Newt orthologue of Growth arrest-specific 6 (NvGas6) is implicated in stress response during newt forelimb regeneration

Authors


Abstract

Red-spotted newts are capable of regenerating various structures and organs through the process of epimorphic regeneration. Receptor tyrosine kinases (RTKs) and their ligands are important for normal cellular development and physiology but most have not yet been characterised during regeneration. We have isolated a newt orthologue of Growth arrest-specific 6 (NvGas6), and examined its expression during forelimb regeneration and within a blastema cell line (B1H1). During limb regeneration, NvGas6 expression increases upon amputation, peaks during maximal blastema cell proliferation, and is subsequently downregulated during redifferentiation. Transcripts are localised to the wound epithelium and distal mesenchymal cells during dedifferentiation and proliferative phases, and scattered within redifferentiating tissues during later stages. In B1H1 cultures, NvGas6 is upregulated under reduced serum conditions and myogenesis. Treatment with mimosine and colchicine or exposure to heat shock or anoxia results in upregulation of NvGas6 expression. Taken together, our findings suggest that during regeneration, NvGas6 expression may be upregulated in response to cellular stress. Developmental Dynamics 235:711–722, 2006. © 2006 Wiley-Liss, Inc.

INTRODUCTION

The red-spotted newt (Notophthalmus viridescens) has the ability to regenerate complex structures such as limb, lens, retina, and jaw, which have been lost through injury or amputation (Brockes, 1997; Tsonis, 2000; Brockes and Kumar, 2002; Nye et al., 2003). Forelimb regeneration begins with tissue injury signalling the rapid migration of epithelial cells to cover the wound surface and form a wound epithelium. Within several days, dedifferentiation, a process whereby differentiated cells lose their specialised cytoplasmic characteristics and revert to a multipotent state, begins in the distal mesodermal stump tissues. The dedifferentiated cells proliferate and converge to form a mesenchymal mound of cells called a blastema. Blastema cells proliferate until they reach a critical mass and then undergo redifferentiation and morphogenesis to reform the limb.

Receptor tyrosine kinase (RTK) signalling pathways govern diverse cellular functions and responses to extracellular cues. Members of the Axl family of RTKs (Axl, Sky, and Mer) are associated with proliferation, apoptosis, migration, and cellular adhesion (McCloskey et al., 1994; Goruppi et al., 1996; Li et al., 1996; Bellosta et al., 1997; Nakano et al., 1997a; Avanzi et al., 1998; Fridell et al., 1998; Scott et al., 2001). The common ligand for the Axl family members is the product of the Growth arrest-specific 6 (Gas6) gene (Godowski et al., 1995; Stitt et al., 1995; Varnum et al., 1995; Mark et al., 1996; Nagata et al., 1996; Chen et al., 1997). Gas6 is a secreted vitamin K–dependent ligand with structural similarity to Protein S of the coagulation cascade (Manfioletti et al., 1993). It was first identified as a gene upregulated during fibroblast serum starvation and cell cycle arrest (Schneider et al., 1988). Gas6 is composed of a γ-carboxyglutamic acid (GLA) domain followed by tandem repeats of epidermal-like growth factor (EGF) modules and 2 laminin G-like (LG) domains. It binds to the Axl family of RTKs (Axl, Sky, and Mer) with differing affinities (Nagata et al., 1996). Gas6 exerts its cellular mitogenic effects via downstream MAP, PI3, and STAT3 kinase pathways (Goruppi et al., 1996, 1997a, 1999, 2001; Li et al., 1996; Yanagita et al., 2001), and also acts synergistically with thrombin in vascular smooth muscle cells (Nakano et al., 1995, 1997b; Melaragno et al., 1998). Independent from its proliferative activities, Gas6 promotes cell survival mainly through the PI3 kinase pathway by prevention or suppression of apoptosis (Nakano et al., 1996; Bellosta et al., 1997; Goruppi et al., 1999; O'Donnell et al., 1999; Chan et al., 2000; Demarchi et al., 2001; Healy et al., 2001; D'Arcangelo et al., 2002). Gas6-Axl interaction is linked in vitro to migration, cell-cell and cell-matrix interactions via the Rac GTPase pathway (Bellosta et al., 1995; McCloskey et al., 1997; Avanzi et al., 1998; Fang et al., 1998; Fridell et al., 1998; Allen et al., 1999, 2002).

Many of the processes associated with Gas6-Axl activity, including cellular migration, apoptosis, stress responses, and cell cycle reentry, occur during dedifferentiation and blastema formation, suggesting an involvement of RTKs in adult newt forelimb regeneration. We report here the cloning and primary characterization of a newt orthologue of mammalian Gas6 (NvGas6). During regeneration, NvGas6 expression is associated with the origin and expansion of the proliferating blastema. In vitro, NvGas6 expression shows dramatic upregulation during myogenesis, in cell cycle arrest assays, and in heat shock and anoxia studies. These results suggest roles for NvGas6 as a stress-response factor.

RESULTS

Isolation of Full-Length NvGas6

RACE (rapid amplification of cDNA ends) was used to isolate the full-length NvGas6 cDNA. The BLASTX analysis and position of the highly conserved propeptide sequence place the putative MET translational start at 429 bp from the 5′ end (see GenBank DQ324381 for complete sequence). BLASTX searches revealed that NvGas6 is highly conserved, with 78% amino acid identity to mammalian Gas6. Multiple alignment with known Gas6 sequences reveals that the major domains are conserved, including key residues in the GLA propeptide sequence and cysteine sites within the EGF_CA modules (Fig. 1).

Figure 1.

Multiple alignment of Gas6 amino acid sequences of Notophthalmus viridescens (red-spotted newt; DQ324381), Xenopus laevis (frog; AAH76835), Danio rerio (zebrafish; NP_956272), Homo sapiens (human; NP_000811), Mus musculus (mouse; NP_062394), and Rattus norvegicus (Norway rat; NP_476441). Amino acids shaded in yellow indicate identity between sequences. High homology is observed between species in the GLA, EGF_CA and LG domains (marked by red, blue, and green boxes, respectively). Blue asterisks identify key conserved hydrophobic residues of the GLA propeptide. Red asterisks denote largely conserved GLU residues. Blue numbers indicate sites of cysteine residues conserved between species.

Nvgas6 Expression Analysis During Limb Regeneration

Real-time quantitative RT-PCR (RT-qPCR) was performed to generate NvGas6 expression profiles during forelimb regeneration. An evaluation of intact limbs, 1, 5, 7, 15, and 21 days post-amputation and palette stage regenerates demonstrated steady upregulation beginning within 24 hr post-amputation, and peaking at 21 days (Fig. 2a). During the wound healing and dedifferentiation stage of regeneration (1-, 5-, and 7-day samples), NvGas6 expression increased approximately 10-, 17-, and 25-fold, respectively, relative to the intact limb. Within the blastema stages of maximal proliferation (15- and 21-day samples), NvGas6 was upregulated 35- and 50-fold relative to intact limbs, respectively. In the palette stage regenerate, NvGas6 was downregulated relative to the 21-day regenerate, but was still high in relation to the intact limb (Fig. 2a). Normalisation to either Nvβ-Actin or NvGAPDH generated equivalent profiles (data not shown).

Figure 2.

NvGas6 real-time quantitative RT-PCR profile during newt forelimb regeneration and B1H1 cell culture conditions. a:NvGas6 is upregulated upon amputation, peaks at 21 days after amputation, and is downregulated at the palette stage. d, day. b: Expression increases from states of proliferation to myogenesis, and decreases upon serum resupplementation. S, subconfluent cells; C, confluent cells; M, 4-day-old myotubes; MC, 8-day-old myotubes; MR, 4-day-old myotubes resupplemented with serum for 4 days. Expression was normalised to NvGAPDH. Error bars denote ± standard deviation. Fold over-expression was calculated as per Livak and Schmittgen (2001).

In situ hybridization provided a spatial analysis of NvGas6 expression during forelimb regeneration. In the intact stump, staining was observed in the epidermal and dermal layers, and there was occasional weak staining in the intact muscles and dermal glands (data not shown). The wound healing stage involves the migration of epithelial cells over the open wound. Strong NvGas6 expression was observed in the wound epithelium, and in the dermal and epidermal layers of the stump (Fig. 3a). By 7 days after amputation, wound healing is complete and the cells are undergoing dedifferentiation. At this stage, NvGas6 expression was found in the dedifferentiating and accumulating blastema cells of the distal stump (Fig. 3b). At 15 days, the base of the regenerate is undergoing extensive dedifferentiation, and the distal tip is comprised of rapidly proliferating blastema cells (Iten and Bryant, 1973). NvGas6 was strongly expressed in a peppering of tissues distal to the bone that included both dedifferentiating and blastema cells (Fig. 3c). The 21-day regenerate is characterised by maximal blastema cell proliferation and a well-defined blastema. The staining in the blastema was very intense and represented a mixture of highly stained and weakly stained cells (Fig. 3d). There was less apparent staining in areas lateral to the bone relative to early stage regenerates. At the palette stage, blastema cell proliferation is rapidly decreasing, and cells are beginning to differentiate into the structures of the complete regenerate. NvGas6 staining was localised to distal mesenchymal cells, differentiating myotubes, and condensing chondrocytes (Fig. 3e).

Figure 3.

NvGas6 in situ hybridisation analysis during newt forelimb regeneration. a: Staining in the 24-hr regenerate. Extensive staining is seen in the wound epithelium (inset). Some non-specific staining is seen in the intact muscle of the stump (based on observations in sense controls). b: Expression in the 7-day regenerate is found in the epithelial layer and the mesenchymal cells of the emerging blastema (inset). c: Staining in the 15-day regenerate is located throughout the blastema (inset), in the dedifferentiating cells at the base of the regenerate, and in the periosteum. d: Staining in the 21-day regenerate, which displays high NvGas6 expression in the blastema (inset). e: Staining at the palette stage regenerate, with NvGas6 transcripts located in condensing cartilage (C), differentiating myotubes (M), and distal blastema cells. f: Sense control.

NvGas6 Expression in B1H1 Cells During Myogenesis

Cells of the newt B1H1 blastema line can be induced to differentiate into myotubes under low serum conditions. Serum resupplementation of myotubes leads to cell cycle re-entry, progression through S phase, and G2 arrest. Extensive characterization of the cellular composition and myogenesis within this polyclonal line has previously been performed (Ferretti and Brockes, 1988; Corcoran and Ferretti, 1997; Vascotto et al., 2005). Subconfluent cultures (<70% confluent) and confluent cultures (>90% confluent) are comprised largely of pleiomorphic cells. Real-time analysis demonstrated that NvGAPDH-normalised NvGas6 expression did not change significantly with cell density (Fig. 2b). Four- and 8-day-old myogenic cultures are comprised of approximately 15 and 25% myotubes, with the remainder primarily pleiomorphic cells. NvGas6 expression increased approximately twofold with the reduction of serum in B1H1 culture. It was subsequently downregulated to near basal levels with cell cycle re-entry induced by the restoration of full serum medium. These serum-resupplemented cultures represent 4-day myotube cultures exposed to growth medium for an additional 4 days, and contain fewer myotubes and a greater proportion of pleiomorphic cells.

NvGas6 Expression in B1H1 Cells Treated With Mimosine and Colchicine

Mimosine and colchicine are potent cell-cycle arrest agents that arrest at the G1 and metaphase stages of mitosis, respectively (Wilson and Friedkin, 1966; Lalande, 1990). Treatment of subconfluent B1H1 cells with either agent resulted in almost 100% cell cycle arrest (Fig. 4a–c). Although differences in cell numbers were not evident at 24 hr after treatment (T-test, P = 0.131, df = 4), by 48 hr there was a significantly lower number of B1H1 cells in the colchicine-treated cultures versus the untreated control (T-test, P = 0.017, df = 4). In addition, B1H1 cultures treated with colchicine had many floating cells that were not seen in either the untreated or mimosine treated cultures (data not shown), and had reductions in the adherent cell population, suggesting that colchicine treatment induces cell death. By 48 hr, approximately 45% of colchicine-treated B1H1 nuclei were in later stages of mitosis, whereas this number was less than 5% in untreated controls and following mimosine treatment (data not shown). The effect of mimosine and colchicine treatment on NvGas6 expression was analysed by RT-qPCR (Fig. 4d). NvGas6 expression increased approximately 10- and 2.5-fold in mimosine- and colchicine-treated B1H1 cells, respectively, relative to untreated controls at 48 hr.

Figure 4.

Effects of mimosine and colchicine treatment on B1H1 cells. BrdU incorporation (identified by arrows) and DAPI staining at 48 hr after treatment in untreated (a), 400 nM mimosine (b), and 20 μg/mL colchicine (c) treated B1H1 cells. Cells were labelled with BrdU for 18 hr in the presence of inhibitors. d: Real-time RT-qPCR expression of NvGas6 in B1H1 cells treated with inhibitors for 48 hr shows higher upregulation in mimosine- versus colchicine-treated cells. Expression was normalised to Nvβ-Actin.

Migration of B1H1 Cells Cultured in Mimosine and Colchicine

To examine whether cell migration in growth-arrested cells could be correlated with NvGas6 expression, “wounds” were scraped into a confluent monolayer after 24 hr in the presence of mimosine or colchicine, and gap closure was measured (Fig. 5). There was no significant difference in the size of the initial gaps generated by the scraping (ANOVA, P = 0.126, df = 2, 6, see Fig. 5). By 48 hr, the relative gap closure was highest in untreated B1H1 cells (Fig. 5). Although gap closure in mimosine-treated cells was slower than untreated, by 48 hr mimosine-treated wounds were not significantly different from controls (T-test, P = 0.131, df = 4). Migration in colchicine-treated cultures was significantly inhibited (T-test, P = 0.017, df = 4), resulting in only 5% gap closure in relation to the untreated control.

Figure 5.

Migration assay of mimosine- and colchicine-treated B1H1 cells showing percent gap closure of untreated-, mimosine-, and colchicine-treated B1H1 cells. The initial sizes of the wounds that were scraped into the monolayer were 433.3 ± 34.6 μm for the untreated cells, 361.6 ± 26.7 μm for colchicine-treated cells, and 420 ± 49.8 μm for mimosine-treated cells. There was no statistical difference in the initial gaps generated by the “wound” (ANOVA, P = 0.126, df = 2,6). By 48 hr, gap closure in mimosine-treated cells is slower, but not significantly different from controls while colchicine-treated cells show significantly reduced migration capabilities (ANOVA, P = 0.002, df = 2,6; Tukey test, P < 0.05).

Expression of NvGas6 in B1H1 Cells Under Heat Shock and Anoxic Conditions

The effects of stress on NvGas6 expression were evaluated by RT-qPCR of cultures under heat shock and anoxic conditions. B1H1 cells are normally maintained at 28°C. Exposure of B1H1 cells to 37°C for 20 min followed by a 2-hr recovery period resulted in a 15-fold increase in NvGas6 expression (Fig. 6). Heat-shocked B1H1 cultures demonstrated apparent normal cell morphologies and did not contain detached cells (data not shown). Cultures maintained under anoxic conditions for 48 hr contained floating cells, potentially representing necrotic or apoptotic bodies (data not shown). The adherent cell population demonstrated a 5-fold increase in NvGas6 expression relative to untreated controls (Fig. 6).

Figure 6.

Real-time quantitative RT-PCR profile of NvGas6 in B1H1 cells during anoxic and heat shock conditions compared to untreated B1H1 cells. B1H1 cells were subjected to 0% O2 for 48 hr and heat-shocked B1H1 cells were exposed to 37°C for 20 min and allowed to recover at 28°C for 2 hr. Expression was normalised to Nvβ-Actin and fold overexpression was calculated as per Livak and Schmittgen (2001).

DISCUSSION

We report the cloning and characterization of the first Gas6 gene identified in the lower tetrapods. NvGas6 is a 2.9-kb single copy gene encoding for one GLA domain, 4 EGF_CA domains, and 2 LG domains. The highly conserved GLA module contains the propeptide sequence that includes a well-conserved phenylalanine site present in all chordate GLA domains and several key hydrophobic residues (Bandyopadhyay et al., 2002; Wang et al., 2003). The propeptide sequence is believed to anchor the substrate to γ-carboxylase, the enzyme responsible for GLU to GLA modification (Furie et al., 1999; Berkner, 2000). Following this domain are 4 tandem EGF_CA modules that are involved in Ca2+-dependent protein–protein interactions (Rao et al., 1995; Stenflo et al., 2000). Within these domains, NvGas6 shows conservation of the 5–6 core cysteines.

The carboxy-terminal region of NvGas6 contains 2 LG domains that are required for RTK binding and proper conformation (Mark et al., 1996; Goruppi et al., 1997b; Tanabe et al., 1997; Sasaki et al., 2002). The first LG domain is highly conserved compared to mammalian Gas6, likely due to its importance in receptor binding (Sasaki et al., 2002). The second domain is putatively required for proper folding and/or receptor interaction (Sasaki et al., 2002). The relative affinity of Gas6 for the specific receptor within the RTK subclass is apparently dictated by this region and varies across species. Thus, while NvGas6 demonstrates great similarity to mammalian Gas6, the receptor subclass that it binds and the pathway that it stimulates remain to be determined and will ultimately dictate its specific function.

NvGas6 Is Expressed Throughout Limb Regeneration and Peaks During the Proliferative Phase

Given our current knowledge and technology, in vivo functional data is difficult to obtain in the newt. Therefore, in order to identify candidate roles for NvGas6 during regeneration, its spatiotemporal expression was analysed during adult newt forelimb regeneration. NvGas6 is moderately upregulated during the wound healing stage, gradually increases during the dedifferentiation stage, and is highly upregulated during the proliferative phase of regeneration. During the redifferentiation stage, the level of NvGas6 expression is reduced but is still high relative to the intact limb.

A multitude of functions have been ascribed to Gas6 in the literature, depending on cell type, context, and environmental conditions. Gas6 binds platelet RTKs and promotes platelet aggregation and secretion (Angelillo-Scherrer et al., 2001). Gas6 and its RTKs are involved in phagocytosis and apoptotic clearance of damaged or dying cells (Nakano et al., 1997a; Ishimoto et al., 2000) and serve as important regulators of macrophage function (Lemke and Lu, 2003). Gas6 has been shown to have anti-apoptotic properties during vascular injury (Melaragno et al., 1998; Yin et al., 2000) or nutrient deprivation (Fleming et al., 1998; Gonos, 1998; Goruppi et al., 1999). Gas6 has further been shown to have roles in proliferation (Goruppi et al., 1996; Loeser et al., 1997), cellular adhesion (Bellosta et al., 1995; McCloskey et al., 1997; Wimmel et al., 2001), and cellular migration (Fang et al., 1998; Fridell et al., 1998; Allen et al., 2002).

The regenerating limb is a heterogeneous dynamic structure with multiple cell and tissue types. Wound healing begins with the formation of a thrombus and the lateral migration of epithelial cells to cover the wound stump (Iten and Bryant, 1973). This is followed by massive apoptosis at the wound-stump interface (Mescher et al., 2000; Vlaskalin et al., 2004), and invasion by macrophages to clear the debris. There is upregulation of genes associated with ECM remodelling, including thrombin, matrix metalloproteinases, and proteases (Miyazaki et al., 1996; Mescher et al., 2000; Kato et al., 2003; Vlaskalin et al., 2004). As dedifferentiation progresses, there is proliferation, accumulation, and aggregation of the mesenchymal cells to generate the blastema. Blastema formation involves proliferation, cellular migration (Hay and Fischman, 1961; Gardiner et al., 1986; Nace and Tassava, 1995), and changes in cellular adhesion (Maier et al., 1986).

Identifying a specific function for NvGas6 on such a complex background is difficult. If we assume that NvGas6 may have multiple roles in the regenerating limb depending on the tissue type, then the expression and spatiotemporal results of the current study do not eliminate any of the potential functions assigned to Gas6 in the literature. However, if we consider a single role for NvGas6 during the regeneration process, then the maximal expression at 21 days post-amputation, when blastema cell proliferation is greatest, eliminates potential roles in platelet aggregation, macrophage regulation, or apoptotic cell clearance. Potential NvGas6 roles in proliferation, migration, and cell survival are discussed below in the context of the in vitro studies.

NvGas6 Is Upregulated in Blastema Cells Under Reduced Serum Concentrations and Following Treatment With Cell Cycle Inhibitors

In order to further narrow down the possible roles for NvGas6 in forelimb regeneration, we conducted in vitro assays with the B1H1 cell line. The cell line was originally isolated from a forelimb blastema, and similar to other blastema-derived cell lines, it retains many of the characteristics of the in vivo blastema (Ferretti and Brockes, 1988; Maier and Miller, 1992). Cells of the B1H1 line withdraw from the cell cycle under reduced serum conditions in vitro, arrest at G1/G0, and undergo myogenesis (Ferretti and Brockes, 1988; Corcoran and Ferretti, 1999; Vascotto et al., 2005). Restoration of serum to the culture medium allows newt myotube nuclei to traverse the S phase and arrest in G2 (Tanaka et al., 1997, 1999). In mammalian systems, Gas6 was first identified as a factor upregulated with cell cycle arrest under conditions of serum and nutrient deprivation in fibroblasts (Schneider et al., 1988). It is perhaps not surprising then that real-time RT-qPCR analysis demonstrated a similar upregulation of NvGas6 under reduced serum conditions and myogenesis in B1H1 cells, and this expression was shown to decrease with serum resupplementation. These results show that NvGas6 expression does not correlate with proliferation of the B1H1 cells, since it was upregulated as cells exited the cell cycle and differentiated.

It is noteworthy that NvGas6 expression is maximal during the height of blastema cell proliferation stages in vivo, and yet is associated with the exiting of B1H1 blastema cells from the cell cycle in vitro. This apparent paradox can be explained if, both in vivo and in vitro, NvGas6 upregulation occurs as a response to cellular stress. In other systems, cellular proliferation is often associated with increased expression of survival and anti-apoptotic factors (Nickoloff et al., 2002; Borgne and Golsteyn, 2003; Calo et al., 2003; Gaur and Aggarwal, 2003; Klein et al., 2003).

It is somewhat difficult to determine whether the in vitro changes in expression of NvGas6 were associated with serum deprivation (and cellular stress) or whether NvGas6 might have been involved in regulating the cell migration and adhesion associated with cellular alignment and fusion during myogenesis. The differentiation assay would suggest a combination of both factors, since NvGas6 was slightly downregulated in the advanced 8-day-old myotube cultures, which remained serum-starved, but in which fusion events were presumably decreasing.

To evaluate whether NvGas6 expression under myogenic conditions was associated with arrest at a particular stage in the cell cycle, B1H1 cultures were treated with the cell cycle inhibitors colchicine and mimosine. Treatment with colchicine results in cell arrest at the prometaphase/metaphase junction and inhibition of cell migration by interfering with microtubule organization (Wilson and Friedkin, 1966; Andreu and Timasheff, 1982). Mimosine reversibly arrests cell cycle division in the G1 phase by interfering with deoxyribonucleotide metabolism and cyclin D1 and CDKI production (Mosca et al., 1992; Gilbert et al., 1995; Wang et al., 1995; Chang et al., 1999). In mammalian cells, mimosine has a marginal affect on migration while colchicine treatment has been demonstrated to completely halt neutrophil motility (Kubens et al., 2001; Niggli, 2003). Colchicine treatment of B1H1 cells resulted in large numbers of dead, floating cells, and caused 2.5-fold upregulation in NvGas6 expression. Mimosine treatment of B1H1 cells inhibited proliferation and resulted in 13-fold upregulation of NvGas6 expression, but there were very few dead cells in the cultures. The results with the myogenesis studies and the cell cycle inhibitors might be consistent with upregulation of NvGas6 during the G0 or G1 phase of the cell cycle. However, if this were the case, we would have expected a smaller percentage of blastema cells in the regenerates to express NvGas6, given that the G1 phase of the cell cycle only accounts for approximately 1/20th of the total cell cycle (Grillo, 1971).

To evaluate whether NvGas6 upregulation during myogenesis was related to cellular migration, gap closure was evaluated in control and in mimosine- and colchicine-treated B1H1 cells. Although mimosine-treated cells migrated slower than untreated cells, they succeeded in closing of the gap, whereas colchicine-treated cells did not migrate. Presumably, if the untreated cells bridged the gap by both proliferation and migration phenomena, and NvGas6 expression was only related to cellular migration, we would have expected control B1H1 cells to express high levels of NvGas6 (comparable to mimosine treatment). NvGas6 expression was higher in mimosine-treated cells and in colchicine-treated cells in comparison to untreated controls. Thus, although we cannot rule out the possibility that migration events may induce NvGas6 expression, it appears that stress-related expression predominates in treated B1H1 cells.

NvGas6 Is Upregulated in B1H1 Cells by Heat Shock and Anoxia

To further evaluate whether NvGas6 expression in arrested or myogenic cells could represent a stress response, B1H1 cells were subjected to heat shock and anoxia conditions. Heat shock causes cellular stress that results in altered gene expression and production of heat shock proteins and can ultimately lead to cell death (Lindquist and Craig, 1988). In newt cells, heat shock at 37°C has been demonstrated to lead to upregulation of NvHsp70 (Billoud et al., 1993; Prudhomme et al., 1997). Heat-shocked B1H1 cells demonstrated 15-fold upregulation in NvGas6 expression compared to normal proliferating B1H1 cells. NvGas6 may be actively transcribed to serve a prosurvival effect under heat shock conditions, as has been demonstrated in mammalian models via the P13K pathway (Nakano et al., 1996; Bellosta et al., 1997; Goruppi et al., 1999; O'Donnell et al., 1999; Chan et al., 2000; Demarchi et al., 2001; Healy et al., 2001; D'Arcangelo et al., 2002).

Similarly, hypoxia, in vitro, leads to cell death and activation of both pro-survival and pro-apoptotic factors (Saikumar et al., 1998; Banasiak et al., 2000). Survival is mediated through the activation of PI3K and MAPK pathways (Alvarez-Tejado et al., 2001; Kayyali et al., 2002; Zhang et al., 2003). Under conditions of anoxia, NvGas6 expression was upregulated 5-fold relative to proliferating subconfluent B1H1 cultures, and was accompanied by the presence of many detached, floating, dead cells in the cultures. Interestingly, colchicine treatment and anoxia, both of which induced massive apoptosis of the B1H1 cells, were associated with a more moderate upregulation of NvGas6 than mimosine or heat shock. The latter two cause less severe and potentially reversible damage. It is possible that NvGas6 is dramatically upregulated to exert its prosurvival role when the cell receives stress signals which are not fatal, and that pro-apoptotic signals predominate when the damage/insult is too great.

CONCLUSIONS

During forelimb regeneration, expression patterns of NvGas6 show a steady rise from the initial injury up to 21 days after amputation and a gradual decline thereafter. Stress assays involving colchicine, mimosine, heat shock, and hypoxia show upregulation of NvGas6 in response to the treatments. Taken altogether, the in vivo expression pattern identified for NvGas6 during regeneration, the in vitro analysis of NvGas6 expression in blastemal B1H1 cells, and published accounts of the destabilized cellular environment generated during regeneration all suggest that NvGas6 expression, during limb regeneration is most likely associated with cellular stress. We cannot, however, discount multiple roles for NvGas6 in different cell types in the regenerating forelimb. Without cloning, characterization, and tissue distribution of the RTKs that bind to NvGas6, and functional studies involving NvGas6 misexpression, a role in proliferation, migration, and cell adhesion cannot be completely discounted, since all these processes are occurring concurrently with NvGas6 upregulation in the regenerating forelimb.

EXPERIMENTAL PROCEDURES

Animal Care and Tissue Sampling

Adult N. viridescens were purchased from Charles D. Sullivan and maintained according to Cameron et al. (2004). Newts were anaesthetised with 0.1% tricaine methanesulphonate (MS222) buffered to pH 7.0 with NaHCO3. Initial amputations were made proximal to the elbow and the tissues were trimmed to yield a flat amputation plane. Intact limbs, and 1-, 5-, 7-, 15-, 21-day and palette-stage regenerating limbs were collected by amputating at the mid-stylopodium and were either immediately flash frozen in Trizol (Invitrogen, La Jolla, CA) and stored at −70°C for RNA extraction or immersed overnight at 4°C in 4% paraformaldehyde for in situ hybridization analysis. Fixed samples were equilibrated in 30% sucrose overnight, and embedded in a 1:1 OCT (Tissue-Tek) and 30% sucrose solution. Samples were sectioned at 16 μm on a Shandon cryostat, mounted onto SuperFrost Plus Slides (Fisher Scientific), and stored at −70°C.

RACE

A 321-bp partial cDNA was identified in a screen of a blastema lambda gt10 phage library using the EGF repeat region of Xenopus Notch (Xotch). Gas6-specific oligonucleotides were synthesized at the 5′ and 3′ end of this sequence (see Table 1). Total RNA was extracted from confluent B1H1 cells using Trizol (Invitrogen), and enriched for poly(A)+ RNA using oligo(dT) cellulose beads (Amersham). A cDNA library was generated using the Marathon cDNA Amplification kit (Clontech, Palo Alto, CA), adapters ligated to the 5′ and 3′ ends of the cDNAs and RACE conducted. PCR products were separated by agarose gel electrophoresis, extracted using the QIAquick Gel Extraction kit (Qiagen, Chatsworth, CA), and subcloned into a pGEM-T Easy plasmid vector (Promega, Madison, WI). Clones were sequenced on an ABI 370 sequencer.

Table 1. Primer Sets for PCR
NvGas6 primers 
RACE 
 Sense5′-CGCGCTGCTCTTTGTCCTGCTA-3′
 Antisense5′-GGTGGCCCTGCTTGGTCTCCT-3′
 Nested sense5′-CGCAGACCCCGCGCAAGGCACTA-3′
 Nested antisense5′-CAATAGTGCCTTGCGCGGGGTCTGC-3′
RT-qPCR 
 Sense5′-TGCCCCTGTCACTATCGCTCATCG-3′
 Antisense5′-TGCCCCGCTATACCATCCACAGTC-3′
Nvβ-Actin 
 Sense5′-CCACTGCTGCTTCTTCATCCTCTC-3′
 Antisense5′-GGGCACCTGAACCGCTCATTG-3′
NvGAPDH 
 Sense5′-TGTGGCGTGACGGCAGAGGTG-3′
 Antisense5′-TCCAAGCGGCAGGTCAGGTCAAC-3′

B1H1 Cell Culture

Differentiation assays.

A newt blastemal cell cline, B1H1, was kindly provided by J. Brockes (University College London, London, UK) and maintained according to Ferretti and Brockes (1988). Myogenesis was induced in cells that were 80 to 90% confluent by reducing the serum in the medium from 10 to 0.5% FBS (Tanaka et al., 1997). Cells were sampled after 4 or 8 days for short- and long-term differentiation cultures, respectively. For serum resupplemented samples, cells were kept for 4 days in low serum, and then the medium was replaced with the normal 10% FBS medium for a further 4 days before sampling (Tanaka et al., 1997).

Cell-cycle inhibitors.

B1H1 cells were grown to 80–90% confluency and treated with either 400 nM mimosine (Wang et al., 1995) or 20 μg/mL colchicine (Albert et al., 1987) for 48 hr. Representative samples were either harvested for RNA extraction or counted at 0, 24, and 48 hr after addition of the inhibitor. Cells were fixed with 70% ethanol/30% 50 mM glycine, stained with 30 nM DAPI, and the nuclei of at least 3 random fields of view were counted per plate. Cells undergoing cell division or DNA synthesis were identified morphologically (condensed, tightly packed chromatin at various stages of alignment at the metaphase/anaphase state) or by BrdU incorporation. For BrdU studies, B1H1 cells were labelled with 10 μM BrdU for 18 hr, fixed, immunostained with 1:20 anti-BrdU (Molecular Probes, Eugene, OR) and 10 μg/mL anti-mouse Alexa Fluor 488 (Molecular Probes), and mounted in 1% n-propylgallate in 50% 1× PBS/50% glycerol.

Cell migration.

B1H1 cells were maintained in normal culture medium or with inhibitors for 24 hr. Parallel transects were generated with a flattened pipette tip, and gap distance was measured at 0, 24, and 48 hr post-scrape using a Zeiss inverted microscope and Axiovision software.

Heat shock and anoxia.

Subconfluent B1H1 cells were treated at 37°C for 20 min and allowed to recover at 28°C (Prudhomme et al., 1997). After 2 hr of recovery, adherent cells were harvested for RNA extraction. For anoxia treatment, subconfluent B1H1 cells were exposed to 0% O2 (MACS-VA500, du Scientific) at 28°C for 48 hr and adherent cells harvested for RNA extraction.

Real-Time RT-qPCR

NvGas6 specific sense and antisense oligonucleotides were generated using Primerselect software (DNAstar). SYBR Green I chemistry-based RT-qPCR was carried out on an iCycler iQ real-time thermocycler (Bio-Rad, Richmond, CA) with 1:1, 1:2, 1:10, 1:20, 1:100, 1:200, and 1:500 dilution of cDNA to determine primer efficiency. Cycling conditions were 95°C for 3 min, and 40 cycles of 94°C for 25 sec, and 63°C for 25 sec. Following a 10-min extension at 72°C, a dissociation curve analysis of 95°C for 60 sec, 55°C for 60 sec, and 80 cycles of 0.5°C increments every 10 sec was performed. An efficiency plot of ΔCT and log cDNA was constructed, with slopes determined to be within the acceptable range. The generation and validation of Nvβ-Actin and NvGAPDH primers have been described elsewhere (Vascotto et al., 2005).

Total RNA was extracted from B1H1 cultures or from flash frozen regenerate tissues using Trizol (Invitrogen). cDNA was generated from DNaseI-treated 2 μg total RNA primed with 500 ng of both oligo(dT) and random hexamers (Promega) in a RT reaction using M-MLV (Invitrogen). RT-qPCR was conducted using a 1:60 dilution of first strand as template. Each reaction was performed in triplicate and 5 μL of each replicate was visualised using agarose gel electrophoresis. Replicates were sampled at the presumptive CT and the expression profiles obtained on the thermocycler were confirmed by gel electrophoresis. ΔΔCT values were calculated relative to subconfluent B1H1 cells for the B1H1 expression profile and relative to intact limbs for the regeneration studies. RT-qPCR was performed at least three times with independent RNA samples.

In Situ Hybridization

A NvGas6 cDNA fragment corresponding to the region downstream of the EGF_CA domain, comprising the loop region and first LG domain, was chosen because it was determined to be a single copy region by Southern blotting (data not shown) and has been used previously for mammalian analyses (Prieto et al., 1999). The cDNA fragment was cloned into pGEM-T Easy (Promega), and digested with SpeI and NcoI to generate the antisense and sense probes for NvGas6, respectively. Digests were purified and 11-digoxygenin dUTP labelled RNA probes (Roche) were generated. For in situ hybridization, slides were hybridised overnight at 65°C in a solution containing a 1:250 dilution of the denatured probe, 1× salt (195 mM NaCl, 8.9 mM Tris HCl, pH 7.5, 1.1 mM Tris base, 5 mM NaH2PO4, 5 mM Na2HPO4, 5 mM EDTA), 50% deionised formamide, 10% dextran sulphate, 1 mg/mL yeast tRNA, and 1× Denhardt's. Slides were washed in 1× SSC, 50% formamide, and 0.1% Tween-20 at 65°C, equilibrated in 1× MABT (0.1 M maleic acid, 0.15 M NaCl, 0.1% Tween-20), and blocked with a solution of 20% heat-inactivated sheep serum (Sigma), 2% blocking reagent (Roche), and 1× MABT. Cells were incubated overnight at 4°C with an alkaline phosphatase-conjugated anti-DIG antibody (Roche) diluted 1:1,500 in the blocking solution, and washed with 1× MABT. Signal was resolved in 10% polyvinyl alcohol, 100 mM NaCl, 100 mM Tris-HCl, pH 9.3, 50 mM MgCl2, 0.1% Tween-20, 0.45 mg/mL NBT (4-nitroblue tetrazolium chloride; Roche), and 0.175 mg/mL BCIP (x-phosphate/5-bromo-4-chloro-3-indolyl-phosphate; Roche). Slides were counterstained with Eosin Y and mounted with Permount.

Acknowledgements

We thank Jeremy Brockes and Patrizia Ferretti (University College London, London, UK) for providing the B1H1 cell line and for helpful discussions on its maintenance, and Adam Baker for technical assistance.

Ancillary