VI, it is strikingly absent from keratinocytes of the wound epithelium at stage IV. We also observed a dramatic difference in TGFβ1 expression among cells of the dermis and blastema. More specifically, fibroblasts of the original dermis and blastema cells/fibroblasts of the regenerating tail at stages V and VI are immunopositive for TGFβ1, whereas blastema cells at stage IV are immunonegative (Fig. 3C). At stage III, the blastema cells are also immunonegative, although various immune cells demonstrate strong TGFβ1 expression (Fig. 3B). Combined, these data indicate that TGFβ1 protein expression is dynamic throughout regeneration.

Figure 3.

Immunohistochemical localization of TGFβ1 protein expression during four stages of gecko tail regeneration as well as original tissue; positive signal is demonstrated by brown staining (DAB chromogen) contrasted to hematoxylin (blue) counterstain. Negative control tissue is shown in the right panel. A, An: Original tissue. B, Bn: Stage III. C, Cn: Stage IV. D, Dn: Stage V. E, En: Stage VI. Scale bar = 100 μm. Arrows marked M = TGFβ1-positive mesenchymal-like cells. Arrows marked I = TGFβ1-positive immune cells. OE, original epithelium; WE, wound epithelium.

Inverse Expression of TGFβ1 and pSMAD2 at Stage IV of Regeneration

To confirm that protein expression of TGFβ1 and pSMAD2 at stage IV of regeneration is largely non-overlapping, we used immunohistochemistry (for TGFβ1) and immunofluorescence (for pSMAD2) on adjacent serial sections (Fig. 4). During re-epithelialization of the wound site, newly formed keratinocytes form an involuting structure known as an epidermal down growth, marking the boundary between the original epidermis and the wound epithelium, thereby framing the blastema (McLean and Vickaryous, 2011) (Fig. 4). TGFβ1 is strongly expressed by cells in original tail tissues (primarily keratinocytes and fibroblasts; Fig. 4A, B), but absent from regenerated tissues (Fig. 4C). In contrast, pSMAD2 expression is relatively limited in cells of the original tissues (Fig. 4D, E), but widespread among cells in the regenerating tissue (Fig. 4F). This result reveals that expression of TGFβ1 and pSMAD2 are non-overlapping in original and regenerate tissues, and suggests that another TGFβ/activin ligand is responsible for the pSMAD2 induction seen in early regenerate tissues.

Figure 4.

Non-overlapping expression of TGFβ1 and pSMAD2 in original and regenerate tissue in stage IV regenerate tails. TGFβ1 expression (A), high-magnification insets (B, C), and negative controls (Bn, Cn). pSMAD2 expression on adjacent tissue section (D) and high-magnification insets (E, F). Scale bars = (A, D) 200 μm, (B, C, E, F) 50 μm.

Activin-βA Is Upregulated in the Blastema During Early Regeneration

Given the absence of TGFβ1 during stage IV of regeneration, we sought to identify the unknown pSMAD2-inducing ligand(s) using a qRT-PCR screen. We first cloned all three TGFβ isoforms (TGFβ1, TGFβ2, and TGFβ3), as well as BMP2 and activin-βA. Then we examined mRNA expression of these ligands using qRT-PCR on tissue from stage IV tails subdivided into regenerate, proximal original, and distal original segments. We determined that mRNA expression of TGFβ1 and TGFβ2 remains relatively constant throughout original and regenerated tissue segments (Fig. 5), consistent with our protein expression data for TGFβ1. Although, TGFβ3 and BMP2 mRNA is upregulated in regenerating tissues, the increases are not statistically significant (Fig. 5). Significantly, however, activin-βA mRNA expression undergoes a ∼15-fold upregulation in regenerating tissues compared to original tissues (Fig. 5). These results point towards activin-A or activin-AB as a candidate ligand responsible for the induction of pSMAD2 during the early stages of tail regeneration.

Figure 5.

Relative mRNA levels of TGFβ family members in segmented stage-IV regenerate tails determined using quantitative real-time polymerase chain reaction (qRT-PCR) and normalized to 18s mRNA levels. Graph depicts the average fold change (±SEM) in segmented tissue sections relative to the most proximal tissue section. A statistically significant increase in activin-βA was observed in the most distal tissue section corresponding to newly regenerate tissue and the position of the blastema. Statistical significance was determined using Kruskal-Wallis non-parametric statistics followed by a Dunns post-test with a p value cutoff of *p < 0.05; n = 3 for biological and technical replicates.

EMT Transcription Factors Snail1 and Snail2 Are Upregulated in Regenerate Tissue

TGFβ/activin ligands participate in a wide range of cellular processes, many of which are likely to play important roles during multi-tissue regeneration. We investigated the TGFβ/activin-induced target genes Snail1, Snail2 (formerly Slug) and Zeb2 at stage IV of regeneration using qRT-PCR. Snail1, Snail2, and Zeb2 are transcription factors known to mediate the epithelial to mesenchymal transition (EMT), a process implicated in driving cell motility as well as stemness (Lim and Thiery, 2012). EMT has been implicated in development, cancer and wound healing but is not known to play a role in multi-tissue regeneration. Interestingly, we observed a significant upregulation of Snail1 and Snail2 (but not Zeb2) in the regenerate tissues of the gecko tail compared to original tissues (Fig. 6). These novel findings, combined with the robust expression of pSMAD2 seen in the blastema, suggest that the EMT could be an important, TGFβ/activin-mediated process that is occurring during multi-tissue regeneration.

Figure 6.

Relative mRNA levels of EMT transcription factors in segmented stage-IV regenerate tails determined using quantitative real-time polymerase chain reaction (qRT-PCR) and normalized to 18s mRNA levels. Graph depicts the average fold change (±SEM) in segmented tissue sections relative to the most proximal tissue section. A statistically significant increase in Snail1 and Snail2 was observed in the most distal tissue section corresponding to newly regenerate tissue and the position of the blastema. Statistical significance was determined using Kruskal-Wallis non-parametric statistics followed by a Dunns post-test with a p value cutoff of *p < 0.05; n = 3 for biological and technical replicates.

DISCUSSION

The TGFβ/Activin Signaling Pathway Is Activated Throughout the Tail Regeneration Program

Among amniotes, one of the most dramatic examples of spontaneous multi-tissue regeneration is the restoration of the lizard tail. Following tail loss (typically in response to the threat of predation), many lizard species are able to recreate a replacement appendage, complete with skeletal support, nerves, muscle, and a spinal cord (Bellairs and Bryant, 1985). We determined that similar to other regeneration-competent species (e.g., zebrafish, axolotl, and Xenopus; Jazawinska et al., 2007; Levesque et al., 2007; Ho and Whitman, 2008), in leopard geckos the TGFβ/activin signaling pathway is activated throughout the regenerative program. As evidenced by robust pSMAD2 expression and nuclear localization, TGFβ/activin signaling occurs throughout blastema formation, regenerative outgrowth, and tissue differentiation, representing regeneration stages III–VI (McLean and Vickaryous, 2011). The co-localization of PCNA and pSMAD2 seen throughout regeneration indicates that large numbers of blastema cells are proliferating and actively responding to TGFβ/activin signals. Whereas TGFβ/activin ligands are classically thought of as antiproliferative cytokines (SMAD2/3 complexes are known to inhibit cell cycle progression; Sandhu et al., 1997; Seoane, 2004; Alexandrow and Moses, 1995), during development and cancer progression the cytostatic functions can be overcome (Massague, 2012). In particular, crosstalk with other signaling pathways has been demonstrated to overcome the cytostatic program through integration at the level of the SMADs (Guo and Wang, 2009; Massague, 2012). For example, SMAD-β-catenin complexes driven by TGFβ and Wnt signaling promote stem cell proliferation and self-renewal by stimulating mitosis, while inhibiting differentiation (Jian et al., 2006). Thus, the ability for SMAD proteins to cooperate with other signals and direct diverse transcriptional effects depending on the cellular context represents a logical explanation for their consistently observed involvement in multi-tissue regeneration.

Activin and Not TGFβ, Is Upregulated During Early Regeneration

Whereas pSMAD2 expression is a reliable readout of canonical TGFβ/activin signaling (Moustakas and Heldin, 2009), it does not distinguish between activating ligand(s). Given the documented roles of TGFβ1 during mammalian wound healing (Ferguson and O'Kane, 2004; Bielefeld et al., 2012), we investigated the spatiotemporal expression of this ligand during gecko tail regeneration. Based on Western blotting and immunohistochemical data, we demonstrate that TGFβ1 is constitutively present in original tail tissue and at most stages of regeneration (stages III, V–VI) but is conspicuously absent from regenerating tissues at stage IV. Classically, TGFβ1 is characterized as a pro-fibrotic cytokine. For example, scar-free wound healing in fetal mammals is characterized by lower levels of TGFβ1 and higher levels of TGFβ3, compared to scar forming wound healing typical of adult mammals (Soo et al., 2003; Larson et al., 2010; Bielefeld et al., 2012; Wynn and Ramalingam, 2012). Furthermore, when scar-free fetal wounds are treated with TGFβ1, scarification occurs (Lin and Adzick, 1996). Therefore, limited expression of TGFβ1 protein during tail regeneration is consistent with the characterization of this process as a scar-free phenomenon (Delorme et al., 2012).

Despite the absence of TGFβ1 expression in regenerating tissue, we still observed a large induction of pSMAD2 among cells of the blastema and wound epithelium, indicating continued activation of the TGFβ/activin pathway. In order to identify the ligands responsible for the SMAD2 phosphorylation, we used qRT-PCR to screen for TGFβ/activin mRNA transcripts. We determined that activin-βA is the only ligand significantly upregulated. Hence, we propose that one or both of the activin-βA-containing ligands (activin A, activin AB) participate as the primary driver(s) of SMAD2 phosphorylation during early stages of tail regeneration.

Previous studies on zebrafish fin regeneration have also reported that activin-βA mRNA is also the only TGFβ/activin ligand significantly upregulated in regenerate fin tissues (Jazwinska et al., 2007). The increase in activin-βA expression, combined with a conspicuous absence of TGFβ1 observed in our study, suggests that the type and levels of the TGFβ/activin isoforms expressed determines regeneration outcome. Furthermore, these findings point towards activin-mediated signaling as an important mediator of both scar-free wound healing and the establishment (or maintenance) of cell proliferation during multi-tissue regeneration. In support of this, studies in chick granulosa cells demonstrated that activin-A treatment cooperates with protein kinase A (PKA) activity to cause a rapid, exclusive induction of SMAD2 mRNA and protein (Schmierer et al., 2003). This results in a shift in the SMAD2/SMAD3 ratio within a responding cell that preferentially increases SMAD2 phosphorylation and SMAD2 mediated transcription (Schmierer et al., 2003). In support of these data, several studies have also demonstrated that SMAD2 and SMAD3 target a subset of separate genes resulting in different transcriptional and cellular effects (Kretschmer et al., 2003; Dzwonek et al., 2009; Petersen et al., 2010). Further, SMAD2 knockout mice are embryonic lethal whereas SMAD3 knockout mice are viable and, following wounding, show reduced ECM deposition, with fewer fibroblasts and a smaller wound area when compared to control mice (Ashcroft et al., 1999). Taken together, these findings suggest that by selectively activating SMAD2, potentially through activin signaling, fibrotic responses can be attenuated. Our work indicates that TGFβ/activin isoforms play unique and dynamic roles during scar-free wound healing. Further investigation of TGFβ/activin signaling during tissue regeneration, particularly on the function of activin in modulating the SMAD2/3 balance and subsequent transcriptional/cellular outcome, is necessary to establish the role of activin in scar-free wound healing.

EMT Transcription Factors Are Upregulated During Tail Regeneration

EMT is a fundamental cellular mechanism during embryonic development, involving the transition of stationary, polarized epithelial cells to a motile, mesenchymal-like phenotype (Kalluri and Weinberg, 2009). Known to be induced by SMAD-mediated TGFβ/activin signaling (Kalluri and Weinberg, 2009), EMT is driven by specific zinc finger and basic helix loop helix (bHLH) transcription factors such as Snail1, Snail2, Zeb1, Zeb2, and Twist (Lim and Thiery, 2012). Accordingly, our observation that Snail1 and Snail2 are significantly upregulated during leopard gecko tail regeneration leads us to propose that an EMT process is likely occurring. Although the exact role of EMT during regeneration remains unclear, it has previously been shown that EMT transcription factors, including Snail1 and Snail2, can induce stem cell traits in transitioning cells (Mani et al., 2008; Guo et al., 2012; Lai et al. 2012; Soleimani et al., 2012). Furthermore, since EMT and EMT-like cellular plasticity is known to drive morphogenesis, organogenesis, and tissue remodeling during normal embryonic development (Kalluri and Weinberg, 2009), a recapitulation of these programs during multi-tissue regeneration is strongly indicated.

EXPERIMENTAL PROCEDURES

Animal Care and Handling

All animals were captive bred and obtained from a commercial supplier (Global Exotic Pets, Kitchener, Ontario, Canada). Leopard geckos were housed individually in standard rat enclosures in the Central Animal Facility at the University of Guelph, following husbandry protocols of McLean and Vickaryous (2011) and standard animal care procedures in accordance with the Canadian Council on Animal Care (CCAC) Guidelines (Animal utilization protocol numbers 09R026 and 10R113, approved by the University of Guelph Animal Care Committee).

Tail Collection and Tissue Segmentation

Tail collection was achieved by autotomy using the protocol of McLean and Vickaryous (2011). Briefly, leopard geckos were manually restrained and tails were firmly pinched at a position halfway between the cloaca and tail tip. Following pinching, the distal portion of the tail is self-detached, whereas the proximal portion is retained. Following autotomy, tails were allowed to regenerate to appropriate stages and autotomized a second time. Secondarily detached tails include the regenerated tissues plus a portion of original tail. Tails were morphologically staged according to the criteria of McLean and Vickaryous (2011). Outwardly, tails at stages I and II (wound healing; 0–8 days following tail loss) demonstrate no signs of regeneration and instead have prominent exudate clots capping the site of tail loss (clots in stage I are incomplete and/or relatively thin; at stage II the clot spans the entire surface and is relatively thick). Stages III and IV (blastema formation; 4–15 days following tail loss) are characterized by the loss of the clot and exposure of the newly formed wound epithelium and underlying blastema (at stage III) and expansion of the new tissue to form a dome-like outgrowth wider than long (length: diameter ratio less than 0.5; stage IV). Stages V to VI (tissue differentiation; 12–30 days following tail loss) involve continued outgrowth of the new tail tissue (length: diameter ratio greater than 0.5 but less than 1.0), scalation and histogenesis of skeletal, muscle, and connective tissues. By stage VII (25+ days following tail loss), the length:diameter ratio of the tail is greater than 1.0. In addition, the tail is pigmented and all the internal tissues have fully differentiated. For the purposes of the current study, we focused on stages III through VI.

Following the second autotomy, tails were segmented (by transverse cuts using a No. 15 scalpel blade) into three domains: regenerated tissue; a segment (∼1.0 cm long) of original tissue immediately adjacent to the regenerated tissue (“distal tissue”); and the next consecutive segment of original tissue more proximal to the body (“proximal tissue”). Each segment was immediately flash frozen in liquid nitrogen and stored at −80°C for protein and RNA extraction.

Western Blot Analysis

Protein was extracted from each segment of individual autotomized tails and homogenization in protein lysis buffer (Cell Signaling, Danvers, MA). For each segment, a 30 μg sample of total protein lysate was resolved on 10–15% polyacrylamide gels under reducing (pSMAD2, SMAD2, α-Tubulin) or non-reducing (TGFβ1) conditions and subsequently transferred to PVDF membranes (Roche, Indianapolis, IN). Membranes were first blocked for 1 hr in TBST containing 5% skim milk powder and then incubated with primary antibody diluted in blocking solution overnight at 4°C. The concentrations of primary antibodies were as follow: 1:1,000 rabbit monoclonal anti-pSMAD2 (Ser465/Ser467) (Cell Signaling), 1:1,000 mouse monoclonal anti-SMAD2 (Cell Signaling), 1:500,000 mouse monoclonal anti-tubulin (Sigma-Aldrich, St. Louis, MO), 1:1,000 rabbit polyclonal anti-TGFβ1 (Santa Cruz Biotechnology, Santa Cruz, CA). Membranes were then washed with TBST, incubated with the appropriate HRP-conjugated secondary antibodies for 1 hr at room temperature, washed again with TBST, and developed with HRP chemiluminescent substrates (Millipore, Billerica, MA) visualized on a ChemiDoc XRS (Bio-Rad, Hercules, CA). Three biological replicates were examined by Western blotting for each protein of interest.

Immunofluorescence and Immunohistochemistry

Fixation and sectioning were performed as previously described (McLean and Vickaryous, 2011). Briefly, tissues were fixed in either 10% neutral buffered formalin (NBF) or 4% paraformaldehyde (PFA) and serially sectioned. Slide-mounted serial sections were then deparaffinized and rehydrated. Sections destined for immunohistochemical analysis were quenched in 3% H2O2 for 30 min. All sections were rinsed with phosphate-buffered saline (PBS). Next, heat-induced epitope retrieval (citrate buffer heated to 90°C for 12 min) was used to unmask antigenic sites. Following antigen retrieval, sections were incubated with appropriate blocking reagent for 1 hr, followed by primary antibody overnight at 4°C. Primary antibody concentrations were 1:150 anti-pSmad2 (Ser 465/467) (Cell Signalling), 1:300 anti-PCNA (Dako, Carpinteria, CA), and 1:500 anti-TGFβ1 (Santa Cruz). Primary antibody was omitted from one section on each slide to function as a negative control. For immunohistochemistry, slides were rinsed in PBS and then all sections were incubated with appropriate biotinylated secondary antibody (Vector Laboratories, Burlingame, CA) at a concentration of 1:500. Slides were then incubated with peroxidase-streptavidin substrate (Vector Laboratories) for 1 hr at room temperature, and stained with 3-3′-diaminobenzidine for 6 min (DAB from Vector Laboratories). Slides were then rinsed in dH20, counterstained in Mayers Hematoxylin for 4 min, and mounted with coverslips. For immunofluorescence, secondary antibody concentrations and incubations were 1:300 Alexa-Fluor goat anti-rabbit 488 (Life Technologies, Carlsbad, CA), or 1:300 Cy3 Goat anti-mouse (Jackson ImmunoResearch Laboratories, West Grove, PA) for 1 hr at room temperature. Following the addition of fluorescent secondary antibodies, slides were incubated with 0.3 μM 4′,6′–diamidino-2-phenylindole (DAPI) (Sigma) diluted in PBS for 4 min, washed twice, and mounted with aqueous mounting media (Dako) and coverslipped.

Identification of Eublepharis macularius cDNA Sequences Using Degenerative PCR

RNA was isolated from segmented tails (stored at −80°C) by homogenization in Trizol (Invitrogen) at cold temperature followed by the Aurum total isolation RNA kit (Bio-Rad). cDNA synthesis was performed using the iScript cDNA synthesis kit (Bio-Rad) with 1 μg total RNA. Degenerate primers were designed by eye based on conserved sequences identified using ClustalW (EMBL-EBI) and publicly available sequences from related species (for primer sequences see Tables 1 and 2). PCR was performed on a CFX-96 thermocycler (Bio-Rad) using SYBR Green I Supermix (Applied Biosystems, Foster City, CA) across various temperatures. Amplicons were run on a 2% agarose gel and correct-sized bands were subsequently extracted using illustra GFX PCR Gel Band Purification Kit (GE Healthcare, Piscataway, NJ). Following extraction, amplicons were either sent for sequencing at the advanced analysis facility of the University of Guelph or cloned into a TOPO-TA cloning vector pGEM-T Easy (Promega, Madison, WI) and then sequenced. Sequence information was then analyzed using NCBI nucleotide blast and ClustalW (EMBL-EBI) to confirm sequence specificity (Table 3). Identified sequences were deposited to NCBI GenBank (numbers to be added once sequences annotated).

Table 1. Primers Used for qRT-PCR
PrimerForward primerReverse primer
TGFβ15-TCGAGTCATGACAGCTACGG-35-GGTACAGCAGGTCCGGATTA-3
TGFβ25-CACAATCCAGCGCCTATTTT-35-TGGCGAAGTTAGGTCTTTGC-3
TGFβ35-AACGCATTGAGCTTTTCCAG-35-CGCACGGTATCTGTGACATC-3
BMP25-CCCCTATCGTGAGATGGATT-35-GAGACCAGCTAGCGTCATCC-3
Activin βA5-CCTGGCACATCTTTCCAGTT-35-GCTGGCTCCAGTTTCTTGAC-3
Snail15-GCAAGAAGCCCAACTACAGC −35-AGCCCTGTGTCCCATACAAG −3
Snail25-ATGAGAGCTACCCAATGCCTG-35-CCAAAGATGAGGGATTATCCAGAGA-3
Zeb25-TCTCAGCCAGAGGAACAAGG-35-GCTCTGGAGACAAGCTTTGG-3
18s5-CGGCTACCACATCCTATGAA-35-TGGAGCTGGAATTACCGCGG-3
Table 2. Primers Uused for cDNA Identification
PrimerForward primerReverse primer
TGFβ15-AAGCGATCGAAGCCATC-35-CGYACRCARCAGTTCTYCTC-3
TGFβ25-ATGCGCAAGAGGATCGAGGCSAT-35-GGTTCRTGDATCCAYTTCCA-3
TGFβ35-GCBYTBTACAACAGCACC-35-CGYACRCARCAGTTCTYCTC-3
Activin βA5-AGGACAAAAGTGACCATCCG-35-GCTGGCTCCAGTTTCTTGAC-3
18s5-CGGCTACCACATCCTATGAA-35-TGGAGCTGGAATTACCGCGG-3
Table 3. cDNA Sequence Analysis
  cDNA sequence max identity % with 
GenecDNA (bp)Anolis carolinensisGallus gallusHomo SapiensGenBank accession
  1. a

    Activin-βA was identified and published during the production of this manuscript by another group.

  2. b

    Currently submitted to GenBank/NCBI awaiting accession numbers.

TGFβ1839717283b
TGFβ2881818278b
TGFβ3749858377b
Activin βAa574878376JN880655.1
18s142999999b

qRT-PCR Analysis

RNA was isolated and cDNA was produced as described above. Appropriate quality control steps were included along the way conforming to the MIQE guidelines (Bustin et al., 2009; Taylor et al., 2010). The cDNA that was obtained was diluted 1:4 in nuclease free water and 4 μl of diluted cDNA was added to reaction mixes (10 μl) containing 5 μl Sso Fast Eva Green Supermix (Bio-Rad) and 500 nM of each primer. Primers were designed using Primer3 (NCBI) and optimized using a temperature gradient and eight-point standard curve to determine PCR efficiency. Acceptable efficiency was deemed between 90 and 110%. All amplicons were determined to be specific by agarose gel analysis and subsequent sequencing. qRT-PCR amplifications were carried out using a CFX-96 (Bio-Rad) as follows: an initial denaturation step 2 min at 95°C, followed by 40 cycles of 5 sec at 95°C and 5 sec at 59°C. Data was expressed as relative gene expression normalized to 18s mRNA, which was determined to be a suitable housekeeping gene using qBase Plus software (BioGazelle, Ghent, Belgium). Kruskal-Wallis non-parametric statistics were performed on qRT-PCR data followed by a Dunns post test with a p value cutoff of p < 0.05. Three biological replicates and three technical replicates were used for all qRT-PCR experiments.

ACKNOWLEDGMENTS

The authors gratefully acknowledge Jodi Morrison, Dr. Brenda Coomber, and Emily Gilbert for extensive technical support and valuable feedback regarding the manuscript. This work is supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) (400419 to A.V.P and 400358 to M.K.V.) R.W.D.G is supported by an Ontario Veterinary College student fellowship.

Ancillary